It was an appropriately hazy afternoon on the day my son Dylan, age two years, seven months, and sixteen days, was diagnosed with autism. The doctor was empathetic but serious when she said that he showed “symptoms consistent with an autism spectrum diagnosis.”
And, she added, Dylan will need 40 hours a week of intensive, one-on-one educational therapy for at least several years.
“Why?” was the only response I could muster to that bombshell.
Many of the “whys” of this condition were not well understood on that day in 2007. A small number of genes were known to be involved in commonly co-occurring conditions such as Fragile X and Rett syndromes, but they explained only a tiny fraction of the total number of people with autism. At that time, twin studies had shown that genetic factors accounted for 60% to 90% of cases. But if autism was largely genetic, where were the rest of those genes?
Significant progress is being made toward answering this essential question.
Much has been learned about the genetic underpinnings of autism and the extent of its heterogeneity. To date, scientists have identified more than 200 genes involved in autism and similar neurodevelopmental conditions. Most of these genes lead to autism primarily through de novo variants — variants that are new in the person and not present in their parents. Understanding what these genes do and how they function in the brain represents a foundational step toward grasping the biology of autism’s core symptoms in order to identify pathways to better therapies.
Over the past six years, I have worked with a team to build SPARK, a study supported by the Simons Foundation that has assembled the largest sequencing data set in autism to date: 35,000 people with the condition, along with thousands of their parents and siblings. Beyond identifying an increasing number of genes involved in autism, the study is also learning more about the kinds of genetic variants that cause autism in its many forms. The picture that is emerging is one of a heterogenous condition that is different in every person because the genetic factors involved differ in every person.
In a new study published in Nature Genetics on Thursday, the SPARK team examined more than 42,000 people with autism (individuals from SPARK plus previously published genetic data) and, in a first for autism genetics, identified four genes associated with autism that are mostly due to inherited variants. We looked for variants in genes that showed non-random patterns of inheritance from parents to children with autism and identified four genes in which inheritance of loss-of-function variants — those that lead to a non-working copy of the gene — were seen more than 50% of the time, but only among children with autism.
These results show that while loss-of-function variants in these four genes do not always lead to autism — they are present in parents and some siblings without autism — they are inherited at higher than expected levels in people with autism and therefore are contributing to the condition.
High concordance of autism in sets of identical twins support the hypothesis that there are inherited genetic variants that lead to autism, in addition to well-established de novo variants, but until now no one has been able to identify them systematically. The inherited variants we identified do not always cause autism, so we hypothesized that people with autism who have these inherited genetic variants are probably not as likely to have additional cognitive impairment, especially compared to people with autism with de novo variants of strong effect. So far, our data strongly supports this hypothesis.
We also found that 70% of autism caused by de novo variants can be explained by genes that have already been discovered. In other words, already identified genes explain 70% of autism cases attributed to de novo variants — a remarkable amount of progress since Dylan was diagnosed in 2007. Our research demonstrates just how important it is to continue to learn more about the genetics of autism, particularly because we think the new inherited variants are starting to explain a different part of the autism spectrum than de novo variants.
Three other studies recently published in Nature Genetics that used SPARK data are helping refine what’s known about the genetics of autism across the spectrum. A study led by the Autism Sequencing Consortium found that genes associated predominantly with developmental conditions — a group of conditions that lead to delayed cognitive and/or atypical physical development in young children — tend to be important in early fetal brain development, while genes predominantly associated with autism are more important in later stages of brain development.
Two other studies looked at the contributions of multiple genetic factors to autism spectrum disorder and showed that the wide variety of clinical presentations in individuals across the spectrum can be explained by the combinations of genetic factors they carry.
Although researchers have yet to identify the complete picture of brain molecules and pathways underlying autism, these new studies show that range of severity across the autism spectrum can be attributed to an array of genetic variants. As always in science, much more remains to be learned. My colleagues and I estimate that the SPARK study will need to include more than 100,000 people with autism, along with as many of their parents and siblings that we can recruit, to better understand the biology of autism across the entire spectrum. It must also include individuals from diverse ancestries to ensure that no one is left behind.
I may never know why Dylan loves a specific Howard Johnson’s off the New York Thruway, but it may someday be possible to understand the biological processes that lead to the intense repetitive interests that interfere with his ability to learn important life skills. That understanding may eventually lead to more effective ways to help Dylan and people like him to reach their full potential for independent and fulfilling lives.
In the future, I hope that everyone diagnosed with autism will be given clearer prognoses and better treatments grounded in a deep understanding of the genetics and molecular biology of autism.
Genetic research is essential to unlocking the complex biology — the whys — of autism. To get there, it will take thousands more people with autism and their families who, by simply spitting in a tube and contributing a DNA sample, will help provide the genetic information needed to answer these whys.
Pamela Feliciano is the scientific director of SPARK and a senior scientist with the Simons Foundation Autism Research Initiative.