Dissertations and Theses

Date of Award


Document Type



Biomedical Engineering

First Advisor

Sihong Wang


Microfluidics, Tumor chip, Cancer immunotherapy, cancer-associated fibroblast, endothelial cells, triple negative breast cancer


In the United States, breast cancer is the first-leading cancer type in women. Current strategies to treat breast cancer mainly rely on surgery and chemotherapy, particularly for those cancer subtypes lacking specific antigens, such as triple-negative breast cancer (TNBC). However, the low survival rate of TNBC is still a significant concern. Immunotherapy holds the promise, including lower toxicity, fewer side effects and more sustained therapeutic efficacy, and may address this issue. Recent studies have shown in some breast cancer subtypes that the presence of tumor infiltrated lymphocytes correlates patients’ survival rate, especially for TNBC patients. In addition, the immune checkpoint ligands, PD-L1, is upregulated in 30-50% of all the breast cancer subtypes. Atezolizumab and pembrolizumab, two antibody drugs targeting the PD-L1/PD-1 pathway, have been approved by FDA in combination with chemo drugs for certain TNBC cases in 2019 and 2020, respectively. These facts reveal the feasibility of immunotherapy in treating breast cancer. However, generation of a proper murine model to mimic human tumor microenvironment (TME), which is composed of a variety of different cell types, for immunotherapy validation is difficult, costly, and complicated. There is therefore an unmet need for novel approaches to replicate TME for pre-screening these drug candidates. Moreover, several studies have demonstrated that the immune response observed in 2D monolayer culture is significantly different from that in 3D culture. Microfluidic platforms further provide a spatial control of cell distribution as well as the TME construction, thereby allowing a more precise control of flow to mimic in vivo transport conditions as closely as possible.

In this study, we developed a three-layer microfluidic cell array platform to reconstruct the physiologically spatial structure of TME by culturing cancer cells with microvascular endothelial cells (ECs) in two compartments, separated by a thin layer with clustered pores to guide flow between the two compartments and allow their interactions. Cancer associated fibroblasts (CAFs), the most abundant stromal cells in TME, were encapsulated with cancer cells in Matrigel to reconstruct TME that is more physiologically relevant. Compared with conventional 2D in vitro assays, in our platform, the spatial configuration of endothelial monolayer on the top channel represents an EC barrier against lymphocyte infiltration to mimic the in vivo scenario occurring during the cancer immunotherapy treatment. Continuous flow in the endothelial channel provided nutrient supply as well as waste removal in TME. It also mimics the mechanical and chemical stimulations of bloodstream. Our 3D microfluidic cell array was further incorporated with pneumatic microfluidic valves to prevent cross-interaction between chambers in order to increase the throughput. In addition, the recirculation circuit, which was constructed by microfluidic check valves, was integrated into our platform to enable on-chip immune cell circulation with unidirectional flow by a syringe pump with functions of infusion and withdraw. Collectively, these designs allow to study effects of tumor-endothelium and tumor-stroma interactions in cancer drug responses and kinetics in a scenario close to real physiological conditions. The demonstration of CD8+ T cell infiltration and cytotoxicity to the cancer cells on this platform further shows this platform’s potential for prescreening cancer immunotherapy candidates in a high throughput manner.

Available for download on Sunday, December 20, 2026