J. Yu1, G. Zhou1, S. Liu1, J. Wu1, R. Sanchez1, D. Dawson1, W. Fisher2, F. C. Brunicardi1 1University Of California – Los Angeles,Surgery Department,Los Angeles, CA, USA 2Baylor College Of Medicine,Houston, TX, USA
Introduction: Recent molecular characterizations of pancreatic ductal adenocarcinoma (PDAC) have identified hundreds of genetic alterations and aberrant gene expression in key biological processes and signaling pathways. However, PDAC global database analyses are insufficient in identifying relevant PDAC biomarkers for individual patients. Our objective was to develop a list of PDAC signature genes that could serve as potential biomarkers for individuals with PDAC using a novel combination of real time genomic analysis and published PDAC databases.
Methods: PDAC and normal matching pancreas specimens (n=8) were procured and processed for genomic sequencing. RNA sequencing libraries were generated using Illumina TruSeq RNA v2.0 and sequenced on an Illumina HiSeq 2500 platform. Sequencing reads were aligned to the human genome hg19 using TopHat and Bowtie. Three microarray gene expression datasets from the GEO database were collected and processed with R and Bioconductor packages. PDAC signature genes were identified by comparing whole genome transcriptome profiles from gene expression microarrays on independent cohorts and microarray platforms. Weighted Gene Co-expression Network Analysis (WGCNA) was then conducted to identify hub genes that highly correlate with PDAC. Hub genes were further validated with RNAseq and QPCR in PDAC and normal tissue control specimens.
Results: 1007 differentially expressed genes were identified with a cutoff (log2 fold-change > 0.5, FDR < 0.01) by comparing whole transcriptome of PDAC with normal matching tissues from three independent cohorts of two microarray platforms. GO-term enrichment analysis revealed that these 1007 genes were significantly enriched in ECM-receptor interaction, focal adhesion, complement and coagulation cascades, and glycolysis and gluconeogenesis pathways. With higher stringency (log2 fold-change > 2 and FDR < 0.01), 83 PDAC signature genes that distinguish PDAC from the normal pancreas were identified. To identify modules of highly co-regulated genes, co-expression gene networks using the transcriptome data of the entire 1007 gene list were constructed, revealing five modules and 12 hub genes with high connectivity and significant correlation with PDAC. Three hub genes, MBOAT2, LAMC2, and TSPAN1, were validated using QPCR of individual PDAC and normal pancreas specimens.
Conclusion: This novel actionable genomic analysis utilizes a comparison of RNAseq in real time to microarray on a large number of PDAC samples and WGCNA using a multifactorial genomic analysis and reveals a list of 83 PDAC signature genes, including MBOAT2, LAMC2 and TSPAN1. The data suggest that this actionable genomic analysis platform will help increase the understanding of signaling pathways and the identification of potential therapeutic targets in real-time for each patient with PDAC.