Annual Report 2020

Searching for the Genetic Roots of Inherited Autism

Autism Research Initiative (SFARI)

Over the past decade, sequencing studies of children with autism and their families have uncovered about 100 high-confidence autism risk genes by searching for de novo variants — spontaneous variants that appear in the child but in neither parent. In one of autism research’s most striking success stories, these findings have provided an explanation for about two-thirds of the portion of autism risk that is associated with de novo variants.

Yet even though about 80 percent of autism risk comes from genetic factors, researchers estimate that only about one in five autism cases can be explained by a de novo variant. “That clearly isn’t the answer for the majority of individuals,” says Wendy Chung, the principal investigator of the Simons Foundation Powering Autism Research for Knowledge (SPARK) initiative. “There’s still something missing.”

When it comes to genetics (as opposed to environmental factors), the “something missing” consists of all the inherited variants that can increase autism risk. This vast portion of the autism risk landscape is much harder to chart than the portion that comes from de novo variants. Whereas an individual typically has just one or two de novo variants in the entire protein-coding portion of their genome, they will typically have tens of thousands of inherited variants, most of which will have nothing to do with autism. “There are so many that it’s hard to know which to pay attention to,” Chung says.

Pamela Feliciano and Wendy Chung review genetic data from SPARK.

Sifting through these variants to figure out which ones increase autism risk requires sequencing tens of thousands of families or possibly even 100,000 families. Until recently, such numbers were out of reach. But SPARK, launched in 2016, is in the process of assembling a cohort of 50,000 individuals with autism and their families. Sequencing studies of SPARK families are now starting to illuminate the inherited portion of autism risk. 

SPARK researchers analyzed the genomes of nearly 10,000 families from SPARK and other publicly available autism genomic data. Their analysis has uncovered the first gene, NAV3, that confers autism risk only through inherited variants and not through de novo variants. “We have very high confidence that this is a true autism risk gene,” says Yufeng Shen, a researcher at Columbia University who carried out the study with Chung and a large group of SPARK collaborators.

Lopsided Transmission. Examining the entirety of inherited autism risk is far too big a task to carry out without a much larger cohort even than the SPARK collection. To make the task more manageable, SPARK researchers restricted their attention to ultra-rare variants — ones that appear in at most 1 in 100,000 people in the general population. The variants that make the strongest contributions to autism risk tend to be these rare ones, since people with autism are less likely than others to have children, which means that their genetic variants are not as likely to be passed down to the next generation. 

Next, researchers narrowed down the list of ultra-rare variants even further by focusing on those that had been computationally predicted to be deleterious — variants that probably destroy or significantly alter the functioning of the protein that the gene encodes.

Rico and Israel Winston participate in the SPARK autism study.

To find the genes in this shortlist that are associated with autism risk, the researchers next looked for what they call a “disequilibrium” in how the gene’s variants are transmitted to the next generation. Since a parent has two copies of each gene, when one copy has a variant, each child has only a 50 percent chance of inheriting the variant. In the case of an autism-linked variant, the children who inherit it are much more likely to show up in the SPARK data than the children who don’t inherit it. This means that within SPARK, transmission of the autism variant will look lopsided: More than 50 percent of parents with the autism variant will have passed it down to their children with autism.

In the case of NAV3, 49 parents had a rare damaging variant, and 40 of them (81 percent) had passed the variant down to their children with autism — much more than 50 percent. “It’s very strong statistical evidence,” Shen says.

The researchers also found some indications of inherited autism risk for another gene, ITSN1. SPARK is now completing a genetic analysis of an additional 10,000 families, and researchers hope that these added data will bolster the evidence for NAV3 and ITSN1 and bring many other inherited autism risk genes to light. “As of today, we have more than doubled our genomic data since our first analysis,” says Pamela Feliciano, SPARK’s scientific director. “I’m confident that the data SPARK is generating will yield even more insights.”

Specific Pathways. The NAV3 protein, which is involved in neuronal migration, is part of a biological process that has long been known to be involved in de novo autism risk. Although this connection to other autism risk genes bolsters the evidence for NAV3, researchers hope that the search for inherited autism risk genes will ultimately lead them to some pathways that are more specific to autism than the ones that have come up in the context of de novo risk. Many of the known de novo risk genes are associated with autism that is compounded by intellectual disability and other neurodevelopmental conditions. Much less is known about the biology of autism in which only the core traits of the condition are present. “There’s a large part of the autism spectrum that we just don’t have answers for,” Chung says.

The genes underlying this type of autism may turn up more abundantly among inherited risk genes than among de novo risk genes. That’s because among people with autism, those who have only the core traits, without intellectual disability, are most likely to have children and pass their variants down to the next generation. Additional research is needed to determine if people with autism spectrum disorder who have inherited mutations in genes such as NAV3 and ITSN1 are less likely to have cognitive impairments.

“How many of the inherited variants we find might elucidate brand-new biology that might have greater specificity to autism?” Chung wonders. “It opens up the possibility that we’re just crossing the threshold into a new dimension.”