Abstract
Objective. Adipocytokines, including leptin and adiponectin, may play an important role in the pathogenesis of rheumatoid arthritis (RA). We investigated the effects of longterm therapeutic tumor necrosis factor (TNF) blockade on adipocytokine concentrations in patients with RA.
Methods. We studied 58 RA patients starting anti-TNF therapy and 58 healthy controls matched for age, sex, and body mass index (BMI). Fasting blood samples were drawn at baseline, 2 weeks, and 6 months after the start of anti-TNF therapy and serum levels of leptin and adiponectin were measured.
Results. Patients with RA had increased adiponectin (p < 0.001) and similar leptin concentrations compared with the controls. Leptin concentrations were significantly higher in patients with high BMI (p < 0.001) and correlated positively with BMI at all timepoints (r > 0.75). In contrast, serum adiponectin tended to be higher in lean RA patients and did not correlate with BMI at any timepoint. There were no clear correlations between serum concentrations of adipocytokines and disease activity (Disease Activity Score 28). Short or longterm TNF blockade alone had no influence on circulating leptin and adiponectin concentrations. Patients treated with anti-TNF and concomitant corticosteroids on a stable basis showed a significant decrease in adiponectin levels after 6 months of therapy (p < 0.025).
Conclusion. In patients with RA, chronic inflammation and its suppression during anti-TNF therapy have limited influence on plasma leptin concentrations, while significantly decreasing circulating adiponectin levels. Our findings question the suggested key role of inflammatory markers in regulating adipocytokine patterns in RA.
Leptin and adiponectin are 2 adipocyte-derived hormones that play a central role in the homeostasis of the energy and glucose metabolism, respectively1. Leptin was initially described as a hormone that regulates food intake and energy balance2. Later studies showed that leptin also stimulates T cell-mediated immunity, cytokine release from monocytes/macrophages, and the differentiation of hematopoietic cells3,4. The role of leptin as immunomodulator in humans is strongly sustained by the increased incidence of severe infections in subjects with genetic leptin insufficiency5 and by the immune system deficiency during starvation and malnutrition, when leptin concentrations are low3. Adiponectin is also synthesized by adipocytes, and one of its main actions is to improve insulin sensitivity6. Serum concentrations of adiponectin are markedly decreased in individuals with visceral obesity and states of insulin resistance7. Both hormones are also regulated by central mechanisms through the hypothalamus8. Like leptin, adiponectin can also modulate inflammatory processes. Studies have indicated that adiponectin has antiinflammatory effects, through the inhibition of nuclear factor-κB (NF-κB) activation in endothelial cells and macrophages9, inhibition of tumor necrosis factor (TNF) production and phagocytic activity of macrophages10, and by inducing production of the antiinflammatory cytokines interleukin 10 (IL-10) and IL-1 receptor antagonist (IL-1RA) by human monocytes/macrophages and dendritic cells11. Nevertheless, certain situations in which adiponectin might have proinflammatory actions have recently been reported. Accordingly, adiponectin can increase IL-6 production from endothelial cells, monocytic cells, and from synovial fibroblasts12. These effects are likely to be strongly related to its different molecular species, with low molecular weight adiponectin being antiinflammatory, whereas high molecular weight and globular adiponectin would be proinflammatory13.
Recently, in vitro4,12 and in vivo14–16 studies have suggested that leptin and adiponectin may play a role in the pathogenesis of rheumatoid arthritis (RA). They may also interfere with atherosclerosis17, which develops more frequently in patients with RA compared to the general population. To date, studies investigating the role of adipokines in RA have been mainly of cross-sectional design14,16,18–20, and little is known about how these hormones behave during the course of the disease or about the effects of therapy with antirheumatic agents, i.e., anti-TNF drugs, on the homeostasis of leptin and adiponectin21,22. Our aim was to investigate potential relations between circulating leptin and adiponectin concentrations and RA disease activity and body weight in a prospective manner. In addition, since TNF is an important determinant of the production of leptin and adiponectin, we investigated whether longterm TNF neutralization therapy modulates the circulating concentrations of these adipokines.
MATERIALS AND METHODS
Patients and controls
Fifty-eight consecutive patients with active RA and 58 healthy controls matched for age, sex, and body mass index (BMI) were enrolled for study. All patients were attending the Sint Maartenskliniek in Nijmegen, The Netherlands, and were about to start TNF-neutralizing therapy with infliximab. Patients had failed at least 2 disease-modifying antirheumatic drugs (DMARD) before starting anti-TNF. All patients fulfilled the American College of Rheumatology criteria and had active disease as defined by a Disease Activity Score23 (DAS28) > 3.2. All had given written informed consent. Patients on therapy with lipid-lowering drugs were excluded because this medication may interfere with activities of several adipocytokines. Infliximab at a dose of 3 mg/kg was administered at baseline, at 2 and 6 weeks, and every 8 weeks thereafter. Study data were collected before and 2 weeks and 6 months after the start of therapy. Stable doses of DMARD and concomitant prednisone (< 10 mg/day, n = 11) were allowed during the study. No DMARD other than methotrexate was used, except for 2 patients who took salazopyrine. Disease activity was measured using the DAS28 score before each infliximab infusion. Demographic and disease characteristics were recorded at baseline and BMI was determined at each visit. Patients were classified according to their BMI as lean or normal weight (BMI < 25), overweight (BMI 25–30), or obese (BMI > 30).
The Regional Medical Ethical Committee approved the study.
Laboratory measurements
Fasting blood samples were collected before each administration of infliximab using vacutainer tubes (Beckton Dickinson, Rutherford, NJ, USA) containing K3-EDTA (1 mg/ml). Blood was centrifuged at 3600 rpm for 8 min at 4°C, supplemented with saccharose as a cryoprotectant (final concentration 6 mg/ml), and frozen at −80°C until assay. Serum levels of leptin and adiponectin were measured using commercial ELISA kits (R&D Systems, Minneapolis, MN, USA) according to the instructions of the manufacturer.
Statistical analysis
For nonparametric within-group values, comparisons were made using the paired Wilcoxon signed-rank test, while the paired Student t test was used in the normally distributed values. Correlations between inflammatory status markers and adipocytokine levels were determined using Spearman test. Friedman’s nonparametric test for more than 2 related samples was used to test for changes in adipocytokine concentrations from baseline during the followup period. Significance was set at the level of 0.05. Values are expressed as mean ± standard deviation (SD), unless stated otherwise.
RESULTS
Characteristics of patients at baseline and changes in disease activity
Baseline characteristics of RA patients and healthy controls are presented in Table 1. As shown, patients with RA had almost 2-fold higher circulating adiponectin concentrations than controls: 4116 ± 598 ng/ml versus 2352 ± 266 ng/ml for men (p < 0.02) and 6017 ± 524 ng/ml versus 3487 ± 298 ng/ml for women (p < 0.0001) (Figure 1A), whereas leptin concentrations did not differ between RA patients and controls (Figure 1B). As expected, DAS28 and erythrocyte sedimentation rate (ESR) values were high in RA patients at baseline and decreased after the first infliximab infusion, remaining low throughout the entire followup period (p < 0.001). We tested whether disease activity and markers of inflammatory status were related to adipocytokine concentrations in patients. We found no correlations of adiponectin, leptin, or leptin adjusted to BMI with DAS28 and ESR, even after correction for patients’ sex: between DAS28 and adiponectin r = 0.19 (p = 0.26), leptin r = −0.02 (p = 0.88), leptin/BMI r = −0.30 (p = 0.11); and between ESR and adiponectin r = 0.21 (p = 0.19), leptin r = −0.06 (p = 0.68), and leptin/BMI r = −0.18 (p = 0.36). Nevertheless, a subanalysis in women with active disease (DAS28 > 3.2) revealed a trend toward an inverse relation between leptin and leptin/BMI on the one hand, and DAS28 and ESR on the other hand, at the majority of timepoints we studied (Table 2). Of note, no relation between leptin concentrations at baseline and treatment with methotrexate or glucocorticoids could be established.
Influence of BMI on adipocytokine levels
We tested the relations between BMI and adipocytokine concentrations measured at several timepoints. Using the BMI classification, 59% of patients had normal weight, 24% were overweight, and 17% were obese. There was no variation in these percentages throughout the followup. We found that leptin concentrations were significantly higher in obese patients compared with normal-weight patients, at all timepoints (Figure 2A). Although leptin concentrations were higher in women, the proportion of women and men was similar in all BMI subgroups, and therefore gender could not explain these results. In addition, in the whole RA group, serum concentrations of leptin correlated positively with BMI at all timepoints (r > 0.62, p < 0.001; Figure 2C). This relation was even more consistent when evaluated separately for men and women: r > 0.74 (p < 0.002) for men and r > 0.78 (p < 0.0001) for women. Leptin also correlated positively with BMI in the control group (r = 0.75, p < 0.0001), in both men (r = 0.69, p < 0.006) and women (r = 0.89, p < 0.0001). In contrast to leptin, serum adiponectin levels were not significantly increased in lean RA patients compared with obese patients (Figure 2B). In addition, there was no correlation between adiponectin levels in RA patients and BMI at any timepoint studied (Figure 2D), whereas a negative correlation was observed in the healthy control group (r = −0.54, p < 0.0001).
Serum adipocytokine levels during TNF blockade
Anti-TNF therapy was initiated in these patients, and the short-term and longterm effects on serum adipocytokine concentrations were assessed after 2 weeks and 6 months of medication, respectively. Serum leptin concentrations were not modified through the followup period (Figure 3A) in men or women (Table 3). Serum adiponectin concentrations dropped significantly after 6 months of anti-TNF therapy (Figure 3B). Further analysis showed that this was the case only among patients with concomitant corticosteroid therapy (n = 11) and not in the others (Figure 3B). In contrast, leptin and the leptin/BMI ratio were not affected by concurrent corticosteroid therapy (Figure 3A). Of note, no association could be found between baseline levels of adipocytokines and the response to anti-TNF therapy.
DISCUSSION
We describe, for the first time in a larger group of patients with RA, that longterm TNF blockade using infliximab did not alter circulating concentration of leptin and adiponectin, except for patients using stable dosage of corticosteroids, in whom longterm TNF blockade significantly decreased adiponectin concentrations. In addition, we found that BMI is an important determinant of leptin levels in RA patients, while inflammation and disease activity had no clear association with serum concentrations of the adipocytokines we studied.
Previous studies suggested that leptin is a proinflammatory mediator that favors the damaging processes characteristic to RA. Initially, leptin concentration was found to be higher in RA patients compared to healthy volunteers, in both the circulation and synovial fluid14, and a positive correlation with disease activity and inflammatory markers was suggested19,24. However, as shown here, chronic inflammation in patients with RA did not have stimulatory effects on serum leptin levels, in contrast with the acute inflammation of sepsis and surgery25. Accordingly, we show that RA patients and healthy controls have similar plasma concentrations of leptin, despite an increased inflammatory status in the first group. In addition, there were no clear relations between plasma leptin and inflammatory status, as assessed by disease activity and ESR. This was supported by the fact that anti-TNF therapy had no effect on plasma leptin concentrations, while decreasing the inflammation. Given these results, which are in accord with other reports from our group and others18,21,22,26,27, we argue that circulatory leptin reflects the role of leptin in RA development, as previously suggested. In line with this, plasma leptin can inversely relate to inflammation in RA22, as we observed in a subgroup of patients with active disease. In addition, the compartment where leptin is produced seem to be of importance, with locally (intraarticularly) produced leptin being likely to be more important for the pathogenesis of RA14,16. BMI, and thus most likely fat tissue, remains the major determinant of plasma leptin concentrations, which is confirmed by studies in RA18 and in other chronic inflammatory conditions6,28,29. Therefore, we hypothesize that in RA, circulating leptin does not reflect and has a limited influence on the intraarticular inflammation. We further consider that in these patients plasma leptin concentrations should be more closely related to the energy metabolism (e.g., nutritional status, cachexia) and susceptibility to infections.
In our study, plasma concentrations of adiponectin were higher in the RA patients than in controls. In contrast to the initial hypothesis that chronic inflammation associated with obesity inhibits adiponectin production, increased adiponectin levels were observed during chronic inflammatory conditions that were unrelated to increased adipose tissue mass20,24. This could be explained by the presence of inflammation-induced catabolic responses, which may raise adiponectin levels in these patients. Moreover, these levels may be positively associated with C-reactive protein concentrations24 and exert proinflammatory actions in a TNF-dependent manner, including stimulation of matrix metalloproteinase-1 synthesis in human synovial fibroblasts and monocyte chemoattractant protein-1 expression in osteoarthritis chondrocytes12,30. However, the proinflammatory activities are related only to high molecular weight and globular adiponectin, whereas low molecular weight adiponectin has antiinflammatory effects31,32. Since no previous study has assessed the presence and the ratios between these 3 forms in patients with RA, the question whether the overall effects of adiponectin in RA are pro- or antiinflammatory remains open and awaits further investigation.
In our study, therapeutic TNF blockade for 6 months was not by itself able to produce changes in circulating adiponectin levels. However, adding anti-TNF treatment to patients that already received stable doses of corticosteroids resulted in a significant decrease of adiponectin 6 months after TNF blockade was initiated. These results differ from a recent report by Härle, et al21, who described a constant decreased adiponectin concentration in patients receiving corticosteroids beginning from baseline, and who observed no additional potential of anti-TNF agents to further diminish these levels. In addition, Serelis, et al33 found in a small group of women with RA that anti-TNF agents raised adiponectin levels after 1 year of followup. A different anti-TNF agent and a longer exposure to either drug may account for these differences. The combined effects of prednisolone and TNF blockade on adiponectin may be of increased importance, since findings on the effects of corticosteroids alone on adiponectin remain contradictory, with suppressor activity observed in vitro34 and no effect in vivo35,36.
Our results suggest that in RA chronic inflammation and its suppression during anti-TNF therapy have limited influence on plasma leptin concentrations, and therefore, they call into question the importance of circulating leptin in the pathogenesis of RA. While circulating adiponectin levels were higher in RA, concentrations dropped significantly 6 months after initiation of TNF blockade in patients receiving concurrent corticosteroid therapy. Depending on which molecular form of adiponectin is diminished, this may have detrimental effects either for the disease itself or by increasing the risk for atherosclerosis and cardiovascular disease17,37 (if antiinflammatory low molecular weight adiponectin is affected), or beneficial consequences (if proinflammatory high molecular weight and globular adiponectin diminishes). Finally, our findings suggest the possibility that other mechanisms than inflammation might be of greater importance for regulating circulating adipocytokine patterns in RA.
Footnotes
-
Dr. Popa was supported by “De Drie Lichten” Foundation, The Netherlands. Dr. Netea was supported by a VIDI Grant from the Netherlands Organization for Scientific Research (NWO).
- Accepted for publication November 27, 2008.