Bcl-3 deficiency protects against dextran-sodium sulphate-induced colitis in the mouse

Authors


Correspondence: R. J. Carmody, Centre for Immunobiology, Institute of Infection, Immunity and Inflammation, College of Medicine, Veterinary and Life Sciences, University of Glasgow, Glasgow G11 6NT, UK.

E-mail: ruaidhri.carmody@glasgow.ac.uk

Summary

Bcl-3 is a member of the IκB family of proteins and is an essential negative regulator of Toll-like receptor-induced responses. Recently, a single nucleotide polymorphism associated with reduced Bcl-3 gene expression has been identified as a potential risk factor for Crohn's disease. Here we report that in contrast to the predictions of single nucleotide polymorphism (SNP) analysis, patients with Crohn's disease and ulcerative colitis demonstrate elevated Bcl-3 mRNA expression relative to healthy individuals. To explore further the potential role of Bcl-3 in inflammatory bowel disease (IBD), we used the dextran-sodium sulphate (DSS)-induced model of colitis in Bcl-3−/− mice. We found that Bcl-3−/− mice were less sensitive to DSS-induced colitis compared to wild-type controls and demonstrated no significant weight loss following treatment. Histological analysis revealed similar levels of oedema and leucocyte infiltration between DSS-treated wild-type and Bcl-3−/− mice, but showed that Bcl-3−/− mice retained colonic tissue architecture which was absent in wild-type mice following DSS treatment. Analysis of the expression of the proinflammatory cytokines interleukin (IL)-1β, tumour necrosis factor (TNF)-α and IL-6 revealed no significant differences between DSS-treated Bcl-3−/− and wild-type mice. Analysis of intestinal epithelial cell proliferation revealed enhanced proliferation in Bcl-3−/− mice, which correlated with preserved tissue architecture. Our results reveal that Bcl-3 has an important role in regulating intestinal epithelial cell proliferation and sensitivity to DSS-induced colitis which is distinct from its role as a negative regulator of inflammation.

Introduction

The nuclear factor (NF)-κB transcription factor family controls the inducible expression of more than 500 genes, including cytokines, chemokines and regulators of cell survival and proliferation [1, 2]. The dual role of NF-κB as a key regulator of inflammation and cell survival makes it a critical factor in the pathogenesis of chronic diseases such as inflammatory bowel disease (IBD). Increased NF-κB activation is observed in the mucosa of IBD patients, and the requirement for NF-κB for the expression of proinflammatory cytokines supports a contributory role for NF-κB in IBD [3, 4]. Indeed, in the interleukin (IL)-10−/− mouse model of colitis, increased activation of NF-κB in myeloid cells is critical for the development of disease, while mice lacking cylindromatosis tumour suppressor (CYLD) or A20, two important negative regulators of NF-κB, show increased sensitivity to dextran sodium sulphate (DSS)-induced colitis [4-7]. Moreover, the pharmacological inhibition of NF-κB by anti-sense oligonucleotides or inhibitory peptides can prevent DSS-induced colitis in mice [8].

Genetic studies have identified an equally important role for NF-κB in maintaining the homeostasis of the intestinal epithelium. Mice lacking the NF-κB essential modulator (NEMO) in intestinal epithelial cells (NEMOΔIEC) develop spontaneous and severe colitis resulting from elevated intestinal epithelial cell apoptosis [4, 9]. A similar phenotype is observed in mice lacking both the IκB kinase α (IKKα) and IKKβ subunits in intestinal epithelial cells (IKKα\βΔIEC), and mice lacking the NF-κB subunit RelA in intestinal epithelial cells are hypersensitive to DSS-induced colitis [4, 10]. Toll-like receptors (TLRs) are the key sensors of microbial products in innate immunity and appear to be critical in initiating NF-κB activation in intestinal epithelial cells. Thus, mice lacking myeloid differentiation primary response gene 88 (MyD88), a key component downstream of a number of TLRs, are also hyper-responsive to DSS-induced colitis [11, 12]. Together, these studies indicate that while NF-κB activity is critical for inflammation in IBD, NF-κB activity in the epithelium is critical for tissue homeostasis and its inhibition can have severe consequences, including the development of IBD. Thus, a further understanding of the regulation of NF-κB during inflammation in the intestine and the contribution of components of the NF-κB pathway to inflammation and epithelial proliferation in the mucosa are critical for the development of effective therapies for IBD.

Bcl-3 is a member of the IκB family of proteins, as determined by sequence homology and the presence of ankyrin repeat domains which mediate interaction with NF-κB dimers [13-15]. Bcl-3 is largely a nuclear protein, and binds only homodimers of the p50 or p52 NF-κB subunits [14]. Interestingly, these two subunits lack a transactivation domain and thus have been regarded generally as repressors of NF-κB transcription when present in the homodimeric form. Bcl-3 is an essential negative regulator of TLR-induced responses. Bcl-3−/− macrophages and mice are hyper-responsive to TLR stimulation, and are defective in lipopolysaccharide tolerance [16]. Recently, a single nucleotide polymorphism (SNP) associated with reduced Bcl-3 gene expression has been identified as a potential risk factor for Crohn's disease (CD) [17]. However, the role of Bcl-3 in IBD has not been investigated to date.

In this study we report that our measurements of Bcl-3 mRNA in patient groups with CD, ulcerative colitis (UC) and healthy individuals reveal elevated Bcl-3 expression associated with IBD, in contrast to the predictions of the single nucleotide polymorphism (SNP) analysis [17]. To explore further the potential role of Bcl-3 in IBD we used the DSS-induced model of colitis in Bcl-3−/− mice. Considering the previously described anti-inflammatory role of Bcl-3, we were surprised to find that Bcl-3−/− mice were less sensitive to DSS-induced colitis. Measurement of the inflammatory response in the colon by analysis of the expression levels of proinflammatory cytokines and the recruitment of T cells, neutrophils, macrophage and dendritic cells revealed no significant differences between DSS-treated Bcl-3−/− and wild-type mice. Analysis of intestinal epithelial cell death and proliferation revealed increased proliferation and regeneration of the epithelium in Bcl-3−/− mice identifying Bcl-3 as an important factor in regulating epithelial cell turnover and sensitivity to colitis. Our study suggests that Bcl-3 may be an effective target for promoting regeneration of the epithelium in the colon.

Materials and methods

Mice

Bcl3−/−C57BL/6 (B6) mice were generated as described previously [15, 16]. All mice were group-housed in individually ventilated cages (IVCs) under specific pathogen-free conditions. Standard housing and environmental conditions were maintained (temperature 21°C, 12 h light/12 h darkness with 50% humidity). Animals were fed sterile standard pellet diet and water ad libitum. Animal husbandry and experimental procedures were approved by the University College Cork Animal Experimentation Ethics Committee (AEEC).

DSS-induced colitis

Mice were administered 2% DSS (45 kDa; TdB Consultancy, Uppsala, Sweden) ad libitum in their drinking water to induce colitis, as described previously [18]. DSS solutions were prepared freshly and administered on a daily basis for 6 days. This was followed by water up to day 8 to induce acute disease. Body weight, stool consistency and posture/fur texture were recorded daily to determine the daily disease activity index (DAI). DAI scoring was assessed blinded with a maximum score of 10, as described previously [18, 19]. DAI scoring combined scoring from weight loss (% change) 0–4, stool consistency 0–4 and posture/fur texture 0–2. Briefly, a percentage weight loss score of 0 = no loss, 1 = 1–3% loss, 2 = 3–6% loss, 3 = 6–9% loss and 4 = greater than 9% loss in body mass. A stool consistency score of 0 = no change, 1 = mild change, 2 = loose stool, 3 = loose stool and rectal bleeding, 4 = diarrhoea and rectal bleeding. A fur and posture score 0 = no change, 1 = mild hunched posture, 2 = hunched posture and reduced movement. Mice were killed at day 8 with colons removed from anus to caecum and washed in phosphate-buffered saline (PBS). Colons were measured and cut longitudinally dividing into the distal and proximal colon. Both proximal and distal colons were weighed and processed for histology, protein and quantitative reverse transcription–polymerase chain reaction (qRT–PCR) analysis.

Colon histology

Distal colons (3 cm) were cut longitudinally and into three sections. One section was rolled in a ‘swiss roll’ fashion and frozen in optimal cutting temperature (OCT) tissue-freezing medium (Tissue Tek, Sakura Finetek, Torrance, CA, USA) using liquid nitrogen. Frozen sections (6 μm) were fixed in ice-cold acetone/ethanol 3:1 solution and stained with haematoxylin and eosin (H&E) according to standard histological staining procedures. Stained sections were analysed and scored using a light microscope (Olympus BX51; Olympus, Hamburg, Germany). Images were captured using Cell F software (Olympus). Images captured are representative of greater than seven fields of view at ×20 magnification per mouse. Histological scoring was performed in a blinded fashion. Scoring of tissue damage was quantified as described previously with a maximum combined score of 12 [19, 20] as follows: 0, no infiltration, no injury, no crypt damage; 1, minor infiltration, mucosal injury, damage at crypt base; 2, moderate infiltration (foci formation), mucosal and submucosal injury, damage at crypt base and centre; 3, severe infiltration, transmural injury, only epithelium intact; and 4, loss of whole crypt and epithelium.

Gene expression analysis

Distal colon tissue gene expression was measured by qRT–PCR. Distal colons (3 cm) were divided into three sections with one section frozen at −80°C in RNAlater (Sigma Aldrich, Dublin, Ireland). Colon tissue samples were thawed on ice and transferred to magNALyser green bead tubes (Roche Applied Sciences, West Sussex, UK) and homogenized using the magNALyser homogenizer three times for 15 s at ×6500 (Roche). Colonic tissue was homogenized in RLT lysis buffer (Qiagen Ltd, Manchester, UK) with homogenized samples centrifuged for 5 min at 4°C at 200 g. Supernatants were stored at −80°C until required. Total RNA was extracted using the RNeasy mini kit (Qiagen). One μg total RNA was used to synthesize cDNA with random hexamer primers using transcriptor reverse transcriptase (Roche). qRT–PCR was performed using the LightCycler 480 (Roche). Primers were designed using the Universal Probe Library system (Roche), as follows: IL-6 (forward = TCTAATTCATATCTTCAACCAAGAGG, reverse = TGGTCCTTAGCCACTCCTTC); tumour necrosis factor (TNF)-α (forward = TCTTCTCATTCCTGCTTGTGG, reverse = GGTCTGGGCCATAGAACTGA); IL-1β (forward = TGTAATGAAAGACGGCACACC, reverse = TCTTCTTTGGGTATTGCTTGG); CXCL1 (forward = AGACTCCAGCCACACTCCAA, reverse = TGACAGCGCAGCTCATTG); IL-22 (forward = TTTCCTGACCAAACTCAGCA, reverse = CTGGATGTTCTGGTCGTCAC); and IL-17A (forward = CAGGGAGAGCTTCATCTGTGT, reverse = GCTGAGCTTTGAGGGATGAT) was measured and normalized to 18S (forward = AAATCAGTTATGGTTCCTTTGGTC, R = GCTCTAGAATTACCACAGTTATCCAA). Gene expression changes were calculated using the 2-ΔΔCT method.

Human tissue arrays (CD/Colitis cDNA Array; Origene, Rockville, MD, USA) were used to measure Bcl-3 expression. Gene expression was measured using the LightCycler 480 system in combination with Taqman gene expression assay for Bcl-3 (Applied Biosystems/Life Technologies, Grand Island, NY, USA). Relative mRNA was calculated using the 2-ΔΔCT method. Transcriptional profiling of CD and UC tissue was performed using a data set of sigmoid biopsy patient samples published by Costello et al. (GEO data set ID GDS1330) [21] (CD n = 10, UC n = 10, normal controls n = 11).

Terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labelling (TUNEL)

The extent of apoptosis in colonic tissue between groups was measured by TUNEL. Six-μm colonic tissue sections were incubated with 3% H2O2 and a 4% diethyl pyrocarbonate (DEPC) solution to eliminate background from both peroxidase and endonuclease enzyme activity in the tissue. The colon sections were incubated for 1 h in a reaction using terminal deoxynucleotidyl transferase (Promega) and fluorescein 12-dUTP (Roche). Nuclei were counterstained with Hoechst (Molecular Probes/Life Technologies, Grand Island, NY, USA). Stained sections were analysed using a fluorescence microscope (Olympus BX51; Olympus). Fluorescence images were captured using Cell F software (Olympus). Images captured are representative of greater than seven fields of view at ×20 magnification per mouse.

Immunofluorescence staining

Frozen sections (6 μm) were fixed in ice-cold acetone/ethanol 3:1 solution and blocked with blocking buffer [10% serum, 5% fish gelatine, 0·05% Tween-20, 1% bovine serum albumin (BSA), 0·1% sodium azide]. Colon sections were incubated with anti-mouse Ki67 (Biolegend, San Diego, CA, USA) and counterstained with Hoechst (Molecular Probes). Stained sections were mounted with Prolong Gold anti-fade mounting medium (Molecular Probes) and visualized using a fluorescence microscope (Olympus BX51; Olympus). Fluorescence images were captured using Cell F software (Olympus). Images captured are representative of greater than seven fields of view at ×20 magnification per mouse.

Western blotting

Frozen distal colon tissue samples were thawed, transferred to magNALyser green bead tubes (Roche) and homogenized using the magNALyser homogenizer three times for 15 s at 6500 g (Roche). Total protein was isolated by lysing the distal colon tissue in RIPA buffer [150 mM NaCl, 50 mM Tris-Cl, pH 7·4, 1% NP-40, 0·25% sodium deoxycholate, 1 mM Na3VO4, 1 mM ethylenediamine tetraacetic acid (EDTA)] supplemented with a protease and phosphatase inhibitor cocktail (Sigma). Total protein was resolved by sodium dodcyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) gels, transferred to polyvinylidene difluoride (PVDF) membrane and immune blotted for cleaved caspase-3 (Cell Signalling, Boston, MA, USA), and β-actin (Sigma).

Statistical analysis

Statistical analysis was determined using one-way analysis of variance (anova)/two-way anova with post-hoc analysis (Tukey's post-hoc test and Bonferroni's post-hoc test). qRT–PCR expression data were calculated using the 2-ΔΔCT followed by unpaired t-test and Mann–Whitney t-test to compare differences between groups. Statistical analysis was performed using GraphPad software (San Diego, CA, USA). Data are represented by mean ± standard error of the mean with P < 0·05 considered statistically significant.

Results

Elevated BCL3 mRNA is found in the colon of IBD patients

To assess the role of Bcl-3 in inflammatory bowel disease we initially analysed the Bcl-3 expression levels from a previously published study which identified a large number of genes associated with inflammatory bowel diseases [21]. In that study, transcriptional profiles were generated from biopsies taken from the sigmoid colon of patients with CD (n = 10) and UC (n = 10) and those of normal controls (n = 11). Our bioinformatics analysis of this data set revealed that Bcl-3 mRNA expression levels were increased significantly in both CD (P < 0·01) and UC (P < 0·05) (Supporting Information, Fig. S1). The elevated Bcl-3 mRNA levels in CD and UC were unexpected, considering that an SNP in the BCL3 locus predicted to reduce expression of Bcl-3 mRNA was associated with CD [17]. To confirm our analysis we next measured Bcl-3 mRNA expression by qRT–PCR in an additional, independent patient cohort of 21 CD, 21 UC and six normal control colon tissue samples. Importantly, this independent analysis of Bcl-3 mRNA expression also revealed a statistically significant increase in Bcl-3 gene expression in CD tissue samples relative to normal healthy controls (P < 0·05) (Fig. 1a). Moreover, the magnitude of increase of Bcl-3 mRNA levels in CD and UC relative to normal controls was similar in our tissue samples and in those contained in the previous microarray analysis. Next we measured Bcl-3 gene expression levels in wild-type mice receiving 6 days treatment with 2% DSS followed by 2 days without DSS to induce colitis. We found an increase in Bcl-3 mRNA in wild-type DSS-treated mice relative to untreated control mice (Fig. 1b). Taken together, these data demonstrate a strong correlation between increased Bcl-3 mRNA expression and colitis in both a murine model and human IBD.

Figure 1.

Bcl-3 expression in inflammatory bowel disease (IBD). (a) Bcl-3 mRNA levels in Crohn's disease (CD, n = 21) and ulcerative colitis (UC, n = 21) tissue samples relative to normal (N, n = 6) controls (P < 0·05 *). (b) Bcl-3 mRNA levels in colon tissue of untreated and dextran-sodium sulphate (DSS)-treated wild-type mice. Mice were treated with 2% DSS for 6 days followed by 2 days on normal drinking water before tissue was harvested. Data are expressed as means ± standard error of the mean. Statistical significance was determined using Mann–Whitney t-tests.

Bcl-3−/− mice are protected against DSS-induced colitis

In order to investigate further the potential role of Bcl-3 in IBD we performed DSS-induced acute colitis in Bcl-3−/− and wild-type littermate controls. Wild-type and Bcl-3−/− mice were treated with 2% DSS in their drinking water for 6 days, after which they were monitored for an additional 2 days, during which time they received normal drinking water. Within 4 days of beginning DSS treatment both Bcl-3−/− and wild-type mice developed characteristic symptoms associated with DSS-induced colitis. These included hunched posture and changes in stool consistency, including rectal bleeding and diarrhoea. By day 8 following DSS treatment wild-type mice had lost greater than 12% of their body weight (day 6; P < 0·01, day 7; P < 0·001, day 8; P < 0·001; Fig. 2a). In contrast, DSS-treated Bcl-3−/− mice did not demonstrate any significant loss of body mass when compared to untreated Bcl-3−/− mice up to 8 days following the initial DSS treatment (Fig. 2a). When rectal bleeding, diarrhoea, hunched posture and weight loss of DSS-treated and -untreated mice were scored and combined to give a DAI score we found that Bcl-3−/− mice develop a significantly less severe form of DSS-induced colitis (Fig. 2b). The reduced disease observed in Bcl-3−/− mice was not a consequence of reduced DSS intake, as water consumption was equivalent between groups during the experiment (data not shown). These data demonstrate clearly that Bcl-3 contributes to colitis.

Figure 2.

Bcl3−/− mice develop milder dextran-sodium sulphate (DSS)-induced colitis. (a) Weight loss (% change) was measured over time with differences between wild-type 2% DSS and wild-type healthy controls found to be statistically significant from days 6–8 (**P < 0·01; ***P < 0·001; ***P < 0·001), respectively. Differences in weight loss between wild-type 2% DSS and Bcl-3−/− healthy controls were found to be significant at day 8 (***P < 0·001). (b) Differences in disease activity index (DAI) scores were observed starting from days 5–8 in both wild-type and Bcl-3−/− DSS groups relative to both healthy controls (***P < 0·001). Differences in weight loss and DAI score were calculated over time by two-way analysis of variance (anova) with Bonferroni's post-hoc test. (c–d) Differences in colon length and distal colon weight between in both wild-type and Bcl-3−/− mice relative to healthy controls were found to be statistically significant (P < 0·05). Data are expressed as mean ± standard error of the mean. Statistical significance was determined using Mann–Whitney t-tests (n = 7/group).

Macroscopic analysis of colon tissue was performed on day 8 after the beginning of DSS treatment. Wild-type DSS-treated mice demonstrated significant shortening of the colon when compared to untreated controls (P < 0·05; Fig. 2c). Surprisingly, a similar degree of colon shortening was observed in DSS-treated Bcl-3−/− mice when compared to untreated Bcl-3−/− controls (Fig. 2c). Moreover, the significantly increased colonic weight of DSS-treated wild-type mice relative to untreated controls was also observed in DSS-treated Bcl-3−/− mice (P < 0·05, Fig. 2d). Thus, although the macroscopic inflammation of colonic tissue was similar in both DSS-treated wild-type and Bcl-3−/− mice, the clinical indices of the DSS-induced colitis, in particular weight loss, were reduced significantly in Bcl-3−/− mice.

Bcl-3−/− mice show reduced tissue pathology following DSS treatment

To investigate further the differences in DSS-induced colitis between wild-type and Bcl-3−/− mice, we performed a histological examination of distal colon tissue sections from untreated and DSS-treated wild-type and Bcl-3−/− mice (Fig. 3a). No differences were observed between untreated wild-type and untreated Bcl-3−/− distal colonic tissue samples by H&E staining. Both wild-type and Bcl-3−/− mice displayed normal epithelial architecture with intact goblet cells and crypts with no discernible inflammatory influx. DSS treatment of wild-type mice induced a dramatic alteration in the colonic mucosal tissue with extensive oedema, large cellular infiltrates and a severe loss of tissue organization with destruction of crypts and loss of goblet cells. Although histological analysis revealed similar levels of oedema and cellular infiltrates in Bcl-3−/− mice, there was significantly less destruction of the tissue architecture following DSS treatment (Fig. 3a). Quantitative histopathological analysis of the distal colon tissue from DSS-treated Bcl-3−/− mice revealed significantly reduced epithelium damage and loss of tissue architecture compared to wild-type mice (Fig. 3b). However, there were no significant differences in the extent of inflammation (Fig. 3c) and the degree of cellular infiltration and oedema (Fig. 3d) between DSS-treated wild-type and Bcl-3−/− mice. This histological analysis provides insight into the reduced weight loss and overall clinical disease score observed in DSS-treated Bcl-3−/− mice relative to wild-type mice, which would appear to result from an intact or regenerated epithelium rather than reduced leucocyte infiltration.

Figure 3.

Reduced histological damage in Bcl-3−/− mice following dextran-sodium sulphate (DSS)-induced colitis. (a) Representative haematoxylin and eosin (H&E) of untreated (control) and DSS-treated wild-type and Bcl-3−/− mice. Histological sections were scored for cellular infiltration (b), extent of injury (c), epithelium and crypt damage (d). DSS-induced cellular infiltration, extent of injury and epithelium and crypt damage scores were increased in wild-type and Bcl-3−/− mice relative to untreated controls (P < 0·05). DSS-treated wild-type mice showed greater crypt and epithelium damage relative to DSS-treated and Bcl-3−/− mice (P < 0·01). Data are representative of greater than seven fields of view per tissue section at ×20 (n = 7 per group).

DSS induces similar levels of cytokines in the colon of Bcl-3−/− and wild-type mice

Although histological analysis showed similar levels of oedema and leucocyte infiltration in DSS-treated wild-type and Bcl-3−/− mice, it is possible that the inflammation may be qualitatively different between these groups. In order to characterize the inflammation associated with DSS-induced colitis in Bcl-3−/− mice, we next measured inflammatory gene expression in distal colon tissue from untreated and DSS-treated wild-type and Bcl-3−/− mice using qRT–PCR. Surprisingly, although Bcl-3 has been described previously as a negative regulator of Toll-like receptor-induced proinflammatory gene expression, we found no significant difference in the expression of TNF-α, IL-6, CXCL1 and IL-1β between DSS-treated wild-type and Bcl-3−/− mice (Fig. 4a).

Figure 4.

Normal inflammatory cytokine expression in mucosal tissue of dextran-sodium sulphate (DSS)-treated Bcl-3−/− mice. Distal colon tissues were analysed for changes immune gene expression relative to the housekeeping gene 18 s rRNA. (a) Increased proinflammatory gene expression was found in the DSS group of wild-type and Bcl-3−/− mice relative to untreated control groups (*P < 0·05). No statistical significance in gene expression changes was observed between wild-type and Bcl-3−/− mice. (b) Increased interleukin (IL)-17A and IL-22 gene expression in DSS-treated wild-type and Bcl-3−/− mice relative to untreated controls. Data are represented as relative mRNA expression as determined by the 2-ΔΔCT method. Statistical significance was calculated using Mann–Whitney t-tests (*P < 0·05). Data are expressed as mean ± standard error of the mean (n = 7).

Recent studies have identified a protective role for the cytokines IL-17A and IL-22 [22-24] in DSS colitis by inducing anti-bacterial peptide expression and epithelial cell regeneration in the colon. To assess any role for these cytokines in the observed resistance of Bcl-3−/− mice to DSS-induced colitis and maintenance of intestinal epithelium we next measured their expression in the colon of wild-type and Bcl-3−/− mice. In line with previous reports, the expression of both IL-17A and IL-22 is induced robustly by DSS treatment in wild-type mice; however, no significant differences in the expression of these cytokines was found between DSS-treated wild-type and Bcl-3−/− mice (Fig. 4b). We next analysed the cellular composition of the leucocyte infiltrates in DSS-treated wild-type and Bcl-3−/− mice using immunofluorescence microscopy and antibodies against the cell surface markers F4/80 (macrophage), CD3 (T Cell), Ly6G (neutrophil) and CD11c (dendritic cells) (Fig. 5a). Quantitative analysis of tissue sections demonstrated recruitment of macrophage, neutrophils and, to a lesser degree, T cells and dendritic cells to the distal colon of DSS-treated mice. No significant differences in the recruitment of these cell types were found between wild-type and Bcl-3−/− mice (Fig. 5b). These data demonstrate that the inflammatory component of DSS-induced colitis is similar between wild-type and Bcl-3−/− mice and suggest that the reduced susceptibility of Bcl-3−/− mice may result from altered epithelial responses to treatment.

Figure 5.

Analysis of inflammatory infiltration during colitis in Bcl-3−/− mice. (a) Immunofluorescent staining of 6 μm distal colon section at ×40 for macrophages (F/480), dentritic cells (CD11c), T cells (CD3) and neutrophils (Ly6G). (b) Quantification of F4/80, CD3, Ly6G- and CD11c-positive cellular populations was performed on seven fields of view per section. Data are representative of scoring of all sections per treatment group; n = 7.

Increased epithelial cell proliferation in DSS-treated Bcl-3−/− mice

Because DSS induces epithelial cell damage to initiate colonic inflammation and colitis we next measured cell death in the colon of wild-type and Bcl-3−/− mice using terminal dUTP nick end labelling (TUNEL) of tissue sections followed by fluorescence microscopy analysis. In both untreated wild-type and untreated Bcl-3−/− mice we observed a small number of TUNEL-positive nuclei in the top of the crypt representing the normal turnover of epithelial cells in this tissue (Fig. 6a). However, following DSS treatment we observed a dramatic increase in TUNEL-positive cells in both wild-type and Bcl-3−/− mice. Quantitative analysis of TUNEL staining demonstrated no significant differences in the number of cells undergoing apoptosis in both groups. Immunoblot analysis of caspase-3 cleavage in colonic tissues also demonstrated a significant increase in DSS-induced apoptosis in wild-type and Bcl-3−/− mice following DSS treatment (Fig. 6b). Densitometric analysis of cleaved caspase-3 levels normalized to β-actin levels revealed no significant difference between wild-type and Bcl-3−/− mice (Supporting Information, Fig. S2). Analysis of the mRNA levels of the apoptotic regulators p53 up-regulated modulator of apoptosis (PUMA), Bcl-XL, cellular inhibitor of apoptosis protein 1/2 (cIAP1/2) and phorbol-12-myristate-13-acetate-induced (NOXA) by qRT–PCR also revealed no significant differences expression between wild-type and Bcl-3−/− mice (Fig. 6c).

Figure 6.

Analysis of cell death during colitis in Bcl-3−/− mice. (a) Terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labelling (TUNEL)-stained distal colon 6 μm section. Arrows indicate TUNEL-positive cells. Quantification of TUNEL-positive cells was performed on seven fields of view per tissue section at ×20 (n = 7 per group). (b) Western blotting for inactive (32 KDa) and active (17 KDa) caspase-3 in distal colon tissue of control and dextran-sodium sulphate (DSS)-treated wild-type and Bcl-3−/− mice. (c) Quantitative polymerase chain reaction (qPCR) analysis of p53 up-regulated modulator of apoptosis (PUMA), Bcl-XL, cellular inhibitor of apoptosis protein 1/2 (cIAP1/2) and phorbol-12-myristate-13-acetate-induced (NOXA) mRNA in distal colon tissue of control and DSS-treated wild-type and Bcl-3−/− mice.

We next assessed epithelial cell proliferation in tissue sections using the cell proliferation marker Ki67. Immunofluorescence microscopy analysis of untreated wild-type and Bcl-3−/− mucosal colonic tissue revealed equivalent numbers of Ki67-positive cells at the base of the crypts (Fig. 7a,b). Ki67 staining was largely absent in wild-type mucosal tissue following DSS treatment and coincided with the extensive destruction and loss of tissue architecture (Fig. 3). In contrast, widespread and strong Ki67 staining was found throughout the crypts of colonic tissue taken from DSS-treated Bcl-3−/− mice, indicating significantly enhanced proliferation of Bcl-3−/− epithelial cells following treatment (Fig. 7a). Immunofluorescence microscopy analysis of Bcl-3 protein in tissue sections was unsuccessful using commercially available antibodies; however, previous studies have demonstrated Bcl-3 mRNA expression in intestinal epithelial cells [25, 26]. Taken together, these data suggest that Bcl-3−/− mice develop less severe clinical and histopathological colitis due to an increase in epithelial proliferation, which leads to regeneration of the damaged epithelium. Our data also demonstrate that this regeneration occurs despite the presence of ongoing inflammation in the colonic mucosa.

Figure 7.

Elevated cellular proliferation in dextran-sodium sulphate (DSS)-treated Bcl-3−/− mice. (a) Ki67 immunofluorescence staining of 6 μm distal colon sections. Data are representative of greater than seven fields of view per tissue section at ×20 (n = 7 per group). Inset, magnified section of Ki67-positive cells indicated by arrowhead. (b) Quantification of Ki67-positive cells was performed on seven fields of view per tissue section at ×20 (n = 7 per group).

Discussion

In this study we investigated the expression of Bcl-3 in human IBD and also the role of Bcl-3 in DSS-induced colitis in the mouse. We found that Bcl-3−/− mice develop less severe colitis compared to littermate control wild-type mice. These findings were unexpected, given the previously described role of Bcl-3 as a negative regulator of inflammatory gene expression [16] and the recent identification of reduced Bcl-3 expression as potential risk factors for CD [17]. However, the resistance of Bcl-3−/− mice to experimentally induced colitis correlates with our analysis of Bcl-3 expression in the colon of IBD patients, which was significantly increased when compared to healthy individuals. It is possible that the identified SNPs may lead to increased Bcl-3 expression rather than decreased expression as predicted. Thus, our findings suggest that increased expression of Bcl-3 rather than reduced expression may be a potential risk factor for IBD. Our study also identifies a novel role for Bcl-3 in regulating intestinal epithelial cell proliferation during DSS-induced colitis.

Analysis of cytokine expression during DSS-induced colitis in Bcl-3−/− mice revealed a robust inflammatory response following DSS treatment characterized by significantly elevated levels of proinflammatory cytokines TNF-α, IL-6 and IL-1β. The levels of these cytokines was similar to wild-type mice, indicating that Bcl-3 does not act as a negative regulator of TNF-α, IL-6 and IL-1β expression in the context of DSS-induced colonic inflammation. Histological analysis supported this observation further, as significant oedema and leucocyte infiltration were present in Bcl-3−/− colonic tissue sections and to a similar degree to that seen in wild-type mice. Furthermore, equivalent composition of cellular infiltrates was observed between wild-type and Bcl-3−/− mice, which demonstrated that the inflammation was qualitatively, as well as quantitatively, similar to wild-type mice. These data suggest that Bcl-3 may not play a significant role in the regulation of inflammation in the colon.

Despite a robust inflammatory response following DSS treatment, the colonic tissue architecture in Bcl-3−/− mice, in particular the epithelial features, remain intact. Following DSS treatment intestinal epithelial cell proliferation in Bcl-3−/− mice was enhanced significantly, whereas in wild-type mice it was absent. The increased proliferation in Bcl-3−/− mice correlates with the maintenance of tissue architecture and structure and suggests that the resistance to DSS-induced colitis of Bcl-3−/− mice results from increased regeneration of the epithelium. It is also noteworthy that Bcl-3 acts a negative regulator of myeloid progenitor proliferation and differentiation, and is essential for limiting granulopoiesis under inflammatory conditions [27]. This study identifies a novel role for Bcl-3 in regulating intestinal epithelial cell proliferation under inflammatory but not homeostatic conditions. Our identification of Bcl-3 as a negative regulator of intestinal epithelial cell proliferation during colitis suggests additional physiological functions for Bcl-3 beyond its role as a negative regulator of proinflammatory gene expression.

The dual role of NF-κB as a key mediator of inflammation and a critical driver of epithelial cell survival and proliferation has rendered it a complex and difficult therapeutic target in IBD. Transgenic mice in which NF-κB activity has been inhibited selectively in the intestinal epithelium develop spontaneous colitis due to failure of the epithelial barrier function, while an increase in intestinal NF-κB activity also leads to severe inflammation [4]. The data obtained in this study, however, suggest that certain regulatory components of the NF-κB pathway such as Bcl-3 may play a more important role in the epithelium rather than the immune system in the colon. We have demonstrated previously that Bcl-3 expression is induced by inflammation [16]. Given that the proliferation of intestinal epithelial cells is normal in Bcl-3−/− mice, it is probable that inflammation-induced expression of Bcl-3 in the epithelium during colitis contributes to the development of disease. Thus, by targeting Bcl-3 it may be possible to enhance epithelial cell proliferation and regeneration without exacerbating inflammation in the intestine. The potential therapeutic benefits to IBD are highlighted by the reduced clinical score and lack of weight loss in DSS-treated Bcl-3−/− mice.

In summary, we describe a novel function for Bcl-3 in regulating epithelial cell proliferation during DSS-induced colitis. The increased epithelial cell proliferation and regeneration in Bcl-3−/− mice supports further a role for NF-κB in maintaining the integrity of the intestinal epithelium. This report suggests that targeting Bcl-3 in colitis may be therapeutically beneficial in IBD through increasing tissue regeneration and repair in the colon without exacerbating the inflammatory response.

Acknowledgements

This research was supported by Science Foundation Ireland (grant no. 08/IN.1/B1843 and CSET grant no. 07/CE/B1368) and the Marie Curie International Re-integration Grant programme.

Disclosures

The authors have no conflicts of interest to declare.

Ancillary