A new microbiome-based diagnostic test, based on reproducible bacteria biomarkers, makes it possible to diagnose inflammatory bowel disease (IBD) in a non-invasive way, by identifying the different metabolic pathways and microbial signatures that characterize ulcerative colitis and Crohn’s disease.

For this achievement we must thank a study carried out by Jiaying Zheng and colleagues from the University of Hong Kong and published in Nature Medicine. Indeed, the researchers identified reproducible microbiota alterations through the analyses of around 4400 faecal samples obtained from patients with IBD and healthy individuals all over the world. Based on this information, they then developed a diagnostic model to discriminate ulcerative colitis (UC) and Crohn’s disease (CD) from controls. The main steps and findings of this study, starting with the characterization of gut microbial alterations associated with IBD, are described below.

Metagenomics data from 4406 samples of 13 IBD cohorts all over the world allowed the identification of 1175 taxa, of which:

  • 674, 637, and 506 bacterial species were assigned to the UC, CD, and control groups, respectively.
  • Patients with UC and CD showed a decrease in microbial diversity and richness if compared to controls.
  • IBD was associated with 2.2% of microbial alteration; age, and gender with 0.28% and 0.29%, respectively.
  • Gut microbiomes showed significant differences at phylum level between controls and IBD patients.

The comparison between UCs and CDs also suggested the presence of disease-specific bacterial species. In particular:

  • 15 species showed an enrichment in UC patients, and 48 a depletion in CD ones.
  • CD patients showed lower levels of Bacteroidetes.
  • Some anti-inflammatory species are depleted in CD patients, while E. coli and Streptococcus species are enriched in the CD but not in UC ones.

With this information, the researchers developed a diagnostic model using 10 bacterial species as biomarkers for UC (four enriched: Gemella morbillorum, B. hansenii, Actinomyces sp. oral taxon 181 and C. spiroforme; and six depleted: C. leptum, F. saccharivorans, G. formicilis, R. torques, Odoribacter splanchnicus, and Bilophila wadsworthia) and 9 for CD (three enriched: B. fragilis, E. coli, and Actinomyces sp. oral taxon 181 and six depleted: R. inulinivorans, B. obeum, Lawsonibacter asaccharolyticus, Roseburia intestinalis, Dorea formicigenerans and Eubacterium sp. CAG: 274). Applying a suitable algorithm, the mentioned species discriminated UCs or CDs from the controls with a 95% confidence interval. 

The study focused then its attention on the metabolic functions, identifying 545 pathways dysregulated in UC and CD patients. In particular:

  • Pathways of amine and polyamine degradation, fatty acid, and lipid biosynthesis showed a marked enrichment in UC and CD patients vs. controls.
  • Biomarkers increased in UC and CD patients correlated with disease-enriched metabolic pathways. On the other hand, depleted biomarkers were negatively correlated.
  • Pathways involved in the amino acid biosynthesis were mainly associated with bacterial species depleted in UCs. Among these, C. leptum, F. saccharivorans, G. formicilis, and R. torques. E. coli, instead, resulted to be the major contributor of this pathway alteration in CDs.

To explore the role of bacterial biomarkers in metabolic functions, researchers also developed a dysfunctional score for each individual by estimating their difference from controls. The positive correlation between diagnostic model and dysfunctional score suggests how the bacterial biomarkers could signalise the metabolic alterations in IBD.

Next, the study evaluated the accuracy and stability of the bacterial biomarkers for IBD. To do so, patients were divided into active and inactive disease status. The results highlight a different abundance of species associated with UCs and CDs in individuals with inactive IBD vs. controls. Although the diagnostic model couldn’t distinguish between active and inactive IBD, it showed great power in the classification of patients with inactive UC or CD with respect to healthy controls. 

The model was then successfully validated using five independent and multi-ethnic datasets, inclusive of the biomarkers reported in the discovery cohort and additional metagenomic datasets. If compared to the traditional test for the IBD screening, the new model developed showed higher sensitivity and specificity. 

The last step of this study was to develop a general IBD model, not only for UC and CD cases. Taking 18 species identified for UCs and CDs, the model was tested and validated against other gastrointestinal disorders, showing higher performances than the traditional faecal calprotectin in distinguishing IBDs from non-IBDs. 

To conclude, this study presented the development of a diagnostic tool for IBDs, pointing out reproducible alterations in the gut microbiota signatures and metabolic functionality mainly associated with ulcerative colitis and Crohn’s disease. This tool aims to facilitate the diagnosis of IBDs in a non-invasive way.