Genetic Factors in Autism

A concerned mother calls my office one morning for advice. “My sister just found out that she’s pregnant,” she says, “and I have a 6 year old son with autism. I’ve heard it’s genetic, is there a test she could have to find out if her baby will have autism?” Inquiries such as these are frequent for professionals involved in autism genetics research. The answers are far from straightforward, and a simple test for diagnostic purposes or prenatal detection is not available in most instances. Explaining the current state of affairs is a complex process, and one that requires some knowledge of the history and methods of autism genetics research.

Since 1977, when the first autism twin study demonstrated higher concordance rates of autism among identical twins than fraternal twins, the evidence for inherited factors in autism has gained widespread recognition among researchers. Based on several decades of work, scientists are now focusing in on specific chromosomal regions that are thought to contain autism-related genes, and are beginning to test candidate genes. The ultimate goal of this research, which is being carried out by researchers around the world, is to identify genes related to autism with the expectation that this information may lead to a better understanding of the disorder, its diagnosis, and its treatment.


In the first systematic and detailed autism twin study, conducted by Dr. Susan Folstein and Dr. Michael Rutter, the rate of concordance was compared between identical twins and fraternal twins. Concordance in this instance refers to the likelihood that if one twin has a diagnosis of autism, the second twin will also have a diagnosis of autism. Because identical twins share 100% of their genes, whereas fraternal twins share on average 50% of their genes, a higher concordance rate among identical twins is evidence for genetic influence. Dr. Folstein and Dr. Rutter found that the concordance rate for autism was significantly higher among the identical twins they studied, and subsequent twin studies have confirmed this finding. In general, the concordance rate for fraternal twins is similar to the 5-8% recurrence rate observed among non-twin siblings. Concordance rates among identical twins are estimated to be approximately 60%, but have been reported to be as high as 95%. The fact that identical twins are not always concordant for autism indicates that there may be non-genetic factors that are important as well, but the high concordance rates are strong evidence for significant genetic influence. The results of family studies, which have shown increased rates of autism among siblings and first degree relatives, are also an indication of the role that inherited factors play in the development of autism.


Evidence for an underlying genetic basis also comes from the many instances in which individuals with autism have been diagnosed with known genetic syndromes. It is estimated that 10-15% of individuals with autism have an underlying medical or genetic diagnosis. There is a known association between autism and fragile X syndrome, which is an X-linked genetic condition that more frequently affects males but may also affect females. Autism is also sometimes seen in association with tuberous sclerosis, a dominantly inherited condition that may lead to seizures, mental retardation, and unusual skin findings. There have been many case reports of individuals with autism who have chromosome abnormalities, most often involving chromosome 15. There have also been case reports of autism in association with neurofibromatosis type 1 (NF1), a dominantly inherited neurological condition, as well as case reports of autism in association with other genetic syndromes. Finally, researchers at Duke University recently reported that some individuals with autism have mutations in the MECP2 gene, which is the gene related to Rett syndrome. When evaluating the possible causes of autism in any individual child, a genetics evaluation should be considered and the above mentioned conditions ruled out. Both fragile X syndrome and MECP2 gene mutations can be tested for through DNA analysis, and the chromosome abnormalities frequently found in individuals with autism can be tested for through a high resolution karyotype and fluorescent in situ hybridization (FISH) for regions on chromosome 15. Tuberous sclerosis and NF1 are typically diagnosed through a physical exam, which includes a woods lamp exam of the skin.


In the majority of individuals with autism, there is as of yet no identifiable genetic cause. Based on all the evidence so far, researchers believe that autism is due to “complex” inheritance. Disorders that are due to complex inheritance do not follow the same predicted patterns of inheritance seen in dominant, recessive or X-linked disorders. Sometimes mutations in several different genes must occur in combination with certain environmental factors, such as exposure to certain chemicals or medications or possibly diet. This type of inheritance is often referred to as multifactorial because many different factors, genetic and/or environmental, are involved. It is estimated that as many as 15 different genes may be related to the occurrence of autism. It is possible that these genes may each have a small effect, in which case multiple gene mutations would be necessary for a child to develop autism. It is also possible that there may be several genes of major effect, but that the specific genes involved differ from family to family.

In order to determine the genes that may be involved, scientists perform what are referred to as “genome screens”. To do this they use maps of the chromosomes (similar to road maps) in order to look for genes. Just as gas stations or restaurants can be used as landmarks when locating a friend’s house, scientists use markers to find a gene. Markers are known regions or “sequences” of DNA along the chromosomes that may differ slightly from person to person or among populations. These differences, or “polymorphisms”, serve as landmarks that can be tested in individuals. In performing a genome screen, researchers look at many different markers throughout the genome, trying to find markers that are consistently found in family members who have a particular disorder, but not in family members without the disorder. These markers are landmarks that identify which chromosome a gene is located on (similar to which street a house is on). Statistical methods can tell a scientist how close these markers are to a gene. Testing additional markers will narrow the search area of the gene (similar to which block a friend’s house is on). Markers that are very close to a gene are said to be “linked” because the marker and the gene are almost always inherited together. Once scientists find a set of markers that are linked to a gene, then they say that they have found linkage. It is important to remember that linkage does not mean that a gene has been identified, but rather that the gene being searched for is somewhere nearby. There have been several published genome screens to date, and additional unpublished screens. Chromosomal regions of interest identified thus far include 2p, 4p, 6q, 7q, 13q, 15q, 19p, and Xq. Additional study will be needed to confirm linkage in these regions, and to narrow down the areas further.


Linkage results from genome screens tell us approximately where on a chromosome a gene is located. Researchers still need to determine the exact location of the gene (similar to finding a house on a particular street). One common method uses candidate genes, which are genes known through previous research to be localized to the region. A gene is called a candidate if the function of it relates in some way to the effect the disorder has on individuals who have the disorder. This laboratory technique is similar to knocking on the door of every house on a block until you find the one your friend lives in. Scientists test the candidate genes for mutations that might cause the disorder. If there are no mutations in the gene of a person who has the disorder, then that candidate gene could not have caused the disorder in that particular person. If all the candidate genes are tested and none are found to be responsible for the disorder, then the researcher studies genes whose functions are not yet known. Many genes may be tested until the correct gene is found. Then comes the long process of understanding how the gene works and why it causes the problems that it does. Recently a number of candidate genes have been under investigation. They include WNT2, RELN, and HOXA1. The WNT2 gene belongs to a group or “family” of genes that all contribute to the development of the central nervous system. The proteins produced by the WNT gene family are dependent upon the proper function of proteins from another gene family, referred to as DVL. A mouse knockout of one DVL gene, Dvl1, leads to features reminiscent of autism, consisting mostly of reduced social interaction. Because of this, and the fact that the WNT2 gene is located in the region of chromosome 7q where evidence for linkage has been found in genome screens, researchers are actively studying WNT2 as a candidate gene. The reelin protein (RELN gene) plays an important role in brain development and the RELN gene is also located on chromosome 7q, either close to or within the same regions where evidence for linkage has been obtained. Additionally, neuroanatomical differences in the brains of autistic individuals are in some ways similar to developmental alterations observed in the brains of “reeler” mice, who have no reelin protein due to the experimental introduction of mutations in the RELN gene. The HOX family of genes modulates other genes during embryonic development, and these genes are similar across different species. Mutations in the HOXA1 gene in mice have been reported to lead to abnormalities in brain development. Similar abnormalities have been reported in some individuals with apparent teratogen induced autism (for example, in utero exposure to thalidomide or alcohol). The HOXA1 gene in humans is located on chromosome 7p.


Getting back to the mother in our phone inquiry, it is clear that there is no simple answer to her question. An important consideration would be to determine whether her 6-year-old has ever been evaluated by a geneticist. Have the various genetic syndromes that are sometimes associated with autism been ruled out? If the answer is no, such an evaluation could provide information that might lead to the availability of specific recurrence estimate information, as well as prenatal diagnosis. If the answer is yes, and the mother’s 6-year-old has not been found to have an identifiable genetic syndrome, a careful review of the entire family history by a genetic counselor knowledgeable about autism would make it possible to give a rough estimate of the chance for recurrence. In such instances, genetic testing is not an option currently, because the many genes thought to be involved have not yet been identified. There is much that is not known at this point in time, but a spirit of collaboration among autism genetics researchers around the world will hopefully lead to new findings and significant improvements in our ability to diagnose and treat autism in the future.

*Based in part on the Duke University Center for Human Genetics web site:


Ongoing Autism Genetics Research at New England Medical Center
There are several ongoing studies at New England Medical Center. All focus on the identification of autism related genes, and therefore require blood specimen collection from participating family members.

· Collaborative Linkage Study of Autism (CLSA)
Principal Investigator: Susan Folstein, MD
The goal of this project, which is funded by the National Institute of Mental Health (NIMH), is to identify autism susceptibility genes for the purpose of improving diagnosis and treatment for autism spectrum conditions (autism, Asperger syndrome and PDD/NOS). We are primarily looking for families that have 2 individuals with autism spectrum diagnoses. This includes families with 2 affected siblings, first cousins, or an uncle (aunt)/nephew (niece) pair. Both individuals in the pair must be at least 4 years of age.

· Autism Language Project
Principal Investigators: Susan Folstein, MD, Helen Tager-Flusberg, PhD, J. Bruce Tomblin, PhD
In addition to the goal of identifying autism susceptibility genes, this project focuses on the study of language characteristics in individuals with autism spectrum conditions and their family members. We are looking for families in the New England area that have at least two children who are between the ages of 6 and 16, at least one of whom has an autism spectrum diagnosis. This project is being conducted in collaboration with researchers from the University of Iowa and Boston University, and is funded by the National Institute of Neurological Disorders and Stroke (NINDS).

· Discordant Sibling Project
Principal Investigator: Susan Santangelo, ScD
This project, which is funded by the March of Dimes, seeks to identify potential autism susceptibility genes by studying individuals with autism and their unaffected siblings. We are looking for families that have both a child with an autism spectrum diagnosis who is at least 4 years of age, as well as a child who is at least 16 years of age and who does not have an autism spectrum diagnosis.

Leave a Reply

Your email address will not be published. Required fields are marked *