Background—Tumour necrosis factor (TNF) α and TNF-β are soluble ligands binding to TNF receptors with similar activities; soluble TNF receptors neutralise TNF activity by acting as inhibitors. Little is known about the cytokine/soluble receptor role in inflammatory bowel disease (IBD).
Aims—To test the hypothesis that an imbalance in secretion between TNF and TNF inhibitors plays a role in gut inflammation in patients with IBD.
Methods—The secretion of TNF-α, TNF-β, and soluble TNF receptors was compared in the culture supernatants of colonic biopsy specimens and isolated lamina propria mononuclear cells from patients with active colonic IBD.
Results—Spontaneous secretion of TNF-α in involved IBD mucosa was higher than in normal control and self limited colitis mucosa. Secretion of TNF-β was higher in patients with Crohn’s disease than in those with ulcerative colitis. Soluble TNF receptor in IBD mucosa inhibited TNF activity. Type 2 soluble receptor release from IBD mucosa was increased in active inflammation; release from lamina propria cells was not increased. Mucosal TNF-α production correlated with severity of disease.
Conclusions—Results showed enhanced secretion of TNF-α but failure to release enhanced amounts of soluble TNF receptor in lamina propria mononuclear cells of patients with IBD. An imbalance in secretion between TNF and TNF inhibitor may be implicated in the pathogenesis of IBD.
- Crohn’s disease
- mucosal immunology
- soluble TNF receptor
- tumour necrosis factor
- ulcerative colitis
Statistics from Altmetric.com
- Crohn’s disease
- mucosal immunology
- soluble TNF receptor
- tumour necrosis factor
- ulcerative colitis
Several factors have been implicated as possible initiating events in inflammatory bowel disease (IBD), including enteric bacteria or bacterial cell wall components. These factors may initiate a sequence of chronic inflammatory processes that are not appropriately down regulated. Normal intestinal immune responses must be carefully regulated so that an inflammatory insult is suppressed by local anti-inflamnmatory mechanisms. It has been hypothesised that chronic inflammation in IBD may be due to an imbalance between inflammatory and anti-inflammatory cytokine production.1 Tumour necrosis factor (TNF) ligands and the natural inhibitor of TNF, soluble TNF receptor (sTNF-R), play a major role in the regulation of inflammatory diseases such as sepsis and rheumatoid arthritis.2 3 The gastrointestinal tract is a prime target of TNF produced by blood mononuclear cells, accumulating 10% of a systemically administered dose.4 Administration of TNF induces extensive bowel haemorrhage and necrosis.5 TNF-α and TNF-β (lymphotoxin α) are soluble ligands binding to TNF receptors with similar effects; sTNF-R neutralises TNF activity.6 TNF mediated apoptosis is evoked by the intracellular death domain of the TNF receptors in cells.7 The two soluble TNF receptors, p55 type 1 (sTNF-R1) and p75 type 2 (sTNF-R2), are derived from the extracellular domains of TNF receptors by enzymatic cleavage.
Recent reports suggest that antibodies against TNF-α are therapeutic in Crohn’s disease.8-10 Serum and stool samples from patients with active Crohn’s disease or ulcerative colitis contained increased concentrations of TNF-α, probably derived from inflammatory cells in the gut lesions.11 12 However, others have reported that serum TNF-α levels and mucosal RNA transcript levels were not elevated in patients with IBD.13 14 Although TNF has been studied in IBD, little is known about the cytokine/soluble receptor role in the pathogenesis of IBD. The aim of this study was to test the hypothesis that an imbalance in secretion between TNF and TNF inhibitors plays an important role in gut inflammation, specifically in patients with IBD. We examined the secretion of TNF-α, TNF-β, and soluble TNF receptors in the culture supernatants of both colonic biopsy specimens and isolated lamina propria mononuclear cells (LPMCs) from patients with Crohn’s disease, ulcerative colitis, and controls.
Patients and methods
Seventy patients with IBD, being treated in the Hospital of Tohoku University School of Medicine, were prospectively included in the study. All patients had active disease, as assessed by colonoscopy. The diagnosis was based on routine clinical, radiological, and endoscopic criteria and confirmed by histological evaluation. Table 1 presents details of the patients with IBD. The activity of Crohn’s disease was assessed according to the criteria of Van Hees et al.15 Index values of 100–150 can be regarded as indicating mild inflammatory activity, values of 150–210 as indicating moderate activity, and values above 210 as indicating severe activity. The severity of ulcerative colitis was assessed according to the classification of Truelove and Witts.16
Normal control tissue samples were obtained from 14 patients with colonic polyps (four women; mean age 43 years, range 35–58). Inflammatory control tissue samples were obtained from eight patients with self limited colitis (three women; mean age 38 years, range 24–45).
Biopsy specimens were obtained by colonoscopy under direct vision from involved and normal appearing colonic sites in patients with Crohn’s disease and ulcerative colitis, and from normal mucosa of patients with a colonic polyp. Colonic mucosal explants were cultured for 24 hours at 37°C with 95% O2 and 5% CO2 by the established procedure.17 18 All culture media consisted of RPMI 1640 medium and 10% fetal calf serum (Life Technologies, Great Island, USA) with penicillin (100 U/ml) and streptomycin (100 μg/ml). After culture for 24 hours, the organ culture medium was collected and used for TNF-α, TNF-β, and sTNF-R assays. In our data, biopsy specimen wet weight was shown to correlate closely with protein content of tissue homogenates.
LAMINA PROPRIA MONONUCLEAR CELL CULTURE
Thirty four patients with IBD admitted for therapeutic bowel resection were studied. Table 2 shows the clinical data and treatment at the time of the study. Disease severity was higher in the ulcerative colitis group than in the Crohn’s disease group. The surgical control group consisted of 10 patients (two women; aged 38–65 years, mean age 52), admitted for bowel resection for adenocarcinoma of the large bowel.
Isolation of lamina propria mononuclear cells (LPMCs) was performed as described previously.18 19 The LPMCs were purified by density gradient centrifugation over Ficoll-Paque (Pharmacia, Uppsala, Sweden). The LPMCs were always 95–98% viable as determined by trypan blue staining. These cells were incubated in 24 well cell culture plates for 24 hours at 37°C in 95% air and 5%CO2 with and without 10 μg/ml of Escherichia colilipopolysaccharide (LPS; 0128 E coli, Sigma, St Louis, USA) as enteric bacterial cell wall components. The supernatant was then collected and used for TNF-α, TNF-β, and sTNF-R assays.
The leucocyte subpopulations present within the isolated LPMCs were analysed by flow cytometry using monoclonal antibodies that recognise CD4, CD8, CD3, and CD11b. No significant differences in the percentages of cell subpopulations were detected, when either involved or non-involved LPMCs from ulcerative colitis mucosa or Crohn’s disease mucosa were compared with controls. These findings are consistent with previous observations.20
IMMUNOASSAYS FOR TNF LIGANDS AND SOLUBLE RECEPTORS
TNF-α levels in the media were measured using a human TNF-α immunoenzymetric assay kit (Medgenix, Fleurus, Belgium). TNF-β levels in supernatant were measured using a human TNF-β enzyme linked immunosorbent assay kit (R&D systems, Minneapolis, USA). The limit of sensitivity of the assays used was 4.4 pg/ml for TNF-α, and 7.0 pg/ml for TNF-β. Anti-TNF-α antibody and anti-TNF-β antibody did not cross react with any other cytokine. Soluble TNF receptor levels in the supernatant were measured using a human soluble TNF receptor type 1 (p55) and type 2 (p75) immunoenzymetric assay kit (Medgenix). The limit of sensitivity of the assays is 1.0 pg/ml for sTNF-R1, and 1.0 pg/ml for sTNF-R2. All samples were analysed in duplicate. Results of cytokine concentrations were expressed according to the weight of the biopsy specimen (pg/mg tissue/ml culture medium) in supernatant of organ culture and according to the number of isolated cells (pg/106 cells/ml culture medium) in the supernatant fluid of mononuclear cell culture.
INHIBITION OF TNF ACTIVITY IN CYTOTOXIC ASSAY
Biological TNF activity was measured by the WEHI cell killing assay.21 The inhibitory activity of the culture media was assessed by its ability to inhibit TNF-α in the WEHI 164 cytotoxic assay. Briefly, mouse WEHI-164 cells (obtained from Tohoku University Cancer Cell Collection) were seeded in 96 well microplates in triplicate at a density of 20 000 cells/well. Culture samples (100 μl) from five patients with IBD were added alone or with the addition of 10 μl of monoclonal antihuman p75 TNF-R2 antibody (5 μg/ml; Genzyme, Cambridge, USA) or p75 sTNF-R2 (2 ng/ml; Medgenix). Culture medium (100 μl) and actinomycin D (1 μg/ml; Wako) was then added after five hours. Monoclonal antihuman p75 TNF-R2 antibody binds specifically to the extracellular domain of human TNF-R2 and recognises soluble forms of TNF-R2. Cell viability was assessed 24 hours later at 37°C and 5% CO2 by tetrazolium salts colorimetric assay (Boehringer Mannheim, Mannheim, Germany).
Results are expressed as median (interquartile range, IQR). The results were compared by the non-parametric Kruskal-Wallis test and the Pearson r correlation test. Significance was accepted as p<0.05 and determined by the Bonferroni (Dunn) post hoc test.
The acquisition and use of human blood and tissue for these studies were approved by the Human Ethics Committee of Tohoku University School of Medicine.
RELEASE OF TNF-α AND TNF-β BY ORGAN CULTURE
Spontaneous release of TNF-α measured in the supernatant fluid of involved Crohn’s colitis mucosa and involved ulcerative colitis mucosa was significantly higher than in normal control mucosa or inflammatory control mucosa (table 3). Non-involved mucosa from patients with Crohn’s disease produced significantly higher amounts of TNF-α than did that from normal controls. Concentrations of TNF-β measured in the medium of involved mucosa from patients with Crohn’s disease were significantly higher than in that from involved mucosa and non-involved mucosa from patients with ulcerative colitis or normal controls.
RELEASE OF SOLUBLE TNF RECEPTOR BY ORGAN CULTURE
Spontaneous release of sTNF-R2 measured in the supernatant fluid of involved Crohn’s colitis mucosa and involved ulcerative colitis mucosa was significantly higher than from normal control mucosa or inflammatory control mucosa (p<0.001; table 3). Secretion of sTNF-R2 tended to correlate with the amount of TNF-α released in mucosa from IBD patients (r=0.719, p<0.001).
When the ratio TNF-α:sTNF-R2 was calculated, an increased production ratio was not observed with involved mucosa of patients with IBD compared with ratios from non-involved mucosa of patients with IBD, control mucosa, and infectious colitis mucosa (fig1).
In the organ culture, sTNF-R1 secretion was no higher in patients with IBD than in controls. No correlation was observed between levels of TNF-α and levels of sTNF-R1 (r=0.32, NS).
RELEASE OF TNF-α AND TNF-β BY ISOLATED CELL CULTURES
Under LPS stimulated conditions, secretion of TNF-α in the supernatant fluid from isolated LPMCs of involved mucosa from Crohn’s disease and ulcerative colitis was significantly higher than from control mucosa or non-involved mucosa from Crohn’s disease and ulcerative colitis (table 4).
Spontaneous TNF-α secretion by isolated LPMCs of involved mucosa from Crohn’s colitis (median 270 (IQR 80–1131) pg/ml) and ulcerative colitis (275 (90–484) pg/ml) was also higher than that produced by non-involved mucosa from Crohn’s disease (30.0 (18–205) pg/ml) and ulcerative colitis (36.0 (31–57.3) pg/ml) or that of control mucosa (26.0 (18–54.3) pg/ml) (p<0.05). Under LPS stimulated conditions, TNF-β secretion was not higher in LPMCs from patients with IBD than from controls (data not shown).
RELEASE OF SOLUBLE TNF RECEPTORS BY ISOLATED CELL CULTURES
Release rates of sTNF-R2 by LPS stimulated LPMCs of the involved mucosa from Crohn’s colitis and ulcerative colitis were not significantly higher than those of non-involved mucosa from Crohn’s disease and ulcerative colitis and control mucosa (table 4). No correlation was observed between levels of TNF-α and levels of sTNF-R2 (r=0.18, NS). Spontaneous sTNF-R2 secretion by isolated LPMCs of involved mucosa from Crohn’s colitis (median 240 (IQR 170–390) pg/ml) and ulcerative colitis (680 (340–1100) pg/ml) was no higher than that produced by non-involved mucosa from Crohn’s disease (260 (210–360) pg/ml) and ulcerative colitis (205 (170–350) pg/ml) or that of control mucosa (940 (780–1100) pg/ml) (p<0.05).
When the ratio of median TNF-α:median sTNF-R2 was calculated, an increased production ratio was observed in LPMCs from involved mucosa of patients with Crohn’s disease and ulcerative colitis (fig 2). In LPS stimulated LPMCs, sTNF-R1 secretion was no higher in patients with IBD than in controls.
RELATION OF MUCOSAL TNF SYSTEM AND SEVERITY OF DISEASE
Clinical disease severity tended to correlate with the secretion of TNF-α in colonic lesions (r=0.581, p<0.0001) and in LPS stimulated LPMCs from patients with IBD (r=0.739, p<0.0001; table 5). No correlation was observed between severity and the levels of TNF-β, sTNF-R1, or sTNF-R2.
EFFECT OF SOLUBLE TNF RECEPTOR ON TNF ACTIVITY
To assess the function of sTNF-R, we incubated culture samples with neutralising concentrations of monoclonal antibody to sTNF-R. TNF cytotoxicity was significantly higher in organ culture samples with sTNF-R2 monoclonal antibody than in samples only (53.6 (46.0–70.5)% versus 26.7 (22.0–34.8)%, p<0.05; fig 3). Furthermore, the effect of inhibitor could be reversed by incubation with human sTNF-R2 (29.7 (4.5–36.0)%). No cytotoxicity to WEHI 164 cells was observed after treatment with monoclonal antibody alone.
Our study showed that a large amount of TNF-α was released by LPMCs from inflamed mucosa of patients with IBD. Breeseet al also reported that TNF-α positive cells increased in the mucosa of inflamed intestine.22When TNF mRNA has been measured, different results have been obtained. Isaacs et al could not detect any TNF-α transcripts in the mucosal samples from patients with IBD.14 It is now known however that most autocrine cytokines and soluble cytokine receptors are derived from the cell surface by enzymatic cleavage; messenger RNA analysis for cytokine and the receptor is therefore not informative, as it does not discriminate between the cell surface form and the soluble form.23Cytokine membrane molecules are enzymatically cleaved from the cell surface and then released into the extracellular medium in the form of soluble fragments.24 In particular, TNF-α, a proinflammatory membrane anchored cytokine, is released from the membrane glycoprotein of metalloprotease disintegrins by TNF-α convertase.25 Several physiological processes are known to regulate the activity of TNF-α with respect to synthesis, processing, and effects on the target cell.26 Therefore, this study of soluble TNF and soluble TNF inhibitor may be important for determination of the pathogenesis of IBD.
Our study showed that LPS stimulation increased TNF-α secretion only twofold in LPMCs of non-involved mucosa and mucosa from controls. Similar results were reported by Reinecker et al.20 The relatively poor response of LPMCs to LPS indicates previous stimulation or a state of relative anergy.27 Our study showed that TNF-α production in LPMCs of severe inflamed mucosa could be enhanced further by LPS stimulation. We suggest that in newly recruited LPMCs, TNF-α production is enhanced by LPS stimulation, as Rugtveitet al report.28
Our bioassay data showed that sTNF-R had a functional effect as a TNF inhibitor in inflamed mucosa. Spontaneous secretion of sTNF-R2 from inflamed mucosa in IBD was significantly higher than in control mucosa, suggesting that sTNF-R2 acts as a naturally occurring damper of the inflammatory process in IBD. LPS (endotoxin), a major component of the outer membrane of Gram negative bacteria, has a number of immunological activities. A release of sTNF-R2, but not sTNF-R1, is enhanced after activation with E coli LPS.29When stimulated by LPS, sTNF-R can induce TNF tolerance, a well controlled active response orchestrated to prevent excessive inflammation.2 29 The LPS stimulated sTNF-R2 secretion by LPMCs obtained from patients with IBD was low. Rugtveitet al also did not observe any significant increase in the low levels of sTNF-R2 produced by LPMCs after any type of stimulation.28 Foley et alreported a decrease in TNF inhibitor in serum of patients with Crohn’s disease compared with those with sarcoidosis or tuberculosis.30 Bouma et alreported that TNF activity was decreased by bioassay in stimulated PBMCs of patients with ulcerative colitis.31 Different quantitative results can be obtained when TNF is measured in biological fluids by bioassay and immunoassay. Our TNF inhibitor bioassay showed the presence of the antigenic form of TNF that cannot be detected by routine bioassay, such as complexes with the soluble receptors. These results suggested that release of the TNF inhibitor from intestinal mononuclear cells is down regulated in IBD. Enteric bacterial derived LPS may initiate a sequence of inflammatory processes that are not appropriately down regulated by TNF inhibitor in IBD. LPS stimulated sTNF-R2 production is inhibited by bactericidal/permeability inducing protein.29 Monajemi et alreported that mucosal levels of bactericidal/permeability inducing protein are increased in IBD.32 Soluble TNF receptors may regulate TNF activity; conversely, at a high TNF:sTNF-R ratio, these can act as carriers for TNF and prolong their activities by stabilising the trimeric structure.29 The LPMCs in IBD lesions in particular had a high TNF:sTNF-R ratio. The intestinal mucosal mononuclear cells in IBD lesions may abrogate the function of TNF inhibitor.
TNF-β plays an important role for gut homing in intestinal lymphocytes and development in intestinal lymphoid organs.33 34 Our results showed that the spontaneous production of TNF-β in the mucosa of patients with Crohn’s disease was increased, but this was not observed in that of patients with ulcerative colitis. Bouma et al reported TNF-β production was increased in stimulated peripheral blood mononuclear cells of patients with Crohn’s disease.31 We found that the normal appearing mucosa in patients with active Crohn’s disease produced larger amounts of TNF-α than normal control mucosa; this was not observed in patients with ulcerative colitis. Reimundet al reported increased production of TNF-α by normal intestinal biopsy specimens from patients with Crohn’s disease.35 These findings suggest that Crohn’s disease and ulcerative colitis have a different pathogenesis. TNF affects the tight junction between epithelial cells and increases the flow of solute between cells and across the epithelium.36 37 Increased intestinal permeability is a consistent finding in patients with active Crohn’s disease.38 These studies are important and the sequence of appearances is shown in the neoterminal ileum, where recurrent lesions progress from normal appearing mucosa of Crohn’s disease by an effect of faecal stream.39
It has been reported that IBD clinical activity is virtually independent of the severity of the mucosal lesions and tissue destruction.40 41 Our study showed that production of TNF-α in the mucosa and the LPMCs was associated with severity of IBD; TNF-α may have a pivotal role in the degree of severity produced. We suggest that a relative deficiency of TNF inhibitor in involved intestinal mononuclear cells may contribute to inflammation, which is relevant to current concepts of immunopathogenesis for IBD. Monoclonal anti-TNF-α antibody treatment was effective for patients with severe graft versus host disease and the gut lesions responded very well.42 Binding of anti-TNF-α antibody resulted in killing of TNF positive cells by antibody dependent cellular toxicity.43 Therefore, it is more likely that anti-TNF induces tolerance in IBD, killing the TNF dysregulatory cells rather than simply neutralising TNF released from these cells.
A recombinant human sTNF-R2-Fc fusion protein has been developed for therapeutic neutralisation of TNF-α in patients with active rheumatoid arthritis.44 We suggest that inhibition of TNF-α may have potential in the treatment of IBD. For example, inhibition of TNF-α by thalidomide was effective in a patient with Crohn’s disease.45 TNF is known to have affinity for heparin and an increase in the amount of sTNF-R shed is caused by heparin.46 Heparin can be protective against the harmful effects of TNF and has been shown to be useful in the treatment of ulcerative colitis.47 Anti-CD4 treatment also results in increased shedding of sTNF-R in patients with Crohn’s disease.48 Further investigation of the inhibitor, such as sTNF-R in autoimmune tissue damage, will allow us to decide the validity of our suggestion.49 50
This study was supported by grants from the Research Committee for Inflammatory Bowel Disease of the Japanese Ministry of Public Welfare
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.