Search

Non-alcoholic fatty liver disease (NAFLD) is one of the most common causes of chronic liver disease, increasing in worldwide prevalence as a result of the obesity epidemic. It manifests in hepatic cells as steatosis with or without lobular inflammation and/or ballooning. Animal and human studies have suggested the gut microbiome contributes to steatosis/NAFLD. The aim of this study was to use an integrated approach with various -omics and clinical data to evaluate the contribution of the gut microbiome to the molecular phenome (hepatic transcriptome, metabonome) of steatosis. Metagenomic (faecal microbiome), transcriptomic (liver biopsy), metabonomic (plasma and urine, 1H-NMR) and clinical data were collected for 56 morbidly obese (BMI >35) women from Italy (n = 31) and Spain (n = 25) who elected for bariatric surgery. Confounder analyses of clinical data were done using linear modelling. Histological examination of liver biopsies was used to grade steatosis. Faecal metagenomes were generated and analysed using the SCalable Automated Metagenomics Pipeline (SCAMP). Differentially expressed genes were identified in hepatic transcriptomes, and analysed using a range of different bioinformatics tools. 1H-NMR data were generated for plasma and urinary metabonomes. Clinical, metagenomic, transcriptomic and metabonomic data were integrated in the context of steatosis using partial Spearman's correlation, taking confounders (age, body mass index and cohort) into account. Steatosis was anti-correlated with microbial gene richness, and correlated with abundance of Proteobacteria. KEGG analyses of metagenomic data suggested increased microbial processing of dietary lipids and amino acids, as well as endotoxin-related processes related to Proteobacteria. Steatosis-associated hepatic transcriptomes were associated with branched-chain amino acid (BCAA) metabolism, endoplasmic reticulum/phagosome, and immune responses associated with non-specific microbial infections. Metabonomic profiles highlighted imbalances in choline metabolism, BCAA metabolism and gut-derived microbial metabolites resulting from metabolism of amino acids. Molecular phenomic signatures were stable and predictive regardless of sample size, and consistent with the microbiome making a significant contribution to the steatosis phenome. Validation studies (in vitro and in vivo) supported findings from the study. There is disruption of the gut-liver axis in steatosis, which can be seen in the gut microbiome, hepatic transcriptome and urinary and plasma metabonomes. Consistency of phenome signatures strongly supports a relationship between microbial amino acid metabolism and microbial gene richness, hepatic gene expression and biofluid metabonomes, and ultimately steatosis.

Objectives: To integrate metagenomic (faecal microbiome), transcriptomic, metabonomic and clinical data to evaluate the contribution of the gut microbiome to the molecular phenome (hepatic transcriptome, plasma and urine metabonome) of non-alcoholic fatty liver disease (NAFLD) independent of clinical confounders in morbidly obese women recruited to the FLORINASH study. Methods: Faecal, liver biopsy, blood and urine samples and data for 28 clinical variables were collected for 56 obese [body mass index (BMI) >35] women from Italy (n = 31) and Spain (n = 25) who elected for bariatric surgery. Confounder analyses of clinical data were done using linear modeling. Histological examination of liver biopsies was used to grade NAFLD (NAFLD activity score: 0, 1, 2, 3). Faecal metagenomes were generated and analysed using the Imperial Metagenomics Pipeline. Differentially expressed genes were identified in hepatic transcriptomes, and analysed using Enrichr, network analyses and Signaling Pathway Impact Analysis. 1H-NMR data were generated for plasma and urinary metabonomes. Clinical, metagenomic, transcriptomic and metabonomic data were integrated using partial Spearman’s correlation, taking confounders (age, body mass index and cohort) into account. Results: NAFLD activity score was anti-correlated with microbial gene richness, and correlated with abundance of Proteobacteria. KEGG analyses of metagenomic data suggested increased microbial processing of dietary lipids and amino acids, as well as endotoxin-related processes related to Proteobacteria. Metabonomic profiles highlighted imbalances in choline metabolism, branched-chain amino acid metabolism and gut-derived microbial metabolites resulting from metabolism of amino acids. NAFLD-associated hepatic transcriptomes were associated with branched-chain amino acid metabolism, endoplasmic reticulum/phagosome, and immune responses associated with microbial infections. Molecular phenomic signatures were stable and predictive regardless of sample size, and consistent with the microbiome making a significant contribution to the NAFLD phenome. Conclusions: Low microbial gene richness is associated with NAFLD. There is disruption of the gut–liver axis in NAFLD, which can be seen in the gut microbiome, hepatic transcriptome and urinary and plasma metabonomes. Consistency of phenome signatures strongly supports a relationship between microbial amino acid metabolism and microbial gene richness, hepatic gene expression and biofluid metabonomes, and ultimately NAFLD activity scores.