Natural killer T cell deficiency in active adult-onset Still's disease: Correlation of deficiency of natural killer T cells with dysfunction of natural killer cells

Authors

  • Sung-Ji Lee,

    1. Chonnam National University Medical School and Chonnam National University Hospital, Gwangju, Republic of Korea
    Search for more papers by this author
    • Dr. S.-J. Lee and Ms Y.-N. Cho contributed equally to this work.

  • Young-Nan Cho,

    1. Chonnam National University Medical School and Chonnam National University Hospital, Gwangju, Republic of Korea
    Search for more papers by this author
    • Dr. S.-J. Lee and Ms Y.-N. Cho contributed equally to this work.

  • Tae-Jong Kim,

    1. Chonnam National University Medical School and Chonnam National University Hospital, Gwangju, Republic of Korea
    Search for more papers by this author
  • Seong-Chang Park,

    1. Chonnam National University Medical School and Chonnam National University Hospital, Gwangju, Republic of Korea
    Search for more papers by this author
  • Dong-Jin Park,

    1. Chonnam National University Medical School and Chonnam National University Hospital, Gwangju, Republic of Korea
    Search for more papers by this author
  • Hye-Mi Jin,

    1. Chonnam National University Medical School and Chonnam National University Hospital, Gwangju, Republic of Korea
    Search for more papers by this author
  • Shin-Seok Lee,

    1. Chonnam National University Medical School and Chonnam National University Hospital, Gwangju, Republic of Korea
    Search for more papers by this author
  • Seung-Jung Kee,

    1. Chonnam National University Medical School and Chonnam National University Hospital, Gwangju, Republic of Korea
    Search for more papers by this author
  • Nacksung Kim,

    1. Chonnam National University Medical School, Gwangju, Republic of Korea
    Search for more papers by this author
  • Dae-Hyun Yoo,

    1. The Hospital for Rheumatic Diseases, Hanyang University, Seoul, Republic of Korea
    Search for more papers by this author
  • Yong-Wook Park

    Corresponding author
    1. Chonnam National University Medical School and Chonnam National University Hospital, Gwangju, Republic of Korea
    • Department of Rheumatology, Research Institute of Medical Sciences, Brain Korea 21, Chonnam National University Medical School and Chonnam National University Hospital, 42 Jebong-ro, Dong-gu, Gwangju 501-757, Republic of Korea
    Search for more papers by this author

Abstract

Objective

To examine the levels and functions of natural killer (NK) and natural killer T (NKT) cells, investigate relationships between NK and NKT cells, and determine the clinical relevance of NKT cell levels in patients with adult-onset Still's disease (AOSD).

Methods

Patients with active untreated AOSD (n = 20) and age- and sex-matched healthy controls (n = 20) were studied. NK and NKT cell levels were measured by flow cytometry. Peripheral blood mononuclear cells were cultured in vitro with α-galactosylceramide (αGalCer). NK cytotoxicity against K562 cells and proliferation indices of NKT cells were estimated by flow cytometry.

Results

Percentages and absolute numbers of NKT cells were significantly lower in the peripheral blood of AOSD patients than in that of healthy controls. Proliferative responses of NKT cells to αGalCer were also lower in patients, and this was found to be due to proinflammatory cytokines and NKT cell apoptosis. In addition, NK cytotoxicity was found to be significantly lower in patients than in healthy controls, but NK cell levels were comparable in the 2 groups. Notably, this NKT cell deficiency was found to be correlated with NK cell dysfunction and to reflect active disease status. Furthermore, αGalCer-mediated NK cytotoxicity, showing the interaction between NK and NKT cells, was significantly lower in AOSD patients than in healthy controls.

Conclusion

These findings demonstrate that NK and NKT cell functions are defective in AOSD patients and suggest that these abnormalities contribute to innate immune dysfunction in AOSD.

Adult-onset Still's disease (AOSD) is an uncommon systemic inflammatory disorder of unknown etiology, which was first described as a distinct clinical syndrome by Bywaters in 1971 (1). Diagnosis is based on clinical and laboratory findings, such as high spiking fever, evanescent skin rash, polyarthralgia, lymphadenopathy, hepatosplenomegaly, leukocytosis, liver enzyme elevation, increased erythrocyte sedimentation rate, and elevated serum ferritin levels (2, 3). The etiology of AOSD is still largely unknown, but a complex interplay of genetic, environmental, and neuropsychogenic factors is believed to lead to the manifestations of the disease (4, 5). Furthermore, similarities between the clinical and laboratory features of systemic juvenile idiopathic arthritis (JIA) (formerly called Still's disease) and AOSD imply that these conditions have similar pathogenic mechanisms. Recently, it has been suggested that the contribution of innate immunity to systemic JIA is prominent and that systemic JIA be classified as an autoinflammatory disorder (6–8).

Natural killer (NK) cells principally participate in innate immunity. They exhibit extensive cytolytic activity and produce a variety of cytokines and chemokines (9, 10). Due to these properties, NK cells play significant roles in tumor immunosurveillance and in the control of viral infections and autoimmune disorders (10–12). Natural killer T (NKT) cells are a subset of T lymphocytes characterized by restricted expression of an invariant T cell receptor (Vα24–Jα18/Vβ11 in humans), and they recognize glycolipid antigens, such as α-galactosylceramide (αGalCer), presented by the class I major histocompatibility complex (MHC)–like molecule CD1d (13, 14). These cells have been implicated in the control of autoimmunity, cancer, and infectious disease (15). Furthermore, αGalCer-activated NKT cells have the ability to modulate innate and adaptive immunity (16). Although NK and NKT cells have distinct lineages, they exhibit striking similarities. For example, they express the same set of NK cell receptor protein 1 and Ly49 receptors and both are able to release massive amounts of cytokines, such as interferon-γ (IFNγ) (NK cells) and IFNγ and interleukin-4 (IL-4) (NKT cells), with extreme celerity without prior sensitization (11, 17, 18). Furthermore, it has been proposed that there is cross-talk between NK and NKT cells (19).

NK cell dysfunction is frequently observed in some human autoimmune diseases (20–22), and global NK cell impairment has been studied extensively in hemophagocytic lymphohistiocytosis (23–25). In addition, previous studies have demonstrated that, as in hemophagocytic lymphohistiocytosis, NK cell function is profoundly depressed in patients with systemic JIA and macrophage activation syndrome (26, 27). However, NKT cell levels and functions have not previously been investigated in systemic autoinflammatory disorders, such as systemic JIA and AOSD. In addition, the relevance of NKT cells to NK cell dysfunction has not been determined. The aim of the present study was to examine the levels and functions of NK and NKT cells in AOSD, to investigate potential relationships between the two, and to determine the clinical relevance of NKT cell levels in AOSD.

PATIENTS AND METHODS

Subjects.

Twenty patients (15 women and 5 men; mean ± SD age 38.7 ± 15.3 years) with active untreated AOSD fulfilling the criteria proposed by Yamaguchi et al (28) were enrolled in this study. The clinical and laboratory characteristics of the patients are summarized in Table 1. Remission was defined as the absence of clinical and laboratory evidence of disease activity for at least 2 consecutive months (3). Twenty age- and sex-matched healthy volunteers (15 women and 5 men; mean ± SD age 35.0 ± 8.9 years) were enrolled as healthy controls. All controls were documented to have no history of autoimmune disease, infectious disease, malignancy, chronic liver or renal disease, or diabetes mellitus. None of the controls had ever received immunosuppressive therapy, and none had fever during the 72 hours prior to enrollment. The study protocol was approved by the Institutional Review Board of Chonnam National University Hospital, and written informed consent was obtained from all participants.

Table 1. Baseline characteristics of the 20 patients with active adult-onset Still's disease*
  • *

    ESR = erythrocyte sedimentation rate; CRP = C-reactive protein; AST = aspartate aminotransferase; ALT = alanine aminotransferase; LDH = lactate dehydrogenase.

Age, median (range) years41 (17–60)
No. male/female5/15
Clinical symptoms, no. (%) 
 Fever20 (100)
 Myalgia15 (75)
 Skin rash14 (70)
 Arthralgia13 (65)
 Lymphadenopathy13 (65)
 Arthritis12 (60)
 Sore throat12 (60)
 Splenomegaly6 (30)
 Hepatomegaly4 (20)
 Pleurisy4 (20)
 Pneumonitis3 (15)
Laboratory variables, median (range) 
 Leukocytes,/μl9,700 (5,400–24,000)
 Neutrophils,/μl8,000 (700–22,500)
 Hemoglobin, gm/dl10.9 (8.9–12.6)
 Platelets, × 103/mm3276 (90–699)
 ESR, mm/hour58 (20–120)
 CRP, mg/dl7.56 (0.32–27)
 Albumin, gm/dl3.3 (2.5–4.3)
 AST, units/liter44 (14–618)
 ALT, units/liter30 (5–403)
 LDH, units/liter770 (470–2,842)
 Ferritin, ng/ml1,932 (173–40,000)

Monoclonal antibodies (mAb) and flow cytometry.

The following mAb and reagents were used in this study: fluorescein isothiocyanate (FITC)– or PerCP-conjugated anti-CD3, FITC-conjugated anti-CD45, allophycocyanin (APC)–conjugated anti-CD56, phycoerythrin (PE)–conjugated anti-6B11, FITC-conjugated annexin V, 7-aminoactinomycin D (7-AAD), APC-conjugated anti-IFNγ, and APC-conjugated mouse IgG isotype control (all from Becton Dickinson), and FITC-conjugated anti-Vβ11 and PE-conjugated anti-Vα24 (both from Immunotech). Cells were stained with combinations of appropriate mAb for 20 minutes at 4°C. Stained cells were analyzed on a FACSCalibur flow cytometer using CellQuest software (BD Biosciences).

Isolation of peripheral blood mononuclear cells (PBMCs) and identification of NK and NKT cells.

Peripheral venous blood samples were collected into heparin-containing tubes, and PBMCs were isolated by density-gradient centrifugation using Ficoll-Paque Plus solution (Amersham Biosciences). NK and NKT cells were identified phenotypically as CD3−CD56+ and CD3+6B11+ cells, respectively, by flow cytometry as previously described (20, 29).

NKT cell proliferation assay.

The proliferative abilities of NKT cells were assayed by flow cytometry as previously described (29). Briefly, freshly isolated PBMCs were suspended in complete media supplemented with 10% fetal bovine serum (FBS; Gibco BRL), seeded in a 24-well plate at 1 × 106 cells/well, and then cultured for 7 days at 37°C in a 5% CO2 humidified incubator in the presence of IL-2 (100 IU/ml; BD PharMingen) and αGalCer (100 ng/ml; Alexis Biochemicals) or 0.1% DMSO as a control. Cells were harvested and stained with FITC-conjugated anti-CD3 and PE-conjugated anti-6B11 mAb. Percentages of CD3+6B11+ NKT cells were determined by flow cytometry using a lymphoid gate. The proliferation index was defined as the ratio between NKT cell percentages on day 7 and day 0; indices are expressed as fold increases. To quantify NKT cell death in culture, cells were stained with FITC-conjugated annexin V, PE-conjugated mAb 6B11, 7-AAD, and APC-conjugated anti-CD3 mAb as previously described (29–31). Percentages of apoptotic (annexin V–positive) and necrotic (7-AAD–positive) NKT cells were measured by flow cytometry on days 0 and 7.

To determine changes in the proliferative responses of NKT cells to αGalCer after cytokine stimulation, freshly isolated PBMCs were stimulated with a proinflammatory cytokine cocktail consisting of IL-1β (10 ng/ml), IL-6 (50 ng/ml), IL-8 (10 ng/ml), IL-18 (100 ng/ml), IFNγ (10 ng/ml), and tumor necrosis factor α (TNFα) (5 ng/ml) (all from PeproTech) for 3 days in the presence or absence of cytokine inhibitors (i.e., blocking antibodies), and then proliferation indices were determined by flow cytometry as described above. Blocking antibodies for cytokines included anti–IL-18 (5 μg/ml; R&D Systems) and anti–IL-1β (5 μg/ml), anti–IL-6 (5 μg/ml), anti–IL-8 (5 μg/ml), anti-IFNγ (5 μg/ml), and anti-TNFα (5 μg/ml) (all from BD Biosciences).

Functional NKT cell assay.

IFNγ expression in NKT cells was detected by intracellular cytokine flow cytometry after stimulation of cells with αGalCer as previously described (29, 32). Briefly, freshly isolated PBMCs (1 × 106/well) were incubated in 500 μl complete media supplemented with 10% FBS and αGalCer (200 ng/ml) for 2 hours. Medium (500 μl) containing brefeldin A (GolgiPlug; BD Biosciences) was then added; the final concentrations of αGalCer and brefeldin A were 100 ng/ml and 10 μg/ml, respectively. After incubation for a further 6 hours, cells were stained with FITC-conjugated anti-Vβ11 and PE-conjugated anti-Vα24 mAb for 20 minutes at 4°C, fixed in 4% paraformaldehyde for 15 minutes at room temperature, and permeabilized with Perm/Wash solution (BD Biosciences) for 10 minutes. Cells were then stained with APC-conjugated anti-IFNγ mAb for 30 minutes at 4°C and analyzed by flow cytometry.

NK cytotoxicity assay.

NK cytotoxicity against K562 cells was assayed by flow cytometry as previously described (20, 33). Briefly, isolated PBMCs were used as effector cells and were cultured for 4 hours at 37°C in complete media supplemented with 10% FBS in a humidified incubator containing 5% CO2. K562 cells (CCL-243; ATCC) were used as target cells. Effector and target cells were mixed in 12 × 75–mm round-bottomed polystyrene tubes (Becton Dickinson) at an effector-to-target (E:T) cell ratio of 20:1. Control tubes including only target cells were also assayed to quantify spontaneous K562 cell death. Tubes were incubated for 4 hours at 37°C in a humidified incubator containing 5% CO2. Mixed effector and target cells were stained with FITC-conjugated anti-CD45 mAb for 20 minutes at 4°C, washed once in phosphate buffered saline (PBS), resuspended in 0.5 ml of PBS containing 20 μl of 1 μg/ml propidium iodide (Becton Dickinson), and incubated for 15 minutes at room temperature. Percentages of dead K562 cells were determined by flow cytometry. NK cytotoxicity was calculated by subtracting the percentages of dead K562 cells in control tubes from the percentages of dead cells in sample tubes.

To determine αGalCer-mediated NK cytotoxicity, freshly isolated PBMCs (1 × 106/well) were stimulated with αGalCer (100 ng/ml) or 0.1% DMSO as a control for 24 hours in the presence or absence of anti–IL-2 (5 μg/ml), anti-IFNγ (5 μg/ml), and/or anti-TNFα (5 μg/ml). NK cytotoxicity was then determined by flow cytometry and expressed as the enhancement ratio, i.e., the ratio of NK cytotoxicity in the presence of αGalCer to NK cytotoxicity in the absence of αGalCer.

Statistical analysis.

The Mann-Whitney U test was used to compare percentages and absolute numbers of NK and NKT cells, proliferation indices of NKT cells, and NK cytotoxicity, as well as percentages of apoptotic and necrotic NKT cells and αGalCer-mediated NK cytotoxicity in AOSD patients versus healthy controls. Relationships between NKT cell levels and NK cytotoxicity were examined using Spearman's correlation coefficient. Wilcoxon's signed rank test was used to compare changes in proliferation indices of NKT cells after cytokine stimulation, in αGalCer-mediated NK cytotoxicity in the presence of cytokine inhibitors, and in NKT cell levels according to disease activity. P values less than 0.05 were considered significant. All statistical analyses were performed using SPSS, version 17.0.

RESULTS

Reduced numbers of circulating NKT cells in patients with AOSD.

The percentages and absolute numbers of NKT cells in the peripheral blood samples from the 20 AOSD patients and the 20 age- and sex-matched healthy controls were determined by flow cytometry. NKT cells were defined as those coexpressing CD3 and 6B11 (Figure 1A). Percentages of NKT cells were significantly lower in patients than in controls (median 0.03% versus 0.09%; P = 0.001) (Figure 1B). Absolute NKT cell numbers were calculated by multiplying NKT cell percentages by total lymphocyte numbers (per microliter) in peripheral blood as described previously (29). The absolute numbers of NKT cells were significantly lower in patients than in controls (median 0.3 cells/μl versus 1.2 cells/μl; P < 0.001) (Figure 1C).

Figure 1.

Decreased circulating natural killer T (NKT) cell numbers in the peripheral blood of adult-onset Still's disease (AOSD) patients. Freshly isolated peripheral blood mononuclear cells from 20 AOSD patients and 20 age- and sex-matched healthy controls (HCs) were stained with PerCP-conjugated anti-CD3 and phycoerythrin-conjugated anti-6B11 monoclonal antibodies and then analyzed by flow cytometry. Percentages of NKT cells were calculated using a lymphoid gate. A, Representative NKT cell percentages as determined by flow cytometry. B and C, NKT cell percentages among peripheral blood lymphocytes (B) and absolute NKT cell numbers (C). Symbols represent individual subjects; horizontal bars show the median. ∗ = P < 0.005; ∗∗ = P < 0.001.

Impaired response of NKT cells to αGalCer in patients with AOSD.

To examine the proliferative effects of αGalCer on NKT cells, PBMCs from the AOSD patients and controls were cocultured with αGalCer for 7 days in the presence of 100 IU/ml IL-2. Proliferation indices were evaluated by flow cytometry as described in Patients and Methods. The percentage of NKT cells among PBMCs from control subjects increased markedly in response to αGalCer (from 0.16% on day 0 to 12.5% on day 7 in a representative subject). In contrast, NKT cells from AOSD patients proliferated only slightly (Figure 2A). Overall proliferation indices were significantly lower in patients than in controls (median 3.7 versus 37.1; P < 0.001) (Figure 2B). To investigate cytokine expression in NKT cells, we examined IFNγ expression at the single-cell level by intracellular cytokine flow cytometry. Percentages of IFNγ+ NKT cells were found to be markedly reduced in AOSD patients (Figure 2C).

Figure 2.

Decreased proliferative responses of NKT cells from AOSD patients to α-galactosylceramide (αGalCer). A and B, Freshly isolated peripheral blood mononuclear cells (PBMCs; 1 × 106/well) from 20 AOSD patients and 20 age- and sex-matched healthy controls were cultured for 7 days in the presence of interleukin-2 (100 IU/ml) and αGalCer (100 ng/ml) or 0.1% DMSO as a control. Representative flow cytometry results are shown (A). The proliferation index was determined as described in Patients and Methods. Symbols represent individual subjects; horizontal bars show the median (B). C, Interferon-γ (IFNγ) expression in Vα24+Vβ11+ NKT cell populations was examined by flow cytometry. Results are representative of 5 experiments. D, NKT cell apoptosis was assessed. Values are the mean ± SEM (n = 5 per group). E, Effects of stimulation with a proinflammatory cytokine cocktail on the proliferative responses of NKT cells to αGalCer were determined. Freshly isolated PBMCs from 6 control subjects were stimulated with a cytokine cocktail for 3 days in the presence or absence of cytokine inhibitors (i.e., blocking antibodies [Abs]) as described in Patients and Methods, and proliferation indices were determined by flow cytometry. Values are the mean ± SEM. ∗ = P < 0.05; ∗∗ = P < 0.001. See Figure 1 for other definitions.

To determine whether the impaired proliferative response of NKT cells to αGalCer was due to a loss of NKT cells in culture, PBMCs from AOSD patients and healthy controls were cultured for 7 days in the presence of IL-2 (100 IU/ml) and αGalCer (100 ng/ml) and then NKT cell apoptosis was evaluated by flow cytometry. Percentages of NKT cells that were apoptotic and necrotic were minimal on day 0, with no significant differences between patients and controls. However, after stimulation with αGalCer for 7 days, NKT apoptosis and necrosis were greater in patients than controls (mean ± SEM 33.9 ± 6.5% versus 10.7 ± 2.2% [P < 0.05] and 31.3 ± 4.7% versus 13.0 ± 1.4% [P < 0.05], respectively) (Figure 2D).

Several groups of investigators have suggested that levels of proinflammatory cytokines, such as IL-1β, IL-6, IL-8, IFNγ, TNFα, and IL-18, are elevated in the sera of patients with active AOSD (34–36). We hypothesized that these high levels of proinflammatory cytokines influence the apparent dysfunction and reduce responsiveness of NKT cells. PBMCs from 6 healthy controls were prestimulated for 3 days with a proinflammatory cytokine cocktail consisting of IL-1β, IL-6, IL-8, IFNγ, TNFα, and IL-18, and proliferation indices were then determined by flow cytometry. Proliferation indices were significantly reduced by stimulation with the proinflammatory cytokine cocktail (mean ± SEM 12.7 ± 2.8 versus 73.1 ± 14.1; P < 0.05). Furthermore, when blocking antibodies were added along with αGalCer and the proinflammatory cytokine cocktail, the stimulation index was higher than that obtained with αGalCer and the proinflammatory cytokine cocktail alone (mean ± SEM 28.7 ± 4.7 versus 12.7 ± 2.8; P < 0.05) (Figure 2E).

Impaired NK cytotoxicity in patients with AOSD.

The percentages and absolute numbers of NK cells in the peripheral blood of AOSD patients and healthy controls were determined by flow cytometry. NK cell percentages were found to be comparable in the patients and controls (median 10.5% and 11.4%, respectively; P = 0.317) (Figure 3A). However, absolute NK cell numbers were slightly lower in patients than in controls (median 114.5/μl versus 172.5/μl; P = 0.04) (Figure 3B).

Figure 3.

Decreased cytotoxicity of NK cells in AOSD patients. A and B, Freshly isolated peripheral blood mononuclear cells from 20 AOSD patients and 20 age- and sex-matched healthy controls were stained with fluorescein isothiocyanate–conjugated anti-CD3 and phycoerythrin-conjugated anti-CD56 monoclonal antibodies and then analyzed by flow cytometry. CD3−CD56+ cell values are expressed as percentages of peripheral blood lymphocytes (A) and as absolute numbers (B). C and D, NK cytotoxicity was determined as described in Patients and Methods and expressed as the percentage of apoptotic K562 cells. Representative flow cytometry results (C) and results in individual subjects (D) are shown. In A, B, and D, symbols represent individual subjects; horizontal bars show the median. ∗ = P < 0.05; ∗∗ = P < 0.001. See Figure 1 for definitions.

To examine the cytotoxic effects of NK cells on K562 cells, we used patient and control PBMCs. NK cytotoxicity was evaluated by flow cytometry at an E:T cell ratio of 20:1. NK cytotoxicity in a representative healthy control subject was 16.6%, compared with 5.9% in a representative patient (Figure 3C). Overall, NK cytotoxicity was significantly reduced in patients (median 5.7%, versus 19.0% in controls; P < 0.001) (Figure 3D).

Correlation between NKT cell deficiency and NK cell dysfunction in patients with AOSD.

To evaluate the relationship between NKT cells and NK cells in AOSD, we investigated associations between NKT cell levels and NK cytotoxicity, using Spearman's rank correlation analysis. The analysis showed that NK cytotoxicity was significantly correlated with NKT cell percentages (rs = 0.61, P = 0.005; n = 20) (Figure 4A) and numbers (rs = 0.47, P = 0.037; n = 20) (Figure 4B). These results suggest that NK cell dysfunction is related to NKT cell deficiency in AOSD.

Figure 4.

Relationship between NK cells and NKT cells in AOSD. A and B, NK cytotoxicity plotted against NKT cell percentages (rs = 0.61, P = 0.005) (A) and absolute numbers (rs = 0.47, P = 0.037) (B) in the 20 patients with AOSD. Symbols represent individual patients. C–E, Effect of α-galactosylceramide (α-GalCer) on NK cytotoxicity. Freshly isolated peripheral blood mononuclear cells (1 × 106/well) were stimulated with αGalCer (100 ng/ml) or 0.1% DMSO as a control for 24 hours in the presence or absence of cytokine inhibitors (i.e., blocking antibodies), and NK cytotoxicity was then determined as described in Patients and Methods. Representative NK cytotoxicity as determined by flow cytometry is shown (C), with values expressed as percentages of apoptotic K562 cells. NK cytotoxicity mediated by αGalCer in 5 AOSD patients and 5 age- and sex-matched healthy controls (D) and NK cytotoxicity mediated by αGalCer in the presence of cytokine inhibitors in 6 healthy controls (E) was determined and was expressed as an enhancement ratio. Values in D and E are the mean ± SEM. ∗ = P < 0.05; ∗∗ = P < 0.01. Anti-TNFα = anti–tumor necrosis factor α; anti-IFNγ = anti–interferon-γ; anti–IL-2 = anti–interleukin-2 (see Figure 1 for other definitions).

Effect of αGalCer on NK cytotoxicity in patients with AOSD.

To assess whether NKT cells have the potential to enhance NK cytotoxicity, PBMCs from 5 AOSD patients and 5 healthy controls were preincubated for 24 hours in the presence or absence of αGalCer, and NK cytotoxicity was evaluated by flow cytometry. In the controls, αGalCer administration resulted in enhancement of NK cytotoxicity against K562 cells, whereas in the patients, the effect of αGalCer on NK cytotoxicity was only marginal (Figure 4C). Overall, αGalCer-mediated NK cytotoxicity was significantly lower in patients than in controls (mean ± SEM enhancement ratio 1.0 ± 0.5 versus 1.7 ± 0.2; P < 0.001) (Figure 4D).

To determine whether αGalCer-mediated NK cytotoxicity is mediated by cytokines produced by NKT cells, PBMCs from 6 healthy controls were stimulated with αGalCer for 24 hours in the presence or absence of anti–IL-2, anti-IFNγ, and/or anti-TNFα, and NK cytotoxicity was determined by flow cytometry. NK cytotoxicity mediated by αGalCer was significantly lower in the presence of the combination of anti–IL-2, anti-IFNγ, and anti-TNFα than in the absence of blocking antibodies (mean ± SEM enhancement ratio 1.0 ± 0.4 versus 1.5 ± 0.1; P < 0.05). Moreover, αGalCer-mediated NK cytotoxicity was found to be significantly reduced only by anti-IFNγ, irrespective of the presence of anti–IL-2 or anti-TNFα (enhancement ratio 1.1 ± 0.1 versus 1.5 ± 0.1; P < 0.05) (Figure 4E).

Changes in NKT cell levels according to disease activity.

Given our observation that peripheral blood NKT cell levels are reduced in patients with active AOSD, we investigated changes in these in relation to disease activity. Six AOSD patients were available for followup examination of NKT cell levels. NKT cell percentages and numbers were found to be greater when the disease was in remission than when it was active (mean ± SD 0.17 ± 0.14% versus 0.05 ± 0.03% [P < 0.05] and 1.12 ± 1.19 cells/μl versus 0.17 ± 0.16 cells/μl [P < 0.05], respectively) (Figures 5A and B).

Figure 5.

Changes in NKT cell levels in AOSD patients. The percentages (A) and absolute numbers (B) of NKT cells in the peripheral blood of 6 AOSD patients during active disease and during remission were determined by flow cytometry. Symbols represent individual patients. ∗ = P < 0.05. See Figure 1 for definitions.

DISCUSSION

This study represents a first attempt to investigate numerical and functional deficiencies of NKT cells in patients with active untreated adult-onset Still's disease. In addition, peripheral blood samples from patients were analyzed to examine NK cell dysfunction, and it was demonstrated that such dysfunction was related to NKT cell deficiency. The potential relationship between NK and NKT cells was further explored using an in vitro αGalCer-mediated NK cytotoxicity assay, which showed impaired NK cytotoxicity in response to αGalCer in AOSD patients. These results suggest that NKT cell deficiency influences NK cell dysfunction in AOSD.

The observation that NKT cell levels and functions are defective has also been reported in some human autoimmune diseases, such as rheumatoid arthritis (37) and systemic lupus erythematosus (20). In accordance with the results of previous studies, the present study demonstrated that circulating NKT cell numbers were reduced and proliferative responses of NKT cells to αGalCer were impaired in AOSD patients. The semi-invariant T cell receptor of NKT cells recognizes glycolipid antigens presented by the non-polymorphic major class I MHC–related protein CD1d (38, 39). Indeed, it has been demonstrated that blocking of CD1d using mAb or a germline deletion inhibits the proliferative response of NKT cells to αGalCer (40, 41). To test the possibility that NKT cell deficiency in AOSD might be due to alterations in CD1d expression, we analyzed the expression levels of CD1d. However, CD1d levels were found to be comparable in patients and healthy controls (data available from the corresponding author upon request).

One possible explanation for NKT cell deficiency in AOSD is systemic inflammation caused by proinflammatory cytokines. Our in vitro stimulation assay showed that proinflammatory cytokines markedly inhibited the proliferative responses of NKT cells to αGalCer and that the response was partially recovered with their cytokine inhibitors, suggesting that proinflammatory cytokines contribute at least in part to NKT cell dysfunction in AOSD. Another possibility is increased NKT cell apoptosis in AOSD patients. Chen et al demonstrated that peripheral blood lymphocytes from AOSD patients are susceptible to spontaneous and IL-18–stimulated apoptosis (42). In the present study, proportions of αGalCer-stimulated apoptotic NKT cells were found to be higher in patients than in healthy controls. Moreover, circulating NKT cell levels were found to be reduced during active disease but increased during remission, which suggests that NKT cell numbers in peripheral blood reflect disease activity.

Using an NK cytotoxicity assay against K562 cells, we demonstrated NK cell dysfunction in patients with AOSD. This is consistent with the report by Villanueva et al of decreased NK cytotoxicity in patients with systemic JIA and macrophage activation syndrome (27). Furthermore, it has been reported that perforin expression in NK cells is down-regulated in systemic JIA (43, 44). Previous studies have demonstrated low proportions of NK cells among the PBMCs of patients with systemic JIA (44, 45), and we examined whether this is related to the observed decreases in NK cytotoxicity. We found overall NK cell proportions among PBMCs to be comparable in AOSD patients and healthy controls. However, NK cells from patients expressed lower levels of perforin (data available from the corresponding author upon request). These observations suggest that suppressed NK cell function in AOSD is a result of an NK cell defect rather than a reduction in NK cell numbers.

Our data also indicate that NK cell dysfunction is related to NKT cell deficiency in AOSD. Correlation analysis showed that NK cytotoxicity was significantly correlated with peripheral blood NKT cell percentages and numbers. Studies have shown that stimulation of NKT cells with αGalCer can rapidly activate NK cells via the production of IFNγ and IL-2, suggesting a relationship between NK and NKT cells (19, 46, 47). In our in vitro experiments, administration of αGalCer to PBMCs from healthy controls induced IFNγ production by NKT cells and enhanced NK cytotoxicity. In addition, in our experiments using blocking antibodies, αGalCer-mediated NK cytotoxicity was found to be predominantly mediated by IFNγ. In contrast, this was not observed in the PBMCs of AOSD patients. These findings suggest that numerical and functional deficiencies of NKT cells contribute to NK cell dysfunction in AOSD. Further studies are needed to identify the molecular mechanism responsible for NKT cell deficiencies in AOSD and to investigate strategies to repair NK and NKT cell dysfunction.

In summary, the present results show that NKT cells are numerically and functionally deficient in AOSD. In addition, we report the novel finding that NK cell dysfunction is related to NKT cell deficiency. These findings provide important information concerning the pathogenesis of AOSD.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting or revising the article for intellectual content, and all authors approved the final version. Dr. Y.-W. Park had full access to all of the data and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. N. Kim, Yoo, Y.-W. Park.

Acquisition of data. S.-J. Lee, Cho, S.-C. Park, D.-J. Park, Jin.

Analysis and interpretation of data. T.-J. Kim, S.-S. Lee, Kee, Y.-W. Park.

Acknowledgements

The authors thank Ms Ee-Seul Park and Ms Mun-Ju Kim (Department of Rheumatology, Chonnam National University Hospital) for technical assistance.

Ancillary