The Endoplasmic Reticulum Stress Sensor IRE1α in Intestinal Epithelial Cells Is Essential for Protecting against Colitis*

Background: Endoplasmic reticulum (ER) stress is implicated in inflammatory bowel disease (IBD) and IRE1α plays a critical role in ER stress. Results: Genetic ablation of Ire1α in intestinal epithelial cells leads to colitis in mice. Conclusion: IRE1α acts as an important defense molecule against IBD. Significance: The finding provides insight into the regulation of intestinal epithelium homeostasis by IRE1α. Intestinal epithelial cells (IECs) have critical roles in maintaining homeostasis of intestinal epithelium. Endoplasmic reticulum (ER) stress is implicated in intestinal epithelium homeostasis and inflammatory bowel disease; however, it remains elusive whether IRE1α, a major sensor of ER stress, is directly involved in these processes. We demonstrate here that genetic ablation of Ire1α in IECs leads to spontaneous colitis in mice. Deletion of IRE1α in IECs results in loss of goblet cells and failure of intestinal epithelial barrier function. IRE1α deficiency induces cell apoptosis through induction of CHOP, the pro-apoptotic protein, and sensitizes cells to lipopolysaccharide, an endotoxin from bacteria. IRE1α deficiency confers upon mice higher susceptibility to chemical-induced colitis. These results suggest that IRE1α functions to maintain the intestinal epithelial homeostasis and plays an important role in defending against inflammation bowel diseases.

Inflammatory bowel disease (IBD) 4 is a chronic inflammatory condition with severe pathology (1-3), which includes two major types, Crohn's disease (CD) and ulcerative colitis (UC). The precise mechanism underlying the pathogenesis of IBD is poorly understood. A common feature of IBD is the loss of intestinal epithelial barrier function due to excessive epithelial cell death, which allows the invasion of bacteria into the submucosa, leading to inflammatory response and barrier disruption (4). The barrier relies on the turnover of intestinal epithelial cells (IECs) originating from the stem cells in the crypts. Under the physiologic state, IECs have a relatively short lifespan and their death is tightly controlled, which is critical for the maintenance of normal barrier function (5). Increased apoptosis of IECs with resultant epithelial barrier defect has a key role in the development of IBD (4,6).
The unfolded protein response allows cells to manage endoplasmic reticulum (ER) stress resulting from accumulation of unfolded and misfolded proteins (7,8). Three ER-localized proteins, inositol-requiring kinase/endonuclease 1 (IRE1), pancreatic ER kinase (PERK), and activating transcription factor 6 constitute the three arms of the unfolded protein response to resolve ER stress. IRE1 is the most evolutionally conserved among the three unfolded protein response arms (9,10). In mammals, IRE1 consists of two structurally related proteins, IRE1␣ and IRE1␤ (11). IRE1 possesses both protein kinase and endoribonuclease activities. Under ER stress, IRE1␣ is activated through trans-autophosphorylation and dimerization/oligomerization, removing a 26-bp nucleotide intron from the mRNA encoding X-box binding protein (XBP) 1 (12). This in turn generates a spliced active form XBP1s, which regulates the expression of genes involved in protein folding, ER-associated degradation, protein quality control, and phospholipid synthesis (12). In response to ER stress, PERK is activated and phosphorylates the eukaryotic translation-initiation factor 2␣ (eIF2␣), leading to shutdown of cellular protein translation (13). Simultaneously, the phosphorylated eIF2␣ stimulates the expression of C/EBP homologous protein (CHOP), a proapoptotic protein (13)(14)(15)(16).
Many studies have shown that ER stress is linked to intestinal inflammation (11,(17)(18)(19)(20). Genetic deletion of Xbp1 in IECs caused spontaneous enteritis in mice (19), whereas deletion of IRE1␤ led to higher susceptibility to dextran sulfate sodium (DSS)-induced colitis (11). Further studies show that IRE1␤ optimizes the production of mucin in goblet cells, indicating that IRE1␤ is involved in ER homeostasis in goblet cells (21). It still remains elusive whether or not IRE1␣ is involved in the homeostatic control of the intestinal epithelium. In this article we report that mice in which Ire1␣ is specifically deleted in IECs develop spontaneous colitis and exhibit increased sensitivity to DSS colitis. Our results demonstrate that IRE1␣ acts as an important defense molecule against IBD, playing a critical role in regulating the integrity and homeostasis of intestinal epithelium.

Materials and Methods
Animals-The Villin-Cre transgenic mice expressing Cre recombinase specifically in intestinal epithelium (22) were obtained from the Model Animal Research Center of Nanjing University. Floxed mice (Ire1␣ flox/flox ), in which the 121-nucleotide exon 2 of the Ern1 (i.e. Ire1␣) allele was flanked with two loxP recombination sites, were generated as described (23). Intestinal epithelia-specific Ire1␣ knock-out mice (Ire1␣ flox/flox -Villin-Cre) were produced by intercrossing the Ire1␣ flox/flox mice with Villin-Cre mice. Mice were housed in laboratory cages at 23 Ϯ 3°C with a humidity of 35 Ϯ 5% under a 12-h dark/light cycle. With free access to a regular chow diet (Shanghai Laboratory Animal Co. Ltd, Shanghai), animals were maintained under a specified pathogen-free condition. All protocols of animal experiments were approved by the Institutional Animal Care and Use Committee at the Institute for Nutritional Sciences.
In Vivo Intestinal Permeability Assay-Age-matched female littermates (18 -20 weeks) were orally administered (0.6 mg/g body weight) a FITC-dextran solution (70 kDa, 80 mg/ml). After 4 h, the mice were sacrificed and blood was obtained by cardiac puncture. Plasma was used for FITC measurement by fluorometry (19). The distribution of FITC-dextran in colon tissues was determined by fluorescence microscopy (24).
In Vitro Permeability Assay-Caco-2 cells were cultured on Transwells with polyester membrane insert (Corning) allowing proper cellular polarization with formation of apical (upper compartment) and basolateral face (lower compartment). The insert was pretreated with DMEM overnight before cell plating. Caco-2 cells were seeded at a density of 0.5 ϫ 10 5 cells/insert. The medium were replaced with fresh medium every 2 days. After 18 days, IRE1␣ siRNA transfection was carried out for 4 days as described (25). After transfection, the fresh medium containing FITC-dextran (10 kDa, 10 g/ml) was added to the upper compartment and incubated at 37°C for 4 h. The aliquots from the bottom compartment were examined for FITC-dextran in a spectrophotometer (excitation 485 nm and emission 530 nm).
Tissue Staining-Colon tissues were formalin-fixed and paraffin-embedded, and hematoxylin eosin (H&E) staining was performed. The periodic acid-Schiff (PAS) staining was performed for detection of goblet cells as described (26). Immunohistochemical staining was done as described (27).
Isolation of Colon Epithelial Cells-Colonic epithelial cells were isolated as described (11). Briefly, the colon was removed and washed with solution "A" (96 mM NaCl, 27 mM sodium citrate, 1.5 mM KCl, 0.8 mM KH 2 PO 4 , 5.6 mM Na 2 HPO 4 , 5,000 units/liter of penicillin, 5 mg/liter of streptomycin, 0.5 mM DTT, and 2 mM phenylmethylsulfonyl fluoride, pH 7.4). Square pieces of tissue were placed in 10 ml of solution A at 37°C for 10 min with gentle shaking to remove the mucus, bacteria, and other lumenal contents. The tissue fragments were then incubated in solution "B" (0.1 mM EDTA, 115 mM NaCl, 25 mM NaHCO 3 , 2.4 mM K 2 HPO 4 , 0.4 mM KH 2 PO 4 , 5,000 units/liters of penicillin, 5 mg/liters of streptomycin, 2.5 mM glutamine, 2 mM phenylmethylsulfonyl fluoride, and 0.5 mM DTT, pH 7.4) at 37°C for 30 min with gentle shaking; the disruption of the mucosa and elution of cells was stopped by adjusting to 1 mM CaCl 2 . Tissue fragments were removed, and cells recovered in the suspension were collected.
Cell Culture and Reagents-Human normal colon epithelial CCD841 cells were maintained in RPMI 1640 medium with 10% FBS, 100 units/ml of penicillin, and 100 g/ml of streptomycin. Colon cancer Caco-2 cells were grown in DMEM with 20% FBS and the above antibiotics. Lipofectamine 2000 was from Invitrogen. Lipopolysaccharide (LPS) (Escherichia coli 055:B5) and FITC-dextran (70 and 10 kDa) were purchased from Sigma. DSS was from MP Biomedicals.
Immunoblotting-Immunoblotting was performed in a standard manner. The ICAM-1 and ATF4 antibodies were purchased from Santa Cruz Biotechnology. Antibodies against the cleaved caspase 3, IRE1␣, eIF2␣, phospho-eIF2␣(Ser-51), and CHOP were products of Cell Signaling Technology. ␤-Actin and IRE1␤ antibodies were from Sigma and Abcam, respectively.
Cell Apoptosis Assay-Apoptosis of cultured cells was measured by FACS using the FITC Annexin V Apoptosis Detection Kit (BD Pharmingen). Apoptotic cells in tissues were detected by staining with a TUNEL kit (Promega).
Statistic Analysis-Statistical analysis was conducted using the unpaired two-tailed Student's t test or two-way analysis of variance (ANOVA) with GraphPad Prism 5.0. Data are mean Ϯ S.E. p Ͻ 0.05 is considered statistically significant.
IRE1␣ Deficiency in IECs Impairs the Intestinal Epithelial Barrier Function-Next, we performed H&E staining of the colon tissue section and observed apparent morphological alterations, such as loss of the goblet cells, distortion, or collapse of lamina propria, as well as lymphocyte infiltration in Ire1␣ Ϫ/Ϫ mice relative to their Ire1␣ ϩ/ϩ littermates ( Fig. 2A 1 ). Deletion of IRE1␣ had little effect on colon crypt number, but it increased the crypt length ( Fig. 2A 2 ). Goblet cells synthesize and secrete mucins that contribute the formation of the gel-like matrix, separating the intestinal epithelium from lumen microorganisms (29). Using PAS staining to detect mucosubstances that are normally restricted to goblet cells, we found that the number of PAS-positive cells was reduced in the colon epithelium of Ire1␣ Ϫ/Ϫ mice (Fig. 2B). The mRNA expression of mucin 2 (Muc2) was also decreased in the colon epithelium of Ire1␣-deficient mice (Fig. 2C). Because reduced goblet cells may lead to failure of epithelial barrier function, we tested this by oral FITC-dextran administration in mice. Microphotography of colon tissue sections showed more FITC-dextran that passed through the epithelial barrier in Ire1␣ Ϫ/Ϫ mice (Fig. 2D), whereas most of the dye was retained at the surface of the barrier in Ire1␣ ϩ/ϩ littermates. Consistently, Ire1␣ Ϫ/Ϫ mice had higher serum levels of FITC-dextran than control littermates (Fig. 2D). We also determined the effects of knockdown of IRE1␣ on paracellular permeability of the Caco-2 cell monolayer. Knockdown of IRE1␣ enhanced the pass of FITC-dextran through the monolayer (Fig. 2E). These results suggest that IRE1␣ deficiency results in loss of goblet cells and disrupts the epithelial barrier function of colon.
We then examined whether IRE1␣ deficiency in IECs altered the expression pattern of pro-inflammatory cytokines in colon mucosa. Analyses by real-time PCR showed marked elevations in the expression of Tnf␣, Il-1␤, and Il-6 in colon mucosa of Ire1␣ Ϫ/Ϫ mice as compared with that in Ire1␣ ϩ/ϩ littermates (Fig. 2F). Immunostaining of the colon sections also revealed higher expression of ICAM-1, a marker of activation of immune responses (30), in Ire1␣ Ϫ/Ϫ mice than in Ire1␣ ϩ/ϩ littermates (Fig. 2G). These data indicate an exacerbated inflammatory state as a result of IRE1␣ deficiency in IECs.
Loss of IRE1␣ Promotes Cell Apoptosis in Colon Epithelium-To know how IRE1␣ deficiency impairs colon epithelial homeostasis, we examined Xbp-1 mRNA splicing by real-time PCR in isolated primary colon epithelial cells. The ratio of spliced to unspliced Xbp-1 mRNA was decreased in colon epithelial cells from Ire1␣ Ϫ/Ϫ mice (Fig. 3A). The results suggest that IRE1␤ cannot compensate for the absence of IRE1␣ in the splicing of Xbp1 mRNA. JNK is activated by IRE1␣ (31). We presumed that IRE1␣ deletion might decrease the phosphorylation of JNK. As expected, the phosphorylation levels of JNK in colon epithelial cells of Ire1␣ Ϫ/Ϫ mice were reduced (Fig. 3B). Whereas a decreased Xbp1 mRNA splicing and JNK phosphorylation was detected (Fig. 3, A and B), a dramatic increase of phosphorylation of eIF2␣ was observed (Fig. 3C), implying an activation of PERK signaling. The expression of CHOP, a downstream target of this signaling was also up-regulated in colon epithelial cells of Ire1␣ Ϫ/Ϫ mice (Fig. 3, C and D). This indicates an exacerbated state of ER stress, which may induce cell death (13)(14)(15)(16). TUNEL stain of the colon tissue sections revealed an apparent increase in cell death in colon epithelium of Ire1␣ Ϫ/Ϫ mice (Fig. 3, E and F). In accord with this, increased cleavage of caspase 3 was found in colon epithelial cells of Ire1␣ Ϫ/Ϫ mice (Fig.  3G).
Next, we determined the effect of IRE1␣ knockdown in human normal colon epithelial CCD841 cells. siRNA-directed suppression of IRE1␣ expression resulted in increased eIF2␣ phosphorylation and up-regulated expression of CHOP (Fig.  4A), in parallel with increased apoptotic cells (Fig. 4B), suggesting that IRE1␣ deletion augmented the ER stress-associated death of colon epithelial cells. Similar results were obtained with Caco-2 cells (Fig. 4, A and B). The results prompt us to hypothesize that IRE1␣ in colon epithelium may act as a microbial sensor and defend against the detrimental effects of intes-tinal bacteria. LPS, an endotoxin produced by bacteria, is known to induce ER stress (32)(33)(34). It can directly cause intestinal epithelial cell apoptosis (35) and failure of intestinal barrier (36). Thus, we used CCD841 cells to determine whether IRE1␣ could affect the cytotoxic activity of LPS. Although LPS treatment induced cell apoptosis, knockdown of IRE1␣ further amplified cell apoptosis induced by LPS (Fig. 4B), indicating that IRE1␣ deficiency could sensitize CCD841 cells to LPS challenge. LPS treatment increased phosphorylation of eIF2␣ and expression of CHOP (Fig. 4C), and IRE1␣ knockdown together with LPS enhanced these effects (Fig. 4D). To further investigate the possible mechanism, we considered the involvement of CHOP, the critical molecule that has been known to mediate ER stress-induced apoptosis (13)(14)(15)(16). Notably, knockdown by siRNAs  of CHOP expression (Fig. 4E) blunted significantly the increases of apoptotic CCD841 cells resulting from IRE1␣ knockdown (Fig.  4F). Moreover, the repression of CHOP also abolished the IRE1␣ deficiency enhancement of LPS-promoted apoptosis (Fig. 4G). These results indicate that IRE1␣ can exert cytoprotective actions against toxins from bacteria, likely through managing ER stressinduced apoptosis in colon epithelia.
The above results indicate that IRE1␣ deletion leads to cell apoptosis (Fig. 3, E and F) and decrease of Muc-2 mRNA (Fig.  2C). We determined whether the decrease of Muc-2 mRNA is caused by cell apoptosis. We examined Caco-2 cells that produce Muc2. We found that knockdown of IRE1␣ induced apoptosis and decreased the Muc2 mRNA level (Fig. 4H). Inhibition of cell apoptosis by caspases inhibitor Z-VAD prevented the decrease of Muc-2 mRNA (Fig. 4H).

Discussion
Under ER stress, the IRE1␣-XBP1 pathway plays a pivotal role in restoring ER homeostasis for cell survival. Deletion of Xbp1 in mice IECs led to spontaneous enteritis (19). Although deletion of Ire1␤ did not result in IBD spontaneously, it sensitized mice to DSS-induced colitis (11). These promoted us to determine the possible role of IRE1␣ in intestinal epithelium homeostasis. We demonstrate here that IEC-specific deletion of IRE1␣ in mice caused spontaneous colitis and the female mice with IRE1␣ ablation are more susceptible to colitis. Our findings suggest that IRE1␣ exerts critical actions in maintaining the homeostasis of intestinal epithelium.
Our results show that deletion of IRE1␣ increased eIF2␣ phosphorylation and CHOP expression (Fig. 3), indicating the activation of the PERK pathway, another arm of unfolded protein response. Under persistent ER stress, PERK signaling induces the expression of CHOP (38), a key mediator of ER stress-associated apoptosis (14). It is well established that unresolved ER stress leads to cell apoptosis (13). The Ire1␣ Ϫ/Ϫ mice had more apoptotic cells in the colon epithelium, along with activated PERK signaling (Fig. 3). Thus, cell apoptosis in IRE1␣-deficient colon epithelium might result, at least in part, from the up-regulated expression of CHOP. Supporting this notion, our in vitro experiments showed that suppression of CHOP blunted IRE1␣ deficiency induced cell apoptosis (Fig. 4). These results suggest that IRE1␣ deficiency activates the pro-apoptotic PERK/CHOP pathway. Nonetheless, other mechanisms might also be involved, as chronic ER stress can impair cellular homeostasis through ER calcium leakage, mitochondrial damage, oxidative stress, and caspases activation (13,39). We noted that deletion of IRE1␣ led to a decrease of Xbp-1 splicing in knock-out mice (Fig. 3A). Spliced XBP-1 functions to promote cell survival. Thus, the impaired Xbp-1 splicing may also contribute to the observed phenotypes.
Intestinal bacteria are associated with bowel inflammation (40). Host recognition of bacteria is achieved through communications between epithelial cells and microbial components such as LPS, the main bacterial product that triggers immune responses. LPS may provoke chronic inflammation to damage colon epithelial cells, leading to failure of barrier function. We found that LPS activated eIF2␣ and induced CHOP (Fig. 4). The underlying mechanism is not clear, which needs further study. Our results show that loss of IRE1␣ rendered cells more susceptible to LPS challenge, and repression of CHOP could dampen the IRE1␣ deficiency amplified LPS induction of apoptosis of colon epithelial cells (Fig. 4). Thus, IRE1␣ may act to protect intestinal epithelial cells from LPS-elicited damage, contributing to the intestinal epithelial homeostasis and barrier function. Deletion of IRE1␣ led to increased inflammatory factors including TNF␣ that may also contribute to increased apoptosis of the intestinal epithelial cells.
Goblet cells are critical in epithelial defense against luminal stimulants and pathogens (41,42). They produce and secrete mucins to lubricate and cover the intestinal epithelium. Loss of FIGURE 5. Ire1␣ ؊/؊ mice are more susceptible to DSS-induced colitis. A-E, age-matched (6 -8 weeks) male littermates were treated with DSS (2% in drinking water) for 9 days as described under "Materials and Methods." A, body weight was monitored at the indicated time. Data are mean Ϯ S.E. Ire1␣ ϩ/ϩ (n ϭ 9), Ire1␣ ϩ/Ϫ (n ϭ 11), and Ire1␣ Ϫ/Ϫ (n ϭ 9). p values are indicated for statistic analysis of Ire1␣ ϩ/ϩ versus Ire1␣ Ϫ/Ϫ . B, bleeding score. Data are mean Ϯ S.E. p values are indicated for statistic analysis of Ire1␣ ϩ/ϩ versus Ire1␣ Ϫ/Ϫ . C, measurement of colon length. The upper panel shows representatives of colons and the lower panel shows statistic analysis of colon length. Data are mean Ϯ S.E. D, H&E staining of colon tissue sections. E, PAS staining of colon tissue sections. F-H, antibiotic treatment abrogates the differences in susceptibility to DSS colitis. The experiments were performed as in A, except that a higher dose of DSS (3.5%) was used. Neomycin sulfate (1.5 g/liter) and metronidazole (1.5 g/liter) were added in drinking water over the examination. F, body weight; G, colon length; H, bleeding score. Ire1␣ ϩ/ϩ (n ϭ 5), Ire1␣ Ϫ/Ϫ (n ϭ 4). I, body weight recovery for mice after DSS exposure. Age-matched male mice (6 -8 weeks) were exposed to 2.5% DSS for 5 days, followed by normal water (n ϭ 4 per genotype). Body weight was monitored. Two-way analysis of variance was employed for statistic analysis. Scale bar ϭ 100 m.
goblet cells could be one of the causes of barrier dysfunction. In human IBD and animal colitis, disease progression is associated with the depletion of goblet cells (43,44). We show here that deletion of IRE1␣ resulted in reduction of goblet cells (Fig. 2). This might be, at least in part, attributable to the exacerbated ER stress caused by IRE1␣ deficiency.
IBD is known to occur in genetically susceptible individuals under certain influences of environmental factors (45). ER stress is linked to the pathogenic progression of IBD (46). Single-nucleotide polymorphisms within the XBP1 gene have been shown to confer a higher risk for IBD in the human population (19). Similarly, studies of potential genetic association of IRE1␣ with IBD will provide valuable insights into the role of ER stress in the pathogenic progression of this disease.
Our results show that genetic ablation of Ire1␣ in IECs activates PERK signaling and leads to colitis in mice, which supports the consensus that ER stress is linked to intestinal inflammation. IECs are emerging as critical mediators of inflammatory and immune responses in mucosal tissues. Our findings demonstrate that IRE1␣ is essential for epithelial protection, suggesting that it may serve as a key regulator of intestinal epithelium homeostasis. IRE1␣ loss in IECs causes persistent and excessive ER stress that can trigger inflammation and disrupts intestinal epithelial functions. Further elucidation of the mechanisms by which IRE1␣ protects the intestinal epithelial integrity and homeostasis may open avenues for new therapeutic strategies against IBD.