P048 Deep phenotyping of the peripheral immune cell compartment in Crohn‘s Disease

Haag, L.M.(1)*;Walling, S.(2);Kunkel, D.(3);Hecker, J.(2);Letizia, M.(2);Huck, A.(2);Weidinger, C.(2);Glauben, R.(2);Kuehl, A.(4);Siegmund, B.(2);

(1)Charité Universitätsmedizin Berlin, Department of Gastroenterology- Infectiology and Rheumatology, Berlin, Germany;(2)Charité Universitätsmedizin Berlin, Department of Gastroenterology- Infectious Diseases and Rheumatology, Berlin, Germany;(3)Berlin institute of Health, BIH Flow & Mass Cytometry Core Facility, Berlin, Germany;(4)Charité Universitätsmedizin Berlin, iPATH.Berlin – Core Unit Immunpathologie für Experimentelle Modelle, Berlin, Germany;


Crohn’s disease (CD) is characterized by chronic inflammation that can discontinuously affect any segment of the gastrointestinal tract. New findings concerning clinical behaviour, epidemiology, genetics, and gut microbiota suggest that CD in the ileum (iCD) and the colon (cCD) should be considered as two distinct subtypes of IBD. As systemic immune cell signatures differ between CD and UC, we propose a similar strategy for the differentiation between iCD and cCD. The aim of this study is to analyse and compare peripheral blood mononuclear cell (PBMC) subsets of iCD and cCD patients. For a robust analysis of the peripheral immune compartment in inflammatory bowel diseases (IBD), PBMC subsets of ileocolonic CD (icCD), ulcerative colitis (UC) and healthy controls (HC) were included as well.


In this study a total of 50 adults, 10 subjects per subgroup  – iCD, cCD, icCD, UC and HC – were  included. From each individual, both naïve and PMA/Ionomycin stimulated samples were stained for phenotypic and functional characterization using mass cytometry. For data analysis pre-gating was performed in Cytobank followed by debarcoding, compensation, normalization, clustering, and analysis on differential abundances edge R (DA_edgeR) and differential states limma (DS_limma) with CATALYST in R.


FlowSOM allowed us to cluster CD45+ PBMCs of all samples into 16 major cell subtypes as visualized via UMAP (Fig. 1). DA testing showed significant differences between iCD and cCD, iCD and icCD but not cCD and icCD (Fig 2). Significantly different abundances among CD subgroups and HC included cluster 1 CD33+ myeloid cells, 7 CD45RO+ CD16+ cells, 8 CD11b+ cells, 9 CD4+ IL7R+ T cells and 12 CD27+ CD38+ CD45RA+ B cells (Fig. 3). 

When focussing on differences between iCD, cCD and icCD, DS testing revealed significantly higher IFNgamma and TNFalpha production in cluster (CD4+ IL7R+ T cells), 15 (CD8+ T cells) and 16 (CD8+ CD45RO+ T cells) in iCD compared to icCD. Moreover, cluster 9 showed significantly higher levels of IFNgamma in iCD compared to cCD, and for TNFalpha cluster 16 revealed significantly higher levels in cCD compared to icCD (Fig. 4).

Fig. 1 UMAP clustering of stimulated cells.

Fig. 2 Differential analysis of stimulated samples with DA_edgeR and DS_limma.

Fig. 3 Relative population abundances with DA_edge R testing.

Fig. 4 Relative state marker abundance within clusters with DS_limma testing.


Deep phenotyping of the peripheral immune cell compartment in CD might serve as a confirmatory marker for disease location and therefore as predictor of clinical behaviour.