Metabolomic Analysis Uncovers Lipid and Amino Acid Metabolism Disturbance During the Development of Ascites in Alcoholic Liver Disease

Abstract

Ascites is one of the most common complications of cirrhosis, and there is a dearth of knowledge about ascites-related pathologic metabolism. In this study, 122 alcoholic liver disease (ALD) patients, including 49 cases without ascites, 18 cases with mild-ascites, and 55 cases with large-ascites (1) were established according to the International Ascites Club (2), and untargeted metabolomics coupled with pattern recognition approaches were performed to profile and extract metabolite signatures. A total of 553 metabolites were uniquely discovered in patients with ascites, of which 136 metabolites had been annotated in the human metabolome database. Principal component analysis (PCA) analysis was used to further identify 21 ascites-related fingerprints. The eigenmetabolite calculated by reducing the dimensions of the 21 metabolites could be used to effectively identify those ALD patients with or without ascites. The eigenmetabolite showed a decreasing trend during ascites production and accumulation and was negatively related to the disease progress. These metabolic fingerprints mainly belong to the metabolites in lipid metabolism and the amino acid pathway. The results imply that lipid and amino acid metabolism disturbance may play a critical role in the development of ascites in ALD patients and could be a potent prognosis marker.

Keywords: alcohol liver disease; amino acid metabolism disturbance; ascites; lipid metabolism disturbance; untargeted metabolomic.

FIGURE 1

Metabolomics analysis of ALD patients with/without ascites. (A–D) PCA and orthogonal partial least squares discrimination analysis (OPLS-DA) plots with all metabolic variables in NA, MA, and LA patients under positive and negative models of mass spectrometry, respectively. (E) Metabolites with significant differences in MA and LA, in comparison to NA. There are 553 metabolites uniquely associated with ascites, with 136 metabolites annotated in the Human Metabolome Database. (F) PCA plot of MA + LA vs. NA with 136 annotated metabolites related to ascites. The QC samples are clustered together in the PCA plots, indicating stability and technical reproducibility.

FIGURE 2

Selection of metabolic fingerprints uniquely expressed in ALD patients with ascites. (Ai) Hierarchical cluster analysis of the AUCs and p value assessing the discriminating accuracy of each of the 136 metabolites in differentiating ALD with or without ascites. (Aii) The corresponding metabolites’ AUC and P values in accessing MA or LA relative to NA. The vertical violet bar identifies the 21-metabolite clusters highly associated with ascites and the ascites-associated metabolite fingerprints. (B) The decreasing trend of the eigenmetabolite (21 fingerprint metabolites) in ALD patients without ascites and different stages of ascites.

FIGURE 3

Use of metabolic fingerprints to unravel underlying pathophysiological changes during the progression of ascites. The Nightingale rose diagram of ascites-related metabolic fingerprints. The value of the relative intensity of each metabolite was normalized within the two groups. The whole set of metabolites in the fingerprint was classified into different metabolic pathways, including fatty acid metabolism, phospholipid metabolism, steroids and hormones, amino acids, and proteolysis.

FIGURE 4

Metabolic pathways involved in ALD patients with ascites. Our data suggested the dysregulation in the fatty acid metabolism pathway, amino acid metabolism pathway, and steroid and hormone biosynthesis pathway during ALD-related ascites progression.