Ultrasmall MnOx Nanodots Catalyze Glucose for Reactive Oxygen Species‐Dependent Sequential Anti‐Infection and Regeneration Therapy

The management of diabetic wounds poses significant challenges due to persistent bacterial infections and chronic inflammation caused by hyperglycemia. Herein, a sequential two‐phase treatment strategy involving a reactive oxygen species (ROS) burst in the first phase for anti‐infection is proposed, followed by a benign level of ROS in the second phase for wound regeneration. To this end, ultra‐small manganese oxide nanodots (BM‐NDs) are incorporated into a gelatin methacrylamide (GelMA) hydrogel via a ROS‐responsive linker to form GelMA@BM dressing. The BM‐NDs catalyze a self‐cascade reaction that decomposes glucose into hydrogen peroxide, generates hydroxyl radicals (·OH), and simultaneously depletes glutathione. Upon application on diabetic wounds, BM‐NDs are rapidly released from the hydrogel due to endogenous ROS exposure, leading to high levels of ·OH that effectively eliminate bacteria and promote macrophage polarization to M1 phenotype, thereby facilitating phagocytosis of bacteria. With the consumption of glucose and degradation of BM‐NDs, ROS in the wound area declines to a benign level, which stimulates polarization of M2 macrophages and promotes wound healing. This two‐phase treatment strategy based on GelMA@BM dressing demonstrates potent antibacterial and pro‐healing efficacy, showcasing its potential for clinical translation.


Introduction
The prevalence of diabetes is expected to rise to 439 million individuals by 2030, [1] with 15% of these patients developing diabetic foot ulcers over their lifetime. [2]ortality and disability rates associated with diabetes-related chronic wounds surpass those of most cancers, with an annual mortality rate of 11% for diabetic foot ulcers and five-year mortality rate of 22%. [3]Consequently, addressing diabetesrelated chronic wounds is of significant socioeconomic and clinical importance.Among the various environmental and intrinsic factors associated with diabetic chronic wounds, the persistent presence of microbial load due to antibiotic resistance [4] and microbiome variation [5] is an immediate critical factor.Furthermore, the ongoing inflammatory state resulting from primary hyperglycemia inhibits the normal healing process of diabetic wounds. [6]Specifically, elevated glucose levels can exacerbate oxidative stress in the wound area by increasing the generation of reactive oxygen species (ROS) from endogenous cells. [6]Therefore, effective The management of diabetic wounds poses significant challenges due to persistent bacterial infections and chronic inflammation caused by hyperglycemia.Herein, a sequential two-phase treatment strategy involving a reactive oxygen species (ROS) burst in the first phase for anti-infection is proposed, followed by a benign level of ROS in the second phase for wound regeneration.To this end, ultra-small manganese oxide nanodots (BM-NDs) are incorporated into a gelatin methacrylamide (GelMA) hydrogel via a ROS-responsive linker to form GelMA@BM dressing.The BM-NDs catalyze a self-cascade reaction that decomposes glucose into hydrogen peroxide, generates hydroxyl radicals (•OH), and simultaneously depletes glutathione.Upon application on diabetic wounds, BM-NDs are rapidly released from the hydrogel due to endogenous ROS exposure, leading to high levels of •OH that effectively eliminate bacteria and promote macrophage polarization to M1 phenotype, thereby facilitating phagocytosis of bacteria.With the consumption of glucose and degradation of BM-NDs, ROS in the wound area declines to a benign level, which stimulates polarization of M2 macrophages and promotes wound healing.This two-phase treatment strategy based on GelMA@BM dressing demonstrates potent antibacterial and pro-healing efficacy, showcasing its potential for clinical translation.management of bacterial infections and dynamic manipulation of the intricate microenvironment in chronic wounds are two essential elements for achieving favorable outcomes in diabetic wound healing.
Macrophages play a critical role in the immune response and can exhibit distinct functional phenotypes depending on the microenvironment. [7]M1 macrophages are proinflammatory and predominate during the first three days of wound healing, while M2 macrophages are considered prohealing and dominate thereafter. [8]However, in a hyperglycemic environment, the transition from M1 to M2 macrophage polarization is limited, contributing to the chronic nature of diabetic wounds. [9]ROS can influence macrophage polarization in two different patterns, with moderate to high levels of ROS promoting M1 polarization, while relatively low levels of ROS inducing M2 macrophage polarization. [10]Previous studies have employed high levels of ROS for bactericidal purposes [11] or emphasized ROS removal to alleviate inflammation [12] in the treatment of diabetic infectious wounds.However, the importance of sequential dominance of M1 and M2 macrophages in physiological repair processes is often overlooked.Therefore, we propose that the management of diabetic infectious wounds should follow a two-step sequential process: 1) immediate local glucose consumption and ROS generation to activate M1 macrophages and eliminate infection and 2) establishment of a proregenerative microenvironment dominated by M2 macrophage polarization.Nevertheless, achieving such a dynamic therapeutic strategy with a simple method remains challenging.Therefore, advanced therapeutic strategies that can precisely control ROS levels and regulate M1/M2 macrophage polarization are needed to effectively manage diabetic infectious wounds.
15][16] Specifically, peroxidase (POD)-mimicking nanozymes can catalyze hydrogen peroxide (H 2 O 2 ) to generate hydroxyl radical (•OH), thereby increasing the local ROS level, while superoxide dismutase-mimicking nanozymes can scavenge superoxide radicals into less harmful molecules, thereby reducing the local ROS level.Among various nanozymes, manganese-based nanozymes have garnered significant attention due to their multi-enzymatic property, pH-responsive nature, and biodegradability. [17]As a transitional metal, manganese oxides (MnO x ) nanomaterials can initiate a Fenton-like reaction under acidic conditions to generate •OH, which has been widely explored for tumor therapy. [18]n our recent work, we have developed Mn-based nanohybrids with "four-in-one" enzyme-mimicking activities that are capable of self-supplying H 2 O 2 and generating •OH and oxygen, while simultaneously depleting intracellular glutathione (GSH), ultimately inducing ferroptosis of tumor cells. [19]Notably, the ultrasmall MnO x nanodots exhibit glucose oxidase (GOx)-mimicking properties, enabling the efficient consumption of glucose. [19]This unique property of the MnO x nanodots makes them particularly well-suited for use in diabetic wound microenvironments, where they could potentially function as a glucose-driven therapeutic agent.
Therefore, in this study, we propose a novel two-phase therapeutic strategy for the management of diabetic infectious wounds through dynamic regulation of ROS levels in the wound microenvironment using ultrasmall bovine serum albumin (BSA)-stabilized MnO x nanodots (BM-NDs).To achieve this, BM-NDs were incorporated into a photopolymerized GelMA hydrogel [20] (referred to as GelMA@BM) through a ROS-responsive linker, the benzeneboronic acid pinacol ester group. [21]Upon exposure to the diabetic wound, BM-NDs were rapidly released in response to endogenous ROS stimuli, initiating a self-cascade catalytic reaction that consumed excess glucose, generated •OH, and simultaneously depleted GSH.The resulting high levels of ROS effectively destructed bacteria and polarized macrophages into the M1 phenotype, thereby synergistically eliminating infection.With consumption of the local glucose and degradation of BM-NDs, the level of ROS was downregulated in the wound area, creating a favorable microenvironment for M2 macrophage polarization to promote angiogenesis.Ultimately, this approach mimics the physiological healing process of diabetic infectious wounds (Scheme 1).Considering its straightforward and easily executed protocol, as well as its proven efficacy in eliminating wound infection and promoting tissue regeneration, we anticipate that this two-phase treatment strategy based on GelMA@BM dressing holds significant promise for clinical translation in the management of diabetic infectious wounds.

Preparation, Characterization, and Multiple Enzyme-Like Activities of BM-NDs
The ultrasmall BM-NDs were synthesized following a straightforward biomineralization procedure, [19] in which BSA served as both a reductant and template for reducing potassium permanganate (KMnO 4 ) and accommodating the resulting MnO x nanodots.The strong oxidizing properties of KMnO 4 resulted in the reduction of Mn 7þ to various valences of Mn (Mn 4þ/3þ/2þ ) in the oxides.Among these, Mn 4þ exhibited the most favorable oxidation properties, suggesting the potential for highest catalytic activity. [22]To optimize the ratio of Mn 4þ and the enzymatic activity of the BM-NDs, we varied the pH value of BSA solution to 5.5, 7.0, and 8.5, resulting in the formation of MnO x with distinct ratios of Mn 4þ (denoted as BM-NDs 5.5 , BM-NDs 7.0 , and BM-NDs 8.5 , respectively) (Figure 1A).The synthesized BM-NDs were characterized to be ultrasmall, with a hydrodynamic diameter of less than 5 nm, and demonstrated stability in a physiological-mimicking environment (Figure 1B and S1, Supporting Information).BM-NDs 5.5 , BM-NDs 7.0 , and BM-NDs 8.5 were then analyzed using X-ray photoelectron spectroscopy (XPS) to determine the valence states of the manganese.The XPS spectra revealed distinct peaks corresponding to C1s, O1s, N1s, S2p, and Mn2p, indicating that the BM-NDs were composed of BSA and MnO x (Figure 1C and S2, Supporting Information).Notably, BM-NDs 8.5 exhibited a higher content of Mn 4þ (55.18% vs 31.68% and 36.97%) and a relatively lower amount of Mn 3þ (29.27% vs 50.41% and 45.31%) compared to BM-NDs 5.5 and BM-NDs 7.0 , as observed in the XPS analysis (Figure 1D,E and S3, Supporting Information).These findings suggest that BM-NDs 8.5 may exhibit the highest enzymatic activity.To test this hypothesis, we first evaluated the GOx-mimicking activity of BM-NDs 5.5 , BM-NDs 7.0 , and BM-NDs 8.5 using a dye indicator-based glucose oxidase assay kit.The results demonstrated that all BM-NDs exhibited distinct GOx-mimicking activity, which is consistent with our previous report. [19]Importantly, BM-NDs 8.5 exhibited 1.24-fold and 1.29-fold higher catalytic activity than those of BM-NDs 5.5 and BM-NDs 7.0 at a concentration of 30 μg Mn mL À1 , respectively (Figure 1F).We further investigated the POD-mimicking activity of BM-NDs based on a chromogenic substrate, 3,3 0 ,5,5 0 -tetramethylbenzidine (TMB).Our results revealed that all BM-NDs exhibited robust POD-like properties, with BM-NDs 8.5 showing 1.33fold and 1.31-fold higher enzymatic activity compared to BM-NDs 5.5 and BM-NDs 7.0 at a concentration of 30 μg Mn mL À1 , respectively, which is consistent with the observation of GOxmimicking activity (Figure 1G).Based on these findings, it can be concluded that the valence state of manganese in BM-NDs can be regulated by adjusting the pH value of BSA solution.The enhanced GOx-and POD-mimicking activities of BM-NDs 8.5 were attributed to the higher proportion of Mn 4þ .Therefore, in the subsequent experiments, we employed BM-NDs 8.5 as the optimized nanozyme, and all subsequent mentions of BM-NDs refer specifically to BM-NDs 8.5 unless otherwise specified.
We further verified the GOx-and POD-mimicking activities of BM-NDs using electron paramagnetic resonance (EPR) spectroscopy in the presence of glucose (30 mM).The spectrum of BM-NDs þ glucose exhibited a characteristic four-peak pattern of •OH with an intensity ratio of 1:2:2:1, [23] whereas glucose alone did not show such a pattern (Figure S4, Supporting Information).This observation indicated that BM-NDs could catalyze glucose to generate H 2 O 2 and subsequently decompose it into •OH.Moreover, BM-NDs exhibited a notable GSHOx-mimicking property, effectively depleting GSH under mild acidic Scheme 1. Schematic illustration of the time-phased anti-infection and regeneration of diabetic wound by modulating local ROS levels using GelMA@BM dressing.microenvironment (pH 6.4), similar to the pH conditions observed in infected tissues. [24]In contrast, the activity of BM-NDs was negligible under neutral conditions (pH 7.4) (Figure S5, Supporting Information).Thus, the pH-dependent GSHOxmimicking activity of BM-NDs enables its specific application in antibacterial therapy.
The size and morphology of BM-NDs were characterized using transmission electron microscopy (TEM), revealing a uniform distribution of nanodots with a size smaller than 5 nm (Figure 2A).Elemental mapping analysis demonstrated a homogeneous distribution of Mn element in the BSA corona (Figure 2A).Furthermore, the powder X-ray diffraction (XRD) pattern of BM-NDs showed broad amorphous features with a peak at 21.06°, corresponding to the (101) plane of MnO 2 (PDF#42-1316) (Figure 2B).These findings confirm the successful synthesis of ultrasmall BM-NDs through a facile BSA-directed biomineralization procedure.

Preparation of ROS-Responsive GelMA@BM
Gelatin is a commonly used biomaterial in tissue regeneration studies due to its remarkable biocompatibility and modifiability. [25]elMA, a derivative of gelatin that has been modified with alkene groups through photo-cross-linking technology, has been demonstrated to possess outstanding mechanical properties such as toughness, stretchability, and adhesiveness.These properties allow the hydrogel to maintain its compactness over the wound morphology, while also facilitate ease of wound dressing  5.5 , BM-NDs 7.0 , BM-NDs). 8.5 B) Hydrodynamic size distribution of BM-NDs.C) XPS spectrum of MN-NDs 8.5 .D) XPS spectra of Mn2p for BM-NDs 8.5 .E) The ratios of Mn 4þ and Mn 3þ in BM-NDs 5.5 , BM-NDs 7.0 , BM-NDs 8.5 , respectively.F) GOx-mimicking activity of BM-NDs (n = 3).G) POD-mimicking activity of BM-NDs characterized by chromogen TMB (n = 3).*** P < 0.001.preparation and application. [26]Therefore, in this study, we utilized GelMA hydrogel as a carrier to stabilize and load BM-NDs.To enhance the stability of BM-NDs within GelMA and simultaneously enable their burst release upon in vivo application, we employed a ROS-responsive linker to covalently link BM-NDs with GelMA molecules, as reported previously. [21]To achieve this, 4-carboxyphenylboronic acid was used to modify BM-NDs, while 3-amino-1,2-propanediol was used to modify GelMA hydrogel (Figure S6, Supporting Information).The modified BM-NDs and GelMA hydrogel were then co-incubated overnight to form GelMA@BM (Scheme 1).Scanning electron microscopy (SEM) images revealed that GelMA@BM exhibited an interconnected porous structure inside the photo-cross-linked hydrogel, which could serve as a protective covering and facilitate the regulation of wound bioliquids [27] (Figure 2C).Importantly, the conjugation of BM-NDs had minimal impact on the microstructure of the GelMA hydrogel, but slightly increased the surface roughness, indicating the presence of BM-NDs (Figure 2C and S7, Supporting Information).Energy-dispersive spectroscopy (EDS) analysis and elemental mapping results demonstrated the existence and uniform distribution of BM-NDs in the GelMA hydrogel (Figure 2D,E).
The GelMA@BM hydrogel was designed to incorporate a benzeneboronic acid pinacol ester group, which enabled the ROS-responsive release of BM-NDs in the wound area.To investigate the release behavior of BM-NDs from the hydrogel, we incubated the GelMA@BM hydrogel (500 μL) with equal volume of phosphate buffer solution (PBS, pH 7.4) or H 2 O 2 solution (100 μM) for 0.5 h and then quantified the released Mn content in the eluents using inductively coupled plasma-mass spectrometry (ICP-MS).As a control, we also prepared the GelMA þ BM hydrogel by incubating unmodified BM-NDs with GelMA hydrogel overnight.The results demonstrated that BM-NDs could be rapidly released from GelMA þ BM hydrogel to PBS, while negligible amounts of BM-NDs were observed in the eluent of GelMA@BM hydrogel (Figure 2F and S8, Supporting Information).However, after exposure to H 2 O 2 , the release rate of BM-NDs from GelMA@BM hydrogel was similar to that of the GelMA þ BM group (Figure 2F and S8, Supporting Information), indicating the triggered release of BM-NDs through the hydrolysis of ROS linker.These findings provide compelling evidence for the use of GelMA@BM hydrogel as a promising wound dressing material for the treatment of diabetic chronic wounds.

BM-NDs-Induced Cascade Reactions for Multidrug-Resistant (MDR) Bacterial Killing
Based on the confirmed GOx-, POD-, and GSHOx-mimicking activities of BM-NDs, we hypothesized that these nanodots could effectively combat bacterial infection through high levels of ROS.To test this hypothesis, we exposed Gram-negative MDR Escherichia coli (E.coli) and Gram-positive MDR Staphylococcus aureus (S. aureus) to gradient concentrations of BM-NDs (0, 1, 3, 10, and 30 μg Mn mL À1 ) under low-glucose (5.5 mM) or high-glucose (30 mM) conditions for 6 h, then assessed the bacterial activity using the plate count method with diluted bacterial suspensions (10 À4 -fold, 100 μL).Our results showed that BM-NDs could significantly inhibit bacterial growth under both glucose conditions, and an even more significant antibacterial activity of BM-NDs was observed under high-glucose conditions at each concentration examined (Figure 3A,B), where more abundant ROS was generated due to the ample glucose substrate.As the concentration of BM-NDs increased from 1 to 10 μg mL À1 , the antibacterial activity gradually increased.However, further increasing the concentration of BM-NDs to 30 μg mL À1 failed to demonstrate superior antibacterial effects compared to 10 μg mL À1 of BM-NDs (Figure 3A,B), likely due to the rate of ROS generation being limited under the current reaction conditions and treatment duration.To further evaluate the concentration-and glucose-dependent antimicrobial activity of BM-NDs, live/dead bacterial staining assays were performed.In the absence of BM-NDs, both MDR E. coli and MDR S. aureus showed prevalent green fluorescence (indicating the live bacteria) and negligible red fluorescence (indicating the dead bacteria) (Figure 3C).As the concentration of BM-NDs increased to 10 μg mL À1 , both bacteria under low-glucose conditions showed enhanced red fluorescence, which was even more pronounced under high-glucose conditions.The fluorescence of bacteria treated with 30 μg mL À1 of BM-NDs was comparable to that of bacteria treated with 10 μg mL À1 of BM-NDs, which is consistent with the results obtained from the bacterial plate count assay.Consequently, the optimal antimicrobial concentration of BM-NDs for subsequent experiments was determined to be 10 μg mL À1 .
The bacterial cell membrane functions not only as a protective barrier and material exchange site but also as a crucial cellular structure that houses the respiratory chain, which is essential for bacterial survival. [28]Previous studies have demonstrated that the bactericidal effect of high levels of ROS is attributed, in part, to lipid peroxidation and respiratory chain damage occurring on the cell membrane. [29]To gain further insights into the antibacterial activity of BM-NDs, we conducted SEM to observe the micro-morphology of both MDR bacterial strains following treatment with 10 μg mL À1 of BM-NDs under low-and high-glucose conditions (5.5, 30 mM).Compared to the smooth and intact cell membrane observed in the control group, treatment with BM-NDs resulted in wrinkles and collapses in the bacteria under low-glucose conditions.Notably, under high-glucose conditions, the bacteria treated with BM-NDs exhibited even more severe damage, including cell dehiscence (Figure 3D).These findings provide preliminary confirmation of the potent antibacterial properties of BM-NDs in high-glucose environments.
We subsequently investigated the intracellular levels of GSH and ROS in MDR bacteria treated with BM-NDs (0, 1, 3, 10, and 30 μg Mn mL À1 ) under high-glucose conditions (30 mM).The ROS levels in the bacteria increased in a concentrationdependent manner upon treatment with increasing concentrations of BM-NDs, while the GSH content showed a concentration-dependent decrease, with no further changes observed beyond a concentration of 10 μg mL À1 (Figure 4A,B).Additionally, bacteria treated with BM-NDs (10 μg Mn mL À1 ) under high-glucose conditions (30 mM) exhibited intense green fluorescence emitted by 2,7-dichlorofluorescein (DCFH) probes, [30] indicating high levels of intracellular ROS generated by BM-NDs treatment (Figure 4C).These findings provide further support for the role of BM-NDs in inducing intracellular ROS overload and suggest a potential mechanism underlying ROS-induced damage to bacterial cell membranes.
The proposed mechanism for the antimicrobial activity of BM-NDs involves the generation of ROS and depletion of GSH through the self-cascade enzymatic activities of BM-NDs, as discussed previously.To further investigate this hypothesis, we conducted GSH-rescue and ROS-quenching experiments by supplementing GSH (400 μM), catalase (500 U mL À1 ), or isopropanol (1% w/v) to BM-NDs (10 μg Mn mL À1 ) treatment under high-glucose conditions (30 mM) for further antibacteria experiments.The results demonstrated that the antibacterial efficacy of BM-NDs significantly decreased after the addition of catalase (scavenger of H 2 O 2 ) or isopropanol (scavenger of •OH) in the culture media, as evidenced by denser colony formation and a more intense green fluorescence signal in the live/dead staining experiment compared to the BM-NDs group.These findings reveal the critical roles of H 2 O 2 and •OH in the antimicrobial process (Figure 4D-F).In addition, the antibacterial efficiency of BM-NDs exhibited a moderate reduction upon the application of exogenous GSH to bacteria, indicating the significance of GSH depletion in the antimicrobial effect.Furthermore, SEM images revealed that the micromorphological damages of both MDR bacteria were significantly alleviated with the addition of GSH.Meanwhile, BM-NDs nearly lost antibacterial activity after the administration of H 2 O 2 or •OH scavengers, as indicated by the smooth and intact micromorphology of both MDR bacterial strains (Figure 4G).Our results were consistent with a previous study that revealed the effectiveness of the antibacterial strategy of ROS generation combined with GSH consumption. [31]

Macrophage Polarization Regulated by BM-NDs
In the context of diabetes mellitus, impaired macrophage polarization from M1 to M2 due to the hyperglycemic microenvironment is a critical factor that adversely affects wound healing. [32]It is noteworthy that both the type and timing of macrophage polarization are equally important in the context of wound healing, particularly in diabetic infectious wounds.The sequential utilization of the anti-infection activity of M1 macrophages [33] and the proregeneration property of M2 macrophages [34] can significantly affect the therapeutic outcome in such cases.Previous studies have demonstrated that hyperglycemia-induced high levels of ROS are key factors promoting chronic M1 macrophage polarization in diabetic wounds, [10] whereas a benign level of ROS is crucial for the M2 polarization phenotype. [35]Accordingly, we aimed to investigate the role of BM-NDs-generated high levels of ROS in activating M1 polarization phenotype and phagocytosis activity of macrophages in a high-glucose environment.Additionally, we sought to determine whether low concentrations of glucose and BM-NDs, indicative of benign levels of ROS, could activate the M2 polarization phenotype and promote the proangiogenic potential of macrophages.To ensure the safety of using BM-NDs in macrophage polarization experiments, we initially assessed their cytotoxicity by conducting a CCK-8 assay on RAW264.7 cells treated by BM-NDs (0, 1, 3, 10, and 30 μg Mn mL À1 ) under low-glucose (5.5 mM) or high-glucose (30 mM) conditions.The results determined that cellular proliferation activity was not significantly affected by BM-NDs up to the optimal antibacterial concentration of 10 μg mL À1 , possibly attributed to the more complete oxidation-reduction system of eukaryotic cells (Figure 5A).As BM-NDs may potentially regulate macrophage polarization through their influence on intracellular ROS levels, we next measured the intracellular ROS levels of RAW264.7 cells treated with different concentrations of BM-NDs (0, 1, and 10 μg Mn mL À1 ) in both glucose environments (5.5 and 30 mM) using 2 0 ,7 0 -dichlorodihydrofluorescein diacetate (DCFH-DA) probes.In the absence of BM-NDs, we observed minimal intracellular ROS in macrophages cultured under low-glucose conditions.In contrast, when subjected to high-glucose conditions that mimic chronic inflammation in diabetic wounds, there was a noteworthy elevation of intracellular ROS.The observed increase was found to be even higher than that observed in macrophages treated with a low concentration of BM-NDs under low-glucose conditions (Glu lo BM-NDs lo ) (Figure 5B,C).Moreover, the concentration of BM-NDs closely correlated with an increase in ROS levels in macrophages under both low-and high-glucose conditions (Figure 5B,C).It was noteworthy that BM-NDs induced a greater amount of intracellular ROS in high glucose concentrations compared to those in low glucose concentrations.Specifically, the ROS levels in macrophages treated with a high concentration of BM-NDs under high-glucose conditions (Glu hi BM-NDs hi ) were found to be the highest among all groups, as evidenced by the most intense green fluorescence (Figure 5B,C).Based on the intracellular ROS levels of RAW264.7 cells across all groups, our findings suggested that macrophages subjected to Glu lo BM-NDs lo conditions had the potential to exhibit an M2 phenotype, while those treated under Glu hi BM-NDs hi conditions would demonstrate a marked M1 phenotype.
To validate our hypotheses, we examined the transcriptional expression of polarization-related genes in macrophages from each treatment group.Our results demonstrated that the expression of M1 polarization-related genes in macrophages under the Glu lo BM-NDs lo environment remained unchanged, while the expression of M2 polarization-related genes significantly increased (Figure S9, Supporting Information).Conversely, macrophages incubated under the Glu hi BM-NDs hi condition exhibited significant upregulation of M1-related genes, which was consistent with the observed variation in intracellular ROS levels.Notably, the expression of M2-related genes in macrophages from these groups was not significantly upregulated, and the expression of MRC1 and TGF-β genes in the Glu hi BM-NDs hi group decreased (Figure S9, Supporting Information).To further investigate the impact of different concentrations of BM-NDs (0, 1, and 10 μg Mn mL À1 ) on macrophage polarization under varying glucose concentrations (5.5 and 30 mM), flow cytometry analysis was conducted to examine specific membrane proteins associated with polarization.The treated macrophages were co-stained with PE-conjugated CD86 (an M1 marker) [36] and APC-conjugated CD206 (an M2 marker) [37] antibodies.The flow cytometry results showed that only macrophages in the Glu lo BM-NDs lo group exhibited a significant increase in CD206 expression, confirming the M2 macrophage phenotype (Figure 5D,E).In contrast, the expression of CD86 was upregulated to varying degrees in the macrophages of the other treatment groups, particularly the Glu hi BM-NDs hi  treatment group.Rhodamine-labeled phalloidin revealed the small, round morphology of M0 macrophages and their colony-forming characteristics in the absence of high levels of glucose and BM-NDs (Figure 5F,G).The expression of polarizationrelated markers was minimal in these cells.Macrophages in the Glu lo BM-NDs lo group exhibited similar cell size and assembling characteristics as M0 macrophages, but they were mainly spindle shaped with highly expressed M2 polarization marker Arg-1, confirmed by intense green fluorescence (Figure 5F,G).In other groups, the size of macrophages increased significantly, and their morphology showed highly heterogeneous pleomorphism (Figure 5F,G), which is indicative of the M1 phenotype. [38]otably, the expression of iNOS was significantly upregulated in macrophages in the Glu hi BM-NDs hi group, consistent with qPCR and flow cytometry analysis results.To summarize, our results provided evidence that high concentrations of BM-NDs could induce M1 polarization of macrophages in a high-glucose environment, mimicking the initial stages of BM-NDs treatment in diabetic wounds.Conversely, low concentrations of BM-NDs were found to activate M2 macrophages in a low-glucose environment, reflecting the later stages of treatment characterized by local glucose consumption and degradation of BM-NDs.
As previously discussed, macrophages have the potential to promote diabetic wound repair through their bactericidal activity as M1 macrophages and proangiogenic potential as M2 macrophages.Therefore, we investigated the effects of glucose and BM-NDs on the properties of M1 and M2 macrophages, specifically their ability to phagocytose bacteria and promote angiogenesis in vitro.Initially, we compared the bactericidal ability of macrophages treated with varying concentrations of BM-NDs (0, 1, and 10 μg Mn mL) in a high-glucose environment (30 mM).Our findings revealed that macrophages treated with BM-NDs exhibited enhanced bactericidal activity compared to untreated macrophages.Furthermore, we observed that as the concentration of BM-NDs increased, macrophages demonstrated an enhanced ability to phagocytose MDR E. coli and S. aureus, as evidenced by an accumulation of propidium iodide (PI)-labeled dead bacteria within macrophages (Figure 6A,B).Additionally, bio-TEM images revealed that macrophages treated with high concentrations of BM-NDs in high-glucose environments efficiently phagocytosed MDR bacteria and disrupted their morphology in phagosomes (Figure 6C).These results confirmed superior antimicrobial function of M1 macrophages polarized in an environment with high concentrations of BM-NDs and glucose.In addition, we investigated the proangiogenic function of M2 macrophages induced by low concentrations of BM-NDs and glucose.Macrophages were treated with varying concentrations of BM-NDs (0, 1, and 10 μg Mn mL À1 ) in a low-glucose environment (5.5 mM), followed by culturing in complete medium for additional 48 h.The resulting conditioned medium (CM) was collected and subsequently used to incubate the EA.hy926 endothelial cell line seeded on the Matrigel.The results of the tube formation assay indicated that the CM obtained from M2 macrophages in the Glu lo BM-NDs lo group elicited a significant increase in the total tube length and branch points as compared to those derived from the control groups of M0 macrophages and M1 macrophages that were polarized from the Glu lo BM-NDs hi group (Figure 6D,E).Furthermore, the results of the Transwell assay revealed that the macrophages from Glu lo BM-NDs lo group induced a significantly greater mobilization of endothelial cells, thereby highlighting their superior provasculogenic potential (Figure 6F,G, S10, Supporting Information).These results provided further evidence of the therapeutic potential of BM-NDs in promoting diabetic wound repair through their ability to modulate the time-phased polarization of macrophage phenotypes.

Diabetic Infectious Wound Healing by GelMA@BM Dressing
First, we established a type II diabetes model in Sprague-Dawley rats by administering a high-sugar, high-fat diet, and streptozotocin injection. [39]Circular skin defects with an 8-mm diameter were then created on the back of the diabetic rats using a skin biopsy punch, and MDR S. aureus (100 μL, OD = 1) was inoculated onto the wounds to establish a diabetic infectious wound model.Next, we applied GelMA@BM dressing (10 μg mL À1 ) to the wounds, with PBS, plain GelMA hydrogel, and vancomycinloaded GelMA (GelMA þ VMC) (20 μg mL À1 ) as controls.We recorded the wound morphology immediately after modeling and at 2, 4, and 8 days after surgery.Two days post-operation, the wounds in the PBS and plain GelMA hydrogel groups exhibited poor scabbing and evident abscess formation, with no reduction in wound dimension; the GelMA þ VMC group showed a slight reduction in wound dimension, likely attributed to the antibiotic effect of vancomycin (Figure 7A-C).Remarkably, GelMA@BM inhibited the formation of infectious abscess, and the wounds exhibited significant contraction and formation of scabs (Figure 7A-C).These findings strongly suggested that BM-NDs were rapidly released into the wound area, exerting an anti-infective effect.At the conclusion of the treatments (day 8), the GelMA þ VMC group showed no evident signs of infection but still had unclosed wounds, whereas the GelMA@BM hydrogel dressing almost fully healed the diabetic infectious wounds.In contrast, the PBS and GelMA groups resulted in chronic wounds that remained unhealed (Figure 7A-C).In addition, the infected tissues were harvested 8 days after the treatment and homogenized for the blood agar plate assay.The results demonstrated the presence of bacterial infection in both the PBS and GelMA groups, whereas the administration of vancomycin and BM-NDs in the hydrogel dressing markedly mitigated the infection (Figure 7D,E).
Compared to vancomycin, the superior therapeutic efficacy of BM-NDs on diabetic infectious wounds could be attributed, in part, to the consumption of local pathological glucose.To confirm this, we measured the glucose content in the callus using the tissue homogenate 2 and 8 days after treatment.Two days after treatment, GelMA@BM significantly reduced the local glucose levels to a normal physiological level.By the end of the 8-day treatment, although there was a slight relapse, the glucose content remained significantly lower than that of other groups (Figure 7F).Eight days after bacterial inoculation, the infected tissues were collected for histological detection.Hematoxylin and eosin (H&E) and Masson's trichrome staining of the wound tissues demonstrated that both the PBS and GelMA groups exhibited incomplete epithelial layers, along with the presence of abscesses or infected necrotic tissues.The wounds in the GelMA þ VMC group exhibited a preferable healing tendency, despite some discontinuity in the epidermal and dermal tissues.In contrast, the wounds treated with GelMA@BM were completely healed (Figure 7G).Combined with the in vitro results, we inferred that the time-phased antibacterial and wound healing effects of GelMA@BM hydrogel were attributed to the ROS-mediated M1 macrophage domination in the early stage, followed by M2 macrophage polarization in the late stage of treatment.
To further test our hypothesis, the infected tissues were collected and prepared as frozen sections 2 and 8 days after  treatment for ROS detection and immunofluorescence staining.ROS levels in the infected tissues were first assessed using a dihydroethidium probe two days after the treatments.The results showed that GelMA@BM treatment resulted in a higher production of ROS, as evidenced by the intense red fluorescence observed within the wound, compared to other treatments (Figure 8A,B).Furthermore, immunofluorescence double staining revealed that the GelMA@BM hydrogel dressing significantly accumulated M1 macrophages (CD68 and iNOS co-staining) [40] (Figure 8C,D), which coincided with the high local ROS content.This suggested the ability of GelMA@BM to highly activate M1 macrophage-mediated nonspecific immunity.In contrast, the number of mature M1 macrophages in the wounds treated with vancomycin was slightly reduced compared to the PBS and GelMA groups (Figure 8C,D), which may be related to the controlled infection.Moreover, following the 2-day treatment period, no significant M2 polarization was detected in the local tissues of any group (Figure S11A,B, Supporting Information).This finding aligns with the physiological healing process, where M2 macrophages predominantly exert their prohealing role during the late stages. [8]Eight days after the treatments, the ROS content was significantly lower in the late stage of wound healing upon GelMA@BM hydrogel application, compared to other groups, indicating the alleviation of the chronic inflammatory environment caused by diabetes (Figure 8E,F).Simultaneously, the local tissues in the GelMA@BM group exhibited a notable decrease in M1 polarization (Figure S11C, D, Supporting Information), which aligns with the findings obtained from ROS level measurements.Additionally, the abundance of M2 macrophages (CD68 and Arg-1 co-staining [41] ) in the wounds of the GelMA@BM group was significantly higher than that of other groups (Figure 8G,H), elucidating the unique potential of BM-NDs in promoting tissue repair compared to vancomycin.To assess the formation of local neovascularization, we conducted CD31 and α-SMA double-labeled immunofluorescence staining 8 days after the treatment. [42]The GelMA@BM treatment group exhibited more robust angiogenesis within the wounds (Figure 8I,J), which could accelerate the regeneration process of local tissues.Moreover, the H&E staining of major organs in the GelMA@BM group demonstrated no obvious pathological changes, indicating the reliable biosafety of the treatment (Figure S12, Supporting Information).All these results highlighted the effectiveness of our sequential glucose-driven treatment strategy for diabetic infectious wounds.

Conclusions
In summary, we have developed a novel two-phase treatment strategy to effectively manage diabetic wounds by successively regulating local ROS levels.Our approach involved incorporating ultrasmall BM-NDs into GelMA hydrogel via a ROS-responsive linker.The BM-NDs exhibited multienzyme properties that triggered a self-cascade reaction, leading to the decomposition of glucose and generation of ROS, while simultaneously depleting GSH.Upon application to the wound, GelMA@BM dressing initiated a controlled ROS burst in the first phase, effectively combating persistent bacterial infections and promoting M1 macrophage polarization for bacterial phagocytosis.In the second phase, with the depletion of local glucose and degradation of BM-NDs, the dressing maintained a benign level of ROS to stimulate M2 macrophage polarization for wound healing.Our study demonstrated the potent antibacterial and prohealing efficacy of this approach, which has significant clinical implications for improving patient outcomes and reducing healthcare costs.

Experimental Section
Synthesis and Characterizations of BM-NDs: BSA-coated MnO x nanodots with different valence states of Mn were prepared according to our previous study with minor modifications. [19]Briefly, the pH value of BSA solutions (250 mg, 5 mL) was adjusted to 5.5, 7.0, and 8.5 using Na 2 CO 3 /HCl, followed by the dropwise addition of KMnO 4 (400 μL, 100 mM), and the solutions were continuously stirred (500 rpm) for 2 h at 37 °C using a magnetic heating agitator (DLAB, China).After being purified by the ultrafiltration (MWCO = 100 K), the products were lyophilized for subsequent characterizations and usages.The products synthesized under the pH of 5.5, 7.0, and 8.5 were abbreviated as BM-NDs 5.5 , BM-NDs, 7.0 and BM-NDs 8.5 , respectively.
Dynamic light scattering was performed to measure the hydrodynamic size of BM-NDs on Zetasizer Nano ZSP (Malvern Panalytical, UK).The XPS was conducted on K-Alpha XPS System (Thermo Fisher Scientific, USA) for the determination of the valence states of manganese.The morphology and elemental mapping of BM-NDs were evaluated using the TEM, high-resolution TEM, and EDS on Talos F200X G2 (FEI, USA).The XRD was performed on D8 Advance Da Vinci (Bruker, Germany).
GOx-Mimicking Activity: A commercial glucose oxidase activity assay kit (ab219924, Abcam) was utilized for the assessment of the GOx-mimicking activity of BM-NDs.Briefly, 50 μL of BM-NDs 5.5 , BM-NDs 7.0 , or BM-NDs 8.5 with different concentrations (0, 1, 2, 5, 10, and 30 μg Mn mL À1 ) were added to a 96-well plate.Then 50 μL of glucose oxidase reaction mix was added to each well to make the total assay volume of 100 μL.After mixing well, the plate was incubated for 10 min at 37 °C protected from light, then the optical density (OD) values at 570 nm were monitored using a microplate reader (Molecular Devices, USA) to represent the GOx activity of the samples.
POD-Mimicking Activity: The POD-mimicking activity of BM-NDs was tested using the TMB substrate solution (P0292, Beyotime).After the incubation of BM-NDs (20 μL; 0, 1, 2, 5, 10, and 30 μg Mn mL À1 ) in each group (BM-NDs 5.5 , BM-NDs 7.0 , and BM-NDs 8.5 ) with the TMB chromogen solution (200 μL) for 10 min, the OD values at 650 nm were recorded to represent the POD-mimicking activity of BM-NDs.The •OH generation was further evaluated by the EPR spectroscopy (Bruker, Germany).Briefly, the BM-NDs 8.5 (10 μg Mn mL À1 ) were first incubated with glucose (30 mM) for 10 min, with the glucose solution (30 mM) as the control group.The treated samples (30 μL) were then mixed with DMPO (100 mM, 30 μL) and put in capillary tubes for EPR measurements at 1 G modulation amplitude and 100 G sweep width with 6.325 mW power (Bruker EMX plus, Germany).
In vitro Antibacterial Activity of BM-NDs: MDR E. coli (Rosetta-gami2 (DE3)) and MDR S. aureus (ATCC43300) were used for the evaluation of the antibacterial performance of BM-NDs.The bacteria were cultured in Trypticase Soy Broth Medium (LA0020, Solarbio) at 37 °C for 12 h.After washed with saline for three times, the bacteria suspensions were placed in a 96-well plate.Subsequently, the bacteria were incubated with BM-NDs in different concentrations (0, 1, 3, 10, and 30 μg Mn mL À1 ) under low (5.5 mM) and high (30 mM) glucose conditions.In another series of experiments, bacteria were incubated with BM-NDs in a therapeutic concentration (10 μg Mn mL À1 ) under the high-glucose condition in the supplement of exogenous GSH (400 μM), H 2 O 2 scavenger (catalase; E3289, Sigma Aldrich; 500 U mL À1 ), or •OH scavenger (isopropanol; 563 935, Sigma Aldrich; 1% w/v).After incubated at 37 °C for 6 h, the bacteria suspension (diluted by 10 À4 folds, 100 μL) was spread on the blood agar plates (B530120, Sangon Biotech) culturing at 37 °C.After 24 h, the bacterial colonies on the plate were analyzed.For live/dead staining (L7007, Thermo Fisher Scientific) analysis, 50 μL of bacteria suspension was mixed with SYTO 9 (1 μL, 6 μM) and propidium iodide (PI) (1 μL, 30 μM) for 15 min protected from light.Then the stained bacteria were washed with saline and imaged by confocal laser scanning microscope (Leica, Germany).The bacterial morphology was characterized by SEM (Hitachi, Japan).After antibacterial experiments, the bacteria were fixed with glutaraldehyde (2.5% w/v) and sequentially dehydrated with gradient ethanol solutions.The obtained samples were observed after supercritical drying and sputter coating with gold.
Macrophage Response to BM-NDs in vitro: Macrophage-like cell line RAW264.7 (SCSP-5036, National Collection of Authenticated Cell Cultures) was utilized for evaluating the biocompatibility and polarization-inducing activity of BM-NDs.The viability of RAW264.7 cells upon BM-NDs treatment was determined using the CCK-8 assay.Briefly, macrophages were inoculated in 96-well plates (5000 cells well À1 ) and incubated with a series of concentrations of BM-NDs (0, 1, 3, 10, and 30 μg Mn mL À1 ) under different glucose conditions.After incubation for 24 h, 90 μL of fresh culture medium and 10 μL of CCK-8 reagent (CK04, Dojindo) were mixed into each well for incubation for another 2 h.Then, the OD value at 450 nm was measured using a microplate reader.The absorbance obtained from the mixed 100 μL of CCK-8 reagent in blank wells was defined as a blank control and was subtracted to obtain the corrected values to calculate the relative cell viability.Moreover, the macrophages were stimulated with different concentrations of BM-NDs (0, 1, and 10 μg Mn mL À1 ) for 48 h in the presence or absence of high glucose.After treatment, total RNA was extracted, and cDNA was obtained using the RNA Purification and Reverse Transcription Kits (B0004D and A0010, EZBioscience).qPCR was then performed using SYBR Green qPCR Master Mix (A0001, EZBioscience) to examine the transcriptional expressions of polarization-related genes.Relative expression levels were calculated with GAPDH as the reference gene for normalization.For flow cytometry analysis, induced macrophages were fixed, blocked, and then incubated with APC-conjugated CD206 (1:100; 17-2061-82, eBioscience) and PE-conjugated CD86 (1:100; 12-0862-82, eBioscience) antibodies for 30 min.The expression levels of CD86 (M1 polarization marker) and CD206 (M2 polarization marker) were detected using a Flow Cytometer (BD Biosciences, USA).F-actin staining combined with immunofluorescence was conducted to further determine the morphology and polarization markers expression of induced RAW264.7 cells.After BM-NDs treatment, the macrophages were fixed, blocked, and incubated with the primary antibody overnight at 4 °C, followed by incubation with an FITC-conjugated secondary antibody (1:1000; ab6717, Abcam) at 25 °C for 2 h.F-actin was stained with TRITC Phalloidin (100 nm; 40734ES75, Yeasen), and nuclei were counterstained with DAPI (1 μg mL À1 ; 40727ES10, Yeasen).The primary antibodies used in this study included anti-iNOS (1:500; ab178945, Abcam) and anti-Arg-1 (1:100; 93 668, Cell Signaling Technology).
Intracellular ROS and GSH Content Measurement: DCFH-DA (S0033S, Beyotime) was used as an oxidant-sensitive fluorescence probe to detect the intracellular ROS.After BM-NDs treatment, the bacteria and macrophages were incubated with DCFH-DA (10 μM) at 37 °C for 30 min protected from light, then washed and observed using a confocal laser scanning microscope.A Micro Reduced GSH Assay Kit (KTB1600, Abbkine) was used to examine the intracellular GSH content according to the manufacturer's instructions.Briefly, the bacteria were resuspended in the Extraction Buffer followed by freeze-thawing steps to obtain the supernatant.Then the samples were incubated with Assay Buffer for 2 min, and the OD value at 412 nm was recorded.The absorbance obtained from the blank wells was subtracted to obtain the corrected values for calculating the relative GSH content.
Antibacterial Activity of BM-NDs-induced Macrophages: RAW264.7 cells were cultured with a special medium (CM-0190, Procell) at 37 °C in a humidified atmosphere with 5% CO 2 .Upon BM-NDs treatment under the high-glucose condition for 48 h, MDR E. coli and S. aureus solutions (OD = 1) were suspended (1: 100) into the medium for 3 h.After discarding the supernatant and washing 3 times with PBS, PI (40755ES64, Yeasen) was used to stain dead bacteria, and DAPI was used to stain cell nuclei.The morphology of macrophages with endocytosed bacteria was observed by TEM (FEI, USA).Briefly, the fixed macrophages were stained with osmium tetroxide, washed with ultrapure water, and dehydrated with a series of gradient ethanol and acetone solutions before embedded in epoxy resin.Then, the ultrathin sections were prepared and stain with uranyl acetate and lead citrate for TEM imaging.
Angiogenesis Assays in vitro: The RAW264.7 cells were treated with different concentrations of BM-NDs (0, 1, and 10 μg Mn mL À1 ) under low-glucose condition.After 2 days of induction, the medium was replaced with a complete medium for a subsequent 2 days of incubation.The CM from each group was harvested for further angiogenesis assays to determine the proangiogenic potential of macrophages.For the tube formation assay, the endothelial cell line Ea.hy926 cells (20 000 cells/well) were seeded onto 96-well plates coated with Matrigel (356 231, Corning) and treated with CM from each group for 6 h.Then the cells were stained with Calcein-AM (5 μM; 40719ES80, Yeasen) and observed using a fluorescence microscope (Leica, Germany).For the cell migration assay, endothelial cells (20 000 cells well À1 ) were loaded into the upper chamber of a Transwell plate (3422, Corning) and co-incubated with treated macrophages in the lower chamber from different groups for 12 h.With unmigrated cells removed by cotton swabs, the migrated cells that passed through the 8-μm pores were fixed and stained with crystal violet (0.5% w/v; 60506ES60, Yeasen) for observation.
In vivo Diabetic Infectious Wound Healing Evaluation: All animal procedures were approved by the Animal Research Committee of Shanghai Sixth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (DWSY2023-0048).After two weeks of high-fat feed, the male Sprague-Dawley rats (approximately 200 g) were intraperitoneally injected with streptozotocin (10 mg mL À1 in 0.01 M citrate buffer; V900890, Sigma Aldrich) after fasting for 12 h.Two weeks later, individuals with plasma glucose levels above 16.7 mM over two continuous detections were defined as type II diabetic rats.After anesthesia and hair removal, fullthickness wounds were created using a biopsy punch (diameter: 8 mm) on their backs, and MDR S. aureus (100 μL, OD = 1) was inoculated on the wounds to establish the diabetic infectious wound model.PBS solution, plain GelMA hydrogel, vancomycin (20 μg mL À1 ; V2002, Sigma Aldrich)-loaded GelMA, and GelMA@BM (10 μg mL À1 ) were severally placed onto the wounds for experimental treatment.Wound photos were obtained immediately after modeling and 2, 4, and 8 days after surgery.Wound healing rate was calculated by the following equation where R and S 8 are the wound healing rate and wound area at day 8, respectively.S 0 is the initial wound area.On the 8th day posttreatment, the related wound tissues were harvested and homogenized in normal saline.100 μL of the diluted solution (10 À4 folds) was poured onto blood agar plate for culturing to assess the in vivo antibacterial capacity of the hydrogels.After 2 and 8 days of treatment, tissue was taken from the wound and homogenized for the measurement of local glucose concentration using a commercial glucose assay kit (BC2505, Solarbio).Histological Analysis: At different time points (2 and 8 days) after treatments, the wound tissue was collected, fixed, dehydrated using a graded alcohol series, paraffin-embedded, sliced into 5 μm-thick sections, and stained with H&E (G1120, Solarbio) and Masson's trichrome (G1346, Solarbio) for histological evaluation following the manufacturer's instructions.For immunofluorescence staining, tissue samples were collected 2 and 8 days after treatments, dehydrated in 30% sucrose solution for 24 h, embedded in OCT, and finally sliced into 10 μm-thick sections.The ROS levels were determined by DHE (50102ES02, Yeasen) staining.M1 and M2 macrophage polarization within tissues after GelMA@BM treatment were evaluated using CD68þiNOS and CD68 þ Arg1 immunofluorescence double staining, respectively, at the therapeutic day 2 and 8. Local neovascularization in the late stage of treatment (8th day) was examined by the CD31 þ α-SMA double-labeled immunofluorescence staining.Briefly, the sections were incubated overnight at 4 °C with primary antibodies after blocking, followed by incubation with fluorescent-labeled secondary antibodies (1:200; ab97035, ab150077, Abcam) at room temperature for 2 h.DAPI was used to stain the cell nuclei, and the sections were observed under a fluorescence microscope.The primary antibodies utilized in this study included anti-CD68 (1:1000; ab955, Abcam), anti-iNOS (1:500; ab178945, Abcam), anti-Arg-1 (1:500; ab133543, Abcam), anti-CD31 (1:100; ab64543, Abcam), and anti-α-SMA (1:500; ab124964, Abcam).After 8 days of treatment, major organs (heart, liver, spleen, lung, kidney, and brain) were collected from the rats in PBS and GelMA@BM treatment groups.Paraffin sections were prepared as described above for H&E staining (G1120, Solarbio) to examine the latent pathological changes.
Statistical Analysis: All experimental data are presented as mean AE standard deviation.Statistical analysis was performed using Student's t-test for comparisons between two groups, or one-way analysis of variance (ANOVA) followed by Tukey's post hoc test for comparisons among multiple groups, utilizing GraphPad Prism 8 software (GraphPad Software, USA).Differences were considered significant with a P value less than 0.05.

Figure 3 .
Figure 3. Antibacterial performance of different concentrations of BM-NDs in vitro.A) Photographs of bacterial colonies and B) bacterial viability analysis of MDR E. coli and S. aureus treated with different concentrations of BM-NDs under different glucose conditions (n = 3).C) Live/dead fluorescence staining of MDR E. coli and S. aureus.Scale bar: 25 μm.D) SEM images of MDR E. coli and S. aureus.Scale bar: 1 μm.*P < 0.05, ** P < 0.01, ## P < 0.01 vs 3 μg mL À1 group, NS P > 0.05.

Figure 4 .
Figure 4. Roles of GSH and ROS in the antibacterial performance of BM-NDs.A) The intracellular GSH content of MDR E. coli and S. aureus treated with BM-NDs under high-glucose conditions (n = 3).B) The intracellular ROS level and C) fluorescence staining of MDR E. coli and S. aureus using DCFH-DA as the probe (n = 3).Scale bar: 25 μm.D) Photographs of bacterial colonies and E) viability analysis of MDR E. coli and S. aureus treated with BM-NDs along with GSH, catalase, or isopropanol (n = 3).F) Live/dead fluorescence staining images of MDR E. coli and S. aureus.Scale bar: 25 μm.G) SEM images of MDR E. coli and S. aureus.Scale bar: 1 μm.*P < 0.05, ** P < 0.01, # P < 0.05 vs 3 μg mL À1 group, NS P > 0.05.

Figure 5 .
Figure 5.Effect of BM-NDs on RAW 264.7 macrophages.A) Cell viability of macrophages treated with different concentrations of BM-NDs under different glucose conditions (n = 3).B) The fluorescence staining and C) intracellular level evaluation of ROS in macrophages using DCFH-DA as the probe.Scale bar: 100 μm (n = 3).D) Scatter diagram and E) histogram overlay of flow cytometry demonstrating the polarization state of macrophages.F) Fluorescence staining images of M1 macrophages using iNOS as the specific marker.Scale bar: 100 μm.G) Fluorescence staining images of M2 macrophages using Arg-1 as the specific marker.Scale bar: 100 μm.*P < 0.05, ** P < 0.01, NS P > 0.05.

Figure 6 .
Figure 6.Antibacterial performance and proangiogenic potential of macrophages treated with BM-NDs.A) Fluorescence imaging and B) quantitative analysis of intracellular dead bacteria with PI staining (n = 3).Scale bar: 50 μm.C) TEM characterization of intracellular dead MDR bacteria in macrophages treated with BM-NDs.Scale bar: 2 μm.D) Fluorescence imaging and E) quantitative analysis of endothelial tube formation of endothelial cells incubated with conditioned medium from BM-NDs-treated macrophages (n = 3).Scale bar: 200 μm.F) Schematic illustration of endothelial cell migration in a transwell system.G) Quantitative analysis of endothelial cells migration treated with conditioned medium from BM-NDs-treated macrophages (n = 3).*P < 0.05, ** P < 0.01, NS P > 0.05.

Figure 7 .
Figure 7. GelMA@BM treatment for diabetic infectious wound healing.A) Representative photographs, B) schematic diagram, and C) quantitative analysis of the wound area treated with PBS, blank hydrogel, vancomycin-loaded hydrogel, and BM-NDs-integrated hydrogel (n = 9).Scale bar: 2 mm.D) Photographs of bacterial colonies and E) viability analysis of bacteria from tissues of different treatment groups (n = 3).F) Local glucose concentrations in tissues of different treatment groups (n = 3).G) H&E and Masson staining of wound tissues from different treatment groups.Scale bar: 1 mm.*P < 0.05, ** P < 0.01, NS P > 0.05.

Figure 8 .
Figure8.ROS level, macrophage polarization, and angiogenesis within diabetic infectious wound treated with GelMA@BM hydrogel.A) Fluorescence staining and B) quantitative analysis of ROS in tissues from different treatment groups determined at 2 days posttreatment (n = 3).Scale bar: 200 μm.C) Fluorescence staining and D) quantitative analysis of M1 macrophages using iNOS and CD68 as the specific markers at 2 days posttreatment (n = 3).Scale bar: 100 μm.E) Fluorescence staining and F) quantitative analysis of ROS in tissues from different treatment groups determined at 8 days posttreatment (n = 3).Scale bar: 200 μm.G) Fluorescence staining and H) quantitative analysis of M2 macrophages using Arg-1 and CD68 as the specific markers at 8 days posttreatment (n = 3).Scale bar: 100 μm.I) Fluorescence staining and J) quantitative analysis of angiogenesis using CD31 and α-SMA as the specific markers at 8 days posttreatment (n = 3).Scale bar: 100 μm.*P < 0.05, ** P < 0.01, NS P > 0.05.