To investigate the impact of STAT-3–mediated regulation on Th17 differentiation in patients with rheumatoid arthritis (RA).
To investigate the impact of STAT-3–mediated regulation on Th17 differentiation in patients with rheumatoid arthritis (RA).
CD4+ T cells isolated from peripheral blood (PB) and synovial fluid (SF) were stimulated to differentiate into Th17 cells or Treg cells. The activity of STAT-3 was knocked down by transfecting CD4+ T cells with small interfering RNA (siRNA). After 3 days in culture, the proportions of Th17 cells and Treg cells were measured by flow cytometry, and the production of interleukin-17 (IL-17) was measured by reverse transcriptase–polymerase chain reaction and enzyme-linked immunosorbent assay.
The levels of IL-17, IL-6, IL-23, IL-1, and tumor necrosis factor α were significantly higher in RA SF and synovial tissue than in SF and synovial tissue from osteoarthritis patients. In RA synovial tissue, the expression of STAT-3 increased in proportion to the severity of synovitis, as shown by stromal cellularity, intimal hyperplasia, and inflammatory infiltration. The degree of Th17 differentiation was highest in RA SF, followed by RA PB, and lowest in normal subjects. In CD4+ T cells, transfection with STAT-3 siRNA prevented Th17 differentiation of mononuclear cells from RA PB and SF but increased the proportion of Treg cells. In contrast, inhibition of STAT-5, the transcription factor for Treg cells, increased the proportion of Th17 cells and reduced that of Treg cells.
Our findings indicate that modulation of STAT-3 in CD4+ T cells affects the differentiation of Th17 cells and Treg cells in patients with RA. This role of STAT-3 in RA synovial T cells may provide a new therapeutic target for the management of RA.
CD4+ T cells can differentiate into various subsets, including both pathogenic cells and suppressors of inflammation in autoimmune diseases. Earlier understanding of T cell differentiation led to the classification of rheumatoid arthritis (RA) as a Th1-mediated disease (1–3) until the recent acceptance of Th17 cells as the major players in RA pathogenesis (4, 5). The pathogenic role of Th17 cells has been elucidated by characterizing their production of interleukin-17 (IL-17), which facilitates inflammation. An increase in the proportion of Th17 cells in peripheral blood (PB) is compensated for by decreased Treg cell differentiation in animal experiments only. Overall, Th17 and its cytokine IL-17 cause a constellation of symptoms of RA, which range from aggravation of synovitis (6) and perichondral damage (7) to bone destruction (8). However, the exact mechanism responsible for Th17-derived pathogenicity has not been clearly defined in human autoimmune diseases.
Genetic mutations of STAT-3 in patients with hyperimmunoglobulinemia E syndrome first suggested the involvement of this transcription factor in the differentiation of Th17 cells (9–12); however, information about the pathogenesis and peripheral differentiation of Th17 cells in human autoimmune diseases is limited. In RA, Th17 differentiation requires IL-6, an inflammatory cytokine whose level increases significantly in RA patients (13). IL-21 has also been shown to differentiate CD4+ T cells into Th17 cells (14). Interestingly, both IL-6 and IL-21 act via the JAK/STAT pathway, suggesting that the downstream activation of STAT-3 is the merging point of these proinflammatory effectors in the process of Th17 differentiation.
In this study, we measured the levels of cytokines responsible for Th17 differentiation in RA, and we compared the efficiency of these cytokines in inducing Th17 differentiation from CD4+ T cells in RA patients and in healthy individuals. Using small interfering RNA (siRNA)–mediated inhibition, we also investigated the functional outcome of disrupting STAT-3 and STAT-5, two transcription factors known to control the differentiation of Th17 cells and Treg cells, respectively. Our results suggest that STAT-3 plays a pivotal role in the process of determining T helper cell differentiation in RA T cells.
Anti-CD3 (clone OKT3) and anti-CD28 (clone CK248) were purchased from BD PharMingen. Phorbol myristate acetate (PMA) and ionomycin were purchased from Sigma-Aldrich. Anti–IL-17, anti-FoxP3, anti-CD4, anti-CD25, anti-CD45RO, and anti–interferon-γ (anti-IFNγ) were purchased from eBioscience.
Cytokine levels were analyzed by enzyme-linked immunosorbent assay (ELISA) in synovial fluid (SF) and PB samples from patients with RA (n = 31) who fulfilled the American College of Rheumatology (ACR) 1987 revised criteria (15) and from patients with osteoarthritis (OA) (n = 25) from the outpatient clinic of the Department of Rheumatology, Seoul St. Mary's Hospital. To be eligible for inclusion, OA patients had to be diagnosed as having primary knee OA according to the ACR criteria (16). Mononuclear cells from the SF and PB were obtained from an additional 10 RA patients who fulfilled the ACR 1987 revised criteria. Ten age- and sex-matched healthy donors were included as normal controls. Informed consent was obtained from all patients and healthy donors, and the experimental protocol was approved by the Catholic University of Korea Human Research Ethics Committee.
Peripheral blood mononuclear cells (PBMCs) were isolated from healthy donors and RA patients by Ficoll-Hypaque (Amersham Pharmacia Biotech) density-gradient centrifugation. The cells were washed 3 times with sterile phosphate buffered saline and resuspended in RPMI 1640 (Life Technologies) supplemented with 10% fetal bovine serum, 2 mM L-glutamine, and 1% penicillin–streptomycin (“complete medium”). Naive or memory CD4+ T cells were further purified using a human naive or memory CD4+ T cell isolation kit (Miltenyi Biotec). The purity of CD45RA+CD4+ cells was nearly 90%. Naive or total CD4+ T cells were activated with plate-bound anti-CD3 (1 μg/ml) and soluble anti-CD28 (1 μg/ml) together with antigen-presenting cells (APCs). APCs were prepared from human PBMCs that did not contain lymphocytes after 5,000 rads irradiation. CD4+ T cells were induced to differentiate into Th0 (anti-CD3 and anti-CD28) and various other subtypes by stimulation with the appropriate cytokines; for example, Th17 cells were induced by stimulation with 10 ng/ml IL-6, 10 μg/ml anti-IFNγ, 10 μg/ml anti–IL-4, 5 ng/ml IL-23, and 2 ng/ml transforming growth factor β (TGFβ), and Treg cells were induced by stimulation with 20 ng/ml TGFβ and 5 ng/ml IL-2. Th17 cells were also induced with an additional 5 ng/ml tumor necrosis factor α (TNFα) and 10 ng/ml IL-1β in some experiments. After replacing half of the media on day 3, cells were restimulated with half the amount of cytokines and cultured continuously until day 6. TGFβ was purchased from PeproTech, and other stimulating cytokines were purchased from R&D Systems.
Cytokine-producing cells were identified by intracellular staining using fluorescein isothiocyanate (FITC)–conjugated anti-IFNγ, FITC-conjugated FoxP3, phycoerythrin-conjugated anti-human IL-17, PerCP-conjugated anti-human CD4, allophycocyanin-conjugated anti-human CD45RO, and allophycocyanin-conjugated anti-CD25. Briefly, cells were stimulated with 50 ng/ml PMA and 500 ng/ml ionomycin for 5 hours, and GolgiStop (BD Biosciences) was added. Cells were fixed in Cytofix/Cytoperm, permeated with 0.1% saponin, stained with fluorescent antibodies, and analyzed on a FACSCalibur flow cytometer (BD Biosciences). CellQuest software (BD Biosciences) was used for data acquisition, and FlowJo software version 4.5 (Tree Star) was used for analysis. IL-22 secretion was measured using a human IL-22 ELISA development kit from PeproTech. Secretion of IL-17, IFNγ, IL-4, TGFβ, and IL-10 in the culture supernatants was measured by sandwich ELISA (R&D Systems).
We used a 3-component score to grade the histologic severity of synovitis. Each of the 3 components (intimal hyperplasia, stromal cellularity, and inflammatory infiltration) was graded on a scale of 0–3, yielding a final score of 0–9 (17). Synovitis was scored by an independent pathologist. Images of RA synovium immunostained with STAT-3 were analyzed using the TissueFAXS system (TissueGnostics). All images were analyzed using the analysis software HistoQuest (TissueGnostics), which allows quantification of the total cell number based on color separation and a nucleus detection algorithm. The algorithm used measures the intensity around each individual nucleus in an area of variable size preset by the user. All values representing each individual cell were plotted automatically in a scatterplot based on the values of cell size (y value) and the intensity of staining (x value).
CD4+ T cells were incubated for 5 days with various concentrations of Th0, Th1, Th2, Th17, and Treg cells. After incubation, mRNA was extracted using RNAzol B according to the recommendations of the manufacturer (Biotecx). Reverse transcription of 2 μg of total mRNA was performed at 42°C using the SuperScript reverse transcription system (Takara). PCR amplification of complementary DNA aliquots was performed by adding 2.5 mM deoxyribonucleotide phosphates, 2.5 units Taq DNA polymerase (Takara), and 0.25 μM sense and antisense primers. The reaction was conducted in PCR buffer (1.5 mM MgCl2, 50 mM KCl, and 10 mM Tris HCl, pH 8.3) in a total volume of 25 μl, and processed in a DNA thermal cycler (PerkinElmer Cetus). PCR products were run on a 2% agarose gel and stained with ethidium bromide. The results are expressed as the ratio of target PCR product relative to that of β-actin.
Reverse transcription of corresponding mRNAs was performed as in conventional RT-PCR. PCR amplification and analysis were conducted using a LightCycler 2.0 instrument with version 4.0 software (Roche). All reactions were performed with LightCycler FastStart DNA SYBR Green I Master Mix (Roche). Amplification conditions comprised an initial preincubation at 95°C for 10 minutes, followed by amplification of the target DNA for 45 cycles at 95°C for 10 seconds, at target annealing temperatures for 10 seconds, and at 72°C for 10 seconds. Melting curve analysis was performed immediately after amplification at a linear temperature transition rate of 0.1°C from 65°C to 95°C. The results were systematically normalized to the expression levels of the reference genes.
Human STAT-3 and STAT-5 siRNA were designed by Dharmacon. STAT-5 siRNA was designed to target 2 forms of STAT-5: STAT-5A and STAT-5B. CD4+ T cells were purified from human PBMCs, plated in 24-well plates, and transfected with 100 nM siRNA or 100 nM nonspecific control siRNA conjugated with Alexa Fluor 488 (Qiagen). The transfection efficiency was monitored using HiPerFect transfection reagent, according to the recommendations of the manufacturer (Qiagen). The cells were incubated with siRNA for 30 minutes and then stimulated with a combination of cytokines to induce Th17 and Treg cell differentiation for 48 hours, after which APCs were added. CD4+ T cells were activated with plate-bound anti-CD3 and anti-CD28 for 6 days, after which the T cells were restimulated with PMA and ionomycin, and then harvested for analysis.
Data are expressed as the mean ± SEM. Data were analyzed using Student's t-test for matched pairs. Analysis of variance was used to analyze multiple mean values. P values less than 0.05 were considered significant.
We used ELISAs to measure the concentrations of IL-17 signature cytokines in RA SF, OA SF, and RA PB. IL-17 levels were higher in RA SF than in OA SF (Figure 1A). IL-17, IL-21, and IL-22 cytokines produced by Th17 cells were preferentially expressed in RA synovium (Figure 1B). IL-6 levels were higher in RA SF than in OA SF (Figure 1C). In samples from RA patients, IL-6 levels were much higher in the SF than in PB (P < 0.01). IL-6 and other cytokines that may induce Th17 differentiation (IL-1β, IL-23, and TNFα) also stained positively in RA synovium (Figure 1D).
The proportion of IL-17–producing CD4+ T cells was greater in RA SFMCs than in RA PBMCs. However, the Th17 cells were too infrequent in the untreated state for statistical analysis (Figure 2A). Upon stimulation with 10 μg/ml anti-CD3 or 25 μg/ml PMA, the proportion of Th17 cells increased by 2-fold in both RA PBMCs and RA SFMCs (Figures 2A and B). When stimulated to induce Th17 differentiation, RA SFMCs produced significantly more IL-17 than did RA PBMCs (Figure 2C).
The proportion of phosphorylated STAT-3 protein was higher in RA synovial T cells than in RA PBMCs or normal PBMCs (Figure 3A). Immunostaining showed that the expression of both IL-17 and STAT-3 was more prominent in synovial tissue from RA patients than in that from OA patients (Figure 3B). We measured STAT-3 production in additional RA synovia obtained from the tissue bank in the Seoul St. Mary's Hospital Department of Pathology. The brown staining of STAT-3 was measured with a computer program that automatically calculates the areas positive for the target molecule. In this analysis, STAT-3 expression correlated positively with the synovitis score (Figures 3C and D). There were fewer IL-17–positive spots than STAT-3–positive spots in the RA synovium (Figure 3E). Dual immunostaining for CD4 and STAT-3 showed double-positive cells scattered in the RA synovium. A similar distribution was observed for cells that were double-positive for CD4 and IL-17. To examine the localization of cells with phosphorylated STAT-3, RA SFMCs were concentrated by cytospin and analyzed by confocal microscopy. The distribution of STAT-3 correlated with the distribution of CD4+ T cells and IL-17 (Figure 3F).
To assess the role of STAT-3 in the differentiation of Th17 cells and Treg cells, we used siRNA-based knockdown of STAT-3 mRNA. Transfection with STAT-3 siRNA significantly decreased the proportion of IL-17–producing cells among the CD4+ T cell population isolated from RA PB and RA SF (Figures 4A and D). In contrast, inhibition of STAT-3 increased the proportion of cells positive for FoxP3, the key transcription factor for Treg cell differentiation.
We also examined the expression of mRNA for STAT-3, the Th17 transcription factor retinoic acid receptor–related orphan nuclear receptor c (RORc), and FoxP3 in STAT-3–knockdown cells. No significant changes in STAT-3, RORc, or FoxP3 expression were observed in unstimulated Th0 cells, but STAT-3 siRNA exerted opposing effects on the expression of these mRNAs in cells stimulated toward Th17 differentiation (Figures 4B and E). STAT-3 siRNA suppressed the expression of RORc and STAT-3 in CD4+ T cells from RA PB and RA SF and increased the expression of FoxP3 in CD4+ T cells from RA PB.
To investigate the effect of STAT-3 inhibition on the differentiation of Th17 subsets, we measured IL-17, IL-22, and IL-10 production by Th17 cells. IL-17 and IL-22 production were decreased by STAT-3 inhibition in Th17 cells from both RA PB and RA SF (Figures 4C and F). The production of IL-10 remained at the basal level in Th17 cells and was not changed significantly by STAT-3 knockdown.
We next compared the changes in the proportion of Th17 and Treg cell subtypes among CD4+ T cells isolated from RA SF, RA PB, and normal PB. STAT-3 siRNA decreased the proportion of IL-17–producing Th17 cells in all 3 samples and had the greatest effect in cells from RA SF (Figure 5A). In contrast, STAT-3 inhibition significantly increased FoxP3 expression (Figure 5B). STAT-3 siRNA significantly decreased the mean fluorescence intensity (MFI) of IL-17 in RA SF (P < 0.05) (Figure 5C). STAT-3 siRNA increased the MFI of FoxP3 in the 3 compartments of normal PB, RA PB, and RA SF (P < 0.05) (Figure 5D). Transfection of CD4+ T cells with STAT-3 siRNA prevented Th17 differentiation in RA PBMCs and RA SFMCs, whereas the proportion of Treg cells was increased.
STAT-5 is required for the induction of FoxP3 expression in Treg cells. We examined the effect of suppressing STAT-5 on the differentiation of the IL-17–producing or FoxP3-producing population in CD4+ T cells in the PB of RA patients and healthy controls. Isolated CD4+ cells were stimulated with a combination of cytokines to induce the differentiation of Th17 cells (TGFβ plus IL-6 plus IL-23) or Treg cells (TGFβ plus IL-2). The proportions of IL-17–producing and FoxP3-producing cells changed in opposite directions after treatment with 2 kinds of STAT-5 siRNA (STAT-5A and STAT-5B). The proportion of IL-17–producing cells increased in CD4+ cells stimulated toward Th17 and Treg cell differentiation, but the proportion of FoxP3-positive cells decreased in both conditions (Figures 6A and B).
Finally, we examined the expression of RORc and FoxP3 after treatment with 2 kinds of STAT-5 siRNA (STAT-5A and STAT-5B). Suppression of STAT-5 decreased the expression of FoxP3 mRNA in Treg cells but increased that of RORc in Th17 cells (Figure 6C). STAT-5 knockdown increased the production of IL-17 and IL-22 in Th17 cells but decreased the amount of IL-10 in Treg cells (Figure 6D). No significant changes in the production of Th1 versus Th2 signature cytokines were induced by STAT-3 or STAT-5 siRNA (data not shown).
The outdated concept of Th1 and Th2 skewing was previously used to categorize RA as a typical Th1 disease. However, recent identification of diverse subsets of T cells, including Th17 cells (4, 5), Treg cells (18, 19), and follicular helper T cells, suggested that Th17, rather than Th1, is the pathogenic culprit in RA. Treg cells are believed to play the role of suppressing pathogenic processes associated with RA. Recent studies suggest that commitment to each subtype is not irrevocable, and interconversion between Th17 cells and Treg cells is possible (20).
Unfortunately, the conditions for T helper cell differentiation in patients with RA are poorly understood and are more controversial than those in animal models of arthritis. For example, TGFβ is crucial for Th17 differentiation in a mouse model (4), but contradictory effects of TGFβ, ranging from stimulation to inhibition, have been reported in humans (21, 22). Researchers recently agreed to designate IL-6 and TGFβ as key factors for Th17 differentiation. It is believed that IL-2 and TGFβ together can induce naive T cells to differentiate into Treg cells and that the addition of TGFβ together with IL-6 leads to Th17 differentiation.
It is intriguing to note that in the present study, some cells stained positive for both IL-17 and FoxP3 in both the Th17 and Treg cell subsets. These double-positive cells obtained from RA SF were observed by confocal microscopy. Although these cells expressed both IL-17 and FoxP3, the degree of expression varied between cells (data not shown). The presence of such double-positive cells may be clinically significant for several reasons. First, even though Th17 cells and Treg cells have opposite functions, their developmental and differentiation processes may be closely linked. The existence of cells expressing both IL-17 and FoxP3 renders a theoretical rationale for the regulation of Th17 cell and Treg cell homeostasis by a single cellular signal transduction.
Second, the greater number of cells in a transitional state may explain why a single cytokine may be sufficient for Th17 differentiation. Considering that the IL-6 level is 100 times higher in RA SF than in serum, this characteristic of the RA intraarticular environment may favor Th17 differentiation. We believe that the increased expression and phosphorylation of STAT-3 in RA synovial T cells is triggered by the elevated concentration of IL-6.
Third, if transitional cells that display characteristics of both Treg cells and Th17 cells do exist, one may postulate that Treg cells, which are reactive for autoantigens, may play a key pathogenic role by differentiating into Th17 cells.
The intraarticular environment of the inflamed RA joint has abundant IL-1 (23) and TNFα (24), and these cytokines may act as the instigators to skew T cell differentiation toward Th17. Intriguingly, comparison of the levels of cytokines induced by IL-1β and/or TNFα between PB and synovium from an RA patient showed that IL-6 concentration was >10-fold higher in the synovium than in PB. In contrast, the synovial level of TGFβ in an RA patient was significantly lower than the circulating level. With regard to migrating T cells, these observations indicate that circulating CD4+ T cells exposed to high levels of TGFβ would encounter a marked decrease in TGFβ concentration upon entering the joint space in RA. This hypothesis is also compatible with the findings of a recent study showing that environments with a decreased TGFβ signal favor Th17 differentiation (22). Within the intraarticular milieu, the direct effect of TGFβ would be diminished, and instead, the alternative increase in IL-6 concentration would stimulate CD4+ T cells and transmit signals required for Th17 differentiation.
Upon IL-6 stimulation, the target cell launches an intracellular signal transduction cascade via the JAK/STAT pathway, leading to the activation of STAT-3. STAT-3 has been described as a critical transcription factor that regulates the expression of diseases related to chronic cell proliferation, such as cancer (25). The activation of STAT-3 is also involved in the pathogenesis of RA by steering the abnormal activation, automaticity, and prolonged survival of synovial cells. The increased expression of STAT-3 in RA synovium was also evidenced in this study, and the severity of synovitis was positively related to STAT-3 expression. Strong STAT-3 expression was observed in both CD4+ T cells and synovial cells, suggesting that STAT-3 activation endows synovial cells with tumor-like characteristics, allowing their hyperproliferation, prolonged survival, and the infiltration of surrounding joint tissue. Our results suggest that the counterdifferentiation of Th17 cells and Treg cells via STAT-3 regulation further strengthens the possibility of Th17–Treg cell conversion during T helper cell differentiation. Instead of categorizing T cells into distinct subsets, a more accurate understanding of T cells might be to describe them on a spectrum, with Th17 cells and Treg cells on either polarity, in which the T cell skewing reflects the gradient of STAT-3 expression.
Both Treg cells and Th17 cells produce a distinct panoply of transcription factors related to their differentiation (4, 5). For example, STAT-3 and RORc are involved in Th17 differentiation, whereas STAT-5 and FoxP3 are responsible for the differentiation of Treg cells (26). Our study showed that the inhibition of STAT-3 increased the signals from TGFβ and STAT-5, leading to the increased differentiation of Treg cells. This pattern of differentiation is easily deducible from previous findings. This observation suggests the existence of cross-talk between STAT-3 and STAT-5. There is accumulating evidence of possible cross-talk between downstream signals of TGFβ and IL-6 (27). For example, FoxP3-positive T cells inhibit Th17 differentiation via STAT-3 regulation (28), and TGFβ is down-regulated by the STAT-3–mediated inhibition of the Smad pathway or by the direct binding of FoxP3 to RORc (29–31). The conventional concept of a duel between Th1 and Th2 is now being replaced by the concept of the interaction of the multiple players Th1, Th2, Th17, and Treg cells.
Th17 cells are difficult to identify because they comprise a small subset of CD4+ T cells under physiologic conditions. In this study, we assessed the effects of STAT-3 modulation in cultured Th17 cells by priming them with a combination of signature cytokines known to trigger Th17 differentiation. Such treatment is equivalent to the intraarticular environment of RA synovium. RA synovial T cells exposed to elevated levels of IL-6, TNFα, and IL-1β would have a greater potential toward Th17 differentiation. Interestingly, siRNA-mediated inhibition of STAT-3 also increased the activity of STAT-5 and FoxP3. The opposing effects of STAT-3 and STAT-5 have been observed in animal models, but our result is the first to report such duality in human disease. It would be intriguing to investigate in detail the process of mutual regulation between STAT-3 and STAT-5, particularly how knocking down STAT-3 is related to the activation of STAT-5. Identification of the signal pathways involved in this process would provide novel target molecules for the control of RA joint inflammation and for innovative customized therapies.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Kim had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Ju, Heo, Cho, Moon, Kwok, K.-S. Park, S.-H. Park, Kim.
Acquisition of data. Ju, Heo, Cho, Jhun, J.-S. Park, Lee, Oh.
Analysis and interpretation of data. Ju, Heo, Cho, Lee, Oh, Moon, Kwok, K.-S. Park, S.-H. Park, Kim.