Scientists claim cancer cells have lost the ordered patterns of methylation demonstrated by normal tissues and demonstrate chaotic methylation across the genome that could help them adapt to changing environments in growing tumors and facilitate metastasis. Research by a Johns Hopkins University-led team has found that in contrast to normal tissues that demonstrate specific patterns of methylation at CpG islands or in large blocks of DNA, cells from a range of different tumor types showed vastly differing methylation patterns at the same sites, which they termed differentially-methylated regions (cDMRs).
In colon cancer this was evidenced by loss of methylation stability both at the boundaries of CpG islands and the presence of blocks of hypomethylation affecting over half the genome, which together led to significant variability in gene expression. Andrew P. Feinberg, M.D., professor of molecular medicine and director of the Center for Epigenetics at the Johns Hopkins University School of Medicine’s Institute for Basic Biomedical Sciences, and colleagues, report their findings in Nature Genetics. In their paper, titled “Increased methylation variation in epigenetic domains across cancer types,” the authors point out that their findings could indicate that the use of nonspecific DNA methylation inhibitors as cancer therapy could have unwanted effects by activating tumor-promoting genes in hypomethylated blocks.
Work by Dr. Feinberg’s team has previously demonstrated that colon cancers exhibit changes in the degree of methylation at regions of lower CpG density near CpG islands. These cDMRs corresponded in the main to the same regions that show DNA methylation variation in normal spleen, liver, and brain tissues, or tissue-specific DMRs (tDMRs), suggesting their involvement in cell differentiation in normal tissues.
To investigate cancer-related changes in global methylation further, the researchers designed an array to analyze over 151 cDMRs that were consistently altered across colon cancer, and compared methylation within these regions in another 290 samples including matched normal and cancer samples from colon, breast, lung, thyroid, and Wilms’ tumor. They found that almost all of the cDMRs were altered across all cancers tested. “Specifically, the cDMRs showed increased stochastic variation in methylation level within each tumor type, suggesting a generalized disruption of the integrity of the cancer epigenome,” the team suggests.
To look further at this phenomenon, the team carried out genome-scale bisulfite sequencing of three colorectal cancers, matched normal colonic mucosa and two adenomatous polyps. The study was designed to obtain methylation estimates with enough precision to detect differences of 20% methylation. The results demonstrated a significant loss of methylation stability in colon cancer, which involved CpG islands and their boundary regions (shores), and also identified large blocks (5010 kb) of hypomethylation. “Remarkably, these hypomethylated blocks in cancer corresponded to more than half of the genome, even after accounting for the number of CpG sites within the blocks,” the authors note.
Over 5,800 hypermethylated and 4,300 hypomethylated small DMRs (less than 5 kb) were also identified. Interestingly, the research comfirmed previous findings that hypermethylated cDMRs are enriched in CpG islands, whereas hypomethylated cDMRs are enriched at CpG island shores. The thousands of small DMRs frequently involved loss of boundaries of DNA methylation at the edge of CpG islands, shifting of DNA methylation boundaries, or the creation of novel hypomethylated regions in CG-dense regions that are not normally recognized as islands. The knock-in effect of this variability in cancer cell methylation was both an increase in gene silencing and a substantial enrichment of genes with increased expression variability in the methylated blocks, Dr. Feinberg and colleagues state.
These data underscore the importance of hypomethylated CpG island shores in cancer, the authors note: “Shores associated with hypomethylation and gene overexpression in cancer are enriched for cell cycle related genes, suggesting a role in the unregulated growth that characterizes cancer.” Regions of altered methylation in cancer were also found to match those in normal tissues associated with controlling cell differentiation into specific cell types. “Targeting those regions might help the cells become more normal,” suggests Rafael Irazarry, Ph.D., professor of biostatistics at the Johns Hopkins University Bloomberg School of Public Health and lead author of the Nature Genetics paper.
From the cancer perspective, methylation chaos is helpful because it means tumors can turn genes on and off in an uncontrolled way, increasing their adaptability, Dr. Feinberg adds. This indicates that increased epigenetic heterogeneity in cancer at cDMRs may play a role in the ability of cancer cells to adapt rapidly to changing tissue environments, such as increased oxygen in regions of neovascularisation, then decreased oxygen with necrosis, or metastasis to a new intercellular milieu.
Current efforts to exploit DNA methylation for cancer screening have focused on identifying narrowly defined cancer-specific profiles. However, the Johns Hopkins research suggests broader evaluation of the cancer epigenome may be more relevant. “Given the importance of boundary regions for both small DMRs and large blocks identified in this study, it will be important to focus future epigenetic investigations on the boundaries of blocks and CpG islands (shores) and on genetic or epigenetic changes in genes encoding factors that interact with them,” the authors conclude.