Fig 1.
Schematic illustration and images of neonatal blood-brain barrier on a chip (B3C).
Schematic illustration of B3C showing the tissue compartment in the center of the device surrounded by two independent vascular channels with flow access openings. The dimensions of vascular channels are 200 μm x 100 μm x 2762 μm (width x height x length) and the dimensions of tissue compartment are 1575 μm x 100 μm (diameter x height). Vascular channels are in communication with the tissue compartment through a series of 3μm porous interface (pore dimensions are: 3μm x 3μm x 100 μm, width x height x length, spaced every 50 μm) along the length of the vascular channels (A). Schematic illustration of cell culture in B3C device showing one of two vascular channels (blue) with endothelial cells lining the channel walls, the tissue compartment (red) containing astrocytes, and the porous interface (white) separating the vascular channel and tissue compartment (B). The B3C device is assembled on a microscope glass slide with plastic tubes (dark blue) allowing access to individual vascular channels and the tissue compartment (C).
Fig 2.
Passage of fluorescent dextran from the vascular channel to tissue compartment of B3C under shear flow.
Permeability of Texas Red 40 kDa dextran from vascular channel to the tissue compartment in a cell-free B3C after 5 min (A), 15 min (B), 30 min (C), 60 min (D) and 120 min (E) from the initiation of flow in vascular channel. Normalized tissue intensity in a cell-free B3C increases linearly with time in the tissue compartment (F).
Fig 3.
Neonatal RBEC cultured under flow in the vascular channel of B3C form a complete lumen.
Full view of B3C device showing RBEC cultured in the two vascular channels (A). Magnified view of inset from panel (A) showing a section of the vascular channel with cultured RBEC (B). 3D reconstruction of confocal images of neonatal RBEC cultured in B3C stained with f-actin (green) and Draq5 (red) after 5 days in culture maintained under flow of 0.01 μl/min in RBEC medium (C)—(F); images are shown with a Y-axis rotation of 0, 140, 210 and 330 degrees in (C), (D), (E) and (F) respectively. All images were acquired after 5 days of 0.01 μl/min of flow of RBEC medium over cultured RBEC in the vascular channel of B3C.
Fig 4.
Tight junction formation by neonatal RBEC under static and flow conditions as indicated by immunofluorescence staining of ZO-1.
RBEC cultured for 5 days under static conditions were stained with ZO-1 (red) and Hoechst 33342 (blue) (A). RBEC cultured for 5 days under flow of RBEC medium in B3C were stained with ZO-1 (red) and Hoechst 33342 (blue) (B). RBEC cultured under flow of ACM for 5 days in B3C were stained with ZO-1 (green) and Hoechst 33342 (blue) (C). Bright field image of B3C showing RBEC in the vascular channels after 5 days of co-culture with rat astrocytes in the tissue compartment of B3C (D). Immunofluorescence staining of RBEC in vascular channel for ZO-1 (green) and astrocytes in tissue compartment for GFAP (red) (E). Magnified view of interface with pores from panel E showing staining of cells inside the pores, ZO-1 (green), GFAP (red) and nuclei (blue) (F).
Fig 5.
Real-time analysis of passage of fluorescent dextran from the vascular to tissue compartment of B3C under shear flow.
Compared to cell-free B3C (A) or RBEC alone (B), presence of ACM (C) and co-culture with rat astrocytes (D) improves barrier function of neonatal RBEC in B3C as detected by the passage of Texas Red 40 kDa dextran from the vascular channel to the tissue compartment of B3C under shear flow. The passage of Texas Red 40 kDa dextran was monitored and imaged by Fluorescent microscopy (Nikon TE200). Representative images acquired 60 min after the initiation of flow in the vascular channel (A-D). Quantification of permeability shows that RBEC cultured in the presence of ACM or RBEC co-cultured with astrocytes exhibit significant reduction in the permeability of 40 kDa dextran compared to RBEC alone (E). Coefficient of variance of measured intensities of fluorescent dextran at 12 regularly spaced ROIs in the tissue compartment immediately adjacent to the vascular channel at 60 min after the initiation of the flow was used as an index of permeability heterogeneity. Variation in permeability is lowest in cell-free B3C but increases as the microenvironment of B3C becomes more realistic (F). (*** indicates significant difference p<0.001, one-way ANOVA with Tukey’s multiple comparison test, n = 3–4 experiments per group).
Fig 6.
ACM enhances the barrier properties of neonatal RBEC more significantly in B3C as compared to transwell.
Presence of ACM increases electrical resistance of neonatal RBEC in both B3C (A) and transwell (B), the electrical resistance measurements are from day 5. Presence of ACM increases resistance more significantly in B3C as compared to the transwell model (C). Please note that the units of electrical resistance for B3C and transwell are different as noted in the results section. [* indicates significant difference p<0.05, ** indicates significant difference p<0.01, *** indicates significant difference p<0.001, one-way ANOVA (panels A and B) or two-way ANOVA (panel C) with Tukey’s multiple comparison test; n = 3 experiments per group].
Fig 7.
B3C exhibits significantly improved barrier function compared to the transwell model and closely approximates the permeability of neonatal in vivo BBB.
B3C and Transwell BBB were constructed with neonatal RBEC in the presence of ACM. Permeability of 40 kDa dextran in B3C is significantly lower than transwell but not significantly different from that of in vivo BBB in neonatal rats. (** indicates significant difference p<0.01, *** indicates significant difference p<0.001, one-way ANOVA with Tukey’s multiple comparison test; n = 3–4 experiments per group).
Fig 8.
Neonatal RBEC exhibit distinct barrier structure and function compared to adult RBEC in B3C.
Neonatal RBEC alone (A) and neonatal RBEC + ACM (B) exhibit distinctly weaker and less organized ZO-1 staining (arrows) compared to adult RBEC alone (C) and adult RBEC + ACM (D). In the presence of ACM, neonatal RBEC exhibit discontinuous bands of ZO-1 staining (B) compared to neonatal RBEC in the absence of ACM where ZO-1 staining is discontinuous and granular (A). Presence of ACM has a significantly larger impact on both permeability (E) and electrical resistance (F) in neonatal RBEC compared to adult RBEC. Inset panels show higher magnification of white squared regions. (** indicates significant difference p<0.01, student’s t-test; n = 3–4 experiments per group).