Negative impact of Interleukin-9 on synovial regulatory T cells in rheumatoid arthritis

Abstract

In Rheumatoid Arthritis (RA), regulatory T cells Tregs) have been found to be enriched in the synovial fluid. Despite their accumulation, they are unable to suppress synovial inflammation. Recently, we showed the synovial enrichment of interleukin-9 (IL-9) producing helper T cells and its positive correlation with disease activity. Therefore, we investigated the impact of IL-9 on synovial Tregs in RA. Here, we confirmed high synovial Tregs in RA patients, however these cells were functionally impaired in terms of suppressive cytokine production (IL-10 and TGF-β. Abrogating IL-9/IL-9 receptor interaction could restore the suppressive cytokine production of synovial Tregs and reduce the synovial inflammatory T cells producing IFN-γ, TNF-α,IL-17. However, blocking these inflammatory cytokines failed to show any effect on IL-9 producing T cells, highlighting IL-9’s hierarchy in the inflammatory network. Thus, we propose that blocking IL-9 might dampen synovial inflammation by restoring Tregs function and inhibiting inflammatory T cells.

1. Introduction

Rheumatoid arthritis (RA), an inflammatory bone disease is characterized by chronic synovitis with progressive joint destruction [1,2]. Enrichment of T helper (Th)-1 (IFN-γ+, TNF-α+) and Th-17 (IL-17+) cells in the synovial fluid (SF) is well documented in literature and plays a pivotal role in synovitis [3,4]. Another crucial subset of Th cells are regulatory T (Treg) cells, which play an important role in restraining the chronic or aberrant inflammation, thus preventing the bystander tissue damage in the autoimmune diseases including RA [5]. Altered Treg cell number and/or their functional impairment have been ascribed to the pathogenesis of autoimmune disease [6,7]. Adoptive transfer of Treg cells have been shown to ameliorate inflammation in the settings of experimental models of various autoimmune diseases like lupus, colitis etc. [8–12]. In RA, conflicting reports are available with respect to altered number or impaired function of Treg cells as a pathogenic contributor [13−16]. Most of the studies in RA on Treg cells have analysed circulatory Treg cells in the periphery which do not reflect the status of Treg at the RA pathological site. It has become apparent that the ability of Treg cells to dampen the inflammatory T cells is modified not only due to altered ratio of Treg / Teff cells, but due to functional impairment of locally accumulated Treg cells by inflammatory cytokine milieu. Multiple cytokines including IL-6, TNF, IL-17 and IL-21 have been implicated in T cell impairment [14,17–18].

Inflammatory cytokines are central to the pathogenesis of RA, however, plausible impairment of Treg cell function have also been implicated in mediating chronic inflammation. Therefore, understanding i) the role of locally accumulated T cells producing inflammatory cytokine ii) functional incapacitation of Treg cells and iii) their impact in perpetuating the inflammatory process is of paramount importance.

Recently, we and others have observed higher frequency of Interleukin-9 (IL-9) producing Th-9 cells in the peripheral blood (PB) especially synovial fluid (SF) of RA patients, which correlated with the disease activity status [19–21]. IL-9, a member of the common gamma chain (γc) receptor cytokine family, mediates pleiotropic effect by binding through a heterodimeric IL-9 receptor [22–23]. We previously demonstrated that IL-9 facilitates the survival of neutrophils, enhances osteoclastogenesis and the function of inflammatory T cells producing IFN-γ, TNF-α and IL-17 in RA patients indicating its role in the pathogenesis [19,24]. Also, studies in animal models and patients have revealed a pathogenic role for IL-9 in various autoimmune diseases [22]. However, the impact of synovial IL-9 of RA patients on the functional impairment of locally accumulated Treg cell remains poorly understood. Here, we attempted to dissect out the impact of IL-9 on the FOXP3+ Treg cells and its implication in the pathogenesis.

In this study, we demonstrate enrichment of Treg cells in the synovial fluid (SF) in RA patients relative to peripheral blood (PB). However locally enriched Treg cells are functionally impaired leading to their inability to suppress inflammation in RA. We also show a higher frequency of IL-9R expressing T cell subsets in SF compared to PB. Blocking the IL-9 pathway restored the function of SF derived Treg cells resulting in inhibition of inflammatory T cells producing IFN-γ, TNF-α and IL-17; whereas supplementation of IL-9 dampened the suppressive cytokine production of Treg cells. Taken together our results strongly indicate that inhibiting IL-9 may rescue the Treg cell function, thus suppressing the inflammatory T cells in RA. We propose that blocking IL-9 axis might constitute an important therapeutic strategy in clinical management of RA.

2. Method

2.1. Study subjects

The study was cross-sectional and was conducted with approval from the Institute Ethics Committee (IEC-490/01.09.2017). After obtaining informed consent, 60 Rheumatoid Arthritis (RA) and 10 Osteoarthritis (OA) patients were recruited from Rheumatology and Orthopaedics OPD of AIIMS, New Delhi. Diagnosis of RA patients was made on the basis of ACR criteria 1987. All recruited RA patients were assigned active disease if the DAS is >3.2. For this study, we collected synovial fluid (SF) and autologous peripheral blood (PB) from RA patients during RA flares (medical visit due to worsening of disease activity). Clinical details are summarized in Supplementary Table la. OA was diagnosed clinically and radiologically if symptoms of joint pain, swelling and limitation of movements are present with narrowing of joint space with negative result for Rheumatoid factor and anti-cyclic citrullinated peptide (CCP) autoantibodies. Demographic details of RA and OA patients are presented in Supplementary Table 1b.

Another group of active RA patients (N = 5) and JAK inhibitor treated RA patient (N = 2) were recruited after obtaining consent at the Division of Rheumatology outpatient clinic of the Heidelberg University Hospital under approval S-969/2020 from the Ethics Committee of the Heidelberg University. In addition, 3 age and sex matched healthy subjects were also recruited. Patient disease activity was based on the DAS28-CRP score: DAS28 < 2.6 remission; 2.6 < DAS28 < 3.2 intermediate; DAS28 > 3.2 active.

Exclusion criteria for recruited RA Patients were: a) active cancer and/or ongoing anti-cancer therapy; b) pregnancy or breast-feeding; c) have a current or recent clinically serious viral, bacterial, fungal, or parasitic infection; d) vaccinated in the past 4 weeks; e) have a history of active hepatitis B virus (HBV), hepatitis C virus (HCV), or human immunodeficiency virus (HIV) and f) did not provide written consent.

2.2. Data and sample collection

Study variables were collected from the day of patient enrolment from the Department of Rheumatology and Department of orthopaedics. PB and SF samples were obtained from the Department of Rheumatology and Department of Orthopaedics. Paired PB of patients with RA was also collected on the same day. Only PB was collected from OA patients. The clinical information of the enrolled subjects was collected at the Department of Transplant Immunology and Immunogenetics, AIIMS, New Delhi. Specimens for cell culture were collected in heparinized tubes (BD, USA); whereas plasma samples for ELISA were collected in EDTA vials. Peripheral blood mononuclear cells (PBMCs) and Synovial Fluid mononuclear cells (SFMCs) were isolated and were processed immediately for ex vivo and in vitro assay.

2.3. Cell separation and cell culture

SFMC and PBMCs were isolated using Lymphoprep (Axis-Shield, Oslo, Norway) density gradient centrifugation. Cells were resuspended in RPMI-1640 containing L-Glutamine, and HEPES (Thermo Fischer Scientific, MA USA) supplemented with, 10% FBS (Thermo Fischer Scientific, MA USA) and 100 U/ml penicillin, 100 μg/ml streptomycin (Sigma-Aldrich, St. Louis, MO, USA). Cells were kept in humidified 5% CO₂ incubator at 37 °C.

2.4. Cell stimulation

For stimulation, cells were cultured in the presence of Dynabeads™ Human T-Activator CD3/CD28 (Thermo Fisher Scientific) supplemented with rIL-2 (30 U/ml) (Thermo Fisher Scientific) for 48 or 72 has indicated. For intracellular cytokine analysis, cells were cultured (0.5x10⁶ cells/ml) for 48 h and 10 μg/ml of Brefeldin A (Sigma Aldrich, MO, USA) was added during the last 12 h.

For IL-9 blocking experiments, purified NA/LE mouse anti human IL-9R alpha (AH9R7) was used (BD Biosciences, NJ, USA). Recombinant IL-9 was used from Peprotech (PeproTech NJ, USA). Neutralizing antibodies for IL-17 (0.2 μg/ml), TNF-α (10 μg/ml) and IFN-γ (l0 μg/ml) were procured from R&D system (MN, USA), BioLegend (CA, USA) and BD Biosciences (NJ, USA). Tofacitinib citrate was procured from MedChemExpress LLC (NJ, USA). We obtained anti-TGFβ (50 μg/ml) and anti-IL-10 (0.3 μg/ml) neutralizing antibodies from ThermoFisher and R&D Systems.

2.5. Flow cytometry (FC)

For surface staining, 0.5 × 10⁶ cells were washed with phosphate-buffered saline (PBS) and stained using PE-Cy5 anti-CD4 (RPA-T4), PE anti-CD25 (M-A251), PE anti-human LAP (TGF-β1) (TW4-2F8), PE anti-IL9R (AH9R7), PE/Cy7 anti-human CD279 (PD1), PE anti-human CD152 (CTLA-4). Cells were then incubated for 30 min.

For intracellular staining, cells were stained for various intracellular molecules using PE/Cy7 anti-IFN-γ (4S.B4), PE/Cy7 anti-TNF-α (Mab11), PE anti-IL-9 (MH9D1), FITC anti-FOXP3(206D), FITC anti-IL-17 (eBio64Dec17), PE/Cy7 anti-human IL-10 (JES3-9D7), PE anti-Ki67 (B56), PE anti-Helios (22F6) after fixation and permeabilization with appropriate buffer (BD Biosciences and e-biosciences, San Diego, CA, USA)). Flow cytometry was performed using LSR Fortessa X-20 and analysis was done with Flowjo v10.

For studying IL-9 induced activation of STAT1, 3, 5 and AKT1/2. Enriched CD3 was stimulated with rIL-9 for 15, 30 and 60 mins. Cells were then fixed (Cyto-Fast^™ Fix/Perm Buffer Set; BioLegend) and stained for surface marker BV421 anti-CD3, Alexa Fluor700 anti-CD4, BV605 anti-CD25 and PerCpCy5.5 anti-CD127 (all from BioLegend). Cells were then permeabilized with (Cyto-Fast™ Fix/Perm Buffer Set; BioLegend) and stained for intra cellular molecules with APC antiFOXP3 (BioLegend), anti-pSTATl (Tyr701), anti-pSTAT3 (Tyr705), anti-pSTAT5 (Y694) and anti-pAKTl/2 (Ser-473) (antibody from Cell Signaling Technology, MA, USA). Secondary labelling was done by anti-rabbit B8515 antibody. Flow cytometry was performed using BD FAC-Symphony^™ and analysis was done with Flowjo v10. We used Fluorescence Minus One (FMO) control for all our experiments for setting appropriate gates.

2.6. Cell proliferation

CellTrace™Violet Cell Proliferation Kit (ThermoFischer Scientific) was used for staining the cell as per the manufacturer’s recommendation. Cells were stimulated for 5 days with Dynabeads™ Human T-Activator CD3/CD28 (Thermo Fisher Scientific) supplemented with rIL-2 (30 U/ml) (Thermo Fisher Scientific) in presence and absence of anti-human IL-9R alpha (AH9R7) antibody. After stimulation, cells were analysed using flow cytometry.

2.7. Real time PCR

A total of 0.5 × 10⁶ cells was seeded per well and then stimulated as indicated. RNA was extracted using GeneJET RNA Purification Kit (Thermo Fisher Scientific), according to the manufacturers’ protocol. cDNA was prepared by using Revert Aid First strand cDNA synthesis kit (Thermo Fisher Scientific). Quantitative RT-PCR was performed using Powerup Sybr Master Mix (Applied Biosystem, MA, USA) with the primers listed in Supplementary Table 2 using the QuantStudio 5 Real-Time PCR Systems (Applied Biosystems). An initial enzyme activation step of 2 min at 50 °C and denaturation step of 5 min at 95 °C was common for all genes, but the cycle for annealing was different. For TNFA, IFNG, TUBA1A (40 cycles at 95 °C for 15 s and at 52 °C for 1 min); 119, ACTB (40 cycles at 95 °C for 15 s and at 56 °C for 1 min); IL17, IL9R and GAPDH (40 cycles at 95 °C for 15 s and at 60 °C for 1 min). As normalization control GAPDH, ACTB and TUBA1A was used, fold change was calculated using 2-^ΔΔcT

Micro RNA (miRNA) was extracted using miRNeasy Mini Kit (Qia-gen, Hilden, Germany) and cDNA was prepared by using miScript IIRT-Kit (Qiagen). Quantitative RT-PCR was performed using miScript SYBR® Green PCR Kit (Qiagen) with commercially available primers for miR-155 and RNU6-2 (Qiagen). As normalization control RNU6-2 was used, fold change was calculated using 2^-ΔΔCT.

2.8. Cytokine ELISA

Levels of soluble IL-9 was assessed in plasma (OA, RA) and SF (RA) by IL-9 ELISA kit (Biolegends) as per the manufacturer’s protocol.

2.9. Statistical analysis

Continuous variables were expressed either as means ± SDs when normally distributed or median with interquartile range when non-normally distributed. The student’s t-test or Wilcoxon rank-sum test was performed to compare normally and non-normally distributed continuous variables, respectively. For the effect size i.e. difference, 95% CIs were calculated. Statistical significance was considered at P < 0.05 (*** p < 0.001; **P < 0.005; *P < 0.05). All statistical analyses were performed using Graphpad Prism 9.

3. Results

3.1. Accumulation of regulatory T cells at the disease site in RA

Participation of T lymphocytes in the pathogenesis of RA is known for several years [25]. An important subset of T cells and a component of the immune regulatory network are Treg cells. In RA, there are several but conflicting reports about the frequency of Treg cells [13−16]. To determine, the distribution of Treg cells between the peripheral and site of inflammation, we analysed the frequency of the Treg cells in the paired samples of RA patients; synovial fluid (SF) of RA patients relative to their autologous paired peripheral blood (PB). We also compared the frequency of Treg cells in PB of RA patients with PB of OA patients, which were the disease control group. We observed that the frequency of Treg cells was significantly higher in the SF compared to the paired PB derived, suggesting an enrichment of Treg cells at the disease site in RA (Fig. 1 A and B). However, no significant difference was observed between the frequency of Treg cells in the PB of OA and RA patients (Fig. 1 A and B; Supplementary Table 3).

Fig. 1. Enrichment of regulatory T cells at the disease site of RA patients.

(A - D) Flow cytometry based comparative ex vivo analysis of regulatory T (Treg) cells derived from peripheral blood (PB) of OA patients, Synovial fluid (SF) and autologous PB of RA patients. First, single cells were gated, then lymphocytes were gated on the basis of their scatter profile (FSC/SSC) followed by gating on CD4⁺ CD25⁺ cells. Representative histogram showing FOXP3+ Treg cells in PB of OA patient, PB and SF on gated CD4 + CD25+ cell population in RA patients. (B) Cumulative plots showing frequency of CD4⁺CD25⁺Foxp3⁺ cells in PB of OA; PB and autologous SF of RA patients {Mean ± SD, N = 10 (OA), n = 20 (RA); **** p < 0.0001, ns -non significant}. (C) Cumulative plots showing frequency of CD4⁺ CD25⁺ FOXP3⁺ Helios⁺ cells in PB of OA; PB and autologous SF of RA patients. {Mean ± SD, N = 10; **P < 0.005}. (D, E) SFMCs and PBMCs from RA patients were stimulated with Dynabeads™ Human T-Activator CD3/CD28 for 48 h followed by detection of IL-10 and TGF-β on Treg cells using Flow cytometry. Lymphocytes were gated on the basis of their scatter profile (FSC/SSC) and further gated on CD4⁺. On gated CD4+ cells, the frequency of TGF- and IL-10 expressing FOXP3+ cells were assessed using Flow cytometry. Bar diagram showing significant decrease in the frequency of TGF-β (D) and IL-10 (E) expressing Treg cells at the local site (SF) compared to the autologous PB of patients (Mean ± SD; N = 10; *** p < 0.001}. We randomly selected the patients for performing experiments.

We were intrigued to investigate the reason for active inflammation at the pathologic site in spite of Treg cell accumulation. Therefore, we checked the expression of Helios, a member of the Ikaros family of transcription factors, shown to be critical for mediating suppressive functions of Tregs [26]. We observed a higher frequency of Helios expressing Treg cells in the PB compared to the paired SF of RA patients suggesting functional incapacitation of SF enriched Treg cells (Fig. 1 C; Supplementary Fig. 1 A; Supplementary Table 3). We further analysed if the Treg cells at the SF presented further defects in the production of regulatory immune mediators. In RA, we observed a significantly lower frequency of Treg cells expressing transforming growth factor beta (TGF-β) and IL-10 (Fig. 1 D and E; Supplementary Fig. 1B and 1C) relative to peripheral Treg cells after TCR engagement. These observations suggest that SF enriched Treg cells might be functionally incapacitated due to the diminished production of suppressive cytokines. Also, these observations hint that SF accumulated Treg cells might be compromised due to the influence of local inflammatory milieu.

3.2. Elevated IL-9R expressing regulatory T cells at the site of inflammation

IL-9 has been observed to be present in high concentrations in the SF in RA [19]. Thus, we quantified the concentration of IL-9 in RA patients compared to OA patients. We observed increased levels of soluble IL-9 in the PB of RA patients compared to OA patients. Moreover, the IL-9 concentration was much higher in the SF than PB of RA patients (Fig. 2 A; Supplementary Table 3). This observation was in accordance with our earlier observation, that SF derived CD4 cells secrete more IL-9 with TCR engagement compared to the autologous PB derived CD4 cells [19].

Fig. 2. Elevated IL-9 at the disease site of RA patients and it impact on Regulatory T cells.

(A) Measurement of IL-9 concentration in the synovial fluid (SF) and autologous peripheral blood (PB) of RA patients and PB of OA patients using ELISA {Mean ± SD, N = 10 (OA), N = 20 (RA); **** p < 0.0001; **P < 0.005}. (B, C) Relative expression of IL9 (B) and its receptor IL9R (C) in PBMC of OA and RA; SFMC of RA using realtime PCR. (Mean ± SD; N = 10; **** p < 0.0001; **P < 0.005). (D) FC based ex vivo analysis of IL-9R expressing Treg cells at disease site and PB of RA patients; PB of OA patients. First, lymphocytes were gated on the basis of their scatter profile (FSC/SSC) followed by gating on CD4⁺ CD25⁺ cells. Bar graph showing cumulative frequency of CD4⁺CD25⁺ FOXP3⁺ IL.-9R⁺ in PBMC of OA; PBMC & SFMC of RA patient (Mean ± SD; N = 10; **P < 0.005). (E, F) Influence of IL-9 on PB derived Treg cells was investigated. Mononuclear cells (MNCs) derived from PB of RA patients were stimulated with anti-CD3/CD28 conjugated dynabeads for 48 h in presence or absence of 20 ng/ml of recombinant IL-9 (rIL-9) followed by detection of IL-10 and TGF-β on Treg cells using FC. Bar diagram displaying significantly decreased percentage frequency of CD4⁺FOXP3⁺TGF-β ⁺ and CD4⁺Foxp3⁺1L-10⁺ cells with rIL-9 (Mean ± SD; N = 10; **P < 0.005). Patients were randomly selected for our experiments.

We next examined the gene expression of IL-9 and its receptor IL-9R. The expression of IL-9 and IL-9R in synovial fluid mononuclear cells (SFMCs) was much higher relative to peripheral blood mononuclear cells (PBMCs) (Fig. 2 B and C; Supplementary Table 3). SF samples from OA patients were not collected in our study due to two primary reasons: significantly lower soluble IL-9 levels compared to RA patients (Fig. 2A) and fewer infiltrating cells in OA synovial fluid, which hindered further experimentation on their phenotype and functions. Our study primarily focused on investigating the impact of IL-9 in the context of RA, where IL-9 levels are elevated, and synovial inflammation is more pronounced.

As Treg cells are enriched in RA SF compared to PB, we evaluated the frequency of IL-9 receptor (IL-9R) expressing Treg cells between the two compartments. Increased frequency of IL-9R expressing Treg cells was present in SF compared to PB (Fig. 2D; Supplementary Fig. 2 A, Supplementary Table 3). Additionally, increased frequency of IL-9R expressing T effector cells (CD4⁺ CD25⁺ FOXP3^-) was observed in SF compared to PB (Supplementary Fig. 2 A and B). These observations clearly indicate accumulation of IL-9R expressing CD4 cells at the inflammatory site.

To understand the effect of IL-9 on the functions of PB derived Treg cells, we evaluated the anti-inflammatory cytokine production of Treg cells from RA patients following TCR engagement in the presence and absence of recombinant IL-9 (rIL-9, 20 ng/ml). The chosen concentration of rIL-9 was based on the dose-dependent effect of rIL-9 on the frequency of suppressive cytokine (IL-10 & TGF-β) producing Treg cells (Supplementary Fig. 2C). A significant decrease in frequency of TGF-β and IL-10 expressing Treg cells was noted (Fig. 2 G and H; Supplementary Fig. 2D). We also checked the effect of IL-9 on the proliferation of Treg cells using Ki-67, a nuclear antigen marker selectively expressed in proliferating cells, and observed that IL-9 supplementation reduced the frequency of Ki67⁺Treg cells derived from PB of the RA patients (Supplementary Fig. 2 E and F). Our findings suggest that IL-9 may impair the immune suppressive cytokine production and the proliferative capacity of Treg cells in RA patients.

3.3. IL-9 negatively influences the function of synovial fluid derived Treg cells

To understand the impact of IL-9 on SF derived Treg cell function we used an antibody against IL-9R to abrogate the interaction between IL-9 with its receptor IL-9R. SFMCs were stimulated in the presence and absence of IL-9R blockade followed by flow cytometry-based evaluation of Treg cell function. (Fig. 3 A - F). Blocking IL-9/IL-9R interaction enhanced the proliferation of SF derived Treg cells (Fig. 3 A; Supplmentary Fig. 3 A and B). This was further substantiated by CellTrace CFSE cell proliferation assay (Fig. 3 B). Treg cells mediate their suppressive function either contact independent involving suppressive cytokines (IL-10 and TGF-β) or contact dependent involving immune inhibitory molecules like programmed cell death protein 1 (PD-1) and cytotoxic T lymphocyte associated protein 4 (CTLA-4). With IL-9R blockade, we observed an increase in the frequency of TGF-β (Fig. 3 C; Supplementary Fig. 3 A and C) and IL-10 producing Treg cells (Fig. 3 D; Supplementary Fig. 3 A and D). We also observed a significant increase in the frequency of PD1⁺ Treg cells and CTLA⁺ Treg cells with IL-9R blockade (Fig. 3 E and F; Supplementary Fig. 3 A, E and F). Together, our data revealed that abrogation of signaling through IL-9R re-established the immune suppressive capacity of SF derived Treg cells in RA.

Fig. 3. Effect of in vitro blocking IL-9 receptor (IL-9R) on Treg cell irnmunosuppressive cytokine production and proliferation.

(A-F) Mononuclear cells (MNCs) derived from SF of RA patients were stimulated with anti-CD3/CD28 conjugated dynabeads in presence or absence of 10 μg/ml of IL-9R blocking antibody. For analysing suppressive cytokine production, cells were stimulated for 48 h, whereas for proliferation assay, cells were stimulated for 72 h. (A) For analysis of Treg cell proliferation, lymphocytes were gated on the basis of their scatter profile (FSC/SSC) and further gated on CD4⁺ T cells followed by gating FOXP3⁺Ki67⁺ cells. Cumulative Plots showing frequency of CD4 + FOXP3 + Ki67⁺ (N = 8; *** p < 0.001; **P < 0.005; *P < 0.05). (B) CFSE stained cells were stimulated with anti-CD3/CD28 conjugated dynabeads in presence or absence of 10 μg/ml of IL-9R blocking antibody for 5 days. The histogram showing the proliferation of CD4⁺CD25⁺FOXP3⁺cells. One representative out of three experiments. Proliferated cells showed decrease in fluorescence. (C, D) On gated CD4+, the frequency of TGF-β and IL-10 producing FOXP3₊ cells were assessed. Cumulative Plots showing frequency of CD4⁺FOXP3⁺TGF-β⁺ (C) and CD4⁺FOXP3 + IL-10+ (D) (N = 10; *** p < 0.001; **P < 0.005; *P < 0.05). (E, F) Mononuclear cells (MNCs) derived from SF of RA patients were stimulated for 48 h with anti-CD3/CD28 conjugated dynabeads in presence or absence of 10 μg/ml of IL-9R blocking antibody followed by detection of PDl and CTLA-4 on Treg cells using FC. Lymphocytes were gated on the basis of their scatter profile (FSC/SSC) and further gated on CD4₊. On gated CD4+ cells, the frequency of PDl (E) and CTLA-4 (F) expressing FOXP3+ cells were assessed using FC. Cumulative Plots showing frequency of CD4⁺FOXP3⁺PD1⁺(n = 10; **P < 0.005). (E), CD4⁺FOXP3⁺CTLA4⁺ (F) (n = 8; *** p < 0.001). We randomly selected the patients for our experiments.

3.4. Enumeration of inflammatory cytokine producing CD4+ T cells derived from SF

Enrichment of inflammatory T cells (IFN-γ, TNF-α and IL-17) in SF is well documented in literature and plays a pivotal role in synovitis. Thus, we investigated the influence of blocking IL-9-signaling on CD4 expressing IFN-γ, TNF-α and IL-17 (Fig. 4A-F). IL-9R blockade along with TCR engagement significantly reduced the frequencies of TNF-α (Fig. 4A; Supplementary Fig. 4A and B), IFN-γ (Fig. 4C; Supplementary Fig. 4A and C) and IL-17 (Fig. 4E; Supplementary Fig. 4A and D) secreting CD4 cells. These results were further verified by the analysis of the respective mRNA expression, which confirmed a reduction in the expression of TNFA (Fig. 4B), IFNG (Fig. 4 D) and IL17 (Fig. 4F) with IL-9R blockade. Together, these observations suggest that blocking signaling through IL-9R is able to reduce the expression of pro inflammatory cytokines. Association of increased expression of miRNA-155 (miR-155) with synovial inflammation has been documented in RA [27]. Thus, we investigated the impact of blocking IL-9R on miR-155 expression. We observed a reduction in the expression of miR-155 in SFMCs following the IL-9R blockade (Fig. 4 G), which may contribute in reduction of inflammation. Since, we observed that IL-9 has the potential to modulate the expression of T cell derived pro-inflammatory cytokines. We next investigated the impact of TNF-α, IFN-γ and IL-17 on IL-9-expressing CD4 cells. SF derived cells were TCR engaged in the presence or absence of neutralizing antibodies against TNF-α, IFN-γ and IL-17. No significant changes in the frequencies of CD4 cells expressing IL-9 were observed in the presence of these neutralizing antibodies (Fig. 4 H; Supplementary Fig. 4E). These observations hint that IL-9 plays a dominant role in modulating the expression of TNF-α, IFN-α and IL-17 in RA CD4 T cells.

Fig. 4. Impact of blocking IL-9R on frequency of inflammatory CD4 cells.

(A-F) Mononuclear cells (MNCs) derived from SF of RA patients were stimulated with anti-CD3/CD28 conjugated dynabeads for 48 h in presence or absence of 10 μg/ml of IL-9R blocking antibody followed by intracellular detection of IFN-γ, 1NF-α and IL-17. Lymphocytes were gated on the basis of their scatter profile (FSC/SSC) followed by gating of CD4⁺T cells and then further gating CD4⁺TNF-α⁺, CD4⁺IFN-γ⁺, and CD4⁺IL-17⁺T cells. For mRNA expression mononuclear cells (MNCs) derived from SF of RA patients were stimulated with anti-CD3/CD28 conjugated dynabeads for 16 h in presence or absence of 10 μg/ml of IL-9R blocking antibody. This was followed by RNA isolation, cDNA preparation and subsequently RT-PCR for IFNG, TNF and IL17. Bar diagram displaying significantly decreased percentage frequency of CD4⁺IFN-γ ⁺cells (A), CD4⁺TNF-α ⁺cells (C) and CD4⁺IL17⁺cells (E) with IL-9R blocking antibody treatment (N = 7; *P < 0.05).). (B, D, F) Graphs show quantitative RT-PCR analysis of mRNA of TNFA, IFNG, and IL17 expression. The data are presented as fold change relative to the expression of untreated cells at the same time point (N = 7; **P < 0.005; *P < 0.05).). (G) Mononuclear cells (MNCs) derived from SF of RA patients were stimulated with anti-CD3/CD28 conjugated dynabeads for 72 h in presence or absence of 10 μg/ml of IL-9R blocking antibody or isotype control antibody. This was followed by RNA isolation, cDNA preparation and subsequently RT-PCR for miRNA 155. Graphs show quantitative RT-PCR analysis of155 expression normalized to RNU6-2 expression. The data are presented as fold change relative to the expression ofuntreated cells at the same time point (mean ± SD; N = 7; **** p < 0.0001). (H)Mononuclear cells (MNCs) derived from SF of RA patients were stimulated with anti-CD3/CD28 conjugated dynabeads for 48 h in presence or absence of anti -IFN-γ (10 μg/ml) antibody, anti-TNF-α(l0 μg/ml) antibody and anti-IL-17 (0.2 μg/ml) antibody followed by detection of IL-9 producing CD4 T cells. First, lymphocytes were gated on the basis of their scatter profile (FSC/SSC) followed by gating CD4⁺IL-9⁺. Cumulative data represented frequency of CD4⁺IL-9⁺cells (N = 7; ns -non significant). Statistical analysis was performed using 1 way-Anova test. Patients were randomly selected for performing experiments.

3.5. Interleukin − 9 mediates its effect through activation of STATl, 3, 5 and AKT

To understand the mechanism by which IL-9 plays a contrasting effect on effector T cells and Treg cells we stimulated CD3 enriched cells derived from PB of RA patients with rIL-9 for 15mins, 30 mins and 60 mins, and quantified the phosphorylation of Signal Transducer and Activator of Transcription (STAT)1 (Tyr701), STAT3 (Tyr705), STAT5 (Tyr694) and AKT (Ser473). Enhanced phosphorylation was observed in all proteins after 15 min of stimulation in Teff and Treg cells. The phosphorylation of STAT1, 3 and AKT remained for 30 min of stimulation with rIL-9, but the phosphorylation of STATS was diminished at 30 min compared to 15 min (Fig. 5A-I). After 60 min, phosphorylation of STAT1, 3, 5 and AKT in either Teff or Treg cells returned to baseline values. We did not observe any difference in the kinetics of phosphorylation of STAT1,3,5 and AKT between Teff and Treg cells. We observed similar kinetics of rIL-9 induced phosphorylation for STAT1, STAT3, STATS, and AKT between Teff and Treg cells in healthy controls (Supplementary Fig. 5A, B, C and D).

Fig. 5. IL-9 mediated activation of signaling molecules in effector T cell and regulatory T cells.

(A - I) Enriched CD3 cells derived from peripheral blood of RA patients were stimulated with recombinant IL-9 (20 ng/ml) for 15 mins, 30 mins and 60 mins. Lymphocytes were gated on the basis of their scatter profile (FSC/SSC) followed by gating of CD3⁺T cells and then further gating of CD4⁺ CD25⁺ cells. On gated CD4⁺ CD25⁺ cells, further gating was performed on FOXP3⁺ CD127⁻ (regulatory T cells) and FOXP3^- CD127⁺(effector T cells). (B) Representative histogram showing pSTAT3+ on gated regulatory T (Treg) cells and effector T cells (Teff cells) population in RA patients. (C) Cumulative plots showing the frequency of pSTATl (mean± SD; N = 5). (D) Representative histogram showing pSTAT3+ on gated regulatory T (Treg) cells and effector T cells (Teff cells) population in RA patients. (E) Cumulative plots showing the frequency of pSTAT3 (mean ± SD; N = 5). (F) Representative histogram showing pSTAT5+ on gated regulatory T (Treg) cells and effector T cells (Teff cells) population in RA patients. (G) Cumulative plots showing the frequency of pSTAT5 (mean ± SD; N = 5). (H) Representative histogram showing pAKT+ on gated regulatory T (Treg) cells and effector T cells (Teff cells) population in RA patients. (I) Cumulative plots showing the frequency of pAKT positive cells on Treg and Teff cells. (mean ± SD; N = 5). We randomly selected the patients for our experiments.

Phosphorylation of AKT at Ser-473 is essential for the functions of Teff cells. However, AKT phosphorylation at Ser473 has been proven to reduce the function of Treg cells [28]. Interestingly, we observed the phosphorylation of AKT in Treg cells with IL-9 stimulation, which might explain the mechanism behind the negative influence of IL-9 on Treg cells. our observations provide clear evidence of the direct impact of IL-9 on T cell subsets, as indicated by the activation of STAT1, STAT3, STATS, and AKT. To ascertain whether IL-9’s influence on T effector cells is mediated directly or indirectly through Treg cells, we conducted experiments using blocking antibodies against IL-10 and TGF-β to compromise the suppressive activity of Treg cells. our findings revealed that IL-9 was able to promote IFN-γ, TNF-α, and IL-17-expressing T effector cells, even in the presence of blocking antibodies against IL-10 and TGF-β (Supplementary Fig. 5 E). This suggests a direct effect of IL-9 on T effector cells on one hand and on the other hand IL-9 promotes inflammatory T cells by compromising the suppressive cytokine production of Treg cells.

3.6. Tofacitinib inhibits the Interleukin-9 mediated effect on inflammatory T cells

IL-9 mediates its effect through the activation of Janus kinase (JAK1) and JAK3. Tofacitinib, which is used for treatment of RA patients with moderate to severe RA disease, is a potent, selective JAK inhibitor that preferentially inhibits JAK 1 and JAK3. Therefore, we investigated the effect of tofacitinib on IL-9 mediated effect on SF derived inflammatory T cells. Cells derived from SF of RA patients were pre-treated with tofacitinib prior to TCR engagement. Tofacitinib treatment reduced the IL-9-mediated increase in the T cells producing IFN-γ (Fig. 6 A; Supplementary Fig. 6 A and B), TNF-α (Fig. 6 B; Supplementary Fig. 6 A and C) and IL-17 (Fig. 6 C; Supplementary Fig. 6 A and D).

Fig. 6. Impact of Tofacitinib on IL-9 induced inflammatory CD4 cells.

(A-C) Mononuclear cells (MNCs) derived from SF of RA patients were pre-treated with Tofacitinib (5 μM) for 30 mins and then stimulated (STIM) with anti-CD3/CD28 conjugated dynabeads for 48 has indicated with or without recombinant IL-9 (20 ng/ml) followed by intracellular detection of IFN-γ, TNF-α and IL-17. Lymphocytes were gated based on their scatter profile (FSC/SSC) followed by gating of CD4⁺T cells and then further gating CD4⁺ IFN-γ ⁺ (A), CD4⁺TNF-α ⁺ (B), and CD4⁺IL-17⁺ (C) T cells. Cumulative plots showing the frequency of CD4⁺ IFN-γ⁺ (A), CD4⁺TNF-α⁺ (B), and CD4⁺IL-17⁺ (C) T cells (mean ± SD; N = 6; **P < 0.005; *P < 0.05).

As we observed that IL-9 activates STAT1, 3, 5 and AKT in Teff and Treg cells. We checked whether IL-9 can induce the activation of STAT1, 3, 5 and AKT in RA patients undergoing treatment with JAK inhibitors (Supplementary Fig. 7 A-E). Indeed, we observed that IL-9 failed to induce activation of STAT1, 3, 5 and AKT after 15, 30 and 60 min of stimulation in Treg and Teff cells derived from RA patients undergoing intervention with JAK inhibitor. Our results indicate that JAK inhibitors like Tofacitinib have the potential to inhibit the proinflammatory effect of IL-9.

4. Discussion

Despite IL-9 was discovered and known for more than three decades, its role in several inflammatory diseases remains unclear. IL-9 was extensively studied in the context of Th2 associated diseases such as helminth infestation and allergies and thought to be a Th2 associated cytokine [29–30]. However, recent studies show the relevance of IL-9 in the pathogenesis of various autoimmune inflammatory conditions like RA, colitis, uveitis, multiple sclerosis, psoriatic arthritis and pulmonary sarcoidosis [19,31–38].

Previously, in RA we demonstrated elevated levels of IL-9 and enrichment of IL-9 producing T cells (Th-9) in the SF of RA patients [19]. We also showed that IL-9 facilitates chronic inflammation in RA by prolonging neutrophil survival and enhancing the function of effector T cells producing inflammatory cytokines IFN-γ, TNFα, IL-17 [19]. Here, we confirm our previous results of elevated IL-9 and additionally demonstrate elevated expression of IL-9 and its receptor on T cells derived from SF of RA patients. As IL-9 is a growth factor for various cell types, we intended to unravel how the heightened expression of IL-9/IL-9R axis could contribute to synovial inflammation in RA patients.

Higher frequency of FOXP3+ Treg cells are observed in chronic inflammatory diseases and thought to be a compensatory mechanism to maintain immune homeostasis and thus avoiding excessive bystander damage to host tissue. Accordingly, we found elevated number of Treg cells at the SF of RA patients compared to the PB, substantiating earlier reports [13,39]. Such active synovial inflammation in RA despite Treg cell enrichment intrigued us to evaluate the suppressive capacity of locally accumulated Treg cells. Interestingly, the SF derived Treg cells from RA patients were functionally impaired relative to the PB derived Treg cells in terms of suppressive cytokine production (IL-10 and TGF-β). In recent years, it has become apparent that in the pathological conditions, the ability of Treg cells to elicit effective immune suppression gets modified not just by altered ratio of effector / suppressor cells, but also due to incapacitation of their suppressive function in local inflammatory cytokine milieu. There are reports in RA clearly indicating that the functional incapacitation of Treg cells by inflammatory cytokines such as TNF and IL-6 [40–42]. Elevated immunosuppressive function of Treg cells was observed in RA patients with anti-TNF-α treatment suggesting its impact on modulation of regulatory T cell functions [14]. Thus, we intended to evaluate if IL-9 plays any role in functional incapacitation of synovial Treg cells in RA. Indeed, we observed that circulatory Treg cells from RA patients lost their capacity to produce suppressive cytokines (IL-10 and TGF-β) in the presence of IL-9. On the contrary, blocking the IL-9R pathway could rescue the immune-suppressive cytokines (IL-10 and TGF-β) production of synovial Treg cells of RA patients, which were otherwise very weak producer of IL-10 and TGF-β. Also, IL-9 blockade enhanced the expression of co-inhibitory molecules (PD1 and CTLA4) which are critical for mediating their immunosuppressive function. Furthermore, we showed that inhibiting the IL-9 pathway reduced the frequency of inflammatory (TNFα, IFN γ and IL-17) CD4 cells. However, blocking TNFα, IFNγ and IL-17 failed to show any effect on CD4 producing IL-9 cells indicating that IL-9 is hierarchical dominant over TNFα, IFN-γ and IL-17 in the cytokine network in RA. We also noted similar dominance of IL-9 in Pulmonary Sarcoidosis [38]. We observed simultaneous rescue of Treg function and dampening of inflammatory T cells in RA. Our results strongly indicate the possibility of rescue of immunosuppressive function of Treg cells and thus ameliorating inflammatory T cells in RA patients through abrogating IL-9 pathway. Similar to our observation, in animal model of myasthenia gravis, IL-9 neutralization resulted in an increased in the frequency of the Treg cells, coupled with a decrease in the frequency of pathogenic Thl cells and disease severity [43]. Also, we showed that IL-9R blockade can decrease the expression of miR-155, which is critical for mediating Th-17 inflammation as miR-155-knockout mice are resistant to collagen-induced arthritis [27]. Taken together, our study shows heightened signaling through the IL-9/IL-9R axis in RA patients, which might dampen the Treg mediated immune suppression and thus rendering inflammatory T cells in the joints of RA patients unchecked. In other words, inhibiting the IL-9 pathway can rescue the Treg cell function and control the inflammatory T cells.

Here, we demonstrate that the IL-9 mediated effect on T cell subsets (Teff and Treg cells) is dependent on the activation of STAT1 (Tyr701), STAT3 (Tyr705), STAT5 (Tyr694) and AKT (Ser473). Phosphorylation of STAT5 and STAT3 facilitates inflammatory Thl (IFN-γ and TNF-α) and Th17 (IL-17), whereas phosphorylation of STATS promotes Treg cell function [45]. Here, we observed that IL-9 induced activation of STAT1 and STAT3 was for a longer duration (till 30 min) compared to activation of STAT5 (till 15 min). This suggests that IL-9 promotes inflammation by maintaining the pro-inflammatory intracellular signaling intermediates (STAT1 and 3) for longer periods compared to activation of STAT 5 which supports Treg cell function. Activation of AKT (Ser473) plays a T cell subset specific function. In Teff cells, AKT activation promotes cell survival and proliferation. In contrast, in Treg cells, AKT activation reduces their immune-suppressive function [28]. This suggests that IL-9 mediated negative impact on Treg cells in RA is possibly due to the activation of AKT in Treg cells. Very recently, we have shown similar impact of IL-9 on mouse Treg and Teff cells [46]. Thus, our present study provides a possible mechanism by which IL-9 can impair the functions of Treg cells and promote inflammatory T cells in RA. Additionally, we showed that JAK-inhibitor tofacitinib, which is used to treat RA patients, was able to reduce the pro-inflammatory effect of IL-9. Our results provide evidence that one of the putative ways by which tofacitinib reduces inflammation in RA patients, might be through inhibiting the pro-inflammatory effect of IL-9. The limitation of this study is that we could not study the IL-9 induced transcriptome signature of Teff and Treg cells, which could have provided better mechanistic insight of IL-9 mediated effect. Our past and ongoing investigations into the source of IL-9 among lymphocytes and monocytes have revealed that CD4 cells are the main contributors to IL-9 in RA [19]. However, comprehensive future research is needed to identify and characterize the specific immune cell subsets responsible for pathogenic IL-9 production in RA.

In summary, we provide evidence that blocking IL-9 pathway could inhibit synovial inflammation by restoring the suppressive functions of synovial Treg cells which in turn can inhibit the synovial inflammatory T cells from RA patients (Fig. 7). Thus, inhibiting IL-9 pathway may constitute a promising therapy for RA as it has the potential to rejuvenate Treg cells which can dampen inflammation in RA. A very recent report showed that IL-9 neutralization reduced disease severity in collagen Induced Arthritis (CIA) model [47]. In addition, prophylactic treatment with anti-IL-9 monoclonal antibodies delayed the onset of arthritis in the CIA model [47].

Fig. 7.

Schematic diagram of the effect of IL-9 blocking on restoration of sy novial Treg cells and reduction in inflammatory T cells. (Upper figure) At the pathological site, there is enrichment of effector T (Teff) and regulatory T(Treg) cells. These T cell subsets show higher expression of IL-9 receptor (IL-9R). However, Treg cells of synovial fluid are functionally compromised in terms of suppressive cytokine (IL-10 and TGF-β) production. (Lower Figure) Blocking IL-9 receptor reduced the inflammatory T cells secreting Tumour Necrosis Factor alpha (TNF-α), Interferon gamma (IFN-γ) and interleukin −17 (IL-17); and restored the suppressive cytokine production by Treg cells derived from syno vial fluid.

Thus, our results along with recent reports strongly suggest an important role for IL-9 in RA pathogenesis and blocking IL-9/IL-9R axis may constitute novel therapeutics approaches in RA.

Data availability

The data that support the findings of this study are available in the manuscript and can be obtained from the corresponding author upon request.