Blood-brain barrier model on a microfluidic chip for the study of tumor cell extravasation

Abstract

With up to 40% of cancer patients showing metastatic lesions to the brain and a 30% five-year survival rate post-diagnosis, secondary tumors to the brain are a leading cause of cancer-related deaths. Understanding the mechanisms of tumor cell extravasation at the brain is therefore crucial to the development of therapeutic agents targeting this step in cancer metastasis, and to the overall improvement of cancer survival rates . Investigating the interactions between tumor cells and brain stroma is of particular interest due to the site's unique microenvironment. In fact, the interface between brain and blood, known as the blood-brain barrier (BBB), is the tightest endothelial barrier in humans. The presence of tight junctions between brain endothelial cells, coupled with the spatial organization of pericytes and astrocytes around the vasculature, restrict the entry of most solutes and cells into the brain. Yet, the brain constitutes a common metastatic site to many primary cancers originating from the lung, breast and skin. This suggests that tumor cells must employ specific mechanisms to cross the blood-brain barrier. While in vitro models aimed at replicating the human blood-brain barrier exist, most are limited in their physiological relevance. In fact, the majority of these platforms rely on a monolayer of human brain endothelial cells in contact with pencytes, astrocytes and neurons. While this approach focuses on incorporating the relevant cell types of the brain microenvironment, it fails to accurately replicate the geometry of brain capillaries, the barrier tightness of the BBB, and the juxtacrine and paracrine signaling events occurring between brain endothelial cells and stromal cells during vasculogenesis. To integrate these features into a physiologically relevant blood-brain barrier model, we designed an in vitro microvascular network platform formed via vasculogenesis, using endothelial cells derived from human induced pluripotent stem cells, primary human brain pericytes, and primary human brain astrocytes. The vasculatures formed with brain pericytes and astrocytes exhibit decreased cross-section areas, increased endothelial cell-cell tight junction expression and basement membrane deposition, as well as reduced and more physiologically relevant values of vessel permeability, compared to the vasculatures formed with endothelial cells alone. The addition of pericytes and astrocytes in the vascular system was also coupled with increased extravasation efficiencies of different tumor cell subpopulations, despite the lower permeability values measured in this BBB model. Moreover, an increase in the extravasation potential of metastasized breast tumor cells collected from the brain was recorded with the addition of pericytes and astrocytes, with respect to the parental breast tumor cell line. These results were not observed in metastasized breast tumor cells collected from the lung, thus validating our BBB model and providing useful insight into the role of pericytes and astrocytes in extravasation. Our microfluidic platform certainly provides advantages over the current state-of-the-art in vitro blood-brain barrier models. While being more physiologically relevant than most in vitro platforms when it comes to geometry, barrier function and juxtacrine/paracrine signaling between the relevant cell types, our model provides a robust platform to understand tumor cell-brain stromal cell interactions during extravasation.