Abstract
Tumours develop in the context of a complex environment composed of various immune and non-immune stromal components that define the tumour microenvironment (TME). Evidence of interactions between tumours and their TME has emerged over the past decades. In particular, features of the TME, such as the adaptive immune system, can sculpt the genomic landscape of growing tumours. At the same time, alterations in some cancer driver genes can promote immune escape by modulating the composition of the TME or by conferring resistance to immune killing. Importantly, such an interplay can impact response to therapies that target the immune contexture of the TME, particularly immune checkpoint blockade (ICB) therapy. Colorectal cancer (CRC) is an exemplar of a cancer type where all these complex tumour-immune interactions are observed and where the tumour’s genetic profile perfectly segregates with its immune profile. This clear separation in tumour and immune features is also reflected clinically, since sensitivity to ICB in CRC is largely prevalent among cases with microsatellite instability (MSI), owing to their high neoantigen load and immune-rich profile. Nevertheless, more than half of MSI CRCs do not benefit from ICB treatment, suggesting that other factors contribute to response determination. Therefore, CRC provides an excellent opportunity to study how the interplay between tumours and their immune TME impacts response to ICB.In this thesis, I investigate the properties of cancer driver genes involved in the tumour-immune interplay (i.e. TME cancer drivers) and investigate how such an interplay impacts CRC response to ICB. First, I present the Network of Cancer Genes (NCG), a comprehensive repository of cancer driver genes and their systems-level properties. Using NCG, I undertake a systematic analysis of the alteration pattern and the systems-level properties of a curated list of TME cancer drivers. I find that TME cancer drivers form a core of cancer genes that tends to acquire inactivating somatic alterations and that shows the strongest systems-level property profile among all cancer drivers. This strong property profile, particularly in terms of protein-protein interactions, reflects the central position of TME cancer drivers in the cell as well as their involvement in diverse biological processes, part of which include immunomodulatory functions. These results therefore provide a systems-level rationale for the involvement of some cancer driver genes in the interplay between tumours and their immune TME. In CRC, TME cancer drivers are preferentially altered among all CRC-specific drivers, emphasising the prevalence of driver gene-mediated tumour-immune interactions in this cancer type.
Second, I present our immunogenomic profile of CRC response to ICB. Specifically, we leverage clinical data along with multi-regional and multi-omic data from CRC patients who were subsequently treated with anti-PD1 therapy. We elucidate the involvement of the WNT pathway in inducing resistance to ICB via immune exclusion in both non-hypermutated and hypermutated CRCs. We also clarify the limits of tumour mutational burden in predicting response and uncover further response predictors, including clonal neoantigens, clonally expanded T cells, and high infiltration of CD74+ macrophages. Using spatial data from IMC, we show that these CD74+ macrophages form clusters within the TME of hypermutated CRCs and uncover interactions between PDL1+ CD74+ macrophages and cytotoxic or proliferating PD1+ CD8+ T cells. We propose that this co-localisation pattern is key for CRC response to anti-PD1 treatment.
Altogether, this thesis provides novel insights into the properties of cancer drivers involved in the tumour-immune cross-talk and provides novel tumour- and immune- related determinants of CRC response to ICB that could serve as biomarkers to improve patient stratification.
Date of Award | 1 May 2022 |
---|---|
Original language | English |
Awarding Institution |
|
Supervisor | Francesca Ciccarelli (Supervisor) & Kathleen Steinhofel (Supervisor) |