A Murine Skin Infection Model Capable of Differentiating the Dermatopathology of Community-Associated MRSA Strain USA300 from Other MRSA Strains

USA300 is a predominant and highly virulent community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) strain that is a leading cause of skin and soft tissue infections. We established a murine intradermal infection model capable of demonstrating dermatopathological differences between USA300 and other MRSA strains. In this model, USA300 induced dermonecrosis, uniformly presenting as extensive open lesions with a histologically documented profound inflammatory cell infiltrate extending below the subcutis. In contrast, USA400 and a colonizing control strain M92 caused only localized non-ulcerated skin infections associated with a mild focal inflammatory infiltrate. It was also determined that the dermonecrosis induced by USA300 was associated with significantly increased neutrophil recruitment, inhibition of an antibacterial response, and increased production of cytokines/chemokines associated with disease severity. These results suggest that induction of severe skin lesions by USA300 is related to over-activation of neutrophils, inhibition of host antibacterial responses, and selective alteration of host cytokine/chemokine profiles.


Introduction
Methicillin-resistant Staphylococcus aureus (MRSA) is a leading cause of infection worldwide. It was traditionally considered a nosocomial pathogen in healthcare facilities (known as hospital-associated MRSA (HA-MRSA)). However, community-associated MRSA (CA-MRSA) emerged in the 1990s, and quickly increased and then replaced HA-MRSA in the United States and Canada. Prior to 2000, the predominant CA-MRSA strain in North America was pulsotype USA400 (ST1-MRSA-IV). It was quickly replaced by pulsotype USA300 (ST8-MRSA-IV), which is the strain responsible for the increase in CA-MRSA infections in North America [1,2], as well as globally [3][4][5]. USA300 causes primarily acute bacterial skin and skin structure infections (ABSSSIs) [6], with an estimated 44.6% of ABSSSIs in North America caused by S. aureus [7], and the majority of those caused by USA300 [8].

MRSA Strains and Their Phenotypic and Genotypic Characterization
USA300-C2406 was isolated from a patient with a lethal case of necrotizing pneumonia during our local CA-MRSA strain USA300 outbreak in Calgary in 2004 [21]. Control strain USA400-CMRSA7 was provided by the National Microbiology Laboratory, Health Canada, Winnipeg, Canada [22]. Control strain M92 was found as a nasal colonizer of staff in our local hospitals, but was never associated with invasive infection. Phenotypic and genotypic characterization of the isolates was done as previously described [23]. Whole genome sequencing was performed for the isolates using Pacific Biosciences (PacBio, Menlo Park, CA, USA) RSII sequencing technology, at the McGill University Genome Quebec Innovation Center. Genome assembly was accomplished using the hierarchical genome assembly process (HGAP v. 2.3.0.140936.p5), with the complete genome of C2406 (GenBank No.: PRJNA345240; CP019590.1), CMRSA7 (GenBank No.: PRJNA362898), and M92 (GenBank No.: PRJNA319679; CP015447.1) submitted to GenBank and annotated using the prokaryotic genome annotation pipeline [24,25]. Virulence gene profiles for the isolates, including 17 toxin genes, 12 adhesin genes, and 4 exoenzyme genes, were identified and analyzed through the strain whole genome sequences.

Murine Skin Infection Model
Single colonies of each bacterial strain were grown overnight in brain-heart infusion (BHI) broth at 37 • C, followed by subculture in 50 mL BHI at 37 • C until the optical density at 600 nm (OD600) was 0.7. Bacteria were centrifuged and the pellet washed twice with saline, then resuspended in 0.5 volume of saline (roughly 2 × 10 8 CFU/mL). Animal infection experiments were performed at the Animal Resource Centre at the University of Calgary, in accordance with institutional and national guidelines of the Canadian Council on Animal Care, under protocol numbers M06074, M09115, and AC13-0076 (approved by the Animal Care Committee, University of Calgary). The experiments were repeated 3 times. Female BALB/c mice (Charles River Laboratories, Inc., Wilmington, MA, USA), aged 6 to 8 weeks, were shaved in the intrascapular region with electrical clippers prior to injection. Fifteen mice were assigned to each of the 4 groups and injected intradermally with 1 × 10 7 CFU/50 µL of the appropriate strain or mock-infected control (saline control), in the center of the shaved area. Mice were carefully monitored for skin (wound) infection and signs of distress, and euthanized on days 4, 7, and 15~17. The skin lesion areas (mm 2 ) were estimated by multiplying the width (mm) and length (mm) of the lesion. For each mouse, the whole spleen was aseptically harvested and homogenized in 1 ml sterile saline. Quantitative cultures [serial dilutions and spread on Tryptic soy agar (TSA)plates] were performed to determine bacterial load, as splenic bacterial load was used as an indicator of systemic MRSA dissemination. Full thickness biopsies of the core cutaneous lesions were performed, fixed in 10% neutral buffered formalin, and subjected to detailed histopathological examination.

Histology
The lesional skin tissue was cut into tissue blocks by the Airway Inflammation Research Group (AIRG) Histology Services at the University of Calgary. Tissue sections 4 µm thick were affixed to microscope slides and deparaffinized. The sections were stained with hematoxylin and eosin (H&E), Gram, and chloracetate esterase staining for routine histology, to identify bacteria, and to highlight neutrophils, respectively. For the H&E staining (NovaUltra TM H&E Stain Kit, Rockville, MD, USA), fixed slides were deparaffinized by 2 changes of xylene for 5 min, and re-hydrated in 2 changes of 100% alcohol for 5 min each, then 95% and 70% alcohol for 1 min each. The slides were stained with Mayer's hematoxylin solution for 2 min, washed with running tap water for 2 min, and rinsed in 95% alcohol. The slides were further counterstained with eosin solution for 45 s, and rinsed in 95% alcohol. The stained slides were dehydrated through 2 changes of 100% alcohol, 5 min each, and cleared with 2 changes of xylene, 5 min each. Finally, the slides were mounted with xylene-based mounting medium. For the Gram staining (BD Gram Stain Kits, Franklin Lakes, NJ, USA), fixed slides were deparaffinized and re-hydrated as described above. The slides were stained with Gram crystal violet for 1 min, and flooded with Gram iodine for 1 min. The slides were decolorized by Gram decolorizer, then counterstained with Gram safranin for 1 min. For the chloracetate esterase staining, deparaffinized sections were stained with the freshly prepared staining solution (5.0 mg naphthol AS-D chloracetate, 5 mL N-N dimethylformamide (NNDMF), 6 drops 4.0% sodium nitrite, 6 drops 4% new fuchsin in 47.5 mL phosphate buffer) for 25 min, followed by counterstaining with hematoxylin nuclear stain. The stained slides were dehydrated and cleared as described above.

Myeloperoxidase (MPO) Assay
The tissue myeloperoxidase activity assay was performed as an index of neutrophil recruitment, as previously described [26]. The test was repeated twice, with at least 5 mice from each group and each time point included from each experiment. Briefly, the skin was cut and weighed prior to homogenization in a 0.5% hexadecyltrimethylammonium bromide phosphate-buffered (pH 6.0) solution using a polytron PT1300D homogenizer (Kinematica, Lucerne, Switzerland). The homogenates were centrifuged at 14,000 rpm for 5 min at 4 • C in a microcentrifuge and five aliquots of each supernatant were transferred into 96-well plates, followed by the addition of a 3, 3 -dimethoxybenzidine and 1% hydrogen peroxide solution. Standard dilutions of pure myeloperoxidase were also tested for their activity to construct a standard curve (OD as a function of units of enzyme activity). Optical density readings at 450 nm were taken at 1 min (which corresponds to the linear portion of the enzymatic reaction) using a Spectra Max Plus plate reader using SOFTmax Pro v. 3.0 software (Molecular Devices Corp., Sunnyvale, CA, USA). Myeloperoxidase activity was expressed as units of enzyme per gram of tissue.

Spinning Disk Confocal Microscopy
An amount of 10 7 /CFU of bacteria were injected intradermally into the lateral abdominal skin, to avoid forming lesions on the mid-line. Examination of the microcirculation of the lateral abdominal skin was prepared for microscopy in order to determine the recruitment of immune cells in lesions, as described [27]. Flank skin microvasculature was visualized using a spinning disk confocal microscope, using an Olympus BX51 upright microscope with an x20/0.95 XLUM Plan Fl water immersion objective. The microscope was equipped with a confocal light path (WaveFx; Quorum, ON, Canada) based on a modified Yokogawa CSU-10 head (Yokogawa Electric Corporation, Tokyo, Japan). Anti-CD4-FITC (L3T4; BD Biosciences; 3 µg/mouse) and anti-mouse Gr-1 efluo660 (RB6-8C5; eBioscience, CA, USA; 2 µg/mouse) and anti-mouse CD31 coupled to Alexa647 (390; BD Biosciences conjugated via an Invitrogen protein labeling kit: 10 µg/mouse) were injected intravenously (IV) into BALB/c mice to image CD4+ T lymphocytes, neutrophils, and endothelial cells, respectively. Both 491 and 640 nm laser excitation wavelengths (Cobalt, Stockholm, Sweden) were used in rapid succession and visualized with the appropriate long-pass filters (520 +/− 35 nm and 640 +/− 40 nm, respectively, Semrock, IL, USA). Typical exposure time for both excitation wavelengths was 900 ms. A 512 × 512 pixel backthinned electron-multiplying charge-coupled device camera (C9100-13, Hamamatsu, Japan) was used for fluorescence detection, with Volocity Acquisition software (Improvision, MA, USA) used to drive the confocal microscope. CD4 T lymphocyte and neutrophil rolling and adhesion were simultaneously assessed in 5-10 random fields of view in the postcapillary venules of the skin. The adherent cells were defined as cells which remained in the same location for 30 s in 100 µm of the venule, while the emigrated cells were defined as the cells which stayed outside of the vasculature in the field of view (FOV).

In Vivo Neutrophil Depletion
To determine the role of neutrophils in USA300-C2406 infection, a depletion process was employed using the antibody RB6-8C5 (BioXCell, West Lebanon, NH, USA), which mainly depletes neutrophils, at different time points in mice with a USA300-C2406. There were at least 3 mice for each group and each time point. As noted above, mice were injected with 10 7 CFU of USA300-C2406 intradermally in the intrascapular region on day 0. Mice in the early depletion groups received an intraperitoneal (i.p.) injection of 200 µg RB6-8C5 24 h before infection, followed by injections every 48~72 h. Mice in the late depletion groups received an i.p. injection of 200 µg RB6-8C5 24 h after infection, followed by injections every 48~72 h. Mice in the control group were injected i.p. with 200 µg of rat IgG2 isotype control monoclonal antibody (BioXCell), on the same schedule with the depletion antibody RB6-8C5. All mice were carefully monitored (including obtaining weights), then euthanized on days 4, 7, and 14 post-infection. Lung, liver, and spleen were harvested and homogenized in 1 mL sterile saline, then quantitative cultures (serial dilutions and spread on TSA plates) were performed to determine bacterial load, as described above. The skin samples were processed into tissue sections and affixed to microscope slides for Gram staining, as described above.

Mouse Antibacterial Response PCR Array
To assess whether the bacterial strains could inhibit host immune responses, the transcriptional levels of 84 host genes related to the antibacterial response were determined in local skin using a Qiagen mouse antibacterial response PCR array (Qiagen Inc., Germantown, MD, USA). Included were genes involved in Toll-like receptor (TLR) signaling, NODlike receptor (NLR) signaling, apoptosis, inflammatory response, cytokines/chemokines, and antimicrobial peptides. Samples from at least 3 mice in each group on day 4 were assessed. RNA from lesion samples was isolated using the RNeasy Plus Mini Kit (Qiagen Inc.). Ten milligrams of each sample were homogenized in RLT buffer. The contaminating DNA was removed with gDNA eliminator spin columns. RNA was purified by washing with RW1 buffer, and collected with the RNeasy spin column. RNA was further purified and reverse transcribed to cDNA using the RT2 First Strand Kit (Qiagen Inc.). Eight microliters of raw RNA were mixed with 2 µL buffer GE and incubated for 5 min at 42 • C to eliminate contaminating DNA. Four microliters of buffer BC3, 1 µL control P2, 2 µL RE3 reverse transcriptase mix, and 3 µL RNase-free water were added and the reaction was incubated at 42 • C for 15 min. The reaction was stopped by incubation at 95 • C for 5 min. The expression of 84 host genes in the mouse antibacterial response PCR array (Qiagen Inc.) was assessed by qRT-PCR using RT2 SYBR Green qPCR Mastermix (Qiagen Inc.) on a CFX96 Real-Time Detection System (Bio-Rad, Hercules, CA, USA). Thermal cycle conditions were performed as described for the RT2 Profiler PCR Array (Qiagen Inc.). One hundred and two microliters of the cDNA synthesis reaction were mixed with 1350 µL 2× RT2 SYBR green mastermix and 1248 µL RNase-free water. Twenty-five microliters were then aliquoted into each well of the RT2 Profiler PCR Array. The PCR plates were heated to 95 • C for 15 min, followed by 40 cycles of denaturation-extension (95 • C for 15 s and 60 • C for 1 min). Because the transcription of housekeeping genes during wound forming and healing varies [28], expression of the 84 genes was normalized to the mean of the 5 most stably expressing housekeeping genes (actb, b2m, gapdh, gusb, and hsp90ab1). The PCR array data were further uploaded to the Qiagen RT2 Profiler PCR Array Data Analysis Center (http://www.qiagen.com/geneglobe. Samples). Ct = 35 was set as the cut-off Ct value, while 5 housekeeping genes were selected as normalization factors. Relative target gene expression was calculated according to the ∆∆Ct method [29], in which the fold difference in expression was 2 -∆∆Ct .

Cytokine and Chemokine Assay
The production of various cytokines and chemokines in local skin samples was determined using a Luminex 32-plex by Eve Technologies (University of Calgary, AB, Canada). Sigma protease inhibitor 1:100 buffer. The supernatant was collected and cytokines and chemokines were separated into several groups for analysis, as follows: (1) protective cytokines and chemokines whose depletion results in severe infection in a mouse skin infection model [30,31]; (2) cytokines and chemokines related to S. aureus disease severity, but which lack depletion experiments to confirm their role in a mouse skin infection model [32][33][34][35][36][37][38]; (3) growth factors which are important for wound healing;(4) cytokines and chemokines related to the Th1 and Th2 response; and (5) cytokines and chemokines whose roles are unclear in MRSA infection. The latter was further divided into 2 sub-groups; one sub-group generally believed to be related to tissue injury, infection, and allergic disease and the other believed to be related to immune cell development, virus infection, and aging.

Statistical Analysis
All analyses were performed using SPSS 20.0 and Prism v. 5. Differences in CFU, diameter of lesion, and MPO activity were evaluated using one way ANOVA. The Qiagen RT2 Profiler PCR Array Data Analysis Center was used to analyze the Qiagen PCR Array. p values of p < 0.05 were considered statistically significant.

Genotypic and Phenotypic Characterization and Virulence Gene Profiles of MRSA Strains
The genotype, antibiotic resistance ( Figure 1A), and virulence factor profiles ( Figure 1B) of USA300-C2406 were compared to those of the control strains. Colonization strain M92 belonged to ST239-MRSA-III, carried a non-typeable spa gene, and was Panton-Valentine leucocidin (PVL) negative and agr type I. USA400-CMRSA7 belonged to pulsotype USA400 (ST1-MRSA-IVa), was PVL positive, belonged to spa type t128, and agr type III. USA300-C2406 belonged to pulsotype USA300 (ST8-MRSA-IVa), was PVL positive, belonged to spa type t008, and agr type I. Phenotypic tests indicated that USA400-CMRSA7 was sensitive to erythromycin and ciprofloxacin, unlike USA300-C2406, which was resistant to both. Likewise, M92 was resistant to clindamycin, gentamicin, and tetracycline, while USA300-C2406 was sensitive to these agents. Virulence factors were also assessed and compared between the strains. They were similar, with the differences being that USA300-C2406 carried the chp gene and CMRSA7 carried the sea and seh genes. was sensitive to erythromycin and ciprofloxacin, unlike USA300-C2406, which was resistant to both. Likewise, M92 was resistant to clindamycin, gentamicin, and tetracycline, while USA300-C2406 was sensitive to these agents. Virulence factors were also assessed and compared between the strains. They were similar, with the differences being that USA300-C2406 carried the chp gene and CMRSA7 carried the sea and seh genes.

USA300 Induced Extensive Open Lesions (Dermonecrosis) while M92 and USA400 Caused Localized Infection
Our mouse skin infection model differentiated USA300-C2406 hypervirulence from the control strains in that it demonstrated more severe skin lesions than M92 and USA400-CMRSA7 ( Figure 2). Intradermal inoculation in either M92 or USA400-CMRSA7 resulted in confined dermal abscesses with average areas of 15.5 mm 2 and 16.36 mm 2 , respectively. The sizes of these abscesses remained relatively unchanged from day 1 to day 7. Healing began after day 7, with full recovery on day 14~17. In contrast, USA300-C2406 caused abscesses with an average area of 51.40 mm 2 (p < 0.01 compared with M92 and USA400-CMRSA7) on day 1. This was followed by the development of an ulcerated open wound with underlying necrosis that reached its maximum size on day 5. The size of the lesion did not change significantly from day 4 to day 7. The wound started to heal after day 7, . sea/c/d/e/g/h/i/j/k/q, staphylococcal enterotoxin A/C/D/E/G/H/I/J/K/Q; tst, toxic shock syndrome toxin; chp, chemotaxis inhibitory protein; scn, staphylococcal complement inhibitory protein; hla/b/d/g, α/β/ δ/γ-hemolysin; clfA, clumping factor; fnbA/B, fibronectin adhesive molecule A/B; cnaA/B, collagen adhesive molecule A/B; sdrC/D/E, putative adhesin; bbp, bone sialoprotein adhesin; ebpS, elastin adhesin; map, major histocompatibility complex class II analog protein; ica, polysaccharide intercellular adhesin. coa, coagulase; V8, serine protease; hysA, hyaluronidase; sak, staphylokinase; +, positive; -, negative.

USA300 Induced Extensive Open Lesions (Dermonecrosis) while M92 and USA400 Caused Localized Infection
Our mouse skin infection model differentiated USA300-C2406 hypervirulence from the control strains in that it demonstrated more severe skin lesions than M92 and USA400-CMRSA7 ( Figure 2). Intradermal inoculation in either M92 or USA400-CMRSA7 resulted in confined dermal abscesses with average areas of 15.5 mm 2 and 16.36 mm 2 , respectively. The sizes of these abscesses remained relatively unchanged from day 1 to day 7. Healing began after day 7, with full recovery on day 14~17. In contrast, USA300-C2406 caused abscesses with an average area of 51.40 mm 2 (p < 0.01 compared with M92 and USA400-CMRSA7) on day 1. This was followed by the development of an ulcerated open wound with underlying necrosis that reached its maximum size on day 5. The size of the lesion did not change significantly from day 4 to day 7. The wound started to heal after day 7, with recovery on day 14~17. Cultures of the spleen demonstrated that there was no systemic infection in this model.  The average lesion area (mm 2 ) for M92 (blue), USA400-CMRSA7 (green), and USA300-C2406 (red) at each time point was expressed as mean + SEM. The experiment was repeated twice, with 5 mice in each time point and each group. Not significant between USA400-CMRSA7 and M92. Statistical comparisons between USA300-C2406 and M92 or USA400-CMRSA7 were indicated with asterisk: ** p <0.01.

Profound Inflammatory Cell Infiltration by USA300
The histological changes induced by various MRSA strains were assessed on day 4 since this was the point when the skin lesions reached their maximum diameter/area. The severe tissue damage induced by USA300-C2406 was associated with more disseminated inflammation when compared with that caused by M92 and USA400-CMRSA7. As shown in the Figure 3, M92-infected skin was associated with mild changes in the epidermis and dermis (panel A). The inflammatory cell infiltration extended into the adipose tissue, superficial muscle, and fascia (D and G). USA400-CMRSA7 infection formed a well-circumscribed area of skin necrosis (B), which was associated with a mild inflammatory cell in- The average lesion area (mm 2 ) for M92 (blue), USA400-CMRSA7 (green), and USA300-C2406 (red) at each time point was expressed as mean + SEM. The experiment was repeated twice, with 5 mice in each time point and each group. Not significant between USA400-CMRSA7 and M92. Statistical comparisons between USA300-C2406 and M92 or USA400-CMRSA7 were indicated with asterisk: ** p < 0.01.

Profound Inflammatory Cell Infiltration by USA300
The histological changes induced by various MRSA strains were assessed on day 4 since this was the point when the skin lesions reached their maximum diameter/area. The severe tissue damage induced by USA300-C2406 was associated with more disseminated inflammation when compared with that caused by M92 and USA400-CMRSA7. As shown in the Figure 3, M92-infected skin was associated with mild changes in the epidermis and dermis (panel A). The inflammatory cell infiltration extended into the adipose tissue, superficial muscle, and fascia (D and G). USA400-CMRSA7 infection formed a wellcircumscribed area of skin necrosis (B), which was associated with a mild inflammatory cell infiltration in the surrounding adipose tissue, superficial muscle, and fascia (E and Microorganisms 2021, 9, 287 8 of 23 H). In contrast, USA300-C2406 formed extensive lesions with ulceration (C) and dense neutrophil-rich infiltrates that extended into the dermis, adipose tissue, fascia, and skeletal muscle (F and I). The necrotic debris was located at the edge and bottom (fascial level) of the ulcer and the disseminated inflammatory cells infiltrated not only the skin but also the skeletal muscle beneath the fascial plane. phil-rich infiltrates that extended into the dermis, adipose tissue, fascia, and skeletal muscle (F and I). The necrotic debris was located at the edge and bottom (fascial level) of the ulcer and the disseminated inflammatory cells infiltrated not only the skin but also the skeletal muscle beneath the fascial plane. In contrast, USA300-C2406 induced ulceration, more tissue necrosis, and a denser, more deep-seated inflammatory cell infiltrate that extended into the skeletal muscle (F and I).

Primary Infiltrating Inflammatory Cell Type Was the Neutrophil
An esterase stain was done on samples collected on day 4 ( Figure 4). This confirmed the histological findings of a neutrophil-rich inflammatory infiltrate in the case of USA300-C2406 that extended deeper than was the case with M92 and USA400-CMRSA7. In the former, a dense neutrophilic infiltrate was located not only in the skin but also in the deep skeletal muscle.

Primary Infiltrating Inflammatory Cell Type Was the Neutrophil
An esterase stain was done on samples collected on day 4 ( Figure 4). This confirmed the histological findings of a neutrophil-rich inflammatory infiltrate in the case of USA300-C2406 that extended deeper than was the case with M92 and USA400-CMRSA7. In the former, a dense neutrophilic infiltrate was located not only in the skin but also in the deep skeletal muscle.

USA300 Induced Prolonged Periods of Excessive Neutrophil Infiltration
An MPO assay was employed to quantify neutrophil infiltration in local skin samples ( Figure 5). All three MRSA strains induced higher levels of MPO activity relative to the mock-infected control. On day 4, M92, USA400-CMRSA7, and USA300-C2406 induced 386.15 U/g, 276.75 U/g, and 244.85 U/g, respectively (p > 0.05). On day 7, however, the MPO activity of M92 and USA400-CMRSA7 was 150.88 U/g and 85.97 U/g, respectively, while the MPO activity of USA300-C2406 reached 415.61 U/g (p = 0.007 and 0.001 compared with M92 and CMRSA7, respectively). On day 17, the MPO level of M92, USA400-CMRSA7, and USA300-C2406 decreased to 3.87 U/g, 5.00 U/g, and 16.96 U/g, respectively, with no significant differences noted (p > 0.05).  . Positive cells are labeled with arrows. The skin lesion induced by USA300-C2406 presented a denser and deeper neutrophil infiltrate than did that of USA400-CMRSA7 and M92.

USA300 Induced Prolonged Periods of Excessive Neutrophil Infiltration
An MPO assay was employed to quantify neutrophil infiltration in local skin samples ( Figure 5). All three MRSA strains induced higher levels of MPO activity relative to the mock-infected control. On day 4, M92, USA400-CMRSA7, and USA300-C2406 induced 386.15 U/g, 276.75 U/g, and 244.85 U/g, respectively (p > 0.05). On day 7, however, the MPO activity of M92 and USA400-CMRSA7 was 150.88 U/g and 85.97 U/g, respectively, while the MPO activity of USA300-C2406 reached 415.61 U/g (p = 0.007 and 0.001 compared with M92 and CMRSA7, respectively). On day 17, the MPO level of M92, USA400-

Significantly Greater Neutrophil (But Not CD4 T Cell) Adhesion and Emigration Tr by USA300
Neutrophil adherence and emigration were assessed using spinning disk c microscopy ( Figure 6). Infected mice had elevated numbers of adherent and em neutrophils as compared to mock-infected controls, with greater neutrophil adhes emigration induced by USA300-C2406 in comparison to what was seen in the ot strains. On day 1, M92 and USA400-CMRSA7 induced 3.00 cells/100 µm and 8.00 c µm, respectively, while USA300-C2406 induced 11.60 cells/100 µm (p = 0.055 an compared with USA400-CMRSA7 and M92, respectively). On day 7, the number trophils induced by all three strains was reduced; M92 and USA400-CMRSA7 i 0.60 cells/100 µm and 3.60 cells/100 µm, respectively, and USA300-C2406 induc cells/100 µm, which was significantly greater than M92 and USA400-CMRSA7 (p and 0.001, respectively). On day 14, there was no significant difference among th strains. In terms of neutrophil emigration, on day 1, M92 and USA400-CMRSA7 i 8.40 cells/field and 24.56 cells/field, respectively, while USA300-C2406 induce cells/field (p = 0.000 compared with both M92 and CMRSA7). On day 7, M92 and U CMRSA7 induced 2.00 cells/field and 10.60 cells/field, respectively, while USA30 induced 20.20 cells/field (p = 0.000 compared with both M92 and USA400-CMRS day 14, there was no significance among the three strains. No significant trend served for CD4+ T cell adhesion and emigration among the three strains.

Significantly Greater Neutrophil (But Not CD4 T Cell) Adhesion and Emigration Triggered by USA300
Neutrophil adherence and emigration were assessed using spinning disk confocal microscopy ( Figure 6). Infected mice had elevated numbers of adherent and emigrated neutrophils as compared to mock-infected controls, with greater neutrophil adhesion and emigration induced by USA300-C2406 in comparison to what was seen in the other two strains. On day 1, M92 and USA400-CMRSA7 induced 3.00 cells/100 µm and 8.00 cells/100 µm, respectively, while USA300-C2406 induced 11.60 cells/100 µm (p = 0.055 and 0.000 compared with USA400-CMRSA7 and M92, respectively). On day 7, the number of neutrophils induced by all three strains was reduced; M92 and USA400-CMRSA7 induced 0.60 cells/100 µm and 3.60 cells/100 µm, respectively, and USA300-C2406 induced 7.40 cells/100 µm, which was significantly greater than M92 and USA400-CMRSA7 (p = 0.000 and 0.001, respectively). On day 14, there was no significant difference among the three strains. In terms of neutrophil emigration, on day 1, M92 and USA400-CMRSA7 induced 8.40 cells/field and 24.56 cells/field, respectively, while USA300-C2406 induced 50.60 cells/field (p = 0.000 compared with both M92 and CMRSA7). On day 7, M92 and USA400-CMRSA7 induced 2.00 cells/field and 10.60 cells/field, respectively, while USA300-C2406 induced 20.20 cells/field (p = 0.000 compared with both M92 and USA400-CMRSA7). On day 14, there was no significance among the three strains. No significant trend was observed for CD4+ T cell adhesion and emigration among the three strains.

Co-localization of Higher Bacterial Load with Neutrophil Infiltration in USA300 Infection
To determine if neutrophils were targeting bacteria, tissue slides (from day 4 lesions) were stained with Gram stain to determine bacterial localization (Figure 7). Deeper bacterial dissemination in USA300-C2406-infected tissue was noted, with bacteria found

Co-localization of Higher Bacterial Load with Neutrophil Infiltration in USA300 Infection
To determine if neutrophils were targeting bacteria, tissue slides (from day 4 lesions) were stained with Gram stain to determine bacterial localization (Figure 7). Deeper bacterial dissemination in USA300-C2406-infected tissue was noted, with bacteria found throughout the skin tissue and invading the skeletal muscle. In contrast, bacteria were predominantly located in the abscess of lesions caused by M92 and USA400-CMRSA7. There was a direct correlation between the presence of bacteria and the presence of neutrophils.
Microorganisms 2021, 9,287 12 of 22 throughout the skin tissue and invading the skeletal muscle. In contrast, bacteria were predominantly located in the abscess of lesions caused by M92 and USA400-CMRSA7. There was a direct correlation between the presence of bacteria and the presence of neutrophils.

Neutrophil Depletion Resulted in a More Severe Infection
We depleted the neutrophils in mice infected with USA300-C2406 to examine the role they played in disease progression. We observed that early neutrophil depletion (24 h before infection) resulted in a significantly more severe infection as compared to the isotype control (Figure 8). With the isotype control group, the wound formed on day 1 and remained constant with an average area of 42.53 mm 2 until day 10, at which time it gradually decreased and was healed around day 17. Similarly, with late neutrophil depletion (24 h after infection), the average wound area on day 1 was 65.67 mm 2 , which was not statistically different from the isotype control group (p = 0.216). The wound maintained an average area of 50 mm 2 until day 8, at which time it decreased, similar to the isotype control. When neutrophils were depleted prior to infection, however, the mice showed an

Neutrophil Depletion Resulted in a More Severe Infection
We depleted the neutrophils in mice infected with USA300-C2406 to examine the role they played in disease progression. We observed that early neutrophil depletion (24 h before infection) resulted in a significantly more severe infection as compared to the isotype control (Figure 8). With the isotype control group, the wound formed on day 1 and remained constant with an average area of 42.53 mm 2 until day 10, at which time it gradually decreased and was healed around day 17. Similarly, with late neutrophil depletion (24 h after infection), the average wound area on day 1 was 65.67 mm 2 , which was not statistically different from the isotype control group (p = 0.216). The wound maintained an average area of 50 mm 2 until day 8, at which time it decreased, similar to the isotype control. When neutrophils were depleted prior to infection, however, the mice showed an enlarged wound with an average area of 164.37 mm 2 on day 1 (p = 0.001 and = 0.005, compared with isotype control and late depletion group, respectively), which remained constant until day 11, then gradually reduced in size and healed by day 17 (Figure 8 B).
icroorganisms 2021, 9,287 13 of 22 enlarged wound with an average area of 164.37 mm 2 on day 1 (p = 0.001 and = 0.005, compared with isotype control and late depletion group, respectively), which remained constant until day 11, then gradually reduced in size and healed by day 17 (Figure 8 B).

Neutrophil Depletion Resulted in Bacterial Dissemination and Invasive Infection
To investigate the cause of increased tissue damage following neutrophil depletion, we used Gram staining to identify bacteria in local skin ( Figure 9A), and organ culture to identify bacterial load in the internal organs ( Figure 9B). Gram stains indicated that there were more bacteria present in the tissue in both early and late neutrophil depletion groups as compared to the isotype control group, and that the bacteria invaded deeply into skeletal muscle in depletion groups, but were limited to the skin in the isotype control group. Organ cultures demonstrated that the bacterial load in the lung, liver, and spleen was less than 1 CFU/mg in the control group, whereas the bacterial load from the early and late depletion groups ranged from 5-100 CFU/mg. Weight change during infection was used as a marker for infection severity (Figure 9C), and showed that the late depletion group had significantly more weight loss from days 3 to 4, while the early depletion group had significantly more weight loss from days 1 to 8.

Neutrophil Depletion Resulted in Bacterial Dissemination and Invasive Infection
To investigate the cause of increased tissue damage following neutrophil depletion, we used Gram staining to identify bacteria in local skin ( Figure 9A), and organ culture to identify bacterial load in the internal organs ( Figure 9B). Gram stains indicated that there were more bacteria present in the tissue in both early and late neutrophil depletion groups as compared to the isotype control group, and that the bacteria invaded deeply into skeletal muscle in depletion groups, but were limited to the skin in the isotype control group. Organ cultures demonstrated that the bacterial load in the lung, liver, and spleen was less than 1 CFU/mg in the control group, whereas the bacterial load from the early and late depletion groups ranged from 5-100 CFU/mg. Weight change during infection was used as a marker for infection severity (Figure 9C), and showed that the late depletion group had significantly more weight loss from days 3 to 4, while the early depletion group had significantly more weight loss from days 1 to 8. staining indicates that there were high bacterial loads when neutrophils were depleted. (B) USA300-C2406 bacterial counts (CFU/mg) recovered from lung, liver, and spleen at day 4, 7, and 17 time points, in early neutrophil depletion (gray bars), late neutrophil depletion (black bars), and isotype control (white bars) groups. (C) Percentage of weight change following infection with USA300-C2406 in the isotype (green), early neutrophil depletion (red), and late neutrophil depletion groups (blue), showing that severe weight loss was associated with neutrophil depletion. Weight changes were expressed as mean + SEM. Experiments with 9 mice in each group. Statistical comparisons between mock infection and early depletion (black asterisk), mock infection and late depletion (orange asterisk), early and late depletion (grey asterisk); *p < 0.05; **p < 0.01.

Unique Pattern of Molecular Transcription Results from Infection with USA300
Because USA300 infections were associated with extensive neutrophil infiltration and increased bacterial load, we hypothesized that USA300 might employ a mechanism designed to modify or inhibit the antibacterial response in the mice. Using an antibacterial response PCR array, we examined the expression of factors such as Toll-like receptor signaling, NOD-like receptor (NLR) signaling, bacterial pattern recognition receptors (PRRs), signaling downstream of antibacterial responses, apoptosis, inflammatory responses, cytokines/chemokines, and antimicrobial peptides in local skin samples. Our results demonstrated that MRSA USA300 infection could activate the transcription of most of the components involved in the antibacterial responses. However, when compared with USA400-CMRSA7 and M92, USA300-C2406 induced lower degrees of upregulation.
There was, however, no statistically significant difference in the transcription of most of these aforementioned genes, with the exception of il-18. With il-18, the transcription in USA400-CMRSA7 and M92 infected mice was close to that of the mock-infected control, while USA300-C2406 induced a 2.69-fold downregulation (p = 0.0211). The level of transcription in USA300-C2406-infected mice was downregulated 3.22-and 3.39-fold (p = 0.011 and 0.030) when compared with USA400-CMRSA7 and M92, respectively.
The detailed list of transcriptional levels for genes related to each response is summarized in Supplementary Table S1.

Discussion
There are several mouse models that have been used to study S. aureus infections [14][15][16][17][18][19]. Most of the models, however, were either designed to study the mechanisms of S. aureus virulence by comparing mutants to their corresponding wild type strains, or to study more invasive infections similar to what is seen in hospital-acquired diseases. S. aureus USA300 is the most common pathogen responsible for skin and soft tissue infections (SSTIs), which differ from wound or surgical site infections in that they generally begin as small lesions with areas of necrosis [36,39,40]. In this study, building on previously described models, we established a mouse skin infection model that mimics the clinical presentations seen with CA-MRSA SSTIs, and which is able to differentiate between two CA-MRSA strains, USA400 and USA300, as well as the colonization control strain, M92. With this mouse intradermal infection model, M92 and USA400-CMRSA7 caused localized infections, while USA300-C2406 caused ulceration with necrosis. The model was also able to show that the severe infections induced by USA300 were associated with a specific pattern of immune response, as well as an increased bacterial burden.
USA300 induced ulcer formation on day 1, reaching a maximum size and severity on day 4, remaining constant until day 7, and then healing by day 14-17. The USA300 infection was marked and lasted longer with more pronounced inflammation and a higher bacterial burden than the other MRSA infections, suggesting that USA300 might employ a mechanism to modify/inhibit the host's immune response. The role of neutrophils in MRSA infection is, however, complex. We showed that neutrophil depletion resulted in more severe tissue damage and invasive infections, possibly from loss of the protective role of neutrophils, resulting in uncontrolled bacterial infections. Neutrophils play an important role in the response to S. aureus SSTIs [41][42][43], however, extensive infiltration and activation of neutrophils at the site of infection is believed to cause chronic inflammation, impaired injury repair, and loss of organ function [44].
Transcriptional PCR array results further supported the notion that USA300 is capable of modulating the mouse immune system and neutrophil response. Il-18, which has been shown to restore neutrophil phagocytosis during severe skin damage [20], and Il-12, which is related to natural killer NK cell and macrophage defensive function in mice [45], each showed decreased transcriptional levels in the USA300-infected mice. USA300 also induced higher transcription of slpi, which was shown to be involved with the inhibition of neutrophil extracellular trap (NET) formation [46]. In addition, several factors involved with Toll-like and NOD-like receptor signaling showed decreased transcriptional levels in USA300-infected mice. Recognition of S. aureus by Toll-like and NOD-like receptors is required for the activation of protective host responses, such as the production of inflammatory cytokines and chemokines, and is important in the defense against S. aureusinduced SSTIs [47,48]. It was demonstrated that some staphylococcal virulence factors, such as staphylococcal TIR domain protein (TirS) that can inhibit TLR signaling via molecular mimicry [49], and staphylococcal superantigen-like protein 3 (SSL3) that can negatively interfere with TLR2 recognition and heterodimer formation [50,51], can inhibit and block TLR2-mediated inflammatory responses and impair the cytokine production, as well as neutrophil and macrophage activity, and consequently inhibit polymorphonuclear leukocyte (PMN) infiltration [52]. Taken together, these results indicate that USA300 might possess a mechanism to inhibit the defensive immune responses, especially involving neutrophils. However, the PCR array assayed the overall transcription of factors in whole skin samples, which includes various immune cells, keratinocytes, and other structural cells. This may not represent the transcriptional level for specific cell types and may explain why there was no difference in mpo transcription among the various strains and mock-infected control.
Cytokine and chemokine profiles in local lesions were consistent with our hypothesis that USA300 is modulating the mouse immune response. IL-17A, IL-1α, and IL-1β are believed to be protective during mouse MRSA skin infections [30,31]. The cytokines not only play a role in neutrophil recruitment, but can also promote the production of antimicrobial peptides (such as β-defensins and cathelicidins) [53]. In this mouse skin infection model, USA300 induced the same level of production for these cytokines on day 4 when compared with the other strains, yet more neutrophils were recruited, and there was a lower level of antimicrobial peptide production in USA300-infected mice. USA300 appears to be suppressing certain factors involved with the protective response. However, infection with USA300 did trigger increased production of cytokines and chemokines related to severity, including IFNy, GM-CSF, TNFa, IL-4, and IL-6. These are believed to be related to severe infections [35,36]. Other cytokines and chemokines such as IL-5 [54], LIF [55], LIX, MCP-1 [56], MIP-1a, and MIP-1b [57], which are believed to be involved in tissue damage, infection, and allergic diseases, also showed increased production in USA300 infections. Interestingly, the expression of MIG (also related to tissue damage, infection, or allergic diseases), was the only member of this group whose expression in USA300 infections was less than that in infections with the other MRSA strains. This phenomenon deserves further study. Macrophages expressing MIG (also known as CXCL9) are believed to protect against abscess formation [58] and promote tissue repair. Our USA300 resulted in lower production of MIG on day 7 and 14 when compared with other strains, which could indicate that USA300 is inhibiting MIG to increase tissue damage. We also recognize that the level of infiltration was very different between the different strains, as shown in this study. This could complicate the transcriptomic analysis as the composition of the material used for analysis was different with more or less immune cells. For example, a lack of a statistically significant difference in IL-17a, IL1a, and IL-1b should be viewed in light of the fact that there was more immune cell infiltration in the lesions for USA300 than the other strains. Further study by using tissue type-specific control genes is needed to address this issue.

Conclusions
Our results demonstrated that USA300 infections induce an intense neutrophilic response, inhibits host antibacterial responses, selectively inhibits the production of protective cytokines, and selectively activates the production of pro-inflammatory cytokines and chemokines. This unique pattern of immune responses could cause increased tissue damage at the site of the wound and increased bacterial persistence, which would facilitate the spread of USA300. This may provide some insight into why USA300 is more dominant and more virulent than other MRSA strains.