Microbiota-Dependent Induction of Colonic Cyp27b1 Is Associated With Colonic Inflammation: Implications of Locally Produced 1,25-Dihydroxyvitamin D₃ in Inflammatory Regulation in the Colon

Abstract

Our recent studies demonstrated that intestinal epithelial vitamin D receptor (VDR) signaling plays a critical role in regulating colonic inflammation by protecting epithelial barrier integrity. Epithelial VDR is downregulated in colitis, but how mucosal inflammation affects local 1,25-dihydroxyvitamin D₃ [1,25(OH)₂D₃] production is unknown. Here we showed that cytochrome P450 27b1 (Cyp27b1), a cytochrome P450 enzyme necessary for 1,25(OH)₂D₃ biosynthesis, is highly induced in colonic mucosa in inflammatory response. Although VDR is reduced in colon biopsies from patients with ulcerative colitis, Cyp27b1 is markedly upregulated in these samples. Colon mucosal Cyp27b1 was also markedly induced in an experimental colitis mouse model, and this local Cyp27b1 induction and colonic inflammation required the presence of commensal bacteria. Vitamin D deficiency further exaggerated colonic Cyp27b1 induction and aggravated colonic inflammation in mice. In HCT116 cells, lipopolysaccharide or tumor necrosis factor-α treatment induced Cyp27b1 in time- and dose-dependent manners, and the induced Cyp27b1 was enzymatically active. The inflammation-induced upregulation of Cyp27b1 was mediated by nuclear factor κB. Collectively these data suggest that induction of colonic epithelial Cyp27b1, which is expected to increase local production of 1,25(OH)₂D₃, is a protective mechanism that partially compensates for the downregulation of epithelial VDR during colonic inflammation. Increased local 1,25(OH)₂D₃ maintains 1,25(OH)₂D₃-VDR signaling to protect the mucosal barrier and reduce colonic inflammation.

Colonic Cyp27b1 is highly induced in colonic inflammation, suggesting a protective role of locally produced vitamin D hormone in controlling colonic inflammation.

Cyp27b1 is the cytochrome P450 enzyme 25-hydroxyvitamin D 1α-hydroxylase that regulates the biosynthesis of 1,25-dihydroxyvitamin D₃ [1,25(OH)₂D₃], the active hormonal metabolite of vitamin D (1). The biological activity of 1,25(OH)₂D₃ is mediated by the vitamin D receptor (VDR), a nuclear hormone receptor (2). The majority of the body’s vitamin D content is derived from photosynthesis in the skin that is catalyzed by ultraviolet light irradiation (3). Vitamin D is converted to 1,25(OH)₂D₃ after two steps of hydroxylation: 25-hydroxylation catalyzed by hepatic 25-hydroxylase and 1α-hydroxylation catalyzed by cytochrome P450 27b1 (Cyp27b1) (1). Although Cyp27b1 expression is detected in multiple tissues in the body, it is believed that renal Cyp27b1 in the proximal tubular cells is the predominant contributor to systemically circulating 1,25(OH)₂D₃, whereas Cyp27b1 in other tissues, such as the skin, placenta, prostate, breast, macrophages, pancreas, vasculature, and colon, controls only the local concentration of 1,25(OH)₂D₃ that exerts intracrine or paracrine actions (4). It is well known that renal Cyp27b1, hence the circulating 1,25(OH)₂D₃ level, is stimulated by parathyroid hormone and suppressed by FGF23, whereas extrarenal Cyp27b1 is influenced by other factors such as proinflammatory cytokine interferon-γ (IFN-γ) (5).

The small and large intestines have the most abundant VDR expression in the body. It is well established that the main physiological function of VDR in the gastrointestinal tract is to regulate calcium absorption in the duodenum, which is crucial for the skeletal health (6). Compared with the VDR, the expression level of Cyp27b1 in the gut is very low; thus the calcium transport activity of the VDR is believed to be activated mainly by systemic 1,25(OH)₂D₃. Until recently the role of the VDR in the colon is unclear. Recent studies, including ours, have demonstrated that colonic epithelial VDR plays a key role in colon homeostasis (7–10). Colonic epithelial VDR controls colonic inflammation by protecting the integrity of the mucosal barrier, which separates harmful luminal substances such as microorganisms, toxins, and antigens from the body. We have shown that epithelial VDR not only inhibits excessive epithelial cell apoptosis by blocking PUMA–mediated proapoptotic pathway but also prevents tight junction dysfunction in colonic inflammation by suppressing the activation of the myosin light chain kinase (MLCK)–myosin light chain (MLC) pathway (8, 9), which regulates tight junction permeability (11). Colonic VDR levels are downregulated substantially during inflammation in mice and humans (8, 9), which is consistent with the barrier-protective role of VDR in the gut.

Despite the remarkable advance in our understanding of the colonic VDR in recent years, little is known about the role of the colonic Cyp27b1 and locally produced 1,25(OH)₂D₃ within the colon. Interestingly, a recent study reported that steady-state Cyp27b1 gene transcription in intestinal epithelial cells requires the presence of microbiota, suggesting that the commensal bacteria promote the production of local 1,25(OH)₂D₃, probably for the purpose of maintaining intestinal homeostasis (12). This observation prompted us to hypothesize that the synthesis of 1,25(OH)₂D₃ catalyzed by local Cyp27b1 plays a role in the regulation of colonic inflammation. Here we tested this hypothesis in patients with inflammatory bowel disease (IBD) and a mouse model of experimental colitis. We showed that colonic Cyp27b1 is dramatically induced in human patients and mice in response to colonic inflammation.

Experimental Procedures

Human biopsies

Colonic mucosal biopsies were obtained from patients with active ulcerative colitis at Shengjing Hospital of China Medical University. Biopsies were collected from the inflamed region and the adjacent normal tissue in each patient during endoscopic examination. Study subjects were recruited, with written informed consent from the participants or their guardians. All human studies were approved by the Institutional Ethical Committee of Shengjing Hospital, China Medical University (Protocol 2014PS145K). The collected biopsies were immediately frozen with liquid nitrogen for later RNA and protein lysate preparations or placed in 4% formaldehyde in phosphate-buffered saline (pH 7.2) for later histological analyses.

Animal studies

C57BL/6 mice (6 to 8 weeks old) were induced to colitis with 2,4,6-trinitrobenzene sulfonic acid (TNBS), as previously reported (8, 13). Mouse body weight was monitored and disease score assessed daily. Mice were usually euthanized on day 3 after TNBS treatment. Colons were collected immediately after euthanasia and fixed in 10% formalin for histology. Mucosa was scraped to isolate total RNAs or proteins. Colonic epithelial cells were purified according to a protocol previously published (14). To assess the effect of commensal bacteria on Cyp27b1 induction and mucosal inflammation, an antibiotic-induced microbiota-depleted (AIMD) model was established as described previously (12, 15). Mice were treated with a cocktail of antibiotics in drinking water (ampicillin 1 g/L, vancomycin 500 mg/L, neomycin sulfate 1 g/L, and metronidazole 1 g/L) for 4 weeks. Mice were subjected to TNBS treatment at the end of the third week. In a group of mice, lipopolysaccharide (LPS; 10 mg/mL) was added to the antibiotic cocktail at the end of the third week before TNBS treatment. Controls were treated with regular water. To study the effect of vitamin D deficiency, 3-week-old male C57BL/6 mice were placed on a vitamin D–deficient (D-VitD; <25 IU/kg vitamin D₃) or vitamin D–sufficient (S-VitD; 1000 IU/kg vitamin D₃) diet (Xietong, Jiangsu, China) for 8 weeks. These were AIN93G-based diets that contained 3850 kcal/kg, with protein 20%, carbohydrate 64%, and fat 16% kcal. To avoid direct ultraviolet light exposure, these mice were housed in a dark room equipped with a red light. Experimental colitis was then induced in these mice with TNBS at the end of week 8. Both male and female mice were used in the study. These animal studies were approved by the Institutional Ethical Committee of China Medical University.

Measurement of 25-hydroxyvitamin D

Mouse serum 25-hydroxyvitamin D concentrations were measured with an EIA kit (Eagle Biosciences, Nashua, NH) according to the manufacturer’s instructions. Colorimetric reading was performed using a Tecan M200 microplate reader system (Tecan Group, Mannedorf, Switzerland).

Histology and immunostaining

Freshly dissected colon or colon biopsies were fixed overnight with 4% formaldehyde in phosphate-buffered saline (pH 7.2) and embedded in paraffin wax. Mouse colons were embedded as “Swiss roll” (16). Colonic morphology was examined by hematoxylin and eosin staining, and histological scores were graded according to a previously published system (17, 18). For immunostaining, slides were incubated with anti–tumor necrosis factor-α (TNF-α) or anti-Cyp27b1 antibodies (Santa Cruz Biotechnology, Dallas, TX), followed by fluorescein isothiocyanate–conjugated or horseradish peroxidase–conjugated secondary antibodies. Slides were examined under a Leica DFC425 fluorescence microscope.

Cell culture

HCT116 cells were cultured in Dulbecco’s modified Eagle medium supplemented with 10% fetal bovine serum at 37°C and 5% CO₂. The cells were treated with different doses of LPS or TNF-α (0 to 100 ng/mL) for 24 hours, or treated with LPS or TNF-α (100 ng/mL) for different times (0 to 48 hours) before VDR, TLR4, Cyp24a1, Cyp27b1, and CD14 expression were analyzed. To assess the activity of induced Cyp27b1, the cells were treated with LPS in the presence or absence of 1,25(OH)₂D₃ (20 nM) or 25(OH)D (50 nM) for 24 hours before we assessed TNF-α induction by reverse transcription polymerase chain reaction (RT-PCR). In a similar experiment, the cells were transfected with a scramble or Cyp27b1-specific small interfering RNA (siRNA) (5′TAGGCACAAGACCAAGGTATA3′) to silence Cyp27b1, as described previously (19), before exposure to 1,25(OH)₂D₃ or 25(OH)D. In separate experiments, the cells were pretreated with nuclear factor κB (NF-κB) inhibitor BAY 11-7082 (20 ng/mL) or AP-1 inhibitor T-5224 (20 ng/mL, MedChemExpress, NJ) for 24 hours, followed by 12-hour TNF-α or LPS treatment before assessment of Cyp27b1 expression by Western blotting. The cells were also transfected with a plasmid-expressing inhibitor of nuclear factor–κB subunit β (IKKβ), followed by BAY 11-7082 treatment before measuring Cyp27b1 protein levels. In another experiment, cells were pretreated with anti–TNF-α neutralizing antibodies (4 µg/mL; R&D Systems, Minneapolis, MN) for 12 hours before being stimulated with LPS (100 ng/mL) for 24 hours.

Western blot

Western blot analyses were carried out as described previously (20). The following antibodies were used: anti-phospho-MLC from Cell Signaling Technology (Beverly, MA); anti-long MLCK from Abcam (Cambridge, MA); and anti-β-actin, anti-VDR, anti–TNF-α, anti-Cyp27b1, and anti-TLR4 from Santa Cruz Biotechnology (Dallas, TX).

RT-PCR

Total RNAs were extracted with TRIzol reagent (Invitrogen). First-strand complementary DNA templates were synthesized with a PrimeScript RT reagent kit (TaKaRa, Mountain View, CA). Real-time polymerase chain reaction (PCR) was carried out with an SYBR Premix Ex kit (TaKaRa) in a Bio-RAD IQ5 real-time system. Relative amount of transcripts was calculated by the 2^−ΔΔCt formula, with glyceraldehyde 3-phosphate dehydrogenase (GAPDH) used as an internal control. PCR primers are listed in Table 1.

Table 1.

Nucleotide Sequences of PCR Primers

Primer Name	Forward (5′-3′)	Reverse (5′-3′)
mVDR	GATGCCCACCACAAGACCTA	CGGTTCCATCATGTCCAGTG
mCYP27B1	TCAGCAGGCATCGCAGAAC	GCATTGGATCCTGAGGAATGA
mCYP24A1	GGTTATCTCCGGGGTGGAGT	AGTGGCCAATGAGCACGC
mTLR4	ACCAATGCATGGATCAGAAA	GTCTCCACAGCCACCAGATT
mTNFα	TCAGCCTCTTCTCATTCCTG	CAGGCTTGTCACTCGAATTT
mIFNγ	GCGTCATTGAATCACACCTG	TGAGCTCATTGAATGCTTGG
mIL-1β	CCAAAAGATGAAGGGCTGCT	ACAGAGGATGGGCTCTTCTT
mIL-6	CCTCTCTGCAAGAGACTTCCA	AGAATTGCCATTGCACAACTCT
mIL-17	TCCCTCTGTGATCTGGGAAG	AGCATCTTCTCGACCCTGAA
mIL-23p19	AATAATGTGCCCCGTATCCA	CATGGGGCTATCAGGGAGTA
mMCP-1	CAAGAAGGAATGGGTCCAGA	TGAGGTGGTTGTGGAAAAGG
mCD14	GAAGCAGATCTGGGGCAGTT	CGCAGGGCTCCGAATAGAAT
mGAPDH	TGTGTCCGTCGTGGATCTGA	CCTGCTTCACCACCTTCTTGA
hVDR	GACTTTGACCGGAACGTGCCC	CATCATGCCGATGTCCACACA
hCYP27B1	CAGACAAAGACATTCATGTGGG	GTTGATGCTCCTTTCAGGTAC
hCYP24A1	CAAACCGTGGAAGGCCTATC	AGTCTTCCCCTTCCAGGATCA
hTLR4	AGAACTGCAGGTGCTGGATT	AAACTCTGGATGGGGTTTCC
hCD14	CGCTCCGAGATGCATGTG	AACGACAGATTGAGGGAGTTCAG
hGAPDH	ACCACAGTCCATGCCATCAC	TCCACCACCCTGTTGCTGTA

Statistical analysis

Data values were presented as means ± standard deviations. Statistical analyses were carried out via unpaired two-tailed Student t test for two data comparisons or via one-way analysis of variance with a Student-–Newman–Keuls post hoc test for multiple data comparisons. Correlation analyses were performed via two-tailed Pearson correlation test with 95% confidence intervals. A P ≤ 0.05 is considered statistically significant.

Results

Colon mucosal Cyp27b1 is markedly upregulated in patients with IBD

Because we previously found that colonic epithelial VDR was downregulated in biopsies from patients with IBD (8), we further assessed the expression of colonic Cyp27b1 in the patients. For this study we collected colon mucosal biopsies from the inflamed lesions and adjacent normal tissues from a new cohort of patients with active ulcerative colitis. Western blot analyses showed that whereas VDR expression was decreased by ~50% in the inflamed lesions, as we reported previously (8), TNF-α and Cyp27b1 were markedly induced in the lesions compared with the adjacent normal tissues in these patients (Fig. 1A and 1B). The increase of Cyp27b1 was unexpected. Immunostaining showed that a large number of infiltrated leukocytes in the lamina propria were strongly positive for TNF-α production in the inflamed lesion (Fig. 1C), suggesting that the increase in TNF-α protein was generated mostly from leukocytes. Crypt epithelial cells, on the other hand, were strongly positive for Cyp27b1 staining in the lesion (Fig. 1D), suggesting that the induction of colonic Cyp27b1 was contributed mostly by the epithelial cells.

Figure 1.

Cyp27b1 is markedly upregulated in colonic biopsies from patients with ulcerative colitis. (A) Western blot analyses of biopsies of inflamed lesion (L) and adjacent normal tissues (N) from seven patients with ulcerative colitis using the antibodies indicated. (B) Relative protein levels in the lesions compared with the normal tissues, determined by densitometric quantitation of protein bands on the blots. *P < 0.05, ***P < 0.001 vs normal; n = 7. (C) Immunostaining with anti–TNF-α antibody. Note TNF-α–positive cells in the lamina propria of the lesion sections. Magnification ×100. (D) Immunostaining with anti-Cyp27b1 antibody. Arrows indicate examples of Cyp27b1-negative (left panel) and Cyp27b1-positive (right panel) epithelial cells. Magnification ×400.

Cyp27b1 is dramatically induced in colon mucosa in experimental colitis model

To confirm the induction of Cyp27b1 in the inflamed colon, we examined the mouse model of TNBS-induced colitis. As expected, 3 days after TNBS instillation, mucosal VDR was markedly suppressed in the distal colon; in contrast, Cyp27b1, which was barely detectable in untreated controls, was dramatically induced in the same region after TNBS instillation (Fig. 2A and 2B). We previously reported that a treatment with paricalcitol, an active vitamin D analog, was able to prevent the reduction of mucosal VDR and attenuate colonic inflammation in the TNBS colitis model (9). Reanalyses of the colon mucosal samples from that study via real-time RT-PCR showed that paricalcitol treatment not only prevented the reduction of Vdr transcript but also blocked the induction of Cyp27b1 and Tlr4 transcripts caused by TNBS treatment (Fig. 2C). As expected from previous studies (21, 22), paricalcitol treatment induced Cyp24a1 and Cd14 in the colon in normal mice, but these inductions were attenuated in TNBS-treated mice (Fig. 2C). These observations suggest that, like VDR reduction, Cyp27b1 induction is a consequence of colonic inflammation.

Figure 2.

Induction of colon Cyp27b1 in TNBS colitis model. (A) Western blot analyses of colonic mucosal VDR and Cyp27b1 in control and TNBS-treated mice on day 3. (B) Densitometric semiquantitation of the Western blot data. *P < 0.05, ***P < 0.001 vs corresponding control; n = 4–5. (C) Real-time RT-PCR quantitation of Vdr, Cyp27b1, Cyp24a1, Tlr4, and Cd14 transcripts in colonic mucosa from four groups of mice: control, TNBS-treated, paricalcitol-treated, TNBS + paricalcitol (Pari). *P < 0.05; ***P < 0.001 vs the rest; ^#P < 0.05 vs control and paricalcitol or TNBS. n = 5 in each group.

Microbiota are necessary for colonic Cyp27b1 induction and colitis development

Because microbiota play an important role in colonic inflammation, we studied the relationship between microbiota and colonic Cyp27b1 upregulation, using mice depleted of commensals by antibiotic treatment. As shown in Fig. 3, compared with control mice with normal microflora, the AIMD mice exhibited less body weight loss after TNBS instillation, whereas after 1 week of LPS supplementation the weight loss of AIMD mice was comparable to that of the TNBS group without antibiotic treatment (Fig. 3A). Microbiota depletion led to significant reduction in clinical scores and histological damage of the colon and marked attenuation of colitis development, but the severity of these colonic phenotypes was restored after 1 week of LPS supplementation (Fig. 3B–3D). These observations indicate that microbiota, specifically bacterial endotoxins, are necessary for the development of colitis in the TNBS model. As expected, microbiota depletion eliminated the induction of proinflammatory cytokines [TNF-α, interleukin (IL)-1β, IL-6, and IFN-γ] in the colonic mucosa after TNBS instillation, whereas LPS was able to restore these proinflammatory inductions (Fig. 3E). Importantly, microbiota depletion completely eliminated mucosal VDR suppression and Cyp27b1 upregulation at both messenger RNA (mRNA) and protein levels after TNBS instillation, and these changes were completely reversed after LPS supplementation (Fig. 3E and 3F). The expression pattern of MLCK, phospho-MLC, and PUMA in the mucosa reflected the expected changes in tight junction permeability and apoptotic status in the mucosal barrier under these conditions (Fig. 3F).

Figure 3.

Microbiota are necessary for the induction of colon Cyp27b1 in colonic inflammation. (A) Body weight changes in five groups of mice from 0 to 3 days after TNBS instillation: control (Ctrl), AIMD, TNBS, AIMD + TNBS, and AIMD + LPS + TNBS. *P < 0.05, ** P < 0.01, *** P < 0.001 vs Ctrl; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 vs AIMD + TNBS; n = 5 in each group. (B) Clinical score and (C) histological score of the colonic sections, from TNBS, AIMD + TNBS, and AIMD + LPS + TNBS mice; *P < 0.05, **P < 0.01 vs the rest. (D) Representative hematoxylin and eosin sections of distal colons from the indicated groups of mice. Magnification, ×400. (E) Real-time RT-PCR quantitation of Vdr, Cyp27b1, Cyp24a1, Tnfa, Il1b, Il6, and Ifng transcripts in colonic mucosa from these five groups of mice. *P < 0.05, **P < 0.01, ***P < 0.001 vs Ctrl; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 vs AIMD + TNBS. (F) Western blot analysis of colonic mucosal VDR, Cyp27b1, MLCK, p-MLC, and PUMA in the indicated groups of mice.

Vitamin D deficiency aggravates colonic inflammation and exacerbates Cyp27b1 induction

Vitamin D deficiency is common in people with IBD and is thought to be an environmental factor for IBD. We therefore assessed the effect of vitamin D deficiency on colitis development and Cyp27b1 expression in the TNBS model. Vitamin D deficiency was induced by placing newly weaned mice on a vitamin D–deficient diet for 2 months, which was confirmed by low serum 25(OH)D concentrations (~20 ng/mL) in these mice (Fig. 4A). Compared with S-VitD mice, D-VitD mice showed more severe body weight loss (Fig. 4B) and developed more severe clinical scores and histological damage in the colon after TNBS instillation (Fig. 4C–4E). Real-time RT-PCR quantitation revealed that inductions of mucosal proinflammatory cytokines (TNF-α, interferon-γ, IL-1β, IL-6, IL-17, IL-23, and MCP-1) were more robust in D-VitD mice than in S-VitD mice (Fig. 4F). These observations confirmed that vitamin D deficiency promotes colonic inflammation and colitis development. Western blot analyses of mucosal lysates showed that VDR was suppressed, whereas Cyp27b1, MLCK, p-MLC, and PUMA were induced after TNBS instillation as expected; however, these changes were much more robust in vitamin D deficiency (Fig. 4G). These data are consistent with the notion that colonic inflammation simultaneously suppresses VDR and upregulates Cyp27b1 in the mucosa.

Figure 4.

Vitamin D deficiency exaggerates colonic Cyp27b1 induction in TNBS model. (A) Serum 25-hydroxyvitamin D₃ levels in four groups of mice: S-VitD, D-VitD, S-VitD + TNBS, D-VitD + TNBS; **P < 0.01 vs S-VitD control (Ctrl). (B) Body weight changes after TNBS instillation in the indicated groups of mice. *P < 0.05, **P < 0.01, ***P < 0.001 vs S-VitD Ctrl. (C) Clinical score from S-VitD + TNBS, D-VitD + TNBS mice and (D) histological score of the colonic sections from S-VitD + TNBS, D-VitD + TNBS mice; *P < 0.05. n = 5 in each group. (E) Representative hematoxylin and eosin sections of distal colons from these four groups of mice. Magnification, ×100. (F) Real-time RT-PCR quantitation of Tnfa, Ifng, Il1b, Il6, Il17, Il12, and Mcp1 transcripts in colonic mucosa from the indicated groups of mice. **P < 0.01, ***P < 0.001 vs S-VitD Ctrl. (G) Western blot analysis of colonic mucosal VDR, Cyp27b1, MLCK, p-MLC, and PUMA in these four groups of mice.

Enzymatically active Cyp27b1 is induced by LPS or TNF-α in colonic epithelial cells

To confirm that Cyp27b1 is induced by inflammation, we studied HCT116 cells, a human colonic cancer epithelial cell line. We first assessed whether LPS regulated Cyp27b1 in these cells. As shown in Fig. 5, LPS suppressed Vdr and Cyp24a1 transcripts but markedly stimulated Cyp27b1, Tlr4, and Cd14 transcripts in time- and dose-dependent manners (Fig. 5A and 5B). Cyp24a1 is an enzyme that initiates the catabolic pathway to inactivate 1,25(OH)₂D₃ (1). Therefore, LPS treatment strongly favored 1,25(OH)₂D₃ synthesis. TLR4 and CD14 are the receptor and coreceptor for LPS, and both are known to be upregulated by LPS (23), consistent with our finding. At the protein level, VDR was highly expressed in HCT116 cells, but TLR4 and Cyp27b1 proteins were barely detectable at baseline. Consistently with the mRNA data, LPS inhibited VDR but upregulated TLR4 and Cyp27b1 in a time-dependent manner (Fig. 5C and 5D). Thus, consistent with the in vivo observation, the changes of VDR and Cyp27b1 in cultured colonic epithelial cells were in opposite directions under LPS treatment.

Figure 5.

LPS induces Cyp27b1 in HCT116 cells. (A) Time course of LPS effects on Vdr, Cyp27b1, Cyp24a1, Tlr4, and Cd14 transcripts in HCT116 cells. (B) Dose-dependent effects of LPS on Vdr, Cyp27b1, Cyp24a1, Tlr4, and Cd14 transcripts in HCT116 cells. *P < 0.05, **P < 0.01 vs corresponding 0 hour or 0 ng/mL. (C) Time-dependent changes in TLR4, VDR, and Cyp27b1 proteins in HCT116 after LPS treatment. (D) Time course curve of TLR4, VDR, and Cyp27b1 proteins in HCT116 after LPS treatment. Pearson correlation coefficient (r) and P values for each curve are presented. (E–G) Assessment of Cyp27b1 activity to convert 25(OH)D₃ to 1,25(OH)₂D₃ in HCT116 cells. HCT116 cells were treated with LPS in the presence of 1,25(OH)₂D₃ or 25(OH)D₃ for 16 hours, and (E) Tnfa transcript was measured by RT-PCR; or (F) after the cells were transfected with a scramble or Cyp27b1-specific siRNA to knock down Cyp27b1, (G) a similar experiment was performed.

To address whether the induced Cyp27b1 is enzymatically active, we assessed TNF-α expression, known to be stimulated by LPS and inhibited by 1,25(OH)₂D₃. As expected, LPS markedly stimulated Tnfa transcript and this stimulation was suppressed by 1,25(OH)₂D₃. Interestingly, Tnfa mRNA upregulation was also blocked in the presence of 25-hydroxyvitamin D (25(OH)D₃), an inactive metabolite, indicating that 25(OH)D₃ was converted to 1,25(OH)₂D₃ by Cyp27b1 to inhibit Tnfa transcription (Fig. 5E). Moreover, when Cyp27b1 was silenced using a Cyp27b1-specific siRNA (Fig. 5F), the ability of 25(OH)D₃ to suppress Tnfa expression was lost (Fig. 5G), confirming the specificity of Cyp27b1 in the conversion of 25(OH)D₃ to 1,25(OH)₂D₃. Together, these observations indicate that Cyp27b1 induced in colonic epithelial cells is enzymatically active.

We then asked whether TNF-α, a key proinflammatory cytokine involved in colitis development, regulates Cyp27b1 in colonic epithelial cells. Consistent with our previous observation (24), TNF-α markedly suppressed VDR in HCT116 cells in dose- and time-dependent manners. Simultaneously, TNF-α also markedly upregulated Cyp27b1 in these cells (Fig. 6A–6D).

Figure 6.

TNF-α induces Cyp27b1 in HCT116 cells. (A) Dose-dependent effects of TNF-α on VDR and Cyp27b1 proteins in HCT116 cells. (B) Relative VDR and Cyp27b1 protein levels plotted against TNF-α dose. (C) Time course of LPS effects on VDR and Cyp27b1 proteins in HCT116 cells. (D) Relative VDR and Cyp27b1 protein levels plotted against TNF-α treatment time. Pearson correlation coefficient (r) and P values for each curve are presented.

LPS or TNF-α induction of Cyp27b1 is NF-κB dependent

Previous studies identified one NF-κB and two AP-1 binding sites in the mouse Cyp27b1 gene promoter that are responsive to microbiota (12). Therefore, we assessed whether NF-κB and AP-1 were involved in the induction of Cyp27b1 by LPS or TNF-α in HCT116 cells. As shown in Fig. 7, LPS or TNF-α stimulation of Cyp27b1 was markedly attenuated in the presence of BAY 11-7082, a NF-κB inhibitor (Fig. 7A and 7B); in contrast, T-5224, an AP-1 inhibitor compound, had little effect on LPS- or TNF-α−induced Cyp27b1 upregulation (Fig. 7C and 7D). Furthermore, when the cells were transfected with an IKKβ-expressing plasmid, which is known to activate the NF-κB signaling pathway, Cyp27b1 was markedly induced, but this IKKβ induction of Cyp27b1 was blocked by BAY 11-7082 treatment (Fig. 7E). These results suggest that both LPS and TNF-α upregulate Cyp27b1 through activation of NF-κB. Because LPS is able to induce TNF-α production, we asked whether LPS stimulation of Cyp27b1 is mediated by TNF-α. Indeed, the induction of Cyp27b1 and MLCK, as well as the repression of VDR, that was initiated by LPS was almost completely abrogated in the presence of anti–TNF-α neutralizing antibody; however, the induction of TLR4 was not affected (Fig. 7F). These observations indicate that LPS acts through TNF-α to regulate VDR, Cyp27b1, and MLCK, but not TLR4, in HCT116 cells. In this case, TNF-α probably works in an autocrine or paracrine manner.

Figure 7.

Cyp27b1 induction by LPS or TNF-α is mediated by NF-κB. HCT116 cells were treated with (A, C) 100 ng/mL LPS or (B, D)100 ng/mL TNF-α overnight in the presence or absence of (A, B) 20 ng/mL BAY 11-7082 or (C, D) 20 ng/mL T-5224 as indicated. Cyp27b1 protein levels were measured by Western blotting. (E) HCT116 cells were transfected with an IKKβ-expressing plasmid, followed by BAY 11-7082 treatment for 24 hours before assessing Cyp27b1 protein levels. (F) HCT116 cells were treated with LPS (100 ng/mL) overnight in the presence or absence of neutralizing anti–TNF-α antibody (4 μg/mL). VDR, Cyp27b1, MLCK, and TLR4 proteins were assessed by Western blotting after these treatments.

Discussion

Cyp27b1 is an important enzyme in the vitamin D endocrine system because it catalyzes the biosynthesis of 1,25(OH)₂D₃, the active hormonal metabolite of vitamin D. Whereas renal Cyp27b1 regulates systemic 1,25(OH)₂D₃ levels, extrarenal Cyp27b1 usually controls local 1,25(OH)₂D₃ concentrations to perform local functions. In this regard, one good example is macrophage Cyp27b1, which is highly induced in mycobacterial infection via Toll-like receptor–mediated pathways, leading to antimicrobial peptide production (25). Although a few early studies reported the expression of Cyp27b1 in human colons (26, 27), very little is known about colonic Cyp27b1. In this study we showed that although local Cyp27b1 expression in the colon is low under normal conditions, it is highly induced during colonic inflammation both in human patients with IBD and in a mouse model of experimental colitis. Upregulation of enzymatically active Cyp27b1 was also seen in HCT116 cells when treated with LPS or TNF-α, confirming that epithelial Cyp27b1 induction is regulated by inflammation. On the other hand, colonic Cyp24a1, the enzyme that catalyzes inactivation of 1,25(OH)₂D₃, is suppressed in inflammation, and the underlying mechanism is unknown. Therefore, colonic inflammation results in a microenvironment that favors local 1,25(OH)₂D₃ production and maintenance. In fact, previous studies reported elevated circulating 1,25(OH)₂D₃ levels and colonic Cyp27b1 expression in a subset of patients with Crohn disease (28), which is consistent with our findings. The human biopsies used in our study were from patients with ulcerative colitis, and the TNBS model is thought to mimic Crohn disease. Therefore, it appears that colonic inflammation stimulates local 1,25(OH)₂D₃ synthesis regardless of Crohn disease or ulcerative colitis. One thing that is unclear, however, is whether the local increase in 1,25(OH)₂D₃ production in the colon contributes to the systemic 1,25(OH)₂D₃ levels.

Interestingly, the induction of colonic Cyp27b1 is inversely correlated with VDR levels. Given the ample evidence demonstrating that intestinal epithelial VDR signaling plays a critical role in controlling colonic inflammation (7–9), we postulated that colonic Cyp27b1 induction, which increases local 1,25(OH)₂D₃ production in the colon, is a protective response to partially compensate for the reduction of epithelial VDR so as to maintain vitamin D–VDR signaling during colonic inflammation. Indeed, increased 1,25(OH)₂D₃ ligand is able to promote VDR occupancy in the genome even in the face of low VDR levels. Epithelial Cyp27b1 induction in colonic inflammation is not surprising, because similar protective responses have been reported for other barrier-protecting proteins such as hypoxia-inducible factor. Intestinal epithelial hypoxia-inducible factor protects the mucosal barrier in experimental colitis and is also highly induced in the TNBS model (29).

Through serendipity by Western blot analysis we noticed the dramatic induction (>3.5-fold) of Cyp27b1 in the biopsies from a cohort of patients with IBD, together with the induction of TNF-α (~fourfold) and decrease (by ~50%) of VDR in these samples. TNF-α appeared to be produced mainly from the infiltrated leukocytes, whereas Cyp27b1 upregulation was seen mainly in the epithelial cells. In agreement with these human observations, we confirmed similar expression patterns of Cyp27b1 and VDR in mouse colon mucosa of the TNBS experimental colitis model. These observations prompted speculation that TNF-α secreted from lamina propria leukocytes drives Cyp27b1 and VDR expression in opposite directions in mucosal epithelial cells during inflammation. In fact, in previous work we reported a ~50% reduction in epithelial VDR in the lesions of patients with IBD (8, 24), and this downregulation is attributed largely to a TNF-α-induced increase of miR-346, which blocks VDR translation (24). In the current study we demonstrated that TNF-α upregulates epithelial Cyp27b1 by an NF-κB dependent mechanism in colonic epithelial cells. It is unclear why the expression levels of Cyp27b1 and VDR are inversely correlated in the colon during an inflammatory reaction, but the notion that Cyp27b1 induction may serve the purpose of maintaining 1,25(OH)₂D₃-VDR signaling to protect the mucosal barrier is supported by an early report, which showed that global Cyp27b1-null mice developed more severe colonic inflammation than normal mice in a dextran sulfate sodium–induced colitis model (30). Arguably this concept still needs more experimental proof, particularly in mouse models with Cyp27b1 deletion only in intestinal epithelial cells. In addition to the epithelial cells, Cyp27b1 immunostaining data from human biopsies (Fig. 1D) suggest that the nonepithelial components in the colon may also have Cyp27b1 induction. In fact, a previous study has reported Cyp27b1 expression in CD8⁺ T cells (31). Therefore, we cannot exclude the possibility that Cyp27b1 induced in colonic lymphocytes also contributes to the local production of 1,25(OH)₂D₃ during inflammation.

An interesting finding in this study is the need for commensal bacteria for Cyp27b1 induction during colonic inflammatory response. We showed here that depletion of the microbiota by antibiotic treatment completely blocked the upregulation of Cyp27b1 and attenuated colitis development, but Cyp27b1 induction could be completely restored with LPS supplementation. LPS does not seem to directly upregulate epithelial Cyp27b1, however. Rather, LPS stimulates TNF-α secretion from epithelial cells, which then induces Cyp27b1 by an autocrine or paracrine mechanism. It is conceivable that in vivo, when colonic inflammation is initiated by a triggering event, such as TNBS in the mouse model, a compromised mucosal barrier allows commensal bacteria or bacterial endotoxin components to cross the barrier and stimulate lamina propria leukocytes and epithelial cells, which release proinflammatory cytokines such as TNF-α to alter VDR and Cyp27b1 expression in the epithelial and nonepithelial cells. Whereas inflammation-induced VDR downregulation promotes progression of colonic inflammation, Cyp27b1 upregulation counterbalances this detrimental effect, without which, as seen in Cyp27b1-null mice (30), inflammation is expected to be more severe. This effect may be an intrinsic protective homeostatic mechanism in the body.

The notion that colonic VDR and Cyp27b1 are regulated by inflammation is also supported by the observation that vitamin D deficiency exacerbated the opposite changes of these two proteins after TNBS stimulation, because vitamin D deficiency led to more severe colonic inflammation and aggravated colitis. Despite the profound increase of Cyp27b1 in vitamin D deficiency, it is conceivable that the lack of 25-hydroxyvitamin D substrate would decrease the ability of the induced Cyp27b1 to protect the mucosal barrier, which is reflected by the observation that vitamin D deficiency promoted inflammation and colitis in these mice. Overall, the implication of this work is that in colitis development the locally produced 1,25(OH)₂D₃ may play a role that was previously unrecognized in maintaining 1,25(OH)₂D₃ VDR signaling to protect the integrity of the mucosal epithelial barrier. In addition, locally synthesized 1,25(OH)₂D₃ can also act on the immune components of the colon, such as T helper 17 cells and regulatory T cells, to suppress colonic inflammation. It is well established that 1,25(OH)₂D₃ inhibits T helper 17 cells–mediated immune response and promotes regulatory T cell–mediated anti-inflammatory response (32, 33). In this context, it is important to keep a normal vitamin D status so that sufficient substrate can be supplied to Cyp27b1, which produces local 1,25(OH)₂D₃ to maintain the protective VDR signaling in the colon.

Abbreviations:

1,25(OH)₂D₃

1,25-dihydroxyvitamin D₃

AIMD

antibiotic-induced microbiota-depleted

Cyp27b1

cytochrome P450 27b1

D-VitD

vitamin D–deficient

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

IBD

inflammatory bowel disease

IFN

interferon

IKKβ

inhibitor of nuclear factor κB subunit β

IL

interleukin

LPS

lipopolysaccharide

MLC

myosin light chain

MLCK

myosin light chain kinase

mRNA

messenger RNA

NF-κB

nuclear factor κB

PCR

polymerase chain reaction

RT-PCR

reverse transcription polymerase chain reaction

siRNA

small interfering RNA

S-VitD

vitamin D–sufficient

TNBS

2,4,6-trinitrobenzene sulfonic acid

TNF-α

tumor necrosis factor-α

VDR

vitamin D receptor.

Appendix. Antibody Table

Peptide/Protein Target	Antigen Sequence (If Known)	Name of Antibody	Manufacturer, Catalog No.	Species Raised in; Monoclonal or Polyclonal	Dilution Used	RRID
		β-Actin	Santa Cruz, sc-47778	Mouse; monoclonal	WB (1:1000)	AB_626632
		TNF-α	Santa Cruz, sc-52746	Mouse; monoclonal	WB (1:1000) IHC (1:200)	AB_630344
		VDR	Santa Cruz, sc-13133	Mouse; monoclonal	WB (1:1000) IHC (1:500)	AB_628040
		Cyp27b1	Santa Cruz, sc-67261	Rabbit; polyclonal	WB (1:1000) IHC (1:200)	AB_2089287
		TLR4	Santa Cruz, sc-293072	Mouse; monoclonal	WB (1:1000)	AB_10611320
		L-MLCK	Abcam, ab76092	Rabbit; monoclonal	WB (1:1000)	AB_1524000
		P-MLC	Cell Signaling, 3674	Rabbit; polyclonal	WB (1:1000)	AB_2147464
		PUMA	Cell Signaling, 7467	Rabbit; monoclonal	WB (1:1000)	AB_10829605

Abbreviations: IHC, immunohistochemistry; RRID, Research Resource Identifier; WB, Western blot.