A regulatory effect of IL-21 on T follicular helper-like cell and B cell in rheumatoid arthritis

Serum samples were collected from active RA patients (n = 120) admitted to the ward of The Affiliated Drum Tower Hospital of Nanjing Medical University, from July 2010 to September 2012. All patients fulfilled the American College of Rheumatology criteria for the classification of RA and they had no other autoimmune or systemic diseases. None of these patients was pregnant or menopausal at the time of the study. Age- and sex-matched healthy controls (HC, n = 80) were obtained from the medical examination center. Serum samples were stored at -80°C until used. The study protocol was approved by the ethics committee of The Affiliated Drum Tower Hospital of Nanjing Medical University. Written informed consents were obtained from all patients and controls. Detailed clinical characteristics and laboratory features are shown in Table 1.

Serum IL-21 levels in RA patients and HC were measured by human IL-21 ELISA kits (Biolegend, San Diego, CA, USA) according to the instructions of the manufacturer. The plate was read at 450 nm and sensitivity of the ELISA kits used in the experiment was 16 pg/ml.

All patients were followed up to obtain clinical data on age, sex, disease duration, number of swollen and tender joints, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), anti-CCP antibodies, RF-IgM, RF-IgG and RF-IgA.

ESR was evaluated by the Westergren method. Values ≤ 15 mm/h for men and 20 mm/h for women were considered normal. CRP was examined by the immunonephelometry method and a value > 8 mg/l was considered positive. Anti-CCP antibody, RF-IgM, RF-IgG and RF-IgA were tested by ELISA, with normal ranges of 0 to 5 RU/ml for RF-IgM, and 0 to 20 RU/ml for RF-IgG and RF-IgA. The DAS28 was calculated as previously described [30].

Peripheral blood mononuclear cells (PBMC) were isolated from active RA patients or HC using Ficoll density-gradient centrifugation. The cell suspension was washed three times in phosphate- buffered saline (PBS) and then labeled with the following monoclonal antibodies: phycoerythrin (PE)-conjugated anti-IL21R (R&D Systems, Minneapolis, MN, USA), fluorescein isothiocyanate (FITC)- or allophycocyanin (APC)-conjugated anti-CD19, Pecy7-conjugated anti-CD40, PE-conjugated anti-CD25, FITC-conjugated anti-CD69, FITC-conjugated anti-CD4, APC-conjugated anti-CXCR5 and PerCP-Cy5.5-conjugated anti-PD-1 (BD Biosciences, Franklin Lakes, NJ, USA). For surface marker staining, cells were maintained in the dark at 4°C for 30 minutes and then washed twice in PBS. IL-21R on Tfh-like cells and B cells and activation surface markers expression of B cells were analyzed by flow cytometry.

PBMC from RA patients or HC were cultured in RPMI 1640 supplemented with 10% fetal calf serum (FCS) and antibiotics (penicillin 100 IU/ml, streptomycin 100 μg/ml; Invitrogen, Camarillo, CA, USA) in a humidified atmosphere of 5% CO₂ at 37°C. For the studies of IL-21R and B cell activation, PBMC (1*10⁶/well) were added to 96-well plates directly with or without 100 ng/ml recombinant human IL-21 (Abcam, Cambridge, MA, USA). For proliferation studies, PBMC (1*10⁶/well) were treated with 3 μg/ml anti-CD40 antibody (eBioscience, San Diego, CA, USA), 50 ng/ml rIL-4 (PeproTech Inc, Rocky Hill, NJ, USA) and 100 ng/ml rIL-21 (Abcam). For apoptosis studies, PBMC (1*10⁶/well) were stimulated with 3 μg/ml anti-CD40 antibody (eBioscience), and 100 ng/ml rIL-21 (Abcam). For differentiation to plasmablast studies, PBMC (1*10⁶/well) were treated with 3 μg/ml anti-CD40 antibody (eBioscience), 5 μg/ml F(ab)2 goat anti-human IgM (Jackson ImmunoResearch, West Grove, PA, USA) and 100 ng/ml rIL-21 (Abcam). To investigate the possible mechanism of IL-21, 10 ug/ml anti-IL-21R antibody (Biolegend) was included in cell cultures.

Serum cytokines were analyzed using Quantibody Human TH17 Array 1 (QAH-TH17-1, RayBiotech, Norcross, GA, USA), according to the manufacturer's specification. Each sample was prepared in triplicate. An Axon scanner 4000B with GenePix software was used to collect fluorescence intensities.

PBMC from RA patients and HC were labeled with 5 μM carboxyfluorescein diacetate succinimidyl ester (CFSE) (Invitrogen) in PBS for 10 minutes at 37°C. An excess of ice-cold RPMI 1640 with 10% FCS was added to the cells to quench the reaction and cells were washed extensively. CFSE-labeled cells (1*10⁶/well) were cultured according to above-mentioned methods. Following 4 days of culture, cells were collected and then stained with APC-conjugated anti-CD19 (BD Biosciences). B cell proliferation was determined by flow cytometry analysis of CFSE fluorescence intensity. To detect apoptotic cells, cultured PBMC were collected and resuspended in 100 μl of 1 × binding buffer (10 mM HEPES (pH 7.4), 140 mM NaCl and 2.5 mM CaCl₂) and stained with 5 μl of FITC-conjugated Annexin V (BD PharMingen, San Diego, CA, USA) for 15 minutes at room temperature in the dark and then analyzed by flow cytometry.

After 4 days' culture, secreted IgG and IgM in the culture supernatants were quantitated by ELISA (MABTECH, Nacka Strand, Sweden) according to the instructions of the manufacturer. PE-conjugated anti-CD138 (BD Biosciences) expression was analyzed in cultured cells using flow cytometry.

Data were summarized as means ± standard error of the mean (SEM). Statistical significance was performed by Student's t-test and the correlation coefficient between serum IL-21 levels and clinical features in RA patients were analyzed by Spearman's rank test. All statistical analyses were performed using GraphPad Prism software (Graph-Pad, San Diego, CA, USA). A P-value < 0.05 was considered significantly different.

We measured the serum IL-21 levels in RA patients (n = 120) and age-matched HC (n = 80) by ELISA and revealed that RA patients had significantly higher IL-21 levels than HC subjects (197.60 ± 32.57 vs. 59.10 ± 3.45 pg/ml, P < 0.01; Figure 1a).

Then the correlations of IL-21 concentrations with clinical activity and auto-antibody levels were analyzed in the high IL-21 group. IL-21 concentrations > 120.74 pg/ml (mean + 2SD of the HC group) were defined as high IL-21. We found that RA patients with elevated IL-21 levels had a higher DAS28 (r = 0.44, P = 0.006; Figure 1b). In addition, high IL-21 levels were positively correlated with anti-CCP antibody levels (r = 0.36, P = 0.02; Figure 1c), but not with the levels of RF-IgM (r = 0.24, P = 0.11), RF-IgG (r = 0.31, P = 0.06) and RF-IgA (r = 0.17, P = 0.29; Figure 1d).

As IL-21 was related to disease activity, we assessed whether drug treatment had any influence on the concentrations of IL-21. RA patients were divided into four groups: 1) patients taking disease modifying anti-rheumatic drugs (DMARDs) including a single DMARD and combination DMARDs; 2) patients taking corticosteroid (CS) including a single CS and CS plus non-DMARDs; 3) patients taking CS plus DMARDs, and 4) patients taking other treatments, including non-steroidal anti-inflammatory drugs (NSAIDs), traditional Chinese medicine, and patients never taking any medication. As expected, IL-21 levels were lowest in patients treated with CS plus DMARDs (133.50 ± 18.22 pg/ml; Figure 1e) and the highest levels of IL-21 were observed in group 4) who were not receiving CS or DMARDs (270.60 ± 71.93 pg/ml; Figure 1e), although this was not statistically significant.

Recent studies found that Tfh cells could provide help for B cells and allow formation of long-lived antibody responses [10, 17, 31]. First, the frequencies of circulating CXCR5⁺PD-1⁺CD4⁺ Tfh-like cells were examined. As shown in Figure 2a, the frequencies of circulating Tfh-like cells were significantly upregulated in peripheral blood (PB) of RA patients (5.71% ± 0.53% vs. 2.32% ± 0.13%, P < 0.01; Figure 2a). Serum IL-21 levels in these subjects are shown in Additional file 1a. Furthermore, the correlations of high proportions of Tfh-like cells in RA patients, with clinical activity and auto-antibody production were analyzed. The frequencies of Tfh-like cells > 4.12% (mean + 2SD of the HC group) were defined as high proportions of Tfh-like cells. We found that patients with upregulated frequencies of Tfh-like cells had a higher DAS28 (r = 0.39, P = 0.02; Figure 2b). In addition, the percentages of Tfh-like cells were positively correlated with anti-CCP antibodies (r = 0.36, P = 0.04; Figure 2c), but not with RF-IgM (r = 0.10, P = 0.58), RF-IgG (r = 0.33, P = 0.06) and RF-IgA (r = 0.12, P = 0.51; Figure 2d) among the RA patients with high frequency of Tfh-like cells.

Because the source of IL-21 was Tfh cells, Th17 cells and NK T cells, we next investigated the correlation between IL-21 levels and the percentages of Tfh-like cells in the group of RA patients with high IL-21. We found that IL-21 levels positively correlated with the percentages of Tfh-like cells (r = 0.49, P = 0.04; Figure 3a). Flow cytometry analyses showed that IL-21 was derived from both CXCR5⁺CD4⁺ T cells and CXCR5^-CD4⁺ T cells (1.17% ± 0.27% vs. 1.07% ± 0.25%, P > 0.05; Figure 3b).

To investigate a direct function of IL-21 on Tfh-like cells in RA, we examined IL-21R expression on Tfh-like cells. IL-21R expression on Tfh-like cells in RA patients was substantially augmented compared to HC (6.64% ± 0.97% vs. 3.03% ± 0.40%, P < 0.05; Figure 3c), but this phenomenon was not detected on CXCR5^-CD4⁺T cells (0.62% ± 0.11% vs. 1.00% ± 0.19%, P > 0.05; Figure 3d). Serum IL-21 levels of these subjects are shown in Additional file 1b. Given that IL-21R expression was significantly increased in PB of RA patients, we observed the functional relevance of IL-21 to IL-21R expression in RA. However, patients with elevated IL-21 levels did not have any relationship with IL-21R expression on Tfh-like cells (r = 0.28, P = 0.29; Figure 3e). We then cultured PBMC with rIL-21 in vitro and IL-21R expression was determined. Figure 3f demonstrated that IL-21R expression on Tfh-like cells remained unchanged both in RA patients (23.05% ± 2.98% vs. 24.16% ± 3.07%, P > 0.05; Figure 3f) and in HC (17.36% ± 2.19% vs. 18.59% ± 2.23%, P > 0.05; Figure 3f). This is consistent with in vivo results. Serum IL-21 levels in these subjects are shown in Additional file 1c.

First, we found that IL-21R expression was significantly upregulated on B cells in RA patients (54.88% ± 2.64% vs. 38.39% ± 2.35%, P < 0.01; Figure 4a). Serum IL-21 levels of these subjects are shown in Additional file 2a. Then we cultured PBMC with rIL-21 in vitro and IL-21R expression was determined. Figure 4b showed that IL-21R expression on B cells was significantly enhanced in response to rIL-21 in RA patients (29.11% ± 4.11% vs.38.13% ± 4.72%, P < 0.05; Figure 4b) but not in HC (19.84% ± 2.68% vs. 21.71% ± 2.95%, P > 0.05). Serum IL-21 levels in these subjects were shown in Additional file 2b.

Next, we investigated whether IL-21 was able to regulate B cell activation. We first found that the expressions of B cell activation markers (CD25: 56.34% ± 2.53% vs. 28.78% ± 2.29%, P < 0.01; CD69: 1.51% ± 0.24% vs. 0.71% ± 0.18%, P < 0.05; CD40: 87.85% ± 2.53% vs. 77.13% ± 4.12%, P < 0.05; Figure 4c) were significantly upregulated in RA patients, suggesting B cells are in an activated state in RA. Serum IL-21 levels in these subjects are shown in Additional file 2c. Thus, we supposed that B cell activation was partly due to elevated IL-21 in RA. Then we cultured PBMC in the presence and absence of rIL-21 and the expressions of those markers on B cells were examined. As shown in Figure 4d, inclusion of rIL-21 led to higher expressions of CD25, CD69 and CD40 in RA patients and this response was reversed by anti-IL-21R antibodies (CD25: 43.48% ± 2.63% vs. 61.63% ± 2.63% vs. 57.74% ± 2.89%, P < 0.05; CD69: 1.48% ± 0.26% vs. 2.85% ± 0.43% vs. 2.30% ± 0.48%, P < 0.05; CD40: 73.14% ± 1.09% vs. 81.77% ± 1.77% vs. 75.22% ± 1.64%, P < 0.05; Figure 4d). However, rIL-21 only slightly upregulated CD25 expression on B cells in HC and had no effect on the expression of CD69 and CD40 (CD25: 26.92% ± 1.82% vs. 34.71% ± 2.50% vs. 29.14% ± 1.76%, P < 0.05; CD69: 0.73% ± 0.23% vs. 0.80% ± 0.22% vs. 0.79% ± 0.21%, P > 0.05; CD40: 70.37% ± 1.30% vs. 71.83% ± 1.63% vs. 70.16% ± 1.42%, P > 0.05; Figure 4d). Serum IL-21 levels in these subjects are shown in Additional file 2d. To further explore why IL-21 exposure was prone to activating B cells in RA, we performed microarray analysis of serum from RA patients. We found that several kinds of inflammatory factors were increased (Figure 4e), suggesting an inflammatory microenvironment in RA. Therefore, it indicates that under an inflammatory microenvironment in RA patients, IL-21 is sensitive to activation of B cells.

As previous reports suggested that IL-21 could co-stimulate mature B cells to proliferate in mice [32, 33], we studied whether it could promote B cell proliferation in RA patients. Figure 5a demonstrates that in the presence of anti-CD40 and rIL-4, rIL-21 had more effect on B cell proliferation in RA, evidenced by dilution of CFSE. This effect was reversed by anti-IL-21R antibodies in both RA patients (41.65% ± 3.00% vs. 52.82% ± 4.03% vs. 46.50% ± 2.72%, P < 0.05; Figure 5a) and HC (28.87% ± 3.66% vs. 38.12% ± 2.66% vs. 33.13% ± 3.59%, P < 0.05; Figure 5a). Serum IL-21 levels in these subjects are shown in Additional file 2e. Unexpectedly, IL-21 failed to influence B cell apoptosis in RA patients (24.41% ± 4.82% vs. 24.30% ± 4.89%, P > 0.05; Figure 5b), whereas it was able to enhance B cell apoptosis in HC (10.21% ± 1.06% vs. 15.37% ± 2.29%, P < 0.05; Figure 5b). These data indicate different effects of IL-21 on B cell apoptosis in RA patients and HC. Serum IL-21 levels in these subjects are shown in Additional file 2f.

As large amounts of RF and anti-CCP antibody appeared in the serum of RA patients, we next observed the effect of IL-21 on the production of IgG and IgM by B cells in RA patients and HC. We showed that rIL-21 upregulated the frequencies of plasmablasts (CD138⁺CD19^low/+) both in RA patients (7.30% ± 1.01% vs. 9.24% ± 1.08%, P < 0.05; Figure 5c) and HC (3.26% ± 0.66% vs. 4.12% ± 0.93%, P < 0.05; Figure 5c). However, IL-21-mediated B cell differentiation was stronger in RA patients than in HC. Furthermore, IL-21 induced B cells to produce higher levels of IgG and IgM in RA patients (IgG: 297.20 ± 24.77 vs. 353.70 ± 29.30 ng/ml, P < 0.05; IgM: 12.43 ± 4.39 vs. 20.47 ± 9.73 ng/ml, P > 0.05; Figure 5d) compared to HC (IgG: 221.60 ± 18.19 vs. 258.50 ± 24.51 ng/ml, P < 0.05; IgM: 3.29 ± 0.18 vs. 3.03 ± 0.19 ng/ml, P > 0.05; Figure 5d), although there were no marked differences in the secretion of IgM. Serum IL-21 levels in these subjects are shown in Additional file 2g.