Bile multi-omics analysis classifies lipid species and microbial peptides predictive of carcinoma of gallbladder

Histopathological examination is the gold standard for detection of gallstone (GS) or gallbladder carcinoma (CAGB). Bile concentrated in the gallbladder (GB) is expected to recapitulate metagenomics and molecular changes associated with development of CAGB.

Bile samples were screened for lipidomics and metaproteome (metagenomics) signatures capable of early detection of cancer in GB anomalies. Analysis of the training cohort (n = 87) showed that metastability of bile was reduced in CAGB (p < 0.05). CAGB bile showed significant alteration of lipidome and microbiome as indicated by multivariate partial least squares regression analysis and alpha-diversity and beta-diversity indexes (p < 0.05). Significant reduction of lipid species and increase in bacterial taxa were found to be associated with patients with CAGB, CAGB with GS, and GS (p < 0.05, log fold change >1.5). A multimodular correlation network created using weighted lipid/metaproteomic correlation network analysis showed striking associations between lipid and metaproteomic modules and functionality. CAGB-linked metaproteomic modules/functionality directly correlated with lipid modules, species, clinical parameters, and bile acid profile (p < 0.05). Increased bacterial taxa (Leptospira, Salmonella enterica, Mycoplasma gallisepticum) and their functionality showed a direct correlation with lipid classes such as lysophosphatidylinositol, ceramide 1-phosphate, and lysophosphatidylethanolamine and development of CAGB (r²  > 0.85). Lipid/metaproteomic signature-based probability of detection for CAGB was > 90%, whereas that for GS was > 80% (p < 0.05). Validation of eight lipid species using four machine learning algorithms in two separate cohorts (n = 38; bile [test cohort 1] and paired plasma [test cohort 2]) showed accuracy (99%) and sensitivity/specificity (>98%) for CAGB detection.

Bile samples of patients with CAGB showed significant reduction in lipid species and increase in bacterial taxa. Our study identifies a core set of bile lipidome and metaproteome signatures which may offer universal utility for early diagnosis of CAGB.