Abstract Diabetic wounds are characterized by chronic inflammation, reduced angiogenesis, and insufficient collagen deposition, leading to impaired healing. Extracellular vesicles (EVs) derived from adipose-derived mesenchymal stem cells (ADSC) offer a promising cell-free therapeutic strategy, yet their efficacy and immunomodulation can be enhanced through bioactivation. In this study, we developed calcium silicate (CS)-stimulated ADSC-derived EVs (CSEV) incorporated into collagen hydrogels to create a sustained-release system for promoting diabetic wound healing. CSEV exhibited enhanced protein content, surface marker expression, and bioactive cargo enriched with pro-angiogenic and anti-inflammatory factors. In vitro, CSEV-loaded collagen significantly reduced reactive oxygen species production, promoted cell proliferation and migration compared to standard EV-loaded collagen. Cytokine profiling revealed the upregulation of anti-inflammatory cytokines and extracellular matrix components, highlighting their immunomodulatory and regenerative potential. In vivo, histological evaluation of diabetic rabbit models treated with CSEV-loaded collagen revealed superior reepithelialization and organized collagen deposition, indicating accelerated wound closure. These findings underscore the potential of CSEV-loaded collagen hydrogels as an innovative and effective therapeutic platform for enhancing diabetic wound healing by simultaneously addressing inflammation and tissue regeneration. Graphical abstract [44]graphic file with name 12951_2025_3097_Figa_HTML.jpg Keywords: Extracellular vesicles, Calcium silicate, Immunomodulation, Diabetic wound healing Introduction Diabetes mellitus (DM), a chronic metabolic disorder, constitutes a global health concern affecting millions of individuals and is associated with numerous complications [[45]1]. In addition to the challenges in glycemic control, increased infection susceptibility, and impaired angiogenesis are major issues accompanying DM. The pathophysiology of DM is complex, involving factors including peripheral neuropathy, vascular insufficiency, and impaired wound healing [[46]2]. In the diabetic state, the normal wound healing processes are disrupted, marked by prolonged inflammation, defective angiogenesis, and reduced collagen deposition [[47]3]. Consequently, these alterations result in chronic, non-healing wounds that are resistant to conventional treatments. Traditional approaches to diabetic wound care, such as wound debridement, infection control, offloading, and advanced dressings [[48]4], frequently produce suboptimal outcomes, underscoring the need for novel therapeutic strategies. Advances in wound dressing materials, such as nanofibers or multi-layered scaffolds, have demonstrated the potential to integrate angiogenesis promotion, antibacterial properties, and tissue regeneration through controlled drug release [[49]5, [50]6]. At the same time, addressing the chronic inflammation in diabetic wounds remains critical, highlighting the importance of immunomodulatory therapies to complement these innovations. In this context, regenerative medicine approaches, particularly those involving stem cells and their secreted factors, have garnered considerable interest in diabetic wound healing [[51]7]. ADSCs have shown promise in this context, as they can differentiate into various cell types and secrete paracrine factors that support wound repair [[52]8]. However, the direct use of stem cells in clinical applications faces several challenges, including potential tumorigenicity, immune rejection, and logistical issues related to cell preservation and delivery. Extracellular vesicles (EVs) are nano-sized, membrane-bound vesicles playing pivotal roles in mediating intercellular communication by transferring bioactive molecules between cells [[53]9]. They carry diverse cargoes of proteins, lipids, and nucleic acids-including mRNAs, microRNAs, and other signaling molecules, enabling them to influence a wide range of cellular processes in recipient cells, such as proliferation, migration, differentiation, and immune modulation [[54]10–[55]13]. Owing to their ability to naturally encapsulate and deliver therapeutic cargo, EVs have gained significant attention as promising cell-free therapeutics in regenerative medicine, particularly for wound-healing applications [[56]14]. In particular, mesenchymal stem cell (MSC)-derived EVs have been shown to promote critical aspects of wound healing, including enhancing cell proliferation, stimulating migration, and fostering angiogenesis, processes essential for tissue repair [[57]15, [58]16]. MSC-EVs can modulate the inflammatory response by reducing proinflammatory cytokines and promoting a more regenerative environment, which is crucial for the impaired healing observed in diabetic wounds [[59]17]. EVs have been shown to carry specific factors that are beneficial for wound healing [[60]18, [61]19]. MSC-derived EVs frequently contain pro-angiogenic factors such as VEGF, FGF, and TGF-β, which directly contribute to neovascularization, a critical step in restoring blood flow to damaged tissue in diabetic wounds [[62]20]. Additionally, they may carry extracellular matrix (ECM)-modulating enzymes, promote wound bed remodeling, and thus facilitate improved healing outcomes [[63]21]. Calcium silicate (CS), a well-established inorganic compound in biomaterials research, has recently gained renewed attention in regenerative medicine for its biocompatibility and capacity to influence cellular behavior [[64]22–[65]24]. CS enhances the proliferation and differentiation of various cell types, including fibroblasts and endothelial cells, both of which are critical for wound healing [[66]25–[67]27]. Additionally, CS-stimulated adipose-derived stem cells enhance angiogenesis and accelerate skin wound healing by significantly boosting cell proliferation, migration, and blood vessel formation. Moreover, releasing Calcium (Ca) and Silicon (Si) ions from the CS creates a favorable microenvironment that activates cells to promote EV secretion [[68]28–[69]30]. We have previously developed a CS-activated gelatin methacrylate hydrogel that accelerates the proliferation and differentiation of human dermal fibroblasts, highlighting its application potential for skin wound repair [[70]31]. In this study, we utilized CS-stimulated ADSC to produce EVs (termed CSEV) and further evaluated the benefits of CSEV for diabetic wound treatment. As illustrated in Fig. [71]1, the FiberCell system was used to enable the scalability of CSEV production for therapeutic applications. The physical and biological properties of CSEV were characterized through an in vitro diabetic human dermal fibroblast (HDF) model and an in vivo diabetic rabbit wound model. Additionally, the miRNA and protein content related to the regulation of angiogenesis, tissue regeneration, and inflammation were identified, supporting the therapeutic potential of CSEV for diabetic wound treatment. Fig. 1. [72]Fig. 1 [73]Open in a new tab Schematic representation of the experimental process and therapeutic application of CSEV for diabetic wound healing. The top-left panel illustrates the stimulation of ADSC by calcium (Ca) and silicon (Si) ions to produce EVs enriched with growth factors and anti-inflammatory cytokines. The lower-left panel shows the FiberCell System used for large-scale production of CSEV. The right section outlines the application of CSEV-loaded collagen hydrogel on diabetic wounds, with the progression from an untreated wound to a repaired state. The treatment aims to support wound healing, angiogenesis, hair follicle neogenesis, and reduce inflammation Materials and methods Preparation and characteristics of calcium silicate Calcium oxide (Sigma-Aldrich, St. Louis, MO, USA), silicon dioxide (Sigma-Aldrich), and aluminum oxide (Sigma-Aldrich) powders were sequentially added to a beaker at ratios of 70%, 25%, and 5%, respectively. The mixture was placed in a high-temperature sintering furnace, where it was gradually heated at a rate of 10℃/min until it reached 1400℃. This temperature was maintained for 2 h before the furnace was allowed to cool naturally to room temperature. Subsequently, the sintered powder was mixed with anhydrous ethanol in a specified ratio and ground in a planetary ball mill at a rotational speed of 300 rpm for 8 h. The resultant solution was poured into a beaker and placed in an oven at 100℃ overnight to ensure complete ethanol evaporation. For the crystalline phase analysis, X-ray diffraction (XRD) was employed (Bruker D8 SSS, Karlsruhe, Germany) to determine the crystal structure through characteristic peaks. Scanning was performed continuously at a rate of 1 °/min with the scan range set from 20° to 50°. The surface morphology of the calcium silicate was examined using field-emission scanning electron microscopy (FE-SEM, JEOL JSM-7800 F, Tokyo, Japan). CS extract preparation To prepare the CS extract solution, the powder was immersed in a commercial basal medium (Mesenchymal Stem Cells Medium, MSCM, ScienCell, CA, USA) at a ratio of 1 g/40 mL. The solution was incubated in a 37℃-water bath overnight and filtered through a 0.22 μm pore filter to obtain the CS extract solution needed for further experiments. The released Ca and Si ions concentration of the CS extract solution was analyzed by inductively coupled plasma-atomic emission spectrometry (PerkinElmer OPT 1MA 3000DV; Shelton, CT, USA). ADSC culture and behaviors Human adipose-derived mesenchymal stem cells (ADSC, ScienCell) were purchased from ScienCell. The cells were cultured in MSCM and maintained in a humidified incubator at 37℃ with 5% CO[2]. The culture medium was replenished every three days, and the cells were passaged as needed based on their growth status. Cells between passages 4 and 9 were used in all experiments in this study. To evaluate the effect of CS extract on ADSC, we cultured the cells in 6 wells (5*10^4 cells/well) and replaced the culture medium with CS extract after 1 day. After different culture time points, cell number was assessed using PrestoBlue^® Cell Viability Reagent. In addition, we collected the cells and used ELISA to analyze the expression of angiogenic-related proteins. Each experimental condition was tested in sextuplicate. EV production and purification The ADSC were seeded into hollow-fiber bioreactors (FiberCell Systems, Inc., Frederick, MD, USA) with MSCM or CS extract-contained MSCM and collected medium every 3 d. The EVs were isolated using a tangential flow filtration system (MWCO 10KD, mPES hollow fiber, Spectrum, mounted with a MAP-TFF System, LEF Science) and filtered through membranes with a molecular weight cut-off of 100 kDa. All EVs were filtered with a 0.22-µm filter paper, resuspended in phosphate-buffered saline (PBS, Invitrogen, Carlsbad, CA, USA), and stored at 4℃ for immediate use. EVs obtained from MSCM were labeled EV, while those obtained from cultures stimulated with CS extracts were labeled CSEV. EVs characterization To determine the size and concentration of EVs, Nanoparticle Tracking Analysis (NTA) was performed using a clean and calibrated ZetaView^® instrument. Samples were filtered through a 0.22-µm membrane filter and diluted in PBS to achieve a final volume of 1 mL, ensuring a 50–200 particle count per field of view to obtain accurate size distribution curves and concentration data. The zeta potential measurements were performed at two positions for each sample to ensure consistency. Transmission electron microscopy (TEM; JEOL JEM-1400, Japan) was employed for morphological examination. Samples were fixed with glutaraldehyde for 10 min and deposited on copper grids. The grids were stained with 2% phosphotungstic acid and washed thrice with deionized water to remove the excess liquid. After preparation, all samples were placed in a 37℃ oven to dry before TEM imaging. The total protein content of different EVs was measured using a BCA protein assay kit. Additionally, EVs displaying CD9, CD63, and CD81 expression were quantified using an ExoCounter (JVCKENWOOD Corporation, Yokosuka, Japan). The optical disk was attached to a removal plate with 16 wells for EV injection. Each well was pre-coated overnight with 5 mg/L of anti-CD9, anti-CD63, or anti-CD81 antibodies in a coating buffer. After buffer removal and washing with 0.05% Tween 20 (Invitrogen, CA, USA) in PBS (PBS-T), a blocking solution (0.1% casein in PBS-T) was added and incubated for 30 min at 37℃. EVs (50 µL) were added to each well and incubated for 2 h at 37 ℃. Approximately 1 mg of anti-CD9, anti-CD63, or anti-CD63 antibody-conjugated beads was added to each well and incubated for 90 min at 37℃. Wells were washed with PBS-T and deionized water before drying in a thermostatic oven at 37℃ for 10 min, then quantified using the ExoCounter. To analyze the proteins extracted from EVs, samples were lysed and followed by centrifugation at 13,000 rpm for 15 min at 4℃ to clarify the lysate. The supernatant was mixed with protein sample buffer and denatured in a 95℃ dry bath for 10 min. Each sample, containing 20 µg of protein, was subjected to electrophoresis on 10% SDS-PAGE gel and transferred onto a PVDF membrane, followed by blocking with 5% bovine serum albumin for 1 h at room temperature and incubated with the primary antibodies, including anti-CD63 (ab239686, Abcam, Cambridge, MA, USA), anti-CD9 (ab239685, Abcam), anti-Alix (ab88743, Abcam), anti-HSP70 (ab47454, Abcam), anti-TSG101 (ab133586, Abcam), and anti-β-actin (A5441, Sigma-Aldrich) overnight at 4℃. Subsequently, after hybridization with a secondary antibody, bolts were visualized using an ECL kit and quantified using ImageJ software (US National Institutes of Health, MD, USA). To analyze the expression profiles of CD73 and CD146 on the EV surface, EVs were counted using a CytoFLEX nano (Beckman Coulter, Brea, CA, USA), with unmodified EVs as a background control. Each experimental condition was tested in sextuplicate. In vitro diabetes cell model establishment Human dermal fibroblasts (HDF, ScienCell) and human umbilical vein endothelial cells (HUVEC, ScienCell) were used to establish an in vitro diabetes model. HDFs were seeded in 48-well plates at a density of approximately 10⁴ cells/well and cultured in Fibroblast Medium (FM, ScienCell). HUVEC were cultured in endothelial cell medium (ECM, ScienCell) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (P/S) at 37℃ in a humidified atmosphere of 5% CO₂. To mimic the diabetic microenvironment in vitro, FM or ECM containing various glucose concentrations (8, 13.5, 19.5, and 25 mM) was prepared by adding glucose powder to the respective medium. These glucose-enriched media were used to culture HDF and HUVEC. After culturing for different time points, cell viability under varying glucose concentrations was assessed using the PrestoBlue^® Cell viability reagent. Cell cycle detection To verify the successful induction of diabetes-like conditions, cells were seeded in 6-well plates at a density of 10^5 cells/well. Cells were cultured in media containing 8 mM glucose (NG) or 25 mM glucose (HG). Cells were harvested on days 0, 1, and 5 post-seeding, washed once with PBS, centrifuged at 1,500 rpm to remove the supernatant, and gently resuspended in 70% ice-cold ethanol to dissociate the cell clumps. Overnight fixing at 4℃ and subsequent ethanol removal by centrifugation at 1,500 rpm was performed, and 100 µL of PI/RNase Staining Buffer was added for DNA staining. After incubation at 37℃ for 40 min in the dark, the samples were analyzed using a BD flow cytometer (BD Biosciences, NJ, USA). We counted the total cell number for each sample as 10,000 cells, and the percentage of cells in each phase of the cell cycle was calculated. All data were analyzed using FlowJo version 10.0 (Tree Star, Ashland, OR, USA). EVs uptake measurement HDF cultured under different conditions were seeded into 6-well plates or 8-well µ-slide (ibidi GmBH, Martinsried, Germany). After 24 h of culture, the experimental groups were treated with EVs labeled using ExoGlow™-Protein EV Labelling Kit (System Biosciences, Palo Alto, CA, USA). For ExoGlow labeling, the dye was mixed with an equal volume of the EV sample for 5 min in the dark, followed by centrifugation to pellet labeled EVs. The supernatant was carefully aspirated and the EV pellet was resuspended in PBS. After adding labeled EVs, the cells were cultured for the designated experimental period. For the 6-well plates, the cells were collected, washed, and analyzed by flow cytometry to calculate the uptake rate. For 8-well chamber slides, the cells were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100. The samples were then stained overnight at 4℃ with DAPI (Invitrogen) and Phalloidin-594 (Invitrogen), both diluted at a 1:1,000 ratio in PBS. Confocal microscopy was used to capture fluorescent images and observe the cellular uptake of EVs. Reactive oxygen species detection Cells cultured under the same conditions as those used in the cell cycle experiments were harvested using standard procedures and centrifuged for further use. To prepare the working solution for the total reactive oxygen species (ROS, Invitrogen) detection reagent was diluted with buffer at a ratio of 1:5000 following manufacturer’s instructions. Then, a 500 µL aliquot was thoroughly mixed with the cell pellet and incubated at 37℃ for 30 min in the dark. Samples were analyzed using flow cytometry, and all data were processed using FlowJo software. Migration assay The culture-insert 2 Well system (ibidi GmBH) was assessed cell migration. HDF were evenly seeded into the two chambers at a density of 5 × 10^4 cells/chamber. After 24 h incubation period, the insert was carefully removed to create a cell-free gap. The initial images of the gap area were captured using a microscope (Olympus BX53, Melville, NY, USA) under bright-field illumination. The cells were cultured with different EVs and then allowed to migrate for 24 h, with imaging of their movement. Migration analysis was performed using ImageJ software to quantify the cell-free area at each time point. The migration rate was calculated as the percentage reduction in the gap area over 24 h, using the formula: Migration rate (%) = [(Initial gap area - Final gap area) / Initial gap area] × 100. Angiogenesis assay To evaluate the angiogenic potential, HUVEC were co-cultured with EVs, then VEGF expression was quantified using an ELISA kit according to the manufacturer’s protocol. Supernatants collected on days 0 and 3 were centrifuged at 2,000 g for 10 min, diluted 1:2 with assay buffer and analyzed. Absorbance at 450 nm was measured, and VEGF concentrations were calculated using the standard curve. Then, the HUVEC was assessed through an in vitro tube formation assay using Matrigel (BD Biosciences, San Jose, CA, USA). A volume of 30 µL of Matrigel was added to each well of a 96-well plate and allowed to solidify at 37℃ for 1 h. HUVECs were then seeded onto the Matrigel-coated wells at a density of 10⁴ cells per well. EV or CSEV were added to the wells, and tube formation was imaged using the EVOS M7000 Imaging System (Thermo Fisher Scientific, CA, USA) at 6 h post-seeding. Next generation sequencing miRNA was extracted from EVs using the PureLink™ RNA Mini Kit and quantified with the Qubit™ RNA High Sensitivity assay. Subsequently, miRNA sequencing analysis was conducted, and the resulting data were aligned with databases. Differentially expressed gene (DEG) analysis was performed to identify miRNA showing significant differences (p < 0.05). Each group’s samples were analyzed three times. mRanda software was utilized to predict genomic targets for miRNA, with parameters set at S ≥ 150 and ΔG ≤ -30 kcal/mol, and 5’ primers were matched to the animal’s genome. Gene Ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) were employed to analyze the functions and sequences of different mRNA target genes. To quantify changes in cellular miRNA expression levels, the Log2 fold change was calculated to determine differences between each group. Collagen hydrogel fabrication and characterization To prepare a sterile 1% collagen hydrogel (Col), pre-weighed solid collagen sheets (Horien, Taichung, Taiwan) were placed in a laminar flow hood and exposed to UV light for 30 min. Subsequently, the sheets were cut into small pieces and transferred to beakers containing PBS. The mixture was placed in a 4℃ environment and continuously stirred using a magnetic stirrer for three days to ensure complete dissolution. Finally, the solution was filtered and stored in a desiccator until subsequent experiments. The structural morphology of the hydrogels was analyzed using a thermal field-emission scanning electron microscope. The solidified hydrogel samples were cut to appropriate sizes and placed on a specimen stage containing a mounting medium and frozen in liquid nitrogen before being transferred to the instrument, where they were fractured and subjected to a vacuum. Then, cross-sectional morphologies and pore structures of the hydrogels were examined. Degradation tests were performed by immersing hydrogel samples (10 × 10 mm) in PBS at 37℃. Samples were removed at determined days, dried under vacuum, and weighed to determine weight loss. The degradation rate was calculated based on the percentage difference between initial and final weights. The hydrogel was initially placed on a pre-cooled (4℃) rheometer sensing platform. The parallel-plate geometry was then lowered to a gap of 0.5 mm from the sensing platform. Using a modular compact rheometer (MCR302, Anton Paar, Graz, Austria), measurements were conducted at a frequency of 1 Hz and 0.5% strain, with temperature gradually increasing from 4℃ to 37℃. The storage modulus (G’) and loss modulus (G’’) were recorded as a function of time. The resulting data were plotted for analysis. Hemolysis assay The hemolysis rate was measured to evaluate the blood compatibility of collagen hydrogels loaded with different types of EVs. First, 1% collagen hydrogels were prepared and loaded with or without EVs. These hydrogels were immersed in normal saline for 30 min prior to testing. Fresh blood from New Zealand White rabbits was then collected and added to each hydrogel sample. After incubation at 37℃ for 1 h, the hydrogels were removed, and the mixtures were centrifuged at 2000 rpm for 5 min. The absorbance of the supernatant was then measured at 545 nm to quantify hemolysis. Additionally, the morphology of red blood cells (RBC) in the presence of each hydrogel formulation was observed using a microscope. Diabetic wound healing assessment in vivo The in vivo wound healing experimental procedures employed in the present study were approved by the Animal Experimental Ethics Committee of China Medical University (CMUIACUC-2022-486). The male New Zealand White rabbits (8 weeks old) were randomly divided into four groups. Alloxan monohydrate (≥ 98%, Sigma-Aldrich) was used to induce diabetes. A 5% Alloxan solution was prepared by dissolving 5.26 g of the powder in 100 mL of sterile saline (pH = 7) and filtered through a 0.22 μm membrane under an ice bath before immediate intravenous injection. To establish the diabetic rabbit model, the rabbits were weighed and their blood glucose levels were measured before the procedure. After anesthesia, a 5% alloxan solution was administered via the marginal ear vein at a dose of 200 mg/kg body weight [[74]32]. One week after alloxan administration, blood glucose levels were measured to confirm the establishment of the diabetic model. Rabbits with blood glucose levels exceeding 200 mg/dL were considered diabetic and selected for subsequent skin defect experiments. Blood glucose levels were subsequently measured every two days to monitor the progress of the experiment. Then, the diabetic rabbits were weighed, and their blood glucose levels were recorded. After administering anesthesia, the surgical site of the upper back area was completely hairless, preventing interference with wound observation during the experiment. Using a template, three 1.5 cm square areas were marked on the skin, with a minimum distance of 3.5 cm between each site. Rabbits were divided into four experimental groups: control (Ctl), pure collagen (Col), collagen with EV (EV@Col), and collagen with CSEV (CSEV@Col). Each experimental condition was tested in sextuplicate. The skin at each marked site was excised to the dermal layer, and the fascia was removed from the wound area. After excision, the initial images of the wounds were captured. Pre-prepared Col corresponding to each group were applied to the wound areas, followed by photo imaging. The Col were allowed to solidify at body temperature to prevent adhesion to the gauze. Once the Col was fully cross-linked, gauze was secured over the wounds using breathable medical tape. Subsequently, the rabbits were fitted with protective clothing to prevent them from biting the wound sites. Immunohistochemistry The rabbits were sacrificed, and regenerated skin was excised, collected, and excised along the margins and fixed in 4% paraformaldehyde on days 14 and 21. Subsequently, the tissues were embedded in paraffin and sectioned into 6 μm thick slices for hematoxylin and eosin (HE) staining, Masson’s trichrome (MT) staining, and Picrosirius red (PSR) staining. ImageJ software analyzed the quantification of IHC. In order to detect angiogenesis in the wound, the tissue sections were stained with CD31 and the protein levels of CD31 were analyzed using ELISA. Evaluation of tissue inflammation After treatment for 14 and 21 days, skin samples were assembled to evaluate tissue inflammation. To estimate the scope of IL-6 and TNF-𝛼, the protein levels of IL-6 and TNF-𝛼 were analyzed using ELISA. Additionally, tissue sections were analyzed by IHC staining with IL-6 and TNF-𝛼 antibodies, and IHC results were represented as the mean density (IOD/Area) of stained regions. Statistical analysis All experimental data in this study are presented as mean ± standard deviation (SD). Statistical analyses were conducted using GraphPad Prism (version 9.0), applying one-way analysis of variance (ANOVA). A p-value of less than 0.05 (*) was considered statistically significant. Results and discussion Characterization of calcium silicate ceramic and ion release effects on ADSC proliferation The XRD pattern of the synthesized CS exhibited characteristic peaks at 2θ values of 21.9°, 29.6°, 32.7°, and 47.6° (Fig. [75]2A), which corresponded to the crystalline phases of dicalcium silicate (C2S). The peaks at 31.2° and 34.2° are known to represent the tricalcium silicate (C3S) phases [[76]33]. Therefore, the sharpness of these peaks indicated a well-crystallized structure, confirming the successful synthesis of CS. The presence of these distinct peaks aligned with the standard reference patterns for CS, verifying the phase composition and crystallinity (Fig. [77]2B). In terms of morphology, SEM micrographs of the CS ceramics revealed a flaky and angular morphology. This structural appearance is consistent with our previous findings, confirming the reproducibility of the microstructure. In addition to their morphological characteristics, the concentrations of Ca and Si ions released from CS into the medium were quantified after one day of immersion (Fig. [78]2C). The results indicate that although there was a significant release of both ions, the Ca concentration decreased after the initial release. This may be due to the precipitation of Ca ions to form calcium phosphate or other mineral phases on the ceramic surface, which is common in CS-based materials [[79]34]. ADSC cultured in a medium conditioned with CS exhibited a marked increase in cell proliferation compared to the control group (Ctl) without CS exposure (Fig. [80]2D). This enhancement in cellular activity is likely due to the bioactive properties of the ions released from CS, suggesting that the ion-rich environment created by the ceramics promotes cell growth and vitality. Moreover, Si ions have been linked to enhanced collagen synthesis and fibroblast proliferation, further supporting their role in tissue regeneration. The observed increase in proliferation suggests that the CS extracts could be a promising candidate for wound healing and tissue regeneration applications, in which fibroblast activity is critical. Results harvested from the cytokine array indicated the upregulation of VEGF, FGF-2, HGF, Ang-1, and IL-10 in the CS-treated group compared to the Ctl (Fig. [81]2E). These factors are closely linked to angiogenesis, wound healing, and anti-inflammatory processes, further emphasizing the therapeutic potential of CS-conditioned environments for enhancing ADSC-mediated tissue regeneration. Fig. 2. [82]Fig. 2 [83]Open in a new tab XRD pattern of CS with peaks corresponding to dicalcium silicate (C2S) and tricalcium silicate (C3S) phases. (B) SEM image of CS showing its flaky and angular structure. (C) The concentration of Ca and Si ions released from CS after 1 day of immersion. (D) The proliferation of ADSC cultured in a CS-conditioned medium at different time points compared to Ctl groups. Protein expression related to angiogenesis in ADSC was analyzed using a protein array after 1 day of treatment with CS extracts, utilizing three separate samples for testing. Data are presented as mean ± SD, and * indicates statistical significance (p < 0.05). (E) Heatmap showing the expression levels of cytokines and growth factors, including VEGF, FGF-2, HGF, Ang-1, IL-10, IL-1RA, TGF-β, and collagen I (Col I), in ADSC cultured with or without CS-conditioned medium. Three replicates were used for each condition to assess protein expression Characterization of EVs from ADSC cultured with CS extract medium Figure [84]3 presents a comprehensive characterization of the EVs released by ADSC cultured in either standard medium (EV) or medium supplemented with CS extract (CSEV). TEM images (Fig. [85]3A) depicted the morphological characteristics of the EVs, with EV and CSEV exhibiting a typical spherical shape and well-defined lipid bilayer. No significant morphological differences were observed between the two groups. The uniformity in shape and size suggests that CS extract introduction into the medium does not alter the fundamental morphology of the secreted EVs. The particle size distribution, analyzed using NTA, further confirmed this observation (Fig. [86]3B), as both EV and CSEV exhibited similar size profiles. These findings align with previous studies demonstrating that modifications to culture conditions typically do not affect the general morphology of EVs, but rather influence their cargo and bioactive properties [[87]28]. Xu et al. found that EV size remains consistent despite alterations in the culture conditions of the EV source cells, highlighting that functional differences in EVs are more likely associated with changes in EV content rather than size [[88]35]. The zeta potential measurements in Fig. [89]3C provide insights into the EV surface charge. The CSEV group exhibited a more negative zeta potential than the EV group, indicating a more stable colloidal suspension. A higher negative charge can reduce vesicle aggregation, thereby enhancing colloidal stability [[90]36]. This increased stability is particularly beneficial for therapeutic applications because it improves EV dispersion in solution and enhances cellular uptake efficiency. The more negative zeta potential observed in the CSEV group may be partially attributed to the presence of integrins, which carry a net negative charge [[91]37]. Our previous work has demonstrated that CS treatment stimulates an upregulation of integrins on the cell surface [[92]38, [93]39]. These surface integrins are likely incorporated into EVs during secretion, contributing to the increased zeta potential. This enrichment of surface integrins might reflect the bioactive effects of CS on cellular behavior and EV surface composition, potentially enhancing their therapeutic applications. Figure [94]3D shows that CSEV displayed a significantly higher total protein content than EV. This increase in protein content suggests that cells cultured in the CS extract medium may secrete more biologically active or protein-rich EVs, potentially enhancing their functional properties for downstream applications. Fig. 3. [95]Fig. 3 [96]Open in a new tab Comparison of EVs secreted by ADSC under normal culture conditions (EV) and CS extraction solution (CSEV). (A) TEM images of EV and CSEV with spherical morphology and well-defined lipid bilayers. The scale bar is 100 nm. (B) Particle size distribution of EV and CSEV analyzed by NTA. (C) Zeta potential of EV and CSEV showing surface charge differences. (D) Total protein content of EV and CSEV per 10^9 particles. (E-G) Expression ratios of CD9, CD63, and CD81 surface markers in EV and CSEV. (H) Western blot analysis of EV-specific markers CD63, CD9, Alix, HSP70, and TSG101, with β-tubulin as a loading control. Data are presented as mean ± SD, and * indicates statistical significance (p < 0.05). (I) The proportions of CD73 and CD146 on different EV surfaces The proportion of EVs with expression of the surface markers CD9, CD63, and CD81, which are key members of the tetraspanin family and commonly used to characterize EVs, are shown in Fig. [97]3E-G. The data indicate that CSEV exhibited significantly higher expression levels of all three markers to EV. Specifically, CSEV showed a marked increase in the ratios of these tetraspanins, including CD9, CD63, and CD81. These findings suggest that the CS extract may improve both the yield and quality of EV production, potentially due to the bioactive effects of the Ca and Si ions released from the CS extract. Ion-enriched environments are known to stimulate cellular secretion pathways, leading to a higher production of EVs with improved bioactivity [[98]40]. Additionally, the ion-enriched environment created by CS stimulates cellular secretion pathways, leading to increased EV production. This process may also modulate the bioactive cargo of EVs, such as miRNAs and proteins that regulate oxidative stress pathways, further amplifying their therapeutic potential. Additionally, western blot analysis revealed enhanced levels of EV-specific markers, including CD63, CD9, Alix, HSP70, and TSG101 in CSEV (Fig. [99]3H), with β-tubulin serving as a loading control. Elevated tetraspanin levels may enhance EV stability and facilitate their interaction with recipient cells, aligning with strategies to optimize EV surface properties for therapeutic applications [[100]41]. This suggests that EVs derived from cells cultured with CS extract exhibit higher expression of markers and may possess enhanced functional properties due to their enriched protein cargo. In addition, we also analyzed the unique surface markers of EVs secreted by ADSC, such as CD73 and CD146. The presence of CD73/CD146 (Fig. [101]3I) double-positive EVs, with notably higher concentrations in CSEV group than in the EV group. CD73 + EVs demonstrated HUVEC to promote angiogenesis by activating the ADO/A2AR signaling axis [[102]42]. The effect of CD146 + EVs on restoring and stabilizing blood vessel regeneration after injury [[103]43]. The in vitro diabetic microenvironment model To mimic the “diabetic microenvironment” in vitro, a high-glucose medium (25 mM) was prepared by adding D-anhydrous glucose powder into a Fibroblast Medium (FM, 8 mM glucose) and was used for cell culture. As shown in Fig. [104]4A, the viability of HDF was assessed at various glucose concentrations (8, 13.5, 19, and 25 mM) for 1, 3, and 5 d. The data revealed a clear trend in which higher glucose levels led to reduced cell viability, with the most significant decrease observed at 25 mM. The impact of glucose was also time-dependent, with the most substantial decline in viability occurring after 5 days of culture. The in vitro diabetic model using high glucose concentration has been widely validated in previous studies, particularly in investigating diabetes-related cellular responses [[105]44]. These findings suggest that elevated glucose levels, which simulate hyperglycemic conditions, exert a cytotoxic effect on HDF, reducing survival rates over time. This is consistent with chronic hyperglycemia inducing oxidative stress and apoptosis in various cell types, contributing to decreased cell viability [[106]45].The analysis of cell cycle distribution, presented in Fig. [107]4B, compares HDF cultured under NG (8 mM) and HG (25 mM) conditions on days 0, 1, and 5. Under NG conditions, the cell cycle phases remained consistent with stable proportions of cells in the G1, S, and G2 phases across time points. In contrast, HG conditions induced pronounced arrest in the G1 phase on day 1, with a sharp decline in the S phase population. After 5 days of culture under HG conditions, most cells were arrested in the G1 phase, with minimal progression to the G2 phase. Glucose-induced cell cycle arrest is a well-documented phenomenon, often associated with increased ROS production and the activation of stress pathways that halt cell proliferation [[108]46]. High glucose conditions, such as 25 mM glucose treatment, have been shown to induce G1 phase arrest in fibroblasts through activation of the p38 MAPK pathway and the ATM/p53/p21 signaling cascade [[109]47]. HG environments can disrupt cell cycle progression by increasing p21 and p27 expression, leading to G1 phase arrest [[110]48]. The cell cycle arrest observed under high glucose conditions may contribute to reduced cell viability, as it impairs the ability of cells to proliferate and progress through the normal cell cycle. Fig. 4. [111]Fig. 4 [112]Open in a new tab (A) Cell viability of HDF cultured under different glucose concentrations (8 mM, 13.5 mM, 19 mM, and 25 mM) for 1, 3, and 5 days. (B) Cell cycle distribution of HDF under normal glucose (NG, 8 mM) and high glucose (HG, 25 mM) conditions at Day 0, Day 1, and Day 5, assessed by flow cytometry. Data are presented as mean ± SD, and * indicates statistical significance (p < 0.05). (C) Fluorescence microscopy images of F-actin (red), EVs (green), and nuclei (blue) in HDF after 24 and 48 h of EV or CSEV uptake. Scale bar is 50 μm. Flow cytometry analysis of EV and CSEV uptake in HDF, quantified as the percentage of cells positive for FITC fluorescence. (D) Flow cytometry analysis of ROS levels in HDFs treated with EV, CSEV, or control (Ctl). (E) Wound healing assay showing HDF migration after treatment with control, EV, or CSEV at 0 and 24 h. Scale bar is 150 μm. Quantification of HDF migration rate after 24 h of treatment. Data are presented as mean ± SD, and * indicates statistical significance compared to Ctl (p < 0.05), and # indicates statistical significance between EV and CSEV (p < 0.05) Enhanced cellular uptake and behavior of HDF following CSEV treatment Given the differences observed in protein content and surface marker expression between CSEV and standard EV, it was essential to evaluate whether these differences translated into enhanced cellular uptake. Fluorescence images (Fig. [113]4C) revealed that HDF exposed to CSEV showed a significantly stronger green fluorescence signal at 24 h and 48 h than those treated with standard EV. This heightened fluorescence indicates a higher level of CSEV internalization. Notably, the EVs were observed to localize predominantly around the perinuclear region, suggesting successful internalization rather than surface binding. The co-localization of EVs (green) with F-actin (red) and nuclei (blue) in the merged images confirmed the successful uptake and integration of these vesicles into the HDF cellular structure. The increased uptake efficiency of CSEV may result from their modified surface characteristics or enriched bioactive content, which facilitate their internalization by cells. Consistent with the cellular uptake, flow cytometry analysis showed that 86.3% of HDF internalized CSEV, compared to 66.0% for standard EV. The higher uptake efficiency of CSEV may be attributed to their modified surface characteristics and enriched bioactive content, making them more readily internalized by cells. In addition, the presence of surface markers such as CD9, CD63, and CD81, along with integrins, can modulate the interaction between EVs and recipient cells to facilitate uptake [[114]49]. Moreover, the bioactive ions, particularly calcium and silicon, from the CS extract may participate in cell membrane modulation, thereby promoting vesicle internalization [[115]29]. Flow cytometry analysis results in Fig. [116]4D show the production of ROS in HDF following the uptake of CSEV and EV. Cells treated with CSEV exhibited lower ROS production (16.6%) than those treated with EV (37.4%), while the control group showed the highest ROS levels (92.1%). Ca and Si ions play a role in reducing oxidative stress by modulating cellular antioxidant pathways and enhancing the activity of enzymes such as superoxide dismutase (SOD) and catalase [[117]50]. Therefore, the diminished ROS levels may be attributed to the antioxidant properties of the CS extract components. Moreover, CS-contained biomaterials have been shown to decrease ROS production in the wound environment, accelerating healing and reducing inflammation [[118]51]. Accordingly, the above results suggest that CSEV are more effectively internalized by HDF and provide a protective advantage by reducing oxidative stress. This has significant implications for their therapeutic applications, particularly in conditions where oxidative damage plays a critical role in disease progression. To assess the impact of EV on cell migration, a wound-healing assay was performed on HDF cultured in media supplemented with these vesicles. The wound-healing assay images in Fig. [119]4E show the migration of HDF at 0 and 24 h in media containing no vesicles (Ctl), standard EV, or CSEV. At 0 h, a clear wound gap was visible in all groups. After 24 h, gap closure was the most pronounced in the CSEV group, followed by the EV group, while the control group showed the least closure. Quantitative analysis of the migration rate, depicted in Fig. [120]5E, supports these observations. HDF treated with CSEV demonstrated a significantly higher migration rate than those treated with standard EV or the control group. Specifically, the migration rate in the CSEV group exceeded 90%, which was markedly higher than the 70% rate observed in the EV group. The control group exhibited the lowest migration rate, highlighting the superior effectiveness of CSEV in promoting cell migration. EVs derived from MSC can enhance fibroblast migration by delivering bioactive cargo that promotes cell motility and extracellular matrix remodeling [[121]52, [122]53]. Moreover, CSEV likely contain elevated levels of key regulatory proteins, such as TGF-β and VEGF, which are associated with enhanced cell migration and angiogenesis [[123]54]. Therefore, we propose that the enhanced migration in response to CSEV may be due to their enriched content of bioactive molecules, such as growth factors, cytokines, and miRNAs, all of which are known to stimulate cellular processes critical for wound healing and tissue repair [[124]55]. Fig. 5. [125]Fig. 5 [126]Open in a new tab (A) Cell viability of HUVEC cultured under different glucose concentrations (8 mM, 13.5 mM, 19 mM, and 25 mM) for 1, 3, and 5 days. (B) Cell cycle distribution of HUVEC under normal glucose (8 mM) and high glucose (25 mM) conditions at Day 0, Day 1, and Day 5, analyzed by flow cytometry. (C) Fluorescence microscopy images of F-actin (red), EVs (green), and nuclei (blue) in HUVEC after incubation with EV or CSEV for 24 h. Scale bar: 100 μm. (D) VEGF concentration in HUVEC treated with EVs over 3 days. (E) Tube formation assay of HUVEC treated with EVs visualized after 6 h. Scale bar: 500 μm. Quantification of branch points (F) and total tube length (G) in the tube formation assay. Data are presented as mean ± SD, and * indicates statistical significance compared to Ctl (p < 0.05), # indicates statistical significance between EV and CSEV (p < 0.05) Analysis of HUVEC behavior under diabetic microenvironment following EVs uptake HUVEC viability decreased with increasing glucose concentrations (8, 13.5, 19, and 25 mM) over a 5-day culture period (Fig. [127]5A). At day 1, the cells maintained relatively high viability across all conditions. However, prolonged exposure to high glucose levels, particularly 25 mM, led to a significant decline in cell viability by day 5. This trend suggests that hyperglycemic conditions impair cellular function and viability over time, consistent with the cytotoxic effects of chronic hyperglycemia reported in vascular cells. The cell cycle profiles of HUVEC treated with 8 mM (NG) and 25 mM (HG) glucose concentrations are depicted in Fig. [128]5B. Under NG conditions, the distribution of cells across G1, S, and G2 phases remained consistent over 5 days. In contrast, HG treatment resulted in G1 phase arrest, with a marked reduction in the S phase population as early as day 1. By day 5, the majority of cells were arrested in G1, indicating disrupted cell cycle progression, likely due to glucose-induced oxidative stress and activation of cell cycle regulatory pathways. This finding is similar to our earlier observations with HDFs, where chronic hyperglycemia was shown to induce oxidative stress, impair cellular function, and reduce viability, highlighting the detrimental effects of diabetes on vascular endothelial cells. This glucose-induced cell cycle arrest highlights the adverse impacts of hyperglycemia on vascular cells, as it halts proliferation and repair processes critical for maintaining vascular integrity, a key issue in diabetes-related vascular diseases [[129]56]. To assess the impact of EVs on HUVEC, the results of HUVEC exposed to EV and CSEV under HG conditions (Fig. [130]5C). Cells treated with CSEV exhibited stronger green fluorescence signals compared to those treated with standard EV, indicating enhanced uptake. The co-localization of EVs (green) with F-actin (red) and nuclei (blue) confirms successful internalization of the vesicles. As shown in Fig. [131]5D, VEGF secretion was significantly higher in HUVEC treated with CSEV compared to EV and control groups. On day 3, the CSEV group exhibited the highest VEGF levels. VEGF plays a central role in angiogenesis by promoting endothelial cell proliferation and migration, processes that are impaired under diabetic conditions due to oxidative stress and inflammation [[132]56]. The ability of CSEV to upregulate VEGF highlights their potential to restore angiogenic balance and counteract diabetes-induced vascular dysfunction. The tube formation assay under HG conditions further demonstrated the pro-angiogenic capacity of CSEV. As shown in Fig. [133]5E, HUVEC cultured with CSEV formed significantly more extensive and organized tubular networks compared to those cultured with EV or in the control medium. Quantitative analysis (Fig. [134]5F and G) revealed the highest number of branch points and total loops in the CSEV group. These findings underscore the capacity of CSEV to promote endothelial cell functionality critical for vascular network formation. NGS and miRNA profiling of EVs To identify the key miRNAs in CSEV that play a beneficial role, a comparative analysis of their respective miRNA expression profiles was performed. The heatmap (Fig. [135]6A) shows global transcriptomic differences between EV and CSEV. The expression profiles indicated that a large number of genes were upregulated in CSEV, suggesting activation of specific signaling pathways. Compared with the standard EV, CSEV has a total of 273 differentially expressed genes, including 114 upregulated mRNAs and 159 downregulated mRNAs (Fig. [136]6B). Figure [137]6C shows the results of KEGG pathway enrichment analysis of DEGs between CSEV and EV. The results show that WNT, MAPK, and focal adhesion pathways were prominent, suggesting the potential influence of CSEV on cell proliferation, differentiation, and migration. Figure [138]6D presents a GO term enrichment analysis, categorizing DEGs by biological processes, cellular components, and molecular functions. The GO terms identified include nucleus, cytosol, and cytoplasm, highlighting intracellular localization. Terms related to metal ions, ATP, and DNA binding were significant, indicating the influence of CSEV on various intracellular processes. These GO terms align with KEGG pathways, further supporting the unique capacity of CSEV to modulate key cellular functions, particularly in metabolism and signaling. This finding reinforces previous studies indicating that bioactive materials can regulate EV cargo, thereby influencing diverse cellular activities [[139]54]. Fig. 6. [140]Fig. 6 [141]Open in a new tab Transcriptomic and miRNA profiling of EV and CSEV. (A) Heatmap of differentially expressed genes (DEGs) between EV and CSEV. (B) Volcano plot showing upregulated (red) and downregulated (blue) genes in CSEV compared to EV. (C) KEGG pathway enrichment analysis of DEGs, highlighting key pathways involved in CSEV effects. (D) GO term enrichment analysis categorizes DEGs based on biological processes, cellular components, and molecular functions. (E) Upregulated miRNAs in CSEV associating with angiogenesis, anti-inflammation, and wound healing. Data are presented as log2 fold changes (log FC) Furthermore, analysis of the miRNA profile revealed significantly alter levels of specific miRNAs in CSEV, which are involved in three crucial biological processes: angiogenesis (miR-31, miR-130, miR-210, miR-150, miR-4488, miR-181, and miR-185-5p), anti-inflammatory (miR-183, miR-207, and miR-101), and wound healing (miR-452-5p, miR-222, and miR-211) (Fig. [142]6E). Among them, miR-31 and miR-130 were most notably up-regulated in CSEV. Previous studies have demonstrated that miR-31 modulates angiogenesis by regulating mitofusin-2 and HIF-1 pathways [[143]57]. Yan et al. revealed that milk-derived exosomal miR-31-5p accelerates diabetic wound healing by promoting vascular network formation and increasing blood vessel density [[144]58]. In terms of anti-inflammatory effects, CSEV exhibited significant upregulation of miR-183, miR-207, and miR-101 compared to EV controls. These miRNAs may reduce inflammation through inhibition of the TGF-β/Smad/TLR3 pathway [[145]59]. and have been implicated in promoting M2 macrophage polarization and anti-inflammatory responses [[146]60]. Taken together, the unique miRNA cargo of CSEV may contain key regulatory factors involved in angiogenesis, inflammation, and wound healing. Physicochemical properties of collagen enhance the therapeutic potential of CSEV-loaded collagen for wound healing and inflammation modulation The FTIR spectrum of the collagen matrix exhibits a broad absorption band around 3300 cm⁻¹, attributed to the O-H and N-H stretching vibrations. Peaks near 2920 cm⁻¹ correspond to C-H stretching vibrations, while the characteristic amide I and amide II bands appear around 1650 cm⁻¹ and 1550 cm⁻¹, respectively, indicating the presence of collagen’s protein structure. Additionally, the absorption band near 1240 cm⁻¹ represents C-N stretching, and the band near 1020 cm⁻¹ corresponds to C-O-C vibrations, further confirming the presence of functional groups typical of collagen (Fig. [147]7A) [[148]61]. These prominent peaks confirm the presence of the collagen’s functional groups, indicating that the collagen matrix maintains its characteristic biochemical properties. Figure [149]7B demonstrates the thermal responsiveness of the collagen matrix, transitioning from a solution state to gelation at 37℃. The rheological properties reveal an increase in the storage modulus (G’) over time, surpassing the loss modulus (G”), indicating successful gelation and the formation of a stable network structure. This mechanical behavior supports the matrix’s suitability for loading and sustained release of EVs, ensuring the material’s ability to maintain structural integrity during use. In addition, the evolution of the storage modulus (G’) and loss modulus (G’’) of pre-cured (4℃) and cured (37℃) collagen hydrogels as a function of shear strain is shown in Fig. [150]7C. The cured hydrogel exhibits significantly higher mechanical strength, elasticity, and flexibility, as indicated by the increase in G’ from approximately 6 Pa to 100–150 Pa, the loss factor from 0.95 to 1.47, and the expansion of the linear viscoelastic region from 10–40%. Figure [151]7D evaluates the hemocompatibility of the collagen matrix through a hemolysis assay. The absorbance at 545 nm for PBS, Col, EV, and CSEV groups was minimal, indicating negligible hemolysis, whereas the H[2]O group showed significantly higher absorbance, serving as a positive control. Additionally, the microscopic images of RBC morphology under different treatment conditions. No morphological alterations were observed in each group, further confirming the biocompatibility of the collagen matrix and its formulations. The absence of RBC damage underscores the matrix’s safety profile, making it suitable for in vivo applications. These results confirm that the collagen matrix and its EV formulations are hemocompatible and safe for biomedical use, rheological, and biocompatibility features of the collagen matrix, demonstrating its potential for use as a delivery platform for EVs in wound healing and inflammation modulation. Fig. 7. [152]Fig. 7 [153]Open in a new tab Physicochemical properties of collagen and its enhanced therapeutic potential of CSEV for wound healing. (A) FTIR spectrum of the collagen matrix showing characteristic peaks. The oscillatory rheological experiment of (B) temperature sweep and (C) strain sweep of the collagen matrix. (D) Hemolysis assay assessing the hemolytic activity of PBS, Col, EV-loaded Col (EV), and CSEV-loaded Col (CSEV), with water (H[2]O) as the positive control. Scale bar: 75 μm. (E) Schematic of the experimental setup for EV or CSEV encapsulated in collagen, cultured with diabetic HDF. (F) Residual weight of collagen matrix over 7 days of immersion, indicating its structural stability. (G) SEM images of the collagen matrix on Day 0 and Day 7. Scale bar is 5 μm (H) Cumulative release profile of EV and CSEV from the collagen matrix over 14 days. (I) Fluorescence microscopy images of HDF cultured on EV- or CSEV-loaded collagen at different time points (Day 0, 1, 3, and 7). Scale bar is 25 μm. (J) ROS generation in HDF cultured on collagen matrices loaded with EV or CSEV. (K) HDF proliferation on collagen matrices loaded with EV or CSEV over 7 days. Data are presented as mean ± SD, and * indicates statistical significance compared to control (Ctl) (p < 0.05), # indicates statistical significance between EV and CSEV (p < 0.05) Figure [154]7E illustrates the experimental setup where EVs are embedded in the collagen matrix placed in the upper insert of the coculture system, while diabetic HDF are seeded in the bottom chamber. During culture for 7 days, the weight loss of Col is about 20.7 ± 5.6% (Fig. [155]7F). SEM images further confirm the structural integrity of the collagen matrix in Fig. [156]7G. The porous nature of the matrix is visible, with interconnected pores providing ample space for EV encapsulation and release. As shown in Fig. [157]7H, the release profile of EVs from the collagen matrix demonstrates a steady release over 14 days, confirming that the collagen matrix effectively prolongs the release of both vesicle types. The numerical results extracted from Fig. [158]7 are presented in Table [159]1. The controlled release over time suggests that collagen is an ideal carrier for EVs, ensuring their prolonged bioavailability for therapeutic applications. In addition, the release of growth factors also achieves the effect of continuously stimulating cell growth, angiogenesis and wound healing [[160]62]. Furthermore, fluorescence imaging indicates that while the EVs content within the collagen matrix decreases with soaking time, a substantial amount of EVs remains present on day 7 (Fig. [161]7I). These findings align with previous studies showing that collagen matrices effectively act as carriers for the controlled release of bioactive molecules, including EVs, due to their structural integrity and biocompatibility [[162]63]. Table 1. The behaviors of EV-loaded col after immersion at various times. Four Immersion time Degradation (%) EV release (E + 10) Pore size (µm) 0 0.0 ± 0.0 0.0 ± 0.0 2.1 ± 1.4 3 10.0 ± 3.9 3.3 ± 0.3 - 7 20.7 ± 5.6 4.4 ± 0.3 4.2 ± 1.2 [163]Open in a new tab Three replicates were considered for the immersion test. The results were reported as the mean ± standard deviation Figure [164]7J presents the ROS generation analysis of HDF cultured on collagen matrices containing EVs. The CSEV-loaded collagen group exhibited significantly lower ROS levels compared to both the EV-loaded and control groups (collagen without EVs). This reduction in oxidative stress aligns with earlier observations of direct CSEV treatment reducing ROS levels in cells. The antioxidant properties of miRNAs and other bioactive components within CSEV likely contribute to this protective effect, creating a more favorable cellular environment that mitigates damage and promotes healthy cellular functions. These findings are consistent with previous studies demonstrating that EVs can alleviate oxidative stress-induced cellular injury, further highlighting their therapeutic potential [[165]64]. Finally, as shown in Fig. [166]7K, HDF proliferation was significantly enhanced in the presence of CSEV-loaded collagen compared to the other groups. Over a 7-day period, the CSEV group exhibited the highest proliferation rates, particularly on day 7, indicating their superior regenerative capacity. While standard EV also promoted cell proliferation, their effects were less pronounced than those of CSEV. These results underscore the enhanced therapeutic potential of CSEV-loaded collagen, driven by their sustained release and enriched bioactive content. Enhanced therapeutic potential of CSEV-loaded collagen for wound healing and inflammation modulation Cytokine arrays and related analyses were conducted to investigate the influence of EV- and CSEV-loaded collagen, respectively, on HDF secretory activity. The heat map in Fig. [167]8A summarizes the differential secretion of cytokines and growth factors from HDF cultured on collagen loaded with EVs. Notably, a superior upregulation of proangiogenic factors such as VEGF-A, EGF, and FGF-2 in the CSEV group indicates that CSEV has a more pronounced effect on promoting angiogenesis than standard EV. Additionally, the secretion of anti-inflammatory marker IL-10, along with growth factors PDGF-BB and IGF-I, was elevated in the CSEV group, reinforcing that CSEV contributes to a favorable cellular microenvironment by modulating growth and inflammatory responses in HDF. Additionally, cytokine modulation, particularly the upregulation of IL-10, is crucial in promoting the regenerative response, diminishing inflammation, and accelerating tissue repair [[168]65]. Fig. 8. [169]Fig. 8 [170]Open in a new tab Analysis of cytokine and extracellular matrix production in cells treated with EVs. (A) Heatmap representing the secretion levels of various cytokines and growth factors, including Col I, VEGF-A, EGF, and IL-10, among others, in the CSEV, EV, and control groups. The secretion of IL-1β (B), TNF-α (C), IL-10 (D), and Collagen I (Col I) (E) in the CSEV, EV, and control groups. Data are presented as mean ± SD, and * indicates statistical significance compared to Ctl (p < 0.05), and # indicates statistical significance between EV and CSEV (p < 0.05) Both of IL-1β and TNF-α are major pro-inflammatory cytokines. As shown in Fig. [171]8B and C, the CSEV group exhibited a substantial reduction in the levels of IL-1β and TNF-α compared to the standard EV and control groups, suggesting a potent anti-inflammatory effect of CSEV. This reduction in proinflammatory cytokines is consistent with the well-established role of EVs in mitigating inflammatory responses, potentially through the delivery of specific miRNAs that inhibit NF-κB and other inflammatory signaling pathways. In contrast, the secretion of IL-10, a crucial anti-inflammatory cytokine, was significantly higher in the CSEV group than in the standard EV and control groups. This increase in IL-10 production highlights the superior capacity of CSEV to promote anti-inflammatory signaling (Fig. [172]8D). Furthermore, Fig. [173]8E demonstrates that the production of Collagen I (Col I), a key extracellular matrix component crucial for wound healing and tissue regeneration, was significantly higher in the CSEV group than in the EV- and control groups. This suggests that CSEV are more effective at promoting extracellular matrix production, further supporting their regenerative potential. PDGF-BB and IGF-I are growth factors stimulating fibroblast activity and collagen production. The increase in Col I secretion suggests a correlation with the upregulation of PDGF-BB and IGF-I. The enriched KEGG pathways highlighted crucial signaling cascades, particularly PI3K-Akt and MAPK pathways, which aligned with GO terms associated with cellular proliferation and tissue regeneration, which demonstrated elevated secretion of anti-inflammatory cytokines, collectively supporting CSEV’s beneficial role in modulating inflammatory responses during wound healing [[174]66]. Histological evaluation of wound healing in diabetic rabbits treated with CSEV-loaded collagen The establishment of diabetic animal models has been extensively studied and validated for investigating diabetes-related wound healing complications. Among various methods, alloxan-induced diabetes in rabbits has emerged as a well-established and reliable approach, with numerous studies demonstrating its effectiveness in creating stable hyperglycemic conditions [[175]67]. The optimal protocol for diabetes induction, including the dosage and administration route, has been thoroughly investigated, with 200 mg/kg body weight via marginal ear vein administration being widely accepted as the standard procedure [[176]32]. This specific dosage has been shown to effectively destroy pancreatic β-cells while maintaining acceptable survival rates in experimental animals [[177]68]. The healing process of skin wounds in diabetic rabbit models was analyzed using HE, MT, and PSR staining on post-surgery days 14 and 21 (Fig. [178]9A). Higher magnification images of the HE-stained sections at days 14 and 21 provide more detailed insights into re-epithelialization and granulation tissue formation (Fig. [179]9B). The HE stains depict the epidermal layer and the formation of granulation tissue in different treatment groups. On days 14 and 21, the control and collagen-only groups (Col) showed incomplete epithelial coverage with sparse granulation tissue, indicating slow wound healing. In contrast, the EV@Col and CSEV@Col groups displayed more advanced healing, with the CSEV@Col group exhibiting the most extensive re-epithelialization on day 14 and a significantly thickened epithelial layer on day 21. Additionally, sweat glands, capillaries, and new hair follicles were clearly observed in the CSEV@Col group, as indicated by red arrows in Fig. [180]9B. This highlights its superior regenerative effect. Fig. 9. [181]Fig. 9 [182]Open in a new tab Histological evaluation of skin wound healing in diabetic rabbits treated with collagen loaded with EV or CSEV at 14- and 21-days post-surgery. HE, MT, and PSR staining were performed to assess re-epithelialization, collagen deposition, and collagen fiber organization in the control (Ctl), collagen-only (Col), EV@Col, and CSEV@Col groups. Scale bar is 500 μm. (B) Higher magnification images of HE-stained sections at 14- and 21-days post-surgery. Red arrows indicate sweat glands, capillaries, and hair follicles observed in the CSEV@Col group. Scale bar is 100 μm. Quantitative analysis of (C) the wound area length and (D) collagen fraction (%) at days 14 and 21 (n = 6). *indicates a significant difference (p < 0.05) The statistical data for the wound length also confirmed that the CSEV@Col group had the shortest wound length, followed by the EV@Col, collagen hydrogel, and control groups (Fig. [183]9C). These results indicate that CSEV@Col demonstrated good biocompatibility and accelerated the healing process in a normal wound healing model. MT assesses collagen formation and fibrosis. On day 14, the wounds in the control group showed minimal collagen fibers and were primarily covered with eschar. In contrast, the wounds in the CSEV@Col treatment group displayed abundant and relatively well-organized collagen fibers. On day 21, newly formed collagen fibers were observed in all groups, with the CSEV@Col group showing the highest collagen volume fraction compared to the others. These findings highlight the impact of CSEV@Col on fibrosis and tissue repair, which are critical for wound stability and strength [[184]69]. Quantitative analysis of the collagen fraction (%) is presented in Fig. [185]9D. The results demonstrate that the CSEV@Col group had the highest collagen content at both day 14 and day 21, significantly exceeding that of the EV@Col, Col, and Ctl groups. As an EV sustained-release system, CSEV@Col exhibited the greatest potency in enhancing collagen deposition and accelerating skin regeneration. Picrosirius Red staining under polarized light provides insights into collagen deposition and maturation of collagen fibers. The images revealed that on day 14, the CSEV@Col and EV@Col groups induced a greater deposition of collagen than other groups. The intensity and organization of type I and type III collagen fibers were more pronounced in the CSEV@Col group, with denser, well-aligned fibers observed by day 21, indicating enhanced tissue remodeling and structural integrity. Angiogenesis is an essential part of the entire process of wound healing. CD31, as a marker of endothelial cells, was detected by IHC and ELISA to assess the newly formed vessels in an in vivo diabetic wound model. The CSEV@Col group exhibited a significantly higher number of blood vessels compared to the other groups (Fig. [186]10A and C). This enhanced vascularization is likely linked to the bioactive miRNAs enriched in CSEV, including miR-31, miR-150, and miR-210, which are known to promote cell viability, migration, and angiogenesis [[187]70, [188]71]. In addition, persistent inflammation was also an important reason for delayed wound healing, which resulted in the inability to transition between the inflammatory and proliferative stages. Therefore, reducing the level of inflammation is beneficial for the wound to move to the proliferative stage. The IHC satin demonstrated that CSEV@Col could reduce the expression of TNF-𝛼 (Fig. [189]10D and E) and IL-6 (Fig. [190]10G and H), thus the inflammatory response was alleviated. The results of ELISA also affirmed the same observations (Fig. [191]10F and I). CSEV contains many miRNAs (such as miR-183 and miR-101), which have been proven to have great potential in regulating the differentiation and development of macrophages, thereby inhibiting the activation of inflammatory signaling pathways [[192]59]. This supports the potential therapeutic advantage of CSEV-loaded collagen dressings in improving wound healing outcomes, particularly in diabetic conditions where tissue regeneration is compromised. Fig. 10. [193]Fig. 10 [194]Open in a new tab In vivo wound healing evaluation of the diabetic wound model. (A) IHC images, (B) quantification, and (C) ELISA results of CD31 at day 14 and day 21. (D) IHC images, (E) quantification, and (F) ELISA results of TNF-𝛼 at day 14 and day 21. (G) IHC images, (H) quantification, and (I) ELISA results of IL-6 at day 14 and day 21. *indicates a significant difference (p < 0.05). Scale bar is 150 μm Conclusion The findings of this study demonstrate that CSEV, incorporated into collagen hydrogel dressings, significantly enhances diabetic wound healing. CSEV exhibit superior bioactivity as evidenced by their increased protein content, enhanced cellular uptake, and modulation of inflammatory and regenerative pathways. The sustained release of CSEV from the collagen matrix supports long-term therapeutic efficacy, promotes angiogenesis, reduces oxidative stress, and accelerates re-epithelialization and collagen deposition in vivo. These results highlight the potential of CSEV-loaded collagen dressings as a novel therapeutic strategy for improving outcomes in diabetic wound care, particularly in cases where traditional treatments are insufficient owing to impaired tissue regeneration. Acknowledgements