*6.1. Epigenetic Drugs as Monotherapies in CM* 6.1.1. DNMTi

DNMTis are potent antineoplastic compounds able to reverse the DNA hypermethylation status of TSGs. Depending on their mode of action, DNMTIs can be divided into two classes: cytosine analogue inhibitors and non-nucleotide analogue inhibitors [239].

Cytosine analogues such as azacytidine, decitabine, zebularine, guadecitabine (SGI-110), fazarabine, and pseudois cytidine, may replace C-5 of cytosine with N-5 into the DNA or RNA backbone, leading to DNMTs degradation, DNA damage, and subsequently, apoptosis [239]. Although they gained FDA approval for the treatment of hematologic malignancies, azacitidine and decitabine showed disappointing results in CM and other cancer patients with solid tumors when employed as monotherapy in Phase I/II clinical trials [240]. The clinical benefits seen in patients with solid tumors exposed to high doses of azacytidine and decitabine are offset by severe adverse effects, such as hematological toxicity [203]. The lowest dose at which decitabine was shown to be effective in treating a solid tumor was 50 mg/m<sup>2</sup> but the reported side effects were severe [241]. Interestingly, the incidence of hematotoxicity may be reduced if these nucleoside analogues are administered by hepatic arterial infusion and not via the intravenous route [242]. Moreover, it seems that low doses of DNMTis (~20 mg/m<sup>2</sup> ) can minimize toxicity and side effects while making tumor cells more sensitive to immunotherapeutic or chemotherapeutic agents [242]. Further clinical research revealed that cytosine analogs can induce ERVs and CTA expression in cancer patients so that cancer cells end up expressing a plethora of neoantigens that can be targeted by immunotherapy [243]. Moreover, it has been reported that DNMTi exposure can induce the activation of transposable elements such as Alu or LINEs, leading to a state of viral mimicry in which treated tumor cells translate the induced expression as caused by an exogenous viral infection, ultimately triggering an innate immune response [203]. Combining DNMTi with chemotherapy is another intensively investigated strategy to improve

the clinical management of CM patients. A Phase I/II study combining decitabine with temozolomide in patients with metastatic melanoma showed promising results in terms of efficiency and safety profile (NCT00715793) [244]. The study reported that administration of decitabine in combination with temozolomide re-sensitizes CM patients to temozolomide by dual modulation of DNA repair machinery due to depletion of DNA repair protein O6-methylguanine-DNA-methyltransferase (MGMT) and subsequent induction of DNA mismatch repair (MMR) pathway in temozolomide-refractory melanoma cells [244].

Other cytosine analogues that are gaining considerable research interest for clinical use in CM are guadecitabine and zebularine. Although these next-generation DNMTis act similarly with azacytidine and decitabine, they have a longer half-life and better bioavailability when compared to their predecessors [237,239]. Guadecitabine has been tested in over 30 clinical trials in cancers, including certain phase III trials (https://clinicaltrials.gov). In particular, for CM, there are no clinical trials in which guadecitabine is tested as monotherapy, there are only two ongoing clinical trials, but in which it is administered in conjunction with immunotherapy (NCT02608437, NCT04250246). Zebularine is another secondgeneration DNMTi that acts by sequestering DNMTs after incorporation into DNA, interfering with their catalytic activities [245]. Preclinical studies in CM highlighted that zebularine treatments can reverse the mechanisms developed by the tumor to get rid of anti-tumoral immunity, priming it for the action of immunotherapies. That is, zebularine allows for the re-expression of intercellular adhesion molecule-1 (ICAM-1) on tumor endothelial cells, which results in restored leukocyte-endothelial cell adhesion and enhanced leukocyte infiltration [216]. Zebularine remains to be further evaluated in clinical trials in CM.

Another class of DNMTi is the class of non-nucleotide analogue inhibitors. This class encompasses small molecules that block the interaction of DNMTs with target sequences either by binding to the catalytic site of DNMTs or by binding to CpG-enriched sequences; however, the antineoplastic effects of non-nucleotide analog inhibitors are inferior to those of cytosine analog inhibitors. They include several compounds such as hydralazine, epigallocatechin gallate (EGCG), and disulfiram [239]. Disulfiram is one of the best-studied non-nucleotide analog inhibitors in cancers, and implicitly in CM. In numerous in vitro experimental studies, disulfiram and its metabolites were extremely potent in blocking tumor growth and inducing apoptosis in tumor cells, but with the disadvantage of strong cytotoxic effects. However, no clinical reports attest to a measurable efficacy of disulfiram as a mono-therapeutic agent against solid tumors to date, most likely due to the low circulating bioavailability of this compound [246]. Due to the potent antiapoptotic activity of disulfiram in tumor cells, it remains a promising approach when administered in conjunction with other pharmacological approaches such as chemotherapy, targeted therapy, radiotherapy, or immunotherapy. However, to minimize the cytotoxic effects and increase the therapeutic benefit of cancer and CM patients, extensive testing of disulfiram is needed in experimental studies and clinical combinatorial trials [246].

#### 6.1.2. HDACi

HDACis are a class of compounds capable of rectifying the aberrant acetylation status of histone and non-histone proteins in cancers to orchestrate TSG activation. In addition, cancer cells display an increased sensitivity to HDACi-induced apoptosis, making those epigenetic agents promising anti-cancer strategies [239]. The FDA has approved four HDACis for use in cancer patients. Three of these compounds are the hydroxamic acids vorinostat, belinostat, and panobinostat, which have been described as non-specific HDAC inhibitors targeting all classical HDACs (classes I, II, and IV). In contrast, the fourth HDACi that gained FDA approval, romidepsin, also known as FK228 or depsipeptide, has been reported to act specifically on Class I HDACs [247]. Other HDACis that have been tested in vitro or are in current clinical trials are well characterized by the scientific literature [203]. Proposed mechanisms of action for HDACis include cell cycle arrest by p53-dependent or -independent induction of the p21CIP/WAF1 axis, downregulation of oncogenes, enhanced

ROS production, autophagy induction, as well as suppression of genes involved in cell survival and EMT programs [245,248].

Several HDACis, including both specific (class I HDAC inhibitor entinostat) and non-specific agents (panobinostat), have recently been tested in phase I/II clinical trials in metastatic CM patients; yet, they have shown reduced efficacy and poor tolerability as monotherapies [249,250]. The most common side effects associated with HDAC inhibition are hematotoxicity, fatigue, nausea, and hyperglycemia [229]. Valproic acid (VA) administration also led to disappointing results in CM clinical cohorts [227]. This was mainly because valproic acid requires a long period, up to several weeks, to reach full dose in most patients, making it almost ineffective in the case of aggressive and rapidly dividing CM tumors. VA was linked to serious adverse effects, such as grade 3 neurological toxicity and intracerebral hemorrhage [227]. In a phase II clinical trial, vorinostat (suberoyl-anilide-hydroxamic acid: SAHA) demonstrated some early responses in patients with advanced CM; however, for the majority of these patients, the disease state was stable, and vorinostat did not meet its primary endpoint of response [251]. As expected, vorinostat therapy was associated with significant side effects, such as lymphopenia, fatigue, and nausea [251]. Surprisingly, quisinostat, an HDACi with increased affinity for class I and II HDACs, showed strong antineoplastic activities and an acceptable safety profile in phase I clinical study of metastatic melanoma patients, calling for further investigation of its clinical efficiency and tolerability in different doses and combinations [252].

In recent years, considerable attention has been paid to sulforaphane (SFN), a natural compound found in cruciferous vegetables, such as broccoli, cauliflower, cabbage, and Brussels sprouts. Being a natural compound, SFN brings with it the advantage of being well-tolerated in human subjects [245]. Additionally, in CM preclinical studies, SFN consumption was linked with the inhibition of UV-induced inflammation and tumorigenesis [253,254]. It was further reported that SFN, through its inhibition of HDACs properties, can regulate apoptotic programs and cell survival pathways to induce cell arrest and apoptosis in melanoma cells [255]. Other mechanisms by which SFN can control tumor growth and evolution in CM cells are by downregulating the metastasis-promoting enzymes MMP-9 and sulfatase-2 [256,257]. However, a pilot study with 17 patients with at least two atypical nevi and a prior history of melanoma revealed that SFN treatment is relevant for chemoprevention in this clinical condition [228]. SFN treatment at 200 µmol daily for 28 days resulted in a significant reduction of the proinflammatory cytokines IL-10 (CXCL10), MCP-1 (CCL-2), MIG (CXCL9), and MIP-1β, and overexpression of tumor suppressor decorin. These encouraging results support that further testing of SFN should be performed in phase II clinical trials at higher doses and over a longer period, ideally accompanied by morphological, histopathological, and molecular analyses of excised tumors [228]. Other HDACis such as trichostatin A (TSA) and suberic bishydroxamate (SBHA) are currently in the preclinical development phase for CM treatment [229].

## 6.1.3. Next-Generation Epigenetic Agents

Novel epigenetic targets identified in cancers have given rise to the next generation of epigenetic therapies that include enhancer of zest homolog 2 inhibitors (EZH2i), bromodomain, and extra-terminal domain inhibitors (BETi), inhibitors of the demethylases JMJD3 and JARID1B, as well as many other compounds [202]. The observation that EZH2 is often overexpressed in cancers led to the development of certain small molecule agents targeting this histone methyltransferase such as EPZ-6438 (tazemetostat), GSK2816126, and CPI-1205 [239,258]; these drugs are currently in early clinical studies reporting clinical responses with acceptable tolerability in many cancers [258,259]. EZH2 is also amplified in CM, and increased levels have been linked with aggressive disease and poor prognosis [77]. Notably, the use of EZH2i in the treatment of wild-type and mutant melanoma cell lines led to a reduction in cell proliferation and growth, legitimating EZH2 as a veritable target in melanoma cells [170]. Another study highlighted that EZH2 depletion may interfere with cancer cell proliferation both in vitro and in vivo via the induction of the p21/CDKN1A-

mediated senescence [124]. Moreover, conditional ablation of EZH2 in a melanoma mouse model impaired tumor growth and attenuated metastasis without affecting the functionality of normal melanocytes [76]. These effects were recapitulated by pharmacological inhibition of EZH2, suggesting that EZH2 inhibitors are promising therapeutic approaches for the treatment of CM patients [76].

Another class of epi-therapies that is gaining considerable research interest in cancers is that of BET inhibitors (BETi). BETis exert their anti-neoplastic activities in tumor cells by preventing BETs-acetylated histones interaction [239]. So far, several BETis, such as thienodiazepine JQ1, I-BET151/GSK1210151A, I-BET762/GSK525762, ABBV-075, PLX51107, ODM-207, and ZEN003694 have shown encouraging clinical outcomes with tolerable toxicity and increased efficiency in various tumor types [260,261]. BRD4 is a notorious BET family member that is consistently reported to be amplified or overexpressed in human melanoma lines and primary tumors [262]. Notably, one study suggested that BETi treatments significantly affected melanoma cell proliferation in vitro and tumor growth and metastasis in vivo due to the downregulation of genes involved in cell cycle progression (SKP2, ERK1, and c-MYC) and concomitant accumulation of cyclin-dependent kinase inhibitors (p21 and p27). These effects were also reported upon individual silencing of BRD4. However, the efficiency of BETi was not affected by the BRAF or NRAS mutational status, suggesting that BRD4 inhibitors may be a promising approach for the treatment of CM patients that are not eligible for targeted therapy [134]. NHWD-870, a novel BRD4 inhibitor that suppresses the secretion of macrophage colony-stimulating factor (CSF)-1 by tumor cells and disrupts the BRD4/HIF-1α axis, is currently under preclinical investigation for the treatment of CM [233]. Another BETi, I-BET151, triggered the selective inhibition of the NF-κB signaling pathway in melanoma cells, regulating genes involved in inflammation (VEGF, CCL-20) and cell cycle progression (CDK6) and suppressing the production of IL-6 and IL-8 via BRD2 displacement [263]. PLX51107, a next-generation BETi, impaired tumor growth to differing degrees in BRAF V600E syngeneic mouse models, effects that were linked to the influx of TAMs. Tumors that were poorly responsive to BET inhibition displayed an increased influx of pro-tumoral macrophages and the addition of PLX3397 resulted in improved response rates to PLX51107, offering a novel combination therapy for metastatic CM patients [234]. In another study, PLX51107 slowed the growth of mouse BRAF V600E melanoma tumors by inducing the CD8<sup>+</sup> T cell-mediated anti-tumor effects; moreover, PLX51107 proved to be an effective second-line therapy for CM tumors that harbored resistance to PD1/PDL1 checkpoint inhibition [264]. While certain BETis have recently started to be evaluated in phase I clinical trials in CM patients [232], novel pharmacological inhibitors continue to be tailored to target other classes of epigenetic enzymes, such as JMJD3 and JARID1B histone demethylases [265,266]. Therefore, the continuous enrichment of the anti-cancer armamentarium with next-generation therapies and their exploration either alone or in conjunction with traditional therapies is expected to positively impact CM therapeutic management and open up new opportunities for precision medicine in those patients.
