Metabolic Profiling of the Children’s Autism Metabolome Project (CAMP)

Friday, May 12, 2017: 5:00 PM-6:30 PM
Golden Gate Ballroom (Marriott Marquis Hotel)
R. Burrier, J. King, A. M. Smith, R. Alexandridis, P. West, D. Sugden, M. Ludwig, L. Feuling and E. Donley, Stemina Biomarker Discovery, Madison, WI

ASD is a complex spectrum of neurodevelopmental disorders with heterogeneous underlying genetic, metabolic, and environmental causes. The CAMP (ClinicalTrials.gov Identifier NCT02548442) study, the largest clinical study using metabolomics based methodologies, is being conducted to validate biomarkers identified in three previous studies of banked blood samples and to better understand the metabolism of children with ASD. Metabolomic analyses of ASD may be useful in identifying biomarkers that can provide insight into the role that biochemical disorders, the gut microbiome, dietary and environmental factors play in ASD. The metabolic signatures of ASD can be useful in parsing the broad autistic spectrum into more homogeneous and clinically significant subtypes. Better understanding of metabolic subtypes or metabotypes of ASD can lead to early diagnosis, development and selection of more precise therapeutic intervention, as well as better understand of the efficacy of current interventions within metabotypes.


Conduct broad, discovery-based metabolomics profiling of CAMP subjects to reveal predictive metabolic signatures and discover metabolic subtypes of ASD. Identify metabotypes which can identify at least 50% of the ASD patient samples with a positive predictive value (PPV) of greater than 90%. Validate the metabolites identified as potential biomarkers of metabotypes of ASD in previous studies in the CAMP study.


Plasma was obtained in sodium heparin tubes from children aged 18 to 48 months. Samples were analyzed using 4 orthogonal LC/HRMS-based methods as well as GC-MS to measure a broad range of metabolites through both targeted and non-targeted metabolomics methods. The patient samples were split into a training set of samples utilized for discovery profiling and a test set used for evaluation of the diagnostic signatures discovered in the training set. Univariate, multivariate and machine learning methods were applied to the training set to identify the most predictive set of metabolic features capable of classifying plasma samples as being from ASD or typically developing (TD) children. The molecular signatures were evaluated in the test set to determine their classification performance.


The metabolomics analysis of the 1500 CAMP subject samples is currently underway. In three separate clinical studies, comprising nearly 500 ASD and TD individuals, we demonstrated that metabolic profiling can be used to identify biomarkers associated with ASD as well as to elucidate metabotypes of ASD. These signatures contained both previously reported metabolic changes associated with ASD as well as novel, unreported changes. This poster will present results from the analysis of CAMP samples and potential of metabolites to discriminate ASD from non-ASD individuals.

Conclusions: Applying a paradigm such as this to identify metabolic signatures associated with ASD and elucidate their biochemical implications may be useful in developing diagnostic tests to detect ASD in children at an earlier age and for improving outcomes through personalized treatment. This approach will provide novel information on the biochemical mechanisms involved in ASD and the potential to identify targets for new therapies designed to treat the unique metabolic profile of the metabotype of the individual.