Understanding Autism: Baby's Delivery May Be Key

Quieting nerve cells in the brain during birth in mice seems to protect against autism, according to a new study.

A girl with autism embraces a horse at a specialized riding school in Germany.


Autism is as mysterious a disease as it is tragic. While there is clearly a genetic component to some forms of the illness, it remains unclear how many genes are involved and what role the environment may play in causing the varying degrees of cognitive and social impairment classified as autism spectrum disorders (ASD).

New research published February 6 in the journal Science points to another possible player in this complex of causes: the interplay of hormones and chemicals affecting the fetus during birth. (See "One Thing We Know About Autism: Vaccines Aren't to Blame.")

In a study led by Yehezkel Ben-Ari of the Mediterranean Institute of Neurobiology in Marseille, France, strains of pregnant mice modeled to give birth to offspring showing autistic behaviors were given a drug to lower their levels of chloride. Their offspring, which would normally have been expected to show signs of the disease, were born with no symptoms of autism.

An earlier study in mice in 2006  showed that oxytocin, a hormone that triggers contractions, also acts as a critical "switch" affecting the function of a neurotransmitter called GABA. Usually, GABA excites neurons in the growing brain of the fetus. But during the birth process, oxytocin makes GABA quiet the neurons instead, protecting the fetus's vulnerable brain. When this switch doesn't happen, the neurons remain in their excited state, and chloride, a key signaling molecule, builds to higher-than-normal levels. The mice are born with autism. (See also "Oxytocin: The 'Love Hormone' Might Also Help in Autism.")

In the study announced today, the same group of researchers administered a drug called bumetanide, which reduces chloride levels, to two kinds of pregnant rats modeled for autism. Their offspring were born with normal brain function. Conversely, blocking oxytocin in normal pregnant mice resulted in offspring showing autistic-like behaviors.

The result of the rodent study adds further hope to a 2012 clinical trial of 60 human children aged 3 to 11 years old who were given bumetanide. The drug reduced the severity of autism symptoms in the children.

The rodent study does not suggest a way of treating humans prenatally to prevent autism, however, since there is no way to identify fetuses that will later develop the disease as young children. But it provides information to develop further research in animals and in humans.

"If this result is confirmed," says Ben-Ari, "then it implies that what we need is as early diagnosis as possible. We know that the earlier the treatment, the better the success."

The search for biomarkers that might some day identify autism in the uterus—brain waves, blood or urine tests, or clues in amniotic fluid—is an active area of research, says Andrew Zimmerman, a pediatrician at the Lurie Center for Autism at Massachusetts General Hospital in Boston.