January 10, 2013

It is difficult to overstate the complexity of autism.

Two children born with identical genetic mutations can have profoundly different outcomes—one with typical neural and social development, the other unable to communicate with the world around him.

One in 88 children in the U.S. has autism, according to the CDC’s Autism and Developmental Disabilities Monitoring (ADDM) Network. Hundreds of different genetic abnormalities may contribute to—or inhibit—autism. Many of the known anomalies cluster around genes that control the development of neurons and synapses, but many others are in regions of the genome that are still poorly understood.

Layered on top of the bewildering genetics, scientists are only beginning to understand how environmental factors may affect this condition. Autism, which is characterized by impaired social communication, fixated interests and repetitive behaviors, may occur alone or as part of other developmental syndromes, such as Fragile X, Rett’s, and tuberous sclerosis, conditions that also include intellectual delays, epilepsy and a number of gastrointestinal disorders.

Unraveling all of this is like trying to solve a riddle encrypted in a cipher printed on a jigsaw puzzle—with extra pieces mixed in that don’t quite match.


This complexity presents fascinating challenges for researchers, but it also makes diagnosing and treating children with autism vexingly difficult. Even with the latest genetic testing, pediatricians still can’t predict whether an at-risk child will develop autism, or how severe symptoms will be. And while some genetic analogs for autism have been reversed in animal models, there is currently no treatment for any of the core symptoms of autism for people with the disorder.  

Despite its daunting complexity, however, new technologies and techniques and an emphasis on collaborative research are helping scientists unravel the fundamental biological basis for autism, and providing suggestions for new treatments. 

A hub for autism research

In 2006, realizing that the challenges of autism’s complexity were too great to solve piecemeal, a group of researchers and patient advocates in Boston banded together to form the Autism Consortium. The Consortium, which brings together researchers from Harvard, Boston University, MIT, Tufts, UMASS Medical School and affiliated hospitals and research institutes, affords a platform for collaboration, enabling diverse groups to share samples of patient and animal data. The organization hosts an annual meeting that highlights biomedical and clinical advances.

“The history of science isn’t just a history of great ideas, it’s a history of new tools,” said Steve Hyman, HMS professor of neurobiology and Harvard professor of stem cell and regenerative biology, at a Consortium meeting at HMS last fall.

“Galileo’s novel use of the telescope allowed him to revolutionize our understanding of the workings of the solar system,” said Hyman. “Galileo was a really smart guy, but without the tool, no new conception of the universe, no excommunication—none of that stuff.”

Those new tools—including genomic resources, engineering tools, cellular models, and organizational models and partnerships to facilitate data sharing and collaboration—are opening windows for understanding the biological bases for autism.

One challenge to understanding the biology of autism is purely physiological: the disorder is hidden in the folds of the brain. Researchers can’t biopsy living brain tissue, and the image resolution of many imaging techniques is severely limited. Because autism is a disease of neural systems and circuits and a disorder of human social behavior, studying nerve cells in the lab and animal models provides limited insight.

Nicholas Lange, HMS associate professor of psychiatry at McLean Hospital and Harvard School of Public Health associate professor of biostatistics, has worked on brain imaging techniques that are beginning to shed some light on the physiology of the brains of people with autism. This is another area with a great deal of unmapped territory to explore. In a recent essay in Nature, Lange emphasized that we still need to better understand the great breadth and depth of healthy typical and atypical neurobiology and the diversity of the types of brain features in autism, and that without that foundational biology all attempts at diagnosing autism by brain imaging are futile.

“Brain scans, neuronal subtyping and genetics are the signposts showing us where and how to look, see, ponder, hypothesize and test,” Lange said. “We start where we are and delve more deeply into the biological etiology and treatment of autism.”

Some members of the Consortium are planning to grow neurons from stem cells as a way of examining how neurons from patients with different genomic profiles develop and operate.

Others are developing new techniques to look at how the neurons perform at the synaptic level, using high-resolution, high-speed imaging to visualize a single pulse of neurotransmitters in order to see what happens each time a neuron fires.

“You need that excruciating, exquisite level of detail so you can watch signals in real time,” said Mriganka Sur, director of the Simons Center for the Social Brain at MIT. Researchers will also need to scale these tiny neuronal interactions to see how they interrelate in complex neural networks, he added.

“One of the greatest challenges is determining how these genes map onto the development of the social brain,” Sur said.

Synaptic tuning

One reason autism is so complex is that the process of brain development itself is incredibly dynamic, according to Christopher Walsh, chief of the Division of Genetics at Boston Children’s Hospital and HMS Bullard Professor of Pediatrics and Neurology.

Walsh began his career as a developmental neurobiologist and focused on genetics as a method of getting to the root of neural development. In 2008, working with samples of consanguineous Middle Eastern families to find recessive genes related to developmental disorders, Walsh identified a handful of new genetic components of autism, including both protein encoding genes and the codons surrounding genes that act as on-off switches for genes.

Michael Greenberg, chair of the neurobiology department at HMS and Nathan Marsh Pusey Professor of Neurobiology, had previously identified those genes as key players in experience-dependent learning.

“When neurons become active, they turn on a lot of genes,” Walsh said. When a person is learning a new behavior, repeated experience activates genes that, over time, strengthen some connections in the brain and weaken others. This building and pruning of neural pathways shapes the networks that record and enable learned behaviors.

This kind of learning is essential to developing the complex social behaviors impaired in autism, and it turns out that several of the genomic anomalies associated with autism are related to regulators that signal these synaptic shaping and pruning genes when to turn on and off.

If the genes involved in making proteins to build those connections don’t turn on at the right time, or if they shut down too soon, they might not build an important connection; if they stay on too long, connections can be overly strengthened.

“You could think of it as synaptic tuning,” Walsh said. “When a synapse fires, it doesn’t just fire once, it fires repeatedly at a certain frequency. “Hearing” the right tune is what stimulates the next synapse,” he said. “In autism, for certain kinds of complex tasks, like social communication, you might say that you’ve lost that tuning mechanism.”

Walsh said a variety of drugs under development to treat synaptic deficiencies in far-flung disorders from Fragile X to Alzheimer’s might prove effective in treating autism.

Gene hunting

A handful of genetic anomalies related to autism have been identified, but some researchers believe that more than 100 genes may contribute to autism, and that those genes are just the beginning of a series of complex biochemical pathways, many of which remain uncharted territory.

Researchers in the Greenberg lab are studying the genetic and molecular machinery that builds brains; that is, the proteins that coax embryonic stem cells into becoming neurons as well as the switches that help promote and prune synapses. When this developmental process goes awry, conditions such as autism can result. For example, Greenberg discovered how the loss of a key enzyme related to the molding of synapses could explain the devastating developmental deficits that occur in Angelman syndrome, which causes a constellation of developmental problems in children, including mental disability and, in some cases, autism. 

There are more than 300 genes that trigger these synaptic changes. Many of the genes that Greenberg’s molecular and cellular studies have identified as playing important roles in the modification of synaptic connections have been identified as sites for genetic anomalies that may contribute to autism. Researchers probing these links include Walsh and Mark Daly, HMS associate professor of medicine at Massachusetts General Hospital’s Center for Human Genetic Research.

 “One piece of the puzzle isn’t enough to open up the biology and crack something as complex as autism,” said Daly, who is also director of computational biology for the Medical and Population Genetics Program of the Broad Institute of Harvard and MIT.

Working with the samples available through the Autism Consortium, Daly’s lab has now sequenced the genomes of hundreds of families, identifying several novel genetic contributions to autism.

 “By this time next year, we will have several thousand families sequenced. We’ll have a much bigger set of those puzzle pieces,” Daly said.

Recent technical advances have made gene sequencing much more accessible. A decade ago it took 10 years and cost $1 billion to sequence a human genome. Now it can be done in one week or less for a few thousand dollars. Still, the work of understanding sequenced genomes remains challenging, Daly said.

Differences in the genomes of any two healthy individuals are vast. When researchers sort through all that unique data, in billions of combinations of bases in the genetic code, identifying the differences that are characteristic of autism is no simple task.

“It’s a huge computational challenge,” Daly said. “We can’t yet answer basic questions for families about the potential severity of their children’s illnesses, let alone offer any treatment.”

Finding—and learning from—biomarkers

One way to solve the autism puzzle is to identify the genes that are implicated in the disorder and then study the pathways and processes that those genes regulate.

Another way is to identify measurable biological features characteristic of autism, and to compare those with the same features in the biology of typical development. If genetics are the pieces of the jigsaw puzzle, biomarkers are part of the picture on the box that shows you what the pieces will look like when they’re put together.

HMS researchers at Boston Children’s Hospital, for example, have developed a blood test for autism spectrum disorders that outperforms existing genetic tests. It also presents evidence that abnormal immunologic activity affecting brain development may help explain some of autism’s origins.

The same is true for many other projects that autism researchers are undertaking to define how the disorder manifests itself physiologically; for example, measuring electrical activity in the brain, monitoring stress levels and predicting the onset of seizures by measuring skin conductivity.

The research not only leads to more objective diagnostic tools, it can also provide crucial insight into the basic biology of the brain, which remains one of the great black boxes of biomedical science. 

As complex as the challenges are, researchers say that great progress is being made in sorting out the roles of genes and gene regulators, of toxins and the immune system, of synaptic pruning and stress. Researchers are also building a basic understanding of the function of the social brain from the molecular to the macro level, and beginning to see how the pieces fit together.

 “All of these things are in sight now,” said Gerald Fischbach, director of the Simons Foundation Autism Research Initiative. “We have the tools to do this.”