All animal experiments in the present study were approved by the Medical Ethics Committee of the Chinese PLA General Hospital (Beijing, China) in compliance with the Guidelines for Care and Use of Animals in Research (45).

A total of 18 adult female Japanese large-ear white rabbits (age, 3–6 months; weight, ~3 kg; purchased from the Laboratory Animal Center of the Academy of Military Medical Sciences, Beijing, China) were used for this study. The rabbits were acclimated to standard housing and fed ad libitum under constant temperature (22°C) and humidity (45%) with a 12-h light/dark cycle.

S. aureus strain RN6390 constitutively expressing green fluorescent protein (obtained the from Chinese PLA Institute for Disease Control and Prevention, Beijing, China) was utilized for wound inoculation. S. aureus was grown overnight and subcultured in Luria-Bertani broth (AOBOX Biotechnology Co. Ltd., Beijing, China) at 37°C until log phase. Cells were harvested by centrifugation at 4°C (5,000 × g), and washed three times with phosphate-buffered saline (PBS). The final bacterial resuspension was diluted with PBS to an optical density of 1.0 at 600 nm equivalent to 10⁵ colony-forming units (CFUs)/µl empirically (11).

The wounding and bacterial inoculation protocol was based on the previously published wound model with minor modifications (12). Briefly, rabbits were anesthetized by intramuscular injection of a ketamine (45 mg/kg; Gutian Pharma Co., Ltd., Fujian, China) and xylazine (5 mg/kg; Huamu Animal Health Product Co., Ltd., Jilin, China) mixture. Ears were shaved and sterilized twice with 70% ethanol. Following local anesthesia with 1% lidocaine (Yimin Pharmaceutical Co., Ltd, Beijing, China), six standardized 8 mm diameter full-thickness dermal wounds were created by an experienced surgeon on each ventral ear down to the perichondrium using a scalpel. Following hemostasis, the wounds were dressed with semiocclusive IV3000 Transparent Adhesive Film Dressing (Smith & Nephew Healthcare Ltd., Hull, UK). On postoperative day (POD) 3, each wound was inoculated with 1×10⁶ CFUs of S. aureus at a volume of 10 µl. Planktonic bacteria were allowed to proliferate in vivo under the semiocclusive transparent film for a minimum of 6 h to ensure bacterial adhesion and colonization (12,41,46,47).

All rabbit ears were used to create acute S. aureus-contaminated wounds (6 wounds/ear; 36 ears). For each animal, the two ears were respectively and randomly assigned to the ‘untreated control group’ and the ‘NPWT group’ (18 ears/group), with the 6 wounds on each ear following the same protocol (32).

The wounds were treated under the previously published protocol with minor modifications (48). Briefly, 6 h postinoculation on POD 3, the wounds were dressed with a standard NPWT dressing (consisting of polyvinyl alcohol foam, semiocclusive transparent dressing and suction tube; Wuhan VSD Medical Science & Technology Co., Ltd., Wuhan, China) trimmed in advance to the appropriate size. The suction tube was then connected to a vacuum pump device (provided by Professor Lei Hu, Beihang University, Beijing, China). Wounds treated with NPWT were subjected to continuous negative pressure at −125 mmHg throughout the study (41,47). Dressings were checked daily and changed on PODs 4, 6, 8 and 10, as recommended for infected wounds by the manufacturer. Animals were sacrificed via an overdose of intravenous pentobarbital sodium (100–240 mg/kg; Sigma-Aldrich, St. Louis, MI, USA) on PODs 4 (n=8), 6 (n=2), 8 (n=2), 10 (n=2) and 13 (n=4). Wounds were harvested using an 8-mm dermal biopsy punch (Miltex, Inc., York, PA, USA) for PIA, eDNA, viable bacterial count and scanning electron microscopy (SEM) analyses, or a scalpel for epifluorescence and histological analyses.

The dorsal side of the wound was removed to eliminate the interference of non-specific bacteria outside of the wound surface (32). The tissue samples were placed in centrifuge tubes with 1 ml PBS, and sonicated for 2 min to remove bacterial biofilms from the tissue (32,49). The insoluble material was discarded by centrifugation at 4°C (13,400 × g). In total, 0.5 ml supernatant was used for the measurement of PIA, and the remainder was used for eDNA extraction. An improved Elson-Morgan assay was performed subsequently to measure the levels of PIA (13,19,20,50). Briefly, 0.5 ml supernatant was supplemented with 0.1 ml potassium tetraborate (0.8 mol/l; Vetec; Sigma-Aldrich) in a test tube. Tubes were heated in a boiling water bath for 3 min and were subsequently cooled using tap water. A total of 3 ml Ehrlich's reagent (Vetec; Sigma-Aldrich) was added and, immediately after mixing, the tubes were placed in a water bath at 37°C for 20 min. The reaction was terminated by cooling the tubes with tap water. The results were expressed as the absorbance of the final reaction solution at 585 nm as determined using a spectrophotometer (GeneQuant 1300; GE Healthcare Life Sciences, Logan, UT, USA) (50).

The remaining supernatant was used for the extraction of eDNA using the TIANamp Micro DNA kit (Tiangen Biotech Co., Ltd., Beijing, China) according to the manufacturer's protocol. The eDNA levels/wound were expressed as the DNA concentration, quantified using a Qubit® 2.0 fluorometer with a Qubit® dsDNA BR Assay kit (both Invitrogen; Thermo Fisher Scientific, Inc., Waltham, MA, USA) according to the manufacturer's protocol (30).

Following 2.5% glutaraldehyde and 1% osmium tetroxide fixation, the tissue samples were dehydrated through a series of graded ethanol (30, 50, 70, 80, 90 and 100%) and dried with a critical point dryer (HCP-2; Hitachi, Ltd., Tokyo, Japan) by flooding with liquid carbon dioxide at 5°C for 20 min and raising the temperature to the critical point (~35°C). Subsequently, samples were mounted using double-sided tape and coated with gold in an auto sputter coater (E-1010; Hitachi, Ltd., Tokyo, Japan). Imaging of the tissue samples was performed by SEM (S-3400N; Hitachi, Ltd., Tokyo, Japan) operated at the scanning voltage of 15 kV.

The wounds were bisected at the largest diameter and embedded in optimal cutting temperature (O.C.T.) Compound (Sakura Finetek USA, Inc., Torrance, CA, USA), snap-frozen, and stored in liquid nitrogen until cryosectioning. Tissue sections (6 µm) were obtained with a Leica CM1950 freezing microtome (Leica Microsystems GmbH, Wetzlar, Germany). Visualization of biofilm matrix was accomplished by staining with Concanavalin A Texas Red® Conjugate (Invitrogen; Thermo Fisher Scientific, Inc.), specific for S. aureus exopolysaccharides, for 15 min in the dark at room temperature (12). The tissue sections were then rinsed three times in PBS and incubated with 4′,6-diamidino-2-phenylindole dilactate (Invitrogen; Thermo Fisher Scientific, Inc.) for 5 min to visualize the host cells. Fluorescence microscopy was performed with an Olympus BX51 microscope (Olympus Corporation, Tokyo, Japan).

Wounds were pretreated as described in the protocol of PIA measurement. Tissue specimens were homogenized into 1 ml suspension with sterile PBS, followed by sonication for 2 min to disrupt bacterial aggregates in the biofilm. Subsequently, the homogenates were serially diluted with sterile PBS (ranging from 10⁻¹, 10⁻², 10⁻³ to 10⁻⁶ times the concentration of the homogenates), plated on Staphylococcus Isolation Agar (Hardy Diagnostics, Santa Maria, CA, USA) and incubated at 37°C for 24 h. A standard colony-counting method was conducted and the results were expressed as the logarithm of CFUs/wound (12,41).

From POD 0 (the day of wounding), images of the wounds were captured with a digital camera (IXUSi; Canon, Inc., Tokyo, Japan) on each day of dressing change. Wound size was then determined by quantifying the wound area using Image-Pro Plus version 6.0 software (Media Cybernetics, Inc., Rockville, MD, USA). The rate of wound closure was expressed as a percentage of the initial wound area (51).

Wounds were bisected at the maximum diameter and fixed in 10% neutral formalin (Yili Fine Chemicals Co., Ltd., Beijing, China). The tissue samples were then embedded in paraffin, sectioned into 5-µm sections and stained with hematoxylin and eosin. The observation of the tissue sections was performed with an Olympus MVX10 macro-microscope (Olympus Corporation). Images of the stained wounds were captured and analyzed using Image-Pro Plus software, version 6.0. The morphometric parameters included the epithelial and granulation gaps (the distance between the leading edges of newly formed epithelial or granulation tissue) and the epithelial or granulation area (the sum of the area of newly formed epithelial or granulation tissue on the two sides of the wound bed) (28). The measurements were carried out by two independent observers who were blinded to the treatment, and an average result was calculated.

Data are reported in graphical form as mean ± standard deviation. Student's t-test (two-tailed and paired) was used to compare the PIA/eDNA content, viable bacterial counts and histological parameters between NPWT-treated wounds and their untreated control. Wound closure was analyzed using repeated-measures analysis of variance, followed by the least significant difference post-hoc test to evaluate the statistical difference between groups at each time point. All analyses were performed using SPSS software, version 17.0 (SPSS Inc., Chicago, IL, USA) at a significance level of α=0.05.

With the previously established rabbit ear wound model, S. aureus-contaminated wounds were treated with or without NPWT from 6 h postinoculation. For the main components of S. aureus biofilm, namely PIA and eDNA, NPWT resulted in a significant reduction in their levels in the wound beds (PIA, P<0.05; eDNA, P<0.01) (Fig. 1). This reduction could be visualized through SEM (Fig. 2). The untreated wounds manifested a relatively intact biofilm structure (Fig. 2A) with large amounts of cocci-shaped S. aureus embedded within the lattice-like extracellular matrix (arrows, Fig. 2B). Conversely, wounds treaded with NPWT showed comparable amounts of bacteria but a lack of extracellular matrix (Fig. 2C), which was more visible at a higher magnification (Fig. 2D). To verify the results, epifluorescence microscopy was performed, showing a mature biofilm with large amounts of exopolysaccharide matrix surrounding the bacteria in untreated wounds (Fig. 3A). By comparison, bacteria in the NPWT group spread over the wound beds with sparse exopolysaccharide matrix (Fig. 3B). An uncolonized area of the wound tissue showed background staining of Concanavalin A, and served as a negative control (Fig. 3C).

Viable bacterial counts were measured over time to investigate the effect of early application of NPWT on bacterial burden (Fig. 4). On POD 4 and 6, the bacterial load in the wounds treated with NPWT was not significantly different from that in the untreated controls with ~10⁷ CFUs/wound. However, from POD 8 there was a significant decline of the bacterial load in the NPWT group (POD 8, P<0.05; POD 10, P<0.01). At the end of the study (POD 13), NPWT resulted in a significant reduction in the bacterial count by two-log fold compared with the untreated wounds (P<0.001).

The reduction in bacterial burden secondary to NPWT correlated with a synchronous enhancement in wound healing. Gross appearance on POD 13 manifested a marked difference between the two groups. Compared with the film-like exudates overlying the untreated wounds (Fig. 5A), NPWT accelerated wound closure and epithelialization, with clean granulation tissue beds (Fig. 5B). The rate of wound closure in the NPWT group was significantly higher compared with that of the untreated group (Fig. 5C). Images of histological tissue sections showed increased amounts of new epithelial and granulation tissue in NPWT-treated wounds (Fig. 6B) compared with untreated controls (Fig. 6A). Trends in wound healing were quantified by measuring the epithelial and granulation gaps and areas. NPWT led to a significant reduction in epithelial and granulation gaps (P<0.001) and increase in new epithelial (P<0.01) and granulation (P<0.001) areas compared with the untreated wounds (Fig. 6C and D).