The worldwide spread of antibiotic resistance in bacterial pathogens has become a major public health problem. Factors presumed to prevent or slow the evolution of antibiotic resistance include the limitation of mutations conferring the resistance and the fitness costs associated with the resistance (Andersson 2006; Read and Huijben 2009; MacLean et al. 2010; Andersson and Hughes 2011; Cantón and Morosini 2011; Hermsen et al. 2012). However, antibiotic-resistant strains often emerge very rapidly under drug treatment, due to the typically large population sizes of bacteria and the presence of strains with intrinsic or inducible high mutation rates (Chopra et al. 2003; Henrichfreise et al. 2007; Couce and Blazquez 2009; Kohanski et al. 2010; Perron et al. 2010; Weigand and Sundin 2012). Meanwhile, it has been shown that the fitness costs of antibiotic resistance may sometimes be compensated for by additional mutations, and thus, the antibiotic-resistant strains not only spread quickly during drug treatment, but also often remain persistent or are only very slowly outcompeted by their susceptible relatives after drug use has been reduced (Schrag et al. 1997; Levin et al. 2000; Maisnier-Patin et al. 2002; Gagneux et al. 2006; Perron et al. 2010; Andersson and Hughes 2011).
In the past decade, the use of bacteriophages (phages) has received much attention as an alternative to antibiotic therapy, although bacteria may also readily evolve resistance to phages (Levin and Bull 1996; Barrow and Soothill 1997; Chanishvili et al. 2001; Summers 2001; Thiel 2004; Cairns and Payne 2008; Kutateladze and Adamia 2010; Kutter et al. 2010; Monk et al. 2010; Pirnay et al. 2011; Escobar-Páramo et al. 2012). Recently, combined use of antibiotics and phages has been shown to greatly reduce the chance of resistance evolution as there is typically little cross-resistance to phages and antibiotics (Chanishvili et al. 2001; Kutateladze and Adamia 2010; Zhang and Buckling 2012). Resistance to phages may also impose fitness costs on the bacteria, in particular when the bacteria and the phages show an evolutionary arms race in defense and counter defense (Levin and Bull 1996; Bohannan et al. 1999; Bohannan and Lenski 2000; Buckling et al. 2006; Brockhurst et al. 2007; Perron et al. 2007; Forde et al. 2008; Koskella et al. 2012). We hypothesize that treatment with phages may impact the evolution of bacterial resistance to antibiotics, as the fitness costs of resistance to phages may add to those of antibiotic resistance, further reducing the growth performance of antibiotic-resistant bacteria.
Here, we use an in vitro experimental system to address whether exposure to phages impacts the evolution of high-level bacterial antibiotic resistance (resistance to high-dose antibiotics) in heterogeneous drug environments. Mutations conferring resistance to low-dose antibiotics are often of high supply rates; whereas resistance to high-dose antibiotics, which is more medically relevant, may only be conferred by a combination of multiple mutations (Weinreich et al. 2006; Lozovsky et al. 2009; Read et al. 2011; Toprak et al. 2012). However, such a mutation supply limitation for high-level antibiotic resistance can be relaxed in spatially or temporally heterogeneous drug environments, where the low-dose drug environments can select for low-level antibiotic-resistant mutants that function as stepping-stones for the evolution of high-level resistance (Baquero and Negri 1997; Olofsson et al. 2005; Perron et al. 2006; Couce and Blazquez 2009; Cantón and Morosini 2011; Greulich et al. 2012; Habets and Brockhurst 2012; Hermsen et al. 2012). In such environments, fitness costs suffered by the antibiotic-resistant mutants may be an important constraint on high-level resistance evolution; and exposure to phages may reduce the level of antibiotic resistance by imposing further fitness costs (of resistance to phages) on the bacteria. In this study, we also examined the reversion of antibiotic resistance after drug use being terminated. The bacterium Pseudomonas fluorescens SBW25 and its lytic phage SBW25Φ2 were used as the experimental system. The bacterium and the phage can exhibit an antagonistic coevolution. On most occasions, the phage cannot drive the bacterial populations to very low densities or extinct (Vogwill et al. 2009; Escobar-Páramo et al. 2012; Zhang and Buckling 2012; Harrison et al. 2013), and thus, the phage alone does not function as a good antibacterial. However, resistance to the phage causes significant fitness costs on the bacterium (Buckling et al. 2006; Lopez-Pascua and Buckling 2008), and such fitness costs may impact the evolution of bacterial resistance to antibiotics.