While many autistic children show signs of autism from birth, some seem to develop normally but then regress after their first year.
This suggests a trigger in the environment. Scientist now think it may well be both: something in the environment may trigger the disease in children who are genetically susceptible.
Paul Shattock, a British biochemist, has found evidence that digestive problems in a child may set off autism
(New Scientist, 20 June 1998, p 42).
He proposes that some children can't break down cow's milk and wheat into amino acids, leaving "rogue" peptides in their systems that wreak havoc by mimicking the brain's opiates and neurotransmitters.
In Andrew Wakefield's controversial version of this theory the MMR vaccine causes inflammation of the gut, making it "leaky" and allowing these rogue peptides into the body.
But while he claims that the measles virus from MMR can be found in inflamed gut tissue, other studies using more sensitive methods have failed to find the virus.
A Japanese team has evidence that Wakefield's methods are actually detecting a human protein that resembles a measles protein.
Another theory is that autism is an autoimmune disease, in which the body starts attacking a key substance in the brain.
Vijendra Singh, a neuroimmunologist at Utah State University in Logan, has found high levels of "auto-antibodies" against myelin basic protein in autistic children.
Myelin is the fatty sheath that protects and insulates nerve fibres in the brain and is installed at record rate during early childhood brain development.
Singh doesn't stop there. He thinks measles triggers the autoimmune reaction. He says that autistic children tend to have a very strong reaction to the measles virus, producing seven times more antibodies than normal.
In a study that has been submitted for publication, he claims that 53 percent of 80 autistic children but none of 60 other children had abnormal antibodies to the measles component of MMR.
"What we're finding is that many autistic children have an inappropriate response to MMR," he says.
Imposible to know which antigens
Singh thinks that an atypical measles infection, perhaps in children with a particular genetic vulnerability, might lead to autism. He believes it might be possible to identify children who shouldn't be given the vaccine.
Elizabeth Miller of the Public Health Laboratory Service in London says it's impossible to know which antigens in the vaccine supposedly trigger production of antibodies in patients.
She points out that an exhaustive study by Phil
Minor of the US National Institute for Biological Standards and Control has shown that there's nothing resembling myelin basic protein in the MMR vaccine.
"A dispassionate look at the published scientific evidence pionts to the overwhelming safety of the vaccine and no link to autism," she says.
Even Singh isn't suggesting that vaccination be stopped. "It's the best preventative medicine we have today," he said. -New Scientist