Immunological profiling of patients with ulcerative colitis leads to identification of two inflammatory conditions and CD1a as a disease marker

Abstract

Background: Conventional approaches to understand mechanisms underlying the development of pathological manifestations in ulcerative colitis (UC) mostly rely on identification of certain cell types and cytokines followed by verification of their roles in vitro and in vivo. In light of the highly dynamic processes in UC, requiring the cross talk of immune cells, epithelial-, endothelial-, muscle cells and fibrocytes, this approach might neglect temporal and spatial connectivity of individually differing inflammatory responses.

Methods: We undertook a more holistic approach whereby we designed a flow cytometric analysis- and ELISA panel and determined the immunological profiles of UC patients in comparison to Non UC donors. This panel consisted of B-cells, T-cells, macrophages, monocytes, NK- and NK T-cells and subtypes thereof, the cytokines TGFß1 and HGF, the chemokine TARC and periostin. Blood was collected from 41 UC patients and 30 non-UC donors. Isolated PBMC were subjected to flow cytometric analysis and sera were analyzed by ELISA. Data were analysed by cluster- and correlation analysis. To corroborate that the identified cells reflected the inflammatory condition in the colon of UC patients, leucocytes were isolated from colons of UC patients and subjected to the same flow cytometric analysis.

Results: Immunological profiling followed by cluster- and correlation analysis led to the identification of two inflammatory conditions: An 'acute' condition characterized by adaptive immune cells as plasma cells, TSLPR expressing CD11b+ macrophages, CD64 and CCR2 expressing CD14+ monocytes, HGF and TARC and a 'remodeling' condition signified by NK T-cells and TLSPR expressing CD14+ monocytes, TGFß1 and periostin. ROC analysis identified TARC and TGFß1 as biological markers with high potential to discriminate between these two conditions (Δ = -6687.72 ng/ml; p = 1E-04; AUC = 0.87). In addition, CD1a+ CD11b+ macrophages (Δ = 17.73% CD1a+ CD11b+; p = 5E-04; AUC = 0.86) and CD1a+ CD14+ monocytes (Δ = 20.35; p = 0.02, AUC = 0.75) were identified as markers with high potential to discriminate between UC and Non UC donors. CD1a+ CD11b+ macrophages and NK T-cells were found to be significantly increased in inflamed colons of UC patients as compared to non-UC control samples (p = 0.02).

Conclusions: Immunological profiling of UC patients might improve our understanding of the pathology underlying individual manifestations and phases of the disease. This might lead to the development of novel diagnostics and therapeutic interventions adapted to individual needs and different phases of the disease. In addition, it might result in stratification of patients for clinical trials.

Keywords: Biomarker; CD1a; Correlation analysis; HGF; Immune-profiling; TARC; TGFß1; Ulcerative colitis.

Fig. 1

CD1a expressing macrophages and monocytes as disease specific markers. a Exemplary flow cytometric analysis of CD11b+ CD1a and CD14+ CD1a populations in PBMC from a Non UC donor and a UC patient. b ROC curve for CD11b CD1a+ (Non UC n = 31, UC n = 40) and CD14+ CD1a (Non UC n = 9, UC n = 27) in differentiating UC and Non UC patients

Fig. 2

Immunological profiling leads to identification of distinct subgroups of UC patients. PBMC and serum were isolated from UC patients and subjected to flow cytometric analysis and ELISA, respectively. A Hierachically clustered heatmap based on frequencies of immune cells and serum levels of TGFß1, HGF and TARC. (linkage hierarchical cluster with euclidean distance). Simple clinical colitis activating index (SCCAI). B Boxplot analysis of concentrations of TARC, HGF and TGFß1 and plasma cells and TSLPR expressing CD11b+ macrophages and monocytes in dependence of diagnosis or treatment with TNFα-blockers and glucocorticoids. (UC patients n = 30, treated with TNFα blocker n = 16, Glucocorticoid n = 9, mesalazine n = 16, colectomized patients n = 5. C ROC curve for TARC (a) and TGFß1 (b) in differentiating group I+ II and III. TARC: 95% CI value 0.74–0.97, 18 cases treated with TNFα-blocker, 23 cases no treatment with TNFα-blockers; TGFß1: 95% CI value: 0.59–0.9, 18 cases treated with TNFα-blocker, 24 cases no treatment with TNFα-blockers

Fig. 3

Correlation analysis of the clinical activity score (SCCAI) with subtypes of immune cells and serum factors depicted as scatter plots. (method = Spearman, numbers display Spearman rank-order correlation coefficients (rho-values) and p values

Fig. 4

Correlation- and interrelation analysis of TARC, HGF, TGFß1 and periostin depicted as scatter plots. (method = Spearman, numbers display Spearman rank-order correlation coefficients (rho-values) and p values

Fig. 5

Correlation- and interrelation analysis of CD11b+ TSLPR+ macrophages and CD14+ TSLPR+ monocytes with subtypes of immune cells and serum factors depicted as scatter plots. (method = Spearman, numbers display Spearman rank-order correlation coefficients (rho-values) and p values

Fig. 6

Correlation- and interrelation analysis of NK T-cells and ILC2 with subtypes of immune cells and serum factors depicted as scatter plots. (method = Spearman, numbers display Spearman rank-order correlation coefficients (rho-values) and p values

Fig. 7

Correlation analysis CD11b+ CD1a+ macrophages with serum factors depicted as scatter plots. (method = Spearman, numbers display Spearman rank-order correlation coefficients (rho-values) and p values

Fig. 8

Comparison of frequency of leucocytes in colon samples of UC and Non UC patients. Leucocytes were isolated from colons of UC patients (n = 4) and Non UC patients (n = 5) undergoing colectomy. a Frequency of leucocyte; b Frequency of respective parent. Bars represent mean values and SD. A two-sided t-test and a significance level = 0.05 was used to compare groups