Abstract 2183: Expression patterns in pancreatic intraepithelial neoplasms

Abstract

Background: Pancreatic cancer remains the fourth leading cause of cancer-related death in the United States with a 5-year survival rate of 5%. This oppressively-low survival is owed largely to the fact that the disease is rarely diagnosed before reaching an advanced stage. The broad majority of pancreatic tumors are thought to arise from pancreatic intraepithelial neoplasia (PanINs). These may exist for many years in a patient prior to diagnosis, but there remains no reliable, non-invasive method to screen patients for PanINs. One of the main challenges in identifying suitable biomarkers in the expression landscape of PanINs is the difficulty in obtaining pure RNA of usable quality/quantity from such small and RNAse-rich lesions. To this end, we have developed a sample processing strategy for the preservation and isolation of PanIN RNA through laser microdissection. Our aim is to sequence and compare these samples of PanIN RNA in order to identify biomarkers for early detection. Methods: Gross dissections of fresh patient pancreatic tissue were mounted on activated PEN slides and processed for microdissection through our rapid workflow. Total RNA was extracted from 13 laser microdissected patient samples (1 normal, 2 PanIN1s, 3 PanIN2s, and 7 PanIN3s) and sequenced on a SOLiD v. 5500 Wildfire. Total RNA was extracted from an additional 8 laser microdissected samples (2 normals, 2 PanIN1s, 2 PanIN2s, and 2 PanIN3s) and sequenced on an Illumina HiSeq 2500. In both sets, a cutoff RIN score of 7.0 was used for RNA quality. Data from the two sets of samples were combined using ComBat for batch removal. Lists of differentially-expressed genes in comparisons between the different stages of PanINs and normal epithelium were generated using limma with an empirical Bayes method and fitted intensity trend. Results: Using cutoffs of adjusted p-value 2.32, we find that a number of genes in the limma-generated lists are significantly upregulated in PanINs compared to normal epithelium: 33, 44, and 42 in PanIN1s, 2s, and 3s respectively. Many of the highly overexpressed genes in PanINs are previously-identified oncogenes in pancreatic cancer and PanINs, including REG4, CLDN18, and PPARG. At least one of the most highly overexpressed genes in PanIN3s is novel and has intriguing biomarker and functional implications. We have confirmed its overexpression in patient PanINs and tumors through IHC staining and are in the process of functional studies. Conclusions/Future directions: In this study, we describe a workflow for the isolation of pure, high-quality RNA from PanIN lesions. We have identified a few novel targets that we are in the process of validating in patient tumors and conducting functional studies on. Additionally, we are beginning to explore alternative transcription and gene fusion data in our samples for the possibility of additional targets. It is our hope that one or more of these may show promise as a screening or treatment target for pancreatic cancer at its earliest stages. Citation Format: Michael Ayars, Eileen O'Sullivan, Michael Goggins. Expression patterns in pancreatic intraepithelial neoplasms. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 2183. doi:10.1158/1538-7445.AM2015-2183