Igf2bp1 Is Required for Full Induction of Ptgs2 mRNA in Colonic Mesenchymal Stem Cells in Mice

GASTROENTEROLOGY 2012;143:110 –121

Igf2bp1 Is Required for Full Induction of Ptgs2 mRNA in Colonic Mesenchymal Stem Cells in Mice NICHOLAS A. MANIERI, MONICA R. DRYLEWICZ, HIROYUKI MIYOSHI, and THADDEUS S. STAPPENBECK Department of Pathology and Immunology, Washington University in St. Louis School of Medicine, St. Louis, Missouri

See editorial on page 19.

BASIC AND TRANSLATIONAL AT

BACKGROUND & AIMS: Prostaglandin-endoperoxide synthase (Ptgs)2 is an enzyme involved in prostaglandin production during the response to mucosal damage. Its expression is regulated, in part, by messenger RNA (mRNA)-binding proteins that control the stability of Ptgs2 mRNA. We used a precise system of colonic injury and repair to identify Ptgs2 mRNA-binding proteins. METHODS: We used endoscopy-guided mucosal excision to create focal injury sites in colons of mice. Wound beds from wild-type, Ptgs2⫺/⫺, Ptgs2⫹/-, and Myd88⫺/⫺ mice were analyzed at 2-day intervals after injury for aspects of repair and Ptgs2 expression. We used cultured colonic mesenchymal stem cells (cMSCs) that express Ptgs2 to identify and analyze molecules that regulate Ptgs2 expression. RESULTS: Ptgs2⫺/⫺ mice had defects in wound repair, validating the biopsy technique as a system to study the regulation of Ptgs2. Ptgs2⫹/- mice had similar defects in wound healing, so full induction of Ptgs2 is required for wound repair. In wild-type mice, levels of Ptgs2 mRNA increased significantly in the wound bed 2 and 4 days after injury; the highest levels of Ptgs2 were observed in cMSCs. In a functional short hairpin RNA knockdown screen, we identified Igf2bp1, a VICKZ (Vg1 RNA binding protein, Insulin-like growth factor II mRNA binding protein 1, Coding region determinant-binding protein, KH domain containing protein overexpressed in cancer, and Zipcode-binding protein-1) mRNA-binding protein, as a regulator of Ptgs2 expression in cMSCs. Igf2bp1 also interacted physically with Ptgs2 mRNA. Igf2bp1 expression was induced exclusively in wound-bed cMSCs, and full induction of Ptgs2 and Igf2bp1 during repair required Myd88. CONCLUSIONS: We identified Igf2bp1 as a regulator of Ptgs2 mRNA in mice. Igf2bp1 is required for full induction of Ptgs2 mRNA in cMSCs. Keywords: Cox-2; Imp-1; ZBP1; Colorectal Cancer; Inflammatory Bowel Disease.

P

rostaglandin-endoperoxide synthase 2 (Ptgs2) is an important enzyme for localized prostaglandin production during the response to mucosal damage and the pathogenesis of colorectal cancer.1 Ptgs2 and Ptgs1 are isoforms that are the 2 rate-limiting enzymes in the synthesis of all prostaglandins, which affect downstream re-

sponses such as angiogenesis, proliferation, and apoptosis.1 Ptgs1 is widely expressed and constitutive, whereas Ptgs2 is focally inducible with limited constitutive expression (primarily colon and testes).2 During inflammation, damage signals and cytokines signal through Myd88 to activate nuclear factor-␬B and induce Ptgs2 transcription.3 Specialized mRNA binding/stabilizing proteins bind the AU-rich elements (ARE) in the 3’-untranslated region of Ptgs2 messenger RNA (mRNA) and prevent its degradation.4 Additional, unknown mRNA-binding proteins likely regulate Ptgs2 expression because mutations in the 3’-untranslated region outside of the ARE regions also affect Ptgs2 stability.5 To study the regulation and effects of Ptgs2 in vivo, we and others have used various mouse models of intestinal injury including dextran sodium sulfate,6 –9 2,4,6-trinitrobenzene sulfonic acid,10 and irradiation.11 Here, Ptgs2 affects epithelial proliferation and resolution of inflammation. Despite the requirement for Ptgs2 during repair, understanding the mechanisms that control Ptgs2 expression has been challenging because of the variable timing and location of injuries in these models. To discover novel Ptgs2 regulatory proteins expressed after mucosal damage, we created small, focal injury sites in the mouse colon using endoscopy-guided forceps.12–14 The precise timing and location of injuries in this system allowed us to investigate the cellular source, target, and timing of specific genes that regulate Ptgs2 expression. We previously showed that repair after biopsy injury proceeds reproducibly in defined stages.12 In the first phase, neutrophils are recruited to the wound and an immature protective epithelial barrier (wound associated epithelium [WAE]) forms over the nascent wound bed. This WAE layer consists of a single layer of postmitotic cells that emerge from crypts that are adjacent to the wound and migrate toward the center of the wound bed.12 The second phase of healing involves increased epithelial Abbreviations used in this paper: ␣-SMA, ␣-smooth muscle actin; ARE, AU-rich elements; cDNA, complementary DNA; cMSC, colonic mesenchymal stem cell; GFP, green fluorescent protein; Igf2bp1, insulin-like growth factor 2 binding protein 1; mRNA, messenger RNA; OCT, optimal cutting temperature compound; PBS, phosphate-buffered saline; PG, prostaglandin; Ptgs2, prostaglandin-endoperoxide synthase 2; qRT-PCR, quantitative reverse-transcription polymerase chain reaction; shRNA, short hairpin RNA; WAE, wound-associated epithelial cell; WT, wild-type; ZO-1, zonula occludens 1. © 2012 by the AGA Institute 0016-5085/$36.00 http://dx.doi.org/10.1053/j.gastro.2012.03.037

proliferation in the adjacent crypts as well as expansion of the wound bed mesenchyme by the influx and/or proliferation of macrophages and additional stromal cells.12 In the final stage, the wound bed is replaced by new crypts. In this study, we were able to use this precise injury system to identify a novel Ptgs2 mRNA-binding protein, insulin-like growth factor 2 binding protein 1 (Igf2bp1). We found that maximal Ptgs2 expression was required for proper mucosal healing, and that the highest levels of Ptgs2 expression were localized to colonic mesenchymal stem cells (cMSCs). By using a short hairpin RNA (shRNA) knockdown screen in cultured cMSCs, we showed that Igf2bp1, a VICKZ mRNA binding protein, was necessary for maximal expression of Ptgs2 mRNA. Igf2bp1 interacted with Ptgs2 mRNA and expression of both genes was induced in wound bed cMSCs in a Myd88dependent manner. These data show that Igf2bp1 was necessary for the maximal Ptgs2 expression in cMSCs that was required for proper colonic injury repair.

Materials and Methods Mice Animal experiments were performed in accordance with approved protocols from the Washington University School of Medicine Animal Studies Committee. Ptgs2⫹/- mice15 were bred to generate littermate wild-type (WT), Ptgs2⫹/⫺ and Ptgs2⫺/⫺ progeny for experiments. Myd88⫹/⫺ mice16 from Jackson Laboratories (Bar Harbor, Maine) were bred to generate Myd88⫺/⫺ and Myd88⫹/⫺ littermate controls. All mice were on a C57Bl/6 background.

Colonic Biopsy We used a high-resolution miniaturized colonoscope system to visualize the lumen of the colon and discretely injure the mucosal layer in 10- to 16-week-old anesthetized mice. After inflating the colon with phosphate-buffered saline (PBS), we inserted 3F flexible biopsy forceps into the sheath adjacent to the camera. We removed 3–5 full-thickness areas of the entire mucosa and submucosa that were distributed along the dorsal side of the colon. All genetically modified mice were injured with the same technique. For this study, we evaluated wounds that averaged approximately 1 mm2, which is equivalent to removal of approximately 250 –300 crypts.

Colonic Tissue Preparation Wounded mice were killed 2– 6 days after injury and each wound was frozen individually in optimal cutting temperature compound (OCT) to make frozen sections (see the Supplementary Materials and Methods section for detailed descriptions).

In Situ Hybridization A complementary DNA (cDNA) clone for Ptgs2 (clone ID: 30059181) was purchased from Thermo Fisher Scientific (Waltham, MA) to create digoxigenin-labeled antisense RNA probes (see the Supplementary Materials and Methods section for detailed descriptions).

IGF2BP1 REGULATES PTGS2 MRNA IN CMSCS

111

Immunofluorescence Staining was performed as previously published17 (see the Supplementary Materials and Methods section for detailed descriptions).

RNA Isolation and Quantitative ReverseTranscription Polymerase Chain Reaction For wound RNA isolation, the wound bed mucosa or adjacent uninjured mucosa was dissected with forceps under the guidance of a whole-mount microscope, leaving the underlying muscularis propria intact. RNA was isolated using a NucleoSpin RNA II kit (Machery-Nagel, Duren, Germany). cDNAs were synthesized using Superscript III reverse transcriptase (Invitrogen, Carlsbad, CA). Quantitative reverse-transcription polymerase chain reaction (RT-PCR) reactions using SYBR-green master mix (Clontech Laboratories, Mountain View, CA) were analyzed on an Eppendorf (Hauppauge, NY) realplex Mastercycler. RNA from wounds and uninjured mucosa were normalized to glyceraldehyde-3-phosphate dehydrogenase. The Supplementary Materials and Methods section contains the primer sequences.

Immunoprecipitation and RNA Isolation in NIH3T3 Cells The open reading frame of Igf2bp1 was cloned into p3XFLAG-CMV-14 (Sigma-Aldrich, St. Louis, MO) (see the Supplementary Materials and Methods section for detailed descriptions).

shRNA Transfection and Immunoblotting HEK293T cells were transfected with Mission shRNA constructs specific for Igf2bp1 or nontargeting controls (SigmaAldrich, St Louis, MO). Lentivirus obtained from transfected cells was used to infect cMSCs for knockdown of specific genes. The Supplementary Materials and Methods section contains detailed descriptions.

Cell Isolation and Culture Conditions Colonic mesenchymal stem cell isolation and culture was performed as previously reported17 (see the Supplementary Materials and Methods section).

Statistics All statistical analyses were performed using GraphPad Prism 3.0 (GraphPad Software, La Jolla, CA). Significant differences were evaluated by the Student t test, analysis of variance (ANOVA), followed by the Tukey multiple comparison test or the Fisher exact test, as noted.

Results Two Functional Ptgs2 Alleles Are Required for Colonic Mucosal Wound Repair To determine if Ptgs2 expression was required for colonic mucosal repair of biopsy injuries, we assessed the gross and microscopic features of wound beds from Ptgs2⫺/⫺ and control WT littermate mice at 2-day intervals after injury. Starting at day 6 post-injury (second stage of repair), Ptgs2⫺/⫺ mice displayed severe defects that included loss of muscularis propria and an incomplete WAE barrier (Figure 1A–E). Specifically, the muscularis propria showed complete loss of ␣-smooth muscle actin (␣-SMA)

BASIC AND TRANSLATIONAL AT

July 2012

112

MANIERI ET AL

GASTROENTEROLOGY Vol. 143, No. 1

BASIC AND TRANSLATIONAL AT

Figure 1. Two functional alleles of Ptgs2 are required for proper mucosal healing. (A) Colonic sections from WT, (B) Ptgs2⫹/-, and (C) Ptgs2⫺/⫺ mice 6 days post-injury were stained with antibodies against ␣-SMA (red), ␤-catenin (green), and bis-benzamide (nuclei, blue). Dotted yellow lines, muscularis propria; white dashed lines, adjacent crypts (AC); yellow arrowheads, epithelial barrier interruptions. Scale bars: 100 ␮m. (D) Graph of the percentage loss of ␣-SMA underlying day 6 wound beds (gap length/wound bed length) that included data points (n ⱖ 6/genotype), median bar, and significance by one-way ANOVA and post hoc Tukey test: F2,18 ⫽ 10.33. **P ⬍ .01. (E) Graph of the percentage of wound bed sections with incomplete epithelial coverage at day 6 post-injury. Wound bed sections were scored as either fully or incompletely covered by epithelial cells (␤catenin–positive). The Fisher exact test for the resulting data contingency table compared incomplete restitution in WT (2/14) with Ptgs2⫹/- (14/17) (**P ⬍ .001), and WT with Ptgs2⫺/⫺ (7/9) (**P ⬍ .001).

IGF2BP1 REGULATES PTGS2 MRNA IN CMSCS

113

BASIC AND TRANSLATIONAL AT

July 2012

Figure 2. Ptgs2 is induced in colonic wounds in a spatial and temporal manner. (A) Graph of Ptgs2 mRNA expression (qRT-PCR) in wounds at days 2, 4, and 6 post-injury (baseline, uninjured mucosa; n ⱖ 5 wounds/day). Data are displayed as means ⫾ standard error of the mean and significance by one-way ANOVA and post hoc Tukey test: F2,18 ⫽ 5.33. *P ⬍ .05. (B) Colonic sections from WT mice at days 2, 4, and 6 post-injury stained by in situ hybridization for Ptgs2. Solid black boxes, insets (right panels). Arrows indicate additional scattered Ptgs2 high-expressing cells outside of the boxed area at day 2. The adjacent crypts (AC) and muscularis propria were outlined in red dashed lines. Scale bars: 200 ␮m (inset scale bars: 50 ␮m). (C) Graph of the number of Ptgs2 high-expressing cells per wound section. Data are displayed as means ⫾ standard error of the mean and significance was determined by one-way ANOVA and post hoc Tukey test: F2,9 ⫽ 11.6. *P ⬍ .05. (D) Colonic sections from WT mice at days 2, 4, and 6 post-injury stained with anti-Ptgs2 antisera (green) and bis-benzamide (blue). White boxes, insets (right panels). Yellow dashed lines mark adjacent crypts (AC) and muscularis propria. Scale bars, 100 ␮m (inset scale bars: 25 ␮m). (E) Cartoon depicting the typical positional and temporal relationship of Ptgs2 high-expressing cells in the wound bed post-injury.

underlying wound beds of Ptgs2⫺/⫺ but not WT mice (Figure 1A–D). The damaged muscularis propria was replaced by granulation tissue comprising endothelial cells and myeloid cells that extended through the serosa (Supplemental Figure 1A–F). This phenotype resembled the histology of perforated human colonic ulcers that can occur after prolonged nonsteroidal anti-inflammatory drug use.18 Importantly, the ␣-SMA staining pattern was intact in the muscularis propria at day 2 post-injury in Ptgs2⫺/⫺ mice, indicating that this phenotype was a delayed biological response to injury and was not caused by a more fragile mucosa in the Ptgs2⫺/⫺ mice (Supplementary Figure 1G–I).

The epithelial barrier at day 6 post-injury also was defective in Ptgs2⫺/⫺ mice because WAE cells incompletely covered these wounds. WAE cells are claudin-4 –positive, zonula occludens 1 (ZO-1)–negative, post-mitotic epithelial cells that emerge from adjacent crypts and completely migrate across the wound bed during the first stage of repair.12 In contrast to WT mice, most Ptgs2⫺/⫺ wounds were covered only partially by an epithelial barrier at day 6 post-injury as shown by gross morphologic examination (Supplementary Figure 2) and by immunofluorescence microscopy (Figure 1A, C, and E). In addition, the epithelial cells that partially covered Ptgs2⫺/⫺ wounds were still immature WAE cells, whereas WT wounds were covered

114

MANIERI ET AL

GASTROENTEROLOGY Vol. 143, No. 1

BASIC AND TRANSLATIONAL AT

by a complete layer of mature surface epithelial cells (Supplementary Figure 3A–D). There were also defects in the final stage of healing in Ptgs2⫺/⫺ wounds because they contained fewer regenerated crypts at 14 days post-injury (Supplementary Figure 4A–D). In summary, the complex repair defects in Ptgs2⫺/⫺ mice were observed in the latter phases of repair. Interestingly, by these parameters, littermate Ptgs2⫹/mice also showed defective wound repair (Figure 1B). At day 6 post-injury, Ptgs2⫹/- mice showed loss of the muscularis propria underlying the wound bed (Figure 1D) and an incomplete epithelial layer (Figure 1E). These data showed that loss of just one functional Ptgs2 allele had severe effects on colonic mesenchymal and epithelial regeneration. Because this injury system was sensitive to a 2-fold decrease of gene dosage, we proposed that it would be useful to discover novel proteins that regulate maximal Ptgs2 expression.

Ptgs2 Expression Is Increased Transiently and Focally During Colonic Wound Repair Given the profound wound repair defects in Ptgs2⫹/- mice, we next evaluated the timing and location of Ptgs2 expression in WT colonic wounds during 2-day intervals after injury. We first quantified Ptgs2 mRNA levels by qRT-PCR using mRNAs isolated from the mucosa of biopsy wounds. Compared with uninjured colonic mucosa, Ptgs2 levels peaked during the first stage of repair (an average of 1640-fold higher at day 2 and 2170fold higher at day 4), and subsequently were down-regulated during the second stage of repair (only a 93-fold increase at day 6) (Figure 2A). In contrast, Ptgs1 mRNA levels did not differ between wounds and adjacent mucosa (Supplementary Figure 5). Interestingly, Ptgs2 mRNA expression also was significantly lower in the wound beds of Ptgs2⫹/- mice compared with WT mice at day 4 post-injury (Supplementary Figure 6). This intermediate expression of Ptgs2 in Ptgs2⫹/- mice was not sufficient for proper wound repair (Figure 1D and E), supporting the conclusion that maximal induction of Ptgs2 is a critical component of mucosal repair. We then performed in situ hybridization for Ptgs2 in WT mice at 2-day intervals after injury. Because we wanted to identify cells that expressed the highest levels of Ptgs2, we limited the exposure time of the detection reagent. Throughout injury, cells expressing high levels of Ptgs2 were detected only within the mesenchyme of the wound bed and were not observed in either the underlying muscularis propria or crypts adjacent to the wound bed (Figure 2B). The Ptgs2 high-expressing cells were scattered in the wound bed at day 2 post-injury and were clustered beneath WAE cells at days 4 and 6 post-injury (Figure 2B).

IGF2BP1 REGULATES PTGS2 MRNA IN CMSCS

115

At day 6 post-injury, the number of Ptgs2 high-expressing cells was greatly diminished (Figure 2B). Thus, the cells with the highest levels of Ptgs2 mRNA in the upper wound during the first phase of repair were not detected during the second stage of repair. We corroborated the in situ findings by immunofluorescence localization of Ptgs2 protein in wound bed sections. We used dilutions of primary antibody (1:5000) that did not detect constitutive Ptgs2 expression in cMSCs located outside of the wound bed or other cell types known to express lower levels of Ptgs2.7 In WT wound beds, we found a similar spatial and temporal pattern of Ptgs2 protein induction (Figure 2A). The number of Ptgs2 high-expressing cells in the wound bed was significantly higher at days 2 and 4 post-injury as compared with day 6 post-injury (Figure 2C). As with in situ hybridization, Ptgs2 high-expressing cells detected by immunofluorescence were present only in the wound bed mesenchyme (Figure 2D). As additional controls, we performed in parallel in situ hybridization and immunofluorescence studies for localization of Ptgs2 mRNA and protein within wounds of Ptgs2⫺/⫺ mice. We found no detectable signal in Ptgs2⫺/⫺ mice for either assay (Supplementary Figure 7). Thus, Ptgs2 was maximally induced in upper wound bed mesenchymal cells during the first stage of repair (Figure 2E).

Ptgs2 High-Expressing Cells in the Wound Bed Are cMSCs To identify novel Ptgs2 regulators during colonic repair, we first defined the mesenchymal cell lineages that produced the highest levels of Ptgs2. Candidates included myeloid cells,19 endothelial cells,20 colonic mesenchymal stem cells,7 and myofibroblasts.21 Multilabel immunofluorescence imaging of Ptgs2 high-expressing cells and various lineage markers at day 4 post-injury (Figure 3) showed that these cells were neither endothelial nor hematopoietic lineages because they did not co-label with either CD31 or CD45, respectively (Figure 3A). Instead, Ptgs2 high-expressing cells showed dual labeling for multiple cMSC markers17 including CD44, CD29, and CD54 (Figure 3B). Additional studies at days 2 and 6 post-injury showed that the Ptgs2 high-expressing cells showed dual labeling with these same markers (data not shown). Because myofibroblasts and cMSCs share many of these markers, we evaluated definitive myofibroblast markers. We did not detect ␣-SMA cells in cells that expressed high levels of Ptgs2 (Figure 3A). In addition, interferon-␥ is expressed in wounds,12 which can down-regulate ␣-SMA22 and increase major histocompatibility complex class II on myofibroblasts.23 Therefore, we showed that Ptgs2 highexpressing cells also were major histocompatibility com-

4™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™ Figure 3. Ptgs2 high-expressing cells in colonic wound beds are cMSCs. (A) Sections of WT wounds at day 4 post-injury labeled with anti-Ptgs2 (red) and lineage markers (green) for endothelial cells (anti-CD31), hematopoietic cells (anti-CD45), and mesenchymal cells (␣-SMA and major histocompatibility complex [MHC] class II). (B) WT day 4 wounds labeled with anti-Ptgs2 (red) and cMSC markers anti-CD44, anti-CD29, and anti-CD54. Yellow boxes, insets (right). White dashed lines, WAE basal surface. Scale bars, 50 ␮m (inset scale bars: 10 ␮m).

BASIC AND TRANSLATIONAL AT

July 2012

116

MANIERI ET AL

GASTROENTEROLOGY Vol. 143, No. 1

Figure 4. Igf2bp1 is required for maximal Ptgs2 expression in cMSCs. (A) Graph depicting the relative mRNA expression (means ⫾ standard error of the mean) of candidate Ptgs2 regulators after shRNA knockdown (two most efficient for each gene) vs control shRNA cells. (B) Graph depicting the relative expression of Ptgs2 mRNA in cMSCs that contain knockdown of candidate mRNA regulators (n ⫽ 4). Data are displayed as means ⫾ standard error of the mean and significance was determined by one-way ANOVA and the post hoc Tukey test: F4,15 ⫽ 20.25. **P ⬍ .01, ***P ⬍ .001.

BASIC AND TRANSLATIONAL AT

plex class II–negative in the wound bed to exclude resting or activated myofibroblasts as Ptgs2 high-expressing cells (Figure 3A). We hypothesized that cMSCs appeared in wound beds in part by migration from adjacent uninjured tissue. We injected green fluorescent protein (GFP)-expressing cultured cMSCs into the mucosa adjacent to wounds. We found that cMSCs migrated into wounds, indicating these cells have the potential to migrate from the adjacent mucosa (Supplementary Figure 8). As a control for our isolation technique, we showed that Fstl1, a marker that distinguishes myofibroblasts from intestinal MSCs,24 was expressed in cultured cMSCs but not MIC 216 cells (Supplementary Figure 9). To show that the defects seen in Ptgs2⫺/⫺ mice were owing to loss of prostaglandin synthesis, we injected stable analogues of prostaglandin (PG)E2 and PGI2 into Ptgs2⫺/⫺ mice twice daily. This procedure rescued the muscle degradation phenotype at day 6 post-injury (Supplementary Figure 10). We chose these 2 prostaglandins because cMSCs express detectable levels of prostaglandin synthases for PGE2 and PGI2 but not for PGD2 or PGF2␣.17 Because the Ptgs2 high-expressing cells in wounds were consistent with cMCSs, we further investigated cultured cMSCs for novel regulators of Ptgs2.

Igf2bp1 Interacts With Ptgs2 mRNA Our previous studies in cMSCs showed that mRNAbinding proteins including CUG triplet repeat RNA-binding protein 2 (CUGbp2) play a major role in the post-transcriptional regulation of Ptgs2 expression through mRNA stabilization.17,25 However, the ARE-binding protein CUGbp2 does not account for complete stabilization of Ptgs2 mRNA in cMSCs.17 Furthermore, previous reviews have proposed that additional unknown mRNA-binding proteins will regulate Ptgs2 mRNA in a cooperative fashion.4 We hypothesized that additional transacting protein(s) outside of the ARE-binding family will interact with and

regulate Ptgs2 mRNA in cMSCs. Because cMSCs express 10-fold higher levels of Ptgs2 than bone marrow MSCs,17 we analyzed microarray data for these 2 cell types17 to identify candidate mRNA-binding proteins that preferentially were expressed in cMSCs. This screen identified known ARE-binding proteins (HuR, CUGbp1, and CUGbp2) as well as candidate Ptgs2 mRNA-binding proteins (Rbm3, Rbms3, Rbpms, and Igf2bp1). We functionally screened the non– ARE-binding protein candidates individually by shRNA-mediated knockdown in cMSCs. We obtained multiple shRNAs for each gene, and the 2 shRNAs that had the greatest knockdown of the target gene (at least 80% knockdown efficiency) were used in further experiments (Figure 4A). Of all these candidates, we found that Igf2bp1-deficient cMSCs expressed the lowest levels of Ptgs2 mRNA (Figure 4B). Igf2bp1 was of further interest because it is an mRNAbinding protein26 that had not been shown previously to interact with Ptgs2 mRNA. We next tested if Igf2bp1 interacted with Ptgs2 mRNA. We immunoprecipitated FLAG-tagged Igf2bp1 and eluted Igf2bp1 from the pellet fraction with 3X FLAG peptide. We then immunoblotted total protein, eluate, and pellet fractions to confirm immunoprecipitation of correct size proteins (Figure 5A). Protein was degraded in each fraction and the associated mRNAs were quantified by quantitative RT-PCR (qRT-PCR) (Figure 5B). Ptgs2 mRNA was enriched in the eluate as compared with the total fraction, indicating immunoprecipitation of Igf2bp1 with Ptgs2 mRNA (Figure 5B). As a control, 18S RNA was not enriched in the eluate. These data show the interaction of Igf2bp1 protein and Ptgs2 mRNA. We then designed 2 FLAG-tagged truncation mutants of Igf2bp1, one containing the 2 RNA-recognition motifs (RRM) in the amino terminus (RRM-FLAG) and a second containing the 4 K Homology (KH) RNA-binding motifs in the carboxy terminus (KH-FLAG) (Figure 5C). We observed that KH-FLAG similarly enriched for Ptgs2 mRNA

IGF2BP1 REGULATES PTGS2 MRNA IN CMSCS

117

BASIC AND TRANSLATIONAL AT

July 2012

Figure 5. Igf2bp1 interacts with Ptgs2 mRNA. (A) Immunoblot of mock-transfected (control) and Igf2bp1-FLAG–transfected cells for anti-FLAG– horseradish peroxidase. T, total fraction before immunoprecipitation; E, elution fraction post-3X-FLAG incubation; P, pellet fraction post-elution. (B) Graph of the relative RNA fractions from transfected vs mock-transfected cells. Data are displayed as means ⫾ standard error of the mean with significance by one-way ANOVA and the post hoc Tukey test: F3,16 ⫽ 14.04. **P ⬍ .01. (C) Graphic representation of the Igf2bp1 gene displaying domains and constructs. Amino acids were numbered at intervals. A C-terminal FLAG tag was included for RRM-FLAG (amino acid 185), KH-FLAG (amino acid 577), and Igf2bp1-FLAG. (D) Immunoblot with anti-FLAG– horseradish peroxidase for control, and cells transfected with Igf2bp1-FLAG and deletion mutants. (E) Graph of the relative expression of RNAs in transfected vs mock-transfected cells. Data are displayed as means ⫾ standard error of the mean with significance by one-way ANOVA and the post hoc Tukey test: F11,28 ⫽ 6.14. *P ⬍ .05, **P ⬍ .01. Immunoblots represent ⬎3 experiments. In the graphs, N ⫽ 3– 4 experiments.

whereas RRM-FLAG did not (Figure 5D and E). These data suggested that the KH domains of Igf2bp1 were required for interaction with Ptgs2 mRNA.

Igf2bp1 Is Induced in Ptgs2 High-Expressing cMSCs During the Early Phase of Repair We next wanted to determine the spatial and temporal expression of Igf2bp1 in wound-associated cMSCs as well as the mechanism that controls its expression. We found that Igf2bp1 mRNA was enriched in the wound mucosa at days 2 and 4 post-injury (but not day 6 postinjury) as compared with uninjured mucosa (Figure 6A). This temporal pattern of Igf2bp1 induction/reduction was similar to Ptgs2 expression during repair. We also

assessed whether Igf2bp1 co-localized with Ptgs2 highexpressing cMSCs by multilabel immunofluorescence and found that 100% of the Igf2bp1 signal co-labeled with Ptgs2 high-expressing cMSCs (Figure 6B). Thus, Igf2bp1 was expressed during wound repair at the appropriate time (first phase of healing) and location (cMSCs) to regulate maximal Ptgs2 expression.

Myd88 Signaling Is Required for Induction of Igf2bp1 in cMSCs The temporal and cellular co-induction of Igf2bp1 and Ptgs2 suggested that these genes may be regulated by the same pathway. One candidate that regulates Ptgs2

118

MANIERI ET AL

GASTROENTEROLOGY Vol. 143, No. 1

BASIC AND TRANSLATIONAL AT

expression in various cell types is Myd88 signaling.3 Therefore, we hypothesized that Myd88 was required for induction of Igf2bp1 and maximal Ptgs2 expression in cMSCs. We first biopsy-injured Myd88⫺/⫺ mice and found a significant repair defect as indicated by loss of ␣-SMA expression in the muscularis propria at day 6 post-injury (Figure 7A). This phenotype was similar to wounded Ptgs2⫹/- mice (Figure 1D), suggesting that Myd88 signaling was required for maximal Ptgs2 expression. Therefore, we next evaluated Ptgs2 expression in wounds 2 days postinjury and found that the magnitude of Ptgs2 induction was 30-fold lower in Myd88⫺/⫺ compared with WT mice (Figure 7B). We hypothesized that the dramatic difference in Ptgs2 expression in wounded Myd88⫺/⫺ mice was owing to defective Ptgs2 regulation in Ptgs2 high-expressing cMSCs. In support of this hypothesis, we found that in contrast to WT littermate controls, Igf2bp1 was not induced in Myd88⫺/⫺ mice at day 2 post-injury (Figure 7C). To test whether cell-intrinsic Myd88 signaling was required in cMSCs, we isolated cMSCs and found significantly lower expression levels of Igf2bp1 in Myd88⫺/⫺ compared with WT cMSCs (Figure 7D). From these data, we conclude that Myd88 signaling was required for induction of Igf2bp1 expression in cMSCs. Because Ptgs2 also can be induced in macrophages by a Myd88-dependent, cell-intrinsic mechanism,3 we tested whether Myd88 signaling was required for the expression of Igf2bp1 in this cell type. We treated WT bone marrow– derived macrophages with lipopolysaccharide, which dramatically induced Ptgs2 mRNA expression as expected (Figure 7E). We found that Igf2bp1 expression was not detectable in either lipopolysaccharide-treated or untreated macrophages. Thus, not all cell types capable of expressing Ptgs2 also express Igf2bp1 (Figure 7F). Therefore, the Myd88-dependent up-regulation of Igf2bp1 may be a specific mechanism in MSCs.

Discussion Igf2bp1 Is a Novel Regulator of Ptgs2 in cMSCs Igf2bp1 is a VICKZ mRNA-binding protein and its homologs found in other organisms have multiple names (Vg1 RNA binding protein, Insulin-like growth factor II mRNA binding protein 1, Coding region determinant-binding protein, KH domain containing protein overexpressed in cancer, and Zipcode-binding

Figure 6. Igf2bp1 is expressed in Ptgs2 high-expressing cMSCs during the first phase of wound repair. (A) Graph of relative Igf2bp1 mRNA expression (biopsy-injured compared with uninjured mucosa) at days 2, 4, and 6 post-injury (n ⱖ 5 wounds/day). Data are displayed as means ⫾ standard error of the mean with the Student t test comparing injured with uninjured mucosa at each day. *P ⬍ .05. (B) A section of WT colon 2 days post-biopsy was stained with anti-Ptgs2 (red) and goat antiIgf2bp1 (green) antibodies. Individual and merged channels are shown. White dashed lines, WAE cells basal surface. Yellow box, inset. Scale bars: 50 ␮m (inset scale bars: 10 ␮m).

IGF2BP1 REGULATES PTGS2 MRNA IN CMSCS

119

Figure 7. Myd88 is required for Igf2bp1 expression in cMSCs. (A) Graph of loss of ␣-SMA in WT and Myd88⫺/⫺ sections 6 days post-injury (n ⱖ 5 per genotype). (B and C) Graphs showing relative (B) Ptgs2 or (C) Igf2bp1 mRNA expression comparing day 2 wounds with uninjured mucosa (n ⱖ 4 per genotype) for WT and Myd88⫺/⫺ mice. (D) Graph of Igf2bp1 mRNA relative expression comparing cultured Myd88⫺/⫺ and WT cMSCs (n ⫽ 3 experiments with 3– 4 replicates/sample). (E) Graph of relative Ptgs2 and Igf2bp1 mRNA expression comparing unstimulated bone marrow macrophages (Macs) with stimulated macrophages (Macs ⫹ lipopolysaccharide [LPS]) (n ⫽ 4/treatment). Data are displayed as means ⫾ standard error of the mean. Student t test: *P ⬍ .05; **P ⬍ .01. (F) Proposed model of the role of Igf2bp1 in cMSCs after biopsy injury. In WT mice, Myd88 signaling induced Igf2bp1 expression in cMSCs during the early phase of mucosal repair. Igf2bp1 interacted with Ptgs2 mRNA and was required for maximal Ptgs2 expression. Maximal expression of Ptgs2 was required for proper mucosal healing observed during the later phase of repair.

protein-1).26 ZBP-1 was identified as a zip-code binding factor that localizes ␤-actin mRNA to the leading edge of chicken embryo fibroblasts.27 Igf2bp1 regulates target mRNAs by several complex mechanisms,26 including the following: (1) direct binding to target mRNAs that affect their stability or translation,4 (2) scaffolding with other mRNA-binding proteins,28 and (3) controlling mRNA intracellular localization through the formation of Igf2bp1-containing ribonucleoprotein granules (complex structures of mRNAs and proteins). Because Igf2bp1 has both shared and unique mechanisms of regulation compared with the known AREbinding proteins, Igf2bp1 will be a key tool to elucidate principles that mediate the overall regulation of Ptgs2 in cMSCs. The expression pattern of Igf2bp1 and the phenotype of Igf2bp1-deficient mice show that Igf2bp1 is important in development.29,30 Igf2bp1 is an oncofetal protein that is expressed in embryos and various cancers but not adult tissues. Igf2bp1-deficient mice had dwarfism and impaired gut formation,30 thus, colonic biopsy of Igf2bp1-deficient mice was not feasible. An inducible Igf2bp1 deletion is required to test its function in the adult. Igf2bp1 is overexpressed in certain cancers including 81% of colorectal carcinomas.31 Because Ptgs2 is important for the progression of colon cancer and Igf2bp1 frequently is expressed in these cancers, the mechanism of the Ptgs2/Igf2bp1 interaction is highly relevant. In addi-

tion, because Igf2bp1 is virtually undetectable in healthy tissue, its interaction with and stabilization of Ptgs2 makes it an intriguing candidate target for drug development. Drugs inhibiting the interaction of Igf2bp1 with Ptgs2 mRNA may potentially have fewer side effects than traditional nonsteroidal anti-inflammatory drugs or Ptgs2 inhibitors.

MSCs as Mobilizable Sources of Ptgs2 During Repair Although many cell types can express Ptgs2 in the colon, we showed that cMSCs expressed the highest detectable levels of Ptgs2 mRNA and protein after mucosal biopsy. Because Ptgs2⫹/- mice had defective healing similar to Ptgs2⫺/⫺ mice, we concluded that maximal expression of Ptgs2 was necessary for proper mucosal repair. Previous studies identified numerous cell types capable of expressing Ptgs2,19 –21 and lineage-specific Ptgs2 knockouts showed that myeloid and endothelial expression of Ptgs2 was necessary for proper intestinal healing after dextran sodium sulfate treatment.8 We acknowledge that cMSCs are not the only cell type that expresses Ptgs2 during intestinal repair, and these previous studies highlight the importance of multiple Ptgs2-expressing cells, including myofibroblasts,32 during healing. The relative contributions of different cell types are unknown. However, we show that cMSCs express the highest levels of Ptgs2 after

BASIC AND TRANSLATIONAL AT

July 2012

120

MANIERI ET AL

BASIC AND TRANSLATIONAL AT

severe mucosal injury; therefore the regulation of Ptgs2 in these cells is highly relevant to mucosal healing. MSCs are an emerging cellular therapy for inflammatory diseases, such as inflammatory bowel disease.33 MSCs have at least 2 possible functions after intestinal injury, as follows: (1) differentiation into stromal lineages to replace damaged tissue and (2) expression of immunomodulatory molecules, including prostaglandins.33 The location of Ptgs2 high-expressing cMSCs directly beneath the epithelium may be functionally relevant because prostaglandins have a short half-life and act within a small area. Also, Ptgs2 is expressed during the first phase of repair when WAE cells begin migrating, and Ptgs2⫺/⫺ mice have a defective epithelial barrier. We were able to rescue the muscle degradation phenotype in Ptgs2⫺/⫺ mice, showing that the main function of Ptgs2 is the production of prostaglandins. We also were able to inject cMSCs adjacent to wounds and show that they were capable of migrating toward sites of inflammation from adjacent uninjured tissue. We believe this may be one major method of homing for cMSCs, although we cannot rule out the migration of cMSCs from the bone marrow. The accepted markers for MSCs are a constellation of positive markers including CD29, CD44, and CD54, as well as negative lineage markers.33,34 A challenge in the field is that there is no single marker that only marks cMSCs. Because Igf2bp1 protein was found only in cMSCs in wound beds, we will investigate further whether Igf2bp1 may be such a marker. In conclusion, we show the importance of maximal expression of Ptgs2 during mucosal repair, and identify Igf2bp1 as a novel mRNA-binding protein necessary for maximal Ptgs2 expression in cMSCs. Our study has implications for understanding complex mechanisms of Ptgs2 regulation by mRNA-binding proteins. Additional work will be needed to determine how Igf2bp1 regulates Ptgs2 expression and whether it interacts with ARE-binding proteins in cMSCs or affects Ptgs2 localization.

GASTROENTEROLOGY Vol. 143, No. 1

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

Supplementary Material Note: To access the supplementary material accompanying this article visit the online version of Gastroenterology at www.gastrojournal.org, and at http:// dx.doi.org/10.1053/j.gastro.2012.03.037.

20.

21.

References 1. Wang D, Mann JR, DuBois RN. The role of prostaglandins and other eicosanoids in the gastrointestinal tract. Gastroenterology 2005;128:1445–1461. 2. Ishikawa T, Jain NK, Taketo MM, et al. Imaging cyclooxygenase-2 (Cox-2) gene expression in living animals with a luciferase knock-in reporter gene. Mol Imaging Biol 2006;8:171–187. 3. Wu KK. Control of cyclooxygenase-2 transcriptional activation by pro-inflammatory mediators. Prostaglandins Leukot Essent Fatty Acids 2005;72:89 –93. 4. Barreau C, Paillard L, Osborne HB. AU-rich elements and associated factors: are there unifying principles? Nucleic Acids Res 2005;33:7138 –7150. 5. Cok SJ, Morrison AR. The 3’-untranslated region of murine cyclooxygenase-2 contains multiple regulatory elements that alter mes-

22.

23.

24.

25.

sage stability and translational efficiency. J Biol Chem 2001; 276:23179 –23185. Morteau O, Morham SG, Sellon R, et al. Impaired mucosal defense to acute colonic injury in mice lacking cyclooxygenase-1 or cyclooxygenase-2. J Clin Invest 2000;105:469 – 478. Brown SL, Riehl TE, Walker MR, et al. Myd88-dependent positioning of Ptgs2-expressing stromal cells maintains colonic epithelial proliferation during injury. J Clin Invest 2007;117:258 –269. Ishikawa T, Oshima M, Herschman HR. Cox-2 deletion in myeloid and endothelial cells, but not in epithelial cells, exacerbates murine colitis. Carcinogenesis 2011;32:417– 426. Zheng L, Riehl TE, Stenson WF. Regulation of colonic epithelial repair in mice by Toll-like receptors and hyaluronic acid. Gastroenterology 2009;137:2041–2051. Zhang L, Lu YM, Dong XY. Effects and mechanism of the selective COX-2 inhibitor, celecoxib, on rat colitis induced by trinitrobenzene sulfonic acid. Chin J Dig Dis 2004;5:110 –114. Riehl TE, Newberry RD, Lorenz RG, et al. TNFR1 mediates the radioprotective effects of lipopolysaccharide in the mouse intestine. Am J Physiol Gastrointest Liver Physiol 2004;286:G166 – G173. Seno H, Miyoshi H, Brown SL, et al. Efficient colonic mucosal wound repair requires Trem2 signaling. Proc Natl Acad Sci U S A 2009;106:256 –261. Pickert G, Neufert C, Leppkes M, et al. STAT3 links IL-22 signaling in intestinal epithelial cells to mucosal wound healing. J Exp Med 2009;206:1465–1472. Normand S, Delanoye-Crespin A, Bressenot A, et al. Nod-like receptor pyrin domain-containing protein 6 (NLRP6) controls epithelial self-renewal and colorectal carcinogenesis upon injury. Proc Natl Acad Sci U S A 2011;108:9601–9606. Morham SG, Langenbach R, Loftin CD, et al. Prostaglandin synthase 2 gene disruption causes severe renal pathology in the mouse. Cell 1995;83:473– 482. Adachi O, Kawai T, Takeda K, et al. Targeted disruption of the MyD88 gene results in loss of IL-1- and IL-18-mediated function. Immunity 1998;9:143–150. Walker MR, Brown SL, Riehl TE, et al. Growth factor regulation of prostaglandin-endoperoxide synthase 2 (Ptgs2) expression in colonic mesenchymal stem cells. J Biol Chem 2010;285:5026 – 5039. Robinson MH, Wheatley T, Leach IH. Nonsteroidal antiinflammatory drug-induced colonic stricture. An unusual cause of large bowel obstruction and perforation. Dig Dis Sci 1995;40:315–319. Pouliot M, Baillargeon J, Lee JC, et al. Inhibition of prostaglandin endoperoxide synthase-2 expression in stimulated human monocytes by inhibitors of p38 mitogen-activated protein kinase. J Immunol 1997;158:4930 – 4937. Jones MK, Wang H, Peskar BM, et al. Inhibition of angiogenesis by nonsteroidal anti-inflammatory drugs: insight into mechanisms and implications for cancer growth and ulcer healing. Nat Med 1999;5:1418 –1423. Mifflin RC, Saada JI, Di Mari JF, et al. Regulation of COX-2 expression in human intestinal myofibroblasts: mechanisms of IL-1mediated induction. Am J Physiol Cell Physiol 2002;282:C824 – C834. Desmouliere A, Rubbia-Brandt L, Abdiu A, et al. Alpha-smooth muscle actin is expressed in a subpopulation of cultured and cloned fibroblasts and is modulated by gamma-interferon. Exp Cell Res 1992;201:64 –73. Saada JI, Pinchuk IV, Barrera CA, et al. Subepithelial myofibroblasts are novel nonprofessional APCs in the human colonic mucosa. J Immunol 2006;177:5968 –5979. Geske MJ, Zhang X, Patel KK, et al. Fgf9 signaling regulates small intestinal elongation and mesenchymal development. Development 2008;135:2959 –2968. Mukhopadhyay D, Houchen CW, Kennedy S, et al. Coupled mRNA stabilization and translational silencing of cyclooxygenase-2 by a novel RNA binding protein, CUGBP2. Mol Cell 2003;11:113–126.

26. Yisraeli JK. VICKZ proteins: a multi-talented family of regulatory RNA-binding proteins. Biol Cell 2005;97:87–96. 27. Ross AF, Oleynikov Y, Kislauskis EH, et al. Characterization of a beta-actin mRNA zipcode-binding protein. Mol Cell Biol 1997;17: 2158 –2165. 28. Weidensdorfer D, Stohr N, Baude A, et al. Control of c-myc mRNA stability by IGF2BP1-associated cytoplasmic RNPs. RNA 2009;15:104– 115. 29. Yaniv K, Yisraeli JK. The involvement of a conserved family of RNA binding proteins in embryonic development and carcinogenesis. Gene 2002;287:49 –54. 30. Hansen TV, Hammer NA, Nielsen J, et al. Dwarfism and impaired gut development in insulin-like growth factor II mRNA-binding protein 1-deficient mice. Mol Cell Biol 2004;24:4448 – 4464. 31. Ross J, Lemm I, Berberet B. Overexpression of an mRNA-binding protein in human colorectal cancer. Oncogene 2001;20:6544–6550. 32. Mifflin RC, Pinchuk IV, Saada JI, et al. Intestinal myofibroblasts: targets for stem cell therapy. Am J Physiol Gastrointest Liver Physiol 2011;300:G684 –G696. 33. Manieri NA, Stappenbeck TS. Mesenchymal stem cell therapy of intestinal disease: are their effects systemic or localized? Curr Opin Gastroenterol 2011;27:119 –124. 34. Dominici M, Le Blanc K, Mueller I, et al. Minimal criteria for defining multipotent mesenchymal stromal cells. The International Society for Cellular Therapy position statement. Cytotherapy 2006;8:315–317.

IGF2BP1 REGULATES PTGS2 MRNA IN CMSCS

121

Received November 23, 2011. Accepted March 20, 2012. Reprint requests Address requests for reprints to: Thaddeus Stappenbeck, 660 S. Euclid Avenue, Box 8118, St. Louis, Missouri 63110. e-mail: [email protected]; fax: ⴙ1 314 362 7487. Acknowledgments The authors thank Dr William Stenson for critically reading the manuscript and Dr Deborah Rubin for providing the MIC 216 cells. Nicholas A. Manieri and Monica R. Drylewicz designed and performed the experiments and wrote the manuscript; Hiroyuki Miyoshi performed experiments; and Thaddeus Stappenbeck oversaw the project/manuscript. N.A.M. and M.R.D. contributed equally to this work. Conflicts of Interest The authors disclose no conflicts. Funding Supported by R01 DK071619, P30-DK52574 (Washington University Digestive Disease Research Core), National Institutes of Health/National Heart, Lung, and Blood Institute T32 HL007317 Predoctoral Pulmonary and Critical Care Training Grant and Pew Scholars in the Biomedical Sciences.

BASIC AND TRANSLATIONAL AT

July 2012

121.e1

MANIERI ET AL

Supplementary Materials and Methods Colonic Tissue Preparation For fixed-frozen sections, wounded mice were killed and perfused transcardially with 4% paraformaldehyde (PFA). After dissection, the colon was inflated with 4% PFA, opened longitudinally, and pinned flat in 4% PFA overnight. The following day the fixed colons were incubated in 20% sucrose–PBS for 6 hours. For frozen sections, colons were removed, washed in PBS, and placed in OCT compound (Sakura, Torrance, CA). For both procedures, each wound bed, including 1–2 mm of adjacent uninjured area, was removed with a razor blade and frozen in OCT compound. For each individual wound, the material was oriented so that serial 5-␮m– thick sections were obtained in a proximal to distal manner with respect to the orientation of the in vivo colon. Each wound bed was sectioned completely. To standardize our analysis, we evaluated the 5–10 sections in the central (largest) portion of a given wound bed as determined by light microscopic views of unstained sections. Serially numbered, unstained sections were stored at ⫺80°C until use. One of these sections was stained with H&E to confirm orientation.

In Situ Hybridization A cDNA clone for Ptgs2 (clone ID: 30059181) was purchased from Thermo Fisher Scientific. Digoxigeninlabeled antisense RNA probes were synthesized with T7 RNA polymerase (New England Biolabs, Ipswich, MA) and DIG RNA Labeling Mix (Roche, Indianapolis, IN) and purified with NucAway Spin Columns (Life Technologies, Carlsbad, CA). Fixed frozen sections (5-␮m thick) were fixed in 4% PFA for 10 minutes, treated with Protease K for 15 minutes (Viagen, Austin, TX), incubated in TEA and acetic anhydride for 10 minutes, and hybridized with RNA probes overnight. Sections then were treated with RNase in 2xSSC for 30 minutes (Sigma) and incubated with anti-digoxigenin-alkaline phosphatase Fab fragments for 1 hour (Roche). Specific signals were visualized by incubating sections for 1.5 hours with nitro blue tetrazolium chloride and 5-bromo-4-chloro-3-indolyl phosphate (Roche) and sections were counterstained with methyl green.

Immunofluorescence Fixed-frozen sections of wounds used for in situ hybridization also were used for immunofluorescence. Sections were fixed in 4% PFA, boiled in 10 mmol/L citrate buffer, rinsed in PBS, blocked in 3% bovine serum albumin/0.1% Triton-X for 20 minutes, and incubated with primary antibody for 1 hour. Slides then were rinsed with PBS, incubated with secondary antibody followed by bis-benzimide staining, and mounted with Mowiol 4-88 (EMD Chemicals, Billerica, MA). Sections were viewed with a Zeiss (Oberkochen, Germany) Axiovert 200 with an Axiocam MRM camera. Antibodies used were poly-

GASTROENTEROLOGY Vol. 143, No. 1

clonal rabbit anti-Ptgs2 (Thermo Scientific Lab Vision), goat anti-Igf2bp1 (E-20; Santa Cruz Biotechnology, Santa Cruz, CA), mouse monoclonal anti–␣-SMA conjugated to Cy3 (1A4; Sigma), rat monoclonal anti-CD29 (KMI6; BD Biosciences, San Jose, CA), rat monoclonal anti-CD31 (MEC 13.3; BD Biosciences), rat monoclonal anti-CD44 (IM7; BD Biosciences), rat monoclonal antiCD45 (30-F11; BD Biosciences), rat monoclonal antiCD54 (KAT-1; eBioscience, San Diego, CA), rat monoclonal anti-KI67 (TEC-3; DAKO, Glostrup, Denmark), monoclonal rat anti-F4/80 (CI:A3-1; Caltag Laboratories, Burlingame, CA), goat polyclonal anti– claudin 4 (C-18; Santa Cruz), rabbit polyclonal anti–ZO-1 (Invitrogen), rat monoclonal anti–major histocompatibility complex class II (M5/114.15.2; eBioscience), and rabbit polyclonal anti– ␤-catenin (Sigma). All primary antibodies were used at 1:100 dilution except anti-Ptgs2 (1:5,000 dilution). For Ptgs2, claudin-4, and ZO-1 staining, sections from unfixed colons were acetone-fixed (Ptgs2) or methanol-fixed (ZO-1 and claudin-4) for 5 minutes, rinsed in PBS, blocked, and incubated with primary and secondary antibodies without antigen retrieval. For all unconjugated antibodies, appropriate donkey anti-rat, anti-rabbit, or anti-goat secondary antibodies conjugated to Alexa Fluor 488 or Alexa Fluor 594 (Invitrogen) were used at 1:500 dilution to visualize staining.

Primers for qRT-PCR The primers for qRT-PCR were as follows: glyceraldehyde-3-phosphate dehydrogenase: AGGTCGGTGTGAACGGATTTG, TGTAGACCATGTAGTTGAGGTCA; Ptgs2: TGCCTGGTCTGATGATGTATG, GCCCTTCACGTTATTGCAGATG; Igf2bp1: CGGCAACCTCAACGAGAGT, GTAGCCGGATTTGACCAAGAA; Rbm3: CTTCGTAGGAGGGCTCAACTT, CTCCCGGTCCTTGACAACAAC; Rbms3: CAGCAGGAACGATTCAATCCC, AGGTATCAGCAACAACACCTTC; Rbpms: CCCGTAGGCTTTGTCAGTTTT, GTAAAGTGCAGGTACTGTGAGC; Ptgs1: AGTGCGGTCCAACCTTATCC, GCAGAATGCGAGTATAGTAGCTC; and Fstl1: CACGGCGAGGAGGAACCTA, TCTTGCCATTACTGCCACACA.

Immunoprecipitation and RNA Isolation in NIH3T3 Cells The open reading frame of Igf2bp1 was cloned (primers: forward: cccgaattcatgaacaagctttacatcggc; reverse: cccggatcccttcctccgagcctg) into p3XFLAG-CMV-14 (SigmaAldrich). Igf2bp1 truncation mutants RRM (residues 1–185; primers: Igf2bp1 F and cccggatcctgctgccacgggcg) and KH (residues 195–577; primers: cccgaattcatgatccctctccggctcc and Igf2bp1 R) were similarly cloned. NIH3T3 cells were plated in 10-cm dishes and transfected using FuGene HD reagent (Roche). Transfected cells were exposed to 300 mJ, 254 nm UV, scraped, pelleted, and lysed using polysome lysis buffer (100 mmol/L KCl, 5 mmol/L MgCl2, 10 mmol/L HEPES) supplemented with 0.05% NP-40, 2 mmol/L vana-

July 2012

dyl ribonucleoside complex, 100 U RNase inhibitor, 1:100 phosphatase inhibitor cocktails 2 and 3 (Sigma), and 1:40 protease inhibitor cocktail (Sigma). Lysates were precleared with mouse IgG-conjugated Protein A beads and then incubated overnight at 4°C with M2 antibody-conjugated agarose beads (Sigma). Beads were washed 3 times with polysome lysis buffer and 3 times with polysome lysis buffer–500 mmol/L NaCl. When noted, FLAG-tagged proteins were eluted using 100 ␮g/mL “3X FLAG” peptide (Sigma). Beads/pellet, total protein, and eluate fractions then either were boiled in Laemmli’s buffer to analyze protein or incubated at 55°C with 30 ␮g proteinase K and 0.1% sodium dodecyl sulfate in 100 ␮L Tris-buffered saline for 30 minutes, phenol-chloroform extracted, and ethanol precipitated to isolate RNA.

shRNA Transfection HEK293T cells were transfected with Mission shRNA constructs specific for Igf2bp1 or nontargeting control (Igf2bp1 clones NM_009951.2-1611s1c1, NM_ 009951.2-1874s1c1; Sigma) and lentiviral packaging plasmids using FuGene HD reagent (Roche). The next day, cell media was refreshed and allowed to accumulate virus for 24 hours. Viral-containing supernatant was used undiluted on plated cMSCs for 24 hours before adding puromycin-containing media for selection. shRNAs specific for the candidates or control shRNA were transfected into HEK293T cells along with lentiviral packaging vectors to produce lentivirus that then was used to infect cMSCs in culture. After 1–2 passages, expression of the targeted gene and Ptgs2 were assessed by qRT-PCR.

Immunoblotting Cells were lysed in RIPA buffer (Sigma) containing protease and phosphatase inhibitor cocktails (Sigma). Lysates were mixed 1:1 with 2X Laemmli’s and boiled, run on 10% or 12% Tris-Gly gels (BioRad, Hercules, CA), transferred to a 0.45-␮m nitrocellulose membrane, and blocked in 5% milk in 0.05% Tween 20 Tris-buffered saline overnight at 4°C. Blots were incubated for 1 hour with horseradish-peroxidase– conjugated M2 antibody (Sigma) before development using the SuperSignal West Dura chemiluminescent kit (Thermo Fisher Scientific).

Cell Isolation and Culture Conditions Colons were minced and treated with collagenase type 1 (from Clostridium histolyticum; Invitrogen). Released cells were cultured until confluent in 10 mmol/L HEPESlow glucose Dulbecco’s modified Eagle medium with 10% fetal bovine serum, and media was changed every 3 days. All experiments were performed on confluent cells. NIH3T3 cells and MIC 216 cells were cultured similarly. Bone marrow– derived macrophages were cultured from bone marrow flushed from the femurs and tibias of a single mouse into 30 mL of Dulbecco’s modified Eagle medium. The cell suspension was split into 2 equal por-

IGF2BP1 REGULATES PTGS2 MRNA IN CMSCS

121.e2

tions and 35 mL of a specific conditioned media was added to each aliquot. The cell suspensions then were pipetted into 10 ⫻ 15 cm, nontissue culture, Petri dishes (5 plates, 10 mL/plate) and the cells were incubated in a standard tissue culture incubator for 7 days to stimulate differentiation into macrophages by incubating with macrophage differentiation media. Differentiation media consisted of Dulbecco’s modified Eagle media supplemented with 10% fetal bovine serum, 20% L-cell supernatant (from L929 fibroblasts), 5% horse serum, 1% L-glutamine, 1% sodium pyruvate, 1:100 of 1 mol/L HEPES, 1 ␮L/mL of 10,000 U/mL penicillin/streptomycin with media changed on days 4 and 7. For lipopolysaccharide experiments, macrophages were treated with EDTA to remove them from the plate. The cells then were re-plated at 1.25 ⫻ 105 cells/mL for 18 hours overnight and then treated for 1 hour with 10 ng/mL lipopolysaccharide from Escherichia coli (Sigma-Aldrich) or fresh conditioned media.

Injection of GFP-Expressing cMSCs Adjacent to Wounds cMSCs were isolated from CMV immediate enhancer/␤-actin promoter (CAG)-GFP mice obtained from Jackson Laboratory. To set-up equipment for cMSC injection into the mucosa, a 30-gauge needle was superglued to one end of an approximately 30-cm long piece of 26-gauge polytetrafluoroethylene thin-wall tubing (Zeus, Orangeburg, SC). This then was inserted through the sheath adjacent to the colonoscope camera. A 25-gauge needle was attached to the other end of the tubing outside the colonoscope. This needle then was attached to a syringe containing 1 ⫻ 106 cMSCs. Six hours post-biopsy, the 30-gauge needle tip was positioned adjacent to wounds and inserted directly into the mucosa, but not through the muscularis propria. cMSCs then slowly were injected to maintain cMSCs at the site of injection. Mice were killed 4 – 6 days post-injection to observe cMSC migration into wounds. Bright field and fluorescent images were taken at the time of death, and wounds were frozen in OCT and cut into 5-␮m sections. Sections were stained with bis-benzimide to label nuclei and GFP-expressing cells were observed within wound beds.

Injection of Stable Analogues of PGE2 and PGI2 Stable analogues of PGE2 and PGI2 (16,16-dimethyl prostaglandin E2 and Iloprost, respectively) were purchased from Cayman Chemical (Ann Arbor, MI). The prostaglandins were diluted into sterile 5% sodium bicarbonate immediately before use. Twice-daily injections of Iloprost (200 ␮g/kg body weight) and 16,16-dimethyl prostaglandin E2 (100 ␮g/kg body weight) were injected intraperitoneally into biopsy-injured Ptgs2⫺/⫺ mice for 6 days. Mice were killed at 6 days post-biopsy and wounds were processed for sectioning and immunofluorescence microscopy.

121.e3

MANIERI ET AL

GASTROENTEROLOGY Vol. 143, No. 1

Supplementary Figure 1. Two functional alleles of Ptgs2 are required to prevent accumulation of granulation tissue throughout the muscularis propria at day 6 post-biopsy. Representative sections from (A) WT, (B) Ptgs2⫹/-, and (C) Ptgs2⫺/⫺ mice 6 days after colonic biopsy were stained with anti-F4/80 antisera to detect macrophages. The yellow boxed regions indicated muscularis propria. Representative sections from (D) WT, (E) Ptgs2⫹/-, and (F) Ptgs2⫺/⫺ mouse colons 6 days after biopsy injury stained with anti–␤-catenin antisera (green) to detect epithelial cells, anti-CD31 antisera (red) to detect endothelial cells, and bis-benzamide to label nuclei (blue). The yellow boxed regions indicated muscularis propria. Representative sections from (G) WT, (H) Ptgs2⫹/-, and (I) Ptgs2⫺/⫺ mouse colons 2 days after biopsy injury stained with anti–␤-catenin antisera (green) to label epithelium, anti–␣-SMA antisera (red) to detect muscularis propria, and bis-benzamide to label nuclei (blue). Scale bar: 100 ␮m.

July 2012

IGF2BP1 REGULATES PTGS2 MRNA IN CMSCS

121.e4

Supplementary Figure 2. Injured Ptgs2⫺/⫺ mice have incomplete migration of WAE cells 6 days post-injury. Whole-mount images of WT (left) and Ptgs2⫺/⫺ (right) colons pinned in a proximal to distal orientation 6 days post-injury. The initial injury sites are outlined in solid lines. Epithelial cells completely cover the center of WT wounds, whereas Ptgs2⫺/⫺mice have an incomplete epithelial layer covering the center of wounds as indicated by black arrowheads. To generate sections for histologic analysis, wounds were oriented and cut in a proximal to distal manner and sections continually were cut in 5-␮m increments. The sections with the largest area of wound bed (dashed lines) were considered the center of wounds and used for immunofluorescence staining.

121.e5

MANIERI ET AL

GASTROENTEROLOGY Vol. 143, No. 1

July 2012

IGF2BP1 REGULATES PTGS2 MRNA IN CMSCS

121.e6

Supplementary Figure 4. Ptgs2⫺/⫺ mice have defective crypt regeneration at day 14 post-biopsy. Sections of (A) WT, (B) Ptgs2⫹/-, and (C) Ptgs2⫺/⫺ colons 14 days after biopsy injury were stained with H&E. (D) Graph of the largest distance between crypts at day 14 post-injury in WT, Ptgs2⫹/-, and Ptgs2⫺/⫺ mice (n ⫽ 4 wounds per genotype). Scale bar: 200 ␮m. Data were displayed as means ⫾ standard error of the mean and one-way ANOVA with the post hoc Tukey test: F2,8 ⫽ 5.82. *P ⬍ .05.

Supplementary Figure 3. Ptgs2⫺/⫺mice have immature WAE cells 6 days post-injury. (A) Sections from WT and (B) Ptgs2⫺/⫺mice 6 days post-injury were stained with claudin-4 (red) to label immature WAE cells and ␤-catenin (green) to label all epithelial cells. Wound sections from WT mice have rare claudin-4 –positive cells, indicated by white arrowheads. Wound sections from Ptgs2⫺/⫺mice are covered by claudin-4 –positive immature epithelial cells (white bar), indicating an immature epithelial layer. Merged and individual channels are shown. Scale bars: 100 ␮m. (C) Sections from WT and (D) Ptgs2⫺/⫺ mice 6 days post-injury were stained with ZO-1 (green) to observe tight junction formation. WT mice have punctate ZO-1 expression in epithelial cells covering the wound, whereas Ptgs2⫺/⫺ mice lack ZO-1 expression in WAE cells, indicating a lack of tight junctions and maturation. White boxes indicate the area for magnification in the inset below. Scale bars: 100 ␮m; inset scale bars: 15 ␮m.

121.e7

MANIERI ET AL

Supplementary Figure 5. Ptgs1 mRNA is not increased significantly in wounds compared with uninjured mucosa. Graph of the relative expression of Ptgs1 mRNA isolated from WT littermate control wounds and adjacent uninjured areas at days 2, 4, and 6 post-injury (n ⫽ 4 – 6 wounds per genotype). There were no significant differences comparing times post-injury by ANOVA and no significant differences between injured and uninjured mucosa at each time point by the Student t test. Data were displayed as means (ratio of injured to uninjured) ⫾ standard error of the mean.

Supplementary Figure 6. Ptgs2⫹/- mice have submaximal expression of Ptgs2 at day 4 post-injury compared with WT mice. Graph of the relative expression of Ptgs2 using RNA isolated from WT littermate control and Ptgs2⫹/- wounds 4 days post-injury (n ⫽ 5 wounds per genotype). Data were displayed as means ⫾ SEM and the Student t test. **P ⬍ .01.

GASTROENTEROLOGY Vol. 143, No. 1

Supplementary Figure 7. No detectable staining in Ptgs2⫺/⫺ mouse colons labeled with anti-Ptgs2 antisera. Section of a colon from a Ptgs2⫺/⫺ mouse taken 2 days after biopsy injury. The section was stained with anti-Ptgs2 antisera (green) and bis-benzamide to label nuclei (blue). The wound bed is outlined in white and the muscularis mucosa is outlined in red. There is no background nonspecific staining in Ptgs2⫺/⫺ mice stained with Ptgs2 antisera.

July 2012

IGF2BP1 REGULATES PTGS2 MRNA IN CMSCS

121.e8

Supplementary Figure 8. GFP-expressing cMSCs specifically migrate into sites of colonic injury from adjacent uninjured mucosa. (A) Bright field and fluorescent or (B) fluorescent images were taken 4 days after colonic biopsy and endoscopy-guided injection of GFP-expressing cMSCs. The initial injection site was determined by the highest density of GFP-expressing cells. A halo of GFP signal showed directionality toward the wound bed, but not in the opposite direction, suggesting directed chemotaxis of cMSCs. (C) Bright field and fluorescent or (D) fluorescent images of the same wound photographed in panels A and B. GFP-expressing cMSCs were found only in an area of the wound closest to the injection site, but not in areas of the wound that were more distant. This suggests that cMSCs can migrate directly into wounds at a certain distance. (E and F) A section of colon was stained with bis-benzamide to label nuclei 6 days post-biopsy and post-injection of cMSCs. GFP-expressing cMSCs can be seen throughout the wound bed as in panel E, but were not found at sites further from the wound on the same section as in panel F.

121.e9

MANIERI ET AL

Supplementary Figure 9. Isolated cMSCs are not identical to myofibroblasts. Graph of relative expression of Fstl1 in cMSCs isolated from wild-type mice and MIC 216 cells, a myofibroblast cell line. Data were displayed as means ⫾ standard error of the mean and the Student t test. ***P ⬍ .001; n ⫽ 6 wells per cell type.

GASTROENTEROLOGY Vol. 143, No. 1

July 2012

IGF2BP1 REGULATES PTGS2 MRNA IN CMSCS

121.e10

Supplementary Figure 10. Intraperitoneal injections of prostaglandins rescue the muscle degradation phenotype in Ptgs2⫺/⫺ mice. Sections of colons from Ptgs2⫺/⫺ mice injected with (A) stable analogues of PGE2 and PGI2 or (B) vehicle 6 days post-biopsy. The sections were stained with anti–␣-SMA antisera (red) to detect muscularis propria, anti–␤-catenin (green) to detect the epithelium, and bis-benzamide to label nuclei (blue). (C) Graph of the percentage loss of ␣-SMA underlying day 6 wound beds (gap length/wound bed length) in prostaglandin-treated or vehicle-treated mice (n ⱖ 3/treatment). Data were displayed with a median bar. Student t test: ***P ⬍ .001.