Multi-dimensional analyses identify genes of high priority for pancreatic cancer research

Abstract

Genomic profiling has unveiled the molecular subtypes and mutational landscape of pancreatic ductal adenocarcinoma (PDAC). However, there is a knowledge gap on the consistency of gene expression across PDAC tumors profiled in independent studies and this limits follow up research. To facilitate novel drug target prioritization and biomarker discovery, we investigated the most consistently expressed genes in human PDAC. We identified ~4,000 genes highly or lowly expressed in at least 4 of 5 microarrays (adjusted P<0.05) and validated their expression pattern in additional datasets, bulk tumor and single-cell RNA sequencing samples. Over 50% of the genes were previously uncharacterized in PDAC; many correlated with proliferation, metastasis, mutation, tumor grade, and ~41% predicted overall survival. We identified 185 high-priority targets (notably in cell cycle and glycolysis) whose inhibition suppressed PDAC cell viability in multiple RNA interference datasets and these genes predicted treatment in mouse models. Our results represent important milestone in the quest for mechanisms, drug targets and biomarkers in PDAC, and originate from an adaptable analytical concept that can aid discovery in other cancers. HighlightsO_LIIdentifies ~4,000 consistent genes across PDAC microarrays, >50% of which have not been studied C_LIO_LIGlycolysis and cell cycle are the most consistent processes in PDAC C_LIO_LIHeterogeneous pathways underlie or correlate with clinicopathological variables C_LIO_LIIdentifies 205 genes with similar expression pattern in PDAC tissues and peripheral blood C_LIO_LIHighlights 185 upregulated genes that are high priority therapeutic targets in PDAC C_LI