Activation of growth-promoting oncogenes, largely driven by genetic alterations, is a key step during tumorigenesis. Meanwhile, tumors are not only genetically, but also epigenetically, distinct from their tissues of origin. Yet, little is known regarding the extent to which epigenetic mechanisms can activate oncogenes. DNA methylation is the most extensively documented epigenetic modification that can influence cell fate and gene expression. For example, promoter hypermethylation that silences tumor suppressor genes is a key epigenetic event in tumorigenesis. In addition, gene-body hypomethylation is positively correlated with gene expression, although the causal relationship remains to be established. Recently, our group and that of Dr. Bing Ren independently reported broad (e.g. >3.5-kb) under methylated regions (UMRs), termed DNA methylation canyons (Nature Genetics 2014) or valleys (Cell 2013), which in most cases span promoters and gene bodies. Canyons exhibit very low levels of methylation (<10%) in almost all normal cells. Canyon-associated genes (~1,100 in each cell type) are enriched in developmental regulators and homeobox genes, many of which exhibit low or no expression in normal cells. However, despite numerous follow-up studies (i.e., ~540 citations for both papers), the functional role of canyons in gene expression and tumorigenesis remains poorly understood. Nevertheless, our preliminary data indicated that canyons found in normal cells are prone to gene-body, but not promoter, hypermethylation in tumors, with the hypermethylated canyon genes surprisingly enriched in oncogenes. Furthermore, our dCas9-mediated DNA methylation editing experiments revealed an unexpected causal role of gene-body canyon hypermethylation in oncogene activation. Therefore, we hypothesize that we can use cancer DNA methylation data from public domains such as The Cancer Genome Atlas (TCGA) and the International Human Epigenomics Consortium (IHEC), as well as powerful bioinformatics algorithms, to identify genes with cancer-specific hypermethylated canyons in thousands of tumors. We further hypothesize that we can test the functional roles of the hypermethylated canyons in cell and animal models using a dCas9-mediated genome-wide methylation screen. The proposed work is expected to identify hypermethylated canyons that can activate oncogenes and promote tumor growth in vitro and in vivo. Thus, this work can fundamentally transform our understanding about how DNA methylation regulates gene expression and tumorigenesis.

Public Health Relevance

DNA methylation canyon, i.e. broad (i.e. >3.5 kb) under-methylated region (UMR), is emerging as a novel epigenetic feature in the mammalian genome, but its cause-and-effect relationships with gene expression and tumorigenesis remain poorly understood. We propose to reveal hypermethylated canyon target genes, and test their roles in regulating oncogene activation and tumor growth in vitro and in vivo, by taking advantage of existing DNA methylation data from thousands of tumors across major cancer types. This will provide valuable insights into the personalized diagnosis and treatment of many human cancers. 1

National Institute of Health (NIH)
National Cancer Institute (NCI)
Research Project (R01)
Project #
Application #
Study Section
Genomics, Computational Biology and Technology Study Section (GCAT)
Program Officer
Li, Jerry
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Baylor College of Medicine
Anatomy/Cell Biology
Schools of Medicine
United States
Zip Code
Su, Jianzhong; Huang, Yung-Hsin; Cui, Xiaodong et al. (2018) Homeobox oncogene activation by pan-cancer DNA hypermethylation. Genome Biol 19:108