Cell Lines and Cell Culture

The induced pluripotent stem cells (iPSC), SIGi001-A-2 (66540357, EBiSC, England), were cultured in mTeSRTM1 culture medium (05850, StemCell Technologies, France) in 6 wells plate pre-coated with 0.333 mg/ml Matrigel Growth Factor Reduced (354230, Corning, Netherlands) for 1 h at 37 °C. The primary human brain vascular pericytes, HBVP (1200, ScienCell Research Laboratories, USA), were cultured in complete pericyte medium (1201, ScienCell Research Laboratories, USA) in flasks pre-coated with 2 μg/cm2 poly-l-lysine (P4707, Sigma-Aldrich, Germany). All cells were kept at 37 °C, in a 5% CO2 atmosphere. For differentiation of iPSC into iPSC-derived BMEC, a two-step protocol was followed as reported in Katt et al. (2016). 300.000/9.5 cm2 iPSC were cultured in mTeSRTM1. Differentiation was initiated by growing 60–70% confluent colonies with unconditioned medium for 5 or 6 days. The unconditioned medium consisted of DMEM/F12 (12660012, Thermo Fisher, Switzerland) supplemented with 20% KnockOut™ Serum Replacement (10828010, Thermo Fisher, Switzerland), 1% non-essential amino acids (11140050, Thermo Fisher, Switzerland), 0.5% L-glutamine (G7513, Sigma-Aldrich, Germany), and 0.836 μM beta-mercaptoethanol (31350010, Thermo Fisher, Switzerland). Subsequently, the medium was switched to endothelial cell serum-free medium (11111044, Thermo Fisher, Switzerland) supplemented with 1% human platelet-poor derived serum (P2918, Sigma-Aldrich, Germany), 20 ng/ml basic Fibroblast Growth Factor (bFGF; GF003, Sigma-Aldrich, Switzerland), and 10 μM all-trans retinoic acid (RA; R2625, Sigma-Aldrich, Germany). Cells remained in complete endothelial cell medium for 2 days before being subcultured in the system of choice as reported below.

Characterization of iPSC and iPSC-Derived BMEC in 2D Cell Culture

For cell evaluation in 2D, 20.000/0.33 cm2 iPSC and 100.000/0.33 cm2 iPSC-derived BMEC were seeded in vessels pre-coated with 0.333 mg/ml Matrigel Growth Factor Reduced for 1 h at 37 °C or with a 50/50 (v/v) mixture of 100 μg/ml collagen IV (C5533, Sigma-Aldrich, Germany) and 50 μg/ml fibronectin (F1056, Sigma-Aldrich, Germany) at 37 °C overnight, respectively. For fluorescence imaging analysis, CELLview™ Slides Greiner Bio-One (7.543.979, Huberlab, Switzerland) were used. Pluripotency of iPSC was evaluated at the gene expression level by using the TaqMan® hPSC Scorecard™ Panel service offered by Life Technologies Corporation. For analysis of pluripotency at the protein level, cells were fixed and stained with the Pluripotent Stem Cell 4-Marker Immunocytochemistry Kit (A24881, Thermo Fisher, Switzerland) according to the manufacturer’s protocol. The ability of iPSC to differentiate into the three germ layers was investigated by using the STEMdiff™ Trilineage Differentiation Kit (05230, Stemcell technologies, France) according to the manufacturer’s protocol. Cells were fixed on day 5 for mesoderm and endoderm or day 7 for ectoderm and labeled with specific lineage primary antibodies (Table 1).

Table 1 List of primary antibodies used for three germ layers characterizationEstablishment of Static- and Dynamic-Based In Vitro BBB Models

For the static-based BBB model, 6.5 mm transwells with 0.4 μm pore polycarbonate membrane inserts (10147291, Fisher Scientific, Switzerland) were used. For the co-culture model, the bottom side of the membrane was pre-coated with 2 μg/cm2 poly-l-lysine at 37 °C for 1 h, followed by coating of the top side with a 50/50 (v/v) mixture of 100 μg/ml collagen IV and 60 μg/ml fibronectin at 37 °C overnight. 300.000/0.33 cm2 iPSC-derived BMEC and 20.000/0.33 cm2 HBVP were seeded on the apical and bottom sides of the membrane, respectively, and co-cultured for 2 days. For the dynamic-based BBB model, the microfluidic 3D cell culture OrganoPlate® 3-lane (4004-400-B, Mimetas, Netherlands) was used. According to the manufacturer’s protocol, 4 mg/ml collagen I-based extracellular matrix (ECM) gel (3447-020-01, R&D systems, Switzerland) was dispensed in the gel inlet of the chip at 37 °C for 1 h. Then, coating of the top channel was performed with a 50/50 (v/v) mixture of 100 μg/ml collagen IV and 60 μg/ml fibronectin at 37 °C overnight. Therefore, 50.000/chip iPSC-derived BMEC were seeded in the top channel and the plate was incubated flat for 5 h before placing it on an interval rocker switching between a + 7° and − 7° inclination every 8 min, allowing a bidirectional flow-induced shear stress of ~ 1.2 dyne/cm2 for 2 days (Wevers et al. 2018).

Assessment of the Barrier Tightness

Transendothelial electrical resistance (TEER) measurement was performed by placing the transwells into a CellZscope system (nanoAnalytics, GmbH) and values were recorded in real-time and analyzed with the CellZscope software.

Permeability was measured after exposure to the following molecules: Lucifer yellow (LY) CH dipotassium salt (L0144, Sigma-Aldrich, Germany) or Alexa Fluor™ 488 IgG (A11078, Thermo Fisher, Switzerland). Under static conditions, 100 µl of 100 µg/ml LY solution was added to the apical chamber and the plate was placed on an orbital shaker at 100 rpm, for 1 h. Under dynamic conditions, 500 µg/ml LY or 200 µg/ml Alexa Fluor™ 488 IgG was perfused in the top channel of the chip according to the manufacturer’s protocol “Barrier Integrity Assay in the OrganoPlate®” for up to 6 h or 4 h, respectively. Fluorescence of the basolateral compartments was read with a FlexStation 3 microplate reader (Molecular Device, LCC) at excitation/emission 428/540 nm for LY, and 495/519 nm for Alexa Fluor™ 488 IgG. Apparent permeability (Papp) was calculated in cm/s according to the following equation: Papp (cm/s) = VB/(ACAO) × (ΔCB/ΔT), where VB is the volume in the basolateral chamber (0.6 cm3), A is the surface area of the filter (0.33 cm2), CAO is the initial concentration in the apical chamber, and ΔCB/ΔT is the change of concentration in the basolateral chamber over time. Under dynamic conditions, VB and A were adjusted accordingly (0.04 cm3 and 0.005 cm2, respectively).

For hyperosmotic BBB opening, iPSC-derived BMEC were pre-incubated with 1.4 M mannitol (M4125, Sigma-Aldrich, Germany) at 37 °C for 15 min followed by perfusion of 500 µg/ml LY solution for up to 6 h. The effect of mannitol on the barrier was monitored by fluorescence imaging with a Zeiss Colibri 7 LED system and by quantification of Papp of LY as reported above.

Immunocytochemistry Analysis

For immunocytochemistry, cells were fixed with 4% paraformaldehyde for 10 min, permeabilized with 0.1% Triton X-100 for 5 min, and blocked with 3% bovine serum albumin (BSA; A2153, Sigma-Aldrich, Germany) for 1 h at RT. Cells were incubated with primary antibodies overnight at 4 °C, and then with fluorescent secondary antibodies for 60 min at RT (Table 2). Nuclei were counterstained with DAPI for 2 min (2D culture) or 15 min (3D culture) at RT. Pictures were taken with a Zeiss Colibri 7 LED system or with a confocal laser scanning microscope (Olympus, FV3000).

Table 2 List of primary antibodies used for immunofluorescence stainingGene Expression Analysis

For gene expression analysis, total RNA was extracted from at least 5 pooled biological replicates according to the manufacturer’s protocol “TRizol Reagent life technologies™”. RNA was then purified with the RNA cleanup kit (74204, Qiagen, Switzerland) and reverse transcribed to cDNA. Expression was evaluated by qRT-PCR using TaqMan probes (Thermo Fisher, Switzerland) specific for the gene of interest: ABCB1 (Hs00184500_m1), ABCC1 (Hs00219905_m1), ABCG2 (Hs01053790_m1), SLC2A1 (Hs00892681_m1), TRFC (Hs00951083_m1) and B2M (Hs00187842_m1) as the housekeeping gene. Results were reported as ΔCt values (Ct gene of interest–Ct housekeeping gene).

Functional Evaluation of P-glycoprotein and Transferrin Receptor

To assess the activity of the efflux pump, P-glycoprotein (P-gp), iPSC-derived BMEC were pre-incubated with 10 µM inhibitor, Cyclosporin A (CsA; 30024, Sigma-Aldrich, Germany), for 30 min, and subsequently exposed to 10 µM substrate, Rhodamine 123 (R123; 83702, Sigma-Aldrich, Germany), for 1 h with or without the inhibitor. Since CsA was reconstituted in ethanol, the same amount of ethanol was added to the samples without inhibitor as control. Efflux was allowed for 1 h and cells were imaged with a Zeiss Colibri 7 LED system before lysis with 1% Triton X-100 for 15 min. Fluorescence was read with a FlexStation 3 microplate reader at excitation/emission 511/534 nm.

For transferrin uptake assay, iPSC-derived BMEC were pre-incubated with 10 mg/ml unlabeled transferrin (T3309, Sigma-Aldrich, Germany) for 20 min, and subsequently exposed to 500 µg/ml Alexa Fluor™ 488 transferrin (T13342, Thermo Fisher, Switzerland) for 4 h with or without unlabeled transferrin. Fluorescence pictures were taken with a confocal laser scanning microscope.

For transferrin transcytosis assay, iPSC-derived BMEC were exposed to a mixture of 500 µg/ml unlabeled transferrin and 500 µg/ml BSA for 2 h. The amount of transferrin in the bottom channel was quantified by ELISA according to the manufacturer’s protocols (Transferrin ELISA kit, Invitrogen, EHTF, Switzerland). To quantify the amount of BSA, an ELISA assay was set up using the following antibodies: sheep polyclonal, antigen affinity purified (A10-113A, Bethyl, Switzerland) and sheep polyclonal, antigen affinity purified HRP (A10-113P, Bethyl, Switzerland). Data were generated using the SoftMax Pro software, and Papp was calculated as described above.

Investigation of Antibody Transcytosis

The integrity of the barrier was evaluated by incubating iPSC-derived BMEC with 500 µg/ml Texas Red Dextran (D182, Thermo Fisher, Switzerland) for 1 h according to the manufacturer’s protocol “Barrier Integrity Assay in the OrganoPlate®”. Leakage of Texas Red Dextran from the top to the gel channel was evaluated by qualitative imaging analysis. Chips that did not allow the passage of the dye were considered leak-tight and used for antibody transcytosis. Therefore, a mixture of 1.25 µM mouse anti-human transferrin receptor, MEM-189 (NB500-493, Novus Biological, Switzerland), and 1.25 µM sheep anti-BSA (A10-113, Bethyl, Switzerland) antibodies was perfused in the top channel according to the manufacturer’s protocol “Antibody transcytosis assay in the OrganoPlate®”. Texas Red Dextran was also added to the mixture to monitor the barrier integrity at the end of the incubation time. After 2 h, cell culture medium was collected from the bottom channel, and the amount of antibody was quantified by ELISA according to the instructions of the mouse IgG ELISA kit (ab151276, Abcam, UK) and sheep IgG ELISA kit (ab190546, Abcam, UK).

Statistical Analysis

Data are presented as median with interquartile or as mean ± SD for at least three chips per condition, unless stated otherwise. Statistics were performed using GraphPad Prism version 8 (San Diego, CA, USA) or Microsoft Excel (Version 2306). Homogeneity of variance was assessed with a two-sample F-test for variances when two groups were analyzed or with Bartlett’s or Brown-Forsythe tests when more groups were analyzed. Normality was assessed with Shapiro–Wilk test.

When normality was met, t test was applied when two groups were analyzed or Welch ANOVA or one-way ANOVA tests were applied when more groups were analyzed. When normality was not met, Mann–Whitney or Wilcoxon matched-pairs signed rank tests were applied when two groups were analyzed or Kruskal–Wallis was applied when more groups were analyzed. Longitudinal data were analyzed with RM one-way ANOVA test.

The adopted test is specified in the caption of each graph. Data were considered significant if p ≤ 0.05; exact p values and details on the statistical analysis are reported in the supplementary information.