Document Type : Review article
Introduction
Recent studies indicate that Osteoarthritis (OA) affects more than 500 million individuals, representing more than 3.3% of the world's population 1. OA leads to chronic pain, progressive joint tissue damage, stiffness, and substantial loss of mobility. The economic complications are not only due to direct healthcare costs, but also because of productivity loss and long-term disability 2. Because cartilage is a nonvascular tissue with limited self-repair capability, developing durable regenerative strategies remains a major clinical challenge. Existing therapeutic strategies for OA mainly focus on symptom management instead of structural regeneration of cartilage tissue. Conventional treatments, in-cluding Nonsteroidal Anti-Inflammatory Drugs (NSAIDs), corticosteroid injections, hyaluronic acid supplementation, and physical rehabilitation, may briefly improve pain and recover joint function; nevertheless, they fail to avert progressive cartilage degeneration. Surgical interventions such as microfracture, osteochondral grafting, and Autologous Chondrocyte Implantation (ACI) have established partial regenerative potential, yet regularly result in fibrocartilage formation rather than mechanically durable hyaline cartilage 3. In advanced disease stages, total joint arthroplasty remains the decisive treatment option, in spite of concerns regarding implant longevity, revision surgery, and postoperative complications. Consequently, at present none of the available therapies consistently restores native cartilage architecture or effectively halts OA progression, highlighting the crucial need for disease-modifying regenerative strategies.
To overcome these limitations, regenerative medicine approaches based on cell therapy have been widely investigated. Mesenchymal Stem Cells (MSCs), induced Pluripotent Stem Cells (iPSCs), and autologous chondrocytes have shown potential to promote cartilage regeneration because of their differentiation capacity and paracrine activity. Among these, MSCs derived from bone marrow, adipose tissue, umbilical cord, and synovium are the most broadly studied because they can modulate inflammation, support extracellular matrix synthesis, and promote tissue repair 4,5. However, several serious barriers continue to restrict the clinical translation of cell-based therapies, including low cell survival after transplantation, donor-to-donor variability, limited engraftment efficiency, phenotypic instability, risk of unwanted differentiation or tumorigenicity, and potential immunogenic responses. Challenges associated with large-scale manufacturing, storage, and regulatory standardization further complicate their clinical implementation. These limitations have increased attention toward alternative cell-free regenerative strategies capable of preserving therapeutic value while minimizing the risks related to live-cell transplantation.
Extracellular Vesicles (EVs) have recently emerged as promising cell-free therapeutic candidates for cartilage repair and OA management. EVs are membrane-bound nanoscale vesicles naturally secreted by nearly all cell types and play vital roles in intercellular communication through the transfer of proteins, lipids, and nucleic acids 6,7. Among EV subtypes, exosomes have attracted particular attention because of their small size, intrinsic biocompatibility, and ability to deliver bioactive cargoes that regulate inflammation, apoptosis, and tissue regeneration. Collective evidence suggests that MSC-derived EVs can reproduce many of the beneficial paracrine effects of their parental cells while demonstrating lower immunogenicity, improved biosafety, and reduced tumorigenic risk compared with conventional cell therapies 8. Preclinical studies have shown that EVs can reduce inflammatory signaling, enhance chondrocyte proliferation, stimulate extracellular matrix synthesis, and promote cartilage repair in OA models 4.
Despite these advantages, various significant limitations continue to delay the clinical translation of EV-based therapies. Natural EVs show considerable heterogeneity in their composition, biological activity, and therapeutic potency depending on donor cell source, isolation method, and culture conditions. Besides, challenges associated with inefficient targeting, rapid clearance, limited cargo-loading capacity, and lack of standardized manufacturing protocols limit their reproducibility and therapeutic efficacy. To address these barriers, growing attention has focused on engineering strategies designed to boost EV stability, targeting specificity, cargo delivery, and regenerative efficiency.
In this review, we critically examine engineered EV-based strategies for cartilage regeneration and OA therapy with a specific focus on their translational readiness. Unlike previous reviews that mainly summarize the biological effects of MSC-derived EVs, this article integrates EV source heterogeneity, engineering approaches, cargo-loading strategies, delivery platforms, and regulatory barriers into a single translational framework. We distinguish between in vitro evidence, preclinical animal data, and emerging early-stage clinical findings to avoid overestimating the current maturity of EV-based therapies. By emphasizing engineered EVs, precision-oriented source selection, manufacturing standardization, and clinical implementation challenges, this review aims to define the key scientific and regulatory milestones required before EV-based interventions can progress toward routine OA treatment.
Heterogeneity of EVs
EV Subtypes: EVs represent a diverse and heterogeneous group of membrane-bound particles released by cells. They are usually classified into three main categories, exosomes, microvesicles, and apoptotic bodies, based on their size, mode of biogenesis, and molecular characteristics (Figure 1) 9. Exosomes (30–150 nm) originate from the endosomal pathway and are released via multivesicular bodies. Microvesicles (100–1000 nm) are formed through direct budding of the plasma membrane. Apoptotic bodies (500–2000 nm) are released during programmed cell death and contain cellular fragments.
Although EV subtypes share the ability to transport proteins, lipids, and nucleic acids, they differ substantially in biogenesis, molecular composition, and biological function. Exosomes are the most extensively investigated subtype in regenerative medicine because of their stability, small size, and enrichment in signaling molecules involved in tissue repair and immunomodulation. In contrast, microvesicles exhibit greater heterogeneity because they directly originate from plasma membrane budding and often reflect the physiological state of the parental cell. Apoptotic bodies are less commonly explored for therapeutic applications because they may contain fragmented intracellular components associated with cell death and inflammatory signaling.
The selection of a source for therapeutic EVs should be practical and sustainable to support clinical applications. However, because EV composition and therapeutic activity are strongly influenced by donor-cell origin, culture conditions, and isolation methods, there remains no consensus regarding the optimal EV source for regenerative applications.
Functional Variations across EV Sources: EV therapeutic activity varies substantially depending on donor-cell origin, as cellular source influences EV cargo composition, immunomodulatory properties, and regenerative potential. In OA research, MSC-derived EVs have received particular attention because MSCs can be obtained from multiple tissues (Table 1), including adipose tissue 10, bone marrow 11, umbilical cord 12, synovium 13, dental follicles 14, induced Pluripotent Stem Cells (iPSCs) 15, amniotic fluid 16, human progenitor endothelium 17, and endometrial stem cells 18. These EVs can serve as carriers for bioactive molecules involved in cartilage regeneration and tissue repair. Recent studies have shown that MSC-derived EVs perform better than do MSCs themselves in reducing cartilage breakdown, joint inflammation, and pain, even though the overall effects are moderate. Interestingly, treatment with MSCs alone sometimes leads to worse joint outcomes, especially in patients with metabolically driven OA. These findings, which are mainly derived from preclinical and early translational studies, suggest that MSC-EVs may offer a safer cell-free alternative to MSC transplantation; however, their superiority in human OA remains to be confirmed in adequately powered clinical trials 19. Furthermore, EVs can be obtained from a variety of other sources, such as macrophages 20; biological fluids, such as urine or semen; chondrogenic progenitor cells; synovial fluid; fibroblast-like synoviocytes; bovine milk; marine organisms; and plant-derived EVs 21.
In addition to MSC-derived EVs, tissue-specific EV populations have attracted attention because they may better preserve cartilage-associated signaling properties and microenvironmental compatibility. Unlike TGF-β treatment, Chondrocyte-derived EVs (CC-EVs) enhance the chondrogenic differentiation and proliferation of Human Umbilical Cord Mesenchymal Stem Cells (HUCMSCs) without inducing hypertrophy or fibrosis. CC-EVs increased key chondrogenic markers, including ACAN, COL2A, and SOX-9, and aided in cartilage repair. The increase in autophagosomes indicates that CC-EVs facilitate cartilage regeneration through autophagy activation, underscoring their potential in cell-free articular cartilage repair strategies 22.
Table 1. Summary of in vivo and in vitro studies using MSC-derived EVs for OA and cartilage repair
|
EV source |
Therapeutic effect |
Model used |
Origin |
Method (EVs isolation, characterization, dosage, etc.) |
Ref. |
|
Bone Marrow MSC-EVs |
BM-MSC-EVs decreased inflammation and increased proteoglycan and type II collagen production in OA chondrocytes |
TNF-α-stimulated OA chondrocytes (in vitro) |
Human |
Isolation by differential centrifugation, characterization by Western blot (CD9, CD63), NTA, and immunoelectron microscopy. Dosage: Equivalent of 500x103 BM-MSC cells (up to ~1.8x10^9 particles per dose). Injection: In vitro culture and coculture with OA chondrocytes |
[1] |
|
Equine BM-MSC-EVs reduced MMP-13 expression in chondrocytes under pro-inflammatory conditions |
Equine chondrocytes (in vitro) |
Equine |
Isolation by ultracentrifugation, characterization by TEM, NTA, CD9 immunostaining. Dosage: 13,333 EVs per chondrocyte |
[2] |
|
|
BM-MSC-EVs modulate macrophage polarization, lower oxidative stress, and protect mitochondria via PINK1/Parkin pathway |
Rat OA model |
Human |
Isolation by ultracentrifugation, characterization by TEM, nanoparticle tracking analysis (NTA), Western blot for protein markers (PINK1, Parkin, Akt), JC-1 staining. Dosage: 100 μg/ml for macrophage treatment, 50 μl for rat intra-articular injection |
[3] |
|
|
LIPUS enhances MSC exosome release via autophagy, improving chondrocyte function and cartilage repair |
Rat bone marrow MSCs and rat OA knee model |
Rat |
MSC Isolation: Bone marrow from SD rats cultured in DMEM +10% FBS. Autophagy: Activated by rapamycin (10 μM), inhibited by 3-MA (10 μM). Exosome Isolation: Ultracentrifugation. LIPUS: 3 MHz, 50 mW/cm², for 20 mins/day. Coculture: MSC + OA chondrocytes. Histopathology: Safranin-O staining, Mankin scoring. Injection: MSC + GW4869 (10 μM). Analysis: Western blot, NTA |
[4] |
|
|
Cryogel (0.3% HA) + BM-MSC-Exos enhanced ECM production, GAG synthesis, and cartilage repair |
Rabbit cartilage defect model |
New Zealand rabbit |
Exosome isolation by ultracentrifugation, Characterized by NTA and TEM, Dosage: 106 particles/ml, Cryogel seeding, Injection into cartilage defects |
[5] |
|
|
Reduced MMP13, increased COL2, improved behavior, lowered CGRP/iNOS in DRG |
OA Rat Model |
Rat |
Exosome isolation via ultracentrifugation, Characterization by TEM, Zetasizer, Western blot, Dosage: 40 μg/week, Injection into knee joint cavity, Fluorescence imaging for uptake |
[6] |
|
|
Exosomes and MPs reduced joint damage, restored ECM markers, suppressed inflammation |
CIOA Mouse Model |
Mouse |
EV isolation via differential centrifugation, Characterization by DLS, Nano Tracking Analysis, Size: 96 nm for Exos, 223 nm for MPs, Dosage: Exosomes (250 ng), MPs (500 ng), Injection into joint cavity |
[7] |
|
|
Improved pain, suppressed abnormal nerve/vessel formation in the LBP mouse model |
LFJ OA Mouse Model |
Mouse |
EV isolation via differential ultracentrifugation, Characterization by TEM and Western Blot (CD9, CD63, TSG101 markers), Dosage: 200 μg exosomes, injection via the tail vein |
[8] |
|
|
Combined therapy improved histological and mechanical repair vs. HA alone |
Rabbit osteochondral defect model |
Rabbit |
Exosomes isolated from MSCs (isolation and characterization method N/A), combined with HA, three intra-articular injections at 0, 7, and 14 days, evaluated at 6 and 12 weeks, macroscopic, histologic, and biomechanical analysis |
[9] |
|
|
Improved function in 132 joints of Navy veterans with no adverse events |
OA in the Navy SEALs |
Human |
Sample Size: 33 Navy SEAL veterans Follow-up: 12 hr, 24 hr, 48 hr, 2 weeks, 6 weeks, 3 months, 6 months, 1 year |
[10] |
|
|
A single EVIP dose yielded 75% improvement in a grade III OA athlete |
Grade III OA patient |
Human |
Patient: 51-year-old male athlete with a history of increasing left medial and patellofemoral knee pain, previously failed treatments |
[11] |
Contd. Table 1. Summary of in vivo and in vitro studies using MSC-derived EVs for OA and cartilage repair
|
EV source |
Therapeutic effect |
Model used |
Origin |
Method (EVs isolation, characterization, dosage, etc.) |
Ref. |
|
Adipose MSC-EVs |
ADSC-EVs enhance BMSC migration, proliferation, and chondrogenic/osteogenic differentiation |
Rat MSCs (in vitro), murine model |
Rat |
MSC Isolation: ADSCs, BMSCs, and SMSCs were isolated from rat adipose, bone marrow, and synovium, respectively. differentiation of MSCs. |
[12] |
|
AD-MSC-EVs and UC-MSC-EVs repaired cartilage and bone in the pig OA model; AT-MSCs were most effective |
Humanized pig OA model |
Human |
ISOLATION: Ultracentrifugation, CHARACTERIZATION: NTA, Flow Cytometry Intra-articular Injection |
[13] |
|
|
Human AD-MSC-EVs reduce cartilage degradation by miR-376c-3p targeting WNT3/WNT9a and inhibiting WNT/β catenin |
In vitro and in vivo models |
Human |
Exosome Isolation from hADSCs, Bioinformatics, Luciferase Reporter Assay, Safranine O-Fast Green Cartilage Staining, Masson Staining, Immunohistochemistry, Immunofluorescence |
[14] |
|
|
H-ApoEVs from hypoxia-treated adipose MSCs enhance proliferation and M2 macrophage polarization, improving cartilage regeneration with scaffolds |
3D-printed ECM scaffold in vitro and in vivo |
Human |
ApoEVs and H-ApoEVs isolated from adipose MSCs under normoxic and hypoxic conditions, characterized by TEM, DLS, Western Blotting, NTA, and DiO-labeling; injection via Gel/ECM scaffold in the knee joint cavity |
[15] |
|
|
Combined IA therapy in sheep reduced lameness and improved OA scores |
Meniscectomy Sheep Model |
Sheep |
Exosomes isolated from Ad-MSCs, characterized by flow cytometry (CD63, CD81 positive), particle size (88.7 nm), zeta potential (-1.4 mV), and ultracentrifugation; injection via IA route at specified intervals (HA and exosomes) |
[16] |
|
|
Umbilical Cord MSC-EVs |
hUCMSC-EVs boost BMSC proliferation and chondrogenesis via miR-181c-5p, suppressing SMAD7 and enhancing BMP2 signaling |
In vivo cartilage repair model |
Human |
EVs isolated by ultracentrifugation from hUCMSCs; characterized by TEM, NTA, and Western blot (CD9, CD81, CD63); miR-181c-5p delivered via EVs to BMSCs |
[17] |
|
miR-7704-modified EVs promote collagen II and reduce MMP13, preserving cartilage and motor function in OA mouse model |
Mouse OA model |
Human |
EVs were isolated from HUCMSCs cultured in serum-free medium using ExoQuick. Characterization by Western blot (CD9, CD63, CD81), NTA, and TEM. Injection of EVs (1×10⁷ particles/mL) into OA mice via intra-articular route |
[18] |
|
|
hWJMSC-EVs deliver ITGB1 to activate TGF-β/Smad2/3 pathway, enhancing cartilage repair after microfracture |
Rabbit model |
Human |
In vitro: Coculture with hBMSCs and chondrocytes treated with hWJMSC-EVs at a concentration of 50 µg/ml. In vivo: Injection of hWJMSC-EVs at various concentrations into rabbit joints post-MF surgery. Evaluation via micro-CT, histology, immunohistochemistry, and Western blot |
[19] |
|
|
Combining PRP with hUCMSC-EVs improves cartilage regeneration and motor function by increasing collagen II and reducing IL-1β |
Mouse OA model |
Human |
In vivo: EVs extracted from HUCMSCs cultured with 5% PRP were injected into the OA mice model. In vitro: Differentiation and characterization of HUCMSCs treated with PRP. Evaluation of motor function, cartilage integrity, and protein expression (type II collagen, aggrecan, IL-1β) using rotarod test, histology, ICRS scoring, and immunohistochemistry |
[20] |
|
|
3D-culture-derived exosomes had better biological activity and stability than 2D-derived exosomes |
Chondrocytes in vitro |
Human |
Isolation: U-MSCs were cultured in a hollow-fiber bioreactor, conditioned medium collected, and exosomes purified via ultracentrifugation. Characterization: Transmission electron microscopy for cup-shaped morphology; Western blot for exosome markers (CD63, CD81, TSG101); Nanosight analysis for particle size (~120 nm). Dosage: 10 μg/ml of exosomes used in in vitro assays; 500 μl of exosome suspension (1 × 10¹⁰ particles/ml) for in vivo injections. Injection: Exosomes were injected IA into the rabbit knee joints with cartilage defects |
[21] |
Contd. Table 1. Summary of in vivo and in vitro studies using MSC-derived EVs for OA and cartilage repair
|
EV source |
Therapeutic effect |
Model used |
Origin |
Method (EVs isolation, characterization, dosage, etc.) |
Ref. |
|
|
|
|
|
|
|
|
Synovial MSC-EVs |
Sy-MSC-EVs deliver miR-26a-5p to reduce inflammation and apoptosis, targeting PTEN |
Chondrocytes and OA mouse model |
Human |
Isolation: not specified |
[22] |
|
Sy-MSC-EVs promote chondrocyte proliferation and reduce apoptosis and inflammation via miR-130b-3p targeting LRP12 and AKT/β-catenin pathway |
Chondrocytes and OA mouse model |
Human |
Isolation: EVs were isolated from SMSCs and identified. |
[23] |
|
|
Exosomes reversed IL-β-induced downregulation of COL2A1 and ACAN, inhibited MMP13/ADAMTS5, and corrected autophagy dysfunction |
Chondrocytes in vitro |
Human |
Isolation using Ribo™ Exo isolation reagent from culture supernatants of SMSCs grown in Exo-free medium for 72 hours. After centrifugation at 2000 g and 1500 g, the exosome pellets were resuspended in PBS. Exosome characterization was done by TEM and DLS. Western blot confirmed the presence of exosome markers CD63 and HSP70 and the absence of Calnexin. In vitro, exosomes were applied to IL-1β-induced mouse chondrocytes at a concentration of 10 µg/ml for 24 hours. For in vivo studies, 10 µl of SMSC-Exo (100 µg/ml) was injected into the knee joints of OA mice for three consecutive weeks |
[24] |
|
|
miR-140-5p-enriched SMSC-Exos enhanced chondrocyte proliferation, migration, and ECM maintenance |
Chondrocytes in vitro, Rat in vivo |
Human |
Isolation after miR-140-5p transfection using ultracentrifugation. Exosomes were characterized by DLS, TEM, and Western blot analysis, identifying key markers such as CD63, CD81, and Alix. In vitro, exosome treatment was performed on articular chondrocytes (ACs) to assess proliferation, migration, and ECM secretion. In vivo cartilage OA rat protection was assessed through histology, Safranin-O and Fast Green staining, and immunohistochemical analysis |
[25] |
|
|
Reduced ECM breakdown and cartilage degradation in the OA rat model |
Rat OA model |
Rat |
(SMSC-Exo) isolated via ultracentrifugation. Exosome Characterization via TEM, DLS, and Western blot |
[24] |
|
|
Dental Follicle Cell EVs |
DFC-sEVs improved TMJ chondrocyte activity and protected cartilage by downregulating HIF-1α/2α and catabolic markers |
Rat TMJ-OA model |
Human |
DFC-sEV were isolated from DFC using ultrafiltration and characterized by TEM, NTA, and Western blot for surface markers (CD9, CD81, HSP90, TSG101). The size range was 50-200 nm. For in vivo treatment, 50-100 μg/ml DFC-sEV was injected intra-articularly into TMJ-OA rats for 4-8 weeks, showing therapeutic effects |
[26] |
|
Embryonic Stem Cell EVs |
ESC-MSC-EVs reduce cartilage damage and matrix degradation by upregulating COL II and suppressing ADAMTS5 |
DMM OA mouse model |
Mouse |
Exosome isolation from ESC-MSCs |
[27] |
|
ESC-MSCs and their EVs improved cognition and mobility in older dogs without adverse effects |
Dog |
Human |
Isolation: TFF, concentrated ~10x. |
[28] |
|
|
iPSC-EVs |
IPSC-EVs showed higher effectiveness than SM-MSC-EVs in reducing joint damage and enhancing chondrocyte repair |
In vitro and in vivo OA models |
Human |
Isolation: Ultrafiltration method, characterized by Tunable Resistive Pulse Sensing (TRPS), TEM, Western blot (CD9, CD63, TSG101). |
[29] |
|
IPSC-EVs suppress T cells and enhance M2 macrophage polarization, similar to the regenerative impact of hUCMSC-EVs |
In vitro comparative study |
Human |
Isolation: Conditioned medium from iMSCs, differential centrifugation, TFF |
[30] |
|
|
MSC-Exos + HA |
Improved cartilage and bone regeneration in porcine OA model; safe and effective |
Porcine osteochondral defect |
Human |
Isolation: Exosomes isolated from MSCs. Characterization: MRI, histology, biomechanical tests, and micro-CT. Dosage: 2 ml IA injection on day 0, 8, and 15 postsurgery |
[31] |
Contd. Table 1. Summary of in vivo and in vitro studies using MSC-derived EVs for OA and cartilage repair
|
EV source |
Therapeutic effect |
Model used |
Origin |
Method (EVs isolation, characterization, dosage, etc.) |
Ref. |
|
Amniotic Fluid Stem Cell EVs |
Induced hyaline-like cartilage, improved pain, and modulated macrophages via TGF-β |
MIA-induced OA model |
Human |
Exosome Isolation: Commercial kit. Characterization: Identified markers, HGF, TGF-β, IDO. Dosage: Exo injection into the knee joints of animals, followed by histology and behavioral scoring up to 3 weeks |
[32] |
|
Improved joint space and pain in OA and spine degeneration patients |
OA & spinal degeneration |
Human |
Isolation, Characterization: not specified Dosage: 4 trillion exosomes in 1.5 ml matrix (Case 1 and Case 3: half IA + half IV; Case 2: all IV) |
[33] |
|
|
hPESC-derived EVs |
Reduced pain and improved MRI cartilage quality in OA patients |
Grade II/III knee OA patients |
Human |
Isolation: Conditioned medium (CM) from hPESCs. Characterization: GFs, CKs, and EVs. Dosage: 2 ml IA injection per patient |
[34] |
Abbreviations: BM-MSC: Bone marrow–derived mesenchymal stem cell, ADSC: Adipose-derived stem cell, UC-MSC: Umbilical cord–derived mesenchymal stem cell, WJMSC: Wharton’s jelly–derived mesenchymal stem cell, Sy-MSC: Synovial-derived mesenchymal stem cell, BMSC: Bone marrow stem cell, SMSC: Synovium-derived mesenchymal stem cell, DFC: Dental follicle cell, ESC: Embryonic stem cell, iPSC: Induced pluripotent stem cell, hPESC: Human progenitor endothelial stem cell, EV: Extracellular vesicle, sEV: Small extracellular vesicle, Exo: Exosome, MPs: Microparticles, ECM: Extracellular matrix, HA: Hyaluronic acid, DRG: Dorsal root ganglion, OA: Osteoarthritis, LBP: Low back pain, LFJ: Lumbar facet joint, CIOA: Collagenase-induced OA, LIPUS: Low-intensity pulsed ultrasound, HGF: Hepatocyte growth factor, TGF-β: Transforming growth factor beta, BMP2: Bone morphogenetic protein 2, MMP: Matrix metalloproteinase, COL2: Type II collagen, ACAN: Aggrecan, miR: MicroRNA, NTA: Nanoparticle tracking analysis, TEM: Transmission electron microscopy, DLS: Dynamic light scattering, TFF: Tangential flow filtration, IV: Intravenous, IA: Intra-articular, PRP: Platelet-rich plasma, MF: Microfracture.
EVs from synovial cells, including Cartilage Progenitor Cell-derived EVs (CPC-EVs), protect IL-1β-stimulated chondrocytes by promoting cartilage matrix production and reducing inflammation, mainly by inhibiting STAT3 activation. Proteomic analysis revealed that CPC-EVs contain STAT3 regulatory proteins that are transferable to chondrocytes. To enhance therapeutic effectiveness, CPC-EVs were modified with a cationic peptide. The modified EVs significantly reduced OA while assisting in matrix repair in ex vivo cartilage explants 23.
Beyond mammalian cell-derived EVs, various emerging nontraditional EV sources are currently being investigated for regenerative and anti-inflammatory applications. For instance, plant-derived EVs are safe, natural, and eco-friendly nanocarriers with significant potential in regenerative medicine. They offer advantages over synthetic nanoparticles, including lower toxicity and enhanced cellular uptake. The anti-inflammatory and immunomodulatory effects of these vesicles make them promising therapeutic agents. For example, EVs derived from grapefruit juice 21, tomatoes, and lemons 24 have demonstrated potential in treating OA.
EVs from Antler Blastema (deer) Progenitor Cells (ABPC-EVs) demonstrate emerging regenerative potential. EVs from deer stem cells have been shown to slow the aging of human mesenchymal stem cells and enhance cartilage repair. Treatment with these EVs reversed senescence-related changes in MSCs and modulated the expression of aging-associated genes, suggesting their potential in regenerative and antiaging therapies 25.
Bacterial Extracellular Vesicles (BEVs) have recently emerged as promising nanocarriers for OA therapy, primarily owing to their unique ability to modulate the gut–joint axis. Owing to their nanoscale size, low toxicity, high drug-loading capacity, and excellent biocompatibility, BEVs can efficiently deliver bioactive molecules to regulate both the intestinal and joint microenvironments, opening new opportunities for OA treatment 26.
Source selection is one of the most important determinants of EV reproducibility and translational feasibility. MSC-derived EVs are the most widely investigated because of their immunomodulatory and chondroprotective properties, but their composition varies according to tissue origin, donor age, health status, passage number, and culture conditions. Chondrocyte- and cartilage progenitor cell-derived EVs may provide more cartilage-specific signals, yet their scalability and donor availability remain limited. Platelet-Rich Plasma (PRP)-derived EVs are clinically attractive, because PRP is already used in orthopedic practice, but their composition is highly patient-dependent and may vary with platelet preparation protocols. Plant- or food-derived EV-like nanoparticles may offer advantages in scalability and oral delivery, but their mechanisms, molecular identity, and regulatory classification remain less defined. Therefore, the optimal EV source for OA therapy should not be selected only according to regenerative potency, but also according to reproducibility, safety, scalability, manufacturing feasibility, and compatibility with regulatory quality-control requirements.
Engineering Strategies for EV-Based Cartilage Repair and Artificial EV Platforms
Although native EVs exhibit considerable regenerative potential, their therapeutic application remains limited by insufficient targeting specificity, rapid clearance, restricted cargo-loading capacity, and biological heterogeneity 27. To address these limitations, multiple engineering strategies have been developed to enhance EV stability, targeting efficiency, cargo delivery, and regenerative performance. These approaches generally include donor-cell engineering, direct surface modification, cargo-loading techniques, and the development of artificial or hybrid EV platforms 28.
Donor-cell engineering represents an indirect strategy in which parental cells are genetically or environmentally modified to alter EV cargo composition and biological activity. One study increased the number of EVs from human urine-derived stem cells via the overexpression of miR-140. Compared with standard EVs, the engineered version more effectively improved chondrocyte function and increased the production of key cartilage matrix proteins by targeting Vascular Endothelial Growth Factor type A (VEGFA). These EVs promote cartilage repair and subchondral bone regeneration, highlighting their therapeutic potential 29.
Hypoxic preconditioning of MSCs is a promising method for enhancing the therapeutic potential of their EVs in OA models. Hypoxic-derived EVs (Hypo-sEVs) are more effective than normoxic sEVs in enhancing chondrocyte growth, migration, and survival. This improved effect is associated with increased levels of miR-216a-5p in Hypo-sEVs, which regulate the JAK2/STAT3 pathway involved in cartilage health. The role of HIF-1α in this process highlights hypoxic preconditioning as a powerful, nongenetic method to increase EV bioactivity 30.
Table 2. Representative EV cargos and their mechanisms in promoting cartilage repair
|
Cargo type |
RNA/Protein |
EV Source |
Target/Pathway |
Therapeutic Effect |
Ref. |
|
miRNA |
miR-206 |
BMSC-derived exosomes |
ELF3 |
Enhances osteoblast proliferation, reduces inflammation, and apoptosis |
[1] |
|
miR-148a |
Milk-derived EVs |
Catabolic and inflammatory pathways |
Protects cartilage by maintaining homeostasis |
[2] |
|
|
miR-214-3p |
SFB-derived exosomes |
NF-κB signaling pathway |
Reduces inflammation and apoptosis, protects cartilage |
[3] |
|
|
miR-9-5p |
Bone marrow MSC-derived exosomes |
Syndecan-1 (SDC1) |
Reduces inflammation, oxidative stress, and cartilage degradation |
[4] |
|
|
miR-127-3p |
MSC-derived exosomes |
CDH11, Wnt/β-catenin pathway |
Suppresses cartilage degradation and joint inflammation |
[5] |
|
|
miR-31 |
Synovial MSC-derived exosomes (SMSC-EVs) |
KDM2A/E2F1/PTTG1 axis |
Enhances chondrocyte function and alleviates cartilage damage and inflammation |
[6] |
|
|
miR-136-5p |
MSC-derived exosomes |
ELF3 |
Protects against chondrocyte degeneration after joint injury |
[7] |
|
|
miR-100-5p |
Infrapatellar fat pad MSC-derived exosomes |
mTOR signaling pathway |
Protects cartilage and improves gait in OA models |
[8] |
|
|
miR-129-5p |
Human synovial MSC-derived exosomes |
HMGB1 |
Reduces inflammation and apoptosis in chondrocytes |
[9] |
|
|
miR-130b-3p |
EVs (unspecified) |
LRP12/AKT/β-catenin signaling |
Protects against inflammation and ECM degradation in chondrocytes |
[10] |
|
|
miR-181c-5p |
Bone marrow MSCs |
SMAD7 |
Promotes BMP2-induced cartilage repair |
[11] |
|
|
miR-135b |
MSC-derived exosomes |
Sp1 transcription factor |
Enhances chondrocyte proliferation |
[12] |
|
|
miR-92a-3p |
MSC-derived exosomes (overexpressing) |
WNT5A |
Enhances chondrogenesis and reduces cartilage degradation |
[13] |
|
|
miR-150-3p |
H–FLS–derived EVs |
Trim14/NF-κB/IFNβ axis |
Maintains joint homeostasis and reduces inflammation |
[14] |
|
|
miR-320 |
MSC-derived exosomes |
MMP-13, IL-1β |
Regulates ECM remodeling and inflammation in chondrogenesis |
[15] |
|
|
miR-140-5p |
Human synovial MSC-derived exosomes (overexpressing) |
Not specified |
Promotes cartilage repair and delays OA progression |
[16] |
|
|
miR-143 |
Exosomes from curcumin-treated MSCs |
NF-κB, ROCK1 |
Enhances chondrocyte viability, reduces apoptosis, and restores miR-143 expression |
[17] |
|
|
miR-124 |
Exosomes from curcumin-treated MSCs |
NF-κB, ROCK1 |
Restores miR-124 expression, suppresses inflammation, and cartilage degradation |
[18] |
|
|
circRNA |
circRNA_0001236 |
Exosomes |
Sox9, MMP13 via miR-3677-3p |
Facilitates chondrogenesis and alleviates OA symptoms |
[19] |
|
circSERPINE2 |
Exosomes |
miR-495/TGFBR2 pathway |
Suppresses apoptosis and inflammation in OA |
[20] |
|
|
hsa_circ_0005567 |
Exosomes |
miR-495/ATG14 axis |
Restores autophagy and protects cartilage |
[21] |
|
|
circHIPK3 |
Exosomes |
miR-124/SOX8 |
Regulates chondrocyte behavior |
[22] |
|
|
lncRNA |
PCGEM1 |
Synovial fluid-derived exosomes |
Not specified |
Biomarker and stage indicator in OA |
[23] |
|
KLF3-AS1 |
hMSC-derived exosomes |
KLF3-AS1/miR-206/GIT1 axis |
Reduces chondrocyte apoptosis and promotes cartilage repair |
[24] |
|
|
TRAF1-4:1 |
RA-FLS-derived exosomes |
miR-27a-3p/CXCL1 |
Impairs cartilage by ECM degradation and inflammation |
[25] |
|
|
H19 |
Umbilical cord MSC-derived exosomes |
miR-29b-3p/FoxO3 |
Enhances osteochondral regeneration and reduces apoptosis |
[26] |
|
|
H19 |
Fibroblast-like synoviocyte-derived exosomes |
miR-106b-5p/TIMP2 |
Mitigates OA progression |
[27] |
Contd. Table 2. Representative EV cargos and their mechanisms in promoting cartilage repair
|
Cargo type |
RNA/Protein |
EV Source |
Target/Pathway |
Therapeutic Effect |
Ref. |
|
Protein |
scSOX9 |
Cell-penetrating recombinant protein |
Not specified |
Promotes hyaline-like cartilage formation in a rabbit model |
[28] |
|
FGF18 (via CRISPR activation) |
Targeted exosomes (CAP/FGF18-hyEXO) |
FGF18 gene activation |
Enhances cartilage regeneration, reduces inflammation, and ECM degradation |
[29] |
|
|
Cas9 |
Chondrocyte-targeted exosomes fused with liposomes (CAP-Exo) |
MMP-13 |
Reduces cartilage breakdown via gene editing |
[30] |
|
|
MBPs |
MSC-derived exosomes |
Not specified |
Support cartilage repair and attenuate inflammation in osteoarthritic joints |
[31] |
|
|
Synthetic molecules and drugs |
Dexamethasone sodium phosphate |
Folic acid–modified exosomes (FPC-Exo/Dex) |
Inflamed joints in RA |
Anti-inflammatory, preserves bone/cartilage in CIA model |
[32] |
|
Kartogenin |
E7-peptide engineered exosomes (E7-Exo) |
MSC targeting, enhanced KGN intracellular dispersion |
Promotes cartilage regeneration in OA |
[33] |
|
|
Curcumin |
Exosomes from curcumin-treated MSCs |
NF-κB and ROCK1 (via miR-143 and miR-124) |
Enhances chondrocyte viability and reduces apoptosis in OA models |
[17] |
Abbreviations: BMSC: Bone marrow–derived mesenchymal stem cell, MSC: Mesenchymal stem cell, SMSC: Synovial mesenchymal stem cell, hMSC: Human mesenchymal stem cell, UC-MSC: Umbilical cord–derived mesenchymal stem cell, FLS: Fibroblast-like synoviocyte, RA-FLS: Rheumatoid arthritis fibroblast-like synoviocyte, H–FLS: Human fibroblast-like synoviocyte, EV: Extracellular vesicle, Exo: Exosome, OA: Osteoarthritis, RA: Rheumatoid arthritis, ECM: Extracellular matrix, BMP2: Bone morphogenetic protein 2, MMP: Matrix metalloproteinase, NF-κB: Nuclear factor kappa-light-chain-enhancer of activated B cells, ROCK1: Rho-associated protein kinase 1, CAP: Cartilage-affinity peptide, CAP-Exo: Chondrocyte-targeted exosome, CAP/FGF18-hyEXO: Cartilage-targeted exosome engineered for FGF18 activation, CIA: Collagen-induced arthritis, scSOX9: Single-chain SOX9 protein, MBPs: Matrix-binding proteins, FPC-Exo/Dex: Folic acid–modified exosome loaded with dexamethasone, E7-Exo: E7-peptide–engineered exosome, KGN: Kartogenin, miR: MicroRNA, circRNA: Circular RNA, lncRNA: Long non-coding RNA, Sox9: SRY-box transcription factor 9, SMAD7: Mothers against decapentaplegic homolog 7, HMGB1: High mobility group box 1, ATG14: Autophagy-related protein 14, TGFBR2: Transforming growth factor beta receptor 2, FoxO3: Fork-head box O3, TIMP2: Tissue inhibitor of metalloproteinases 2, GIT1: G-protein-coupled receptor kinase-interacting protein 1, IFNβ: Interferon beta, LRP12: Low-density lipoprotein receptor-related protein 12, SDC1: Syndecan-1, ELF3: E74-like ETS transcription factor 3, mTOR: Mechanistic target of rapamycin.
In parallel, therapeutic enhancement can also be achieved through incorporation of bioactive molecules or drugs into EV systems to improve regenerative and anti-inflammatory efficacy. Another innovative approach combines curcumin with small EV (sEVs) from adipose-derived MSCs to create sEV-CUR, a hybrid therapy. Compared with free curcumin or unmodified sEVs, sEV-CUR had more substantial protective effects on OA models. It enhances chondrocyte survival under oxidative stress and, in mice, reduces joint damage and cell death, highlighting its potential for cartilage protection and regeneration 31.
Beyond natural EV modification, artificial and hybrid EV platforms have emerged as alternative strategies designed to improve scalability, targeting precision, and cargo-loading flexibility. While exogenous Hyaluronic Acid (HA) injections relieve pain in OA patients, their effects are short-lived. A new approach utilizes engineered, pH-responsive EVs that carry Hyaluronan Synthase 2 (HAS2) to reprogram chondrocytes to produce high-molecular-weight HA internally 32.
Although PRP is widely used for healing due to its growth factors, PRP-EVs may promote angiogenesis during tissue repair, thereby supporting nutrient and oxygen delivery to injured areas and facilitating waste removal. In orthopedic conditions such as knee OA, PRP-derived EVs show therapeutic potential. They increase chondrocyte proliferation and reduce apoptosis by activating the Wnt/β-catenin pathway, which is essential for joint stability 33.
Another recent study examined the limitations of EVs in treating temporomandibular joint OA, suggesting that Artificial Cell-Derived Vesicles (ACDVs) may offer a superior alternative. When created from stem cells via extrusion techniques, ACDVs resemble EVs in structure but provide a greater yield of particles and proteins 34.
A stable hybrid system that integrates EVs and lipid nanoparticles through ethanol-mediated fusion retains the bioactive properties and targeting capabilities of both elements. The final hybrid particle contained nicotinamide-loaded, Col2A1 antibody-modified liposomes alongside TGF-β1-overexpressing EVs. This hybrid particle exhibited significant cartilage-protective and anti-inflammatory properties, featuring precise targeting and prolonged retention at OA sites 35.
Surface Modification for Targeted Delivery
One major strategy for improving EV therapeutic efficacy involves surface functionalization with targeting ligands that enhance tissue-specific delivery 36. This strategy is particularly important in cartilage-targeted therapies, where natural EVs face a significant delivery barrier due to their negatively charged membrane. This leads to electrostatic repulsion from the anionic cartilage matrix, thereby reducing its retention and uptake 37.
In response, charge-reversal engineering strategies have been developed through the surface decoration of EVs with positively charged motifs of avidin or peptide carriers rich in arginine. These surface-modified EVs were then loaded with Interleukin-1 Receptor Antagonist (IL-1RA), a disease-modifying agent for OA. These EVs exhibited increased retention in joints and enhanced therapeutic delivery, indicating an essential advantage in penetrating cartilage barriers 38. In another strategy, bio-conjugated EVs with a Cartilage-binding Peptide (CAP) are loaded with MMP13-targeting siRNA, an enzyme responsible for cartilage matrix degradation. This engineered delivery system, known as CAP-Exo/siMMP13, effectively targeted chondrocytes, inhibited MMP13 expression, and induced the expression of cartilage-specific markers in an OA rat model. This outcome further underscores the potential of EV-mediated gene silencing for the safe and therapeutic treatment of OA 39. Another study used a collagen-binding domain peptide attached to the exosomal membrane protein LAMP-2B to bind miR-21-loaded EVs to collagen-I scaffolds. This helps improve the local retention of vesicles at sites of cartilage lesions, enhancing therapeutic efficacy 40.
Furthermore, dendritic cell-derived EVs tend to be distributed throughout the body, with notable accumulation in the kidney. To enhance site-specific delivery, a chondrocyte-affinity peptide that effectively delivers miR-140 into cartilage was fused to the EV surface, significantly improving its retention within the articular cavity. This modification not only confines the EVs to the joint space, but also promotes their penetration into cartilage, enabling targeted delivery to chondrocytes and thereby enhancing therapeutic efficacy 41. One study incorporated a glycosylation motif at the N-terminus of the peptide–LAMP–2B construct to increase its stability. Additionally, a peptide was introduced for targeted delivery to Synovial Fluid-derived Mesenchymal Stem Cells (SF-MSCs). EVs engineered in this way and produced from dendritic cells were subsequently loaded with kartogenin. When applied in vivo and in vitro, these engineered EVs selectively targeted SF-MSCs and significantly increased the differentiation of MSCs into chondrocytes and promoted cartilage regeneration 42.
Compared with unmodified EVs, surface-engineered vesicles generally demonstrate enhanced cartilage targeting, prolonged intra-articular retention, and improved therapeutic efficacy. However, these modifications may increase manufacturing complexity, production costs, and potential immunogenicity, which remain important considerations for large-scale clinical translation.
Loading of Different Cargo Types for Enhanced Therapeutic Efficacy
Among the molecules delivered by EVs, RNA-based therapeutics, such as microRNAs (miRNAs), small interfering RNAs (siRNAs), and long noncoding RNAs (lncRNAs), are significant 43. Additionally, short hairpin RNAs (shRNAs) and circular RNAs (circRNAs) have gained considerable attention for their ability to modulate gene expression and their potential application in treating a wide range of medical conditions (Table 2) 44.
miRNAs are small, noncoding, single-stranded RNAs, typically 19–24 nucleotides in length, that regulate gene expression at the post-transcriptional level. They are essential for normal physiological development and are implicated in various biological processes 45. miRNAs originate from precursor molecules that form a characteristic hairpin-shaped loop structure 46. Circular RNAs (circRNAs) are conserved, single-stranded RNA molecules with a closed-loop structure. Found across many species, they play essential roles in gene regulation by acting as transcriptional regulators, microRNA sponges, and, occasionally, templates for protein synthesis 47. Like those of microRNAs, the expression patterns of specific circular RNAs (circRNAs) have also been associated with the progression of OA 48. Long noncoding RNAs (lncRNAs) represent a significant class of RNA transcripts that are longer than 200 nucleotides and do not encode proteins. Although over 18,000 lncRNAs have been discovered, only a small number of them have established functions. Characterized lncRNAs participate in various regulatory processes, including gene expression, chromatin remodeling, and protein activity. Importantly, they demonstrate high specificity to cell types and play a crucial role in immune regulation 49.
Proteins, in addition to recombinant protein drugs, have good therapeutic potential because they adjust inflammatory pathways, balance cartilage catabolism and anabolism, and support tissue repair. However, their rapid clearance and short intra-articular half-life pose significant barriers to their clinical application, highlighting the need for effective and sustained drug delivery systems to improve therapeutic outcomes 50. Advancements in genetic engineering have expanded the use of protein-based therapeutics, which are valued for their high efficacy, specificity, and low toxicity. However, their immunogenicity can limit their clinical application, which may trigger unwanted immune responses and allergic reactions 51. In addition to the other cargoes discussed previously, EVs can be engineered to carry synthetic small molecules and common related drugs, offering a versatile platform for targeted drug delivery in regenerative medicine.
Different loading strategies offer distinct advantages. Passive loading approaches are technically simple and better preserve EV membrane integrity; however, they often exhibit poor encapsulation efficiency and limited control over cargo concentration. In contrast, active loading methods provide greater loading efficiency and cargo control but may compromise EV structural stability and biological activity. Therefore, selecting an appropriate loading strategy requires balancing loading efficiency, vesicle integrity, cargo type, and scalability requirements.
Overall, these studies demonstrate that engineering strategies can significantly enhance EV targeting, cargo delivery, and therapeutic efficacy, although differences in methodology and lack of standardization remain important challenges.
Engineered EVs exert their therapeutic effects through coordinated modulation of key signaling pathways involved in cartilage homeostasis. For instance, EV-mediated delivery of miRNAs and proteins frequently targets pathways such as JAK/STAT, Wnt/β-catenin, and TGF-β/Smad, which regulate chondrocyte proliferation, apoptosis, and extracellular matrix synthesis. Furthermore, suppression of inflammatory mediators such as IL-1β and TNF-α appears to be a central mechanism through which EVs attenuate cartilage degradation. These findings highlight that the therapeutic efficacy of engineered EVs is not merely dependent on cargo delivery, but on their ability to reprogram cellular signaling networks toward a regenerative phenotype.
EV Loading, Cargo Release, and Cellular Uptake
Therapeutic agents can be loaded into EVs through either active or passive encapsulation methods, each providing different levels of efficiency and cargo stability within the vesicles (Figure 2). Passive cargo-loading techniques provide a straightforward approach for incorporating therapeutic agents into EVs. This process involves culturing cells with the drug of interest, allowing the cells to naturally take up the drug, and then integrating it into the resulting EVs. The resulting drug-loaded EVs are then released into the culture medium. However, it usually results in low loading efficiency and offers limited control over the amount of drug packaged into the EVs. It tends to be more effective for lipophilic compounds that readily cross cellular and vesicular membranes. Nonetheless, this method may not be suitable for all therapeutic agents 52. An alternative to passive loading is active loading, where drugs are introduced directly into pre-isolated EVs. This approach typically involves techniques that temporarily increase EV membrane permeability, including sonication, heat shock, electroporation, detergent incubation, or the use of saponins. Active loading gen-
erally increases drug encapsulation efficiency and offers more precise control over the amount of drug incorporated. However, it can be technically demanding and time-consuming, and improper handling may compromise the structural integrity of the EVs. Regardless of the loading method, removing unincorporated drugs is essential and is usually achieved through purification techniques such as ultracentrifugation, size exclusion chromatography, or ultrafiltration 53. Additionally, genetic engineering provides a strategy for incorporating specific genetic materials into EVs. This involves modifying the donor cells to express the desired RNAs or proteins, which are then packaged into EVs during biogenesis. Although this technique allows the targeted delivery of nucleic acids or proteins, it relies on advanced molecular biology tools and may not be practical for all therapeutic applications. Each approach presents advantages and limitations, and selecting the most suitable method depends on variables such as the drug’s properties, the EV source, and the therapeutic target 52.
Strategies for loading bioactive cargo into EVs can be broadly classified into pre-loading (endogenous) and post-loading (exogenous) approaches. In the pre-loading approach, parental cells are exposed to therapeutic molecules (e.g., proteins, RNAs, plasmids, drugs), which are naturally incorporated into EVs during biogenesis and subsequently secreted. In the post-loading approach, isolated EVs are engineered to carry cargoes through techniques such as transfection, electroporation, incubation, sonication, freeze–thaw cycling, extrusion, detergent permeabilization using agents such as saponin, or fusion with liposomes to form chimeric EVs. These complementary strategies enhance the versatility of EVs as delivery vehicles for nucleic acids, proteins, and small molecules in both experimental and therapeutic applications.
Consequently, the efficient therapeutic use of EVs depends on effectively loading functional molecules into vesicles and delivering them precisely to target cells. To increase the bioavailability and performance of EV-based treatments, cargo-loading strategies and cellular uptake mechanisms must be optimized to maximize their effectiveness. Once released, EVs can interact with recipient cells through several pathways, including endocytosis, direct fusion with the plasma membrane, or ligand and receptor interactions that trigger cell-to-cell signaling. Endocytosis is the most common uptake mechanism and is mediated by various routes, such as clathrin-dependent and caveolin-dependent pathways, macropinocytosis, and phagocytosis 54. A fusion-based mechanism enables the rapid delivery of exosomal contents into the cytoplasm and is considered more efficient than the endocytic route 55. Nonetheless, via the uptake route, EV cargo must ultimately be released into the cytoplasm to perform its biological functions. Otherwise, cargoes delivered through endocytosis can be degraded in lysosomes or removed from the cell, thus diminishing therapeutic efficacy 56. Finally, while EV‒cell interactions depend primarily on the surface features of both vesicles and recipient cells, the exact nature and specificity of these interactions remain a subject of ongoing investigation. It is still debated whether EV uptake is a directed process or occurs randomly 57.
In the studies reviewed, dosing strategies for EV-based therapies vary widely and are often determined based on preliminary cytotoxicity assays or prior literature. However, the lack of standardized dose–response frameworks and inconsistent reporting of IC₅₀ or therapeutic thresholds remains a major limitation in the field. Establishing dose standardization will be essential for improving reproducibility and clinical translation.
2D and 3D culture
The biological activity and therapeutic efficiency of EVs are strongly influenced by the microenvironment in which parental cells are cultured. Conventional two-Dimensional (2D) culture systems fail to precisely replicate the native cartilage microenvironment, whereas three-Dimensional (3D) culture platforms better mimic physiological cell–matrix interactions, and mechanical stimuli. Consequently, increasing attention has focused on 3D culture systems and biomaterial-based scaffolds to enhance EV production and regenerative potency.
A recent study revealed that cryogels made from extracellular matrix components such as gelatin, chondroitin sulfate, and hyaluronic acid can promote cartilage repair, particularly when combined with EVs from BMSCs. Among the formulations tested, cryogels containing 0.3% HA provided the optimal balance of strength and cell compatibility. Laboratory and animal studies confirmed that EVs and EV-infused cryogels enhanced cartilage matrix production. However, EV-loaded cryogels result in the most organized tissue structure and effective regeneration, positioning them as promising tools for cartilage repair 58. Additionally, in another study, a bilayer cryogel scaffold featuring a cartilage-like layer (chitosan-gelatin-chondroitin sulfate) and a bone-like layer (nanohydroxyapatite-gelatin) was developed to repair osteochondral defects. Chondrocyte-derived EVs enhanced regeneration, with approximately 80% released within 72 hr, while chondroitin sulfate was gradually released over a week. The combination of EVs and scaffold extract improved chondrocyte growth and migration in vitro, indicating potential for osteochondral tissue repair 59.
A study examined umbilical-derived EVs under conventional 2D culture and 3D microgravity conditions for cartilage repair. Both types of EVs promoted chondrocyte proliferation, migration, and matrix synthesis and reduced apoptosis. However, EVs derived from 3D culture exhibited superior biological activity, higher yields, and an enhanced ability to maintain chondrocyte phenotypic stability. This effect was at tributed mainly to improved regulation of matrix homeostasis through the TGF-β1/Smad2/3 signaling pathway, suggesting that 3D-derived U-MSC-EVs may offer greater therapeutic efficiency in cartilage regeneration (Figure 3) 60. Recent advances in the 3D bioprinting of hydrogels, especially those loaded with U-MSC-EVs, have shown potential for improving tissue regeneration and wound healing. The combination of EVs with 3D bioprinted hydrogels (3D-BPH-Exos) provides a novel platform for OA treatment by reducing the levels of inflammatory mediators such as IL-1β, TNF-α, and IFN-γ; preserving mitochondrial function; and enhancing chondrogenesis in joint-resident cells. Loading these EVs with chondrogenic agents, such as kartogenin, may offer a more effective, localized, and safer alternative to conventional anti-inflammatory or cell-based therapies. This emerging strategy emphasizes the potential of 3D-BPH-Exos as a next-generation immunomodulatory and regenerative approach for OA management 61. A different study revealed that culturing U-MSCs in a 3D environment with human platelet lysate significantly enhances the production and therapeutic efficacy of 3D-EVs. Compared with 2D-derived EVs, 3D-EVs promote the proliferation of MSCs and chondrocytes, suppress inflammatory and catabolic responses, and improve cartilage repair in vivo when administered through a hydrogel. The regenerative benefits were associated with increased synthesis of type II collagen and decreased expression of MMP13, underscoring 3D-EVs as a promising approach for effective EV production and cartilage regeneration 62.
Translational challenges, regulatory considerations, and future perspectives
While engineered EVs have established promising regenerative and immunomodulatory effects in preclinical OA models, numerous biological, manufacturing, regulatory, and economic barriers continue to limit their clinical translation. The effective integration of EV-based therapies into orthopedic practice will necessitate scalable manufacturing systems, and clear regulatory frameworks. Additionally, patient heterogeneity and disease-stage variability require important considerations that may significantly influence therapeutic sensitivity.
An important challenge involves the biological heterogeneity of OA itself. OA is increasingly recognized as a multifactorial and heterogeneous disease with distinct inflammatory, metabolic, traumatic, and age-associated phenotypes. Consequently, therapeutic responsiveness to EV-based interventions may differ significantly among patient populations. Recent evidence suggests that early-stage or moderate OA, where viable chondrocytes and partially preserved cartilage architecture remain present, is more likely to benefit from regenerative EV-based therapies.
In contrast, advanced OA characterized by extensive cartilage loss, severe subchondral bone remodeling, and chronic inflammation may require combination strategies involving biomaterials, surgical intervention, or joint replacement. Furthermore, obesity-associated and metabolically driven OA phenotypes may respond differently because of different inflammatory microenvironments. Therefore, future clinical translation will likely depend on personalized therapeutic approaches to improve patient satisfaction.
A significant challenge in harnessing EVs for clinical applications lies in their isolation and characterization, as current methods vary considerably in terms of efficiency, yield, and purity. These factors directly affect reproducibility across studies and therapeutic reliability. Techniques such as ultracentrifugation, size-exclusion chromatography, and polymer-based precipitation offer unique advantages and limitations; however, none have yet been universally accepted as the gold standard 63.
On the other hand, owing to their nanoscale size, EVs are too small to be visualized through conventional light microscopy, and reliable characterization of EVs is essential for evaluating their size, concentration, and molecular content, which can significantly influence their therapeutic efficacy. Comprehensive physical, proteomic, and genomic analyses are therefore required to accurately assess EV bioactive cargoes and their functional properties. Variability in these parameters can lead to diverse clinical outcomes, highlighting the critical need for robust quality-control standards to ensure that only high-quality, reproducible EV preparations are employed in therapeutic settings 64.
EVs offer benefits over whole-cell therapies, including reduced immunogenicity, a lower risk of tumor formation, and improved stability in storage. However, progress toward approved OA treatments is slow due to challenges such as low yield, limited scalability, and high heterogeneity, which were discussed above. To improve clinical application, standardized purification methods are essential, along with the cautious selection of EV sources to ensure consistent therapeutic quality. Additionally, understanding the biodistribution of EVs and their effects on non-cartilaginous tissues is crucial for safety and targeting. Addressing scientific and regulatory challenges and conducting thorough preclinical and clinical studies are vital for integrating EV-based therapies into OA treatment 65.
Regulatory authorities, including the European Medicines Agency (EMA) and the U.S. Food and Drug Administration (FDA), currently classify human EV-based products as biological therapeutics, requiring compliance with stringent safety, quality, and efficacy criteria 66. Regulatory standards for EV-based therapies vary worldwide, with differences in product classification, extraction methods, manufacturing requirements, and quality-control expectations across regions such as the United States, Europe, Japan, South Korea, and Taiwan. Despite these regional differences, the shared objective is to ensure that EV products meet rigorous standards for quality, safety, and efficacy. These regulations emphasize the importance of detailed documentation, traceability, and control throughout the product lifecycle, from raw material sourcing to final distribution and clinical use 67.
From a practical translational perspective, engineered EV products must satisfy several manufacturing and quality-control requirements before clinical implementation. GMP-compliant production requires standardized donor-cell sourcing, serum-free or xeno-free culture conditions, validated isolation methods, sterile processing, and reproducible storage protocols. Batch-to-batch variability remains a major barrier because EV yield, size distribution, cargo composition, and biological potency can change according to donor characteristics, passage number, culture conditions, and purification methods. Therefore, quality-control panels should include particle concentration, size distribution, morphology, protein and RNA markers, purity indicators, sterility, endotoxin testing, residual DNA or reagent assessment, and stability after storage. In addition, potency assays must be functionally linked to the intended therapeutic mechanism, such as suppression of inflammatory mediators, reduction of MMP13 or ADAMTS5 activity, enhancement of COL2A1 and ACAN expression, or promotion of chondrocyte survival under inflammatory stress. Without validated potency assays and release criteria, regulatory approval and inter-study reproducibility will remain limited.
Long-term safety evaluation is also essential because EVs can interact with multiple joint and non-joint cell populations. Potential concerns include undesired immune modulation, off-target biodistribution, pro-fibrotic responses, altered synovial activity, ectopic tissue effects, and unknown consequences of repeated intra-articular dosing. These risks may be amplified in engineered EVs carrying exogenous RNAs, drugs, targeting peptides, or hybrid nanoparticle components. Therefore, future studies should evaluate biodistribution, retention time, degradation, repeat-dose toxicity, immunogenicity, and long-term joint structural outcomes before clinical translation.
Economic concerns may also influence the extensive acceptance of EV-based therapies. Compared with conventional treatments such as corticosteroid injections, hyaluronic acid supplementation, or PRP, engineered EV production currently requires sophisticated isolation technologies, quality-control systems, and specialized manufacturing facilities, which considerably increase treatment costs. In addition, the cost-effectiveness of EV-based therapies compared with existing treatments remains unclear and may represent a barrier to large-scale clinical implementation.
Cartilage repair is rapidly evolving through bioengineering, precision medicine, and combinatorial strategies like EVs, bioactive scaffolds, and gene therapies. While these advances show great promise for true regeneration, rigorous clinical trials, standardized protocols, and regulatory clarity are still needed to translate innovations into safe, patient-specific treatments 68.
At present, most EV-based therapies remain in preclinical or early clinical trial stages. Widespread clinical application is likely to require several years, depending on the resolution of current challenges related to standardization, safety, and regulatory approval.
The future of OA treatment holds significant promise through the incorporation of EVs and cutting-edge technologies such as machine learning and artificial intelligence to facilitate diagnosis. In recent studies, machine learning models have been used to diagnose OA at its earliest stages via untargeted metabolomics data from EV isolates obtained via liquid chromatography‒tandem mass spectrometry. This innovative approach highlights the possibilities of combining EV-based therapies with advanced AI-driven diagnostics, paving the way for more effective and timely interventions in the fight against OA 69.
Although engineered EVs offer significant regenerative potential, overcoming current scientific, manufacturing, and regulatory limitations will determine whether these therapies can successfully transition from experimental platforms to routine clinical practice.
Conclusion
OA is a highly prevalent and disabling joint disorder for which current therapeutic strategies mainly provide symptomatic relief rather than restoring damaged cartilage. This limitation has driven growing interest in regenerative approaches capable of targeting the causal pathological mechanisms of disease progression. Among developing strategies, EVs, particularly those derived from MSCs, have gained significant attention as a promising cell-free regenerative platform. Through their immunomodulatory, anti-inflammatory, and pro-regenerative properties, EVs may help restore chondrocyte homeostasis, enhance anabolic activity, and promote extracellular matrix synthesis, thus supporting cartilage repair and joint tissue protection.
Recent advances in bioengineering have further expanded the therapeutic potential of EV-based therapies. As mentioned, surface engineering strategies improve tissue targeting and retention. Moreover, cargo-loading approaches enhance therapeutic efficacy through the delivery of bioactive molecules, while cellular preconditioning methods strengthen the biological activity of EVs under pathological microenvironments. Collectively, these innovations represent a vital step toward more accurate and effective regenerative interventions for OA.
A key outcome measure for future EV-based OA studies should be the quality of regenerated tissue. Many interventions described as "cartilage regeneration" may actually produce fibrocartilage-like repair tissue rather than native hyaline cartilage. Therefore, future studies should distinguish between symptomatic improvement, matrix deposition, fibrocartilage repair, hyaline-like cartilage formation, and restoration of native cartilage architecture using histological, biomechanical, and molecular criteria.
Nevertheless, several critical challenges continue to hinder clinical translation. Variability in EV source, isolation, purification, and characterization contributes to significant heterogeneity and limits reproducibility across studies. In addition, uncertainties regarding optimal dosing, biodistribution, long-term safety, storage stability, and scalable manufacturing remain major barriers to regulatory approval and widespread clinical application.
Although preclinical studies have demonstrated encouraging regenerative and chondroprotective effects, clinical evidence remains limited, and the long-term therapeutic efficacy of engineered EVs in human OA has yet to be fully established. Future progress in this field may focus on standardized production protocols, deeper mechanistic insights, and robust clinical trials designed to evaluate safety, efficacy, and translational viability. Integration of EVs with biomaterials, tissue engineering platforms, and advanced nanobiotechnology may further enhance their regenerative capacity and therapeutic precision.
In conclusion, engineered EVs represent a rapidly developing and promising frontier in cell-free regenerative medicine for OA. While important scientific, manufacturing, and regulatory challenges remain uncertain, continued advances in EV engineering and translational research may ultimately support the development of more targeted disease-modifying therapies for OA if current translational barriers are resolved.
Data availability statement
No data was used for the research described in the article.
Conflict of Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Funding: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
