Although exosome molecular composition varies with the cell of origin, microenvironment, developmental stage, epigenetic landscape, and precise biogenetic pathway, exosomes share several conserved features. An exosome comprises an external “shell” and an internal “cargo.” The limiting membrane displays a lipid-raft-like architecture enriched in cholesterol, sphingomyelin, and ceramide, which facilitates vesicular trafficking within the cytosol (18). Canonical surface markers include heat-shock proteins, tetraspanins (CD82, CD81, CD63, and CD9), and major histocompatibility complex (MHC) molecules (19). TSG101, which binds ubiquitinated cargo proteins, serves as a hallmark of endosomal sorting. Initially regarded as metabolic waste, exosome contents are now known to encompass abundant nucleic acids, lipids, and proteins, including long non-coding RNA, microRNA, and mRNA, that play critical roles in intercellular communication and immune regulation.

Exosome biogenesis and secretion constitute a tightly regulated, multi-step process initiated by the inward budding of the plasma membrane to form early endosomes. These compartments subsequently mature into multivesicular bodies (MVBs) through a secondary invagination process that requires coordinated action of key molecular machinery, including the endosomal sorting complex required for transport (ESCRT), particularly ESCRT-III, which mediates vesicle budding and scission (20). While ESCRT-dependent mechanisms predominate in most cell types, alternative ESCRT-independent pathways involving sphingomyelinase-mediated ceramide generation can also facilitate exosome formation. Following their biogenesis, exosomes are secreted through three primary mechanisms: (i) fusion of MVBs packed with exosomes to the plasma membrane; (ii) direct budding outward from the plasma membrane; (iii) discharge from intracellular plasma-membrane-connected compartments (IPMCs) after relief of export restrictions (21, 22). Sustained mTORC1 activation inhibits exosome secretion, while mTORC1 blockade promotes it, with both processes being linked to autophagy (23).

In 1996, immunologists first observed that B lymphocytes transformed by the Epstein-Barr virus could produce exosomes via fusion between MVBs and the plasma membrane (24). Subsequent studies revealed that numerous immune and non-immune cells, including T cells, B cells, dendritic cells (DCs), and macrophages, release exosomes capable of mediating immune activation or suppression (25). Exosomal immunoactivity affects both innate and adaptive immunity by modulating antigen presentation, T cell activation, regulatory T cell polarization, immunosuppression, and anti-inflammatory pathways. Exosomes play a vital role in activating and enhancing immune responses via antigen presentation. Professional antigen-presenting cells, including macrophages, B cells, and dendritic cells, release exosomes containing abundant costimulatory signals and MHC class I/II molecules. These exosome-associated peptide antigens are essential for regulating immune function (26).

To contextualize the roles of exosomal cargos in OA pathogenesis, it is essential to examine how epigenetic alterations mediated by these molecules drive disease progression. The molecular payload enclosed within exosomes is intimately involved in the immunomodulation of osteoarthritis. Throughout disease onset and progression, epigenetic alterations, including miRNA repression, histone modification and DNA methylation (52), perturb multiple transcriptional programs and the synthesis of proteolytic factors that govern the anabolic–catabolic equilibrium, such as ADAMTS5, MMP-13 and RUNX2. Among these epigenetic regulators, miRNAs have attracted particular attention in OA pathobiology (53).

Within the joint, miRNAs exert their effects by targeting key mediators of cartilage homeostasis. Current evidence shows that miR-140 is down-regulated in osteoarthritis cartilage and in chondrocytes stimulated with IL-1β, underscoring its chondroprotective function (54). Mice lacking miR-140 exhibit proteoglycan loss and cartilage fibrillation, whereas transgenic over-expression confers resistance to experimental arthritis (55). This regulatory role is further emphasized by the inverse relationship between miR-140 and catabolic enzymes. In osteoarthritis tissues, low level of miR-140 expression is typically accompanied by increased MMP-13 and ADAMTS5, thereby accelerating disease progression (56, 57). Mechanistically, miR-140 exerts its protective effects primarily by directly targeting the 3′-UTR of ADAMTS5 and MMP-13 mRNA, suppressing their translation and thus attenuating matrix degradation (58). Furthermore, miR-140 negatively regulates RALA and SMAD3, which concurrently modulates chondrocyte hypertrophy and reduces catabolic gene expression (59). Exosomal delivery of miR−140 to chondrocytes enhances COL2A1 expression while simultaneously inhibiting IL−1β-induced activation of the NF-κB pathway, leading to reduced synthesis of pro-inflammatory mediators such as inducible nitric oxide synthase (iNOS) and cyclo−oxygenase−2 (COX-2) (60, 61). Moreover, miR−140-containing exosomes prevent oxidative stress-induced apoptosis by targeting P38 MAPK signaling in articular cartilage (62). Together, these multifaceted mechanisms underscore the central role of exosomal miR−140 in restoring the anabolic-catabolic balance within the joint microenvironment and protecting against OA-associated cartilage erosion. Additional miRNAs, including miR-139 and miR-9, similarly disrupt the anabolic–catabolic balance, precipitating extracellular-matrix breakdown and chondrocyte injury (63).

The dysregulation of exosomal miRNAs in OA is not merely a bystander phenomenon but actively contributes to disease pathology. The miRNA composition of exosomes differs markedly between patients with OA and healthy individuals (64). miRNA profiling of synovial-fluid–derived exosomes reveal the miR-200c overexpression in OA, which suppresses ZEB1 and consequently diminishes type II collagen synthesis (65). Further evidence of exosomal miRNA dysregulation comes from inflammatory stimulation experiments: IL-1β stimulation of normal synovial fibroblasts up-regulates 340 and down-regulates 24 distinct miRNAs relative to unstimulated controls, whereas exosomes released by OA chondrocytes contain 22 up-regulated and 29 down-regulated miRNAs compared with those from normal chondrocytes (66). Collectively, these findings highlight the bidirectional relationship between exosomal cargos and OA progression. Such changes reflect both an altered joint microenvironment and the host’s immunoregulatory response; they, in turn, feed back into disease control—for example, elevated miR-140 dampens local inflammatory cytokine release and mitigates immune activation (67). Exosomes thus occupy a pivotal position in the immunological circuitry of OA.

Beyond immunoregulation, exosomes intersect with several other facets of OA pathogenesis, notably synovial angiogenesis. Exosomes released by synovial fibroblasts augment vascular endothelial growth factor (VEGF) secretion, thereby stimulating angiogenesis and driving pathological progression (68, 69). Human umbilical-vein endothelial cells cultured with these exosomes display enhanced migration, tube formation and overall angiogenic capacity. Chondrocyte-derived exosomes are also implicated in osteophyte formation (68). Vesicles 20 – 200 nm in diameter, present within nascent cartilage and bone outgrowths, share exosomal features and carry mediators such as bone morphogenetic proteins (BMPs) that are indispensable for calcification and osteophyte development (70). Exosomes further influence chondrocyte metabolism. Reduced mitochondrial mass in human OA chondrocytes relative to healthy controls, whereas elevated reactive oxygen species (ROS), signifying concurrent mitochondrial dysfunction and ROS accumulation (71–73). Exosome treatment restores mitochondrial integrity, evidenced by increased intracellular ATP, while lowering ROS levels. Intra-articular administration of exosomes can therefore attenuate OA progression by rectifying chondrocyte metabolic defects (74).

The therapeutic potential of MSC-derived exosomes in OA hinges on their ability to reprogram the inflammatory microenvironment. MSC-derived exosomes contain bioactive components, including trophic factors and apoptosis inhibitors, that modulate the injury microenvironment by shifting the balance from pro-inflammatory to anti-inflammatory responses (75). This shift is particularly relevant in OA, where synovial macrophages play a central role in disease progression (76, 77). In OA pathogenesis, both clinical observations and experimental models demonstrate significant inflammatory cell accumulation within the synovial tissue, particularly highlighting the crucial role of synovial macrophages (78). Notably, macrophage polarization dictates their functional impact: during inflammatory processes, macrophages undergo functional polarization into two distinct subsets: the pro-inflammatory M1 phenotype and the anti-inflammatory M2 phenotype, which play counterregulatory roles in disease progression (79).

MSC-derived exosomes directly influence macrophage polarization to attenuate inflammation. Exosomes derived from MSCs, particularly those containing elevated levels of miR-135b, inhibit MAPK6 expression, which facilitates the polarization of synovial macrophages toward the M2 phenotype and subsequently reduces cartilage degeneration (80, 81). Another key mechanism involves the suppression of NF-κB signaling, a master regulator of inflammation. Targeted inhibition of NF−κB is considered a promising strategy for controlling OA−associated inflammation (82). Upon stimulation with inflammatory mediators, NF-κB translocates to the nucleus and up regulates a repertoire of inflammatory genes encoding proteins, including COX-2, MMPs, and iNOS, culminating in chondrocyte death and exacerbation of OA pathology. MSC exosomes counteract this process via specific miRNAs: phosphorylation−dependent degradation of IκB−α is a pivotal step in NF−κB activation; MSC−derived exosomes carrying miR−147b inhibit TNF−α and IL−1β- driven expression of inflammatory mediators and prevent IκB−α degradation (83, 84) ( Supplementary Table S1 ).

Multiple RNAs contained in MSC exosomes modulate inflammatory signaling in OA. For example, miR−222 targets HDAC4 and thereby down−regulates MMP−13 protein (85, 86); miR−199a−5p lowers IL−6 and TNF−α levels, limiting inflammation and cartilage destruction; miR−140−5p targets Toll−like receptor 4 (TLR4), restraining proliferation of synovial fibroblasts and reducing IL−6 and IL−8 secretion, thus fostering cartilage regeneration (87, 88). Additional miRNAs contribute to the resolution of oxidative stress and inflammation. miR−9−5p down−regulates SDC1, diminishing expression of IL−1, IL−6, TNF−α, MMP−13, alkaline phosphatase (ALP), cartilage oligomeric matrix protein (COMP) and C−reactive protein (CRP), while increasing superoxide dismutase (SOD), NO, malondialdehyde (MDA), iNOS and COX−2, collectively alleviating cartilage injury and curbing inflammatory and oxidative stress damage (89, 90). Beyond miRNAs, long non-coding RNAs also play a role: the long non−coding RNA MALAT1 up−regulates miR−19b via the Wnt/β−catenin and NF−κB pathways, protecting chondrocytes from lipopolysaccharide−induced inflammatory injury (91). In addition, miR−181c, miR−146a, and miR−21 contained in MSC exosomes can reverse the pathological inflammatory milieu characteristic of OA (92, 93) ( Figure 1B ).

The reparative function of MSCs and their exosomes can be modulated by diverse pre−conditioning strategies, encompassing both biomaterial−based and physical interventions. Biomaterials such as hyaluronic acid, sodium−alginate Janus microspheres and related carriers enhance MSC adhesion to cartilage and enable targeted delivery that accelerates cartilage repair (94, 95). Recent advances in exosome-based OA therapies highlight innovative strategies for cartilage repair. Preconditioning MSCs with cytokines and biomaterials, such as hyaluronic acid and Janus microspheres, enhances exosome yield and therapeutic efficacy (64, 96). Engineered exosomes, such as CRISPR/Cas9-loaded CAP-modified hybrids (CAP/FGF18-hyEXO) and fucoidan-primed exosomes (F-MSCs-Exo), promote chondrogenesis and autophagy via miR-146b-5p (97). Hydrogel encapsulation enables sustained, targeted delivery. Conversely, pathogenic FLS-derived exosomes exacerbate OA via HIF1A-driven glycolysis, which is reversible with 2-DG (98). Macrophage-derived miR-26b-5p exosomes and placental exosomes further modulate inflammation and anabolism (50). These approaches underscore exosomes’ potential in precision OA therapy through engineering, priming, and advanced delivery systems.