Pericyte-Assisted Vascular Lumen Organization in a Novel Dynamic Human Blood-Brain Barrier-on-Chip Model

Abstract

Organ-on-Chip (OoC) technology provides a powerful platform for neurovascular research, enabling the precise replication of the blood-brain barrier (BBB) microenvironment, including its 3D architecture and the influence of dynamic blood flow. This study introduces a novel microfluidic device designed to investigate the morphological and structural adaptations of human brain endothelial cells (ECs) within narrow, square-shaped microchannels that closely mimic the microvessels of the brain's microcirculation. The endothelial microchannels are layered above a microchamber filled with Matrigel and abluminal vascular cells, enhancing cell-cell interactions across the BBB interface. The system integrates co-culture with pericytes and astrocytes while subjecting brain ECs to physiologically relevant pulsatile flow. The findings reveal that the morphology and cytoskeletal organization of brain ECs are distinctly influenced by pulsatile flow depending on the presence of pericytes and astrocytes. Specifically, in the absence of perivascular support, brain ECs exhibit a stretched morphology with prominent actin stress fibers, while co-culture with pericytes and astrocytes promotes endothelial rearrangement, leading to lumen formation and enhanced barrier properties. This study highlights the essential role of perivascular cells in modulating endothelial responses under microvascular confinement and physiologically relevant flow. These insights advance in vitro models of the neurovascular unit and BBB mechanobiology.

Keywords: HCMEC/d3; bbb‐on‐chip; blood‐brain barrier; organ‐on‐chip; pulsating flow.

Figure 1

Schematic representation of two microfluidic chip designs: A) “enclosed” chip and B) “free” chip. Both configurations consist of a PDMS layer with microchannels, a thin porous PC membrane, and a second PDMS layer forming a milli‐well. In the “enclosed” chip, the milli‐well is open and sealed thanks to a plastic housing, while in the “free” chip, the milli‐well is closed, allowing for distinct experimental setups. The milli‐well represents the brain side of the hBBB model, enabling the static culture of pericytes and astrocytes, whereas the microchannels simulate the blood side, allowing brain ECs to be cultured under flow conditions.

Figure 2

ECs monoculture in the microfluidic device. A) Schematic representation of ECs distribution within the chip, with the gray box indicating the region observed under confocal microscopy, referred to as “microchannels view.” B) Confocal image of F‐actin‐stained ECs within the microchannels, with insets i, ii, iii) showing magnified regions of cells featuring aligned, prominent actin stress fibers. C) 3D projections illustrate ECs adherence on both sides of the interposed membrane, with a tendency for alignment along the microchannel interface. (Scale bar: 50 µm).

Figure 3

Vascular Lumen in the hBBB model. A) Schematic representation of ECs, pericytes, and astrocytes distribution within the hBBB model, with the gray box indicating the region observed under microscopy, referred to as “microchannels view.” B) Widefield image of microchannels in the hBBB model, with insets i, ii) showing magnified regions of walls lined by ECs partially covered by pericytes. Staining markers include Anti‐α‐SMA for Pericytes (magenta), DAPI for nuclei (blue), and FITC‐phalloidin for F‐actin (green). C) 3D rendering of confocal z‐stack highlights the uniform distribution of F‐actin (green) in cells along the entire surface of the microchannel walls, forming a well‐defined lumen, with a top view (D) on a single channel showing the organization of actin in the cortical area of cells. E) 3D rendering of confocal z‐stack highlights the distribution of α‐SMA‐stained pericytes (magenta), closely interacting with and partially surrounding the lumen within the microchannels, with a top view (F) on a single channel showing pericytes along the microchannels. (scale bar: 50 µm).

Figure 4

TJs and Astrocytic Projections in the hBBB Model. A) Confocal images of ECs within the microchannels stained for ZO‐1, DAPI, and F‐actin. Insets i, ii) highlight magnified regions where ZO‐1 displays a characteristic zipper‐like arrangement (white arrows), indicative of TJ organization. B) Depth map reconstructed from a confocal z‐stack of F‐actin‐stained cells, showing astrocytic projections (white arrows) extending from the milli‐well toward the microchannels, demonstrating their close spatial relationship with the endothelial lumen. (Scale bar: 50 µm).

Figure 5

OrientationJ‐based analysis of HCMEC/d3 in the hBBB model. A) Source image. B) Gradient X image. C) Gradient Y image. D) Coherency map. E) Hue–saturation–brightness (HSB) color‐coded map together with the color wheel. F) Local distribution of the orientations investigated by the ellipse measure method.

Figure 6

OrientationJ‐based analysis of HCMEC/d3 monoculture with microchannels. A) Source image. B) Gradient X image. C) Gradient Y image. D) Coherency map. E) Hue–saturation–brightness (HSB) color‐coded map together with the color wheel. F) Local distribution of the orientations investigated by the ellipse measure method.

Figure 7

Permeability Comparison:BBB Model versus In Vivo. The apparent permeability coefficients (Papp (cm ⁻¹s)) of our hBBB model (blue bars) closely align with those measured in rat brains (orange bars) for Dextran‐70K and Dextran‐20K, demonstrating comparable permeability within the same order of magnitude.