Micro-engineering a platform to reconstruct physiology and functionality of the human brain microvasculature in vitro

摘要

A predominant unsolved challenge in tissue engineering is the need of a robust technique for producing vascular networks, particularly when modeling human brain tissue. The availability of reliable in vitro human brain microvasculature models would advance our understanding of its function and would provide a platform for high-throughput drug screening. Current strategies for modeling vascularized brain tissue suffer from limitations such as (1) culturing non-human cell lines, (2) limited multi-cell co-culture, and (3) the effects of neighboring physiologically unrealistic rigid polymeric surfaces, such as solid membranes. We demonstrate a new micro-engineered platform that can address these shortcomings. Specifically, we have designed and prototyped a molding system to enable the precise casting of ~100 μm-diameter coaxial hydrogel structures laden with the requisite cells to mimic a vascular lumen. Here we demonstrate that a fine wire with diameter ~130 μm or a needle with outer diameter ~300 μm can be used as a temporary mold insert, and agarose-collagen composite matrix can be cast around these inserts and thermally gelled. When the wire or needle is retracted under the precise positional control afforded by our system, a microchannel is formed which is then seeded with human microvascular endothelial cells. After seven days of culture these cells produce an apparently confluent monolayer on the channel walls. In principle, this platform could be used to create multilayered cellular structures. By arranging a fine wire and a hollow needle coaxially, three distinct zones could be defined in the model: first, the bulk gel surrounding the needle; then, after needle retraction, a cylindrical shell of matrix; and finally, after retraction of the wire, a lumen. Each zone could be independently cell-seeded. To this end, we have also successfully 3D cultured human astrocytes and SY5Y glial cells in our agarose-collagen matrix. Our approach ultimately promises scalabl