Application of high-throughput cancer genomic approaches for the molecular characterization, patient stratification and identification of evolutionary patterns in different cancer types

Cancer genomics refers to the study of the genome and transcriptome from tumour cells and their normal host cells. It aims to understand the genetic basis of tumour cell proliferation and evolution of cancers driven by somatic mutations and selection pressures from the tumour microenvironment, the immune system and therapeutic interventions. Until the emergence of next generation sequencing (NGS) technologies, microarrays were the fundamental tool used in cancer genomics. Using these high-throughput platforms, the characterization and stratification of cancer patients was initially based on the molecular (transcriptional, genomic and epigenetic) profiles derived from single bulk tumour samples. Recently, cancer genomics has further interrogated the inherent complexity of tumours by sequencing multiple cancer specimens from each patient. This approach has uncovered a high degree of heterogeneity within tumours resulting from the Darwinian character of cancer which also explains treatment resistance and is key for personalised treatments. This introductory section presents my contribution to the cancer genomics field in particular on breast cancer, glioblastoma multiforme (GBM), colorectal carcinoma (CRC) and prostate cancer. I demonstrated the potential of miRNA profiling to characterise established breast cancer cell lines and primary tumours. Further, I designed and performed a microarray platform comparison study that sustained the transcriptomic and genomic profiling of the largest breast cancer cohort to date. More recently, I uncovered different evolutionary patterns of treatment-naïve breast cancer patients which are recapitulated in plasma cell-free tumour DNA (ctDNA). Unfortunately, the potential use of ctDNA is limited in GBM patients, thus tumour tissue specimens are necessary to perform brain cancer genomic studies. Using a meticulous multisampling scheme to collect spatially distinct GBM tumour fragments I identified high levels of heterogeneity within individual tumour masses at the genomic and transcriptomic levels. Interestingly, a proportion of GBM patients also bear tumour cells in the subventricular zone (SVZ) away from the primary tumoir mass and for the first time I showed that some GBM tumours grow out from the SVZ to develop large tumour masses while others grow in the SVZ from previously established primary tumours. In addition, I have shown that there are also GBM cells infiltrating the normal brain parenchyma. These residual cancer cells and those in the SVZ, indicate a putative source of treatment resistance. In colorectal cancer, resistance to cetuximab treatment is nearly inevitable. Using multi-region and longitudinal biopsies I uncovered resistant cancer cells which pre-existed cetuximab therapy and resistant cells which emerged and expanded during treatment. Importantly, treatment response in metastatic CRC patients can be recapitulated using patient derived organoids (PDOs) which could be implemented for personalised medicine. Intriguingly, despite a high metastatic potential, CRC might have a slow progression rate as I uncovered from the disease chronology of one CRC patient. Lastly, I validated a recurrent epigenetic alteration in prostate cancer and demonstrated that it is a very early event on the progression of this disease. In summary, this introductory section will describe the application of novel methodologies based on standard cancer genomics tools which have shed light into the molecular classification and evolutionary patterns of different cancer types. The applicability of these findings to personalised cancer treatments will also be discussed