Novel insights into the pathogenesis and treatment of NRAS mutant melanoma

Abstract

Introduction:

NRAS was the first mutated oncogene identified in melanoma and is currently the second most common driver mutation in this malignancy. For patients with NRAS^mutant advanced stage melanoma refractory to immunotherapy or with contraindications to immune-based regimens, there are few therapeutic options including low-efficacy chemotherapy regimens and binimetinib monotherapy. Here, we review recent advances in preclinical studies of molecular targets for NRAS mutant melanoma as well as the failures and successes of early-phase clinical trials. While there are no targeted therapies for NRAS-driven melanoma, there is great promise in approaches combining MEK inhibition with inhibitors of the focal adhesion kinase (FAK), inhibitors of autophagy pathways, and pan-RAF inhibitors.

Areas Covered:

This review surveys new developments in all aspects of disease pathogenesis and potential treatment – including those that have failed, stalled, or progressed through various phases of preclinical and clinical development.

Expert Opinion:

There are no currently approved targeted therapies for BRAF wild-type melanoma patients harboring NRAS driver mutations though an array of agents are in early phase clinical trials. The diverse strategies taken exploit combined MAP kinase signaling blockade with inhibition of cell cycle mediators, inhibition of the autophagy pathway, and alteration of kinases involved in actin cytoskeleton signaling. Future advances of developmental therapeutics into late stage trials may yield new options beyond immunotherapy for patients with advanced stage disease and NRAS mutation status.

1. Introduction

Mutations in the RAS family of genes act as oncogenic drivers in 15–36% of all melanomas, with most mutations localizing to NRAS^Q61[1–11]. Downstream signaling through NRAS is critical to diverse cellular processes involved in tumorigenesis including cell proliferation, metabolism, and survival. RAS has been considered by most to be an “undruggable” target based on the fact that this membrane-tethered intracellular protein forms extremely tight bonds between cycling active RAS-GTP bound and inactivated RAS-GDP forms [12]. Though immunotherapy has revolutionized the landscape of treatment for advanced melanoma, a considerable number of patients harboring NRAS mutant tumors fail to see disease improvement or develop dose-limiting autoimmune toxicities to checkpoint inhibitors. [13–17]. Despite the fact that NRAS was the first oncogene identified in melanoma, patients with this subset of mutations have yet to experience the benefit of breakthroughs in targeted therapy that have revolutionized the treatment of BRAF^V600-mutant melanoma. In this review, we provide an update on the role of NRAS in melanoma tumorigenesis and highlight key developments in preclinical and clinical studies of emerging therapeutics.

2. The role of NRAS in the pathogenesis of melanoma

The RAS proto-oncogene family is one of the oldest and most studied in oncology[3,18–22]. NRAS mutations are the second most frequent driver mutation in cutaneous melanoma[23,24]. Mutations in the RAS family of genes occur in a quarter of all human cancers and approximately the same fraction of melanomas, with over 130 unique missense mutations reported in the literature [25–28]. NRAS is one of three proteins encoded by the RAS family of genes in humans, with the two other family members being HRAS and KRAS [29,30]. All proteins transition between active RAS-GTP bound and inactive RAS-GDP forms in a cycle that is mediated by guanine nucleotide exchange factor (GEF) and GTPase-activating protein catalysis (GAPs) and each also harbors intrinsic guanosine 5’-triphosphate (GTP) binding activity. Activation of the cell pathway by upstream growth factor receptors leads to downstream mitogen activated kinase (MAPK), RAF, MEK, and ERK phosphorylation as well as activation of phosphatidylinositol-3- kinase (Protein kinase B)/AKT pathways. NRAS activation of downstream MAP-kinase family effectors occurs through dimerization of BRAF or CRAF homodimers or BRAF-CRAF heterodimers. Oncogenic BRAF mutations, in comparison, drive MAPK signaling through a BRAF^V600 mutant monomer that is blocked by BRAF inhibitors. Transcription factors activated downstream of NRAS-MAPK include serum response factor (SRF), JUN, activating transcription factor 2 (ATF2), and NF-κB, ultimately lead to increased cyclin D1 expression and progression through the G1 phase of the cell cycle. Further shifts of signaling pathway networks downstream of these effectors creates a cellular state of pro-survival and proliferative activity.

Compared to other solid tumors featuring driver mutations in RAS family genes, the landscape of mutations in human melanoma predominantly involves NRAS, with HRAS and KRAS mutations found in only 1% of human melanomas. Two decades of bulk and single-cell resolution tumor sequence data have painted a picture of near uniform mutual exclusivity of oncogenic BRAF and NRAS driver mutations in treatment-naïve melanoma [31–40]. NRAS mutations appear to also occur independently of other aberrant signaling pathway mutations including mutations in phosphatase and tensin homolog (PTEN)[41,42]. In contrast to cutaneous melanoma, NRAS mutations are seldomly seen in uveal melanoma[43].

Though canonical driver[44] mutations in NRAS are causally related to tumor formation in model systems, the development of human cutaneous melanoma from normal skin and precursor lesions is multifactorial. Human melanomas featuring these mutations arise in a complex series of steps at the intersection of environmental risk factors including UV exposure as well as germline genetic variants. Sun exposure (from chronic intermittent environmental sources) has been shown to be the major environmental risk factor for WHO pathway I-III melanomas (including superficial spreading, lentigo maligna, desmoplastic histologic subtypes) and nodular melanomas[45]. The role of UV radiation in NRAS-mutant melanoma pathogenesis is complex, as exposure to sun damage is not required for development of genetic alterations at the NRAS locus demonstrated by the acral and mucosal subtypes as well as melanomas arising in a congenital nevus (WHO V, VI, and VII). Non-sun exposed melanomas – including acral lentiginous melanomas – also harbor NRAS mutations but less frequently than sun-associated melanomas[46]. For UV-associated melanomas, irradiation appears to mediate base pair mutagenesis in-vitro and in-vivo within the confines of the NRAS gene. UV exposure produces signature cyclobutane pyrimidine dimers at codons 12, 13, and 61. Compared to the heterogeneity of UV signature mutations seen at other loci including TP53, CDKN2A (p16^INK4A) and PTEN/MMAC1, UV-associated lesions that fall within NRAS are positioned in a highly stereotypical manner at just a few genomic hot spots [1]. Interactions between UV exposure and tumor genetic background appear to be codon-specific, with transgenic NRAS^Q61R but not NRAS^G12D mice developing increased rates of tumor formation after UV treatment[47].

The most common mutations in NRAS fall in exon 1 and 2 at codons 12, 13, and 61 with strong evidence for allele specificity in driving tumorigenesis[48]. Biochemical studies into several of the most common oncogenic NRAS proteins demonstrate a diverse range of mechanisms involved in oncogenic protein activity. Of all NRAS mutations in melanoma, the vast majority (~84%) occur in codon 61, with mutations at this codon displaying a non-UV signature mutation[49]. Mutations at the NRAS^Q61 site dampen catalytic GTPase activity, shifting towards a GTP-bound active NRAS conformation that drives downstream cell signaling through the MAPK pathway[50]. In contrast, codon 12 and 13 mutations in all RAS proteins lead to inactive GTPase domains. Conditional knock-in mouse models have shown that NRAS^Q61R but not NRAS^G12D promotes tumorigenesis through downstream activation of PI3K and MAPK signaling in a p16^INK4a loss of function melanoma model[51]. Similarly, a study of seventy-seven cutaneous melanoma metastases revealed that p16^INK4A promoter methylation was more frequent in NRAS-mutant tumors compared to NRAS wild-type[52]. Collectively, these data suggest possible cooperativity between p16^INK4A and NRAS mutations in driving melanoma formation, which has translated to emerging treatment strategies currently being studied which are discussed in later sections.

We now have evidence at the genomic resolution of single human melanocytes that supports the involvement of NRAS at the most proximal steps of the original Clark model of melanoma formation[53,54]. Melanocytes isolated from normal skin obtained from human cadavers and freshly discarded from surgery adjacent to melanomas carry mutations in NRAS^Q61 [55]. In studies of the progressive evolution of precursor lesions to cutaneous melanoma using microdissection and DNA sequencing of formalin fixed paraffin embedded biopsy specimens, mutations in NRAS are early and intermediate steps that occur in parallel to BRAF prior to later genomic events such as alterations in the TERT promoter[56]. Tertiary genomic alterations including copy number changes subsequently develop and push nevi further along the path of melanoma after awakening melanocytes from states of oncogene-induced senescence[57–59]. There is strong evidence that premalignant melanocytic nevi – including those carrying NRAS mutations – progress to form melanoma though there are also isolated reports of melanomas arising from nevi with discordant mutational profiles[59–61]. Benign and premalignant melanocytic lesions including congenital nevi carry mutations in NRAS but not BRAF, whereas histologically similar acquired nevi harbor BRAF^V600E mutations more frequently than alterations in NRAS[62–64].

Copy number variations (CNV) of NRAS also play a role in tumorigenesis and carry prognostic significance. NRAS amplifications have been shown to occur in 3–14% of patients, and NRAS gains – along with tumor thickness and AJCC stage – are independent prognostic factors for melanoma[65]. From a therapeutic perspective, CNV status may be associated with response to MEK inhibitors (MEKi) from data obtained from human xenograft melanoma models[66].

3. NRAS mutation status and patient prognosis

NRAS mutation status for melanoma appears to confer distinct histopathologic and prognostic implications, though such data must be carefully examined in aggregate given the numerous single and multi-center retrospective studies in the literature in contrast to a much larger pool of prospectively obtained data from BRAF-mutant patients. The preponderance of such data suggests that NRAS status carries a worse prognosis for melanoma. Other studies have reported mixed data on the association between NRAS mutation status and overall survival, with Omholt et al., 2003 and Ururel et al., 2007 indicating that overall survival (OS) for NRAS^mutant melanoma patients was significantly decreased compared to those of BRAF^mutant or NRAS/BRAF^{wild type} tumor status[67]. NRAS mutation status has been associated with higher rates of mitoses, thicker primary tumors, and shorter tumor-specific survival compared to BRAF[33,68]. In a meta-analysis, melanomas with NRAS mutation more frequently developed from areas of chronic sun damage than those with BRAF mutations and were found more often at the limbs [69,70]. As more trials using next generation sequencing and targeted sequencing panels accrue patients and molecular testing expands in clinical practice, larger pools of prospectively obtained data will shed light on the clinical utility of genetic alterations at this locus for potential prognostication[71,72]. The Melanoma Institute of Australia and Novartis are exploring the utility of a 28 gene panel for assessment of type and frequency of BRAF and NRAS aberrations in a 1,000 patient study (MatchMel; NCT02645149) that will document response to targeted therapies including, trametinib, CDK4/6 inhibitors and MEK inhibitors.

4. From the bench to bedside: key preclinical developments in therapeutic targeting

Several groups have reported success in preclinical studies of small molecule and antibody-based approaches to directly inhibit NRAS signaling as well as blockade of signaling pathways including RAS-RAF-MEK-ERK. The reinvigoration of the field of direct, allele-specific RAS inhibitors by recent breakthroughs in targeting KRAS bodes well for the future generation of agents that may one day prove similarly effective for NRAS mutant melanoma. NRAS and its effectors pose a difficult target to inhibit with its intracellular location and redundancies in downstream signaling pathways. Several small molecule inhibitors of NRAS, MEK, ERK, cell cycle regulators, and post-translational modifiers have been shown to disrupt tumorigenesis in vitro and in vivo. Genome-wide screens continue to identify new cellular pathways that are critical for cell survival and proliferation in the background of oncogenic NRAS signaling and thus are strong candidates for therapeutic targeting[73]. To date, creative attempts at effectively blocking activation downstream from NRAS have not translated well in clinical trials. The literature is abundant in emerging avenues of potential pharmacologic blockade.

4.1. Direct targeting of NRAS and RAS effectors

Diverse approaches to pharmacologic inhibition of NRAS have been reported, ranging from molecules that directly bind conserved protein residues to inhibitors of other mediators of the GTP exchange cycle including SOS1 and antibody based approaches targeting RAS directly. Many efforts have focused on disruption of the C-terminal hypervariable region (HVR) interaction involved in tethering of the RAS protein to the inner plasma membrane leaflet. Compound 3144 has been shown to bind Asp38 to block RAS effectors from interacting with all isoforms of wild-type KRAS, NRAS, and HRAS while simultaneously inhibiting tumor growth in-vivo[74]. Such pan-RAS inhibitors have toxicity profiles that are in line with the biology of extremely conserved signaling pathways. Genetic deletion of all three RAS isoforms in mice produces an embryonic lethal phenotype. Alternatively, agents such as 4,6-dichloro-2-methyl3-aminoethyl-indole (DCAI) weakly bind RAS at a site of interaction between RAS and GEF and inhibits SOS1-mediated guanine exchange. Families of molecules in this DCAI class similarly have been shown to inhibit interaction of all RAS isoforms with important binding partners and effector proteins including CRAF, guanine nucleotide dissociation stimulator (RalGDS), and PI3K. Sotorasib (AMG 510), a small molecule that specifically and irreversibly binds the switch II pocket region of the KRAS^G12C protein, has been successful in recent clinical studies of other solid tumors with hope that the same paradigm can eventually be applied to NRAS mutant melanoma[75,76]. This molecule has demonstrated objective clinical responses in both lung and colon cancer for tumors expressing the KRAS^G12C alteration. Other KRAS^G12C inhibitors (MRTX849, MRTX1257, GDC6036, LY3499446, and JNJ74699157) are being tested in trials for non-melanoma solid tumors[77]. In recent years, several antibody[78] and protein-based[79] approaches have shown promise in direct inhibition of RAS in cell culture and animal models, including a recent study published using an IgG-based approach of pan-RAS blockade in human cell lines and mice [80].

4.2. MEK inhibitors

MEK1/2 phosphorylation occurs downstream of NRAS activation and blockade of this critical kinase junction has been explored in various preclinical settings both as monotherapy and combination therapy. Three FDA approved MEK kinase inhibitors – binimetinib, cobimetinib, and trametinib, are currently indicated for use in melanomas with BRAF^V600 mutations though to date no clinical trial has identified a MEK monotherapy agent as effective for NRAS-mutant solid tumors[81,82]. Outside of melanoma, MEK inhibition is being employed under the compassionate use setting by dermatologists for the treatment of large and giant congenital melanocytic nevi, which frequently harbor NRAS driver mutations[83,84]. These premalignant growths, defined as lesions >20 cm in size, are associated with extracutaneous complications such as leptomeningeal involvement and increased risk for developing melanomas of the skin and central nervous system even after definitive surgical resection of the primary lesion[85]. It is clear from emerging studies that MEK blockade can lead to downstream pathway reactivation in a manner analogous to paradoxical reactivation in the setting of BRAF inhibition. Preclinical xenograft studies of CI-1040 and PD0325901 showed partial response in mice with NRAS mutant melanoma[86,87]. Initial clinical data from a phase I studies of PD-0325901 showed promise but was halted in further development, as neurotoxicity was observed in a subset of patients[88,89]. Agents that can re-sensitize cells to MEK inhibition are being explored in the preclinical setting. Drug combinations that that act on endoplastic reticulum stress pathways including BRAF inhibitors in combination with MEKi have shown viability in preclinical studies[90]. A recent study showed that low doses of the bacteria compound anisomycin restores NRAS-driven melanoma sensitivity to MEKi through activation of the P38α-MAPK14 axis [91]. Further developments in clinical trials for MEK inhibitors are discussed in a subsequent section.

4.3. ERK inhibitors

For many years, ERK inhibitors were pursued as the most attractive inhibitors in MAP kinase mutant tumors. ERK phosphorylation downstream of MEK is the next step in the MAP kinase signaling pathway (Figure 1). Recent developments in the understanding ERK signaling in the context of NRAS-mutant melanoma has led to an interest ERK5 (also referred to as Big MAP Kinase or BMK1) as a promising therapeutic target. Inhibition of ERK5 was shown to re-sensitize NRAS mutant melanoma cells to MEK inhibition. The paradigm of MEK inhibitors leading to compensatory ERK5 phosphorylation and downstream signaling has now been shown in cell culture models for KRAS mutant colon carcinoma as well as several different NRAS mutant human melanoma cell lines[92,93]. Combined ERK5- inhibition with vemurafenib (BRAF inhibitor) has been shown to disrupt nuclear translocation of ERK5 in V600E mutant cells, while vemurafenib alone did not fully impede ERK5 activity[94]. In light of these recent advances in pursuing ERK5 as a novel target for NRAS mutant melanoma multiple pharmaceutical companies are pursuing this drug discovery pipeline[95].

Fig 1.

A diverse spectrum of signaling pathways currently being explored as targets in preclinical and early phase clinical studies of NRAS mutant melanoma

4.4. Cell cycle and mitotic regulators and effectors

Cyclin-dependent kinases (CDKs) are serine-threonine protein kinases that are involved in progression through the cell cycle via phosphorylation-based regulation of phase-specific cyclin proteins. Roughly 90% of melanomas studied have an alteration in a component of this pathway[38,96,97]. Their dysregulation in oncogenic NRAS-melanoma cells leads to unperturbed cell cycle progression and tumorigenesis as well as metastatic dissemination. Cyclins, CDKs and associated regulators are implicated in the pathogenesis of melanoma include CDK4, CDK5[98], CDK6, CDK11, the CDK4-CCND1 inhibitor CDKN2A (p16^INK4A), G1 phase-predominant cyclin D (encoded by CCND1), Aurora kinase A, the retinoblastoma protein (Rb)[99] and the forkhead family transcription factor FOXM1[38,100]. CDK11 is up-regulated in several BRAF- and NRAS-mutant human melanoma cell lines and has been shown to be critical for cell survival in a subset of melanoma cell lines[101]. Knockdown with siRNA in BRAF and NRAS-mutant melanoma in vitro reduced cell survival and disrupted cell cycle progression in both mitotically active and quiescent cell lines. In silico models of signaling pathway networks disrupted by NRAS^Q61K suggest that combined MEKi with CDK4 inhibitors provides added benefit for disruption of downstream oncogenic signaling[102]. Similarly, a potentially synergistic effect exists with combined CDK inhibition and immune checkpoint therapy in preclinical models[103–105] as well as in a small cohort of patients treated with anti-PD-1 therapy harboring CDK4 gain-of-function mutations or amplification associated with resistance to immune checkpoint inhibitor (ICI) blockade[106].

The PI3K/AKT/mTOR pathway has been explored as a preclinical target given the role of this intracellular signaling cascade in suppression of apoptosis and promotion of cell proliferation. Inhibition of MEK combined with inhibition of PI3K/mTOR produces a synergistic effect in human NRAS mutant melanoma cell lines and cause xenograft models of NRAS melanomas to decrease in size[107]. Clinical developments in targeting this pathway are discussed in a subsequent section.

4.5. Inhibition of MAP kinase in combination with autophagy blockade.

Trials using agents targeting the autophagy pathway have been initiated based on preclinical findings that MAPK pathway inhibition of RAS/RAF/MEK/ERK signaling elicits heighted amounts of cell autophagy, a process of intracellular recycling that protects pancreatic cancer cells lines (PDA) from the cytotoxic effects of KRAS pathway inhibition[108,109]. Mechanistically, inhibition of MEK1/2 leads to activation of the LKB1/AMPK/ULK1 signaling axis, a key regulator of autophagy. ERK inhibition does the same to autophagy increasing its activity to 3–20 times baseline. Furthermore, combined inhibition of MEK1/2 plus inhibition of autophagy displays synergistic anti-proliferative effects against PDA cell lines in vitro, and promotes regression of xenograft patient-derived PDA tumors in mice. ERK inhibition leads to a greater dependency on autophagy. Activity has been observed preclinically in BRAF resistant BRAF^V600 with colon and melanoma and KRAS^mutant pancreatic and NRAS^mutant melanoma. Furthermore, ERK inhibition likely induces damage to mitochondrial function and autophagy can block the mitochondrial compensatory activity. Remarkably, the T/HCQ (trametinib/hydroxychloroquine) combination resulted in substantial reduction in overall tumor burden, CA19–9 tumor marker, and resolution of debilitating cancer pain in a single patient on a compassionate use protocol[109].

4.6. Attempts at targeting post-translational modification of RAS

Alterations in post-translational modification in addition to acquired secondary mutations may frequently drive drug resistance in melanoma. Post-transcriptional and post-translational changes may be playing a role. Farnesylation and palmitoylation have been explored as possible targets of novel therapies. NRAS plasma membrane localization is dependent on a cycle of palmitoylation and depalmitolyation mediated by the enzymes acyl protein thioesterases 1 and 2 (APT-1, APT2), respectively[110–112]. Palmostatin B inhibits NRAS^G12D hematopoietic cells in vitro through a mechanism that disrupts physiologic NRAS membrane localization. RAS family GEF and GAPs are also being explored in preclinical studies, though there have been no breakthroughs with targeting these GTPase enzyme mediators for NRAS melanoma[113]. Post-translational protein modification pathways including ubiquitin-mediated degradation offer another avenue of attack. The E3 ubiquitin ligase encoded by MDM2 is selectively inhibited by the small molecule, PDX1826, which in combination with CDK4/6i therapy resulted in decreased melanoma growth in vivo and in patient-derived xenografts[114].

The trafficking of RAS from a nascent polypeptide in the Golgi apparatus to a mature, functional membrane-tethered protein requires several tightly orchestrated biochemical steps that – when disrupted – can interfere with downstream signaling pathway activity. NRAS requires farnesylation for translocation to the plasma membrane inner leaflet. Data from RAS-mutant breast and lymphoid cancer studies first showed the promise of farnesyl transferase inhibition as an approach for untethering oncogenic RAS from activating downstream signaling activities. Lonafarnib and tipifarnib failed to meet efficacy endpoints in NRAS and KRAS-mutant cancers. In a recent phase II clinical trial, the farnesyltransferase inhibitor (FTI) R115777 showed no clinical response in 14 patients[115]. Outside of melanoma, Tipifarnib showed promise in an AML trial with 3/54 patients developing complete remission, a phase II study for HRAS mutant metastatic urothelial carcinoma patients in 5/12 patients, and is currently being explored in the context of KRAS mutant lung cancer. Mechanistically, these shortcomings have been explained by compensatory geranylgeranyltransferase activity that manifests during farnesyltransferase blockade to provide an alternative pathway for producing biochemical tethering of RAS proteins to the cell membrane[116]. A new preclinical approach of targeting depalmitoylation was successful in NRAS-mutant melanoma cell lines and a mouse xenograft model. RAS intracellular trafficking following depalmitoylation and the dynamic nature of these pathways and built-in biological redundancies may offset potential clinical efficacy of this class of emerging therapeutics [117,118]. To our knowledge, none of the above avenues of lipid biochemistry-based or ubiquitin pathway-targeted modalities have moved forward into clinical trials, but there is extensive activity in the medicinal chemistry and preclinical literature.

4.7. Targeting TERT, heat shock proteins, and mediators of genomic stability

The human telomerase catalytic subunit, TERT, and its adjacent gene regulatory elements are involved in the pathogenesis of melanoma and have recently been explored as a potential therapeutic target in preclinical models of NRAS-mutant melanoma. In NRAS-mutant human melanoma cell lines, silencing of TERT with short hairpin RNA (shRNA) targeting induced apoptosis within 7–10 days by a mechanism that appears to involve telomeric DNA damage and non-telomeric DNA damage pathways. Treatment of cells with the telomere uncapping molecule 6-thio-dG induced cell death at 7–14 days, similar to the timeline of shRNA downregulation[119]. Compensatory increases in enzymes involved in response to reactive oxygen species (ROS) – including SOD2 – were observed in parallel to treatment with 6-thio-dG. The authors speculated that dual inhibition of mitochondria would exploit the vulnerability of TERT-starved NRAS-mutant melanoma cells to oxidative damage. By combining 6-thio-dG and Gamitrinib, a mitochondrial disrupting agent that downregulates SOD2, the authors observed enhanced NRAS-mutant melanoma cell death and increased generation of ROS. This points to a potential therapeutic vulnerability of NRAS-mutant melanoma of potential interest in exploring in future clinical trials.

The heat shock protein family, including HSP90, has also been explored separately as a potential therapeutic target. These proteins, essential to cell housekeeping functions including proteostasis, are expressed in higher amounts in tumor tissue compared to benign melanocytic nevi. HSP90 in melanoma cell lines is involved in diverse processes including clearance of debris for antigen presentation during cell death, cell migration, pro-survival pathways, and angiogenesis. Studies into the inhibition of heat shock proteins has revealed potential utility of small molecule inhibitors including XL888[120], a compound with specific activity against HSP90, decreased RAF/MEK/ERK/CDK4 and decreased PI3K and AKT signaling[121]. Inhibition of HSP90 in human melanoma cells leads to reprogramming that restores cell pigmentation and melanin synthesis[122]. There are two clinical trials for XL888 that are active and/or recruiting with inclusion criteria encompassing melanoma (NCT02721459, NCT01657591). Neither trial is enrolling patients based on NRAS mutation status and are instead designed as dose escalation regimens for XL888 with vemurafenib plus cobimetinib (NCT02721459) and vemurafenib with XL888 (NCT01657591) for patients with cytology or histology-confirmed diagnoses of AJCC Stage IIIB/C and IV BRAF^V600 mutant melanoma.

4.8. RNA-based strategies

Though few advances have been made in the last several years for this therapeutic modality for melanoma, the leaps in innovation in small interfering (siRNA) and RNA interference (RNAi) is promising for melanoma therapy. This technology relies on intracellular activity of exogenously delivered short oligonucleotides approximately 20 base pairs in length that trigger degradation of target gene transcripts prior to host cell mRNA translation. Such small interfering RNA for melanoma and other cancers have already been explored in preclinical studies showing effective knockdown of STAT3 and BRAF in mice, in vitro melanocytes, and ex vivo porcine skin [123,124]. Previously, technical hurdles surrounding nanoparticle-mediated delivery of shRNA presented obstacles for movement of these agents into the clinic[125]. The unprecedented success of two RNA-based vaccines containing COVID-19 spike protein sequences packaged into biocompatible delivery systems capable of efficient cellular uptake bodes well for future progress using this modality to target human melanomas [126–128]. In a 2010 phase I study including melanoma patients, systemic delivery of RNAi was shown to produce successful targeting of the RRM2 transcript at the site of tumor though no recent results have been published after the study’s termination in 2012 (NCT00689065) [129]. The melanoma-associated transcription factor encoded by MITF has also been successfully downregulated at the RNA level by siRNA in melanoma xenografts obtained from patients[130].

5. The current landscape of clinical trials for NRAS mutant melanoma

Compared to targeted therapy for BRAF^V600 only one agent – the MEK inhibitor binimetinib – has an NCCN guideline indication for use in the setting of NRAS mutant melanoma. For this reason, NRAS mutation status is unknown for the vast majority of advanced stage melanoma patients, and many in this population would likely have initiated standard-of-care mutation-agnostic immune based regimens prior to consideration for NRAS-directed trial enrollment. In contrast, BRAF mutation status is known for nearly all melanoma patients given the wide use of FDA approved agents in the last decade for metastatic disease and more recently in the adjuvant setting. Since BRAF is downstream of NRAS, in theory FDA approved BRAF inhibitors such as vemurafenib, dabrafenib or encorafenib could be effective against oncogenic NRAS; however, this is not the case due to the different mechanism of activation seen in BRAF^V600 mutant and NRAS^Q61 mutant melanomas. In fact in NRAS-mutant melanoma, BRAF type I inhibitors lead to paradoxical and hyper-activation of downstream MEK-ERK1/2 by their activation of RAF dimers leading to tumor cell proliferation in-vitro [131–134]. BRAF^V600 inhibition is effective only when BRAF^V600 monomers drive downstream activation rather than dimers. This is the basis for development the “paradox-breaking” class of BRAF inhibitors, which are profiled below. It is not yet established if candidates such as PLX8394 or other pan-RAF inhibitors are clinically effective against NRAS-mutant melanoma[135]. The approaches described below have either been proposed, are in early phase trials, or are actively recruiting to assess efficacy and further evaluation of clinical endpoints in phase III trials.

5.1. MEK inhibitor monotherapy

Early observations in trials enrolling BRAF^V600 -wild type melanoma showed some clinical activity for trametinib against NRAS melanoma. But the clinical activity was very limited and there was insufficient enthusiasm from the sponsor for further development. Subsequently binimetinib, another selective MEK inhibitor was evaluated in a phase I/II trial with an objective clinical response rate of 20% in NRAS mutant melanoma patients. This small phase I/II trial enrolled only 28 melanoma patients including some with BRAF^V600-mutant melanoma.

Clinical responses in this small phase I/II led to a large phase III trial of binimetinib randomized against dacarbazine, enrolling 402 patients in a 2:1 randomization [136]. Dacarbazine had been the standard control in trials prior to the approval of immune checkpoint inhibitors. The trial demonstrated a modest PFS advantage (2.8 mos vs 1.5 mos, p < 0.001) for Binimetinib over dacarbazine but no improvement in OS[136,137]. When this analysis was restricted to patients enrolled late in the trial when prior immunotherapy was allowed, the effect size was more impressive, as 85 patients who received prior immunotherapy were randomized with 57 patients receiving binimetinib vs 28 patients randomized to dacarbazine. The PFS was (5.5 vs 1.8 months) for binimetinib versus dacarbazine with a HR 0.46 (95% CI 0.26–0.81). With the present patterns of clinical practice, all NRAS melanoma patients have prior immunotherapy to targeted therapy the later findings may be more relevant. For now, this MEK inhibitor therapy is not a standard single agent therapy for NRAS^mutant melanoma despite current NCCN 2B recommendations for use of binimetinib in this patient population.

5.2. Cell cycle and MEK inhibition

Cell cycle inhibition with a cyclin-dependent kinase (CDK) CDK4/6 inhibitor in combination with MEK inhibitor appeared to mimic the anti-tumor effect of direct NRAS inhibition in several animal models, and other preclinical evidence demonstrated synergy of MEK and CDK4/6 inhibition against NRAS-mutant melanomas. Targeting this axis with the aim to enhance the benefit in those melanomas characterized by NRAS-activating mutations (G12, G13, and Q61) has been evaluated in a phase 1b/2 study, where investigators assessed the combination of Ribociclib, an orally available small-molecule inhibitor of CDK4/6 and binimetinib, an orally available ATP noncompetitive, highly selective inhibitor of MEK1/2 (NCT01781572). The phase 1b/2 clinical trial enrolled 61 patients in the dose escalation phase in order to define a recommended phase II dose. In the dose-expansion phase, 41 patients were treated with binimetinib 45 mg BID combined with ribociclib 200 mg QD on a 28-day schedule. The most common reasons for discontinuation in the dose escalation phase were progressive disease (23 patients [56.1%]) and AEs (11 patients [26.8%]). In the dose seeking phase, 6 patients (20.7%) had a PR, 14 patients (48.3%) had SD as best response, and 4 patients (13.8%) had PD with 6 patients reported as unevaluable. The confirmed ORR was 20.7% (95% CI: 8.0, 39.7), based on the investigator’s assessment. In the dose expansion phase with 41 patients, the confirmed ORR was 19.5% (95% CI: 8.8, 34.9), with all 8 patients achieving PR. The response was listed as unknown for 6 (14.6%) of the patients and they were not included in the response data. Median duration of response was 10.3 months (95% CI: 4.1, not estimable). Overall, the majority of patients achieved disease response or stabilization, with a confirmed disease control rate (DCR) of 70.7% (95% CI: 54.5, 83.9).

Tumors with alterations in CDK4, CCND1, and CDKN2A, whose gene products act at the G1 cell cycle checkpoint, are implicated as potentially responsive to agents such as the CDK4/6 inhibitor Palbociclib. Indeed, the response rate in patients with alterations of cell cycle regulators in NRAS-mutant melanoma was higher than in patients without such alterations, ORR of 32.5% (95% CI: 18.6, 49.1, n = 40) and 10% (95% CI: 2.8, 23.7, n = 40), respectively. Post-hoc analysis revealed a significant difference of the ORRs of both subgroups (p =0.014). Future plans for this regimen are still under discussion; however, this may be restricted to genetically selected patients and toxicity may limit its further development.

As previously discussed, preclinical studies of PI3K inhibition have led to trials exploring combined MEK and PI3K pathway targeting. A non-randomized multicenter phase II study of combination trametinib and GSK2141795 enrolling n=10 NRAS wild-type and n=10 NRAS mutant patients yielded no objective response in either cohort (NCT01941927). The treatment regimen was not well tolerated, with 4 patients of 20 enrolled developing a grade 3 or 4 rash[138]. No recent trials targeting PI3K have been initiated since publication of these results in 2018.

5.3. ERK inhibitors

Ulixertinib (BVD-523) is a highly potent, selective, reversible, ATP-competitive ERK1/2 inhibitor that has been shown to reduce tumor growth and induce tumor regression in BRAF and RAS-mutant xenograft models[139]. Furthermore, single-agent ulixertinib was found to inhibit tumor growth in human xenograft models that were cross-resistant to both BRAF and MEK Inhibitors. An open-label, first-in-human study (NCT01781429) of oral ulixertinib to identify both the maximum tolerated dose (MTD) and the recommended phase II dose (RP2D) was conducted[140]. The study aims included assessment of the pharmacokinetic and pharmacodynamic properties of ulixertinib then proceed to an expansion phase for several cohorts defined genetically. Preliminary efficacy has been reported in patients with genetic aberrations in BRAF – including the V600 and non-V600 positions –NRAS, and MEK mutated melanoma. Seventeen patients with NRAS mutated melanoma were evaluable for response. Three patients (18%) achieved a PR, 6 had stable disease, and 8 had disease progression as best response.

With an orally bioavailable ERK inhibitor MK-8353, early phase studies were conducted to determine anti-tumor pharmacodynamic endpoints, dosing, and schedule. MK-8353 was evaluated in a phase I clinical trial in patients with advanced solid tumors. Forty-eight patients were enrolled in the study of which twenty-six patients were in the dose expansion phase[141]. Adverse events included diarrhea (44%), fatigue (40%), nausea (32%), and rash (28%). Dose-limiting toxicity was observed in the 400mg and 800mg dose cohorts. Sufficient exposure to MK-8353 correlated with biological activity in preclinical data. Three of fifteen patients evaluable for treatment response in the MK-8353–001 study had partial response, all with BRAF^V600-mutant melanomas. In this study, four patients had NRAS mutant melanoma and the best response reported in this subset was stable disease while the other three developed progression at first tumor evaluation.

5.4. FAK inhibition combined with MEK inhibition

Emerging science has shown that focal adhesion kinase (FAK) inhibitors potentially overcome a tumor fibrotic barrier which leads to immune tolerance, boosting multiple treatment modalities including immunotherapy. GSK2256098 is a potent, oral, reversible inhibitor of FAK-tyrosine kinase activity measured by the autophosphorylation site (Tyr 397) of FAK. Based on preclinical evidence of synergistic growth inhibition and cell death between GSK2256098 and trametinib (an oral, small molecule inhibitor of MEK1/2) in mesothelioma cell lines. The combination of FAK and MEK inhibitors may be superior to the effect of FAK inhibitor alone or MEK inhibitor alone[142]. FAK inhibition by IN10018 displayed anti-tumor effects against a NRAS-mutated metastatic melanoma in a presented report. InxMed just initiated its Phase IB clinical trial involving its investigational FAK inhibitor IN10018 as a monotherapy and in combination with the MEKi cobimetinib (NCT04109456). The open-label, Phase IB trial will be conducted in patients with NRAS melanoma and uveal melanoma at numerous study sites in the US and Australia. Its aim is to evaluate the safety, tolerability, and activity of the therapies. The trial will enroll up to 52 patients. One of the exploratory aims of the trial is to examine other potential predictive biomarkers of response, such as the level of phosphorylated-FAK.

5.5. Pan-RAF inhibitors

The pan-RAF inhibitors as a class have moved from preclinical to clinical phases of development following years of searching for a solution to paradoxical reactivation of MAP kinase signaling secondary to first and early-generation BRAF inhibitors[143]. Their utility may carry into patients harboring NRAS driver mutations based on the common overactivation downstream of RAS signaling that occurs in both forms of tumors[144]. PLX8394 (NCT02428712) is a parodoxic breaker for BRAF ^V600-mutant cancers[145]. It is able to prevent the secondary activation of the RAS pathway that occurs with type 1 BRAF inhibitor activation of ERK; its application was widely anticipated in the field[146]. The compound was believed to be active against any alterations upstream of RAF signaling – including those found in NRAS mutant tumors and growth factor driven malignancies. It has entered phase 1 trials and initial results have not been completely reported.

Another study has been designed to define the toxicity, pharmacodynamics, and pharmacokinetics of LXH254 (pan-RAF inhibitor) in combination with other signal inhibitors. LXH254 has minimal paradoxical activation and will be combined with inhibitors of CDK4/6 (Ribociclib), MEK (Trametinib), ERK (TLL462). The study will begin accrual at the end of 2020 in only BRAF- or RAS-mutant melanomas (NCT04417621).

LY3009120 is another pan-RAF inhibitor which preclinically does not lead to paradoxical activation of MAPK signaling[147]. The primary objective of a first-in-human phase I study (NCT02014116) is to determine the recommended phase II dose (RP2D) of LY3009120 in patients with advanced cancer. Secondary objectives are to look at a subset of patients’ patients with advanced or metastatic melanoma, non-small cell lung cancer (NSCLC) and colorectal carrying BRAF or NRAS mutations and non–small cell lung cancer (NSCLC) or colorectal cancers carrying KRAS or BRAF mutations. A recently published report from this trial with 6 melanoma patients (n=5 with NRAS mutation) showed minimal clinical activity, with no patients meeting criteria for complete or partial clinical response [148].

5.6. RAF dimer inhibition

Lifirafenib (BGB-283) is a novel, first-in-class, investigational RAF dimer inhibitor with potent, reversible inhibition of wildtype ARAF, BRAF, CRAF, and BRAF^V600E as well as EGFR and KRAS[149]. Preclinical studies have suggested that lifirafenib leads to a greater number of responses in BRAF–mutated cancers than first-generation BRAF inhibitors, including vemurafenib and dabrafenib. In a very recent phase I trial that included 35 patients (10 with NSCLC, 18 with melanoma in the dose expansion phase) there was no clinical activity of this agent found in the melanoma subgroups, which included patients with NRAS mutant tumors [150]. Currently, one additional open-label phase 1b trial of BGB-283 in combination with MEK inhibition is enrolling patients to establish safety and tolerability for patients with a variety of solid tumors including melanoma (NCT03905148).

5.7. Immune-based therapeutic strategies

Immune checkpoint inhibitor therapy is the initial approach to most melanomas including all of those with mutations in their NRAS melanoma. A comprehensive review of immune checkpoint therapy is beyond the scope of this review. The major challenge faced by the field is that the majority of such patients do not respond or develop resistance to ICI after initial benefit and have few if any options outside of enrollment in a clinical trial. NRAS tumor biology may provide opportunities in combination with current therapeutic strategies based on T-cell immunity though the landscape of current clinical data does not yet paint a clear picture. Early preclinical data suggests that tumors with NRAS driver mutations have a distinct tumor microenvironment that is differentially responsive to ICI compared to that of BRAF mutant melanomas[151–153]. Additionally, trials of ICI backbone regimens including anti-PD1/PD-L1 blockade in combination with inhibitors of the MAP kinase (NCT02027961; durvalumab, dabrafenib, and trametinib) that are enrolling and testing for NRAS mutation status are beginning to report promising data on tolerability and immunologic correlates of therapy response (n=11 with NRAS-mutant status; 27.2% PR median PFS 4.9 months; 95% CI 3.0–5.5 and 5.9 months; 95% CI: 2.4–11.1 in cohorts B and C respectively)[154]. However a subsequent randomized phase III trial with atezolizumab (anti-PD-L1) and cobimetinib in RAS mutated melanoma (n=71; 32% of cobimetinib plus atezolizumab and n=104; 46.4% of pembrolizumab arms with NRAS mutation status) showed no improvement over pembrolizumab alone[155].

More prospective studies are required to bring clarity to the possible predictive value of NRAS tumor biology for patients on immunotherapy as this area of the literature is still in its infancy.Several recent retrospective studies have produced conflicting results. Multivariable analysis of a single-center retrospective study of 230 patients with anti-PD1 with or without anti-CTLA4 therapy for metastatic or unresectable melanoma showed that combination ICI was associated with improved overall survival in the NRAS (n=69, HR 0.24, p=0.003) and BRAFV600E/K (n=86, HR 0.47, p=0.024) subgroup whereas the BRAF and NRAS wild-type tumor subgroup (n=94) did not demonstrate a significant difference in overall survival[156]. Interestingly, another recent retrospective study of n=162 NRAS-mutant and BRAF wild-type and n=169 NRAS and BRAF wild-type patients found no relationship between NRAS mutation status and response to ICI when examining overall response rate, disease control rate, progression free survival, or overall survival[157].

Besides agents that act at T cell checkpoints, emerging evidence suggests that immunotherapies involving oncolytic viral systems may soon reach this patient population[158]. Combination T-VEC with BRAF inhibition and triple therapy T-VEC, and MEKi showed promise in a preclinical study in a NRASQ⁶¹BRAF^wildtype genetic background [159].

6. Conclusion

We have made great strides in understanding the role of NRAS in melanoma development since the gene was first sequenced in these tumors in 1984 – the same year that Wallace Clark proposed his multi-step model of melanoma development. Molecular alterations in NRAS occur at the first steps of melanocyte progression in normal, chronically UV-exposed skin at key codons. Aberrations in the second most common driver mutation of all sequenced melanomas carry distinct prognostic implications in addition to potentially targetable molecular pathways. Despite strong efforts, including a recently postponed multi-institutional RAS Initiative started by the National Cancer Institute, there are no successful therapeutic inhibitors of mutant NRAS like there are for other melanoma drivers including BRAF^V600E though our exploration of the molecular signaling events has led to early phase clinical trials for inhibitors of cell cycle kinases, heat shock proteins, kinases regulating the actin cytoskeleton, cell autophagy mediators, and old players such as downstream RAF, MEK, and ERK kinases. MEK inhibition alone is clearly ineffective in advanced NRAS melanoma. Human melanomas have resisted attempts to target conserved alterations in the MAPK pathway including RAS farnesylation and monotherapies that block lone RAF monomers. In the next decade of clinical research, there is hope that these early phase trials will show some success for targeted therapies in these patients. In parallel, there will undoubtedly be improvements in identifying strategies for mitigating resistance to immune based therapies and their toxicities which are currently the first line of therapeutic options for patients with advanced and metastatic disease harboring these driver mutations. Now – more than ever – the field is closer to bringing molecular-based therapies to patients harboring this “undruggable” mutation.

7. Expert opinion

After nearly four decades of research into the underlying biology and potential treatment strategies for targeting NRAS-mutant melanoma, no agent or combination of agents has been successful enough in late-stage clinical trials to secure FDA approval. We have outlined the many reasons for this slow pace of discovery – namely the complex biology of oncogenic signaling with many compensatory escape routes following blockade of RAS/RAF/MAPK as well as the difficulty of developing agents that can directly alter the activity of the elusive NRAS^Q61 substrate. The next decade of research holds great promise, and we anticipate that clinical success of direct inhibitors of KRAS^G12C in other solid tumors will translate into preclinical and clinical studies of similarly designed molecules that can inhibit NRAS^Q61 in human melanomas. There has been an explosion of FDA approvals and new trials for immune-based regimens for advanced melanomas of all genomic subtypes that has overshadowed continued research into targeted therapies for NRAS melanoma. It is likely that some of the strategies described above will advance into late stage trials and eventually demonstrate clinical benefit for this patient subgroup. All approaches must be vetted in pre-clinical testing and the results must be very compelling. The clinical trials must be designed in order to glean the most information as early as possible to allow investigators to better understand both the failures and successes. Finally, randomized designs may provide early information that favors one regimen over another regimen. With an expanding armamentarium of drugs of varying mechanisms and degree of efficacy, the field will soon face the challenge of choosing the right agent to develop and identifying biomarkers to predict response and potential toxicity prior to treatment.

Table 1.

Clinical trials relevant to patients with NRAS-mutant melanoma genetic status currently in active or recruiting phases.

Agent	Study Design	Intervention	Patient Setting	NCT ID
MEK162 (MEKi)	Phase II, Open-label	45 mg bid MEK162 BRAFV600E; 45 mg bid MEK162 NRAS mutant; 60 mg bid MEK162 BRAFV600	Locally advanced, unresectable, or metastatic malignant cutaneous melanoma harboring BRAFV600 or NRAS	NCT01320085
Trametinib (MEKi) + hydroxychloroquine	Phase Ib/II Open label	2 mg PO trametinib, dose escalation HCQ	Metastatic NRAS mutant melanoma	NCT03979651
HL-085 (MEKi)	Phase I/II open-label dose escalation	Various; dose escalation	AJCC 7 Stage III or IV histologically confirmed unresectable melanoma with NRAS mutation	NCT03973151
FCN-159 (MEKi)	Phase Ia/Ib open label	Various; dose escalation	AJCC 7 Stage III or IV histologically confirmed unresectable melanoma with NRAS mutation	NCT03932253
LXH254 (RAFi)	Phase II, randomized open-label	LXH254+LTT462, or LXH254+trametinib, or LXH254+ribociclib	Unresectable or metastatic melanoma with NRAS mutation with prior anti-PD-1/PD-L1 single agent or in combination with anti-CTLA-4 or with RAFi single agent or in combination with MEKi targeted therapy with progressive disease on previous systemic agents	NCT04417621
Trametinib (MEKi) + Dabrafenib (RAFi)	Phase Ib/II Open label	Trametinib 2 mg qd + dabrafenib 50 mg bid	AJCC Stage III, IV melanoma with BRAF V600 WT or NRAS mutant status	NCT04059224
Various; NRAS mutation matched to binimetinib	Phase II	Binimetinib	Solid tumors and hematologic malignancies with targetable mutations	NCT02465060
Lifirfenib (BGB-283) and PD-0325901 (Mirdametinib)	Phase Ib	Various; dose escalation and dose finding	Advanced/metastatic, unresectable tumors with disease progression	NCT03905148

Table 2.

Summary of selected clinical trials for agents targeting NRAS mutant melanoma with complete or partially complete data reporting and published results.

Agent	Study Design	Results	Grade 3–4 irAE	NCT ID
MEK162 vs dacarbazine	Open label phase III; randomized to MEK162 or dacarbazine	PFS 2.8 months (95% CI 2.8–3.6) in the binimetinib group and 1.5 months (1.5–1.7) in the dacarbazine group	Increased CPK (19% MEK162 vs 0% dacarbazine), hypertension (7 vs 2%), anemia (2 vs 5%)	NCT01763164
Ribociclib (LEE011) +MEK162	Phase Ib dose escalation, phase II dose expansion	Ib ORR 20.7% (95% CI: 8.0, 39.7); dose expansion phase II ORR 19.5% (95% CI: 8.8, 34.9) Disease control rate (DCR) of 70.7% (95% CI: 54.5, 83.9)	Reported in 33–76% of various phase 1b/phase 2 arms; Phase 2 common serious AE’s include GI (22%), infection (22%)	NCT01781572
Ulixertinib (BVD-523)	Open label phase I; 3+3 dose escalation with 6 dose expansion groups	RP2D of 600 mg bid determined from dose escalation n=27 dosed from 10 to 900 mg bid	Total with any G3–4 irAE: 56 of 135 (41%); rash (17%), diarrhea (7%), anemia (4%), fatigue (4%)	NCT01781429
MK-8353	Phase I dose escalation with phase 2 dose confirmation	Well tolerated up to 400 mg bid; no RP2D established; Early evidence of disease stability in patients with NRAS mutant status but no antitumor response.	Diarrhea (16%), fatigue (4%), rash (8%), nausea/vomiting (8%), increased bilirubin (16%), anemia (4%), pruritus (4%), acute renal failure (4%)	NCT01358331
LY300912 (pan-RAF)	Phase I open label part A dose escalation, part B dose confirmation	RP2D established at 300 mg bid; no CR or PR reported	30 (59%) overall; most common included stomatitis, fatigue, pain, increased ALT/AST increased bilirubin, arthralgia, myalgia, dermatitis acneiform, myalgia	NCT02014116
Lifirafenib (BGB-283)	Phase I dose expansion	RP2D established at 30 mg/d; no clinical activity found in the melanoma subgroup	Hypertension (n=23; 17.6%), fatigue (n=13; 9.9%)	NCT02610361