Quinolone efflux pumps play a central role in emergence of fluoroquinolone resistance in Streptococcus pneumoniae

Abstract

The preferential use of older antimicrobial agents is, in general, sound public health policy and is meant to maintain susceptibility to newer agents. In the case of fluoroquinolones, however, this strategy is flawed and may actually hasten the spread of Streptococcus pneumoniae strains resistant to newer members of the class. In a mouse thigh infection model, we were unable to isolate clones of pneumococci resistant to the newer fluoroquinolone levofloxacin at 2 x or 4 x the baseline MIC. An initial exposure in vivo to the older agent, ciprofloxacin, allowed straightforward selection of clones resistant to levofloxacin in a subsequent experiment. The original ciprofloxacin exposure generated clones without changes in the parC/E and gyrA/B quinolone target sites almost exclusively but did allow overexpression of a reserpine-responsive pump. While this caused only minimal change in the levofloxacin MIC (0.6 mg/liter to 0.8 mg/liter), it allowed a major change in the mutational frequency to resistance for levofloxacin (<1/10(8.5) to approximately 1/10(4.5)), which allowed levofloxacin-resistant clones to be isolated in a subsequent in vivo experiment. The reason underlying ciprofloxacin's propensity to select for pump-overexpressed clones is likely related to its hydrophilicity. To preserve the susceptibility of Streptococcus pneumoniae to newer members of the class of quinolones, use of ciprofloxacin for community-acquired respiratory infections should be minimized.

FIG. 1.

(A) Dose-response relationship between levofloxacin and changes in S. pneumoniae (strain AMC-058) density in thigh muscles (mean ± 1 standard deviation) following ∼10⁸-CFU/thigh inoculation. In this evaluation, all doses were given on a once-daily basis. In neither experiment (A and B) were any drug-resistant mutants recovered by direct plating on levofloxacin-contaning plates. (B) Dose fractionation experiments were performed (QD refers to the whole dose given once; BID refers to half the dose given every 12 h twice; QID refers to one-quarter of the dose given every 6 h). The data are displayed with the AUC/MIC ratio as the independent variable. The peak/MIC ratio and time > MIC are also evaluated as independent variables, but AUC/MIC displayed the best fit of the model to the data. The inoculum was 6.5 log₁₀ CFU/thigh. When tested by analysis of variance, no difference was seen between q24h (the whole dose once), q12h (a half dose q12h), and q6h (a quarter dose q6h) administration of the same total daily dose.

FIG. 2.

Comparison of efflux rates of EtBr from the parent strain (AMC-058) and its isogenic efflux-mediated-resistance daughter (RC2). The change in fluorescence is normalized to the initial observed levels for direct comparison of initial EtBr efflux rates. The time scale is in seconds. The error bars indicate standard deviations.

FIG. 3.

In vivo outcome study results for S. pneumoniae AMC-058 with pharmacodynamically equivalent exposures to ciprofloxacin (CIP) and levofloxacin (LVX). Total bacterial population densities (black) enumerated on drug-free plates and drug-resistant-subpopulation (CIPr and LVXr) densities (gray) isolated on 3 × MIC-containing plates are shown for ciprofloxacin (A) and levofloxacin (B). Levofloxacin-exposed organisms were plated on levofloxacin-containing plates, and ciprofloxacin-exposed organisms were plated on ciprofloxacin-containing plates. In panel B, no clones resistant to levofloxacin were isolated; this is displayed by plotting at the detection limit. The error bars indicate standard deviations.

FIG. 4.

In vivo outcome study results for S. pneumoniae RC2 exposed to pharmacodynamically equivalent exposures of ciprofloxacin (CIP) and levofloxacin (LVX). Total bacterial population densities (black) enumerated on drug-free plates and drug-resistant-subpopulation (CIPr and LVXr) densities (gray) isolated on 3 × MIC-containing plates are shown for ciprofloxacin (A) and levofloxacin (B). Levofloxacin-exposed organisms were plated on levofloxacin-containing plates, and ciprofloxacin-exposed organisms were plated on ciprofloxacin-containing plates. The error bars indicate standard deviations.