The composition of the gut microbiota varies significantly across different Immune-Mediated Inflammatory Diseases (IMIDs), affecting diverse body systems such as the gastrointestinal tract, central nervous system, and synovium. This research aimed to identify patterns of gut microbiota dysbiosis in IMIDs compared to healthy controls and among different IMID cohorts.
Key Findings of Gut Microbiota Dysbiosis in IMIDs
This study revealed significant differences in stool microbial profiles between individuals with IMIDs and healthy controls. These differences were identified through differential abundance analyses and machine learning using the random forest algorithm. Consistent dysbiosis patterns were observed across multiple IMIDs, along with specific taxonomic biomarkers associated with particular diseases or disease groups. Notably, a lower than expected abundance of Gram-negative taxa was observed across all cohorts. The study focused primarily on Gram-positive bacteria due to this observation but also validated findings using the complete dataset (including both Gram-positive and Gram-negative bacteria).
Figure 1: Principal Coordinates Analysis (PCoA) plot illustrating the beta diversity of gut microbiota based on Bray-Curtis dissimilarity.
Microbial Diversity in IMID Patients
Reduced gut microbial diversity was a consistent finding in all IMID patients compared to healthy controls, as measured by Shannon and Simpson indices. This observation aligns with previous studies on gut microbiota in IMIDs. Significant differences in diversity were also noted between disease cohorts, particularly when comparing Crohn’s Disease (CD) and Rheumatoid Arthritis (RA) to other IMIDs. While previous studies have assumed similar microbial shifts across IMIDs, this study highlights distinct variations in diversity patterns.
Figure 2: Box plots showing alpha diversity metrics (Shannon and Simpson indices) for various IMID cohorts and healthy controls.
Identifying Taxonomic Biomarkers for IMIDs
Several genera and operational taxonomic units (OTUs) were consistently overabundant in IMIDs compared to healthy controls, suggesting their potential role in inflammatory disease. Actinomyces and Eggerthella spp., implicated in inflammatory conditions, were consistently overabundant in IMID patients. Other taxa, such as Roseburia, known for its butyrate-producing and anti-inflammatory properties, were significantly reduced in IMIDs, pointing to a potential protective role against disease.
Machine Learning Classification of IMIDs
A random forest approach was employed to classify diseases versus healthy controls using OTU or genus abundance as features. The performance of classifiers using only Gram-positive data was comparable to those using the complete dataset. Classifiers for CD versus healthy controls performed well, exceeding previously reported results for Inflammatory Bowel Disease (IBD) versus healthy controls. However, multi-disease classifiers performed poorly, suggesting similarities in gut microbiota composition between different diseases.
Study Limitations and Future Directions
A significant limitation of this study was the unusually low abundance of Gram-negative microorganisms, potentially due to suboptimal storage conditions. However, focusing on Gram-positive bacteria and validating results with the complete dataset mitigated this limitation. Other limitations included the use of stool samples instead of mucosa and the exclusive focus on bacteria. Lack of detailed patient data regarding disease activity, treatment, and lifestyle factors also limited the analysis. Future studies should address these limitations by incorporating larger cohorts, detailed patient information, and broader microbial analysis including fungi and viruses. Investigating the interplay between gut microbiota and systemic immune responses in different IMIDs is crucial for understanding disease pathogenesis and developing targeted therapies.
