Uveal Melanoma and GNA11 Mutations: a new piece added to the puzzle

Uveal melanoma is the most common primary malignant intraocular tumor of adults, with an age-adjusted incidence of 4.3 per million/yr in the United States (Singh and Topham, 2003). Uveal melanoma can originate from neural crest-derived melanocytes of the iris (3%), ciliary body (7%), and choroid (90%) (Landreville et al., 2008). Approximately half of all primary uveal melanomas metastasize through hematogenous dissemination, primarily to the liver and less frequently to extrahepatic organs, such as lung, bone, and skin. Metastatic disease only rarely responds to treatment and is usually fatal within 6–9 months of the onset of symptoms. Although various host factors have been indicated as modifiers of risk for uveal melanoma development, e.g. congenital ocular melanocytosis (nevus of ‘Ota’ when the eyelid is affected), uveal nevus, light color eye, fair color skin, and inability to tan (Landreville et al., 2008), environmental factors rendering predisposition to this disease remain mostly elusive (Singh and Topham, 2003). A prime example of a controversial risk determinant for uveal melanomagenesis is occupational/recreational sunlight ultraviolet (UV) exposure. Genetic features associated with the pathogenesis and metastasis of uveal melanoma include monosomy 3, gain of chromosome 8q and chromosome 6p, and loss of chromosome 8p. Gene-expression profiling of primary uveal melanomas has revealed two major transcriptomic subgroups with prognostic significance, including class 1 tumors (competent for melanocytic differentiation and incompetent for metastasis) and class 2 tumors (incompetent for melanocytic differentiation and competent for metastasis). In contrast to cutaneous melanomas, uveal melanomas do not contain mutations in BRAF or NRAS (Landreville et al., 2008), suggesting that melanocyte precursors are different for subgroups of melanocytic tumors and/or that different molecular events are involved in their transformation. GNAQ and GNA11 encode members of the q class of heterotrimeric G-protein α subunits (Gαq and Gα11, respectively) involved in mediating signals between G-protein-coupled receptors and downstream effectors, such as protein kinases A and C (Neves et al., 2002). Stimulation of the α subunit of G-proteins by an activated G-protein-coupled receptor renders these molecules bound to guanosine triphosphate (GTP), which, in turn, leads to a cascade of signaling events. Mutations of GNAQ and GNA11 that cause substitutions of specific glutamine or arginine residues in the respectively coded proteins, which contact the GTP molecules, can result in blocking of the intrinsic GTPase activity and constitutive G-protein activation (Neves et al., 2002). Recently, Van Raamsdonk et al. (Van Raamsdonk et al., 2009) investigated the mutation status of GNAQ and GNA11 in a wide range of benign and malignant melanocytic neoplasms by sequencing the coding regions of the respective genes. They found somatic mutations in GNAQ, but not in GNA11, in 83% of blue nevi, 50% of malignant blue nevi, and 46% of uveal melanomas. These mutations occurred exclusively in codon 209 of GNAQ with the majority predicting substitution of glutamine to leucine (Q209L), and to proline (Q209P) (66% and 30%, respectively). In vitro, transfection of hTERT/CDK4R24C/p53DD immortalized melanocytes with GNAQQ209L induced anchorage-independent growth, and in vivo, tumorigenicity experiments with nude mice transfected with GNAQQ209L melanocytes showed formation of heavily pigmented tumors at the injection site. GNAQQ209L induced mitogen-activated protein kinase (MAPK) activation in primary normal human melanocytes, hTERT/CDK4R24C/p53DD melanocytes, and 293T cells, whereas siRNA-mediated knockdown of GNAQ in the GNAQQ209L mutant uveal melanoma cell lines OMM1.3 and Mel202 resulted in MAPK inhibition, apoptosis, and reduced ability to grow in soft agar. The overall results confirmed that somatic mutations occurring exclusively in codon 209 of GNAQ cause constitutive activation of this G-protein, thus, turning this protein into a dominantly acting oncogene (Van Raamsdonk et al., 2009). In a new study published in the December second issue of ‘The New England Journal of Medicine’, Van Raamsdonk et al. extended their investigations by reexamining the mutation frequency and pattern of GNA11, which is 91% identical to GNAQ at the amino acid level, in a broad spectrum of melanocytic neoplasias. Here, they report a progressively increased frequency of mutations in codon 209 of GNA11 from 6.5% in blue nevi, to 31.9% in primary uveal melanomas, to 56.5% in uveal melanoma metastases. The frequency of codon 209 mutations in GNAQ in the respective tissues, however, followed a reverse pattern, including 54.7% in blue nevi, 44.8% in primary uveal melanomas, and 21.7% in uveal melanoma metastases; this pattern was significantly different from that found in GNA11 (P < 0.001). Similarly to GNAQ mutations (Van Raamsdonk et al., 2009), the GNA11 mutations in codon 209 were exclusively predictive of glutamine substitution by leucine (Q209L), and by proline (Q209P) (97.3% and 2.7%, respectively). These mutations occurred as a result of single- or tandem-base substitutions in codon 209 (CAG) as follows: A→T (94.5%), A→C (2.7%), AG→TA (1.4%), and AG→TT (1.4%) (see, Fig. 1). Somatic mutations in GNA11 and GNAQ in uveal melanocytic neoplasms, and in BRAF in cutaneous melanomas. Potential sites of mutations in codons 209 and 183 in GNA11 and GNAQ, and in codon 600 in BRAF are indicated. The red-color font indicates the mutated codons. Base substitutions and the resulting amino acid changes are shown. WT, wild type. Furthermore, there were mutations in codon 183 (CGC) of GNA11 predicting substitution of arginine to cysteine (R183C) or to histidine (R183H) in 1% of blue nevi, 2.1% of primary uveal melanomas, and 5.9% of uveal melanoma metastases. All but one of these mutations were caused by C→T transitions in codon 183 and CC→TT transitions in codons 182–183 of GNA11; the remainder being caused by a single G→A transition in codon 183, which would appear as a C→T transition on the reverse DNA strand. Likewise, 1% of blue nevi, 2.8% of primary uveal melanomas, and 5.9% of uveal melanoma metastases harbored mutations in codon 183 (CGA) of GNAQ, which were predictive of arginine-to-glutamine substitution (R183Q). The mutations affecting codon 183 of GNAQ were exclusively caused by G→A transitions. Together, all mutations occurring at codon 183 of both GNA11 and GNAQ were cytosine-to-thymine transitions either on the forward or on the reverse DNA strand (see, Fig. 1). These cytosine-to-thymine transitions are the hallmark of mutations induced by the UVB (λ = 280–320 nm) component of sunlight (Besaratinia and Pfeifer, 2008). Whereas the sequence contexts of codon 183 in GNA11 and GNAQ genes allow the signature mutations of sunlight UV to manifest phenotypically, the sequence compositions of codon 209 in both genes prevent phenotypic expression of characteristic UVB-induced mutations, as C→T transitions occurring in this codon in GNA11 or GNAQ would result in stop codons or silent mutations. Overall, GNA11 and GNAQ mutations in uveal melanomas are reminiscent of BRAF mutations in cutaneous melanomas (Besaratinia and Pfeifer, 2008). Up to 70% of cutaneous melanomas harbor mutations in the BRAF oncogene, predominantly in or around codon 600 of exon 15, which are predictive of valine substitutions (BRAFV600) (Besaratinia and Pfeifer, 2008). Uniquely, 29% of these mutations are tandem-base substitutions on the same allele, an extremely rare type of mutation in non-melanocytic tumors (see, Fig. 1). However, BRAFV600 variants in cutaneous melanomas do not bear the signature mutations of sunlight UVB (i.e. single or tandem cytosine-to-thymine transitions), despite the fact that sunlight UV irradiation has been implicated in the etiology of human skin melanoma (Besaratinia and Pfeifer, 2008). Scrutiny of the mutation spectra of GNA11 and GNAQ in uveal melanomas reveals two distinct types of events: (i) UV-specific mutations at codon 182–183 and (ii) selection-driven mutations at codon 209; the latter having resemblance to BRAFV600 mutations in cutaneous melanomas. More specifically, in all cases, mutations involve a CpA dinucleotide and tandem mutations are seen at this dinucleotide (Besaratinia and Pfeifer, 2008; Van Raamsdonk et al., 2009), suggesting that there is an unidentified dimeric photoproduct at CpA or TpG (other strand), perhaps related to melanin or tyrosine photochemistry in uveal and cutaneous melanocytes. In the new study by Van Raamsdonk et al., except for one single blue nevus that harbored a concomitant GNA11Q209L and GNAQR183H mutations, all other samples carried mutations at codons 183 and 209 on a mutually exclusive basis. In total, 63.2% of all blue nevi, 81.6% of all primary uveal melanomas, and 90.0% of all uveal melanoma metastases had oncogenic mutations in either codon 209 or codon 183 of GNA11 or GNAQ. Of significance, the frequency of GNA11 mutations in primary uveal melanomas arising in the ciliochoroidal region was significantly elevated when compared to tumors originating from other anatomical locations (P = 0.048). From a prognostic standpoint, uveal melanomas stemming from the ciliochoroidal region have a poor outcome (Landreville et al., 2008). Nonetheless, there was no statistically significant difference in overall survival or disease-free survival between patients whose tumors harbored GNA11 mutations and those who had GNAQ-mutated tumors and those who had no mutation in either gene. However, the small number of subjects (n = 81) available for this analysis may have prevented eliciting a statistically significant result. Tumorigenicity experiments in immunocompromised mice injected with melan-a-cells transduced with either GNA11Q209L or GNA11R183C variants showed decreased latency and increased penetrance for the former variant. The dominant oncogenicity of the GNA11Q209L variant manifested as rapidly growing tumors that developed at each of the six injection sites for the mice injected with this variant, when compared to latent tumors, which developed in only three of eight injection sites for the mice injected with the GNA11R183C variant. In addition, all mice injected with cells transduced with GNA11Q209L variant, but not with GNAQQ209L variant, developed metastatic tumors in lung and liver. The absolute propensity of the GNA11Q209L variant for metastasis is in agreement with the progressive frequency of mutations in codon 209 of GNA11 in blue nevi, to primary uveal melanomas, to uveal melanoma metastases. In view of these findings, it is tempting to speculate that GNA11 mutations play a superseding role over GNAQ mutations in the genesis and metastasis of uveal melanoma. Finally, Western blot analysis of melan-a-cells transduced with the GNA11Q209L variant showed activation of the MAPK pathway, as indicated by increased levels of phosphorylated extracellular signal-regulated kinase (ERK). Although GNAQQ209L mutant cell lines are highly sensitive to inhibitors of MAPK/ERK-activating kinase (MEK; a component of the MAPK pathway), the unavailability of uveal melanoma cell lines with GNA11 mutation precluded verification of a similar sensitivity of the GNA11Q209L mutant cells to MEK inhibitors. Given the close functionalities of Gαq and Gα11, it is likely, however, that the activation of the MAPK pathway consequent to GNA11Q209 mutations in uveal melanoma cells can be counteracted by treatment with MEK inhibitors. Altogether, the prevalence of GNA11 and GNAQ mutations in uveal melanomas suggests that future therapeutic strategies should center on targeting either the mutated Gα11 and Gαq proteins themselves or the downstream oncogenic-signaling pathway(s) affected by mutations in the GNAQ and GNA11 genes.