A systems biology approach to signalling pathway and gene network regulation modelling in Mastitis

Abstract

Mastitis, an inflammation in the mammary gland, is one of three major diseases in the dairy industry. One in three cows will encounter the disease which is also a problem in humans and other species. While E. coli bacterial infections lead to acute mastitis, S. aureus lead to chronic mastitis. Dynamics of the regulation and identification of differentially expressed genes between two bacterial infections are important factors for understanding mastitis and assist in the development of pharmaceutical and breeding targets. Previous studies have identified differentially expressed genes. However, they have not compared expression between two bacterial infections over time. Neither have the dynamics of the signalling and gene network regulation that leads to the differential expressions been investigated. This thesis aims to provide new insight into immune defence in mastitis by analysing dynamics of the signalling and gene regulation in mammary epithelial cells. The main focus is to develop a mathematical model of the signalling and gene network regulation in mastitis. First, the genes differentially regulated between the two clinical presentations of the disease are identified. Time series microarray experiments of E. coli and S. aureus challenged mammary epithelial cells are analysed, and confirm that each type of mastitis has a significantly different gene expression time profile from healthy cells. The differentially expressed gene time profiles are then compared between the bacterial challenges. RANTES is identified as the key cytokine which is responsible for two distinctly different time profiles between the bacterial challenges. In this second part the mathematical model is developed and a systems biology approach applied to investigate the complex dynamics of signalling proteins and gene network regulation of three different cytokines (RANTES, IL8 and TNFα) in mastitis. A modification to a conversion method allows us to use relative microarray expression data in the model. The method opens up a large amount of datasets for use in future modelling. The model explains signalling and gene network regulation of three cytokines in acute mastitis. No fit could be found for the S. aureus experimental data indicating that there is a difference in the regulatory mechanisms between the two types of mastitis. In the third part sensitivity analysis is used to investigate the role of parameters on the model output. The analysis reveals that each cytokine is sensitive to specific parameter changes. This indicates different dynamics in the regulatory mechanism. As a result, pharmaceutical and breeding targets need to be evaluated in the context of all cytokines to prevent undesirable side effects. The importance of modelling prior to experimental design is also revealed; each cytokine has a specific time frame for the most informative experimental measurement. In the fourth part robustness analysis is used to investigate the role of the bacterial load on the model output. Robustness analyses indicate that robustness does not originate in the nuclear NFκB time profile and is specific for each cytokine. Finally, future directions of the model and biological experiments are discussed.