Skip to Content

Solving the Autism Puzzle

For years scientists searched fruitlessly for the causes of autism by looking for genes shared by families prone to the disorder. Now researchers taking a new approach have begun to unlock its secrets.
December 18, 2014

His name was David. He was 10 years old and, to put it bluntly, compellingly weird—especially in the buttoned-down, groomed normality of suburban Long Island in the early 1960s. At the time, Michael Wigler was a ninth-grade student in Garden City, and he liked to hang out at the home of his girlfriend. That’s where he encountered David, her younger brother. Half a century later, he still can’t get the boy out of his mind.

“He was just like from another planet—it was like meeting an alien,” says Wigler, who ended up a little further east on Long Island as a geneticist at Cold Spring Harbor Laboratory. “He was so different from anybody I had ever met before. First of all, he threw his arms about a lot. And then he moved his head around a lot and would never look at you when he talked to you. And he had an uncanny knowledge of baseball statistics. And I just thought, you know, ‘Boy, this guy is really different. I mean, he’s not just a little different. He’s very different.’”

In the 1950s and 1960s, children like David were pretty much anomalies without a name. Long after becoming a prominent cancer researcher, Wigler would mention him to colleagues, students, postdocs, writers, almost anyone. As one of those postdocs later recalled, “At the time, autism existed; they just didn’t call it autism, so Mike didn’t know this kid had that particular disorder.” Nonetheless, Wigler had become fascinated by the biological mystery that might explain such aberrant behavior. “I think it’s probably what got me interested in genetics,” he says.

Wigler, now 67, indeed devoted his career to genetics, establishing a reputation as one of the most original and productive thinkers in cancer research. So it was a bit of a surprise when, about 10 years ago, he jumped into autism research. Even more surprising has been what he and a few other maverick geneticists began to find.

Nicole Appel
Vintage Tools

One of the things Wigler had seen in cancer is that the disease usually arises because of spontaneous mutations. Rather than lurking in the population for generations and passing from ancestors to descendants, as in classic Mendelian illnesses like Huntington’s disease, these noninherited mutations popped up in one generation. They were fresh new changes in the DNA—de novo mutations, in the jargon of geneticists. As a cancer researcher, Wigler developed new techniques for identifying them, and that led to another surprise. Some of these new mutations were often stunningly complex—not just little typos in the DNA, but enormous chunks of duplicated or missing text, which often created unstable, mistake-prone regions in chromosomes.

All that—the memory of David, his successes in understanding cancer genetics, and the resulting realization that a focus on inheritance might miss some of the most significant disease-causing genes—served as background when, in the spring of 2003, Wigler received a phone call from James Simons, a wealthy hedge fund manager and cofounder (with his wife) of the Simons Foundation, whose daughter had been diagnosed with an autism spectrum disorder. The foundation had received a grant proposal for a research project, and Simons asked Wigler if he would be willing to evaluate it.

The researchers had proposed hunting for autism genes using conventional methods to look for inherited mutations passed down through families. Wigler didn’t mince his words. “I thought they were looking the wrong way,” he says now. “And I didn’t want to see all this wasted effort.”

Wigler, still fascinated by the boy he’d met some 40 years earlier, threw his own hat in the ring. “Autism?” he recalls telling Simons. “Autism? Iwant to work on autism.”

Beginning with a paper in Science in 2007 and culminating with a report published in Nature last October, Wigler’s group and its collaborators have written a dramatically different story about the genetic origins of autism spectrum disorders—a story so unexpected and “out of left field,” as Wigler puts it, that many other genetic researchers refused to believe it at first. Wigler and his colleagues have shown that many cases of autism seem to arise from rare de novo mutations—new wrinkles in the fabric of DNA that are not inherited in the traditional way but arise as last-minute glitches during the process in which a parent’s sperm or egg cells form.

Importantly, these rare mutations exert big effects on neurological development and function. Wigler’s methods have allowed researchers to zero in on numerous genes that are damaged in people with autism and begin to classify subtypes according to the genes involved. And they have begun to take the next step: using the specific genes as clues, they are working to identify critical pathways that may shed light on how the disorder works and suggest possible therapies.

Jessica Park
The Duke University Chapel

Publishing Errors

It shouldn’t be surprising that the genetics of autism make for an extremely difficult puzzle. After all, autism disorders cover a spectrum characterized by everything from atypical yet highly functional behavior to severe intellectual disability—a jumble of excitation and withdrawal, stunning intellectual capacity and severe mental disability, kinetic explosions of movement and repetitive actions, and other symptoms seen to varying degrees in different people. And yet much current research is predicated on the belief that the tiniest aberration at the level of genes, in the wrong place at the wrong time in development, can produce the kinds of aberrant behavior that are the hallmark of autism: social awkwardness and repetitive thinking and actions.

Since the disorder was first described in 1943, by Leo ­Kanner of Johns Hopkins, people have been vexed by its complex and paradoxical nature. Researchers have put forward a series of hypotheses that have not survived scientific scrutiny, attributing it to everything from emotionally remote mothers to ingredients in childhood vaccines. Genetics had always been an obvious route to explore, because it was known that autism often runs in families. So researchers have spent years gathering data on affected families and looking for suspicious mutations passed down from parent to child.

Geneticists pored over genomes in search of small shared errors in the DNA that were seen frequently enough to explain the disorder. But overall, these attempts were consistently uninformative; to use Wigler’s characterization, they were “worthless.” Though the search turned up a few common genetic variants found in people with autism, each of these variants has only an insignificant effect. The effort to find the genetic causes of autism by this strategy was “a total failure,” says ­Gerald ­Fischbach, scientific director of the Simons ­Foundation.

That was precisely the point that Wigler made to James Simons when the foundation sought his advice. Wigler wanted to take the opposite approach: look for new mutations that were not shared by parents and children. Although extremely rare, these mutations were often very disruptive, creating devastating effects in a single generation; identifying them would be a much more effective way to discern which genes are especially important in autism. So Wigler urged the Simons Foundation to find families in which only one child had autism, while the parents and siblings did not. Thanks to their cancer research, he and his colleagues had already developed the technology to spot newly arising mutations, and it looked like a more powerful way to identify key autism-related genes, too.

David Barth
Vogels (Birds)

Wigler’s move into autism came at an important juncture in the biology of development disorders. It was one thing to implicate new mutations in cancer, a disease that often results from genetic insults to a person’s DNA over a lifetime. It was quite another to suggest that de novo mutations played a major role in diseases that develop early in life. But scientists led by Wigler and a few others, including Evan Eichler at the University of Washington, had begun to find that the genome itself was not what previous researchers had envisioned. While the Human Genome Project had presented genomic DNA as a single thread of letters (the “sequence”), and researchers had then catalogued variations consisting primarily of thousands of small differences of a letter or two, “new school” geneticists were finding oddities: huge duplications, gaping holes, and vast tracts of repetitive segments, known collectively as copy number variants. “Let’s suppose you buy a book,” Wigler says. “We’re used to getting books where the cover’s on right, the pages are in order, and they tell a continuous story. But imagine a publisher that duplicated his pages, dropped some pages, changed the order of the pages. That’s what happens in the human genome. That’s copy number variation.”

This form of mutation turns out to appear with surprising frequency in the human genetic text. Wigler’s group first glimpsed the phenomenon in cancer cells, but his hunch was that similar “publishing” errors might also play a role in diseases like autism. Sure enough, when the researchers examined the genomes of people with autism, they often found weird, large-scale duplications or deletions of DNA—mutations not present in the mother or father. The fact that they were not inherited strongly suggested that they were recent corruptions of the genetic text, almost certainly arising in the sperm or egg cells of the parents.

As more families participated in the research, and as technologies for identifying mutations improved, this body of work painted a new picture of the genetics of autism (indeed, the genetics of neurocognitive disorders more generally), confirming that de novo mutations and copy number variations account for many cases of the disorder. And these mutations seem to be especially prevalent in genes that affect neurological development and cognition.

Jessica Park
The Chrysler Building with Perihelion and Transit of Venus #2

In October, Wigler’s group—with collaborators including Eichler at the University of Washington and Matthew State at the University of California, San Francisco—identified up to 300 genes potentially related to autism. Twenty-seven of them confer a significantly heightened risk when disrupted by these rare new mutations. Each specific de novo mutation is rare enough to be found in less than 1 percent of the autism population, but collectively they may account for 50 percent of all cases of autism, says the Simons Foundation’s Fischbach.

Some of these genes are active in the earliest weeks of prenatal brain development; others kick into gear after birth. Some affect the function of synapses, the junctions between nerve cells; others affect the way DNA is packaged (and activated) within cells. One gene, CHD8, previously linked by Eichler’s group to children with a severe form of autism, has also been linked to schizophrenia and intellectual disability. Subtypes of autism seem to be associated with mutations in certain genes, which may begin to explain such long-standing mysteries as why some cases of autism produce severe symptoms while others cause more modest behavioral tics.

The findings also provide insight into just why autism is so common. “Let me highlight a critical point, and one of the biggest insights to come from the genetics of autism,” says ­Jonathan Sebat, a professor at the University of California, San Diego, who previously worked in Wigler’s lab and helped to reveal this new genetic landscape. “We did not fully appreciate how plastic the genome is, in the sense of how much new mutation there is. The genome is mutating, evolving constantly, and there’s a steady influx of new mutations in the population. Every child born has roughly 60 new changes in their DNA sequence, and [one in] every 50 children born have at least one large rearrangement. This is a really significant contributor to developmental disorders.”

Another surprising discovery is that certain regions of the human genome seem especially prone to disruption. Not only do some of these genetic “hot spots” seem to be linked to many forms of autism, but some of them have a deep and significant evolutionary history. If you trace them back in time, as Evan Eichler’s laboratory has begun to do, you can begin to glimpse the emergence of precisely the traits that distinguish humans from all other animals. “It’s kind of a crazy idea,” Eichler says, “but it’s like autism is the price we pay for having an evolved human species.”

Copy number variations in one specific hot spot on the short arm of chromosome 16, for example, have been associated with autism. By comparing the DNA of chimpanzees, orangutans, a Neanderthal, and a Denisovan (another archaic human) with the genomes of more than 2,500 contemporary humans, including many with autism, Xander Nuttle, a member of Eichler’s group, has been able to watch this area on the chromosome undergo dramatic changes through evolutionary history. known as BOLA2 that seems to promote instability. Nonhuman primates have at most two copies of the gene; Neanderthals have two; contemporary humans have anywhere from three to 14, and the multiple copies of the gene appear in virtually every sample the researchers have looked at. This suggests that the extra copies of the BOLA2 gene, which predispose people to neurodevelopmental disorders like autism, must also confer some genetic benefit to the human species. Otherwise, evolutionary pressure would have scrubbed the duplications out of the genome. In other words, the same duplications that can lead to autism may also create what Eichler calls genetic “nurseries” in which new gene variants arise that enhance cognition or some other human trait.

At a meeting of the American Society of Human Genetics last fall, Nuttle reported that this mutation-prone region, which contains more than two dozen genes related to neurocognitive function, lies adjacent to an intriguing gene

“The evolutionary twist on this whole story,” says Eichler, “is that our genome is really set up to fail, in the sense that we’re prone to delete and duplicate. The flip side of it is that that selective disadvantage is offset by the emergence of novel genes that have conferred an advantage to us cognitively.”

Jeroen Pomp
Dieren en Planten (Animals and Plants)

Diagnosing Hope

Despite the recent advances in autism genetics, there hasn’t been much difference at the treatment level. Thomas Insel, director of the National Institute of Mental Health, put the new findings in perspective in an interview with a reporter from the Simons Foundation at the Society for Neuroscience meeting last November. “This has been an incredible period of discovery,” said Insel, “but families are looking for interventions, not papers.”

As genetic researchers identify more genes involved in autism, they are beginning to classify autism cases according to their association with particular mutations. Eichler’s team, for example, recently gathered a group of patients with a mutation in the CHD8 gene. And “lo and behold,” Eichler says, the individuals shared many symptoms: 73 percent, for example, had severe gastrointestinal problems (CHD8, the researchers subsequently discovered, is also active in the gut). Such findings may in turn point to gene-specific interventions someday. The long-range hope is that as more rare mutations associated with autism are uncovered, the affected genes will tend to converge in ways that suggest molecular pathways critical to neurological development and function. Researchers are quick to point out that de novo mutations are only part of the autism story. Scientists continue to hunt for inherited mutations and common variations that may also play important roles. But by using de novo mutations to spotlight some of the genes involved, Wigler and others have provided renewed hope for the field. Indeed, though Wigler concedes that there is “a long way to go” before genetic findings translate into useful medicines, he sees therapeutic possibilities in the very nature of those mutations. “Because the kids that have autism have one bad gene and one good gene, I think there should be ways of getting that good gene to be more active, and probably reversing things,” he says.

The genetic findings also suggest that even more dramatic (and ethically provocative) forms of therapy may be possible in the more distant future. “For many of the genes that we now think are important for autism, the genes are essentially [active] at eight to 16 weeks of development,” says Eichler. “So you have to not only make a diagnosis early, but some people argue that you have to intervene early in order to make a big difference.” And because many of the genes in question are also related to intelligence, Wigler says, it will be tempting to harness emerging technologies like prenatal genome analysis and precise new gene-editing tools as part of broader interventions in cognitive development. “It’s a little dangerous to tap into it,” he adds, “because we’re getting to designer babies and the Gattaca world. The autism world does bring us face to face with some science fiction stuff.”

As urgently as Wigler wants to understand the puzzle of autism, even he abides by certain limitations on his curiosity. Asked if he had ever been tempted to reconnect with David, the autistic boy who inspired his original interest in the disease, he practically recoiled. “No,” he said quickly. “That would be intruding.” But he still can’t stop talking about his old girlfriend’s brother with something like awe. “It wasn’t like he was trying to be different, you know? He wasn’t,” he said. “If anything, he was probably doing the opposite. But he was just really different. And it was an amazing thing.”

Stephen S. Hall, a science writer based in New York, teaches science communication and journalism at New York University. His last story for MIT Technology Review was “Neuroscience’s New Toolbox.”

Keep Reading

Most Popular

It’s time to retire the term “user”

The proliferation of AI means we need a new word.

The problem with plug-in hybrids? Their drivers.

Plug-in hybrids are often sold as a transition to EVs, but new data from Europe shows we’re still underestimating the emissions they produce.

Sam Altman says helpful agents are poised to become AI’s killer function

Open AI’s CEO says we won’t need new hardware or lots more training data to get there.

An AI startup made a hyperrealistic deepfake of me that’s so good it’s scary

Synthesia's new technology is impressive but raises big questions about a world where we increasingly can’t tell what’s real.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at with a list of newsletters you’d like to receive.