Biology paper

37 Microbiology insights 2016:9 Evaluating the Effect of Oxygen Concentrations on Antibiotic Sensitivity, Growth, and Biofilm Formation of Human Pathogens shilpi g upta, n ozrin laskar and Daniel E. Kadouri Depar tment of Oral Biology, Rutgers School of Dental Medicine, Newark, NJ, USA.

ABSTR ACT: Standard antimicrobial susceptibilit y tests are performed in vitro under normal room oxygen conditions to predict the in vivo effectiveness of antimicrobial therapy. The aim of this study was to conduct a comprehensive analysis of the effect of different oxygen levels on the antibiotic susceptibilit y of t wo strains of Staphylococcus aureus , Pseudomonas aeruginosa, and Klebsiella pneumoniae . It was found that anoxic conditions caused reduced sensitivit y of bacteria to aminoglycoside antibiotics in four of six bacteria used in the study. In addition, oxygen limitation decreased the susceptibilit y of P. aeruginosa strains and K. pneumoniae strains to piperacillin/tazobactam and azithromycin, respectively. In contrast, five of six bacteria became more susceptible to tetracycline antibiotics under oxygen-limiting conditions. Our data highlight the importance of considering the potential in vivo oxygen levels within the infection site when setting susceptibilit y breakpoints for evaluating the therapeutic potential of a drug and its effect on antibiotic sensitivit y of the pathogen.

K E Y WOR DS: oxygen concentration, Staphylococcus aureus , Klebsiella pneumoniae, Pseudomonas aeruginosa , antibiotic sensitivit y, minimum inhibitor y concentration CITATION: Introduction Since their introduction in the early 1940s, antibiotics have saved millions of lives and are considered a marvel of modern medicine. However, in the past few years, there has been an alarming increase in antibiotic resistance. 1 The most common microorganisms developing resistance and responsible for two- thirds of healthcare-associated infections have been reported to belong to the “ESK APE” group of pathogens, which include the gram-negative bacteria Acinetobacter baumannii , Enterobacter species, Klebsiella pneumoniae , and Pseudomonas aeruginosa and the gram-positive bacteria Enterococcus faecium and Staphylococcus aureus .2,3 To determine efficacy of an antibiotic or possible drug resistance in a specific pathogen, it is required to isolate and examine the antibiotic susceptibility of the pathogen in a clinical microbiology laboratory. Antimicrobial susceptibility tests are performed in vitro to predict the in vivo effective - ness of antimicrobial therapy and help guide the choice and dosage of the antibiotic. The most commonly used testing methods include broth dilution, disk diffusion, and gradient diffusion methods, along with the use of automated instru - ment systems. 4 The key parameter used as a measure of anti - biotic sensitivity is minimum inhibitory concentration (MIC).

It is calculated as the lowest concentration of the antibiotic required to inhibit the visible growth of a microorganism after overnight incubation. 5,6 One of the methods for MIC determination on agar medium is the use of E-test antibiotic strips. 7 The strips have a predefined antibiotic concentration gradient and the MIC values are determined based on where the zone of inhibition intersects the strip. The E-test has been found to be fairly accurate and comparable with the other con - ventional susceptibility tests. 8,9 The standard antibiotic susceptibility tests (ASTs) are routinely performed on bacteria grown planktonically or on agar plates under normal ambient room oxygen conditions. However, oxygen levels could be low under clinically relevant environments such as in burn wounds, lungs of cystic fibrosis patients, intra-abdominal abscess, the oral cavity, and others, contributing to increased pathogen persistence. 10 –1 2 Further - more, reduced oxygen levels might facilitate biofilm formation for pathogens such as Pseudomonas and Staphylococcus , contrib - uting toward their increasing tolerance to traditionally recom - mended antibiotics. 13 ,14 On the other hand, hyperbaric oxygen therapy (HBO 2) and oxygen therapy 15,16 could increase the exposure of a pathogen to high levels of oxygen, which again might alter the antibiotic susceptibility from that measured under ambient oxygen levels. 17, 18 The purpose of this study was to conduct a compre - hensive analysis of the effect of oxygen on antibiotic sus - ceptibility of three key human pathogens. In this study, a Journal name: Microbiology Insights Journal t y pe: Original Research Ye a r : 2 0 16 Volume: 9 Running head verso: Gupta et al Running head recto: Effect of ox ygen on antibiotic susceptibilit y of human pathogens Gupta et al 38 Microbiology insights 2016:9 system was developed, which allowed experiments to be conducted in enriched oxygen environments. Together with other oxygen-limiting systems, the MICs of 14 antibiot - ics, representing 7 different classes, were determined using both laboratory strains and clinical isolates of S. aureus, K. pneumoniae , and P. aeruginosa . Our data show that oxygen levels can greatly alter the antibiotic sensitivity of the patho - gen and should be taken into consideration when setting up susceptibility breakpoints and evaluating the therapeutic potential of a drug.

Materials and Methods Bacterial strains and culture conditions. The follow - ing bacteria were used in the study: S. aureus FPR3757, a multidrug-resistant strain USA300, 19 S. aureus SH1000, 20 K. pneumoniae A Z1169, 21 K. pneumoniae ATCC 33495, P. aeruginosa U C BP P-PA 14 , 22 and P. aeruginosa PaA, a clini - cal isolate from keratitis patients. 23 The bacteria represent both relatively fresh clinical isolates and commercially obtained laboratory strains. Bacteria were grown routinely in Luria Broth (LB) media at 37 °C. Antibiotic sensitivity test. Fourteen antibiotics were chosen to represent different classes and modes of action, which included the following: b-lactam antibiotics (merope - nem, doripenem, ampicillin, and piperacillin/tazobactam), glycopeptide (vancomycin), tetracyclines (tetracycline and tigecycline), aminoglycosides (gentamicin, amikacin, tobra - mycin, and kanamycin), macrolides (azithromycin), fluoro - quinolones (ciprofloxacin), and rifampicin. Although not all tested antibiotics are clinically relevant for each of the bac - teria evaluated, to maintain consistency, all of the 14 selected antibiotics were tested on all the bacteria. Antibiotic sensi - tivity assays were performed using E-test antibiotic strips (bioMerieux), according to the manufacturer’s protocol, with some modifications. In brief, a single colony of each bacterial strain was inoculated overnight in LB. Thereafter, the cultures were washed and diluted in phosphate-buffered saline (PBS) to bring the concentrations to an optical density (OD 600 ) of 0.1, corresponding to 1  ×  10 8, 4  ×  10 8, and 6  ×  10 7 colony forming units/mL (CFU/mL) for S. aureus , P. aeruginosa , and K. pneumoniae , respectively. A total of 100  µL of S. aureus or K. pneumoniae cells were placed onto a 20  mm Petri plate containing tryptic soy agar (1.5% agar; TSA). P. aeruginosa cells were placed onto plates containing LB supplemented with 1.5% agar and 1% KNO 3. The medium used in the pres - ent study was selected as it allowed the growth of bacteria in all oxygen conditions. Sterile glass beads were used to spread the inoculums on the plates and produce an evenly distrib - uted lawn. Once the agar surface was completely dry, E-test antibiotic strips were placed on top of the microbial lawn with sterile forceps. Plates were placed in the appropriate oxygen environment and incubated at 37 °C for 24 hours. Experi - ments were conducted three times in triplicate. MIC val - ues were determined according to manufacturer’s guidelines (E-test Antimicrobial Susceptibility Testing, 2012), which specified values at the point of complete inhibition of all growth. Antibiotic concentrations on the strips used for meropenem, doripenem, ciprofloxacin, and rifampicin were 0.0 03–32  µg/mL. The remaining antibiotics that were tested had antibiotic gradients 0.016–256  µg /m L . Oxygen growth conditions. Experiments were con - ducted in five oxygen conditions. For anoxic growth envi - ronment (0% O 2), plates were placed in a CoyLab anaerobic chamber (Coy Laboratory Products) using anoxic gas mix of 10 % H 2, 10% CO 2, and 80% N 2. Hypoxic conditions with low oxygen levels (7% –9% O 2, 5% –8% CO 2) were obtained by placing the plates in a sealed Mitsubishi™ AnaeroPack™ 2.5  L Rectangular Jar system containing an AnaeroPack™- MicroAero Gas Generator pack (Mitsubishi Gas Chemical America Inc). For normoxic room oxygen environment (20.8% O2), plates were placed in a standard benchtop incuba - tor (V WR). A benchtop CO 2 incubator was used to obtain enriched CO 2 environment of 5.5%, while maintaining ambient room oxygen levels (20.8%). Finally, for hyperoxic oxygen environment with elevated O 2 levels (95% –99% O 2), a modification of our gasbag system was used (Fig. 1). 24 Figure 1. Closed container system with valve used to maintain hyperoxic gas environment. Plates were placed in a 1 L airtight container. Air was removed using a vacuum tube connected to a standard laboratory vacuum gas tap. Pure oxygen was inserted via a PVC tube connected to a compressed gas cylinder and the container was incubated at 37 °c. Effect of oxygen on antibiotic susceptibility of human pathogens 39 Microbiology insights 2016:9 A 1 L polypropylene airtight container with a sealing O-ring was used (Fisherbrand™ Infecon™ 3000 Infectious Substance Shipper Kit). Two Luer Stopcock valves were placed on 3 mL syringes, which were inserted into the lid and secured using clear silicone sealant. One valve was connected via a clear poly - vinyl chloride (PVC) tube to a standard laboratory vacuum gas tap. The second valve was connected to a gas regulator attached to a compressed gas cylinder containing 99.99% pure medi - cal grade oxygen (Gts-Welco). To conduct the experiment, plates were inserted into the jar and the jar was sealed. Air was removed using the vacuum tube. Thereafter, the vacuum tube was closed and pure oxygen was inserted (final Psi read - ing of ~8). The oxygen was removed once more via vacuum and reintroduced. This gas-flushing procedure was performed five times in order to remove all ambient air. The jars were placed at 37 °C and incubated for 24 hours. No change in gas pressure was measured, conforming that the jar was airtight.

Oxygen levels were confirmed by using a Traceable® Portable Dissolved Oxygen Meter (Fisher Scientific). Growth cur ve and biofilm formation. To m e a s u r e microbial growth in each oxygen condition, each bacterial strain was inoculated overnight in LB. Thereafter, the cells were washed and diluted in PBS to bring the concentrations to an OD 600 of 0.1. The cells were diluted 1:10 in fresh TSB (for S. aureus and K. pneumoniae ) or LB/KNO 3 media (for P. aeruginosa ). The cultures were placed in 96-well micro titer wells (120  µL per well). Each plate was placed at a different oxygen condition, one plate for each time point in a separate jar.

At each time point (0, 10, 20, 28, and 46 hours), the plates were removed and the growth was measured at 600 nm (OD 600 ) using a Synergy H1 Hybrid Multi-Mode Microplate Reader (BioTek).

For biofilm assays, the plates were prepared as described above.

After 24 hours of incubation, plates were removed, washed in order to detach loosely attached cells, and stained with 0.1% crystal violet (CV). CV was solubilized using acetic acid 50% (v/v), and relative biofilm biomass was assayed by measuring the optical density of the CV solution at 600  nm (OD 600 ).25 The experiments were conducted twice in quadruplicate. Statistical analysis. GraphPad Prism 6 (GraphPad Software, Inc.) was used to perform one-way analysis of vari - ance followed by Tukey’s multiple comparison test. Results were considered significant at P-value 0.05. Results Antibiotic sensitivity tests. To measure the effect of different oxygen levels on antibiotic efficacy, MIC tests were conducted using E-test strips. As anticipated, the MIC values varied greatly among the different bacteria and conditions.

Furthermore, large differences in susceptibility were mea - sured ranging from 2-fold to a greater than 30-fold for differ - ent oxygen environments when compared to the MIC values under normal room ambient air incubation. Compared to room O 2 levels, anoxic conditions altered S. aureus strain FPR 3757 susceptibility in 80% of the tested antibiotics. Increased MIC values were measured for meropenem (11-fold), doripenem (20-fold), ampicillin (6-fold), piperacillin/tazobactam (5-fold), kanamycin (7-fold), gentamicin (10-fold), amikacin (12-fold), and tobramycin (23-fold). In comparison, 2–3-fold decreases in MIC values for tetracycline and tigecycline were observed. On the other hand, hyperoxic incubation did not produce any notable change in the MIC values. Hypoxic environment and elevated CO 2 conditions increased the sensitivity of bacteria to piper - acillin/tazobactam 2–3-fold. In contrast, hypoxia decreased the efficacy of amikacin and tobramycin, increasing the MICs 3- and 6-fold, respectively (Table 1). S. aureus strain SH1000 demonstrated similar reductions in susceptibility to aminoglycosides under anoxic conditions indicated by increases in MIC values by 12-, 18-, and 26-fold for gentamicin, amikacin, and tobramycin, respectively.

Hypoxic incubation as well as incubation under high O 2 envi - ronment decreased the efficacy of gentamicin and azithromy - cin by increasing the MICs 3- to 5-fold. Additionally, there was a 2-fold reduction in MIC values for piperacillin/tazo - bactam under hypoxic incubation and for meropenem under high CO 2 conditions, as compared to values under normal ambient air (Table 2). Investigating the sensitivity of K. pneumoniae strain AZ1169 to the various antibiotics, we found that 50% of the antibiotics had altered efficacy under oxygen-limiting condi - tions. Of the 14 antibiotics tested, anoxic incubation made the pathogen more susceptible to tetracycline and ciprofloxacin by 2–3-fold. A 3-fold increase in MIC values was measured for gentamicin, kanamycin, and azithromycin, whereas, a 10-fold increase in MIC values was measured for tobramycin.

A hypoxic environment shifted the MIC value up by 3-fold for tobramycin and over 15-fold for azithromycin moving the value above the maximum concentration imbedded on the strip. On the other hand, a hyperoxic environment decreased the MIC value for ampicillin by almost 2-fold, in contrast to a 3-fold increase for azithromycin (Table 3). K. pneumoniae ATCC strain 33495 showed a simi - lar decline in sensitivity to five of the antibiotics under anoxic incubation, with approximately 3-fold increases in MIC values for gentamicin, amikacin, and kanamycin, and 5- and 8-fold increases for azithromycin and tobramycin, respectively. In contrast, susceptibility toward tigecycline increased 2-fold under anoxic condition. Hypoxic incuba - tion increased the MIC values for gentamicin, kanamycin, and ciprofloxacin by around 2.5-fold, but caused a notable 30-fold increase in MIC value for azithromycin, sugges - tive of a drastic decrease in efficacy under oxygen-limiting conditions. Incubation under elevated CO 2 environment increased the MIC values for kanamycin and azithromy - cin by 2- and 13.5-folds, respectively. In addition, hyper - oxic environment increased the MIC values of tobramycin, kanamycin, azithromycin, and ciprofloxacin by 2-, 2.5-, 33-, and 8.5-folds, respectively (Table 4). Gupta et al 40 Microbiology insights 2016:9 Table 1. Antibiotic sensitivity of S. aureus FPR 3757 under various oxygen conditions. NO. ANTIBIOTIC NORMOXIA ANOXIA HYPOXIA ELEVATED CO 2 HYPEROXIA 1 Meropenem 0.25 ± 0.00 3 ± 0.00 0.09 ± 0.00 0.09 ± 0.00 0.42 ± 0.07 2 Doripenem 0 .13 ± 0.00 2.75 ± 0.96 0 .11 ± 0 .12 0.04 ± 0.01 0 .15 ± 0.04 3 Vancomycin 9.33 ± 2.31 6 ± 0.00 6.67 ± 2.31 6.67 ± 2.31 5.33 ± 1.15 4 Ampicillin 0.5 ± 0.00 3.67 ± 0.58 0.23 ± 0.03 0 .15 ± 0.04 0.83 ± 0.58 5 Piperacillin/tazobactam 4 ± 0.00 24 ± 0.00 1.33 ± 0.29 1 ± 0.00 1. 5 ± 0.00 6 te t r a c y c l i n e 21. 3 3 ± 4.62 6 ± 0.00 10.67 ± 2.31 10.67 ± 2.31 10.67 ± 2.31 7 tigecycline 0.75 ± 0.00 0 .19 ± 0.00 0.29 ± 0.08 0.46 ± 0.07 0.42 ± 0.29 8 Gentamicin 0.63 ± 0 .14 6.67 ± 1.15 1. 8 3 ± 0.29 1. 3 3 ± 0.29 1. 5 6 ± 0.59 9 Amikacin 4 ± 0.00 53.33 ± 9.24 16 ± 6.93 5.33 ± 1.15 6.67 ± 1.15 10 Tobramycin 0.38 ± 0.00 9.33 ± 2.31 2.67 ± 1.15 1 ± 0.00 1 ± 0.00 11 Kanamycin* 4 ± 0.00 32 ± 0.00 6 ± 0.00 6 ± 0.00 6 ± 0.00 12 Azithromycin 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 13 Ciprofloxacin 32 ± 0.00 18.67 ± 4.62 32 ± 0.00 32 ± 0.00 32 ± 0.00 14 Rifampicin 0.01 ± 0.00 0 ± 0.00 0 ± 0.00 0.01 ± 0.00 0.01 ± 0.00 Notes: Data represent M ic ( µg/mL) values. Each experiment was conducted three times, with each value representing the mean and standard deviation. Values in bold represent 2-fold or higher differences in the values compared to that measured under normal room oxygen conditions. Bold gray boxes represent 5 -fold or higher differences in the values compared to that measured under normal room oxygen conditions. The MIC values with “ ” symbol are above the maximum concentration on the E-test strips. *Data represent values from only two experiments, as the antibiotic became unavailable from the manufacturer.

Table 2. Antibiotic sensitivity of S. aureus sh1000 under varying oxygen conditions. NO. ANTIBIOTIC NORMOXIA ANOXIA HYPOXIA ELEVATED CO 2 HYPEROXIA 1 Meropenem 0 .13 ± 0.00 0.09 ± 0.00 0.05 ± 0.01 0.04 ± 0.01 0 .1 ± 0.02 2 Doripenem 0.04 ± 0.01 0.03 ± 0.00 0.02 ± 0.00 0.01 ± 0.00 0.04 ± 0.01 3 Vancomycin 10.67 ± 2.31 8 ± 0.00 8 ± 0.00 8 ± 0.00 8 ± 0.00 4 Ampicillin 0 .11 ± 0.02 0 .13 ± 0.00 0.08 ± 0.02 0.05 ± 0.01 0 .11 ± 0.02 5 Piperacillin/tazobactam 0.46 ± 0.07 0.38 ± 0.00 0 .13 ± 0.00 0.23 ± 0.03 0.75 ± 0.00 6 te t r a c y c l i n e 0.63 ± 0 .14 0.58 ± 0.29 0.5 ± 0.00 0.46 ± 0.07 0.38 ± 0.00 7 tigecycline 0.38 ± 0.00 0.32 ±0 .11 0.42 ± 0 .14 0.27 ± 0.08 0.23 ± 0.03 8 Gentamicin 0.25 ± 0.00 3.33 ± 0.58 1 ± 0.00 0.58 ± 0 .14 1 ± 0.00 9 Amikacin 3 ± 0.00 58.67 ± 33.3 3.33 ± 0.58 2.67 ± 0.58 4.67 ± 1.15 10 Tobramycin 0.29 ± 0.08 8 ± 0.00 0.67 ± 0.29 0.5 ± 0.00 0.75 ± 0.00 11 Kanamycin* – – – – – 12 Azithromycin 2 ± 0.00 2 ± 0.00 13.33 ± 2.31 4.67 ± 1.15 6.67 ± 2.31 13 Ciprofloxacin 0.5 ± 0.00 0.21 ± 0.03 0.5 ± 0.00 0.46 ± 0.07 0.58 ± 0 .14 14 Rifampicin 0.02 ± 0.00 0.01 ± 0.00 0.01 ± 0.00 0.01 ± 0.00 0.01 ± 0.00 Notes: Data represent M ic ( µg/mL) values. Each experiment was conducted three times, with each value representing the mean and standard deviation. Values in bold represent 2-fold or higher differences in the values compared to that measured under normal room oxygen conditions. Bold gray boxes represent 5 -fold or higher differences in the values compared to that measured under normal room oxygen conditions. The MIC values with “ ” symbol are above the maximum concentration on the E-test strips. *Antibiotic was not available from the manufacturer. Antibiotic sensitivity testing on P. aeruginosa strain Pa14 showed the highest sensitivity to tetracycline under anoxic condition, which was 5-fold more than that measured under normal ambient oxygen conditions. However, oxygen limi - tation increased MIC values of amikacin by approximately 2-fold, but remarkably shifted the MIC for piperacillin/tazo - bactam combination to values above the maximum concentra - tion on the E-test strip. CO 2-enriched environment led to a 3-fold decrease in MIC values for ampicillin. Finally, hyper - oxic environment decreased the sensitivity of the bacterium to around one-third of the antibiotics, with 3- and 4-fold increases in MIC values measured for amikacin and cipro - floxacin, respectively (Table 5). P. aeruginosa strain PaA demonstrated higher suscepti - bility to one-third of the test antibiotics under oxygen-limit - ing conditions. MIC values for tetracycline and tigecycline in Effect of oxygen on antibiotic susceptibility of human pathogens 41 Microbiology insights 2016:9 Table 3. Antibiotic sensitivity of K. pneumoniae A Z1169 under varying oxygen conditions. NO. ANTIBIOTIC NORMOXIA ANOXIA HYPOXIA ELEVATED CO 2 HYPEROXIA 1 Meropenem 0.02 ± 0.00 0.01 ± 0.00 0.02 ± 0.00 0.01 ± 0.00 0.01 ± 0.00 2 Doripenem 0.03 ± 0.01 0.01 ± 0.00 0.01 ± 0.00 0.02 ± 0.00 0.01 ± 0.00 3 Vancomycin 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 4 Ampicillin 160 ± 55.43 256 ± 0.00 256 ± 0.00 256 ± 0.00 56 ± 11.13 5 Piperacillin/tazobactam 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 6 te t r a c y c l i n e 12 ± 0.00 4 ± 0.00 12 ± 0.00 12 ± 0.00 8 ± 0.00 7 tigecycline 4 ± 0.00 3 ± 0.00 6 ± 0.00 6 ± 0.00 6 ± 0.00 8 Gentamicin 0.75 ± 0.00 3.33 ± 0.58 3 ± 0.00 1. 8 3 ± 0.29 1. 8 3 ± 0.29 9 Amikacin 3 ± 0.00 8 ± 3.46 5 ± 0.00 4 ± 1.73 4.67 ± 1.15 10 Tobramycin 0.75 ± 0.00 8 ± 3.46 3 ± 0.00 1. 8 3 ± 0.29 1. 5 ± 0.00 11 Kanamycin 2 ± 0.00 8 ± 0.00 4.67 ± 1.15 3.33 ± 0.58 4 ± 0.00 12 Azithromycin 16 ± 0.00 64 ± 0.00 256 ± 0.00 64 ± 0.00 64 ± 0.00 13 Ciprofloxacin 32 ± 0.00 6.67 ± 1.15 32 ± 0.00 32 ± 0.00 32 ± 0.00 14 Rifampicin 32 ± 0.00 26.67 ± 4.62 32 ± 0.00 32 ± 0.00 32 ± 0.00 Notes: Data represent M ic ( µg/mL) values. Each experiment was conducted three times, with each value representing the mean and standard deviation. Values in bold represent 2-fold or higher differences in the values compared to that measured under normal room oxygen conditions. Bold gray boxes represent 5 -fold or higher differences in the values compared to that measured under normal room oxygen conditions. The MIC values with “ ” symbol are above the maximum concentration on the E-test strips.

Table 4. Antibiotic sensitivity of K. pneumoniae ATCC 33495 under varying oxygen conditions. NO. ANTIBIOTIC NORMOXIA ANOXIA HYPOXIA ELEVATED CO 2 HYPEROXIA 1 Meropenem 0.03 ± 0.01 0.02 ± 0.00 0.03 ± 0.00 0.03 ± 0.01 0.03 ± 0.01 2 Doripenem 0.05 ± 0.00 0.03 ± 0.00 0.03 ± 0.01 0.04 ± 0.02 0.03 ± 0.00 3 Vancomycin 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 4 Ampicillin 3.5 ± 0.58 3.67 ± 0.58 6.75 ± 4 .11 5.33 ± 1.15 5.33 ± 2.31 5 Piperacillin/Tazobactam 4 ± 0.00 3.67 ± 0.58 6 ± 0.00 6.67 ± 2.31 8 ± 0.00 6 te t r a c y c l i n e 42.67 ± 9.24 21. 3 3 ± 4.62 32 ± 0.00 10 6.67 ± 18.48 96 ± 0.00 7 tigecycline 0.75 ± 0.00 0.25 ± 0.00 1 ± 0.00 1 ± 0.00 1. 5 ± 0.00 8 Gentamicin 0.75 ± 0.00 3 ± 0.00 2.67 ± 0.58 1.67 ± 0.29 1. 8 3 ± 0.29 9 Amikacin 2 ± 0.00 9.33 ± 2.31 4 ± 0.00 3.67 ± 0.58 4 ± 0.00 10 Tobramycin 0.67 ± 0 .14 6 ± 0.00 3 ± 0.00 2.33 ± 0.58 2 ± 0.87 11 Kanamycin 1. 27 ± 1.4 8 5.33 ± 2.31 4.67 ± 1.15 4 ± 1.73 4.67 ± 1.15 12 Azithromycin 3.67 ± 0.58 21.33 ± 4.62 117. 3 3 ± 18.48 53.33 ± 18.48 128 ± 0.00 13 Ciprofloxacin 0.02 ± 0.00 0.04 ± 0.01 0.07 ± 0.02 0.04 ± 0.01 0 .19 ± 0.00 14 Rifampicin 32 ± 0.00 32 ± 0.00 32 ± 0.00 32 ± 0.00 32 ± 0.00 Notes: Data represent M ic ( µg/mL) values. Each experiment was conducted three times, with each value representing the mean and standard deviation. Values in bold represent 2-fold or higher differences in the values compared to that measured under normal room oxygen conditions. Bold gray boxes represent 5 -fold or higher differences in the values compared to that measured under normal room oxygen conditions. The MIC values with “ ” symbol are above the maximum concentration on the E-test strips.

normoxic conditions were found to be over the maximum con - centration on the strip. However, anoxic, hypoxic, and hyperoxic environments reduced the values by at least 8- to 12-, 7-, and 4- to 8-folds, respectively. Anoxic environment decreased the MIC values for meropenem and rifampicin by more than 2-fold, in contrast to an increase of 2-fold for amikacin and notably more for piperacillin/tazobactam, suggestive of a considerable decrease in bacterial susceptibility as compared to that under normal ambient air incubation. Elevated CO 2 potentiated the activity of carbapenems and tigecycline by decreasing the MICs 3-fold (Table 6). Microbial growth. To determine if there is a correlation between sensitivity of the pathogens to various antibiotics under different oxygen conditions and the growth capability of Gupta et al 42 Microbiology insights 2016:9 Table 5. Antibiotic sensitivity of P. aeruginosa Pa14 under varying oxygen conditions. NO. ANTIBIOTIC NORMOXIA ANOXIA HYPOXIA ELEVATED CO 2 HYPEROXIA 1 Meropenem 0 .13 ± 0.00 0 .11 ± 0.02 0 .13 ± 0.00 0 .11 ± 0.02 0 .1 9 ± 0.00 2 Doripenem 0 .1 ± 0.04 0 .13 ± 0.00 0 .1 9 ± 0.00 0 .13 ± 0.00 0 .17 ± 0.04 3 Vancomycin 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 4 Ampicillin* 24 24 32 8 64 5 Piperacillin/tazobactam 2 ± 0.00 256 ± 0.00 1. 5 ± 0.00 1.17 ± 0.29 2.67 ± 0.58 6 te t r a c y c l i n e 85.33 ± 18.48 13.33 ± 2.31 48 ± 0.00 3 7. 3 3 ± 9.24 42.67 ± 9.24 7 tigecycline 24 ± 0.00 18.67 ± 4.62 12 ± 0.00 29.33 ± 4.62 24 ± 0.00 8 Gentamicin 2.67 ± 0.58 3.33 ± 0.58 3.33 ± 0.00 2.67 ± 0.58 5.33 ± 1.15 9 Amikacin 8 ± 0.00 26.67 ± 4.62 13. 3 3 ± 0.00 9.33 ± 2.31 32 ± 0.00 10 Tobramycin 1. 5 ± 0.00 6.67 ± 2.31 4 ± 0.00 2.67 ± 0.58 4.67 ± 1.15 11 Kanamycin** – – – – – 12 Azithromycin 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 13 Ciprofloxacin 0 .13 ± 0.00 0 .13 ± 0.00 0.29 ± 0.08 0.34 ± 0.08 0.75 ± 0.00 14 Rifampicin 32 ± 0.00 32 ± 0.00 32 ± 0.00 32 ± 0.00 32 ± 0.00 Notes: Data represent M ic ( µg/mL) values. Each experiment was conducted three times, with each value representing the mean and standard deviation. Values in bold represent 2-fold or higher differences in the values compared to that measured under normal room oxygen conditions. Bold gray boxes represent 5 -fold or higher differences in the values compared to that measured under normal room oxygen conditions. The MIC values with “ ” symbol are above the maximum concentration on the E-test strips. *Data represent values from only one experiment as the manufacturer discontinued the antibiotic. **Antibiotic was not available from the manufacturer.

Table 6. Antibiotic sensitivity of P. aeruginosa PaA under varying oxygen conditions. NO. ANTIBIOTIC NORMOXIA ANOXIA HYPOXIA ELEVATED CO 2 HYPEROXIA 1 Meropenem 0.29 ± 0.08 0.08 ± 0.02 0.06 ± 0.01 0.06 ± 0.00 0 .15 ± 0.04 2 Doripenem 0.32 ± 0 .11 0 .17 ± 0.07 0 .17 ± 0.04 0.08 ± 0.02 0.25 ± 0.00 3 Vancomycin 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 4 Ampicillin* – – – – – 5 Piperacillin/tazobactam 1. 8 3 ± 0.29 256 ± 0.00 2 ± 0.00 1. 5 ± 0.00 1. 5 ± 0.00 6 te t r a c y c l i n e 256 ± 0.00 18.67 ± 4.62 32 ± 0.00 256 ± 0.00 58.67 ± 9.24 7 tigecycline 256 ± 0.00 26.67 ± 4.62 32 ± 0.00 58.67 ± 9.24 29.33 ± 4.62 8 Gentamicin 5.33 ± 1.15 14.67 ± 2.31 10.67 ± 2.31 7. 3 3 ± 1.15 12 ± 4.00 9 Amikacin 26.67 ± 4.62 85.33 ± 36.95 42.67 ± 9.24 26.67 ± 4.62 3 7. 3 3 ± 9.24 10 Tobramycin 3.33 ± 0.58 8 ± 0.00 6.67 ± 1.15 4.75 ± 1. 5 0 4 ± 0.00 11 Kanamycin* – – – – – 12 Azithromycin 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 256 ± 0.00 13 Ciprofloxacin 0 .1 9 ± 0.00 0 .15 ± 0.04 0.38 ± 0.00 0.23 ± 0.03 0.46 ± 0.07 14 Rifampicin 32 ± 0.00 13.33 ± 2.31 32 ± 0.00 32 ± 0.00 32 ± 0.00 Notes: Data represent M ic ( µg/mL) values. Each experiment was conducted three times, with each value representing the mean and standard deviation. Values in bold represent 2-fold or higher differences in the values compared to that measured under normal room oxygen conditions. Bold gray boxes represent 5 -fold or higher differences in the values compared to that measured under normal room oxygen conditions. The MIC values with “ ” symbol are above the maximum concentration on the E-test strips. *Antibiotic was not available from the manufacturer.

different bacteria, optical density of the bacterial cultures was measured after 0, 10, 20, 28, and 46 hours of incubation under different oxygen environments. For S. aureus, both FPR 3757 and SH1000 displayed the highest growth under hyperoxic environment, followed by ambient environment, least being under the oxygen-limiting conditions. For both the strains of K. pneumoniae , anoxic environment was the least favorable, with all other conditions being similar in facilitating growth as compared to growth under normal room air. Growth curves for both the strains of P. aeruginosa indicate least growth under anoxic conditions, but Pa14 strain had the highest optical den - sity under hyperoxia, in contrast to a hypoxic environment favoring the growth of PaA (Fig. 2). In conclusion, differences in growth of the bacteria under varied environments were Effect of oxygen on antibiotic susceptibility of human pathogens 43 Microbiology insights 2016:9 observed, with anoxic conditions being the least favorable for all the pathogens tested. Biofilm formation. As biofilm formation greatly influ - ences antibiotic tolerance, 26 we measured the ability of each bacterium to form biofilm under conditions similar to those under which the antibiotic sensitivity tests were performed.

To this end, biofilm formation was measured after 24 hours of incubation under the different oxygen conditions. For one-third of the bacteria tested, a significant ( P  0.0001) pos - itive correlation between oxygen levels and biofilm formation was observed, with the highest biofilm biomass seen for S. aureus SH1000 and P. aeruginosa Pa14 in 100% oxygen environments, as compared to that under ambient air envi - ronment. In contrast, P. aeruginosa strain PaA showed a decrease in biofilm formation under low as well as elevated oxygen environments (Fig. 3). Figure 2. Effect of ambient oxygen levels on microbial growth. Ninety-six-well plates were inoculated with S. aureus FPR 3757, S. aureus sh1000, K. pneumoniae A Z11 6 9 , K. pneumoniae ATCC 33495, P. aeruginosa Pa14, and P. aeruginosa PaA. Plates were placed under varying oxygen conditions (five plates for each condition containing all six bacteria). At each time point (0, 10, 20, 28, and 46 hours), one plate was removed from its designated incubation chamber and growth was monitored by the change in culture turbidity measured as optical density at 600 nm (OD 600). Each value represents the mean of 8 wells. Error bars are shown as one standard deviation. Gupta et al 44 Microbiology insights 2016:9 Discussion In recent years, nonprudent use of antibiotics has contributed to the surge of multidrug-resistant infections. 27 In order to make accurate therapeutic decisions, it might be important to determine antibiotic susceptibility under clinically relevant environmental conditions. Standard ASTs are done under normal room oxygen conditions, despite the fact that differ - ent sites of infection in the body have different oxygen and carbon dioxide concentrations. 28 The purpose of this study was to conduct a comprehensive analysis of the effect of oxygen on susceptibility of key human pathogens to a range of antibiotics. Aminoglycosides are the most commonly used broad- spectrum antibiotics that inhibit bacterial protein synthesis by an energy-dependent mechanism for transport into the bacterial cells. 29 We observed that anoxic conditions caused the most notable reduction of the susceptibilities of S. aureus and K. pneumoniae strains to aminoglycoside antibiotics (Tables 1–4). This observation was in accordance with previ - ous studies, which indicate that the bacterial uptake of these antibiotics is oxygen dependent, and thereby an anoxic envi - ronment markedly curtails their efficacy. 14 , 3 0 In addition, we observed that incubation under anoxic environment led to substantial reduction in sensitivity of S. aureus strain FPR3757 to the b-lactam group of antibiotics (Table 1). b-Lactam antibiotics are known to inhibit bacterial cell wall synthesis by binding to penicillin-binding proteins (PBPs) on the cell membrane. 31 Oxygen deprivation may alter the expression of PBPs or decrease the affinity of the drugs for the PBP enzymes, leading to decreased sensitivity. How - ever, this correlation is not well documented. Furthermore, K. pneumoniae strains demonstrated reduced sensitivity to the macrolide antibiotic, azithromycin, under anoxic conditions (Tables 3 and 4). Macrolide antibiotics are bacterial protein synthesis inhibitors. Reduction in efficacy of these antimicro - bial agents suggests a modification of cell metabolic signaling pathways under oxygen-deficient environment; however, the mechanism is not clear. 32 Piperacillin/tazobactam is a b-lactam / b-lactamase inhibitor antibiotic combination that inhibits bacterial cell wall synthesis. In this study, we found that both the P. aeruginosa strains show an increase in resistance to piper - acillin/tazobactam under anoxic conditions (Tables 5 and 6). Possible reason for resistance could be altered membrane permeability to the drug under oxygen-deprived condi - tions. 33 In addition, an overexpression of multidrug efflux pump protein might lead to development of resistance. 34 Our data were not consistent with a previous study, which found no change in MIC values for piperacillin/tazobactam under anoxic conditions. 35 The different results in our study com - pared to the published work could be attributed to different bacterial strains used in the two studies. Piperacillin/tazo - bactam E-tests manufactured between December 2012 and October 2015 were recalled by manufacturer for issues in the results provided. However, the purpose of this study was to determine the impact of oxygen levels on antimicrobial sus - ceptibility and not to obtain actual MIC values to be used for clinical applications. In contrast to observed reduction in aminoglycoside, b-lactam, and macrolide antibiotic efficacies under anoxic environment, tetracycline and tigecycline antibiotics were found to be more effective against majority of the pathogens under limited oxygen conditions. There is a lack of docu - mented studies investigating the effect of oxygen on tetracy - cline antibiotics. To our knowledge, it has not been previously shown that tetracycline antibiotics might be more effective under oxygen-limiting conditions. We found that the MIC values for P. aeruginosa strains decreased by greater extent compared to those for the other pathogens. P. aeruginosa strain PaA was resistant to tetracycline antibiotics at normal air incubation, but became susceptible under altered oxygen concentrations (Table 6). Mechanisms attributed to tetracy - cline resistance include its energy-dependent efflux, ribosomal protection, and oxygen-modulated chemical alteration of the antibiotic. 36,37 Hence, under depleted oxygen environment, there might be downregulation of these pathways, leading to increased susceptibility. In addition, oxygen limitation might favor loss of antibiotic resistance genes due to elevated meta - bolic burden. 38 However, we found an increase in suscepti - bilit y of P. aeruginosa PaA to tetracyclines under hyperoxic conditions as well, the mechanism for which is not clear. These Figure 3. Effect of ambient oxygen levels on biofilm formation. Ninety- six-well plates were inoculated with S. aureus FPR 3757 (FPR 3757), S. aureus SH1000 (SH1000), K. pneumoniae A Z1169 (A Z1169), K. pneumoniae ATCC 33495 (ATCC 33495), P. aeruginosa Pa14 (Pa14), and P. aeruginosa PaA (PaA). Plates were placed under varying oxygen conditions. Data represent the amount of CV staining following 24 hours of incubation and measured at 600 nm (OD 600). Each value represents the mean of 8 wells. Error bars are shown as one standard deviation.

Asterisks indicate significant differences ( P  0.0001) as compared to values under normal ambient air. Effect of oxygen on antibiotic susceptibility of human pathogens 45 Microbiology insights 2016:9 production. However, there might be additional in vivo differ - ences in sensitivities depending on the best growth conditions available for a pathogen and ambient environment at the site of infection. Hyperoxia has been proposed to increase the antimicro - bial efficacy of some of the antibiotics as well as to help restore normal oxygen tension in ischemic tissues. 18 Although expo - sure to elevated oxygen alone might be bacteriostatic, a combi - nation therapy with antimicrobials is recommended. 17 Thus, it is important to determine the changes in bacterial susceptibil - ity to antimicrobials in the presence of elevated oxygen levels.

For this study, we have modified and developed a low-cost system that allows experiments to be conducted in a variety of ambient gas environments (Fig. 1). Since the container is sealed, it could be used with gases that are considered unsafe for use in a standard laboratory environment, such as elevated oxygen levels. The ability to insert Petri dishes and standard 6–96-well plates allows experiments to be conducted using standard laboratory protocols. Finally, the small size of the unit allows compatibility with standard laborator y equipment such as incubators and shakers. In this study, MICs were measured using E-test as the simplest estimate of antibacterial effect. Being a gradient dif - fusion test, E-test strips make it possible to determine MIC values between the conventional two-fold dilution values suggested by breakpoints. 44 It is important to note that the AST-like protocols used in this study were for research pur - poses only, and the MIC values obtained might not reflect actual drug concentration for therapeutic use. It is essential that the clinical applications be based on the most current breakpoints from international organizations such as CLSI a n d E U C A S T. 45,46 Conclusion In this study, we have conducted a thorough analysis of the effects of different oxygen environments on susceptibility of medically relevant bacteria to an extensive range of antibiotics.

We found that oxygen limitation decreases the sensitivity of the pathogens to most of the antibiotics. We also observed that enriched oxygen environment might favor growth and biofilm formation of some pathogens, but not necessarily reflect onto a significant difference in antimicrobial efficacy. We conclude that use of clinically relevant oxygen environments should be a parameter in antimicrobial susceptibility testing and the breakpoints should be set accordingly. This would help phy - sicians make better therapeutic decisions by predicting more accurately the susceptibility of the pathogens in vivo, thereby leading to improved clinical outcome.

Acknowledgments The authors would like to thank Robert Shanks from the University of Pittsburg and Nancy Connell from New Jersey Medical School, for reviewing the work and for their con - structive comments. findings emphasize the importance of relevant oxygen levels for AST. In vitro sensitivity testing under normoxia might suggest resistance of the pathogen to tetracycline antibiotics, but in fact, the existent clinical environment might be oxygen deprived and the antibiotic may actually be effective. In comparison to oxygen limitation, testing suscepti - bility of the pathogens to antimicrobials at high CO 2 and O2 conditions did not show any consistent trend for most of the pathogens tested. The only noteworthy finding was the decreased susceptibility of both the K. pneumoniae strains to azithromycin under high CO 2 and high O 2 concentra - tions. Similar reductions in sensitivity were found under limited oxygen environments (Tables 3 and 4). These find - ings suggest that normoxic environment is best suited for azithromycin efficacy. Target site modification or efflux of the drug in response to altered oxygen environments may be some of the possible factors responsible for reduction in efficacy of the drug. 39 – 41 In order to correlate antibiotic efficacy to growth pat - terns of the pathogens in each oxygen environment, growth curve assays were performed. As anticipated, there were dif - ferences in growth of the bacteria under limited or enriched O2 environments in comparison to that under normal room oxygen levels. Hyperoxia facilitated growth of 50% of the pathogens tested, whereas anoxic and hypoxic environments were least favorable for all the pathogens (Fig. 2). However, when comparing the antibiotic MICs to the growth of these bacteria under the different environments, there does not seem to be a significant correlation between the environment that is most favorable for the growth of a bacterium and its susceptibility toward the antibiotics under those conditions.

For example, hyperoxic conditions favored growth of both the strains of S. aureus. However, SH1000 strain showed a decreased sensitivity to gentamicin and azithromycin, whereas FPR3757 did not show any change in the MIC val - ues under elevated oxygen conditions, as compared to those under normal incubation. It is widely believed that biofilm formation protects bac - teria from antibiotic challenges and increases their tolerance to antimicrobials, contributing to the chronic nature of infec - tions. 42,43 Biofilm formation assay in the current study dem - onstrated that one-third of the pathogens tested differ in their ability to form biofilms under varied oxygen environments.

We found that S. aureus SH1000 and P. aeruginosa Pa14 showed a higher biofilm biomass buildup under hyperoxic environment (Fig. 3). This increase can be linked to a higher growth of these pathogens under elevated oxygen environ - ment, although no such association was found with respect to antibiotic sensitivity. Hence, the influence of different oxygen environments on biofilm formation by the pathogens does not necessarily translate into differences in their susceptibilities to antibiotics. This underlines the fact that the differences observed in MICs under various oxygen environments are not reflective of mere differences in growth rates and/or biofilm Gupta et al 46 Microbiology insights 2016:9 Author Contributions Conceived and designed the experiments: DEK. Analyzed the data: SG, NL, and DEK. Wrote the first draft of the manu - script: SG and DEK. Contributed to the writing of the manu - script: SG, NL, and DEK. Agreed with manuscript results and conclusions: SG, NL, and DEK. Jointly developed the structure and arguments for the paper: SG and DEK. Made critical revisions and approved the final version: SG and DEK.

All the authors reviewed and approved the final manuscript.

REFERENCES 1. Bebell LM, Muiru AN. Antibiotic use and emerging resistance: how can resource-limited countries turn the tide? Glob Heart . 2014;9(3):347–358. 2. R ice LB. Federal funding for the study of antimicrobial resistance in nosocomial pathogens: no ESK APE. J Infect Dis . 20 08;197(8):1079 –1081. 3. Bassetti M, Merelli M, Temperoni C, Astilean A. New antibiotics for bad bugs: where are we? Ann Clin Microbiol Antimicrob . 2013;12:22. 4. Jorgensen JH, Ferraro MJ. Antimicrobial susceptibilit y testing: a review of general principles and contemporar y practices. Clin Infect Dis . 2 0 0 9;49(11):1749 –1755. 5. Andrews JM. Determination of minimum inhibitor y concentrations. J Antimicrob Chemother . 20 01;48(suppl 1):5–16. 6. European Committee for Antimicrobial Susceptibilit y Testing (EUCAST) of the European Societ y of Clinical Microbiolog y and Infectious Diseases (ESC - MID). EUCAST Definitive Document E.DEF 2.1, August 20 0 0: determi - nation of antimicrobial susceptibilit y test breakpoints. Clin Microbiol Infect . 20 0 0;6(10):570 –572. 7. Baker CN, Stocker SA, Culver DH, Thornsberr y C. Comparison of the E Test to agar dilution, broth microdilution, and agar diff usion susceptibilit y testing tech - niques by using a special challenge set of bacteria. J Clin Microbiol . 1991; 29(3): 533 –538. 8. Huang MB, Baker CN, Banerjee S, Tenover FC. Accuracy of the E test for deter - mining antimicrobial susceptibilities of Staphylococci , Enterococci , Campylobacter jejuni, and gram-negative bacteria resistant to antimicrobial agents. J Clin Microbiol . 1992;30(12):3243–3248. 9. Joyce LF, Downes J, Stockman K, Andrew JH. Comparison of five methods, including the PDM Epsilometer test (E test), for antimicrobial susceptibilit y testing of Pseudomonas aeruginosa . J Clin Microbiol . 1992;30(10):2709–2713. 10. Schobert M, Tielen P. Contribution of ox ygen-limiting conditions to persistent infection of Pseudomonas aeruginosa . Future Microbiol . 2010;5(4):603– 621. 11. Eltzschig HK, Carmeliet P. Hy poxia and inflammation. N Engl J Med . 2011; 364(7):656–665. 12. Starnes HF Jr, Warren RS, Jeevanandam M, et al. Tumor necrosis factor and the acute metabolic response to tissue injur y in man. J Clin Invest . 1988;82(4):1321–1325. 13. Borriello G, Werner E, Roe F, K im A M, Ehrlich GD, Stewart PS. Ox ygen limi - tation contributes to antibiotic tolerance of Pseudomonas aeruginosa in biofilms. Antimicrob Agents Chemother . 20 04;48(7):2659–2664. 14. Harrell LJ, Evans JB. Effect of anaerobiosis on antimicrobial susceptibilit y of staphylococci . Antimicrob Agents Chemother . 1977;11(6):1077–1078. 15. Woo KY, Coutts PM, Sibbald RG. Continuous topical ox ygen for the treatment of chronic wounds: a pilot study. Adv Skin Wound Care . 2012;25(12):543–547. 16. Roeck l-Wiedmann I, Bennett M, K ranke P. Systematic review of hy perbaric ox ygen in the management of chronic wounds. Br J Surg . 20 05;92(1):24 –32. 17. Bornside GH. Enhancement of antibiotic activit y against Staphylococcus aureus by exposure to hy perbaric ox ygen. Appl Microbiol . 1967;15(5):102 0 –1024. 18. Muhvich K H, Park MK, Myers R A, Marzella L. Hy peroxia and the antimi - crobial susceptibilit y of Escherichia coli and Pseudomonas aeruginosa . Antimicrob Agents Chemother . 1989;33(9):1526–1530. 19. Diep BA, Gill SR, Chang RF, et al. Complete genome sequence of USA30 0, an epidemic clone of communit y-acquired meticillin-resistant Staphylococcus aureus . Lancet . 2 0 0 6;367(951 2):731–739. 20. Horsburgh MJ, Aish JL, W hite IJ, Shaw L, Lithgow JK, Foster SJ. sigmaB mod - ulates virulence determinant expression and stress resistance: characterization of a functional rsbU strain derived from Staphylococcus aureus 8325-4. J Bacteriol . 2 0 02;184(19):5457–54 67. 21. Kadouri DE, To K, Shanks R M, Doi Y. Predator y bacteria: a potential ally against multidrug-resistant Gram-negative pathogens. PLoS One . 2013;8(5):e63397. 22. Rahme LG, Stevens EJ, Wolfort SF, Shao J, Tompk ins RG, Ausubel FM. Com - mon virulence factors for bacterial pathogenicit y in plants and animals. Science . 1995; 268(5219):1899 –19 02 . 23. Shanks R M, Davra V R, Romanowsk i EG, et al. An eye to a k ill: using predator y bacteria to control Gram-negative pathogens associated with ocular infections. PLoS One . 2013;8(6):e66723. 24. Kadouri DE, Tran A. Measurement of predation and biofilm formation under different ambient ox ygen conditions using a simple gasbag-based system. Appl Environ Microbiol . 2013;79(17):526 4 –5271. 25. Merritt JH, Kadouri DE, O’Toole GA. Growing and analyzing static biofilms. Curr Protoc Microbiol . 20 05; Chapter 1:Unit 1B1. 26. Davies D. Understanding biofilm resistance to antibacterial agents. Nat Rev Drug Discov . 20 03;2(2):114 –122. 27. Lee CR, Lee JH, Kang LW, Jeong BC, Lee SH. Educational effectiveness, tar - get, and content for prudent antibiotic use. Biomed Res Int . 2015;2015:214021. 28. Davis JC, Hunt TK. Problem Wounds: The Role of Oxygen . New York, N Y: Elsevier; 1988. 29. Davis BD. Mechanism of bactericidal action of aminoglycosides. Microbiol Rev . 1987;51(3):341–350. 30. Tack KJ, Sabath LD. Increased minimum inhibitor y concentrations with anaero - biasis for tobramycin, gentamicin, and amikacin, compared to latamoxef, piper - acillin, chloramphenicol, and clindamycin. Chemotherapy . 1985;31(3):204 –210. 31. Neu HC. Penicillin-binding proteins and beta-lactamases: their effects on the use of cephalosporins and other new beta-lactams. Curr Clin Top Infect Dis . 1987; 8:37– 61. 32. Shima K, Szaszak M, Solbach W, Gieffers J, Rupp J. Impact of a low-ox ygen environment on the efficacy of antimicrobials against intracellular Chlamydia trachomatis . Antimicrob Agents Chemother . 2011;55(5):2319–2324. 33. Gotoh N. [Antibiotic resistance caused by membrane impermeabilit y and multi - drug efflu x systems]. Nihon Rinsho . 2001;59(4):712–718. 34. Schaible B, Taylor CT, Schaffer K. Hy poxia increases antibiotic resistance in Pseudomonas aeruginosa through altering the composition of multidrug efflu x pumps. Antimicrob Agents Chemother . 2012;56(4):2114 –2118. 35. Field TR, W hite A, Elborn JS, Tunney MM. Effect of ox ygen limitation on the in vitro antimicrobial susceptibilit y of clinical isolates of Pseudomonas aerugi - nosa grown planktonically and as biofilms. Eur J Clin Microbiol Infect Dis . 20 05; 24(10):677– 687. 36. Salyers A A, Speer BS, Shoemaker NB. New perspectives in tetracycline resis - tance. Mol Microbiol . 1990;4(1):151–156. 37. Grossman TH. Tetracycline antibiotics and resistance. Cold Spring Harb Perspect Med . 2016;6(4):a025387. 38. Rysz M, Mansfield W R, Fortner JD, A lvarez PJJ. Tetracycline resistance gene maintenance under var ying bacterial grow th rate, substrate and ox ygen availabil - it y, and tetracycline concentration. Environ Sci Technol . 2013;47(13):6995–70 01. 39. Goldman RC, Scaglione F. The macrolide-bacterium interaction and its biologi - cal basis. Curr Drug Targets Infect Disord . 2004;4(3):241–260. 40. Taha N, Araj GF, Wak im R H, et al. Genot y pes and serot y pe distribution of macrolide resistant invasive and non-invasive Streptococcus pneumoniae isolates from Lebanon. Ann Clin Microbiol Antimicrob . 2012;11:2. 41. Li X Z, Plesiat P, Nikaido H. The challenge of efflu x-mediated antibiotic resis - tance in Gram-negative bacteria. Clin Microbiol Rev . 2015;28(2):337–418. 42. Moreau-Marquis S, Stanton BA, O’Toole GA. Pseudomonas aeruginosa biofilm formation in the cystic fibrosis air way. Pulm Pharmacol Ther . 20 08; 21(4):595–599. 43. Vuotto C, Longo F, Balice MP, Donelli G, Varaldo PE. Antibiotic resistance related to biofilm formation in Klebsiella pneumoniae . Pathogens . 2 014 ; 3(3): 743–758. 44. Jorgensen JH, Turnidge JD. Susceptibilit y test methods: dilution and disk dif - fusion methods. In: Jorgensen J, Pfaller M, Carroll K, Funke G, Landr y M, R ichter S, Warnock D, eds. Manual of Clinical Microbiology, Eleventh Edition . Washington, DC: ASM Press; 2015:1253–1273. 45. Coyle MB; American Societ y for Microbiolog y. Manual of Antimicrobial Suscep - tibility Testing . Washington, DC: American Societ y for Microbiolog y; 20 05. 46. Turnidge J, Paterson DL. Setting and revising antibacterial susceptibilit y break - points. Clin Microbiol Rev . 2007;20(3):391–408.