Therapeutic Effects of Mesenchymal/Stromal Stem Cells and Their Derived Extracellular Vesicles in Rheumatoid Arthritis Authors: Dimitrios Tsiapalis, Achilleas Floudas, Tobias Tertel, Verena Boerger, Bernd Giebel, Douglas J Veale, Ursula Fearon, Lorraine O’Driscoll Published: 2023 ABSTRACT AbstractCurrently available therapies for rheumatoid arthritis (RA) are inadequate to alleviate the inflammation and reduce joint damage. While the immune-regulatory effect of human mesenchymal/stromal stem cells (MSCs) extracellular vesicles (EVs) has been tested in many inflammation-related diseases, little is known regarding their effect on patients with RA. Thus, we assessed the effect of human MSCs and MSC-EVs (from naïve or IFN-β-primed MSCs) on CD4+ T cells from patients with RA. Moreover, we investigated the effect of MSC-EVs on RA patients-derived synovial fibroblasts (FLS). MSC-EVs were prepared using a PEG precipitation followed by ultracentrifugation-based protocol. Applied to RA CD4+ T cells, EVs from IFN-β-primed MSCs, suppressed the expression of more key RA-associated cytokines (IL-4, GM-CSF IFN-γ, IL-2, TNF-α), and decreased CD4+ T-cell polyfunctionality than MSCs or EVs from naïve MSCs. MSCs mediated a slight decrease in the frequency of T-regulatory cells, while MSC-EVs rescued the frequency of T-regulatory cells. MSCs significantly inhibited CD4+ T-cell proliferation (P < .05), while no inhibition was observed in response to EV preparations. EVs from IFN-β-primed MSCs inhibited (P < .01) RA FLS migration and downregulated (P < .05) RA FLS surface markers CD34 and HLA-DR. Collectively, we demonstrated the immune-modulatory function of MSCs and their derived EVs in RA CD4+ T cells, which could be further enhanced by priming MSCs with IFN-β. Moreover, EVs from IFN-β-primed MSCs more efficiently inhibit RA FLS migration, and expression of RA FLS-related surface markers, suggesting these EVs as a potent therapy for RA.Keywords: rheumatoid arthritis, mesenchymal/stromal stem cells, extracellular vesicles, CD4+ T cells, RA synovial fibroblasts Introduction Rheumatoid arthritis (RA) is a chronic progressive autoimmune disease characterized by a spectrum of clinical manifestations leading to damage to the joint, increased comorbidities and mortality.1 Importantly, RA can be severely detrimental to the patients’ quality of life and is associated with high economical and societal costs. There is no cure for RA and current therapeutic strategies using corticosteroids, non-biologic disease-modifying anti-rheumatic drugs (DMARDs) or biological DMARDs focus on treating the symptoms of the disease but have many side-effects.2-4 In addition, a significant proportion of patients have suboptimal or no response to current therapies, thus there is a need for more stratified effective treatment to alleviate the inflammation associated with the pathogenesis of all RA patients and to inhibit damage to the joint. T cells infiltrating the joint are key contributors to RA pathogenesis and synovial inflammation, resulting in joint damage and loss of function.5 Formation of synovial B cell-T cell aggregates leads to more aggressive synovial inflammation and erosive disease in RA.6 Importantly, polyfunctional T-cell responses characterized by the simultaneous expression of several proinflammatory cytokines have been linked to disease activity and response in patients with psoriatic arthritis, Down syndrome-associated arthritis and RA.7-10 A recent study by Floudas et al. identified metabolically active, polyfunctional T-cell subpopulations that are resistant to regulatory T-cell-mediated suppression in RA patient synovial biopsies.11 Interestingly, these cells are present in the joints of individuals prior to the clinical manifestation of inflammation, further highlighting the importance of polyfunctional T cells in the pathogenesis of RA.11 One of the most abundant cells of the inflamed joint are the synovial fibroblasts (FLS). Under homeostatic conditions, FLS are responsible for the production of the extracellular matrix, lubrication of the joint, and protection of the cartilage. However, FLS become dysregulated and hyperproliferative in RA and adopt an aggressive proinflammatory phenotype that significantly contributes to synovial inflammation and joint destruction.12 Specifically, RA FLS have increased proinflammatory cytokine production and migrative/invasive capacity, which lead to cartilage degradation and subsequent bone erosions.13 Therefore, being able to simultaneously reduce T-cell polyfunctionality and limit the migration capacity of synovial fibroblasts could offer novel opportunities for therapeutic intervention. The immune-modulatory properties of mesenchymal/stromal stem cells (MSCs) have been widely assessed in various in vitro and in vivo settings.14-22 In RA in vivo models, administration of MSCs significantly improved the severity of symptoms by modulating the inflammatory response and by reducing the damage to bone and cartilage.23,24 Despite MSCs great potential, their application in clinics is lagging behind due to concerns relating to heterogeneity of MSCs resulted to donor variations, stemness stability, inadvertent differentiation, and limited expansion during the preparation of MSCs, homing capacity of MSCs, risk of rejection and transformation of MSCs into malignant sarcoma cells at the state of application.25,26 Many, if not all, of the beneficial effects of MSCs are attributed to their paracrine action through the release of extracellular vesicles (EVs), suggesting that EVs can represent the therapeutic potency of MSCs, bypassing the concerns with the use of MSCs.27,28 Importantly, compared to their cells of origin, MSC-EVs can be easily stored, sterilized, and distributed for clinical use.29 Recent indications suggest that the beneficial effects of MSC-EVs can be further augmented by modifying the way in which MSCs are cultured (hypoxia,30,31 mechanical stimulation,32 3D culture,33 priming with pharmacological and chemical agents34,35 and priming with cytokines36-40). Among other cytokines, it was demonstrated that interferon beta (IFN-β) can enhance the immune-modulatory properties of MSCs, by increasing the expression of molecules secreted by MSCs that have immune-modulatory activities.41 Accordingly, we hypothesized whether priming the MSCs with IFN-β can enhance the immune-regulatory potential of MSC-EVs. While the immune-regulatory effect of MSC-EVs has been assessed in many diseases, little is known of their potential with respect to their effect on RA pathogenic mechanisms. Therefore, in this study, we aimed to assess and compare the effect of MSCs, MCS-EVs, and IFN-β-primed MSC-EVs on the expression and release of inflammatory cytokines from CD4+ T cells associated with the pathogenesis of RA; the frequency of T regulatory cells; T-cell polyfunctionality, and the proliferation of CD4+ T cells. In addition, the effect of MSC-EVs (primed and nonprimed) on RA FLS was evaluated. Materials and Methods Study Approval Peripheral blood and synovial tissue samples were collected from RA patients that were recruited from the Rheumatology Department, St. Vincent’s University Hospital, UCD. All patients gave fully informed written consent approved by the institutional Ethics Committee, and research was performed in accordance with the Declaration of Helsinki. Arthroscopies were performed under local anesthetic using a Wolf 2.7 mm needle, and synovial tissue biopsies were obtained from the site of inflammation under direct visualization as previously described.42 RA Peripheral Blood Mononuclear Cells Isolation and Culture of Primary RA FLS PBMCs from RA patients were obtained by density gradient centrifugation (Lymphoprep, Norway). PBMCs were washed 3 times with phosphate-buffered saline (PBS, Sigma Aldrich, Ireland) and suspended in in Rosewell Park Memorial Institute medium (RPMI, Sigma Aldrich) supplemented with 10% foetal bovine serum (FBS, Sigma Aldrich) and 1 IU/mL penicillin/0.1 mg/mL streptomycin (1% P/S, Sigma Aldrich). To isolate primary RA FLS, RA synovial biopsies were digested with 1 mg/mL collagenase type 1 (Worthington Biochemical, Freehold, NJ, USA) in RPMI-1640 (Gibco-BRL, UK) for 4 hours at 37°C in humidified air with 5% CO2. Dissociated cells were grown to approximately 80% confluency in RPMI-1640 (Gibco-BRL, UK), 10% FBS (Sigma Aldrich), and 1% P/S (Sigma Aldrich) before passaging. These RA FLS were used at passages 3 to 5. Human Bone Marrow MSCs Isolation and Culture MSCs were raised from human bone marrow samples of 2 healthy donors as previously described43 after obtaining ethical approvals and patient consent. Obtained MSCs were cultured at a density of 250 000 in T-175 culture flasks (Corning, USA) in low glucose Dulbecco’s-modified Eagle medium (DMEM) (Sigma Aldrich) supplemented with 2 mM L-glutamine (Sigma Aldrich), 1% P/S (Sigma Aldrich), 5 IU/mL heparin (Sigma Aldrich), and 10% human platelet lysate (hPL; STEMCELL Technologies, UK) and maintained at 37°C in a humidified atmosphere, with 5% CO2. The medium was changed every 3 to 4 days. Characterization of MSCs Fluorescence-activated cell sorting (FACS) analysis using anti-CD73, anti-CD44, anti-CD90, and anti-CD105 antibodies as positive markers and anti-CD31, anti-CD34, and anti-CD45 antibodies as negative markers was performed at the time of media harvesting for the collection of EV samples (Supplementary Materials). Collection of EVs From MSC-Conditioned Media At passage 6 and at ~80% confluency, cells were primed with 1500 UI/mL IFN-β (Merck Millipore, Ireland) for 1 hour in hPL-free DMEM as previously described.41 Primed with IFN-β and non-primed MSCs were washed with PBS and cultured for an additional 48 hours in media containing EVs-depleted hPL (hPL was EV depleted by ultracentrifugation at 120 000 × g for 18 hours at 4°C; Ti70 rotor, κ-factor 298.0, Beckman Coulter, Ireland), and the CM was collected. To remove cells, CM (100 mL) was centrifuged at 2000 × g for 10 minutes. To remove debris and large-sized EVs, supernatants were centrifuged at 10 000 × g for 45 minutes at 4°C in a Sorvall ST 8 small benchtop centrifuge using an HIGHConic III fixed angle rotor (κ-factor 2488, Thermo Fisher Scientific, Ireland), followed by filtration using a 0.2-µm pore filter membrane (Thermo Fisher Scientific). Next, small-sized EVs were isolated using the polyethylene glycol 6000 (PEG) precipitation protocol as previously described43-45 followed by ultracentrifugation at 100 000 × g for 91 minutes (Ti70 rotor, κ-factor 298.0, Beckman Coulter, Ireland) at 4°C. MSC-EV samples were resuspended in 0.5 mL of 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES)/0.9% sodium chloride (NaCl, Thermo Fisher Scientific) and stored at−80°C. Characterization of MSC-EVs MSC-EV samples were characterized according to the current MISEV guidelines.46 Particle concentration and particle size of MSC-EV samples were measured using nanoparticle tracking analysis (NTA; NS500, NanoSight, UK). Aliquots of MSC-EV samples were injected into the NTA system under constant flow conditions (flow rate = 50), and videos of the particles were recorded and analyzed using NTA 3.1.54 software. Results were presented as the concentration of particles/mL normalized to 1 million (M) cells. The protein concentration of MSC-EV samples was quantified using the Bradford assay using the Bio-Rad protein assay dye reagent (Bio-Rad, UK) according to manufacturer’s instructions. Results are normalized to 1 M cells. Immunoblotting analysis for CD63, syntenin-1, CD105, and calnexin on MSC-EVs and MSCs lysates was performed as previously described.47 Detailed information is provided in Supplementary Materials. Transmission Electron Microscopy (TEM) was performed as previously described.48 10 µL of MSC-EV samples were placed on formvar carbon-coated nickel grids (Ted Pella, Inc, USA) and settled for 10 minutes. A droplet of paraformaldehyde (PFA, 2%) was placed on parafilm, the grid placed on top for 10 minutes, contrasted in 2% uranyl acetate (BDH Industries, India), and imaged by JEOL TEM-2100. Imaging Flow Cytometry (IFCM) analysis for CD9 and CD63 of MSC-EV samples was performed as previously described,49 and detailed staining protocol is provided in Supplementary Materials. The gating strategy used for the MSC-EVs analyses is presented in Supplementary Fig. S1. Results indicated the number of small-sized objects/mL normalized to 1 M cells. Coculture of RA PBMCs With MSCs or MSC-EVs MSCs at passage 6 were cultured in 96-multiwell flat bottom culture plates (Corning-Costar, Ireland) at the density of 25 000 cells/well in DMEM (Sigma Aldrich) supplemented with 10% hPL (STEMCELL Technologies, UK). After 24 hours for cell adhesion, the medium was removed and replaced with RA PBMCs at 250 000 cells/well in RPMI (Sigma Aldrich) supplemented with 10% FBS (Sigma Aldrich). T-cell stimulation was achieved by adding anti-CD3 (clone: OKT3, 1 µg/mL) and anti-CD28 (clone: 28.2, 2 µg/mL) (eBioscience, USA) for 3 days. The effect of MSC-EV samples was studied with the same procedure and time course, by adding MSC-EV samples (5 µl in 200 µl of RA PBMC culture media) 1 hour after seeding the PBMCs, according to the concentration of particles/mL produced by 25 000 MSCs as determined by NTA (Table 1). Table 1.Quantity and number of particles/mL of MSC-EV samples added in PBMC cultures.MSC-EV samplesQuantity of MSC-EVs samples in 200 µl of culture mediaNumber of particles/mL produced by 25 000 MSCsMSC-EVs first donor5 µl4 × 108 particles/mLIFN-β-primed MSC-EVs first donor5 µl3.1 × 108 particles/mLMSC-EVs second donor5 µl3.6 × 108 particles/mLIFN-β β primed MSC-EVs first donor5 µl2,87 × 108 particles/mL Quantity and number of particles/mL of MSC-EV samples added in PBMC cultures. Flow Cytometry Analysis for the Assessment of T-Cell Cytokine Expression and T-Regulatory Cells For the detection of cytokine expression, T cells were stimulated with Cell Stimulation Cocktail (phorbol myristate acetate/ionomycin) (eBioscience) for 1 hour prior to the addition of 0.05 μg/mL brefeldin A and Monensin (Thermo Fisher Scientific) for 4 hours of incubation. Then, cells were stained with antibodies against extracellular targets; anti-CD45, anti-CD3, anti-CD8, and anti-CD4. Cells were fixed and permeabilized using the intracellular FOXP3 staining kit (eBioscience) as per the manufacturer’s instructions. Following the fixation and permeabilization step, cells were washed in perm buffer and incubated with antibodies against intracellular targets for 30 minutes at 4°C with antibodies against intracellular targets; anti-interleukin-2 (IL-2), anti-tumor necrosis factor alpha (TNF-α), anti-interferon gamma (IFN-γ), anti-interleukin-17A (IL-17A), anti-interleukin-22 (IL-22), anti-granulocyte macrophage-colony stimulating factor (GM-CSF), and anti-interleukin-4 (IL-4). Detailed staining of T cells is provided in Supplementary Materials. For the detection of T regulatory cells, T cells were stimulated with Cell Stimulation Cocktail (phorbol myristate acetate/ionomycin) (eBioscience) for 1 hour and stained for viability and extracellular markers; anti-CD45, anti-CD3, anti-CD8, anti-CD4, anti-CD25, and anti-CD127, followed by staining against the forkhead box P3 (FOXP3). Results were presented as % fold change to the no treatment condition. T-Cell Proliferation T-cell proliferation assay was performed as previously described.11 Briefly, RA PBMCs (coculture with MSCs and MSC-EVs) at 250 000 cells/well in 96-well plates were labeled with cell trace violet (CTV, Thermo Fisher Scientific) according to manufacturer’s instructions followed by 5-day stimulation with anti-CD3 (clone: OKT3, 1 µg/mL) and anti-CD28 (clone: 28.2, 2 µg/mL) (eBioscience). Cells were then stained for viability and combinations of extracellular antibodies; anti-CD5, anti-CD8, and anti-CD4. Cells were fixed 1% PFA for 10 minutes at 4°C prior to acquisition on LSR Fortessa cytometer (BD, UK). Analysis was performed on FlowJo (V.10) using the Proliferation platform of FlowJo (V.10). Simplified Presentation of Incredibly Complex Evaluations Analysis The simultaneous production of multiple cytokines by T cells was evaluated by utilization of Simplified Presentation of Incredibly Complex Evaluations (SPICE) is a supervised dimensionality reduction algorithm that is based on Boolean gating of flow cytometric data. The pie segments represent the percentage of cells that produce a specific combination of cytokines, while the pie arcs indicate the cytokines that are produced. Cytokine Quantification The release of cytokines in CD4+ T-cell-free supernatants was performed using a Mesoscale Discovery (MSD, USA) multiplexed assay for 9 cytokines (IL-9, IL-4, IL-21, IL-10, IL-2, IL-17A, GM-CSF, IFN-γ, and TNF-α) according to manufacturer’s instructions. Results were presented as % fold change to the no treatment condition. Migration Assay RA FLS (25 000) were seeded in 24-well plates for 24 hours and serum starved as previously described. RA FLS were then stimulated with TNF-α (1 ng/mL; R&D systems, UK) for 24 hours and then a single scratch wound was induced through the middle of each well with a sterile 200 µL pipette tip and followed by treatment with the MSC-EV samples (15 µL in 600 mL of RA FLS culture media). RA FLS migration across the wound margins was assessed and photographed using a phase-contrast microscope. Wound closure was assessed by measuring the area at 0 hour and after 24 hours using ImageJ software. Three images per group per RA FLS patient were analyzed. Statistical Analysis Statistical analysis was performed using Prism V.9 (GraphPad Software, Inc., CA). Unpaired 2-tailed standard Student’s t test were to compare 2 groups. One-way ANOVA followed by Dunnett’s post-test was used for analysis of multiple comparison groups. Results are expressed as mean of a minimum of 3 independent experiments (except when noted otherwise) ± SEM. Statistical significance was set at *P < .05; **P < .01; ***P < .001. Results IFN-β-Primed and Non-Primed MSCs Release Comparable Amounts of EVs Here, we aimed to access the impact of EVs prepared from conditioned media of primary BM-derived MSCs (from 2 donors), either raised conventionally or in the presence of IFN-β, on the RA pathogenic mechanisms. To this end, we at first characterized the impact of IFN-β on the phenotype of cultured MSCs. At the time of media harvesting to proceed with the collection of MSC-EV samples, we used flow cytometry to characterize the BM-derived cells (naïve and primed with IFN-β). Cells raised from either naïve or primed with IFN-β exhibited similar immunophenotype to MSCs as revealed by the expression of positive stromal markers CD73, CD44, CD90, and CD105 and being negative for the endothelial (CD31) and hematopoietic (CD34 and CD45) markers (Supplementary Fig. S1), indicating that IFN-β did not recognizably alter MSC phenotypes. For the functional studies, we prepared EVs from conditioned media by PEG precipitation method followed by ultracentrifugation. We showed that the concentration of particles/mL/1 M cells was higher (P < .05) in the EVs derived from the naïve MSCs than those primed with IFN-β from both donors of MSCs, while the mean size of primed with IFN-β MSC-EVs was larger than those from the naïve MSC-EVs (Table 2, Supplementary Fig. S2). Consistent to the NTA results, protein concentration analysis revealed significantly (P < .05) higher protein contents in the EVs from naïve MSCs from first donor and marginally higher from the second donor (Table 2). At TEM analysis, EVs samples derived from either naïve or primed MSCs exhibited the typical round morphology of EVs with a peripheral membrane surrounding as indicated by the electron-dense content (Fig. 1A). Immunoblotting analysis demonstrated no apparent differences in the presence of the EV-associated transmembrane protein CD63, the MSC-related marker CD105, the EV-associated cytosolic protein Syntenin-1, and the absence of the negative EV marker Calnexin in EV samples derived from both culture conditions (Fig. 1B). Analysis for CD9 and CD63-positive MSC-EVs using IFCM (Supplementary Fig. S3) showed no apparent differences in the number of small-sized CD9 between naïve and IFN-β-primed MSCs. Instead, higher number of CD63 small-sized EVs was observed in the EV samples derived from naïve MSCs as indicated by the number of objects/mL/1 M cells (Fig. 1C). Table 2.Characterization of EV populations from naïve and IFN-β-primed MSCs by NTA and Bradford assay. NTA analysis showed significant higher particles concentration (particles/mL/1 M cells) in EV samples collected from naïve MSCs from both donors as compared to those collected from IFN-β-primed MSCs. The size of the particles was significant higher on EV samples from IFN-β-primed MSCs. Results summarize the mean ± SEM of n = 6 repeat measurements (videos) of each EV population from 2 donors of MSCs and normalised to EV particles released from 1 M cells. Bradford assay revealed significant higher protein concentration/1M cells in µg/mL. Results summarize the mean ± SEM for n = 3 repeat measurements for Bradford assay of each MSC-EV sample from 2 donors of MSCs. Comparisons between groups were performed with unpaired 2-tailed standard Student’s t test. *P < .05.MSC-EVsParticles/ml/1M cellsDiameter (nm)Protein conc./1 M cells (µg/mL)MSC-EVs (first donor)1.60E + 10 ± 6.05 × 108*118.2 ± 2.114.77 ± 1.39*IFN-β-primed MSC-EVs (first donor)1.24E + 10 ± 7.83 × 108127.2 ± 2.9*10.14 ± 0.90MSC-EVs (second donor)1.44E + 10 ± 8.30 × 108*116.5 ± 1.511.40 ± 0.25IFN-β-primed MSC-EVs (second donor)1.15E + 10 ± 8.22 × 108131.4 ± 3.3*10.71 ± 1.22 Characterization of EV populations from naïve and IFN-β-primed MSCs by NTA and Bradford assay. NTA analysis showed significant higher particles concentration (particles/mL/1 M cells) in EV samples collected from naïve MSCs from both donors as compared to those collected from IFN-β-primed MSCs. The size of the particles was significant higher on EV samples from IFN-β-primed MSCs. Results summarize the mean ± SEM of n = 6 repeat measurements (videos) of each EV population from 2 donors of MSCs and normalised to EV particles released from 1 M cells. Bradford assay revealed significant higher protein concentration/1M cells in µg/mL. Results summarize the mean ± SEM for n = 3 repeat measurements for Bradford assay of each MSC-EV sample from 2 donors of MSCs. Comparisons between groups were performed with unpaired 2-tailed standard Student’s t test. *P < .05. Characterization of EV populations from naïve and IFN-β-primed MSCs by TEM, immunoblotting and IFCM analysis. (A) Representative TEM images revealed the round morphology of MSC-EVs from both culture conditions. Scale bar = 100 nm. (B) Immunoblotting analysis showed the presence of CD105, CD63, and Syntenin-1 and the absence of calnexin in the MSC-EV samples. (C) IFCM analysis showed no differences in the number of CD9+ small-sized EVs between the different MSC-EV populations, while higher number of CD63+ small-sized EVs was noticed in EV samples from naïve MSCs. Data represent the number of small-sized objects/mL/1 M cells after the sum of 2 technical runs in Amnis ImageStream per sample. Hence, priming of MSCs with IFN-β does not result in any recognized phenotypic changes neither at the cellular nor the MSC-EV level. However, NTA data imply that IFN-β-primed MSCs release less EVs being larger in size than non-primed MSCs. EVs Derived From IFN-β MSCs Suppress RA-Associated Cytokines Expression and T-Cell Polyfunctionality More Effectively Than MSCs and EVs From Conventional MSC Cultures To evaluate the immune-regulatory effect of MSCs and MSC-EVs (naïve and primed with IFN-β) on RA CD+ T cells (Supplementary Fig. S4A), we first analyzed by flow cytometry the expression of inflammatory cytokines associated with the pathogenesis of RA. Coculture of MSCs with RA CD4+ T cells tended to downregulate the expression of IFN-γ, TNF-α, and IL-4 as compared to the nontreated CD4+ T cells (Fig. 2A, B). The addition of MSC-EVs to the RA CD4+ T cells tended to decrease the levels of IFN-γ and IL-22 and IL-4 (Fig. 2A, B). EV samples from IFN-β-primed MSCs significantly reduced the expression of IL-4 (P < .01), while trends toward downregulation in the expression of the cytokines IL-2, TNF-α, IFN-γ, IL-17A, and GM-CSF were observed (Fig. 2A, B). We then assessed the populations of T regulatory cells as distinguished by the expression of CD4 + CD127lowCD25 + FOXP3 + T cells (Fig. 2C, Supplementary Fig. S4B). MSCs treatment tended to decrease the frequency of T regulatory cells, while EV samples (from naïve and IFN-β-primed MSCs) did not change the T regulatory cells populations as compared to the non-treated CD4+ T cells (Fig. 2C). Characterization of CD4+ T cells after MSC and MSC-EV (from naïve or IFN-β-primed MSCs) treatment. (A) Representative flow cytometric analysis for the identification of the inflammatory cytokine IL-2 expressed in CD4+ T cells. (B) Cumulative data for the CD4+ T-cell frequency of the indicated inflammatory cytokines from RA patients showed that EVs from IFN-β-primed MSCs downregulated the expression of more RA-associated cytokines than MSCs or EVs from naïve MSCs. (C) Representative flow cytometric analysis for the identification of the frequency of T-regulatory cells (CD4+CD127lowCD25+FOXP3+) from RA patients showed higher populations of T-regulatory cells when treated with MSC-EVs than MSCs. Data are presented as the mean ± SEM of the % fold change of the expression of each inflammatory cytokine and CD4+CD127lowCD25+FOXP3+ cells as compared to the non-treated conditions. CD4+ T cells from 3 RA patients treated with 2 different donors of MSCs and MSC-EV samples (from naïve or IFN-β-primed MSCs) were used. Paired one-way ANOVA followed by Dunnett’s post-test was used for analysis of multiple comparison groups. *P < .05; **P < .01. To assess RA CD4+ T-cell polyfunctionality, supervised simplified presentation of incredibly complex evaluations (SPICE) algorithm analysis based on flow cytometric data gating was performed. SPICE is a supervised dimensionality reduction algorithm based on Boolean gating of flow cytometric data. The pie segments represent the percentage of cells that express a combination of cytokines while the pie arcs indicate the cytokines that are produced. Therefore, overlapping arcs indicate that multiple cytokines are expressed by the same T cells simultaneously. Interestingly, we showed that EV samples from IFN-β-primed MSCs reduced the coexpression of inflammatory cytokines (Fig. 3A). In addition, analysis of the frequency of RA CD4+ T cells expressing at least 3 (3+), 4 (4+), and 5 (5+) cytokines simultaneously, revealed that both MSCs and MSC-EV samples decreased the polyfunctionality of CD+ T cells of RA patients (Fig. 3B). Analysis of CD4+ T-cell polyfunctionality after MSC and MSC-EV (from naïve or IFN-β-primed MSCs) treatment. (A) SPICE analysis of flow cytometric data for the identification of polyfunctional CD4+ T cells showed that EVs from IFN-β-primed MSCs decreased more efficiently T-cell polyfunctionality than MSCs or EVs from naïve MSCs. (B) Frequency of CD4+ T cells not expressing inflammatory cytokines and frequency of CD4+ T cells expressing at least 3 (3+), 4 (4+), and 5 (5+) cytokines simultaneously after SPICE analysis. Data are presented as the mean ± SEM of the % polyfunctionality after MSC and MSC-EV treatment (from 2 MSC donors) on CD4+ T cells from 3 RA patients. One-way ANOVA followed by Dunnett’s post-test was used for analysis of multiple comparison groups. Analysis of CD4 T-cell proliferation using the CTV assay, demonstrated that MSCs (P < .05) inhibited T-cell proliferation. Moreover, a slight inhibition of T-cell proliferation using the EV samples (from naïve and IFN-β-primed MSCs) was observed (Supplementary Fig. S5). Collectively, we showed that EVs released from IFN-β MSCs suppressed more effectively the expression of inflammatory cytokines from stimulated RA CD4+ T cells, rescued the frequency of T regulatory cells, and decreased T-cell polyfunctionality than MSCs or EVs derived from naïve MSCs. EVs Derived From IFN-β MSCs Decreased the Release of More RA-Associated Cytokines From CD4+ T cells Than MSCs and EVs From Conventional MSC Cultures To further evaluate and confirm the immune-regulatory potential of MSCs and their derived EVs as was evidenced from flow cytometry analysis, we assessed the release of cytokines in the supernatants of RA CD4+ T cells, using the MSD assay. Coculture of MSCs with RA CD4+ T cells significantly inhibited the release of the cytokines IL-9 (P < .001) and TNF-α (P < .05), while a trend toward downregulation of cytokines IL-4, IL-10, and IFN-γ was observed as compared to the non-treated RA CD4+ T cells (Fig. 4). EV samples from naïve MSCs significantly downregulated the release of IL-9 (P < .05), IL-4 (P < .01), and IL-2 (P < .01), and a trend toward reduced release of IL-21, IL-10, GM-CSF, and IFN-γ was observed. EV samples from primed with IFN-β MSCs inhibited the release of IL-21 (P < .01) and IL-2 (P < .05), with a marked reduction in the expression of IL-9, IL-4, IL-10, GM-CSF, IFN-γ, and TNF-α (Fig. 4). Therefore, we demonstrated that EVs primed with IFN-β MSCs have a more inhibitory effect on key cytokines that are implicated in pathogenesis of RA compared to MSCs or EVs from naïve MSCs. MSD analysis of cytokines released by CD4+ T cells. Quantification of the release of inflammatory cytokines from CD4+ T cells after MSD analysis showed that EVs from IFN-β-primed MSCs reduced the release of more cytokines than MSCs or EVs from naïve MSCs. Data are presented as the mean ± SEM of the % fold change of the release of each cytokine as compared to the control non-treated conditions. CD4+ T cells from 3 RA patients treated with 2 different donors of MSCs and MSC-EV samples (from naïve or IFN-β-primed MSCs) were used. Paired one-way ANOVA followed by Dunnett’s post-test was used for analysis of multiple comparison groups. *P < .05; **P < .01; ***P < .001. EVs Derived From IFN-β MSCs Inhibited More Effectively RA FLS Migration To evaluate the effect of MSC-EV samples on the migration of RA FLS, a wound repair scratch assay was performed. EV samples derived from naïve MSCs showed a slight inhibitory effect on RA FLS migration (28.2%), while a significant inhibition (P < .01) of wound closure was demonstrated when EVs collected from IFN-β-primed MSCs were added to RA FLS (23.19%) after 24 hours as compared to the nontreated RA FLS (35.4%) (Fig. 5). Assessment of the effect of MSC-EV samples (from naïve or IFN-β-primed MSCs) on RA FLS. (A) Representative images of scratch assay from each experimental group (the resulting interfaces are plotted in yellow) and analysis of % migration after 24 hours of MSC-EV treatment showed that EVs from IFN-β-primed MSCs inhibited the migration of RA FLS. Data are presented as the mean ± SEM of 3 images per experimental group per each RA FLS (from 3 patients) and per EV samples (from naïve or IFN-β-primed MSCs derived from 2 donors). Paired one-way ANOVA followed by Dunnett’s post-test was used for analysis of multiple comparison groups. *P < .05; **P < .01. Scale bar = 100 µm. Discussion Here, for the first time, we assessed the immune-regulatory effect of human MSCs and EV samples (from conventional or IFN-β-primed MSC cultures) on CD4+ T cells from RA patients. We showed that EVs from IFN-β-primed MSCs more effectively decreased the expression and release of RA-associated cytokines, retain the T regulatory cell populations and decreased T-cell polyfunctionality compared to MSCs and EVs from conventional MSCs. Moreover, they inhibited RA FLS migration. MSC-EV preparations were generated using PEG precipitation followed by ultracentrifugation-based protocol, previously established as a scalable and reproducible method for the separation/enrichment of highly pure EVs.45,50,51 This method of EV collection was also utilized in the first successful treatment of an acute steroid-refractory Graft-versus-Host Disease using MSC-EVs.43 We demonstrated that the mean size of EV samples derived from IFN-β MSCs was higher in diameter than the EVs from naïve MSCs. This observation comes in accordance with a previous study when MSCs were primed with IFN-γ.40 A possible explanation of higher in diameter EVs from IFN-β MSCs than from naïve MSCs could attribute on the uptake and incorporation of IFN-β into the EVs and further investigation is needed. Moreover, decreased particle number and protein content of EV samples collected from IFN-β MSCs was noticed as was evidenced after NTA and Bradford analysis. A similar trend toward a decrease in protein content of EV samples collected from human umbilical cord MSCs primed with IFN-γ36 and in the number of particles/ MSCs ratio collected from human BM-MSCs primed with IFN-γ52 were previously reported, and it requires further investigation to validate the effect of IFN family of cytokines in the release of EVs derived from MSCs. Numerous studies have investigated the beneficial effect of EVs derived from MSCs as cell-free approaches in the amelioration of inflammatory and autoimmune diseases, as previously reviewed.27,53-55 A handful of studies have assessed the influence of MSC-EVs on activated T cells in RA with variable degrees of efficacy.56-60 The divergence of results may be attributed on the different protocols for the collection and purification of EVs, the source and the species of MSCs and stimulated T cells used. While the above studies assessed the effect of EV samples from either human or murine MSCs on murine models of RA, here we described the effect of human MSC-EV samples on RA patient-derived T cells. In addition, the different methodology and experimental read-out in assessing the immune-modulatory effect of MSC-EVs on RA cells the type of stimulation, and the dose of EV administered can generate variable results. Primarily, RA has been described as a T helper 1 (Th1)-cell-mediated disorder and is driven by a population of T cells producing inflammatory cytokines.61,62 Therefore, in the present experimental set up, we examined the effect of MSCs and their derived EVs on the expression and release of a panel of pro- and anti-inflammatory cytokines associated with the pathogenesis of RA. Importantly, we demonstrated that MSCs and their derived EV samples appeared to induce similar changes in the expression of the inflammatory cytokines, indicating that these microparticles could reproduce the anti-inflammatory effect of their cells of origin. Previous studies assessing the immune properties of MSC-EVs, have indicated that specific MSC-EV cargoes such as miRNAs (ie, miR-34a,59 miR-223,63 miRNA-181c,64 miRNA-150-5p,65 miRNA-181a,66 and microRNA-27b67) play a significant role in regulating inflammatory responses. In addition, proteomic analysis on the proteomic signature of MSC-EVs, support the importance of protein cargo in the modulation of the inflammation (ie, collagen 1 and 6, and CD73).68 Thus, we proposed that potential mechanism by which these EVs collected in our study convey their therapeutic function may be via their miRNA and/or protein cargo. However, to establish if that is so, profiling of the EVs’ content would be necessary, and the direct effects of key molecules identified singly and in combination evaluated to definitive determine main contributors to this benefit. Interestingly, priming the MSCs with IFN-β enhanced the anti-inflammatory properties of MSC-EVs. It was reported that IFN-β induced the expression of secretory leukocyte protease inhibitor (Slpi) and hepatocyte growth factor (HGF), through the phosphorylation of signal transducer and activator of transcription 1 (STAT1) and STAT3, 2 molecules that can increase the immune-modulatory properties of MSCs.41 In addition, Slpi has been reported to play a crucial role in reducing inflammatory and adaptive immune responses suppressing the secretion of inflammatory cytokines by Th1 cells and inhibiting CD4+ T-cell proliferation.69 Furthermore, HGF is recognized as a key player in the regulation of immune responses mainly by dampening down the inflammatory responses through the inhibition of nuclear factor kappa B (NF-κB) transactivating activity.70–72 In this study, we also analyzed the effect of MSCs and MSC-EV treatment on CD4+ T-cell polyfunctionality. Notably and in accordance with our previous observations of analyzing the expression of each cytokine separately, treatment with EVs from IFN-β-primed MSCs reduced more effectively the T-cell polyfunctionality as was evidenced by the % of CD4+ T cells not expressing inflammatory cytokines, and the decrease of the % CD4+ T-cell polyfunctionality, while MSCs and MSC-EVs could only reduce the % polyfunctionality. Previous studies have highlighted the importance of T-cell polyfunctionality in RA showing that they are strongly associated with disease activity and response, in addition to being resistant to Treg suppression.10,11 Thus, rather than focusing on the expression of single cytokines from T cells of RA patients, we suggest that analysis of T-cell polyfunctionality should be also considered to assess the immune-regulatory function of MSC-EVs. Previous studies suggested that MSC-EVs exerted their immune-modulatory function by regulating the populations of T regulatory cells.73,74 In accordance with previous in vitro study using murine RA CD4+ T cells,56 we demonstrated that MSCs reduced the frequency of T regulatory cells. In contrast with the aforementioned study, MSC-EVs did not significantly increase the frequency of T regulatory cells; a slight increase was noticed only with EVs samples from IFN-β-primed MSCs. This divergent response could be reflected by the different species of PBMCs used (from RA murine models vs. RA patient PBMCs), the dose of each treatment, the heterogeneity of RA patients’ samples, and the stimulation method. However, when comparing the effect of MSCs and MSC-EVs, we showed that EVs from naïve MSCs tended to increase, and EVs from IFN-β-primed MSCs significantly enhanced the frequency of T regulatory cells than MSCs. This observation might be explained due to the increased levels of transforming growth factor beta (TGF-β), presented in the MSC derived EVs,75 which has been reported to stimulate the frequency of T regulatory cells.76 While MSC-EV did not affect Treg cell polarization, addition of EVs in a Treg suppression assay or pretreatment of Treg with EV prior to their utilization for autologous effector T-cell suppression may highlight previously unappreciated function of EV in enhancing Treg suppressive capacity. In agreement with previous studies,56,73,77,78 we demonstrated that MSCs in coculture could significantly inhibit the proliferation of CD4+ T cells. Although the proliferation index was slightly lower that non-treated CD4+ T cells after MSC-EV treatment, EV samples (from naïve or primed MSCs) did not significantly inhibit T-cell proliferation suggesting that MSCs contain specific cargo not presented in their derived EVs as previously discussed.52,56,73 We also demonstrated that both MSC and MSC-EV treatment inhibit the release of proinflammatory IL-2179 and the pleiotropic IL-9. The role of IL-9 in inflammatory diseases is controversial with both anti- and proinflammatory function.80–82 Previous studies have demonstrated that IL-9 promotes inflammation in acute models,83,84 while increased levels of IL-9 was found in the serum of patients with RA,85,86 suggesting a potent role in the pathogenesis of disease. Thus, inhibition of IL-9 release may result in impaired inflammation. Finally, the tendency toward decreased release of the anti-inflammatory cytokine IL-10 may indicate that MSCs and their EVs act as suppressors of the release of most cytokines from RA CD4+ T cells and requires further investigation. Finally, we reported that EVs derived from naïve MSCs inhibited the migration of RA FLS. It has been previously reported that EV samples from BM-MSCs contained specific cargo (ie, miR-34a), which caused inhibition of RA FLS proliferation in vivo in a mouse model of RA.59 Notably, we demonstrated that EV samples from IFN-β-primed MSCs exhibited a stronger inhibition of RA FLS migration than control and EVs from naïve MSCs conditions suggesting that these EV populations as a potent therapy for RA. Limitations of the Study Despite the substantial work conducted that, in our opinion, has provided insights on the therapeutic effect of human MSCs and MSC-EVs on RA, we feel obliged to also point out certain limitations. First, we performed these experiments under standard culture conditions, without separating MSCs with T cells via membranes. We anticipate, however, that building on this work, future studies using of Boyden chambers could be beneficial for examining soluble factors released from the MSCs. Although we proved that MSCs either in conventional cultures or in the presence of IFN-β maintained their phenotype in terms of the expression of stromal markers (CD73, CD44, CD90 and CD105) and being negative for the endothelial (CD31) and hematopoietic (CD34 and CD45), we anticipate that trilineage differentiation potential of MSCs should be performed. Previous studies have shown that long-culture of MSCs resulted in loss of differentiation potential,87 which was more evident after passage 8,88 however, even with some limitations, the effect of EVs was always based on direct comparison of EVs from naïve and primed MSCs collected at the passage 6. In addition, in this study, we selected PEG precipitation followed by ultracentrifugation for the collection of EVs. We agree that different methods of EVs collection can have different effects on the EVs, and PEG precipitation followed by ultracentrifugation may not be the purification method of choice if molecular profiling of pure EV samples is intended.45 Crucially, the reason why, among a range of methods, we selected PEG precipitation followed by ultracentrifugation was because this specific method was used to collect the MSC-EVs that were then applied in what we understand to have been the first successful trial of MSC-EVs in a patient with host versus graft disease.43 Thus, to be able to advance from human clinical studies where a precedent had been set with Regulatory Authorities concerning use of MSC-EVs, and as we had to select one method for EVs collection, then the method that had been accepted previously seemed most relevant to apply in this study. Conclusion MSC-EVs are getting increased attention in inflammatory-associated diseases; potentially surpassing that of their cells of origin. Here, we provided new insights on the immune-modulatory function of MSCs and their derived EVs in RA. In addition, we demonstrated that the beneficial effect of MSC-EVs can be enhanced by priming MSCs with IFN-β. Also, our findings suggested that EV samples from IFN-β-primed MSCs more efficiently inhibit RA FLS migration, opening new therapeutic avenues for the treatment of RA. Supplementary Material Click here for additional data file. Acknowledgements We thank the Flow Cytometry Facility, Trinity Biomedical Sciences Institute, funded by Science Foundation Ireland for using IFCM and flow cytometry instrumentation. Contributor Information Dimitrios Tsiapalis, School of Pharmacy and Pharmaceutical Sciences & Trinity Biomedical Sciences Institute, Trinity College Dublin and Trinity St. James’s Cancer Institute, Dublin, Ireland. Achilleas Floudas, Molecular Rheumatology, Trinity Biomedical Sciences Institute, Trinity College Dublin, Dublin, Ireland; EULAR Centre of Excellence, Centre for Arthritis and Rheumatic Diseases, St Vincent’s University Hospital, UCD, Dublin, Ireland. Tobias Tertel, Institute for Transfusion Medicine, University Hospital Essen, University of Duisburg-Essen, Essen, Germany. Verena Boerger, Institute for Transfusion Medicine, University Hospital Essen, University of Duisburg-Essen, Essen, Germany. Bernd Giebel, Institute for Transfusion Medicine, University Hospital Essen, University of Duisburg-Essen, Essen, Germany. Douglas J Veale, EULAR Centre of Excellence, Centre for Arthritis and Rheumatic Diseases, St Vincent’s University Hospital, UCD, Dublin, Ireland. Ursula Fearon, Molecular Rheumatology, Trinity Biomedical Sciences Institute, Trinity College Dublin, Dublin, Ireland; EULAR Centre of Excellence, Centre for Arthritis and Rheumatic Diseases, St Vincent’s University Hospital, UCD, Dublin, Ireland. Lorraine O’Driscoll, School of Pharmacy and Pharmaceutical Sciences & Trinity Biomedical Sciences Institute, Trinity College Dublin and Trinity St. James’s Cancer Institute, Dublin, Ireland. Funding This work was supported by the European Union and Horizon 2020 research and innovation action under grant agreement no. 814495-EVPRO (www.evpro-implant.eu). Conflict of Interest B.G. is a founding director of Exosla Ltd. The other authors declare no potential Conflict of Interest. Data Availability The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. Author Contributions D.T.: designed the study, carried out experiments, analyzed the data and wrote the manuscript. A.F.: isolated PBMCs from RA patients, isolated FLS from RA synovial biopsies, carried out experiments, analyzed data and discussed the manuscript. T.T.: optimized and provided IFCM protocols and discussed the manuscript. V.B.: provided BM-MSCs, experimental protocols and discussed the manuscript. B.G.: provided BM-MSCs, experimental protocols and discussed the manuscript. D.J.V.: provided peripheral blood and synovial biopsies from RA patients and discussed the manuscript. U.F.: provided peripheral blood and synovial biopsies from RA patients and discussed the manuscript. L.O.: designed and supervised the study, acquired the funding, wrote, and discussed the manuscript. References Associated Data Supplementary Materials Click here for additional data file. Data Availability Statement The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.