205 Discussion In this study, we showdistinct mucosal host–microbe interactions in intestinal tissue frompatients with IBD. Mucosal gene expression patterns in IBD are mainly determined by tissue location and inflammatory status and systematically demonstrate upregulation of distinct inflammationassociated genes, even in endoscopically non-inflamed tissue. Subsequently, we observed that the mucosal microbiota composition in patients is marked by high inter-individual variability. The main focus of our analyses, however, was integrative analysis of both data entities, which allowed us to comprehensively uncover many host–microbe associations, both on component level and as individual associations in IBD. Furthermore, we identify specific transcriptional networks that are significantly altered in patients with fibrostenotic CD and patients using TNF-α-antagonists and observe that these associations depend on the degree of mucosal dysbiosis. Finally, we show that mucosal microbiota are significantly associated with intestinal cell type composition, in particular with epithelial cells, macrophages and NK cells. Tissue location and inflammatory status have the greatest impact on the variation in mucosal gene expression patterns. Enriched genes are mainly represented by those involved in pathophysiological pathways relevant to IBD, e.g. interleukin and interferon signaling and ECM remodeling. Patients with CD and UC show striking differences, e.g. Notch-1 signaling pathways are upregulated in CD, while genes involved in nutrient absorption and lipid metabolism are downregulated. Activation of Notch-1 signaling has been associated with improved mucosal barrier function, driven by lamina propria-residing CD4+-T-lymphocytes that induce intestinal epithelial cell differentiation.17 Notch-1 signaling more efficiently spreads within CD intestinal epithelia, as compared to UC or control epithelia. Notch-1 is not only implicated in IBD, it also confers protection against the development of colorectal carcinoma via p53 signaling, thereby promoting cell cycle arrests and cellular apoptosis.18,88,89 Since UC patients with long-lasting colonic inflammation have a higher risk of developing IBD-associated colorectal carcinoma, we hypothesize that downregulation of Notch-1 in these patients may potentially be involved in carcinogenesis. Analysis of mucosal microbiota in patients with IBD reveals reduced alpha-diversity, microbial dissimilarity and marked intra-individual variability that is particularly strong in CD but still present to a lesser extent in UC. Given the large heterogeneity in IBD and the fact that compositional differences are largely attributable to individual phenotypic factors, cautious interpretation is warranted when associating mucosal microbial profiles to disease phenotypes or outcomes, rendering them inappropriate for diagnostic purposes. These observations corroborate those of previously published mucosal 16S studies in IBD.13,21,22 Moreover, our findings align with a recent prospectivemeta-analysis study that concluded there is sparse evidence for additional population structure in mucosal microbiomes in IBD, e.g. microbiota-driven discrete disease subtypes within IBD.90 Sparse-CCA analysis was performed to capture the key pathway–bacteria interactions. These include numerous inverse associations between bifidobacteria and expression of genes involved in fatty acid metabolism, which align well with previously published data from animal studies demonstrating anti-inflammatory and anti-lipogenic effects of Bifidobacterium treatment on chemically-induced intestinal inflammation.29,31,32 For example, treatment with Bifidobacterium Mucosal host-microbe interactions in IBD
RkJQdWJsaXNoZXIy MjY0ODMw