Genetics in Psychiatry: The Case of Autism

Thomas H. Wassink, M.D.
University of Iowa Hospitals and Clinics

First Published: Spring 2002
Last Revised: August 2003
Peer Review Status: Internally Peer Reviewed

History: Autism is a disorder of disturbed brain development that often leads to life-long mental disability. Its symptoms fall into three primary categories: 1) impairment in the development and use of language; 2) deficits in social interaction; and 3) ritualistic or repetitive behaviors. It affects nearly one out of every 1,000 individuals.

Autism is usually first noticed by parents or clinicians before two years of age when behaviors, such as onset of speech and social activity, would first appear in normally developing children. While the causes of autism still remain a mystery, it has become clear that genetic factors contribute significantly to the occurrence of the disorder. Siblings of autistic children, for example, are 25 to 50 times more likely than randomly chosen individuals to develop autism. While this could be due to a shared home environment, twin studies show that a genetically identical twin of someone with autism is also much more likely to have the disorder than a nonidentical (fraternal) twin. These studies and others also show that the genetic basis of autism is likely to arise from complex interactions of several genes, making identification of those genes difficult.

Some clues about potential autism genes come from epidemiological observations: 75% of children with autism have mental retardation; boys are affected four times more often than girls; 5% to 10% of patients have an identifiable chromosomal abnormality; and, in another 5%, the autism is attributable to a medical disorder with a known genetic basis, such as tuberous sclerosis or the fragile X syndrome. In most cases, however, no such objective links exist, and therefore, in our attempt to find genes, we rely on linkage and candidate gene studies.

New facts: Linkage studies, which involve families with more than one affected person, identify chromosomal regions that might contain disease genes. They take advantage of two properties of the human genome: polymorphisms and recombination.

All of the cells in the human body have two identical copies of the 22 nonsex chromosomes (autosomes). The full identity is only apparent, however. There are short stretches of DNA scattered throughout the chromosome that, instead of being absolutely identical on both paired chromosomes, are slightly different. These stretches are called polymorphisms. They are passed down from parents to children and can be detected (genotyped). Recombination refers to the exchange of genetic material across chromosomes at meiosis, so that the single copy of each chromosome that a parent passes on is actually a mixture of their two original chromosomes.

Using these genomic properties, a linkage study searches for a link between genotyped polymorphisms and inheritance of a disease through families. Hundreds of polymorphisms can be genotyped throughout the genome. Some of them will by chance be near the mutation that causes disease, while, either because they are on a different chromosome or because of recombination, the vast majority will not. Statistical analysis combines the genotyping data with the families' disease pedigrees to tell us where in the chromosomes a disease-causing mutation might be located.

Relevant findings: In a recent linkage study conducted by a research group at UI College of Medicine, 350 polymorphic markers in 75 families with at least two autistic children were studied, and two chromosomal regions that were linked (i.e., appeared likely to contain autism-related genes) were identified. One of these regions, located on the long arm of chromosome 7, had been identified in previous studies of both autism and isolated language impairment. We accessed the human genome sequence through the internet and used computational bio-informatics tools that we have developed to cull through and identify all the known genes in our region of interest. This region was broad, containing many genes. Therefore, we narrowed the field of candidates by examining their patterns of expression and function using information again made available through internet-based resources. We selected genes that were expressed in the developing brain and whose functions might plausibly be related to the deficits seen in autism.

One of these genes was named WNT2. There are more than a dozen WNT-type genes that are expressed during development in a variety of tissues. Previous studies in mice, frogs, and fish have demonstrated important contributions of various WNT genes to the development and patterning of the vertebrate central nervous system.

Further piquing our interest in WNT2 was a report describing a mouse engineered to be deficient for a gene called disheveled 1 (Dvl1). Transmission of the WNT signal is dependent on DVL proteins, and the Dvl1 knockout mouse exhibited strikingly reduced social interaction, characterized by a lack of huddling for warmth during sleep, the absence of normal interactive grooming of cage mates, and diminished mothering behaviors. Lastly, WNT2 was found immediately adjacent to RAY1, a gene that was interrupted by a chromosomal breakpoint in one autistic patient but that had otherwise not been shown to be involved in autism.

These findings led us to examine WNT2 as a possible autism-related gene. We did this by: 1) testing whether WNT2 was expressed in human brain cells; 2) screening the coding sequence for variations that would affect the constitution of the WNT2 protein; and 3) searching for additional polymorphisms close to WNT2 that we would use to do more powerful linkage-type analyses.

We obtained positive results for all three of these tests. We first showed that WNT2 was expressed in human thalamus cells. We then found nucleotide changes that altered the amino acid composition of the WNT2 protein in two out of 135 unrelated autistic individuals. When we examined WNT2 in the families of these two individuals, the changes (mutations) were shown to be transmitted from one parent to the autistic children but not to the unaffected children ( Fig. 1). Furthermore, the mutations were not found in 160 non-autistic individuals. Lastly, we identified two genetic markers very close to WNT2, and when we genotyped them, we found that one was in linkage disequilibrium with autism. Linkage disequilibrium is a screening tool related to linkage that helps to narrow linked intervals and hone in on disease genes.

While this converging evidence suggests that WNT2 might be an autism-related gene, we have interpreted it cautiously. Only two out of 135 families with autism had mutations in the gene, and the linkage disequilibrium, although real, was not dramatically strong. Numerous similar findings have been reported in autism for other genes, only to fail to be replicated or confirmed by other research groups.

We are, nonetheless, pursuing a variety of approaches to follow up these findings. First, we can insert copies of both normal and mutated WNT2 genes into test tube-based cellular systems and, by comparing them, determine whether the mutations affect the production or the activity of WNT2 protein. Second, we can design WNT2 knockout mice where the WNT2 protein deficiency is restricted solely to the brain and then see whether these mice have behavioral abnormalities related to autism. Third, we can examine whether any of the numerous genes that interact with WNT genes (such as the DVL genes) contribute to autism heritability. Lastly, we can examine the WNT2 gene in other samples of autistic subjects to look for similar changes to those we have reported.

All of these projects are now in progress and we eagerly await the results. The clinical implications of these data are, unfortunately, still limited. It is not appropriate yet, for example, to test autistic individuals for mutations in WNT2, nor do these findings suggest potential treatments. Rather, they primarily serve to guide our growing effort to untangle the mystery of this complex and tragic disorder.

Figure 1: A WNT2 Mutation in One Family.

See related Provider Textbooks about Psychiatry.

See related Provider Topics Autism, Child and Teen Health, Food, Nutrition and Metabolism, Genetic Disorders, Genetics/Birth Defects, Mental Health and Behavior or Psychiatry.

See related Patient Topics Autism, Child and Teen Health, Food, Nutrition and Metabolism, Genetics/Birth Defects or Mental Health and Behavior.

All contents copyright © 1992-2004 the Author(s) and The University of Iowa. All rights reserved.

http://www.vh.org/pediatric/provider/psychiatry/autism/index.html