More than one ball in the air
Paths to new treatments for autism
Editor’s Note: This story was written in 2003. Morgan is now in middle school, and her sister Allison is in college, majoring in comprehensive special education.
That was Morgan several years ago—vivacious, engaged, talkative. Then something went wrong.
“She just stopped looking at me,” her mother, Tammy Vice, recalls. “She started echoing back words instead of saying things on her own. It was like something was taking her away, and we didn’t have any idea what it was.”
More than a year and seven professionals later, that “something” got a name. “By then we already knew,” Vice says. Morgan had autism. Her official diagnosis is Asperger disorder, one of five developmental disorders that make up the “autistic spectrum.”
All children with an autism spectrum disorder share core deficits—abnormal social behavior, impaired communication, and restricted and repetitive behaviors—though the severity and constellation of symptoms vary dramatically. Intellectual function ranges from profound mental retardation to above average intelligence as measured on IQ tests. Some children, like Morgan, appear to develop typically for the first year or two and then stop, or regress. Others show signs of autism from early infancy.
This baffling variability in symptoms and their onset has added to the complexity of diagnosing, treating, and understanding autism. Today, 60 years after it was first described, autism is still a mysterious disorder. It’s clear that brain development goes awry, but why and how exactly are open questions. There are currently no biological markers—genes or blood proteins—that can be used to diagnose the disorder or predict who will suffer from it, and there are no cures.
One thing is clear—autism is not a rare disorder. It affects as many as one in 250 children, four times as many boys as girls, across all racial groups. The number of children with autism spectrum disorders appears to have skyrocketed in the last 10 years, and although this finding is controversial, it has sparked a sense of urgency and an influx of federal and private research funding for autism research. The increased support is bringing renewed energy to efforts to define brain regions that are affected by autism and to identify autism susceptibility genes and environmental “triggers” for the disorder, and it is making some researchers optimistic.
“Ultimately we want a biologic cure and prevention for this disorder, and that’s going to happen,” says Dr. Nancy J. Minshew, director of a Collaborative Program of Excellence in Autism at the University of Pittsburgh. That cure may be several decades off, Minshew acknowledges, but research findings along the way are improving early diagnosis and treatment options, giving children with autism spectrum disorders the best chance for a typical life.
Exploding a myth
The 1950s and 60s saw the rise of theories that parents caused autism by being too “cold” and failing to psychologically bond with their children. It wasn’t until the 1970s that the tide began to turn, with prominent research groups formulating diagnostic criteria for autism and speculating that it had biological underpinnings.
Then came hard proof.
Dr. Susan Folstein had just completed her residency in psychiatry when she joined Dr. Michael Rutter’s group at the Institute of Psychiatry in London. She wanted research experience and took on a project to study autism in twins. Crisscrossing the English countryside, Folstein examined and interviewed as many twins as she and Rutter could find, one or both of whom had autism. They found that identical twins were much more likely to both be affected than were fraternal twins—evidence that autism had genetic roots.
The findings, published in Nature in 1977, “brought to a clear end the period of time when people thought of autism as something that was caused by parents,” says Folstein, professor of Psychiatry at Tufts-New England Medical Center. “The study was also the impetus for many other family studies of autism.”
Folstein and Rutter’s initial twin study and others since then offer the most compelling evidence for the high heritability of autism. Identical twins have a greater than 50 percent chance of both being autistic. The relative risk to siblings—nine to 45 times the risk to the general population—is higher than many other complex disorders like diabetes, asthma, and schizophrenia. But the genetic factors contributing to autism have remained elusive.
Jonathan L. Haines, Ph.D., director of the Vanderbilt Center for Human Genetics Research, began collaborating with Folstein in the mid-1990s. Compared to other diseases and disorders he was working on at the time, “it looked like the genetics of autism should be relatively easy to solve,” he recalls. “Big mistake.”
“The data are suggesting that there are many genes and that they are probably interacting to cause this disorder,” Haines says. “It’s incredibly complex.” That makes sense, he adds, given the sheer variation in the expression of the disorder, in the symptoms that children have.
To search for those proverbial needles in the haystack—the small group of genes that contributes to autism out of the 30,000 or so genes that make up the human genome—investigators use a combination of two approaches. They scan randomly through the entire genome in families with at least two affected individuals, called multiplex families, looking for DNA regions with high similarity in the people with autism. And they examine “candidate genes,” which because of their biological function are suspected of playing a role in the developmental changes that cause autism.
The approaches have pointed to many different chromosomal regions—areas on chromosomes 2, 3, 7, 15, and 19 that are linked to the disorder in some sets of families—and to some candidate genes. “There’s a lot of disagreement about those regions, and there are no confirmed genes,” Haines says. “No one’s been able to say, ‘This is a gene that causes or influences autism.’
“One could argue that the field is in a great state of confusion right now,” he adds. “But out of that confusion is going to come some real progress in the next year or two.”
The power of genetics
One reason for Haines’ optimism: the development of the Autism Genome Project, a consortium of autism genetics researchers promoted by the National Alliance for Autism Research. If NAAR, a parent-founded advocacy group, is able to secure funding for the giant genetics initiative, up to 1,200 multiplex families will be available for study.
“With those families, we’re going to do a very high density genome screen,” Haines says. “That will be the definitive genome scan for autism, because it will include virtually all of the multiplex families in the world.”
At the same time, autism genetics researchers are trying to define some of the complexity of the disorder by sub-grouping patients according to their symptoms and by characterizing autistic traits in their parents and other family members. Folstein traces this idea back to her twin studies in England, when she noticed that co-twins and other family members had subtle symptoms of autism—compulsions, language problems, or social awkwardness.
“We realized that maybe we shouldn’t be looking for genes for autism, but maybe we should be looking for genes for the kinds of compulsions, for example, that you see in patients with autism and also in their parents,” Folstein says.
She and Dr. Joseph Piven, professor of Psychiatry at the University of North Carolina, divided their group of multiplex families into those in which the autistic children had very poor language and those who did not, and then they considered parents “affected” if they had a history of language problems. They thought it might be possible to improve the linkage signal—the finding that affected individuals share a particular chromosomal region—and that’s what happened, Folstein says. Other investigators soon followed suit.
For example, James S. Sutcliffe, Ph.D., assistant professor of Molecular Physiology & Biophysics at Vanderbilt, and his colleagues recently demonstrated that linkage to a region of chromosome 15 improved, or got stronger, in a subset of autistic patients with savant skills—extraordinary abilities in areas such as rote memorization, calculation, and mechanical achievement. Duke University investigators led by Margaret A. Pericak-Vance, Ph.D. found stronger linkage to the same chromosomal region in a subgroup of autistic patients who exhibit repetitive compulsions and extreme difficulty with changes to their daily routine.
“It seems like every time we use one aspect of the autism phenotype, one or another of the chromosomal regions that we suspect tend to make themselves better known, give better signals,” Folstein says. “This strategy is allowing us to disentangle the condition into its component parts, which we hope have a connection with the component genes.”
Combining this approach with the pooled resources of the Autism Genome Project may offer the best hope yet for making sense of the genetics of autism.
“The number of families we’ll be looking at collectively is so large that we will finally have the statistical power to ask these kinds of questions that essentially come down to statistics,” Sutcliffe says. “It will ultimately be much easier to find the genes, and I would be surprised if within the next two years, someone hasn’t identified the first autism gene.”
The hope, these genetics researchers agree, is that finding the genes that cause autism or increase an individual’s susceptibility for the disease will improve diagnostic capabilities and pave the way for new biologically-based treatments, perhaps even preventions.
A problem with wiring
Proceeding in lockstep with the search for autism genes have been efforts to understand the neurobiology of the disorder. Attention has focused on defining the brain regions affected by autism, with the hope that knowing which brain regions are affected and how will guide diagnosis and treatment strategies.
Autopsies and brain imaging have revealed that brains of individuals with autism are larger than normal, on average, and that there are alterations in the brainstem, cerebellum, and “limbic” structures, like the amygdala and the hippocampus, which are involved in emotional processing. But no clear picture of an “autistic” brain has emerged.
“Could you hand a CT scan to a neurologist and say, ‘Does this child have autism based on your knowledge of structure?’ The answer is no,” says Stephen M. Camarata, Ph.D., deputy director of Research on Communication and Learning at the Vanderbilt Kennedy Center for Research on Human Development. “Clearly something is wrong in the brain, but we suspect it’s not going to be a gross anatomical difference. More likely, it’s going to involve interactions among different areas of the brain and how these areas integrate information.”
Pittsburgh’s Minshew calls autism a disorder of complex information processing. “People with autism can hear and remember information, but they have trouble making sense of it,” she says. “It’s a generalized brain phenomenon where complex circuitry and higher order cognitive abilities supported by that circuitry fail to develop.” She points out that all areas of the brain—including those controlling skilled motor movements—are affected.
Minshew and her colleagues first observed the disconnect between basic skills and higher order skills during behavioral testing. They have recently confirmed, using functional magnetic resonance imaging (fMRI), that basic brain circuitry, but not higher order circuitry, is intact in individuals with autism.
Brain development during the first two years of life sets the stage for the complex circuitry and information processing capabilities of the mature brain. An acceleration of neuronal growth during this critical period, and/or a failure of the mechanisms that normally “prune” away unnecessary neurons, could lead to disarray, Minshew says.
Using neuroanatomy and powerful new techniques like fMRI to get at the regions of brain dysfunction in autism is part of a progressive search for the cognitive and brain basis of behavior, Minshew says. “How well we understand behavior makes an enormous difference in how well we can intervene,” she says. “We are most effective at changing behavior when we understand why someone’s doing what they’re doing.”
Finding what works
Behavioral and educational interventions for children with autism have come a long way since the days when diagnosis was accompanied by a suggestion for institutionalization, but even now it is unclear which treatment will work best for a given child.
“If a parent comes to me and asks me which therapies she should use, I have no way to answer that question,” says Paul J. Yoder, Ph.D., professor of Special Education and an investigator in the Vanderbilt Kennedy Center. “It would be useful to know which treatments might be counterproductive to use together, or which ones might be synergistic. And we’re not just talking about educational treatments, it’s pharmacological treatments too.” About 50 percent of children with autism take medication to lessen anxiety or control serious behavioral disturbances like self-injury, aggression and hyperactivity.
Yoder and fellow Kennedy Center investigator Wendy L. Stone, Ph.D., professor of Pediatrics and Psychology, are collaborating on a study to compare two behavioral treatments. Complicating this kind of research, Yoder says, is the fact that children are often involved in multiple types of educational intervention in addition to the research study, making it difficult to control all the variables. “We don’t have good measures of the quality of those other treatments,” he says.
Despite the uncertainty about which treatments might work best, educational and behavioral interventions are effective, and the sooner they’re started, the better. “The brain is more plastic at young ages,” Stone says, “and I think there has been enough intervention research to suggest that young children who get intervention early have better outcomes than when the intervention is started later.”
To that end, Stone and collaborators have developed and are studying the STAT, Screening Tool for Autism in Two-year-olds, and they’re attempting to make it work for even younger children. The advantage of the STAT over traditional diagnostic tests, Stone says, is that it takes less time and doesn’t require advanced psychological training to administer. Because it can be more available to people in the community, “it promotes an awareness of what the early signs of autism are—the early social communicative deficits that everybody should be looking for.”
The STAT also offers the advantage of guiding treatment decisions by allowing the tester to directly interact with children and observe their strengths and weaknesses. “You can go right from the information on the STAT to designing appropriate educational activities for these children,” Stone says.
Eventually, treatment guidance may come from genetic and neuroimaging profiles, Yoder says, but those days are a long way off. “It’s exciting to me that as a nation, we’re finally spending serious money on learning how to treat children with autism, how to educate them, what to do with their day-to-day moments,” he says. “There are no easy solutions to this problem.”
This summer, Morgan Vice spent three weeks at a camp run by TRIAD, the Treatment and Research Institute for Autism Spectrum Disorders at Vanderbilt. The camp is one type of intervention effort: When school’s out, it keeps children with autism on a schedule and focuses on teaching them social skills—greetings, compliments, conversations. Now in its third year, the camp is seeing tremendous gains among its campers, says director Misty Ballew.
The slogan printed on this summer’s camp T-shirts said, “I believe I can fly,” and although it had no relationship to Morgan’s toddlerhood experiences, the phrase seemed especially fitting. Now nine, Morgan is “coming back, bit by bit,” Tammy Vice says. “When all this happened, my goals for Morgan did change,” she says, “but my dreams for her are the same. I want her to be happy and successful, where she is.”
Fly, Morgan, fly.