Frequent Switching of Polycomb Repressive Marks andDNA Hypermethylation in the PC3 Prostate Cancer CellLine

摘要

A key observation in these presented results was the mag-nitude of hypermethylation in PC3 cells, which occurred atalmost all of the inactive CpG island genes. A second keyobservation was that when PRC marks were depleted, aPC3 hypermethylation was observed. One possible modelthat can explain these observations assumes that DNAmethylation in PC3 cells sweeps all CpGs that are notphysically protected (or "masked") (Figure 3). Accordingto this "passive" methylation model, transcription initiation complexes at the TSS [RNA Polll and related chromatinmarks, can serve as one masking mechanism (Figure3a), and Polycomb repressive complexes can serve asanother (Figure 3b). This view was also supported byrecent data on DNA methylation patterns of mouse ESCs.It is important to note that even if Polycomb complexesmask loci from de novo DNA methylation, it is possiblethat PRC activity (or recruitment) is not masked by DNAmethylation (Figure 3c). This may be the case in some ofthe de novo PRC targets we observe near genes that arerepressed in PC3. Regardless of the underlying mechanism, the data outlinea fundamental change in the epigenetic programming ofPC3 cells, in which many genes that are originally silencedby PRCs acquire DNA methylation as an alternative silenc-ing mechanism, and other genes are de novo repressed byPolycomb recruitment, in agreement with another recentstudy. If the PC3 regulatory program has stem cell fea-tures, these must be achieved through a variation on thenormal epigenetic mechanisms for immortality and pluri-potency, a variation that may be critically less flexible. The applicability of these observations to prostate tumorsand cancer cells in general should be further investigated.Even though there are limitations to the use of short-termcultures of epithelial cells for these studies, they at leastrepresent a relatively homogeneous population of the nor-mal cell counterpart of PC3 cells. The results of the studystress the importance of an integrative approach for thecharacterization of the cancer epigenome and demon-strate the benefits of using a coherent experimental sys-tem to map multiple epigenetic factors. According to Gal-Yam et al., it will be intriguing to apply the methodologydeveloped here to additional cell systems and epigeneticmarks and to use a more extensive genomic coverage touncover additional processes that contribute to the epi-genetic reprogramming of cancer cells.