Abnormal T cell responses to bacterial superantigens in Behçet's disease (BD)

Abstract

This study examines the nature of T cell hypersensitivity in BD. Highly purified T cells from 32 BD patients, from 29 rheumatoid arthritis (RA) patients and from 14 healthy individuals were cultured with various concentrations of Staphylococcal enterotoxins (SE) B and C₁ in the presence of monocytes for 5 days, after which the production of interferon-gamma (IFN-γ) was assessed. High concentrations of SE (1 ng/ml) stimulated BD T cells as well as control T cells to produce comparably high amounts of IFN-γ, whereas low concentrations of SE (1 pg/ml) stimulated BD T cells much more effectively than normal or RA T cells. The hypersensitivity of BD T cells to low concentrations of SEC₁ was restored with RA monocytes instead of BD monocytes, whereas BD monocytes could not elicit the SEC₁-induced IFN-γ production of RA T cells. Moreover, there were no significant differences between BD T cells and RA T cells in monocyte-independent IFN-γ production stimulated with low or high concentrations of immobilized anti-CD3, or in the monocyte-mediated enhancement of IFN-γ production stimulated with a low concentration of immobilized anti-CD3. These results confirm that T cell hypersensitivity is not confined to certain specific antigens in BD. More importantly, the data strongly suggest that abnormalities in signal transduction triggered by perturbation of T cell receptors, but not in that induced by cross-linking of CD3 molecules nor in that delivered through costimulation molecules, play an important role in the pathogenesis of BD.

INTRODUCTION

BD is characterized by recurrent aphthous stomatitis, uveitis, genital ulcers and skin lesions, the aetiology and pathogenesis of which still remain uncertain [1,2]. In recent years, greater attention has been paid to the role of certain strains of streptococci as an aetiologic agent [3]. Thus, BD patients have a significantly higher incidence of tonsillitis and dental caries, in which certain strains of streptococci are involved [4]. In addition, systemic symptoms of BD could be induced after treatment of dental caries or even by intracutaneous injection of streptococcal antigens [3,4]. On the other hand, there have been several lines of evidence suggesting the involvement of abnormal T lymphocyte activation in the pathogenesis of BD, including the enhanced expression of HLA class II antigen in T cells infiltrating in mucosal ulceration [5] and the therapeutic efficacy of cyclosporin A (CsA), a strong T cell inhibitor, in uveitis attacks in BD [6]. These observations led to the hypothesis that T cells from BD patients respond specifically to certain strains of streptococci to produce certain cytokines that mediate the inflammatory responses in BD. However, we have also shown that BD T cells are stimulated by a variety of bacterial antigens, including streptococcal antigens, Escherichia coli-derived antigens, and Staphylococcus aureus, to produce IFN-γ [7,8]. In addition, recent studies have shown that lymphocytes from BD patients are stimulated with several epitopes of mycobacterial and homologous human 65-kD heat shock protein (hsp) peptides more effectively than those from control subjects [9]. It is therefore likely that T cell hypersensitivity in BD might not be confined to certain bacterial antigens and that intrinsic abnormalities in T cells that lead to activation upon suboptimal stimulation might play an important role in the pathogenesis of BD.

The current studies were therefore undertaken to explore the nature of such intrinsic abnormalities of BD T cells using superantigens such as Staphylococcal enterotoxins (SE) and immobilized anti-CD3 MoAb. The results indicate that T cells from patients with BD were much more effectively stimulated with low concentrations (1–10 pg/ml) of SEB and SEC₁ to produce higher amounts of IFN-γ than those from patients with rheumatoid arthritis (RA) or from normal healthy donors. More importantly, there was no significant difference in T cell IFN-γ production upon stimulation with low concentrations of immobilized anti-CD3 between BD patients and RA patients. The data therefore suggest that abnormalities in signal transduction triggered through T cell receptor (TCR) β-chains, but not in that induced by cross-linking of CD3 molecules, might play an important role in the T cell hypersensitivity in BD.

PATIENTS AND METHODS

Patients

Thirty-two patients with BD, who satisfied the diagnostic criteria of the international study group for BD [10], were studied (13 males and 19 females, aged 20–73 years). Of the 32 patients, 10 were taking colchicine (0.5–1.0 mg/day), eight were taking prednisolone (2.5–10.0 mg/day), one was taking methotrexate (7.5 mg/week) and two were taking azathioprine (25–50 mg/day) along with prednisolone (15–20 mg/day). Fourteen healthy volunteers (11 males and three females, aged 20–40 years), who were found to have no signs or symptoms of BD or any other disease, were studied as normal controls. In addition, 29 patients with RA, who satisfied the American College of Rheumatology 1987 revised criteria for the disease [11], were studied as disease controls (eight males and 21 females, aged 21–80 years). Of the 29 patients, four were receiving monthly injections of gold sodium thiomalate, four were taking methotrexate (7.5 mg/week), five were taking d-penicillamine (200–300 mg/day), four were taking salazosulfapyridine (1.0–2.0 g/day), two were taking bucillamine (100–200 mg/day), and four were taking bredinin (150 mg/day). Thirteen of the 29 patients were taking prednisolone (2.5–10.0 mg/day). All the above individuals gave informed consent.

Reagents

SEB and SEC₁ were purchased from Serva (Heidelberg, Germany). A murine IgG2a MoAb to the human CD3 complex, 64.1, was a generous gift of Dr P. E. Lipsky (University of Texas South-Western Medical Center, Dallas, TX).

Cell preparation

Peripheral blood mononuclear cells (PBMC) were obtained from patients with BD or RA patients or healthy adult volunteers by centrifugation of heparinized venous blood over sodium diatrizoate/Ficoll gradients. Monocytes were prepared from PBMC by adherence to glass Petri dishes as previously described [7]. The adherent cells dislodged from the glass surface were 80–90% monocytes as assessed by staining for non-specific esterase activity. The non-adherent cells were further depleted of monocytes and natural killer (NK) cells by incubation with 5 mml-leucine methyl ester HCl (Sigma Chemical Co., St Louis, MO) in serum-free RPMI 1640 (Gibco, Grand Island, NY) as described [12]. The treated cell population was washed twice and then incubated with neuraminidase-treated sheep erythrocytes. After centrifugation on diatrizoate/Ficoll gradients, the sedimented neuraminidase-treated sheep erythrocyte rosette-forming cells were treated with isotonic NH₄Cl to lyse the neuraminidase-treated sheep erythrocytes. The T cell populations thus obtained contained < 0.1% esterase-positive cells, < 1% NK cells, < 0.1% CD19⁺ cells and > 96% CD3⁺ cells.

Cell culture for induction of IFN-γ production

All cultures were carried out in RPMI 1640 medium supplemented with penicillin G 100 U/ml, streptomycin 10 μg/ml, l-glutamine 0.3 mg/ml, and 10% fetal bovine serum (FBS; Gibco). T cells (1 × 10⁵/well) were cultured with autologous monocytes (1 × 10⁴/well) in U-bottomed 96-well microtitre plates (no. 3799; Costar, Cambridge, MA) in a total volume of 200 μl in the presence or absence of various concentrations of SE. In some experiments, T cells (1 × 10⁵/well) were cultured alone in wells with various concentrations of immobilized anti-CD3 in flat-bottomed 96-well microtitre plates (no. 3596; Costar) [13]. The cells were incubated for 5 days at 37°C in a humidified atmosphere of 5% CO₂ and 95% air. Then the supernatants were harvested and kept frozen at −20°C until they were assayed for IFN-γ content.

Measurement of IFN-γ

IFN-γ contents in the supernatants were assayed using a solid-phase enzyme immunoassay, as previously described [7,8]. Briefly, wells of a 96-well microtitre plate (Cooke; Dynatech, Alexandria, VA) coated with rabbit anti-human IFN-γ (Hayashibara, Okayama, Japan) were incubated with cell-free culture supernatants or various concentrations of recombinant IFN-γ (Chemicon, El Segundo, CA) in PBS containing 1% bovine serum albumin (BSA; Miles, Elkhart, IN). Bound IFN-γ was detected with peroxidase-conjugated rabbit anti-human IFN-γ. The detection limit of the assay was ≈ 2.0 U/ml of IFN-γ. The assay was specific for natural and recombinant human IFN-γ.

Immunofluorescence staining of cell surface markers and analysis by flow cytometry

PBMC obtained from 14 patients with BD and 11 healthy volunteers, who were randomly selected, were analysed for expression of a variety of Vβ chains of TCR by flow cytometry. Briefly, PBMC were washed three times with PBS containing 2% normal human AB serum and 0.1% sodium azide (staining buffer). The cells were reacted in suspension by incubating for 30 min at 4°C with saturating concentrations of FITC-conjugated MoAb to human Vβ5, Vβ6.7, Vβ8 and Vβ12 families (Diversi-T αβ TCR Screening Panel 1A; T Cell Diagnostics Inc., Cambridge, MA) and PE-conjugated anti-CD3 MoAb (Leu-4; Becton Dickinson, Mountain View, CA). After the cells had been washed three times with staining buffer, they were fixed with 1% paraformaldehyde in PBS pH 7.4 for > 5 min at room temperature. The cells were analysed using an EPICS profile flow cytometer (Coulter, Hialeah, FL) equipped with an argon-ion laser at 488 nm. The flow cytometer was calibrated with Immuno-Check (Coulter) by a computer program. EPICS profile software was used to generate the plots. Gating on the forward and side scatter measurement was used to identify viable lymphocytes. The data generated by flow cytometry, based on 5000 lymphocytes defined by the scatter gates, were collected on dual-parameter scattergrams. In conjunction with logarithmic amplifiers, the scattergrams encompassed a four-decade range of fluorescence intensity. The percentages of cells staining positively for each MoAb were determined by integration of cells above a specified fluorescence channel, calculated in relation to isotype-matched control MoAb. The percentages of cells positively stained for Vβ5, Vβ6.7, Vβ8 and Vβ12 within the CD3⁺ T cell population were calculated.

RESULTS

Enhanced in vitro T cell IFN-γ production upon stimulation with low concentrations of staphylococcal enterotoxins in patients with BD

The initial experiments compared the capacity of BD T cells and control T cells to produce IFN-γ upon stimulation with a variety of concentrations of SEC₁. As can be seen in Fig. 1, a high concentration of SEC₁ (1 ng/ml) stimulated T cells of a representative BD patient as well as those of a representative healthy individual to produce comparable amounts of IFN-γ. By contrast, lower concentrations of SEC₁ (1–10 pg/ml) exclusively stimulated T cells from a BD patient. It was thus suggested that BD T cells might be hypersensitive to SEC₁ at a concentration so low that it does not stimulate T cells of individuals without BD.

Fig. 1.

T cell production of IFN-γ induced by various concentrations of Staphylococcal enterotoxin (SE) C₁. T cells (1 × 10⁵/well) from a representative patient with BD or from a representative normal individual were cultured with autologous monocytes (1 × 10⁴/well) in the presence of various concentrations of SEC₁. After 5 days of incubation, the supernatants were harvested and assayed for IFN-γ.

Next, experiments were therefore designed to address this question of the capacity of T cells from 13 BD patients, nine RA patients, and 14 healthy individuals to produce IFN-γ upon stimulation with low concentrations of SEB or SEC₁. As summarized in Fig. 2, T cell production of IFN-γ in the absence of SEB or SEC₁(presumably induced by autologous mixed lymphocyte reactions) was not significantly different among the three groups. However, T cell production of IFN-γ in the presence of a low concentration of SEB or SEC₁ (1 pg/ml) was significantly elevated in BD compared with that in RA or in healthy individuals. The results indicate that BD T cells are activated with suboptimal stimuli that do not significantly activate T cells in healthy or disease controls. Moreover, the data also confirm that the hypersensitivity of T cells is not confined to certain specific antigens in BD.

Fig. 2.

T cell production of IFN-γ induced by low concentrations of Staphylococcal enterotoxin (SE) B and C₁. T cells (1 × 10⁵/well) from 13 BD patients, nine rheumatoid arthritis (RA) patients, or 14 normal individuals were cultured with autologous monocytes (1 × 10⁴/well) in the presence or absence of SEB or SEC₁ (1 pg/ml). After 5 days of incubation, the supernatants were harvested and assayed for IFN-γ. Results are shown in box plots. Statistical analysis was done by Mann–Whitney U-test.

Abnormalities in T cells are involved in the hypersensitivity of T cells to low concentrations of SE in BD

T cells from patients with BD were stimulated with low concentrations of SEB or SEC₁ to produce IFN-γ much more effectively than those of healthy or disease controls. Stimulation of T cells with SE requires the presence of accessory cells such as monocytes [14]. It was thus uncertain which of T cells or monocytes might play a role in the hypersensitive responses to low concentrations of SE in BD. As can be seen in Table 1, however, elevated IFN-γ production of BD T cells stimulated with SEC₁ (10 pg/ml) was restored in the presence of RA monocytes instead of BD monocytes, whereas BD monocytes were not able to support SEC₁-stimulated IFN-γ production of RA T cells. The results therefore indicate that abnormalities of T cells, but not those in monocytes, are involved in the hypersensitive responses to low concentrations of SE in BD.

Table 1.

Hypersensitivity of T cells to low concentrations of Staphylococcal enterotoxin (SE) C₁in BD is due to abnormal T cell functions

Differential effects of low concentrations of SEC₁and immobilized anti-CD3 on T cell responses in BD

Recent studies indicate that there is physical dissociation of TCR and CD3 after stimulation through TCR ligation [15]. In order to delineate the pathway involved in the hypersensitive responses of BD T cells, experiments were then designed in which BD T cells and RA T cells were stimulated with low concentrations of SEC₁ in the presence of autologous monocytes or with various concentrations of immobilized anti-CD3 in the absence of monocytes. As shown in Fig. 3, BD T cells produced much higher amounts of IFN-γ than RA T cells upon stimulation with low concentrations of SEC₁, whereas both BD T cells and RA T cells responded comparably to any concentrations of immobilized anti-CD3. Figure 4 summarizes the IFN-γ production of T cells from 11 patients with BD and from nine patients with RA upon stimulation with high or low concentrations of immobilized anti-CD3 or with low concentrations of SEC₁. Again, there were no significant differences in IFN-γ production stimulated with low as well as high concentrations of immobilized anti-CD3 between BD T cells and RA T cells, whereas BD T cells produced much higher amounts of IFN-γ upon stimulation with SEC₁ than RA T cells. Collectively, the results strongly suggest that abnormalities in the signal transduction triggered by perturbation of β-chains of TCR, but not in that induced by cross-linking of the CD3 molecules, may be involved in the hypersensitivity of T cells in BD.

Fig. 3.

Differential effects of Staphylococcal enterotoxin (SE) C₁ and immobilized anti-CD3 on the IFN-γ production of T cells. T cells (1 × 10⁵/well) from BD patients or rheumatoid arthritis (RA) patients were cultured alone with various concentrations of immobilized anti-CD3 or cultured with autologous monocytes (1 × 10⁴/well) and SEC₁. After 5 days of incubation, the supernatants were harvested and assayed for IFN-γ.

Fig. 4.

Differential effects of Staphylococcal enterotoxin (SE) C₁ and immobilized anti-CD3 on IFN-γ production of T cells. T cells (1 × 10⁵/well) from 11 BD patients or nine rheumatoid arthritis (RA) patients were cultured alone with high or low concentrations of immobilized anti-CD3 or cultured with autologous monocytes and SEC₁ (10 pg/ml). After 5 days of incubation, the supernatants were harvested and assayed for IFN-γ. Results are shown in box plots. Statistical analysis was done by Mann–Whitney U-test.

It was also possible that abnormalities in the signal transduction through costimulation molecules mediated by T cell–monocyte interactions might be responsible for the hypersensitive T cell responses to low concentrations of SE in BD. Final experiments were carried out to address this point, in which BD T cells and RA T cells were stimulated with a low concentration of anti-CD3 in the presence or absence of autologous monocytes. Figure 5 summarizes IFN-γ production of T cells from nine patients with BD and nine patients with RA upon stimulation with a low concentration of immobilized anti-CD3 in the presence or absence of autologous monocytes. Irrespective of the presence of autologous monocytes, there were no significant differences in IFN-γ production between BD T cells and RA T cells. Thus, the enhancement of IFN-γ production of T cells by autologous monocytes in BD patients was comparable to that in RA patients (stimulation index 2.66 ± 1.13 (mean ± s.d.) and 3.36 ± 2.14, respectively). The results indicate that costimulatory signals provided by accessory cells in BD are comparable to those in RA, ruling out the possibility that abnormalities in costimulatory signals might be responsible for the hypersensitive responses of BD T cells. The data therefore confirm that abnormal signal transduction through β-chains of TCR is involved in the hypersensitivity of T cells in BD and therefore in the pathogenesis of BD.

Fig. 5.

Comparable effects of monocytes on IFN-γ production of T cells between BD and rheumatoid arthritis (RA). T cells (1 × 10⁵/well) from nine BD patients or nine RA patients were cultured with or without autologous monocytes (1 × 10⁴/well) in the presence of a low concentration of immobilized anti-CD3 (0.125 μg/ml). After 5 days of incubation, the supernatants were harvested and assayed for IFN-γ. (a) IFN-γ production shown in box plots. (b) Stimulation index (SI) as calculated by IFN-γ production in the presence of monocytes/IFN-γ production in the absence of monocytes. Statistical analysis was done by Mann–Whitney U-test.

Expression of TCR Vβ chains on unstimulated T cells from patients with BD and normal healthy individuals

It has been demonstrated that Staphylococcal superantigens show selectivity for various Vβ specificities. Thus, previous studies disclosed that SEB and SEC₁ show a major predilection for Vβ5 and Vβ12 [14]. It was therefore plausible that the hypersensitivity of BD T cells for SEB and SEC₁might be due to an over-representation of these Vβ-specific T cells within the PBMC of BD. Experiments were therefore carried out to address this point. As can be seen in Fig. 6, there were no significant differences in the percentages of T cells expressing Vβ5, Vβ6.7, Vβ8 and Vβ12 between BD patients and normal healthy individuals. The data thus remove the possibility that abnormal T cell responses in BD might be a result of biased expression of certain Vβ specificities.

Fig. 6.

Expression of T cell receptor Vβ chains on unstimulated T cells. Peripheral blood mononuclear cells (PBMC) from 14 BD patients (BD) and 10 normal healthy individuals (N) were stained with FITC-conjugated MoAb to human Vβ5, Vβ6.7, Vβ8 and Vβ12 and PE-conjugated anti-CD3. Immunofluorescent staining fixation, and analysis were performed as described under Patients and Methods. The percentages of cells positively stained for Vβ5, Vβ6.7, Vβ8 and Vβ12 within the CD3⁺ T cell population were calculated. Statistical analysis was done by Student's t-test.

DISCUSSION

We have previously disclosed that BD T cells are hypersensitive to various bacterial antigens, including Streptococcus sanguis-related antigens, E. coli- derived antigens, and Staph. aureus Cowan I, to produce higher amounts of IFN-γ and IL-6, compared with T cells from healthy or disease controls [7,8]. It has been recently shown that BD T cells are hypersensitive to mycobacterial and homologous human 65-kD hsp peptides [9]. The present study has extended the investigation of T cell hypersensitivity in BD using Staphylococcal superantigens SEB and SEC₁, which can stimulate a broader spectrum of T cell repertoires than nominal antigens [16]. It should be noted that at their optimal concentrations SE strongly stimulate T cells from normal healthy individuals [14,16]. The results in the present study, however, have demonstrated that T cells from BD patients were stimulated to produce IFN-γ even with very low concentrations of SEB as well as SEC₁, that were not able to stimulate T cells from normal or disease controls. Moreover, it has also been revealed that BD T cells do not show over-representation of Vβ5 and Vβ12 specificities, which have been shown to be preferentially stimulated with SEB and SEC₁ [14]. Our data thus confirm that the hypersensitivity of T cells from patients with BD is not confined to certain specific antigens, but rather suggest that T cell hypersensitivity might be due to intrinsic abnormalities in a broad spectrum of T cells in BD. More importantly, the concentrations of SE used in the present study were so small as to be achieved in physiological conditions. Thus, the hypersensitivity of BD T cells to various bacterial antigens, including superantigens, may well account for the characteristic clinical features of BD, including the induction of systemic symptoms after infections, treatment of dental caries, operation, or trauma [1–3].

The current studies include three major experiments: one for the comparison of the responses of T cells from BD patients, RA patients, and healthy donors to low concentrations of SEB and SEC₁, another for comparison of the responses of T cells from BD patients and RA patients to SEC₁ and immobilized anti-CD3, and the other for comparison of the effects of monocytes on T cell IFN-γ production in cultures stimulated with a low concentration of immobilized anti-CD3. It should be pointed out that the BD and RA patients used in each experiment were randomly selected on an availability basis without any bias, although three BD patients happened to be used in two of the three experiments. Moreover, there was no significant difference in the use of steroid, immunosuppressant, anti-rheumatic drugs, or colchicine for BD patients and RA patients among the three experiments. Therefore, the conclusion that BD T cells are hypersensitive to low concentrations of SEC₁ has been repeatedly verified by two of the three independent experiments in the present study. It should also be mentioned that T cells from two BD patients who were taking azathioprine and prednisolone (15–20 mg/day) responded to low concentrations of SEC₁ and SEB as effectively as those from BD patients without prednisolone and cytotoxic drugs (data not specified). Collectively, it is suggested that the medication of the patients might have no serious influence on the results in the current studies.

Previous studies have also suggested that Strep. sanguis-related antigens (KTH-1 antigens), to which T cells from patients with BD are hypersensitive, have characteristics of superantigens [7]. Thus, the hypersensitivity of T cells from BD patients required the binding of the KTH-1 antigens to monocytes, but not necessarily processing of the antigens by monocytes [7]. Of note, lipopolysaccharide (LPS) could not elicit IFN-γ production of BD T cells, although it markedly stimulated IL-6 production of BD monocytes, indicating that activation of monocytes alone is not sufficient for stimulation of the cytokine production of BD T cells [8]. It is also possible that BD monocytes have prominent capacities to bind and/or present superantigens compared with those from normal or disease controls. However, the hypersensitivity of BD T cells to SEC₁ was restored with the presence of RA monocytes, whereas BD monocytes were not able to elicit the hypersensitivity of RA T cells to SEC₁. The results therefore indicate that T cell hypersensitivity to the superantigens in BD is a result of intrinsic abnormalities in T cells. Moreover, the data suggest that T cell hypersensitivity in BD is directed to such antigens that can directly interact with T cells.

There were no significant differences in responses to either low or high concentrations of immobilized anti-CD3 in the absence of monocytes between BD T cells and RA T cells. In addition, there were no significant differences in the enhancement of IFN-γ production by monocytes between BD and RA. It is therefore unlikely that abnormalities in costimulatory signals might be responsible for the hypersensitive responses of BD T cells to low concentrations of SE. Of note, a number of recent reports suggest that the signal transduction pathway triggered by TCR might be different from that induced by CD3 [15,17]. Thus, pretreatment with anti-CD4 MoAb has been shown to inhibit anti-TCR MoAb-mediated activation but fail to inhibit activation with anti-CD3 MoAb [18]. Moreover, association of the tyrosine kinase p56^lck with CD4 inhibited stimulation of T cell clone with anti-TCR MoAb, but did not alter the effect of anti-CD3 MoAb treatment [19]. Finally, it has been recently shown that wortmannin, a phosphatidylinositol 3-kinase inhibitor, blocks antigen-mediated, but not anti-CD3-induced, activation of murine CD4⁺ T cells, even in the absence of influences of costimulation molecules B7-1 and B7-2 [20]. Collectively, our data suggest that BD T cells may have defects that lead to abnormalities in signal transduction pathway triggered through TCR by superantigens, but not in that triggered through CD3 molecules or through costimulation molecules. Although the current study has characterized BD as that with abnormal signal transduction through TCR, further studies designed to explore such abnormalities in the complexity of molecular associations involved in TCR signalling would be important to elucidate completely the nature of abnormal T cell functions in BD.