Alternating antibiotic treatments constrain evolutionary paths to multidrug resistance - PubMed (original) (raw)

Alternating antibiotic treatments constrain evolutionary paths to multidrug resistance

Seungsoo Kim et al. Proc Natl Acad Sci U S A. 2014.

Abstract

Alternating antibiotic therapy, in which pairs of drugs are cycled during treatment, has been suggested as a means to inhibit the evolution of de novo resistance while avoiding the toxicity associated with more traditional combination therapy. However, it remains unclear under which conditions and by what means such alternating treatments impede the evolution of resistance. Here, we tracked multistep evolution of resistance in replicate populations of Staphylococcus aureus during 22 d of continuously increasing single-, mixed-, and alternating-drug treatment. In all three tested drug pairs, the alternating treatment reduced the overall rate of resistance by slowing the acquisition of resistance to one of the two component drugs, sometimes as effectively as mixed treatment. This slower rate of evolution is reflected in the genome-wide mutational profiles; under alternating treatments, bacteria acquire mutations in different genes than under corresponding single-drug treatments. To test whether this observed constraint on adaptive paths reflects trade-offs in which resistance to one drug is accompanied by sensitivity to a second drug, we profiled many single-step mutants for cross-resistance. Indeed, the average cross-resistance of single-step mutants can help predict whether or not evolution was slower in alternating drugs. Together, these results show that despite the complex evolutionary landscape of multidrug resistance, alternating-drug therapy can slow evolution by constraining the mutational paths toward resistance.

Keywords: antibiotic resistance; collateral sensitivity; drug cycling; experimental evolution; multidrug therapy.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.

Fig. 1.

Experimental evolution of antibiotic resistance under multidrug treatments. (A) Each population of S. aureus was inoculated into a series of wells with a gradient of drug concentrations. After 20 h, the well with the highest drug concentration permitting bacterial growth (OD600 > 0.2) was used to inoculate the next cultures. This procedure was repeated for 22 d in 10 replicate populations per drug treatment. (B) Three antibiotics and their pairwise combinations were studied: TMP (blue), NEO (green), and CIP (red). (C) For each drug pair, we tested each single drug individually, a mixed treatment using a fixed ratio of the two drugs, and alternating treatments with daily switching between the two single drugs, starting with either drug.

Fig. 2.

Fig. 2.

Alternating drugs slows evolution of resistance to one of the two drugs. For each of the three drug pairs, (A) TMP-NEO, (B) NEO-CIP, and (C) TMP-CIP, the mean resistance (MIC) to each of the two drugs is shown both as a function of the time exposed to that drug (Left) and as a distribution at the final time point, 11 d of exposure to that drug (Right), for the single- (TMP in blue, NEO in green, and CIP in red), alternating- (black), and mixed-drug treatments (gray; resistance to single drugs measured at only final time point). Each line indicates the mean resistance of 10 single-drug or 20 alternating-drug independent populations, as inferred from the well chosen for propagation during evolution (individual trajectories vary) (Fig. S1). Error bars indicate SEM. Day 1 represents the resistance after 2 d of evolution for half of the alternating-drug populations, and may not match that of single-drug populations. Histograms reflect phenotypic measurements following 1 d of growth in the absence of drug (Materials and Methods) (may not match exactly with inferred resistance), and triangles indicate mean final phenotyped resistance. Downward-pointing colored arrows indicate statistically significant differences (P < 0.05, two-sample t test) between the final resistance levels of the single and alternating drug treatments. In all drug pairs, resistance to one of the two drugs was slower in the alternating-drug treatment than in the single-drug treatment.

Fig. 3.

Fig. 3.

Alternating drugs constrains evolution. (A) Genes with mutations in at least four populations are shown with their frequency in each of nine treatments: trimethoprim (T, shown twice), neomycin (N), ciprofloxacin (C), their pairwise alternating treatments (double-headed arrows), and mixed treatments (plus signs). Known resistance genes are indicated at the right. (B) Genes with significant enrichment (P < 0.01) for mutations in alternating treatment compared with single drugs. Black horizontal lines indicate expected mutation frequency in alternating treatment (sum of frequencies in single drugs), and statistical significance relative to this expectation was assessed with Fisher’s exact test. *P < 0.01, **P < 0.0001. (C) Genes with significant depletion (P < 0.01) of mutations in alternating treatment. *P < 0.01. Genes with only locus tag names are indicated by the numerical portion of the locus name.

Fig. 4.

Fig. 4.

Cross-resistance explains efficacy of alternating drug treatments. (A) Single-step mutants were generated by culturing wild-type cells in liquid media with antibiotic for 20 h, then isolated by pooling wells that grew to turbidity, and then plating these mutants on agar plates with enough antibiotic to prevent growth of the wild-type cells. At least 40 resistant colonies were selected for each drug and then phenotyped in all three drugs (Materials and Methods). (B) Cross-resistance of single-step TMP-, NEO-, and CIP-resistant mutants. Histograms show the distribution of resistance level (MIC) to the three drugs relative to the ancestor. The measurements of selected resistance are shown in colors and those of cross-resistance are shown in gray. Dashed lines indicate the mean resistance. Colored arrows indicate significant cross-resistance (mean change in log2 MIC > 0.25), and colored horizontal bars indicate no significant cross-resistance (mean change in log2 MIC < 0.25). (C) Cross-resistance correctly predicts effect of alternating drugs on both component drugs for the TMP-NEO and TMP-CIP drug pairs, and on NEO for the NEO-CIP drug pair (the only inconsistency is effect on CIP in NEO-CIP drug pair). Arrows indicate change in rate of evolution of resistance in alternating drugs compared with single-drug treatment, as observed during evolution experiments (Fig. 2) and as predicted by cross-resistance measurements (B). Horizontal bars indicate a prediction or observation of no change in rate of evolution of resistance.

Comment in

Similar articles

Cited by

References

    1. Taubes G. The bacteria fight back. Science. 2008;321(5887):356–361. - PubMed
    1. Levy SB, Marshall B. Antibacterial resistance worldwide: Causes, challenges and responses. Nat Med. 2004;10(12) Suppl:S122–S129. - PubMed
    1. Lipsitch M, Bergstrom CT, Levin BR. The epidemiology of antibiotic resistance in hospitals: Paradoxes and prescriptions. Proc Natl Acad Sci USA. 2000;97(4):1938–1943. - PMC - PubMed
    1. US Department of Health and Human Services . Antibiotic Resistance Threats in the United States 2013. Atlanta: Centers for Disease Control and Prevention; 2013.
    1. Lieberman TD, et al. Parallel bacterial evolution within multiple patients identifies candidate pathogenicity genes. Nat Genet. 2011;43(12):1275–1280. - PMC - PubMed

Publication types

MeSH terms

Substances

LinkOut - more resources