The region near the brain vasculature in human brain tumors, called the perivascular niche (PVN), is an important microenvironment for the maintenance of brain tumor stem-like cells (BTSCs), the development of resistance to chemo or targeted therapies, and the path for tumor infiltration to distant regions in the whole brain, leading to incurable diseases. Current in vitro models such as 2D cell cultures or 3D tumor spheroids do not contain this niche environment. Mouse models of brain tumors can recapitulate some aspects of the PVN, but have challenges in terms of costly assays, low throughput, and lack of the ability for high-resolution live cell tracking of BTSC dynamics. Herein, we propose to develop a tissue-engineered 3D microvascular niche-on-a- chip model that can incorporate primary brain tumor cells from patients in order to bridge this gap between in vitro and in vivo models. Our pilot study has demonstrated the success in co-culture of patient-derived glioblastoma cells and microvasculature in a microfluidic gel system and observed preferential localization of BTSCs in the PVN. Comparing ex vivo dynamics of individual tumor cells on-chip to single-cell transcriptomes across 10 patients further revealed a correlation between perivascular localization and transcriptional subtypes. In this project, we propose to further examine tumor cell migration and localization using a larger cohort of patient specimens and compare the results to pathological and clinical data, aiming to develop it into an ex vivo functional assay for patient prognosis and subclassification (Aim 1). We will apply scRNA-seq to the same samples to generate correlative data to identify subtypes associated with distinct ex vivo dynamics in the tissue-engineered PVN model, which can help elucidate the molecular mechanisms of PVN in tumor cell fate and invasion (Aim 2). Finally, we will investigate the response of tumor cells in PVN to chemo and targeted therapies administered through the perfusable microvascular network to assess the potential to perform personalized drug test and therapeutic stratification (Aim 3). This project will lead to a novel tissue-engineered microsystem to not only study the biology of PVN in human brain tumor development but also develop new assays for ex vivo test of human tumor cells for precision medicine.
This project will lead to a tissue-engineered model to mimic the microvascular niche in human brain tumor. It can be used to elucidate the biological mechanism in brain tumor stem cell maintanence, tumor infiltration, and functional heterogeneity. It also has the potential utility for personalized drug testing and therapeutic stratification.