MicroRNA-493-5p engineered exosomes delivered via piezoelectric microneedles for epigenetic modulation of macrophages in diabetic wound healing

TL;DR

A new system uses engineered exosomes and piezoelectric microneedles to improve healing in diabetic wounds by reprogramming immune cells.

Contribution

A novel exosome delivery platform using miR-493-5p and piezoelectric microneedles for epigenetic modulation of macrophages in diabetic wound healing.

Findings

MiR-493-5p promotes M2 macrophage polarization via lactate-induced histone lactylation.

The AgGOx@GelMAZnO MN@EXO@miR system accelerates diabetic wound closure through synergistic immunomodulation.

The system reduces ROS and inflammation while promoting re-epithelialization and neovascularization in vivo.

Abstract

Diabetic foot ulcers, affecting millions worldwide, face impaired healing due to dysregulated macrophage polarization. However, the epigenetic mechanisms underlying aberrant macrophage polarization remain to be elucidated. This study introduces a multifunctional, exosome-based delivery platform that combines miR-493-5p–engineered M2 macrophage exosomes with piezoelectric GelMA microneedles to reprogram macrophage metabolism and epigenetics for diabetic wound healing. Engineered EXO@miR-493-5p are embedded in GelMA microneedles (MN) and delivered via a ZnO piezoelectric substrate with a nanosilver/GOx coating to provide antibacterial and antioxidant benefits. Ultrasound-induced electrostimulation enhances exosome deposition and endocytic uptake, enabling sustained, localized cargo release. Mechanistically, miR-493-5p targets HDAC1 to amplify histone H3K18 lactylation, activating the STAT6 axis and driving metabolic reprogramming toward M2 polarization with upregulation of Arg1. In vitro, EXO@miR-493-5p promote M2 markers and angiogenesis. In vivo, they accelerate wound closure, promote re-epithelialization, collagen deposition, and neovascularization, while reducing ROS and inflammation. The integrated platform offers a translatable, epigenetic-metabolic strategy for chronic diabetic wounds.

Our work presents an Ag/GOx-loaded GelMA/ZnO microneedle system for engineered exosome (EXO@miR-493-5p) delivery. Upon ultrasonic induced piezoelectric effect, the system provides efficient delivery of exosomes in diabetic wounds. This orchestrates immunomodulation by reprogramming macrophages from a pro-inflammatory M1 to a reparative M2 phenotype, concurrently promoting angiogenesis, which synergistically enhances wound healing.Image 1

•MiR-493-5p promotes M2 macrophage polarization via lactate-induced histone lactylation.•MiR-493-5p targets HDAC1 to enhance H3K18 lactylation and activate STAT6 for metabolic reprogramming.•Piezoelectric hydrogel microneedles integrate antibacterial, antioxidant, and electrostimulation for accelerated wound healing.•AgGOx@GelMAZnO MN@EXO@miR system promotes diabetic wound closure through synergistic immunomodulation.

MiR-493-5p promotes M2 macrophage polarization via lactate-induced histone lactylation.

MiR-493-5p targets HDAC1 to enhance H3K18 lactylation and activate STAT6 for metabolic reprogramming.

Piezoelectric hydrogel microneedles integrate antibacterial, antioxidant, and electrostimulation for accelerated wound healing.

AgGOx@GelMAZnO MN@EXO@miR system promotes diabetic wound closure through synergistic immunomodulation.

Figures

Ag/GOx@GelMA/ZnO MN^@EXO@miR^ accelerates DW healing in vivo. (A) Schematic of diabetic wound model and treatment. (B) Photographs of wounds of C57BL/6J mice at different time points after different treatments investigated in this study. (C, D) Quantitative analysis of the relative wound areas at different times (n = 3). (E) Representative H&E staining images of wounds on day 14. Scale bar, 2.5 mm (top) and 200 μm (enlarged). (F) Representative Masson Trichrome staining images of wounds on day 14. Scale bar, 2.5 mm (top) and 200 μm (enlarged). (G-H) Quantitative analysis of epithelial and epid

G–I](#fig8), J), evidencing accelerated angiogenesis. Collectively, these findings establish that Ag/GOx@GelMA/ZnO MN^@EXO@miR^ facilitates diabetic wound healing through coordinated mechanisms: inducing macrophage phenotypic transition, attenuating inflammatory and oxidative stress responses, and promoting vascular regeneration.

Full text

Introduction

1

Diabetic foot ulcers (DFU) refer to wounds on the feet that develop in patients with type 1 or type 2 diabetes, which affect about 18.6 million people worldwide and cause lower extremity amputations in 20% DFU patients of moderate to severe infections [1,2]. Current therapeutic strategies emphasize multidisciplinary approaches, including surgical debridement, offloading, glycemic control, revascularization for peripheral artery disease as well as applying advanced wound therapies, such as bioactive dressings, stem cell applications, and platelet-rich plasma [3]. Despite these interventions, challenges persist due to microbial resistance and impaired tissue repair mechanisms [4]. A critical barrier lies in modulating macrophage polarization during the inflammatory phase [5]. In diabetic microenvironments, macrophages often remain pro-inflammatory (M1-phenotype) rather than transitioning to reparative M2-states, and current attempts to pharmacologically promote M2-polarization face hurdles, including off-target effects and insufficient spatial-temporal control within heterogeneous wound beds [6,7]. Addressing these mechanistic complexities require innovative biomaterial-based delivery systems and precision immunomodulation strategies (see Scheme 1).Scheme 1. Illustration of the fabrication of Ag/GOx@GelMA/ZnO MN^@EXO@miR^ and wound treatment in a diabetic mouse model. The Ag/GOx@GelMA/ZnO MN^@EXO@miR^ system enables sustained release of engineered exosomes encapsulating miR-493-5p to diabetic wounds, synergistically enhancing tissue regeneration through efficient delivery of engineered exosomes, reprogramming macrophage polarization from pro-inflammatory M1 to anti-inflammatory M2 phenotypes, and promoting angiogenesis, thereby orchestrating immunomodulation and reparative functions in the wound microenvironment.Scheme 1

Emerging evidence indicates lactate promotes wound healing by enhancing M2 polarization through histone lactylation (H3K18la), which epigenetically activates pro-reparative genes like Arg1 and VEGF [8]. Tissue injury elevates lactate levels, driving H3K18 lactylation to transcriptionally activate repair genes and promote M2 macrophage phenotypes in wound contexts, as supported by models of injury and metabolic stress [9]. Mechanistically, lactate-driven lactylation synergizes with miRNAs to regulate metabolic reprogramming, thereby modulating macrophage plasticity [10]. Histone lactylation enhances chromatin accessibility and integrates with metabolic pathways, shifting glycolysis-dependent reprogramming to support macrophage adaptation and plasticity during healing responses [11]. This miRNA-lactylation crosstalk reshapes the inflammatory milieu, shifting the wound microenvironment from pro-inflammatory to pro-regenerative [12]. Disrupted lactylation in chronic wounds sustains inflammation, reinforcing this axis as a modifiable target for microenvironmental normalization and accelerated repair. Hence, targeting miRNA-lactic acid axis could amplify M2 polarization and accelerate healing in chronic wounds, offering a dual epigenetic-metabolic therapeutic strategy.

Current strategies for modulating tissue and cellular functions via miRNAs remain confronted by significant bottlenecks, primarily due to their inherent instability in extracellular environments and suboptimal cellular internalization efficiency [13]. Exosome-based miRNA delivery has emerged as a widely adopted approach in tissue engineering, leveraging the natural lipid bilayer of exosomes (30–150 nm vesicles) to preserve miRNA integrity against enzymatic degradation [14]. Critically, exosomes exploit endogenous cellular tropism through surface adhesion molecules, facilitating receptor-mediated membrane fusion and efficient cytosolic delivery of encapsulated miRNAs [15]. This intrinsic targeting mechanism enhances the bioavailability and functional potency of therapeutic miRNAs, enabling precise regulation of gene expression and biological effects in recipient cells [16]. M2 macrophage-derived exosomes (M2-Exos) play a pivotal role in tissue repair by reprogramming pro-inflammatory M1 macrophages toward M2 phenotypes via immunomodulatory miRNAs (such as miR-21-3p and miR-30a-5p), thereby resolving chronic inflammation and enhancing angiogenesis [17]. Engineered exosomes offer distinct advantages for miRNA delivery, including low immunogenicity, intrinsic stability, and precise targeting via surface modifications, enabling efficient suppression of inflammation and promotion of tissue regeneration in diabetic wounds [18]. However, challenges persist in clinical translation, such as insufficient exosomal stability and low fusion efficiency due to membrane lipid heterogeneity, necessitating further optimization enhance cellular uptake and therapeutic consistency [19]. Therefore, the application of exosomes in diabetic wound healing necessitates a rationally designed delivery platform.

Hydrogel microneedles (HMNs) are transdermal drug delivery platforms composed of swellable polymers that form continuous microchannels upon skin insertion, enabling controlled release of therapeutic agents. For exosome delivery, HMNs offer distinct advantages, including high biocompatibility, sustained cargo release, and protection of exosomal integrity while minimizing systemic immunogenicity [16]. Recent advances propose piezoelectric HMNs as an innovative strategy to promote tissue regeneration and enhance exosome delivery efficiency [20,21]. Piezoelectric materials convert mechanical energy (e.g., ultrasound) into electrical energy via the piezoelectric effect, generating localized electric fields; optimizing these fields spatially enhances exosome enrichment in target cells, thereby improving delivery precision and efficiency [22]. This approach synergizes can well address challenges such as low fusion efficiency and exosome instability. Hence, constructing piezoelectric-engineered systems may revolutionize exosome-based therapies for chronic wounds or regenerative medicine by optimizing intracellular delivery dynamics.

Considering that the repair of damaged tissues by EXOs requires sustained-release delivery in situ to achieve good therapeutic effects, this study aimed to develop an MNP-based delivery system effectively delivering engineered M2-exos to treat chronic wound healing. In this study, we developed engineered exosomes derived from M2 macrophages, which were loaded with miR-493-5p to modulate the lactylation-metabolic reprogramming axis. These engineered exosomes were encapsulated within gelatin methacryloyl (GelMA) hydrogel microneedles, with the microneedle tips coated with a nano-silver/glucose oxidase (GOx) composite layer. To further enhance functionality, a ZnO-based piezoelectric substrate was integrated into the microneedle base. This design leverages synergistic antibacterial and antioxidant properties, combined with electrostimulation, to minimize structural degradation of exosomes while promoting their delivery to target cells. Our experimental data established that Ag/GOx@GelMA/ZnO MN^@EXO@miR^ significantly induces macrophage polarization into the M2 phenotype and promotes the angiogenesis process of endothelial cells, which exhibits both anti-inflammatory and regenerative capabilities. And our Ag/GOx@GelMA/ZnO MN^@EXO@miR^ was proved to accelerate wound healing processes in diabetic murine models (Schematic 1).

Materials and methods

2

Materials

2.1

Gelatin methacryloyl (GelMA) was procured from Sigma-Aldrich (St. Louis MO USA; Cat# GMA-100) as the hydrogel matrix for microneedle fabrication. Nano-silver particles (20 nm) and glucose oxidase (GOx) were supplied by Sigma-Aldrich (Cat# 576,832 and G0543, respectively). Zinc oxide (ZnO) nanoparticles (30 nm) were sourced from US Research Nanomaterials (Houston TX USA; Cat# US3090). Exosomes isolated from IL-4-induced M2 macrophages were loaded with miR-493-5p using electroporation. RAW264.7 cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) (Gibco, Thermo Fisher Scientific; Cat# 12,430,054 and 16,140,071). Scanning electron microscopy (SEM, Zeiss Merlin) and laser confocal microscopy (Leica TCS SP8) were utilized for structural characterization, while nanoparticle tracking analysis (NTA, Malvern NanoSight NS300) quantified exosome size distribution. A bicinchoninic acid (BCA) assay kit (Thermo Fisher Scientific; Cat# 23,225) was employed to analyze protein release kinetics.

Exosome isolation and characterization

2.2

Exosomes were isolated from IL-4-induced M2 macrophages (20 ng/mL IL-4; PeproTech, #214-14) via differential ultracentrifugation. Conditioned media were centrifuged at 300×g for 10 min (4 °C) to remove cells, followed by 2000×g for 20 min to eliminate debris, and 100,000×g for 70 min (Optima XE-100, Beckman Coulter) to pellet exosomes. Pellets were resuspended in PBS and purified by 100,000×g centrifugation for 70 min. For engineering, miR-493-5p (RiboBio) was loaded into exosomes via electroporation (Gene Pulser Xcell, Bio-Rad; 350 V, 150 μF). ≈50% of the miRNA was found to be encapsulated within the liposomes. The encapsulation efficiency of miRNA in liposomes was determined spectrophotometrically using a NanoDrop One (Thermo Scientific, USA). Briefly, liposome suspensions were separated from free miRNA by centrifugation, and the supernatant (free miRNA) was quantified at 260 nm. Total miRNA was measured after lysing the liposomes with Triton X-100. Encapsulation efficiency (EE) was calculated as EE (%) = [(Total miRNA − Free miRNA)/Total miRNA] × 100. All measurements were performed in triplicate. Particle size distribution was quantified by nanoparticle tracking analysis (ZetaView PMX 120, Particle Metrix). Morphology was assessed via transmission electron microscopy (HT7800, Hitachi). Exosomal markers (CD63: Abcam, #ab217345; CD9: Abcam, #ab263019) and absence of cellular contaminants (CD206: Proteintech, #60143-1-Ig) were confirmed by western blot.

Microneedle fabrication

2.3

Gelatin methacryloyl (GelMA, EFL, #EFL-GM-60) hydrogel was prepared as 15% (w/v) solution in PBS with 0.25% (w/v) lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP Sigma-Aldrich, #900889) photoinitiator. The solution was pipetted into polydimethylsiloxane molds (10 × 10 array, needle height: 920 μm, base width: 350 μm) and centrifuged at 2000×g for 5 min to remove bubbles. Photocrosslinking used 405 nm UV light (EFL, #EFL-UV400) at 10 mW/cm^2^ for 30 s. For functionalization, microneedle tips were dip-coated in nanosilver/glucose oxidase suspension (Ag/GOx, Sigma-Aldrich, #900521/#G7141) and air-dried. The ZnO piezoelectric substrate (Sigma-Aldrich, #544906) was integrated into the base layer. Engineered exosomes (20 μg/mL in PBS) were loaded into microneedles by vacuum-assisted infusion.

Structural characterization

2.4

Microneedle morphology was analyzed by scanning electron microscopy (Regulus 8100, Hitachi). Mechanical strength was tested via compression assay (Instron 5944, 0.5 mm/min). Exosome distribution was confirmed by confocal microscopy (LSM 900, Zeiss) after DiI labeling (Thermo Fisher, #D3911). Piezoelectric output was measured under ultrasound stimulation (1 MHz, 1 m pulses).

Biocompatibility evaluation

2.5

For hydrogel cytotoxicity assessment, HUVECs were cultured with extracts from Ag/GOx@GelMA/ZnO MN formulations prepared by immersing microneedles in complete DMEM (Gibco, #11995065) supplemented with 10% FBS for 0-72 h at 37 °C. Cellular viability was quantified using CCK-8 assays (Dojindo, #CK04) where 6 × 10^3^ HUVECs per well were exposed to extract solutions for 24 h, followed by 1-h incubation with CCK-8 reagent and absorbance measurement at 450 nm. Live/dead staining employed Calcein AM/PI (Thermo Fisher, #L3224) according to manufacturer protocols with fluorescence imaging using an IX53 microscope (Olympus).

Exosome uptake evaluation

2.6

Endocytosis of engineered exosomes was assessed in RAW264.7 macrophages. Exosomes were labeled with 10 μM DiI (Thermo Fisher, #D3911) at 37 °C for 15 min, purified via ultracentrifugation (100,000×g, 70 min), and incubated with macrophages (20 μg/mL, 24 h). Cells were fixed with 4% paraformaldehyde, stained with DAPI (Thermo Fisher, #D1306) and phalloidin (Thermo Fisher, #A12379), and imaged by confocal microscopy (LSM 900, Zeiss).

Exosome release measure

2.7

For exosome release profiling, DiI-labeled exosomes were incorporated into GelMA prepolymer (15% w/v) before photocrosslinking. Samples were immersed in PBS (800 μL, 37 °C) with daily medium replacement. Released exosome quantities were determined via BCA assay (Thermo Fisher, #23225) using: Cumulative Release (%) = [Σ(C_n_ - Cctrl_n_) × V + (C_n_ - Cctrl_n_) × V_0_]/M × 100 where C_n_ = daily concentration, Cctrl_n_ = control concentration, V = replacement volume (400 μL), V_0_ = initial volume (800 μL), M = total loaded exosomes (50 μg).

Macrophage culture and viability assays

2.8

RAW264.7 cells obtained from the Chinese National Cell Bank were cultured in Dulbecco's Modified Eagle Medium containing 10% fetal bovine serum under standard incubation conditions (37 °C, 5% CO2, 95% humidity). For viability assessment, cells were seeded in 96-well plates at 6 × 10^3^ cells/well and allowed to adhere for 12 h. Following adhesion, experimental treatments including lactate (10 mM concentration) or miR-493-5p-engineered exosomes were administered to respective wells. Post-treatment incubation continued for 24 h under identical culture conditions. Cellular viability was quantitatively evaluated using the CCK-8 assay kit according to the manufacturer's established protocol, with absorbance measurements recorded at 450 nm. All procedures maintained strict adherence to aseptic technique throughout experimental execution.

RAW264.7 transduction

2.9

Transfection of RAW264.7 cells was performed using Lipofectamine LTX reagent (Thermo Fisher Scientific, #A12621) per manufacturer protocol. Plasmids were purchased from Genechem for HDAC1 overexpression. Transfection sequences were shown in Supplementary Table 1.

Flow cytometric analysis

2.10

Macrophage polarization was assessed using antibodies from eBioscience: cells were stained with FITC-anti-CD11b, APC-anti-F4/80, and PE-anti-CD86 at 4 °C for 30 min. Following permeabilization, PerCP/Cyanine5.5-anti-CD206 was applied for an additional 30 min at 4 °C. Cells were washed with cold PBS and analyzed using a flow cytometer.

Transwell assays

2.11

HUVECs (5 × 10^4^ cells per condition) were suspended in serum-depleted DMEM and plated in the upper compartments of 24-well transmembrane chambers (8 μm pore size). The lower chambers received 600 μL of growth medium supplemented with 10% fetal bovine serum as chemoattractant. After 24-h incubation under standard conditions, non-migrated cells were removed from the upper membrane surface. Migrated cells on the lower surface were sequentially processed through PBS washing, fixation in 4% paraformaldehyde solution for 30 min, and staining with 0.5% crystal violet solution for equivalent duration. Migration was quantified by imaging three random microscopic fields per well under bright-field optics and performing automated cell counting with image analysis software.

Scratch-wound healing assays

2.12

HUVECs were seeded into 6-well culture plates (Corning #3516, Corning NY) at 5 × 10^5^ cells/well in complete 1640 medium and incubated until achieving 90-100% confluency (37 °C, 5% CO_2_). Confluent endothelial monolayers established in 6-well plates were mechanically wounded using sterile pipette tips. After removing dislodged cellular material with PBS washes, wound margins were documented at initial (0 h), intermediate (12 h, 24 h), and final (36 h) timepoints using phase-contrast microscopy. Wound closure kinetics were calculated as percentage reduction in wound area relative to baseline measurements using specialized image analysis algorithms.

Tube formation assay

2.13

HUVECs were culture in ECM (Shanghai Zhong Qiao Xin Zhou Biotechnology Co.,Ltd, #ZQ-1304) and then seeded on basement membrane matrix-coated surfaces in 96-well plates (2 × 10^4^ cells per well). Following 45-min initial adhesion, tube network development proceeded for 6 h under physiological culture conditions. Tube formation was quantified by measuring cumulative tube length and branch point frequency in three representative microscopic fields using vascular morphometry software modules.

In vitro cytotoxicity

2.14

The Calcein-AM/PI double staining kit (Sigma-Aldrich) was applied to evaluate the viability of HUVECs cultured in RPMI 1640 with hydrogel extract liquid. Cellular viability following material exposure was evaluated using fluorescent live/dead staining protocols. Cells cultured in extract-containing media were stained at 24-h intervals (24, 48, 72 h) and visualized by fluorescence microscopy to determine cytotoxicity profiles.

Quantitative PCR (qPCR) analysis

2.15

Total RNA was extracted with TRIzol reagent (Invitrogen). Reverse transcription employed HiScript III RT SuperMix (Vazyme, #R323). Quantitative PCR utilized 2× SYBR Green qPCR Mix (Invitrogen, #A25742) with gene-specific primers (Supplementary Table 2). β-Actin served as the endogenous control. Data analysis used the ΔΔCt method.

Western blot analysis

2.16

Cells were lysed in RIPA buffer containing protease/phosphatase inhibitors (NCM Biotech, #P1005). Proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes (Millipore, #ISEQ00010). After blocking with 5% skim milk, membranes were incubated overnight at 4 °C with primary antibodies against: Arg1 (Proteintech, #16001-1-AP), iNOS (Proteintech, #18985-1-AP), IL-1β (Proteintech, #16806-1-AP), PKM2 (ABclonal, #A19053), SLC7A11 (ABclonal, #A24146), HIF-1α (ABclonal, #A11945), and β-actin (Cell Signaling Technology, #4970). HRP-conjugated secondary antibodies (Aspen, #AS1107) were applied for 4h at 4 °C. Detection used the ChemiDoc XRS + System (Bio-Rad).

RNA sequencing analysis

2.17

Total RNA was isolated from RAW264.7 cells using TRIzol reagent (Invitrogen, #15596018) according to manufacturer specifications. RNA integrity (RIN >8.0), concentration, and purity were determined using an Agilent 2100 Bioanalyzer with RNA Nano chips (Agilent Technologies). Sequencing libraries were prepared from 1 μg total RNA using the TruSeq Stranded mRNA Library Prep Kit (Illumina, #20020594). Paired-end sequencing (150 bp) was performed on an Illumina NovaSeq 6000 platform (Macrogen). Differential gene expression analysis employed the DESeq2 package (v1.38.3) with thresholds of |log2FC| ≥ 1 and adjusted p < 0.05. Functional enrichment of GO terms was conducted using DAVID Bioinformatics Resources (v6.8). Heatmap visualizations were generated with Morpheus matrix analysis tools.

Measurement of OCR and ECAR

2.18

Cellular bioenergetics were assessed using Seahorse XFp Analyzers (Agilent Technologies). For oxygen consumption rate (OCR) measurements, RAW264.7 cells transfected with agomiR-493-5p or controls were seeded in XFp plates and analyzed in XF Base Medium (Agilent, #103334-100) supplemented with 25 mM glucose and 4 mM glutamine. The Mitochondrial Stress Test Kit (Agilent, #103010-100) was employed with sequential injections of 1 μM oligomycin, 1 μM FCCP, and 0.5 μM rotenone/antimycin A. For extracellular acidification rate (ECAR), cells were equilibrated in glutamine-supplemented base medium without glucose, with Glycolysis Stress Test Kit (Agilent, #103020-100) reagents (10 mM glucose, 1 μM oligomycin, 50 mM 2-DG) injected sequentially. All values were normalized to cellular protein content quantified via BCA assay. Data processing utilized Wave Desktop Software (Agilent, v2.6.0.31).

Diabetic mouse model generation

2.19

All experimental protocols were approved by the Wenzhou Institute of University of Chinese Academy of Sciences (WIUCAS24121902). Male C57BL/6J mice (6-week-old) were fed a high-fat diet for 4 weeks before receiving intraperitoneal streptozotocin injections (STZ; Sigma-Aldrich, #S0130; 40 mg/kg daily for 7 consecutive days). Diabetes induction was confirmed by fasting blood glucose >16.7 mmol/L in three sequential measurements.

In vivo diabetic wound healing assessment

2.20

STZ-induced diabetic mice were anesthetized with pentobarbital sodium (50 mg/kg; Sigma-Aldrich). Full-thickness excisional wounds (1.0 × 1.0 cm) were created on the dorsum. Mice received topical applications (200 μL) of experimental formulations on days 0, 3, 5, 7, 10, and 14 post-wounding (n = 8 per group). Wounds were covered with transparent dressings (Tegaderm™ Film, 3M). Wound areas were documented using digital calipers and quantified via ImageJ using: Closure (%) = [(A_0_ - A_n_)/A_0_] × 100, where A_0_ represents initial wound area and A_n_ denotes area at timepoint n.

Histological and immunostaining analyses

2.21

Wound tissues were fixed in 4% paraformaldehyde, paraffin-embedded, and sectioned at 4 μm thickness. Sections underwent Hematoxylin & Eosin (H&E) and Masson's trichrome staining following standard protocols, with whole-slide imaging performed using a PANNORAMIC Flash scanner (3DHISTECH). For angiogenesis assessment, day 14 sections underwent antigen retrieval in citrate buffer (pH 6.0), blocking with 10% goat serum (30 min), and overnight incubation at 4 °C with anti-CD31 antibody (1:100; Abcam, #ab28364). After PBS washes, sections were incubated with HRP-conjugated secondary antibody (Aspen, #AS1107) for 1 h at 25 °C, developed with DAB substrate, and counterstained with hematoxylin. CD31^+^ microvessels (2-10 μm diameter) were quantified across five random fields per sample using an Olympus IX53 microscope. Macrophage polarization was evaluated via immunohistochemistry for Arg1 (Proteintech, #16001-1-AP) and iNOS (Proteintech, #18985-1-AP) using identical antigen retrieval/blocking conditions, with visualization employing Olympus CX31 and Zeiss fluorescence microscopes.

Small animal Doppler

2.22

On postoperative day 10, wound perfusion was quantitatively assessed using a laser speckle contrast imaging (LSCI) system (PSI-ZR PeriCam system, PERIMED Ltd, Stockholm, Sweden). Perfusion images were acquired under standardized conditions with an 785-nm near-infrared laser. Blood perfusion was calculated and expressed as perfusion units. The mean perfusion unit (MPU) ratio was determined by correlating the MPU within the wound area (ROI-1) to the MPU of adjacent uninjured tissue (ROI-Total), consistent with established methodology described in Results Section 2.8 and Fig. 8F.

Quantification of exosome internalization under ultrasound-induced electrostimulation

2.23

RAW264.7 macrophages (2 × 10^5^ cells/well) were seeded in 12-well plates and cultured overnight in DMEM containing 10% FBS. Exosomes engineered with miR-493-5p were labeled with DiI (10 μM, 37 °C, 15 min), washed, and re-isolated by ultracentrifugation (100,000×g, 70 min). GelMA MN@EXO@miR and Ag/GOx@GelMA/ZnO MN@EXO@miR patches were loaded with equivalent amounts of DiI-EXO (20 μg per patch). MN patches were gently applied to the cell monolayer for 10 min to allow microneedle insertion, followed by incubation for 4 h. For the ultrasound group, Ag/GOx@GelMA/ZnO MN@EXO@miR-treated cells were exposed to pulsed ultrasound (1 MHz, 1 mms pulse width, 0.5 s interval, 0.5 W/cm^2^) for 5 min immediately after MN application. Cells were then washed with cold PBS, detached with EDTA‐free trypsin, and resuspended in PBS containing 2% FBS. DiI fluorescence was quantified using a flow cytometer (488/561-nm excitation; PE or appropriate channel). Mean fluorescence intensity of Dil-positive cells was used to compare exosome uptake among groups.

Statistical analysis

2.24

Quantitative data are presented as mean ± SD. Statistical significance between groups was determined using Student's two-sided t-test, with significance thresholds defined as P value < 0.05; ∗∗P value < 0.01; ∗∗∗P value < 0.001 per; ∗∗∗∗P value < 0.0001 per group by unpaired t-test. This analytical approach aligns with all figure legends and methodological descriptions within the manuscript.

Results and discussion

3

Lactate mediates M2 polarization of macrophages and accelerates wound repair

3.1

Lactate (LA), a key metabolite derived from glycolysis, regulates macrophage polarization by promoting the pro-inflammatory M1 phenotype and disrupting the shift to the pro-healing M2 phenotype in diabetic wounds [23]. Modulating lactate metabolism can restore M2 polarization, improving the wound microenvironment and enhancing tissue repair in diabetes by influencing pathways like glycolysis and lactylation [24]. Targeting lactate thus offers a novel therapeutic approach for diabetic wound management.

To demonstrate the role of lactate in macrophage polarization and wound healing processes, we established a concentration gradient of LA to treat RAW 264.7 cells. Following 24-h treatment, macrophage polarization was assessed using western blotting, immunofluorescence, and flow cytometry. And we found that 10 mM LA showed the best efficiency in inducing M2 polarization in macrophages (Fig. S1). Subsequently, we further proved the time-dependent induction of M2 macrophage polarization by 10 mM lactate through western blotting, immunofluorescence, and flow cytometry. The extent of M2 polarization increased progressively with prolonged induction time (Fig. 1A–F). To further explore the role of LA in wound healing process, 10 mM LA was intravenously administered on days 0, 3 and 7 post-wounding, showing that the wound closure rate was promoted compared with blank group (Fig. 1G–I). Histological analysis of skin specimens showed that wounds treated with 10 mM LA exhibited complete re-epithelialization and dermal thickening compared with blank group (Fig. S2). Masson's trichrome staining further revealed marked elevation in collagen deposition within the 10 mM LA group (Fig. S2). Then, immunofluorescence staining was performed on 14-day wound samples. Compared to the blank group, a decrease in the M1 macrophage marker iNOS and an increase in the M2 macrophage marker Arg1 were observed in 10 mM LA -treated wounds, indicating that lactate improved macrophage polarization during wound healing (Fig. 1J and K). These results indicate that LA improves macrophage polarization during the wound healing process. Investigating the metabolic and molecular mechanisms underlying lactate-induced M2 macrophage polarization may provide therapeutic targets for diabetic wound healing.Fig. 1Lactic acid mediates M2 polarization of macrophages compared to the sham group in dose dependent manner. (A, B) Western blot was used to assess the level of Arg1 and iNOS on RAW 264.7 cells from the Blank group and 10 mM LA group. (C, D) Immunofluorescence staining and quantitative analysis of Arg1 and iNOS in RAW264.7 cells from the Blank group and 10 mM LA group; Scale bar: 100 μm. (E, F) The percentage of CD86^+^ and CD206+ cells among the CD11b + F4/80+ RAW 264.7 cells from the Blank group and 10 mM LA group. (G-I) Photographs and quantitative analysis of wounds of C57BL/6J mice at different time points a from the Blank group and 10 mM LA group; Scale bar: 5 mm. (J, K) Immunofluorescence staining and quantitative analysis of at 7 days post-wounded using anti-Arg1and anti-iNOS; Scale bar: 200 μm. Data are presented as mean ± SD and statistical significance was analyzed via Student's two-sided t-test. ∗P value < 0.05; ∗∗P value < 0.01; ∗∗∗P value < 0.001 per; ∗∗∗∗P value < 0.0001 per group by unpaired t-test.Fig. 1

miR-493-5p is enriched in LA-treated microphage and is responsible for M2 differentiation

3.2

The precise roles of microRNAs (miRNAs) in mediating lactate-dependent regulation of macrophage polarization and protein lactylation remain largely undefined. Recent studies have shown that lactate can influence macrophage polarization through multiple mechanisms, including modulating miRNA expression and inducing histone lactylation modifications [25,26]. Targeting miRNAs to modulate macrophage polarization and lactylation signaling represents a novel therapeutic strategy for promoting diabetic wound repair by resolving sustained inflammation and impaired phenotypic switching [27]. This approach holds promise for rectifying the dysregulated macrophage functions central to diabetic wound chronicity.

Previous research has demonstrated that lactate treatment can significantly alter miRNA expression patterns in macrophages, affecting their polarization state [28]. To define the expression profile associated with LA treatment, we conducted an RNA sequencing analysis of miRNAs of RAW264.7 cells from the Blank group and 10 mM LA group. Compared to the Blank group, 43 and 71 miRNAs were significantly upregulated and downregulated in RAW264.7 cells post LA treatment, respectively. Among them, precursor miR-493-5p was the most significantly expressed gene (Fig. 2A and B). And our qRT-PCR result also verified that miR-493-5p was significantly upregulated under the treatment of LA (Fig. 2C). Pervious studies have shown that miR-493-5p was a negative regulator of excessive inflammation. miR-493-5p directs anti-inflammatory action through targeting TAB2 in an LPS model, attenuating NF-κB-driven inflammatory injury [29]. miR-493-5p also participate in the biology process of re-epithelialization, granulation tissue formation, and remodeling [30]. Hence, we decided to further investigate miR-493-5p.Fig. 2miR-493-5p is enriched in LA-treated microphage and is responsible for M2 differentiation. (A) small RNA-seq analysis and volcano plot showing genes or miRNAs with a cut-off fold-change of ≥4 or ≤ −4 and a p value of <0.01. (C) qRT-PCR detection of the relative abundance of miR-493-5p in RAW264.7 cells treated with or without LA. (D) KEGG pathway enrichment analysis of the DEGs in RAW264.7 cells treated with antagomiR-493-5p transfection. (E) GO classification of genes categorized by cellular component, molecular function, and biological process in RAW264.7 cells treated with antagomiR-493-5p transfection. (F) Selected GSEA enrichment score curve of histone modification genes in RAW264.7 cells treated with antagomiR-493-5p transfection. (G) Dual-luciferase reporter assay to test miR-493-5p binding to wild-type and mutated HDAC1 3′ UTR. WT, wild-type; Mut, mutated; miR-NC, negative control. (H) Venn diagram showing the miR-493-5p targets from miRWalk, TargetScan, miRDB database and Lactylation-related genes. (I) qRT-PCR analysis of suspected miR-493-5p target (HDAC1). (J) Western blot of HDAC1 expression level in RAW264.7 cells with indicated treatments. (K) qRT-PCR analysis and western blot of HDAC1 expression in raw264.7 transfected with NC plasmid or HDAC1 plasmid. (L) Western blot was used to assess the level of Arg1 and iNOS on RAW 264.7 cells with indicated treatments. (M) The percentage of CD86^+^ and CD206+ cells among the CD11b + F4/80+ RAW 264.7 cells with indicated treatments. (N) Immunofluorescence staining and quantitative analysis of Arg1 and iNOS in RAW264.7 cells from the Blank group and 10 mM LA group; Scale bar: 100 μm. (O) Quantitative analysis of western blot results. (P) Quantitative analysis of flow cytometry. (Q) Quantitative analysis of immunofluorescence results. Data are presented as mean ± SD and statistical significance was analyzed via Student's two-sided t-test. ∗P value < 0.05; ∗∗P value < 0.01; ∗∗∗P value < 0.001 per; ∗∗∗∗P value < 0.0001 per group by unpaired t-test.Fig. 2

We collected RAW264.7 cells, including those treated with antagomiR-493-5p transfection, for transcriptome sequencing. Based on transcriptome sequencing, we demonstrated that miR-493-5p transfection enhances histone modification (GO analysis, Fig. 2D) and chromatin remodeling pathways (KEGG analysis, Fig. 2E), with GESA revealing significant enrichment of histone modification genes, consistent with augmented epigenetic regulation (Fig. 2F).

To identify novel downstream effectors, computational target prediction platforms (miRWalk, TargetScan, and miRDB) were integrated with histone lactylation-associated gene sets for miR-493-5p target screening, leading to the selection of HDAC1 (Fig. 2H). The HDAC1 3′ UTR sequence was screened in the IntaRNA database (https://rna.informatik.uni-freiburg.de/IntaRNA/) to identify the miR-493-5p binding site, in which the identical sequence in the miR-493-5p seed region binds to HDAC1. Subsequently, we cloned the 3′UTRs of the mutant target in a dual luciferase assay system to further determine whether miR-493-5p targeted HDAC1 directly (Fig. 2G).Subsequent validation in RAW264.7 cells transfected with antagomiR-493-5p or antagomiR-NC involved quantitative assessment of HDAC1 expression changes via western blotting and qRT-PCR. Results demonstrated that HDAC1—a key regulator of histone delactylation—exhibited downregulation in response to LA treatment. Conversely, antagomiR-493-5p overexpression significantly restored HDAC1 expression in RAW 264.7 cells (Fig. 2I and J). To further elucidate whether miR-493-5p regulates macrophage polarization phenotypes through HDAC1, RAW264.7 cells were transfected with NC plasmid or HDAC1 plasmid, respectively. Then, macrophage polarization was assessed using western blotting, immunofluorescence, and flow cytometry. We validated the transduction efficiency of HDAC1 plasmids through qRT-PCR and WB (Fig. 2K, L). Results from western blotting manifested that Arg-1 expression was dramatically lessened by overexpressing HDAC1. Instead, the expression of iNOS exhibited an increase after overexpressing HDAC1 (Fig. 2L, O). Flow cytometry detection demonstrated that miR-493-5p-overexpression led to M2 polarization of RAW264.7 cells while overexpressing HDAC1 resulted in a higher ratio of CD86^+^ M1 phenotype and a lower proportion of CD206+ M2 phenotype in miR-493-5p-overexpression RAW264.7 cells (Fig. 2M, P). Immunofluorescence results similarly demonstrated that miR-493-5p induced M2 polarization of macrophages, whereas this effect was reversed upon HDAC1 overexpression (Fig. 2N, Q). Collectively, our experimental data demonstrate that LA-treated macrophages exhibit upregulated expression of miR-493-5p. This microRNA promotes M2 polarization of macrophages, an effect potentially mediated via HDAC1 downstream signaling pathways. miR-493-5p might serve as a master modulator of epigenetic plasticity, potentially through direct or indirect targeting of transcriptional co-activators.

miR-493-5p targets HDAC1 to drive M2 macrophage polarization via Histone Lactylation-Dependent STAT6 activation

3.3

To identify downstream mechanism of miR-493-5p in regulating M2 polarization of macrophages, we collected RAW264.7 cells under the treatment of NC plasmid or HDAC1 plasmid transfection for transcriptome sequencing. Based on transcriptome sequencing, we demonstrated that cellular lipid metabolic process was significantly attenuated in HDAC1 plasmid transfected cells via GO analysis (Fig. 3A). This observation aligns with emerging evidence that metabolic shifts are crucial for macrophage functional plasticity [31]. KEGG analysis revealed suppression of the JAK-STAT signaling pathway following HDAC1 overexpression (Fig. 3B). The JAK-STAT pathway is known to be central to cytokine-mediated polarization signals, particularly through STAT6 activation during M2 polarization [12]. Furthermore, GSEA confirmed enrichment of STAT6 pathway genes in the control group, highlighting its pivotal role in this regulatory mechanism (Fig. 3C). Notably, STAT6 has been shown to directly regulate genes involved in lipid metabolism during alternative activation of macrophages [32]. These findings collectively suggest that HDAC1, acting downstream of miR-493-5p, likely modulates macrophage polarization by activating STAT6 signaling pathway.Fig. 3miR-493-5P Targets HDAC1 to Drive M2 Macrophage Polarization via Histone Lactylation-Dependent STAT6 Activation. (A) GO classification of genes categorized by cellular component, molecular function, and biological process in RAW264.7 cells treated with HDAC1 plasmid transfection. (B) KEGG pathway enrichment analysis of the DEGs in RAW264.7 cells treated with HDAC1 plasmid transfection. (C) Selected GSEA enrichment score curve of glycolytic metabolism genes in RAW264.7 cells treated with HDAC1 plasmid transfection. (D) The fluorescence intensity ratio of STAT6 immunostaining between cell nuclei and cytosol was analyzed. (E, F) Western blots and analysis of STAT6, Lamin B and GAPDH (n = 3). (G)Western blotting analysis shows the levels of Pan Kla and H3K18la in RAW264.7 cells with indicated treatments. (H, I) Western blot was used to assess the level of STAT6 pathway proteins. (J, K) Seahorse analysis of oxygen consumption rates (OCR) (n = 3) and extracellular acidification rate (ECAR) (n = 3) in control and HDAC1 overexpression RAW264.7 cells after agomiR-493-5p transfection. Data are presented as mean ± SD and statistical significance was analyzed via Student's two-sided t-test. ∗P value < 0.05; ∗∗P value < 0.01; ∗∗∗P value < 0.001 per; ∗∗∗∗P value < 0.0001 per group by unpaired t-test.Fig. 3

Given that HDAC1 regulates histone delactylation—a process critically involved in nuclear translocation of key transcription factors and chromatin accessibility—we further investigated its role in STAT6 subcellular localization. To address this, confocal microscopy and nuclear/cytoplasmic fractionation followed by Western blotting were employed to assess STAT6 nuclear translocation dynamics. Laser confocal imaging showed that agomiR-493-5p transfection resulted in an increase of the nuclear location of STAT6, and HDAC1 overexpression suppressed the nuclear location of STAT6 in macrophages (Fig. 3D). Similar results were observed by nuclear/cytoplasmic fractionation (Fig. 3E and F). Emerging evidence indicates that LA triggers epigenetic reprogramming via histone lactylation, driving enhanced M2 macrophage polarization—a process modulated by HDAC1 [33,34]. In agomiR-493-5p-transfected macrophages, western blotting detected markedly elevated histone lactylation. Conversely, HDAC1 overexpression decreased H3K18la modification levels (Fig. 3G). To further investigate whether miR-493-5p regulates the STAT6 pathway via HDAC1, we performed western blotting analysis on the aforementioned treatment groups. The results demonstrated that transfection with agomiR-493-5p significantly increased the ratio of phosphorylated STAT6 to total STAT6 (pSTAT6/STAT6), accompanied by elevated protein levels of PPARγ and CPT1A—key downstream effectors associated with lipid metabolic reprogramming. These findings suggest that miR-493-5p promotes M2 macrophage polarization by activating the STAT6 signaling pathway. Notably, the activation of this pathway was attenuated upon transfection with HDAC1 plasmid, further supporting the dependency of miR-493-5p-mediated STAT6 regulation on HDAC1 (Fig. 3H and I). We also performed a ChIP-PCR assay to confirm H3K18la presence at STAT6 promoter regions across various treatments. Our result revealed that RAW264.7 cells treated of agomiR-493-5p transfection exhibited significantly increased H3K18la occupancy at STAT6 promoter regions compared to the control group, which could be reversed by the overexpression of HDAC1 (Fig. S3A). To further elucidate the functional necessity of STAT6, we performed STAT6 silencing in macrophages via siRNA in the presence of agomiR-493-5p. As shown in Fig. S3B–E, STAT6 depletion markedly suppressed the agomiR-493-5p-induced downregulation of iNOS and upregulation of Arg1, preventing the miR-493-5p-driven phenotypic shift from M1 to M2 macrophages, as evidenced by both Western blot and immunofluorescence analyses. Furthermore, STAT6 knockdown substantially reversed the miR-493-5p-mediated increases in p-STAT6/STAT6 ratio, CPT1A, PPARγ, and H3K18la protein levels (Fig. S3F–G), with corresponding changes in mRNA expression profiles (Fig. S3H). These findings collectively indicate that STAT6 activation is requisite for miR-493-5p/HDAC1-dependent epigenetic modifications and metabolic reprogramming of macrophages, establishing STAT6 as a pivotal hub integrating miRNA signaling with both chromatin remodeling and metabolic reprogramming pathways.

Next, the Seahorse Extracellular Flux analyzer was used to analyze bioenergetic function in RAW 264.7 cells. In agreement with RAW 264.7 cells polarization results in Fig. 2, we observed a significant increase in the oxygen consumption rate (OCR) and ATP production in RAW264.7 cells under the treatment of agomiR-493-5p transfection. However, the impact of agomiR-493-5p transfection was reversed in HDAC1 overexpression group (Fig. 3J). Conversely, agomiR-493-5p transfection decreased the extracellular acidification rate (ECAR) and glycolytic capacity in RAW264.7 cells, while HDAC1 overexpression inhibited the miR-493-5p-induced decrease in ECAR and glycolytic capacity (Fig. 3K).

We further test the functional necessity of direct link between this metabolic shift and the epigenetic modifier lactate. We applied lactate assay kit quantified intracellular lactate concentrations. As shown in Fig. S4A, EXO treatment modestly decreased lactate levels compared with Blank, whereas EXO@miR-493-5p induced a significantly greater reduction in intracellular lactate. Then, we evaluated proteins involved in lactate production and transport. Western blot analysis demonstrated that LDHA and the lactate transporters MCT4 were downregulated upon EXO treatment, with the most pronounced decreases observed in the EXO@miR-493-5p group. And MCT1 were upregulated upon EXO treatment, with the most pronounced increases observed in the EXO@miR-493-5p group (Fig. S4B and C). This pattern is consistent with a switch from glycolytic M1 to oxidative M2 metabolism. MCT4 is a high‐affinity, predominantly efflux transporter induced under high glycolytic flux to prevent intracellular acidification, typical of M1 macrophages [35]. In contrast, MCT1 is a bidirectional, lower-affinity transporter that maintains lactate homeostasis and can import extracellular lactate under moderate intracellular concentrations, supporting histone lactylation–dependent M2 gene programs [33]. LDHA/MCT4 downregulation limits excessive glycolytic lactate, while MCT1 upregulation allows controlled lactate import to sustain the intracellular pool required for epigenetic lactylation during M2 polarization [26].

These results suggest that HDAC1-mediated downstream STAT6 signaling pathway is critical for the miR-493-5p-induced metabolism shift in macrophages.

Fabrication and characterization of engineered M2-EXOs

3.4

Previous experimental evidence indicates that miR-493-5p orchestrates macrophage metabolic reprogramming by modulating HDAC1-mediated histone delactylation—a newly identified post-translational modification wherein lactate covalently binds to histone lysine residues, altering gene transcription profiles. This lactate-driven epigenetic modification has been shown to regulate macrophage polarization through transcriptional activation of reparative genes [33]. The epigenetic regulation subsequently activates the STAT6 signaling pathway (a critical mediator of anti-inflammatory responses and macrophage polarization), thereby enhancing M2-like phenotypic switching in diabetic wounds. Collectively, these actions demonstrate miR-493-5p′s potential to ameliorate chronic inflammation and promote tissue regeneration in diabetic wound healing.

However, the inherent structural instability of miRNAs—owing to rapid degradation by serum nucleases and inefficient cellular internalization—severely limits their bioavailability and target engagement in vivo. This limitation is particularly problematic in diabetic wounds, where excessive inflammation further compromises miRNA stability [36]. To address these delivery challenges and functionally validate miR-493-5p′s efficacy in wound repair, we engineered an exosome (EXO)-based delivery platform. Specifically, exosomes (40–150 nm vesicles) were isolated from IL-4-induced M2 macrophages via ultracentrifugation, leveraging their intrinsic biocompatibility, low immunogenicity, and natural tropism toward inflammatory sites. The choice of M2 macrophage-derived exosomes was strategic, as they inherently carry miRNAs that promote tissue repair and have demonstrated efficacy in diabetic wound models [37]. Subsequently, electroporation was employed to encapsulate miR-493-5p into exosomes (Fig. 4A). This strategy exploits exosomes' ability to protect payloads from enzymatic degradation and facilitate endosomal escape, thereby ensuring precise intracellular delivery of miR-493-5p to wound-resident cells. Then M2-Exos and Engineered M2-Exos were characterized for morphology, size, and surface marker expression. TEM imaging revealed that both M2-Exos and Exo@miR-493-5p displayed saucer-shaped or hemispherical morphology with a concave side (Fig. 4B). Nanoparticle tracking analysis (NTA) showed an average particle size of 142.1 nm and a concentration of 4.0 E+6 particles mL^−1^ of M2 EXO, and an average particle size of 145.3 nm and a concentration of 4.3 E+6 particles mL^−1^ of Exo@miR-493-5p (Fig. 4C). To test whether the exosomes could be phagocytized into cells to exert their effects, uptaketest of the two types of exosomes was applied in HUVEC and RAW264.7 cells observed via laser confocal microscopy, which showed that both exosome types could be well (Fig. 4D and E). Western blotting confirmed the presence of exosome markers (CD63 and CD9) and the M2 Mϕ phenotypic marker CD206 was absent, while the M2 Mϕ showed CD206 without presence of exosome markers (CD63 and CD9) (Fig. 4F). Subsequently, robust validation of miR-493-5p loading into engineered exosomes was confirmed via qRT-PCR (Fig. 4G). Then, RAW264.7 cells were treated with two distinct exosome groups: non-engineered exosomes (control group) and miR-493-5p-loaded exosomes (experimental group), both at a concentration of 20 μg/mL for 24 h to ensure cellular uptake. Total RNA was isolated from treated cells and qRT-PCR analysis revealed a significant increase in intracellular miR-493-5p levels exclusively in cells treated with engineered exosomes, confirming the delivery ability of the therapeutic miRNA (Fig. 4H). This result aligns with the established role of exosomes in protecting RNA cargos from degradation and facilitating endosomal escape, thereby enhancing miRNA bioavailability in recipient cells. Together, in this part we applied exosome delivery system to overcomes miRNA instability limitations, with qRT-PCR confirming efficient intracellular transfer of miR-493-5p, demonstrating its potential for enhanced diabetic wound therapy.Fig. 4Fabrication and characterization of Engineered M2-EXO. (A) Flowchart of the fabrication of engineered M2-EXO. (B) Nonengineered M2 exosome and engineered miR-N20 exosome morphology and diameter distribution, as detected by TEM. (C) Nanoparticle tracking analysis (NTA) of the range of particle size distribution of Nonengineered M2 exosome and engineered miR-493-5p exosome. (D, E) Exosome internalization by HUVECs and RAW264.7 cells. (F) Exosome surface markers detected by western blotting. (G) qRT-PCR analysis of miR-493-5p content in exosomes. (H) qRT-PCR analysis of miR-493-5p content in RAW264.7 cells under EXO or EXO@miR-493-5p treatment. Data are presented as mean ± SD and statistical significance was analyzed via Student's two-sided t-test. ∗P value < 0.05; ∗∗P value < 0.01; ∗∗∗P value < 0.001 per; ∗∗∗∗P value < 0.0001 per group by unpaired t-test.Fig. 4

Multiple functions of engineered miR-493-5p exosomes

3.5

To determine the efficacy of the EXO@miR-493-5p in diabetic wound healing, we first performed in vitro experiments. Western blotting results showed that EXO@miR-493-5p treatment resulted in an increased expression of cyclin D1 and cyclin D3 participating in cell cycle transition in HUVECs (Fig. 5A). And cck8 test showed that engineered miR-493-5p exosomes promoted cell proliferation (Fig. 5B). Results from the wound scratch and transwell migration assay also indicated that Exo@miR-493-5p treatment significantly restrained the mobilization ability of HUVECs. Exo@miR-493-5p markedly accelerated the rate of cell scratch closure and enhanced HUVECs mobilization to the lower chamber when compared to the Exo group (Fig. 5C–F). Similarly, we observed that the total tube length and the number of tube branch points formed by HUVECs was enhanced when cocultured with Exo@miR-493-5p (Fig. 5G–I). These results implyed the importance of proliferation facilitation in pro-angiogenic mechanisms of this phytochemical.Fig. 5Multiple functions of engineered miR-N20 exosomes. (A)Western blotting and quantitative analysis on HUVECs with indicated treatments. (B) CCK8 to test the proliferation of HUVECs cells with indicated treatments. (C,D) Scratch test and quantitative analysis to evaluate migration of HUVECs. (E, F) Transwell test and quantitative analysis on HUVECs with indicated treatments; Scale bar: 250 μm. (G-I)Tube formation test and quantitative analysis on HUVECs with indicated treatments; Scale bar: 100 μm. (J) Western blot was used to assess the level of Arg1 and iNOS on RAW 264.7 cells with indicated treatments. (K) Immunofluorescence staining and quantitative analysis of Arg1 and iNOS in RAW264.7 cells from the Blank group and 10 mM LA group; Scale bar: 100 μm. (L) The percentage of CD86^+^ and CD206+ cells among the CD11b + F4/80+ RAW 264.7 cells with indicated treatments. (M) Quantitative analysis of western blot results. (N) Quantitative analysis of immunofluorescence results. (O) Quantitative analysis of flow cytometry. Data are presented as mean ± SD and statistical significance was analyzed via Student's two-sided t-test. ∗P value < 0.05; ∗∗P value < 0.01; ∗∗∗P value < 0.001 per; ∗∗∗∗P value < 0.0001 per group by unpaired t-test.Fig. 5

Then, macrophage polarization was assessed using western blotting, immunofluorescence, and flow cytometry. Results from western blotting manifested that Arg-1 expression exhibited an increasing and the expression of iNOS was dramatically lessened by EXO@miR-493-5p (Fig. 5J–M). Immunofluorescence results similarly demonstrated that EXO@miR-493-5 induced M2 polarization of macrophages (Fig. 5K–N). Flow cytometry detection demonstrated that Exo@miR-493-5p treatment led to a higher ratio of CD206+ M2 phenotype and a lower proportion of CD86^+^ M1 phenotype in RAW264.7 cells (Fig. 5L–O). Additionally, western blot analysis confirmed that EXO@miR-493-5 enhanced H3K18 lactylation and activated the STAT6 signaling pathway (Fig. S5).

Collectively, our in vitro evidence demonstrates that EXO@miR-493-5p orchestrates dual therapeutic mechanisms: potently enhancing endothelial proliferation, migration, and tubulogenesis to drive angiogenesis while concurrently reprogramming macrophages toward pro-healing M2 polarization to resolve inflammation, thereby substantiating its translational potential for diabetic wound therapy.

A key limitation of our in vitro work is the exclusive use of the transformed RAW264.7 monocytic/macrophage cell line rather than primary BMDMs, which may differ in signaling complexity, basal gene expression, and the dynamic range of M2 markers such as Ym1 [38]. Nonetheless, RAW264.7 cells represent a well-established model for IL-4–driven M2 polarization and provided the scalability and reproducibility required for multi-omics, metabolic, and repeated molecular assays [39]. To strengthen biological validity, M2-like polarization was verified using multiple complementary markers (Arg1/iNOS, CD206/CD86) rather than a single readout, and the central miR-493-5p–HDAC1–H3K18la–STAT6 axis identified in vitro was corroborated, supporting the effects of miR-493-5p and EXO^@miR−493−5p^ despite the inherent limitations of this cell line–based system.

Preparation and characterization of ag/gox@gelma/ZnO MN@EXO@miR

3.6

To achieve localized delivery of exosomes to diabetic wounds, we introduced biocompatible gelatin methacryloyl (GelMA) hydrogel microneedles. GelMA has emerged as an ideal biomaterial for transdermal drug delivery due to its tunable mechanical properties, biocompatibility, and ability to sustain release of therapeutic cargo [22,40]. The fabrication process of these microneedles is illustrated in the accompanying figure (Fig. 6A). Given that elevated reactive oxygen species (ROS) and bacterial presence in the diabetic wound microenvironment can compromise exosome integrity, we applied a surface coating of nanosilver and glucose oxidase (GOx) to the microneedles. This dual-functional coating synergistically addresses two major pathological features of diabetic wounds: persistent infection and oxidative stress [41], which mitigates ROS- and bacteria-induced damage to the exosomes. Furthermore, current literature demonstrates that piezoelectric materials can enhance tissue repair by generating microvoltages through mechanical stimulation [42]. Concurrently, studies indicate that low-intensity electrical currents can improve exosome delivery efficiency [43]. Building on this evidence, we incorporated a ZnO piezoelectric base layer into the microneedle design to leverage these effects. The integration of piezoelectric components represents an innovative approach to combine mechanical and electrical cues for enhanced wound therapy [44]. Morphological evaluation disclosed the intact tip shape and sharp point of each needle, orderly aligned on the backing layer in a 10 × 10 array, with a needle height, base width and tip-to-tip spacing of 920, 350 and 800 μm respectively (Fig. 6B, C, D). 3D reconstructive fluorescence images uncovered that Dil-probed EXO was uniformly distributed in the hydrogel MN (Fig. 6E). The viscoelastic properties of the skin reduce the penetration capability of microneedles (MNs), which therefore must possess sufficient mechanical strength to effectively penetrate the skin. Studies have indicated that the insertion force required for microneedles to penetrate the skin barrier is approximately 0.098 N per needle, and the mechanical strength of the microneedles must exceed this value to effectively penetrate the stratum corneum [45]. The mechanical strength of obtained MN patch was determined and the force curve showed that the fracture force was approximately 2 N/needle, signifying outstanding mechanical characters sufficient to pass through skin barrier without breaking (Fig. 6F). The penetration performance was confirmed by penetration testing carried out on mouse skin (Fig. S6A). Benefiting from the sharp conical geometry and appropriate length of each microneedle, the delivery vehicle could traverse the stratum corneum and reach the dermis, as demonstrated by histological examination (Fig. S6B). The structure and morphological changes of individual needle tips were observed under SEM. The needle shows a porous structure and volume expansion, with a large amount of EXO, which showed good uniformity (Fig. 6G). Energy-dispersive X-ray spectroscopy (EDS) analysis confirmed that the microneedle matrix is primarily composed of organic material containing carbon (C), nitrogen (N), and oxygen (O), consistent with the expected composition of gelatin methacrylate (GelMA) (Fig. 6H). With respect to the degradation and release property, we observed similarly cumulative release of EXO from these two MN systems, and incorporation of Ag/GOx functional coatings and ZnO-reinforced substrates did not significantly alter the degradation and release properties (Fig. 6I and J). To measure the electrical output of the piezoelectric scaffold upon mechanical stimulation, the drive signal for US (frequency 1 MHz) was set as a pulse width of 1 mms and a pulse interval of 0.5 s. The open-circuit voltage generated by Ag/GOx@GelMA/ZnO MN^@EXO@miR^ scaffolds under US excitation was 40 mV (Fig. 6K). To directly validate the effect of ultrasound-induced piezoelectric electrostimulation on exosome internalization, we quantified Dil-labeled EXO@miR uptake in RAW264.7 macrophages by flow cytometry. Macrophages were treated with GelMA MN^@EXO@miR^, Ag/GOx@GelMA/ZnO MN^@EXO@miR^, or Ag/GOx@GelMA/ZnO MN^@EXO@miR^ under ultrasound stimulation. As shown by the overlaid histograms (Fig. S6C), Ag/GOx@GelMA/ZnO MN^@EXO@miR^ produced a moderate rightward shift in DiI fluorescence compared with GelMA MN^@EXO@miR^, indicating improved exosome delivery by the piezoelectric microneedles. Notably, application of ultrasound to Ag/GOx@GelMA/ZnO MN^@EXO@miR^ resulted in a further marked right shift and increase in mean fluorescence intensity, demonstrating significantly higher exosome internalization than either GelMA MN@EXO@miR or Ag/GOx@GelMA/ZnO MN^@EXO@miR^ without ultrasound. These results provide direct biological evidence that ultrasound-activated piezoelectric electrostimulation facilitates macrophage uptake of engineered exosomes, thereby supporting the enhanced delivery performance of the Ag/GOx@GelMA/ZnO MN^@EXO@miR^ system.Fig. 6Preparation and characterization of Ag/GOx@GelMA/ZnO MN^@EXO@miR^. (A) Fabrication procedure of Ag/GOx@GelMA/ZnO MN^@EXO@miR^. (B,C) The appearance of GelMA MN and Ag/GOx@GelMA/ZnO MN patch under a low-magnification field of view. (D)Scanning electron microscopy (SEM) images to verify the needle shape of Ag/GOx@GelMA/ZnO MN; Scale bar: 1000 μm, 500 μm, 300 μm. (E) Confocal microscopy image of GelMA MN (red) loading EXO (green). (F) Force-displacement curve from compression of GelMA MN and Ag/GOx@GelMA/ZnO MN. (G) SEM images to verify the internal structure of Ag/GOx@GelMA/ZnO MN; Scale bar: 150 μm, 50 μm. (H) EDS to verify the element composition of Ag/GOx@GelMA/ZnO MN; Scale bar: 50 μm. (I) Degradation curve of GelMA MN and Ag/GOx@GelMA/ZnO MN determined by dry weight. (J) EXO release curve of GelMA MN and Ag/GOx@GelMA/ZnO MN analyzed using the bicinchoninic acid (BCA) assay. (K) Piezoelectric effect test of Ag/GOx@GelMA/ZnO MN. (L) Physical and electron microscopy images of MRSA and E. coli cultured for 12 h; Scale bar: 1 cm, 500 nm. (M) DFCH-DA staining of HUVECs treated with liquid GelMA MN or Ag/GOx@GelMA/ZnO MN extracts. (N) Live and dead cell assays of HUVECs treated with liquid GelMA MN or Ag/GOx@GelMA/ZnO MN extracts.Fig. 6

To evaluate the antibacterial efficacy of distinct MN formulations, extracts derived from each MN group were co-cultured with bacteria. The results showed that the Ag/GOx@GelMA/ZnO MN^@EXO@miR^ group exhibited excellent antimicrobial capacity against methicillin-resistant Staphylococcus aureus (MRSA) and Escherichia coli, respectively (Fig. 6L). DCFH-DA fluorescence staining revealed that Ag/GOx@GelMA/ZnO MN exhibited significant reactive oxygen species (ROS) scavenging activity. This potent antioxidant capacity aligns with the synergistic effects of silver nanoparticle (AgNP)-catalyzed glutathione peroxidase (GPx)-like activity and ZnO-mediated superoxide dismutation, which collectively neutralize hydroxyl radicals (·OH), superoxide anions, and hydrogen peroxide through redox cycling mechanisms (Fig. 6M). Furthermore, live/dead cell co-staining (calcein AM/propidium iodide) demonstrated minimal cytotoxicity of Ag/GOx@GelMA/ZnO MN in HUVECs after 72-h exposure, confirming no significant inhibition of proliferation (Fig. 6N). Besides, HE staining results of in vivo experiments also demonstrated that Ag/GOx@GelMA/ZnO MN exerted minimal toxic effects on major organs (Fig. S7).

Diabetic wounds remain notoriously difficult to heal due to chronic inflammation, persistent oxidative stress, hypoxia, and bacterial infections, which collectively impair tissue regeneration processes [46,47]. Based on these comprehensive evaluations, our multifunctional microneedle platform—integrating Ag/GOx coating for ROS-scavenging/bactericidal capacities and ZnO piezoelectric layers for enhanced exosome delivery—demonstrates robust mechanical integrity, sustained release kinetics, potent antibacterial/antioxidant efficacy, and excellent biocompatibility, thereby offering a promising localized therapeutic strategy for diabetic wound healing.

Ag/GOx@GelMA/ZnO MN@EXO@miR accelerates DW healing in vivo

3.7

The effect of Ag/GOx@GelMA/ZnO MN^@EXO@miR^ on wound healing in vivo was investigated in HFD/STZ diabetic mice (Fig. 7A). Full-thickness skin wounds were created on the dorsum of diabetic mice, in which Ag/GOx@GelMA/ZnO MN^@EXO@miR^ was introduced with GelMA MN group, GelMA MN^@EXO^ group and GelMA MN^@EXO@miR^ group serving as comparison. Compared to the blank control, treatment with GelMN MN alone provided only a marginal improvement in wound healing rate in diabetic mice. However, incorporating M2 exosomes into GelMN MN significantly accelerated wound closure versus GelMN MN monotherapy. Notably, the group receiving engineered miR-493-5p exosomes outperformed those treated with non-engineered M2 macrophage-derived exosomes. Strikingly, the optimal therapeutic outcome was achieved by supplementing the microneedle system with a multifunctional Ag/glucose oxidase (Ag/GOx) coating (for synergistic antibacterial and glycemic control) and a ZnO piezoelectric underlayer (enabling electrostimulation-enhanced tissue regeneration), which collectively promoted near-complete wound healing by day 14 (Fig. 7B–D). To evaluate the therapeutic effects on cutaneous regeneration, histological analysis of skin specimens was performed at two-week post-injury. Compared with controls (blank, GelMA MN, GelMA MN^@EXO^, and GelMA MN^@EXO@miR^ group), wounds treated with Ag/GOx@GelMA/ZnO MN^@EXO@miR^ exhibited complete re-epithelialization and dermal thickening, indicating enhanced tissue reconstruction (Fig. 7E–G, H). Masson's trichrome staining further revealed marked elevation in collagen deposition within the Ag/GOx@GelMA/ZnO MN^@EXO@miR^ group. Mechanistically, this extracellular matrix remodeling was contingent on EXO@miR delivery via the Ag/GOx@GelMA/ZnO MN platform (Fig. 8F–I). Collectively, these data substantiate the efficacy of Ag/GOx@GelMA/ZnO MN^@EXO@miR^ in accelerating diabetic wound repair.Fig. 7. Ag/GOx@GelMA/ZnO MN^@EXO@miR^ accelerates DW healing in vivo. (A) Schematic of diabetic wound model and treatment. (B) Photographs of wounds of C57BL/6J mice at different time points after different treatments investigated in this study. (C, D) Quantitative analysis of the relative wound areas at different times (n = 3). (E) Representative H&E staining images of wounds on day 14. Scale bar, 2.5 mm (top) and 200 μm (enlarged). (F) Representative Masson Trichrome staining images of wounds on day 14. Scale bar, 2.5 mm (top) and 200 μm (enlarged). (G-H) Quantitative analysis of epithelial and epidermal thickness in H&E staining. (I) Quantitative analysis of collagen deposition in Masson Trichrome staining. Data are presented as mean ± SD and statistical significance was analyzed via Student's two-sided t-test. ∗P value < 0.05; ∗∗P value < 0.01; ∗∗∗P value < 0.001 per; ∗∗∗∗P value < 0.0001 per group by unpaired t-test.Fig. 7. Fig. 8Ag/GOx@GelMA/ZnO MN^@EXO@miR^ accelerates DW healing in vivo through promoting M2 polarization and angiogenesis. (A) Immunofluorescence staining was performed at 7 days post-wounded using anti-Arg-1 and anti-iNOS; Scale bar: 200 μm. (B) Reactive oxygen species (ROS) concentrations within the wound site on the seventh day were quantified using dihydroethidium (DHE) with indicated treatments; Scale bar: 200 μm. (C, D) Quantitative of Arg-1 and iNOS on day 7 in response to indicated treatments (n = 3). (E) Quantitative of ROS on day 7 in response to indicated treatments (n = 3). (F) The blood flow of wound area on day 10 was measured using the laser speckle contrast imaging system. (G) Immunohistochemical staining was performed at 10 days post-wounded using anti-CD31 and anti-α-SMA. (H) Quantitative of mean perfusion unit (MPU) ratios on day 10 in response to indicated treatments (n = 3). (I, J) Quantitative of CD31 and α-SMA on day 10 with indicated treatments (n = 3). Data are presented as mean ± SD and statistical significance was analyzed via Student's two-sided t-test. ∗P value < 0.05; ∗∗P value < 0.01; ∗∗∗P value < 0.001 per; ∗∗∗∗P value < 0.0001 per group by unpaired t-test.Fig. 8

Ag/GOx@GelMA/ZnO MN@EXO@miR accelerates DW healing in vivo through promoting M2 polarization and angiogenesis

3.8

Immunofluorescence analysis of diabetic wounds revealed that Ag/GOx@GelMA/ZnO MN^@EXO@miR^ treatment significantly upregulated Arg1 expression while suppressing iNOS levels, demonstrating its capacity to modulate the inflammatory microenvironment by promoting macrophage polarization towards an anti-inflammatory phenotype (Fig. 8A). Dihydroethidium (DHE) staining indicated a substantial decrease in intracellular reactive oxygen species (ROS) following intervention with exosome-incorporated microneedles (GelMA MN^@EXO^ or GelMA MN^@EXO@miR^), with the Ag/GOx@GelMA/ZnO MN^@EXO@miR^ exhibiting optimal ROS-scavenging efficacy (Fig. 8B). Quantitative analysis on day 7 confirmed these synergistic regulatory effects on Arg1, iNOS, and ROS levels across treatment groups (Fig. 8C–E). Doppler assessment of wound perfusion demonstrated significantly elevated mean perfusion unit (MPU) ratios in microneedle-treated cohorts, particularly in the Ag/GOx@GelMA/ZnO MN^@EXO@miR^ group (Fig. 8F–H). Furthermore, this composite system markedly enhanced expression of angiogenic markers CD31 and α-SMA (Fig. 8G–I, J), evidencing accelerated angiogenesis. Collectively, these findings establish that Ag/GOx@GelMA/ZnO MN^@EXO@miR^ facilitates diabetic wound healing through coordinated mechanisms: inducing macrophage phenotypic transition, attenuating inflammatory and oxidative stress responses, and promoting vascular regeneration.

In our study, the miRNA-493-5p–engineered exosomes demonstrated superior functional outcomes in relevant cellular models, which we attribute to efficient cargo packaging and natural cell-to-cell transfer dynamics. Currently, drugs derived from natural plant sources, synthetic short peptides, and other wound-active agents are being delivered through various platforms, such as hydrogels and nanoparticles, in increasingly diverse ways to promote wound healing [48,49]. Compared with natural plant-derived agents and other wound-active drugs, the exosome–miRNA microneedle system in this manuscript offers several distinctive advantages [14]. Plant phytochemicals and growth factors often suffer from poor stability, rapid clearance, and pleiotropic, poorly targeted actions, even when incorporated into hydrogels or nanoparticles [50]. Synthetic short peptides can be rationally designed for pro-angiogenic or immunomodulatory effects, yet they are still vulnerable to proteolysis and frequently lack cell-specific targeting, so their spatiotemporal activity in heterogeneous diabetic wounds is difficult to control [51,52]. By contrast, the present strategy uses M2 macrophage–derived exosomes as a biological nanocarrier for a single, well-defined synthetic miRNA (miR-493-5p), achieving protected cargo transport, intrinsic tropism to inflammatory sites, and receptor-mediated uptake [53]. Encapsulation in GelMA hydrogel microneedle arrays further enables minimally invasive, localized, sustained release, while the Ag/GOx coating and ZnO piezoelectric substrate add antibacterial, antioxidant, and electrostimulation-enhanced delivery functions. Potential disadvantages include manufacturing complexity, higher cost, and standardization of exosome production and safety, which need to be further explored.

Conclusions

4

This study presents a piezoelectric microneedle system delivering miR-493-5p-engineered exosomes (Ag/GOx@GelMA/ZnO MN^@EXO@miR^) for diabetic wound healing. The exosomes inhibit HDAC1 to enhance histone H3K18 lactylation, activating STAT6 signaling to drive metabolic reprogramming and M2 macrophage polarization. The multifunctional platform combines antibacterial Ag/GOx coating, sustained-release GelMA, and ultrasound-responsive ZnO electrostimulation to optimize exosome delivery. In diabetic murine models, the system accelerated wound closure through enhanced collagen deposition, re-epithelialization, angiogenesis, and M2 enrichment while reducing inflammation and ROS. This integrated strategy validates a novel epigenetic-metabolic axis for chronic wound therapy.

Ethical approval

All experimental protocols were approved by the Wenzhou Institute of University of Chinese Academy of Sciences (WIUCAS24121902).

CRediT authorship contribution statement

Tao Chen: Methodology, Project administration, Resources. Lizhi Ouyang: Methodology, Project administration, Resources. Bobin Mi: Software, Supervision, Validation. Yishen Sheng: Resources, Software, Supervision. Fawwaz AL-Smadi: Supervision, Validation, Visualization. Hui Xu: Investigation, Methodology, Project administration. Yaosen Wu: Investigation, Supervision, Validation. Xiaolei Zhang: Methodology, Project administration, Resources. Kailiang Zhou: Investigation, Methodology, Project administration. Aimin Wu: Supervision, Validation, Visualization. Xiangyang Wang: Formal analysis, Funding acquisition. Guohui Liu: Data curation, Formal analysis, Funding acquisition. Wenqian Zhang: Visualization, Writing – original draft, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.