Extracellular vesicles from human adipose-derived stem cells relieve pain and inflammation in a rat model of knee osteoarthritis
Woo Sung Kim, Chang Hee Woo, Kyoung Soo Lee, Young Chan Choi, Ye Eun Yun, Ji Suk Choi, Yong Woo Cho
TL;DR
Extracellular vesicles from human fat stem cells reduce pain and inflammation in a rat model of knee osteoarthritis, suggesting a new cell-free treatment option.
Contribution
This study demonstrates the pain-relieving and cartilage-protective effects of hASC-EVs in an OA rat model, offering a novel cell-free therapeutic strategy.
Findings
hASC-EVs significantly suppressed inflammation and pain markers in human osteoarthritic chondrocytes.
In vivo, hASC-EVs improved pain behavior and preserved cartilage in OA rats.
hASC-EVs down-regulated key PI3K/Akt signaling genes associated with inflammation.
Abstract
Inflammatory pain is a hallmark symptom of osteoarthritis (OA), characterized by spontaneous hypersensitivity resulting from tissue damage and chronic inflammation. This study investigates the pain-relieving and cartilage-protective potential of extracellular vesicles (EVs) derived from human adipose-derived stem cells (hASCs) as a cell-free therapeutic approach for OA. hASC-EVs were isolated via multi-filtrations based on tangential flow filtration (TFF) and characterized using transmission electron microscopy (TEM), nanoparticle tracking analysis (NTA), dynamic light scattering (DLS), zeta potential measurement, flow cytometry and Liquid chromatography-mass spectrometry (LC-MS/MS)-based proteomic analysis. An in vitro inflammatory OA model was established by treating human osteoarthritic chondrocytes (HC-OA) with interleukin-1β (IL-1β). The expression of inflammation- and…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8- —Korean Fund For Regenerative Medicine (KFRM)
- —Ministry of Trade, Industry & Energy (MOTIE), Korea Planning & Evaluation Institute of Industrial Technology (KEIT)
- —National Research Foundation of Korea (NRF)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsExtracellular vesicles in disease · Osteoarthritis Treatment and Mechanisms · Mesenchymal stem cell research
Background
Osteoarthritis (OA) is one of the most prevalent joint disorders globally and poses an increasing public health burden, particularly in aging populations. It is characterized by progressive degeneration of articular cartilage, resulting in chronic pain, local inflammation, and impaired joint mobility [1, 2]. Among the clinical manifestations of OA, persistent pain and joint stiffness are particularly debilitating, significantly impairing patients’ quality of life across all disease stages. Current clinical management of OA is largely focused on alleviating pain, with non-steroidal anti-inflammatory drugs (NSAIDs) being the most widely prescribed pharmacological agents [3]. Other medications including aspirin, acetaminophen, and centrally acting agents such as benzodiazepines, are also used to modulate pain transmission and reduce discomfort [4]. However, these medications often provide only transient relief and are associated with considerable side effects, including gastrointestinal complications such as ulcers, especially with prolonged use [5].
OA-related pain arises from multifactorial mechanisms, involving nociceptive and neuropathic components, inflammatory responses, structural damage to joint tissues [6–9]. Injury to cartilage, subchondral bone, and synovium triggers the release of pro-inflammatory cytokines, including interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α), which sensitize peripheral nociceptors and modulate pain perception [10–12]. These inflammatory mediators can also activate nociceptive neurons by engaging receptors such as the transient receptor potential ankyrin 1 (TRPA1) channel, leading to increased neuronal excitability, lowered pain thresholds, and heightened pain sensitivity (hyperalgesia) [13–17]. In parallel, overexpression of matrix metalloproteinases (MMPs) contributes to both cartilage degeneration and inflammation, thereby exacerbating OA-related pain. Recent evidence suggests a mechanistic link between MMP activity and OA pain, proposing MMP inhibition as a promising therapeutic target [18–20].
Extracellular vesicles (EVs) derived from mesenchymal stem cells (MSCs) have emerged as a cell-free therapeutic approach for OA. MSC-derived EVs carry a cargo of functional biomolecules-such as microRNAs, messenger RNAs (mRNAs), protein, and lipids-that mediate intercellular communication and contribute to immunomodulation, tissue repair, and anti-inflammatory effects [21–23]. Growing evidence suggests that MSC-EVs can mitigate inflammation and supports tissue regeneration in several degenerative diseases, including OA [24–29]. We previously reported that human adipose-derived stem cells (hASC)-EVs attenuate OA progression and preserve extracellular matrix homeostasis, indicating their applicability as a therapeutic option for OA [30]. In present study, we investigated the analgesic potential of hASC-EVs, with a focus on their modulation of inflammatory and nociceptive signaling within the OA joint. Human ASC-derived EVs were employed as part of a translational research strategy aimed at evaluating their therapeutic relevance in preclinical OA models. We characterized the physicochemical characteristics of hASC-EVs and evaluated their analgesic effects and underlying mechanisms using both in vitro and in vivo OA models.
Methods
hASCs culture
Primary hASCs were obtained from Matica Biolabs (Seongnam, Korea). Cells were maintained up to passage 5 in Minimum Essential Medium α (MEM α; Life Technologies, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS), 10 µg/mL gentamicin, and 10 ng/mL basic fibroblast growth factor (bFGF) at 37 ℃ in a humidified atmosphere of 5% carbon dioxide (CO_2_). Upon reaching 80–90% confluence, cells were washed and cultured in serum-free, phenol red-free Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 1% GlutaMAX™ and 1% sodium pyruvate for 24 h to collect conditioned medium (CM).
EV isolation and characterization
hASC-EVs were isolated from hASC-derived CM using a tangential-flow filtration (TFF) system (Repligen, USA) with a 300 kDa molecular-weight cut-off (MWCO) filter, as previously reported protocols [30]. Briefly, collected CM was first passed through a 0.22 μm filter to remove cell debris and large particles. The CM was then continuously circulated through the membrane filtration system at an operating flow rate of 100–300 mL/min, with a permeate aspiration rate of 10–50 mL/min. During the process, transmembrane pressure, concentration factor, and flux were monitored in real-time to ensure optimal filtration efficiency. Following concentration, diafiltration was performed by adding phosphate buffered saline (PBS), and the TFF process was repeated to remove residual soluble proteins and low molecular weight contaminants. The final concentrated and purified hASC-EVs were recovered in a volume of approximately 10–20 mL and stored at − 70 °C until further use. hASC-EVs were quantified using a Micro BCA Protein Assay Kit (Thermo Scientific, Rockford, IL, USA), and particle size and count were determined using nanoparticle tracking analyzer (NTA, Nanosight LM100, Malvern Instruments, UK). NTA setting included: particle/frame 10–100, three capture per sample, 30-second capture duration, detection threshold of 3, and screen gain of 10.
Transmission electron microscopy (TEM)
EVs were fixed with 0.5% glutaraldehyde at 4 ℃ overnight, centrifuged (13,000 × g, 3 min), and dehydrated with absolute ethanol. Samples were loaded onto Formvar-carbon coated copper grids (TED PELLA, Inc, CA, USA), stained with 1% phosphotungstic acid, and visualized using a JEM-2100 F electron microscope (JEOL Ltd, Japan). For cryo-TEM, EVs were applied to lacey carbon grid (Electron Microscopy Science, PA, USA), rapidly frozen in liquid nitrogen, and imaged using a Tecnai F20 Twin microscope.
Dynamic light scattering (DLS) and zeta potential
EV size distribution and surface charge (zeta potential) were analyzed by Zetasizer Nano ZS90 (Malvern, UK). EVs were diluted in PBS for size measurements and in deionized water for zeta potential. Measurements were performed in sextuplicate and expressed as mean values.
Flow cytometry analysis
EV surface markers were analyzed using the ExoStep™ kit (Immunostep, Spain) according to the manufacturer’s instructions. EVs were captured on CD63-coated magnetic beads and stained with antibodies against CD9, CD63, CD81 (Immunostep, Spain), GM130 (Novus biologicals, USA), and calnexin (Novus biologicals, USA), followed by streptavidin-PE detection. Samples were analyzed using a NovoCyte Flow Cytometer (ACEA Biosciences, USA).
Proteomic analysis via liquid chromatography-mass spectrometry (LC-MS/MS)
Proteomic profiling of hASC-EVs was conducted by BERITIS Inc (Seoul, Korea). EV proteins were labeled with Tandem Mass Tag (TMT) reagents according to the manufacturer’s protocol. A total of 100 µg of the pooled peptide sample was subjected to analysis using a Thermo Dionex Ultimate 3000 Nano LC/MS/MS system coupled with an Orbitrap Exploris 480 mass spectrometer. The analysis was conducted using three independent biological replicates (n = 3). Identified peptides were matched to the UniProt protein sequence database (https://www.uniprot.org), and gene ontology (GO) and protein-protein interaction (PPI) analyses were performed (p < 0.05).
Cellular uptake assay
EVs were labeled with PKH67 dye (Sigma, USA) and filtered through EV spin columns (MWCO 3000 Da; Invitrogen, USA). Human OA chondrocytes (HC-OA; Cell Applications, USA) were derived from human articular cartilage obtained from donors diagnosed with osteoarthritis, as defined by the vendor. HC-OA were seeded (1 × 10^5^ cells/dish) and incubated with 1 × 10^8^ particles/mL of labeled EVs for 12 h. After 4’,6-diamidino-2-phenylindole (DAPI) nuclei staining (Invitrogen, USA), cellular uptake was visualized by confocal laser scanning microscopy (CLSM; Zeiss, Germany).
hASC-EV treatment in IL-1β-stimulated human OA chondrocytes
HC-OAs were seeded (2 × 10^5^ cells/well in 6-well plates) and incubated in growth medium (GM) for 24 h. Cells were stimulated with recombinant human IL-1β (10 ng/mL; R&D Systems, USA) with or without hASC-EVs (5 × 10^7^, 1 × 10^8^ particles/mL) for 48 h.
Quantitative polymerase chain reaction (qPCR)
Total RNA was extracted using the RNA-spin™ total RNA extraction kit (iNtRON Biotechnology, Korea) and cDNA was synthesized using the RT^2^ first strand kit (Qiagen, Germany). mRNA levels of TRPA1, cyclooxygenase-2 (COX-2), MMP-2, MMP-3, MMP-9, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were quantified using gene-specific primers (Qiagen, Germany) and a Stratagene Mx3000P system (Agilent Technologies, USA). Gene expression data were normalized to GAPDH mRNA levels.
Phosphoinositide 3-kinase/protein kinase B (PI3K/Akt) pathway array
To assess signaling modulation, the Human PI3K-Akt Signaling Pathway Array L1 (AAH-BLG-AKT, Raybiotech, USA), a glass slide-based antibody array detecting 307 human proteins, was used. Protein lysates were extracted from IL-1β stimulated HC-OAs with or without hASC-EVs (5 × 10^7^ particles/mL) using 1× cell lysis buffer (Raybiotech). Total protein concentration was determined by a protein assay, and 248.4 µg of protein per sample was applied to the array (n = 1, analyzed in technical duplicate). Arrays were blocked, incubated with equal protein lysates, and probed with biotinylated antibodies and Cy3-conjugated streptavidin. Fluorescence signals were scanned using an Axon GenePix laser scanner and quantified with GenePix Pro software. Signal intensities were normalized to internal controls.
Monosodium iodoacetate (MIA)-induced OA rat model
All animal procedures were approved by the Institutional Animal Care and Use Committee of Catholic University of Korea (CUMC-2021-0260-01) and the experiment has been reported in line with the ARRIVE guidelines 2.0. Seven-week-old male Wistar rats were purchased from Central Lab Animal, Inc. (Seoul, Korea) and acclimated for one week prior to experimentation. Animals were housed under controlled environment conditions (temperature of 20–26 °C, humidity 30–70%) with free access to standard chow and water.
A total of 23 rats were used in this study. The wild-type (WT) group included 3 rats, while experimental group-vehicle, hASC-EV low dose (hASC-EV L), hASC-EV high dose (hASC-EV H), and celecoxib (positive control)-each included 5 rats. OA was induced in 20 rats by a single intra-articular injection of MIA (Sigma-Aldrich, USA) dissolved in 0.9% sterile saline at concentration of 60 mg/mL. A total volume of 50 µL was injected into the right knee joint cavity under brief isoflurane (2–3%) inhalation anesthesia using a 26-gauge needle. The MIA-induced OA rat model was selected because it reproduces the key pathological features of human OA. Three days after MIA injection, the rats were randomly divided into four groups and treated with intra-articular injections of hASC-EVs (0.5, 1.0 × 10^8^ particles/50 µL, once weekly for three weeks) or vehicle control. Celecoxib (30 mg/kg) was orally administered once daily for three weeks as a positive. Randomization was stratified according to baseline pain-threshold measurements to minimize intergroup variability. The sample size was determined based on prior experience with this model and ethical considerations to minimize animal use. All animals were retained for the final analyses, and all behavioral and histological assessments were conducted by investigators blinded to the treatment allocation. On day 24, the animals were euthanized by CO₂ inhalation. CO₂ was introduced into the animal chamber at a flow rate of 30% volume displacement per minute until at least one minute after the loss of consciousness. Death was confirmed by the absence of a heartbeat and corneal reflex. Following euthanasia, the knee joints were harvested for subsequent analysis.
Pain-related behavioral assessment
Pain-related behavioral assessment was designated as the primary outcome measure and served as the key functional indicator of OA-induced pain. Paw withdrawal latency (PWL) and paw withdrawal threshold (PWT) were measured using a dynamic plantar aesthesiometer (Ugo Basile, Italy). Rats were acclimated to the testing environment before each session and stimulated with a gradually increasing mechanical force (0–50 g over 10 s). The latency and threshold to paw withdrawal were automatically recorded. Weight-bearing asymmetry was assessed using an Incapacitance tester (Model 600, IITC, USA). Measurements were conducted twice per week, and the weight-bearing ratio was calculated as follows:
\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\% Weight \, bearing = \left( {Right \, limb \, weight /Total \, hind \, limb \, weight} \right) \times 100$$\end{document}Histological and immunohistochemical (IHC) analysis
Knee joints were fixed in 10% paraformaldehyde (7 days), decalcified in 10% ethylenediaminetetraacetic acid (EDTA) solution (30 days), paraffin-embedded, and sagittally sectioned. Sections were stained with safranin O and scored using the osteoarthritis research society international (OARSI) grading system [31]. IHC was performed with antibodies against IL-1β (1:400 dilution, Novus), MMP-13 (1:50 dilution, Abcam), CD86 (1:200, Bioss) and collagen Ⅱ antibody (1:300 dilutions, Abcam). Sections were counterstained with hematoxylin. Positively stained cells were examined at 400× magnification, corresponding to a high-power field (HPF). For IL-1β, MMP-13, and CD86, expressions levels were quantified as the absolute number of positively stained cells per HPF within the synovial tissue adjacent to the cartilage. Collagen type II expression was quantified as the percentage of positively stained areas relative to the total cartilage area. All quantitative analyses were performed using three representative sections per group.
Statistical analysis
Data were presented as the mean ± standard deviation (SD). Statistical analyses were performed using GraphPad Prism software (version 8.0; GraphPad Software, San Diego, CA, USA). Comparisons among multiple groups were performed using two-way analysis of variance (ANOVA) followed by Bonferroni’s post hoc test or the Kruskal-Wallis test followed by Dunn’s post hoc test, as appropriate. For comparisons between two groups, the Mann-Whitney U test was used. Statistical significance was indicated as follows: ^^p < 0.05; ^^p < 0.01; ^^p < 0.001; ^****^p < 0.0001; ns (p > 0.05): non-significant.
Results
Characterization of hASC-EVs
hASC-EVs were successfully isolated using a TFF-based multi-filtration system and characterized through a series of analytical techniques (Fig. 1). TEM image revealed that the EVs exhibited a typical round, bilayered spherical morphology (Fig. 1A). DLS showed that the average particle size was 88 nm, and the zeta potential was − 26 mV, indicating stable dispersion (Fig. 1B, C). Flow cytometry analysis confirmed the presence of EV surface markers (CD9, CD63, and CD81), while endoplasmic reticulum and golgi markers (Calnexin and GM130) were not detected, verifying the purity of the EV preparations (Fig. 1D).
Fig. 1. Characterization of extracellular vesicles from human adipose-derived stem cells (hASC-EVs). A Classic and cryogenic transmission electron microscopy (TEM) images. Scale bars, 50 nm (left) and 100 nm (middle and right). B Representative particle size and distribution measured by dynamic light scattering (DLS). The mean particle size was 88 nm. C Zeta surface potential measurements (n = 6). The mean surface charge was − 26 mV. D Flow cytometric analyses of EV surface markers (CD9, CD63 and CD81) and cellular contaminants (GM130 and calnexin)
Bioinformatics analysis of hASC-EVs
Proteomic profiling of hASC-EVs using LC-MS/MS identified an average of approximately 3,200 proteins. To elucidate the functional roles of these proteins, GO enrichment analysis was performed, covering three major categories: biological process (BP), cellular component (CC), and molecular function (MF) (Fig. 2). In the BP category, enriched terms included “response to stress” and “response to cytokine”, indicating the potential of hASC-EVs to modulate inflammatory pathways. In the CC category, EV-associated terms such as “vesicle” and “extracellular exosome” were prominently represented, supporting the vesicular origin of the protein cargo. In the MF, terms such as “protein binding” and “RNA binding” were enriched, suggesting a functional role in post-transcriptional regulation. Of note, GO terms related to “NIK/NF-κB signaling” were identified among the enriched categories, suggesting a potential association with inflammation-related signaling pathway.
Fig. 2. Proteomic analysis of hASC-EVs. A Gene ontology (GO) enrichment analysis of EV proteins categorized by biological process (BP), cellular components (CC), and molecular functions (MF) Significantly over-represented GO terms were selected from the total GO categories (n = 3, p < 0.05). B Protein–protein interaction (PPI) network analysis conducted using public databases
To further interpret the functional landscape, a PPI network was constructed using publicly available databases (Fig. 2B). Network clustering and annotation revealed that hASC-EV proteins are involved in multiple biological functions, including inflammatory pain regulation, MMP inhibition, macrophage inflammatory regulation, and cartilage regeneration. Notably, the presence of tissue inhibitors of metalloproteinases (TIMPs) within the EV proteome supports their potential role in suppressing matrix degradation and inflammatory signaling through MMP inhibition [32, 33].
Cellular uptake and anti-inflammatory effects of hASC-EVs in human OA chondrocytes
To assess whether hASC-EVs are effectively internalized by HC-OAs, cells were incubated with PKH67-labeled EVs (1 × 10^8^ particles/mL) for 12 h. Confocal microscopy revealed prominent green fluorescence signals within the cytoplasm and perinuclear region, confirming successful EV uptake (Fig. 3A).
Fig. 3. Anti-inflammatory effect of hASC-EVs on IL-1β stimulated human chondrocytes (HC-OA). A Confocal images of HC-OAs treated with PKH67-labeled hASC-EVs, showing intracellular uptake. Scale bar, 20 μm. B–F The mRNA expression of TRPA1, COX-2, MMP-2, MMP-3, and MMP-9 was quantified using qPCR and normalized to GAPDH. Data are presented as means ± SD (n = 3). ns, not significant (p > 0.05), ^^p < 0.05, ^^p < 0.01, ^^p < 0.001, ^****^p < 0.0001 vs. IL-1β only group
To evaluate the anti-inflammatory and pain-related effects of hASC-EVs, HC-OAs were stimulated with IL-1β (10 ng/mL) in the presence or absence of hASC-EVs (0.5 and 1 × 10^8^ particles/mL) for 48 h. qPCR analysis demonstrated that IL-1β stimulation significantly upregulated the expression of pain- and inflammation-associated genes, including TRPA1, COX-2, MMP-2, MMP-3, and MMP-9. Co-treatment with hASC-EVs led to a marked and dose-dependent reduction in the expression of these genes (Fig. 3B–F), indicating attenuation of catabolic and nociceptive responses in inflammatory chondrocytes. In addition, western blot analysis showed that hASC-EV treatment reduced nuclear factor-kappa B (NF-κB) phosphorylation in IL-1β-stimulated HC-OAs (Supplementary Fig. S2), further supporting the anti-inflammatory activity of hASC-EVs at the cellular level.
Modulation of PI3K/Akt signaling pathway by hASC-EVs
To assess the regulatory effects of hASC-EVs on the PI3K/Akt signaling pathway in inflammatory chondrocytes, a targeted protein array was employed to profile 307 key proteins involved in OA, inflammation, and pain pathways. Upon IL-1β stimulation, a significant upregulation of phosphorylated AKT1, AKT3, and upstream components phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA) and phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit beta (PIK3CB) were observed, indicating activation of the PI3K/Akt pathway (Fig. 4; Table 1). This was accompanied by increased expression of pro-inflammatory mediators, such as interleukin-6 (IL-6), and signaling molecules including insulin receptor (INSR) and insulin receptor substrate 1 (IRS1). Conversely, IL-1β treatment led to a downregulation of protective factors such as B-cell lymphoma 2 (BCL2), insulin like growth factor-1 (IGF1), and cyclin D1 (CCND1), which are associated with anti-inflammatory responses, cartilage homeostasis, and cell survival [34–36]. These results underscore the deleterious impact of inflammatory stimulation on chondrocyte viability and tissue integrity.
Fig. 4. Modulation of PI3K/Akt/NF-κB pathway by hASC-EVs. A Schematic representation of PI3K/Akt pathway and its role in OA progression. B Heatmap of differentially expressed proteins following hASC-EV treatment (5 × 10^7^ p/mL), showing distinct patterns of downregulated and upregulated proteins associated with inflammatory signaling and chondroprotective pathways
Table 1. Differential expression of PK3K/Akt pathway-related proteins in response to treatments (log_2_ fold change)ProteinControl (A)IL-1βtreatment (B)hASC-EV5 × 10^7^ p/mL (C)log_2_(FC) (A vs B)log_2_(FC) (B vs C)AKT17098849975270.26− 0.175AKT318,91132,19018,8270.767− 0.774CDKN1A870611,18690160.362− 0.311COL6A27315988178580.434− 0.331CSF315,98420,72218,0030.374− 0.203EGF2665383535450.525− 0.113EGFR1643376528051.197− 0.425EIF4EBP12073264421090.351− 0.326ERBB31490209312220.49− 0.776FGF211864581471.301− 1.637FGF72377298125640.327− 0.217FGF81988369333870.893− 0.125FN11380901044252.707− 1.026GHR3515072640.532− 0.942HSP90B111,87921,70217,1210.869− 0.342IKBKG1975296018390.584− 0.686IL623,33334,53231,2470.566− 0.144INS12,32317,64512,7170.518− 0.472INSR24,81632,68227,9240.397− 0.227IRS12292337021220.556− 0.668ITGA53231390330800.273− 0.341LAMB119,90824,50520,1300.3− 0.284LPAR116312082622.894− 2.207MAP2K12654812360.863− 1.028NTRK140,44361,30824,9210.6− 1.299NTRK21461243211780.735− 1.046OSM89714741560.717− 3.24PIK3CA4137434539870.071− 0.124PIK3CB1763286923130.703− 0.31PRL2543480634440.919− 0.481RELA3732476632420.353− 0.556THBS279412758540.683− 0.579TEK147516559370.166− 0.82TNC2595261111.022− 2.25BCL2137411741322− 0.2270.171CCND1281143574− 0.982.009HGF918773874− 0.2480.178IGF1393136503894− 0.1070.094PIK3R137638378− 3.3113.319THBS421189751198− 1.120.297
Treatment with hASC-EVs significantly attenuated the IL-1β-induced activation of the PI3K/Akt pathway. Expression levels of AKT1, AKT3, PIK3CA and PIK3CB were markedly reduced compared to IL-1β treatment alone, suggesting suppression of hyperactive PI3K/Akt signaling. Furthermore, hASC-EVs reversed the downregulation of BCL2, IGF1, and CCND1, restoring their expression to near-baseline levels (Fig. 4B). The decreased expression of INSR, IRS1, and IL-6 further supports the anti-inflammatory and cytoprotective effects of hASC-EVs [37]. Collectively, these data indicate that hASC-EVs modulate PI3K/Akt/NF-κB pathway activity by repressing pro-inflammatory signaling while reactivating genes critical for chondroprotection and cell survival, offering a molecular basis for their therapeutic potential in OA.
hASC-EVs alleviate pain in MIA-induced OA rats
To evaluate the analgesic effect of hASC-EVs in vivo, a MIA-induced OA rat model was employed. Rats were intra-articularly injected with hASC-EVs (0.5, 1 × 10^8^ particles, once weekly for three weeks, n = 5), celecoxib (30 mg/kg, oral, daily for three weeks, n = 5), or vehicle (PBS, once weekly for three weeks, n = 5) following MIA induction (Fig. 5A). Pain-related behavioral parameters, including PWT, PWL, and hind limb weight-bearing asymmetry, were measured biweekly for 24 days. MIA injection noticeably reduced PWL, PWT, and weight-bearing capacity compared to baseline and healthy controls, reflecting the induction of OA-associated mechanical hypersensitivity. Treatment with hASC-EVs led to a significant improvement in both PWL (hASC-EV L: 11.78 s; hASC-EV H: 12.25 s; Celecoxib: 10.63 s; Vehicle: 8.97 s) and PWT (hASC-EV L: 23.75 g; hASC-EV H: 24.77 g; Celecoxib: 21.48 g; Vehicle: 18.18 g) beginning on day 6 post-injection. Notably, hASC-EV-treated groups exhibited superior or comparable analgesic effects relative to the celecoxib group. In the weight-bearing test, significant improvements were observed as early as day 11 in the high-dose EV group, with all hASC-EV-treated rats displaying significantly improved weight distribution between the hind limbs by day 14 (hASC-EV L: 38.48%; hASC-EV H: 38.12%; Celecoxib: 39.18%; Vehicle: 34%) (Fig. 5B). These findings suggest that hASC-EVs effectively attenuate OA-related nociception and mechanical allodynia.
Fig. 5. Pain-relieving effects of hASC-EVs in a MIA-induced OA rat model. A Experimental timeline of intra-articular hASC-EV administration and oral celecoxib treatment. B Behavioral assessments of pain using paw withdrawal threshold (PWT), paw withdrawal latency (PWL), and weight bearing tests from baseline to day 24. Data are presented as means ± SD (WT, n = 3; others, n = 5). Statistical significance was determined relative to the vehicle group (^^p < 0.05, ^^p < 0.01, ^^p < 0.001, ^****^p < 0.0001). WT, wild type; hASC-EV L: 5 × 10^7^ particles; hASC-EV H: 1 × 10^8^ particles. Units: s, second; g, gram-force
hASC-EVs attenuate synovial inflammation in MIA-induced OA rats
To evaluate the anti-inflammatory effects of hASC-EVs, immunohistochemical analysis of synovial tissue was performed. IL-1β and MMP-13, key mediators of synovial inflammation and cartilage degradation, were highly elevated in the vehicle-treated group. In addition, CD86, a representative marker of pro-inflammatory (M1) macrophages, was also significantly increased in the vehicle group. In contrast, hASC-EV treatment significantly reduced the expression of IL-1β and MMP-13 and CD86 in synovial tissue, indicating attenuation of inflammatory and macrophage-associated responses (Fig. 6A-F). Notably, celecoxib treatment did not result in a comparable reduction in these markers, suggesting a more pronounced suppression of synovial inflammatory responses by hASC-EVs under the experimental conditions tested.
Fig. 6. Suppression of synovial inflammation by hASC-EVs. A, B Representative IHC images and quantification of IL-1β-positive cells in synovium (hASC-EV L: 190.6; hASC-EV H: 129.2; Celecoxib: 122.4; Vehicle: 278.2). Scale bars, 100 μm. C, D IHC and quantification of MMP13 expression in synovium (hASC-EV L: 248.8; hASC-EV H: 203.4; Celecoxib: 207.8; Vehicle: 353.8). Scale bars, 100 μm. E, F IHC and quantification of CD86-positive cells (hASC-EV L: 158.4; hASC-EV H: 153.8; Celecoxib: 221.6; Vehicle: 288.2). Scale bars, 100 μm. Data are presented as means ± SD (WT, n = 3; others, n = 5). Statistical significance was determined relative to the vehicle group (^^p < 0.05, ^^p < 0.01, ^^p < 0.001, ^****^p < 0.0001). WT, wild type; hASC-EV L: 5 × 10^7^ particles; hASC-EV H: 1 × 10^8^ particles
hASC-EVs preserve cartilage integrity in MIA-induced OA rats
Histological analysis of joint sections using OARSI grading criteria (Supplementary Table S1) revealed pronounced cartilage surface erosion and matrix breakdown in the vehicle-treated group. In hASC-EV-treated rats, cartilage structure was notably preserved, with lower OARSI scores compared to the vehicle group (Fig. 7A, B, Supplementary Fig. S3A). These findings indicate that hASC-EVs effectively mitigate cartilage degradation in OA. Further analysis of cartilage matrix integrity focused on collagen type II, the principal structural component of articular cartilage. Immunohistochemical staining demonstrated a reduction in collagen type II in the vehicle group. Conversely, all hASC-EV-treated groups exhibited robust preservation of collagen type II in the cartilage matrix, even in areas of damage (Fig. 7C, D, Supplementary Fig. S3B). Collectively, these results demonstrate that hASC-EVs exert dual protective effects in OA by suppressing inflammatory mediators in synovial tissue and preventing cartilage matrix degradation, thus preserving joint integrity.
Fig. 7. Cartilage-protective effects of hASC-EVs in MIA-induced OA rats. A,** B** Histological analysis of knee joint stained with Safranin-O. Cartilage degradation scored using OARSI system (hASC-EV L: 4.95; hASC-EV H: 4; Celecoxib: 3.9; Vehicle: 5.3). Scale bars, 100 μm. C, D IHC analysis of type II collagen preservation in joint cartilage (hASC-EV L: 23.65%; hASC-EV H: 23.02%; Celecoxib: 30.83%; Vehicle: 2.76%). Scale bars, 100 μm. Data are presented as means ± SD (WT, n = 3; others, n = 5). Statistical significance was determined relative to the vehicle group (^^p < 0.05, ^**^p < 0.01, ^***^p < 0.0001). WT, wild type; hASC-EV L: 5 × 10^7^ particles; hASC-EV H: 1 × 10^8^ particles
Discussion
OA is not only a degenerative disease of cartilage and synovium but also a major cause of chronic pain, representing an unmet medical need of critical importance. Pain in OA arises from a complex interplay of inflammatory mediators, peripheral sensitization, and joint structural damage [38], yet current treatment options are largely limited to symptomatic relief via NSAIDs or corticosteroids, which have limited efficacy and adverse effect profiles [5, 39]. In this study, our findings reveal that hASC-EVs offer a non-pharmacologic and multi-targeted therapeutic strategy with potent pain-relieving efficacy, as demonstrated in both cellular and animal models. Proteomic profiling revealed that hASC-EVs encapsulate proteins involved in key inflammatory signaling pathways, including NF-κB, and are enriched in TIMPs, which are known to inhibit MMP-driven cartilage degradation (Supplementary Fig. S1). In vitro, hASC-EVs significantly attenuated IL-1β-induced upregulation of pain-associated genes in human OA chondrocytes, including TRPA1, a known ion channel involved in mechanical allodynia and inflammatory pain [40, 41]. In addition, COX-2, a key enzyme driving prostaglandin-mediated nociceptive signaling [42, 43], was robustly suppressed by EV treatment in a dose-dependent manner. These findings suggest that hASC-EVs interfere with both nociceptor sensitization and inflammatory mediator production, offering a dual mechanism for pain alleviation. Importantly, suppression of MMP-2, -3, and − 9 may also indirectly contribute to pain reduction by preserving extracellular matrix integrity and limiting joint tissue degradation that sensitizes nociceptors [44–46]. Mechanistically, our findings suggest that hASC-EVs modulate hyperactivation of the PI3K/Akt/NF-κB signaling axis, a central driver of inflammatory responses and chondrocyte catabolism. Proteomic analysis indicated that IL-1β stimulation was associated with upregulation of key PI3K/Akt-related molecules, including PIK3CA/B, AKT1/3, IRS1, and IL-6, all of which have been implicated in inflammation-induced hyperalgesia and OA progression [47–49]. In contrast, hASC-EVs treatment downregulated these pro-inflammatory signaling components while restoring the expression of anti-inflammatory and pro-survival factors such as IGF1, BCL2, and CCND1 (Fig. 4), which are known to rebalance pro- and anti-nociceptive signaling pathway [34–36]. Consistent with these proteomic changes, hASC-EVs significantly attenuated NF-κB phosphorylation in IL-1β-stimulated HC-OAs (Supplementary Fig. S2), providing functional support for the pathway identified by proteomic analysis.
Given the well-established role of NF-κB as a key regulator of inflammatory gene expression-including MMPs and pro-inflammatory cytokines [50, 51], these finding support a biologically relevant association between the proteomic changes induced by hASC-EV treatment and the observed anti-inflammatory and analgesic effects. Although direct in vivo assessment of PI3K/Akt signaling was beyond the scope of this study, the suppression of NF-κB activation observed in vitro provides functional support for the relevance of this pathway in mediating the therapeutic effects of hASC-EVs. Meanwhile, the therapeutic relevance of this analgesic effect was clearly demonstrated in vivo. In the MIA-induced OA rat model—widely used for evaluating pain phenotypes resembling human OA [14, 18]—treatment with hASC-EVs led to rapid and sustained improvements in PWL and PWT, two gold-standard readouts of mechanical hyperalgesia. Remarkably, EV-treated animals exhibited superior or equivalent pain relief compared to celecoxib, a COX-2 selective NSAID commonly used in clinical settings. This superiority is particularly noteworthy given that celecoxib primarily targets downstream prostaglandin synthesis with some reported disease-modifying effect [42, 43], whereas hASC-EVs act upstream, modulating inflammatory cascades and neuroimmune interactions, thus enabling a more comprehensive disease-modifying effect [52–55]. Furthermore, hASC-EVs significantly improved hind limb weight-bearing asymmetry, a functional pain-related behavior reflecting spontaneous discomfort and mobility impairment. The early onset of analgesia (as early as day 6) and the dose-dependent response pattern strengthen the translational potential of EV-based interventions. Histological and immunohistochemical analyses revealed that hASC-EVs effectively reduced the expression of IL-1β, MMP-13 and CD86 in synovial tissues while preserving collagen type II in articular cartilage, an effect that was not observed with celecoxib. The reduction in CD86 expression in the hASC-EV-treated group suggests suppression of pro-inflammatory (M1) macrophage activation and a partial rebalancing of the inflammatory microenvironment within the OA joint. These in vivo findings are consistent with our in vitro results showing inhibition of the TIMP-MMP axis and attenuation of NF-κB signaling following hASC-EV treatment, supporting a mechanistic link between EV-mediated immunomodulation and cartilage-protective effects. Importantly, hASC-EVs suppressed inflammatory cytokines and reduced cartilage degradation, collectively contributing to pain amelioration. These findings align with our proteomic and bioinformatics analyses, which implicated hASC-EV-associated proteins in immunomodulation, MMP regulation, and cartilage repair. Taken together, our results underscore the utility of hASC-EVs as a cell-free, multi-targeted therapeutic modality capable of addressing both pain and structural degeneration in OA. Unlike NSAIDs that focus solely on symptomatic relief via COX-2 inhibition, hASC-EVs modulate upstream signaling networks, enabling a more comprehensive disease-modifying effect. Given the limitations of current pharmacotherapies in treating OA pain, particularly in patients with comorbidities or NSAID intolerance, the development of EV-based therapeutics offers a promising alternative. Their nanoscale size, low immunogenicity, and ability to cross biological barriers may provide added advantages in reaching inflamed and innervated joint tissues. This study represents an early preclinical step toward the translational development of hASC-EV-based therapy, and several limitations should be considered. First, the use of human ASC-derived EVs in a rat OA model involves cross-species administration, which may influence specific ligand-receptor interactions despite the conserved nature of many EV-associated regulatory molecules. Future studies using species-matched EVs or advanced translational models will help further validate these findings. Second, although pain-related behavioral outcomes and inflammatory signaling pathways were assessed, direct analysis of nociceptive neuropeptides such as substance P or nerve growth factor (NGF) in the synovial fluid was beyond the scope of the present study and should be incorporated in future mechanistic investigations. Finally, expanded studies in larger cohorts and additional preclinical models, including large-animal OA models, will be important to further support the clinical translation of hASC-EVs.
Conclusions
hASC-EVs represent a potent, multi-modal therapeutic strategy for OA pain. By targeting both inflammatory and nociceptive signaling pathways, most notably through modulation of the PI3K/Akt axis and suppression of TRPA1 and COX-2, hASC-EVs effectively attenuate joint inflammation, protect cartilage integrity, and relieve mechanical hypersensitivity. Unlike conventional analgesics that primarily offer symptomatic relief, hASC-EVs exhibit disease-modifying properties by concurrently addressing the upstream drivers and downstream manifestations of OA-related pain. These findings highlight the potential of hASC-EVs as a next-generation, cell-free therapeutic candidate for managing chronic joint pain and altering the course of degenerative joint diseases.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary Material 1
Supplementary Material 2
The reference list from the paper itself. Each links out to its DOI / PubMed record.