Colonic mucosal gene expression in inflammatory bowel disease: From whole genome to REG gene family expression
MetadataShow full item record
Inflammatory bowel disease (IBD) is a group of chronic inflammatory diseases of the digestive tract. The two main forms of IBD are ulcerative colitis (UC) and Crohn’s disease (CD). As IBD are chronic diseases, often with an onset at a young age, there are many patients enduring the often debilitating symptoms associated with the disease for decades. The aetiology is largely unknown, and the pathology is poorly understood. Treatment alternatives are few, and with varying degree of efficacy. Today approximately 2.2 million Europeans live with an IBD diagnosis. While UC only occurs in the colon, CD can affect any part of the digestive tract. UC usually affects the rectum and the distal parts of the colon, expanding in continuity towards the small bowel. The extent of inflammation varies from only affecting the rectum (then called ulcerative proctitis) to the whole colon from rectum to caecum (then called total colitis). In contrast to UC, CD inflammation is not limited to the colon and often appears as patches of inflamed tissue interspersed by stretches completely spared from inflammation. The inflammation is often deeper than what is seen in UC, and spreads into the intestinal wall, sometimes leading to the formation of fissures and fistulas. Both UC and CD can have extraintestinal manifestations. The current understanding of the aetiology of IBD is that it is a result of an aberrant inflammatory response, partly due to genetic variation making the affected individual susceptible to the disease, possibly resulting in lowered tolerance towards otherwise benign stimuli. What initiates this first IBD inflammation is unknown, but both the microbial and nutritional environment are suggested as initiating factors. In paper I, the mRNA gene expression of both UC and CD colonic mucosa is explored using whole genome gene expression microarrays. The expression levels of both inflamed (CD and UC) and un-inflamed (CDU and UCU) samples are contrasted against the expression levels found in normal control samples (N) and the expression profiles of both diseases are identified. Surprisingly, although both diseases display several thousands of genes differentially expressed from normal controls, there is almost no significant difference between the gene expression in inflamed UC and CD mucosa. There was little difference between un-inflamed IBD mucosa and normal controls, with only the contrast CDU vs. N showing a few differentially expressed genes. The lack of difference between CD and UC profiles is surprising seen in light of the many previous studies describing such differences on a colonic mucosal gene expression level. To elucidate this discrepancy further, the data from ten different data sets from whole genome analyses of colonic mucosa were used in a metaanalysis. All data sets were analysed using the same contrasts and statistical methods used in the initial analysis, and resulting sets of gene lists were compared. The hierarchical clustering of gene lists showed no clustering of UC and CD gene lists, instead grouping gene lists originating from the same studies closely. This observation lends itself to the initial observation of little detectable difference between the expression pattern in UC and CD colonic mucosa. The meta-analysis focused further on finding consensus over all studies. By focusing on consensus, we hoped to reduce the data from all data set to a subset of information confirmed by reproduction in several studies. Furthermore, by focusing the efforts towards subjects known to be of great importance in IBD, namely T helper cells and antimicrobial peptides, we wanted to find sound, reproducible expression profiles describing these subjects. T helper cells are important parts of the adaptive immune response, and have been extensively studied in the IBD setting. The T helper cells can be grouped in four, depending on their function; Th1, Th2, Th17 and Treg. Previously, UC and CD were thought to differ in this respect, with CD seen as a Th1-driven disease, and UC as an atypical Th2 disease. Today, however, this view is being revised, and many see both UC and CD inflammation as Th1/Th17-dominated. Our analysis confirms the latter view, again showing similar expression profiles for UC and CD over all sets for genes related to T helper cell differentiation. The expression profiles of both diseases are dominated by Th1 and Th17-related expression, with some Treg and little Th2-related expression. The two diseases are similar in most respects, with one interesting exception. The IL23-subunit, IL23A consistently showed differential expression in UC gene lists and not in CD. This is particularly interesting seen in connection with previous research underlining the role of genetic variation the IL23 receptor in IBD. Antimicrobial peptides are peptides produced in the mucosal lining of the digestive system, released as constituents of the mucus lining. These peptides act as a first line of defence, controlling the microbial load of the mucus lining through a diverse array of antimicrobial functions. Our analysis confirmed a previous analysis of these peptides, but over the whole set of available data. Most antimicrobial peptides are either unchanged or highly expressed in IBD, with the exception of DEFB1 and LEAP2. This down-regulation is also previously reported, and was then attributed to the loss of differentiated colonocytes in IBD mucosa. However, the expression of DEFB1 has also previously been demonstrated as controlled by PPARG, and our data shows a strong correlation between the expression of DEFB1 and PPARG, suggesting that the latter explanation is the effect seen in IBD. Finally, the initial analysis of our data showed a group of genes in the REG gene family as highly up-regulated. This observation was explored further in papers II and III. In paper II, immunohistochemical (IHC) methods were used to evaluate the protein expression pattern of the REG proteins REG1A and REG4. The REG family proteins are highly expressed in many settings, but their function is poorly understood. We used sections from inflamed and un-inflamed colonic mucosa, as well as inflamed mucosa from non-IBD inflammation (antibiotic-induced pseudomembraneous colitis) and healthy controls, in a series of IHC labelling reactions using antibodies against REG1A, REG4, serotonin (a marker for neuroendocrine cells) and DEFA6 ( a marker for Paneth cells). All inflamed tissues showed expression of both REG1A and REG4, demonstrating that the REG expression is not IBD-specific, but possibly rather a general response during colonic inflammation. Also some un-inflamed IBD samples showed REG positivity. REG1A was most highly expressed towards the base of the colonic crypts, most strongly in Paneth cells. REG4 was expressed in the upper half of the crypts in inflamed samples, and seemed to be excreted as constituents of the mucus formed in goblet cells. REG4 positivity was also observed in serotonin-positive cells in both inflamed and un-inflamed samples. In paper III we continued our exploration of REG expression in IBD. We here perform the first comprehensive study demonstrating the expression localisation of all human REGs in colon mucosa, forming a sound background for further exploration of REG protein function. In situ hybridization (ISH) was used to detect the location of REG mRNA expression in a similar set of samples as in paper III. We utilized a new protocol for ISH (RNAscope from Advanced Cell Diagnostics), with improved specificity, sensitivity and ease-of-use compared to traditional methods. By using ISH we circumvent the problem with antibody availability encountered in paper II, while also opening for the detection of possible differences in protein and mRNA expression localization. As in paper II, DEFA6 and serotonin were used as cell type markers. The analysis confirmed the expression pattern of REG1A and REG4 on an mRNA level. REG1B and REG3A displayed an expression pattern similar to REG1A, possibly slightly more restricted to the deeper parts of the crypts. REG1A, REG1B and REG3A expression is suggested to be dependent of Paneth cell metaplasia, as DEFA6-positive cells resided in the base of almost all crypts where epithelial cells were positive for these REGs. Seen together, the data from papers II and III suggest that REG4 is expressed in well differentiated colonocytes in the upper half of the colonic crypts. This expression pattern fits with what would be expected by a migration stimulating factor for differentiating epithelial cells moving towards the lumen from the proliferative zone in the crypt base, a function previously suggested for REG4. However, the fact that REG4 is secreted as a constituent of mucus fits with what would be expected for an AMP, an action that meets what would be expected based on REG4 amino acid sequence and the fact that antimicrobial effect has been demonstrated for other REGs previously.