Formation of Pseudomonas aeruginosa Biofilms in Full-thickness Scald Burn Wounds in Rats

Using Sprague-Dawley rats (350–450 g; n = 61) and the recently updated Walker-Mason rat scald burn model, we demonstrated that Pseudomonas aeruginosa readily formed biofilms within full-thickness burn wounds. Following the burn, wounds were surface-inoculated with P. aeruginosa in phosphate-buffered saline (PBS), while sterile PBS was used for controls. On post-burn days 1, 3, 7, and 11, animals were euthanized and samples collected for quantitative bacteriology, bacterial gene expression, complete blood cell counts, histology, and myeloperoxidase activity. Robust biofilm infections developed in the full-thickness burn wounds inoculated with 1 × 104 CFU of P. aeruginosa. Both histology and scanning electron microscopy showed the pathogen throughout the histologic cross-sections of burned skin. Quantigene analysis revealed significant upregulation of alginate and pellicle biofilm matrix genes of P. aeruginosa within the burn eschar. Additionally, expression of P. aeruginosa proteases and siderophores increased significantly in the burn wound environment. Interestingly, the host’s neutrophil response to the pathogen was not elevated in either the eschar or circulating blood when compared to the control burn. This new full-thickness burn biofilm infection model will be used to test new anti-biofilm therapies that may be deployed with soldiers in combat for immediate use at the site of burn injury on the battlefield.

One of the greatest challenges facing the burn clinic is the complication of bacterial infection within the burn wound that may lead to more severe disease states, including sepsis. Scarcity of new antimicrobials, specifically anti-biofilm agents, further escalates the challenge posed by resistant microorganisms in treating burn wound infections. Several different animal models that seek to recapitulate the hallmarks of clinical burn wound infection are used to assist the development of novel antimicrobial therapies. The three most common burn wound infection in vivo models include pigs, mice, and rats [1][2][3][4][5] . Porcine models are often the best substitute for human skin, but their use for screening vast librarys of antimicrobials are hampered by high costs and ethical implications 3 . Mice offer the greatest range of benefits for characterizing aspects of the host's response to burn infection, based on genetic mutants, but are often limited to acute (less than 48 hours post inoculation) burn wound infection before rapid onset of sepsis and death, potentially due to the requirement for injection of the high bacterial inoculums beneath the burn eschar to induce the burn infection [6][7][8] . Others have also included formulation of the pathogen in a biopolymer matrix to drive the infection towards a more chronic biofilm-like state 9,10 . However, these inoculation and exposure routes do not recapitulate what occurs in burn clinics and fail to mimic the natural development of infection in full-thickness burn injuries. Rats have played a significant role in the development and transition of antimicrobial burn wound treatments to clinics since the 1960's, when the original Walker-Mason rat scald burn and surface infection model was described [11][12][13][14][15][16] .
The Walker-Mason rat scald burn model displays the classical features of invasive burn wound infection, typically observed with Pseudomonas aeruginosa using a natural surface route of infection, albeit with high inoculum levels of the pathogen, after induction of the burn injury. Importantly, studies using this model directly led to the development and clinical adoption of the 'gold standard' antimicrobial treatments for severe burn

Results
We characterized the kinetics of P. aeruginosa infection and biofilm formation in full-thickness scald burns using the modified Walker-Mason burn model in Sprague-Dawley rats. Figure 1 shows representative gross appearance of the burn wound with and without P. aeruginosa infection over the course of the study. Exudate became apparent by POD 3 in the wounds inoculated with P. aeruginosa. By POD 7 and 11, areas of tissue necrosis within the burn wound could be easily observed in both inoculum groups compared to the controls. The un-inoculated wounds did show contamination of the eschar based on the discoloration of the burned area. Given the severity of the wound infection, an expected decline in body weight was observed in both inoculum groups, but generally leveled off or began to recover during the study with the extra administrations of pain medication and Lactated Ringers on POD's 3 and 7. The body weights of the animals over the course of the infection are shown in Supplementary Fig. 1. Quantitative bacteriology. Viable CFU counts isolated from the tissue biopsies are shown in Fig. 2. A similar increase in bacterial number was obtained on P. aeruginosa isolation agar ( Fig. 2A) and standard blood agar (Fig. 2B) as seen previously in partial-thickness burn wounds infected with P. aeruginosa 22 . By POD 7, a plateau in number of microorganisms, P. aeruginosa or native flora, per gram of wound tissue was observed. Large variability in the 1 × 10 3 CFU/wound inoculum was present in the counts obtained from P. aeruginosa isolation agar due to lack of P. aeruginosa recovery in the second run of the infection study. The results of the viable counts were confirmed via qPCR for bacterial DNA isolated from the homogenate of the same biopsy punches, shown in Fig. 3A,B. PCR analysis showed near perfect mirroring of the viable CFU counts. As a result of the variability within the 1 × 10 3 inoculum, the P. aeruginosa percentage of total bacterial DNA recovered from the wound sample was greatly reduced compared to the 1 × 10 4 inoculum, shown in Fig. 3C. Representative gross images of full-thickness burn wounds inoculated with P. aeruginosa (1 × 10 3 or 1 × 10 4 CFU/wound) show significantly more wound exudate and areas of necrosis compared to wounds receiving sterile PBS (Un-inoculated). Table 1 documents the results of the Signal TM Blood Culture System from the second and third runs of the study. Unfortunately, the Blood Culture System was not employed during the first experimental run, therefore blood culture results are not available from those animals. No animal showed positive results on POD 1. One of four animals in both inoculum groups showed positive blood cultures for P. aeruginosa on POD 3. By POD 7, 25% of 1 × 10 3 CFU inoculum group was positive for P. aeruginosa in the blood stream, while 50% of the 1 × 10 4 CFU inoculum group showed the pathogen in the blood. By POD 11, half of the animals in the 1 × 10 3 CFU inoculum group showed P. aeruginosa and Staphylococcus sciui ss sciuri growth in the blood, while all animals in the 1 × 10 4 CFU inoculum group were positive for P. aeruginosa or Staphylococcus xylosus in the blood. These data demonstrate that P. aeruginosa burn wound infection may lead to bacteremia by P. aeruginosa or other native flora 30,31 [Sanjar et al. "Identification of Metagenomics Structure and Function Associated with Temporal Changes in Rat (Rattus norvegicus) Skin Microbiome during Health and Cutaneous Burn" in press] common to the rat skin that transgress to the blood stream within a week of the burn, if left untreated. Importantly, animals that received a burn but were not inoculated with P. aeruginosa did not show the same invasive burn wound infection and presence of any bacterial cells, P. aeruginosa or native microflora, in the blood stream. Maximum bacterial load of full-thickness burn wounds was achieved within 7 days. The viable counts from P. aeruginosa isolation agar (A) and non-selective agar (B) increase with time following the burn. P. aeruginosa was not cultured from un-inoculated control wounds. The 1 × 10 3 inoculum showed pronounced variability in the P. aeruginosa counts due to lack of recovery during the second experimental run. The 1 × 10 4 inoculum exhibited superior reproducibility, in terms of bacterial recovery, between all three experimental runs. Regardless of inoculation amount, by POD 11, all groups on the non-selective agar medium obtained a plateau of ~1 × 10 9 CFU/g wound tissue of either P. aeruginosa or native flora. Figure 3. qPCR confirmed the P. aeruginosa and total bacterial loads within the burn eschar over 11 days following burn injury. The results of the qPCR analysis mirrored the data obtained in the viable counts for P. aeruginosa (A) and total bacterial cells (B). As a percentage of the total bacterial load (C), within three days post-burn and infection P. aeruginosa dominated the burn wounds inoculated with 1 × 10 4 CFU/wound. As expected, wounds inoculated with 1 × 10 3 CFU/wound showed greater variability of P. aeruginosa percentage of bacterial load due to its lack of recovery in the second experimental run.
www.nature.com/scientificreports www.nature.com/scientificreports/ Histology. We evaluated histologic cross-sections of the burn wound inoculated with P. aeruginosa by scanning electron microscopy (SEM). Figure 4A shows both 1,000× and 10,000× magnification views obtained by SEM of a wound inoculated with 1 × 10 4 CFU of P. aeruginosa over the course of the study. On POD 1, few if any bacterial cells were observed, but large numbers were seen penetrating into the eschar by POD 3. Similar to our work in partial-thickness burns, the bacterial cells were often associated with the collagen fibers of the burned tissue. We also examined the appearance of the eschar surface with SEM on POD 11. Figure 4B shows areas of both exposed bacterial cells and smoothed matrix like material that may be a combination of host proteins/tissues and bacterial exopolysaccharides of a wound inoculated with 1 × 10 4 CFU of P. aeruginosa. Additionally, upon high magnification (10,000×), the layers of bacterial cells could be easily seen in the cracks on the surface of the eschar, potentially a product of the sample dehydration during the preparation for SEM imaging. Unlike the wounds infected with P. aeruginosa, SEM images of control burn wounds inoculated with PBS, Supplementary  Fig. 2, do not show the presence of bacterial cells associated with the collagen fibers or within a matrix on the wound surface.
Micrographs of histologic sections from 1 × 10 4 inoculated wounds on POD 7 stained with H&E and Giemsa or PNA-FISH are shown in Fig. 5. Micrographs of histologic sections from control un-inoculated wounds are  www.nature.com/scientificreports www.nature.com/scientificreports/ shown in Supplementary Fig. 3 and an overview of the tissue sections are displayed in Supplementary Fig. 4. High magnification images taken with the 63× objective of three areas in the tissue section include the epidermis, the middle of the dermis, and of the deep dermis near the panniculus carnosus. The combination of the brightfield stains allows for identification of the tissue structure while also highlighting the penetration of the pathogen throughout all levels of the burned eschar. PNA-FISH, specific for P. aeruginosa, confirmed the identity of the pathogen within the tissue sections. Respective micrographs of the un-inoculated wounds on POD 7 after the burn are shown in Supplementary Fig. 3. The un-inoculated wounds did show the presence of the native flora, mostly gram-positive cocci, within the histological sections stained with H&E and Giemsa, but lacked the presence of the P. aeruginosa and microcolonies observed in the inoculated wounds stained with PNA-FISH.
Biofilm gene expression. Expression patterns of important biofilm associated genes of P. aeruginosa over the 11-day experiment are shown in Fig. 6 with results shown as a fold change of the planktonic inoculum. Alginate genes (algD, alg8, and algE) displayed in Fig. 6A-C show roughly 5-10 fold increases in expression levels as compared to the inoculum. Plateaus in algD and algE gene expression were reached by POD 7 after the burn. Alg8 showed higher expression levels in the 1 × 10 3 CFU/wound inoculated group as compared to the 1 × 10 4 CFU/wound group, which were uniformly elevated ~10 fold higher than the inoculum throughout the experiment. Similar expression increases were observed in the pelB, pelC, and pelD (pellicle biofilm) genes over the course of the study as depicted in Fig. 6D-F. Roughly, 5-15 fold increases in pellicle biofilm gene expression were detected in the burn wound compared to the inoculum, although pelC and pelD showed increased variability in the 1 × 10 3 inoculum. We measured the expression of three classical P. aeruginosa virulence genes (protease genes lasA, lasB, and iron starvation sigma factor gene pvdS) in the burn wound tissue, shown in Fig. 6G-I. Increased expression, between 10-30 fold over the inoculum, were measured over the course of the study for each virulence gene. Conversely, biofilm polysaccharide genes pslA and pslC did not show similar increases in expression level upon in vivo inoculation as seen in Fig. 6J,K, and may be reduced compared to the planktonic inoculum.
Host neutrophil response. The local tissue and systemic neutrophil response to the burn and P. aeruginosa infection are shown in Fig. 7. MPO activity confined to the burn wound, shown in Fig. 7A, was the same between all inoculation groups on each day. Generally, the MPO activity increased with time after the burn, but addition of P. aeruginosa did not further elevate Neutrophil MPO activity in the local tissue, regardless of inoculation level. Similar results were observed on a systemic level in the Neutrophil counts obtained from cardiac puncture as seen www.nature.com/scientificreports www.nature.com/scientificreports/ in Fig. 7B with only the 1 × 10 4 CFU/wound inoculum showing increased circulating Neutrophils as compared to the un-inoculated burn wounds on POD 11.

Discussion
We previously updated the Walker-Mason rat scald burn model to document the growth of P. aeruginosa biofilms in deep partial-thickness burn wounds 22 . This current study builds on that foundation, extending our P. aeruginosa infected animal models to include full-thickness burn injuries. In our burn time optimization study, reported previously, we identified that the time required to achieve a uniform full-thickness burn in Sprague-Dawley rats was 6 s at 99 °C. To develop biofilm infections in the partial-thickness burn model, we inoculated the wounds with three different levels of P. aeruginosa, namely 1 × 10 3 , 1 × 10 4 , 1 × 10 5 CFU/wound. Given the concerns on animal survival with large (1 × 10 6 CFU/wound) P. aeruginosa inoculums in full-thickness burn injuries shown in earlier studies 11,12,32 , we limited the inoculum to only 1 × 10 3 and 1 × 10 4 CFU/wound. The reduced pathogen numbers avoided the lethality associated with greater bacterial inoculums as documented in previous literature [11][12][13]32 . We also included additional administration of proactive pain relief (Buprenorphine SR LAB) and lactated Ringers www.nature.com/scientificreports www.nature.com/scientificreports/ resuscitation to all animals remaining in the study beyond PODs 3 and 7. These additional treatments were provided to every animal in those groups in order to maintain uniformity between the different study groups.
Similar to partial-thickness burn wounds, we observed robust infections develop in all the full-thickness scald burns, plateauing above 1 × 10 8 CFU/g of wound tissue. The wounds inoculated with sterile PBS showed the growth of the native microflora in the burn eschar without any culturable P. aeruginosa over the 11 days of the study. Compared to partial-thickness burn wounds infected with P. aeruginosa, less tissue separation and sloughing was observed in the full-thickness burn wounds on PODs 7 and 11. However, the full-thickness burns infected with P. aeruginosa showed a greater degree of tissue necrosis as seen in the gross wound images in Fig. 1. Interestingly, the wounds inoculated with P. aeruginosa showed greater variability in recoverable CFU's of P. aeruginosa compared to our previous work with partial-thickness burns. In the second experimental run, the bacterium was notably absent on both P. aeruginosa isolation and standard blood agar plates of the 1 × 10 4 CFU inoculum on POD 1 and the 1 × 10 3 CFU inoculum on PODs 3, 7, and 11. This was also confirmed by qPCR for the bacterial DNA indicating that this was not a plating error. However, the bacterium was observed in histological sections of those samples as well as at low levels in end point PCR for the oprL gene from different biopsy punches from the same burn wound. Presumably, in that experimental run the native flora may have suppressed the growth or recovery of P. aeruginosa, because the first and third runs showed greater recovery of the pathogen from the biopsy punch samples of the 1 × 10 3 CFU/wound inoculum group. Therefore, sampling errors could be a contributing factor to the recovery of the pathogen in that study group. Conversely, the 1 × 10 4 CFU/wound inoculum showed better consistency between all three runs, suggesting an optimal infectious dose of P. aeruginosa to surface infect full-thickness burn wounds without the mortality observed in previous studies 11,12 . This inoculation method is important because it did not require injection of the pathogen under the burn eschar to establish the biofilm infection, which is commonly used in many murine and rodent burn infection studies 6,7,9,10 . Additionally, the inoculum was spread over the burn surface as planktonic cells and did not require the use of a biomaterial 10 . This model recapitulated the invasive P. aeruginosa burn wound infection using a much lower inoculum of the pathogen than previously used, potentially mimicking a more clinically relevant burn infection model than others described 11,12 .
We utilized a new assay that directly measures mRNA to determine expressional changes of important biofilm and virulence associated genes of P. aeruginosa in the context of full-thickness burn wound infection. The QuantiGene technology selectively binds mRNA and amplifies the signal of the capture beads specific for the gene of interest. This method avoids the typical challenges associated with reverse transcription of mRNA into cDNA used in qRT-PCR, but instead directly binds to the target mRNA produced by the bacterial cells. The results were expressed as fold change over the planktonic inoculum. The genes involved in alginate biosynthesis (algD and alg8) and export (algE) increased throughout the infection. This indicated that P. aeruginosa produced and exported the exopolysaccharide alginate, an important constituent of its mature biofilm matrix. Similar increases occurred in genes associated with biofilm matrix component pel (pelB, pelC, and pelD). However, little to no increase in psl exopolysaccharide (pslA and pslC) production occurred in the infection. This may be due to the competence this clinical P. aeruginosa strain has for producing psl. Similar differences in pel and psl expression have been reported in laboratory P. aeruginosa strains PAO1 and PA14 24,33-36 . These two P. aeruginosa strains generally express different biofilm matrix components, namely psl (PAO1) and pel (PA14). The strain used in this study may possibly be more pel dominant or potentially deficient for producing psl. We previously characterized www.nature.com/scientificreports www.nature.com/scientificreports/ the the genome sequence of this clinically isolated pathogen 37 , but work is currently ongoing to further understand the in vitro characteristics of this pathogen namely its biofilm formation, motility, and antibiotic resistance genes compared to several other clinical and laboratory isolates of P. aeruginosa (manuscript in preparation). Lastly, expression of three virulence genes (lasA, lasB, and pvdS) all showed increases upon in vivo infection, indicating adaptation to the wound environment by increasing production of siderophores to liberate iron from host tissues and elastase to degrade the host tissue matrix [38][39][40][41] . Scavenging bound iron from the blood and tissues within the burn wound is an important step for P. aeruginosa biofilm formation 42,43 . Taken together the increased expression of biofilm exopolysaccharides along with factors that degrade host tissues and liberate important nutrients indicates active biofilm development by P. aeruginosa in full-thickness burn wounds.
Initial characterization of the host response to the tissue damage resulting from the full-thickness burn and subsequent P. aeruginosa infection suggests marked differences from partial-thickness burns. Our work in deep partial-thickness burns showed both a local and systemic, inoculum-dependent, neutrophil response to the P. aeruginosa infection. Interestingly, in the full-thickness burn, a similar elevation of neutrophil activity with P. aeruginosa infection, regardless of inoculum level, above the burn trauma itself was notably absent. A general trend of increased MPO activity was observed as time progressed in response to the burn, but not necessarily to the presence of P. aeruginosa in the burn eschar. Additionally, only on POD 11 was an increase in systemic neutrophil counts seen in only the 1 × 10 4 CFU/wound inoculum group as compared to all the other inoculation groups. The increased neutrophil counts were accompanied with a large standard deviation, because the two samples from the final experimental run of the 1 × 10 4 CFU/wound inoculum clotted before they could be analyzed, while the first run showed lower counts (~10 × 10 3 neutrophils/µl) and the second run had higher counts (~34 × 10 3 neutrophils/µl). Comparatively, in the partial-thickness burn with P. aeruginosa infection, we previously reported a trend of increased systemic neutrophils by POD 7 and significant increases on POD 11 22 . The lack of neutrophil response to the subsequent infection of full-thickness burns with P. aeruginosa may be due to the lack of blood flow or circulation in the burn eschar. In the deep partial-thickness burn model, roughly the bottom third of the dermis remained vitalized, while the full-thickness burn dermis was devoid of blood circulation. In larger TBSA rat burn models, slow accumulation of neutrophils into full-thickness burned skin, due to limited number of blood vessel beneath the burned dermis, has previously been shown 44 . Similarly, sequestration of neutrophils into distal injury sites (polyurethane sponges or bacterial skin lesions) from the burn was reduced with concomitant presence of full-thickness burn injuries 45,46 . These intrinsic features of the burn wound, albiet in larger TBSA burns, may have reduced the host's ability to interact with P. aeruginosa or native microflora in the burn injury, thereby shielding the invading microbes from the innate immune response. Further, in-depth analysis (local and systemic cytokine expression, histologic scoring of infected verus un-inoculated tissue sections, circulating damage/pathogen associated molecular patterns, and systemic immune cell counts) of the host response to both partial-and full-thickness burns in the context of P. aeruginosa biofilm burn wound infection will be reported in a future manuscript.
In this report, we detail the course of P. aeruginosa infection in a 10% TBSA full-thickness scald burn with Sprague-Daweley rats. Robust infections developed using relatively low inoculums of 1 × 10 4 P. aeruginosa cells spread over the surface of the burn wound instead of injecting the inoculum beneath the burn eschar. The bacterial infection resulted in the natural development of P. aeruginosa biofilms in the burn eschar, coupled with systemic infection by 11 days post burn and inoculation. This work further enhances our portfolio of rat scald burn biofilm infection models for testing the effectiveness of novel and easily deployable anti-biofilm treatments in the context of both severe partial-and full-thickness burn wounds that afflict combat soldiers on the battlefield.

Materials and Methods
Animal ethics statement. Research

Full-thickness burn and infection. Full-thickness burns were made on the dorsum of 61 male
Sprague-Dawley rats using the method we previously described 22 . The burn area averaged ~10% of the total body surface area based on a body weight range between 350-450 g in Meeh's formula, = A kW 2 3 , where A is total body surface area (cm 2 ), k is Meeh's constant (9.46), and W is body weight (g) 47 . In brief, the animals were acclimated to the facility for 2 weeks prior to the burn. The day before the burn, the dorsal surface of anesthetized (Forane, Baxter Healthcare Corporation, Deerfield, IL) rats was shaved and depilated (Nair, Church & Dwight Co, Ewing, NJ). Additionally, 1.2 mg/kg of Buprenorphine SR LAB (Zoopharm Pharmacy) was subcutaneously administered for proactive pain management. Post-shaving, the rats were housed individually until the end of the experiment.
Immediately prior to the burn, rats were anesthetized with 2.5% isoflurane (Forane) for 15 minutes in an induction chamber followed by application of eye lube (Artificial Tears Ointment, Akorn, Inc., Lake Forrest, IL). Pulse rate and blood oxygen levels (2500A Vet Pulse Oximeter, Nonin Medical, Inc., Plymouth, MN) were monitored throughout the procedure. The anesthetized rats were positioned into a custom burn template, which was lowered into 99 °C water bath for 6 seconds. We previously showed that this exposure time produced full-thickness burns in Sprague-Dawley rats 22 . Immediately post-burn, the burn area was blotted on damp (room temperature) paper towels to remove excess water and the first of four fluid resuscitations (warm 4 ml Lactated Ringers Solution (LRS) based on Parkland's Formula) given via intraperitoneal injection. The remaining resuscitations were given within 36 hours following the burn. Burn wounds were imaged using a Nikon D90 with Nikkor AF-S lens (18-105 mm, A:3.5-5.6 G) following the burn and on the day of euthanasia.
Immediately after the burn and first resuscitation, P. aeruginosa strain 12-4-4(59) suspended in 1 × phosphate-buffered saline (PBS) was inoculated on the burn surface at 1 × 10 3 or 1 × 10 4 CFU/wound 11,37 . The bacterial cells were grown overnight at 37 °C in Trypticase Soy Broth (TSB; Becton, Dickerson and Company, Sparks, MD) and subcultured the day of the burn in fresh TSB to mid-logarithm growth phase. Prior to inoculation, the bacterial cells were centrifuged at 3,000 × g for 15 minutes at 4 °C and re-suspended in PBS to the correct cell density. A 100 µl P. aeruginosa suspension was spread over the burn area with a sterile pipette tip. Un-inoculated control wounds received 100 µl of sterile PBS. In total 61 animals comprised the study, with 12 serving as un-inoculated controls and 24 in each P. aeruginosa inoculum group. The study was repeated three times. Each experimental repeat consisted of one third of the animals in each group, distributed between four end points, post-operative day (POD) 1, 3, 7, and 11. One animal in the 1 × 10 4 CFU/wound inoculum required a replacement due to early removal from the study on POD 5.
To prevent disruption of the burn surface by the animals, the wounds were sealed with Tegaderm TM Film (3M Health Care, St. Paul, MN) using NOTAPE professional silicone bonding adhesive (Vapon, Inc. Fairfield, NJ). N-terface Wound Contact Layer (Winfield Laboratories, Inc. Richardson, TX) separated the burn surface and the Tegaderm TM Film. Rats were additionally placed into composite rat jackets, described previously for the duration of the study 22 . A ThermoCare Portable Animal Intensive Care Unit (Daisy Products LLC, Paso Robles, CA) set to 37 °C was used to recover the animals after the burn until the effects of the anesthesia wore off. Additional fluid resuscitation (LRS given intraperitoneally) and pain relief (Buprenorphine SR LAB given subcutaneously) beyond the initial 36 hours following the burn were provided on PODs 3 and 7 to rats remaining in the study beyond those time points. Daily evaluation of body weight and symptoms of pain/distress were made throughout the study. Upon removal of the Tegaderm and N-terface dressings, the wounds were imaged and entire burn area excised from the dorsum. Three (cranial, middle, and caudal) separate histologic cross-sections of the burn area were fixed in 10% buffered formalin in PBS (Fisher Diagnostics, Kalamazoo, MI) for 48 hours and processed for routine embedding in paraffin wax. Biopsy punch (7 mm) samples of the burn area were isolated for quantitative bacteriology, scanning electron microscopy, myeloperoxidase (MPO) activity, and bacterial mRNA analysis. Internal organs (spleen, liver, kidney, and lungs) were also recovered. Tissue and organ samples, except those for quantitative bacteriology, were flash frozen in liquid Nitrogen and kept at −80 °C prior to analysis.
Quantification of bacterial load. The samples (biopsy punches) used to determine the bacterial load within the burn tissue were homogenized in 1.0 ml of PBS in MagNA Lyser Green Beads tubes (Roche Diagnostics GmbH, Mannheim, Germany) using a FastPrep ® -24 Tissue Homogenizer (MP Biomedicals, LLC. Santa Ana, CA). The homogenized samples were plated in duplicate on Trypticase soy agar containing 5% sheep's blood (Becton, Dickinson and Co. Sparks MD) and P. aeruginosa isolation agar (Hardy Diagnostics, Santa Maria, CA). Viable bacterial colonies were counted and plotted as log 10 (CFU/g wound tissue) ± Standard Deviation.
Bacterial DNA was also isolated from the same biopsy punch samples for viable CFU quantification using PCR as previously described 22 . Fifty microliters from each of the four-biopsy sample tissue homogenates per animal were pooled and DNA isolated using the DNeasy Blood & Tissue Kit (Qiagen, Valencia, CA). To quantify bacterial load in the burn eschar, primers and probes for outer membrane lipoprotein oprL of P. aeruginosa 49 and universal primers and probes for the 16S rDNA of gram-negative bacterial cells were used 50 . Primer and probe sequences are listed in Supplemental Table 1. Standard curves were generated from mid-log growth phase cultures of P. aeruginosa. Genomic DNA concentration was measured using a Quant-iT ds DNA BR Assay Kit (Invitrogen, Carlsbad, CA) and confirmed using gel separation. All samples were run in triplicate on a StepOne Plus Real-Time PCR System (Applied Biosystems, Inc. Foster City, CA) using a total volume of 20 µl of TaqMan Gene Expression Master Mix (Applied Biosystems). Concentrations of forward and reverse primers and TaqMan MGB probe were set to 100 nM along with 5 µl of sample template. Target DNA was amplified using a standard qPCR protocol of a 5 minute 95 °C denaturation step, followed by 40 PCR amplification cycles (95 °C for 15 s and 60 °C for 60 s). The resulting data was analyzed using the StepOne software provided with the instrument. The genome copy number was calculated from the amount of genomic (P. aeruginosa or total) DNA and normalized by the weight of the tissue to provide CFU/g tissue.
Histology. Thin 5 µm sections of fixed rat burn eschar were stained with hematoxylin-eosin (H&E) or H&E with Giemsa to evaluate tissue damage and P. aeruginosa invasion into the tissue. Images were taken with both the 20× and 63× objectives of a Leica Aperio Versa 200 slide scanner (Leica Biosystems, Inc. Buffalo Grove, IL). Additionally, peptide nucleic acid (PNA)-fluorescent in situ hybridization (FISH) was used to confirm the presence of P. aeruginosa deep within the burned tissue section using a standardized method in our laboratory 51 . The QuantiGene plex gene expression assay for key P. aeruginosa biofilm and virulence genes. We used the QuantiGene Plex Assay (Assay ID M17102001, Affymetrix, Inc. Santa Clara, CA) to determine the expression level of key biofilm structural genes (algD, alg8, algE, pelB, pelC, pelD, pslA, and pslC) and virulence genes (lasA, lasB, and pvdS) using three distinct housekeeping genes (proC, fabD, and rpoD) following the manufacturer's procedure for Frozen Tissue Homogenates. Briefly, samples were pulverized under liquid Nitrogen using a Bessman Tissue Pulverizer (Spectrum, Inc. Rancho Dominguez, CA) and lysed for 15 minutes at 65 °C with lysis mixture (cold TES Buffer with Lysozyme and Proteinase K Homogenization solution). The supernatants of the tissue homogenates were stored at −80 °C, after centrifuging at 4,000 rpm for 5 minutes, until analysis. The tissue homogenates were thawed and incubated at 37 °C for 30 minutes prior to mixing with the Working Bead Mix, containing Proteinase K and the Capture Beads, in a hybridization plate and incubated at 54 °C for 18 hours with gentle shaking at 600 rpm. After the overnight incubation, the sample was transferred to the Magnetic Separation Plate. Using the Hand-held Magnetic Plate Washer, the samples were washed 3× with Wash Buffer in between 1 hour incubations at 50 °C with Pre-Amplifier Solution, Amplifier Solution, and Label Probe Solution. After the Label Probe Solution, SAPE Working Reagent was added for 30 minutes at room temperature. After washing the plate 3× with SAPE Wash Buffer, 130 µl of SAPE Wash Buffer was added to each sample and mixed for 3 minutes at 800 rpm followed by immediate reading on a Bio-Plex 200 System with Bio-Plex Manager TM Software Version 6.1 Build 727 (BioRAD Laboratories, Inc. Hercules, CA).

Quantification of myeloperoxidase activity in burn tissue. A Bessman Tissue Pulverizer was used
to fracture biopsy punch samples under liquid Nitrogen. The samples were further homogenized as directed by the Fluoro MPO Myeloperoxidase Detection Kit (Cell Technology, Inc. Mountain View, CA). Briefly, an IKA T10 basic Ultra Turrax tissue homogenizer (IKA Works, Inc. Wilmington, NC) was used to process the samples in MPO homogenization buffer, provided in the kit. The samples were then centrifuged at 12,000 × g for 20 minutes at 4 °C and re-suspended in MPO solubilization buffer containing 0.5% Hexadecyltrimethylammonium (Sigma-Aldrich, Inc. St. Louis, MO). The re-suspended samples were homogenized again and sonicated using a Sonic Dismembrator Model 100 (Fisher Scientific, Kalamazoo, MI) for 30 s followed by two cycles of freezing/ thawing before isolating the supernatant. Equal volumes of sample or MPO enzyme standard were mixed with MPO reaction cocktail and incubated in black opaque, clear bottom, microtiter plates for 30 minutes at room temperature. Fluorescence (530/590 nm) was measured using a BioTek Synergy HT with Gen5 Software (BioTek Instruments, Inc. Winooski, VT). MPO activity of each sample was calculated from the MPO enzyme standard curve and plotted as MPO activity in milliUnits (mU) ± Standard Deviaiton. Statistical analysis. All data were plotted and analyzed using GraphPad Prism 7.03 (GraphPad Software, Inc. San Diego, CA). Comparisons were performed using two-way ANOVA with Tukey's correction for multiple comparison and an α set to 0.05. Data was graphed as the mean ± Standard Deviation (SD).

Disclaimer.
The opinions or assertions contained herein are the private views of the author and not to be construed as official or as reflecting the views of the Department of the Army or the Department of Defense.

Data Availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.