Mir-146a regula la interacción entre células epiteliales intestinales, microbiota y estímulos inflamatorios

Resumen

1. INTRODUCTION The intestinal homeostasis depends on the interaction between bacteria and the intestinal epithelium. This fact is clearly manifested in intestinal inflammatory diseases in which dysbiosis is known to play a pathogenic role (1) In addition, dysbiosis has been related to systemic diseases such as fatty-liver disease, obesity or diabetes (2,3) The current view indicates that the intestinal immune system and the microbiota maintain a collaborative alliance in which the microbiota modulates the immune system and, in turn, the latter tolerates microbiota and fights off invasive pathogenic bacteria (4). This partnership is possible to a great extent due to non-specific receptor in innate immune cells that recognize bacterial antigens. Among these receptors, TLRs are the best characterized, together with the retinoic acid-inducible gene I (RIG-I)-like receptors and NLRs (5,6). Intestinal epithelial cells (IECs) are considered components of the intestinal innate immune system that play a major role in the cross-talk with the microbiota and the development of tolerance. In fact, IECs express TLRs, but under homeostatic conditions expression is low, and the effects of their ligands are correspondingly attenuated (7). Nevertheless, under inflammatory conditions TLR expression is increased and contributes to inflammation and immune tolerance (8). In the last few years microRNAs (miRNAs) have been shown to be involved in the regulation of the inflammatory response by microbiota-derived antigens. These small non-coding RNAs regulate gene expression and, as such, are involved in the pathogenesis of several diseases (9,10). A general mechanism of action of miRNAs involves the regulation of protein expression both at the transcriptional and translational level. Among the miRNAs associated to the inflammatory response and the microbiota, miR-146a is strongly induced by endotoxin through the stimulation of TLR4 in different cell types (11,12). Furthermore, it has been shown to participate in the regulation of the immune response limiting it to prevent its overstimulation (13–15). The response of miR-146a knockout mice to endotoxin has given proof of the negative feedback regulatory loop that controls pro-inflammatory signaling, as miR-146a knockout mice are hyperresponsive to lipopolysaccharide. Thus, miR-146a is considered a negative effector of the innate immune response (14). Although there are not many studies assessing the role of miR-146a in intestinal inflammation, a study indicated that miR-146a also protects against intestinal ischemia/reperfusion injury (16). Nevertheless, late findings have introduced controversy regarding the role of miRNA-146a in the intestinal inflammation. A recent study indicates that knockout animals are resistant to DSS induced colitis, a chemical model in which the intestinal barrier function is disrupted. The current explanation for this apparent contradiction is based on the ability of miR-146a to inhibit both inflammatory and intestinal barrier related genes (14). In this study we aimed to better understand the regulation of miR-146a in intestinal inflammation studying its expression in animal models of colitis, its role in intestinal inflammation and the maintenance of immune tolerance by IECs. Characterization of the function of miR-146a in intestinal inflammation and tolerance will contribute to better understand these processes and their contribution to the pathogenesis of inflammatory diseases like inflammatory bowel disease and even of systemic diseases in which a role for intestinal microbiota has been suggested. 2. AIMS Based on all the above, we proposed 5 main objetives in this doctoral thesis: 1. To determine changes in miRNA expression in human intestinal epithelial cells, depending on the stage of cell differentiation and inflammation. 2. To study miR-146a expression in three animal models of colitis, that differ in the involvement of the intestinal mucosal barrier and, accordingly, in the level of contact with the intestinal microbiota. 3. To study the effect of bacterial antigens and cytokines on the expression of miR-146a using IEC lines. 4. To confirm the hypothesis that miR-146a may contribute to prevent overstimulation of the immune response. 5. To study the effect of bacterial antigens and cytokines on the expression of miR-146a using organoids from mouse small intestine. 3. MATERIAL AND METHODS To carry out these objectives, we used a wide variety of techniques, including cell and tissue culture and isolation techniques, real time PCR, microarray, histology techniques, shRNA gene silencing, Western Blot, overexpression of has-miR-146a-5p and ELISA as well as different experimental models of colitis (TNBS, DSS and lymphocyte transfer model). Differences among means were tested for statistical significance using Student’s t-test when two groups were compared or by one-way ANOVA and a posteriori Fisher-LSD test on preselected pairs. Analyses were carried out with the SigmaStat 3.5 program (Jandel, San Jose, CA, USA). Differences were considered significant at p < 0.05. The functional analysis of microarrays was carried out by the program MeV (MultiExperiment Viewer) which allows to analyze microarray data incorporating algorithms for clustering, visualization, classification and statistical analysis in heat maps. 4. RESULTS • miRNAs: differentiation and inflammation We used a specific microarray analysis for miRNAs involved in differentiation which includes 84 miRNAs. We cultured Caco-2 cells at 2 (preconfluence), 7 (confluence) and 17 (postconfluence) days for testing differentiation. The monolayers were cultured for the last 24 hours under proinflammatory conditions adding Il-1β and TNF to the medium. Compared with the control sample (monolayers at day 2), 40 miRNAs in the sample collected at 7 days and 75 miRNAs in the sample collected at 17 days were overexpressed (Figure 2). The clustering of miRNAs by families as miRBase (http://www.mirbase.org/) indicated that between miRNAs whose expression is augmented with differentiation are those belonging to families let7, miR1, miR15 or miR10 (Figure 3). Moreover, of the 84 miRNAs studied, only miR302 was inhibited in differentiated cells. Meanwhile, IL-1β and TNF induced variations in expression in most miRNAs analyzed (Figure 1). Among those genes induced by TNF and IL-1β include the families of miR10, miR15 and family members and cluster mir17 miR17-92 (Figure 4). In figure 4 they are clustered miRNAs that were more overexpressed. It is important to note that the most overexpressed miRNA was miRNA 146a, being one of the reasons why we decided to continue studying its role in inflammatory bowel disease. •miR-146a is overexpressed in several models of colitis We assessed the expression of miR-146a in two different and widely used chemically induced colitis models, namely TNBS colitis in rats and DSS colitis in mice. In addition, we evaluated miR-146a levels in Rag1-/- mice transferred with CD4+CD62L+ T cells from regular donor mice i.e. lymphocyte transfer colitis. In the chemically induced colitis models the inflammatory reaction starts immediately or shortly after the administration of the colitogenic agent (TNBS or DSS) and is not strictly chronic (i.e. animals heal with time), while the transfer model is characterized by a slow, insidious onset (6-8 weeks) and a true chronic course. In addition, the latter model is lymphocyte driven as human inflammatory bowel disease (IBD), while both TNBS and DSS colitis develop virtually normally in the absence of lymphocytes. Colonic expression of miR-146a as determined by qRT-PCR was upregulated in TNBS and DSS models but not in the transfer model (Figures 9, 14 and 17). • miR-146a is overexpressed in differentiated intestinal epithelial cells Crypt and villus cells from mouse jejunum were isolated and the expression of miR-146a was measured by qRT-PCR (Figure 17a). Levels were found to be up-regulated in villi when compared to crypts. As expected, the expression of iAP, a marker of IEC differentiation, followed the same pattern (Figure 17b). • Proinflammatory cytokines and the bacterial components LPS and flagellin induce miR-146a expression in intestinal epithelial cells Caco-2 cells (human adenocarcinoma colonic cells) miR-146a expression was sharply upregulated in response to IL-1β (fold change: 49.22 p<0.05 vs. C; n=4). Similarly, quiescent IEC-18 cells (nontumoral rat ileal cells) displayed low expression of miR-46a, which shifted to a marked increase after stimulation with LPS (a TLR4 ligand), flagellin (a TLR5 ligand) or TNF (Figure 20). Conversely, no significant effect of CpG DNA (TLR9 ligand), peptidoglycan (TLR2 ligand), muramyldipeptide (NOD2 ligand) or IL-1β was noted. The expression of TLR 2, 4, 5 and 9 and NOD-2 in IEC-18 cells was also studied and was directly correlated with the observed stimulation of miR-146a, being TLR4 and TLR5 highly expressed (Figure 22). As a positive control of the effect of IEC stimulation, MCP-1 and IL-8 production were measured in IEC-18 and Caco-2 cells, respectively. The pattern of cytokine secretion paralleled that of miR-146a induction, as it was confirmed in IL-1β treated Caco-2 cells (269.13 ± 63.43 vs 3.56 ± 1.34 pg/ml, p<0,05 vs C; n=4) and in IEC-18 exposed to LPS, flagellin and TNF, but not to other stimuli (Figure 21). •Inflammatory mediators such as IL-1β and TNFα induce miR-146a expression in organoids obtained from jejunum crypts of mice. To test the effect of inflammation in a third model, closer to a physiological model than IEC-18 cells, we used a recently developed 3D intestinal culture system allowing the generation and multiplication of “mini-guts” (termed organoids) from mouse small intestinal epithelial stem cells (17). We first studied the effect of proinflammatory cytokines (IL-1β, TNFα and IFNɣ) on the secretion of GROα/CXCL1 in mouse jejunum organoids. After incubation for 24 hours of the organoids grown for one week in MatrigelTM with the respective cytokines. In general the addition of cytokines to the culture medium induced secretion GROα/CXCL1, in particular the results showed a significant increase in secretion of this cytokine when added IFNɣ and TNFα (Figure 30). •Signaling through TLR4/MyD88 induces miR-146a expression in intestinal epithelial Cells MyD88 is a TLR adapter protein used by most TLRs to activate transcription factor NF- κB with the consequent induction of proinflammatory cytokines. To confirm that TLR ligand evoked induction of miR-146a depends on MyD88, gene knockdown with interference RNA technology was applied to IEC-18 cells. Silencing MyD88 expression in IEC-18 cells resulted in the inhibition of miR-146a associated with a decrease of MCP-1 induction by LPS (Figures 25 a-c). In addition, TLR4 expression was similarly knocked down in IEC-18 cells, resulting as expected in an attenuated MCP-1 response to LPS, and downregulated miR-146a expression (Figures 25 b-d). • Overexpression of miR-146a inhibits cytokine production in response to LPS and IL-1β in intestinal epithelial cells A role limiting the immune response to prevent overstimulation has been attributed to miR-146a (13) In order to test the validity of this hypothesis in intestinal epithelial cells, we overexpressed miR-146a in both IEC-18 and Caco-2 cells and studied the effect on cytokine production (Figure 26). Transfection of a small RNA mimic of miR-146a was employed for this purpose. Overexpression of miR-146a in IEC-18 inhibited the production of MCP-1 and GROα in the culture medium after 24 h of incubation both in basal conditions and after LPS stimulation (Figure 26a). The studies with Caco-2 cells stimulated with IL-1β confirmed these results, showing a decreased production of IL-8 and MCP-1 in cells overexpressing miR-146a after IL-1β stimulation and in basal conditions (Figure 26b). 5. CONCLUSIONS 1. The expression of miRNAs is regulated in differentiation and proliferation processes in the intestinal epithelial cells Caco-2. These same miRNAs are also regulated by proinflammatory cytokines. In particular, in differentiated cells, ie with low proliferative levels, the expression of most of the miRNAs analyzed in this study is increased. This is the case of tumor suppressors miRNAs such as family let7, miR1, miR10 or miR15. Moreover, of the 84 miRNAs studied, only miR302, whose expression has been related to cell reprogramming, was inhibited in differentiated cells. IL-1β and TNF induce the expression of tumor suppressor miRNAs belonging to the families miR10, miR15 and miR17, while inhibit family let7 expression. This observed link between specific miRNAs and processes of inflammation and epithelial differentiation suggests an association between inflammation and cancer, which is consistent with the literature. 2. miR-146a is involved in maintaining the intestinal barrier function. This conclusion is based on: a) miR-146a expression is induced under conditions of proliferation and in response to proinflammatory cytokines IL-1β and TNF in intestinal epithelial cells and colonic inflammation in vivo in which there is epithelial damage. b) miR-146a expression is induced by self molecules of intestinal microbiota in human and murine intestinal cell lines and small intestinal mouse organoid, it includes LPS and flagelin (bacterial antigens and TLR4 and TLR5 ligands) c) miR-146 helps the innate tolerance inhibiting proinflammatory cytokine expression in response to bacterial antigens.