Image Segmentation and Modelling of Host-Pathogen Dynamics of Salmonella

Abstract

The pathogenic bacterium Salmonella enterica has recently been shown to hyper-replicate (HR) within the cytoplasm of epithelial cells. To better understand the mechanisms of HR, time-lapse fluorescence microscopy images, taken by our experimental collaborators, of Salmonella infecting HeLa cells were segmented. We have developed a robust approach to identifying the presence of the HR phenotype cell-by-cell and, across three independent experiments, found that HR fraction is independent of bacterial load. Related to Salmonella invasion dynamics, but separate from the hyper-replication phenomenon, a mean-field model of infection was developed. The model involves a set of coupled differential equations governed by rates of interactions between a population of hosts and pathogens. Through fitting the model to data from previously published experiments, we have demonstrated that the Salmonella-HeLa system may be simply describable with time-independent rates, regardless of experimental conditions.