Microenvironment responsive nanocomposite hydrogel with NIR photothermal therapy, vascularization and anti-inflammation for diabetic infected wound healing

Abstract

Bacterial infection, excessive inflammation and damaging blood vessels network are the major factors to delay the healing of diabetic ulcer. At present, most of wound repair materials are passive and can't response to the wound microenvironment, resulting in a low utilization of bioactive substances and hence a poor therapeutic effect. Therefore, it's essential to design an intelligent wound dressing responsive to the wound microenvironment to achieve the release of drugs on-demand on the basis of multifunctionality. In this work, metformin-laden CuPDA NPs composite hydrogel (Met@ CuPDA NPs/HG) was fabricated by dynamic phenylborate bonding of gelatin modified by dopamine (Gel-DA), Cu-loaded polydopamine nanoparticles (CuPDA NPs) with hyaluronic acid modified by phenyl boronate acid (HA-PBA), which possessed good injectability, self-healing, adhesive and DPPH scavenging performance. The slow release of metformin was achieved by the interaction with CuPDA NPs, boric groups (B–N coordination) and the constraint of hydrogel network. Metformin had a pH and glucose responsive release behavior to treat different wound microenvironment intelligently. Moreover, CuPDA NPs endowed the hydrogel excellent photothermal responsiveness to kill bacteria of >95% within 10 min and also the slow release of Cu²⁺ to protect wound from infection for a long time. Met@ CuPDA NPs/HG also recruited cells to a certain direction and promoted vascularization by releasing Cu²⁺. More importantly, Met@CuPDA NPs/HG effectively decreased the inflammation by eliminating ROS and inhibiting the activation of NF-κB pathway. Animal experiments demonstrated that Met@CuPDA NPs/HG significantly promoted wound healing of diabetic SD rats by killing bacteria, inhibiting inflammation, improving angiogenesis and accelerating the deposition of ECM and collagen. Therefore, Met@CuPDA NPs/HG had a great application potential for diabetic wound healing.

Graphical abstract

Highlights

• •A metformin-laden CuPDA NPs composite hydrogel (Met@ CuPDA NPs/HG) was fabricated.
• •The hydrogel had a microenvironment response to release metformin intelligently.
• •The hydrogel eradicated bacteria by photothermal responsiveness and the slow release of Cu²⁺.
• •The hydrogel promoted vascularization and decreased the inflammation.
• •Met@ CuPDA NPs/HG had a significant promotion on diabetic infected wound healing.

1. Introduction

Diabetic ulcer is a common complication of diabetes and becomes more severe with the number of diabetes increasing in recent years [1,2]. As a serious chronic wound, diabetic ulcer is usually accompanied by a complex pathological microenvironment, such as high glucose levels, bacterial infection, high oxidative stress, excessive inflammation, damaging blood vessels network and even large defected skins [3], so its treatment remains a big clinical challenge in the worldwide. The fundamental for treating diabetic wounds is to regulate the pathological microenvironment to a normal condition. However, the existing treatment for diabetic ulcer in the clinic was dominated by debridement, negative pressure drainage and traditional dressings [4], which had the disadvantages of complex process and limited clinical efficacy [5]. Owing to the strength on the regulation of microenvironment, lots of biomaterials have been widely developed for the treatment of diabetic ulcer. In order to achieve a better repair of diabetic wound, it's critical to design multifunctional biomaterials according to the feature of diabetic wound.

Compared to other wounds, diabetic wound is more susceptible to infection due to the high levels of glucose in diabetes. Clinical cases show that bacterial infection is a major factor to aggravate diabetic wounds and even amputate, so the materials for diabetic wound with an intrinsic antibacterial property are crucial [6]. Recently, owing to the merits of broad-spectrum and efficient antibacterial property, good controllability, non-invasiveness and prevention of multidrug resistance bacteria, photothermal therapy (PTT) treatment has been used to eradicate bacteria and biofilm efficiently by producing hyperthermia [[7], [8], [9]]. Many inorganic materials (such as gold nanoparticle, graphene, carbon nanotube and transition metal dichalcogenide) and organic materials (small molecular agents such as Indocyanine Green (ICG), IR-825, and polymers such as polypyrrole, polyaniline, dopamine (PDA)) have been reported as PTT biomaterials [[10], [11], [12]]. As a melanin analogue naturally existing in the body [13,14], PDA has the most extensive application for PTT due to the intriguing properties such as simple synthetic method [15,16], good anti-oxidation property [17], adhesive performance [18] and biocompatibility. More importantly, due to the chelation interaction with metal ions, PDA is also a good carrier [19], which can slowly release metal ions into the wound and avoid the biotoxicity due to their excessive usage [20,21]. Recently, bioactive metal ions, such as Cu, Mg, Zn, Ag, have been demonstrated to play a key role on tissue regeneration [[22], [23], [24]]. Our previous work demonstrated Cu²⁺ promoted vascularization by activating HiF-1α to mimic low oxygen microenvironment [25]. Cu²⁺ also increased and stabilized the expression of keratin and collagen to accelerate wound healing [3,26]. Furthermore, Cu²⁺ has a desirable antibacterial efficacy [27,28], which can protect wound from infection for a long time. On the whole, Cu-loaded dopamine nanoparticles (CuPDA NPs) can eradicate bacteria or biofilm quickly and also avoid bacterial infection in the entire repair process, and promote wound healing by releasing Cu²⁺, so it is a potential way to establish multifunctional biomaterials based on CuPDA NPs.

Except from bacterial infection, excessive inflammation is also the primary factor to prolong diabetic wound healing. In general, the healing process of diabetic wounds usually stagnates in inflammation phase, which is mainly regulated by macrophages [29]. Macrophages can't polarize from pro-inflammation phenotype (M1) to pro-regeneration phenotype (M2) [[30], [31], [32]]. Metformin is a commonly oral hypoglycemic drug in the clinic. When applied to local delivery towards the wound, metformin can accelerate the glucose consumption of skins (non-insulin-dependent tissues) and increase the sensitivity of peripheral tissues to insulin, hence reducing the level of glucose at the wound sites [33]. Furthermore, it's reported that metformin could also decrease inflammation and oxidative stress by mediating the phenotypic shift of macrophages [34], reducing the production of mitochondrial reactive oxygen species (ROS) and preventing the activation of inflammation related signal pathways [4,35,36]. Thus, a proper concentration metformin with an effective delivery way is beneficial for the healing of diabetic wound.

At present, different materials are applied to achieve the delivery of drug or bioactive ingredients. However, most of them are passive and don't adapt well with the dynamic process of wound repair. Therefore, it's essential to develop an intelligent material responsive to the wound microenvironment in different phases. Due to the advantages of desirable moisture retention, good biocompatibility, designable physical structure, hydrogels are widely applied as wound dressing [37]. More importantly, hydrogels are also optimum drug carriers and can achieve the intelligent release of drugs by introducing structure which are responsive to different conditions (pH, ROS, Glucose, ect.) [38,39]. As the main components of extracellular matrix (ECM) [40,41], gelatin (Gel) and hyaluronic acid (HA) can simulate wound microenvironment better and provide the scaffold for the adhesion and migration of cells [42,43]. Additionally, HA with high molecular weight can also reduce inflammation [44], beneficial to wound repair. However, the mechanical strength of hydrogels based on gelatin and hyaluronic acid needs to be improved. The graft of reactive groups on the backbone of natural polymer is the direct method to reinforce the mechanical strength of hydrogels. Previous works [[45], [46], [47], [48]] showed that boric acid ester bonds formed by diols and boronic acid have many advantages of simple modifying method, mild reaction condition and good adhesive performance. Besides, boric acid ester bonds have a rapid recovery rate of breakage-formation and can be destroyed by low pH and glucose, thereby responsive to the microenvironment of diabetic ulcer. Moreover, some studies indicated that boronic acid could also form B–N coordination bonds with the drugs containing nitrogen to prolong the release cycle of drugs [[49], [50], [51], [52]]. So, hydrogel crosslinked by boric acid ester bonds can achieve the loading and release of bioactive substances intelligently and improve their utilization.

In this work, we firstly synthesized Cu-loaded polydopamine nanoparticles (CuPDA NPs) to realize the loading of Cu in situ. Afterwards, Metformin and CuPDA NPs were incorporated into an ECM-mimicking hydrogel to obtain a multifunctional nanocomposite hydrogel (Met@CuPDA NPs/HG) (Scheme 1). The hydrogel possessed good injectability, self-healing and adhesive performance, 1-diphenyl-2-picrylhydrazyl (DPPH) clearance and NIR light-responsiveness. Additionally, the release of metformin had good pH and glucose responsiveness, achieving the intelligent treatment of diabetic wound. In addition, owing to the good photothermal effect and the slowly released Cu²⁺, Met@CuPDA NPs/HG had excellent antibacterial performance. More importantly, Met@CuPDA NPs/HG also had good cytocompatibility, anti-oxidative stress, ant-inflammation and angiogenic capacities. Based on above multiple excellent performance, Met@CuPDA NPs/HG was further demonstrated a significant promotion on wound healing in vivo.

Scheme 1.

The preparation of Met@CuPDA NPs/HG hydrogel and its application on diabetic SD rat with full-thickness skin defect and infection.

2. Results and discussion

2.1. The physicochemical properties of PDA NPs and CuPDA NPs

PDA NPs were prepared by the oxidative polymerization and self-assembly of dopamine (DA) under the alkaline condition. In general, the least reaction time to form PDA NPs was 12 h. However, Cu²⁺ was added to form the coordination with groups in PDA including catechol, amino, carboxy, imine etc., which increased the local concentration of the oligomers and accelerated the reaction rate, so the reactive time was reduced to 3 h. SEM images (Fig. 1a) showed that PDA NPs and CuPDA NPs had a regular spherical structure. The analysis of particle size (Fig. 1b) indicated that the particles diameter of PDA NPs and CuPDA NPs was 250 nm approximately and had a small polydispersity index (0.201 for PDA NPs and 0.163 for CuPDA NPs). In addition, zeta potential test (Fig. 1c) showed that PDA NPs had a negative potential in PBS (−35.8 mV) and the surface potential of CuPDA NPs increased to −28.4 mV owing to the introduction of Cu²⁺. After incubating with metformin, the surface potential of CuPDA NPs further increased to −20.6 mV due to the positive potential of metformin, indicating the electrostatic interaction occurred between CuPDA NPs and metformin. EDS (Fig. S2) analysis further demonstrated the successful introduction of Cu, and Table S1 showed that the mass ratio of Cu in CuPDA NPs was 3.71 wt%. XPS (Fig. 1d) illustrated that two peaks at 933 eV and 953 eV attributed to the Cu 2p_3/2 and Cu 2p_1/2 orbital respectively, appeared in the spectrum of CuPDA NPs, but didn't in the spectrum of PDA NPs, suggesting Cu existed with divalence in CuPDA NPs [53]. FT-IR spectra of PDA NPs and CuPDA NPs were displayed in Fig. S1a. Both of PDA NPs and CuPDA NPs had the broad peaks at 1610-550 cm⁻¹, attributed to the deformation and extension of aromatic ring [54]. Due to the chelation with Cu²⁺, the characteristic peaks of the oxygen on catechol groups shifted from 1508 cm⁻¹ to 1471 cm⁻¹. The XRD patterns (Fig. 1e) displayed that both of PDA NPs and CuPDA NPs had an amorphous structure, indicating Cu²⁺ existed with a free state in PDA NPs and could be released effectively. According to above analysis, it was demonstrated that PDA NPs and CuPDA NPs were successfully obtained.

Fig. 1.

(a) Morphology of PDA NPs and CuPDA NPs. (b) Particle diameter distribution of PDA NPs and CuPDA NPs. (c) Zeta potential of PDA NPs, CuPDA NPs and Met@ CuPDA NPs. (d) XPS of PDA NPs and CuPDA NPs. (e) XRD of PDA NPs and CuPDA NPs.

2.2. The structure and morphology of hydrogels

Firstly, HA-PBA and Gel-DA were synthesized, and then the chemical structure before and after modification was tested by FT-IR. FT-IR spectra of HA-PBA (Fig. 2c) showed that the peaks at 1459 cm⁻¹ and 1340 cm⁻¹ were assigned to the benzene ring and B–O respectively, which confirmed the grafting of PBA groups, and the peaks at 1275 cm⁻¹ were ascribed to the C–N stretching in amide band, illustrating the undergoing of the amide reaction. The substitution degree of PBA groups was approximately 28.5%, tested in our previous work [30]. Fig. 2c also exhibited that the characteristic absorption peak of catechol groups on DA appeared at 1282 cm⁻¹ in the spectra of Gel-DA, suggesting the grafting of DA on the backbone of Gel. Furthermore, the FT-IR of the hydrogels was applied to observe the chemical structure change of HA-PBA and Gel-DA in the process of gelation. Fig. 2d displayed that the peaks at 1650 cm⁻¹, 1540 cm⁻¹ and 1287 cm⁻¹ in the spectra of HG respectively represented the absorption peak of amide I band, amide II band and amide III band [43]. Additionally, a weak absorption peak in 1334 cm⁻¹ attributed to asymmetric extension of B–O–C appeared in HG [55], indicating the reaction happened between boric acid groups in HA-PBA and catechol groups in Gel-DA. Furthermore, Fig. S1b illustrated that there was no difference in FT-IR of HG, PDA NPs/HG, CuPDA NPs/HG and Met@CuPDA NPs/HG owing to the overlapping of characteristic peaks.

Fig. 2.

(a) Photograph of HG, 1PDA NPs/HG, 1CuPDA NPs/HG and Met@ 1CuPDA NPs/HG. (b) Pore morphology and surface topography of HG, 1PDA NPs/HG and 1CuPDA NPs/HG. (c, d) FTIR of HA-PBA, Gel-DA and HG. (e) DPPH scavenging of HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, 1CuPDA NPs/HG and 2CuPDA NPs/HG. (f) The release of metformin from Met@1CuPDA NPs/HG at pH = 7.4, pH = 7.4 + 4 g/L Glucose, pH = 5.0 and pH = 5.0 + 4 g/L Glucose. (g) The release of metformin from Met/HG and Met@ 1CuPDA NPs/HG at pH = 7.4. (h) The release of Cu²⁺ from Met@1CuPDA NPs/HG at pH = 7.4.

Fig. 2a showed the appearance of HG, PDA NPs/HG, CuPDA NPs/HG and Met@ CuPDA NPs/HG hydrogels. It was seen that PDA NPs/HG, CuPDA NPs/HG and Met@CuPDA NPs/HG hydrogels had a black appearance due to the adding of PDA NPs and CuPDA NPs. Moreover, SEM was conducted to observe the pore structure of hydrogels, displayed in Fig. 2b. The results showed that PDA NPs/HG had a smaller pore diameter than HG, and CuPDA NPs/HG had the smallest pore diameter. The reasons are the imine and residual catechol groups in PDA NPs could react with –NH₂ groups in Gel and PBA groups in HA-PBA to enhance the crosslinking density. Besides, CuPDA NPs further increased the crosslinking density of hydrogels because of the chelation of Cu²⁺ with DA in Gel-DA and –COOH groups in HA-PBA. In the meantime, it also could be found from Fig. 2b that the surface of HG was smooth, and but that of PDA NPs/HG and CuPDA NPs/HG were rough, and PDA NPs and CuPDA NPs were evenly dispersed in hydrogels.

2.3. The release behavior of metformin and Cu²⁺

Phenyl boronate acid groups can form boronate ester bonds with orthodiol groups in glucose, which forms a competitive relation with the dopamine groups in Gel-DA and thus destroys the hydrogels. Additionally, boronate ester bonds are easily broke under the acid condition. So, as the main structure of the hydrogels, boronate ester bonds have a strong sensitivity to pH and glucose levels [45], and the release rate of metformin greatly depends on the microenvironment. Diabetic wounds are usually at a high glucose levels and weak acid condition for a long time, so it's essential to study the release of metformin under the high glucose levels and weak acid solution. Fig. 2f indicated that Met@CuPDA NPs/HG hydrogels had the least release of metformin at pH = 7.4 (44% at 24 h). The weak acid or high glucose levels strikingly accelerated the release rate (52% and 50% at 24 h respectively). Besides, under the condition of weak acid and high glucose levels, metformin had the quickest release with the release ratio of 57% at 24 h. Therefore, applied to diabetic wounds, Met@CuPDA NPs/HG has a rapid release of metformin due to the high glucose levels and low pH at the wound sites. As the wound repair proceeds, glucose level and pH at the wound sites gradually recover to a normal condition, and the release of metformin becomes slower. In short, metformin has a responsive release based on wound microenvironment to perform its anti-inflammation and glucose-reducing efficiently, achieving the intelligent treatment of diabetic wound.

Meanwhile, we also investigated the release rate of Met/HG and Met@CuPDA NPs/HG at pH = 7.4 to study the loading of CuPDA NPs to metformin, exhibited in Fig. 2g. It was found that metformin in Met@CuPDA NPs/HG had a slower release than that in Met/HG, suggesting CuPDA NPs could realize the loading of metformin by forming interaction (schiff base, hydrogen bonds or electrostatic interactions) with CuPDA NPs and hence prolong the release cycle of metformin. The zeta potential change of CuPDA NPs can also confirm this conclusion after incubating with metformin. It could be figured out from Fig. 2h that Cu²⁺ was slowly released from Met@CuPDA NPs/HG to play its bioactivity better.

2.4. DPPH scavenging activity

Excessive free radicals at the chronic wound sites usually lead to a strong oxidative stress and even large skin defect, delaying wound healing. Thus, wound dressings with the property of free radical elimination are advantageous for wound healing. DPPH is a stable nitrogen-centered chromogenic radical, and is widely applied to test the radical elimination performance of materials. The DPPH scavenging ability of the hydrogels was showed in Fig. 2e. It was figured out that HG hydrogel had a clearance ratio of 56.54% on DPPH, resulting from DA in Gel-DA. Compared to HG, PDA NPs/HG, CuPDA NPs/HG and Met@CuPDA NPs/HG had a higher eliminative effect on DPPH (>85%) due to the introduction of PDA NPs and CuPDA NPs. Thus, the hydrogels had a good DPPH scavenging activity, originating from the synergetic effect of DA groups in Gel-DA and PDA NPs.

2.5. Injectability, self-healing and rheology behavior

The hydrogels were injected out from a 16G needle (Fig. 3a) to form the desirable shape like “SCUT”, which could be explained by the typical shear-thinning behavior (Fig. 3e). The self-healing property can avoid the ineffectiveness of hydrogel dressing for joint wounds such as ankles, wrists etc. Due to its structure breakage. Firstly, two pieces of hydrogels were placed together. After contacting for 1 min, the recombined hydrogel didn't fracture after bearing its weight and even pulling towards to the two sides, suggesting an excellent self-healing performance. Furthermore, the rheology test was performed to observe the self-healing behavior of the hydrogels from a micro perspective. The transition of sol-gel was firstly determined by the strain amplitude sweep test. Fig. 3f indicated that the intersection of the storage modulus (G′) and loss modulus (G′′) appeared at 700% approximately for both of HG and CuPDA NPs/HG. Besides, it also could be seen that CuPDA NPs/HG had a larger modulus than HG, further confirming that CuPDA NPs increased the crosslinking density of the hydrogel. On the basis of the strain amplitude sweep test, a continuous step-strain sweep at 1% and 1000% was conducted to assess the self-healing property of HG and CuPDA NPs/HG (Fig. 2g). Under the alternative of the high and low strain for 3 times, the modulus of the hydrogels recovered with no loss within 60 s, indicating a good self-healing behavior. This could be explained by the dynamic network of the hydrogel. In general, the hydrogels owned a good injectability and self-healing performance.

Fig. 3.

(a) Image of injectability, self-healing of 1CuPDA NPs/HG and adhesive performance of 1CuPDA NPs/HG to human's fingers, porcine skin, plastic, glass and metal. (b) Photograph of test device for adhesive strength (c, d) Adhesive strength of HG, 1PDA NPs/HG and 1CuPDA NPs/HG to porcine skin. (e) The shear-thinning behavior test of HG and 1CuPDA NPs/HG (f) Strain amplitude sweep test of HG and 1CuPDA NPs/HG. (g) Self-healing performance of HG and 1CuPDA NPs/HG under the alternative strain of 1% and 1000%.

2.6. Adhesive performance

A good adhesive performance can avoid dressing from falling off wounds, especially for wounds in the joint, and play their performance for a long time. The adhesive property was assessed quantitatively and qualitatively, displayed in Fig. 3a. The hydrogels could adhere on the joint of human fingers and porcine skin, and didn't fall after subjecting to repetitive stretching and bending. Besides, the hydrogels adhered on and lifted the big vitreous bottle, centrifuge tube and tweezer, indicating a good adhesive performance to glass, plastic and metal substrates. Subsequently, lap shear test (Fig. 3b) was conducted to quantize the adhesive strength of the hydrogels to porcine skin. The result is showed in Fig. 3c and d. The adhesive strength of HG, PDA NPs/HG and CuPDA NPs/HG was 20.13, 26.08 and 30.01 KPa. The reasons for a good wet adhesive performance are as followed. Firstly, the hydrogels are consisted of hyaluronic acid and gelatin to quickly absorb the water in the porcine skin. Subsequently, dopamine forms the physical and chemical interaction (such as hydrogen bonds, schiff base bonds etc.) with porcine skin to achieve a good adhesive performance. Moreover, PDA NPs/HG and CuPDA NPs/HG had a higher adhesive strength than HG, attributed to the introduction of PDA NPs and CuPDA NPs. Noticeably, the adhesive strength of CuPDA NPs/HG to porcine skin was higher than that of PDA NPs/HG, which could be explained that Cu²⁺ protected catechol groups from oxidation and also enhanced the cohesive force of the hydrogel by chelation.

2.7. The NIR photothermal responsiveness

Hydrogels with good photothermal conversion efficiency can generate hyperthermia to eradicate the bacterial and even biofilms to prevent wound infection. Furthermore, the elimination of bacteria or biofilms is beneficial to perform the bioactivity of the drugs or bioactive metals subsequently. Thus, the photothermal responsiveness of hydrogels was tested by exposing to near-infrared (NIR) laser with the intensity of 0.5 W/cm² or 1 W/cm². The results are showed in Fig. 4. As Fig. 4a displayed, HG had no evident change on temperature after irradiating with NIR laser of 0.5 W/cm² for 10 min. Compared to HG, the hydrogels with PDA and CuPDA NPs had a striking temperature variation, and the highest temperature increased from 45.8 °C to 59.6 °C with the concentration of CuPDA NPs from 1 mg/mL to 2 mg/mL. Fig. 4b is the temperature change curve of the hydrogels. It was evident that the photothermal responsiveness of the hydrogels was adjustable and raised with the increase of laser intensity. However, when laser intensity increased to 1 W/cm², the temperature of the hydrogels raised to above 60 °C. As 55 °C is the optimal temperature to kill bacteria with no damage to cells, 1 mg/mL PDA NPs and CuPDA NPs was chosen for the preparation of the hydrogels and the laser intensity of 0.5 W/cm² was applied for the subsequent experiment. Moreover, the photothermal effect of the hydrogels was measured for 4 times to evaluate the recyclability. Fig. 4c displayed that the highest temperature of CuPDA NPs/HG have a little change after 4 cycles, indicating a stable recyclable NIR responsiveness. Taken together, the hydrogels possessed a good NIR responsiveness and thus had a great potential to prevent wound from bacterial infection.

Fig. 4.

(a) Thermal imaging photograph of HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, 1CuPDA NPs/HG and 2CuPDA NPs/HG after irradiating with NIR for 10 min. (b) Temperature change of HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, 1CuPDA NPs/HG and 2CuPDA NPs/HG after irradiating with NIR for 10 min. (c) Temperature change of 1CuPDA NPs/HG after repeatedly irradiating and cooling for 4 times.

2.8. In vitro antibacterial property

Wound infection usually leads to a great many of complications, and is also the main reason to cause the amputation of patients with diabetic ulcers in the clinic. So, hydrogels with good antibacterial performance are very essential for wound healing, especially for diabetic wound. Thus, the antibacterial performance of the hydrogels was assessed by the spread-plate method after incubating with Escherichia coli (E. coli), Staphylococcus aureus (S.aureus) and multiple-resistant Staphylococcus aureus (MRSA), and the results were exhibited in Fig. 5. Firstly, the antibacterial activity of the hydrogels after exposure to NIR for 10 min was studied. It's found from Fig. 5a that the hydrogel with PDA NPs or CuPDA NPs had a few bacterial colonies and but a great many of bacterial colonies existed in the blank and HG groups, indicating that the hydrogels with PDA NPs or CuPDA NPs had a good antibacterial performance owing to the generation of hyperthermia. The quantitative results (Fig. 5b) showed that the hydrogels with PDA NPs and CuPDA NPs had a killing bacteria rate of above 95% and nevertheless HG had almost no bactericidal effect, further demonstrating the mentioned conclusion.

Fig. 5.

(a) Photo of survival E. coli, S.aureus and MRSA after incubation with HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, and 1CuPDA NPs/HG with irradiating with NIR for 10 min. (b) The antibacterial rate of HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, and 1CuPDA NPs/HG with NIR irradiating for 10 min. (c) Photo of survival E. coli, S.aureus and MRSA after incubation with HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, and 1CuPDA NPs/HG for 2 h without NIR irradiation. (d) The antibacterial rate of HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, and 1CuPDA NPs/HG without NIR irradiating.

The antibacterial performance from the photothermal effect usually is efficient and but temporary. However, due to the open wound microenvironment, it's essential for hydrogel dressing to provide a persistent antibacterial property for the wounds. In our work, Cu²⁺ was slowly released from the hydrogels to overcome the above problem and thus the antibacterial activity of Cu²⁺ was further evaluated. It could be figured out from Fig. 5c that CuPDA NPs/HG group had fewer bacterial colonies compared to the Blank, HG and PDA NPs/HG groups, and the bacteria colonies decreased with the increase of CuPDA NPs from 0.5 mg/mL to 1 mg/mL. Besides, Fig. 5d further confirmed that the introduction of Cu²⁺ endowed the hydrogels a striking antibacterial activity. And the antibacterial rate of 1 mg/mL CuPDA NPs/HG was approximately 86% for MRSA and even above 91% for E. coli and S.aureus. Therefore, the CuPDA NPs/HG protected wound from infection by NIR thermal-responsiveness to efficiently eradicate existing bacterial and biofilms, and the slow release of Cu²⁺ to provide antibacterial performance for a long time.

2.9. Cytocompatibility

A good biocompatibility is the prerequisite of hydrogels to be applied in vivo, so cells compatibility of the hydrogels is assessed by LIVE/DEAD® Viability/Cytotoxicity Kit assay, and the densities of fibroblast cells (L929s) and human umbilical vein endothelial cells (HUVECs) after co-cultivation with the hydrogels for 1, 2 and 3 days are showed in Fig. 6. For L929s, there was no significant difference on cells density on day 1. However, on day 2, the introduction of PDA NPs and CuPDA NPs negatively affected the cells growth. Noticeably, it was easy to find that 1 mg/mL CuPDA NPs/HG group had a promotive effect on cells proliferation compared to other hydrogels groups. After incubating for 3 days, it was figured out that the cells density had a tendency of increased firstly and then decreased with the increase of the content of CuPDA NPs, attributed to the toxicity of excessive Cu²⁺. Moreover, the cells number in 1 mg/mL CuPDA NPs/HG group was more than that in the other groups, illustrating that an appropriate concentration of Cu²⁺ had a positive effect on cells proliferation. Similarly, 1 mg/mL CuPDA NPs/HG had the best effect to promote HUVECs proliferation. Live/Dead staining result of HUVECs is in accordance with that of Cytotoxicity Kit assay, further confirming this conclusion. Therefore, 1 mg/mL CuPDA NPs/HG had the best cytocompatibility and was chosen for subsequent experiments.

Fig. 6.

(a) The viability of L929s after co-cultivation with HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, 1CuPDA NPs/HG, 2 CuPDA NPs/HG and Met@1CuPDA NPs/HG for 1, 2 and 3 days. (b) The viability of HUVECs after co-cultivation with HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, 1CuPDA NPs/HG, 2 CuPDA NPs/HG and Met@1CuPDA NPs/HG for 1, 2 and 3 days. (c) Live/Dead staining images of HUVECs after co-incubating with HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, 1CuPDA NPs/HG, 2CuPDA NPs/HG and Met@1CuPDA NPs/HG for 3 days.

2.10. Cells recruitment property

A severe chronic wound is usually accompanied by a large skin defect, so the recruitment of healing-related cells from peripheral tissues into the wound is beneficial to accelerate wound repair. L929s and HUVECs play a critical role on ECM, collagen deposition and vascularization respectively, so the recruitment of the hydrogels to L929s and HUVECs was performed and the results were displayed in Fig. 7. It could be seen that the recruitment ability of the hydrogels to L929s (Fig. 7a) and HUVECs (Fig. 7b) were similar. There were very few cells recruited into the lower chamber in the blank, HG and PDA NPs/HG groups. In contrast, the hydrogels with CuPDA NPs had more cells in the lower chamber and the number of cells increased with the concentration of CuPDA NPs increasing, suggesting Cu²⁺ provided a stimulation on cells to migrate into a particular direction. Besides, metformin didn't affect cell recruitment of CuPDA NPs/HG. The quantitative results (Fig. 7c and d) further solidified the conclusion. Thus, CuPDA NPs/HG could recruit L929s and HUVECs into the wound by the release of Cu²⁺, facilitating wound healing.

Fig. 7.

(a, b) Images of L929s and HUVECs recruited from the upper chamber to the lower chamber after co-incubating with HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, 1CuPDA NPs/HG and Met@1CuPDA NPs/HG for 12 h. (c, d) The quantitative data of L929s and HUVECs in the lower chamber.

2.11. Angiogenic performance

In the process of skin repair, lots of nutrients are needed, so the reconstruction of blood vessels network in the early is crucial for wound healing. HUVECs are the key cells to rebuild blood vessel network. Cells recruitment experiment confirmed that Cu²⁺ could recruit HUVECs into the wound to facilitate vascularization. Furthermore, the regulation of the hydrogels to the expression of proteins and genes in HUVECs was also investigated. CD31 and VE-cadherin are the marked proteins expressed by HUVECs in the process of angiogenesis. After co-cultivating with the hydrogels for 3 days, CD31 and VE-cadherin in HUVECs were stained, showed in Fig. 8a. It was evident that there was little expression of CD31 in the Blank, HG and PDA NPs/HG groups. Nevertheless, due to the release of Cu²⁺, a remarkable expression of CD31 emerged in 1CuPDA NPs/HG and Met@1CuPDA NPs/HG groups, indicating a good vascularization. And metformin had little effect on the expression of CD31. Similar results are observed in the expression of VE-cadherin in HUVECs. In consequence, 1CuPDA NPs/HG could enhance the expression of CD31 and VE-cadherin by the release of Cu²⁺ and thus promote the vascularization.

Fig. 8.

(a) Immunofluorescence staining images of CD31 and VE-cadherin in HUVECs after co-cultivating with HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, 1CuPDA NPs/HG and Met@1CuPDA NPs/HG for 3 days. (b) The expression quantity of angiogenic related genes (Angiogenin, eNOS, VEGF, HIF-1α and KDR) in HUVECs after co-cultivating with HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, 1CuPDA NPs/HG and Met@1CuPDA NPs/HG for 3 days.

Afterwards, RT-PCR was conducted to quantify the expression of angiogenic related genes (VEGF, KDR, eNOs, Angiogenin, HiF-1α). As Fig. 8b exhibited, 1CuPDA NPs/HG and Met@1CuPDA NPs/HG groups had a higher expression of VEGF, KDR, eNOs, Angiogenin and HiF-1α than the other groups, further demonstrating angiogenic performance of Cu²⁺. In the meantime, metformin didn't adversely affect on the expression of vascular genes. Hypoxia-inducible factor-1α (HIF-1α) is an important transcription factor in hypoxia and angiogenesis. Our previous work [25]confirmed that Cu²⁺ promoted the expression of vascular related genes by activating HIF-1α signaling pathway and thereby facilitated the rebuild of vessels network. In this work, the same conclusion was obtained.

2.12. The elimination activity of intracellular ROS

Owing to the adverse microenvironment (bacterial infection and high glucose levels) of diabetic wound, excessive ROS are expressed at the wound sites, which has a destructive effect on cells or tissues and causes a strong oxidative stress. The strong oxidative stress hinders the reconstruction of blood vessels network by damaging HUVECs. Thus, HUVECs were treated by 100 μM H₂O₂ to mimic oxidative stress in vitro and then intracellular ROS expression in HUVECs was stained by DCFH-DA to detect the ROS elimination of the hydrogels. The results are displayed in Fig. 9a. The positive control group had a stronger ROS expression than the negative control group, illustrating the successful induction of oxidative stress. In contrast, HG group had a weaker ROS expression than the positive group, originated from the ROS scavenging property of DA groups on Gel-DA. Owing to the ROS scavenging property of PDA NPs, the intensity of ROS expression in 1PDA NPs/HG groups became less. It was worth noting that CuPDA NPs/HG group had a more expression of ROS than HG group, maybe attributed to the pro-inflammation of Cu²⁺. In addition, Met@1CuPDA NPs/HG group had the least ROS expression, and the intensity of ROS expression was similar and even weaker than that in negative control group, indicating the best ROS elimination activity. It could be explained by the inhibition of metformin to mitochondrial ROS [36]. Thus, Met@1CuPDA NPs/HG had an excellent ROS eliminative property owing to the synergetic effect of Gel-DA and Metformin.

Fig. 9.

(a) The expression of ROS in HUVECs after co-cultivating with HG, 1PDA NPs/HG, 0.5CuPDA NPs/HG, 1CuPDA NPs/HG and Met@1CuPDA NPs/HG for 6 h under the condition of 100 μM H₂O₂. (b) The immunofluorescence staining images of p65 proteins in RAW 264.7 after inducing by LPS for 8 h and then co-cultivating with HG, 1PDA NPs/HG, 1CuPDA NPs/HG and Met@1CuPDA NPs/HG for 3d. (c) The expression quantity of pro-inflammation genes (TNF-α, iNOS and IL-6) and anti-inflammation genes (IL-10 and Arginase) in RAW 264.7 after inducing by LPS for 8 h and then co-cultivating with HG, 1PDA NPs/HG, 1CuPDA NPs/HG and Met@1CuPDA NPs/HG for 3d.

2.13. Anti-inflammation performance

The healing of diabetic ulcers usually stagnates in inflammation stage which is mediated by macrophages. The phenotypes transformation of macrophages from M1(pro-inflammation phenotype) to M2 (pro-proliferation phenotype) is key for promoting wound healing from inflammation stage to proliferation stage. Therefore, we assessed anti-inflammation property of the hydrogels by testing the expression of related protein and genes in macrophages (RAW264.7) and the results were showed in Fig. 9b and c. NF-κB proteins are usually expressed in the extracellular matrix (ECM), and but transferred into cell nucleus once NF-κB signal pathway is activated. p65 protein is a main component of NF-κB protein, so the location of p65 protein in macrophages can be used to detect if NF-κB signaling pathway is activated. Firstly, macrophages were treated by lipopolysaccharide (LPS) for 8 h to activate NF-κB signaling pathway. Fig. 9b showed that lots of p65 proteins were expressed in ECM for the negative control group, and but in cell nucleus for the positive control group, indicating the activation of NF-κB signaling pathway and the successful induction of inflammation. After incubation with the hydrogels for 3d, p65 protein in the nucleus significantly decreased. The inhibition of HG to nucleus transfer of p65 proteins derived from hyaluronic acid, which inhibited the activation of the TLR4 signaling pathway and further prevented the activation of downstream NF-κB signaling pathway [56,57]. Noticeably, due to the pro-inflammation of Cu²⁺ [25,58], 1CuPDA NPs/HG group had more expression of p65 protein in the cell nucleus than 1PDA NPs/HG group. Nevertheless, after the introduction of metformin, Met@1CuPDA NPs had the least nucleus transfer of p65 proteins, attributed to the excellent inhibition of metformin to NF-κB signaling pathway. Thus, Met@1CuPDA NPs had the best anti-inflammation activity by the synergetic effect of hyaluronic acid and metformin.

Subsequently, RT-PCR was preformed to quantify the expression of pro-inflammatory genes (TNF-α, iNOS and IL-6) and anti-inflammatory genes (IL-10, Arginase). It was figured out from Fig. 9c that the expression of TNF-α, iNOS and IL-6 in the positive control group significantly increased, indicating the successful polarization of macrophages to M1 phenotype. After co-cultivation with the hydrogels for 3d, TNF-α, iNOS and IL-6 had a less expression than the positive control group. More importantly, Met@1CuPDA NPs had the least expression of TNF-α, iNOS and IL-6, and even the expression of IL-6 was less than that in the negative control group. Besides, the expression of anti-inflammatory genes (IL-10 and Arginase), especially for Arginase, in HG and 1PDA NPs/HG groups was more than that in the positive control group, ascribed to the anti-inflammation of hyaluronic acid. 1CuPDA NPs/HG group had a lower expression of IL-10 and Arginase than the positive control group due to the introduction of Cu²⁺. However, the expression of IL-10 and Arginase in Met@1CuPDA NPs/HG group was strikingly higher than that in the positive control group, confirmed the excellent anti-inflammatory effect of metformin, in accordance with the immunofluorescence staining result of p65 protein. Thus, the above results demonstrated that Met@1CuPDA NPs/HG achieved a good anti-inflammation by inhibiting the activation of NF-κB signaling pathway.

2.14. In vivo diabetic wound healing

Full-thickness skin defects infection model of diabetic SD rats was applied to assess the repair effect of the hydrogels in vivo. HG, 1PDA NPs/HG and Met@1CuPDA NPs/HG were set as the experimental groups and the Blank group was the wounds without any treatment. The wounds were photographed on day 0, 3, 7 and 14, displayed in Fig. 10a. The quantitative results (Fig. 10b) were analyzed to further evaluate wound closure rate and Fig. 10c visualized the process of wound shrinkage. It's obvious that the hydrogels matched and adhered on the wounds well. Fig. S3 showed that PDA NPs/HG and Met@1CuPDA NPs/HG had a striking temperature rise within 10min after exposure to 808 nm NIR (0.5W/cm²), and but the Blank and HG groups didn't, suggesting the introduction of PDA NPs or CuPDA NPs/HG endowed the hydrogels a good photothermal responsiveness in vivo. Furthermore, the wounds in the Blank and HG groups had an evident abscess on day 3, illustrating a serious wound infection. However, 1PDA NPs/HG and Met@1CuPDA NPs/HG groups didn't have an obvious signal of wound infection, owing to the excellent antibacterial performance from the hyperthermia generated by NIR photothermal responsiveness and the release of Cu²⁺. Besides, the wounds had a significant shrinkage in the 1PDA NPs/HG (about 37%) and Met@1CuPDA NPs/HG (about 59%). On day 7, all groups had an evident wound closure. The size of wounds in Met@1CuPDA NPs/HG groups was smaller (approximately 11%) than the other groups and that in the Blank group was the largest (approximately 37%). On day 14, the wounds in Met@1CuPDA NPs/HG group closed completely, and but Blank, HG and 1PDA NPs/HG groups still had a wound residual rate of 26%, 22% and 12% respectively. So, it was concluded that Met@1CuPDA NPs/HG avoided wound infection and accelerated wound closure.

Fig. 10.

(a) Photographs of wounds in diabetic SD rats with full-thickness skin defect and infection after treatment by HG, 1PDA NPs/HG and Met@1CuPDA NPs/HG for 0, 3, 7, 14 days. (b) The wounds residual rate after treatment by HG, 1PDA NPs/HG and Met@1CuPDA NPs/HG for 0, 3, 7, 14 days. (c) Visual images of wounds shrinkage process.

A rapid wound closure doesn't mean a good repair of skin tissues, so it's essential to evaluate the quality of newborn skin tissues by histological staining. Firstly, Hematoxylin and eosin (H&E) staining was conducted to observe the deposition of extracellular matrix (ECM), exhibited in Fig. 11a. On day 3, Met@1CuPDA NPs/HG group had a large amount of ECM deposition. However, the quantity of ECM in Blank, HG and 1PDA NPs/HG groups was very little. On day 7, the Blank and HG groups had an uneven ECM deposition, and there was the deposition of sparse ECM in the partial areas. In contrast, ECM deposition in 1PDA NPs/HG and Met@1CuPDA NPs/HG groups was even and dense, and Met@1CuPDA NPs/HG group had the densest deposition of ECM. On day 14, it was obviously seen that re-epithelization appeared in HG, 1PDA NPs/HG and Met@1CuPDA NPs/HG groups and but didn't in the blank group. Fig. S4a displayed the quantitative results of ECM deposition. The results showed that Met@1CuPDA NPs/HG group had the most ECM deposition. Besides, compared to other groups, Met@1CuPDA NPs/HG group had the regeneration of skin appendages (hair follicles and sebaceous glands), indicating the best healing effect.

Fig. 11.

(a) Images of H&E staining. The green arrows represent skin appendages such as hair follicles, sebaceous glands etc. (b) Images of Masson trichrome staining.

Further, Masson trichrome staining (MTS) was used to observe the effect of collagen deposition and the results were showed in Fig. 11b. With the time extending, the blue in all groups became more, indicating the collagen deposition increased. Among them, Met@1CuPDA NPs/HG group had a significant collagen deposition on day 3 and but the Blank group had a little. On day 7, collagen deposition in the Blank group was very sparse. By contrast, collagen deposition in HG, 1PDA NPs/HG and Met@1CuPDA NPs/HG groups was denser and that in Met@1CuPDA NPs/HG was densest. On day 14, collagen in the Blank group was loose and disorganized, and that in HG and 1PDA NPs/HG groups was denser and more regular than the blank group. Noticeably, collagen in Met@1CuPDA NPs/HG group was with a parallel arrangement, suggesting more mature collagen fibers. The quantitative result is exhibited in Fig. S4b and further confirms the above analysis. Therefore, Met@1CuPDA NPs/HG promoted wound healing with high quality by accelerating the deposition of ECM and collagen.

Due to the key role of blood vessels on wound healing, immunohistochemical staining of CD31 was applied to evaluate the regeneration of blood vessels on day 7, showed in Fig. 12a. It's obvious that the Blank and HG groups had the regeneration of many small blood vessels on day 7. Nevertheless, the blood vessels in Met@1CuPDA NPs/HG group were bigger and more mature than these in other groups. Furthermore, the quantitative data (Fig. S4c) indicated that the number of blood vessels in Met@1CuPDA NPs/HG group was more than that in other groups. So, Met@1CuPDA NPs/HG had a significant promotion on vascularization to accelerate wound healing.

Fig. 12.

(a, b, c) Immunohistochemical staining images of CD31(a), iNOS (b) and Arginase (c).

Additionally, immunohistochemical staining of iNOS and Arginase was exerted to observe the inflammation of newborn skin tissues on day 3 and 7, exhibited in Fig. 12b and c. On day 3, the Blank group had a strong expression of iNOS and a weak expression of Arginase, suggesting the wound was in inflammation. Compared to the Blank group, the expression of iNOS in HG group was weaker and the expression of Arginase was stronger, attributing to the anti-inflammation of hyaluronic acid and the elimination of DA to ROS. Moreover, due to the good scavenging ability of PDA NPs to ROS, 1PDA NPs/HG group had a lower inflammation than HG group on day 3, embodied in a lower expression of iNOS and a higher expression of Arginase. Noticeably, Met@1CuPDA NPs/HG group had the least expression of iNOS and the strongest expression of Arginase, illustrating a weak inflammation. After treatment for 7 d, there was still marked expression of iNOS in the Blank group and but the expression of Arginase increased in contrast to that on 3d, implying an alleviative inflammation. The expression of iNOS in all hydrogel groups had a striking decrease. The expression of Arginase in HG and 1PDA NPs/HG groups was still high. It's interesting that both of iNOS and Arginase expression in Met@1CuPDA NPs/HG group were very little. The reason for this phenomenon is that the number of macrophages decreases and the process of wound repair switches from the inflammation stage to the proliferation stage. The statistic results of iNOS (Fig. S4d) and Arginase (Fig. S4e) expression in newborn wound tissues illustrated the same conclusion. Thus, Met@1CuPDA NPs/HG decreased the inflammation and facilitated the process of wound repair to proliferation stage, which accelerated wound healing.

3. Conclusions

In summary, Met@ CuPDA NPs/HG hydrogel with a good injectability, self-healing, adhesive and DPPH scavenging performance was fabricated, and the active components, CuPDA NPs and metformin, achieved antibacterial, anti-inflammatory and angiogenic performance to promote wound healing. Metformin had a pH and glucose responsive release to treat different diabetic wound intelligently and improve the utilization of metformin. Moreover, the efficient photothermal responsiveness of Met@1CuPDA NPs/HG generated hyperthermia to kill bacteria of >95% for 10 min, and the release of Cu²⁺ from CuPDA NPs also had a good antibacterial rate (>85%) to avoid wound infection for long time. Cu²⁺ also recruited L929s and HUVECs into wound and promoted angiogenesis to accelerate wound healing. DA groups and CuPDA NPs eliminated excessive ROS. Besides, metformin and hyaluronic acid inhibited the activation of NF-κB signaling pathway, which hence greatly decreased the inflammation. In vivo animal experiment further demonstrated that Met@ CuPDA NPs/HG facilitated diabetic wound healing by avoiding wound infection, decreasing inflammation, accelerating the formation of blood vessels and the deposition of ECM and collagen. Therefore, Met@ CuPDA NPs/HG provides an alternative for the treatment of diabetic wound.

4. Experimental section

4.1. Materials

Sodium hyaluronic acid (HA) (100–150 kDa) was provided by Shanghai Yuanye Bio-Technology Co., Ltd (Shanghai, China). 2-morpholinoethanesulfonic acid (MES), Dopamine, 4-(4,6-Dimethoxy-1,3,5-triazin-2-yl)-4-methylmorpholinium chloride (DMTMM), CuCl₂·2H₂O, gelatin and streptozotocin (STZ) were purchased from Shanghai Aladdin Industrial Co., Ltd (Shanghai, China). Luria-Bertani agar (LB agar) and Luria-Bertani broth (LB broth) were purchased from Guangzhou HuanKai Biology Technology Co., LTD. 4% formaldehyde was obtained from Beijing Lanjieke Technology Co., LTD. All chemical reagents above were analytical grade. Tris(hydroxymethyl)methyl aminomethane (Tris), 1, 1-diphenyl-2-picrylhydrazyl (DPPH), N-hydroxysuccinimide (NHS) and 1-(3-Dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDC) were obtained from Shanghai Macklin Biochemical Co., Ltd (Shanghai, China).

4.2. Preparation of polydopamine nanoparticles PDA NPs and Cu-loaded PDA NPs

PDA NPs were prepared by the oxidative polymerization and self-assembly of dopamine (DA) under the alkaline condition. Cu²⁺ was added to chelate with the active groups (-NH₂, –COOH, catechol groups etc.) in PDA to prepare CuPDA NPs. The specific process was as followed. DA was dissolved in deionized water with a concentration of 1% w/v. CuCl₂·2H₂O was added into the mixtures and stirred for 1 h to fully chelate with DA. Subsequently, 1% tris solution was poured into the mixture and then constantly stirred for another 2 h. After that, the crude product was collected by centrifuging and washed with deionized for three times. Finally, the product was obtained after drying at 50 °C for 12 h in a vacuum drying oven. Similarly, DA was added into 1% tris solution and stirred for 12 h at room temperature to obtain PDA NPs. The morphology was observed by scanning electron microscope (SEM) and the content of elements was measured by Energy Disperse Spectroscopy (EDS) analysis.

4.3. The synthesis of hyaluronic acid modified by phenyl boronate acid (HA-PBA) and gelatin modified by dopamine (Gel-DA)

HA-PBA was synthesized according to our previous work [30]. Gel-DA was synthesized by amidation reaction between –COOH in gelatin and –NH₂ in dopamine hydrochloride. In brief, 1g gelatin was dissolved in 100 mL MES buffer solution (50 mM) at 50 °C and then the mixtures were cooled to room temperature. After that, a certain of EDC and NHS were added and fully dissolved into the mixtures to activate –COOH in gelatin. Subsequently, dopamine hydrochloride was added and stirred for 12 h at room temperature. Finally, the mixtures were loaded into dialyzed bags (MWCO 8000–14000) to remove the residues in deionized water with a certain hydrochloric acid (pH = 5) for 3d and were lyophilized to obtain Gel-DA. ¹HNMR and FT-IR were applied to characterize the chemical structure of Gel-DA.

4.4. The preparation of hydrogels

HA-PBA was dissolved in PBS with a concentration of 3% w/w. Afterwards, Gel-DA was dissolved in PBS with a concentration of 4% w/w, and then 0.5 M NaOH solution was added to adjust pH of Gel-DA solution to 7.4. Next, Gel-DA solution was poured into HA-PBA solution with the volume ratio of 1:1 under stirring. Hydrogel formed within 1 min, named as HG.

Similarly, after sonicating for 30 min, PDA NPs or CuPDA NPs was evenly dispersed in PBS according to the concentration of 1 mg/mL. HA-PBA was dissolved in the above mixtures, and then mixed with Gel-DA solution to obtain hydrogels, named as 1PDA NPs/HG or 1CuPDA NPs/HG respectively. And hydrogels with 0.5 and 2 mg/mL CuPDA NPs were named as 0.5CuPDA NPs/HG and 2CuPDA NPs/HG respectively. Noticeably, CuPDA NPs were evenly dispersed in PBS with 1 mg/mL metformin by sonicating for 30 min and then incubated for 1 h in constant temperature shaker (37 °C, 120 rpm) to make metformin fully interact with CuPDA NPs. Afterwards, HA-PBA was dissolved in the above mixtures and mixed with Gel-DA solution to get Met@CuPDA NPs/HG hydrogel. The hydrogels were lyophilized to observe the pore structure and the dispersity of CuPDA NP by SEM and analyze the chemical structure by FT-IR spectra.

4.5. The pH and glucose responsive release behavior of metformin from the Met@CuPDA NPs/HG and Met@HG

Met@CuPDA NPs/HG and Met@HG were chosen to test the release of metformin. Met@HG was prepared by mixing HA-PBA solution with 1 mg/mL metformin and Gel-DA solution. The steps are as followed. Firstly, 1 g Met@CuPDA NPs/HG and 1 g Met@HG were immersed into 20 mL citric acid-Na₂HPO₄ buffer solution (pH = 7.4), and then a certain volume of the supernatant was taken out at the different time. Meanwhile, the same volume buffer solution was supplemented. In the end, the release concentration of metformin was tested by the ultraviolet and visible Spectroscopy (UV-VIS). Further, according to the same process, Met@CuPDA NPs/HG was immersed in pH = 5.5 citric acid-Na₂HPO₄ buffer solution, pH = 7.4 citric acid-Na₂HPO₄ buffer solution with 4 g/L glucose and pH = 5.5 citric acid-Na₂HPO₄ buffer solution with 4 g/L glucose to test the release of metformin under acid or high glucose levels or acid and high glucose levels.

4.6. The release behavior of Cu²⁺ from Met@CuPDA NPs/HG

According to the same process with testing the release of metformin, the release concentration of Cu²⁺ was tested by inductively coupled plasma optical emission spectrometer (ICP-OES, PerkinElmer PinAAcle 900T, Germany).

4.7. The scavenging of DPPH

DPPH was dissolved in absolute ethanol according to the concentration of 0.1 mmol/L. Afterwards, 0.1 g hydrogels were immersed in 1 mL DPPH solution and then incubated for 1 h in a constant temperature shaker (37 °C, 120 rpm). In the end, the absorbance of supernatant at 517 nm was tested by a microplate reader (Thermo-Fisher, Varioskan Flash 3001).

4.8. Rheological analysis

The dynamic mechanical analysis of the hydrogels was assessed by an air oscillating rheometer (HAAKE, MARS iQ air), with the angular frequency of 10 rad/s at 37 °C. The viscosity variation of the hydrogels was performed to test the shear-thinning behavior with a shear rate ranging from 0.1 to 100s⁻¹. Dynamic strain sweep measurements were carried out at a constant frequency of 10 rad/s, and the strain was set in the range of 0.1–1000%. After that, the self-healing behavior of the hydrogels was evaluated by the alternate step-strain sweep test under a constant frequency (10 rad/s) and the amplitude oscillatory strains were switched from low strain (1%) to high strain (1000%) for 3 times with the interval of 60 s.

4.9. Injectability and self-healing performance

The hydrogels were injected out from a 16G needle to test the injectability. Subsequently, the hydrogel was cut into two pieces and then contacted for 1 min. The healing effect was assessed by observing if the healing hydrogel fractured after taking up and then pulling towards to two sides.

4.10. Adhesive, antibacterial performance in vitro

The measurement was conducted according to our previous work [30]. The specific process was showed in SI.

4.11. The photothermal performance

The photothermal performance of the hydrogels was tested by the irradiation of near-infrared (NIR) laser. In brief, the hydrogels were placed in a plastic Petri dish and then exposed to 808 nm NIR laser with the intensity of 0.5 and 1 W/cm² for 10 min. In the meantime, the temperature variation of the hydrogels was recorded and photographed with thermal imager (FOTRIC 220s, America). Next, NIR laser was withdrew and the hydrogels were cooled to the room temperature. Afterwards, NIR laser (808 nm, 1 W/cm²) irradiated the hydrogels for 10 min again and the process was repeated for 4 times to tested the repeatable NIR-responsiveness.

4.12. Cytocompatibility test in vitro

L929s and HUVECs were used to evaluate the cytotoxicity of the hydrogels. Briefly, cells were seeded in 48-well plate with the density of 1 × 10⁴ cells/well and cultivated for 12 h to adhere fully. After that, the hydrogels were placed into the corresponding wells and co-incubated with cells for 1, 2 and 3 days. Finally, cells were colored by CCK8 assays (CCK8, Dojindo), and then the OD values at 450 nm were detected by a microplate reader (Thermo-Fisher, Varioskan Flash 3001). The blank group was the cells without any treatment. In the meantime, HUVECs were dyed with Live/Dead staining kit (Calcein-AM/PI, Invitrogen) and cells viability was observed by inverted fluorescence microscope (Axio Observer.7, ZEISS, Germany).

4.13. Trans-well experiment, intracellular ROS elimination in HUVECs and immunofluorescent staining of CD31 and VE-cad in HUVECs

The recruitment effect of the hydrogels to HUVECs and L929s was assessed by 24-well trans-well chambers (Corning, USA). And the effect of the hydrogels on the elimination of intracellular ROS and the expression of CD31 and VE-cad in HUVECs were also investigated. The detailed process was exhibited in SI.

4.14. Immunofluorescent staining of NF-κB inflammatory signaling pathway in RAW264.7

When NF-κB signal pathway is inactivate, NF-κB protein is expressed in extracellular matrix (ECM). However, once NF-κB signal pathway is activated, NF-κB protein will transfer to cell nucleus. p65 protein is a main component of NF-κB protein, so the immunofluorescent staining of p65 protein can be applied to locate NF-κB protein in cells and determine if NF-κB signal pathway activates. Specifically, macrophages (RAW264.7) were seeded on confocal Petri dish with the density of 2.0 × 10⁴ cells/dish. After incubating for overnight, complete medium with 100 ng/mL lipopolysaccharide (LPS) was added to the Petri dish and incubated with cells for 8 h to activate the NF-κB pathway. After rinsing with PBS for 3 times, the hydrogels were added into and co-cultivated with macrophages for 3d. Subsequently, the cells were dealt with by 4% paraformaldehyde and then 0.1% Triton X-100 (Biofroxx). 3% bovine serum albumin (BSA, Biofroxx) was used to block the cells for 1 h at the room temperature. After that, the cells incubated with primary antibody of p65 (1:100 dilution, Abcam, England) and then Goat anti-Rabbit IgG Cross-Adsorbed Secondary Antibody Alexa Fluor® 568 conjugate (1:500, Thermo Fisher, China). In the end, the cells incubated with DAPI staining (Beyotime, China) to dye cell nuclei and then were observed by a laser scanning confocal microscope (TCS SP8, Leica, Germany). The negative control group was the cells without any treatment and the positive control group was the cells treated by LPS in the absence of the hydrogels.

4.15. RT-PCR of angiogenic and inflammation genes

The mRNA of angiogenic and inflammation genes was extracted from HUVECs and RAW264.7 respectively and transcribed reversely into complementary DNA (cDNA) according to the previous method. The operation steps and calculative methods were displayed in SI.

4.16. Wound healing of diabetic SD rats with full-thickness skin defects and infection

Sprague-Dawley (SD) male rats were chosen to evaluate the promotive effect of the hydrogels on wound healing in vivo. All animal procedures were conducted in accordance with the Guidelines for Care and use of Laboratory Animals of South China University of Technology and approved by the Animal Ethics Committee of South China University of Technology. Firstly, diabetic SD rats model was built by intraperitoneally injecting high dose of streptozotocin into SD rats. After 3 days, SD rats with the glucose level of over 16.7 mmol/L for 2 weeks were used for subsequent experiment. The blood sugar levels in SD rats were tested every three days, and remained over 16.7 mmol/L in the entire process of animal experiment.

15 male diabetic rats were randomly divided into 3 groups. Firstly, SD rats were anesthetized and shaved the hair of the dorsal. Afterwards, four full-thickness round wounds with the diameter of 10 mm were created on the dorsal by biopsy punches and then10 μL of S. aureus (ATCC 6538p, 1.0 × 10⁸ CFU/mL) was dropped in the wounds. After being infected for 30 min, the wounds were treated by HG, DA NPs/HG and Met@CuPDA NPs/HG and the wounds without any treatment were set as the blank group. Subsequently, the hydrogels were exposed to near-infrared light (NIR) laser (0.5W/cm²) for 10 min, and the temperature change of the hydrogels was recorded by infrared imager. After treatment for 0, 3, 7, 14 days, the wounds were photographed to record the size of wound. The wounds closure rate was analyzed by Image J software and calculated according to the below formula (n = 3, 7 or 14).

$$Woundclosurerate=\frac{\mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a}\left(\mathrm{D}\mathrm{a}\mathrm{y}0\right)-\mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a}\left(\mathrm{D}\mathrm{a}\mathrm{y}\mathrm{n}\right)}{\mathrm{a}\mathrm{r}\mathrm{e}\mathrm{a}\left(\mathrm{D}\mathrm{a}\mathrm{y}0\right)}\times 100\%$$

4.17. Histologist analysis

After curing for 3, 7 and 14 days, SD rats were executed to harvest the newborn skin tissues. Subsequently, skin tissues were fixed on 4% paraformaldehyde for at least 24 h and then embedded in paraffin blocks. Hematoxylin-Eosin (H&E) and Masson trichrome staining were applied to analyze wound repair effect from the deposition of ECM and collagen. Moreover, the immunofluorescent staining of CD31 was performed to observe the formation of blood vessels and the immunofluorescent staining of iNOs and Arginase were conducted to monitor the inflammation at the wound sites.

4.18. Statistical analysis

All quantitative data are displayed as mean ± standard deviation. A one-way ANOVA followed by a Tukey test for means comparison was performed to evaluate the level of significance. The value of P＜0.05 was believed to be significant in statistics. The symbols *, **, and *** respectively represented a significant level at 0.05, 0.01 and 0.001. All experiments were performed for at least 3 times to ensure the repeatability of the results.

Ethics approval and consent to participate

All animal procedures were conducted in accordance with the Guidelines for Care and use of Laboratory Animals of South China University of Technology and approved by the Animal Ethics Committee of South China University of Technology.

CRediT authorship contribution statement

Shuangli Zhu: Methodology, Formal analysis, Writing – original draft, Preparation, Writing – review & editing. Bangjiao Zhao: Investigation, Methodology. Maocai Li, Hao Wang and Jiayi Zhu: Research resources/reagents. Qingtao Li: Conceptualization, Technical supports. Huichang Gao: Conceptualization, Technical supports. Qi Feng: Investigation, Methodology, Writing – review & editing. Xiaodong Cao: Conceptualization, Writing – review & editing, Funding acquisition, Project administration.

Declaration of competing 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.

Appendix A. Supplementary data

The following is the Supplementary data to this article.