Isoliquiritigenin micellar microneedle for pH monitoring and diabetic wound healing

Abstract

Chronic inflammation, oxidative stress, and insufficient angiogenesis hinder wound healing in patients with diabetes, necessitating long-term management strategies. In this study, we developed pH-responsive hydrogel microneedles (LSI-GCA) based on Lycium barbarum polysaccharide stearate micelles that could achieve real-time pH monitoring and intelligent drug release for effective wound management. LSI-GCA was prepared through hydrophobic modification and isoliquiritigenin (ISO) loading, followed by its combination with a chitosan/anthocyanin backing layer. LSI-GCA exhibited excellent biocompatibility and antibacterial and antioxidant properties by activating the NRF2 pathway and inhibiting pro-inflammatory factors. Animal experiments confirmed that LSI-GCA significantly accelerated wound healing in a diabetic model and promoted angiogenesis, collagen deposition, and M2 macrophage polarization. Genomic and network pharmacological analyses revealed a multi-target synergistic mechanism involving the modulation of EGFR/VEGF signaling to promote proliferation, inhibition of inflammatory pathway (NF-κB), and repair of DNA damage through upregulation of BRCA1/2. This study provides an integrated "monitoring-treatment" strategy for diabetic wounds, offering great potential for clinical transformation and personalized treatment.

Graphical abstract

1. Introduction

The global prevalence of diabetes continues to rise, accompanied by an increasing incidence of chronic wounds. Epidemiological data show that 19–34 % of patients with diabetes develop nonhealing wounds [1,2], primarily due to excessive oxidative stress and impaired vascular regeneration, both of which hinder effective tissue repair. Unlike acute wounds, which are acidic (pH 4–6) as a result of neutrophil activity, chronic diabetic ulcers tend to be alkaline (pH 7–9) due to tissue necrosis and microbial colonization [3,4]. This alkalinity promotes matrix degradation and bacterial growth. Thus, real-time pH monitoring is essential for assessing infection risk and the stage of wound healing.

Polymer micelles form spontaneously when amphiphilic block copolymers self-assemble in aqueous solutions [5,6]. Because of the co-existence of both hydrophilic and hydrophobic chains, polymer micelles can improve the solubility of poorly water-soluble drugs and enhance their bioavailability [7]. In addition, a polymer micelle drug delivery system can increase drug stability, reduce toxicity, and provide target delivery under specific conditions [8]. Lycium barbarum polysaccharides (LBP) have been shown to enhance the function of various immune cells, promote the generation of cytokines, and exhibit in vitro antioxidant capacity [9]. However, their strong water solubility is not suitable for loading and delivering hydrophobic drugs. To address this issue, researchers have modified LBP with hydrophobic molecules, such as stearic acid (SA) [10,11]. Isoliquiritigenin (ISO), a natural isoflavone extracted from licorice, exhibits antioxidant, anti-inflammatory, and antibacterial properties [12,13]. It can regulate the TLR4/NF-κB signaling pathway to induce antioxidant effects and may further modulate the inflammatory response by influencing cytokine levels, such as IL-13 [14,15].

Microneedles are a transdermal drug delivery system capable of continuously delivering drugs. They physically penetrate the stratum corneum, allowing for direct access to deeper layers of the skin. This mechanism enhances drug absorption while minimizing pain and discomfort compared to traditional needle-based injections [16,17]. However, hydrophobic drugs have a low hydrogel affinity, which can lead to poor drug-loading efficiency and rapid drug release. To solve this problem, hydrophobic drugs are encapsulated in cyclodextrin, micelles, or liposomes before being loaded into hydrogels [18,19]. Stimulus-responsive hydrogel drug delivery systems can be designed based on the specific microenvironments of diseases and treatment needs. The stimuli for these systems can be exogenous (e.g., light, electric, or magnetic) or endogenous (e.g., temperature or pH) [20]. Hydrogel microneedles incorporating functional materials and drugs play key roles in wound healing by promoting hemostasis, preventing infection, and accelerating tissue repair.

Gelatin Methacryloyl (GelMA) is a photosensitive biomaterial synthesized by reacting methacrylic anhydride (MA) with gelatin (Gel) [21]. It exhibits excellent biocompatibility and degradability [22] and can mimic the natural extracellular matrix, supporting cell adhesion, growth, differentiation, and migration. As a result, GelMA is widely used for the regeneration and repair of soft and hard tissues. Chitosan is a natural polysaccharide derived from shells of shrimp, crabs, lobsters, and other animals [23]. It is widely used in medical dressings, disinfectants, and other applications due to its broad-spectrum antibacterial activity that inhibits the growth of various bacteria and fungi [24].

Anthocyanins, also known as flower pigments, are water-soluble natural pigments widely present in plants [25,26]. They belong to the bioflavonoid class and are typically obtained through the hydrolysis of anthocyanin glycosides. Like other bioflavonoids, anthocyanins exhibit significant free radical-scavenging and antioxidant capabilities [27]. Notably, the color of anthocyanins changes in response to environmental pH, shifting from red at pH levels below 7 to purple between pH 7 and 8. When the pH exceeds 11, the solution turns blue. During the healing process of chronic wounds in patients with diabetes, the pH of the microenvironment undergoes dynamic changes, gradually shifting from the weakly acidic state of acute wounds to the weakly alkaline state of chronic wounds [28]. Therefore, the pH-responsive color-changing property of anthocyanins makes them a promising candidate for real-time pH monitoring. This feature is expected to provide an intuitive and convenient means for assessing the healing status of chronic diabetic wounds [29,30].

In this study, we developed an intelligent microneedle system (LISI-GCA) that integrated pH monitoring and treatment functions. This system incorporated ISO into hydrophobically modified LBP stearate polymer micelles (LBP-SA) and constructed microneedle tips through photo-crosslinked GelMA to achieve efficient drug delivery. The dorsal layer consists of a chitosan/anthocyanin complex, enabling real-time pH monitoring in the wound environment. The experiments showed that LSI-GCA significantly promoted targeted drug release, achieving a cumulative drug release rate of 80 % within 12 h in an alkaline solution. The porous structure of LSI-GCA enhanced the antioxidant capacity and accelerated the migration of vascular endothelial cells and angiogenesis. Genomic and network pharmacological analyses revealed the multi-target synergistic effects of ISO, including inhibition of TNF-α/IL-6 inflammatory axis, activation of NRF2/ARE pathway, and regulation of EGFR/VEGF signaling. This microneedle system integrates pH-responsive monitoring and intelligent drug delivery functions, providing a new "diagnosis and treatment" strategy for diabetic wound management. It holds significant potential for clinical transformation and personalized treatment (Fig. 1).

Fig. 1.

Schematic diagram of the preparation principle and action mechanism of the LSI-GCA microneedle system. The drug-loaded micelles (LBP-SA-ISO) were formed by stearic acid (SA)-modified LBP loaded with ISO and cross-linked with GelMA to prepare microneedle tips. The backing layer was composed of a chitosan/anthocyanin complex. The system enables wound microenvironment monitoring via anthocyanin color change and intelligent drug release, with ISO accelerating drug release under alkaline conditions through pH responsiveness. It also promotes diabetic wound healing through antioxidant, anti-inflammatory, and pro-angiogenic effects. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

2. Materials and methods

2.1. Materials

Stearic acid (SA) and chitosan were purchased from Maclin (Shanghai). Lycium barbarum polysaccharide (LBP), isoliquiritigenin (ISO), and anthocyanidin were purchased from Source Leaves (Shanghai). GelMA was obtained from MedChemExpress, and high-glucose DMEM medium was supplied by Sewell. Calcein-AM/PI double staining reagent for live and dead cells, CCK-8 cell proliferation detection kit, DCFH-DA reactive oxygen species detection probe, JC-1 mitochondrial membrane potential detection probe, and bacterial live/dead staining kit (DMAO/PI) were purchased from Biyantian. Human umbilical vein endothelial cells (HUVEC), mouse fibroblasts (3T3), and mouse macrophage RAW264.7 cells were supplied by Wuhan Pulisai Biotechnology Co., Ltd. Polyclonal antibodies against CD206 and CD86, plus CoraLite488-conjugated goat anti-rabbit IgG(H + L), were bought from Proteintech. Rabbit polyclonal antibodies for NRF2, EGFR, TNF-α and IL-6, as well as 594-conjugated goat anti-rabbit IgG (H + L), were obtained from ABclonal. Six-week-old male ICR mice were purchased from Hangzhou Paisiao Biotechnology Co., Ltd.

2.2. Network pharmacology and molecular docking

Potential targets of ISO were predicted using PharmMapper and SwissTargetPrediction and then intersected with diabetic wound-related targets obtained from GeneCards/OMIM to identify the key targets (ESR1, EGFR, and IGF1R). The STRING-based PPI network and Cytoscape topology analyses prioritized the core nodes. GO/KEGG enrichment via DAVID highlighted the effects of ISO on the regulation of inflammatory and angiogenic pathways. AutoDock Vina was used to validate the binding of ISO to targets (e.g., TNF-α and IL-6), and the molecular interactions were visualized by PyMOL. Three independent replicates were conducted, confirming the multi-target synergy of ISO via NF-κB inhibition, VEGF activation, and antioxidant pathways.

2.3. Hydrophobic modification of LBP-SA and construction of LBP-SA-ISO

An appropriate amount of SA was dissolved in 8 mL of dimethyl sulfoxide (DMSO) with N, N'-dicyclohexylcarbodiimide (DCC). The solution was stirred in a water bath for 1 h to activate SA. An appropriate amount of LBP and 4-Dimethylaminopyridine (DAMP) were added to DMSO and stirred for 48 h to synthesize LBP-SA. DMSO was replaced with deionized water using the solution displacement method, and the LBP-SA solid powder was obtained by freeze-drying. The feeding molar ratio of SA:DCC: LBP polysaccharide: DAMP was 10:0.15:1:0.25. LBP-SA was characterized by ultraviolet (UV) spectroscopy, infrared (IR) spectroscopy, and ¹H nuclear magnetic resonance (NMR) spectroscopy.

First, 5 mg of LBP-SA and 1 mg of ISO were accurately weighed and dissolved in an appropriate volume of DMSO to form a homogeneous solution after stirring. Subsequently, the mixed solution was transferred into a dialysis bag and dialyzed against deionized water (ddH₂O). Following 24 h of magnetic stirring, the LBP-SA-ISO nanoparticles were successfully obtained. The particle size and morphology of the nanoparticles were further characterized using transmission electron microscopy (TEM).

2.4. Preparation and characterization of LSI-GCA

LBP-SA-ISO and 20 % GleMA (0.5 % lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP)) were used as needle tip contents and vacuum-dried for 15 min, followed by UV irradiation for 8 min. The back of the needle tip was treated with a 1 % chitosan and anthocyanin solution (0.5 % LAP), followed by vacuum-drying for 15 min and UV irradiation for 10 min. Finally, the tip was sealed and dried in blue silica gel spheres for 2 h and then de-molded to obtain LSI-GCA. The structure of LSI-GCA was observed using an optical microscope and a scanning electron microscope (SEM), and the changes of LSI-GCA before and after drug release were compared. The tensile strength of the microneedles was determined using an electronic universal material testing machine.

2.5. Investigation on the pH response of LSI-GCA

The pH-dependent release profile of LSI-GCA was evaluated by dialysis in PBS buffers (pH 5–8, 37 °C), with samples collected at predetermined intervals (0, 0.25, 0.5, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, and 12 h). At each time point, 1.0 mL of dialysate was withdrawn and replaced with fresh PBS to maintain sink conditions, followed by absorbance measurement at 367 nm to construct release kinetics. To assess pH responsiveness, LSI-GCA was incubated in PBS (pH 5–10), demonstrating distinct color variations that correlated with pH changes. Furthermore, reversible cycling tests were conducted on the response of LSI-GCA to pH changes. First, we added LSI-GCA to PBS buffer with pH 5 and record the initial color. We added an appropriate alkaline solution, adjust the pH value of the solution to 8, and observe and record the color change. The solution was adjusted back to pH 5 with an appropriate acidic solution and the color recovery situation was recorded. To evaluate the pH sensing ability of LSI-GCA in vivo, we conducted a controlled experiment using a diabetic mouse model. Two days after the treatment, we systematically recorded and compared the chromaticity responses in the treated skin areas between the LSI-GCA experimental group and the PBS control group.

2.6. Biocompatibility test

Four experimental groups were established: untreated control, GCA, LS-GCA, and LSI-GCA microneedles. For cytocompatibility testing, material extracts were prepared by incubation in DMEM (37 °C, 24 h), filtered (0.22 μm), and assessed using calcein AM/PI staining and CCK-8 assays (n = 3). Blood compatibility was evaluated by measuring hemolysis rates (<5 % threshold) after incubating extracts with heparinized whole blood (37 °C, 60 min). For in vivo safety, mice received material treatments or saline (control), followed by histopathological examination of major organs via H&E staining. All experiments were performed in triplicate, demonstrating the materials' biocompatibility and safety profile.

2.7. Cell migration, tube formation, and proliferation experiments

The migration of HUVECs was tested using a standard scratch assay. After seeding in a 6-well plate to 90 % confluence, scratches were made on the well. GCA, LS-GCA, and LSI-GCA extracts were added, and wound closure was monitored at 0, 6, 12, and 24 h and quantified using ImageJ software. To assess angiogenesis, HUVECs in the extract solutions were seeded, and their tubular structures were observed after 8 h of incubation. For cell proliferation tests, HUVECs treated with the extracts were subjected to antibody incubation, DAPI staining, and Ki67 positive-rate analysis using confocal microscopy. Blank controls were included in all experiments.

2.8. Antibacterial experiment

The antibacterial efficacy of GCA, LS-GCA, and LSI-GCA was comprehensively evaluated through a multi-dimensional approach, including plate coating counting, live/dead staining (DMAO/PI), membrane potential analysis (DiSC3(5) probe), biofilm clearance (crystal violet staining), and protein leakage assays (Coomassie brilliant blue). Bacterial suspensions (1 × 10⁶ CFU/mL) were co-cultured with the extracts for 24 h, followed by serial dilution and plating to quantify CFUs, while fluorescence microscopy distinguished live (green) and dead (red) bacteria. Membrane potential changes were assessed via fluorescence intensity after DiSC3(5) staining, and biofilm disruption was measured by crystal violet absorbance at 570 nm. Protein leakage, indicating membrane damage, was quantified at 595 nm using a standard curve. In vivo validation employed bacterial infection models, with therapeutic effects assessed through clinical symptoms, tissue bacterial loads, and inflammatory markers. PBS served as the blank control in all experiments, which were performed in triplicate to ensure reproducibility.

2.9. Antioxidant experiment

The antioxidant capacities of the GCA, LS-GCA, and LS-GCA microneedles were determined using multiple methods. The extract was incubated with 20 % H₂O₂ to observe bubble formation. The intracellular ROS level was quantified by a DCFH-DA probe (10 μM) using fluorescence microscopy (488/525 nm), flow cytometry (FITC), and a microplate reader (485/535 nm), and the signal intensity was related to the ROS concentration. The mitochondrial membrane potential was measured using JC-1 probe analysis. The red/green ratio was calculated based on the red (intact cells) and green (depolarized cells) fluorescence. Lipopolysaccharide (LPS)-treated cells were used as positive controls (n = 3). All fluorescence experiments were conducted in the dark, and the data were analyzed using ImageJ and FlowJo. The results confirmed dose-dependent ROS clearance and mitochondrial protection, thereby validating the antioxidant efficacy of the microneedles.

2.10. LSI-GCA regulates the polarization of RAW264.7

RAW264.7 cells (5 × 10³ per well) were plated in confocal dishes. To trigger M1 polarization, they were exposed to GCA, LS-GCA, LSI-GCA extracts, or LPS (100 ng/mL) at 37 °C for 24 h. After fixation, permeabilization and blocking, cells were left with primary antibodies against M1 markers (CD86, iNOS) and M2 markers (CD206, Arginase-1) overnight. Post-washing, specimens were incubated for 1 h with Alexa Fluor 488-linked goat anti-rabbit IgG (H + L) secondary antibody (1:500 dilution), then nuclei were counterstained with DAPI. Confocal microscopy measured fluorescence intensity, analyzing five random fields per group. All tests ran in triplicate following standard protocols to ensure consistent results.

2.11. Cell immunofluorescence experiment

Immunofluorescence was used here to investigate NRF2's subcellular localization in RAW264.7 macrophages and HUVECs, along with EGFR expression in HUVECs. Cells were exposed to GCA, LS-GCA, or LSI-GCA extracts for 24 h, with LPS as positive control. After fixation, permeabilization and blocking, specimens were incubated with anti-NRF2 (1:200) and anti-EGFR primary antibodies at 4 °C overnight. Next, cells were stained with Alexa Fluor 594-conjugated goat anti-rabbit IgG (H + L) secondary antibody (1:500) at room temperature, followed by DAPI counterstaining for nuclei. Confocal microscopy visualized NRF2 distribution, while ImageJ software calculated nuclear-to-cytoplasmic fluorescence intensity ratios to evaluate NRF2 translocation. All tests were run in triplicate to confirm result consistency.

2.12. Western blot

Western blotting was used to detect the protein expression of EGFR, NRF2, TNF-α, and IL-6 in HUVECs. These cells had been treated for 24 h with GCA, LS-GCA, LSI-GCA extracts, or LPS (as control). The experimental steps were as follows: First, cell samples were lysed and their proteins quantified. Next, proteins were separated via sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). After transmembrane transfer, primary antibody incubation was carried out. For secondary antibody detection, HRP-labeled AffiniPure goat anti-rabbit IgG (H + L) antibody was used at a 1:500 dilution and incubated accordingly. Band intensity was analyzed with ImageJ software, and normalized using β-actin as the internal control. This experiment adhered strictly to standard operating protocols and was repeated three times. Data were analyzed via GraphPad Prism software to guarantee result reliability and reproducibility.

2.13. Establishment and treatment of chronic wound models for diabetes

All animal experiments followed ethical guidelines (Approval No. RY202505171). Diabetic mice induced by STZ, bearing 1 cm full-thickness dorsal wounds, were randomly assigned to four treatment groups (n = 8 each): GCA, LS-GCA, LSI-GCA, and saline control. Daily monitoring of wound closure was done, with quantification via ImageJ. Tissue samples taken on days 3, 6, 9, and 12 underwent H&E staining to assess healing stages. On day 12, comprehensive evaluations included Masson's trichrome staining for collagen buildup, immunohistochemistry for inflammatory markers (IL-6, TNF-α), and immunofluorescence to detect myofibroblasts (α-SMA), angiogenesis (CD31), and macrophage polarization (F4/80, CD206).

2.14. Gene sequencing

RNA sequencing was performed by Shanghai Ouyi Biomedical Technology Co., Ltd., with all experimental processes conducted in accordance with industry-standard operation protocols. The screening criteria for differentially expressed genes (DEGs) were set as: a Benjamini-Hochberg-corrected p value (FDR) < 0.05 and a gene-expression fold-change (|log2FC|) ≥ 1. For the screened DEGs, KEGG pathway enrichment and GO functional classification were performed to comprehensively elucidate the biological processes, molecular functions, and pathway involvement of these genes. A protein interaction network was constructed based on the string database (https://string -db.org/). The Cytoscape software (version 3.9.1) was used for topological parameter analysis to identify core hub genes and key regulatory modules, thus uncovering the synergistic mechanism of genes associated with apoptosis inhibition.

2.15. Statistical analysis

All data are expressed as mean ± SD, with a minimum sample size of n ≥ 3. One-way analysis of variance (ANOVA) was used for statistical comparisons between groups. A P-value <0.05 was defined as statistically significant.

3. Results and discussion

3.1. Research on network pharmacology and molecular docking

The potential targets of ISO were predicted using PharmMapper and SwissTargetPrediction, and their intersections with chronic wound-related targets in diabetes were analyzed from the GeneCards and OMIM databases to identify 45 common targets (Fig. 2a). The PPI network constructed by the STRING database shows that the core targets include EGFR, IGF1R, PPARG, IL-6, TNF-α, etc. (Fig. 2b). These targets form functional modules through highly interconnected nodes, suggesting that they collaborate to regulate the key biological processes of chronic wound healing in diabetes. KEGG enrichment analysis revealed significant enrichment of DEGs in the NF-κB, VEGF, IL-17, and chemokine signaling pathways (Fig. 2c). GO analysis showed DEGs were primarily involved in negative regulation of inflammatory responses, angiogenesis, extracellular matrix organization, and antioxidant activity (Fig. 2d). ISO bubble plots, covering biological processes, cellular components, and molecular functions, visualize the enrichment analysis results. Network pharmacology findings of ISO are comprehensively presented, spanning target relationships, network construction, functional enrichment, and pathway analysis. Pathway analysis demonstrates that ISO triggers gene expression, modulates vascular permeability, and enhances cell migration, proliferation, and survival via the VEGF and EGFR pathways—pathways critical for normal vascular development and homeostasis maintenance (Figs. S1, S2, S3). Molecular docking of the core targets was carried out through AutoDock Vina, and the binding energies were as follows: IL-6: −1.05 kcal/mol, IL-10: −2.99 kcal/mol, TNF-α: −2.88 kcal/mol, PPARG: −2.02 kcal/mol, EGFR: −2.27 kcal/mol, IGF1R: −1.93 kcal/mol, ESR1: −2.54 kcal/mol, CXCR4: −1.16 kcal/mol, PIK3R1: −1.65 kcal/mol, and IL-17: 1.55 kcal/mol (Fig. 2e). The molecular docking results reveal that ISO promotes the healing of diabetic chronic wounds through the synergistic action of multiple targets (EGFR, TNF-α, and VEGF). Therefore, this study focused on introducing LSI-GCA as a delivery system for ISO to evaluate the effect of the pH-responsive hydrogel microneedle in promoting diabetic wound healing.

Fig. 2.

Network pharmacology and molecular docking analysis of ISO in diabetic wounds. (a) Venn diagram of drug targets (blue) versus disease targets (yellow). (b) Protein-protein interaction (PPI) network based on intersection targets. (c) Gene ontology (GO) enrichment analysis. (d) KEGG pathway analysis. (e) Molecular docking results. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.2. Characterization of LBP-SA and LBP-SA-ISO

Electron microscopy observations revealed that the blank micelles LBP-SA and drug-loaded micelles LBP-SA-ISO exhibited regular spherical morphologies with minimal agglomeration (Fig. 3a). Their particle sizes were 180 and 230 nm, respectively, and their absolute values of zeta potentials were greater than 20 mV (Fig. 3b, 3c, S4a). According to the absorbance-concentration standard curve of ISO, the encapsulation efficiency of LBP-SA-ISO was 95 %, and the drug loading was 0.94 mg (Fig. S4b and S4c). LBP-SA-ISO showed a larger particle size and a higher absolute value of zeta potential than LBP-SA. This is because ISO loading introduces additional steric hindrance and surface charges, improving the stability of the micelles.

Fig. 3.

Characterization and pH response of LBP-SA and LBP-SA-ISO. (a) Transmission electron microscopy (TEM) images of blank micelles (LBP-SA) and drug-loaded micelles (LBP-SA-Iso), scale bar: 200 nm. (b–c) Particle size distribution and Dundahl effect of LBP-SA and LBP-SA-ISO. (d–f) Ultraviolet (UV), nuclear magnetic resonance (NMR), and infrared (IR) spectra of LBP-SA. (g–h) Appearance and scanning electron microscopic (SEM) images of LSI-GCA. (i) Different colors on the back of the microneedle when immersed in buffer solutions with pH values ranging from 5 to 10. (j–k) Cumulative release curves of ISO from LBP-SA-ISO and LSI-GCA in buffer solutions at pH 5 and pH 8. (l) Mechanical properties of microneedles. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

The structural characteristics of LBP-SA and LBP-SA-ISO were systematically analyzed using multidimensional characterization techniques. Their UV spectra exhibited characteristic absorption peaks of LBP-SA in the range of 200–800 nm, confirming the existence of the π→π transition of the benzene ring and the n→π transition of the carboxyl group of SA (Fig. 3d). ¹H NMR analysis showed chemical shifts for the benzene ring proton of SA (δ 6.5–8.0 ppm) and the sugar ring proton of LBP (δ 3.0–5.5 ppm), confirming that a stable complex formed by hydrogen bonding and hydrophobic interactions between SA and LBP (Fig. 3e). The changes in peak position revealed the intermolecular interaction between SA and LBP. In the IR spectra, the carboxyl peak (1700 cm⁻¹) in SA and the broadening of the O-H peak in the LB-SA complex confirmed the formation of a hydrogen bond network, while the peaks in the fingerprint region (1000-1300 cm⁻¹) suggested that there might be an interaction between the ester bonds and the C-O bonds of SA (Fig. 3f). These results elucidate the mechanism of ester bond formation in LBP-SA at the molecular level.

3.3. Characteristics of LSI-GCA

LSI-GCA was prepared using GleMA as the needle tip and chitosan/anthocyanin as the back layer. The microneedle tips were clearly observed at different angles and magnifications, forming a regular array with a uniform distribution (Fig. 3g). LSI-GCA had a porous structure, which increased the specific surface area of the microneedles, providing spatial advantages to increase the contact area and enhance the action effects on tissues such as the skin (Fig. 3h), the microstructure collapse of LSI-GCA after drug release (Fig. S4d). The load-displacement curve (Fig. 3l) showed that the microneedles could withstand a loading force of 0.2–0.8 N within the displacement range of 0–10 mm. This mechanical property indicates that the microneedles have sufficient rigidity to penetrate the stratum corneum of the skin (which usually requires 0.1–0.5 N) while maintaining adequate flexibility to minimize the risk of fracture during use.

3.4. The pH response capability of LSI-GCA

The color of the LSI-GCA microneedles varied from pink to brown at different pH values (pH = 5–10), indicating that the anthocyanins in LSI-GCA microneedles were sensitive to pH changes (Fig. 3i). The cumulative release rate from LBP-SA-ISO and LSI-GCA reached approximately 60 % within 12 h at pH 8, significantly higher than that at pH 5 (approximately 40 %) (Fig. 3j, 3k). The release curves were fitted to the first-order kinetic model, zero-order kinetic model, and Higuchi model. For the LBP-SA-ISO at pH = 6, the correlation coefficients (R²) of the fittings were 0.91, 0.66, and 0.79 for the first-order, zero-order, and Higuchi models, respectively, indicating that the release followed the first-order kinetic release mode, where the release rate was proportional to the drug concentration. At pH = 6, LBP-SA-ISO showed a higher R² value of 0.89 for the Higuchi model compared to the zero-order kinetics (R² = 0.76) and the first-order kinetics (R² = 0.78), suggesting that the drug release mechanism is diffusion-dominated. Similarly, for LSI-GCA, the release at pH = 8 conformed to the first-order kinetic pattern (concentration-dominating), while the drug was released via a diffusion mechanism at pH = 6 (Fig. S5). Reversible cycling tests confirmed LSI-GCA's excellent color-switching reversibility between pink (pH 5) and blue-purple (pH 8), maintaining stability over multiple cycles (Fig. S6a). In diabetic mice, LSI-GCA showed clear color changes corresponding to actual skin pH variations, demonstrating reliable in vivo pH sensing capability (Fig. S6b).

This pH-dependent release behavior may arise from the ionization states of the pH-sensitive groups in the microneedle materials, including the carboxyl groups of stearic acid (SA) and amino groups of GelMA [31]. Due to the lack of potentiometric titration equipment and time constraints, we were unable to directly determine their pKa values. However, existing literature provides consistent references for their typical dissociation constants: the pKa of SA carboxyl groups is generally reported to be 4.75–5.19 in aqueous systems [32], while the pKa of GelMA amino groups is inferred to be 8.5–8.8 based on the isoelectric point of type A gelatin and related methacrylation studies. Within our experimental pH range (5.0–8.0), SA carboxyl groups would undergo significant deprotonation (-COOH→-COO^-) as pH exceeds their pKa, enhancing matrix hydrophilicity and accelerating micelle disassembly for drug release [33]. In contrast, GelMA amino groups remain weakly protonated within this range (below their pKa), avoiding interference with the carboxyl-mediated pH response [34]. In addition, the molecular structure rearrangement of anthocyanins at different pHs leads to pH-sensitive color change.

3.5. Biocompatibility

Live-dead staining of HUVECs and 3T3 cells revealed no significant increase in dead cells in GCA, LS-GCA, and LSI-GCA treatment groups compared to controls, while live cell counts remained consistently high (Fig. S7a–d). This indicates that when the materials come into direct contact with cells, they do not cause significant cell death, affirming their cytocompatibility. The CCK-8 results showed that the cell viabilities of the GCA, LS-GCA, and LSI-GCA groups were close to that of the control group, indicating that the materials had minimal impact on normal physiological functions and the cells maintained high activity (Fig. S7f and S7g). In the hemolysis experiment, the hemolysis rates of the PBS, GCA, LS-GCA, and LSI-GCA groups were extremely low (<5 %) with no significant differences among them, indicating that the materials did not cause red blood cell rupture and had good blood compatibility (Fig. S7e). HE staining of the heart, liver, spleen, lungs, and kidneys of treated mice showed no lesions (Fig. S8). In conclusion, these materials demonstrate excellent biocompatibility at both cellular and blood levels, laying a solid foundation for their subsequent applications in biomedical fields.

3.6. Cell proliferation and angiogenesis

In this study, the effects of LSI-GCA on cell migration, angiogenesis, and proliferation were evaluated. The experimental results showed that LSI-GCA exhibited significant triple biological effects. The cell migration promotion rate was 60 % (Fig. 4a, 4b, S9), which was twice that of the control group. The number of vascular connection points induced by LSI-GCA far exceeded that of the control group (Fig. 4c, 4d). Furthermore, the Ki67 test confirmed the proliferation-promoting ability of LSI-GCA, in which the density of positive cells was three times higher than that of the control (Fig. 4e, 4f). The three-dimensional porous structure of LSI-GCA provides an ideal scaffold for cell migration and angiogenesis. Notably, the trend of the three biological effects followed a consistent order (LSI-GCA > LS-GCA > GCA > control), indicating a clear dose-dependent relationship between the degree of modification of the material and the release of ISO. These findings provide important inspiration for the development of new tissue engineering materials. By precisely regulating the physicochemical properties, it is possible to achieve synergistic regulation of various cellular behaviors.

Fig. 4.

Angiogenesis, cell proliferation, and antimicrobial capacity of LSI-GCA. (a–b) Cell migration assay and statistical results of LSI-GCA in HUVECs. Scale: 100 μm (c–d) HUVEC ring formation assay and statistical results promoted by LSI-GCA. Scale: 100 μm (e–f) Ki67 immunofluorescence staining. (DAPI: blue, Ki67: green) Scale: 100 μm (g) Plate coating experiments for Escherichia coli and Staphylococcus aureus. Scale: 2 cm (h–i) Statistical analysis of the results of plate coating experiments compared with the control group. (j) Live-death staining of Escherichia coli and Staphylococcus aureus. Scale: 100 μm (k–l) Statistical analysis of the results of live and dead staining experiments compared with the control group. (m–o) Analysis of the antibacterial effect and statistical results of LSI-GCA on animal infection models. Scale: 2 cm Data are presented as mean ± standard deviation (n = 3). (Statistical differences: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, NS, no statistical significance). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.7. Antibacterial effects

In the bacterial plate coating experiment, the E. coli and Staphylococcus aureus colonies in the control group grew vigorously, whereas the number of colonies in the GCA, LS-GCA, and LSI-GCA groups decreased significantly, especially in the LSI-GCA group (Fig. 4g, 4h, 4i). The results of live and dead bacterial staining indicated that the control group was primarily composed of live bacteria. After treatment, the number of dead bacteria (red fluorescence) increased, and the proportion of red fluorescence in the LSI-GCA treatment group was the highest (Fig. 4j, 4k, 4l). Biofilm removal assays (crystal violet staining, OD570) demonstrated that GCA, LS-GCA, and LSI-GCA significantly reduced biofilm formation compared to Control, with LSI-GCA being the most effective (Fig. S10a and S10d). Protein leakage assays (OD595) confirmed that the extracts disrupted bacterial membrane integrity, inducing intracellular protein release, with LSI-GCA causing the highest leakage (Fig. S10b and S10e). Membrane potential analysis (DiSC3(5) probe) revealed that all three extracts significantly increased bacterial fluorescence intensity compared to PBS, with LSI-GCA showing the strongest effect (Fig. S10c and S10f). In a murine infection model, GCA, LS-GCA, and LSI-GCA significantly alleviated clinical symptoms and reduced bacterial load compared to Control, with LSI-GCA exhibiting the strongest antibacterial effect (Fig. 4m, 4n, 4o). Overall, GCA, LS-GCA, and LSI-GCA can inhibit the growth of E. coli and Staphylococcus aureus and promote the death of bacteria by destroying their morphological structure. Among them, LSI-GCA exhibits the best antibacterial effect due to the presence of ISO, which has been reported to be effective against a variety of bacteria and plant pathogens [35]. This antibacterial mechanism may be related to the ability of ISO to disrupt bacterial cell membranes and enzymatic activity. When ISO is loaded onto nanoparticles, a synergistic effect may further enhance its antibacterial properties.

3.8. Antioxidant effects

The antioxidant performance of the GCA, LS-GCA, and LSI-GCA microneedles was systematically evaluated through in vitro experiments using DCFH-DA and JC-1 probes. First, the in vitro antioxidant experiments showed that a few small bubbles were generated in the GCA and LS-GCA groups, whereas larger bubbles appeared in the LSI-GCA group (Fig. S11). The DCFH-DA results showed that LPS, GCA, LS-GCA, and LSI-GCA treatments significantly reduced intracellular ROS levels compared to the control group. The lowest ROS generation was detected in the LSI-GCA group, and this trend was confirmed by fluorescence microscopy, quantitative flow cytometry, and a microplate reader (Fig. 5a–e). The JC-1 probe experiment revealed that LSI-GCA treatment significantly decreased the relative fluorescence intensity and increased the red/green ratio, suggesting that LSI-GCA could maintain the stability of the mitochondrial membrane potential most effectively (Fig. 5f, 5g). These results indicate that microneedles exert antioxidant effects by eliminating ROS and stabilizing mitochondrial function, and loading ISO significantly enhances the antioxidant efficacy of the material (LSI-GCA > LS-GCA > GCA).

Fig. 5.

Antioxidant activity of LSI-GCA. (a) Fluorescence distribution of intracellular ROS (green: DCFH-DA). Scale bar: 50 μm. (b) ROS detected by flow cytometry (FITC channel). (c–e) Comparative analysis of the fluorescence microscopy, flow cytometry (FITC channel), and microplate reader (485/535 nm) results. (f) Mitochondrial membrane potential measured by JC-1 (red represents normal membrane potential, green represents depolarization). Scale bar: 100 μm. (g) Statistical results of JC-1 fluorescence measurement ratio. Data are presented as mean ± standard deviation (n = 3). (Statistical differences: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, NS, no statistical significance). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.9. Microregulation of macrophage polarization

RAW264.7 cells treated with LPS showed significant polarization. However, when cells were co-incubated with the extracts of GCA, LS-GCA, and LSI-GCA, cell polarization was significantly inhibited. As a specific marker of M1 macrophage polarization, the CD86 and iNOS expression level can effectively indicate the degree of M1 polarization [24]. The expression of CD206 and Arginase 1 are important indicators of the polarization of M2 macrophages [25]. As shown in the results, the LPS-stimulated control group displayed markedly higher fluorescence intensities for CD86 and iNOS (Fig. 6a, 6b, S12a, S12b). In contrast, treatment with GCA, LS-GCA, and LSI-GCA led to a significant decrease in CD86 fluorescence and a notable increase in CD206 and Arginase 1 fluorescence signals, with LSI-GCA exerting the most prominent regulatory effect (Fig. 6c, 6d, S12c, S12d)). These findings indicate that LSI-GCA can regulate cellular inflammatory responses by suppressing macrophage polarization toward the pro-inflammatory M1 subtype and enhancing polarization toward the anti-inflammatory M2 subtype.

Fig. 6.

Anti-inflammatory mechanism and regulation of polarization of macrophages by LSI-GCA. (a) Immunofluorescence staining for M1 marker CD86 (green: CD86, blue: DAPI). Scale bar: 100 μm. (b) Quantitative statistics of CD86. (c) Immunofluorescence staining of M2 marker CD206 (green: CD206, blue: DAPI). Scale bar: 100 μm. (d) Quantitative statistics of CD206. (e) NRF2 nuclear translocation assay (red: NRF2, blue: DAPI). Scale bar: 100 μm. (f) Quantitative statistics of NRF2 nuclear translocation assay. Data are presented as mean ± standard deviation (n = 3). (Statistical differences: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, NS, no statistical significance). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.10. In vitro anti-inflammatory and angiogenic promotion mechanisms

In RAW264.7 cells and HUVECs, both immunofluorescence and Western blot results showed that the expression of NRF2 was significantly upregulated after GCA, LS-GCA, and LSI-GCA treatments. NRF2 is a key transcription factor involved in oxidative stress. After activation, it regulates the expression of a series of downstream antioxidant proteins, enhances the antioxidant capacity of cells, and reduces the inflammatory damage caused by oxidative stress (Fig. 6, Fig. 7a-d, 7e(ii), 7g). In HUVECs, microneedle treatment significantly downregulated the expression of inflammatory factors TNF-α and IL-6. As shown in the Western blot results and quantitative analysis, the protein bands of these two inflammatory factors became thinner in the treatment group, and their expression levels decreased, thereby alleviating the inflammatory response (Fig. 7e(ⅲ-ⅳ), 7h, 7i). GCA, LS-GCA, and LSI-GCA treatments also inhibited EGFR expression in HUVECs (Fig. 7e(i), 7f). Excessive EGFR activation is associated with abnormal angiogenesis. The inhibition of its expression may help regulate angiogenesis-related signaling pathways and maintain a normal angiogenic state. While not directly depicted in the figure, prior relevant research implies that LSI-GCA might activate the extracellular signal-regulated kinase (ERK1/2) and heme oxygenase-1 (HO-1) signaling pathways through reactive oxygen species production and enhanced secretion of VEGF. Additionally, LSI-GCA may enhance the migration and proliferation of endothelial cells [36], thereby further promoting angiogenesis. Overall, these results suggest that GCA, LS-GCA, and LSI-GCA may exert antioxidant and anti-inflammatory effects by regulating related pathways, such as NRF2, EGFR, and inflammatory factors.

Fig. 7.

Anti-inflammatory and pro-angiogenic action mechanisms of LSI-GCA. (a) Confocal microscopy images of NRF2 in HUVECs (red: NRF2; blue: DAPI). Scale bar: 100 μm. (b) Quantitative assessment of NRF2 fluorescence intensity. (c) Confocal microscopy images of EGFR in HUVECs (red: EGFR; blue: DAPI). Scale bar: 100 μm. (d) Quantitative assessment of EGFR fluorescence intensity. (e) Western blot analyses showing the protein expression levels of NRF2, EGFR, TNF-α, and IL-6 in HUVECs, with β-actin serving as the internal reference. (f–i) Comparative analyses of NRF2, EGFR, TNF-α, and IL-6 expression. Data are expressed as mean ± standard deviation (n = 3). (Statistical significance: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; NS, not significant). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.11. Promotion of wound healing

STZ-induced diabetic mice were administered GAC, LS-GAC, and LSI-GAC treatments (Fig. 8a). Wound healing was delayed in the control group, whereas the LSI-GCA group exhibited a significantly higher wound closure rate on day 12 post-treatment compared to the controls (Fig. 8b, 8c, S13a, S13b). Specifically, the wound defect width was reduced to 50 % in the LSI-GCA group, versus 100 % in the control group, demonstrating that LSI-GCA markedly promotes wound healing. HE staining showed that compared with the control group, the tissues of the GCA and LS-GCA treatment groups exhibited accelerated repair on the 3rd, 6th and 9th days, confirming their effectiveness in promoting wound tissue repair and reconstruction (Fig. 8d, 8e, S14). Masson's trichrome staining showed that the collagen fiber contents in the GCA, LS-GCA, and LSI-GCA treatment groups increased significantly. The deposition amount increased by 40 %, and the more orderly arrangement was conducive to structural reconstruction and strength recovery of the tissue at the wound site (Fig. S13c and S13d). The expressions of inflammatory factors TNF-α and IL-6 in the GCA, LS-GCA, and LSI-GCA treatment groups were significantly lower than those in the control group. The reduction in inflammatory factors can alleviate the local inflammatory response and create a favorable microenvironment for wound healing (Fig. 9a, 9b, 9e, 9f).

Fig. 8.

Wound healing efficacy of LSI-GCA in a diabetic mouse model. (a) Animal modeling and experimental treatment protocols. (b) Wound healing photographs taken on days 0, 3, 5, 7, 9, and 12. Scale bar: 5 mm. (c) Wound healing process simulation. (d) HE staining of mouse skin tissue adjacent to wounds. Scale bar: 1 mm. (e) Quantitative analysis of wound width based on HE staining. Data are expressed as mean ± standard deviation (n = 3). (Statistical significance: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; NS, not significant).

Fig. 9.

LSI-GCA enhances wound angiogenesis and downregulates inflammatory factor expression. (a) TNF-α expression in skin wound tissues. Scale bar: 100 μm. (b) IL-6 expression in skin wound tissues. Scale bar: 100 μm. (c) Immunofluorescence staining of CD31 and α-SMA. Scale bar: 100 μm. (d) Immunofluorescence staining of CD206 and F4/80. Scale bar: 100 μm. (e) Quantitative analysis of TNF-α expression. (f) Quantitative analysis of IL-6 expression. (g) Quantitative analysis of CD31 expression. (h) Quantitative analysis of CD206 expression. Data are expressed as mean ± standard deviation (n = 3). (Statistical significance: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; NS, not significant).

Immunofluorescence assays revealed that the treatment groups had significantly higher expression of CD31, a marker associated with angiogenesis, than the control group (Fig. 9c, 9g). Elevated CD31 expression indicates that microtargeting can stimulate angiogenesis at the wound site, thereby supplying nutrients and oxygen to support healing. α-smooth muscle actin (α-SMA), a specific marker for vascular smooth muscle cells, is critical for regulating vascular maturation and preserving vascular structural stability. Additionally, in the GCA, LS-GCA, and LISI-GCA treatment groups, we observed a marked upregulation of CD206 (M2 macrophage marker) accompanied by altered expression patterns of F4/80 (Fig. 9d, 9h). These results suggest that microneedles can inhibit inflammatory response and accelerate tissue repair by promoting macrophage polarization toward the anti-inflammatory M2 phenotype.

3.12. Tissue genomics analysis

This study compared the differential gene expression and mechanism of action between a diabetic wound treatment group (LSI-GAC) and a control group through histogenomic analysis. The results showed that in the LSI-GAC group, angiogenesis-related genes (VEGF, Ang-1) and antioxidant genes (NRF2, HO-1) were significantly upregulated, while pro-inflammatory factors (TNF-α, IL-6) and apoptosis-related genes (Caspase-3, Bax) were significantly downregulated (Fig. 10a, 10b, 10c). GO and KEGG enrichment analyses indicated that JS-Z significantly improved healing efficiency by activating the cell cycle pathway (Cyclin D1/CDK4), inhibiting the inflammatory response, enhancing DNA repair (upregulated BRCA1/2), promoting extracellular matrix remodeling (upregulated COL1A1), and synergistically optimizing the immune microenvironment (downregulated PD-L1) (Fig. 10d, 10e). Experimental verification showed that in the LSI-GAC group, the expressions of vascular markers CD31 and α-SMA increased by 20 % and 15 %, respectively. The amount of collagen deposition increased by 40 %, and the ratio of NRF2/β-actin increased as well, confirming that the healing was accelerated through multi-dimensional promotion of angiogenesis, anti-inflammation, and anti-oxidation. The changes in gene expression profiles suggest that LSI-GAC may reshape the diabetic wound repair process through metabolic reprogramming (such as activation of retinal metabolic genes), cell migration mediated by calcium signaling pathways, and the synergistic effect of EGFR/VEGF signaling. This study provides a theoretical basis for targeted multi-pathway intervention strategies.

Fig. 10.

Transcriptome analysis of RNA sequences in skin wound tissues. (a) Volcano plot showing differentially expressed genes (DEGs) in skin wound tissues. (b) Quantitative comparison of DEGs between control and treatment groups. (c) Heatmap displaying DEGs detected in control and treatment groups. (d) GO (biological process) enrichment analysis of DEGs. (e) Statistical summary of DEG enrichment in KEGG pathways.

4. Conclusion

In this study, we present a successfully synthesized hydrogel microneedle (LSI-GCA) based on LBP micelles. By loading ISO and leveraging the pH-responsive characteristics of anthocyanins, LSI-GCA exhibits pH-dependent color changes and controlled drug release behavior. It also activates angiogenesis, promotes collagen deposition, regulates macrophage polarization toward type M2, and reshapes the healing microenvironment. Genomic and network pharmacological analyses further reveal a multitarget synergy mechanism. The innovative design of LSI-GCA provides an integrated "monitoring-treatment" strategy for diabetic wound management, demonstrating great potential for clinical transformation. In the future, the long-term safety of microneedles and their applicability in other types of chronic wound healing should be explored.

CRediT authorship contribution statement

Huaqian Xue: Experimental operation, Data collection, Writing – original draft, Validation. Chen Zhang: Writing – review & editing, Methodology. Dini Lin: Investigation, Data Curation. Qiancheng Gu: Software. Chuchu Sun: Visualization, Data Curation. Xiufei Lin: Validation. Chi Zhang: Supervision. Lanjie Lei: Project administration, Funding acquisition. Liangle Liu: Project administration, Funding acquisition.

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.

Appendix A. Supplementary data

The following is the supplementary data to this article:

Data availability

The authors do not have permission to share data.