Abstract LB-24: Molecular subtypes of pancreatic adenocarcinoma with different biological characteristics and therapeutical responses

Abstract Pancreatic ductal adenocarcinoma (PDAC) is an invariably lethal disease. Due to its late clinical presentation, surgical cure is attempted in only a minority, and resected patients usually recur quickly with incurable, metastatic disease. These attributes may reflect the variable and often disappointing responses seen when deploying the standard treatment, gemicitabine alone or in combination with other chemotherapeutic agents in unselected PDAC populations. Studies in other solid tumors have shown that heterogeneity in therapeutic responses can be anticipated by molecular differences between tumors, and targeting drugs specific to tumor subtypes in which they are predicted to be selectively effective can indeed improve treatment. Seeking to extend this new paradigm of personalized medicine to PDAC, we performed an integrative analysis of transcriptional profiles of human primary PDAC tumors (n = 143) and cell lines (n = 35). We have defined 62 gene (PDAssigner) signature that classifies the human PDAC tumors into three subtypes (classical, QM-PDAC and exocrine-like tumors) and present evidence that two (classical and QM-PDAC) of these three are present in human and mouse PDAC cell lines. Interestingly, we found that PDAC subtype classification is an independent predictor of overall survival, and patients with classical PDAC subtype tumors have significantly better survival than those with QM-PDA tumors. In addition, the expression of GATA6 (a transcriptional factor involved in the embryonic pancreas development) and KRAS (mutated in majority of PDACs) are functionally associated with classical PDAC subtype relative to QM-PDAC subtype. Intriguingly, treatment of PDAC cell lines with gemcitabine or erlotinib shows significantly differential and opposing responses between classical and QM-PDAC subtypes. Overall, we describe here a broadly integrative approach to PDAC subtype discovery and analysis that identifies gene expression signature with prognostic significance in primary patient samples and couples them with biological information and subtype specific drug responses. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr LB-24. doi:10.1158/1538-7445.AM2011-LB-24