To study the effects of reactive oxygen species (ROS)-responsive antibacterial microneedles (MNs) on the full-thickness skin defect wounds with bacterial colonization in diabetic mice. Experimental research methods were adopted. The ROS-responsive crosslinker N1-(4-boronobenzyl)-N3-(4-boronophenyl)-N1, N1, N3, N3-tetramethylpropane-1,3-diaminium (TSPBA) was first synthesized, and then the polyvinyl alcohol (PVA)-TSPBA MNs, PVA-ε-polylysine (ε-PL)-TSPBA MNs, PVA-TSPBA-sodium hyaluronate (SH) MNs, and PVA-ε-PL-TSPBA-SH MNs were prepared by mixing corresponding ingredients, respectively. The PVA-TSPBA MNs were placed in pure phosphate buffer solution (PBS) and PBS containing hydrogen peroxide, respectively. The degradation of MNs immersed for 0 (immediately), 3, 7, and 10 days was observed to indicate their ROS responsiveness. The standard strains of () and cultured in Luria-Bertani medium containing hydrogen peroxide were divided according to the random number table (the same grouping method below) into blank control group (without any treatment, the same below) and 0 g/L ε-PL group, 1.0 g/L ε-PL group, 5.0 g/L ε-PL group, and 10.0 g/L ε-PL group with which PVA-ε-PL-TSPBA MNs containing the corresponding concentration of ε-PL were co-cultured, respectively. Bacterial growth was observed after 24 h of culture, and the relative survival rate of bacteria was calculated (=3). The mouse fibroblast cell line 3T3 cells at logarithmic growth stage (the same growth stage below) were divided into blank control group and 0 g/L ε-PL group, 1.0 g /L ε-PL group, 5.0 g /L ε-PL group, and 10.0 g /L ε-PL group in which cells were cultured in medium with the extract from PVA-ε-PL-TSPBA MNs containing the corresponding concentration of ε-PL, respectively. Cell growth was observed after 24 h of culture by optical microscopy, and the relative survival rate of cells was detected and calculated by cell counting kit 8 (CCK-8) assay to indicate the cytotoxicity (=6). Both PVA-TSPBA MNs and PVA-TSPBA-SH MNs were taken, the morphology of the two kinds of MNs was observed by optical microscopy, and the mechanical properties of the two kinds of MNs were tested by microcomputer controlled electronic universal testing machine (denoted as critical force, =6). Six male BALB/c mice aged 6-8 weeks (the same gender and age below) were divided into PVA-TSPBA group and PVA-TSPBA-SH group, with 3 mice in each group. After pressing the skin on the back of mice vertically with the corresponding MNs for 1 minute, the skin condition was observed at 0, 10, and 20 min after pressing. Another batch of 3T3 cells were divided into blank control group, 0 g/L ε-PL group and simple 5.0 g/L ε-PL group which were cultured with the extract of PVA-ε-PL-TSPBA MNs containing the corresponding concentration of ε-PL, and 5.0 g/L ε-PL+SH group which were cultured with the extract of PVA-ε-PL-TSPBA-SH MNs with 5.0 g/L ε-PL. The CCK-8 assay was performed to detect and calculate the relative survival rate of cells cultured for 24, 48, and 72 h to indicate the cell proliferation activity (=6). Eighteen BALB/c mice were induced into diabetic mice model by high-sugar and high-fat diet combined with streptozotocin injection and then divided into sterile dressing group, 0 g/L ε-PL+SH group, and 5.0 g/L ε-PL+SH group, with 6 mice in each group. A full-thickness skin defect wound was made on the back of each mouse, and solution was added to make a full-thickness skin defect wound with bacterial colonization model for diabetic mouse. The wounds of mice in 0 g/L ε-PL+SH group and 5.0 g/L ε-PL+SH group were covered with PVA-ε-PL-TSPBA-SH MNs with the corresponding concentration of ε-PL, and the wounds of mice in the 3 groups were all covered with sterile surgical dressings. The wound healing was observed on post injury day (PID) 0, 3, 7, and 12, and the wound healing rate on PID 3, 7, and 12 was calculated. On PID 12, the skin tissue of the wound and the wound margin were stained with hematoxylin and eosin to observe the growth of new epithelium and the infiltration of inflammatory cells. Data were statistically analyzed with one-way analysis of variance, analysis of variance for repeated measurement, Mann-Whitney test, and Bonferroni test. With the extension of the immersion time, the PVA-TSPBA MNs in PBS containing hydrogen peroxide gradually dissolved and completely degraded after 10 days of immersion. The PVA-TSPBA MNs in pure PBS only swelled but did not dissolve. After 24 h of culture, there was no growth of in 5.0 g/L ε-PL group or 10.0 g/L ε-PL group, and there was no growth of in 10.0 g/L ε-PL group. The relative survival rate of was significantly lower in 1.0 g/L ε-PL group, 5.0 g/L ε-PL group, and 10.0 g/L ε-PL group than in blank control group (<0.05 or <0.01). The relative survival rate of was significantly lower in 5.0 g/L ε-PL group and 10.0 g/L ε-PL group than in blank control group (<0.01). After 24 h of culture, the cells in blank control group, 0 g/L ε-PL group, 1.0 g/L ε-PL group, 5.0 g/L ε-PL group, and 10.0 g/L ε-PL group all grew well, and the relative survival rate of cells was similar among the groups (>0.05). The needle bodies of PVA-TSPBA MNs and PVA-TSPBA-SH MNs were both quadrangular pyramid-shaped and neatly arranged, and the needle bodies of PVA-TSPBA-SH MNs was more three-dimensional and more angular. The critical force of PVA-TSPBA-SH MNs was significantly higher than that of PVA-TSPBA MNs (=3.317, <0.01). The MNs in PVA-TSPBA+SH group penetrated the skin of mice at 0 min after pressing, and the pinholes partially disappeared after 10 min and completely disappeared after 20 min, while the MNs in PVA-TSPBA group failed to penetrate the skin of mice. After 24, 48, and 72 h of culture, the proliferation activity of the cells in 5.0 g/L ε-PL+SH group was significantly higher than that of blank control group (<0.05 or <0.01). In sterile dressing group, the wounds of mice healed slowly and exuded more. The wound healing speed of mice in 0 g/L ε-PL+SH group was similar to that of sterile dressing group in the early stage but was faster than that of sterile dressing group in the later stage, with moderate exudation. The wound healing of mice in 5.0 g/L ε-PL+SH group was faster than that in the other two groups, with less exudation. The wound healing rates of mice in 5.0 g/L ε-PL+SH group were (40.6±4.2)%, (64.3±4.1)%, and (95.8±2.4)% on PID 3, 7, and 12, which were significantly higher than (20.4±2.7)%, (38.9±2.2)%, and (59.1±6.2)% in sterile dressing group and (21.6±2.6)%, (44.0±1.7)%, and (82.2±5.3)% in 0 g/L ε-PL+SH group (<0.01). The wound healing rates of mice in 0 g/L ε-PL+SH group on PID 7 and 12 were significantly higher than those in sterile dressing group (<0.05 or <0.01). On PID 12, the wounds of mice in 5.0 g/L ε-PL+SH group were almost completely epithelialized with less inflammatory cell infiltration, the wounds of mice in 0 g/L ε-PL+SH group were partially epithelialized with a large number of inflammatory cell infiltration, and no obvious epithelialization but a large number of inflammatory cell infiltration was found in the wounds of mice in sterile dressing group. The composite MNs prepared by TSPBA, PVA, ε-PL, and SH can successfully penetrate mouse skin and slowly respond to ROS in the wound to resolve and release antibacterial substances, inhibit bacterial colonization, and promote the repair of full-thickness skin defect wounds with bacterial colonization in diabetic mice.