Causal Relationship Between Gut Microbiota and Liver Cirrhosis: 16S rRNA Sequencing and Mendelian Randomization Analyses

Abstract

Background and aims: Accumulating evidence highlights the association between the gut microbiota and liver cirrhosis. However, the role of the gut microbiota in liver cirrhosis remains unclear.

Methods: We first assessed the differences in the composition of the bacterial community between CCl4-induced liver cirrhosis and control mice using 16S rRNA sequencing. We then performed a two-sample Mendelian randomization (MR) analysis to reveal the underlying causal relationship between the gut microbiota and liver cirrhosis. Causal relationships were analyzed using primary inverse variance weighting (IVW) and other supplemental MR methods. Furthermore, fecal samples from liver cirrhosis patients and healthy controls were collected to validate the results of the MR analysis.

Results: Analysis of 16S rRNA sequencing indicated significant differences in gut microbiota composition between the cirrhosis and control groups. IVW analyses suggested that Alphaproteobacteria, Bacillales, NB1n, Rhodospirillales, Dorea, Lachnospiraceae, and Rhodospirillaceae were positively correlated with the risk of liver cirrhosis, whereas Butyricicoccus, Hungatella, Marvinbryantia, and Lactobacillaceae displayed the opposite effects. However, the weighted median and MR-PRESSO estimates further showed that only Butyricicoccus and Marvinbryantia presented stable negative associations with liver cirrhosis. No significant heterogeneity or horizontal pleiotropy was observed in the sensitivity analysis. Furthermore, the result of 16S rRNA sequencing also showed that healthy controls had a higher relative abundance of Butyricicoccus and Marvinbryantia than liver cirrhosis patients.

Conclusions: Our study provides new causal evidence for the link between gut microbiota and liver cirrhosis, which may contribute to the discovery of novel strategies to prevent liver cirrhosis.

Keywords: 16S rRNA gene sequencing; Causality; Gut microbiota; Liver cirrhosis; Two-sample Mendelian randomization study.

Fig. 1. Study design and workflow of MR analysis.

GWAS, genome-wide association study; IVW, inverse variance weighting; LD, linkage disequilibrium; MR, Mendelian randomization; MR PRESSO, MR pleiotropy residual sum and outlier; SNP, single nucleotide polymorphism

Fig. 2. Construction of CCl4-induced liver cirrhosis mice model.

(A) The experimental design diagram presentation of CCl4-induced liver fibrosis. (B) Body weight changes of the two groups. (C) Liver weight/Body weight of the two groups at the 6th and 8th week after intraperitoneal injection of CCl4. (D–E) Representative images of HE and Masson staining of liver sections. (F) Serum ALT, AST and ALP level. ***p<0.001, ****p<0.0001. CCl4, carbon tetrachloride; HE, hematoxylin-eosin; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase.

Fig. 3. Comparisons of alpha and beta diversity between control and CCl4-induced liver cirrhosis mice.

(A) Alpha diversity of two groups using Shannon, Simpson, Chao1, Observed species, Faith’s PD, Pielou’s evenness, and Good’s coverage indices. (B) Beta diversity shown by principal coordinate analysis. (C) Comparative analysis of the gut microbiota between control and CCl4-induced liver cirrhosis by Heat map analysis. CCl4, carbon tetrachloride.

Fig. 4. Metabolic pathway difference analysis.

(A–B) Functional prediction based on the KEGG database. (C–D) Metabolic pathway difference analysis based on the MetaCyc database.

Fig. 5. Causal associations between two gut microbiota with the risks of liver fibrosis.

(A–D) Forest plots of (A) Butyricicoccus and (B) Marvinbryantia. (C–D) Scatter plots of (C) Butyricicoccus and (D) Marvinbryantia. IVW, inverse variance weighting; MR, Mendelian randomization; MR PRESSO, MR pleiotropy residual sum and outlier.

Fig. 6. Bar plots to validate MR.

(A–B) Comparative relative abundance of genes_Butyricicoccus (A) and genes_Marvinbryantia (B) in healthy controls and liver cirrhosis patients. *p<0.05.