ORIGINAL ARTICLE

Opposite Effects of Interferon Regulatory Factor 1 and on the of Epithelial Cells Induced by TNF-a in Inflammatory Bowel Disease Ruihan Tang, MD, PhD,* Guang Yang, MD, PhD,*,† Shenghong Zhang, MD, PhD,* Changyou Wu, MD, PhD,‡ and Minhu Chen, MD, PhD* Downloaded from https://academic.oup.com/ibdjournal/article/20/11/1950/4578913 by guest on 23 September 2021

Background: Inflammatory bowel disease (IBD) is characterized by a damaged intestinal epithelium barrier. Interferon regulatory factor 1 (IRF1) and osteopontin (OPN) regulate survival and growth in a variety of circumstances but their effects on the intestinal epithelium have not been elucidated. In this study, we sought to determine the effects of OPN on intestinal epithelial cells under conditions of (TNF)-a–induced inflammation and whether IRF1 regulates OPN expression, the activation of downstream pathways, and inflammatory responses. Methods: The expression levels of OPN and IRF1 were assessed by immunohistochemical analyses of human IBD and experimental mouse colitis. The effects of IRF1 and OPN on inflammatory responses were investigated in vitro in NCM460 and Caco-2 cells stimulated by TNF-a. Changes in p-AKT, p-P38, and p-ERK levels were quantified by western blotting assays. The regulation of OPN expression by IRF1 was determined by luciferase activity and chromatin immunoprecipitation assays. Results: IRF1 was upregulated in human IBD and in the colon epithelium of mice with dextran sulfate sodium–induced colitis. Additionally, IRF1 was correlated with high-sensitivity C-reactive , erythrocyte sedimentation rate, Crohn’s disease activity index, Crohn’s disease endoscopic index of severity, and simple endoscopic score for Crohn’s disease in Crohn’s disease and with high-sensitivity C-reactive protein, erythrocyte sedimentation rate, Mayo score, Baron score, modified Baron score, Rachmilewitz score, ulcerative colitis endoscopic index of severity, ulcerative colitis colonoscopic index of severity, and disease duration in ulcerative colitis. The expression of OPN was significantly decreased in patients with IBD compared with controls and in dextran sulfate sodium–induced experimental colitis and was also inversely correlated with clinical and endoscopic activities in both Crohn’s disease and ulcerative colitis. TNF-a treatment upregulated IRF1 and diminished OPN in both NCM460 and Caco-2 cells. The overexpression of OPN and rhOPN ameliorated the apoptosis induced by TNF-a, whereas the overexpression of IRF1 aggravated apoptosis, indicating opposite effects of OPN and IRF1 in inflamed epithelial cells. The luciferase and chromatin immunoprecipitation assays showed that IRF1 transcriptionally modulated the expression of OPN. TNF-a inhibited the OPN-induced upregulation of p-ERK, p-P38, and p-AKT. Conclusions: Our data suggest that during intestinal inflammation, the TNF-a–mediated activation of IRF1 is related to the subsequent suppression of OPN expression, further reducing p-AKT, p-P38, and p-ERK activities and resulting in aggravation of the injury to intestinal epithelial cells. (Inflamm Bowel Dis 2014;20:1950–1961) Key Words: TNF-a,inflammatory bowel diseases, osteopontin, IRF1

nflammatory bowel disease (IBD), which includes Crohn’sdis- I ease (CD) and ulcerative colitis (UC), is characterized by chronic inflammation and epithelial barrier damage.1,2 Accumulating evi- Received for publication June 3, 2014; Accepted July 13, 2014. a From the *Department of Gastroenterology, the First Affiliated Hospital, Sun dence has shown that tumor necrosis factor (TNF)- acts as a key Yat-sen University, Guangzhou, China; †Department of Medical Imaging and Inter- stimulator in both cell survival and proapoptotic pathways, exerting ventional Radiology, Sun Yat-sen University Cancer Center, State Key Laboratory of its impact by regulating proinflammatory to induce tissue Oncology in South China, Guangzhou, China; and ‡Department of Immunology, 3–5 a Zhongshan School of Medicine, Sun Yat-Sen University, Guangzhou, China. degradation or protection. Elevated levels of TNF- in the intes- M. Chen was supported by grants from the National Natural Science tinal tissue and serum have been well established to be associated Foundation of China (grant no. 81270473), the Key Project of the National 12th with the severity and extent of both CD and UC.6,7 Additionally, five-year Research Program of China (grant no. 2012BAI06B03), the Scientific and fi Technological Planning of Guangzhou (grant no. 2011YZ-00004), and the clinical trials on anti-TNF therapies have shown great ef cacy in Fundamental Research Funds for Sun Yat-sen University (grant no. 13YKJC01). inducing IBD mucosal healing.8,9 However, the detailed pathways The authors have no conflicts of interest to disclose. and mechanisms require further elucidation. Reprints: Minhu Chen, MD, PhD, Department of Gastroenterology, the First IRF1, an interferon regulatory factor, was originally identi- Affiliated Hospital of Sun Yat-Sen University, 58 Zhongshan II Road, Guangzhou fi 510080, China (e-mail: [email protected]). ed as a key regulator of interferon-stimulated . IRF1 serves Copyright © 2014 Crohn’s & Colitis Foundation of America, Inc. as a in many biological processes, such as 10,11 DOI 10.1097/MIB.0000000000000192 inflammation and tumor proliferation. Several studies in murine Published online 9 September 2014. colitis models have demonstrated that TNF-a significantly induces

1950 | www.ibdjournal.org Inflamm Bowel Dis Volume 20, Number 11, November 2014 Inflamm Bowel Dis Volume 20, Number 11, November 2014 Opposite Effects of IRF 1 and Osteopontin the expression of IRF1.12,13 Only limited efforts have been Participants and Disease Activity directed toward the investigation of IRF1 about IBD. One report Between January 2010 and April 2014, patients diagnosed on suggested that conditional knockout of the IRF1 protected the basis of the standard clinical, endoscopic, and histological mice from chronic inflammation,14 whereas another report criteria of IBD at the Gastroenterology Departments of the First showed that IRF1 exhibited few anti-inflammatory effects in Affiliated Hospital of Sun-Yat Sen University were included. The dextran sulfate sodium (DSS)–induced colitis.15 The expression study was approved by the hospital ethics committees, and written of IRF1 in patients with IBD is still unclear. A recent study informed consents were obtained. The were obtained concluded that there was no obvious change in IRF1 expression during endoscopy or surgery. Blood samples for the measurement in patients compared with controls.16 However, another study of high-sensitivity C-reactive protein and erythrocyte sedimentation observed increased expression of IRF1 in 72% of patients with rate were taken within 1 week before or after endoscopy. Patients Downloaded from https://academic.oup.com/ibdjournal/article/20/11/1950/4578913 by guest on 23 September 2021 CD among study cohorts.17 Still, little is known about how IRF1 who had infectious colitis and colorectal cancer were excluded. interacts with multiple cytokines and activates different signal- Colon biopsies were obtained from healthy volunteers during ing pathways in the context of IBD. endoscopy or incised from unaffected areas surgically removed Osteopontin (OPN) is a secreted arginine-glycine-aspartic from patients with colon cancer. The clinical disease activity was acid–containing phosphoprotein that is ubiquitously expressed in assessed by the measurement of the Crohn’s disease activity index a variety of tissues.18 By binding to integrins and CD44 variants, for CD and of the Mayo score for UC by a physician who was it is active in many biological processes, including wound heal- blinded to the colonoscopic appearance of the patient.30,31 Endos- ing, tumorigenesis, inflammation, and immune responses.19,20 copies were performed and graded according to the Crohn’sdisease OPN is pivotal in maintaining cell survival by inhibiting apopto- endoscopic index of severity and simple endoscopic score for sis. The downregulation of OPN sensitizes cells to apoptosis.21 Crohn’s disease scoring systems in patients with CD and according OPN, a soluble , was also found to be involved in var- to the Baron, modified Baron, Rachmilewitz, ulcerative colitis – ious inflammatory diseases and autoimmune disorders.22 25 OPN endoscopic index of severity, and ulcerative colitis colonoscopic primarily serves as a factor that fosters survival in different sit- index of severity scores for UC.32–38 To avoid bias, all gastroenter- uations. However, whether OPN has pro- or anti-inflammatory ologists performing the endoscopies were unaware of the results of – effects in IBD remains controversial.26 28 The regulation of OPN the Crohn’s disease activity index, Mayo score, high-sensitivity C- expression in the inflamed mucosa of patients with IBD has yet to reactive protein, or erythrocyte sedimentation rate. be investigated. A recent report found that IRF1 siRNA resulted in Mice and Experimental Colitis decreased expression of OPN in cardiac fibroblasts.29 However, Male C57BL/6 mice (n ¼ 24, age 6–8 wk, weight 18–20 g) the influence of IRF1 on the expression of OPN in other cells or were obtained from Guangdong Medical Laboratory Animal Center tissues has not been reported to date. We speculated that (Guangzhou, Guangdong, China). The mice were randomly assigned increased TNF-a in the inflamed intestinal mucosa might sup- to the experimental group (n ¼ 12) or the control group (n ¼ 12). The press OPN gene expression at the transcriptional level by stim- mice in the experimental group were subjected to acute colitis induced ulating IRF1 regulation. In this study, we investigated the by 2% (wt/vol) DSS (MW 36,000–50,000; MP Biochemicals, Solon, association between IRF1/OPN levels in epithelial cells and OH) in their drinking water for 7 days.39 The control mice were given clinical indices of disease activity in patients with IBD. Addi- distilled water. On day 8, the mice were euthanized, and their colons tionally, we revealed subtle crosstalk among TNF, IRF1, and were collected for analysis.40 The colon length was measured before OPN signaling and detected cytoprotective effects of OPN histological evaluation. The colon epithelium was isolated by scraping against TNF-a on growth and apoptosis in intestinal epithelial the mucosal surface with a glass slide. RNA and protein were ex- cells (IECs), further confirming the influenceofIRF1onOPN tracted for reverse-transcriptase polymerase chain reaction (PCR) and expression at the transcriptional level. western blot analysis. The remaining colon tissues were Swiss-rolled, formalin-fixed, and paraffin-embedded for histological and immuno- histochemical analyses. All animal use was approved and monitored MATERIALS AND METHODS by the hospital’s ethics committees. Materials Cell Culture and Transfection Recombinant human OPN was purchased from PeproTech The Caco-2 cell line was a gift from Dr. Jianjun Tang of the (Rocky Hill, NJ), prepared in water and stored in aliquots of State Key Laboratory of Oncology in Southern China, Sun Yat-Sen appropriate volumes at 2808C until use. The antibodies used were University, Guangzhou, China. The Caco-2 cells were cultured in anti-OPN (Assay Biotechnology, Sunnyvale, CA) and anti-IRF1 Dulbecco’s Modified Eagle’s Medium supplemented with heat- (Santa Cruz, Dallas, TX), as well as anti-p-AKT, anti-AKT, anti- inactivated fetal calf serum (10%) and penicillin (100 U/mL) or p-EKR, anti-P38 and anti-actin (Beyotime, Nantong, China). The streptomycin (100 mg/mL) in 5% CO2 at 378C. NCM460 cells were siRNAs against OPN and IRF1 were synthesized (Genepharma, obtained from Jennio Biotechnology (Guangzhou, China) and cul- Suzhou, China). tured in M3:10 TM medium. This cell line is considered to be

www.ibdjournal.org | 1951 Tang et al Inflamm Bowel Dis Volume 20, Number 11, November 2014

a noncancerous colonic epithelial cell line. The transfection of plas- with anti-IRF1 antibodies at 48C for 15 hours, the cell cultures were mids and siRNA was performed using Lipofectamine 3000 (Invitro- washed 3 times with PBS. The cells were then incubated with the gen, Carlsbad, CA) according to the manufacturer’s instructions. secondary antibody for 1 hour at room temperature and observed through an inverted fluorescence microscope after being washed 3 MTT Assay times with PBS. The cell lines were seeded in 96-well plates at a density of 5000 cells per well in a volume of 150 mL of cell culture medium RNA Extraction and per well for 24 hours before drug exposure. OPN and TNF-a were Reverse-transcriptase PCR added into the wells at different concentrations in triplicate. The Total RNA was extracted from the cultured cells or colonic plates were incubated at 378Cwith5%CO for 48 hours. Once the 2 biopsies from the patients with IBD and controls using RNAiso Downloaded from https://academic.oup.com/ibdjournal/article/20/11/1950/4578913 by guest on 23 September 2021 transfection was performed, TNF-a was added to the wells 24 hours Plus (Takara, Otsu, Japan). Total RNA was reverse transcribed after transfection. A 20-mL sample of MTT solution (5 g/L, dis- using the Reverse Transcription System (Promega, Madison, WI) solved in phosphate-buffered saline [PBS]) was added to each well, according to the manufacturer’s instructions. PCR was performed and the plates were incubated at 378C for an additional 4 hours. The using PCR Mix (TransGen Biotech, Beijing, China). supernatant was discarded, and 150 mL of dimethyl sulfoxide was added to dissolve the formazan product. The absorbance values at Western Blot Analysis

490 nm (A490) were determined using a multiwell plate reader Caco-2 and NCM460 cells were exposed to various experi- (Tecan, Männedorf, Switzerland). mental conditions for the indicated times before being harvested and lysed for protein extraction using sodium dodecyl sulfate (Sigma- Flow Cytometry Analysis Aldrich, St. Louis, MO). The protein concentrations were determined The samples were analyzed using a flow cytometer (Beck- using the Bio-Rad Protein Assay Kit (Bio-Rad, Hercules, CA). The man Coulter, Urbana, IL). To detect apoptotic cells, labeling tests blots were visualized using the enhanced chemiluminescence detec- involving both propidium iodide and annexin-V were performed tion system (Amersham, Piscataway, NY), and b-actin was used as using an Annexin-V staining (Invitrogen) according to the a loading control. manufacturer’s instructions. Luciferase Reporter Assay Immunohistochemistry NCM460 cells were seeded at 5 · 103 per well in 96-well An immunohistochemical analysis was performed to evaluate plates 24 hours before transfection. A 683-bp (2365to+318) the expression levels of OPN and IRF1 in paraffin-embedded fragment of the region of OPN was inserted into the sections. The general processes were as follows: Paraffin sections pGL3-Basic luciferase reporter vector (Promega). The cells were were baked for deparaffinization and washed. After assessing the cotransfected with 0.1 mgofthefirefly luciferase reporter construct, endogenous peroxidase activity, the sections were incubated with 3% 0.01 mg of the pRL-TK Renilla luciferase reporter plasmid (Prom- H2O2 for 15 minutes at room temperature and washed with distilled ega), and the pcDNA3.1-IRF1 vector using Lipofectamine 3000 water and PBS. Then, the sections were treated with ethylenediami- (Invitrogen). The luciferase activity was examined using a dual- netetraacetic acid in a microwave oven for 10 minutes for luciferase reporter assay system (Promega) according to the manu- retrieval. After subsequent blocking with 1% bovine serum albumin facturer’s instructions, and the signal was normalized to the internal for 10 minutes, the sections were incubated with rabbit anti-human Renilla control to determine the efficiency of the transfection. OPN antibodies (1:100) overnight at 48C and then incubated with goat anti-rabbit EnVision working solution for 30 minutes at room Chromatin Immunoprecipitation temperature. Horseradish peroxidase activity was detected NCM460 cells were used for chromatin immunoprecipita- by adding the chromogenic substrate 3,30-diaminobenzidine tetra- tion with the EZ ChIP kit (Millipore, Billerica, MA) according to chloride (DAB; Dako, Copenhagen, Denmark) for 30 seconds, the manufacturer’s instructions. After elution and purification, the whichresultedinbrownstaining.Allthestepswerefollowedby3 recovered immunoprecipitated DNA samples were used for PCR, washes with PBS for 5 minutes each. Negative controls were treated with the primers 50-GTGGCCTACTTTACATACC-30 (forward) with PBS under the same conditions, rather than with primary anti- and 50-GGACTATATGTGTGTTAAC-30 (reverse) to amplify bodies. Image Pro Plus 6.0 (Media Cybernetics, Rockville, MD) was a 185-bp segment of the OPN promoter containing the potential used in calculating the optical density of IRF1 and OPN. IRF1 binding sites (2284 to 2272). The PCR products were analyzed by agarose gel electrophoresis. Immunofluorescence After being treated with TNF-a for 6 hours, the cells were Statistical Analysis washed with PBS and fixed for 20 minutes at room temperature with The results for numerical data are presented as mean 6 SD. 4% paraformaldehyde. They were washed again with PBS and per- The Mann–Whitney U test was used to explore the associations meabilized with Triton X-100 for 10 minutes at room temperature. between the nonparametric numerical data in 2 independent groups, Then, the cells were washed with PBS and incubated with 3% bovine and the t test was used for parametric numerical data. The associ- serum albumin to block the nonspecific binding sites. After incubation ation between 2 variables was assessed by Spearman’srank

1952 | www.ibdjournal.org Inflamm Bowel Dis Volume 20, Number 11, November 2014 Opposite Effects of IRF 1 and Osteopontin

TABLE 1. Demographic and Clinical Characteristics of the Study Population

CD UC Controls

N 211 100 110 Male, n (%) 136 (64.5) 62 (62.0) 61 (55.5) Age, yr 30.1 6 11.1 (11–65) 43.2 6 14.3 (14–70) 34.5 6 6.0 (23–49) Disease duration, mo 37.0 (1–252) 52.5 (12–132) Age at diagnosis, yr 26 (10–65) ——

Disease phenotype, n (%) Downloaded from https://academic.oup.com/ibdjournal/article/20/11/1950/4578913 by guest on 23 September 2021 B1 169 (80.1) —— B1p 7 (3.3) —— B2 34 (16.1) —— B2p 6 (2.8) —— B3 8 (3.8) —— B3p 7 (3.3) —— Disease location, n (%) L1 52 (24.6) —— L2 82 (38.9) —— L3 77 (36.5) —— Disease extent, n (%) Proctosigmoiditis — 52 (52.0) — Left-sided colitis — 19 (19.0) — Pancolitis — 29 (29.0) — IBD-related surgery history, n (%) 37 (17.5) 11 (11.0) — , n (%) 5-ASA 50 (23.7) 82 (82.0) — Corticosteroids 91 (43.1) 33 (33.0) — 92 (43.6) 17 (17.0) — 6-Mercaptopurine 14 (6.6) 1 (1.0) — 9 (4.3) 0 (0.0) — 18 (8.5) 0 (0.0) — Anti-TNF-a therapy 16 (7.6) 5 (5.0) — No 5 (2.4) 4 (4.0) —

Normally distributed data are presented as mean 6 SD, whereas nonparametric data are presented as median, minimal, and maximal values. Locations: L1 ¼ terminal ileum; L2 ¼ colon; L3 ¼ ileocolon. Patients with L4 (isolated upper disease) were excluded. Behaviors: B1 ¼ nonstricturing; B2 ¼ stricturing, B3 ¼ penetrating; P ¼ perianal disease. ASA, aminosalicylic acid.

correlation coefficient (r) for nonparametric correlations. Two-sided (CD: n ¼ 211, 174 endoscopic samples and 37 surgical P values ,0.05 were considered statistically significant. resection samples; UC: n ¼ 100, 91 endoscopic biopsy samples and 9 surgical resection samples), and 70 normal surgical resections and 40 intestinal biopsies were used as controls (n ¼ 110). Increased RESULTS expression of IRF1 in the IECs was observed in both the CD and the UC patients compared with the controls (Fig. 1B). This observation Expression Levels of IRF1 and OPN Were was confirmed by quantitative analysis. The average optical density Inversely Correlated in Patients with (AOD) of IRF1 was significantly higher in the CD and UC groups Active CD or UC than that in the control group (n ¼ 211, AOD ¼ 0.171 6 0.063; A total of 211 patients with CD, 100 patients with UC, and n ¼ 100, AOD ¼ 0.151 6 0.049; and n ¼ 110, AOD ¼ 0.101 6 110 controls were enrolled (Table 1). Immunohistochemical 0.047, respectively; P , 0.001) (Fig. 1C). The AOD of IRF1 was analyses was performed to examine the expression and localiza- correlated with several indicators of disease activity in both CD and tion of IRF1 and OPN in the biopsies from the patients with IBD UC, suggesting that the IRF1 level is associated with IBD disease

www.ibdjournal.org | 1953 Tang et al Inflamm Bowel Dis Volume 20, Number 11, November 2014 Downloaded from https://academic.oup.com/ibdjournal/article/20/11/1950/4578913 by guest on 23 September 2021

FIGURE 1. The expression levels of IRF1 and OPN in human IBD. A, The levels of OPN and IRF1 messenger RNA were determined by reverse-transcriptase PCR. The relative levels were quantified through the gray values. The data indicated that the level of IRF1 messenger RNA was increased, whereas the level of OPN messenger RNA was decreased in both UC and CD tissues (P , 0.001) (Mann–Whitney U test). B, Representative immunostaining of IRF1 and OPN in normal intestinal and IBD-inflamed tissues. OPN expression was robust in normal intestinal epithelium and low or even undetected in IBD-damaged epithelium, whereas IRF1 showed a greater expression in epithelium of IBD tissues compared with normal tissues. C, The average optical densities of OPN were significantly higher in the CD and UC groups than in the control group, P , 0.001. In contrast to OPN expression, the average optical densities of IRF1 were much higher in IBD tissues than in normal tissues, P , 0.001. Furthermore, there were inverse correlations of the IRF1 levels with the OPN levels in both CD (r ¼ 20.593, P , 0.001) and UC (r ¼ 20.691, P , 0.001) epithelium. ***P , 0.001 versus control.

1954 | www.ibdjournal.org Inflamm Bowel Dis Volume 20, Number 11, November 2014 Opposite Effects of IRF 1 and Osteopontin

and CD (n ¼ 30) tissues compared with normal tissues (n ¼ 30, TABLE 2. Correlation of the Expression of IRF1 and P , 0.001) (Fig. 1A). OPN with Disease Activity Indicators in CD In addition, the expression of OPN showed an inverse ¼ 2 , IRF1 OPN association with the IRF1 level in CD (r 0.593, P 0.01) and UC (r ¼ 20.691, P , 0.01) (Fig. 1C), indicating that these 2 r P r P factors may have different effects on IEC injury in IBD. hs-CRP 0.488 ,0.01 20.416 ,0.01 ESR 0.377 ,0.01 20.366 ,0.01 IRF1 Expression Was Elevated, Whereas OPN CDAI 0.461 ,0.01 20.385 ,0.01 Expression Was Depleted in CDEIS 0.402 ,0.01 20.436 ,0.01 Experimental Colitis Downloaded from https://academic.oup.com/ibdjournal/article/20/11/1950/4578913 by guest on 23 September 2021 SES-CD 0.336 ,0.01 20.410 ,0.01 To explore the role of IRF1 and OPN in the IBD context in vivo, we induced acute colitis in C57BL/6 mice using 2% DSS CDAI, Crohn’s disease activity index; CDEIS, Crohn’s disease endoscopic index of in their drinking water. The mice developed severe colitis with severity; ESR, erythrocyte sedimentation rate; hs-CRP, high-sensitivity C-reactive pro- apparent weight loss and bloody diarrhea. The colon lengths of ’ fi tein; r, Spearman s rank correlation coef cient; SES-CD, simple endoscopic score for the mice with colitis were significantly decreased (controls 108 6 Crohn’s disease. 3.6 mg versus DSS-induced colitis 64 6 5.7 mg; P , 0.01), and mucosal inflammation, edema, and thickening were observed. Additionally, hematoxylin–eosin staining revealed crypt epithelial activity (Tables 2 and 3). IRF1 staining was also correlated with disruption, ulceration, and diffuse infiltration of inflammatory ¼ ¼ disease duration in the patients with UC (r 0.532, P 0.032) cells in the mice with colitis. In contrast, the control mice ex- (Tables 2 and 3). hibited no detectable inflammation. In contrast, our quantitative analyses showed that OPN was Western blotting analyses of mucosal scrapings were per- fl highly expressed in nonin ammatory IECs; the depletion of OPN was formed to detect both IRF1 and OPN expression. The results showed slightly more severe in the CD group than in the UC group (Fig. 1B). much more robust expression of IRF1 in the experimental group The AOD of OPN was much lower in the CD and UC groups than than in the control group, whereas the expression of OPN was ¼ ¼ 6 ¼ in the control group (n 211, AOD 0.125 0.064; n 100, weaker in the experimental group than in the control group (Fig. 2B). ¼ 6 ¼ ¼ 6 AOD 0.149 0.059; and n 110, AOD 0.230 0.057, The immunohistochemical analysis results demonstrated that IRF1 , respectively; P 0.001) (Fig. 1C). Additionally, the expression of expression was lacking in the epithelial cells in the unchallenged OPN correlated well with several disease activity indices of IBD colon, except for occasional staining in the cytoplasm (AOD ¼ (Tables 2 and 3). 0.154 6 0.029) (Fig. 2A), whereas the expression of OPN was Consistent with expression level assessed by immunohis- diffuse and intense in the epithelial cells (AOD ¼ 0.428 6 tochemical analyses, an elevated level of IRF1 and a decreased 0.063). In contrast, IRF1 expression was elevated in the compro- ¼ level of OPN messenger RNA were observed in both UC (n 30) mised epithelium (AOD ¼ 0.390 6 0.054), whereas OPN staining was nearly undetectable (AOD ¼ 0.175 6 0.029) (Fig. 2A). Addi- tionally, AOD analysis suggested an inverse association between the TABLE 3. Correlation of the Expression of IRF1 and ¼ 2 ¼ OPN with Disease Activity Indicators in UC expression of IRF1 and OPN (r 0.643, P 0.024) (Fig. 2C).

IRF1 OPN Roles of IRF1 and OPN in the Effects of TNF-a on the Viability and Apoptosis of Intestinal r P R P Epithelial Cells hs-CRP 0.555 ,0.01 20.707 ,0.01 The viability of the Caco-2 cell lines was measured by the MTT ESR 0.505 ,0.01 20.581 ,0.01 assay after treatment with rhOPN and TNF-a. As shown in Figure 3A, Mayo UC 0.393 ,0.01 20.489 ,0.01 after the treatment of Caco-2 cells with rhOPN of different concen- Baron score 0.262 0.01 20.296 ,0.01 trations and TNF-a (50 ng/mL), the viability of the Caco-2 cells Modified Baron score 0.205 0.04 20.261 0.01 increased compared with that of cells treated with TNF-a (50 ng/ Rachmilewitz score 0.256 0.01 20.352 ,0.01 mL) alone. This result indicated that rhOPN counteracted the cytotoxic UCEIS score 0.233 ,0.01 20.274 0.01 effects of TNF-a on the Caco-2 cells. The constructed IRF1 and OPN UCCIS score 0.340 0.02 20.424 ,0.01 expression vectors and siRNAs against IRF1 or OPN were also used Disease duration 0.599 0.01 20.537 0.02 to examine these effects on NCM460 cells (Fig. 3B). The overexpres- sion of IRF1 made it easier for TNF-a to decrease the viability of the ESR, erythrocyte sedimentation rate; hs-CRP, high-sensitivity C-reactive protein; Mayo NCM460 cells, whereas the overexpression of OPN counter- ’ fi UC, Mayo score for ulcerative colitis; r, Spearman s rank correlation coef cient; UCCIS, acted this effect of TNF-a (Fig. 3C, D). After being exposed ulcerative colitis colonoscopic index of severity; UCEIS, ulcerative colitis endoscopic a index of severity. to TNF- for 12 hours, the level of apoptosis among the cultured cells was analyzed by flow cytometry and the terminal dexynucleotidyl

www.ibdjournal.org | 1955 Tang et al Inflamm Bowel Dis Volume 20, Number 11, November 2014 Downloaded from https://academic.oup.com/ibdjournal/article/20/11/1950/4578913 by guest on 23 September 2021

FIGURE 2. The expression locations and levels of IRF1 and OPN in experimental colitis. A, Representative immunostaining of IRF1 and OPN in colon tissues of controls and DSS-induced colitis. OPN expression was intense in epithelium of normal colon tissues (AOD ¼ 0.428 6 0.063) and much lower in inflamed epithelium (AOD ¼ 0.175 6 0.029, P , 0.001). In contrast, IRF1 showed a greater expression in challenged colon epithelium (AOD ¼ 0.390 6 0.054) compared with intact epithelium (AOD ¼ 0.154 6 0.029, P , 0.001). B, Western blotting analyses of representative colon samples showed that the expression of OPN was decreased, whereas the expression of IRF1 was elevated in DSS-induced colitis (P , 0.001) (Mann–Whitney U test). C, The inverse correlations of the IRF1 levels with the OPN levels in DSS-induced colitis was significant (r ¼ 20.643, P ¼ 0.024). ***P , 0.001 versus control. transferase(TdT)-mediated dUTP nick end labeling (TUNEL) IRF1 Transcriptionally Downregulated OPN assay. As shown in Figure 4A, the proportion of apoptotic cells Expression in Intestinal Epithelial Cells in was decreased in the rhOPN- and TNF-a-treated groups compared Response to TNF-a with the TNF-a-treated group. The overexpression of OPN in the To determine whether OPN is inhibited by IRF1, we NCM460 cells treated with TNF-a also decreased the number of performed a western blot assay and found that the overexpression apoptotic cells compared with the number in the control group, of IRF1 decreased the level of OPN (Fig. 6A). Luciferase reporter whereas the overexpression of IRF1 elevated the number of apo- gene assays were performed by transfecting the IECs with con- ptotic cells (Fig. 4B, C). structs containing regions of the human OPN promoter followed by treatment with TNF-a. The data suggested that TNF-a could TNF-a Stimulated IRF1 Expression and significantly decrease the promoter activity of OPN by ;45% Decreased OPN Expression in Intestinal (P , 0.01) compared with the control group (Fig. 6B). Further- Epithelial Cells more, dual luciferase assays were performed on extracts from Caco-2 and NCM460 cells were exposed to TNF-a at differ- NCM460 cells that were cotransfected with the luciferase reporter ent doses and times. The IRF1 and OPN expression levels were plasmid and either IRF1 or the pcDNA3.1 empty vector. As detected by western blotting (Fig. 5A). TNF-a significantly elevated shown in Figure 6B, the overexpression of IRF1 decreased the expression of IRF1 in the Caco-2 and NCM460 cells in a dose- OPN promoter activity compared with the control group. dependent manner. In contrast, TNF-a inhibited the expression of To investigate the physical interaction between IRF1 and OPN in the Caco-2 and NCM460 cells. The elevated expression of the gene-regulatory regions on OPN, we performed a chromatin IRF1 was further confirmed by immunofluorescence (Fig. 5B). immunoprecipitation (CHiP) assay with NCM460 cells. An

1956 | www.ibdjournal.org Inflamm Bowel Dis Volume 20, Number 11, November 2014 Opposite Effects of IRF 1 and Osteopontin Downloaded from https://academic.oup.com/ibdjournal/article/20/11/1950/4578913 by guest on 23 September 2021

FIGURE 3. Roles of IRF1 and OPN in the effect of TNF-a on the viability of intestinal epithelial cells. A, The rhOPN counteracted the cytotoxic effect of TNF-a on Caco-2 cells’ viability in a dose-dependent manner. B, The siRNAs against IRF1 and OPN effectively reduced the IRF1 and OPN levels in NCM460 cells, whereas IRF1 and OPN recombinant vectors significantly elevated the OPN and IRF1 expression levels. C, Overexpression of IRF1 facilitated the cytotoxic effects of TNF-a (1.75 6 0.09 versus 1.33 6 0.19, P , 0.05), whereas IRF1 knockdown showed no obvious influence on this effect. D, The downregulation of OPN by OPN siRNA facilitated the cytotoxic effects of TNF-a on NCM460 cells’ viability (1.76 6 0.11 versus 1.39 6 0.19, P , 0.05), and OPN overexpression counteracted the cytotoxic effects of TNF-a (1.50 6 0.13 versus 1.89 6 0.21, P , 0.05). *P , 0.05 versus control. online program was used to analyze the potential IRF1 binding enhanced the binding of IRF1 to the promoter region of OPN, as the motif in the OPN promoter region. The 683-bp promoter region was IRF1-specific DNA band was much stronger for TNF-a-treated cloned into the pGL3-Basic luciferase reporter plasmid. The epithelial cells compared with untreated cells (Fig. 6D). template used for PCR was immunoprecipitated from the DNA/ protein complex with an anti-IRF1 antibody, and potential IRF1 TNF-a Compromised OPN-induced binding sites were found in the amplified region. As shown in Prosurvival Signals of p-P44/42, p-P38, Figure 6C, we determined that the binding was specific because and p-AKT astrongbandwasobtainedbyChIPusingtheIRF1antibody, OPN has been reported to elevate the p-P44/42, p-P38, whereas amplification in the immunoglobulin G control group and p-AKT signals, which are all involved in cell proliferation. showed no bands. Furthermore, we found that TNF-a significantly In this study, we detected these factors to examine the influence

www.ibdjournal.org | 1957 Tang et al Inflamm Bowel Dis Volume 20, Number 11, November 2014 Downloaded from https://academic.oup.com/ibdjournal/article/20/11/1950/4578913 by guest on 23 September 2021

FIGURE 4. Effects of OPN and IRF1 on the TNF-a–induced apoptosis of intestinal epithelial cells. A, The rhOPN reduced the TNF-a–induced apoptosis of NCM460 (11.5 6 5.5% versus 16.1 6 6.1%, P , 0.05) cells as demonstrated by flow cytometric analysis. *P , 0.05 versus non-rhOPN treatment. B, OPN overexpression decreased TNF-a–induced apoptotic cells (2.8 6 8.0% versus 6.7 6 8.8%, P , 0.05) and IRF1 overexpression increased the level of apoptotic cells (13.4 6 7.4% versus 6.7 6 8.8%, P , 0.05). *P , 0.05 versus empty vector transfection. C, TUNEL assays indicated that OPN overexpression decreased the cell apoptosis induced by TNF-a (4.9 6 2.5% versus 9.6 6 1.3%, P , 0.05), whereas IRF1 overexpression facilitated the TNF-a–induced apoptosis of NCM460 cells (15.5 6 1.9% versus 9.6 6 1.3%, P , 0.05). *P , 0.05 versus empty vector. of TNF-a on the effects of OPN on these signals. Our results function may be critical to both the initial pathogenesis and indicated that TNF-a inhibited any OPN-induced elevation of p- reactivation processes of chronic intestinal diseases.1,2 Increased P44/42, p-P38, and p-AKT levels. Our data also showed that rates of intestinal epithelial shedding are well known to be com- TNF-a could further reduce the levels of p-P44/42, p-P38, and monly associated with increased mucosal TNF-a production, and p-AKT, which were downregulated by OPN siRNA in NCM460 they can be alleviated by anti-TNF therapies, indicating the cen- cells (Fig. 6E). tral role of TNF-a in IBD pathogenesis.8,9 In this study, we dem- onstrated a novel downstream pathway of TNF-a. IRF1 is the first identified interferon regulatory factor. DISCUSSION Accumulating evidence has revealed that IRF1 acts as an Apparent lesions in the intestinal mucosa caused by important transcriptional factor in many biological processes, compromised intestinal barrier function represent a common including inflammation, apoptosis, and tumorigenesis.10,11 IRF1 feature of IBD. In CD and UC, disease reactivation is paralleled hasbeenreportedtobenormallyexpressedatlowlevelsin by the level of damage to barrier integrity. Disrupted barrier normal tissues, although its expression can be stimulated by

1958 | www.ibdjournal.org Inflamm Bowel Dis Volume 20, Number 11, November 2014 Opposite Effects of IRF 1 and Osteopontin Downloaded from https://academic.oup.com/ibdjournal/article/20/11/1950/4578913 by guest on 23 September 2021

FIGURE 5. TNF-a decreased the OPN level and stimulated IRF1 expression in intestinal epithelial cells. A, Western blotting assays indicated that TNF-a reduced OPN levels in a time- and dose-dependent manner in Caco-2 and NCM460 cells. B, Immunofluorescence further verified that TNF-a obviously elevated IRF1 expression in Caco-2 and NCM460 cells.

TNF-a. Furthermore, IRF1 acts as an initial signaling response anti-inflammatory role in IBD.26–28 The data and their subsequent to TNF-a and transcriptionally modulates the amplitude and significance are not well defined. Some reports have indicated that character of various cytokines.12,13 Thus, IRF1 stimulation by OPN expression is elevated in IBD promoting inflammatory pro- TNF-a might represent a positive feedback mechanism under cesses, whereas others have found that OPN is decreased in the inflammatory conditions. However, recent reports have IECs but elevated in the lamina propria exudative cells.26 OPN described the quantity of IRF1 in challenged IECs in a contro- has been regarded as an anti-apoptotic factor in several patholog- versial manner; how it modulates other genes in IBD remains ical events, such as activation-induced cell death in challenged under investigation.16,17 In our study, we found that the IRF1 and endothelial cells.19,42 Our data showed that OPN expression was elevated in IBD samples compared with the levels were significantly decreased in IECs in human IBD samples control groups. Furthermore, in a tractable experimental mouse and DSS-induced mouse colitis. This result was consistent with colitis model, IRF1 expression was significantly enhanced in the data from cultured IECs with reduced OPN expression after challenged epithelial cells. Additionally, we demonstrated that being treated with TNF-a. Our data indicated that OPN overex- TNF-a stimulated IRF1 expression in cultured IECs. This pression protected cells from TNF-a cytotoxicity. Together with TNF-a–induced apoptosis was primed by preliminarily trans- the observation that OPN siRNA aggravated apoptosis in cultured fected IECs with the IRF1 expression vector. Interestingly, the IECs, these data positioned OPN as an important participant in the change in IRF1 among TNF-a–challenged epithelial cells epithelial response to injury and inflammation. The roles of OPN seemed to be negatively correlated with the OPN levels in the in IBD remain controversial28; one possible explanation for these IBD tissues. Given the opposite trends in the expression levels divergent results is a biphasic response of OPN in the acute and of IRF1 and OPN, we speculated that OPN might be regulated chronic phases of inflammation. OPN promotes survival in acute by IRF1 in IECs. inflammation but is cytotoxic in sustained inflammation. In future In contrast, TNF-a promotes the depletion of OPN in the studies, we will examine the longer term effects of OPN. Alter- epithelial cells. Prosurvival roles for OPN have been described in natively, different OPN isoforms, as well as even stronger effects a subset of cancer cell studies.41 However, there has been consid- of other existing signals involving OPN, may be important in erable debate as to whether OPN plays a proinflammatory or determining whether OPN promotes survival or apoptosis.

www.ibdjournal.org | 1959 Tang et al Inflamm Bowel Dis Volume 20, Number 11, November 2014 Downloaded from https://academic.oup.com/ibdjournal/article/20/11/1950/4578913 by guest on 23 September 2021

FIGURE 6. IRF1 regulated the expression of OPN, and OPN regulated the TNF-a induced changes in p-AKT, p-P38, and p-ERK in intestinal epithelial cells. A, Western blotting showed that the overexpression of IRF1 inhibited OPN expression in NCM460 cells. B, The luciferase activity assay confirmed that TNF-a and IRF1 decreased OPN promoter activity. Luciferase reporter assays assessing OPN promoter activity were performed after the cotransfection of NCM460 cells with constructs derived from the OPN promoter region and pcDNA3.1-IRF1 for 24 hours or exposure to TNF-a for 24 hours. The results are shown relative to the luciferase activity. Independent experiments were repeated 3 times. *P , 0.05 versus control. **P , 0.01 versus control. C, A chromatin immunoprecipitation assay was performed to show the in vivo binding of IRF1 to the OPN promoter region. An anti-IRF1 antibody or normal immunoglobulin G was used to precipitate sonicated chromatin from NCM460 cells. The PCR products were resolved on a 1.5% agarose gel; DNA from the lysates before immunoprecipitation was used as a positive input control. D, A ChIP assay further determined that TNF-a increased the binding of IRF1 to the promoter of OPN. E, NCM460 cells were transfected with OPN siRNA or an OPN-expressing vector for 24 hours and then treated with TNF-a for 12 hours. Our data indicated that TNF-a inhibited the OPN-induced changes in the p-AKT, p-P38, and p-P44/42 signals. Additionally, TNF-a facilitated OPN knockdown, resulting in decreased p-AKT, p-P38, and p-P44/42 signals. F, Schematic representation of a new possible mechanism of TNF-a-induced signals in intestinal epithelial cells.

A previous study found that IRF1 is involved in IL-18– Several publications have shown that the plasma or mucosal inducedOPNexpressionincardiacfibroblasts.29 However, to OPN or IRF1 levels are correlated with the clinical disease activity the best of our knowledge, modulation of the OPN level by in CD or UC patients.28,43 However, studies assessing the correla- IRF1 has not been reported in either normal epithelial cells or tion between endoscopic disease activity and mucosal OPN/IRF1 epithelial cells in IBD. In this study, the TNF-a–stimulated have not been reported. We applied multiple endoscopic score IRF1 expression and decreased OPN expression in the IECs in systems to evaluate the endoscopic activity of patients with IBD. cell culture, increased IRF1 levels, and decreased OPN expres- All the scores correlated well with each other and with the Crohn’s sion were also observed in the human IBD intestinal epithelium. disease activity index or Mayo indices. We demonstrated that OPN The luciferase activity and ChIP assays confirmed that IRF1 was strongly correlated with the endoscopic grade of colonic epi- transcriptionally inhibited the OPN expression in the IECs. thelial degradation, which suggested that OPN might be a biomarker OPN expression was associated with the p-P38, p-ERK, and to assess the epithelial cell status in active IBD. p-Akt levels. Our data demonstrated that TNF-a further regu- In conclusion, our study provides a novel mechanism of lated the p-P38, p-ERK, and p-Akt levels while reducing OPN TNF-a–induced damage to IECs. The results presented here expression. identify IRF1 as a TNF-inducible transcript factor and position

1960 | www.ibdjournal.org Inflamm Bowel Dis Volume 20, Number 11, November 2014 Opposite Effects of IRF 1 and Osteopontin

its downstream effector OPN as a protector against apoptosis in 20. Liaw L, Birk DE, Ballas CB, et al. Altered wound healing in mice lacking – response to injury and inflammation. TNF-a decreased the a functional osteopontin gene (spp1). J Clin Invest. 1998;101:1468 1478. 21. Lin YH, Yang-Yen HF. The osteopontin-CD44 survival signal involves expression of OPN; thus, it was able to facilitate damage to activation of the phosphatidylinositol 3-kinase/Akt signaling pathway. the IECs by elevating the IRF1 level (Fig. 6F). Increasing our J Biol Chem. 2001;276:46024–46030. understanding of the TNF-a signaling axis and the biological 22. Miyazaki Y, Tashiro T, Higuchi Y, et al. Expression of osteopontin in a cell line and in transgenic mice with pulmonary fibrosis functions of OPN in the context of endoscopic recurrent CD or resulting from the lung expression of a tumor necrosis factor-alpha trans- UC may aid in the discovery of new targets for IBD treatment in gene. Ann N Y Acad Sci. 1995;760:334–341. the future. 23. O’Regan AW, Hayden JM, Body S, et al. Abnormal pulmonary granu- loma formation in osteopontin-deficient mice. Am J Respir Crit Care Med. 2001;164:2243–2247.

24. Yumoto K, Ishijima M, Rittling SR, et al. Osteopontin deficiency protects Downloaded from https://academic.oup.com/ibdjournal/article/20/11/1950/4578913 by guest on 23 September 2021 ACKNOWLEDGMENTS joints against destruction in anti-type II collagen antibody-induced arthri- Author contributions: R. Tang and G. Yang contributed tis in mice. Proc Natl Acad Sci U S A. 2002;99:4556–4561. equally to this work. 25. Morales-Ibanez O, Dominguez M, Ki SH, et al. Human and experimental evidence supporting a role for osteopontin in alcoholic . Hepatol- ogy. 2013;58:1742–1756. 26. Chen F, Liu H, Shen Q, et al. Osteopontin: participation in inflammation REFERENCES or mucosal protection in inflammatory bowel diseases? Dig Dis Sci. 2013; 1. Xavier RJ, Podolsky DK. Unravelling the pathogenesis of inflammatory 58:1569–1580. bowel disease. Nature. 2007;448:427–434. 27. da Silva AP, Ellen RP, Sorensen ES, et al. Osteopontin attenuation of 2. Khor B, Gardet A, Xavier RJ. Genetics and pathogenesis of inflammatory dextran sulfate sodium-induced colitis in mice. Lab Invest. 2009;89: bowel disease. Nature. 2011;474:307–317. 1169–1181. 3. Wang J, Fu YX. Tumor necrosis factor family members and inflammatory 28. Sato T, Nakai T, Tamura N, et al. Osteopontin/Eta-1 upregulated in bowel disease. Immunol Rev. 2005;204:144–155. Crohn’s disease regulates the Th1 immune response. Gut. 2005;54: 4. Frey MR, Edelblum KL, Mullane MT, et al. The ErbB4 1254–1262. receptor is required for colon epithelial cell survival in the presence of 29. Yu Q, Vazquez R, Khojeini EV, et al. IL-18 induction of osteopontin TNF. Gastroenterology. 2009;136:217–226. mediates cardiac fibrosis and diastolic dysfunction in mice. Am J Physiol 5. Hilliard VC, Frey MR, Dempsey PJ, et al. TNF-alpha converting enzyme- Heart Circ Physiol. 2009;297:H76–H85. mediated ErbB4 transactivation by TNF promotes colonic epithelial cell ’ survival. Am J Physiol Gastrointest Liver Physiol. 2011;301:G338–G346. 30. Best WR, Becktel JM, Singleton JW, et al. Development of a Crohn s disease activity index. National Cooperative Crohn’s Disease Study. Gas- 6. Avdagic N, Babic N, Seremet M, et al. Tumor necrosis factor-alpha serum – level in assessment of disease activity in inflammatory bowel diseases. troenterology. 1976;70:439 444. Med Glas (Zenica). 2013;10:211–216. 31. Schroeder KW, Tremaine WJ, Ilstrup DM. Coated oral 5-aminosalicylic fl acid therapy for mildly to moderately active ulcerative colitis. A random- 7. Murata Y, Ishiguro Y, Itoh J, et al. The role of proin ammatory and – immunoregulatory cytokines in the pathogenesis of ulcerative colitis. ized study. N Engl J Med. 1987;317:1625 1629. J Gastroenterol. 1995;30(suppl 8):56–60. 32. Mary JY, Modigliani R. Development and validation of an endoscopic ’ 8. Smith K. IBD: early use of anti-TNF agents might give a helping hand to index of the severity for Crohn s disease: a prospective multicentre study. ’ fl paediatric patients with Crohn’s disease. Nat Rev Gastroenterol Hepatol. Groupe d Etudes Therapeutiques des Affections In ammatoires du Tube – 2013;10:691. Digestif (GETAID). Gut. 1989;30:983 989. ’ 9. Ben-Horin S, Chowers Y. Tailoring anti-TNF therapy in IBD: drug levels 33. Daperno M, D Haens G, Van Assche G, et al. Development and valida- fi ’ and disease activity. Nat Rev Gastroenterol Hepatol. 2014;11:243–255. tion of a new, simpli ed endoscopic activity score for Crohn s disease: the – 10. Taniguchi T, Ogasawara K, Takaoka A, et al. IRF family of transcription SES-CD. Gastrointest Endosc. 2004;60:505 512. factors as regulators of defense. Annu Rev Immunol. 2001;19:623–655. 34. Baron JH, Connell AM, Lennard-Jones JE. Variation between observers in – 11. Chen F, Song J, Di J, et al. IRF1 suppresses Ki-67 promoter activity describing mucosal appearances in proctocolitis. Br Med J. 1964;1:89 92. through interfering with Sp1 activation. Tumour Biol. 2012;33: 35. Feagan BG, Greenberg GR, Wild G, et al. Treatment of ulcerative colitis 2217–2225. with a humanized antibody to the alpha4beta7 integrin. N Engl J Med. 12. Yarilina A, Park-Min KH, Antoniv T, et al. TNF activates an IRF1- 2005;352:2499–2507. dependent autocrine loop leading to sustained expression of 36. Rachmilewitz D. Coated mesalazine (5-aminosalicylic acid) versus sul- and STAT1-dependent type I interferon-response genes. Nat Immunol. phasalazine in the treatment of active ulcerative colitis: a randomised trial. 2008;9:378–387. BMJ. 1989;298:82–86. 13. Venkatesh D, Ernandez T, Rosetti F, et al. Endothelial TNF receptor 2 37. Travis SP, Schnell D, Krzeski P, et al. Developing an instrument to assess induces IRF1 transcription factor-dependent interferon-beta autocrine sig- the endoscopic severity of ulcerative colitis: the Ulcerative Colitis Endo- naling to promote recruitment. Immunity. 2013;38:1025–1037. scopic Index of Severity (UCEIS). Gut. 2012;61:535–542. 14. Tada Y, Ho A, Matsuyama T, et al. Reduced incidence and severity of 38. Samuel S, Bruining DH, Loftus EV Jr, et al. Validation of the ulcerative antigen-induced autoimmune diseases in mice lacking interferon regula- colitis colonoscopic index of severity and its correlation with disease tory factor-1. J Exp Med. 1997;185:231–238. activity measures. Clin Gastroenterol Hepatol. 2013;11:49–54. e1. 15. Mannick EE, Cote RL, Schurr JR, et al. Altered phenotype of dextran 39. Wirtz S, Neufert C, Weigmann B, et al. Chemically induced mouse mod- sulfate sodium colitis in interferon regulatory factor-1 knock-out mice. els of intestinal inflammation. Nat Protoc. 2007;2:541–546. J Gastroenterol Hepatol. 2005;20:371–380. 40. Cooper HS, Murthy SN, Shah RS, et al. Clinicopathologic study of dex- 16. Mudter J, Yu J, Zufferey C, et al. IRF4 regulates IL-17A promoter activity tran sulfate sodium experimental murine colitis. Lab Invest. 1993;69: and controls RORgammat-dependent Th17 colitis in vivo. Inflamm Bowel 238–249. Dis. 2011;17:1343–1358. 41. Agrawal D, Chen T, Irby R, et al. Osteopontin identified as lead marker of 17. Huff CD, Witherspoon DJ, Zhang Y, et al. Crohn’s disease and genetic colon cancer progression, using pooled sample expression profiling. J Natl hitchhiking at IBD5. Mol Biol Evol. 2012;29:101–111. Cancer Inst. 2002;94:513–521. 18. Senger DR, Asch BB, Smith BD, et al. A secreted phosphoprotein marker 42. Standal T, Borset M, Sundan A. Role of osteopontin in adhesion, for neoplastic transformation of both epithelial and fibroblastic cells. migration, cell survival and bone remodeling. Exp Oncol. 2004;26: Nature. 1983;302:714–715. 179–184. 19. Denhardt DT, Noda M, O’Regan AW, et al. Osteopontin as a means to 43. Clavell M, Correa-Gracian H, Liu Z, et al. Detection of interferon regu- cope with environmental insults: regulation of inflammation, tissue re- latory factor-1 in lamina propria mononuclear cells in Crohn’s disease. modeling, and cell survival. J Clin Invest. 2001;107:1055–1061. J Pediatr Gastroenterol Nutr. 2000;30:43–47.

www.ibdjournal.org | 1961