Integrated analysis of the transcriptome and its interaction with the metabolome in metabolic associated fatty liver disease: Gut microbiome signatures, correlation networks, and effect of PNPLA3 genotype.

Interactions between communities of the gut microbiome and with the host could affect the onset and progression of metabolic associated fatty liver disease (MAFLD), and can be useful as new diagnostic and prognostic biomarkers. In this study, we performed a multi-omics approach to unravel gut microbiome signatures from 32 biopsy-proven patients (10 simple steatosis -SS- and 22 steatohepatitis -SH-) and 19 healthy volunteers (HV). Human and microbial transcripts were differentially identified between groups (MAFLD vs. HV/SH vs. SS), and analyzed for weighted correlation networks together with previously detected metabolites from the same set of samples. We observed that expression of Desulfobacteraceae bacterium, methanogenic archaea, Mushu phage, opportunistic pathogenic fungi Fusarium proliferatum and Candida sorbophila, protozoa Blastocystis spp. and Fonticula alba were upregulated in MAFLD and SH. Desulfobacteraceae bacterium and Mushu phage were hub species in the onset of MAFLD, whereas the activity of Fonticula alba, Faecalibacterium prausnitzii, and Mushu phage act as key regulators of the progression to SH. A combination of clinical, metabolomic, and transcriptomic parameters showed the highest predictive capacity for MAFLD and SH (AUC = 0.96). In conclusion, faecal microbiome markers from several community members contribute to the switch in signatures characteristic of MAFLD and its progression towards SH.

Plasma and stool metabolomics to identify microbiota derived-biomarkers of metabolic dysfunction-associated fatty liver disease: effect of PNPLA3 genotype.

INTRODUCTION: Non-invasive biomarkers are needed for metabolic dysfunction-associated fatty liver disease (MAFLD), especially for patients at risk of disease progression in high-prevalence areas. The microbiota and its metabolites represent a niche for MAFLD biomarker discovery. However, studies are not reproducible as the microbiota is variable. OBJECTIVES: We aimed to identify microbiota-derived metabolomic biomarkers that may contribute to the higher MAFLD prevalence and different disease severity in Latin America, where data is scarce. METHODS: We compared the plasma and stool metabolomes, gene patatin-like phospholipase domain-containing 3 (PNPLA3) rs738409 single nucleotide polymorphism (SNP), diet, demographic and clinical data of 33 patients (12 simple steatosis and 21 steatohepatitis) and 19 healthy volunteers (HV). The potential predictive utility of the identified biomarkers for MAFLD diagnosis and progression was evaluated by logistic regression modelling and ROC curves. RESULTS: Twenty-four (22 in plasma and 2 in stool) out of 424 metabolites differed among groups. Plasma triglyceride (TG) levels were higher among MAFLD patients, whereas plasma phosphatidylcholine (PC) and lysoPC levels were lower among HV. The PNPLA3 risk genotype was related to higher plasma levels of eicosenoic acid or fatty acid 20:1 (FA(20:1)). Body mass index and plasma levels of PCaaC24:0, FA(20:1) and TG (16:1_34:1) showed the best AUROC for MAFLD diagnosis, whereas steatosis and steatohepatitis could be discriminated with plasma levels of PCaaC24:0 and PCaeC40:1. CONCLUSION: This study identified for the first time MAFLD potential non-invasive biomarkers in a Latin American population. The association of PNPLA3 genotype with FA(20:1) suggests a novel metabolic pathway influencing MAFLD pathogenesis.