Q&A with John Constantino: Fixing the biomarker conundrum | Spectrum
Inna / Adobe Stock
Professor, Washington University in St. Louis
The most common forms of autism – those that do not occur as part of a rare genetic syndrome – are highly heritable. But despite this clearly biological basis, disease variability has hampered the search for biomarkers that could be used to improve diagnosis, track prognosis, and identify new therapies.
Part of the problem could be that the biological factors responsible for whether a person has autism are different from those that affect the severity of the condition, says John Constantino, professor of psychiatry at Washington University in St. Louis , Missouri.
In a comment published in Molecular Autism in April, Constantino introduces a new way to differentiate between different types of biomarkers. To identify those that reflect the origins of autism, researchers need to study the behavioral traits that precede and predict a diagnosis – and they need to look not just at autistic people, but at the general population.
Constantino spoke to Spectrum about the “biomarker puzzle” for autism and how researchers can do it better.
Spectrum: How do researchers traditionally define and seek biomarkers in autism, and what is the mystery of how this has evolved?
Johannes Constantino: Biomarkers link a biological signature – for example a variation at the gene and DNA level, brain signatures from neuroimaging or serum markers – with a neuropsychiatric characteristic or result. One result could be a diagnosis of autism, or it could be the severity of the autism. But the puzzle is: do you associate your biomarker with a cause of autism or an effect of autism?
Based on our work with identical twins and families, we believe that researchers sometimes assume that their biomarkers are related to causes of autism, when those biomarkers actually signal effects of autism, and therefore biomarker associations may not always be what They seem.
S: How did your research address this potential problem?
JC: The story goes back to a paper we published in 2017. Colleagues at Emory University found that in infants and babies who have a sibling with autism, social visual distancing was a strong predictor of autism. When we analyzed this biomarker in twins in the general population, we found that social visual distancing is extremely highly inherited – and that what predicted autism in later-born siblings was also present in some of the typically developing children. This suggested that social visual disconnection is highly genetic and may be necessary, but not sufficient, to cause autism.
So we started thinking about what else could combine with social visual distancing to create autism. We found three more predictors that could be identified before the core signs of autism manifest themselves: autism-like features passed from parents to children, signs of attention-deficit hyperactivity disorder (ADHD) in babies, and subtle impairments in motor coordination. Not only were these four factors highly heritable, but they were all independent of one another – no overlap.
S: Why is that important?
JC: If multiple things are inherited but there is no correlation between them in an epidemiological sample, they are very likely to be tapping into different genetic factors. We looked at these four factors in the general population and found that together they determine 60 percent of the baby [autism-like behaviors]. We also showed that, taken together, they are responsible for a similar recurrence of autism in families. Hence, we are likely missing additional factors that would make up the other 40 percent. However, we believe that certain combinations of these inherited predictors, especially when severe, cause autism. This is the model of autism that scientists should consider.
S: Can you explain a little more about the model of autism you are proposing?
JC: The idea is that these factors act as levers. They can occur in different permutations and are present in the general population. There are young children who walk around with little social visual engagement, but if they don’t have extreme atypicals in any of the other factors, they may not have autism. The children with autism seem to be the ones with combinations of extremes.
If we could measure all of these markers or levers very early, before autism manifests itself, and give them a little nudge, we could potentially throw a child off the path that leads to autism. However, it could be that these combinations of developmental commitments are themselves effects of earlier developmental forces taking place in the womb (and not after birth) and then we need to take a different approach to intervention.
S: According to this model, what happens after autism manifests, and what does that say about biomarkers?
JC: That’s the second part of the story. In identical twins, if one twin has autism, the other twin will also have autism 95 percent of the time. But we were surprised to learn that they can have remarkably different degrees of autism severity, with correlation coefficients suggesting that only 5 to 10 percent of the cause of the severity is shared between two identical twins. A previous report noted this, but the results were not often incorporated into considerations for inheritance in autism, and it was not common in twin studies to quantify severity in a standardized way. If a pair of identical twins raised together differ in some trait of interest, the difference must be attributed to non-shared environmental factors.
S: What could that be?
JC: Three things can cause a pair of identical twins to be different from one another. One is the measurement error. Another is a very influential event that occurred in one twin but not in another – for example, a perinatal stroke or child abuse. The third thing is what are known as stochastic influences – random, minor events that would have no effect on a typically developing person, but on someone with a neurodevelopmental disorder [might have] much higher consequences. It’s like a gust of wind at the wrong time, and your brain moves in direction X instead of Y for reasons we don’t understand. We believe that stochastic influences can greatly affect brain development if your brain is vulnerable to a condition like autism is. What we have learned from studies of identical twins is that a child’s autism can be more or less severe, almost randomly, and that there can be differences in severity over time with age.
S: What would you advise researchers looking for autism biomarkers?
JC: First, we should distinguish inherited familial autism from autism caused by de novo mutations. De novo mutations are not inherited, and there is a difference in the types of autism caused by such mutations, being predominantly associated with intellectual disability, while hereditary autism is much more commonly (not exclusively) associated with one intellectual development in the typical field. One appeal to the field is to separate inherited (multiplex) from sporadic (simplex) cases in study samples, recognizing that they represent potentially different pathways to autism and that inherited autistic syndromes are most common in the population.
Also, when trying to link a biomarker to the cause of inherited forms of autism, the most insightful way to look at the association with predictors of autism before the condition manifests itself may be the most revealing. A biomarker’s relationship to autism can change after autism becomes apparent. In parallel, we need to examine traits that predict autism not just in families affected by autism but in families of the general population.
After all, when conducting a case-control study comparing children with autism and typically developing children, the controls all have the same levers of developmental liability – they just aren’t that atypical for this trait. Many unaffected controls have significant levels of one or more of the characteristics that confer autism risk – perhaps more in the 70th percentile than the 95th percentile. But if you don’t measure or control them, you are, by definition, jeopardizing the statistical power to distinguish cases from controls based on measurements of biomarkers that are primarily related to those underlying traits.