Abstract

Aberrant DNA hypermethylation is a near universal hallmark of human cancer.  Upon DNA replication, these methylation profiles are copied onto the newly-synthesized DNA strand by DNA methytransferase 1 (DNMT1).Reversal of DNA methyl marks by a hypomethylating agent such as decitabine offers treatment for cancers such as acute myeloid leukemia.  However, considerable drawbacks exist, including IV administration, poor PK, lack of selectivity and incorporation into DNA.  This indirect, irreversible inhibition of the entire DNMT family (DNMT1, 3a and 3b) induces significant dose-limiting toxicity, preventing sufficient target engagement for maximal demethylation. The past few decades have seen considerable interest in the development of potent, selective DNMT1 inhibitors with little, if any, success. Through an industry / not-for-profit collaboration, we have delivered agents which overcome these limitations.  A  high-throughput screen and robust screening cascade revealed a single molecule which was found to be non-DNA incorporating and highly selective for DNMT1 over DNMT3a or DNMT3b.  Optimization of the compound led to compounds that induced decreases in global DNA methylation, activated silenced genes, and inhibited cancer cell growth.  In vivo investigations with these agents decreased DNA methylation and decreased tumor growth at well-tolerated doses, without the toxicity observed with decitabine.