Ligand-dependent activity engineering of Cu-MOFs based on biomimetic strategies for nanozyme-decorated smart hydrogels for therapy of inflammatory bone defects

Abstract

Reconstructing inflammatory bone defects using stem cell-based therapy faces challenges posed by the oxidative stress microenvironment. Herewith, inspired by the response mechanism of the intracellular antioxidant defense systems (IADS), we propose MOF nanozymes that mimic the structure of horseshoe crab hemocyanin and find that a ligand engineering strategy can modulate the SOD- and CAT-mimicking activities of MOF-based nanozymes. Through in vitro screening, the most efficient Cu-Imi nanozymes were selected and incorporated into a smart hydrogel (composed of oxidized dextran and dopamine-functionalized gelatin) for inflammatory bone defect treatment. Studies demonstrate that smart hydrogels are able to respond to the acidic microenvironment induced by oxidative stress and early tissue injury by intelligently releasing biomimetic nanozymes, which can effectively eliminate bacteria, alleviate inflammation, and reduce ROS levels. In addition, the smart hydrogel has been demonstrated to effectively sustain the viability of stem cells and osteogenic differentiation in environments where there is an elevated level of ROS, and has the potential to promote angiogenesis and modulate the inflammatory microenvironment during the process of rat bone tissue regeneration. It is hypothesised that this smart hydrogel-like material decorated with antioxidant enzymes offers a promising avenue for treating various inflammation-associated disorders, including arthritis, chronic wounds and bone fractures.

---

1. Introduction

Severe bone defects caused by accidental trauma, bacterial infection, and malignant tumour resection usually cannot be repaired on their own during a patient's lifetime.¹ Bone grafts have been widely used clinically to enhance bone regeneration and repair, and they hold promising prospects in treating severe bone defects, driven by the in-depth research of novel biomaterials that are easy to manufacture and widely applicable. Tissue regeneration is a dynamic process that involves two-way interactions between cells and their surrounding matrix,² requiring the design of biomaterials that can adapt to and respond to changes in the microenvironment with positive feedback. Bone defects resulting from trauma are often accompanied by local microvascular rupture and prolonged inflammation, causing a sustained increase in reactive oxygen species (ROS) and limitations in stem cell function.³ Current approaches employing biomineralized hydrogels also encounter some key challenges in regulating redox homeostasis and anti-inflammatory responses, thus significantly limiting the applicability of biomineralized hydrogels in the regeneration of inflammatory bone defects.⁴ The current strategic challenges in the repair of inflammatory bone defects^5,6 have led to an increasing interest in the development of smart hydrogels that remodel the oxidative stress microenvironment.⁷

Tissue regeneration in bone defects occurs primarily through intramembranous osteogenesis, in which endogenous stem cells differentiate into osteoblasts and subsequently form complex skeletal structures.⁸ However, the therapeutic efficacy of endogenous stem cell infiltration is often hampered by accumulated reactive oxygen species (ROS)⁹ and inadequate oxygen levels.¹⁰ Particularly in inflammatory bone defects, elevated ROS levels exacerbate local tissue damage and cause chronic inflammation.^9,11 Therefore, removing excessive ROS to create a favorable microenvironment for endogenous stem cells has become a promising strategy to enhance bone regeneration in the management of inflammatory bone defects.¹² Among the types of ROS, superoxide and hydrogen peroxide (H₂O₂) play vital roles.¹³ For instance, H₂O₂ not only has a long half-life but also easily diffuses across lipid membranes and serves as a precursor to the highly destructive hydroxyl radicals. Therefore, regulating intracellular ROS levels, particularly H₂O₂ and superoxide, below their toxic threshold is essential for protecting cells from oxidative damage.¹⁴ In biological systems, several antioxidant enzymes, including superoxide dismutase (SOD) and catalase (CAT),¹⁵ are capable of converting superoxide and H₂O₂ into oxygen (O₂) and water (H₂O), thereby limiting harmful effects while providing oxygen to the damaged tissue microenvironment.¹⁶ Although SOD and CAT play critical roles in maintaining cellular redox balance, they share common limitations inherent in most natural enzymes, such as low stability and poor availability, which restrict their practical biomedical applications.¹⁷ To overcome these limitations, significant efforts have been devoted to developing SOD and CAT mimetics. Among these mimetics, nanozymes – nanomaterials with enzyme-like activity – have attracted widespread attention in recent years due to their unique advantages, such as low cost,¹⁸ large surface area,¹⁹ and robustness under extreme conditions.²⁰

MOFs are a class of porous materials constructed from metal nodes and organic ligands.²¹ Due to their highly tunable structures and large surface areas, MOFs have found widespread applications in gas adsorption,²² chemical sensing²³ and heterogeneous catalysis.²⁴ Recently, biomimetic MOFs have garnered significant attention due to their promising potential in mimicking natural enzymes,¹⁴ such as glutathione peroxidase (GPx)²⁵ and catalase (CAT).²⁶ However, despite considerable progress, there are no reports on MOFs specifically designed to mimic SOD and CAT, nor have they been explored for use in inflammatory bone defect therapy. Inspired by the metal–ligand coordination in hemocyanin from horseshoe crab blood, we hypothesize that the shared metal–ligand coordination between natural metal enzymes and MOFs could enable the rational tuning of MOF enzyme-mimetic activity through ligand engineering strategies. Specifically, we present the first demonstration of the controlled regulation of SOD and CAT mimetic activities in copper-based MOFs (Cu-MOFs), including Cu-Im, Cu-mIm, and Cu-Imi, by varying the number of methyl substitutions in the imidazole ligands. These substitutions alter the electronic properties of the ligands, thereby modulating the enzyme-like activities of the MOFs.

Treatment of inflammatory bone defects with a high oxidative stress microenvironment based on bionic mineralized hydrogels poses significant challenges. To address this, we selected the most active Cu-Imi nanozyme and incorporated it into a smart hydrogel system for inflammatory bone defect therapy. This hydrogel is responsive to the low pH environment typically encountered during early stages of tissue injury. It features a dual-network structure composed of oxidized dextran, dopamine-functionalized gelatin and borax, which facilitates the intelligent release of the nanozyme (an overview is provided in Scheme 1). In vitro studies demonstrated that the Gel-DA/ODex^0.5 smart hydrogels exhibited remarkable antibacterial activity, reactive oxygen species (ROS) clearance, immune modulation, as well as the ability to promote angiogenesis and osteogenic differentiation., MOF-based nanozymes and smart hydrogels constructed in this study based on bionic and intelligent strategies not only provide a guiding strategy for the design of MOF-based nanozymes but also show promise as a new class of inflammatory bone repair materials that can provide an effective and intelligent solution for bone defect therapy.

Scheme 1 Hemocyanin-mimicking Cu-MOF nanozyme synthesis, smart hydrogel preparation and inflammatory bone defect therapy: (A) schematic of the synthesis of rationally designed SOD- and CAT-mimicking Cu-MOF nanozymes for inflammatory bone defect therapy; (B) ligand-dependent adsorption energies of MOF-based nanozymes; (C) preparation of the smart hydrogel; and (D) multiple biological functions of smart hydrogels including ROS scavenging protecting endogenous stem cells, immunomodulation and antimicrobial as well as applications and mechanisms in the therapy of inflammatory bone defects.

2. Materials and methods

2.1. Materials

Imidazole (Im), 2-methylimidazole (mIm), 1,2-dimethylimidazole (Imi), Cu (NO₃)₂·3H₂O, lipopolysaccharide (LPS), gelatin (CAS 9000-70-8, gel strength 240 g Bloom, pharmaceutical grade) and dextran (CAS 9004-54-0, MW 20 000) were purchased from Adamas (China). Dopamine hydrochloride (DA) was purchased from Aladdin. Analytical reagents NaOH, ethylene glycol, hydrogen peroxide, sodium periodate and methanol were from Kelong Chemical Reagent Factory (Chengdu, China). DDPH (didodecyldimethylammonium bromide), NBT (nitroblue tetrazolium), riboflavin and methionine were obtained from McLean Biochemistry Technology Co. (Shanghai, China).

2.1.1. Synthesis of Cu-Im, Cu-mIm and Cu-Imi MOF-based nanozymes. First, 4 mmol of imidazole was dissolved in 30 mL of 1 : 1 methanol : water solution and then 30 mL of 1 mmol Cu (NO₃)₂·3H₂O (prepared in methanol : water = 1 : 1) was added. After that, 0.1 M NaOH solution (prepared in methanol : water = 1 : 1) was added and the pH of the solution was adjusted to 8. The reaction mixture was transferred to a flask and stirred at 37 °C for 6 h. Finally, the reaction products were collected and separated by centrifugation, and the precipitates were washed several times with the 1 : 1 methanol : water solution, and then dried under vacuum to obtain the Cu-Im MOF-based nanozymes. The Cu-mIm and Cu-Imi MOF-based nanozymes were prepared using 2-methylimidazole and 1,2-dimethylimidazole, respectively, and did not require pH adjustment.

2.1.2. Synthesis of ODex. The preparation of ODex followed previously reported methods.²⁷ Briefly, 5.0 g of dextran was dissolved in 50 mL of deionized water, to which 4.68 g of NaIO₄ was added. The mixture was stirred at room temperature in the dark for 5 h. The reaction was quenched by adding 10 mL of ethylene glycol and allowing it to proceed for 2 h. The resulting solution was dialyzed against deionized water (molecular weight cutoff: 3500) for 3 days, after which the product was lyophilized to yield ODex.

2.1.3. Synthesis of Gel-DA. Gel-DA was synthesized according to previously reported methods.²⁸ Specifically, gelatin (5 g) was dissolved in 400 mL of degassed MES buffer (50 mM, pH 4.5) at 37 °C. EDC (600 mg) and NHS (450 mg) were then added to the solution. Separately, dopamine hydrochloride (300 mg) was dissolved in 5 mL of MES buffer (50 mM, pH 3.3) and subsequently added to the reaction mixture under nitrogen protection, stirring at 600 rpm for 24 hours at 37 °C in the dark. The resulting solution was purified by dialysis (MWCO 12 000 Da) against deionized water for 3 days. Finally, Gel-DA was lyophilized and stored at −20 °C.

2.1.4. Synthesis of smart hydrogels. To prepare blank hydrogels (Gel-DA/ODex), 15 wv% Gel-DA was stirred in 0.03 M borax solution, while 15 wv% ODex was dissolved in PBS at 37 °C for 0.5 h. The two solutions were then combined, poured into molds, and left to set for 3 h to form the gel. Briefly, the smart hydrogel Gel-DA/ODex^X (the article also simplifies it to GOD^X) was prepared by mixing 15 wv% Gel-DA with 15 wv% ODex and varying amounts of Cu-Imi MOF-based nanozymes (X = 0.1% or 0.3% or 0.5%).

2.2. Characterization

The morphology of the cross-section of various hydrogel samples was characterized using a scanning electron microscope (SEM, JSM-5900LV, Japan). Elemental distribution analysis of the hydrogel sample was performed via energy dispersive X-ray spectroscopy (EDS, Octane Elect Super, USA). X-ray photoelectron spectroscopy (XPS, Thermo Fisher K-ALPHA, USA) was employed to analyze the surface elements of Cu-Im, Cu-mIm and Cu-Imi. The crystalline structure of these samples (Cu-Im, Cu-mIm and Cu-Imi) was confirmed by X-ray diffraction (XRD, Rigaku, Japan). Nuclear magnetic resonance (NMR) spectroscopy (Bruker AV III HD 400 MHz, Germany) was used to characterize the synthesis of ODex and Gel-DA. The UV-visible (UV-vis) spectra of Cu-Im, Cu-mIm and Cu-Imi were obtained using a UV-2600 spectrophotometer (Shimadzu, Japan). FTIR spectra of Cu-Im, Cu-mIm, Cu-Imi, ODex, Gel-DA and the hydrogel were collected on a Nicolet 6700 FTIR spectrometer (Thermo Nicolet, Madison, WI). Thermogravimetric analysis (TGA) was conducted on samples (Cu-Im, Cu-mIm and Cu-Imi) using a thermogravimetric analyzer (METTLER TOLEDO, Switzerland). The specific surface area and pore size distribution of the samples were determined using an automated gas adsorption analyzer (Micromeritics ASAP 2460, USA).

2.3. Calculations

Density functional theory (DFT) calculations were conducted using the DMol 3 module, applying the Perdew–Burke–Ernzerhof (PBE)²⁹ exchange–correlation functional within the generalized gradient approximation (GGA). Before proceeding with subsequent computations, full structural optimization was performed for all geometries, ensuring convergence criteria of 1 × 10⁻⁵ Ha for energy, a maximum force of 0.002 Ha Å⁻¹, a maximum displacement of 0.005 Å and a step size not exceeding 0.3 Å. The structural unit of the MOFs (Cu-Im, Cu-mIm and Cu-Imi) was isolated as a cluster molecule for simulation purposes. Adsorption energy (E_ads) was calculated using the relation E_ads = E_ad/sub − E_ad − E_sub, where E_ad/sub, E_ad and E_sub represent the adsorption energies of the adsorbate/substrate complex, the adsorbate/MOFs and the adsorbate on a clean substrate, respectively.

2.4. Scavenging free radicals of MOF-based nanozymes in vitro

2.4.1. ˙O₂⁻ scavenging activity. The efficiency of ˙O₂⁻ scavenging was determined using nitro blue tetrazolium chloride (NBT) as a probe. The solution produced by the reaction between NBT and ˙O₂⁻ shows a characteristic peak at 560 nm due to the production of a blue color substance (formazan). The assay solution typically contains riboflavin (6.67 μM), methionine (4.33 mM) and NBT (25 μM), as well as various concentrations of Cu-Im, Cu-mIm or Cu-Imi samples in a PBS buffer solution (pH 7.35, 0.01 M). Then, the mixture was irradiated with a 30 W UV lamp for 5 min. Finally, the absorbance of formazan was determined by using a UV-vis spectrophotometer. The ˙O₂⁻ scavenging efficiency was calculated using the formula [(A₁ − A₂)/(A₁ − A₀)] × 100%, where A₀ is the absorbance of a single NBT, and A₁ and A₂ are the absorbance of formazan in the absence and presence of samples (Cu-Im or Cu-mIm or Cu-Imi), respectively.

2.4.2. H₂O₂ scavenging ability. Our study also tested the scavenging ability of H₂O₂. Briefly, 10 mL of H₂O₂ solution (100 μM) was prepared, to which different concentrations of samples (Cu-Im or Cu-mIm or Cu-Imi) were added (2 h). Then, the supernatant of each group was mixed with 2 mL of Ti(SO₄)₂ solution (obtained from the mixture of 1.33 mL of 5 wv% Ti(SO₄)₂ and 16.66 mL of 2 M H₂SO₄ in 100 mL of deionized water) for 30 min, and the supernatant was analyzed using a UV-visible spectrophotometer to determine its absorbance at 405 nm. H₂O₂ scavenging was calculated by using the following equation: H₂O₂ scavenging = [(A₀ − A₁)/A₀] × 100%, where A₀ represents the absorbance of the mixture without samples (Cu-Im or Cu-mIm or Cu-Imi, control sample) and A₁ is the absorbance of the sample in the experimental group.

2.4.3. DPPH free radical scavenging activity. Different concentrations of samples (Cu-Im or Cu-mIm or Cu-Imi) were added to DPPH ethanol solution (0.1 mM, 10 mL) and then incubated for 40 min in the dark at 37 °C. The absorbance of solution at 517 nm was recorded using a UV-visible spectrophotometer, and the DPPH scavenging ability was calculated using the following formula: DPPH scavenging rate = [(D₀ − D₁)/D₀] × 100%, where D₀ is the absorbance of the experimental group and D₁ is the absorbance of DPPH.

2.4.4. SOD-like activity of Cu-Imi. The activity of Cu-Imi was assessed using a superoxide dismutase assay kit (WST-1, Solarbio, Beijing, China). The SOD activity of Cu-Imi at different concentrations was expressed as the inhibition rate of the WST-1 reaction, according to the protocol provided by the manufacturer.

2.5. Electrochemical measurement

The redox peaks of the samples (Cu-Im, Cu-mIm or Cu-Imi) were evaluated using electrochemical methods with a standard three-electrode system. The system consisted of an Ag/Ag⁺ reference electrode, a platinum wire counter electrode and a glassy carbon electrode (GCE) as the working electrode. Prior to modification, the bare GCE was mechanically polished to a mirror finish with 0.03 μm alumina slurry, followed by thorough rinsing with anhydrous ethanol and double-distilled water. A 3 mg sample of Cu-Im, Cu-mIm or Cu-Imi was dispersed in a 1 mL solution of methanol and water (1 : 1) via sonication for 60 minutes. Then, 5 μL of this suspension was deposited onto the GCE surface and dried under infrared light to create the modified electrode. After purging the electrolyte solution with N₂ for 3 minutes, cyclic voltammetry was performed at a scan rate of 10 mV s⁻¹ to record the redox behavior of the samples (Cu-Im, Cu-mIm or Cu-Imi).

2.6. Swelling test

The hydrogels were initially weighed (W₀) and completely immersed in PBS and then they were taken out at different times (0.5 h, 1 h, 3 h, 6 h, 12 h, 24 h, 36 h, 48 h and 60 h) at 37 °C. Water was gently absorbed from the surface of hydrogel with filter paper. After that, they were weighed again (W_t). The swelling rate calculated as follows: swelling rate (%) = (W_t − W₀)/W₀ × 100%.

2.7. Mechanical properties test

Compression testing was performed using an electronic universal testing machine (Instron 5967, USA). Hydrogels, with a height of 20 mm and a diameter of 12 mm, were prepared and subjected to a compression test with a maximum strain limit of 60% of the original height at a rate of 5 mm min⁻¹.

2.8. In vitro enzyme-responsive release

The method for studying hydrogel substance release was optimized based on previous research.³⁰ In brief, the specific absorption peak of Cu-Imi (350 nm) was determined via UV spectroscopy. The absorbance of Cu-Imi at 350 nm was measured across a range of concentrations, and a standard curve was constructed through spectrophotometric analysis. Hydrogel samples of Gel-DA/ODex and Gel-DA/ODex^0.5 (0.2 g) were placed in centrifuge tubes containing 20 mL of PBS, with the pH adjusted to 5.5 using 0.01 M HCl to simulate the microenvironment typical of early tissue injury (e.g., broken bones, skin wounds, torn muscles, etc.). At various time points, equal volumes (2 mL) of the PBS solution were collected, and the absorbance at 350 nm was measured using spectrophotometry. Finally, the release amount at each time point was quantified by reading the corresponding values on the standard curve.

2.9. Antibacterial activity test

The inhibitory effects of samples (Cu-Im or Cu-mIm or Cu-Imi) on E. coli (typical Gram-negative bacteria) and S. aureus (typical Gram-positive bacteria) were tested by the disc diffusion method. First, 120 μL of bacterial suspension (10⁶ CFU mL⁻¹) was inoculated on agar plates. Then, the samples (Cu-Im or Cu-mIm or Cu-Imi, diameter: 9 mm) were placed on the surface of agar plates. After incubation of 24 h at 37 °C, the zone of inhibition was recorded. In addition, the inhibitory activity of the hydrogels was evaluated by the plate counting method. First, UV-sterilized hydrogel samples (5 × 5 × 10 mm³) were immersed in a plate containing 1000 μL of bacterial suspension (10⁶ CFU mL⁻¹), followed by gentle shaking (180 rpm) for 4 h at 37 °C. SEM was used to examine the morphology of bacteria (bacteria were fixed with 5% glutaraldehyde and dehydrated with ethanol for observation) incubated with the hydrogel to verify the antibacterial effects of the hydrogel. Subsequently, 120 μL of the co-culture solution was taken and spread on an agar plate, and photographed after 24 hours.

2.10. In vitro cell experiment

2.10.1. Cell culture. The culture medium consisted of 89% basal medium, 10% fetal bovine serum and 1% penicillin–streptomycin (all from Gibco, USA). Cells were maintained in an incubator at 37 °C with 5% CO₂ and 95% relative humidity. The medium was refreshed every 2 days, with PBS washes performed before each replacement. Human umbilical vein endothelial cells (HUVECs), RAW264.7 cells and rat bone marrow mesenchymal stem cells (rBMSCs) were sourced from West China Hospital, Sichuan University.

2.10.2. In vitro rBMSC and HUVEC proliferation assay. All specimens were sterilized by γ-rays. Sterilized hydrogel discs (1 mm in diameter, 1 mm in thickness) were placed in 48-well culture plates, followed by seeding rBMSCs at a density of 1 × 10⁴ cells per well. Cell viability on the hydrogel surface at various time points was assessed using the CCK-8 assay. The basic steps of the cell compatibility assay for HUEVCs were similar to those for rBMSCs. The percentage of relative growth rate (RGR) of cells was used to express the cytocompatibility of samples, and was calculated as follows:

RGR (%) = OD₁/OD₀ × 100%

where OD₁ is the optical density of cell cultures with different smart hydrogels and OD₀ is the optical density of cells cultured in medium without the hydrogel.

Sterilized hydrogels (diameter, 2 mm; thickness, 2 mm) were co-cultured with rBMSCs or HUVECs inoculated into 24-well plates. On days 1 and 3, after the addition of 10 μg mL⁻¹ fluorescein diacetyl (FDA), the cells were incubated for 30 min and photographed under an inverted fluorescence microscope (Nikon Corporation, Japan). The HUVECs were stimulated using the ROS kit drug (Beyotime Biotechnology, China) and then co-cultured with hydrogels (1 mm in diameter, 1 mm in thickness). The culture medium from HUVECs on day 3 was collected, and the VEGF concentration was measured using an enzyme-linked immunosorbent assay (ELISA) kit (Beyotime, China).

2.10.3. Effect of hydrogels on RAW 264.7. Experiments were performed according to the method provided in the ELISA kit from Biyuntian Biologicals, China. Briefly, RAW 264.7 macrophages (1 × 10⁵ cells per mL) were first pretreated with lipopolysaccharide (LPS, 1 μg mL⁻¹) for 2 h and then co-incubated with different hydrogels (diameter, 2 mm; thickness, 2 mm) for 72 h in 24-well culture plates. The total amount of IL-1β, TNF-α, Arg-1 and IL-10 was detected using an enzyme-linked immunosorbent assay (ELISA) kit.

2.10.4. Cell morphology observation. Cells were stimulated to mimic the oxidative stress microenvironment using the ROS kit drug (Beyotime Biotechnology, China), and the cytoskeleton of rBMSCs, HUVECs, and RAW 264.7 was stained for cell morphology after 3 days of co-culture with hydrogels (diameter, 1 mm; thickness, 1 mm). These cells were fixed with 4% paraformaldehyde for 3 h. Subsequently, the cells were stained with DAPI for nuclear staining and with phalloidin for cytoplasmic staining. The fluorescence images of cells were obtained by using a confocal laser scanning microscope (CLSM, NSTORM and A1, Nikon, equipped with NIS-Elements Viewer)

2.10.5. In vitro scratch assay. The rBMSCs were seeded and cultured in a 24-well plate at a density of 3 × 10⁶ cells per mL for 24 h, cells were stimulated to mimic the oxidative stress microenvironment using the ROS kit drug (Beyotime Biotechnology, China), and then a sterile 200 μL pipet tip was used to carefully scrape fine linear scratches on confluent cells. After washing thoroughly with PBS to remove residual cellular debris, the hydrogel samples (diameter, 2 mm; thickness, 2 mm) and medium were added to each well of the 24-well plate. After culturing for 24 h, the hydrogel samples were taken out, the cells in the 24-well plate were washed three times with PBS, and the cell movement was observed under a microscope. The basic steps of the HUVEC cell scratch assay were similar to those of the rBMSC cell scratch assay.

2.10.6. Intracellular ROS scavenging activity. Intracellular ROS production was measured using a ROS kit (Beyotime Biotechnology, China) according to the manufacturer's instructions. Briefly, H₂O₂ acts as an exogenous oxidative stress inducer for macrophage RAW 264.7. After 12 h of co-culture with the hydrogel, the culture medium was replaced with 2,7-dichlorodihydrofluorescein diacetate (DCFH-DA, 10 μM), which can be oxidized to fluorescence by ROS, and then co-cultured with the cells for 25 min. Cells were observed with an inverted fluorescence microscope, and all images were taken under the same conditions, and the fluorescence was quantified with Image J software. Meanwhile, the effect of ROS on the viability of Raw 264.7 cells was detected using the CCK-8 kit.

2.10.7. Osteoimmunomodulatory activity of smart hydrogels. To investigate the bone immune modulation ability of the smart hydrogel, we seeded rBMSCs at a density of 1 × 10⁵ cells per well in a 12-well plate. After 24 hours of culture, cells were stimulated to mimic the oxidative stress microenvironment using the ROS kit drug (Beyotime Biotechnology, China), and the medium was replaced with a 1 : 2 mixture of the conditioned medium obtained in Section 2.10.3 and osteogenic medium (containing 10 mM beta-glycerophosphate and 50 mg mL⁻¹ ascorbic acid) for osteogenesis studies. The bone immune modulation activity of the samples was confirmed by ALP staining and ARS staining.

After 14 days of co-culture and induction, the cells were fixed by 4% paraformaldehyde for 30 min. Then the ALP activity was detected by ALP staining using an ALP staining kit (Beyotime Biotechnology, China). After incubating rBMSCs with the ALP chromogenic agent for 30 minutes under dark conditions, the results were examined under a microscope.

The calcium nodules generated by the cells were stained by an ARS kit. After 14 and 21 days of co-culture, the cells were washed with PBS and fixed with 4% paraformaldehyde for 30 min, followed by ARS staining for 15 min; the calcium nodules were observed by using a microscope. To ascertain the value of ARS staining, treated cells were co-incubated with 10% (v/v) acetic acid for 12 hours. The resulting coculture supernatant was then collected and subjected to centrifugation to remove solid particles, after which 10% ammonium hydroxide was added. The absorbance of the supernatant was measured at 405 nm. Furthermore, the levels of osteogenesis related cytokines were determined using an enzyme-linked immunosorbent assay (ELISA, Beyotime Biotechnology, China) following a 14-day incubation period.

2.11. In vivo study

All animal experiments were ethically approved by the Animal Ethics Committee of Sichuan University (No. SCU44-2502-03) and conducted in accordance with the National Research Council Guidelines for the Care and Use of Laboratory Animals. The animals in this study were Sprague Dawley (SD) rats (male, 6–8 weeks, 250–300 g).

2.11.1. Inflammatory rat femoral defect model and implantation of materials. Before surgery, rats were housed in separate cages for 1 week to allow them to adapt to the environment. Anesthesia was performed by inhalation of isoflurane. An inflammatory bone defect model was also established by injecting lipopolysaccharide (LPS) dissolved in saline (1 mg mL⁻¹, 0.03 mL) in rats. Eighteen rats were randomly divided into 3 groups (6 rats in each group): control ROS up group, Gel-DA/ODex ROS up group and Gel-DA/ODex^0.5 ROS up group. A bone defect of 1 mm in diameter and 2 mm in depth was then created on the femoral condyles of the rats by a minimally invasive method. After modeling the bone defect, different hydrogel samples were carefully implanted in the defects. The animals without implants in the bone defect were used as a control group to assess natural self-repair. Bone wax was applied to seal the bone defect with a hydrogel sample implanted, and finally, the subcutaneous tissue and skin were sutured. After 4 and 8 weeks, the rats were euthanized and dissected separately. Then, they were fixed/preserved in a 4% paraformaldehyde solution for follow-up experiments.

2.11.2. Micro-CT analysis. Paraformaldehyde-fixed samples were studied by microtomographic (scanco viva CT80) observations to analyze the repair of bone tissue defects.

2.11.3. Histological analysis. Samples were decalcified, embedded, sectioned and stained with hematoxylin and eosin (H&E) staining and Masson's trichrome stain for histological evaluation.

2.11.4. Immunohistochemical and immunofluorescent staining. Angiogenesis, an early stage of tissue regeneration, primarily occurs approximately four weeks following injury. Four weeks after implantation, immunohistochemistry was employed to assess the levels of CD31 within the samples for analyzing the pro-angiogenic effect of the smart hydrogel, while immunofluorescent staining was utilized to determine TNF-α and CD206 expression for analyzing the bone immune modulation capabilities of the smart hydrogel.

2.12. Statistical analysis

All quantitative results are shown as mean ± standard deviation (SD); P < 0.05 was considered as statistically significant (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). Data are mean values ± sd (n ≥ 3). GraphPad Prism 8.0 software was used to perform the Student's t-test in two groups and one-way ANOVA with a Bonferroni-corrected t-test in more than two groups.

3. Results and discussion

3.1. Preparation and characterization of Cu-MOF nanozymes

Hemocyanin, the oxygen-transport protein in the blood of arthropods³¹ such as horseshoe crabs, exhibits diverse catalytic functions, including polyphenol oxidase, catalase and antibacterial activities under specific conditions. The hemocyanin molecule consists of a coordination combination of Cu ions and nitrogen on the peptide chain of an amino acid.³² Inspired by hemocyanin in horseshoe crab blood, we developed biomimetic hemocyanin-mimic MOF-based nanozymes with a Cu–N coordination structure. These nanozymes (Cu-Im, Cu-mIm, and Cu-Imi), formed by coordinating copper atoms with nitrogen from imidazole ligands (Im, mIm, and Imi), exhibit superoxide dismutase and catalase activities.

The hemocyanin-mimic Cu-MOF (Cu-Im, Cu-mIm and Cu-Imi) nanozymes were synthesized via a hydrothermal method, specifically, by the coordination of copper metal nodes and imidazole-like ligands (Im, mIm and Imi). Various characterization techniques were performed for Cu-MOFs (Cu-Im, Cu-mIm and Cu-Imi). The microscopic morphology of Cu-Im is a nanoflower as in Fig. 1A, Cu-mIm is a 2D sheet (Fig. 1B), and Cu-Imi possesses a three-dimensional structure that is composed of stacks (Fig. 1C). To identify the factors causing the morphological differences among the three MOFs, the secondary structural units of the MOFs were simulated. Cu-Im (Fig. 1D), Cu-mIm (Fig. 1E) and Cu-Imi (Fig. 1F) are composed of a single metal center coordinated with four nitrogen atoms from imidazole-based ligands (Im, mIm and Imi). Differences in these secondary structural units may result in variations in coordination modes and coordination kinetics between the metal ions and ligands, ultimately affecting the morphology of the MOFs. The structure of Cu-MOFs was further characterized using X-ray diffraction (XRD) analysis. As shown in Fig. 1G, the XRD patterns of Cu-MOFs display multiple sharp diffraction peaks, indicating the formation of a well-defined crystalline structure. The minor differences in peak positions suggest similar coordination forms among the MOFs. The leftward shift of the peak positions of Cu-mIm and Cu-Imi compared to Cu-Im may be attributed to the different crystalline morphology (larger crystal face spacing).³³ The thermogravimetric analysis (TGA; SI, Fig. S1A) indicated that the stability of all the Cu-MOFs (Cu-Im, Cu-mIm and Cu-Imi) could be guaranteed to at least 199.8 °C; the thermal stability gradually increases with the number of methyl groups. Moreover, the UV/Vis absorption spectra of Cu-MOFs (Cu-Im, Cu-mIm and Cu-Imi) were also recorded and depicted in Fig. S1B. In agreement with previous studies,³⁴ the specific peaks of Cu-MOFs around about 205 and 260 nm are mainly caused by C=N on the imidazole ring.

Fig. 1 Structural characteristics of MOF nanozymes: SEM images show (A) Cu-Im, (B) Cu-mIm and (C) Cu-Imi; simulated structures of possible secondary building units include (D) Cu-Im, (E) Cu-mIm and (F) Cu-Imi; the analysis includes (G) XRD patterns, (H) Fourier-transform infrared (FTIR) spectra, (I) XPS survey spectra, and high-resolution XPS spectra of Cu 2p for (J) Cu-Im, (K) Cu-mIm and (L) Cu-Imi.

In order to directly demonstrate the coordination between the imidazole ligand (Im, mIm and Imi) and Cu²⁺, the chemical structure of Cu-MOFs was characterized by using Fourier transform infrared (FTIR) spectral analysis and X-ray photoelectron spectroscopy (XPS). As shown in Fig. 1H, the broad peak around 3437 cm⁻¹ is attributed to –OH stretching vibration (H₂O contained in nanozymes). The peaks at 880 cm⁻¹ and 1346 cm⁻¹ are assigned to the in-plane bending of the imidazole ring.³⁵ Absorption bands at 777 cm⁻¹ and 679 cm⁻¹ correspond to the out-of-plane bending of the imidazole ring, while the peak at 1627 cm⁻¹ is attributed to the stretching vibration of the C = N bond.³⁶ Significantly, the absorption band appearing at 432 cm⁻¹ is attributed to the stretching vibration of the Cu–N bond in the Cu-MOFs.³⁷ XPS analysis confirmed the presence of the elements Cu, C and N in all Cu-MOFs (Fig. 1I). As shown in Fig. 1J, the Cu 2p spectrum exhibits a high binding energy peak at 935.1 eV, which is attributed to Cu²⁺, while the peak at 933.7 eV is assigned to Cu⁺. Additionally, a characteristic Cu²⁺ satellite peak is observed at 955.2 eV.³⁸ Overall, Cu-Im contains both Cu⁺ and Cu²⁺ states, consistent with the oxidation states of copper in hemocyanin found in horseshoe crabs. The chemical composition of Cu-mIm (Fig. 1K) and Cu-Imi (Fig. 1L) was further characterized using XPS, which revealed strong characteristic peaks of Cu 2p in both Cu-mIm and Cu-Imi. As expected, they both exhibit typical high-resolution spectra for elements such as C and N. Beyond structural consistency with the oxidation states of Cu in hemocyanin, the presence of both Cu²⁺ and Cu⁺ in these materials may also contribute to enhanced nanozyme activity.³⁹ Based on the XPS analysis results shown in Table S1, the proportions of Cu(I) and Cu(II) in Cu-Im are 2.88% and 97.12%, respectively; the proportions of Cu(I) and Cu(II) in Cu-mIm are respectively 7.69% and 92.31%; the proportions of Cu(I) and Cu(II) in Cu-Imi are 16.19% and 73.81%, respectively. It can be observed that the ratio of Cu(I) to Cu(II) in Cu-Imi is more balanced, suggesting its potentially superior catalytic activity.

3.2. SOD-mimic, CAT-mimic and DPPH elimination activity assays

Reactive oxygen species (ROS), such as superoxide radicals (˙O₂⁻) and hydrogen peroxide (H₂O₂), have been demonstrated to play a role in a variety of intracellular functions.⁴⁰ Tissue injury typically results in cellular redox imbalance, leading to elevated ROS levels, which cause oxidative damage to living cells and significantly impair tissue regeneration. After successfully synthesizing the Cu-MOF nanozymes, we investigated their SOD-mimicking catalytic activity. Briefly, ˙O₂⁻ is generated via the reaction between methionine and riboflavin, and NBT is used as a probe, which reacts with ˙O₂⁻ to form a purple compound (Fig. S2B); the absorbance change at 560 nm is measured to determine the SOD-mimicking catalytic activity of Cu-MOF nanozymes. As illustrated in Fig. 2A, an increase in the concentration of Cu-MOFs resulted in enhanced ˙O₂⁻ scavenging capacity, indicating a concentration-dependent scavenging effect. Additionally, the SOD-mimicking catalytic activity at the same concentration followed the order Cu-Imi > Cu-mIm > Cu-Im. At a concentration of 25 μg mL⁻¹, the Cu-Imi nanozyme effectively scavenged ˙O₂⁻ radicals (achieving a scavenging rate greater than 60%), demonstrating its remarkable SOD-mimic activity, as shown in Fig. 2 B and Fig. S2B. And the SOD activity of Cu-Imi was determined to be 1280.78 U mg⁻¹, suggesting its satisfactory SOD activity (Fig. 2K).

Fig. 2 Assessment of MOF-based nanozymes’ activity: (A) SOD-mimic activity assessment; (B) absorption spectrum of ˙O₂⁻ chromogenic response in the absence and presence of 25 μg mL⁻¹ Cu-Imi; (C) cyclic voltammogram of Cu-Im; (D) CAT-mimic activity assessment; (E) absorption spectrum of H₂O₂ chromogenic response in the absence and presence of 25 μg mL⁻¹ Cu-Imi; (F) cyclic voltammogram of Cu-mIm; (G) DPPH elimination efficiency; (H) absorption spectrum of DPPH in the absence and presence of 15 μg mL⁻¹ Cu-Imi; (I) cyclic voltammogram of Cu-Imi; (J) schematic of the mechanism of SOD- and CAT-mimicking activities of Cu-Imi MOF-based nanozymes; (K) SOD-mimic activity of Cu-Imi; and (L) radar plots comparing the enzyme-mimicking activities of Cu-Im, Cu-mIm and Cu-Imi. Data are mean values ± sd (n ≥ 3).

Typically, ROS generated at the site of tissue damage also includes H₂O₂. To evaluate the CAT-mimic activity of Cu-MOFs (Cu-Im, Cu-mIm and Cu-Imi), we used the titanium sulfate colorimetric method to assess H₂O₂ levels in the presence of Cu-MOFs. Fig. 2D shows that as the concentration of Cu-MOFs increases, the H₂O₂ level decreases, indicating that the CAT activity of Cu-Imi is concentration-dependent, reaching a relative activity of 57.12% at 25 μg mL⁻¹. We also conducted a comparison between Cu-Imi, Cu-mIm and Cu-Im, ultimately demonstrating that Cu-Imi exhibits the highest level of CAT-mimicking activity among them. In addition, absorption spectra (Fig. 2E) were also used to detect the CAT-mimicking activity of Cu-Imi. As shown in Fig. 2E, pure H₂O₂ displays a characteristic absorption peak at 405 nm. In the presence of 25 μg mL⁻¹ Cu-Imi, the absorption intensity significantly decreased, indicating that Cu-Imi demonstrates excellent CAT-mimicking activity. In addition, we verified the ability of Cu-MOFs to scavenge DPPH (Fig. 2 G and H), and the results showed that Cu-Imi exhibited a superior scavenging ability compared to both Cu-Im and Cu-mIm. The potential mechanism underlying the ultra-high SOD- and CAT-mimic activities of Cu-Imi is shown in Fig. 2J and A, and all the results indicate that Cu-Imi has optimal enzyme-like activities.

To investigate the origin of the superior activity exhibited by Cu-MOFs (Fig. 2C, F and I), CV was used to examine the redox activity of Cu-MOFs. The study found that, at the same scan rate, Cu-Imi exhibited stronger redox peaks,⁴¹ indicating that Cu-Imi possesses superior catalytic performance compared to Cu-Im and Cu-mIm, which is consistent with our experimental results (Fig. 2L). The Cu-MOFs (Cu-Im, Cu-mIm and Cu-Imi) all exhibit a hemocyanin-mimic structure, and their 3D network architecture allows metal active sites to interact with substrates. Therefore, the observed differences in the catalytic activity of Cu-MOFs can be attributed to changes in the electronic effects at the metal active sites,⁴² resulting from the varying number of methyl groups on the imidazole rings. XPS analysis of Fig. 1I–L indicates that, although both Cu(I) and Cu (II) oxidation states are present in all Cu-MOFs, the higher number of methyl substituents in Cu-Imi reduces the oxidation state of Cu compared to Cu-Im and Cu-mIm. As a result, the lower oxidation state of Cu in Cu-Imi enhances its reactivity with ˙O₂⁻ and H₂O₂, thus increasing its catalytic activity. In summary, we present, for the first time, a strategy for preparing dual SOD and CAT mimetic nanozymes by precisely tuning the number of methyl groups, and have identified high-performance Cu-Imi as a candidate for further biological studies.

3.3. Density functional theory (DFT) study of the CAT-like activity of Cu-MOFs

To gain deeper insights into the influence of the metal center and ligand in MOF-based nanozymes on mimicking CAT-like activity, we performed DFT calculations on the adsorption of H₂O₂, a characteristic substrate of CAT (SI). Prior to these DFT calculations, the structures of the three Cu-MOFs and adsorbing molecules were fully optimized. As shown in Fig. 3A, all three Cu-MOFs exhibit a biomimetic coordination structure similar to that of hemocyanin. However, the Cu–N bond length (Fig. 3B) in Cu-Im (2.009 Å) and Cu-mIm (2.072 Å) is shorter than that in Cu-Imi (2.105 Å). Additionally, the metal active sites in Cu-Im and Cu-mIm have a lower electron density compared to that in Cu-Imi. All of the aforementioned factors may enhance the interaction between the Cu active site of Cu-Imi and substrate molecules. As shown in Fig. 3C, the adsorption energies (E_ads) of H₂O₂ on the three Cu-MOFs were calculated to be −1.79 eV for Cu-Im, −3.36 eV for Cu-mIm and −5.89 eV for Cu-Imi, with Cu-Imi having the lowest E_ads value. Although periodic boundary conditions of these MOFs might be considered to obtain more precise E_ads values during these calculations, the observed differences in E_ads can still explain the distinct interactions between Cu-Im, Cu-mIm and Cu-Imi with the substrate (H₂O₂). It is well known that E_ads is indicative of catalytic activity,⁴³ where a more negative E_ads implies stronger substrate adsorption. The calculations confirm that Cu-Imi exhibits the highest adsorption capacity for the substrate, consistent with experimental findings (Fig. 2). Thus, we are the first to discover that the CAT-like activity of MOF-based nanozymes can be regulated by precisely tuning the number of methyl groups in the organic ligands. This strategy has been validated through both experimental and computational methods in our study, and it demonstrates potential broad applicability.

Fig. 3 Cu-MOF structure and adsorption calculations: (A) schematic illustration of H₂O₂ adsorption in the Cu-MOF structure; (B) bond length and charge analysis of Cu–N within Cu-MOFs; and (C) adsorption energies for H₂O₂ adsorption on Cu-MOF surfaces.

3.4. Structural and morphological analysis

Dextran (Dex), a homopolysaccharide composed of glucose units, can undergo oxidation of its hydroxyl groups into aldehyde groups (Fig. 4A), enabling further cross-linking with amine groups through Schiff base reactions. In this study, ODex was synthesized by treating Dex with sodium periodate. The successful oxidation was confirmed by the ¹H NMR spectrum of ODex, which showed multiple new signals in the range of 5.0–6.0 ppm,²⁷ attributed to the formation of hemiacetal structures involving aldehyde groups (Fig. S3A). Additionally, the FT-IR spectrum (Fig. S3B) of ODex displayed a characteristic aldehyde peak at 1738 cm⁻¹ in comparison to Dex, further verifying the successful synthesis of ODex.⁴⁴ As demonstrated in Fig. 4B, the schematic illustrates the composition process of Gel-DA by gelatin and dopamine in the presence of EDC/NHS. The chemical structure of Gel-DA was characterized by ¹H NMR and FT-IR spectroscopy. In the ¹H NMR spectrum, the peak at 2.75 ppm was assigned to methylene protons adjacent to the phenyl group of dopamine (Fig. S3C). This peak was observed in the Gel-DA spectrum but was absent in the spectrum of pure gelatin. In the FT-IR spectrum of Gel-DA (Fig. S3D), the peak at 3030 cm⁻¹ was attributed to the aromatic –C–H groups of the dopamine moiety.⁴⁵ The amide stretching peak at 1640 cm⁻¹ confirmed the formation of the HN–CO bond, verifying the successful chemical conjugation between dopamine and gelatin.²⁸ In addition, O–H stretching at 3560–3250 cm⁻¹ was observed in the gelatin and dopamine samples,⁴⁶ as well as Gel-DA, due to the appearance of –OH groups.

Fig. 4 (A) Schematic diagram of Dex oxidation; (B) schematic of the composition process of Gel-DA by gelatin and dopamine; (C) FTIR spectra of the smart hydrogel; (D) SEM images of the cross-section of the smart hydrogel; (E) EDS elemental analysis of smart hydrogels; (F) hydrogel compression stress–strain curve; (G) hydrogel compression strength; and (H) hydrogel compression modulus. Data are mean values ± sd (n ≥ 3).

The sol–gel of the hydrogel is shown in Fig. S4A and B. The FT-IR spectrum of the smart hydrogel (Fig. 4C) exhibited characteristic peaks of amide bonding at 1635 and 1239 cm⁻¹. The peaks observed at 1532 and 1445 cm⁻¹ were attributed to C=N stretching vibrations,⁴⁶ confirming the formation of Schiff base linkages within the hydrogel. Furthermore, the introduction of Cu-Imi led to hydrogen bond formation within the hydrogel network, which resulted in a red shift of the hydroxyl (OH stretching vibration) frequency to a lower wavenumber (from 3372 cm⁻¹ to 3360 cm⁻¹). These FT-IR results indicate the successful incorporation of Cu-Imi into the hydrogel network. The internal structure of the hydrogel was investigated using SEM. The image in Fig. 4D reveals a uniformly distributed 3D porous architecture with interconnected macropores, indicating a good structural stability, which is crucial for effective diffusion for nutrients and waste in a 3D environment, and cell motility. Compared with Gel-DA/ODex, the incorporation of Cu-Imi had minimal impact on the microstructure of the smart hydrogel, which maintained its macroporous structure. Fig. 4E shows the EDS mapping of different elements (C, N, O and Cu) in the Gel-DA/ODex^0.5 smart hydrogels. The uniform distribution of C, N, O and Cu throughout the cross-section suggests good compatibility between the hydrogel matrix and Cu-Imi.

3.5. Swelling tests, pH-responsive release and mechanical tests

The rate of swelling and the equilibrium state are crucial indicators for assessing the performance of hydrogel materials. The hydrogel (Gel-DA/ODex) exhibited a rapid water absorption rate, achieving 6.15% swelling within 10 minutes and reaching 15.16% swelling within 60 minutes (Fig. S4C). The smart hydrogel (Gel-DA/ODex^0.5) exhibited improved anti-swelling properties, achieving a swelling ratio of only 10.18% at 60 minutes – an enhancement of 32.84% compared with the hydrogel (Gel-DA/ODex). Both hydrogels reached equilibrium swelling at approximately 60 hours (Fig. S4D), with the hydrogel (Gel-DA/ODex) achieving a swelling ratio of 59.38%, while the smart hydrogel (Gel-DA/ODex^0.5) reached 49.09%. An appropriate swelling ratio is beneficial for the hydrogel to absorb tissue fluids at the inflammatory bone defect site and conform to the shape of the defect.⁴⁷ The enhanced anti-swelling properties of the smart hydrogel are primarily attributed to the incorporation of Cu-Imi.

According to existing studies, the local microenvironment pH in early tissue injuries,⁴⁸ such as bone fractures, skin wounds and muscle tears, is approximately 5.5. To address these characteristics, we designed a novel smart hydrogel with pH-responsive properties. This hydrogel can rapidly release nanozymes at low pH to reduce local ROS levels, provide rapid antibacterial activity and shorten inflammation duration. We adjusted the pH of the PBS solution (pH 7.35) to 5.5 by 0.01 M hydrochloric acid to simulate the local pH at an early tissue injury site to test the responsive release behavior of the smart hydrogel. As shown in Fig. S5, the smart hydrogel exhibited rapid nanozyme release behavior at pH 5.5. During the 12-hour experiment, the smart hydrogel released 1.71 times more nanozyme under early injury conditions (pH 5.5) than under normal conditions (pH 7.35), with cumulative release still 1.65 times higher at 14 days. In this study, the smart hydrogel (Gel-DA/ODex^0.5) was cross-linked by Schiff base formation between –CHO and –NH₂ groups. Studies have shown that Schiff base bonds are labile under acidic conditions, and their stability decreases as pH decreases. The breaking of these Schiff base linkages leads to the dissociation of the hydrogel structure, which facilitates the release of Cu-Imi nanozymes from both the surface and deeper regions of the hydrogel; a similar release behavior was also observed by Li et al.⁴⁹

The mechanical testing results demonstrated that the compressive strength of the smart hydrogel significantly increased with the addition of Cu-Imi, showing an enhancement of 453.72% (Fig. 4F). Specifically, the compressive strength and modulus of the Gel-DA/ODex hydrogel were approximately 16.64 kPa and 106 kPa (Fig. 4G and H), respectively, while the compressive strength and modulus of the smart hydrogel Gel-DA/ODex^0.5 increased by 453.72% (92.14 kPa) and 206.23% (324.6 kPa), respectively. This improvement can be attributed to the Cu-Imi nanozyme inhibiting the mobility of the hydrogel chains. Typically, the reinforcing effect of nanofillers largely depends on their dispersion and interaction with the polymer matrix.²³ In consideration of the findings from the FT-IR, SEM and EDS analyses (see Fig. 4C, D and E), the significant enhancement in the mechanical properties of the smart hydrogel can be attributed to the presence of robust electrostatic interactions between the well-dispersed Cu-Imi and the hydrogel matrix, which effectively suppress phase separation and facilitate efficient stress transfer at the interface. Similarly, Liu et al. reported that encapsulating bimetal–organic framework-derived Mn@Co₃O₄@Pt in a composite hydrogel formed from phenylboronic acid-modified sodium alginate (Alg-PBA) and polyvinyl alcohol (PVA) also led to an increase in the hydrogel's compressive modulus.⁵⁰ Consequently, the incorporation of Cu-Imi confers enhanced mechanical performance and structural stability to the smart hydrogel.

3.6. Antibacterial activity evaluation

Tissue repair materials with exceptional antimicrobial activity can significantly reduce the risk of secondary infections, which are commonly encountered during surgical procedures due to bacterial contamination. The antibacterial activities of Cu-Im, Cu-mIm and Cu-Imi against common pathogens E. coli (Gram-negative) and S. aureus (Gram-positive) were evaluated using the disc diffusion method. As shown in Fig. S6, Cu-Im, Cu-mIm and Cu-Imi exhibited significant inhibition against both bacterial strains. The superior antibacterial performance of the Cu-MOFs (Cu-Im, Cu-mIm and Cu-Imi) can be primarily attributed to two factors: the release of copper ions, which effectively disrupts bacterial cell membranes, and the antimicrobial properties of the imidazole ligands.

To evaluate the antibacterial activity of the smart hydrogel, the co-culture experiments were conducted with bacteria. Fig. 5A provides a qualitative visualisation of the antibacterial effects of different hydrogels against E. coli and S. aureus. The results clearly demonstrate that the antibacterial activity of the smart hydrogels increases with the increase in Cu-Imi content, as evidenced by the significant reduction in bacterial colonies in the culture plates co-incubated with Gel-DA/ODex^0.5. Quantitative assessments of the antibacterial activity against both bacterial strains are presented in Fig. 5C and D. Notably, the antibacterial rates of Gel-DA/ODex^0.5 reached 98.7% for E. coli and 99.87% for S. aureus. This exceptional antibacterial performance is primarily attributed to the incorporation of Cu-Imi, which endows the hydrogel with bactericidal capabilities (Fig. S6). To further investigate the antibacterial mechanism of the smart hydrogel, SEM imaging was performed on bacteria co-cultured with the hydrogel (Fig. 5B). SEM images of the control and Gel-DA/ODex groups revealed intact and undamaged bacterial structures. However, after incubation with Gel-DA/ODex^0.5, significant morphological changes were observed in E. coli and S. aureus. The bacterial cell membranes exhibited severe damage, including depressions and ruptures. These findings underscore the excellent antibacterial activity of the smart hydrogel and its potential to mitigate bacterial infections during surgical procedures.

Fig. 5 Evaluation of the antibacterial effect of samples in vitro: (A) bacterial inhibition of smart hydrogels against E. coli and S. aureus; (B) morphology of bacteria after co-culturing with different smart hydrogels; (C) bacterial inhibition rate of smart hydrogels against E. coli; (D) bacterial inhibition rate of smart hydrogels against S. aureus; data are mean values ± sd (n ≥ 3); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

3.7. In vitro cytocompatibility assessment

3.7.1. The effect of the smart hydrogel on HUVECs and rBMSCs. Biocompatibility is crucial for bone repair materials. We evaluated the biocompatibility of hydrogels by co-culturing them with rBMSCs or HUVECs and quantifying cell viability using a CCK-8 assay. Fig. 6A shows rBMSC viability at different time points. The relative growth rates (RGR) of rBMSCs cultured with all hydrogels were higher than 94% at various time points, indicating excellent biocompatibility. Live cell imaging using FDA staining further confirmed the high cell viability of various samples (Fig. 6D). The rBMSC population increased significantly over time, suggesting that the hydrogels provided a favorable environment for cell growth. Additionally, the cytocompatibility of the hydrogels toward HUVECs was also evaluated, and an RGR value as high as 99% was obtained across all groups (Fig. 6B). Thus, our hydrogels offer a conducive environment for the growth and proliferation of rBMSCs and HUVECs, promoting osteogenesis and angiogenesis.

Fig. 6 Biocompatibility evaluation of the hydrogel: (A) cell viability of BMSCs cocultured with various smart hydrogels; (B) cell viability of HUVECs cocultured with various smart hydrogels; (C) quantification of the migration area of BMSCs; (D) fluorescent staining images of rBMSCs; (E) fluorescence images of rBMSC DNA/RNA damage (marked by white circles); (F) fluorescence images of HUVEC DNA/RNA damage (marked by white circles); actin was stained by phalloidin conjugated tetramethylrhodamine (red) and the cell nucleus were stained by DAPI (blue). Data are mean values ± sd (n ≥ 3); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

In inflammatory bone defects, elevated ROS levels directly damage the morphology and DNA/RNA of stem cells,⁵¹ which severely affects the osteogenic differentiation of stem cells, and this could be observed by CLSM. Under H₂O₂ stimulation, the control and Gel-DA/ODex groups showed significant changes in stem cell morphology, with triangular and rounded shapes and varying degrees of DNA/RNA damage (Fig. 6E, marked by white circles). Compared to control and Gel-DA/ODex groups, CLSM results showed that BMSCs cultured with Gel-DA/ODex^0.5 hydrogels exhibited larger cytoplasmic spreading areas and more prominent actin filaments, presenting a polygonal osteoblast-like morphology (Fig. 6E). Interestingly, when rBMSCs transitioned from an elongated shape to a more osteoblast-like polygonal form with abundant filopodia, they were considered to have superior osteogenic differentiation activity. Thus, Gel-DA/ODex^0.5 hydrogels have a higher potential to induce early osteogenic differentiation in oxidative stress microenvironments compared to Gel-DA/ODex hydrogels.

Under H₂O₂ stimulation to mimic the oxidative stress microenvironment of inflammatory bone defects, the control and Gel-DA/ODex groups showed severe deformation of cell morphology, minimal actin, and visibly impaired DNA/RNA (Fig. 6F, marked by white circles). As shown in Fig. 6F, the HUVECs co-cultured with Gel-DA/ODex and Gel-DA/ODex^0.5 exhibited normal cytoskeletal and nuclear morphology, displaying a correct morphological phenotype, indicating that the hydrogels provide an ideal growth environment for HUVECs. However, compared with the control, the Gel-DA/ODex^0.5 group showed more spread cytoplasm and richer actin filaments, typically indicative of enhanced cellular activity.

3.7.2. Scratch assay. High ROS in inflammatory bone defects severely affects the migratory activity of stem cells and thus prevents osteogenic differentiation. We used H₂O₂ to stimulate rBMSCs to simulate the oxidative stress microenvironment of inflammatory bone defects and performed scratch experiments to evaluate the effect of smart hydrogels on the migration of rBMSCs. As shown in Fig. 6C and Fig. S7A, the control group and the Gel-DA/ODex group exhibited limited migration, whereas rBMSCs in the Gel-DA/ODex^0.5 hydrogel group demonstrated the greatest migration distance, approximately 6.86 times that of the control group (Fig. S7A). The results of scratch assay indicated that the Gel-DA/ODex^0.5 hydrogel significantly enhanced the migration of rBMSCs in the oxidative stress microenvironment, which could be attributed to the ROS scavenging effect of Cu-Imi nanozymes to enhance cell viability.

3.7.3. Response of RAW264.7 cells (macrophage) to smart hydrogels. Prolonged inflammation can negatively impact the local microenvironment at the bone defect site, inhibiting new bone formation. Therefore, reducing the inflammatory response and shortening the duration of inflammation are critical for promoting bone tissue regeneration. Macrophages, as the primary defenders of the host immune system, play a pivotal role in innate immunity.⁵² As shown in Fig. 7B, we activated RAW 264.7 cells with H₂O₂ and co-cultured them with hydrogels to observe morphological changes of RAW 264.7 cells, and there was clear nuclear damage (marked by white circles). H₂O₂-activated RAW 264.7 cells displayed significant morphological changes, transitioning from a round shape to a polygonal form with increased filamentous pseudopodia, whereas most RAW 264.7 cells in the Gel-DA/ODex^0.5 group maintained a rounded shape with minimal filamentous pseudopodia, indicating that smart hydrogels can effectively scavenge ROS reducing macrophage stimulation.

Fig. 7 Effects of smart hydrogels on RAW 264.7 in vitro: (A) diagram of the osteogenesis mechanism of the smart hydrogel in an oxidative stress microenvironment; (B) fluorescence images of RAW 264.7 DNA/RNA damage (marked by white circles); (C) the cytoprotection ability of smart hydrogels; (D) relative fluorescence intensity of RAW264.7 treated with different smart hydrogels stained by DCFH-DA; (E) fluorograms for clearing of ROS within macrophages by smart hydrogels; inflammatory cytokines were detected by ELISA: (F) TNF-α; (G) IL-1β; (H) IL-10; and (I) Arg-1. Data are mean values ± sd (n ≥ 3); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

3.7.4. Cellular antioxidation performance and immunomodulation. Before exploring the potential biological applications of the Gel-DA/ODex^0.5 smart hydrogel, its excellent SOD and CAT mimic activities encouraged us to assess its cytoprotective effects and intracellular ROS scavenging ability (Fig. 7A). The ROS removal capability of the hydrogel was evaluated in macrophages according to the manufacturer's protocol (Beyotime Biotechnology, China). As shown in Fig. 7D and E, the smart hydrogels effectively reduced intracellular ROS levels in macrophages, achieving an ROS clearance rate of approximately 93.7% compared with the H₂O₂ group. Furthermore, as shown in Fig. 7C, macrophages co-incubated with Gel-DA/ODex^0.5 maintained a RGR of over 90%, whereas H₂O₂-treated macrophages showed a reduction in RGR to 51.38%. These results indicate that Gel-DA/ODex^0.5 protects cells from H₂O₂-induced oxidative stress. In summary, these findings demonstrate that the Gel-DA/ODex^0.5 smart hydrogel not only exhibits good biocompatibility but also effectively mimics SOD and CAT, protecting cells from oxidative damage.

The inability of conventional biomineralized hydrogels⁵³ to scavenge ROS and modulate the immune response limits their use in the treatment of inflammatory bone defects. We further evaluated the effect of the smart hydrogels Gel-DA/ODex0.5 on macrophage polarization from the M1 to M2 phenotype. Macrophages generally exist in a quiescent M0 state and switch to the activated M1 state upon bacterial infection or other stimuli, releasing pro-inflammatory cytokines (such as IL-1β and TNF-α). Alternatively, macrophages can adopt the activated M2 phenotype, promoting tissue regeneration by releasing anti-inflammatory cytokines (such as IL-10 and Arg-1). Upon LPS treatment, the secretion of TNF-α by macrophages increased by 4.62-fold (Fig. 7F) and IL-1β secretion by macrophages increased by 5.41-fold (Fig. 7G). As previously reported, these elevated markers confirm LPS-induced macrophage inflammation and M1 polarization. Following treatment with Gel-DA/ODex^0.5, the expression levels of pro-inflammatory M1 markers were significantly reduced, while the expression levels of anti-inflammatory M2 markers (IL-10 and Arg-1) were significantly increased (Fig. 7H and I). In contrast, the LPS + Gel-DA/ODex group showed no significant reduction in the secretion of pro-inflammatory cytokines IL-1β and TNF-α, nor an increase in anti-inflammatory cytokines IL-10 and Arg-1 compared with the LPS group. It is speculated that the regulation of macrophage polarization from M1 to M2 in the Gel-DA/ODex^0.5 group is attributed to the addition of Cu-Imi nanozymes that mimic hemocyanin, effectively scavenging ROS and reducing macrophage stimulation. Therefore, Gel-DA/ODex^0.5 successfully induced M1-to-M2 macrophage polarization and demonstrated satisfactory anti-inflammatory activity in vitro.

3.7.5. Osteogenic activity of smart hydrogels in the inflammatory bone defect environment in vitro. The rBMSCs were cultured in macrophage-conditioned medium (while using H₂O₂ stimulation to mimic the oxidative stress microenvironment of inflammatory bone defects) to assess the osteoimmunomodulatory activity of the smart hydrogels (Fig. 8A). After 14 days of culture in conditioned medium, ALP staining (Fig. 8B) and quantitative analysis of ALP activity (Fig. 8C) were performed. The results showed that the Gel-DA/ODex^0.5 group exhibited significantly higher ALP expression compared to other groups, indicating superior early osteogenic induction activity. Additionally, Alizarin Red S (ARS) staining was conducted on Day 14 and Day 21 of culture to assess calcium deposition in the extracellular matrix, a key marker of late-stage osteogenic differentiation. As shown in Fig. 8D and F, qualitative and quantitative analyses revealed that the Gel-DA/ODex^0.5 group had the most pronounced mineralized calcium nodules and the deepest ARS staining on both Day 14 and Day 21. These findings suggest that Gel-DA/ODex^0.5 promotes osteogenesis under inflammatory conditions, demonstrating enhanced osteogenic activity in both early (ALP staining, Fig. 8B) and late (ARS staining, Fig. 8C) stages. This may be attributed to the fact that the addition of a bionic nanozyme can effectively scavenge ROS (Fig. 7E) and reduce the inflammatory response (Fig. 7F–I).

Fig. 8 Evaluation of osteogenesis in the microenvironment of oxidative stress: (A) rBMSC culture process; (B) ALP staining on day 14; (C) quantitative analysis of ALP activity; (D) ARS staining on day 14 and day 21; (E) VEGF measurement via ELISA; (F) quantitative analysis of ARS; measurement of osteogenic cytokines via ELISA: (G) OCN, (H) COL-1, (I) BMP-2, (J) OPN, and (K) RUNX-2. Data are mean values ± sd (n ≥ 3); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

To further investigate the mechanism underlying the osteogenic activity of Gel-DA/ODex^0.5, the expression levels of osteogenesis-related factors in rBMSCs were measured via ELISA, including OCN (Fig. 8G), COL-1 (Fig. 8H), BMP-2 (Fig. 8I), OPN (Fig. 8J) and RUNX-2 (Fig. 8K). The expression of these five key osteogenic markers was significantly higher in the Gel-DA/ODex^0.5 group compared to the control and Gel-DA/ODex groups, confirming the promoting effect of Cu-Imi nanozymes on osteogenic differentiation. These findings are consistent with the results from ALP and ARS staining, suggesting that Gel-DA/ODex^0.5 induces an osteoimmune microenvironment by scavenging ROS and promoting macrophage M1 to M2 polarization. The Gel-DA/ODex^0.5 hydrogel exhibits excellent osteogenic properties in the microenvironment of inflammation and oxidative stress.

3.7.6. Response of HUVECs to smart hydrogels. The Gel-DA/ODex^0.5 smart hydrogels can eliminate excess ROS and release oxygen during the early stages of repair; local ROS levels are balanced through the hydrogel's scavenging action. However, the process of bone tissue repair is lengthy. The sustained supply of oxygen and nutrients required for long-term bone tissue regeneration still depends on the formation of new blood vessels. Fortunately, research has found that increased oxygen concentration can promote the formation of new blood vessels,⁵⁴ which in turn transport oxygen and nutrients, promote bone repair, and form a virtuous cycle. So theoretically, the Gel-DA/ODex^0.5 hydrogels should be able to promote angiogenesis. We assessed the effect of Gel-DA/ODex^0.5 hydrogels on HUVEC migration in an oxidative stress microenvironment, and the results showed that the migration distance in the Gel-DA/ODex^0.5 group was approximately two times higher than that in the control and Gel-DA/ODex groups (Fig. S8A and B). These findings suggest that Gel-DA/ODex^0.5 enhances endothelial cell activity and induces cell migration.

To more directly demonstrate the angiogenesis-promoting effect of Gel-DA/ODex^0.5, we measured the secretion of VEGF by HUVECs in different groups via ELISA. As shown in Fig. 8E, the VEGF level in the Gel-DA/ODex^0.5 group was approximately 2.17-fold that of the control group and 2.02-fold that of the Gel-DA/ODex group, providing direct evidence for the angiogenesis-promoting effect of the smart hydrogels. Notably, the enhancement of endothelial cell viability and the elevation of VEGF levels are sustained, thereby promoting the transport of oxygen and nutrients. This effect is consistent with the ability of Gel-DA/ODex^0.5 to scavenge intracellular ROS from early-stage tissue damage and release oxygen, creating a positive feedback loop that further promotes bone tissue regeneration.

3.8. In vivo bone regeneration evaluation of smart hydrogels

3.8.1 Micro-CT analysis and histological observation. We established a circular defect (1 mm in diameter and 2 mm in depth) in the femur of rats using a minimally invasive technique to evaluate the therapeutic efficacy of the intelligent hydrogel on inflammatory bone defects (see Section 2.11.1 for details of the animal model). The 3D micro-CT results are shown in Fig. 9A. The control group underwent self-repair with minimal bone volume formation and the presence of a large bone defect (week 8), indicating a severe lack of bone regeneration capacity in the absence of treatment (Fig. 9B, the bone volume/total volume was 20.09%). In contrast, the Gel-DA/ODex^0.5 group significantly promoted bone regeneration after the release of Cu-Imi nanozymes for treatment (Fig. 9B, the bone volume/total volume was 44.93%). Notably, bone mineral density (BMD) is crucial for evaluating new bone regeneration, and as shown in Fig. 9C, the BMD value in the Gel-DA/ODex^0.5 group was approximately 1.5 times higher than that in the Gel-DA/ODex group and approximately 1.56 times higher than that in the control group. Therefore, it is hypothesized that the excellent therapeutic ability of Gel-DA/ODex^0.5 for inflammatory bone defects is due to the smart hydrogels's ability to release Cu-Imi nanozymes with ROS scavenging and bone immune modulation functions.

Fig. 9 Evaluation of the therapeutic efficacy of smart hydrogels for inflammatory bone defects: (A) micro-CT images; (B) BV/TV at 8 weeks; (C) BMD at 8 weeks; (D) HE staining; and (E) Masson staining. Data are mean values ± sd (n ≥ 3); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

To further analyze the integration of new bone and host bone, hematoxylin and eosin staining (H&E) and Masson trichrome staining were used for verification. As shown in Fig. 9D, the thickness of the regenerated bone in the Gel-DA/ODex^0.5 group was significantly greater than that in the other two groups, with more new mineralized bone formation. Notably, the collagen is considered an important component of bone.⁵⁵ Collagen deposition in the repaired inflammatory bone defect was assessed using Masson's trichrome staining (Fig. 9E). As shown in the figure, a substantial amount of collagen (blue) was observed in the Gel-DA/ODex^0.5 group, with collagen thickness at week 8 far exceeding that of the other two groups. Additionally, during the H&E staining process, we clearly observed typical new blood vessel formation in the new bone area of the Gel-DA/ODex^0.5 group (Fig. 9D), with neovascularization appearing more organized and surrounded by mature bone tissue. The results indicate that the body is incapable of effectively repairing severe inflammatory bone defects independently, and even after an eight-week recovery period, cavitations persist within the area of the inflammatory bone defect, where there is minimal bone tissue formation (Fig. 9A, D, and E). However, after treating with the smart hydrogels Gel-DA/ODex^0.5, bone regeneration was significantly promoted, effectively repairing the inflammatory bone defect. This suggests that in vivo, the smart hydrogel Gel-DA/ODex^0.5 possesses excellent capability for inflammatory bone defect repair, consistent with the results obtained from in vitro experiments.

3.8.2. Osteoimmunomodulation and vasculogenesis of the smart hydrogels in vivo. In view of satisfactory results of using the Gel-DA/ODex^0.5 smart hydrogel (abbreviated as GOD^0.5 in the image) for the repair of bone defects, we further investigated the immune mechanisms and angiogenic effects of this hydrogel in vivo. Inflammatory bone defect healing is closely related to immune modulation. M1 macrophages accumulate at the inflammatory bone defect site and secrete inflammatory cytokines and cytokines, creating a persistent inflammatory microenvironment that severely impairs bone defect remodeling. In contrast, M2 macrophages not only secrete anti-inflammatory factors to suppress inflammation, but also promote tissue repair and remodeling. We validated the bone immune modulation function of the smart hydrogels using fluorescence imaging. In our study, TNF-α (a marker for M1 macrophages) was detected in M1 macrophages at the defect site, while CD206 was identified in M2 macrophages in the same region. As shown in Fig. 10A–D, the proportion of M1 macrophages significantly decreased in the smart hydrogel group, while the proportion of M2 macrophages increased, indicating the excellent bone immune modulation capabilities of the smart hydrogel. Angiogenesis occurs primarily 2–4 weeks after injury during the early stages of tissue regeneration, with denser blood vessels enhancing oxygen and nutrient delivery to the bone defect site, thereby promoting bone remodeling. As shown in Fig. 10E and F, after 4 weeks of implantation, the number of CD31-positive cells in the Gel-DA/ODex^0.5 group was significantly higher than that in the other two groups, strongly supporting that Gel-DA/ODex^0.5 accelerates angiogenesis during the early stages of tissue regeneration. These results indicate that Gel-DA/ODex^0.5 has a significant impact on the phenotypic transition of macrophages from M1 to M2 type, modulates the immune microenvironment, promotes angiogenesis and accelerates bone remodeling, thus confirming our findings and hypothesis.

Fig. 10 Immunomodulation and angiogenesis in the oxidative stress microenvironment: (A) immunofluorescent staining of TNF-α; (B) quantification of TNF-α using imageJ; (C) immunofluorescent staining of CD206; (D) quantification of CD206 using imageJ; (E) representative immunohistochemical staining of CD31; and (F) quantification of CD31 using imageJ. Data are mean values ± sd (n ≥ 3); *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

4. Conclusion

In conclusion, by mimicking the natural IADS mechanism of efficient redox reactions, we present an efficient SOD- and CAT-inspired design inspired by the horseshoe crab hemocyanin, complexed with a double-crosslinked hydrogel consisting of oxidized dextran and dopamine-functionalized gelatin, to remodel the oxidative stress microenvironment and for inflammatory bone defect therapy using Gel-DA/ODex^0.5. Our study verified that the structure–activity relationship theory derived from protein engineering can be applied to the design of MOF nanozymes by modulating ligands to obtain MOF nanozymes with high adsorption energy and high catalytic efficiency for ROS. Thus, the smart hydrogel Gel-DA/ODex0.5 supports metabolic processes by intelligently releasing the Cu-Imi nanozyme to block oxidative stress-induced DNA damage, protects cellular activity, induces macrophage polarization from M1 to M2, modulates the immune microenvironment, and maintains stem cell viability and osteogenic differentiation. Strikingly, the smart hydrogel Gel-DA/ODex^0.5 demonstrated effective remodeling of the oxidative stress microenvironment in vivo and repair of inflammatory bone defects during regeneration of inflammatory bone defects.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its SI. Supplementary information: Partial characterization. See DOI: https://doi.org/10.1039/d5tb01686j.

Acknowledgements

This work was financially supported by “From 0 to 1” innovative research projects of Sichuan University (2022SCUH0045); the Key Research and Development Program of Sichuan Province (2023NSFSC1001); and the Key Research and Development Program of Chengdu Supported by Chengdu Science and Technology Program (2024-YF05-0062-SN). The authors extend their gratitude to Ceshigo Research Service (https://www.ceshigo.com) for providing invaluable assistance with the XPS analysis.

References

1. Y. Xu, C. Xu, L. He, J. Zhou, T. Chen, L. Ouyang, X. Guo, Y. Qu, Z. Luo and D. Duan, Bioact. Mater., 2022, 16, 271–284.
2. J. Wang, C. Zhang, W. Zhang, W. Liu, Y. Guo, P. Dang, L. Wei, H. Zhao, X. Deng, S. Guo and L. Chen, Adv. Funct. Mater., 2023, 34, 2306319.
3. Y. Huang, Z. Du, K. Li, W. Jing, P. Wei, B. Zhao, Y. Yu, Q. Cai and X. Yang, Adv. Fiber Mater., 2022, 4, 894–907.
4. M. Schweikle, S. H. Bjornoy, A. T. J. van Helvoort, H. J. Haugen, P. Sikorski and H. Tiainen, Acta Biomater., 2019, 90, 132–145.
5. L. Ma, Y. Cheng, X. Feng, X. Zhang, J. Lei, H. Wang, Y. Xu, B. Tong, D. Zhu, D. Wu, X. Zhou, H. Liang, K. Zhao, K. Wang, L. Tan, Y. Zhao and C. Yang, Adv. Mater., 2024, 36, e2307846.
6. T. Wang, M. Bai, W. Geng, M. Adeli, L. Ye and C. Cheng, Nat. Commun., 2025, 16, 856.
7. P. Zhang, X. Liu, P. Guo, X. Li, Z. He, Z. Li, M. J. Stoddart, S. Grad, W. Tian, D. Chen, X. Zou, Z. Zhou and S. Liu, Bioact. Mater., 2021, 6, 3097–3108.
8. X. Zhang, W. Jiang, C. Xie, X. Wu, Q. Ren, F. Wang, X. Shen, Y. Hong, H. Wu, Y. Liao, Y. Zhang, R. Liang, W. Sun, Y. Gu, T. Zhang, Y. Chen, W. Wei, S. Zhang, W. Zou and H. Ouyang, Nat. Commun., 2022, 13, 5211.
9. C. Deng, Q. Zhou, M. Zhang, T. Li, H. Chen, C. Xu, Q. Feng, X. Wang, F. Yin, Y. Cheng and C. Wu, Adv. Sci., 2022, 9, e2105727.
10. L. Tian, J. Wang, Y. Tang, P. Liu, Q. Cao, Y. Liu, X. Li, X. Chen, X. Zhu and X. Zhang, Composites, Part B, 2025, 301.
11. A. Rimessi, M. Previati, F. Nigro, M. R. Wieckowski and P. Pinton, Int. J. Biochem. Cell Biol., 2016, 81, 281–293.
12. L. Ferrucci and E. Fabbri, Nat. Rev. Cardiol., 2018, 15, 505–522.
13. Y. Yang, Z. Li, X. Fan, C. Jiang, J. Wang, Y. Rastegar-Kashkooli, T. J. Wang, J. Wang, M. Wang, N. Cheng, X. Yuan, X. Chen, B. Jiang and J. Wang, ACS Nano, 2024, 18, 16450–16467.
14. R. Tian, H. Ma, W. Ye, Y. Li, S. Wang, Z. Zhang, S. Liu, M. Zang, J. Hou, J. Xu, Q. Luo, H. Sun, F. Bai, Y. Yang and J. Liu, Adv. Funct. Mater., 2022, 32.
15. Y. Zhang, W. Liu, G. Wei, Q. Liu, G. Shao, X. Gu, X. Cui, Z. Zhou, Y. Wang, S. Zhao, F. Muhammad, S. Li, T. Li, Y. Du and H. Wei, ACS Nano, 2024, 18, 9019–9030.
16. Z. Chen, B. Li, Y. Zhang, X. Shang, C. Ma, W. Gao and Z. Zhu, Mater. Horiz., 2025, 12, 5957–5974.
17. S. Dong, Y. Dong, T. Jia, S. Liu, J. Liu, D. Yang, F. He, S. Gai, P. Yang and J. Lin, Adv. Mater., 2020, 32, e2002439.
18. Z.-Z. Xu, B.-F. Xu, A. P. Tanjung, L. Zhang, A.-J. Wang, P. Song, L.-P. Mei and J.-J. Feng, Sens. Actuators, B, 2024, 412.
19. J. Zhang, J. Wang, J. Liao, Y. Lin, C. Zheng and J. Liu, ACS Appl. Mater. Interfaces, 2021, 13, 50236–50245.
20. L. Zhang, H. Wang and X. Qu, Adv. Mater., 2024, 36, e2211147.
21. J. Zhang, M. Guo, Q. He, Z. Zhang, B. Wu, H. Wu, R. Li, Q. Zhang, Y. Tang, Y. Lin and Y. Jin, Small, 2024, 20, e2310675.
22. H. Wu, Y. S. Chua, V. Krungleviciute, M. Tyagi, P. Chen, T. Yildirim and W. Zhou, J. Am. Chem. Soc., 2013, 135, 10525–10532.
23. S. Feng, Q. Tang, Z. Xu, K. Huang, H. Li and Z. Zou, Food Hydrocolloids, 2023, 135.
24. J. Liang, Z. Liang, R. Zou and Y. Zhao, Adv. Mater., 2017, 29.
25. Z. He, X. Huang, C. Wang, X. Li, Y. Liu, Z. Zhou, S. Wang, F. Zhang, Z. Wang, O. Jacobson, J. J. Zhu, G. Yu, Y. Dai and X. Chen, Angew. Chem., Int. Ed., 2019, 58, 8752–8756.
26. K. Xiang, H. Wu, Y. Liu, S. Wang, X. Li, B. Yang, Y. Zhang, L. Ma, G. Lu, L. He, Q. Ni and L. Zhang, Theranostics, 2023, 13, 2721–2733.
27. Z. Ge, L. Long, F. Zhang, R. Dong, Z. Chen, S. Tang, L. Yang and Y. Wang, Int. J. Biol. Macromol., 2024, 273, 132740.
28. X. Sun, J. Yang, J. Ma, T. Wang, X. Zhao, D. Zhu, W. Jin, K. Zhang, X. Sun, Y. Shen, N. Xie, F. Yang, X. Shang, S. Li, X. Zhou, C. He, D. Zhang and J. Wang, Biofabrication, 2023, 15.
29. S. Ji, B. Jiang, H. Hao, Y. Chen, J. Dong, Y. Mao, Z. Zhang, R. Gao, W. Chen, R. Zhang, Q. Liang, H. Li, S. Liu, Y. Wang, Q. Zhang, L. Gu, D. Duan, M. Liang, D. Wang, X. Yan and Y. Li, Nat. Catal., 2021, 4, 407–417.
30. Y. Xu, Y. Luo, Z. Weng, H. Xu, W. Zhang, Q. Li, H. Liu, L. Liu, Y. Wang, X. Liu, L. Liao and X. Wang, ACS Nano, 2023, 17, 18732–18746.
31. H. T.-T. Karen, A. Magnus and J. E. Carpenter, Chem. Rev., 1994, 94, 44106–44935.
32. M. Ashrafuzzaman, M. H. Razu, N.-N. Showva, T. A. Bondhon, M. Moniruzzaman, S. A. R. Rahman, M. R. I. Rabby, F. Akter, M. Khan and J. Kolla, Cell. Microbiol., 2022, 2022, 1–17.
33. C. F. Holder and R. E. Schaak, ACS Nano, 2019, 13, 7359–7365.
34. R. Majidi, A. Farhadi, I. Danaee, N. B. Panah, D. Zarei and S. Nikmanesh, Prog. Org. Coat., 2023, 183.
35. Y. Ye, W. Guo, L. Wang, Z. Li, Z. Song, J. Chen, Z. Zhang, S. Xiang and B. Chen, J. Am. Chem. Soc., 2017, 139, 15604–15607.
36. P. Bhadane, P. Mahato, D. Menon, B. K. Satpathy, L. Wu, S. Chakraborty, P. Goyal, I. Lynch and S. K. Misra, Environ. Sci.: Nano, 2024, 11, 2385–2396.
37. C. Li, J. Shen, K. Wu and N. Yang, Small, 2022, 18, e2106607.
38. M. Li, J. Chen, W. Wu, Y. Fang and S. Dong, J. Am. Chem. Soc., 2020, 142, 15569–15574.
39. A. Bakandritsos, R. G. Kadam, P. Kumar, G. Zoppellaro, M. Medved, J. Tucek, T. Montini, O. Tomanec, P. Andryskova, B. Drahos, R. S. Varma, M. Otyepka, M. B. Gawande, P. Fornasiero and R. Zboril, Adv. Mater., 2019, 31, e1900323.
40. L. Ma, J. J. Zheng, N. Zhou, R. Zhang, L. Fang, Y. Yang, X. Gao, C. Chen, X. Yan and K. Fan, Nat. Commun., 2024, 15, 233.
41. Z. Wei, G. Peng, Y. Zhao, S. Chen, R. Wang, H. Mao, Y. Xie and C. Zhao, ACS Nano, 2022, 16, 18329–18343.
42. J. Wu, Y. Yu, Y. Cheng, C. Cheng, Y. Zhang, B. Jiang, X. Zhao, L. Miao and H. Wei, Angew. Chem., Int. Ed., 2021, 60, 1227–1234.
43. W. Yang, X. Yang, J. Jia, C. Hou, H. Gao, Y. Mao, C. Wang, J. Lin and X. Luo, Appl. Catal., B, 2019, 244, 1096–1102.
44. G. Bai, C. Niu, X. Liang, L. Li, Z. Wei, K. Chen, K. Bohinc and X. Guo, Int. J. Biol. Macromol., 2024, 267, 131288.
45. B. Zhao, Y. Zhang, D. Li, X. Mo and J. Pan, Acta Biomater., 2022, 151, 235–253.
46. H. Chen, X. Chen, B. Wang, H. Deng, X. Liu, S. Peng, Q. Zhang and Y. Yan, Colloids Surf., A, 2023, 677.
47. P. Yu, X. Peng, H. Sun, Q. Xin, H. Kang, P. Wang, Y. Zhao, X. Xu, G. Zhou, J. Xie and J. Li, Mater. Horiz., 2024, 11, 5352–5365.
48. M. D. Young Cheol Woo, S. Seog Park, A. R. Subieta and M. D. T. J. Brennan, Anesthesiology, 2004, 101, 468–475.
49. Y. Li, C. Liu, Y. Tan, K. Xu, C. Lu and P. Wang, Carbohydr. Polym., 2014, 110, 87–94.
50. S. Liu, M. Lu, M. Zhang, X. Sun, B. Luo and Y. Wu, ACS Nano, 2025, 19, 2677–2694.
51. M. P. Murphy, H. Bayir, V. Belousov, C. J. Chang, K. J. A. Davies, M. J. Davies, T. P. Dick, T. Finkel, H. J. Forman, Y. Janssen-Heininger, D. Gems, V. E. Kagan, B. Kalyanaraman, N. G. Larsson, G. L. Milne, T. Nystrom, H. E. Poulsen, R. Radi, H. Van Remmen, P. T. Schumacker, P. J. Thornalley, S. Toyokuni, C. C. Winterbourn, H. Yin and B. Halliwell, Nat. Metab., 2022, 4, 651–662.
52. D. G. DeNardo and B. Ruffell, Nat. Rev. Immunol., 2019, 19, 369–382.
53. N. Rauner, M. Meuris, S. Dech, J. Godde and J. C. Tiller, Acta Biomater., 2014, 10, 3942–3951.
54. S. Das, S. Singh, J. M. Dowding, S. Oommen, A. Kumar, T. X. Sayle, S. Saraf, C. R. Patra, N. E. Vlahakis, D. C. Sayle, W. T. Self and S. Seal, Biomaterials, 2012, 33, 7746–7755.
55. A. M. Ferreira, P. Gentile, V. Chiono and G. Ciardelli, Acta Biomater., 2012, 8, 3191–3200.

---