Abstract LB-315: Comprehensive genomic characterization of cutaneous melanoma

Abstract

Abstract Melanoma has emerged as the latest success story for both immuno- and genomics-guided targeted therapy following the advent of both monoclonal antibodies directed toward the inhibitory immune receptors CTLA-4, PD-1, and PD-L1, as well as BRAF inhibitors targeting mutant BRAF found in over 50% of melanoma patients. To characterize the genomic alterations in cutaneous melanoma, The Cancer Genome Atlas Project (TCGA) analyzed exome sequence, DNA structural variations (copy number and translocations), promoter methylation, mRNA, miRNA, and protein expression in primary and metastatic melanoma samples originating from non-glabrous skin in up to 336 samples. Interestingly, clustering analysis of RNA-seq data demonstrated 3 expression module subgroups of melanomas, categorized by elevated lymphocytic, keratin, and melanocyte differentiation markers. Consistent with previous findings, the high lymphocytic expression subgroup was found to possess a better overall survival. Mutation analysis from exome sequencing identified a broad spectrum of samples possessing a signature of UV mutagenesis dominated by C > T transitions at dipyrimidines. We observed a mean somatic mutation rate of 17 mutations per megabase (Mb), almost doubling the highest mutation rate reported previously for a cancer type analyzed by TCGA. The InVEx algorithm identified 13 significantly mutated genes (P. Bonferroni <0.05, Q <0.1), which included: known melanoma oncogenes and tumor suppressors (BRAF, NRAS, CDKN2A, TP53, and PTEN); genes previously implicated in melanoma, but found here for the first time to be significantly mutated (including the RAS-GAP NF1, IDH1 and RB1); recently implicated melanoma genes possessing either characteristic hotspot or loss of function mutations (RAC1, MAP2K1, PPP6C, and ARID2); and finally, a gene not previously implicated in melanoma, the RNA helicase DDX3X. Over 80% of samples possessed at least one mutation in the significantly mutated genes BRAF, NRAS, or NF1 involved in the MAPK pathway. Importantly, hotspot mutations in BRAF were found to be anti-correlated with both NRAS (Fisher's exact test p-value = 6.461e-16) and NF1 mutations (p-value 1.888e-06). Given this finding, we propose melanoma can be categorized into four genetic subgroups distinguished by MAPK driver mutations: BRAF, NRAS, NF1, and triple-wild type melanomas. Integrative analysis with additional platforms revealed enrichment of copy number gains in KIT, PDGFRA, KDR (VEGFR2), CCND1,CDK4, MDM2, and TERT in the triple wild-type melanomas. Integration of reverse phase protein array data also revealed differentially downstream signaling of the MAPK pathway in these four genetic subgroups. In summary, insights provided by this analysis have shed light on the different genetic mechanisms leading to cutaneous melanoma. Citation Format: Ian R. Watson, Chang-Jiun Wu, Jeffrey E. Gershenwald, Lynda Chin, On Behalf of the Melanoma Analysis Working Group and The Cancer Genome Analysis Research Network. Comprehensive genomic characterization of cutaneous melanoma. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr LB-315. doi:10.1158/1538-7445.AM2014-LB-315