Preparation and Characterization of HAZ@Fu.
Biocompatible and porous ZIF-8 was synthesized by a reaction between 2-methylimidazole (2-MIM) and Zn2+ salts in DI water. ZIF@Fu was prepared by a one-step synthesis method. First, Fu was mixed with Zn2+ salts, and the sulfonic acid groups in Fu formed coordination bonds with zinc ions. After further assembly with 2-MIM, the Fu molecules were mostly embedded into the framework to obtain a high drug payload of ~ 16.1%.
The morphology, particle size and surface potential of the NPs were analyzed by SEM and DLS. As shown in Fig. 2A, both ZIF-8 and ZIF@Fu NPs showed a uniform rhombic dodecahedron structure with a sharp edge, while the HA-coated HAZ and HAZ@Fu showed smoother edges of the dodecahedron. DLS showed that the hydrodynamic diameters of all the NPs were larger than those observed in the SEM images (Fig. 2C and Table 1). The average hydrodynamic diameter of ZIF-8 was 116.0 ± 0.7 nm, and both loading with Fu and coating with HA were associated with an increase in diameter; of the diameters of HAZ, ZIF@Fu and HAZ@Fu NPs were 168.5 ± 1.5, 222.5 ± 3.6 and 283.1 ± 8.4 nm, respectively. The PDIs of all the NPs were less than 0.3, indicating their good dispersibility. ZIF was positively charged with a zeta potential of 10.10 ± 1.15 mV (Fig. 2D) but changed to a negative charge after Fu encapsulation or HA coating due to their electronegative groups. The elemental composition of HAZ@Fu NPs using TEM element mappings (Fig. 2B) confirmed the existence of the specific sulfur element that originated from Fu molecules, revealing the successful loading of Fu molecules. The crystalline structure of NPs was characterized by XRD, as shown in Fig. 2E, with HAZ, ZIF@Fu and HAZ@Fu exhibiting characteristic peaks for a crystalline structure similar to that of ZIF, with 2θ values of 12.8 and 18.25°, while ZIF@Fu and HAZ@Fu showed a new peak at 13.8°, and overall, the ratio of peak width to peak height decreased, also confirming the successful loading of Fu molecules and smoothing of the dodecahedron edges. Fourier transform infrared spectroscopy (FTIR) analysis of various compositions was conducted. As shown in Fig. 2F, both the crude Fu and the depolymerized fucoidan with low molecular weight (LMWF, Fu) showed characteristic peaks at 840, 1049, 1248 and 1630 cm-1, which were attributed to bending vibrations of C-O-S, stretching vibrations of C-O-C, asymmetric stretching vibrations of S = O bonds in sulfate moieties, and bending vibrations of O-H groups, respectively. The peak intensity at 1220–1270 cm-1 was higher for Fu, indicating a higher content of sulfate esters. ZIF@Fu showed corresponding peak changes at approximately 1049, 1248 and 1630 cm-1 compared with ZIF, indicating sufficient Fu loading. Similar changes were also observed in the comparison between the FTIR spectra of ZIF and HAZ (Fig. 2G). To verify the pH-responsive degradation ability of ZIF-8 NPs, the Zinc Colorimetric Assay Kit was applied to determine the concentration of zinc ions. The HAZ@Fu NPs were incubated with PBS at pH 7.4 and pH 6.0. Approximately 82.66% of the Zn2+ was released at pH 7.4 and only 56.49% at pH 6.0 after 8 h (Fig. 2H). The HAZ@Fu NPs maintained the pH sensitivity of ZIF-8 and could quickly release most cargos in an acidic environment. The cell viability of RAW 264.7 cells treated with various concentrations of Fu and Fu-loaded NPs for 24 h was also determined by the CCK-8 assay (Fig. S1). All MOF-based NPs at a concentration of 80 µg/mL presented satisfactory biocompatibility.
To further verify the antibacterial activity of Fu, MOF and the Fu-loaded MOF in vitro, a colony forming unit (CFU) assay was applied. Crude Fu, Fu, ZIF, and ZIF@Fu at different concentrations were cocultured with MRSA (> 1 × 106 CFU/mL) in LB broth for 16 h, serially diluted (100×) in sterile PBS and inoculated onto LB agar plates, which were further incubated at 37°C for 24 h (Fig. 2I). The ZIF@Fu group was also examined (Fig. S2). The crude Fu at a concentration of 7.5 mg/mL still showed an unreliable antibacterial effect. The MBCs of Fu, ZIF, HAZ, ZIF@Fu and HAZ@Fu were ~ 30, 75, 75, 30 and 30 µg/mL, respectively. Also, zone of inhibition (ZOI) of Crude Fu, Fu, ZIF and ZIF@Fu against MRSA was detected. As shown in Fig. 2J, the Crude Fu showed no obvious ZOI at high concentrations, the ZOI diameters of Fu, ZIF and ZIF@Fu (at an equivalent concentration of 3 mg/mL, which contained 90 µg in filter papers) for MRSA were 18.10 ± 0.72, 10.60 ± 0.49 and 20.64 ± 1.08 mm, respectively, the ZOI diameter of ZIF@Fu group was larger than Fu and ZIF group (P < 0.05), which confirmed with result of the MBCs test. All these results demonstrated that both Fu and ZIF had excellent antibacterial effects, while their combination in the drug-loading system achieved a synergistic effect and reduced the effective antibacterial concentration, which may be because they both have the ability to damage the bacterial capsule, resulting in a significant impact on bacterial permeability.
Table 1
Hydrodynamic Size and Zeta Potential of all NPs.
Sample | Size (nm) | PDI | Zeta potential (mV) |
ZIF-8 | 116.0 ± 0.7 | 0.088 ± 0.023 | + 10.10 ± 1.15 |
HAZ | 168.5 ± 1.5 | 0.151 ± 0.012 | -12.10 ± 1.10 |
ZIF@Fu | 222.5 ± 3.6 | 0.054 ± 0.031 | -26.57 ± 0.55 |
HAZ@Fu | 271.1 ± 8.4 | 0.225 ± 0.013 | -38.97 ± 0.47 |
Synthesis and Characterization of HAZ@Fu MN.
The biocompatible photocrosslinked GelMA hydrogel was chosen as the material for MN tips, in which HAZ@Fu NPs were loaded. PVA was chosen as the backing layer for the MNs due to its biosafety, mechanical properties and rapid dissolution. Using a commercial polydimethylsiloxane (PDMS) MN mold (Fig. 3A), such MNs were fabricated by a two-step template replication method (Figs. 1 and 3B). Figure 2C, D presents photographic images of the MN patch, which was composed of a 15 × 15 MN array. The SEM images show that the MN tips had a pyramidal shape and were aligned in an orderly manner on the backing layer (Fig. 3E). The length of the tip base and height were 290 and 650 µm, respectively. The sharp pyramidal structure ensured that MNs could be inserted deep into the skin quickly and noninvasively, which was further test in mouse skin. As shown in Fig. S3A-C, MN was inserted into the area below the dermis of the mouse skin and stained by methylene blue, indicating that strong mechanical force of MN could penetrate the skin dermis. The HAZ@Fu MN patch was cocultured with MRSA in 1 mL of PBS. As shown in Fig. 3F, obvious distortion of MRSA cells at 0.5 h and disintegration of the bacterial capsule of MRSA at 2 h were observed, indicating the excellent antibacterial activity of HAZ@Fu MN in vitro. To investigate the HAZ@Fu MN release process, the MNs were immersed in PBS, and the concentration of zinc ions was monitored over time. The measured release profile displayed sustained release, in which a cumulative release rate of 82.27 ± 8.36% was obtained at 2 h (Fig. 3G). When MN patche was inserted into the LB agar plate at 37°C, the patch was almost completely dissolved within 2 h (Fig. S4). Which completely consistent with the release profile of HAZ@Fu MN. To evaluate the mechanical properties of the MN patches, the maximum strain capacity of their needle tips was tested by an electronic tension testing machine. The MNs were first placed on a horizontally positioned fixed station with tips facing a force sensor that slowly approached the MNs. Then, the force measurements started when the sensor touched the MN tips and lasted until the measured value was 100 N. The force‒displacement curve of each sample was recorded (Fig. 3H), and the Young’s modulus and the Axial Fracture Force were obtained when the sensor traveled 0.3 mm as the stiffness index of the MN patches (Fig. 3I and S5A). The Young’s modulus of MNs was increased with the increase of GelMA concentration, indicating the enhancement of the mechanical strength of the MNs (P < 0.05). When the GelMA concentration reached 10%, the MNs could withstand compressive forces of 0.17 ± 0.013 N/needle, which was strong enough to realize successful skin puncture, and the loading of NPs hardly changed their mechanical strength (P > 0.05). To evaluate the antibacterial activity of the drug-loaded MNs, the ZIF MN, HAZ MN, ZIF@Fu MN, HAZ@Fu MN and Van MN patches were cultured with MRSA at 37°C for 24 h. From Fig. 3J and Fig. S5B, it noticed that ZIF@Fu MN, HAZ@Fu MN and Van MN caused noticeable ZOI in MRSA with diameters of 22.27 ± 0.40, 23.56 ± 0.67 and 25.02 ± 0.59 mm, respectively. In contrast, both ZIF MN and HAZ MN patch could inhibit bacterial growth in the patch coverage area, the diameters of ZOI significantly smaller than that of ZIF@Fu MN, HAZ@Fu MN and Van MN group (P < 0.05), bare MN without any drugs did not present antibacterial effect. Therefore, the HAZ@Fu MNs presented great antibacterial effect close to that of Van MNs.
Cellular Uptake Efficiency of NPs.
We demonstrated the efficiency of HAZ@FITC NP uptake by the M1-polarized RAW cell group and normal RAW cell group. The rates of M1 polarization were verified through flow cytometry (FCM). As shown in Fig. S6, the level of the M1 marker CD86 significantly increased after RAW 264.7 cells were treated with LPS (100 ng/mL) for 24 h (22.6–65.8%). Correspondingly, only the M1 polarized RAW cells + HAZ@FITC group showed obvious FITC signals at 1 h (Fig. 4A). The FITC fluorescence signal circled around the DAPI-stained cell nucleus, demonstrating that the NPs were rapidly internalized into the cytoplasm of the cells. The FITC fluorescence signal intensity of the M1 polarized RAW cells in the HAZ@FITC group was highest at all timepoints. Additionally, the fluorescence intensity of every group increased from 1 h to 3 h, suggesting that the effect was time dependent. We also quantified the uptake efficiency through FCM (Fig. 4B, C). The M1 polarized RAW cells + HAZ@FITC group also showed the highest FITC fluorescence intensity (77.10 ± 2.91% at 1 h and 90.20 ± 4.94% at 3 h), while that of the M1 polarized RAW cells + ZIF@FITC group was 58.23 ± 5.49% at 1 h and 68.8 ± 6.88% at 3 h. It is inferred that with the upregulation of CD44 expression in M1 polarized cells, more HA-coated nanoparticles could be taken up.
Effects of Fu and HAZ@Fu NPs on the Polarization of Macrophages.
Bone marrow-derived macrophages (BMMs) expressing the M1 marker CD86 and the M2 marker CD206 were detected by flow cytometry (FCM) to investigate the polarization of macrophages. Figure 4D, F show that the M2 phenotype of BMMs was activated by Fu and Fu-containing NPs after 24 h of coculture, as the percentages of cells expressing the CD206 marker in total cells cultured with Fu at concentrations of 15 µg/mL (33.67 ± 3.42%) and 30 µg/mL (49.87 ± 1.54%) or with 80 µg/mL HAZ@Fu NPs (42.73 ± 1.80%) were apparently higher than those in the control group (17.63 ± 2.06%, P < 0.05). Moreover, Fu could inhibit the M1 polarization of BMMs induced by LPS. As shown in Fig. 4E and G, the percentage of cells expressing the CD86 marker in the total cells cultured with LPS (100 ng/mL for 24 h) was 65.00 ± 4.81%, while that of the LPS + 15 µg/mL Fu group was 50.53 ± 1.71%, that of the LPS + 30 µg/mL group was 25.97 ± 5.77% and that of the LPS + HAZ@Fu NPs 80 µg/mL group was 30.33 ± 4.91%. Both Fu and Fu-loaded NPs could activate M2 polarization and inhibit M1 polarization of macrophages in a concentration-dependent manner. Significantly, 80 µg/mL HAZ@Fu NPs contained Fu at only ~ 10.48 µg/mL but showed a similar effect as 30 µg/mL Fu in inhibiting M1 polarization, which may be due to the M1 phenotype macrophage exhibiting higher uptake of HA-coated NPs.
Intracellular Fate and Intracellular Activity of ZIF and HAZ NPs.
To explore the possibility of HAZ NP colocalization with lysosomes, RAW 264.7 cells were cultured with ZIF@FITC or HAZ@FITC NPs for 4 h and then stained with Lyso-tracker Red DND-99. Confocal microscopy images were recorded on living cells. Additionally, we used the Plot Profile feature of ImageJ software to describe the distribution of the fluorescence signal. Since the uptake efficiency had been determined and NP uptake by RAW cells was found to continuously increase within 1 to 3 h, we chose 4 h (over 3 h) as the timepoint. As shown in Fig. 5A, B, there was a significant correlation between the HAZ@FITC signal and lysosomes with a PCC value of 0.72 ± 0.10. The intensity images also showed that the peaks of red and green fluorescence signals were highly coincident, while the ZIF@FITC group showed no significant correlation with lysosomal fluorescence, with a PCC value of 0.42 ± 0.09. These results confirmed the capability of HAZ NPs to accumulate in the lysosomes of RAW 264.7 cells. To further investigate the ability of HAZ NPs to provide subcellular targeting to intracellular MRSA, RAW 264.7 cell infection was achieved by coculture with MRSA (USA 300, transferred by GFP or mCherrymCherry) for 1 h. Then, we added FCS-free DMEM containing 2×MIC vancomycin and incubated the cells for 1 h to remove any extracellular bacteria. Then, the medium was removed, and the cells were twice with PBS. Infected cells were treated with HAZ@FITC for 3 h, washed with PBS and observed by confocal microscopy (Fig. 5C and Fig. S7A). There was a significant correlation between the FITC signal and MRSA (labelled with mCherry), and the intensity images also confirmed this finding. Pearson’s correlation coefficient (PCC) is an index commonly used to quantify the degree of colocalization between different fluorescence signals. The PCC of the above fluorescence signals was calculated by ImageJ software (6 images in total, obtained from three independent experiments). The graph in Fig. 5D shows colocalization with FITC@HAZ for lysosomes (0.72 ± 0.10) and MRSA (0.77 ± 0.19); the values for both were higher than 0.7, indicating strong colocalization. Moreover, the colocalization of MRSA (labelled with GFP) with lysosomes was also detected by confocal imaging (Fig. S7B), and the results showed that almost all the internalized bacteria entered the lysosomes in the cell (indicated with arrows). These results all suggest that HAZ NPs were efficient in targeting intracellular MRSA in macrophages.
To assess the intracellular activity of HAZ@Fu NPs, RAW 264.7 cells infected with MRSA (ATCC 43300) were treated with Fu, ZIF NPs, HAZ NPs, ZIF@Fu NPs and HAZ@Fu NPs (all groups at an equivalent concentration of 60 µg/mL) for 2 or 5 h. Then, the cells were lysed with 0.025% Triton X, and lysates were harvested and serially diluted in sterile PBS, inoculated onto LB agar plates and incubated for 24 h. Digital images of the plates were obtained, and viable cell counts were determined (Fig. 5E, F). After 2 h of incubation, HAZ@Fu showed the most significant antimicrobial activity (P < 0.05). After 5 h of incubation, both ZIF@Fu and HAZ@Fu showed significant antimicrobial activity against intracellular MRSA, and we also noticed that HAZ@Fu was more efficient in treating intracellular MRSA than ZIF@Fu. The pure Fu showed a limited effect on intracellular MRSA, while ZIF and HAZ showed a higher efficiency compared with the control group (P < 0.05). This suggests that the high efficiency of HAZ@Fu in treating intracellular MRSA may be due to the following: Ⅰ) the appropriate size of NPs could facilitate cell uptake; Ⅱ) both Fu and ZIF have antibacterial effects, while their combination in the drug-loading system achieved a synergistic effect, reducing the effective antibacterial concentration; and Ⅲ) HA coating further facilitated intracellular bacterial targeting.
Evaluation of MRSA-Infected Wound Healing In Vivo.
In vivo experiments were conducted in female BALB/c mice aged 6–8 weeks to verify the practical value of HAZ@Fu NPs for infected wound healing. We created 0.8 cm round wounds on the backs of the mice, and then 50 µL PBS containing MRSA was inoculated onto the wound to establish a full-thickness infected cutaneous defect mouse model. The whole experimental procedure is illustrated in Fig. 6A. The mice were randomly divided into 6 groups, and one group that did not receive treatment served as the control. The other five groups were treated with ZIF MN, HAZ MN, ZIF@Fu MN and HAZ@Fu MN patches, and the Van MN group (pure vancomycin loaded via MNs) was used for direct comparison between HAZ@Fu NPs and a clinically effective antibiotic for MRSA. Wound photographs were taken on day 0 and day 4, and all groups were photographed at constant intervals (2 days) until the wounds of the treatment groups exhibited basic healing (Fig. 6B). We measured and analyzed the wound closure rate of every group, and quantitative analysis of the relative wound area was also performed (Fig. 6C and D). From the dynamic changes in wound morphology, the round wounds were nearly healed after 10 days in the HAZ@Fu MN and Van MN groups, while the wounds showed a slower healing rate and scabs could be clearly seen in the ZIF MN, HAZ MN and control groups. The wound area of each group was reduced to some extent on the 4th day, and the relative wound areas in the HAZ@Fu MN and ZIF@Fu MN groups were 58.60 ± 12.18% and 63.11 ± 12.91%, respectively. These values were further decreased to 12.08 ± 4.38% and 16.9 ± 7.32% on the 8th day, which were significantly lower than those in the control group (42.74 ± 12.95%). On the 10th day, the wounds of the HAZ@Fu MN, ZIF@Fu MN and Van MN groups were nearly healed, with complete shedding of epidermal scabs, indicating that the epithelial tissue regeneration process was almost finished, and the HAZ@Fu MNs achieved a wound closure rate over 90%. The HAZ@Fu MNs were significantly more effective than the HAZ MNs and control (P < 0.05), which may be due to the loading of Fu, which reduces the concentration required for antibacterial activity and provides biological activity to promote wound tissue regeneration.
The in vivo antibacterial efficiency was also determined. On day 6, we harvested the skin tissue containing the whole wound area from all groups for tissue homogenization, and the homogenate of each group was serially diluted in sterile PBS and inoculated onto LB agar plates. Digital images of the plates were obtained, and viable cell counts were determined. As shown in Fig. 6E, the number of colony forming units (CFUs) clearly decreased after treatment with HAZ@Fu MNs, ZIF@Fu MNs and Van MNs. As shown in Fig. 6F, all treatment groups showed obvious differences compared with the control group (P < 0.05), and the HAZ@Fu MN group showed no obvious difference compared to the Van MN group (P > 0.05), while ZIF@Fu had a lower bacterial killing efficiency than vancomycin in vivo (P < 0.05). After the wounds were treated for 10 days, Giemsa staining of tissue slices was performed. Consistent with the bacterial counts of the tissue homogenates, the control, HAZ MN and ZIF MN groups showed abundant bacteria in the tissue (Fig. 6G).
To explore the reconstruction of epidermal tissue, histological tests were carried out on the 10th day of the in vivo experiment. Hematoxylin and eosin (H&E) staining was used to investigate the regeneration of wound beds, granulation formation, and epithelial formation processes. The thickness of granulation tissues in every group was also dectecd (Fig. S8). As shown in Fig. 7A, the new stratum corneum in the HAZ@Fu MN, ZIF@Fu MN and Van MN groups was completely formed, the epidermal scab had fallen off, and the granulation tissue thickness was obviously smaller than that in the control and MOF MN groups. Meanwhile, many inflammatory cells (with a high ratio of nucleus to cytoplasm, with predominantly nuclear staining observed in the wound field) and residual scabs were observed in the control group and single MOF MN groups. Quantitatively, the granulation tissue thickness was significantly reduced with HAZ@Fu MN treatment (440.60 ± 20.36 µm) compared to the control (997.80 ± 170.30 µm) and single MOF MN groups (Fig. 7C), indicating that the regulation of inflammation and tissue regeneration was nearly complete. Collagen synthesis, deposition, and directional alignment are crucial processes in tissue remodeling and are also important phases in wound healing. Masson staining was performed. As shown in Fig. 7B, a large area of new collagen stained dark blue was directionally aligned in the neonatal epithelial tissue in the ZIF@Fu and HAZ@Fu groups, and the coverage of collagen (Fig. 7D) reached 39.87 ± 1.41% and 35.18 ± 5.92%, respectively, more than twice the value in the control and single MOF MN groups and significantly higher than that in the Van MN group (24.08 ± 5.59%, P < 0.05). This demonstrated an elevated amount of collagen deposition and improved tissue remodeling.
Investigation of Tissue Regeneration and Inflammation after Treatments.
To investigate the biological mechanism underlying the wound healing process, inflammation and angiogenesis were also evaluated. For this, immunohistochemistry staining of CD206 and IL-6 and double immunofluorescence staining of CD31 plus α-smooth muscle actin (α-SMA) were carried out. As shown in Fig. 8A and B, high amounts of secreted IL-6 and a small number of CD206 positive cells were distributed in the control group (IL-6 positive area of 33.00 ± 1.61% and CD206 positive cell of 10.16 ± 3.64%) and single MOF MN groups, indicating that a strong inflammatory response was still present in these groups. Little IL-6 secretion and large numbers of CD206-positive cells were observed in the HAZ@Fu MN group (IL-6-positive area of 11.52 ± 2.49% and CD206-positive cells of 28.73 ± 3.8%), indicating few signs of an inflammatory response. The ZIF@Fu MN group, another Fu-loaded MOF MN group, showed a larger amount of IL-6 secretion (20.52 ± 2.86%, P < 0.05) and fewer CD206-positive cells (25.82 ± 4.21%, P > 0.05), indicating the positive effect of HA in the treatment of MRSA-infected wounds. The neovascularization degree is also an important healing indicator that can be evaluated by immunofluorescence staining of CD31 (endothelial cell markers) and α-SMA (fibroblast markers). As revealed in Fig. 7C and F, the new blood vessel density in the wound bed was found to increase significantly in the HAZ@Fu and ZIF@Fu MN groups, and the coverage of CD31% reached 3.63 ± 0.17% in the HAZ@Fu group and 2.93 ± 0.59% in the ZIF@Fu MN group. Since a small amount of IL-6 secretion, moderate number of CD206-positive cells and moderate coverage of CD31 were observed in the Van MN group (16.96 ± 5.08%, 16.73 ± 2.59% and 2.18 ± 0.21%, respectively), the positive effect of HAZ@Fu NPs in modulating the inflammatory response and neovascularization process was demonstrated. A low density of CD31 markers was observed in the control and single MOF MN groups, which may be attributed to the inhibitory effect of neovascularization caused by MRSA infection at the wound bed. Collectively, while the HAZ@Fu MN group exhibited an anti-MRSA effect similar to that of the Van MN group, it also showed enhanced wound closure, epithelial regeneration, neovascularization, and anti-inflammatory effects compared with all other groups.