Inside Salk; Salk Institute
Home > News & Press > InsideSalk > 02|11 Issue > Unlocking the Mystery of Autism
Carol Marchetto, Fred H. Gage, Alysson Muotri

From left to right: Carol Marchetto, Fred H. Gage, Alysson Muotri



The pictures of her first birthday party show a smiling little girl, excitedly reaching for the candles on her cake. A year later, she is unable to sit up or grasp with her fingers. Rett syndrome, a devastating brain disorder, has thrown her development into reverse.

Triggered by a tiny genetic flaw, the rare disease afflicts girls almost exclusively. The symptoms start to emerge just as they are beginning to walk and talk. Seemingly out of the blue, normal development slows down, and eventually the infants regress, progressively losing speech and motor skills. Toddlers who used to bask in their parents' attention often become withdrawn and anxious and avoid eye contact. Instead of reaching for toys, they wring their hands over and over again, leading researchers to classify Rett syndrome as one of the autism spectrum disorders. But unlike most forms of autism, which have no single known cause, almost all cases of Rett syndrome can be traced to defects in a single gene known as MeCP2.

"Because of similar symptoms and shared genetic links, Rett syndrome is sometimes considered a "Rosetta Stone" that can help us to understand other developmental neurological disorders such as autism and schizophrenia," says Fred H. Gage, a professor in the Salk's Laboratory of Genetics and holder of the Vi and John Adler Chair for Research on Age-Related Neurodegenerative Diseases.

In two recently published landmark papers, he and his team reported on some startling discoveries that are bound to change the way people think about psychiatric disease. One features "jumping genes," the other "autistic" brain cells in a dish.

Gage had long been interested in so-called L1 retrotransposons, restless bits of DNA that can freely move about the genome, earning them the nickname "jumping genes." Almost all of them are marooned at a permanent spot by mutations rendering them dysfunctional, but in humans a hundred or so are free to paste copies of themselves into random stretches of DNA.

Brain Image

L1 retrotransposons are particularly active in the cerebellum (shown in purple), which plays an important role in motor control. Green dots represent brain cells with new L1 insertions.

In fact, over past millennia, so many copies have squeezed themselves into the human genome that they now make up 17 percent of a cell's DNA. While most of them have landed in a genomic no-man's land, once in a while, a retrotransposon can insert itself close enough to a functional gene to disrupt its function and cause disease. For example, L1 insertions created the mutations underlying hemophilia and Duchenne muscular dystrophy.

One of the best documented and arguably most beautiful examples of retrotransposons wreaking havoc on a gene's function is the merle coat pattern, which gives Australian shepherds their characteristic patchy coat and stunningly blue eyes. The merle patterning was born when aretrotransposon slipped into SILV, a gene important in mammalian pigmentation.

Most retrotransposons are active in sperm and eggs, explaining how new insertions can be passed on to succeeding generations. But until recently, scientists had no clue that these restless genes might be active elsewhere. Then last year, Gage and his team demonstrated that L1 retrotransposons are on the move in the human brain. They add hundreds of extra copies to the genome of neurons, randomly rearranging our mental structure.

He believes that the addition of these so-called mobile elements is a potential mechanism to create the neural diversity that makes each person unique. "It could also help explain the differences in identical twins," he says. "Even though they are clones, they have their own personalities."

But what if the reshuffling goes wrong? There is a fine line between spawning variety in our minds and eroding the foundation of our mental architecture. "Mobile elements can't just be let loose in the brain; their activity has to be strictly controlled," says Gage.

Maybe not surprisingly, L1 elements are only active during a short window of time: the early stages of brain cell development. Once brain stem cells are committed to spending the rest of their lives as neurons, for example, the L1 elements will cease mobility.

"In my mind, this restricted time frame immediately raised the question of how this process is regulated and what happens when it is derailed," explains Carol Marchetto, a postdoctoral researcher on the Gage team. She and her husband, Alysson Muotri, Ph.D., who started his career in the Gage lab and now leads his own research group at the University of California, San Diego School of Medicine, decided to look for the answer.

Carol Marchetto

Carol Marchetto

Their attention was quickly drawn to MeCP2. It is one of several so-called methyl-CpG-binding proteins, which are best known as gene silencers. They turn off genes by binding to nearby regulatory regions of DNA. Not only is MeCP2 protein incredibly abundant in neurons, it also binds to a stretch of DNA that is known to control the activity of L1 retrotransposons.

What if?

When Marchetto and Muotri discovered a few years ago that mutations in MeCP2 had been linked to Rett syndrome, they couldn't help but think, What if unregulated jumping genes played a role in neurological disorders?

First, the researchers wanted to know whether MeCP2 interfered with the ability of an artificial L1 element to move around in cultured neuronal stem cells. To find the tiny mobile sequence in a sea of genomic DNA, they added to it a molecular tracer—a green fluorescing protein—that lit up the host cell whenever a copy had broken free and migrated to a new spot.

And sure enough, without MeCP2, the cells lit up like a beacon, giving away the presence of roving L1 copies. With MeCP2 around, nothing much happened. What's more, in the brains of mice that had been engineered in the same way, L1 activity increased up to six-fold compared to normal control mice.

Intrigued, Marchetto and Muotri wondered whether the same was happening in the brain cells of patients with Rett syndrome. To find out, they first generated so-called induced pluripotent stem (iPS) cells d erived from Rett patients. They started with skin biopsies taken from affected individuals and a healthy control. By exposing the skin cells to four reprogramming factors, they turned back the clock, triggering the cells to look and act like embryonic stem cells.

When Marchetto introduced the modified L1 element and coaxed the iPS cells down the developmental path leading to neurons, she found that the mobility in cells derived from Rett patients was almost twice as high as that in normal cells.

Making an artificial element jump in a Petri dish was one thing, but what about endogenous L1 retrotransposons that make their home in brain cells of patients with Rett syndrome? To search for evidence that mutations in MeCP2 set endogenous L1s free as well, Gage's team developed a new technique that allowed them to detect the minute increase in DNA content caused by the additional copies inserted into the genome.

"A single L1 retrotransposon, which is only about 6,000 base pairs long, resembles the proverbial needle in a haystack, when compared to the 3 billion base pairs of the human genome," says Marchetto.

The painstaking search paid off. The researchers found what they had been looking for: the number of L1 copies in the brains of Rett patients was significantly higher than in the brains of non-affected individuals.

"This is the first time that we can show a connection between genomic stability and a mental disorder," says Gage. But he cautions that the high rates of neuronal transpositions in MeCP2-deficient mice and Rett patients may be a consequence, rather than the cause, of the disease.

"Nonetheless, new somatic insertions, especially during early developmental stages, may play a role at later stages of the disease and could explain the baffling variability of autistic symptoms observed in Rett syndrome patients," he adds. Since MeCP2 abnormalities have been linked to autism and other mental disorders, the insights gained from studying this gene may be relevant beyond Rett syndrome.

"There is certainly a genetic component to Rett syndrome and other psychiatric disorders, but it may not be the only thing that's relevant," says Muotri. "Somatic insertions and alterations caused by L1 elements could play a significant but underestimated role in other neurodevelopmental diseases because they are hard to detect."

The only way to find out is to decrease the activity of L1 elements in mouse models and to see whether that changes the animals' behavior. "If it does, we can then search for drugs that modulate L1 activity in humans," says Gage.

Autism in a dish

The ability to obtain iPS cells from patients' skin cells and to differentiate them into the cell type damaged by the disease gave Marchetto and Muotri a first glimpse at the connection between hyperactive jumping genes and Rett syndrome.

In a separate study, recently published in the journal Cell, they used these cells to replicate autism in the lab and study the molecular pathogenesis of the disease.

In the past, scientists had been limited to studying the brains of people with autistic spectrum disorders via imaging technologies or postmortem brain tissues. Both of these strategies are limited, however, and do not allow the researchers to perform experiments in live human neurons.

"It is quite amazing that we can recapitulate a psychiatric disease in a Petri dish," says Gage. "Being able to study Rett neurons in a dish allowed us to identify subtle alterations in the functionality of the neuronal circuitry that we never had access to before."

At first, Rett-derived iPS cells were indistinguishable from their normal counterparts. It was only after Marchetto had patiently coaxed the iPS cells to develop into fully functioning neurons—a process that can take up to several months—that she was able to discern differences between the two.

Neurons carrying the MeCP2 mutations had smaller cell bodies and a reduced number of synapses and dendritic spines—specialized structures that enable cell-to-cell communication—as well as electrophysiological defects, indicating that things start to go wrong early in development.

"Mental disease and particularly autism still carry the stigma of bad parenting," says Muotri. "Our results show very clearly that autism is a biological disease that is caused by a developmental defect directly affecting brain cells."

When the researchers treated the diseased cells with a drug that was able to alleviate some of the autism symptoms in mice, the abnormalities in the neurons were reversed. "This finding suggests that the autistic phenotype is not permanent and could be at least partially reversible," says Gage.

Because children don't often develop Rett syndrome until they are six to 18 months old, the discovery suggests there may be a window of opportunity for early diagnosis and preventive therapies before the disease develops.

Often it's hard to test autism treatments in animals because it's difficult to see the physical manifestations of the disorder; researchers can't observe the impaired social interactions and communication that are the hallmarks of the disease in humans.

"We now know that we can use disease-specific iPS cells to re-create mental disorders and start looking for new drugs based on measureable molecular defects," says Muotri.