An All‐in‐One Nanoplatform for Synergistic Anti‐Infection and Healing of Diabetic Wounds via Photothermal‐Gas Therapy

ABSTRACT

Diabetic wounds represent a significant clinical challenge, characterized by bacterial biofilm formation, persistent inflammation, and oxidative stress. Although photothermal therapy (PTT) and gas therapy (e.g., CO, H₂) are promising alternatives to conventional antibiotics, existing monotherapeutic strategies fail to address the multifactorial pathological characteristics of these wounds, which is a critical research gap. To fill this gap, an integrated multimodal nanoplatform (CO/AB@MPDA) by co‐loading dodecacarbonyltriiron (FeCO, a CO donor) and ammonia borane (AB, a H₂ donor) into mesoporous polydopamine (MPDA) carriers was developed. The innovation lied in its synergistic design. It enabled near‐infrared (NIR)‐triggered CO release and pH‐responsive H₂ generation, simultaneously targeting infections, inflammation, and microenvironment dysregulation. In vitro and in vivo experiments confirmed its potent antibacterial/biofilm‐disrupting activity, high biosafety, reactive oxygen species (ROS) scavenging, anti‐inflammatory properties, as well as superior ability to accelerate diabetic infected wound healing. In summary, the CO/AB@MPDA nanosystem overcomes the limitations of monomodal therapies through synergistic PTT‐gas therapy, providing an innovative and comprehensive therapeutic strategy for diabetic wound infections.

A multifunctional nanoplatform is developed for diabetic infected wounds. Under NIR light, it releases CO and H₂, eliminates bacteria, scavenges ROS, and promotes macrophage polarization to accelerate wound healing.

1. Introduction

Diabetes has emerged as a major and growing public health challenge worldwide, with its prevalence rising most rapidly in low‐ and middle‐income countries [1]. In 2022, approximately 830 million people were living with diabetes globally, a figure projected to reach around one billion by 2030. Diabetic wounds, a serious complication of the disease, are particularly difficult to manage, often resulting in persistent pain and imposing considerable economic and psychological burdens on patients [2, 3]. These wounds can lead to anxiety, depression, and a marked decline in quality of life. As a result, the effective management of wound healing in individuals with diabetes remains a persistent and complex challenge in clinical practice [4].

Hyperglycemia is a defining characteristic of the diabetic wound microenvironment [5, 6, 7]. The elevated glucose levels render these wounds more susceptible to vascular impairment compared to normoglycemic wounds, thereby promoting substantial bacterial colonization. Clinical evidence consistently identifies bacterial infection as a principal contributor to diabetic wound progression and subsequent amputations [8, 9]. Compromised skin barrier function in diabetic patients further exacerbates infection risk. Moreover, the high‐glucose milieu provides an abundant carbon source for pathogens such as S. aureus, facilitating robust biofilm formation [10, 11]. Biofilms represent a major cause of treatment failure in chronic wound infections [12], as they impede antibiotic penetration through their extracellular polymeric substance (EPS) matrix and promote a persister cell phenotype that enhances antibiotic tolerance [13, 14]. Additionally, biofilms continuously release virulence factors, including elastase LasB [15, 16, 17], which contribute to extracellular matrix degradation and activation of the TLR signaling pathway, leading to overexpression of pro‐inflammatory cytokines such as TNF‐α and IL‐6 [16, 17]. This cascade of events further disrupts wound microenvironment homeostasis, often resulting in treatment failure when chronic wounds develop secondary infections.

In the dermis, persistent hyperglycemia leads to the formation of advanced glycation end products (AGEs) [18, 19]. Accumulated clinical evidence indicates that AGEs deposition in the extracellular matrix (ECM) correlates closely with the severity of diabetic wounds [19]. AGEs exert detrimental effects primarily by inducing cellular dysfunction, sustaining inflammation, enhancing oxidative stress, and impairing nitric oxide (NO) production [18]. Consequently, AGEs accumulation is now recognized as a major contributor to impaired healing in diabetic wounds [20, 21]. The interaction between AGEs and their receptor (RAGE) appears to impede macrophage phenotype switching [22, 23]. Under physiological wound healing conditions, a well‐orchestrated transition occurs from the inflammatory phase to the tissue regeneration stage. However, in diabetic wounds, this process is disrupted, characterized by sustained M1 macrophage persistence and a notable scarcity or absence of M2 macrophages. The hyperglycemic microenvironment promotes AGEs accumulation, which, upon binding to RAGE, triggers NADPH oxidase‐dependent ROS burst. This oxidative stress response leads to both direct cellular damage and overactivation of the NF‐κB pathway. These events collectively inhibit fibroblast proliferation and vascular endothelial migration while promoting macrophage polarization toward the pro‐inflammatory M1 phenotype, thereby impeding inflammation resolution and tissue remodeling.

Biofilms exacerbate chronic inflammation through persistent interaction with components of the host immune system, including neutrophils, macrophages, and inflammatory cytokines [24]. The immunosuppressive effects mediated by the AGEs–RAGE axis, combined with the physical barrier properties of biofilms, establish a self‐sustaining vicious cycle [25, 26], which significantly diminishes the efficacy of conventional debridement and antibiotic therapies. Consequently, there is an urgent need to develop more effective treatment strategies capable of simultaneously combating bacterial infections, modulating the hyperinflammatory microenvironment, and accelerating diabetic wound healing. Such integrated therapeutic approaches are essential for improving clinical outcomes and quality of life in affected patients.

The pathological complexity of diabetic wound healing has promoted the development of novel therapeutic strategies represented by gas therapy [27, 28]. Different from traditional single‐target drugs, gas molecules can regulate the wound microenvironment through multiple signaling pathways and concentration‐dependent biological effects, showing unique advantages in chronic wound repair. Among various bioactive gases, CO and H₂ have been widely studied due to their excellent biosafety and specific pharmacological activities [29, 30, 31]. Low‐dose CO exhibits remarkable anti‐inflammatory and antibacterial effects, while high‐dose CO can directly kill drug‐resistant bacteria [32, 33, 34]. Meanwhile, H₂ selectively scavenges cytotoxic hydroxyl radicals (·OH) while preserving functional reactive oxygen species such as H₂O₂, thus maintaining essential immune defenses [35, 36]. Although gas therapy shows great potential, its clinical application is still limited by unstable gas release, uncontrollable concentration, and a single therapeutic function.

To overcome these limitations, we designed a multifunctional nanoplatform co‐loaded with CO and H₂ donors for synergistic treatment of diabetic infected wounds. MPDA was selected as the carrier owing to its good biocompatibility, high drug loading efficiency, and excellent photothermal conversion performance. FeCO was used as a photothermal‐responsive CO donor, which could achieve on‐demand gas release under NIR irradiation. AB was introduced as a pH‐sensitive H₂ donor to realize sustained hydrogen production in the acidic infected wound microenvironment. Compared with traditional gas delivery systems with single function or single gas release, this platform integrated photothermal therapy, dual‐gas synergistic therapy, and immune microenvironment regulation. As shown in Scheme 1, the obtained CO/AB@MPDA nanoparticles (NPs) realized NIR/pH dual‐responsive gas release. Under NIR irradiation, CO was rapidly released to eliminate bacteria and inhibit biofilm formation. Meanwhile, acidic microenvironment‐triggered H₂ sustained release reduced oxidative stress and promoted the transformation of macrophages to the anti‐inflammatory M2 phenotype. This multi‐mechanism synergistic strategy effectively improves the complex pathological microenvironment of diabetic wounds, providing a new idea for the development of intelligent wound repair materials, which shows great potential for clinical translation in the treatment of diabetic infected wounds.

SCHEME 1.

Schematic illustration of the synthesis of CO/AB@MPDA NPs and therapeutic mechanism in treating infected diabetic wounds.

2. Experimental Section

2.1. Materials

1, 3, 5‐Trimethylbenzene (TMB), dopamine hydrochloride, FeCO, sodium bicarbonate, L‐glutamine, tert‐butanol, citric acid, and trisodium citrate were purchased from Aladdin Co., Ltd. Anhydrous ethanol, trichloromethane, acetone, and ammonia solution were obtained from Chuan Dong Chemical Co., Ltd. Pluronic F127, hemoglobin (HB), methylene blue (MB), and AB were supplied by Macklin Co., Ltd. Mueller–Hinton broth (MHB) culture medium, hematoxylin and eosin (H&E) staining kit, Masson's trichrome staining kit, and agarose were provided by Solabio Co., Ltd. Crystal violet staining solution, SYTO9/PI staining reagent, and ROS detection kit (DCFH‐DA) were obtained from Biyuntian Biotech Co., Ltd.

2.2. Characterization

The morphology and structure of NPs were characterized using transmission electron microscopy (TEM; Talos F200S, Thermo Fisher Scientific, Czech Republic) coupled with Super‐X energy‐dispersive spectroscopy (EDS) at an accelerating voltage of 200 kV. Hydrodynamic size distribution and surface charge measurements were conducted via dynamic light scattering and zeta potential analysis (Omni multi‐angle particle analyzer). The crystalline phase and elemental valence states of CO/AB@MPDA were determined by X‐ray photoelectron spectroscopy (XPS; PHI 5400, Physical Electronics, USA). Fourier‐transform infrared spectroscopy (FT‐IR; Thermo Scientific iS50 Nicolet, USA) was employed to verify the loading efficiency of FeCO and AB within the composite. For photothermal performance assessment, temperature elevation of CO/AB@MPDA suspensions under 808 nm laser irradiation (MR.FCL.808.T2‐2 W.MM.YF.OEM) was monitored in real time using an infrared thermal camera (FLIR E64501). CO and H₂ release profiles were quantified through thermogravimetric analysis (TGA2) and validated via a gas‐phase UV–vis spectrophotometer (SpecordⓇ210).

2.3. Preparation of the Samples

MPDA NPs with large‐pore structures were synthesized following an established nanoemulsion assembly approach [37]. Specifically, 50 mL of deionized water and 50 mL of anhydrous ethanol were mixed in a 250 mL reaction flask. Then, 1000 mg of Pluronic F127 and 500 mg of dopamine hydrochloride were added to the mixture under ultrasonication until a homogeneous solution was obtained. Under continuous stirring at 600 rpm, 2 mL of TMB was introduced dropwise, followed by stirring for 30 min at room temperature. While maintaining the same stirring speed, 5 mL of ammonia solution was added dropwise, and the reaction was allowed to proceed for 180 min. The resulting product was equally divided into 50 mL centrifuge tubes, and 10 mL of acetone was added to each tube to induce demulsification. After centrifugation at 11 000 rpm for 15 min, the supernatant was carefully discarded. Template removal was accomplished via three successive washing steps using an ethanol:acetone (2:1, v/v) solution, with each step involving 30 min of ultrasonication. The purified MPDA NPs were finally stored in anhydrous ethanol, and the solid content was determined prior to use.

The CO@MPDA and CO/AB@MPDA NPs were prepared as follows. Initially, deionized water was degassed by boiling under vacuum for 2 h to obtain oxygen‐free water. Separately, 2 mg of MPDA NPs were dispersed in 10 µL of chloroform, and 2 mg of FeCO was dissolved in 10 µL of chloroform. Under a nitrogen atmosphere, the two solutions were combined in a 25 mL three‐necked flask. The mixture was ultrasonicated while 4 mL of oxygen‐free water was slowly injected using a 5 mL syringe. Ultrasonication was maintained for an additional 5 min after injection. The resulting suspension was equally aliquoted into 2 mL centrifuge tubes and centrifuged at 11 000 rpm for 5 min to collect the product. After removal of the supernatant, the precipitate was washed three times with chloroform. The final CO@MPDA NPs were resuspended in anhydrous ethanol. For the synthesis of CO/AB@MPDA NPs, the as‐prepared CO@MPDA NPs were dispersed in 5 mL of oxygen‐free water containing 10 mg of AB under nitrogen protection. The reaction proceeded overnight at room temperature with stirring at 400 rpm. The CO/AB@MPDA NPs were collected by centrifugation (11 000 rpm, 10 min) and stored in anhydrous ethanol for further use.

2.4. Photothermal Properties of NPs In Vitro

The photothermal conversion performance of the NPs was evaluated under 808 nm laser irradiation. The NPs were uniformly dispersed in deionized water via ultrasonication (1 mL per well) in 48‐well plates. Real‐time temperature variations were recorded using a digital thermometer with a thermocouple probe during 5 min of NIR exposure. The following experimental sets were performed. (1) Photothermal heating profiles under different NIR power densities (0.5, 1.0, and 1.5 W/cm²); (2) Concentration‐dependent temperature rise (0.1, 0.2, and 0.5 mg/mL) at a fixed power density of 1.0 W/cm²; (3) Three repeated laser on/off cycles for CO/AB@MPDA NPs (0.2 mg/mL) at 1.0 W/cm² to assess photothermal stability.

2.5. The Release of CO/H₂

The gas‐releasing activity was quantified using an allyl fluorescein ether (AFE)–Pd(II) fluorescent probe [38]. A 5.0 mL PBS solution (10 mm, pH 7.4, containing 0.5% DMSO) containing both the probe (5 µm) and PdCl₂ (5 µm) was prepared. CO/AB@MPDA NPs (200 µg/mL) were added to this solution and divided into five aliquots. Following irradiation with an 808 nm NIR laser (1 W/cm²) for 1–5 min, fluorescence emission spectra were recorded using a spectrofluorometer (λex = 488 nm, λem = 500–600 nm).

CO release was quantified spectrophotometrically using reduced hemoglobin (r‐Hb) according to the Beer–Lambert principle [33]. Deoxygenated r‐Hb solution (6 µm in 500 µL PBS) was treated with excess sodium dithionite (0.1%) and purged with N₂ for 10 min. After adding CO/AB@MPDA dispersion (500 µL deoxygenated PBS), the mixture was irradiated with 808 nm NIR laser (1.0 W/cm²) for 20 min. UV–vis spectra were recorded to monitor the conversion of r‐Hb to carboxyhemoglobin (r‐Hb‐CO), characterized by characteristic absorbance bands at 430 nm (r‐Hb) and 410 nm (r‐Hb‐CO). CO concentration was calculated using the Equation:

$$C_{CO}=\frac{528.6\times l_{410nm}-304\times l_{430nm}}{216.5\times l_{410nm}+442.4\times l_{430nm}}\times C_{HB}$$

where C_CO and C_Hb represent CO and hemoglobin concentrations, respectively, and A₄₁₀/A₄₃₀ denote absorbance values at specified wavelengths.

H₂ quantification employed a catalytic reduction assay where Pt reduces MB to colorless leucomethylene blue (MBH₂). An MB standard curve was first established spectrophotometrically (664 nm). CO/AB@MPDA NPs (10 mm) in pH‐varied solutions were reacted with MB‐Pt probe solution (300 µg/mL) for designated durations. Absorbance measurements at 664 nm quantified MB consumption, with H₂ release calculated from the standard curve based on the reduction‐dependent absorbance decrease.

2.6. In Vitro Antibacterial Activity of NPs

The antibacterial activities of CO/AB@MPDA NPs against S. aureus and E. coli were assessed by colony counting. A 2×4 factorial design was employed, comprising eight experimental groups: Control, MPDA, CO@MPDA, CO/AB@MPDA, Control + NIR, MPDA + NIR, CO@MPDA + NIR, CO/AB@MPDA + NIR.

Bacterial suspensions were combined with the corresponding NPs and seeded into 48‐well plates. Specifically, 100 µL of each material and 0.5 mL of bacterial suspension (1×10⁶ CFU/mL) were added to each well. For NIR‐treated groups, samples were irradiated with an 808 nm laser (1 W/cm²) for 10 min. All plates were then incubated at 37°C for 6 h under aerobic conditions. After incubation, each sample was serially diluted to 10⁴ in sterile PBS within a biosafety cabinet. A 100 µL aliquot of the diluted suspension was spread onto agar plates and incubated overnight at 37°C. Representative colonies were photographed, and the number of colonies was counted to determine the bacterial inhibition rate.

S. aureus and E. coli were cultured in MHB at 37°C with shaking until the logarithmic growth phase was reached (OD₆₀₀ ≈ 0.6). Bacterial cells were harvested by centrifugation (5,000 rpm, 5 min), washed three times with PBS, and resuspended to a final concentration of 1 × 10⁶ CFU/mL. The bacterial suspensions were co‐incubated with the nanomaterials at 37°C for 6 h; samples in the NIR‐treated groups were irradiated with an 808 nm laser (1 W/cm²) for 10 min. For live/dead staining, 100 µL aliquots of the bacterial suspensions were transferred to a 96‐well plate and stained with SYTO 9 (5 µm) and propidium iodide (PI, 30 µm) in the dark for 15 min. After centrifugation (5000 rpm, 5 min) and two washes with PBS, the pellets were resuspended in 100 µL of PBS. A 20 µL aliquot of each suspension was mounted on a glass slide for immediate observation under a confocal laser scanning microscope (CLSM). SYTO 9 and PI were excited at 488 and 561 nm, with emission signals collected in the ranges of 500–550 nm (green) and 600–650 nm (red), respectively. The bacterial lethality was calculated as follows:

$$\begin{array}{cc}& \mathrm{Lethality}\left(\%\right)\hfill \\ & =\left(\mathrm{Red}\mathrm{fluorescence}\mathrm{area}/\mathrm{Total}\mathrm{fluorescence}\mathrm{area}\right)\times 100\hfill \end{array}$$

Bacterial suspensions of S. aureus or E. coli (1×10⁶ CFU/mL) were incubated with test materials (100 µL) in 48‐well plates containing 0.5 mL of culture medium. During the 6 h incubation at 37°C, samples designated for photothermal treatment were exposed to NIR irradiation (808 nm, 1 W/cm²). After incubation, bacterial cells were collected by centrifugation and transferred onto silicon wafers pre‐positioned in 24‐well plates. The specimens were fixed with 4% paraformaldehyde at 4°C for 2 h, followed by sequential dehydration through an ethanol gradient series (25%, 50%, 75%, and 100%; 15 min per concentration) and subsequent tert‐butanol replacement. Finally, the samples were sputter‐coated with a thin gold layer and examined under SEM to characterize morphological alterations.

Mix the S. aureus suspension (1 × 10⁶ CFU/mL) with MPDA, CO@MPDA, and CO/AB@MPDA NPs, and the corresponding groups were irradiated with an 808 nm NIR laser. The incubated samples were centrifuged at 11 000 rpm for 10 min, and the total amount of leaked protein in the resulting supernatant was determined using a BCA assay kit, with the absorbance measured at 595 nm. For the o‐nitrophenyl‐β‐D‐galactopyranoside (ONPG) hydrolysis assay, bacteria were collected by centrifugation after treatment, gently resuspended in PBS, and then incubated with ONPG substrate at 37°C in the dark. After the reaction was terminated, the absorbance of the supernatant at 405 nm was measured to assess the release of intracellular β‐galactosidase, thereby confirming the disruption of bacterial cell membranes.

Bacterial suspensions (1 × 10⁶ CFU/mL) were co‐incubated with NPs under NIR irradiation (808 nm, 1 W/cm²) at 37°C for 6 h. The samples were then stained with 5 µm DCFH‐DA in the dark for 15 min. After centrifugation (5,000 rpm, 5 min) and two washes with PBS, the pellets were resuspended in 100 µL of PBS. A 20 µL aliquot of each suspension was mounted on a glass slide for immediate observation under CLSM. The mean fluorescence intensity was quantified using ImageJ software and subjected to statistical analysis.

The anti‐biofilm activity of the NPs was evaluated through crystal violet staining and live/dead biofilm imaging. S. aureus suspension (1 × 10⁸ CFU/mL) was inoculated into 24‐well plates and incubated at 37°C in a shaking incubator for 12 h to allow biofilm formation. After removing the bacterial solution, each well was supplemented with PBS‐diluted MHB medium and the respective NPs formulations. Following 10 min of NIR (1 W/cm²) for the light‐treated groups, the plates were further incubated at 37°C with shaking for 12 h. The resulting biofilms were quantified using crystal violet staining. After discarding the supernatant and washing with PBS, the biofilms were fixed and stained with 200 µL of crystal violet solution (1% w/v) for 30 min. Excess stain was removed by PBS rinsing, and the bound dye was solubilized with 200 µL of 95% ethanol. A 100 µL aliquot of the solution was transferred to a 96‐well plate, and the absorbance at 570 nm was measured using a microplate reader to calculate the biofilm disruption rate. For live/dead staining, pre‐formed biofilms were co‐cultured with NPs as described above. The biofilms were then stained with SYTO 9 (5 µm) and propidium iodide (30 µm) in the dark for 15 min. After gentle washing with PBS to remove excess dye, the stained biofilms were visualized using the 3D layer scanning mode of CLSM.

2.7. Cell Viability and Live/Dead Assay

L‐929 fibroblasts were seeded in 24‐well plates at a density of 5 × 10⁴ cells per well in 500 µL of high‐glucose DMEM and cultured for 24 h at 37°C under 5% CO₂. For the CCK‐8 assay, cells were treated with CO/AB@MPDA NPs at concentrations of 0, 25, 50, 100, and 200 µg/mL (n = 5) for 24 and 72 h. After treatment, the medium was replaced with serum‐free medium containing 10% CCK‐8 reagent, followed by incubation at 37°C for 60 min. Absorbance was measured at 450 nm using a microplate reader. For live/dead staining, cells treated with Control, MPDA, CO@MPDA, and CO/AB@MPDA NPs (n = 3) for 24 and 72 h were stained with 100 µm propidium iodide (PI) and 100 µm fluorescein diacetate (FDA) in the dark at 37°C for 10 min. After three washes with PBS, the cells were imaged using an inverted fluorescence microscope.

2.8. Assessment of Intracellular ROS

L‐929 cells seeded in 24‐well plates were treated with Control, MPDA, CO@MPDA, and CO/AB@MPDA NPs (n = 3) and then exposed to 300 µm H₂O₂ to induce oxidative stress. After 24 h of incubation (37°C, 5% CO₂), the cells were washed with PBS and stained with 10 µm DCFH‐DA for 10 min at 37°C in the dark. After three washes with PBS, intracellular ROS levels were assessed using either CLSM or flow cytometry following trypsinization.

2.9. Analysis of Macrophage Phenotype

RAW264.7 cells were stimulated with LPS (100 ng/mL) and co‐treated with Control, MPDA, CO@MPDA, and CO/AB@MPDA NPs in high‐glucose DMEM for 72 h (37°C, 5% CO₂; n = 3). For immunofluorescence staining, cells were fixed with 4% paraformaldehyde (PFA) for 10 min, permeabilized with 0.3% Triton X‐100 (if required) for 5 min, and then incubated overnight at 4°C with anti‐CD86 or anti‐CD206 primary antibodies. After washing, the cells were labeled with corresponding fluorescent secondary antibodies at 37°C for 2 h, counterstained with DAPI for 5 min, and finally mounted in 95% glycerol for observation under a CLSM. For flow cytometry analysis, cells were harvested by trypsinization, followed by fixation and permeabilization. Subsequently, the cells were stained with directly conjugated anti‐CD86 or anti‐CD206 antibodies at 37°C for 30 min and analyzed using a flow cytometer.

2.10. Establishment of Type 2 Diabetes Mellitus (T2DM) Rat Model and In Vivo Antibacterial Evaluation

The animal surgery operation was carried out strictly in accordance with the existing law and the accompanying ethical standards. It received approval from Chongqing University's Laboratory Animal Welfare and Ethics Committee (CQU‐IACUC‐RE‐202507‐005). Insulin resistance, a hallmark of T2DM, typically coexists with β‐cell dysfunction. To recapitulate this pathophysiology, we established a T2DM Sprague–Dawley (SD) rat model by combining high‐fat diet (HFD)‐induced insulin resistance and partial β‐cell impairment via multiple low‐dose streptozotocin (STZ) injections. Male SD rats (mean weight 200 g) were housed under standard conditions in compliance with the guidelines of the Institutional Animal Ethics Committee of Chongqing University. The animals were fed a high‐fat/high‐sucrose diet for two weeks, after which they were fasted for 16 h and then intraperitoneally injected with STZ (30 mg/kg) dissolved in citrate buffer (pH 4.2–4.5, 10 mg/mL). Rats exhibiting fasting blood glucose levels >16.67 mmol/L on day 7 post‐injection were selected for subsequent experiments. Full‐thickness infected skin wounds (10 mm in diameter) were created on the shaved dorsum of each rat and inoculated with 50 µL of S. aureus suspension (1×10⁸ CFU/mL). The rats were randomly assigned to one of eight experimental groups (Control, MPDA, CO@MPDA, CO/AB@MPDA, each with or without NIR treatment; n = 15). On days 1 and 2, wounds were topically treated with 100 µL of the corresponding test samples, and those in the NIR groups were exposed to an 808 nm laser (1 W/cm², 10 min). On day 3, wound swabs were collected from five rats per group, cultured in Mueller–Hinton broth at 37°C for 12 h, serially diluted, plated on agar, and CFUs were enumerated after 12 h of incubation. On days 3, 7, and 14 post‐wounding, tissue samples were excised, fixed in 4% PFA, embedded in paraffin, and sectioned for immunofluorescence staining of TNF‐α (day 3), VEGF, CD86, and CD206 (day 7), as well as AGEs (day 14).

2.11. Evaluation of Biosafety In Vivo

H&E staining was performed on major organs (heart, liver, spleen, lung, and kidney) collected on day 14 post‐treatment from experimental rats, diabetic controls, and normal controls, following the same protocol used for skin tissue sections. For hematological assessment, venous blood was collected via cardiac puncture under anesthesia from all groups and subjected to complete blood count analysis as well as liver and kidney function tests. Hemocompatibility was evaluated via hemolysis assay. Briefly, red blood cells (RBCs) isolated from SD rat blood were washed with PBS, diluted to 5% (v/v), and incubated with CO/AB@MPDA at 37°C for 1 h. After centrifugation (1000 rpm), the absorbance of supernatants at 540 nm was measured, with deionized water and PBS as positive and negative controls, respectively.

2.12. Statistical Analysis

All experimental data are expressed as means ± standard deviations (SD) with n ≥ 3 independent replicates. Statistical analyses were performed using GraphPad Prism 8.0. Significance between groups was determined by one‐way ANOVA and t‐tests (p < 0.05, p < 0.01, p < 0.001; NS = not significant).

3. Results and Discussion

3.1. Synthesis and Characterization of CO/AB@MPDA NPs

As schematically depicted in Scheme 1, CO/AB@MPDA NPs were fabricated through a sequential loading strategy. TEM imaging (Figure S1) revealed that the pristine MPDA NPs exhibited a spherical morphology with a well‐defined mesoporous structure. Upon incorporation of FeCO to form CO@MPDA NPs, the mesopores appeared occluded (Figure S2), yet EDS elemental mapping confirmed a homogeneous distribution of iron throughout the NPs (Figure S3). Subsequent loading of AB via hydrogen bonding restored the mesoporous architecture (Figure 1A), which could be attributed to the hydrophobic encapsulation of FeCO within the MPDA pores and the hydrophilic crystallization of AB on the NPs surfaces. HRTEM analysis (Figure S4) showed lattice fringes localized specifically at the interfacial regions of CO/AB@MPDA NPs, indicating the presence of crystalline phases that were absent in CO@MPDA NPs. Selected area electron diffraction analysis (Figure S5) identified these crystalline regions as AB, matching the reference pattern (space group C2/M, PDF#12‐0484‐B10H14; ∠AOB tolerance ≤ 5°). EDS elemental mapping (Figure 1A) further verified the co‐localization of Fe and B within the NPs, although the boron background signal was intensified due to beam‐induced decomposition of AB during measurement. Despite this, EDS line scanning profiles (Figure S6) clearly revealed distinct enrichment of both B and Fe in the NPs relative to the background, providing preliminary validation of the successful synthesis of CO/AB@MPDA NPs. As seen from Figure S7, the hydrated particle sizes of the three particles were 145.4 ± 13.2, 155.9 ± 8.8, and 181.3 ± 10.9 nm, respectively. The gradual increase in hydrodynamic particle size further confirms the successful loading of FeCO and AB onto the NPs. Zeta potential measurements indicated a slight positive shift for CO@MPDA NPs compared to MPDA (Figure 1B). However, accurate quantification for CO/AB@MPDA NPs was impeded by the generation of hydrogen microbubbles during the measurement process. FTIR spectroscopy (Figure 1C) confirmed successful surface functionalization. Characteristic B–H stretching vibrations at 2370 cm⁻¹ were observed in both AB@MPDA and CO/AB@MPDA NPs, while the C = O stretching band of FeCO at 1680 cm⁻¹ was present in CO@MPDA and CO/AB@MPDA NPs. Notably, the intensity of the C = O peak was significantly enhanced in the CO/AB@MPDA composite. XPS survey scans (Figure 1D) identified the presence of C, N, O, Fe, and B in CO/AB@MPDA NPs. High‐resolution Fe2p spectra (Figure 1F) exhibited characteristic peaks at 711.0 and 724.0 eV, consistent with encapsulated FeCO. The B 1s spectra (Figure 1G) showed bonding signatures in the range of 192–193 eV, confirming the successful conjugation of AB. Deconvolution of the C1s spectra (Figure 1E and Figure S8) further elucidated the chemical environments, revealing contributions from C–H/C–C (284.4 eV), C–N (285.4 eV), C–O (286.3 eV), and a notably enhanced C = O (287.7 eV) intensity, indicative of coordination interactions within the composite structure. Subsequently, the stability of the prepared NPs was further evaluated. As shown in Figure S9A, MPDA, CO@MPDA and CO/AB@MPDA could disperse well and remain stable in pure water, PBS buffer, and DMEM medium containing 10% fetal bovine serum (FBS), with no obvious aggregation observed within seven days. The hydrodynamic size of CO/AB@MPDA after 7 days of dispersion was further measured. As presented in Figure S9B, compared with the initial state on day 1, only a slight increase in hydrodynamic diameter was detected for the NPs in different media after 7 days. This slight change might be attributed to weak particle aggregation, while the overall dispersion state remained stable. Those findings demonstrated that the as‐prepared NPs possessed favorable long‐term dispersion stability.

FIGURE 1.

(A) TEM image and element distribution map of CO/AB@MPDA NPs (scale bar: 100 nm); (B) Zeta potential of MPDA and CO@MPDA NPs; (C) FTIR spectra of standard samples for MPDA NPs, CO@MPDA NPs, AB@MPDA NPs, CO/AB@MPDA NPs, AB and FeCO; (D) XPS scan of CO/AB@MPDA NPs; High‐resolution XPS scan of (E) C1s, (F) Fe2p, and (G) B1s obtained from CO/AB@MPDA NPs.

As shown in Figure 2A, CO/AB@MPDA suspensions exhibited a clear laser power‐dependent temperature increase under 808 nm NIR irradiation, with temperature rises of +21.4°C (0.5 W/cm²), +25.5°C (1.0 W/cm²), +32.7°C (1.5 W/cm²), and +40.7°C (2.0 W/cm²) after 10 min of exposure. A concentration‐dependent photothermal effect was also observed under 1.0 W/cm² irradiation for 10 min (Figure 2B), resulting in temperature increases of +8.2°C (25 µg/mL), +10.5°C (50 µg/mL), +15.8°C (100 µg/mL), +20.0°C (200 µg/mL), and +26.2°C (300 µg/mL). These concentration‐ and power‐dependent heating profiles were visually corroborated by infrared thermal imaging (Figure 2C). The photothermal conversion efficiency, calculated from the heating‐cooling curves (Figure 2D), reached 20.3%. Excellent photostability was confirmed, as less than 5% attenuation in temperature elevation was observed over three consecutive NIR on/off cycles.

FIGURE 2.

(A) Temperature elevation profiles of CO/AB@MPDA NPs (200 µg/mL) under 808 nm NIR irradiation at different power densities; (B) Concentration‐dependent temperature changes under 808 nm NIR irradiation (1.0 W/cm²); (C) Infrared thermal images of PBS and CO/AB@MPDA NPs suspension during NIR irradiation; (D) Photothermal stability was assessed by temperature changes over three on/off cycles of NIR irradiation (200 µg/mL); (E) UV–vis spectra showing the conversion of reduced r‐Hb to r‐Hb‐CO for CO detection; (F) CO release profile from CO/AB@MPDA NPs (200 µg/mL) with or without NIR irradiation; (G) H₂ release at pH 7.4 and (H) pH 5.8, showing pH‐dependent release behavior; (I) Quantitative comparison of H₂ release from CO/AB@MPDA NPs (200 µg/mL) under different pH conditions over 24 h.

Based on the thermal instability of FeCO and the hydrolytic susceptibility of AB, the dual‐stimuli‐responsive gas release behavior of the NPs were characterized. Upon photothermal activation, FeCO undergoes thermal decarbonylation, in which localized heating from the MPDA carrier cleaves the Fe─CO bonds to rapidly release CO in a controlled, on‐demand manner [33]. In contrast, AB hydrolysis generated H₂ through an acid‐catalyzed cleavage of B─H bonds, a process greatly accelerated in the proton‐rich, acidic microenvironment typical of bacterial‐infected tissues. Such distinct response mechanisms enable precise spatiotemporal gas release triggered by external light and internal pathological microenvironment [39]. Thermogravimetric analysis (TGA, Figure S10A) of CO/AB@MPDA NPs revealed a two‐stage mass loss: 6.5% between 80°C and 110°C and 9.88% between 110°C and 140°C, corresponding to the decomposition of AB and FeCO, respectively. These transitions were further corroborated by distinct DTG peaks. Qualitative confirmation of NIR‐triggered CO release was obtained using an AFE‐Pd(II) fluorescence assay (Figure S10B), in which the reductive gas reduced Pd(II) to Pd(0), initiating a Tsuji–Trost‐type allylic substitution that liberated the fluorophore. UV–vis spectroscopy verified the successful loading of FeCO onto MPDA NPs, as indicated by a characteristic absorption band at approximately 600 nm (Figure S11). After repeated washing cycles, the absorbance of the supernatant was measured, revealing a loading capacity of 1.8 mg of FeCO per 2.0 mg of NPs, corresponding to a loading efficiency of 47.4%.

NIR‐triggered CO release was monitored in real time using a standard hemoglobin assay, which quantifies CO by spectrophotometrically tracking the conversion of reduced r‐Hb to r‐Hb‐CO, characterized by distinct absorption maxima at 430 and 410 nm, respectively. Under 808 nm NIR irradiation at 1 W/cm² (Figure 2E), CO/AB@MPDA NPs exhibited rapid CO release, reaching 2.4 µm within 10 min, followed by a marked slowdown upon laser cessation (Figure 2F). This on‐demand release behavior confirmed the feasibility of NIR‐controllable CO delivery, which was expected to facilitate targeted antibacterial outcomes. The H₂ release from CO/AB@MPDA NPs was quantitatively assessed using an MB‐Pt probe. This method was based on the H₂‐induced catalytic reduction of MB to colorless MBH₂ on Pt surfaces, monitored via the characteristic absorption peak at 664 nm (Figure S12). Under physiological conditions (pH 7.4), the NPs released H₂ slowly over 24 h (Figure 2G). In contrast, under acidic infection‐mimicking conditions (pH 5.8), H₂ generation was significantly accelerated (Figure 2H), demonstrating proton‐enhanced hydrolysis as the underlying pH‐responsive mechanism. Quantitative analysis of H₂ release over 24 h under different pH conditions was summarized in Figure 2I, clearly illustrating the H⁺‐regulated release behavior and adaptive properties of the NPs. Together, these results validated the capacity of CO/AB@MPDA NPs for microenvironment‐driven, stimulus‐responsive gas release.

Thus, we engineered mesoporous polydopamine co‐loaded with iron carbonyl and ammonia borane, yielding the composite denoted as CO/AB@MPDA NPs. The physicochemical properties of the resulting NPs were systematically characterized using TEM, XPS, XRD, and FTIR spectroscopy. Furthermore, the material exhibited notable photothermal conversion efficiency and stimuli‐triggered gas release capabilities, confirming its potential for therapeutic applications.

3.2. In Vitro Assessment of Antibacterial Efficacy of CO/AB@MPDA NPs

The antimicrobial performance of CO/AB@MPDA NPs against S. aureus and E. coli was systematically evaluated through minimum inhibitory concentration assays, plate counting, SEM, and live/dead staining. As shown in Figure S13A,B, dose‐dependent antibacterial effects were observed under NIR irradiation (1 W/cm²). At a concentration of 200 µg/mL, the NPs exhibited potent bactericidal activity, achieving inhibition rates of 95.7% against S. aureus and 97.5% against E. coli. Plate counting assays further revealed distinct bactericidal efficacies across the experimental groups (Figure 3A, B). For S. aureus (Figure S14), the inhibition rates were as follows: PBS (0.9%), PBS + NIR (3.1%), MPDA (8.8%), MPDA + NIR (45.3%), CO@MPDA (33.2%), CO@MPDA + NIR (91.3%), CO/AB@MPDA (37.3%), and CO/AB@MPDA + NIR (97.1%). A similar trend was observed for E. coli (Figure S15): PBS (1.2%), PBS + NIR (8.2%), MPDA (11.4%), MPDA + NIR (58.4%), CO@MPDA (44.2%), CO@MPDA + NIR (93.6%), CO/AB@MPDA (46.8%), and CO/AB@MPDA + NIR (98.4%). Three key patterns emerged from the above data. (1) A significant photothermal enhancement was observed (Δ > 42% for MPDA‐based groups under NIR, p < 0.001), confirming the essential role of photothermal therapy. E. coli exhibited higher thermosensitivity, likely due to its thin peptidoglycan layer and metabolically active outer membrane. (2) CO‐releasing groups (CO@MPDA + NIR vs. CO@MPDA: Δ58.1%, p < 0.001) demonstrated a synergistic gas‐photothermal bactericidal mechanism. (3) CO/AB@MPDA + NIR achieved near‐complete eradication (97.1%–98.4%) through concurrent NIR‐triggered CO release and proton‐responsive H₂ liberation, both of which disrupted respiratory chain function. The overall higher efficacy against E. coli (+3.3% mean inhibition) aligned with the inherent thermal vulnerability of Gram‐negative bacteria.

FIGURE 3.

(A) Inhibition of S. aureus and (B) E. coli by different samples; (C, D) Live/dead fluorescence staining of S. aureus and E. coli after treatment (scale bar: 25 µm); (E, F) SEM images showing morphological changes in S. aureus and E. coli (scale bar: 1 µm); (G) Intracellular ROS detection in bacteria and (H) quantitative analysis of the mean fluorescence intensity (scale bar: 25 µm). Data are presented as mean ± SD (n = 5); ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001.

Live/dead staining with SYTO9/PI corroborated these results, showing markedly increased red fluorescence (dead cells) in NIR‐irradiated groups (Figure 3C,D). Bimodal fluorescence quantification (Figure S16) indicated significantly higher mortality in photothermally activated CO@MPDA (Δviability > 55% vs. dark controls) and CO/AB@MPDA groups (Δ > 60%), consistent with plate counting data. Notably, the minimal difference (Δ < 5%) between CO@MPDA + NIR and CO/AB@MPDA + NIR against both species suggested that bactericidal activity was primarily driven by NIR‐triggered CO burst release, which disrupted respiratory chain function, while H₂ liberation played a negligible role as an upstream effector. It was consistent with the limited H₂ yield under experimental conditions and its redundant contribution to respiratory inhibition compared to CO.

SEM imaging revealed critical bacterial membrane damage induced by photothermal‐gas synergy (Figure 3E,F). Both S. aureus and E. coli retained intact, smooth surfaces in PBS and MPDA groups. In contrast, dark‐treated CO@MPDA and CO/AB@MPDA groups showed subtle surface buckling, attributable to osmolarity imbalance caused by respiratory chain disruption from CO/H₂ accumulation. Under NIR irradiation, CO@MPDA and CO/AB@MPDA groups exhibited drastic membrane collapse and cytoplasmic leakage, accompanied by extensive deformation of surviving cells. These morphological changes conclusively demonstrated the combined effect of photothermal disruption and gas‐mediated cytotoxicity. Subsequently, bacterial protein leakage assay and ONPG hydrolysis assay were adopted to further verify the damage effect of photothermal effect and CO on bacterial membrane integrity. The results revealed a distinct increase of total protein content in the bacterial supernatant (Figure S17A). Meanwhile, the CO/AB@MPDA + NIR group exhibited markedly higher absorbance at 405 nm (Figure S17B), which implied massive release of intracellular β‐galactosidase. These results confirmed that the combined treatment of CO/AB@MPDA+NIR irradiation could kill bacteria by disrupting bacterial membrane integrity and triggering the leakage of intracellular macromolecules. High levels of CO could disrupt redox homeostasis during bacterial cellular respiration. In terms of CO‐mediated antibacterial effects, such metabolic disturbance further triggers the accumulation of endogenous ROS [40]. The DCFH‐DA probe was adopted to evaluate intracellular ROS levels in bacteria. As shown in Figure 3G,H, obvious ROS fluorescence signals were observed in all groups treated with NIR irradiation. Notably, the CO@MPDA and CO/AB@MPDA groups displayed much higher fluorescence intensity. These results fully demonstrate that sustained CO release interferes with bacterial respiratory metabolism and redox balance, thereby inducing excessive ROS accumulation and accelerating bacterial death.

The efficacy of different samples in inhibiting and disrupting S. aureus biofilms was assessed using crystal violet staining and 3D confocal fluorescence imaging. As shown in Figure 4A, biofilm integrity was visualized via SYTO 9/PI dual staining. Quantitative analysis of live/dead staining (Figure 4B) revealed that biofilm structures in the control and MPDA groups remained dense and structurally intact, with predominantly green fluorescence (viable bacteria), indicating minimal impact on bacterial metabolic activity. Under dark conditions, localized red fluorescence regions (bacterial mortality <40%) were observed within the biofilm, though the overall architecture remained largely undisturbed. These findings were consistent with prior antibacterial results, suggesting that sustained release of CO/H₂ alone could induce partial bacterial death but has limited efficacy in disrupting the biofilm matrix. In contrast, the combination of photothermal treatment with CO/H₂ release resulted in extensive red fluorescence coverage (90%–96%), indicative of widespread bacterial death and effective disintegration of the biofilm's three‐dimensional structure. The biofilm‐disrupting capability of the NPs was further confirmed by crystal violet assays. As shown in Figure 4C and quantified in Figure 4D, the residual biomass after co‐culture with different materials was statistically analyzed. The control group showed similar OD values with or without NIR irradiation (2.5 ± 0.3), confirming that the biofilm structure remained largely unaffected. In contrast, a marked reduction in biofilm biomass was observed in the NPs treated groups, with a clear photothermal‐triggered enhancement. Specifically, the CO/AB@MPDA NPs group under NIR irradiation exhibited a final OD value of 0.8 ± 0.1, corresponding to only 32% ± 0.4% of the biomass retained relative to the control, underscoring the material's potent biofilm disruption ability. In addition, the inhibitory effect of the NPs on early‐stage biofilm formation was investigated (Figures S18 and S19). Unlike the disruption assays, these inhibition studies evaluated the ability of the NPs to prevent biofilm development after incubation with high‐concentration bacterial suspensions. This model in vitro simulated the potential of nanomaterials to suppress biofilm recurrence following wound debridement. The results demonstrated that the NPs effectively inhibited biofilm formation, supporting their potential use in further studies in vivo.

FIGURE 4.

(A) 3D images and (B) live/dead fluorescence area distribution of S. aureus biofilms; (C) Representative images and (D) quantitative analysis of crystal violet‐stained biofilms; (E) CLSM images and (F) corresponding fluorescence intensity analysis of CO probe‐stained biofilms under different treatments. Data are presented as mean ± SD (n = 5); ^*** p < 0.001. Scale bar: 25 µm.

To further elucidate the antibacterial mechanism of CO/AB@MPDA NPs, a CO fluorescent probe was employed to evaluate the fluorescence intensity after co‐culture with biofilms (Figure 4E,F). This approach validated the gas release capability of the NPs and established the connection between CO/H₂ release and photothermal synergy. The control and MPDA groups, which lacked carbonyl iron and ammonia borane, showed only weak or background fluorescence signals, confirming the absence of interference from probe autofluorescence or non‐specific binding. The CO@MPDA group exhibited low fluorescence intensity in the absence of NIR irradiation, indicating minimal CO release under dark conditions. However, upon NIR irradiation, CO@MPDA demonstrated significantly enhanced fluorescence intensity, confirming the photothermal‐triggered CO release mechanism. In the case of the CO/AB@MPDA group, the similar redox potentials of H₂ and CO enabled both gases to reduce Pd²⁺ and activate the fluorescent probe. The increased fluorescence intensity observed in this group reflects the combined release of CO and H₂, demonstrating that CO/AB@MPDA NPs could effectively release both gases under the dual regulation of photothermal activation and the acidic microenvironment generated during co‐culture with high‐concentration bacteria.

3.3. In Vitro Biocompatibility and Immunomodulatory Evaluation

The cytocompatibility of CO/AB@MPDA NPs was systematically assessed using CCK‐8 assays and live/dead staining in L929 fibroblasts. CCK‐8 results (Figure 5A,B) revealed negligible cytotoxicity after 24 h of exposure. Following prolonged incubation for 72 h, cell viability remained above 90% (90.0% ± 1.0%) even at the highest NP concentration tested (200 µg/mL), indicating a well‐controlled and time‐dependent cytotoxic profile. Hemolysis tests performed at various NPs concentrations (Figure S20) showed no notable hemolytic activity, further supporting the excellent hemocompatibility of CO/AB@MPDA NPs. Live/dead staining results (Figure 5C) were consistent with the CCK‐8 findings. After 24 h treatment with 200 µg/mL NPs, similar green fluorescence intensities were observed across groups, with only sporadic red fluorescence signals. After 72 h, robust cell proliferation was evident, accompanied by minimal cell death. Together, these data conclusively demonstrated the favorable biocompatibility of CO/AB@MPDA NPs at concentrations up to 200 µg/mL over a 72 h exposure period.

FIGURE 5.

(A, B) Viability of L929 cells after treatment with NPs for 1 and 3 days; (C) Live/dead staining of L929 cells co‐cultured with different formulations for 1 and 3 days (scale bar: 400 µm); (D, E) Immunofluorescence staining and flow cytometry analysis of CD206 expression in RAW264.7 macrophages; (F, G) Immunofluorescence staining and flow cytometry analysis of CD86 expression in RAW264.7 macrophages (scale bar: 50 µm); (H, I) Intracellular ROS scavenging capacity in L929 cells assessed by fluorescence imaging and flow cytometry (scale bar: 200 µm).

In diabetic wounds, where persistent M1 macrophage polarization (pro‐inflammatory, TNF/LPS‐driven) critically impeded the healing process [41], so the immunomodulatory potential of CO/AB@MPDA NPs was evaluated using RAW264.7 macrophages. Flow cytometry and immunofluorescence staining (Figure 5D–G) revealed a pronounced shift in macrophage polarization markers. NP‐treated groups showed significant upregulation of CD206⁺ M2 macrophages along with downregulation of CD86⁺ M1 phenotypes compared to control groups. The underlying regulatory mechanism was closely related to the synergistic effects of CO and H₂ released from the NPs. CO specifically inhibited the activation of the NF‐κB signaling pathway, which was a key cascade mediating M1 polarization and pro‐inflammatory cytokine secretion, thereby reducing the expression of pro‐inflammatory factors and blocking excessive M1 activation [40]. Meanwhile, H₂, as a selective antioxidant, alleviated intracellular oxidative stress, further weakening the pro‐inflammatory microenvironment and promoting the expression of M2‐related markers [42]. This dual‐gas mediated signaling regulation directly drived the phenotypic switch of macrophages. These findings indicated that sustained release of CO and H₂ from the NPs effectively promoted repolarization from the M1 to the M2 phenotype, thereby steering the immune microenvironment toward an anti‐inflammatory and pro‐regenerative state. Within the complex diabetic wound environment—where oxidative stress and chronic inflammation jointly hindered tissue regeneration [43]—the intracellular ROS‐scavenging capacity of the NPs was further assessed using a DCFH‐DA probe. Fluorescence imaging and flow cytometry (Figure 5H,I) showed a strong ROS burst in H₂O₂‐treated control groups (300 µm), as indicated by intense green fluorescence. By contrast, all NPs treated groups exhibited markedly reduced fluorescence intensity, demonstrating a significant ROS‐eliminating capacity that was essential for alleviating oxidative damage in the diabetic microenvironment.

3.4. In Vivo Evaluation of Antibacterial Efficacy and Wound Healing

The antibacterial and tissue‐repair functions of the NPs were further validated through animal experiments. Full‐thickness infected wounds were established on the dorsum of diabetic rats and subjected to different treatments. Wound areas were photographed and statistically analyzed throughout the healing process. As shown in Figure 6A,B, the initial wound areas (10 mm diameter) were comparable across all groups on day 1. By day 7, the NIR‐activated CO/AB@MPDA group exhibited over 70% wound closure, with newly formed pink epithelial tissue showing no exudation or scab formation. By day 14, the CO/AB@MPDA + NIR group achieved complete re‐epithelialization, and the resulting scar displayed elasticity approaching that of normal skin. H&E staining of wound sections collected on days 7 and 14 was performed to evaluate tissue remodeling (Figure 6E). On day 7, wounds were predominantly in the proliferative phase. The control group exhibited a pronounced wound gap and the thinnest granulation tissue, consistent with an impaired healing microenvironment. In contrast, the CO/AB@MPDA + NIR group showed enhanced granulation tissue formation and re‐epithelialization, with a more continuous wound surface and significantly greater thickness of both granulation and epithelial layers compared to other groups. By day 14, the healing process had entered the remodeling phase. Quantification of the remaining unhealed areas revealed that the CO/AB@MPDA + NIR group had the smallest area of incomplete repair, underscoring the outstanding wound‐healing promotion capacity of the NPs. Corresponding quantitative data are provided in Figure S22. Infrared thermal imaging confirmed that wounds treated with CO/AB@MPDA NPs (200 µg/mL) reached a controlled temperature of approximately 45°C during 10 min of NIR irradiation (Figure S21), which aligned with the requirements for mild photothermal therapy. Quantitative bacteriological analysis on day 3 (Figure 6C,D) showed a significant reduction in bacterial load in the photothermally activated groups, particularly in the CO@MPDA + NIR and CO/AB@MPDA + NIR groups. This in vivo antibacterial efficacy was consistent with prior results. Together, these findings demonstrated the synergistic therapeutic effect of the NPs under NIR activation: mild photothermal sterilization combined with accelerated tissue regeneration.

FIGURE 6.

(A) Representative photographs of wound closure under different treatments (initial wound diameter: 10 mm); (B) Quantitative analysis of wound area ratio expressed as A/A₀, where A and A₀ represent wound area and initial wound area, respectively; (C) The antibacterial ability of NPs on wounds using plate counting method; (D) Statistical analysis of bacterial inhibition rates; (E) H&E staining images of wound sections at different healing stages. Data are presented as mean ± SD (n = 3); ^*** p < 0.001.

Masson's trichrome staining of diabetic wounds on days 7 and 14 revealed enhanced tissue structural repair in the MPDA + NIR group compared with the control group (Figure 7A). The treated wounds displayed a more compact extracellular matrix and restored dermal‐epidermal junction integrity. Quantitative analysis further demonstrated a significant increase in collagen deposition in the CO/AB@MPDA + NIR group (Figure S23). These results confirmed that the NPs facilitated tissue remodeling under NIR irradiation, thereby promoting structural and functional wound recovery through gas‐mediated anti‐inflammatory mechanisms.

FIGURE 7.

(A) Masson's trichrome staining of wound sections on days 7 and 14 (scale bar: 200 µm); Immunofluorescence staining and corresponding quantitative analysis of (B, F) VEGF, (C, G) CD86, (D, H) CD206, and (E, I) AGEs in wound tissues (scale bar: 100 µm). Data are presented as mean ± SD; ^** p < 0.01, ^*** p < 0.001.

Early ROS and TNF‐α levels during wound healing were evaluated (Figure S24). DHE fluorescence imaging (Figure S24A,C) revealed a significant reduction in ROS levels in NPs treated wounds, effectively mitigating diabetes‐aggravated oxidative damage that typically impairs fibroblast proliferation. Concurrently, TNF‐α immunofluorescence quantification (Figure S24B,D) demonstrated decreased expression of this inflammatory cytokine in the NPs treated group, thereby disrupting the vicious cycle of chronic inflammation that inhibits cell migration and collagen synthesis. As shown in Figure 7B, during the proliferative phase of healing (day 7), VEGF expression was markedly upregulated in the groups receiving CO/H₂‐releasing NPs, effectively promoting angiogenesis and alleviating the characteristic hypoxia and nutrient deficiency in diabetic wounds. Macrophage polarization analysis further confirmed that the treatment successfully shifted the in vivo macrophage phenotype from pro‐inflammatory M1 to pro‐reparative M2 (Figure 7C,D), thereby reversing the M1‐dominant pathological state that impeded diabetic wound healing. Notably, these effects collectively contributed to metabolic reprogramming within the wound microenvironment. By day 14, NPs treated wounds exhibited reduced accumulation of AGEs (Figure 7E), attributable to improved insulin sensitivity resulting from effective ROS scavenging and suppression of non‐enzymatic glycation. In summary, NPs enabled gas therapy systemically remodeled the diabetic wound microenvironment through four coordinated mechanisms: oxidative stress alleviation, inflammation resolution, angiogenesis promotion, and metabolic normalization. The insulin‐sensitizing effect mediated by CO/H₂ release primarily stems from synergistic ROS scavenging and anti‐inflammatory actions, collectively driving efficient diabetic wound repair [44, 45].

3.5. Biosafety Assessment

The biosafety of nanomaterials was a critical consideration for their clinical translation. Body weight monitoring during the treatment period (Figure S25) further supported the therapeutic benefits. Diabetic control rats exhibited progressive weight loss, whereas NP‐treated groups maintained weight trajectories close to normal baseline levels, suggesting systemic metabolic improvement. Histopathological evaluation of major organs (heart, liver, spleen, lung, and kidney) collected on day 14 (Figure S26) revealed no signs of structural damage or pathological changes in any treatment group. In addition, complete blood count analysis (Figure S27) showed all hematological parameters within normal ranges after NPs treatment. These results collectively demonstrated the favorable biosafety profile of CO/AB@MPDA NPs, supporting their potential for future clinical application in the management of diabetic wound infections.

4. Conclusion

This study developed a multifunctional CO/AB@MPDA nanoplatform through a nanoemulsion‐assisted assembly strategy using MPDA as the carrier. The amphiphilic nature of MPDA enabled simultaneous incorporation of the hydrophobic CO donor (FeCO) and the hydrophilic H₂ donor (AB). Under NIR irradiation, the fabricated NPs exhibited synergistic photothermal effects and triggered rapid CO release, leading to potent antibacterial activity. In the acidic infected wound microenvironment, pH‐responsive H₂ release further enhanced antioxidant capacity and promoted inflammation resolution. Comprehensive biosafety assessment, including hemolysis tests (<5% hemolysis), hematological analysis, and histopathological examination, confirmed minimal systemic toxicity. Together, in vitro and in vivo validation demonstrated that the CO/AB@MPDA system offers triple therapeutic modalities (antimicrobial action, anti‐inflammatory regulation, and pro‐healing functionality), presenting a promising strategy for clinical management of diabetic wounds. While its large‐scale production process and clinical applicability still require further optimization, the design idea of the CO/AB@MPDA nanoplatform can serve as a theoretical reference for developing new‐generation intelligent tissue repair materials, facilitating the transition of antibacterial nanomaterials from basic laboratory research to practical clinical use.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adhm71403‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.