The test uses machine learning to recognize autoantibody patterns in mother’s plasma that are highly associated with autism
In a recent study conducted by the UC Davis MIND Institute, researchers discovered specific associations between maternal autoantibody patterns and increased risk of a child developing maternal autoantibody-related autism spectrum disorder (MAR ASD). By using these recurring patterns as potential biomarkers as inputs for machine learning, clinicians and doctors can test any mother’s plasma to observe whether the same patterns are present and make an early prediction of the chances that their child will develop ASD, even before a mother’s pregnancy.
Judy Van de Water—a professor at UC Davis Internal Medicine, Associate Director of UC Davis MIND Institute and lead author of the study—explained the impact this test could have on autism diagnosis.
“Right now, the average time for diagnosis is really three and a half to four years old, but by then the plasticity of the brain—to change how the brain is developing—is already complete,” Van de Water said. “But if we can diagnose really early, we can have a greater impact with earlier interventions.”
At a young age, brain development in children can be greatly impacted by their environment and interactions, an idea known as brain plasticity. Thus, being able to diagnose and intervene early has huge implications, including lessening the impacts of autism, according to Van de Water.
The research team worked to achieve early diagnosis by first comparing autoantibodies present in a mother whose child developed MAR ASD to those of a mother whose child developed normally. Alexandra Ramirez, a graduate student in UC Davis’ Graduate Immunology Group working in Van de Water’s research team and the first author of the paper, explained the process in which autoantibodies can attack fetal brains.
“We all have autoantibodies to some degree, but during pregnancy they can cross the mother’s placenta and the baby’s blood-brain barrier and attack certain targets in the fetal brain,” Ramirez said.
Autoantibodies are a type of immune response that attacks an individual’s own cells or proteins, and if a mother’s autoantibodies can target components of the fetal brain, alterations in neural development are inevitable.
Van de Water’s research team retrieved plasma samples from mothers of affected children and mothers of unaffected children from a comprehensive CHARGE study, which had accumulated a large sample of enrolled mothers and collected mass amounts of data on various environmental factors and risk exposures that could affect their child’s development. The team then identified through machine learning which autoantibody combinations and patterns were observed at high frequencies in mothers of children with ASD.
The researchers then built and validated a test of 100% accuracy that could recognize these autoantibody patterns that can alter fetal brain development in different mothers with ASD children. Irva Hertz-Picciotto, a principal investigator of the CHARGE study who previously collaborated with Van de Water, described the emotional stress parents may experience while learning of their child’s risk for autism.
“I remember talking to one family,” Hertz-Picciotto said. “They were thinking about getting pregnant and having another child and there’s an elevated risk of having a child [with ASD]. The father asked, ‘Well, how much?’ and I said, ‘Somewhere around 20%.’ The father was shocked, expecting it to be between 3 to 5% [chance of the child developing ASD]. It can be very devastating and draining for the family providing a child with intensive care.”
Ramirez explained that autism affects individuals in nearly every community throughout the U.S., highlighting that people with autism still face prejudice in communities where stigma is predominant.
“Autism is a disorder that affects 1 in 59 children in the U.S.,” said Ramirez. “And the tendency is still on the rise. We really do not know much about autism. In the more underrepresented communities, there is a lot of stigma on mental health, so just talking about it—this is a biological process—we need to educate ourselves and the community about how this is something that is real that we don’t really understand much about.”
Not only is academic discourse around this diverse and heterogeneous disorder of autism necessary, but so is a humane discussion of how we can better support families who oftentimes are left devastated thinking their child is different from the rest.
Though it will be a long time before this maternal biomarkers test can enter the market, both Ramierez and Van de Water note that it is a big step forward and a promising result, laying the foundation for further refinement and testing—as well as a much earlier diagnosis—that could change the whole process of ASD intervention and benefit families in the long run.
“As we move toward early risk detection as something more commonly administered, we can reduce the impact that autism has on both the family and the individual, so that life might be more comfortable for everyone as the early interventions can greatly lessen the severity of autism in the child,” Van de Water said.
Written by: Brandon Nguyen — science@theaggie.org
