The surplus synapses – places where neurons connect and communicate – are due to a lack of “pruning” that normally occurs early in life.
In mice with autistic traits, scientists were able to restore synaptic pruning and reduce symptoms with a drug used to suppress the immune systems of transplant patients.
The drug, rapamycin, has side effects that make it unsuitable as an autism treatment. But the discovery opens up exciting possibilities for other therapies based on synaptic pruning.
Excessive synapses could be a fundamental causal factor behind autism, the scientists believe.
“This is an important finding that could lead to a novel and much-needed therapeutic strategy for autism,” said Professor Jeffrey Lieberman, chair of psychiatry at Columbia University Medical Center in New York, where the research took place.
Autism, which affects around 500,000 people in the UK, covers a range of behavioural disorders that reduce the ability of sufferers to communicate with and relate to other people.
It is believed to be triggered by a combination of genetic and environmental factors that impact on the developing brain.
During normal brain development, a burst of synapse formation occurs in infancy, especially in the cortex – a region closely linked to autistic behaviour.
Pruning removes more than half of these cortical synapses by late adolescence.
Synapses are known to be affected by many genes linked to autism, leading to speculation about the role they play in the condition.
The Columbia University scientists examined the brains of 26 autistic children and young people aged two to 20 who had died from a variety of causes.
Brains from 22 non-autistic children were used as a comparison.
Synaptic density in samples of brain tissue was measured by counting the numbers of tiny branching spines that make up neural connections.
By late childhood, spine density had dropped by about half in the “healthy” brains, but by only 16% in the brains of autistic individuals.
“It’s the first time that anyone has looked for, and seen, a lack of pruning during development of children with autism, although lower numbers of synapses in some brain areas have been detected in brains from older patients and in mice with autistic-like behaviours,” said lead researcher Professor David Sulzer.
Synaptic pruning is driven by a cell-degrading process known as autophagy, from the Greek for self-eating.
In laboratory mice, the pruning problem was traced to a protein called mTOR which when over-active suppressed the “self-eating” ability of brain cells.
Large amounts of the over-active protein were also present in the brains of autism sufferers, the scientists reported in the journal Neuron.
Rapamycin inhibited mTOR and restored normal autophagy and synaptic pruning in the mice, even at a late stage when the animals were already displaying signs of autism.
Prof Sulzer said: “The fact that we can see changes in behaviour suggests that autism may still be treatable after a child is diagnosed, if we can find a better drug.”
He added: “W hile people usually think of learning as requiring formation of new synapses, the removal of inappropriate synapses may be just as important.
“What’s remarkable about the findings is that hundreds of genes have been linked to autism, but almost all of our human subjects had overactive mTOR and decreased autophagy, and all appear to have a lack of normal synaptic pruning.
“This says that many, perhaps the majority, of genes may converge onto this mTOR/autophagy pathway, the same way that many tributaries all lead into the Mississippi River. Overactive mTOR and reduced autophagy, by blocking normal synaptic pruning that may underlie learning appropriate behaviour, may be a unifying feature of autism.”
Dr Alan Packer, from US research charity the Simons Foundation, which co-funded the study, said: ” It is possible that screening for mTOR and autophagic activity will provide a means to diagnose some features of autism, and normalising these pathways might help to treat synaptic dysfunction and treat the disease.”
Carol Povey, director of the National Autistic Society’s Centre for Autism, said: “This interesting research may help develop our understanding of the complex brain differences that exist between people with autism and those who do not have the condition.
“However, the suggestion that a drug could be developed to ‘treat’ autism should be treated with caution. Aside from considering the ethical implications, we question whether it’s possible to extrapolate the effects of medication on mice that exhibit supposedly autism-like behaviour to humans who actually have the disability.”