Bone marrow mesenchymal stem cell-derived exosomes improve pyroptosis and mitochondrial integrity through miR-515-5p-mediated TLR4/NLRP3/GSDMD axis in rheumatoid arthritis Authors: Dongfeng Cai, Chao Zhong, Zixiao Yang, Jimo Li, Song Hong Published: 2025 ABSTRACT BackgroundBone marrow mesenchymal stem cell (BMSC) therapy can significantly improve the outcomes of rheumatoid arthritis (RA). This study explores the protective role of BMSC-derived exosomes (BMSCs-Exos) in RA through modulation of pyroptosis and mitochondrial integrity via the microRNA (miR)-515-5p/Toll-like receptor 4 (TLR4)/NOD-like receptor protein 3 (NLRP3)/gasdermin D (GSDMD) pathway.MethodsExosomes were isolated from rat BMSCs, with or without miR-515-5p transfection. Exosomes were identified and analyzed through transmission electron microscopy, tunable resistive pulse sensing, and protein profiling via Western blot analysis. An in vitro RA model was established by stimulating RA fibroblast-like synoviocytes (RA-FLSs) with interleukin-1β (IL-1β). Co-culture of RA-FLSs with miR-515-5p-enriched BMSCs-Exos was used to evaluate inflammation, extracellular matrix (ECM) adhesion, migration, and invasion. Dual-luciferase reporter and RNA immunoprecipitation assays were performed to validate the targeting relationship between miR-515-5p and TLR4. Pyroptosis, reactive oxygen species (ROS) generation, and mitochondrial function were assessed. In vivo effects were confirmed using the collagen-induced arthritis (CIA) rat model.ResultsIn RA-FLSs, BMSCs-Exos suppressed ECM adhesion, migration, and invasion, and attenuated IL-1β-induced inflammation through the TLR4/NLRP3/GSDMD pathway. BMSCs-Exos inhibited pyroptosis and improved mitochondrial function. Inhibition of miR-515-5p reduced cell viability, caused morphological changes, elevated cytosolic calcium (Ca²⁺), and increased mitochondrial ROS, activating caspase-dependent apoptosis and TLR4/NLRP3/GSDMD-mediated pyroptosis. In CIA rats, BMSCs-Exo treatment significantly alleviated joint damage, reduced pro-inflammatory cytokines, and protected against bone erosion.ConclusionBMSCs-Exos ameliorate RA progression by secreting miR-515-5p, which targets the TLR4/NLRP3/GSDMD pathway, thereby inhibiting pyroptosis and preserving mitochondrial homeostasis in RA-FLSs.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13075-025-03679-5. Introduction Rheumatoid arthritis (RA) is a chronic, systemic autoimmune disease characterized by persistent synovial inflammation, progressive joint destruction, and functional disability [1]. It affects approximately 0.5–1% of the global population, disproportionately impacting women and the elderly [2]. Central to RA pathogenesis is the hyperactivation of the immune system, resulting in excessive production of pro-inflammatory cytokines within the synovial microenvironment [3]. A pivotal player in RA progression is the fibroblast-like synoviocyte (RA-FLS), a specialized mesenchymal-derived cell type located within the synovial lining [4]. RA-FLSs directly cause the pathogenicity of RA, such as joint deformities [5]. These pathological behaviors contribute to the chronicity of synovitis and joint destruction, positioning RA-FLSs as critical therapeutic targets. Among the various forms of programmed cell death, pyroptosis has emerged as a key mediator of inflammation and tissue injury in RA. Pyroptosis is a lytic and pro-inflammatory form of cell death, triggered by inflammasome activation and subsequent cleavage of gasdermin D (GSDMD) [6]. Concurrently, mitochondrial dysfunction, characterized by elevated mitochondrial reactive oxygen species (mtROS), disrupted membrane potential, and adenosine triphosphate (ATP) depletion, further amplifies inflammatory signaling and cell death in RA-FLSs [7]. The preservation of mitochondrial homeostasis is therefore essential to maintaining cellular integrity and mitigating inflammation in RA. Bone marrow mesenchymal stem cells (BMSCs) have garnered attention for their regenerative and immunomodulatory properties, which not only differentiate into various lineages but also secrete a plethora of bioactive factors that influence the immune milieu [8]. The therapeutic effects of BMSCs are largely mediated through their secreted extracellular vesicles, particularly exosomes (Exos) [9]. BMSCs-derived exosomes (BMSCs-Exos) are nanosized vesicles (30–150 nm) that carry a cargo of proteins, lipids, and nucleic acids, including microRNAs (miRNAs), and are capable of modulating recipient cell functions through intercellular communication [10]. miR-515-5p has been implicated in diverse cellular processes including inflammation, mitochondrial regulation, and cell death [11], although its specific role in RA remains poorly understood. Recent findings suggest that in RA, TLR4 expression is upregulated in RA-FLSs and synovial tissues, contributing to the activation of downstream signaling cascades, including the NLRP3 inflammasome and subsequent pyroptosis [12]. Therefore, targeting the miR-515-5p/TLR4 axis represents a promising therapeutic avenue for modulating RA. We attempt to analyze the protective effects of miR-515-5p-enriched BMSCs-Exos in RA by investigating their impact on RA-FLS behavior, pyroptosis, and mitochondrial integrity via the TLR4/NLRP3/GSDMD pathway. Methods Cell culture BMSCs were harvested from the femoral and tibial bones of Sprague-Dawley (SD) rats aged 3 to 4 weeks. Rats were anesthetized with 3% pentobarbital sodium and euthanized by cervical dislocation. Femurs and tibias were harvested, and bone marrow cavities were flushed with DMEM-LG. Large tissue fragments were filtered through a 150-mesh nylon filter to obtain BMSCs. The single-cell suspension of isolated BMSCs was centrifuged at 3000 g for 10 min. Subsequently, the cells were seeded into culture dishes (1 × 106 cells/mL) and cultured in LG-DMEM medium containing 10% fetal bovine serum (FBS) and supplemented with 100 U/mL of penicillin/streptomycin. The cells were then incubated in a humidified incubator at 37 °C with 5% CO2. After 3 days, adherent cells were defined as passage 0 (P0) BMSCs. When cells reached 80–90% confluency, they were passaged. Passage 5 (P5) BMSCs were collected and analyzed for expression of surface markers CD90 (ab307737, Abcam, Cambridge, UK), CD45 (ab10558, Abcam), CD29 (ab27947 Abcam), and CD34 (ab316277, Abcam) using FACS Canto system (BD Biosciences, San Jose, CA, USA). The single-cell suspension was adjusted to a concentration of 1 × 107 cells/mL using PBS buffer and incubated at 4 °C for 30 min in the dark, with anti-IgG as the isotype control. The cells were washed twice with PBS and then subjected to machine-based detection. For differentiation assays, passage 3 (P3) BMSCs at 90% confluence and seeded in 6-well plates at 1 × 105 cells/mL and were induced with adipogenic or osteogenic differentiation media. After 14 days, adipogenic differentiation was assessed by Oil Red O staining (Sigma-Aldrich, St. Louis, MO, USA). After 21 days, osteogenic differentiation was detected by Alizarin Red staining (Sigma-Aldrich), and chondrogenic differentiation was detected by Alcian Blue staining. Isolation of BMSC-Exos Exosomes were isolated from conditioned media (CM) of BMSCs using a modified differential ultracentrifugation protocol as previously described [13]. The supernatant was collected under 300 × g centrifugation for 10 min to remove dead cells, and then 2000 × g centrifugation for 10 min to remove cell debris. The supernatant was filtered through a Steritop™ 0.22 μm sterile filter (Millipore, Billerica, MA, USA). Exosomes were pelleted by ultracentrifugation at 10,000 × g for 2 h at 4 °C, resuspended in 15 mL PBS, then further concentrated using a centrifugal filter device at 4000 × g at 4 °C to a final volume of approximately 200 µL. The isolated Exos were either used immediately or stored at −80℃ for subsequent analysis. Exo characterization BMSCs-Exo obtained through separation were identified by the following methods: Transmission electron microscopy (TEM) used a Tecnai 12 microscope. Briefly, 20 µL Exos suspension was placed on copper grids for 3 min, blotted with filter paper, stained with 30 µL phosphotungstic acid (pH 6.8), air-dried under incandescent light, and imaged. Nanoparticle tracking analysis (NTA) was performed on the ZetaView system (Particle Metrix, Germany). Data were analyzed with ZetaView software version 8.04.02. Western blot was performed to detect the protein expression levels of the surface antigens CD63 (ab217345), TSG101 (ab125011), and the negative marker Calnexin (ab22595) on MSCs-Exos. All antibodies were purchased from Abcam. Exo uptake assay To evaluate exosome uptake by RA-FLSs, the PKH26 Red Fluorescent Cell Linker Kit (Sigma-Aldrich) was used. BMSCs-Exos were resuspended in 1 mL of Diluent C, mixed with 4 µL PKH26 dye, and incubated with Exo solution at room temperature for 4 min. Excess dye was quenched by adding 2 mL 1% bovine serum albumin (BSA) and incubated for 1 min in the dark at room temperature. Labeled exosomes were pelleted by centrifugation at 10,000 × g for 70 min, washed with PBS, and resuspended in PBS. PKH26-labeled exosomes were co-cultured with RA-FLSs for 12 h. Subsequently, the cells were fixed with 4% paraformaldehyde, washed with PBS, and then stained with DAPI. The cells were observed under a laser scanning confocal microscope (Olympus, Tokyo, Japan). Acetylcholinesterase (AChE) activity Exos extracted by the multi-step ultracentrifugation method were diluted to 120 µL with PBS. In each well of a 96-well plate, 42.5 µL of this solution was added, followed by the addition of equal volumes of DTNB solution (0.1 mmol/L) and thioacetyl iodide solution (1.25 mmol/L), making a final volume of 300 µL. After 30 min of incubation, the optical density (OD) was measured at 412 nm using a microplate reader. In vitro RA-FLS model establishment and treatment Rat RA-FLSs (ATCC, Manassas, VA, USA) were cultured in high-glucose DMEM (90%) with 10% FBS, 100 µg/mL streptomycin, and 100 U/mL penicillin (Cell Biologics, Chicago, IL, USA). Cultures were maintained in a humidified incubator at 37 °C, 5% CO₂, and 95% relative humidity. To simulate ischemia-like conditions and specific factor treatments in vitro, RA-FLS density was adjusted to 80–90% confluence in 6-well plates based on the growth rate of RA-FLSs before treatments. Experimental groups were designed as follows: Control group (RA-FLSs cultured under normal conditions), IL-1β group (RA-FLSs treated with 10 ng/mL IL-1β to induce the RA-FLS model), IL-1β + BMSCs-Exo group (PKH26-labeled BMSC-Exos resuspended in complete medium, adjusted to 100 µg/mL, and added to IL-1β-treated RA-FLSs), IL-1β + BMSCs-Exo + GW4869 group (BMSCs pretreated with GW4869, Exo secretion inhibitor, and added alongside IL-1β and BMSC-Exos), and IL-1β + GW4869 group (BMSCs pretreated with GW4869 before IL-1β treatment). miR-505-5p mimics and inhibitors were purchased from GenePharma (Shanghai, China). Transfection was performed using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). Briefly, 4 µg of target plasmid and 10 µL of Lipofectamine 2000 were diluted separately in 250 µL Opti-MEM (Gibco, Waltham, MA, USA), incubated for 5 min at room temperature, mixed gently, incubated additional 20 min, and then added to cells. After 8 h, the medium was replaced with a fresh complete medium. Chemical reagents used included TAK242 (HY-11109), Caspase-1 inhibitor VX-765 (50 nM, HY-13205), and MitoTEMPO (100 mM, HY-112879), all purchased from MedChemExpress (Monmouth Junction, NJ, USA). Except for control cells, all other cells were treated with 10 ng/mL IL-1β for 24 hours. Enzyme-Linked Immunosorbent Assay (ELISA) Using ELISA kits from Beyotime (Shanghai, China), IL-1β (PI301), IL-18 (PI555), TNF-α (PT512), and IL-6 (PI522) expression was measured. Matrix metalloproteinases MMP1 (E-EL-R0617) and MMP13 (E-EL-R0045) were measured using kits from Elabscience (Wuhan, China). The supernatant was collected directly by centrifugation and then processed according to the kit instructions. The absorbance at a wavelength of 450 nm was measured using a microplate reader. A standard curve was constructed, and the concentrations of each factor were calculated. The experiment was independently repeated three times. Transwell assays Transfected cells were harvested and resuspended in serum-free medium. For migration assays, 1 × 10⁵ cells were seeded into Boyden chambers with 8.0 μm pore membranes. For invasion assays, 1 × 10⁵ cells were seeded into chambers precoated with 150 µg Matrigel (BD Biosciences). Chambers were placed into 24-well plates and incubated for 24 hours with the medium containing 10% FBS. Non-migrated cells on the upper membrane surface were removed; cells on the lower surface were fixed, stained with 0.05% crystal violet, and then counted under an Olympus microscope. Hoechst 33,342/Propidium Iodide (PI) staining Cell slides were fixed in 4% paraformaldehyde for 15 minutes at room temperature, washed three times with PBS, and then stained with Hoechst 33,342 (Sigma-Aldrich) for 5 minutes. Slides were sealed with an antifade mounting medium. PI (Sigma-Aldrich) staining was performed by incubating samples for 20 minutes, followed by fluorescence microscopy imaging and counting. TEM Cells were fixed with 2.5% glutaraldehyde in the dark for 2 minutes, scraped, and collected by centrifugation at 500 g for 5 min. Following the removal of supernatants, cell pellets were immersed in TEM fixative. After 30 minutess at room temperature, samples were stored at 4 °C. Cell pellets were embedded in 1% agarose to form blocks. Agarose-cell blocks were fixed in 0.1% osmium tetroxide for 2 hours, dehydrated through graded ethanol series, incubated three times for 10 min each in 90% acetone, and embedded in epoxy resin. Samples were examined and imaged using a Hitachi HT7800 transmission electron microscope. RT-qPCR Total RNA was extracted from tissue using TRIzol reagent (Invitrogen). cDNA was synthesized using the PrimeScript RT reagent kit (Takara, Kusatsu, Japan). For miRNA detection, reverse transcription to cDNA was performed using the miRNA First Strand cDNA Synthesis (Tailing Reaction) kit (B532451-0020, Sangon Biotech, Shanghai, China). The reaction system was 20 µL. The reaction conditions were 37 °C for 60 minutes, followed by 85 °C for 5 minutes. The SYBR® Premix Ex Taq II (Perfect Real Time) kit (DRR081, Takara, Japan) was used for sample addition. The qRT-PCR was performed on a real-time fluorescence quantitative PCR instrument (ABI 7500, Applied Biosystems, Foster City, CA, USA). Each sample was set up in triplicate wells. Ct values were recorded, and GAPDH or U6 was used as the internal reference. Relative expression was calculated by the 2-ΔΔCt method. Primer sequences are listed in Table 1. Table 1Primer sequencesGeneSequences miR-515-5p F: 5’-TTCTCCAAAAGAAAGCACTTTCTG-3’R: 5’-CTCGCTTCGGCAGCACA-3’ TLR4 F: 5’-GATCTACTCACTTTACCATA-3’R: 5’-GCTAATCGAGGCTACGACT-3’ U6 F: 5’-GCGCGTCGTGAAGCGTTC-3’R: 5’-GTGCAGGGTCCGAGG-3’ GAPDH F: 5’-CACATGGCCTCCAAGGAGTAA-3’R: 5’-GAGGGTCTCTCTCTTCCTCTTGT-3’ Table 1Primer sequencesGeneSequences miR-515-5p F: 5’-TTCTCCAAAAGAAAGCACTTTCTG-3’R: 5’-CTCGCTTCGGCAGCACA-3’ TLR4 F: 5’-GATCTACTCACTTTACCATA-3’R: 5’-GCTAATCGAGGCTACGACT-3’ U6 F: 5’-GCGCGTCGTGAAGCGTTC-3’R: 5’-GTGCAGGGTCCGAGG-3’ GAPDH F: 5’-CACATGGCCTCCAAGGAGTAA-3’R: 5’-GAGGGTCTCTCTCTTCCTCTTGT-3’ Primer sequences Western Blot (WB) Total protein was extracted, incubated on ice for 30 minutes, and centrifuged. Protein concentration was determined using a BCA assay kit. Fifty micrograms of protein were mixed with 2× SDS loading buffer and boiled at 100 °C for 5 minutes. Samples were separated by SDS-PAGE and transferred to PVDF membranes by wet transfer. Membranes were sealed in 5% skim milk at room temperature for 1 hour. The membranes were then incubated overnight at 4 °C with diluted primary antibodies against rabbit anti-Collagen I (1:1000, ab270993), anti-Fibronectin (1:500, ab268020), anti-TLR4 (1:1000, #AF7017), anti-NLRP3 (1:1000, #DF15549), anti-cleaved caspase-1 (1:200, ab238972), anti-cleaved IL-1β (1:500, ab254360), anti-GSDMD-N (1:1000, ab215203), and anti-GAPDH (1:2000, ab181602) as the loading control. All antibodies were purchased from Abcam. Membranes were incubated with HRP-conjugated goat anti-rabbit IgG HRP secondary antibody (1:2000, ab97051, Abcam). Following washes with TBST, chemiluminescence detection was performed using the ECL kit (BB-3501, Ameshame, UK). The bands were visualized using the Bio-Rad image analysis system (Bio-Rad, USA). The band intensities were analyzed using Image Pro Plus 6.0 software (Media Cybernetics, Inc.). The relative protein expression levels were represented by the ratio of the grayscale value of the target protein band to that of the GAPDH band. Membranes were imaged, and band intensities were quantified. Experiments were repeated independently three times. Immunofluorescence RA-FLSs were seeded at a density of 5 × 104 cells per dish. The cells were fixed with 4% paraformaldehyde for 30 minutes and then washed with 1 × PBS. Subsequently, the cells were permeabilized with 0.2% Triton X-100 for 15 minutes and blocked with 5% BSA. Primary antibodies anti-NLRP3 (ab263899) and anti-ASC (ab309497) were applied overnight at 4°C. After washing, cells were incubated with Alexa Fluor 488 or 594-conjugated secondary antibodies (1:200 dilution) for 30 min. Cells were washed with PBS, stained with DAPI for 15 minutes, and washed twice with PBS. Images were acquired using a confocal laser scanning microscope (Leica, Wetzlar, Germany). For phalloidin immunofluorescence, RA-FLSs were fixed, permeabilized, blocked with goat serum, and incubated with phalloidin overnight at 4 °C. Cells were incubated with IgG (1:100), counterstained with DAPI (1 µg/mL) for 5 min to detect fluorescence, and photographed. ROS content detection ROS levels were detected using the ROS fluorescent probe DCFH-DA (Sigma). Tissue or cell supernatants were incubated with 2 mL of 10 µmol/L DHE probe at 37 °C in the dark for 30 minutes. DCFH-DA is oxidized into highly fluorescent DCFH, which accumulates in the cytoplasm. The fluorescence images were captured using a fluorescence microscope (Olympus BX53, Japan). JC-1 fluorescence detection The mitochondrial membrane potential (MMP) was measured using JC-1 (Thermo Fisher Scientific, Waltham, MA, USA). Cells were seeded at 5 × 103 cells per well in 96-well plates and cultured overnight. After transfection and 24-hour incubation, cells were stained with 10 µM JC-1 solution at 37 °C in the dark for 1 hour. Fluorescence was observed under a fluorescence microscope. Intracellular calcium ion Intracellular free calcium ion concentration was measured using Fluo-3/AM (Invitrogen). Cells cultured on slides were incubated with 5 µM Fluo-3/AM at 37 °C for 30 min, washed with fresh DMEM, and scanned using a confocal laser scanning microscope (LCSSP-2, Leica). Cells were excited with a 488 nm krypton/argon laser and emission fluorescence captured at 526 nm. Images were recorded 10–20 s after fluorescence intensity stabilized, and average fluorescence intensity was used to estimate intracellular calcium changes. MitoSOX fluorescent probe mtROS production was assessed with the MitoSOX Red fluorescent probe. Cells were seeded in 12-well plates at 5 × 104 cells/well and cultured overnight. The next day, cells were stained with 5 µM MitoSOX Red probe for 30 minutes, followed by incubation at 37 °C for 1 hour in the dark. Cells were stained with MitoSOX Red or Fluo-3/AM for 30 minutes to 1 hour and analyzed by flow cytometry (Becton-Dickinson, Bedford, MA, USA). Lactate dehydrogenase (LDH) cytotoxicity LDH release was detected to assess cell death using a colorimetric LDH cytotoxicity detection kit (A020-1–2, Jiancheng Bioengineering, Nanjing, China) strictly following the manufacturer’s instructions. After incubation, supernatants were collected, and absorbance was measured at 490 nm using a microplate reader (Thermo Fisher Scientific) to evaluate cytotoxicity. ATP assay Cells were thoroughly lysed and centrifuged at 10,000 g for 5 min at 4 °C. The supernatant was stored on ice. ATP levels were measured using an ATP bioluminescence assay kit (Beyotime) on a multifunctional microplate reader SpectraMax M2e (Molecular Devices, San Jose, CA, USA). All experiments were repeated independently three times. Construction of CIA rat model SD rats (6–8-week-old; 200–205 g) were purchased from Hunan Slac Experimental Animal (Changsha, China) with Animal License No. SYXK (Xiang) 2024−0019. All rats were housed with a 12-hour light/dark cycle, with free access to food and water. Rats fasted for 12 hours and water-deprived for 4 hours prior to surgery. Procedures were approved by Animal Care and Use Committee of Affiliated Hospital of Zunyi Medical University. (Ethics approval number: KLL-2022-346) To establish the CIA model, we followed previously described methods [14]. For primary immunization, bovine type II collagen (2 mg/mL, Chondrex, Woodinville, WA, USA) and Complete Freund’s Adjuvant (Sigma-Aldrich) were mixed on ice to form an emulsion at a final concentration of 1 mg/mL. This 200 µL emulsion was injected subcutaneously into the tail. A booster immunization of 100 µL was given one week later. Two weeks after the primary immunization, the CIA rat model was successfully established. Negative control mice were not immunized but received physiological saline injections to serve as a baseline without experimental intervention. Rats were randomly divided into groups, each containing 6 animals: Control group (rats injected with normal saline only), CIA group (rats injected with CII-IFA emulsion), Exo-miR-inhibitor group (rats injected intra-articularly with BMSC-Exos transfected with 40 µg/100 µL miR-505-5p inhibitor, twice weekly until the end of the experiment), and Exo-miR-mimic group (rats injected intra-articularly with BMSC-Exos transfected with 40 µg/100 µL miR-505-5p mimic, twice weekly until the end of the experiment). The treatment lasted 5 weeks. Two months after the first injection, rats were euthanized by intraperitoneal injection of pentobarbital sodium (200 mg/kg) following tail vein blood collection. For histological analysis, ankle and knee joints were fixed in 4% paraformaldehyde solution and decalcified with 14% EDTA solution. Samples were stored at −80 °C for subsequent experiments. BMSCs-Exos were labeled with the fluorescent dye Dil (Invitrogen) according to the manufacturer’s protocol and incubated with RA-FLSs for 12 hours. Cell nuclei were stained with DAPI. Observations were made using confocal fluorescence microscopy. Assessment of joint arthritis Joint arthritis scoring was performed after the second immunization. The severity of arthritis was assessed weekly by measuring the thickness of each rat’s paw using calipers. Starting from day 21 post-CII immunization, arthritis scores were evaluated three times per week. The score ranged from 0 (no arthritis) to 4 (severe arthritis), as described previously [15]. Higher scores indicated more severe injuries in joint arthritis. Micro-CT imaging Micro-CT imaging of joints was conducted using a micro-CT scanner at 70 kV, 200 µA, and resolution of 10.2 μm. The knee joint was scanned with the lowest point of the lateral tibial growth plate as the baseline. The bone marrow cavity area 3 mm thick below the baseline was designated as the region of interest (ROI) for three-dimensional reconstruction. For ankle joint scanning, 1451 consecutive slices covering a 14.8 mm thick area of the calcaneus, namely the tarsal region, were set as the ROI for three-dimensional reconstruction. The three-dimensional images were generated by NRecon software. Bone mineral density (BMD), bone surface-to-volume ratio (BS/BV), trabecular separation (Tb.Sp), and trabecular thickness (Tb.Th) were evaluated. Histopathological examination Ankle joint tissues were fixed in 10% formalin for 24 hours, decalcified in formic acid for 30 days, embedded in paraffin, and sectioned. HE staining was used to score inflammation: 0 = no inflammation; 1 = mild synovial inflammation; 2 = mild synovitis; 3 = moderate synovitis; 4 = extensive inflammatory cell infiltration in synovium. Safranin O-fast green staining was performed to assess cartilage damage with the following scores: 0 = no destruction; 1 = mild reduction; 2 = moderate loss; 3 = severe loss; 4 = no staining. Quantitative scoring was performed on the hind paws of rats to calculate total scores for each group and analyze inter-group differences. Immunohistochemistry Sections were boiled in 0.01 M citrate buffer for 15 min, cooled to room temperature, and washed with PBS. Sections were blocked with goat serum for 20 min at room temperature, and excess liquid was removed. Primary antibody Ly-6G (ab25377, 1:500, Abcam,) was applied for immunostaining. Sections were incubated for 1 h at room temperature and then washed with PBS. Secondary antibody goat anti-rabbit IgG HRP (ab97051, 1:2000, Abcam) was applied for 1 hour at room temperature. After washing, SP staining was performed at 37 °C for 30 minutes, followed by PBS wash. After 3,3’-Diaminobenzidine (DAB) staining for 5–10 minutes, the reaction was terminated by rinsing with tap water for 10 min. Sections were counterstained with hematoxylin for 2 minutes, differentiated with hydrochloric acid alcohol, washed with tap water for 10 minutes, dehydrated, cleared, and mounted with neutral gum. Statistical analysis All raw data were statistically analyzed using GraphPad Prism 8.0 software (GraphPad Prism Software, CA, USA). Data are presented as mean ± standard deviation (SD). Two-group comparisons were analyzed by Student’s t-test. Multiple-group comparisons were conducted using one-way analysis of variance (ANOVA), followed by Tukey’s multiple comparisons test. A p-value < 0.05 was considered statistically significant. Results Isolation and characterization of BMSCs-Exos BMSCs were isolated from SD rats and characterized to investigate the effects of BMSC-Exos on RA. Under inverted microscopy, BMSCs displayed adherent growth with spindle-shaped or polygonal morphology and abundant, evenly distributed cytoplasm (Fig. 1A). Flow cytometry analysis showed that the percentage of CD29-positive cells was 97.16 ± 5.98%, while the percentage of CD90-positive cells was 95.69 ± 7.32%. In contrast, the percentage of CD34-positive cells was only 0.75 ± 0.11%, and the percentage of CD45-positive cells was 1.12 ± 0.23%. These specific quantitative data clearly and intuitively demonstrated that CD29 and CD90 were highly expressed, while CD34 and CD45 were extremely lowly expressed (Fig. 1B). Tri-lineage differentiation assays confirmed the osteogenic, adipogenic, and chondrogenic differentiation potentials of the isolated cells (Fig. 1C-E), confirming they were BMSCs. Fig. 1Isolation and identification of BMSCs. A Bright-field microscopy observation of BMSCs morphology after isolation from tibial tissue of SD rats (200×); B Flow cytometry analysis detected expression of surface antigens CD29, CD90, CD45, and CD34 in BMSCs, and quantitative analysis; C–E Bright-field microscopy observation of BMSCs after 14 and 21 days of culture, with Alizarin Red S, Alcian Blue, and Oil Red O staining to evaluate osteogenic, chondrogenic, and adipogenic differentiation, respectively (200×) Isolation and identification of BMSCs. A Bright-field microscopy observation of BMSCs morphology after isolation from tibial tissue of SD rats (200×); B Flow cytometry analysis detected expression of surface antigens CD29, CD90, CD45, and CD34 in BMSCs, and quantitative analysis; C–E Bright-field microscopy observation of BMSCs after 14 and 21 days of culture, with Alizarin Red S, Alcian Blue, and Oil Red O staining to evaluate osteogenic, chondrogenic, and adipogenic differentiation, respectively (200×) Given previous reports indicating a protective role of BMSCs-Exos in RA [16, 17], Exos were isolated from BMSC culture supernatants. TEM revealed membrane-bound vesicles approximately 100 nm in diameter, with round or oval morphology and low-density cores (Supplementary Fig. 1 A). NTA showed a size distribution mainly between 70 and 150 nm (Supplementary Fig. 1B). WB confirmed the presence of exosomal surface markers CD63 and TSG101, and the absence of the negative marker Calnexin (Supplementary Fig. 1 C), verifying the successful isolation of BMSC exosomes. BMSCs-Exos suppress adhesion, migration, and invasion of RA-FLSs To explore the role of BMSCs-Exos in RA, FLSs were stimulated with IL-1β (10 ng/mL) (Supplementary Fig. 5) to establish an in vitro RA model. Exos were labeled with PKH26 and co-incubated with RA-FLSs for 12 hours; the presence of red fluorescence within RA-FLSs confirmed exosome internalization (Fig. 2A). To investigate the functional effects of BMSCs-Exos on RA-FLSs, IL-1β-induced RA-FLSs were treated with BMSCs-Exos and the Exos inhibitor GW4869. The specific groups included Control group, IL-1β group, IL-1β + GW4869 group, IL-1β + BMSCs-Exo group, and IL-1β + BMSCs-Exo + GW4869 group. AchE activity assay showed a significant reduction of exosome release after GW4869 treatment (Fig. 2B). Transwell assays demonstrated that BMSCs-Exos significantly inhibited IL-1β-induced RA-FLS migration and invasion, which were reversed by GW4869. At the same time, no significant difference was observed in migration and invasion capabilities between the IL-1β + GW4869 and IL-1β groups (Fig. 2C-D). Fig. 2Inhibition of RA-FLSs adhesion, migration, and invasion by BMSCs-Exos. A Immunofluorescence analysis of RA-FLSs internalization of BMSCs-Exo; B Ach E activity assay to quantify BMSCs-Exo release; C–D Transwell assays to evaluate migration and invasion of RA-FLSs; E–F WB analysis showing representative images and quantification of ECM proteins collagen I and fibronectin; G Immunofluorescence staining of actin cytoskeleton in RA-FLSs. All data are expressed as mean ± SD, N = 3. One-way ANOVA followed by Tukey’s post hoc test was used for statistical analysis. *p < 0.05 vs. Control group, #p < 0.05 vs. IL-1β group, &p < 0.05 vs. IL-1β + BMSCs-Exo group Inhibition of RA-FLSs adhesion, migration, and invasion by BMSCs-Exos. A Immunofluorescence analysis of RA-FLSs internalization of BMSCs-Exo; B Ach E activity assay to quantify BMSCs-Exo release; C–D Transwell assays to evaluate migration and invasion of RA-FLSs; E–F WB analysis showing representative images and quantification of ECM proteins collagen I and fibronectin; G Immunofluorescence staining of actin cytoskeleton in RA-FLSs. All data are expressed as mean ± SD, N = 3. One-way ANOVA followed by Tukey’s post hoc test was used for statistical analysis. *p < 0.05 vs. Control group, #p < 0.05 vs. IL-1β group, &p < 0.05 vs. IL-1β + BMSCs-Exo group WB analysis revealed decreased levels of extracellular matrix (ECM) proteins collagen I and fibronectin following BMSCs-Exo treatment (Fig. 2E-F). Since the F-actin cytoskeleton regulates RA-FLS migration and invasion, immunofluorescence staining using rhodamine-labeled phalloidin was performed. BMSCS-Exo treatment increased the number and thickness of stress fibers while reducing filopodia and lamellipodia formation at the cell periphery (Fig. 2G). These results suggest that BMSCs-Exos promote cytoskeletal remodeling in RA-FLSs, thereby inhibiting their adhesion, migration, and invasion capabilities. miR-515-5p secreted from BMSCs-Exos inhibits IL-1β-induced inflammation in RA-FLSs Considering the key role of inflammation in RA, the effect of BMSCs-Exos on inflammatory responses of RA-FLSs was examined. ELISA results showed that IL-1β stimulation significantly upregulated pro-inflammatory cytokines TNF-α, IL-6, and matrix metalloproteinases MMP1 and MMP13, while BMSCs-Exo treatment markedly suppressed these increases. Furthermore, the addition of GW4869 to RA-FLS not co-cultured with BMSCs-Exo did not affect the secretion of TNF-α, IL-6, MMP1, and MMP13 (Fig. 3A-D). TLR4 was predicted by TargetScan to be a potential target of miR-515-5p. However, whether miR-515-5p is carried by BMSC-Exo and its effects on IL-1β-treated RA-FLS inflammation were unclear. RT-qPCR analysis showed miR-515-5p was highly enriched in BMSCs-Exos compared to BMSC culture supernatant and was significantly reduced in IL-1β-treated RA-FLSs (Fig. 3E-F). miR-515-5p directly targets TLR4 (Supplementary Fig. 2A-C). To verify that BMSCs-Exos exert effects via miR-515-5p, BMSCs-Exos were transfected with miR-515-5p inhibitors or mimics and co-cultured with RA-FLSs. Additionally, RA-FLSs were treated with miR-515-5p mimic after IL-1β induction (Supplementary Fig. 2D-F). Functional assays showed that inhibition of miR-515-5p restored RA-FLS migration and invasion, increased ECM protein expression, and exacerbated inflammatory dysregulation. Exosomes loaded with miR-515-5p mimics exhibited significantly greater efficacy than conventional exosomes in regulating ECM adhesion, migration, and invasion. Furthermore, while direct administration of miR-515-5p mimics or inhibitors also produced some effects, these were less pronounced than those achieved by exosomes loaded with miR-515-5p mimics or inhibitors (Fig. 4A-E). Fig. 3BMSCs-Exos inhibit IL-1β-induced inflammation in RA-FLSs. A–D ELISA analysis of pro-inflammatory cytokines and matrix metalloproteinases (TNF-α, IL-6, MMP1, MMP13) in cell supernatants of each group; E Expression analysis of miR-515-5p in BMSCs-Exos and BMSCs supernatant; F Expression analysis of miR-515-5p in IL-1β-treated RA-FLSs. All data are expressed as mean ± SD, N = 3. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. Control or BMSCs group, #p < 0.05 vs. IL-1β group, &p < 0.05 vs. IL-1β + BMSCs-Exo group BMSCs-Exos inhibit IL-1β-induced inflammation in RA-FLSs. A–D ELISA analysis of pro-inflammatory cytokines and matrix metalloproteinases (TNF-α, IL-6, MMP1, MMP13) in cell supernatants of each group; E Expression analysis of miR-515-5p in BMSCs-Exos and BMSCs supernatant; F Expression analysis of miR-515-5p in IL-1β-treated RA-FLSs. All data are expressed as mean ± SD, N = 3. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. Control or BMSCs group, #p < 0.05 vs. IL-1β group, &p < 0.05 vs. IL-1β + BMSCs-Exo group Fig. 4miR-515-5p inhibition restores RA-FLSs adhesion, migration, and invasion. Co-culture BMSCs-Exo with RA-FLS after transfection with miR-515-5p inhibitor or mimic, or without transfection. A–B Transwell assays to assess migration and invasion of RA-FLSs; C–D WB analysis of representative images and quantification of ECM proteins (collagen I and fibronectin); E ELISA detection of inflammation-related cytokines in cell supernatant. All data are expressed as mean ± SD, N = 3. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. mimic NC group, #p < 0.05 vs. inhibitor NC group miR-515-5p inhibition restores RA-FLSs adhesion, migration, and invasion. Co-culture BMSCs-Exo with RA-FLS after transfection with miR-515-5p inhibitor or mimic, or without transfection. A–B Transwell assays to assess migration and invasion of RA-FLSs; C–D WB analysis of representative images and quantification of ECM proteins (collagen I and fibronectin); E ELISA detection of inflammation-related cytokines in cell supernatant. All data are expressed as mean ± SD, N = 3. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. mimic NC group, #p < 0.05 vs. inhibitor NC group BMSCs-Exos secrete miR-515-5p/TLR4 to repress pyroptosis in RA-FLSs To better understand the effect of miR-515-5p secreted by BMSCs-Exo on pyroptosis of RA-FLSs, Hoechst 33,342/PI staining was performed. After IL-1β induction, fluorescence intensity significantly increased, indicating increased pyroptotic cells. After adding BMSCs-Exos, fluorescence intensity decreased (Fig. 5A). Furthermore, Annexin V-APC and 7-AAD double staining was performed to measure pyroptotic cells. Flow cytometry analysis showed that the percentage of Annexin V+/7-AAD + cells increased in the IL-1β group but significantly reduced by BMSCs-Exo treatment (Fig. 5B-C). In addition, LDH activity, a marker of cell death, was analyzed to assess pyroptosis. Results showed LDH activity was significantly increased in the IL-1β-induced group but was inhibited by BMSCs-Exos (Fig. 5D). TEM observation revealed typical pyroptotic features in IL-1β-treated RA-FLS: cell swelling with large bubbles, decreased cytoplasmic density, and pore formation on the cell membrane, which were all improved after BMSCs-Exo treatment (Fig. 5E). These results indicate that BMSCs-Exos reduce IL-1β-induced RA-FLS pyroptosis. Fig. 5BMSCs-Exos reduce RA-FLS pyroptosis. A Hoechst 33,342/PI staining to detect apoptotic cells in each group; B–C Flow cytometry analysis of pyroptosis ratio using Annexin V + 7AAD double staining; D LDH release detection; E TEM analysis of cellular ultrastructure. All data are expressed as mean ± standard deviation, N = 3. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. Control group, #p < 0.05 vs. IL-1β group, &p < 0.05 vs. IL-1β + BMSCs-Exo group BMSCs-Exos reduce RA-FLS pyroptosis. A Hoechst 33,342/PI staining to detect apoptotic cells in each group; B–C Flow cytometry analysis of pyroptosis ratio using Annexin V + 7AAD double staining; D LDH release detection; E TEM analysis of cellular ultrastructure. All data are expressed as mean ± standard deviation, N = 3. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. Control group, #p < 0.05 vs. IL-1β group, &p < 0.05 vs. IL-1β + BMSCs-Exo group WB analysis showed IL-1β treatment upregulated NLRP3, cleaved-caspase-1, cleaved-IL-1β, and GSDMD-N in RA-FLSs, while BMSCs-Exos significantly reduced the expression of these proteins (Fig. 6A-B). Immunofluorescence staining showed co-localization and enhanced fluorescence of NLRP3 and ASC in IL-1β-treated RA-FLSs, which was significantly inhibited by BMSCs-Exo treatment (Fig. 6C). Finally, ELISA results also demonstrated that BMSCs-Exos suppressed IL-1β and IL-18 in RA-FLSs (Fig. 6D). Fig. 6BMSCs-Exos attenuate IL-1β-induced NLRP3-mediated pyroptosis in RA-FLSs. A–B WB analysis of representative bands for NLRP3 inflammasome marker proteins; C Dual immunofluorescence staining for NLRP3 and ASC colocalization in RA-FLSs after IL-1β exposure (Red: NLRP3, Green: ASC, scale bar = 25 μm); D ELISA analysis of IL-1β and IL-18 levels in cell supernatant. All data are expressed as mean ± standard deviation, N = 3. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. Control group, #p < 0.05 vs. IL-1β group, &p < 0.05 vs. IL-1β + BMSCs-Exo group BMSCs-Exos attenuate IL-1β-induced NLRP3-mediated pyroptosis in RA-FLSs. A–B WB analysis of representative bands for NLRP3 inflammasome marker proteins; C Dual immunofluorescence staining for NLRP3 and ASC colocalization in RA-FLSs after IL-1β exposure (Red: NLRP3, Green: ASC, scale bar = 25 μm); D ELISA analysis of IL-1β and IL-18 levels in cell supernatant. All data are expressed as mean ± standard deviation, N = 3. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. Control group, #p < 0.05 vs. IL-1β group, &p < 0.05 vs. IL-1β + BMSCs-Exo group Further, miR-515-5p inhibitor was transfected into BMSCs-Exos and co-cultured with RA-FLSs. Flow cytometry and ELISA results confirmed that miR-515-5p inhibitor promoted IL-1β-induced Annexin V+/7-AAD + positive cells percentage and LDH release (Supplementary Fig. 3A-C). Additionally, RA-FLSs were pretreated with caspase-1 inhibitor VX-765 or TLR4-specific inhibitor TAK242. The results showed that both VX-765 and TAK242 reversed cell death, reduced LDH release, and decreased the percentage of Annexin V+/7-AAD + cells (Supplementary Fig. 3A-C). Moreover, VX-765 and TAK242 inhibited IL-1β-induced expression of NLRP3, cleaved-caspase-1, cleaved-IL-1β, and GSDMD-N in RA-FLSs treated with miR-515-5p inhibitor (Supplementary Fig. 3D-E). In summary, these data indicate that BMSCs-Exos inhibit TLR4 activation by secreting miR-515-5p, thereby suppressing NLRP3 inflammasome-induced pyroptosis in RA-FLS. BMSCs-Exos improve mitochondrial function of RA-FLSs ROS fluorescence intensity significantly increased in IL-1β-induced RA-FLS, while BMSCs-Exos decreased ROS fluorescence intensity after IL-1β treatment (Fig. 7A-B). mtROS fluorescence intensity significantly increased after IL-1β treatment but reduced in the BMSCs-Exo group (Fig. 7C-D). Furthermore, IL-1β treatment decreased mitochondrial membrane potential in RA-FLS, which was increased after BMSCs-Exo treatment (Fig. 7E-F). Fig. 7BMSCs-Exos inhibit mitochondrial dysfunction in RA-FLSs. A–B DCFDH fluorescence probe detection of ROS levels in cells (scale bar = 25 μm); C–D MitoSOX fluorescence (red) to assess mtROS levels in RA-FLSs after 24 h incubation, with representative images and fluorescence quantification (scale bar = 25 μm); E–F JC-1 kit detection of mitochondrial membrane potential in RA-FLSs (J-aggregates, red; J-monomers, green; scale bar = 25 μm); G–H Fluo-3/AM fluorescence probe to detect calcium levels in IL-1β-treated RA-FLSs; I ATP release detection. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. Control group, #p < 0.05 vs. IL-1β group, &p < 0.05 vs. IL-1β + BMSCs-Exo group. Experiments were repeated independently three times BMSCs-Exos inhibit mitochondrial dysfunction in RA-FLSs. A–B DCFDH fluorescence probe detection of ROS levels in cells (scale bar = 25 μm); C–D MitoSOX fluorescence (red) to assess mtROS levels in RA-FLSs after 24 h incubation, with representative images and fluorescence quantification (scale bar = 25 μm); E–F JC-1 kit detection of mitochondrial membrane potential in RA-FLSs (J-aggregates, red; J-monomers, green; scale bar = 25 μm); G–H Fluo-3/AM fluorescence probe to detect calcium levels in IL-1β-treated RA-FLSs; I ATP release detection. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. Control group, #p < 0.05 vs. IL-1β group, &p < 0.05 vs. IL-1β + BMSCs-Exo group. Experiments were repeated independently three times BMSCs-Exos inhibited IL-1β-induced calcium accumulation and restored ATP levels (Fig. 7G-I). In RA-FLSs co-transfected with BMSCs-Exos and miR-515-5p inhibitor, these effects were reversed by miR-515-5p inhibitor but rescued by treatment with mitochondrial antioxidant MitoTEMPO or TLR4 antagonist TAK242 (Supplementary Fig. 4). BMSCs-Exos alleviate bone destruction in CIA mice by regulating NLRP3 inflammasome via the miR-515-5p/TLR4 axis Finally, we investigated the in vivo effect of Exo-miR-515-5p and its signaling axis in a CIA mouse model. Greater cartilage damage and destruction were found in the CIA and Exo-miR-515-5p inhibitor groups compared to controls, whereas Exo-miR-515-5p mimic treatment alleviated the damage (Fig. 8A-C). Consistent with these findings, joint swelling was evaluated every other day from day 21. Arthritis scores of CIA paws increased progressively, with decreased cartilage thickness and increased synovial inflammation scores, which were ameliorated by Exo-miR-515-5p mimic treatment (Fig. 9A-C). These results suggest Exo-miR-515-5p effectively inhibits RA progression. Micro-CT examination showed that BMSCs-Exo treatment effectively reduced cartilage and bone destruction in CIA mice. Quantitative analysis revealed RA decreased BV/TV, BS/TV, Tb.N, and increased Tb.Sp (Fig. 9D-E). Fig. 8Exo-miR-515-5p alleviates joint destruction in CIA mice. A Representative photos of mouse ankle joints in each group; B–C HE staining and Safranin O staining of joint tissues in mice with different treatments Exo-miR-515-5p alleviates joint destruction in CIA mice. A Representative photos of mouse ankle joints in each group; B–C HE staining and Safranin O staining of joint tissues in mice with different treatments Fig. 9Exo-miR-515-5p attenuates arthritis severity in CIA mice. A Arthritis score of inflamed joints; B Paw thickness of inflamed joints; C Synovitis score analysis; D Micro-CT images of ankle joints in CIA mice; E Quantitative analysis of BV/TV, BS/TV, Tb.Sp, and Tb.N. N = 6. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. Control group, #p < 0.05 vs. CIA group Exo-miR-515-5p attenuates arthritis severity in CIA mice. A Arthritis score of inflamed joints; B Paw thickness of inflamed joints; C Synovitis score analysis; D Micro-CT images of ankle joints in CIA mice; E Quantitative analysis of BV/TV, BS/TV, Tb.Sp, and Tb.N. N = 6. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. Control group, #p < 0.05 vs. CIA group ELISA analysis showed significantly elevated plasma IL-18 and IL-1β levels in the CIA group, which were significantly reduced after Exo-miR-515-5p mimic treatment (Fig. 10A-B). Immunohistochemical staining showed increased expression of TLR4 and pyroptosis-related proteins NLRP3 and GSDMD in the synovial tissue of CIA mice compared to controls. Exo-miR-515-5p mimic treatment reduced expression of TLR4, NLRP3, and GSDMD in synovial tissue (Fig. 10C). Overall, these results demonstrate that BMSCs-Exos alleviate bone destruction in CIA mice by regulating the miR-515-5p/TLR4/NLRP3 inflammasome pathway. Fig. 10Exo-miR-515-5p reduces inflammatory responses in CIA mice. A–B ELISA detection of IL-1β and IL-18 levels in serum of each group; C Immunohistochemical analysis of TLR4, NLRP3, and GSDMD expression in synovial tissue of each group. N = 6. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. Control group, #p < 0.05 vs. CIA group Exo-miR-515-5p reduces inflammatory responses in CIA mice. A–B ELISA detection of IL-1β and IL-18 levels in serum of each group; C Immunohistochemical analysis of TLR4, NLRP3, and GSDMD expression in synovial tissue of each group. N = 6. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs. Control group, #p < 0.05 vs. CIA group Discussion In RA synovium, aberrant activation of inflammasomes and elevated expression of pyroptosis-related proteins have been observed in RA-FLSs, implicating this pathway in the perpetuation of joint inflammation and destruction [18]. Concurrently, the functional state of mitochondria is emerging as a pivotal factor in RA pathogenesis [19]. This interplay between mitochondrial integrity and pyroptosis creates a vicious cycle that exacerbates synovial inflammation and joint damage. Our study extends these findings by focusing on the role of miR-515-5p-enriched BMSCs-Exos in mitochondrial dysfunction and pyroptosis in RA-FLSs via the TLR4/NLRP3/GSDMD axis. Our findings demonstrated that BMSCs-Exos suppressed adhesion, migration, and invasion of RA-FLSs, likely through cytoskeletal remodeling and regulation of ECM protein expression. MMPs are produced by multiple cells and tissues to be involved in ECM degradation and remodeling, which is involved in several facets of RA pathogenesis [20]. BMSCs-Exos decreased the expression of collagen I and fibronectin in RA-FLSs, suggesting reduced ECM adhesion capacity. Cytoskeletal dynamics, particularly the organization of F-actin, are essential for RA-FLS motility and ECM interactions [21]. BMSCs-Exo treatment increased intracellular stress fibers while reducing lamellipodia and filopodia formation, indicating impaired cell spreading and motility. This suggests that BMSCs-Exos induce actin cytoskeletal remodeling, contributing to suppressed invasion and adhesion behaviors. The dysregulation of inflammatory responses is often accompanied by alterations in the stress response within the body [22]. High levels of ROS within cells can induce the activation of the NLRP3 inflammasome [23, 24]. ROS are also closely associated with mitochondrial dysfunction, and excessive mtROS can trigger pyroptosis, thereby regulating Caspase-1/GSDMD-dependent pyroptosis [25, 26]. BMSCs-Exos reduced the ROS fluorescence intensity in cells treated with IL-1β, decreased the production of mtROS, and increased mitochondrial membrane potential. Mitochondrial dysfunction causes cellular energy depletion occurs and alters mitochondrial membrane permeability, and Ca²⁺ influxes into the cytoplasm and mitochondria, inducing pyroptosis [27]. BMSCs-Exos inhibited Ca²⁺ accumulation increased by IL-1β treatment and restored ATP levels. Consistently, Exos from MSCs can rapidly increase ATP and NADH levels within 1 h of reperfusion and significantly reduce oxidative stress [28]. Moreover, MSCs-Exos suppress the production of ROS [29] and inhibit pyroptosis through miRNAs [30]. These findings support our results that BMSCs-Exos relieve IL-1β-induced mitochondrial dysfunction and pyroptosis in RA-FLSs. For the mechanistic study, we found that miR-515-5p in BMSCs-Exos was increased compared with that in BMSCs supernatant. In RA-FLS treated with IL-1β, the expression of miR-515-5p was significantly decreased. Exosomes can carry miRNAs to mediate inflammation, immune regulation, and cell migration in RA [31]. miR-515-5p was reduced in OA cartilage tissues and IL-1β-treated CHON-001 cells [32]. The miR-515-5p inhibitor restored the migration and invasion of RA-FLSs, increased the expression of ECM proteins, exacerbated the dysregulation of the inflammatory response, and promoted the IL-1β-induced proportion of Annexin V+/7-AAD + positive cells and LDH release. It also increased the expression of NLRP3, cleaved-caspase1, cleaved-IL-1β, and GSDMD-N. Similar to our findings, miR-515-5p restoration alleviated IL-1β-induced chondrocyte apoptosis, inflammatory responses, and ECM degradation [33]. Taken together, BMSCs-Exo suppressed inflammation, ECM remodeling, and pyroptosis in RA-FLSs through miR-515-5p targeting TLR4, reducing downstream NLRP3 activation and inflammatory cytokine release. Bioinformatics prediction revealed a potential target relationship between TLR4 and miR-515-5p. The dual-luciferase reporter gene assay and RIP experiments also confirmed that miR-515-5p targeted TLR4. TLR4 was significantly upregulated in the synovial tissues of RA. TLR4 is notably increased in synovial fibroblasts, macrophages, and dendritic cells within RA joints, contributing to the production of pro-inflammatory cytokines and the progression of the disease [34]. As our results showed, pre-treatment with the caspase-1 inhibitor VX-765 or the TLR4-specific inhibitor TAK-242 effectively reversed cell death, reduced LDH release, decreased percentages of Annexin V+/7-AAD + cells, and suppressed the expression of NLRP3, cleaved caspase-1, cleaved IL-1β, and GSDMD-N. Previous evidence reported that TAK-242, a specific TLR4 inhibitor, improved mitochondrial disorders [35, 36]. Samarpita et al. reported that in the adjuvant-induced arthritis rat model, TAK-242 downregulated IL-6 levels and significantly alleviated inflammatory symptoms in the joint tissues [37]. Resatorvid, a specific TLR4 antagonist, mitigated synovial inflammation, cartilage degeneration, and bone destruction in mice with temporomandibular joint osteoarthritis by preventing ROS production and NLRP3-mediated chondrocyte pyroptosis [38]. This highlights a regulatory mechanism where BMSCs-Exos modulate pyroptotic pathways via miR-515-5p/TLR4/NLRP3. A study demonstrated that miR-515-5p directly targets TLR4, leading to the inhibition of the MyD88/NF-κB signaling pathway [39]. Moreover, miR-515-5p modulates FLSs in RA by enhancing cell proliferation and reducing apoptosis through the TLR4/JNK pathway [40]. Additionally, Exosomes derived from BMSCs have been shown to carry miR-145-5p and miR-216a-5p, which can suppress the TLR4/NF-κB signaling pathway [41]. These studies collectively support the concept that miR-515-5p and other exosomal miRNAs from BMSCs can modulate key inflammatory pathways involving TLR4 and NLRP3, offering potential therapeutic avenues for RA and related inflammatory diseases. In vivo, BMSCs-Exos enriched with miR-515-5p alleviated joint inflammation, cartilage erosion, and bone destruction in CIA mice, as evidenced by improved histology, reduced arthritis scores, and Micro-CT parameters (increased BV/TV, Tb.N; reduced Tb.Sp). Plasma levels of IL-18 and IL-1β decreased, while immunohistochemistry showed reduced expression of TLR4, NLRP3, and GSDMD. Previous study demonstrates that BMSCs-Exos carrying miR-378a-5p attenuated synovial histopathological alterations in CIA mice via the IRF1/STAT1 axis, and treatment with these exosomes resulted in decreased arthritis scores, reduced paw thickness, and alleviation of symptoms [42]. This study focused on validating the mechanism by which BMSCs-Exos improve RA by regulating the TLR4/NLRP3/GSDMD pathway via miR-515-5p. Due to limitations in the research timeline and resources, the following key questions remain unexplored. First, optimization potential for administration methods and dosages. This study employed intra-articular injection (40 µg/100 µL, twice weekly). While this approach allowed direct targeting of affected joints and reduced systemic exposure, it required invasive procedures. Furthermore, the dosage was selected based on efficacy validation from preliminary pre-experiments without conducting dose-gradient studies (e.g., 20 µg, 40 µg, 80 µg) to clearly establish the “efficacy-dose relationship. Furthermore, it did not compare efficacy differences with more clinically feasible administration routes such as intravenous or subcutaneous injection. Second, missing pharmacokinetic characteristics. Key parameters such as the distribution, clearance half-life, and targeting efficiency of BMSCs-Exos within the body have yet to be determined. These include whether Exos diffuse to other tissues after intra-articular injection, their retention time in synovial tissue, and the stability of miR-515-5p within Exos. Such data are crucial for designing subsequent clinical dosing regimens. Third, inadequate assessment of long-term safety and immune responses. This study only evaluated rats’ general condition (e.g., body weight, appetite) and major organ function (liver, kidney) during the 5-week treatment period. without conducting long-term follow-up (e.g., at 12 weeks, 24 weeks) to assess potential effects of Exos on immune homeostasis (e.g., whether they induce anti-Exos antibody production or impair normal immune cell function). Additionally, the biosafety of exogenous Exos (e.g., endotoxin residue, cytotoxicity) was not verified. To address the aforementioned limitations, we have outlined clear directions for subsequent research to enhance the foundation for clinical translation. First, subsequent studies will compare the efficacy differences between intra-articular and intravenous injections using the CIA model. By combining micro-CT imaging with in vivo tracing of fluorescently labeled Exos, we will clarify the targeting efficiency of different administration routes. In addition, we will observe various dose gradients and determine the optimal efficacy-minimum dose combination through joint inflammation scoring and quantitative analysis of bone erosion, thereby reducing potential toxic side effects. Second, we will assess the plasma concentration-time curve, tissue distribution ratios (with a focus on synovium, liver, and kidney), and elimination half-life in rats. Meanwhile, we will evaluate the stability of miRNA delivery mediated by Exos by detecting miR-515-5p expression in synovial tissue at different time points via qPCR. Third, we will conduct 24-week long-term intervention to monitor rat body weight, routine blood tests, liver and kidney function, and immune markers (such as the proportion of peripheral blood Treg cells and anti-Exos antibody titers). Pathological sections will be examined to observe structural changes in major organs (heart, liver, kidneys, lungs), ruling out potential toxicity from long-term administration and providing safety data for clinical research. In summary, this study provides new evidence that BMSCs-Exos exert therapeutic effects on RA by delivering miR-515-5p, which directly targets TLR4 to inhibit NLRP3 inflammasome-mediated inflammation and pyroptosis, preserve mitochondrial function, and attenuate cartilage and bone destruction in vivo. Supplementary Information Supplementary Material 1: Supplementary Figure 1. Identification of BMSCs-Exos. (A) TEM observation of the ultrastructure of exosomes (×5000); (B) Nanosight particle size analysis for exosome diameter detection; (C) WB detection of exosome surface markers (positive markers: CD63, TSG101; negative marker: Calnexin). Supplementary Material 1: Supplementary Figure 1. Identification of BMSCs-Exos. (A) TEM observation of the ultrastructure of exosomes (×5000); (B) Nanosight particle size analysis for exosome diameter detection; (C) WB detection of exosome surface markers (positive markers: CD63, TSG101; negative marker: Calnexin). Supplementary Material 2: Supplementary Figure 2. miR-515-5p targets TLR4. (A) Schematic diagram of the binding site between miR-515-5p and TLR4; (B) Dual-luciferase reporter assay to detect the targeting regulation of TLR4 by miR-515-5p; (C) RIP assay to verify the binding between miR-515-5p and TLR4; (D) RT-qPCR detection of miR-515-5p and TLR4 expression in cells; (E–F) WB evaluation of representative bands and quantitative analysis of TLR4 expression in cells. All data are expressed as mean ± standard deviation, N = 3. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs mimic NC group,#p < 0.05 vs inhibitor NC group. Supplementary Material 2: Supplementary Figure 2. miR-515-5p targets TLR4. (A) Schematic diagram of the binding site between miR-515-5p and TLR4; (B) Dual-luciferase reporter assay to detect the targeting regulation of TLR4 by miR-515-5p; (C) RIP assay to verify the binding between miR-515-5p and TLR4; (D) RT-qPCR detection of miR-515-5p and TLR4 expression in cells; (E–F) WB evaluation of representative bands and quantitative analysis of TLR4 expression in cells. All data are expressed as mean ± standard deviation, N = 3. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs mimic NC group,#p < 0.05 vs inhibitor NC group. Supplementary Material 3: Supplementary Figure 3. miR-515-5p/TLR4 inhibits pyroptosis in RA-FLSs via Caspase-1. After transfection of miR-515-5p inhibitor into BMSCs-Exos and co-culture with RA-FLSs, with or without Caspase-1 inhibitor VX-765 (50 nM) or TLR4 inhibitor TAK242: (A–B) Flow cytometry analysis of pyroptosis ratio using Annexin V + 7AAD double staining; (C) LDH release detection; (D–E) WB detection of representative bands and quantification of pyroptosis-related protein expression. All data are expressed as mean ± standard deviation, N = 3. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs inhibitor NC group, #p < 0.05 vs miR-inhibitor group. Supplementary Material 3: Supplementary Figure 3. miR-515-5p/TLR4 inhibits pyroptosis in RA-FLSs via Caspase-1. After transfection of miR-515-5p inhibitor into BMSCs-Exos and co-culture with RA-FLSs, with or without Caspase-1 inhibitor VX-765 (50 nM) or TLR4 inhibitor TAK242: (A–B) Flow cytometry analysis of pyroptosis ratio using Annexin V + 7AAD double staining; (C) LDH release detection; (D–E) WB detection of representative bands and quantification of pyroptosis-related protein expression. All data are expressed as mean ± standard deviation, N = 3. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs inhibitor NC group, #p < 0.05 vs miR-inhibitor group. Supplementary Material 4: Supplementary Figure 4. miR-515-5p inhibition increases mitochondrial dysfunction in RA-FLSs. (A–B) DCFDH fluorescence probe detection of intracellular ROS levels (scale bar = 25 µm); (C–D) MitoSOX fluorescence (red) to assess mtROS levels in RA-FLSs after 24 h incubation, with representative images and fluorescence quantification (scale bar = 25 µm); (E–F) JC-1 kit detection of mitochondrial membrane potential in RA-FLSs (J-aggregates, red; J-monomers, green; scale bar = 25 µm); (G–H) Fluo-3/AM fluorescence probe detection of calcium levels in IL-1β-treated RA-FLSs. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs IL-1β group, #p < 0.05 vs IL-1β + BMSCs-Exo group, &p< 0.05 vs IL-1β + BMSCs-Exo + miR-inhibitor group. Experiments were independently repeated three times. Supplementary Material 4: Supplementary Figure 4. miR-515-5p inhibition increases mitochondrial dysfunction in RA-FLSs. (A–B) DCFDH fluorescence probe detection of intracellular ROS levels (scale bar = 25 µm); (C–D) MitoSOX fluorescence (red) to assess mtROS levels in RA-FLSs after 24 h incubation, with representative images and fluorescence quantification (scale bar = 25 µm); (E–F) JC-1 kit detection of mitochondrial membrane potential in RA-FLSs (J-aggregates, red; J-monomers, green; scale bar = 25 µm); (G–H) Fluo-3/AM fluorescence probe detection of calcium levels in IL-1β-treated RA-FLSs. One-way ANOVA and Tukey post hoc test were used. *p < 0.05 vs IL-1β group, #p < 0.05 vs IL-1β + BMSCs-Exo group, &p< 0.05 vs IL-1β + BMSCs-Exo + miR-inhibitor group. Experiments were independently repeated three times. Supplementary Material 5: Supplementary Figure 5. FLSs were stimulated with IL-1β (10 ng/mL) to establish an in vitro RA model. Supplementary Material 5: Supplementary Figure 5. FLSs were stimulated with IL-1β (10 ng/mL) to establish an in vitro RA model. Supplementary Material 6. Supplementary Material 6.