Researchers Look for Genetic Clues in Autism Researchers scan the genomes of the largest collection of families with multiple cases of autism ever assembled, looking for common factors. Researchers in the journal Nature Genetics reports on the initial stages of a project that is hunting for genetic factors in autism The analysis found a region on chromosome 11 that appears to be shared by many of the individuals studied.
NPR logo

Researchers Look for Genetic Clues in Autism

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript
Researchers Look for Genetic Clues in Autism

Researchers Look for Genetic Clues in Autism

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript


This is TALK OF THE NATION Science Friday. I'm Ira Flatow. Autism rates are on the rise, according to numbers released this month by the Centers for Disease Control and Prevention. The disease may affect one out of every 150 children. That is up from about one in 160 in the last survey period.

Researchers aren't sure why autism rates are rising because they don't exactly know what causes autism. There have been lots of suggestions, ranging from mercury in vaccines to better detection of the disease. The one area that researchers feel will yield a trustworthy clue is in the genetics of the people with the disease.

The case for autism as a hereditary disease, at least in part, got a boost this week. An international team of researchers, part of the Autism Genome Project, published a paper in the journal Nature Genetics that identifies genetic links to the disease, though it did fail to find what you would call or what could be called a single autism gene.

This hour, we're going to take about the new autism numbers, what they mean, how genes may contribute to the disease, along with environmental factors. If you'd like to give us a call, our number is 1-800-989-8255, 1-800-989-TALK. And as always, when we talk about diseases on Science Friday, it's just unprofessional for us to try to comment - or our experts to comment on an individual case that the disease - that you may have or someone you know may have. So please don't ask us to diagnose or talk about it.

Rowland Barrett is the director for the Center for Autism and Developmental Disabilities at the Bradley Hospital in East Providence, Rhode Island and associate professor of psychiatry and human behavior at Brown University Medical School. He joins us today from his office. Welcome to the program, Dr. Barrett.

Dr. ROWLAND BARRETT (Brown University Medical School): Thank you, Ira.

FLATOW: You're welcome.

Dr. Ed Cook is a visiting professor of psychiatry at the University of Illinois at Chicago and a participant in the Autism Genome Project. He joins us today from his office there. Welcome to the program, Dr. Cook.

Dr. EDWIN COOK (University of Illinois at Chicago; Autism Genome Project): Thank you, Ira.

FLATOW: Welcome. Dr. Rowland, Dr. Barrett, let's talk about the CDC study. Give us an idea - help us understand what the numbers mean.

Dr. BARRETT: Well, the most recent CDC report, which was released just this month, in February, sought to improve our estimates of how many children in the U.S. are afflicted with autism and closely related autism-spectrum disorders such as Asperger's Disorder.

The study examined a total of 2,685 children, all of whom were 8 years old. They were from 14 different states, and the CDC researchers, based on reviews of health records and education records, found that on average, one in 150 of the children who were studied showed behavior that was characteristic of and consistent with a diagnosis of an autism-spectrum disorder.

So why is this news, I guess you're asking. Well, likely because for many years, it was thought that the prevalence of autism was approximately one in 2,000. However, about seven years ago, a few epidemiologic studies appeared that indicated the prevalence to be far greater than one in 2,000, and one in 500 was reported.

And subsequently, the CDC partnered with the American Academy of Pediatrics in 2004 and found rates somewhere in between one in 166 and one in 500, but the media picked up on the higher estimate which created concern about a so-called autism epidemic. And subsequently, this study also shows higher prevalence rates than what we thought before.

FLATOW: You said so-called autism epidemic.

Dr. BARRETT: Yeah. I think that, you know, an epidemic medically has a number of different definitions. And it - I guess it can be defined fairly simply as a disease or a disorder that's excessively prevalent. I think that in the case of autism, when we never knew what the true prevalence was, it's hard to say that things are epidemic once we put them under heavier scrutiny.

FLATOW: So you're saying it could have always been this high, we just missed it?

Dr. BARRETT: Well, I mean that to a certain extent, that may be true. But one of the most-important features of the increase in the autism numbers has been a change in nosology. And by that I mean specifically broadening the definition of autism to include closely related groups of autistic-like behavior, such as pervasive development disorder and Asperger's Disorder - which are now included in the data sets and also included in the most recent CDC report that's used to calculate prevalence.

FLATOW: Let's talk about the biological pathways that doctors think could cause the disease. What might be some of them?

Dr. BARRETT: Well certainly in terms of genetics, which I'm sure we'll talk about a great deal. I'm very interested in hearing about that study myself. But you know, there is an existing notion that there are multiple pathways to autism. There's always been a strong, powerful link in terms of genetics that autism is seen as running in families. However, folks have also looked at cytomegalovirus and PKU, a metabolic disorder; and even toxoplasmosis, a bacterial infection that may result in behavior that either, you know, that mimics autism. It looks like autism.

Again, the difficulty in diagnosing autism is that we don't know what autism is, we only know what it looks like. We're absent a biological marker, which is why this Autism Genome Project is so important.

FLATOW: Before I get to the genome project and talk to Dr. Cook, I want to ask you one more question about the continuing drumbeat of people who believe that the mercury preservative in some vaccines, thimerosal, is linked to autism. Any more evidence one way or the other on that?

Dr. BARRETT: Well you know, it's a long story with respect to thimerosal. It started back in the 1930s. It's a preservative, an ethyl-mercury preservative, that is used in vaccines to kill bacteria and fungi. About 15 years ago, it came under increased scrutiny when the EPA revised and lowered the reference doses for safe levels of mercury in the food chain, based on studies in the Faro Islands, where it appeared that neurodevelopmental outcomes in children were being adversely affected by the intake of organic mercury by their mothers.

And so in the U.S., they expanded this to concerns about ethyl mercury, which is a metabolite of mercury. It's used as preservatives in vaccines, thimerosal being one. And subsequent to establishing some studies in around 2000, 2001, they were unable to conclude what a safe level of ethyl mercury was. And so as a prudent measure, they have gone forward, and the Institute of Medicine and the EPA and the FDA have gone forward such that all newly licensed vaccines after 1999 have not contained - must be thimerosal free, and now there are only trace amounts.

Subsequent to this controversy about thimerosal and its link, the Institute of Medicine has concluded on the basis of several international and large-scale studies that there is no causal mechanism.

FLATOW: Ed Cook, let's talk about this, your study, which is looking for a genetic basis for the disease, for autism. Tell us about - why look for genes that contribute to autism? Is there such a good family history here?

Dr. COOK: Well, autism has drawn many of the world's premier geneticists to it because it looks so promising from a scientific opportunity perspective. Obviously, for those of us who are working with families with people with autism and that, that is a wonderful thing. So why would all these geneticists that could be studying something else be drawn to autism? The findings that lead to the interest are that if two parents have a child with autism, then the risk that the next child will have autism is one in 20. And we can put that in the context of the one in 500 number for autism. The one in 150 number relates to the entire autism spectrum. The risk for a second child or a subsequent child - each subsequent child - having autism spectrum is probably closer to 10 percent.

So just focusing on narrow autism, the increased risk to siblings has gone from one in 500 in the population to one in 20, which is a minimum estimate of a 25-fold increase in risk to siblings. We put this in the context of disorders such as Type One Diabetes where the risk is about 15 to one and where the hunt for genes has been quite successful. That's interesting, but it's not sufficient because obviously the same language being spoken runs in family as well. Most compelling evidence is in the twin studies, in which if for identical twins if one has autism, the risk that the other identical twin has autism and narrow autism is now 60 percent.

So we've gone from a risk from the population of one in 500 to the sibling of one in 20 - to now three in five. Now, OK, well that's fine. If it was a shared environmental effect then the non-identical twins or the fraternal twins would also have a three in five risk. But I think what's most interesting in why estimates of inheritability for autism end up in the over 90 percent range is that the non-identical twin of a child with autism is - has actually been estimated at lower than the sibling prevalence. Now I doubt that it's somehow protective. It's probably just an underestimate. But basically, there's no increased risk to the fraternal twin compared to the sibling.

This very low rate in the identical twins is actually a very strong indicator of genetic factors.

FLATOW: All right. We're going to take a short break and come back and talk more with Ed Cook and Rowland Barrett, take your phone calls about autism, delve a little bit more into this genetic connection and what kind of genes have been isolated and what might be the future of that study. So stay with us. We'll be right back after this short break.

(Soundbite of music)


(Soundbite of music)

FLATOW: You're listening to TALK OF THE NATION: SCIENCE FRIDAY. I'm Ira Flatow. We're talking this hour about autism with my guests, Dr. Ed Cook, visiting professor of psychiatry, University of Illinois at Chicago and a participant in the autism genome project. Dr. Rowland Barrett, Director of the Center for Autism and Developmental Disabilities at the Bradley Hospital in East Providence Rhode Island, also associate professor of psychiatry and human behavior at Browning University Medical School. Dr. Cook, I know Dr. Barrett is eager to hear more about this autism genome project. You collected many, many samples over 1,500 families. Is that correct?

Dr. COOK: That's correct.

FLATOW: And you found genetic evidence. Tell us about that genetic evidence. What did you find?

Dr. COOK: Yes, we found evidence on Chromosome 11 where the two children were sharing having autism. They were also sharing markers in a particular region on Chromosome 11. And this is going to require follow-up of finding out what specific variation and what specific genes underlie that peak of linkage on Chromosome 11.

FLATOW: So you haven't found a smoking gun, so to speak - a single gene?

Dr. COOK: We're - we're - we have - we're in the neighborhood and we are now going on a house-by-house search.

FLATOW: Have any houses turned up anything?

Dr. COOK: Well there's some interesting houses in the neighborhood, so we're particularly interested, for example, that right under the peak is a glutamate transporter gene. And we're beginning to pick up some clues that perhaps there may be some theme to some of the various clues that we're picking up in terms of glutamate signaling at neuronal synapses. The other big clue that we found in looking for places where there were holes in the genome or increased pieces of the genome, was a family that had a deletion in one of the two copies of the neurexin-1 gene. The neurexin-1 gene is very interesting because it helps to act to interface with Neuroligon, which had previously been shown to contribute to autism in two families even though these are uncommon events and are not happening in most of the families. They're starting to give us clues to help us understand mechanism of how autism may arise.

And we're focusing on these molecules such as neurexin and neuroligin that contribute to the development of synapses and connections between neurons.

FLATOW: Do you think that eventually all cases of autism then will be eventually be shown to have a genetic basis, I mean, if you just identify the gene?

Dr. COOK: I would expect that it would be likely that all cases would have genetic influence. They're going to fall into differing ranges in terms of whether there are large hits such as what we confirmed in this study the existence of four million base pairs of extra material in the Prader-Willi and Angelman region of Chromosome 15. Those may be the equivalent of pinheads in the system. Many others will have ten smaller risk variance scattered throughout the genome.

And then as I told you before, the inheritability is not 100 percent. So there will probably be a range of how much environmental contribution is also coming in. But I would expect even with environmental contribution, that the genetics are so strong that the risk factors would be interacting. As a matter of fact, we're hopeful that by understanding the genetics better that it will give us a clue into how to approach whatever environmental variance we know is out there.

FLATOW: Can you guess what an environmental influence might be? What kind of, you know, environmental factor?

Dr. COOK: I'm not too good at guessing.

FLATOW: That's good enough for me. Not going to put you on the spot. 1-800-989-8255 is our number. Let's see if we can take a phone call or two. Jeff(ph) in Wooster(ph), Ohio. Hi Jeff.

JEFF (Caller): Hi Ira.

FLATOW: Hi there.

JEFF: Love your show. Been a longtime listener.

FLATOW: Go ahead.

JEFF: Thanks for the show today though. I might - it validates my thoughts on that it is a predisposition to autism - genetic predisposition. My son was diagnosed with autism around two years ago, typical, back in '94, '95, and the issue we had was being diagnosed with autism but then where do you get treatment. So it wasn't until he was 11 - he was nonverbal, or he became nonverbal - completely nonverbal - and then we found the Cleveland Clinic Center for Autism in Cleveland, Ohio and then they began treatment with him and within three months he was signing a song in church, a capella.

So, but that's my concern. And now, of course, he's been in the school for a little over three years now with a great success. But my concern was not only the treatment. You know, we can - like a lot of parents who have children with autism, they're diagnosed, but what's the next step as far as, like, treatment, you know, education. But, you know, what do we - you know, diets? Is this an allergic reaction? Can we remove something from their diets to help them along and help those - help their functioning. So I'll take my answer off the air, or your comments off the air. Thanks.

FLATOW: Thanks for calling. Dr. Barrett, any comment?

Dr. BARRETT: Well, I think that this is not an unusual - not an unusual question. I think that number of parents have the same experience as the caller. And there is a broad consensus that involves the National Academy of Sciences, the American Academy of Pediatrics, child psychiatry, American Psychological Association, among many others, that early and intensive intervention is critical to obtaining a favorable clinical and developmental outcome for these children. However, comprehensive early treatment regiments are very expensive. That would cost about $60,000.00 a year and they're not covered by private insurance. And what I mean by early intervention protocol: The ones that have been proven to be successfully involved anywhere from 20 to 30 to 40 hours a week of treatment, that's delivered in the home, by the way, and involves family members in an integral and critical manner.

In contrast, most states run early intervention programs but most sates allow only for one or two hours of speech and occupational therapy per week. So the options that families have are very limited short of some movement that includes mandating private insurance companies to cover early and intensive therapies that are prevent to be effective, or expanding the autism waiver and autism Medicaid waiver in states that will allow families to participate regardless of their level of income. You know, Ira, the methodology to improve the qualitative of lives of these children and their families through this early and intensive intervention is available. Right now, though, it's a matter of to whom it's available and for how much. And this has to change, if only because in the long run it's going to prove much cheaper, both in terms of the human toll as well as the actual dollars.

FLATOW: Let's go to Charlie in Middletown(ph). Hi Charlie.

CHARLIE (Caller): Hi, thank you. I guess what I want to know is - it may have been covered earlier but I wanted to get it more strictly: I gather that there's something called narrow autism and then there's the autism spectrum. Did I get that right?

FLATOW: Dr. Barrett?

Dr. BARRETT: Yeah, actually you did get that right. Prior to 1994, autism was diagnosed on the basis of criteria put together in 1943 by Dr. Leo Kanner from Johns Hopkins. And subsequent to 1994, closely related groups of children who presented autistic-like behaviors such as were diagnosed as pervasive - having pervasive developmental disorders and Asperger's Disorder - were brought in to the group. So what has to be taken into account here in looking at the prevalence controversy is that the group of children that were allowed into the data steps using the broadened definition of autism had a much higher prevalence rate to begin with then classic autism.

Broadening the definition was not a matter of bringing in a few more kids who were on the cusp of earning a diagnosis of autism. Instead, a very large group was brought into the mix, a group that by some estimates was at least 300 percent - or up to 500 percent larger than the original narrow group that was defined by the original diagnostic category of autism.

CHARLIE: Well okay, that's really - I mean, I was kind of guessing that, but I'm wondering - I'm glad to hear you explain that. What I really would like to know is, I mean, not that you want to - I mean, obviously, you want to treat people with Asperger's Syndrome or whatever they have, but has the prevalence of narrow autism expanded since, say, 1943 - and if so, how much?

Dr. COOK: Well, that's hard - that's a very good question and it's hard to answer because one of the things that happened, for example, was that prior to the early 1990s autism was not a legitimate - what I meant by that, I mean, federally subsidized special education diagnosis.

So prior to - in 1993, there were about 23,000 kids in special education with -or 1994 there were 23,000 in special education with an autism diagnosis. Ten years later after the expanded definition, there were 193,000 kids in there with - in special education with an autism diagnosis although it was expanded. How many of those kids actually had more narrowly defined diagnosis of autism but had been labeled with mental retardation in order to gain the services prior to 1994 and then switched over makes manipulating the data set hard to interpret.

FLATOW: Thank you, Charlie. Dr. Cook, where do you go from here? What do you do with this genetic information at the clinical level? Do you eventually expect to screen children - or prenataly - for possibly having autism?

Dr. COOK: Well, we already know that they are relatively uncommon but some in the one percent or more range findings in genetics that can be tested for by those that would like to be tested even before conception, such as Fragile X disorder, such as duplications of chromosome 15 K11, K13 - The Prader-Willi, Angelman region and can be tested prenataly and can be tested afterwards.

The question is how much - what will that percentage be eventually? Will it remain in the 5 to 10 percent range or will it increase further? My guess is it will increase further but will not ever reach the point at which we'll be able to prenataly diagnose all cases of autism or even have very what you would think of as a single gene test, which certainly keep coming back to this theme of multiple genes, even for the ones that are more major hits.

Again, one of the things that this study found is that we did not find a new small recurring, say, five to 10 percent place were below the microscope level. We are finding deletions or increases in copy number.

FLATOW: Talking about autism this hour in TALK OF THE NATION: SCIENCE FRIDAY from NPR News.

Which tells you what, Dr. Cook?

Dr. COOK: Oh, by not finding those?

FLATOW: Yes, yeah, for those of us who don't know what that means.

Dr. COOK: Well, it tells us that there were likely still be a large amount of autism in which it relates to about 10 or more genes conspiring to cause autism together. And that's a very interesting point that it raises because it means that if it takes 10 genes all to be risk variants to cause autism, number one, they maybe relatively common in the population. And it also suggests that when they're not conspiring the cause severe autism they may be contributing much of the diversity that makes our species so wonderful.

So you have this contrast between something like Fragile X, which is equal to at least eight out of the 10 hits compared to other things where be lots of little risk variants.

FLATOW: Dr. Barrett, do you think that number 150 is the bottom line?

Dr. BARRETT: Yeah, actually, I think that I'm more, you know, it's - there's a lot of overlapped. Part of the problem with, I mean, there's some very good points to the CDC study, you know. For example, it's going to be useful in the future when we go back to those 14 communities and then reevaluate. But it doesn't give us a national prevalence rate, and in some areas, it was considerably lower - the prevalence was considerably lower.

And, you know, again they did this archivally. They didn't actually see the children to the best of my knowledge that they depended on records and it made the diagnosis based on behaviors that are the characteristic of the disorder.

Now, those behaviors are still disabling and those kids still needs special education but often times when you're doing research like this that's archivally based, you end up with a spurious lining. If you went solely on a basis of the number of kids in the United States right now who receives special education services and with the diagnosis of an autism spectrum disorder, the prevalence rate is one in 336 but that does not include kids who were younger than three years old who might have the disorder.

FLATOW: So we - you need a more accurate survey, you're saying?

Dr. BARRETT: Oh, yeah, we're just starting. We're just starting this and all information is welcome.

FLATOW: Anything from scratch that would create new records instead of looking back on the old ones.

Dr. BARRETT: I'm sorry, I…

FLATOW: Any research that's going forward that might start from scratch creating new records depending on maybe unreliable older ones.

Dr. BARRETT: Oh, yeah, there are number of studies that are going to be doing face-to-face interviews with, you know, inter-rator(ph) agreement among clinicians who will be looking to try and establish or - but that's a long and tedious process. And what we're trying to do is get more quickly to do a number, which is why this procedure within these 14 different communities in United States, although not representative of the national prevalence rate, is going to be useful as a baseline for further studies.

FLATOW: (Unintelligible)

Dr. BARRETT: I have a question for Dr. Cook…

FLATOW: You're going to have to hold it because we run out of time. Give him a call.

Dr. BARRETT: Okay.

FLATOW: Okay. I want to thank Rowland Barrett, director of the Center for Autism and Developmental Disabilities at the Bradley Hospital, East Providence, Rhode Island, associate professor of psychiatry and human behavior at Brown University Med School; Ed Cook, visiting professor of psychiatry in the University of Illinois at Chicago and participant in the autism genome project. Always like to bring scientists together to chat with each other. Thank you, gentlemen, for taking time to be with us.

Dr. COOK: Thank you.

Dr. BARRETT: Thank you.

FLATOW: We're going to take a short break. After we come back, we're going to talk about searching for killer asteroids. Stay with us.

I'm Ira Flatow. This is TALK OF THE NATION: SCIENCE FRIDAY from NPR News.

Copyright © 2007 NPR. All rights reserved. Visit our website terms of use and permissions pages at for further information.

NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.