Structural variation in cancer genomes & the identification of personalized DNA-based cancer biomarkers

Cancer is the result of the accumulation of genetic and epigenetic alterations in key genes, which ultimately lead to uncontrolled growth of mutated cells. Genetic alterations range from small base substitutions to large chromosomal structural rearrangements. Many cancer genomes carry tens to hundreds of somatic structural rearrangements, which may have functional consequences. However, the patterns and characterization of somatic rearrangements in human cancers are still in early stages. In this thesis, I describe the genomic landscape of somatic structural rearrangements in human pancreatic cancers. I then investigate the potential mechanisms involved in somatic rearrangement formation and test whether such rearrangements can be used as biomarkers to monitor patient therapy. Chapter 1 - Introduction: This chapter is a broad overview of somatic structural rearrangements in human cancers, motivation for studying them, and their potential use for clinical applications. Chapter 2 – A workflow to increase verification rate of somatic chromosomal rearrangements using high throughput next generation sequencing: I established a high throughput workflow to rapidly verify somatic structural variants and identify the exact location of breakpoints using benchtop next generation sequencing and computational tools. The workflow examined more than 300 predicted breakpoints and identified breakpoint location in more than 80% of somatic events at base level. The results demonstrated that next generation sequencing is comparable to the conventional Sanger sequencing and can complete the verification workflow in a shorter timeframe, enabling rapid validation of events. Chapter 3 – Patterns of somatic breakpoints may indicate repair mechanisms that were active or absent during the generation of genomic rearrangements: This chapter describes the spectrum of somatic rearrangements and breakpoints detected in 120 primary pancreatic ductal adenocarcinomas genomes. The analysis includes characterization of the breakpoint junctions to infer which potential DNA repair processes are occurring. The results revealed that the majority of tumours exhibited repair of chromosome structural breaks using microhomology suggesting that NHEJ is the main mechanism of DNA repair in pancreatic cancer. Tumours with BRCA1 or BRCA2 gene mutations or with a high contribution of a BRCA-like mutational signature showed a higher frequency of somatic breakpoints with microhomology length 1 to 5 bp and lower frequency of breakpoints with a blunt end (0 bp) when compared to tumours harbouring BRCA wild type or low BRCA mutational signature. The similarity in breakpoint characteristics between tumours with BRCA mutation and BRCA mutational signature reinforce previous findings that a subtype of pancreatic tumour might have deficiency in the HR pathway and could respond to PARP1 inhibitors. Furthermore, the analysis of the DNA sequence surrounding the breakpoints revealed strong signal of A+Ts rich regions suggesting that the formation of somatic rearrangements could also be mediated by either retrotransposition activity or chromosomal fragile sites. Taken together, the analyses of breakpoint junctions highlighted that the formation of somatic rearrangements in pancreatic carcinogenesis is complex – potentially with more than one mechanism active within a cancer genome. Chapter 4 – Identification of personalized DNA-based biomarkers for pancreatic cancer and the assessment of whether they can be used to monitor tumour burden and response to chemotherapy: In this chapter, I first evaluated the performance of sequencing and PCR based methods to detect ctDNA. Subsequently, I quantified ctDNA in the serum or plasma of the three pancreatic cancer patients. Little or no detection of ctDNA was observed in the analysed serum samples. For one patient, a tumour specific rearrangement was detected in the serum, and this patient had already presented with metastatic disease. Based on the results, it was hypothesized that ctDNA released from pancreatic cancer might be limited by: i) nature of pancreatic cancer; ii) the physiological location of the pancreas which could influence the amount of ctDNA released in the circulation, as the fragmented tumour DNA might be cleared by the liver and therefore reduced the opportunity to detect ctDNA in the circulation; iii) volume of plasma or serum used to isolate cfDNA; iv) the status of disease progression. In this analysis, the ctDNA was detected only in the sample collected at late stage pancreatic cancer (after the diagnosed of liver metastasis). Thus, in the context of pancreatic cancer disease, the quantification of ctDNA might be more suitable for recurrent disease, which has metastasized from the pancreas. Chapter 5: I concluded my findings throughout my studies in Chapter 2, 3 and 4 with a summary and future directions.