Despite the recent availability of physiologically relevant in vitro cancer models such as organoids and spheroids, two-dimensional (2D) monolayer culture and animal models remain frequently used in preclinical drug development. Current in vitro models used in industry are compatible with high-throughput screening, yet are unable to recapitulate tumor microenvironment (TME) heterogeneity, a critical need that this thesis aimed to address. The objective of this thesis was to engineer a hyaluronic acid (HA) hydrogel to enable three-dimensional (3D) ex vivo culture of patient derived xenografts (PDX) and associated stroma in a perfusable microfluidic platform amenable to high-throughput screening. To achieve this, I investigated the impact of tailoring hydrogel composition and network structure for cancer-relevant fibroblasts in a perfusable microfluidic drug- screening microplate. The first part of this thesis demonstrated the importance of tuning the hydrogel parameters of crosslink density, bioligand functionalization, and network degradability on fibroblast morphology and viability. This work underscored the importance of formulating a hydrogel that could maintain structural integrity for long-term 3D culture without negatively impacting embedded cells, providing a context for rational hydrogel design within the scope of high-throughput microfluidic screening platforms.

After addressing the model extracellular matrix (ECM), I sought to recapitulate stromal encapsulation of prostate cancer cells by building a Core-Shell organization of fibroblasts wrapped by a uniform layer of fibroblasts. The second part of this thesis used this Core-Shell architecture to build a complex four-compartment in vitro metastatic prostate cancer model, consisting of prostate cancer PDXs, bone marrow derived fibroblasts, and monocytes alongside an endothelial microtubule within a perfusable microfluidic drug-screening plate. The representative nature of the model was achieved by the complexity of cell types represented, including heterogeneity inherent to the PDX cancer cell source. The interaction within this in vitro TME can be used to generate insight into how crosstalk between stromal and cancer cell compartments can impact cancer cell survival, a phenomenon that current simple 2D and 3D models do not have the capability to represent. Lastly, this work tested the feasibility of the model for drug screening on a diverse panel of PDXs using clinically relevant anti-cancer drugs.