In vitro Multi-Species Biofilms of Methicillin-Resistant Staphylococcus aureus and Pseudomonas aeruginosa and Their Host Interaction during In vivo Colonization of an Otitis Media Rat Model

Staphylococcus aureus (SA) and Pseudomonas aeruginosa (PA) are known to cause biofilm-related infections. MRSA and PA have been frequently isolated from chronically infected wounds, cystic fibrosis, chronic suppurative otitis media (CSOM), and from indwelling medical devices, and these bacteria co-exist; however, their interaction with each-other or with the host is not well known. In this study, we investigated MRSA and PA multi-species biofilm communities in vitro and their interaction with the host during in vivo colonization using an OM rat-model. In-vitro biofilm formation and in-vivo colonization were studied using CV-microtiter plate assay and OM rat-model respectively. The biofilms were viewed under scanning electron microscope and bacteria were enumerated using cfu counts. The differential gene expressions of rat mucosa colonized with single or multi-species of MRSA or PA were studied using RNA-sequencing of total transcriptome. In multi-species in-vitro biofilms PA partially inhibited SA growth. However, no significant inhibition of MRSA was detected during in-vivo colonization of multi-species in rat bullae. A total of 1,797 genes were significantly (p < 0.05) differentially expressed in MRSA or PA or MRSA + PA colonized rat middle ear mucosa with respect to the control. The poly-microbial colonization of MRSA and PA induced the differential expression of a significant number of genes that are involved in immune response, inflammation, signaling, development, and defense; these were not expressed with single species colonization by either MRSA or PA. Genes involved in defense, immune response, inflammatory response, and developmental process were exclusively up-regulated, and genes that are involved in nervous system signaling, development and transmission, regulation of cell growth and development, anatomical and system development, and cell differentiation were down-regulated after multi-species inoculation. These results indicate that poly-microbial colonization induces a host response that is different from that induced by single species infection.

Staphylococcus aureus (SA) and Pseudomonas aeruginosa (PA) are known to cause biofilm-related infections. MRSA and PA have been frequently isolated from chronically infected wounds, cystic fibrosis, chronic suppurative otitis media (CSOM), and from indwelling medical devices, and these bacteria co-exist; however, their interaction with each-other or with the host is not well known. In this study, we investigated MRSA and PA multi-species biofilm communities in vitro and their interaction with the host during in vivo colonization using an OM rat-model. In-vitro biofilm formation and in-vivo colonization were studied using CV-microtiter plate assay and OM rat-model respectively. The biofilms were viewed under scanning electron microscope and bacteria were enumerated using cfu counts. The differential gene expressions of rat mucosa colonized with single or multi-species of MRSA or PA were studied using RNA-sequencing of total transcriptome. In multi-species in-vitro biofilms PA partially inhibited SA growth. However, no significant inhibition of MRSA was detected during in-vivo colonization of multi-species in rat bullae. A total of 1,797 genes were significantly (p < 0.05) differentially expressed in MRSA or PA or MRSA + PA colonized rat middle ear mucosa with respect to the control. The poly-microbial colonization of MRSA and PA induced the differential expression of a significant number of genes that are involved in immune response, inflammation, signaling, development, and defense; these were not expressed with single species colonization by either MRSA or PA. Genes involved in defense, immune response, inflammatory response, and developmental process were exclusively up-regulated, and genes that are involved in nervous system signaling, development and transmission, regulation of cell growth and development, anatomical and system development, and cell differentiation were down-regulated after multi-species inoculation. These results indicate that poly-microbial colonization induces a host response that is different from that induced by single species infection.

INTRODUCTION
Globally, Staphylococcus aureus and Pseudomonas aeruginosa (PA) are two major opportunistic pathogens that cause community-acquired and nosocomial infections. S. aureus and PA are the most prevalent pathogens that colonize structurally abnormal airways such as those in cystic fibrosis (CF) and other chronic obstructive lung diseases (Lyczak et al., 2002;Hubert et al., 2013). In addition, they are frequently found together in chronic wound infections (Gjødsbøl et al., 2006;Fazli et al., 2009). S. aureus and PA cause biofilm-related infections, and methicillin-resistant S. aureus (MRSA) has emerged as a clinically relevant pathogen because of its resistance to antibiotics and its ability to form biofilms (Chopra et al., 2015). Bacteria within biofilms are difficult to eradicate because being encased in a polymer matrix decreases their susceptibility to antimicrobials and immune defenses; this inherent antimicrobial resistance provides added resistance to antimicrobial therapy and host defense. In addition, during infection, the bacteria that originate in biofilms disperse as planktonic cells, which results in spread to secondary sites and progression of the infection (Hall-Stoodley and Stoodley, 2009;Lister and Horswill, 2014). MRSA and PA have been detected in biofilm-related infections such as chronic suppurative otitis media (CSOM) and chronic middle ear infections (Jung et al., 2009;Kim et al., 2015). S. aureus and PA have been isolated from upper respiratory tract infections including several chronic diseases such as chronic otitis media, cholesteatoma, chronic adenoiditis, chronic sinusitis, post-operative trampansomay, and nasal polyposis (Post et al., 2004;Bendouah et al., 2006;Boase et al., 2013). In chronic rhinosinusitis (CRS) patients, multi-species biofilms have been associated with enhanced mucosal inflammation, more severe osteitis, higher incidence of recurrent infection (Li et al., 2011;Dong et al., 2014), and postoperative outcomes (Singhal et al., 2011), and post-surgery progression (Bendouah et al., 2006). Furthermore, S. aureus and PA have been isolated from multi-species biofilms that are frequently found on indwelling medical devices such as prostheses, stents, implants, catheters, and endotracheal tube (Percival et al., 2015). Although the effect of poly-microbial infections of S. aureus and PA has not been well studied, some studies suggested that such poly-microbial infections are more virulent than single species infections (Hendricks et al., 2001;Pastar et al., 2013). Voggu et al. (2006) reported that in vitro interactions between S. aureus and PA are competitive and result in PA eradicating S. aureus (Voggu et al., 2006). However, it has been reported that during polymicrobial colonization of PA and S. aureus, PA does not completely inhibit the colonization of S. aureus; rather, S. aureus employs numerous defense strategies for its survival in the same ecological niche and grows as a small-colony variant (SCV) (Biswas et al., 2009). As a result of the competitive interactions between S. aureus and PA, an altered colony morphology strains called small colony variants (SCVs) emerges. Those SCVs are more persistent and more antibiotic-resistant strains than normal S. aureus (Nair et al., 2014).
Multi-species biofilm infections have important implications for management because this association will modify the clinical course of the disease, affecting the selection of antimicrobial therapy and the anticipated response to treatment. However, despite the gravity of these multi-species infections, their pathogenesis remains largely unknown. Therefore, in this study, we investigated S. aureus and PA multi-species biofilm communities in vitro, and assessed their interaction with the host during in vivo colonization using an otitis media rat model.

Ethics Statement
All animal experiments were performed in accordance with the guidelines of the Animal Research Committee, Korea University College of Medicine, Seoul, South Korea. The animal experiment protocol was approved by the Institute Review Board of Korea University, Guro Hospital, Seoul, South Korea (Permit Number, KOREA 2016-0019).

Bacterial Strain and Culture Medium
Methicillin resistant S. aureus (MRSA) was purchased from the Culture Collection of Antimicrobial Resistant Microbes (CCARM 3903), Seoul, Korea. P. aeruginosa (ATCC 27853) was obtained from the American Type Culture Collection (Manassas, VA). The selective medium for MRSA was oxacillin resistance screening agar base (ORSAB) and Pseudomonasspecific medium was Pseudomonas agar base (PAB), and was purchased from Oxoid limited (Hampshire, UK). 2-n-heptyl-4hydroxyquinoline-N-oxide (HQNO) was purchased from Santa Cruz Biotechnology (Dallas, Texas, USA).

In vitro Single Species or Multispecies Biofilm Growth of MRSA or PA
Single-or multi-species in vitro biofilm growth of MRSA and PA was established in 24-well (flat-bottom) polystyrene tissue culture plates (BD Falcon, Sparks, MD, USA) using a static model and a previously described procedure (Christensen et al., 1985;Yadav et al., 2015b). The biofilm biomass was quantified using a crystal violet (CV) microtiter plate assay, and the bacterial loads within biofilms were enumerated by colony forming unit (CFU) counts. MRSA or PA cell suspensions (1 × 10 7 ), individually or in combination in TSB media, were inoculated (1 mL) in 24-well polystyrene plates. The plates were incubated at 37 • C for 24 h. After incubation, medium was discarded, and plates were gently washed with 1 mL sterile water. Thereafter, plates were air-dried and stained with 200 µL CV (0.1%) for 15 min. Excess stain was decanted, and plates were washed three times with sterile distilled water. The biofilm was dissolved in 1 mL (95%) ethanol and the optical density (OD) at 570 nm was measured in an automatic spectrophotometer. All experiments were performed in triplicate and the average was calculated. The experiments were repeated three times.
Alternatively, MRSA, PA, or combinations of both species were grown in TSB medium under the same conditions. CFUs were counted to quantify the number of viable cells growing in the biofilms. Biofilms were dissolved with sonication at 50 W for 10 s, serially diluted, and plated on selective medium, specifically ORSAB or PAB with CN supplement, to determine CFU-values.
Biofilm Growth of SCVs of MRSA S. aureus forms SCVs in the presence of HQNO produced by PA (Mitchell et al., 2010). To evaluate biofilm formation capability of SCVs of MRSA (3903), biofilms were grown with (20 µg/mL) HQNO, whereas the control biofilm was grown with DMSO supplementation. Biofilm biomass was quantified using a CVmicrotiter plate assay as described above. The morphological variations in SCV biofilms grown with HQNO were examined using a scanning electron microscope (SEM).

In vivo Single Species or Multispecies Colonization of MRSA or PA in the Rat Middle Ear
In vivo single species or multispecies colonization by MRSA or PA was assessed using a previously described otitis media (OM) model (Yadav et al., 2012(Yadav et al., , 2014(Yadav et al., , 2015a. Forty pathogen-free Sprague-Dawley rats weighing 150-200 g were obtained from Koatech (Pyeongtaek, South Korea). Before the experiment, all animals were examined for abnormalities in the middle ear and were housed in an infection-free zone for 2 weeks. The rats were divided into four experimental groups. Group 1 included rats inoculated with MRSA only (n = 11); Group 2 was rats inoculated with PA only (n = 11); Group 3 included rats inoculated with a mixed culture of MRSA + PA (n = 11); and Group 4 was rats inoculated with media only (vehicle control; n = 7). The bacterial cell suspensions were prepared in TSB medium. Cell suspension (50 µL) containing 1 × 10 7 CFU of either single species MRSA, PA, or mixed culture of the two species was injected into the middle ear cavity through the tympanic membrane of the right ear using a tuberculin syringe and a 27-gauge needle. The animals were euthanized using a 1:1 combination of nitrous oxide and oxygen. One week after inoculation, the rats were sacrificed using carbon dioxide inhalation, and the bulla was aseptically acquired. The tympanic membrane was removed, and the ears were irrigated to remove planktonic bacteria. The bullae from each group were dissected and cleaned, and the middle ear was visualized and photographed. The bullae were homogenized, serially diluted, and plated on selective plates to obtain CFU counts. For SEM analysis, the bullae were preserved in SEM solution.

Visualization of In-vitro Biofilms and
In-vivo Colonization of MRSA or PA Single Species or Multi-Species Using SEM MRSA or PA single species or multispecies in vitro biofilms grown in 24-well tissue culture plates for 24 h were analyzed by SEM. After a 24-h incubation, the medium with planktonic cells were discarded, and the plates were gently washed twice with sterile PBS. Samples were pre-fixed by immersion in 2% glutaraldehyde and 2.5% paraformaldehyde solution, and postfixed for 2 h in 1% osmic acid dissolved in PBS. Samples were treated with a graded series of ethanol and t-butyl alcohol. The samples were freeze-dried in a freeze dryer (ES-2030 Hitachi, Tokyo, Japan), followed by platinum coating using an IB-5 ion coater (Eiko, Kanagawa, Japan), and specimens were visualized using a S-4700 field emission scanning electron microscope (Hitachi).
Similarly, for SEM analysis of in vivo colonization by MRSA or PA, the rat bullae from representative groups were dissected and cleaned of unwanted tissue, and the middle ear was carefully opened. The bullae were preserved in glutaraldehyde and paraformaldehyde solution. The rest of the protocol (prefix, post-fix, dehydration, etc.) was the same as that described previously for in vitro biofilms.
Total RNA Extraction and Transcriptome Sequencing Using Rat Mucosa Colonized With Single or Multi-Species of MRSA or PA Rats from representative groups were sacrificed as described above, and the bullae were acquired. Bullae were immediately dissected, and the middle ear was exposed and preserved in RNAlater (Qiagen, Hilden, Germany). The mucosal membrane of the bullae preserved in RNAlater solution was carefully scraped and preserved in fresh RNAlater solution until RNA extraction. Total RNA was isolated using a Qiagen RNeasy kit (Qiagen, Hilden, Germany) in accordance with the manufacturer's instructions. RNA was quantified using a nano-drop, and the RNA quality was assessed by analysis of rRNA band integrity using an Agilent RNA 6000 Nano kit (Agilent Technologies, Palo Alto, CA, USA). Ahead of cDNA library construction, 2 µg of total RNA and magnetic beads with Oligo (dT) were used to enrich poly(A)-containing mRNA. Subsequently, the purified mRNA was disrupted into short fragments, and double-stranded cDNA was immediately synthesized. The cDNA was subjected to end-repair and poly(A) addition, and was connected with sequencing adapters using the TruSeq RNA sample prep kit (Illumina, CA, USA). Suitable fragments that were automatically purified using a BluePippin 2% agarose gel cassette (Sage Science, MA, USA) were selected as templates for PCR amplification. The final library sizes and qualities were evaluated electrophoretically using an Agilent high sensitivity DNA kit (Agilent Technologies, CA, USA) and the fragment was determined to be between 350 and 450 bp. Subsequently, the library was sequenced using an Illumina HiSeq2500 sequencer (Illumina, CA, USA).
Low quality reads were filtered according to the following criteria; reads containing more than 10% of skipped bases (marked as "N"s), reads containing more than 40% of bases with quality scores <20, and reads with average quality scores <20. The entire filtering process was performed using the inhouse scripts. Filtered reads were mapped to the human reference genome (Ensembl release 72) (Flicek et al., 2013) using aligner STAR v.2.3.0e (Dobin et al., 2013).
Gene expression levels were measured with Cufflinks v2.1.1 (Trapnell et al., 2010) using the gene annotation database of Ensembl release 72. The non-coding gene region was removed with the mask option. To improve the accuracy of measurement, multi-read-correction and fragbias-correct options were applied, whereas all other options were set to default values. For differential expression analysis, gene level count data were generated using the HT Seq-count v0.5.4p3 (Anders et al., 2015) tool with the option "-m intersection-nonempty" and -r option, considering paired-end sequences.
Based on the calculated read count data, differentially expressed genes (DEG) were identified using the R package TCC (Sun et al., 2013). The TCC package applies robust normalization strategies to compare tag count data. Normalization factors were calculated using the iterative DEGES/edgeR method. Q-values were calculated based on the p-value using the p.adjust function of the R package with default parameter settings. DEGs were identified based on a q-value threshold <0.05. Gene ontology (GO) and KEGG pathway analysis of the DEGs were performed using STRING version 10.

Quantification of Gene Expression Using Real-Time RT-PCR
RNA-sequencing gene expression results were confirmed by real-time RT-PCR. Eighteen DEGs, and one house-keeping gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), were analyzed by real-time RT-PCR. The gene sequences were downloaded from the NCBI gene bank, and primers were designed using primer express software ( Table 1). cDNA synthesis was performed using a PrimeScript first strand cDNA synthesis kit according to the manufacturer's instructions (TaKaRa Bio Inc., Kusatsu, Japan). The amplification reaction was performed in an ABI Prism 7300 real-time PCR system (Applied Biosystems, Foster, CA, USA) in a total volume of 20 µL. The reaction mixture consisted of 10 µL of 2× SYBR  Green PCR Master Mix (Roche Applied Science, Indianapolis, IN, USA), 2.5 pmol of each forward and reverse primer, and 4 µL cDNA. The PCR reaction mixture was incubated as follows: initial denaturation at 95 • C for 10 min, followed by 45 cycles of DNA denaturation at 95 • C for 15 s, primer annealing and extension at 56 • C for 1 min, and a final extension step at 72 • C for 5 min. A control reaction without reverse transcriptase was included in each real-time RT-PCR experiment to check for cDNA contamination with genomic DNA. The relative quantification of gene expression in rat mucosa inoculated with S. aureus, PA, or mixed culture, with respect to the type of media, was performed using the 2 − CT method (Livak and Schmittgen, 2001), and the gene expression values were normalized to those of the housekeeping gene, GAPDH.

Statistical Analysis
Values were calculated as the mean of individual experiments that were performed in triplicate, and compared to those of the control groups. Differences between mean values were assessed using a Student's t-test. Statistical significance was set at p < 0.05.

In vitro Single Species or Multispecies Biofilm Growth of MRSA or PA
To assess the growth of MRSA and PA in single and multi-species biofilms, in vitro single and multi-species biofilms were grown on polystyrene plates that allowed for bacterial attachment and biofilm formation. The biofilms were quantified using a CVmicrotiter plate assay after removing planktonic cells. The CVmicrotiter plate assay detected significantly (p < 0.05) increased biofilm biomass in multi-species biofilms of MRSA and PA, compared to single species biofilms of either MRSA or PA ( Figure 1A). The OD 570 -values of single species biofilms of MRSA and PA were 0.7 and 1.8, respectively; however, that of the multi-species biofilm was 2.0. These results indicated that total biomass of multi-species biofilms were higher than that of single species biofilms, which might be because of the accumulation of extracellular polysaccharide (EPS) and cells from both species. Thus, both species contributed to increased biofilm biomass.
To further evaluate the viable bacteria in single or multispecies biofilms, CFUs were enumerated. The CFU counts showed significantly higher numbers of bacteria in single species biofilms than in multi-species biofilms of either MRSA or PA ( Figure 1B). In single species biofilms, the CFU counts for MRSA and PA were 1 × 10 8 and 4 × 10 9 , respectively. In multispecies biofilms, PA inhibited the growth of MRSA. The CFU counts of MRSA and PA in multi-species biofilms were 1 × 10 6 and 8 × 10 8 , respectively. The CFU counts for PA were significantly (p < 0.05, approximately 2.6 log values) higher than those of MRSA. These results indicated that PA and MRSA could exist together in multi-species in vitro biofilms. However, the decreased CFU counts for MRSA indicated that PA partially inhibited the growth of MRSA in multi-species biofilms.
The treatment of pre-established biofilm of MRSA with DNase I and Proteinase K significantly (p < 0.05) decreased biofilm biomass in compare to control biofilms ( Figure 1C). However, no effect of sodium metaperiodate was detected. These results indicate that MRSA biofilms contains e-DNA and proteins, and the MRSA produce ica-independent biofilms. Previous studies reported no effect of sodium metaperiodate on MRSA biofilms (Fitzpatrick et al., 2005). Treatment of PA biofilms with DNase I and proteinase K significantly (p < 0.05) reduced biofilms biomass. However, very low biofilm eradication activity of alginate lyase was detected on pre-established biofilms of PA ( Figure 1D). PA is known to produces at least three types of exopolysaccharides. The mucoid PA strains predominantly produce alginate exopolysaccharide (Hentzer et al., 2001), and non-mucoid strains produce the Pel and Psl polysaccharides for biofilm formation (Wozniak et al., 2003). ATCC 27853 is a standard non-mucoid strain (Li et al., 2001). These results indicate that PA (ATCC 27853) biofilms contains DNA and proteins along with polysaccharides primarily the Pel and Psl.
MRSA or PA single-species or multi-species in vitro biofilms were analyzed using SEM. SEM analysis revealed that singlespecies biofilms were compact, and cells were connected to the bottom of the plate and to each other (Figures 2A,B). The multi-species biofilms were also compact, and the bacteria were interconnected to each other and to the surface. In multi-species biofilms, the biofilm debris or may be EPS was visible that were not visible in single species biofilms of either MRSA or PA ( Figure 2C). As expected in multispecies biofilms, fewer S. aureus bacteria were visible than PA, indicating that PA partially inhibited MRSA growth in the biofilm state.

SCVs of MRSA Enhance Biofilm Formation
The in vitro biofilm formation ability of MRSA 3903 SCVs was evaluated by culturing biofilms in the presence or absence of HQNO. CV-microtiter plate assay results showed no significant decrease in MRSA biofilm biomass in the presence of 20 µg HQNO (Figure 3). SEM images revealed no significant difference between SCV biofilms and phenotypically normal MRSA biofilms (Figure 4). These results indicate that MRSA 3903 could form SCVs, and that MRSA could exist in multi-species biofilms with PA.

FIGURE 3 | In vitro biofilm growth of methicillin-resistant
Staphylococcus aureus (MRSA) small colony variant. In vitro biofilm growth of MRSA supplemented with HQNO (20 µg/mL) or without HQNO. The biofilm biomass was detected by a crystal violet microtiter plate assay. Error bars represent standard deviation from the mean value.
In vivo Colonization of MRSA and PA Alone or Together in the Rat Middle Ear Rats were sacrificed 1 week after inoculation, and bullae were acquired. The representative rat bullae were dissected and photographed (Figures 5A-D). The rat bullae inoculated with media only (vehicle control) were clear with no sign of mucosal inflammation or bacterial colonization ( Figure 5A). However, the rat bullae inoculated with MRSA or PA alone or together were filled with a sticky fluid and showed signs of bacterial colonization. The mucosa was thick, swollen, and showed signs of severe inflammation (Figures 5B-D). The mean CFU counts for the rat bullae inoculated with MRSA and PA alone were 2.63 × 10 5 (SD = 1.07 × 10 5 ) and 4.50 × 10 5 (SD = 7.07×10 4 ), respectively. A higher number of PA CFUs, compared to MRSA CFUs, was recovered from rat bullae inoculated with either MRSA or PA alone; however, this difference was not significant. The CFU counts for MRSA and PA in the rat bullae inoculated with both species were 1.34 × 10 5 (SD = 1.1 × 10 5 ) and 1.10 × 10 5 (SD = 7.3 × 10 4 ), respectively ( Figure 5E). No significant difference in CFU counts between MRSA and PA were detected in multi-species-inoculated rat bullae. However, the CFU counts for PA alone were significantly high than those of PA in multi-species inoculations.  control middle ear, the ciliated epithelium in the hypotympanum area and Eustachian tube orifice were intact (Figures 6A-C).

Visualization of In vivo Colonization of MRSA and PA Alone or in Combination
In the MRSA, PA, or MRSA + PA treated groups, the entire middle ear was covered with cell debris or biofilm EPS. The tips of the cilia were invisible due to the deposition of cell debris (Figures 6D-L).
Transcriptome Sequencing of Rat Mucosa Inoculated with MRSA, PA, or a Combination of the Two Total raw data reads in all samples ranged from 50 to 65 million, with an average of approximately 55 million raw reads per sample. The average mRNA insert size detected by sequencing and mapping was 159 bp. The uniquely mapped reads ranged between ∼50 and 55 million; there were ∼5 million mapped reads and ∼6 million unmapped reads ( Figure 7A). There were ∼58 million total reads, 55 million genome reads, and 43 million gene reads (Figure 7B). The gene reads were ∼20 K, known genes were ∼5 K, known plus new isoforms were ∼12 K, and novel genes were ∼5 K (Figure 7C). A total of 1797 genes were significantly (p < 0.05) differentially expressed in all three treatments with respect to the control (media). Among them, 973 were upregulated > 1-fold and 824 were downregulated > 1-fold.

Differential Gene Expression
A total of 360, 634, and 504 genes were significantly (p < 0.05) upregulated in multi-species, MRSA and PA, colonized rat middle ear mucosae ( Figure 8A). Among 360 upregulated genes in multi-species-colonized rat mucosae, 113 (31%) were similar to those of MRSA, and 22 (6.1%) were similar to those of PA. The three treatments shared 163 genes whose expression was upregulated. Moreover, 62 genes were exclusively upregulated during multi-species colonization. Among them, the 42 genes upregulated by ≥1.5-fold and encoding proteins of known functions are shown in Table 2. A total of 580, 401, and 481 genes were significantly (p < 0.05) downregulated in multi-species-, MRSA-, and PA-colonized rat middle ear mucosae ( Figure 8B). Among the 580 downregulated genes after multi-species colonization, 75 genes (12.9%) were similar to those associated with MRSA, and 168 (28.9%) genes were similar to those associated with PA. Among 160 genes exclusively downregulated after multi-species colonization, 88 genes that were downregulated by ≥1.5-fold and encoding known proteins are shown in Table 3. The gene ontology (GO) of upregulated and downregulated genes in multi-species inoculated rat mucosae is shown in Figures 9A,B. The GO analysis of 42 protein-encoding genes that were upregulated (≥1.5-fold) after multi-species colonization revealed that genes involved in defense, immune response, inflammatory response, and developmental process were exclusively upregulated after multi-species inoculation ( Figure 9A). The GO analysis of 88 protein-encoding genes that were downregulated (≥1.5fold) revealed that genes that are involved in nervous system signaling, development and transmission, regulation of cell growth and development, anatomical and system development, and cell differentiation were significantly downregulated after multi-species inoculation ( Figure 9B). Based on KEGG pathway analysis of upregulated genes after multi-species inoculation, genes involved in phagosome, antigen processing and presentation, influenza A infection, Staphylococcus infection, herpes virus infection, toxoplasmosis, and tuberculosis (and other) pathway genes were evaluated (Figure 10). These results revealed that colonization of the rat middle ear with MRSA or PA alone or in combination regulates a significant number of genes. In addition, multi-species colonization with MRSA and PA induced a different set of genes that were not induced by either MRSA or PA alone. This shows that the host response is different after multi-species colonization, which could enhance the pathogenicity of the bacteria.

Quantification of Gene Expression Using Real-Time RT-PCR
The results of real time-PCR were in complete agreement with RNA sequencing gene expression results. The expression of RPS9, MMP12, ITGB2, CST7, SCTR, APOC1, CD69, and CCNO genes was significantly (p < 0.05) upregulated (by >1.5-fold) in rat mucosae colonized with MRSA or PA alone or in combination based on both real time RT-PCR and RNA sequencing (Table 4). Similarly, the expression of KERA, GPD1, ACTA1, EEF1A2, HRC, FOXD1, MYH1, SYNPO2L, TNMD, and MYOM2, based on real time RT-PCR and RNA sequencing, was significantly (p < 0.05) decreased by more than 1.5-fold (Table 4).

DISCUSSION
S. aureus and PA are known to cause biofilm-related infections and are frequently found together in chronically infected wounds, CF, CSOM, and on indwelling medical devices such as prostheses, stents, implants, catheters, and endotracheal tubes (Chole and Faddis, 2002;Hubert et al., 2013;Percival et al., 2015). Previous studies using a wound infection model suggested that poly-microbial infections of MRSA and PA were more virulent than single species infections and that synergism exists between these species, which increases virulence (Hendricks et al., 2001;Pastar et al., 2013). In various types of chronic ear infections, MRSA and PA have been frequently isolated, and these bacteria co-exist; however, their interaction with each-other or with the host was previously unknown (Klein and Chan, 2010;Kim et al., 2015). In this study, we investigated MRSA and PA multi-species biofilm communities in vitro and their interaction with the host during in vivo colonization using an OM rat model.
CV microtiter plate assay results showed significantly increased biofilm biomass in multi-species biofilms of MRSA and PA in comparison with that of single species biofilms of either MRSA or PA. It was speculated that the matrix produced by the two bacterial species and cellular components of the bacteria contributes to increased biomass. It is suggested that Staphylococcus and Pseudomonas biofilm matrix consist of extracellular DNA (e-DNA) along with exopolysaccharides (Whitchurch et al., 2002;Eckhart et al., 2007). Staphylococcal biofilm accumulation is mediated by polysaccharide intercellular adhesion (PIA), modulated by gene products encoded by ica operon (Cramton et al., 1999), and can also form icaindependent biofilm, mediated by proteins binding of the  (Fitzpatrick et al., 2005). While the PA biofilms matrix consists of alginate (a negatively charged polymer of guluronic and mannuronic acid), Psl (a neutral polysaccharide consisting of a pentasaccharide repeat containing glucose, mannose, and rhamnose), and Pel (positively charged exopolysaccharide composed of partially acetylated 1→4 glycosidic linkages of N-acetylgalactosamine and N-acetylglucosamine) (Franklin et al., 2011). However, MRSA CFU counts were significantly lower (by >2-fold) than those of PA. These results are in agreement with previous studies that demonstrated that S. aureus and PA are frequently found together in human infections, and that PA kills S. aureus when the two species are grown together in nutrient rich medium in vitro (Palmer et al., 2005(Palmer et al., , 2007Voggu et al., 2006). The killing of S. aureus has been attributed to various exoproducts of PA such as LasA protease (Mansito et al., 1987), HQNO (Hoffman et al., 2006), pel and psl products (Qin et al., 2009), and pyocyanin (Dietrich et al., 2006). SEM images also confirmed the presence of few MRSA bacteria in multi-species biofilms; however, increased EPS matrix was visible in multi-species biofilms, compared to that in single species biofilms. Many studies have suggested that S. aureus can overcome this suppression and escape PA-mediated cell death by growing as SCVs, which are characterized as a respiration-defective subpopulation with an altered phenotype (Biswas et al., 2009;Nair et al., 2014). Moreover, S. aureus forms SCVs in the presence of HQNO produced by PA (Hoffman et al., 2006). In addition, these SCVs that were shown to form under stress of HQNO formed robust in vitro biofilms through the activation of SigB and the repression of agr (Mitchell et al., 2010). Our results also showed robust biofilm formation by MRSA 3903 in the presence of HQNO, when compared to that in the control. SEM results provided further evidence of significant biofilm formation in the presence of HQNO. This process has been suggested to be due to the activation of SigB and the repression of agr, which upregulates various cell surface proteins such as FnBA, which are involved in biofilm formation (Mitchell et al., 2008(Mitchell et al., , 2009(Mitchell et al., , 2010. S. aureus and PA frequently colonize the upper respiratory tract and cause several chronic diseases including chronic OM, cholesteatoma, chronic adenoiditis, chronic sinusitis, postoperative trampansomay, and nasal polyposis (Bendouah et al., 2006;Boase et al., 2013). Furthermore, MRSA and PA have been frequently isolated from chronic ear infections such as CSOM (Jung et al., 2009;Kim et al., 2015). In this study, we used the OM rat model to study the interaction between MRSA and PA in vivo. We previously used this model to study the colonization of S. aureus and Streptococcus pneumoniae separately in the rat middle ear (Yadav et al., 2012(Yadav et al., , 2014(Yadav et al., , 2015a. Herein, we first evaluated the utility of the OM rat model for multi-species colonization. The results showed consistent recovery of MRSA or PA from the rat middle ear, which was associated with significant inflammation. We choose the OM rat model over other animal models because the rat middle ear is a natural habitat for S. aureus and PA, and because inflammation in the middle ear mucosa (due to bacterial challenge or infection) results in severe swelling that can be easily visualized by SEM; this tissue can also be easily recovered for gene expression studies.
The middle ear is a natural habitat for many bacteria including normal body flora or opportunistic pathogens. Both S. aureus and PA have been isolated from middle ear infections (Brook, 2003). Our in vivo results showed marked recovery of MRSA  or PA after multi-species and single species inoculation with significant inflammation in the rat middle ear. In addition, no significant differences in CFU counts between MRSA and PA were observed after multi-species inoculation. These results indicate that MRSA and PA successfully colonized the rat middle ear with no significant inhibition of either species. Using wound infection and rabbit models, DeLeon et al. (2014) reported no inhibition of S. aureus in the presence of PA during multispecies in vivo colonization (DeLeon et al., 2014). Other studies also showed a lack of inhibition in the absence of nutrient rich medium, which could be due to the presence of SCVs of S. aureus (McNamara and Proctor, 2000;Filkins et al., 2015). Inflammation in the middle ear is a sign of infection; our results also detected significant inflammation in this tissue after inoculation with either MRSA or PA alone or in combination. The significant number of bacteria recovered (based on CFU counts) indicates that multi-species bacterial colonization was established in the middle ear. SEM analysis also confirmed the presence of biofilmlike structures that were deposited on the cilia of the middle ear after single or multi-species inoculation. However, those of MRSA or PA alone were indistinguishable. Similarly, Kaya et al. (2013) also detected poly-microbial biofilms of S. aureus and PA in the middle ear (Saunders et al., 2011;Kaya et al., 2013). In CRS patients, multi-species biofilms have been associated with increased mucosal inflammation, more severe osteitis, higher incidence of recurrent infection (Li et al., 2011;Dong et al., 2014) and postoperative outcome (Singhal et al., 2011), and post-surgery progressions (Bendouah et al., 2006). To analyze differential gene expression in rat middle ear mucosae inoculated with MRSA or PA alone or in combination, RNA sequencing of the total transcriptome was performed. RNAseq results detected significantly increased expression of defense-, immune-, response-, and inflammatory-related genes after multi-species colonization and decreased expression of signaling, development, and communication pathway genes. Interestingly, a total of 122 genes (62 upregulated and 160 downregulated) were exclusively differentially regulated by multi-species colonization, and these genes were not expressed with either MRSA or PA inoculation. Gene pathway analysis revealed that those genes are involved in the phagosome, antigen processing and presentation, influenza A infection, Staphylococcus infection, herpes virus infection, toxoplasmosis, and tuberculosis among others.
FCN2, CFI, KLRD1, KLRB1F, UBD, CCL5, and GAL were significantly upregulated after multi-species inoculation and were either downregulated or insignificantly upregulated with MRSA or PA single inoculation. The ficolin-2-encoding gene FCN2, was upregulated by 1.52-fold (P = 0.04) with multispecies infection and was downregulated after inoculation with MRSA or PA alone. Ficolin activates the lectin pathway of complement by specifically binding to lipoteichoic acid, a cell wall constituent of gram-positive bacteria, and initiates an innate anti-microbial immune response (Lynch et al., 2004). The CFI gene was significantly upregulated in multi-species infections and was downregulated with MRSA or PA alone. CFI encodes complement factor I, which is involved in complement activation during cochlear responses to acoustic trauma . The KLRD1 and KLRB1F genes encode the killer cell lectin-like receptor subfamily K, member 1 protein and killer cell lectin like receptor B1, which are involved in the activation of immune responses and cytotoxicity, were upregulated by 2.1-and 2.0-fold, respectively, after mixed culture inoculation, and were downregulated with MRSA or PA alone (Taniguchi et al., 2015). The UBD gene, encoding the protein ubiquitin D, which is involved in cytokine-induced apoptosis in rat, was upregulated by mixed cultures and MRSA and was downregulated with PA alone (Brozzi et al., 2016).   Similarly, the CCL5 gene encodes chemokine (C-C Motif) ligand 5 protein. These chemokine proteins are involved in immunoregulatory and inflammatory processes. CCL5 is an essential factor for the induction and maintenance of protective pneumococcal immunity (Palaniappan et al., 2006). The GAL gene encodes galanin peptides, and was upregulated by 1.9-fold or repressed with SA or PA, respectively. The galanin protein is involved in post-traumatic stress disorder and mild blast-induced traumatic brain injury (Kawa et al., 2016). The upregulation of important immune response-and cytotoxicity-related genes after multispecies inoculation indicates an enhanced host response, when compared to that after inoculation with MRSA and PA alone. DLECI, ST8SIA2, and SSC5d are important genes, which were downregulated with multi-species inoculation and either upregulated or insignificantly downregulated with MRSA or PA alone. The DLEC1 gene encodes the deleted in lung and esophageal cancer 1 protein (DLECI) that has tumor suppressive activities. The downregulation of this gene has been observed in several human cancers including lung, esophageal, renal, and FIGURE 9 | Gene ontology (GO) of differentially expressed genes in rat middle ear mucosa colonized with methicillin-resistant Staphylococcus aureus (MRSA) or Pseudomonas aeruginosa (PA) as a single or multi-species inoculum. (A) GO category of significantly (p < 0.05) upregulated genes (by ≥1.5-fold) with multi-species colonization. (B) GO category of significantly (p < 0.05) downregulated genes (≥1.5-fold) after multi-species colonization.
head and neck squamous cell carcinoma (Zhang et al., 2015). Similarly, the ST8SIA2 gene encodes alpha-2,8-sialyltransferase 2, which is involved in neuronal plasticity (Shaw et al., 2014). The SSC5D gene encodes scavenger receptor cysteine rich family member with 5 domains protein, which induces bacterial and fungal aggregation and subsequent inhibition of PAMPinduced cytokine release, and was also downregulated after multi-species inoculation. These results indicate that multispecies colonization triggers host responses that are different from those induced by infection with MRSA or PA alone. It has been previously reported that poly-microbial biofilms of S. aureus and PA, in a chronic wound infection model, significantly impair wound healing relative to that observed with their singlespecies biofilm counterparts. Multi-species biofilms also trigger enhanced host inflammatory responses through the expression of IL-1β and TNF-α (Seth et al., 2012). Similarly, Pastar et al. (2013) also detected significantly decreased re-epithelization with mixed species biofilms, via the suppression of keratinocyte growth factor-1. Moreover, in poly-microbial wound infections, the presence of PA induced the expression of S. aureus virulence factors including Panton-Valentine leucocidin and α-hemolysin (Pastar et al., 2013).

CONCLUSION
This study demonstrated that MRSA and PA could coexist in poly-microbial biofilms in vitro, and colonize the rat middle ear in vivo. The poly-microbial colonization of MRSA and PA induced the differential expression of a significant number of genes that are involved in immune response, inflammation, signaling, development, and defense; these were not expressed with single species colonization by either MSA or PA. These results indicate that poly-microbial colonization induces a host response that is different from FIGURE 10 | KEGG pathway analysis using string showing significantly (p < 0.05) upregulated genes (by ≥1.5-fold) after multi-species colonization of the rat middle ear mucosa. that induced by single species infection. It is probable that poly-microbial infections were more virulent than single species infections. Thus, poly-microbial infections could modify the clinical course of disease, affecting the selection of antimicrobial therapy and the anticipated response to treatment.

AUTHOR CONTRIBUTIONS
MY, designed research, did experiment, results analysis wrote manuscript. SC, evaluated results, chemical supply, facility arrangement, check manuscript. YG, protocol approval, animal work design, arrange reagent, and chemical facility. GI, protocol design, manuscript reading, manuscript review. JS, Protocol design, results analysis, manuscript review.

FUNDING
This study was supported by Korean Health Technology Research and Development Project, Ministry of Health and Welfare (HI15C2835), South Korea.