DNA METHYLATION ALTERATIONS AS FUNCTIONALLY RELEVANT MECHANISMS, PROMISING TUMOUR BIOMARKERS AND THERAPEUTIC TARGETS

DNA methylation alterations are frequent early events in cancer. A global loss of DNA methylation and a focal hypermethylation of CpG islands (CGIs), mostly located at gene promoter regions, occur in cancer cells. Cancer DNA methylation changes present several features making them potential biomarkers for different applications, such as cancer risk definition, early tumour detection, tumours stratification, prognosis, prediction of therapy response and monitoring of disease evolution including early detection of tumour recurrence. Methylation alterations can be also trace in cell-free circulating tumour DNA, allowing their detection in non-invasive matrices such as urine, stool, serum or plasma. Although several studies have been focused on the identification of methylation-based biomarkers, very few of them have been introduced in clinical practice. Therefore, biomarkers of proven clinical utility that can support or even replace the current diagnostic methods of many cancers consisting in invasive procedures, such as tumour biopsy, are still lacking. DNA methylation alterations are associated with changes in gene expression patterns but the relationship between these two mechanisms is still not clear and needs to be elucidated. In fact, while promoter hypermethylation of some genes, defined as tumour suppressor genes, is associated to gene downregulation, the majority of genes targeted for DNA methylation in cancer are genes expressed at background or very low levels in the normal tissues from which tumours originate. In this last case, some studies have not found a decrease in gene expression levels, while others have detected a further gene downregulation. Moreover, while hypermethylation of promoter-associated CGIs is usually associated with gene downregulation, hypermethylation of CGIs located at gene bodies have been either negatively or positively associated with gene expression. Moreover, since DNA methylation is a reversible epigenetic change, it represents a potential cancer therapeutic target. In fact, methylation status can be edited and reverted to a DNA methylation pattern characteristic of normal cells, possibly restoring a normal cell phenotype. This thesis project aims to identify new potential methylation-based biomarkers in different cancers, to investigate the association between methylation and gene expression and to test the application of CRISPR-dCas9 tool for targeted-methylation editing. A genome-wide approach for the selection of the most informative biomarkers have been employed in all the tumours analysed, including two solid cancers, colorectal cancer (CRC) and biliary tract cancer (BTC) and one blood cancer, chronic lymphocytic leukaemia (CLL). Potential methylation biomarkers have been identified in all the cancers analysed and validated in publicly available datasets. In CRC, the potential utility of two selected methylation alterations for CRC detection through non-invasive methods has been demonstrated. In CLL, methylation of a CGI located in the gene body of SHANK1 has been proposed as a biomarker for prediction of the disease years prior diagnosis, CLL diagnosis and prognosis. Finally, specific-BTC methylation alterations showing high specificity and sensitivity have been identified in the BTC study. The analysis of different types of cancer has allowed to discover that methylation alterations of CGIs associated with protocadherin (PCDH) gene cluster are frequent common events in different solid tumours but not in CLL. Gene expression analyses have confirmed a negative correlation between hypermethylation and gene expression, also for the SHANK1-associated CGIs located in the gene body. Moreover, our results showed that normally low-expressed genes that are hypermethylated in cancer are further downregulated. Finally, preliminary results of CRISPR-dCas9 tool confirmed its utility for targeted methylation editing and its possible use as a cancer therapeutic strategy.