Gold nanocluster-complexed dressings for instant closure of bacteria-infected wound

Abstract

Rapid closure is crucial for effective wound healing. However, current wound dressings often fall short in this aspect. Herein, we present a light-responsive wound dressing (A@MC) composed of a heat-curable methyl cellulose (MC) hydrogel and aminophenylboronic acid (ABA)-modified gold nanoclusters (AGNCs). The A@MC is injectable and can conform, connect, and adhere to wounds of irregular shapes before gelation. Upon near-infrared (NIR) light irradiation, the photothermal effect of the AGNCs triggers an instant gelation and contraction of the dressing and simultaneously a sharp shrinkage of the wound. It achieves a shrinkage rate up to 46% within 5 min for bacteria-infected wound in mice. Moreover, the contraction of A@MC can achieve a rapid release (16%, 5 min) of the AGNCs, which can bind to bacteria through ABA-lipopolysaccharide (LPS) interaction, displaying a targeted antibacterial therapy. The A@MC realizes an accelerated healing (complete healing within 14 days) of multidrug-resistant Pseudomonas aeruginosa (P.a)-infected wounds.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12951-026-04697-4.

Introduction

Wound contraction plays a crucial role in the healing process, and achieving rapid yet moderate shrinkage offers notable benefits for accelerated recovery [1–2]. By promptly reducing the wound area and depth, tissue exposure to the external environment is limited, which decreases the amount of tissue that must regenerate and shortens the distance for epidermal cell migration [3–4]. This facilitates fast wound closure and quick reestablishment of the epidermal barrier [5–6]. Recent research has increasingly focused on active contractive materials to respond to body temperature, such as poly(N-isopropylacrylamide) (PNIPAM) and its copolymers [7], elastin-like polypeptides (ELPs) [8], and Pluronic [9]. These materials not only shrink upon heating but also scavenge reactive oxygen species (ROS) and regulate signaling pathways like EGFR/AKT to promote fibroblast proliferation and collagen production [10]. However, in infected wounds, contraction becomes more complex [11]. While it may provide some protective benefits, it can also heighten the risk of infection. Thus, the combination of wound contraction and antibacterial materials is essential for treating infected wounds.

In recent years, a wide range of metal-based nanomaterials have shown potent antibacterial activity against wound infections caused by multidrug-resistant (MDR) bacteria [12–15]. Among these nanomaterials, gold nanomaterials stand out for their unique physicochemical properties, excellent biocompatibility, and multiple antibacterial mechanisms [16–19]. In previous study, we successfully developed gold nanoclusters (GNCs) that exhibited potent antibacterial activity against MDR bacteria [20]. Additionally, upon activation by near-infrared (NIR) light, GNCs can generate localized high temperatures (> 50 °C) for photothermal therapy (PTT) [21–23]. While the PTT effects enhance the antibacterial efficacy of GNCs, non-selective therapy can inevitably damage host tissues. Thus, the GNCs that can target bacteria are desirable. Inspired by the strong affinity of boronic acid groups for polysaccharides, our previous work demonstrated that aminophenylboronic acid (ABA)-modified gold nanoparticles (A-AuNPs) could recognize bacteria through the interaction between ABA and lipopolysaccharide (LPS) [24], showing a pronounced calorimetric response between LPS and A-AuNPs with a dissociation constant of 2.29 µM [25–26]. However, the PTT capability of the previously reported GNC has not been fully explored. In this study, by adjusting synthesis method, we enable the AGNC to have a high photothermal effects responding to low intensity NIR light while maintaining its bacterial targeting and bactericidal activity.

Hydrogels offer notable benefits for wound healing by maintaining a moist environment, promoting hemostasis, and enabling controlled drug release [27–29]. These features help minimize the need for extended medical interventions and significantly reduce the pain and secondary damage commonly caused by traditional wound dressings. PNIPAM offers tunable lower critical solution temperature but is non-degradable and has poor biocompatibility [30]. ELPs provide excellent biocompatibility and no immunogenicity, but they are costly and mechanically weak [31]. Pluronic possesses favorable thermogelation properties but is prone to dilution and rapid drug release [32]. In contrast, methyl cellulose (MC) is widely used because of its excellent biocompatibility [33–35], easy preparation, good skin adhesion, adjustable gelation temperature, wide availability, and low cost [36]. MC demonstrates excellent moisture absorption and retention capabilities, helping to maintain a moist wound environment and promote autolytic debridement [37]. Its porous structure, which closely resembles the extracellular matrix, enhances cell migration and accelerates healing [38–39]. Additionally, its temperature-sensitive properties allow the hydrogel to solidify at body temperature, a process that can be further enhanced by the photothermal properties of the incorporated antibacterial agents, enabling in situ gelation and reducing the risk of dressing displacement [40–41].

Herein, we propose an AGNC-incorporated MC hydrogel wound dressing (A@MC) that responds to light irradiation by simultaneously inducing rapid wound contraction through a thermally triggered curing mechanism and exerting potent antibacterial effects (Scheme 1). When applied to an irregularly shaped wound, the liquid-like A@MC rapidly spreaded across the entire wound. AGNCs contain abundant hydroxyl groups can bind to wound surface protein or collagen by forming hydrogen bonds between them, enabling A@MC to quickly transform into an adhesive hydrogel. Under low-energy NIR irradiation, the photothermal effect of AGNCs enhanced the adhesion between dressing and the tissue by disrupting the hydration layer of MC and exposing previously hidden hydrophobic methyl groups. This process simultaneously promoted a rapid gel contraction and a passive wound shrinkage. In addition, light-triggered contraction expelled AGNC from the shrinking gel network into the wound tissue, creating a high local concentration of AGNCs. The released AGNCs selectively bond to bacteria through ABA-mediated interactions with lipopolysaccharide, enabling targeted antibacterial therapy [25]. Thus, the photothermal action of AGNC drives stronger tissue adhesion, gel contraction, and synergistic antibacterial effects. The combined effects of rapid wound shrinkage and bacteria-targeted antibacterial action made A@MC an excellent dressing for accelerated and safe healing of infected wounds (Scheme 1). Compared with previous studies, the prominent innovation of this work lies in the thermosensitive contractility of A@MC hydrogels and reduced inflammatory responses. A@MC enables hydrogel contraction simultaneously during photothermal polymerization, and effectively adheres wound tissue to the dressing via photothermal effect, thus achieving photoresponsive wound contraction. In addition, the AGNCs released from the hydrogel displayed bacterial targeting ability, which can realize targeted antibacterial activity while accelerating wound healing. Moreover, A@MC can not only kill bacteria but also reduce wound inflammation and systemic inflammatory responses. Thus, A@MC paves a way for accelerating wound closure by combining physical wound contraction and inflammatory environment control.

Scheme 1.

Schematic illustration of AGNCs-complexed MC hydrogel which can lead to wound contraction and realize targeted antibacterial effects to accelerate wound healing

Results and discussion

Characterization of AGNCs

We synthesized aminophenylboronic acid (ABA)-modified gold nanoclusters (AGNCs) through the reduction of HAuCl₄ with NaBH₄, based on our previously established methods [19]. Transmission electron microscopy (TEM) revealed that the AGNCs were monodisperse, exhibiting a uniform spherical shape with an average diameter of approximately 3 nm (Fig. 1A). High-resolution TEM further confirmed the crystalline structure of the AGNCs, showing a lattice spacing of 2.008 Å, which corresponds to the (200) crystal planes of gold. Dynamic light scattering (DLS) analysis showed a single and sharp peak, confirming that AGNCs are uniformly and stably dispersed in water (Figure S1). Given the generally negative surface charge of bacterial cell walls and membranes, AGNCs with a positive zeta potential of + 4.8 mV are well suited for electrostatic interactions with the bacterial surfaces. To validate the elemental composition of AGNCs, X-ray photoelectron spectroscopy (XPS) was performed. The Au 4f binding energies were around 83 eV and 87 eV (Fig. 1B-C), indicating the successful formation of gold nanoclusters (GNCs). XPS analysis further revealed the presence of mixed valence states in AGNCs (Fig. 1C), including Au⁰ (82.40 eV and 86.15 eV), Au⁺ (83.79 eV and 86.43 eV), and Au³⁺ (90.01 eV and 93.90 eV). The presence of N 1s peaks additionally confirmed the successful modification of ABA to the GNCs (Figure S2). AGNCs exhibited a spectrum similar to that of ABA, with a characteristic peak at 1094 cm^-1. This peak corresponds to the shift from the boric acid peak at 1342 cm^-1 to the boronate ester peak (Figure S3), also confirming the successful modification of the GNCs. AGNCs exhibited a strong fluorescence emission peak at 1070 nm when excited at 808 nm (Figure S4), which lies within the biological NIR-II window. As a result, AGNCs can effectively convert the absorbed laser energy into heat. At a low power density of 0.5 W/cm², even a low concentration (50 µg/mL) of AGNC resulted in a temperature rise of around 20 °C within 5 min (Fig. 1D). Notably, at 200 µg/mL, the temperature increased from 20.3 °C to 69.1 °C after 10 min of irradiation (Fig. 1D). AGNCs at varying concentrations (50, 100, 200, and 400 µg/mL) showed a concentration-dependent photothermal effect when irradiated with a 808 nm laser at 0.5 W/cm² for 10 min (Fig. 1E). Based on our previous methods [20], the photothermal conversion efficiency of AGNCs was calculated to be approximately 53.3% (Figure S5). We further evaluated the photothermal reversibility and stability of AGNCs across multiple irradiation cycles. The AGNCs at a concentration of 100 µg/mL consistently retained their photothermal performance throughout 6 repeated heating-cooling processes (Fig. 1F).

Fig. 1.

Characterization of AGNCs. (A) TEM images of GNCs and AGNCs, with the inset highlighting the distinct crystalline structures. (B) XPS spectra showing mixed valence states of Au 4f in GNCs. (C) XPS spectra showing mixed valence states of Au 4f in AGNCs. (D) Thermal images of AGNCs at different concentrations captured at various time intervals during exposure to a 808 nm laser (0.5 W/cm²). (E) Temperature elevation of AGNCs at different concentrations under 808 nm NIR laser irradiation (0.5 W/cm²). (F) Temperature profiles of AGNCs during 6 consecutive heating-cooling cycles

Antibacterial activity of AGNCs

We evaluated the antibacterial activity of AGNCs against various Gram-negative bacteria and their corresponding MDR strains. Without light irradiation, the minimum inhibitory concentration (MIC) of AGNCs ranged from 5 to 20 µg/mL (Table S1). Among them, AGNCs exhibited the strongest antibacterial effect against Pseudomonas aeruginosa (P. a) and its MDR strain, which were selected for subsequent experiments. We evaluated the bactericidal activity of AGNCs combined with PTT. After 5 min of irradiation, the colony forming units of P. a were rapidly destroyed (Figure S6), which was comparable to that of antibiotics and faster than that of AGNCs without light exposure. We also evaluated bactericidal effects of NIR irradiation, and NIR exposure alone did not exhibit bactericidal activity (Figure S7). We also used scanning electron microscopy (SEM) to investigate the morphological change of P. a and MDR P. a after different treatments. The bacteria treated with AGNCs exhibited significant structural damage, including noticeable aggregation and collapse, indicating compromised bacterial integrity (Fig. 2A). Exposure to NIR light further amplified these effects in both P. a and MDR P. a. To further clarify the antibacterial effects of NIR irradiation, we monitored temperature changes in the culture medium. Under low power irradiation at 0.5 W/cm² for 5 min, AGNCs induced a noticeable temperature increase, whereas NIR irradiation alone caused no significant change in LB medium (Figure S8). We also evaluated antibacterial effects of NIR irradiation alone. Compared with the non-irradiated group, NIR exposure alone did not exhibit antibacterial activity (Figure S9). Additionally, the inhibition zone assay confirmed that NIR irradiation markedly enhanced the antibacterial activity of the AGNCs (Figure S10, Table S2). Owing to the excellent PTT properties of AGNCs, their antibiofilm activity was further evaluated. At a concentration of 100 µg/mL, up to 97% of MDR P. a biofilm was removed under NIR irradiation, whereas approximately 85% biofilm was removed without irradiation (Figure S11), indicating that AGNCs can disrupt the biofilm and the photothermal effect significantly enhances bacterial eradication efficiency. To further explore the antibacterial mechanism of AGNCs combined with PTT, we assessed bacterial membrane permeability using SYTO9/PI staining and measured total ROS levels with a commercial ROS detection kit. In the control group, bacteria remained viable, displaying uniform and intense green fluorescence. AGNCs without NIR irradiation had only a mild impact on membrane integrity, with about 60% of bacteria showing reduced viability (Fig. 2B-C). Remarkably, a 5 min NIR irradiation increased the killing efficiency up to 90%, indicating significantly destroys of the membrane integrity (Figure S12). AGNCs treatment also led to a clear rise in ROS levels in both P. a and MDR P. a, with ROS production increasing sharply after irradiation (Figure S13). These results indicate that AGNCs kill Gram-negative bacteria by disrupting the membrane and generating ROS, which are further enhanced by the photothermal effect.

Fig. 2.

Antibacterial activities of AGNCs combined with the PTT and the crosslinking mechanism of A@MC hydrogels. (A) SEM images illustrating morphological changes in bacteria following treatment with AGNCs with or without NIR irradiation (0.5 W/cm², 10 min). Assessment of (B) P. a and (C) MDR P. a viability following AGNC treatment with or without NIR irradiation, using SYTO 9 (green, live cells) and propidium iodide (red, dead cells) staining to evaluate membrane permeability. NIR irradiation condition was set as 0.5 W/cm² for 10 min. (D) Schematic illustration showing the fabrication process and the crosslinking behavior of A@MC hydrogels. (E) Temperature-dependent sol to gel transition of MC and photothermal-triggered gelation of A@MC at various irradiation time points. (F) SEM images of MC and A@MC hydrogels. (G) FT-IR spectra of MC and A@MC hydrogels. (H) Cumulative release profile of AGNCs from A@MC hydrogel with or without 5 min of NIR irradiation over time

Characterizations of A@MC hydrogel

MC is a widely used pharmaceutical excipient that exhibits reverse thermal gelation in water. In this study, the A@MC hydrogels were formed through dynamic boronate ester bonds between the boric acid groups on AGNC and the hydroxyl groups of MC, endowing the material with injectability and self-healing properties. Upon irradiation, the heat generated by AGNCs disrupted the hydration layer of MC, exposing previously hidden hydrophobic methyl groups and driving the transition into a stable, crosslinked gel structure (Fig. 2D). MC remained in a sol state at low temperatures (30–40 °C) and transforms to a gel state when the temperature rose to a critical range, typically between 50 and 60 °C (Fig. 2E). AGNCs exhibited strong photothermal effects, and their incorporation into MC (A@MC) facilitated rapid gel formation when exposed to NIR irradiation. At 0.5 W/cm², A@MC showed a temperature increase of approximately 30 °C within 5 min, and was completely solidified (Fig. 2E). After the laser was turned off, the temperature of A@MC rapidly decreased and returned close to room temperature (Figure S14A-B). In contrast, the MC without AGNCs exhibited almost no temperature change under the same conditions (Figure S14B). We further evaluated the photothermal reversibility and stability of A@MC over multiple irradiation cycles, and the hydrogel consistently maintained its photothermal performance across six repeated heating and cooling cycles (Figure S14C).

We employed SEM to examine the microscopic structure and surface texture of freeze-dried MC and A@MC hydrogels. Both hydrogels exhibited three-dimensional lattice framework (Fig. 2F), closely resembling the extracellular matrix, which is advantageous for absorbing wound exudate and supporting cell growth. At higher magnification, the MC hydrogel showed a smooth internal surface, whereas A@MC presented a rougher inner texture, likely attributed to the presence of AGNCs (Fig. 2F). We analyzed the chemical composition of MC and A@MC using Fourier transform infrared spectroscopy (FT-IR). Both samples displayed similar peak positions (Fig. 2G), with characteristic peaks at 3409 cm^-1 and 2905 cm^-1 corresponding to the stretching vibrations of O-H and CH₃ groups. Peaks observed at 1700 and 900 cm^-1 represent the stretching vibrations of C = C and C-O-C bonds, along with C-H bending vibrations. After the incorporation of AGNCs, new peaks appeared at 1373 cm^-1 and 468 cm^-1 which was attributed to boronate ester and Au-N bonds respectively, confirming the successful crosslink of AGNCs with the MC hydrogel (Fig. 2G). The minimal changes in the remaining peaks suggest that AGNC incorporation does not significantly alter the overall chemical structure of the hydrogel matrix. To investigate the effects of different pH values and ionic strengths on gelation and self-healing behavior of the hydrogel, we recorded the gelation time occurred under various conditions. Within the pH range of 4–10, the gelation time of MC remained largely unchanged, whereas increasing ionic strength significantly shortened the gelation time from 10 min to 3.5 min (Figure S15). A@MC have the same trend under different pH and ionic strengths (Figure S15). A rheometer was further used to evaluate the oscillatory strain of the gels. At a frequency of 1 Hz and within a strain range of 0.1–1000%, the storage modulus (G′) of both MC and A@MC hydrogels remained higher than the loss modulus (G″), indicating typical viscoelastic characteristics and good elastic behavior (Figure S16). Although the storage modulus (G′) of MC was higher than that of A@MC, the strain at which the G′ and G″ curves intersected was greater for A@MC, indicating enhanced structural stability and improved resistance to deformation. The incorporation of AGNCs increased the crosslinking degree of the MC hydrogel, thereby improving its mechanical properties (Figure S16). In addition, the larger gap between G′ and G″ for A@MC reflects stronger elastic dominance and a more stable network structure, whereas the smaller gap observed for MC indicates a greater viscous component and a relatively looser structure. From the frequency sweep curves of G′ and G″ for the A@MC hydrogel at a strain of 1%, the trend showed G’’ > G’ at low frequencies and G’ > G’’ at high frequencies across the frequency range of 0–50 Hz, further confirming the good elastic properties of the A@MC hydrogel (Figure S17). The self-healing performance of A@MC hydrogels with different pH values was evaluated by cutting the gel at the center. Fracture-healing experiments showed that the bisected A@MC hydrogel could self-heal within 30 min without visible cracks and with good shape recovery (Figure S18). When stretched with tweezers, the healed hydrogel could be extended to around 3 times its original length without fracture at the interface, further confirming the excellent self-healing capability of the A@MC hydrogel. Also, the self-healing ability of the A@MC hydrogel was further evaluated using a rheological recovery test. A strain amplitude sweep was performed by alternating strains of 1% and 500%. At a high strain of 500%, G′′ rapidly decreased from 383 to 274 Pa. When the strain was reduced to 1%, both G′ and G″ recovered to their original values even after five repeated cycles, demonstrating good elastic behavior and reconstruction of the network (Figure S19). The adhesive strength of the hydrogels was evaluated using a shear lap test, and the stress increased from 4.61 to 6.06 kPa after incorporation of AGNCs (Figure S20). In the hydrogel swelling ratio test, both MC and A@MC hydrogels exhibited a rapid increase in swelling within the first 60 min and then maintained a stable swollen state thereafter (Figure S21). After different NIR irradiation times of 1, 2, and 5 min, A@MC reached distinct temperatures, resulting in different gelation states. Over 180 min, the swelling ratios of all hydrogels gradually increased and then reached a plateau (Figure S22). The prolonged NIR irradiation time enhanced the crosslinking of A@MC, thereby slowing water absorption and ultimately reducing the swelling ratio (Figure S22). The release behavior of AGNCs from A@MC hydrogels was also evaluated. In A@MC, AGNCs are present in excess. Thus, in addition to the particles tethered to MC fiber network through boronate ester bonds, more AGNCs are freely distributed in the hydrogel. After 5 min of NIR irradiation upon gelation, the cumulative release of AGNCs from A@MC reached approximately 16%, which was almost identical to that of the control group without NIR irradiation. Only a slight reduction in release was observed in the irradiated group relative to the non-irradiated group during the later stage of 60–240 min, and more than 80% of the incorporated AGNCs were released from A@MC after 360 min of incubation (Fig. 2H). The underlying cause of this observation is that photothermal curing of the A@MC hydrogel may partially restrict the release of the fiber-tethered AGNCs. However, hydrogel contraction simultaneously generates a driving force that pushes internal fluid outward, facilitating the release of freely distributed AGNCs within the pores as the fluid is expelled. This demonstrates that A@MC hydrogels provide an initial burst release followed by a sustained release of AGNCs, effectively maintaining antibacterial activity over time.

Due to the natural tension of skin, wound edges frequently undergo stress concentration, which disrupts the healing process and delays closure. When applied, the A@MC dressing adhered to proteins or collagen of the wound surface through hydrogen bonding and van der Waals interactions, and its temperature-responsive contraction reshaped the mechanical environment around the wound, promoting rapid wound healing (Fig. 3A). During the photothermal curing process, the A@MC hydrogel underwent rapid shrinkage, reaching a shrinkage rate of up to 68.90% after just 5 min of NIR irradiation (Fig. 3B, Table S3). In contrast, MC gels without gold nanomaterials remained in a sol state, as they failed to reach the temperature required for gelation and lack the anchors between short gel chains, resulting in shrinkage rates of less than 5% (Fig. 3B, Table S3). Additionally, the A@MC hydrogels demonstrated strong adhesion to various moist tissues, including the heart, liver, spleen, lungs, kidneys, and skin, indicating their suitability for adapting to the complex conditions of infected wounds (Fig. 3C). Since the matrix of the A@MC is composed of MC, which contains abundant hydroxyl groups, it readily forms hydrogen bonds with polyhydroxylated components in skin tissue, such as hyaluronic acid, protein or collagen. Upon light-induced heating from the AGNCs, partial water evaporation and MC curing promoted the A@MC to shrink and adhere more tightly to the skin. In contrast, MC alone adhered weakly and, without a thermal trigger for solidification, cannot establish stable adhesion (Fig. 3D). Remarkably, under the photothermal curing process, the A@MC hydrogel triggered rapid contraction of the entire wound, similar to its in vitro shrinkage behavior, reaching a shrinkage rate of 46.78% after just 5 min of light irradiation. In comparison, MC without gold nanomaterials failed to undergo gelation and exhibited minimal natural shrinkage associated with the skin, with a shrinkage rate of less than 10% (Fig. 3E, Table S4). We also evaluated the performance of A@MC hydrogels on wounds of different shapes, including round, triangular, square, and crescent-shaped wounds. Under NIR light irradiation, the AGNCs-generated photothermal effect effectively triggered MC gelation (Fig. 3F), enabling the hydrogel to conform precisely to wounds of various geometries. Additionally, wounds of different shapes exhibited varying degrees of contraction, with irregularly shaped wounds showing more pronounced shrinkage compared to the circular ones. Triangular wounds showed the most significant contraction. This process not only forms a protective barrier and promotes rapid wound closure but also initiates localized PTT to effectively eliminate bacteria. Thus, A@MC hydrogels performed as designed in gelation in situ.

Fig. 3.

Characterization of A@MC hydrogels. (A) Schematic illustration depicting how active wound contraction and hydrogel adhesion reshape the wound’s mechanical environment to accelerate closure. (B) Shrinkage comparison of MC and A@MC hydrogels before and after NIR irradiation. (C) Evaluation of A@MC hydrogel adhesion on various tissues. (D) Detachment behavior of MC and A@MC hydrogels from skin before and after NIR irradiation. (E) Comparison of MC and A@MC hydrogel shrinkage on mouse skin before and after NIR exposure. (F) Application of A@MC hydrogel to mouse wounds of various shapes, showing wound contraction before and after NIR treatment. NIR irradiation conditions: 808 nm, 0.5 W/cm², 5 min

Biosafety of AGNCs and A@MC hydrogel

Biocompatibility is a critical factor for the future biomedical application of nanomaterials. To evaluate the effect of AGNCs on cell proliferation, NIH-3T3 (mouse embryonic fibroblasts) and HMEC-1 (human dermal microvascular endothelial) cells were incubated with AGNCs for 24 h, with or without NIR irradiation, and assessed using the CCK-8 assay. Compared to cells without any treatment, AGNCs had minimal impact on cell viability without NIR irradiation (Fig. 4A-B). Under NIR irradiation, the viability of AGNC-treated cells slightly decreased but remained above 85%. Live/Dead staining assay showed high density, uniform morphology, and strong green fluorescence of the AGNC-treated (200 µg/mL) cells, which was consistent with the CCK-8 assay results (Fig. 4C-D). Both CCK-8 assay and Live/Dead staining assay indicate good biocompatibility of the AGNCs.

Fig. 4.

Biocompatibility assessment of AGNCs and A@MC hydrogels. Cell viability of (A) NIH-3T3 and (B) HMEC-1 cells after treatment with AGNCs with or without NIR irradiation, measured by OD_450nm using the CCK-8 assay. (C) Confocal images of NIH-3T3 and (D) HMEC-1 cells following AGNCs treatment with or without NIR irradiation (scale bar, 50 μm). (E) Thermal imaging of mice treated with A@MC hydrogels before and after NIR irradiation (808 nm, 0.5 W/cm²) for 5 min

We co-incubated NIH-3T3 cells with the solidified A@MC hydrogel and assessed their viability using the CCK-8 assay. While cell proliferation was slightly reduced in the early phase following NIR irradiation, it recovered to the levels comparable to those of non-irradiated cells at later stages (Figure S23). This indicates that the A@MC hydrogel combined with light irradiation does not induce significant cytotoxicity. We also evaluated the effect of AGNCs and A@MC hydrogel on red blood cells using a hemolysis test. Even at a high concentration of 200 µg/mL (40 times the MIC), AGNCs showed no signs of hemolysis (Figure S24). Similarly, both MC and A@MC hydrogels caused no significant hemolytic damage (Figure S25), indicating the excellent blood compatibility of A@MC hydrogel for potential applications. We also evaluated the photothermal curing capability of A@MC hydrogels on the wound. Upon 5 min of low-power NIR irradiation (808 nm, 0.5 W/cm²), the surface temperature of the A@MC hydrogel increased from 29.9 °C to 55.9 °C. We further quantified the proportion of the wound area within different temperature ranges. At 5 min, the region with a temperature exceeding 50 °C accounted for 0.88% of the total wound area and was confined to the central area of the hydrogel, while the area above 45 °C accounted for less than 8% (Table S5). The majority of the wound surface remained within the 32.5 to 37.5 °C range and surrounded the high-temperature area, preventing heat from spreading to surrounding tissues. After the laser was turned off, the temperature rapidly decreased to 35.8 °C within 1 min and returned to normal body temperature within approximately 3 min (Fig. 4E). These results indicate that the transient temperature elevation and localized high temperature can cause negligible damage to the surrounding tissues. Moreover, the A@MC hydrogels fully solidified on the skin, closely matching the contours of the wound, whereas the uncured MC remained in a sol state as it did not reach the required temperature for gelation (Figure S26).

Therapeutic efficacy of A@MC hydrogels in treating infected wounds

To evaluate the therapeutic effects of A@MC hydrogel combined with PTT for infected wounds, a full-thickness MDR P.a-infected skin wound model was established in mice. We evaluated the infection-curing effect of the A@MC hydrogel in mice, setting gauze and MC hydrogel loaded with carbapenems (C@MC) as the controls. Upon NIR irradiation (808 nm, 0.5 W/cm²), A@MC rapidly reached the curing temperature, solidifying and closely conforming to the wound edges (Fig. 5A, Figure S27, 0 day), whereas the C@MC hydrogel remained in an uncured state under the same treatment. So, the A@MC combined NIR irradiation acts as a protective barrier, shielding the wound from external contaminants and providing comprehensive protection. Despite the presence of exudate in the infected wound, A@MC could still rapidly solidify upon light activation, although the shrinkage was slightly reduced. This process led to an immediate wound area reduction of about 25%, whereas other groups showed no wound contraction (Fig. 5A, Figure S27, day 0). Throughout the treatment period, we monitored and recorded the body weight of experimental mice every day. Wound healing progress and residual bacterial load in the tissue were assessed on days 3, 7, 10, and 14 following the different treatments. As a result of successful infection, the mice exhibited signs of illness and reduced food intake, leading to an initial weight loss of approximately 10% within the first 3 days (Figure S28). In the later stages, body weight gradually increased and stabilized. In the MDR P. a-infected group, the weight gain was slightly slower due to the high pathogenicity and treatment resistance of the infection. In the P. a infection model, all treatment groups showed significant wound size reduction, with A@MC combined with NIR irradiation promoting faster healing than other treatments at every time point (Figure S29-S30). By day 7, the A@MC group achieved a 50% wound closure, comparable to the antibiotic-loaded hydrogel group (C@MC), and this rate further improved to 80% when A@MC was combined with NIR irradiation (Figure S29-S30). In the MDR P.a infection model, by day 7, MC hydrogels complexed with different drugs led to smaller wound areas and faster healing compared to traditional gauze. A@MC demonstrated superior performance over C@MC, which may be ascribed to the distinct effectiveness of AGNCs and carbapenems in treating drug-resistant bacteria (Figure S31). By day 14, A@MC combined with NIR irradiation achieved nearly complete wound healing, reaching up to 100% closure (Fig. 5B). Furthermore, we evaluated bacterial residues in infected tissues at different time points following various treatments. In both susceptive and drug-resistant bacteria-caused infections, the group treated with A@MC combined with NIR irradiation exhibited a marked reduction in bacterial load throughout the healing process (Fig. 5C, Figure S32-S33). Remarkably, by day 7, MDR P. a bacterial levels had significantly declined and the bacteria were completely eradicated by day 14 (Figure S34), consistent with the observed wound healing progress. Thus, A@MC demonstrates synergistic bactericidal ability to promote wound healing.

Fig. 5.

Therapeutic effects of A@MC hydrogel on MDR P.a-infected wounds in mice. (A) Representative photographs showing wound healing progression under various treatments on day 3, 7, 10, and 14 post-treatment (scale bar, 1 cm). (B) Wound closure ratios across different treatment groups over time. (C) Quantification of bacterial load in MDR P.a-infected tissues following different treatments (scale bar, 1 cm). (D) H&E staining and (E) Masson’s trichromatic staining of infected wound tissues under various treatments (scale bar, 50 μm) was performed for histological evaluation

To evaluate the healing outcomes following various treatments, we performed H&E staining to examine wound beds, granulation tissue, and epithelial formation, and used Masson’s trichrome staining to assess collagen deposition. On day 3, all groups showed the presence of inflammatory cells such as neutrophils and lymphocytes, confirming successful infection establishment (Fig. 5D, Figure S35). In the case of P. a infection, by day 7, the wounds treated with MC-based hydrogels loaded with different materials exhibited earlier epithelial layer formation compared to the gauze-treated group, suggesting that the matrix structure of the MC hydrogel enhances re-epithelialization more effectively than traditional dressings (Figure S35). By day 14, the A@MC hydrogel combined with PTT exhibited a wound healing effect comparable to that of antibiotic treatment (C@MC). Additionally, NIR irradiation promoted a greater presence of elongated fibroblasts, which support the development of mature fibrous granulation tissue (Figure S36). In the case of MDR P. a infection, the A@MC hydrogel, both with and without NIR irradiation, demonstrated superior wound healing compared to antibiotic treatment, while the gauze group exhibited worsening ulceration (Fig. 5D). By day 7, significant inflammation was observed in the gauze and C@MC groups, whereas the A@MC hydrogel notably reduced the inflammatory response due to its potent antibacterial activity. By day 10, collagen distribution in the A@MC-treated wounds became more uniform and extended into the epidermis, indicating effective tissue regeneration (Fig. 5E). After 14 days, the A@MC-treated wounds showed the most effective healing, characterized by substantial collagen deposition, epithelial regeneration, and the formation of skin appendages. Under NIR irradiation, the photothermal properties of the incorporated AGNCs enable the A@MC hydrogel to fully solidify, forming a protective barrier that shields the wound from external bacterial contamination. This photothermal effect also synergistically enhances the antibacterial efficacy to enhance the wound healing. All observations aligned with the wound healing progression and bacterial residue analysis. Unlike the passive coverage provided by traditional dressings, the MC matrix in the A@MC hydrogel contains abundant hydrogen bonding interactions that enhance adhesion between the hydrogel and the surrounding wound tissue. Under low power light irradiation (0.5 W/cm², 5 min), AGNCs within the gel generate photothermal heating, inducing shrinkage of the hydrogel network. This contraction exerts a physical force that actively pulls the wound edges, resulting in rapid wound closure. Meanwhile, AGNCs released at the wound site can disrupt bacterial cell membranes and intracellular proteins through the photothermal effect, promoting wound healing. In addition, we evaluated changes in inflammatory markers in serum during the treatment process. Procalcitonin (PCT) and serum amyloid A (SAA) levels increased markedly on day 3 after infection with different bacteria. By day 14, these inflammatory markers in the treated group returned to normal ranges, indicating a significant anti-inflammatory effect of A@MC (Table S6). We further evaluated the levels of inflammatory factors, including interleukin (IL)-6 and tumor necrosis factor (TNF)-α, in wound tissues using immunohistochemical staining. The expression of IL-6 and TNF-α in the A@MC group was lower than those in other groups, indicating the antimicrobial and anti-inflammatory effect of the A@MC (Figure S37). A@MC suppressed the overexpression of TNF-α and IL-6, helping to shift the wound microenvironment from a highly inflammatory, difficult to heal state to a low inflammatory, pro-repair condition. Through light activation, the system simultaneously enables physical closure through contraction and chemical debridement through antibacterial activity, thereby promoting rapid wound healing. Additionally, we evaluated the safety of the MC-based hydrogel during the treatment period. Despite some fluctuations, the biochemical indicators of the mice remained within the normal range (Table S7). Histological analysis of major organs after treatment showed intact tissue structures without noticeable damage (Figure S38), confirming favorable biocompatibility of the hydrogel and systemic safety.

We further compared the performance of A@MC and C@MC in treating P. a and MDR P. a-caused wound infections. For both infection models, A@MC with NIR irradiation promoted active wound contraction through MC gel solidification, achieving a healing rate of about 50% on day 3. By day 7, A@MC and C@MC showed similar therapeutic effects against P. a. However, in MDR P. a infection, A@MC demonstrated superior efficacy, improving the wound healing rate by approximately 21% compared with C@MC, highlighting its ability to overcome the limitations of antibiotics against drug-resistant bacteria. After 10 days of treatment, the wounds infected with MDR P. a showed more than 75% healing under A@MC treatment, even without NIR irradiation. However, the introduction of light irradiation was critical, as it reduced the bacterial count by ten-fold, fully activating the photothermal therapy effect of AGNCs and enabling more effective bacterial clearance. Histological analysis and biochemical indicators further confirmed that the photothermal effect of AGNCs promoted early wound closure through physical contraction, controlled infection through photothermal sterilization during the middle stage, and reduced inflammation while enhancing skin appendage remodeling in the late stage. These findings demonstrate that A@MC combines the rapid gelation and wound contraction properties of MC with the photothermal antibacterial function of AGNCs, providing a multi-stage synergistic therapeutic advantage.

Conclusions

In this study, we integrated AGNCs with methyl cellulose hydrogel (A@MC) and connected it to wounds of irregular shapes before gelation by hydrogen bonds and Van der Waals forces. Upon near-infrared (NIR) light irradiation, the photothermal effect of the AGNCs triggered a rapid connection of the dressing with the wound tissues and simultaneously the gelation of the dressing, causing an instant contraction of the dressing and subsequently a sharp shrinkage of the wound. A contraction rate up to 46% was achieved within 5 min in mouse wounds. The released AGNCs bound to bacteria through ABA-LPS interactions, enabling targeted antibacterial therapy. In vitro and in vivo studies using P. a and MDR P. a-infected models revealed that A@MC suppressed the overexpression of TNF-α and IL-6 and significantly accelerated wound healing with up to 100% closure within 14 days, which is much shorter than other wounds including uninfected wounds (usually > 21 days). Thus, the key novelty of this work lies in the thermosensitive contractility of A@MC hydrogels and reduced inflammatory responses. Therefore, A@MC represents a promising therapeutic approach for rapid healing of MDR bacteria-infected wounds.

Materials and methods

Materials

Tetrachloroauric acid (HAuCl₄·3H₂O) was sourced from Sinopharm Chemical Reagent Co., Ltd. Aminoboric acid hydrochloride (ABA, MW = 173.4) and glutathione (GSH, MW = 307.32) were purchased from Sigma (USA). Methyl cellulose (MC, 400 mPa•s) and sodium borohydride (NaBH₄) were obtained from Shanghai Aladdin Biotechnology Co., Ltd. Dulbecco’s Modified Eagle Medium (DMEM), fetal bovine serum (FBS), and penicillin-streptomycin (PS) were acquired from Gibco (USA). Phosphate-buffered saline (PBS) and lysogeny broth (LB) were provided by Beijing Solarbio Technology Co., Ltd. Calcein AM, propidium iodide (PI), hematoxylin-eosin (HE) staining solution, and Masson’s trichrome staining solution were obtained from Servicebio (Wuhan Xavier Biotechnology Co., Ltd.). The Live/Dead Bacterial Double Stain Kit (SYTO 9/PI) was sourced from Dalian Meilun Biotechnology Co., Ltd. The Cell Counting Kit-8 (CCK-8) was purchased from Beijing Lanjieke Technology Co., and 4% paraformaldehyde was supplied by Nanchang Benro Biotechnology Co., Ltd.

Synthesis of AGNCs

ABA-modified gold nanoclusters (AGNCs) were synthesized via a one-step aqueous-phase reduction of HAuCl₄·3H₂O using NaBH₄, following our previously reported methods [19]. Briefly, 2 mM ABA and 1.2 mM HAuCl₄·3H₂O were dissolved in 5 mL of ultrapure water, resulting in a color change from colorless to pale yellow. Then, 100 µL of freshly prepared NaBH₄ solution (4.3 mg/mL in 0.2 M NaOH) was added dropwise. The solution was stirred at 800 rpm at room temperature in the dark for 3 h. The reaction mixture turned from clear to dark brown. ABA modification introduced amino functional groups onto the surface of the GNCs. For control GNCs, the same procedure was followed adding glutathione. The AGNC solution was purified using a dialysis membrane (MW 3500 Da, Solarbio) for 24 h to remove unreacted components. Finally, the AGNCs were filtered through a 0.22 μm membrane (Millipore) for sterilization and stored at 4 °C for later use.

Characterization of GNCs and AGNCs

To evaluate the composition and structure of AGNCs, we conducted a series of comprehensive characterizations. Dynamic light scattering (DLS) and zeta potential measurements were performed using a Particle Size and Zeta Potential Analyzer (Brookhaven Instruments, NanoBrook Omni, America) to assess particle dispersity and surface charge. Optical properties, including absorbance and fluorescence, were measured using a UV-vis spectrophotometer (Shimadzu UV-2450, Japan) and an EnSpire multimode reader (PerkinElmer, USA). The morphology of AGNCs was observed via a transmission electron microscopy (TEM, FEI TECNAI G2 F20, USA). Elemental composition was analyzed using an X-ray photoelectron spectroscopy (XPS, JEOL, Japan). Additionally, the concentration of AGNCs was determined by inductively coupled plasma (ICP) analysis using an iCAP 6300 system (Thermo Fisher Scientific, USA).

To evaluate the photothermal properties of AGNCs, we assessed their temperature response under 808 nm near-infrared (NIR) laser irradiation. Various concentrations of AGNCs (200 µL; 50, 100, 200, and 400 µg/mL) were placed in Corning 96-well plates (Costar 3599) and exposed to an 808 nm NIR laser (0.5 W/cm²; Changchun Laser Photonics Technology Co., Ltd.) for 10 min. The samples were then allowed to cool at room temperature for 10 min. Temperature changes were recorded every 30 s over a total of 20 min using an infrared thermal camera (Testo 865, Germany). To examine photothermal stability, a cycling test was performed by irradiating 200 µL of AGNCs solution (100 µg/mL) through six repeated on/off cycles (10 min each) under the same NIR laser. Photothermal conversion efficiency (η) was calculated from the recorded temperature data following established methods reported in previous studies [20].

where m represents the mass of the AGNPs solution, C_water (4.2 J/g • ^oC) is the specific heat capacity of water. T_max and T_surr refer to the maximum and ambient temperatures, respectively, while τ_s is the time constant related to heat transfer. Q_Dis indicates the amount of heat lost to the solvent and container during laser exposure. I denotes the laser power, and A₈₀₈ is the absorbance of AGNPs solution at 808 nm.

Antibacterial effects of AGNCs in vitro

We selected various bacterial strains and their corresponding multidrug-resistant (MDR) isolates, all sourced from the American Type Culture Collection (ATCC). The bacteria were cultured in a shaker at 260 rpm and 37 °C for 4 h to reach the logarithmic growth phase. The minimum inhibitory concentrations (MICs) of AGNCs were determined using the broth microdilution method, following the Clinical and Laboratory Standards Institute (CLSI) guidelines. In brief, bacterial suspensions were adjusted to 10⁵ CFU/mL and added to Corning 96-well plates. AGNCs were serially diluted (2- to 128-fold) in 100 µL of LB medium. A broth-only group was included as a blank control. Equal volumes of the bacterial suspension and AGNCs solution were combined at a 1:1 ratio, then incubated at 37 °C for 24 h. The optical density at 600 nm (OD_600nm) was measured using a multimode microplate reader (TECAN Infinite 200 Pro, Switzerland). To further evaluate the antibacterial efficacy of AGNCs, we performed an inhibition zone and bactericidal assay. A bacterial suspension (100 µL, 10⁵ CFU/mL) was spread on agar plates. Sterile discs (approximately 1 cm in diameter) were placed on the surface of the agar, and 5 µL of LB, antibiotic solution, or AGNCs (with or without 808 nm NIR laser irradiation at 0.5 W/cm² for 5 min) was applied to each disc. The plates were then inverted and incubated at 37 °C for 24 h. The inhibition zones were observed and recorded using a gel imaging system (Universal Hood II, Bio-Rad Laboratories). The antibacterial effects of AGNCs, with or without 808 nm NIR laser irradiation, were evaluated at various time points (0, 2, 4, 8, 12, and 24 h). At each interval, bacterial suspensions were collected and spread onto agar plates. Following a 24-h incubation at 37 °C, the bacterial growth was assessed and documented. Crystal violet staining was used to assess antibiofilm activity. Bacterial suspensions were diluted to 1 × 10⁵ CFU/mL and cultured into 96-well plates. Different concentrations of AGNCs (100, 50, 25, 12.5, and 6.25 µg/mL) were added, with or without NIR irradiation (808 nm, 5 min). After 48 h of incubation, the medium in each well was removed, and the wells were gently rinsed with PBS. Subsequently, 100 µL of 0.3% crystal violet solution was added and incubated in the dark for 30 min to stain the biofilm. 200 µL of 95% ethanol was added, followed by shaking and incubation for 30 min for decolorization. The optical density at 590 nm was measured to quantify the biofilm content. The biofilm survival rate was calculated using the formula:

where OD_sample is the absorbance values of bacterial culture medium with nanomaterials treatment under different concentration, and OD_control is the absorbance values of untreated bacterial culture medium.

To explore the antibacterial mechanism, bacteria were treated with AGNCs (100 µg/mL) for 4 h, with or without NIR irradiation (808 nm, 0.5 W/cm², 10 min). After treatment, bacterial cells were washed three times with PBS and fixed in 2.5% glutaraldehyde at 4 °C for 4 h. The morphology of the bacteria was examined using a scanning electron microscopy (SEM, SU8200, Japan). To assess reactive oxygen species (ROS) production, bacteria in the logarithmic growth phase were treated with AGNCs (with or without 808 nm NIR irradiation) for 6 h at 37 °C. A PBS-treated group served as the control. The cells were then incubated with DCFH-DA (10 µM) for 60 min, and fluorescence intensity was measured (excitation: 488 nm; emission: 525 nm). For bacterial viability analysis, cells from different treatments were centrifuged at 8000 rpm for 3 min, stained with SYTO9 and PI for 30 min, and observed using a confocal microscope (Dragonfly, Andor).

Preparation and the characterization of A@MC

Considering that the gelation temperature of MC is approximately 50–60 °C, the concentration of AGNCs in A@MC must exceed 150 µg/mL to achieve temperatures above 60 °C. In addition, a sufficient amount of AGNCs contributes to enhanced antibacterial activity. Therefore, AGNCs (400 µg/mL) were mixed with MC (150 mg/mL) at an optimized volume ratio of 4:6, resulting in a final AGNC concentration of 160 µg/mL in A@MC. MC without AGNCs served as the control. To observe the gelation process, the MC gel was placed in a water bath at various temperatures (30, 40, 50, and 60 °C). To assess the effect of AGNCs on MC gel solidification, A@MC hydrogels were exposed to 808 nm NIR laser irradiation (0.5 W/cm²) for different durations (1, 3, 5, and 10 min). We measured the gelation time of MC and A@MC under different pH values and ionic strengths. Successful gelation was defined as the point at which the mixture no longer flowed. For physicochemical characterization, MC and A@MC hydrogels were freeze-dried using a lyophilizer (FD-1 A-50+, Bo Medical Kang Instrument Co.) for 12 h. Fourier transform infrared (FT-IR) spectroscopy (Bruker Tensor 27, Germany) was used to analyze the chemical composition and molecular structure of the hydrogels. SEM was employed to examine the internal microstructure of the hydrogels.

A rheometer (ARES-G2, TA Instruments, USA) was used to evaluate the elastic properties of MC and A@MC hydrogels. The temperature was maintained at 25 °C. The variations in storage modulus (G′) and loss modulus (G″) were measured as a function of strain (0.01%-1000%) at a constant frequency of 1 Hz, and as a function of frequency (0.01–50 Hz) at a constant strain of 1%. The self-healing ability of the hydrogels was evaluated according to previous studies [42–43]. Briefly, 1 g of prepared A@MC hydrogel in different pH value was cut in the middle, and the separated pieces were then brought together. The rejoining process was observed at different time points. Strain amplitude sweep tests were performed to determine the critical strain of A@MC hydrogels. In addition, G′ and G″ were recorded over five cycles by alternating oscillatory strains of 1% and 500% at a frequency of 1 Hz and a temperature of 25 °C. The adhesive strength of the MC and A@MC hydrogels was qualitatively evaluated by lap shear testing on pig skin based on a previous study [44]. The hydrogel was applied between two pig skin samples and the lap shear test was performed in tensile mode using an electronic universal testing machine (CMT2102, MTS), and the maximum tensile strength was recorded.

Water absorption (WA) was assessed by immersing the hydrogels in ultrapure water and weighing them at 1, 5, 10, 20, 30, 60, 120, and 180 min. The WA rate was calculated using the formula:

where W_wet is the weight after water absorption, and W_dry is the initial dry weight.

AGNCs exhibited a characteristic absorption peak at 556 nm (Figure S39). The release behavior of the A@MC hydrogel was evaluated by quantifying AGNCs in the supernatant by measuring optical density at 556 nm. A@MC samples were immersed in ultrapure water, and at predetermined time points (1, 5, 10, 20, 30, 60, 120, 180, 240, 300, and 360 min), 100 µL of solution was collected. The optical density of the supernatant at 556 nm was measured, and the cumulative release of AGNCs from A@MC was calculated using the following equation:

where Cₙ represents the concentration of AGNCs in the supernatant at each time point, and C₀ is the initial concentration of AGNCs. All experiments were performed in triplicate.

To evaluate the shrinkage induced by photothermal effects, A@MC and MC droplets were placed on a slide, and their initial areas were recorded. The contraction rate of the hydrogel was evaluated based on changes in area of the hydrogel size. Hydrogel samples of uniform size (S_initial) were prepared, and after NIR irradiation for 5 min, the final area (S_final) was measured. Each sample was measured three times, and the average value was used. The contraction rate of the hydrogel was calculated using the following equation:

The experimental mice (~ 20 g) were obtained from Spivey Biotechnology Co. To verify that A@MC contraction exhibits a similar effect in mouse skin, 0.5 cm wounds were created on the backs of mice and treated with either A@MC or MC. Experimental procedures similar to the in vitro tests were performed, including photographing and recording the wound size before and after irradiation with an 808 nm NIR laser, followed by calculating the contraction rate. To assess the adaptability of A@MC to various wound shapes, round, triangular, square, and crescent-shaped wounds were created on the backs of mice. Following 5 min of irradiation, the condition of the gel was observed and recorded, and the shrinkage rate was calculated. To assess the adhesive properties of A@MC, freshly harvested murine organs (liver, heart, spleen, lung, kidney) and skin were placed in contact with 1-cm A@MC hydrogel samples, followed by exposure to 808 nm NIR laser irradiation (0.5 W/cm²) for 5 min. After irradiation, the samples were lifted into a suspended position, and photographs were taken to visually evaluate the adhesive performance. To further investigate the effect of AGNCs on the photothermal-induced gelation of A@MC at wound sites, 1-cm-sized wounds were created on the backs of mice. Each wound was treated with 0.5 mL of either MC or A@MC, followed by irradiation with a 808 nm NIR laser (0.5 W/cm²) for 5 min and then cooled for another 5 min. Temperature changes were recorded at 1-min intervals, and the extent of gel curing was visually assessed and photographed after 10 min.

Biosafety of AGNCs and A@MC

The cytotoxicity of AGNCs was assessed using the Cell Counting Kit-8 (CCK-8) assay. Briefly, Mouse Embryonic Fibroblast Cells (NIH-3T3) and Human Microvascular Endothelial Cells (HMEC-1) (3 × 10³ cells/well) were seeded in 96-well plates and exposed to various treatments, including DMEM (control), AGNCs at different concentrations (12.5, 25, 50, 100, and 200 µg/mL), with or without 808 nm NIR laser irradiation (0.5 W/cm² for 5 min), followed by 24-h incubation. Then, 10 µL of CCK-8 solution was added to each well and incubated in the dark for 2 h. Cell viability was determined by measuring optical density at 450 nm (OD_450nm) using a microplate reader. The cytotoxicity of A@MC was evaluated using the same procedure. Additionally, Calcein AM/PI staining was used to observe morphological changes in NIH-3T3 and HMEC-1 cells after the same AGNC treatments, visualized under a confocal microscope. Additionally, to assess the biocompatibility of AGNCs and A@MC, a hemolysis assay was conducted. Fresh mouse blood was centrifuged at 1500 rpm for 10 min to separate the plasma, which was then discarded. The erythrocytes were washed three times with PBS to remove residual plasma proteins and impurities, then resuspended in PBS. The erythrocyte suspension (final concentration of 2%) was treated with AGNCs at varying concentrations (25, 50, 100, and 200 µg/mL), PBS (negative control), or deionized water (positive control) in a final volume of 1 mL. The samples were incubated at room temperature for 3 h, followed by centrifugation at 1500 rpm for 10 min. The supernatant was collected, and absorbance at 540 nm (OD_540nm) was measured using a microplate reader to evaluate hemolysis. The hemocompatibility of MC and A@MC was assessed using the same method. All experiments were performed in triplicate to ensure consistency and reliability.

Infection wound model and histological analysis

P. a is a common pathogen in skin injuries such as burns, chronic ulcers, and infection wounds, while MDR P. a highlights the growing challenge of antibiotic resistance. Animal models replicate key clinical features such as biofilm formation, inflammation, and delayed healing, providing a realistic platform to study anti-infection strategies. They also enable dynamic evaluation of wound healing by linking dressing contraction, drug release, inflammatory responses, and tissue regeneration. To evaluate the therapeutic efficacy of A@MC hydrogels in treating infected wounds, we established wound models infected with P.a and MDR P.a. All animal experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals and approved by the Animal Care and Use Committee (ACUC) of the National Center for Nanoscience and Technology (NCNST). We anesthetized the mice by intraperitoneal injection of pelltobarbitalum natricum (30 mg/kg). Mice were euthanized by CO₂ inhalation followed by cervical dislocation to ensure death. Four circular wounds (~ 1 cm in diameter) were created on the backs of mice, and each wound was infected with 100 µL of P.a or MDR P. a suspension (10⁶ CFU/mL) for 30 min. The hydrogel was applied to fully cover the wound without overflow, with an optimal weight of approximately 0.2 g. The amount of hydrogel was kept constant to ensure experimental consistency and reproducibility. Treatments included gauze, MC hydrogel loaded with carbapenems (C@MC), and A@MC with or without 808 nm NIR laser irradiation (5 min, 0.5 W/cm²). The antibiotic-loaded hydrogel (C@MC) was prepared using the same procedure as A@MC, with AGNCs replaced by carbapenems. Body weight was monitored daily using a precision scale (SF-100, Zhejiang Jinyun Weichuang Electronic Co., Ltd.). Wound healing was documented by photographing and measuring the wound areas on days 0, 3, 7, 10, and 14. At each time point, wound tissues were collected for further analysis. Each wound sample was split into two parts: one half was homogenized in LB medium, plated on LB agar, incubated for 24 h, and the bacterial colonies were imaged and counted using a gel imaging system. The other half was fixed in 4% paraformaldehyde for 24 h, followed by dehydration, embedding, paraffin sectioning, and frozen sectioning for histological analysis. Hematoxylin and eosin (H&E) and Masson’s trichrome staining were performed to evaluate tissue regeneration using an upright optical microscope (Nikon DS-U3, Nikon, Japan), and digital scans were acquired with a tissue scanner (Pannoramic 250 FLASH, 3DHISTECH). Sections were immersed in xylene for 15 min and then in ethanol for 5 min. Antigen retrieval was performed by heating the sections in citric acid buffer (pH 6.0) for 15 min, followed by washing in PBS (pH 7.4) on a shaker for 15 min. The sections were then incubated with 3% hydrogen peroxide at room temperature in the dark for 25 min. After blocking with 3% BSA for 30 min at room temperature, the sections were incubated with primary antibody overnight at 4 °C and subsequently with secondary antibody for 50 min at room temperature. The sections were developed using freshly prepared DAB solution and counterstained with hematoxylin for 3 min. The stained sections were observed under a microscope and images were analyzed. Nuclei stained with hematoxylin appeared blue, while positive DAB expression appeared brownish yellow. To assess systemic toxicity during treatment, major organs (liver, heart, spleen, lungs, kidneys, and brain) were collected and processed for H&E staining. Additionally, blood samples were centrifuged at 4000 rpm for 10 min, and serum was isolated for biochemical analysis following the manufacturer’s instructions.

Statistical analysis

All experiments were performed with at least three independent replicates. Data are presented as mean ± standard deviation (mean ± SD). Statistical significance between two groups was evaluated using a two sample t-test, with P-values less than 0.05 considered statistically significant. All analyses were conducted using GraphPad Prism 9. * represents P < 0.05, ** represents P < 0.01, *** represents P < 0.001, **** represents P < 0.0001, and ns: no significance.

Supplementary Information

Supplementary material 1. ^{(10.8MB, docx)}

Author contributions

Shanchen Cai: Writing-original draft, Visualization, Methodology, Formal analysis. Wenfu Zheng: Conceptualization, Formal analysis, Writing-review & editing. Ting Lin: Software, Data curation. Siyuan Zhang: Formal analysis, Data curation. Lanye He: Formal analysis. Mali Zu: Writing-review & editing. Le Wang: Conceptualization, Project administration, Supervision, Writing-review & editing, Funding acquisition. Shanchen Cai and Wenfu Zheng with co-first authorship based on equal contributions.

Funding

This work was financially supported by the National Natural Science Foundation of China (22304065, U23A20686), Natural Science Foundation of Jiangxi Province of China (20242BAB20088), the Project of Academic and Technical Leaders in Major Disciplines in Jiangxi Province (20243BCE51158), and Jiangxi University of Chinese Medicine School-level Science and Technology Innovation Team Development Program (CXTD22005).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Consent for publication

All authors of this study agreed to publish.

Competing interests

The authors declare no competing interests.