Search in the Abstract Database

Search Abstracts 2012

* = Presenting author

P074. Mucosal mRNA expression profiling from the terminal ileum and colon reveals under‑expression of claudin 8, a tight junction molecule, as potentially causal in ulcerative colitis


P.J. Smith1, A.P. Levine1, G.W. Sewell1, N.R. O'Shea1, R. Vega2, S.L. Bloom2, A.M. Smith1, A.W. Segal1

1University College London, Division of Medicine, London, United Kingdom; 2University College London Hospitals NHS Foundation Trust, Department of Gastroenterology, London, United Kingdom



Background: Intestinal barrier dysfunction plays an important role in the pathogenesis of ulcerative colitis (UC). We investigated mRNA profiles of mucosa from the colon and terminal ileum, in patients with UC and controls (HC) to identify genes that might be implicated in the pathogenesis of the disease.

Methods: Mucosal biopsies were taken from 24 quiescent UC patients (Mayo score <3) and 33 HCs undergoing colonoscopy. Patients were on no treatment or on 5‑aminosalicylates ± azathioprine. HCs were patients without organic disease. Parallel biopsies were taken for RNA extraction and histology from macroscopically non-inflamed mucosa in the terminal ileum (TI), ascending, descending and sigmoid colon, and rectum. cRNA was hybridised to Illumina HumanHT v12.0 Expression Beadchips.

Expression data were log transformed and normalised. Probes with a detection p‑value <0.01 were analysed. Comparing 85 biopsies from HCs and 68 biopsies from UC patients across the colon, the data for each bowel location were adjusted to the mean HC rectal expression level. Where multiple biopsies were taken from the same individual, the adjusted data across all biopsies for that individual were averaged. T tests between groups and outlier analysis (p < 0.005, fold change (FC) >1.5) were performed using proprietary software.

Results: Of the ∼30K probes analysed, the two most significantly under-expressed in UC in the colon were those of claudin 8 (CLDN8) with FCs 2.94 (p = 1.29×10−5) and 3.45 (p = 3.92×10−5). The expression of claudin 8 increased distally in the colon, whereas claudins 3, 7 and 23 were highly, and uniformly, expressed throughout the colon and were normal in UC. Outlier analysis between HC and UC showed CLDN8 to be significantly under-expressed in 25%-40% of UC patients at all 4 colonic sites. There were no CLDN8 UC outliers in the TI. Most of these outlier patients demonstrated consistent levels of under-expression throughout the colon, and their histology revealed microscopic inflammation. Other patients with histologically active disease had normal CLDN8 expression.

Conclusions: CLDN8 is significantly under-expressed in the UC colon. Outlier analysis has also identified a group of patients in whom CLDN8 is grossly under-expressed. Low expression of CLDN8 in UC could be secondary to inflammation, although the evidence presented here is against this. Reduced levels of CLDN8 could lead to a weak and permeable mucosa predisposing to UC by reducing barrier resistance and allowing penetration by microbes.