Synergistic antibacterial and anti-biofilm activities of resveratrol and polymyxin B against multidrug-resistant Pseudomonas aeruginosa

Bacterial infection caused by multidrug-resistant Pseudomonas aeruginosa has become a challenge in clinical practice. Polymyxins are used as the last resort agent for otherwise untreatable Gram-negative bacteria, including multidrug-resistant P.aeruginosa. However, pharmacodynamic (PD) and pharmacokinetic (PK) data on polymyxins suggest that polymyxin monotherapy is unlikely to generate reliably efficacious plasma concentrations. Also, polymyxin resistance has been frequently reported, especially among multidrug-resistant P.aeruginosa, which further limits its clinical use. A strategy for improving the antibacterial activity of polymyxins and preventing the development of polymyxin resistance is to use polymyxins in combination with other agents. In this study, we have demonstrated that resveratrol, a well tolerated compound, has synergistic effects when tested in vitro with polymyxin B on antibacterial and anti-biofilm activities. However, its’ systemic use is limited as the required high plasma levels of resveratrol are not achievable. This suggests that it could be a partner for the combination therapy of polymyxin B in the treatment of topical bacterial infection caused by MDR P.aeruginosa.


Introduction
Pseudomonas aeruginosa is Gram-negative bacterium that widely exists in various ecological environments, and one of the most common opportunistic pathogens that cause clinical infections [1,2]. In 2017, it has been recognized as a priority pathogen for research and development of new antibiotics by the World Health Organization [3]. In the past years, with the increasing use of antibiotics in clinical practice, antibiotic resistance has become even more serious for P.aeruginosa [4,5]. Especially in recent years, with the frequent emergence of MDR strains, the treatment of P.aeruginosa infection has become a big challenge faced by clinicians [6,7].
Polymyxin B, an old antibiotic that had been neglected for several decades due to its relatively poor clinical efficacy and high risk of nephrotoxicity, was reemployed in clinical practice in recent years because of the high prevalence of MDR bacteria [8][9][10]. It is even regarded as the last resort agent for the treatment of MDR bacterial infections, even though this antibiotic is far from being an ideal antimicrobial agent [11,12]. Unfortunately, polymyxin B resistance has emerged in various bacterial species and been frequently reported in recent years [13,14]. Given the increasingly serious antibiotic resistance of pathogenic bacteria and the current resistance to polymyxin B, polymyxin B has been used in combination with other active antimicrobial agents when it is possible, with the aim to improve the antimicrobial effect and prevent the emergence of polymyxin B resistance [15,16].
Resveratrol is a naturally occurring polyphenolic compound that has received attention for its health benefits, including anticancer, antiaging and anti-inflammatory [17,18]. This compound has also shown its potential value in antibacterial therapy. For example, it has been reported to enhance the antimicrobial efficacy of aminoglycosides against Staphylococcus aureus, to rescue the antibacterial activity of chlorhexidine against carbapenem-resistant Acinetobacter baumannii, and to increase the bactericidal effect of polymyxin B against Klebsiella pneumoniae and Escherichia coli [19][20][21]. However, the potential activity of resveratrol on the antibacterial property of polymyxin B against P.aeruginosa is still unknown.
Here we investigate the synergistic effect of resveratrol and polymyxin B on anti-biofilm formation and the bactericidal effect against P.aeruginosa with or without established biofilm, with the intent to find a partner for combination therapy with polymyxin B for the treatment of MDR P.aeruginosa infection.

Material and methods
Bacterial isolates and chemicals 7 different isolates of MDR P.aeruginosa, including 5 polymyxin B susceptible and 2 resistant isolates, were specially selected from a strain repository at the Department of Microbiology, Shiyan Renmin Hospital (Hubei, China). All the isolates were obtained from different inpatients and specimens and identified using MALDI-TOF mass spectrometry (MS). P. aeruginosa ATCC 27853 was used as the control strain. Polymyxin B MICs were interpreted accordingly to the EUCAST clinical breakpoints, version 12.0 (https://www.eucast.org/). Before each experiment, the isolates were cultured onto M-H agar plate and incubated at 37°C for 24 h, and then one colony was selected for further study. Both resveratrol and polymyxin B used in this study were purchased from Solarbio (Beijing, China) and dissolved with sterile water to prepare stock solution with the original concentration of 20 mg/ml for polymyxin B and 100 mg/ml for resveratrol. A small amount of dimethyl sulfoxide (DMSO) was used to assist the dissolution of resveratrol and DMSO control groups were always included for each experiment.

Susceptibility testing and genetic analysis
MICs of resveratrol and polymyxin B were determined with the Broth microdilution method according to the Clinical and Laboratory Standards Institute (CLSI) guideline (https://clsi.org/). Briefly, bacterial cell suspension prepared with cation-regulated Mueller-Hinton broth (CAMHB) was dispensed into wells of 96-well microtiter plate and made the final bacterial concentration of 5 × 10 5 CFU/ml for each well. Polymyxin B and resveratrol were separately added into those wells to get the desired concentration, ranging from 32 μg ml −1 to 0.125 μg ml −1 with two-fold dilution for polymyxin B, and from 512 μg ml −1 to 8 μg ml −1 with twofold dilution for resveratrol. The plates were read after incubating at 37°C for 24 h, and the MIC was defined as the lowest concentration of drug required to visibly inhibit the growth of bacteria. PCR was performed on two polymyxin B resistant isolates with specially designed primers to detect five polymyxin B-resistance related genes including PmrA, PmrB, PhoP, PhoQ, and mcr-1, and the final products were sequenced and analyzed by Biomed (Beijing, China).

Checkerboard studies
Basing on the susceptibility testing result, checkerboard assay was further conducted to determine the interaction between polymyxin B and resveratrol on antibacterial activity. In this experiment, polymyxin B and resveratrol were prepared by two-fold serial dilutions and mixed to produce different concentration combinations. The result was read after incubated at 37°C for 24 h and interpreted as follows: FICI ≤ 0.5, synergistic effects; 0.5< FICI ≤ 1.0, additive effects; 1.0< FICI ≤ 2.0, no interaction; and FICI > 2.0, antagonistic effects. The result was confirmed with at least three independent experiments.

Time-kill assays
Time-kill studies were performed as described previously with slight modification with all isolates [22]. Resveratrol used in this experiment was maintained at the concentration of 64 μg/ml and the concentration of polymyxin B for each isolate was determined as the 1/4 MIC of polymyxin B. Bacteria suspension of 5 × 10 5 CFU/ml was freshly prepared and incubated with polymyxin B and resveratrol alone or in combination at 37°C. Cell counts were determined at the time points of 0, 1, 2, 4, 6, 8, 24 h respectively, by plating 100 μl of the culture at different time points on M-H agar plates with appropriate dilution. Synergistic activity was defined as ≥ 2log10 decrease in CFU/ml by two-drug combination group compared with every single drug group [23].

Biofilm formation
Bacterial cell suspensions of approximately 5 × 10 5 CFU/ml were freshly prepared with LB broth and dispensed into wells of 96-well microtiter plate. After static incubation at 37°C for 48 h, the liquid phase was discarded and the wells were gently washed with PBS for five times. Biofilm was fixed by adding 99% methanol to each well for 30 min. The fixative was then removed and the plate was allowed to air dry before the addition of crystal violet for 15 min to stain the adherent biofilm. The well was washed five times with PBS to remove free cells and 100 μl of 7% acetic acid was then added to solubilize the dye combined by biofilm. Absorbance, which reflected the established biofilm, was measured at 595 nm on a reader. All of those isolates were ranked according to their OD value from large to small, and the top three isolates were selected for further study.

Biofilm formation inhibition analysis
Three isolates with high biofilm formation capability were selected and biofilm formation analysis was performed as described above. Bacterial cell suspension of approximately 5 × 10 5 CFU/ml was dispensed into wells of 96-well microtiter plate and mixed with polymyxin B and resveratrol alone or in combination. To ensure that the inhibitory function is mainly on biofilm formation rather than on bacteria growth, the subinhibitory concentration of drug that has less than 10% influence on bacteria growth was employed in this experiment. The influence on bacteria growth was measured by incubating the bacteria suspension of 5 × 10 5 CFU/ml with polymyxin B and resveratrol alone or in combination at 37°C for 48 h and calculated as the decreased percentage of cell number compared with the control group. Resveratrol concentration of 8 μg ml −1 was employed in this experiment, which has less than 10% influence on bacteria growth. While the concentration of polymyxin B for each isolate was determined, by incubating the bacterial suspension of 5 × 10 5 CFU/ml with a series of diluted polymyxin B in combination with 8 μg ml −1 resveratrol, as the highest polymyxin B concentration of those combinations which has less than 10% influence on bacteria growth. With each newly prepared bacterial suspension, two experiments were performed simultaneously, one was for inhibitory function on biofilm formation analysis, the other for influence on bacteria growth analysis. After incubating at 37°C for 48 h, OD value and cell count were measured respectively, and the result was confirmed with at last three independent experiments.

Bactericidal activity analysis with established biofilm
The isolate with the highest biofilm formation capability was selected for this experiment. Biofilm was cultivated as described above and the cells protected by established biofilm were obtained by discarding the content and gently washing the wells with sterilized PBS solution five times. Change in bacteria sensitivity of this isolate to polymyxin B was confirmed by comparing the influence of polymyxin B on the growth of cells with or without established biofilm. Briefly, 24 wells with established biofilm were divided into two groups. One group was for calculating the average quantity of cells attached to the bottom of well. 100 μl sterilized PBS solution was added into each well, and a sterile pipette tip was used to scrape and blow the bottom of the well to release the adherent cells and mix the suspension. The cell number was counted by plating the suspension onto M-H agar plates with appropriate dilution. The other group was for analyzing the sensitivity of those cells protected by established biofilm. In total, 100 μl cation-regulated Mueller-Hinton broth with 8 μg ml −1 polymyxin B was added into each well and incubated at 37°C for 24 h. Then total cell number in each well, including both the adherent and planktonic cells, was counted by plating onto M-H agar plates with appropriate dilution. When the mean number of cells attached to the bottom of well with established biofilm had been obtained with the former group, 100 μl freshly prepared bacterial suspension, which could provide approximately similar (equal or small above but no more than 10%) amount of cells as the calculated result, was mixed with 8 μg ml −1 polymyxin B and added into wells of 96-well microtiter plate. After incubated at 37°C for 24 h, the total cell number was counted and compared with that of the group with established biofilm.
To measure bactericidal activity of polymyxin B and resveratrol against bacterial cells protected by biofilm, every 12 wells with established biofilm from the same batch were divided into a group. In total, 100 μl cation-regulated Mueller-Hinton broth, mixed with polymyxin B and resveratrol alone or in combination, was added into each well. Given the decreased sensitivity of cells with established biofilm, relatively high concentration of drugs, 32 μg ml −1 for polymyxin B and 128 μg ml −1 for resveratrol, were employed in this experiment. After incubating at 37°C for 24 h, the total cell number was counted and statistical analysis was performed. The result of this experiment was confirmed with three independent experiments.

Statistical analysis
Every performance was done in triplicate at least. Statistical analysis was carried out with Student's t test and P < 0.01 was considered as significant.

Bacterial isolates
Among 7 MDR P.aeruginosa isolates selected for this study, as well as one control isolate ATCC27853, 6 isolates (PA2, PA4, PA5, PA6, PA7, ATCC27853) were sensitive to polymyxin B and 2 isolates (PA1, PA3) were resistant (Table 1). PmrA, PmrB, PhoP, PhoQ, and mcr-1 are five genes that have been reported to be closely associated with polymyxin B resistance. PCR and gene sequencing were performed with two polymyxin B resistant isolates for those five genes. PmrA mutation (L71R) was found in both isolates, while PmrB mutations (Y345H) and PhoQ mutation (V260G) in PA1 (Table 2). No mutation in PhoP was found in either isolate. Gene mcr-1, which has been proved to lead to polymyxin B resistance, could not be detected in either of those two isolates, suggesting that this gene may be unrelated to polymyxin B resistance in this study. Biofilm formation is another characteristic of P.aeruginosa, which can lead to antibiotic resistance, but not all P.aeruginosa isolates have the same biofilm formation capability. In this study, biofilm formation capability was measured as the OD value of the crystal violet adhered by established biofilm, and the top three isolates were PA2, PA1, and PA7, with the OD595 of 1.351, 0.950, and 0.802 respectively, compared with the OD595 of 0.119 for the negative control strain PA4 (Table 1).

Antibacterial activity of resveratrol and its synergistic effect with polymyxin B
To assess the combinatorial effect of resveratrol and polymyxin B against MDR P.aeruginosa, we first determined the MIC of resveratrol and polymyxin B respectively using the broth microdilution method. The MIC values of resveratrol for all isolates were >512 μg ml −1 , indicating the lack of intrinsic bactericidal activity. However, resveratrol could effectively increase the sensitivity of those isolates to polymyxin B, with the MIC of polymyxin B reduced by 4-8 fold with 64 μg ml −1 resveratrol. Also, the synergistic effect of resveratrol with polymyxin B could be observed with the concentration of ≥64 μg ml −1 for resveratrol with all isolates with the FIC index of <0.5 (Table 1). It is likely that 32 μg ml −1 be the minimum concentration required by resveratrol for combination therapy since below this concentration (16 μg ml −1 and 8 μg ml −1 ), no synergistic effect with polymyxin B can be detected with any tested isolate.

Synergistic activity of resveratrol and polymyxin B in time-kill assays
To further confirm the combination effect of resveratrol and polymyxin B on bactericidal activity, time-kill assay was performed with the inoculums of 5 × 10 5 CFU/ml incubated with 1/4 MIC polymyxin B, alone or in combination with 64 μg ml −1 resveratrol. According to the checkerboard assay results, the combination of 1/4 MIC polymyxin B and 64 μg ml −1 resveratrol, with which the FICI is <0.5, can be expected to provide synergistic effect on bactericidal activity. This was shown with all MDR P.aeruginosa isolates tested in this study (as shown in Fig. 1). When compared with the control group, the single-drug treatment group showed little or no time-dependent bactericidal activity, as bacterial load did not decrease even after 24 h treatment. While in comparison, for all tested MDR P.aeruginosa isolates, cells could be effectively eradicated by the combination treatment of 1/4 MIC polymyxin B and 64 μg ml −1 resveratrol within 2-8 h. For strain ATCC27853, though the combination of the chosen concentration of polymyxin B and resveratrol could not effectively eradicate bacterial cells, a decrease in cell number   Fig. 1). For all isolates, but not strain ATCC27853, in this study, more than 7log10 decrease in CFU/ml can be obtained with the combination treatment compared with the single treatment after 24 h, which supports the concept that resveratrol and polymyxin B have synergistic effect on bactericidal activity against MDR P.aeruginosa.

Combination effect of resveratrol and polymyxin B on inhibiting biofilm formation
Biofilm formation provides P.aeruginosa isolates with protection against antibiotics and could lead to antibiotic treatment failure. To investigate if resveratrol has combination effect with polymyxin B on inhibiting biofilm formation, three isolates (PA1, PA2, PA7) with high biofilm formation capability were selected. The subinhibitory concentration of resveratrol and polymyxin B, defined as the concentration that has less than 10% influence on bacteria growth compared with that of the untreated group, was employed for each group. Biofilm was cultivated with the bacterial suspension mixed with a subinhibitory concentration of resveratrol and polymyxin B alone or in combination and measured as the OD value of crystal violet adhered by established biofilm. At the same time, the influence of the chosen concentration of resveratrol and polymyxin B used alone or in combination on bacteria growth was monitored with the same bacteria suspension as used for biofilm formation. After incubated at 37°C for 48 h, significant decrease in OD value (P < 0.01), but no more than 10% decrease in cell number, can be observed in all three isolates for the combination group compared with that of the polymyxin B or resveratrol single treatment group (Fig. 2). This result demonstrates that resveratrol and polymyxin B have combination effect on inhibiting biofilm formation.

Resveratrol increases the bactericidal activity of polymyxin B against the cells protected by established biofilm
Since it has been reported that biofilm can protect bacterial cells from antibiotic pressure, we examined if resveratrol can increase the bactericidal activity of polymyxin B against cells with established biofilm [24,25]. In this experiment, biofilm was cultivated with the isolate PA2, which has the highest biofilm formation capability, and biofilm influence on the sensitivity of this isolate to polymyxin B was measured by comparing the growth of bacteria with or without established biofilm. Strictly following the operating procedures described in the methods section, we got approximately 1.5 × 10 8 CFU/ml of cells attached to the bottom of each well. After incubation with 8 μg/ml polymyxin B for 24 h, the total number of cells for the group with established biofilm reached approximately 2.3 × 10 9 CFU/ml, but just 1.7 × 10 5 CFU/ml for the group without established biofilm. About 4log10 increase in CFU/ml supports the concept that established biofilm leads to the decreased sensitivity of PA2 to polymyxin B. When bacterial cells with established biofilm were incubated with 128 μg ml −1 resveratrol and 32 μg ml −1 polymyxin B alone or in combination for 24 h, cell counts for the group treated with polymyxin B in combination with resveratrol were significantly smaller than that for the group treated with polymyxin B alone (P < 0.01), as well as for the group treated with the same concentration of resveratrol alone (P < 0.01). These results demonstrate that the bactericidal activity of polymyxin B against those cells protected by established biofilm can be increased by resveratrol (Fig. 3).

Discussion
In recent few years, with the high prevalence of MDR P.aeruginosa, and the limited number of antibacterial agents that can be available for MDR bacterial infection, the treatment of P.aeruginosa infection is challenging [26].
Polymyxin B is one of the rare antibiotic agents that can be used for the treatment of MDR Gram-negative bacterial infection. However, in recent few years, polymyxin B resistance has been frequently reported [27]. Though the polymyxin B resistance mechanism of P.aeruginosa is still not completely understand, a common appreciated one is the specific modification of the lipid A component of the outer membrane lipopolysaccharide (LPS) which is the initial target of polymyxin B [28]. It has been reported that some specific mutations in PhoP/PhoQ and PmrA/PmrB two-component system (TCS) can lead to the increased LPS modification, which in turn results in polymyxin B resistance in P.aeruginosa [29]. Another well-known mechanism is associated with the presence of the gene, mcr-1. Gene mcr-1 encodes a lipid A phosphoethanolamine transferase that can catalyze the modification of lipid A moiety on bacterial lipopolysaccharide (LPS), so the plasmid carrying mcr-1 gene can also cause polymyxin resistance in P.aeruginosa [30]. Mutations of PmrB, PmrA, and PhoQ can be detected in both polymyxin B resistant isolates, suggesting that the conclusion of this study is likely to be suitable for those types of P.aeruginosa. However, mcr-1 could not be found in either of those two isolates. The high risk of nephrotoxicity and increasing resistance has seriously restricted the clinical application of polymyxin B [10,31]. Undoubtedly, any drug that can improve the sensitivity of MDR P.aeruginosa to polymyxin B can greatly benefit patients. Given this consideration, resveratrol was selected as a potential candidate, since it is well tolerated and has been reported to have synergistic effect with different antibiotics against different bacteria species [32,33]. In this study, both checkerboard assay and time-kill assay have demonstrated that resveratrol and polymyxin B have synergistic effect against MDR P.aeruginosa, which indicates that, with the assistance of resveratrol, better bactericidal effect can be expected for polymyxin B therapy.
However, one limitation should be considered for clinical use of resveratrol is that, due to the relatively poor bioavailability, concentration of resveratrol as employed in this study is difficult to be maintained in bloodstream. Actually, according to the current available data, the highest bloodstream concentration of resveratrol that can be achieved is just about 534 ng ml −1 [34]. Considering that resveratrol at the concentration below 32 μg /ml shows no synergistic effect with polymyxin B against MDR P.aeruginosa, we infer that, with the purpose of providing synergistic effect with polymyxin B, the achievable bloodstream concentration of resveratrol should be increased with some new strategies, which deserves further study.
Even though the achievable concentration of resveratrol in bloodstream has been greatly restricted by its relatively low bioavailability, there is no such concern for topical use, and the high concentration of resveratrol can still be achieved with the methods such as incorporated in dressing for wound infection, made into aerosolized solution for respiratory infection, contained in sustained-released gel for focal body cavity infection, etc. Actually, the availability of some of those methods has been confirmed with many animal models or human trials [35][36][37]. The feasibility of topical use of resveratrol, incorporating with the synergistic effect with polymyxin B as confirmed in this study, suggests that polymyxin B used in combination with resveratrol may be a prospective therapeutic alternative for the treatment of topical MDR P.aeruginosa infection. While this strategy can be supported by the fact that ointment made with polymyxin B, neomycin and bacitracin has been effectively used in clinical practice for skin infections [38]. It is likely that a similar topical formulation developed with polymyxin B and resveratrol can be expected for the treatment of topical infections of MDR P.aeruginosa.
Biofilm can protect bacteria from antibiotic pressure, and even impedes phagocytosis [23,39]. So antibiofilm activity has been regarded as an important index for evaluating the antibacterial activity of antibiotics against those cells with high biofilm formation capability [40,41]. In this study, we have demonstrated that polymyxin B and resveratrol have combination effect on inhibiting the biofilm formation of P.aeruginosa. This result indicates that, when used in combination with resveratrol, a lower concentration of polymyxin B may be required for inhibiting biofilm formation.
One concern of the combination antibiofilm strategy is for the respiratory P.aeruginosa infection where bacterial cells are more likely to colonize and form biofilm and relative lower polymyxin B concentration can be achieved with routine administration. Systemic use of polymyxin B, incorporating with nebulization therapy of resveratrol, can be expected to getting better antibacterial efficacy against this kind of infection than polymyxin B monotherapy. This strategy is possible a breakthrough for the treatment of Fig. 3 a Growth of PA2 cells with (Biofilm bacteria group) or without (Planktonic bacteria group) established biofilm after exposing to 8 μg/mL polymyxin B for 24 h. b Combination effect of polymyxin B and resveratrol against the growth of PA2 cells with established biofilm. PB, 32 μg/mL polymyxin B; Res, 128 μg/ml resveratrol; PB + Res, 32 μg/mL polymyxin B + 128 μg/ml resveratrol. *There were significant differences for this group compared with each other group (P < 0.01) chronic respiratory infection caused by MDR P.aeruginosa, so long as the therapeutic concentration of resveratrol can be achieved in pulmonary airway and the related pulmonary toxicity can be tolerated, which deserves further research to determine if feasible.
Established biofilm is frequently observed with chronic wound P.aeruginosa infection and results in poor wound healing [42,43]. Since antibiotic resistance can be greatly increased by biofilm, bacterial cells, which may be initially deemed to be susceptible to polymyxin B by the official MIC breakpoint-based criterion, can still not be effectively eliminated by polymyxin B with the conventionally recommended does strategy, when the bacterial cells have been protected by established biofilm [44]. Even though the loading of polymyxin B for systemic administration has been greatly limited by the high risk of nephrotoxicity, there is no such restriction for topical use, and high concentration of polymyxin B can still be used topically for the treatment of bacterial infections [45,46]. In this study, polymyxin B of 32 μg ml −1 was employed together with resveratrol of 128 μg ml −1 against those bacterial cells protected by established biofilm, and the result has demonstrated that combination therapy has better anti-bacterial effect compared with the polymyxin B monotherapy. This result suggests that if polymyxin B is used in combination with resveratrol in the treatment of bacterial infection caused by MDR P.aeruginosa with established biofilm, a better clinical efficacy could be expected. This strategy can be helpful for those patients, such as with burns, pressure ulcers, and diabetic foot, that frequently suffer from the ineradicable MDR P.aeruginosa infection caused by established biofilm [43,47,48].

Conclusions
Though lacking intrinsic bactericidal activity, resveratrol may still be an ideal partner for combination therapy of polymyxin B against MDR P.aeruginosa, since it can be well tolerated, and greatly increases the sensitivity of MDR P.aeruginosa to polymyxin B. Besides, it can improve the antimicrobial property of polymyxin B against MDR P.aeruginosa from various aspects, such as inhibiting biofilm formation and killing the cells protected by established biofilm. It can be projected that resveratrol has the potential to increase the clinical efficacy of polymyxin B in the treatment of topical infections caused by MDR P.aeruginosa.