Epigenetic gene silencing in human cancers: The roles of DNA methyltransferase 1 and double-stranded RNA

DNA methylation refers to the covalent modification of cytocines at the 5 position by a methyl group. In the human genome, such modification predominantly occurs in a CpG dinucleotide context and is catalyzed by DNA methyltransferases (DNMTs) 1, 3a, and 3b. Clusters of CpG dinucleotides, termed CpG islands, are commonly present in promoter regions, and hypermethylation of promoter CpG islands correlates tightly with transcriptional silencing of the downstream genes. This epigenetic phenomenon is especially important in cancers because the promoter regions of many tumor suppressor genes have been found to be aberrantly hypermethylated and silenced. DNMTI is the maintenance methyltransferase that is solely responsible for the faithful replication of DNA methylation patterns at each round of cell division in the murine system. Therefore, DNMT1 has been hypothesized as a logical target for therapy in an attempt to reverse tumor suppressor hypermethylation and silencing. However, genetic deletion studies in the human colorectal cancer cell line, HCT116, showed that depletion of DNMT1 alone is insufficient to cause demethylation of tumor suppressor genes. Significant demethylation and gene re-expression were only observed after both DNMT1 and 3b were disrupted in these cells, suggesting that DNMT1 and DNMT3b cooperate to maintain hypermethylation. This finding was criticized for the low frequency at which DNMTI-/- clones arose and was directly challenged by data obtained using antisense and RNA-interference (RNAi) based techniques to knockdown DNMT1 in the same HCT116 cells. The latter study showed that the dense methylation at the CDKN2A promoter is relieved after only 10 days of DNMT1 depletion. This work began as an effort to independently verify using RNAi based techniques, the requirement for DNMT1 in HCT116 cells to maintain DNA hypermethylation, and was extended to further test this hypothesis in other cancer cell types. Systematic depletion of DNMT1 by RNAi based methods in several different cancer cell lines revealed a cell type specific reliance on DNMT1 to maintain the full hypermethylation pattern. Colorectal cancer cells (HCT116, SW480, and SW48) and T24 bladder cancer cells exhibited lack of dependency on DNMT1 for survival and maintenance of DNA hypermethylation, while T47D breast cancer cells clearly required DNMT1 for both cellular growth and DNA hypermethylation maintenance. Another aspect of this work focuses on understanding the mechanism of initiation of tumor suppressor hypermethylation. In plants, double-stranded RNA (dsRNA) species homologous to the promoter region have been shown to cause transcriptional silencing and DNA methylation of the target sequence. We tested the possibility that dsRNA targeting only the promoter region could induce epigenetic silencing and therefore be a potential trigger of promoter hypermethylation in cancers. We found that short dsRNAs targeted to the CDH1 promoter region could induce transcriptional gene silencing, which was accompanied by a conversion of histone modifications from active to inactive state. Although this dsRNA-dependant transcriptional silencing (RdTS) did not result in the immediate initiation of DNA methylation at the target promoter, RdTS still could potentially lead to DNA methylation if silencing was prolonged.