Christiane Goulart and Miriam Barlow and colleagues have published an important study in this month’s PLOS One entitled “Designing Antibiotic Cycling Strategies by Determining and Understanding Local Adaptive Landscapes.”
Because of the growing problem of antibiotic resistance, antibiotic cycling has been proposed as a strategy to preserve antibiotic susceptibility of disease-causing microbes. Unfortunately, these efforts have largely been disappointing. In 2004 Carl Bergstrom argued that antibiotic cycling was likely to be ineffective in the hospital setting, based on a computational approach informed by natural selection. At that time, Bergstrom left open the possibility that a cycling regime could work, but he warned that the theoretical underpinnings of such an approach had not been demonstrated.
Goulart, Barlow and colleagues have satisfied that precondition with their compelling recent study. Miriam Barlow’s group at University of California at Merced has been studying the evolution of beta-lactamases, the most commonly encountered bacterial enzymes responsible for antibiotic resistance, since the early 2000s. In this recent work, Goulart et al. propose that earlier antibiotic cycling strategies failed to slow resistance evolution because they relied on random genetic drift for reversion to ancestral alleles and because they used unrelated antibiotics. In a proof of principle study, Goulart et al. show that selection-informed cycling can exploit fitness differences conferred by beta-lactamase gene variants, favoring reversion to the ancestral resistance allele, known as blaTEM-1.
In their recent study, Goulart et al. examined bacteria with combinations of the four mutations that separate the ancestral blaTEM-1 from the evolutionarily derived variants blaTEM-50 and blaTEM-85. The researchers generated adaptive landscapes based on the fitness differences of bacteria with sequential changes in the mutations that distinguish the ancestral from the derived beta-lactamases.
The researchers were able to demonstrate pathways in adaptive landscapes showing step-wise increments in fitness between ancestral blaTEM-1 and blaTEM-85 under selection pressure from certain antibiotics. When a single antibiotic was considered by itself, one or two pathways to fixation of blaTEM-85 were apparent. However, when fluctuating levels of multiple antibiotics were modeled, there were over 15,000 pathways to fixation of blaTEM-85. These results match Bergstrom’s (2004) findings that random changes in the deployment of antibiotics (which may more accurately reflect actual prescribing patterns) speeds up resistance evolution.
In an exciting finding, Goulart et al. were able to identify a multitude of pathways in which the direction of evolution could be reversed, resulting in reversion to the ancestral allele, blaTEM1. These involved cycling of various combinations of antibiotics. They write: “These results indicate that there are numerous routes for resistance to be reversed when those three antibiotics are cycled, which is an indication that this approach is robust.”
The sophisticated approach that Goulart et al. demonstrate here – using empirical measurements of fitness of bacteria exposed to antibiotics in combination with computational modeling using adaptive landscapes – holds promise in generating testable strategies of antibiotic cycling. Such techniques are an absolute necessity in light of current trends of resistance evolution in our hospitals and clinics.
Read the original study here: Designing Antibiotic Cycling Strategies by Determining and Understanding Local Adaptive Landscapes