Heteroresistance (HR) is a phenomenon in which minority populations of antibiotic resistant bacteria are maintained in populations dominated by cells susceptible to that antibiotic. We have shown that HR is often undetected by clinical diagnostics, but that the subpopulations of resistant cells present in HR can lead to in vivo treatment failure. It is critical to gain a thorough understanding of HR in order to design more effective and sensitive diagnostics for its detection, and to guide clinical treatment. One major unaddressed area of investigation is which parameters control the dynamics of the resistant subpopulations in HR. The goal of Project 3 is to elucidate the dynamics of resistant subpopulations through experimental testing, supported by quantitative modeling of the pharmaco-, population- and evolutionary dynamics of heteroresistant bacteria. Importantly, this will include studies of both their dynamics upon antibiotic treatment as well as the reversion of the population toward susceptibility upon removal of the drug. Toward this end we will develop and analyze the properties of mathematical models of HR based on the mechanisms of heteroresistance derived from molecular, genetic and single-cell microfluidic studies in Projects 1 and 2. The parameters used for the numerical analyses of the properties of these models will be estimated with clinical isolates of HR Enterobacteriaceae (Enterobacter, Escherichia, Klebsiella) and Acinetobacter baumannii obtained from Core B and studied in depth in Projects 1 and 2. For each subpopulation we will estimate: (i) parameters of comprehensive pharmacodynamic functions, (ii) the rates of transition between susceptible and resistant states, and (iii) the fitness costs of these resistant states. Using Hollow Fiber Bioreactors, batch culture, and microfluidics, we will evaluate how well the models, with independent estimates of their parameters, fit the pharmacodynamic of HR-bacteria confronted with antibiotics and, with continuous culture devices, how well these models account for the dynamics of drug treatment of heteroresistant infections. In serial transfer culture we will estimate the rates of transition to baseline susceptible states following the removal of the antibiotics. Based on the results of these experiments, and in an iterative process, the models will be modified to make them more accurate and predictive analogs of the pharmacodynamics of HR. For each HR isolate studied, we will also perform experiments to determine if the frequency of the resistant subpopulations change as a consequence of antibiotic-mediated selection (i.e. if the baseline frequency of the resistant cells increases), and elucidate if there are conditions under which HR will be replaced by permanent resistance. The results from this research will for the first time, provide a broad and detailed understanding of the dynamics of HR, facilitating an understanding of the parameters that control the frequency of the resistant subpopulations. These studies will have a major impact on the development of diagnostic procedures to detect HR and the design of protocols for treating infections with bacteria that exhibit HR to the treating antibiotic.
