RESUMO
INTRODUCTION: It is challenging to establish the mechanisms involved in the variety of well-defined clinical phenotypes in autism spectrum disorder (ASD) and the pathways involved in their pathogeneses. OBJECTIVES: The aim of the present study was to evaluate the metabolomic profiles of children with ASD subclassified by mental regression (AR) phenotype and with no regression (ANR). METHODS: The present study was a cross-sectional case-control study. Thirty children aged 2-6 years with ASD were included: 15 with ANR and 15 with AR. In addition, a control group of 30 normally developing children was selected and matched to the ASD group by sex and age. Plasma samples were analyzed with a metabolomics single platform methodology based on liquid chromatography-mass spectrometry. Univariate and multivariate analysis, including orthogonal partial least squares-discriminant analysis modeling and Shared-and-Unique-Structures plots, were performed using MetaboAnalyst 4.0 and SIMCA-P 15. The primary endpoint was the metabolic signature profiling among healthy children and autistic children and their subgroups. RESULTS: Metabolomic profiles of 30 healthy children, 15 ANR and 15 AR were compared. Several differences between healthy children and children with ASD were detected, involving mainly amino acid, lipid and nicotinamide metabolism. Furthermore, we report subtle differences between the ANR and AR groups. CONCLUSIONS: In this study, we report, for the first time, the plasmatic metabolomic profiles of children with ASD, including two different phenotypes based on mental regression status. The use of a liquid chromatography-mass spectrometry platform approach for metabolomics in ASD children using plasma appears to be very efficient and adds further support to previous findings in urine. Furthermore, the present study documents several changes related to amino acid, NAD+ and lipid metabolism that, in some cases, such as arginine and glutamate pathway alterations, seem to be associated with the AR phenotype. Further targeted analyses are needed in a larger cohort to validate the results presented herein.