Mammalian, plant, and gut microbiota-derived extracellular vesicles as emerging therapeutics for bone diseases Authors: Guanchen Yin, Hao Wang, Yi Liu, Zhenyan Wang, Xu Lin, Qisheng Cheng, Hang Gao Published: 2026 ABSTRACT Bone disorders, including osteoporosis and osteoarthritis, represent a growing global health burden, yet current therapies remain limited by poor targeting efficiency and significant adverse effects. Extracellular vesicles (EVs), nanoscale lipid bilayer particles that mediate intercellular communication, have emerged as promising endogenous nanocarriers with intrinsic biocompatibility and regulatory capacity. This review summarises recent advances in EVs-based therapeutic strategies for bone diseases, focusing on EVs derived from mammalian cells, plants, and the gut microbiota. Particular emphasis is placed on gut microbiota-derived EVs, which have gained increasing attention for their role in regulating bone homeostasis through the “gut-bone axis”. By comparing the biological characteristics, advantages, limitations, and disease-specific applicability of EVs from these distinct sources, we highlight their complementary therapeutic potential. Overall, this review provides a concise framework for the rational development and clinical translation of EVs-based precision therapies for bone disorders. Introduction Bone disorders—including osteoporosis (OP), osteoarthritis (OA), rheumatoid arthritis (RA), fractures, and avascular necrosis of the femoral head—are increasingly prevalent worldwide and constitute a major cause of disability. Global burden of disease analyses identify musculoskeletal disorders as one of the leading contributors to years lived with disability, reflecting the substantial clinical and socioeconomic impact of these conditions. Population aging and lifestyle changes further exacerbate this burden, underscoring the need for more effective therapeutic strategies (GBD, 2021 Diseases and Injuries Collaborators, 2024; GBD, 2021 Other Musculoskeletal Disorders Collaborators, 2023). Current treatments rely primarily on pharmacological intervention and surgery, yet their long-term efficacy is limited. In OP, anti-resorptive and anabolic agents act unidirectionally on bone remodeling and often fail to restore physiological balance, while prolonged use is associated with poor adherence and serious adverse effects (Ayers et al., 2023; Bellido, 2024; Reid and Billington, 2022). OA management remains largely symptomatic, with non-steroidal anti-inflammatory drugs and joint replacement surgery constrained by systemic toxicity, surgical risk, and limited disease-modifying capacity (Zeng et al., 2018; Richardson et al., 2023). Similar challenges are evident across other bone pathologies, highlighting the unmet need for targeted, mechanism-based therapies (Serhal et al., 2020; Depypere et al., 2020). In addition to traditional drug and surgical treatments, therapies related to the “gut-bone axis” are also gaining increasing attention. The traditional “gut-bone axis” primarily involves the direct regulation of bone metabolism by gut microbiota metabolites (e.g., SCFAs), as well as indirect regulation via the host’s immune and endocrine systems. Based on these mechanisms, traditional “gut-bone axis” therapies, such as dietary interventions, probiotics, and fecal microbiota transplantation (FMT), have been applied to the treatment of bone diseases (Liu et al., 2025). However, despite their therapeutic potential, these interventions still face significant clinical challenges, including the risk of colonization failure and systemic infection (Yadegar et al., 2024). In recent years, addressing the clinical challenges of bone regeneration has driven the extensive exploration of advanced biomaterials and therapeutic strategies to enhance bone regeneration and implant osseointegration. Significant progress has been made in engineering functional microenvironments for bone repair, including the development of biomimetic periosteum (Du et al., 2025), tissue-adhesive barrier membranes for guided bone regeneration (GBR) (Poos et al., 2025), and the application of nanotechnology to optimize orthopedic implants and composite scaffolds (Liang et al., 2024). In this context, extracellular vesicles (EVs) have emerged as a highly promising new alternative and adjunctive therapeutic approach in the field of orthopedics (Kumar et al., 2024; Liu Z. et al., 2023; Fang et al., 2024; Liu H. et al., 2022). Nearly all eukaryotic and prokaryotic cells secrete EVs, which are nanoscale particles (typically 30–1,000 nm in diameter) enclosed by a robust lipid bilayer and serve as fundamental mediators of intercellular and interorgan communication. This stable membrane structure effectively protects the complex bioactive substances within (including nucleic acids, proteins, lipids, and metabolites) from enzymatic degradation, facilitating their functional delivery to recipient cells via endocytosis, direct membrane fusion, or receptor-ligand interactions. Compared to synthetic nanocarriers, naturally derived EVs exhibit excellent biocompatibility, extremely low immunogenicity, and an inherent ability to evade immune clearance and traverse biological barriers. Consequently, they serve as highly effective and safe delivery vehicles for the treatment of systemic and localized skeletal disorders (Welsh et al., 2024; Huang et al., 2023; van Niel et al., 2022; Yin et al., 2022). Growing evidence implicates EVs in the regulation of bone metabolism, inflammation, and regeneration (Liu et al., 2021; Thomas et al., 2024; Chen et al., 2022). While mammalian cell-derived and plant-derived EVs have been extensively studied, EVs derived from the gut microbiota have only recently attracted attention (Fan and Pedersen, 2021). The gut microbiota plays a central role in skeletal homeostasis through the “gut-bone axis”, integrating immune, metabolic, and endocrine signaling (Welsh et al., 2024; Huang et al., 2023). Notably, gut microbiota-derived EVs(GM-EVs) can cross the intestinal barrier, enter systemic circulation, and influence distant tissues, including bone, thereby representing a critical mediator of microbiota-host crosstalk (Chen et al., 2022; Jiao et al., 2025). In this review, we provide a concise and comparative overview of EVs derived from mammalian cells, plants, and the gut microbiota in bone diseases. We highlight the emerging role of GM-EVs, contrast the strengths and limitations of EVs from different origins, and discuss their translational potential. This synthesis aims to inform the rational design of EVs-based therapeutic strategies for bone disorders. Mammalian cell-derived extracellular vesicles and their role in bone diseases Mammalian cell-derived EVs (MEVs), released by mesenchymal stem cells (MSCs), immune cells, endothelial cells, and bone-resident cells, are key regulators of skeletal homeostasis (Giancaterino and Boi, 2023). These EVs regulate osteoblast (OB) differentiation, osteoclast (OC) activity, and osteocyte signaling, thereby coordinating the processes of bone formation and resorption (Ra et al., 2010; Huang Y. et al., 2021; Lu et al., 2019). For example, studies demonstrate that EVs derived from bone marrow mesenchymal stem cells (BMSCs) can deliver lncTUG1 to OB, sequestering miR-22-5p and upregulating Anxa8 expression, thereby promoting OB differentiation and accelerating fracture healing (Li et al., 2023). Furthermore, BMSCs-derived EVs containing miR-206 enhance OB proliferation and differentiation in OA by downregulating Elf3, contributing to the attenuation of OA progression (Huang Y. et al., 2021). Similarly, EVs derived from adipose-derived MSCs (ADSCs) alleviate diabetic OP by inhibiting NLRP3 inflammasome activation in OC (Zhang L. et al., 2021). In addition, Davies et al. found that EVs released by mineralized OB promote MSCs differentiation into OB by carrying Annexin A1, A2, and A6, and bridging collagen type VI (Davies et al., 2017). MEVs exert bidirectional effects on bone metabolism. EVs derived from young or healthy cells generally promote osteogenesis, suppress osteoclastogenesis, and facilitate fracture repair, whereas EVs originating from aged, senescent, or diseased cells may impair osteogenic differentiation, enhance adipogenesis, and exacerbate bone loss (Wang et al., 2024; Ni et al., 2019; Lv et al., 2022; Zhao et al., 2023). Evidence indicates that EVs isolated from the plasma of OP patients suppress MSCs differentiation into OB via miR-1246. Furthermore, they promote OC activation by presenting surface RANKL and upregulating RANKL expression in OB, shifting the remodeling equilibrium toward excessive resorption (Pepe et al., 2022). Beyond metabolic regulation, MEVs modulate inflammatory and oxidative stress responses by shaping macrophage polarization, regulating T-cell subsets, and controlling key pathways such as NF-κB, PI3K-AKT, STAT3, and Nrf2 (Li et al., 2022; Yuan et al., 2023; Chen et al., 2024; Li Y. et al., 2024). For instance, EVs derived from human umbilical cord MSCs (hUCMSCs) deliver miR-122-5p, miR-148a-3p and key proteins including A2M and ALB to activate the PI3K-Akt pathway. This activation suppresses NF-κB nuclear translocation, promotes M2 macrophage polarization, reduces pro-inflammatory cytokines including IL-6, IL-1β and TNF-α, and increases anti-inflammatory cytokines including IL-10 and TGF-β, thereby alleviating OA (40). Similarly, EVs derived from H2O2-pretreated ADSCs significantly reduce ROS levels and oxidative damage in diabetic BMSCs by activating the Nrf2/HO-1 pathway. This alleviates cellular senescence and enhances osteogenic differentiation, promoting the repair of diabetic bone defects (Li Y. et al., 2024). In addition, MEVs support tissue regeneration by promoting angiogenesis, inhibiting apoptosis, remodeling extracellular matrix, and activating resident progenitor cells (Chew et al., 2019; Yan et al., 2021; Kuang et al., 2019; Liu et al., 2017; Xu et al., 2019; Wu et al., 2023). Specifically, umbilical cord MSC-derived EVs deliver miR-21 to target PTEN mRNA, suppressing PTEN protein expression. This alleviates the inhibition of the AKT pathway and promotes AKT phosphorylation, leading to reduced expression of apoptotic proteins such as cleaved Caspase-3 and Bax and increased expression of the anti-apoptotic protein Bcl-2. Consequently, these EVs attenuate osteocyte apoptosis and ameliorate the severity of osteonecrosis of the femoral head in rat models (Kuang et al., 2019). In addition,MSC-derived EVs enhance endothelial cell proliferation, migration, and tube formation capacity by activating the PI3K/AKT pathway, thus promoting angiogenesis and offering a promising strategy for the prevention and treatment of osteonecrosis of the femoral head (Liu et al., 2017). Owing to their high biocompatibility, intrinsic targeting capacity, and cargo versatility, MEVs represent a promising platform for precision therapies in chronic bone diseases (as summarized in Figure 1; Table 1). However, translating positive preclinical findings into clinical practice requires careful consideration. A major limitation is that current mechanistic studies rely heavily on small animal models, which cannot fully replicate the complex long-term dynamics of skeletal remodeling in humans. Furthermore, the inherent batch-to-batch variability in cell culture and the lack of standardized protocols for isolating clinical-grade MEVs remain key bottlenecks that must be overcome before large-scale clinical translation can be achieved. MEVs in the Regulation of Bone Homeostasis. Taken together, this figure illustrates the crucial bidirectional regulatory role of MEVs in skeletal homeostasis: they actively drive OB-mediated bone formation and tissue/angiogenic repair, while simultaneously suppressing osteoclastogenesis, adipogenic differentiation, and macrophage-mediated inflammatory and oxidative stress responses. TABLE 1Summary of key cargo molecules in MEVs and their functions and regulatory mechanisms in skeletal diseases.EVs typeEVs sourceMolecular mediators in the EVsFunction and mechanism of actionReferencesMEVsBMSCsLncTUG1Competitively binds miR-22-5p to upregulate Anxa8, promoting OBs differentiation and fracture healing Li et al. (2023) BMSCsmiR-206Downregulates Elf3 to enhance OBs proliferation/differentiation and delay OA progression Huang Y et al. (2021) U-MSCsLncRNA H19Promotes chondrocyte proliferation and matrix synthesis while inhibiting apoptosis Chen et al. (2024) U-MSCsmiR-21Inhibits PTEN to activate AKT, decreasing Bax/Caspase-3 and increasing Bcl-2 to reduce osteocyte apoptosis Li Y et al. (2024) U-MSCsmiR-122-5p, miR-148a-3p, A2M, ALBActivates the PI3K-Akt pathway, inhibits NF-κB, and promotes M2 macrophage polarization, thereby alleviating OA Lv et al. (2022) TDSCsTGF-βActivates TGF-β/SMAD3 to promote tenogenic differentiation and stem cell proliferation for tendon regeneration Yan et al. (2021) EPCsmiR-222-3pDownregulates SOCS3 to activate JAK2/STAT3, promoting M2 macrophage polarization and alleviating spinal cord inflammation Zhao et al. (2023) EPCs-Inhibiting OBs ferroptosis to improve hormone-induced OP and osteonecrosis Lu et al. (2019) OBsAnnexin A1/A2/A6Binds to type VI collagen, thereby promoting the differentiation of MSCs into OBs Davies et al. (2017) MEVsOCsmiR-106a-5pDownregulates Fam134a, enhances osteogenic differentiation of BMSCs, and thereby promotes bone regeneration Kuang et al. (2019) ABMCmiR-483-5pTargets PPARγ to promote BMSCs adipogenesis, inducing bone-fat imbalance and OP Wang et al. (2024) OA-ChsmiR-449a-5pInhibits ATG4B to block autophagy, inducing mitochondrial ROS and NLRP3 inflammasome activation to exacerbate OA. Ni et al. (2019) OA-ChsLncRNA OANCTInhibits m6A demethylation to stabilize PIK3R5 mRNA, activating PI3K/AKT/mTOR and M1 macrophage polarization to exacerbate OA Lv et al. (2022) dNPCsmiR-27a-3pActivates the PPARγ/NF-κB/PI3K/AKT pathway, induces M1 macrophage polarization, thereby exacerbating intervertebral disc degeneration Zhao et al. (2023) Sen-Ost-Reducing TPM1 expression promotes the adipogenic differentiation of BMSCs Wang et al. (2024) BMSCs, Bone Marrow Mesenchymal Stem Cells; U-MSCs, Mesenchymal Stem Cells; TDSCs, Tendon-Derived Stem Cells; EPCs, Endothelial Progenitor Cells; OBs, Osteoblasts; OCs, Osteoclasts; ABMC, Aged Bone Matrix Cells; OA-Chs, Osteoarthritis Chondrocytes; dNPCs, Degenerated Nucleus Pulposus Cells; Sen-Ost, Senescent Osteocytes; LncTUG1, Long Non-Coding RNA TUG1; Anxa8, Annexin A8; Elf3, E74-Like Factor 3; OA, Osteoarthritis; LncRNA H19, Long Non-Coding RNA H19; PTEN, Phosphatase and Tensin Homolog; AKT, Protein Kinase B; Caspase-3, Cysteine Aspartate Protease-3; Bax, Bcl-2-Associated X Protein; Bcl-2, B-Cell Lymphoma 2; A2M, Alpha-2-Macroglobulin; ALB, Albumin; SOCS3, Suppressor of Cytokine Signaling 3; PPARγ, Peroxisome Proliferator-Activated Receptor Gamma; ATG4B, Autophagy-Related Gene 4B; ROS, Reactive Oxygen Species; m6A, N6-Methyladenosine; TPM1, Tropomyosin-1. Summary of key cargo molecules in MEVs and their functions and regulatory mechanisms in skeletal diseases. BMSCs, Bone Marrow Mesenchymal Stem Cells; U-MSCs, Mesenchymal Stem Cells; TDSCs, Tendon-Derived Stem Cells; EPCs, Endothelial Progenitor Cells; OBs, Osteoblasts; OCs, Osteoclasts; ABMC, Aged Bone Matrix Cells; OA-Chs, Osteoarthritis Chondrocytes; dNPCs, Degenerated Nucleus Pulposus Cells; Sen-Ost, Senescent Osteocytes; LncTUG1, Long Non-Coding RNA TUG1; Anxa8, Annexin A8; Elf3, E74-Like Factor 3; OA, Osteoarthritis; LncRNA H19, Long Non-Coding RNA H19; PTEN, Phosphatase and Tensin Homolog; AKT, Protein Kinase B; Caspase-3, Cysteine Aspartate Protease-3; Bax, Bcl-2-Associated X Protein; Bcl-2, B-Cell Lymphoma 2; A2M, Alpha-2-Macroglobulin; ALB, Albumin; SOCS3, Suppressor of Cytokine Signaling 3; PPARγ, Peroxisome Proliferator-Activated Receptor Gamma; ATG4B, Autophagy-Related Gene 4B; ROS, Reactive Oxygen Species; m6A, N6-Methyladenosine; TPM1, Tropomyosin-1. Plant EVs and bone disorders Plant-derived EVs (PDEVs), naturally secreted by edible and medicinal plants, encapsulate a rich repertoire of plant-specific phytochemicals, lipids, proteins, and regulatory RNAs. Endowed with intrinsic anti-inflammatory, antioxidant, and regenerative properties, PDEVs are emerging as a highly compelling natural nanotherapeutic strategy for the management of bone disorders (Cui et al., 2020; Lian et al., 2022). In bone diseases, PDEVs promote osteogenic differentiation, inhibit OC activation, and rebalance bone remodeling by activating signaling pathways such as BMP-Runx2, MAPK, and Smad (Zhan W et al., 2023; Hwang et al., 2023; Cao et al., 2024; Park et al., 2023; Seo et al., 2023). Specifically, yam-derived EVs (YNVs) activate the BMP-2/p-p38-dependent Runx2 signaling pathway, thereby promoting OB proliferation, differentiation, and mineralization, ultimately improving bone mineral density and microarchitectural integrity in osteoporotic mice (Hwang et al., 2023). Furthermore, euphorbia-originated EVs (MOEVLPs) promote the proliferation of osteoblastic precursor cells (MC3T3-E1) by activating the MAPK signaling pathway, leading to enhanced bone formation and improved trabecular structure in OP models (Cao et al., 2024). Concurrently, PDEVs attenuate inflammation and oxidative stress by suppressing NF-κB and NLRP3 inflammasome activation, activating Nrf2-dependent antioxidant responses, and modulating macrophage polarization (Kim et al., 2023; De Robertis et al., 2020; Liu C. et al., 2022; Zeng et al., 2024; Lou et al., 2023). In particular, ginger-derived EVs relieve Keap1-mediated inhibition of Nrf2 and promote Nrf2 nuclear translocation, thereby downregulating inflammatory factors including IL-6 and TNF-α and upregulating antioxidant genes such as heme oxygenase-1 (HO-1) and NAD(P)H quinone dehydrogenase 1 (NQO1) to reduce ROS and oxidative damage (Zeng et al., 2024). In tandem, grapefruit-derived EVs improve the chondrocyte microenvironment by downregulating inflammatory genes including COX2 and PTGS2 and upregulating antioxidant genes including superoxide dismutase 2 (SOD2) and glutathione peroxidase (GPX), exerting combined anti-inflammatory and antioxidant effects (Rashidi et al., 2025). Although direct evidence for PDEVs-mediated bone regeneration remains limited, their ability to improve the inflammatory, oxidative, and angiogenic microenvironment indirectly supports skeletal repair (Rashidi et al., 2025; Zu et al., 2021). Furthermore, grapefruit-derived EVs accelerate wound healing through their antioxidant effects by reducing ROS generation. They also stimulate extracellular matrix (ECM) production by upregulating COL1A1, fibronectin, and other related genes, and promote angiogenesis by enhancing the tube-forming capacity of vascular endothelial cells (Savcı et al., 2021). Similarly, Aloe vera-derived EVs exhibit anti-inflammatory properties by suppressing mRNA expression of pro-inflammatory cytokines including IL-6 and IL-1β, promote angiogenesis by improving endothelial cell tubulogenesis, and enhance fibroblast proliferation and migration, collectively offering therapeutic value in chronic wound healing (Kim and Park, 2022). PDEVs are therefore well suited as adjunctive therapies for multifactorial bone disorders, such as OA and OP, particularly where inflammation and oxidative stress are dominant drivers of pathology (as summarized in Figure 2; Table 2). Although the multifunctionality of PDEVs has been well documented in vitro and in animal models, there remains a significant gap in understanding their direct clinical relevance in targeted bone therapy. Furthermore, current research primarily relies on local administration or nonspecific systemic delivery. Whether orally administered PDEVs can naturally bypass the gastrointestinal barrier and reach therapeutic concentrations in distant skeletal sites remains an unproven hypothesis. These highlight key areas for future research. PDEVs in the Regulation of Bone Homeostasis. Collectively, this figure highlights the comprehensive therapeutic potential of PDEVs, demonstrating their ability to orchestrate bone and cartilage regeneration, promote angiogenesis for tissue repair, and fundamentally improve the local microenvironment through potent, multi-targeted anti-inflammatory and antioxidant signaling. TABLE 2Summary of key cargo molecules in PDEVs and their functions and regulatory mechanisms in skeletal diseases.EVs typeEVs sourceMolecular mediators in the EVsFunction and mechanism of actionReferencesPDEVsPDEVsEuphorbia-Activates MAPK/CREB/RSK1 signaling to promote MC3T3-E1 proliferation and ameliorate OP Cao et al. (2024) Yam-Activates BMP-2/p-p38/Runx2 pathway to enhance OBs proliferation, differentiation, and mineralization Hwang et al. (2023) Plum-Upregulates BMP-2/Smad-1 and p-p38/p-JNK to induce Runx2/Osterix-mediated osteogenesis; downregulates NFATc1/c-Fos to block RANKL-induced osteoclastogenesis Park et al. (2023) Puerariae lobata-Increases Runx2 levels to drive ALP/OCN expression and promote BMSC osteogenic differentiation Zhan H et al. (2023) Ginseng-Inhibits IκBα/JNK/ERK phosphorylation to downregulate NFATc1/c-Fos and block OCs maturation Seo et al. (2023) Fucosylated MSCsmiR-146b-5pBlocks PI3K/AKT/mTOR to reduce COX-2/iNOS, inhibit M1 macrophages polarization, and alleviate OA Lou et al. (2023) Ginger-Inhibits Keap1 to promote Nrf2 nuclear translocation, upregulating HO-1/NQO1 and reducing oxidative damage Zeng et al. (2024) Tea-Reduces ROS and pro-inflammatory factors while promoting IL-10 secretion Zu et al. (2021) Blueberry-Upregulates HO-1/NQO1 and reduces ROS levels De Robertis et al. (2020) Grapefruit-Reduces ROS and upregulates COL1A1/Fibronectin to promote ECM formation and angiogenesis Savcı et al. (2021) Grapefruit-Downregulates COX2/PTGS2 and upregulates SOD2/GPX to improve chondrocyte microenvironment Rashidi et al. (2025) Aloe vera-Suppresses IL-6/IL-1β, promotes angiogenesis, and enhances fibroblast migration for chronic wound healing Kim and Park (2022) MAPK, Mitogen-Activated Protein Kinase; CREB, cAMP Response Element-Binding Protein; RSK1, Ribosomal S6 Kinase 1; MC3T3-E1, Mouse Osteoblast Cell Line; BMP-2, Bone Morphogenetic Protein 2; p38, p38 Mitogen-Activated Protein Kinase; Runx2, Runt-Related Transcription Factor 2; OBs, Osteoblasts; OCs, Osteoclasts; Smad1, Mothers Against Decapentaplegic Homolog 1; JNK, c-Jun N-Terminal Kinase; Osterix, Sp7 Transcription Factor; ALP, Alkaline Phosphatase; OPN, Osteopontin; OCs, Osteoclasts; NFATc1, Nuclear Factor of Activated T Cells 1; c-Fos, Proto-Oncogene c-Fos; PPAR-γ, Peroxisome Proliferator-Activated Receptor Gamma; OCN, Osteocalcin; BMSCs, Bone Marrow Mesenchymal Stem Cells; TRAP, Tartrate-Resistant Acid Phosphatase; OSCAR, Osteoclast-Associated Receptor; MSCs, Mesenchymal Stem Cells; COX-2, Cyclooxygenase-2; iNOS, Inducible Nitric Oxide Synthase; OA, Osteoarthritis; HO-1, Heme Oxygenase 1; NQO1, NAD(P)H Quinone Dehydrogenase 1; ROS, Reactive Oxygen Species; COL1A1, Collagen Type I Alpha 1; ECM, Extracellular Matrix; PTGS2, Prostaglandin-Endoperoxide Synthase 2; SOD2, Superoxide Dismutase 2; GPX, Glutathione Peroxidase. Summary of key cargo molecules in PDEVs and their functions and regulatory mechanisms in skeletal diseases. MAPK, Mitogen-Activated Protein Kinase; CREB, cAMP Response Element-Binding Protein; RSK1, Ribosomal S6 Kinase 1; MC3T3-E1, Mouse Osteoblast Cell Line; BMP-2, Bone Morphogenetic Protein 2; p38, p38 Mitogen-Activated Protein Kinase; Runx2, Runt-Related Transcription Factor 2; OBs, Osteoblasts; OCs, Osteoclasts; Smad1, Mothers Against Decapentaplegic Homolog 1; JNK, c-Jun N-Terminal Kinase; Osterix, Sp7 Transcription Factor; ALP, Alkaline Phosphatase; OPN, Osteopontin; OCs, Osteoclasts; NFATc1, Nuclear Factor of Activated T Cells 1; c-Fos, Proto-Oncogene c-Fos; PPAR-γ, Peroxisome Proliferator-Activated Receptor Gamma; OCN, Osteocalcin; BMSCs, Bone Marrow Mesenchymal Stem Cells; TRAP, Tartrate-Resistant Acid Phosphatase; OSCAR, Osteoclast-Associated Receptor; MSCs, Mesenchymal Stem Cells; COX-2, Cyclooxygenase-2; iNOS, Inducible Nitric Oxide Synthase; OA, Osteoarthritis; HO-1, Heme Oxygenase 1; NQO1, NAD(P)H Quinone Dehydrogenase 1; ROS, Reactive Oxygen Species; COL1A1, Collagen Type I Alpha 1; ECM, Extracellular Matrix; PTGS2, Prostaglandin-Endoperoxide Synthase 2; SOD2, Superoxide Dismutase 2; GPX, Glutathione Peroxidase. Gut microbiota EVs and bone disorders Bone disorders are often associated with intestinal dysbiosis. For instance, OP patients show elevated levels of Clostridium, Ruminococcaceae, and Megamonas but reduced Roseburia and Weissella (Yang et al., 2022). Similarly, OA patients exhibit increases in Roseburia and Butyricicoccus alongside decreases in Bacteroides spp. and Faecalibacterium prausnitzii (Wang W. et al., 2025). EVs derived from the gut microbiota have recently emerged as critical mediators of microbiota-host communication in bone diseases (Jones et al., 2018). By crossing the intestinal barrier and entering systemic circulation, GM-EVs influence distant skeletal tissues through immune, metabolic, and endocrine pathways (Chen et al., 2022). Beneficial GM-EVs promote OB activity, suppress osteoclastogenesis, and preserve bone mass, whereas EVs derived from pathogenic or dysbiotic bacteria can induce inflammation, disrupt immune balance, and accelerate bone loss (Liu et al., 2021; Chen et al., 2022; Wang et al., 2022; Bielaszewska et al., 2018). For instance, EVs from the beneficial bacterium Akkermansia muciniphila accumulate in bone tissue, promoting mineralization and inhibiting OC, which improves bone density in ovariectomized mice. Conversely, reduced abundance of beneficial bacteria diminishes these protective effects (Liu et al., 2021). GM-EVs regulate bone health by modulating inflammatory signaling (e.g., TLR-NF-κB) and immune cell differentiation (e.g., Treg/Th17 balance). EVs from the probiotic Lactobacillus johnsonii inhibit mTORC1 and promote anti-inflammatory M2 macrophage polarization, thereby alleviating OA-associated cartilage damage (Liu R. et al., 2023). In RA, EVs from the probiotic Propionibacterium freudenreichii mitigate RA by reducing collagen-specific antibodies, suppressing pro-inflammatory cytokines, and inhibiting OC formation and bone erosion (Woo et al., 2024). In contrast, Fusobacterium nucleatum EVs deliver the virulence factor FadA to synovial macrophages, activating the Rab5a/YB-1 axis to drive IL-6 and TNF-α production and tissue destruction (Hong et al., 2023). They are also involved in metabolic pathways mediated by short-chain fatty acids (SCFAs), serotonin, bile acids, and trimethylamine N-oxide (TMAO) (Peng et al., 2021; Yang et al., 2024; Hu et al., 2022). Beyond direct skeletal effects, GM-EVs maintain intestinal barrier integrity and shape microbial ecology, thereby preventing systemic inflammation that contributes to bone degeneration (Thoo et al., 2019; Ciccia et al., 2017; Li et al., 2019; Liang et al., 2022). Intestinal dysbiosis and barrier dysfunction allow the translocation of pathogens and toxins into the systemic circulation, affecting distal tissues including bone (Yan et al., 2016; Guido et al., 2021). Lactobacillus fermentum EVs increase microbial diversity and the abundance of beneficial genera (Gao et al., 2026). Akkermansia muciniphila EVs facilitate intraspecies competition by inhibiting rival clades and modulating host immunity to reinforce niche dominance (Hong et al., 2025). Furthermore, EVs from probiotic strains like E. coli Nissle 1917 (EcN) enhance tight junction proteins (ZO-1, claudin-14), increase transepithelial electrical resistance (TEER), and induce anti-inflammatory cytokines and antimicrobial peptides (hBD-2), thereby strengthening the barrier (Alvarez et al., 2016). Emerging evidence further suggests that GM-EVs participate in multi-organ communication within the “brain-gut-bone axis”, offering a novel paradigm for systemic regulation of skeletal homeostasis (Li R. et al., 2024; Zhang YW. et al., 2021; Bravo et al., 2011; Siva Venkatesh et al., 2024). The “brain-gut-bone axis” is a complex, integrated network linking the central nervous system (CNS), gut microbiota, and skeletal system via multi-directional communication, involving neural, immune, endocrine, and metabolic pathways (Li R. et al., 2024; Zhang YW. et al., 2021). However, when evaluating this complex cross-talk, it is pivotal to distinguish clinical observational correlations from mechanistically proven causalities. Gut-Brain Interactions: Clinical studies often report associations between intestinal dysbiosis and neurological disorders, such as the co-occurrence of depression and inflammatory bowel disease (Huang X. et al., 2021). Beyond these correlations, experimental models provide evidence for the causal relationships in this bi-directional communication. For instance, Lactobacillus rhamnosus JB-1 modulates neurobehavior via vagal pathways (Bravo et al., 2011). This specific link is supported by vagotomy models, in which severing the vagus nerve prevents these microbe-induced behavioral changes (Sgritta et al., 2019). Conversely, the brain influences gut microbiota through sympathetic nerve signaling to intestinal immune cells, protecting against dysbiosis (Gabanyi et al., 2016). Moreover, direct intervention studies have revealed the effects of microbial metabolites: exogenous supplementation with SCFAs attenuates neuroinflammation and regulates brain function (Siva Venkatesh et al., 2024). Brain-Bone Interactions: Epidemiological studies have long established a positive correlation between chronic psychological stress (or depression) and decreased bone mineral density (Zhang YW. et al., 2021). Beyond these correlations, experimental models further reveal the causal mechanisms underlying this interaction. Specifically, the central nervous system regulates bone metabolism via specific neurotransmitters (e.g., norepinephrine, neuropeptide Y). Mice deficient in sympathetic β2-adrenergic receptors exhibit increased bone mass, demonstrating a causal inhibitory effect of sympathetic tone on OB activity (Otto et al., 2020). Reciprocally, bone functions as an endocrine organ. Direct administration of bone-derived mediators, such as osteocalcin (OCN) and FGF18, alleviates motor deficits in parkinson’s models, indicating a functional bone-to-brain feedback loop (Guo et al., 2018; Guo et al., 2017). As previously detailed, the “gut-bone axis” highlights the gut’s influence on skeletal health. These three axes are functionally interlinked, forming a synergistic regulatory network where each component acts as both a “signal source” and “target effector”. Intestinal microbiota-derived EVs are crucial carriers for precise inter-organ signal transfer within this “brain-gut-bone axis”. Their protective membrane structure allows them to traverse physiological barriers like the intestinal mucosa and blood-brain barrier, overcoming issues of signal degradation and poor targeting seen with free metabolites. For instance, Lactobacillus EVs promote neurite growth and neuroprotection more effectively and stably than free metabolites in alzheimer’s models (Kim et al., 2024). Although mechanistic understanding and safety profiling remain incomplete, existing research has shown that the primary safety concerns associated with GM-EVs stem from the pathogen-associated molecular patterns (PAMPs) they carry, particularly lipopolysaccharides (LPS) derived from the outer membranes of Gram-negative bacteria (Chen et al., 2025). Given that the gut microbiota is a complex ecosystem composed of both beneficial bacteria and potentially pathogenic bacteria, GM-EV preparations may inevitably contain vesicles derived from Gram-negative pathogenic commensals with high LPS loads. When exposed in their entirety, there is a risk of inducing systemic inflammation or localized excessive inflammation (Broz, 2016). It is important to note that this risk is strain-dependent. Therefore, rigorous LPS removal or strain screening is essential to ensure the safety of GM-EVs therapy (Liang et al., 2025). Furthermore, caution is warranted when translating GM-EVs therapy into clinical practice. First, most of the current insights into the underlying mechanisms are derived from animal models, whose microbiome composition and immune responses differ significantly from those of humans. Second, animal studies typically use high concentrations of GM-EVs isolated in vitro, which may not accurately reflect the physiological doses that cross the human intestinal barrier. Finally, given the highly individualized nature of the human gut microbiome, the universality of the efficacy of single-strain GM-EV interventions remains a major challenge to be addressed in future clinical trials. GM-EVs represent a promising therapeutic avenue for metabolic and inflammatory bone disorders (as summarized in Figure 3; Table 3). It must be emphasized that direct experimental evidence demonstrating that GM-EVs regulate bone specifically through the CNS is currently lacking. Most mechanistic claims regarding this multi-organ axis are speculative and inferred from separate “gut-brain axis”, “brain-bone axis ”, and “gut-bone axis”. However, based on the established individual axes, an indirect “gut-brain-bone” pathway is mechanistically plausible: GM-EVs may initially modulate CNS function, which subsequently transmits altered neural signals to regulate bone metabolism (Li R. et al., 2024). Although an indirect “gut-brain-bone” pathway is mechanistically plausible, rigorous in vivo tracing and targeted knockout models are required to definitively prove this causal link. Furthermore, validating this multi-organ axis warrants further research to identify specific EVs cargos and their downstream neural targets. GM-EVs in the Regulation of Bone Homeostasis. In summary, this figure elucidates the complex multi-organ crosstalk within the “brain-gut-bone axis”, revealing how beneficial and pathogenic GM-EVs bidirectionally dictate bone remodeling by systematically integrating gut-derived metabolites (e.g., SCFAs, 5-HT, TMAO), immune responses, and central neuroendocrine signaling pathways. TABLE 3Summary of key cargo molecules in GM-EVs and their functions and regulatory mechanisms in skeletal diseases.EVs typeEVs sourceMolecular mediators in the EVsFunction and mechanism of actionReferencesGM-EVsGM-EVs Akkermansia muciniphila -Promotes bone mineralization and inhibits osteoclast activity to improve bone density Liu et al. (2021) Lactobacillus animalis -Promotes bone formation/angiogenesis and inhibits apoptosis to alleviate hormone-induced osteonecrosis Chen et al. (2022) Lactobacillus johnsonii -Inhibits mTORC1 to promote M2 macrophage polarization and ameliorate OA Liu R et al. (2023) Escherichia coli SOST siRNAActivates WNT pathway to promote OB differentiation and increase bone density Liu H et al. (2023) E. coli Nissle 1917 -Increases IL-10, reduces pro-inflammatory cytokines, and enhances tight junction proteins/TEER/hBD-2 to strengthen intestinal barrier Alvarez et al. (2016) E. coli O157:H7 FlagellinActivates NF-κB via TLR5 to induce IL-8 production and intestinal inflammation Bielaszewska et al. (2018) Propionibacterium freudenreichii -Inhibits pro-inflammatory factors and OCs formation to alleviat RA Woo et al. (2024) Proteus mirabilis -Downregulates miR-96-5p and upregulates Abca1 to induce OC apoptosis and mitochondrial dysfunction, improving OP Wang et al. (2022) Fusobacterium nucleatum FadAActivates Rab5a/YB-1 to promote IL-6/TNF-α and exacerbate OA; activates RIPK1 to induce oxidative stress and epithelial necroptosis, disrupting intestinal barrier Hong et al. (2023) Pseudomonas aeruginosa Quorum-sensing molecules (e.g., PQS)Participates in metabolic exchange and disrupts intestinal barrier integrity Liang et al. (2022) mTORC1, Mammalian Target of Rapamycin Complex 1; OA, Osteoarthritis; SOST siRNA, Sclerostin Small Interfering RNA; OB, Osteoblast; TEER, Transepithelial Electrical Resistance; hBD-2, Human Beta-Defensin 2; TLR5, Toll-Like Receptor 5; OCs, Osteoclasts; RA, Rheumatoid Arthritis; Abca1, ATP-Binding Cassette Subfamily A Member 1; OP, Osteoporosis; FadA, Fusobacterium Adhesin A; PQS, Pseudomonas Quinolone Signal. Summary of key cargo molecules in GM-EVs and their functions and regulatory mechanisms in skeletal diseases. mTORC1, Mammalian Target of Rapamycin Complex 1; OA, Osteoarthritis; SOST siRNA, Sclerostin Small Interfering RNA; OB, Osteoblast; TEER, Transepithelial Electrical Resistance; hBD-2, Human Beta-Defensin 2; TLR5, Toll-Like Receptor 5; OCs, Osteoclasts; RA, Rheumatoid Arthritis; Abca1, ATP-Binding Cassette Subfamily A Member 1; OP, Osteoporosis; FadA, Fusobacterium Adhesin A; PQS, Pseudomonas Quinolone Signal. Comparison of EVs from different sources EVs derived from mammalian cells, plants, and the gut microbiota share a common role as carriers of bioactive signals but differ substantially in origin, mechanism of action, and therapeutic applicability (as summarized in Table 4). MEVs offer high targeting precision and direct regenerative capacity, making them suitable for personalized and localised bone repair (Garcia et al., 2016; Kim et al., 2026; Zou et al., 2023). PDEVs provide scalable, low-toxicity, and pleiotropic anti-inflammatory and antioxidant effects, supporting their use as adjunctive therapies (Zeng et al., 2024; Zhang et al., 2025). GM-EVs uniquely exploit cross-system regulation via the “gut-bone axis”, enabling indirect but sustained modulation of systemic bone homeostasis (Liu H. et al., 2023; Cheung et al., 2022; Lu et al., 2025; Zu et al., 2024). TABLE 4Comparative analysis of mammalian-derived, plant-derived, and gut microbiota-derived EVs in bone health regulation.EVs sourceKey advantagesLimitationsApplicable bone disordersReferencesMEVsLow immunogenicity & High biocompatibilityNatural “homing” ability to injury sitesHigh potential for personalized engineeringLow natural yield & High costIsolation/purification is non-standardized and difficultChronic & Localized repair: OA, Chronic bone defects, OP, Bone tumors (Less practical for acute conditions) Yin et al. (2022), Thoo et al. (2019), Li R et al. (2024) PDEVsHigh yield, scalable, & cost-effectiveLow immunogenicity and no ethical concernsHigh GI stability (suitable for oral delivery)Intrinsic bioactivity (no loading needed)Low targeting specificityHigh batch-to-batch variationUndefined surface markers and isolation protocolsMulti-mechanistic & Adjunctive therapy: OA, Rotator cuff injuries, OP (Unclear potential for large defects/tumors) Liu et al. (2017), Lian et al. (2022), Zhang Y. W. et al. (2021) GM-EVsSystemic regulation via “gut-bone axis”High oral bioavailabilityStrain specificityHigh adaptability for drug loadingPotential toxicity/pathogenicityBioavailability and absorption pathways are poorly understoodLack of unified isolation protocolsSystemic & Metabolic Bone Diseases: OP (gut dysbiosis linked), Steroid-induced osteonecrosis, Early OA, Inflammatory Bowel Disease-associated bone loss, Autoimmune bone diseases (e.g., RA) Liu C et al. (2022), Liu et al. (2021), Chen et al. (2022), Bravo et al. (2011), Siva Venkatesh et al. (2024), Huang X et al. (2021), Sgritta et al. (2019), Gabanyi et al. (2016) GI, gastrointestinal; OA, osteoarthritis; OP, osteoporosis; RA, rheumatoid arthritis. Comparative analysis of mammalian-derived, plant-derived, and gut microbiota-derived EVs in bone health regulation. GI, gastrointestinal; OA, osteoarthritis; OP, osteoporosis; RA, rheumatoid arthritis. Currently, the main methods for isolating EVs include ultracentrifugation (UC), density gradient centrifugation (DGC), size-exclusion chromatography (SEC), tangential flow filtration (TFF), and immunoaffinity capture (Welsh et al., 2024). However, methods for isolating EVs vary depending on their source. MEVs are typically isolated using established UC, SEC, TFF protocols, which are relatively well-standardized (Auquière et al., 2025; Aliakbari et al., 2024). PDEVs require additional purification steps to remove interfering substances such as cellulose and pectin, often involving a combination of SEC and UC (Lv et al., 2024; Bokka et al., 2020). The isolation of GM-EVs poses the greatest challenge: these vesicles must be isolated from complex fecal matrices or culture media rich in heterologous proteins and LPS, and typically require a combination of DGC and ultrafiltration to achieve sufficient purity (Wu et al., 2024; Northrop-Albrecht et al., 2022). In addition, clinical translation requires selecting the appropriate route of administration based on the specific type of EVs. EVs can be administered via various routes, including intravenous, intraperitoneal, oral and local administration (such as intratumoral or intra-articular injection) (Su et al., 2025). In the treatment of skeletal disorders, MEVs can be delivered via local intra-articular injection or scaffold implantation to maximize local retention and minimize systemic clearance (Li et al., 2025; Huang and Xie, 2025). Due to their gastrointestinal stability, PDEVs are suitable for oral administration. Research evidence suggests that PDEVs can regulate intestinal homeostasis and the gut microbiota, which may enable non-invasive delivery via the “gut-bone axis” (Zhan et al., 2026; Guan et al., 2025). GM-EVs, particularly those derived from probiotics, naturally possess the ability to cross the intestinal barrier and are therefore equally suitable for oral administration (Liang et al., 2025; Wang K. et al., 2025). Although intravenous injection is widely used, EVs primarily accumulate in organs such as the liver, lungs and spleen, and their half-life is relatively short (Su et al., 2025). Therefore, choosing the appropriate route of administration is crucial. Future translation of EVs-based therapies will require standardized isolation and characterization protocols, improved understanding of in vivo biodistribution and targeting, and advanced engineering strategies to enhance safety and efficacy. Integrating synthetic biology, materials science, and precision medicine approaches will be essential to unlock the full therapeutic potential of EVs in bone diseases. Conclusion EVs derived from mammalian cells, plants, and the gut microbiota represent a versatile and biologically compatible therapeutic platform for bone diseases. Despite their shared function as carriers of bioactive molecules, these EVs operate through distinct regulatory paradigms. MEVs primarily enable local, high-precision modulation of bone metabolism, inflammation, and tissue regeneration. PDEVs exert pleiotropic anti-inflammatory and antioxidant effects that improve the bone microenvironment and support repair processes. GM-EVs uniquely integrate immune, metabolic, and microbial signals through the “gut-bone axis”, offering a systemic and cross-organ approach to skeletal regulation. Together, these complementary EVs sources establish a unified framework for EVs-based precision therapy in bone disorders, ranging from local degenerative diseases to systemic metabolic and inflammatory conditions. However, significant challenges remain, including the lack of standardized isolation and characterization methods, incomplete understanding of in vivo biodistribution and targeting, and unresolved safety concerns, particularly for microbiota-derived EVs. In addition, the long-term storage stability of EVs remains a significant barrier to clinical translation. Repeated freeze-thaw cycles and improper storage conditions can compromise vesicle integrity, reduce particle concentration, and diminish biological activity (Ahmadian et al., 2024). To overcome the aforementioned limitations in biodistribution and therapeutic efficacy, engineered EVs have been actively developed and investigated as a promising strategy (Malekian et al., 2023). Current modification strategies primarily involve surface modification, cargo loading and targeting peptides. Surface modification can be achieved via before-separation modification (e.g., genetic engineering, metabolic engineering, and direct parent cell membrane engineering) or post-separation modification (e.g., physical and chemical modifications). Before-separation modification preserves native EVs structure but requires complex cell engineering and suffers from batch variability, whereas post-separation modification offers greater flexibility but may compromise EVs integrity and surface functionality (Hu et al., 2025). For cargo loading, exogenous methods (e.g., electroporation, sonication, freeze-thaw, extrusion) are widely used; among them, freeze-thaw and hypotonic dialysis show superior loading efficiency for macromolecules like mRNA and proteins, yet all exogenous techniques face significant cargo loss, low reproducibility, and potential damage to EVs membranes. Endogenous loading via producer cells maintains EVs naturalness but offers poor controllability and is unsuitable for synthetic drugs (Zhan H et al., 2023; Mendonca et al., 2025). Targeting peptides involve displaying specific short peptides on the surface of EVs, enabling them to actively recognize and accumulate at lesion sites. Depending on the EVs membrane proteins utilized, a diverse array of targeting peptides can be displayed (e.g., RGD, GE11, RVG, iRGD). While peptides are low in immunogenicity and easy to synthesize, their optimal conformation, surface density, and in vivo stability remain challenging, and improper conjugation may impair EVs targeting (Song et al., 2022). Future progress will depend on elucidating source-specific mechanisms, identifying key functional cargoes, and optimizing EVs engineering strategies to enhance targeting accuracy and therapeutic efficacy. However, translating these mechanistic insights into clinical practice requires overcoming key translational bottlenecks, particularly the need for rigorous standardized protocols for EVs isolation, characterization, and storage, as well as the development of source-specific delivery strategies. In addition, advances at the intersection of extracellular vesicle biology, synthetic biology, and materials science are expected to accelerate clinical translation, positioning EVs-based therapies as a next-generation strategy for the treatment of bone diseases.