Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder with a wide spectrum of impacts that primarily affect behavioral, communication, and social skills. ASD was once considered a rare condition, with prevalence recorded as 3-4 in 10,000 individuals in the 1970s. As of now, about 1% of the global population is estimated to have autism. This represents a 20 to 30-fold increase in just the past few decades.

Ongoing research is trying to understand the drivers for this increase, which may include changes in diagnostic criteria, increased awareness and screening, and even environmental factors. One key area that may help with diagnosis is the discovery of biomarkers related to ASD. Biomarkers provide an objective way to identify and measure biological differences for numerous diagnostic purposes, as well as to predict therapeutic potential.

While the methods for an ASD diagnosis have evolved over time, the criteria used in the DSM5 are still primarily behavioral and often subjective. Identifying effective treatments is also a challenge, given the variability of symptoms and condition severity (Figure 2)  

Researchers have noted that various biomarkers have been playing important roles in early cancer detection and the identification of neurodegenerative diseases like Alzheimer’s and Parkinson’s. They are now finding that autism-related biomarkers can be used to screen patients at high risk of autism, assist in earlier diagnosis for more effective intervention, and classify patients into subgroups to better predict responses to different medications. These advancements can provide critical insights for clinicians and better overall care for patients.

We examined the CAS Content Collection^TM, the largest human-curated repository of scientific information, and found a marked increase in publications relating to ASD biomarkers over the last decade (see Figure 3).  

Which types of biomarkers can be used to identify autism?

ASD biomarkers can be categorized as non-molecular and molecular. As seen in Figure 4, molecular biomarkers are much more common in the literature (92% of the total documents), while non-molecular biomarkers represent a much smaller portion (8%). Non-molecular biomarkers are considered subjective, whereas molecular biomarkers are more objective and measurable.  

Figure 3 shows the document publication trends of identified autism related biomarkers. We also observed numerous documents related to possible biomarkers which yet need to be confirmed as ASD biomarkers in the future with more comprehensive studies.

Molecular biomarkers of autism

Numerous types of molecular ASD biomarkers are being researched. Let’s examine the groupings based on their prominence in the literature, via the CAS Content Collection:

• Genetic biomarkers: ASD appears to have high heritability, and as a result, researchers are investigating many potential genetic biomarkers. Studies have shown that genetic factors can be traced to more than 50% of autism cases, and to date over 400 genes have been reported as associated with ASD phenotypes. We identified leading genes associated with ASD publications in the CAS Content Collection (see Figure 5)  

The gene most associated with ASD is the FMR-1 gene (fragile X mental retardation 1 gene), which is connected to Fragile X syndrome (FXS). About 50% of the FXS males and 20% of the FXS females meet the DSM-5 diagnostic criteria for ASD. Single nucleotide polymorphisms (SNPs) in exons 5, 10, and 11 are also shown to be associated with ASD. Some other key genes linked to ASD include PTCHD1, HOX, CHD2, CHD8, FOXP2, SHANK3, OXTR, PTEN, CHD2, and NL3. These genes play various roles in brain development and function, influencing the behavioral, communicative, and cognitive abilities of individuals with ASD.

The complexity of autism genetics suggests there may be multiple pathways contributing to it. This implies not just a simple inheritance pattern but complex interactions between different genes and environmental factors. There is now increasing evidence suggesting that the bulk of genetic risk associated with ASD may be due to SNPs and copy number variants, rather than specific single-gene mutations or other monogenic syndromes. Genetic markers are promising early diagnostic tools and may identify genetic pathways to target pharmacological interventions. Furthermore, SNPs can potentially be used to stratify ASD patients based on symptom severity.

• Immunologic biomarkers: Several meta-analysis studies have identified predictive markers like maternal fetal brain-directed autoantibodies and folate receptor alpha autoantibody. Researchers have also discovered elevated levels of inflammatory markers in ASD individuals like interferon-γ, interleukin-1β, interleukin-6, tumor necrosis factor-α, and chemokines. Nonetheless, it is still not clear whether neuroinflammation is a causative factor of ASD or  a downstream effect due to other biological processes, such as oxidative stress.  

• Metabolic biomarkers: ASD is associated with many variations in metabolism, but the exact correlations of these metabolic disturbances with pathophysiology of the disorder are not fully understood. However, a number of metabolic markers are emerging as potential diagnostics.  

Several large-scale liquid chromatography/mass spectrometry-based metabolomic studies have shown that individuals with ASD have altered purine metabolism, ATP-related purinergic signaling, fatty acid oxidation, and increased physiologic stress molecules like lactate, glycerol, cholesterol, and ceramides. Changes in metabolic pathways related to energy, neurotransmitters, and branched-chain amino acids (BCAA), as well as aminoacyl-tRNA biosynthesis are also shown to be associated with ASD severity. Abnormal activities of the electron transport chain and impaired energy metabolism, owing to dysfunctional mitochondrial metabolism, also contributes to ASD.

Recent efforts in the field are focused on identifying the relationship between metabolic biomarker clusters, their inter-connectivity, and using them in autism sub-grouping. Metabolomic studies are focused on using certain molecules, such as suramin (CAS RN: 145-63-1) or sulforaphane (CAS RN: 4478-93-7), as pharmacological interventions to treat ASD. Metabolomic studies can also help determine therapy approaches in ASD individuals; for example, preliminary evidence suggests that children with low carnitine or low BCAA-associated phenotypes may benefit from carnitine or BCAA supplementation.

• Oxidative stress biomarkers: Several studies have established a correlation between ASD and physiological abnormalities linked to oxidative stress, such as increased concentration of lipid peroxidation products and reduced glutathione reserve capacity in brain tissue. Subsequently, byproducts of lipid peroxidation like malondialdehyde (MDA; CAS RN: 542-78-9), 4-hydroxy-2-nonenal (CAS RN: 75899-68-2), and F2-isoprostane are emerging as potential biomarkers for early diagnosis of ASD.  

There is also a possible positive correlation between elevated plasma MDA levels and ASD severity. Another important and dependable emerging ASD biomarker with high sensitivity and repeatability for ASD is the GSH/GSSG redox ratio. These potential biomarkers could be used for early diagnosis and evaluation , as well as assistance in pharmacological or nutritional treatment interventions. By enhancing the functionality of the antioxidant enzyme system, it is possible to lessen the oxidative stress damage experienced by ASD patients, thereby potentially reducing symptom severity.

• Epigenetic biomarkers: Along with genetic components, it is equally important to understand the role of epigenetic regulators in ASD. Studies have indicated an association between differential DNA methylation patterns and ASD-related genes. Another histone acetylome-wide association study has identified a collective acetylome signature in ≥68% of ASD cases in the prefrontal and temporal cortex. Cell-free circulating microRNAs have been associated with levels of ASD behaviors are now being evaluated as diagnostic ASD markers thanks to in-depth profiling.  

• Gut microbiome biomarkers: In the past few years, understanding the role of gut microbiota in ASD pathogenesis has become a prominent area of research. Studies have indicated that there is a dysregulation in the composition of gut microbiota in ASD individuals like an increase in Bacteroidetes phylum or Firmicutes phylum or a decrease in Coprococcus and Bifidobacterium genera. Notably, within the Bacteroides family and Lachnospiraceae, the reduced microbial quantities correlate with neurodevelopmental levels and behavioral symptoms like social impairment in patients with ASD.  

The activity of the gut microbiota generates various metabolites which play a crucial role in regulating brain function, such as short-chain fatty acids and propionic acid. Furthermore, research has revealed a close relationship between the gut microbiota composition and neurotransmitters like gamma-aminobutyric acid (GABA; CAS RN: 56-12-2), serotonin (CAS RN: 50-67-9), and dopamine (CAS RN: 51-61-6). The imbalance in gut microbiota composition hence has a vital role in pathogenesis of ASD.

‍Non-molecular biomarkers of autism

While there are fewer non-molecular biomarkers being researched, these measurements can provide a well-rounded examination and assist in ASD diagnosis:

• Neuroimaging biomarkers: There are two types of neuroimaging biomarkers that can help with ASD diagnosis. The first is structural imaging. Multiple longitudinal MRI studies have identified connections between ASD and structural variations in the brain like excessive expansion of cortical surface at 6–12 months or anomalies in Broca’s area and Wernicke’s area. Such studies provide prospective imaging biomarkers for early diagnosis.

The other type of neuroimaging biomarker is functional imaging. Resting-state functional MRI (fMRI) plays a pivotal role in capturing the functional connectivity and activity within the brain in the presence or absence of specific tasks, thereby providing an in-depth insight into the neural mechanisms of ASD. Like structural imaging, functional imaging may also be utilized for early diagnosis.

• Neurophysiological biomarkers: Electroencephalography (EEG) is the most used tool to detect brain activity and analyze neurophysiological conditions. Certain frequencies or patterns and features on EEG readings were shown to be correlated with ASD diagnoses done by clinicians. These studies suggest that EEG measurements can be used to derive certain helpful digital biomarkers which can serve as early diagnosis markers of ASD.
  

• Eye tracking biomarkers: These refer to a method of identifying indicators of ASD by observing and analyzing individuals’ eye movement patterns while they view, focus, and track visual stimuli. Several studies have indicated different patterns of visual attention in ASD individuals compared to control groups. These differences are also closely correlated with the social impairments in ASD individuals.  

Clinical trials could lead to autism biomarker breakthroughs

As of November 2024, there were 59 observational trials and 107 interventional trials for ASD biomarkers. Among the interventional trials, 40 are in Phase I or Phase II, and three have reached the later phases:

Table 1: Late-phase clinical trials of ASD biomarkers. Source: ClinicalTrials.gov.

Clinically accepted biomarkers may be identified for wider usage soon. However, the biggest challenge with ASD biomarkers has been reproducibility across studies owing to the heterogenicity present in ASD individuals.  

Identifying control populations of children for studies is also difficult. The ideal control population for biomarker studies would be children the same age as the ASD cohort but with developmental delays unrelated to ASD. Usually, the control population included in such studies are non-sibling, typically developing children, which is not always a clinically relevant population.

It will take time and effort to surmount these hurdles to clinical research, but the promising data we have about ASD biomarkers means that better diagnosis and treatment options may be near. With the prevalence of ASD diagnosis increasing at alarming rates that aren’t yet fully understood, it’s vital for researchers and clinicians to find more effective methods to diagnose ASD and identify which treatments are right for different patients. The quantifiable nature of biomarkers may provide clinicians with the crucial tools to do so.