Impact of resistance selection and mutant growth fitness on the relative efficacies of streptomycin and levofloxacin for plague therapy

Abstract

Yersinia pestis, the bacterium that causes plague, is a potential agent of biowarfare and bioterrorism. The aminoglycoside antibiotic streptomycin is the gold standard for treatment. However, this recommendation is based on scant animal and clinical data. We used an in vitro pharmacodynamic infection model to compare the efficacies of 10-day regimens of streptomycin versus the fluoroquinolone antibiotic levofloxacin for the treatment of Y. pestis infection and to evaluate for emergence of resistance. The human serum concentration-time profiles for standard clinical regimens of 1 g of streptomycin given every 12 h and 500 mg of levofloxacin given every 24 h were simulated. The growth fitness of drug-resistant mutants was examined in neutropenic and immunocompetent mouse thigh infection models. In the in vitro infection system, untreated bacteria grew from 10(7) to 10(10) CFU/ml. Streptomycin therapy caused a 10(5) CFU/ml reduction in the number of bacteria over 24 h, followed by regrowth with streptomycin-resistant mutants. Levofloxacin resulted in a 10(7) CFU/ml reduction in the number of bacteria within 12 h, ultimately sterilizing the culture without resistance selection. In both the normal and neutropenic mouse infection models, streptomycin-resistant and wild-type strains were equally fit. However, 90% of levofloxacin-resistant isolates, cultured from the control in vitro infection arm, did not proliferate in the mouse models. Thus, the fluoroquinolone antibiotic levofloxacin was superior to streptomycin in our in vitro infection model. The majority of levofloxacin-resistant mutants were less fit than streptomycin-resistant and wild-type Y. pestis.

FIG. 1.

Schematic of the in vitro pharmacodynamic hollow-fiber infection model.

FIG. 2.

Pharmacokinetic simulations for (A) levofloxacin given orally at 500 mg every 24 h and (B) streptomycin given i.v. at 1 g every 12 h. The solid lines represent the targeted antibiotic concentration-time profiles, and the circles represent the measured drug concentrations over the first 48 h of the study.

FIG. 3.

Effects of streptomycin and levofloxacin therapies on the total Y. pestis population and the mutant populations with ≥3× MICs to (A) streptomycin and (B) levofloxacin. Streptomycin was given to simulate the human serum concentration-time profile for 1 g i.v. every 12 h, and levofloxacin dosing simulated the human serum concentration-time profile for levofloxacin given orally at 500 mg every 24 h. The studies with streptomycin and levofloxacin therapy were conducted simultaneously. The results are presented in two graphs for clarity.

FIG. 4.

Growth fitness of the parent Y. pestis strain ΔCO92 (streptomycin MIC, 2 mg/liter; levofloxacin MIC, 0.06 mg/liter), 2 of 10 streptomycin-resistant mutants (streptomycin MIC, 8 mg/liter), and 3 of 10 levofloxacin-resistant (MIC, 1 mg/liter) mutants in (A) neutropenic and (B) immune normal mouse thigh infection models. The error (standard deviation) bars are not shown for clarity. At the 7-day time point, the difference in bacterial densities in tissues of mice infected with the parent strain and each of the less fit levofloxacin-resistant mutants was significant (P < 0.01).