Integrative analysis of metabolome and gut microbiota in diet-induced hyperlipidemic rats treated with berberine compounds

Abstract

Background: Hyperlipidemia is a major component of metabolic syndrome, and often predicts cardiovascular diseases. We developed a new therapeutic agent berberine compounds (BC), consisting of berberine, oryzanol and vitamin B6, and determined their anti-hyperlipidemia activity and underlying mechanisms.

Methods: Male Wistar rats were fed a high fat diet (HFD) to induce hyperlipidemia, and then given BC orally for 4 weeks. Body weight and food intake were recorded weekly, and lipid profiles in serum were determined biochemically. Metabolites in serum, urine, liver and feces were analyzed by GC-MS, and the structure of microbiota was determined by 16S rDNA sequencing.

Results: Lipid lowering was observed in the hyperlipidemic rats upon BC treatment without apparent adverse side effects. Metabolomics analysis indicated that the BC treatment resulted in increased pyruvic acid, serotonin, and ketogenic and glycogenic amino acid levels in the serum, increased pyridoxine and 4-pyridoxic acid in the urine, decreased hypotaurine and methionine in the liver, and increased putrescine and decreased deoxycholate and lithocholate in feces. The BC treatment also resulted in an enrichment of beneficial bacteria (e.g. Bacteroides, Blautia) and a decrease in Escherichia.

Conclusions: The lipid lowering effect of BC treatment in hyperlipidemic rats is associated with a global change in the metabolism of lipids, carbohydrates and amino acids, as well as the structure of microbiota.

Keywords: Berberine compounds; Gut microbiota; Hyperlipidemia; Integrative metabolomics.

Fig. 1

Changes of serum parameters in rats. Male Wistar rats (8 weeks of age) placed on HFD or chow diet for 8 weeks, HFD rats were divided into BC treated and untreated groups, with the 4-week intervention. At the end of experiment, blood was collected, and serum was separated, the lipid profiles were analyzed. a TC, b LDL-c, c HDL-c, d TG, e FFA. The data were presented as the mean ± SD. ***P < 0.001, compared with chow group; ^# P < 0.05, ^## P < 0.01 and ^### P < 0.001, compared with untreated (HFD) group

Fig. 2

The metabolic profiles in the serum. The HFD induced hyperlipidemic rats were treated with or without BC for 4-week, serum was collected, and metabolomics analysis was made by GC/MS. a PCA score plots serum samples from BC treated group and untreated group; b scores plots of PLS-DA between untreated group and BC treated group; c Pearson’s correlations of the quantities of the 35 metabolites determined from serum samples; d heat map showing the fold change (FC) of 19 metabolites. Shades of green represent FC decrease while red represent FC increase; e simplified draft illustrating perturbed pathways involved; f differential metabolites between the groups in glycolysis and citrate cycle; g differential metabolites between groups in glycolysis and citrate cycle; h differential metabolites between the groups in fatty acid biosynthesis. Values were showed as mean peak intensities ± SEM, *P < 0.05, **P < 0.01 and ***P < 0.001, compared with untreated (HFD) group

Fig. 3

The metabolic profiles in the urine. The HFD induced hyperlipidemic rats were treated with or without BC for 4-week, urine was collected, and metabolomics analysis was made by GC/MS. a PCA score plots of urine samples from BC treated group and untreated group; b scores plots of OPLS-DA between untreated group and BC treated group; c Pearson’s correlations of the quantities of the 35 metabolites determined from rat urine samples; d heat map showing the FC of 19 metabolites. Shades of blue represent FC decrease while red represent FC increase; e simplified draft illustrating perturbed pathways involved; f differential metabolites between groups in vitamin B₆ metabolism; g differential metabolites between groups in phenylalanine metabolism. Values were showed as mean peak intensities ± SEM. *P < 0.05, **P < 0.01 and ***P < 0.001, compared with untreated (HFD) group

Fig. 4

The metabolic profiles in liver tissue. The HFD induced hyperlipidemic rats were treated with or without BC for 4-week, liver tissues were collected, and metabolomics analysis was made by GC/MS. a PCA score plots of liver samples from BC treated group and untreated group; b scores plots of OPLS-DA between untreated group and BC treated group; c Pearson’s correlations of the quantities of the 13 metabolites determined from rat liver samples; d heat map showing the FC of 19 metabolites. Shades of green represent FC decrease while red represent FC increase; e simplified draft illustrating perturbed pathways involved; f differential metabolites between groups in the liver. Values were showed as mean peak intensities ± SEM. *P < 0.05; **P < 0.01, compared with untreated (HFD) group

Fig. 5

The metabolic profiles in feces. The HFD induced hyperlipidemic rats were treated with or without BC for 4-week, feces were collected, and metabolomics analysis was made by GC/MS. a PCA score plots of the feces from BC treated group and untreated group; b scores plots of PLS-DA between untreated group and BC treated group; c Pearson’s correlations of the quantities of the 67 metabolites determined from the feces samples; d heat map showing the FC of 19 metabolites, shades of green represent FC decrease while red represent FC increase; e simplified draft illustrating perturbed pathways involved; f differential metabolites between the groups in tyrosine metabolism; g differential metabolites between the groups in β-alanine metabolism; h differential metabolites between the groups in arginine and proline metabolism; i differential metabolites between the groups in secondary bile acid biosynthesis. Values were showed as mean peak intensities ± SEM. *P < 0.05, **P < 0.01 and ***P < 0.001, compared with untreated (HFD) group

Fig. 6

The structure of micorbiota in the feces. Heat map of key OTUs indicating genus-level changes among the groups. The relative abundance of each genus was indicated by a gradient of color from green (low abundance) to red (high abundance). Complete linkage clustering of samples was based on the genus composition and abundance. a Chow group versus HFD group; b BC treated (HFD) group versus untreated group