Research article
Yüklə 0,79 Mb. Pdf görüntüsü
|
|
combination of doxycycline (2mg/L) with polyamino-isoprenic compounds (10 μM) on twenty clinical isolates recently isolated from the HIA-Laveran (see M&M section and S1 Table ). Parent derivatives 1 and 2 showed only a slight efficacy, restoring doxycycline susceptibility in only 27% and 6% of the isolates, respectively. Compounds 4 restored the doxycycline suscep- tibility of 88% of the isolates ( Fig 2 ), whereas compound 3 led to a total susceptibility of all the considered isolates with a MIC ratio of 32 ( S1 Table
). We concluded that compound 3 was the best compound and suitable for a more detailed study. The same experiment was performed using chloramphenicol (4 mg/L) in combination with 3 (10 μM final concentration). Amongst the 20 clinical isolates, the growth of 15 (75%) was impaired, while 5 (25%) were resistant (data Fig 1. Structure of polyamino-isoprenic derivatives 1 –4 identified in this study. LogD have been determined by using chemical simulation software Marvin Sketch 5.11.3. doi:10.1371/journal.pone.0154490.g001 Table 2. WST1 cytotoxicity test of the compounds. Compound CHO
Fibroblast IC 50 ( μM)*
IC 50 ( μM)* 1 808 +/- 68.1 138 +/- 18.6 2 235.5 +/- 28.9 Nd ** 3 320.5 +/- 41.2 Nd ** 4 363 +/- 20.3 136 +/- 7.9 Doxorubicin 48.3 +/- 6.8 12.4 +/- 2.1 * mean +/- Standard deviations, data are obtained from 2 independent experiments ** Not determined. doi:10.1371/journal.pone.0154490.t002 Inhibitors of Antibiotic Resistance in Pseudomonas PLOS ONE | DOI:10.1371/journal.pone.0154490 May 6, 2016 6 / 16
not shown). These experiments were performed to study the effect of our compounds on the natural resistance of P. aeruginosa. In addition, we were wondering if compound 3 could also increase susceptibility to other antibiotic families including ß-Lactams, fluoroquinolones and aminoglycosides, that are representative of the main classes of anti-pseudomonal agents. Experiments were performed with the ß-lactams ceftazidime and ticarcillin and ciprofloxacin and amikacin. Results are summarized in S3 Table . We found that, neither compound 3 nor PAßN (used as a control) showed significant synergy with ceftazidime on all the strains tested. A slight effect however, was observed in the combination of 3 with ticarcillin on all the strains. No synergistic effect of 3 with ciprofloxacin was observed with the exception of PA0-7H (that overproduces MexEF) where a 4 times decrease of MIC was measured. Conversely, a surprising increase of the MIC of amikacin was caused by 3 on every strains, this could be attributed to Fig 2. Susceptibility of clinical isolates of P. aeruginosa to treatment with doxycycline in combination with various polyamino-isoprenic derivatives. The percentage of susceptible isolates to doxycycline (2 mg/L) supplemented with one of the compounds (10 μM) was determined for each combination. grey bar, black bar, dark grey bar, and white filled bar, correspond to compounds 1, 2, 3 and 4 respectively. Error bars represent standard deviations of 3 independent experiments. doi:10.1371/journal.pone.0154490.g002 Inhibitors of Antibiotic Resistance in Pseudomonas PLOS ONE | DOI:10.1371/journal.pone.0154490 May 6, 2016 7 / 16
the deleterious effect of 3 on the proton gradient (see below), indeed the influx of aminoglyco- sides into bacteria is depending on the proton gradient through the inner membrane [ 28 ]. To
better characterize compound 3, we first searched for the minimum concentration of com- pound that allowed killing of bacteria by doxycycline, and PAßN was used as the reference molecule. Determination of the minimum efficient concentration of compound 3. PA βN is the most characterized efflux pump inhibitor (EPI) in P. aeruginosa. Initially considered a promis- ing alternative to prevent antibiotic resistance, its use as a prospective drug was discontinued because of its toxicity [ 29 ]. Nevertheless, it remains a useful tool for studying resistance of Gram-negative bacteria. The efficiency of compound 3 was further compared with that of PAßN.
First, we performed a checkerboard assay to analyze the synergy between doxycycline and the polyamino-isoprenic compound and between doxycycline and PA βN. For both, compound 3 and PA βN we observed a synergistic inhibition of growth when combined with doxycycline with a FIC Index of 0.09 ( S2 Table ). Nevertheless, to more precisely refine these results, we fixed the concentration of interest of doxycycline at 2 mg/L (see above) and tested a gradient of concentrations for PAßN and 3. PAßN is generally used at a concentration greater than 10 μM to inhibit the resistance of P. aeruginosa with respect to various antibiotics [ 30 ], and, to our knowledge, no data have been reported in the literature on its use in combination with doxycy- cline. The percentage of growth inhibition was monitored as a function of compound concen- tration, and the results are presented in Fig 3
. Under these conditions, it clearly appears that at least a 30 μM (15.6 mg/L) concentration of PAßN is necessary to efficiently inhibit the growth of P. aeruginosa in the presence of sub-inhibitory concentration of doxycycline, which is in agreement with the efficacy of PAßN to restore fluoroquinolone susceptibility of Pseudomonas [ 30 ]. Conversely, a concentration of approximately 3 μM (1.22 mg/L) of 3 was sufficient to sig- nificantly inhibit PA01 growth (Graphically determined on Fig 3
) and confirmed with the checkerboard assay ( S2 Table ). According to these data, 3 appears to be a stronger adjuvant of doxycycline than does PAßN in vitro. This first set of assays was performed on the reference strain PA01, which exhibits only a natural resistance towards doxycycline. Action of compound 3 on efflux pumps We hypothesized that if compound 3 targets the intrinsic resistance, it should also be effective on strains with acquired resistance, such as through the efflux mechanisms which are well known to contribute to the MDR phenotype [ 31 ]. The strains overproducing the major efflux pumps of P. aeruginosa used in this experiment are listed in Table 1
. All of the strains grew well on the agar medium containing 2 μg/mL of doxycycline ( Fig 4
). In contrast, in the pres- ence of both, doxycycline and 3, PA01 or strains overproducing either MexCD or MexEF cells growth was observed when a dilution of 10 −2 was plated on MH. Cells overproducing either MexAB or MexXY grew at a higher dilution (10 −4 ). Nevertheless, in all cases, no growth was observed above a dilution of 10 −5 . Moreover, the cells from the ΔmexB strain did not grow at any of the dilutions, in accordance with the contribution of MexAB-OprM to cycline resistance [ 15
susceptibility of Pseudomonas to other antibiotics. We thus decided to test chloramphenicol because P. aeruginosa also exhibit a high level of intrinsic resistance to this drug ( Table 1 ). Sim-
ilar results were obtained at least for strains PA01 and CMZ091 and to a lesser extent for PA01-ERY
r . For PT629 and PA0-7H a slight effect was also observed, ( Fig 4b ). These data sug- gest that the selected compound was able to efficiently affect resistance to different classes of antibiotic. Inhibitors of Antibiotic Resistance in Pseudomonas PLOS ONE | DOI:10.1371/journal.pone.0154490 May 6, 2016 8 / 16
Taken together, these results demonstrate that the combination of doxycycline and 3 increases susceptibility of all the strains of P. aeruginosa overproducing the major efflux pumps, suggesting a possible inhibition of this resistance mechanism. According to these data and to the hyper-susceptibility of the mexB deficient mutant observed in Fig 4 , we hypothesized that 3 could inhibit the efflux pumps of P. aeruginosa. However, at this time, we could not exclude the possibility that increased permeability might also explain the results obtained. To better understand whether these compounds were actually able to inhibit efflux, we developed a real-time efflux assay for Pseudomonas that was previously described for Entero- bacteriaceae [ 17 ]. As shown in Fig 5A , the strain PA01 was able to efficiently expel 75% of the preloaded dye, immediately after glucose addition, thus demonstrating a strong activity of dye efflux in this strain. When compound 3 was added before glucose addition, a concentration- dependent inhibition of efflux was observed, and a concentration of 125 μM of 3 completely abolished the transportation of dye ( Fig 5
). The same experiments were performed with all the four compounds, in the same range of concentrations, and the results are summarized in Fig 5B
pound tested, including PAßN, efflux is inhibited in a dose-dependent manner. One can men- tion that compound 3 is the most efficient, which is consistent with the data obtained in the susceptibility assays described in Fig 2
. Recent studies have shown that the mode of action of PAßN includes not only efflux inhibi- tion but also outer membrane destabilization [ 32 , 33 ]. Considering that our compounds are polyamine derivatives and that polyamines are known to disrupt membranes [ 34 ], one cannot exclude that the inhibition of glucose-triggered efflux of the dye described in Fig 5
could result, at least in part, from the disruption of the outer membrane of P. aeruginosa and consequently Fig 3. Comparison of the synergistic activity of PAßN and compound 3 with doxycycline. Bacterial growth corresponding to each concentration of compound was calculated as a percentage of bacterial growth under the same conditions without the addition of compound. Black filled circles, PAßN; open squares, Compound 3. Error bars represent standard deviations of 3 independent experiments. doi:10.1371/journal.pone.0154490.g003 Inhibitors of Antibiotic Resistance in Pseudomonas PLOS ONE | DOI:10.1371/journal.pone.0154490 May 6, 2016 9 / 16
Fig 4. Compound 3 activity against P. aeruginosa strains over-expressing efflux pumps. The ability of bacteria to grow under various conditions was assessed in solid media containing, in a: 2 mg/L doxycycline (left) or a combination of 2 mg/L doxycycline and 10 μM compound 3 (right) and in b: 4 mg/L chloramphenicol (left) or a combination of 4 mg/L chloramphenicol and 10 μM compound 3 (right). 5 μL of logarithmic dilutions of an overnight culture of each strain were spotted on the plates. From top to bottom, PA01, PT629, PA01-ERY R , PA0-7H, CMZ091 and FB1. doi:10.1371/journal.pone.0154490.g004 Inhibitors of Antibiotic Resistance in Pseudomonas PLOS ONE | DOI:10.1371/journal.pone.0154490 May 6, 2016 10 / 16
from the destabilization of the interactions between the partners of the RND complex that were shown necessary to ensure transport in vitro [ 35 ]. To evaluate the potential ability of 3 to permeabilize the outer membrane, we monitored the hydrolysis rate of the chromogenic β-lac-
tam, nitrocefin, by the cells of P. aeruginosa PA01. The polymyxin-B derivative PMBN was previously described as having a strong permeation activity on the outer membrane of Gram- negative bacteria [ 24 ]. Therefore, we compared 3 with PMBN in permeation and efflux inhibi- tion assays. We observed ( Fig 6
), that the permeation efficacy of 3 at 15.6 μM was significantly Fig 5. Inhibition of the efflux of the 1,2 ’ dinaphthylamine dye by the compound 3. Efflux was triggered after 200 s by the addition of 50 mM glucose (arrow). The intensity of fluorescence emission for 1,2 ’ dinaphthylamine is given in relative fluorescence units (RFU). (a) Concentration-dependent inhibition of 1,2 ’ dinaphthylamine efflux by compound 3 after glucose addition (arrow). Compound concentration ( μM) is indicated in front of each green curve; as a control, an experiment without addition of compound and glucose (-glc, red curve) was also included; (b) percent of efflux inhibition obtained for each compound and PAßN as indicated. Error bars represent standard deviations of 2 independent experiments. * indicates statistically significant difference in values (p < 0.05) compared to PAßN. doi:10.1371/journal.pone.0154490.g005 Inhibitors of Antibiotic Resistance in Pseudomonas PLOS ONE | DOI:10.1371/journal.pone.0154490 May 6, 2016 11 / 16
identical to 31.25 μM of PMBN (as indicated with bracket on Fig 6 ; p value 0.05). At the same concentrations, 3 already showed a fifty percent efflux inhibition, while PMBN was quite inefficient. The same analysis can be drawn for the respective concentrations of 31.25 μM of
compound 3 and 125 μM of PMBN. In this case one can observe a 10 percent permeation activ- ity for both, and a 60 percent efflux inhibition for 3 compared to 30 percent for PMBN ( Fig 6
). In addition, we investigated the activity of 3 on PA01 using a fluorescence assay that monitors cytoplasmic membrane depolarization [ 36 ]. The compound 3 dissipated the proton motive force ( S3 Fig
). This result may explain the antagonism of 3 with the aminoglycoside antibiotic amikacin ( S3 Table ), which requires the component of the transmembrane electrochemical gradient for cell penetration [ 28 ]. Discussion Here, we describe the study of new chemosensitizers that are able to increase the susceptibility to doxycycline of intrinsically resistant P. aeruginosa. P. aeruginosa is one of the most prevalent human pathogens and is known to be difficult to eradicate using the current therapeutic strate- gies. This is due to its natural antibiotic resistance and its strong ability to acquire additional mechanisms of resistance. In the present study, we aimed to identify a combination drug and new molecule that could be effective on both wild type and clinical strains. We used doxycy- cline as the antibiotic of choice because it is not used as a human therapeutic to treat infections with P. aeruginosa due to its low susceptibility. A preliminary study had previously demon- strated that polyamino-geranic compounds, in combination with doxycycline, were able to restore susceptibility of Enterobacteriaceae to antibiotics [ 17 ]. Fig 6. Membrane permeation assays. The effect on the PAO1 strain of compound 3 (black-filled and black-hatched bars) and PMBN (grey-filled and grey-hatched bars) were compared at concentrations ranging from 15.6 μM to 125 μM, as indicated. The relative effect corresponds either to efflux inhibition (filled-bars) or to outer-membrane permeation (hatched-bars). Error bars represent standard deviations of 2 independent experiments. * indicates statistically significant (p 0.05) differences of relative effect (efflux inhibition/outer membrane permeation) for compound 3 treatment compared to the respective PMBN treatment. Brackets indicate statistically comparable permeation effects. doi:10.1371/journal.pone.0154490.g006 Inhibitors of Antibiotic Resistance in Pseudomonas PLOS ONE | DOI:10.1371/journal.pone.0154490 May 6, 2016 12 / 16
A prerequisite screening of original molecules was performed in a high-content procedure on the reference strain PA01 using a 10 μM concentration of our bioactive compounds [ 36 ] and a concentration of 4 mg/L of doxycycline. Although several compounds showed activity by decreasing bacterial growth in combination with doxycycline, we focused our studies on com- pound 3, which completely abolished bacterial growth. Using the checkerboard assay, the FIC index of the compound 3-doxycycline combination was determined to be 0.09, demonstrating a strong synergy between the molecules. Additionally, we observed that derivative 3 strongly decreased the resistance of MDR strains to both doxycycline and chloramphenicol, which belong to different antibiotic families. This observation suggests that this polyamino-isoprenic molecule may directly impair the natural resistance mechanisms of P. aeruginosa. This com- pound is a polyamino-farnesyl molecule and possesses a tri-isoprenyl group that could help the molecule pass through the outer membrane of P. aeruginosa, which is known to be strongly impermeable due to the presence of a highly hydrophobic lipid bilayer, in addition to a posi- tively charged spermine moiety that could interact with the negatively charged outer mem- brane of the bacteria. In comparison with PAßN, the most studied chemosensitizer of P. aeruginosa toward antibiotics [ 30 ], compound 3 appeared to be approximately 10 times more potent ( Fig 2
). Interestingly, the same type of assay performed against Enterobacteriaceae iden- tified not derivative 3 but compound 2 as the most potent (data not shown). These two com- pounds differ in both the size of the isoprenyl moiety and the structure of the involved amino group (
Fig 1 ). Taking into consideration the LogD parameter, which reflects the true behavior and bioavailability of an ionizable compound in a solution at a given pH, a significant correla- tion with anti-pseudomonal activity was observed. Indeed, derivative 3 displaying the highest LogD values (e.g., the most hydrophobic derivative) also possesses higher activity than com- pounds 1, 2, 4. Moreover, it clearly appears ( S2 Fig ) that derivative 3 is ionized as tri or tetra protonated species at a physiological pH enhancing its ability to interact with the negative charges present at the surface of the outer membrane of bacteria. When tested in association with doxycycline against a series of clinical isolates, the high potency of derivative 3 was confirmed, with all of the clinical isolates being killed. To determine the cytotoxicity of these compounds, they were tested on Chinese hamster ovary (CHO) cells and exhibited an IC50 of over 100 μM, which is at least 10 times the concen- tration used in this study ( Table 2 ). However, we also observed a higher toxicity for human fibroblasts of compounds 1 and 4. Consequently, we cannot exclude the same increase for com- pounds 2 and 3. In addition, in all of the assays performed, we observed that the compounds are less cytotoxic than the control molecule doxorubicin. To better understand the mode of action of compounds, we performed real-time efflux assays. We observed a dose response effect for each compound, with compound 3 being the most efficient at inhibiting the 1,2 ’-DNA efflux (Figs 5 and 6 ). As described in Fig 5 , a relatively high concentration of 3 (31.25 μM) is needed to inhibit 50% of 1,2’DNA efflux. This might result from the setting up of a real-time assay that requires a high number of bacteria (2.5x10 8 cfu/mL) compared to MIC determination (5x10 5 cfu/mL) combined to an immediate observ- able effect. This compound was further assayed for its ability to permeabilize the outer mem- brane of P. aeruginosa. Compared to PMBN, compound 3 partially destabilizes the outer membrane, and strongly inhibits efflux pumps, ( Fig 6
). Taken together, these data strongly sug- gested that although 3 destabilizes the outer membrane, this effect cannot completely explain the efflux inhibition observed. In addition, we observed that 3 induced a significant depolariza- tion of the inner membrane of PA01 ( S3 Table ). Taken together, according to the proton gradi- ent dependence of the RND-efflux pumps, these data suggest that 3 inhibits efflux in P. aeruginosa by depleting the energy of the pumps. Inhibitors of Antibiotic Resistance in Pseudomonas PLOS ONE | DOI:10.1371/journal.pone.0154490 May 6, 2016 13 / 16
Nowadays, there is no doubt that the emergence and the dissemination of bacterial resis- tance towards a drug is directly related to its rate of use [ 37 ]. The continuous development of drug resistance mechanisms linked to drug usage is also exemplified by the observed decrease in chloramphenicol resistance in countries where its use is strongly controlled [ 38 ]. Here, we have determined that an association between doxycycline and a polyamino-isoprenic com- pound killed P. aeruginosa reference strain, efflux overproducers and clinical isolates. To our knowledge, it is the first report of such an activity and may open a new field of investigation, associating pharmaco-chemistry and microbiology to develop molecules with better efficiency on bacteria and decreased toxicity. Since it is necessary to continuously develop alternate solu- Yüklə 0,79 Mb. Dostları ilə paylaş:
|