The genetic mystery of why some people develop autism

BBC News

16-04-2025

Genetic factors are thought to play a major role in the development of autism – but for decades what they are has proven elusive. Now scientists are starting to uncover clues.
Until the 1970s, the prevailing belief in psychiatry was that autism was a consequence of bad parenting. In the 1940s, the Austrian psychiatrist Leo Kanner had coined the controversial "refrigerator mother" theory suggesting that autism arose from early childhood trauma, created by mothers who were cold, uncaring and rejected their children.
Daniel Geschwind, a neuroscience and genetics professor at the University of California, Los Angeles (UCLA), says that this is now rightly recognised as being deeply damaging and wrong – but it took the better part of three decades for Kanner's theory to be debunked. It was not until 1977, when a couple of psychiatrists carried out a landmark study demonstrating that autism often runs in identical twins, that a more nuanced and accurate picture of autism's origins began to emerge.
That 1977 study was the first time that a genetic component of autism had been identified. Research has since shown that when one identical twin is autistic, the likelihood that the other twin will be too can be more than 90%. Meanwhile, the chances of fraternal twins of the same sex each sharing a diagnosis of autism are around 34%. These levels are substantially higher than the typical rate of occurrence among the wider population, of around 2.8%.
It is now widely accepted that there is a strong genetic component to autism. But which genes are involved and how their expression is influenced by other factors are only just starting to be unravelled.
Tiny differences
Even after the twin study in 1977, it would take several more decades for the full subtleties of the interaction between autism and the human genome to become apparent.
Between any two individuals, the amount of genetic variation is around 0.1%, meaning that approximately one letter or base pair out of every 1,000 in their DNA will be different. "Sometimes these variations have no effect at all," says Thomas Bourgeron, a neuroscience professor at the Institut Pasteur in Paris. "Sometimes they have a little effect, and sometimes they have a super strong effect."
Currently, "super strong" variations have been identified in up to 20% of all cases of autism, with a single mutation in a single gene being largely responsible for driving critical neurodevelopmental differences. The role of these single gene mutations and how they arise is one of the most heavily studied areas in autism research, because as Bourgeron explains, they often result in severe and life-limiting disabilities.
"This is not like the autism you see in the movies," Bourgeron says. "If you're born with one of these major mutations, there's a high likelihood you'll end up with intellectual disability or motor delay [the ability to coordinate muscle groups] or epileptic encephalopathy. It has a major impact on their quality of life and their family in most cases."
So far scientists have identified at least 100 genes where these mutations can occur. Bourgeron himself made one of the first discoveries in March 2003 when he identified two gene mutations linked to autism. Each impacted proteins involved in synaptogenesis, the process of forming connections between neurons in the brain. It was a major breakthrough, although it barely made a ripple in the media at the time, with Bourgeron recalling how former US President George W Bush had recently declared war on Iraq.
But more discoveries were to come, including mutations in the Shank3 gene which are estimated to occur in less than 1% of people with autism. We now know that some of these mutations are known as de novo variants, which means that they occur through random chance in a developing embryo and aren't present in the blood DNA of either the mother or father. Geschwind describes de novo variants as being akin to a "bolt of lightning", that is both unexpected and rare.
However, in other cases, these mutations can have been passed on by one of the parents, even if both appear to be neurotypical, a more complex phenomenon which researchers have only begun to understand in the past decade.
"You might wonder, if an autistic child has inherited a rare gene mutation from one of their parents, why doesn't the parent have autism too?" says Geschwind. "What seems to happen is that in the parent, it's not sufficient to be causal, but in the child, that major gene mutation combines additively with other, less individually impactful gene variants to drive neurodevelopment differences," he says.
Of course, there are also thought to be environmental factors involved in the development of autism – even among identical twins where one has been diagnosed, 10% of the time the other one will not be.
Historically, identifying the environmental factors behind autism has led to pseudoscientific beliefs such as the idea – now widely debunked – that certain vaccines might be involved. Now the US Health Secretary Robert F Kennedy Jr has pledged a massive research effort to identify the causes of autism before September 2025. This includes hiring the vaccine sceptic David Geier as a data analyst at the US Department of Health & Human Services. The Autism Society of America have expressed concerns that the plans are unrealistic, as well as potentially harmful and misleading.
According to the US National Institutes of Health (NIH), potential non-genetic causes of autism include prenatal exposure to air pollution and certain pesticides, extreme prematurity, and birth difficulties leading to oxygen deprivation in the baby's brain, among other factors.
Early development
Today genetic research is leading progress into how neurodevelopment can lead to autism. It appears that many of these genes become functional during the formation of the cortex – the wrinkly outer layer of the brain responsible for many high-level functions, including memory, problem-solving and thinking.
This critical part of brain development occurs in the foetus as it is developing in the womb, and according to Geschwind, peaks somewhere between 12 and 24 weeks. "You can think of these mutations as disrupting the normal patterns of development, knocking development off of its normal track so to speak and maybe onto another tributary, instead of the normal, neurotypical pattern of development," says Geschwind.
Because they cause such severe disability, the information about these gene mutations has enabled parents to form support groups, for example the FamilieSCN2A Foundation which serves as a community for families of autistic children where the autism diagnosis has been linked to a genetic change in the SCN2A gene. Discussions have also been held regarding the idea of using such genetic information to influence future reproductive decisions.
"If it's a de novo variant, then you can tell the parents that the risk would be low [of having another child with the same neurodevelopmental challenges], because there is a limited contribution from inherited factors, if they decided to have subsequent children," says Geschwind. "We can also give the family a sense of the spectrum of how their child might develop over time, and for parents of a two-year-old that's non-verbal and has some walking delay, they want to know what to expect."
But while this can offer huge benefits for these families, the concept of genetic research is not viewed with universal positivity across the autistic community. Autism is a vast spectrum, ranging from those with severe impairments in physical and mental development which will never allow them to live independently, to others with far fewer support needs who view their autism as an identity and advantage, and oppose depictions of autism as a disorder.
Because of this, for some autistic people, their families and a number of academic researchers, the compilation of genetic data has yielded ongoing concerns about how it might be used.
A complex picture
In the past half century, genetics studies have shown that in the majority of autistic people, their neurodiversity arises through the additive effects of hundreds or even thousands of relatively common gene variants which they have inherited from both parents.
These gene variants exist throughout the population of both neurotypical and neurodivergent people, and the individual contribution of any one of these genes to neurodevelopment is negligible. But in combination, they have a significant effect on the wiring of the brain. Bourgeron says that it is not uncommon for one or both parents, who carry some of these gene variants, to display autistic traits such as a preference for order, difficulties in detecting emotions, and being hyperaware of patterns; but unlike their child, these traits do not manifest to such a significant degree that they themselves could be diagnosed as autistic.
Over the last 20 years, autism researchers have devised some ingenious ways of identifying some of these more subtle variants. In the early 2000s, Simon Baron-Cohen, a professor of psychology and psychiatry at the University of Cambridge, and colleagues, devised a test called Reading the Mind in the Eyes. This is intended to assess a person's ability to detect emotions such as looking playful, comforting, irritated or bored – based on a photograph which shows only the person's eyes.
The idea is that poorer performance on the test indicates a higher likelihood of a person being autistic. "Autistic individuals have a different way of looking at the face, and they seem to get more information from a person's mouth," says Bourgeron. "Neurotypical individuals get more information from the eyes."
More recently, in partnership with the DNA testing site 23andMe, which agreed to host the Reading the Mind in the Eyes test on their website, Bourgeron and Baron-Cohen were able to gather data on the abilities of more than 88,000 people to read thoughts and emotions from a person's eyes, and compare this performance with their genetic information. Through this dataset, they were able to identify large groups of gene variants associated with poorer emotion recognition, many of which are thought to be carried by autistic people.
Other research studies have found that common gene variants associated with autism tend to be negatively correlated with empathy or social communication. But they are positively correlated with the ability to analyse and construct systems as well as rules and routines. Most intriguingly, they are also often linked to higher educational attainment, along with greater spatial or mathematical or artistic abilities. "This perhaps explains why these genetic variants, which come from very distant ancestors, have remained in the population throughout human history," says Geschwind.
Geschwind and Baron-Cohen are now embarking on a project to try to understand whether some of the common gene variants linked with autism can explain why autism seems to be more prevalent in men, and why autistic women are thought to be more adept at masking their neurodivergent traits compared with autistic men.
"The likelihood is that differences in male and female brain development and function make the men more susceptible and the women protected from the genetic susceptibility to autism to some degree, but we don't fully understand that yet," says Geschwind.
However, some experts believe that autism may be much more common in women than is currently thought, and that the experiences of this group are being overlooked.
Geschwind suggests that understanding sex differences in autism could help identify protective factors which could be used as future treatments, yet this very concept remains deeply divisive and reflective of one of the core underlying tensions within autism research. While some scientists pursue treatments, other researchers and some autistic people believe that autism is not a disorder to be fixed, but an identity and a shared experience.
"Autism isn't a biological phenomenon that has to be tested for, and where you get a categorical outcome or prognosis," says Sue Fletcher-Watson, professor of developmental psychology at the University of Edinburgh. "It isn't something, like cancer, that is universally agreed to be bad and for which everyone wants a cure. In my opinion, it never will be."
In particular, Fletcher-Watson says that many autistic people fear that the ultimate outcome of autism genetic research will be a prenatal test, which could pose an existential threat to autism. Beginning in 2005, one activist created the Autistic Genocide Clock, stating that if such a test existed, it would represent a continuation of historic attempts to eliminate minority groups. Two decades later, such fears remain.
"Genetic researchers on the whole have done little to listen to, and address, the fears of the autistic community regarding security and future use of genetic data," says Fletcher-Watson. These fears are heightened by political contexts, she says, such as the strength of certain far-right parties, which make the possibility of eugenic use of genetic data seem much more real.
Prenatal tests are already established practice in the UK for conditions caused by having an extra copy of a chromosome in some or all of the body's cells. These include Down's syndrome (where there's an extra copy of chromosome 21), Edward syndrome (where there's an extra copy of chromosome 18), and Patau syndrome (where there's an extra copy of chromosome 13), and in some countries like Iceland, termination rates following a positive screen are close to 100%," she says.
A broad spectrum
Joseph Buxbaum, a psychiatry professor at the Icahn School of Medicine at Mount Sinai, who founded the Autism Sequencing Consortium, an international group of scientists who share samples and genetic data, feels that some autistic activists are missing the point.
"When I get challenged by somebody who says, 'Well I have autism and I don't think I need be researched,' I'm like, 'Well, how about somebody who has no language, an IQ of 50 and will never be able to live alone and unsupervised,'" says Buxbaum. "What are your thoughts about that person? So, when I think about interventions, I'm thinking about these people, as opposed to someone who has trouble maintaining eye contact, unusual interests and conflicts in social situations."
Geschwind agrees, also citing the marked differences which exist across the autistic spectrum. "The majority of the spectrum is a condition that has to be accommodated like any other disability," he says. However, he adds that another cohort – those who are more severely impacted – would warrant treatment. "These are different things," he says.
To try and better stratify the broad spectrum of autistic traits, the Lancet Commission formally recognised the term "profound autism" in 2021, as a way of describing autistic people who are unable to advocate for themselves and are likely to require 24-hour support throughout their lives. Since then, a variety of clinical trials have begun, all using various therapeutic strategies to try and target the single genes underpinning physical and intellectual disability in different individuals with profound autism.
The main idea for these treatments surrounds the fact that all of us have two copies or alleles, or variants, of every single gene, one from each parent. A recent study from Geschwind's lab leveraged the understanding that most of the de novo gene mutations linked with profound autism only knock out one of these copies, suggesting it might be possible to reduce the degree of disability by boosting the unaffected copy. "That means you have one unaffected copy, [the activity of which] which we showed could be turned up to compensate," says Geschwind.
Bourgeron has recently been running a clinical trial using the metal lithium to boost a version of the Shank3 gene in autistic children who are known to have Shank3 mutations. In the future, Geschwind suggests that a technology such as Crispr, which allows scientists to edit a person's DNA, could be used to intervene at an even earlier stage of life. For example, gene therapy could be delivered to unborn babies found to have various mutations, while they are still in the womb. "We've recently figured out a way of doing this," he says. "It might not fully correct the impacted gene, but it could at least partially correct it."
The FDA has recently granted approval for the US-based biotechnology company Jaguar Gene Therapy to run a clinical trial where a gene therapy is administered to autistic children with a Shank3 gene mutation along with a co-occurring genetic condition called Phelan-McDermid syndrome which affects development, speech and behaviour.
"This trial is only possible because all the children participating have genetic diagnoses," says Buxbaum. "And because researchers at Mount Sinai and elsewhere have spent the past 15 years studying how these children develop when they have these mutations. We can then use this natural history data as a control in the study."
But while such trials could undoubtedly result in enormous benefits for the children involved and their families, Fletcher-Watson is still sceptical about their depiction as therapies for autism, profound or otherwise. She would prefer to see them characterised as treatments for intellectual disability.
"I believe that when people talk about these single-gene cases of autism, they are being disingenuous," says Fletcher-Watson. "They are talking about single-gene causes of intellectual disability, perhaps many of whom are also autistic. But there is funding available for research to address autism, active parent campaign groups and all sorts of resources, in a way that there is not for intellectual disability."
At the same time, Fletcher-Watson is more optimistic about the potential of genetic research to devise novel treatments for some of the co-occurring conditions with which autistic people are often diagnosed, including epilepsy, sleep disturbances, OCD and gastrointestinal disorders.
Bourgeron now coordinates a European project on risk, resilience and developmental diversity in mental health, collaborating with autistic people and their families to better understand why autism rarely comes in isolation, and what makes different individuals prone to these conditions.
At the same time, Bourgeron says that we also need to better recognise neurodiversity and reduce stigma relating to autism. "I think that for us as geneticists we need to go back to the needs of each person" he says. "Some autistic individuals with Shank3 mutations are so severely impacted that they need round the clock care. For others, they might only need specific support at school.
"Overall, we need to do a better job of recognising neurodiversity, and doing everything we can to make sure that people who function differently from the majority can flourish in our societies."
--
For trusted insights into better health and wellbeing rooted in science, sign up to the Health Fix newsletter, while The Essential List delivers a handpicked selection of features and insights.
For more science, technology, environment and health stories from the BBC, follow us on Facebook, X and Instagram.