In Vitro Modeling of Blood-Brain Barrier with Human iPSC-Derived Endothelial Cells, Pericytes, Neurons, and Astrocytes via Notch Signaling.

Differentiation of iPSCs into ECs, Pericytes, Neurons, and Astrocytes(A) Schematic representation of the EC and pericyte induction protocol.(B) FACS analysis for mesoderm markers, KDR and PDGFRα. Positive cells appeared at 4 and 5 days after differentiation. The percentage of KDR-positive cells in total cells are indicated.(C and F) FACS analysis for an EC marker, CD31, and pericyte marker, PDGFRβ. Positive cell appearance at 9 days (C) and 12 days (F) after differentiation. Percentages of CD31-positive cells and PDGFRβ-positive cells are indicated.(D) A phase-contrast image of differentiated ECs and immunostaining for CD31 and VE-CADHERIN at 9 days after differentiation. Scale bar, 200 μm.(E) LDL uptake assay for ECs at 9 days after differentiation. Scale bar, 200 μm.(G) Immunostaining of purified pericytes at 12 days after differentiation for αSMA, SM22α, NG2, and CALPONIN. Scale bar, 200 μm.(H) Tube formation assay for ECs at 12 days after differentiation. A phase-contrast image and immunostaining for CD31 and αSMA. Scale bar, 200 μm.(I) Schematic representation of the neuron and astrocyte induction protocol.(J) Phase-contrast image at 11 days after differentiation and immunostaining for NESTIN and TUJ1 at 16 days after differentiation. Scale bar, 200 μm.(K–M) Double immunostaining for TUJ1 and GFAP at 32 days (K, left panel), 48 days (K, right panel), 86 days (L, left panel), 100 days (L, right panel), and 145 days (M) after differentiation. Scale bars, 200 μm (K, L) and 100 μm (M).(N) qPCR for the mRNA expressions of Oct3/4, Pax6, MAP2, and GFAP during neuron and astrocyte differentiation (n = 1). mRNA expression on undifferentiated hiPSCs was set as 1.0.

Generation of ciBECs Using Four Cell Populations Derived from iPSCs(A) Schematic of the co-culture system with four lineages derived from iPSCs for ciBEC generation.(B) A phase-contrast image at 2 days after co-culture. Asterisks, ECs; arrows, endfeet of astrocytes attached to ECs. Scale bar, 200 μm.(C) Double immunostaining for CD31 and GFAP (left panel) or TUJ1 (right panel) at 5 days after co-culture. Scale bars, 200 μm.(D) qPCR for the mRNA expressions of BBB-specific transporters and receptors in purified CD31-positive ECs (n = 6 independent experiments), ciBECs (n = 7 independent experiments), and immortalized cell lines, hCMEC/D3 (n = 3 independent experiments) and HUVEC (n = 3 independent experiments) (∗p < 0.05 versus ECs). mRNA expression on ECs was set as 1.0.(E) Double immunostaining for CD31 and BCRP (upper panels) or PGP (bottom panels). Scale bars, 200 μm.

Induction of ciBECs via Notch Activation(A) Pharmacological approach using six inhibitors to study the mechanism of BBB specification. qPCR for the mRNA expressions of the BBB-specific transporters CAT3, ABCA1, MFSD2A, BCRP, PGP, and MRP5 (n = 3 independent experiments; ∗p < 0.05, ∗∗p < 0.01 versus ECs, #p < 0.05, ##p < 0.01 versus ciBECs). mRNA expression in ECs was set as 1.0.(B) Double immunostaining for CD31 and NICD at 5 days after co-culture. Scale bars, 50 μm.

Dll1 in Neurons Activates Notch Signaling to Generate ciBECs(A) qPCR for the mRNA expressions of CD31, SM22α, MAP2, and GFAP in purified ECs, pericytes, neurons, and astrocytes by FACS (n = 3 independent experiments). mRNA expression on ECs was set as 1.0.(B) qPCR for the mRNA expressions of Notch receptors Notch1, Notch2, Notch3, and Notch4 and Notch ligands Dll1, Dll4, Jagged1, and Jagged2 (n = 3 independent experiments). mRNA expression in ECs was set as 1.0.(C) Loss-of-function using Dll1 siRNA. qPCR for the mRNA expressions of the BBB-specific transporters CAT3, ABCA1, MFSD2A, BCRP, PGP, and MRP5 (n = 3 independent experiments; ∗p < 0.05, ∗∗p < 0.01 versus ECs, #p < 0.05, ##p < 0.01 versus ECs + astrocytes and neurons treated with siCon). mRNA expression in ECs was set as 1.0.

BBB Model Using ciBECs Derived from iPSCs(A) Schematic of the two-compartment BBB model. iPSC-derived ciBECs were seeded onto a Transwell filter coated with fibronectin and co-cultured with iPSC-derived astrocytes to analyze BBB properties.(B) Double immunostaining for CD31 and ZO-1 (upper panels; scale bar, 25 μm) or CLAUDIN5 (bottom panels; scale bar, 10 μm).(C) Measurement of TEER at 7 days after plating onto the Transwell filter. iPSC-derived ciBECs responded to soluble factors from astrocytes (n = 4 independent experiments; ∗p < 0.05, ∗∗p < 0.01 versus ECs without astrocytes, ##p < 0.01 versus ciBECs without astrocytes).(D) Transmission electron microscopy of iPSC-derived ciBECs after co-culture with iPSC-derived astrocytes for 24 hr. M, Transwell filter. Right panel shows higher magnification of the content in the dashed box. Scale bars, 2 μm (left panel) and 750 nm (right panel).(E) Permeability assay using fluorescein-Na/FITC-labeled dextran (n = 3 independent experiments; ∗∗p < 0.01 versus ECs with fluorescein-Na, ##p < 0.01 versus ECs with 4-kDa FITC-labeled dextran).

Screening System of Drug Permeability with BBB Model(A) PGP functional efflux assay using rhodamine-123 at 7 days after plating onto a Transwell filter co-cultured with astrocytes. Cell accumulation ratio of rhodamine-123 with verapamil treatment and vehicle (0.5% DMSO) (n = 3 independent experiments; ∗∗p < 0.01 versus ECs).(B) Representative results of nanoLC-MS/MS. Upper panel shows the intensity of caffeine; lower panel shows the intensity of epinastine. Gray data, 0.1 μM calibration samples; black data, samples after 10 min, blue data, samples after 20 min; green data, samples after 30 min; red data, samples after 40 min.(C) Drug permeability measured by nanoLC-MS/MS. We selected ten drugs with known permeability to test our BBB model (n = 3 independent experiments). CNS-positive drugs could cross the BBB into the CNS by lipid-mediated free diffusion. CNS-negative drugs had low BBB permeability for efflux transporter substrates.