In-depth multiomic characterization of the effects of obesity in high-fat diet-fed mice

Abstract

High-fat diet (HFD)-fed mice have been widely used in the clinical investigation of obesity. However, the long-term effect of HFD on gut microbiota and metabolites, plasma and liver metabolomics, colonic and liver transcriptomics remain largely unknown. In this study, 6-week-old C57BL/6J male mice fed with HFD for 14 weeks showed increased obesity-related indexes including alanine aminotransferase, aspartate aminotransferase, total cholesterol, total triglyceride, free fatty acids, lipopolysaccharides, IL-6, and TNFα. Furthermore, microbial diversity and richness were also significantly decreased. In the colon, genes involved in tryptophan metabolism, PPAR signaling pathway, cholesterol metabolism, and lipid localization and transport, were upregulated. While in the liver, MAPK signaling and unsaturated fatty acid biosynthesis were upregulated. Metabolomic analyses revealed decreased levels of glycerophospholipids and fatty acyl, but increased amino acids, coenzymes and vitamins, and organic acids in the colon, suggesting high absorption of oxidized lipids, while acyl-carnitine, lysophosphatidylcholine, lysophosphatidylethanolamine, and oxidized lipids were reduced in the liver, suggesting a more active lipid metabolism. Finally, correlation analyses revealed a positive correlation between gut microbiota and metabolites and the expression of genes associated with lipid localization, absorption, and transport in the colon, and nutrients and energy metabolism in the liver. Taken together, our results provide a comprehensive characterization of long-term HFD-induced obesity in mice.

Keywords: colonic gene expression; gut microbiome; high‐fat diet‐fed obesity; liver‐specific gene expression; metabolomics; multiomics.

Fig. 1

Changes in the gut microbiota composition in mice fed with and without HFD. (A) PCoA based on Bray–Curtis's distance showing significant differences in the gut microbial composition between two groups at genus level. (B) Alpha diversity calculated on Shannon, Simpson, and Inverse Simpson indexes showed a significant decrease of the gut microbial genera in HFD‐fed mice. Significantly different phyla and genera (C) and species (D) between two groups. The color key in C and D suggested the occurrence rate of each phylum, genera and species in HFD‐fed and control group. The columns in C and D represent the effect size of the corresponding species and the color of columns (orange and blue) represent the gut microbial enrichment between two groups. P < 0.05 was considered of significance.

Fig. 2

Differential analysis of colon metabolic signatures of HFD‐fed and control mice. (A) PCoA based on Bray–Curtis's distance showed significant difference in the colonic metabolites between two groups of mice. (B) Volcano plot showed the differences in colonic metabolites between two groups. Red and blue dots represent the significantly altered colonic metabolites in control and HFD‐fed mice, respectively (P < 0.05). (C) KEGG pathways involved for the significantly different colonic metabolites (P < 0.05). The columns represent the number of metabolites. (D) Significantly different colonic metabolites between two groups (P < 0.01). The columns represent the effect size of metabolites. The color of the columns in C and D represents enrichment of colonic metabolites between two groups.

Fig. 3

Colonic gene expression in HFD‐fed and control mice. (A) PCoA showing significant differences in the colonic gene expression of the two groups of mice (P < 0.05). Red and blue dots represent the genes highly abundant in HFD‐fed and control mice, respectively. (B) GO annotation and KEGG pathway analysis of the significant colonic DEGs between two groups (P < 0.05). Colors of the bars meant different levels of GO including CC, MF and BP, and KEGG. The color of the text distinguishes the two groups. (C) Correlation analysis between the significantly different gut microbiota and colonic DEGs involved in lipid absorption, transport, and metabolism. Red and blue squares meant positive and negative correlations, respectively. *P < 0.05; ⁺ P < 0.01.

Fig. 4

Differences in liver gene expression between HFD‐fed and control mice. (A) Volcano plot showed significantly different liver genes between two groups. (B) GO annotation and KEGG pathway analysis of the significant liver DEGs between two groups (P < 0.05). The color of bars suggested the different levels of GO (CC, MF and BP) and KEGG, the color of the axis text suggested two different groups. (C) Association analysis between the significantly different gut microbiota and the DEGs that involved in lipid metabolism in the HFD‐fed mice. Red and blue squares meant positive and negative correlations, respectively. *P < 0.05; ⁺ P < 0.01.

Fig. 5

Seven KEGG pathways involved in liver lipid metabolism by liver differential genes between two groups. (A) Fatty acid biosynthesis. (B) Fatty acid elongation; (C) Biosynthesis of unsaturated fatty acids; (D) Four KEGG pathways involved by Peroxisome; (E) AMPK signaling pathway; (F) PPAR signaling pathway. The front and back of the arrows are compounds, next to the arrows are the genes involved in this step. The compounds in the rectangular circle represent those metabolites detected by untargeted metabolomics. Bold Italic font in bracket: Significantly different DEGs between two groups (P < 0.05); Red font in bracket: Significantly abundant in related group (P < 0.05); Black font in bracket: Higher in related group (P ≥ 0.05).

Fig. 6

Heatmap showed the correlation analysis between the significantly different microbial species and the significantly different liver genes involved in lipid metabolism. Red and blue squares meant positive and negative correlations, respectively. *P < 0.05; ⁺ P < 0.01.