Molecular basis for aberrant de novo DNA methylation in cancer

PROJECT SUMMARY
Methylation of cytosine in the context of CpG dinucleotides is an essential and highly conserved chromatin
modification. It plays a critical role in gene regulation, cell differentiation and organismal development. Aberrant
patterns of DNA methylation are commonly observed in human diseases including cancer. CpG islands (CGIs),
regions with a high frequency of CpG sites found at the promoters of ~40% mammalian genes, are devoid of
DNA methylation in normal tissues but become hypermethylated in tumors, leading to the silencing of many
tumor suppressor genes. However, the molecular mechanisms underlying CGI hypermethylation in cancer
remain poorly understood, and therefore therapeutic strategies that specifically target CGI hypermethylation are
lacking. DNA methylation is established by the DNMT3 family of de novo DNA methyltransferases. The DNMT3
family includes two catalytically active members, DNMT3A and DNMT3B. DNMT3A has two splice variants
DNMT3A1 and DNMT3A2. While DNMT3A2 and DNMT3B are expressed during early development, DNMT3A1
is the main de novo DNA methyltransferase expressed in differentiated tissues. We and others have previously
demonstrated that DNMT3A1 contains regulatory domains interacting with histone post-translational
modifications (PTMs) to guide its genomic targeting to regions marked by H3K36 di- and tri-methylation
(H3K36me2/3). Our collaborative preliminary studies have identified an additional functional domain unique to
DNMT3A1 that facilitates its interaction with histone H2AK119 mono-ubiquitination (H2AK119Ub), a PTM
catalyzed by Polycomb Repressive Complex 1 (PRC1). These findings resonate with meta-analyses of patient
tumor samples, which revealed a strong correlation between CGIs that gain methylation in cancers and those
regulated by Polycomb complexes in normal tissues. We will pursue a hypothesis that the redistribution of
DNMT3A1 from H3K36 methylation to H2AK119ub-marked CGIs drives CGI hypermethylation and neoplastic
transformation. To test this hypothesis, we will employ a multidisciplinary approach that leverages the
complementary expertise of Armache and Lu labs, spanning from chromatin biochemistry, structural study to
cancer biology and epigenomics. These studies will reveal the molecular mechanisms of DNMT3A1 regulation
by H3K36 methylation and H2AK119Ub and uncover how the balance between these PTMs mediates the
recruitment, activity and function of DNMT3A1 in healthy tissues and tumors. Expected results will provide an
enhanced understanding of the dynamics, cause and consequence of CGI hypermethylation – a molecular
hallmark of human cancers that remains a key focus of the field. Furthermore, they will lay the foundation for
developing inhibitors that target the interaction of DNMT3A1 with H2AK119Ub to specifically reverse cancer-
associated CGI hypermethylation, which are expected to show less toxicity compared to current FDA-approved
unspecific DNA hypomethylating agents.

Grant Number: 5R01CA266978-05
NIH Institute/Center: NIH
Principal Investigator: Karim Armache