In most cases, autism is caused by a combination of genetic factors, but some cases, such as Fragile X syndrome, a rare disorder with autism-like symptoms, can be traced to a variation in a single gene that causes overproduction of proteins in brain synapses, the connectors that allow brain cells or neurons to communicate with one another. Now a new study led by the same MIT neuroscientist who made that discovery, finds that tuberous sclerosis, another rare disease that leads to autism and intellectual disability, is caused by a malfunction at the opposite end of the spectrum: underproduction of the synaptic proteins.

Mark Bear, the Picower Professor of Neuroscience and a member of the Picower Institute for Learning and Memory at Massachusetts Institute of Technology (MIT), and colleagues write about their findings in the 23 November online issue of Nature.

It seems puzzling that underproduction of synaptic proteins and overproduction of those same proteins lead to the same disorder, but it does fit into the idea that autism is caused by a wide range of problems to do with brain synapses, as Bear tells the press in a statement:

“The general concept is that appropriate brain function occurs within a very narrow physiological range that is tightly maintained.”

“If you exceed that range in either direction, you have an impairment that can manifest as this constellation of symptoms, which very frequently go together – autism spectrum disorder, intellectual disability and epilepsy,” he adds.

The study also offers a caution for drug developers making drugs that target the cellular origins of autism: they will have to be tailored to individual patients to make sure they do more good than harm.

Phase III trials of drugs to treat Fragile X syndrome are already under way.

The journey that led Bear to study autism and Fragile X syndrome started when he was investigating mGluR5, a receptor found on the surface of brain cells or neurons that plays a key role in sending signals between two neurons communicating across the synapse. The neuron sending the signal is called the presynaptic neuron and the neuron receiving the signal is called the postsynaptic neuron.

(To put this in context, it is sobering to remember that the human brain contains billions of neurons and trillions of synapses, that each neuron can connect to thousands of synapses and that from this hugely complex interconnected signalling system emerge the brain functions than control memory, emotions, learning, movement and sensing.)

The presynaptic neuron sends a signal to the postsynaptic neuron by releasing glutamate, a neurotransmitter that diffuses across the synapse and binds to the mGluR5 receptor on the postsynaptic neuron. When this happens, it triggers the production of new synaptic proteins. Fragile X protein (FMRP) acts as a brake on this protein synthesis, as Bear explains:

“The appropriate level of protein synthesis is generated by a balance between stimulation by mGluR5 and repression by FMRP.”

Loss of FMRP results in overproduction of synaptic proteins, which leads to Fragile X syndrome and its symptoms: seizures, autistic behavior and learning disability. Bear and others have already established that blocking mGluR5 in mice can reverse those symptoms.

After this, the resarchers started to wonder what might happen if mGluR5 were overactive: would it cause other autism-like syndromes? That is when they turned their investigation to tuberous sclerosis (TSC). But they were not expecting to find what they did: in mice with TSC, synapses have too little protein, so when treated with a drug that inhibits mGluR5, they did not improve. But when treated with a drug that stimulated it, they did.

It appears that Fragile X and TSC are “mirror images” of each other, says Bear. One is a case of too much protein synthesis where blocking mGluR5 reverses the symptoms, but the other is a case of too little protein synthesis, and symptoms only impove with stimulation of mGluR5.

In the Nature paper, he and his colleagues conclude:

“Thus, normal synaptic plasticity and cognition occur within an optimal range of metabotropic glutamate-receptor-mediated protein synthesis, and deviations in either direction can lead to shared behavioural impairments.”

The big message of this study appears to be that not all cases of what may look like similar autism will respond to the same treatment.

“This study identified one functional axis, and it will be important to know where a patient lies on this axis to devise the therapy that will be effective,” says Bear.

“If you have a disorder of too little protein synthesis, you don’t want to inhibit the neurotransmitter receptor that stimulates protein synthesis, and vice versa,” he adds.

In a way, this is not a surprise, because it echoes what we now know about treatments for bipolar disorder and schizophrenia, that also have varied origins.

Bear says they would also like to be able to go on from these rare cases of autism, that account perhaps for 10% of cases, to help the other 90% of people with autism of unknown cause.

Developing and approving drugs that block or stimulate mGluR5 may help scientists identify which autistic patients respond to which drugs, and aid the development of biomarkers in a field where there are currently no good tests for finding which genetic markers patients may have.

Melissa Ramocki, an assistant professor of pediatric neurology at Baylor College of Medicine was not involved in Bear’s study. She said finding out how a given mutation behaves at the molecular level will help tailor treatment to the individual patient, and studies like this are very important in that respect. They are “exactly the kind of work that needs to be done to understand the molecular mechanisms, because the treatments will be so diverse,” she says.

In the meantime, Bear and his team are looking at what happens to mGluR5 activity in other single-gene disorders such as Angelman syndrome and Rett syndrome, and they are also trying to identify the detailed steps in the mGluR5 synaptic protein synthesis pathway.

Written by Catharine Paddock PhD