Rack1 maintains intestinal homeostasis by protecting the integrity of the epithelial barrier

Abstract

Previously, we generated mouse models of Rack1 deficiency to identify key functions for Rack1 in regulating growth of intestinal epithelia: suppressing crypt cell proliferation and regeneration, promoting differentiation and apoptosis, and repressing development of neoplasia. However, other than low body weight, we did not detect an overt phenotype in mice constitutively deleted of Rack1 in intestinal epithelia (vil-Cre:Rack1^fl/fl mice), presumably because Rack1 was deleted in <10% of the total surface area of the epithelia. To assess the effect of Rack1 loss throughout the entire intestinal epithelia, we generated another mouse model of Rack1 deficiency, vil-Cre-ER^T2:Rack1^fl/fl. Within 5–10 days of the initial tamoxifen treatment, the mice lost over 20% of their body weight, developed severe diarrhea that for some was bloody, became critically ill, and died, if not euthanized. Necropsies revealed mildly distended, fluid-, gas-, and sometimes blood-filled loops of small and large bowel, inguinal lymphadenopathy, and thrombocytosis. Rack1 was deleted in nearly 100% of the epithelia in both the small intestine and colon when assessed by immunofluorescent or immunoblot analyses. Rack1 expression in other tissues and organs was not different than in control mice, indicating tissue specificity of the recombination. Histopathology revealed a patchy, erosive, hemorrhagic, inflammatory enterocolitis with denuded, sloughed off surface epithelium, and crypt hyperplasia. These results suggest a protective function for Rack1 in maintaining the integrity of intestinal epithelia and for survival.

NEW & NOTEWORTHY Our findings reveal a novel function for Rack1 in maintaining intestinal homeostasis by protecting the epithelial barrier. Rack1 loss results in a patchy, erosive, hemorrhagic, inflammatory enterocolitis, which resembles that of inflammatory bowel diseases (IBD) in humans. Understanding mechanisms that protect barrier function in normal intestine and how loss of that protection contributes to the pathogenesis of IBD could lead to improved therapies for these and other erosive diseases of the gastrointestinal tract.

INTRODUCTION

Rack1 is an evolutionarily conserved member of the tryptophan-aspartate repeat (WD-repeat) family of proteins and shares significant homology with the β-subunit of G proteins (reviewed in Refs. 1, 6, 13). It has seven β-propeller blades, which serve as binding sites for multiple interacting partners, enabling Rack1 to act as a scaffolding protein where key signaling complexes assemble. Rack1 interacts with the cytoplasmic tails of cell surface receptors, facilitating cross talk and leading to activation/deactivation of downstream kinases that control biologic processes. Rack1 also interacts with the ribosomal machinery and shuttles proteins within the cell, anchoring them at specific subcellular locations and regulating their activity. As a result, Rack1 is a key mediator of various pathways and contributes to numerous aspects of cellular function. In vitro studies show that Rack1 is involved in diverse processes including protein translation, cell growth, cell cycle progression, cytokinesis, apoptosis, cell survival, cell adhesion, migration, and stress responses (reviewed in Refs. 1, 6, 13).

Our in vitro studies show that Rack1 regulates growth of colon cells, partly by inhibiting Src activity at key cell cycle checkpoints, in apoptotic and cell survival pathways, and at cell-cell adhesions (14–18, 21).

However, how Rack1 functions in vivo in intestinal epithelia of higher animals is an unanswered and important question with broad and profound implications to the regulation of cell growth and death during health and disease.

We hypothesized that Rack1 regulates growth of intestinal epithelial cells in vivo, as it does in vitro. To test this, we generated mouse models of Rack1 deficiency (vil-Cre:Rack1^fl/fl and CAG-Cre-ER^TM:Rack1^fl/fl) and identified key functions for Rack1 in regulating growth of intestinal epithelia: suppressing crypt cell proliferation and regeneration, promoting differentiation and apoptosis, and repressing development of neoplasia (5).

However, other than low body weight, we did not detect an overt phenotype in mice constitutively deleted of Rack1 in intestinal epithelia (vil-Cre:Rack1^fl/fl mice), presumably because Rack1 was deleted in <10% of the total surface area of the epithelia. In an attempt to assess the phenotype of Rack1 loss and thereby the function of Rack1 in intestinal epithelia, we generated another mouse model, vil-Cre-ER^T2:Rack1^fl/fl, where Rack1 was inducibly deleted only in intestinal epithelia and in nearly 100% of the epithelia throughout the small intestine and colon. Rack1 loss results in a patchy, erosive, hemorrhagic, inflammatory enterocolitis. We identified a protective function for Rack1 in maintaining the integrity of intestinal epithelia and for survival.

MATERIALS AND METHODS

Mice.

Mice were bred and maintained at the Stanford Veterinary Service Center. The animal protocol and procedures for the studies were approved by the Stanford Institutional Animal Care and Use Committee, known as the Administrative Panel on Laboratory Animal Care.

Generation of mice with Rack1 deleted in intestinal epithelia.

To target Rack1 deletion to intestinal epithelia, floxed-Rack1 mice (5) were bred to those expressing a tamoxifen-dependent Cre under the control of the Villin promoter (vil-Cre-ER^T2; Ref. 7) to generate vil-Cre-ER^T2:Rack1^fl/fl mice. Offspring were examined by PCR analyses of tail DNA for the presence of vil-Cre-ER^T2 and floxed-Rack1.

Vil-Cre-ER^T2:Rack1^fl/fl mice and littermates (vil-Cre-ER^T2:Rack1^+/fl, Rack1^+/+, Rack1^+/fl, and/or Rack1^fl/fl mice) ranging in age from 1.0 to 6.8 mo (all but 1 were between 1.0 and 4.5 mo, and the average age was 2.9 mo), were treated with tamoxifen at a dose of 0.2 g/kg via oral gavage every 2–3 days, unless otherwise stated (2). Mice were euthanized at various time points after initiation of tamoxifen feeding (range: 4−14 days later) and after receiving two to four total doses (usually 3). Necropsies were performed by board-certified veterinarians in the Stanford Veterinary Service Center.

Isolation and lysis of intestinal epithelial cells, tissue homogenization, and immunoblot analysis.

To assess Rack1 expression in intestinal epithelial cells, intestines were opened longitudinally, rinsed twice in cold PBS, and then incubated in cold PBS supplemented with 10 mM EDTA on ice for 30 min. Epithelial sheets were then released by vigorously agitation in PBS for 5 min and collected by centrifugation at 300 g for 2 min after removal of the intestinal muscle layers with forceps (22). Release of epithelial sheets, including crypts and villi, was confirmed microscopically. Cells were lysed in NP-40 buffer (4). To assess Rack1 expression in pancreas and liver, organs were minced and homogenized in NP-40 buffer. All lysates were cleared by centrifugation at 15,000 rpm for immunoblot analyses. Immunoblot analyses for Rack1, tubulin, and GAPDH were performed following standard techniques for detection and analyses using the Odyssey scanner and the Odyssey Image Studio software.

Tissue harvesting and immunofluorescent and histopathologic analyses.

Mice were euthanized in a standard chamber with carbon dioxide administered through a controlled gas cylinder. Intestines were removed en bloc, flushed intraluminally once with cold PBS unless otherwise indicated, coiled into Swiss rolls, and fixed in phosphate-buffered formalin (10%) for 24 h. Fixed tissues were embedded in paraffin and cut into 4-μm sections and stained with hematoxylin and eosin (H&E) or periodic acid-Schiff (PAS) by the Stanford Comparative Medicine Animal Histology Service Center. For immunofluorescent (IF) staining, antigen retrieval was carried out by soaking deparaffinized and rehydrated sections in subboiling 10 mM Tris base/1 mM EDTA buffer (pH 9) for 10 min. IF staining was otherwise carried out using standard techniques. Light and fluorescent images were captured using a Nikon Eclipse E600 microscope.

Reagents.

Primary antibodies were as follows: mouse monoclonal antibody (mAb) Rack1 (B-3) and rabbit mucin 2 (H-300; Santa Cruz Biotechnology), mouse mAb PCNA (BD Transduction Laboratories), mouse mAb α-tubulin (T6074; Sigma-Aldrich), rabbit mAb Rack1 (EPR7388), synaptophysin (EP1098Y; Abcam), rabbit lysozyme (180039; Invitrogen), β-tubulin (ab6046), and mAb GAPDH (ab9484; Abcam). Secondary antibodies were as follows: Alexa Fluor 680 goat-anti-mouse IgG (H⁺L) highly cross-adsorbed secondary Ab (Invitrogen, Molecular Probes), IRDye800-conjugated affinity purified anti-rabbit IgG (H&L, goat; Rockland), and Alexa Fluor 488 goat anti-mouse IgG, 488 goat anti-rabbit IgG, 568 goat anti-rabbit IgG, 594 goat anti-rat IgG, and 594 goat anti-mouse IgG (Invitrogen, Molecular Probes). Other reagents were Hoechst 33342, VectorShield mounting medium (Vector Laboratories), and Tamoxifen base (Sigma).

RESULTS

Inducible deletion of Rack1 throughout the entire intestinal epithelia results in weight loss, diarrhea, and death.

Other than slightly low body weight, we did not detect a phenotype in mice that were constitutively deleted of Rack1 in intestinal epithelia (vil-Cre:Rack1^fl/flor Rack1^−i/−i), presumably because Rack1 was deleted in <10% of the total surface area of the epithelia (5).

In an attempt to further study the phenotype of Rack1 loss in just intestinal epithelia, we generated another mouse model of Rack1 deficiency: vil-Cre-ER^T2:Rack1^fl/fl or inducible Rack1^−i/−i. Within 2–3 days of the initial tamoxifen treatment, the vil-Cre-ER^T2:Rack1^fl/fl mice began losing weight (Fig. 1A) and developed diarrhea (Fig. 1B). By day 6 most had lost 20% of their body weight (Fig. 1A) and had severe diarrhea (Fig. 1B), which for some became bloody. By days 5–10, they became critically ill and died, if not euthanized. Tamoxifen-fed littermate controls (vil-Cre-ER^T2:Rack1^+/fl, Rack1^fl/fl, Rack1^+/fl, and Rack1^+/+ mice) did not suffer significant weight loss or diarrhea (Fig. 1, A and B).

Fig. 1.

Inducible deletion of Rack1 throughout the entire intestinal epithelia results in weight loss and diarrhea. vil-Cre-ER^T2:Rack1^fl/fl mice and littermates (vil-Cre-ER^T2:Rack1^+/fl, Rack1^+/+, Rack1^+/fl, and/or Rack1^fl/fl) were fed tamoxifen (0.2 g/kg) via oral gavage every 2–3 days, as described in materials and methods. A: change in body weight following tamoxifen induction. Mice were weighed daily; body weight is expressed relative to the weight on the day that tamoxifen feeding was initiated. B: change in stool water content following tamoxifen induction. Stool was collected daily in individually caged mice, weighed, and dried overnight at 95°C. The percentage of stool water content was calculated as a ratio of wet minus dry weight to wet weight. A and B: data shown represent the mean values ± SE for tamoxifen-fed vil-Cre-ER^T2:Rack1^fl/fl mice (ko; n = 8), littermate controls: vil-Cre-ER^T2:Rack1^+/fl mice (het; n = 7), or wild-type mice (wt; Rack1^+/+, Rack1^+/fl or Rack1^fl/fl, n = 5). C and D: effect of tamoxifen feeding of vil-Cre-ER^T2:Rack1^fl/fl mice on Rack1 expression in intestinal epithelia and other organs of the digestive tract. Epithelial fractions of small intestine or colon (C) or tissue lysates of pancreas or liver (D) were subjected to immunoblot analyses using antibodies that recognize Rack1 (36 kDa) and tubulin, as described in materials and methods. C: analyses of Rack1 levels in epithelial fractions from the entire small intestine (SB) of vil-Cre:Rack1^+/fl control mice (C1 and C2), or from regions of small intestine (duodenum, jejunum, or ileum) or colon of a vil-Cre-ER^T2:Rack1^fl/fl mouse (KO) or its littermate controls: vil-Cre-ER^T2:Rack1^+/fl (C3) or Rack1^+/fl (C4). Total lysate protein loaded: SB, duodenum, jejunum, or ileum: 20 μg; colon: 2.5 μg. While the epithelial fractions are highly enriched for epithelial cells, some cells from the lamina propria could be present. Quantitative analyses of Rack1 levels in epithelial fractions from small intestine are expressed relative to those of the first vil-Cre:Rack1^+/fl control (C1), and Rack1 levels in epithelial fractions from colon are expressed relative to those of the vil-Cre-ER^T2:Rack1^+/fl littermate control (C3). For each of 2 experiments, 4 sections of intestine were analyzed (duodenum, jejunum, ileum, and colon) from a vil-Cre-ER^T2:Rack1^fl/fl mouse and 2 littermate control mice and the entire small intestine was analyzed from 2 additional vil-Cre:Rack1^+/fl control mice. Tubulin levels are shown at bottom. D: analyses of Rack1 levels in tissue lysates of pancreas (left) or liver (right) in a vil-Cre-ER^T2:Rack1^fl/fl mouse (KO) or its littermate controls vil-Cre-ER^T2:Rack1^+/fl (C3) or Rack1^+/fl (C4). Total lysate protein loaded: pancreas: 20 μg; liver: 30 μg. Rack1 levels in pancreatic or liver cells are expressed relative to those of the vil-Cre-ER^T2:Rack1^+/fl littermate control (C3). Data shown for pancreas and liver are representative of 2 experiments. Tubulin levels are shown at bottom.

Necropsies performed on euthanized tamoxifen-fed vil-Cre-ER^T2:Rack1^fl/fl mice revealed mildly distended, fluid-, gas-, and sometimes blood-filled loops of small and large bowel (data not shown), inguinal lymphadenopathy, and thrombocytosis: for example, the platelet count of 1 mouse was 1,695 (normal 675–1,338). Necropsies of other organs of the tamoxifen-fed vil-Cre-ER^T2:Rack1^fl/fl mice and of all organs of the littermate control mice were normal (Fig. 2 and data not shown).

Fig. 2.

Following Cre activation in vil-Cre-ER^T2:Rack1^fl/fl mice, Rack1 deletion occurs only in intestinal epithelia and not in other epithelia, tissues, or organs. A 6-wk-old tamoxifen-fed vil-Cre-ER^T2:Rack1^fl/fl (inducible Rack1^−i/−i) mouse and a Rack1^+/fl littermate control were euthanized and tissues/organs were formalin-fixed, paraffin-embedded, and sectioned, as described in materials and methods. Necropsies were performed by Stanford Veterinarians. Immunofluorescent (IF) staining for Rack1 (green) was performed on sections from 15 organs/tissues (11 shown), as indicated. Nuclei were identified by Hoechst 33342 fluorescent staining of DNA (blue). Images were captured using a Nikon Eclipse E600 microscope at ×200 magnification, except for those of the cerebellum and pancreas combined with the small bowel (SB), which were captured at ×100 magnification. Scale bar = 100 μm. Data are representative of up to 10 fields examined for each organ.

Immunoblot analyses of epithelial fractions of small intestine and colon revealed that Rack1 levels were nearly absent in intestinal epithelia (duodenum, jejunum, ileum, and colon) of tamoxifen-fed vil-Cre-ER^T2:Rack1^fl/fl mice, whereas Rack1 levels in their littermate controls were similar to those of other control mice (Fig. 1C). In contrast, Rack1 levels in the pancreas and liver of induced Rack1^−i/−i mice were equivalent to those of littermate controls (Fig. 1D), indicating tissue specificity of the recombination.

Cre activation in vil-Cre-ER^T2:Rack1^fl/fl mice (Rack1^−i/−i mice) results in deletion of Rack1 only in intestinal epithelia and not in other epithelia, tissues, or organs of the mice.

Because some Cre strains exhibit significant, unexpected, and unreported excision activity/deletion outside of their targeted tissue or cell types (11), we assessed Rack1 expression in other organs and tissues of a Rack1^−i/−i mouse. We observed that Rack1 expression in skin, lung, brain (choroid plexus, hippocampus, cortex, and cerebellum), uterus, tongue, esophagus, and pancreas was not different than in a littermate control (Fig. 2), confirming the specificity of the targeted deletion.

Rack1 was highly expressed in many types of epithelia: simple columnar (e.g., in small bronchi of the lung and in uterine endometrium), simple cuboidal (e.g., in the choroid plexus and in terminal bronchioles of the lung), stratified squamous (e.g in the esophagus and tongue), and epidermis of the skin (Fig. 2). Rack1 was also highly expressed in the pancreas and specific neurons of the cortex, hippocampus, and cerebellum.

In contrast, Rack1 expression was low in the liver and kidneys of the Rack1^−i/−i mouse and was not different than that in the control mouse (data not shown). Similarly, Rack1 expression was fairly low in all types of muscle (intestine, uterus, skeletal, and heart) of the Rack1^−i/−i mouse and was not different than that in the control mouse: for example, for uterus compare the side-by-side panels in Fig. 2 and for skeletal muscle compare the muscle underlying the tonque in Fig. 2.

Rack1 loss throughout the intestinal epithelia results in a patchy, erosive, hemorrhagic, inflammatory enterocolitis.

Immunofluorescent analyses of Rack1 expression in tamoxifen-fed vil-Cre-ER^T2:Rack1^fl/fl mice [see Figs. 3A, 4A, 5 (expression not shown), 6, 7, 8 (expression not shown), and 9A] confirmed the results of the immunoblot analyses that Rack1 was deleted in nearly 100% of the intestinal epithelia throughout the entire small and large bowel. The lack of mosaicism demonstrated consistent Cre activity and thereby near universal Rack1 deletion in intestinal epithelia.

Fig. 3.

Rack1 loss throughout the intestinal epithelia results in a patchy, erosive, hemorrhagic, inflammatory enteritis. A 6-wk-old vil-Cre-ER^T2:Rack1^fl/fl mouse and a Rack1^fl/fl littermate control were euthanized 6 days after initiating tamoxifen treatment, when the induced vil-Cre-ER^T2:Rack1^fl/fl (inducible Rack1^−i/−i) mouse developed bloody diarrhea and weight loss. Mouse necropsies were performed by a Stanford veterinarian. IF staining for Rack1 (top) and hematoxylin and eosin (H&E) analysis (bottom) of small intestine were performed. The loss of the muscularis in the control mouse (A, left) was an artifact of fixation and was not seen in other control mice. Slightly enhanced staining of Rack1 in lamina propria near Rack1-deleted epithelia was observed (A, top right). Images were captured at ×100 (A, top and middle), ×200 (A, bottom) and ×400 (B) magnification. Scale bar = 100 μm. Stars: eroded surface epithelia; arrowheads: vacuolated, poorly differentiated, epithelial cells; dashed lines: surround inflammatory cell infiltrates in the lamina propria.

Fig. 4.

Rack1 loss throughout the intestinal epithelia results in a patchy, erosive, hemorrhagic, inflammatory colitis. IF staining for Rack1 (A, top) and H&E analysis (A, bottom, and B) of the colon were performed on the 6-wk-old vil-Cre-ER^T2:Rack1^fl/fl (inducible Rack1^−i/−i) mouse and the Rack1^fl/fl littermate control shown in Fig. 3. Images were captured at ×200 (A) and ×400 (B) magnification. Scale bar = 100 μm.

Fig. 5.

The severity of the enterocolitis in the Rack1-deficient mice is fully revealed when resected intestines are flushed before fixation. vil-Cre-ER^T2:Rack1^fl/fl mice (inducible Rack1^−i/−i), and littermate controls: vil-Cre-ER^T2:Rack1^+/fl (inducible Rack1^+/−i), Rack1^fl/fl, Rack1^+/+, or Rack1^+/fl were tamoxifen fed every 2–3 days and euthanized when the induced Rack1^−i/−i mice became ill with bloody diarrhea and weight loss, and the resected intestines were gently flushed once intraluminally with PBS before fixation in formalin, as described in materials and methods. H&E analyses were performed on small intestine (duodenum, jejunum, and ileum) and colon sections. Shown is the histopathology from a 7-mo-old induced Rack1^−i/−i mouse (A, right) and a Rack1^fl/fl littermate control (A, left) and a 3-mo old induced Rack1^−i/−i mouse (B, right) and a Rack1^+/−i littermate control (B, left). Scale bar = 100 μm. Images were captured at ×100 magnification. Data are representative of up to 50 fields examined for each mouse.

Fig. 6.

Rack1 deficiency increases the number of proliferating cells in intestinal crypts. A: IF costaining for Rack1 (green) and PCNA (red) in the small intestine (top) and colon (bottom) of an induced Rack1^−i/−i mouse (right) and a Rack1^fl/fl littermate control (left). Images were captured at ×400 magnification. Scale bar = 100 μm. Data are representative of up to 50 fields examined for each mouse. B: quantitative analysis of the number of PCNA-positive cells/crypt in the small intestine (top) or colon (bottom) of the induced Rack1^−i/−i mouse (RK−, red bar) and the Rack1^fl/fl control (RK+, blue bar). Data represent mean values ± SE for PCNA-positive cells/crypt in the small intestine (RK+, n = 116 crypts counted; RK-, n = 50) and colon (RK+, n = 68; RK−, n = 51).

Fig. 7.

Rack1 loss reduces the number of fully differentiated enteroendocrine and goblet cells. A: IF costaining for Rack1 (green, left) and synaptophysin (red, right) in the small intestine of an induced Rack1^−i/−i mouse (bottom) and a Rack1^fl/fl littermate control (top) who had received 2 doses of tamoxifen, 4 days apart. In other fields of control mice, synaptophysin-positive cells were seen spanning the entire crypt-villus axis. B: IF costaining for Rack1 (left, green) and mucin 2 (right, red) in the colon (rows 1 and 2) and the small intestine (rows 3 and 4) of induced (ind) Rack1^−i/−i mice (rows 2 and 4) and littermate controls: tamoxifen-fed Rack1^fl/fl (row 1) or vil-Cre-ER^T2:Rack1^+/fl (Rack1^+/−, row 3). DNA is labeled by Hoechst 33342 fluorescent staining (blue). Scale bar = 100 μm. Images were captured at ×100 magnification. Data are representative of up to 50 fields examined for each mouse. C and D: quantitative analysis of the number of synaptophysin-positive cells/villus (C) or mucin 2-positive cells/colonic crypt (D) in the induced Rack1^−i/−i mice (RK−, red bar) and the Rack1^fl/fl controls (RK+, blue bar). Data represent mean values ± SE for synaptophysin-positive cells/villus (RK+, n = 74 villi counted; RK−, n = 154 intact villi counted), or Muc2-positive cells/crypt (RK+, n = 49; RK−, n = 74).

Fig. 8.

Rack1 deficiency results in loss of fully differentiated goblets cells and expansion of the Paneth cell population. A and B: periodic acid-Schiff staining (red) of goblet and Paneth cells in the small intestine (A) and of goblet cells in the colon (B) in an induced Rack1^−i/−i mouse (right) or littermate controls (left): Rack1^fl/fl (A) or Rack1^+/+ (B). Scale bar = 100 μm. Images were captured at ×100 magnification. Data are representative of up to 50 fields examined from each of 2 mice.

Fig. 9.

Rack1 loss increases the number of Paneth cells in crypts. A and B: IF costaining (A) for Rack1 (top, green) and lysozyme (bottom, red) or H&E staining (B) of the small intestine of an induced Rack1^−i/−i mouse (right) and a Rack1^fl/fl littermate control (left). Scale bar = 100 μm. Images were captured at ×400 magnification. Data are representative of up to 50 fields examined from 2 mice (A) or 10 mice (B).

In contrast, Rack1 expression was normal in intestinal epithelia of littermate control mice: tamoxifen-fed Rack1^fl/fl mice (for example, as shown in Figs. 3A, 4A, 6, 7, and 9A), vil-Cre-ER^T2:Rack1^+/fl (Fig. 7B), Rack1^+/+ (not shown), and Rack1^+/fl (not shown).

The histopathology of the Rack1^−i/−i mice revealed a patchy, erosive, hemorrhagic, inflammatory enteritis (Fig. 3) and colitis (Fig. 4). In the small intestine, there was villus blunting, villus fusion, striking elongation of crypts, and increased numbers of crypts per villus structure, only some of which could be accounted for by villus fusion (Fig. 3A, right). There were multifocal erosions in surface epithelia (stars), villus tip epithelial cell vacuolation (arrowheads), and increased inflammatory cell infiltrates in the lamina propria (within the dashed lines). In the base of the crypts, the epithelial cells were basophilic and the degree of basophilia persisted as the cells exited the crypt zone. The cells appeared cuboidal, vacuolated, and poorly differentiated as they migrated up the villus (Fig. 3A, bottom right, and Fig. 3B).

In the colon of Rack1^−i/−i mice, there was erosion of the surface epithelia and marked reduction in the number of fully differentiated goblet cells (Fig. 4, A and B). The epithelial cells remained basophilic, cuboidal, and poorly differentiated and lacked the obvious goblet cell morphology as they migrated toward the surface (Figs. 4, A and B, 5A, and 8B, right). Muc2 staining revealed that the large white spaces (Fig. 4) did not contain mucin 2 (Fig. 7B, row 2, right).

When the resected intestines of the Rack1^−i/−i mice were flushed once intraluminally with PBS before fixation in formalin, there was strikingly more destruction of the surface epithelium, with severely denuded/sloughed off epithelium, disruption of villus tips, and extrusion of the lamina propria into the lumen (Fig. 5, A and B, right), than was observed in resected intestines that were not flushed with PBS before fixation (Figs. 3 and 4). Perhaps flushing makes the enterocolitis appear more severe than it is by washing off the eroded damaged epithelia and villus tips. Collectively, our results show that Rack1 loss renders the epithelium more susceptible to barrier injury.

The gross pathologic findings of diarrhea that was sometimes bloody (seen in the cages of individually housed mice, extruding from the anus and in some resected intestines), together with the microscopic findings of red blood cells seen in/near the eroded surface epithelia of both the small and large bowel (Figs. 5 and 9B), demonstrated the hemorrhagic nature of the enterocolitis.

To determine whether age of the mice affected the phenotype, we compared the ages and histopathology of the 23 induced Rack1^−i/−i mice. The age range was 1.0 to 6.8 mo, with all but 1 between 1.0 and 4.5 mo. The average age was 2.9 mo. We observed no significant difference in severity of the enterocolitis between the older and younger mice. Shown is the histopathology of one of the youngest, a 6 wk old (Figs. 3 and 4); the oldest, a 7 mo old (Fig. 5A); and an average age of 3 mo old (Fig. 5B). While all had severe enterocolitis, the 3-mo old (intermediate age) appeared to have the most severe disease. Thus the age of the mice did not appear to affect the phenotype.

To determine whether the moribund state of the critically ill Rack1^−i/−i mice contributed to the phenotype, we euthanized/analyzed some mice before they became seriously ill. While all Rack1^−i/−i mice developed diarrhea and weight loss within 2–3 days after initiating tamoxifen treatment (Fig. 1, A and B), most did not become seriously ill until 5–10 days after initiating treatment. When mice were euthanized/analyzed before they became critically ill or moribund, enterocolitis was detected (data not shown). Thus the enterocolitis predated the moribund state. This indicates that the moribund state was not the cause of the enterocolitis but rather the result of it.

To assess for possible tamoxifen-induced intestinal toxicity, we analyzed the histology of the small intestine and colon from tamoxifen-fed littermate control mice: Rack1^fl/fl mice (Figs. 3–9), vil-Cre-ER^T2:Rack1^+/fl (Figs. 5B and 7B, left), Rack1^+/+ (Fig. 8B, left), and Rack1^+/fl (not shown). We observed normal intestinal histology in all of these mice, indicating that there was no detectable tamoxifen-induced toxicity.

Rack1 deficiency increases the number of proliferating cells in intestinal crypts.

Previously, we showed that Rack1 regulates growth of colon cells in vitro, partly by inhibiting Src activity at key cell cycle checkpoints, in apoptotic and cell survival pathways, and at cell-cell adhesions (14–18, 21). Utilizing a mouse model of constitutive Rack1 deletion, we showed that Rack1 also suppresses crypt cell proliferation and regeneration in vivo (5). To assess this in induced Rack1^−i/−i mice, IF analyses were performed using PCNA as a marker for proliferating crypt cells. We observed that Rack1-deleted crypts in the small intestine or colon of induced Rack1^−i/−i mice (like those of constitutive Rack1^−i/−i mice), contained approximately three times more PCNA-positive cells than did Rack1-expressing crypts of littermate controls (Fig. 6, A and B; Ref. 5).

Because of the striking increase in proliferating cells in Rack1-deleted crypts, we then stained for Sox9 and CD44, which mark populations of crypt cells that encompass stem and progenitor cells (8–10). We observed that Rack1-deleted crypts in the small intestine of Rack1^−i/−i mice (like those of constitutive Rack1^−i/−i mice) contained many more CD44-positive and Sox9-positive cells than did Rack1-expressing crypts in littermate controls (data not shown). Collectively, our results from PCNA, CD44, and Sox9 staining suggest that Rack1 deficiency either promotes crypt cell proliferation and/or results in a partial block to proliferating cells exiting the crypt zone.

Rack1 deficiency results in loss of fully differentiated goblet and enteroendocrine cells and expansion of the Paneth cell population.

Utilizing a mouse model of constitutive Rack1 deletion, we showed that Rack1 promotes differentiation of crypt cells into fully differentiated goblet and enteroendocrine secretory cell lineages and inhibits differentiation into Paneth cells (5). Based on these results, we hypothesized that Rack1 deficiency in induced Rack1^−i/−i mice would result in loss of fully differentiated goblet and enteroendocrine cells and expansion of the Paneth cell population. To test this, we assessed the effect of Rack1 loss on cell differentiation using markers specific for each cell lineage. We observed that Rack1-deleted crypts in both the small intestine and colon of induced Rack1^−i/−i mice contained significantly fewer goblet cells, as detected by H&E staining (Figs. 4, 5A, and 9B), IF staining for mucin 2 (Fig. 7, B and D), and PAS staining (Fig. 8, A and B) than did Rack1-expressing crypts of littermate controls. These results were similar to those observed in constitutive Rack1^−i/−i mice (5).

To assess for the effect of Rack1 loss on differentiation of enteroendocrine cells, we analyzed synaptophysin expression in the small intestine of induced Rack1^−i/−i mice in regions of intact epithelia and villi that had not been destroyed/denuded/sloughed off by severe erosive enteritis. We observed that in these regions of Rack1-deleted (but otherwise well preserved) villi there were significantly fewer synaptophysin-positive enteroendocrine cells than in Rack1-expressing villi of littermate controls (Fig. 7A, compare the top and bottom right, and Fig. 7C). These results were similar to those observed in constitutive Rack1^−i/−i mice, where erosion of the surface epithelia does not occur (5). Collectively, these results indicate that the diminished numbers of synaptophysin-positive enteroendocrine cells observed in Rack1^−i/−i mice is not due to loss/erosion of surface epithelium.

Based on our finding that Rack1 loss induces an increase in CD44- and Sox9-positive crypt cells (data not shown and Ref. 5), which encompass stem and progenitor cells, we hypothesized that Rack1 loss would induce differentiation into Paneth cells, which support the stem cell niche. We observed that Rack1-deleted crypts of induced Rack1^−i/−i mice contained a patchy but striking increase in the number of Paneth cells, as detected by IF staining for lysozyme (Fig. 9A, bottom right), H&E staining (Fig. 9B, bottom right), where Paneth cells are identified by the numerous, eosinophilic, refractile, cytoplasmic granules that are characteristic of these cells, and by PAS staining (Fig. 8A, top right). While PAS stains both goblet and Paneth cells, the striking increase in PAS-positive cells observed in Rack1-deleted crypts in the small intestine of Rack1^−i/−i mice (Fig. 8A, top right) likely marks Paneth cells because mucin staining revealed few, if any, goblet cells in Rack1-depleted crypts of small intestine (Fig. 7B, bottom right). These results were similar to those in constitutive Rack1^−i/−i mice, where lysozyme staining also revealed expansion of the Paneth cell population (5). Collectively, our results show that Rack1 deficiency results in loss of fully differentiated goblet and enteroendocrine cells and expansion of the Paneth cell population.

DISCUSSION

In this study we identify a novel function of Rack1 in vertebrates: maintaining intestinal homeostasis by protecting the integrity of the intestinal epithelium. Clearly Rack1 provides an important protective function for the intestine as is evident by our finding that Rack1 deletion in nearly 100% of the intestinal epithelia of induced Rack1^−i/−i mice results in a severe erosive enterocolitis (Figs. 3–9) and death. The severe destruction of the surface epithelium, disruption of villus tips, and extrusion of the lamina propria that were observed following Rack1deletion (Figs. 3–9) would profoundly impact absorption and secretion. Because the mice become critically ill and die within 1–2 wk of tamoxifen induction, this limits further studies on Rack1’s influence on intestinal function. In regards to the only constitutive mouse model of intestinal Rack1 deficiency that we generated (Rack1^−i/−i; Ref. 5), Rack1 was deleted in such a small percentage of the total surface area of the intestinal epithelia (<10%) that this would have little, if any, impact on overall gut function.

Each mouse model of Rack1 deficiency that we generated allows us to study unique aspects of Rack1 function that cannot be studied in other mouse models. For example, because Rack1 is deleted in <10% of the surface area of the epithelia of Rack1^−i/−i mice, they live a normal lifespan, and this allows for analysis of Rack1’s role in injury-induced apoptosis and regeneration (the mice can withstand radiation injury) and suppression of neoplasia but does not allow for analysis of the full phenotype (5). In contrast, because Rack1 is deleted in nearly 100% of the induced Rack1^−i/−i mice, the full phenotype of severe enterocolitis emerges (Figs. 3–5), but the mice become so critically ill with bloody diarrhea and weight loss that they die within 1–2 wk of initiation of tamoxifen treatment, thus precluding analysis of other Rack1 functions e.g., injury-induced apoptosis and regeneration (the mice cannot withstand/survive radiation) or development of neoplasia.

In Rack1-deleted epithelia of the small intestine of induced Rack1^−i/−i mice, the epithelial cells were basophilic and crowded at the base of the crypts, which were elongated/hyperplastic (Fig. 3A, bottom right). This degree of basophilia persisted as the cells exited the crypt zone. The cells appeared cuboidal, vacuolated, and poorly differentiated as they migrated up the villus (Fig. 3B). Similarly in the colon, there was crypt elongation and the cells remained basophilic, cuboidal, and poorly differentiated and lacked the obvious goblet cell morphology as they migrated toward the surface (Figs. 4A, 4B, 5A, 7B, and 8B, right). These changes suggest abnormal differentiation and could explain, in part, the reduced numbers of fully differentiated goblet and enteroendocrine cells seen in Rack1-deleted epithelia (Figs. 3B, 4A, 4B, 5A, 7B, and 8B; and Ref. 5). There could also be a partial block to proliferating cells exiting the crypt zone to explain the crypt elongation/hyperplasia and the reduced numbers of fully differentiated cells.

Rack1 is involved in a large number of diverse biological functions and works in many ways (by shuttling proteins around the cells, anchoring proteins at particular subcellular locations, and stabilizing or inhibiting the activity of its enzymatic binding partners) and at many subcellular sites (e.g., ribosomes, inner plasma membrane, and nucleus; reviewed in Refs. 1, 6, 13). Thus, while Rack1 itself is fairly ubiquitously expressed, its biologic function may depend more upon the subcellular localization, translocation, and compartmentalization of both Rack1 and its binding partners than on the overall expression levels of Rack1. Further studies are needed to assess the sites and mechanisms by which Rack1 works in intestinal epithelial cells in vivo.

Interestingly, our in vitro studies show that one function of Rack1 in intestinal epithelial cells is to promote cell-cell adhesion by regulating E-cadherin tyrosine phosphorylation and endocytosis and by diverting E-cadherin from a degradative to a recycling pathway (21). Perhaps this is one mechanism by which Rack1 maintains the integrity of the intestinal epithelial barrier in vivo. However, our initial IF studies revealed that E-cadherin expression and subcellular localization in Rack1-depleted epithelia of induced Rack1^−i/−i mice (at least in areas that remained after severe destruction/denudation of the surface epithelia) were not different from those in Rack1-expressing epithelia of controls (data now shown). It remains possible that alterations in E-cadherin stability/function in cell-cell adhesions occur but cannot be detected by IF analyses.

The intestine and many organs of the body maintain epithelial homeostasis by employing a mucosal barrier, which is required for lubrication and for barrier function against outside pathogens. Goblet cells are specialized secretory cells that are largely responsible for secreting components of this mucosal barrier and represent a key component of the innate defense system (reviewed in Ref. 19). Goblet cells secrete mucins (predominantly mucin 2), trefoil factors, and other proteins, all of which contribute to the mucus layer protecting the epithelium. Loss or dysfunction of intestinal goblet cells and the protective barrier that they provide has been implicated in chronic infectious and inflammatory bowel diseases. Using three complementary approaches (H&E, IF staining for mucin 2, and PAS staining), we observed a striking loss of differentiated goblet cells in both the small intestine and colon of Rack1-deficient mice (Figs. 4, 5, 7, 8, and 9B). We think that loss of goblet cells and the protective barrier that they provide is a major factor contributing to the development of the severe enterocolitis and thereby the death of the Rack1-deficient mice. Our results suggest a key role for Rack1 in maintaining intestinal homeostasis and protecting the integrity of the epithelial barrier by maintaining a stable population of fully differentiated goblets cells.

Crohn’s disease and ulcerative colitis are genetically common chronic inflammatory bowel diseases the pathophysiology of which is poorly understood (reviewed in Refs. 12, 20). Ulcerative colitis is an erosive mucosal disease that affects only the colon whereas Crohn’s is a patchy, ulcerated, exudative, transmural disease that affects both the small intestine (enteritis) and colon (colitis) (3). Understanding mechanisms that protect barrier function in normal intestine and how loss of that protection contributes to the pathogenesis of inflammatory bowel diseases could lead to improved therapies for these diseases.

The severe, erosive, hemorrhagic, and inflammatory enterocolitis resulting from intestinal Rack1 deficiency (Figs. 3–5) resembles the destruction seen in inflammatory bowel diseases. The patchy involvement of both the small intestine and colon in the Rack1-deficient mice resembles Crohn’s, whereas the erosive disease confined to the mucosa (epithelia and lamina propria) resembles ulcerative colitis. Understanding mechanisms by which Rack1 protects the integrity of the epithelium could shed light on the underlying pathophysiology of inflammatory bowel diseases and lead to novel therapies for these and other erosive diseases of the gastrointestinal tract.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-43743 (to C. A. Cartwright) and DK-56339.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Z.-F.C. and C.A.C. conceived and designed research; Z.-F.C. performed experiments; Z.-F.C. and C.A.C. analyzed data; Z.-F.C. and C.A.C. interpreted results of experiments; Z.-F.C. and C.A.C. prepared figures; Z.-F.C. and C.A.C. drafted manuscript; Z.-F.C. and C.A.C. edited and revised manuscript; Z.-F.C. and C.A.C. approved final version of manuscript.