Opportunistic detection of Fusobacterium nucleatum as a marker for the early gut microbial dysbiosis.

backgroundThe essential roles of gut microbiome have been emphasized in modulating human health and disease. Fusobacterium nucleatum (F. nucleatum), an obligate Gram-negative microorganism residing in oral cavity, gastrointestinal tract and elsewhere, has been recently considered as a potential oncobacterium associated with human cancers. However, the consequence of its enrichment was not extensively explored in terms of microbial homeostasis and stability at the early stage of disease development.

resultOur analysis on longitudinal metagenomic data generated by the Integrative Human Microbiome Project (iHMP) showed that F. nucleatum was frequently found in inflammatory bowel diseases (IBD) subjects with reduced microbial diversity. Using non-parametric logarithmic linear discriminant analysis (LDA) effect size (LEfSe) algorithm, 12 IBD- and 14 non-IBD-specific bacterial species were identified in the fecal metagenome and the IBD-specific ones were over-represented in the F. nucleatum-experienced subjects during long-term surveillance. In addition, F. nucleatum experience severely abrogated intra-personal stability of microbiome in IBD patients and induced highly variable gut microbiome between subjects. From the longitudinal comparison between microbial distributions prior and posterior to F. nucleatum detection, 41 species could be proposed as indicative "classifiers" for dysbiotic gut state. By multiple logistic regression models established on these classifiers, the high probability of experiencing F. nucleatum was significantly correlated with decreased alpha-diversity and increased number of biomarker species for IBD and colorectal cancer (CRC). Finally, microbial clustering confirmed that biomarker species for IBD and non-IBD conditions as well as CRC signature markers were well distinguishable and could be utilized for explaining gut symbiosis and dysbiosis.

conclusionF. nucleatum opportunistically appeared under early dysbiotic condition in gut, and discriminative classifier species associated with F. nucleatum were successfully applied to predict microbial alterations in both IBD and non-IBD conditions. Our prediction model and microbial classifier biomarkers for estimating gut dysbiosis should provide a novel aspect of microbial homeostasis/dynamics and useful information on non-invasive biomarker screening.