Intestinal protein expression profile identifies inflammatory bowel disease and predicts relapse

Abstract

To date, most studies have applied individual factors as indicators of disease classification and prognosis. The aim of this study is to determine whether clustering analysis of protein expression profiles in intestinal epithelia improves classification and prognosis in patients with inflammatory bowel disease (IBD). One hundred and twenty Crohn's disease (CD) patients, 117 ulcerative colitis (UC) patients and 120 cases of nonspecific colitis provided intestinal biopsy samples for tissue microarray (TMA). Both unsupervised and supervised analyses were used for evaluation of clustering and association with relapse. There was a significant concordance between cluster groups based on immunostaining data of TMA and clinical classification in distinguishing IBD from nonspecific colitis (kappa= 0.498, p<0.001). CD27, CD70, CD40, TRAF3, TRAF4 and TRAF2 presented similar immunostaining features, which were different from clusters of CD154, CD80 and TRAF5. Moreover, higher expression of TRAF2 was a predictor of relapse in patients with UC (p=0.006).Thus, protein expression profiles can distinguish IBD and nonspecific colitis, and combination analysis protein expression profiles show that TRAF2 can predict relapse of UC.

Keywords: Crohn’s disease; classification; relapse; ulcerative colitis.

Figure 1

Protein expression profiles studied by immunohistochemistry on tissue microarrays. A. HE stain of paraffin blocks with 0.6 mm tissue cores. B. Examples of immunohistochemistry staining for 10 proteins. Magnification ×200 or ×400. CD, Crohn’s disease; UC, ulcerative colitis; NC, nonspecific.

Figure 2

Hierarchical clustering analysis of protein expression profiles in non-specific colitis (A), Crohn’s disease (B) and ulcerative colitis (C) as measured by tissue microarray. Graphical representation of hierarchical clustering results based on expression profiles of proteins. Rows, samples; columns, proteins. Protein expression scores are depicted according to a color scale: red, positive staining; green, negative staining; black, zero; gray, missing data. Dendrograms of samples (to the left of matrix) and proteins (above matrix) represent overall similarities in expression profiles. In the dendrogram, the length of branch between two elements reflects their degree of relatedness. A trend towards cluster of nonspecific colitis (purple dendrograms to the left of matrix, zoomed in A) is shown to classify patients with inflammatory bowel disease and nonspecific colitis. Two major protein clusters are identified (above matrix).