Abstract 778: Potential drug targets for adenoid cystic carcinoma elucidated by proteogenomic analysis

Background: Adenoid cystic carcinoma (ACC) is a rare cancer of secretory glands accounting for 10% of salivary gland malignancies and 1% of head and neck cancers. About 1,200 new cases are diagnosed annually in the United States. ACC is typically chemoresistant and clinical trials of multiple targeted agents found few responders. Studies of ACC oncogenesis have described frequent fusion of MYB-NFIB genes and other infrequent genomic mutations. ACC has not been proteomically characterized. We hypothesized that proteogenomic analysis of ACC tumor tissues would identify clinically relevant molecular differences between ACC and other tumor types. The objective of this study was to identify altered disease pathways and potential drug targets in ACC. Methods: Clinical tumor samples of ACC and squamous cell carcinoma (SCC) of the head and neck were analyzed using the GPS Cancer diagnostic test, which includes whole genome and transcriptome sequencing, and proteomic expression analysis using mass spectrometry. To quantify proteins, tumor areas of formalin-fixed paraffin-embedded tissue sections were marked by a pathologist, microdissected and solubilized. The resulting lysate was analyzed with mass spectrometry to quantitate 30 clinically relevant proteins. Samples were further analyzed with a global proteomics platform intended to discover targetable protein biomarkers. Proteins that were overexpressed or underexpressed (defined as a 1.5-fold difference between ACC and SCC) were subjected to pathway analysis to identify perturbed pathways and potential drug targets. Finally, the mRNA expression in ACC tumors was compared with RNA-seq data from various solid tumor types using the k-nearest neighbors algorithm. Results: In 14 tumor samples, unsupervised hierarchical clustering analysis of 4,002 proteins revealed a clear separation between ACC (n=8) and SCC (n=6) tissues. Four of 8 ACC samples harbored an MYB-NFIB fusion and single samples had fusions of MYBL-NFIB or AHI1-NFIB. Mutational burden in ACC and SCC samples was 1.35 and 3.53 mutations per megabase, respectively. Pathway analysis found enrichment of ACC genes in essential cell cycle processes. CDK6 protein expression was 4-fold higher in ACC samples than SCC samples by mass spectrometry (p=0.0036) and 3-fold higher at the mRNA level. Expression of p16 protein was 3-fold lower in ACC than in SCC (p=0.0289) and corresponding p16 mRNA levels were 17-fold lower in ACC than SCC. All ACC samples harbored intact retinoblastoma (RB1) gene. Expression of mRNA in 15 ACC samples was compared with that of 25 different tumor types from TCGA and the authors9 clinical laboratory; breast cancer was found to be the nearest neighbor to ACC. Conclusions: Proteogenomic analysis revealed CDK6 overexpression in ACC. The combination of CDK6 overexpression, p16 underexpression and RB1 proficiency suggests that ACC tumors may respond to treatment with CDK6 inhibitors. Citation Format: Sheeno P. Thyparambil, Yeoun Jin Kim, Andrew G. Chambers, Dongyao Yan, Shankar Sellappan, Andrew J. Sedgewick, Yulia Newton, J Zachary Sanborn, Charles J. Vaske, Stephen C. Benz, Fabiola Cecchi, Hyunseok Kang, Todd A. Hembrough. Potential drug targets for adenoid cystic carcinoma elucidated by proteogenomic analysis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 778.