Y-RBCEVs accelerate fracture repair through osteogenesis and VCAN-mediated CD44/PI3K/AKT-dependent macrophage polarization Authors: Jiahao Wang, Xiaokang Zhang, Qi Zhang, Jiheng Xiao, Siteng Li, Jianpeng Chen, Chengbin Huang, Yang Luo, Zeming Shan, Ziping Chen, Ling Wang, Yingze Zhang Published: 2026 ABSTRACT AbstractObjectiveTo investigate the effects of young human red blood cell-derived extracellular vesicles (Y-RBCEVs) on fracture healing.MethodsAssociations of RBC-related indicators with bone mineral density and mortality in fracture populations were analyzed using NHANES data (1999–2020). A murine femoral fracture model was established to investigate RBC transfusion effects. Y-RBCEVs and O-RBCEVs were isolated to examine their effects on osteogenic differentiation and BMM polarization in vitro. In vivo distribution was assessed by IVIS, and therapeutic efficacy was evaluated in fracture mice. Proteomic sequencing identified differentially expressed proteins between Y-RBCEVs and O-RBCEVs, and key protein function was investigated by transcriptomic analysis and in vitro validation.ResultsThe observational study revealed associations between RBC count and bone mineral density with fracture populations mortality. RBC transfusion modulated the expression of osteogenic and M2 macrophage polarization-related genes. In vitro, RBCEVs were efficiently internalized by BMSCs and BMMs, with Y-RBCEVs enhancing osteogenic activity in BMSCs and promoting M2 macrophage polarization, thereby partially restoring osteogenesis impaired by inflammatory conditions. In vivo, Y-RBCEVs accumulated at fracture sites and enhanced bone regeneration without detectable toxicity, accompanied by increased OCN and CD206 expression. Proteomic analysis identified VCAN as one of the potential functional cargos. Mechanistically, Y-RBCEV-derived VCAN interacted with CD44 on macrophages and was associated with activation of the PI3K/AKT pathway, contributing to M2 polarization.ConclusionY-RBCEVs enhance fracture repair by modulating osteogenesis and macrophage polarization, with VCAN-associated CD44/PI3K/AKT signaling contributing to their osteo-immunomodulatory effects. These findings highlight RBCEVs as a potential therapeutic strategy for bone regeneration.Keywords: Red blood cell-derived extracellular vesicles, Fracture healing, Macrophage polarization, VCAN, PI3K/AKT signaling pathway Introduction Red blood cells are among the most frequently used blood products in clinical transfusion therapy. Transfusion of red blood cells can effectively correct anemia in patients with trauma and fractures, thereby improving clinical outcomes—for example, by shortening fracture-healing time and lowering the risk of infection [1,2]. Research indicates that blood transfusion enhances oxygen delivery to local tissues and modulates systemic immunity [3], thereby promoting fracture healing and tissue repair. However, aside from their well-known oxygen-carrying function, red blood cells also serve as carriers rich in bioactive substances. Yet their potential additional roles in fracture repair remain largely unexplored. Extracellular vesicles (EVs) have emerged as critical mediators of intercellular communication, capable of precisely modulating recipient cell function [[4], [5], [6], [7]]. In the context of bone repair, EVs derived from mesenchymal stem cells and platelets enhance healing by promoting angiogenesis and modulating immune responses [[8], [9], [10], [11]]. Similarly, M2 macrophage-derived EVs accelerate fracture healing through anti-inflammatory effects and the stimulation of osteogenic differentiation [12,13]. These observations raise the question of whether red blood cells (RBCs) influence distant tissues via their own EVs, red blood cell-derived extracellular vesicles (RBCEVs). To date, research on RBCEVs has primarily focused on their utility as drug delivery vehicles, capitalizing on their translational advantages: clinical safety, ease of isolation, and high yield [[14], [15], [16], [17]]. For example, intramuscularly injected RBCEVs can deliver siRNA to treat cancer cachexia-induced muscle atrophy, a strategy noted for its favorable safety profile, low immunogenicity, and enhanced bioavailability [18]. Engineered RBCEVs have also been designed as dual-delivery systems to co-administer doxorubicin and siRNA, enabling combined chemotherapy and gene regulation in acute myeloid leukemia [19]. Besides, RBCEVs can be taken up by macrophages, release heme intracellularly, inhibit pro-inflammatory cytokine production, and suppress foam cell formation [20], suggesting a potential role in immunomodulation. Nevertheless, whether RBCEVs influence bone metabolism or contribute therapeutically to fracture repair remains unexplored. Fracture healing is greatly influenced by two coordinated biological processes: the migration and osteogenic differentiation of bone marrow-derived mesenchymal stem cells (BMSCs) [[21], [22], [23]], and the dynamic polarization of macrophages within the local immune microenvironment from a pro-inflammatory (M1) to a reparative (M2) phenotype [24,25]. BMSCs are recruited to the fracture site, where they proliferate and differentiate into osteoblasts and chondrocytes, forming the callus that bridges the fracture gap. Concurrently, the timely shift from M1-to M2-dominant macrophage populations facilitates healing by promoting stem cell recruitment, angiogenesis, and tissue remodeling [[26], [27], [28]]. Building on this framework, we hypothesized that RBCEVs contribute to fracture repair by regulating BMSC osteogenesis and macrophage polarization. Using in vitro and in vivo models, we investigated the therapeutic effects and molecular mechanisms of RBCEVs. Our findings provide a potential vesicle-mediated explanation for the association between RBC transfusion and fracture recovery, highlighting RBCEVs as promising candidates for bone regeneration. Methods Cells and animals 13-week-old male mice were purchased from Sipeifu Biotech (Beijing, China) and housed in the Animal Experiment Center of The Medical University. All animal experiments were performed with the approval of the Medical Ethics Committee of The Medical University (Ethics Approval No. Z2024-039-1). To obtain bone marrow-derived macrophages (BMM) and bone marrow mesenchymal stem cells (BMSC), 4-6-week-old male C57BL/6J mice were used. Mice were euthanized and disinfected, and the tibias and femurs were exposed and isolated. Muscle tissues were removed, both ends of the femurs and tibias were cut off, and bone marrow was flushed out using a syringe filled with MEMα complete medium (PM150421B, Wuhan Procell Biotechnology Co., Ltd.). The cell suspension was pipetted, filtered through a 70 μm mesh filter, resuspended, and cultured in 10 cm cell culture dishes. For BMSC isolation, the culture supernatant was discarded after 12 h of cell adherence, fresh MEMα complete medium was added, and cells were further cultured until reaching 80% confluence for subsequent experiments. For BMM isolation, non-adherent cells in the supernatant were collected after 24 h of cell adherence and cultured in DMEM/F-12 complete medium (PM150310B, Wuhan Procell Biotechnology Co., Ltd.) supplemented with 25 ng/mL M-CSF (315-02-10UG, Gibco). The medium was changed every two days, and cells were used for experiments once they reached 80% confluence. Isolation and characterization of RBCEVs Red blood cells (RBCs) were isolated from fresh whole blood samples obtained from volunteers at the Third Hospital of Hebei Medical University, with the approval of the Medical Ethics Committee of the University (Ethics Approval No. 2019-021-1). After obtaining ethical approval and informed consent, volunteers were recruited according to age groups, including young adults (18–35 years), middle-aged adults (36–60 years), and older adults (61–75 years). All participants were generally healthy individuals without known chronic diseases or clinically significant medical conditions. Participants were excluded if they had acute infection, fever, common cold, diarrhea, or other inflammatory conditions within the previous 2 weeks; autoimmune diseases; malignant tumors; hematological disorders; or long-term immune dysfunction. These criteria ensure that the collected samples reflected the physiological characteristics associated with aging rather than alterations caused by acute or chronic pathological conditions. Peripheral whole blood was collected from 6 healthy volunteers in each of three age groups. For RBCEVs isolation, whole blood was centrifuged at 4000 rpm for 5 min to remove plasma. Red blood cells were washed twice with PBS to eliminate adsorbed plasma proteins. The washed RBCs were incubated in DMEM medium supplemented with 0.1 mg/mL CaCl2 (C7250, Beijing Solarbio Science & Technology Co., Ltd.) and 10 μM Calcium ionophore I (HY-136460, MedChemExpress, China) at 37°C for 24 h. The supernatant was collected and centrifuged repeatedly at 4000 rpm for 5 min to obtain a clarified supernatant, which was then concentrated using a 30 kDa ultrafiltration tube. Exosome isolation reagent (EXOTC10A-1, System Biosciences) was added at one-fifth the volume of the concentrate, mixed thoroughly, and incubated overnight at 4°C. After centrifugation at 1500g for 30 min at room temperature, the pellet was collected and diluted with normal saline. The mixture was centrifuged at 4000 g for 20 min to remove cell debris. The supernatant was further centrifuged at 10000 g for 20 min to remove large vesicles, and the final supernatant was collected as purified RBCEVs. Quantification was performed using a BCA protein assay kit (BL367A, Biosharp, China). The morphology, particle size, zeta potential, and expression of exosomal marker proteins (CD9 (+), TSG101 (+), and Hemoglobin (HBA) (+)) were characterized by scanning electron microscopy (SEM), nanoparticle tracking analysis (NTA), and Western blot. The uptake of RBCEVs by BMSC and BMM was assessed using a cellular uptake assay. RBCEVs were aliquoted and stored at −80°C. Y-RBCEVs were defined as extracellular vesicles derived from healthy young adult donors aged 18–35 years. O-RBCEVs were defined as extracellular vesicles derived from healthy older adult donors aged 61–75 years. Exosome internalization assay BMM and BMSC were seeded in glass-bottom cell culture dishes (801021, NEST Biotechnology) at a density of 50,000 cells per dish. After overnight incubation, RBCEVs labeled with Dio green fluorescent dye (HY-D0969, MedChemExpress, China) were added to each dish at a concentration of 100 μg/mL. Following 4 h of uptake, the supernatant was discarded, and uninternalized RBCEVs were removed by washing with PBS. After fixation, nuclei were stained with DAPI working solution (HY-D0814, MedChemExpress, China) to blue. After washing with PBS, cellular uptake was observed under a confocal microscope (Nikon Instruments, Melville, NY, USA). The fluorescence intensity of internalized RBCEVs was quantified using ImageJ. In addition, BMM and BMSC were seeded in 6-well plates at a density of 150,000 cells per well and pretreated for 1 h with inhibitors targeting different endocytic pathways, including 50 μM Dynasore (dynamin activity inhibitor, HY-15304, MedChemExpress, China), 50 μM EIPA (macropinocytosis inhibitor, HY-101840, MedChemExpress, China), and 3 mM MβCD (lipid raft inhibitor, HY-101461, MedChemExpress, China). After pretreatment, RBCEVs labeled with DiD blue membrane dye (HY-D1028, MedChemExpress, China) were added to each well at 100 μg/mL. Cells were harvested at 1 h, 2 h, and 4 h after incubation, and the percentage of uptake-positive cells was determined by flow cytometry. In vitro proliferation and migration assays EdU proliferation assay BMSC were seeded into 24-well plates at a density of 1 × 105 cells per well. After cell adherence, the medium was replaced with serum-free medium, and 100 μg/mL Y-RBCEVs, O-RBCEVs, or 100 ng/mL VCAN were added for 24 h of treatment. Following treatment, the medium was changed to complete medium containing the 5-ethynyl-2′-deoxyuridine (EdU) labeling reagent, and the cells were incubated for 6 h in the dark. After fixation and permeabilization, cell proliferation was detected using the BeyoClick™ EdU Cell Proliferation Kit with AF594 (C0078L, Beyotime). Nuclei were stained with Hoechst 33342 (C1022, Beyotime). Images were observed under a fluorescence microscope (IX73, Olympus Corporation, Japan) and quantitatively analyzed using ImageJ. Wound scratch assay BMSC were seeded in 6-well plates. When cell density reached 90%, a scratch of approximately 1 mm width was made in the cell monolayer. Loose cells were removed by PBS washing, and the medium was replaced with fresh complete medium containing 2% serum. Cells were treated with 100 μg/mL Y-RBCEVs, O-RBCEVs, or 100 ng/mL VCAN for 24 h. The initial scratch width at 0 h was recorded, and cells were cultured continuously for 24 h with photographic documentation. Scratch widths were measured in ImageJ, and the healing rate was calculated. Transwell assay BMSC were seeded into the upper chambers of 24-well Transwell plates using serum-free medium at a density of 10,000 cells per well. Complete serum-containing medium was added to the lower chambers, and 100 μg/mL Y-RBCEVs, O-RBCEVs, or 100 ng/mL VCAN were added to the upper compartment according to the experimental groups for 24 h of treatment. Cells were fixed with 4% paraformaldehyde, stained with 0.1% crystal violet for 10 min, and washed thoroughly with PBS. Cell migration was observed under an inverted microscope, and quantification was performed using ImageJ software. In vitro osteogenesis and macrophage polarization assays BMSC were seeded into 24-well plates. When cell density reached more than 80%, osteogenic differentiation was induced using osteogenic differentiation medium (MUXMX-90021, Cyagen, OriCell) for 7 days, followed by alkaline phosphatase staining (C3250S, Beyotime). 100 μg/mL of Y-RBCEVs or O-RBCEVs were added into the differentiation medium. After 21 days of osteogenic induction, osteogenic calcium nodules were stained with alizarin red staining solution (T240924C301, OriCell). For Western blot detection, cell samples were collected for protein extraction on day 14 of osteogenic induction. For the lipopolysaccharide (LPS)-induced macrophage inflammation model, BMM were treated with 50 ng/mL LPS (L8880, Beijing Solarbio Science & Technology Co., Ltd.) for 24 h. The medium was then replaced with fresh medium containing 100 μg/mL of Y-RBCEVs or O-RBCEVs, and cells were cultured for an additional 24 h. After treatment, the expression of target genes or proteins was determined by qPCR, flow cytometry, immunofluorescence, and Western blot. After LPS stimulation, cells were treated with PBS, RBCVEVs, or indicated drugs, following which culture supernatants were used as conditioned media. To investigate the indirect regulatory effect of macrophage polarization on osteogenic differentiation, conditioned medium from treated BMM were collected, mixed with osteogenic differentiation medium at a 1:1 vol ratio, and used for osteogenic induction. For CD44 blockade, BMM were treated with LPS for 24 h. The medium was then replaced with fresh medium, and 10 μg/mL anti-CD44 antibody (A2174, Selleck) was added for 1 h of pretreatment. Afterward, the medium was changed to fresh medium supplemented with Y-RBCEVs or VCAN. For AKT signaling blockade, 1 μM MK2206 (HY-108232, MedChemExpress, China) was co-administered with Y-RBCEVs or VCAN to BMM for 24 h. For VCAN blockade in BMM and BMSC treatment, Y-RBCEVs were pre-incubated with 200 ng/mL VCAN-neutralizing antibody (Cell Signaling Technology, 98586) for 1 h and then added together to cells. The proteins distinguished by proteomic sequencing analysis were administrated in osteogenesis and macrophages polarization. The detailed information was as follows: EPB41(Hubei Ipodix Biotechnology Co.,Ltd. PA2000-716DB), VSIG8(MedChemExpress, China, HY-P71426), OLFM1(Wuhan Fine Biotech Co., Ltd. P6504), FGFR4(MedChemExpress, China, HY-P73057), PROZ(Shanghai Lianmai Biotechnology Co., Ltd. LM736Mu01), VCAN(MedChemExpress, China, HY-P702616), PTPRG(MedChemExpress, China, HY-P701970), TANGO2(Hubei Ipodix Biotechnology Co.,Ltd. PA2000-62122), and DBH(Hubei Ipodix Biotechnology Co.,Ltd. PA1000-317DB). Immune fluorescence staining After being washed and fixed, cells were permeabilized with 0.1% Triton X-100 (T8200, Beijing Solarbio Science & Technology Co., Ltd.) at room temperature for 20 min, followed by blocking with 5% bovine serum albumin (BSA) (4240GR100, BioFroxx) at 37°C for 1 h. After washing with PBST, appropriately diluted primary antibodies were added and incubated overnight at 4°C. Primary antibodies were as follows: CD86 Rabbit antibody (YT7823, Immunoway, 1:100), CD206 Rabbit antibody (YT5640, Immunoway, 1:100), and VCAN Rabbit antibody (30599-1-AP, Proteintech, 1:100). After incubation, cells were washed three times with PBST, then incubated with appropriately diluted secondary antibody Goat anti-Rabbit IgG (H + L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 594 (A-11012, Invitrogen, 1:2000) and Actin-Tracker Green-488 (C2201S, Beyotime, 1:200) at room temperature for 2 h. The staining solution was discarded, and DAPI staining solution (HY-D0814, MedChemExpress, China) was added and incubated for 10 min at room temperature in the dark. After thorough washing with PBS, cells were observed under a fluorescence microscope. Flow cytometry detection For in vitro experiments, cells were treated, washed thoroughly, collected into flow tubes, and stained directly. For in vivo experiments, callus tissues were harvested and prepared into single-cell suspensions. Red blood cells were removed using red blood cell lysis buffer (R1010, Beijing Solarbio Science & Technology Co., Ltd.). After cell counting, an appropriate number of cells were transferred to flow tubes for further staining. For macrophage polarization detection, TruStain FcX™ PLUS (anti-mouse CD16/32) (156604, Biolegend) was added at a ratio of 1 μg per 106 cells, followed by incubation on ice for 10 min. Surface marker antibodies were then added and incubated on ice for 1 h, and cells were washed three times with pre-chilled PBS.BD Cytofix™ Fixation Buffer (554655, BD Biosciences) was added for permeabilization on ice for 40 min. Cells were washed three times with a 1:10 dilution of BD Perm/Wash™ Buffer (10×) (554723, BD Biosciences), then incubated with intracellular antibodies on ice for 1 h. After final washing with 1× Wash Buffer, samples were analyzed on a CytoFLEX LX flow cytometer (Beckman Coulter, China). For macrophage polarization detection, the antibodies used were as follows: PerCP CD86 Antibody (105025, Biolegend, 1 μg/106 cells), PE anti-mouse F4/80 (111603, Biolegend, 1 μg/106 cells), and APC CD206 antibody (321109, Biolegend, 1 μg/106 cells). For hematopoietic differentiation detection, the antibodies used were as follows: APC Sca-1 (17-5981-83, eBioscience, 1.5 μL/106 cells), BV421 CD117 (562609, BD Biosciences, 1 μL/106 cells), PE CD34 (551387, BD Biosciences, 1 μL/106 cells), FITC CD16/32 (561728, BD Biosciences, 1 μL/106 cells), and PerCP-Cy5.5 lineage (561317, BD Biosciences, 20 μL/106 cells). QPCR Total RNA was extracted from cells and peripheral blood samples using an RNA extraction kit (DP419, Tiangen Biotech, Beijing). RNA concentration was quantified using the NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA). mRNA was reverse-transcribed into cDNA using a reverse transcription kit (ZS-M14003, Zhongshi TONTRU, China), and quantitative real-time PCR was performed on a LightCycler 480 II system using Superbrilliant® 5×Fast SYBR Green qPCR Mix (ZS-M13002, Zhongshi TONTRU, China). Gene expression levels were calculated following the 2−ΔΔCt method and normalized to the reference gene GAPDH. Primer sequences were listed as follows. PrimerSequencesGAPDH-F5′-GGTTGTCTCCTGCGACTTCA-3′GAPDH-R5′-TGGTCCAGGGTTTCTTACTCC-3′TNF-α-F5′-AGCCGATGGGTTGTACCTTG-3′TNF-α-R5′-ATAGCAAATCGGCTGACGGT-3′IFN-γ-F5′-ATGAACGCTACACACTGCATC-3′IFN-γ-R5′-CCATCCTTTTGCCAGTTCCTC-3′CD86-F5′-TCTCCACGGAAACAGCATCT-3′CD86-R5′-CTTACGGAAGCACCCACGAT-3′ARG1-F5′-AAGAATGGAAGAGTCAGTGTGG-3′ARG1-R5′-GGGAGTGTTGATGTCAGTGTG-3′IL4-F5′-CTGTCTGATTTTGCTGTTGGTG-3′IL4-R5′-GATGTAGTCAGAGAAGCAGGTG-3′CD206-F5′-CCTATGAAAATTGGGCTTACGG-3′CD206-R5′-CTGACAAATCCAGTTGTTGAGG-3Col1-F5′-CTGGCGGTTCAGGTCCAAT-3′Col1-R5′-CCACGTCTCACCATTGGGG-3′Runx2-F5′-ATGCTTCATTCGCCTCACAAA-3′Runx2-R5′-GCACTCACTGACTCGGTTGG-3′OCN-F5′-CTGACCTCACAGATCCCAAGC-3′OCN-R5′-TGGTCTGATAGCTCGTCACAAG-3′ Immune blotting Total protein was extracted from RBCEVs or cell samples by lysis on ice using RIPA lysis buffer (P0013B, Beyotime). Membrane proteins were isolated and extracted using a Membrane and Cytosol Protein Extraction Kit (PC202, Yamei, China). Protein concentration was determined using a BCA protein assay kit (BL521A, Biosharp, China). After SDS-PAGE electrophoresis, membrane transfer, and blocking, primary and secondary antibodies were sequentially incubated. Protein bands were visualized using ultra-sensitive ECL chemiluminescence reagent (BL520B, Biosharp, China) and detected with a ChemiDoc™ imaging system (Bio-Rad). Primary antibodies used were as follows: TSG101 Rabbit antibody (28283-1-AP, Proteintech, China), CD9 Rabbit antibody (20597-1-AP, Proteintech, China), HBA (YN2315, Immunoway, 1:2000), VCAN Rabbit antibody (30599-1-AP, Proteintech, China), PTPRG Rabbit antibody (PA5-15524, Invitrogen, USA), VSIG8 Rabbit antibody (K007223P, Beijing Solarbio Science & Technology Co., Ltd.), ATP1A1 Rabbit antibody (14418-1-AP, Proteintech, China), ACTIN Mouse antibody (66009-1-Ig, Proteintech, China), GAPDH Rabbit antibody (10494-1-AP, Proteintech, China), PI3 Kinase Mouse antibody (60225-1-Ig, Proteintech, China), Phospho-PI3 Kinase Rabbit antibody (HY-P80846, MedChemExpress, China), Phospho-AKT Rabbit antibody (29163-1-AP, Proteintech, China), AKT Rabbit antibody (10176-2-AP, Proteintech, China), iNOS Rabbit antibody (22226-1-AP, Proteintech, China), ARG1 Rabbit antibody (ab96183, Abcam, China), CD44 Rabbit antibody (30854-1-AP, Proteintech, China). Secondary antibodies used were as follows: Multi-rAb® HRP-Goat Anti-Rabbit Recombinant Secondary Antibody (H + L) (RGAR001, Proteintech, China), Multi-rAb® HRP-Goat Anti-Mouse Recombinant Secondary Antibody (H + L) (RGAM001, Proteintech, China). Evaluation of RBCEVs treatments in vivo To investigate the regulatory effects of RBCEVs on the fracture repair process in vivo, this study used the femoral fracture model with intramedullary fixation, a well-established and highly reproducible animal model that has been widely applied to evaluate the effects of different interventions on fracture healing [29,30]. Thirteen-week-old male C57BL/6J mice were anesthetized and fixed in the supine position on a surgical table. After skin preparation and disinfection, an incision was made along the lateral skin of the left femur, and muscles were bluntly separated along the intermuscular space to expose the femur. A sterile No. 11 scalpel was used to create a fracture at the midshaft of the femur, followed by internal fixation with a sterile 25-gauge stainless steel needle inserted through the intercondylar fossa of the distal femur. Muscles and skin were sutured layer by layer. According to different treatments, mice were divided into the NS group, the Y-RBCEVs group, and the O-RBCEVs group. Mice were treated three times per week at a dose of 100 μg per 100 μL per administration for 2 consecutive weeks. 2-consecutive-weeks is a commonly used time point for evaluating fracture healing in mice, with previous studies assessing bone regeneration through micro-CT, histological staining, and immunohistochemical analyses at this stage [13,31,32]. After treatment, the mice were euthanized, and the femoral tissues were harvested. Micro-CT (Bruker, Germany) was performed to evaluate bone callus mass, and bone volume/total volume (BV/TV), bone surface area/bone volume (BS/BV), trabecular number (Tb.N), trabecular separation (Tb.Sp), and trabecular thickness (Tb.Th) were quantified. In addition, hematoxylin–eosin (HE) staining and safranin O-fast green staining were performed to observe morphological changes at the femoral metaphysis. Callus (callus area/total tissue area) and new bone (new bone area/total tissue area) were quantified by ImageJ software. Immunofluorescence staining was conducted to detect the expression of the osteogenic marker OCN and the anti-inflammatory macrophage marker CD206 at the metaphysis and proximal femur. Femoral callus tissues were collected, and flow cytometry was used to analyze the proportion of macrophage polarization and hematopoietic cell differentiation in the callus. Histological examination Briefly, femoral tissues were fixed, decalcified, embedded in paraffin, and sectioned. After dewaxing with xylene and rehydration with graded ethanol, sections were stained using an H&E staining kit (G1076-500 mL, Servicebio, China) and a Safranin-O and Fast Green Stain Kit (G1371, Beijing Solarbio Science & Technology Co., Ltd.). Subsequently, the sections were dehydrated in graded ethanol, cleared in xylene, and mounted in neutral balsam. After air-drying at room temperature, sections were observed under a light microscope. Immunofluorescence staining of tissue As previously described, callus tissue sections were dewaxed, rehydrated, and subjected to antigen retrieval. Sections were blocked with Normal Goat Serum (SL038, Beijing Solarbio Science & Technology Co., Ltd.) and permeabilized with 0.5% Triton X-100. After washing with TBST, sections were incubated with primary antibodies overnight at 4°C. Following TBST washing, fluorescently labeled secondary antibodies were applied. The antibodies used were as follows: F4/80 Rat antibody (GB12027-100, Servicebio, China), CD206 Rabbit antibody (YT5640, Immunoway), Osteocalcin Rabbit antibody (GB11233-100, Servicebio, China), Goat anti-Rabbit IgG (H + L) Cross-Adsorbed Secondary Antibody (A-11008, Invitrogen, Alexa Fluor™ 488), Goat anti-Rat IgG (H + L) Highly Cross-Adsorbed Secondary Antibody (A48264, Invitrogen, Alexa Fluor™ Plus 594). After thorough washing with TBST, sections were counterstained and mounted with DAPI-containing mounting medium (G1407-25 ML, Servicebio, China). Images were observed under a fluorescence microscope. Fluorescent secondary antibodies labeled the target proteins, and nuclei were stained blue. The relative percentage of fluorescent-positive areas was quantified using ImageJ. In vivo RBCEVs distribution and biosafety evaluation assays Femoral fracture mouse models were established as described above. RBCEVs were labeled with the cell membrane fluorescent dye DID, and PBS solution containing an equivalent amount of DID was prepared. The stained exosomes were washed and collected using 30 kDa ultrafiltration tubes. Fracture mice to be observed were divided into three groups: DID group, DID-labeled Y-RBCEVs group, and DID-labeled O-RBCEVs group, with 6 mice per group. Each mouse was injected with 100 μL of the corresponding solution. Fluorescence distribution images of bilateral femurs were collected at 6 h, 12 h, 24 h, 36 h, 48 h, and 72 h post-injection. After in vivo imaging at 72 h, major organs, including the heart, liver, spleen, lung, kidney, and the left lower limb, were harvested for ex vivo fluorescence analysis. To evaluate the in vivo biosafety of RBCEVs, mice were intravenously treated with PBS, Y-RBCEVs, or O-RBCEVs. After treatment, peripheral whole blood was collected and submitted to Ruiweiyouke Biotechnology Co., Ltd. (Hangzhou, China) for the detection of biochemical indicators, including urea (UREA), uric acid (UA), albumin (ALB), alanine aminotransferase (ALT), creatinine (CRE), and aspartate aminotransferase (AST). In addition, major organs, including the heart, liver, spleen, lungs, and kidneys, were harvested and stained with HE to assess whether exosome treatment altered histological morphology in mice. Proteomic and RNA sequencing analysis Proteomic analysis was used to distinguish protein contents between Y-RBCEVs and O-RBCEVs. Briefly, the total protein from exosome samples was extracted using RIPA. Sequencing analysis was performed on the timsTOF platform with assistance from OE Biotech (China). RNA sequencing was used to detect differentially expressed mRNAs in cells treated with or without VCAN after LPS treatment. The total RNA was extracted for the construction of cDNA libraries. RNA sequencing analysis was performed on the Illumina Novaseq 6000 platform with the assistance of OE Biotech, Inc. (Shanghai, China). Prediction and analysis of protein-protein interaction targets Interacting proteins of VCAN were predicted using the online databases GENEMANIA (https://genemania.org), STRING (https://cn.string-db.org), BIOGRID (https://thebiogrid.org), HINT (http://hitpredict.org), and Hitpredict (https://www.hitpredict.org/). CO-IP Cells in each group were collected and lysed in lysis buffer containing protease inhibitors on ice for 30 min. After centrifugation at 12,000 g for 15 min at 4°C, the supernatant was transferred to a new centrifuge tube. The protein supernatant was incubated with CD44 antibody and Protein A/G Magnetic Beads (HY-K0202, MedChemExpress, China) with constant mixing on a vertical shaker at 4°C overnight. An equal amount of Rabbit Immunoglobulin G (IgG) (A7058-2 mg, Beyotime) was used as a negative control. Beads were harvested using a magnetic stand, the supernatant was discarded, and the beads were washed three times with pre-chilled washing buffer to remove unbound nonspecific proteins. 2×SDS loading buffer was added to the beads, followed by boiling in a water bath for 5 min to dissociate protein complexes from the beads. Protein expression was detected by Western blot and quantified using ImageJ. National health and Nutrition Examination Survey (NHANES) based analysis Population The observational study used publicly available data from the NHANES database (https://www.cdc.gov/nchs/nhanes). Participants from 10 NHANES survey cycles (1999-2000, 2001-2002, 2003-2004, 2005-2006, 2007-2008, 2009-2010, 2011-2012, 2013-2014, 2015-2016, 2017-2020) were included. All NHANES protocols received ethical approval from the National Center for Health Statistics (NCHS) Ethics Review Board. Written informed consent was obtained from each participant before the survey. Mortality records were obtained from the National Death Index (NDI) (https://www.cdc.gov/nchs/ndi/). Variables and measurements Femur bone mineral density (BMD, gm/cm2) was measured by the Dual-energy x-ray absorptiometry method. The examinations were conducted by trained and certified radiology technologists. The complete blood count (CBC) data recording information of red blood cells (RBCs), hemoglobin (HGB), and Hematocrit (HCT) were detected based on the Beckman Coulter methodology in the Beckman Coulter DxH 800 instrument in the NHANES mobile examination center. The laboratory staff members were formally trained and observed during operations to ensure the quality of their work. Fracture population was defined as those with self-reported hip, wrist, or spine fractures in the home questionnaire administered using the Computer-Assisted Personal Interview system. Information on age, gender, race, educational level, marital status, body mass index (BMI), total energy intake, physical activity, smoking status, alcohol consumption, diabetes, and hypertension for participants was collected and used as covariates in the statistical analysis. The race information included Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black, and other Race. The education levels included Less Than 9th Grade, 9-11th Grade, High School Graduate, Some College or AA degree, College Graduate or above, and Refused. The marital statuses were married, widowed, divorced, separated, never married, and living with a partner. The BMI index was calculated by dividing weight by the square of height (kg/m2). Total energy intake values were obtained based on the baseline 24h dietary recall assessment. Physical activity was assessed through a systematic evaluation of performance. The alcohol drinking and smoking states were determined according to the questionnaire answers. A systolic blood pressure of more than 140 mmHg or a diastolic blood pressure of more than 90 mmHg was used to confirm the diagnosis of hypertension. Diabetes was diagnosed following questionnaire answers, fasting plasma glucose levels≥7.0 mmol/L, or usage of antidiabetic medications. Statistical analysis All results shown in cellular and animal experiments were reported as mean ± standard deviation (mean ± SD). Statistical analyses were conducted on at least 3 biological replicates. ImageJ software was used to quantify protein expression levels and calculate the positive area. Data were statistically analyzed using GraphPad Prism software (version 8.0). A Student's t-test was used for comparisons between two groups, and a one-way analysis of variance (ANOVA) was used for comparisons among more than two groups. A P value < 0.05 was considered statistically significant. For the observational study, participants were grouped by sex, and the associations between RBC-related variables and BMD or mortality were analyzed separately. The differences in variables between groups were analyzed using the Chi-square test for categorical data and the Kruskal-Wallis test for quantitative data. The quantile logistic regression method was used, grouping participants according to RBC-relevant variables. In the Non-adjusted model, no variables were adjusted. In the adjusted model, the analysis was adjusted for age, race, educational level, marital status, BMI, moderate work activity, hypertension, diabetes, total energy intake, smoking, and drinking. In addition, threshold-effect analysis and smooth curve fitting were conducted to visualize the relationship further. P < 0.05 was considered statistically significant for all statistical analyses. EmpowerStats (https://www.empowerstats.net/cn/) and the related R package (https://www.r-project.org/) were used together for data analysis. The Package R was used for reading raw data, and the Package R was used for data analysis. Results In this study, NHANES data were used to investigate the associations of RBC-related indicators with BMD and mortality in fracture populations. As shown in Fig. 1A, 45,496 participants from 10 NHANES cycles (1999–2020) with complete BMD measurements and hematological data (RBC count, HGB, and HCT) were included, comprising 23,768 males and 21,728 females. Among them, 3294 individuals with fracture, available mortality data, and RBC-related parameters were identified (1796 males and 1498 females). Baseline characteristics are summarized in Supplementary Tables 1 and 4 Before adjustment, higher RBC count was associated with increased BMD in both sexes. Higher HGB and HCT levels were positively associated with BMD in males but showed inverse associations in females. After multivariable adjustment, males in the Q2 RBC group exhibited higher BMD than those in Q1, whereas no significant association was observed among female RBC groups. In males, Q2–Q3 HGB and Q2 HCT groups showed higher BMD than Q1. Conversely, females in the highest HGB quartile (Q4, 14.1–19.6 g/dL) exhibited lower BMD compared with Q1. In the fracture population, higher RBC counts were associated with lower risks of all-cause mortality among males in the Q2 and Q3 groups, while no significant associations were observed in females. HGB and HCT were not significantly associated with mortality in either sex (Supplementary Tables 2 and 5). Association of RBC-related indicators with BMD and fracture repair. (A) Flowchart of the NHANES-based observational study. (B) Smooth curve analyses showing the associations of RBC-related indicators (RBC, HGB, and HCT) with BMD in the general U.S. population and all-cause mortality in fracture populations. Inflection points and P values from threshold effect analyses are indicated. (C) Schematic illustration of RBC transfusion in a murine femoral fracture model. (D) Expression of osteogenic markers (COL-I, RUNX2, and OCN) and macrophage polarization-related genes (IFN-γ, TNF-α, CD86, IL-4, CD206, and Arg1) in peripheral blood after treatment. Data are presented as mean ± SD. *P < 0.05; NS, not significant. Smooth curve fitting and threshold analyses (Supplementary Tables 3 and 6; Fig. 1B) revealed nonlinear associations between RBC-related indicators, BMD, and mortality. RBC count was associated with larger BMD before reaching inflection points (4.66 × 1012 cells/mL in males and 3.83 × 1012 cells/mL in females), followed by saturation effects. Similarly, higher RBC count was associated with lower mortality risk in male fracture patients below the inflection point (4.47 × 1012 cells/mL). HGB showed threshold effects in males, with higher levels associated with higher BMD and lower risks of mortality before 14.1 g/dL, followed by saturation. No significant threshold effects of HGB were observed in females or fracture mortality. HCT showed similar threshold effects in males, with inflection points of 41.5% for BMD and 41.3% for mortality. In females, HCT levels above 44.1% and 45.4% were associated with lower BMD and increased mortality risk, respectively (normal HCT range in females: 36%-45%). Fig. 1C illustrates the RBC transfusion protocol in femoral fracture mice. RBC transfusion increased the expression of osteogenic markers (COL-I, OCN, and RUNX2) and M2 macrophage markers (CD206, IL-4, and Arg1) in peripheral blood, while showing no significant effects on pro-inflammatory markers (CD86, IFN-γ, and TNF-α) (Fig. 1D). These findings suggest that RBCs may modulate systemic immune responses and bone repair. To characterize RBCEVs and explore their biological functions in bone regeneration, Y-RBCEVs and O-RBCEVs were isolated and analyzed. SEM revealed that both types of RBCEVs exhibited typical spherical or cup-shaped morphology (Fig. 2A). NTA showed that Y-RBCEVs and O-RBCEVs had average diameters of approximately 100 nm and 120 nm, respectively (Fig. 2B). Western blot analysis confirmed the expression of EV markers (TSG101 and CD9), the RBC-specific marker HBA, and the lack of cell marker ACTIN in both RBCEVs groups (Fig. 2C). Internalization assays in Fig. 2D and E confirmed that both Y-RBCEVs and O-RBCEVs were efficiently internalized by BMSCs and BMMs, with no significant difference in uptake efficiency between the two groups (Fig. 2D–F). Time-course analysis showed increased RBCEV uptake by both cell types, with BMMs exhibiting higher uptake efficiency than BMSCs. Mechanistic analysis using endocytosis inhibitors indicated that dynamin-dependent pathways contributed substantially to RBCEV internalization, while micropinocytosis (inhibited by EIPA) and lipid raft-mediated pathways (inhibited by MβCD) played limited roles (Supplementary Fig. 1). Isolation and functional characterization of RBCEVs. (A) Morphology of Y-RBCEVs and O-RBCEVs analyzed by scanning electron microscopy (SEM). Scale bar = 100 nm. (B) Particle size distribution determined by nanoparticle tracking analysis (NTA). (C) Expression of RBCEV markers (TSG101, CD9, and HBA) detected by Western blotting. (D-E) Uptake of Y-RBCEVs and O-RBCEVs by bone marrow-derived mesenchymal stem cells (BMSCs) and macrophages (BMMs). Scale bar = 10 μm. (F) Quantification of fluorescence intensity by ImageJ. (G) EdU assay showing the effects of RBCEVs on BMSC proliferation. Scale bar = 200 μm. (H-I) Wound healing and Transwell assays evaluating BMSC migration. Scale bar = 200 μm. (J-L) Quantification of EdU-positive cells, migration area, and migrated cells. (M − N) ALP and ARS staining evaluating osteogenic differentiation of BMSCs. (O-P) Quantification of ALP activity and ARS staining. (Q) qPCR analysis of osteogenic genes (RUNX2, OCN, and COL-Ⅰ) in BMSCs after RBCEV treatment. Data are presented as mean ± SD. *P < 0.05; NS, not significant. Functionally, Y-RBCEVs significantly enhanced BMSC proliferation, migration, and osteogenic differentiation, as demonstrated by increased EdU-positive cells, accelerated wound closure, enhanced Transwell migration, elevated ALP activity, increased mineralized nodule formation, and upregulated expression of osteogenic markers (RUNX2, OCN, and COL-Ⅰ) (Fig. 2G–Q). In contrast, O-RBCEVs showed no significant effect on osteogenic properties. These findings indicate that aging alters RBCEV functional activity, with Y-RBCEVs exhibiting stronger osteogenic regulatory capacity. To determine whether RBCEVs regulate bone metabolism through macrophage polarization, BMMs were treated with Y-RBCEVs or O-RBCEVs under basal or LPS-induced inflammatory conditions. As shown in Fig. 3A, under basal conditions, Y-RBCEVs slightly reduced the proportion of CD86+ BMMs (3.27% to 2.5%) and increased CD206+ BMMs (20.2% to 31%). LPS stimulation increased CD86+ cells from 3.27% to 10.6% and decreased CD206+ cells from 20.2% to 15%. Y-RBCEV treatment partially reversed this inflammatory polarization by reducing CD86+ cells from 10.6% to 7.65% and increasing CD206+ cells from 15% to 27.2%. In contrast, O-RBCEVs showed no significant effects on CD86+ or CD206+ BMM proportions under either condition. Y-RBCEVs regulate macrophage polarization and indirectly promote osteogenic differentiation. (A) Flow cytometric analysis of BMM polarization after RBCEV treatment under basal and LPS-induced inflammatory conditions. (B-C) qPCR analysis of M1 markers (IFN-γ, TNF-α, and CD86) and M2 markers (Arg1, IL-4, and CD206) in BMMs. (D-E) Immunofluorescence staining of CD86 and CD206 in LPS-stimulated BMMs. Target proteins are shown in red, F-actin in green, and nuclei in blue. Scale bar = 200 μm. (F) ALP and ARS staining of BMSCs treated with macrophage-conditioned medium. Scale bar = 100 μm. (G-H) Quantification of CD86 and CD206 fluorescence intensity. (I-J) Quantification of ALP activity and ARS staining. (K) qPCR analysis of osteogenic gene expression (RUNX2, OCN, and COL-Ⅰ) in BMSCs after conditioned medium treatment. Data are presented as mean ± SD. *P < 0.05; NS, not significant. Consistently, qPCR analysis (Fig. 3B and C) showed that Y-RBCEVs decreased M1-related genes (TNF-α and CD86) and increased M2-related genes (ARG1 and CD206) under basal conditions. Under LPS stimulation, Y-RBCEVs suppressed IFN-γ, TNF-α, and CD86 expression while enhancing ARG1, IL-4, and CD206 expression. O-RBCEVs only reduced CD86 expression under inflammatory conditions, with limited effects on other polarization markers. Immunofluorescence staining further confirmed that Y-RBCEVs decreased CD86 and increased CD206 protein expression in LPS-stimulated BMMs, whereas O-RBCEVs showed only a mild inhibitory effect on CD86 expression (Fig. 3D–E, G-H). To investigate whether RBCEVs-mediated macrophage modulation affects osteogenesis, conditioned media from treated BMMs were collected and applied to BMSCs (Fig. 3F–I-K). Conditioned medium from LPS-stimulated BMMs (CMLPS) suppressed BMSC osteogenic differentiation, as indicated by reduced ALP activity, calcium nodule formation, and RUNX2, OCN, and COL-Ⅰ expression. CM from O-RBCEV-treated BMMs (CMLPS + O) showed limited improvement, with slight increases in ALP activity and OCN expression but without statistical significance. In contrast, CM from Y-RBCEV-treated BMMs (CMLPS + Y) significantly restored ALP activity, mineralization, and osteogenic gene expression, partially reversing the inhibitory effects of inflammatory macrophages on BMSC differentiation. An in vivo imaging system (IVIS) was used to evaluate the distribution of RBCEVs in femoral fracture mice. As shown in Fig. 4A and B, DiD-labeled Y-RBCEVs and O-RBCEVs accumulated at the fracture site after tail vein injection. Fluorescence signals in both RBCEV groups gradually increased from 6 h to 36 h and were significantly higher than those in the free DiD group at 24 h and 36 h. Both Y-RBCEVs and O-RBCEVs reached peak accumulation at 36 h, followed by a gradual decline. No significant difference was observed between the two RBCEV groups at any time point. In contrast, the fluorescence signal in the DiD group gradually decreased after injection, indicating rapid clearance of free dye. Ex vivo imaging at 72 h (Fig. 4C and D) showed that RBCEVs exhibited slightly higher fluorescence accumulation in the liver compared with the DiD group, although the difference was not statistically significant. No obvious differences were observed in other major organs. Importantly, fluorescence intensity in isolated femurs remained significantly higher in both RBCEV groups than in the DiD group, confirming the retention of RBCEVs at fracture sites. In vivo distribution of RBCEVs after systemic administration. (A) Time-dependent distribution of DiD-labeled Y-RBCEVs and O-RBCEVs in mice with femoral fractures following tail vein injection (6, 12, 24, 36, 48, and 72 h). (B) Quantification of fluorescence intensity at the fracture sites over time. (C-D) Ex vivo fluorescence imaging and quantification of major organs at 72 h post-injection. Data are presented as mean ± SD. *P < 0.05; NS, not significant. To evaluate the biosafety of RBCEVs, histological examination showed no detectable pathological changes in major organs after treatment (Fig. S2A). Serum biochemical analysis, including urea (UREA), uric acid (UA), albumin (ALB), alanine aminotransferase (ALT), creatinine (CRE), and aspartate aminotransferase (AST), revealed no significant alterations in liver or kidney function (Fig. S2B). Collectively, these findings demonstrate that systemically administered RBCEVs accumulate at fracture sites and exhibit favorable in vivo biocompatibility during the treatment period. As shown in Fig. 5A, a mouse model of femoral fracture was established to investigate further the therapeutic effect of RBCEVs on femoral fracture repair in vivo. Micro-CT analysis in Fig. 5B–G-K demonstrated that Y-RBCEVs treatment significantly increased BV/TV, Tb. N, and Tb. Th, and markedly decreased Tb. Sp, but had no significant effect on BS/BV. O-RBCEVs exerted no significant influence on any of the bone morphometric parameters. HE staining (Fig. 5C–L) showed that the callus volume was markedly larger in the Y-RBCEVs group. Compared with the control group, the Y-RBCEVs group displayed significantly more new bone tissue (stained blue-green by SO/FG) (Fig. 5D–M). Immunofluorescence staining of callus tissues in Fig. 5E–F, N-O revealed that Y-RBCEVs treatment significantly upregulated the expression of the osteogenic marker protein OCN and increased the expression of CD206 in the callus, whereas O-RBCEVs treatment had no significant effects on OCN and CD206 expression. In vivo evaluation of RBCEVs in a murine femoral fracture model. (A) Schematic illustration of RBCEV administration in fracture mice. (B) Micro-CT analysis of callus formation after Y-RBCEV or O-RBCEV treatment. (G-K) Quantification of BV/TV, BS/BV, Tb.Th, Tb.N, and Tb.Sp. (C) HE staining of fracture callus tissues. Scale bars = 500 μm and 100 μm. (L) Quantification of callus area. (D) Safranin O/Fast Green staining of fracture callus tissues. Scale bars = 500 μm and 100 μm. (M) Quantification of new bone area. (E) Immunofluorescence staining of OCN in callus tissues. (N) Quantification of OCN-positive area. (F) Immunofluorescence staining of CD206 in callus tissues. (O) Quantification of CD206-positive area. Scale bars = 200 μm and 100 μm. Data are presented as mean ± SD. *P < 0.05; NS, not significant. Single-cell suspensions were prepared from fracture callus tissues and analyzed by flow cytometry to evaluate the effects of RBCEVs on hematopoietic populations and macrophage polarization. As shown in Fig. 6A–E, neither Y-RBCEVs nor O-RBCEVs significantly affected the proportions of LSK, CMP, or GMP populations, suggesting that RBCEVs had limited effects on hematopoietic progenitor differentiation. Further analysis of macrophage polarization (Fig. 6F–H) revealed that Y-RBCEVs significantly increased the proportion of CD206+ macrophages without altering CD86+ macrophage proportions in callus tissues. In contrast, O-RBCEVs showed no significant effects on macrophage polarization. Changes in hematopoietic differentiation and macrophage polarization in callus tissues after RBCEVs treatment. (A) Gating strategy for hematopoietic cells in callus tissues. (HSC: Hematopoietic Stem Cells; LSK: Lineage−, Sca-1+, c-Kit+; CMP: Common Myeloid Progenitors; GMP: Granulocyte-Monocyte Progenitors). (B) Flow cytometric analysis of hematopoietic cell populations after RBCEV treatment. Quantification of (C) LSK, (D) GMP, and (E) CMP cell proportions. (F) Flow cytometric analysis of macrophage polarization in callus tissues after RBCEV treatment. Quantification of (G) CD86+ and (H) CD206+ macrophage proportions. Data are presented as mean ± SD. *P < 0.05; NS, not significant. To investigate the molecular basis underlying the distinct functions of Y-RBCEVs and O-RBCEVs, comparative proteomic analysis was performed. PCA showed clear separation between the two RBCEV groups, and hierarchical clustering further demonstrated distinct protein expression patterns (Fig. 7A and B). Compared with O-RBCEVs, Y-RBCEVs contained 214 upregulated and 21 downregulated proteins (Fig. 7C). The top 10 differentially expressed proteins based on fold change included PADI4, TANGO2, OLFM1, PTPRG, DBH, VSIG8, PROZ, EPB41, VCAN, and FGFR4 (Fig. 7D). GO and KEGG enrichment analyses revealed that differentially expressed proteins were mainly associated with plasma-related functions, including fibrinolysis, complement activation, platelet aggregation, and coagulation. Cellular component analysis showed enrichment in extracellular regions and EV-associated components (Fig. 7E). Analysis of the top 50 altered proteins further identified enrichment in biological processes related to protein degradation, proteasome activity, osteogenic differentiation, integrin signaling, and regulation of protein tyrosine phosphorylation (Fig. 7F). KEGG analysis demonstrated associations with proteasome pathways, extracellular matrix–receptor interactions, and cell adhesion signaling (Fig. 7G and H). Several bioactive proteins enriched in Y-RBCEVs were also identified as potential contributors to fracture repair (Supplementary Fig. 11A). Osteogenesis-associated cargos, including SPARC, TGFBI, COMP, and CLEC3B, have been implicated in bone matrix formation, extracellular matrix remodeling, osteogenic differentiation, and mineralization [[33], [34], [35], [36]]. In addition, THBS4 and TNC, involved in angiogenesis and periosteal stem cell regulation, respectively, were enriched in Y-RBCEVs [37,38]. Y-RBCEVs also showed higher abundance of immunoregulatory proteins, including CD163, CHI3L1, and TGFB1, whereas inflammation-associated proteins such as IL36G, S100A7L2, and CFP were relatively decreased compared with O-RBCEVs [[39], [40], [41], [42], [43]]. These findings suggest that the beneficial effects of Y-RBCEVs on BMSC osteogenic differentiation, macrophage polarization, and fracture healing are unlikely to be mediated by a single molecule, but rather by the synergistic effects of multiple bioactive cargos with osteogenic and immunomodulatory functions. Proteomic profiling identifies differentially expressed proteins between Y-RBCEVs and O-RBCEVs. (A) Principal component analysis (PCA) scatter plot of proteomic mass spectrometry data from Y-RBCEVs and O-RBCEVs. (B) Heatmap showing hierarchical clustering of differentially expressed proteins. (C) Number of differentially expressed proteins between the two groups. (D) Volcano plot highlighting significantly altered proteins (P < 0.05, log2(FC) > 1). (E-H) Functional enrichment analyses of differentially expressed proteins, including Gene Ontology (GO) enrichment (E-F) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis (G-H). Data are presented as mean ± SD. *P < 0.05; NS, not significant. Previous studies have reported that PADI4 promotes M1 macrophage polarization and exacerbates inflammatory responses [44,45]. However, the roles of other top-10 proteins, including EPB41, VSIG8, OLFM1, FGFR4, PROZ, VCAN, PTPRG, TANGO2, and DBH, in regulating macrophage differentiation and bone metabolism remain unclear. Therefore, this study investigated the functions of these proteins. As shown in Figures S3–10, 8A-B, and 9F, recombinant VCAN, VSIG8, and PTPRG enhanced osteogenic-related gene expression, suppressed M1-associated genes, and increased M2-associated gene expression in LPS-stimulated BMMs. These findings indicate that VCAN, VSIG8, and PTPRG may represent functional cargos contributing to the immunomodulatory effects of Y-RBCEVs. To identify age-related changes in these candidate cargos, VCAN, VSIG8, and PTPRG levels were compared in RBCEVs derived from donors of different age groups (18–35, 35–60, and 61–75 years). Among these proteins, only VCAN exhibited an age-dependent reduction, with significantly lower levels in RBCEVs from elderly donors (61–75 years) compared with young donors (18–35 years) (Fig. 8C–E). This finding suggests that reduced VCAN abundance may contribute to the functional decline of O-RBCEVs. Furthermore, Y-RBCEV treatment increased VCAN levels in both total protein and membrane fractions of BMMs and BMSC (Fig. 8D–F, and Supplementary Fig. 11B and E). Immunofluorescence analysis further demonstrated increased VCAN localization in the cytoplasm and cell membrane following Y-RBCEV treatment (Fig. 8G and H), supporting the uptake and intracellular distribution of EV-derived VCAN. Flow cytometry was then performed to determine the direct effects of VCAN on macrophage polarization. Under basal conditions, VCAN (50 and 100 ng/mL) reduced the proportion of CD86+ BMMs, while 100 ng/mL VCAN increased CD206+ BMMs. Under LPS-induced inflammatory conditions, both concentrations of VCAN significantly decreased CD86+ cells and increased CD206+ cells (Fig. 8I–K), demonstrating that VCAN promotes macrophage polarization toward an M2 phenotype. Validation of VCAN as a functional cargo of Y-RBCEVs in macrophage polarization. (A-B) qPCR analysis of macrophage polarization-related genes (TNF-α, IFN-γ, CD86, IL-4, Arg1, and CD206) after VCAN treatment under (A) basal and (B) LPS-induced inflammatory conditions. (C) Western blot analysis of VCAN, VSIG8, and PTPRG levels in Y-RBCEVs derived from donors of different ages, with TSG101 as the loading control. (E) Quantification of protein expression. (D) Western blot analysis of VCAN expression in BMM membrane and total protein fractions after Y-RBCEV treatment. ATP1A1 and ACTIN were used as membrane and total protein controls, respectively. (F) Quantification of VCAN expression. (G) Immunofluorescence staining of VCAN localization in BMMs after Y-RBCEV treatment. (H) Quantification of fluorescence intensity. (I) Flow cytometric analysis of macrophage polarization after VCAN treatment under basal or LPS-induced conditions. (J-K) Quantification of CD86+ and CD206+ BMM proportions. Data are presented as mean ± SD. *P < 0.05; NS, not significant. Immunofluorescence staining (Fig. 9A–C) further confirmed that VCAN treatment (50 and 100 ng/mL) reduced CD86 expression and increased CD206 expression in BMMs under both basal and LPS-induced inflammatory conditions. Conditioned medium collected from LPS-stimulated BMMs treated with VCAN was subsequently applied to BMSCs during osteogenic induction. As shown in Fig. 9D–I, VCAN-modulated BMM conditioned medium significantly restored LPS-impaired ALP activity, mineralized nodule formation, and expression of osteogenic markers (RUNX2, OCN, and COL-I) in BMSCs. To further determine whether VCAN is required for the biological effects of Y-RBCEVs, a VCAN-specific neutralizing antibody was used to block VCAN function on Y-RBCEVs. Compared with Y-RBCEV treatment alone, VCAN blockade significantly partially attenuated the osteogenic-promoting and immunomodulatory effects of Y-RBCEVs. Specifically, VCAN neutralization partially reduced OCN and RUNX2 expression (Supplementary Fig. 11C and F), decreased ALP activity, and impaired calcium nodule formation as indicated by ARS staining (Supplementary Fig. 11H–I). Under inflammatory conditions, VCAN blockade partially increased the expression of pro-inflammatory markers iNOS and CD86, while decreasing anti-inflammatory markers Arg1 and CD206 (Supplementary Figs. 11D,G,J-K). These findings indicate that VCAN contributes to Y-RBCEV-mediated macrophage M2 polarization and subsequent enhancement of BMSC osteogenic differentiation, supporting VCAN as an important bioactive cargo underlying the regenerative effects of Y-RBCEVs. VCAN modulates macrophage polarization and indirectly enhances BMSC osteogenic differentiation. (A) Immunofluorescence staining of CD86 and CD206 expression in BMMs treated with VCAN under basal or LPS-induced inflammatory conditions. CD86/CD206 are shown in red, F-actin in green, and nuclei in blue. Scale bar = 200 μm. (B-C) Quantification of CD86 and CD206 fluorescence intensity, respectively. (D-E) ALP and ARS staining of BMSCs treated with conditioned medium from VCAN-modulated BMMs under inflammatory conditions. Scale bar = 100 μm. (H-I) Quantification of ALP activity and ARS staining. (F) qPCR analysis of osteogenic marker genes (RUNX2, OCN, and COL-I) in BMSCs treated with VCAN. (G) qPCR analysis of osteogenic gene expression in BMSCs treated with conditioned medium from VCAN-modulated BMMs. Data are presented as mean ± SD. *P < 0.05; NS, not significant. In addition, the effects of VCAN on BMSC proliferation and migration were evaluated using EdU proliferation assays, transwell migration assays, and wound healing assays. As shown in Supplementary Fig. 12, treatment with 100 ng/mL VCAN significantly increased BMSC proliferation, whereas VCAN had no significant effect on BMSC migration capacity. To elucidate how Y-RBCEV-derived VCAN regulates macrophage polarization, potential VCAN-interacting proteins were predicted using STRING, GeneMANIA, HitPredict, BioGRID, and HINT databases. As shown in Fig. 10A, CD44 was identified as the only overlapping candidate among these databases. Co-immunoprecipitation analysis further confirmed that both recombinant VCAN and Y-RBCEV treatment enhanced the interaction between VCAN and CD44 in BMMs (Fig. 10B). The functional involvement of CD44 was subsequently examined. CD44 neutralization significantly attenuated the effects of VCAN and Y-RBCEVs on macrophage polarization under LPS-induced inflammatory conditions (Fig. 10C–E). Specifically, anti-CD44 treatment reversed the suppression of M1-associated genes (TNF-α and IFN-γ), reduced the induction of M2-associated genes (IL-4 and ARG1), and partially restored CD86 expression while decreasing CD206 expression. These findings indicate that CD44 is required for the immunomodulatory effects of VCAN and Y-RBCEVs. CD44/PI3K/AKT signaling mediates the effects of Y-RBCEV-derived VCAN on macrophage polarization. (A) Venn diagram showing potential VCAN-interacting proteins predicted by STRING, GeneMANIA, HitPredict, BioGRID, and HINT databases. (B) Co-immunoprecipitation (Co-IP) analysis of VCAN-CD44 interaction in BMMs after VCAN or Y-RBCEV treatment. (C) qPCR analysis of inflammatory and polarization-related genes (TNF-α, IFN-γ, IL-4, and ARG1) after CD44 neutralization in VCAN- or Y-RBCEV-treated BMMs. (D) Immunofluorescence staining of CD86 and CD206 expression after CD44 blockade. Scale bar = 200 μm. (E) Quantification of fluorescence intensity. (F) PCA of transcriptomic profiles from LPS-stimulated BMMs treated with VCAN. (G) Number of differentially expressed genes identified by transcriptomic analysis (P < 0.05, |log2FC| > 1). (H) KEGG pathway enrichment analysis of differentially expressed genes. (I) Western blot analysis of PI3K/AKT pathway activation and macrophage polarization markers (p-PI3K, PI3K, p-AKT, AKT, ARG1, and iNOS) following VCAN or Y-RBCEV treatment with or without CD44 blockade. (J) Quantification of protein expression. (K) Western blot analysis of macrophage polarization-related proteins after AKT inhibition by MK2206. (L) Quantification of protein expression. (M) Immunofluorescence staining of CD86 and CD206 expression after MK2206 treatment. Scale bar = 200 μm. (N) Quantification of fluorescence intensity. Data are presented as mean ± SD. *P < 0.05; NS, not significant. To further investigate downstream signaling, transcriptomic analysis revealed distinct gene expression patterns between LPS-stimulated BMMs with and without VCAN treatment (Fig. 10F). Compared with the LPS group, VCAN treatment resulted in 26 upregulated and 24 downregulated genes (Fig. 10G). KEGG enrichment analysis identified significant enrichment of the PI3K-AKT signaling pathway among the differentially expressed genes (Fig. 10H). Western blot analysis demonstrated that Y-RBCEVs and VCAN increased PI3K and AKT phosphorylation, enhanced ARG1 expression, and reduced iNOS expression in LPS-stimulated BMMs (Fig. 10I and J). These effects were attenuated by CD44 neutralization, confirming the involvement of CD44-mediated PI3K/AKT activation. Furthermore, inhibition of AKT phosphorylation using MK2206 partially reversed the effects of Y-RBCEVs and VCAN on AKT activation, Arg1 induction, and iNOS suppression (Fig. 10K and L). Consistently, immunofluorescence analysis showed that MK2206 partially abolished the Y-RBCEV- and VCAN-induced increase in CD206 and reduction in CD86 expression (Fig. 10M − N). Collectively, these findings demonstrate that Y-RBCEV-derived VCAN interacts with CD44 to activate PI3K/AKT signaling, thereby promoting macrophage polarization toward an anti-inflammatory phenotype. Discussion This study elucidated the potential role and underlying mechanisms of RBCEVs in fracture repair through clinical observational study, proteomic and transcriptomic profiling, and comprehensive in vitro and in vivo investigations. Analysis of the NHANES database revealed an association between RBC-related indicators, BMD, and mortality in fracture populations. Complementary murine fracture experiments further suggested a potential link between RBC transfusion and enhanced fracture repair, providing the rationale for investigating RBCEVs as functional mediators of RBC-derived effects. Subsequent experiments demonstrated that RBCEVs promote fracture healing through coordinated osteogenic and immunomodulatory effects. Y-RBCEVs directly enhanced BMSC proliferation and osteogenic differentiation, while also modulating the macrophage polarization through bioactive cargos. Among these cargos, VCAN was identified as an important functional protein that interacts with CD44 on macrophages and activates PI3K/AKT signaling, thereby promoting macrophage polarization toward an anti-inflammatory phenotype and facilitating bone regeneration. Therefore, this study provides new insights into the biological functions of RBCEVs in fracture repair and highlights their potential as a promising biological therapeutic strategy for bone regeneration. Although the role and underlying mechanisms of RBCs in fracture repair remain incompletely understood, accumulating evidence suggests an association between RBC-related parameters and bone metabolism. A retrospective study of mandibular angle osteotomy patients reported that higher preoperative RBC count and hemoglobin levels were associated with improved postoperative bone regeneration, with each 1 × 1012/L increase in RBC count and each 1 g/L increase in hemoglobin corresponding to 2.558-fold and 1.036-fold higher probabilities of bone regeneration, respectively [46]. In addition, NHANES-based studies have shown that increased red blood cell distribution width (RDW) is inversely associated with femoral neck BMD in men and postmenopausal women [47], whereas RBC folate levels are positively associated with lumbar BMD in postmenopausal women [48]. These findings indicate a close relationship between RBCs and bone metabolism and repair. In this study, large-scale NHANES analysis further demonstrated associations of RBC count, hemoglobin (HGB), and hematocrit (HCT) with BMD and mortality in fracture populations. The inverse association observed in the female HGB Q4 subgroup may reflect characteristics of individuals with relatively high HGB levels, residual confounding, or population-specific factors. Notably, HGB values in this subgroup ranged from 14.1 to 19.6 g/dL, with some exceeding the commonly referenced upper limit for females (15 g/dL), which may partially contribute to this observation. However, given the observational nature of NHANES, these findings require further validation and do not establish causality. Instead, this analysis provided clinical evidence supporting the hypothesis that RBC-derived factors may participate in fracture repair, motivating subsequent functional investigations. Few studies have explored how RBCs influence bone metabolism. In this study, we investigated RBCEVs as potential mediators linking RBCs to bone repair. Extracellular vesicles are important regulators of intercellular communication that transfer bioactive cargos, including proteins, lipids, and nucleic acids, to modulate immune responses and tissue homeostasis [[49], [50], [51]]. Previous studies have demonstrated that EVs from various cell sources can promote fracture healing by enhancing osteogenesis and angiogenesis through cargo delivery [52,53]. However, the role of RBCEVs in bone regeneration and osteoimmune regulation remains largely unexplored. Unlike EVs derived from nucleated cells, mature RBCs lack nuclei and most intracellular organelles, resulting in a relatively simple cargo profile and low immunogenicity [[54], [55], [56]]. Nevertheless, RBCEVs retain the ability to transport functional molecules, including non-coding RNAs and proteins, to recipient cells [54,57] Moreover, RBCEVs can be obtained from peripheral blood, providing a potentially scalable source of derived vesicles [20]. Previous studies have shown that RBC-derived apoptotic vesicles deliver carbonic anhydrase 1 to BMSCs to enhance osteogenic differentiation [58], while RBCEVs-mediated heme delivery suppresses inflammatory responses in macrophages and reduces foam cell formation in atherosclerosis [20]. In this study, we demonstrated that Y-RBCEVs promote BMSC osteogenic differentiation and induce anti-inflammatory macrophage polarization in vitro and in vivo, thereby enhancing fracture repair. Although the improvement in individual micro-CT parameters was moderate, consistent changes across multiple bone parameters (BV/TV, Tb.N, Tb.Th, and Tb.Sp), together with histological and immunological evidence, support the biological significance of Y-RBCEV-mediated bone regeneration. These findings highlight RBCEVs as promising candidates for future therapeutic strategies in bone repair. Previous studies have reported that hematopoietic regulation contributes to fracture healing by shaping the bone marrow immune microenvironment. Alterations in hematopoietic stem and progenitor cell populations, myeloid lineage differentiation, and subsequent macrophage polarization may influence local immune responses and bone regeneration [59]. Therefore, we examined hematopoietic and macrophage populations in fracture callus tissues after RBCEV treatment. However, neither Y-RBCEVs nor O-RBCEVs significantly affected the proportions of HSC/LSK, CMP, or GMP populations, suggesting limited effects on hematopoietic lineage distribution. In contrast, Y-RBCEVs markedly increased the proportion of M2-polarized macrophages at fracture sites. These findings indicate that RBCEVs may primarily regulate the functional phenotype of macrophages rather than altering upstream hematopoietic differentiation. Through modulation of macrophage polarization, Y-RBCEVs may reshape the local immune microenvironment and facilitate bone regeneration. RBCEVs were observed to accumulate at fracture sites; however, the mechanisms underlying this localization remain unclear. This process may involve both passive retention and active interactions. Fracture-induced inflammation, increased vascular permeability, and tissue barrier disruption may promote EV deposition within injured tissues [60,61]. In addition, proteomic analysis revealed enrichment of Y-RBCEVs in proteins associated with focal adhesion, ECM-receptor interaction, and integrin-mediated signaling, suggesting their potential involvement in EV–tissue interactions. Therefore, RBCEV enrichment at fracture sites may result from a combination of injury-associated changes and cargo-mediated interactions with the local microenvironment. Further studies using molecular blocking approaches are required to identify the specific surface molecules responsible for RBCEV localization and retention. Among the differentially expressed cargos identified by proteomic profiling, VCAN was selected for further mechanistic investigation because of its marked enrichment in Y-RBCEVs and potential roles in immune regulation. Previous studies have shown that VCAN can regulate macrophage function in inflammatory and tumor-associated contexts [62,63]; however, its role in bone immune regulation remains largely unknown. Our findings provide new evidence that VCAN may participate in osteoimmune regulation by linking extracellular vesicle-derived signals with macrophage functional regulation. Importantly, although VCAN exhibited significant immunomodulatory activity, the therapeutic effects of Y-RBCEVs are unlikely to be attributed to VCAN alone. Proteomic profiling revealed extensive age-associated alterations in RBCEV cargo compositions, suggesting that the functional differences between Y-RBCEVs and O-RBCEVs may result from coordinated changes in multiple bioactive components. In addition to regulating macrophage polarization and BMSC osteogenesis, these differential cargos may contribute to other aspects of fracture repair, such as extracellular matrix remodeling, angiogenic responses, and tissue regeneration. Moreover, in addition to VCAN, preliminary validation indicated that other candidate proteins, such as PTPRG and VSIG8, may also influence osteogenic and inflammatory responses. These findings suggest that Y-RBCEVs represent a complex biological system in which multiple cargos collectively shape the regenerative microenvironment, rather than functioning through a single dominant molecule. However, the individual and cooperative functions of these differential proteins remain largely unexplored. In addition, EVs biological activity is not solely determined by protein components, and age-related alterations in non-protein cargos, including miRNAs and lipids, may also contribute to the functional difference between Y-RBCEVs and O-RBCEVs. Future studies integrating multi-omics analyses and targeted functional approaches are required to comprehensively elucidate how RBCEV cargo remodeling regulates bone regeneration and fracture repair. Therefore, the VCAN-CD44/PI3K-AKT axis identified in this study should be considered as one representative mechanism rather than the sole explanation for the regenerative activity of Y-RBCEVs. This study has several strengths. First, by integrating population-based observational analysis, proteomic profiling, and comprehensive in vitro and in vivo experiments, this study provides new insights into the potential biological link between RBC-derived components and fracture repair. The findings suggest that RBCEVs may represent important mediators underlying the beneficial effects associated with RBCs and provide a new perspective for understanding the role of circulating vesicles in bone regeneration and osteoimmunity. Second, this study highlights the translational potential of Y-RBCEVs as a cell-free therapeutic strategy by demonstrating their capacity to modulate osteogenic and immune responses. Furthermore, the identification of age-associated differences in RBCEV cargo composition provides potential molecular candidates for future studies investigating biomarkers of fracture healing and regenerative capacity. This study also has several limitations. First, the observational analyses based on NHANES cannot establish causal relationships between RBC-related parameters, BMD, and fracture outcomes. Instead, these findings were used for hypothesis generation and provided clinical relevance for subsequent RBCEV investigations. Second, although this study identified the VCAN-CD44/PI3K-AKT axis as an important mechanism underlying the immunomodulatory effects of Y-RBCEVs, RBCEV-mediated fracture repair is likely regulated by multiple functional cargos. The independent and synergistic roles of other differential proteins and non-protein components, including miRNAs and lipids, remain to be investigated. Third, this study mainly focused on the regulation of macrophage polarization and BMSC osteogenic differentiation. Whether RBCEVs influence other cell populations involved in fracture healing, such as endothelial cells, periosteal progenitor cells, or other immune subsets, requires further exploration. Finally, the use of human-derived RBCEVs in murine models may not fully recapitulate interactions within the human physiological environment and could affect immune responses, biodistribution, metabolism, and therapeutic outcomes. In addition, species-specific differences in EVs uptake and receptor interactions cannot be completely excluded. Future studies using humanized or immune-compatible models may further improve the translational relevance of RBCEV-based therapies. Conclusion Taken together, this study reveals that Y-RBCEVs promote fracture repair through the coordinated actions of multiple bioactive cargos. Among them, VCAN was identified as an important functional mediator that interacts with CD44 and activates PI3K/AKT signaling to promote anti-inflammatory macrophage polarization, representing one mechanism underlying the regenerative effects of Y-RBCEVs. These findings highlight the potential of RBCEVs as a promising cell-free therapeutic platform for bone repair. Ethics approval and consent to participate All animal experiments were performed with the approval of the Medical Ethics Committee of The Third Hospital of Hebei Medical University (Ethics Approval No. Z2024-039-1). Red blood cells (RBCs) were isolated from fresh whole blood samples obtained from young and elderly volunteers at the Third Hospital of Hebei Medical University, with the approval of the Medical Ethics Committee of the Third Hospital of Hebei Medical University (Ethics Approval No. 2019-021-1). Funding This work was supported by the Natural Science Foundation of China (32130052), Hengrui Hebei Innovation Development Medical Collaborative Initiative (HR202502052) and Hebei Natural Science Foundation General Program (H2024206071). CRediT authorship contribution statement Jiahao Wang: Conceptualization, Investigation, Writing – original draft. Xiaokang Zhang: Investigation, Writing – original draft. Qi Zhang: Investigation. Jiheng Xiao: Investigation. Siteng Li: Investigation. Jianpeng Chen: Investigation. Chengbin Huang: Investigation. Yang Luo: Investigation. Zeming Shan: Investigation. Ziping Chen: Investigation. Ling Wang: Conceptualization, Writing – review & editing. Yingze Zhang: Conceptualization, Writing – review & editing. Declaration of competing 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. Footnotes Peer review under the responsibility of editorial board of Bioactive Materials. Supplementary data to this article can be found online at https://doi.org/10.1016/j.bioactmat.2026.07.047. Contributor Information Ling Wang, Email: wangling2021@hebmu.edu.cn. Yingze Zhang, Email: yzling_liu@163.com, YZZ_NKU@126.com. Supplementary data The following are the Supplementary data to this article. Multimedia component 1 Multimedia component 2 References Associated Data Supplementary Materials Multimedia component 1 Multimedia component 2