The Met@PPS was synthesized following previous methods (Liu et al., 2020). Hydrophilic-hydrophobic interactions and supramolecular polymerization between porphyrin molecules were started by gradually adding water to the polymer’s solvent solution. Typically, 10 mg of polymer and 0.1/0.2/0.5 mg metformin were dissolved in 5 mL of DMF, and, after sonication for 10 min, filtered through a nylon filter with a pore size of 22 µg. Then, 10 mL of deionized water was slowly added, dropwise, to the above solution. This was followed by transfer to a dialysis bag with a molecular weight cut-off of 3,500 Da and dialysis in water for 48 h to remove DMF to obtain an aqueous solution of polymer vesicles. The volume of the assembly solution was used to calculate the vesicle concentration. Met@PPS aqueous solution (40 μg mL⁻¹) was dispersed by ultrasonication for 10 min; then, HAuCl₄ (10 mM, equivalent to 15 wt% Au) was added to the solution and stirred for 5 min in the dark, followed by reaction at 25 °C under visible light (λ > 420 nm, 300 W mercury lamp) for 1 h. The color of the solution deepened as the gold ions were reduced. After photoreaction, excess ions were removed by centrifugation and washing with deionized water for three times to obtain Janus-shaped Met@J-AuPPS.

The sample preparation process was carried out as follows: the assembly solution (0.1 mg mL^-1) was dropped on a clean silicon wafer, and, after air-drying, gold ions were sprayed on the surface of the sample. Then, SEM images were acquired with a Nova NanoSEM 450 field emission scanning electron microscope (FEI Corporation, United States).

The sample preparation process was as follows: the assembly solution (0.1 mg mL^-1) was dropped on a 200-mesh carbon film, and then air-dried for TEM photography on a Tecnai G2 biotype transmission electron microscope (Thermofisher, United States).

The UV-Vis absorption spectra of polymers, assemblies, and composite structures were tested on a Lambda 35 UV-Vis absorption spectrometer (PerkinElmer, United States). The sample solution was dispersed in a 1 × 1 cm quartz cell and the test temperature was 298 K.

Five to 10 mg of sample was dissolved in 0.6 mL of deuterated reagent (DMSO-d ₆ ), and then tested with a AVANCE III HD 400 nuclear magnetic resonance spectrometer (Bruker, United States). DMSO or TMS was used as the internal standard, and the test temperature was 298 K.

Gel permeation chromatography of ZnTHPG was performed on a HLC-8320GPC gel permeation chromatograph (Tosoh Corporation, Japan). A solution with a polymer concentration of 10 mg mL^-1 was prepared. After dissolving and standing overnight, it was filtrated through a PTFE filter with a pore size of 0.22 µg. The test temperature was 298 K, DMF was the eluent, and the reference molecule was polystyrene.

The particle sizes of the self-assembled polymer vesicles and Met@AuPPS were characterized by DLS on a Nanoparticle Size and Zeta Potential Analyzer ZS90 (Malvern Instruments Ltd, UK). The measuring angle was 90° and test temperature was 298 K.

The thermal properties of PPS, J-AuPPS, and Met@J-AuPPS, as well as the thermal imaging effects of the samples in vivo, were captured by an infrared thermal imager (FOTRIC220 s, China).

The metformin (Met) ELISA Assay Kit (Zhenke, Shanghai) was used to determine the metformin content released after NIR irradiation of the Met@J-AuPPS solution. The sample wells were supplemented with 50 µL of Met@J-AuPPS solution irradiated by laser for different durations (1, 2, 3, 4, and 5 min), followed by the addition of working solution. The plate was covered with film and incubated at 37 °C for 15 min prior to measurement of OD values at a wavelength of 450 nm using the ELX‐800 absorbance microplate reader. The concentration of metformin in the sample was calculated based on the standard curve obtained from the standard products.

HUVECs (The cell lines present in this study were obtained from National Collection of Authenticated Cell Cultures) were used to determine the cytotoxicity of Met@J-AuPPS using the CCK-8 kit (Dojindo, Japan). Briefly, cells, at a concentration of 1 × 10⁴ cells/well, were cultured with Met@J-AuPPS (400 μg mL⁻¹) in a 96-well-plate for 24, 48, and 72 h. PBS was set as a control group. After washing with fresh PBS twice, the cells were stained with 10% CCK-8 solution for 4 h. The relative cell density was determined by a microplate reader at a wavelength of 450 nm.

To investigate cellular damage by Met@J-AuPPS under NIR irradiation, 1 × 10⁴ cells/well HUVECs were co-incubated with 400 μg mL^-1 Met@J-AuPPS, and then NIR-irradiated (1,000 mW cm^-2) for 5 min. Next, the cells were rinsed with PBS twice and re-cultured with fresh culture medium. On day 1, day 2, and day 3, the cell viability was determined by CCK-8 assay, as described above.

For this assay, 1 × 10⁶ HUVECs per well were seeded in 6‐well plates. When the confluence of cells reached 80%, the 200 µL pipette tip was used to scrape the cells in a line to create wounds. Met, J-AuPPS, and Met@J-AuPPS were added to each well and irradiated with NIR (1,000 mW cm^-2) for 5 min. Images were captured by inverted microscopy at 0 and 48 h after wounding. The migration area was quantified using ImageJ software.

For the transwell assay, 2 × 10⁴ HUVECs per well were seeded in the upper chamber of 24‐mm Transwell chambers (Corning #3412, United States) with 200 µL serum‐free DMEM medium. Met, J-AuPPS, and Met@J-AuPPS were added to each well and irradiated with NIR (1,000 mW cm^-2) for 5 min. The lower chamber was filled with DMEM medium containing 10% FBS. After incubation for 24 h, cells in the upper chamber were fixed with 4% paraformaldehyde solution and stained with 0.1% crystal violet for 20 min. Subsequently, the chambers were air‐dried and images were captured. Migrated cells were quantified using ImageJ software.

The effect of Met@J-AuPPS on the tube-forming ability of HUVECs was investigated. The matrigel was thawed overnight at 4 °C, and sterile tips and angiogenesis slides (Ibidi, Germany) were precooled. The next day, the matrigel was quickly added to the angiogenesis slides and placed in a cell incubator for 60 min at 37 °C. Then, 8 × 10³ HUVECs per well were seeded in angiogenesis slides. Next, Met, J-AuPPS, and Met@J-AuPPS were added to each well and irradiated with NIR (1,000 mW cm^-2) for 5 min. Angiogenesis slides were observed at 4, 6, 8, 10, and 12 h, and photographed using an optical microscope. Tubes were quantified using ImageJ software.

Methicillin-resistant S. aureus (MRSA, ATCC43300) and extended-spectrum β-lactamase-producing E. coli (ESBL E. coli, ATCC35218) were selected as the two types of model bacteria. Single colonies were transplanted into tryptic soy broth (TSB) and grown at 37 °C overnight. The harvested cultures were centrifuged and adjusted to 10⁸ CFU mL^-1. After that, 10 μL bacterial suspension was mixed with 1 mL PBS (as control), J-AuPPS (400 μg mL^-1), or Met@J-AuPPS (400 μg mL^-1) in centrifuge tubes with or without 808 nm NIR irradiation (1,000 mW cm⁻²) for 10 min. Then, 10 μL of serially diluted bacterial solution was spread on a sheep blood agar plate and incubated overnight at 37 °C. Next, the bacterial colonies were counted.

As-prepared bacterial solution was added into a 24-well plate and cultured for 48 h. The plate was washed with PBS, and J-AuPPS or Met@J-AuPPS were added to each well and irradiated with NIR (1,000 mW cm^-2) for 10 min. After washing again, the biofilm was fixed with 4% paraformaldehyde solution for 10 min and stained with 0.1% crystal violet for 20 min. Then, 3% glacial acetic acid solution was used for decolorization of crystal violet in the biofilm. The OD value of each well was determined at 595 nm wavelength.

A sterile pure titanium disk was immersed into as-prepared bacterial solution and co-incubated for 48 h at 37 °C. Then, titanium was gently rinsed with aseptic PBS twice and re-immersed into PBS (control), J-AuPPS and Met@J-AuPPS under NIR laser irradiation (1,000 mW cm^-2) for 10 min. Next, the titanium disk was rinsed with PBS twice. A Bacterial Live/Dead Staining Kit was used to visualize the ratio of live cells/dead cells in the biofilm. Briefly, as-prepared titanium disks were immersed in a 1 mL mixture of STYO9 and PI (1: 1,000 v/v in saline) in a 24-well-plate. After co-culturing for 30 min, the live/dead cells encapsulated in biofilm were observed with CLSM.

The as-treated titanium disks were fixed with 2.5% glutaraldehyde solution at 4 °C overnight, followed by sequential dehydration in a serial dilution of ethanol (30, 50, 70, 85, 90, 95% and 100%, v/v) for 10 min, according to standard procedures. Next, titanium disks were freeze-dried for 24 h prior to scanning electron microscopy.

All animal experimental procedures were performed in accordance with guidelines, and approved by the Ethics Committee of Shanghai Jiaotong University School of Medicine Affiliated Shanghai Sixth People’s Hospital (No.DWLL205-0010). A diabetic infected wound model was constructed using SD rats (∼250 g), and all rats were anesthetized by an intraperitoneal injection of pentobarbital sodium at a dose of 40 mg/kg. A diabetic rat model was established with Streptozocin (STZ) by intraperitoneal injection. A circular wound (1 cm) was created on the dorsal area of each rat after 2-week feeding. The wound was inoculated with MRSA bacteria (1 × 10⁷ CFU/mL), and a biofilm was formed after 2 days to construct a wound infection model in diabetic rats.

The rats were then randomly divided into eight groups: NIR (+) groups: PBS, Met, J-AuPPS, and Met@J-AuPPS groups; NIR (−) groups: PBS, Met, J-AuPPS, and Met@J-AuPPS groups. Next, 50 μL of samples were added on the wound with or without NIR laser irradiation (1,000 mW cm⁻²) for 5 min, and local temperature changes were recorded. The skin wounds of rats were observed and photographed on days 3, 6, 9, and 12. On day 6 and day 12, rats were euthanized for further analysis. Skin tissues as well as major organs, including heart, liver, spleen, lung, and kidney, were collected for H&E staining, Masson staining, Giemsa staining, and immunofluorescence staining.

Prism nine software (GraphPad, US) and Origin 9 (OriginLab, US) were used for data analyses. The data were presented as the mean ± SD. T-test or one-way ANOVA was used for statistical analysis. A value of P < 0.05 was considered to indicate statistical significance, indicated as *P < 0.05, **P < 0.01, and ***P < 0.001.

First, ZnTHPG was synthesized according to a previous study (Liu et al., 2020). The ¹H NMR spectrum of ZnTHPG is shown in Supplementary Figure S1. Subsequently, the assembly procedure was then started by gradually adding water to the ZnTHPG/N,N-dimethylformamide (DMF) solution of the polymer in order to begin supramolecular polymerization and hydrophilic-hydrophobic interactions between porphyrin molecules. During the synthesis of PPS, we added metformin to the polymer solution to prepare metformin-loaded PPS (Met@PPS). The proportions of metformin were 1%, 2%, and 5%, respectively.

Transmission electron microscopy (TEM), scanning electron microscopy (SEM), and dynamic light scattering (DLS) were used to analyze the morphology and structure of the as-prepared PPS and Met@PPS(Figure 1A,B; Supplementary Figure S2). As shown in Figure 1, PPS, 1%Met@PPS, 2%Met@PPS, and 5%Met@PPS all formed smooth vesicle structures with uniform size. Some vesicles aggregated, which may have been due to the hydrophilic structure of the PPS vesicle surface. The DLS results showed that the vesicle size was similar between the four groups, indicating that the concentration of metformin had no significant effect on the uniformity and distribution of vesicle size. These results suggest that metformin does not interfere with the self-assembly of ZnTHPG. Therefore, we speculate that the properties of ZnTHPG loaded with metformin should be the same as those of hollow PPS, which provides the foundation for the synthesis and application of subsequent materials.

The J-AuPPS and Met@J-AuPPS heterostructures were then obtained by introducing HAuCl₄ as an Au source into the PPS aqueous solution at 25 °C while being exposed to visible light. The TEM imaging validated the Janus nanostructure (Figure 1C). We counted more than 200 nanoparticles after the photoreaction. As can be seen from the statistical results, only one Au nanoparticle grew on most of the vesicles (Supplementary Figure S3). To verify whether the metformin was successfully loaded, we characterized materials by UV-Vis absorption spectroscopy (Figure 1D). The absorption peak of Met@J-AuPPS was significantly higher than that of J-AuPPS at a wavelength of 400–550 nm, which proved that the three groups of Met@J-AuPPS were well loaded with metformin, indicating the feasibility and reliability of it.

Compared to other nanostructures, we have shown that J-AuPPS exhibits improved photothermal conversion efficiency (48.4%) and near-field enhancement effect (NFE) in the NIR spectrum (Chen et al., 2023b). According to the infrared thermal images, when exposed to an 808 nm laser, Met@J-AuPPS showed a comparable temperature increase from 29.5 °C to 59.8 °C to J-AuPPS (27.4 °C–63.1 °C) (Figure 1E,F). Additionally, after six cycles of heating and cooling, J-AuPPS and Met@J-AuPPS retained good photothermal characteristics (Figure 1G). These results confirm that, as expected, the loading of metformin does not adversely affect its original photothermal properties and photothermal conversion efficiency, and indicate that Met@J-AuPPS has satisfactory photothermal properties. Next, we tested the metformin release capacity of 1% Met@J-AuPPS, 2% Met@J-AuPPS, and 5% Met@J-AuPPS (Figure 1H). Met@J-AuPPS showed higher metformin release rate within 1 min, and almost all metformin was released within 5 min.

Good biocompatibility is a prerequisite for the application of Met@J-AuPPS in diabetic wound healing. The cytotoxicity of J-AuPPS to cells was comparatively low when the feeding concentration was less than 400 μg mL⁻¹, in our previous research (Chen et al., 2023b). J-AuPPS and 1%/2%/5% Met@J-AuPPS (400 μg mL⁻¹) were respectively co-cultured with HUVECs. CCK-8 assay revealed no significant difference in cell activity between the groups after 24, 48, and 72 h (Supplementary Figure S4A). Moreover, NIR-activated J-AuPPS and Initially, Met@J-AuPPS caused enormous cell death, although this effect was partly lost after being cultured for 2 and 3 days (Supplementary Figure S4B). These findings suggest that J-AuPPS and Met@J-AuPPS can cause damage to cells due to photothermal effects induced by NIR laser irradiation, but the damage is controlled and transient. More cells still maintain normal proliferative activity to eliminate this damage. These results demonstrate that Met@J-AuPPS has further utilization prospects in biological applications. Based on the above results, 5%Met@J-AuPPS(400 μg mL⁻¹) was selected for follow-up studies.

The effect of Met@J-AuPPS on HUVECs migration was evaluated by cell scratch test (Figure 2A). Compared with PBS and J-AuPPS, the number of migrated HUVECs after Met@J-AuPPS (NIR-triggered) treatment was significantly increased, showing similar effects as metformin treatment. The quantitative results of cell migration showed that the cell migration ratio of Met@J-AuPPS was significantly better than that of J-AuPPS (P < 0.01) (Figure 2D). The effect of Met@J-AuPPS was confirmed again by transwell migration test (Figures 2B,E). The ability of Met@J-AuPPS to induce tube formation by HUVECs was further tested, and the number of meshes and nodes was counted. As shown in Figures 2C F,G and only a small number of tubules were formed in the PBS group. Met@J-AuPPS induced the formation of more distinct and interconnected tubule structures by HUVECs. In summary, Met@J-AuPPS has ideal biocompatibility and good ability to promote angiogenesis in vitro.

Next, J-AuPPS and Met@J-AuPPS were evaluated for their antibacterial efficacy against MRSA and ESBL E. coli. The spread-plate method was used to assess the impact of J-AuPPS and Met@J-AuPPS on bacterial viability. As shown in Figures 3A–C, the NIR-activated J-AuPPS group nearly entirely scavenged MRSA or ESBL E. coli (bacterial viability <1%), indicating that PTT has a desirable antibacterial impact against microorganisms that are resistant to many drugs. Meanwhile, Met@J-AuPPS + NIR group showed similar results for antibacterial efficiency. The results reveal that the present nanomaterials kill both gram-negative and -positive bacteria by generating high local temperature through PTT, and metformin has no adverse effect on the PTT antibacterial properties.

Adhesion and colonization of bacteria further lead to the formation of biofilms. Antibiotic therapy and the host immune response are both highly tolerated by bacteria in biofilms (Yelin and Kishony, 2018). Therefore, the anti-biofilm properties of Met@J-AuPPS and J-AuPPS were assessed. Subsequently, following a 48-h biofilm-forming culture, MRSA and ESBL E. coli were exposed to NIR laser irradiation while being treated with J-AuPPS and Met@J-AuPPS. Using crystal violet staining, the biomass was examined to assess each sample’s anti-biofilm efficacy (Figures 3D–F). The biofilm elimination efficiency of the J-AuPPS and Met@J-AuPPS groups with NIR irradiation was 45.8% and 56.4%, respectively, according to the normalization of staining findings. Moreover, SEM was used to visualize the morphological changes of biofilms on the titanium (Ti) surface. The NIR-activated J-AuPPS and Met@J-AuPPS groups only had sporadic bacteria, but the control group had a dense bacterial cluster (Figure 3G). In addition, the bacterial live/dead (green/red) staining assay was used to assess the therapeutic efficacy of J-AuPPS and Met@J-AuPPS. The mature biofilms were thick and undamaged in control groups, which were stained green (Figure 3H). The NIR-activated J-AuPPS and Met@J-AuPPS groups, on the other hand, showed greater red fluorescence and a sparser biofilm, indicating that the biofilm structure was disrupted down and that the sealed bacteria were significantly reduced. These findings showed that the anti-biofilm properties of Met@J-AuPPS and NIR-activated J-AuPPS are comparable in vitro.

Based on the results that Met@J-AuPPS demonstrated ideal in vitro biocompatibility and excellent in vitro antibacterial ability, the ability of Met@J-AuPPS to accelerate infected wound healing was evaluated by establishing a MRSA-infected full-thickness skin defect diabetic rat model. The progression is illustrated in Figure 4A. The treatments were divided into eight groups: PBS, Met, J-AuPPS, Met@J-AuPPS, PBS + NIR, Met + NIR, J-AuPPS + NIR, Met@J-AuPPS + NIR(Figures 4B,C). Wound photographs were taken on the 3rd, 6th, 9th, and 12th day to assess the degree of wound healing. The Met@J-AuPPS + NIR group showed the best healing status without visible scarring (Figure 4D). The relative wound area with time (Figure 4E) showed that the healing rates of wounds in the Met, Met + NIR, J-AuPPS + NIR, and Met@J-AuPPS + NIR groups were much faster than those in other groups. Notably, the relative wound area of the Met@J-AuPPS + NIR group was the smallest among all the groups at day 12.

These results were confirmed by H&E and Masson staining, which demonstrated the excellent performance of Met@J-AuPPS in accelerating diabetic wound healing (Figures 5A,B). More skin defects and inflammatory cells were found in the PBS, J-AuPPS, Met@J-AuPPS, and PBS + NIR groups. In comparison, more complete skin structure, fewer inflammatory cells, and more collagen fibers were found in the Met, Met + NIR, J-AuPPS + NIR, and Met@J-AuPPS + NIR groups, especially in the Met@J-AuPPS + NIR group. After the sixth day of wound healing, the bacterial residue was assessed by Giemsa staining (Figure 5C). Numerous bacterial aggregates were seen in the -NIR, PBS + NIR, and Met + NIR groups, suggesting a serious infection. The antibacterial properties of J-AuPPS and Met@J-AuPPS under NIR irradiation contributed to the nearly total elimination of the bacteria in the J-AuPPS + NIR and Met@J-AuPPS + NIR groups.

We then performed immunostaining with CD31/α-SMA/VEGF, which were angiogenesis markers. It was observed that the relative coverage area of CD31/α-SMA/VEGF was higher in the wound areas of the Met@J-AuPPS + NIR groups (Figure 5D), which was attributed to the angiogenic ability of metformin and the non-infectious environment. In contrast, other groups showed a lower relative coverage area, indicating an impaired angiogenic state with increased inflammation. Finally, the vital organs were harvested and stained with H&E for biocompatible assessment (Figure 5E).