Potential novel therapeutic strategies in cystic fibrosis: antimicrobial and anti-biofilm activity of natural and designed α-helical peptides against Staphylococcus aureus, Pseudomonas aeruginosa, and Stenotrophomonas maltophilia

Background Treatment of cystic fibrosis-associated lung infections is hampered by the presence of multi-drug resistant pathogens, many of which are also strong biofilm producers. Antimicrobial peptides, essential components of innate immunity in humans and animals, exhibit relevant in vitro antimicrobial activity although they tend not to select for resistant strains. Results Three α-helical antimicrobial peptides, BMAP-27 and BMAP-28 of bovine origin, and the artificial P19(9/B) peptide were tested, comparatively to Tobramycin, for their in vitro antibacterial and anti-biofilm activity against 15 Staphylococcus aureus, 25 Pseudomonas aeruginosa, and 27 Stenotrophomonas maltophilia strains from cystic fibrosis patients. All assays were carried out in physical-chemical experimental conditions simulating a cystic fibrosis lung. All peptides showed a potent and rapid bactericidal activity against most P. aeruginosa, S. maltophilia and S. aureus strains tested, at levels generally higher than those exhibited by Tobramycin and significantly reduced biofilm formation of all the bacterial species tested, although less effectively than Tobramycin did. On the contrary, the viability-reducing activity of antimicrobial peptides against preformed P. aeruginosa biofilms was comparable to and, in some cases, higher than that showed by Tobramycin. Conclusions The activity shown by α-helical peptides against planktonic and biofilm cells makes them promising “lead compounds” for future development of novel drugs for therapeutic treatment of cystic fibrosis lung disease.


Background
Physicians treating patients with cystic fibrosis (CF) are increasingly faced with infections caused by multidrugresistant strains. Pseudomonas aeruginosa and Staphylococcus aureus are the most common bacterial pathogens isolated from the CF respiratory tract where they cause persistent infections associated with a more rapid decline in lung function and survival [1,2]. In recent years, however, there has been an increasing number of reports on potentially emerging and challenging pathogens, probably due to improved laboratory detection strategies and to selective pressure exerted on bacterial populations by the antipseudomonal antibiotic therapy [2]. In this respect, both the overall prevalence and incidence of intrinsically antibiotic-resistant Stenotrophomonas maltophilia isolations from CF respiratory tract secretions have been recently reported [3][4][5].
Efforts to treat CF infections are also hampered by the high microbial adaptation to the CF pulmonary environment, resulting in an increased ability to form biofilms intrinsically resistant to therapeutically important antibiotics such as aminoglycosides, fluoroquinolones, and tetracycline [6][7][8][9][10].
Novel antimicrobial agents that could replace or complement current therapies are consequently needed to fight chronic infections in CF patients.
Antimicrobial peptides (AMPs) are naturally occurring molecules of the innate immune system that play an important role in the host defence of animals and plants [11][12][13]. Over the last years, natural AMPs have attracted considerable interest for the development of novel antibiotics for several reasons [14,15]: i) the broad activity spectrum, comprised multiply antibiotic-resistant bacteria; ii) the relative selectivity towards their targets (microbial membranes); iii) the rapid mechanism of action; and, above all, iv) the low frequency in selecting resistant strains. Although the antimicrobial activity of AMPs has been extensively reported in literature [13][14][15][16][17], only few studies have been reported with respect to CF pathogens [18][19][20][21].
Hence, in an attempt to evaluate the therapeutic potential of AMPs in the management of CF lung infections, for the first time in the present study three cationic α-helical AMPs -two cathelicidins of bovine origin (BMAP-27, BMAP-28) and the artificial peptide P19(9/B) -were tested for their in vitro antibacterial effectiveness, as well as their in vitro anti-biofilm activity, against selected S. aureus, P. aeruginosa, and S. maltophilia strains collected from CF patients. The efficacy of the AMPs was compared to that of Tobramycin, selected as the antibiotic of choice used for chronic suppressive therapy in CF patients.
Since the conditions present in the CF patients' airway surface liquid could counteract the potency of antibiotics such as Tobramycin [22,23], in the present study all in vitro antimicrobial assays were carried out under experimental conditions simulating the physical-chemical properties observed in CF lung environment [24][25][26].

Phenotypic features and clonal relatedness of CF strains
A total of 9 out of 25 P. aeruginosa strains tested showed mucoid phenotype on MHA, while 3 exhibited SCV phenotype. Among 15 S. aureus isolates tested, 7 were methicillin-resistant.

In vitro activity of AMPs and Tobramycin against planktonic cells: MIC, MBC
In order to determine the efficacy of AMPs, the antimicrobial activity was measured against 67 CF clinical isolates, and results are summarized in Table 1. Overall, BMAP-28 showed the widest activity spectrum among AMPs tested, as suggested by MIC 90 and MBC 90 values (16 μg/ml, for both), although all of them exhibited a species-specific activity. In fact, although AMPs showed comparable activity against P. aeruginosa, BMAP-28 was found to be more active than P19(9/B) against S. maltophilia, and resulted the best active AMP against S. aureus (MIC 90 : 32 μg/ml; MBC 90 : 32 μg/ml). Compared to AMPs, Tobramycin exhibited a lower activity (MIC 90 and MBC 90 : >64 μg/ml) regardless of the species considered. Killing quotient values, calculated as MBC/MIC ratio, were < 4 for all AMPs, as well as for Tobramycin, clearly suggesting a bactericidal activity. No differences in susceptibility levels to AMPs were found with regard to phenotype (mucoid, SCV, MRSA), pulsotype, or susceptibility to Tobramycin (data not shown).
MIC and MBC values obtained under CLSIrecommended or "CF-like" experimental conditions (see Materials and Methods section) are shown in Table 2. Comparative evaluation of these values showed that mean MIC CF-like /MIC CLSI and MBC CF-like /MBC CLSI values obtained for Tobramycin (23.9 and 15.6, respectively) were significantly higher than those observed for BMAP-27 (1.5 and 1.2, respectively; p < 0.001), BMAP-28 (0.5 and 0.5, respectively; p < 0.001), and P19(9/B) (2.8 and 2.9, respectively; p < 0.001), regardless of species tested, indicating a reduced antibiotic activity of Tobramycin in CF-like conditions.

Bactericidal kinetics
Time-killing results have been summarized in Figure 1. BMAP-27, BMAP-28, and P19(9/B) exerted a rapid bactericidal activity against P. aeruginosa, reducing the number of viable bacterial cells of at least 3 logs within 60 min of exposure. However, the bactericidal effect of BMAP-28 against P. aeruginosa was incomplete for two (Pa6 and Pa22) of the three strains tested, allowing bacterial regrowth after 24-h incubation, although at levels lower than those observed for untreated control. In parallel experiments, Tobramycin showed only a bacteriostatic effect against P. aeruginosa, causing no more than 1-log reduction in viable count after 24 h. BMAP-27, BMAP-28 and P19(9/B) exerted bactericidal activity also against S. maltophilia, although with streaking strain-specific differences. Particularly, BMAP-28 exhibited only bacteriostatic effect against Sm192 strain, while P19(9/ B) showed a rapid bactericidal effect against Sm138 strain, causing more than a 4-log reduction in viable count after 10 min-exposure. Tobramycin exhibited a late (after 24-h exposure) bactericidal effect only against Sm138 strain.
AMPs activity against S. aureus was significantly strainspecific, ranging from the rapid bactericidal activity of BMAP-28 against Sa10 strain, to the bacteriostatic effect of P19(9/B) and BMAP-28 against Sa4 strain. Tobramycin   showed a bactericidal effect against all S. aureus strains tested, although allowing bacterial regrowth of Sa4 strain after 2-h exposure.

In vitro activity of AMPs and Tobramycin against biofilm
All CF strains were screened for biofilm forming ability on polystyrene. A significantly higher proportion of biofilm producer strains was found in P. aeruginosa and S. aureus, compared to S. maltophilia (96 and 80% vs 55%, respectively; p < 0.01) (data not shown). However, efficiency in biofilm formation was significantly higher in P. aeruginosa than in S. aureus, as suggested by median biofilm amounts produced (0.162 vs 0.109, respectively; p < 0.01) (data not shown).   To determine if AMPs could be prophylactically used to prevent biofilm formation, we tested the effect of AMPs and Tobramycin at sub-inhibitory concentrations (1/2x, 1/4x, and 1/8xMIC) against biofilm formation ( Figure 2). Tobramycin at 1/2x and 1/4xMIC caused a significantly higher reduction in biofilm-forming ability of S. maltophilia and S. aureus, in comparison with the three AMPs. This effect was more relevant with S. aureus, being observed also at 1/8xMIC. Tobramycin showed to be more effective than BMAP-27 against P. aeruginosa at concentrations equal to 1/4x and 1/8xMIC. The activity of Tobramycin in reducing biofilm formation was not related to drug susceptibility (data not shown). Among AMPs, BMAP-28 and P19(9/B) at 1/2xMIC were significantly more active compared to BMAP-27, and BMAP-28 at 1/4xMIC was significantly more active than other AMPs against S. aureus.
We further evaluated AMPs as potential therapeutics for CF by testing their efficacy against preformed biofilms. To this, BMAP-27, BMAP-28, P19(9/B), and Tobramycin at 1xMIC and at bactericidal concentrations (5x, and 10xMIC) were assayed against preformed (24 h) biofilms by six representative P. aeruginosa strains selected for high biofilm formation ability ( Figure 3).
The activity of AMPs and Tobramycin against preformed biofilms resulted to be similar in 5 out of 6 strains tested, causing a highly significant reduction of biofilm viability compared to the controls (biofilm not exposed; p < 0.0001), regardless of the concentrations tested ( Figure 3). AMPs showed to be active at all concentrations, also against biofilms formed by P. aeruginosa Pa32, against which Tobramycin was effective only at the highest concentration used (10xMIC). The activity of Tobramycin against preformed biofilms was not related to drug susceptibility (data not shown).

Discussion
This study was aimed at verifying the potential of some α-helical AMPs as lead compounds for the development of novel antimicrobials to treat lung disease in CF patients. To this, we tested the in vitro susceptibility of P. aeruginosa, S. maltophilia and S. aureus CF isolates to the naturally occurring AMPs BMAP-27 and BMAP-28, as well as the rationally designed P19(B/9), and we compared their effectiveness with that of Tobramycin, the antibiotic of choice for the inhalation therapy of chronic airway infections in CF patients.
Effort to treat CF are also hampered by the conditions present in patients' airway surface liquid where the accumulation of large volumes of viscous sputum (mucus) providing bacteria with a nutritionally rich growth environment composed of host-and bacterial-derived factors which deeply change their phenotype and possibly their susceptibility against AMPs [31]. Therefore, to accurately judge the feasibility of these peptides as potential anti-infectives in the context of CF, in this study we investigated the activity of AMPs under some CF-like experimental conditions, including acidic pH, reduced O 2 tension, and a chemically defined medium mimicking the nutritional composition of CF sputum [24][25][26].
These conditions allow pathogens to assume a physiology similar to that shown in vivo in the CF lung [24] and constitute a more realistic model to assay their sensitivity to AMPs.
Evaluation of MIC and MBC values, as well as timekilling assays against planktonic forms of different CF isolates of P. aeruginosa, S. maltophilia, and S. aureus, have shown that all three AMPs are highly active in vitro against most tested strains, although BMAP-28 showed the widest spectrum of activity. It is noteworthy that all the three peptides exhibited an activity higher than Tobramycin. This observation is even more evident when considering the molar concentration (μM) of each compound rather than that by weight (μg/ml), given that the peptides tested are at least six folds heavier than Tobramycin.
The poor activity showed by Tobramycin is probably due to the experimental conditions used in this study, as suggested by comparative evaluation of MIC values observed in both "CF-like" and CLSI-recommended conditions. On the contrary, the activity of AMPs tested resulted to be slightly enhanced (BMAP-28), unaffected (BMAP-27), or slightly reduced [P19(9/B)] in "CF-like" conditions, compared to CLSI-recommended ones, so they can be considered to be quite robust and medium insensitive.
MBC/MIC ratio clearly indicated that all AMPs exert a bactericidal effect against the CF isolates, in agreement with the known capability of BMAP-27, BMAP-28 and P19(B/9) to kill target cells by rapid permeabilization of their membranes [28]. Results of killing kinetic assays confirmed this mode of action, although bactericidal activity against S. aureus and S. maltophilia was straindependent. Again, the potency of AMPs was overall comparable or higher than that showed by Tobramycin.
Due to the different mechanism of action showed by AMPs and Tobramycin, we investigated the potential synergy between them. Interestingly, Tobramycin exhibited synergy with both BMAP-27 and P19(9/B) against planktonic S. aureus Sa4 and Sa10 strains, both resistant to Tobramycin, thus suggesting that at least in these cases both AMPs may overcome resistance to Tobramycin by facilitating the internalization of the aminoglycoside into the bacterial cells. Further studies on a more representative number of S. aureus strains will be mandatory to understand the mechanism of this synergy and the Figure 3 Activity of AMPs at bactericidal concentrations against preformed P. aeruginosa biofilms. BMAP-27, BMAP-28, P19(9/B), and Tobramycin were tested at 1x (white bars), 5x (gray bars), and 10xMIC (black bars) against preformed biofilm by 6 P. aeruginosa CF strains. Results are expressed as percentage of biofilm' viability compared to control (not exposed, 100% viability). ** p < 0.0001, Fisher's exact test.
feasibility to use these AMPs in association with traditional antibiotic treatments.
Within the CF lung, pathogens cells grow as biofilms, which are inherently recalcitrant to antimicrobial treatment and host response [32]. Even worse, it has recently been reported that some antibiotics may even stimulate biofilm formation at subinhibitory concentrations [7]. Biofilm resistance is mainly due to the slow growth rate and low metabolic activity of bacteria in such community. For these reasons, AMPs whose mechanism of action makes them active also on non-growing bacteria, should be able to efficiently inhibit or prevent biofilm formation.
Our results in fact indicate that the three α-helical peptides were all able to reduce biofilm formation, although generally at a less extent than Tobramycin. In particular, all peptides reduced the capacity of P. aeruginosa, S. maltophilia and S. aureus to form biofilms when used at sub-inhibitory concentrations, with the strongest effects at about 1/2xMIC values, while Tobramycin was efficacious also at lower concentrations (1/4x, and 1/8x MIC). This effect was particular evident with the isolates of S. aureus. Interestingly, no planktonic growth inhibition was observed at concentrations able to reduce biofilm formation, and also AMPs with poor killing capacity against some planktonic cells showed antibiofilm effects. These observations suggest that BMAP-27, BMAP-28 and P19(9/B) may interfere with biofilm formation by different mechanisms other than direct antimicrobial activity similarly to what observed with the human cathelicidin LL-37 [33], and recently reviewed by Batoni et al. [34].
Most CF patients are infected by P. aeruginosa whose persistence is due to the formation of antibiotic resistant biofilms in the lung [35]. Our results showed that BMAP-27, BMAP-28, and P19(9/B) were also as effective as Tobramycin in reducing cell viability of preformed biofilms formed by selected strains of P. aeruginosa. At MIC concentrations, and even more at 5xMIC values, the two cathelicidins caused highly significant reduction of biofilm viability of all six strains of P. aeruginosa whereas Tobramycin showed comparable results only for five isolates. It has previously been reported that extracellular DNA is an important biofilm component [36], and that in P. aeruginosa it is involved in cell-cell attachment and biofilm development [37]. Due to the high affinity of cationic AMPs for DNA [38], it may be presumed that this binding might facilitate the detachment or disruption of otherwise-stable biofilm structures.

Conclusions
The overall results of this study shed new insights on the antibacterial properties of α-helical peptides, allowing the selection of those with the best properties to cope with lung pathogens associated to CF. BMAP-27, BMAP-28 and also the rationally designed P19(9/B) may thus be considered useful not only as lead compounds for the development of novel antibiotics but also for compounds that may counteract bacterial biofilm formation and eradicate preformed biofilms, reflecting the modern understanding of the role of biofilm formation in chronic CF infections. However, before applying these molecules in the future for early prophylactic and therapeutic treatment of CF lung disease, further in vitro studies (against other CF pathogens, such as Burkholderia cepacia, and fungi), as well as in vivo studies are needed to evaluate their therapeutic potential.

Bacterial strains
Overall, 67 antibiotic-resistant bacterial strains were tested in the present study: 15 S. aureus, 25 P. aeruginosa, and 27 S. maltophilia. Strains were collected from respiratory specimens obtained from patients admitted to the CF Operative Unit, "Bambino Gesù" Children's Hospital and Research Institute of Rome. Identification to species level was carried out by both manual (API System; bioMérieux, Marcy-L'Etoile, France) and automated (BD Phoenix; Becton, Dickinson and Company, Buccinasco, Milan, Italy) biochemical test-based systems. Each isolate was collected from a single patient and resistant to at least three of the following groups of antibiotics: β-lactams with or without β-lactamase inhibitor, aminoglycosides, fluoroquinolones, folate-pathway inhibitors (trimethoprim-sulphamethoxazole), tetracyclines, and macrolides. Strains were stored at −80°C in a Microbank system (Biolife Italiana S.r.l., Milan, Italy) and subcultured in Trypticase Soya broth (Oxoid S.p.A., Milan, Italy), then twice on Mueller-Hinton agar (MHA; Oxoid S.p.A) prior to the use in this study.

Phenotypic and genotypic characterization of CF strains
All strains grown on MHA were checked for mucoid phenotype and the emergence of small-colony variants (SCVs). Further, they were screened for their susceptibility to antibiotics by agar-based disk diffusion assay, according to the CLSI criteria [39], and by the Etest following the manufacturer's instructions assays (Biolife Italiana S.r.l.; Milan, Italy).
All CF strains tested in this study were genotyped by Pulsed-Field Gel Electrophoresis (PFGE) analysis in order to gain clue on genetic relatedness of strains. DNA was prepared in agarose plugs for chromosomal macrorestriction analysis as previously described [40,41]. For S. aureus isolates, agarose plugs were digested with enzyme SmaI (40U). DNA from P. aeruginosa and S. maltophilia isolates was digested using XbaI (30U). PFGE profiles were visually interpreted following the interpretative criteria previously described [27,40]: in particular, isolates with indistinguishable PFGE patterns were assigned to the same PFGE subtype; for S. aureus, isolates differing by 1 to 4 bands were assigned to different PFGE subtypes within the same PFGE type; for S. maltophilia and P. aeruginosa, isolates were assigned to the same PFGE type with different PFGE subtypes when they differed by 1 to 3 bands.

"CF-like" experimental conditions
In order to simulate the physical-chemical properties observed in CF lung environment [24][25][26], all in vitro antimicrobial assays against planktonic (MIC, MBC, time-kill kinetics, synergy testing) and sessile (biofilm formation, preformed biofilms) cells were performed in "CF-like" conditions: i) under reduced oxygen concentration (5% CO 2 ); ii) at acidic pH (6.8); and iii) in a chemically defined "synthetic CF sputum medium" (SCFM), that mimics the nutritional composition of CF sputum [24]. SCFM was prepared by using Casamino Acids Vitamin Assay (BD Difco) mixture containing each amino acid at concentration not significantly different from that originally described by Palmer and coworkers [24], except for a reduced amount of glycine and ornithine, which were therefore added from ad hoc prepared stock solutions to reach their required concentration.

Susceptibility testing
MICs and MBCs were determined by microdilution technique, in accordance with CLSI M100-S20 protocol [39], with some modifications. Briefly, serial two-fold dilutions (64 to 0.12 μg/ml) of each AMP and Tobramycin (Sigma-Aldrich S.r.l.; Milan; Italy) were prepared in SCFM at a volume of 100 μl/well in 96-well microtiter plates (Bibby-Sterilin Italia S.r.l.; Milan, Italy). Each well was then inoculated with 5 μl of a standardized inoculum, corresponding to a final test concentration of about 0.5-1 × 10 5 CFU/well. After incubation at 37°C for 24 h, the MIC was read as the lowest concentration of the test agent that completely inhibited visible growth. To measure the MBC, 100 μl of broth from clear wells were plated on MHA plates, and incubated at 37°C for 24 h. MBC was defined as the lowest concentration of the test agent killing of at least 99.99% of the original inoculum.
To evaluate the impact of "CF-like" experimental conditions on the antimicrobial activity of AMPs and Tobramycin, a set of PFGE-unrelated isolates representative for different levels of susceptibility to Tobramycin (4 P. aeruginosa, 3 S. maltophilia, and 4 S. aureus) was also tested for MIC and MBC values determined under standard CLSI-recommended conditions (i.e., aerobic atmosphere, cation-adjusted Mueller-Hinton broth, and pH 7.2).

Time-killing assay
Kinetics of AMPs' and Tobramycin' activity was evaluated by using the broth macrodilution method against three representative isolates within each tested species. Briefly, the standardized inoculum (1x10 5 CFU/mL) was exposed to the test agent at 1xMIC in SCFM, and incubated at 37°C. After 10 min, 30 min and 1, 2, and 24-h of incubation, aliquots of each sample were diluted and plated onto MHA, then the viable counts determined after 24-h of incubation at 37°C. Killing curves were constructed by plotting the log CFU/mL versus time.

Synergy testing
The activity of each AMP combined to Tobramycin against CF strains was evaluated by checkerboard technique by using 96-well polystyrene microplate (Kartell S. p.A., Noviglio, Milan, Italy). Briefly, concentrations of multiple compounds (range: 64-0.12 μg/ml) were combined in standard MIC format along with 5 × 10 5 CFU/ ml of tested. Inoculated microplates were incubated at 37°C for 24 h under 5% CO 2 . At the end of the incubation, for each combination interaction a Fractional Inhibitory Concentration (FIC) index was calculated as follows: FIC index = Σ (FIC A + FIC B ), where FIC A is the MIC of drug A in the combination/MIC of drug A alone, and FIC B is the MIC of drug B in the combination/MIC of drug B alone. Synergy was defined as a FIC index of ≤0.5, indifference as a FIC index of >0.5 to ≤ 4, and antagonism as a FIC index of > 4.

In vitro activity against biofilm formation
In each well of a 96-well flat-bottom polystyrene tissueculture microtiter plate (Iwaki; Bibby-Sterilin Italia S.r.l.), 5 μl of a standardized inoculum (1-5 × 10 7 CFU/ml) were added to 100 μl of SCFM containing test agent at 1/2x, 1/4x, and 1/8xMIC. After incubation at 37°C for 24 h, non-adherent bacteria were removed by washing twice with 100 μl sterile PBS (pH 7.2; Sigma-Aldrich S.r.l.). Slime and adherent cells were fixed by incubating for 1 h at 60°C, and stained for 5 min at room temperature with 100 μl of 1% crystal violet solution. The wells were then rinsed with distilled water and dried at 37°C for 30 min. Biofilms were destained by treatment with 100 μl of 33% glacial acetic acid for 15 min, and the OD 492 was then measured. The low cut-off was represented by approximately 3 standard deviations above the mean OD 492 of control wells (containing medium alone without bacteria). The percentage of inhibition was calculated as follows: (1 -OD 492 of the test/OD 492 of nontreated control) x 100.
In vitro activity against preformed P. aeruginosa biofilms In vitro activity of AMPs and Tobramycin was evaluated against biofilms formed by 6 P. aeruginosa strains, selected because strong biofilm-producers. Biofilms were allowed to form in each well of a 96-well flat-bottom polystyrene tissue-treated microtiter plate (Iwaki), as described above. Biofilms samples were then exposed to 100 μl of drug-containing SCFM (prepared at 1x, 5x, and 10x MIC). After incubation at 37°C for 24 h, nonadherent bacteria were removed by washing twice with 100 μl sterile PBS (pH 7.2), and biofilm samples were scraped with a pipette tip following 5-min exposure to 100 μl trypsin-EDTA 0.25% (Sigma-Aldrich S.r.l.). Cell suspension was then vortexed for 1 min to break up bacterial clumps. Bacterial counts were assessed by plating serial 10-fold dilutions of the biofilm cell suspension on MHA plates.

Statistical analysis
All experiments were performed at least in triplicate and repeated on two different occasions. Differences between frequencies were assessed by Fisher's exact test. Statistical analysis of results was conducted with GraphPad Prism version 4.00 (GraphPad software Inc.; San Diego, CA, USA), considering as statistically significant a p value of < 0.05.