© Thom Graves
Although arguments remain over whether autism is genuinely on the rise to the astonishing degree reported in places like California, there is general agreement among scientists that the condition has a genetic basis. The search for the underlying gene – or genes – is simultaneously further behind than anyone would like and further ahead than anyone expected.
Major progress has been made since the Cure Autism Now Foundation set up the Autism Genetic Resource Exchange (AGRE) in 1997. "That really pushed things into another realm," says Daniel Geschwind of the University of California, Los Angeles. More recently, the National Alliance for Autism Research (NAAR) created the Autism Genome Project, which has brought together more than 50 institutions, and includes the AGRE dataset. "The rationale behind [the Autism Genome Project] is the recognition that sample size plays an enormous role in the chance to find genes," says Joachim Hallmayer, chair of the executive committee. So far they have typed over 10,000 SNPs from over 1,500 families with multiple affected children, so there are now "billions" of genotypes, according to Hallmayer.
Nevertheless, research is complicated by the nature of the syndrome. Autism, and the broader "autism spectrum disorder," present a range of symptoms easily confused both with retardation and milder learning disabilities.
It's hard to search for a gene when there simply isn't a clear-cut phenotype. "Because of the variability in the phenotype, we know there's variability in the genetics," says Steve Scherer, senior scientist at Toronto's Hospital for Sick Children. And this causes arguments. For example, Scherer says that while he believes neuroligin could be a candidate in some rare cases, other researchers have countered that the family where the gene was found may have mental retardation rather than autism. There's also an overlap between Fragile X syndrome and autism, "[but] most of the people who have autism don't have Fragile X," says Gerard Schellenberg of the University of Washington.
Researchers haven't agreed on how to subdivide clinical presentations, although there is agreement that breaking the disorder down into subcomponents may be the most productive approach to finding candidate genes.
BEYOND SINGLE GENES
So far, much of the best published work has only been able to narrow down chromosomal loci. Definite candidate regions have been found on chromosomes 7 and 17; there are also potential regions on chromosomes 5 and 11, according to Geschwind, whose group found a fine-mapping peak at 17q21.1 "There is quite a bit of evidence for involvement of the serotonin transporter [found on chromosome 17] as a candidate gene," writes Cook in an E-mail.
Researchers at Rutgers University in New Jersey recently replicated findings that implicate the homeobox transcription factor
Other researchers implicate chromosomes 2, 15, and 16, according to NAAR Chief Science Officer Andy Shih and Vlad Kustanovich, researcher liaison for AGRE. Some researchers also consider the X chromosome a possibility, because of the 4:1 male/female ratio of presentation.
Geschwind says, "I think the field is no more than a year off from having very plausible reasons to make certain mouse models." But, he warns, "mice don't have big frontal lobes, so should this turn out to be a frontal lobe disorder, to what extent can one use mouse models to model the behavior?"
Nevertheless, Nobel Laureate Susumu Tonegawa at Massachusetts Institute of Technology is confident that mice should make good models, and he and his colleagues are already creating a mouse to address the Fragile X commonalities. While an