Research ArticleIMMUNOLOGY

Induced, but not natural, regulatory T cells retain phenotype and function following exposure to inflamed synovial fibroblasts

See allHide authors and affiliations

Science Advances  28 Oct 2020:
Vol. 6, no. 44, eabb0606
DOI: 10.1126/sciadv.abb0606

Abstract

Aberrant number and/or dysfunction of CD4+Foxp3+ Regulatory T cells (Tregs) are associated with the pathogenesis of rheumatoid arthritis (RA). A previous study has demonstrated that thymus-derived, natural Tregs (nTregs) prefer to accumulate in inflamed joints and transdifferentiate to TH17 cells under the stimulation of inflamed synovial fibroblasts (SFs). In this study, we made a head-to-head comparison of both Treg subsets and demonstrated that induced Tregs (iTregs), but not nTregs, retained Foxp3 expression and regulatory function on T effector cells (Teffs) after being primed with inflamed SFs. In addition, iTregs inhibited proliferation, inflammatory cytokine production, migration, and invasion ability of collagen-induced arthritis (CIA)–SFs in vitro and in vivo. Moreover, we noted that iTregs directly targeted inflamed SFs to treat autoimmune arthritis, while nTregs failed to do this. Thus, manipulation of the iTreg subset may have a greater potential for prevention or treatment of patients with RA.

INTRODUCTION

Rheumatoid arthritis (RA) is a chronic, progressive, and systemic autoimmune disorder characterized by primary inflammatory synovitis, subsequent articular cartilage destruction, and bone erosion. Inflamed synovial fibroblasts (SFs) are a dominant cell type involved in the pathogenesis of RA (1). These cells have a semitransformed, autoaggressive phenotype characterized by reduced apoptosis, production of matrix-degrading enzymes, secretion of proinflammatory cytokines and receptor activator of nuclear factor-κ B ligand (RANKL) that promote osteoclasts differentiation, and the subsequent bone destruction (2, 3).

It is well recognized that CD4+Foxp3+ regulatory T cells (Tregs) play an essential role in maintaining immune homeostasis and preventing autoimmune diseases (4). Tregs abnormalities in number and/or function may contribute to many autoimmune diseases including RA (5). Tregs are heterogeneous and consist of at least two types, thymus-derived naturally occurring (nTregs) and induced Treg (iTreg) developed in the periphery or in vitro after stimulation with interleukin-1 (IL-2) and transforming growth factor–β (TGF-β) (6). We and others have previously reported that both Treg subsets not only share similarities but also display some differences (7, 8).

Komatsu et al. (9) provided evidence that adoptively transferred nTreg preferentially lost Foxp3 expression, transdifferentiated to pathogenic T helper 17 (TH17) cells mediated by SF-derived IL-6 and accumulated in the inflamed synovium in a collagen-induced arthritis (CIA) model. These ex-Foxp3 TH17 cells accelerated the onset and increased the severity of arthritis. Our hypothesis is that iTregs have different biological characteristics following the priming with inflamed SFs, as compared to nTregs. We aimed to address the hypothesis by directly comparing nTreg and iTreg subsets in inflammatory conditions in vitro and in the established colitis and CIA mouse models in vivo.

RESULTS

iTreg subset retains Foxp3 expression and regulatory function on Teff after being primed with CIA-SFs in vitro

To determine whether iTregs similarly lose Foxp3 and regulatory function as nTregs, we firstly isolated inflamed SF from CIA mice (CIA-SFs, purity > 90%; fig. S1, A to D) and then cocultured nTreg or iTreg subsets with CIA-SFs in a cell-cell contact manner (10:1 ratio of SFs to Treg subsets), as performed in Komatsu et al.’s study (9). We also performed CIA-SF identification after multiple passages. As we expected, the SFs maintained their phenotypic characteristics, such as the expression of vascular cell adhesion molecule–1 (VCAM-1), intercellular adhesion molecule–1, and fibroblast activation protein α (FAPα) (10, 11). We used nTreg cultured under the condition for TH17 cell differentiation as a control. nTregs and iTregs cultured alone expressed a similar high level of Foxp3 and rarely IL-17A (Fig. 1A). nTreg cultured alone under the stimulation of IL-6 or cocultured with inflamed SFs preferentially lost Foxp3 expression and some transdifferentiated into TH17 cells (Fig. 1A), while iTreg treated with CIA-SFs showed no substantial difference when compared to Tregs cultured alone. We also observed that pTreg isolated from lymph node lost Foxp3 expression and transdifferentiated to TH17 cells after being cocultured with CIA-SFs, which is similar to pTreg or naïve CD4+ T cells cultured under TH17 cell differentiation (fig. S2A). When we extended the coculture time from 3 to 7 days, we observed that nTreg cocultured with CIA-SFs sharply lost Foxp3 expression and increasingly transdifferentiated to TH17 cells (fig. S2, B and C). In addition, both nTregs and iTregs rarely expressed various transcript factors of T effector cells (Teffs) which could mediate both acute or chronic inflammatory reactions under conditions of infection or autoimmunity; however, nTregs, but not iTregs, began to express T-bet, RORγt, IL-23R, and Gata-3 mRNA after being exposed to inflamed SFs (fig. S3).

Fig. 1 iTreg subset retains Foxp3 expression and regulatory function on Teff after priming with SFs isolated from CIA mice in vitro.

(A) Treg cells were cocultured with or without CIA-SFs in a cell-cell contact manner (10:1 ratio) for 3 days [namely, Treg and Treg(p), respectively]. nTreg and iTreg cultured under TH17 differentiation condition were used as a control. The expression of IL-17A and Foxp3 on Treg cells was determined by flow cytometry. (B and C) nTreg and iTreg cocultured with or without CIA-SFs for 3 days were collected and then cocultured with carboxyfluorescein diacetate succinimidyl ester (CFSE)–labeled Teff, respectively, at a ratio of 1:1, 1:2, and 1:4 for another 3 days; Teff cells cultured alone were used as a control(baseline). The representative images of the proliferation of CD8+ T cells were showed (Treg: Teff, 1:2). n = 3. *P < 0.05, **P < 0.01, and ***P < 0.001 by two-way analysis of variance (ANOVA). Graphs show the means ± SEM. ns, not significant.

Foxp3 expression and stability usually reflect the functional activity of Tregs. To validate whether Tregs are still functional after being primed with CIA-SFs, we harvested the two Treg subsets after being cocultured with or without CIA-SFs for 3 days and performed a standard functional assay in vitro (12). The results showed that both nTregs and iTregs without being primed with CIA-SFs had similarly excellent suppressive function against Teff proliferation (Fig. 1, B and C). After being primed with CIA-SFs, iTregs maintained their suppressive function, while nTregs notably lost their suppressive capacity (Fig. 1, B and C). We further analyzed whether these differences between nTregs and iTregs also can be seen in cells from RA patient when exposed to SFs from RA patients (RA-SFs; purity, >90%; fig. S1C). Our results suggested that both nTreg and iTreg from RA patient cultured alone expressed a high level of Foxp3 and low level of IL-17A (fig. S4, A and B). Similarly, after being cocultured with RA-SFs, nTreg lost some Foxp3 expression and transdifferentiated to TH17 cells, while iTreg showed no notable difference from iTreg cultured alone. Consistently, both nTreg and iTreg showed strong suppression on CD8 proliferation, RA-SFs pretreatment weakened the suppression function of nTreg but not iTreg on CD8 proliferation (fig. S4C). Performing t-distributed stochastic neighbor embedding (tSNE) analysis by FlowJo also indicated that nTreg subset transdifferentiated more into TH17 and TH1 subsets after being cocultured with CIA-SFs (Fig. 2, A and B). In addition, nTregs underwent a sharp reduction in the expression of CD25 under inflammatory condition (Fig. 2, A and B). The cytotoxic T lymphocyte–associated protein 4 (CTLA-4) expression was significantly lower in nTreg than in iTreg. iTregs, but not nTregs, increased CTLA-4 expression in the presence of CIA-SFs (Fig. 2, A and B). We also observed that IL-6 stimulation decreased the expression of CD39, Helios, Nrp-1, and interferon regulatory transcription factor 4 (IRF4) on nTreg and had no effect on T cell immunoreceptor with Ig and ITIM domains (TIGIT), OX-40, glucocorticoid-induced TNF (tumor necrosis factor) receptor (GITR), Blimp-1, 4-1BB, inducible co-stimulator (ICOS), and CTLA-4 expression when compared to nTreg cultured alone. CIA-SFs priming did not affect the expression of OX-40, Nrp-1, TIGIT, and CIA-SFs on both nTreg and Tregs (fig. S2D). CIA-SFs priming lowered the expression of CD137 (also called 4-1BB), ICOS, Helios, CD39, CTLA-4, and IRF4 and increased the expression of GITR and Blimp-1 by nTreg when compared to nTreg cultured alone (Fig. 2C and fig. S2D). In contrast, CIA-SFs priming did not affect most of these markers expression on iTreg while increasing CD39 and CTLA-4 expression by iTreg (Fig. 2C and fig. S2D). These results further support that nTreg, but not iTreg, encounters a functional decline when exposed to the inflamed synovial tissues (9). Furthermore, under the stimulation of inflammatory microenvironment caused by CIA-SFs, iTreg, but not nTreg, successfully differentiates into effector Treg (eTreg).

Fig. 2 iTreg is more stable than nTreg in the presence of CIA-SFs.

(A and B) Tregs were cocultured with or without CIA-SFs (lower chamber) in a cell-cell contact at a ratio of 10:1 for 3 days. The expression of Treg or TH cell markers was detected using flow cytometry. The coexpression of Treg markers and TH cell markers are shown using t-distribution and Stochastic Neighbor Embedding (tSNE) plots in the left panel (A). The summary data are shown in the right panel (B). (C) Tregs were cocultured with or without CIA-SFs (lower chamber) in a ratio of 1:20 for 7 days. The expression of CD137, Helios, ICOS, CD39, and CTLA-4 by Tregs were analyzed by flow cytometry. n = 3. *P < 0.05, **P < 0.01, and ***P < 0.001 by one-way ANOVA. Graphs show the means ± SEM.

iTreg subset maintains regulatory function on Teff after being primed with CIA-SFs in vivo

The functional activity of Tregs in vitro does not necessarily reflect their functional effect on the disease in vivo. We chose a colitis model because it is a Teff-mediated disease to determine the functional activities of Treg subsets in vivo (13). Naïve CD4+ T cells (0.6 million) from CD90.1+ C57BL/6 mice were injected into Rag1−/− mice alone or together with 0.6 million of either nTreg or iTreg subsets isolated or induced, respectively, from CD90.2+ Foxp3green fluorescent protein (GFP) reporter mice after being primed with or without CIA-SFs on the same day. We found that the colitis model that had only received naïve CD4+ T cells gradually lost weight from day 9 onward and gradually developed typical severe colitis on day 33 (Fig. 3A). Cotransfer the nTreg or iTreg subset markedly suppressed the appearance of colitis, indicating that both Treg subsets are functional in the disease condition in vivo. Recipient mice of either nTreg or iTreg without priming or iTreg with priming slowly gained weight and only showed slight but not notable colitis (Fig. 3A). In contrast, when nTregs were primed with inflamed SFs, they no longer suppressed colitis. Nonetheless, cotransfer of the iTreg subset that had been primed with inflamed SFs still potently controlled the colitis (Fig. 3A).

Fig. 3 iTregs after being primed with CIA-SFs maintain the phenotype and suppressive function on Teffs in vivo in colitis model.

Naïve CD4+ T cells (0.6 million) from CD90.1+ Foxp3GFP reporter C57BL/6 mice were intraperitoneally injected into Rag1−/− mice alone or together with 0.6 million of the two Tregs subsets from CD90.2+ Foxp3GFP reporter DBA1/J mice after being primed with or without CIA-SFs at the same day. Weights of the recipient mice were monitored per 3 days after the cell transfer. Thirty-three days later, single cells from spleen, mLNs, and cLP were harvested. (A) Weight change of the mice from different groups. (B and C) The proportions of Foxp3+ cells and IL-17A+ cells on CD90.1+ cells were determined. (D and E) The proportions of Foxp3+ and IL-17A+ cells on CD90.2+ cells were determined. (F) The colon was subjected to H&E staining (40×; scale bars, 200 μm) and the development of colitis was evaluated blindly by two pathologists. n = 6. *P < 0.05, **P < 0.01, and ***P < 0.001 by one-way ANOVA(C and E) and two-way ANOVA (A). Graphs show the means ± SEM.

As expected, in this colitis model, 30% of CD4+CD90.1+ cells produced IL-17A or interferon-γ (IFN-γ) and 5% produced both IL-17A and IFN-γ in the spleen (Fig. 3, B and C). Similar results were seen in mesenteric lymph nodes (mLNs) and colonic lamina propria (cLP) (Fig. 3C). Both nTreg and iTreg subsets without priming potently suppressed the differentiation of naïve CD4+ T cells (CD90.1+) to TH17 and TH1 cells and inhibited the proliferation of Teff in the spleen, mLNs, and cLP (Fig. 3, B and C). After priming, nTregs almost completely lost their suppressive activity, while iTregs sustained this ability (Fig. 3, B and C).

To trace the fate of Treg subsets in the colitis mice, we detected Foxp3 and IL-17A expression on CD90.2+ Treg following an in vivo transfer. Consistent with the above results, both nTreg and iTreg subsets without priming mostly maintained the Foxp3 expression and resisted TH17 transdifferentiation (Fig. 3, D and E). However, the nTreg subset primed with CIA-SFs significantly lost Foxp3 expression and acquired a TH17/TH1 phenotype, especially in cLP, indicating some nTregs have converted to inflammatory cells (Fig. 3, D and E). Conversely, iTregs primed with inflamed SFs maintained Foxp3 and resisted TH17 conversion (Fig. 3, D and E). Hematoxylin and eosin (H&E) staining showed that the colitis model developed a severe inflammatory cell infiltration, ulcers, edema, and bowel wall thickening (Fig. 3F). Both nTreg and iTreg subsets without priming strongly inhibited inflammatory cell infiltration and kept epithelial cells and goblet cells intact. Severe inflammatory cell infiltration and mucous membrane ulceration were observed with cotransfer of nTregs that had been primed with inflamed CIA-SFs. In sharp contrast, cotransfer of iTregs primed with CIA-SFs exhibited less inflamed pathology and fewer inflammatory cells (Fig. 3F). These results strongly suggest that the iTreg subset has superior stability and better biological activity in down-regulating TH17/TH1 cells differentiation and Teff proliferation in autoimmune and inflammatory conditions, particularly following exposure to inflamed SFs.

iTreg, but not nTreg, maintains therapeutic effect in CIA model after being pretreated with CIA-SFs

Since we have observed that iTreg, but not nTreg, pretreated with CIA-SFs maintained their stability with little transdifferentiated into TH17 cells in vitro and in colitis model, we further hypothesized that iTreg, but not nTreg, being pretreated with CIA-SFs in vitro retained their therapeutic effect in CIA mice. To this end, we pretreated nTreg and iTreg with or without CIA-SFs (SFs to Tregs, 1:10) for 3 days and then intravenously injected these cells (3 × 106 per mouse) into CIA mice on day 27 after immunization when the onset of clinical symptoms of arthritis was observed. We observed that both iTreg and nTreg treatments reduced clinical scores and inhibit bone erosion in CIA mice (Fig. 4, A and B). Mice received nTreg being pretreated with CIA-SFs showed no substantial alleviation when compared to mice receiving no treatment (Fig. 4, A and B). In contrast, the transfer of iTreg being pretreated with CIA-SFs remained markedly decreased the severity of arthritis, showing no notable therapeutic difference to iTreg treatment (Fig. 4, A and B). Notably, the transfer of nTreg and iTreg increased the number of Foxp3+ Tregs while decreased the number of IFN-γ+ TH1 cells in the joint fluid (Fig. 4, C to F). Consistently, infusion of iTregs being pretreated with CIA-SFs also markedly increased the number of Tregs while decreased the number of TH1 cells in the joint fluid, whereas infusion of nTreg being pretreated with CIA-SFs showed no effect on the number of Foxp3+ Tregs and IFN-γ+ TH1 cells in the joint fluid when compared to mice with no treatment (Fig. 4, C to F). We also observed that both iTreg and iTreg being pretreated with CIA-SFs markedly lowered the expression of TH17 cells in lymph node cells (Fig. 4, G and H). While nTreg treatment also showed a reduced on TH17 cells in lymph node cells, infusion of nTreg being pretreated with CIA-SFs showed no effect on TH17 cell expression in lymph node cell of CIA mice (Fig. 4, G and H). These results suggest that iTreg, but not nTreg, retained their therapeutic function in CIA mice after being primed with CIA-SFs.

Fig. 4 iTreg, but not nTreg, maintains therapeutic effect in CIA model after being pretreated with CIA-SFs.

CIA was induced in DBA1/J mice with CII/CFA. Tregs (3 × 106) after being pretreated with or without [namely, Treg and Treg(p) respectively]. CIA-SFs for 3 days (Tregs to CIA-SFs, 10:1) were adoptively transferred into the established CIA mice on day 28 after the immunization, CIA model without any treatment was set as a control. (A) Clinical scores were recorded every 2 to 3 days after the onset of arthritis, and mice were sacrificed on day 50 after immunization. (B) Representative micro–computed tomography imaging of ankle joints and bone volume of the metatarsophalangeal joints from CIA mice of different treatment. (C to F) Cells in the joint fluid were collected and stimulated with PMA and ionomycin for 1 hour and then with brefeldin A, followed by incubation for another 4 hours. The expression of Foxp3, IFN-γ, and IL-17A by these cells was analyzed by flow cytometry. (G and H) Single cells isolated from draining lymph nodes were stimulated with PMA and inomycin for 1 hour and then with brefeldin A, followed by incubation for another 4 hours. The expression of Foxp3 and IL-17A by these cells were analyzed by flow cytometry. n = 5. *P < 0.05, **P < 0.01, and ***P < 0.001 by one-way ANOVA (B, D to F, and H) and two-way ANOVA (A). Graphs show the means ± SEM.

iTregs inhibit the inflammatory activities of CIA-SFs in vitro

Previous studies have reported that proinflammatory cytokines secreted by inflamed SFs, including tumor necrosis factor–α (TNFα), IL-1β, IL-6, matrix metalloproteinase–9 (MMP9), and others, are not only triggers of chronic inflammation but also promoters of pannus formation and cartilage degradation (14). The tumor-like biologic behaviors of inflamed CIA-SFs including proliferation, migration, and invasion also contribute to the severity of arthritis (15). Although previous studies have suggested that Tregs can directly target Teffs, B cells, macrophages, dendritic cells and osteoclast cells (1620), their effects on the activities CIA-SFs remain elusive. To this end, we developed in vitro experiments to determine the regulatory function of Tregs on inflamed SFs. Our result suggested that iTregs, but not nTregs, significantly inhibited the migration of CIA-SFs (Fig. 5, A and B). We also observed that iTreg, but not nTreg, significantly inhibit the invasion of CIA-SFs in both extracellular matrix (ECM) gel and collagen gel condition (Fig. 5, A and B, and fig. S5). In addition, iTregs, but not nTregs, significantly inhibited the proliferation of CIA-SFs (Fig. 5C). To determine whether iTreg subset inhibited cytokine expression or suppressed the expansion of inflamed SFs to indirectly reduce cytokine production, we also conducted a quantitative real-time polymerase chain reaction (PCR) assay to analyze the expression levels of transcription factors of these cytokines. As shown in Fig. 5 (D and E), mRNA levels of IL-6, IL-1β, IL-15, TNFα, MMP9, vascular endothelial growth factor (VEGF), and MMP14 were markedly lower in CIA-SFs treated with iTreg when compared to those CIA-SFs treated with nTreg, which suggested that iTreg diminished the production of proinflammatory cytokines by CIA-SFs. Furthermore, enzyme-linked immunosorbent assay (ELISA) results also suggested that after being cocultured with iTregs but not nTregs, the secretion of inflammatory cytokines TNFα, IL-6, and IL-1β by CIA-SFs was almost completely abrogated (Fig. 5F).

Fig. 5 iTreg subset inhibits the inflammatory activities of CIA-SFs in vitro.

(A and B) CIA-SFs were cocultured with the two Treg subsets in 24-well transwell plate; CIA-SFs cultured alone were used as a negative control. Tregs were resuspended with 10% FBS and seeded in the lower chamber; CIA-SFs were resuspended with medium without FBS and seeded in the upper insert well. After 12 hours, CIA-SFs that passed through the membrane were stained with crystal violet and detected by microscope. For the invasion test, the upper chambers were coated with fresh ECM gel. The subsequent operations were conducted similarly as for migration assay. After 24 hours, CIA-SFs that passed through membrane were stained with crystal violet and detected by microscope. Scale bar, 500 μm. (C) We cultured CIA-SFs in the presence of iTregs or nTregs in 96-well plate; CIA-SFs only were set as a negative control. Three days later, the proliferation of CIA-SFs was tested with CCK-8. (D and E) CIA-SFs were cocultured with nTregs or iTregs in 12-well transwell system. Three days later, CIA-SFs were harvested, and qPCR analysis of relative markers was performed. GAPDH was used as an internal control in all experiments. (F) Cytokines including TNFα, IL-6, and IL-1β in the culture supernatants were measured by ELISA. (G) CIA-SFs were cultured alone or cocultured with transwell-separated Tregs for 3 days and then stimulated CIA-SFs from different groups with TNFα (10 ng/ml) for 30 min before collecting protein. The expression of NF-κB, phosphorylated NF-κB (p–NF-κB), and VCAM-1 by CIA-SFs was detected by Western blot, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control. n = 6. *P < 0.05, **P < 0.01, and ***P < 0.001 by one-way ANOVA. Graphs show the means ± SEM.

It has been previously documented that the osteoclastogenesis in the CIA model requires the binding of RANKL to its receptor RANK (21). Therefore, we also sought to determine whether Treg subsets affect RANKL expression on inflamed SFs. After being cocultured for 3 days, the RANKL and IL-6 expressed by CIA-SFs were detected by flow cytometry. Our result showed that iTregs, but not nTregs, significantly decreased the levels of IL-6 and RANKL (fig. S6A). We further performed experiment to determine on inflamed SF-mediated osteoclastogensis. As shown in fig. S7, osteoclast precursors (OCPs) were cocultured with CIA-SFs in a cell-cell contact manner in the presence of macrophage colony-stimulating factor (M-CSF), with or without Tregs treatment for 3 days. We observed that iTreg, but not nTreg, inhibit inflamed SF-mediated osteoclastogenesis. These results raise the possibility that iTregs might suppress RANKL expression and then decrease osteoclastogenesis and bone erosion in autoimmune arthritis. Moreover, iTregs inhibited the production of IL-6 by CIA-SFs, followed by restraining the excessive inflammatory responses in the synovial tissues.

We have observed that iTregs inhibit the expression of different kinds of inflammatory factors produced by inflamed SFs, and many of these factors are now known to be under the control of the nuclear factor κB (NF-κB) transcription factors (22). Activated NF-κB was suggested to promote the proliferation, migration, and invasion of inflamed SFs, and the presence of activated NF-κB transcription factors has been demonstrated in animal and human inflamed SFs (23). Given that NF-κB signaling activation and VCAM-1 expression are associated with inflammation, cell migration, and invasion (24, 25), we also evaluated whether Treg subsets suppress the expression of phosphorylated NF-κB and VCAM-1in CIA-SFs. We found that iTregs, but not nTreg, significantly inhibited NF-κB phosphorylation and lowered VCAM-1 expression in CIA-SFs (Fig. 5G and fig. S6B).

iTreg subsets exhibit a protective effect on disease and inhibit the inflammatory activities of CIA-SFs in vivo in CIA model

The superior function of iTregs on CIA-SFs in vitro led us to hypothesize that iTreg subset has better function in suppressing proliferation, inflammatory cytokine production, migration, and invasion of the inflamed SFs in vivo. To address this issue, 3 × 106 nTreg or iTreg subsets were adoptively transferred into the established CIA model 14 days after the collagen II/ complete Freund’s adjuvant (CII/CFA) immunization. After infusion for about 24 hours, nTreg and iTreg mainly distributed in lung, liver, spleen, kidney, lymph nodes, and some can migrate to joints of CIA mice (fig. S8, A and B). The incidence and severity of CIA were recorded every 3 days after the cell injection. We observed that infusion of nTregs exhibited a diminished capacity to decrease the incidence and clinical score of CIA, while infusion of iTreg markedly inhibited the incidence, delayed the onset of CIA, and decreased the clinical scores (fig. S8, C and D). We isolated CIA-SFs from CIA mice of different groups on days 45 to 60 after immunization. We found that CIA-SFs isolated from iTreg-treated mice showed decreased migration, invasion, and proliferation abilities when compared to CIA-SF isolated from nTreg-treated mice (fig. S8, E to G). We also observed that iTreg decreased the mRNA expression level of IL-6, TNFα, and IL-1β in CIA-SFs isolated from iTreg-treated CIA mice when compared with nTreg-treated mice. Also, iTreg also lowered the level of IL-15, VEGF, and MMP9 in CIA-SFs when compared with nTreg but lack of statistically significance and inconsistent with the data in vitro. It suggests a more complicated microenvironment and context-specific iTregs function in distinct immune challenge in vivo (fig. S9, A to C). As VCAM-1 and NF-κB are essential for migration and invasion of CIA-SFs (25), we also assessed whether Treg treatment reduced the expression of phosphorylated NF-κB and VCAM-1 by CIA-SFs in vivo. In line with what we observed in vitro, iTregs, but not nTregs, treatment significantly inhibited NF-κB phosphorylation and VCAM-1 expression by CIA-SFs (fig. S9D). These results provide further evidence that iTreg subset has a superior ability to ameliorate CIA severity by inhibiting the activities of CIA-SFs and down-regulating the proliferation, migration and invasion ability, and inflammatory cytokine production.

Inflamed SF reduced the stability of nTreg and transdifferentiated nTreg into TH17 cells mainly via IL-6/IL6R signaling

We further investigated the mechanism by which inflamed SFs lowering the expression of Foxp3 and inducing the conversion TH17 cells in nTregs. We firstly compared cell-cell contact and transwell-separated coculture of Tregs with CIA-SFs to determine whether the effects of CIA-SFs on Tregs could be attributed to soluble mediators. We observed that cell-to-cell contact contributed to Foxp3 loss while it is not essential for TH17 cell conversion (Fig. 6A). Since we have observed that iTreg, but not nTreg, inhibit the production of inflammatory cytokines IL-6, IL-1β, and TNFα, which were previously reported to contribute to TH17 differentiation (26), and receptors for these cytokines expressed differently on nTreg and iTreg (fig. S10), neutralizing antibodies against IL-6, IL-6 receptor, IL-1β, and TNFα were added to the coculture environment of nTreg and CIA-SFs. Our result shows that neutralizing IL-6 almost abolished the conversion of TH17 cells and blocking IL-6 receptor also slightly inhibit the differentiation of TH17, whereas neither neutralization of TNFα nor neutralization of IL-1β inhibited the conversion of TH17 cells from nTreg after being priming with SFs (Fig. 6B). These results indicate that SF-derived IL-6 has a crucial role in the conversion of Foxp3+CD4+ T cells to TH17 cells. We also observed that inhibited IL-6, IL-6 receptor, IL-1β, and TNFα all contribute to the stability of nTreg (Fig. 6B). These results suggest that inflamed SF reduced the stability of nTreg and converted nTreg into TH17 cells mainly via IL-6/IL6R signaling.

Fig. 6 Inflamed SF reduced the stability of nTreg and transdifferentiated nTreg into TH17 cells mainly via IL-6/IL6R signaling.

(A) Tregs were cocultured with CIA-SFs in a cell-cell contact manner or in a transwell-separated manner. Tregs cultured alone were set as a control. Three days later, Tregs were collected and stimulated with PMA and ionomycin for 1 hour and then with brefeldin A, followed by incubation for another 4 hours. The levels of Foxp3 and IL-17A expressed in these cells were analyzed by flow cytometry. (B) nTregs were cocultured with CIA-SFs in the presence of neutralizing antibodies against IL-6, IL-6 receptor (IL-6R), IL-1β, and TNFα for 3 days. nTregs cultured with or without the presence of control immunoglobulin G (IgG) were set as a control. Three days later, nTregs were collected and stimulated with PMA and ionomycin for 1 hour and then with brefeldin A, followed by incubation for another 4 hours. Foxp3 and IL-17A expressed in these cells were analyzed by flow cytometry. n = 3. *P < 0.05, **P < 0.01, and ****P < 0.0001 by one-way ANOVA. Graphs show the means ± SEM.

DISCUSSION

Tregs are responsible for suppressing aberrant Teff autoreactive responses and maintaining immune self-tolerance (27). Functional defects and declines in the proportions of Tregs have been linked to the pathogenesis and development of RA (28), while pathogenic TH17 cells have an important role in the pathogenesis of RA by mediating osteoclastogenesis, synovial neoangiogenesis, and bone erosion. The imbalance between TH17 and Treg is strongly implicated in the onset and progression of RA (29).

In autoimmune arthritis, inflamed SFs constitute most of the cell population of inflamed synovial tissues and play a crucial role in the pathogenesis of RA, although T cells are also indispensable. Previous studies have shown that RA-SFs contribute to pannus formation and synovium damage by migrating and invading the cartilage and bone (30). Whether Tregs directly regulate and affect the biological behavior of inflamed SFs or not remains unknown and is a key issue.

The first study from Komatsu and colleagues revealed that half of the natural CD25+Foxp3+ Treg subset lost Foxp3 expression and acquired a pathogenic TH17 phenotype in CIA synovial fluid. IL-6 secretion by CIA-SFs in the joints drove the conversion of CD25+Foxp3+ T cells into pathogenic TH17 cells, and in turn, IL-17 production by the formerly Foxp3+ TH17 cells induced more IL-6 from the SFs, which establishes a positive feedback loop in the arthritic joints. Moreover, these ex-Foxp3 TH17 cells exerted more aggressive osteoclastogenic capability than TH17 cells derived from naïve CD4+ T cells because ex-Foxp3 TH17 cells produce a considerably higher level of RANKL, which is critical in osteoclastogenesis and bone erosion (31). It partially explains why nTreg subset is less effective in treating autoimmune arthritis (32). We and others provided more evidence that nTregs significantly prevented but were less satisfactory in treating autoimmune arthritis because of their plasticity and conversion to TH17, TH1, or T follicular cells under inflammatory conditions (33, 34).

The iTreg subset represents another population, which mostly not only shares similarities but also displays some differences compared to nTreg (35). Under similar inflammatory conditions, iTregs were resistant to TH17 cells conversion and maintained their phenotype and functional activities in vitro and in vivo (8). Both nTreg and iTreg subsets had a different role in suppressing osteoclastogenesis and bone protection in CIA (20). Recently, while high salt diet diminished the functional activities of nTreg (36), the development and function of iTreg was not comprised (37), demonstrating further differences in biological characteristics between nTreg and iTreg populations.

In this investigation, we have made a previously unidentified observation on the differences of nTregs and iTregs following exposure to inflamed SFs both in vitro and in vivo. While nTregs primed with inflamed SFs significantly lost their Foxp3 and CD25 expression, transdifferentiated to TH17 cells, and failed to suppress Teff proliferation in vitro and in vivo, iTregs induced ex vivo highly maintained Foxp3 and CD25 expression, resisted transdifferentiating to TH17 cells, and maintained their suppressive ability against Teff proliferation. This result further validates previous finding that CD25dim cells convert more into TH17 cells (9). iTregs display more CTLA-4 and even enhance CTLA-4 expression after being exposed to inflamed SFs; it is likely that CTLA-4 involves in the Treg stability. Previous studies have demonstrated that CTLA-4 plays crucial role in the development and function of Tregs (38). In addition, our results also showed that iTreg, but not nTreg, retained the expression of the peculiar markers of eTreg. eTreg plays a vital role in the peripheral immune homeostasis. Diverse eTreg subsets are activated by different antigens and cytokines. Hence, eTreg gains the ability to migrate to specific tissues and exerts its particular types of function on the immune responses (39). Our data indicate that iTreg retains the ability to transdifferentiate into eTreg and maintains wholesome regulatory function under CIA-SFs priming. We also have used two animal models, colitis and CIA, to further demonstrate these differences in vivo.

Another notable observation was that the iTreg, but not nTreg, subset directly targeted inflamed SFs. TNFα, IL-1β, and IL-6 promote the migration and invasion of SFs, and VCAM-1 sustains the synovial inflammation and promotes adherence of peripheral blood cells to endothelial cells (40). RANKL is a critical determinant of osteoclastogenesis (21). We demonstrate that in both in vitro and in vivo experiments, iTregs, but not nTregs, markedly suppressed the production of inflammatory cytokines, proliferation, migration, and invasion of CIA-SFs. Moreover, iTregs, but not nTregs, also inhibited the activation of NF-κB in inflamed SFs. The NF-κB pathway is triggered by proinflammatory cytokines and is a participant in the pathogenic mechanisms of inflamed synovium. A previous study also demonstrated that blocking the NF-κB signaling pathways alleviates the inflammation, bone destruction, and cartilage damage in the joints (25).

These observations raise several possibilities to explain the differences in Treg subsets under arthritic and inflamed SFs conditions. First, both Treg subsets have a different response to IL-6 since we and others have previously reported that iTregs have a lower level of CD126 (one of the IL-6 receptors) relative to nTregs (8). Second, iTregs but not nTregs expressed undetectable SOCS1 and SOCS3 in the response to IL-6 stimulation (41), which promotes TH17 cell differentiation. Third, the iTreg subset that was induced with TGF-β promotes Bcl-2 expression and inhibits the apoptosis of T cells (35). iTregs express a high level of Bcl-2, implicating that iTregs are probably more resistant to apoptosis compared to nTreg (42). Fourth, iTregs may produce more IL-10 than nTregs to suppress the biological activities of the inflamed SF (32). Fifth, iTreg inhibit the production of inflammatory cytokines by inflamed SFs, including TNFα, IL-1β, and IL-6, which promote the stability of iTreg in the presence of inflamed SFs.

CONCLUSION

Together, our results demonstrate that manipulation of the iTreg subset may have potential promise to treat patients with RA in the future.

MATERIALS AND METHODS

Human samples

The study was approved by the medical ethics committees of Institutional Review Boards at Ohio State University and the Medical Ethics Committee of the Third Affiliated Hospital of Sun Yat-sen University. Human blood samples were obtained from the Department of Rheumatology and Immunology at the Wexner Medical center of Ohio State University and the Department of Rheumatology at the Third affiliated Hospital of Sun Yat-sen University. The synovial tissues were obtained from the National Disease Research Interchange or the Department of Rheumatology at the Third Affiliated Hospital of Sun Yat-sen University during knee arthroscopy or synovectomy. The written informed consent was obtained from the donors.

Mice

All of the C57BL/6 and DBA1/J mice from the Jackson laboratory and Nanjing Animal Institute were housed under specific pathogen–free conditions in animal rooms of Penn State Hershey Medical Center, the Ohio State University Wexner Medical Center, and the Sun Yat-sen University. All mice used in this study were 6 to 8 weeks old. All mice were treated by National Institutes of Health (NIH) guidelines for the use of experimental animals with the approval of Penn State University, Ohio State University, and Sun Yat-sen University Committees for the Use and Care of Animals.

Reagents

Dynabeads mouse/human T activator CD3/CD28 for T cell expansion and activation (anti-CD3/CD28–coated beads) were obtained from Gibco (New York, USA), and rhIL-2, rhTGFβ, recombinant mouse IL-6 (rmIL-6), and rmIL-23 were from R&D Systems (Minneapolis, MN). The carboxyfluorescein diacetate succinimidyl ester (CFSE) Cell Division Tracker Kit was purchased from BioLegend (San Diego, CA). Phorbol 12-myristate 13-acetate (PMA), ionomycin, and brefeldin A were purchased from Sigma-Aldrich. Primary antibodies against p–NF-κB, NF-κB, VCAM-1, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and rabbit anti-rat secondary antibody were purchased from Cell Signaling Technology. ELISA kits for TNFα, IL-1β, and IL-6 were from eBioscience. Incomplete Freund’s adjuvant (IFA) was from Difco, MI, USA. The Cell Counting Kit-8 (CCK-8) is from Dojindo Laboratories, Kumamoto, Japan. Killed Mycobacterium tuberculosis strain H37Ra was purchased from Difco. ECM Gel from Engelbreth-Holm-Swarm murine sarcoma (Matrigel) and collagen from rat tail were purchased from Sigma-Aldrich. TRIzol reagent and the AMV First Strand cDNA Synthesis Kit were from Sangon Biotech Co. Ltd., Shanghai, China. Phospho–NF-κB p65 (Ser536) (93H1) rabbit monoclonal antibody (mAb) (catalog no. 3033), NF-κB p65 (C22B4) rabbit mAb (catalog no. 4764), VCAM-1 (D8U5V) rabbit mAb (catalog no. 39036) VCAM-1, and GAPDH (D16H11) XP rabbit mAb were purchased from Cell Signaling Technology.

Naïve CD4+ T cell isolation

Single-cell suspensions were obtained from peripheral lymph nodes and the spleen of DBA1 mice. Splenic erythrocytes were eliminated with red blood cell lysis buffer (Sigma-Aldrich). Total T cells were enriched with nylon wool. Naïve CD4+ T cells that were purified from these naïve CD4+ T cells were purified from these T cells via magnetic cell sorting by auto magnetic cell sorter (MACS) (Miltenyi Biotec, Germany). In brief, total cells labeled with biotin anti-CD8, CD25, B220, CD11b, CD11C, and CD49b antibodies and anti-biotin microbeads were subjected to a depletion followed by a positive selection with CD62L microbeads by auto MACS. The purity of naïve CD4+ T cells were determined as CD4+CD25+CD62L+ with flow cytometry, and the purity of >95% naïve CD4+ T cells was used to induce iTreg. Human naïve CD4+ T cells (CD4+CD25CD45RA+) were isolated from human peripheral blood mononuclear cells (PBMCs) via fluorescence-activated cell sorting; the purity was >99%.

iTregs induction in vitro from naïve CD4+ T cells

For mouse iTreg induction, the naïve CD4+ T cells were cultured in the presence of anti-CD3/CD28–coated beads (cells to beads at a 1:5 ratio), recombinant human IL-2 (rhIL-2; 50 U/ml), and recombinant human TGF-β (rhTGF-β; 2 ng/ml) in 48-well plates for 3 days to induce iTregs. Foxp3 expression was detected by flow cytometry about 60 to 75%. For human iTreg induction, human naïve CD4+ T cells were cultured in the presence of anti-CD3/CD28–coated beads (cells to beads at 1:5 ratio), rhIL-2 (50 U/ml), and rhTGF-β (5 ng/ml) in 48-well plates for 3 days. Foxp3 expression was detected by flow cytometry about 70 to 80%.

nTregs and pTregs isolation

Mouse CD4+CD25+ cells were sorted from the thymus of DBA1/J mice (nTregs) or peripheral lymph node cells (pTregs) with a purity of >95%. Around 75% CD4+CD25+ cells expressed Foxp3 as detected by flow cytometry. CD4+CD25+CD127low (purity of >99%) cells sorted from human PBMCs were used as human nTregs. Foxp3 expression was detected by flow cytometry about 75 to 85%. Fresh mouse nTreg and pTreg were activated with anti-CD3/CD28–coated beads (1:5 to cells) in the presence of rhIL-2 (50 U/ml) for 3 days before cocultured with inflamed SFs when iTreg is under induction. Fresh human nTregs were amplified with anti-CD3/CD28–coated beads (1:3 to cells) in the presence of rhIL-2 (300 U/ml).

Tregs suppressive function in vitro

Lymphocytes were harvested, and Teffs from lymphocytes were enriched by nylon column and subsequently labeled with 1 μM CFSE (CellTrace CFSE Cell Proliferation Kit, Thermo Fisher Scientific). Teffs (0.3 million per well) were seeded in a 96-well plate, and allophycocyanins (APCs) irradiated with 30 gray (0.3 million per well) were added in the culture medium simultaneously. iTregs or nTregs primed with or without CIA-SFs for 3 days were added to some wells at various ratios to Teffs, as indicated in individual experiments. After 72-hour incubation, the proliferation of CD8+ T cells was determined by flow cytometry as previously reported (10).

TH17 cell differentiation

Naïve CD4+ T cells were activated with anti-CD3/CD28 beads (Tregs: beads, 1:3) in the presence of rmIL-1β (10 ng/ml), rmIL-6 (100 ng/ml), rmIL-23 (50 ng/ml), rhTGF-β (10 ng/ml), anti–IFN-γ (5 μg/ml), and anti–IL-4 (5 μg/ml) for 3 days.

Tregs coculture with inflamed SFs

Inflamed SFs were cultured and activated with TNFα (10 ng/ml) stimulation for 24 hours before coculture. For a cell-to-cell contact coculture, iTregs or nTregs (Tregs: SFs, 10:1) were added to the activated inflamed SFs directly and cocultured in the presence of anti-CD3/CD28–coated.

beads (Tregs: beads, 1:3) for 3 days. For transwell assays, Tregs were cultured in the insert well of a 24-well transwell plate (0.4-μm pore, Corning) in the presence of anti-CD3/CD28–coated beads (Tregs: beads, 1:3) and while SFs were cultured in the lower chamber separately to coculture for 3 days. nTregs/pTregs cultured under TH17-polarizing condition (the same condition for TH17 cell differentiated from naïve CD4+ T cells) or stimulated with exogenous rmIL-6 (100 ng/ml) for 3 days were set up as a positive control. Foxp3 and IL-17A expression in Treg subsets were determined by flow cytometry. For in vivo transfer, Tregs after being cocultured with or without CIA-SFs for 3 days were labeled with biotin anti-CD4 antibody and anti-biotin microbeads and then subjected to a positive selection by auto MACS to acquire CD4+ T cells. These CD4+ T cells were used as Tregs for transfer.

Inflamed SF-mediated osteoclastogenesis

Bone marrow cells from DBA1/J mice were labeled with Biotin anti-CD11b antibody and anti-microbeads followed by a positive selection by auto MACS to acquired CD11b + cells (purity > 95%). Then, CD11b+ cells were suspended and cultured in 24 well (1 × 106 cells per well) in α-minimum essential medium culture medium [containing 10% fetal bovine serum (FBS)] in the presence of mouse M-CSF (50 ng/ml) for 3 days, and the adherent cells were used as OCPs. CIA-SFs were added directly to OCPs and then cocultured for 3 days to induce the differentiation of osteoclast. Tregs were seeded in the insert well in the presence of anti-CD3/CD28–coated beads (1:3 to cells) and also added to coculture with OCPs separately. OCPs cultured in the presence of mouse RANKL (50 ng/ml) and mouse M-CSF (50 ng/ml) or cocultured with CIA-SFs only for 3 days were set as a control. To evaluate osteoclast formation, cells were stained with tartrate-resistant acid phosphatase (TRAP) (Sigma-Aldrich) according to the manufacturer’s instructions, and TRAP+ cells were analyzed by microscopy.

Flow cytometry

The following monoclonal antibodies conjugated with fluorescein isothiocyanate, phycoerythrin, PerCP-Cy5.5, APC, or Alexa Fluor 647 were purchased from BioLegend: anti-mouse CD4 (GK1.5), CD25 (3C7), CD62L (MEL-14), PD-1 (29F.1A12), IFN-γ (XMG1.2), IL-17A (TC11-18H10.1), CD11b (M1/70), CD90.2(Thy1), CD39 (Duha59), ICOS (C398.4), OX-40 (OX-86), GITR (YGTTR/65), RANKL (IK22/5), Blimp1(5E7), TIGIT (1G9), IRF4 (IRF4.3E4), CD137 (also called 4-1BB, 17B5), Helios (22F6), NRP-1 (3E12), CTLA-4 (UC10-4B9), IFN-γ (4S.B3), IL-17A (BL168), anti-human CD11b (M1/70), Thy1 (5E10), CD4 (A161A1), CD25 (BC96), CD127 (A019D5), and Foxp3 (206D). For cell surface antigen staining, related antibodies for different antigens were diluted with phosphate-buffered saline (PBS) at 1:200, and cells were resuspended in 200-μl dilution and stained for 15 min in the dark at 4°C. The cells were washed with PBS twice and then resuspended with 500 μl of PBS followed by detecting the expression of related markers by flow cytometry immediately. For intranuclear Foxp3 staining, eBioscience Foxp3/Transcription Factor Staining Buffer Set (eBioscience) was used, and the staining was performed as described in the protocol provided with the kit. For intracellular cytokine staining, cells were stimulated with PMA (0.05 μg/ml) and ionomycin (0.5 μg/ml) (Sigma-Aldrich) for 1 hour and then with brefeldin A (10 μg/ml), followed by incubating these cells for another 4 hours at 37°C in a humidified incubator with 5% CO2. After stimulation, first, cell surface antigen staining was performed, and the cells were fixed with fixation buffer (BioLegend) for 30 min at room temperature in the dark followed by permeabilization with 1× Intracellular Staining Perm Wash Buffer (BioLegend) for 15 min. The fixed/permeabilized cells were resuspended with 200 μl dilution in which monoclonal antibodies to cytokines were diluted at 1:100 in 1× Intracellular Staining Perm Wash Buffer (BioLegend) and stained overnight. For intranuclear and intracellular staining together, we followed the protocol for intranuclear staining. After staining, cells were washed with PBS for at least twice, resuspended in 500-μl PBS, and analyzed by flow cytometry.

Proliferation assay

CIA-SFs were cocultured with or without iTregs or nTregs sorted as described above (ratio 1:10) in a 96-well plate for 3 days. To detect the cell viability of CIA-SFs after being cocultured with the two Treg subsets, CCK-8 (20 μl) was added to each well of the plate, and the cells were incubated at 37°C in 5% CO2 incubator for 2 hours. We measured the absorbance at 450 nm with a microplate reader.

Migration and invasion assay

In case cell-cell contact is not necessary for the effect between Treg subsets and CIA-SFs, they were cocultured in 24-well transwell plate (9). Tregs (2.5 × 105) were seeded in the lower chamber, which was filled with Dulbecco’s modified Eagle’s medium (DMEM) containing with 10% FBS, and then CIA-SFs (2.5 × 104) were cultured in the upper wells with DMEM without FBS. After 12 hours, CIA-SFs that passed through the membrane were fixed with 4% methanol and stained with crystal violet solution for 10 min (43). Photomicrographs of five random fields were obtained, and the numbers of CIA-SFs were counted. Previous study has demonstrated that RA-SF is one of the key players in invading and degrading cartilage without additional stimuli by matrix degradation. CIA-SFs’ invasiveness was determined using ECM gel or collagen gel invasion (1 mg/ml, 100 μl per well) (13). The upper side of the chambers was coated with fresh Matrigel. Tregs (1 × 106) were seeded in the lower chamber, which was filled with DMEM containing 10% FBS as described above, and then CIA-SFs (1 × 105) were cultured in the upper wells with DMEM without FBS for 24 hours. CIA-SFs that passed through the Matrigel membrane were fixed and counted in the microscope as described above.

Tregs tracking in CIA mice

To investigate whether transferred Tregs can migrate into the joints of CIA mice, 3 × 106 of Tregs were fluorescently labeled with the Cell Tracker Red CMTPX dye (Thermo Fisher Scientific) and adoptively infused into CIA mice 14 days after immunization. Twenty-four hours later, different tissues including the heart, lung, liver, spleen, lymph node, kidney, and joint from CIA mice were collected and immediately imaged by an IVIS imaging system to determine the distribution of Tregs.

Induction and assessment of arthritis

CIA was induced in DBA1/J mice through intradermal injection of 100 μl of emulsion of bovine collagen II (2 mg/ml) and Complete Freund’s Adjuvant containing inactivated M. tuberculosis (4 mg/ml) [at 1:1 (v/v) ratio]. A second immunization with 100 μl of IFA was given on day 14 after the first CII/CFA immunization.

For assessment of arthritis, clinical signs of the mice were assessed for every 3 days as follows: 0, normal joints; 0.5, swelling of one or more digits; 1, erythema and mild swelling of the ankle joint; 2, mild erythema and mild swelling involving the entire paw; 3, erythema and moderate swelling involving the entire paw; and 4, erythema and severe swelling involving the entire paw. The scores for four limbs were added together, and the maximal score for each mouse is 16. For the prophylactic function of Treg in CIA model, 3 × 106 nTregs or iTregs were intravenously transferred to DBA/1J mice on day 14 after the first CII/CFA immunization. For the therapeutic function of Treg in CIA model, 3 × 106 CD4+ Tregs with or without being primed with CIA-SFs for 3 days were transferred to DBA/1J after the onset of arthritis was observed, around 30 days after the first CII/CFA immunization.

Induction and assessment of colitis

Naïve CD4+ T cells (0.6 million) were isolated from CD90.1+ Foxp3GFP reporter DBA1/J mice (wild type) and intraperitoneally injected into Rag1−/− mice alone or together with 0.6 × 106 CD4+ Tregs with or without being primed with CIA-SFs for 3 days. The weights of all the Rag1−/− mice were monitored every day. Forty-five days after the transfer, the cells from the spleen, MLNs, and cLP were harvested from the recipient mice and were stimulated with PMA (50 ng/ml) and ionomycin (500 ng/ml). Production of IL-17A and IFN-γ by the CD4+ T cells derived from CD90.1+ mice was determined by flow cytometry.

Isolation of inflamed synovial tissues

For CIA-SFs isolation, CIA mice immunized for about 45 to 60 days were euthanized by cervical dislocation before dissection of inflamed joints. The four limbs of CIA mice were dissected at the joints, respectively. The skin, nail, muscle, and tendon on the specimens were removed as thoroughly as possible. We opened up the joint spaces of paws and ankles and exposed the synovial tissues in the digestive fluid with collagenase type IV (1 mg/ml), followed by strongly shaking for 1 hour at 37°C in the shaker. The SFs in the media were centrifuged at 1400 rpm for 4 min, and the cell pellet was resuspended in 10 ml of fresh media with 20% FBS. The SFs were subcultured at 90% confluence before characterization at passage four. When CIA-SFs were cultured during the passages 4 to 7 (three- to fivefold expansion), CIA-SFs were labeled with anti-mouse CD90.2-biotin antibody and anti-biotin microbeads followed by a positive selection with auto MACS. The purity of CIA-SFs was determined as Thy1+CD11b by flow cytometry as previously reported (purity of >90%) (9).

For RA-SFs isolation, the synovial tissue biopsies from RA patients were finely minced into small pieces and transferred to a tissue culture flask in DMEM (Hyclone Laboratories) supplemented with 20% FBS (Gibco), penicillin (100 U/ml; Gibco), and streptomycin (100 μg/ml; Gibco) in a humidified incubator at 37°C in 5% CO2. RA-SFs were labeled with anti-human Thy1-biotin antibody and anti-biotin microbeads followed by a positive selection with auto MACS. The purity of CIA-SFs was determined as Thy1+CD11b by flow cytometry (purity of >90%).

Histopathology

For knee joints in CIA, both hind limbs from the CIA mice were dissected and incubated in 10% buffered formalin. Then, the specimens were sectioned and stained with H&E. Two pathologists blindly evaluated the global histological changes according to the proliferation of SFs, the infiltration of inflammatory cells, and the thickness and erosions of joint bone and cartilage. For colon in the colitis mice, we cut the 1-mm distal colon incubated in 10% buffered formalin and subsequently trimmed so as to render a transverse section. Subsequently, the specimens were stained with H&E. Two pathologists blindly evaluated the global histological changes according to the inflammatory cell infiltration, edema, ulcer, the intact of the epithelial cells, and goblet cells.

Enzyme-linked immunosorbent assay

The supernatants from coculture medium were subjected to ELISA for measurement of TNFα, IL-6, and IL-1β.

Western blot analysis

The collected CIA-SFs were lysed in the radioimmunoprecipitation assay lysis buffer. We detected the total protein concentration with the bicinchoninic acid protein assay and separated the proteins by using sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and then the membranes were blocked with 5% skim milk for 2 hours at room temperature. The membranes were incubated in antibodies against GAPDH (1:1000), phospho–NF-κB p65 (Ser536) (p–NF-κB, 1:1000) (Cell Signaling Technology, catalog no. 3033), NF-κB p65 (C22B4) rabbit mAb (Cell Signaling Technology, catalog no. 4764) p–NF-κB (1:1000), NF-κB (1:1000), and VCAM-1 (1:1000) overnight at 4°C. The next day, we washed the membranes three times to remove the redundant antibodies and incubated the membranes in horseradish peroxidase–conjugated anti-rabbit secondary antibody (1:1000) for 2 hours at room temperature. For equal loading test, the blotted membranes were stripped and reprobed with primary and secondary antibodies. The autoradiographic films were scanned and quantitated using Quantity One software (Bio-Rad, Hercules, CA).

Quantitative real-time PCR

RNA was harvested from CIA-SFs or Tregs using TRIzol reagent according to the manufacturer’s protocol. RNA quantity and purity were confirmed with a Nanodrop spectrophotometer (Thermo Fisher Scientific). Complementary DNA (cDNA) was synthesized using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems by Life Technologies). PCR reactions were run using the QuantStudioTM 12 K Flex real time PCR system v1.2.2 (Life Technologies TM) with SYBR Green. The following murine primers were used: IL23R (forward: 5′-GGTCCAAGCTGTCAATTCCCTAGGC-3′; reverse: 5′-AGCCCTGGAAATGATGGACGCA-3′), T-bet (forward: 5′-GTCTGGGAAGCTGAGAGTCG-3′; reverse: 5′-GAAGGACAGGAATGGAACA-3′), RORγt (forward: 5′-CATCGACAAGGCCTCCTAGC-3′; reverse: 5′-TTCCACATGTTGGCTGCACA-3′), Gata-3 (forward: 5′-AGGGACATCCTGCGCGAACTGT-3′; reverse: 5′-CATCTTCCGGTTTCGGGTCTGG-3′), IL-6 (forward: 5′-CACAAGTCCGGAGAGGAGAC-3′; reverse: 5′-CAGAATTGCCATTGCACAAC-3′), TNFα (forward: 5′-TTATCTCTCAGCTCCACGCC-3′; reverse: 5′-TGCGCACTGAAAGCATGATC-3′), MMP9 (forward: 5′-GTATGGTCGTGGCTCTAAGC-3′; reverse: 5′-AAAACCCTCTTGGTCTGCGG-3′), IL-1β (forward: 5′-TTCAGGCAGGCAGTATCACTC-3′; reverse: 5′-CCACGGGAAAGACACAGGTAG-3′), IL-15 (forward: 5′-CATATGGAATCCAACTGGA TAGATGTAAGATA-3′; reverse: 5′-CATATGCTCGAGGGACGTGTTGATGAACAT-3′), VEGF (forward: 5′-ACTGGACCCTGGCTTTACTGCT-3′; reverse: 5′-TGATCCGCATGATCTGCATGGTG-3′), and MMP14 (forward: 5′-GTGCCCTAGGCCTACATCCG-3′; reverse: 5′-TTGGGTATCCATCCATCACT-3′).

Statistical analysis

The data were expressed as means ± SEM and analyzed using GraphPad Prism (GraphPad Software, San Diego, CA, USA). Statistical analysis and comparisons were performed using one-way analysis of variance (ANOVA), and two groups with different conditions comparisons were performed using two-way ANOVA. P < 0.05 is considered as statistically significant.

SUPPLEMENTARY MATERIALS

Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/44/eabb0606/DC1

https://creativecommons.org/licenses/by-nc/4.0/

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.

REFERENCES AND NOTES

Acknowledgments: Funding: This work was supported in part by grants from the NIH R01 AR059103, NIH STAR Award, and R61 AR073409 (to S.G.Z.), China National Natural Science Foundation (nos. 81671611 and 81871224 to S.Y.), and Program for Guangdong Introducing Innovative and Entrepreneurial Teams (2016ZT06S252 to F.H.). Author contributions: S.Y. and X.Z. performed most of the experiments; analyzed, acquired, and interpreted the data; and drafted the manuscript. J.C. performed most of the experiments analysis and manuscript revision. J.D., R.L., D.Z., H.Z., Y.X., Y.Li., W.W., J.Z., and J.W. performed some experiments. N.O., W.H., Y.P., H.X., B.S., F.H., and Y. Lu analyzed the data and revised the manuscript. S.G.Z. conceived and designed the study and wrote the manuscript. All coauthors approved the final version. Competing interests: The authors declare that they have competing interest. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.
View Abstract

Stay Connected to Science Advances

Navigate This Article