Pancreatic ductal adenocarcinoma (PDAC) remains one of the most difficult to treat major cancers, with a 5-year survival of less than 5%, and little to no improvement in prognosis over the previous fifty years. The complexity and heterogeneity of the pancreatic tumour and its highly desmoplastic stromal microenvironment present major physical and biochemical barriers to the delivery and efficacy of therapeutics. Improving the success rate of therapeutic interventions for this cancer requires both a detailed understanding of stromal dynamics and drug resistance within the tumour, as well validated in vitro and in vivo models through which novel therapeutic targets can be identified and investigated.

Within this project I focused initially on further developing, validating, and evaluating mechanistically a coculture model of gemcitabine resistance, established previously. I demonstrated that the gemcitabine resistance observed in coculture was a product of cell-cell contact, transient in nature, and related to cell density-mediated signalling processes. I discovered that the mesenchymal-like cell lines thought to drive the resistance effect within the model were epithelial cancer cells in origin, rather than a cancer associated fibroblast (CAF) line, verified through genotype and protein signature experiments.

This finding was expanded through demonstration that true pancreatic CAFs did not induce gemcitabine resistance in the coculture model and hence demonstrated the real challenge in isolating true pancreatic CAFs from genetically engineered mouse PDAC tumour tissue using canonical techniques. Nevertheless, the in vitro coculture model data was congruent with in vivo and clinical gemcitabine efficacy data, suggesting it may have utility as a lower cost model for evaluating efficacy of novel drug combinations with the aim of subsequent translation into the clinic.

The coculture model was used as the foundation of a genome-wide shRNA depletion screen to identify sensitizers of gemcitabine resistance in PDAC in vitro culture. The primary screen and subsequent validation screen identified 19 high confidence genes with potential translational value, of which knockdown sensitised to gemcitabine within this model. This is alongside an expanded list of 444 genes significantly contributing to the resistance phenotype, representing pathways such as MTORC signalling and DNA damage response. This dataset serves as the first shRNA-mediated coculture screen for gemcitabine resistance in PDAC.

My findings lay credence to the value of this model of gemcitabine resistance for preclinical use and provides a robust and voluminous shRNA dataset supporting target identification for the ablation of gemcitabine chemoresistance in PDAC.