logo ecco
Portal

ECCO'25 Abstracts

ECCO'25 Abstracts on the JCC Website

ECCO'25 Abstracts COI Disclosure

ECCO'24 Abstracts

ECCO'24 Abstracts on the JCC Website

ECCO'24 Abstract Book PDF Version

ECCO'24 Abstracts COI Disclosure

Oral presentations: Scientific session 8: IBD Horizons (2020)

OP30 The interplay of microbiome dysbiosis and immune system deregulation in patients with Crohn’s disease

N.S. Seyed Tabib1, C. Caenepeel1, K. Machiels1, S. Verstockt1, B. Verstockt1,2, N. Ardeshir Davani1, J. Sabino1,2, M. Ferrante1,2, S. Vermeire1,2

1Chronic Diseases, Metabolism and Ageing, KU Leuven, Leuven, Belgium, 2Department of Gastroenterology and Hepatology, University Hospitals Leuven, Leuven, Belgium

Background

The perturbation of composition, function, and structure of the gut microbiota known as dysbiosis is a key factor in inflammatory bowel disease (IBD) pathogenesis. There is a crosstalk between the microbiota and the gut immunological niche. To better understand this interaction, we characterised the degree of dysbiosis and dysregulation of the immune proteome in Crohn’s disease (CD) patients to see whether subtypes of patients could be identified.

Methods

We collected faecal and serum samples of 146 CD patients and 63 healthy controls (HC) (Figure 1), and studied microbiota phylogenetic (16S rRNA gene sequencing) and serum proteomic (91 inflammatory proteins OLINK). Microbial dysbiotic index (MDI), defined as the logarithm of the sum of [abundance in organisms increased in CD] over the [abundance of organisms decreased in CD] was calculated and patients were ranked from Q1 (the least dysbiotic state) to Q4 (the most dysbiotic state). For the proteomic score, 32 proteins that correlated (adj. p ≤ 0.01) with faecal calprotectin (FC) were selected. A penalised logistic regression model was trained on these proteins, to distinguish HC from super active (defined as FC ≥ 1800 μg/g). We next developed an inflammatory proteomic score (IPS) defined as the weighted sum of the serum level of inflammatory proteins, using the coefficient value of the regression model as the protein’s weight. Using the IPS score, patients were clustered from Q1 (the least inflammatory state) to Q4 (the most inflammatory state). Statistical analyses were performed in R 3.5.2.

Results

The MDI did not correlate with standard phenotypic subgroups based on the Montreal classification but did positively correlate with C-reactive protein (CRP) and FC level (p ≤ 0.001). The regression model identified 14 proteins [including CCL20, CXCL1, IL-7, IL-17A, FGF-19] distinguishing super active CD patients from HC with accuracy, sensitivity, and specificity of 95.6%, 92.3%, 100%, respectively. IPS positively correlated with CRP and FC level (p ≤ 0.001). Likewise, MDI and IPS-based clusters were significantly different in CRP and FC levels. Different components of the microbiome correlated with the proteome in a subset of samples. For example, fibroblast growth factor 19 (FGF-19) positively correlated with Faecalibacterium and negatively with Fusicatenibacterium. Of note, we observed a significant positive correlation between MDI and IPS (r = 0.33, p ≤ 0.001) (Figure 2).

Conclusion

We were able to define clusters of patients based on molecular characterisation of different players in IBD pathogenesis such as microbiota and proteome. This molecular clustering in a given patient could be considered as a novel therapeutic and personalised approach to IBD. Further validation in larger cohorts is required.