Association of global CpG methylation status with gene expression phenotypes in normal and cancer cells

Association of global CpG methylation status with gene expression phenotypes in normal and cancer cells Anna DE GRASSI Department of Biosciences, Biotechnology and Biopharmaceutics, University of Bari, via Orabona 4, 70125, Bari, Italy Cytosine methylation at CpG dinucleotides is a major mechanism of epigenetic regulation of gene expression, whereas methylation of promoters and demethylation of gene bodies are typically associated with transcriptional repression and vice versa (1). Poor is instead known about the forces that generate variable degrees of global CpG methylation in the genomes of different cells and how the gene expression phenotypes are consequently altered. The ENCODE project has recently furnished and invaluable data platform for this investigation, by systematically performing high-throughput experiments in several human cell lines (2). Forty-five normal and cancer cell lines were collected from the ENCODE platform for which both reduced representation bisulfite sequencing (RRBS) data and exon array data are publicly available. A subset of ~130,000 CpG sites scattered throughout the genome were selected for the high reproducibility among RRBS replicates and were used to estimate the global CpG methylation status of each cell line. Using this estimation, cancer cell lines were distinguishable from primary cell lines with 100% sensitivity and specificity. The global CpG methylation degree was higher in cancer cells than in normal cells in all the genome compartments (exons, introns, promoter and intergenic regions), suggesting that at least one component of the CpG methylation profile of cancer cells is due to a basal and homogeneous hypermethylation along the genome. By comparing the CpG methylation status of cell lines and the expression level of ~20,000 genes in the same cell set, several genes were detected whose expression profile significantly correlates with the global CpG methylation degree. The top positively correlated genes include DNA methyltransferases and genes encoding proteins involved in nucleosome organization and assembly, as well as in histone exchange. Beside furnishing a list of candidate genes whose expression profile associates with the hypermethylation of cancer cells, these results suggest that an increased global CpG methylation status might be determined by the over-expression of DNA methyltransferases, and not by the transcriptional repression of DNA demethylator genes, and by an extensive remodeling of chromatin. References: 1. Ball MP et al. Nat Biotechnol. 27, 361-368 (2009). doi: 10.1038/nbt.1533. 2. Encode Project Consortium. Nature 489, 57-74 (2012). doi: 10.1038/nature11247.