Differential Activity of the Combination of Vancomycin and Amikacin on Planktonic vs. Biofilm-Growing Staphylococcus aureus Bacteria in a Hollow Fiber Infection Model

Abstract

Combining currently available antibiotics to optimize their use is a promising strategy to reduce treatment failures against biofilm-associated infections. Nevertheless, most assays of such combinations have been performed in vitro on planktonic bacteria exposed to constant concentrations of antibiotics over only 24 h and the synergistic effects obtained under these conditions do not necessarily predict the behavior of chronic clinical infections associated with biofilms. To improve the predictivity of in vitro combination assays for bacterial biofilms, we first adapted a previously described Hollow-fiber (HF) infection model by allowing a Staphylococcus aureus biofilm to form before drug exposure. We then mimicked different concentration profiles of amikacin and vancomycin, similar to the free plasma concentration profiles that would be observed in patients treated daily over 5 days. We assessed the ability of the two drugs, alone or in combination, to reduce planktonic and biofilm-embedded bacterial populations, and to prevent the selection of resistance within these populations. Although neither amikacin nor vancomycin exhibited any bactericidal activity on S. aureus in monotherapy, the combination had a synergistic effect and significantly reduced the planktonic bacterial population by -3.0 to -6.0 log₁₀ CFU/mL. In parallel, no obvious advantage of the combination, as compared to amikacin alone, was demonstrated on biofilm-embedded bacteria for which the addition of vancomycin to amikacin only conferred a further maximum reduction of 0.3 log₁₀ CFU/mL. No resistance to vancomycin was ever found whereas a few bacteria less-susceptible to amikacin were systematically detected before treatment. These resistant bacteria, which were rapidly amplified by exposure to amikacin alone, could be maintained at a low level in the biofilm population and even suppressed in the planktonic population by adding vancomycin. In conclusion, by adapting the HF model, we were able to demonstrate the different bactericidal activities of the vancomycin and amikacin combination on planktonic and biofilm-embedded bacterial populations, suggesting that, for biofilm-associated infections, the efficacy of this combination would not be much greater than with amikacin monotherapy. However, adding vancomycin could reduce possible resistance to amikacin and provide a relevant strategy to prevent the selection of antibiotic-resistant bacteria during treatments.

Keywords: Staphylococcus aureus; amikacin; antibiotic combination; antimicrobial resistance; biofilm; hollow-fiber infection model; vancomycin.

FIGURE 1

Diagrammatic representation of the Hollow Fiber Infection Model kindly provided by FiberCell Systems^® (Cadwell, 2015). Bacteria were trapped by the hollow fiber capillaries in the cartridge (see also embedded photo). Drugs were added to the central reservoir and freely circulated through the cartridge and bacteria by means of the Fibercell Systems Duet pump^® (FiberCell Systems, Inc., Frederick, MD, United States). Drug concentrations decreased over time after drug administrations, due to the continuous addition of a diluent (RPMI) by means of another set of pumps (here, Mini Rythmic^® PN+, SMD, Fleury-sur-Orne, France).

FIGURE 2

Expected (blue lines) and observed (red circles) concentration-time profiles in the Hollow Fiber system from D3 to D7 for (A) vancomycin after administrations twice a day with peak concentrations of 18 μg/mL (V18 treatment) and for (B) amikacin after administrations once a day with peak concentrations of 70 μg/mL.

FIGURE 3

Mean ± SD of the bacterial counts (log₁₀ CFU/mL) for planktonic (in orange) and biofilm-embedded bacteria (in blue) at the end of the experiments (D7) for control assays and the different treatments (n = 2 for each antibiotic combination). The BEB population was smaller than the planktonic population in the control experiments, and also after monotherapy with amikacin or vancomycin. In contrast, the BEB populations were 1.2–2.0 log₁₀ CFU/mL higher than the planktonic populations (p < 0.001).

FIGURE 4

Changes in the planktonic bacterial populations (log₁₀ CFU/mL) after exposure to amikacin or vancomycin in monotherapy from D3 to D7. Full circles represent the bacterial counts in the HF model during 5 days of treatment with vancomycin twice a day (V18 treatment, in blue) or amikacin once a day (A70 treatment, in orange). Full red squares represent the bacterial counts of planktonic bacteria growing on agar supplemented with threefold MIC of amikacin over time during A70 treatment. Mean ± SD of the bacterial counts are shown (n = 2 for each treatment).

FIGURE 5

Changes in the planktonic bacterial population (log₁₀ CFU/mL) after exposure to combinations of amikacin and vancomycin from D3 to D7. The marks represent the mean ± SD of the bacterial counts for the different tested treatments [blue: A70 V18 treatment, red: A95 V18 treatment, green: A130 V18 treatment and black: A70 CRIV9 treatment (n = 2 for each antibiotic combination)]. The reduction of the planktonic bacterial population between the 1st day (D3) and the last day (D7) of treatments with combinations of amikacin and vancomycin was significant (p < 0.001).

FIGURE 6

Mean ± SD of the bacterial counts (log₁₀ CFU/mL) of total planktonic bacteria (dark orange) and bacteria growing on agar supplemented with three-times the MIC of amikacin (medium orange) and six-times the MIC of amikacin (light orange) in the control experiment or after 5 days of exposure to different treatments (D7) (n = 2 for each condition).

FIGURE 7

Mean ± SD of the bacterial counts (log₁₀ CFU/mL) of total biofilm-embedded bacteria (dark blue) and of biofilm-embedded bacteria growing on agar supplemented with three-times the MIC of amikacin (blue) and six-times the MIC of amikacin (light blue) in the control experiment or after 5 days of exposure to different treatments (D7) (n = 2 for each condition).