Uveal melanoma is the most common primary intraocular malignant tumor in adults and arises from the transformation of melanocytes in the uveal tract. Even after treatment of the primary tumor, up to 50% of patients succumb to metastatic disease. The liver is the predominant organ of metastasis. There is an important need to provide effective treatment options for advanced stage uveal melanoma. To provide the preclinical basis for new treatments, it is important to understand the molecular underpinnings of the disease. Recent genomic studies have shown that mutations within components of G protein–coupled receptor (GPCR) signaling are early events associated with approximately 98% of uveal melanomas.
Implications: This review discusses the alterations in GPCR signaling components (GNAQ and GNA11), dysregulated GPCR signaling cascades, and viable targeted therapies with the intent to provide insight into new therapeutic strategies in uveal melanoma. Mol Cancer Res; 15(5); 501–6. ©2017 AACR.
Uveal melanoma is a rare cancer with approximately 2,500 new patient diagnoses being reported per year in the United States. Primary uveal melanoma tumors are treated effectively by radiation plaque therapy or enucleation; however, approximately 50% of patients develop metastatic disease, frequently in the liver (1, 2). In some cases, metastases are found decades after successful treatment of the primary tumors and one likely explanation is the dissemination of tumor cells from the primary site followed by cellular dormancy (3). Metastatic uveal melanoma responds poorly to clinically available therapies and patients often succumb within 1 year of diagnosis of the metastases; hence, there is an urgent unmet need for effective therapeutic strategies for advanced uveal melanoma (4). There is a low mutational burden in uveal melanoma tumors, unlike cutaneous melanoma, but identification of mutations in components of G protein–coupled receptor (GPCR) signaling in uveal melanoma tumors may uncover new therapeutic targets in uveal melanoma. These components include GNAQ, GNA11, PLCB4, and CYSLTR2, mutations of which are found in approximately 98% of uveal melanoma. Here, we review the molecular alterations in GPCR pathway components and discuss the therapeutic possibilities directed at targeting GPCR signaling.
GNAQ and GNA11 mutations
GNAQ and GNA11 encode the alpha subunits of guanine nucleotide-binding proteins (G proteins), Gαq and Gα11, respectively. They form a heterotrimeric complex with β and γ subunits and are important intermediates between membrane-bound GPCRs and intracellular signaling cascades. G proteins are normally inactive when bound by guanosine diphosphate (GDP) but agonist activation of GPCRs triggers recruitment of the G proteins to the receptors where a switch from GDP to guanosine triphosphate (GTP) occurs, rendering the G proteins active to bind/stimulate proteins associated with downstream pathways. Normally, the GTPase activity of G proteins hydrolyzes GTP to GDP (5), an inactivation process that is catalyzed by the regulator of G protein signaling (RGS) proteins. GNAQ and GNA11 mutations occur in a mutually exclusive manner in approximately 93% of uveal melanoma tumors [The Cancer Genome Atlas (TCGA): GNAQ and GNA11 mutations detected in ∼50% and ∼43% of uveal melanoma tumors, respectively; ref. 6)]. They are almost exclusively found in exon 5 codon 209 (Q209) although mutations in exon 4 codon 183 (R183) have been determined in a minority of cases (7–9). Common substitutions in GNAQ are glutamine-to-leucine (Q209L) and glutamine-to-proline (Q209P), whereas in GNA11, the most frequent substitution is Q209L. Q209 is crucial for the GTPase activity of G proteins; thus, hydrolysis of GTP is abolished in GNAQ and GNA11 mutants, leading to constitutive activation of the Gαq and Gα11 proteins in uveal melanoma. Interestingly, the levels of Gαq Q209L mutant proteins may be regulated by Ric-8A, a molecular chaperone that contributes to folding of the G protein (10). Deletion of Ric-8A in GNAQ Q209L mutant melanocytes grafted into NSG mice led to a marked reduction in levels of membrane-associated mutant Gαq proteins and inhibition of GNAQ Q209L–driven tumor progression.
GNAQ and GNA11 are the most frequently mutated genes in uveal melanoma; however, their mutations do not correlate with uveal melanoma patient outcome, patient survival, or factors that would indicate high risk of metastasis (9, 11). GNAQ and GNA11 mutations occur at similar frequencies in metastasizing and nonmetastasizing tumors. Similarly, these mutations are not associated with class 1 (low metastatic potential) or class 2 (high metastatic potential) of uveal melanoma tumors (11, 12). It has been shown that the Q209 mutation in GNAQ and GNA11 are found in benign nevi such as blue nevi in addition to primary and metastatic uveal melanoma tumors (7, 8). Particularly, GNAQ is frequently mutated in blue nevi (∼83%). This suggests that mutations in the G proteins are early events in uveal melanoma development. Despite this notion, the GNA11 Q209 mutation is more commonly identified in uveal melanoma metastases (∼57%) and found only in approximately 7% of benign blue nevi, indicating that in comparison with GNAQ, alteration in GNA11 is associated with higher risk of metastasis of uveal melanoma (7, 8, 13).
Effector and receptor mutations: PLCB4 and CYSLTR2
In addition to the GNAQ and GNA11 mutations, other, less common, driver mutations in uveal melanoma have been identified recently by next-generation sequencing. Mutations in phospholipase C β4 (PLCB4) and cysteinyl leukotriene receptor 2 (CYSLTR2) have been identified in 1% and 4% of uveal melanoma tumors, respectively (14, 15) and are mutually exclusive with GNAQ and GNA11 mutations. The alteration in PLCβ4, D630Y, affects the Y-domain of the highly conserved catalytic core of PLCβ4 which controls signal transduction (14). Consistent with this finding, PLCβ4 interacts directly with Gαq (ref. 16; Fig. 1). PLCβ4 mediates signal transduction by catalyzing the conversion of phosphatidylinositol 4,5-bisphosphate (PIP2) in the plasma membrane into inositol 1,4,5-trisphosphate (IP3) and diacylglycerol (DAG). Subsequently, DAG and IP3 activate downstream signaling components such as protein kinase C (PKC). IP3, particularly, translocates into the cytosol where it induces the release of calcium (Ca2+) from the endoplasmic reticulum to activate PKC.
The other alteration in the GPCR pathway in uveal melanoma is at the level of GPCRs. Out of 136 uveal melanoma patient specimens analyzed by Moore and colleagues, 4 samples harbored a leucine to glutamine substitution at codon 129 (Leu129Gln) in the CYSLTR2 gene, and all four samples lacked mutations in GNAQ, GNA11, or PLCB4 (15). CYSLTR2 encodes a seven-transmembrane GPCR that activates Gαq, a finding that is consistent with the mutual exclusive profile of the mutations (17) (Fig. 1). Activation of Gαq by CysLT2R promotes binding of Gαq/Gα11 to PLCβ4. The Leu129Gln mutation is located in the third transmembrane helix of the receptor and promotes ligand-independent activation of the GPCR. Expression of Leu129Gln CYSLTR2 in HEK293 cells increased basal levels of calcium and promoted the growth of melanocyte cell lines in vitro and in vivo (15). Collectively, these findings raise the possibility of targeting mutant forms of GNAQ/GNA11, PLCB4, and CYSLTR2 in uveal melanomas as well as pathways downstream of mutant GNAQ/GNA11.
Dysregulated pathways downstream of mutant GNAQ/GNA11
As the identification of mutations in GPCR signaling components in a high proportion of uveal melanoma tumors, molecular understanding of pathways downstream of Gαq and Gα11 has become crucial for development or discovery of effective treatment options for metastatic uveal melanoma. These pathways include the ERK1/2, Rho/Rac/YAP and PI3K/AKT pathways.
The ERK1/2 pathway usually involves binding of ligands to tyrosine kinase receptors on the cell membrane and activation of downstream intermediates, RAS, RAF, and MEK. MEK phosphorylates and activates ERK1/2, which subsequently undergoes dimerization and translocates into the nucleus to regulate cellular processes including proliferation, survival, differentiation, and apoptosis (18). The ERK1/2 pathway is frequently activated in uveal melanoma with 86% of primary uveal melanoma tumors reported to have elevated ERK1/2 phosphorylation (19). However, unlike cutaneous melanoma, in which BRAF is commonly mutated, mutations in RAS and BRAF are rare in uveal melanoma tumors (20). Only one patient with choroidal melanoma has been shown to harbor the BRAF V600E mutation (21). MEK–ERK1/2 activation in uveal melanoma is expected to be induced rather by mutant Gαq/Gα11 proteins. The G proteins activate PLCβ and PKC, which then stimulates MEK/ERK1/2 (ref. 18; Fig. 1). Transfection of GNAQ Q209L in human melanocytes enhanced phosphorylated ERK1/2 protein levels and was associated with increased anchorage-independent growth (7). Consistently, these results were reversed by siRNA-mediated knockdown of GNAQ which decreased phospho-ERK1/2 levels (7). However, it is noteworthy that in some uveal melanoma cases, ERK1/2 may not be activated by G proteins as a study of 22 uveal melanoma patient tumors did not observe a correlation between GNAQ mutation and ERK1/2 activation although samples with GNAQ mutations had a higher average of the total ERK1/2 expression level compared with GNAQ wild-type tumors (22).
A major role of Gαq/11 is to directly bind and activate PLCβ, but additional effectors and signaling pathways may also play an important role. Particularly relevant to uveal melanoma, Gαq/11 stimulates Rho and Rac small GTPase-mediated signaling through direct binding and activation of several members of the large Rho guanine-nucleotide exchange factor (RhoGEF) family, including p63RhoGEF and Trio (refs. 23, 24; Fig. 1). Trio appears to be a key player in mediating mitogenic signals in uveal melanoma as knockdown of Trio inhibited tumor growth and DNA synthesis in two uveal melanoma cell line models (25). Interestingly, Trio is a very large multi-domain protein that has two GEF domains, one that selectively activates RhoA and one that activates Rac1/RhoG. Notably, activation of both RhoA and Rac1 by Trio are required for Gαq/11 mitogenic signaling (25, 26).
Multiple pathways downstream of Rho/Rac are likely to mediate Trio-dependent cell proliferation in mutant Gαq/11 uveal melanoma cells. Depletion of Trio by siRNA did not inhibit ERK1/2 activation, but instead blocked activation of the MAPKs JNK and p38; JNK and p38 can both regulate the transcription factor AP-1 which controls the transcription of a number of growth-promoting genes (25). In addition, two reports strongly implicate two transcriptional coactivators in the Hippo pathway, Yes-associated protein (YAP) and transcriptional coactivator with PDZ-binding motif (TAZ), as being critical mediators of oncogenic Gαq/11 downstream of Rho/Rac (27, 28). These reports showed that mutationally activated Gαq/11 induced the cytoplasm to nucleus translocation of YAP/TAZ and transcription of several YAP/TAZ-dependent genes that promote cell proliferation (27, 28). Moreover, siRNA depletion of YAP or inhibition of YAP with the pharmacologic inhibitor verteporfin suppressed the growth of uveal melanoma cells in culture and in a mouse model (27, 28). Robust nuclear localization of YAP was also observed in uveal melanocytes in a recently described zebrafish model for uveal melanoma (29). The mechanisms underlying Rho and Rac activation of YAP appear to involve both inhibition of upstream kinases, large tumor suppressor homolog 1 and 2 (LATS1/2), as well as increased sequestration to F-actin of the YAP-binding protein angiomotin (AMOT), resulting in both cases in increased levels of YAP available to translocate to the nucleus (27, 28, 30). The exciting implication of the above studies is that the Trio/Rho/Rac/YAP pathway of oncogenic Gαq/11 signaling suggests numerous new therapeutic targets in uveal melanoma.
ADP-ribosylation factor 6 (ARF6)
Another monomeric small GTPase, ARF6, has been proposed as a key mediator of most pathways activated by oncogenic Gαq/11 (31). Inhibition or depletion of ARF6 in uveal melanoma cells inhibited cell proliferation and the downstream signaling targets PLCβ, ERK1/2, Rho, Rac and YAP. Activation of ARF6 by Gαq/11 appears to be mediated by direct binding of Gαq/11 to the ARF-GEF, GEP100. A provocative mechanism for the role of ARF6 in controlling Gαq/11 signaling was proposed, in which ARF6 promoted localization of mutationally activated Gαq/11 to cytoplasmic vesicles, rather than the plasma membrane, and these cytoplasmic vesicles are the site of oncogenic signaling by Gαq/11 (ref. 31; Fig. 1). Regardless of the exact mechanism of how ARF6 regulates Gαq/11 signaling, ARF6 provides another potential therapeutic target in uveal melanoma.
Phosphatidylinositol (4,5)-bisphosphate 3-kinase (PI3K) catalyzes the formation of phosphatidylinositol (3,4,5)-trisphosphate (PIP3) from PIP2, and PIP3 induces membrane translocation of AKT where it becomes active and signals to promote cell proliferation and survival (18). PI3K signaling is negatively regulated by PTEN, which reverses PIP2 conversion to PIP3. Complete loss of PTEN is identified in 16% of uveal melanoma tumors (n = 75) and weaker immunostaining of PTEN is found in 42.7% of uveal melanoma tumors (32). Patients with PTEN-null tumors are associated with a shorter disease-free survival compared with patients with uveal melanoma tumors expressing normal levels of PTEN (32).
Targeted therapies in uveal melanoma
Patients with metastatic uveal melanoma typically die within one year of diagnosis; hence, there is an urgent unmet need to identify novel approaches for treatment which includes targeted therapies as either monotherapy or in combinational strategies.
As activating mutations in either CYSLTR2, GNAQ, GNA11, or PLCB4 are found in approximately 98% of uveal melanomas, this signaling pathway represents a viable therapeutic target for the treatment of uveal melanoma. One approach to targeting this pathway is the development of specific inhibitors to individual proteins in the pathway. For example, constitutively active GPCRs such as the CysLT2R-L129Q mutant found in uveal melanoma (15) could potentially be targeted using a receptor-specific inverse agonist, which would bind to and stabilize an inactive conformation of the receptor that should no longer activate Gαq/11. Unfortunately, CysLT2R-L129Q is only a viable target in approximately 4% of uveal melanoma patients. A more compelling target is the activated G protein, either Gαq or Gα11, which together are mutated in approximately 93% of uveal melanoma patients. In this regard, there are two potent and specific inhibitors that have been reported for Gαq, YM-254890, and FR900359 or UBO-QIC. YM-254890 is a cyclic depsipeptide isolated from Chromobacerium sp. broth and showed inhibition of Gαq signaling-mediated processes such as ADP-induced platelet aggregation and Ca2+ mobilization through a Gαq-coupled receptor (33). A crystal structure of the Gαq protein, loaded with GDP, in complex with YM-254890 suggests the inhibitor works by allosterically stabilizing the protein in the GDP-bound form, preventing GDP dissociation, and locking it in the inactive state (34). More recently, FR900359, a slightly different cyclic depsipeptide isolated from the Ardisia crenatasims plant, has shown similar pharmacologic activity and specificity against Gαq (35). Docking models and molecular dynamics studies suggest FR900359 inhibits Gαq in an identical manner to YM-254890 (36). While these compounds appear to be potent and selective inhibitors of Gαq (they have not been tested on Gα11), it is unclear whether they would be able to inhibit the constitutively active forms of Gαq typically found in uveal melanoma. For example, while YM-254890 inhibited a Gαq-R183C mutant, it was unable to inhibit Gαq-Q209L (33). Moreover, one also needs to consider Gαq versus Gα11 selectivity when inhibiting these proteins as a nonselective Gαq/11 inhibitor would likely be toxic, as knockout of both Gαq and Gα11 in mice leads to death in utero (37). Finally, one could consider targeting PLCβ4 with a selective inhibitor as this is mutated in approximately 1% of uveal melanoma and is a downstream effector of Gαq/11. However, PLCβ4 is unlikely to be the only downstream target of activated Gαq/11 in uveal melanoma so such an inhibitor might only prove useful in the small number of patients that have a PLCβ4 mutation.
Because of the high incidence of MEK–ERK1/2 activation in uveal melanoma, targeting of components of the Gαq/Gα11–induced ERK1/2 cascade, such as MEK and PKC, has been investigated as a therapeutic approach in uveal melanoma. Inhibition of MEK1/2 either by trametinib, selumetinib, or PD0325901 induced growth arrest and apoptosis in GNAQ/GNA11–mutant uveal melanoma cell lines and xenograft tumors (38–41). Trametinib is FDA-approved and is used in combination with the BRAF inhibitor, dabrafenib, for patients with unresectable or metastatic melanoma harboring BRAF V600E and V600K mutations (42). However, reports from clinical studies indicated variable outcomes for MEK inhibition in uveal melanoma. In a phase I trial involving 16 metastatic uveal melanoma patients, trametinib was shown to have limited efficacy (43). A phase II trial which enrolled 120 advanced uveal melanoma patients showed that selumetinib improved the median progression-free survival (PFS) by approximately 9 weeks compared with chemotherapy (dacarbazine or temozolomide) but only modestly increased median overall survival (44). In the most recent phase III clinical trial (SUMIT), the combination of selumetinib and dacarbazine failed to improve PFS compared with chemotherapy alone and was associated with a low response rate (3–4%; ref. 45). Thus, while preclinical studies support MEK inhibitors as part of a therapeutic approach for metastatic uveal melanoma, it is important to identify agents that in combination enhance the responses of uveal melanoma to MEK inhibition.
The poor clinical response of metastatic uveal melanoma to MEK inhibitors may be in part due to factors produced by the tumor microenvironment. Hepatocyte growth factor (HGF) provided resistance to trametinib growth-inhibitory effects in metastatic uveal melanoma cell lines and targeting of cMET (the receptor for HGF) enhanced the effects of trametinib in metastatic uveal melanoma (40, 46). HGF-CMET signaling induced downregulation of BH3 proapoptotic proteins, BIM and BMF, which correlated with HGF-mediated inhibition of apoptosis in trametinib-treated cultures (46). Importantly, HGF is secreted by quiescent hepatic stellate cells that can be found in the liver microenvironment and phosphorylated/activated cMET is detected in the majority of uveal melanoma liver metastases (40, 46). HGF promotes activation of the PI3K/AKT pathway as a mechanism of resistance to trametinib and the increase in phosphorylation of AKT is mediated by PI3K isoforms α, δ, and γ. Use of a β-sparing PI3K inhibitor, GDC0032, reversed HGF-induced resistance to trametinib (46). These findings support an earlier study which reported the combination of MEK and PI3K inhibitors for uveal melanoma and these agents induce marked early apoptosis in GNAQ-mutant uveal melanoma cell lines compared with the inhibitors alone which have moderate effects on apoptosis (40, 41).
MEK inhibitors have also been tested in combination with PKC inhibitors for uveal melanoma. Inhibition of PKC alone such as by AEB071 or sotrastaurin induced cell-cycle G1 phase arrest and suppressed PKC and ERK1/2 signaling (39). However, these effects were not durable, whereas in combination with MEK inhibitors, PD0325901, or MEK162, sustained inhibition of the ERK1/2 pathway was observed (39). Furthermore, the inhibitors caused a strong synergistic growth-inhibitory effect on uveal melanoma cell lines and marked tumor regression in uveal melanoma xenograft models. Recently, screening of drug combinations involving AEB701 in a large panel of uveal melanoma patient-derived xenografts (PDX) also determined two non-MEK inhibitors that could potentially be used in combination with PKC inhibitors as therapeutic strategies for metastatic uveal melanoma (47). These were the p53-MDM2 inhibitor, CGM097, and the mTORC1 inhibitor, RAD001. AEB701 and CGM097 or AEB701 and RAD001 markedly reduced tumor growth and induced tumor regression (47). These studies provide evidence for a number of targeted therapies that may be evaluated in combination with MEK inhibitors or PKC inhibitors in clinical studies to improve the survival of patients with metastatic uveal melanoma.
There is a lack of effective treatment options for advanced uveal melanoma despite success in the treatment of primary uveal melanoma tumors. While identification of driver mutations in uveal melanoma, such as GNAQ and GNA11, has shed light on potential therapeutic targets, alterations in GPCR signaling occur early in disease progression. Other mutations and their impact on responses of uveal melanoma to GPCR-targeted therapies will need to be explored (48). For example, inactivating mutations of BAP1 have been found in approximately 80% of aggressive uveal melanoma (49) while SF3B1 and EIF1AX were shown to correlate with a favorable prognosis (50). In addition, ongoing investigations will be needed to devise strategies that elicit favorable treatment outcomes using therapies targeting GPCR signaling. As in other forms of advanced-stage melanoma, it is likely that combination approaches will be key to elicit high response rates and the most effective combinations and scheduling of inhibitors will need to be identified.
Disclosure of Potential Conflicts of Interest
A.E. Aplin reports receiving a commercial research grant from Pfizer. No potential conflicts of interest were disclosed by the other authors.
Uveal melanoma research in the Aplin, Benovic, and Wedegaertner laboratories is collaboratively funded by a Dr. Ralph and Marian Falk Medical Research Trust Catalyst Award. In addition, uveal melanoma research in the Aplin lab is supported by a Melanoma Research Alliance Team Science Award, a Cure Ocular Melanoma/Melanoma Research Foundation Established Investigator Award and NIH R01 (CA182635). The Cancer Cell Biology and Signaling (CCBS) program is supported by the Cancer Center Support Grant 5P30CA056036-17. The Wedegaertner laboratory is also funded by NIH/NCI R03 CA202316-01 and a diversity supplement for NIH R01 GM56444-17 while the Benovic laboratory is also supported by NIH P01 HL114471-04.