Resistance Evolution against Phage Combinations Depends on the Timing and Order of Exposure

Abstract

Phage therapy is a promising alternative to chemotherapeutic antibiotics for the treatment of bacterial infections. However, despite recent clinical uses of combinations of phages to treat multidrug-resistant infections, a mechanistic understanding of how bacteria evolve resistance against multiple phages is lacking, limiting our ability to deploy phage combinations optimally. Here, we show, using Pseudomonas aeruginosa and pairs of phages targeting shared or distinct surface receptors, that the timing and order of phage exposure determine the strength, cost, and mutational basis of resistance. Whereas sequential exposure allowed bacteria to acquire multiple resistance mutations effective against both phages, this evolutionary trajectory was prevented by simultaneous exposure, resulting in quantitatively weaker resistance. The order of phage exposure determined the fitness costs of sequential resistance, such that certain sequential orders imposed much higher fitness costs than the same phage pair in the reverse order. Together, these data suggest that phage combinations can be optimized to limit the strength of evolved resistances while maximizing their associated fitness costs to promote the long-term efficacy of phage therapy.IMPORTANCE Globally rising rates of antibiotic resistance have renewed interest in phage therapy where combinations of phages have been successfully used to treat multidrug-resistant infections. To optimize phage therapy, we first need to understand how bacteria evolve resistance against combinations of multiple phages. Here, we use simple laboratory experiments and genome sequencing to show that the timing and order of phage exposure determine the strength, cost, and mutational basis of resistance evolution in the opportunistic pathogen Pseudomonas aeruginosa These findings suggest that phage combinations can be optimized to limit the emergence and persistence of resistance, thereby promoting the long-term usefulness of phage therapy.

Keywords: Pseudomonas aeruginosa; bacteriophage therapy; bacteriophages; evolutionary biology; resistance evolution.

FIG 1

Trade-offs in the strength of resistance to each phage combination. Strength of resistance, measured as reduction in bacterial growth (RBG ± standard error [SE]), such that 1 indicates equal growth in the presence and absence of phage. Selection treatment for focal resistance is indicated by color (see individual keys). Background shading indicates the mean resistance strength for the first step in sequential resistance treatments to highlight changes in resistance to the first-step phage in the second step of sequential selection treatments.

FIG 2

Relative fitness of resistant mutants is determined by selection regime. Fitness is measured as the maximum growth rate achieved during a 48-h growth period and given relative to that of the ancestral genotype; a value above 1 (light gray background) indicates lack of fitness costs associated with resistance mutation(s). Data points indicate the means of three technical replicates (±SE) measured for each set of biological replicates (plot headings indicate ancestral genotype used in the fluctuation test). Selection treatment for focal resistances is indicated by color (see individual keys). The mean relative fitness of simultaneous selection treatments is shown as a dashed purple line, with light purple shading to indicate standard error. Missing data indicate that no resistant mutants were recovered for that replicate treatment.

FIG 3

Treatment regimes determine the frequency and type of resistance mutations selected. Total number of identified mutations per resistant mutant across phage selection treatments (plot headings) under single (A), sequential (B), or simultaneous (C) selection. Colors indicate category of mutational target (see key).

FIG 4

Contrasting fitness costs resulting from specific combinations of single and double mutations. Fitness relative to the ancestor, where a value of 1 indicates equal fitness and a value below 1 indicates a fitness cost associated with labeled single (left panel) or double (right panel) mutations. The dashed line indicates predicted additive cost of two mutations affecting different receptors, given as the sum of the mean cost of individual receptor-specific mutations (gray band; single LPS mutations RF = 0.886, single type IV pilus-specific mutations RF = 0.880).