Molecular characterization of cellular heterogeneity in pancreatic ductal adenocarcinoma by transcriptome profiling

Pancreatic ductal adenocarcinoma (PDAC) is a lethal disease with an extremely aggressive behavior due to both the advanced stage of the disease at diagnosis and the peculiar biology of this tumor. The current classifications of PDAC in different subtypes failed to associate defined molecular properties with specific morphological features and disease outcomes, thus making the histopathological classification of PDAC not informative for patient stratification for therapy choice. In order to profile the transcriptional profiles of morphologically distinct low-grade (LoG) and/or high-grade (HiG) tumor areas in PDACs, we developed a high-throughput approach of laser microdissection (LMD) coupled to RNA sequencing (RNA-seq) and an RNA single molecule fluorescence in-situ hybridization (smFISH). We showed a high efficiency and specificity in the transcriptional profiling of 500-1000 isolated cells using archived formalin‐fixed paraffin‐embedded (FFPE) samples. The current molecular classification scheme largely stratifies patients in classical or basal-like subtypes. Nevertheless, cells with classical and basal-like gene expression programs coexisted in variable proportions in the majority of cases of PDAC. In spite of this large variability both within and among patients, which is indicative of the high degree of heterogeneity in PDAC, we identified different transcriptional states that allowed us to match different gene expression programs with specific morphological features. Moreover, we identified functional protein networks characterized by important hub genes that in principle may represent druggable targets for the distinct morphological and molecular subtypes.