Increased interleukin-18 expression in bone marrow of a patient with systemic juvenile idiopathic arthritis and unrecognized macrophage-activation syndrome



The aberrant induction of proinflammatory cytokines is considered to be crucial in the pathogenesis of systemic juvenile idiopathic arthritis and adult-onset Still's disease. Interleukin-18 (IL-18) in particular has been reported to be a candidate for the key cytokine in both diseases; however, the origin of IL-18 is unclear. To clarify the origin, we investigated specimens from various organs obtained during autopsy of a child with systemic JIA and macrophage activation syndrome, using immunohistochemical staining. Our results showed a high number of cells expressing IL-18 in the bone marrow but not in the other organs. This finding suggests that bone marrow is the origin of increased serum IL-18 and raises the possibility that other proinflammatory cytokines are also induced by IL-18 in bone marrow in this disease. Bone marrow may be an essential organ in the pathogenesis of systemic JIA.

Systemic juvenile idiopathic arthritis (JIA), also known as Still's disease, is one of the most common systemic and chronic inflammatory diseases in childhood and is characterized by chronic arthritis associated with a high spiking fever, a salmon-pink evanescent rash, hepatosplenomegaly, lymphadenopathy, and serositis. The precise etiology of systemic JIA remains unclear, although the activation of macrophages is thought to be crucial. Previous studies have demonstrated that the aberrant induction of proinflammatory cytokines such as interleukin-6 (IL-6), IL-1β, and tumor necrosis factor α (TNFα) may be involved in the pathogenesis of systemic JIA (1). However, it has not yet been elucidated which mediators are the key ones, nor has the origin of increased cytokines in systemic JIA been determined.

Recently, we and other researchers reported highly elevated serum levels of IL-18 (>10,000 pg/ml) in patients with systemic JIA or adult-onset Still's disease (AOSD), which is pathogenetically identical to systemic JIA (2–4). IL-18 strongly stimulates T lymphocytes, natural killer (NK) cells, and macrophages to produce proinflammatory cytokines (5, 6) and is therefore thought to be closely related to the pathogenesis of systemic JIA and AOSD. The cellular source of increased IL-18 synthesis in systemic JIA is unclear, although circulating or tissue monocyte/macrophages seem the most likely candidates in AOSD (3). In the present study, we investigated the source of increased IL-18 synthesis in a patient with systemic JIA.


The patient, a 20-month-old girl, was admitted to the county hospital because of a 2-week history of high spiking fever, rash, and tender arthritis of the extremities. Evaluation for bacterial and viral infections, including cytomegalovirus and Epstein-Barr virus, yielded negative results, and a bone marrow aspirate obtained at that time showed normal myelographic results. The patient was treated with intravenous antibiotics and aspirin for 1 week. However, her condition did not improve, and she was transferred to our hospital. Upon admission to our hospital, laboratory tests gave the following results: white blood cell (WBC) count 15,500/μl, hemoglobin concentration 10.0 gm/dl, platelet count 408,000/μl, serum glutamic oxaloacetic transaminase (SGOT) level 57 units/liter (normal <40), serum glutamic pyruvic transaminase (SGPT) level 37 units/liter (normal <35), lactate dehydrogenase (LDH) concentration 995 units/liter (normal <400), erythrocyte sedimentation rate 108 mm/hour, and serum C-reactive protein level 20.4 mg/dl (normal <0.5). Systemic JIA was diagnosed, and the dosage of aspirin was increased to 40 mg/kg/day.

Despite this week-long therapy, the patient's condition continued to deteriorate, and liver dysfunction and pancytopenia developed. A peripheral smear showed no abnormalities. Aspirin was then discontinued, and a course of parenteral steroids was instituted. However, 1 week later, her transaminase and LDH levels had further increased (SGOT 526 units/liter, SGPT 202 units/liter, and LDH 7,891 units/liter). In addition, the pancytopenia and coagulation disturbance had progressed (WBC count 2,900/μl, hemoglobin concentration 7.9 gm/dl, platelet count 38,000/μl, fibrinogen level 109 mg/dl, and fibrinogen–fibrin degradation product 27.4 μg/ml [normal <1.0 μg/ml]). The next day, before cyclosporin A treatment was initiated, the patient died of a pulmonary hemorrhage. At that time, her serum levels of ferritin and proinflammatory cytokines were extremely elevated, as follows: ferritin 12,178 ng/ml, interferon-γ (IFNγ) 4,900 pg/ml, TNFα 82.3 pg/ml, IL-6 83.3 pg/ml, monocyte colony-stimulating factor 627 pg/ml, and IL-18 210,288 pg/ml (IL-18 enzyme-linked immunoassay kit; MBL, Nagoya, Japan).

A postmortem examination, conducted with informed consent from the patient's parents, revealed hemophagocytosis in the bone marrow. These findings indicated macrophage activation syndrome (MAS), one of the most common and severe complications of systemic JIA (7). The pulmonary hemorrhage observed in this patient might be attributable to the hemorrhagic diathesis mediated by MAS; such a finding is reportedly a characteristic of MAS (7).

In order to investigate the origin of the highly increased level of IL-18, we used immunohistochemical staining with a monoclonal antibody to IL-18 (clone 25-2G; MBL) or CD68 (clone KP1; Dako, Kyoto, Japan) to examine specimens of the liver, spleen, axillary lymph nodes, lung, and bone marrow obtained from this patient. Autopsy specimens from a 22-month-old boy who had died of gastrointestinal bleeding were also studied as a control after informed consent was obtained. The bone marrow specimens from the patient with systemic JIA, compared with those from the control subject (Figure 1A), demonstrated prominent infiltrates of macrophages stained with CD68, and some macrophages showed hemophagocytosis (Figures 1C and D). Furthermore, it is noteworthy that strong IL-18 staining was observed in these bone marrow cells (Figures 1E and F). In contrast, few cells expressing IL-18 were observed in the bone marrow from a control subject (Figure 1B). The infiltration of CD68-positive cells in the patient's liver was not so significant and was localized around the portal tract, while CD68-positive Kupffer cells were widely seen in sinusoids in the control subject (Figures 2A and C). IL-18 was expressed in the liver cells of neither the patient with systemic JIA nor the control subject (Figures 2B and D). In addition, no cells with significant IL-18 staining were observed in the spleen, lung, or lymph nodes of the patient with systemic JIA (data not shown).

Figure 1.

Immunohistochemical staining of bone marrow with anti-CD68 monoclonal antibody or anti–interleukin-18 (anti–IL-18) monoclonal antibody. Bone marrow specimens from the control subject were stained with anti-CD68 (A) and anti–IL-18 (B), as were the specimens from the patient with systemic juvenile idiopathic arthritis (C and D and E and F, respectively). The specimens from the patient were also subject to control staining by replacing the primary antibody with an isotype-matched control reagent (mouse IgG1) (G), and by neutralizing (NT) the anti–IL-18 antibody with recombinant IL-18 protein (H). Brown color indicates positivity. (Original magnification × 100 in A, B, C, E, G, and H; × 400 in D and F; × 1,000 in insets.)

Figure 2.

Immunohistochemical staining of liver specimens with anti-CD68 monoclonal antibody or anti–interleukin-18 (anti–IL-18) monoclonal antibody. Liver specimens from the control subject were stained with anti-CD68 (A) and anti–IL-18 (B), as were those from the patient with systemic juvenile idiopathic arthritis (C and D). (Original magnification × 100.)


Our results suggest that highly elevated serum IL-18 levels originate in bone marrow, raising the possibility that bone marrow contains abundant proinflammatory cytokines induced by IL-18 in this disease. This hypothesis may be supported by reports that several kinds of viruses, such as varicella virus, measles virus, or parvovirus B19, can induce bone marrow suppression that may lead to transient clinical remission in children with intractable systemic JIA (8, 9). Furthermore, Schnedl et al reported that diffuse bone marrow edema was characteristically detected by magnetic resonance imaging and biopsies in patients with AOSD, and that this might be the result of increased permeability of the capillary endothelium in the bone marrow mediated by inflammatory cytokines (10). In addition, quantitative and qualitative abnormalities in bone marrow cells were demonstrated in patients with JIA, and these changes were considered to be consequences of the inflammatory milieu, including cytokines (11). Likewise, the etiologic association between bone marrow cells and rheumatoid arthritis (RA) in adults has been discussed. Remarkable elevations of IL-6, IL-8, and TNFα in bone marrow were reported in patients with RA (12, 13). Another study demonstrated that nurse-like cells from the bone marrow of patients with severe RA have unique characteristics and may play an important role in the pathogenesis of RA (14). The immunologic role of bone marrow cells in the pathogenesis of systemic JIA needs to be elucidated by further studies.

In this study, the significant expression of IL-18 was observed only in bone marrow and not in other organs. We at first supposed that the liver, especially its Kupffer cells, might be the source of the high elevation of IL-18, because Kupffer cells have been shown to produce IL-18 in a mouse model of endotoxin shock (5) and in children with biliary atresia (15). In our patient, however, immunohistochemical staining revealed few infiltrated Kupffer cells and no expression of IL-18 in the liver tissue. Therefore, the origin of IL-18 might differ among diseases. Accordingly, the expression of other cytokines, including IL-6, TNFα, and IFNγ, in liver and bone marrow should be also studied. We cannot exclude the possibility that our findings are a distinctive feature of MAS. However, considering that highly elevated serum IL-18 levels were detected in almost all patients with systemic JIA/AOSD regardless of the presence of MAS (2–4), our results may be commonly observed in patients with systemic JIA or AOSD.

It was recently reported that depressed NK cell functions might account for inadequate control of T cell and macrophage activation in patients with systemic JIA complicated by MAS (16), because NK cells and the perforin-based systems are normally involved in the down-regulation of cellular immune responses (17). Considering that NK cells are strongly stimulated by IL-18, the relationship between the highly elevated IL-18 level and the NK cell dysfunction in patients with systemic JIA is interesting and should be investigated. If NK cells produce no negative feedback despite highly elevated IL-18 levels, the sustained macrophage activation will result in further production of high levels of proinflammatory cytokines. In addition, autologous stem cell transplantation for severe systemic JIA was recently shown to restore the reduced perforin expression in CD8 T cells and NK cells (18) and to induce a significant and drug-free remission of the disease (19, 20). Bone marrow ablation and subsequent reconstitution may remove the activated cells and then allow regrowth of the normal cell population. Therefore, strategies that target activated bone marrow cells, including anti–IL-18 therapy, may lead to a more satisfactory treatment for systemic JIA.

In conclusion, we demonstrated that bone marrow was the probable source of highly elevated IL-18 levels in a patient with systemic JIA. Further investigations are needed to determine the role of “bone marrow IL-18” in the pathogenesis of systemic JIA or AOSD.