Up-Regulation of microRNA-126 May Contribute to Pathogenesis of Ulcerative Colitis via Regulating NF-kappaB Inhibitor IκBα

Background MicroRNAs (miRNAs) are important post-transcriptional regulators. Altered expression of miRNAs has recently demonstrated association with human ulcerative colitis (UC). In this study, we attempted to elucidate the roles of miR-126 in the pathogenesis of UC. Methods Expression of miR-126, miR-21, miR-375 and the potential targets NF-κB inhibitor alpha (IκBα, IKBA or NFKBIA), Polo-like kinase 2 (PLK2) and v-Crk sarcoma virus CT10 oncogene homolog (CRK) were assessed in 52 colonic biopsies from patients with active UC, inactive UC, irritable bowel syndrome (IBS) and from healthy subjects by quantitative RT-PCR and immunofluorescence analyses. Regulation of gene expression by miR-126 was assessed using luciferase reporter construct assays and specific miRNA mimic transfection. Results We found that the expression of miR-126 and miR-21 were significantly increased in active UC group compared to the inactive UC, IBS and healthy control groups (P<0.05). In contrast, the expression of IKBA mRNA and protein was remarkably decreased in the active UC group compared with the other three groups (P<0.05). The expression of miR-126 and IKBA mRNA were inversely correlated in active UC patients (P<0.05). However the expression of miR-375, PLK2 and CRK showed no difference between each group. Furthermore, we demonstrate that endogenous miR-126 and exogenous miR-126 mimic can inhibit IκBα expression. Finally, mutating the miR-126 binding site of the IKBA 3′-UTR reporter construct restored reporter gene expression. Conclusion miR-126 may play roles in UC inflammatory activity by down-regulating the expression of IKBA, an important inhibitor of NF-κB signaling pathway.


Introduction
Ulcerative colitis (UC) is one of two major types of inflammatory bowel diseases (IBD) and is presented as a chronic, relapsing and remitting inflammatory disease limited to the mucosal layer of the rectum and colon. An increasing incidence of UC has recently been reported worldwide in the last 20 years, especially in China [1]. Pathogenesis of IBD is not yet fully understood. It is currently proposed that the pathogenesis of IBD involves genetic susceptibility variation, inappropriate immunoresponse to microbes in the gut and undetermined environmental factors. Furthermore, pathogenesis of UC may vary in different ethnic populations and demographic regions. Genome-wide association studies have identified multiple susceptibility loci with high risk of IBD development [2][3][4][5]. Gene expression studies found a significant set of genes differentially expressed in patients with UC compared to healthy individuals [6][7][8][9]. However, it is still unclear how such differential expression of genes is regulated. microRNAs (miRNAs) have recently been recognized as important post transcriptional regulators. Mature miRNAs are a class of small, non-coding RNA molecules in length of 20-25 nucleotides(nt). miRNAs bind to sequences located in the 39untranslated region (39-UTR) of target mRNAs in the RNAinduced silencing complex and regulate mRNA degradation or repress their translation. miRNAs have been proven to play critical roles in many biological processes such as cell differentiation, proliferation, apoptosis and tumorigenesis [10][11]. Mounting evidence has shown that miRNAs are critically involved in the regulation of inflammatory and immune responses. Inflammatory cytokines and microbial components (flagellin and lipopolysaccharide, LPS) induce expression of miRNAs such as miR-146a, miR-155 and miR-132, via the transcription factor nuclear factor-kappaB (NF-kB) pathway in myeloid cells [12,13]. It has been shown that miR-146a modulates Toll like receptors (TLR) and tumor necrosis factor-a (TNF-a) signaling pathways through a negative feedback loop, by reducing expression of Interleukin-1 receptor-associated kinase 1 (IRAK1), TNF receptor-associated factor 6 (TRAF6) and cyclooxygenase-2 [13][14][15]. miRNAs including miR-155 and miR-150 are shown to influence the fate of immune cells, and to regulate adaptive immune response such as antigen presentation and T-cell receptor signaling [12,16,17].
In this study, we aimed to characterize the roles of miR-126 in the pathogenesis of human UC. We first found that miR-126 was significantly increased in the colons of active UC patients compared to other control groups. We also found that among three potential targets of miR-126, nuclear factor-kappaB inhibitor alpha (IkBa, also named as IKBA, NFKBIA) was decreased and inversely correlated with miR-126 expression in colons with active UC. Furthermore, we demonstrated that endogenous miR-126 and exogenous miR-126 mimic can inhibit IkBa expression. Our findings suggest that miR-126 may play an important role in the regulation of inflammation in chronic inflammatory disease, including UC.

Human Tissues
Colonoscopic pinch biopsies from the sigmoid colon of 52 patients with chronic active UC, chronic inactive UC, irritable bowel syndrome (IBS), and normal, healthy patients undergoing screening colonoscopies were obtained from The Affiliated Hospital of Guangdong Medical College, Zhanjiang, China. Biopsies from all patients were assessed between September 2008 and December 2009 and the clinical characteristics of the patients are indicated in Table 1. The diagnoses of active UC and inactive UC were confirmed by colonoscopy and pathological findings. Disease activities were determined using the Ulcerative Colitis Disease Activity Index (UCDAI) [28], which is a series of qualifiers about the symptoms of ulcerative colitis including stool frequency, rectal bleeding, appearance of the lining of the colon, and a physician rating of disease activity. Each of these items was given a number from 0 to 3, with 3 being the highest for rating of disease activity. Ratings of 0-2 were recorded as remission, 3-6 as mild, 7-10 as moderate,.10 as severe UC respectively. Patients with UCDAI .8 were included in the study.

Isolation of Total RNA
Colonoscopic biopsies or cultured cells were rapidly transferred into 1.0 ml of TRIzol reagent (Invitrogen, Carlsbad, CA) and total RNA was extracted according to the manufacturer's instructions and stored at 280uC until used.

Real-time Quantitative Reverse Transcriptase-polymerase Chain Reaction (qRT-PCR)
The NCode TM VILO TM miRNA cDNA Synthesis Kit(Invitrogen, Carlsbad, CA)and ImProm-II TM Reverse Transcription System(Promege, Madison, USA)were used to reverse-transcribe the miRNA and mRNA, respectively. For qRT-PCR on biopsy tissues, 500 ng of total RNA was converted to cDNA. 2 ml of cDNA from each sample was amplified by real-time fluorescent qPCR, using SYBR Premix Ex Taq TM (Takara Bio, Inc., Tokyo, Japan). The expression of each target miRNA in tissues was calculated relative to U6B, a ubiquitously expressed small nuclear RNA used as internal control. Expression of beta-actin was used as normalization control in mRNA qPCR. For miRNA qPCR, the reverse primer was the NCode miRNA universal qPCR primer (Invitrogen, Carlsbad, CA). Forward miRNA primers and mRNA primers were compound by Sangon Biotech Co, Ltd (Shanghai, China) and are listed in Table 2. A comparative threshold cycle method was used to compare each condition with controls.

Immunofluorescence
Immunofluorescencet staining was performed according to standard protocol (Santa Cruz Biotechnology). Briefly, a 5-mm frozen section from each case was fixed in acetone at 4uC for 10 min and blocked with 10% goat serum albumin at room temperature (RT) for 30 min. the sections were incubated with rabbit anti-human IkBa primary antibody (1:400) (Santa Cruz Biotechnology) overnight at 4uC. After rinsing in phosphatebuffered saline (PBS) for three times, the sections were incubated with secondary polyclonal Cy3 labeled goat anti-rabbit-594 IgG (1:400) (Zhongshan Goldenbridge Biotechnology. Beijing, China) for one hour at RT in dark, and then washed with PBS. PBS replaced anti-IkBa antibody as a negative control. 49,6-diamidino-2-phenylindole (DAPI) was used to counter-stain the nuclei for 5 min at RT. The staining was evaluated on a Leica converted fluorescence microscope and the fluorescence intensity was measured using Image-Pro Plus 6.0 software. We quantified the total Cy3-fluorescence intensity (as shown in red) of the mucosa and submucosal layers as representation of protein expression level of IkBa. Intensity of DAPI staining (blue) was used as an internal normalization control for adjusting fluorescence signals among different slides.
The wild type and mutant sequences were then cloned into the XhoI and NotI sites downstream of the Renilla luciferase gene of the dual-luciferase report vector psiCHECK-2 (Promega Corpo-ration, Madison, USA) according to manufacturer's protocols. The wild type and the mutant constructs were named as psi-IKBA and psi-mutIKBA, respectively. The sequences of the constructs were verified by DNA sequencing. The constructs were transformed in E.coli JM109 and the plasmid DNA was purified using TIANpure MidiPrep kit (Tiangen Biotech, Beijing, China).

Construct Transfection and Luciferase Report Assay
Human colonic cancer cell line HT29 and HCT116 (Beijing Institute For Cancer Research, China) were maintained in DMEM containing 10% FBS at 37uC in a 5% CO 2 incubator. HT29 cells were seeded in 24-well plates overnight before transfections. Luciferase report constructs (100-250 ng/well) were transfected into cells using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) according to the manufacturer's protocols. Cells were harvested at 24 hours after transfection. Renilla and Firefly luciferase activites were measured using the Dual Luciferase Reporter Assay System (Promega Corporation, Madison, USA) according to the manufacturer's instructions. Experiments were performed in quadruplicate wells and datas are represents of three independent experiments. miRNA Mimic Transfection HT29 cells were seeded into 24-well plates at 90% confluence and kept in a 37uC and 5% CO 2 incubator overnight. Varying amounts of miR-126 mimic or negative control mimic (RiBoBio, Guangzhou, China) were transfected into HT29 cells using Lipofectamine 2000. The transfected cells were incubated at 37uC in a 5% CO 2 incubator for 24 hours. Cellular total RNA and protein were harvested separately, and stored in 280uC until used.
To determine the specific effect of miR-126 on 39-UTR of IKBA, the miR-126 mimic and the psi-IKBA luciferase construct were co-transfected into HT29 cells using Lipofectamine 2000. Renilla and Firefly luciferase levels were measured at 24 hours post-transfection using the Dual Luciferase Reporter Assay System. Experiments were performed in quadruplicate wells and datas are represents of three independent experiments.

Western Blot
Western blot was performed according to standard protocol. Briefly, 15 mg protein per samples were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to polyvinylidene fluoride (PVDF) membrane. The membrane was probed with rabbit anti-human IkBa primary antibody (1:1000 dilution. SC-371. Santa Cruz Biotechnology, Santa Cruz, CA, USA) overnight at 4uC. The membrane was then incubated with HRP-labeled Goat Anti-Rabbit IgG (1:1000. Beyotime, Jiangsu, China) at RT for one hour. Mouse anti-human Tubulin antibody was used to detect tubulin as a loading control. The integral of optical density of bands was measured using Quantity One image analysis software (Bio-Red, Hercules, CA, USA).

Statistical Analysis
Experimental results are expressed as mean values 6 standard error. Statistical analysis for qRT-PCR and Immunofluorescence

Ethical Statement
The study was approved by Institutional Review Board of the Affiliated Hospital of Guangdong Medical College. All patients gave written informed consent for their participation.

Differential Expression of miR-126, miR-21 and miR-375 in UC Tissues
A genome-wide microarray screening indicated that miR-126 and miR-21 are increased and miR-375 is decreased in sigmoid colons of active UC patients [18]. To determine whether miR-126, miR-21, and miR-375 are differentially expressed in UC patients from different ethnic groups, we analyzed the expression of these candidate miRNA in Chinese patients with active UC.
A total of 52 sigmoid colon pinch biopsies from patients with histologically confirmed active UC, inactive UC, IBS and normal healthy control subjects were obtained. Clinical characteristics of each patient group are listed in Table 1. We performed quantitative real-time PCR on total RNAs isolated from these biopsies. We found that expressions of miR-126 and miR-21 were significantly increased in active UC tissues compared to healthy control tissues (Figure 1 A and B). The expression of miR-126 was increased 18-fold (P,0.05), while the expression of miR-21 exhibited a 14.7-fold increase (P,0.05), in active UC tissues compared to healthy controls, respectively. However, expressions of miR-126 and miR-21 were not significant changed in either inactive UC or IBS tissues compared to healthy controls (both P.0.05 respectively) ( Figure 1). The results are consistent with findings by previous microarray analysis [18]. However, there was no statistical significant difference in expression of miR-375 among the sample groups tested (P.0.05) ( Figure 1C).

Selection of Potential Downstream Targets of miR-126
In order to identify downstream targets of miR-126, we used three miRNA targets prediction programs including MicroCosm Targets, Targetscan and PicTar. In silico analysis indicated approximately 900 mRNA targets of miR-126. Among these targets, Polo-like kinase 2 (PLK2, a tumor suppressor) and v-Crk sarcoma virus CT10 oncogene homolog (CRK) were shown as targets of miR-126 in previous studies [30][31][32][33]. However, there currently is no evidence that PLK2 and CRK are association with UC.
Interestingly, IkBa, an inhibitor of NF-kB pathway, is also one of the predicted targets of miR-126. A putative miR-126 binding site that encompasses 11 perfectly matched nucleotides was defined at the 39-UTR (un-translated region) of IKBA ( Figure 2).
In an effort to delineate potential downstream targets and further understand the mechanisms of miR-126 over-expression for pathogenesis of UC, we selected three putative targets (IKBA, PLK2, CRK) for further molecular and functional analyses.

Both IKBA mRNA and Protein Decreased in Active UC Tissues
To experimentally identify the targets of miR-126 in colon tissue, we performed qRT-PCR analysis on the expression of IKBA, PLK2, and CRK in total RNA from the same 52 sigmoid colon pinch biopsies. Among the three candidates, we found that only IKBA mRNA was markedly decreased in active UC tissues compared to healthy controls. A 32% decrease in IKBA mRNA expression was observed in active UC tissues relative to healthy controls (P,0.05) ( Figure 3A). However, IkBa expression in either inactive UC or IBS tissues was not significantly changed as compared to healthy control tissues (P.0.05) ( Figure 3A). No significant differences for the expression of PLK2 and CRK mRNAs were observed between active UC tissues and healthy controls ( Figure 3B, C).
To determine whether the expression of IkBa protein was altered in colons with active UC, we further performed immunofluorescent staining on cryosections of sigmoid biopsies within each group. The staining results displayed that IkBa protein was clearly expressed in epithelial cells of normal colonic mucosa ( Figure 4A and Figure S1). It was markedly decreased in active UC tissues, when compared to the other three groups (P,0.05) ( Figure 4A, B and E). However, there was no differential expression of IkBa protein among inactive UC, IBS, and healthy control colon tissues (P.0.05) ( Figure 4B, C, D and E).

Inverse Correlation of IKBA and miR-126 Expression
Our findings showed that miR-126 expression was significantly increased in active UC tissues as compared to healthy controls, inactive UC and IBS patients. In contrast, IkBa expression, mRNA and protein, were decreased markedly in active UC tissues. We found that the expression of miR-126 and IKBA mRNA were inversely correlated in active UC tissues and normal controls (Pearson correlation analysis r = 20.466, P,0.01) ( Figure 5A). Such inverse correlation patterns were also present between miR-126 expression and IkBa protein level in active UC and normal control samples ( Figure 5B). These results indicate that IkBa may be a target of miR-126 in the pathogenesis of active UC.

Inhibition of IKBA Expression by miR-126
To test whether miR-126 regulates endogenous IkBa expression, the miR-126 mimic was transiently transfected into HT29 cells and then IKBA mRNA expression was assessed. IKBA  Figure 6A). However, the negative control mimic did not significantly reduce IKBA mRNA expression. In addition, we also checked the expression of CRK and PLK2 in these cells. The expression of CRK mRNA was not affected by the miR-126 mimic treatment, while PLK2 expression was reduced by about 35% in cells transfected with 20 nM of miR-126 mimic ( Figure 6B, C). These results were consistent with previous reports [30][31][32][33].
Furthermore, we found that the IkBa protein expression in HT29 cells transfected with 20 nM miR-126 mimic, was significantly inhibited by 57% as detected by western blot analysis (Figure 6 D, E). Similar inhibitory effect of miR-126 mimic on IkBa protein expression was observed in the experiments using an additional colonic cell line HCT116 ( Figure S2). These results suggest that miR-126 negatively regulates endogenous IkBa expression in colonic epithelia.

Regulation of IKBA 39-UTR by miR-126
To determine whether miR-126 directly targets the 39-UTR of IKBA mRNA, we transfected HT29 cells with an IKBA 39-UTR luciferase reporter construct containing two wild type miR-126 putative binding sites (psi-IKBA) or a mutant construct bearing mutations on miR-126 binding sites (psi-mutIKBA). Compared to the parental luciferase reporter (psi-vector), the relative luciferase activity of the wild type psi-IKBA showed 35% reduction. Mutations of miR-126 binding sites in the IKBA 39-UTR resulted in a restoration of luciferase activity to 92% of the parental reporter (P,0.01) ( Figure 7A). These results indicate that endogenous miR-126 can regulate IkBa expression by directly targeting its 39-UTR. To further confirm the findings, the miR-126 mimic was cotransfected with the above luciferase reporter constructs into HT29 cells. The miR-126 mimic dramatically reduced (.55%) the luciferase activity of the wild type IKBA 39-UTR reporter construct psi-IKBA, while a negative control mimic had no effect on luciferase activity in any group ( Figure 7B). The inhibitory effect of miR-126 mimic on the wild type psi-IKBA also showed a dose-dependent response, even though a relatively high amount of the wild type psi-IKBA construct was used ( Figure 7C).
However, miR-126 mimic did not reduce the luciferase activity of mutant construct psi-mutIKBA ( Figure 7B), indicating its specific recognition effect. These results further indicate that miR-126 can regulate IkBa expression by directly targeting its 39-UTR.

Discussion
Numerous studies have demonstrated that different subtypes of IBD possess a distinct gene expression profile and identified a number of protein coding transcripts as molecular targets for pathogenesis of human IBD. Compared to protein coding genes, the roles of microRNAs, a subset of non-coding transcripts, in initiation and progression of IBD have not been extensively characterized. The involvement of microRNAs in pathological processes of IBD was first demonstrated in a recent study which revealed a strikingly unique signature of microRNA expression in different subtypes of IBD [18], later confirmed by several studies [19][20][21][22][23][24][25][26][27]. These findings suggest the critical roles of microRNA alterations in the development of chronic inflammatory processes of human IBD. From genome-wide microarray screening, miR-126 was previously identified as one of the up-regulated microRNAs detected in active UC patients in America [18]. In this study, we specifically analyzed alterations of miR-126 in a cohort of UC and IBS patients as well as healthy controls. Among three candidate microRNAs, we found that miR-126 and miR-21 were upregulated in colonic biopsies from patients with active UC, with 18-and 14.7-fold increases respectively. This finding was consistent with a previous study [18] and further suggests that miR-126 is a crucial molecular target for development of UC.
An inconsistent finding between previous genome-wide screening [18] and our current study was observed for miR-375 expression in UC patients, which may be attributed to patient populations with different ethnic and genetic backgrounds.
Our study provides the first evidence that upregulation of miR-126 is a novel mechanism in the development of UC. Although regulation of a number of molecular targets by miR-126 has been studied, involvement of pathways regulated by miR-126 in UC is completely unknown. In cancers, miR-126 has been found to possess paradoxical roles in different tumor types. miR-126 was found to target a tumor suppressor PLK2 and further inhibit   apoptosis and increase the viability of acute myeloid leukemia cells [30]. On the other hand, miR-126 was shown to function as a tumor suppressor targeting on CRK in gastric cancer, mammary cancer and non-small cell lung carcinoma cell lines [31][32][33].
Recently, a few studies suggest that miR-126 is associated with human disorders other than cancers such as inflammation and fibrosis [34][35][36]. miR-126 was shown to down-regulate vascular cell adhesion molecule-1 (VCAM-1) expression and may provide additional control of vascular inflammation [34]. Selective blockade of miR-126 suppresses the asthmatic phenotype by diminishing TH2 responses [35]. These findings indicate that miR-126 regulates different molecular targets in different inflammatory processes.
To characterize the potential mechanism regulated by miR-126 in UC, we performed target scan analysis to identify the molecular targets of miR-126. We found that the 39-UTR of IkBa mRNA possesses a miR-126 targeted sequence. IkBa is an important inhibitor for NF-kB. IkBa binds to the NF-kB complex and further inactivates NF-kB by retaining its cytoplasmic localization. The degradation of IkBa is in response to certain pro-inflammatory cytokine like TNF and results in dissociation of NF-kB from IkBa/p50/p65 complex. Consequentially NF-kB is translocated into the nucleus and activates NF-kB-mediated transcription of genes such as cytokines and chemokines. Constant activation of NF-kB activity has been a well-proven molecular event for human IBD including UC [37][38][39]. Several studies have established IkBa as a key inhibitor for NF-kB activity in IBD [37,40]. Viral IkBa expression vectors have been shown to produce a preventive effect on NF-kB activation in mucosal macrophages and T lymphocytes as a treatment option in IBD [41]. Moreover, Teaflavin-3,39-digallate which stabilizes IkBa from degradation, can inhibit the activity of NF-kB in colon tissues [42].
Here we demonstrated that IkBa is abundantly expressed in normal colonic epithelia, but significantly reduced in active UC tissues. More interestingly, we found a significant inverse correlation between miR-126 over-expression and down-regulation of IkBa in active UC tissues. We then demonstrated that endogenous miR-126 and exogenous miR-126 mimic can inhibit IkBa expression. Finally, mutating the miR-126 binding site in the IKBA 39-UTR in a reporter construct resulted in the restoration of reporter gene expression. Taken together, our results imply that over-expression of miR-126 leads to NF-kB activation via down-regulation of IkBa, which may contribute to the pathogenesis of human UC.
NF-kB mediated pathways are complex biological processes and signaling components including IkBa in this important pathway are regulated by a number of molecules playing essential roles in different cellular functions. It was shown that TNF-a and interleukin-1(IL-1)activated IkappaB kinase-beta (IKK-beta) and further phosphorylate IkBa [43]. In addition, phosphatidylinositol (PI) 39-kinase (PI39K) can also directly lead tyrosine phosphorylation of IkBa in breast cancer cell models [44]. IkBa phosphoration consequentially results in its ubiquitination and degradation, which ultimately leads to activation of NF-kB signaling. Therefore, multiple molecular mechanisms in addition to miR-126 overexpression may potentially contribute to the regulation of IkBa expression in inflammatory processes of UC. Figure S1 Localization of IkBa in human colon tissues. Representative frozen-section was prepared from a healthy control, and stained for IkBa by immunofluorescence. IkBa is detected in the epithelial cells of healthy colon tissues. Red, IkBa; blue, DAPI nuclear staining. Pictures were imaged at 640 magnification on a Leica converted fluorescence microscope. (TIF) Figure S2 Effects of miR-126 mimic on expression of IKBA in HCT116. HCT116 cells were transfected with 20 uM of miR-126 mimic or negative control mimic for 24 hours. (A) Expression of IkBa proteins were detected by Western Blot. Tubulin detection was served as loading reference. (B) The integral of optical density of (A) was measured using Quantity One program and normalized to corresponding density of Tubulin band. Data is presented as mean 6 SEM of three independent experiments (*P,0.05, compared to that of negative control mimic treatment). (TIF) activity (normalized to Firefly luciferase activity) was measured 24 hours post-transfecting. (A) Relative Renilla luciferase activity was reduced in cells transfected with wild type IKBA 39-UTR construct (psi-IKBA) compared to that with the psi-vector (100 ng/well of a 24-well-plate). Relative Renilla luciferase activity was restored in cells transfected with mutant construct (psi-mutIKBA). (*P,0.05, compared to that of psi-vector). (B) Co-transfecting 10 nM miR-126 mimic with various luciferase reporter constructs (100 ng/well) into HT29 cells reduced the relative Renilla luciferase activity. (*P,0.05, compared to that of psi-vector) (C) miR-126 mimic showed dose-dependent effects on relative Renilla luciferase activity in cells transfected with wild type psi-IKBA construct (250 ng/well) for 24 hours. (*P,0.05, **P,0.01, compared to that without miR-126 mimic treatment). Data is presented as mean 6 SEM of three independent experiments. doi:10.1371/journal.pone.0052782.g007