An investigation of genomic instability and its impact on cancer development and heterogeneity

Abstract

Genomic instability (GIN), a genomic state facilitating large scale chromosomal rearrangements, is a hallmark of cancer. GIN can contribute to oncogenesis by disrupting genes, and leading to copy number aberrations (CNAs), the gain or loss of genomic segments. In this thesis I describe two projects linked by the overarching theme of GIN, outlined below:

Project 1: Copy-number aberrations (CNAs) contribute to clonal diversity within cancer, with clinical implications. Breast cancer is one such example, but the effect of CNAs on gene expression in intra-tumour subclonal populations has not been properly characterised. Due to sequencing technology limits and lack of computational methods, it is difficult to assess CNAs at a subclonal level. Here, I have benchmarked the ‘InferCNV’ computational method and used it to infer single cell CNA profiles from 14 primary breast cancer single cell RNA-sequencing (scRNA-seq) datasets. I reveal diverse intratumoural heterogeneity involving at least four subclonal populations per tumour. Finally, I identify subclones with expression/CNA profiles indicative of metastatic potential, involving differential regulation of metastasis associated genes such as MUCL1, BST2 and IGFBP5.

Project 2: High-grade serous ovarian cancer (HGSOC) is characterised by widespread GIN. Drivers of GIN include deficient DNA repair and amplification of Cyclin E1, however no major cause is known for one third of tumours. Deregulation of repetitive elements may contribute to GIN in HGSOC. It is difficult to investigate repetitive elements from sequencing data as they map to multiple places within the genome. I have quantified repetitive RNA in 99 high-grade serous ovarian cancer (HGSOC) and matched control RNA-seq datasets to determine their potential contribution to GIN. I identified retrotransposons which are deregulated in HGSOC, which may have been active during cancer development. Some of these retrotransposons were enriched at structural variant breakpoints, indicating potential causality. Finally, I identified retrotransposon-associated structural variants in proximity to deregulated oncogenes implicated in homologous DNA repair, which may have modulated their expression and contributed to cancer development.

In summary, I have explored both a cause (retrotransposons) and consequence (CNA-based heterogeneity) of GIN in cancer, and shown how GIN can contribute to the modulation of cancer-associated genes which influence cancer development and outcomes.