Pancreatic ductal adenocarcinoma (PDAC) is a deadly disease of the exocrine pancreas, with treatment resistance attributed to the aggressive biology of the tumor microenvironment (TME). Interactions between activated pancreatic stellate cells (PSCs) and pancreatic cancer cells (PCCs) give rise to dense fibrotic masses (stroma) and the concomitant avascularity and hypoxia within the TME alters metabolic activity and mediates infiltration of T cells and chemotherapeutic agents. While many studies have investigated the mechanisms of stroma-mediated treatment resistance, these effects are still not well understood. There is thus an urgent need for more accurate models of the TME stromal compartment.

In the last decade, the synergistic combination of microfluidic technologies with 3D cultures have resulted in Organ-on-a-Chip (OOC) technologies that leverage the microscale manipulation of fluids to achieve precise spatiotemporal control of 3D tissues. However, many of these systems remain in the proof-of-concept stage due to complex fabrication steps, and sealed architectures that significantly limit methods of analysis and data collection. Therefore, more advanced microfluidic platforms for studying the stromal compartment of the PDAC TME are required, that can provide not only optical access but also non-destructive physical access for more detailed assessments of mechanical and metabolic properties.

The main objective of this thesis is to develop and characterize the stromal compartment of a novel PDAC-on-a-Chip model, using an openable multilayer microfluidic architecture that enables traditional image-based analysis as well as on-chip mechanical stiffness testing and sample extraction for small molecule detection. We first tested multilayer and phaseguide-separated channel layouts and combined these features to develop a novel microfluidic architecture that permitted 3D diffusive interactions. A modified version was used to develop a PDAC-on-a-Chip model where hydrogel-embedded PSCs were cultured with PCCs in an openable device that provided direct access to the hydrogel compartment. Our model displayed hallmark characteristics of PDAC including PSC activation, stiffening of the matrix, and collagen content increase. We also identified potential contributions to metabolic pathways via mass spectrometric analysis of extracted gel samples. These advanced platforms have thus demonstrated the potential to deepen our understanding of PDAC through modeling and analysis of more complex TME stromal characteristics.