Preview

PDF, English Download (27MB) | Terms of use

Abstract

Recent efforts to incorporate a range of molecular characteristics into diagnostic guidelines underscore the necessity of comprehending the intricacies of tumour biology for better clinical decision-making. The role of environmental factors and cellular interactions is also increasingly recognised as a vital component in understanding cancer growth and progression. A complex cellular and spatial landscape is particularly evident in gliomas, which represent highly heterogeneous and plastic brain tumours. In this work, I develop a robust computational framework for the analysis of spatial single cell transcriptomics data and utilise it to conduct two studies on the spatial biology of gliomas. In the first study, I conducted a comparative analysis of necrotic tissues in ten patients who were previously diagnosed with glioblastoma and subsequently presented with either tumour progression or radionecrosis. To this end, I generated a spatial single cell transcriptomic atlas consisting of over a million of cells and encompassing several brain-resident cell types and tumour states. The analysis of the annotated data revealed that radionecrotic samples contained abundant tumour cells with downregulated EGFR expression and were infiltrated by macrophages that contributed to gliosis. In contrast, samples with progression contained progenitor-like and cycling tumour cells that maintained high EGFR expression. The study offers invaluable insights into the spatial landscape and cellular interplay in radionecrosis and holds the potential to inform future research aimed at improving diagnostic and therapeutic strategies for glioblastoma patients.

In the second study, I designed a custom gene panel to enable a thorough investigation into the intricacies of cellular and spatial composition of over 300 samples from patients diagnosed with seven different glioma types. The examination of the annotated data set revealed both disparities and commonalities in the tumour expression patterns among adult-type diffuse gliomas and ependymal tumours. The spatial resolution permitted systematic examination of the spatial neighbourhoods that were linked to individual tumour transcription programs. The study establishes the foundation for future research projects in the group that will employ tailored panels and offers a glimpse into the spatial organisation of gliomas. I believe that the generated atlas along with standardised clinical data can facilitate further attempts to identify clinically relevant associations.

To summarise, I posit that the research conducted within the scope of this dissertation stands to provide a valuable basis for future endeavours in the spatial field. Furthermore, the biological insights derived from the generated data can be used to inform a more focused exploration of the intricate biology of necrotic tissue and the broader spatial patterns characterising multiple gliomas in the future.