Loss of TMF/ARA160 Protein Renders Colonic Mucus Refractory to Bacterial Colonization and Diminishes Intestinal Susceptibility to Acute Colitis*

Background: TMF/ARA160 regulates the NF-κB subunit, p65 RelA, under stress conditions. Results: Loss of TMF/ARA160 renders mice colonic mucus refractory to bacterial colonization and diminishes intestinal susceptibility to acute colitis. Conclusion: TMF/ARA160 affects intestinal susceptibility to acute colitis. Significance: Targeting TMF/ARA160 in the colon may attenuate the onset and progression of acute colitis. TMF/ARA160 is a Golgi-associated protein with several cellular functions, among them direction of the NF-κB subunit, p65 RelA, to ubiquitination and proteasomal degradation in stressed cells. We sought to investigate the role of TMF/ARA160 under imposed stress conditions in vivo. TMF−/− and wild-type (WT) mice were treated with the ulcerative agent dextran sulfate sodium (DSS), and the severity of the inflicted acute colitis was determined. TMF−/− mice were found to be significantly less susceptible to DSS-induced colitis, with profoundly less bacterial penetration into the colonic epithelia. Surprisingly, unlike in WT mice, no bacterial colonies were visualized in colons of healthy untreated TMF−/− mice, indicating the constitutive resistance of TMF−/− colonic mucus to bacterial retention and penetration. Gene expression analysis of colon tissues from unchallenged TMF−/− mice revealed 5-fold elevated transcription of the muc2 gene, which encodes the major component of the colonic mucus gel, the MUC2 mucin. Accordingly, the morphology of the colonic mucus in TMF−/− mice was found to differ from the mucus structure in WT colons. The NF-κB subunit, p65, a well known transcription inducer of muc2, was up-regulated significantly in TMF−/− intestinal epithelial cells. However, this did not cause spontaneous inflammation or increased colonic crypt cell proliferation. Collectively, our findings demonstrate that absence of TMF/ARA160 renders the colonic mucus refractory to bacterial colonization and the large intestine less susceptible to the onset of colitis.

TMF/ARA160 is a Golgi-associated protein with several cellular functions, among them direction of the NF-B subunit, p65 RelA, to ubiquitination and proteasomal degradation in stressed cells. We sought to investigate the role of TMF/ ARA160 under imposed stress conditions in vivo. TMF ؊/؊ and wild-type (WT) mice were treated with the ulcerative agent dextran sulfate sodium (DSS), and the severity of the inflicted acute colitis was determined. TMF ؊/؊ mice were found to be significantly less susceptible to DSS-induced colitis, with profoundly less bacterial penetration into the colonic epithelia. Surprisingly, unlike in WT mice, no bacterial colonies were visualized in colons of healthy untreated TMF ؊/؊ mice, indicating the constitutive resistance of TMF ؊/؊ colonic mucus to bacterial retention and penetration. Gene expression analysis of colon tissues from unchallenged TMF ؊/؊ mice revealed 5-fold elevated transcription of the muc2 gene, which encodes the major component of the colonic mucus gel, the MUC2 mucin. Accordingly, the morphology of the colonic mucus in TMF ؊/؊ mice was found to differ from the mucus structure in WT colons. The NF-B subunit, p65, a well known transcription inducer of muc2, was upregulated significantly in TMF ؊/؊ intestinal epithelial cells. However, this did not cause spontaneous inflammation or increased colonic crypt cell proliferation. Collectively, our findings demonstrate that absence of TMF/ARA160 renders the colonic mucus refractory to bacterial colonization and the large intestine less susceptible to the onset of colitis.
TMF/ARA160 is a Golgi-associated protein with various cellular functions. These include a role in retrograde transport processes from endosomes to the Golgi and from the Golgi to the endoplasmic reticulum (1). Additionally, TMF/ARA160 was shown to promote ubiquitination and proteasomal degra-dation, which is induced in cells subjected to stress insults. The targets of the TMF/ARA160 ubiquitination-promoting activity are key prosurvival transcription factors such as Stat3 and the NF-B subunit, p65-RelA (2,3). To gain a deeper insight to the physiological roles of TMF in vivo, we established TMF Ϫ/Ϫ knock-out mice. TMF Ϫ/Ϫ mice, which lack the TMF protein, develop normally, and are healthy, although the males are infertile due to major sperm maturation defects. These defects could be attributed both to the role of TMF/ARA160 in the transport of Golgi-derived granules to the nuclear apical surface, and to its ubiquitination-promoting activity (4).
Inflammatory bowel disease (IBD) 2 comprises two major chronic inflammatory disorders of the gastrointestinal tract, which can also lead to the development of malignancies of this tissue. These are ulcerative colitis (UC) and Crohn disease (5).
UC develops due to malfunctioning of the colonic mucosal barrier and exposure of epithelial and innate immune cells to the luminal bacterial flora and to bacterial toxins (6,7). Protection of the large intestine, which harbors an extensive amount (10 13 -10 14 ) of commensal bacteria, is therefore a formidable challenge. Several components comprise the mucosal/epithelial barrier. These include the extracellular mucus layer facing the lumen, whose main constituent is the evolutionarily preserved glycoprotein MUC2, secreted by the intestinal goblet cells (8). The extracellular barrier is composed of an inner "firmly" adherent mucus layer attached to the intestinal epithelial cell layer and shielding the epithelium from direct contact with bacteria and an outer "loose" non-adherent mucus layer (9). Underneath the firm mucus are the epithelial cell layers which, in addition to the production of mucosal constituents, also serve as an additional barrier restricting the intrusion of pathogens and toxins (10).
Murine models have become essential tools to investigate pathophysiological mechanisms and immunological processes underlying mucosal inflammation, including dextran sodium sulfate (DSS), which when given for a short period induces UC-* This work was supported by a grant from The Israeli Ministry of Health Chief like disease in mice (11). DSS is a physical agent with an intrinsic capacity to disrupt the epithelial cell barrier, causing normal mucosal microfloral substances to activate mucosal macrophages, which in turn produce immunomodulatory cytokines and induce acute and chronic inflammatory UC (12). One of the notions emerging from murine IBD models is that targeted changes in expression of survival factors of intestinal epithelial cells can also affect the vulnerability of this tissue to the induction of inflammatory processes. Accordingly, when the ikk gene encoding the NF-B upstream activator, IKK, is disrupted selectively in intestinal epithelial cells, the intestine becomes extremely sensitive to DSS-induced colitis (13). In agreement with this observation, disruption of the genes encoding either the p50 or p65 (RelA) NF-B subunits in intestinal epithelial cells, renders these cells sensitive to pathogen and chemically induced cell damage and consequent inflammation (14).
Considering the ability of TMF/ARA160 to direct proteasomal degradation of the NF-B subunit p65 and the importance of this factor in the onset of IBD, we sought to investigate the potential role of TMF/ARA160 in DSS-induced colitis by testing disease susceptibility of TMF/ARA160 knock-out mice. We show here that the cellular roles of TMF/ARA160 are redundant under normal intestinal functioning in mice but affect intestinal susceptibility to UC. Thus, manipulation of TMF/ARA160 functions could ameliorate intestinal sensitivity to the onset and progression of UC.

MATERIALS AND METHODS
Mice Handling and Maintenance-All experiments were carried out on female, 8-week-old TMF Ϫ/Ϫ ICR mice (4) and WT littermate controls. Mice were group housed under specific pathogen-free conditions with controlled temperature (25°C) and photoperiod (12:12 h light/dark cycle) and allowed unrestricted access to standard mouse chow and water. All of the protocols were approved by the Institutional Animal Care and Use Committee at Bar-Ilan University.
Scoring Colitis Clinical Activity-Body weights were measured every other day during the DSS treatment period. Presence of blood in excreta was assessed on the last day of DSS treatment. Disease score was determined as was described previously (15). Mice were sacrificed by cervical dislocation 7 days after induction of colitis. The abdominal cavity was exposed by a midline laparotomy, and the entire colon was removed from the caecum to the anus. The length of the colon was measured, and tissue obtained from each colon was stored at Ϫ80°C and processed for further assays. All measurements were preformed in a blinded fashion.
Histological, Immunohistochemical, and Immunofluorescence Analysis-Colons removed from mice were fixed immediately overnight in either 4% formalin for H&E and immunohistochemical staining or in water-free Methanol-Carnoy's fixative (60% dry methanol, 30% chloroform and 10% acetic were incubated for 1 h with the following secondary antibodies: anti-mouse or anti-rabbit Alexa Fluor 488 and anti-rabbit Alexa Fluor 594 (1:150) (Molecular Probes, Invitrogen), and the nuclei were stained with 0.1 g/ml Hoechst (Sigma-Aldrich). For HRP immunohistochemistry, a secondary anti-rabbit HRPconjugated antibody was used according to the manufacturer's instructions (ab80437, Abcam). Nuclei were visualized with hematoxylin (Sigma-Aldrich).
Histological Assessment of Colitis-Histopathological analysis and semi-quantitative scoring was performed by a boardcertified toxicological pathologist according to the scoring system described by Cooper et al. (16), taking into consideration the grades of extension (laterally, along the mucosa and deep into the mucosa, submucosa, and/or muscular layers) of the inflammation and ulceration. Analysis was performed in a blinded fashion.
Myeloperoxidase Activity-Neutrophil infiltration into the colon was quantified by measuring myeloperoxidase activity according to the technique described by Boughton-Smith et al. (17). Results are shown as percent change in kinetics compared with control mice.
Fluorescent in Situ Hybridization-Paraffin sections were dewaxed with toluene and hybridized with a general bacterial probe, EUB 338, conjugated to FITC as described by Johansson et al. (9). Nuclei were stained with 0.1 g/ml Hoechst (Sigma-Aldrich).
Quantification of Colonic Bacteria-DNA isolation from colons was performed using DNeasy ® blood and tissue kit (Qiagen) according to the manufacturer's instructions. A total of 100 ng of DNA was analyzed by PCR for 23 cycles as described by Yan et al. (18). PCR amplification of the GAPDH gene was performed to ensure equal loading. Gels were analyzed for a density plot histogram using NIH ImageJ software.
RNA Isolation and Quantitative Real-time PCR Analysis-RNA from colon tissue was extracted using TRI reagent (Molecular Research Center, ICN) following the manufacturer's instructions. Total RNA (1 g) was reverse-transcribed using SuperScript first-strand synthesis system for qRT-PCR (Invitrogen). Quantitative real-time PCR was performed using the StepOnePlus system with Fast SYBER Green Master Mix and analyzed with StepOne software (Applied Biosystems). Expression levels were normalized to GAPDH and are presented as fold expression change relative to untreated wild-type samples. The primers used are shown in supplemental Table S1.
NF-B p65 RelA Transcription Factor Activity Assay-Isolation of nuclear extracts from mouse colons and NF-B p65 binding activity assay were performed using the NF-B (p65) transcription factor assay kit (Cayman Chemical) according to the manufacturer's instructions.
Statistical Analysis-Data are presented as mean values Ϯ S.E. Statistical analysis was conducted using the Student's t test and is indicated for each figure. The n for each experiment is noted in the figure legend.

TMF Is Highly Expressed in Apical Cytoplasm of Intestinal
Epithelial Cells-TMF/ARA160 has been shown previously to reside in the Golgi apparatus of several mammalian cell lines (1). It was also demonstrated that TMF/ARA160 is abundantly expressed in most murine tissues, with remarkably high levels in the colon (4). To investigate the localization of TMF/ ARA160 in the mouse colon, immunofluorescent staining was performed on tissue sections. Staining was almost exclusive to intestinal epithelial cells (IEC) of all types: enterocytes, goblet cells, and crypt stem cells (Fig. 1, A and B). In these cells, TMF/ ARA160 localized to the apical cytoplasm (Fig. 1, A and B). Utilizing confocal microscopy, we found that TMF does not co-localize with the Golgi apparatus and appears scattered in the apical cytoplasm of enterocytes (Fig. 1, C and D). In the lamina propria, TMF/ARA160 is expressed at relatively low lev-els and is concentrated in a defined area in the cytoplasm of these cells (Fig. 1C) TMF Ϫ/Ϫ Mice Are Less Sensitive to Chemically Induced Colitis-Pathological examination of colon tissue sections from female TMF Ϫ/Ϫ mice of all ages show normal architecture of the colonic epithelium and the interior lamina propria (Fig. 2, A  and B). These results demonstrate that lack of TMF/ARA160 does not cause pathological effects in unchallenged mice colon.
To assess the role of TMF/ARA160 under injury and inflammation in the colon, we used the established model of DSSinduced colitis. WT and TMF Ϫ/Ϫ littermates, at the age of 8 weeks, were administered 3% DSS in drinking water, and control mice were given drinking water without additives. After 7 days, mice were sacrificed, and tissue samples were collected for further processing and examination. Histopathological inspection revealed significantly less tissue damage and ulceration in TMF Ϫ/Ϫ colons as opposed to their WT littermates (Fig. 2, C,  D, and F). Other symptoms of colitis such as weight loss, fecal blood, and colon shortening were also diminished in TMF Ϫ/Ϫtreated mice (Fig. 2, E, G, I, and J). To evaluate the extent of neutrophil infiltration into the mucosa, we measured myeloperoxidase activity in tissue extracts. Although myeloperoxidase levels in WT mice were elevated ϳ5-fold following DSS treatment, as expected, TMF Ϫ/Ϫ treated mice showed only a 2-fold increase in myeloperoxidase activity, relative to their untreated controls (Fig. 2H).
The difference in inflammatory scores between DSS treated WT and TMF Ϫ/Ϫ mice was further manifested by the differences between the expression levels of key proinflammatory agents in the two cohorts. Using qRT-PCR analysis, we showed that the levels of mRNAs encoding IL-1␤, KC, and TNF␣ are significantly lower in TMF Ϫ/Ϫ mice. Notably, although the levels of the above three cytokines were lower in colons of both untreated and DSS-treated TMF Ϫ/Ϫ mice, the level of the IFN␥ transcript was decreased significantly only in colons of DSStreated TMF Ϫ/Ϫ animals (Fig. 3).
To further evaluate the damage to the epithelial barrier following DSS treatment, we analyzed the colonic tissue sections for bacterial penetration by fluorescent in situ hybridization using a general, bacterial-specific 16 S rRNA probe. Although bacteria were abundant in colonic tissues of DSS-treated WT mice, only weak staining was observed in TMF Ϫ/Ϫ -treated animals (Fig. 4, A and B). Accordingly, semi-quantitative PCR analysis using primers directed toward the highly conserved bacterial 16 S rRNA gene sequence confirmed these results, showing significantly lower bacterial presence in the colons of DSS-treated TMF Ϫ/Ϫ mice (Fig. 4, C and D). These results suggest that the lack of TMF in the IEC diminishes barrier disruption and inflammatory damage under conditions of induced acute colitis.

Bacterial Colonies Are Not Detected in Colonic Mucus of TMF
Ϫ/Ϫ Mice-We postulated that the diminished susceptibility of TMF Ϫ/Ϫ mice to DSS induced colitis could result from greater epithelial barrier integrity, which decreases bacterial penetration upon DSS treatment (19). Considering our observation that bacterial penetration of the colonic mucosal layer was diminished significantly in DSS-treated TMF Ϫ/Ϫ mice, we sought to investigate the ability of bacterial cells to colonize the colonic mucus of untreated TMF Ϫ/Ϫ mice. We used a very rapid and aggressive fixation technique, the Carnoy's fixation, which was proven to preserve the mucus layer and its inhabiting bacteria (8). Giemsa differential staining was applied to visualize the bacterial cells. In colons of untreated WT mice, bacterial colonies clearly were visible colonizing the mucus layer. Surprisingly, no bacterial colonies were visible in sections of colon samples from untreated TMF Ϫ/Ϫ mice (Fig. 5, A-D). To confirm and quantify our observations, we used semi-quantitative PCR analysis of the 16 S rRNA gene sequence present in colons of untreated WT and TMF Ϫ/Ϫ mice. This showed significantly lowered bacterial presence in colons of TMF Ϫ/Ϫ mice (Fig. 5, E  and F). Thus, absence of TMF/ARA160 renders the colonic mucus refractory to bacterial colonization.
Elevated Expression of muc2 mRNA and Distinct Architecture of Colonic Mucus in TMF Ϫ/Ϫ Mice-The colonic mucus layers are major constituents of the colonic barrier. Although the firm inner layer physically separates the gut bacteria from the epithelial cell layer, the loose outer layer allows bacterial adherence (20). The decreased bacterial colonization could therefore reflect modified characteristics of the colonic mucus in TMF Ϫ/Ϫ mice.

Diminished Susceptibility to Acute Colitis in TMF ؊/؊ Mice
To validate this notion, we analyzed the expression of an indispensable component of mucosal protective function, the muc2 gene, which is expressed exclusively in goblet cells and encodes the key colonic mucus component, the MUC2 gelforming mucin (21). By use of real-time qRT-PCR analysis, we found a 5-fold increase in the muc2 mRNA level in colons of untreated TMF Ϫ/Ϫ mice comparing to untreated WT mice, and this increase was maintained after DSS treatment as well (Fig. 6A).
Another plausible cause of the decreased bacterial penetration to the colonic mucosal layer of DSS-treated TMF Ϫ/Ϫ mice is the overexpression of antibacterial intestinal peptides of the defensin family (22). However, gene expression analysis in TMF Ϫ/Ϫ mice did not reveal any increase in the levels of mRNAs encoding the antimicrobial peptide mBD3 or the trefoil family protein TFF3 (Fig. 6, B and C), which is thought to maintain colonic epithelium stability (23). Similarly, qRT-PCR analysis of muc5ac mRNA, encoding the MUC5AC mucin, which has been shown to attenuate parasitic pathogen infiltration (24), showed no significant variance between WT and KO mice (Fig. 6D).
The significant increase in muc2 mRNA level in colonic epithelium of untreated TMF Ϫ/Ϫ mice suggests altered characteristics of the colonic mucus in these animals. To investigate this notion, we used Carnoy's fixation, which preserves the mucus layer architecture. Colonic sections were stained with specific Mucin2 antibody to visualize the colonic mucus morphology. As expected (8), staining of colon sections from WT mice showed a thin and densely stained firm inner layer overlaid by a diffuse loose outer mucosal layer (Fig. 7, A and C). However, in colonic sections from TMF Ϫ/Ϫ , we found a uniform, thick, and dense mucus layer (Fig. 7, B and D).
Up-regulation of NF-B p65 RelA Activity in IEC of TMF Ϫ/Ϫ Mice-One of the known transcription activators of the muc2 gene is the transcription factor, NF-B, which has been shown to be targeted to proteasomal degradation by TMF/ARA160 (2,26). Specific immunohistochemical staining revealed a major increase in the protein levels of the p65 NF-B subunit in the IECs of untreated TMF Ϫ/Ϫ mice (Fig. 8, A and B). To examine whether the increased level of the p65 protein leads to an upregulation of NF-B activity, the DNA-binding activity of the NF-B p65 subunit was determined in colons of untreated WT and TMF Ϫ/Ϫ mice using a commercial NF-B p65 transcriptional activity assay kit. This revealed a 2-fold increase in the relative level of NF-B p65 transcriptional activity in TMF Ϫ/Ϫ mice (Fig. 8E). The NF-B transcription regulatory activity was shown to be crucial for colonic epithelial integrity by regulating anti apoptotic genes in IECs (13). We therefore examined mRNA levels of the survival gene bcl-xl, which is a downstream target of NF-B p65 (10). Real-time qRT-PCR analysis revealed a 2-fold increase in the bcl-xl mRNA level in colons of untreated TMF Ϫ/Ϫ mice in comparison with colons of untreated WT mice (Fig. 8F). Because NF-B has been shown to control expression of proteins with anti-apoptotic and pro-proliferative activities (13), we examined whether the increase in NF-B p65 activity leads to an excessive proliferation of TMF Ϫ/Ϫ IECs. Immunohistochemical staining of PCNA, a marker of cell proliferation (28), did not show any difference in PCNA-positive cells in colonic crypts of untreated WT versus TMF Ϫ/Ϫ mice (Fig. 8, C and D).

DISCUSSION
Ulcerative colitis is an IBD, which combines impaired host reaction to commensal bacteria with ulcerative outcomes. This results mainly from the malfunctioning of the colonic mucosal barrier, thereby enabling the exposure of the innate immune cells residing in the colonic lamina propria to the luminal bacterial flora and to bacterial toxins (7,9). Consequent production of inflammatory cytokines initiates inflammatory cascades that lead to colonic tissue damage and to chronic inflammation, which can initiate malignant transformation (29).
In the current study, we show that the TMF/ARA160 protein affects the susceptibility of the colon to the onset of acute colitis. Absence of TMF/ARA160 alleviates the pathoclinical and inflammatory parameters that characterize and define the onset and progression of DSS-induced acute colitis. Because one of these parameters is bacterial penetration of the colonic tissue (30), which is a key step in the development of UC, we investigated whether the mucosal barrier of the colons in TMF Ϫ/Ϫ mice exhibits distinct properties from those of the WT mucosal barrier. Surprisingly, we found that although bacterial colonies inhabit the mucosal layer in colons of WT mice, no bacterial colonies could be detected in the mucus and epithelial layers in colons of TMF Ϫ/Ϫ mice. Accordingly, usage of quantitative methods revealed a significantly decreased amount of bacteria in TMF Ϫ/Ϫ colons in comparison with WT mice. Thus, absence of TMF/ARA160 renders the colonic mucus layer less available to bacterial retention and penetration, and consequently to bacterial colonization.
The refractory nature of TMF Ϫ/Ϫ colonic mucus to bacterial colonization suggests a change in its properties, which could be linked either to elevated accumulation of anti-microbial substances (31) or to increased viscosity (32), reflecting a modifi-cation in its biochemical composition. Gene expression analysis revealed an elevated expression of the muc2 gene, encoding the gel-forming MUC2 mucin, in colons of TMF Ϫ/Ϫ mice. The increased accumulation of MUC2 may thereby affect the biochemical and biophysical characteristics of the TMF Ϫ/Ϫ colonic mucus. In compliance with this notion, we noticed distinct architecture of the colonic mucus in TMF Ϫ/Ϫ mice. Although the WT mucus appeared as expected (8) to be composed of a thin, firm inner layer overlaid by a diffuse, loose outer layer, the TMF Ϫ/Ϫ mucus primarily exhibited only one dense and thick layer. Further biochemical analysis is required for confirming that the uniform colonic mucus of TMF Ϫ/Ϫ mice bears the physical characteristics of the inner firm layer of WT colons. However, the morphological structure observed by the MUC2 staining, combined with the lack of bacterial colonies support our postulated notion. The pivotal role of MUC2 and the colonic mucus layer in protection from the onset of IBD was underscored previously by the severe colonic inflammation, which develops spontaneously in MUC2-deficient mice (33). Furthermore, the lowered levels of the murine cytokine KC and of the antibacterial peptide mBD3, which are produced in response to microbial stimuli (34,35), coincide with the observation that bacterial presence is significantly lower in colons of TMF Ϫ/Ϫ mice.
TMF/ARA160 was shown previously to direct the proteasomal degradation of the NF-B subunit, p65 (2). Immunostaining of colonic sections from TMF Ϫ/Ϫ mice demonstrated an increased accumulation of NF-B p65 in TMF Ϫ/Ϫ IECs, and functional assay documented the up-regulation of the NF-B p65 transcription inducing activity in the TMF Ϫ/Ϫ colon. Because NF-B was shown to activate the transcription of the muc2 gene (5, 6), we propose that absence of TMF/ARA160 Diminished Susceptibility to Acute Colitis in TMF ؊/؊ Mice increases the levels of transcription factors that up-regulate the expression of the muc2 gene. This leads to the elevated accumulation of MUC2, to altered structure and characteristics of the colonic mucus, and to its consequent diminished availability for bacterial colonization. Thus, the mucus of TMF Ϫ/Ϫ mice renders the colonic mucosa a more robust barrier to bacterial penetration and to the onset of colitis. Although this alteration does not affect the normal development and growth of the animals, it does provide them with reduced sensitivity to the onset of acute colitis.
NF-B is a well known modulator of inflammation as well as a survival factor (13). However, recent studies have shown that elevated basal levels of NF-B activity in IECs does not necessarily lead to the onset of inflammation and even has a protective effect in the onset of acute colitis (25,27). Our results show that in colons of TMF Ϫ/Ϫ mice, the elevated level of NF-B p65 activity does not induce spontaneous inflammation or elevated cytokine production. Moreover, we cannot exclude the possibility that other factors such as elevated accumulation of the prosurvival protein Bcl-x L , driven by NF-B in TMF Ϫ/Ϫ IECs (14), also contribute to the reduced intestinal susceptibility of TMF Ϫ/Ϫ mice to DSS-induced acute colitis. Further studies should extend our understanding of the entire regulatory net-work, which is governed by TMF/ARA160 and could affect the susceptibility of mammals to IBD.