The long non-coding RNA landscape of neuroendocrine prostate cancer and its clinical implications, biological dysregulation, and functional impact

Neuroendocrine prostate cancer (NEPC) is a lethal subtype of castration-resistant prostate cancer (CRPC). It can develop de novo from prostate neuroendocrine cells, yet primarily is a treatment-induced phenotype arising from transdifferentiated prostate adenocarcinoma (AD) cells (NEtD). Currently there is an unmet clinical need for predictive biomarkers, therapeutic targets, and more reliable diagnostics. In this dissertation we use the first-in-field patient-derived xenograft model of NEtD, six in vitro CRPC/NEPC models, and ~30,000 PCa patient samples, including 344 NEPC or molecular analogous NEPC samples. We implement a state-of-the-art next-generation sequence analysis pipeline, capable of detecting transcripts at low expression levels to build a comprehensive lncRNA catalog (~N=40,000). Our xenograft model enabled identification of transcriptional changes during NEtD. Our in vitro models were used for functionalization and our patient samples were used to determine clinical relevancy and/or to test for patient survival. In Chapter I, we review lncRNA research in PCa over the last 30 years. We include known genomic structures, mechanisms of actions, roles in PCa progression, and their use in disease management. In Chapter II, we identify a 122-lncRNA signature capable of robustly classifying NEPC from AD, 25 with predictive ability to classify metastatic patients, and 2 (SSTR5-AS1 and LINC00514) capable of stratifying patients more probable to develop metastasis following androgen deprivation therapy (ADT). In Chapter III, we identify two NEPC molecular subtypes driven by lncRNAs FENDRR and GAS5. They also have a predictive ability to stratify ADT patients by clinical outcome. In Chapter IV, we investigate our top candidate NEPC lncRNA H19. We identify the active isoform, determine it is conserved, a dozen associated PCa risk single nucleotide polymorphisms (SNPs) nearby, and NEPC-related TFBS (MYC/MAX) embedded within. H19 was highly sensitive and relatively specific for NEPC. Functionally, we identified associations to invasion, proliferation, the NEPC phenotype, and physical interactions with EZH2. Most importantly H19 is predictive for ADT-patient outcome. Collectively, this thesis constitutes a step forward in understanding the complexity of the transcriptome for NEPC and the NEtD process. The results here will advance our knowledge of clinically relevant lncRNAs involved in cancer progression and treatment resistance.Science, Faculty ofGraduat