Dysfunction of the Polycomb protein RYBP and of 5-methylcytosine oxidases leads to widespread CpG island hypermethylation and cell transformation

Abstract

DNA hypermethylation is a hallmark of cancer and predominantly affects CpG island regions. Although this phenomenon was first described more than three decades ago, its mechanisms have remained unknown. Since CpG island hypermethylation is strongly biased towards Polycomb target genes, we reasoned that dysfunction of Polycomb repression complexes (PRCs) may underlie CpG island hypermethylation. We observed that a few genes coding for components of the PRC1 complex are downregulated in many cancer types. We focused on RYBP, a key activator of variant PRC1 complexes responsible for H2AK119 monoubiquitylation. We inactivated RYBP in nontumorigenic bronchial epithelial cells and observed a limited extent of DNA hypermethylation. Considering that tumors are deficient in 5-methylcytosine oxidase (TET protein) function as documented by substantially reduced levels of 5-hydroxymethylcytosine in all solid tumors, we then inactivated TET1, TET2, and TET3 in bronchial cells, individually and in combination. Using quadruple knockouts of RYBP and all three TET proteins, we observed widespread hypermethylation of H2AK119Ub1-marked CpG islands affecting almost 4,000 target genes. This hypermethylation closely mirrored the DNA hypermethylation landscape observed in human lung tumors. These cells showed aberrant methylation and dysregulation of several cancer-relevant pathways including cell cycle control genes, defects in the Hippo pathway and overexpression of AP-1 transcription factor genes. As a result, the quadruple knockout bronchial cells acquired properties of a transformed phenotype, including efficient growth in soft agar and formation of squamous cell carcinomas in immune-compromised mice. Our data provide a long- sought mechanism for DNA hypermethylation in cancer and explain how such hypermethylation leads to cell transformation. Cancer formation, therefore, is achievable by misregulation of two epigenetic pathways without introduction of cancer driver mutations.