BACKGROUND/AIMS/PATIENTS Glucocorticoid treatment is known to reduce nuclear factor-kappa B (NF-κB)p65 binding activity and activation in lamina propria cells of patients with Crohn’s disease. However, lamina propria cells of glucocorticoid treated patients did not show increased expression of IκBα, and the hypothesised upregulation of IκBα by glucocorticoid treatment has not yet been shown in vivo. To investigate whether cells other than lamina propria localised mononuclear cells contribute to increased IκBα, resection gut specimens from patients matched for Crohn’s disease activity index (CDAI) with or without glucocorticoid treatment were studied, and changes in the NF-κB/IκBα system were determined in the lamina propria as well as in underlying submucosal and endothelial cells.
METHODS Changes in the NF-κB/IκBα system were determined by immunohistochemistry, electrophoretic mobility shift assay, and western blot analysis in resected gut specimens from patients matched for CDAI and van Hees index with or without long term glucocorticoid treatment.
RESULTS Resection gut specimens from patients with Crohn’s disease under glucocorticoid treatment had significantly lower nuclear NF-κBp65 levels in mononuclear, epithelial, and endothelial cells than samples from CDAI and van Hees index matched patients not having glucocorticoid treatment. Nuclear NF-κBp65 showed a strong positive correlation with both the CDAI (r = 1 for both groups) and the van Hees index (r = 0.605 for untreated and r = 0.866 for glucocorticoid treated specimens). Lower nuclear translocation of NF-κBp65 in the glucocorticoid treated group was paralleled by higher IκBα levels in vascular endothelial cells, but not in infiltrating mononuclear cells.
CONCLUSION A comparison of resection gut specimens from untreated and treated CDAI matched patients with Crohn’s disease showed downregulation of NF-κB binding activity and NF-κBp65 expression and cell specific induction of endothelial IκBκ expression in the glucocorticoid treated group. As the two groups showed similar disease activity (CDAI, van Hees index), the activation of the NF-κBp65/IκBα system must be only part of the inflammatory cascade leading to the clinical appearance of Crohn’s disease.
- Crohn’s disease
- transcription factor
Statistics from Altmetric.com
Increased oxidative stress and activation of redox sensitive cellular signalling systems are well recognised key steps in the development of chronic inflammatory disease.1 Several molecular mechanisms have been described that convert the transient appearance of oxygen free radicals into sustained cellular changes and altered gene expression. One ubiquitous mechanism is represented by the redox sensitive transcription factor NF-κB, which has been shown to be activated in response to a number of immune and inflammatory stimuli.2-6 Recently, Crohn’s disease has been shown to be a model system for understanding the biological implications of NF-κB activation.7-13
The NF-κB family of transcription factors consists of the members p105/p50, p100/p52, p65 (RelA), relB, c-Rel, the proto-oncogene v-rel, and the Drosophilamorphogen.2-6 Characteristically, NF-κB proteins are sequestered in the cytoplasm as the result of retention by a class of inhibitory proteins, referred to as the IκB family.4-6 14 15 On stimulation, IκBα is phosphorylated at two serine residues in its N-terminal portion by a large IκB kinase (IKK) complex in which IKKα and IKKβ are associated with the newly identified scaffold proteins NEMO/IKKγ and IKAP and thus physically linked to upstream activators.16 17 Phosphorylation targets IκBα for ubiquitination and subsequent rapid proteolytic degradation,14-17 thereby releasing free NF-κB (preferentially the NF-κB heterodimer p50/p65) which translocates into the nucleus, where it binds to decameric DNA sequences and activates transcription of NF-κB regulated target genes.2-6 As IκBα expression itself is controlled by NF-κBp65, NF-κBp65 activation is shut down by an IκBα dependent autoregulatory feedback loop.2-5 14 15 It has been shown in vitro that newly synthesised IκBα not only retains NF-κBp65 in the cytoplasm but translocates to the nucleus, where it can sequester free NF-κBp65 and promote translocation of NF-κB/IκBα complexes from the nucleus back to the cytoplasm.15 However, recent studies show that this autoregulatory mechanism can be compensated by induction of perpetuated NF-κBp65 mRNA synthesis, which provides a continuously growing pool of free NF-κBp65 which can translocate into the nucleus and thereby overrides the IκBα dependent self regulation.18
Glucocorticoids are steroid hormones with potent immunosuppressive and anti-inflammatory action, frequently used in the treatment of a broad range of chronic diseases such as allergic reactions, asthma, rheumatoid arthritis, systemic lupus erythematosus, transplantation dependent tissue rejection, and inflammatory bowel disease.19-24 The molecular mechanisms underlying the therapeutic effects of glucocorticoids on chronic inflammation are poorly understood. However, there is growing evidence that they exert part of their action by downregulating gene expression by interfering with the activity of transcription factors such as STAT-1,25 NFAT-1,26 AP-1,27 and NF-κB.2 6 28 The highly immunosuppressive and anti-inflammatory action of glucocorticoids is at least in part related to a glucocorticoid dependent repression of NF-κB mediated gene expression.2 6 28
Glucocorticoids mediate their effects by binding to intracellular glucocorticoid receptors (GRs),29 which function as ligand activated nuclear transcriptional regulators.29-31Binding results in conformational changes in the ligand activated GR, which translocates into the nucleus and directly interferes with gene transcription.31 Two putative mechanisms have been proposed to explain GR mediated inhibition of NF-κB: (a) in vitro studies indicate that ligand bound GR can physically interact with the NF-κB subunit p65, thereby preventing transactivation of NF-κB31-35; (b) GR can upregulate IκBα gene transcription and expression possibly by yet undiscovered positive GR responsive elements in the IκBα promoter.36 37 The latter mechanism has been proposed to play a role in attenuating NF-κBp65 activity under chronic conditions.6 Only a few data are available that describe the relative contribution of each mechanism under physiological conditions.6 10
It has recently been reported that IκBα was not upregulated in lamina propria biopsy specimens of patients with Crohn’s disease receiving steroid treatment for seven days, although there was a significant decline in nuclear NF-κBp65.10 The authors suggest that clinical relapse of disease from remission may override an upregulation of IκBα10 and thereby contrasts with the in vitro situation.36 37 Another likely explanation may lie in the fact that the above study10 examined lamina propria specimens only, leaving the question open of what happens to the underlying submucosal and endothelial cells. Furthermore, because of the study design, the patients presumably had a different disease activity after glucocorticoid treatment, making it difficult to define whether changes in the NF-κB/IκBα system are the cause or consequence of the glucocorticoid effect. The study presented here therefore evaluates IκBα expression and NF-κB activation in resection gut specimens from patients with Crohn’s disease with and without glucocorticoid treatment matched with respect to Crohn’s disease activity index (CDAI) and van Hees index. As the two groups show the same disease activity and the same extent of inflammation, the study design allows us to define the effect of glucocorticoid treatment on NF-κBp65 activation independent of other factors that may modulate NF-κBp65 activity.
[γ-32P]ATP (3000 Ci/mmol at 10 Ci/ml), ECL-nitrocellulose membranes, ECL detection reagents, and Hyperfilmx ray films were obtained from Amersham, Braunschweig, Germany. Monoclonal anti-p65 antibodies, specific for active NF-κB, were obtained from Boehringer Mannheim, Mannheim, Germany. Anti-p65 (#sc-109), anti-c-Rel (N, sc-70), and anti-IκBα (C-15, #sc-203) polyclonal antibodies were obtained from Santa Cruz Inc, Heidelberg, Germany. The anti-p105/p50 sera no 1157, 1140, 1613, and 1614 were generously given by Dr N Rice (NCI-Frederick Cancer Research and Development Centre, Frederick, Maryland, USA). Anti-CD3, anti-CD20, and anti-CD68 antibodies were from Dako, Hamburg, Germany. Anti-CD38 antibodies, Cy3 conjugated streptavidin, and Cy2 conjugated donkey anti-sheep antibodies were obtained from Dianova, Hamburg, Germany. Normal human immunglobulins (gamma-venin) were from Behring, Marburg, Germany. Bovine serum albumin (BSA) and Naphthol AS-biphosphate were obtained from Sigma, Deisenhofen, Germany. New Fuchsin was purchased from Merck, Darmstadt, Germany. Poly(dI/dC) was from Pharmacia, Freiburg, Germany. The NF-κB consensus oligonucleotides and the rabbit reticulocyte lysate were purchased from Promega, Heidelberg, Germany.
Table 1 gives a description of the patients, including sex, age, duration of disease, haemoglobin (g/dl), leucocyte concentration (109/l), CDAI, and van Hees index. Table 2 lists their treatments. The patients presented were under medical treatment in the Department of Visceral Surgery in Bruchsal, Germany (H T), which is an academic teaching hospital associated with the University of Heidelberg. The study was approved by the ethics committee of the University of Heidelberg, according to the guidelines of the Helsinki Declaration.
PREPARATION OF TISSUES AND IMMUNOHISTOCHEMISTRY
Resection gut specimens of patients with Crohn’s disease were taken during operation and snap frozen in liquid nitrogen. Depending on the conditions during surgery, time between sampling and freezing in liquid nitrogen varied slightly but never exceeded 20 minutes. Under these conditions, however, it could not be excluded that some intestinal tissue underwent ischaemia before it was frozen. Cryostat sections (5 μm) were mounted and fixed in cold acetone or 2% paraformaldehyde in phosphate buffered saline (pH 7.4) for 10 minutes at room temperature, followed by a wash in Tris buffered saline (TBS, pH 7.4). Primary antibodies (mouse anti-human NF-κBp65, IgG3 isotype) and rabbit anti-human IκBα were diluted 1:10 and 1:20 in TBS, pH 7.4, containing 0.2% BSA and 2.5 mg/ml normal human immunoglobulins (gamma-venin) and applied to the sections for 90 minutes at room temperature. After being washed in TBS/0.2% BSA, sections were incubated with 5% inactivated normal sheep (NF-κBp65) or goat (IκBα) serum to reduce non-specific background. Acetone treated slides were post-fixed with 2% paraformaldehyde before this step. After removal of the blocking solution, a biotinylated secondary antibody, adsorbed against human immunoglobulin (biotin-sheep anti-mouse (1:100); Amersham) or biotin-goat anti-rabbit (1:20; Dako) respectively was added for 30 minutes, followed by a 30 minute incubation with streptavidin complexed with alkaline phosphatase (Dako). After subsequent washes with TBS and a final wash in distilled water, the colour reaction was developed using Naphthol AS-biphosphate and New Fuchsin as chromogen.38 After counterstaining with haematoxylin, sections were mounted in glycerol/gelatin.
The immunohistochemical results were evaluated according to a score depicted in table 3. Coded slides were analysed by one investigator (K T) blinded to the clinical diagnosis and medication. Expression of NF-κBp65 and IκBα was determined on consecutive serial sections in each case. In each slide, the number of positive mononuclear cells was determined in at least five visual fields (magnification × 160) in five different representative areas of the lamina propria as well as in the submucosa. For determination of the endothelial expression, 40 vessels in the lamina propria and the submucosa were analysed in each case. Median values of the respective results were obtained for statistical evaluation.
IN SITU DOUBLE IMMUNOFLUORESCENCE AND LASER SCAN MICROSCOPY
To determine which subsets of mononuclear cells express NF-κBp65 in Crohn’s disease, in situ double immunofluorescence for NF-κBp65 in combination with other cell markers was performed. Sections were incubated for one hour with a mixture of monoclonal mouse anti-NF-κBp65 (IgG3 isotype) and monoclonal mouse anti-CD3 (clone OKT-3, IgG2a isotype; 10 μg/ml), mouse anti-CD20 (clone L-26, IgG2a isotype; dilution 1:50), mouse anti-CD68 (clone EBM11, IgG1 isotye dilution; 1:200), or mouse anti-CD38 (clone CBL 167, IgG2b isotype; dilution 1:50). Incubations with the primary reagents omitted and with the use of a mixture of non-binding monoclonal mouse antibodies of different isotypes served as negative controls. The antibodies were diluted in TBS/0.2%BSA/2.5 mg/ml gamma-venin. TBS/0.2% BSA was used for all washing steps and to dilute the secondary and tertiary reagents. As secondary antibodies, a combination of biotinylated goat anti-mouse IgG3 (dilution 1:50; Southern Biotech, distributed by Bibby Dunn, Asbach, Germany) and sheep anti-mouse IgG1, IgG2a, or IgG2b (dilutions 1:50, 1:100 and 1:50, respectively; Serotec Ltd, Kidlington, UK) plus gamma-venin was used, followed by simultaneous incubation with Cy3 conjugated streptavidin (dilution 1:1000; red fluorescence) and Cy2 conjugated donkey anti-sheep antibodies (dilution 1:100; green fluorescence) for 30 minutes respectively. After the final washes, the sections were mounted with histogel mounting medium (Camon, Wiesbaden, Germany). Slides were viewed with a Laserscan microscope (Ernst Leitz GmbH, Wetzlar, Germany) using 570 nm (red emission) and 508 nm (green emission) filters.
ELECTROPHORETIC MOBILITY SHIFT ASSAYS (EMSA)
EMSA was performed as described previously in detail.39 40 In brief, frozen tissue (0.1 cm × 0.2 cm) was broken down mechanically, transferred to a 50 ml Falcon tube containing 5 ml cold buffer A (10 mM Hepes/KOH, pH 7.9 at 4°C, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM dithiothreitol, 0.2 mM phenylmethanesulphonyl fluoride, 0.6% Nonidet P40) and homogenised with an Ultraturrax homogeniser (Wheaton, Milleville, New Jersey, USA) for one minute. Insoluble material was removed by centrifugation (30 seconds at 400 g (2000 rpm), 4°C) and the supernatant was incubated on ice for 10 minutes before being centrifuged for five minutes at 6000 g (8000 rpm), 4°C. The supernatant was discarded and the nuclear pellet was resuspended in 100 μl buffer B (25% glycerol, 20 mM Hepes/KOH, pH 7.9 at 4°C, 420 mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM dithiothreitol, 0.2 mM phenylmethanesulphonyl fluoride, 2 mM benzamidine, 5 mg/ml leupeptin) and incubated on ice for 20 minutes. Cellular debris was removed by two minutes centrifugation at 18000g (14000 rpm)at 4°C and the supernatant was quick frozen at −80°C. Protein concentrations were determined by the Bradford assay.40a A 10 μg portion of nuclear extract was included in the binding reaction.
Binding to an NF-κB consensus oligonucleotide (5’-AGTTGAGGGGACTTTCCAGG C-3’) was performed in 10 mM Hepes, pH 7.5, containing 0.5 mM EDTA, 100 mM KCl, 2 mM dithiothreitol, 2% glycerol, 4% Ficoll 400, 0.25% Nonidet P40, 1 mg/ml BSA (DNAse free), and 0.1 μg/μl poly(dI/dC) in a total of 20 μl as described.40 Specificity of binding was ascertained by competition with a 160-fold molar excess of unlabelled consensus oligonucleotides and by characterisation with specific polyclonal antibodies (Santa Cruz).
PREPARATION OF RECOMBINANT NF-κBp65-CONTAINING LYSATE
Recombinant NF-κBp65 (rNF-κBp65) was prepared as previously described in detail.41 42 In brief, rNF-κBp65 was synthesised using the T7 promoter of the plasmid Rc-CMV-p65, kindly provided by Dr P A Baeuerle (Tularik, San Francisco, California, USA). NF-κBp65 RNA (3 μg) was added to 70 μl of rabbit reticulocyte lysate to translate proteins in vitro. Efficient translation was monitored in a parallel reaction using [35S]methionine as substrate. A 0.15–20 μl volume of lysate, containing about 0.15–10 μg of in vitro translated protein,42 was used in the EMSA. Unprogrammed lysate without NF-κBp65 served as control.
SEMIQUANTITATIVE DETERMINATION OF NF-κBp65 BINDING ACTIVITY IN EMSA
To quantify the NF-κB signal obtained in each resection gut specimen, rNF-κBp65 was included in each EMSA as previously described in detail.41 Binding reactions with 5, 7.5, and 10 μg rNF-κBp65-containing lysate were included in each gel. The resulting NF-κB signals were quantified by densitometry (Bio-Rad, Munich, Germany) and used to establish a standard curve for each gel. The densitometric value for the NF-κB signal of the resection gut specimens was converted into rNF-κBp65 equivalents by using the internal standard curve for rNF-κBp65.41
WESTERN BLOT ANALYSIS
Cytoplasmic and nuclear fractions were prepared as previously described in detail.43 Cytoplasmic extract (20 μg) or nuclear extract (10 μg) was subjected to sodium dodecyl sulphate/polyacrylamide gel electrophoresis (SDS/PAGE; 12.5% gel), followed by electroblotting on ECL-nitrocellulose membranes. Equal loading was confirmed by Ponceau Red staining. Filters were probed with the respective antisera for p65 (1:100), c-Rel (1:100), p50/p105 (1:2000), and IκBα (1:100) as described previously.40 43 Membranes were washed (2 × 7 min; TBS/0.05% Tween) before the secondary antibody (horseradish peroxidase coupled rabbit IgG, diluted 1:2000 in Blotto A) was added and incubation was continued for 30 minutes at room temperature. Membranes were washed (3 × 15 minutes) as above followed by a last five minute wash in TBS. Immunoreactive proteins were detected with the ECL-western blot system and subsequent autoradiography for two minutes (p65, p50), four minutes (IκBα), or five minutes (c-Rel).
DENSITOMETRIC QUANTIFICATION OF WESTERN BLOT AUTORADIOGRAMS
Signals obtained in western blot analysis were quantified using a GS-700 imaging densitometer (Bio-Rad). Determination of the signal area to be measured and the quantitative evaluation were performed twice for two independent experiments. The mean of the two measurements was taken for statistical analysis.
Values are given as median for evaluation of immunohistochemistry and as mean (SD) for densitometric evaluation of EMSA results and western blot signals. Student’s two tailedt test was used to determine significance. p<0.05 was considered to be statistically significant.
NF-κBp65 BINDING ACTIVITY AND ACTIVATION IN RESECTION GUT SPECIMENS
To determine the influence of glucocorticoid treatment on NF-κB activation and IκBα expression in inflammatory bowel disease, we studied resection gut specimens, derived from 15 patients with Crohn’s disease. Ten patients had never been treated with glucocorticoids, while five had taken glucocorticoids for at least four weeks (tables 1and 2). Five patients (three without and two with glucocorticoid treatment) were taking sulfasalazine derivatives (table 2). Patients with and without treatment were matched for CDAI (196.8v 187, p = 0.69) and van Hees index (265.1v 289.4, p = 0.707).
Consistent with previous studies,10 11 13immunohistochemical staining with an NF-κBp65 antibody that exclusively recognises activated NF-κBp6544 showed positive red staining of numerous infiltrating mononuclear cells in sections obtained from patients without glucocorticoid treatment (fig1A), and staining for activated NF-κBp65 was significantly lower in mononuclear cells in sections from glucocorticoid treated patients (fig1B). When staining was repeated in the absence of haematoxylin, it became evident—that is, in sections from patients without glucocorticoid treatment, almost all NF-κBp65 positive cells displayed nuclear staining, and additional cytoplasmic staining was also noted (fig 1C). Activated NF-κB was present in mononuclear cells of the lamina propria and the submucosa as well as in endothelial and epithelial cells (fig 1C). In contrast, in sections from glucocorticoid treated patients, only a few mononuclear cells in the laminar propria showed positive nuclear staining (fig 1D, arrows). Some additional positive cells displayed weak cytoplasmic expression of NF-κBp65 (fig1D, arrowheads). The lower level of activated NF-κBp65 observed in mononuclear cells of glucocorticoid treated resection gut specimens was confirmed by statistical evaluation (fig 1E).
To identify the mononuclear cells that express activated NF-κBp65 in Crohn’s disease, double immunofluorescence staining for activated NF-κBp65 and CD3, CD20, CD38, and CD68 positive cells was performed on sections from untreated patients (fig 2). Activated NF-κBp65 was detected in CD3 positive T lymphocytes (first row), in CD20 positive B lymphocytes (second row), and CD68 positive mononuclear phagocytes/macrophages (third row) of the lamina propria, indicating that, in the inflamed mucosa, NF-κBp65 is activated in all major subsets of mononuclear cells. In addition, positive staining for activated NF-κBp65 was also observed in CD38 positive cells, presumably representing plasma cells.
Vascular endothelial cells of resection gut specimens without treatment also stained for activated NF-κBp65 (fig 3A), confirming that Crohn’s disease also activates cells other than mononuclear cells to participate in the inflammatory response. Glucocorticoid treatment was paralleled by a decrease in activation of NF-κBp65 in the vessel wall (fig 3B,C).
This prompted us to investigate the NF-κB activation pattern in total resection gut specimens in order to define whether (a) it differed from the NF-κB activation previously described for mononuclear cells in lamina propria biopsy specimens10 13 and (b) whether there were differences in nuclear NF-κBp65 in patients with and without glucocorticoid treatment. When EMSA was performed with nuclear extracts of resection gut specimens, three different concentrations (5, 7.5, and 10 μg) of rNF-κBp65 programmed rabbit reticulocyte lysate were tested in parallel binding reactions and included on each gel as internal standard.41 This allowed us to normalise for intra-assay variation and to perform a semiquantitative determination of the NF-κB binding activity expressing NF-κB binding activity as recombinant NF-κB equivalents (rNF-κB equivalents41). The results for four representative patients without (patient 1 and 2) and with (patient 3 and 4) glucocorticoid treatment are shown in fig4A. Despite being normalised with respect to CDAI and van Hees index, NF-κB binding activity was higher in untreated patients (122.9 rNF-κB equivalents) than in patients taking glucocorticoids (68.7 rNF-κB equivalents; p = 0.051) (fig 4B). Characterisation of proteins contributing to the signal observed in EMSA by supershift experiments showed that in untreated samples the NF-κBp50/NF-κBp65 heterodimer preferentially bound to the NF-κB binding motif, as evidenced by supershifts and reduction in NF-κB binding activity in the presence of anti-p50 and anti-p65 antibodies (patient 2; fig 4C, lanes 2 and 3). Compared with samples of untreated patients, the contribution of NF-κBp65 and NF-κBp50 to the observed shift did not significantly differ in samples of glucocorticoid treated patients (fig 4D, lanes 2 and 3), although NF-κB binding activity was lower in all samples tested (fig 4A,B). However, there was a tendency towards a slight increase in NF-κBp50 in samples from glucocorticoid treated patients as reflected by an increase in supershifted material in the presence of anti-p50 antibodies (patient 3; fig 4D, lane 2). In contrast with recently published results for epithelial and mononuclear cells isolated from inflamed mucosa,13 incubation with c-Rel antibodies did not result in supershifting or significant weakening of the NF-κB complexes (fig 4C,D, lane 5). Only in one sample from a patient receiving glucocorticoid treatment did c-Rel antibodies result in a pronounced reduction in NF-κB binding activity (data not shown). Although we cannot exclude a serological bias of detection, these findings imply that NF-κB-cRel is not a major component of the NF-κB complex detected in resection gut specimens from untreated and glucocorticoid treated patients with Crohn’s disease.
CORRELATION OF DOWNREGULATION OF NF-κBp65 WITH DISEASE ACTIVITY
To extend further the observation of a glucocorticoid dependent switch in the extent of NF-κB activation in CDAI and van Hees index matched Crohn’s disease resection gut specimens, western blot analysis was performed. All probes from untreated patients showed a prominent translocation of NF-κB from the cytoplasm to the nucleus. Although there was a signal for NF-κBp65 in the cytoplasm, stronger NF-κB banding was detected in nuclear extracts of resection gut specimens as shown for two representative patients (patients 2 and 14; fig 5A, table4). In contrast, a strong cytoplasmic NF-κBp65 signal was present in specimens from patients receiving glucocorticoid treatment, while nuclear NF-κBp65 staining was hardly detected. Two representative patients are shown (3 and 15; fig 5B, table 4). When cytoplasmic and nuclear signals of all patients were evaluated by densitometry, the signals obtained for nuclear NF-κBp65 confirmed a decrease in nuclear NF-κB from 10.1 arbitrary NF-κB units in samples from untreated patients to 3.9 arbitrary NF-κB units in samples from glucocorticoid treated patients (fig 5C; table 4; p = 0.0407). To adjust for individual differences in NF-κB expression, we calculated the ratio of nuclear to cytoplasmic NF-κBp65 (table 4), thereby establishing an NF-κB activity coefficient which allowed definition of the actual extent of NF-κB translocation. This coefficient was 1.37 in untreated and 0.497 in glucocorticoid treated patients (fig 5D, table 4; p = 0.0037) and thus confirmed that NF-κBp65 nuclear translocation was lower in patients having glucocorticoid treatment.
The amount of nuclear NF-κBp65 in both groups correlated exactly with the given CDAI (r = 1 for both; fig 6A) and sufficiently with the van Hees inflammation index (r = 0.605 for untreated and 0.866 for treated patients; fig 6B), suggesting a relation between nuclear NF-κBp65 activation and the severity of the disease.
EMSA as well as a previous study13 showed a contribution of other members of the NF-κB family to the observed NF-κB binding activity. The above extracts were therefore tested for NF-κBp50 and NF-κB-cRel. Samples from all patients without glucocorticoid treatment showed only moderate translocation of NF-κBp50 from the cytoplasm to the nucleus (data not shown). In contrast, samples from patients under treatment showed greater nuclear translocation of NF-κBp50 (data not shown), which is consistent with the increased contribution of NF-κBp50 to NF-κB binding activity observed by EMSA (fig 4D). However, nuclear NF-κBp50 showed only a weak negative correlation with the CDAI (data not shown) and therefore may be influenced by additional individual factors not necessarily determined by the outcome of disease or treatment.
Although the above supershifting experiments failed to detect NF-κB-cRel in the NF-κB binding complex (fig 4C,D), western blot analysis showed a slight increase in NF-κB-cRel nuclear translocation in all patients with Crohn’s disease when compared with disease specific controls (data not shown). As the data are in line with a previous report that describes the involvement of NF-κB-cRel in NF-κB complexes of lamina propria mononuclear cells from inflamed mucosa,13 the failure to detect NF-κB-cRel by EMSA may be related to the fact that the antibody used is unable to react with traces of NF-κB-cRel in an NF-κB binding complex that is mainly formed by NF-κB(p50/p65). In western blots, however, no difference was observed in the extent of nuclear translocation of NF-κB-cRel in untreated and glucocorticoid treated samples (data not shown). Consistently, there was no correlation between nuclear NF-κB-cRel and the CDAI (data not shown).
GLUCOCORTICOIDS UPREGULATE IκBα IN VASCULAR ENDOTHELIAL CELLS BUT NOT IN MONONUCLEAR CELLS
Western blot analysis of cytoplasmic extracts of the above resection gut specimens showed that IκBα expression was weak in samples from patients without treatment (fig 7A), while it was stronger in all patients having glucocorticoid treatment (fig 7A). Densitometric evaluation of the signal showed an increase from 7.13 relative IκBα units in untreated samples to 13.07 relative IκBα units in glucocorticoid treated samples (p = 0.022) (fig 7B).
These findings are in contrast with previous reports, in which steroid treatment for seven days did not change IκBα expression levels in lamina propria biopsy specimens.10 Therefore the cells contributing to IκBα expression were determined by immunohistochemistry in consecutive sections of the above tested resection gut specimens. Consistent with the previous observation,10 there was no significant difference in the extent of IκBα expression in infiltrating mononuclear cells of patients with Crohn’s disease without (fig 8A) or with (fig 8B) glucocorticoid treatment. Figure 8C summarises the quantitative results and the statistical analysis.
Vascular endothelial cells, however, which stained only weakly positive for IκBα in sections from patients without treatment (fig 9A), showed stronger immunohistochemical expression of IκBα in the glucocorticoid treated group (fig 9B). Statistical analysis confirmed a higher IκBα expression in vascular endothelial cells in the glucocorticoid treated group (fig 9C), which may, at least in part, explain the increase in IκBα observed in western blots of total gut extracts (fig 7A).
Chronic NF-κB activation in Crohn’s disease seems to depend on NF-κBp65, as (a) local application of NF-κBp65 antisense oligonucleotides abrogated clinical and histological signs of mucosal inflammation in mice with experimental colitis,7 (b) NF-κBp65 antisense oligonucleotides reduced production of proinflammatory cytokines in cultured lamina propria macrophages of patients with Crohn’s disease,7 8 and (c) increased expression of NF-κBp65 has been shown in lamina propria biopsy specimens10 13 and full thickness specimens of bowel11 of patients with Crohn’s disease. Glucocorticoids, frequently used in the treatment of Crohn’s disease, have been shown to be effective inhibitors of NF-κBp65 and therefore seem to directly inhibit amplification and perpetuation of the inflammatory response.2 6 10 28 31-37 Several models have been proposed to explain the molecular mechanisms that underlie the inhibitory action of glucocorticoids on NF-κB activation. Recent in vitro studies have reported direct protein-protein interactions between NF-κBp65 and GR, which result in mutual transrepression.31-35 Other investigators suggest that expression of the NF-κB specific inhibitor IκBα is upregulated via a GR responsive element.36 37 At present, however, the relative contribution of each mechanism has not been defined under physiological conditions in human disease.
In order to determine the effect of glucocorticoids on the NF-κBp65/IκBα system, we studied whole gut segments (rather than biopsy specimens) of disease activity matched patients with and without glucocorticoid treatment. As presented here, the mode of glucocorticoid action seems to depend on the underlying cell type. Downregulation of NF-κBp65 was evident in both infiltrating mononuclear cells of the lamina propria and submucosa as well as in vascular endothelial cells. The fate of IκBα, however, was different in different cell types: mononuclear cells located in the lamina propria and submucosa did not show higher levels of IκBα in the glucocorticoid treated group (fig8), which is in line with previous observations in sigmoid biopsy samples of patients with Crohn’s disease.10 In contrast, vascular endothelial cells showed significantly higher IκBα expression in resection gut specimens obtained from patients under glucocorticoid treatment (fig 9). These in vivo results contrast with the in vitro situation: mononuclear cells have been shown to respond to glucocorticoids in vitro with increased IκBα expression while studies in cultured endothelial cells failed to detect glucocorticoid dependent upregulation of IκBα.46 This discrepancy points to an important, but mostly neglected, drawback of in vitro models of human disease: in vitro experiments on cell activation at the transcriptional level are performed over minutes to several hours, while human disease lasts much longer. In vitro experiments to identify the mechanism of glucocorticoid dependent NF-κB downregulation have been performed only for hours and only with single cell types.31-37 46 They may therefore not be sufficient to completely explain the complex physiological situation involving much longer time frames and cell-cell interactions.
It has been proposed that the glucocorticoid mediated increase in IκBα synthesis is long lived34 and thus plays a crucial role in attenuating NF-κBp65 activity under chronic conditions.6 This view is emphasised by the study presented here, as the prominent endothelial IκBα upregulation shown was detected in samples of long term glucocorticoid treated patients, while other studies have so far only looked at the transient effects of glucocorticoid treatment on the NF-κB/IκBκ system in vivo.10 Recent in vitro studies that dissociated glucocorticoid dependent transrepression of NF-κB and glucocorticoid mediated transactivation of IκBα synthesis indicate that increased IκBα synthesis is neither required nor sufficient for glucocorticoid mediated downmodulation of NF-κB activity.34 The authors postulate that the transrepressing function of glucocorticoids is the decisive factor in anti-inflammatory action while the transactivating capacity may account for unwarranted side effects.34 Also the latter cannot be excluded, the study presented here demonstrating that upregulation of IκBα may significantly contribute to the attenuation of NF-κB activation in Crohn’s disease. As depicted in table 2, some of the patients included in this study received the drug sulfasalazine which has been shown to inhibit NF-κB activation by preventing IκBα phosphorylation and subsequent degradation.45 However, the NF-κB activity score in patients that did not receive glucocorticoids but had sulfasalazine treatment was significantly higher (table 4) than the NF-κB activity score in patients under glucocorticoid treatment. This implies that additional mechanisms such as increased IκBα synthesis are present in glucocorticoid treated patients and thus allow further reduction in NF-κB activation. It has recently been reported that induction of NF-κBp65 mRNA perpetuates NF-κB activation under chronic inflammatory conditions.18 Although an excess of newly synthesised NF-κBp65 would activate IκBα synthesis, the cellular defence would be much stronger if large amounts of IκBα were already available as the result of glucocorticoid induced IκBα synthesis. As the untreated and glucocorticoid treated groups investigated here were matched with respect to disease activity (CDAI and van Hees index), we conclude that activation of the NF-κBp65/IκBα system is only part of the inflammatory cascade leading to the clinical appearance of Crohn’s disease. As the patient groups presented here are too small to perform a careful multivariate analysis, larger studies will now be required to define further the effects of glucocorticoid treatment on the activation of the NF-κBp65/IκBα system in patients with Crohn’s disease.
To date, most of the research on inflammatory bowel disease has focused on the role of lamina propria infiltrating inflammatory mononuclear cells.10-13 The above data, however, show that vascular endothelial cells also participate in the inflammatory response by activating the NF-κB/IκBα expression system. It will be important to define whether NF-κB induced endothelial gene expression such as upregulation of the adhesion molecules VCAM-1 and ICAM-1 or the monocyte chemotactic protein-1 is particularly involved in the recruitment of inflammatory cells to the mucosa3 6 47and therefore may actively participate in the maintenance of chronic inflammation in Crohn’s disease. Part of the therapeutic effect of glucocorticoids may therefore be explained by downregulation of endothelial NF-κBp65 and the subsequent prevention of leucocyte recruitment to the inflamed mucosa. To elucidate the role of endothelial cells in Crohn’s disease further, it will be necessary to define ligands and cellular receptors that mediate sustained endothelial NF-κB activation. The ligand-receptor interactions involved would be attractive targets for the development of novel anti-inflammatory drugs and are likely to become a prime focus of future investigations.
The authors wish to thank Dr Nancy Rice (Fredericks, USA) for the gift of the NF-κBp50/p105 antibodies and Dr P A Baeuerle (Tularik Inc, San Francisco, California, USA) for providing the NF-κBp65 plasmid Rc-CMV-p65. The excellent technical assistance of Mrs Christina Koch (immunohistology, reproduction of microphotographs) is acknowledged. This work was supported in part by Deutsche Forschungsgemeinschaft grants SFB no 601 (to PPN) and no 405/B9 (to FA) and a grant from the Deutsche Crohn/Colitis ulcerosa Vereinigung (to PPN). PPN performed this work during the tenure of a Schilling professorship.
↵* KT, AB, and FA contributed equally to this work.
- Abbreviations used in this paper:
- bovine serum albumin
- Crohn’s disease activity index
- electrophoretic mobility shift assay
- glucocorticoid receptor
- nuclear factor-kappa B
- recombinant NF-κBp65
- sodium dodecyl sulphate/polyacrylamide gel electrophoresis
- Tris buffered saline
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.