The general goal of this proposal is to start the analysis of the evolutionary dynamics of antibiotic resistance in order to evaluate strategies for reversing the increase in antibiotic resistance among human pathogens. There are three parts to this proposal that are logically connected, but are separate in both models and experimental systems. The first is the population dynamics of resistance in vitro, the second, the dynamics in vivo, and the third involves modeling the epidemiology of resistance. The first part will investigate the reality of the model for bacterial growth (equation 2 on page 18) where m(A) is the function that describes how an antimicrobial agent at concentration A effects growth. One possible function of m(A) is (n+1)f(A) where f(A) is the fraction of cells killed each cell cycle. The effect of the concentration of antibiotic on growth will be investigated using batch culture. The model is adapted to chemostat culture (equation 5, page 20). The chemostat culture is designed such that there is a large population of resistant cells, with only a small fraction of sensitive cells. When the antibiotic is added, the number of sensitive cells will decline, but the environment of the chemostat, which is determined by the majority population, will not change. It is assumed that the function f(A) will be determined from batch culture and this will then be used to predict the growth rate in chemostats. The antibiotic resistance studied is chromosomal--resistance arising from mutation rather than transfer via a plasmid. The cost of this resistance when competing with the sensitive will be determined and included in the model if required. The second part studies the effect of the immune response and antimicrobial action on bacterial growth in mice. The immune response is divided into two parts--the non-specific and antigen-specific immune responses. These will be distinguished by using SCID mice, which have only the non-specific and isogenic control mice which have both. The growth rate of the bacterial population (L. monocytogenes) will be monitored in the spleen of the mouse. This growth rate will be determined for the two types of mice with various dosing of antibiotics. Here the concentration of antibiotics is a different variable than in the first part. In the first part, the concentration is the concentration experienced by the bacteria; in this part the concentration is the concentration injected into the mouse. In the model for this section, the effect of antibiotics on growth rate is assumed to have an interaction between growth rate and antibiotic concentration, m(r,A). The third part involves development of a model as given in figure 7. This properties of this model will be investigated using the results and data from the other two sections. In particular, Anita will see if the data from S. pneumoniae where there seems to be a threshold of about 40% resistant can be explained.
