Abstract:
Glioblastoma is one of the most fatal tumours to affect the central nervous system. Although
immunotherapies present a promising therapeutic avenue, there has been a relative failure in the
translation of immune-targeted therapies to the clinic. The inherent heterogeneity and complex
immunosuppressive tumour microenvironment represent significant hurdles for the development
of successful immune-targeted therapies. The success of future immunotherapies relies on a robust
foundational understanding of this complex immune contexture. In light of this, the aim of this
thesis was to map the immune landscape of human glioblastoma, with a particular focus on
myeloid cell heterogeneity.
The glioblastoma immune landscape was characterised by immunohistochemical labelling of
neutrophils, T cells, dendritic cells and microglia/tumour-associated macrophages. Microglia and
peripherally invading macrophages were the most abundant immune cells and were thus the focus
of the remainder of the thesis. Microglia and their peripherally invading counterparts, referred to
as tumour-associated macrophages (TAMs), are ontogenetically distinct myeloid populations that
may hold different functions within the tumour. To interrogate these two macrophage populations,
microglia and TAMs were delineated via microglial-specific marker expression using a single-cell
image analysis pipeline. This novel analysis pipeline was subsequently used to spatially map
microglia and TAMs and their functional profiles across tumour macro-and micro-environments.
Microglia were the predominant myeloid population within the tumour periphery, whilst TAMs
were the dominant population within the tumour core and within vascular and hypoxic niches.
Comparison of gene expression profiles revealed that microglia and TAMs hold distinct functional
profiles within the tumour, with microglia displaying a more chemotactic profile and TAMs a
more pro-tumoural profile. However, microglia within the tumour core also upregulated many of
the same tumour-promoting genes as TAMs. Indeed, many of these genes clustered to specific
niches within the tumour, such as the hypoxic niche. These results suggest the phenotypic and
functional profile of macrophages is dependent on spatial location rather than ontogeny alone.
Taken together, the data presented in this thesis demonstrate the heterogeneity of macrophages
within glioblastoma tumours, highlighting that macrophage phenotype and function is strongly
driven by the local tumour microenvironment.