ABSTRACT
Staphylococcus aureus is responsible for difficult-to-treat bone and joint infections (BJIs). This is related to its ability to form biofilm and to be internalized and persist inside osteoblasts. Recently, bacteriophage therapy has emerged as a promising option to improve treatment of such infections, but data on its activity against the specific bacterial lifestyles presented above remain scarce. We evaluated the activity of a combination of three bacteriophages, recently used for compassionate treatment in France, against S. aureus HG001 in a model of staphylococcal biofilm and a model of osteoblasts infection, alone or in association with vancomycin or rifampin. The activity of bacteriophages against biofilm-embedded S. aureus was dose dependent. In addition, synergistic effects were observed when bacteriophages were combined with antibiotics used at the lowest concentrations. Phage penetration into osteoblasts was observed only when the cells were infected, suggesting a S. aureus-dependent Trojan horse mechanism for internalization. The intracellular bacterial count of bacteria in infected osteoblasts treated with bacteriophages as well as with vancomycin was significantly higher than in cells treated with lysostaphin, used as a control condition, owing to the absence of intracellular activity and the rapid killing of bacteria released after the death of infected cells. These results suggest that bacteriophages are both inactive in the intracellular compartment after being internalized in infected osteoblasts and present a delayed killing effect on bacteria released after cell lysis into the extracellular compartment, which avoids preventing them from infecting other osteoblasts. The combination of bacteriophages tested was highly active against S. aureus embedded in biofilm but showed no activity against intracellular bacteria in the cell model used.
INTRODUCTION
Staphylococcus aureus is the first causative agent of bone and joint infections (BJIs) and is responsible for particularly difficult to treat infections related to its wide panel of virulence factors, among which some specifically promote a chronic course of infection (1–3). For instance, the formation of biofilm plays a key role in the pathophysiology of chronic staphylococcal BJIs by limiting the activity of antibiotics and immune response (4, 5). In addition, S. aureus is able to invade and persist within bone cells, in particular osteoblasts (6). The poor diffusion in osteoblastic cells and the low activity of some antibiotics currently used to treat BJIs on internalized bacteria are likely to contribute to bacterial persistence (7). Thus, despite an appropriate antibiotic treatment and surgical procedure, there is a high rate of infection relapse underlining the need for new therapeutic strategies. In this context, phage therapy has recently reemerged as a promising option to improve the treatment of chronic BJIs (8).
Bacteriophages are viruses that specifically infect bacteria. Discovered in 1917 by Felix d’Hérelle, they were considered in the following years as an interesting antimicrobial therapy (9). However, their use was quickly abandoned in Western Europe following the discovery of antibiotics, while, despite the absence of large and controlled clinical trials, they remained a popular treatment during the 20th century in Eastern Europe among others for patients with BJIs (10–12). Recently, 3 studies using animal models have highlighted the potential of bacteriophages in combination with antibiotics to prevent and treat S. aureus BJIs (13–15). In addition, with the approval of the French National Agency for Medicines and Health Products Safety (Agence nationale de sécurité du médicament et des produits de santé [ANSM], Paris, France), bacteriophages have also been used successfully in humans as a salvage therapy for severe periprosthetic joint infections (16, 17). This efficacy is supported by very few in vitro studies that have demonstrated the particular activity of bacteriophages against biofilm (18–20). The production of enzymes by bacteriophages that are able to alter the biofilm matrix favors their activity as well as those of the antibiotics used in association with bacteriophages (21). Conversely, the activity of bacteriophages against the second reservoir, formed by the bacteria internalized in osteoblasts, has yet to be explored.
In the present study, we evaluated the activity of a combination of three bacteriophages, previously used in association with antibiotics in a clinical case of chronic infection reported by Ferry et al. (17), against an S. aureus reference strain. To this end, we tested the bacteriophages alone and in association with antibiotics classically used to treat BJIs (vancomycin and rifampin) against S. aureus in a model of biofilm and in a model of intracellular infection of human osteoblasts.
RESULTS
Bacteriophages are highly active on planktonic S. aureus HG001.The efficiency of plating (EOP) score of the combination of the three bacteriophages against planktonic S. aureus HG001 was 0.22, indicating that the bacteriophages were strong killers (EOP ≥ 0.1). In addition, bacteriophages also had a bactericidal activity against S. aureus HG001 in liquid medium at both tested multiplicity of infection (MOI) values (Fig. 1).
S. aureus HG001 growth inhibition in liquid medium using a combination of three bacteriophages. The growth inhibition of a bacterial inoculum of 107 CFU/ml was assessed in the presence of bacteriophages at a multiplicity of infection (MOI) of 1 or 10 bacteriophages per bacteria. Error bars represent the standard deviation of the data from three independent experiments performed in triplicates.
Bacteriophages are active on biofilm-embedded S. aureus HG001 in a dose-dependent manner and can potentiate the activity of antibiotics.S. aureus HG001 biofilm was exposed for 24 h to the combination of the three bacteriophages at different concentrations (from 105 to 108 PFU/ml). The antibacterial activity of bacteriophages was dose dependent (Fig. 2): the greatest reduction in viable bacteria count was observed for the highest concentrations of bacteriophages (107 and 108 PFU/ml) with a reduction of 3.1 and 3.6 log, respectively, compared to that of the untreated condition (P < 0.001). Exposure to lower concentrations induced a moderate bactericidal effect at 106 PFU/ml (reduction of 1.4 log; P < 0.001) or no significant effect at 105 PFU/ml.
Viable bacteria counts after 24-h exposure of 24-h-old S. aureus HG001 biofilms to various bacteriophage concentrations associated or not with antibiotics (vancomycin [A] or rifampin [B]). Antibiotics were used at the following concentrations: 1.5 mg/liter (MIC), 6 mg/liter (Cbone = 4× MIC), and 15 mg/liter (10× MIC) for vancomycin and 0.016 mg/liter (MIC), 0.16 mg/liter (10× MIC), and 6 mg/liter (Cbone = 375× MIC) for rifampin. For monotherapy conditions, significant variations of the viable bacteria counts after 24 h of treatment compared to those of the untreated condition are denoted (Mann-Whitney test; **, P < 0.01; ***, P < 0.001), and for bitherapy conditions, significant (2-way ANOVA; P < 0.05) synergistic interactions (S) between bacteriophages and antibiotics are indicated. Error bars represent the standard deviation of the data from three independent experiments performed in triplicates.
The association of bacteriophages at the concentration of 105 PFU/ml with vancomycin used at the usual bone concentration (6 mg/liter) or at 10× MIC (15 mg/liter) was synergistic (Fig. 2A). In addition, a synergistic effect between rifampin and bacteriophages was observed when they were used at the lowest concentrations (105 PFU/ml and MIC of 0.016 mg/liter) (Fig. 2B). For higher rifampin concentrations, the activity of rifampin alone was too high to demonstrate any synergy with bacteriophages.
Bacteriophages can be internalized into infected MG63 osteoblastic cells but are not active against intracellular S. aureus HG001.In order to specifically investigate the capacity of bacteriophages to control the intracellular S. aureus infection, MG63 osteoblastic cells were first infected with HG001 and then exposed to lysostaphin to kill residual extracellular S. aureus before being treated with bacteriophages in the presence of lysostaphin throughout the incubation of cells (lysostaphin protection assay). There was no significant reduction of the intracellular count of bacteria after 24 h of treatment with bacteriophages at 107 or 109 PFU/ml compared to that of the same cells treated only with lysostaphin; no bacteriophages were detected in lysates of infected cells treated by bacteriophages.
As previously suggested by Kaur et al., bacteriophage penetration into the intracellular compartment may be facilitated by bacteria (22). Thus, we assumed that lysostaphin could interfere with the internalization of bacteriophages in osteoblasts by killing bacteria released upon host cell lysis, avoiding reinfection of new host cells by these released bacteria. In this context, for subsequent experiments, lysostaphin was used only for 1 h after the bacterial internalization step. Then, the infected cells were exposed for 24 h to bacteriophages and/or antibiotics. Using this protocol, as expected, the intracellular count of bacteria was significantly lower after 24 h of exposure to rifampin than to lysostaphin, which is active only on extracellular bacteria (P < 0.0001). Using the same protocol, exposure of infected cells to bacteriophages alone for 24 h led to a significantly lower intracellular count of bacteria than that in the untreated condition (P < 0.0001). However, the residual intracellular count of bacteria was significantly greater than that recovered in cells treated for 24 h with lysostaphin alone (P < 0.0001), suggesting both that bacteriophages had no intracellular activity and that their extracellular activity was too slow to prevent reinfection of osteoblasts by bacteria released after cell lysis. Of note, the inoculum found in cells treated with bacteriophages at 109 PFU/ml was significantly lower than that found in cells exposed to bacteriophages at 107 PFU/ml (P = 0.04), suggesting that less bacteria were reinternalized when the highest concentration of bacteriophages was used and that the control of the extracellular bacteria by bacteriophages was concentration dependent. In addition, activities of bacteriophages and vancomycin were not significantly different. Of note, the intracellular count of bacteria recovered in cells treated with the association of vancomycin and bacteriophages was significantly lower than that in cells treated with vancomycin or bacteriophages alone (P < 0.001) and not significantly different from that in cells treated with lysostaphin. This result suggests that the combination of vancomycin and bacteriophages allowed a better control of released bacteria by infected cells in the extracellular media but that both agents had no intracellular activity as does lysostaphin. The number of bacteria recovered in cells treated with rifampin alone was not significantly different than that of cells treated with rifampin and bacteriophages (P = 0.49) (Fig. 3).
S. aureus HG001 intraosteoblastic amount of bacteria after 24 h of treatment by bacteriophages and/or antibiotics. Antibiotics were used at their usual bone concentration (6 mg/liter for both antibiotics), and lysostaphin was used at 10 mg/liter. The activities of the different treatments against internalized S. aureus were compared using the Mann-Whitney U test (*, P < 0.05; ***, P < 0.001). Error bars represent the standard deviation of the data from three independent experiments performed in triplicates.
In order to investigate the replication of bacteriophages in this model and their penetration inside osteoblasts, titrations were performed in cellular supernatants and in noninfected or infected MG63 lysates after 24 h of incubation with bacteriophages combined or not with antibiotics. The ratios between the bacteriophage titer after 24 h and the initial titer were higher with infected cells than with noninfected cells (P = 0.0006 and 0.02 for bacteriophages used at 107 and 109 PFU/ml, respectively), likely in relation to the replication of bacteriophages in bacteria released after the lysis of the infected cells. This difference was not observed when bacteriophages were combined with rifampin, contrary to vancomycin, which is likely to be due to its intense bactericidal activity (Fig. 4A). After 24 h, a significantly greater number of bacteriophages was recovered in the lysates of infected cells (7.9 × 105 PFU/ml when treated with a concentration of 107 PFU/ml and 3.6 × 106 PFU/ml when treated with a concentration of 109 PFU/ml) than in lysates of noninfected cells (P < 0.001). The titers of bacteriophages in infected osteoblastic cells were significantly lower when combined with antibiotics, suggesting that the penetration occurs when bacteria released in the extracellular medium reinfect cells (Fig. 4B). Finally, using electron microscopy, we have been able to observe numerous bacteriophages around bacteria inside infected osteoblasts exposed to bacteriophages (Fig. 5).
Bacteriophage titrations in the extracellular and intracellular compartments. (A) The variation of the bacteriophage titers in the osteoblastic culture supernatants of noninfected and infected cells was assessed using the ratio between the bacteriophage titer after 24 h and the initial bacteriophage titer. Significant differences compared to the control condition (same concentration of bacteriophage applied to noninfected cells) are indicated as follows: *, P < 0.05; ***, P < 0.001. (B) Intracellular bacteriophage titers. Significant differences of the intracellular titer compared to that of the control condition (same concentration of bacteriophages applied to noninfected cells) are denoted as follows: **, P < 0.01; ***, P < 0.001. Antibiotics were used at their usual bone concentration (6 mg/liter for both antibiotics). Error bars represent the standard deviation of the data from three independent experiments performed in triplicates.
Electron microscopy observations of MG63 cells infected by S. aureus HG001 and treated during 24 h with bacteriophage. (A) Infected MG63 cell; S. aureus in a vacuole are indicated by arrows. (B) Enlarged view of panel A; bacteriophages surrounding S. aureus can be seen inside cells.
DISCUSSION
In the present study, we tested various concentrations of bacteriophages and compared their activity when used alone or in association with rifampin and vancomycin, antibiotics currently used to treat BJIs. We chose these antibiotics as comparison tools for the evaluation of bacteriophage activity, as rifampin is of particular interest for BJI treatment due to its well-known intracellular and antibiofilm activity, whereas vancomycin is conversely considered to be poorly active on these bacterial reservoirs (7, 23). The results obtained showed that the activity of bacteriophages against biofilm-embedded staphylococci was concentration dependent and that they were strongly active from 107 PFU/ml. The combination of bacteriophages with antibiotics used at low concentrations was synergistic, while both agents had no or only low bactericidal activity at these concentrations when used alone. This synergistic effect on mature biofilm is noteworthy in the context of BJI treatment. Because of the poor diffusion of antibiotics, their concentration in some locations of the bone can be low, and the association with bacteriophages in this context may contribute to increasing their activity. The data presented herein corroborate those of previous studies showing that bacteriophages can enhance the activity of antibiotics (18–20, 24).
We also investigated the bactericidal activity of the combination of the three bacteriophages against S. aureus HG001 after internalization in nonprofessional phagocytic cells, namely, osteoblasts. Bacteriophage titrations in cellular lysates, as well as electron microscopy observations, suggested that they were able to get inside cells and that their internalization in cells was not passive and was likely to be achieved using S. aureus as a Trojan horse. Nevertheless, our results show that after getting access to the intracellular compartment, bacteriophages were inactive on intracellular S. aureus. Only two other studies have evaluated the intracellular activity of bacteriophages. They provided contradictory results. Zhang et al. claimed that the anti-S. aureus vB_SauM_JS25 bacteriophage was able to penetrate into mammary epithelial cells and was able to reduce the intracellular inoculum (25). However, the authors did not compare the activity of this bacteriophage to that of an antimicrobial agent only active against extracellular bacteria, such as lysostaphin. Thus, it is difficult to affirm that the decrease of the intracellular inoculum after exposure to bacteriophages compared to that of the untreated condition was due to the intracellular activity of bacteriophages and not to the natural clearance of the intracellular bacteria in such cells. Conversely, Kaur et al. reported that the bacteriophage MR-5 did not impact the natural intracellular killing of S. aureus by macrophages (22). Interestingly, the authors indicated that the preadsorption of the bacteriophage onto viable bacteria before applying them to infected macrophages allowed a reduction of the intracellular bacterial inoculum. The authors hypothesized, as we do from the data presented herein, that bacteriophages could be shuttled inside macrophages thanks to bacteria.
The absence of intracellular activity of bacteriophages in the cellular model used herein, despite their intracellular location, could be explained by two phenomena. First, as the bactericidal effect of bacteriophages requires bacterial multiplication, their intracellular activity is possibly influenced by the bacterial dormancy inside osteoblasts. There is currently no published data describing the metabolic activity of S. aureus inside osteoblasts, which likely varies from one strain to another. Furthermore, Rollin et al. showed, using an endothelial cell model (another type of nonprofessional phagocytic cell), that the intracellular fate of S. aureus varied from one endothelial cell to another; ranging from rapid bacterial proliferation leading to cell lysis to prolonged bacterial dormancy (26). Moreover, the low intracellular pH may also affect the activity of bacteriophages (27). Dupieux et al. recently reported the influence of acid pH on the intracellular activity of some antibiotics in osteoblasts, namely, daptomycin and oxacillin (28). The concomitant use of alkalinizing agents (such as hydroxychloroquine, used to increase the activity of antibiotics during the treatment of infections caused by an intracellular bacteria named Coxiella burnetii) could be an interesting way to restore the intracellular activity of bacteriophages (29).
We acknowledge some limitations of our study. We only tested the activity of bacteriophages against one reference S. aureus strain. Further work should be performed including more strains to assess if their activity could be influenced by the biofilm matrix composition or the intracellular replication of S. aureus, which may vary from one strain to another. Moreover, it could be interesting to evaluate if synergistic effects could be obtained by combining bacteriophages with other antibiotics commonly prescribed to treat S. aureus BJIs, such as oxacillin, macrolides, or fluoroquinolones. Finally, we were not able to use a standardized definition of “synergy,” as these methods were developed for the antibiotic susceptibility testing of planktonic bacteria (30).
In conclusion, the combination of the three bacteriophages tested in the present study was highly active against S. aureus in mature biofilm but had no activity against bacteria internalized in osteoblasts. A synergistic effect was highlighted when associating bacteriophages with antibiotics, suggesting that they could be an interesting option as an adjuvant therapy for the treatment of S. aureus BJIs. Further studies using animal models of BJI and well-conducted clinical trials are needed to further evaluate phage therapy and its positioning in the management of these infections.
MATERIALS AND METHODS
Bacterial strains and bacteriophages.The methicillin-susceptible Staphylococcus aureus strain HG001 is a reference strain (31) previously used by Valour et al. (7) to test the activity of antibiotics against intracellular staphylococci and was used in the present study to assess the bactericidal activity of bacteriophages alone or in association with antibiotics. S. aureus SH50-2 is a clinical strain (isolated in France from a bone and joint infection) and was used in the present study for bacteriophage amplification and titrations.
Before each experiment, all strains were grown on Columbia blood agar (COS; bioMérieux, Marcy-l’Etoile, France) for 24 h at 36°C. Three colonies were then used to inoculate brain heart infusion (BHI) medium (bioMérieux) before incubation overnight. The HG001 strain had MICs (determined by microdilution according to EUCAST guidelines) of 1.5 mg/liter for vancomycin and 0.016 mg/liter for rifampin.
Three bacteriophages, PP1493, PP1815, and PP1957, produced and purified by Pherecydes Pharma (Romainville, France), were used and mixed extemporaneously before each experiment in equal proportions and at the indicated concentration. When needed, bacteriophage titration was performed using a spot test assay. For this, 10-fold serial dilutions (from 10−1 to 10−8) of bacteriophage suspensions in lysogeny broth (LB Lennox; Fisher Scientific, Hampton, NH, USA) were prepared, and 5 μl of these dilutions was spotted on an agar prepared immediately before use by mixing 25 ml of LB 0.75% agar medium and 2 ml of an exponential S. aureus SH50-2 culture. After 24 h of incubation at 36°C, the number of PFU was visually determined.
S. aureus HG001 susceptibility to bacteriophages.S. aureus HG001 susceptibility to the combination of the three bacteriophages was evaluated using two complementary methods. First, the efficiency of plating (EOP) score was calculated. The latter is the ratio between the bacteriophage titer obtained with the HG001 strain divided by the titer obtained with the SH50-2 reference strain used for the production of bacteriophages; the closer this score is to 1, the more efficient the bacteriophages are (strong killers, EOP > 0.1; intermediate killers, 0.005 < EOP < 0.09; weak killers, EOP < 0.005) (32). Second, the growth inhibition of S. aureus by bacteriophages in liquid medium was assessed by culturing bacteria in a 96-well plate at a starting concentration of 107 CFU/ml with or without bacteriophages at two concentrations (multiplicity of infection [MOI] of 1 or 10 bacteriophages per bacteria). The bacterial growth at 36°C was monitored by measuring the optical density at 600 nm (OD600).
Assessment of bacteriophages bactericidal activity against biofilm-embedded S. aureus HG001.An overnight culture of S. aureus HG001 in liquid BHI was adjusted to an OD600 of 1.0 and was then diluted 100-fold in BHI (Becton, Dickinson GmbH, Heidelberg, Germany). Then, a 96-well polystyrene plate (Greiner Bio-One, Kremsmünster, Austria) was inoculated with 100 μl of bacterial suspension per well and incubated at 36°C for 24 h to allow mature biofilm formation. After supernatants were removed, the biofilm was rinsed three times with saline water and further incubated for 24 h with 200 μl of antibiotic and/or bacteriophages. A range of 3 concentrations of antibiotics (corresponding to the MIC of the strain, 10-fold MIC, and the usual intraosseous concentration reached in humans when using standard therapeutic doses [7]) and 4 concentrations of bacteriophages (105 to 108 PFU/ml of each bacteriophage) were used to assess potential dose effects. Then, the biofilm was again rinsed before being resuspended in 100 μl of phosphate-buffered saline (PBS) by scraping the wells with sterile pipette tips followed by 10 min of sonication (40 Hz, BactoSonic; Bandelin, Berlin, Germany). Then, the residual viable number of bacteria after exposure to bacteriophages and or antibiotics was determined by plating 10-fold serial dilutions of the resulting bacterial suspensions on COS agar.
Intracellular infection of osteoblasts and assessment of bacteriophage activity against internalized S. aureus HG001.Bacteriophage activity against internalized S. aureus HG001 was assessed in an in vitro model of osteoblast infection adapted from a previously published protocol (7). All cell culture reagents were purchased from Gibco (Thermo Fischer Scientific, Waltham, MA, USA). MG63 osteoblastic cells (ATCC CRL-1427TM; LGC Standards, Molsheim, France) were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (culture medium) and 100 U/ml penicillin and 100 mg/liter streptomycin at 36°C in a 5% CO2 atmosphere. They were passaged once a week and used up to passage 20 after thawing. Prior to infection experiments, cells were seeded at 50,000 cells per well into 24-well tissue culture plates (Falcon, Le Pont de Claix, France) and incubated during 48 h in culture medium with antibiotics.
Intracellular infection of MG63 cells was performed using a bacterial suspension whose concentration was determined as follows. The day before the infection of osteoblastic cells, 2 ml of an overnight bacterial culture in BHI was centrifuged, and the pellet was resuspended in 10 ml of cell culture medium without antibiotics. Serial dilutions of this suspension were plated on Trypticase soy agar (TSA) (bioMérieux) using an easySpiral device (Interscience, Saint-Nom-la-Bretèche, France) for bacterial enumeration. Meanwhile, the bacterial suspension was kept overnight at 4°C. The next day, this suspension was adjusted at an MOI of 10 bacteria/cell and added to the osteoblastic cell culture wells for 2 h at 36°C to allow bacterial internalization. Cells were then washed twice with phosphate-buffered saline (PBS) and were treated with culture medium containing 10 mg/liter lysostaphin (Sigma-Aldrich, Saint-Louis, MO, US) to kill all remaining extracellular bacteria without affecting intracellular bacteria (33). After 1 h, cells were washed twice, and the cells from three wells were lysed in sterile water over 15 min to determine the number of internalized bacteria at this time, defined as time zero. Then, incubation was carried on for 24 h with 1 ml of culture medium containing bacteriophages and/or an antibiotic. Two concentrations of the combination of three bacteriophages were tested as follows: 107 PFU/ml and 109 PFU/ml of each bacteriophage. The antibiotics were used at their usual local intraosseous concentration (6 mg/liter for both antibiotics) (7).
In the first set of experiments, in order to evaluate specifically the intracellular bacteriophage activity, lysostaphin at 10 mg/liter was also added to the culture medium to kill the bacteria released upon host cell lysis throughout the experiment, thus preventing these bacteria from reinfecting new host cells. The absence of effect of lysostaphin against bacteriophage titer stability was verified. After 24 h of incubation in the presence of lysostaphin with or without bacteriophages, the culture supernatants were removed. Then, the cells were washed twice with PBS, treated with citric buffer (40 mM citric acid in PBS, pH 3) for 2 min in order to inactivate any residual phage particles remaining on the surface of the cells, and washed twice again (25). The number of viable intracellular bacteria was determined after lysis of the cells in sterile water during 15 min. In addition, bacteriophage titrations were performed in both cellular culture supernatants and in the cell lysates.
In the second set of experiments, after the 1-h treatment of infected cells with lysostaphin (10 mg/liter), cells were rinsed and incubation was carried out with bacteriophages and/or antibiotics but without lysostaphin to be sure that lysostaphin was not interfering with bacteriophage penetration inside infected osteoblasts. Thus, in the absence of any treatment, bacteria may escape from infected cells and multiply in the extracellular medium before being reinternalized in other cells. This protocol allowed the evaluation of the natural course of intracellular infection, i.e., the effect of bacteriophage and/or antibiotics on intracellular bacteria and extracellular bacteria released following the death of infected cells.
Electron microscopy analysis.After 24 h of incubation of cells with bacteriophages as described above (second set of experiments), cells were washed and fixed in 2% glutaraldehyde. Then, they were washed three times with a buffered solution containing 1 volume of saccharose 0.4 M and 1 volume of cacodylate 0.2 M (pH 7.4) for 1 h at 4°C and postfixed with a solution containing 1 volume of osmium 2% and 1 volume of cacodylate 0.3 M (pH 7.4) for 1h at 4°C. Cells were then dehydrated using an ethanol gradient (5 min in ethanol 30%, 50%, 70%, and 95% and 3 times in absolute ethanol for 10 min). Impregnation was performed with an epon mix. Inclusion was obtained by polymerization at 60°C for 72 h. Ultrathin sections (approximately 70 nm thick) were cut on a UC7 ultramicrotome (Leica Microsystems, Wetzlar, Germany) mounted on 200 mesh copper grids coated with 1:1,000 polylysine, stabilized for 1 day at room temperature, and contrasted with uranyl acetate and lead citrate. Sections were examined using a 1400-JEM transmission electron microscope (Jeol, Tokyo, Japan).
Statistical analysis.Each experiment was performed three times independently in technical triplicates. For cellular experiments, variables were compared using the nonparametric Mann-Whitney U test. A P value of <0.05 was considered significant. For biofilm eradication data analysis, remaining viable inocula after 24 h of treatment, initially determined in CFU/ml, were log transformed. Effects of bacteriophages or antibiotics in monotherapy were compared to the untreated control condition using the Mann-Whitney U test. Significant variations (P < 0.05) of the viable bacteria counts after 24 h of treatment by bacteriophages or antibiotics alone compared to the untreated condition were denoted with asterisks in Fig. 2. Two-way analysis of variance (ANOVA) and Bonferroni post tests were used to compare the effects of combinations of antibiotics and bacteriophages to that of the antibiotic or bacteriophages in monotherapy at the same concentration. A synergy was defined by an effect of the combination significantly greater (P < 0.05) than that of both agents used alone and was denoted with the sign “S” in Fig. 2. All analyses were performed using Prism software (GraphPad, San Diego, CA, USA).
ACKNOWLEDGMENTS
We thank Elisabeth Errazuriz-Cerda and the CIQLE team (Lyon, France) for the preparation and the observation of samples by transmission electron microscopy.
Pherecydes Pharma provided the bacteriophages used in this study.
Cindy Fevre and Mathieu Medina are employed by Pherecydes Pharma.
FOOTNOTES
- Received 5 November 2019.
- Returned for modification 23 November 2019.
- Accepted 18 December 2019.
- Accepted manuscript posted online 23 December 2019.
- Copyright © 2020 American Society for Microbiology.