Epigenetic inhibitors for the precision treatment of ARID1A-mutant ovarian cancers: what are the next steps?

Histone acetylation plays a key role in gene transcription, with acetylation generally associated with elevated gene transcription and deacetylation associated with gene repression. Epigenetic modifiers such as histone acetyltransferases (HATs) and histone deacetylases (HDACs) regulate chromatin accessibility. HDACs are critical repressors of gene transcription by enzymatically removing acetyl groups from both histones and non-histone substrates. In general, HDAC activity promotes chromatin compaction and HAT activity leads to increased nucleosome accessibility. Suberoylanilide hydroxamic acid (SAHA/vorinostat), a pan-HDAC inhibitor, was approved by United States Food and Drug Administration (FDA) in 2006 for cutaneous T-cell lymphoma ( 11 ). Several other HDAC isoform-selective inhibitors and pan-HDAC inhibitors have also been approved or tested in clinical studies. HDAC inhibitor monotherapy has been approved for patients with hematological tumors, while their effect in solid tumors are marginal in randomized trials ( 12 ). More recent studies tend to focus on the combination of HDAC inhibitors with other anticancer therapies. According to Clinicaltrials.gov numerous active clinical trials are evaluating combinatorial approaches with vorinostat in a variety of disease indications such as, melanoma, glioma, leukemia, and lung cancers.

The field of medical oncology has shifted its focus on how to best employ immune-based therapeutics, which is highlighted by the significant number of new clinical trials evaluating immunomodulating agents (i.e. PD-1/PD-L1 checkpoint inhibitors). In the context of chemoresistant EOC, a clinical trial evaluating nivolumab (PD-1 antibody) in 20 patients observed a complete response in one of the two OCCC patients ( 25 ), but whether ARID1A -mutation was a contributing factor to this response is unclear. A recent study found that loss of ARID1A led to inactivation of mismatch mediated DNA repair ( 26 ). ARID1A-induced mismatch repair deficiencies promoted increased sensitivity to PD-L1 targeting agents. In addition, several in vitro studies demonstrate that mutations in ARID1A and/or SWI/SNF complexes lead to aberrant immune response and regulation ( 27 ). Taken together, these studies suggest that a precision medicine approach in ARID1A -mutated cancer by combining HDAC or EZH2 inhibitors with immune-oncology agents could provide a targeted and highly effective therapeutic strategy.

To develop precision medicine approaches based on the patient’s genetic background further research is needed to optimize methodologies to detect and describe pathogenic mutations. Loss or decreased ARID1A expression is observed in the precursor lesion of OCCC, endometriosis, suggesting that ARID1A mutation is an early and potentially initiating event ( 28 ). This highlights a target population for attempting to detect ARID1A -mutations prior to the development of OCCC. Utilizing cell-free DNA (cfDNA) from women with endometriosis to perform targeted sequencing of ARID1A could result in earlier diagnosis and tailored therapeutic approaches; however, the level of detection of these mutational events needs improvement. Alternatively, identifying a secreted protein or metabolite that correlates to ARID1A -mutational status could also prove to be an effective approach to identifying patients and tailoring treatment. Utilizing a proteomics in OCCC cell lines Goldman et al. found that the cholesterol biosynthesis and the mevalonate pathway are differentially regulated upon ARID1A- mutations ( 29 ). As a path forward there is evidence that combining protein or metabolic markers with cfDNA could improve rates of early cancer diagnosis. For example, in pancreatic cancer, combining a protein biomarker (CA19–9) and cfDNA (KRAS mutations) improved the sensitivity of early cancer detection ( 30 ). These studies suggest that combining cfDNA ARID1A testing with protein or metabolic biomarkers could improve early OCCC diagnosis.

ARID1A and SWI/SNF-subunit mutations lead to epigenetic vulnerabilities that are targetable through EZH2 and/or HDAC inhibition. Although further characterization and research into epigenetic vulnerabilities is required because there are clinically applicable epigenetic inhibitors and some that are FDA-approved that could be utilized to treat ARID1A -mutated tumors. As these precision medicine strategies move forward several areas of research are needed including efficiently identifying mutations in patients, therapeutic dosing, and potential immunomodulatory combinations ( Fig. 1 ).

Lack of genomic instability and loss of SWI/SNF epigenetic regulation in OCCC represents a potential precision medicine target. SWI/SNF complexes maintain transcriptional equilibrium. For example, SWI/SNF complexes canonically activate transcription by promoting increase chromatin accessibility and antagonizing the polycomb repressive complex 2 (PRC2) catalytic subunit, enhancer of zeste homolog 2 (EZH2) ( 13 , 14 ). PRC2 represses transcription through the addition of methyl groups onto histone H3 lysine 27 (H3K27Me3), which leads to chromatin compaction ( 15 ). Previously, it was reported EZH2 inhibition is synthetically lethal in SWI/SNF-mutated cancers because of the antagonistic roles played by SWI/SNF and PRC2 in regulating the same target genes ( 13 ). Specific SWI/SNF subunit mutations that convey an EZH2-dependent synthetic lethality include, ARID1A , SMARCB1 , and SMARCA4 ( 13 , 16 , 17 ). Consequentially, several clinically applicable EZH2 inhibitors have been developed. Recently, ARID1A -mutated OCCC cells were described to be highly sensitive to the inhibition of HDAC2 and HDAC6 ( 18 , 19 ). HDAC2 mechanistically interacts with EZH2 and functions as a co-repressor of EZH2 to suppress the expression of EZH2/ARID1A target tumor suppressor genes such as PIK3IP1 ( 19 ). Indeed, ARID1A mutation confers sensitivity to the pan-HDAC inhibitor, SAHA, in both in vitro and in vivo OCCC models ( 19 ). HDAC2 inhibition selectively inhibited growth of ARID1A -mutated cells compared to ARID1A -wildtype cells and there were no significant additive effects induced by concomitant knockdown of EZH2 and HDAC2, which reinforces the hypothesis that these factors have similar functions ( 19 ). Based on the TCGA, ARID1A and TP53 mutations are mutually exclusive across several cancer types, highlighting a potential functional overlap between ARID1A and TP53 and a therapeutic vulnerability in ARID1A -mutated cancers ( 20 , 21 ). Consistently, a novel HDAC6:p53 regulatory axis was observed to be ARID1A -dependent. ARID1A inactivation upregulates HDAC6 and HDAC6 then deacetylates p53 on lysine 120 (p53K120Ac), a pro-apoptotic post-translational modification ( 20 ). A clinically applicable HDAC6 inhibitor, ACY1215, selectively induced apoptosis in ARID1A -deficient tumors compared to ARID1A -wildtype tumors. In OCCC cell line (TOV-21G [ ARID1A -mutated] compared to RMG1 [ ARID1A -wildtype]) mouse xenografts and in a genetically engineered transgenic (Arid1a flox/flox ,Pi3kca H1047R ) mouse model ACY1215 induced a more robust apoptotic response in ARID1A -mutated tumors ( 20 ). Further investigations are needed to elucidate the in vivo potential of HDAC inhibitors, especially in immune competent models. In in vitro and in vivo OCCC models, ARID1A -mutation conveys increased sensitivity to epigenetic directed therapies, but response rates to epigenetic directed therapies in clinical trials of solid tumors have not been as robust ( 22 , 23 ). Notably, one of the first epigenetic-directed therapies, 5-Azacitidine, a DNA hypomethylating agent in clinical trials against leukemia was shown to be highly toxic ( 24 ). However, almost 40 years after 5-Aza’s initial discovery it was FDA-approved for myelodysplastic syndrome because of further investigation with alternative dosing. In most in vitro studies with epigenetic-directed therapies, cells are treated for 7 to 12 days before apoptotic induction or proliferation inhibition are appreciated ( 13 , 19 ). Therefore, the clinical dosing of HDAC or EZH2 inhibitors in ARID1A -mutated cancer should be empirically optimized. These findings may lead to the development of the first effective targeted therapeutic strategy for ARID1A -mutated cancers by repurposing FDA-approved pan-HDAC inhibitors or clinically applicable HDAC6 or EZH2 inhibitors.