vnewjay

A link between DNA methylation and cancer was first demonstrated in 1983, when it was shown that the genomes of cancer cells are hypomethylated relative to their normal counterparts. Hypomethylation in tumour cells is primarily due to the loss of methylation from repetitive regions of the genome6, and the resulting genomic instability is a hallmark of tumour cells. Rearrangements that involve the large block of pericentromeric heterochromatin on chromosome, for example, are among the most frequent genomic instabilities in many tumour types. Reactivation of transposon promoters following demethylation might also contribute to aberrant gene regulation in cancer by TRANSCRIPTIONAL INTERFERENCE or the generation of antisense transcripts. Loss of genomic methylation is a frequent and early event in cancer, and correlates with disease severity and metastatic potential in many
tumour types.
Gene-specific effects of hypomethylation also occur. For example, the melanoma antigen (MAGE) family of cancer–testis genes, which encode tumour antigens of unknown function, are frequently demethylated and re-expressed in cancer19. Demethylation accompanied by increased expression has been reported for the S100 calcium binding protein A4 (S100A4) gene in colon cancer20, the serine protease inhibitor gene SERPINB5 (also known as maspin) in gastric cancer21, and the putative oncogene γ-synuclein (SNCG) in breast and ovarian cancers22. Global demethylation early in tumorigenesis might predispose cells to genomic instability and further genetic changes, whereas gene-specific demethylation could be a later event that allows tumour cells to adapt to their local environment and promotes metastasis.
Research on genome-wide demethylation in cancer cells has been largely overshadowed by studies of gene-specific hypermethylation events, which occur concomitantly with the  hypomethylation events discussed above. Aberrant hypermethylation in cancer usually occurs at CpG islands, most of which are unmethylated in normal somatic cells, and the resulting changes in chromatin structure (such as histone hypoacetylation) effectively silence transcription. Indeed, a distinct subset of many tumour types has a CpG-island-methylator phenotype, which has been defined as a 3–5 fold increase in the frequency of aberrant hypermethylation events. Genes involved in cell-cycle regulation, tumour cell invasion, DNA repair, chromatin remodelling, cell signalling, transcription and apoptosis are known to become aberrantly hypermethylated and silenced in nearly every tumour type. This provides tumour cells with a growth advantage, increases their genetic instability (allowing them to acquire further advantageous genetic changes), and allows them to metastasize. In tumours with a well-defined progression, such as colon cancer, aberrant hypermethylation is detectable in the earliest precursor lesions, indicating that it directly contributes to transformation and is not a late event that arises from genetic alterations.
Use of RESTRICTION LANDMARK GENOMIC SCANNING to analyze the methylation status of 1,184 CpG islands from 98 tumour samples showed that de novo methylation of CpG islands is widespread in tumour cells. In this study, the extent of methylation varied between individual tumours and tumour types, and an average of 608 CpG islands were aberrantly hypermethylated. Because the hypermethylation of CpG islands is relatively rare in normal cells, is an early event in transformation, and robust assays can detect methylated DNA in bodily fluids, it represents a good potential biomarker for early cancer detection. The underlying cause of methylation defects in cancer remains unknown, but possible mechanisms are discussed at the end of this review. DNA methylation and cancer are also related through the loss of imprinted methylation patterns in many tumours, as discussed below.