Tower-shaped super-adhesive microneedles loaded with vitamin A-containing Bletilla striata vesicles: temporally regulated Inflammation resolution and ECM remodeling for scarless wound healing

Abstract

Pathological scars, resulting from abnormal collagen metabolism during wound healing, pose clinical challenges due to the high recurrence rates of conventional treatments. Therefore, early-stage modulation of the wound healing process is emerging as a pivotal strategy for preventing and treating pathological scars, necessitating precise coordination of vascular regeneration, inflammation regulation, cell migration, and collagen remodeling. Here, we developed a 3D-printed tower-shaped microneedle system (BSEVs@VA-HP MNs) loaded with vitamin A (VA)-containing Bletilla striata vesicles (BSEVs@VA) for functionally staged effects supported by release kinetics in scarless wound healing. The BSEVs@VA-HP MNs demonstrate exceptional mechanical strength, making them highly effective in penetrating the skin, adhering to tissues, and enzyme-responsive degradation. In rat wound models and rabbit ear scar models, the BSEVs@VA-HP MNs gradually dissolve and release BSEVs, which passively target and are internalized by cells. These BSEVs reprogram macrophages toward an M2 phenotype, mediating anti-inflammatory effects while simultaneously promoting angiogenesis and cell migration. Upon cellular internalization, lysosomes degrade BSEVs, liberating VA, which regulates the type I/III collagen balance during tissue remodeling, thereby suppressing myofibroblast activation and facilitating scarless wound healing. This study integrates natural plant-derived vesicles with 3D-printed bioinspired microneedle technology to establish a synergistic, stage-coordinated therapeutic strategy for scarless wound healing.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04462-7.

Introduction

Excessive scars resulting from wound healing represent pathological outcomes of the body’s trauma repair process [1]. They pose risks of long-term pain, itching, restricted mobility, recurrent ulcers, and even malignant transformation [2]. Scar formation arises from fibroblast hyperproliferation coupled with dysregulated collagen metabolism, ultimately causing excessive growth of fibrous connective tissue [3]. Treatment of pathological scars after their formation is associated with high recurrence rates. For instance, hypertrophic scars occur in 40%-70% of surgical procedures and can reach up to 91% after burn injuries, depending on wound depth [4, 5]. Although early intervention during wound healing is widely recognized as an effective anti-scarring strategy, the sequential progression of scar formation through inflammatory, proliferative, and maturation phases complicates the precise timing of anti-scarring drug administration. Administering drugs too early may suppress inflammation and delay healing, while administering them too late may fail to inhibit scar formation. Therefore, developing a drug delivery platform capable of sequentially exerting anti-inflammatory and anti-scarring effects during wound healing is crucial for scar inhibition.

Vitamin A (VA) has distinct functions in regulating wound repair and suppressing scar development by promoting normal differentiation of keratinocytes and fibroblasts, thereby reducing excessive collagen synthesis and improving scar appearance [6]. However, the clinical application of VA is limited by poor water solubility, vulnerability to oxidation, and suboptimal bioavailability [7]. Nanocarrier technology has become a viable approach for enhancing VA delivery, yet traditional nanosystems (e.g., liposomes, micelles, nanoparticles) face drawbacks such as limited biocompatibility, high production costs, and potential immunogenicity [8–11]. Plant-derived extracellular vesicles (PEVs) represent an innovative class of natural nanocarriers with distinct advantages, including high biocompatibility, low immunogenicity, and inherent bioactivity, enabling precise cross-species delivery of bioactive molecules [12]. For example, extracellular vesicles derived from Platycodon grandiflorus have demonstrated efficacy in delivering saponins and inhibiting triple-negative breast cancer growth [13]. Similarly, lemon-derived vesicles, following membrane modification, effectively deliver chemotherapeutic agents (e.g., doxorubicin) to targeted tumor sites [14]. These findings suggest that PEVs represent a promising advanced drug delivery platform.

As a traditional Chinese medicinal herb with hemostatic and tissue-repair properties, Bletilla striata (B. striata) has been widely used in wound care due to its ability to promote tissue regeneration and attenuate inflammatory responses [15–17]. Building on the advantageous properties of PEVs, the isolation of Bletilla striata-derived extracellular vesicles (BSEVs) preserves the natural pharmacological activities of B. striata, including anti-inflammatory and wound-healing effects, while also serving as an effective carrier for VA. However, an effective drug delivery system for VA-loaded BSEVs that enables sustained therapeutic effects at wound sites remains lacking [18].

Among various drug delivery systems, microneedles (MNs) have emerged as a promising transdermal tool for scar management, offering advantages such as efficient drug delivery, promotion of collagen realignment, and high patient compliance [19]. However, conventional MNs fabricated using molding techniques face limitations, including poor conformity to irregular scar morphology, susceptibility to skin extrusion, and fixed drug-loading capacity. The integration of 3D printing with MN fabrication enables the flexible design of diverse MN geometries (e.g., spear-shaped, pyramid-shaped, and barbed MNs) using photopolymerizable polymers as inks, thereby significantly enhancing transdermal drug delivery performance through improved skin penetration, enhanced tissue adhesion, and better adaptation to wound contours [20].

Herein, we propose a synergistic, stage-coordinated delivery strategy using a 3D-printed tower-shaped microneedle (BSEVs@VA MNs) patch. This patch precisely delivers VA-loaded Bletilla striata extracellular vesicles (BSEVs) to wound tissues. This approach leverages the anti-inflammatory and repair-promoting effects of BSEVs and the anti-scarring efficacy of VA, optimizing the therapeutic sequence from anti-inflammation to anti-scarring. As depicted in Scheme 1, BSEVs@VA MNs rapidly dissolve upon skin penetration, releasing BSEVs@VA into the wound microenvironment. Due to their inherent targeting ability, BSEVs are internalized by local cells, reprogramming macrophages toward the M2 phenotype, suppressing pro-inflammatory cytokines, and promoting angiogenesis and fibroblast migration. Lysosomal degradation of BSEVs releases VA, which regulates the collagen I/III ratio, inhibits myofibroblast activation, and synergistically enhances extracellular matrix (ECM) remodeling. This staged delivery creates a coordinated regulatory network involving inflammation resolution, tissue regeneration, and matrix remodeling, effectively promoting wound healing and preventing scar formation. Previously, MNs fabricated via conventional molding methods have combined a hydrophilic GelMA shell and hydrophobic PGLADMA core. These systems sequentially release mangiferin (rapid) and exosomes (sustained), aligning with different wound-healing phases at the tissue level [21]. In contrast, this study utilizes 3D printing to fabricate bioinspired tower-shaped MNs with excellent mechanical strength, enhanced tissue adhesion, and enzyme-responsive degradation. Collectively, this study presents a sequential therapeutic strategy for scarless wound healing through synergistic and stage-coordinated delivery, highlighting its potential in scar-free wound management.

Scheme 1 .

Preparation and mechanism of BSEVs@VA-HP MNs for scarless wound healing. (A) Fabrication process of BSEVs@VA-HP MNs. (B) Synergistic anti-inflammatory and anti-scarring effects

Experimental materials and procedures

Reagents and materials

Fresh Bletilla striata (Thunb.) Reichb. f. was sourced from Sichuan Kangyangjian Biotechnology Co., Ltd. (Sichuan, China). VA was procured from Chengdu DeSiTe Biological Technology Co., Ltd. (Sichuan, China). Hyaluronic acid was obtained from Foshan Kepai Chemical Co., Ltd. (Foshan, China). Polyethylene glycol diacrylate 600 (PEGDA600) was acquired from Yongqinquan Intelligent Equipment Co., Ltd. (Suzhou, China). Methacrylated hyaluronic acid (HAMA) was synthesized according to previously reported methods [22]. L929 cells, HUVECs, and RAW264.7 cells were purchased from Guangzhou Huatuo Biotech Co., Ltd. Primary antibodies against RXR, SOX9, iNOS, CD86, α-SMA, IL-6, COL-I, COL-III, IL-17, and MMP19 were obtained from Wuhan Servicebio Technology Co., Ltd. (Wuhan, China). The Csrp2 antibody was acquired from Zenbio Biotechnology Co., Ltd. SPF male SD rats (180–220 g) used in this study were obtained from SPF (Beijing) Biotechnology Co., Ltd. Ethical approval for this research and all related experimental procedures was granted by the Ethics Committee of Southwest Minzu University (Approval No.: SMU202401116). All chemicals used in this research were of analytical grade and required no further purification.

Isolation of BSEVs and preparation of BSEVs@VA

BSEVs were isolated from fresh Bletilla striata rhizomes. Freshly harvested rhizomes were thoroughly washed, roots were excised, and the rhizomes ground into homogenates. These homogenates were filtered through a coarse sieve to remove large debris. The filtrate underwent sequential purification using differential centrifugation and microporous filtration. The clarified supernatant was collected and processed by ultracentrifugation to isolate BSEVs. Purified BSEVs were resuspended in PBS, aliquoted, and stored at -80 °C for subsequent experiments.

BSEVs@VA were prepared by co-incubation. The protein concentration of freshly isolated BSEVs was quantified using a bicinchoninic acid (BCA) assay kit (Boster Biological Technology; Catalog# AR0146) according to the manufacturer’s instructions. VA was precisely weighed and mixed with BSEVs at a feed ratio of 1:1. The mixture was incubated at 37 °C with continuous shaking (100 rpm) for 2 h. Unencapsulated free VA was removed by ultracentrifugation (150,000 × g, 2 h, 4 °C). The supernatant was discarded, and the precipitate was resuspended in PBS to obtain BSEVs@VA (VA concentration: 2.79 ± 0.22 mg/mL).

Characterization of BSEVs and BSEVs@VA

The structural morphology of BSEVs and BSEVs@VA was examined by transmission electron microscopy (TEM; n = 3). Zeta potentials were measured using dynamic light scattering (DLS; Malvern Panalytical, UK; n = 3). Nanoparticle tracking analysis (NTA) was performed to determine particle size, concentration, and visualize motion (n = 3) [23]. Protein concentrations in BSEVs and BSEVs@VA were quantified using a BCA assay. SDS-PAGE was used to identify protein molecular weights (n = 3). Proteomic analysis (conducted by Novogene Co., Ltd.) characterized the protein composition and functional properties of BSEVs (n = 3). Encapsulation efficiency (EE) and loading capacity (LE) of VA in BSEVs were quantified by UV spectrophotometry (n = 3). To evaluate batch-to-batch reproducibility, three independent batches of BSEVs@VA were prepared using identical protocols. Particle size, polydispersity index (PDI), and zeta potential were measured for each batch by DLS. EE and LE of VA were also determined using UV spectrophotometry. The consistency across batches was assessed by calculating the relative standard deviation (RSD) for each parameter.

Preparation and characterization of BSEVs@VA-HP MNs

Preparation of BSEVs@VA-HP MNs

The photocurable printing ink was prepared by dissolving HAMA, PEGDA600, and 0.2% (w/v) LAP photoinitiator in PBS buffer. After homogenization, the BSEVs@VA complex was incorporated to obtain the final printing ink. A tower-shaped microneedle template was designed using Cinema 4D software. The prepared ink was then poured into the vat of the printer (EFL-BP86). The cover was closed, and the mixture was allowed to stand for 5 min to remove air bubbles. The target model was imported into the printing system. The slicing layer height was set to 50 μm. The light intensity was set to 12, and the exposure time was set to 13 s. Printing was then initiated. After printing, the MNs were removed, and residual ink was gently removed using paper. The samples were dried at room temperature for 24 h, followed by post-curing under 365 nm UV light (10 mW/cm²) for 10 min to obtain BSEVs@VA-HP MNs. Blank MNs (HP MNs), BSEVs-loaded microneedles (BSEVs-HP MNs), and VA-loaded microneedles (VA-HP MNs) were prepared using the same procedure.

Characterization of BSEVs@VA-HP MNs

The microstructure and surface morphology of BSEVs@VA-HP MNs were examined using field-emission scanning electron microscopy (SEM; GeminiSEM 360, ZEISS, Germany; n = 3). Before observation, samples were fixed on conductive adhesive and coated with a thin gold layer (10 nm) under vacuum. Images were acquired at an accelerating voltage of 5 kV, a working distance of 8 mm, and magnifications of 35 × , 50 × , and 200 × . Fourier transform infrared spectroscopy (FTIR; IRTracer-100, SHIMADZU, Japan; n = 3) was used to analyze chemical functional groups. Samples were mixed with spectroscopic-grade KBr (1:100, w/w) and pressed into transparent pellets. Spff 4000–400 cm⁻¹ at a resolution of 4 cm⁻¹ (32 scans) using a pure KBr background. X-ray diffraction (XRD; SmartLab, Rigaku, Japan; n = 3) was performed to analyze crystalline structure. Samples were mounted on a zero-background holder and scanned from 5° to 90° (2θ) at 2°/min (step size 0.02°) using Cu Kα radiation (40 kV, 40 mA). Diffraction patterns were analyzed using Origin software.

Mechanical strength of BSEVs@VA-HP MNs

Adequate mechanical strength is essential for MNs to effectively penetrate the stratum corneum and deliver drugs [24]. Insufficient strength limits clinical applications, despite favorable drug release profiles [25, 26]. The mechanical properties of four MN groups (HP MNs, BSEVs-HP MNs, VA-HP MNs, and BSEVs@VA-HP MNs) were assessed using a texture analyzer (Rapid TA + , Tengba Instrument Technology Co., Ltd., Shanghai, China). MN arrays (tips upward) were placed on the analyzer, and the probe was gradually lowered until tip fracture occurred. Each test was performed in triplicate, and data were recorded for mechanical strength evaluation.

Insertion performance of BSEVs@VA-HP MNs

To evaluate the skin penetration capability of BSEVs@VA-HP MNs, SD rats were first anesthetized. Hair from dorsal areas was removed by close clipping, followed by depilatory cream application. MNs were then applied to the dorsal skin under constant pressure for 5 min. Skin tissues from penetration sites were fixed in 4% paraformaldehyde, embedded in paraffin, sectioned, and stained with H&E (n = 3).

In Vivo dissolution performance of BSEVs@VA-HP MNs

A defining feature of dissolving MNs is their gradual dissolution and drug release after skin penetration [27]. To evaluate the in vivo dissolution of BSEVs@VA-HP MNs, MNs were inserted into anesthetized SD rats. MNs were retrieved at 0, 10, and 30 min post-insertion (n = 3 per time point) and examined by optical microscopy to assess dissolution status.

In Vitro drug release and degradation of BSEVs@VA-HP MNs

BSEVs@VA-HP MNs were immersed in PBS (pH 7.4) and incubated under continuous shaking (100 rpm, 37 °C). To maintain sink conditions, 2 mL of medium was sampled at designated intervals and replaced immediately with fresh PBS. Sample absorbance was measured at 325 nm by UV–Vis spectroscopy (n = 3). Cumulative release profiles were plotted based on absorbance values calibrated against a standard curve of VA solution.

HP MNs, BSEVs-HP MNs, VA-HP MNs, and BSEVs@VA-HP MNs were freeze-dried and precisely weighed (M₀). MNs were soaked in PBS (pH 7.4) containing 500 µg/L hyaluronidase and incubated (37 °C, 100 rpm). Over 8 days, MNs were retrieved daily, freeze-dried, and reweighed (M_t, n = 3 per group). The degradation rate was calculated using the following equation:

Tissue adhesion assessment of BSEVs@VA-HP MNs

Tissue adhesion properties of BSEVs@VA-HP MNs were compared with smooth conical MNs using chicken muscle tissue (n = 3). MNs were inserted to a depth of 3 mm for 60 s and then withdrawn. The maximum upward tissue displacement during retraction was quantified as the adhesion strength.

Cytotoxicity evaluation of BSEVs@VA-HP MNs on L929 cells

Cytotoxicity of BSEVs@VA-HP MNs was evaluated on RAW264.7, L929, and HUVEC cells using the CCK-8 assay. Cells were seeded at 10⁴ cells/well in 96-well plates and incubated for 24 h. After adherence, cells were treated with fresh medium containing various concentrations (0.1, 0.3, 0.6, and 1 µg/mL) of the four MN formulations for another 24 h (n = 3, triplicate wells per concentration). Finally, 10 µL CCK-8 reagent was added per well. After 1 h of incubation, absorbance at 450 nm was measured to calculate cell viability.

Cellular uptake of BSEVs in L929 cells

To assess BSEVs cellular uptake, coumarin-6-labeled BSEVs were incubated with L929 fibroblasts for 1, 2, and 4 h (n = 3). Uptake efficiency was quantified by flow cytometry (FlowJo software) after cells were resuspended in PBS (n = 3 per time point). After DAPI staining, fluorescence microscopy was used to visualize BSEVs uptake morphology in L929 cells, and fluorescence intensity was quantified using ImageJ software (n = 3 random fields per sample).

Assessment of intracellular ROS content in L929 cells

Intracellular antioxidant capacity was assessed using the DCFH-DA probe to detect ROS. L929 cells were seeded in 6-well plates for 24 h (n = 3), then exposed to 0.3 mM H₂O₂ for 30 min as the oxidative stress control group. Cells were subsequently treated for 30 min with extracts from HP MNs, BSEVs-HP MNs, VA-HP MNs, and BSEVs@VA-HP MNs (all containing 0.3 mM H₂O₂; VA concentration was adjusted to 1 µg/mL in relevant extracts). Cellular ROS levels (green fluorescence) were visualized by inverted fluorescence microscopy and quantified with ImageJ software (n = 3 random fields per well).

Regulation of macrophage polarization in RAW264.7 cells

RAW264.7 cells were seeded in 12-well plates and incubated for 24 h. Cells were then divided into groups (n = 3 wells per group): untreated control, LPS-only (500 ng/mL, inflammatory control), and treatment groups (500 ng/mL LPS + HP MNs, BSEVs-HP MNs, VA-HP MNs, or BSEVs@VA-HP MNs extracts). After 6 h of stimulation, cells and supernatants were collected.

Cells were resuspended in staining buffer and incubated for 30 min in the dark with macrophage polarization markers (APC-CD86 for M1, PE-CD206 for M2). Polarization status was assessed using flow cytometry (FlowJo software, n = 3).

Measurement of inflammatory factors in RAW264.7 cells

Levels of TNF-α, IL-6, IL-1β, and IL-10 in RAW264.7 cell supernatants were quantified using commercial ELISA kits according to the manufacturers’ protocols (n = 3).

Angiogenesis assay in HUVECs cells

An in vitro angiogenesis assay assessed the potential of BSEVs@VA-HP MNs to enhance wound healing through neovascularization. Briefly, HUVECs were seeded onto a pre-formed Matrigel layer and incubated with HP MNs, BSEVs-HP MNs, VA-HP MNs, or BSEVs@VA-HP MNs for 6 h (n = 3 wells per group). The vascular networks formed were stained with Calcein-AM, visualized by fluorescence microscopy, and quantitatively analyzed (total tube length and branch points) using ImageJ software (n = 3 random fields per well).

Cell migration assay in L929 cells

A scratch assay evaluated the wound-healing efficacy of BSEVs@VA-HP MNs by measuring L929 fibroblast migration. Briefly, a scratch was introduced into a ~ 90% confluent monolayer, followed by gentle rinsing with PBS. Cells were treated with HP MNs, BSEVs-HP MNs, VA-HP MNs, or BSEVs@VA-HP MNs (n = 3 wells per group). Cell migration was monitored by Calcein-AM staining at 0 h and 24 h post-scratch. Images were captured using inverted fluorescence microscopy (n = 3 random fields per well).

Therapeutic effects on full-thickness skin defect model in rats

All experiments strictly complied with protocols approved by the Animal Ethics Committee of Southwest Minzu University (Approval No.: SMU-202401116). Male Sprague–Dawley (SD) rats (200–220 g) were anesthetized with 3% pentobarbital sodium (30 mg/kg). After dorsal hair removal, full-thickness wounds (10 mm diameter) were created and randomly divided into five groups: Control, HP MNs, BSEVs-HP MNs, VA-HP MNs, and BSEVs@VA-HP MNs (n = 6 rats per group).

On days 3, 7, and 14 post-treatment, wounds were photographed, and wound areas were measured to calculate wound closure rates (n = 6 wounds/group/time point). Wound tissues collected at these intervals underwent histological staining (H&E, Masson, and Sirius red; n = 3). On day 7 post-injury, immunofluorescence staining assessed macrophage polarization (iNOS-labeled M1 and CD206-labeled M2 phenotypes; n = 6 random fields/section). On day 14, immunofluorescence was used to evaluate collagen deposition (α-SMA, COL-I, COL-III) and hair follicle regeneration (RXR, SOX9; n = 3 random fields/section).

RNA sequencing and bioinformatic analysis

On day 14 post-treatment, full-thickness skin tissues were harvested from wounds in the Control and BSEVs@VA-HP MNs groups (n = 3/group). Total RNA was extracted using the TRIzol method, and sequencing libraries were prepared following poly(A) enrichment. Sequencing was conducted by Novogene (Beijing, China) on an Illumina NovaSeq 6000 platform. Differentially expressed genes (DEGs) were identified using DESeq2 with criteria (|log₂FC|> 1, adjusted p-value < 0.05). Subsequently, GO and KEGG enrichment analyses were performed using the R package clusterProfiler.

Quantitative proteomic analysis

On day 14 post-treatment, full-thickness skin tissues were collected from Control and BSEVs@VA-HP MNs groups (n = 3/group) for proteomic profiling. Proteins were extracted, digested with trypsin, desalted, and fractionated before LC–MS/MS analysis using an Orbitrap Astral mass spectrometer in DIA mode. Mass spectrometry data were processed against the Rattus norvegicus UniProt database using specialized software. Differentially expressed proteins (DEPs) were identified (|fold change|> 1.5, adjusted p-value < 0.05) and subjected to comprehensive bioinformatic analysis and functional annotation. Transcriptome–proteome correlation analyses were performed on the Novogene cloud platform to identify key pathways mediating the scarless healing effects of BSEVs@VA-HP MNs.

Western blot assay in vivo

Following electrophoresis, proteins were transferred onto PVDF membranes. Membranes were incubated overnight (4 °C) with primary antibodies (IL-17, MMP19, Csrp2, GAPDH). After thorough washing, membranes were incubated with HRP-conjugated secondary antibodies for 1 h at room temperature. Immunoreactive bands were imaged, and intensities quantified using AIWBwell™ software (n = 3).

Biosafety evaluation of BSEVs@VA-HP MNs

An in vitro hemolysis assay evaluated the hemocompatibility of the MNs. Briefly, RBCs isolated from SD rats were washed and prepared as a 5% (v/v) suspension (n = 3). HP MNs, BSEVs-HP MNs, VA-HP MNs, and BSEVs@VA-HP MNs were co-incubated with the RBC suspension at 37 °C for 1 h. Normal saline served as the negative control and Triton X-100 as the positive control (n = 3 replicate wells per group). After centrifugation, supernatant absorbance was measured at 540 nm, and hemolysis rate was calculated using the formula:

where A_s, A_n, and A_p represent absorbances of the sample, negative control, and positive control, respectively.

To further assess biosafety, blood samples were collected from rats at day 14 post-treatment. Major organs (heart, liver, spleen, lungs, kidneys) were harvested on day 14 and processed for H&E staining (n = 3 tissue sections per organ per rat). Histological sections were scanned and imaged using a PANNORAMIC MIDI digital slide scanner (3DHISTECH).

Anti-scarring effect of BSEVs@VA-HP MNs

Female New Zealand large-eared rabbits (2–3 months old, 2–2.5 kg) were randomly divided into groups (n = 6). Rabbits were anesthetized using 3% sodium pentobarbital. After routine skin disinfection, scar models were established on both ears by needle pricking and removal of the perichondrium using a dissector. Following hemostasis, wounds received local treatment with PBS, HP MNs, BSEVs-HP MNs, VA-HP MNs, or BSEVs@VA-HP MNs. Wounds were photographed on days 0, 7, 14, and 28 post-modeling. On day 28, rabbits were sacrificed, and wound tissues were fixed in 4% PFA. Tissues were embedded in paraffin, sectioned, and stained (H&E, Masson’s trichrome, Sirius Red; n = 3). Sequential immunofluorescence staining (α-SMA, COL-I, COL-III) was performed (n = 3 random fields per section). Sections were scanned (PANNORAMIC MIDI 1, 3DHISTECH) and analyzed using ImageJ software.

Statistical analysis

All data are presented as mean ± SD from at least three independent experiments. Differences between two groups were evaluated using Student’s t-test, and multiple groups were compared using one-way ANOVA with Tukey’s post hoc test. P < 0.05 was considered statistically significant.

Results and discussion

Synthesis and characterization of BSEVs and BSEVs@VA

Plant-derived vesicles possess nanoscale bilayer membrane structures that enable inherent trans-membrane capability [28]. These vesicles can carry biomolecules such as proteins and nucleic acids and effectively encapsulate lipophilic small-molecule drugs through hydrophobic interactions [29]. Figure 1A shows a TEM image of BSEVs, clearly revealing their defined bilayer membrane structure. Figure 1B demonstrates that BSEVs retained their intact spherical morphology and membrane integrity after VA loading [30]. As illustrated in Fig. 1C, BSEVs exhibited an average diameter of 161.2 ± 0.7 nm, slightly decreasing to 160.6 ± 0.4 nm after VA loading (BSEVs@VA, Fig. 1D, n = 3). This minimal reduction indicates that VA primarily incorporates into the lipid bilayer rather than the aqueous lumen, consistent with hydrophobic interaction-based loading [31]. Figure 1E shows a zeta potential shift from −15.6 mV (BSEVs) to −13.3 mV (BSEVs@VA), likely due to partial charge shielding by VA molecules. The maintained negative charge indicates good colloidal stability [32]. Figure 1F reveals comparable protein profiles between BSEVs and BSEVs@VA, with major bands below 130 kDa, confirming VA loading did not alter protein composition. Particle-to-protein ratio (PPR) is widely accepted for assessing EV purity [33]; a high ratio indicates most proteins are vesicle-associated rather than free contaminants [34]. The particle concentration of BSEVs, measured by NTA, was 9.15 × 10¹⁰ ± 7.12 × 10⁸ particles/mL, and protein concentration by BCA assay was 2.49 ± 0.43 mg/mL. The calculated PPR was 3.68 ± 0.67 × 10⁷ particles/µg protein (Figure S1). This value aligns with reported purified EVs (e.g., 1.12 × 10⁷ for SEC-purified EVs [35]), confirming sufficient purity for downstream applications. The encapsulation efficiency (EE) of VA was 62.06 ± 4.83% (n = 3), and loading efficiency (LE) was 7.13 ± 1.54%, indicating the suitability of BSEVs as a VA carrier (Figure S2). Batch-to-batch reproducibility of BSEVs@VA was assessed with three independent preparations (Table S1). All batches demonstrated consistent physicochemical properties, including particle size, polydispersity index (PDI), zeta potential, EE, and LE. Relative standard deviations (RSD) were below 8% for particle size (1.28%), zeta potential (3.92%), EE (5.44%), and LE (7.79%), while PDI showed a slightly higher RSD of 13.38%. These results confirm acceptable reproducibility and stability, supporting BSEVs as reliable VA nanocarriers. FTIR analysis (Figure S3) confirmed successful VA loading, as BSEVs@VA exhibited characteristic peaks at 1620 cm⁻¹ (C = C stretch), 960 cm⁻¹ (alkene structure), and 720 cm⁻¹ (C–H stretch, cyclohexene). Release profiles (Figure S4A) showed sustained VA release in PBS (pH 7.4), achieving 79 ± 8.08% cumulative release over 72 h (n = 3). Enhanced release (87.19 ± 1.55%, n = 3) was observed in Triton X-100-containing PBS (Figure S4B), confirming VA incorporation within vesicle structures rather than surface adsorption.

Fig. 1.

Characterization of BSEVs and BSEVs@VA. (A) TEM images of BSEVs. (B) TEM images of BSEVs@VA. (C) Size distribution of BSEVs (n = 3). (D) Size distribution of BSEVs@VA (n = 3). (E) Zeta potentials of BSEVs and BSEVs@VA (n = 3). (F) SDS-PAGE electrophoresis of BSEVs and BSEVs@VA. (G) Venn diagram of BSEVs proteins. (H) Number of proteins identified in BSEVs. (I) GO annotation of proteins in BSEVs. (J) KEGG pathway enrichment of BSEVs proteins

Additionally, we conducted a quantitative proteomic analysis of Bletilla striata extracellular vesicles (BSEVs) to clarify their protein composition and potential functions. Proteomic analysis identified 19 proteins (Fig. 1H, n = 3). Venn diagram analysis (Fig. 1G) indicated that these proteins are annotated across multiple databases, including GO, KEGG, IPR, KOG, and subcellular localization, highlighting their functional diversity.

Figure 1I shows that the identified proteins were enriched in critical biological processes and molecular functions, including proton transport, ATP metabolic processes, and ATP binding. These results suggest BSEVs possess inherent capabilities to mediate transmembrane energy transduction and maintain cellular energy homeostasis, which are fundamental for tissue repair [36]. Additionally, the presence of chloroplast-derived proteins (PsbD and PsbE) associated with photosystem II implies potential antioxidant activity [37], possibly reducing oxidative stress within the wound microenvironment.

KEGG pathway analysis (Fig. 1J) revealed significant enrichment in pathways critical for tissue regeneration, including carbohydrate metabolism, energy metabolism, and translation. These pathways are integral to cellular energy supply, biosynthetic capacity, and intercellular communication, which are crucial during inflammatory, proliferative, and remodeling phases of wound healing [38].

These results demonstrate that BSEVs exhibit dual functionality, serving both as effective VA carriers and possessing intrinsic physiological activities directly involved in wound healing processes.

Preparation and characterization of BSEVs@VA-HP MNs

To prepare MNs, a mixture containing PEGDA600, HAMA, BSEVs@VA, 0.25% LAP, and 0.05% UVAW was prepared and photocrosslinked into well-defined tower-shaped structures using a digital light processing (DLP) printer (Figs. 2A and S5). Magnified imaging (Fig. 2B) showed that BSEVs@VA-HP MNs had a yellow, tower-shaped geometry. SEM imaging (Fig. 2E) further confirmed their structural integrity, displaying clearly defined, thread-like surface textures and precise dimensions: height 750 ± 12 µm, base diameter 350 ± 8 µm, and inter-needle distance 500 ± 10 µm (n = 3).

Fig. 2.

Preparation and characterization of BSEVs@VA-HP MNs. (A) Schematic illustration of 3D printing for fabricating BSEVs@VA-HP MNs. (B) Magnified view of BSEVs@VA-HP MNs. (C) FTIR spectra of BSEVs@VA-HP MNs. (D) XRD patterns of BSEVs@VA-HP MNs. (E) SEM images of BSEVs@VA-HP MNs. (F) Mechanical force characterization of BSEVs@VA-HP MNs. (G) Skin surface changes after BSEVs@VA-HP MN application and H&E staining of punctured rat skin (n = 3). (H) Dissolution images of BSEVs@VA-HP MNs in rat skin at different time points (n = 3). (I) Representative images of adhesion properties of BSEVs@VA-HP MNs and barbless MNs. (J) Quantitative adhesion evaluation of BSEVs@VA-HP MNs and barbless MNs (n = 3)

FTIR (Fig. 2C) and XRD (Fig. 2D) analyses collectively confirmed successful integration of BSEVs and VA into the MNs (n = 3). FTIR spectra demonstrated the chemical integrity of all components due to VA pre-encapsulation within BSEVs. XRD patterns indicated modifications in the crystalline structure of the matrix, affirming the structural and compositional integrity of the BSEVs@VA-HP MNs.

Previous studies reported that microneedle hardness must reach at least 0.098 N/needle to penetrate the skin [39]. All microneedle formulations in this study exhibited mechanical strength (1.91–3.95 N/needle) significantly above this threshold (Fig. 2F, n = 3). The highest strength was observed in MNs loaded with VA and BSEVs, indicating these additives enhanced polymer cross-linking and increased hardness [40]. These findings confirm sufficient mechanical integrity for effective skin insertion, supporting their use as robust localized drug-delivery platforms.

H&E staining (Fig. 2G) verified successful skin penetration by MNs. The resulting micropores closed within 30 min without causing erythema or edema, confirming rapid self-healing properties and excellent biocompatibility (n = 3).

Rapid dissolution of MNs is a prerequisite for precise drug release [41]. BSEVs@VA-HP MNs quickly dissolved in skin tissue, achieving nearly complete dissolution within 30 min (Fig. 2H, n = 3), enabling prompt drug release. The VA release profile exhibited a biphasic pattern, with an initial burst (44.32 ± 7.7% within 24 h) followed by sustained release (80.17 ± 7.69% cumulative release by day 8, Figure S6). It provides a stable drug-release source for BSEVs@VA-HP MNs to achieve functionally staged effects supported by release kinetics. However, direct causality between drug release kinetics and biological effects has not been fully confirmed. Therefore, we describe these effects as functionally staged effects supported by release kinetics and include a schematic timeline (Figure S6) to reflect this limitation.

Muscle tissue, with its dense structure and abundant collagen and elastin fibers, is an ideal model for evaluating microneedle adhesion, as it more closely mimics the mechanical properties of human skin and underlying tissues [42]. Adhesion testing in fibrous chicken muscle tissue showed that tower-shaped MNs (BSEVs@VA-HP MNs) adhered to the tissue upon withdrawal, forming a visible bulge, whereas smooth conical MNs (barbless MNs) remained flat and did not lift the tissue (Fig. 2I, n = 3). Quantitative analysis demonstrated that BSEVs@VA-HP MNs achieved a maximum withdrawal distance of 0.6 ± 0.03 cm, which was fivefold greater than that of barbless MNs (0.1 ± 0.02 cm; Fig. 2J, n = 3), indicating that geometric design plays a critical role in achieving strong tissue adhesion through mechanical interlocking. Although enhanced adhesion is beneficial for prolonged drug delivery and reducing patch dislodgement, it may raise concerns regarding removal, including potential tissue damage or discomfort. However, several factors mitigate these risks. The tower-shaped design relies on passive mechanical interlocking rather than barbed structures or chemical adhesives, reducing the risk of irreversible damage [43, 44]. Moreover, the microneedle matrix consists of biodegradable materials (HAMA and PEGDA) that undergo gradual enzymatic degradation, allowing interlocking forces to diminish over time and enabling atraumatic removal [45]. Additionally, the observed withdrawal distance (0.6 cm) reflects tissue elasticity rather than permanent deformation, indicating a compliant interaction mechanism [46]. Thus, the tower-shaped design achieves a balance between adhesion and safety by ensuring stable retention during treatment while permitting safe removal or bioabsorption upon degradation. Future studies will further evaluate the safety of repeated in vivo application and removal.

An appropriate degradation rate is essential for wound dressings [47]. All four microneedle types exhibited rapid degradation, with more than 89% mass loss within 8 days (Figure S7, n = 3), which is attributed to enzymatic degradation of the HAMA matrix by endogenous hyaluronidase in the wound microenvironment [48]. These findings provide insight into the relationship between enzyme-responsive degradation and physiological wound healing. Hyaluronidase is widely present in skin tissues [49] and can selectively degrade the HAMA matrix, potentially accelerating microneedle degradation compared with non-enzymatic conditions such as PBS [50]. This enzyme-responsive behavior may influence in vivo drug release kinetics. Although in vitro release studies in PBS showed a biphasic pattern, actual in vivo release may be faster due to accelerated enzymatic degradation [51]. Direct comparison between in vitro and in vivo release remains challenging. Further studies using enzyme-containing media or ex vivo skin models are required to better predict in vivo behavior [52]. Nevertheless, the observed biphasic release pattern provides valuable information on intrinsic release characteristics, and degradation results indicate that the system can respond to the enzymatic wound microenvironment.

These results indicate that the 3D-printed BSEVs@VA-HP MNs exhibit key functional properties, including mechanical robustness, rapid dissolution, biphasic drug release, enhanced adhesion, and enzymatic biodegradability. These features enable efficient drug release and sustained therapeutic action at the wound site. For clinical translation, however, several additional considerations must be addressed. Reproducibility and scalability: batch-to-batch consistency was confirmed for BSEVs@VA (RSD < 8%), though reproducibility data for unloaded BSEVs remain pending. Establishing standardized isolation protocols and scalable manufacturing techniques, such as tangential flow filtration, will be essential for ensuring consistent product quality [53-56]. Storage stability: humidity and temperature could affect microneedle integrity due to hygroscopic materials [57, 58]; incorporating cryoprotectants (e.g., trehalose or sucrose) may preserve bioactivity of BSEVs and VA [59, 60]. Future studies should assess stability under various conditions to determine shelf-life. Regulatory considerations: plant-derived EVs currently lack a clearly defined regulatory pathway compared to mammalian exosomes [61], necessitating comprehensive characterization and early consultation with regulatory authorities [62]. Clinical application with repeated microneedle usage: although single application was evaluated in this study, chronic wounds may require repeated dosing, which has shown safety in clinical trials with transient erythema being the primary adverse event [63, 64]. Future large animal studies and clinical trials will assess safety and tolerability for repeated applications. Addressing these translational aspects through scalable manufacturing, validated storage protocols, regulatory consultations, and comprehensive safety studies will be crucial for advancing the BSEVs@VA-HP MNs system toward clinical application.

Cytotoxicity activities and cellular uptake of BSEVs@VA-HP MNs

Clinical wound dressings require excellent biocompatibility [65]. BSEVs@VA-HP MNs exhibited excellent cytocompatibility, with cell viability greater than 85% (VA dosage: 1 μg/mL) in L929, HUVEC, and RAW264.7 cells, showing no significant cytotoxicity (Figure S8, n = 3). Other microneedle formulations at the same dosage also confirmed their potential as safe wound-dressing materials.

Due to their intact membrane structure, vesicles can efficiently cross cell membranes, facilitating intracellular drug delivery [66]. Flow cytometry confirmed time-dependent cellular uptake, with significantly increased fluorescence after 4 h (Figure S9, n = 3), supporting progressive internalization of BSEVs. BSEVs@C6 (47.77 ± 1.43) and BSEVs@C6-HP MNs (54.19 ± 2.86) both enhanced green fluorescence in L929 cells after 6 h compared to free C6 (16.88 ± 2.87) (Fig. 3A, n = 3), confirming effective cellular uptake. BSEVs@C6-HP MNs showed the highest fluorescence intensity. This phenomenon may result from cross-linked materials stabilizing vesicles and reducing C6 quenching, maintaining efficient uptake (Fig. 3B, n = 3) [67].

Fig. 3.

Evaluation of cellular uptake, antioxidant, and anti-inflammatory effects of BSEVs@VA-HP MNs. (A) Representative uptake images of BSEVs@C6-HP MNs in L929 cells. (B) Quantitative analysis of cellular uptake (n = 3). (C) Fluorescence images of DCFH-DA H₂O₂-treated L929 cells. (D) Quantification of DCFH-DA signal intensity (n = 3). (E) Flow cytometry analysis of RAW264.7 macrophage polarization (n = 3). (F) CD86 expression (n = 3). (G) CD206 expression (n = 3). (H) CD86/CD206 ratio (n = 3). (I-L) Quantification of TNF-α, IL-1β, IL-6, and IL-10 by ELISA (n = 3)

Elimination of ROS, regulation of macrophage polarization, and balancing of inflammatory factors

Excessive inflammation and ROS during wound healing hinder tissue repair and promote scarring [68]. Under H₂O₂-induced oxidative stress, BSEVs@VA-HP MNs significantly reduced DCFH-DA fluorescence (Fig. 3C). Quantitative analysis (Fig. 3D, n = 3) showed sharply increased ROS in the H₂O₂ group (53.83 ± 2.66) compared to control (0.50 ± 0.45). HP MNs alone had minimal effect (51.78 ± 1.45), whereas BSEVs-HP MNs (21.13 ± 2.66) and VA-HP MNs (27.53 ± 0.80) moderately reduced ROS. Notably, BSEVs@VA-HP MNs showed the strongest inhibition (15.06 ± 1.14, P < 0.0001 vs. H₂O₂), suggesting synergistic enhancement of ROS-scavenging by BSEVs and VA, effectively suppressing oxidative stress.

Notably, as macrophages are key regulators of the inflammatory phase [69], the anti-inflammatory activity of BSEVs@VA-HP MNs was evaluated by treating LPS-stimulated RAW264.7 cells for 6 h. Flow cytometry showed that BSEVs@VA-HP MNs increased M2 markers in LPS-stimulated macrophages, indicating a phenotypic shift (Fig. 3E, n = 3). The BSEVs@VA-HP MNs group showed a marked reduction in CD86 + M1 cells (3.33 ± 0.57%, Fig. 3F), an increase in CD206 + M2 cells (43.26 ± 2.5%, Fig. 3G), and a CD86/CD206 ratio (0.08 ± 0.02) close to the control (Fig. 3H). BSEVs and VA synergistically reversed the LPS-induced increase in the CD86/CD206 ratio. The combination (BSEVs@VA-HP MNs) showed a much stronger effect on promoting M2 polarization than either BSEVs-HP MNs or VA-HP MNs alone.

ELISA was used to quantify key inflammatory cytokines (Fig. 3I-L, n = 3). For pro-inflammatory cytokines, the LPS group showed increased levels of TNF-α (154.30 ± 12.84), IL-1β (26.30 ± 4.40), and IL-6 (29.79 ± 2.97) compared with the control (TNF-α: 87.51 ± 4.43; IL-1β: 14.41 ± 1.07; IL-6: 17.15 ± 0.32). Treatment with BSEVs-HP MNs reduced TNF-α (121.91 ± 12.67), IL-1β (18.16 ± 2.37), and IL-6 (22.46 ± 2.25). VA-HP MNs produced similar reductions (TNF-α: 115.12 ± 7.88; IL-1β: 19.25 ± 1.39; IL-6: 23.51 ± 0.63). For the anti-inflammatory cytokine IL-10, the LPS group (49.41 ± 8.41) was lower than the control (77.79 ± 6.82). Both BSEVs-HP MNs (79.69 ± 4.69) and VA-HP MNs (79.83 ± 3.95) restored IL-10 to near control levels. These results indicate that BSEVs strongly suppress pro-inflammatory cytokines, while VA shows comparable anti-inflammatory effects. Notably, BSEVs@VA-HP MNs achieved the strongest anti-inflammatory response, indicating that combined BSEVs and VA are required for optimal efficacy.

Promotion of angiogenesis and cell migration

Angiogenesis is essential for wound regeneration. It supplies nutrients and oxygen, removes waste, and guides tissue repair [70]. The angiogenesis assay in HUVECs showed that BSEVs@VA-HP MNs significantly promoted tube formation, with increased vascular junctions and total capillary length compared with the control (Fig. 4A-D, n = 3). Quantitative analysis (n = 3) showed vascular junction numbers of 352.0 ± 10.5, 712.3 ± 7.8, 2505.7 ± 100.1, 1238.0 ± 76.7, and 3311.3 ± 209.3 for Control, HP MNs, BSEVs-HP MNs, VA-HP MNs, and BSEVs@VA-HP MNs, respectively. Total capillary lengths were 21,885.7 ± 1094.0, 41,510.0 ± 1256.5, 75,439.7 ± 860.5, 52,083.7 ± 954.1, and 82,951.0 ± 1055.5, respectively. Both indices showed stronger angiogenesis in BSEVs-containing groups than in the VA-alone group. The pro-angiogenic effect was mainly driven by BSEVs, with VA providing synergistic enhancement during tissue remodeling, confirming the dominant role of BSEVs in angiogenesis.

Fig. 4.

Angiogenesis and cell migration assays. (A) Representative images of tube formation assay in HUVECs. (B) Schematic diagram illustrating angiogenesis assay. (C) Quantification of angiogenic node counts (n = 3). (D) Quantitative evaluation of total angiogenic structure length (n = 3). (E) Schematic diagram of L929 cell migration assay (n = 3). (F) Quantification of cell migration rates at 24 h (n = 3). (G) Representative images of L929 scratch assay (0 h and 24 h)

Furthermore, fibroblast migration to injury sites replenishes cellular populations, synthesizes extracellular matrix, and facilitates wound contraction and closure, accelerating wound healing [71]. Results showed that BSEVs@VA-HP MNs significantly enhanced fibroblast migration, achieving the highest wound closure rate (78.36 ± 2.64%) among all groups. This rate was markedly higher than the control (52.38 ± 2.56%), HP MNs (56.10 ± 1.26%), VA-HP MNs (66.50 ± 2.51%), and BSEVs-HP MNs (73.65 ± 0.75%) (Fig. 4E-G, n = 3). Notably, BSEVs-containing groups displayed significantly higher closure rates than the VA-alone group, suggesting that BSEVs primarily accelerate fibroblast migration, while VA provides a synergistic enhancement. This synergy was further confirmed by the optimal outcomes observed in the BSEVs@VA-HP MNs group.

Scarless wound-healing effects of BSEVs@VA-HP MNs in rat full-thickness skin defect models

These findings suggest BSEVs@VA-HP MNs hold promise for wound repair and scar reduction (n = 6). The BSEVs@VA-HP MNs group demonstrated superior wound healing, achieving a closure rate of 95% ± 1.0% on day 14, significantly higher than the control (83% ± 3.0%) (Fig. 5A-C). This enhanced healing is primarily due to the combined effects of BSEVs and VA.

Fig. 5.

Evaluation of scarless wound healing using a rat full-thickness skin injury model. (A) Schematic illustration of the wound therapy. (B) Representative images of wound healing in SD rats from different treatment groups at days 0, 3, 7, and 14 post-treatment. (C) Quantitative analysis of wound healing rates at indicated time points (n = 6). (D) Quantification of epidermal thickness index (ETI) on day 14 (n = 6). (E) Quantification of collagen I/III ratio (n = 6). (F) Representative images of H&E and Sirius Red staining of skin tissue sections on day 14. Red arrows indicate blood vessels. Polarized light: collagen I (orange-yellow), collagen III (green)

Re-epithelialization directly indicates wound closure progress, and the epidermal thickness index (ETI) characterizes wound healing [72]. H&E staining showed complete re-epithelialization in all groups on day 14. The BSEVs@VA-HP MNs group (98.33 ± 3.27) exhibited ETI values close to the normal group (102.83 ± 1.72), while the control group (355.50 ± 35.86) showed significant epithelial hypertrophy. HP MNs (217.67 ± 31.26), BSEVs-HP MNs (170.00 ± 16.00), and VA-HP MNs (120.67 ± 14.76) partially improved abnormal hyperplasia (Fig. 5D, F, n = 6). Combined BSEVs and VA effectively restored ETI to normal levels, outperforming single components. Together with neovascularization, BSEVs@VA-HP MNs promoted epidermal regeneration and reduced scarring.

Excessive collagen deposition drives scar formation, thus achieving a collagen ratio similar to normal skin is essential for scarless repair [73, 74]. Sirius Red staining and polarization microscopy accurately distinguish collagen type I (yellow) from type III (green), assessing fibrosis severity and scar type [75]. Results revealed significantly increased collagen I/III ratio in the control (1.41 ± 0.11) compared with normal (0.78 ± 0.02). HP MNs (0.93 ± 0.03), BSEVs-HP MNs (0.84 ± 0.08), and VA-HP MNs (0.85 ± 0.02) reduced this ratio to varying degrees. The combination (BSEVs@VA-HP MNs) synergistically normalized the ratio (0.74 ± 0.03), promoting scarless wound healing (Fig. 5E, n = 6). Masson staining results of collagen deposition are shown in Figure S10 (n = 6). In the control group, abundant collagen accumulated around granulation tissue, promoting scar formation. In contrast, collagen fibers in the BSEVs@VA-HP MNs group were orderly and flattened, supporting scarless repair.

Immunofluorescence staining analysis

During the inflammatory phase of wound healing, M1 macrophages release pro-inflammatory factors to recruit innate immune cells and enhance host defense. However, excessive inflammation delays wound repair. Later, macrophages transform into the M2 phenotype, secreting anti-inflammatory factors to facilitate tissue remodeling [76–78]. Figure S11 (n = 6) shows macrophage distribution and polarization in repaired skin tissues on day 7. The expression of iNOS (red, an M1 macrophage marker) was elevated in the control group but significantly reduced in the BSEVs@VA-HP MNs group. CD206 (green, an M2 macrophage marker) was expressed in all groups, but treatment with BSEVs@VA-HP MNs (37.51 ± 0.83) markedly increased its expression compared to controls (28.77 ± 1.43). This demonstrates that BSEVs@VA-HP MNs promoted macrophage polarization from M1 to M2 phenotype. Additionally, the inflammatory factor IL-6 showed minimal expression in the BSEVs@VA-HP MNs group, confirming effective inflammation resolution.

During the proliferative phase, myofibroblasts contract wounds using α-SMA stress fibers, effectively reducing wound area. Subsequent programmed apoptosis prevents excessive activation, accumulation, and scar formation [79–81]. Myofibroblasts synthesize and deposit collagen types I and III. Excessive collagen I deposition is strongly linked to pathological scarring. In contrast, scarless healing predominantly involves collagen III synthesis, promoting orderly fiber arrangement and serving as a scaffold for tissue regeneration [82, 83].

This study examined α-SMA, COL-I, and COL-III expression on days 7 and 14 post-wounding. α-SMA expression decreased over time, with the BSEVs@VA-HP group (0.49 ± 0.13) significantly lower than the control (11.04 ± 1.83) by day 14, indicating increased myofibroblast apoptosis. COL-I (related to skin strength) levels progressively declined, with BSEVs-HP (15.45 ± 0.96), VA-HP MNs (16.57 ± 2.20), and BSEVs@VA-HP MNs (15.97 ± 2.42) significantly lower than controls by day 14. COL-III (responsible for skin elasticity) slightly increased across all groups, with the highest elevation observed in the BSEVs@VA-HP group. The collagen I/III ratio dynamically decreased, reaching its lowest value in the BSEVs@VA-HP group (Control: 1.89 ± 0.06, HP MNs: 1.78 ± 0.19, BSEVs-HP MNs: 0.58 ± 0.12, VA-HP MNs: 0.68 ± 0.02, BSEVs@VA-HP MNs: 0.46 ± 0.09) (Fig. 6A, C-F, n = 3). These findings demonstrate that BSEVs@VA-HP mediates functionally staged effects supported by release kinetics on fibroblast activity and collagen expression, thereby promoting orderly tissue regeneration and reducing scar formation.

Fig. 6.

Immunofluorescence analysis of wound tissues on day 14. (A) Representative images of α-SMA, COL-I, and COL-III at days 7 and 14. (B) Representative images of RXR and SOX9 (n = 3). (C) Quantification of α-SMA intensity (n = 3). (D) Quantification of COL-I intensity (n = 3). (E) Quantification of COL-III (n = 3). (F) COL-I/COL-III ratio analysis (n = 3). (G) Quantification of RXR expression (n = 3). (H) Quantification of SOX9 expression (n = 3)

During skin injury repair, retinoid X receptor (RXR), a key nuclear receptor of VA, regulates the expression of genes such as SOX9. SOX9 promotes differentiation of hair follicle stem cells into hair follicle cells, thereby enhancing hair regrowth. Coordinated regulation of RXR and SOX9 maintains skin homeostasis and normal hair cycling, ultimately promoting scarless healing. As shown in Fig. 6B, the expression levels of RXR and SOX9 increased over time. Notably, VA alone (VA-HP MNs) induced slightly higher expression of RXR (13.89 ± 0.66) and SOX9 (1.44 ± 0.05) than BSEVs alone (BSEVs-HP MNs: RXR 13.54 ± 1.26; SOX9 1.42 ± 0.06), suggesting that VA is the primary activator of this pathway. However, the combination group (BSEVs@VA-HP MNs) achieved the highest expression levels (RXR 15.78 ± 1.66; SOX9 1.58 ± 0.17), exceeding both single treatments (Fig. 6B, G-H). This result highlights the role of BSEVs as an efficient delivery system that enhances targeted and sustained VA release, thereby maximizing therapeutic efficacy. These findings indicate that VA drives RXR/SOX9 activation, whereas BSEVs optimize its bioavailability and effectiveness. The synergy between precise delivery by BSEVs and biological activity of VA promotes hair follicle differentiation, restores skin function, and facilitates scarless healing.

Transcriptomic analysis revealed a pro-regenerative gene expression signature

To elucidate the anti-scarring mechanism, transcriptomic analysis was performed focusing on genes involved in extracellular matrix (ECM) remodeling. BSEVs@VA-HP MNs induced significant differential gene expression compared with controls. Volcano plot analysis identified 119 upregulated and 84 downregulated genes (|log₂FC|≥ 1, adjusted p < 0.05) (Fig. 7A). The heatmap showed clear separation between groups based on DEG profiles (Figure S12). GO analysis indicated that downregulated genes were associated with immunosuppression and ECM activity, whereas upregulated genes were related to tissue morphogenesis and cell migration (Figure S13). KEGG pathway analysis demonstrated that MNs suppressed the PI3K-Akt inflammatory pathway and activated the Wnt signaling pathway, promoting hair follicle regeneration and structural remodeling (Fig. 7B). ECM-focused analysis confirmed downregulation of pro-fibrotic genes (e.g., Col27a1, Ccl5) and upregulation of remodeling-associated genes (e.g., Camk2b, MMP19), forming a regulatory network that limits pathological ECM deposition (Fig. 7C).

Fig. 7.

Multi-omics and Western blot analyses reveal the scarless healing mechanism of BSEVs@VA-HP MNs. (A) Volcano plot of transcriptomic DEGs. (B) KEGG pathway enrichment of DEGs. (C) Heatmap of ECM-related genes. (D) KEGG enrichment of DEPs. (E) Subcellular localization of DEPs. (F) GSEA of GO molecular function (actin binding). (G) GSEA of KEGG (cytoskeleton in muscle cells). (H) GSEA of IPR (tubulin FtsZ, 2-layer sandwich domain). (I) Sankey diagram showing DEG and DEP correlations. (J) Nine-quadrant plot integrating transcriptomic and proteomic analyses (Red: both upregulated; Yellow: DEGs up, DEPs down; Green: DEGs down, DEPs up; Blue: both downregulated; Gray: below log₂FC threshold). (K) Representative Western blot images of IL-17, MMP19, and Csrp2. (L) Quantitative analysis of IL-17, MMP19, and Csrp2 expression (n = 3)

Proteomic analysis confirmed functional pathways at the protein level

Proteomic analysis identified 359 upregulated and 261 downregulated proteins (|log₂FC|≥ 1.2, adjusted p-value < 0.05) (Figure S14). Proteins showed clear clustering within groups (Figure S15) and distinct separation between groups (Figure S16). GO enrichment implicated processes such as small molecule metabolism, intracellular organelle function, and metalloexopeptidase activity in promoting scarless repair (Figure S17). KEGG enrichment highlighted the upregulation of the glycosaminoglycan biosynthesis pathway, enhancing ECM stability, and downregulation of glycine, serine, threonine metabolism, and PPAR signaling pathways, collectively forming a coordinated “matrix remodeling-inflammatory regulation” network (Fig. 7D). Subcellular localization and GSEA analyses indicated these proteins modulate cytoskeletal dynamics (e.g., actin binding, muscle cell cytoskeleton), promoting orderly cell migration and tissue regeneration (Fig. 7E-H).

Transcriptome-protein correlation analysis of key proteins involved in ECM remodeling and Western blot verification

To evaluate mRNA-protein correlations, a nine-quadrant plot was constructed (Fig. 7J). Most molecule pairs showed concordant regulation, confirming transcriptional control as a primary driver. The Sankey diagram (Fig. 7I) visualized functional pathways, revealing strong, coordinated upregulation of cell adhesion molecules and cytoskeleton in muscle cells pathways, driving organized repair, while the PPAR signaling pathway was synchronously downregulated. The focal adhesion pathway emerged as a central regulatory hub, integrating signals for coordinated ECM remodeling.

To experimentally validate central predictions, Western blot analysis was performed on three key proteins: MMP19 (ECM remodeling), Csrp2 (cytoskeletal regulation in muscle cells), and IL-17 (pro-inflammatory mediator). Results (Fig. 7K-L, n = 3) showed significant upregulation of MMP19 and Csrp2, and downregulation of IL-17 in wounds treated with BSEVs@VA-HP MNs. These findings matched integrated analyses, confirming MMP19 and Csrp2 upregulation, facilitating tissue restructuring and controlled cell migration, and validating IL-17 downregulation, suppressing pro-inflammatory signaling. Collectively, this evidence supports the anti-scarring mechanism whereby BSEVs@VA-HP MNs inhibit inflammation (IL-17 downregulation), enhance cell migration (Csrp2 upregulation), and remodel ECM (MMP19 upregulation), thus achieving scarless healing via functionally staged effects supported by release kinetics with coordinated anti-inflammatory and tissue remodeling responses.

In addition, hemocompatibility and histological evaluation of major organs confirmed excellent biosafety. Hemolysis rates remained below the 5% biomaterial threshold, and no tissue damage or toxicity was observed in vital organs, highlighting the clinical potential of BSEVs@VA-HP MNs (Figures S18, S19, n = 3).

BSEVs@VA-HP MNs inhibited scar formation in rabbit ear scar model

Although rat wound models are widely used, their healing pattern differs from human skin [84]. In contrast, the rabbit ear model is recognized as a suitable model for human scar formation [85]. Therefore, this model was used to further evaluate the anti-scarring efficacy of BSEVs@VA-HP MNs. As shown in Fig. 8A-B, rabbits treated with BSEVs@VA-HP MNs for 28 days exhibited improved wound healing compared with the control (n = 6). The PBS group developed red, raised, firm, and smooth scars with invasion into adjacent tissues. The HP MNs group showed similar outcomes to the control. Persistent erythema was observed in the BSEVs-HP MNs and VA-HP MNs groups, but was absent in the BSEVs@VA-HP MNs group. These findings indicate that scar inhibition results from the synergistic effects of BSEVs and VA, consistent with rat model results.

Fig. 8.

Evaluation of scarless wound healing using a rabbit ear scar model. (A) Schematic diagram of rabbit ear wound treatments. (B) Representative images of rabbit ear wounds from different treatment groups over time. Scale bar = 2 mm. (C) Representative images of H&E, Masson’s trichrome, and Sirius Red staining at day 28. (D) Quantitative analysis of scar elevation index (SEI, n = 3). (E) Quantitative analysis of collagen deposition (blue area) by Masson staining (n = 3). (F) COL-I/COL-III ratio determined by Sirius Red staining (n = 3). (G) Quantitative analysis of α-SMA expression by immunofluorescence (n = 3). (H) COL-I/COL-III expression quantified by immunofluorescence (n = 3). (I) Representative immunofluorescence images showing α-SMA and COL-I (red) / COL-III (green) expression at day 28

H&E, Masson, and Sirius Red staining were performed to evaluate dermal proliferation. The control group showed increased dermal thickness with obvious tissue protrusion (SEI > 105%), whereas the HP MNs group showed reduced dermal growth (SEI < 95%). In contrast, scar hyperplasia was significantly inhibited in the BSEVs@VA-HP MNs group (95% < SEI < 105%) (Fig. 8C-D, n = 3) [86]. Masson and Sirius Red staining further showed that collagen fibers in the BSEVs@VA-HP MNs group were more orderly arranged, with a normalized COL-I/III ratio similar to the normal group (Fig. 8C, E, F).

Immunofluorescence analysis showed reduced expression of pro-fibrotic markers (α-SMA and COL-I/COL-III) in fibroblasts. Quantitative analysis demonstrated that α-SMA expression in the control group (141.65 ± 3.07) was higher than in the normal group (107.96 ± 0.86). Treatment with HP MNs (132.32 ± 5.00), BSEVs-HP MNs (119.34 ± 6.18), VA-HP MNs (119.36 ± 3.77), and BSEVs@VA-HP MNs (105.95 ± 3.08) reduced α-SMA levels, with the BSEVs@VA-HP MNs group showing the greatest reduction, approaching normal levels. Similarly, the COL-I/III ratio in the control group (2.81 ± 0.10) was higher than in the normal group (1.53 ± 0.03). Treatment with HP MNs (2.06 ± 0.03), BSEVs-HP MNs (2.10 ± 0.04), VA-HP MNs (1.75 ± 0.04), and BSEVs@VA-HP MNs (1.15 ± 0.01) significantly reduced this ratio (Fig. 9G-I). Notably, VA-HP MNs showed a stronger effect than BSEVs-HP MNs, suggesting that VA primarily regulates collagen remodeling, while BSEVs play a supportive role. The BSEVs@VA-HP MNs group achieved the most normalized ratio, confirming a synergistic effect in ECM regulation. These findings are consistent with results from the rat full-thickness wound model, confirming inhibition of fibroblast activation and pro-fibrotic marker expression.

In summary, the rabbit ear model further confirms that BSEVs@VA-HP MNs reduce scar formation by regulating fibroblast activity, downregulating α-SMA expression, and decreasing the COL-I/COL-III ratio. This effect inhibits myofibroblast overactivation, reduces abnormal collagen deposition, promotes organized collagen alignment, and enhances ECM remodeling, ultimately achieving scarless healing.

Conclusion

This study introduces an innovative strategy for scarless wound healing through a 3D-printed, tower-shaped microneedle patch (BSEVs@VA-HP MNs) loaded with VA-containing Bletilla striata extracellular vesicles (BSEVs@VA). The MNs fabricated via 3D printing exhibited exceptional mechanical strength and robust tissue adhesion, ensuring effective skin penetration and sustained local drug delivery. BSEVs released through enzyme-responsive degradation contributed to anti-inflammatory effects by scavenging reactive oxygen species and promoting M2 macrophage polarization during the inflammatory phase. Subsequent lysosomal degradation of BSEVs released encapsulated VA, facilitating extracellular matrix remodeling by normalizing collagen type I/III ratios and activating RXR signaling pathways during the tissue remodeling phase, alongside hair follicle regeneration. Collectively, this study demonstrates a synergistic, stage-coordinated therapeutic approach with strong potential for clinical translation in scarless wound management.

Author contributions

S. E.: Conceptualization, Methodology, Formal analysis, Investigation, Writing – original draft; G.T.: Methodology, Formal analysis, Investigation, Writing – review & editing; Q. M.: Investigation, Writing – original draft; J.H.: Conceptualization, Methodology, Data curation; H. G.: Methodology; K. G.: Supervision; P.Z.: Conceptualization; J. Z.: Methodology; X.W.: Methodology, Writing – review & editing, Supervision; R.Z.: Resources, Funding acquisition, Project administration.

Funding

This work was supported by The National Natural Science Foundation of China (Grant No. 82474353), the Science and Technology Projects of Xizang Autonomous Region, China (Grant No. XZ202601ZY0175), the General project of the Sichuan Provincial Administration of TCM (Grant No. 25MSZX358), and the Scientific and Technological Innovation Team for Qinghai-Tibetan Plateau Research in Southwest Minzu University (Grant No. 2024CXTD23).

Data availability

The datasets supporting the conclusions of this article are included within the article and its additional file. They are available on request from the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.