Intestinal metabolomics in premature infants with late-onset sepsis

Abstract

We aimed to investigate the characteristics of intestinal metabolomics and non-invasive biomarkers for early diagnosis of late-onset sepsis (LOS) by analyzing gut metabolites in preterm infants with LOS. We collected stool samples from septic and healthy preterm infants for analysis by liquid chromatography-mass spectrometry. 123 different metabolites were identified and 13 pathways were mainly involved. Glycine, serine, and threonine metabolism; glyoxylate and dicarboxylic acid metabolism; glutathione metabolism; primary bile acid biosynthesis; steroid synthesis; pentose and glucuronic acid interconversion may be involved in the pathogenesis of LOS in preterm infants. The significant changes of N-Methyldopamine, cellulose, glycine, gamma-Glutamyltryptophan, N-Ribosylnicotinamide and 1alpha, 25-dihydroxycholecalciferol showed specific diagnostic values and as non-invasive biomarkers for LOS.

Figure 1

Fecal samples collection process diagram.

Figure 2

Identification of total metabolites and metabolic pathways. (a) Identification of metabolites. The abscissa is the number of metabolites, the ordinate is the type of metabolites, and the different colors represent the biological role of metabolites. (b) Identification of metabolic pathways. The different colors represent the different types of metabolic pathways in which the metabolites are involved, the abscissa represents the number of metabolites, and the ordinate represents the biological functions they exert.

Figure 3

The score plot of the OPLS-DA model. (a) The OPLS-DA score map. Green points represent the LOS group, and blue points represent the control group. (b) Permutation test. Green boxes represent the R² value and blue boxes represent the Q² value.

Figure 4

Screening of differential metabolites. (a) The volcano map. Red represents up-regulated metabolites, and blue represents down-regulated metabolites. (b) Heat map. Each column represents a sample with different colors indicating different intensities, with colors from blue to red indicating low to high intensities. (c) First 30 VIP score plot. A represents the LOS group, and C represents the control group.

Figure 5

Comparison of ROC curves and intensity of differential metabolites. (a) Differential metabolite ROC curve and Cut-off value. (b) Comparison of different metabolite intensities between the LOS and control groups.

Figure 6

Metabolic pathway analysis. (a) Metabolic pathway map of differential metabolites. Each point represents the metabolic pathway. The size of the point is proportional to the impact value. The color of the point represents the P value of the metabolic pathway. (b) Schematic representation of metabolic pathways. Green indicates down-regulated metabolites, and red indicates up-regulated metabolites.