Study offers a unified biological model to explain the causes of autism

A new University of California San Diego School of Medicine study offers a unified biological model to explain how genetic predispositions and environmental exposures converge to cause autism spectrum disorder (ASD). The study, published in Mitochondrion on Dec. 9, 2025, describes a "three-hit" metabolic signaling model that reframes autism as a treatable disorder of cellular communication and energy metabolism. The model also suggests that as many as half of all autism cases might be prevented or reduced with prenatal and early-life interventions.

Our findings suggest that autism is not the inevitable result of any one gene or exposure, but the outcome of a series of biological interactions, many of which can be modified. By understanding how these genetic and environmental factors stack to alter a child's developmental trajectory, we can start to imagine preventive care and new approaches to treatment that were previously thought impossible."

Robert K. Naviaux, M.D., Ph.D., study author, professor of medicine, pediatrics and pathology at UC San Diego School of Medicine

The three-hit model, developed from more than a decade of systems biology research, proposes that autism develops when three conditions align:

• Genetic predisposition: Inherited genes can make mitochondria and certain cellular signaling pathways unusually sensitive to change.

• Early trigger: Environmental exposures - such as maternal or early infant infection, immune stress or pollution - can activate a universal cellular stress response, called the cell danger response (CDR).

• Prolonged activation: When this cellular stress response stays switched on for too long - typically because of repeated or ongoing exposure to stressors from late pregnancy through the first two to three years of life - it can interfere with normal brain development and lead to ASD.

At the center of this model is the CDR, a metabolic process that helps cells heal from injury or infection, respond to threats and adapt to changing conditions. The CDR is normally short-lived: it turns on to promote healing and turns off once the danger has passed. However, when the response becomes chronic - due to persistent stressors or inherited hypersensitivity - it can disrupt cellular communication and alter mitochondrial function. This happens through changes in extracellular ATP (eATP) – related purinergic signaling - chemical signals that cells use to communicate stress and coordinate healing - which can interfere with how brain circuits form during early life and contribute to the core features of autism.

"Behavior has a chemical basis," said Naviaux. "The CDR regulates that chemistry. When it remains activated too long, it diverts the body's resources from normal growth and development toward cellular defense, leaving fewer resources for the developing brain."

This systems-level framework integrates decades of findings about autism - from mitochondrial and immune dysfunction to gut microbiome changes and sensory hypersensitivity - into a single biological narrative. It helps explain why both genes and environment play roles in autism risk, and why neither alone is sufficient to cause the condition.

Naviaux argues that this perspective shifts the search away from a single "autism gene" toward understanding how diverse stressors converge on common biochemical pathways. "The same signaling systems that allow cells to respond to injury or infection also regulate the formation of neural circuits in early development," he said.

Because the second and third "hits" - environmental triggers and prolonged activation of the CDR - are potentially reversible, early detection and intervention could dramatically reduce autism risk.

To illustrate how multiple metabolic hits can stack to cause disease, Naviaux compares autism to phenylketonuria (PKU), a classical genetic disorder that causes intellectual disability if untreated. PKU also follows the three-hit metabolic model: if detected and treated early, 95% of affected children develop normally despite carrying a disease-causing gene. Similarly, Naviaux estimates that if pregnancies and infants at highest risk for autism can be identified and supported early, 40–50% of cases might be prevented or significantly improved.

Potential strategies include presymptomatic screening, such as maternal metabolomic profiling, autoantibody testing and specialized newborn analyses to identify at-risk children before symptoms appear. 

The study arrives amid rising autism prevalence and ongoing debate over its causes. By reframing ASD as a neurometabolic and neuroimmune condition - rather than a strictly genetic or behavioral one - Naviaux hopes to bridge scientific silos and encourage new collaborations in prevention and therapy.

Future research, he said, should focus on refining diagnostic tools that can detect metabolic stress before symptoms appear and on testing therapies that rebalance the body's energy and signaling systems. Naviaux calls for the development of new antipurinergic drugs to regulate the abnormal ATP signaling that triggers and maintains the CDR. Larger, multi-site clinical trials are also needed to evaluate these new drugs and metabolic support strategies in children with ASD. He also advocates for prenatal and early-life screening programs that combine genetics, metabolomics and environmental data to identify at-risk families sooner.

Together, these efforts could help determine whether calming the cell danger response can prevent or reduce autism's most disabling features.

"Understanding autism through the lens of metabolic signaling doesn't just change how we think about the condition - it changes what we can do about it," said Naviaux. "If we can recognize and calm the cellular stress response before it becomes chronic, we may be able to improve or even prevent some of the most disabling symptoms."