Pharmacologic Targeting of the Cancer Epigenome

Abstract

Epigenetic dysregulation is increasingly appreciated as a hallmark of cancer, including disease initiation, maintenance, and therapy resistance. As a result, there have been advances in the development and evaluation of epigenetic therapies for cancer, revealing significant promise, but also challenges. Three epigenetic inhibitor classes are approved in the United States, and many more are currently undergoing clinical investigation. In this Review, we discuss recent developments for each epigenetic drug class and their implications for therapy, as well as highlight new insights into the role of epigenetics in cancer.

Epigenetics of Cancer

Epigenetics is broadly defined by the non-genetic biological processes underlying the regulation of chromatin architecture with no changes to the primary DNA sequence. Epigenetic regulation is paramount to establishing the gene expression patterns that define cell identity, differentiation, and response to environmental stimuli¹. Dysregulation of epigenetic processes has long been appreciated as a core feature of cancer, with mutations in genes encoding epigenetic regulatory proteins among the most frequently observed². In recent years, manifestations of epigenetic dysregulation, such as phenotype plasticity and non-genetic epigenetic reprogramming, have also been incorporated as fundamental hallmarks of cancers³.

The process of interpreting DNA-encoded information to guide cellular phenotypes involves a complex network of DNA elements, transcription factors, epigenome modifiers, and biophysical interactions. DNA is packaged tightly in nucleosomes consisting of 8 histone protein subunits that can be modified to facilitate regions of high transcription (euchromatin) or low transcription (heterochromatin)⁴. During transcription, nucleosome positioning is adjusted through ATP-dependent SWI/SNF chromatin remodeling complexes, allowing access to cis-regulatory elements on DNA, such as promoters and enhancers, to which trans-regulatory transcription factors can bind in a sequence-dependent manner to modulate transcription through RNA polymerase II activity and its associated complexes (e.g., CDK9/CCNT1 (P-TEFb), Mediator, etc.)^5,6. Among the most studied epigenetic mechanisms regulating transcription are post-translational modifications (PTMs), either on cytosines of DNA or on amino acids of histone tails of nucleosomes, that facilitate the recruitment of chromatin modifiers, transcription factors, or compact the DNA⁷. Historically, epigenetic regulation has been defined by families of enzymatic writers and erasers that maintain the balance of deposition and removal of histone tail PTMs and the protein readers that recognize these PTMs⁸. Other epigenetic regulators, including scaffolding proteins that facilitate histone regulation, such as Menin, or chromatin remodelers (e.g., the BAF complex), have emerged as therapeutic targets in cancer, but do not fit cleanly into the reader, writer, and eraser categories.

Therapeutic Targeting of Epigenetic Proteins

Epigenetic proteins that regulate gene expression are frequently mutated in cancer, and as their mechanisms are increasingly being elucidated, therapeutic agents that alter their function are being developed. Here, we survey the history and recent advances in targeting epigenetic proteins in cancer, describing Food and Drug Administration (FDA)-approved small molecules and those in clinical trials. Currently, epigenetic inhibitors representing three distinct classes are approved by the FDA in the United States: DNA methyltransferase (DNMT), Histone deacetylase (HDAC), and Enhancer of Zeste Homolog-2 (EZH2) inhibitors. However, at least ten other classes of epigenetic inhibitors are currently being investigated in clinical trials (Figure 1). The development of these epigenetic chemical modulators has significantly expanded our understanding of how cancer cells utilize epigenetic and transcriptional machinery to promote their growth and survival.

Figure 1: Epigenetic Targets in Cancer.

Major pharmacologic drug classes that target epigenetic processes and are currently in clinical testing or approved by government regulatory agencies. * indicates FDA approval.

Targeting Common Essential Chromatin Regulators

Some epigenetic regulators are common essential genes and their inhibition mimics cytotoxic chemotherapy while other epigenetic regulators may be selective dependencies or synthetic lethal in specific cancer subsets. While these differences had to be empirically determined in the past, genome-scale loss-of-function screening efforts in large panels of cancer cell lines have generated data sets, such as the Cancer Dependency Map (DepMap)⁹, in which genes can be broadly classified as common essential, selectively enriched, or non-essential across over 1000 cancer cell line models. Using these data, we can predict whether response to knockout of a particular epigenetic gene and its small molecule therapies by proxy are more likely to perform like a cytotoxic chemotherapy (e.g., TOP1 inhibitors) versus a targeted therapy (e.g., EGFR inhibitor)¹⁰. Common essential genes, including DNA methyltransferases (e.g., DNMT1), histone deacetylases (e.g., HDAC3) and bromodomain and extra-terminal domain (BET) proteins (e.g., BRD4), are among current epigenetic targets of interest for clinical development.

DNA Methyltransferase Inhibitors and Histone Deacetylase Inhibitors

DNMT inhibitors and HDAC inhibitors belong to the first developed classes of small molecules targeting epigenetic regulators and have been extensively studied and reviewed elsewhere^11,12. The clinical impact of these drugs has been notable, but almost completely limited to hematologic malignancies. DNMT inhibitors 5’-azacytidine and decitabine are FDA approved for myelodysplastic syndrome, 5’-azacytidine for juvenile myelomonocytic leukemia, and 5’-azacytidine and decitabine are commonly used off-label for treatment of acute myeloid leukemia (AML) and are FDA approved in combination with venetoclax in older adults diagnosed with AML. HDAC inhibitors are FDA approved for cutaneous T-cell lymphomas (vorinostat, romidepsin)¹³, peripheral T-cell lymphoma (belinostat)¹⁴, and multiple myeloma (panobinostat),¹⁵ and many other clinical candidate HDAC inhibitors are being studied in solid and hematologic malignancies (Figure 1 and Table 1).

Table 1:

Summary of ongoing and completed clinical trials for epigenetic drugs.

Drug	Disease State	In combination with	Study Number	Phase	Status	Reason for withdrawal/termination	Reference articles (PMID)
DNMT inhibitors
5’-azacytidine *	MDS, juvenile myelomonocytic leukemia
Decitabine *	MDS
guadecitabine** (SGI-110)	AML	Standard of care	NCT02920008	Phase 3	Completed
	MDS, Chronic myelomonocytic leukemia		NCT02907359	Phase 3	Completed
	AML		NCT02348489	Phase 3	Completed		37276510
HDAC inhibitors
romidepsin *	Cutaneous T-cell lymphoma				FDA approved
vorinostat *	Cutaneous T-cell lymphoma				FDA approved
belinostat *	Peripheral T-cell lymphoma				FDA approved
panobinostat *	MM				FDA approved
tucidinostat (chidamide) **	Peripheral T-cell lymphoma	azacitidine and CHOP	NCT05075460	Phase 3	Not yet recruiting
	DLBCL	R-CHOP	NCT04231448	Phase 3	Active, not recruiting
	ER+ Breast cancer	exemestane	NCT02482753	Phase 3	Completed		31036468
	Periperal T cell lymphoma		NCT04668690	Phase 3	Not yet recruiting
	ER+ Breast cancer	exemestane	NCT05253066	Phase 3	Not yet recruiting
	Ph-like ALL	dasatinib	NCT03564470	Phase 3	Recruiting
	DLBCL or T cell lymphoma	cladribine, gemcitabine and busulfan or BCNU, etoposide, cytarabine, and melphalan.	NCT05466318	Phase 3	Recruiting
	ER+/HER2- advanced Breast cancer after CDK4/6 failure		NCT05335473	Phase 3	Recruiting
	Advanced CRC	tucidinostat + sintilimab + bev vs. second-line FOLFIRI + Bev	NCT05768503	Phase 3	Recruiting
	T-cell lymphoma	Compared versus SHR2554	NCT06122389	Phase 3	Recruiting
	Stage III/IV extranodal NK/T-cell lymphoma	tucidinostat + anti-PD1 + pegaspargase	NCT06255795	Phase 3	Not yet recruiting
entinostat (MS-275) **	ER+/HER2- breast cancer	exemestane	NCT03538171	Phase 3	Active, not recruiting
	ER+/HER2- breast cancer	exemestane	NCT02115282	Phase 3	Active, not recruiting
Citarinostat (ACY-241)	Advanced solid malignancies	paclitaxel	NCT02551185	Phase 1	Completed		35070991
	NSCLC	nivolumab	NCT02635061	Phase 1	Active, not recruiting
	MM	pomalidomide and dexamethasone	NCT02400242	Phase 1	Active, not recruiting
	Melanoma	ipilimumab and nivolumab	NCT02935790	Phase 1	Completed
	Smoldering MM	PVX-410 + citarinostat +/− lenalidomide	NCT02886065	Phase 1	Active, not recruiting
KA2507	Biliary tract cancer		NCT04186156	Phase 2	Withdrawn	Study not progressing
	Advanced solid malignancies		NCT03008018	Phase 1	Completed		33947698
abexinostat (PCI-24781)	DLBCL, Mantle Cell lymphoma	ibrutinib	NCT03939182	Phase 1/2	Active, not recruiting
	Follicular lymphoma		NCT03934567	Phase 2	Recruiting
	Follicular lymphoma		NCT03600441	Phase 2	Active, not recruiting
	DLBCL		NCT03936153	Phase 2	Recruiting
	Breast and gynecologic cancers	palbociclib and fulvestrant	NCT04498520	Phase 1	Withdrawn	Funding
	Renal cell carcinoma	pazopanib	NCT03592472	Phase 3	Recruiting
	Advanced solid malignancies	pembrolizumab	NCT03590054	Phase 1	Active, not recruiting
	Non-Hodgkins lymphoma		NCT04024696	Phase1/2	Active, not recruiting
	Sarcoma	doxorubicin	NCT01027910	Phase 1/2	Completed		25536954
	Non-Hodgkins lymphoma, MM, leukemia		NCT01149668	Phase 1	Completed
	Advanced solid malignancies	pazopanib	NCT01543763	Phase 1	Active, not recruiting
	Non-Hodgkins lymphoma, Mantle cell lymphoma		NCT00724984	Phase 1/2	Completed		26482040
	Advanced solid malignancies		NCT00473577	Phase 1	Completed
	Recurrent glioma	Temodar + abexinostat	NCT05698524	Phase 1	Recruiting
pracinostat (SB939)	MDS	azacitidine	NCT03151304	Phase 2	Terminated	Sponsor decision
	myelofibrosis	ruxolitinib	NCT02267278	Phase 2	Completed		30632841
	AML	azacitadine	NCT01912274	Phase 2	Completed		30760466
	MDS	azacitadine or decitabine	NCT01993641	Phase 2	Completed
	MDS	azacitidine	NCT01873703	Phase 2	Completed		28094841
	AML	gemtuzumab	NCT03848754	Phase 1	Completed		36401944
	AML	azacitadine	NCT03151408	Phase 3	Terminated	Lack of Efficacy
	translocation-associated sarcomas		NCT01112384	Phase 2	Completed		25632070
	Advanced solid and hematologic malignancies		NCT00741234	Phase 1	Completed		28841236
	CRPC		NCT01075308	Phase 2	Completed		25983041
	Myelofibrosis		NCT01200498	Phase 2	Completed
	Pediatric solid and hematologic tumors		NCT01184274	Phase 1	Completed
	Advanced solid malignancies		NCT00504296	Phase 1	Completed
fimepinostat (CUDC-907)	DIPG, high-grade glioma, and medulloblastoma		NCT03893487	Phase 1	Active, not recruiting
	DLBCL, HGBL		NCT01742988	Phase 1	Completed		28860342
	Metastatic and locally advanced thyroid cancer		NCT03002623	Phase 2	Terminated
	Advanced Cancers		NCT02307240	Phase 1	Completed
	RR DLBCL		NCT02674750	Phase 2	Completed
	Children and young adults with RR solid tumors and lymphoma		NCT02909777	Phase 1	Active, not recruiting
quisinostat (JNJ-26481585)	epithelial ovarian cancer	paclitaxel + carboplatin	NCT02948075	Phase 2	Completed
	T cell cutaneous lymphoma		NCT01486277	Phase 2	Completed
	SCLC and ovarian cancer	standard chemotherapy	NCT02728492	Phase 1	Completed
	Leukemias and MDS		NCT00676728	Phase 1	Terminated	Sponsor decision
	Advanced solid malignancies or lymphoma		NCT00677105	Phase 1	Completed
	MM	bortezomib and dexamethasone	NCT01464112	Phase 1	Completed
ricolinostat (ACY-1215)	MM	pomalidomide and dexamethasone	NCT01997840	Phase 1/2	Active, not recruiting
	Lymphoid malignancies		NCT02091063	Phase 1/2	Completed		33458921
	Breast cancer	nab-paclitaxel	NCT02632071	Phase 1	Completed		36585452
	MM	bortezomib and dexamethasone	NCT01323751	Phase 1/2	Completed
	MM	pomalidomide and dexamethasone	NCT02189343	Phase 1	Completed
	Cholangiocarcinoma		NCT02856568	Phase 1	Withdrawn	Site dropped study
	Gynecologic cancers	paclitaxel and bevacizumab	NCT02661815	Phase 1	Terminated	Sponsor decision
	Chronic lymphocytic leukemia	ibrutinib or idelalisib	NCT02787369	Phase 1	Active, not recruiting
	MM	lenalidomide and dexamethasone	NCT01583283	Phase 1	Completed		27646843
nanatinostat (VRx-3996)	EBV-positive lymphomas	valganciclovir	NCT05011058	Phase 2	Recruiting
	EBV-positive solid tumors	valganciclovir and pembrolizumab	NCT05166577	Phase 1/2	Recruiting
	EBV-positive lymphomas	valganciclovir	NCT03397706	Phase 1/2	Recruiting
	Advanced Cancers		NCT06302140	Phase 1	Active, not recruiting
ivaltinostat (CG200745)	pancreatic adenocarcinoma	capecitabine	NCT05249101	Phase 1/2	Recruiting
	Advanced pancreatic cancer	Gemcitabine + erlotinib	NCT02737228	Phase1/2	Unknown
Bocodepsin (OKI-179)	Advanced Solid Tumors		NCT03931681	Phase 1	Completed
		MEK inhibition	NCT05340621	Phase 2	Active, not recruiting
CUDC-101	Advanced solid malignancies		NCT01702285	Phase 1/2	Terminated	Unknown
	Advanced solid malignancies		NCT00728793	Phase 1	Completed
	advanced head and neck, gastric, breast, liver, and NSCLC tumors		NCT01171924	Phase 1	Completed
	Head and neck cancers	radiotherapy and cisplatin	NCT01384799	Phase 1	Completed		25573383
MPT0E028	Advanced solid malignancies		NCT02350868	Phase 1	Completed
BET inhibitors
Trotabresib (CC-90010)	Astrocytoma, GBM		NCT04047303	Phase 1	Active, not recruiting
	non-hodgkins lymphoma, advanced solid tumors		NCT03220347	Phase 1	Active, not recruiting
	GBM	temozolomide +/− radiotherapy	NCT04324840	Phase 1	Active, not recruiting
	pediatric solid tumors (Neuroblastoma, medulloblastoma, sarcomas) and lymphomas		NCT03936465	Phase 1	Recruiting
	HER2+ metastatic breast cancer	Vinorelbine + radiation	NCT06137651	Phase 1/2	Not yet recruiting
BMS-986158	pediatric solid tumors (Neuroblastoma, medulloblastoma, sarcomas) and lymphomas		NCT03936465	Phase 1	Recruiting
	Myelofibrosis	alone or with ruxolitinib or fedratinib	NCT04817007	Phase 1/2	Active, not recruiting
	Advanced cancers	+/− nivolumab	NCT02419417	Phase 1/2	Completed		36077617
	MM	CC-92480	NCT05372354	Phase 1/2	Recruiting
PLX51107	AML, MDS	azacitidine	NCT04022785	Phase 1	Completed
	Advanced cancers; lymphoma, AML; MDS		NCT02683395	Phase 1	Terminated	Sponsor Decision
AZD5153	Malignant solid tumors (Ovarian, Breast, Pancreatic, Prostate) and Lymphoma	+/−olaparib	NCT03205176	Phase 1	Completed
	Non hodgkins Lymphoma, DLBCL	acalabrutinib	NCT03527147	Phase 1	Completed
	AML	venetoclax	NCT03013998	Phase 1/2	Recruiting
SF1126	Hepatocellular carcinoma	nivolumab	NCT03059147	Phase 1	Terminated	Lack of recruitment
	Neuroblastoma		NCT02337309	Phase 1	Terminated	Low accrual
	Metastatic Squamous Neck Cancer With Occult Primary Squamous Cell Carcinoma		NCT02644122	Phase 2	Terminated	Slow enrollment
	Advanced Cancers		NCT00907205	Phase 1	Completed
Mivebresib (ABBV-075)	Breast cancer, NSCLC, AML, MM, prostate cancer, SCLC, Non-Hodgkins Lymphoma		NCT02391480	Phase 1	Completed		33934351
	Myelofibrosis	+/− navitoclax or ruxolitinib	NCT04480086	Phase 1	Terminated
Molibresib (GSK525762)	AML, Non-Hodgkins lymphoma, MM		NCT01943851	Phase 1	Completed		36350312
	Ras-mutated cancers (SCLC, NSCLC, colorectal cancer, pancreatic cancer)	trametinib	NCT03266159	Phase 2	Withdrawn	Unknown
	ER+/HER2- breast cancer	fulvestrant	NCT02964507	Phase 1	Terminated
	NMC, SCLC, NSCLC, colorectal cancer, neuroblastoma, CRPC, breast cancer, MYCN-driven solid tumors		NCT01587703	Phase 1	Completed		34724226
	CRPC	androgen deprivation	NCT03150056	Phase 1	Terminated	protocol-defined futility
	Solid tumors and lymphoma	entinostat	NCT03925428	Phase 1	Withdrawn	Unknown
	NMC	chemotherapy	NCT04116359	Phase 1/2	Withdrawn	Unknown
birabresib (MK-8628/OTX015)	NMC, Breast Cancer, NSCLC, CRPC		NCT02698176	Phase 1	Terminated	Limited efficacy
	NMC, Breast Cancer, NSCLC, CRPC, Pancreatic		NCT02259114	Phase 1	Completed		29733771
	AML, DLBCL		NCT02698189	Phase 1	Terminated
	AML, ALL, DLBCL, MM		NCT01713582	Phase 1	Completed		27063978
	GBM		NCT02296476	Phase 2	Terminated	Limited efficacy
	Newly diagnosed AML not candidate for standard chemotherapy	azacitidine	NCT02303782	Phase 1/2	Withdrawn
BI 894999	Advanced solid tumors or DLBCL		NCT02516553	Phase 1	Completed		37556913
BAY 1238097	Solid tumors		NCT02369029	Phase 1	Terminated	Toxicity	30711772
pelabresib (CPI-0610)	MPNST		NCT02986919	Phase 2	Terminated	Low enrollment
	MM		NCT02157636	Phase 1	Completed
	Lymphoma		NCT01949883	Phase 1	Completed		36923307
	Advanced malignancies		NCT05391022	Phase 1	Completed
	Myelofibrosis	+/− ruxolitinib	NCT02158858	Phase 2	Active, not recruiting
	Myelofibrosis	+/− ruxolitinib	NCT04603495	Phase 3	Active, Not recruiting
INCB054329	Advanced malignancies		NCT02431260	Phase1/2	Terminated	PK variability	31527168
INCB057643	Myelofibrosis	+/− ruxolitinib	NCT04279847	Phase 1	Recruiting
	Advanced solid malignancies	select agents	NCT02711137	Phase 1/2	Terminated	Safety concerns	31527168
	Advanced solid malignancies	pembrolizumab and epacadostat	NCT02959437	Phase 1/2	Terminated	Sponsor Decision	37087488
FT-1101	AML, MDS, NHL	+/− azacitidine	NCT02543879	Phase 1	Completed
ODM-207	Advanced solid malignancies		NCT03035591	Phase 1/2	Completed		32989226
ZEN-3694	Advanced solid malignancies	chemotherapy + capecitabine	NCT05803382	Phase 1	Recruiting
	Metastatic CRPC	enzalutamide, pembrolizumab	NCT04471974	Phase 2	Recruiting
	Advanced solid malignancies	Talazoparib	NCT05327010	Phase 2	Recruiting
	Recurrent ovarian and endometrial cancers	M1774	NCT05950464	Phase 1b	Recruiting
	TNBC	Pembrolizumab, standard chemotherapy	NCT05422794	Phase 1b	Recruiting
	Advanced solid tumors with RAS alterations and TNBC	Binimetinib	NCT05111561	Phase 1	Recruiting
	NMC	Etoposide + Cisplatin	NCT05019716	Phase1/2	Recruiting
	NMC, other solid tumors	abemaciclib	NCT05372640	Phase 1	Recruiting
	Advanced solid tumors and lymphomas	Entinostat	NCT05053971	Phase1/2	Recruiting
	TNBC	Talazoparib	NCT03901469	Phase 2	Active. Not recruiting
	Solid tumors	Nivolumab +/− ipilimumab	NCT04840589	Phase1/1b	Recruiting
	CRPC	Enzalutamide	NCT04986423	Phase 2b	Recruiting
	CRPC		NCT02705469	Phase 1	Completed
	CRPC	Enzalutamide	NCT02711956	Phase 1b/2a	Completed		32694156
	Squamous cell lung cancer		NCT05607108	Phase 2	Recruiting
	Ovarian cancer, Peritoneal cancer, fallopian tube cancer	talazoparib	NCT05071937	Phase 2	Recruiting
	CRC	in addition to chemo + cetuximab + encorafenib	NCT06102902	Phase 1	Recruiting
	Advanced solid malignancies	Plus Niraparib	NCT06161493	Phase 1	Recruiting
RO6870810	MM	+/− daratumumab	NCT03068351	Phase 1b	Completed
	Advanced ovarian or TNBC	atezolizumab	NCT03292172	Phase 1b	Terminated	Portfolio prioritization
	AML, MDS		NCT02308761	Phase 1	Completed		33586590
	Advanced solid tumors		NCT01987362	Phase 1	Completed		33311588
	DLBCL, high-grade B-cell lymphoma	Venetoclax +/− rituximab	NCT03255096	Phase 1b	Completed		34581757
EZH2 inhibitors
Tazemetostat *	epithelioid sarcoma and EZH2 mutant follicular lymphoma				FDA approved
	ARID1A mutant cancers		NCT05023655	Phase 2	Recruiting
	malignant mesothelioma		NCT02860286	Phase 2	Completed		35588752
	MPNST		NCT04917042	Phase 2	Recruiting
	INI1-negative solid tumors or synovial sarcoma		NCT02601950	Phase 2	Active, not recruiting
	INI1-negative solid tumors or synovial sarcoma		NCT02601937	Phase 1	Completed
	B-cell NHL with EZH2 mutations		NCT03456726	Phase 2	Completed		34159682
	Gynecologic cancers		NCT03348631	Phase 2	Active, not recruiting
GSK2816126	DLBCL, Follicular lymphoma, MM		NCT02082977	Phase 1	Terminated	No efficacy at MTD	31471312
PF-06821497	SCLC, Follicular lymphoma, CRPC		NCT03460977	Phase 1	Recruiting
SHR2554	Relapsed/Refractory mature lymphoid malignancies		NCT03603951	Phase 1	Recruiting
	CRPC	SHR3680	NCT03741712	Phase 1/2	Terminated	Sponsor decision
	Advanced solid tumors and B-cell lymphomas	SHR1701	NCT04407741	Phase 1/2	Recruiting
	Relapsed/Refractory NHL	SHR1701	NCT05896046	Phase 1/2	Recruiting
	Relapsed/Refractory PTCL	Umbrella study	NCT05559008	Phase 1/2	Recruiting
	Advanced Breast Cancer	Umbrella study	NCT04355858	Phase 2	Recruiting
	TNBC	Umbrella study	NCT03805399	Phase 1/2	Recruiting
	T-cell lymphoma	Compared versus chidamide	NCT06122389	Phase 3	Recruiting
	Peripheral T-cell lymphoma (treatment-naïve)	With cheomtherapy	NCT06173999	Phase 1b/2	Not yet recruiting
CPI-1205	B-cell Lymphomas		NCT02395601	Phase 1	Completed
	Advanced Solid tumors		NCT03525795	Phase 1/2	Completed
	CRPC		NCT03480646	Phase 1/2	Unknown
Valemetostat (DS-3201b)	Sinonasal cancer, squamous NSCLC, Squamous Cell Carcinoma, Head and Neck Cancers	Pembrolizumab	NCT05879484	Phase 1/2	Not yet recruiting
	Relapsed/Refractory PTCL, adult T cell leukemia/lymphoma		NCT04703192	Phase 2	Active, not recruiting
	B-cell lymphoma		NCT04842877	Phase 2	Active, not recruiting
	T-cell leukemia/lymphoma		NCT04102150	Phase 2	Active, not recruiting
	prostate, urothelial, renal cell carcinomas	ipilimumab	NCT04388852	Phase 1	Recruiting
	FL	Rituximab, Lenalidomide	NCT05683171	Phase 1/2	Recruiting
	HER2-low breast cancer	Trastuzumab	NCT05633979	Phase 1	Recruiting
	Hepatocellular carcinoma	With atezolizumab and bevacizumab	NCT06294548	Phase 1b/2	Not yet recruiting
	Advanced solid tumors	With DXd ADC	NCT06244485	Phase 1	Not yet recruiting
	Relapsed or refractor NHL	With CC-99282	NCT03930953	Phase 1/2	Recruiting
HH2853	NHL, advanced solid tumors		NCT04390737	Phase 1/2	Recruiting
EED inhibitor
MAK683	DLBCL and solid tumors		NCT02900651	Phase 1/2	Active, not recruiting
APG-5918	Advanced Solid Tumors and Lymphomas including SMARCB1 mutated sarcomas and EZH2 mutant B cell lymphomas		NCT05415098	Phase 1	Recruiting
ORIC-944	Metastatic Prostate Cancer		NCT05413421	Phase 1	Recruiting
SMARCA2/4 inhibitor
FHD-286	Metastatic Uveal Melanoma		NCT04879017	Phase 1	Active, not recruiting
	AML, MDS		NCT04891757	Phase 1	Recruiting
SMARCA2 degrader
PRT3789	Advanced solid tumors with SMARCA4 mutation	Monotherapy and in como with docetaxel	NCT05639761	Phase 1	Recruiting
BRD9 degrader
FHD-609	Synovial Sarcoma		NCT04965753	Phase 1	Active, not recruiting
CFT8634	SMARCB1-perturbed cancers		NCT05355753	Phase1/2	Terminated	No efficacy in highly treated patients
DOT1L inhibitor
pinometostat (EPZ-5676)	MLL-r leukemias	Azacitidine	NCT03701295	Phase 1	completed
	MLL-r AML	Daunorubicin, Cytarabine	NCT03724084	Phase 1/2	Terminated	Unknown
	MLL-r leukemias		NCT02141828	Phase 1	Completed
	MLL-r leukemias		NCT01684150	Phase 1	Completed		29724899
Menin inhibitors
BMF-219	AML, ALL With KMT2A/MLL1r, NPM1		NCT05153330	Phase 1	Recruiting
	KRAS mutant CRC, NSCLC, PDAC		NCT05631574	Phase 1	Recruiting
Revumenib (SNDX-5613)	Acute Leukemia		NCT05406817	Phase 1	Recruiting
	Acute Leukemia	Chemotherapy	NCT05326516	Phase 1	Active, not recruiting
	AML with MLL Rearrangement or NPM1 Mutation	Cobicistat	NCT04065399	Phase 1/2	Recruiting
	AML	Venetoclax, ASTX727	NCT05360160	Phase 1/2	Recruiting
	Untreated AML - genome screening master study	Many drugs of which SNDX-5613 is one	NCT03013998	Phase 1/2	Recruiting
	AML	Combination with 7+3+midostaurin	NCT06313437	Phase 1	Not yet recruiting
	Leukemias associated with HOX cluster genes		NCT06229912	Phase 2	Not yet recruiting
	Pediatric and young adult AML		NCT06177067	Phase 1	Not yet recruiting
	AML	In combination with BCL2 inhibitor	NCT06284486	Phase 2	Not yet recruiting
	CRC and other solid tumors		NCT05731947	Phase 1/2	Recruiting
	AML		NCT06226571	Phase 1	Not yet recruiting
	AML	Standard chemotherapy for newly diagnosed patients with NPM1 or MLL alterations	NCT05886049	Phase 1b	Recruiting
	AML	And giltertinib in FLT3 mutated AML with concurrent MLL or NPM1 alterations	NCT06222580	Phase 1	Not yet recruiting
	AML		NCT05761171	Phase 2	Active, not recruiting
Ziftomenib (KO-539)	relapsed or refractory AML		NCT04067336	Phase 1/2	Recruiting
	AML	Combination with venetoclax/azacitidine, venetoclax, or 7+3	NCT05735184	Phase 1	Recruiting
	AML	Various combinations	NCT06001788	Phase 1	Recruiting
JNJ-75276617	Acute Leukemia		NCT04811560	Phase 1	Recruiting
	AML with MLL Rearrangement or NPM1 Mutation	Venetoclax, Azacitidine	NCT05453903	Phase 1/2	Recruiting
	AML with MLL Rearrangement, NPM1 Mutation, Nucleoporin Mutation	Chemotherapy	NCT05521087	Phase 1	Not yet recruiting
DSP-5336	Acute Leukemia		NCT04988555	Phase 1/2	Recruiting
DS-1594b	Acute Leukemia	Azacitidine, Venetoclax, mini-HCVD	NCT04752163	Phase 1/2	Completed
P300/CBP inhibitors
FT-7051	CRPC		NCT04575766	Phase 1	Terminated	Sponsor decision
Inobrodib (CCS1477)	NHL, MM, AML, MDS	alone or with cancer-specific regimens	NCT04068597	Phase 1/2a	Recruiting
	CRPC, NSCLC, Breast cancer, advanced solid tumors	alone or in combination with select cancer-specific drug regimens	NCT03568656	Phase 1/2a	Recruiting
EP31670/NEO2734	CRPC, NMC		NCT05488548	Phase 1	Recruiting
KAT6A inhibitor
PF-07248144	ER+ / HER2- breast cancer, NSCLC, CRPC	Fulvestrant, Letrozole, Palbociclib	NCT04606446	Phase 1	Recruiting
LSD1 inhibitors
Iadademstat (ORY-1001)	FLT3 mutant AML	Gilteritinib	NCT05546580	Phase 1	Recruiting
	Relapsed SCLC	In combination with paclitaxel	NCT05420636	Phase 2	Recruiting
	SCLC	In combination with atezolizumab or durvalumab	NCT06287775	Phase 1/2	Not yet recruiting
GSK2879552	MDS	Azacitidine	NCT02929498	Phase 1	Terminated	The risk benefit in MDS does not favor continuation of the study
	SCLC		NCT02034123	Phase 1	Terminated	The risk benefit in SCLC does not favor continuation of the study	31260835
	AML	All-Trans Retinoic Acid (ATRA)	NCT02177812	Phase 1	Terminated	The risk benefit in relapsed refractory AML does not favor continuation of the study
Bomedemstat (IMG-7289 / MK-3543)	Myelofibrosis		NCT05223920	Phase 2	Active. Not recruiting
	Myelofibrosis	Ruxolitinib	NCT05569538	Phase 2	Recruiting
	SCLC	Atezolizumab	NCT05191797	Phase 1/2	Recruiting
	AML	Venetoclax	NCT05597306	Phase 1	Recruiting
	Advanced myeloid malignancies	With or without ATRA	NCT02842827	Phase 1/2	Completed
INCB059872	Ewing Sarcoma		NCT03514407	Phase 1	Terminated	Strategic Business Decision
	Solid Tumors, Hematologic Malignancies	ATRA, Azacitidine, Nivolumab	NCT02712905	Phase 1/2	Terminated	Strategic Business Decision
	Solid Tumors	Pembrolizumab, Epacadostat	NCT02959437	Phase 1/2	Terminated	Study terminated by Sponsor
Seclidemstat (SP-2577)	Solid Tumors		NCT03895684	Phase 1	Completed
	Gynecologic Cancers with SWI/SNF pathway mutations	Pembrolizumab	NCT04611139	Phase 1	Withdrawn	Salarius discontinued support
	Ewing or Ewing-related Sarcomas	Cyclophosphamide, Topotecan	NCT03600649	Phase 1	Active. Not recruiting
	Ewing or Ewing-related Sarcomas		NCT05266196	Phase 1/2	Enrolling by invitation
	MDS, Chronic Myelomonocytic Leukemia	Azacitidine	NCT04734990	Phase 1/2	Recruiting
CC-90011	Prostate Cancer	Abiraterone, Prednisone	NCT04628988	Phase 1	Completed
	SCLC, Squamous NSCLC	Nivolumab	NCT04350463	Phase 2	Completed
	SCLC	Cisplatin, Etoposide, Carboplatin, Nivolumab	NCT03850067	Phase 1	Active, not recruiting
	AML	Venetoclax, Azacitidine	NCT04748848	Phase 1	Terminated	Business objectives have changed
	Solid Tumors, Non-Hodgkin’s Lymphomas	Rifampicin, Itraconazole	NCT02875223	Phase 1	Active, not recruiting
NSD2 inhibitors
KTX-1001	t(4;14) and NSD2 mutant multiple myeloma		NCT05651932	Phase 1	Recruiting

Acronym Disease
DLBCL	Diffuse large B-cell lymphoma
AML	Acute Myeloid Leukemia
MDS	Myelodysplastic syndromes
MM	Multiple myeloma
NSCLC	Non-Small Cell Lung Cancer
SCLC	Small Cell Lung Cancer
CRPC	Castration-resistant prostate cancer
NMC	NUT midline carcinoma
ALL	Acute lymphocytic leukemia
GBM	Glioblastoma Multiforme
MPNST	Malignant Peripheral Nerve Sheath Tumor
DIPG	Diffuse intrinsic pontine glioma
HGBL	High-grade B-cell lymphoma
PTCL	Peripheral T-cell cutaneous lymphoma

^*=FDA-approved drugs, Only approved indications are shown

^**=investigational drugs for which there are ≥ 20 ongoing clinical trials. Only phase 3 trials shown in this table for these drugs.

The limited utility of DNMT and HDAC inhibitors for treating solid tumors highlights the double-edged sword of targeting common essential genes, which have a narrow therapeutic window in part because target proteins are crucial for the survival of both healthy cells and cancer cells. Minimizing toxicity of single-agent DNMT inhibitors may be difficult due the essentiality of DNMT1. However, selective targeting of HDACs may be achievable due to the large number of redundancies between HDACs enzymes. The therapeutic efficacy of HDAC inhibitors is likely mediated through inhibition of Class I HDACs (HDACs1-3,8), although the specificity for individual HDAC members varies between small molecules and off-target effects may partially mediate additional therapeutic effects^16,17. Only HDAC3 is predicted to be common essential and plays a crucial role in chromatin remodeling and gene regulation in most cells/tissues. In contrast HDAC1 and HDAC2 are paralogs that induce cell death only when knocked out (or inhibited) together¹⁸. Therefore, selective inhibition of either of the paralogs HDAC1 or HDAC2 could provide therapeutic benefit while minimizing toxicity, as some disease lineages have shown enriched dependency on a single paralog^10,19,. Further, recent compounds, such as TNG260, could inhibit HDACs in certain chromatin-modifying complexes (e.g., CoREST), which may overcome some toxicity by enabling an additional level of specificity.

Bromodomain and Extra-terminal Domain Protein Inhibitors

The BET protein family consists of members BRD2, BRD3, BRD4, and BRDT, although BRD4 is most investigated as a cancer target. BRD4 is a common essential gene, suggesting that pan-BET inhibitors would result in significant off-tumor toxicity²⁰. BRD4 binds acetylated lysine residues, recruits the transcriptional elongation complex P-TEFb, and enables contact with the Mediator complex to facilitate transcription through RNA polymerase II²¹. Oncoproteins, such as c-Myc or N-Myc, have been shown to co-localize with BRD4 to drive cancer cell proliferation. Inhibition of BET proteins via the blocking of their bromodomains blocks the expression of MYC²² or MYCN²³ mRNA itself, and collapses oncogenic transcriptional networks, suggesting a therapeutic window for BET inhibition. BET inhibition also downregulates transcriptional networks associated with other oncogenic transcription factors such as EWS-FLI1²⁴, BRD3/4-NUT²⁵, and MLL9-AF9²⁶, suggesting broad utility for cancers dependent on transcription factor-driven proliferation.

The evaluation of BET inhibitors has demonstrated a requirement of BRD4 to maintain oncogenic transcription, and activity is observed in numerous preclinical models of cancer^25,27. Several highly potent and selective BET inhibitors have been developed for study in clinical trials (Figure 1 and Table 1). BET inhibitors have a clear genetic indication for their use in BRD3/4-NUT fusions in midline NUT carcinomas, and clinical trials have demonstrated clinical benefit for a subset of these patients^28,29. Responses, however, have generally been transient, and BET inhibitors have faced the challenge of dose-limiting hematological toxicity, particularly thrombocytopenia, limiting their activity as single agents in most other malignancies^30,31. Therefore, caution is required in the choice of combination therapies with BET inhibitors, particularly those with overlapping hematologic toxicities.

Targeting Cancer-Selective Dependencies and Synthetic Lethalities Involving Chromatin Regulation

The gold standard for therapeutic targets is cancer-selective or synthetic lethal dependencies for which reliance can be predicated based on cancer context. Somatic mutations resulting in gain-of-function (GOF) of epigenetic modifiers are among the most enriched groups of mutated proteins, and specific mutations have been well covered in other reviews³². The high frequency makes these activating mutations attractive therapeutic targets, and indeed there is an FDA approval in this case (EZH2 inhibitors in EZH2 mutant follicular lymphoma)³³. Synthetic lethal-like dependencies provide another target class. The loss of a protein’s function, either through loss-of-function (LOF) mutations or epigenetic silencing, can lead to reliance on the function of another protein, often a protein paralog or a downstream signaling mechanism, to maintain viability. For example, epigenetic targets, such as P300, are of interest in cancers harboring mutations in its histone acetyltransferase paralog CREBBP³⁴. Further, recent data suggest DNMT3A mutations or TET2 mutations in AML cells predispose to response to 5-azacytidine^35,36.

Chromatin complexes often maintain epigenetic control by balancing two opposing activities. The inactivation of one epigenetic complex can create a reliance on the unopposed activity of another chromatin complex network. In the sections below, we will discuss evidence of such “epigenetic antagonism” where LOF of one epigenetic protein leads to dependence on an alternative epigenetic complex (Figure 2A). For example, EZH2 inhibitors are approved for such an instance (e.g., SMARCB1-deficient epithelioid sarcomas)³⁷. In sum, these cancer-selective and synthetic lethal relationships exploit a vulnerability that is unique to the cancer and might mitigate off-tumor toxicities.

Figure 2: Therapeutic Targeting of the Epigenome.

A) Wild-type (WT) BAF complex evicts PRC1/PRC2 from chromatin to establish open chromatin regions, while loss of BAF activity (e.g., through loss-of-function (LOF) mutations) results in unopposed PRC2 signaling. B) Epigenetic proteins participate in oncogenic signaling by associating with oncogenic transcription factors.

Polycomb Repressive Complex Inhibitors

With recent advances in our understanding of the connection between genetic drivers of cancer and epigenetic regulation, the Polycomb Repressive Complex 2 (PRC2) has emerged as an important therapeutic target. PRC2, comprised of the core protein subunits EED, SUZ12, and EZH2, is responsible for the mono-, di-, and tri-methylation of histone 3 lysine 27 (H3K27)^38,39. Polycomb-dependent gene repression is a critical determinant of lineage commitment, differentiation, and cell identity.

The functional role of PRC2 in oncogenic signaling is complex, with both oncogenic and tumor suppressive activity observed depending on the cell of origin. GOF mutations in the EZH2 subunit are observed in DLBCL, follicular lymphomas, and in rare subsets of hepatosplenic T-cell lymphomas, melanomas, and Ewing sarcomas^40–44. EZH2 overexpression also occurs in numerous cancers and is associated with poorer prognosis^45,46. As such, EZH2 knockout has been shown to induce differentiation and reduce cancer cell viability in many pre-clinical models^45,47–49. PRC2 is opposed by trithorax protein complexes, including the chromatin remodeling BRG1/BRM-associated factor (BAF) complex. The BAF complex quickly evicts PRC2 that is reversible with loss of BAF activity (Figure 2A)⁵⁰. LOF mutations in certain members of the BAF complex, including SMARCB1 and ARID1A mutations lead to unopposed PRC2 activity that results in dependence on PRC2^49–52. These are all examples where inhibition of EZH2 by small molecules are shown to be beneficial. In contrast, deletions or LOF mutations in EED, EZH2, or SUZ12 have been observed in approximately 25% of T-cell acute lymphoblastic leukemias⁵³, almost half of early T-cell precursor acute lymphoblastic leukemias⁵⁴, in some myeloid malignancies^55,56, and in the majority of malignant peripheral nerve sheath tumors⁵⁷. Lysine-to-methionine mutations in histone 3 (H3K27M) in pediatric diffuse midline gliomas also alter EZH2 activity⁵⁸. The distinctive roles of PRC2 in different tumor lineages suggest tissue-specific functions of PRC2.

Small molecule discovery programs to identify inhibitors of PRC2 have been extensive. The SET domain of EZH2 utilizes the cofactor S-adenosyl-methionine (SAM) to facilitate H3K27 methylation and EZH2 inhibitors compete with SAM to inhibit PRC2 activity in both EZH2 wild-type and mutant cell lines⁴⁷. The EZH2 inhibitor tazemetostat has been approved in the United States for EZH2 mutant follicular lymphoma and epithelioid sarcomas with LOF SMARCB1 mutations based on positive clinical studies^33,37. While LOF EZH2 mutations are observed in T cell lymphoblastic leukemias and myeloid malignancies,^53,59 there has been only one case reported of a secondary T-cell malignancy during treatment with an EZH2 inhibitor. Nevertheless, as these inhibitors become more commonly used it will be important to report any identified cases of secondary malignancy. Other EZH2 inhibitors are also being evaluated in clinical studies (Figure 1 and Table 1). Interestingly, valemetostat inhibits both EZH1 and EZH2 and is suggested to induce greater antitumor effects than EZH2 alone⁶⁰. In addition to EZH2 inhibitors, first-in-class EED inhibitors bind and block the H3K27me3 recognition pocket of EED to disrupt protein-protein interactions between PRC2 subunits, including both EHZ1 and EZH2 (Figure 1 and Table 1). MAK683 is currently being evaluated in DLBCL and nasopharyngeal carcinoma, with early reports suggesting signs of activity in DLBCL⁶¹. ORIC-944 is currently being studied in metastatic prostate cancer, and APG-5918 is being evaluated across numerous advanced solid tumors, including SMARCB1-mutant sarcomas and EZH2-mutant B-cell lymphomas.

BAF Complex Inhibitors

The ATP-dependent chromatin remodeling BAF complex (SWI/SNF complex in yeast) is crucial for the regulation of gene expression and differentiation and antagonizes the repressive PRC2 complex.⁵⁰ The BAF family consists of three large, multimeric protein complexes: canonical BAF (cBAF), polybromo-associated BAF (pBAF), and non-canonical BAF (ncBAF)⁶². Each complex includes one of two DNA-stimulated ATPase catalytic subunits (SMARCA4 or SMARCA2), which are responsible for nucleosome displacement and transcriptional activation. A deficiency in either SMARCA2 or SMARCA4 results in a synthetic lethal paralog related dependency on the other subunit^63,64. Mutations in BAF are abundant in cancer; many subunits of the BAF complex can be altered, including BRD9, ARID1A/B, PBRM1, ARID2, SMARCC1, SMARCC2, SS18, and SMARCB1. Mutations and alterations in BAF complex family members result in unique residual complex compositions that alter chromatin accessibility and dynamics, which likely underlies their oncogenic potential⁶⁵. Alternatively, these alterations can incur synthetic lethal-like dependencies in various cancers.

A small molecule allosteric inhibitor of the SMARCA2/4 ATPases, FHD-286 (patent US 2023/0138480), is currently in clinical evaluation for AML and myelodysplastic syndrome (NCT04891757) and metastatic uveal melanoma (NCT04879017) (Figure 1 and Table 1). The clinical trial evaluating patients with AML was halted after a patient receiving FHD-286 showed significant toxicity from differentiation syndrome, likely due to on-target inhibition of SMARCA2/4, suggesting SMARCA2/4 may hold promise as a therapeutic target in AML. However, because the inhibitor targets both SMARCA2 and SMARCA4 paralogs, it raises questions as to whether prolonged inhibition will be cytotoxic to all cells, including healthy tissue. Ideally, an inhibitor would selectively target either SMARCA2 or SMARCA4 to best leverage the synthetic lethal relationship, such compounds are now in active development (Figure 1 and Table 1)^66,67. Further, BAF subunits, including SMARCA2 and SMARCA4, are mutated in as many as 20% of cancers and are considered tumor suppressors in many contexts; thus more work is warranted with regards to evaluating the risk of secondary malignancy with SMARCA2/4 inhibitors⁶⁸.

While targeting the obligate ATPase activity of the BAF complex is predicted to be lethal to all cells, targeting selective subunits that are dependencies in certain contexts seems promising. Recent work has demonstrated that perturbation of the cBAF complex can lead to dependency on the ncBAF complex and more specifically on the ncBAF complex subunit, BRD9.^69,70 In this case, biallelic inactivation of SMARCB1, a subunit of the cBAF complex, portends a dependency on BRD9 in malignant rhabdoid tumors^69,70. In an alternate mechanism, synovial sarcoma is defined by a rearrangement that creates the SS18-SSX fusion protein, which displaces and degrades SMARCB1, mimicking LOF SMARCB1⁷¹. In both cases, SMARCB1 LOF is compensated for by the ncBAF complex, thereby resulting in a synthetic lethality on the ncBAF complex subunit BRD9^69,72. The clinical candidate PROTACs CFT8634⁷³ and FHD-609 (patent US 2023/0142883) degrade BRD9 (Figure 1) and are currently being studied in SS18-SSX fusion synovial sarcoma and other tumors with SMARCB1-loss (Table 1). Notably, enrollment was recently paused for FHD-609 due to reported QTc prolongation, a risk factor for cardiac arrythmia, which was observed at higher doses of the drug.

Targeting of Oncoprotein Complexes

The most direct and effective approach to alter oncogenic gene expression programs is to directly target the activity of the mutant/altered transcription factors that drive cancer. However, drugging these proteins remains challenging. One method to circumvent this problem is to target the critical co-factors and chromatin complexes that are necessary for their oncogenic function (Figure 2B). This is the case in a subset of acute leukemias with KMT2A (MLL1) rearrangements. While KMT2A itself is a histone methyltransferase, the fusion oncoprotein excludes the SET domain of KMT2A, activating leukemic gene expression independent of histone methylation⁷⁴, typically due to rearrangements with mediators of transcriptional elongation such as AF4, AF9, ENL, and AF10⁷⁵ To date, several indirect targeting strategies have been developed to block the KMT2A-fusion protein through targeting of its cofactors. These include small molecules targeting Menin⁷⁶ and DOT1L,⁷⁷ which reverse leukemic gene expression and exhibit therapeutic effects in pre-clinical leukemia models. While the concept of targeting leukemic gene expression through inhibition of chromatin complexes is well developed in KMT2A-r AML and ALL, it is becoming increasingly clear that similar mechanisms are at play in other leukemias, including NPM1 mutant and NUP98-rearranged AML^78,79.

DOT1L Inhibitors

The histone methyltransferase DOT1L complexes with numerous proteins and is responsible for the mono-, di-, and trimethylation of H3K79^80,81. DOT1L-dependent H3K79 methylation is thought to be required for transcriptional elongation at active gene loci⁸⁰. DOT1L is a non-SET domain enzyme that utilizes SAM as a cofactor to perform its catalytic function⁸². A subset of acute leukemias harboring KMT2A-r or NPM1 mutations is defined by the overexpression of transcription factors in the HOXA/B cluster and their co-factor MEIS1. These transcription factors reprogram the epigenome to promote leukemogenesis⁸³. The most prominent known role for DOT1L is in KMT2A-r leukemias where aberrant H3K79me2 at KMT2A-r-bound sites is a central feature. The small molecule pinometostat potently inhibits DOT1L activity through occupation of the SAM binding site, and induced differentiation of KMT2A-r leukemia and reduced KMT2A-r leukemia in preclinical models⁸⁴. Despite the multiple pre-clinical DOT1L inhibitors studied, only pinometostat has been tested in the clinic (Figure 1 and Table 1). An early phase clinical trial demonstrated proof-of-concept clinical remissions as a single agent and general tolerability⁸⁵. However, response rates have been modest and the continuous 28-day infusion of pinometostat is burdensome. Thus, if DOT1L inhibition is to be pursued as a pharmacologic strategy moving forward, pharmacologic inhibitors with more robust pharmacokinetic properties will be required.

Menin Inhibitors

Menin, initially reported as a scaffolding protein and recently implicated as a reader of H3K79me2⁸⁶, is encoded by the Multiple Endocrine Neoplasia 1 (MEN1) gene and is ubiquitously expressed across tissues. Menin binds KMT2A/MLL1 via its N-terminus Menin binding domain and is an essential subunit of the MLL1/2 COMPASS-like complex, a H3K4 histone methyltransferase. The MLL1/2 COMPASS-like complex is required for the maintenance of MEIS1 and the HOX gene expression in AML, and inhibition of Menin significantly reduced tumor burden and induced differentiation in pre-clinical models^79,87. Numerous studies have shown that subsets of genetically defined acute leukemias, namely those with KMT2A-rearrangements, NPM1 mutations or NUP98-rearrangements, strongly depend on Menin to promote MEIS1/HOXA transcription^78,79,87.

The strong support for Menin inhibitors in hematological malignancies has rapidly expanded the number of clinical studies evaluating their safety and efficacy (Table 1). Numerous small molecule inhibitors, which bind the Menin binding pocket of KMT2A, thereby blocking the physical interaction between Menin and KMT2A have been developed (Figure 1A)^88,89. In acute leukemias, this disruption evicts Menin/KMT2A from chromatin and typically downregulates transcription of HOXA genes and MEIS1 in acute leukemias that are dependent on Menin in both the KMT2A-r and KMT2A wild-type settings^79,88. Early clinical data for revumenib showed 59% overall response rates in KMT2A-r leukemias and 36% in NPM1 mutant leukemias⁹⁰. Early clinical reports for ziftomenib showed an overall response rate of 42% and 75% in patients that exhibit differentiation syndrome⁹¹. Menin inhibitors are orally bioavailable and reported as generally well tolerated. However, serious adverse events have been identified, including asymptomatic prolongation of the QTc interval as a dose-limiting toxicity specifically with revumenib⁹⁰. Differentiation syndrome, managed by corticosteroids and hydroxyurea, was also reported for both revumenib and ziftomenib, and is consistent with on-target activity of agents that induce AML differentiation⁹⁰. Menin inhibition has been shown to have activity in some pre-clinical solid tumor contexts^92–94, but clinical trials in these disease contexts have not yet been reported.

Recent evidence from the AUGMENT-101 (NCT04065399) study in patients treated with revumenib suggests that resistance to Menin inhibition occurs after approximately two cycles of treatment⁹⁵. Patients harboring MEN1 mutations at residues M327, G331, T349, and S160 demonstrated reduced drug affinity across multiple inhibitors in vitro without affecting the Menin-KMT2A interaction. Interestingly, all these mutations observed in patients were predicted both using unbiased base-editing CRISPR-Cas9 based screening of the MEN1 gene in vitro and in PDX samples treated in vivo. This is the first epigenetic therapy to our knowledge whereby clinical resistance in patients has been demonstrated to be mediated through somatic mutations in the drug target. These data provide gold standard evidence for the on-target mechanism of drug action and evidence for strong dependency of these disease subsets on Menin. As with resistance to tyrosine kinase inhibitors, overcoming resistance to first-generation Menin inhibitors by increasing affinity for the altered binding pocket with second-generation drugs may be possible.

Targeting Non-Essential Genes Involved in Transcription

The epigenetic cancer dependencies described above have fueled the development of highly selective and potent inhibitors targeting epigenetic complexes. Other epigenetic targets, such as the histone acetylase P300 and histone demethylase LSD1, demonstrate indiscriminate efficacy across numerous cancer lineages, although do not quite cross the threshold of being considered pan-essential, and thus may demonstrate fewer dose-limiting toxicities than cytotoxic chemotherapies.

Histone Acetyltransferase Inhibitors

The acetylation of histones is a dynamic process that is tightly regulated by the activity of histone acetyltransferases (HATs) and HDACs. Acetylation of lysines, including H3K9 and H3K18 at gene promoters and H3K27 at gene enhancers and promoters, is associated with gene transcription. Enzymatic HAT subunits are part of larger multimeric complexes, and the other members of the complex can impact HAT specificity and localization⁹⁶. There are three main HAT complex groups that regulate the acetylation of core histones: MYST (KAT5, KAT6A/B, KAT7, KAT8), GCN5/PCAF (KAT2A/KAT2B), and CBP/EP300 (CREBBP/EP300). While HATs were originally discovered for their histone acetyltransferase activity, this has become a misnomer as a growing number of non-histone targets of these acetyltransferases have been discovered⁹⁷.

Targeting HATs is a relatively new strategy compared to HDACs, but nonetheless there is growing evidence these targets could have a role in cancer therapy. The CBP/P300 family acetylates H3K18 and H3K27 and has also received interest in cancers dependent on super-enhancer driven transcriptional programs. CBP/P300 function as coactivators of transcription factors, such as androgen receptor in prostate cancer⁹⁸. EP300 knockout alone has broad activity across numerous cancer lineages, but is not considered pan-essential, perhaps due to a partial compensation from CBP activity as CREBBP mutations are associated with stronger EP300 dependency³⁴. Confirming the synthetic lethal relationship, CREBBP dependency is pronounced in DLBCL harboring damaging mutations in EP300^34,99. However, while both proteins acetylate histones, they likely also have some non-redundant roles, as genetic knockout of EP300 is notably less tolerated than knockout of CREBBP¹⁰⁰. There are two pharmacologic classes of CBP/P300 inhibitors, enzymatic HAT inhibitors and small molecule bromodomain inhibitors that block CBP/P300 activity^101,102. Head-to-head comparisons of HAT versus bromodomain inhibitors have yet to be performed. It is likely that off-target effects specific to each class of compounds could enable greater activity in specific subsets of disease. For example, CBP/P300 bromodomain inhibitors may inhibit other bromodomain-containing proteins (e.g., BRD4) and could be more effective in BET-inhibitor response contexts such as midline NUT carcinoma¹⁰³. Despite showing pre-clinical efficacy, HAT domain inhibitors have not yet been successfully bridged to clinical trials. In contrast, two CBP/P300 bromodomain inhibitors are being tested in clinical trials. CCS1477 (inobrodib) is currently in Phase I clinical trials for both solid and hematologic malignancies and FT-7051 is being studied for metastatic castration-resistant prostate cancer (Figure 1 and Table 1).

P300-selective, CBP-selective, and pan-targeting PROTACs, resulting in proteolysis of the target protein, have also been developed^100,104,105. As compared to the catalytic inhibitor, a P300-selective PROTAC demonstrated rapid apoptotic cell death rather than cell cycle arrest, suggesting that P300 has catalytic-independent functions that can only be targeted through degradation¹⁰⁰. Thus, deploying new pharmacologic modalities (i.e., degradation) may provide unique functions as compared to their catalytic inhibitor counterparts and would more closely resemble phenotypes observed with CRISPR knockouts (e.g., DepMap). However, PROTAC structures are much larger than typical small molecule catalytic inhibitors and thus may suffer from challenges related to solubility, permeability, and pharmacokinetics.

While the therapeutic utility of CBP/P300 is effective across numerous cancers due to the requirement for their activity to maintain cell viability, other HATs, such as the KAT6A/KAT6B enzymes in the MYST family, have an enriched dependency based on cancer lineage and/or genetics. Recurrent translocations involving KAT6A and KAT6B occur in AML,¹⁰⁶ and KAT6A has been shown to be a critical driver of the stem cell-like properties in AML¹⁰⁷. Amplifications of KAT6A and/or KAT6B also occur in solid tumors, including breast, colon, lung, endometrial, and ovarian cancers and medulloblastoma^108–111. KAT6A/KAT6B are required for suppression of senescence through locus-specific acetylation of H3K9 and H3K23, although the mechanism by which this occurs and whether the KAT6A/B inhibition phenotype is mediated through histone acetylation is still unclear^112,113. Selective catalytic inhibitors targeting KAT6A/B have now been reported^112,114, and the KAT6A/B inhibitor PF-07248144 is being investigated in ER+ breast cancer (Figure 1 and Table 1).

The remaining class of HATs include KAT2A/KAT2B and are commonly associated with acetylation of H3K9 and H3K14^115,116. While inhibitors targeting KAT2A/KAT2B have not yet entered clinical trials, there is pre-clinical evidence that KAT2A HAT activity is a key activator of MYC transcription in some MYC-driven cancers^117,118 and we speculate it will likely be an activator in some MYCN-driven cancers. In addition to catalytic and bromodomain inhibitors, more recently, drug discovery efforts against KAT2A/B HATs have expanded to include PROTACs, which have now been investigated pre-clinically¹¹⁹.

Histone Demethylase Inhibitors

Histone demethylases can broadly be divided into two families: flavin adenine dinucleotide (FAD)-dependent (KDM1) and 2-OG-dependent JmjC domain-containing histone demethylases¹²⁰. Of the 20+ histone demethylases, only chemistry efforts targeting the FAD-dependent family member LSD1/KDM1A have successfully transitioned into clinical trials. LSD1 complexes with the NuRD and CoREST complexes to demethylate H3K4me1/me2^121,122 and with nuclear hormone receptors to demethylate H3K9me1/me2¹²³, resulting in inactivation or activation of transcription, respectively. Inhibition of LSD1 is efficacious in pre-clinical models of AML and small cell lung cancer (SCLC),^124–126 and DepMap predicts the greatest dependency on LSD1 in AML across the over 1000 cancer cell line models screened⁹. A subset of SCLCs that are defined by neuroendocrine-like lineage and highly express ASCL1 are strongly responsive to LSD1 inhibitors by decreasing neuroendocrine expression patterns and increasing NOTCH signaling in preclinical studies¹²⁵. Similarly, LSD1 promotes stemness features and lineage plasticity in prostate cancer¹²⁷, suggesting that histone demethylase inhibitors may have a broad role in preventing mechanisms of neuroendocrine plasticity across numerous cancer types.

Pre-clinical success of LSD1 inhibitors in SCLC and AML advanced the testing of potent LSD1 inhibitors in clinical studies (Table 1). Encouraging data was seen in relapsed/refractory AML with Iadademstat¹²⁸. Clinical trials evaluating Iadademstat in tumors with neuroendocrine plasticity, such as prostate cancer and SCLC, are set to begin shortly. Bomedemstat is being evaluated for efficacy in combination with anti-PD1 immunotherapy as upfront therapy in SCLC and in combination with venetoclax in AML. Not all clinically evaluated LSD1 inhibitors have shown promising signs of efficacy, however. Clinical trials evaluating GSK2879552 were terminated due to significant adverse events, encephalopathy in the SCLC trial, and limited efficacy in SCLC and AML¹²⁹. Encephalopathy has not been reported with any other LSD1 inhibitors so it may represent a compound-specific, off-target effect.

Emerging Pre-clinical Epigenetic Targets

While this review provides detailed pharmacologic and clinical data supporting the use of epigenetic inhibitors being tested in the clinic, there are several epigenetic targets emerging as promising candidates in pre-clinical studies.

The Histone Ubiquitylase Complex PRC1

The Polycomb Repressive Complex 1 (PRC1) complex, which is responsible for the ubiquitylation of lysine 119 on histone 2A (H2AK119Ub), commonly co-localizes with the PRC2 complex to facilitate gene repression¹³⁰. PRC1 contains one of two E3 ubiquitin ligase RING1 paralogues, RING1A or RING1B, and is subdivided into either canonical or non-canonical complexes depending on whether the PRC1 complex contains a chromobox (CBX) protein or one of two paralogs, RYBP or YAF2, respectively (Figure 3). RING1A/B interacts with one or more PCGF proteins to enhance its ubiquitin ligase activity^131,132. Canonical PRC1 commonly complexes with PCGF2/4, while each of four non-canonical complexes PRC1.1, PRC1.3/PRC1.5, or PRC1.6, are defined by containing PCGF1, PCGF3/5, or PCGF6, respectively¹³³.

Figure 3: New Epigenetic Targets in Cancer.

Schematic showing select complexes (PRC1, ENL within the super elongation complex, and ASH1L) that are of interest for the development of new therapeutics.

PRC1 activity has been increasingly recognized as important in cancer¹³², although the contribution of individual PRC1 complexes is not well understood. Genomic alterations in PRC1 are relatively rare, except for alterations in BCOR and its paralog BCORL1, subunits of the PRC1.1 complex. Loss-of-function BCOR/BCORL1 mutations are observed in myeloid malignancies¹³⁴, T-cell lymphomas¹³⁵, and cancers of the central nervous system¹³⁶. BCOR genomic rearrangements are particularly enriched in pediatric solid tumors. Most clear cell sarcomas of the kidney and primitive mesenchymal myxoid tumors of infancy harbor BCOR with internal tandem duplications¹³⁶. BCOR-CCNB3, BCOR-MAML3, ZC3H7B-BCOR, BCOR-RARE, BCOR-CREBBP and EP300-BCOR fusions are also observed in various undifferentiated Ewing-like sarcomas, leukemias, and malignancies of the central nervous system^136–139. Notably these fusions are very rare and not well documented, so the function of BCOR in these disease states is unknown. Most fusions retain BCL6 and PCGF1 binding domains, suggesting that these domains are likely still important for their function; however careful dissection of these fusions is needed to further understand their mechanism.

Due the complexity of the PRC1 family, selective and potent inhibitors of canonical and variant PRC1 are needed to discern their individual contributions in cancer. Inhibitors of RING1 show response rates in leukemia models in vitro¹⁴⁰. Currently, there is no strong clinical rationale for targeting RING1 in leukemia, although it has been suggested that RING1 regulates self-renewal potential of both normal and leukemia cells¹⁴¹. Therefore, it will be important to study whether the combined loss of RING1A and RING1B is tolerated in healthy tissues due to loss of compensatory paralog function. RING1A/B integrate into every PRC1 variant complex, and therefore RING1 inhibitors cannot differentiate functions of the different complexes. USP7 inhibitors, largely reported to be active in TP53 wildtype malignancies by destabilizing MDM2, have also been described as destabilizing the PRC1.1 complex through disrupting PCGF1/RING1 interactions¹⁴². Numerous inhibitors targeting CBX proteins, including CBX2¹⁴³, CBX4/7¹⁴⁴, CBX6¹⁴⁵, and CBX8^146,147 have also been identified. DepMap does not show strong single gene dependency on any of the CBX proteins, perhaps due to redundancies between CBX paralogs. Multiplexed CRISPR-Cas9 studies evaluating the combinatorial effect of double knockout of CBX proteins would of interest to evaluate the utility of non-selective CBX inhibitors in diverse cancer lineages.

Chromatin reader ENL

ENL/MLLT1, a histone lysine acetylation reader, promotes transcription through interactions with the super elongation complex and regulation of chromatin remodeling and gene expression via binding to modified histones (Figure 3). Two groups simultaneously identified a role for the ENL_YEATS domain in the control of leukemogenic gene expression in KMT2A-r AML^148,149. Chromosomal translocations of KMT2A-ENL are an initiator of leukemogenesis and the fusion protein has been shown to be a viable therapeutic target¹⁵⁰. Likewise, recurrent hotspot mutations in the YEATS domain of ENL have been described in Wilms tumor, a pediatric kidney cancer, and have been demonstrated to drive hyperactivation of transcription, supporting oncogenesis^151,152. Since being identified as a potential therapeutic target in leukemia, multiple peptidomimetics and small molecule inhibitors of ENL, which displace ENL from chromatin by blocking its YEATS domain interaction with acetylated histones, have been developed^153–160. ENL degraders have also been developed^155,161. Interestingly, both disruption of the interaction between the ENL YEATS domain and acetylated histones and depletion of the protein significantly suppress leukemia progression with minimal and reversable effects on normal hematopoiesis^157,158. It is important to note that these drugs target the YEATS domain more generally, inhibiting both ENL and its paralog AF9, which is required for normal hematopoiesis¹⁶². Interestingly, AF9 is not a dependency in KMT2A-r leukemias and most KMT2A-AF9 proteins lack the AF9 YEATS domain^148–150. Thus, ENL inhibitors likely are not effective by selective inhibition of AF9 or the KMT2A-AF9 fusion, but rather wild-type ENL may be operating through a Menin-like mechanism whereby ENL inhibitors impair wild-type ENL function and subsequent transcriptional elongation. A therapeutic window might be enhanced if the ENL-containing oncogenic signaling complex that specifically regulates KMT2A-r-dependent genes is distinct from the endogenous ENL-containing complex, similar to recent observations for DOT1L¹⁶³. Although preclinical data and probe molecules developed to date support the translation of ENL inhibitors for the treatment of KMT2A-r leukemia, DepMap data does not show as strong of a dependency as for Menin. Thus, these compounds may not translate to as strong of a clinical response. It will be interesting to test ENL inhibitors in combination with Menin inhibitors to evaluate whether they exhibit synergism.

Histone Methyltransferase ASH1L

Another emerging target is ASH1L, a histone methyltransferase that positively regulates gene expression via methylation of H3K36 and recruitment of KMT2A to chromatin (Figure 3). Based on these molecular functions, ASH1L was found to be required for the initiation and maintenance of KMT2A-r leukemia^164,165. ASH1L also has been implicated in the pathogenesis of other AMLs with high HOX gene expression, and it is overexpressed in thyroid, breast, and liver cancers^166–168, making ASH1L an attractive target for drug discovery. The first-in-class small molecule inhibitor of ASH1L, AS-99, has been reported to induce cell death and differentiation, suppress KMT2A fusion driven gene expression, and reduce leukemia burden in vivo¹⁶⁹. Like ENL, ASH1L does not show as strong of a dependency in DepMap for hematologic malignancies as compared to Menin. However, it does show a strong enrichment for some solid malignancies including ovarian and endometrial cancers, suggesting it might be translated in other diseases. It will be exciting to watch this class of drugs develop into clinical molecules for the treatment of cancer; however, caution should be taken as LOF mutations and haploinsufficiency of ASH1L are linked to many developmental disorders, including autism¹⁷⁰.

Therapeutic considerations

Non-genetic Reprogramming and Plasticity

While genetic drivers are critical to both oncogenesis and drug resistance, non-genetic mechanisms are being increasingly recognized as drivers of tumor relapse^171–174. Much like genetic heterogeneity, cancers often are composed of heterogenous populations of cells that exist in unique transcriptional states. Although it is controversial whether cancer cell states that are associated with therapy resistance pre-exist or transition when exposed to therapy, it is established that tumors can quickly and reversibly alter their cell state by epigenetic and transcriptional reprogramming (Figure 4A)^175,176. Tumor cells that persist in the presence of therapeutic pressure can arise from cells derived from different transcriptional states^177,178. This intrinsic ability of cancer cells to present a different lineage state can promote metastasis and chemoresistance^175,179. The ability of tumor cells to quickly transit between epigenetic states may be partially explained by the presence of bivalent promoters, that is the paradoxical presence of both active H3K4me3 and repressive H3K27me3 marks at promoters that are transcriptionally repressed but poised for transcriptional activation¹⁸⁰. Treatment-refractory breast cancer cells have been shown to utilize bivalent promoters to withstand therapeutic stress, and bivalent promoters are enriched in relapsed cancers^181–183. Similarly, enhancer rewiring has been demonstrated in preclinical studies to render resistance to chromatin modifying compounds such as BET inhibitors¹⁸⁴. While signaling cues that promote epigenetic plasticity are not well known, it is likely that tumor cell extrinsic and intrinsic factors can promote epigenetic reprogramming¹⁸⁴. For example, extrinsic factors in the tumor microenvironment promote residual cell survival and clonal expansion^185,186. It will be important to use advanced lineage tracing, single-cell sequencing platforms, and cell barcoding technologies to increase clarity as to how cells pass through cell states to adapt to cell intrinsic and extrinsic pressures.

Figure 4: Epigenetic/Lineage Plasticity as a mechanism of therapeutic resistance.

A) Drug-tolerant tumor cells harbor the intrinsic ability to reversibly alter their gene expression programs to overcome drug treatments. B) Cancer cells can co-opt lineage-associated differentiation states to overcome therapeutic or environmental stressors. C) Inhibition of an epigenetic protein can select for the upregulated activity of an alternative epigenetic pathway.

Despite inter-tumoral variability within cancers arising from a specific tissue, cancers of the same lineage mostly cluster together in expression profiles, likely by retaining transcriptional features of their tissue of origin^187,188. However, a feature of some cancers is the intrinsic ability to leverage developmentally linked cell states to promote tumor growth, metastasis, and evasion of therapeutic interventions (Figure 4B)¹⁸⁹. It has long been observed that epithelial cancers can undergo an epithelial-to-mesenchymal transition, a state more resistant to cytotoxic, targeted, and immunotherapies^190,191. A similar adrenergic-to-mesenchymal transition has been observed in neuroblastoma that can also mediate resistance to cytotoxic and immune-therapies^192–195. Lineage plasticity is also observed in the neuroendocrine trans-differentiation of a subset of prostate cancers¹⁹⁶, Ewing sarcoma¹⁹⁷, conversion of non-small cell lung cancers to small cell lung cancers^198,199, and phenotype switching between “proliferative” and “invasive” phenotypes in melanoma²⁰⁰. Similarly, pediatric ALLs have been shown to “switch” to a myeloid lineage when treated with CD19 CAR T cells²⁰¹. Some interventions have been shown to alter the differentiation trajectories of cancers into therapy-sensitive states, such as methotrexate in melanoma²⁰², EZH2 inhibitors in prostate cancer^194,203, and KDM5A/B inhibitors in lung and breast cancer^172,204,205. Our ability to target drug-tolerant phenotypic states could be paramount to the eradication of residual tumor cells.

Activation of Alternate Epigenetic Pathways Following Treatment with Epigenetic Inhibitors

LOF mutations that occur naturally in epigenetic complexes can result in changes in the activity of alternate epigenetic pathways due to compensation or a release of regulation. It is likely that treatment with epigenetic inhibitors will produce similar changes in the activity of epigenetic pathways. For example, a recent study suggests that malignant rhabdoid tumors overcome EZH2 inhibition via loss of activity of NSD1, a H3K36me3 methyltransferase²⁰⁶. Knockout of the H3K36me3 demethylase, KDM2B, restored EZH2 sensitivity. In another study, it was found that Menin inhibition releases the KMT2A complex from active genes, inducing a re-localization of the KMT2A complex to bivalent chromatin (Figure 4C)²⁰⁷. The combination of EZH2 and Menin inhibitors reduced cell viability more than either inhibitor alone²⁰⁷. Thus, combinatorial epigenetic therapies that prevent changes in compensatory epigenetic pathways may be more efficacious in clinical trials than single agents. We must also be aware that epigenetic inhibition may in some contexts induce cell states that are more aggressive or that promote secondary malignancies^208,209. Further, LOF in alternate pathways may also mediate resistance. For example, NCOR1 and HDAC3 knockout was found to confer resistance to P300/CBP inhibitors; thus, altered histone acetylation dynamics may affect the response to clinical P300/CBP inhibitors²¹⁰. In sum, these data provide a strong rationale for a better understanding of how cancer cells rewire epigenetic pathways following epigenetic inhibition and suggests that combinatorial epigenetic inhibition may be necessary to achieve durable responses.

Utilizing Epigenetic Inhibitors as Therapy “Boosters”

While sustained therapeutic doses of some single agent epigenetic therapies may be difficult to achieve in humans, emerging evidence suggests that low-dose or short interval delivery of epigenetic therapies may be useful as an adjunctive therapy to enhance the activity of chemotherapy, radiotherapy, and immunotherapy. For example, tolerated doses of vorinostat in combination with the radiotherapeutic 131-metaiodobenzylguanidine (MIBG) in patients with neuroblastoma were associated with increased response rates as compared to either therapy alone²¹¹. The biology of this synergism remains unknown. It is possible that vorinostat prevents the downregulation of the transporter that is required for MIBG uptake, promotes further DNA damage, or alters the cell state to one that is more amenable to radiation-related killing. EZH2 inhibitors have also been found to sensitize lung cancer to standard chemotherapy²¹² or promote increased expression of tumor cell antigens such as GD2^194,213,214 or MHC expression^207,215.

Epigenetics and phase transition

Emerging data suggest that the physicochemical properties generated by organization of proteins into membrane-less compartments also affect the structure and function of chromatin^216–218. Chromatin modifiers, transcription factors²¹⁹, and histones²²⁰ are highly enriched for intrinsically disordered regions that can promote liquid-liquid phase separation. We are just beginning to understand how condensate biology affects partitioning of small molecules, such as epigenetic inhibitors, into specific epigenetic compartments and how these biophysical properties regulate epigenetic mechanisms^221,222. For example, the BET inhibitor JQ1 preferentially concentrates in BRD4 condensates, suggesting localization within the cell may be important for pharmacodynamic response²²². Further, fusion oncoproteins can harbor intrinsically disordered regions (IDRs) and are prone to generate phase-separated compartments that can modulate transcriptional activity and promote leukemogenesis²²³. A recent study experimentally validated that 58% of fusion oncoproteins, particularly those localized to the nucleus, can form condensates²²⁴. Using machine learning, the authors predict that as many as 67% of fusion oncoproteins may be able to form condensates. It has been shown, for example, that NUP98 fusions form condensates that drive leukemogenic transcriptional programs²²³. Further, driving concentrations of fusion oncoproteins through loss of regulators also disrupted oncogenic signaling activity by altering condensate properties^225,226. This suggests a “Goldilocks” principle by which fusion oncoprotein levels must be fine-tuned and either low or high expression disrupts condensate formation and oncogenic activity. Thus, potential avenues of targeting “undruggable” fusion oncoproteins may be through targeting IDRs to modulate condensate properties.

Summary and Future Directions

New insights into the biology of cancer epigenetics have created a strong foundation for the development of small molecule inhibitors targeting at least thirteen unique classes of epigenetic targets (Figure 1). Excitingly, advances in medicinal chemistry have expanded the toolkit to target previously non-targetable proteins. The slate of new small molecule inhibitors includes molecular protein degraders (e.g., BRD9), allosteric inhibitors of catalytic function (e.g., EED, SMARCA2/4) and disruptors of protein-protein interactions (e.g., Menin-KMT2A). Selective degradation of protein targets via PROTACs may be particularly useful, as such degradation may enable distinction between catalytic-dependent versus scaffolding functions of epigenetic complexes. However, caution is warranted in interpreting the phenotype of degraders in the absence of orthogonal models (e.g., CRISPR knockout). A pre-print suggests that degradation of BRD4 also forms neo-amino-terminal peptides that neutralize Inhibitor of Apoptosis proteins, subsequently inducing cell death²²⁷. This may explain why cell death was observed with P300/CBP degraders and not the HAT or bromodomain CBP/P300 inhibitors¹⁰⁰. The discovery of natural glue-like degraders, such as lenalidomide and related derivatives, that degrade transcription factors such as IKZF1 and IKZF3, offers the promise of degrading other transcription factors, including transcription factor fusions. Beyond targeting tumor-intrinsic epigenetic regulation described in this review, epigenetic inhibitors may also hold promise in promoting response to immunotherapy such as by preventing T cell exhaustion, modulating immune cell populations in the tumor microenvironment, or promoting antigen presentation^228–230.

As described, the development of new classes of compounds, such as ENL, PRC1 (RING1), and ASH1L inhibitors shows promise in AML and could be applied to other disease states. Additionally, a first-in-class NSD2 inhibitor, KTX-1001, will soon go into clinical trial for testing in t(4;14) and NSD2 gain-of-function (E1099K) mutant multiple myeloma (Figure 1 and Table 1). While we used a narrow perspective to define drugs that target epigenetic processes in this review to only those inhibitors that perturb the function of proteins that regulate chromatin/transcription factor function, there are other epigenetic-adjacent inhibitors also currently in clinical trials. These include those targeting the m6A methyltransferase METTL3, the kinases CDK7 and CDK9, PRMT5, and IDH1/2.

Despite much promise for epigenetic therapies in cancer, response rates in humans remain poor for solid tumors. In these patients, this may be due to dose-limiting toxicities that prevent therapeutic dosing (e.g., HDAC inhibitors) or inadequate drug concentrations within the tumor. Thus, optimization of pharmacokinetics and limiting off-tumor toxicity will be paramount to establish durable and tolerated therapy responses in patients. Therapeutic boosting with low-dose epigenetic inhibitors is another implementation strategy for epigenetic inhibitors in the clinic. Further, in vitro studies with solid tumor models suggest that epigenetic heterogeneity and adaptation may contribute to rapid drug resistance in the absence of clear genetic drivers of disease, providing another barrier to effective treatment of solid tumors. The implementation of single-cell spatial transcriptomics and epigenomics will hopefully shed light on whether cellular interactions or location can influence the response to epigenetic inhibitors in solid tumors. As we accumulate more information on encoded epigenetic, genetic, and drug sensitivity properties of cancers, we can integrate these data to design rationally guided strategies to treat and provide curative options for patients.