Abstract
Unlike in other cancers, a high infiltration of macrophages in colorectal cancer has been associated with improved patient prognosis. However, controversy lies over the exact effect macrophages have in altering colorectal tumour progression given conflicting data on the subject.
Myeloid populations were assessed by flow cytometry in both tumour and non- tumour bowel (NTB) tissue of colorectal cancer patients. This revealed a heterogeneous macrophage population ranging from gut conditioned to a more conventional monocyte-derived macrophage phenotype. The frequency of gut conditioned macrophage subsets was decreased in tumour compared to matched NTB tissue, whereas the less gut-conditioned phenotypes were increased in the tumour. The frequency of macrophages expressing high levels of CD206 and CD163, both associated with an anti-inflammatory phenotype, was also decreased in the tumour tissue compared to NTB.
Microenvironment strongly affects macrophage phenotype. The tumour microenvironment is often vastly different to that of the adjacent non- transformed tissue. This is due to major alterations in cytokine production by the tissue and local immune cells, in addition to significant changes to metabolic processes associated with nutrient and oxygen deprivation. Patient peripheral blood mononuclear cells (PBMCs) were cultured in both tumour and NTB conditioned media, which were assessed for cytokine content using a multiplex assay. Conditioned medias that contained high levels of IL-6 and/or TNF-α better supported an anti-inflammatory macrophage phenotype. The effect of glucose availability on macrophage phenotype was also assessed. Macrophages cultured in the absence of glucose had a more anti-inflammatory phenotype than those cultured under normal glucose conditions.
Mass cytometry data was also collected for colorectal tumour and NTB tissue. These datasets were assessed using a variety of dimensionality reduction and cluster based, network analysis techniques to look at the immune system as a whole, rather than the traditional compartmentalised approach. A population of putatively suppressive CD25+ myeloid cells present in the tumour tissue was discovered using a network analysis algorithm. Further, this approach highlighted the need to overcome the current dogma regarding the compartmentalisation of immune cells, and consider them as an inter-connected system in our analyses.
Together, work in this thesis revealed a more activated, highly heterogeneous, macrophage population in colorectal tumours compared to adjacent NTB. Moreover, factors present during monocyte to macrophage maturation had the greatest impact on macrophage phenotype. Finally, taking a system wide approach to studying immune populations using network analyses revealed unique myeloid populations and highlighted the necessity of these analytical methods for studying complex immune environments.