Abstract
Introduction
T
The IE has been studied mostly in staphylococci using β-lactam antibiotics, where the IE is typically induced by the antibiotic hydrolysis by bacterial β-lactamases.18,22 Classically, comparing the conventional MIC with MIC of higher inocula assesses the IE. However, to the best of our knowledge, no studies have assessed the importance of the IE in the selection of antimicrobial resistance during prolonged antimicrobial exposure.
In this study, we investigated the IE of a methicillin-resistant Staphylococcus aureus (MRSA) and its importance in the selection of antimicrobial resistance to vancomycin, daptomycin, and linezolid in vitro. These three antibiotics have different modes of action and are commonly used for the treatment of MRSA infections. Evolution of MICs of MRSA with low (104 CFU/ml), medium (106 CFU/ml), and high (108 CFU/ml) inocula was assessed during 15 days of cycling experiments with stepwise twofold increasing concentrations of either antibiotic. In parallel, conventional MICs were performed at days 1, 5, 10, and 15 for bacteria of each inoculum to discriminate IE and resistance selection.
Materials and Methods
Bacterial strain and growth conditions
An MRSA ATCC 43300 strain was used for this study. Bacteria were stored at −80°C in a cryovial bead preservation system (Microbank; Pro-Laboratory Diagnostics, Richmond Hill, Canada). Overnight cultures for bacterial inocula were prepared in MHB (Oxoid, Basingstoke, United Kingdom) from a single colony picked on a brain–heart infusion agar plate and incubated at 37°C for 16 hr.
Antimicrobial agents
Vancomycin was purchased from Teva Pharm AG (Aesch, Switzerland), daptomycin from Novartis Pharma AG (Bern, Switzerland), and linezolid from Pfizer AG (Zürich, Switzerland). For daptomycin studies, MHB was supplemented with Ca++ to a final concentration of 50 μg/ml.
Susceptibility studies
MICs at the beginning of the experiment were determined by broth macrodilution in MHB according to standard guidelines. 4 Briefly, series of tubes containing 2 ml of MHB with twofold increasing antibiotic concentrations were inoculated with a defined bacterial inoculum and incubated for 16 hr at 37°C. A tube without antibiotic was used as positive growth control. MICs were defined as the lowest concentrations of drug inhibiting visual bacterial growth.
Conventional MICs were determined with a standard inoculum of 105 CFU/ml. MICs for low (MICL), medium (MICM), and high (MICH) inocula were determined with inocula of 104, 106, and 108 CFU/ml, respectively. Inocula sizes were verified by conventional viable counts throughout the experiments.
IE and antibiotic resistance selection
For each drug, comparison of MICL, MICM, and MICH at day 1 indicated whether there was or not an IE. Then, to assess the role of the inoculum size on the selection of antibiotic resistance, the three different inocula (low, medium, or high) were maintained through 15 days of antibiotic exposure in a cycling experiment.
Starting at day 1 for each drug and each inoculum, bacteria from the tube containing the highest concentration of drug still showing visual turbidity after 16 hr of incubation at 37°C were used to inoculate the new series of tubes for the next cycle with the adequate inoculum (104, 106, or 108 CFU/ml). Typically, for each antibiotic and inoculum, the range of antibiotic concentrations (in twofold serial dilutions) tested at the next cycle encompassed the MIC measured at the previous cycle by three to four dilutions below and above. A tube without antibiotic was included as positive growth control. Importantly, to prepare the inoculum for each cycle, bacteria were washed three times in NaCl 0.9% before inoculation in order to avoid any antibiotic carryover effect. MICL, MICM, and MICH were then measured daily. All experiments were performed in triplicate.
To monitor the selection of antibiotic resistance for each inoculum size, conventional MICs were also performed at days 1, 5, 10, and 15 using a standard inoculum of 105 CFU/ml. The MICs were then classified into susceptible (S) or resistant (R) categories according to EUCAST clinical breakpoints for vancomycin (S≤2 μg/ml; R>2 μg/ml), daptomycin (S≤1 μg/ml; R>1 μg/ml), and linezolid (S≤4 μg/ml; R>4 μg/ml).
Results
Conventional susceptibility studies
Conventional MICs (105 CFU/ml inoculum) of the MRSA ATCC 43300 to vancomycin, daptomycin, and linezolid at the beginning of the experiment were 1, 0.25, and 1 μg/ml, respectively.
IE and selection of vancomycin resistance
MIC ranges at day 1 for vancomycin were MICL=1 μg/ml, MICM=1–2 μg/ml, and MICH=2 μg/ml, indicating a small IE with a two times increase in MIC between the low and high inocula.
During the 15 days of vancomycin exposure, MICs increased for all three inocula in all experiments (Fig. 1). MIC ranges at day 15 were MICL=2–4 μg/ml, MICM=4–8 μg/ml, and MICH=4–16 μg/ml.
Vancomycin, daptomycin, and linezolid minimum inhibitory concentrations (MICs) of low (104 CFU/ml), medium (106 CFU/ml), and high (108 CFU/ml) inocula of Staphylococcus aureus ATCC 43300 during 15 days of cycling. The results of three independent cycling experiments are shown.
Conventional MICs at days 1, 5, 10, and 15 for the three experiments are shown in the Table 1. In one of the replicates, bacteria of the high inoculum were resistant since day 5, whereas in two of the replicates, bacteria of the medium and high inocula were resistant since day 10. In one of the replicates bacteria of all three inoculum sizes were resistant at day 15.
MICs were then classified as susceptible (S) or resistant (R; in bold) according to EUCAST breakpoints for vancomycin (S≤2 μg/ml; R>2 μg/ml), daptomycin (S≤1 μg/ml; R>1 μg/ml), and linezolid (S≤4 μg/ml; R>4 μg/ml).
EUCAST, European Union Committee on Antimicrobial Susceptibility Testing; MIC, minimum inhibitory concentration; nd, not done.
IE and selection of daptomycin resistance
MIC ranges at day 1 for daptomycin were MICL=0.25 μg/ml, MICM=0.25–0.5 μg/ml, and MICH=2 μg/ml, indicating an important IE with an eight times increase in MIC between the low and high inocula.
During the 15 days of cycling, MICs increased for all three inocula in all experiments, but more significantly for the medium and high inocula (Fig. 1). MIC ranges at day 15 were MICL=0.5–2 μg/ml, MICM=8–128 μg/ml, and MICH=64–256 μg/ml. This indicated a rapid and drastic diminution in the daptomycin susceptibility of the MRSA along the cycling, especially when a medium or high inoculum was tested.
Conventional MICs at days 1, 5, 10, and 15 for the three experiments are shown in Table 1. In all experiments, bacteria of the high inoculum were already resistant since day 5. Bacteria of the medium inocula were resistant at days 5, 10, or 15 depending on the replicate. In one of the replicates, bacteria of the low inoculum were resistant at day 15.
IE and selection of linezolid resistance
MIC ranges at day 1 for linezolid were MICL=1 μg/ml and MICM=1–2 μg/ml, indicating that there was no initial IE for this drug. According to our experimental setting for MIC determination (i.e., inhibition of visual growth after 16–18 hr of incubation at 37°C), no MICH could be determined when the high inoculum was used, even with concentrations of linezolid up to 2,048 μg/ml. Even if a decrease in turbidity was observed in comparison to the control without antibiotic, there was no complete inhibition of growth for the high inoculum. Therefore, tubes with a high inoculum were not further tested during the cycling experiment. Of note, this was repeated with two other staphylococcal strains, yielding the same elevated MIC when the high inoculum was used (data not shown).
During the 15 days of cycling, MICs remained low in all experiments, with only a small increase in the MIC of the medium inoculum in the first and third experiments (Fig. 1). MIC ranges at day 15 were MICL=1 μg/ml and MICM=2–4 μg/ml in the first experiment and MICL=1 μg/ml and MICM=2 μg/ml in the second experiment.
Conventional MICs at days 1, 5, 10, and 15 for the three experiments are summarized in Table 1 and indicated that bacteria of the low and medium inocula remained susceptible to linezolid throughout the assay.
Discussion
Since the U.S. food and drug administration first approved the use of vancomycin in 1958, it has been the antibiotic of choice for the treatment of invasive MRSA infections.13,15 As could be expected, isolates exhibiting reduced susceptibility to vancomycin have evolved.16,21 New anti-MRSA drugs, such as daptomycin5,10,24,25 and linezolid,8,12 have been developed but resistance toward these antimicrobials has also been described, particularly in patients receiving prolonged treatment durations or with high bacterial load infections.2,9,11,26,28,29 In our study, we wanted to assess the importance of the inoculum size and duration of antibiotic exposure to resistance development of these three drugs.
Our results indicated that vancomycin exhibited only a moderate initial IE in MRSA, but that a prolonged stepwise exposure of bacteria to vancomycin was sufficient to select for vancomycin resistance independently of the initial inoculum size at the end of the experiment.
When daptomycin was used, a significant IE was observed throughout the cycling experiment. More importantly, when the medium or high inocula were maintained throughout the 15 days of daptomycin exposure, drug susceptibility rapidly diminished and high daptomycin resistance levels appeared. In contrast, the low inoculum consistently remained at a low MIC level. This indicates that a minimal bacterial density was required for daptomycin resistance development.
Regarding linezolid, stepwise exposure of the low and medium inocula showed that the MRSA kept its susceptibility toward linezolid throughout the 15 days of the experiment. On the other hand, resistance development could not be assessed for the high inoculum since even high concentrations of the drug did not inhibit visual growth. This could be explained by a slow bacteriostatic activity of linezolid that allows bacteria to go through a few replication cycles before coming to a halt. 30
The exact mechanisms underlying the IE are not fully understood but probably include a reduced effective concentration of the antibiotic typically by denaturing enzymes or binding of the antibiotic to other cell structures than its target, resulting in a reduced ratio of available drug molecules per target. 27 Vancomycin and daptomycin resistance mechanisms essentially imply the progressive accumulation of various single-nucleotide mutations in genes involved in cell envelope synthesis or central regulators,1,14 whereas linezolid resistance arises through mutations in the 23S rRNA or acquisition of a cfr gene. 17 We observed for both daptomycin and vancomycin a correlation between inoculum size and duration of antibiotic exposure in the selection of antibiotic resistance. This was especially the case for daptomycin at medium and high inocula. Interestingly, a number of reports show the selection of daptomycin resistance under treatment, and this seems to be especially the case for high inoculum infections.6,10,12,19 For the low inocula, daptomycin exhibits a rapid bactericidal effect that probably sufficiently lowers the probability of resistance development since daptomycin spontaneous resistance rate is inferior to 1×10−10 in vitro. 20
Importantly, when the inoculum was low no selection of resistance was observed except for vancomycin. This highlights the importance of surgical drainage or infection source removal in high bacterial density infections, such as abscesses, endocarditis, or implant infections, where bacterial density can reach 109 CFU/ml of pus or gram of tissue, or the need to use antibiotic combinations in order to prevent the emergence of resistance. 7
Footnotes
Acknowledgments
The authors thank Dr. José M. Entenza for helpful comments on the article. This study was supported in part by an unrestricted educational grant, Antibacterial Europe ASPIRE Research Awards 2010, from Pfizer International Operations (to A.B.).
Disclosure Statement
Pfizer International Operations, which funded in part this work through an unrestricted educational grant, has not played any decision-making role in the research presented in this study.