Hydrogel microneedles loaded with engineered extracellular vesicles accelerate diabetic wound healing by promoting angiogenesis and modulating macrophage polarization Authors: Shaoyihan Fang, Jingjie Min, Zhenzhen Chen, Shuangyan Li, Yulu Zhou, Jingwen An, Zhongyi Chen, Dewu Liu Published: 2026 ABSTRACT AbstractOxidative stress, impaired angiogenesis, and persistent inflammation contribute to delayed diabetic wound healing. In this study, ALDH2-loaded liposomes were associated with extracellular vesicles derived from human amniotic mesenchymal stem cells to generate an ALDH2@Lipo/EV hybrid formulation, which possess pro-angiogenic, antioxidant, and anti-inflammatory properties that modulate mitochondrial homeostasis in cells within diabetic wounds. To enhance the penetration efficiency of EVs in the wound, ALDH2@Lipo/EV were loaded into methacrylated gelatin (GelMA) hydrogel to fabricate the ALDH2@Lipo/EV-MN patch system. The microneedles technology enabled precise and efficient delivery of ALDH2@Lipo/EV, ensuring an optimal depth of delivery for maximum therapeutic efficacy. Mechanistically, ALDH2@Lipo/EV promoted angiogenesis in vitro and alleviated oxidative damage and inflammation in macrophages by inhibiting the activation of dynamin-related protein 1 (Drp1) and activating the PINK1/Parkin signaling pathway. Furthermore, the ALDH2@Lipo/EV-MN patch was shown to enhance diabetic wound healing in a mouse model. This study explored the functions and mechanisms of the ALDH2@Lipo/EV-MNs system, providing a new strategy for the treatment of refractory diabetic wounds.Keywords: Diabetic wound healing, Engineered extracellular vesicles, Microneedles, Mesenchymal stem cell-derived extracellular vesicles Introduction According to statistics, the global number of individuals with diabetes has surpassed 800 million and is exhibiting a staggering trend of continuous growth [1]. Without effective intervention, this number is projected to reach 1.31 billion by 2050 [2]. Among these patients, approximately 19% to 34% are severely affected by diabetic foot ulcers (DFUs) [3]. DFUs impose a substantial burden on healthcare systems and the broader economy due to their intractability, high recurrence rate, and elevated amputation risk [4,5]. Therefore, there is an urgent need for more effective therapeutic strategies to address such chronic non-healing wounds. Mitophagy is a cellular process that maintains homeostasis by eliminating damaged or dysfunctional mitochondria to regulate mitochondrial quality and quantity [6]. However, due to the persistent hyperglycemic injury and intramitochondrial oxidative stress characteristic of the diabetic wound microenvironment, the balance between mitochondrial fission and fusion is disrupted. This leads to the accumulation of impaired mitochondria, exacerbating mitochondrial damage and further compromising mitophagy [7,8]. Drp1, a key regulator of mitochondrial fission, plays a central role in signaling cascades governing cell differentiation, metabolism, proliferation, and cell death [9,10]. Excessive Drp1-mediated mitochondrial fission promotes the generation of mitochondrial reactive oxygen species (mtROS), further disrupting the fusion-fission equilibrium, aggravating mitochondrial injury, and suppressing mitophagy in diabetes-associated pathologies [11,12]. Therefore, modulation of Drp1 activity may represent an attractive therapeutic strategy for the treatment of diabetic wounds. Our previous research has identified aldehyde dehydrogenase 2 (ALDH2) as a critical molecule influencing mitophagy in diabetic wounds [13]. ALDH2 is the mitochondrial isoform of aldehyde dehydrogenase, and its mutational inactivation is closely related to type 2 diabetes mellitus [14,15]. ALDH2 deficiency has been associated with impaired autophagic flux, increased mitochondrial oxidative stress, pancreatic β-cell dysfunction, and disturbed glucose homeostasis, thereby contributing to type 2 diabetes and related tissue injury [16]. However, free recombinant ALDH2 has limited cellular uptake, intracellular stability, and mitochondrial delivery, which restricts its therapeutic application in diabetic wounds. Extracellular vesicles (EVs) are a heterogeneous group of lipid bilayer-delimited particles spanning the nanometer-to-micrometer size range [17], Owing to their natural lipid bilayer, favorable biocompatibility, cargo-protective capacity, and potential to cross biological barriers, EVs have attracted considerable interest as drug-delivery vehicles [18,19]. EVs can be loaded with ALDH2 for synergistic treatment of diabetic wounds. Nevertheless, the complex wound microenvironment can compromise EV integrity, and the formation of eschar and scar tissue at the wound site impedes effective EV delivery [20,21]. Therefore, constructing an efficient local delivery system is necessary to ensure the stable delivery of EVs to diabetic wounds. Microneedles are a novel drug delivery system comprising an array of micron-sized needles. Compared to traditional wound dressings, microneedles possess sufficient mechanical strength to penetrate wound barriers, significantly increasing transdermal drug permeation [22,23]. Specifically, hydrogel-based microneedle patches offer unique advantages for wound healing and tissue regeneration owing to their biodegradability and excellent biocompatibility [24,25]. More broadly, biomimetic scaffold design and biointerface engineering can actively modulate cell–material interactions and the local regenerative microenvironment, thereby promoting tissue repair and regeneration [[26], [27], [28]]. Based on these considerations, we loaded ALDH2 recombinant protein-carrying EVs (ALDH2@Lipo/EV) into methacrylated gelatin (GelMA) hydrogel to fabricate an ALDH2@Lipo/EV-MN patch for diabetic wound treatment. This study investigated the effects and mechanisms of ALDH2@Lipo/EV through in vitro experiments. The results demonstrated that ALDH2@Lipo/EV can inhibit the phosphorylation and activation of Drp1 (dynamin-related protein 1) and active the PINK1/Parkin signaling pathway, thereby facilitating mitophagy. Furthermore, in the diabetic wound mouse model, we validated the biocompatibility of the ALDH2@Lipo/EV-MN patch and its ability to promote angiogenesis, regulate macrophage polarization, and stimulate the release of anti-inflammatory cytokines. These findings confirm that the ALDH2@Lipo/EV-MN patch system is safe and enhances wound healing efficiency, suggesting that this microneedle system offers a promising therapeutic approach for the application of regenerative medicine in diabetic wounds (Fig. 1). Schematic illustration of the study. Preliminary bioinformatics analysis combined with machine-learning approaches was employed to screen for the key mitophagy-related gene ALDH2 in clinical DFU datasets. Differential expression of ALDH2 at both gene and protein levels was validated using post-surgical skin tissues, and its functional role was further confirmed experimentally. ALDH2 recombinant protein was loaded into liposomes and subsequently integrated into hAMSC-derived EVs to generate ALDH2@Lipo/EV. These engineered EVs were then incorporated into microneedles to fabricate the ALDH2@Lipo/EV-MN patch. Functional studies demonstrated that ALDH2@Lipo/EV promoted HUVEC activity and modulated macrophage polarization and the mitophagy phenotype in vitro. Furthermore, in vivo studies using a diabetic mouse wound model confirmed that the ALDH2@Lipo/EV-MN patch effectively accelerates diabetic wound healing. Materials and methods Isolation and identification of human amniotic mesenchymal stem cells Postpartum discarded placentas were collected, and the amniotic membrane was mechanically peeled from the placenta. After removing the adherent chorionic layer, the amniotic membrane was digested using trypsin and collagenase. The digest was passed through a sieve to remove large tissue fragments and cell clumps, and the filtrate was centrifuged to obtain a pellet of human amniotic mesenchymal stem cells (hAMSCs). The cells were cultured in serum-free medium (Dakewe, Shenzhen, China). When the cells reached 80% confluence, they were collected and induced to undergo osteogenic and adipogenic differentiation according to the instructions of the kit (Fuyuanbio, Shanghai, China). Osteogenic induction was terminated on day 21, and Alizarin Red S staining revealed red-stained nodules under low-power microscopy, which were identified as calcium nodules. Adipogenic differentiation was terminated on day 28, and Oil Red O staining showed intracellular red spherical lipid droplets under microscopy, confirming the osteogenic and adipogenic differentiation potential of hAMSCs. Passage-3 cells were collected, and the expression of MSC-specific surface antigens CD73, CD90, CD105, and HLA-DR was assessed using the corresponding antibodies and isotype controls (BD, USA). The cells were stained at 4°C for 30 min in the dark, and the expression of these markers was analyzed by flow cytometry (NovoCyte Advanteon, Agilent, China). Preparation of ALDH2-loaded extracellular vesicles (ALDH2@Lipo/EV) The conditioned medium from hAMSC cultures was collected. Extracellular vesicles (EVs) were isolated from the supernatant using tangential flow filtration (100 kDa cutoff; Sartorius, Germany), followed by extraction with an EV isolation kit (Umibio, Shanghai, China) according to the manufacturer's protocol. The preparation of ALDH2@Lipo/EV was achieved through a three-step process. Briefly, liposomes were first synthesized by dissolving egg yolk lecithin, cholesterol, and DSPE-PEG-COOH (Sigma-Aldrich) at a mass ratio of 1:1:8 in chloroform (Aladdin, Shanghai, China). The components were dissolved in chloroform (Aladdin, Shanghai, China), and ultrasonicated to form a homogeneous solution, and the chloroform was subsequently removed by rotary evaporation at 60°C. The resulting lipid film was vacuum-dried for 2 h, rehydrated with deionized water, and shaken at 37°C for 30 min. Finally, the suspension was sonicated to obtain unilamellar liposomes. Next, the ALDH2 recombinant protein (Cusabio, Wuhan, China) solution was mixed with the liposome suspension at a volume ratio of 1:3 and incubated at 37°C for 30 min. The mixture was then sequentially extruded through a liposome extruder to achieve size uniformity. The mixture was centrifuged at 4°C to collect the liposome precipitate loaded with the recombinant protein. In the final step, the isolated EVs were mixed with the ALDH2-loaded liposomes at a volume ratio of 1:2 and incubated at 37°C with gentle shaking (100 rpm) for 2 h to facilitate membrane fusion. Following fusion, the mixture was centrifuged at 10,000 rpm for 20 min at 4°C. The resulting pellet, consisting of the hybrid nanovesicles (ALDH2@Lipo/EV), was resuspended in phosphate-buffered saline (PBS). Characterization of ALDH2@Lipo/EV The morphology of the EVs was examined using transmission electron microscopy (TEM, Tecnai G2 F20, FEI, USA). A 10 μL sample was deposited onto a carbon-coated copper grid and negatively stained with 10 μL of 2% phosphotungstic acid solution for 1 min prior to imaging. Particle size distributions were characterized by nanoparticle tracking analysis (NTA; NanoSight NS300, Malvern) and dynamic light scattering (DLS; Zetasizer Ultra, Malvern). The zeta potential and UV–visible absorption spectra were measured using a Zetasizer Ultra (Malvern) and a UV–visible spectrophotometer, respectively. Western blotting was performed to detect the EV-associated membrane markers CD9, CD63, and CD81 and the cytosolic EV marker TSG101, as well as to verify the loading of ALDH2. Preparation of ALDH2@Lipo/EV microneedles (MNs) patch The microneedle (MN) patch was fabricated using a two-step vacuum-assisted method. First, an 8 mg/mL LAP photoinitiator solution (EFL, China) and a 15% GelMA solution were prepared. After degassing by centrifugation, the GelMA solution containing ALDH2@Lipo/EV was dispensed into the cavities of a polydimethylsiloxane (PDMS) mold featuring an array of microneedle-shaped voids. The mold was placed in a vacuum drying oven to evacuate air bubbles and then subjected to horizontal centrifugation. The formulation was concentrated by heating at 30°C for 5 h, and the casting process was repeated to increase the loading of ALDH2@Lipo/EV. Subsequently, gelation was initiated by exposure to a 405 nm UV light source, completing the fabrication of the needle body. The backing layer of the MN patch was composed of pullulan. A 25% (w/v) pullulan solution was prepared, centrifuged to ensure homogeneity, and then allowed to stand protected from light at 37°C for 36 h. Once the backing layer had dried, the microneedle array was carefully detached from the PDMS mold, yielding the final ALDH2@Lipo/EV-MN patch. Morphology and elemental analysis of MN patch The microneedle patch was sputter-coated with gold and examined using scanning electron microscopy (SEM; MIRA4, Tescan, Czechia) under high vacuum mode. Images were captured at 100× magnification, with the focus positioned on the microneedle tip region. The accelerating voltage was set to 15 kV, with a live time of 50 s, a count rate of ≥800 cps, and a dead time of ≤25%. Energy-dispersive X-ray spectroscopy (EDS) mapping was performed using a field-emission transmission electron microscope (Ultim Max65, Oxford, UK) to determine the elemental composition of the microneedles. The atomic percentage of each element was calculated using ZAF correction. Mechanical strength test The force-displacement curve was generated using a universal testing machine (AGX-V2, Shimadzu, Japan). The base of the microneedle patch was fixed onto a metal stage with double-sided conductive tape. A flat-ended tungsten carbide indenter with a diameter of 100 μm was employed under a quasi-static loading-holding-unloading mode. The loading rate was set at 2 mN/s, with the maximum load adjusted to 1.2 times the estimated fracture force of the sample. A holding time of 5 s was applied, and the unloading rate was kept consistent with the loading rate. Data were recorded at a sampling frequency of 500 Hz. In vitro biocompatibility of ALDH2@Lipo/EV-MNs In vitro cytotoxicity of ALDH2@Lipo/EV-MNs was evaluated using HUVECs and THP-1 cells. The ALDH2@Lipo/EV-MN patch was placed in a Transwell insert and co-cultured with cells seeded in plates for 24 h to assess biocompatibility. An untreated group served as the negative control. Cell viability was determined using the Live/Dead assay (Beyotime, Shanghai, China). Labeling and internalization of ALDH2@Lipo/EV Labeling and cellular uptake of ALDH2@Lipo/EV were performed according to the manufacturer's instructions for the PKH26 staining kit (Sigma-Aldrich). Briefly, 100 μL of ALDH2@Lipo/EV at a concentration of 100 μg/mL was incubated with 4 μL of PKH26 dye for 3 min at room temperature in the dark. The labeling reaction was terminated by adding an equal volume of 1% bovine serum albumin (BSA). The mixture was then transferred to a 100 kDa ultrafiltration unit (Amicon® Ultra, Millipore, USA) and centrifuged at 3000 × g for 10 min to remove unincorporated dye. When cells reached 60% confluence, 20 μL of PKH26-labeled ALDH2@Lipo/EV was added to the cells, which were then incubated for 8 h in the dark. Cells were then fixed with 4% paraformaldehyde at room temperature, followed by cytoskeletal and DAPI staining. Cellular uptake was visualized using a confocal fluorescence microscope. Release of ALDH2@Lipo/EV from MNs ALDH2@Lipo/EV were labeled with PKH67 fluorescent dye (Yeasen, China) and subsequently incorporated into the microneedles (MNs). Under sterile conditions, the ALDH2@Lipo/EV-MNs were separately immersed in centrifuge tubes containing either phosphate-buffered saline (PBS) or PBS supplemented with 33.3 mM glucose. The tubes were incubated in a thermostatic shaker at 37°C and 100 rpm. At predetermined time points, 200 μL of the release medium was collected and immediately replaced with an equal volume of the corresponding fresh medium. The fluorescence intensity of PKH67 in the collected supernatants was measured using a multimode microplate reader (Varioskan LUX, Thermo Fisher Scientific, USA), and the cumulative release of EVs@ALDH2 was subsequently calculated. Cell culture THP-1 cells were obtained from the Cell Bank of the Chinese Academy of Sciences (China) and cultured in RPMI 1640 medium (Gibco, Thermo Fisher, USA) supplemented with 10% fetal bovine serum (FBS). To induce differentiation into M0 macrophages, THP-1 cells were primed with 10 ng/mL phorbol 12-myristate 13-acetate (PMA) for 24 h. Primary human umbilical vein endothelial cells (HUVECs) were cultured in endothelial cell medium (ScienCell, USA). Primary human dermal fibroblasts (HDFs) were cultured in DMEM supplemented with 10% FBS. High-glucose (HG) medium was defined as complete medium containing 33.3 mM glucose (Sigma-Aldrich), whereas normal-glucose (NG) medium contained 5.5 mM glucose. Macrophages were pretreated with 20 μM Mdivi-1 for 1 h before the indicated treatments. MG149 was administered to macrophages at a final concentration of 50 μM during the indicated treatment period. EdU proliferation assay After the indicated treatments, cells were incubated with EdU working solution for 4 h. Following incubation in the dark, cells were fixed with 4% paraformaldehyde, and nuclei were counterstained with DAPI. After washing, samples were visualized under a fluorescence microscope, and images of three randomly selected fields were captured. Transwell migration analysis Cells were cultured under HG and NG conditions for 7 days. Subsequently, cells were harvested and seeded into the upper chamber of Transwell inserts. Conditioned media from different treatment groups were added to the lower chamber. After 24 h of incubation, the inserts were removed, and cells that had migrated to the lower surface of the membrane were fixed with 4% paraformaldehyde and stained with crystal violet. Non-migrated cells on the upper surface of the membrane were gently removed with a cotton swab. Migrated cells were then counted under a microscope. Scratch experiment Cells subjected to different treatments were seeded into the two compartments of a culture insert (Corning, USA). After cell attachment, the insert was gently removed, and the wells were rinsed twice with pre-warmed serum-free medium. Wound healing was assessed by capturing images at 0 h and 24 h post-insert removal. Tube formation analysis Following different treatments, cells were seeded into a 24-well plate pre-coated with Matrigel (Corning, USA). Cells were cultured in serum-free medium for 4 h, and tube formation was observed under a light microscope every hour. The tube formation efficiency was then quantified. Intracellular ROS detection The DCFH-DA probe working solution (Beyotime, Shanghai, China) was prepared and added to the cell culture plate. After incubation in the dark for 20 min, cells were observed and photographed using an inverted fluorescence microscope. Mitochondrial membrane potential assay The JC-1 staining working solution (Beyotime, Shanghai, China) was prepared and mixed with the cell culture medium at a 1:1 ratio. After incubation in the dark, cells were observed and photographed using an inverted fluorescence microscope. In vitro characterization of macrophage polarization THP-1 cells (2 × 105 cells/well) were seeded into 24-well plates and primed with PMA for 24 h to induce differentiation into M0 macrophages. Cells were then assigned to the indicated treatment groups and treated for 24 h, with the HG group serving as a positive control. For immunofluorescence analysis, cells were fixed with 4% paraformaldehyde, permeabilized with Triton X-100, and blocked with 5% BSA. Subsequently, cells were incubated overnight at 4°C with primary antibodies against the M1 macrophage marker (iNOS) and the M2 macrophage marker (CD206), followed by incubation with corresponding fluorophore-conjugated secondary antibodies. Nuclei were counterstained with DAPI. Expression of these markers was visualized using a confocal microscope. Total RNA was extracted from macrophages in each group using the TRIzol reagent. RNA concentration and purity were determined using a NanoDrop 2000 spectrophotometer, and cDNA was synthesized. The transcript levels of pro-inflammatory mediators (IL1B, IL6, and TNF), the anti-inflammatory mediator IL10, and the proangiogenic factor VEGFA were quantified by qPCR. Gene expression was normalized to β-actin using the 2−ΔΔCt method. Lentiviral transduction and validation THP-1 cells (1 × 106 cells/well) were seeded into 12-well plates and co-incubated with 3 mL of lentivirus at a titer of 1 × 107 TU/mL for 3 h. After 24 h, cell morphology and fluorescence intensity were observed. THP-1 cells that were untransduced and those transduced with lentivirus (5 × 105 cells/well) were then seeded into 6-well plates and primed with PMA for 24 h to induce differentiation into M0 macrophages. All cells were subsequently treated with HG medium for 24 h. Total RNA and total protein were extracted to validate ALDH2 gene and protein expression by qPCR and Western blot, respectively. Immunofluorescence staining Following treatment, cells were fixed with 4% paraformaldehyde, permeabilized with Triton X-100, and blocked with 5% BSA. The cells were then incubated overnight at 4°C with primary antibodies, followed by incubation with corresponding fluorophore-conjugated secondary antibodies. Nuclei were counterstained with DAPI. Protein expression was examined using a confocal microscope. Transmission electron microscopy analysis of mitochondrial ultrastructure Following different treatments, macrophages were fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.4) for 2 h at room temperature. The cells were collected, pre-embedded in 2% low-melting-point agarose, and postfixed with 1% osmium tetroxide for 1 h. After graded ethanol dehydration and epoxy resin embedding, ultrathin sections (70–80 nm) were prepared, stained with uranyl acetate and lead citrate, and examined by transmission electron microscopy at 80 kV. Analysis of mitophagy-related proteins expression Following the indicated treatments, total cellular proteins were extracted and subjected to Western blot analysis to determine total Drp1 expression, Drp1 phosphorylation at Ser616, and the expression of mitophagy-related proteins, including PINK1, Parkin, p62, and LC3. Target protein expression was normalized to β-actin, while phosphorylated proteins were normalized to their corresponding total proteins. Mouse wound model establishment and treatment A wound model was established on the dorsum of diabetic mice to evaluate the in vivo effects of ALDH2@Lipo/EV-MNs. Eight-week-old male db/db mice and wild-type C57BL/6J mice used in this study were purchased from GemPharmatech Co., Ltd. (Jiangsu, China). All animal experiments were approved by the Animal Care and Use Committee of the First Affiliated Hospital of Nanchang University (Approval No. CDYFY-IACUC-202407QR066). Wild-type C57BL/6J mice served as the nondiabetic wound control group (n = 6), whereas db/db mice were randomly assigned to four experimental groups (n = 6 per group). On day 7 after wound induction, half of the mice in each group were randomly selected and euthanized (n = 3), and dorsal skin tissues were harvested. The remaining mice were euthanized on day 14 (n = 3 per group) for tissue collection. Wound closure was calculated as follows: wound closure (%) = [(A0 − At)/A0] × 100%, where A0 and At represent the wound areas on day 0 and the indicated day, respectively. Histological analyses, including hematoxylin and eosin (H&E) staining, Masson's trichrome staining, and CD31 immunofluorescence staining, were performed to assess wound pathology and angiogenesis. Excised dorsal skin tissues were also subjected to ex vivo microneedle insertion tests. The in vivo toxicity of ALDH2@Lipo/EV-MNs was evaluated by histological examination of the major organs, including the brain, heart, liver, spleen, lungs, and kidneys. To assess the immunomodulatory activity of ALDH2@Lipo/EV-MNs in vivo, wound tissues were collected on day 7. Macrophage polarization within the wound tissue was evaluated by immunofluorescence staining for inducible nitric oxide synthase (iNOS) and CD206, whereas the inflammatory status was evaluated by immunohistochemical staining for IL-6, IL-10, TNF-α, and VEGFA. Data analysis Quantitative data in this study were analyzed using Microsoft Excel 2021 and GraphPad Prism 9 software. Each experimental group contained at least three technical replicates, and results were expressed as mean ± SD. Differences between two groups and among multiple groups were analyzed using Student's t-test and one-way analysis of variance (ANOVA), respectively. In all figures, data are presented as mean ± SD, with significant differences indicated as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 and ns: not significant. Results ALDH2 identification and functional validation We previously employed bioinformatics analysis combined with machine-learning approaches to screen for mitophagy-related molecules in diabetic foot ulcers (DFUs) and successfully identified ALDH2 as a target molecule [13]. The expression of ALDH2 in normal skin and diabetic wound skin was validated using clinical tissue samples (Fig. 2a). Our preliminary studies indicated that ALDH2 expression was downregulated in diabetic wounds. THP-1 cells were transduced with lentivirus overexpressing ALDH2 (Fig. 2b), and strong GFP fluorescence indicating ALDH2 overexpression was observed in THP-1 cells under microscopy (Fig. 2c). Bioinformatics screening, identification, and functional validation of ALDH2. (a) Schematic diagram illustrating the bioinformatics analysis combined with machine-learning approaches for screening key mitophagy-related molecules in DFU. (b) Schematic representation of lentiviral transduction for ALDH2 overexpression in THP-1 cells. (c) Representative fluorescence microscopy image showing GFP expression in THP-1 cells following successful lentiviral transduction. (d) Quantitative real-time PCR analysis of ALDH2 gene expression levels in THP-1 cells transduced with different lentiviruses. (e) Western blot bands showing protein expression of ALDH2 and Drp1 in macrophages from the control and overexpression groups, with the corresponding quantitative analysis shown in (f). (g) Representative immunofluorescence images of Drp1 and TOM20 co-staining in macrophages from the control and overexpression groups. (h) Quantitative analysis of Drp1 expression levels. (Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 and ns: not significant.). To further validate successful transduction and overexpression of ALDH2 in THP-1 cells, THP-1 cells in the control and overexpression groups were induced with PMA to differentiate into M0 macrophages and then cultured in HG medium for 24 h. ALDH2 gene and protein expression were validated by qPCR and Western blot, confirming successful transduction and ALDH2 expression in THP-1 cells (Fig. 2d–f). Immunofluorescence staining for Drp1 and TOM20 was performed on macrophages from control and overexpression groups. Fluorescence imaging revealed that Drp1 expression within mitochondria was significantly reduced in ALDH2-overexpressing macrophages compared to the control group (Fig. 2g and h). hAMSCs express MSC-specific markers and undergo multilineage differentiation Human amniotic mesenchymal stem cells (hAMSCs) have emerged as a promising cell source for regenerative medicine applications due to their low or absent immunogenicity, lack of tumorigenicity, absence of ethical concerns, rapid proliferative capacity, and potent paracrine activity [29,30]. hAMSCs secrete abundant extracellular vesicles (EVs). As natural nanoparticles, EVs exhibit superior in vivo tolerance and lack immunogenicity compared to synthetic nanoparticles, thus garnering significant attention for their development as targeted drug delivery systems [31]. The morphology of cultured hAMSCs, observed under light microscopy, revealed a long spindle-shaped or polygonal appearance, as shown in Fig. 3a. To further characterize the hAMSCs, multilineage differentiation assays and flow cytometric analysis were performed. As demonstrated in Fig. 3b and c, the hAMSCs successfully differentiated into adipocytes and osteoblasts, confirming their multilineage differentiation potential. Flow cytometric analysis (Fig. 3d) showed that the hAMSCs expressed the MSC surface markers CD73, CD90, and CD105, but did not express HLA-DR, further confirming their identity as mesenchymal stem cells. Characterization of hAMSCs, hAMSCs-EVs, and ALDH2@Lipo/EV. (a) Representative morphology of hAMSCs. (b) Following successful adipogenic differentiation, distinct red lipid droplets were visualized within the cytoplasm. (c) Following successful osteogenic differentiation, formation of orange-red mineralized nodules was observed. (d) Flow cytometric analysis of positive and negative surface markers expressed on hAMSCs. (e) Schematic illustration of the preparation process for ALDH2@Lipo/EV. (f) TEM images depicting the size and morphology of liposomes, EVs and ALDH2@Lipo/EV. (g) Zeta potential measurements of liposomes, EVs, free ALDH2 protein and ALDH2@Lipo/EV. (h)DLS analysis of liposomes, EVs and ALDH2@Lipo/EV. (i) Particle size distribution analysis of EVs and ALDH2@Lipo/EV. (j) UV-vis absorbance spectra of EVs and ALDH2@Lipo/EV. (k) Western blot analysis confirming the presence of ALDH2 and EV-specific markers. Characterization of hAMSCs-EVs and ALDH2@Lipo/EV As illustrated in Fig. 3e, EVs were isolated from the supernatant of hAMSC culture medium using tangential flow filtration combined with a precipitation-based exosome isolation kit, followed by the preparation of ALDH2@Lipo/EV. The fabrication of ALDH2@Lipo/EV involved three steps: liposome synthesis, loading of ALDH2 protein into liposomes, and hybridization of EVs with the protein-loaded liposomes. The ALDH2 loading content was indirectly calculated to be 18.2 μg of ALDH2 per mg of EV protein by quantifying unbound ALDH2 in the ultrafiltrate obtained after ultrafiltration of the hybrid vesicle suspension. TEM images revealed that the liposomes exhibited a uniform spherical morphology with clearly defined bilayer membranes, whereas the vesicles isolated from hAMSCs displayed the characteristic discoid morphology, confirming the successful isolation of EVs. Notably, ALDH2@Lipo/EV retained the typical morphology of EVs, and no obvious aggregation was observed, indicating good colloidal stability of the nanocarrier (Fig. 3f).The zeta potentials of free ALDH2, liposomes, EVs, and ALDH2@Lipo/EV were measured under identical buffer conditions. The liposomes and EVs exhibited zeta potentials of approximately −17.3 ± 1.11 mV and −15.4 ± 1.17 mV, respectively, whereas free ALDH2 showed a zeta potential of −26.3 ± 0.21 mV. The zeta potential of ALDH2@Lipo/EV was approximately −24.7 ± 3.45 mV, representing a further negative shift relative to those of the liposomes and EVs and approaching the surface charge characteristics of free ALDH2 (Fig. 3g). DLS analysis showed that the liposomes and EVs had mean hydrodynamic diameters of approximately 147 and 142 nm, respectively, whereas the mean hydrodynamic diameter of ALDH2@Lipo/EV increased to approximately 178.8 nm. This increase in particle size was consistent with particle restructuring following liposome–EV hybridization and ALDH2 loading (Fig. 3h). To further confirm the successful loading of ALDH2, nanoparticle tracking analysis (NTA) was performed to determine the particle-size distributions of EVs and ALDH2@Lipo/EV. The results showed that the modal particle size of EVs was 164.65 ± 1.35 nm, whereas that of ALDH2@Lipo/EV increased to 184.28 ± 1.73 nm (Fig. 3i). The ultraviolet absorption spectra of EVs and ALDH2@Lipo/EV were subsequently measured using a UV–visible spectrophotometer. ALDH2@Lipo/EV exhibited a markedly increased absorbance at 280 nm compared with EVs (Fig. 3j). Western blotting revealed no detectable ALDH2 band in the unloaded EV group, whereas a distinct band was observed at the expected molecular weight in the ALDH2@Lipo/EV group. Furthermore, Western blot analysis confirmed the presence of key EV markers, including CD9, CD63, CD81, and TSG101, in both EVs and ALDH2@Lipo/EV, all of which were highly expressed in MSC-derived EVs (Fig. 3k). These results collectively validated the characteristic morphology, size distribution, surface charge, and specific markers of the extracellular vesicles, demonstrating the successful isolation of EVs and efficient loading of ALDH2. Cellular uptake assays demonstrated that ALDH2@Lipo/EV could be internalized by human dermal fibroblasts, human umbilical vein endothelial cells, and THP-1-derived macrophages, indicating its ability to enter cell types involved in diabetic wound repair and providing a cellular basis for its subsequent biological effects (Fig. S1). ALDH2@Lipo/EV promotes angiogenesis in vitro Angiogenesis and the inflammatory response regulated by macrophages play pivotal roles in promoting wound healing [32,33]. Newly formed blood vessels continuously supply essential oxygen and nutrients for tissue metabolism during the healing process [34]. In this study, primary HUVECs were establish an in vitro cell model of angiogenesis in the context of diabetic wounds. Cells were cultured in medium containing 33.3 mM glucose (high glucose, HG group), with 5.5 mM glucose (normal glucose, NG group) serving as the control. HUVECs were identified as vascular endothelial cells via immunofluorescence staining and flow cytometry using the vascular-specific marker CD31, with CD45 as a negative control (Fig. 4a and b). ALDH2@Lipo/EV promotes angiogenesis in vitro. (a) Immunofluorescence identification of the vascular endothelial cell marker CD31 in primary HUVECs. (b) Flow cytometric analysis of positive and negative surface markers on HUVECs. (c) Representative images of the wound healing assay in HUVECs following different treatments. (d) Quantitative analysis of HUVEC migration efficiency at 0 h and 24 h post-treatment, as determined by the wound healing assay. (e) Representative images from the EdU assay assessing HUVEC proliferation in different groups. (f) Quantitative analysis of HUVEC proliferative activity in different groups. (g) Representative images from the Transwell assay evaluating the chemotactic capacity of HUVECs in each group. (h) Quantification of migrated HUVECs in each treatment group. (i) Representative images of the tube formation assay in HUVECs after different treatments. (j) Quantitative analysis of the number of branch points and total tube length formed by HUVECs under different treatment conditions. (Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 and ns: not significant.). We observed that the migratory capacity of HUVECs was significantly reduced in the Control group compared to the NG group, indicating that the hyperglycemic microenvironment impairs cell migration. Notably, treatment with ALDH2@Lipo/EV significantly enhanced the migratory ability of HUVECs relative to the Control group (Fig. 4c and d). In the EdU proliferation assay, cell proliferative activity was markedly inhibited by the hyperglycemic microenvironment in the Control group compared to the NG group. However, compared to the Control group, the ALDH2@Lipo/EV group exhibited an average increase of 20.37% in cell proliferation rate (p < 0.05) (Fig. 4e and f). Transwell assays demonstrated that treatment with EVs and, more prominently, ALDH2@Lipo/EV significantly increased the number of migrating HUVECs compared to the Control group (Fig. 4g and h). Furthermore, in vitro tube formation assays (Fig. 4i and j) revealed that, compared to the Control group, ALDH2@Lipo/EV treatment increased the total tube length by 1.74-fold (p < 0.05) and increased the number of branch points by 3.27-fold (p < 0.05). These experimental results collectively confirm that ALDH2@Lipo/EV treatment significantly enhance the proliferation, migration, and tube formation abilities of vascular endothelial cells. ALDH2@Lipo/EV modulates macrophage polarization and restores mitochondrial homeostasis Appropriate inflammatory responses facilitate the clearance of pathogens and necrotic tissue from wounds while also promoting the release of cytokines, such as IL-10 and VEGF-A, which regulate the repair process [35]. Macrophages play an indispensable role in this process [36]. However, the hyperglycemic and oxidative stress microenvironment in diabetic wounds impedes the anti-inflammatory differentiation of macrophages, thereby delaying wound healing [37,38]. Consequently, regulating the switch between pro-inflammatory and anti-inflammatory phenotypes of macrophages is critical for cutaneous wound repair. Macrophages were loaded with the DCFH-DA fluorescent probe to assess intracellular ROS levels across different treatment groups. The results showed significantly elevated ROS levels in the control group compared to the NG group, indicating increased oxidative stress induced by the hyperglycemic microenvironment. This elevation was not ameliorated in the EVs group. Notably, ALDH2@Lipo/EV effectively reduced ROS levels, bringing them close to those observed in the NG group (Fig. 5a). To investigate the impact of the hyperglycemic and oxidative stress microenvironment on macrophage polarization, immunofluorescence staining for the polarization markers iNOS and CD206 was performed. Immunofluorescence images (Fig. 5b–c, d-e) confirmed that CD206 expression was lowest in the HG group and was significantly increased by ALDH2@Lipo/EV treatment. In contrast, iNOS expression was highest in the HG group and was significantly reduced by ALDH2@Lipo/EV treatment. These observations indicate a shift in macrophage phenotype, with ALDH2@Lipo/EV treatment promoting a more anti-inflammatory macrophage phenotype within the hyperglycemic microenvironment. This phenotypic shift is crucial for wound healing. ALDH2@Lipo/EV treatment reduced IL1B, IL6, and TNF expression while increasing IL10 and VEGFA expression (Fig. 5f). Effects of ALDH2@Lipo/EV on oxidative stress, macrophage polarization, and mitochondrial homeostasis. (a) Representative fluorescence images showing ROS levels in macrophages following different treatments. (b-c) Representative immunofluorescence images of iNOS and CD206 staining in macrophages after various treatments. (d-e) Quantitative analysis of iNOS and CD206 expression levels in macrophages from each treatment group. (f) qPCR analysis of inflammatory cytokine gene expression levels in macrophages from different treatment groups. (g) TEM imaging of the intracellular ultrastructure of macrophages. (h) Representative images of JC-1 staining assessing mitochondrial membrane potential changes in macrophages following different treatments. (i) Quantitative analysis of mitochondrial membrane potential levels in macrophages from each group. (Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 and ns: not significant.). To further investigate mitochondrial homeostasis, TEM was used to examine mitochondrial ultrastructure in macrophages following the indicated treatments, while mitochondrial membrane potential was assessed using JC-1 staining. In the NG group, mitochondria exhibited well-defined outer membrane boundaries, relatively intact cristae, and a uniformly electron-dense matrix, indicating normal mitochondrial ultrastructure. Mitochondria-containing autolysosomal structures were also frequently observed. In the Control group, some mitochondria displayed swollen cristae, suggesting mitochondrial injury. Compared with the Control group, mitochondria in the EVs group exhibited a more orderly cristae arrangement. Treatment with ALDH2@Lipo/EV alleviated mitochondrial cristae swelling and increased the presence of mitophagy-associated structures, consistent with enhanced mitophagic activity (Fig. 5g). JC-1 staining further demonstrated a marked reduction in mitochondrial membrane potential in macrophages in the Control group. Although the EVs group showed an increase in JC-1 aggregates, the difference was not statistically significant. In contrast, ALDH2@Lipo/EV treatment reduced both the abundance and fluorescence intensity of JC-1 monomers, indicating a significant restoration of mitochondrial membrane potential (Fig. 5h and i). To further investigate the mechanisms underlying the altered mitophagy-associated phenotype, Western blot analysis was performed to evaluate total Drp1 expression, Drp1 phosphorylation at Ser616, and the expression of proteins associated with the PINK1/Parkin pathway in macrophages following different treatments. First, macrophages cultured under high-glucose conditions were treated with Mdivi-1 to pharmacologically inhibit Drp1. The levels of Drp1 Ser616, PINK1, Parkin, p62, and LC3B were subsequently examined. Compared with the NG group, the Control group exhibited increased p62 expression and a reduced LC3-II/LC3-I ratio, suggesting impaired autophagic substrate degradation and decreased mitophagy-associated activity under high-glucose conditions. Mdivi-1 treatment reduced Drp1 phosphorylation at Ser616, increased PINK1 and Parkin expression, decreased p62 accumulation, and elevated the LC3-II/LC3-I ratio. These findings indicate that pharmacological inhibition of Drp1 activation at Ser616 may improve mitophagy-associated processes impaired by high-glucose exposure (Fig. 6a and b). Effects of ALDH2@Lipo/EV on the expression of mitophagy-related proteins in macrophages. (a, c, e) Representative Western blot images. (b, d, f) Densitometric quantification of corresponding protein levels. The expression levels of target proteins were normalized to β-actin, whereas phosphorylated protein levels were normalized to their corresponding total protein levels. (Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 and ns: not significant.). As shown in Fig. 6c, EV treatment partially ameliorated these alterations compared with the Control group, although the effects were relatively modest. In contrast, ALDH2@Lipo/EV treatment markedly reduced p62 expression and increased the LC3-II/LC3-I ratio. Together with the upregulation of PINK1 and Parkin and the reduction in Drp1 Ser616 phosphorylation, these results suggest that ALDH2@Lipo/EV may enhance mitophagy-associated activity by suppressing excessive Drp1 activation and promoting PINK1/Parkin-related signaling (Fig. 6c and d). To further assess the involvement of PINK1/Parkin signaling in the effects of ALDH2@Lipo/EV, MG149 was used to pharmacologically interfere with PINK1-related mitophagy signaling. Western blot analysis showed that MG149 partially attenuated the ALDH2@Lipo/EV-induced increases in PINK1 and Parkin expression. Compared with the ALDH2@Lipo/EV group, combined MG149 treatment increased p62 accumulation and reduced the LC3-II/LC3-I ratio. In addition, Drp1 phosphorylation at Ser616 was increased following MG149 treatment, suggesting a potential functional interaction between Drp1 activation and PINK1/Parkin-mediated mitophagy. However, these findings do not establish a direct bidirectional regulatory relationship between the two processes (Fig. 6e and f). Collectively, these results suggest that ALDH2@Lipo/EV modulates macrophage polarization, reduces intracellular ROS accumulation, and improves mitochondrial quality control under high-glucose conditions. The Drp1–PINK1/Parkin axis may contribute to the ALDH2@Lipo/EV-mediated enhancement of mitophagy-associated activity and restoration of mitochondrial homeostasis. ALDH2@Lipo/EV-MNs fabrication and characterization Diabetic wounds are characterized by substantial exudate and the formation of eschar, which hinder efficient drug penetration [39]. The MN patch circumvents these obstacles by penetrating the stratum corneum barrier and creating transient microchannels in the skin, thereby significantly enhancing transdermal drug delivery. In this study, we incorporated ALDH2@Lipo/EV into a GelMA hydrogel-based MN system to protect them from the inflammatory microenvironment and ensure their effective and efficient delivery. ALDH2@Lipo/EV were mixed with the GelMA hydrogel solution and cast into a polydimethylsiloxane (PDMS) mold. The mixture was subjected to repeated vacuum aspiration and drying, followed by crosslinking and solidification upon exposure to focused 405 nm UV light. Subsequently, a 25% pullulan solution was uniformly applied onto the mold to form the backing layer of the MN patch. After thorough drying, the patch was carefully peeled from the mold, yielding the final ALDH2@Lipo/EV-MN patch (Fig. 7a). Based on the concentration of ALDH2@Lipo/EV added to the hydrogel precursor solution, the theoretical ALDH2@Lipo/EV loading content of the microneedle patch was calculated to be 0.4 wt%, corresponding to approximately 4 mg of ALDH2@Lipo/EV per gram of microneedle patch. The calculated encapsulation efficiency was approximately 92% based on the amount of ALDH2@Lipo/EV released into the medium within 24 h under 37°C and 100 rpm shaking conditions. Fabrication and characterization of ALDH2@Lipo/EV-MN patch. (a) Schematic illustration depicting the fabrication process of ALDH2@Lipo/EV-MN patch. (b) High-resolution photograph displaying the appearance and morphology of the fabricated ALDH2@Lipo/EV-MN patch. (c) Scanning electron microscopy (SEM) image revealing the microstructure of the ALDH2@Lipo/EV-MN patch. (d) Elemental mapping showing the composition and distribution of carbon (C), nitrogen (N), oxygen (O), phosphorus (P), and sulfur (S) within the ALDH2@Lipo/EV-MN patch. (e) EDS elemental mapping and the corresponding summed EDS spectrum of the microneedle bodies in the ALDH2@Lipo/EV-MN patch. (f) Force-displacement curve obtained from mechanical testing of the ALDH2@Lipo/EV-MN patch. (g) Sustained release profile of ALDH2@Lipo/EV from the microneedle patch in PBS and PBS supplemented with 33.3 mM glucose. Observation of the MN patch revealed a well-ordered 12 × 12 array of microneedles (Fig. 7b). SEM imaging showed that the microneedles were conical, with a height of 780 μm, a base diameter of 400 μm, and a center-to-center spacing of 850 μm. The microneedles were arranged in a regular array on the backing layer, a configuration designed to facilitate skin penetration (Fig. 7c). Elemental mapping showed a relatively uniform distribution of C, N, O, P, and S across the analyzed region of the microneedles. (Fig. 7d and e). Mechanical testing of the needle tips revealed a significant increase in force at a displacement of 1.5 mm, indicative of the critical force required for stratum corneum penetration (≈0.9 N/needle). The force-displacement curve exhibited fluctuations and a subsequent sharp drop at ≈4.8 N/needle, with the maximum insertion force reached at a displacement of 1.75 mm. These results demonstrate that the microneedles possess excellent mechanical integrity and can withstand forces far greater than those required for skin insertion. The initial low-force region suggests adequate tip elasticity for gentle skin contact, aligning with the principles of minimally invasive drug delivery. The subsequent rise in force confirms that the microneedles maintain their functionality after deformation, preventing collapse post-insertion (Fig. 7f). To assess the release kinetics of ALDH2@Lipo/EV-MNs, ALDH2@Lipo/EV were labeled with PKH67 dye prior to patch fabrication. The resulting patches were immersed in phosphate-buffered saline (PBS) or PBS supplemented with 33.3 mM glucose. Release profiles were generated by monitoring the absorbance of the supernatant at 490 nm. The cumulative release rates of EVs@ALDH2 from the microneedle patches over 21 days were 64.77% in PBS and 63.4% in high-glucose PBS, respectively, demonstrating sustained release over a period compatible with the prolonged healing process of diabetic wounds (Fig. 7g). Biocompatibility of ALDH2@Lipo/EV-MNs As shown in Fig. 8a, HUVECs and macrophages treated with extracts from blank MNs and ALDH2@Lipo/EV-MNs were subjected to Live/Dead staining to assess cell viability, with untreated cells serving as the negative control group. The results demonstrated that, compared to the Control group, neither blank MNs nor ALDH2@Lipo/EV-MNs reduced cell viability (Fig. 8b and c), indicating their biocompatibility. In vitro and in vivo biocompatibility of ALDH2@Lipo/EV-MNs. (a) Schematic illustration of live/dead fluorescence staining of HUVECs or macrophages following no treatment or co-culture with blank MN or ALDH2@Lipo/EV-MNs. (b) Live/dead staining assessing the viability of HUVECs and macrophages under the indicated conditions. (c) Quantitative analysis of cell viability in different treatment groups for HUVECs and macrophages. (d) Schematic illustration of the major organs in mice. (e) Gross photographs of major organs from mice. (f) Representative histological cross-sectional images of excised mouse skin following microneedle insertion. (g) H&E staining of major organs harvested from mice in the ALDH2@Lipo/EV-MNs treatment group. (Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 and ns: not significant.). As shown in Fig. 8d, mice in the ALDH2@Lipo/EV-MNs treatment group were euthanized, and their major organs were harvested. Representative gross images of the collected organs are presented in Fig. 8e. To evaluate the skin penetration capability of the microneedle patches, the patches were inserted into excised mouse skin, followed by histological examination of transverse skin sections. H&E staining revealed well-defined microchannels within the tissue, confirming the effective penetration of the microneedles (Fig. 8f). The brain, heart, liver, spleen, lungs, and kidneys were further subjected to gross morphological and histopathological evaluation using H&E staining. No obvious pathological abnormalities were observed, supporting the short-term in vivo biocompatibility and tolerability of ALDH2@Lipo/EV-MNs for diabetic wound treatment (Fig. 8g). ALDH2@Lipo/EV-MNs promote diabetic wound healing in mice Based on the observed in vitro pro-angiogenic, antioxidant, and anti-inflammatory activities of ALDH2@Lipo/EV, we conducted an in vivo study by incorporating them into a hydrogel microneedle patch to validate their therapeutic potential in a diabetic mouse wound model. Following one week of acclimatization, a standardized full-thickness skin wound (10 mm) was created on the dorsum of wild-type C57BL/6J (WT) and db/db mice. Beginning on day 0 after wounding, the wounds were treated with saline, blank MNs, EVs-MNs, or ALDH2@Lipo/EV-MNs. Throughout the study, wounds were monitored, photographed, and measured at designated time points (Fig. 9a). ALDH2@Lipo/EV-MN patch promotes diabetic wound healing in mice. (a) Schematic illustration of the animal experiment, outlining the establishment, treatment, and processing of the diabetic wound model in mice. (b) Representative macroscopic images of dorsal wounds in mice from each group over time throughout the treatment period, together with corresponding image-analysis simulations of wound closure. (c) Quantitative analysis of the relative wound area for each group at different time points. (Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 and ns: not significant.). At the end of the 14-day observation period, treatment with blank MNs showed no improvement compared to the Control group. Wounds treated with EVs-MNs exhibited a modest reduction in area relative to the Control group. Notably, recovery was further enhanced in the ALDH2@Lipo/EV-MNs group. Fig. 9b and c shows that wounds treated with ALDH2@Lipo/EV-MNs appeared drier and cleaner, exhibited less exudation, and closed more rapidly than wounds in the other diabetic groups. These compelling findings highlight the promising therapeutic efficacy of ALDH2@Lipo/EV-MNs in accelerating diabetic wound healing in mice. Histological analysis On day 7, half of the mice were randomly euthanized, and wound skin tissues were harvested. On day 14, the remaining mice were euthanized, and skin tissues from the wounds of all treatment groups were collected for histological examination. H&E staining was performed to evaluate re-epithelialization, epidermal architecture, and granulation-tissue formation. As shown in Fig. 10a and b, on days 7 and 14, no significant improvement in epidermal proliferation was observed in the MNs group compared to the Control group. The EVs-MNs group promoted epidermal migration relative to the MNs group, while treatment with ALDH2@Lipo/EV-MNs resulted in the greatest increase in wound epidermal thickness. Given the importance of collagen deposition and remodeling in wound healing, Masson's trichrome staining was performed. The results demonstrated that on both day 7 and 14 post-wounding, ALDH2@Lipo/EV-MNs-treated wounds exhibited denser collagen deposition and a more organized collagen-fiber arrangement than those in all other db/db groups (Fig. 10c and d). Histological analysis of diabetic wound healing promoted by the ALDH2@Lipo/EV-MN patch in mice. (a) Representative HE staining images of mouse wound skin tissues on day 7 and day 14. (b) Quantitative analysis of epidermal thickness measured from HE-stained wound skin tissues on day 7 and day 14. (c) Representative Masson's trichrome staining results of mouse wound skin tissues on day 7 and day 14. (d) Quantitative analysis of collagen deposition from Masson's trichrome-stained wound skin tissues on day 7 and day 14. (e) Representative immunofluorescence images of CD31 expression in mouse wound skin tissues from different treatment groups on day 14. (f) Quantitative analysis of relative CD31 fluorescence intensity in mouse wound skin tissues from different treatment groups on day 14. Immunofluorescence staining was employed to detect the expression of specific macrophage polarization markers—iNOS and CD206—in tissues on day 7, and the expression of platelet endothelial cell adhesion molecule-1 (CD31) in skin tissues on day 14. As illustrated in Fig. 10e and f, the Control and MNs groups showed the lowest CD31 fluorescence intensity, indicating slower angiogenesis. The EVs-MNs group displayed a modest increase in CD31 expression compared to the Control group, but this effect was less pronounced than in the ALDH2@Lipo/EV-MNs group. These findings confirm the pro-angiogenic effect of ALDH2@Lipo/EV-MNs during wound healing. Fluorescence imaging revealed that in the Control, MNs, and EVs-MNs groups, the majority of macrophages differentiated towards the M1 phenotype during the wound inflammatory phase. In contrast, immunofluorescence staining of wound skin tissues from the ALDH2@Lipo/EV-MNs treatment group showed significantly increased expression of the M2 macrophage marker CD206 compared to the Control group (Fig. 11a and b). To further corroborate these findings, immunohistochemical staining for cytokines was performed on wound tissues harvested on day 7, and positive expression rates were quantified. ALDH2@Lipo/EV-MNs treatment downregulated the pro-inflammatory cytokines IL-6 and TNF-α, while concurrently upregulating the anti-inflammatory cytokine IL-10 and vascular endothelial growth factor (VEGF-A), highlighting the immunomodulatory potential of ALDH2@Lipo/EV-MNs (Fig. 11c and d). Immunomodulatory activity of the ALDH2@Lipo/EV-MN patch in diabetic wounds in mice. (a) Representative immunofluorescence co-staining images of F4/80 with iNOS or CD206 in mouse wound skin tissues from different treatment groups on day 7. (b) Quantitative analysis of immunofluorescence co-staining for F4/80 and iNOS/CD206 in mouse wound skin tissues following different treatments on day 7. (c) Quantitative analysis of inflammatory and proangiogenic marker expression in wound skin tissues from each treatment group on day 7. (d) Quantitative analysis of inflammatory cytokine expression in wound tissues on day 7. (Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 and ns: not significant.). Discussion Diabetic wounds present a hostile microenvironment characterized by persistent exudation, oxidative stress, chronic inflammation, proteolytic activity, and dysregulated extracellular matrix remodeling, which may reduce the stability and local availability of protein therapeutics and nanoscale carriers. In this study, ALDH2-loaded liposomes were fused with hAMSC-derived EVs to generate ALDH2@Lipo/EV and were subsequently incorporated into photocrosslinked GelMA microneedles. The microneedle array was designed to create transient microchannels, whereas the GelMA network served as a local depot for cargo protection and sustained release. The observed mechanical properties and sustained release profile in PBS support the feasibility of this delivery concept. However, the actual insertion depth, intratissue distribution, and in vivo retention of ALDH2@Lipo/EV remain to be directly quantified. Unlike multifunctional microneedles that rely on metal ions, photothermal stimulation, or microenvironment-responsive reactions, ALDH2@Lipo/EV-MNs integrate a biologically derived hybrid vesicle with a biodegradable and photocrosslinkable hydrogel matrix. This represents a distinct design strategy rather than demonstrated superiority over existing systems. The hybrid vesicles combine the protein-loading capability of liposomes with the potential cell-interaction advantages of EV-derived membranes, while GelMA provides localized and sustained delivery. The in vitro and in vivo findings suggest that ALDH2@Lipo/EV-MNs promote diabetic wound repair through coordinated effects on angiogenesis, oxidative stress, macrophage polarization state, and Drp1–PINK1/Parkin-associated mitochondrial quality control. Nevertheless, several issues require further investigation. The enzymatic activity of ALDH2 should be quantified after liposomal loading, EV fusion, photocrosslinking, storage, and release. Release kinetics should be evaluated under simulated diabetic wound conditions, together with in vivo tracking of tissue retention, spatial distribution, and cell-specific uptake. Hydrated mechanical strength and insertion efficiency should also be examined in ex vivo and in vivo diabetic wound tissues. Moreover, free ALDH2, ALDH2-loaded liposome, and non-fused liposome/EV controls are required to clarify the specific contribution of hybridization. Comparisons or combination studies with standard-of-care treatments, together with infected-wound, large-animal, dose-optimization, and long-term safety studies, would further strengthen the translational relevance of this system. Conclusion In this study, we developed an engineered extracellular vesicle-based GelMA microneedle platform (ALDH2@Lipo/EV-MNs) for diabetic wound repair. This system enabled localized and sustained delivery of ALDH2-loaded hybrid EVs, promoting angiogenesis, reducing oxidative stress, and regulating macrophage polarization. Mechanistically, ALDH2@Lipo/EV improved mitochondrial homeostasis by modulating Drp1 activation and PINK1/Parkin-associated mitophagy processes under diabetic conditions. Overall, this study presents a biomaterial-enabled EV delivery strategy that integrates biological regulation with localized microneedle administration, providing a potential approach for the treatment of chronic diabetic wounds. CRediT authorship contribution statement Shaoyihan Fang: Conceptualization, Investigation, Methodology, Validation, Writing – original draft. Jingjie Min: Formal analysis, Methodology, Writing – original draft. Zhenzhen Chen: Formal analysis, Methodology, Writing – original draft. Shuangyan Li: Methodology, Writing – original draft. Yulu Zhou: Investigation. Jingwen An: Investigation. Zhongyi Chen: Investigation. Dewu Liu: Funding acquisition, Resources, Supervision, Writing – review & editing. Declaration of competing interests The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:Dewu Liu reports financial support was provided by National Natural Science Foundation of China. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements This work was supported by the National Natural Science Foundation of China (82460447, 82160378). Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.mtbio.2026.103555. Supplementary data The following are the supplementary data to this article: Data availability Data will be made available on request. References Associated Data Data Availability Statement Data will be made available on request.