Integrated analysis of microbiome and host transcriptome reveals correlations between gut microbiota and clinical outcomes in HBV-related hepatocellular carcinoma.

BACKGROUND

The gut-liver axis plays a pivotal role in the pathogenesis of hepatocellular carcinoma (HCC). However, the correlations between the gut microbiome and the liver tumor transcriptome in patients with HCC and the impact of the gut microbiota on clinical outcome are less well-understood.

METHODS

Fecal samples collected from HBV-related HCC patients (n = 113) and healthy volunteers (n = 100) were subjected to 16S rRNA sequencing of the microbiome. After a rigorous selection process, 32 paired tumor and adjacent non-tumor liver tissues from the HCC group were subjected to next-generation sequencing (NGS) RNA-seq. The datasets were analyzed individually and integrated with clinical characteristics for combined analysis using bioinformatics approaches. We further verified the potential of the gut microbiota to predict clinical outcome by a random forest model and a support vector machine model.

RESULTS

We found that Bacteroides, Lachnospiracea incertae sedis, and Clostridium XIVa were enriched in HCC patients with a high tumor burden. By integrating the microbiome and transcriptome, we identified 31 robust associations between the above three genera and well-characterized genes, indicating possible mechanistic relationships in tumor immune microenvironment. Clinical characteristics and database analysis suggested that serum bile acids may be important communication mediators between these three genera and the host transcriptome. Finally, among these three genera, six important microbial markers associated with tumor immune microenvironment or bile acid metabolism showed the potential to predict clinical outcome (AUC = 81%).

CONCLUSIONS

This study revealed that changes in tumor immune microenvironment caused by the gut microbiota via serum bile acids may be important factors associated with tumor burden and adverse clinical outcome. Gut microbes can be used as biomarkers of clinical features and outcomes, and the microbe-associated transcripts of host tumors can partly explain how gut microbiota promotes HCC pathogenesis.