D-LL-31 enhances biofilm-eradicating effect of currently used antibiotics for chronic rhinosinusitis and its immunomodulatory activity on human lung epithelial cells

Chronic rhinosinusitis (CRS) is a chronic disease that involves long-term inflammation of the nasal cavity and paranasal sinuses. Bacterial biofilms present on the sinus mucosa of certain patients reportedly exhibit resistance against traditional antibiotics, as evidenced by relapse, resulting in severe disease. The aim of this study was to determine the killing activity of human cathelicidin antimicrobial peptides (LL-37, LL-31) and their D-enantiomers (D-LL-37, D-LL-31), alone and in combination with conventional antibiotics (amoxicillin; AMX and tobramycin; TOB), against bacteria grown as biofilm, and to investigate the biological activities of the peptides on human lung epithelial cells. D-LL-31 was the most effective peptide against bacteria under biofilm-stimulating conditions based on IC50 values. The synergistic effect of D-LL-31 with AMX and TOB decreased the IC50 values of antibiotics by 16-fold and could eliminate the biofilm matrix in all tested bacterial strains. D-LL-31 did not cause cytotoxic effects in A549 cells at 25 μM after 24 h of incubation. Moreover, a cytokine array indicated that there was no significant induction of the cytokines involving in immunopathogenesis of CRS in the presence of D-LL-31. However, a tissue-remodeling-associated protein was observed that may prevent the progression of nasal polyposis in CRS patients. Therefore, a combination of D-LL-31 with AMX or TOB may improve the efficacy of currently used antibiotics to kill biofilm-embedded bacteria and eliminate the biofilm matrix. This combination might be clinically applicable for treatment of patients with biofilm-associated CRS.


Introduction
Chronic rhinosinusitis (CRS) is defined as chronic infection and inflammation of the upper airways involving the nasal cavity and paranasal sinuses with a duration of at least 12 In this study, the bactericidal effects of LL-37, LL-31 and their D-enantiomeric forms (D-LL-37 and D-LL31) against bacteria isolated from CRS patients grown as biofilm were investigated alone and in combination with conventional antibiotics. We found that D-LL-31 has a strong killing effect against all tested bacteria under biofilm-stimulating conditions. It also enhanced antimicrobial activity of the commonly used antibiotics to combat biofilmembedded cells and to eradicate the biofilm matrix. Additionally, D-LL-31 does not cause the induction of the cytokines involved in immunopathogenesis of CRS on human lung epithelial cells.

Peptide synthesis
The human cathelicidin peptides LL-37, LL-31 and D-enantiomers (D-LL-37 and D-LL-31) ( Table 1) were synthesized by solid-phase peptide synthesis using fluoren-9-ylmethoxycarbonyl (Fmoc) chemistry with a Siro II synthesizer (Biotage, Uppsala Sweden) according to the manufacturer's protocol. Labeling of D-LL-31 with 5,6-carboxytetramethylrhodamine (TAMRA) was carried out in-synthesis using an additional Fmoc-Ahx-OH (NovaBiochem) at the N-terminus. Peptides were purified to at least 95% purity by preparative reversed-phase HPLC on a Dionex Ultimate 3000 system (Thermo Scientific, Breda, the Netherlands). The authenticity was confirmed by mass spectrometry with a Microflex LRF MALDI-TOF (Bruker Daltonik GmbH, Bremen, Germany) essentially as described previously [29,30]. Molar concentrations were calculated based on their weight.

Ethics statement
All patients provided signed consent in accordance with the ethical principles relating to biomedical research involving human subjects as adopted by the 18 th World Medical Association General Assembly, Helsinki (1964) and the ICH Good Clinical Practice Guidelines. This study was approved by the Center for Ethics in Human Research, Khon Kaen University, reference number: HE561075.

Sample collection and bacterial identification
Washings from sinus irrigation and sinus mucosa samples were collected from patients with chronic rhinosinusitis, defined according to the criteria of the 2003 Chronic Rhinosinusitis Task Force [31]. These patients were potential candidates for endoscopic sinus surgery in the otolaryngology clinic, Srinagarind Hospital, Khon Kaen University, Thailand.

Biofilm observation on sinus mucosal tissues of CRS patients
Biofilm architecture on sinus mucosal tissues of CRS patients was observed using scanning electron microscopy (SEM). The tissue samples were taken from diseased maxillary or ethmoid sinuses during functional endoscopic sinus surgery. Tissues were fixed by 4% paraformaldehyde/1.25% glutaraldehyde in PBS with 4% sucrose, pH 7.2, and stored at 4˚C. After three PBS washes, samples were post-fixed with 2% osmium tetroxide (OsO 4 ) on a rotator for 1 h. The dehydration of the tissue samples was subsequently performed using a graded series of 70% to 100% ethanol concentrations. Samples were critical-point dried in CO 2 , mounted on scanning electron microscopy stubs, sputter coated with gold palladium (K500X sputter coater, EMITECH, Quorum technologies LTD, UK), and examined using a field emission gun scanning electron microscope (S-3000N, HITACHI, Japan).

Bacterial strains and growth conditions
Three clinical bacterial isolates, Klebsiella pneumoniae, Pseudomonas aeruginosa and Staphylococcus epidermidis, were selected from patients no. 1, 3 and 5, respectively (Table 2) according

Antimicrobial activity of antibiotics and peptides against planktonic bacteria
Each bacterial suspension (5 x 10 5 CFU/ml) in 1 mM potassium phosphate buffer (PPB) (pH 7.0) was added to an equal volume of each antibiotic (AMX and TOB) and AMP (LL-37, LL-31, D-LL-37 and D-LL-31) to reach final concentrations of 1, 5 and 10 μM and incubated at 37˚C for 1 and 2 h. A bacterial suspension without peptide served as a control. The samples were then serially diluted and cultivated on NA (Himedia 1 ) and incubated for 24 h to allow colony counting. The percentage killing by each AMP was calculated using the formula [1 − (CFU sample/CFU control)] × 100% [15,36]. Each experiment was performed independently three times in duplicate.

Determination of the biofilm-forming capacity
The biofilm-forming capacity of each strain was examined using the biofilm crystal violet staining assay (on 2-day biofilm) as previously described [37]. Two hundred microlitres of bacterial suspension (10 8 CFU/ml) were added into a sterile 96-well flat-bottom plate (Nunclon™) and the plate was incubated at 37˚C for 3 h. Cell-free medium was used as a control. Weakly adhered planktonic cells were removed and fresh medium was added, then the plate was incubated at 37˚C for 21 h. The biofilms were washed with sterile distilled water, fresh medium was added and the plate was incubated at 37˚C for 24 h. The adherent biofilms were then fixed with 99% (v/v) methanol, air dried and stained with 2% (w/v) crystal violet for 5 min. Subsequently, 2-day biofilms were solubilized with 33% (v/v) glacial acetic acid. The absorbance was spectrophotometrically measured at the wavelength of 620 nm using a microplate reader (Sun-rise™). Each experiment was performed independently three times in 8 replicates.

Comparison of the antimicrobial activity of peptides and antibiotics against bacteria grown as biofilm
The effects of both conventional antibiotics and AMPs against bacteria under biofilm-stimulating condition were determined using a transferable solid phase (TSP) pin lid, which resembles the 'Calgary' biofilm device as previously described with some modifications [38]. Each strain (10 7 CFU/ml) was added to sterile 96-well plates (Nunclon™), then covered by a TSP pin lid (NUNC™, Roskilde, Denmark), and further incubated at 37˚C for 24 h. Following the period of incubation, biofilms formed on the TSP pin surfaces were washed by sterilized distilled water and challenged with AMX in a range from 134 to 34418 μM (50-25000 μg/ml) and 0.05 to 100 μM of LL-37, LL-31, D-LL-37 and D-LL-31 in 1 mM PPB. Additionally, the effect of TOB (2-836 μM or 1-391 μg/ml) against P. aeruginosa biofilm was also investigated. Then the plates were statically incubated at 37˚C for 24 h. A well free of antimicrobial agent was used as a control. After 24 h, bacterial survival was determined using the plate count technique on NA (Himedia 1 ). The percentage killing of each antimicrobial agent was calculated using the formula [1 − (CFU sample/CFU control)] × 100% [15,19]. The concentration of each agent required to kill 50% of cells in each tested bacterial strain (IC 50 value) was determined by linear regression of plots. Each experiment was performed independently three times in duplicate.

The synergistic activity of D-LL-31 with antibiotics against bacteria grown as biofilm
After the IC 50 value of each antimicrobial agent was calculated, D-LL-31 (the most effective peptide) was then selected to determine its synergistic activity with conventional antibiotics (AMX and TOB) using the broth microdilution checkerboard technique with some modification [15,39]. Briefly, 24 h biofilms on TSP pin lids were mixed with the antibiotics (AMX or TOB) and D-LL-31 diluted in 1 mM PPB to final concentrations of IC 50 down to 1/16 of the IC 50 values of each agent and incubated at 37˚C for 24 h (Table 4). Bacterial viability in each interaction was determined using the plate count technique. The fractional inhibitory concentration index (FICI) values were calculated and interpreted in terms of synergism/antagonism as previously described [40]. Each experiment was performed independently three times in duplicate.

The effect of D-LL-31 alone and in combination with antibiotics on biofilm matrix
The 1 μg/ml)) were used. All tested conditions were incubated at 37˚C for 24 h. A well free of antimicrobial agent was used as a control. Following incubation, biofilms were fixed with 2.5% glutaraldehyde (Sigma-Aldrich) at 25˚C for 3 h. The fixed bacterial biofilms on glass coverslips were stained with 50 μg/ml Alexa Fluor 1 488 (Invitrogen, Carlsbad, CA) for 30 min. This stain reacts with exopolysaccharide matrixes of the biofilm. The glass coverslips were then mounted with 80% glycerol and the stained cells were photographed using a confocal laser-scanning microscope (CLSM) (TCS SP8-Leica Microsystems, Wetzlar, Germany). The biofilm biomass was used as a quality-control parameter to confirm the reduction of biofilm matrix using COMSTAT analysis (BioCentrum-DTU, Lungby, Denmark). Each experiment was performed independently three times in duplicate.

The localization of peptide on biofilm matrix
The localization of TAMRA-labeled D-LL-31 on biofilm matrix was investigated using the AAA-model as mentioned above. One-day biofilm was fixed with 2.5% glutaraldehyde (Sigma-Aldrich) at 25˚C for 3 h and stained with 50 μg/ml of Alexa Fluor 1 488 (Invitrogen) for 30 min. Subsequently, TAMRA-labeled D-LL-31 at the IC 50 concentration was added onto the biofilm architecture of each tested strain for 5 min, and then the biofilm was washed 3 times using sterile distilled water. The localization of D-LL-31 on biofilm matrix was visualized by CLSM (TCS SP8-Leica Microsystems). Each experiment was performed independently three times in duplicate.

Determination of the cytotoxicity of D-LL-31 for A549 cells
Human lung epithelial cells (A549, ATCC 1 , CCL-185™) were used to investigate the cytotoxicity of D-LL-31 as previously observed [42]. Briefly, A549 cells were maintained in RPMI-1640 (HyClone™, Marlborough, MA) with 10% fetal bovine serum (FBS) (HyClone™) and sub-cultured by standard methods using trypsin/EDTA (0.25% trypsin, 0.1% EDTA) (Corning™, Corning, NY). A standard curve of A549 cells (100-50000 cells/well) was constructed as previously described with some modification [43]. For MTT assay, cells were plated at a density of 2.5×10 4 cells/well in 96-well plates and grown overnight. D-LL-31 was diluted in PPB and added to culture medium to final concentrations ranging from 0.05 to 100 μM. Prepared A549 cells were then exposed to the diluted peptides for 24 h at 37˚C, 5% CO 2 atmosphere. Negative controls received either culture medium or PPB. Cells with 2% dimethyl sulfoxide (DMSO) (Sigma-Aldrich) were used as positive control. Cytotoxicity was assessed using MTT (0.5 mg/ ml) according to the manufacturer's instructions (Sigma-Aldrich). Absorbance was measured at a wavelength of 570 nm with background subtraction at 630-690 nm. Then absorbance values from the MTT were converted to cells/well using a standard curve. Cell viabilities were calculated using the formula; [mean of cells in treated well]/[mean of cells in control well] × 100%. Each experiment was performed independently three times in triplicate.

Detection of cytokine release from D-LL-31-treated A549 cells
The immunomodulatory effect of D-LL-31 on A549 human lung epithelial cells was investigated using cytokine antibody arrays. A total of 5.5 × 10 6 A549 cells were seeded into a T25 tissue culture flask (Corning™) and grown overnight at 37˚C, 5% CO 2 . The cells were incubated with 10 μM of D-LL-31 at 37˚C, 5% CO 2 for 24 h (This concentration spanned the IC 50 values of the peptide against all tested bacterial strains grown as biofilm and showed no cytotoxicity in A549 cells). Cells in medium without peptide were used as a control. After incubation, the secreted cytokines in culture supernatant were detected using a human angiogenesis antibody array kit (AAH-ANG-1000-4, Norcross, GA) according to the manufacturer's instructions.
The images were taken using a chemiluminescence imaging system and the densitometry data were calculated using ImageQuant TL Software version 8.1 (Amersham™ Imager 600, GE Healthcare Life Sciences, Chicago, IL). Each experiment was performed independently two times in duplicate.

Statistical analysis
The percentage of killing activity of all tested agents against bacteria grown as biofilm, biofilm biomass, cytotoxicity and densitometry data of cytokine expression are presented as mean ± standard deviation (SD). Comparisons between the average percentage killing activity of each AMP at the same concentration in all tested agents were analyzed using one-way ANOVA. The statistical significance of cell viabilities and pixel density ratio in each treatment compared with controls was determined by Student's t test. All statistical testing was performed using the SPSS software, version 16.0 (Chicago, IL).

Isolation and identification of bacteria from CRS patients
Characteristics of the 20 patients (ten males and ten females) with CRS are shown in Table 2.
The ages ranged from 13 to 81 years (average 53 years). None was pregnant or had underlying diseases; diabetes mellitus, cancer or AIDS. Of 20 clinical specimens, ten yielded positive cultures on at least one of the media that were used. Sixteen microorganisms were isolated from maxillary or ethmoid sinus washings, 11 of which (69%) were Gram-positive and 5 were Gram-negative bacteria. About 75% (12/16) of the isolates were classified as facultative anaerobic bacteria and 25% (4/16) as obligate aerobes or aerobic bacteria.

The presences of biofilms on sinus mucosal tissues of CRS patients
Investigation using SEM revealed evidence of biofilm formation in about 75% (15/20) of sinus mucosal tissues samples from CRS patients, as shown in Table 2. Scanning electron micrographs of mucosal tissues with the indication of biofilms are shown Fig 1. The presence of a biofilm was defined on the basis of visible bacterial aggregation located on the surface of the epithelium. Large biofilm aggregates were observed (Fig 1C and 1D) and are distinct from the appearance of ciliated epithelium from samples that yielded bacteria in culture but without evidence of biofilm ( Fig 1A). Coccus-and bacillus-shaped elements were observed on the biofilm structure ( Fig 1D, white arrow). Biofilm obtained from CRS mucosal tissues showed strands of extracellular materials between the cells, which might represent extracellular DNA (eDNA) from bacterial and human cells (Fig 1B).

Antimicrobial activity of antibiotics and AMPs against all tested bacteria in planktonic form
The MIC and MBC values of conventional antibiotics against bacteria isolated from CRS patients are shown in Table 3. Resistance to AMX was observed in all strains. P. aeruginosa in planktonic form was susceptible TOB. All tested peptides, at 1 μM, exhibited � 95% killing activity against all tested isolates after 1 h of incubation (Fig 2). All peptides at 5 μM completely killed the initial inoculum after 2 h. All concentrations of AMX showed a low killing ability (about 0.5-18.5%) toward all bacterial strains in planktonic form during 1 and 2 h of incubation. TOB at concentrations of 1 and 5 μM showed � 92% killing effect against the clinical isolate P. aeruginosa at 1-2 h, but 10 μM was required to kill all bacterial cells (Fig 2C) while only 1 μM TOB showed � 98% killing effect against the reference strain P. aeruginosa ATCC 27853 at 1 h and both 5 and 10 μM killed all bacterial cells within 1 h (Fig 2D). Interestingly, only D-LL-31 (1 μM) caused rapid killing of all tested strains: bacterial cells at 5 x 10 5 CFU/ml were completely killed within 1 h of the peptide treatment.

The quantification of bacterial biofilm formation
The crystal violet assay (using 2-day biofilm) was performed to determine the biofilm-formation capacity of each bacterial stain. P. aeruginosa showed the greatest tendency for biofilm

PLOS ONE
D-LL-31 enhances biofilm-eradicating effect of antibiotic

Killing effect of all tested agents against bacteria under biofilm-stimulating condition
All agents displayed a clear dose-and strain-dependent killing activity against all tested bacteria grown as biofilm (Fig 3). The highest concentration of AMX (68418 μM or 25000 μg/ml) exhibited 70%, 75%, and 90% killing activity against P. aeruginosa, S. epidermidis and K. pneumoniae biofilm, respectively (Fig 3A). Biofilm of P. aeruginosa was found to be sensitive to TOB (Fig 3B). Among all peptides, D-LL-31 exhibited the strongest killing activity against all bacterial strains (Fig 3C). The killing activities of D-LL-31 at concentrations 1.56-25 μM were significantly higher than all tested peptides (P < 0.05) against S. epidermidis biofilm. However, there were no statistically significant differences in killing activity of D-LL-31 against biofilm of K. pneumoniae and P. aeruginosa when compared to the other peptides. The IC 50 values of each agent was further investigated (Table 4). D-LL-31 exhibited the lowest IC 50 values against all bacteria in biofilm at concentration ranging from 0.5 to 4 μM. IC 50 values of AMX against biofilm of P. aeruginosa, K. pneumoniae, and S. epidermidis were higher than those of D-LL-31 (about 11000-, 13000-and 66000-fold, respectively). In contrast, the IC 50 values of TOB and D-LL-31 toward P. aeruginosa biofilm were similar. However, these findings also indicated that D-LL-31 was more effective than the L-form and the LL-37 variants against all bacteria grown as biofilm.

Synergistic effect of D-LL-31 with antibiotics
The most effective antimicrobial peptide, D-LL-31, showed synergistic interactions with all tested antibiotics against bacteria grown as biofilm based on FICI � 0.5 ( Table 5). The combination of D-LL-31 with antibiotics exhibited 50% killing against each bacterial strain at IC 50 values that were at least 8-to 16-fold lower than those of either agent alone. A 16-fold reduction in IC 50 values of both D-LL-31 and AMX were found against P. aeruginosa and S. epidermidis biofilm (FICI = 0.125). The interaction of D-LL-31 with AMX and TOB reduced the IC 50 values of the antibiotics by 8-to 16-fold against K. pneumoniae and P. aeruginosa biofilm, respectively (FICI = 0.188).

Effect of D-LL-31 alone and in combination with antibiotics against biofilm matrix under static conditions
The green fluorescence of Alexa Fluor 1 488 was used to represent the biofilm matrix of all tested bacteria after exposure to D-LL-31 alone or in combination with antibiotics (Fig 4). A combination of D-LL-31 with AMX led to a distinct decrease in biofilm matrix for all tested strains relative to controls. A statistically significant reduction of P. aeruginosa biofilm matrix occurred when treated with a combination of D-LL-31 and TOB (Fig 4C). Moreover, a greater reduction of biofilm matrix was observed in all tested strains after challenge with D-LL-31 alone than after challenge with antibiotic alone.
To confirm the biofilm-disrupting effect of the agents, biofilm biomass (μm 3 /μm 2 ) was measured. Significant decreases of biomass were observed following D-LL-31 treatment alone (P < 0.01) and in combination with antibiotics (P < 0.001) relative to controls. There were no statistically significant effects of either antibiotic tested against K. pneumoniae and P. aeruginosa biofilm. However, AMX had a significant effect on S. epidermidis (P < 0.05).

Localization of D-LL-31 on biofilm matrix
The localization of the peptide on biofilm architecture was observed using TAMRA-labeled D-LL-31. Image slices were obtained at 2-μm increments moving from the top of the biofilm matrix through the substrate. The stacked sections of biofilm matrix of all tested bacterial strains in the presence of peptide are shown in Fig 5. A rapid binding of D-LL-31 to biofilm matrix of all tested bacteria was observed within 5 min of incubation. D-LL-31 could be detected in all stacks of K. pneumoniae and S. epidermidis biofilms (Fig 5A and 5B). The peptide accumulated mainly in the top layers of P. aeruginosa biofilm (Fig 5C), but some could also be seen in the middle layers.

Cytotoxicity of D-LL-31 on human lung epithelial cells
A cell-survival assay was performed to determine the cytotoxicity of D-LL-31 on human lung epithelial cells as shown in Fig 6. More than 90% of cells remained viable after treatment with 12.5 μM D-LL-31, and more than 80% after treatment with 25 μM, indicating that D-LL- However, treatment with � 50 μM D-LL-31 resulted in more than 90% cytotoxicity of A549 cells.

Expression of cytokines and angiogenesis-related proteins
The expression of angiogenesis-related proteins and inflammatory cytokines in the presence of D-LL-31 was determined using a cytokine array membrane assay. The pixel density ratio of each identified protein is shown in Fig 7A and 7B. Epithelial-derived neutrophil-activating peptide 78 (ENA-78) was significantly secreted in the presence of D-LL-31 (P < 0.05) relative to controls (Fig 7A). There was no significant change in levels of cytokines that might impact CRS, such as IL-4, IL-8, TNF-α and growth-related oncogene-a, b and g (GRO-a, b and g). The expression of IFN-γ, an important cytokine in the immunopathogenesis of CRSsNP, was not found after incubation with D-LL-31. Moreover, treatment with D-LL-31 did not induce expression of IL-6, a marker of inflammation in CRSwNP. In addition, TIMP-2 was significantly elevated in culture supernatant with D-LL-31 (P < 0.05) (Fig 7B).

Discussion
Bacterial infection play an important role in the pathogenesis, cause inflammation and prescribe antimicrobial therapy in CRS [44,45]. Moreover, biofilm on mucosal tissues of CRS patients have demonstrated the significant role in the pathogenesis and disease progression of CRS [46] and its association with antibiotic resistance [10,11]. In this study, the enhancement of biofilm-eradicating effect of currently used antibiotics for CRS treatment with cathelicidinderived peptides and D-enantiomers were observed. Various bacterial species were isolated from maxillary and ethmoid sinus irrigation samples from CRS patients. Of these, S. AMX is a reasonable first-line antibiotic for treatment of sinus infections in many geographic areas [49]. In addition, TOB is recommended to treat P. aeruginosa infection [50]. Our antimicrobial susceptibility results showed that all bacterial strains isolated from CRS patients in planktonic form were resistant to AMX, as previously reported [51]. Despite their resistance to AMX, we observed that the bacteria were sensitive to all AMPs tested. We were able to demonstrate that the D-enantiomeric LL-31 peptide had a higher antibacterial potency against all tested bacteria in planktonic culture than the other AMPs (LL-37, LL-31 and D-LL-37) and currently used antibiotics (AMX and TOB).
Long-term antibiotic therapy in CRS patients may be one of the reasons of the antibioticresistance. Moreover, microbial biofilm can be up to 1000 times more resistant to antibiotic treatment than its planktonic form in CRS patients [34,52]. Thus, we investigated the killing capacity of the tested agents against bacteria grown as biofilm. Our findings are consistent with Kifer D et al [53], given the greater resistance to AMX than their planktonic counterparts. However, all bacterial isolates grown as biofilm tended to be more sensitive to all peptides, especially D-LL-31. This protease-resistant peptide displayed a killing activity comparable to that of TOB against P. aeruginosa biofilm and greater biofilm-reduction properties than TOB. The mode of action of AMPs relies on their capacity to disrupt bacterial membranes, even in slow-growing or dormant biofilm-forming cells [8,54]. Thus, AMPs are better able to combat bacterial cells during biofilm-stimulating conditions. Several studies have shown that AMPs may penetrate into the biofilm matrix and kill the persister cells, a great advantage over conventional antibiotics [8,15].
The biofilm matrix accounts for up to 90% of the total biofilm mass and acts as a barrier for antibiotic diffusion [55], which largely explains why bacteria in biofilms are so resistant to antibiotics [8]. We have shown previously that D-LL-31 may disrupt the biofilm matrix leading to enhanced access by ceftazidime to kill bacterial cells, ultimately resulting in synergistic interaction [15]. Here, TAMRA-labelled D-LL-31 was used to observe the localization of the peptide before it exerted its effect. D-LL-31 was observed in all layers of K. pneumoniae and S. epidermidis biofilm matrix, whereas a small amount of peptide was seen in the mid-layers of P. aeruginosa biofilm (most being in the surface layer). Probably as a consequence, IC 50 values against P. aeruginosa were higher than other tested bacterial strains. Bacterial species differ in their biofilm matrix components [56]. For instance, the ability of P. aeruginosa to synthesize large amounts of the anionic polysaccharide alginate and eDNA probably reduces its susceptibility to magainin II and cecropin P1 and might entrap AMPs (positively charged molecules) before they can reach their bacterial target [57]. Additionally, the positively charged exopolymers (Pel and Psl) in P. aeruginosa biofilm can also cause electrostatic repulsion of the cationic polypeptide antibiotics (polymyxin B and colistin) [58], as also observed for human β-defensin-3 (hBD-3) and LL-37 against S. epidermidis [59]. Nevertheless, our results still indicate the rapid penetration and strong killing activity of D-LL-31 against all tested isolates in a strain-, biofilm-forming ability-, and matrix component-dependent manner. We can conclude that consideration of the interaction capacity of AMPs with biofilm matrix and interference of biofilmspecific molecules will be an essential part of the development of peptide-based therapeutics.
There is much evidence that AMPs in combination with common antibiotics often enhance the efficacy of the antibiotic and decrease development of antibiotic resistance [60]. In addition, previous reports have shown that biofilm-related infections in CRS patients are more difficult to clear, prone to relapse, inherently resistant to antibiotics and might lead to repeated surgery [32,34]. In our study, D-LL-31 showed synergistic effects with conventional antibiotics (AMX and TOB) against embedded bacterial cells within biofilm in all tested isolates, leading to greatly reduced IC 50 values (as much as 16-fold lower) and restoration of the efficacy of the antibiotics. The lowest quantity of biofilm matrix was observed after incubation with a combination of D-LL-31 and antibiotics in all isolates tested. Indeed, D-LL-31 alone could reduce biofilm matrix to a greater extent than did either antibiotic. We hypothesize that the improved antibiotic activity in the presence of D-LL-31 is likely due to the interference with intracellular signals for biofilm formation [61], down-regulation of genes essential for biofilm development [62], reduction of biofilm matrix synthesis [63] and biofilm disruption by the AMPs [15,64]. Consequently, these may recognize the possible mechanisms that perturbation of matrix synthesis and disruption of biofilm architecture caused by AMPs plays a key role in allowing increased access of currently used antibiotics toward biofilm-embedded bacteria, as also previously expected [15,64]. These findings demonstrate that enhanced activities of antibiotic in combination with D-LL-31 to eradicate the biofilm matrix and increase the bactericidal efficacy against embedded cells within biofilm may reduce the relapse cases and number of patients for revision endoscopic sinus surgery.
CRS is an inflammatory disorder disease [1,3]. Given that, we should avoid the use of any treatment that could induce expression of the major cytokines involved in immunopathogenesis of CRS [28]. Thus, the immunomodulatory effect of D-LL-31 was investigated. Previous research has indicated that elevated expression of IL-4, IFN-γ and IL-6 on airway epithelial cells of CRS patients resulted in decreased epithelial barrier function, a phenomenon that might account in part for the progression of CRS [21,22]. We did not detect IFN-γ and IL-6 in the presence of D-LL-31, while low levels of IL-4 were detected both in the presence and absence of the peptide. Although the neutrophil-attracting chemokines, IL-8 and GRO-a, b and g, increased following D-LL-31 treatment, this increase was not statistically significant. We also found high levels of ENA-78 in the culture supernatants with D-LL-31. Rudack and colleagues recently noted that IL-8 and ENA-78 appear to be of secondary importance for the chemotaxis of neutrophils in CRS [65]. Another recent study found that LL-37 causes increased IL-6 and IL-8 release from human nasal cells [66]; while, IL-6 (a marker of inflammation in CRSwNP) was not observed after incubation with D-LL-31. In addition, TNF-α has been thought to play a role in the pathogenesis of CRS via the recruitment of B cells [23]. We did not detect TNF-α expression after treatment with D-LL-31. Based on our results, D-LL-31 does not appear to induce the cytokines that play a prominent role in ongoing inflammatory reactions in patients with CRS.
Several studies have shown that a balance between MMPs and their regulators (TIMPs) controls the remodeling and repairing of airway ECM [24]. The elevated expression of MMP-2, -7 and -9, important endopeptidases for degrading the ECM, has been considered to play important roles in the pathogenesis of nasal polyposis in CRS patients [24][25][26]. Significantly decreased expression of TIMP-1 and -2 has been found in patients with CRS, failing to counterbalance activity of MMPs [25,27]. This imbalance contributes to the excessive degradation of ECM, formation of pseudocysts, albumin deposition and edema [28]. Interestingly, TIMP -2 was up-regulated in presence of D-LL-31. Perhaps D-LL-31 has a novel role in tissue remodeling, which may prevent the progression and development of nasal polyposis in patients with CRS.
Successful development of any pharmaceutical substance requires minimal or no toxicity. D-LL-31 exhibited no cytotoxicity on human lung epithelial cells after 24 h of incubation. Additionally, D-LL-31 exhibits only marginal hemolytic activity against hRBCs (less than 1%) [15].

Conclusion
In summary, our study demonstrated that D-LL-31 displayed strong broad-spectrum activity against bacteria in planktonic condition. Furthermore, this peptide exhibited a high potential to kill bacterial cells under biofilm-stimulating conditions when compared to all tested agents. Increased biofilm matrix eradication was observed when D-LL-31 was combined with conventional antibiotics. These findings suggested that D-LL-31 is an interesting candidate peptide for further development as an antibacterial or antibiofilm agent for use alone, or to enhance the efficacy of commonly used antibiotics to combat biofilm-related CRS. Co-treatment with this peptide may permit reduction in antibiotic concentrations, possible side effects and disease severity. D-LL-31 was not toxic to human lung epithelia cells. No cytokines involved in the immunopathogenesis of CRS were induced after treatment with D-LL-31. A role for D-LL-31 in the remodeling and repairing of airway extracellular matrix was also observed. Therefore, D-LL-31 is a feasible therapeutic supplement for applying to the biofilm-associated infection of CRS patients.
Supporting information S1 Table. Antimicrobial susceptibility test results of reference strain and representative bacteria used from CRS patients. Bacterial suspensions were incubated with antibiotics for 24 h and the results were interpreted according to BD BBL™ Sensi-Disc™ antimicrobial susceptibility test discs. (DOCX)