Immunoregulatory defects of Vα24⁺Vβ11⁺ NKT cells in development of Wegener's granulomatosis and relapsing polychondritis

Abstract

The frequency of either CD4^–8^– (double negative; DN) or CD4⁺ Vα24⁺Vβ11⁺ NKT cells, the expression of CD1d and the binding of CD1d-tetramer loaded with α-galactosylceramide (α-GalCer) to NKT cells were analysed in peripheral blood mononuclear cells (PBMCs) of patients with Wegener's granulomatosis (WG), relapsing polychondritis (RP) and healthy subjects (HS). DN and CD4⁺ Vα24⁺Vβ11⁺ NKT cells as well as CD1d-α-GalCer tetramer-positive NKT cells, were significantly decreased in number in both WG and RP patients compared to those from HS. When cytokine profiles were analysed in these PBMCs upon stimulation with phorbol ester and calcium ionophore, CD4⁺ T cells from patients with WG and RP exhibited a Th1 bias, whereas CD4⁺ NKT cells from WG patients in remission showed a Th2 bias. These findings suggest that NKT cells (especially CD4⁺ NKT cells) play a regulatory role in Th1 autoimmunity in patients with WG and RP. The reduction in NKT cell counts appears to be associated with the low responsiveness to α-GalCer. The dysfunction of NKT cells to recognize ligands such as α-GalCer may also contribute to the defects observed in NKT cells from WG and RP patients.

INTRODUCTION

Wegener's granulomatosis (WG) is a systemic vasculitic disease characterized by necrotizing granulomas and vasculitis of arterioles and venules. The majority of WG patients develop lesions in the nasal or paranasal sinuses [1] and may also manifest lung and kidney involvement during the course of the disease. The presence of cytoplasmic pattern antineutrophil cytoplasmic antibodies (C-ANCA) in the serum of WG patients was first reported by van der Woude et al. [2] and is now recognized to be highly specific for the disease, especially in the active phase. The cause of this disease remains unknown, although the major target antigen of ANCA in WG is identified as proteinase 3 [3]. Studies of T cells from peripheral blood mononuclear cells (PBMCs) and the granulomatous lesions of WG patients showed that the T cells were skewed toward a Th1 type [4,5]. However, again the mechanism underlying this T cell deviation is unclear.

Relapsing polychondritis (RP) is a rare inflammatory disease that leads systemically and predominantly to destruction of the cartilaginous tissue [6]. Tracheolaryngeal symptoms develop later in the course of disease that lead to coughing, dysphonia and, in some cases, fatal respiratory distress. RP may also cause inner ear involvement, ocular inflammation and vasculitis. It has been suggested that a certain autoimmunity is involved in RP and several antibodies to primary target antigen candidates such as type II collagen [7] and matrilin 1 [8,9] have been reported. However, it remains elusive whether these autoantibodies to type II collagen and/or matrilin 1 are related directly to the development of RP. No reports have shown T cell involvement in RP.

Natural killer T cells (NKT cells) represent a novel lymphocyte lineage distinct from conventional T, B or NK cells. The majority of human NKT cells express a single invariant Vα24JαQ chain paired preferentially with Vβ11 [10,11]. Although the natural antigens for NKT cells have not been identified, a glycolipid, α-galactosylceramide (α-GalCer), has been shown to act as a ligand for NKT cells [12–14]. Antigens are presented to NKT cells by an MHC-like molecule, CD1d, on antigen-presenting cells (APCs) [15]. NKT cells are composed mainly of CD4⁺ and CD4^–8^– double negative (DN) populations, and the CD4⁺ NKT population produces Th2 cytokines [such as interleukin (IL)-4] in healthy subjects (HS) [16–18]. It has been postulated that NKT cells are involved in the regulation of various autoimmune disease models such as non-obese diabetic (NOD) mouse and experimental autoimmune encephalomyelitis (EAE) [19–22]. In general human autoimmune diseases, including systemic sclerosis (SSc), systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), insulin-dependent diabetes mellitus (IDDM) and multiple sclerosis (MS), selective decreases in NKT cell number and function have been reported [23–27].

The present study was designed primarily to examine the frequency of CD4⁺ and DN NKT cells in patients with WG and RP by means of a newly established detection system using a monoclonal antibody (MoAb) against Vα24 and an α-GalCer-loaded mouse CD1d-tetramer (α-GalCer-tetramer) [28,29]. Furthermore, we analysed the responsiveness of NKT cells to α-GalCer and the production of Th1/Th2 cytokine by PBMC, CD4⁺ T cells and CD4⁺ NKT cells. We show herein that the proportions of CD4⁺ and DN Vα24⁺Vβ11⁺ NKT or α-GalCer-tetramer⁺ NKT cells are reduced in patients with WG and RP compared with HS. CD4⁺ T cells from these patients exhibited a Th1 bias, while CD4⁺ NKT cells from WG patients in remission showed a Th2 bias. The functional status of NKT cells in WG and RP is discussed in relation to other autoimmune diseases.

MATERIALS AND METHODS

Patients and HS

Twelve WG patients (mean ± s.d., age 60·7 ± 11·8) diagnosed according to the American College of Rheumatology criteria and the Chapel Hill definition for the classification of WG [30,31], and 10 RP patients (age 49·8 ± 12·8) diagnosed according to the criteria defined by Michet et al. [6] were examined. Four WG patients and five RP patients had active inflammatory disease. Eleven of the 12 WG patients showed positive titres of PR-3 ANCA or C-ANCA at the time of diagnosis. The clinical findings of the patients are summarized in Tables 1 and 2. Twelve HS (age 45·7 ± 12·8) served as controls. All patients and HS provided written informed consent.

Table 1.

Clinical findings of the WG patients

Case	Sex/age	Disease duration (years)	Disease activity	Disease manifestations	Treatment (mg)*	Duration of treatment (years)	PR3-ANCA titre	BVAS/WG
1	61/F	0·2	Active	E	P (10)	0·2	81	3
2	65/M	12	Remission	E + L	P (10)	10	–	0
3	69/M	0·1	Remission	E	P (5)	0·1	52	0
4	80/M	10	Active	E + K	–	9	300	5
5	47/M	16	Remission	E + K	P (5)	10	55	0
6	57/F	1	Active	E + L	P (15), AZP (25)	0·5	C +	4
7	62/M	22	Remission	E + L	P (5)	18	–	0
8	72/F	3	Remission	L	P (8)	2	–	0
9	46/F	10	Remission	E + L + K	P (12·5)	9	–	0
10	56/M	10	Remission	E + K	P (15)	8	–	0
11	72/F	18	Remission	E	P (7)	15	–	0
12	41/M	8	Active	E + L	P (20), CYC (50)	0·5	15	5

E, upper respiratory tract; L, lung; K, kidney; P, prednisolone; AZP, azathioprine; CYC, cyclophosphamide; C +, C-ANCA positive; BVAS/WG, Birmingham Vasculitis Score for Wegener's granulomatosis.

^*Doses of treatment at the time of sampling.

Table 2.

Clinical findings of the RP patients

Case	Sex/age	Disease duration (years)	Disease activity	Disease manifestations	Treatment (mg)	Duration of treatment (years)	ESR (mm/ h)
1	47/F	2	Remission	au + n + ey + ar + ie	P (10), AZP (–)	0·5	8
2	66/F	7	Remission	n + ey + ie	P (11)	6	9
3	33/M	1	Remission	au + ey + ar	–	–	26
4	48/F	0·1	Active	au + ey + n + tr + ie	P (17)	0	92
5	43/F	4	Remission	au + ey + ie	P (7)	4	8
6	61/F	4	Remission	au + ey + ar	–	4	8
7	39/F	1	Active	n + ar + tr	P (9)	0·7	13
8	55/M	2	Active	au + ar + ie + tr	P (20)	1	34
9	36/M	0·3	Active	au + ey + n	P (10), CYC (25)	0	57
10	70/M	0·3	Active	au + ey + ie	–	–	120

ESR = erythrocyte sedimentation rate; au = auricle; n = nose; ey = eye; ar = arthritis; tr = trachea, i.e. = inner ear; P = prednisolone; AZP = azathioprine; CYC = cyclophosphamide.

Flow cytometric analysis

Fresh PBMCs were isolated from heparinized blood by Ficoll–Conray separation and then stained with a combination of the following MoAb conjugates: fluorescein isothiocyanate (FITC)-anti-TCR Vα24 (C15, mouse IgG1), -anti-interferon (IFN)-γ (45·15, mouse IgG1) and -mouse IgG1 MoAb (isotype control: 679·1Mc7); phycoerythrin (PE)-anti-TCR Vβ11 (C21, mouse IgG2), -anti-IL-4 (4D9, mouse IgG1), -mouse IgG1 (679·1Mc7) and -mouse IgG2 MoAb (isotype control: U7·27); allophycocyanin (APC)-anti-CD3 (UCHT1, mouse IgG1), and -mouse IgG1 MoAb (679·1Mc7); biotinylated (biotin)-anti-CD4 (13B8·2, mouse IgG1), -anti-CD8 (B9·11, mouse IgG1), -mouse IgG1 MoAb (isotype control: MOPC21) (all from Immunotech, Marseille, France). IntraPrep™ permeabilization reagent was also obtained from Immunotech. Streptavidin–peridinin chlorophyll (PerCP), purified-anti-CD1d (CD1d42, mouse IgG1), PE-anti-CD14 (mouse IgG2), biotin-anti-CD19 (HIB19, mouse IgG1), APC-anti-CD11c (B-ly6, mouse IgG1) and -anti-IFN-γ MoAb (B27, mouse IgG1) were purchased from Pharmingen (San Diego, CA, USA). FITC-conjugated Zenon™ one mouse IgG1-labelling kits were purchased from Molecular Probes (Eugene, OR, USA). Cytometric bead array (CBA) kit was purchased from Becton Dickinson (San Jose, CA, USA). Mouse CD1d-α-GalCer tetramers were prepared as described previously [29]. Biotin CD1d-BSP was incubated with α-GalCer in 0·5% Tween 20/H₂O overnight at room temperature and then at 37°C for a further 2 h. Streptavidin–APC (Molecular Probes) was added to loaded CD1d-BSP and incubated at 4°C in the dark for 2 h followed by incubation at room temperature for 3 h. PBMCs (1 × 10⁶) were incubated with FITC-anti-Vα24, PE-anti-Vβ11, and then incubated with 1 µl of CD1d-α-GalCer tetramers for 20 min on ice. Lymphocytes positive for CD3, Vα24 and Vβ11 or positive for Vα24 and CD1d-α-GalCer tetramer were determined to be NKT cells.

Detection of intracellular cytokines of CD4⁺ T cells and CD4⁺ NKT cells

PBMCs (5 × 10⁵ cells/ml) were suspended in 24-well culture plates with RPMI-1640 medium (Sigma, St Louis, MO, USA) supplemented with 10% heat inactivated fetal calf serum (FCS), 0·05 mm 2-mercaptoethanol, penicillin (100 U/ml), and streptomycin (100 µg/ml). Cells were stimulated with 50 ng/ml phorbol myristate acetate (PMA: Sigma) and 500 ng/ml A23187 (Calbiochem, La Jolla, CA, USA) for 4 h at 37°C in 5% CO₂ in the presence of monensin (2 µm) [32]. For immunostaining of cytoplasmic IL-4 and IFN-γ, cells were washed twice in phosphate buffered saline (PBS) and then stained with FITC-anti-Vα24, biotin-anti-CD4 and streptavidin–Per-CP. After staining of surface markers, cells were fixed with IntraPrep™ permeabilization reagent 1 containing 5·5% formaldehyde for 15 min on ice. After washing in PBS, cells were resuspended with permeabilization reagent, PBS-buffered saponin-based permeabilizing and lysing medium. After 5 min, APC-anti-IFN-γ and PE-anti-IL-4 MoAb were added and incubated for 15 min on ice. Because of the low frequency of NKT cells, at least 300 000 cells were acquired from each sample for analysis.

Cytokine production

PBMCs (5 × 10⁵ cells/ml) were stimulated with 50 ng/ml PMA and 500 ng/ml A23187 for 24 h at 37°C. After stimulation, culture supernatants were collected and frozen at −80°C. A panel of six human cytokines (IFN-γ, tumour necrosis factor (TNF)-α, IL-2, IL-4, IL-5 and IL-10) were measured simultaneously in a single sample (50 µl) using the CBA kit according to the manufacturer's protocol. Cytokine ratios were calculated to determine the Th 2 versus Th 1 balance of the cytokines.

Expansion of NKT cells

PBMCs (1 × 10⁶ cells/ml) from patients or HS were cultured in 24-well plates in the presence of 100 ng/ml α-GalCer and 50 units/ml recombinant human IL-2 (rhIL-2) (provided by Pharmaceutical Research Division, Takeda Chemical Industries, Osaka, Japan), with 0·1% dimethylsulphoxide (DMSO) used as a control. α-GalCer was synthesized and provided by the Pharmaceutical Research Laboratories, Kirin Brewery (Tokyo, Japan). The stained cells were analysed by FACScalibur (Becton-Dickinson, Mountain View, CA, USA) with cellquest software.

Statistical analyses

Data were analysed by the Mann–Whitney U-test. For comparison of α-GalCer responder individuals and non-responder individuals, Fisher's exact probability test was used. P < 0·05 was considered significant.

RESULTS

Reduction of DN and CD4⁺ NKT cell numbers in PBMCs of patients with WG and RP

The frequency of Vα24⁺Vβ11⁺ T cells that represent the major NKT cell population was examined in the CD4⁺ or DN T cell population, and in total PBMCs from WG and RP patients and HS. At least 100 000 cells were acquired from each sample for analysis because NKT cells constitute a rare subset in PBMC. As shown in the representative FACS profile (Fig. 1a), the proportion of Vα24⁺Vβ11⁺ cells was decreased in a WG patient compared to that in a HS. The mean percentage ± s.e.m. of NKT cells in the DN or CD4⁺ T cell population was 0·427 ± 0·104% or 0·031 ± 0·007%, respectively, in HS (Fig. 1b). The mean percentage of NKT cells in DN or CD4⁺ T cells of WG patients (0·07 ± 0·021% or 0·007 ± 0·003, respectively) was significantly lower than that in HS. When the same analyses were performed with RP patients, the frequencies 0·22 ± 0·11% NKT cells in DN T cells; and 0·005 ± 0·001% NKT cells in CD4⁺ T cells were also significantly lower compared to those in HS. The significant reductions of both DN and CD4⁺ NKT cell frequencies were again demonstrated clearly in total lymphocytes of WG and RP patients (Fig. 1c). We then calculated the number of NKT cells per 10⁵ cells gated as lymphocytes and found that both DN and CD4⁺ NKT cell numbers were significantly lower in WG and RP patient groups compared to those in HS (data not shown). Vα24⁺Vβ11⁺ T cells do not necessarily represent cells that respond to α-GalCer in the contex of CD1d. Therefore, we analysed frequencies of CD1d-α-GalCer tetramer-positive cells in PBMCs from HS, RP and WG groups. As shown in Fig. 2a, a small proportion of tetramer and Vα24 double positive cells were detected in PBMCs from both HS and WG patients. When the mean frequency of Vα24⁺/α-GalCer-tetramer⁺ cells was compared among the three groups, these NKT fractions were found to be significantly decreased in WG and RP patients compared with frequencies in HS (P < 0·01 and P < 0·05, respectively; Fig. 2b). These findings are basically consistent with those obtained with Vα24⁺Vβ11⁺ cells (Fig. 1).

Fig. 1.

Flow cytometric analysis of NKT cells from HS, WG and RP patients. (a) Representative FACS profiles of CD3⁺, DN, Vα24⁺, Vβ11⁺ NKT cells from a HS and a WG patient are shown. Proportions of Vα24⁺ Vβ11⁺ NKT cells are indicated. (b) Proportion of NKT cells in CD3⁺ DN T cell population (upper) and in CD3⁺ CD4⁺ T cell population (bottom). (c) Mean proportion of DN NKT cells (upper) or CD4⁺ NKT cells (bottom) in total lymphocyte-gated cells. *P < 0·05, **P < 0·01, ***P < 0·001 versus HS.

Fig. 2.

Identification of NKT cells by α-GalCer-loaded CD1d tetramer (α-GalCer-tetramer) staining. PBMCs from HS, patients with WG and RP were stained with APC conjugated with either the CD1d-α-GalCer-tetramer or the unloaded CD1d tetramer in combination with the FITC conjugated anti-Vα24 MoAb. Proportions of Vα24⁺α-GalCer-tetramer⁺ cells (boxed) are indicated. (a) Representative FACS profile of PBMC stained with anti-Vα24 and α-GalCer-tetramer in HS and WG. (b) Mean proportions of Vα24⁺α-GalCer-tetramer⁺ cells in total lymphocytes from HS (n = 7), WG (n = 8) and RP (n = 7). The data represent mean ± s.e.m. *P < 0·05, **P < 0·01 versus HS.

IFN-γ /IL-4 balance in CD4⁺ T cells and CD4⁺ NKT cells from WG and RP patients

Next, the ability of T and NKT cells to produce IFN-γ and IL-4 was analysed using fresh PBMCs from HS, WG and RP patients. PBMCs were stimulated with PMA and A23187, and the levels of intracellular IFN-γ and IL-4 were determined. A representative result of the flow cytometric analysis of CD4⁺ T cells from a HS and a RP patient is shown in Fig. 3a. A large proportion of IFN-γ⁺ CD4⁺ T cells was noted in the RP patient (16·8%) compared to that in the HS (8·7%). Almost the same proportions of IL-4⁺ CD4⁺ T cells were seen in the RP patient and the HS patient (2·0% and 2·4%, respectively). The ratio of IFN-γ producing cells (%)/IL-4 producing cells (%) in the CD4⁺ T cell population was calculated in three groups (HS, RP and WG) and is presented in Fig. 3b. The mean IFN-γ⁺/IL-4⁺ ratios in CD4⁺ T cells from patients with WG and RP were significantly higher than the ratio in HS (P < 0·01 and P < 0·05, respectively; Fig. 3b).

Fig. 3.

Intracellular staining of IFN-γ and IL-4 in CD4⁺ T cells and CD4⁺ NKT cells from HS, RP and WG patients. (a) IFN-γ positive and/or IL-4 positive cells in CD3⁺ CD4⁺ T cells from HS and RP. Proportions of IFN-γ⁺ cells and IL-4⁺ cells are indicated. (b) Mean ratio of IFN-γ⁺/IL-4⁺ cells in the CD4⁺ T cell population. The data represent mean ± s.e.m. *P < 0·05, **P < 0·01 versus HS. (c) IFN-γ and/or IL-4 positive cells in CD4⁺ Vα24⁺ NKT cells. A representative FACS profile of HS. Proportions in each quadrant are indicated. (d) Ratio of IFN-γ and IL-4 producing cells in CD4⁺ NKT cells from WG patients in remission (n = 5), or active phase (n = 4), RP patients in either remission (n = 4) or active phase (n = 4) and HS (n = 9). *P < 0·05 versus HS.

Then, CD4⁺ Vα24⁺ cells (largely NKT cells) were analysed for production of IL-4 and IFN-γ. For this analysis 300 000–1000 000 cells were acquired from each sample and analysed using four-colour staining. A representative profile of IL-4 and IFN-γ expression in CD4⁺ Vα24⁺ cells from a HS is shown in Fig. 3c. A similar proportion of IL-4⁺ or IFN-γ⁺ cells was detected. Figure 3d shows ratios of IFN-γ⁺ (%) to IL-4⁺ (%) cells in CD4⁺ NKT populations from HS, RP and WG groups. RP and WG patients were divided into remission and active disease groups on the basis of clinical examination. Interestingly, the mean IFN-γ⁺/IL-4⁺ ratio in CD4⁺ NKT cells from WG patients in remission was significantly lower compared with that of HS (P < 0·05; Fig. 3d). The mean ratio in RP remission patients also showed a slight reduction compared to that in HS. On the other hand, the mean IFN-γ⁺/IL-4⁺ ratio in active phase RP or WG patients was higher than that in corresponding remission phase patients. These results suggest that CD4⁺ NKT cells are Th2 biased in patients with WG, and probably RP, in remission. In contrast, it seems that the CD4⁺ NKT cells from these patients in the active phase are biased towards Th1, although the differences were not statistically significant.

Cytokine production by PBMCs from WG and RP patients

In the next set of experiments, we examined production of various cytokines by stimulated PBMCs from HS, WG and RP patients. Whole PBMCs were stimulated with PMA/A23187, and 24 h later supernatant cytokine levels were evaluated by flow cytometry with CBA. As shown in Fig. 4a, significantly higher IFN-γ and TNF-α levels were detected in PBMC culture supernatants from WG patients than those of HS (P = 0·01 and P = 0·016, respectively). Similar, but not significant, increases in the levels of these cytokines were seen in PBMC cultures from RP patients. No differences in the concentration of IL-4 (Fig. 4a) or other cytokines (IL-2, -5, -10) (data not shown) were observed between these patients and controls.

Fig. 4.

Production of various cytokines. (a) PBMCs were cultured with PMA and A23187 for 24 h and the concentrations of cytokines produced were measured in PBMCs from HS (n = 10) and patients with WG (n = 8) and RP (n = 8). Each column and bar represents mean ± s.e.m. *P < 0·05, **P < 0·01 versus HS. (b) Ratios of various cytokine/IL-10 concentrations in HS and patients with WG and RP. (c) Ratios of various cytokine/IFN-γ concentrations in HS and patients with WG and RP. *P < 0·05, **P < 0·01 versus HS.

It has generally been considered that the relative balance between various cytokines is more important than the absolute concentrations of the cytokines in an individual for understanding the pathophysiological status [33]. We utilized the results of six cytokine concentrations quantified with a new flow cytometry-based method to calculate the mean ratios between opposing sets of cytokines {[(concentrations of various cytokines (pg/ml)]/[concentration of IL-10 (pg/ml)] and [concentrations of various cytokines (pg/ml)]/[concentration of IFN-γ (pg/ml)]}[33]. Figure 4b shows that the mean IFN-γ/IL-10 ratio in WG patients was significantly higher than that in HS. The TNF-α/IL-10 ratio also appeared to be higher in WG. On the other hand, the IL-2, IL-4, IL-5 and IL-10/IFN-γ ratios in WG patients were significantly lower compared to the respective values in HS (Fig. 4c). These results suggest that the cytokine profiles in WG patients are Th1-predominant as reported in previous studies [4,5]. On the other hand, only a slight increase in the IFN-γ/IL-10 ratio was noted in RP patients. No differences were seen in the ratios of the other cytokines versus either IL-10 or IFN-γ in RP patients (Figs 4b.c).

Expression of CD1d molecules on APCs

NKT cells are positively selected by CD1d molecules on haematopoietic cells [34,35]. We examined the possibility that the significant reduction in the NKT cell population in WG and RP patients was related to altered CD1d expression. Thus, the proportion of CD1d-expressing B cells (CD19⁺ CD11c^–) or monocytes (CD14⁺ CD11c⁺) in PBMCs from either HS or patients was analysed and compared. The mean fluorescence intensity (MFI) of CD1d expression was also quantified, and the MFI ratio (MFI of CD1d/MFI of a mouse IgG1 isotype-control) was calculated. When the mean proportion of CD1d⁺ cells and the mean MFI ratio in the B cell population were compared among HS, RP and WG groups, no significant differences were seen (data not shown). On the other hand, the mean percentage of CD1d⁺ cells in the monocyte population was significantly lower in RP patients than in HS (86·1 ± 4·5%versus 95·8 ± 1·8%, respectively, P < 0·05). The MFI ratio (MFI of CD1d/MFI of an isotype control) in RP patients (2·68 ± 0·25%) was also slightly lower than that in HS (3·24 ± 0·26%) (P = 0·14). On the other hand, no differences in both the proportion of CD1d positive cells and the MFI ratio were seen in the monocyte population of WG patients compared to those of HS (data not shown).

Reactivity of NKT cells to α-GalCer

To assess whether the reduction of NKT cell numbers in WG and RP patients is associated with the responsiveness of NKT cells, we analysed the reactivity of NKT cells to α-GalCer. PBMCs from patients with WG (n = 10), RP (n = 9) or from HS (n = 8) were co-cultured with either α-GalCer or DMSO, a control vehicle, in the presence of rhIL-2 for 10 days. Figure 5a illustrates a representative flow cytometric analysis of HS PBMCs before and after stimulation with α-GalCer. Upon stimulation with α-GalCer Vα24⁺ Vβ11⁺ NKT cells showed considerable proliferation and formed a distinct population (0·4–10·25%) at day 10. Figure 5b shows the number of DN NKT cells per fraction (10⁵ lymphocytes) in eight HS before culture (pretreatment) and after culture in the presence of α-GalCer or vehicle alone. As shown in this figure, in the presence of α-GalCer and rhIL-2 DN NKT cells expanded from 3·3–43 to 245–3679 per 10⁵ lymphocytes in six of eight HS (13·5–427·5-fold increase). Two HS (nos 2 and 7) showed no considerable proliferation. On the basis of this finding, it appeared that two clearly different groups constituted HS. Thus, the responder group was defined as having a greater than 15-fold expansion of DN NKT cell numbers (proliferation rate = DN NKT cell numbers after culture/DN NKT cell numbers before culture) and the non-responder group as having less than a 15-fold expansion, in accordance with a previous report [27]. When WG and RP patients were examined in this way, only two of 10 WG patients (6, 7·3–228, 641, respectively), and two of nine RP patients (1·8, 15·3–148, 938, respectively) were found to be responders (Fig. 5b).

Fig. 5.

Proliferation of NKT cells upon stimulation with α-GalCer. (a) A representative result of a responder HS. DN NKT cells increased more than 20-fold 10 days after culture with α-GalCer and rhIL-2. (b) α-GalCer responders and non-responders in DN NKT populations from HS and patients with WG and RP. PBMCs (1 × 10⁶) from eight HS, 10 patients with WG and nine patients with RP were cultured with α-GalCer or vehicle alone in the presence of rhIL-2. The numbers of DN NKT cells/10⁵ lymphocytes before (pretreatment) and after culture with α-GalCer or vehicle are shown. The numbers on the x-axis indicate the number allocated to each HS or patient. *α-GalCer responder. (c) The number of CD4⁺ NKT cells/10⁵ lymphocytes before (pretreatment) and after culture with α-GalCer or vehicle. *α-GalCer responder. (d) Proportion of α-GalCer-tetramer⁺ cells in Vα24⁺Vβ11⁺ NKT cells. HS, and WG and RP patients were grouped into α-GalCer responder (res) and non-responder groups (non-res). **P < 0·01.

We then analysed responsiveness of CD4⁺ NKT cells to α-GalCer. Figure 5c shows that the proliferation rate of CD4⁺ NKT cells is significantly higher compared to that of DN NKT cells. Thus, we defined the responder group as having greater than 100-fold expansion and the non-responder group as less than 100-fold with regard to CD4⁺ NKT cell numbers. As shown in Fig. 5c, six of eight HS are responders (from 1·6–41·9 to 684–6656). On the other hand only two of 10 WG patients (from 7·4, 14·5–951, 1224, respectively), and three of nine RP patients (from 0–9·3 to 747–1429) were regarded as responders. These findings suggest that the proportions of the responder population are less in RP and WG patients than in HS, although the differences between HS and these patient groups are not statistically significant (in RP patients, P = 0·057, 0·153, and in WG patients, P = 0·054, 0·054, for DN and CD4⁺ NKT cells, respectively).

Because NKT cells responding to α-GalCer appear to compose a certain proportion of the whole NKT population, we quantified the frequency of α-GalCer-tetramer positive cells in the Vα24⁺Vβ11⁺ population of PBMCs from responder and non-responder groups (collected from both HS and patients). As shown in Fig. 5d, the mean proportion of α-GalCer-tetramer positive cells was significantly lower in α-GalCer non-responder than responder individuals (P < 0·01).

DISCUSSION

In IDDM and MS patients and mouse models of these diseases, it has been reported that Th1-predominant tissue damage is related to a loss of Vα24⁺Vβ11⁺ NKT cells [20–22,24,26]. In a mouse model of IDDM, adoptive transfer of NKT cell-enriched populations or activation of Vα14⁺ invariant NKT cells with α-GalCer protected against the development of IDDM [19]. In addition, Miyamoto et al. [20] reported that activation of Vα14⁺ invariant NKT cells with OCH, an analogue of α-GalCer, protected against EAE development in mice. A decrease in circulating Vα24⁺ NKT cell numbers has also been reported in RA, SSc and a wide variety of autoimmune diseases [23,25,27]. These observations support the idea that NKT cells may function as regulatory T cells and that their quantitative and/or functional defects are associated either directly or indirectly with the development of these autoimmune diseases [36,37].

In the present study, we showed for the first time that numbers of CD4⁺ Vα24⁺Vβ11⁺ NKT cells and DN Vα24⁺Vβ11⁺ NKT cells were reduced significantly in WG, a Th1-mediated autoimmune disease, and in RP patients. We also examined the frequency of NKT cells using α-GalCer-tetramer staining and showed that Vα24⁺/α-GalCer-tetramer⁺ cell numbers were reduced in these patients. Lee et al. [38] reported that the number of NKT cells was normal in patients with IDDM, including discordant twin pairs, by using a combination of α-GalCer-tetramer and anti-Vα24 MoAb staining. On the basis of this finding, these authors claimed that NKT cells might not be involved in pathogenesis of IDDM. On the other hand, Araki et al. [39] showed that frequency of NKT cells was decreased in patients with MS in remission by using a combination of α-GalCer-tetramer and anti-Vβ11 MoAb staining. In these MS patients, the frequency of α-GalCer-tetramer⁺ cells was correlated with that of Vα24⁺Vβ11⁺ NKT cells. It is unclear whether these differences are attributable to the different pathogenesis of these diseases.

RA is a common chronic inflammatory disease and thought to be a Th1-predominant disease. It has been suggested that some RP patients carry antibodies to type II collagen (CII) [40], a putative RA-associated autoantigen. An animal model for RP was established in mice by immunization with CII [7]. In RA, it has been reported that T cells, especially CD4⁺ T cells infiltrating into the joint, play a pathogenic role in situ[41]. Thus far, no report has described the functions of CD4⁺ T cells in RP patients. In the present study, we showed that upon stimulation with PMA and A23187 CD4⁺ T cells produced large amounts of IFN-γ in both RP and WG patients compared with HS. Whole PBMCs from WG and RP patients also showed significantly and slightly increased production of IFN-γ, respectively, but normal levels of IL-4, IL-5 and IL-10. The Th1-predominant state of WG patients was demonstrated more clearly when ratios of various cytokines against IL-10 or IFN-γ were compared (Fig. 4b,c). These findings suggest that both WG and RP patients are in a Th1-predominant state.

Recently, using CD1d-tetramer staining it was shown that CD4⁺ but not CD4^– NKT cells were the predominant source of Th2 cytokines in healthy individuals [17]. Araki et al. [38] reported that CD4⁺ (but not CD4^–) NKT cells from MS patients in remission exhibited a Th2 bias, and suggested that CD4⁺ NKT cells played a role in regulating Th1 autoimmunity. We have demonstrated herein that CD4⁺ NKT cells from WG and RP patients in remission were biased towards a Th2 phenotype by evaluating the balance between IL-4⁺ and IFN-γ⁺ cells in PBMCs after stimulation with PMA plus A23187. These findings, along with the cytokine secretion profiles of CD4⁺ T cells, suggest that NKT cells in WG and RP patients in remission have the potential to modulate the Th1-mediated autoimmune response despite their reduced frequency in PBMC. However, we must analyse increasing numbers of patients in both active and remission phase in further studies to determine precisely the relationship between cytokine balance and clinical manifestation.

Next, we attempted to examine the possible mechanism(s) underlying the significant reduction in NKT cell numbers in PBMCs from WG and RP patients. When the proportion of CD1d⁺ cells in B cells or monocytes and the expression levels of CD1d on these potential APCs were quantified, the frequency of CD1d⁺ cells in the monocyte population of RP patients was found to be significantly lower compared with that in HS. Although the differences were not statistically significant, we also found lower MFI levels of CD1d in RP patients compared to HS. However, sequencing of the complete CD1d gene including its transcriptional regulatory regions may be necessary to clarify basis for the observed low frequency of CD1d⁺ monocytes and to determine the relationship between the frequency of CD1d⁺ monocytes and that of NKT cells in RP patients [24].

Human NKT cells are activated vigorously by a synthetic glycolipid antigen, α-GalCer, as observed in other species [12–14]. Kojo et al. [27] reported that NKT cells in patients with autoimmune diseases such as RA, SLE and SSc could be categorized into two groups according to their reactivity to α-GalCer: the responder and non-responder. Furthermore, they suggested that an insufficient amount of natural ligands may, in part, be associated with the defective NKT cell development observed. In the present study, we observed expansion of NKT cells in a third to a quarter of WG and RP patients upon stimulation with α-GalCer. It appeared that these responder proportions were lower compared with those in HS. Thus, the reduction of NKT cell numbers in patients is probably attributable to insufficient positive selection of NKT cells in response to α-GalCer. It remains to be elucidated whether the defective positive selection in these patients is related to the expression level of natural ligands in vivo or whether other mechanisms operate.

In the final set of experiments, we calculated the proportion of Vα24⁺α-GalCer-tetramer⁺ cells in the Vα24⁺Vβ11⁺ NKT cell population from both responder and non-responder HS and patients. Of interest, the mean proportion of α-GalCer-tetramer⁺ cells was significantly lower in the α-GalCer non-responder group compared to the responder group (Fig. 5d). Thus, it seems that a considerable fraction of Vα24⁺Vβ11⁺ NKT cells cannot bind to α-GalCer plus CD1d in the non-responder group. This finding explains, at least partially, why PBMCs of the non-responder group were unable to proliferate upon stimulation with α-GalCer in vitro. Furthermore, we propose that quantification of the proportion of α-GalCer-tetramer⁺ cells in the Vα24⁺Vβ11⁺ NKT cell population may be a useful predictor of NKT cell responsiveness to α-GalCer in an individual.

In summary, this study showed that NKT cells (especially CD4⁺ NKT cells) might regulate the development of WG and RP by controlling the cytokine balance. Thus, NKT cells may serve as therapeutic targets whose number and/or functions could be modulated, thereby eventually leading to regulation of the development of these diseases.