Project 458925

Pancreatic cancer is a lethal disease that typically presents at the advanced or metastatic stage. Chemotherapy remains the mainstay of treatment, but finding a targeted therapy has been a challenge due to the therapeutic intractability of the four major driver genes (KRAS, TP53, CDKN2A, SMAD4) and high diversity in secondary mutations. How each tumor's individual genomic landscape affects its response to therapy is unknown. Patient derived organoids are a novel 3-D cell culture model which can be grown from patient tumors to capture the genome, transcriptome, and patient responses to chemotherapy. Most organoid work has focused on predicting patient responses to standard chemotherapy or a small number of targeted therapies, but there has been no effort to date to systematically profile a large number of agents with integrated genomic and transcriptomic data to identify novel drug-genome associations specific to pancreatic cancer. We hypothesize that high-throughput drug screening can identify unique molecular vulnerabilities with targetable potential in pancreatic cancer. We have developed and validated a novel organoid drug screening platform capable of testing ~3000 drugs in organoid models within 8-10 weeks of diagnosis. We propose to screen organoids derived from metastatic tumors against a custom drug library and correlate organoid responses with whole genome and RNA sequencing data to uncover personalized predictors of drug response. We aim to identify subtype-specific therapies and chemosensitizing agents for these highly resistant tumors. This work will develop a resource of functional models and identify new drug targets to overcome chemoresistance in metastatic pancreatic cancer.