A migrasome-based osteoinductive strategy: reprogramming the bone microenvironment for accelerated coupling of angiogenesis and osteogenesis

Abstract

Abstract

The spatiotemporal coupling of osteogenesis and angiogenesis mediated by type H vessels represents a fundamental mechanism in bone formation. Although macrophages, the key immunomodulatory cells in the bone microenvironment, are known to participate in the regulation of type H vessels, the underlying mechanisms still remain insufficiently understood. The role of migrasomes, a newly discovered class of substrate-anchored extracellular vesicles, in macrophage-type H vessel crosstalk during bone regeneration were investigated in this study. M2 macrophage-derived migrasomes (RAW-MS) were produced by stimulating RAW 264.7 cells with fibronectin (FN). In vitro experiments indicated that RAW-MS were internalized by recipient cells, thereby promoting anti-inflammatory macrophage polarization, enhancing angiogenic activity, and facilitating osteogenic differentiation. Proteomic analysis revealed that RAW-MS were enriched with numerous angiogenesis-associated proteins. Transcriptome sequencing and subsequently in vitro experiments demonstrated that RAW-MS activated the endothelial tip cells phenotype and sprouting angiogenesis via TGFβ1/Smad2 signaling pathway. When incorporated into GelMA hydrogels, RAW-MS significantly improved vascularized bone regeneration in a critical-sized rat cranial defect model. Both in vitro and in vivo investigations consistently showed that RAW-MS enhanced the coupling of angiogenesis and osteogenesis accompanied by an increased density of type H vessel formation through upregulation of the TGFβ1/Smad2 signaling. In conclusion, this study highlighted the potential of migrasomes as innovative signaling vehicles for manipulating the regenerative microenvironment in tissue engineering.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04619-4.

Introduction

The repair of critical bone defects caused by trauma, tumors, infections, or congenital anomalies remains a significant clinical challenge. These defects are frequently associated with severe complications, including nonunion, malunion, and deep infection [1]. A key limiting factor in critical bone defects is inadequate vascular supply [2]. Bone is a highly vascularized organ. The blood vessels within bone tissue play a crucial role in bone development, remodeling and regeneration. The vascular network is densely and intricately distributed throughout the bone, providing essential nutrients to the bone tissue and removing metabolic waste, thereby maintaining a stable and favorable microenvironment for bone homeostasis [3]. Both bone tissue and blood vessels possess endocrine functions, enabling them to regulate adjacent or distant tissues or organs, creating a bidirectional promotional effect. During the regeneration process following bone injury, osteogenesis and angiogenesis exhibit a tight spatiotemporal coordination, which known as “osteogenic-angiogenic coupling” [4].

The key mechanism of osteogenic-angiogenic coupling is associated with a subtype of capillaries named type H vessels. Type H vessels are mainly located near the metaphyseal growth plate and the periosteum and endosteum of the metaphysis, characterized by the dual-positive expression of CD31 and Endomucin (EMCN) [5]. These vessels are densely surrounded by osteoprogenitor cells highly expressing the transcription factor Osterix. Studies have shown that endothelial cells (ECs) of type H vessels promote the proliferation and differentiation of osteoprogenitor cells, as well as the maturation and hypertrophy of chondrocytes, by secreting many factors such as Noggin [6]. Correspondingly, osteoprogenitor cells enhance the sprouting and formation of new blood vessels by secreting large amounts of SLIT3 and VEGFA [7]. Overall, the numerous paracrine cytokines activate multiple signaling pathways in bone marrow mesenchymal stromal cells (BMSCs) and ECs, such as classic Wnt, PI3K-AKT, TGFβ, Notch signaling pathways, thereby promote the coupling of osteogenesis and angiogenesis [8–11]. In addition to secretory proteins, extracellular vesicles (EVs) are also served as another important medium in intercellular communication [12–16].

EVs are indeed lipid bilayer-enclosed nanoparticles released by virtually all cell types which are generally divided into three categories: exosomes, microvesicles and apoptotic bodies [17]. It was indicated that exosomes from BMSC with 7 days osteogenic differentiation may enhance CD31^hi Emcn^hi ECs angiogenesis [18]. Both Liu et al. and Wang et al. found that the exosomes derived from endothelial cell triggered a positive feedback loop in osteogenesis-angiogenesis coupling [11, 16].

Recently, a kind of newly discovered vesicular organelle named migrasomes formed on the retraction fibers (RFs) of migrating cell has garnered considerable attention in biomedical research field [9]. When the cell migrating, migrasomes are formed through the translocation of cytoplasmic material into the RFs. With the RFs rupturing, the migrasomes release at the moved location and are taken up by the recipient cells, enabling lateral transfer of biomaterial and information [19]. Migrasomes have been proved as a principal secretion route in migrating cells, which recruit cellular contents and enrich with signaling molecules such as lipid, RNA, chemokines, cytokines and angiogenic factors [20, 21]. Therefore, migrasomes are considered as new cellular messengers, which realize precious delivery and play critical roles in physiological processes including angiogenesis. It is reported that monocytes with highly migratory ability deposited migrasomes enriched in pro-angiogenic factors, thereby orchestrating capillary growth and promoting angiogenesis [22].

Besides ECs-BMSCs intercellular communication, the interactions between macrophage and ECs also regulate the formation of type H vessels and bone tissue remodeling [23]. Upon injure or surgery, monocytes in the blood were promptly recruited to the regenerative site and differentiated into macrophages by chemokines and specific receptors [24]. The macrophages release a serious of cytokines, vasoactive agents, matrix metalloproteinases, and growth factors to induce vascular and bone tissue remodeling [25, 26]. For example, the macrophage can secrete PDGF-BB to recruit both endothelial and osteoblast precursor cells to support the growth of type H vessels [6]. Moreover, several studies showed that the macrophage derived EVs also participated in formation of type H vessels[14, 15, 27].

As migrative active cells, macrophages have been proven that secreted migrasomes to communicate with other cells in the microenvironment during migration, thereby influencing their biological activities [28]. However, it is still not clear whether these migrasomes from macrophage participate in the formation of the type H vessel and coupling of angio/osteogenesis. Herein, macrophage-derived migrasomes were successfully extracted and the contents were detected by proteomics analysis. Moreover, the effect of the macrophage-derived migrasomes on type H vessel and angiogenesis-osteogenesis coupling was firstly investigated, which was conducive to uncover a new mechanism of cell communication in bone regeneration.

Materials and methods

Cell culture

Human bone mesenchymal stomal cell (BMSCs), Human Umbilical Vein Endothelial Cells (HUVECs), Mouse mononuclear macrophage (RAW 264.7) cells and human monocytic leukemia (THP-1) cells were grown in Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco Co., USA) containing 10% fetal bovine serum (FBS, Gibco Co., USA) and 1% penicillin-streptomycin (P/S, Gibco Co., USA), and the medium was changed each two days. All cells were cultured at 37 °C with 5% CO₂. To facilitate the differentiation of THP-1 cells into M0 macrophages (MΦ), 100 ng/ml of Phorbol 12-myristate 13-acetate (PMA, Solarbio, China) were added and co-cultured with them for 24 h. Once the cells reached 80%–90% confluence, they were first washed by phosphate-buffered saline (PBS, Servicebio, China) detached using TrypLE (Gibco Co., USA), neutralized with complete DMEM, and then centrifuged.

Observation of RAW-MS in situ with SEM

Cells grown on glass coverslips with a diameter of 15 mm were placed into a 24-well plate and incubated with 10 µg/ml FN (Solarbio, China) for 1 h. After discarding the FN solution, RAW 264.7 cells were seeded at a density of 5 × 10⁴ cells per well and incubated for 12 h. The cells were fixed by 2.5% glutaraldehyde (GA, Servicebio, China) for 30 min, and then dehydrated with graded ethanol (30, 50, 70, 90, 100%) for 5 min orderly. After then, freeze drying was performed. Finally, the morphology of migrasomes producing from RAW 264.7 cells (RAW-MS) was observed under a scanning electron microscope (SEM, Nexsa, Thermo Fisher Scientific, USA).

Observation of RAW-MS in situ with CLSM

RAW 264.7 cells were cultured in confocal dishes for 12 h after incubating with 10 µg/ml FN for 1 h. After washing by PBS twice, the cells were fixed by 2.5% GA for 30 min and stained with Alexa Fluor^® 647 conjugated WGA (Solarbio, China) for 1 h. Subsequently, the migrating RAW 264.7 cells and the derived migrasomes in situ were observed in confocal laser scanning microscope (CLSM, LSM980, Zeiss, Germany).

Observation of RAW-MS in situ with TEM

The cells were first cultivated on plates coated with 10 µg/ml FN. After culturing, the cells were fixed using 2.5% GA and subsequently post-fixed in 2% osmium tetroxide. The samples were then dehydrated through a graded series of ethanol and acetone solutions before being embedded in epoxy resin. Ultrathin sections, ranging from 60 to 80 nm in thickness, were prepared using a diamond knife and collected onto Formvar-coated copper grids (150 mesh). These sections were stained with 5% uranyl acetate for 15 min, followed by 0.1% lead citrate for 5 min. The prepared samples were examined using a Hitachi HT-7800 transmission electron microscope (TEM, Hitachi Ltd, Tokyo, Japan) operating at an acceleration voltage of 80 kV.

Isolation and purification of RAW-MS

Cells were cultured in 150 mm dishes pre-coated with 10 µg/ml FN for 12 h. Following incubation, the cells and migrasomes adhered to the plates were detached using 0.25% TrypLE and subsequently transferred into 50 ml tubes, with all procedures carried out at 4 °C. To eliminate cells and larger debris, the suspension was subjected to sequential centrifugation steps: first at 1,000 g for 10 min, followed by 4,000 g for 20 min. The resulting supernatant was further processed to isolate crude migrasomes by centrifugation at 20,000 g for 30 min. The pelleted migrasomes were washed twice with PBS and then resuspended in 100–200 µL of PBS.

For further purification, the crude migrasome pellets were re-suspended and layered onto a multi-step Optiprep density gradient (Axis-Shield, UK) (5%, 10%, 15%, 20% [sample], 25%, 30%, 35%, and 40%) and centrifuged at 150,000 g for 4 h at 4 °C. After centrifugation, gradient fractions were carefully collected and diluted with 500 µL of PBS. These fractions were then centrifuged at 20,000 g for 30 min at 4 °C. The resulting pellet was washed once with PBS and centrifuged at 2,000 g for 10 min at 4 °C. Finally, the supernatant was collected and subjected to centrifugation at 20,000 g for 30 min at 4 °C to obtain highly purified migrasomes for subsequent analysis. The procedure above was adopted from the reference [9]. The process above was exhibited in Fig. 1I.

Fig. 1.

Characterization of migrasomes. (A) The morphological characteristics of migrating RAW 264.7 observed by SEM. White arrow: Migrasomes, yellow arrow: Retraction fibers. (B) A representative immunofluorescence image of RAW 264.7 cells marked by WGA. White arrow: Migrasomes, yellow arrow: Retraction fibers. (C) A representative image of EVs/migrasomes in situ observed by TEM. Yellow arrow: Retraction fibers, red triangle: Unreleased migrasomes, white arrow: Released migrasomes, black arrow: MDNPs. (D) Representative negative staining image of extracted migrasomes observed by TEM, yellow arrow: Retraction fibers. (E) Representative image of ultrathin section of extracted migrasomes observed by TEM, black arrow: Migrasomes, white arrow: MDNPs. (F) Western blotting of the expression of TSPAN4, NDST1, CPQ, EOGT, TSG101, CD63, GM130, Histone H3 and GAPDH. (G) NTA analysis of particle size distribution of migrasomes. (H) The zeta potential of migrasomes. (I) Schematic illustration of the refined migrasomes purification procedure from cell medium

Characterization of RAW-MS

For the morphological evaluation of extracted RAW-MS, The TEM was employed to observe the structural characteristics of migrasomes through negativing staining and ultrathin sections. For negativing staining preparation, the extracted samples were suspended in PBS and applied to carbon-coated electron microscopy grids. After allowing the grids to incubate at room temperature for 5 min, they were subjected to negative staining with a 2% phosphotungstic acid solution for 30 s. For ultrathin sections preparation, the extracted sample were initially fixed in a solution containing 2.5% glutaraldehyde. They were then dehydrated using a graded acetone ethanol series, and embedded in epoxy resin (SPI Inc., Westchester, PA, USA). Ultrathin sections with thicknesses ranging from 80 to 90 nm were stained for 15 min with 5% uranyl acetate, followed by 5 min with 0.1% lead citrate. The samples were both then analyzed using TEM, and the electron micrographs were captured for further examination.

The particle size distribution and zeta potential of RAW-MS were measured by nanoparticle tracking analysis (Zetasizer Nano ZS, Malvern Instruments Ltd, Malvern, UK) in accordance with the instructions.

RAW264.7 cells exposed to 10 µg/ml FN and the isolated RAW-MS were lysed using RIPA buffer (Beyotime, China) containing proteinase inhibitor cocktail. Afterward, protein concentrations from both RAW264.7 cells and RAW-MS were determined using a BCA Protein Assay Kit (Beyotime Biotechnology Inc., China). To authenticate RAW-MS, the positive markers of migrasomes: bifunctional heparan sulfate N-deacetylase/N-sulfotransferase 1 (NDST1, 1:1000, Proteintech, China), carboxypeptidase Q (CPQ, 1:1000, Abclonal Technology, China), EGF domain-specific O-linked Nacetylglucosamine transferase (EOGT, 1:1000, Abclonal Technology, China), tetraspanin-4 (TSPAN4, 1:1000, Bioss, China) and negative markers tumour susceptibility gene 101 (TSG101, 1:1000, Affinity, China), GM130 (1:1000, Affinity, China), Histone H3 (1:2000, Affinity, China), CD63 (1:1000, Affinity, China) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH, 1:2000, Affinity, China) were detected by western blotting.

Cellular uptake of RAW-MS

RAW-MS were fluorescently labeled with WGA-Alexa Fluor^® 647 dye and resuspended in PBS at a concentration of 5 × 10¹⁰ particles per ml. BMSCs, HUVECs and MΦ were seeded at a density of 5 × 10⁴ cells per well in a 24-well tissue culture plate. After 6, 12, and 24 h of incubation, the uptake ratio was observed by CLSM and detected by flow cytometry. To further confirm the active phagocytosis of macrophages towards migrasomes, High-Content screening (Opera Phenix^™, Revvity, USA) was performed by labeling both migrasomes and cells.

Flow cytometry analysis

M0 MΦ were seeded at 1 × 10⁵ cells per well in 6-well plates and treated with 20 µg/ml RAW-MS. After 12 h of culture, the proportions of M1 (CD68⁺/CD86⁺/CD206⁻) and M2 (CD68⁺/CD86⁻/CD206⁺) macrophage subsets were quantified using flow cytometry. For cell surface marker staining, cells were washed with PBS containing 1% bovine serum albumin (BSA, Gibco Co., USA) and incubated with specific antibodies (eBioscience, USA) for 40 min at 4 °C in the dark. Cells were then resuspended in PBS with 2% paraformaldehyde (PFA, Solarbio, China). For intracellular staining, cells were fixed, permeabilized, and stained according to the manufacturer’s instructions. All the data were analyzed using FlowJo V10.

EDU staining

Cell proliferation was evaluated using a 5-Ethynyl-2’-deoxyuridine (EDU) assay kit (Beyotime Biotechnology Inc., China) following the manufacturer’s protocol. Briefly, cells were plated in 24-well plates at 2.0 × 10⁴ cells per well and grown for 24 h. EDU was then added to the culture at a concentration of 50 µM and incubated. Subsequent staining was performed using Apollo dye and a DNA counterstain. Images were captured and evaluated under a CLSM.

Scratch wound healing assay

HUVECs were cultured to 80–90% confluence in 6-well plates and synchronized in low-serum medium (1% FBS) to suppress proliferation. A straight scratch was made in the cell monolayer using a pipette tip. After removing non-adherent cells with PBS, the remaining cells were co-cultured with sample extracts for 6 and 12 h, stained with calcein AM (Solarbio, China) and observed under fluorescence microscope (Zeiss, Germany).

Transwell migration assay

HUVECs (1 × 10⁴ cells/well) were seeded in the upper chamber of a transwell insert (8 μm pore size, Corning, USA) with serum-free medium, while the lower chamber contained complete medium supplemented with RAW-MS. After 24 h of co-cultivation, cells that migrated to the lower chamber were fixed with 4% PFA for 15 min and stained with 0.4% crystal violet (Solarbio, China) for 20 min. Migrated cells were imaged at five random fields using an optical microscope (Zeiss, Germany), and cell counts were quantified using ImageJ software.

Tube formation assay

A 24-well plate was coated with 100 µL of Matrigel Matrix (Corning, USA) and incubated with sample extracts at 37 °C for 30 min. After gelation, HUVECs were seeded onto the gel at a density of 3 × 10⁴ cells/cm² in RAW-MS containing medium. Following 6–12 h of incubation, the vascular-like structures formed by HUVECs were stained with calcein AM. The tube formation results were imaged using a fluorescence microscope.

Alkaline phosphatase (ALP) staining

To assess osteogenic differentiation, BMSCs were seeded in 24-well plates and cultured in growth medium until 80–90% confluence. The medium was then replaced with osteogenic induction medium (OM), consisting of complete medium supplemented with 1% β-glycerophosphate, 0.05% ascorbic acid, and 0.01% dexamethasone. In the MS group, RAW-MS were added simultaneously. After 7 days of culture, ALP activity was qualitatively and quantitatively assessed using a BCIP/NBT ALP color development kit (Beyotime Biotechnology Inc., China) and an ALP assay kit (Beyotime Biotechnology Inc., China), respectively.

Alizarin red staining

BMSCs were cultured and induced following the same protocol as ALP staining. Calcium deposition was evaluated at 21 days using Alizarin Red S staining. Briefly, cells were washed twice with PBS, fixed for 30 min, and stained with 0.1% Alizarin Red S (Beyotime Biotechnology Inc., China) for 30 min, followed by a rinse with distilled water. ALP activity and mineralized nodules were visualized and photographed under an optical microscope.

Immunofluorescence (IF) staining of cells

After incubation according to the experimental groups, the HUVECs, BMSCs and MΦ were fixed, permeabilized, blocked, and incubated with primary antibodies against osteocalcin (OCN, 1:1000, Servicebio, China), bone morphogenetic protein 2 (BMP-2, 1:1000, Affinity, China), CD34 (1:1000, Servicebio, China), delta-like ligand 4 (DLL4, 1:1000, Affinity, China) at 4 ℃ overnight. Next, the cells were co-incubated with corresponding Alexa Fluor^® 594 conjugated secondary antibodies (1:200, Solarbio, China) at RT for 1 h. As for the cell morphology observation, the cells were fixed with 4% PFA and stained with the phalloidin-Alexa Fluor 488 (1:200, Solarbio, China). Afterwards, the cells were stained with DAPI. Finally, the cells were characterized with CLSM. The mean fluorescence intensity was quantified using ImageJ software.

Real-time quantitative polymerase chain reaction (RT-PCR)

HUVECs, BMSCs and MΦ were seeded onto 6-well plates and incubated for 24 h in accordance with the relevant experimental groups and ctrl group. Total RNA was extracted from cells using an RNA-Quick Purification Kit (Accurate Biology, China) after 7 days cultivation. Complementary DNA (cDNA) was synthesized from the RNA using a PrimeScript RT Reagent Kit (Takara Biotechnology, Japan). Amplification and fluorescence detection were performed using a real-time PCR system. The mRNA expression levels were normalized to the housekeeping gene GAPDH, and relative quantification was calculated using the 2^−ΔΔCT method. Primer sequences used in the study are provided in Table. S1.

Western blot

Total protein was extracted from each group, and concentrations were determined with a BCA protein assay kit. Proteins were separated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE, Genscript, China) and transferred onto polyvinylidene difluoride (PVDF, Millipore, USA) membranes. The membranes were blocked for 1 h at room temperature with 5% non-fat milk in TBST. They were then incubated with primary antibodies (1:1000) overnight at 4 °C. After washing three times with TBST, the membranes were incubated with horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG secondary antibodies (1:5000, Affinity, China) for 1 h at room temperature. Protein bands were detected using an enhanced chemiluminescence (ECL, Millipore, USA) reagent.

Besides migrasomes-specific antibodies mentioned above, primary antibodies against CD86 (1:1000, HUABIO, China) and CD206 (1:1000, HUABIO, China) were used to detect polarization of MΦ. Primary antibodies against vascular endothelial growth factor (VEGF, 1:1000, Affinity, China), CD31 (1:1000, Servicebio, China), BMP-2 (1:1000, Affinity, China) and runt-related transcription factor 2 (RUNX2, 1:1000, Affinity, China) were respectively applied to assess angiogenic and osteogenic ability. Antibodies against transforming growth factor β1 (TGFβ1, 1:1000, HUABIO, China), Sma- and Mad-related proteins (Smad2, 1:1000, HUABIO, China) and pSmad2 (1:1000, HUABIO, China) were adopted to examined the TGFβ signaling pathway level. Moreover, antibodies against kinase insert domain receptor (KDR, 1:1000, Affinity, China), CD34 (1:1000, Servicebio, China) and DLL4 (1:1000, Affinity, China) were supplied for activating tip cells phenotype evaluation.

Proteomics analysis for RAW-MS

RAW-MS and RAW 264.7 cell lysates were prepared using lysis buffer followed by 5 min of sonication on ice. After centrifugation at 15,000 × g for 10 min at 4 °C, protein concentrations in the supernatants were determined using a BCA Kit (Beyotime, China). The protein samples underwent digestion and desalination. For mass spectrometry, peptides were separated using a Vanquish Neo UHPLC system and analyzed on an Orbitrap Astral mass spectrometer (Thermo Fisher Scientific, USA) using Data-Independent Acquisition (DIA) mode. DIA data were processed via DIA-NN software (v1.8.1) using a library-free approach. To ensure high-confidence identification, both precursor and protein identifications were filtered at a 1% False Discovery Rate (FDR). For comparative analysis, proteins of interest were defined as those significantly upregulated in the RAW-MS group compared to the cell lysate control, applying a threshold of log₂(Fold Chang) > 1 and P < 0.05. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium with the dataset identifier PXD078332.

RNA sequencing analysis

HUVECs were plated in 6-well plates and treated with 0–20 µg/mL RAW-MS for 48 h. Total RNA was isolated using RNAzol reagent (Invitrogen, USA). RNA sequencing libraries were prepared and subjected to transcriptome analysis on the Illumina HiSeq^™ platform. Differentially expressed genes were identified using a threshold of log₁.₅(Fold Change) > 1 with false discovery rate (FDR) < 0.05 for subsequent analysis. The raw RNA-seq generated in this study have been deposited in the Genome Sequence Archive (GSA) at the China National Center for Bioinformation (CNCB) under accession number HRA017710.

Inhibiting TGFβ1/Smad2 signaling pathway

Small interfering RNAs (siRNAs) targeting the human TGFβ1 gene from Suzhou GenePharma Co., Ltd. (Suzhou, China) were designed and synthesized to investigate the role of TGFβ1/Smad2 signaling pathway in RAW-MS mediated angiogenic induction of ECs. Three specific siRNA candidates were utilized to ensure effective gene knockdown. The sense and antisense sequences are summarized in Table. S2 and a non-targeting scrambled siRNA sequence was used as a negative control (NC) in all experiments.

Briefly, When the HUVECs reached 50–60% confluence, transient transfection was performed using Lipofectamine 3000 (Invitrogen, USA) according to the manufacturer’s instructions. To evaluate the silencing efficiency, cells were harvested 72 h post-transfection for protein analysis.

Finally, the HUVECs (1.0 × 10⁵ cells/well) in NC group either si- TGFβ1 group were seeded on 6-well plates and cultured with 20 µg/ml RAW-MS.

Animals and experimental protocol

A critical-sized rat cranial defect model was used to evaluate bone regeneration. Ten male SD rats (8-week-old, 250–280 g) were randomly allocated into two groups (n = 5 per group) using a random number table. Sample size (n = 5) was determined based on a power analysis to ensure a power of 0.80 with an alpha level of 0.05. Gelatin methacrylate (GelMA), a photopolymerizable derivative of gelatin, was utilized to stabilize particles at the bone defect site. The GelMA (EFL-GM-90) was provided by the Intelligent Manufacturing Research Institute of Suzhou, China. Under isoflurane anesthesia, a midline incision approximately 15 mm in length was made along the skull’s dorsal surface. Following blunt dissection, a circular defect with a 5 mm diameter was created on the left side of the central suture using a trephine drill. GelMA hydrogel, either alone or loaded with RAW-MS, was implanted to fully cover the defect area. All animal experiments were approved by the Animal Ethics Committee of Sun Yat-sen University (Approval No. SYSU-IACUC-2023-000059) and conducted in accordance with applicable institutional and national guidelines.

At 4 weeks post-surgery, rats were euthanized, and cranial bones were harvested and fixed in 4% PFA for 48 h. The samples were analyzed using micro-computed tomography (micro-CT) at high resolution (55 kV, 135 µA, 8 W) with a threshold range of 212–1000. Quantitative metrics included bone volume fraction (BV/TV), connectivity density (Conn.D), trabecular spacing (Tb.Sp), trabecular number (Tb.N), and trabecular thickness (Tb.Th). To minimize bias, micro-CT reconstruction and histological assessments were performed by two independent investigators who were blinded to the group allocation.

Histological analysis

For histological analysis, the cranial bone samples were decalcified in EDTA solution for one month, embedded in paraffin, and sectioned at 5 μm thickness. Hematoxylin and eosin (H&E) staining and Masson’s trichrome staining were performed.

Immunohistochemical (IHC) staining was conducted using anti-OCN (1:200) and anti-RUNX2 (1:200) antibodies to identify osteoprogenitor cells, while anti-CD31 (1:200) antibodies and anti-VEGF (1:200) antibodies were used to label blood vessels. Images were acquired using a light optical microscope. The staining protocols were as follows: after dewaxing and rehydration, antigen retrieval was performed by incubating sections in citrate buffer (pH 6.5) at 80 °C. After that, sections were incubated with primary antibodies at 4 °C overnight, followed by HRP-conjugated goat anti-mouse/rabbit secondary antibodies for 30 min. Finally, sections were treated with 3,3′-diaminobenzidine (DAB, Servicebio, China) and counterstained with hematoxylin.

For tissue IF staining, the procedure followed the protocol described in ‘Immunofluorescence staining of cells’ part with minor modifications. Briefly, after antigen retrieval, sections were permeabilized and blocked before overnight incubation at 4 °C with antibodies against CD31 (1:200), EMCN (1:200, Bioss, China), CD86 (1:200, Affinity, China), or CD206 (1:200, HUABIO, China). After secondary antibody incubation and DAPI counterstaining, images were acquired via CLSM. Quantitative analysis was conducted using ImageJ.

Statistical analysis

All in vitro experiments were performed with at least three independent biological replicates (n ≥ 3), while in vivo studies used 5 rats per group (n = 5). Data are presented as mean ± standard deviation (SD). Statistical analysis was performed using GraphPad Prism 9.0. For comparisons between two groups, a two-tailed Student’s t-test was used. For multiple comparisons, one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test was applied for normally distributed data. Normality was assessed using the Shapiro-Wilk test. Non-normally distributed data were analyzed using the Kruskal-Wallis test. Significance levels were defined as *P < 0.05, **P < 0.01, ***P < 0.001, and NS (not significant).

Results

Characterization of RAW-MS

Recent research has demonstrated that RAW 264.7 cells, owing to their pronounced migratory capabilities, actively released migrasomes [28]. The SEM analysis revealed an abundance of RFs characterized by elongated tubular formations at the cellular base, with vesicular structures emerging at their extremities and junctions after treating with FN (Fig. 1A). Additionally, certain migrasomes were observed to detach from the RFs through a process of self-rupture. In contrast, control cells showed substantially fewer RFs and migrasomes. Fluorescently conjugated WGA, known for its specific affinity to sialic acid and N-acetyl-D-glucosamine, has been utilized to identify migrasomes within cellular environments [29]. Consequently, WGA was employed as a migrasomes-specific fluorescent probe in the study. Observations of the FN group from Fig. 1B indicated the presence of numerous migrasomes emitting intense fluorescence at the RFs’ tips and intersections, which could either form bead-like chains or remain as discrete entities. In comparison, cells in Ctrl group demonstrated markedly shorter RFs and reduced migrasome counts. In term of morphology, macrophages in the control group maintained spherical shapes with minimal pseudopodia, whereas FN treatment induced extensive cellular spreading and spindle-like morphogenesis (Fig. 1A-B). TEM revealed fundamental distinctions between exosomes and migrasomes (Fig. 1C). Ctrl cells exhibited 100 nm exosomes sprouting out via membrane fusion. Migrating FN-treated RAW 264.7 cells displayed larger (approximately 500 nm) migrasomes tethered to RF termini or dispersed peripherally through RF rupture. The image of high magnification corroborated the presence of intraluminal nanovesicles inside the migrasomes at the size of 500 nm. The term ‘pomegranate body’ aptly always describes migrasomes due to their resemblance to an open pomegranate and the variable quantity of nanovesicles they encapsulate [30]. In addition, most exosomes-like migrasomes-derived nanoparticles (MDNP) also released out from the migrasomes.

To characterize the migrasomes isolated from RAW 264.7 via gradient centrifugation, TEM, western blotting, and NTA were employed. The extracted migrasomes exhibited a monolayer configuration with a predominantly spherical morphology, some of which remained affixed to RFs (Fig. 1D). The TEM image of ultra-thin section clearly showed ‘the pomegranate body’ – the typical structural features of migrasomes (Fig. 1E). Moreover, the images also showed empty migrasomes and the releasing MDNPs, agreeing to the previous studies that migrasomes remained structurally stable for only approximately 400 min and tended to rupture, releasing MDNPs upon maturation or detachment from the cell [9].

Western blot analysis was performed to verify the identity and purity of the isolated RAW-MS (Fig. 1F). TSPAN4 is recognized as essential for the migrasomes formation but was also present on exosomes [31]. NDST1, PIGK, CPQ and EOGT were highly enriched in migrasomes, but were either absent or barely detectable in exosomes [32]. The results confirmed that all three migrasome-specific markers were strongly expressed in the RAW-MS fraction. To further assess purity, potential contaminants were tested. The exosomal markers TSG101 and CD63 were notably absent in the migrasome samples [28, 33]. Furthermore, the isolated fraction showed no expression of intracellular organelle marker GM130, nuclear marker Histone H3 or cytosolic protein GAPDH, which indicated that the extraction was free from non-specific intracellular components. Collectively, these data demonstrate that our isolation protocol yields highly purified and molecularly defined migrasomes.

NTA data (Fig. 1G-H) indicated that the migrasomes had an average diameter of 217.3 nm and a zeta potential of −29.43 ± 1.05 mV. Size distribution of collected migrasomes ranged broadly (10–800 nm), peaking at 175 nm. This contrasted with literature reports of larger diameters (500–3000 nm) [34], potentially attributable to copurification of releasing MDNPs and migrasomes during gradient separation. Liu et al. demonstrated that MDNPs serve as functional subunits of migrasomes; their smaller size enables them to travel through the circulatory system and exert biological effects in locations inaccessible to intact migrasomes [35]. Therefore, in subsequent experiments, MDNPs were not separated from migrasomes but were instead studied as an integrated entity.

Cellular uptake of RAW-MS

To explore the internalization of RAW-MS, BMSCs, HUVECs and MΦ were employed as target cells for in vitro uptake experiments. Flow cytometry was utilized to determine the proportion of cells internalizing migrasomes. The uptake rate of RAW-MS showed a time-dependent increase after 6, 12, and 24 h of incubation in RAW-MS-containing medium (Fig. 2A-C). By 24 h, nearly all cells had internalized RAW-MS. Notably, a subset of cells exhibited strong positivity after 24 h, whereas minimal uptake was observed at 6 and 12 h, suggesting that RAW-MS can accumulate and persist intracellularly. Fluorescent staining was further employed to visualize and quantify RAW-MS uptake (Fig. 2D-I). WGA-labeled RAW-MS displayed a characteristic granular morphology and progressively accumulated in the cytoplasm over time. The High-Content system vividly presented the phagocytosis process of cells toward to the labeled migrasomes (Fig. 2J-L and Movie S1-3). The results above all confirmed that BMSCs, HUVECs as well as MΦ efficiently internalized and retained RAW-MS, which suggested that the RAW-MS may play important role in intercellular communication of angio-osteogenesis coupling.

Fig. 2.

Cellular uptake of RAW-MS by BMSCs, HUVECs and MΦ. (A-C) Flow cytometry images and corresponding quantification results of cellular uptake of migrasomes by (A) HUVECs, (B) BMSCs and (C) MΦ, (n = 3, one-way ANOVA test). (D-I) Representative fluorescent images and statistic calculation of migrasomes uptake by (D-E) HUVECs, (F-G) BMSCs and (H-I) MΦ (n = 3, one-way ANOVA test). (J-L) Sequential time-lapse images showing the phagocytosis of migrasomes: Migrasomes (red), cells (green), and nuclei (blue) (J: HUVECs K: BMSCs L: MΦ). *P < 0.05, **P < 0.01, ***P < 0.001

M2 polarization of MΦ treated with RAW-MS

Initially, the polarization state of RAW 264.7 macrophages following 12 h co-culture with FN was evaluated. The results in Fig. S1 revealed that FN exposure modestly shifted RAW 264.7 cells from the undifferentiated M0 state (F4/80⁺CD86⁻CD206⁻) toward an M2-like profile (F4/80⁺CD86⁻CD206⁺), concomitant with migrasome production.

The assessments of macrophage polarization were all performed 12 h after the addition of RAW-MS. The immunomodulatory capacity of RAW-MS was also investigated in Fig. 3. IF staining results (Fig. 3A-B) demonstrated that MΦ in MS group had significantly elevated CD206 expression than that in Ctrl group. Conversely, mean fluorescence intensity of CD86 displayed a non-significant reduction in the MS-treated group compared to controls. Morphologically, MΦ in Ctrl group maintained a round or oval morphology with minimal pseudopodia, whereas MΦ in MS group adopted elongated or spindle-like shapes. Flow cytometric quantification (Fig. 3C-D) confirmed this polarization shift: M2 macrophages (CD68⁺CD86⁻CD206⁺) increased significantly (from 1.29% ± 0.138% to 16.90% ± 0.141%), whereas M1 populations (CD68⁺CD86⁺CD206⁻) decreased markedly (from 8.72% ± 0.514% to 1.68% ± 0.123%) following RAW-MS treatment. RT-PCR analysis (Fig. 3E) corroborated these findings, showing upregulated transcription of M2-associated markers (CD206, Arg-1, IL-10) relative to M1 signature genes (CD80, CD86, iNOS) in the MS group. Moreover, western blotting results (Fig. 3F-G) also showed MS group had higher expression of CD206 and lower expression of CD86 in the comparison to Ctrl group. Collectively, these data demonstrated that that RAW-MS directly trigger M0-to-M2 transition, evidenced by the rapid, synchronized upregulation of M2 markers at both mRNA and protein levels.

Fig. 3.

Polarization of MΦ culturing with RAW-MS. (A) Polarization of MΦ detected by confocal microscope and (B) Fluorescence intensity analysis of CD86 and CD206 (n = 3, t-test). (C) Polarization of MΦ detected via flow cytometry and (D) Quantification of the percentage of M1 (CD68⁺/CD86⁺/CD206^-) and M2 (CD68⁺/CD86^-/CD206⁺) MΦ (n = 3, t-test). (E) qPCR analysis of M1 and M2 polarization phenotype-related gene expression in MΦ (n = 3, t-test). (F) Western blotting images of CD86 and CD206 markers expression and (G) Quantified protein levels of CD86 and CD206 (n = 3, t-test). *P < 0.05, **P < 0.01, ***P < 0.001, NS: No significant

Effect of RAW-MS on angiogenesis and osteogenic differentiation in vitro

The EDU assay demonstrated that RAW-MS significantly enhanced the early proliferation of HUVECs (Fig. S2A-B). The scratch assay is a convenient and available method of assessing cell migrative capability in vitro [36]. The results (Fig. 4A-B) revealed that the migration rate of HUVECs in the MS group was markedly higher than that in the control group at both 6 h and 12 h. Consistent with these findings, transwell experiments further confirmed the enhanced migratory capacity of HUVECs treated with RAW-MS (Fig. 4C-D). Additionally, the capillary-like tube formation assay was often conducted to evaluate angiogenesis ability of ECs in vitro [37]. The results showed that the MS group exhibited increased numbers of nodes, and junctions, compared to the control group, suggesting that RAW-MS facilitated vascular network formation in vitro (Fig. 4E-F). The RT-PCR analysis (Fig. 4G) revealed elevated expression of angiogenesis-related genes, including HIF-1α, CD31, EMCN, vWF, and VEGF, in HUVECs cultured with RAW-MS for 3 days. Western blot analysis (Fig. 4H-I) further supported these observations, demonstrating upregulated protein levels of HIF-1α and VEGF. These findings suggested that RAW-MS may stimulate angiogenesis through the HIF-1α/VEGF signaling pathway.

Fig. 4.

Effect of RAW-MS on angiogenesis in vitro. (A) Representative images of cell scratch assay and (B) Quantitative statistics of wound healing area (n = 3, t-test). (C) Representative images of migration assay and (D) Quantitative statistics of transmigrated cells (n = 4, t-test). (E) Representative images of tube formation assay and (F) Quantitative statistics of nodes and junctions (n = 4, t-test). (G) RT-PCR analysis of angiogenic related genes (n = 3, t-test). (H) Western blotting images of CD31 and VEGF protein expression and (I) the quantified protein levels of CD31 and VEGF (n = 3, t-test). *P < 0.05, **P < 0.01, ***P < 0.001

The positive effect of RAW-MS on the BMSCs proliferation was also shown by the EDU results in Fig. S2C-D. ALP staining (Fig. 5A-B) demonstrated that RAW-MS enhanced early osteogenic differentiation. Alizarin red staining further confirmed its positive role in promoting calcium nodule formation (Fig. 5C-D). IF staining of osteogenic markers, including OCN and BMP-2, revealed that RAW-MS significantly increased osteogenic differentiation in BMSCs (Fig. 5E-F). Moreover, RT-PCR and western blotting analyses showed upregulated expression of osteogenic genes (ALP, RUNX2, OPN, OCN, BMP-2 and Osterix) and proteins (RUNX2 and BMP-2), further validating the pro-osteogenic effects of RAW-MS on BMSCs in vitro (Fig. 5G-H).

Fig. 5.

Effect of RAW-MS on osteogenesis in vitro. (A) Representative images of ALP staining and (B) ALP activity assay (n = 3, t-test). (C) Representative images of Alizarin Red staining and (D) Quantitative statistics for ECM mineralization (n = 3, t-test). (E) Representative IF images of BMP-2 and OCN and (F) corresponding fluorescence intensity analysis (n = 3, t-test). (G) RT-PCR analysis of angiogenic related genes (n = 3, t-test). (H) Western blotting images of RUNx2 and BMP-2 protein expression and (I) The quantified protein levels of RUNx2 and BMP-2 (n = 3, t-test). *P < 0.05, **P < 0.01, ***P < 0.001

Quantitative proteomics analysis of RAW-MS and RNA-seq analysis of the effect of RAW-MS on HUVEC

It is well established that the monocyte-derived migrasomes are enriched with pro-angiogenic factors and modulate the angiogenic behavior of target cells [22]. To investigate this phenomenon, the proteomic analysis on RAW-MS was performed with RAW 264.7 cell lysates as a comparative control. Volcano plot and heatmap analyses revealed a distinct proteomic signature in migrasomes, where 4,371 proteins were significantly enriched in RAW-MS compared to donor cells (Fig. 6A-B). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis (Fig. 6C) identified the PI3K-Akt signaling pathway, a master regulator of angiogenesis, as one of the most significantly represented terms [38]. The proteins enriched in PI3K-Akt signaling pathway within RAW-MS were performed in Fig. 6D. Specifically, 115 functional proteins associated with the PI3K-Akt axis were encapsulated within RAW-MS. The Voronoi treemap (Fig. 6D) further illustrated the hierarchical distribution of these proteins, highlighting an abundance of key adhesion molecules such as Integrin subunits and vitronectin (VTN). It also revealed signaling transducers including Rac1/2, Jak1, and Gnb1/2, as well as other protein involving in cell cycle and endoplasmic reticulum. Furthermore, targeted analysis of the RAW-MS cargo (Fig. S3A) confirmed the enrichment of cytokines and chemokines, such as TGFβ1 and CXCL2.

Fig. 6.

Quantitative proteomics of RAW-MS and RNA-seq of HUVEC (A) Volcano map of regulated proteins of RAW-MS vs. RAW 264.7 cells. (B) Heatmap of regulated proteins of RAW-MS vs. RAW 264.7 cells. (C) KEGG classification of enriched proteins within RAW-MS. (D) Voronoi treemap analysis of proteins enriched in the PI3K-Akt signaling pathway within RAW-MS. (E) Volcano map of regulated genes of HUVEC treated with RAW-MS in RNA-seq. (F) GO enrichment results, BP: Biological process, CC: Cellular component, MF: Molecular function. (G) KEGG enrichment results

To comprehensively investigate the mechanisms by which RAW-MS regulates angiogenesis, the gene expression profiles of HUVECs treated with RAW-MS (0 vs. 20 µg/mL) for 24 h analyzed using RNA-seq. Compared to the control group, 593 statistically significant differentially expressed genes (DEGs) were identified (|Fold Change| ≥ 1.5; False discovery rate ≤ 0.05), comprising 376 upregulated and 226 downregulated genes (Fig. 6E). Gene Ontology (GO) enrichment analysis indicated that RAW-MS treatment primarily affected “regulation of angiogenesis” and “regulation of vasculature development” (Fig. 6F). Furthermore, KEGG enrichment analysis demonstrated significant enrichment of RAW-MS-induced DEGs in the TGFβ signaling pathway (Fig. 6G). Heatmap visualization in Fig. Sconfirmed that RAW-MS upregulated key genes within the TGFβ signaling pathway (ID2, ID3, NOG, MYC, THBS1) while downregulating LTBP1 and TNF expression (Fig. S3B).

Promotion of RAW-MS on angiogenic sprouting via TGFβ1/Smad2 signaling pathway

To exclude the potential confounding effects of FN coating, RAW-MS derived from macrophages cultured on untreated surfaces (n-MS group) retained their intrinsic capacity to promote angiogenesis and osteogenesis (Fig. S5A-D), as well as to activate the TGFβ1 signaling pathway were verified (Fig. S5E–F). Having confirmed that these effects were intrinsic to the RAW-MS, the specific role of TGFβ1/Smad2 signaling were next investigated. The activation of TGFβ was found to induce the phosphorylation of downstream Smad transcription factor [39]. Therefore, the functional role of the TGFβ1/Smad2 signaling pathway in RAW-MS-mediated angiogenesis promotion was assessed. Based on the validation in Fig. S4, si-TGFβ1–3 was selected to silence TGFβ1 expression in ECs prior to co-culture with RAW-MS. Quantitative western blot analysis revealed that RAW-MS treatment elevated TGFβ1 protein levels by approximately 22.6% and increased Smad2 phosphorylation by 32.0% relative to NC group. Conversely, the knockdown of TGFβ1 suppressed basal TGFβ1 expression by 47.8% and decreased pSmad2/Smad2 levels by 53.3%. (Fig. 7A-B), suggesting that RAW-MS rely on endogenous TGFβ1 to activate the canonical Smad pathway.

Fig. 7.

Examination of the effects of TGFβ1/Smad2 signaling on RAW-MS promotion for angiogenesis. (A) Western blotting images of TGFβ1, pSmad2/Smad2 protein expression and (B) The quantified protein levels analysis (n = 3, one-way ANOVA test). (C) Representative images of cell scratch assay and (D) Quantitative statistics of wound healing area (n = 3, one-way ANOVA test). (E) Representative images of migration assay and (F) Quantitative statistics of transmigrated cells (n = 3, one-way ANOVA test). (G) Representative images of tube formation assay and (H) Quantitative statistics of nodes and junctions (n = 3, one-way ANOVA test). (I) Representative IF images of CD34/DLL4 dual positive EC and (J) Corresponding fluorescence intensity analysis (n = 3, one-way ANOVA test). (K) Western blotting images of KDR, CD34, DLL4 expression and (L) The quantified protein levels analysis (n = 3, one-way ANOVA test). *P < 0.05, **P < 0.01, ***P < 0.001

Functional assays confirmed that TGFβ1 blockade significantly attenuated the pro-migratory effects of RAW-MS on ECs, a critical initiation step in angiogenesis. Both scratch wound healing assays (Fig. 7C-D) demonstrated impaired EC migration following TGFβ1 inhibition. The migratory capacity of ECs was further evaluated using Transwell assays. The results of Fig. 7E showed crystal violet-stained ECs that migrated to the underside of the membrane. Quantitative analysis (Fig. 7F) confirmed that TGFβ1 inhibition largely abolished the RAW-MS-induced enhancement of EC migration, consistent with the scratch wound healing results. abolished following TGFβ1 inhibition, consistent with the results of the scratch wound healing assays. Furthermore, tube formation assays indicated that TGFβ1-silenced ECs exhibited impaired angiogenic capacity, characterized by a significant reduction in tubular network nodes and junctions (Fig. 7G-H).

Mechanistically, IF images, RT-PCR and western blotting analysis provided additional further supported these findings: RAW-MS treatment markedly promoted a ‘tip cell’ phenotype, as evidenced by the increased expression of markers CD34, DLL4, and KDR. Notably, this pro-angiogenic activation was effectively neutralized by TGFβ1 knockdown (Fig. 7I-L and Fig. S6). These results collectively indicated that RAW-MS promoted angiogenesis via the activation of the TGFβ1/Smad2 signaling axis.

Promotion of RAW-MS on vascularized osteogenesis in vivo

This investigation employed murine calvarial defect models to evaluate the effects of RAW-MS on enhancing angiogenesis during bone regeneration. GelMA hydrogel commonly served as the primary scaffold material for defect repair [40]. Surgical implantation introduced either GelMA alone or RAW-MS/GelMA nanocomposites into critical-sized defects (Fig. 8A). Following defect creation, UV-crosslinked GelMA or RAW-MS/GelMA scaffolds were precisely delivered into osseous voids (Fig. 8B). Post-implantation micro-CT assessment at 4 weeks demonstrated substantially augmented bone formation in RAW-MS-treated specimens versus controls. 3D reconstructions visually confirmed reduced defect dimensions in the RAW-MS cohort relative to untreated rats (Fig. 8C). Quantitative evaluation revealed significantly elevated BV/TV in the RAW-MS group (42.09 ± 15.38%) compared to controls (10.53 ± 6.27%) (Fig. 8D). Trabecular microarchitecture analysis further indicated enhanced Conn.D and diminished Tb.Sp in RAW-MS-treated defects, though Tb.N and Tb.Th showed no statistical differences.

Fig. 8.

Micro-CT and immunohistochemistry analysis of rat critical-sized cranial bone defect model. (A) 3D fluorescence image of WGA-labeled GelMA-MS hydrogel. (B) Gross images of the surgical process. (C) Representative micro-CT 3D reconstruction images and (D) quantitative results of the BV/TV, Conn.D, Tb.N, Tb.Th and Tb.Sp (n = 5, t-test). (E-F) Representative (E) H&E and (F) Masson staining images of cranial bone, CF: Collagenous fiber, NB: New bone, MB: Mineralized bone. (G-H) Representative IHC staining images of (G) RUNx2, OCN and (H) CD31, VEGF. (I) Representative IF images of type H vessel, white arrow: Type H vessel. (J) Representative IF images of macrophage polarization, red triangle: CD86⁺ (M1), green triangle: CD206⁺ (M2). *P < 0.05, **P < 0.01, ***P < 0.001, NS: No significant

Fibrosis reflects the inflammation stage. Moderate fibrosis confined the further aggravation and expansion of inflammation but uncontrollable fibrotic encapsulation typically jeopardized tissue integration and function [41]. Histological evaluation via H & E and Masson staining corroborated denser and more mature bone tissue in the MS group, whereas the Ctrl group displayed predominantly fibrous structures (Fig. 8E-F). IHC staining of osteogenic and angiogenic markers (Fig. 8G-H and Fig. S7A-D) demonstrated intensified expression of mature osteoblast marker OCN and osteogenic transcription factor RUNx2 within neo-bone regions of RAW-MS-treated specimens [42]. Additionally, CD31-positive blood vessels were more numerous and larger in the MS group compared to controls. Furthermore, CD31⁺ micro-vessels exhibited greater density and luminal diameter in the experimental group. Notably, VEGFA - the principal regulator of type H vasculature - displayed upregulated expression with perivascular localization in the RAW-MS cohort.

Type H vessels, identified by co-expression of EMCN and CD31 (yellowish areas in Fig. 8I and Fig. S8A), are known to play a critical role in coupling angiogenesis and osteogenesis. In the Ctrl group, these vessels were primarily localized along the periphery of nascent bone, whereas the MS group exhibited a more extensive and uniform distribution throughout the newly formed bone. Quantitative analysis (Fig. S7E-F) confirmed a significant increase in type H vessel density in the MS group, consistent with the observed enhancement in vascularized osteogenesis. While controls exhibited predominant M1 polarization, RAW-MS treatment significantly promoted M2 macrophage activation (Fig. 8J and Fig. S8B). Quantitative analysis further confirmed that the differences in CD86 and CD206 expression between the two groups were statistically significant (Fig. S7G-H).

Discussions

Migrasomes is a unique type of extracellular vesicle released by migrating cells, constituting a significant focus in contemporary biomedical research. In this study, SEM, IF and TEM were employed to observe canonical migrasome biogenesis (Fig. 1). Although basal migrasome production occurred in untreated macrophages, FN stimulation markedly augmented migrasomes secretion. FN, an essential extracellular matrix (ECM) glycoprotein, provides structural and biochemical support to adjacent cells [43]. Crucially, FN-integrin α5β1 engagement serves as a well-established nexus for focal adhesion assembly and signal transduction [44, 45], a fundamental perquisite for migrasomes formation [46]. Consequently, augmenting ECM-integrin interactions and cellular adhesion enhances migrasomes biogenesis. Supporting this mechanism, Li et al. demonstrated that nano-topographic structures potentiate migrasomes generation by facilitating ECM protein adsorption and subsequent integrin receptor engagement [47].

Furthermore, FN exposure induced polarization of M0 macrophages toward an M2 phenotype. This observation aligns with established evidence that FN-functionalized biomaterials attenuate macrophage inflammatory responses and mitigate oxidative stress [48–50]. Specifically, FN adsorption onto hydrophilic titanium dioxide surfaces promotes anti-inflammatory M2 polarization via integrin-mediated signaling [51]. Gao et al. further proposed that FN may suppress NF-κB signaling, thereby reducing cellular stress and favoring M2 polarization [49]. Conventionally, macrophage-derived EVs mediate pathological and therapeutic processes across diverse disease contexts [52]. Notably, EVs originating from M2-polarized macrophages exhibit enhanced anti-inflammatory and pro-regenerative capacities [15, 53, 54]. Hence, these findings suggest that migrasomes, particularly those secreted by M2-polarized macrophages, may similarly facilitate tissue regeneration processes.

Subsequently, it is found that the RAW-MS could promote the M2 polarization of the targeted macrophages and improve angiogenic activity as well as osteogenic differentiation. Consistent with previous researches, the loaded macrophages-derived EVs could directly effectively induce reparative macrophage polarization [15, 55, 56]. During bone regeneration, timely transition to the M2 phenotype is essential to alleviate the foreign body response (FBR) since excessive M1 macrophage activation disrupts the local immune niche and promote abnormal fibrosis [15]. The vivo experiment showed that the bone defect in Ctrl group presented more fibrous tissue as well as higher ratio of M1 macrophages in the late phase of bone defect repair (Fig. 8). It agreed with the point that promoting the polarization of anti-inflammatory M2 macrophages at an earlier point benefited for the neo-bone regeneration [57, 58].

Proteomic analysis identified multiple pro-angiogenic factors within RAW-MS, with significant enrichment in the PI3K-Akt signaling pathway. The observed pro-angiogenic and osteogenic effects of RAW-MS in vitro likely represent the collective action of these protein components. Notably, integrins function as key cell surface receptors mediating cellular adhesion [59]. Integrins are essential for migraosmes formation. During migrasome formation, integrins are first targeted to the ends or branch points of retraction fibers to form integrin foci, which would later grow into migrasomes [60]. Emerging evidence indicates that integrin β1-enriched exosomes from Co²⁺-stimulated macrophages enhance endothelial migration and tubule-genesis independently of VEGF [61]. Similarly, Schwann cell-derived exosomes carrying integrin β1 promote angiogenesis and neurological recovery [62]. Among candidate factors, SPP1/osteopontin is particularly noteworthy. Single-cell RNA-seq highlighted that restoring impaired angiogenesis with supplement of SPP1 was a critical determinant of critical size bone defects reconstruction [63]. Consistently, Chun Yao et al. demonstrated macrophage-mediated angiogenesis via SPP1-dependent regulation of endothelial cell subsets following spinal cord injury [64].

In the KEGG analysis of RNA-sequencing, the TGFβ1/Smad2 signaling pathway had arisen special attention because its close relationship with migrasomes has been reported [65]. Moreover, the study of Abdullah Faqeer et al. reported SPP1-rich EVs from osteoclasts activated TGFβ signaling in MSCs to drive osteogenic differentiation [66]. Therefore, our work corroborated that RAW-MS may facilitate angiogenesis through dual-pathway activation of the TGFβ1/Smad2 axis. This synergy between directly encapsulated TGFβ1 cytokines and synergistic protein cargoes like SPP1 ensured the robust activation of downstream regenerative programs (Fig. 7 and Fig. S3). TGFβ1 is the predominant member of the TGFβ superfamily consisting of 33 members. TGFβ binding to the receptor triggers phosphorylation of intracellular R-Smads, translocating to the nucleus, complexing with transcriptional factor and modulating the expression of target genes [67]. In recent years, increasing attention has been paid to the connection between TGFβ signaling pathway and angiogenesis. For example, TGFβ1 participates in angiogenesis by regulating the stability of capillaries [68]. Maharaj et al. demonstrated that activated TGFβ1 induces VEGF production in endothelial [69]. Considering given that high-concentration VEGF initiates vascular sprouting [70], TGFβ1 may indirectly stimulate tip cell activation via VEGF induction. Therapeutically, RAW-MS-loaded GelMA hydrogels enhanced bone regeneration by promoting type H vessel formation-coupling angiogenesis with osteogenesis. Yukihiro Kohara identified macrophage-EC crosstalk via TGFβ signaling as essential for type H angiogenesis during wound healing [71]. Notably, type H endothelial cells express elevated TGFβ1/TGFβ3 levels compared to type L ECs, as established by type H vessel pioneer Kusumbe et al [72]. In recent years, the complex crosstalk between type H vessel ECs and macrophages has garnered significant interest [73]. Migrasomes likely modulate type H vessel formation through TGFβ signaling upregulation, though the definitive requirement of these vessels for bone repair remains to be established via targeted inhibition. Regardless, the recruitment of type H endothelium stands as a key characteristic of the migrasome-fostered niche, reinforcing their potential as novel signaling vectors in osteo-angiogenic coupling.

Compared to conventional EVs like exosomes, migrasomes offer distinct spatial and functional advantages for tissue regeneration. While exosomes diffuse freely into the systemic circulation, migrasomes remain uniquely anchored to the ECM [9, 74]. Recent literature suggests that this structural stability provides a ‘spatial controllable’ paradigm for bone cell crosstalk, overcoming the limitations of diffusive EVs that lack precise localization [75]. In bone repair, this anchorage-dependent property ensures precise, localized delivery of high-concentration signals directly to the defect site, thereby reducing off-target effects and maximizing therapeutic potency [20, 22].

Furthermore, migrasomes serve as stable signaling hubs that provide essential ‘regional cues’ to coordinate multiple cell types in the bone niche [21, 35]. Due to their large volume and ‘pomegranate-like’ structure, migrasomes can encapsulate more complex bioactive cargo than traditional EVs, including functional proteins and even organelles, which may enhance their signaling efficiency in osteo-angiogenesis coupling [76, 77]. Previous research confirmed that substrate surface properties could finely regulate migrasome biogenesis [47]. Therefore, engineered biomaterial surfaces may enable precise control over migrasome production, offering a new strategy for targeted signal delivery during vascularized bone regeneration.

In this study, the simplified in vitro environment may not fully replicate the complex mechanical and biochemical cues of the native bone niche. Our future research would focus on developing biomimetic scaffolds to achieve more precise spatiotemporal control over migrasomes release for clinical translation in large-scale bone defect repair.

Conclusion

In summary, this study demonstrated that fibronectin (FN) stimulation enhances the production of migrasomes derived from M2-polarized macrophages (RAW-MS). These vesicles significantly promoted macrophage polarization toward an M2 phenotype and facilitated angiogenesis-osteogenesis coupling by enhancing type H vessel formation in a critical-sized rat cranial defect model. Mechanistically, RAW-MS promoted tip cell activation and sprouting angiogenesis through upregulation of the TGFβ signaling pathway. As a newly identified class of extracellular vesicles, migrasomes uniquely convey both molecular cargo and spatial information, positioning them as a novel and promising paradigm for advanced tissue engineering strategies.

Author contributions

L.L. and W.J. conceived and designed the experiments. L.L., W.J., J.S. performed the most of the experiments, analyzed data and wrote the original manuscript. L.R., L. Y., L. S., and H.Y. revised and edited the manuscript. Z.W. and Y. D. provided resources and supervision. All authors read and approved the final manuscript.

Funding

This work was financially supported by the Natural Science Foundation of Guangdong Province (grant number: 2024A1515012918) and National Natural Science Foundation of China (grant number: 82373255).

Data availability

Data will be made available on request.

Declarations

Ethics approval and consent to participate

All animal experiments were approved by the Animal Ethics Committee of Sun Yat-sen University (Approval No. SYSU-IACUC-2023-000059) and conducted in accordance with applicable institutional and national guidelines.

Competing interests

The authors declare no competing interests.