Defining immune system function in humans

Abstract

In this thesis I design two novel methods for the analysis of human immunological data from blood and solid tissue respectively. I have developed a method capable of identifying poorly characterised disease-associated genes and defining mechanisms controlling their expression. By exploring multiple cohorts of public transcriptome data, I showed that CST7 is upregulated in the blood during a diverse set of inflammatory conditions. Interestingly, this upregulation was neutrophil-specific and is not induced by microbial products or cytokines commonly associated with inflammation but is associated with type I interferon signalling. In this chapter, I demonstrate the value of publicly available transcriptome data in knowledge generation and potential biomarker discovery. I have also developed a protocol for the single-cell analysis of lesions in multiplex images. I developed a metric called tCPI that can quantify how tight or loose the distribution of cells is in the lesion centre. An additional metric was developed called immCPI that measures the overall distribution of each immune cell population relative to the lesion centre. I developed a final method called lesion neighbourhood analysis that can quantify the relative location of hundreds of individual lesions within a tissue section and can identify patterns in their distributions. Together this workflow provides a robust tool for spatial quantification at the lesion level. I applied this lesion analysis protocol to lung tissue resection samples from patients with Tuberculosis. Interestingly, lesions did not fall into distinct clusters based on composition. Using both principal component data and tCPI, I grouped lesions into 4 types based on composition and cell distribution. Using immCPI revealed that the intra-lesion location macrophages / monocytes and B cells is dependent on lesion type. Lesion neighbourhood analysis revealed that B cell enriched lesions tend to locate twice as close to a necrotising lesion compared to other lesions. The robust analytic power of these new approaches offers important tools in the study of human immunology and will contribute to furthering our understanding of inflammatory diseases.