|
The Early Origins of Autism New research into the causes of this baffling disorder is focusing on genes that control the development of the brain. by Patricia M. Rodier Autism has been mystifying scientists for more than half a century. The complex behavioral disorder encompasses a wide variety of symptoms, most of which usually appear before a child turns three. Children with autism are unable to interpret the emotional states of others, failing to recognize anger, sorrow or manipulative intent. Their language skills are often limited, and they find it difficult to initiate or sustain conversations. They also frequently exhibit an intense preoccupation with a single subject, activity or gesture. These behaviors can be incredibly debilitating. How can you be included in a typical classroom if you can't be dissuaded from banging your head on your desk? How can you make friends if your overriding interest is in calendars? When children with autism also suffer from mental retardation -- as most of them do -- the prognosis is even worse. Intensive behavioral therapy improves the outcome for many patients, but their symptoms can make it impossible for them to live independently, even if they have normal IQ's. |
 |
||||
|
I became involved in the search for autism's causes relatively recently -- and almost by accident. As an embryologist, I previously focused on various birth defects of the brain. In 1994 I attended a remarkable presentation at a scientific conference on research into birth defects. Two pediatric ophthalmologists, Marilyn T. Miller of the University of Illinois at Chicago and Kerstin Strömland Of Göteborg University in Sweden, described a surprising outcome from a study investigating eye motility problems in victims of thalidomide, the morning-sickness drug that caused an epidemic of birth defects in the 1960's. The study's subjects were adults who had been exposed to the drug while still in the womb. After examining these people, Miller and Strömland made an observation that had somehow eluded previous researchers: about 5 percent of the thalidomide victims had autism, which is about 30 times higher than the rate among general population. When I heard these results, I felt a shock of recognition, a feeling so powerful that I actually became dizzy and began to hyperventilate. In the effort to identify autism's causes, researchers had |
long sought to pinpoint exactly when the disorder begins. Previous speculation had focused on late gestation or early postnatal life as the time of origin, but there was no evidence to backup either hypothesis. The connection with thalidomide suddenly threw a brilliant new light on the subject. It suggested that
|
The biological basis for autism, however, has been elusive -- an unfortunate circumstance, because such an understanding could enable researchers to identify the leading risk factors for autism and possibly to design new treatments for the condition. By examining the inheritance of the disorder, researchers have shown that autism runs in families, thought not in a clear-cut way. Siblings of people with autism have a 3 to 8 percent chance of being diagnosed with the same disorder. This is much greater than the 0.16 percent risk in the general population but much less than the 50 percent chance that would characterize a genetic disease caused by a single dominant mutation (in which one faulty gene inherited from one parent is sufficient to cause the disorder) or the 25 percent chance that would characterize a single recessive mutation (in which a copy of the faulty gene must be inherited from each parent). The results fit best with models in which variants of several genes contribute to the outcome. To complicate matters further relatives of people with autism may fail to meet all the criteria for the disorder but still have some of its symptoms. | ||
 |
||||
|
For embryologists, nothing tells us so much about what happened to an embryo as knowing when it happened. In the case of thalidomide-induced autism, the critical period is much earlier than many investigators would have guessed. Very few neurons form as early as the fourth week of gestation, and most are motor neurons of the cranial nerves, the one that operate the muscles of the eyes, ears, face, jaw, throat and tongue. The cell bodies of these neurons are located in the brain stem, the region between the spinal cord and the rest of the brain. Because these motor neurons develop at the same time as the external ears, one might predict that the thalidomide victims with autism would also suffer from dysfunctions of the cranial nerves. Miller and Strömland confirmed this prediction they found that all the subjects with autism had abnormalities of eye movement or facial expression, or both. 
| |||
|
||||
|
| The next logical question was, "Are the cases of autism after thalidomide exposure similar to cases of unknown cause, or are they different?" Aside from their behavioral symptoms, people with autism have often been described not only as normal in appearance but as unusually attractive. They are certainly normal in stature, with normal-to-large heads. The few studies that have tested nonbehavioral features of people with autism, however, have concluded that there are indeed minor physical and neurological anomalies in many cases, and they are the same ones noted in thalidomide-induced autism. For example, minor malformation of the external ears -- notably posterior rotation, in which |
the top of the ear is tilted backward more than 15 degrees -- are more common in children with autism than in typically developing children, children with mental retardation or siblings of children with autism. Dysfunctions of the eye movement had been associated with autism before the thalidomide study, and lack of facial expression is one of the behaviors used to diagnose the condition.
|
nerve but also has secondary effects on later brain development. That is, the injury to the brain stem might somehow interfere with the proper development or wiring of other brain regions, including those involved in higher-level functions such as speech, resulting in the behavioral symptoms of autism. Or perhaps the ear malformations and cranial nerve dysfunction are only side effects of an injury that we don't understand. whatever the true situation may be, the anomalies in patients with autism of unknown cause were much the same as the anomalies in the thalidomide victims with autism. The conclusion was clear: many cases of autism, if not all, are initiated very early in gestation. |
| The region of the brain of the brain implicated by the thalidomide study--the brain stem-- is one that has rarely been considered in studies of autism or in studies of other kinds of congenital brain damage, for that matter. On a simplistic level, neurobiologists associate the brain stem with the most basic functions: breathing, eating, balance, motor coordination and so forth. Many of the behaviors disturbed in autism, such as language, planning and interpretation of social cues, are believed to be controlled by higher-level regions of the brain, such as the cerebral cortex and the hippocampus in the forebrain. Yet some symptoms common in autism--lack of facial expression, hypersensitivity to touch and sound, and sleep disturbances--do sound like ones more likely to originate in the brain regions associated with basic functions. Furthermore, the most consistently observed abnormality in the brains of people with autism is not a change in the forebrain but a reduction in the number of neurons in the cerebellum, a large processing center of the hindbrain that has long been known to have critical functions in the control of muscle movement. One reason for scientists' confusion about the brain regions involved in autism may be that our assumptions about where functions are controlled are shaky. For example, the laboratory group led by Eric Courshesne of the University of California at San Diego has shown that parts of the cerebellum are activated during certain tasks requiring high-level cognitive processing. Another difficulty is that the symptoms of autism are so complex. If simpler behavioral abnormalities could be shown to be diagnostic of the disorder, researchers might have a better chance of identifying their source in the nervous system. |
In 1995 our research team had the opportunity to follow up on the thalidomide study by examining the brain stem of a person with autism. The tissue samples came from the autopsy of a young woman who had suffered from autism of unknown cause: she had died in the 1970's, but fortunately the samples of her brain tissue had been preserved. When we examined the woman's brain stem, we were struck by the near absence of two structures: the facial nucleus, which controls the muscles of facial expression, and the superior olive, which is relay station for auditory information. Both structures arise from the same segment of the embryo's neural tube, the organ that develops into the central nervous system. Counts of the facial neurons in the woman's brain showed only about 400 cells, whereas counts of the facial neurons in a control brain showed 9,000. Overall, the woman's brain was normal in size; in fact, it was slightly heavier than an average brain. I hypothesized that the brain stem was lacking only the specific neurons already identified -- those in the facial nucleus and the superior olive--and to test that idea I decided to measure the distances between a number of neuroanatomical landmarks. I was surprised to discover that my hypothesis was absolutely wrong. Although the side-to-side measures were indeed normal, the front-to-back measures were astonishingly reduces in the brain stem of the woman with autism. It was as though a band of tissue had been cut out of the brain stem, and the two remaining pieces had been knit back together with no seam where the tissue was missing. For the second time in my life, I felt a powerful shock of recognition. I heard a roaring in my ears, my vision dimmed, and I felt as though my head might |
explode. The shock was not generated by the unexpected result but by the realization that I had seen this pattern before, in paper that showed pictures of abnormal mouse brains.
What had altered the brains of these mice? It was not exposure to thalidomide or any of the other environmental factors associated with autism but the elimination of the function of a gene. These were transgenic "knockout" mice, engineered to lack the expression of the gene known as Hoxa1 so that researchers could study the gene's role in early development. The obvious question was, "Could this be on of the genes involved in autism?" The literature supported the idea that Hoxa1 was an excellent candidate for autism research. The studies of knockout mice showed that Hoxa1 plays a central role in development of the brain stem. Groups in Salt Lake City and London had studied different knockout strains with similar results. |
|
They found that the gene is active in the brain stem when the first meurins are forming--the same period that Miller and Strömland had identified as the time when thalidomide caused autism. Hoxa1 produces a type of protein called a transcription factor, which modulates the activity of other genes. What is more, Hoxa1 is not active in any tissue after early embryogenesis. If a gene is active throughout life, as many are, altered function of that gene usually leads to problems that increase with age. A gene active only during development is a better candidate to explain a congenital disability like autism, which seem to be stable after childhood.
Hoxa1 is what geneticists call a "highly conserved" gene, meaning that the sequence of nucleotides that make up its DNA has changed little over the course of evolution. We assume that this is a characteristic of genes that are critical to survival: they suffer mutations as other genes do, but most changes are likely to be fatal, so they are rarely passed on to subsequent generations. Although many other genes appear in several forms--for example, the genes that encode eye color or blood type--highly conserved genes are not commonly found in multiple versions (also known as polymorphic alleles, or allelic variants). The fact that no one had ever discovered a variant of Hoxa1 in any mammalian species suggested that my colleagues and I might have trouble finding one in cases of autism. On the other hand, it seemed likely that if a variant allele could be found, it might well be one of the triggers for the development of the disorder.
|
The good news is that we have identified two variant allels of HOXA1. One has a minor deviation in the sequence of one of the gene's exons, meaning that the protein encoded by the variant gene is slightly different from the protein encoded by the normal gene. We have studied this newly discovered allele in detail, measuring its prevalence among various groups of people to determine if it plays a role in causing autism. (The other variant allele is more difficult to investigate because it involves a change in the physical structures of the gene's DNA.) We found that the rate of the variant allele among people with autism was significantly higher than the rate of their family members who do not have the disorder and the rate among unrelated individuals without the disorder. The difference is much greater than would be expected by chance. The bad news is that, just as the family studies had predicted, HOXA1 is only one of many genes involved in the spectrum of autism disorders. Furthermore, the allele that we have studied in detail is variably expressed--its presence does not guarantee that autism will arise. Preliminarily data indicate that the variant allele occurs in about 20 percent of the people who do not have autism and in about 40 percent of those who do. The allele approximately doubles the risk of the developing the condition. But in about 60 percent of people with autism, the allele is not present, meaning that other genetic factors must be contributing to the disorder. To pin down those factors, we must continue searching for other variants in HOXA1, because most genetic result from many different deviant alleles of the same gene. Variations in other genes involved in early development may also predispose their carriers to autism. We have already discovered a variant allele of HOXB1, a gene on chromosome 17 that is derived from the same ancestral source as HOXA1 and has similar functions in development of the brain stem, but its effect in autism appears to minor. Other investigators are scrutinizing candidate regions on chromosome 15 and on another part of chromosome 7. Although researches are focusing on alleles that increase the risk of autism, other alleles may decrease the risk. These could help explain the variable expression of the spectrum of autism-related disorders. | Even a minimal understanding of the genetic basis of autism would be of great value. For example, researchers could transfer the alleles associated with autism from humans to mice, engineering them to be genetically susceptible to the disorder. By exposing these mice to substances suspected of increasing the risk of autism, we would be able to study the interaction of environmental factors with genetic background and perhaps compile an expanded list of substances that women need to avoid during early pregnancy. What is more, by examining the development of these genetically engineered mice, we could learn more about the brain damage that underlies autism. If researchers can determine exactly what is wrong with the brains of people with autism, they may be able to suggest drug therapies or other treatments that could ameliorate the effects of the damage. Devising a genetic test for autism--similar to the current tests for cystic fibrosis, sickle cell anemia and other diseases--would be a much more difficult task. Because so many genes appear to be involved in the disorder, one cannot accurately predict the odds of having a child with autism by simply testing for one or two variant alleles in the parents. Tests might be developed, however, for the siblings of people with autism, who often fear that their own children will inherit the disorder. Clinicians could look for a set of well-established genetic risk factors in both the family member with autism and the unaffected sibling. If the person with autism has several high risk alleles, whereas the sibling does not, the sibling would at least be reassured that his or her offspring would not be subject to the known risks within his or her family. Nothing will make the search for autism's causes simple. But every risk factor that we are able to identify takes away some of the mystery. More important, new data spawn new hypotheses. Just as the thalidomide results drew attention to the brain stem and to the HOXA1 gene, new data from developmental genetics, behavioral studies, brain imaging and many other sources can be expected to produce more welcome shocks of recognition for investigators of autism. In time, their work may help alleviate the terrible suffering caused by the disorder. |