Selenium‑based nanotherapeutics promote healing of infected diabetic wounds via photothermal‑driven macrophage immunometabolic reprogramming and the NRF2/HIF‑1α axis

Abstract

Infected wounds heal considerably slowly among patients with diabetes because of biofilm barriers and immune dysregulation. Therefore, treatment strategies must simultaneously eliminate infection and accelerate healing. In this study, we present a dual-mechanism nanoplatform composed of indocyanine green (ICG)-loaded lentinan (LNT)‑functionalized selenium nanoparticles (ICG@LNT‑SeNPs). This platform disrupts the pathological cycle by integrating near‑infrared‑triggered photothermal eradication of methicillin-resistant Staphylococcus aureus (MRSA) biofilms with selenium‑driven immunometabolic reprogramming of macrophages. The selenium (Se) component facilitates metabolic adaptation and functional remodeling of pro‑inflammatory macrophages by activating the NRF2/HIF‑1α axis, inducing transition into an M2 phenotype, which is characterized by elevated expression of repair‑associated factors (e.g., Il10 and Arg1). ICG@LNT‑SeNPs also enhance the defensive properties of cellular antioxidants by upregulating the expression of Secisbp2 (a key regulator of selenoprotein biosynthesis), Gpx4, and Txnrd2. This combined antibacterial‑metabolic‑immune mechanism causes vascular endothelial cells and fibroblasts to migrate, enhancing subsequent collagen deposition and angiogenesis and, thus, accelerating wound closure. This mechanism demonstrated more mature tissue reconstruction in diabetic mice with MRSA-infected wounds. Overall, the proposed SeNPs–based therapeutic strategy promotes infected diabetic wounds healing through macrophage immunometabolic reprogramming. The findings may help identify new targets and provide insights for developing promising management approaches for chronic wounds.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12951-026-04293-6.

Introduction

Chronic diabetic wounds, especially those complicated due to methicillin-resistant Staphylococcus aureus (MRSA) biofilm infection, can increase the chances of adverse clinical outcomes including amputation and mortality [1–3]. The potent biofilm-forming capacity of MRSA enables resistance to antibiotics and evasion of immune clearance, making it a core driver of persistent infection [1, 4, 5]. This persistent biofilm burden, along with the chronic hyperglycemic environment, fuels a self-perpetuating oxidative–inflammatory loop. Sustained reactive oxygen species (ROS) production from infiltrating immune cells disrupts redox homeostasis and prolongs tissue injury [3, 4, 6–9]. Excess oxidative stress can also bias macrophage programs toward a pro-inflammatory M1 state while suppressing reparative M2 features that support healing [10–16]. Thus, we hypothesized that this dual pathology, namely persistent infection and maladaptive macrophage programs, would be reflected in human diabetic wounds. Re-analysis of single-cell transcriptomic data (GSE268834) revealed inflammatory macrophage signatures and sustained activation of NF-κB and TNF signaling programs, supporting our hypothesis and motivating the therapeutic strategy illustrated in Fig. 1. This co-occurrence may help explain why therapies that target either bacterial burden or inflammation alone often yield incomplete resolution, and it supports combined strategies that disrupt biofilms while restoring repair-supportive macrophage function. Because macrophage reprogramming is closely linked to redox balance, restoring redox homeostasis represents a rational entry point. In this context, nuclear factor erythroid 2-related factor 2 (NRF2) coordinates antioxidant gene programs, while selenoproteins, including glutathione peroxidases (GPXs) and thioredoxin reductases (TXNRDs), act as key redox effectors and metabolic regulators [17–22]. Accordingly, an effective therapeutic strategy may need to activate NRF2 signaling while supporting selenoprotein expression and function. This combined direction can strengthen macrophage antioxidant capacity and promote immunometabolic reprogramming toward repair. However, engineering a therapeutic platform that achieves this dual and coordinated modulation within the complex wound microenvironment remains a significant challenge.

Fig. 1.

Therapeutic strategy and single-cell transcriptomic landscape of diabetic wounds. (A) Schematic illustration of the proposed treatment approach. A multifunctional nanoplatform (ICG@LNT-SeNPs) combines photothermal biofilm disruption with selenium-mediated immunometabolic reprogramming targeting the NRF2/HIF-1α axis to reverse the chronic oxidative–inflammatory cycle and promote diabetic wound healing. (B) UMAP visualization of all cells colored by annotated major cell types. (C) Dot plot showing the expression of canonical marker genes used for cell-type annotation; color intensity indicates average expression and dot size represents the percentage of expressing cells. (D) UMAP projection of macrophages after re-clustering to define functional subsets. (E) Expression profiles of representative subset markers, including the pro-inflammatory marker IL1B and reparative markers TGFB1, APOE, and ITGA6. (F) Relative abundance of macrophage subsets in non-diabetic and diabetic wounds, highlighting the expansion of the IL1B⁺ subset in diabetic wounds. (G) Pathway enrichment analysis of genes upregulated in the IL1B⁺ macrophage subset, showing sustained activation of inflammatory signaling pathways, including cytokine–cytokine receptor interaction, NF-κB, and TNF signaling

Nanotechnology offers multiple strategies for wound management. Numerous studies have developed promising nanoformulations delivering antimicrobial agents or modulators of the inflammatory microenvironment [8, 9, 23–28]. In particular, near-infrared (NIR) light-responsive platforms, such as indocyanine green (ICG)-based photothermal and photodynamic systems, enable spatiotemporally controlled bacterial eradication [29–31]. However, the reactive oxygen species generated by these treatments to eliminate pathogens may also cause oxidative damage to host cells, thereby hindering tissue repair [32]. Therefore, an ideal therapeutic strategy should not only clear pathogens but also actively promote the healing process. Selenium nanoparticles (SeNPs) are widely utilized for their antioxidant and immunomodulatory properties [21, 22, 33–37]. and their ability to mimic the activity of enzymes such as glutathione peroxidase. In addition, SeNPs contain bioactive Se, which facilitates the biosynthesis of selenoproteins, leading to a pro-healing microenvironment [5, 22, 38–45]. In this study, we developed a dual-functional nanoplatform, ICG@LNT-SeNPs, rationally engineered to couple a rapid photothermal antibacterial module (ICG) with a sustained redox-regulatory core (Se). This design enables temporally coordinated pathogen eradication and subsequent host microenvironment correction. By integrating NIR-triggered biofilm disruption with selenium-mediated immunometabolic reprogramming, the platform aims to simultaneously eliminate bacterial burden and restore immune homeostasis in diabetic wounds (Scheme 1). Upon NIR irradiation, we show that it enables rapid photothermal clearance of MRSA biofilms. More importantly, we demonstrate that the selenium component orchestrates macrophage immunometabolic reprogramming by co-activating the NRF2 and HIF-1α axes, shifting macrophages toward a pro-reparative phenotype that drives angiogenesis and matrix remodeling.

Scheme 1.

Schematic diagram illustrating the preparation and working mechanism of multifunctional ICG@LNT-SeNPs for accelerating diabetic wound healing

Results

Synthesis, characterization, and photothermal properties of ICG@LNT-SeNPs

ICG@LNT-SeNPs were synthesized in a single‑step process. Dynamic light scattering (DLS) indicated a hydrodynamic size of about 115 nm (Fig. 2A). The particles showed good colloidal stability, maintaining this size for over 10 days (Fig. 2B). Fourier‑transform infrared (FTIR) spectroscopy provided evidence of successful composite formation (Fig. 2F). The spectrum retained characteristic bands of LNT‑SeNPs while displaying new signals from ICG, including C–C stretching (∼1400 cm⁻¹) and S = O vibrations. UV‑vis‑NIR spectra of the nanocomposite indicated ICG loading, with sharp peaks corresponding to free ICG (∼710 and 780 nm) merged into a broadened band (Fig. 2C). Upon ICG incorporation, the surface charge changed from − 20.5 mV (LNT‑SeNPs) to − 27.4 mV (Fig. 2D). As indicated by transmission electron microscopy (TEM) profiles, uniform spherical particles (< 100 nm) appeared, consistent with DLS data (Fig. 2E). Furthermore, a core‑shell‑like structure was confirmed by elemental mapping, which showed sulfur (S) and nitrogen (N) signals from ICG overlapping the selenium (Se) core (Fig. 2G). X‑ray photoelectron spectroscopy (XPS) analysis revealed electrostatic interaction between ICG and the nanoparticle surface, as indicated by changes in binding energy in the Se 3d and S 2p regions (Fig. 2H). TEM confirmed that nanoparticle morphology remained stable even after 12–18 h in human wound exudate (Fig. S2A). Selenium speciation analysis further indicated a gradual transformation over 36 h. A sustained‑release profile was observed, with decreasing elemental Se content and increasing concentrations of bioactive seleno‑amino acids and oxidized Se(+ IV)/Se(+ VI) species (Fig. S2B, C). In photothermal performance analysis, free ICG produced a concentration‑dependent temperature rise under 808-nm irradiation (1.0 W cm⁻²) (Fig. S2D). By contrast, ICG@LNT-SeNPs showed a stronger photothermal response under identical conditions (Fig. 2I), with higher steady-state temperatures and LNT-SeNPs alone producing negligible heating (Fig. 2J). These results indicate that the nanocarrier enhances the photothermal stability of ICG, likely by reducing dye aggregation and photobleaching, consistent with previous reports [46].

Fig. 2.

Synthesis and physicochemical characterization of the ICG@LNT-SeNPs nanoplatform. (A) Hydrodynamic particle size distributions of LNT-SeNPs and ICG@LNT-SeNPs measured by DLS. (B) Colloidal stability of ICG@LNT-SeNPs in water over 10 days. (C) UV-vis-NIR absorption spectra of free ICG, SeNPs, and ICG@LNT-SeNPs. (D) Zeta potential of LNT-SeNPs and ICG@LNT-SeNPs. (E) TEM images of ICG@LNT-SeNPs showing uniform spherical morphology with diameters < 100 nm. Scale bar, 100–200 nm. (F) FTIR spectra of ICG, LNT-SeNPs, and ICG@LNT-SeNPs. The appearance of characteristic ICG peaks (e.g., C–C at 1400 cm⁻¹, S = O) in the composite confirms structural integration. (G) Elemental mapping showing the spatial distribution of Se, S, and N. (H) High-resolution XPS spectra of Se 3d (right) and S 2p (left). The binding energy shifts and characteristic peaks confirm the chemical state of selenium and the presence of ICG-associated sulfur. (I) Concentration-dependent photothermal heating curves of ICG@LNT-SeNPs under 808-nm laser irradiation (1.0 W cm⁻²). (J) Photothermal heating profiles of PBS, LNT-SeNPs, free ICG, and ICG@LNT-SeNPs under NIR irradiation for 5 min. Data are presented as mean ± SD (n = 3)

Assessment of bacterial and biofilm activity in vitro

The antibacterial activity of ICG@LNT-SeNPs against MRSA was evaluated in vitro. For planktonic bacteria, a colony-forming unit (CFU) assay was performed under standardized conditions: ICG concentration was matched between free ICG and ICG@LNT-SeNPs groups, with vancomycin (2 µg mL⁻¹) as a control, and all treatments lasted 6 h. Without NIR irradiation, both PBS and free ICG showed negligible activity, while LNT-SeNPs and ICG@LNT-SeNPs in the dark caused only moderate growth inhibition (Fig. 3A, B). In contrast, upon NIR irradiation, ICG@LNT-SeNPs achieved significantly stronger bactericidal efficacy than free ICG, indicating a synergistic effect. Under these conditions, its potency was comparable to vancomycin (Fig. 3B), and the photothermal killing was dose-dependent (Fig. S3A, B). These findings were validated through live/dead staining. Specifically, PBS-treated samples exhibited green fluorescence (live), whereas ICG@LNT-SeNPs + NIR resulted in intense red fluorescence (dead). This difference was also derived quantitatively (Fig. 3C, D). Consistent with photothermal observations, scanning electron microscopy (SEM) revealed direct morphological damage, including severe membrane rupture and collapse in bacteria treated with ICG@LNT-SeNPs + NIR in contrast with untreated bacteria, which exhibited smooth surfaces (Fig. 3E). The platform also effectively disrupted established biofilms. Crystal violet staining showed a substantial biomass reduction only in the ICG@LNT-SeNPs + NIR group (Fig. 3F). Consistently, confocal microscopy (CLSM) imaging demonstrated that, unlike PBS, free ICG, or LNT-SeNPs treatments (which left biofilms intact and viable), ICG@LNT-SeNPs + NIR induced extensive bacterial death throughout the biofilm depth. Vancomycin, by comparison, produced only limited killing, likely due to poor penetration (Fig. 3G). Taken together, these results demonstrate that NIR-activated ICG@LNT-SeNPs exert potent, broad-spectrum antibacterial activity against both planktonic and biofilm-embedded MRSA.

Fig. 3.

In vitro antibacterial and anti-biofilm properties of ICG@LNT-SeNPs. (A) Representative agar plate images showing MRSA colonies after treatment with different materials, (with or without 808-nm near-infrared light irradiation, 1.0 W cm⁻², 10 min). (B) Quantitative analysis of bacterial survival based on CFU counts. (C) Live/dead staining of planktonic MRSA using SYTO 9 and propidium iodide. Green fluorescence indicates viable bacteria with intact membranes, and red fluorescence indicates dead bacteria with compromised membranes (scale bar, 100 μm). (D) Quantification of bacterial killing rates derived from fluorescence assays. (E) SEM image of the morphological integrity of MRSA. Note the collapsed cell membranes and cellular debris in the ICG@LNT-SeNPs + NIR group (scale bar, 1 μm). (F) Crystal violet-stained bacterial biofilm in 96-well plates. The intensity of the violet color correlates qualitatively with the remaining biofilm biomass, showing visible clearance in the ICG@LNT-SeNPs + NIR group. (G) 3D-CLSM images of biofilms stained with live/dead dyes, visualizing the spatial distribution of viable (green) and dead (red) bacteria within the biofilm matrix (scale bar, 100 μm). Data are presented as mean ± SD (n = 3 biological replicates). Data were analyzed by two-way ANOVA with Tukey’s post hoc test. ***P < 0.001, ****P < 0.0001

Transcriptomic profiling reveals that ICG@LNT-SeNPs rewire macrophage phenotype via the selenoprotein axis and redox modulation

Building on single‑cell evidence of a sustained pro‑inflammatory macrophage state in diabetic wounds, we evaluated whether ICG@LNT‑SeNPs could reprogram macrophage plasticity in vitro. Using an lipopolysaccharide (LPS)‑stimulated RAW264.7 model, we found that both LNT‑SeNPs and ICG@LNT‑SeNPs significantly and similarly reduced intracellular ROS accumulation (Fig. 4A, B), confirming that the antioxidant effect originates from the selenium component. The transcriptional changes underlying this shift were analyzed through RNA sequencing, wherein LPS stimulation markedly altered global gene expression relative to PBS‑treated controls. These changes were largely reversed by ICG@LNT‑SeNPs treatment (Fig. 4C). Differentially expressed genes included core antioxidant genes (Hmox1, Nqo1, Gclc) and key factors in selenoprotein synthesis, such as the insertion factor Secisbp2 and the selenoproteins Gpx4 and Txnrd2 (Fig. 4D). The antioxidant response was accompanied by a shift toward a reparative phenotype, indicated by increased expression of M2 markers (Il10, Arg1), the chemokine Ccl24, and the pro‑angiogenic regulator Hif1a along with its target Vegfa. As revealed by pathway analysis, ICG@LNT‑SeNPs broadly modulated immune and stress‑related signaling, with notable downregulation of TNF and NF‑κB pathways (Fig. 4E). Thus, selenium‑driven immunometabolic reprogramming extended from simple ROS scavenging to correcting the dysregulated pseudohypoxic response common in chronic inflammation at both gene and protein levels. Quantitative real-time PCR (qPCR) verified upregulation of antioxidant (Gpx4, Txnrd2), M2‑associated (Arg1, Cd206, Il10), and angiogenic (Vegfa) genes (Fig. S4). Western blot analysis showed increased protein levels of selenoproteins GPX4 and TXNRD2 (Fig. 4F). NRF2 protein was post‑translationally stabilized without significant mRNA change, confirming activation through the KEAP1‑NRF2 axis. Selenium treatment also restored normal hypoxic signaling. While LPS induced pathological HIF-1α accumulation under normoxic conditions (pseudohypoxia), selenium intervention attenuated this abnormal stabilization and restored HIF-1α toward physiological turnover levels, rather than suppressing its normal hypoxia-responsive function. Together, these results demonstrate a dual regulatory mechanism: reinforcing the NRF2-selenoprotein axis while alleviating pseudohypoxia-associated HIF-1α stabilization and restoring physiological HIF-1α turnover, thereby recalibrating inflammation-driven metabolic bias rather than inhibiting canonical hypoxia signaling.

Fig. 4.

ICG@LNT-SeNPs attenuate oxidative stress and reshape macrophage transcriptional programs in RAW264.7 cells. (A) Representative fluorescence images of intracellular ROS levels visualized by DCFH-DA staining (green). From top to bottom: DCFH channel, bright-field channel, and merged images. Treatments include PBS (unstimulated), LPS (100 ng mL⁻¹), LPS + LNT-SeNPs, and LPS + ICG@LNT-SeNPs. Reduced green fluorescence is observed in nanoparticle-treated groups (scale bar, 100 μm). (B) Quantification of mean fluorescence intensity (MFI) representing relative intracellular ROS levels. LNT-SeNPs and ICG@LNT-SeNPs show comparable ROS-scavenging capacity. (C) Volcano plot illustrating transcriptional changes induced by LPS stimulation relative to PBS controls. (D) Heatmap of key genes involved in antioxidant defense, selenoprotein biosynthesis, and macrophage polarization. Red indicates upregulation and blue indicates downregulation. The ICG@LNT-SeNPs group shows restoration of antioxidant gene expression and induction of M2-associated markers. (E) KEGG pathway analysis revealed altered immune-related signaling after treatment with LNT-SeNPs and ICG@LNT-SeNPs. (F) Protein levels of the transcription factors HIF-1α and NRF2, along with their downstream selenoprotein targets TXNRD2 and GPX4, were assessed by Western blot in RAW264.7 cells. β-Actin serves as the loading control. ICG@LNT-SeNPs treatment increased NRF2 protein abundance and attenuated LPS-induced HIF-1α accumulation, alongside higher GPX4 and TXNRD2 levels, consistent with engagement of antioxidant and repair-associated programs at the protein level. Data in (B) are presented as mean ± SD from n = 3 biologically independent samples. Statistical significance was assessed by one-way ANOVA with Tukey’s multiple-comparisons test. **P < 0.01 versus LPS; ns, not significant between LNT-SeNPs and ICG@LNT-SeNPs

ICG@LNT-SeNPs potentiates the migratory capacity of key reparative cells

In addition to bacterial clearance, resolution of impaired tissue repair, which is characterized by angiogenic deficiency and delayed stromal reconstruction, is critical for diabetic wound recovery. Given that ICG@LNT-SeNPs restore intracellular redox balance (Fig. 4), this redox-normalized microenvironment was hypothesized to enhance the functional capacity of key reparative cells. Therefore, cell migration assays were performed using human umbilical vein endothelial cells (HUVECs) and NIH3T3 fibroblasts.

In scratch wound-healing assays, control groups treated with PBS or free ICG exhibited limited wound closure at 24 h. By contrast, both Se-containing formulations (LNT-SeNPs and ICG@LNT-SeNPs) significantly accelerated gap closure in both cell types (Fig. 5A, C). Quantitative analysis confirmed that ICG@LNT-SeNPs robustly promoted cell motility compared to the PBS control (Fig. 5B, D), likely because of the ability of selenium to scavenge excess reactive oxygen species. Consequently, the cytoskeletal dynamics essential for movement are preserved, as confirmed by a Transwell migration assay. In line with scratch wound findings, ICG@LNT-SeNP treatment substantially elevated the HUVEC population that migrated to the lower chamber (Fig. 5E, F). Overall, the proposed ICG@LNT-SeNPs platform can enhance the migration of both fibroblasts and endothelial cells through intrinsic, photothermal-independent bioactivity. This provides a cellular basis for potentially accelerating granulation tissue formation and re‑epithelialization during wound healing.

Fig. 5.

ICG@LNT-SeNPs accelerates reparative migration of fibroblasts and endothelial cells. (A) Representative images of scratch wound-healing assays in HUVECs at 0 h and 24 h after treatment. The enhanced gap closure in the Se-containing groups compared to the PBS control (scale bar, 200 µm). (B) Quantitative analysis of the relative migration rate for HUVECs, showing accelerated wound closure in the ICG@LNT-SeNPs group. (C) Representative scratch wound images of NIH3T3 showing enhanced endothelial motility following ICG@LNT-SeNPs treatment (scale bar, 200 µm). (D) Quantification of NIH3T3 migration rates at 24 h. (E) Representative images of transmigrated HUVECs on the lower surface of Transwell membranes stained with crystal violet (scale bar, 200 μm). (F) Quantification of the number of migrated HUVECs per field, confirming the pro-angiogenic potential of the nanoplatform. Data are presented as mean ± SD (n = 3 biologically independent experiments). Statistical significance was determined by one-way ANOVA followed by Tukey’s post hoc test. *P < 0.05, **P < 0.01, ****P < 0.0001 vs. PBS control

NIR-activated ICG@LNT-SeNPs promote healing of MRSA-infected diabetic wounds in mice

A murine model of MRSA-infected diabetic wounds was established to evaluate the in vivo therapeutic efficacy of ICG@LNT-SeNPs, including wound closure, bacterial clearance, and tissue repair.

Synergistic photothermal and metabolic therapy accelerates wound closure

As shown in Fig. 6A, infrared thermal imaging revealed that both the ICG@LNT-SeNPs-treated group and the ICG-treated group exhibited a significant increase in wound temperature under 808-nm laser irradiation, whereas no notable temperature rise was observed in the PBS or LNT-SeNPs-treated groups. This demonstrates that the photothermal effect is effective for antibacterial applications, supported by both theoretical and experimental evidence. Visual assessment of wounds revealed delayed healing in the PBS and free ICG groups, characterized by persistent inflammation and incomplete closure at day 14 (Fig. 6B). Treatment with LNT-SeNPs alone or free ICG combined with NIR light showed moderate improvement. The most rapid healing occurred in wounds that received ICG@LNT-SeNPs + NIR irradiation, with near-complete re-epithelialization and minimal scarring by the endpoint (Fig. 6B), as also confirmed by quantitative measurements (Fig. 6D). Thus, combining immediate photothermal sterilization with sustained selenium-mediated metabolic regulation achieves superior healing outcomes compared to individual treatments.

Fig. 6.

In vivo therapeutic efficacy of ICG@LNT-SeNPs against MRSA-infected diabetic wounds. (A) The infrared thermal images of MRSA-infected diabetic wounds mice after treatment with PBS, LNT-SeNPs and ICG, ICG@LNT-SeNPs during 2 min of laser irradiation (808 nm, 1.0 W cm-2), respectively. For each treatment group, images shown in the same row were obtained continuously from a same representative mouse. (B) Representative macroscopic images of wounds on days 0, 1, 4, 7, 11, and 14 after treatment. (C) Representative agar plates showing bacterial colonies cultured from wound exudates on day 4, corresponding to the early infection phase. (D) Quantitative analysis of wound closure on day 14. (E) Quantification of bacterial burden (CFU) in wound exudates on day 4. (F) Representative H&E-stained histological sections of wound tissues on day 7, showing reduced inflammatory cell infiltration. Scale bar, 1 mm. (G) Masson’s trichrome staining of regenerated tissue on day 14 to visualize collagen deposition (blue). Scale bar, 1 mm. (H) Collagen volume fraction was quantified from Masson’s trichrome-stained sections. Values represent mean ± SD (n = 3 independent biological samples). Statistical comparisons for (E) were performed using two-way ANOVA; panels (D) and (H) were analyzed by one-way ANOVA, each followed by Tukey’s post hoc test. ***P < 0.001, ****P < 0.0001 versus the PBS control group

Efficient eradication of bacterial burden

To assess bacterial burden in wound exudates, we observed CFU counts (Fig. 6C, E), noting high bacterial loads in the PBS and ICG-treated (without NIR) groups that confirmed ongoing infection. By contrast, LNT‑SeNPs and ICG@LNT‑SeNPs reduced bacterial counts even without NIR, highlighting that the intrinsic antibacterial activity of selenium is high. Further, ICG@LNT‑SeNP treatment followed by 808-nm irradiation drastically eliminated viable bacteria, achieving negligible CFU counts (Fig. 6E). The exceptional outcomes of this treatment can be attributed to an immediate photothermal reduction of bacterial load followed by prolonged antibacterial activity driven by sustained selenium release.

Modulation of the immune microenvironment and resolution of inflammation

At day 7, wound tissue was examined to determine the modulatory action of nanoplatform toward the stalled inflammatory phase. As revealed by qPCR analysis, a distinct shift occurred in the immune microenvironment because of the effective suppression of the inflammatory phase due to ICG@LNT-SeNPs + NIR treatment (Fig. S5), unlike the sustained pro-inflammatory state in PBS and free ICG groups (high Cd86, Il1b, Il6, Tnf). Moreover, histological analysis confirmed that the integrated therapy strategy promoted repair and elevated expression of the anti-inflammatory cytokine Il10 and the angiogenic factor Vegfa. H&E-stained sections from the ICG@LNT-SeNPs + NIR group demonstrated considerably reduced inflammatory infiltration and swelling and enhanced organization in granulation tissue. Overall, combination therapy reduces inflammation as well as actively redirects the wound immune microenvironment from a chronic inflammatory state toward a pro-regenerative one.

Enhanced collagen deposition and matrix maturity

Wound sections were stained with Masson’s trichrome to visualize collagen deposition in wound tissue at later healing stages (Fig. 6G). Quantitative analysis revealed the lowest collagen content in the PBS and ICG groups(Fig. 6H), reflecting poor tissue repair. In contrast, collagen levels were significantly higher in wounds treated with LNT‑SeNPs or ICG@LNT‑SeNPs, with the most pronounced increase observed in the ICG@LNT‑SeNPs + NIR group. These results indicate that selenium nanoparticles effectively enhance the synthesis and deposition of collagen, a major extracellular matrix component, thereby improving the structural quality of healed tissue.

Orchestrating a pro-regenerative microenvironment to drive angiogenesis and tissue remodeling

On day 14 post-surgery, the immune-stromal state of healed tissue was evaluated by immunofluorescence. Cytokine analysis showed a stepwise reduction in IL-6 across treatment groups, with the lowest level in the ICG@LNT-SeNPs + NIR group (Fig. 7A, D), indicating effective resolution of acute inflammation. In contrast, levels of the anti‑inflammatory cytokine IL‑10 were notably higher in all groups containing selenium compared to the PBS control. The LNT‑SeNPs group showed the strongest IL‑10 signal, with the ICG@LNT‑SeNPs + NIR group close behind (Fig. 7A, E). This consistent increase underscores the role of selenium in stabilizing immune regulation and preventing prolonged inflammation. We next examined macrophage polarization (Fig. 7B, F, G). Among all groups, the LNT‑SeNPs group displayed the most favorable regenerative profile, with the lowest level of the M1 marker CD86 and the highest level of the M2 marker CD206. These results confirm that selenium itself can modulate macrophage behavior by restraining the pro‑inflammatory M1 state while promoting the reparative M2 phenotype. In the combination treatment group (ICG@LNT‑SeNPs + NIR), the expression pattern reflected a more advanced stage of tissue repair. Although CD86 was higher than in the LNT-SeNPs group (yet still lower than PBS) and CD206 was slightly below the level seen with LNT-SeNPs alone, this group maintained strong VEGF-A expression. VEGF-A was highest in the ICG@LNT-SeNPs + NIR group, significantly exceeding all other conditions (Fig. 7C, H). Together, these results suggest that while selenium nanoparticles optimize the immune baseline, photothermal therapy accelerates its functional translation into enhanced vascularization and tissue closure.

Fig. 7.

Immunofluorescence analysis reveals that ICG@LNT-SeNPs orchestrate an M2-dominant, pro-angiogenic microenvironment in MRSA-infected diabetic wounds at day 14. (A–C) Representative immunofluorescence images of wound tissue sections stained for (A) inflammatory cytokines IL-6 (green) and anti-inflammatory cytokine IL-10 (red); (B) macrophage polarization markers CD86 (M1 phenotype, green) and CD206 (M2 phenotype, red); and (C) the angiogenic factor VEGF-A (red). Nuclei were counterstained with DAPI (blue). Scale bars, 100 μm. (D–H) Quantitative analysis of the mean fluorescence intensity (MFI) for (D) IL-6, (E) IL-10, (F) CD86, (G) CD206, and (H) VEGF-A. Data are presented as mean ± SD. Sample sizes are n = 5 biologically independent samples for (H) and n = 3 for (D–G). Statistical significance was determined by one-way ANOVA followed by Tukey’s post hoc test. *P < 0.05, **P < 0.01, ***P < 0.001 vs. the PBS control group; ns, not significant

Biocompatibility assessment: ensuring safety for translational application

To evaluate translational safety of the nanoplatform, hemocompatibility, cytocompatibility, and in vivo systemic toxicity were systematically assessed. Hemocompatibility was first examined using a hemolysis assay. As shown in Fig. S6A, B, ICG@LNT-SeNPs induced minimal hemolytic activity at therapeutic concentrations. Even at the highest tested dose (12.8 µg mL⁻¹, calculated by Se content), the hemolysis rate remained at approximately 4%, considerably lower than the accepted safety threshold of 5% defined by ISO 10993-4. By contrast, the Triton X-100 positive control caused complete erythrocyte membrane disruption. In vitro cytocompatibility was evaluated in three representative cell types relevant to wound healing: RAW264.7 macrophages, NIH3T3 fibroblasts, and HUVECs. CCK-8 assays showed high cellular viability across all cell lines, with viability exceeding 90% at selenium concentrations up to 12.8 µg mL⁻¹ (Fig. S7A–C), indicating negligible impact on cellular metabolic activity. Systemic toxicity was further assessed in vivo by histological examination of major organs, including the heart, liver, spleen, and kidneys, collected on day 14. H&E staining revealed no evidence of acute inflammation, necrosis, or structural abnormalities in any treatment group compared with healthy controls (Fig. S8). Collectively, these evaluations of hemocompatibility, cytocompatibility, and in vivo histology demonstrate the favorable biocompatibility profile of ICG@LNT‑SeNPs and support its safety for topical therapeutic application.

Discussion

Chronic diabetic wounds complicated by multidrug-resistant bacterial biofilms represent a major clinical challenge, sustained by a self-perpetuating cycle of infection, redox imbalance, and inflammation [1, 2]. Conventional antibiotic therapies frequently fail because of limited biofilm penetration and inability to reverse the arrested healing state. While NIR-responsive nanoplatforms offer a physical means to disrupt biofilms [29–31], photothermal stress alone can damage newly forming tissue and exacerbate oxidative injury [32]. Therefore, effective therapy must extend beyond bacterial eradication to actively support host tissue repair. In this work, ICG@LNT-SeNPs were developed as a dual-functional nanoplatform that integrates rapid photothermal sterilization with selenium-driven immunometabolic reprogramming.

Efficient biofilm eradication is a prerequisite for effective wound healing. ICG@LNT-SeNPs produced potent bactericidal activity against both planktonic and biofilm-resident MRSA under NIR irradiation (Fig. 3). Mechanistically, this effect arises from a dual-action process. Localized photothermal heating disrupts the biofilm matrix and compromises bacterial membranes (Fig. 3E) [47], enabling sustained bacteriostatic pressure mediated by the selenium component. This combined strategy outperformed monotherapies in comparative studies. Vancomycin is a commonly employed antibiotic, but its efficacy against biofilm-related infections is often hampered by slow efficacy and poor penetration. In contrast, under the tested experimental conditions, our nanosystem achieved rapid sterilization within a short treatment period, which may offer a potential advantage in accelerating bacterial load reduction in recalcitrant, antibiotic-resistant infections [48, 49].

In addition to sterilization, the nanoplatform exerts a more fundamental effect through immune reprogramming. Transcriptomic analysis showed that selenium does not function solely as a passive ROS scavenger but acts as an active metabolic regulator [50]. In macrophages, ICG@LNT-SeNPs reduced oxidative stress while reinforcing endogenous antioxidant systems. Western blot analysis demonstrated significant upregulation of the selenoproteins GPX4 and TXNRD2 (Fig. 4F), which are critical for limiting lipid peroxidation and preventing ferroptotic cell death. Preservation of these pathways supports macrophage viability and function under oxidative stress conditions typical of diabetic wounds [51, 52].

A key observation was the divergent regulation of NRF2- and HIF-1α–associated programs (Fig. 4D). In diabetic wounds, as supported by re-analysis of scRNA-seq data (Fig. 1), excessive ROS can stabilize HIF-1α despite normoxia (pseudohypoxia), a phenomenon referred to as pseudohypoxia, in which inflammatory redox imbalance drives pathological protein stabilization independent of true oxygen deprivation [53]. Aligning with this feature, HIF-1α protein accumulated abnormally in the LPS group. Notably, ICG@LNT-SeNPs markedly reduced HIF-1α protein abundance (Fig. 4F), indicating alleviation of pseudohypoxic stress and restoration of physiological HIF-1α turnover rather than inhibition of its canonical hypoxia-responsive function. Although Nfe2l2 mRNA levels remained unchanged, NRF2 protein abundance increased markedly (Fig. 4F), supporting a post-translational mechanism in which selenium stabilizes NRF2 by preventing KEAP1-dependent degradation rather than inducing transcription. Importantly, NRF2 activation may mechanistically intersect with HIF-1α regulation through redox control. Excessive ROS are known to inhibit prolyl hydroxylase domain (PHD) enzymes, thereby stabilizing HIF-1α under normoxic conditions (pseudohypoxia) [54]. By restoring intracellular redox balance and reinforcing antioxidant defenses, NRF2 activation can facilitate PHD activity and promote physiological hydroxylation-dependent turnover of HIF-1α. In this context, selenium-mediated NRF2 stabilization likely contributes indirectly to normalization of HIF-1α stability by alleviating ROS-driven pseudohypoxic signaling. This redox-dependent crosstalk provides an integrated explanation for the coordinated reinforcement of antioxidant capacity and attenuation of pathological HIF-1α accumulation observed in our study [54, 55]. Through this shift, the therapy uncouples the wound environment from HIF-1α-driven inflammatory signaling. Meanwhile, angiogenesis appears to be maintained via NRF2-linked Vegfa transcriptional activation and/or M2-associated secretion, independent of pathological HIF-1α accumulation. The observed high VEGF-A expression together with elevated IL-10 levels (Fig. 7) corroborates this reprogramming, converting a stalled inflammatory state into a regenerative process [56]. Although our mechanistic analyses primarily focused on macrophage reprogramming, NRF2 signaling is not restricted to macrophages. Previous studies have demonstrated that NRF2 regulates redox homeostasis and inflammatory effector functions in multiple immune cell types, including neutrophils and T lymphocytes. For instance, NRF2 activation has been shown to modulate neutrophil oxidative responses and inflammatory signaling, including NADPH oxidase-associated pathways during bacterial infection [57]. Moreover, NRF2 has been reported to influence CD4⁺ T-cell activation and metabolic reprogramming by modulating glucose and glutamine utilization [58]. Therefore, selenium-mediated reinforcement of the NRF2-dependent antioxidant axis may represent a broader immunoregulatory principle within the wound microenvironment rather than a strictly macrophage-specific mechanism. Nevertheless, the present study experimentally validates this pathway in macrophages, and further investigation is required to determine the extent of NRF2 axis modulation in other immune populations during diabetic wound repair. To support translational relevance, we next examined tissue repair in vivo. Temporal analysis revealed an orderly healing sequence. During the early phase (days 4–7), effective infection control and inflammation resolution predominated, marked by downregulation of the M1 macrophage marker Cd86, reduced Il1b expression, and diminished inflammatory cell infiltration (Fig. 6). Rapid attenuation of the inflammatory response created a permissive environment for progression into the proliferative phase (day 14), characterized by enhanced collagen deposition and increased neovascularization (Fig. 7). Sustained elevation of IL-10 and VEGF-A at this stage confirms successful reprogramming of the wound milieu from a stalled chronic condition toward an active regenerative state [59]. Selenium further contributes to translational feasibility through favorable biocompatibility. As an essential dietary micronutrient, selenium exhibits a well-established safety profile. Unlike heavy metal–based nanozymes, SeNP degradation products are essential trace elements that can be metabolized by the host, reducing concerns related to long-term accumulation and toxicity [60–62].

Overall, ICG@LNT-SeNPs redefine diabetic wound therapy by coupling antibacterial intervention with pro-regenerative modulation. The therapeutic benefit of this combined strategy does not rely on uniform synergy across all endpoints. Instead, photothermal treatment mainly provides rapid bacterial reduction in the early phase, whereas selenium predominantly restores redox balance and supports immune-mediated repair. These distinct yet coordinated functions together contribute to the overall improvement in wound healing. Integration of photothermal bacterial eradication with protein-level regulation of the selenoprotein axis disrupts the chronic inflammatory loop, eliminates infection, and restores endogenous repair capacity. This strategy provides a mechanistically grounded promising approach for treating refractory diabetic wounds.

Conclusion

This study presents ICG@LNT-SeNPs as a nanotherapeutic platform that addresses the coupled challenges of persistent infection and impaired tissue regeneration in diabetic wounds. In addition to demonstrating near-infrared light-triggered photothermal bactericidal activity, the work elucidates how selenium-based nanomaterials coordinate host immunometabolic reprogramming across multiple regulatory layers. Consistent with experimentally obtained gene expression profiles and protein stability data, activation of the NRF2–selenoprotein axis redirects HIF-1α signaling from a pathological inflammatory state toward a pro‑angiogenic and pro‑repair state. This mechanism disrupts the chronic infection–inflammation loop while restoring endogenous repair capacity. Within this integrated framework, ICG functions as the rapid physical antibacterial module, enabling light-triggered biofilm disruption and immediate reduction of bacterial burden. In contrast, selenium serves as the sustained redox-regulatory core, mediating immunometabolic reprogramming through modulation of the NRF2/HIF-1α axis. The sequential coordination of rapid pathogen clearance followed by microenvironmental correction generates a synergistic therapeutic effect rather than a simple additive combination. In summary, ICG@LNT‑SeNPs represent a viable therapeutic approach for diabetic wound management. The findings define a strategy for nanotherapeutic design that prioritizes active reprogramming of the host’s immune metabolism alongside rapid pathogen clearance. Rather than merely delivering antimicrobial agents, this strategy integrates rapid physical pathogen clearance with sustained host-directed reprogramming, actively engaging and redirecting intrinsic repair mechanisms.

Materials and methods

Bioinformatic analysis of Public Single-Cell RNA-Sequencing Data

From the Gene Expression Omnibus we downloaded a publicly available human wound scRNA-seq dataset (GSE268834) and reanalyzed it in R (v4.4.2) with the Seurat package. This dataset comprises single-cell transcriptomic profiles derived from chronic diabetic foot ulcer tissues collected from the lower extremities, compared with non-diabetic skin controls. After the standard quality control, normalization, dimensionality reduction, batch integration, clustering and cell-type annotation, the macrophage population was further subsetted for secondary subclustering and characterization. Detailed workflow information including exact software versions, filtering thresholds, integration parameters, and downstream enrichment methodologies were provided in the Supplemental Methods.

Synthesis of ICG@LNT-SeNPs

ICG@LNT-SeNPs were synthesized with a one-pot reduction. Briefly, a mixture was prepared by dissolving ICG (200–500 µg mL⁻¹), sodium selenite (5 mM), and lentinan (2 mg mL⁻¹) in deionized water. Ascorbic acid (20 mM) was added dropwise to initiate the reaction (final volume: 5 mL). The mixture was mixed at RT for 8–12 h. After that, the resulting suspension was dialyzed against deionized water for 48–72 h to remove unreacted precursors and stored at 4 °C for further use.

Material characterization and biotransformation analysis

The morphology of the nanosystem was characterized using transmission electron microscopy (TEM; JEOL, USA) and dynamic light scattering (DLS; Malvern, UK) was used to determine the particle size distribution and zeta potential. Chemical composition, structural integrity, and ICG loading were verified by UV-vis spectroscopy (Varian, USA), Fourier-transform infrared spectroscopy (FTIR; Bruker, Germany), and X-ray photoelectron spectroscopy (XPS; Thermo ESCALAB 250Xi, USA). To assess the stability and metabolic fate of the nanoplatform in a physiological environment, selenium content and its metabolite speciation were quantitatively analyzed via high-performance liquid chromatography–inductively coupled plasma mass spectrometry (HPLC-ICP-MS; Thermo Fisher Scientific, USA).

Bacterial strains, cell culture, and animal experiments

The MRSA strain was obtained from the BeNa Culture Collection (Beijing, China). RAW264.7 murine macrophages (Cat. No. SNLM-112) were purchased from Guangzhou RuiShu Biotechnology Co., Ltd. NIH3T3 fibroblasts and human umbilical vein endothelial cells (HUVECs) were obtained from the Cell Resource Center of the Dermatology Hospital, Southern Medical University. Cells were cultured in DMEM supplemented with 10% fetal bovine serum (FBS; Thermo Fisher Scientific, Cat. No. 10091148) and 1% penicillin–streptomycin (Proteintech, Cat. No. PR40022) at 37 °C in a humidified atmosphere containing 5% CO₂. All cell lines were routinely tested for mycoplasma contamination and were negative.

Six-week-old male C57BL/6J mice were purchased from Zhuhai Bestone Biotechnology Co., Ltd. and housed under specific pathogen-free conditions with standard acclimation. All animal procedures were approved by the Animal Ethics Committee of Southern Medical University (Approval No. SMU-202503120 A) and conducted in accordance with institutional guidelines and the Guide for the Care and Use of Laboratory Animals. Human wound exudates were collected from diabetic patients with approval from the Medical Ethics Committee of Southern Medical University (Approval No. KY-2026-004). Written informed consent was obtained from all participants, and all procedures complied with the Declaration of Helsinki.

In vitro antimicrobial activity assessment

Bacterial culture and treatment protocol

MRSA was cultured in LB broth to the mid-logarithmic phase (~ 1 × 10⁸ CFU mL⁻¹) and diluted to the required working concentrations. For all assays, bacterial suspensions were incubated with PBS, free ICG (8 µg mL⁻¹), LNT-SeNPs (12.8 µg mL⁻¹ Se), ICG@LNT-SeNPs (selenium-equivalent concentration), or vancomycin (2 µg mL⁻¹) at 37 °C for 6 h. Samples assigned to NIR+ conditions were then exposed to an 808-nm laser (1.0 W cm⁻²) for 10 min, while NIR− groups were maintained in the dark.

CFU enumeration

Bactericidal efficacy was quantified by CFU enumeration. Treated bacterial suspensions (initial density ~ 1 × 10⁶ CFU mL⁻¹) were serially diluted in sterile PBS and plated on LB agar. Plates were incubated at 37 °C for 18–24 h, after which colonies were counted to calculate bacterial survival rates. All experiments were performed in triplicate.

Live/Dead staining and SEM analysis

Bacterial viability was assessed using the LIVE/DEAD BacLight Bacterial Viability Kit (Invitrogen), containing SYTO 9 (green, live) and propidium iodide (PI; red, dead). Stained samples were incubated in the dark for 20 min and imaged using a confocal laser scanning microscope (CLSM; Nikon A1R). For morphological analysis through scanning electron microscopy (SEM), bacterial pellets were collected by centrifugation, fixed with 2.5% glutaraldehyde at 4 °C overnight, and dehydrated through a graded ethanol series (30%–100%). After critical-point drying and gold sputter-coating, bacterial morphology was examined using a field-emission scanning electron microscope operated at 5 kV.

In vitro anti-biofilm assay

Mature MRSA biofilms were established by culturing bacterial suspensions (~ 1 × 10⁷ CFU mL⁻¹) in 96-well plates at 37 °C for 48 h. After removal of planktonic cells by washing, biofilms were treated with PBS, vancomycin, free ICG, LNT-SeNPs, or ICG@LNT-SeNPs for 6 h. NIR+ groups were then irradiated with an 808-nm laser (1.0 W cm⁻², 10 min), while NIR− groups were kept in the dark. Biofilm viability was evaluated by SYTO 9/PI staining followed by three-dimensional CLSM to visualize the spatial distribution of live and dead bacteria. Simultaneously, duplicate wells were stained with crystal violet to assess biofilm biomass at the macroscopic level. Representative images were acquired for qualitative evaluation of biofilm disruption.

In vitro oxidative stress and mechanism analysis

Intracellular ROS detection

RAW264.7 macrophages were stimulated with lipopolysaccharide (LPS; 100 ng mL⁻¹) for 24 h to induce oxidative stress, followed by treatment with PBS, LNT-SeNPs, or ICG@LNT-SeNPs for 6 h. Intracellular ROS levels were assessed using a 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) assay kit (Solarbio). Cells were incubated with 10 µM DCFH-DA for 30 min and imaged by confocal microscopy (Ex/Em: 488/525 nm). Fluorescence intensity was quantified using ImageJ.

Transcriptome sequencing (RNA-seq)

Cells were treated as described above. Total RNA was extracted using TRIzol reagent (Invitrogen), and samples with an RIN ≥ 8.0 were used for library construction. Sequencing was performed on an Illumina platform with 150-bp paired-end reads. Clean reads were aligned to the mouse reference genome (GRCm39) using HISAT2. Differentially expressed genes (DEGs) were identified with DESeq2 (|log₂FC| ≥ 1, FDR < 0.05) and subjected to GO and KEGG pathway enrichment analyses.

Quantitative real-time PCR (qPCR)

Total RNA was reverse-transcribed using HiScript IV SuperMix (Vazyme). qPCR was performed on a CFX384 real-time system (Bio-Rad) using ChamQ Universal SYBR Master Mix (Vazyme). Relative mRNA expression of target genes (Nfe2l2, Gpx4, Arg1, Cd206, Hif1a, Vegfa, and Il10) was calculated using the 2^−ΔΔCt method and normalized to Actb. Primer sequences are provided in Supplementary Table S1.

Western blot analysis

Protein expression was analyzed by standard Western blotting. Cell lysates were separated by SDS–PAGE and transferred to PVDF membranes. Membranes were incubated overnight at 4 °C with primary antibodies against GPX4 (1:2000, Servicebio), TXNRD2 (1:1000, Proteintech), NRF2 (1:2000, Affinity), and HIF-1α (1:1000, Affinity). β-Actin (1:20000, Proteintech) served as the loading control. Immunoreactive bands were visualized using enhanced chemiluminescence on a Clinx ChemiScope 6100 system.

Cell migration assays

Scratch assay

HUVECs and NIH3T3 fibroblasts were cultured to confluence in 6-well plates. A linear scratch was generated, and cells were incubated with the indicated formulations in low-serum medium (2% FBS). Images were acquired at 0 and 24 h. Relative wound closure was calculated as (W₀ − W₂₄)/W₀ × 100%, where W₀ and W₂₄ represent wound widths at 0 and 24 h, respectively.

Transwell assay

HUVECs (1 × 10⁴ cells per insert) were seeded into the upper chambers of Transwell inserts (8.0-µm pore size). The lower chambers contained medium supplemented with 20% FBS and the indicated treatments. After 24 h, migrated cells on the lower membrane surface were stained with 0.1% crystal violet and quantified from five random fields per insert.

In vivo evaluation of therapeutic efficacy

Diabetic infected wound model and treatment

All animal procedures were approved by the Animal Ethics Committee of Southern Medical University (Approval No. SMU-202503120 A). Type 2 diabetes was induced in C57BL/6J mice by intraperitoneal injection of streptozotocin (STZ; 50 mg kg⁻¹) for five consecutive days. Mice with fasting blood glucose levels > 11.1 mmol L⁻¹ were included. A 6-mm full-thickness excisional wound was created on the dorsal skin and inoculated with MRSA (1 × 10⁶ CFU).

Mice were randomized into four treatment groups: PBS, free ICG, LNT-SeNPs, and ICG@LNT-SeNPs. Each group was further divided into NIR− (dark) and NIR+ (irradiated) subgroups (n = 6). Treatments were applied topically once daily. For NIR+ subgroups, wounds were irradiated with an 808-nm laser (1.0 W cm⁻², 5 min) once daily for the first 3 days. Photothermal imaging method: On each MRSA-induced diabeticmouse, four full-thickness wounds were created on the dorsum and randomly assigned to the four treatments (PBS, LNT-SeNPs, free ICG, or ICG@LNT-SeNPs, n = 3). Temperature changes and IR thermal images ofthe wound in differently treated mice were recorded by an IRthermal imaging camera (Testo 882, Germany) after receiving 808 nm laser irradiation (1.0 W cm⁻²) for different time.And the raw data was stored in the IRSoft BMT format and subsequently exported as JPEG images using Testo IRSoft 5.2 software.

Evaluation metrics

Wound closure was monitored photographically at designated time points. On day 4, wound exudates were collected for CFU-based bacterial burden analysis. On day 7, a subset of animals was euthanized for histological evaluation (H&E staining) and qPCR analysis of inflammatory markers. On day 14, remaining animals were euthanized, and healed tissues were harvested for Masson’s trichrome staining and immunofluorescence analysis of IL-6, IL-10, VEGF-A, CD86, and CD206. Major organs were collected for assessment of systemic toxicity.

Statistical analysis

Data are presented as mean ± standard deviation. Differences between groups were analyzed using one-way ANOVA, two-way ANOVA, or Student’s t-test, as appropriate, with GraphPad Prism software version 10.1 for Windows. A p-value of < 0.05, < 0.01, < 0.001, or < 0.0001 was considered statistically significant.

Supplementary Information

Below is the link to the electronic supplementary material.

Supplementary Material ^{(6.4MB, docx)}

Author contributions

Y C, L Z, Z Z: Conceptualization, Methodology, Writing-original draft, Visualization. W L: Conceptualization, Investigation. Z X: Formal analysis. K L: Data curation. S.S. Li: Writing-review & editing. G C: Conceptualization. S.Q. Li: Data curation. X L: Methodology. J Z: Methodology. T C, R C: Conceptualization, Supervision.

Funding

The financial support for this research was provided by the General Program of Natural Science Foundation of Guangdong Province (No.2023A1515010015), Educational Reform Research Project on Resident Standardized Training at the Dermatology Hospital of Southern Medical University (Nos.ZP202401 and ZP202404) and Science and Technology Project for Social Development in Dongguan City (No.20221800905452), and Guangzhou Science and Technology Plan Basic and Applied Basic Research Project (2024A04J4248).

Data availability

The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Animal Ethics Committee of Southern Medical University (Approval No. SMU-202503120 A). All animal procedures were performed under national and institutional ethic guidelines for the care and use of laboratory animals.Human wound exudates were collected from diabetic patients with approval from the Medical Ethics Committee of Southern Medical University (Approval No. KY-2026-004).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.