Prosecuting the HMTome

The human genome encodes more than 70 enzymes that catalyze the methylation of lysine or arginine residues on proteins. These protein lysine methyltransferases (PKMTs) and protein arginine methyltransferases (PRMTs) are collectively referred to as the human histone methyltransferase (HMT) target class or the human HMTome.

In surveying the epigenetic enzymes of the human genome, HMT family of enzymes stand out as particularly rich source of validated, tractable drug targets. The family is large, these enzymes share a common catalytic mechanism, and a number of HMTs have been demonstrated to play critical roles in cancer and other serious human diseases.

The scientific leadership of Epizyme has therefore chosen to initially focus our efforts on the discovery of small molecule inhibitors of HMTs as novel therapeutic agents for the treatment of human cancers.

(click for enlargement)

Figure 1 - Family tree for the human SET-domain protein lysine methyltransferases (PKMTs), the largest family of human HMTs. The family is divided into four major clusters, indicated by the color-coding in this illustration. There are fifty-two human PKMTs; fifty-one SET-domain enzymes and one non-SET-domain enzyme (DOT1L). This figure was redrawn using data from Copeland et al. (manuscript submitted for publication).

As the names imply, the PKMTs transfer a methyl group onto specific lysine residues of substrate proteins and the PRMTs transfer a methyl group onto specific arginine residues of their substrate proteins. Both of these classes of HMTs share in common a universal methyl donating cofactor, S-adenosyl methionine (SAM). Thus, there is a clear analogy between the protein kinases (a well know family of tractable drug targets for cancer) and HMTs: Both constitute families of enzymes that catalyze group transfer reactions utilizing a common, donor cofactor – adenosine triphosphate (ATP) in the case of kinases and SAM in the case of HMTs. Furthermore, the structural diversity found among the ATP binding pockets of kinases provides an almost universal locus for drug molecule binding within this enzyme family and enables selectivity for specific kinase targets. As a result, many kinase inhibitors are in current clinical use and/or clinical trials for the treatment of cancers and other human diseases. We believe this will also be the case for HMT targets.

Therefore, our strategy at Epizyme is to target the HMTs as a family of SAM-utilizing enzymes, making full use of lessons learned from kinases as drug targets and exploiting technological platforms that allow parallel processing of multiple enzymes of similar mechanism.

Figure 2 - From one cofactor, many inhibitors of different enzymes can be discovered. In the case of protein kinases, the human genome encodes >500 enzymes that selectively phosphorylate different protein substrates, all using the same phosphate donor, ATP. The ATP binding pocket of different kinases have unique molecular architectures, and this has been exploited to discover a large number of kinase-selective inhibitors that are in clinical use or clinical trials today. In an analogous manner, all of the enzymes of the human HMTome utilize a common methyl-donating cofactor, SAM (S-adenosyl-methionine).

The chemical and biochemical platforms in place at Epizyme are designed to effectively and efficiently identify small molecule inhibitors of HMTs as starting points for drug discovery efforts on select targets. At the same time, our biological platforms are designed to explore the role of the HMTome enzymes in disease processes and to define the impact of selective disruption of HMTome enzyme activity on cancer cells and other cell types of interest.

We have identified specific enzymes among the collection of human HMTs (the human HMTome) that have very strong validation as cancer targets and are using the tools of modern biology to understand fully the pathogenic roles of these enzymes in cancer and the consequences of inhibiting them. We are also developing tools to prosecute the entire HMTome as a broad enzyme class and are taking advantage of the efficiencies and economies of enzyme class-focused medicinal chemistry efforts.

Figure 3– Epizyme’s R&D Overarching Strategy: Prosecuting human HMTs as a Target Class.

Chemistry Platforms
Our chemistry platforms are being developed to exploit the commonality of catalytic mechanism seen within the HMT target class to create enriched chemical libraries of small molecule chemotypes targeting this catalytic commonality. At the same time, these libraries are continuously expanded and embellished to derive target selectivity through chemical diversification.

In particular our expertise in enzyme-targeted drug discovery - including structural biology and molecular modeling - is enabling us to develop proprietary libraries that are enriched for HMT inhibitors.

Biochemistry Platforms
We are also developing a suite of biochemical platforms, taking advantage of proprietary assay formats that allow us to process arrays of HMT family members simultaneously, rather than in the traditional approach of one target at a time. This allows us to exploit fully the diversity of our chemical platforms with maximum effectiveness and efficiency, by parallel processing of lead optimization activities on multiple targets.

Biology Platforms
The HMTome contains well over 70 epizymes; yet only a portion of these are likely to play a causal role in serious human diseases. Hence, an initial objective of our HMTome biology platforms is to provide a rational, scientifically rigorous basis for target prioritization and selection. Through a combination of internal research and external collaborations our biology group is focused on the disease relevance of specific HMT targets and the impact of modulating their catalytic activity on pathogenesis and disease progression. For these targets, our biology organization is committed to understanding the physiological niche and specific mechanisms of each enzyme that we target for drug discovery efforts.

Chemical Biology Platforms
Finally, our chemical biology platforms are providing technologies through which we can assess questions of target occupancy, compound selectivity, HMT composition and HMT variation among drug-sensitive and drug-resistant samples of various origins. In this manner, we can address drug-target interactions in situ for samples from cell culture, animal tissues and patient samples during lead optimization, candidate development and, perhaps, during translation into the clinic. These platforms are supplementing more traditional approaches to pharmacodynamic and patient stratification biomarker discovery.