Hyperexpression of CD40 Ligand (CD154) in Inflammatory Bowel Disease and Its Contribution to Pathogenic Cytokine Production¹

Crohn’s disease (CD)³ and ulcerative colitis (UC) are the two major forms of chronic inflammatory bowel disease (IBD). Although their etiopathology remains unknown, increasing evidence has outlined that immune mechanisms play an important role in the pathogenesis (1, 2, 3). Both CD and UC are characterized by leukocytic infiltrates in inflamed intestinal mucosa, especially by activated CD25⁺ T cells, B cells, and RFD9⁺ CD68⁺ macrophages (1, 2, 3). These activated leukocytes have various effector functions in the mucosa, which include secretion of Th1/Th2 cytokines by T cells as well as TNF and IL-12 production by macrophages (4, 5, 6). Despite evidence for CD4⁺ T cell-mediated or -dependent immune responses in inflamed mucosa of IBD (4, 6, 7, 8), the exact role of these cells in the disease pathogenesis and the initiating event for their activation still remain to be investigated.

CD40 ligand (CD40L or CD154) is a 33-kDa transmembrane glycoprotein that is transiently expressed on the surface of activated T cells, predominantly the CD4⁺ T cells (9, 10). Its receptor, CD40, is found on B cells, dendritic cells, monocytes, and some other cell types (11). The cognate interaction of CD40L on activated T cells with CD40 on B cells is essential for B cell proliferation, differentiation, Ig production, and isotype switch (9, 10, 11). Expression of CD40 on monocytes, macrophages, and dendritic cells relates to the role of CD40 signaling in cellular immune responses (12, 13). CD40 ligation of monocytes and macrophages induces cytokine production, tumoricidal activity, and rescue from apoptosis (14, 15, 16, 17). Ligation of dendritic cell CD40 enhances their differentiation and activation, with enhanced expression of costimulatory molecules such as CD58, CD80, and CD86, increased cytokine production, and inhibition of apoptosis (18). Furthermore, CD40 signaling can induce expression of ICAM, VCAM-1, and E-selectin on endothelial cells, suggesting that signaling through CD40 during T cell-endothelial cell interactions may be an important step in leukocyte recruitment (19). In vivo studies have indicated the importance of the CD40-CD40L interactions in the generation of humoral immune responses and germinal center formation (20, 21), in the priming and activation of Ag-specific T cells (22), in the temporal activation of macrophages (23), as well as in the protective cell-mediated immune responses through T cell-mediated macrophage activation against intracellular parasite infections such as Pneumocystis, Listeria monocytogenes, and Leishmania (24, 25, 26).

A number of studies have focused on the impact of CD40-CD40L interactions on the immunopathogenesis of human autoimmune diseases, such as systemic lupus erythematosus (SLE) (27, 28), rheumatoid arthritis (RA) (29), multiple sclerosis (30), Graves’ disease (31), as well as on allograft rejection (32). CD40L expression has been shown to be significantly increased on peripheral blood T cells in SLE and RA or synovial fluid T cells in RA and can be maintained and stabilized at a relatively high level under stimulatory conditions. CD40L expressed by these T cells induces B cell Ig production and triggers dendritic cell IL-12 expression (27, 29). Administration of mAb directed against CD40L has been shown to effectively prevent inflammation in experimental allergic encephalomyelitis (30), murine lupus nephritis (33), murine arthritis induced by immunization with type II collagen (34), and murine experimental autoimmune thyroiditis (35), and to dramatically prolong murine skin and cardiac allograft survival (36). All have come to the conclusion that CD40-CD40L interactions play a central role in immune responses.

There is some evidence that CD40-CD40L interactions might also play a role in the pathogenesis of IBD. An in vivo experiment has shown that anti-CD40L mAb used in 2,4,6-trinitrobenzene sulfonic acid-induced colitis can effectively prevent mucosal inflammation and IFN-γ production by lamina propria CD4⁺ T cells (37). CD40L transgenic mice with the highest transgene copy numbers were shown to acquire a lethal IBD marked by infiltration of CD40L⁺ T cells and CD40⁺ cells into diseased tissues (38). However, little is as far known about the involvement of CD40-CD40L interactions in the immune responses of human IBD. We here investigated whether T cells from the lamina propria of IBD patients might induce IL-12 and TNF production by monocytes through CD40-CD40L interactions.

Fourteen surgical patients with CD were studied. This group included 10 men and four women, aged 24–46 years. Nine patients had pure ileal involvement, and five had ileocolic involvement. Colonic tissues were also obtained from 10 surgical patients with UC undergoing colectomy. This group consisted of seven men and three women, aged 26–51 years. Of all patients with IBD (n = 24), 19 were receiving sulfasalazine or an oral aminosalicylic acid preparation at the time of operation, and 5 patients were on no treatment. None of them received immunosuppressive drugs such as corticosteroids, cyclosporin A, or azathioprine, and none of them had HIV infection. The preoperative diagnosis of IBD was based upon classical clinical, radiological, and endoscopic features and was finally confirmed by histological examination of the resection specimens. Indications for bowel resection in CD were the presence of fistulae (n = 3), abscess formation (n = 1), and stenosis with clinical signs of obstruction (n = 10); in UC the indications for surgery were mostly therapy-resistant inflammatory colitis. In addition, 14 samples of intestinal tissue (seven from ileum and seven from colon) were obtained from surgical specimens of 12 patients (nine men and three women, aged 38–58 years) undergoing right hemicolectomy for carcinoma, which were macroscopically normal and remote from areas of disease involvement, and which were identified by histology as being normal.

All surgical specimens were immediately transferred to the laboratory of pathology in a sterile container filled with ice-chilled RPMI 1640 solution. Samples were taken from macroscopically diseased areas in IBD patients. The samples (5 × 8-mm mucosal area each) for immunohistochemistry (n = 11 in CD, n = 6 in UC, n = 7 in normal ileum, and n = 7 in normal colon) were embedded immediately in OCT compound (Miles, Elkhart, IN) and snap-frozen in liquid nitrogen-precooled 2-methylbutane (ACROS, Fair Lawn, NJ) at −70°C. The frozen samples were further stored at −70°C until cryostat sectioning. The samples (2 × 3-cm mucosal area each) for isolating lamina propria lymphocytes were also obtained. Additional samples were taken for routine pathological diagnosis based on light microscopic examination of hematoxylin and eosin-stained sections of formalin fixed, paraffin-embedded materials.

In parallel to those surgical specimens, heparinized peripheral blood samples were obtained from outpatients with active CD and UC, including 10 cases with CD (seven men and three women, aged 21–42 years) and 10 cases with UC (six men and four women, aged 26–47 years). The diagnosis was established by conventional clinical features and histological criteria. Only five patients with CD and four patients with UC were treated with sulfasalazine or an oral aminosalicylic acid preparation. None received immunosuppressants such as corticosteroids and cyclosporin A. In addition, blood samples from 10 healthy volunteers (seven men and three women, aged 28–38 years) were also taken for comparison.

The mAbs used in this study included anti-CD3 mAb OKT3 (mouse IgG2a; American Type Culture Collection, Manassas, VA), anti-CD40 mAb 5D12 (39) (mouse IgG2b; PanGenetics, Amsterdam, the Netherlands), and anti-CD40L mAb M90 (40) (mouse IgG1; Immunex, Seattle, WA). FITC-conjugated mouse anti-human CD40L mAb (clone 24-31, mouse IgG1) was obtained from Ancell (Bayport, MN). Isotype-matched control mAb mouse-IgG-FITC was obtained from PharMingen (San Diego, CA). For flow cytometry, all mAbs were used at optimal saturating concentrations as recommended by the manufacturers.

Cells were cultured in RPMI 1640 medium (Boehringer Ingelheim BioWhittaker, Heidelberg, Germany) supplemented with 0.3 mg/ml l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin (Boehringer Ingelheim BioWhittaker), 4 U/ml polymyxin B, 5 μg/ml amphotericin B (Sigma, St. Louis, MO), and 10% iron-supplemented bovine calf serum (HyClone, Logan, UT) at 37°C in 5% CO₂-humidified atmosphere.

PBMC from healthy adult volunteers and IBD patients were isolated from heparinized blood over Ficoll-Hypaque (Lymphoprep; Nycomed, Oslo, Norway) gradients. The cell suspension was washed three times in PBS free of calcium and magnesium. Viable cells were determined by trypan blue exclusion and counted with a hemocytometer or Coulter counter (Harpenden, Herts, England). PBMC were then resuspended at 5 × 10⁶ cells/ml in complete culture medium consisting of RPMI 1640 supplemented with 2 mM l-glutamine, penicillin (100 U/ml), streptomycin (100 μg/ml), and 10% bovine calf serum. Monocytes were removed by cold agglutination by rotating tubes for 30 min at 4°C. The agglutinated monocytes were sedimented, and the remaining cells were further purified using T cell lympho-kwik (One Lambda, Los Angeles, CA) supplemented with complement-fixing anti-NK and anti-monocyte mAbs, as previously reported (41). The resultant peripheral blood T cell (PB-T) preparations were >98% CD3⁺ and <1% CD16⁺ cells, as determined by flow cytometry (FACSort, Becton Dickinson, San Jose, CA).

Monocytes used in this study were obtained at several occasions from a single healthy donor. PBMC were isolated according to the method as described above, then placed in 24-well flat-bottom plates (Nunc, Roskilde, Denmark) in RPMI 1640 medium without serum. The cell concentrations were prepared according to numbers of CD14⁺ cells (2.5 × 10⁵ cells in 1 ml/well), as predetermined by flow cytometric analysis. After 2 h of adherence in a 5% CO₂ air-humidified atmosphere at 37°C, the nonadherent cells were aspirated, and the wells were washed three times with PBS. The adherent cells were cultured with RPMI 1640 medium supplemented with 2 mM l-glutamine, penicillin (100 U/ml), streptomycin (100 μg/ml), and 10% bovine calf serum. Viability was >98%, and the cell preparation contained >85% CD14⁺ monocytes, as assessed by flow cytometric analysis.

Lamina propria mononuclear cells (LPMC) were first isolated from both inflamed and control mucosa according to a method as described by Bull and Bookman (42). Briefly, the mucosa was first dissected from the underlying musculature and washed thoroughly with cold PBS to remove all debris and blood. The epithelium was removed from the lamina propria by incubation with 2 mM DTT and 1 mM EDTA in PBS at 37°C for 2 × 20 min under gentle shaking. Tissues were subsequently washed in PBS three times, minced into 2 × 2-mm pieces, and digested with collagenase A (1 mg/ml) and DNase (5 μg/ml) (Boehringer Mannheim, Mannheim, Germany) at 37°C, 5% CO₂ for 3–4 h. Lamina propria cells were pooled and passed through a 70-μm cell strainer (Becton Dickinson) four times. To this end, pellets (lamina propria cells) were resuspended in 5 ml of 40% Percoll, layered over 5 ml of 100% Percoll (Pharmacia, Uppsala, Sweden), and centrifuged at 600 × g for 20 min at room temperature. LPMC were collected from the interphase. Then, LP-T cells from LPMC were prepared according to the method described above.

For stimulation with immobilized anti-CD3, culture wells (24-well plates; Nunc) were precoated with anti-CD3 at a final concentration of 5 μg/ml in 300 μl PBS for 4 h at 37°C. Before being used, wells were washed three times with 1-ml portions of PBS to remove unbound mAb. Purified PB-T or LP-T cells (5 × 10⁵ cells in 1 ml/well) were cultured in a 37°C humidified atmosphere of 5% CO₂ for varying time periods, i.e., 16, 24, and 48 h. Cells were then harvested and assessed for the expression of CD40L by staining with either anti-CD40L-FITC or isotype-matched control mAb for 30 min at 4°C. After two washes with PBS, the cells were fixed with 0.5 ml 1% paraformadehyde in saline and analyzed on a FACSort using Lysis II software.

Peripheral blood monocytes from a healthy donor were cultured at a final concentration of 2.5 × 10⁵/ml, alone or with LP-T cells from IBD patients or control patients at the T cell to monocyte ratio of 2:1. After 48 h of culture, the supernatants of these cultures were harvested and stored at −70°C for IL-12 and TNF assay. To further ascertain whether CD40L expression on LP-T cells was critical to induce monocyte cytokine production, blocking experiments were set up in which purified monocytes were cocultured with LP-T cells in the presence of either anti-CD40L mAb M90 (10 μg/ml) or anti-CD40 mAb 5D12 (10 μg/ml). IFN-γ (Boehringer Mannheim) was added to the cultures where indicated at the final concentration of 1000 U/ml.

TNF was assayed by sandwich ELISA using matched Ab pairs (BioSource, Europe, Fleurus, Belgium). The sensitivity of this assay was 10 pg/ml. IL-12 was detected by sandwich ELISA employing anti-IL-12 mAb (clone 24945.11, mouse IgG1) as a capture Ab in combination with the biotinylated anti-human IL-12 detection Ab (human IL-12 specific goat IgG) (R&D Systems, Minneapolis, MN), which specifically detects the bioactive p70 heterodimer. The sensitivity of this assay was 10 pg/ml.

Cryostat sections (5 μm) of gut tissue were cut and stained immunohistochemically using an avidin-biotin-peroxidase complex technique for the presence of CD40 and CD40L. All procedures were conducted at room temperature in a humidified chamber. Briefly, cryostat sections were air-dried and fixed in acetone for 10 min and afterward rinsed in PBS (pH 7.2). They were first incubated for 30 min with either an anti-CD40 mAb 5D12 or an anti-CD40L mAb M90, both at the working concentration of 10 μg/ml. The sections were then incubated with the second Ab, biotin-labeled rabbit anti-mouse IgG (Dako, Glostrup, Denmark; 1:400 dilution), for 30 min. To efficiently block endogenous peroxidase, sections were also incubated in methanol solution containing 0.3% (v/v) H₂O₂ for 30 min. After three washes with PBS, the avidin-biotin-peroxidase complex (Dako) was added, and sections were incubated for 30 min. The reaction product was developed using a solution of 0.05 M acetate buffer (pH 4.9) containing 0.05% 3-amino-9-ethylcarbazole (Janssen, Beerse, Belgium) and 0.01% H₂O₂ as substrates for 10 min. After color development, sections were counterstained with hematoxylin and mounted in glycerol medium (BDH, Dorset, U.K.). Negative control experiments were performed by incubating sections with irrelevant isotype-matched mouse IgG and by omitting the primary Ab.

In addition to the immunohistochemical stainings, routine H&E stainings were also performed on cryostat sections. Routinely and immunohistochemically stained sections were analyzed randomly and scored for the presence of inflammation. To determine the frequency of immunohistochemical staining cells, 5 high power fields (magnification ×40) of the mucosa were randomly selected. CD40⁺ and CD40L⁺ cells as well as the total number of stromal cells were counted in the lamina propria. The percentage of positive cells was recorded as: [(positive cells)/(total stromal cells)] × 100.

Double staining for CD40 and CD20 (B cells) or CD68 (activated macrophages) as well as for CD40L and CD4 (CD4⁺ T cells) or CD8 (CD8⁺ T cells) was performed by combining the peroxidase technique (brown reaction product) and the alkaline phosphatase-anti-alkaline phosphatase technique (blue reaction production) (43). Endogenous alkaline phosphatase was blocked by levamisole. The brown reaction was obtained using 3,3′-diaminobenzidine (Sigma), while the blue reaction was observed using Fast Blue BB salt (4-benzoylamino 2,5 diethyoxybenzene-diazonium chloride; Sigma). Anti-CD20 mAb L26 and anti-CD68 mAb KP1 (Dako) were used for the identification of B cells and macrophages, respectively. Anti-CD4 mAb MT310 and anti-CD8 mAb DK25 (Dako) were applied for the identification of CD4⁺ and CD8⁺ T cells, respectively.

All values in this study were expressed as mean ± SEM. Patients and controls were compared using the two-tailed Student’s t test, while comparisons of cytokine productions by monocytes in the presence or absence of T cells, or in the presence or absence of mAb to CD40L or CD40, were made using the Wilcoxon test for paired samples. A value of p < 0.05 was considered statistically significant.

Despite evidence for CD4⁺ T cell-mediated or -dependent immune responses in the inflamed mucosa of IBD, the exact role of these cells in disease pathogenesis still remains to be investigated. To study the potential role of LP-T cells in the induction of cytokine production by gut macrophages, we have isolated LP-T cells. In previous studies human monocytes were shown to express CD40 (14, 15). Nonautologous peripheral blood monocytes from a healthy donor were isolated and incubated with LP-T cells from either inflamed mucosa of IBD patients or normal mucosa of control patients without any further stimulus. Freshly isolated monocytes express low levels of CD40, and this CD40 is up-regulated during culture. As shown in Fig. 1, A and C, only low levels of IL-12 and TNF were measured in the supernatants of the monocytes cultured alone or incubated with control LP-T cells. In contrast, when cocultured with LP-T cells from inflamed mucosa of IBD patients, monocytes produced significantly higher amounts of IL-12 and TNF (p < 0.005). LP-T cells from all groups cultured alone did not produce any IL-12 and TNF under nonstimulatory conditions (data not shown). LP-T cells were also isolated from uninvolved tissues of two patients with CD and two patients with UC. These LP-T cells, similar to control LP-T cells, could induce only low levels of TNF and IL-12 production by monocytes, and the levels were significantly lower than those produced by monocytes cocultured with IBD LP-T cells from inflamed areas (data not shown). Our data indicate that LP-T cells from inflamed mucosa of IBD have a high capacity to induce monocyte cytokine production.

IFN-γ has been shown to prime monocytes for production of IL-1β, IL-6, IL-8, TNF, and IL-12 and to up-regulate CD40 expression (14, 15, 44). IFN-γ also plays an important role in IBD, especially in CD (1, 2, 3). Therefore, it was of interest to assess whether IFN-γ could enhance monocyte IL-12 and TNF production when cocultured with IBD LP-T cells. IFN-γ (1000 U/ml) was added to the aforementioned coculture system, and, after 48 h of culture, the cell supernatants were collected and assayed for IL-12 and TNF. Fig. 1 shows that IFN-γ significantly increased the production of IL-12 and TNF by monocytes either cultured alone or cocultured with LP-T cells from all groups (p < 0.05). Also in the presence of IFN-γ, the levels of IL-12 and TNF in the monocyte and IBD LP-T cell cocultures were significantly higher than those in the monocyte and control LP-T cell cocultures (p < 0.05).

To analyze whether interactions of CD40 and its ligand were involved in delivering this Th cell-dependent signal to monocytes, the anti-CD40L mAb M90 and anti-CD40 mAb 5D12 were added to the cultures to know whether these Abs could interfere with the ability of IBD LP-T to induce monocyte IL-12 and TNF synthesis. Both Abs can block CD40-CD40L interactions (39, 40). In a series of 13 experiments (Fig. 2), the inclusion of M90 (10 μg/ml) resulted in a 39–100% inhibition of IL-12 and a 34–100% inhibition of TNF production. In contrast, inclusion of 5D12 at the concentration of 10 μg/ml was also associated with a 26–100% inhibition of IL-12 and a 33–100% inhibition of TNF production (Fig. 2). The data confirm functional CD40L expression on LP-T cells from inflamed mucosa and suggest that LP-T cells from inflamed mucosa directly induce monocyte IL-12 and TNF synthesis in a CD40L-dependent manner.

The levels of CD40L expression on LP-T cells from IBD patients and controls were then studied by flow cytometric analysis. As shown in Figs. 3 A and 4, CD40L expression by freshly isolated LP-T cells from IBD patients was significantly higher, albeit at low levels, as compared with control LP-T cells (p < 0.005). LP-T cells were then activated with immobilized anti-CD3, and CD40L expression was determined at various time intervals of 16, 24, and 48 h. In both groups, maximum levels of CD40L expression were reached at (or before) 16 h of activation. In control subjects, CD40L expression was shown to decrease gradually to near baseline at 48 h of culture. In IBD patients, CD40L expression was maintained at significantly higher levels on activated LP-T cells at each time point of culture as compared with controls (p < 0.0005 at each time point in CD group; p < 0.05 at 16 and 24 h and p < 0.0005 at 48 h in UC group). In three cases of CD and two cases of UC, expression was still detected on 10.6 ± 2.1% of LP-T cells after 72 h of culture. LP-T cells were also isolated from uninvolved gut mucosa of two CD patients and two UC patients. Expression of CD40L on these freshly isolated LP-T cells and the kinetics of its expression after stimulation with immobilized anti-CD3 were similar to those in control patients (data not shown).

PB-T cells from another group of IBD patients and healthy subjects were subsequently activated with immobilized anti-CD3, and CD40L expression was determined at various time intervals of 16, 24, and 48 h on a FACSort. As shown in Fig. 3 B, no CD40L expression was found on resting PB-T cells from all groups. Activated T cells from all groups reached a maximum level of CD40L expression at 16 h following immobilized anti-CD3 stimulation, with a decline of CD40L expression thereafter. CD40L expression in normal subjects was shown to decrease gradually and returned to resting levels at 48 h of culture. The levels and kinetics of CD40L expression on normal PB-T cells were similar to those previously reported for human PB-T cells stimulated with immobilized anti-CD3 (45, 46). Importantly, expression of CD40L on activated T cells from IBD patients was demonstrated to be significantly higher at 16, 24, and 48 h of culture as compared with healthy controls (p = 0.032 at each time point).

In the next series of experiments, the expression of CD40L in situ was examined. As shown in Fig. 5, the number of CD40L⁺ cells in the lamina propria was significantly higher in inflamed ileum from patients with CD and in inflamed colon from patients with UC compared with normal controls (37.4 ± 3.0 in CD, 30.6 ± 4.1 in UC, 5.4 ± 2.4 in normal ileum, and 6.1 ± 2.2 in normal colon, p < 0.0001). No difference was seen between CD and UC. In the sections from diseased areas of CD patients, large numbers of positive cells were present in the mucosa where they tended to be more numerous in the lower third of the lamina propria. Many positive cells were also found in the submucosa, and some of them were present in the muscularis propria, the septum between the internal and external muscle layers and the serosa. These positive cells manifested a dot-like pattern or punctuate paranuclear staining, while a small number of positive cells exhibited a membrane-staining pattern. Interestingly, CD40L⁺ cells were preferentially located in the mantle zones of lymphoid follicles and were rare in the germinal centers. Intraepithelial lymphocytes were faintly positive. In the cryostat sections from inflamed areas of UC patients, the CD40L⁺ cells showed a diffuse distribution in the lamina propria, but only faint expression in submucosa, muscularis, and subserosa. In CD, the distribution was rather patchy and transmural. In contrast, only very few positive cells were detected in the control tissue sections, and they were entirely absent in the submucosa and muscularis from these samples. Intestinal epithelial cells were consistently negative for CD40L in all specimens. Additionally, in three samples of normal mucosa from active CD and two samples from inactive CD only few CD40L⁺ cells were found in the lamina propria, and the pattern looked similar to normal ileum (data not shown).

In situ expression of CD40 in gut tissue from CD, UC, and control patients is also shown in Fig. 5. The results indicated the presence of immunodetectable CD40 in all samples of diseased areas from both CD and UC patients. The number of CD40⁺ cells in the lamina propria was significantly higher in inflamed ileum from patients with CD and in inflamed colon from patients with UC compared with the normal ileum and colon (39.8 ± 7.5 in CD, 41.4 ± 8.1 in UC, 4.6 ± 1.7 in normal ileum, and 6.1 ± 2.0 in normal colon, p < 0.0001), and no difference between CD and UC was observed. Looking at the distribution in inflamed areas of IBD, it appeared that the positive cells were present in the mucosal stroma in a diffuse pattern, as single cells and occasionally in small clusters, which could be associated with glandular epithelium, mainly at the base of glands. In addition, numerous CD40⁺ cells were present in the submucosa and muscularis propria (mainly in the external layer), and lower numbers in the subserosal connective tissue in CD. CD40⁺ cells were found abundantly in the submucosa and the deeper bowel wall wherever lymphoid follicles were present. Interestingly, CD40⁺ cells were mainly concentrated in the germinal centers of active follicles and the number gradually decreased toward the mantle zones. They could also be present in clusters in the connective tissue septum between the internal and external muscle layers, laying close to intrinsic ganglia. In contrast, only few CD40⁺ cells were found in the submucosa, muscularis, and subserosa on the cryostat sections of UC. A small population of CD40⁺ cells was also found in the mucosa but not in the submucosa, muscularis, and subserosa of normal ileum and colon. Intestinal epithelial cells and endothelial cells were consistently negative for CD40 in samples from both diseased and nondiseased or normal areas.

Two patients with active CD with high scores for expression of CD40 and CD40L were selected to identify the phenotype of these positive cells. Double staining for CD40L and CD4 or CD8 showed that the majority of CD40L⁺ cells concomitantly expressed CD4, and some of them showed CD8, but not all CD4⁺ and CD8⁺ T cells showed CD40L positivity. The double positive cells were mainly situated in the lamina propria, and some were found in submucosa, muscularis propria, and serosa, similar to the repertoire of CD40L single staining. In contrast, double staining for CD40 and CD20 (B cells) or CD68 (macrophages) indicated that in the sections from patients with CD, CD40⁺ cells were mainly B cells in the lymphoid follicles and in the lamina propria. Some macrophages in the lamina propria were also CD40⁺ (Fig. 6).

In this study, we have demonstrated that isolated LP-T cells from IBD patients in a coculture system with normal monocytes could induce secretion of TNF and IL-12 by monocytes in a CD40/CD40L-dependent way. Moreover, increased expression of both CD40 and CD40L was found by immunohistochemistry and FACS analysis in inflamed mucosa of IBD patients. Therefore, the interactions of CD40L⁺ T cells with CD40⁺ target cells may play a role in the cytokine production and the activation of B cells and macrophages in IBD lesions. The cytokines that were shown here to be induced by CD40-CD40L interactions, namely IL-12 and TNF, are considered as important proinflammatory mediators in the immunopathogenesis of IBD (5, 6). IL-12 plays a critical role in regulating the balance between the Th1 and Th2 cells and in promoting Th1 responses and cell-mediated immunity (47). It plays a pivotal role in mediating expansion of CD4⁺ T cells of the Th1 subtype and leads to marked IFN-γ production as a major factor in the immunopathogenesis of CD (6). TNF has been recognized as a highly proinflammatory protein that is involved in the induction of fever, insulin resistance, bone resorption, anaemia, sepsis, and in the activation of granulocytes and B and T cells (48). The role of endogenous TNF in regulating immune responses has also been confirmed in TNF-deficient mice, characterized by the high susceptibility to systemic or gastrointestinal candidiasis correlated with an impaired Th1 responses (49). Administration of anti-TNF or anti-IL-12 mAb has been shown to effectively prevent inflammation in the intestine (50, 51, 52). Our data, in combination with the report by Clegg et al. (38) that CD40L transgenic mice with the highest transgene copy numbers acquire a lethal IBD marked by the infiltration of CD40L⁺ T cells and CD40⁺ cells into diseased tissues, suggest that CD40-CD40L interactions participate in the immunopathogenesis of human IBD.

Both B cells and monocytes/macrophages are known to express CD40, and among T cells mainly CD4⁺ T cells express CD40L (9, 10, 11, 12). In the present studies, double staining confirms that the majority of B cells and some of macrophages in IBD lesions are CD40⁺, while the majority of CD4⁺ T cells and some of CD8⁺ T cells are CD40L⁺ in inflamed mucosa of IBD. Our findings on CD40L expression by isolated LP-T cells, together with the immunohistochemical results, thus indicate that a subpopulation of both CD4⁺ and CD8⁺ T cells are activated to express CD40L in vivo and suggest that the LP-T cells continuously produce CD40L. Prolonged CD40L expression reflects persistent activation of Th cells in the IBD patients with active disease.

In the inflamed mucosa of IBD, CD4⁺ T cells and macrophages are the most abundant cells, and both cell types secrete cytokines that play an important role in cell recruitment, proliferation, and tissue damage. The ability of T cell contact-dependent signaling to induce monocyte/macrophage activation has been well documented. Ligation of CD40 on monocytes with CD40L on activated T cells is critical for the T cell-contact-dependent monocyte NO and cytokine production, i.e., IL-1β, IL-6, IL-8, TNF, and IL-12 (14, 15, 16). Both activated T cells and a Th1 clone derived from CD40L-deficient mice failed to induce IL-12 production from splenic APCs or peritoneal macrophages in vitro (17). CD40L-deficient mice show markedly impaired production of TNF, NO, IL-12, and IFN-γ, correlating with impaired cell-mediated immune responses against an intracellular parasite (e.g., Leishmania major) and enhanced susceptibility to infection (23, 24, 25, 26). In this study, freshly isolated LP-T cells from inflamed mucosa of IBD patients could induce high levels of TNF and IL-12 secretion by normal monocytes. Most importantly, mAbs that block the CD40-CD40L interactions also significantly reduced cytokine production by the monocytes. Taken together, our data not only show that CD40L is functionally expressed on LP-T cells from inflamed mucosa, but also imply that the cognate interactions between LP-T cells and macrophages via the CD40-CD40L costimulatory pathway may play an important role in pathogenic cytokine secretion (IL-12 and TNF) in IBD mucosa. Ligation by CD40L expressed by these activated T cells with CD40⁺ cells might also play an important role in B cell clonal expansion and proinflammatory cytokine production such as TNF and IL-12 (53, 54, 55). Moreover, numerous experiments have demonstrated that stimulation of monocytes through the CD40 receptor induces profound phenotypic changes. CD40L signaling induces APCs (e.g., B cells, macrophages, dendritic cells) to express CD80 (B7.1) and CD86 (B7.2) molecules, which interact with CD28 on activated T cells to amplify the response of the T cells (11, 15, 18). We have also confirmed this in cocultures of LP-T cells with normal monocytes (our unpublished observations). Such increased CD80 and CD86 expression in inflamed mucosa of IBD patients was previously demonstrated by immunohistochemical studies (43, 56). Moreover, CD40 signaling can costimulate T cells to secrete proinflammatory mediators (57). Increased expression of CD40L was also found on CD8⁺ T cells in inflamed mucosa of IBD, and CD40L⁺ CD8⁺ T cells, like CD40L⁺ CD4⁺ T cells, thus might also play an important part in immune responses by inducing B cell growth and differentiation and macrophage activation in a CD40L-dependent fashion (58).

The present data on CD40L expression by PB-T cells also indicate that the disease process in IBD is not limited to the involved gut tissue. PB-T cells, despite being CD40L negative, could also express CD40L more strongly after activation than the control T cells. Expression was also more prolonged, and similar results have been reported for SLE and RA (27, 28, 29). Increased expression of CD40L has been shown by freshly isolated T cells from peripheral blood and/or synovial fluid in RA (29). After in vitro activation with PMA/ionomycin, these T cells maintain a higher level of cell-surface expression of CD40L through 24 and 48 h of culture. On the basis of these results, we suggest that the inflammatory process in IBD is not restricted to the gut.

Some reports have shown that the tissue levels of TNF are higher in CD than in UC, and that anti-TNF treatment is more effective in CD (5, 51). Our results demonstrate that CD40L expression is similarly increased on LP-T cells from inflamed mucosa of both CD and UC patients, and we raise the possibility that this CD40L at least contributes to proinflammatory cytokine production. To explain the difference in TNF, one should take consideration that cytokines produced by T cells might have importantly modulatory effects on cytokine secretion by macrophages. Indeed, LP-CD4⁺ T cells from inflamed CD mucosa produce more IFN-γ, while LP-CD4⁺ T cells from inflamed UC mucosa secrete higher level of the Th2 cytokine IL-5 when stimulated in vitro via the CD2/CD28 pathway (4). While IFN-γ enhances TNF and IL-12 production, Th2 cytokines such as IL-4 and IL-10 have inhibitory effects. In vivo cytokine production by macrophages in IBD intestinal mucosa may thus be modulated by other local factors, in addition to CD40-CD40L interactions. Therefore, the exact role of CD40-CD40L interactions in IBD pathophysiology is still an unresolved question, and in vivo studies on the effect of blocking CD40-CD40L signaling seem warranted.

Over recent years, CD40-CD40L interactions have been shown to play an important role in many organ-specific T cell-mediated experimental autoimmune diseases such as experimental allergic encephalomyelitis (30), murine lupus (33), and collagen-induced arthritis (34). Treatment of mice with anti-CD40L mAb has been shown to effectively prevent development of disease in all of the above-mentioned animal models, with a reduction or elimination of damage to target tissues or infiltration of leukocytes into target tissues. In addition, this Ab has also been shown to effectively prevent colitis in a 2,4,6-trinitrobenzene sulfonic acid-induced animal model (37). Information obtained in this study sheds some light on the pathogenic mechanisms involved in human IBD. The persistence of CD40L expression on mucosal T cells will potentially initiate an inflammatory cascade, and CD40L itself is likely to be an important therapeutic target for treatment of IBD. Therefore, blocking of CD40-CD40L interactions with anti-CD40L mAb may have a beneficial therapeutic effect in the inflammatory cascade involved in IBD.

We thank Dr. R. Armitage (Immunex, Seattle, WA) for providing us with the anti-CD40L mAb M90 and M. Adé for expert technical assistance in performing the ELISA.