“The genetics of autism are beginning to be consistent with a working model of cortex-to-cortex dysfunction,” said Daniel Geschwind, M.D., Ph.D., at APA’s 2013 annual meeting in San Francisco. “And we now think that a lot of what happens occurs prenatally, during brain development.”
Genetic studies on autism have proliferated in recent years, said Geschwind, a professor of neurology and psychiatry at the University of California, Los Angeles. More papers were published in the last six years than in the previous 30, he said. In addition, large-scale collaborations are under way, increasing the number of patients providing data. The Autism Genetic Resource Exchange began in 1998 and now contains data from 1,500 families, 6,000 individuals.
As a result, more than 50 genetic causes of autism have been discovered so far, and about 1,000 exomes (the coding parts of the genome) have been identified.
Thus, there is no single “autism gene.” Even the most frequent genes have an effect size of less than 1 percent, he said. “A hundred children could represent a hundred different cases.”
Daniel Geschwind, M.D., Ph.D., says that autism is part of a normal distribution on a continuum of human behavior, overlapping with other neurodevelopmental disorders of higher human cognition.
Autism is part of a normal distribution on a continuum of human behavior, he said. It is a clinically heterogeneous diagnosis, overlapping with other neurodevelopmental disorders of higher human cognition. Common variants have small effects, working not in strict Mendelian fashion but by increasing risk.
“Autism is a syndrome, not an etiology,” said Geschwind. “Autistic behavior is shared by dozens of other disorders, including Tourette’s syndrome, Fragile X, untreated PKU, and others. A large part of autism is a collection of rare diseases.”
He added, “We can talk about genes now, but we still can’t create a risk profile for an individual child.”
Geschwind hypothesized that polymorphisms disrupt normal functions of circuits in areas involved with implicit learning—where people learn language and social cognition.
As an example of current research, he described one mouse model now under study for its role in producing neuropathology and pathophysiology comparable to autism in humans. Mice with the CNTNAP2 gene knocked out exhibit normal motor coordination, but increased social dysfunction, repetitive behavior, hyperactivity, and hypersensitivity to sensory stimulation. They also have deficits in ultrasonic communication with cage mates.
These mice respond to treatment with risperidone, which reverses hyperactivity (as it does in some severe human autism cases), and reduces perseverative behavior and grooming (a mark of anxiety). The drug has no effect on social behavior, but that symptom does respond to oxytocin and vasopressin, he said. “We now think that a number of genes will affect the oxytocin system, so we may be able to identify a subset of patients who respond to oxytocin.”
Genes alone don’t account for autism cases, although how environmental influences interact with genes remains murky. Several risk factors are known or have been hypothesized. Inflammatory cytokines that cross the prenatal blood-brain barrier are a plausible, if unproven, possible source of abnormalities in synapto genesis. Advanced parental age shows an epidemiological relationship, increasing risk threefold from ages 20 to 50. Assortive mating may play a role. Prenatal folic acid supplementation appears to decrease risk.
Unknown aspects of life in the modern world may also play a role, just as the “hygiene hypothesis” appears to influence incidence of asthma, said Geschwind. “Whatever the environmental factor or factors, it must be fairly common around the world.”
Yet another line of attack makes use of systems biology. Advanced mathematical tools help researchers look not at genes alone but at network function and structure. One such study has found 17 modules—high-traffic locations analogous to airline hubs—that go up and down in brain development. Two modules contained rare genome variants and were up-regulated starting at 10 to 12 weeks in gestation and remained up through mid-fetal development.
“This is the period encompassing cortical development, covering the birth and fate of cortical projection neurons,” he said. Three other modules corresponded to synaptic development and function. These modules were not enriched in the siblings of autism patients, and intellectual disability genes were randomly distributed across the modules, suggesting a pattern specific to autism.
“Despite etiological heterogeneity, transcriptional profiling and integrative genomic analyses suggest common pathways on which we can capitalize genomic findings to create in vitro and in vivo models and use systems biology approaches to identify potential convergent molecular pathways,” said Geschwind. “We hope that knowing the mechanisms will lead us to biological pathways and thus to treatments.” ■