Towards optimising polymyxin therapy against infections caused by multidrug-resistant Acinetobacter baumannii

The increasing incidence of infections caused by Gram-negative ‘superbugs’ is emerging as one of the most significant threats confronting public health in the modern era. This crisis has been compounded by the decline in antimicrobial drug discovery research over the last 30 years that has resulted in a lack of novel antimicrobial agents in late-stage clinical development. It is against this backdrop that polymyxins, first introduced in the 1950s, have emerged as last-line therapeutic agents for the treatment of recalcitrant infections caused by Gram-negative pathogens, such as Acinetobacter baumannii. As a result of having been introduced prior to modern drug development practices, polymyxin dosing strategies have not been adequately investigated or optimised. This thesis combines in vitro experimental models, mathematical modelling and high-throughput sequencing (HTS) to characterise polymyxin activity against A. baumannii, and address the dearth of information surrounding the link between polymyxin pharmacokinetics (PK) and pharmacodynamics (PD). In addressing the aims and hypotheses of this thesis, new methods were developed for quantifying polymyxins in bacterial growth media, and mathematically defining antimicrobial activity in in vitro time-kill studies. Further, the bactericidal activity of polymyxin was found to be dependent on the rapid attainment of therapeutic concentrations. These findings carry significant clinical implications, given that the two clinically-used polymyxins, colistin and polymyxin B, exhibit markedly different concentration-time profiles in critically-ill patients. The gradual accumulation of colistin, seen in patients following administration of its inactive prodrug colistin methanesulphonate, resulted in little antimicrobial activity. Notably, these studies consistently highlighted the development of polymyxin resistance during treatment, even with aggressive dosing strategies. Mechanism-based mathematical modelling implicated the involvement of adaptive (i.e. non-stable) and constitutive (i.e. stable) polymyxin resistance in the phenotype observed. Further in vitro experimental studies, aided by HTS to characterise genomes and transcriptomes of polymyxin-resistant A. baumannii, confirmed the presence of non-stable and stable resistance. Together, the results presented in this thesis demonstrate the utility of combining in vitro and in silico strategies for investigating antimicrobial PK/PD. Our findings also point to combination therapy as an avenue for addressing antimicrobial resistance