Identification of Neurotransmitter-Associated Metabotypes: Further Stratification of the Children’s Autism Metabolome Project ASD Subjects.

Background: Autism spectrum disorder (ASD) is biologically and behaviorally heterogeneous and is associated with a diverse array of underlying genetic, metabolic, and environmental factors. We conducted the Children’s Autism Metabolome Project (CAMP, ClinicalTrials.gov Identifier: NCT02548442) comprised of 1100 children to identify altered metabolism associated with ASD. Utilizing a metabotyping approach we created metabolic tests which identify subpopulations of ASD subjects. We recently published (Smith et al., 2018) Amino Acid Dysregulation Metabotypes (AADMs) associated with dysregulation of the amino acids (AA) glutamine, glycine, and ornithine in ratios with branched-chain amino acids (BCAAs). The AADMs were present in 17% of the CAMP subjects with ASD. Metabotype stratification of ASD may provide more biochemically homogenous populations that, in turn, offer the potential for more tailored pharmacological, behavioral, and dietary interventions.

Objectives: We sought to: 1) identify additional metabotypes described by plasma amine containing metabolites associated with neurotransmitters; 2) determine whether neurotransmitter-related metabotypes describe an independent population of ASD subjects that is additive to AADMs; and 3) optimize the inclusion of diagnostic metabotypes to create a battery of tests that increase overall diagnostic performance of the amine-based metabolic tests.

Methods:

The Autism Diagnostic Observations Schedule-Second Version (ADOS-2) was performed by research reliable clinicians to confirm ASD diagnoses. A training set of CAMP subjects (ASD=253, TYP=85) was utilized to identify metabotypes and create metabolic tests. The reproducibility of the tests was evaluated in an independent test set of CAMP subjects (ASD=263, TYP=79). Plasma samples from these subjects were analyzed using a quantitative mass spectrometry-based assay for amine containing metabolites. A heuristic algorithm was applied to the training set to identify biomarkers able to discriminate ASD subpopulations using a diagnostic threshold to define the subpopulation. The diagnostic thresholds were used to create panels of tests for metabotype populations and then the reproducibility of the metabotype was evaluated in the test set of subjects.

Results:

Reproducible metabotypes containing the neurotransmitter-associated metabolites kynurenine and glycine were identified in the training and test sets. The kynurenine-based metabotype also includes, ornithine, glutamic acid, and ethanolamine and identifies 8.5% of CAMP ASD subjects. The glycine-based metabotype also includes asparagine, phenylalanine, and lysine and identifies 12% of CAMP ASD subjects. Both neurotransmitter metabotypes identified ASD subjects with greater than 95% specificity. These additional metabotypes increased the overall diagnostic sensitivity from 17% defined solely by the AADMs to about 30% when all metabotypes are combined into the test battery.

Conclusions: Continued analysis of the CAMP study samples using metabotyping techniques has identified additional neurotransmitter-related metabotypes of ASD. Stratifying ASD based on metabotypes offers an opportunity to identify children earlier at risk for ASD as well as the potential for targeting therapies based on the metabotypes for more precise and individualized treatment.