Epigenetic therapies targeting histone lysine methylation: complex mechanisms and clinical challenges

Abstract

As epigenetic therapies continue to gain ground as potential treatment strategies for cancer and other diseases, compounds that target histone lysine methylation and the enzyme complexes represent a major frontier for therapeutic development. Clinically viable therapies targeting the activities of histone lysine methyltransferases (HKMT) and demethylases (HKDMs) have only recently begun to emerge following FDA approval of the EZH2 inhibitor tazemetostat in 2020 and remain limited to compounds targeting the well-studied SET domain-containing HKMTs and their opposing HKDMs. These include the H3K27 methyltransferases EZH2/EZH1, the singular H3K79 methyltransferase DOT1L, and the H3K4 methyltransferase MLL1/COMPASS as well as H3K9 and H3K36 methyltransferases. They additionally include the H3K4/9-preferential demethylase LSD1 and the H3K4-, H3K27-, and H3K36-preferential KDM5, KDM6, and KDM2 demethylase subfamilies, respectively. This Review discusses the results of recent clinical and preclinical studies relevant to all of these existing and potential therapies. It provides an update on advancements in therapeutic development, as well as more basic molecular understanding, within the past 5 years approximately. It also offers a perspective on histone lysine methylation that departs from the long-predominant "histone code" metaphor, emphasizing complex-disrupting inhibitors and proximity-based approaches rather than catalytic domain inhibitors in the outlook for future therapeutic development.

Figure 1. An overview of histone lysine methylation.

(A) Histones are the protein subunits of nucleosomes, essential chromatin structures in which approximately 147 bp of genomic DNA is wound around a histone core. Lysine residues within the globular core or the unstructured tails of histone proteins can be mono, di, or trimethylated (me1, -2, or -3) by the catalytic activity of histone lysine methyltransferase (HKMT) enzymes. (B) MLL1/COMPASS (left) and other HKMT in the trithorax/COMPASS family deposit histone 3 lysine 4 (H3K4) methylation “marks” that are considered to act as positive regulators of gene expression, while EZH2 and other HKMT in the polycomb family deposit H3K27 methylation that is considered to facilitate transcriptional repression. Subunits of the core WRAD module (present in all COMPASS complexes) are shown with labels in Figure 3A. (C) Methylated histone lysines can recruit a plethora of different methylation, residue, and context-specific effector proteins to regulate gene expression. SGF29, PHF1, and EED are examples of histone methylation-binding effector proteins. SGF29 recruits a version of the PRC2 complex by binding H3K36me3 via its Tudor domain. PHF1 recruits the SAGA transcriptional coactivator complex by binding H3K4me3 via its Tudor domain. The effector function of the PRC2 complex subunit EED, which binds to H3K27me3 via its WD40 repeat domain, is illustrated in Figure 2A. (D) Methylation is removed by the activity of lysine demethylase (HKDM) enzymes.

Figure 2. An emerging inhibitor class indirectly targets EZH2 activity via the PRC2 subunit EED.

The PRC2 complex catalytic subunit EZH2 is a target for cancer therapy because its H3K27me3-depositing activity silences the expression of tumor suppressor genes. However, new targets are needed to overcome acquired resistance to existing EZH2 inhibitors. (A) Baseline EZH2 activity is enabled by interaction with the PRC2 subunit EED, which can also bind to H3K27me3, stimulating 10- to 20-fold greater EZH2 activity toward H3K27 on neighboring nucleosomes to facilitate “spreading” of this repressive modification. (B) EZH2 inhibitors such as the FDA-approved compound tazemetostat directly target the catalytic activity of EZH2 via the EZH2 SET domain. (C) EZH2 activity can be indirectly targeted via small molecules that disrupt EED’s EZH2-interacting pocket or its H3K27me3-binding pocket. (D) EED-targeting molecules can be adapted into proteolysis-targeting chimeras (PROTACs) by fusion via a linker peptide to a ligand of an E3 ubiquitin ligase, which targets EED for proteasomal degradation and may also lead to degradation of EZH2 and other PRC2 subunits.

Figure 3. Noncatalytic inhibitors evict MLL1 and its oncofusion partners from chromatin by disrupting the menin-MLL1 interface.

(A) The COMPASS methyltransferase MLL1 (left) epigenetically modifies histone 3 lysine 4 with dimethylation and trimethylation (H3K4me2/3) via the catalytic activity of its SET domain. MLL1/COMPASS contains MLL1N and MLL1C subunits, products of MLL1 cleavage by taspase1. Menin tethers the MLL1/COMPASS complex to chromatin via its interaction with LEDGF. It was recently reported that menin also binds chromatin via recognition of histone 3 lysine 79 dimethylation (H3K79me2), a modification that is exclusively deposited by the methyltransferase DOT1L, which is a frequent oncofusion partner of MLL1 in leukemias driven by 11q23 translocations. The SET domain–containing C-terminal region of MLL1 is lost in MLL1 oncofusion proteins. (B) Disruption-based inhibitors target an N-terminal domain of MLL1 that binds menin. Inhibitors that disrupt the menin-MLL1 interface can evict both MLL1/COMPASS and MLL1 oncofusion proteins from chromatin. (C) MLL1 oncofusion proteins, which lack the taspase1 cleavage site present in normal MLL1, form more stable chromatin-associated complexes than normal MLL1/COMPASS (left). Stabilizing MLL1/COMPASS (e.g., by targeting taspase1 activity) is a potential strategy for rebalancing chromatin occupancy in favor of normal MLL1/COMPASS (right).