A drug given to pregnant mice with models of autism prevents autistic behavior in their offspring, according to a new report, and though the drug could not be administered prenatally in humans - there is no way to screen for autism in human fetuses - clinical trials of this drug administered later in development, in young children who have already developed autistic symptoms, have showm promise. 

The causes of autism spectrum disorder, or ASD, are complex and not well understood.

Prolonged excitation of brain neurons seems partly to blame. GABA -- the main inhibitory neurotransmitter in the brain -- has drawn attention. GABA typically excites neurons in the brain of a growing fetus and then quiets them during birth, a switch mediated by oxytocin from the mother, and one that has a protective effect during the vulnerable birthing process. 

But in autism, this switch doesn't happen. The neurons remain excited. This is because chloride – a key signaling molecule -- builds to higher concentrations than it should inside neurons.

In a 2012 clinical trial in which a chloride-lowering drug was administered to more than 50 three- to eleven-year olds with autism or Asperger syndrome, the drug, bumetanide, was shown to reduce the severity of both autism and Asperger syndrome, with few side effects.

Now, to better understand if bumetanide affects the cellular processes that underlay the GABA switch in the way they believed it did (by lowering chloride levels), and to determine whether restoring the switch alone could reduce autistic symptoms, Roman Tyzio and colleagues (several of whom were involved in the clinical trial) studied two rodent models of the disorder -- one caused by a gene and one caused by an environmental factor (namely, exposure to a drug called valproate in utero).

Oxytocin didn't signal from mother to baby in either rodent model, the researchers found, and as a result, chloride built to higher concentrations than it should have inside fetal neurons. By injecting the mothers with bumetanide, however, the researchers were able to reduce chloride levels to their appropriate amount -- and in turn, to restore the GABA switch mechanism (it went from excitatory to inhibitory, as expected).

Critically, offspring exposed to this treatment (administered one day before delivery) didn't demonstrate traits of autism.

"This study validates the clinical trials using bumetanide to reduce chloride and restore strong GABAergic inhibition in humans with autism," said the study's senior author, Yehezkel Ben-Ari, of Aix-Marseille University in France and Inserm Emeritus Research Director at the Mediterranean Institute of Neurobiology. He is referring to the 2012 clinical trial done in more than 50 young children. Now, additional clinical trials with this drug are being pursued in Europe. 'This is one of the most promising clinical strategies for ASD currently available,' Dr. Ben-Ari said.

"The observation that a single treatment of the mother before delivery prevents the expression of … features of autism in offspring illustrates the importance of conditions at delivery," he added, "and the amazing long-term priming consequences of a wrong start." In other words, chloride levels during delivery appear to need to be low.

Together, the work of Ben-Ari and colleagues suggests that abnormal chloride levels may cause the defective GABA switch associated with autism, and that these chloride levels may be treatable beyond infancy.

"I think that an early diagnosis of ASD coupled with a drug such as bumetanide or other regulators acting to reduce aberrant brain activities that perturb neuronal activities are likely future therapies," Dr. Ben-Ari said.

He added that communicating with a child suspected to have autism as early as possible is important, too. "Behavioral treatments might be reinforced by these pharmaceutical treatments and might facilitate their success, restoring communication with the child at an early age."