Development of a High-Throughput ex-Vivo Burn Wound Model Using Porcine Skin, and Its Application to Evaluate New Approaches to Control Wound Infection

Abstract

Biofilm formation in wounds is considered a major barrier to successful treatment, and has been associated with the transition of wounds to a chronic non-healing state. Here, we present a novel laboratory model of wound biofilm formation using ex-vivo porcine skin and a custom burn wound array device. The model supports high-throughput studies of biofilm formation and is compatible with a range of established methods for monitoring bacterial growth, biofilm formation, and gene expression. We demonstrate the use of this model by evaluating the potential for bacteriophage to control biofilm formation by Staphylococcus aureus, and for population density dependant expression of S. aureus virulence factors (regulated by the Accessory Gene Regulator, agr) to signal clinically relevant wound infection. Enumeration of colony forming units and metabolic activity using the XTT assay, confirmed growth of bacteria in wounds and showed a significant reduction in viable cells after phage treatment. Confocal laser scanning microscopy confirmed the growth of biofilms in wounds, and showed phage treatment could significantly reduce the formation of these communities. Evaluation of agr activity by qRT-PCR showed an increase in activity during growth in wound models for most strains. Activation of a prototype infection-responsive dressing designed to provide a visual signal of wound infection, was related to increased agr activity. In all assays, excellent reproducibility was observed between replicates using this model.

Keywords: Staphylococcus aureus; accessory gene regulator; bacteriophage therapy; biofilm; ex-vivo burn wound model; infection; infection responsive materials.

Figure 1

Burn Wound Array Device. The Burn Wound Array Device consists of an array of 24 brass pins that rests in a temperature controlled heated block. Pins are heated to the set temperature and the device signals the user when this has been reached and tool is ready for use. (A) Heated block and temperature controller with 24 pin array inserted and heating. (B) 24 pin array resting on a section of porcine skin to generate thermal injuries.

Figure 2

Ex-vivo burn wound model workflow. Graphical representation showing main stages of set-up and operation of the ex-vivo porcine burn wound model used in this study. (I) Sections of porcine skin are initially prepared by shaving and surface disinfection in 70% ethanol solution; (II, III) The BWAD (see Figure 1) is used to generate an array of consistent partial thickness burn injuries on porcine skin; (IV) If isolation of wounds is required, individual wounds may be excised using a punch biopsy and transferred into wells of tissue culture plates. Wound arrays may be used directly on skin sections without excision. (V–VIII) Individual wounds are inoculated with test organisms to simulate infection and wound biofilm formation. Potential treatments may be evaluated against a subset of infected wounds in the array. (IX) Infected wounds are incubated under required conditions and biofilm formation, bacterial growth, and impact on interventions tested can be evaluated as appropriate. Subsequent methods demonstrated in this study are: Recovery and enumeration of viable cells, quantification of metabolic activity via XXT reduction, the selective staining of biofilms followed by direct imaging of biofilms through CLSM and measurement of fluorescence intensity, extraction of RNA, and quantification of gene expression by qRT-PCR.

Figure 3

Activation of infection responsive dressings. Examples of non-activated (OFF) and activated (ON) prototype infection responsive dressings originally described by Thet et al. (2015). Dressings consist of carboxyfluorescein loaded lipid vesicles embedded in an agarose matrix. At high concentrations in vesicles, carboxyfluorescein exhibits self-quenching properties that suppresses visible color. Lysis of vesicles by bacteria liberates the dye which diffuses through the dressing leading to development of a bright green color visible to the naked eye. In examples above, dressings incubated for 6 h in wound models with PBS or S. aureus RN6911 (which is deficient in the Accessory Gene Regulator agr), show no dressing activation. In contrast, dressings incubated with MRSA252 and RN6390B (parental strain of RN6911 with functional agr) show clear dressing activation.

Figure 4

Evaluation of simulated burn wounds. To confirm that the BWAD was able to generate representative thermal injuries in ex-vivo porcine skin sections, areas of skin in contact with the BWAD after heating to 100°C were examined by histology and directly by environmental scanning electron microscopy (ESEM), and compared to uninjured skin. (A) Thin sections of undamaged and burned skin subjected to H&E staining and histological analysis (x100 magnification). The stratum corneum (SC), Epidermis (E), and dermis (D) are clearly visible and discernible on the undamaged skin sections. In contrast, these structures are disrupted in skin section from BWAD exposed skin, consistent with partial thickness burn injuries. (B) ESEM of the surface of undamaged porcine skin and areas in contact with the BWAD. Scale bar = 20 μm.

Figure 5

Morphology of bacteriophage used in wound models. The previously isolated phage DRA88 (Alves et al., 2014), and the newly isolated phage SAB4328-A were examined by transmission electron microscopy. Characteristics of the DRA88 virion were congruent with membership of the Myoviridae family as previously described, while characteristics of SAB4328-A were in keeping with membership of the Sipohoviridae family. Scale bar represents 100 nm.

Figure 6

Impact of bacteriophage treatment on simulated wound infection. The potential for bacteriophage to control the growth of bacteria in burn wounds was evaluated using the ex-vivo porcine skin model as outlined in Figure 2. The effects of phage treatment were assessed by recovery and enumeration of viable cells, as well as estimates of cellular metabolic activity as measured by the XTT assay, in phage treated vs. untreated wounds inoculated with the multidrug resistant S. aureus MRSA252. (A) Evaluation of a single phage treatment. Wounds were treated with a single dose of phage 4 h after inoculation and levels of viable bacterial cells assessed after a total of 24 h incubation. (B) Evaluation of two phage treatments. Wounds were treated with an initial dose of phage 4 h after inoculation, and again after 24 h of incubation. Levels of viable bacterial cells were assessed after a total of 48 h incubation. Data shows the mean of at least three independent replicates, and error bars show standard error of the mean. Statistically significant differences are denoted by: ^*P ≤ 0.05, ^****P < 0.0001.

Figure 7

Visualization of biofilm formation in wound models and impact of phage therapy. To confirm biofilm formation in the ex-vivo wound model and assess the impact of phage treatment, excised wounds were stained with the general DNA stain DAPI, and WGA-488 fluorescent-lectin conjugate (Wheat Germ Agglutinin with Alexa Flour 488) which binds to poly-N-acetlyglucosamine residues in S. aureus biofilm matrix. (A) Visualization of infected and uninfected wounds by confocal laser scanning microscopy. DAPI staining appears as blue, while WGA-488 staining appears as green areas which are indicative of S. aureus biofilm formation. Control, untreated skin PBS only; +MRSA252, wounds inoculated with S. aureus MRSA252-Rif^R only; +MRSA252+phage, wounds inoculated with S. aureus MRSA252-Rif^R and subject to phage treatment 4 h post inoculation. (B,C) The intensity of fluorescence signal from DAPI or WGA-488 staining measured during CLSM imaging of skin sections. Charts show mean values from three randomly selected regions of skin, in each of three independent replicates. Error bars show standard error of the mean. Statistically significant differences in fluorescence intensity are given by: ^****P < 0.0001 vs. Control; ^####P < 0.0001 vs. MRSA252 only. All images and measurements were taken after 24 h incubation.

Figure 8

Evaluation of agr activity and activation of infection-responsive dressing materials. To understand how S. aureus virulence gene regulation was influenced by growth in the porcine wound model, and relate this to activation of prototype infection responsive dressings, the activity of the Accessory Gene Regulator (agr) was measured in 28 strains (Table 1). Agr activity was assessed by qRT-PCR of RNAIII relative to the housekeeping gene gyrB. (A) Heatmap showing RNAIII:gyrB relative gene expression in each strain between time points assessed: Initial inoculum (T0) and after 6 h in growth in wound models (T6). Shading of cells indicates fold-difference in RNAIII:gyrB expression as indicated by the associated scale. Strains in bold denote those not activating the prototype dressing during the assay. (B) Shows mean RNAIII:gyrB relative expression in all strains, in initial inoculum (T0) and after 6 h growth in wound models (T6). (C) Shows mean RNAIII:gyrB relative expression after 6 h growth in models for strains activating prototype infection-responsive dressings (ON), compared to strains that did not elicit a color change in dressings (OFF). Examples of activated and non-activated dressings are provided in Figure 3. All values show the means of three replicate experiments. Error bars show standard error of the mean. ^**P ≤ 0.02, ^****P < 0.0001.