Increasing levels of circulating Th17 cells and interleukin-17 in rheumatoid arthritis patients with an inadequate response to anti-TNF-α therapy

A total of 48 consecutive patients (39 females and 9 males; mean age ± SD 50.1 ± 13.5 years), who fulfilled the 1987 revised criteria of the American College of Rheumatology for RA [30], were evaluated before and six months after anti-TNF-α therapy. All patients remained in active disease in spite of treatment with methotrexate (MTX) and other disease-modifying anti-rheumatic drugs (DMARDs), for whom anti-TNF-α therapy was initiated based on the British Society for Rheumatology guidelines [31]. Fourteen patients received etanercept at a dose of 25 mg twice weekly and 34 patients received adalimumab at a dose of 40 mg every other week in combination with a stable dose of MTX of 7.5 to 15 mg weekly. Corticosteroids (≦10 mg/day) and non-steroid anti-inflammatory drugs (NSAIDs) were allowed but were given at stable doses for at least four weeks before and during the six-month anti-TNF-α therapy. Disease activity was assessed by the 28-joint disease activity score (DAS28) [32]. The therapeutic response was evaluated six months after anti-TNF-α therapy was started, according to the European League Against Rheumatism (EULAR) response criteria [33]. The patients were categorized into good, moderate or non-responders based on the amount of change in the DAS28 and the level of DAS28 reached. Good responders are defined as patients who have a decrease in DAS28 from baseline (ΔDAS28) of >1.2 and a DAS28 at the sixth month of < 3.2; moderate responders have either ΔDAS28 of >1.2 and a DAS28 at the sixth month of ≧3.2 or ΔDAS28 of 0.6 to 1.2 and a DAS28 at the sixth month of < 5.1; and non-responders are those who have either ΔDAS28 of < 0.6 or a DAS28 at the sixth month of ≧5.1 [33]. To obtain a better analysis, we combined good responders and moderate responders into EULAR responders. Twelve age- and sex-matched healthy volunteers (10 females and 2 males, 47.6 ± 8.1 years), who had no rheumatic disease, were used as normal controls. Blood samples were collected at baseline (before starting anti-TNF-α therapy) and six months after anti-TNF-α therapy. The Ethics Committee of Taichung Veterans General Hospital approved this study and the written consent of each participant was obtained.

In order to detect circulating Th17 cells, phycoerythrin (PE)-conjugated anti-IL-17 (eBioscience, San Diego, CA, USA) and Phycoerythrin-Cyanin 5 (PC5)-conjugated anti-CD4 (Beckman Coulter, Marseilles, France) were quantified using flow cytometry according to the manufacturer's protocol and a technique previously described [34, 35]. Briefly, aliquots of 1,000 μl of the sterile heparinized whole blood were stimulated with a combination of 25 ng/ml of phorbol myristate acetate and 1 μg/ml of ionomycin (Sigma, Deisenhofen, Germany) and cultured for one hour at 37°C in a humidified 5% CO₂ incubator. Whole blood was treated with 10 μg/ml of Brefeldin A (Sigma, Germany) to inhibit intracellular protein transport. Activated cultures of blood samples were washed in wash buffer (phosphate buffered saline, 5% foetal bovine serum, 0.1% sodium azide; Merck, Darmstadt, Germany) and then stained with 20 μl of PC5-conjugated CD4-specific monoclonal antibody (mAb) (Beckman Coulter, Marseilles, France) for 15 minutes at room temperature (RT). Erythrocytes were lysed by adding 2 ml of fluorescence-activated cell sorter (FACS) lysing solution (Becton Dickinson, Lincoln Park, NJ, USA). After five minutes of incubation, the samples were centrifuged and washed with 0.1% BSA-PBS, and subsequently fixed with 100 μl Reagent 1 (Beckman Coulter, Marseilles, France) for 10 minutes. After washing, the pellet was incubated with 100 μl Reagent 2, saponin (Beckman Coulter, Marseilles, France) for five minutes at RT in the dark. The samples were washed twice with 0.1% BSA-PBS and then incubated with PE-conjugated IL-17-specific mAb (eBiosciences, San Diego, CA, USA) for 30 minutes at RT in the dark. An isotype control IgG1-PE (eBiosciences, USA) was used for the IL-17 staining at RT in the dark. After staining, the cells were washed and immediately analysed using flow cytometry (Beckman Coulter, USA). Lymphocytes were gated on the basis of forward- and side- scatter properties and at least 10,000 CD4⁺ cells were analysed. The results were analysed using Expo32 software (Beckman Coulter, Miami, FL, USA).

Serum levels of IL-6, IL-17, IL-21, IL-23 and TNF-α were determined in 48 RA patients at baseline and after six months of anti-TNF-α therapy, and in 12 healthy controls using enzyme-linked immunosorbent assay (ELISA) according to the manufacturer's instructions (eBiosciences, USA).

Determination of the anti-CCP antibody was performed by ELISA using a commercial kit (INOVA Diagnostics Inc., San Diego, CA, USA). A result was considered positive for anti-CCP antibodies if the titer was above 20 IU/ml. Serum levels of RF-IgM were measured by nephelometry (Dade Behring Inc., Newark, DE, USA). A result was considered positive for RF when the concentration was above 15 IU/ml.

The results are presented as the mean ± SD or median (interquartile range, IQR). The non-parametric Mann-Whitney U test was used for between-group comparisons of serum levels of IL-6, IL-17, IL-21, IL-23 and TNF-α, and for the percentages of circulating Th17 cells. The correlation coefficient was obtained by the non-parametric Spearman's rank correlation test. The Wilcoxon signed rank test was used to compare the percentages of circulating Th17 cells and serum levels of Th17-related cytokines during follow-up for the RA patients after anti-TNF-α therapy. A probability of less than 0.05 was considered significant.

As illustrated in Table 1, the majority of RA patients were female and all patients had active disease (DAS28, mean ± SD, 7.22 ± 0.79) before starting anti-TNF-α therapy. After anti-TNF-α therapy, 36 (75.0%) patients were EULAR responders, including 20 (41.7%) good responders and 16 (33.3%) moderate responders, whereas 12 (25.0%) were non-responders. Although non-responders had higher erythrocyte sedimentation rates (ESR) and RF titers compared to responders, the differences did not reach statistical significance. There were also no significant differences in the baseline demographic data, positive rate of RF and anti-CCP antibodies, daily dose of corticosteroids, or the proportion of patients previously using DMARDs between responders and non-responders.

Representative examples of flow cytometric dot-plots of intracellular IL-17 staining in Th cells obtained from the PB of one RA patient and from a healthy control are shown in Figure 1 A. Significantly higher baseline frequencies of circulating Th17 cells were observed in RA patients (median 1.11%, IQR 0.50% to 2.05%) than in healthy controls (median 0.12%, IQR 0.05% to 0.18%; P < 0.001, Figure 1B). Among the RA patients, significantly higher baseline frequencies of circulating Th17 cells were observed in EULAR non-responders than in responders (P < 0.01, Figure 2A).

As shown in Figure 1, the baseline serum levels of Th17-related cytokines including IL-17, IL-6, IL-21 and IL-23, were significantly higher in RA patients than in healthy controls. Among the RA patients, significantly higher IL-17 levels were observed in non-responders than in responders (P < 0.01, Figure 2 and Table 2). Although non-responders seemed to have higher levels of IL-6 and TNF-α when compared to the levels in responders, the differences did not reach statistical significance. There was also no significant difference in the levels of Th17-related cytokines between the etanercept-treated group and the adalimumab-treated group (data not shown).

The frequencies of circulating Th17 cells were positively correlated with DAS28 scores (r = 0.294, P < 0.05), IL-17 levels (r = 0.737, P < 0.001), IL-6 levels (r = 0.347, P < 0.05), and TNF-α levels (r = 0.495, P < 0.001) in RA patients before the anti-TNF-α therapy was started. Serum IL-6 levels were also positively correlated with IL-17 levels (r = 0.303, P < 0.05) in RA patients. After six months of anti-TNF-α therapy, the frequencies of circulating Th17 cells remained positively correlated with DAS28 scores (r = 0.375, P < 0.01), IL-17 levels (r = 0.727, P < 0.001), IL-6 levels (r = 0.311, P < 0.05), and TNF-α levels (r = 0.362, P < 0.05).

As illustrated in Figure 2 and Table 2, the mean levels of circulating Th17 cells and IL-17 significantly declined after six-month anti-TNF-α therapy (1.13% vs. 0.79%; 43.1 pg/ml vs. 27.8 pg/ml; respectively, both P < 0.001), in parallel with the decrease in DAS28 (7.18 vs. 5.21, P < 0.001) and in anti-CCP titers (58.7 ± 9.6 vs. 51.8 ± 8.7 IU/ml, P < 0.005) in responders. Serum levels of IL-6, IL-21, IL-23 and TNF-α were also significantly decreased after anti-TNF-α therapy in responders. In contrast, the mean levels of circulating Th17 cells and IL-17 significantly increased (2.94% vs. 4.23%; 92.1 pg/ml vs. 148.6 pg/ml; respectively, both P < 0.05), while TNF-α levels decreased after anti-TNF-α therapy (29.2 pg/ml vs. 15.2 pg/ml, P < 0.005) in non-responders. There were no significant changes in RF titers after anti-TNF-α therapy in either responders or non-responders.