Systemic-onset juvenile idiopathic arthritis (JIA) is an autoimmune disease characterized by arthritis and systemic features. Its pathogenesis is still largely unknown. It is characterized immunologically by natural killer (NK) cell dysfunction and cytokine signatures that predominantly feature interleukin-1 (IL-1), IL-6, and IL-18. Since IL-18 can drive NK cell function, we examined how the high plasma levels of this cytokine are related to the documented NK cell failure in these patients.
The phenotype and function of NK cells from 10 healthy control subjects, 15 patients with polyarticular JIA, and 15 patients with systemic-onset JIA were characterized by staining and functional assays in vitro. IL-18 ligand binding was visualized by fluorescence microscopy. Phosphorylation of several MAP kinases and the IL-18 receptor β (IL-18Rβ) were visualized by Western blotting.
IL-18 from the plasma of systemic-onset JIA patients stimulated the activation of NK cells from healthy controls and bound its cognate receptor. However, NK cells from systemic-onset JIA patients failed to up-regulate cell-mediated killing molecules, such as perforin and interferon-γ, after IL-18 stimulation. Furthermore, treatment with IL-18 did not induce the phosphorylation of receptor-activated MAP kinases in NK cells. Alternate activation of NK cells by IL-12 induced NK cell cytotoxicity. We observed no additive effect of IL-18 in combination with IL-12 in systemic-onset JIA patients. Immunoprecipitation of IL-18Rβ showed that NK cells from systemic-onset JIA could not phosphorylate this receptor after IL-18 stimulation.
The mechanism of the impaired NK cell function in systemic-onset JIA involves a defect in IL-18Rβ phosphorylation. This observation has major implications for the understanding and, ultimately, the treatment of systemic-onset JIA.
Systemic-onset juvenile idiopathic arthritis (JIA), an autoimmune disease characterized by arthritis and systemic features, such as spiking fever, skin rash, generalized lymphadenopathy, hepatosplenomegaly, and serositis (1), was first described by the British pediatrician George F. Still in 1897 (2). The pathogenesis of systemic-onset JIA differs from that of other types of JIA in several respects, such as the lack of association with HLA type (3) and the absence of autoantibodies or autoreactive T cells (1). In fact, systemic-onset JIA has similarities to autoinflammatory diseases, as exemplified by a central role of the innate immune system and by the cytokines involved (e.g., interleukin-1 [IL-1], IL-6, and IL-18) (4). These cytokines are thought to be responsible for at least part of the clinical symptoms of the disease. Moreover, the blocking of both IL-1 and IL-6 has been shown to be efficacious in the treatment of systemic-onset JIA (5–7).
Besides the unrestrained production of cytokines, a role of natural killer (NK) cells in the pathogenesis of systemic-onset JIA has been suggested (8, 9). A key molecule in the granule-mediated exocytosis pathway in both NK cells and cytotoxic T lymphocytes (CTLs) is perforin. In familial hemophagocytic lymphohistiocytosis (FHL), a rare genetic disease, mutations are found in genes encoding perforin and related proteins that are involved in direct cell-mediated killing or cellular processes in NK cells (10, 11). The clinical features of FHL bear a marked resemblance to a common complication of systemic-onset JIA, namely, the macrophage activation syndrome (MAS) (1, 4). This resemblance between FHL and MAS is underscored by the observation of decreased NK cell function as well as decreased levels of perforin in CTLs and NK cells isolated from patients with systemic-onset JIA (8, 12, 13). It is still unclear what drives the cytokine production in relation to the observed NK cell dysfunction in systemic-onset JIA.
Recently, we showed that elevated levels of IL-18 in the plasma of patients with systemic-onset JIA correlated with disease activity (14). When IL-18 is bound to IL-18 receptor α (IL-18Rα), the ligand-binding chain of the receptor, IL-18Rβ is then recruited to form a high-affinity heterotrimeric complex with IL-18Rα and IL-18. Complex formation results in phosphorylation of the IL-18R β-chain, leading to recruitment of intracellular adaptor molecules (myeloid differentiation factor 88, interleukin-1 receptor–associated kinase, and tumor necrosis factor receptor−associated factor 6), and resulting in activation of ERK, JNK, and p38 MAP kinases (15–17). When NK cells are stimulated with IL-18, this cytokine can also potently induce interferon-γ (IFNγ) and increase NK cell activity.
We questioned why, despite increased circulating levels of IL-18, patients with systemic-onset JIA have decreased levels of perforin in CTLs and impaired NK cell function. We assumed that a defect in the IL-18/NK cell axis might contribute to the immunologic abnormalities found in systemic-onset JIA and set out to explore the functionality of NK cells in relation to IL-18. We found that although the increased levels of IL-18 in the plasma of patients with systemic-onset JIA is functional, NK cells in these patients show impaired IL-18–mediated activation due to defective phosphorylation of the IL-18R β-chain. This was not observed in patients with polyarticular JIA or in healthy control subjects. This finding has major implications for understanding and eventually treating systemic-onset JIA.
PATIENTS AND METHODS
All patients with JIA were classified according to the criteria of the International League of Associations for Rheumatology (18), and all had active disease, as previously defined (19). Blood samples from 15 patients with polyarticular JIA, 15 patients with systemic-onset JIA, and 10 age- and sex-matched healthy control subjects were collected into tubes containing heparin. All patients were seen at the outpatient clinic of the University Medical Centre Utrecht (Table 1). This study had full ethical approval from the local Institutional Review Board, and informed consent was obtained either from the patients (ages 12 years and older) or the patients' parents.
Table 1. Characteristics of the JIA study patients and controls*
Heparinized blood samples were centrifuged, and cell-free plasma samples were obtained and stored frozen at –80°C until they were analyzed. Subsequently, blood samples were washed 3 times with phosphate buffered saline (PBS) to remove residual autologous plasma. Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll density-gradient centrifugation (Pharmacia, Uppsala, Sweden) and cultured as described below.
Recombinant human IL-18 (MBL, Woburn, MA) was used in all cultures at various concentrations. All cell cultures were performed in RPMI 1640 (culture medium) supplemented with 2 mmoles/liter of L-glutamine, 100 units/ml of penicillin/streptomycin, and 5% volume/volume heat-inactivated fetal calf serum (all from Gibco Invitrogen, Grand Island, NY). Control experiments were performed using IL-18 in combination with 2 ng/ml of IL-12 (R&D Systems, Abingdon, UK).
Primary NK cells were enriched using magnetic cell separation techniques (negative selection; BD Biosciences, San Jose, CA) according to the manufacturer's instructions. Next, NK cells (CD3–CD16+CD56dim) were sorted with a FACSAria cell sorter (BD Biosciences). Cells (4 × 105 PBMCs/well or 4 × 104 primary NK cells/well) were cultured for 16 hours in the presence or absence of IL-18. During the final 4 hours, GolgiStop (BD Biosciences) was added to samples that were cultured for measurement of intracellular IFNγ production. If applicable, supernatants were harvested and stored at −80°C until they were used in the analyses.
Monoclonal antibodies to CD3, CD8, CD14, CD16, CD19, CD28, CD56, CD107a, CD161, CD314 (NKG2-D), CD2, CD244, CD335 (NKp46), CD337 (NKp30), perforin, and IFNγ were purchased from BD Biosciences. IL-18Rα and IL-18Rβ were obtained from R&D Systems. A monoclonal antibody against IL-18 was purchased from MBL. Extracellular and intracellular staining with various combinations of monoclonal antibodies was performed according to the manufacturer's protocol (BD Biosciences). During culture, granular exocytosis of NK cells and CTLs was assessed by staining for CD107a, as described previously (20, 21). Cells were analyzed using a FACSCalibur instrument and analyzed with CellQuest Pro software (both from BD Biosciences).
Determination of NK cell cytotoxicity.
Cytotoxicity of NK cells was assessed in standard 51Cr-release assays using the NK-specific K562 cell line as target cells (8). PBMCs were seeded in 96-well round-bottomed plates and precultured for 16 hours as described above, with and without various concentrations of IL-18. Target cells were then labeled with 100 μCi of NaCrO4 (Amersham Biosciences, Buckinghamshire, UK), and 4 × 104 labeled target cells were added to the cell culture and incubated for an additional 4 hours. All assays were performed at a 40:1 ratio of effector cells to target cells, which in pilot experiments, was shown to be in the linear range of this assay. Supernatant was harvested and counted in an automated gamma counter. Results are shown as the percentage release of chromium, which was calculated as follows: (counts per minute in experimental sample minus cpm spontaneous release) divided by (cpm maximum release minus cpm spontaneous release).
Measurement of extracellular protein levels.
Levels of various cytokines in plasma samples and in supernatants were measured using a multiplex immunoassay as described previously (22, 23). Measurements and data analysis of all assays were performed using the Bio-Plex system in combination with the Bio-Plex Manager software version 4.1, using 5 parametric curve fitting (all from Bio-Rad, Hercules, CA).
Perforin (Diaclone, Besançon, France) and IL-18 binding protein α (IL-18BPα; R&D Systems) were measured using enzyme-linked immunosorbent assay kits according to the manufacturer's instructions. The level of free, biologically active IL-18 in the IL-18/IL-18BPα complex was calculated based on the mass action law, using a dissociation constant of 0.4 nM and a stoichiometric ratio of 1:1 (24).
Determination of RNA expression.
Total RNA was isolated using TriPure isolation reagent (Roche Diagnostics, Mannheim, Germany) according to the manufacturer's instructions. First-strand complementary DNA (cDNA) was synthesized, and polymerase chain reactions were performed using LightCycler in combination with a FastStart DNA Master SYBR Green I kit (Roche Diagnostics), as previously described (25). A pool of cDNA was used as the standard, and normalization was performed using the constitutively expressed gene β2-microglobulin (β2m). The following primers were used: for β2m, 5′-CCAGCAGAGAATGGAAAGTC-3′ (forward) and 5′-GATGCTGCTTACATGTCT-3′ (reverse); for the perforin gene PRF1, 5′-CCGCTTCTCCATACGGGATTC-3′ (forward) and 5′-GCAGCAGCAGGAGAAGGAT-3′ (reverse); and for the T-bet gene, 5′-GGGAAACTAAAGCTCACAAAC-3′ (forward) and 5′-CCCCAAGGAATTGACAGTTG-3′ (reverse). All primers were obtained from TIB Molbiol (Berlin, Germany). Semiquantitative levels of PRF1 and T-bet were expressed as a percentage of the corresponding gene expression in the cDNA pool.
Immunoprecipitation and Western blotting.
Immunoprecipitation of IL-18Rβ was performed after cells were stimulated for 15 minutes in the presence or absence of 100 ng/ml of IL-18. Next, cells were lysed in radioimmunoprecipitation assay buffer (Sigma-Aldrich, Munich, Germany) supplemented with a protease inhibitor cocktail (Roche Diagnostics). Sequential immunoprecipitation was performed with protein A/G beads (Santa Cruz Biotechnology, Santa Cruz, CA), first with a combination mouse IgG (Santa Cruz Biotechnology) and goat IgG (Jackson ImmunoResearch, Newmarket, UK) and then with IL-18Rβ (R&D Systems). The immunoprecipitates were then dissolved in Laemmli buffer (Bio-Rad) and analyzed by Western blotting. Blots were incubated with antibody directed against phosphorylated tyrosine (BD Bioscience) or with antibody directed against IL-18Rβ (R&D Systems).
Phosphorylation of several MAP kinases was also assessed in whole NK cell lysates by Western blotting using antibodies against total and phosphorylated JNK, ERK-1/2, or p38 MAP kinase: p38, phospho-p38, and pERK-1/2 antibodies were from Cell Signaling Technology (Beverly, MA), pJNK and JNK antibodies were from Santa Cruz Biotechnology, and ERK-1/2 antibodies were from Upstate Biotechnology (Lake Placid, NY). Blots were subsequently incubated with peroxidase-conjugated secondary antibodies (DakoCytomation, Carpinteria, CA). Enhanced chemiluminescence was used according to the manufacturer's instructions (Amersham Biosciences). Western blots were analyzed using ImageJ software (NIH Image, National Institutes of Health, Bethesda, MD; online at: http://rsb.info.nih.gov/ij/).
Slides for fluorescence microscopy were precoated with poly-L-lysine (Sigma-Aldrich) for 30 minutes at 37°C and rinsed with PBS. After staining of PBMCs with CD3, CD16, and CD56 (all of which were mouse antibodies), NK cells (CD3–CD16+CD56dim) were sorted with a FACSAria instrument (BD Biosciences). Next, the cells were coated onto the slides and simultaneously stimulated in the presence or absence of IL-18 (100 ng/ml) for 1 hour at 37°C. After washing, the cells were fixed with 3% paraformaldehyde in PBS (Sigma-Aldrich). The slides were then incubated with IL-18Rβ antibody or an appropriate isotype control antibody (R&D Systems), followed by a secondary antibody (Invitrogen). Cells were embedded with mounting media for fluorescence microscopy (Hoechst, Frankfurt, Germany) and analyzed by confocal fluorescence microscopy. For each slide, 100 cells were counted and assessed for clustering of IL-18Rβ.
All results are expressed as the mean ± SD. Differences between variables were determined by Student's t-test. P values less than 0.05 were considered significant. All statistical analyses were performed using SPSS software version 12.0.1 (SPSS, Chicago, IL).
Plasma levels of IL-18.
Patients with systemic-onset JIA had significantly increased plasma levels of IL-18 as compared with polyarticular JIA patients and with healthy controls (P < 0.001 for each comparison) (Table 1), similar to findings previously described (14). The antibodies used for detection of IL-18 in plasma detect both biologically active IL-18 and IL-18 bound to IL-18BP. It is therefore not possible to distinguish between biologically active or biologically inactive bound IL-18 (26). Consequently, the amount of free, biologically active IL-18 was calculated in all study subjects using the IL-18BP levels that were measured in plasma (Table 1). No difference was found in plasma IL-18BP levels resulting in significantly higher levels of free IL-18 in plasma from systemic-onset JIA patients as compared with that from both the polyarticular JIA patients and the healthy controls (P < 0.001) (Table 1).
Relationship between increased IL-18 levels and defective NK cell function in systemic-onset JIA.
Next, we questioned how this increase in free IL-18 was related to the documented defective NK cell function in systemic-onset JIA (12). We first investigated the expression profiles of NK cell receptors as well as the NK cell numbers. A (nonsignificant) decrease in NK cell numbers was found in patients with systemic-onset JIA (mean ± SD 3.9 ± 3.5%; P = 0.064) compared with the healthy controls (7.9 ± 4.8%) and with the polyarticular JIA patients (7.8 ± 4.9%). Forty percent of the systemic-onset JIA patients lacked a CD56high population, as has previously been shown in another cohort of systemic-onset JIA patients (12); therefore, the CD56dim population was used in further analyses.
No difference was observed in the expression of CD2, CD244, or CD337 (NKp30). However, systemic-onset JIA patients had significantly fewer NK cells and lower expression of CD161, CD16 (Fc receptor γ IIIa), CD314 (NKG2D), and CD335 (NKp46) as compared with healthy controls (P < 0.05). (Data obtained on the phenotype of the NK cell receptor are available upon request from the corresponding author.)
Taken together, these data show a down-regulation of various lectin, immunoglobulin, and natural cytotoxicity receptors on NK cells isolated from systemic-onset JIA patients, indicating a different activation status of these cells.
Binding of IL-18 on NK cells.
The next question we addressed was whether the binding of IL-18 to its receptor on NK cells was hampered in any way. Thus, we determined both IL-18R expression and whether exogenous IL-18 could bind to its receptor on NK cells from patients with systemic-onset JIA. We analyzed IL-18Rα and IL-18Rβ expression on freshly isolated PBMCs from 5 subjects in each group. No difference was observed in IL-18 receptor expression on various cell populations analyzed (data not shown).
Subsequently, to determine the binding of IL-18 to its receptor, purified NK cells were stimulated with IL-18 and harvested after sequential time points. Cells were stained with an antibody against IL-18 (clone 125-2H), which has a shared binding epitope on the IL-18R β-chain. When IL-18 is bound to the receptor complex, this antibody is not able to bind IL-18, resulting in a lower ligand binding of the antibody that is visualized as reduced staining for this antibody (27). We found that exogenous IL-18 bound to IL-18R on NK cells from systemic-onset JIA patients to a similar degree as in healthy controls and polyarticular JIA patients, as shown by a decrease in antibody binding (Figure 1A). (Data obtained from the IL-18/IL-18R binding studies are available upon request from the corresponding author.)
To confirm these findings, IL-18Rβ expression on NK cells was also analyzed by fluorescence microscopy. After IL-18 stimulation of cells from healthy control subjects and systemic-onset JIA patients, IL-18Rβ was found to be dispersed throughout the cell membrane and to form clusters (Figure 1B), with similar percentages of NK cells showing clustering of IL-18Rβ (mean ± SD 91.6 ± 7.8% in healthy controls and 89.7 ± 9.9% in systemic-onset JIA patients). Thus, neither deficient expression of IL-18 nor hampered binding of IL-18 on its receptor on NK cells explains the NK cell dysfunction in patients with systemic-onset JIA.
Induction of NK cell cytolytic activity by IL-18.
Next, we investigated the ability of IL-18 to induce NK cell cytolytic activity. NK cells obtained from systemic-onset JIA patients displayed a significant decrease in NK cell cytolytic activity (mean ± SD 10.2 ± 5.2%) as compared with healthy controls (31.3 ± 7.1%; P < 0.01) (Figure 2A). We found no correlation between the levels of IL-18 and the numbers of NK cells (data not shown).
After IL-18 stimulation, there was no appreciable increase in NK cell cytolytic activity in cells from systemic-onset JIA patients, whereas cytolytic activity was significantly increased in cells from healthy controls (154% increase from baseline with 100 ng/ml of IL-18) as well as from polyarticular JIA patients (148% increase from baseline with 100 ng/ml) (Figure 2A).
Similar experiments were performed using purified primary NK cells from 3 subjects from each group. Comparable results were found after stimulation with 100 ng/ml of IL-18, showing a 1.7-fold increase (169%) in cytolytic activity in healthy controls and a 1.5-fold increase (147%) in polyarticular JIA patients, but no increase in the cytolytic effect after stimulation of cells from systemic-onset JIA patients.
Control experiments were performed in which cells were incubated with IL-12 alone or with IL-12 in combination with IL-18. The cytolytic activity of NK cells from healthy controls increased from 50.2 ± 13.8% to 70.5 ± 29.3% after incubation with IL-12 in combination with IL-18. Cells from polyarticular JIA patients showed a similar increase, from 43.6 ± 10.1% with IL-12 alone to 55.5 ± 18.3% with IL-12 plus IL-18. The cytolytic activity of NK cells from systemic-onset JIA patients increased after stimulation with IL-12 alone (21.7 ± 6.1%), but no additive effect was observed after stimulation with IL-12 in combination with IL-18 (21.3 ± 9.1%). Taken together, these results demonstrate that NK cells from systemic-onset JIA patients are incapable of responding to exogenous IL-18 stimulation.
Analysis of the modulation of NK cell function by free IL-18 in plasma.
Next, we questioned whether, among other molecules, free IL-18 in the plasma of systemic-onset JIA patients might be able to modulate NK cell function. To this end, we coincubated PBMCs from healthy controls, polyarticular JIA patients, and systemic-onset JIA patients with increasing concentrations of pooled plasma from systemic-onset JIA patients and assessed its cytolytic capacity. Cells from healthy controls and from polyarticular JIA patients showed a significant (P < 0.01 and P < 0.05, respectively), dose-dependent increase in cytolytic activity, whereas cells from systemic-onset JIA patients did not show an increased cytolytic response (Figure 2B).
We also studied plasma samples from 9 patients with systemic-onset JIA whose followup sample was obtained after at least 3 months and during a period when there had been no change in their medications. IL-18–mediated NK cell cytolytic activity did not change over time (Figure 2C). IL-18 levels decreased in systemic-onset JIA patients who achieved remission (P < 0.05) (data not shown), as has previously been shown (14), although they still exceeded the concentrations in healthy controls by as much as 100 times. One possible explanation for this finding is that although IL-18 can bind to its cognate receptor, IL-18 signaling is defective as a consequence of chronic stimulation of receptors due to the high plasma levels of IL-18 in patients with systemic-onset JIA.
To investigate this possibility, we mimicked chronically elevated IL-18 levels by incubating PBMCs from healthy controls (n = 5) and polyarticular JIA patients (n = 5) for up to 48 hours with 100 ng/ml of IL-18. Cells were then washed and stimulated for an additional 16 hours with fresh IL-18 (100 ng/ml), and NK cell cytolytic function was analyzed. Both the healthy control subjects and the polyarticular JIA patients showed increased cytolytic activity, a finding that contrasted with those in NK cells from patients with systemic-onset JIA, in which IL-18 did not potentiate cytolysis. (Data obtained from chronic stimulation of NK cells from patients with systemic-onset JIA with IL-18 are available upon request from the corresponding author.)
Expression of killing effector molecules.
We sought to determine whether the unresponsiveness of cytolytic NK cell activity to exogenous IL-18 was caused by decreased expression of killing effector molecules, such as perforin and IFNγ. In these analyses, intracellular expression of perforin and IFNγ was assessed after stimulation with IL-18.
We found that NK cells from polyarticular JIA patients and systemic-onset JIA patients had a significantly lower intracellular perforin content than did those from healthy controls (P < 0.05 and P < 0.001, respectively) (Figure 3A). When NK cells from healthy controls and polyarticular JIA patients, but not systemic-onset JIA patients, were stimulated with IL-18, intracellular perforin expression increased in a dose-dependent manner (Figure 3B).
Since perforin is posttranscriptionally regulated, the expression of messenger RNA (mRNA) for the perforin gene PRF1 was measured in primary NK cell cultures after stimulation with IL-18. There was a 10-fold up-regulation of PRF1 mRNA from healthy controls and polyarticular JIA patients, but not systemic-onset JIA patients (Figure 3C). Supernatants from these NK cell cultures were also analyzed for perforin release, and we found that the secreted perforin content was significantly higher in both healthy controls and polyarticular JIA patients than in systemic-onset JIA patients (P < 0.05 at 10 ng/ml and 100 ng/ml of IL-18) (Figure 3D).
To confirm whether the IL-18–stimulated release of perforin was NK cell–derived, NK cell granular exocytosis was assessed by analyzing the expression of CD107a. Increased expression of CD107a was observed in NK cells, but not CTLs, from healthy controls and from polyarticular JIA patients after stimulation with IL-18, indicating granular release specifically by NK cells but not CTLs. This finding is consistent with previous observations, since aside from IL-18, CD107a expression requires costimulation with other cytokines in order to be activated for the granular exocytosis of perforin (28). In contrast, NK cells from systemic-onset JIA patients demonstrated no increased CD107a expression after stimulation with IL-18 (Figures 3E and F).
Presence of other cytolytic responder cells following IL-18 stimulation.
Cytotoxic killing by NK cells can be stimulated by the presence of exogenous IFNγ. Since CTLs are also potent IFNγ producers, we investigated whether NK cells are the only cytolytic responder cells following stimulation with IL-18. The intracellular IFNγ content in NK cells from healthy controls and polyarticular JIA patients increased in a dose-dependent manner after IL-18 stimulation, whereas neither cell population from systemic-onset JIA patients showed induction of IFNγ (Figures 4A and B). Significantly reduced levels of IFNγ were found in systemic-onset JIA patient cell cultures that had been stimulated with graded concentrations of IL-18 as compared with the healthy controls and with polyarticular JIA patients (P < 0.05 at 10 ng/ml of IL-18 and P < 0.001 at 100 ng/ml of IL-18 for comparison with both the healthy controls and the polyarticular JIA patients) (Figure 4C).
Since only a small percentage of NK cells (±6%) produced IFNγ after stimulation with IL-18 (Figure 4), we performed control experiments in which PBMCs were incubated with IL-18 in combination with IL-12. Under these conditions, the levels of extracellular IFNγ were significantly increased. However, intracellular staining revealed the production of IFNγ by CTLs. (Data concerning the production of IFNγ by CTLs after coincubation of IL-18 with IL-12 are available upon request from the corresponding author.) To confirm the decreased production of IFNγ, we also assessed the expression of T-bet in primary NK cell cultures. A dose-dependent increase in T-bet was found in healthy controls and polyarticular JIA patients, but not systemic-onset JIA patients, after stimulation with IL-18 (Figure 4D).
Since the cytolytic activity of NK cells isolated from systemic-onset JIA patients did not increase upon the addition of exogenous IL-18, we questioned whether IL-18 intracellular signaling was functional. Purified NK cells were stimulated with IL-18, and cell lysates were analyzed for phosphorylation of p38, JNK, and ERK-1/2 MAP kinases. Unstimulated NK cells from systemic-onset JIA patients showed significantly increased phosphorylation of ERK-1/2 compared with both healthy controls and polyarticular JIA patients (P < 0.05 for both) (Figures 5A and B). Unlike the results in cells from healthy controls and polyarticular JIA patients, stimulation with IL-18 did not increase the phosphorylation of ERK-1/2 and JNK in NK cells from patients with systemic-onset JIA. Phosphorylation of p38 MAP kinase remained unchanged in cells from all 3 groups of study subjects after stimulation (Figures 5A and B). Thus, IL-18 stimulation does not result in the activation of MAP kinase in NK cells from patients with systemic-onset JIA. In other words, the binding of IL-18 to its receptor on NK cells from patients with systemic-onset JIA failed to bring about downstream signaling.
Phosphorylation of IL-18Rβ following IL-18 binding.
Ligand binding causes tyrosine phosphorylation of the IL-18Rβ intracellular domain, resulting in activation of MAP kinases (19). Since the IL-18–induced phosphorylation of ERK-1/2 and JNK MAP kinases is defective in NK cells from patients with systemic-onset JIA while IL-18 can still bind to its receptor (Figure 1), we questioned whether IL-18 binding results in the phosphorylation of IL-18Rβ. After stimulation with IL-18, immunoprecipitation with an IL-18Rβ antibody was performed, then blotted and analyzed for tyrosine phosphorylation. In contrast to the findings in healthy controls and polyarticular JIA patients, stimulation with IL-18 did not result in phosphorylation of the IL-18Rβ in systemic-onset JIA patients (Figure 5C). Thus, disturbed NK cell function in systemic-onset JIA patients is caused by a defect in IL-18 receptor function.
Systemic-onset JIA is an autoinflammatory-like disease with characteristic clinical features resembling those of hereditary autoinflammatory disorders (1, 29). Unlike the classic monogenetic autoinflammatory diseases, systemic-onset JIA is far more common and is expected to have a more complex multifactorial pathogenesis. So far, abnormalities have been found in 2 broad areas: increased cytokine-inducing capacity and dysfunction of NK cells. We show herein that these 2 areas are linked with regard to IL-18 and NK cell dysfunction, since CD3–CD16+CD56dim NK cells from patients with systemic-onset JIA were found to be refractory to IL-18 stimulation because of defective IL-18 receptor signal transduction.
Clearly, this defect prohibited the activation of NK cells by IL-18 in patients with systemic-onset JIA. The IL-18 receptor complex consists of an α-chain that is responsible for binding extracellular IL-18 and a nonbinding signal-transducing β-chain, both of which belong to the IL-1 receptor family (15). When IL-18 binds to the IL-18R α-chain, the IL-18R β-chain is recruited, inducing intracellular signaling pathways that are shared with IL-1R (15, 16).
In T and B cells, activation of this signaling pathway requires additional costimulation with IL-12 to yield cell activation (28). We have established the existence of an NK cell population that has the capacity to respond to IL-18 to induce IFNγ without costimulation with IL-12 and can thus activate the innate immune system. Patients with systemic-onset JIA have increased plasma levels of IL-18, but the levels of other potent NK cell stimulators, such as IL-2, IL-12, and IL-15, are in the normal range (14). When plasma samples from patients with systemic-onset JIA were added to NK cells from either healthy controls or patients with polyarticular JIA, NK cell function was augmented, indicating the presence of stimulatory molecules. Clinical symptoms arising during the pathogenesis of systemic-onset JIA can be explained by the high levels of IL-18, as reported in studies of humans as well as mice (30–33), but the question remains, what causes the induction of extremely high IL-18 levels in the plasma of patients with systemic-onset JIA?
IL-18, a member of the IL-1 cytokine superfamily, is stored in inflammasomes as precursor protein and, upon activation, is cleaved by caspase 1 in a manner similar to that of IL-1β and IL-33 to yield active cytokine (34–36). This cytokine is one of the most effective at regulating NK cell activity (37, 38). IL-1 has been suggested to play a role in the pathogenesis of systemic-onset JIA, and indeed, treatment with an IL-1 receptor antagonist has been reported to be clinically beneficial (6). Since the plasma levels of IL-1 and the numbers of gene transcripts of the IL-1 family are normal in systemic-onset JIA (6, 14, 39), it is not expected that IL-18 levels would be increased through hyperactivity of caspase 1. This hypothesis is underscored by the fact that no changes in the levels of IL-18 or pro–IL-1β were observed in a cohort of systemic-onset JIA patients before and after treatment with anti–IL-1 receptor antagonist (5).
Until now, no mutations have been found in cytokines, cytokine receptors, or downstream signal transduction molecules of the IL-1 family. Polymorphisms in the IL-18 promoter are related to increased levels of IL-18 in systemic-onset JIA patients, but even without these polymorphisms, IL-18 levels are still well above normal in these patients (40). Consistent with a recent gene expression profiling study (41) showing an overall down-regulation of the NK cell receptor profile in patients with systemic-onset JIA, we found a decreased expression profile of NK cell receptors on the CD3–CD16+CD56dim NK cell subset in systemic-onset JIA patients, while approximately one-third of these patients lacked the CD3–CD16+CD56bright NK cell subset in the circulating lymphocyte population. One explanation for these differentially activated NK cell populations could be that in systemic-onset JIA patients, there is expansion of one or more specific NK cell populations, as seen in other arthritic diseases, such as rheumatoid arthritis, spondylarthritis, and enthesitis-related arthritis (42). The low expression profile of NK cell receptors and reduced perforin content therefore contribute to diminished NK cell function and a different activation state of NK cells in systemic-onset JIA.
T-bet is a key transcription factor that regulates IFNγ gene expression in NK cells following activation by either JNK, ERK-1/2, or p38 MAP kinases (34, 43, 44). In addition, activation of the perforin lytic signaling pathway requires ERK-1/2 signaling (45, 46). We found increased basal activity of pERK in systemic-onset JIA patients; however, we were unable to find a causal link between IL-18 and activation of this MAP kinase. The high degree of spontaneous phosphorylation of ERK-1/2 seen in unstimulated cells from systemic-onset JIA patients suggests a different activation state of NK cells in systemic-onset JIA, which is possibly linked to the perforin pathway. Previous studies have also shown a direct link between phosphorylated ERK-1/2 and several genes that encode dysregulated IL-1R/Toll-like receptor–mediated inflammation, as well as NK cell receptors, in systemic-onset JIA (39, 41).
In conclusion, we have shown that the nonfunctional IL-18/NK cell axis is due to defective IL-18 receptor function in patients with systemic-onset JIA. Further research will be required to determine whether this defect represents a reversible tolerogenic effect due to the high levels of IL-18 or whether it represents a possible genetic defect either in the IL-18 receptor itself or in essential signaling molecules downstream of the IL-18 receptor. We believe that this novel finding is of importance for the understanding of the complex immune pathogenic mechanisms at play in systemic-onset JIA and may lead to new avenues for therapeutic intervention.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. de Jager had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. De Jager, Vastert, Kuis, Coffer, Prakken.
Acquisition of data. De Jager, Vastert, Beekman, Wulffraat.
Analysis and interpretation of data. De Jager, Beekman, Wulffraat, Kuis, Coffer, Prakken.
The authors thank Marianne Boes for critical reading of the manuscript.