Metabolomics study finds biomarkers predicting autism in newborns

By Dr. Liji Thomas, MDReviewed by Benedette Cuffari, M.Sc.May 14 2024

A recent study published in the journal Communications Biology utilizes metabolomics in newborns to identify markers that might predict the occurrence of autism spectrum disorder (ASD).

Study: Metabolic network analysis of pre-ASD newborns and 5-year-old children with autism spectrum disorder. Image Credit: Vink Fan / Shutterstock.com

Biomarkers for ASD

Children with ASD have difficulties in social interactions, language, and restricted or repetitive interests or behaviors. Even with treatment, only 20% live independently as adults after a childhood ASD diagnosis.

Prior research has identified metabolic and biochemical markers for ASD in children and adults, varying with age, sex, and symptom severity. Many of these markers are involved in the structure and function of the brain, immune system, autonomic nervous system, and microbiome. Nevertheless, no single genetic or environmental factor accounts for all ASD cases among children.

Genes do not work in isolation, and that polygenic and gene-environment interactions are the dominant contributors to the development of ASD.”

The CDR model

The cell danger response (CDR) model depicts metabolic pathways connecting environmental and genetic stressors to altered development and ASD. The CDR flows from the point of impact of the stressor outwards, following various changes in metabolic, inflammatory, autonomic, endocrine, and neurological responses to these injuries or stresses.

ASD is more likely to follow CDR when stressors operate in intrauterine life or early childhood. These impact four areas that are part of the CDR, including the mitochondria, oxidative stress, innate immunity, and microbiomes. Extracellular adenosine triphosphate (eATP) is the fundamental regulator in all CDR pathways.

ATP as a signaling molecule

ATP is the energy currency for all life on earth. About 90% of ATP is generated within the intracellular mitochondria and is used for all metabolic pathways.

Outside the cell, eATP functions as an information molecule. To this end, eATP binds to purine-responsive receptors on the cell to warn of danger, altered metabolism, and induce a generalized CDR response.

eATP is one of the most powerful signaling molecules known, capable of binding to receptors found on every cell in the body.”  

Beginning with innate immunity activation, the sequence continues through acute localized responses to trauma or infection that ultimately become remote organ-level or systemic responses. In some cases, this could affect human neurodevelopment.

ATP in ASD metabolism

Dysregulated purine metabolism and purinergic signaling in response to ATP have been identified in experimental and human studies of ATP and confirmed in multi-omics analyses. The role of eATP is key to multiple aspects of neurological development altered in ASD, including mast cells and microglia, neuronal sensitization, and neuroplasticity.

Mitochondria generate ATP and are crucial for processing data, providing early warning, and initiating timely responses to changes in the environment. Mitochondria conduct almost 800 metabolic reactions independently, including those involved in child development, growth and differentiation, healing, stress adaptation, and aging, over half of which are regulated by ATP and its congeners.

Chronic mitochondrial dysfunction in ASD impairs metabolic pathways and gene expression, thereby disrupting neurodevelopment trajectories.

What did the study show?

Infants in the pre-ASD or typically developing (TD) groups did not exhibit any differences in their exposure to environmental factors during pregnancy and infancy. About 50% of children in the pre-ASD group exhibited regression of development at one or more points as compared to 2% in the TD group. The average age at ASD diagnosis was 3.3 years.

Metabolites were increased above the mean level in the newborn ASD cohort and continued to increase by more than half by five years as compared to the newborn cohort. These metabolites included stress molecules and the purine 7-methylguanine that caps newly formed messenger ribonucleic acid (mRNA),  

In newborns, the most significant increase was observed with four sphingolipids, with a corresponding decrease in sphingomyelins, their source molecules. Similarly, 7-methylguanosine was increased, and guanine decreased.

Conversely, metabolites that were decreased in the newborn cohort were reduced by 120% more by five years of age. These included antioxidants, neurotransmitters like dopamine, and one-carbon molecules.

Among five-year-olds, several phospholipids were increased, whereas cardiolipins involved in the production of mitochondria and ATP decreased. Purine 7-methylguanine levels remained high, while several vitamins and serotonin were reduced.

Differentiating ASD from TD newborns

Using six or seven of the identified biomarkers, pre-ASD was distinguished from TD newborns and five-year-olds with an accuracy of 75% and 90%, respectively. Several important classes of metabolites changed their trajectories between birth and five years.

Bile acids, phosphatidylserine (PS), phosphatidylcholine (PC) lipids, and sphingomyelins decreased with age, whereas purines and fatty acid oxidation levels were unaffected. Comparatively, the levels of mRNA capping purines and several lipids, such as acyl-carnitine linoleylcarnitine, increased.

Exploring network interactions between metabolites in TD newborns and TD five-year-olds showed an 18-fold reversal of the ratio of positive to negative correlations from 5.5 to 0.3 in purine metabolic pathways. In ASD, the expected reversal failed to occur, thus indicating failed development.

Neuronal γ-aminobutyric acid (GABA) signaling typically reverses from net excitatory at birth to inhibitory at two to three years of age. This accompanies a reduced vulnerability to environmental factors and a concomitant reduction in the risk of ASD.

Negative correlations with purines were lost with time in the ceramide and phospholipid hubs. The eicosanoid hub exhibited fourfold higher positive and threefold higher negative correlations in ASD as compared to TD.

Despite similar positive-to-negative correlation ratios, there were qualitative differences in the ASD hypercorrelator hub between study groups. For example, asparagine, which mediates mitochondrial signaling pathways for cell growth, was negatively correlated with eicosanoids at multiple points.

The TD hypercorrelator hub showed different positive and negative correlations. Lipids accounted for 13 of the top 15 metabolites in the TD hypercorrelator hub but lost 90% of their correlations in the ASD metabolome.

Metabolic growth rate

In the TD cohort, Vnet, a measure of metabolic growth rate, increased by 173% between birth and five years of age, whereas Vnet was stable in the pre-ASD cohort, thus indicating development arrest. The low connectivity in the metabolic network in ASD might be due to CDR signaling, which inhibits remote signal reception, causing impaired coordination of chemical signals across the body’s various systems.

Potential ASD mechanisms

The current study identified the most prominent changes in children who developed ASD by five years of age as affecting specific groups of complex lipids. About 80% of the metabolic shift was traceable to 14 metabolic pathways observed both in the newborn pre-ASD and five-year ASD cohorts.

Ceramides are lipids that can cause cell death and loss of mitochondrial function. The loss of negative correlations between ceramides and purines leads to their accumulation in ASD. The result is mitochondrial dysfunction and apoptosis of many cells, even without lethal exposures.

The primary impact of this correlation was reflected as lower anti-inflammatory activity, less antioxidant reserve, and more stress response activity, all of which increased with age. Repeated activation of the CDR might cause increased oxygen utilization within the mitochondria.

With higher dissolved oxygen in the cell, cellular membranes undergo oxidative damage. Although this response allows for excess dissolved oxygen to be sequestered, it also stiffens the membranes, limits mitochondrial function and synaptogenesis, and delays responses to environmental stressors in ASD.

The metabolic changes found in children with ASD were not the result of cell dysfunction or damage. Instead, the measured changes were the result of normal physiologic and neurodevelopmental responses to metabolic signals that cells received in ASD that were not being sent in typically developing children.”

Conclusions

The study findings confirm that ASD is linked to metabolic profiles that are distinct from those of TD children, though varying by age, sex, and disease severity. These changes are reflected in the abnormal neurobiology of ASD.

Taken together, the data may indicate that the failure of normal reversal of the purine network causes failure to reverse the GABA-ergic network. The loss of inhibitory connections reduces natural dampening, thereby allowing excessive excitatory calcium signaling in the ASD network.

Therefore, cells tend to remain excited and respond excessively to sensory signals in ASD. This could explain the need for an unchanging routine with ASD children to avoid anxiety induced by unexpected changes.

Future studies can utilize these findings, as well as those obtained from previous reports, to generate better screening tools for newborns and infants to identify those at risk for ASD. This could aid in early detection and intervention for affected children, ultimately improving patient outcomes and reducing the incidence of ASD.