The crosstalk between microbiota and metabolites in AP mice: an analysis based on metagenomics and untargeted metabolomics

Abstract

Background and purpose: Microbiome dysfunction is known to aggravate acute pancreatitis (AP); however, the relationship between this dysfunction and metabolite alterations is not fully understood. This study explored the crosstalk between the microbiome and metabolites in AP mice.

Methods: Experimental AP models were established by injecting C57/BL mice with seven doses of cerulein and one dose of lipopolysaccharide (LPS). Metagenomics and untargeted metabolomics were used to identify systemic disturbances in the microbiome and metabolites, respectively, during the progression of AP.

Results: The gut microbiome of AP mice primarily included Firmicutes, Bacteroidetes, Actinobacteria, and Proteobacteria, and "core microbiota" characterized by an increase in Proteobacteria and a decrease in Actinobacteria. The Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis found that significantly different microbes were involved in several signaling networks. Untargeted metabolomics identified 872 metabolites, of which lipids and lipid-like molecules were the most impacted. An integrated analysis of metagenomics and metabolomics indicated that acetate kinase (ackA) gene expression was associated with various gut microbiota, including Alistipes, Butyricimonas, and Lactobacillus, and was strongly correlated with the metabolite daphnoretin. The functional gene, O-acetyl-L-serine sulfhydrylase (cysK), was associated with Alistipes, Jeotgalicoccus, and Lactobacillus, and linked to bufalin and phlorobenzophenone metabolite production.

Conclusion: This study identified the relationship between the gut microbiome and metabolite levels during AP, especially the Lactobacillus-, Alistipes-, and Butyricimonas-associated functional genes, ackA and cysK. Expression of these genes was significantly correlated to the production of the anti-inflammatory and antitumor metabolites daphnoretin and bufalin.

Keywords: acute pancreatitis; gut microbiota; metabolites; metabolomics; metagenomics.

Figure 1

Study design and workflow.

Figure 2

Establishment of the AP mouse model. (A) Schematic diagram of the experimental model. (B) Histopathological changes of the pancreas (HE, ×200). (C) Pancreatic histologic scores in the Ctrl and AP groups. (D, E) Amylase and lipase serum levels. (F–H) TNF-α, IL-1β, and IL-6 expression in the pancreas. Data are presented as the mean ± SEM; **P<0.01 vs. the Ctrl group, ***P<0.001 vs. the Ctrl group.

Figure 3

Differences in the gut microbiota of the Ctrl and AP groups. (A) PCoA analysis of the gut microbiota structure and abundance in the Ctrl and AP mice. (B, C) Relative abundance of the gut microbiota at the phylum and genus levels. (D) LEfSe analysis of the Ctrl and AP mice using a cladogram.

Figure 4

KEGG pathway enrichment.

Figure 5

Analysis of untargeted metabolomics. (A, B) The reproducibility of the experiment was evaluated using a population sample principal component analysis (PCA-QC). (C) The proportion of identified metabolites in each chemical class. (D, E) Volcano plots show fold changes in metabolite levels.

Figure 6

Bioinformatic analysis of differential metabolites. (A, B) A heatmap shows the top 30 differentially produced metabolites. (C, D) Chord diagrams show the coregulatory relationships between the various metabolites. (E) KEGG analysis of enrichment pathways.

Figure 7

Integrated analysis of metagenomics and metabolomics. (A) The PCoA map. (B) A Sankey diagram showing the correlation of species–functional genes–metabolites.