HuR Enhances Early Restitution of the Intestinal Epithelium by Increasing Cdc42 Translation

ABSTRACT

The mammalian intestinal mucosa exhibits a spectrum of responses after acute injury and repairs itself rapidly to restore the epithelial integrity. The RNA-binding protein HuR regulates the stability and translation of target mRNAs and is involved in many aspects of gut epithelium homeostasis, but its exact role in the regulation of mucosal repair after injury remains unknown. We show here that HuR is essential for early intestinal epithelial restitution by increasing the expression of cell division control protein 42 (Cdc42) at the posttranscriptional level. HuR bound to the Cdc42 mRNA via its 3′ untranslated region, and this association specifically enhanced Cdc42 translation without an effect on the Cdc42 mRNA level. Intestinal epithelium-specific HuR knockout not only decreased Cdc42 levels in mucosal tissues, but it also inhibited repair of damaged mucosa induced by mesenteric ischemia/reperfusion in the small intestine and by dextran sulfate sodium in the colon. Furthermore, Cdc42 silencing prevented HuR-mediated stimulation of cell migration over the wounded area by altering the subcellular distribution of F-actin. These results indicate that HuR promotes early intestinal mucosal repair after injury by increasing Cdc42 translation and demonstrate the importance of HuR deficiency in the pathogenesis of delayed mucosal healing in certain pathological conditions.

INTRODUCTION

The mammalian intestinal epithelium is a rapidly self-renewing tissue in the body. It functions as a dynamic physical barrier that separates mucosal tissues from a wide array of luminal noxious substances and microorganisms, in addition to its essential role in digestion and absorption. After damage to the epithelium, the intestinal mucosa exhibits a spectrum of responses and repairs itself quickly to restore epithelial integrity (1). In stressful environments, successful repair of wounds and ulcers requires epithelial cells to rapidly change gene expression patterns to regulate cell survival, migration, proliferation, and differentiation (2, 3). Early epithelial restitution is a primary repair modality in the gastrointestinal tract, and it rapidly reseals superficial wounds by migrating visible remaining epithelial cells from areas adjacent to the injured surface to cover the wounded area, followed by proliferation and differentiation (4, 5). This early rapid mucosal reepithelialization after mucosal injury is a complex process that includes the flattening, spreading, migrating, and repolarizing of differentiated columnar epithelial cells but is independent of cell proliferation (2, 6). Defective regulation of early mucosal restitution underlies various critical pathological states such as massive mucosal injury and hemorrhage, delayed wound healing, disruption of epithelial integrity, and epithelial barrier dysfunction (2, 7).

The gene expression programs that control the intestinal mucosal repair and epithelial homeostasis are tightly regulated at the transcriptional level, but the essential contribution of posttranscriptional events is also increasingly recognized (8–11). In particular, changes in mRNA stability and translation play an important role in the regulation of gut mucosal regeneration, protection, and permeability in response to stress (3, 10, 12, 13). The turnover and/or translational regulation of mRNAs involves the interaction of specific mRNA sequences (cis elements) with specific trans-acting factors such as RNA-binding proteins (RBPs) and noncoding RNAs (ncRNAs; particularly microRNAs) (3, 14, 15). RBPs and microRNAs (miRNAs) directly interact with cis elements on the mRNA, frequently at the 3′ untranslated region (3′ UTR), and regulate the stability and translation of target transcripts (16, 17). Long ncRNAs (lncRNAs) can regulate gene expression at multiple levels via distinct mechanisms, including chromatin remodeling, transcriptional and posttranscriptional control, and protein metabolism (18). An increasing body of evidence indicates that homeostasis of the gut epithelium is controlled by RBPs and ncRNAs acting in concert to regulate gene expression synergistically or antagonistically (8). For example, the RBP HuR directly binds to the lncRNA H19 and prevents miRNA 675 (miR-675) processing from H19, thus rescuing expression of tight junctions and gut barrier function in cells overexpressing H19 (19). In contrast, HuR and lncRNA SPRY4-IT1 increase TJ expression at the posttranscriptional level and promote the gut barrier function synergistically by enhancing the association of SPRY4-IT1 with the TJ mRNAs (20). Moreover, HuR competes with miR-195 to modulate Stim1 mRNA stability antagonistically (21), whereas the RBP CUG-binding protein 1 (CUGBP1) and miR-222 inhibit translation of cyclin-dependent kinase 4 synergistically (22).

HuR (encoded by the Elavl1 gene) is among the most prominent translation and turnover regulatory RBPs, and it generally associates with AU-/U-rich elements located in the 3′ UTRs of labile mRNAs (23, 24). Although constitutive HuR inactivation in vivo is lethal to embryos (5), tissue-specific conditional HuR knockout mice have proven a useful tool to investigate the in vivo functions of HuR under biological and pathological conditions. It has been reported that myeloid deletion of HuR exacerbates the production of proinflammatory cytokines and increases the sensitivity to acute inflammatory reactions such as endotoxemia (25), whereas HuR deletion in germ cells leads to male but not female sterility (26). HuR is crucial for the maintenance of hematopoietic stem cells (27), the selection and chemotaxis of T cells (24), and B-cell development and activation/differentiation following an antigen encounter (28). Our previous study (29) and findings from others (30) have shown that conditional deletion of HuR in intestinal epithelial cells (IECs) inhibits mucosal growth in the small intestine and reduces tumor development by targeting multiple genes including the Wnt coreceptor Lrp6. Here, we study the role of HuR in the regulation of intestinal mucosal repair after acute injury and demonstrate that HuR enhances early epithelial restitution by activating translation of the mRNA that encodes the small GTP-binding protein Cdc42.

RESULTS

HuR interacts with the Cdc42 mRNA and enhances its translation.

The Cdc42 mRNA is a potential target of HuR, since it contains several computationally predicted HuR-binding sites in its 3′ UTR (see Fig. S1 in the supplemental material), using standard online software (TargetScan and RNA22). To examine the interaction of Cdc42 mRNA with HuR, ribonucleoprotein (RNP) immunoprecipitation (IP) assays were performed using anti-HuR antibody under conditions that preserved RNP integrity (31). After incubation with anti-HuR antibody or IgG, RNA was isolated from the IP material and then subject to reverse transcription (RT), followed by either conventional PCR or quantitative PCR (Q-PCR) analyses. As shown in Fig. 1Aa, the Cdc42 PCR products were highly enriched in HuR samples compared with control IgG samples. The abundance of the Myc PCR product was also examined and served as a positive control (data not shown), since Myc mRNA is a known target of HuR (32), whereas the detection of Gapdh mRNA, a nonspecific contaminating housekeeping transcript (not a target of HuR), served to monitor the evenness of sample input, as reported previously (29, 33). HuR also associated with the mRNAs encoding Rac1 and RhoA but did not preferentially interact with RhoB and RhoC mRNAs (see Fig. S2 in the supplemental material. To examine whether HuR directly interacts with the 5′ UTR, coding region (CR), or 3′ UTR of the Cdc42 mRNA, “HuR/Cdc42 mRNA” associations were further tested by using biotinylated transcripts which spanned the Cdc42 mRNA regions (Fig. 1B, schematic). After incubation with cytoplasmic lysates, the interaction between the biotinylated Cdc42 transcripts and HuR was examined by biotin pulldown, followed by Western blotting (34, 35). As shown, HuR only associated with the Cdc42 3′ UTR but not to 5′ UTR or CR transcripts. Studies mapping of the interaction between HuR and Cdc42 3′ UTR were also performed by testing the interaction of partial biotinylated transcripts spanning the Cdc42 3′ UTR with HuR and showed that HuR interacted with both fragments 3′UTR-F1 and 3′UTR-F2 that contained hits of the HuR signature motif. Consistent with observations in cultured IECs, HuR/Cdc42 mRNA complexes were enriched in the small intestinal mucosa isolated from control littermates but not in the mucosa from IE-HuR^−/− mice, as examined by ribonuclear immunoprecipitation (RIP)/Q-PCR analysis (Fig. 1C). These results indicate that the Cdc42 mRNA is a novel target of HuR in the intestinal epithelium and that HuR specifically binds to the Cdc42 3′ UTR.

FIG 1.

HuR directly interacts with the 3′ UTR of Cdc42 mRNA. (A) The association of endogenous HuR with endogenous Cdc42 mRNA in cultured IEC-6 cells was examined by RIP using either anti-HuR antibody (Ab) or control IgG, followed by RT-Q-PCR analysis. (a) Cdc42 mRNA in HuR IP, as measured by RT-PCR (left) and RT-Q-PCR (right) analyses. (b) Levels of total input mRNAs. Values are means ± the SEM from triplicate samples. *, P < 0.05 (compared to IgG IP). (B) HuR immunoblots using the pulldown materials by biotinylated transcripts of Cdc42 5′ UTR, CR, and 3′ UTR. (Left) Schematic representation of various biotinylated Cdc42 transcripts used in this study. Cytoplasmic lysates were incubated with 6 μg of biotinylated Cdc42 5′ UTR, CR, or 3′ UTR or fragments of Cdc42 3′ UTR for 30 min at 25°C, and the resulting RNP complexes were pulled down by streptavidin-coated beads. The presence of HuR in the pulldown material was assayed by Western blotting. GAPDH in the pulldown material was also detected and served as a negative control. (C) The association of HuR with the Cdc42 mRNA in small intestinal mucosa in littermates and IE-HuR^−/− mice was measured by RIP followed by RT-Q-PCR analysis. (Left) Cdc42 mRNA in HuR IP; (right) input mRNA. Values are means ± the SEM (n = 4). * and +, P < 0.05 (compared to IgG and littermates, respectively).

To investigate the functional consequence of the HuR/Cdc42 mRNA association, the levels of HuR were decreased by transfecting IEC-6 cells with small interfering RNA (siRNA) targeting the HuR mRNA (siHuR) or increased by infection with an adenovirus expressing HuR (AdHuR). Consistent with our previous studies (32, 33), transfection with siHuR decreased HuR protein levels by >90% (Fig. 2A, top). This reduction was specific, since other RBPs, such as CUGBP1 and TIAR, were not affected in HuR-silenced cells, as reported previously (29, 33). HuR silencing specifically lowered Cdc42 protein levels by >95% (Fig. 2A, second row) (n = 3; P < 0.05) and slightly increased the levels of RhoB, although it failed to alter the expression levels of other GTP-binding proteins, including Rac1, RhoA, and RhoC. The reduction in Cdc42 levels likely occurred at the translation level, since silencing HuR did not decrease the levels of Cdc42 mRNA (Fig. 2B) or its stability (not shown) but repressed the rate of nascent Cdc42 protein synthesis (Fig. 2C). The levels of newly synthesized Cdc42 protein in HuR-silenced cells decreased by ∼85% (n = 3; P < 0.05) compared with those for cells transfected with C-siRNA. To further define the role of HuR in the regulation of Cdc42 translation, we examined the relative distribution of Cdc42 mRNA in individual fractions from polyribosome gradients after HuR silencing. Although decreasing the levels of HuR did not affect global polysomal profiles (data not shown), the abundance of Cdc42 mRNA associated with actively translating components of the gradient (fractions 6 to 9) decreased dramatically in HuR-silenced cells, where a significant shift of Cdc42 mRNA was observed toward low-translating parts of the gradient (fractions 3 and 4) (Fig. 2D, top panel). In contrast, Gapdh mRNA, which is not a target of HuR, distributed similarly in both groups (Fig. 2D, bottom panel).

FIG 2.

HuR silencing inhibits Cdc42 translation. (A) Immunoblots of HuR, Cdc42, and other GTPases in cells transfected with siHuR or C-siRNA. Western blot analysis was carried out 48 h later. (B) Levels of Cdc42 mRNA as measured by RT-Q-PCR analysis in cells treated as described in panel A. Values are means ± the SEM (n = 3). (C) Newly synthesized Cdc42 protein as measured by l-azidohomoalanine (AHA) incorporation assays. (D) Distributions of Cdc42 (top) and Gapdh (bottom) mRNAs in each gradient fraction of polysomal profiles prepared from cells described in panel A. (E) Levels of reporter activities as measured by analysis of Cdc42 5′-UTR, CR, and 3′-UTR luciferase reporters after HuR silencing. (Top) Schematic of plasmids of different chimeric firefly luciferase Cdc42 reporters. *, P < 0.05 (compared to C-siRNA).

To examine whether the effect of HuR on Cdc42 translation through the Cdc42 mRNA 3′ UTR, we used a firefly luciferase reporter gene construct containing the Cdc42 5′ UTR, CR, or 3′ UTR and the negative-control vector pGL3-Luc (Fig. 2E, schematic). A plasmid expressing Renilla luciferase (Fig. 2E, RLuc) was also cotransfected as an internal control for normalization of firefly luciferase. To distinguish translational output from changes in mRNA turnover, the luciferase activities were normalized to luciferase mRNA levels to assess the translational efficiency (the “translation index”). HuR silencing by transfection with siHuR decreased the levels of luciferase reporter activity when cells were transfected with the Luc-3′UTR (containing the full-length Cdc42 3′ UTR) but not with the Luc-5′UTR or Luc-CR. On the other hand, ectopic overexpression of HuR by infection with the AdHuR increased the levels of Cdc42 protein by ∼3.2-fold (Fig. 3A) (n = 3; P < 0.05), although it did not alter Cdc42 mRNA content (Fig. 3B). This induction in Cdc42 levels was due to the stimulation of Cdc42 translation (Fig. 3C) via interaction with its 3′ UTR (Fig. 3D). Together, these results indicate that HuR enhances Cdc42 expression by enhancing its translation through the Cdc42 3′ UTR.

FIG 3.

Ectopic HuR overexpression enhances Cdc42 translation. (A) Immunoblots of HuR and Cdc42 in cells infected with the recombinant adenoviral vector encoding HuR cDNA (HuR) or adenoviral vector lacking HuR cDNA (Vector) and analyzed 48 h later. Protein levels were examined by Western blotting and equal loading was monitored by assessing GAPDH levels. (B) Levels of the Cdc42 mRNA in cells treated as described in panel A. Values are means ± the SEM (n = 3). (C) Newly synthesized Cdc42 protein as measured by l-azidohomoalanine (AHA) incorporation assays. (D) Activities of Cdc42 5′-UTR, CR, and 3′-UTR luciferase reporters after HuR overexpression. *, P < 0.05 (compared to vector).

Intestinal epithelium-specific HuR deletion delays mucosal repair in the small intestine by decreasing Cdc42 levels.

To determine the in vivo function of HuR-mediated Cdc42 expression in the intestinal epithelium, we used IE-HuR^−/− mice that we recently generated. As reported previously (19, 29), HuR levels in the small intestinal (Fig. 4A, left) and colonic mucosa (data not shown) were undetectable in IE-HuR^−/− mice but were at wild-type levels in stomach mucosa, lung, liver, and pancreas. These levels were similar to those observed in control littermates (data not shown). HuR deletion did not alter the expression levels of RBPs CUGBP1 and AUF1 in the intestinal mucosa, although it marginally increased the levels of TIAR protein. Consistent with findings in cultured IECs (Fig. 2A), decreased levels of HuR in the intestinal mucosa by epithelial tissue-specific deletion of the Elavl1 gene inhibited expression of Cdc42 (Fig. 4A, right), as indicated by a dramatic decrease in the levels of Cdc42 protein (by ∼97%; n = 4; P < 0.05). In IE-HuR^−/− mice, mucosal Rac1 levels were unaffected, RhoB expression levels decreased modestly, and RhoC levels increased compared with those observed in control littermates. Immunohistochemical staining assays revealed that Cdc42 was predominantly distributed in epithelial cells of the villous area in control littermates, but its immunoreactivity disappeared completely in the intestinal mucosa of an IE-HuR^−/− mouse (Fig. 4B, bottom). As shown in results from studies in vitro (Fig. 2B), there were no differences in the levels of Cdc42 mRNA in the intestinal mucosa between IE-HuR^−/− mice and control littermates (Fig. 4C), further supporting the role of HuR in the translational regulation of Cdc42 expression.

FIG 4.

HuR deletion in IECs decreases the levels of intestinal mucosal Cdc42 in vivo. (A) Immunoblots of HuR, other RBPs (left), Cdc42, and other GTPases (right) in the small intestinal mucosa obtained from control littermate and IE-HuR^−/− mice. (B) Immunohistochemical staining of HuR (top, shown in dark brown) and Cdc42 (bottom, shown in green) in small intestinal mucosa. Scale bars, 50 μm. (C) Levels of mRNAs encoding Cdc42, Rac1, RhoB, and RhoC in the small intestinal mucosa as measured by RT-Q-PCR analysis. Values are means ± the SEM (n = 5).

To examine whether the reduction of Cdc42 expression after HuR deletion affected mucosal repair in the small intestine, a mesenteric ischemia/reperfusion (I/R)-induced mucosal injury model was used in this study. Both littermates and IE-HuR^−/− mice subjected to mesenteric I/R exhibited signs of significant mucosal injury and erosions in the small intestine, as indicated by severe sloughed cells, denuded villi with dilated capillaries, and by frank hemorrhage microscopically (Fig. 5A, middle) and swollen and edematous mucosa with areas of red steaks macroscopically (data not shown). Surprisingly, HuR deletion did not enhance the digress of I/R-induced mucosal injury, since the injury scores in IE-HuR^−/− mice were similar to those observed in littermates when measured immediately after mesenteric I/R (Fig. 5B). However, HuR-deficient epithelium displayed a significant delay in the process of mucosal repair after injury. In littermate mice, the mucosa repaired quickly and the epithelial integrity was almost completely restored 6 h after I/R. In contrast, this process was inhibited by HuR deletion, since the mucosal surface remained discontinuous, showing sloughed cells and debris in IE-HuR^−/− mice at the same time after I/R. Importantly, the levels of Cdc42 protein in the small intestinal mucosa increased remarkably in littermate mice after I/R-induced injury and remained elevated for additional 6 h thereafter (Fig. 5C and D), but this induction in mucosal Cdc42 abundance by I/R-induced stress was lost in IE-HuR^−/− mice. On the other hand, there were no differences in the response patterns of Rac1 and RhoB expression after exposure to I/R between IE-HuR^−/− mice and littermates, although the levels of mucosal RhoC decreased 6 h after I/R in HuR-deficient mice. These results indicate that targeted deletion of HuR in IECs delays mucosal repair in the small intestine at least partially by decreasing Cdc42 expression.

FIG 5.

HuR deletion in IECs delays mucosal repair in the small intestine by decreasing Cdc42 levels. (A) Micrographs of the small intestinal mucosa taken after sham treatment or I/R treatment (mesenteric ischemia for 30 min, followed by reperfusion for 2 h) or 6 h post-I/R treatment for littermates (a) and IE-HuR^−/− mice (b). Arrows indicate injury and erosion. Scale bars, 50 μm. (B) Quantitative of the mucosal injury data in mice described for panel A. Values are the means ± the SEM (n = 5). * and +, P < 0.05 compared with sham treatment and I/R treatment, respectively. (C) Immunoblots of Cdc42, Rac1, and RhoC in WT littermate and IE-HuR^−/− mice exposed to I/R. (D) Quantitative analysis of the immunoblots described in panel C as densitometry (n = 5). *, P < 0.05 (compared to sham-treated animals).

HuR deletion in IECs enhances DSS-induced colitis.

To determine the role of HuR in the regulation of mucosal inflammatory damage in the colon, dextran sulfate sodium (DSS)-induced acute colitis was used as a model. After administration of 3% DSS in drinking water for 5 days, most of the littermates and IE-HuR^−/− mice effectively developed acute colitis, as characterized by epithelial damage, granulocyte infiltration, and bloody diarrhea. As shown in Fig. 6A, however, generally reducing HuR expression in IECs enhanced DSS-induced mucosal injury in the colon, since mucosal inflammation and ulceration increased significantly in IE-HuR^−/− mice compared to control littermates. The sum of histological scores in the colonic mucosa of IE-HuR^−/− mouse was ∼2-fold that observed in littermates when measured immediately after a 5-day DSS treatment (Fig. 6B, middle). There were no significant differences in body weight after treatment with DSS for 5 days, but both littermates and IE-HuR^−/− mice began to lose weight after day 6 of DSS treatment, particularly in HuR-knockout animals (see Fig. S3 in the supplemental material). Therefore, mice were treated with DSS for 5 days in studies examining the effect of HuR deletion on mucosal recovery after injury. In control littermates, the mucosa repaired quickly after starting of water drinking, and histological features of the mucosa were restored to near normal from day 5 onward. However, significant granulocyte infiltration and erosions remained in the mucosas of IE-HuR^−/− mice at the same time after administration of water (Fig. 6A, right). Unlike the expression pattern of Cdc42 in response to I/R injury in the small intestine, DSS-induced mucosal injury in the colon did not increase the levels of Cdc42 protein in IE-HuR^−/− mice or littermates, although the basal levels of colonic mucosal Cdc42 decreased dramatically in IE-HuR^−/− mice compared to those in littermates. These results indicate that HuR deletion not only enhances DSS-induced colitis but also delays recovery from this challenge, suggesting that HuR-mediated Cdc42 expression is required, rather than sufficient, for stimulation of colonic mucosal protection and repair.

FIG 6.

HuR deletion represses the recovery of colonic mucosal injury. (A) Micrographs of the colonic mucosa in littermates (a) and in IE-HuR^−/− mice treated with water (control), 3% DSS for 5 day (DSS), and 3% DSS for 5 days, followed by 5-day water feeding (post DSS) (b). Arrows indicates mucosal inflammatory injury. Scale bars, 50 μm. (B) Quantitation of the mucosal injury data in mice described in panel A. Values are means ± the SEM (n = 5). *, P < 0.05 (compared to control animals).

Cdc42 silencing prevents HuR-mediated stimulation of IEC migration over the wounded area in vitro.

To determine whether HuR-regulated Cdc42 regulates mucosal repair by altering cell migration after wounding, an in vitro model that mimics the early cell division-independent stage of epithelial restitution (4, 36) was used. As shown previously (10, 21), a significant increase in cell migration over the denuded area occurred quickly after wounding. Cell proliferation did not participate in this rapid process, since there were no changes in DNA synthesis when measured 6 h after wounding and inhibition of cell division by treatment with mitomycin C (2 μg/ml) failed to alter the rate of cell migration (data not shown). Cdc42 silencing by transfection with siCdc42 (Fig. 7A) inhibited this early rapid epithelial restitution; the numbers of cells migrating over the denuded area after wounding decreased significantly in siCdc42-transfected cells compared to cells transfected with C-siRNA (Fig. 7B and C). Cdc42 silencing also prevented HuR-induced stimulation of cell migration after wounding (Fig. 7D). Consistent with our previous findings (21), ectopically expressed HuR following infection with the AdHuR increased the numbers of cells migrating over the wounded area, but this HuR-induced induction in cell migration was abolished by Cdc42 silencing. The numbers of cells migrating over the wounded area in cells cotransfected with the AdHuR and siCdc42 were similar to those observed in the control group (Fig. 7E). On the other hand, ectopic expression of Cdc42 in HuR-silenced cells restored cell migration (data not shown). Interestingly, reducing the levels of cellular Cdc42 by HuR silencing (Fig. 2A) decreased the formation of F-actin stress fibers in migrating cells after wounding. A network of long stress fibers that were just beneath the plasma membrane and traversed the cytoplasm was observed in control group (Fig. 7Fa), but these long F-actin stress fibers were sparse or disappeared in HuR-silenced population of cells. There were no distinct F-actin stress fibers in the cytoplasm of siHuR-transfected cell after wounding (Fig. 7Fb). As reported previously (37, 38), Cdc42 silencing by transfection with siCdc42 also decreased the formation of F-actin stress fibers in IECs, similar to what was observed in HuR-silenced cells (data not shown). In sum, these results strongly suggest that control of Cdc42 translation by HuR critically regulates cell migration by modulating the subcellular reorganization of cytoskeleton F-actin after wounding.

FIG 7.

Cdc42 silencing inhibits IEC migration over the denuded area in vitro. (A) Immunoblots of Cdc42 in IEC-6 cells transfected with control siRNA (C-siRNA) or siRNA targeting Cdc42 mRNA (siCdc42) and analyzed 48 h later. (B) Images of cell migration after wounding. (a) C-siRNA; (b) siCdc42. The monolayer was wounded, and plates were photographed immediately or 6 h thereafter. (C) Summarized data in cells described in panel B. Values are means ± the SEM (n = 6). *, P < 0.05 (compared to C-siRNA). (D and E) Cdc42 silencing prevents HuR-mediated stimulation of cell migration after wounding. Cells were transfected with HuR expression vector alone or cotransfected with HuR and siCdc42. After 48 h, the cultures were wounded, and cell migration was examined 6 h after that. Values are means ± the SEM (n = 6). * and +, P < 0.05 (compared to vector alone and HuR [with C-siRNA], respectively). (F) Distribution of F-actin in the migrating edge of cells transfected with siHuR. After 48 h, the cells were fixed, permeabilized, and incubated with anti-F-actin antibody and then with anti-IgG conjugated with FITC. Scale bars, 50 μm. Three experiments were performed that showed similar results.

DISCUSSION

Cdc42 belongs to the Rho family of small GTPases and plays a pivotal role in actin organization, cell migration, and proliferation (37–39), but the exact mechanisms that control cellular Cdc42 abundance are largely unknown. In this study, we show that the Cdc42 mRNA is a novel target of HuR and that HuR positively regulates Cdc42 expression at the posttranscriptional level. As shown, HuR interacted with the Cdc42 mRNA via its 3′ UTR, and this association enhanced Cdc42 translation without effect on its total mRNA level. The present study also provides powerful genetic evidence showing the importance of HuR-mediated Cdc42 expression in the regulation of early intestinal mucosal restitution after acute injury. Intestinal epithelium-specific deletion of HuR not only decreased the levels of mucosal Cdc42, but it also delayed epithelial repair in the small intestine after mesenteric I/R-induced injury. Conditional HuR deletion also exacerbated DSS-induced colitis and inhibited its recovery. In an in vitro model, Cdc42 silencing prevented HuR-induced stimulation of IEC migration over the wounded area after injury. These findings advance our understanding of the molecular mechanisms underlying the regulation of Cdc42 expression and highlight the biological function of HuR in rapid early intestinal epithelial restitution after injury by altering Cdc42 translation.

Through the use of various ectopic reporters spanning the Cdc42 5′ UTR, CR and 3′ UTR with or without the predicted HuR-hit motif, we found that HuR did not associate with the 5′ UTR or CR of the Cdc42 mRNA, but it interacted directly with the Cdc42 3′ UTR that contains four predicted HuR motifs. Within the Cdc42 3′ UTR, both segments F1 and F2 were found to bind to HuR and mediate HuR actions, because both the repression of Cdc42 by HuR silencing and the stimulation of the reporter activity by HuR overexpression occurred only when cells were transfected with the Cdc42-3′UTR luciferase reporter constructs but not with the Cdc42-5′ UTR or Cdc42-CR reporter constructs. These observations are consistent with our previous results (32, 40, 41) and work from other labs (31, 42) that show that HuR commonly interacts with many of its target mRNAs via their 3′ UTRs, in turn affecting their stability and/or translation. In some instances, HuR also associates with the CRs of target mRNAs for its regulatory actions. For example, HuR stabilizes Xiap mRNA by interacting with both Xiap CR and with Xiap 3′ UTR (43). Moreover, HuR silencing specifically decreased the abundance of Cdc42 but did not alter the expression levels of other GTPases (specifically Rac1, RhoA, RhoB, or RhoC), whereas ectopically overexpressed HuR increased Cdc42 levels. Neither HuR silencing nor HuR overexpression changed total Cdc42 mRNA levels, indicating that the HuR/Cdc42 mRNA association only enhances Cdc42 translation but has no effect on Cdc42 mRNA stability. We did not further characterize the specific nucleotides through which HuR interacts with and increases Cdc42 mRNA translation, since those experiments would require more specialized biochemical, crystallographic, and molecular methods than those used here.

The most important finding of this study is likely the discovery that HuR is a physiological regulator of early intestinal mucosal restitution after acute injury, an effect that is mediated, at least in part, via Cdc42. Our results show that IE-HuR^−/− mice exhibited a specific inhibition of Cdc42 translation as evidenced by a decrease in the levels of Cdc42 protein without any change in its mRNA levels in the intestinal mucosal tissue. These in vivo observations are consistent with the findings obtained from cultured IECs (Fig. 2) and show that decreasing the levels of HuR specifically decreased Cdc42 abundance in the mucosa but failed to alter the expression levels of Rac1. Studies in genetic mouse models in which HuR levels are altered in specific tissues/organs have improved our understanding of the physiological roles of HuR in vivo, in some cases contradicting the predictions from in vitro cell biology studies (24, 44, 45). However, in this instance, the in vivo results are supported by previous in vitro findings (21, 32, 33) and demonstrate the essential role of HuR in the maintenance of the intestinal epithelial integrity under pathological stress. We observed significant defects in mucosal repair in IE-HuR^−/− mice, as indicated by a delayed recovery of damaged mucosa induced by mesenteric I/R in the small intestine and by DSS in the colon. The inhibition of the response to repair the intestinal mucosa after conditional HuR deletion in IECs was in keeping with the upregulation by HuR of several cell migration- and proliferation-promoting proteins (21, 29, 46, 47). Our previous studies (21, 32, 33) reveal that HuR stabilizes stim1 mRNA, encoding the Ca²⁺ sensor protein STIM1 (activator of Ca²⁺ influx in IECs), and also stimulates translation of the transcription factor MYC, two proteins that critically stimulate IEC migration and proliferation and enhance gut mucosal repair after injury (4, 48, 49). In support of our current findings, targeted deletion of HuR in IECs decreases the regenerative potential of crypt progenitors in mice exposed to irradiation (29) and represses the recovery of the intestinal barrier function after I/R stress (19). However, the exact roles of RhoB and RhoC in inhibition of intestinal epithelial restitution after injury resulting from HuR deletion remain unclear. The expression patterns of these two GTP-binding proteins in cultured IECs after silencing HuR are inconsistent with those observed in IE-HuR^−/− mice. Although silencing HuR in cultured IECs increased RhoB slightly, target deletion of HuR in mice modestly decreased RhoB. Moreover, HuR silencing did not alter the abundance of RhoC in cultured IECs, while RhoC levels increased in HuR-deficient intestinal mucosa.

In summary, conditional HuR deletion in IECs inhibits early intestinal epithelial restitution at least in part by altering the distribution of F-actin cytoskeleton as a result of lowering Cdc42 abundance. Cultured IECs migrate rapidly over the wounded area after superficial injury (50), which is a coordinated process and depends on many regulatory factors and local environment, including extracellular matrix composition, interactions with other cells, and chemical stimuli (51, 52). Rho-family GTPases act as molecular switches by cycling between inactive GDP and active GTP-bound states of GTPases to regulate the activity of F-actin and are critical for the stimulation of cell migration in response to various stresses (53). In humans, there are >20 members of the Rho family, but the relative contribution of each Rho GTPase depends on the environment and cell type (54, 55). The role of three GTPases, Cdc42, Rac1, and RhoA, in the coordination of the actin cytoskeleton has been extensively investigated in different cell types, including IECs. Cdc42 and Rac1 enhance de novo actin polymerization to form filopodia and lamellipodia, respectively, whereas RhoA functions to rearrange existing actin structures to form stress fibers (56). Decreasing the levels of cellular GTPases or inactivation of their enzyme activity inhibits IEC migration and delays early restitution of the intestinal epithelium after wounding (57, 58). In this study, Cdc42 expression levels decreased specifically in HuR-silenced IEC-6 cells and in the intestinal mucosal tissue of IE-HuR^−/− mouse. Furthermore, Cdc42 silencing or HuR silencing to lower Cdc42Cdc42 decreased the formation of F-actin stress fibers in migrating cells and inhibited epithelial restitution after wounding in vitro. Collectively, our results indicate that HuR enhances early restitution of the intestinal epithelium after acute injury by increasing Cdc42 translation and suggest that deficiencies in HuR levels or function contribute to pathogenic conditions of delayed mucosal healing.

MATERIALS AND METHODS

Chemicals and cell cultures.

The culture medium and fetal bovine serum were purchased from Invitrogen (Carlsbad, CA), and the biochemicals were from Sigma (St. Louis, MO). The antibodies recognizing Cdc42, HuR, Rac1, RhoA, RhoB, RhoC, and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA) and BD Biosciences, and the secondary antibody conjugated to horseradish peroxidase was from Sigma. The IEC-6 cells, derived from normal rat intestinal crypt cells, were purchased from the American Type Culture Collection at passage 13 and maintained under standard culture conditions (32). Passages 15 to 20 were used in experiments, and there were no significant changes of biological function and characterization of IEC-6 cells at passages 15 to 20 (21, 29).

Plasmid construction.

Recombinant adenoviral plasmids containing human HuR cDNA (AdHuR) were constructed by using the Adeno-X expression system (Clontech) as described previously (34). The chimeric firefly luciferase reporter construct containing Cdc42 mRNA was generated as described previously (33). The full-length Cdc42 5′-UTR, CR, and 3′-UTR fragments were subcloned into the pmirGLO Dual-Luciferase miRNA target expression vector (Promega, Madison, WI) to generate the pmirGLO-Luc-Cdc42-CR, pmirGLO-Luc-Cdc42-5′UTR, and pmirGLO-Luc-Cdc42-3′UTR reporter constructs as described previously (29, 33). Transient transfections were conducted using the Lipofectamine reagent as recommended by the manufacturer, and the levels of firefly luciferase activity were normalized to Renilla luciferase activity. All of the primer sequences for generation these constructs are provided in Table S1 in the supplemental material.

Western blot analysis.

Whole-cell lysates were prepared using 2% sodium dodecyl sulfate (SDS), sonicated, and centrifuged at 4°C for 15 min. The supernatants were boiled and size fractionated by SDS-PAGE. After the blots were incubated with primary and secondary antibodies, immunocomplexes were developed by using chemiluminescence.

RT, followed by conventional PCR analysis and real-time Q-PCR analysis.

Total RNA was isolated by using RNeasy minikit (Qiagen, Valencia, CA) and used in reverse transcription and PCR amplification reactions as described previously (12). The levels of Gapdh PCR product were assessed to monitor the evenness in RNA input in RT-PCR samples. RT-Q-PCR analysis was performed using 7500-Fast real-time PCR systems with specific primers, probes, and software (Applied Biosystems, Foster City, CA).

Biotin pulldown assays and RIP analysis.

The synthesis of biotinylated transcripts and measurement of HuR bound to biotinylated RNA were performed as previously described (34, 35). cDNA from IEC-6 cells was used as a template for PCR amplification of the 5′ UTR, CR, and 3′ UTR of Cdc42 mRNA. The 5′ primers contained the T7 RNA polymerase promoter sequence (T7, CCAAGCTTCTAATACGAC-TCACTATAGGGAGA). All sequences of oligonucleotides for preparation of the full-length Cdc42 5′ UTR, CR, and various fragments of 3′ UTR are described in Table S1 in the supplemental material. PCR-amplified products were used as the templates to transcribe biotinylated RNAs by using T7 RNA polymerase in the presence of biotin-CTP, as described previously (21). Biotinylated transcripts were incubated with cytoplasmic lysates for 30 min at room temperature. Complexes were isolated with paramagnetic streptavidin-conjugated Dynabeads (Dynal, Oslo, Norway) and analyzed by Western blotting with anti-HuR antibody.

To assess the association of endogenous HuR with endogenous Cdc42 mRNA, immunoprecipitation (IP) of RNP complexes was carried out as described previously (32, 34). Twenty million cells were collected per sample, and lysates were used for IP for 4 h at room temperature in the presence of excess (30 μg) IP antibody (IgG or anti-HuR). RNA in IP materials was used in RT reactions, followed by PCR and Q-PCR analysis to measure the levels of Cdc42 and Gapdh mRNAs.

Assays of newly translated protein and polysome analysis.

New synthesis of nascent Cdc42 protein was detected by using a Click-iT protein analysis detection kit (Life Technologies, Grand Island, NY) and conducted according to the company's manual. Briefly, cells were incubated in methionine-free medium and then exposed to l-azidohomoalanine (AHA). After mixing cell lysates with the reaction buffer for 20 min, the biotin-alkyne/azide-modified protein complex was pulled down using paramagnetic streptavidin-conjugated Dynabeads. The pulldown material was resolved by SDS–10% PAGE and analyzed by Western immunoblotting analysis with antibodies to Cdc42 or GAPDH.

Polysome analysis was carried out as described previously (59). Briefly, cells at ∼70% confluence were incubated in 0.1 mg/ml cycloheximide and then lifted by scraping in polysome extraction lysis buffer. Nuclei were pelleted, and the resulting supernatant was centrifuged through a 10 to 50% linear sucrose gradient to fractionate cytoplasmic components according to their molecular weights. The eluted fractions were prepared with a fraction collector (Brandel, Gaithersburg, MD), and their quality was monitored at 254 nm using a UV-6 detector (ISCO, Louisville, KY). After RNA in each fraction was extracted, the levels of each individual mRNA were quantified by Q-PCR analysis in each of the fractions.

Measurement of epithelial repair in vitro.

Epithelial injury model and repair assays were performed as described previously (4). Cells were plated at 6.25 × 10⁴/cm² in DMEM containing FBS on 60-mm dishes thinly coated with Matrigel according to the manufacturer's instructions (BD Biosciences) and incubated as described previously for stock cultures (21). Cells were fed on day 2, and the monolayer was wounded by removing part of the monolayer on day 4; repair was assayed 6 h after wounding by using National Institutes of Health (NIH) image analysis. All experiments were carried out in triplicate, and the results were reported as the percentage of wound width covered.

Animal studies.

All animal experiments were performed in accordance with NIH guidelines and were approved by the Institutional Animal Care and Use Committee of University Maryland School of Medicine and Baltimore VA Hospital. The HuR^flox/flox (HuR^fl/fl) and villin-Cre mice were purchased from the Jackson Laboratory, and intestinal epithelial tissue-specific HuR deletion (IE-HuR^−/−) mice were generated by crossing the HuR^fl/fl mouse with a villin-Cre mouse, as described in our previous studies (29, 44). HuR^fl/fl-Cre⁻ mice served as littermate controls. Both IE-HuR^−/− mice and control littermates were housed and handled in a specific-pathogen-free breeding barrier and cared for by trained technicians and veterinarians. To generate the model of intestinal I/R-induced injury, mice were anesthetized by pentobarbital sodium (Nembutal) (5.5 mg/100 g [body weight], administered intraperitoneally), and a midline abdominal incision was performed. Mice were exposed to 30-min superior mesenteric artery ischemia, followed by reperfusion for 2 h as described previously (60). Sham operation for controls only involved laparotomy without mesenteric ischemia. Acute colitis was induced by feeding IE-HuR^−/− mice and control littermates with 3% dextran sulfate sodium (DSS; MP Biomedicals) dissolved in drinking water for 5 consecutive days as reported previously (61). The colonic mucosal tissues were isolated at various times after starting of water drinking for hematoxylin and eosin (H&E) staining.

Histological analysis.

Dissected and opened intestines were mounted onto a solid surface and fixed in formalin and paraffin. Sections 5 μm thick were stained with H&E for general histology. Slides were examined in a blinded fashion by coding them, and only after examination was complete were they decoded. Microscopic damages in the small intestinal mucosa in mice exposed to I/R stress were measured and semiquantified as described previously (10, 62), while histological colitis scoring system was used to determine the extent of the inflammation and damage in the colon of mouse treated with DSS using a previously published grading system (61).

Statistical analysis.

All values were expressed as the means ± the standard errors of the mean (SEM). An unpaired, two-tailed Student t test was used when indicated, with P values <0.05 considered significant. When assessing multiple groups, one-way analysis of variance (ANOVA) was utilized with Tukey's post hoc test (63). The statistical software used was SPSS17.1.