Introduction

Gut microbiota in Crohn's disease (CD) exhibits considerable variability, with conflicting associations with disease severity and response to therapeutic intervention. We aimed to longitudinally investigate the dynamics of microbial composition in a cohort of patients with newly diagnosed, treatment-naïve Crohn's disease (ndCD) over a one-year period.

Aims & Methods

A prospective real-world study was conducted, enrolling patients with ndCD who were treatment-naïve. Treatment decisions were made by physicians with expertise in CD management. Key one-year clinical outcomes included corticosteroid-free clinical remission (CSFR) and biochemical remission (BR), defined as C-reactive protein levels <1mg/dL and fecal calprotectin levels <150ug/g. Paired stool samples were subjected to 16S rRNA gene sequencing (V4 region) at diagnosis and after one year, followed by analysis for microbial diversity and dysbiosis indices. A linear mixed-effect model was constructed to identify factors associated with microbial alterations and specific taxa driving these alterations, while controlling for confounding variables.

Results

Seventy-five patients completed the one-year follow-up (49.3% female, median age 27 years [IQR 22-38.5]); 45 (60%) received biologics (95.5% anti-tumor necrosis factor [TNF]). Rates of CSFR, BR, and combined CSFR-BR at one-year post-diagnosis were 64%, 56%, and 45.3%, respectively. Analysis of microbial composition revealed significant improvements in microbial indices among patients achieving all desired clinical outcomes, characterized by increased microbial diversity (Shannon index, p=0.004, p=0.002, p=0.004, respectively) and decreased dysbiosis (microbial dysbiosis index [MDI], p=0.003, p=0.04, p=0.01, respectively). Such improvements were not observed in patients failing to achieve CSFR, BR or combined CSFR-BR. Notably, patients treated with biologics, irrespective of outcome achievement, exhibited prominent shifts (improvement) in microbial composition (both indices p<0.001), whereas non-treated patients did not demonstrate significant changes (Shannon index, p=0.53; MDI, p=0.81). Furthermore, microbial composition at baseline was more disrupted among patients treated with biologics compared to non-treated patients (Shannon index, p=0.04; MDI, p=0.03). The linear mixed model analysis confirmed a positive association between biologics and favorable changes in microbial indices (Shannon index, p=0.02; MDI, p=0.04), identifying an increase in 20 gut commensal taxa (considered beneficial) and a decrease in Escherichia coli among the biologics-treated subgroup. Conversely, no taxa exhibited significant alterations in the non-biologics treated subgroup.

Conclusion

In a real-world cohort of patients with treatment-naïve ndCD, clinical improvement over one year was associated with improvement in microbial composition. Treatment with anti-TNF agents was associated with improvement in microbial composition irrespective of clinical outcomes. These findings suggest that microbial composition reflects inflammatory activity and may be positively and independently modified by anti-TNF treatment.

Introduction

The human gut microbiome consists of a diverse range of microorganisms, including not only bacteria but also viruses, archaea and fungi. Among these, bacteria have been shown to have a significant impact on the health or disease status of the host. However, despite increasing evidence for the role of fungi in liver disease, inflammatory bowel disease and cancer, little is known about the interactions between bacteria and fungi.

Aims & Methods

In this study, we aimed to investigate in depth the interkingdom relationship between fungal and bacterial communities in the gut of healthy monozygotic and dizygotic twins using a novel merging analysis method. Faecal samples from 106 twin pairs were analysed using 16S rRNA gene sequencing for bacteria and ITS sequencing for fungi. To enable a comprehensive analysis of the interactions between these two microbial communities, we adapted a novel merging normalisation method and linked it to systematic dietary and environmental characteristics.

Results

The implementation of the merging normalisation method allowed for an in-depth comparison of the fungal and bacterial communities independent of the low read counts. In direct comparison, bacteria were significantly more abundant and diverse than fungi, with Bacteroides and Prevotella being the most dominant bacterial genera. Within the fungal community, Candida was the most abundant genus in the cohort, followed by Geotrichum, while Saccharomyces was less abundant. Other fungal genera were detected sporadically between individuals. Most bacterial taxa showed a negative correlation with both Candida and Geotrichum at some level of abundance. For example, Bacteroides showed no antagonism towards Candida when its relative abundance was below 5%, whereas genera such as Prevotella (rho=-0.16, p=0.02), Alistipes (rho=-0.16, p=0.02) and Faecalibacterium (rho=-0.17, p=0.01) showed antagonistic interactions with Candida at relative abundances as low as 0.1%. These correlations were calculated across different abundances of the interacting fungi and bacteria, showing that negative associations increased with higher abundance (up to rho=-0.6 for Alistipes). An interkingdom network analysis further revealed a complex dynamic involving the three most abundant genera - Bacteroides, Prevotella and Candida - which showed mutually antagonistic relationships.

Conclusion

The implementation of the novel, easy-to-use merging analysis provided an in-depth view of the interactions between bacterial and fungal communities in the human gut, which are predominantly antagonistic in nature. The bacterial community appears to use several mechanisms to protect the gut environment from fungal colonisation.

Disclosure

Nothing to disclose