MicroRNA-346 Mediates Tumor Necrosis Factor-α-induced Down-Regulation of Gut Epithelial Vitamin D Receptor in Inflammatory Bowel Diseases

Abstract

Background

We recently reported that the gut epithelial vitamin D receptor (VDR) signaling inhibits colitis through inhibition of intestinal epithelial cell apoptosis, and the level of colonic epithelial VDR is markedly reduced in patients with inflammatory bowel diseases (IBD). VDR down-regulation promotes colitis, but the mechanism underlying VDR down-regulation in IBD is unknown.

Methods

VDR expression was analyzed in colon cancer cells under pro-inflammatory cytokine treatment. VDR as a target of miR-346 was confirmed using colon cancer cell culture. The relationship among inflammation, miR-346 and VDR was assessed in human IBD biopsies and experimental colitis models.

Results

We showed that TNF-α suppresses VDR expression while simultaneously up-regulating miR-346 in human colon cancer cells. Further studies demonstrated that miR-346 inhibits VDR by a specific target sequence in the 3′ untranslated region of the human VDR transcript, and blockade of miR-346 with a hairpin inhibitor abrogates the ability of TNF-α to inhibit VDR, confirming that TNF-α down-regulates VDR by inducing miR-346. Consistently, in human IBD biopsies the reduction of epithelial VDR is associated with increased immune cell infiltration and elevation of TNF-α and miR-346. In an experimental model of colitis mucosal VDR expression is reduced over time with the progression of colitis, inversely correlated with the induction of TNF-α and miR-346 in the mucosa.

Conclusion

These data suggest that during mucosal inflammation TNF-α induces miR-346, which down-regulates epithelial VDR. Mucosal VDR reduction in turn compromises the integrity of the mucosal epithelial barrier, further driving mucosal inflammation and colitis development.

Introduction

The gut mucosal epithelial barrier separates the body from luminal microorganisms and antigens. Dysfunction of the barrier leads to increased translocation of luminal substances to the lamina propria, triggering inflammatory response ^{1; 2}. In fact, impaired gut mucosal barrier is a significant pathogenic factor for inflammatory bowel diseases (IBD) ³, which include ulcerative colitis (UC) and Crohn’s disease (CD). It is known that inflammatory cytokines produced from mucosal inflammation can in turn increase mucosal permeability by altering intercellular tight junction and inducing apoptosis of intestinal epithelial cells (IEC) ^{3; 4}. This vicious cycle of mucosal events drives mucosal inflammation and promotes the development of colitis.

The vitamin D receptor is a nuclear hormone receptor that mediates the biological activities of the vitamin D hormone, 1,25-dihydroxyvitamin D ⁵. VDR is highly expressed in the epithelial cells of the gastrointestinal tract. Recently we demonstrated that the epithelial VDR signaling plays a critical role in maintaining the integrity of the mucosal epithelial barrier by suppressing inflammation-induced IEC apoptosis ⁶. In experimental colitis models, Vdr deletion exaggerates colonic inflammation, whereas transgenic overexpression of VDR in the epithelial cells renders the transgenic mice highly resistant to colitis. This anti-colitic activity of the epithelial VDR signaling is independent of the VDR actions in the non-epithelial immune compartment ⁶. Importantly, we observed that, in both UC and CD patients, epithelial VDR levels are reduced by more than 50% in the lesion, independent of the serum vitamin D status ⁶. Based on the observations from the experimental colitis models and human biopsies we reason that epithelial VDR reduction compromises the gut mucosal barrier and contributes to the development of IBD; however, the molecular mechanism underlying epithelial VDR down-regulation in IBD is unclear. The goal of this study is to investigate this important issue. Our results indicate that epithelial VDR down-regulation is driven by colonic inflammation, and this process is mediated by miR-346.

Experimental procedures

Cell culture, treatment and transfection

HCT116 and HT29 cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum. Cells were treated with TNF-α (100 ng/ml), IL-1β (10 ng/ml) or IL-6 (100 ng/ml) for 0–48 hours, followed by isolation of total RNA or total protein lysates for analyses. In some experiments, cells were pre-treated with actinomycin D (5 nM) before these cytokine treatments. In other experiments, cells were transfected with plasmids or microRNA oligo mimics using Lipofectamine 2000 (Life Technologies) as detailed in each experiment.

Human biopsies

Collections of colonic biopsies from IBD patients and non-IBD controls were reported previously ⁶. The collection of these biopsies was approved by the Institutional Review Board of the University of Chicago (Chicago, Illinois, USA) and by the Institutional Ethical Committee of China Medical University (Shenyang, Liaoning, China), respectively. Study subjects were recruited with written informed consent from participants or their guardians. Biopsies were subjected to histological and immunohistochemical analyses. Total RNAs and lysates were prepared from the biopsies for analyses.

Experimental colitis models

IL-10(−/−) mice, which develop spontaneous chronic intestinal inflammation due to T cell mediated aberrant immune response ^{7; 8}, were obtained from Jackson Laboratory. Colonic mucosal VDR levels were examined at 3 months of age in wild-type and IL-10(−/−) mice. In other experiments, colitis was induced with 2,4,6-trinitrobenzene sulfonic acid (TNBS) in 8–12 week old C57BL/6 mice as previously described ^{9; 6}. After overnight fasting mice were treated under anesthesia with 100 mg/kg TNBS (Sigma) dissolved in 50% alcohol via intrarectal injection using a 1-ml syringe fitted with an 18-gauge stainless steel gavage needle, with 50% alcohol treatment as control. The mice were killed at days 0, 1, 2 and 3 following TNBS treatment, and colonic mucosal levels of VDR protein, TNF-α and miR-346 transcripts were analyzed by Western blotting and real time RT-PCR. All animal studies were approved by the Institutional Animal Care and Use Committee at The University of Chicago.

Histology and immunohistochemical staining

Freshly dissected mouse colon samples or human colon biopsies were fixed overnight with 4% formaldehyde in PBS (pH 7.2), processed and embedded in paraffin wax. Tissues were cut into 4-μm sections. Colonic morphology was examined by standard H&E staining. To examine VDR and inflammatory status, sections were stained with antibodies against VDR, CD4, CD11c (Santa Cruz Biotechnology) or TNF-α (Abcam) as primary antibodies, respectively, followed by incubation with FITC-conjugated anti-IgG antibodies. Antigens were then visualized using a fluorescence microscope.

Western Blot

Proteins were separated by SDS-PAGE and electroblotted onto Immobilon-P membranes. Western blot analyses were carried out as previously described ¹⁰, using antibodies against VDR (Santa Cruz Biotechnology) or β-actin (Sigma) as internal control. Total cell lysates, including cytosolic and nuclear proteins, were used for VDR assessment.

RT-PCR

Total RNAs were extracted using TRIzol reagents (Life Technologies). First-strand cDNAs were synthesized using a ThermoscriptRT kit (Life Technologies) for mRNA transcript analysis, or using NCode VILO miRNA cDNA Synthesis Kit (Life Technologies) for miR-346 quantitation. Conventional PCR was carried out in a BioRad DNA Engine (BioRad). Real time PCR was performed in a Roche LightCycler 480II Real-Time PCR System using SensiFAST SYBR No-Rox PCR kits (Bioline). Relative amounts of transcripts were calculated using the 2^−ΔΔCt formula. For mRNA transcript analysis, GAPDH was used as the internal control; for microRNA analysis, U6 was used as the internal control for normalization. Each data point was normalized to corresponding control values in data presentation. PCR primer sequences are provided in Table 1.

Table 1.

Primer sequences used in the study

Primer name	Forward (5′-3′)	Revere (5′-3′)
hTNF-α	CGAGTGACAAGCCTGTAGC	GGTGTGGGTGAG-GAGCACAT
hIL-6	AAATGCCAGCCT-GCTGACGAAC	AACAACAATCTGAGGTGCCCATGCTAC
hIL-1β	GCTGAGGAAGATGCTGGTTC	GTGATCGTACAGGTCGCATCG
hGAPDH	ACCACAGTCCATGCCATCAC	TCCACCACCCTGTTGCTGTA
mB2M	ACCGGCCTGTATGCTATCCAGAAA	ATTTCAATGTGAGGCGGGTGGAAC
mTNF–α	TCAGCCTCTTCTCATTCCTG	CAGGCTTGTCACTCGAATTT
mVDR	GATGCCCACCACAAGACCTA	CGGTTCCATCATGTCCAGTG
mmu-mir-346(5p)	TGTCTGCCCGAGTGCCTGCCTCT
U6	GATGACACGCAAATTCGTGAA
hsa-mir-346(5p)	TGTCTGCCCGCATGCCTGCCTCT
hsa-mir-34a	TGGCAGTGTCTTAGCTGGTTGT
hsa-mir-125b	TCCCTGAGACCCTAACTTGTG
hsa-mir-17-3	ACTGCAGTGAAGGCACTTGTAG

Plasmids

A 1.6 kb DNA fragment within 3′ untranslated region (3′UTR) between nucleotides 1444 to 3062 in human VDR cDNA was amplified by PCR using primers 5′TGATCTAGAGTGTTTGGCAATGAGATCTCCTGACTAGGA3′ (forward) and 5′CTGATCTAGAGAGCAAGCTGGACCAATACGTTCTGCAGCT3′ (reverse) and cloned into the XbaI site at the C-terminal downstream of the luciferase reporter gene in pGL3-Luc Control plasmid (Promega). The resultant plasmid was designated as pGL3-Luc-3′UTR. pGL3-Luc-3′UTR-Mut, in which the miR-346 target sequence (5′AGUGGGGGCAGAGGGCAGAGG3′) within the 3′UTR fragment was mutated at the seed sequence (5′AGUGGGGGCAGAaaatgacGG3′), was generated by site-directed mutagenesis. pCDH-miR-346 that expresses human miR-346 sequence was provided by Dr. Y-Y Mo (University of Mississippi Medical Center).

Luciferase reporter assays

HCT116 cells were co-transfected with luciferase reporter and plasmids or hsa-miR-346 oligo mimic using Lipofectamine 2000 as indicated in each experiment. Some transfected cells were treated with 1,25(OH)₂D₃ (20 nM). After 24 hours transfected cells were lysed and luciferase activity determined using the Dual-Luciferase Reporter assay system (Promega) as reported previously ¹¹.

Statistical analysis

Data values were presented as means ± SEM. Statistical comparisons were carried out using unpaired two-tailed Student’s t-test or one-way analysis of variance (ANOVA) as appropriate, with P< 0.05 being considered statistically significant.

Results

To test the hypothesis that inflammation down-regulates VDR expression, we examined the effect of pro-inflammatory cytokines on VDR expression in colon cancer cells. As shown in Figure 1, among the cytokines tested (TNF-α, IL-1β and IL-6), only TNFα substantially suppressed VDR protein levels (by >45%) in HCT116 cells, a human colon cancer cell line (Fig. 1A and B). Further investigations showed that TNFα suppressed VDR expression in dose-dependent (Fig. 1C and D) and time-dependent (Fig. 1E and F) manners. TNF-α also suppressed VDR in HT29 cells, another human colon cancer cell line (Fig. 1G and H), in which TNF-α was able to reduce VDR by more than 60%. These observations are relevant because TNF-α is a major inflammatory cytokine involved in the pathogenesis of IBD, and anti-TNF therapy is widely used in IBD management ^{12; 13}.

Figure 1.

TNF-α suppresses VDR expression. (A and B) HCT116 cells were treated with TNF-α (100 ng/ml), IL-1β (10 ng/ml), IL-6 (100 ng/ml) or PBS control (Ctrl) for 24 hours. Total cell lysates were prepared. VDR expression was measured by Western blotting (A) and quantified by densitometry (B); (C and D) HCT116 cells treated with increasing doses of TNF-α as indicated for 24 hours. VDR expression was quantified by Western blotting (C) and densitometry (D); (E and F) HCT116 cells were treated with 100 ng/ml TNF-α for different times as indicated. VDR expression was quantified by Western blotting (E) and densitometry (F); (G and H) HT29 cells were treated with increasing doses of TNF-α as indicated for 24 hours. VDR expression was quantified by Western blotting (G) and densitometry (H). * P<0.05, ** P<0.01, *** P<0.001 vs. Ctrl or 0. All values are relative to corresponding control or 0 values.

MicroRNAs (miRNAs) are naturally occurring small non-coding RNAs that control target genes by translational repression or mRNA degradation ^{14; 15}. As TNF-α quickly reduced VDR mRNA, we speculated that miRNAs might be involved in the down-regulation of VDR. By in silico analysis (using miRanda: www.microrna.org) we identified several putative miRNA binding sites within the 3′ untranslated region (UTR) of hVDR cDNA, including sites for miR-17-3p, miR-34a and miR-346. Previous studies showed that miR-125b targets hVDR ¹⁶. Thus we studied these four miRNAs. As shown in Figure 2, among these four miRNAs, miR-346 and miR-125b were markedly induced by TNFα in HCT116 cells (Fig. 2A), suggesting that TNF-α might suppress VDR via inducing miR-346 and/or miR-125b. However, when the cells were transfected with miR-125b-specific hairpin inhibitor to inhibit miR-125b, TNF-α was still able to repress VDR; in contrast, when miR-346 was inhibited, TNF-α was no longer able to down-regulate VDR (Fig. 2B and C). These observations strongly suggest that miR-346, not miR-125b, mediates the repressive effect of TNF-α on VDR expression. Consistent with this conclusion, the time course of miR-346 induction by TNF-α closely correlated with the time course of VDR down-regulation by TNF-α in HCT116 cells (Fig. 3A and B). Moreover, when the cells were pre-treated with actinomycin D, which blocks intracellular RNA synthesis, the induction of miR-346 as well as the suppression of VDR by TNF-α was markedly attenuated (Fig. 3C and D), suggesting that new RNA synthesis is required for VDR suppression. It is not surprising that miR-125b is unlikely a major mediator for VDR reduction, because even the previously published study was not able to demonstrate that miR-125b can indeed suppress VDR protein ¹⁶.

Figure 2.

TNF-α inhibits VDR expression via miR-346 induction. (A) HCT116 cells were treated with TNF-α (100 ng/ml) or PBS control for 24 hours. Indicated miRNAs were quantified by real time RT-PCR. ** P<0.01, *** P<0.001. vs. corresponding Ctrl. (B) HCT116 cells were transfected with control, miR-125b-specific or miR-346-specific hairpin inhibitor. After 24 hours, transfected cells were treated with (+) or without (−) 100 ng/ml TNF-α. Total cell lysates were prepared. VDR protein was analyzed by Western blotting. (C) Densitometric quantitation of the Western blot data. In, inhibitor. * P<0.05 vs. non-treated (ND). All values are relative to the control ND value.

Figure 3.

Inverse correlation between miR-346 induction and VDR suppression by TNF-α. (A and B) HCT116 cells were treated with TNF-α (100 ng/ml) for different hours as indicated. Changes of miR-346 (A) and VDR mRNA transcript (B) over time were quantified by qRT-PCR. (C and D) HCT116 cells were pre-treated with actinomycin D (ActD, 5 nM) or DMSO for 30 min followed by TNF-α (100 ng/ml) treatment for 0, 30 and 60 minutes. Levels of miR-346 (C) and VDR transcript (D) were quantified by qRT-PCR. Each experiment was repeated three times with similar results. Data from a single representative experiment are shown. All values are relative to corresponding 0 values.

These results prompted us to confirm that miR-346 can indeed directly target VDR. Figure 4A shows the target sequences of miR-346 in the 3′UTR of human and mouse VDR cDNAs. These sequences are highly conserved between human and mouse and have a high degree of complementarity with human and mouse miR-346s, respectively, particularly within the seed sequence at the 3′ end (Fig. 4A). Transfection of HCT116 cells with hsa-miR-346 oligo mimic substantially reduced hVDR (by >50%), and the potency was comparable to that of hVDR-specific siRNA (si-hVDR) that served as a positive control (Fig. 4B and C). To address whether miR-346 can suppress the transactivating activity of VDR, we co-transfected HCT116 cells with p3xVDRE-Luc reporter plasmid and control miRNA oligo mimic, hsa-miR-346 oligo mimic or hVDR-specific siRNA, followed by 1,25(OH)₂D₃ treatment. While 1,25(OH)₂D₃ stimulated luciferase activity in cells transfected with control miRNA, 1,25(OH)₂D₃ induction of luciferase activity was markedly attenuated in cells transfected with miR-346 or hVDR-siRNA (Fig. 4D), indicating that miR-346 indeed inhibited VDR’s transactivation activity.

Figure 4.

miR-346 inhibits VDR via targeting a 3′UTR site. (A) Alignment of human and mouse miR-346 sequences and corresponding miR-346 target site in the 3′UTR of human and mouse VDR transcripts. (B and C) HCT116 cells were un-transfected (Ctrl or C) or transfected with control miRNA (C-miR), hVDR-specific siRNA (si-hVDR) or miR-346 oligo mimic. Total cell lysates were prepared. VDR protein levels were analyzed by Western blotting (B) and quantified by densitometry (C). ** P<0.01 vs. Ctrl; (D) HCT116 cells were co-transfected with p3xVDRE-Luc and control miRNA (C), hVDR-specific siRNA (si-hVDR) or miR-346 oligo mimic. After 24 hours, the cells were stimulated with ethanol (EtOH) or 1,25(OH)₂D₃ (20 nM), followed by luciferase activity assay. * P<0.05, *** P<0.001 vs. corresponding EtOH. (E) Diagrammatic illustration of hVDR transcript including the protein-coding region (161-1444, black box) and the whole 3′-untranslated region (3′UTR, 1444-4669). Also shown is the construction of pGL3-Luc-3′UTR plasmid. (F) HCT116 cells were co-transfected with pGL3-Luc or pGL3-Luc-3′UTR and control miRNA (C) or miR-346 oligo mimic, followed by luciferase activity assay. ** P<0.01 vs. pGL3-Luc. (G) HCT116 cells were co-transfected with pGL3-Luc-3′UTR (3′UTR) or pGL3-Luc-3′UTR-Mut (3′UTR-Mut) and pCDH or pCDH-miR-346 as indicated. The pCDH-miR-346 plasmid expresses hsa-miR-346. After 24 hours, luciferase activity was quantified in the cell lysates. **P<0.01 vs. corresponding pCDH.

To confirm that miR-346 inhibits VDR via the target site in the 3′UTR, we constructed a luciferase reporter plasmid in which a 3′UTR fragment of hVDR cDNA (1444-3062) that contains the miR-346 target site was cloned to the 3′ downstream of the luciferase gene in pGL3-Luc Control plasmid (Fig. 4E). The resultant pGL3-Luc-3′UTR reporter displayed reduced luciferase activity when co-transfected with hsa-miR-346 mimic in HCT116 cells, in contrast to pGL3-Luc Control plasmid which was not affected by miR-345 mimic (Fig. 4F). Moreover, when cells were co-transfected with pCDH-miR-346 and pGL3-Luc-3′UTR-Mut, a mutant plasmid reporter that carries mutations in the seed sequence, the mutant reporter’s activity was no longer affected by ectopically expressed miR-346, in contrast to pGL3-Luc-3′UTR, whose activity was reduced by pCDH-miR-346 plasmid (Fig. 4G). Together these data demonstrate that VDR is a direct target of miR-346, and TNF-α suppresses VDR by inducing miR-346.

We then assessed the relevance of this regulatory relationship among colonic inflammation, miR-346 and epithelial VDR in IBD patients. As shown by immunostaining, massive infiltration of CD4⁺ and CD11b⁺ cells were detected in the lamina propria of active UC lesion biopsies (Fig. 5A). Similarly, the number of lamina propria TNF-α-producing cells was also dramatically increased in the active colitis lesion (Fig. 5B). Consistently, real time PCR quantitation showed that TNF-α mRNA transcript increased by >8 fold (IL-1β increased by >2 fold) (Fig. 5C), whereas VDR and ZO-1 expression decreased by >50% and >80%, respectively (Fig. 5D) in these lesion biopsies relative to adjacent normal tissues. Indeed, immunostaining data confirmed that the reduction of VDR in the lesion mostly occurred in intestinal epithelial cells (Fig. 5E), and Western blot analysis confirmed that the reduction in VDR protein levels was >50% in the lesions (Fig. 4F and G). Importantly, miR-346 was induced by >2.5 fold in the lesion (Fig. 5H). Therefore, colonic inflammation and TNF-α are inversely correlated with VDR and positively correlated with miR-346 in IBD patients, which is consistent with the notion that TNF-α down-regulates VDR via inducing miR-346.

Figure 5.

Immune cell infiltration and TNF-α expression are correlated with miR-346 induction and VDR suppression in human IBD biopsies. (A) Immunostaining of active UC and adjacent normal colon biopsies with anti-CD4 or anti-CD11b antibodies. Arrows indicated examples of CD4- or CD11b-positive cells. (B) Immunostaining of active UC and adjacent normal colon biopsies with anti-TNF-α antibodies. Arrows indicated examples TNF-α-producing cells. (C) qRT-PCR quantitation of TNF-α and IL-1β transcripts in active UC lesion and adjacent normal colon biopsies. Data are expressed as ratio of lesion to normal tissue. (D) qRT-PCR quantitation of VDR and ZO-1 transcripts in active UC and adjacent normal colon biopsies. Data are expressed as ratio of lesion to normal tissue. (E) Immunostaining of active UC and adjacent normal colon biopsies with anti-VDR antibodies. Arrows indicated epithelial cells. (F) Western blot analysis of VDR protein levels in active UC lesion (Lesion) and adjacent normal colon biopsies (Normal). Each lane represents one individual patient. (G) Quantitative data of the Western blot images. ** P<0.01 vs. Normal. (H) qRT-PCR quantitation of miR-346 in active UC lesion and adjacent normal colon biopsies. * P<0.05 vs. Normal.

We further validated our conclusion using experimental colitis models. IL-10-null mice spontaneously develop chronic intestinal inflammation and colitis due to aberrant immune response ^{7; 8}. As shown in Figure 6, colonic mucosal VDR levels were dramatically suppressed in IL-10-null mice compared to wild-type mice at 4 months of age (Fig. 6A), when the IL-10-null mice had developed severe mucosal inflammation. In TNBS-induced colitis model, severe mucosal inflammation develops within 3 days ⁶. We found that in TNBS-treated mice colonic mucosal VDR protein levels gradually decreased over time in the first 3 days after TNBS treatment (Fig. 6B), which was accompanied by a dramatic, time-dependent induction of TNF-α transcript (Fig. 6C) and miR-346 (Fig. 6D) in the mucosa during this period. Again, these observations are consistent with our conclusion that TNF-α suppresses VDR expression through induction of miR-346.

Figure 6.

Colonic VDR expression is inversely correlated with mucosal TNF-α and miR-346 levels in experimental colitis mouse models. (A) Western blot analysis of colonic mucosal VDR protein in wild-type (WT) and IL-10-null (IL-10KO) mice. (B) Changes of colonic mucosal VDR protein levels in TNBS-treated mice from day 0 (control) to day 3. Total mucosal lysates were used for Western blotting. (C) Changes of colonic mucosal TNF-α transcript levels from day 0 to day 3 in the TNBS mouse colitis model. (D) Changes of colonic mucosal miR-346 levels from day 0 to day 3 in the TBNS mouse colitis model. *** P<0.001 vs. Ctrl. All values are relative to corresponding control values.

Discussion

The VDR is abundantly expressed in the intestine. The classic action of the VDR in the small intestine is to regulate transcellular calcium transport, but its role in the large intestine is less clear. Our recent work demonstrated that colonic epithelial VDR signaling plays a key role in maintaining mucosal barrier integrity and protects against mucosal inflammation ⁶. One of the major actions of epithelial VDR is to suppress inflammation-induced apoptosis of IECs. Consistently with this role, we observed that the colonic VDR level is substantially reduced in patients with IBD ⁶. It is conceivable that epithelial VDR reduction increases the risk of mucosal barrier dysfunction and promotes mucosal inflammation. Thus it is important to understand how mucosal VDR is down-regulated in the context of mucosal inflammation and colitis.

In this report we demonstrated that inflammation plays a key role in the down-regulation of the VDR in colonic epithelial cells. Pro-inflammatory cytokine TNF-α suppresses VDR expression, and the time course of VDR down-regulation is inversely correlated with that of miR-346 up-regulation. We validated that the VDR transcript contains a conserved 3′UTR sequence that is targeted by miR-346 to inhibit VDR expression. Moreover, we showed that colonic VDR down-regulation is inversely correlated with TNF-α and miR-346 induction and lamina propria immune cell infiltration in human IBD biopsies. We also observed that, in experimental models of colitis, colonic VDR expression is gradually reduced with the progression of mucosal inflammation towards the development of colitis, which is closely associated with the induction of TNF-α and miR-346 in this process. Together the in vitro and in vivo evidence strongly supports the notion that TNF-α suppresses epithelial VDR via induction of miR-346.

Pro-inflammatory cytokines play key roles in promoting mucosal inflammation. Among a number of inflammatory cytokines commonly detectable in IBD, our study identified TNF-α as a key cytokine that negatively regulates epithelial VDR expression. Although it is possible that other cytokines not yet tested may also be involved in the down-regulation of VDR, TNF-α as a inflammatory mediator to inhibit VDR makes good sense, because TNF-α plays a critical role in the development of colitis and anti-TNF therapy is a standard therapy in the management of IBD ^{12; 13}. It is known that TNF-α induces mucosal barrier injury by promoting IEC apoptosis and disrupting the intercellular junctions between epithelial cells. PUMA, a pro-apoptotic BCL-2 member, and claudin-2, a tight junction protein, are examples of TNF-α targets in these processes ^17,18. Therefore it is not surprising that VDR is another target for TNF-α to damage the mucosal barrier. Although both miR-125b and miR-346 are stimulated by TNF-α, we showed that only miR-346 is able to inhibit VDR. Although miR-346 is unlikely able to block VDR nuclear translocation, inhibition of VDR protein levels ought to be sufficient to reduce VDR’s activity. Here we were not able to confirm the VDR-inhibiting activity of miR-125b reported previously ¹⁶. In fact, no reduction of VDR protein by miR-125b was reported in this prior work. It appears that miR-125b is not an effective inhibitor of VDR. How TNF-α induces miR-346 is an interesting question that remains to be determined in future studies.

Our previous studies have shown that intestinal VDR signaling inhibits inflammation-induced apoptosis of intestinal epithelial cells via blocking NF-κB activity ⁶, but how mucosal inflammation affects epithelial VDR is unknown. That TNF-α suppresses VDR via miR-346 is thus a novel finding. Given the critical role of VDR in maintaining the integrity of mucosal epithelial barrier, this finding offers a mechanistic insight into how mucosal inflammation induces mucosal barrier damage. Mucosal inflammation down-regulates VDR, which impairs the mucosal epithelial barrier. As a result, VDR reduction in the epithelial cells in turn promotes mucosal inflammation. This vicious cycle is thought to contribute to the progression of mucosal inflammation and promote the development of colitis. It is well known that the vitamin D hormone can induce VDR expression ¹⁹ and, through VDR, suppresses TNF-α production ²⁰. Thus VDR and TNF-α appear to be mutually regulated by the other. Vitamin D therapy raises vitamin D hormone that induces and activates VDR, which shifts the balance to favor inhibition of TNF-α and blockade of IEC apoptosis. This is a mechanism by which vitamin D therapy suppresses mucosal inflammation and ameliorates IBD ^{21; 22}. On the other hand, our data infer that anti-TNF therapy should alleviate the suppression of VDR by TNF-α and restore epithelial VDR levels to protect the barrier. Certainly this prediction requires further confirmation. Interestingly, a recent study reported that miR-346 controls TNF-α secretion in activated macrophages ²³, suggesting a feedback loop existing between TNF-α and miR-346. As the TNF-α-miR-346-VDR axis is not restricted in epithelial cells, the interplays among TNF-α, VDR and miR-346 under various pathophysiological conditions in colonic epithelial cells as well as in other cell types and the resultant biological effects warrant further investigation.