The evolution of resistance to multidrug antibiotic therapies.

Abstract

The purpose of this thesis is to explore the interaction between antibiotics at sub-lethal doses, and E.coli. Initially we focussed on pairwise antibiotic interaction, and the potential to exploit these interactions to minimise antibiotic resistance. In testing the hypothesis that antagonism will slow adaptation by reducing selection for resistance we determined that there are conditions in which this fails to be the case. We furthermore caution against treating drug interactions as anything other than a dynamic property of the bacteria-drug interaction, by showing that the relationship between two drugs may be both synergistic and antagonistic depending on a variety of factors. Whilst exploring the adaptive response to drug combinations we discovered a highly unusual effect of Doxycycline to act as a growth stimulant to E.coli AG100. Chapter 3 and 4 are devoted to determining the nature and mechanism of this stimulation, and analysing any potential genomic changes using whole genome re-sequencing. Having shown that dose response is not always a monotone function of increasing drug dose, in chapter 5 we also look at the dose response in a diffusive context, using a custom built imaging system to show the common non-monotonicity of disk diffusion type assays, that manifest themselves as bullseye patterns of growth. We use a mathematical model to explore the ecological and adaptive reasons for such patterns. Finally in chapter 6 we look at the coevolutionary history of phage and E.coli REL606 strains, by determining trade-offs caused by lambda phage and the sole carbon source maltotriose both utilising the same porin (lamB) for cell entry.