To longitudinally investigate serum and synovial fluid (SF) levels of RANTES and monocyte chemoattractant protein 1 (MCP-1) as well as in vitro migration of mononuclear cells toward SF in patients with juvenile rheumatoid arthritis (JRA).
To longitudinally investigate serum and synovial fluid (SF) levels of RANTES and monocyte chemoattractant protein 1 (MCP-1) as well as in vitro migration of mononuclear cells toward SF in patients with juvenile rheumatoid arthritis (JRA).
Serum and SF levels of RANTES and MCP-1 were determined by enzyme-linked immunosorbent assay. Chemotaxis was performed using the modified Boyden chamber method.
Serum RANTES levels were significantly increased in all onset types of JRA, with the highest levels present in systemic-onset JRA. Serum MCP-1 levels were significantly elevated in patients with systemic-onset JRA and were associated with current systemic features. Although serum levels of RANTES and MCP-1 decreased significantly after treatment, RANTES and MCP-1 levels during disease remission were still significantly higher in JRA patients than in controls. A relationship was found between serum RANTES levels during remission and the duration of clinical remission, with low levels being associated with prolonged clinical remission and high levels with shorter clinical remission. Serum RANTES levels correlated with C-reactive protein concentrations, hemoglobin values, white blood cell (WBC) counts, and platelet counts, whereas serum MCP-1 levels correlated with WBC counts and serum ferritin levels. Levels of RANTES and MCP-1 in SF were elevated as compared with levels in serum. SF chemotactic activity for mononuclear leukocytes was significantly inhibited by either anti-RANTES or anti–MCP-1 antibody.
RANTES is a key molecule in the pathogenesis of all onset groups of JRA, whereas MCP-1 is particularly important in systemic-onset JRA. Serum levels of these CC chemokines represent more highly sensitive markers of disease activity than conventional markers of inflammation.
Juvenile rheumatoid arthritis (JRA), the most common rheumatic disease in children, is characterized by large numbers of infiltrating leukocytes in inflamed joints. Although the molecular signals that control the recruitment of leukocytes to the joints have not been fully characterized, this process is believed to be controlled by certain chemokines (1, 2). Chemokines, a superfamily of small (8–14-kd), structurally related chemotactic cytokines, have been reported to selectively recruit and activate leukocytes at sites of inflammation (1, 3, 4). These chemokines can be divided into 2 major subfamilies, the CXC and CC chemokines. CXC chemokines such as interleukin-8 (IL-8) have been implicated in acute inflammation, since they exert their function mainly on neutrophils, whereas the CC chemokines, including RANTES and monocyte chemoattractant protein 1 (MCP-1), attract and activate a variety of cells, including monocytes, macrophages, lymphocytes, eosinophils, and basophils, and have been implicated in chronic inflammatory disease (1, 3–5).
Recent data from animal models suggest that both RANTES and MCP-1 play important roles in the pathogenesis of arthritis (2, 5–8). In adjuvant-induced arthritis in the rat, increased levels of RANTES have been found in both the blood and the joint, and synovial levels of RANTES have been found to correlate with clinical symptoms of joint inflammation (5, 6). Furthermore, administration of anti-RANTES antibody has been shown to prevent the onset of arthritis and greatly ameliorate arthritis symptoms once the disease develops (6). Similarly, elevated levels of MCP-1 have been found in both the blood and the joint in animal models of arthritis (5, 7). In addition, injection of MCP-1 antagonist was shown to markedly reduce the severity of arthritis and the infiltration of monocytes, and pretreatment with this antagonist could prevent the development of arthritis (7, 8). The pivotal role of RANTES and MCP-1 in arthritis in humans is underlined by the findings of an enhanced production of these two CC chemokines in serum and/or synovial fluid (SF) of adult patients with rheumatoid arthritis (RA) (9–12).
Taken together, these data led us to hypothesize that RANTES and MCP-1 are important in the regulation of inflammation in JRA, and that a determination of RANTES and MCP-1 profiles may be useful for monitoring disease activity in JRA patients. To test this hypothesis, we longitudinally investigated serum and SF levels of RANTES and MCP-1, as well as the in vitro migration of mononuclear cells toward the SF, in patients with JRA. We also investigated correlations of these CC chemokines with subtypes of JRA and with clinical and laboratory parameters of disease activity.
This was a 6-year longitudinal study conducted at Chang Gung Children's Hospital, a tertiary teaching medical center in Taiwan. The estimated prevalence of JRA in Taiwan is 3.8 per 100,000 children who are younger than 16 years, which is on the low end of the spectrum in previous studies throughout the world (0.83–400 per 100,000) (13).
A total of 93 patients (52 girls and 41 boys; mean age 6.9 years [age range 2–15 years]) with JRA were enrolled and prospectively studied over a 6-year period from January 1999 to December 2004. Inclusion criteria were a diagnosis of JRA according to the American College of Rheumatology criteria (14, 15) and a followup period of at least 1 year. The patients were categorized according to JRA onset type, based on the type of disease manifested during the first 6 months after onset: pauciarticular-onset JRA (n = 50 [27 girls and 23 boys]), polyarticular-onset JRA (n = 24 [15 girls and 9 boys]), and systemic-onset JRA (n = 19 [10 girls and 9 boys]). The mean age of the pauciarticular-onset group was 5.7 years, and the disease duration ranged from 6 weeks to 3.6 years. The mean age of the polyarticular-onset group was 8.8 years, and the disease duration ranged from 6 weeks to 4.3 years. The mean age of the systemic-onset group was 7.9 years, and the disease duration ranged from 6 weeks to 4.8 years.
Clinical, demographic, and laboratory characteristics of the patients were recorded at study entry. Clinical assessments were performed by the same experienced pediatric rheumatologist (J-LH) at study entry and monthly thereafter throughout the study period. The following laboratory parameters were measured at entry and bimonthly thereafter: hemoglobin value, white blood cell (WBC) count, platelet count, C-reactive protein (CRP) concentration, erythrocyte sedimentation rate (ESR), and serum ferritin level.
Active disease in the polyarticular- and pauciarticular-onset groups was defined as at least 1 joint that had symptoms of active arthritis (swelling, or if swelling was not present, limitation of movement, accompanied by pain or tenderness on motion, or heat). Active disease in the systemic-onset group was defined as at least 1 joint with active arthritis (as above) or fever of ≥38.5°C at least 4 days a week without definable infection or other identifiable source other than JRA. Patients with active systemic-onset JRA were further categorized according to the persistence or absence of systemic features, which were defined as the presence of a high spiking fever (≥38.5°C) at the time of sampling. Remission was defined as the absence of active synovitis (morning stiffness not exceeding 15 minutes, no fatigue, no joint pain, no joint tenderness, and no joint or tendon sheath swelling), as well as normal ESRs associated with normal serum CRP concentrations in the previous 3 months (16, 17). Disease relapse was defined as a recurrence of active arthritis, or systemic features for patients with systemic-onset disease, following a remission (as defined above).
Patients were treated with nonsteroidal antiinflammatory drugs (n = 93), methotrexate (n = 50), azathioprine (n = 11), hydroxychloroquine (n = 2), cyclosporin A (n = 18), prednisolone (n = 50), or etanercept (n = 2) in varying doses and combinations. In addition, 7 patients with systemic-onset JRA and 2 patients with polyarticular-onset JRA were treated with intravenous methylprednisolone and cyclophosphamide pulse therapy.
We studied a total of 19 SF samples, 12 from patients with pauciarticular-onset JRA, 4 from patients with polyarticular-onset JRA, and 3 from patients with systemic-onset JRA. SF was collected at the time of intraarticular triamcinolone injection. SF samples were kept on ice, centrifuged, and stored at −70°C until tested.
Informed consent was obtained from the patients' parents or guardians before entering the study. The study was approved by the ethics committee of Chang Gung Children's Hospital.
Control serum samples were obtained from 26 age-matched healthy children ages 2–16 years (13 girls and 13 boys; mean age 6.7 years) who were among a group attending the hospital for vaccination. Permission for drawing of extra blood during routine venipuncture was obtained from the parents or guardians of all these children.
Serum and SF samples were frozen in aliquots and stored at −70°C until tested for chemokine content. Levels of RANTES and MCP-1 were measured by sandwich enzyme-linked immunosorbent assay using commercial kits (Quantikine; R&D Systems, Minneapolis, MN) and following the manufacturer's instructions.
Chemotaxis experiments were performed in modified 48-well Boyden chambers. Briefly, PBMCs were isolated from heparinized venous blood obtained from healthy subjects immediately after sampling. Cells were isolated by flotation on Ficoll (Pharmacia, Piscataway, NJ). Mononuclear cells were washed and resuspended in RPMI 1640 medium without L-glutamine (Pharmacia) at a concentration of 2 × 106 cells/ml. The migration chambers (Neuro Probe, Cabin John, MD) had 48 wells in each unit, with a migration area of 8 mm2 per well and a bottom well volume of 25 μl. Polyvinylpyrrolidone-free polycarbonate filters with a pore size of 5 μm were used.
Twenty-five microliters of SF diluted 1:5 in RPMI 1640 medium without L-glutamine, 10−7M fMLP (Sigma, St. Louis, MO), or RPMI 1640 alone was added to the bottom wells in triplicate for each experiment. The PBMCs (50 μl) were applied to each of the top wells. Chemotaxis chambers were incubated for 1 hour at 37°C in a humidified atmosphere containing 5% CO2. The filters were removed, fixed with methanol, and stained with hematoxylin. Mononuclear cells that had migrated through to the bottom of the filters were counted in 5 random high-power fields (400× magnification). Chemotaxis was expressed as the mean number of cells per high-power field (400× magnification).
In neutralization experiments, SF was incubated with either anti-RANTES monoclonal antibody (BioSource International, Camarillo, CA), anti–MCP-1 monoclonal antibody (BioLegend, San Diego, CA), or isotype-matched control antibodies for 1 hour at 37°C. Samples were then subjected to the chemotaxis assay.
Data analysis was performed using the SPSS statistical package version 10.0 for Windows (SPSS, Chicago, IL). For continuous variables, differences among more than 2 groups were assessed by the Kruskal-Wallis test, differences between 2 unpaired groups by the Mann-Whitney test, and differences between 2 paired groups by the Wilcoxon signed rank test. Chi-square tests were used for categorical measures. Pearson's correlation coefficient was used to evaluate correlations between variables. Kaplan-Meier survival curves were used to represent the time to first relapse after clinical remission. P values less than 0.05 were considered statistically significant.
Serum levels of RANTES (mean ± SD 135.64 ± 103.69 ng/ml) and MCP-1 (437 ± 375.14 pg/ml) in patients with active JRA were both significantly higher than those in the healthy controls (18.42 ± 8.26 ng/ml [P < 0.001] and 213.42 ± 122.8 pg/ml [P < 0.001], respectively) (Figures 1A and B). When patients with active JRA were categorized according to onset type, significantly increased levels of RANTES were seen in all 3 JRA onset types as compared with the controls (94.4 ± 69.1 ng/ml in those with pauciarticular-onset, 157.21 ± 97.64 ng/ml in those with polyarticular-onset, and 229.94 ± 132.81 ng/ml in those with systemic-onset; P < 0.01 for all comparisons) (Figure 1C). Among the 3 subtypes of JRA patients, serum RANTES levels were highest in those with systemic-onset disease, with statistically significant differences as compared with polyarticular-onset (P = 0.043) and pauciarticular-onset (P < 0.001) disease.
Serum MCP-1 levels in patients with active systemic-onset JRA (mean ± SD 734.38 ± 597.95 pg/ml) were significantly higher (P < 0.001) than those in healthy controls (213.42 ± 122.8 pg/ml) and those in patients with polyarticular- or pauciarticular-onset JRA (P = 0.009 and P = 0.001, respectively) (Figure 1D). It is worth noting that, although the mean levels of MCP-1 in patients with polyarticular- or pauciarticular-onset JRA tended to be higher than those in the controls, the differences were not statistically significant. Therefore, it is clear that in the case of MCP-1, the difference between patients with active JRA and healthy controls was obviously driven by the elevated levels in patients with systemic-onset JRA.
Among patients with active systemic-onset JRA, serum MCP-1 levels were significantly higher (P = 0.039) in those who had systemic features at the time of sampling (1,076.44 ± 685.66 pg/ml) as compared with those whose systemic features had subsided (392.32 ± 177.58 pg/ml).
Followup studies during periods of active disease and clinical remission showed significantly reduced serum RANTES levels during remission in all 3 JRA subtype groups (P < 0.01 for all comparisons) (Figure 1C). Serum MCP-1 levels were also significantly decreased in patients with systemic-onset JRA during clinical remission as compared with the levels during active disease (P = 0.03) (Figure 1D). Interestingly, we found that despite clinical remission having been achieved in patients with JRA, the serum RANTES levels remained significantly higher (49.88 ± 44.26 ng/ml) than those in age-matched healthy controls (18.42 ± 8.26 ng/ml; P < 0.001), with differences being significant in the systemic-onset (P < 0.001) and polyarticular-onset (P = 0.001) groups, but not in the pauciarticular-onset group (P = 0.162). In patients who were in remission, serum RANTES levels were in the normal range (mean ± 2SD of controls) in only 49%. Similarly, serum MCP-1 levels in patients with systemic-onset JRA during clinical remission tended to be higher than the levels in the controls, and the difference was marginally significant (P = 0.048).
Of note, patients who had a relapse of disease during followup had higher levels of serum RANTES during remission (58.15 ± 44.03 ng/ml) than those who did not have a relapse during the study period (37.48 ± 42.25 ng/ml; P = 0.034). When patients were categorized according to outcome at 6 months after remission, we found that patients with sustained remission (i.e., no evidence of arthritis or systemic features 6 months after the time of sampling) had serum RANTES levels that were significantly lower than the levels in patients with a relapse within 6 months (36.79 ± 36.96 ng/ml versus 70.24 ± 48.14 ng/ml; P < 0.001).
After 6 months of followup, relapse occurred in 57.8% of patients with high serum RANTES levels during remission (>34.94 ng/ml mean + 2SD of the control group), whereas clinical remission continued in 82.5% of patients with normal RANTES levels during remission (odds ratio 6.451 [95% confidence interval 2.355–17.669]; P < 0.001) (Table 1). By Kaplan-Meier analysis, there was a statistically significant difference between these 2 groups with regard to the time to first relapse (P < 0.001) (Figure 2). Moreover, analysis of the correlation between serum RANTES levels during remission and the duration of clinical remission showed a statistically significant inverse correlation (r = −0.342, P = 0.001).
|JRA patient group||No. of patients||No. (%) with sustained remission||No. (%) with brief remission||OR (95% CI)||P|
|Normal serum RANTES||40||33 (82.5)||7 (17.5)||6.451 (2.355–17.669)||<0.001|
|High serum RANTES||45||19 (42.2)||26 (57.8)||1.0 (reference)|
At the time of the highest inflammatory activity during the observation period, as determined by maximal CRP and/or ESR values, serum levels of RANTES correlated closely with CRP concentrations, hemoglobin values, WBC counts, and platelet counts, whereas serum levels of MCP-1 correlated with only WBC counts and serum ferritin levels (Table 2). Among the 3 onset subtypes, the correlation was most significant in patients with systemic-onset JRA. No significant correlations between the RANTES or MCP-1 concentrations and either the number of active joints, the occurrence of uveitis, or antinuclear antibody positivity were found.
Elevated levels of RANTES and MCP-1 were found in 19 SF samples from patients with JRA as compared with serum levels. SF levels of RANTES in patients with systemic-onset JRA (178.56 ± 208.13 ng/ml) tended to be higher than those in patients with polyarticular-onset (88.82 ± 34.62 ng/ml) or pauciarticular-onset (81.33 ± 53.61 ng/ml) JRA, although the differences were not statistically significant. No significant differences in SF MCP-1 levels were found among the 3 JRA onset types (data not shown).
The in vitro migration of PBMCs toward SF correlated strongly with SF levels of RANTES (r = 0.700, P = 0.001) (Figure 3). A trend toward correlation was also observed for SF levels of MCP-1 in patients with systemic-onset JRA, although the small number of SF samples studied in this subgroup (n = 3) did not allow for statistical analysis. A weak correlation was found between SF levels of RANTES and hemoglobin values (r = −0.442, P = 0.066), WBC counts (r = 0.469, P = 0.05), and platelet counts (r = 0.466, P = 0.052). Correlations were not significant for SF levels of MCP-1 and the laboratory parameters of disease activity.
We next examined the contribution of SF levels of RANTES and MCP-1 to the chemotactic activity for PBMCs (Table 3). SF samples from 7 JRA patients were incubated with either anti-RANTES monoclonal antibody or anti–MCP-1 monoclonal antibody or their isotype-matched control antibodies. Preincubation with either antibody did not alter fMLP-induced chemotaxis of PBMCs. Notably, incubation with anti-RANTES antibody resulted in significant suppression of chemotactic activity for PBMCs in all samples assayed (mean 19%, range 15.8–24.6%; P < 0.05). Incubation with anti–MCP-1 antibody significantly decreased chemotactic activity for PBMCs in 6 of the 7 samples (mean 22%, range 9.6–32.1%; P < 0.05).
|Patient||JRA onset type||Anti-RANTES studies||Anti–MCP-1 studies|
|Mean no. of cells/hpf||% suppression||Mean no. of cells/hpf||% suppression|
In this longitudinal study, we investigated serum and SF levels of RANTES and MCP-1 in patients with JRA, as well as the in vitro migration of PBMCs toward SF from JRA patients. Increasing evidence implicates an important role of CC chemokines, particularly RANTES and MCP-1, in the pathogenesis of chronic inflammatory diseases (1, 3–5, 18–20). Recent data from studies of animal models and adults with RA suggest that RANTES and MCP-1 are involved in the pathogenesis of chronic arthritis (2, 5–12). Boiardi et al (9) observed high levels of serum RANTES in a series of adult RA patients during the active stage of disease, and methotrexate treatment significantly lowered the serum RANTES levels. High serum levels of RANTES after 6 months of methotrexate treatment seem to be predictive of radiologic erosions after 1 year in the same patient cohort (9). To the best of our knowledge, this is the first report of a study of RANTES levels in patients with JRA.
The results of our study demonstrate that the different onset types of JRA are all associated with significantly elevated serum levels of RANTES during active disease. After treatment, we observed a significant reduction in serum RANTES levels among the 3 subtypes of JRA when clinical remission had been achieved, suggesting the association of this chemokine with clinical disease activity. This hypothesis is supported by the significant correlations between serum RANTES levels and several conventional parameters of inflammation, including the CRP concentration, hemoglobin value, WBC count, and platelet count. Of particular interest is the observation that the serum RANTES level was equally or more strongly correlated with JRA activity than was either the CRP concentration or the ESR. This was observed at the time of clinical remission in JRA patients, when CRP and ESR values were within the range of normal, but serum RANTES levels remained significantly elevated. This finding is consistent with the recent report on an experimental model of arthritis in rats (6). In that study, Barnes et al (6) described high levels of RANTES in whole blood and joints of Lewis rat during the course of adjuvant-induced arthritis, and the high blood levels of RANTES persisted several weeks after clinical recovery.
Furthermore, in our followup study of JRA patients who fulfilled the criteria for clinical remission, a relationship was found between serum RANTES levels and the duration of clinical remission, with low levels of RANTES being associated with prolonged clinical remission and high levels with shorter clinical remission. Therefore, the elevated concentrations of RANTES during clinical remission suggest immunoactivation even in the absence of clinically apparent disease, possibly indicating an imminent flare of JRA.
The course of JRA can be highly variable. Although some patients experience complete remission and require no further treatment, a subgroup of JRA patients experience a flare of disease shortly after remission, particularly when therapeutic agents are discontinued gradually. The identification of such individuals remains a challenge for pediatric rheumatologists. The monitoring of JRA is traditionally based on clinical observations and determinations of conventional parameters of inflammation, such as the CRP concentration, ESR, WBC count, hemoglobin value, platelet count, and serum ferritin level. While the CRP and ESR values are reliable biochemical indicators that are preferred for monitoring inflammatory activity in adults with RA, neither assessment is sufficient for distinguishing among JRA patients with moderately active disease, JRA patients with inactive disease, and healthy children (21). Surprisingly, the results of our study show that it is possible to predict imminent relapse in JRA patients during clinical remission by determining serum RANTES levels. Specifically, the normalization of serum RANTES levels during clinical remission was associated with a sustained remission in 82.5% of the patients, whereas a relapse of disease occurred within 6 months in 57.8% of the patients in whom high serum RANTES levels persisted during remission.
Taken together, these findings strengthen the conclusion that in patients with JRA, serum RANTES levels reflect the presence of joint inflammation better than does either the CRP concentration or the ESR. Further study is necessary to determine whether these JRA patients with persistently high serum RANTES levels would benefit from continuing aggressive therapy until the RANTES levels return to normal.
The mechanisms underlying the involvement of RANTES in the pathogenesis of JRA remain to be identified and explained. RANTES has been strongly implicated in arthritis, chiefly because of its potent effect on chemotaxis and activation of monocytes and T lymphocytes (22, 23). Current evidence on the pathogenesis of JRA suggests that it is indeed a severe immunoinflammatory reaction in which mononuclear cells are important effector cells. Although the initiating event(s) is not clear, products of mononuclear cells along with abnormal cellular and humoral responses are responsible for the cascade of events that lead to the disease process (24). In the JRA patients in our study, SF levels of RANTES were associated with the in vitro migration of PBMCs, suggesting that RANTES plays a role in the recruitment of mononuclear leukocytes into the inflamed joints. This is supported by the significant inhibition of SF chemotactic activity for PBMCs by the addition of anti-RANTES antibody in our neutralization experiments. An analogous feature was also observed in adult RA patients by Volin et al (25). The role of RANTES in T cell and monocyte recruitment to sites of inflammation is also underlined by the high levels of expression of the chemokine receptor CCR5 (receptor for RANTES, MIP-1α, and MIP-1β) on activated monocytes and T lymphocytes observed in the synovial fluid and tissue of RA patients. This activation can induce a Th1 shift in the immune-mediated response (26). Similarly, in the joints of mice with collagen-induced arthritis an up-regulation of CCR1 and CCR5 (both are receptors for RANTES) was found in the synovium (2).
Taken together, these data indicate the important role played by RANTES in the recruitment of mononuclear cells in chronic arthritis. Of note, increased RANTES expression has been associated not only with juvenile and adult RA, but also with several inflammatory disorders, including asthma, atopic dermatitis, atherosclerosis, glomerulonephritis, endometriosis, allogeneic graft rejection, some neurologic disorders, and certain malignancies (27). These observations support the assumption that RANTES is not specific to chronic arthritis. Accordingly, RANTES might be implicated in mediating the amplification and perpetuation, rather than the initiation, of the chronic synovial inflammation present in JRA patients.
Our neutralization studies indicated that MCP-1 may be one of the major chemokines that contribute to the chemoattraction and retention of mononuclear leukocytes in the inflamed joints of JRA patients. Our examination of MCP-1 also showed that serum levels of MCP-1 were significantly elevated in patients with systemic-onset JRA, similar to the findings of the cross-sectional analysis reported by De Benedetti et al (28). However, that study did not follow up on and compare MCP-1 levels during active disease and remission. We showed that the MCP-1 levels correlated with clinical disease activity, as demonstrated by the significant reduction in the serum MCP-1 levels after treatment. We also found that serum MCP-1 levels were associated with current systemic features at the time of sampling, the WBC count, and the serum ferritin level. In addition, we observed a trend toward a correlation between SF MCP-1 levels and PBMC migration in vitro toward SF from patients with systemic-onset JRA, although the small number of SF samples studied in this subgroup did not allow for statistical analysis. These findings strengthen the conclusion that serum MCP-1 levels are related to disease severity in patients with systemic-onset JRA. Furthermore, elevated MCP-1 levels were still found during periods of remission and reflected immunoactivation, despite the absence of clinical symptoms as well as normalization of laboratory parameters of inflammation. Our study also showed that serum levels of RANTES and MCP-1 varied among different subtypes of JRA. This suggests that CC chemokines have differing roles in the different JRA subtypes and likely reflect JRA subtype heterogeneity.
The reasons for the high levels of expression of RANTES and MCP-1 in JRA remain unclear. The correlation with levels of acute-phase proteins suggests that their production is in fact cytokine driven. It is well known that the production of RANTES and MCP-1 is up-regulated by the proinflammatory cytokines IL-1 and tumor necrosis factor α (TNFα) in a variety of cell types, including cells present in arthritic joints (10–12, 25, 29). The fact that proinflammatory cytokines such as IL-1, IL-6, and TNFα have been found in elevated levels in the blood and SF of patients with JRA (30) indicates that these cytokines may lead to the synthesis and release of RANTES and MCP-1 in these patients. Further study is necessary to determine the regulation of the production of CC chemokines and the expression of corresponding genes in vivo during joint inflammation in patients with JRA.
In summary, we conclude that RANTES is a key molecule in the pathogenesis of all 3 onset types of JRA, whereas MCP-1 appears to be particularly important in systemic-onset JRA. Our findings indicate that RANTES and MCP-1 may be major chemoattractants for mononuclear leukocytes into the JRA joint. Serum CC chemokine levels represent more highly sensitive markers of disease activity than the conventional parameters of inflammation. Of particular interest is that the odds ratio for the probability of a prolonged clinical remission, rather than an early relapse within 6 months, was 6.45 for children in whom serum RANTES levels normalized during remission as compared with those who had persistently high levels of RANTES during remission. It seems reasonable to suggest that RANTES and MCP-1 may represent potential therapeutic targets in the modulation of the inflammation in JRA.