ANTICANCER RESEARCH 37 : 5719-5726 (2017) doi:10.21873/anticanres.12010

Uveal Melanoma: GNAQ and GNA11 Mutations in a Greek Population FILIPPOS PSINAKIS 1, ANASTASIA KATSELI 2, CHRYSSANTHY KOUTSANDREA 1, KONSTANTINA FRANGIA 3, LINA FLORENTIN 2, DESPINA APOSTOLOPOULOU 2, KONSTANTINA DIMAKOPOULOU 4, DIMITRIOS PAPAKONSTANTINOU 5, ELENI GEORGOPOULOU 6 and DIMITRIOS BROUZAS 1

11st Department of Ophthalmology, National and Kapodistrian University of Athens, Athens, Greece; 2ALFA LAB, Molecular Biology and Cytogenetics Center, Leto Maternity Hospital, Athens, Greece; 3HistoBio Diagnosis Pathology Center, Athens, Greece; 4Department of Hygiene, Epidemiology and Medical Statistics, Medical School, National and Kapodistrian University of Athens, Athens, Greece; 5Department of Opthalmology, University of Athens, Georgios Gennimatas General Hospital, Athens, Greece; 6Private Practice, Athens, Greece

Abstract. Background/Aim: Uveal melanoma is the most Uveal melanoma (UM) is the most common primary common primary adult intraocular malignancy. It is known malignancy of the eye, arising from melanocytes of the to have a strong metastatic potential, fatal for the vast choroid, ciliary body and iris. The disruption of specific majority of patients. In recent years, meticulous cytogenetic signaling pathways is considered to be involved in its and molecular profiling has led to precise prognostication, tumorigenesis. One of the main known pathways is mitogen- that unfortunately is not matched by advancements in activated kinase (MAPK)/ERK, known to be adjuvant therapies. subunits alpha Q (GNAQ) and disturbed as a result of G protein subunit alpha Q ( GNAQ ) alpha 11 (GNA11) are two of the major driver that or subunit alpha 11 ( GNA11 ) mutations (1). These mutations contribute to the development of uveal melanoma. appear to play a major role in the development of UM. Understanding their prognostic significance can allow GNAQ and GNA11 are proto oncogenes that encode tailored management and facilitate their use in the on-going closely related Gq alpha subunits of the heterotrimeric Gq quest of targeted uveal melanoma therapies. Materials and protein that activates , leading to a series of Methods: Formalin-fixed, paraffin-embedded specimens were downstream signaling effects, one of which is the activation obtained from 47 patients of Greek origin, with uveal of the MAPK growth signaling pathway. Activation of Gq is melanoma. GNAQ and GNA11 genes were screened for normally terminated by a GTPase activity intrinsic to the G- mutations in exons 4 and 5, by polymerase chain reaction alpha subunit. However, mutations that occur in UM at and Sanger sequencing. Results: The overall mutation amino acid residues glutamine-209 and arginine-183 disable frequency of GNAQ/GNA11 genes was 42.4%. A novel the GTPase activity and prevent inactivation of the protein, mutation c.625_626delinsGC was identified in GNA11. No leading to constitutive activation of the MAPK pathway and correlation was observed between the mutation status and other growth pathways (2). metastasis occurrence or overall survival time of patients. The GNAQ and GNA11 mutations at glutamine-209 and Conclusion: Mutations in GNAQ and GNA11 genes in this arginine-183 are somatic and mutually exclusive (3), with Greek population present frequencies that qualify them as glutamine-209 mutations being much more frequent (4). potential targets for customized therapy. They are found in benign uveal nevi and in the vast majority of UM, with a reported frequency of 80-88% (3, 5). This suggests that perhaps they are initiating events and are not sufficient for full malignant transformation (6). Correspondence to: Anastasia Katseli, ALFA LAB, Molecular Discovering these mutations is part of the major recent Biology and Cytogenetics Center, Leto Maternity Hospital, Athens, Greece. Tel: +30 6948858103, Fax: +30 2106902083, e-mail: advances in molecular genetics that have enabled the gradual [email protected] construction of a detailed molecular landscape of UM. Better understanding of underlying genetic and molecular Key Words: Uveal melanoma, GNAQ, GNA11, Sanger-sequencing. abnormalities implicated in the development and progression

5719 ANTICANCER RESEARCH 37 : 5719-5726 (2017)

Table I. Clinicopathological data of the patients included in this study.

Parameter Value

Total patients 33 Age at enucleation, years 62±12.7 (39-88) Male/female 15 (45.45%)/18 (54.55%) Histopathological cell type/cellular type Spindle 5 (15.15%) Mixed 23 (69.7%) Epithelioid 2 (6.06%) Necrotic 2 (6.06%) Unknown 1 (3.03%) Tumor location Figure 1. Polymerase chain reaction products of exons 4 and 5 of G Chorioid 28 (84.85%) protein subunit alpha 11 (GNA11) and alpha Q (GNAQ) genes, after Iris 1 (3.03%) visualization on ethidium bromide-stained agarose gel. Unknown 4 (12.12%) Metastasis (liver) 8 (24.24) Unknown 3 (9.1%) Survival time post-enucleation, months 48±30.4 each ). All PCR primers (Table II) were designed to include an Data are the mean±SD (range) or n (%). M13 sequence tail to simplify the sequencing reaction setup. PCR reactions for each target were performed with 50-200 ng of genomic DNA as starting material in a 30 μl reaction mix with 500 μM dNTPs (Fermentas-Thermo Scientific Life Science, Milan, of UM provides a great opportunity for improved patient Italy), 0.35 μM primers, 2 mM MgCl 2, 1X Buffer and 2.5 U management and the development of targeted therapies. The Qiagen HotStarTaq DNA Polymerase (Qiagen, Hilden, Germany) estimation of the presence of GNAQ and GNA11 mutations on a Veriti 96-well Thermal Cycler (Applied Biosystems, Life Technologies Corporation, Foster City, CA, USA). Thermal cycling in Greek patients with UM will hopefully contribute to this conditions were as follows: 95˚C for 5 min, followed by 40 cycles ongoing effort. of 95˚C for 40 s, 60˚C for 40 s, 72˚C for 60 s, and a final extension at 72˚C for 10 min. Gel electrophoresis followed in order to ensure Materials and Methods the presence of amplicons in each reaction (Figure 1). Nucleofast96 PCR Plates (Macherey-Nagel, Düren, Germany) were used to This study was conducted at the First Department of Ophthalmology purify the amplification products. of the National and Kapodistrian University of Athens in association with ALFA LAB (Molecular Biology & Cytogenetics Center) and DNA sequencing. Mutational screening was carried out by direct HistoBio Diagnosis Pathology Center. In total, 47 tumor samples sequencing of fragments obtained by PCR using an ABI3130xl or were collected between 1996 and 2016 from patients of Greek ABI3500 Genetic Analyzer (Applied Biosystems, Life Technologies origin with UM, arising mainly from the choroid (Table I). Written Corporation). Sequencing data were analyzed using Sequencing informed consent was obtained from all patients and this study was Analysis v5.3.1 or SeqA6 software, respectively, (Applied approved by the Medical Ethics Committee of the University of Biosystems, Life Technologies Corporation) and Mutation Surveyor Athens (8649/9-4-2012). software (SoftGenetics, State College, PA, USA). Big Dye Terminator v1.1 Cycle Sequencing Kit (Applied Tumor tissue samples and DNA extraction. UM tissue specimens were Biosystems, Life Technologies Corporation) was used to set-up the formalin-fixed and embedded in paraffin blocks. A pathologist marked sequencing reactions, according to the manufacturer’s instructions. the most dense tumor area, which was subsequently macrodissected. The purified PCR products were used as templates. The sequencing These specimens were carefully selected to contain 80-100% neoplastic reaction was performed using the M13 forward and reverse primer cells. Paraffin-embedded tissue specimens (two 5- μm-thick sections cut sequences and the following program on the thermocycler: 30 cycles from each paraffin block) were retrieved from the Pathology of a 10-s denaturation at 96˚C, annealing of the primer at 50˚C for 5 Department of HistoBio Diagnosis Laboratory. DNA was extracted s, and the extension step at 60˚C for 4 min. The sequencing products from formalin-fixed, paraffin-embedded (FFPE) tissue sections, using were purified before capillary electrophoresis to remove MagCore Genomic DNA-FFPE One-Step Kit (RBC Bioscience, New unincorporated ddNTPs using Sephadex columns (Illustra Sephadex Taipei City, Taiwan) with an Automated Nucleic Acid Extractor G-50/fine DNA grade; GE Healthcare, Chicago, IL, USA). MagCore HF16 (RBC Bioscience, New Taipei City, Taiwan). DNA concentration and quality were evaluated using a Nanodrop2000 Statistical analysis. The distribution of all continuous and categorical spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). variables is described by reporting the mean (standard deviation; SD) values; range and frequency and percentage, respectively. The Amplification step. Polymerase chain reactions (PCR) were set-up and samples were analyzed all together and were further analyzed after optimized for the two genes on 9, GNAQ (NM_002072) being subdivided into two groups depending on the storage and GNA11 (NM_002067) and four target sequences (exons 4, 5 of preservation time of FFPE specimens:group A included samples

5720 Psinakis et al : GNAQ and GNA11 Mutations in a Greek Uveal Melanoma Population

Table II. Primer sequences used for mutation analysis of G protein subunit alpha 11 (GNA11) and alpha Q (GNAQ) genes.

Gene Exon Primer sequence PCR product (bp)

GNA11 4 5’ M13F tail-GTGCTGTGTCCCTGTCCTG 3’ 283 5’ M13R tail-GGCAAATGAGCCTCTCAGTG 3’ 5 5’ M13F tail-AGCCGATGTCAGTCTGGTGT 3’ 415 5’ M13R tail-AAGGCAGAGGGAATCAGAGG 3’ GNAQ 4 5’ M13F tail-TCCTTCCCTTTCCGTAGACA 3’ 280 5’ M13R tail-AAGGCATAAAAGCTGGGAAA 3’ 5 5’ M13F tail-TGATCATCGTCATTCAAGAGAA 3’ 316 5’ M13R tail-AAAATGATAATCCATTGCCTGTC 3’

Table III. Distribution of G protein subunit alpha 11 (GNA11) and alpha Q (GNAQ) mutations among the samples from the total cohort and according to sample storage time. Group A: Samples stored for up to 5 years, group B: samples stored for more than 11 years.

Gene Group A (n=19) Group B (n=14) Total samples (n=33)

Mutation Frequency (%) Mutation Frequency (%) Mutation Frequency (%)

GNAQ Q209L (2) 10.5 Q209L (2) 6.1 Q209P (2) 10.5 Q209P (2) 14.3 Q209P (4) 12.1 GNA11 Q209L (6) 31.6 Q209L (1) 7.1 Q209L (7) 21.2 c.625_626delinsGC (1) 5.3 c.625_626delinsGC (1) 3.0 Mutated cases (11) 57.9 (3) 21.4 (14) 42.4 Wild-type cases (8) 42.1 (11) 78.6 (19) 57.6

preserved for up to 5 years, group B included samples preserved for amplified exons 4 and 5 (including mutation hotspot regions- more than 11 years. The associations between mutation status codons 209, 183) of GNAQ and GNA11 genes in order to (mutated GNA11 /exon5 and GNAQ /exon5) and sample groups (A vs. investigate mutations and directly sequenced the purified B) were evaluated by applying chi-square test and z-score test for two proportions. The z-score was used to test the associations PCR products. between tumor mutation status ( GNA11 /exon5 and GNAQ /exon5; Sequencing of both GNAQ and GNA11 genes was only GNA11 /exon5; only GNAQ /exon5) and occurrence of successful in all 33 tumor samples. Among the samples metastasis. This was carried out considering a) all samples screened, the overall mutation frequency was 42.4% (14/33), simultaneously and b) sample group A and B separately. The with 18.2% (6/33) for GNAQ and 24.2% (8/33) for GNA11. Kaplan–Meier method and the log-rank test were used for survival In all samples, the GNAQ or GNA11 mutations were mutually analysis using time from eye enucleation as the endpoint. Correlation exclusive and occurred exclusively at codon 209 (Figure 2). between mutation status and demographic data (gender: male vs. female; age: <52 vs. ≥52 years) was assessed through the chi-square Thirteen of the mutations found were single point test. p-Values were based on two-sided tests and were regarded as substitutions (92.85%) and one case comprised a dual point statistically significant when less than 0.05. All statistical analysis substitution (7.15%). was performed using software STATA v13 (Stata Statistical Mutations affecting codon 209 in GNAQ were c.626A>C/ Software: Release 13; StataCorp., College Station, TX, USA). p.Q209P (four samples, 12.1%) and c.626A>T/ p.Q209L (two samples, 6.1%). These mutations predicted substitution Results of glutamine by proline (p.Q209P) in 67% and by leucine (p.Q209L) in 33% of the samples carrying GNAQ mutations. The present study initially included samples from 47 patients Mutations affecting codon 209 in GNA11 gene were with UM. Out of the available tumor specimens, sufficient c.626A>T/p. Q209L (seven samples, 21.2%) and amount of intact DNA was only obtained from 33 samples c.625_626delinsGC/ p.? (one sample, 3%). These mutations (70.2% yield). The mutational screening of samples with predicted substitution of glutamine by leucine (p.Q209L) in extensively degraded DNA (n=14) resulted in incomplete 87.5% of the samples carrying GNA11 mutations, whereas a information, thus they were excluded from the study. We novel mutation was identified in one sample. The latter

5721 ANTICANCER RESEARCH 37 : 5719-5726 (2017)

Table IV. Distribution of G protein subunit alpha 11 (GNA11) and alpha Q (GNAQ) mutational status among all tumor samples and according to sample storage time. Group A: Samples stored for up to 5 years, group B: samples stored for more than 11 years.

GNA11 exon 5 or Group, N (%) Total p-Value GNAQ exon 5 A B

None 8 (42.1) 11 (78.6) 19 (57.6) 0.036 a* Mutation 11 (57.9) 3 (21.4) 14 (42.4) Total 19 (100) 14 (100) 33 (100)

Comparison of proportions 57.9% 21.4% Z=2.245 0.037 b* aChi-square test and bZ-zcore test for two proportions *Statistically significant result at α=5%.

Table V. Correlation of G protein subunit alpha 11 (GNA11) and alpha Q (GNAQ) mutational status and occurrence of metastasis.

Mutation Metastasis, Total p- Value N (%) Figure 2. Mutations detected in the present study. No Yes

GNA11 exon 5 or GNAQ exon 5 8 (38.1) 4 (40.0) 12 (38.7) 0.920 GNA11 exon 5 4 (19.1) 2 (20.0) 6 (19.4) 0.952 GNAQ exon 5 4 (19.1) 2 (20.0) 6 (19.4) 0.952 mutation has not been previously reported in any somatic cancer mutation database or in the international bibliography and consists of two nucleotide changes within codon 209 (Table III). Low yield of intact DNA during nucleic acid extraction from FFPE samples is a frequent event, due to DNA Table VI . Correlation of G protein subunit alpha 11 (GNA11) and alpha Q (GNAQ) mutational status and occurrence of metastasis according to fragmentation. Among others, storage time of FFPE tissues is sample storage time. Group A: Samples stored for up to 5 years, group a major parameter affecting the efficiency of DNA extraction. B: samples stored for more than 11 years. Indeed, all 14 samples that were excluded from the study had been stored for more than 8 years. Furthermore, the DNA Group Metastasis, N (%) Total p- Value extracted from the long-term samples was of lower quantity No Yes and quality comparing to these of the short-term preserved FFPE samples. This supports the hypothesis that partial or A 7 (53.9) 2 (50.0) 9 (52.9) 0.893 extensive DNA fragmentation may lead to less available B 1 (12.5) 2 (33.3) 3 (21.4) 0.347 amplifiable DNA, thus a higher rate of false-negative results among long-term preserved FFPE samples. Supplementary analysis was performed, after subdividing the samples into two groups according to the storage time. Group A included 19 samples collected from 2012 to 2016, group B included Table VII. Correlation between G protein subunit alpha 11 (GNA11) 14 samples collected from 1996 to 2005. A comparison of the and alpha Q (GNAQ) mutational status and demographic data. mutational frequencies between groups A and B, as well as the total number of samples included in the study, is given in Mutation p- Value Table III. The frequency of mutations in group A was significantly higher (by more than twofold) than in group B GNA11 exon 5 or GNAQ exon 5 0.863 (Table IV). The variation is also valid for individual genes GNA11 exon 5 0.654 and the difference in both cases is likewise substantial. These GNAQ exon 5 0.618 results support the above hypothesis.

5722 Psinakis et al : GNAQ and GNA11 Mutations in a Greek Uveal Melanoma Population

Figure 4. Kaplan –Meier estimate of overall survival in patients harboring G protein subunit alpha 11 (GNA11) and aIpha Q (GNAQ) mutations compared to patients with wild-type tumors according to sample storage time. Group A: Samples stored for up to 5 years; group B: samples stored for more than 11 years).

in group B resulted in alteration of correlation. Nevertheless, mutational status in UM tumors was not correlated to metastasis occurrence in either group ( p> 0.05). In the Kaplan–Meier analysis, the hypothesis that survival among patients with GNA mutations and patients without Figure 3. Kaplan–Meier estimate of overall survival in patients mutations did not differ, was tested. Data analysis was harboring mutation of either G protein subunit alpha 11 (GNA11) or performed on all samples included in the study (data not alpha Q (GNAQ) genes (A), GNAQ (B), and GNA11 (C) compared to patients with wild-type tumors. shown/Figure 3) as well as on groups A and B separately (data not shown /Figure 4). The median overall survival time for UM patients with tumors harboring GNAQ/GNA11 mutations was not significantly altered compared to that of patients with wild-type tumors ( p> 0.05). The mutational status of the tumors was tested for correlation Finally, GNAQ/GNA11 mutations were not significantly with occurrence of metastasis. Table V shows the results associated with certain demographic characteristics of the regarding this association. Among all patients whose tumors patients (sex, age at diagnosis) (Table VII). harbored GNA mutations (n=14), metastasis was detected in four (28.6%), whereas in patients with wild-type tumors (n=19) Discussion metastasis was detected in six (31.6%). No statistically significant correlation was found ( p> 0.05) for either GNA gene. Approximately 80% of UMs in the Caucasian population The same analysis was performed for groups A and B (Table carry activating mutations in either GNAQ or GNA11 , VI) separately to test whether the lower mutational frequency turning these genes into oncogenes (7-9). Mutations in both

5723 ANTICANCER RESEARCH 37 : 5719-5726 (2017) genes are known to be mutually exclusive and are suggested samples which could not be fully sequenced were collected to be the cause of constitutive MAPK pathway activation, from 1996 to 2008. The aforementioned speculation is also leading in turn to cell proliferation, even in the absence of supported by the significant differences in mutation detection external growth stimuli (10). The objective of the present rates between the groups A and B, namely the short-term study was the identification and frequency estimation of preserved and long-term preserved FFPE samples, with overall GNAQ and GNA11 mutations in Greek patients with UM and GNA mutation frequencies of 57.9% and 21.4%, respectively the association of these events with patients’ prognostication. (p= 0.025). The low mutation detection rate in group B is Mutation frequencies of GNAQ and GNA11 were 18.2% probably due to recovery of fewer intact DNA molecules from (6/33) and 24.2% (8/33) respectively, with an overall FFPE specimens. frequency of 42.4% (14/33). Our data do not confirm the One tumor harbored a double mutation in codon 209 of over-representation of GNA mutations in Greek patients, in GNA11 (c.625_626delinsGC), which has not been previously comparison to frequencies found in Caucasian population reported in any cancer database or in the literature. This result studies (3, 5, 11), but they are equivalent to those of other may be due to a single clone with two mutations or to two non-Caucasian ethnic groups (12, 13). It is unclear whether different clones, each harboring one mutation. Sanger these results may allow conclusions concerning the effects sequencing method cannot discriminate between these two of variations in sunlight exposure due to geographical possibilities (26). The reported heterozygous non synonymous differences among the Caucasian population ( i.e. different variants in codon 209 result in either glutamine to alanine latitude) (14). Nevertheless they suggest that the frequency transition (first report) in the same clone, or glutamine to of GNAQ and GNA11 mutations in UM may be more glutamic acid (to our knowledge this is the first report in the dependent on ethnicity and demographic variables than present study) and glutamine to proline transitions (9) in two hitherto considered. GNA mutations are believed to be early distinct subclones. A germline variant was not excluded oncogenic events in the development of UM (8, 15, 16), because normal retinal tissue or a peripheral blood sample however, rare driver mutations in other genes have also been from the patient was not available. Using the in silico analysis reported. Rarely mutated NRAS and BRAF proto-oncogenes tools MutationTaster (http:// www.mutationtaster.org/) and activate the MAPK pathway (17, 18, 19) and genes such as PolyPhen-2 (http://genetics. bwh.harvard.edu/pph2/) both of BRCA1-associated protein 1 ( BAP1 ), splicing factor 3b these transitions resulted in damage to the structure and subunit 1 ( SF3B1 ), eukaryotic translation initiation factor 1A function of the protein. This novel mutation did not affect the X-linked protein ( EIFAX1 ) and telomerase reverse patient’s outcome. It should be highlighted that this is the transcriptase ( TERT ) (5, 20, 21) have been reported to be second report of a double mutated UM case (27) and the first related to UM manifestation. Mutations in these genes were with a double mutation within codon 209 of GNA11 . not tested in the present study; the association of low Comparing our results to those of current literature, the percentage GNA mutations with the presence of driver following points can be addressed. According to current mutations in other genes has not been addressed. literature, all mutations in GNAQ and GNA11 involve the Nevertheless whole-exome sequencing data in UM cases has same two hotspots, p.Q209 and much less frequently R183 (4, indeed identified mutations in relatively few genes, 7, 12, 28). This was also the case in our series, where the indicating low genetic complexity (22, 23). discussed mutations occurred exclusively in codon Q209 of Several biological as well as technical factors limit the both genes. Regarding the GNAQ/GNA11 mutation ratio, accuracy of detection of rare sequence variants. Firstly, the data existing literature presents discordant results (3, 5, 12, 21). In analysis of a single sample may not be representative of the our series the GNAQ/GNA11 ratio was 1/1,3. Higher tumor as a whole, due to high heterogeneity of tumor subclones metastasis rates have been reported in patients with tumors and admixture with normal cells. Furthermore, in the classical with GNA11 mutations, postulating that mutations of GNA11 Sanger sequencing method used in the current study, mutations can be more virulent than those of GNAQ (2, 4, 5, 21). can be detected with a sensitivity of 15-20% of mutant alleles Interestingly, other studies reveal no statistically significant in a background of wild-type alleles, hence lower mutant differences in the prevalence of GNAQ and GNA11 mutations subclone populations may be disregarded. The poor quality of between two subgroups of patients; with and without the DNA extracted from FFPE samples is a major technical metastatic disease (12, 29). In accordance with the former, our factor limiting the detection of somatic mutations. DNA study revealed a higher prevalence of GNA11 mutations in fragmentation varies between FFPE specimens as a result of patients with metastatic disease (1:2 GNAQ/GNA11 ratio). fixation and storage conditions of the samples and has been Further studies are definitely needed, as the small sample size strongly associated with specimen storage time (24, 25). Long- limits the power of this statistical analysis. Finally, we noted term FFPE specimens yield shorter DNA fragments after DNA that the GNAQ p.Q209P (c.626 Α> C) mutation is extraction, thus providing less amplifiable genetic material. predominantly (in a 2:1 ratio) detected in choroidal This suggestion has been verified by the fact that all 14 melanomas that originate from the posterior region of the eye,

5724 Psinakis et al : GNAQ and GNA11 Mutations in a Greek Uveal Melanoma Population which is most exposed to focused light. This may reinforce targeted therapies. Extensive use of next-generation the already suggested hypothesis of light-induced damage sequencing technology should be able overcome any being involved in the pathogenesis of UM (30). technical pitfalls and provide useful data for the cases According to literature, the metastatic phenotype is in without well-defined targetable driver mutations. most cases acquired after loss of one copy of chromosome 3 Summing up, our study concludes that mutations in GNA or BAP1 mutations (2, 3, 20, 28). In the present study we genes occur frequently in Greek patients with UM. The examined the possibility of GNA mutations affecting patient reported mutations frequencies qualify them as potential prognosis as independent risk factors. No statistically targets for personalized therapy, however, they cannot be significant associations were found between GNAQ/GNA11 used as prognostic markers for this disease. mutation status and patient’s age at enucleation, sex or time of metastasis. The overall survival in patients with tumors Acknowledgements harboring GNA mutations was not significantly less than that of those with wild-type tumors. Additionally no correlation The Authors would like to thank HistoBio Diagnosis Pathology was observed between tumor mutation status and the Center and ALFA LAB Molecular Biology Center for their crucial occurrence of metastatic disease. Our data, in accordance support in the completion of this project. with the current literature, indicate that GNAQ and GNA11 mutations are not associated with patient outcome. References The discovery of driver genes in UM will determine the selection of different treatment regiments. Currently, direct 1 Coupland SE, Lake SL, Zeschnigk M and Damato BE: Molecular pathology of uveal melanoma. Eye (Lond) 27(2) : 230-242, 2013. targeting of mutant GNA is not possible, however, 2 Harbour JW and Chao DL: A molecular revolution in uveal new approaches have been developed in an attempt to melanoma; Implications for patient care and targeted therapy. inhibit key downstream effectors of those proteins. The Opthalmology 121(6) : 1281-1288, 2014. MEK/MAPK pathway is an important mediator of activated 3 Decatur CL, Ong E, Garg N, Anbunathan H, Bowcock AM, GNAQ and GNA11; they activate another protein called Field MG and Harbour JW: Driver Mutations in Uveal MEK1 that increases cell division. There is a series of Melanoma; Associations with profile and patient existing drugs that target the MEK1 protein and could outcomes. JAMA Opthalmol 134(7) : 728-733, 2016. 4 Urtatiz O and Van Raamsdonk CD: GNAQ and GNA11 in the therefore target GNA11 and GNAQ indirectly (31, 32). endothelin Signaling Pathway and Melanoma. Front Genet 7: 59, Several MEK inhibitors are currently being tested (33, 34) 2016. and may serve as specific gene-targeted therapy, either as 5 Dono M, Angelini G, Cecconi M, Amaro A, Esposito A I, single agents or in combination with immuno- or Mirisola V, Maric I, Lanza F, Nasciuti F, Viaggi S, Gualco M, chemotherapy, in order to improve survival rates for patients Bandelloni R, Truini M, Coviello DA, Zupo S, Mosci C and with UM. Furthermore blocking of the MAPK pathway by Pfeffer U: Mutation frequencies of GNAQ, GNA11, BAP1, those specific inhibitors appears to be an effective SF3B1, EIF1AX and TERT in uveal melanoma; detection of an therapeutic approach for metastatic UM (35). At present, activating mutation in the TERT gene promoter in a single case of uveal melanoma. BJC 110(4) : 1058-1065, 2014. ClinicalTrials.gov lists 69 trials for metastatic UM, 6 Harbour JW: Genomic, prognostic, and cell signaling advances including some targeting indirectly the mutant GNA tumor in uveal melanoma. Am Soc Clin Oncol Educ Book, pp. 388- clones (36-38). The results are promising but do not yet 391, 2013. equal the success of corresponding therapies for cutaneous 7 Herlyn M and Nathanson KL: Taking the guesswork out of uveal melanoma. melanoma. N Engl J Med 363(23) : 2256-2257, 2010. The present study is to our knowledge the first population 8 Van Raamsdonk C D, Bezrookove V, Green G, Bauer J, Gaugler study on UM patients of Greek origin, regarding the L, O'Brien JM, Simpson EM, Barsh GS and Bastian BC: Frequent somatic mutations of GNAQ in uveal melanoma and underlying genetic variations of UM tumors. Our results blue naevi. Nature 457(7229) : 599-602, 2009. should be interpreted with some caution. Validation and 9 Van Raamsdonk CD, Griewank KG, Crosby MB, Garrido MC, further exploration of these findings in additional and larger Vemula S, Wiesner T, Obenauf AC, Wackernagel W, Green G, cohorts are needed. Longer follow-up for all patients will be Bouvier N, Sozen MM, Baimukanova G, Roy R, Heguy A, required in order to establish whether tumors harboring GNA Dolgalev I, Khanin R, Busam K, Speicher MR, O'Brien J and mutation have a more aggressive phenotype comparing to Bastian BC: Mutations in GNA11 in uveal melanoma. N Engl J wild-type tumors. Further pre-clinical studies will need to be Med 363(23) : 2191-2199, 2010. 10 Zuidervaart W, Van Nieuwpoort F, Stark M, Dijkman R, Packer conducted to identify the presence and evaluate the L, Borgstein AM, Pavey S, Van der Velden P, Out C, Jager MJ, frequency of mutations in other genes of Greek patients with Hayward NK and Gruis NA: Activation of the MAPK pathway UM. Data analysis will determine how all detected mutations is a common event in uveal melanomas although it rarely occurs correlate with clinicopathological features of the tumors as through mutation of BRAF or RAS. BJC 92(11) : 2032-2038, well as the way they affect patient outcome or response to 2005.

5725 ANTICANCER RESEARCH 37 : 5719-5726 (2017)

11 Krantz BA, Dave N, Komatsubara KM, Marr BP and Carvajal 26 Vogelstein B, Papadopoulos N, Velculescu VE, Zhou S, Diaz LA RD: Uveal melanoma; epidemiology, etiology, and treatment of Jr. and Kinzler KW: Cancer Genome Landscapes. Science primary disease, Clin Opthalmol 11 : 279-289, 2017. 339(6127) : 1546-1558, 2013. 12 Xu X, Wei WB, Li B, Gao F, Zhang Z and Jonas JB: Oncogenic 27 Koopmans AE, Vaarwater J, Paridaens D, Naus NC, Kilic E, de GNA Q and GNA11 mutations in uveal melanoma in Chinese. Klein A and Rotterdam Ocular Melanoma Study group: Patient PLOS ONE 9(10) : e109699, 2014. survival in uveal melanoma is not affected by oncogenic 13 Ominato J, Fukuchi T, Sato A, Yamaguchi N, Kobayashi K, Cho mutations in GNAQ and GNA11. BJC 109(2) : 493-496, 2013. H, Oyama T and Ajioka Y: The role of mutation rates of GNAQ 28 Amaro A, Gangemi R, Piaggio F, Angelini G, Barisione G, or GNA11 in case of uveal melanoma in Japan. Appl Ferrini S and Pfeffer U: The biology of uveal melanoma. Cancer Immunohistochem Mol Morphol, 2017. doi: 10.1097/ Metastasis Rev 36(1) : 109-140, 2017. PAI.0000000000000505. [Epub ahead of print] 29 Staby KM, Gravdal K, Mørk SJ, Heegaard S, Vintermyr OK and 14 Vajdic CM, Kricker A, Giblin M, McKenzie J, Aitken J, Giles Krohn J: Prognostic impact of chromosomal aberrations and GG and Armstrong BK: Sun exposure predicts risk of ocular GNAQ, GNA11 and BAP1 mutations in uveal melanoma. Acta melanoma in Australia. Int J Cancer 101 : 175-182, 2002. Opthalmol, 2017. doi: 10.1111/aos.13452. [Epub ahead of print] 15 Zeschnigk M and Lohmann DR: Prognostic testing in uveal 30 de Lange MJ, Razzaq L, Versluis M, Verlinde S, Dogrusöz M, melanoma, Cancer Genomics; Molecular Classification, Böhringer S, Marinkovic M, Luyten GP, de Keizer RJ, de Gruijl Prognosis And Response Prediction. Chapter 3, Springer Science FR, Jager MJ and van der Velden PA: Distribution of GNAQ and and Business Media: Dordrecht, The Netherlands, 79-96, 2013. GNA11 mutation signatures in uveal melanoma points to a light 16 Onken MD, Worley LA, Long MD, Duan S, Council ML, dependent mutation mechanism. PLOS ONE 10(9) : e0138002, Bowcock AM and Harbour JW: Oncogenic mutations in GNAQ 2015. occur early in uveal melanoma. Invest Ophthalmol Vis Sci 31 Ambrosini G, Musi E, Ho AL, de Stanchina E and Schwartz GK: 49(12) : 5230-5234, 2008. Inhibition of mutant GNAQ signaling in uveal melanoma 17 Janssen CS, Sibbett R, Henriquez FL, McKay IC, Kemp EG and induces AMPK-dependent autophagic cell death. Mol Cancer Roberts F: The T1799A point mutation is present in posterior Ther 12(5) : 768-776, 2013. uveal melanoma. BJC 99(10) : 1673-1677, 2008. 32 Khalili JS, Yu X, Wang J, Hayes BC, Davies MA, Lizee G, 18 Maat W, Kilic E, Luyten GPM, de Klein A, Jager MJ, Gruis NA Esmaeli B and Woodman SE: Combination small molecule MEK and Van der Velden PA: Pyrophosphorolysis detects B-RAF and PI3K inhibition enhances uveal melanoma cell death in a mutations in primary uveal melanoma. Invest Ophthalmol Vis mutant GNAQ and GNA11-dependent manner. Clin Cancer Res Sci 49 : 23-27, 2008. 18(16) : 4345-4355, 2012. 19 Henriquez F, Janssen C, Kemp EG and Roberts F: The T1799A 33 Chen X, Wu Q, Tan L, Porter D, Jager MJ, Emery C and Bastian BRAF mutation is present in iris melanoma. Invest Ophthalmol BC: Combined PKC and MEK inhibition in uveal melanoma Vis Sci 48 : 4897-4900, 2007. with GNAQ and GNA11 mutations. Oncogene 33(39) : 4724- 20 Helgadottir H and Höiom V: The genetics of uveal melanoma; 4734, 2014. current insights. Appl Clin Genet 9: 147-155, 2017. 34 Buder K, Gesierich A, Gelbrich G and Goebeler M: Systemic 21 Griewank KG, Van de Nes J, Schilling B, Moll I, Sucker A, treatment of metastatic uveal melanoma; review of literature and Kakavand H, Haydu L E, Asher M, Zimmer L, Hillen U, future perspectives. Cancer Med 2(5) : 674-686, 2013. Thompson JF, Scolyer RA, Schadendorf D and Murali R: 35 Ambrosini G, Pratilas CA, Qin LX, Tadi M, Surriga O, Carvajal Genetic and clinic-pathologic analysis of metastatic uveal RD and Schwartz GK: Identification of unique MEK-dependent melanoma. Mod Pathol 27 : 175-183, 2014. genes in GNAQ mutant uveal melanoma involved in cell 22 Johansson P, Aoude LG, Wadt K, Glasson WJ, Warrier SK, growth, tumor cell invasion, and MEK resistance. Clin. Cancer Hewitt AW, Kiilgaard JF, Heegaard S, Isaacs T, Franchina M, Res 18 : 552-3561, 2012. Ingvar C, Vermeulen T, Whitehead KJ, Schmidt CW, Palmer JM, 36 Wu X, Li J, Zhu M, Fletcher JA and Hodi FS: C Symmons J, Gerdes AM, Jönsson G and Hayward NK: Deep inhibitor AEB071 targets ocular melanoma harboring GNAQ sequencing of uveal melanoma identifies a recurrent mutation in mutations via effects on the PKC/Erk1/2 and PKC/ NF-kappaB PLCB4. Oncotarget 7(4) : 4624-4631, 2016. pathways. Mol Cancer Ther 11(9) : 1905-1914, 2012. 23 Martin M, Maßhöfer L, Temming P, Rahmann S, Metz C, 37 Lyubasyuk V, Ouyang H, Yu F X, Guan K L and Zhang K: YAP Bornfeld N, van de Nes J, Klein-Hitpass L, Hinnebusch AG, inhibition blocks uveal melanogenesis driven by GNAQ or Horsthemke B, Lohmann DR and Zeschnigk M: Exome GNA11 mutations. Mol Cell Oncol 2(1) : e970957, 2015. sequencing identifies recurrent somatic mutations in EIF1AX 38 von Euw E, Atefi M, Attar N, Chu C, Zachariah S, Burgess B and SF3B1 in uvealmelanoma with disomy 3. Nat Genet 45(8) : L, Mok S, Ng C, Wong D J L, Chmielowski B, Lichter D I, 933-936, 2013. Koya R C, McCannel T A, Izmailova E and Ribas A: Antitumor 24 Ludyga N, Grünwald B, Azimzadeh O, Englert S, Höfler H, effects of the investigational selective MEK inhibitor TAK733 Tapio S and Aubele M: Nucleic acids from long-term preserved against cutaneous and uveal melanoma cell lines. Mol Cancer FFPE tissues are suitable for downstream analyses. Virchows 11 : 22, 2012. Arch 460(2) : 131-140, 2012. 25 Gilbert MTP, Haselkorn T, Bunce M, Sanchez JJ, Lucas SB, Jewell LD, Van Marck E and Worobey M: The isolation of nucleic amplification, acids from fixed, paraffin-embedded Received July 12, 2017 tissues which methods are useful when? PLOS One 2(6) : e537, Revised August 3, 2017 2007. Accepted August 9, 2017

5726