Contributed by Michael Karin, August 20, 2008 (received for review July 25, 2008)
NF-κB is a key transcriptional regulator of inflammatory responses, but also controls expression of prosurvival genes, whose products protect tissues from damage and may thus act indirectly in an antiinflammatory fashion. The variable importance of these two distinct NF-κB-controlled responses impacts the potential utility of NF-κB inhibition as a treatment strategy for intractable inflammatory conditions, such as inflammatory bowel disease. Here, we show in murine models that inhibition of IKKβ-dependent NF-κB activation exacerbates acute inflammation, but attenuates chronic inflammatory disease in the intestinal tract. Acute ulcerating inflammation is aggravated because of diminished NF-κB-mediated protection against epithelial cell apoptosis and delayed mucosal regeneration secondary to reduced NF-κB-dependent recruitment of inflammatory cells that secrete cytoprotective factors. In contrast, in IL-10-deficient mice, which serve as a model of chronic T cell-dependent colitis, ablation of IKKβ in the intestinal epithelium has no impact, yet IKKβ deficiency in myeloid cells attenuates inflammation and prolongs survival. These results highlight the striking context and tissue dependence of the proinflammatory and antiapoptotic functions of NF-κB. Our findings caution against the therapeutic use of IKKβ/NF-κB inhibitors in acute inflammatory settings dominated by cell loss and ulceration.
Inflammation is a fundamental physiological process that protects the host against microbial challenges, but can also do great harm if activated inappropriately or excessively. For example, inflammatory bowel disease (IBD) is characterized by persistent inflammation in the colon or small intestine in the apparent absence of pathogenic microbes. The disease commonly follows a chronic relapsing course with clinically quiescent periods followed by bouts of severe intestinal inflammation, which are accompanied by abdominal pain, diarrhea, and weight loss. IBD is thought to result from a combination of genetic and environmental factors, but the underlying causes are probably diverse, depending on specific disease manifestations and patient subsets (1). In the absence of a comprehensive pathophysiologic understanding of the disease, current therapeutic interventions are largely nonspecific and targeted toward attenuating different aspects of the inflammatory process without eliminating its initial triggers.
A key regulator of the inflammatory response is the transcription factor NF-κB, which consists in mammals of homodimers or heterodimers of different NF-κB/Rel proteins (2). NF-κB activates transcription of numerous target genes, many of which encode cytokines and adhesion molecules that orchestrate the influx and activation of leukocyte subsets in sites of infection or tissue damage (3). In addition, NF-κB target genes code for proteins that protect cells against apoptosis and necrosis induced by a wide range of noxious stimuli (4). In unstimulated cells, NF-κB is retained in the cytoplasm by specific inhibitors, the IκB proteins. Stimulation with microbial ligands of Toll-like receptors (TLRs) or the prototypic proinflammatory cytokines, TNF-α or IL-1, induces IκB phosphorylation and ubiquitin-dependent proteasomal degradation resulting in nuclear entry of NF-κB dimers to initiate target gene transcription. IκB phosphorylation is catalyzed by the IκB kinase (IKK) complex composed of two catalytic subunits, IKKα and IKKβ, in conjunction with a scaffolding and regulatory protein, IKKγ/NEMO (5). Whereas IKKα is activated by an only limited set of stimuli, IKKβ activation occurs upon receptor-mediated stimulation by a broad set of microbial or host-derived ligands (2). Inhibition of IKKβ suppresses the production and secretion of the prototypic proinflammatory cytokine TNF-α and attenuates disease in animal models of rheumatoid arthritis, inflammation-induced bone loss, and allergen-induced airway disease (6–9). Furthermore, inhibition of NF-κB by either antisense oligonucleotides directed against RelA/p65 or a small-molecule inhibitor of IKKβ appeared to ameliorate disease in mouse models of intestinal inflammation (10–12), suggesting that NF-κB-directed therapy could be a valuable novel strategy in IBD therapy. In sharp contrast, however, conditional ablation of IKKβ in intestinal epithelial cells (IECs) caused increased inflammation in an acute, chemically induced colitis model (13) and loss of IKKγ in IECs caused spontaneous colitis (14). Moreover, recent findings demonstrating an important antiinflammatory function for NF-κB based on suppression of IL-1β processing and secretion (15) raise concerns about the therapeutic potential of IKKβ inhibitors (16). We undertook the present study to resolve these discrepancies and determine the role of canonical, IKKβ-dependent NF-κB activation in different cell types during acute ulcerating and chronic colitis driven by immune dysregulation rather than epithelial injury.
Loss of IKKβ in epithelial cells renders the cells vulnerable to apoptosis after acute exposure to various stress insults (13, 17, 18). Accordingly, we observed a protective effect of IEC-IKKβ in acute colitis caused by suppression of epithelial cell death early in disease induction and hence maintenance of epithelial integrity and barrier function (13). Cell culture studies have shown that NF-κB can also control epithelial cell migration (19), a process that contributes to epithelial layer healing after acute injury. To determine whether epithelial IKKβ has a role in the early healing phase after acute colitis, we administered dextran sulfate sodium (DSS) in the drinking water to mice lacking IKKβ in IECs (IkkβDIEC mice) and littermate controls for 5 days followed by 16 days of regular drinking water (Fig. 1A). Histological analysis carried out 21 days after initiation of DSS exposure, corresponding to the late healing phase in this colitis model, demonstrated that IEC-specific deletion of IKKβ was associated with more severe mucosal inflammation and greater areas of ulceration in comparison with littermate controls (Fig. 1 B–E).
IKKβ inhibition delays healing in acute ulcerating colitis. (A) Schematic overview of the DSS-induced colitis model. (B–E) Histological score (B), size of ulcerations (C), and representative H&E-stained sections (D and E) of IkkβF/ F and IkkβDIEC colons during the late healing phase of acute colitis, 21 days after the beginning of DSS administration. (F) Schematic overview of ML120B treatment in DSS colitis. (G) EMSA shows inhibition of total NF-κB binding activity in whole colonic extracts of WT mice treated with ML120B (80 mg/kg, twice a day) for 5 days compared with vehicle-treated controls after they had received DSS for 5 days. (H–K) Histological score (H), size of ulcerations (I), and representative H&E-stained sections (J and K) of the colon of WT mice left untreated or treated with ML120B for 5 days after the end of DSS administration. P values were determined by Student's t test and were considered significant when <0.05. (Magnifications: D, E, J, and K, 20×.)
To exclude the possibility that the observed late healing phenotype might have been caused by increased early damage induced by increased apoptosis in IKKβ-deficient enterocytes (13), we treated WT mice after completion of DSS administration with a highly specific IKKβ inhibitor, ML120B (20), or vehicle for 5 days (Fig. 1F). The inhibitor was highly effective in blocking NF-κB activation in the inflamed colon (Fig. 1G). Ten days after the start of DSS application, corresponding to the beginning of the healing phase, mice were examined histologically. Pharmacological IKKβ inhibition, like enterocyte-specific genetic ablation, induced greater areas of ulceration and more severe mucosal inflammation compared with vehicle-treated controls (Fig. 1 H–K). These results strongly support the notion that IKKβ has additional functions in IECs beyond protection against apoptosis and further demonstrate that IKKβ exerts an overall protective function in acute colitis independently of the initial damage.
To unravel the mechanisms responsible for the epithelial healing defect in ML120B-treated animals, we isolated enterocytes and performed a microarray analysis by using Affymetrix MOE430A 2.0 GeneChips (containing >22,000 probe sets that represent ≈14,000 annotated genes). Mice were analyzed on day 7, at which time tissue damage and inflammation is most severe and the process of epithelial healing commences. We focused on decreased gene expression that correlated with IKKβ inhibition to identify critical protective genes. Expression of ≈100 genes was >2.5-fold decreased in IECs of ML120B-treated mice compared with vehicle-treated controls. More than 30% of the identified genes represented known NF-κB target genes, including Cxcl1, Cxcl2, Cxcl5, Cxcl9, Cxcl10, Cxcl20, Ccl5, Ccl8, Icam1, and Pla2g2a (21, 22). Down-regulation of several of these genes by ML120B after DSS treatment was confirmed by real-time PCR analysis (Fig. 2A). However, the products of these NF-κB target genes have proinflammatory functions, so their down-regulation did not explain the increase in inflammation after ML120B treatment.
IKKβ inhibition suppresses NF-κB and STAT3 target gene expression in enterocytes during initiation of healing of acute ulcerating colitis. (A) Expression of NF-κB and STAT3 target genes. Relative mRNA levels were determined by real-time PCR in isolated IECs from mice that had received DSS for 5 days and were either left untreated or treated with ML120B for 2 days. Data are mean ± SE. *, P < 0.05 by t test. (B) HSP70 expression and activation of STAT3 were determined by immunoblot analysis in IECs of mice treated as in A. (C) ML120B has no effect on phosphorylation of STAT3 in colon cancer cells treated with IL-6 for 30 min.
Surprisingly, the microarray analysis revealed a second group of down-regulated genes, including Hspa1a, Tgtp, Ifitm1, Igtp, Ifi47, Iigp, H2-Aa, H2-Ab1, and Socs3, whose expression is known to be controlled primarily by STAT and IRF transcription factors. Of particular interest in this group was Hspa1a, which encodes inducible heat shock protein (HSP) 70 whose epithelial expression provides protective functions during DSS-induced colitis (23). Immunoblot analysis confirmed nearly complete absence of HSP70 expression in enterocytes from ML120B-treated mice after DSS administration (Fig. 2B). Loss of HSP70 expression was paralleled by decreased activation of epithelial STAT3 (Fig. 2B), a key regulator of Hspa1a transcription (24). No differences were observed in STAT1 phosphorylation (data not shown). The attenuation of STAT3 activation was not related to any direct effects of ML120B on epithelial STAT3 signaling (Fig. 2C), consistent with a previous report that ML120B is a highly specific IKKβ inhibitor (20). These results suggest that inhibition of NF-κB indirectly attenuates epithelial STAT3 signaling, resulting in loss of cytoprotective HSP70 in IECs.
To test the hypothesis that IKKβ inhibition reduces mucosal production of STAT3 activators in DSS-induced colitis, we isolated RNA from whole colonic mucosa on day 7 after initiation of DSS administration and examined expression of known STAT3 activators, including type I and II IFNs, as well as IL-6 and IL-10 cytokine family members. No expression differences were detected in mRNAs coding for IL-6, IL-10, IFN-α, IFN-β, or IFN-γ in ML120B-treated and control mice (Fig. 3A and data not shown). However, ML120B inhibited the colitis-associated increase in IL-11 and IL-22 mRNAs in whole colonic mucosa (Fig. 3A). To examine whether IL-11 and IL-22 are directly regulated by NF-κB, we examined mRNA expression in control and IKKβ-deleted bone marrow-derived macrophages (15) before and after LPS stimulation. LPS induced expression of both IL-11 and IL-22 mRNAs in control macrophages. Ablation of IKKβ did not attenuate this induction but rather enhanced it (Fig. 3B), indicating that NF-κB is not required for induction of IL-11 and IL-22 expression. The expression of two known NF-κB targets, IL-6 and IL-10, was inhibited, as expected, in IKKβ-deficient macrophages (Fig. 3B). These data suggest that ML120B did not suppress mucosal IL-11 and IL-22 mRNAs by direct inhibition of their gene expression.
IKKβ inhibition prevents IL-11 and IL-22 expression in colonic mucosa and recruitment of critical inflammatory cells. (A) Relative mRNA levels were determined by real-time PCR in whole colonic mucosa of mice that received DSS for 5 days and were either left untreated or treated with ML120B for 2 days. (B) Relative mRNA levels in bone marrow-derived macrophages of IkkβF/ F and IkkβD mice (15) before and after stimulation with LPS (100 ng/ml) for 4 h. (C) Relative mRNA levels of cell type-specific markers were determined by real-time PCR in whole colonic mucosa of mice that had received DSS for 5 days and were either left untreated or treated with ML120B for 2 days. Data are mean ± SE. *, P < 0.05 by t test.
As an alternative, we considered whether ML120B inhibited the recruitment of cells that produce these cytokines into the mucosa. We therefore assayed the expression of various markers for myeloid, natural killer (NK), and T cells, including CD11c, F4/80, Gr-1, CD49b, CD4, CD8, and FoxP3. Interestingly, ML120B led to a marked decrease in expression of F4/80 and CD4 after DSS administration (Fig. 3C), indicating a decrease in macrophages and CD4+ T cells in the lamina propria of inhibitor-treated mice. On the other hand, expression of the markers for NK cells, CD8+ T cells, and regulatory T cells were not affected, and the neutrophil-specific marker Gr-1 was more highly expressed in ML120B-treated animals (Fig. 3C). The latter indicates increased neutrophil numbers in the colon after ML120B treatment, which is consistent with reports of blood neutrophilia caused by IKKβ inhibition (15, 20). These results, together with the cytokine analyses (Fig. 3 A and B), strongly argue that the ML120B-induced inhibition of IL-11 and IL-22 expression in the mucosa was not due to suppression of these cytokines in macrophages, but rather related to the failure to recruit these cells into the mucosa. Our data suggest a circular paracrine model in which IKKβ inhibition decreases epithelial chemokine induction in colitis, thereby interfering with recruitment of cells that express the key protective cytokines IL-11 and IL-22 (see Fig. 6). These cytokines are in turn required for activation of epithelial STAT3 and induction of normal levels of cytoprotective HSP70 in the epithelium.
To further confirm that the initial recruitment of inflammatory cells is responsible, at least in part, for initiation of the healing process after acute colitis, we took advantage of a model of inducible IKKβ deletion in the intestinal epithelium. Mice carrying a tamoxifen-inducible Cre transgene under control of the epithelium-specific Villin promoter (25) were crossed with floxed Ikkβ mice (26). Oral administration of tamoxifen for 5 days caused effective deletion of IKKβ in the small intestinal and colonic epithelium, but not in lamina propria cells or in any of the other organs tested (Fig. 4A).
Deletion of IKKβ after initial recruitment of inflammatory cells has no effect on healing of acute ulcerating colitis. (A) Selective ablation of IKKβ in IECs throughout the entire intestine (duodenum, jejunum, ileum, colon) of villin-Cre-ERT2/IkkβF/ F mice after induction of Cre-recombinase activity by tamoxifen (TAM). (B) Schematic overview of DSS and tamoxifen application. (C and D) Histological score (C) and size (D) of ulcerations 21 days after the beginning of DSS administration in villin-Cre-ERT2/IkkβF/ F mice, which had received TAM 5 days after the end of DSS administration.
Having established that IKKβ can be effectively deleted in a temporally controlled fashion, we induced acute colitis in IKKβ-proficient mice by DSS administration for 6 days, followed by 4 days on regular water (corresponding to the period of maximal inflammation, recruitment of immune cells, and the beginning of wound healing), after which we started tamoxifen treatment for 5 days to induce epithelial IKKβ deletion (Fig. 4B). In contrast to mice with constitutive IKKβ deletion (Fig. 1 B–E), the mice did not exhibit any difference in regard to mucosal inflammation or epithelial healing 21 days after the start of DSS administration (Fig. 4 C and D). These results suggest that once the critical inflammatory cells are recruited into the mucosa, they can produce cytoprotective cytokines, such as IL-11 and IL-22, independent of epithelial IKKβ and NF-κB. The data further indicate that epithelial IKKβ has two important physiologic functions in acute colitis, cell-autonomous early protection against apoptotic cell death (13) and delayed epithelial protection and healing via recruitment of inflammatory cells that release cytoprotective factors.
Although our studies revealed an important protective function for IKKβ in acute colitis, the observations remained apparently inconsistent with prior studies on the inflammation-promoting function of RelA/p65 (11). A key difference between our and prior studies was the duration of the colitis (acute vs. chronic) and the underlying pathogenic mechanisms, particularly the involvement of T cells. To bridge the gap between the models, we used an injury-independent model of T cell-driven chronic colitis associated with IL-10 deficiency. IL-10-deficient mice develop spontaneous mucosal inflammation and IEC hyperplasia in the colon over a period of months (27). Inflammation was associated with increased NF-κB activity in the entire colon and isolated IECs (Fig. 5 A and B), suggesting that canonical NF-κB activation may also play a role in controlling colitis in this model. To examine this point, we crossed Il10−/− mice with IkkβDIEC mice to yield mice double-deficient for IL-10 and IEC-IKKβ (Il10−/−/IkkβDIEC mice). As controls we used littermates that lacked IL-10 and had floxed Ikkβ but no Cre transgene (Il10−/−/IkkβF/F mice). Both groups were examined for the occurrence and severity of spontaneous colitis. Loss of epithelial IKKβ had no significant impact on the incidence of spontaneous colitis (Fig. 5C) or the severity of mucosal inflammation once animals had developed significant disease (Fig. 5 D–G). Moreover, the time interval during which animals lost 20% of their prior maximal body weight, a clinical measure of the acuteness of disease, was the same in all mice independent of their genotype (Fig. 5 H and I). Thus, despite marked activation of NF-κB in the epithelium in this model of chronic colitis, its function in IECs was dispensable under these conditions. By contrast, crossing of Il10−/− mice with IkkβDmye mice that lack IKKβ selectively in macrophages and neutrophils (13) resulted in significant attenuation of spontaneous colitis (Fig. 5J).
Ablation of myeloid but not epithelial IKKβ ameliorates colitis in Il10−/− mice. (A) Increased NF-κB binding activity in IECs and lamina propria cells (LP) in Il10−/− mice with clinical signs of disease determined by EMSA. Immunoblot analysis (WB) of β-actin was performed to control for using equal amounts of proteins. (B) NF-κB composition determined by a supershift assay in IECs and LP from mice with colitis. (C) Disease-free survival of IkkβF/ F and IkkβDIEC mice on an IL-10-deficient background was followed for 30 weeks after birth. (D and E) Representative H&E-stained sections of IkkβF/ F/Il10−/− and IkkβDIEC/Il10−/− double-mutant mice. (Magnification: 20×.) (F and G) Analysis of crypt depths (F) and numbers of infiltrating cells (G) at the end of the disease-free survival period in IkkβF/ F/Il10−/− and IkkβDIEC/Il10−/− double-mutant mice. (H and I) Time span between age at maximal body weight and the end of the disease-free survival period (H) and degree of weight loss during the last 2 weeks before the end of disease-free survival (I) of IkkβF/ F/Il10−/− and IkkβDIEC/Il10−/− double-mutant mice. (J) Disease-free survival in IkkβDmye mice and IkkβF/ F mice lacking IL-10 was followed for 20 weeks. Note that animals in C and J were kept in different animal facilities, which presumably accounts for the differences in disease-free intervals. (K) Real-time PCR analysis of the indicated mRNAs in IL-10-deficient bone marrow-derived macrophages treated with LPS (100 ng/ml) for the indicated times in the presence or absence of ML120B (30 μM). Data are mean ± SE. *, P < 0.05 by t test.
To determine whether key inflammatory factors known to be important for pathogenesis in this chronic colitis model are affected by deletion of IKKβ in myeloid cells, we isolated bone marrow-derived macrophages from Il10−/− mice. Macrophages were stimulated with LPS in the presence or absence of ML120B, and mRNA levels for several cytokines and other inflammatory mediators were determined. ML120B blocked induction of TNF-α, IL-1β, IL-6, ICAM-1, IL-12p40, and IL-23p19 mRNAs after LPS stimulation (Fig. 5K). Collectively, our data show that, in contrast to acute injury-related colitis, the dominant activity of IKKβ in chronic colitis is proinflammatory and mediated by myeloid cells, whereas its epithelial functions have no impact on the course or severity of inflammation.
This study demonstrates that NF-κB has overall protective functions in acute, ulcerating colitis, but promotes inflammation in chronic colitis that is not associated with rapid and extensive mucosal injury. These results help to reconcile apparently discrepant information from prior reports on the role of NF-κB in controlling intestinal inflammation (11, 12). Thus, mice selectively deficient for IKKβ in IECs exhibit more severe acute colitis (13) and epithelial loss of IKKγ caused spontaneous colitis (14), supporting the notion that IEC IKKβ and NF-κB have an important protective role that prevents injury-induced mucosal inflammation. Our data demonstrate that the protective functions of IKKβ and NF-κB dominate their overall activity in the intestinal mucosa under acute inflammatory conditions. Consistent with this interpretation, treatment with LPS, which strongly activates NF-κB in epithelial and other mucosal cells, attenuates acute chemical colitis and irradiation-induced damage (17, (28). Protection by NF-κB is in part mediated by its direct effects on epithelial cell survival (13). Our results further suggest that NF-κB also protects against damage in a paracrine manner by up-regulating the expression of chemokines that recruit myeloid and T cells (Fig. 6). These cells release cytoprotective factors such as IL-11 and IL-22, whose protective capacity against acute colitis has been demonstrated by direct administration or local gene delivery in WT animals (28, 29). Myeloid cells, in particular, can contribute to protection against acute colitis, because their depletion exacerbated acute DSS-induced colitis (30). One possible mechanism by which NF-κB-dependent myeloid cell recruitment can mediate epithelial protection and healing is through production of IL-11 and IL-22 and other factors that induce epithelial expression of HSP70, which has strong antiapoptotic functions in IECs (31, 32). Indeed transgenic expression of HSP70 in IECs ameliorated acute DSS-induced colitis (23). Hspa1a, which encodes HSP70, is a STAT3 target gene and does not contain NF-κB binding sites in its promoter region and is not known to be induced by IKK activators. However, IL-11 and IL-22 are potent activators of STAT3 and HSP70 expression, suggesting that NF-κB can mediate mucosal protection indirectly by recruiting myeloid and T cells, which produce these cytoprotective cytokines (Fig. 6).
Proposed model of IKKβ actions in acute ulcerating colitis. During DSS-induced colitis, IKKβ and NF-κB are activated in IECs, leading to increased transcription and secretion of chemotactic cytokines for myeloid and T cells. Recruitment of these cells is important for IKKβ-independent mucosal expression of cytokines, including IL-11 and IL-22, that are cytoprotective for epithelial cells through activation of STAT3 and downstream target genes such as HSP70.
In contrast to acute ulcerating colitis, epithelial IKKβ plays no apparent role in chronic colitis driven by immune dysregulation. In the IL-10-deficient model of chronic colitis, disease development is caused by an imbalance of inflammation-causing effector T cells and inflammation-suppressing regulatory T cells, with antigens of the normal microbiota as the driving force (27). Epithelial involvement is secondary to the influx of inflammatory cells and is characterized mostly by hyperproliferation rather than frank ulceration (33). Epithelial erosions are not a central feature of the pathogenesis and, if at all, are superficial and occur very late in the disease process. Under these conditions, IEC-NF-κB was activated but had no tangible impact on disease progression, possibly because of the reduced pathogenetic importance of protection against cell death. Alternatively, prolonged exposure of epithelial cells to inflammatory mediators may render their protection independent of IKKβ/NF-κB, perhaps because other protective pathways are overactivated. In contrast, IKKβ in macrophages and neutrophils promoted inflammation under the same circumstances, indicating that myeloid NF-κB has a key proinflammatory function in this model of chronic colitis. These data are consistent with, and can explain, the overall antiinflammatory effects achieved by administration of antisense oligonucleotides against NF-κB/RelA in a different model of T cell-driven chronic colitis (11).
NF-κB controls the expression of numerous genes encoding chemokines and proinflammatory cytokines and several antiapoptotic proteins (34). Our current and prior studies provide strong evidence that the relative physiologic importance of these two major groups of target genes, within the same tissue and disease process, varies depending on the specific cell types involved and the particular etiologic and pathogenetic mechanisms. Thus, NF-κB-dependent cell survival and chemokine expression is dominant in epithelial cells under acute injury conditions, whereas proinflammatory functions of NF-κB are decisive in chronic situations of immune dysregulation. It should be noted that under most pathophysiologic conditions, including those in IBD patients, these sets of functions are likely to coexist and may even compete with each other. Therefore, the overall effects of IKKβ/NF-κB inhibition in different organ systems and disease processes are not easily predictable under all circumstances, but must be carefully established for successful therapeutic applications.
Our results have implications for the potential therapeutic use of IKKβ inhibitors and add to recently raised concerns about prolonged inhibition of this pathway (15). Although pharmacological inhibition of IKKβ had been suggested to improve acute colitis (12), our results do not support this notion. One possible explanation for these differences is the selectivity of the particular inhibitors used, and the possibility that other inflammatory pathways might be affected nonspecifically. The inhibitor used in our studies, ML120B, displays very high specificity in vivo and its overall effects closely mimic the genetic deletion of IKKβ in enterocytes during DSS-induced colitis as well as those observed after IKKβ ablation in myeloid cells in models of sepsis and endotoxic shock (15). Based on our genetic and pharmacological results, inhibition of IKKβ and NF-κB is likely to exacerbate tissue damage during the acute phase of intestinal inflammation dominated by apoptotic loss of epithelium and subsequent ulceration, and would therefore appear to be contraindicated. In contrast, inhibitors of IKKβ and NF-κB may be beneficial in the chronic phase of intestinal inflammation, when the risk of epithelial cell apoptosis and epithelial ulcerations is reduced or entirely absent, especially if such inhibitors are targeted to myeloid cells.
To generate IkkβDIEC mice, IkkβF/F mice (26) were crossed to Villin-Cre mice (35) and kept on a C57BL/6;129 background. IkkβDmye mice have been described (13). For enterocyte-specific and temporal ablation of Ikkβ, villin-Cre-ERT2 mice (25) were crossed to IkkβF/F. Deletion was induced by five daily oral administrations of 1 mg of tamoxifen in an ethanol/sunflower oil mixture. Il10−/− mice on a C57BL/6 background were obtained from The Jackson Laboratory and intercrossed with IkkβDIEC and IkkβDmye mice. Occurrence of significant spontaneous disease, and thereby the end of disease-free survival, in IL-10-deficient mice was defined as >20% loss of body weight relative to maximal prior weight. Diseased mice were euthanized and examined histologically. All animal procedures were reviewed and approved by the Regierung von Oberbayern and the University of California at San Diego Institutional Animal Care and Use Committee.
To induce acute colitis, mice were given 3% DSS (MP Biomedicals) in their drinking water for 5–6 days, followed by regular drinking water. ML120B, kindly provided by Millenium Pharmaceuticals, was given by oral gavage twice daily in methylcellulose at a concentration of 80 mg/kg. Mice were euthanized on the indicated days, and the colon was removed, fixed in paraformaldehyde, and embedded in paraffin. Severity of colitis was assessed histologically as described (13, 36).
Isolation and culture of bone marrow-derived macrophages, RNA extraction, cDNA synthesis, and real-time PCR were performed as described (15). Primer sequences are available on request. Gene expression profiling of isolated enterocytes was performed with Affymetrix MOE430A 2.0 GeneChips as reported (37). Isolation of IECs, immunoblots, and EMSAs have been described (13). The following antibodies were used: anti-IKKα (Imgenex), anti-IKKβ (UBI), anti-RelA/p65, anti-p50, anti-c-Rel (Santa Cruz), and anti-β-actin (Sigma).
We thank Birgit Wittig, Kristin Retzlaff, and Lucia Hall for excellent technical assistance and expert help with the animal studies. This work was supported by National Institutes of Health Grants DK70867, DK35108, and RR17030, the University of California at San Diego Digestive Diseases Research Development Center (Grant DK80506), and a Jeannik M. Littlefield–American Association for Cancer Research grant (to M.K.). Further support was provided by Wilhelm Sander-Stiftung Grant 2005.146.1 (to M.C.A.) and the Deutsche Forschungsgemeinschaft (Emmy-Noether-Program Grant Gr1916/2–2, SFB 456), Deutsche Krebshilfe Grant 106772, and Fritz-Thyssen-Stiftung Grant 10.05.2.168 (to F.R.G.). M.K. is an American Cancer Society Research Professor.
Author contributions: L.E. and F.R.G. designed research; T.N., A.A.F., S.M.D., and M.C.A. performed research; S.R., M.F.K., and R.M.S. contributed new reagents/analytic tools; L.E., J.M., R.L., M.C.A., and F.R.G. analyzed data; and L.E., M.K., M.C.A., and F.R.G. wrote the paper.
The authors declare no conflict of interest.
