Integration of physiological tissue models and machine learning to understand genomic instability from oncogene activation to cancer initiation

Genes that cause abnormal growth of cells and development of cancer are called oncogenes. While oncogenes often arise from mutation of cellular genes, several types of viruses that can cause cancer, such as human papillomavirus (HPV), bring viral oncogenes into the cells they infect. Oncogenes can induce 'replication stress' (called oncogene-induced replication stress or Oi-RS), which is characterised by the frequent stalling or slowing of DNA replication when chromosomes are copied. This causes more mutations in the chromosomes, which is called genomic instability.

While replication stress is common in established cancers, it has been discovered that it might also occur in pre-cancerous and early cancer tissues, leading to the hypothesis that Oi-RS is an initial driver of cancer but there is currently no direct evidence for this from existing experimental models of cancer development. Moreover, it is not known how early replication stress causes a healthy cell to become cancerous. Discovery of the initiating mechanisms leading to genomic instability and cancer would provide a vital step forward in understanding what causes cells to become cancerous and will be instrumental in improving early detection and treatment of cancer. We propose to use oncogenic HPV infection of human skin cells, the natural target cell of HPV, as a human tissue model system to analyse the precise steps from Oi-RS and genomic instability to cancer initiation.

HPV causes >600,000 cancers per year worldwide that are characterised by high levels of genomic instability. Infection with high-risk types of HPV is a very useful model for Oi-RS and genomic instability during early cancer development, because viral oncogenes rapidly cause replication stress and genomic instability. To date however, no team has yet combined expertise to establish the precise pathways from HPV infection, replication stress and genomic instability, through to cancer formation in relevant human skin models.

Our team comprises a tumour virologist, two cell biologists and a computational scientist. Together, we cross multiple biology disciplines from molecules to human tissue models and will bring diverse methods and technologies to investigate Oi-RS in a model of HPV infection. Our expertise ranges from molecular biochemistry and cell biology to computational biology and the application of artificial intelligence. Importantly, the individual partners all have a fundamental interest in how replication stress and genome instability contribute to cancer development. This project will for the first time bring together distinct backgrounds and disciplines to unravel the molecular events between Oi-RS, genomic instability, and cancer development. We will integrate single cell- and single molecule sequencing technology and machine learning-based detection of replication stress to interrogate the steps from early establishment of HPV in skin cells to the development of invasive cancer cells.

Genome instability is a hallmark of cancer but the causes of genome instability that drive cancinogenesis are poorly defined. Oncogene expression during carcinogenesis is thought to induce oncogene induced replication stress (Oi-RS) which is proposed to induce subsequent mitotic defects and genome instability. However, longitudinal data supporting this hypothesis are sparse due to limited access to appropriate disease progression models. This project will investigate the current major unknowns in oncogene-induced genomic instability using a primary cell-based, physiological model of HPV establishment and oncogenic disease progression. Oncogenic human papillomaviruses (HPV) express viral oncogenes E6 and E7, which induce replication stress and genomic instability, eventually causing carcinomas of the anogenital and oropharyngeal tracts. HPV infection thus provides a tractable model system in which to decipher the pathway from initial Oi-RS to cancer.

The impact of HPV infection on genomic instability has mainly been studied by overexpression of viral oncogenes in primary epithelial, or cancer cell lines, which fail to accurately recapitulate natural pathogenesis of HPV. We will now employ a state-of-the-art 3D primary human tissue model that faithfully recapitulates the HPV infection cycle to dissect the timing and mechanisms of Oi-RS, genomic instability, and oncogenesis. We will integrate state-of-the-art cell and molecular biology methods with single cell and single molecule sequencing technologies and machine learning-based detection of genomic position-resolved replication stress. In depth analysis of these models using methods established outside of the traditional cell biology arena will uncover key driving events in HPV-driven carcinogenesis that will be directly applicable in other oncogene-induced tumours. The findings will define the steps from initial oncogene-induced replication stress to cellular immortalisation and disease progression.