Urinary Metabolomics of Young Italian Autistic Children Supports Abnormal Tryptophan and Purine Metabolism

Background: Autism Spectrum Disorder (ASD) is still diagnosed through behavioral observation, due to a lack of biomarkers, which could instead greatly aid clinicians in providing earlier diagnoses, more timely referrals to behavioral intervention programs and evidence-based prognostic predictions. Metabolomic technologies offer a sensitive means to search human biofluids for metabolite profiles potentially usable as biomarkers for ASD. Initial metabolomic studies, analyzing urines and plasma of ASD and control individuals, suggested that autistic patients may share some metabolic abnormalities, despite several inconsistencies stemming from differences in technology, ethnicity, age range, and definition of “control” status. 

Objectives: Our study aims to detect differences in urinary metabolic patterns between tightly matched young autistic and typically developing children previously shown to differ in urinary p-cresol levels (Altieri et al., 2011) by applying a highly sensitive, accurate and unbiased approach suitable to ensure broad metabolite detection coverage on human urine.  

Methods: ASD-specific urinary metabolomic patterns were explored at an early age in 30 ASD children and 30 controls tightly matched by age, sex, Italian ancestry and city of origin within the country (age range 2-7, M:F=22:8) using hydrophilic interaction chromatography (HILIC)-UHPLC-MS. Metabolites were then subjected to multivariate statistical analysis (MetaboAnalyst 3.0 software) and grouped by metabolic pathway.  

Results: Urinary metabolites displaying the largest differences between young ASD and control children belong to the tryptophan and purine metabolic pathways. Also vitamin B6, riboflavin, phenylalanine-tyrosine-tryptophan biosynthesis, pantothenate and CoA, and pyrimidine metabolism differ significantly. ASD children preferentially transform tryptophan into xanthurenic acid and quinolinic acid (two catabolites of the kynurenine pathway), at the expense of kynurenic acid and especially of melatonin. Also the gut microbiome contributes to altered tryptophan metabolism, yielding increased levels of indolyl 3-acetic acid and indolyl lactate. 

Conclusions: The metabolic pathways most distinctive of young Italian autistic children largely overlap with those found in rodent models of ASD following maternal immune activation or genetic manipulations. These results are consistent with the proposal of a purine-driven Cell Danger Response, accompanied by overproduction of epileptogenic and excitotoxic quinolinic acid, large reductions in melatonin synthesis, and gut dysbiosis. These metabolic abnormalities could conceivably underlie some comorbidities frequently associated with ASD, such as seizures, sleep disorders, and gastrointestinal symptoms, and could contribute to autism severity. Their diagnostic sensitivity, disease-specificity and interethnic variability merit further investigation.