LABORATORY SCIENCES Identification of Candidate Tumor Oncogenes by Integrative Molecular Analysis of Choroidal Melanoma Fine-Needle Aspiration Biopsy Specimens

Tara A. McCannel, MD, PhD; Barry L. Burgess, BSc; Nagesh P. Rao, PhD; Stanley F. Nelson, MD; Bradley R. Straatsma, MD, JD

Objective: To report integrative molecular analysis of gain. Comparative RNA analysis for these 2 groups revealed choroidal melanoma fine-needle aspiration biopsy speci- 49 with greater than 4-fold higher expression and 31 mens to identify candidate tumor oncogenes. genes with greater than 4-fold lower expression in chromo- some3–losstumorsrelativetochromosome6p–gaintumors. Methods: Thirty-one choroidal melanoma fine-needle as- piration biopsy specimens were analyzed using cytopatho- Conclusions: Molecular analysis of choroidal melanoma logic diagnosis of melanoma, fluorescence in situ hybrid- fine-needle aspiration biopsy specimens demonstrated 2 cy- ization for 3, cytogenetic characterization togenetically distinct groups characterized by chromo- (GeneChip Human 250K NSPI Mapping Arrays; Af- some 3 loss or chromosome 6p gain. In –loss fymetrix, Santa Clara, California), and expression pro- melanomas relative to chromosome 6p–gain melanomas, files (GeneChip U133 Plus 2.0 Arrays, Af- integrative RNA analysis revealed genes with higher ex- fymetrix). These analyses were performed by clustering of pression and lower expression and identified several genes cytogenetic aberrations, sorting by chromosome 3 loss and that have not been reported in previous studies. chromosome 6p gain, and comparing gene expression pro- files in chromosome 3 loss– and chromosome 6p–gain tu- Clinical Relevance: Genes differentially expressed be- mors to identify genes with differential expression based tween chromosome 3–loss and chromosome 6p–gain on cytogenetic characteristics. melanomas may provide new knowledge about the bio- logic nature of choroidal melanoma and may contribute Results: Of 31 choroidal melanoma biopsy specimens in- to the development of targeted therapies. cluded in this study, 19 tumors had chromosome 3 loss, and 12 tumors without chromosome 3 loss had chromosome 6p Arch Ophthalmol. 2010;128(9):1170-1177

VEAL MELANOMA, ARISING Tschentscheretal5 reportedthatunsuper- from the choroid, ciliary vised hierarchical cluster analysis of gene ex- body, and iris of the eye, is pression data in primary uveal melanoma themostcommonprimary treated by enucleation identified 2 distinct intraocularmalignantneo- entities characterized by the presence or plasmU in adults. Despite improvements in di- absence of monosomy 3. In a larger series of agnosis and local tumor control, almost half primary uveal melanomas treated by enucle- of the patients with uveal melanoma develop ation,Onkenetal6 alsousedhierarchicalclus- metastasis during long-term follow-up, and ter analysis of gene expression to show 2 melanoma-related mortality rates have been groups of uveal melanomas (class 1 and class virtually unchanged for decades.1,2 2) that predicted greater melanoma-related Author Affiliations: As predictors of metastasis, clinical and death in patients with class 2 melanomas. In Departments of Ophthalmology (Drs McCannel and Straatsma histopathologic factors have been super- subgroup analysis, gain of chromosome 6p and Mr Burgess), Pathology and ceded by cytogenetic and molecular char- was associated with the better prognosis of Laboratory Medicine (Dr Rao), acteristics of the primary tumor. Loss of class1melanomas,whilelossofchromosome and Human Genetics heterozygosity in chromosome 3, which 3 was associated with a poor prognosis.6 (Dr Nelson), David Geffen occurs in approximately 50% of uveal In a further series of choroidal melano- School of Medicine at UCLA, melanomas, is associated with an in- mas treated by enucleation, van Gils et al7 Jules Stein Eye Institute and creased risk for the development of me- combinedclinicaloutcome,cytogeneticdata, Ophthalmic Oncology Center tastasis and for melanoma-related death.3-6 RNA hybridization, and unsupervised hier- (Drs McCannel and Straatsma), and Jonsson Comprehensive Loss of chromosome 3 is usually associ- archical cluster analysis to place tumors into Cancer Center (Drs McCannel, ated with multiplication of chromosome 2 groups based on gene expression and prog- 3,4 Nelson, and Straatsma), 8, 8q, or parts of 8q, and additional cop- nosis. Chromosome 6p gain corresponded University of California, ies of chromosome 8q are significantly re- to a favorable prognosis, and chromosome Los Angeles. lated to reduced survival.4 3 loss (often associated with 6p gain, 8p loss,

(REPRINTED) ARCH OPHTHALMOL / VOL 128 (NO. 9), SEP 2010 WWW.ARCHOPHTHALMOL.COM 1170

©2010 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 and 8q gain) was associated with metastasis and shortened matoxylin-eosin, and subsequently evaluated for cytologic evi- survival. dence of melanoma by the pathologist.12 In 2007, our group described 59 choroidal melano- mas subjected to fine-needle aspiration biopsy (FNAB) INTERPHASE FISH immediately before globe-conserving iodine 125 (125I) plaque brachytherapy (57 patients) or immediately af- To corroborate our genome-mapping array findings, FISH was ter enucleation (2 patients).8 These patients were ana- performed. In a standard manner described elsewhere,8,10,11 a lyzed using fluorescence in situ hybridization (FISH) and Spectrum Orange–conjugated probe (Abbott-Vysis, Des Plaines, high-density whole-genome mapping arrays to show that Illinois) specific for the centromeric region of chromosome 3 was used for interphase FISH. From 100 to 300 hybridization chromosomal aberrations were aligned into 2 discrete signals were manually counted in nonoverlapping nuclei of cells groups: tumors with chromosome 3 loss and tumors with under a fluorescence microscope (Zeiss Axiophot; Zeiss, Jena, chromosome 6p gain. Germany) equipped with a triple filter (4Ј,6Ј-diamidino-2- Focused on choroidal and ciliary body melanoma, this phenylindole, fluorescein isothiocyanate, and Texas Red) at the study presents an integrative molecular analysis of cho- UCLA Clinical Cytogenetics Laboratory. roidal melanoma FNAB specimens. Our analysis com- bines hierarchical clustering of chromosomal aberra- ISOLATION OF DNA AND RNA tions, sorting by chromosome 3 loss and chromosome 6p gain, and comparing of RNA expression profiles to In the operating room, biopsy aspirates were expelled directly identify candidate oncogenes. into a cell reagent (RNAProtect Cell Reagent; QIAGEN Sci- ences Inc, Valencia, California) within 15 seconds of collec- tion, and the needle hubs were rinsed with the same reagent METHODS to maximize recovery. Stabilized pooled aspirates were pel- leted, and DNA and RNA were simultaneously isolated from Among 82 patients with primary choroidal melanoma subjected the pooled sample using a kit (AllPrep DNA/RNA Mini Kit, 125 to FNAB immediately before I plaque brachytherapy or imme- QIAGEN Sciences Inc) per the manufacturer’s instructions. diately (Ͻ5 minutes) after enucleation between May 16, 2006, and July 22, 2008, 31 eyes (31 patients) met the following criteria for inclusion in the study: (1) voluntary consent to participate in the DNA ANALYSIS BY MAPPING ARRAYS research study, (2) cytopathologic diagnosis of melanoma, (3) suf- ficient FISH material for the chromosome 3 centromere, (4) ad- Isolated DNA was quantified using a commercial product equate material for nucleic acid analyses, (5) successful analysis (ND-1000; NanoDrop, Wilmington, Delaware). No DNA sample by cytogenetic characterization (GeneChip Human 250K NSPI was subjected to whole-genome amplification techniques. The Mapping Arrays; Affymetrix, Santa Clara, California), (6) com- DNA copy number was assessed using mapping arrays (Gene- pleted gene expression profiles (GeneChip Human Genome U133 Chip Human 250K NSPI). Probe preparation, hybridization, and Plus 2.0 Arrays, Affymetrix), and (7) a positive finding of cyto- reading were performed by the UCLA DNA Microarray Core Fa- genetic aberration by copy number analysis of the mapping cility according to the standard 96-well protocol provided by Af- array data. Any patient whose biopsy sample was deficient in any fymetrix. Copy number variation was computed using commer- 1 of these 7 criteria was excluded from the study. The most com- cially available software (CNAT version 4.0.1, Affymetrix). mon reasons for exclusion were insufficient RNA (n=20), inad- Chromosomalaberrationfrequencyanalysiswasperformedusing equate material by FISH (n=20), and undetectable cytogenetic Fisher exact test. Chromosomal aberration clustering for each bi- abnormality in chromosome 3 or 6 (n=12). opsy specimen was performed using 1 minus the Pearson prod- uct moment correlation matrix as an input of the mean linkage hi- STUDY CONSENT erarchical clustering to arrive at a dendrogram clustering tree.

This research was approved by the Institutional Review Board of ANALYSIS OF RNA EXPRESSION the University of California, Los Angeles (UCLA). Work was in BY GENE EXPRESSION PROFILES compliance with the Health Insurance Portability and Account- ability Act of 1996. Before treatment, evaluation of each patient The RNA was quantified on a spectrophotometer (NanoDrop) included comprehensive ophthalmic examination, ultrasonogra- and analyzed on a bioanalyzer (2100; Agilent, Santa Clara) for phy, photography, optical coherence tomography, and fluores- integrity. For inclusion in the study, biopsy specimen RNA had cein angiography. All patients underwent systemic evaluation, to meet the following criteria: (1) a yield of greater than 400 usually by an oncologist at the Jonsson Comprehensive Cancer ng of total RNA, (2) a ratio of A260 to A280 (the optical spec- Center at UCLA, and were offered psychological support by a clini- trometer measurements of absorbance at the wavelengths of 260 cal psychologist or social worker with particular expertise in cho- 9 nm and 280 nm) exceeding 1.9, and (3) an RNA integrity num- roidal melanoma. Following treatment, patients were evaluated ber of at least 8.5 as determined using the bioanalyzer. Most every 6 months or more frequently when indicated. samples required 1 round of amplification, and prepared probes The technique of intraoperative transscleral FNAB and 125I 8,10,11 were hybridized to gene expression profiles (GeneChip Hu- plaque brachytherapy has been described elsewhere. In brief, man Genome U133 Plus 2.0 Arrays) at the UCLA DNA Micro- FNAB was performed using a 30-gauge needle via a tangential array Core Facility using the standard Affymetrix protocol. transscleral approach. The biopsy specimen underwent pro- cessing for cytologic evidence of malignant melanoma, FISH for chromosome 3, nucleic acid analyses, and cell culture. INTEGRATIVE COMPARATIVE AND STATISTICAL ANALYSIS CYTOPATHOLOGIC EXAMINATION To determine key overexpressed and underexpressed genes that An aspirate was smeared on glass slides in the operating room. were differentially expressed in chromosome 3–loss and chro- They were immediately fixed in 95% ethanol, stained with he- mosome 6p–gain melanomas, integrative comparative and sta-

(REPRINTED) ARCH OPHTHALMOL / VOL 128 (NO. 9), SEP 2010 WWW.ARCHOPHTHALMOL.COM 1171

©2010 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 Table 1. Clinical Variables Among 31 Patients Who Participated in the Studya

Melanoma, Patient No./Sex/ Diseased Tumor Greatest Basal Age at Treatment, y Eye Height, mmb Diameter, mm Location Treatmentc Mel20-06-043/F/76 R 3.29 6.5 Choroid Brachytherapy Mel20-06-012/M/75 R 6.24 13.0 Choroid Brachytherapy Mel20-06-031/M/56 R 6.87 10.0 Choroid Brachytherapy Mel20-07-061/M/59 L 6.29 12.0 Choroid Brachytherapy Mel20-07-067/M/50 L 6.97 12.0 Choroid Brachytherapy Mel20-07-051/F/39 R 9.83 15.0 Ciliary body with iris component Brachytherapy Mel20-06-037/M/70 L 9.00 17.0 Choroid Enucleation Mel20-07-054/M/37 R 8.23 11.0 Choroid Brachytherapy Mel20-07-073/M/64 R 8.33 17.0 Choroid Brachytherapy Mel20-08-128/F/52 L 7.20 14.5 Choroid Brachytherapy Mel20-07-069/F/85 L 7.02 10.0 Ciliary body with choroid component Brachytherapy Mel20-07-070/F/86 L 8.50 37.7 Ciliary bodyd Enucleation Mel20-08-115/M/84 L 6.44 15.0 Choroid Brachytherapy Mel20-06-039/M/63 L 4.06 10.3 Choroid Brachytherapy Mel20-06-010/M/77 R 7.45 13.0 Choroid with ciliary body component Brachytherapy Mel20-08-131/M/75 L 10.85 12.8 Choroid Brachytherapy Mel20-08-129/F/81 L 8.60 15.0 Ciliary body with iris component Brachytherapy Mel20-06-013/M/65 L 3.84 11.8 Choroid Brachytherapy Mel20-07-048/M/79 L 3.58 8.0 Choroid Brachytherapy Mel20-08-124/M/42 L 8.90 13.0 Choroid Brachytherapy Mel20-06-014/M/72 R 2.95 8.0 Choroid Brachytherapy Mel20-06-046/F/50 R 5.13 11.0 Choroid Brachytherapy Mel20-07-068/M/71 L 3.82 12.0 Choroid Brachytherapy Mel20-07-052/M/53 L 7.21 14.9 Choroid Brachytherapy Mel20-06-009/M/86 L 10.80 15.0 Choroid with ciliary body component Enucleation Mel20-08-127/F/26 R 7.16 15.0 Choroid Brachytherapy Mel20-07-062/M/30 L 3.92 12.0 Choroid Brachytherapy Mel20-07-063/M/78 R 3.29 9.5 Choroid Brachytherapy Mel20-06-040/M/66 L 4.06 9.0 Choroid Brachytherapy Mel20-07-065/M/77 R 12.44 19.3 Choroid with ciliary body component Enucleation Mel20-06-038/F/59 R 4.98 9.0 Choroid Brachytherapy

Abbreviations: F, female; L, left; M, male; R, right. a Patient data are sorted by cytogenetic complexity according to hierarchical clustering used in the Figure (A). Mel20-06-010, Mel20-06-039, Mel20-07-069, and Mel20-07-070 have subsequently died of metastatic disease. b Maximum tumor height as measured by ultrasonography. c Brachytherapy indicates iodine 125 plaque brachytherapy. d Ring melanoma with choroid and iris components and extraocular extension.

tistical analysis was performed. RNA expression data CEL files Before treatment, the mean greatest basal diameter of cho- were generated using Affymetrix software.13 R version 2.6 (http: roidal melanomas was 13.2 mm (range, 6.5-19.3 mm), //www.r-project.org/) and Bioconductor (http://www.bioconductor and the mean apical height was 6.7 mm (range, 3.0-12.4 .org/) were accessed to normalize the CEL files using the Celsius mm). Twenty-nine melanomas were treated with globe- microarray database. Celsius was accessed to quantify and nor- conserving brachytherapy, and 2 melanomas were treated malize each CEL file relative to 50 random samples of the same platform (GeneChip Human Genome U133 Plus 2.0 Arrays) using with enucleation (Table 1). the default variables of robust multi-array average. Thirty-one patients with choroidal melanoma had no Normalized data were imported into dChip (http://biosun1 clinical or radiographic evidence of melanoma metasta- .harvard.edu/complab/dchip/) for differential expression analy- sis or other malignant neoplasm at the time of treat- sis.14,15 Two-group comparative analysis was performed using ment. Follow-up of 31 patients after brachytherapy or selection variables of at least 2-fold change and t test at PϽ.05. enucleation ranged from 6 to 42 months (median follow- Samples were permuted 100 times through dChip to assess the up, 30 months; mean follow-up, 26.1 months). During false discovery rate. that interval, 4 patients developed clinical evidence of liver metastasis, and 2 of these patients died of liver failure. RESULTS CHROMOSOMAL ABERRATIONS PATIENTS WITH PRIMARY CHOROIDAL MELANOMA In 31 FNAB specimens from primary choroidal mela- noma, chromosomal aberrations were identified using The 31 eyes with primary choroidal melanoma were from mapping arrays (GeneChip Human 250K NSPI) and 31 patients (22 men and 9 women), with a mean age of FISH for the centromere of chromosome 3. Tumors 64 years (median age, 66 years; age range, 26-86 years). clustered into 2 distinct categories based on the funda-

(REPRINTED) ARCH OPHTHALMOL / VOL 128 (NO. 9), SEP 2010 WWW.ARCHOPHTHALMOL.COM 1172

©2010 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 A 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 p q p q p q p q p q p q p q p q p q p q p q p q p q p q p q p q p q p q p q p q q q

Mel20-06-043 Mel20-06-012 Mel20-06-031 Mel20-07-061 Mel20-07-067 Mel20-07-051 Mel20-06-037 Mel20-07-054 Mel20-07-073 Mel20-08-128 2x Mel20-07-069 Mel20-07-070 2x Mel20-08-115 2x Mel20-06-039 2x Mel20-06-010 2x Mel20-08-131 2x Patient No. Mel20-08-129 2x Mel20-06-013 Mel20-07-048 Mel20-08-124 Mel20-06-014 2x Mel20-06-046 2x Mel20-07-068 2x Mel20-07-052 2x 2x Mel20-06-009 Mel20-08-127 2x Mel20-07-062 2x 2x Mel20-07-063 Mel20-06-040 Mel20-07-065 Mel20-06-038

B Chromosome 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 p q p q p q p q p q p q p q p q p q p q p q p q p q p q p q p q p q p q p q p q q q Mel20-06-010 2x Mel20-06-039 2x

Patient No. Mel20-07-069 Mel20-07-070 2x

Figure. Chromosomal aberration patterns. A, Among 31 primary choroidal melanomas as determined by cytogenetic characterization (GeneChip Human 250K NSPI Mapping Arrays; Affymetrix, Santa Clara, California) and copy analysis software (CNAT version 4.0.1, Affymetrix). Hierarchical clustering using 1 minus Pearson product moment correlation matrix sorting revealed 2 cytogenetic groups identified by chromosome 3 loss or by chromosome 6p gain. Red boxes indicate whole-arm loss or near–whole-arm loss. Green boxes represent whole-arm gain or near–whole-arm gain. Boxes labeled 2ϫ indicate the detection of 4 or more copies. Four of the primary tumors highlighted in yellow resulted in liver metastasis. B, Among 4 primary tumors that have thus far resulted in liver metastasis. Each primary tumor had chromosome 3 loss and chromosome 8q gain. Losses in 6q, 8p, and 16q were frequent aberrations.

mental aberration of chromosome 3 loss or chromo- GENE EXPRESSION some 6p gain (Figure, A). Nineteen tumors had chro- PROFILES mosome 3 loss, and 12 tumors had chromosome 6p gain in the absence of chromosome 3 loss. Among 19 dChip 2-group comparative analysis identified 903 melanomas with chromosome 3 loss, additional aberra- genes that met the selection criteria, with a median false tions included 14 with chromosome 8q gain. Of these discovery rate of 43 (4.8%) and a 90th percentile of 341 14, 10 also had 8p loss; 3 had 6p gain in addition to (37.8%). Table 2 lists 49 genes with greater than chromosome 3 loss. Among 12 melanomas with chro- 4-fold higher expression in chromosome 3–loss tumors mosome 6p gain in the absence of chromosome 3 loss, 5 relative to chromosome 6p–gain tumors. Table 3 lists tumors had 6q loss, 5 had 8q gain, and 3 had 6p gain as 31 genes with greater than 4-fold lower expression in the sole chromosomal aberration. chromosome 3–loss tumors relative to chromosome Four of 31 patients developed clinical evidence of me- 6p–gain tumors. Several genes are characterized as hav- tastasis during follow-up, and their cytogenetic profiles ing robust expression in only 1 of 2 cytogenetic groups. are shown in the Figure, B. All 4 tumors had chromo- Some of the genes with prominent expressional differ- some 3 loss and chromosome 8q gain, 3 had 8p loss, 3 ences between the 2 groups are associated with the fol- had 6q loss, and 1 had 6p gain. The CNAT version 4.0.1 lowing pathways: G –coupled signaling, calcium analysis revealed that chromosome gains and losses response pathways, cell adhesion marker expression, throughout all samples seemed to be whole arm or al- retinoic acid response pathways, and regulation of pal- most whole arm in each instance. mitoylation (Table 4).

(REPRINTED) ARCH OPHTHALMOL / VOL 128 (NO. 9), SEP 2010 WWW.ARCHOPHTHALMOL.COM 1173

©2010 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 Table 2. Genes With Highly Upregulated Expression in Chromosome 3–Loss Choroidal Melanomasa

Chromosome Chromosome 3 Chromosome 6p Probe Set ID Gene Symbol Location Loss, mean Gain, mean Fold Change P Value 206638_at HTR2B 2q37.1 3247.09 40.95 79.29 Ͻ.001 229823_at . . . 8q22.3 874.25 11.64 75.08 .01 204939_s_at PLN 6q22.31 1376.15 33.28 41.36 Ͻ.001 206137_at RIMS2 8q22.3 1088.06 69.55 15.64 .03 206626_x_at SSX1 Xp11.22 672.41 57.86 11.62 .04 228170_at OLIG1 21q22.11 158.99 14.57 10.91 .03 211425_x_at SSX4 Xp11.23 946.02 92.51 10.23 .02 230650_at . . . 8q13.3 200.41 20.10 9.97 Ͻ.001 205825_at PCSK1 5q15 144.03 15.80 9.12 .01 204897_at PTGER4 5p13.1 2536.72 281.65 9.01 Ͻ.001 229975_at BMPR1B 4q22.3 197.03 22.48 8.77 Ͻ.001 209031_at CADM1 11q23.2 1012.56 119.67 8.46 Ͻ.001 212671_s_at HLA-DQA1 6p21.32 708.67 89.11 7.95 .04 209496_at RARRES2 7q36.1 990.52 132.03 7.50 .01 202158_s_at CUGBP2 10p14 576.75 77.78 7.42 Ͻ.001 227812_at TNFRSF19 13q12.12 606.72 86.56 7.01 Ͻ.001 219895_at FAM70A Xq24 285.25 46.72 6.11 .003 235061_at PPM1K 4q22.1 751.93 123.87 6.07 Ͻ.001 231202_at ALDH1L2 12q23.3 306.30 52.12 5.88 Ͻ.001 202660_at ITPR2 12p11.23 417.53 70.97 5.88 Ͻ.001 235182_at C20orf82 20p12.1 129.82 22.18 5.85 .002 243339_at . . . 5q33.1 72.47 12.54 5.78 .03 238720_at OMG 17q11.2 80.77 14.22 5.68 .001 244463_at ADAM23 2q33.3 202.36 35.89 5.64 .003 204472_at GEM 8q22.1 392.78 69.59 5.64 .001 213245_at ADCY1 7p13 260.61 49.52 5.26 .002 216471_x_at LOC100130648 Xp11.22 197.75 37.66 5.25 .02 201117_s_at CPE 4q32.3 114.83 22.03 5.21 .01 206465_at ACSBG1 15q25.1 139.30 26.82 5.19 .001 206082_at HCP5 6 457.84 91.47 5.01 Ͻ.001 213808_at . . . 2q33.3 239.78 48.20 4.97 Ͻ.001 208451_s_at C4A 6p21.32 484.87 98.43 4.93 Ͻ.001 225056_at SIPA1L2 1q42.2 285.37 58.16 4.91 Ͻ.001 203373_at SOCS2 12q22 338.37 70.26 4.82 Ͻ.001 200953_s_at CCND2 12p13.32 694.07 144.72 4.80 Ͻ.001 213664_at SLC1A1 9p24.2 192.06 41.22 4.66 .002 201525_at APOD 3q29 373.28 81.13 4.60 .03 204018_x_at HBA1 16p13.3 4922.52 1090.50 4.51 .007 213661_at DKFZP586H2123 11p13 295.47 65.87 4.49 .02 230302_at . . . 8q21.13 256.36 57.22 4.48 Ͻ.001 226517_at BCAT1 12p12.1 487.78 108.77 4.48 .04 242964_at . . . 2p23.2 195.95 43.80 4.47 Ͻ.001 224901_at SCD5 4q21.22 418.97 98.20 4.27 .002 216268_s_at JAG1 20p12.2 682.33 161.06 4.24 .001 203060_s_at PAPSS2 10q23.2 579.68 138.84 4.18 .001 239239_at . . . 9q21.11 820.50 198.87 4.13 .02 230441_at PLEKHG4B 5p15.33 114.20 28.05 4.07 Ͻ.001 214180_at MAN1C1 1p36.11 339.60 84.18 4.03 Ͻ.001 212992_at AHNAK2 14q32.33 943.1 235.95 4.01 Ͻ.001

a A 2-group comparative analysis of gene expression profiles (GeneChip Human Genome U133 Plus 2.0 Arrays; Affymetrix, Santa Clara, California) identified 49 genes with greater than 4-fold higher expression in chromosome 3–loss tumors relative to chromosome 6p–gain tumors. Expression values for each gene are the mean expression for 19 chromosome 3–loss melanomas and 12 chromosome 6p–gain melanomas and include only those genes with t test at PϽ.05. The table is sorted in descending order of differential expression. Ellipses indicate not applicable.

2-7,10,16-20 COMMENT creased risk of metastasis. There may be addi- tional clusters of melanoma with normal chromosome An increasing body of evidence shows that choroidal mela- 3 disomy and normal chromosome 6p18 or with aberra- noma cytogenetic aberrations and gene expression pro- tions in chromosome 21 as the only cytogenetic change.2 files cluster into at least 2 groups characterized (1) by On the premise that chromosome 3 loss or chromo- chromosome 3 loss, distinct gene expression profile, and some 6p gain is an early aberration and possibly the ear- increased risk of metastasis with tumor-related death and liest manifested chromosomal aberration, we report (2) by chromosome 6p gain in the absence of chromo- integrative molecular analysis of 31 choroidal melano- some 3 loss, distinct gene expression profile, and de- mas with chromosome 3 loss or with chromosome 6p gain

(REPRINTED) ARCH OPHTHALMOL / VOL 128 (NO. 9), SEP 2010 WWW.ARCHOPHTHALMOL.COM 1174

©2010 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 Table 3. Genes With Highly Downregulated Expression in Chromosome 3–Loss Choroidal Melanomasa

Chromosome Chromosome 3 Chromosome 6p Probe Set ID Gene Symbol Location Loss, mean Gain, mean Fold Change P Value 223572_at HHATL 3p22.1 56.76 2313.04 40.75 .002 210809_s_at POSTN 13q13.3 40.78 1227.46 30.10 .02 205799_s_at SLC3A1 2p21 37.19 663.85 17.85 .03 231535_x_at ROPN1 3q21.1 133.32 1297.42 9.73 .005 232054_at PCDH20 13q21.31 108.24 1012.07 9.35 .009 201839_s_at TACSTD1 2p21 28.71 248.79 8.67 .046 232523_at MEGF10 5q23.2 108.95 827.63 7.60 Ͻ.001 205413_at MPPED2 11p14.1 71.23 515.36 7.24 .001 203786_s_at TPD52L1 6q22.31 112.11 754.89 6.73 .02 238592_at PDLIM3 4q35.1 23.43 150.44 6.42 .047 235238_at SHC4 15q21.1 70.50 437.89 6.21 .001 218332_at BEX1 Xq22.1 98.48 589.79 5.99 .01 230303_at SYNPR 3p14.2 181.70 1017.10 5.60 .02 236936_at . . . 8p11.23 23.37 127.10 5.44 .01 236300_at . . . 12p12.2 135.81 705.28 5.19 .001 242206_at . . . 7q11.21 21.70 108.24 4.99 .008 228245_s_at LOC100132881 12p13.31 622.21 3034.07 4.88 Ͻ.001 234583_at CHL1 3p26.3 18.74 90.67 4.84 .02 207010_at GABRB1 4p12 17.79 86.07 4.84 .04 204584_at L1CAM Xq28 57.16 272.58 4.77 .04 227827_at . . . 4q35.1 204.77 965.99 4.72 Ͻ.001 241017_at TBC1D8 2q11.2 141.85 652.58 4.60 .04 225728_at SORBS2 4q35.1 93.41 413.22 4.42 .01 205514_at ZNF415 19q13.41 64.33 282.46 4.39 Ͻ.001 227181_at LNP1 3q12.2 678.44 2973.63 4.38 .001 227826_s_at . . . 4q35.1 393.97 1671.67 4.24 Ͻ.001 1552365_at SCIN 7p21.3 148.56 625.04 4.21 .03 37892_at COL11A1 1p21.1 73.97 309.20 4.18 .008 232720_at LINGO2 9p21.2 26.30 109.25 4.15 .01 239278_at . . . 20q13.13 60.10 248.71 4.14 .007 214373_at . . . 5p15.33 33.82 136.35 4.03 .006

a A 2-group comparative analysis of gene expression profiles (GeneChip Human Genome U133 Plus 2.0 Arrays; Affymetrix, Santa Clara, California) identified 31 genes with greater than 4-fold lower expression in chromosome 3–loss tumors relative to chromosome 6p–gain tumors. Expression values for each gene are the mean expression for 19 chromosome 3–loss melanomas and 12 chromosome 6p–gain melanomas and include only those genes with t test at PϽ.05. The table is sorted in descending order of differential expression. Ellipses indicate not applicable.

in the absence of chromosome 3 loss sorted by this fun- tumors. Other authors have shown TROY expression to be damental cytogenetic abnormality and analyzed by com- upregulated in metastasizing uveal7 and cutaneous21 mela- parative gene expression; we identify 49 genes with greater nomas and to have limited expression elsewhere in adults. than 4-fold higher expression and 31 genes with greater TROY mRNA levels have been shown in a murine mela- than 4-fold lower expression in chromosome 3–loss mela- noma cell line to be responsive to retinoic acid.21 As a sur- nomas relative to chromosome 6p–gain melanomas face receptor protein, TROY may be a valuable target for (Tables 2 and 3). These genes may have potential roles clinical applications or for study of the underlying bio- in G protein–coupled signaling, calcium response path- logic nature of metastasis. HHATL is a structural ana- ways, cell adhesion marker expression, retinoic acid re- logue of an acyltransferase that appends a palmitic acid to sponse pathways, and regulation of palmitoylation. As sonic hedgehog, enhancing its activity; HHATL possesses such, the differential expression of these genes may con- the same functional domain but has an amino acid change tribute to the greater propensity of chromosome 3–loss that prevents it from binding palmitic acid and may serve tumors to form metastatic lesions in patients and may as a negative regulator of palmitoylation.22 Although highly implicate these pathways as targets for studying the pro- expressed in most chromosome 6p–gain melanomas, gression of this disease. HHATL expression in chromosome 3–loss melanomas is Two genes with major overexpression and underex- low and is silenced in the metastatic-outcome tumors herein. pression in chromosome 3–loss melanomas that have gar- HHATL resides on chromosome 3p22.1 and may repre- nered our interest are tumor necrosis factor receptor su- sent an early expressional alteration. perfamily member 19 (TNFRSF19,orTROY [GenBank Previous groups reporting about gene expression in cho- NM_148957]) on chromosome 13q12.12 and hedgehog roidal melanoma have found similarities and differences acyltransferaselike (HHATL [GenBank NM_020707.3]) on when using various techniques and methods.5-7,23 There- chromosome 3p22.1. TROY messenger RNA (mRNA) is fore, direct comparisons with our data may be impossible. highly expressed in most chromosome 3–loss tumors, in- For example, gene expression profiles of tumors with known cluding the 4 tumors with metastatic outcome mentioned metastatic outcome have been compared with those with- herein, and is virtually unexpressed in chromosome 6p gain out known metastasis. However, our study derives gene ex-

(REPRINTED) ARCH OPHTHALMOL / VOL 128 (NO. 9), SEP 2010 WWW.ARCHOPHTHALMOL.COM 1175

©2010 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 array, which provides a higher degree of resolution than Table 4. Genes of Interest Grouped Into Potentially Relevant a array comparative genomic hybridization, which may ex- Functional Classes of Biologic Activity plain why Tschentscher et al,5 who also derived a list of genes based on cytogenetic differences, did not identify G protein–coupled signaling genes similar to those that we found. Perhaps most rel- HTR2B PTGER4 evant to the understanding of our data is the 2008 study 7 GEM by van Gils et al, with whom we identified several genes Calcium response pathways in common (Table 5). In their gene list, van Gils et al used PLN metastatic outcome and nonmetastatic outcome to arrive CABP7 at their findings. All 3 studies contribute knowledge about RIMS2 potentially relevant genes toward understanding the bio- PCSK1 ITPR2 logic nature of choroidal melanoma. ADCY1 Noteworthy in our series of 31 choroidal melano- TACSTD1 mas, patients with chromosome 3–loss tumors had greater SCIN risk of metastasis and tumor-related death. Four of 19 Cell adhesion marker expression patients with chromosome 3–loss tumor developed clini- CADM1 cal evidence of liver metastasis, and 2 of these patients OMG POSTN died of liver failure within 2 years after primary mela- PCDH20 noma treatment. During short-term posttreatment fol- L1CAM low-up (range, 6-42 months to date), none of 12 pa- CHL1 tients with chromosome 6p–gain melanomas have COL11A1 developed clinical evidence of melanoma metastasis. Retinoic acid response pathways One limitation of this study is the small sample popu- TNFRSF19 lation of 31 patients with choroidal melanomas. An- RARRES2 RARB other limitation is the short posttreatment follow-up Regulation of palmitoylation period (range, 6-42 months). HHATL This study has several strengths. Our analysis is con- sistent in the use of FNAB to obtain tumor samples and a Boldface indicates overexpressed genes in chromosome 3–loss in the rapid processing of tumor samples to minimize the melanomas, and regular font indicates underexpressed genes in chromosome 3–loss melanomas. interval before nucleic acid stabilization. The technique of FNAB allowed inclusion of small- and medium-sized melanomas managed using globe-conserving therapy, Table 5. Significant Genes Identified Herein in Common which may elucidate early changes in the cytogenetic and With Data by van Gils et al7 in a 2008 Studya expressional characteristics of choroidal melanomas. Pro- cessing of the tumor samples in the operating room mini- Gene Symbol Fold Change mized the time between withdrawal of the sample and Increased expression preservation of the transcriptomic state in stabilization HTR2B 74.09 medium. This may account for our discovery of genes PLN 37.23 (such as HHATL) that have not been previously re- RIMS2 15.61 ported. As an additional strength, the simultaneous iso- PTGER4 8.75 lation of DNA and RNA verified the presence of mela- CUGBP2 7.26 TNFRSF19 6.90 noma cytogenetic aberration in the sample used to detect PPMK1 6.26 RNA expression. With study of DNA and RNA in a single ITPR2 5.87 sample, we excluded samples in which the mapping ar- SIPA1L2 4.94 ray detected no aberration and avoided the possible in- JAG1 4.21 clusion of normal or benign tissue in our expressional Decreased expression analyses. A further strength of this study is the consis- ROPN1 9.93 PCDH20 9.29 tent use of mapping arrays (GeneChip Human 250K NSPI) MEG10F 7.51 for high-resolution chromosomal aberration detection. SYNPR 5.60 The consistent use of mapping arrays have additional ana- 273651_x_at 4.44 lytic benefits through alternative integrative expression TBC1D8 4.32 analyses involving examination of single-nucleotide poly- ENTPD1 4.02 morphisms, identification of isodisomies, and bioinfor- matic detection of focal gain and loss. a The table lists genes we identified with a greater than 4-fold expression difference in chromosome 3–loss melanomas relative to chromosome In summary, we used integrative molecular analysis 6p–gain melanomas that were identified in classifier sets by van Gils et al.7 of choroidal melanoma FNAB specimens to demon- strate 2 cytogenetically and biologically distinct mela- noma groups characterized by chromosome 3 loss or by pression based entirely on the cytogenetic differences of chromosome 6p gain. Comparative RNA analysis in chro- chromosome 3 loss and chromosome 6p gain without chro- mosome 3–loss melanomas relative to chromosome 6p– mosome 3 loss. Furthermore, our study used the tech- gain melanomas revealed several unreported genes with nique of whole-genome single-nucleotide polymorphism differential expression that may provide growth and cell

(REPRINTED) ARCH OPHTHALMOL / VOL 128 (NO. 9), SEP 2010 WWW.ARCHOPHTHALMOL.COM 1176

©2010 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021 survival advantages to chromosome 3–loss melanomas; 8. Young TA, Burgess BL, Rao NP, Gorin MB, Straatsma BR. High-density genome these results warrant further study. array is superior to fluorescence in-situ hybridization analysis of monosomy 3 in choroidal melanoma fine needle aspiration biopsy. Mol Vis. 2007;13:2328- 2333. Submitted for Publication: October 30, 2009; final re- 9. Beran TM, McCannel TA, Stanton AL, Straatsma BR, Burgess BL. Reactions to vision received January 22, 2010; accepted April 20, 2010. and desire for prognostic testing in choroidal melanoma patients. J Genet Couns. Correspondence: Tara A. McCannel, MD, PhD, Depart- 2009;18(3):265-274. 10. Young TA, Rao NP, Glasgow BJ, Moral JN, Straatsma BR. Fluorescent in-situ ment of Ophthalmology, David Geffen School of Medi- hybridization for monosomy 3 via 30-gauge fine-needle aspiration biopsy of cho- cine at UCLA, and Jules Stein Eye Institute, 100 Stein roidal melanoma in vivo. Ophthalmology. 2007;114(1):142-146. Plaza, Los Angeles, CA 90095-7000 (TMcCannel@jsei 11. Young TA, Burgess BL, Rao NP, Glasgow BJ, Straatsma BR. Transscleral fine- .ucla.edu). needle aspiration biopsy of macular choroidal melanoma. Am J Ophthalmol. 2008; Author Contributions: Dr McCannel had full access to all 145(2):297-302. 12. Glasgow BJ, Foos RY. Methods in ocular cytology. In: Ocular Cytopathology. Bos- the data in the study and takes responsibility for the in- ton, MA: Butterworth-Heinemann; 1993:1-6. tegrity of the data and the accuracy of the data analysis. 13. Day A, Carlson MRJ, Dong J, O’Connor BD, Nelson SF. Celsius: a community Financial Disclosure: None reported. resource for Affymetrix microarray data. Genome Biol. 2007;8(6):R112. http:// Funding/Support: This study was supported by the www.ncbi.nlm.nih.gov/pmc/articles/PMC2394754/?tool=pubmed. Accessed June 15, 2010. George E. and Ruth Moss Trust and by an unrestricted 14. Li C, Wong WH. Model-based analysis of oligonucleotide arrays: expression in- grant from Research to Prevent Blindness (both to Dr dex computation and outlier detection. Proc Natl Acad Sci U S A. 2001;98(1): McCannel). 31-36. Additional Contributions: Fei Yu, PhD, offered statis- 15. Hosack DA, Dennis G Jr, Sherman BT, Lane HC, Lempicki RA. Identifying bio- tical expertise. Ascia Eskin, MSc, provided microarray ex- logical themes within lists of genes with EASE. Genome Biol. 2003;4(10):R70. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC328459/?tool=pubmed. Ac- pertise. Michael Gorin, MD, PhD, gave thoughtful in- cessed June 16, 2010. sight in reviewing the manuscript. 16. Worley LA, Onken MD, Person E, et al. Transcriptomic versus chromosomal prog- nostic markers and clinical outcome in uveal melanoma. Clin Cancer Res. 2007; 13(5):1466-1471. REFERENCES 17. Hughes S, Damato BE, Giddings I, Hiscott PS, Humphreys J, Houlston RS. Microarray comparative genomic hybridization analysis of intraocular uveal mela- 1. Singh AD, Topham A. Survival rates with uveal melanoma in the United States: nomas identifies distinctive imbalances associated with loss of chromosome 3. 1973-1997. Ophthalmology. 2003;110(5):962-965. Br J Cancer. 2005;93(10):1191-1196. 2. Kujala E, Mäkitie T, Kivelä T. Very long-term prognosis of patients with malig- 18. Ehlers JP, Worley L, Onken MD, Harbour JW. Integrative genomic analysis of nant uveal melanoma. Invest Ophthalmol Vis Sci. 2003;44(11):4651-4659. aneuploidy in uveal melanoma. Clin Cancer Res. 2009;14(1):115-122. 3. Prescher G, Bornfeld N, Friedrichs W, Seeber S, Becher R. Cytogenetics of twelve 19. Parrella P, Sidransky D, Merbs SL. Allelotype of posterior uveal melanoma: im- cases of uveal melanoma and patterns of nonrandom anomalies and isochro- plications for a bifurcated tumor progression pathway. Cancer Res. 1999;59 mosome formation. Cancer Genet Cytogenet. 1995;80(1):40-46. (13):3032-3037. 4. Sisley K, Rennie IG, Parsons MA, et al. Abnormalities of chromosomes 3 and 8 20. Höglund M, Gisselsson D, Hansen GB, et al. Dissecting karyotypic patterns in in posterior uveal melanoma correlate with prognosis. Genes Chromosomes malignant melanomas: temporal clustering of losses and gains in melanoma karyo- Cancer. 1997;19(1):22-28. typic evolution. Int J Cancer. 2004;108(1):57-65. 5. Tschentscher F, Hüsing J, Hölter T, et al. Tumor classification based on gene 21. Spanjaard RA, Whren KM, Graves C, Bhawan J. Tumor necrosis factor receptor expression profiling shows that uveal melanomas with and without monosomy superfamily member TROY is a novel melanoma biomarker and potential thera- 3 represent two distinct entities. Cancer Res. 2003;63(10):2578-2584. peutic target. Int J Cancer. 2007;120(6):1304-1310. 6. Onken MD, Worley LA, Ehlers JP, Harbour JW. Gene expression profiling in uveal 22. Abe Y, Kita Y, Niikura T. Mammalian Gup1, a homolog of Saccharomyces cer- melanoma reveals two molecular classes and predicts metastatic death. Cancer evisiae glycerol uptake/transporter 1, acts as a negative regulator for N-terminal Res. 2004;64(20):7205-7209. palmitoylation of Sonic hedgehog. FEBS J. 2008;275(2):318-331. 7. van Gils W, Lodder EM, Mensink HW, et al. Gene expression profiling in uveal 23. Petrausch U, Martus P, Tönnies H, et al. Significance of gene expression analy- melanoma: two regions on 3p related to prognosis. Invest Ophthalmol Vis Sci. sis in uveal melanoma in comparison to standard risk factors for risk assess- 2008;49(10):4254-4262. ment of subsequent metastases. Eye (Lond). 2008;22(8):997-1007.

Call for Papers

Archives of Ophthalmology, along with JAMA and other Archives specialty journals, will participate in a consor- tium theme issue on infectious diseases/immunology in April 2011. Manuscripts on uveitis, ocular infections, and ocular involvement in systemic infectious and immu- nological diseases received by October 1, 2010, will have the best chance for consideration for this theme issue.

(REPRINTED) ARCH OPHTHALMOL / VOL 128 (NO. 9), SEP 2010 WWW.ARCHOPHTHALMOL.COM 1177

©2010 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/28/2021