The blood-brain barrier (BBB) has a crucial role in sustaining a stable and healthy brain environment. BBB dysfunction is associated with many neurological diseases. We have developed a 3D, stem-cell-derived model of the BBB to investigate cerebrovascular pathology, BBB integrity, and how the BBB is altered by genetics and disease.
The blood-brain barrier (BBB) is a key physiological component of the central nervous system (CNS), maintaining nutrients, clearing waste, and protecting the brain from pathogens. The inherent barrier properties of the BBB pose a challenge for therapeutic drug delivery into the CNS to treat neurological diseases. Impaired BBB function has been related to neurological disease. Cerebral amyloid angiopathy (CAA), the deposition of amyloid in the cerebral vasculature leading to a compromised BBB, is a co-morbidity in most cases of Alzheimer's disease (AD), suggesting that BBB dysfunction or breakdown may be involved in neurodegeneration. Due to limited access to human BBB tissue, the mechanisms that contribute to proper BBB function and BBB degeneration remain unknown. To address these limitations, we have developed a human pluripotent stem cell-derived BBB (iBBB) by incorporating endothelial cells, pericytes, and astrocytes in a 3D matrix. The iBBB self-assembles to recapitulate the anatomy and cellular interactions present in the BBB. Seeding iBBBs with amyloid captures key aspects of CAA. Additionally, the iBBB offers a flexible platform to modulate genetic and environmental factors implicated in cerebrovascular disease and neurodegeneration, to investigate how genetics and lifestyle affect disease risk. Finally, the iBBB can be used for drug screening and medicinal chemistry studies to optimize therapeutic delivery to the CNS. In this protocol, we describe the differentiation of the three types of cells (endothelial cells, pericytes, and astrocytes) arising from human pluripotent stem cells, how to assemble the differentiated cells into the iBBB, and how to model CAA in vitro using exogenous amyloid. This model overcomes the challenge of studying live human brain tissue with a system that has both biological fidelity and experimental flexibility, and enables the interrogation of the human BBB and its role in neurodegeneration.
The blood-brain barrier (BBB) is a key microvascular network separating the central nervous system (CNS) from the periphery to maintain an ideal environment for proper neuronal function. It has a critical role in regulating the influx and efflux of substances into the CNS by maintaining metabolic homeostasis1,2,3,4, clearing waste4,5,6, and protecting the brain from pathogens and toxins7,8.
The primary cell type of the BBB is the endothelial cell (EC). Endothelial cells, derived from the mesoderm lineage, form the walls of the vasculature1,9. Microvascular ECs form tight junctions with each other to greatly decrease the permeability of their membrane10,11,12,13,14 while expressing transporters to facilitate the movement of nutrients into and out of the CNS1,4,12,14. Microvascular ECs are encircled by pericytes (PCs)-mural cells that regulate microvascular function and homeostasis and are critical for regulating the permeability of the BBB to molecules and immune cells15,16,17. The astrocyte, a major glial cell type, is the final cell type comprising the BBB. Astrocyte end-feet wrap around the EC-PC vascular tubes while the cell bodies extend into the brain parenchyma, forming a connection between neurons and vasculature1. Distinct solute and substrate transporters are localized on astrocyte end-feet (e.g., aquaporin 4 [AQP-4]) that have a critical role in BBB function18,19,20,21.
The BBB is critical in maintaining proper brain health function, and dysfunction of the BBB has been reported in many neurological diseases, including Alzheimer's disease (AD)22,23,24,25, multiple sclerosis7,26,27,28, epilepsy29,30, and stroke31,32. It is increasingly recognized that cerebrovascular abnormalities play a central role in neurodegeneration, contributing to increased susceptibility to ischemic and hemorrhagic events. For example, more than 90% of AD patients have cerebral amyloid angiopathy (CAA), a condition characterized by the deposition of amyloid β (Aβ) along the cerebral vasculature. CAA increases BBB permeability and decreases BBB function, leaving the CNS vulnerable to ischemia, hemorrhagic events, and accelerated cognitive decline33.
We recently developed an in vitro model of the human BBB, derived from patient-induced pluripotent stem cells, which incorporates ECs, PCs, and astrocytes encapsulated in a 3D matrix (Figure 1A). The iBBB recapitulates physiologically relevant interactions, including vascular tube formation and localization of astrocyte end-feet with vasculature24. We applied the iBBB to model the susceptibility of CAA mediated by APOE4 (Figure 1B). This enabled us to identify the causal cellular and molecular mechanisms by which APOE4 promotes CAA, and leverage these insights to develop therapeutic strategies that reduce CAA pathology and improve learning and memory in vivo in APOE4 mice24. Here, we provide a detailed protocol and video tutorial for reconstructing the BBB from human iPSCs and modeling CAA in vitro.
1. Differentiating iPSCs into iBBB cells
NOTE: These differentiation protocols have been previously described in Mesentier-Louro et al.34.
2. Assembling the iBBB
3. Inducing cerebral amyloid angiopathy (CAA) with Aβ fibrils
4. Fixing and staining the iBBB
A properly formed iBBB solidifies into a single translucent disc (Figure 3A). It is normal for the iBBB to detach from the surface onto which it was first pipetted after a few days. This cannot be avoided, but is not a major concern to the proper formation of the iBBB if care is taken when changing media to not accidentally aspirate the iBBB. After 24 h, evenly distributed, single cells can be identified under a brightfield microscope (Figure 3B). After 2 weeks, more distinct structures may be visible, although it is difficult to make out with definition (Figure 3C).
The quality of the iBBB is highly dependent on the quality of the differentiation of the input cells. We highly recommend plating some left-over cells in 2D monocultures to fix and stain for cell-type specific markers. iBBBs formed from endothelial cells that are over 70% positive for PECAM1 or VE-cadherin (Figure 2A), pericytes over 95% positive for PDGFRB (Figure 2B), and astrocytes over 95% positive for S100B and CD44 (Figure 2C) are best for successful iBBBs. The results from this quality check are the earliest indicator for high-quality iBBBs.
Physiologically relevant 3D structures should form after 2 weeks of self-assembly. Upon fixing and staining, we see evidence of tube-like structures that stain positive for the endothelial marker PECAM1, which is critical for tight junction formation (Figure 4A). The greatest variability in iBBB formation is the extent of microvasculature formation. In a "worst-case" scenario, the endothelial cell network appears more fragmented or does not extend throughout the iBBB, while in the "best-case" scenario, the vasculature is uniform and branches throughout the iBBB. Endothelial cell differentiation that is over 70% PECAM1-positive forms more consistent networks. Additionally, aquaporin-4, a protein expressed by astrocytes that localizes to the astrocyte endfeet, aligns with PECAM1 staining, indicating that astrocytes extend their endfeet to contact the endothelial cells (Figure 4B). Finally, we expect to see pericytes around the vasculature (Figure 4C).
The primary readout for cerebral amyloid angiopathy (CAA) in iBBBs is the presence of amyloid-β (Aβ). Aβ can be measured using targeted antibodies or by seeding fluorescently labeled Aβ to induce CAA phenotypes (Figure 5). Treatment of iBBBs with Aβ should increase Aβ staining intensity and area, as the cells of the BBB do not express a lot of endogenous Aβ protein. Alternatively, fixed samples can be stained with Thioflavin T to detect amyloid accumulation. Aβ levels are dependent on the genotype of the stem cells used to generate the iBBB, with some Alzheimer's disease and CAA-associated risk factors increasing the amount of Aβ accumulation and staining (Figure 5B, C)24.
Figure 1: An in vitro blood-brain barrier to model cerebral amyloid angiopathy. (A) Schematic representation of iBBB assembly and maturation. After the differentiation of endothelial cells, astrocytes, and pericytes from induced pluripotent stem cells, cells are encapsulated in a gel matrix in a single-cell suspension. Over the course of 2 weeks, the cells self-assemble into vascular units, similar to the structures identified in vivo. (B) Schematic of the cerebral amyloid angiopathy assay. Amyloid-β is added to matured iBBBs for 96 h to induce Aβ aggregation. (C) Diagram showing the side view of an iBBB after seeding in a glass-bottomed well. (D) A graphic of an iBBB prepared for imaging on an inverted microscope. Abbreviations: BBB = blood-brain barrier; iPSCs = induced pluripotent stem cells; iBBB = an in vitro model of the human BBB derived from a patient iPSC-derived BBB model, which incorporates endothelial cells, pericytes, and astrocytes encapsulated in a 3D matrix; Aβ = Amyloid-β. Please click here to view a larger version of this figure.
Figure 2: Validation of iPSC-derived iBBB cells. (A) Representative maximum intensity projection of endothelial cells in 2D monocultures stained with VE-cadherin (green) and PECAM1 (red). (B) Representative images of pericytes in 2D monocultures stained with PDGFRB (green) and NG2 (red). (C) Representative images of astrocytes in 2D monocultures stained with CD44 (top; green) and S100B (bottom; green). All nuclei are stained with Hoechst 33342. Images were taken on a Nikon Eclipse Ti2-E at 20x magnification (A,B) or a Leica Stellaris 8 at 40x magnification (C). All scale bars are 100 µm. Abbreviations: BBB = blood-brain barrier; iPSCs = induced pluripotent stem cells; iBBB = an in vitro model of the human BBB derived from a patient iPSC-derived BBB model, which incorporates endothelial cells, pericytes, and astrocytes encapsulated in a 3D matrix; VE-cadherin = vascular endothelial cadherin; PECAM1 = platelet and endothelial cell adhesion molecule 1; PDGFRB = platelet-derived growth factor receptor beta; NG2 = neuron glial antigen 2; S100B = S100 calcium-binding protein beta. Please click here to view a larger version of this figure.
Figure 3: Assembly of the iBBB. (A) Brightfield image of a 15 µL iBBB 24 h after plating at 2x magnification. (B) Brightfield image of an iBBB 24 h after plating at 10x magnification. (C) Brightfield image of an iBBB 2 weeks after plating at 10x magnification. Scale bar = 1 mm (A), 100 µm (B,C). Images taken on an inverted Nikon Eclipse Ts2R-FL. Abbreviations: BBB = blood-brain barrier; iPSCs = induced pluripotent stem cells; iBBB = an in vitro model of the human BBB derived from a patient iPSC-derived BBB model, which incorporates endothelial cells, pericytes, and astrocytes encapsulated in a 3D matrix. Please click here to view a larger version of this figure.
Figure 4: Cell interactions in the iBBB. Representative images of endothelial cells, pericytes, and astrocytes in the iBBB. (A) Endothelial cells (PECAM1) and astrocytes (S100β). (B) Endothelial cells (PECAM1) and astrocyte endfeet (AQP-4) co-localization. (C) endothelial cells (PECAM1) and pericytes (NG2). All nuclei are stained with Hoechst 33342. Confocal Z-stack images were taken on a Leica Stellaris 8 at 20x magnification (A,B) or a Nikon Eclipse Ti2-E at 20x magnification (C). Scale bars = 200 µm (A), 100 µm (B,C). Abbreviations: BBB = blood-brain barrier; iPSCs = induced pluripotent stem cells; iBBB = an in vitro model of the human BBB derived from a patient iPSC-derived BBB model, which incorporates endothelial cells, pericytes, and astrocytes encapsulated in a 3D matrix; PECAM1 = platelet and endothelial cell adhesion molecule 1; AQP-4 = aquaporin-4; NG2 = neuron glial antigen 2. Please click here to view a larger version of this figure.
Figure 5: Cerebral amyloid angiopathy in vitro. (A) Non-AD iBBBs exposed to conditioned media from control or Aβ-overexpression neurons. 6e10 antibody recognizes Aβ. (B) Representative images of iBBBs from isogenic APOE3 and APOE4 cell lines treated with 20 nM Aβ-FITC1-42 for 96 h. (C) Quantification of amyloid in iBBBs from isogenic APOE3 and APOE4 cell lines treated with 20 nM Aβ-FITC1-40 or Aβ-FITC1-42.Scale bars = 50 µm (A), 10 µm (B). This figure was adapted from Blanchard et al24. Abbreviations: BBB = blood-brain barrier; iPSCs = induced pluripotent stem cells; iBBB = an in vitro model of the human BBB derived from a patient iPSC-derived BBB model, which incorporates endothelial cells, pericytes, and astrocytes encapsulated in a 3D matrix; Aβ = Amyloid-β. APOE = ApolipoproteinE. Please click here to view a larger version of this figure.
Markers | Company | Catalog Number | Dilution | |
Endothelial Cells | PECAM1 (CD31) | R&D Systems | AF806 | 1:500 |
VE-cadherin (CD144) | R&D systems | AF938 | 1:500 | |
ZO-1 | Invitrogen | MA3-39100-A488 | 1:500 | |
Pericytes | PDGFRβ | R&D Systems | AF385 | 1:500 |
NG2 | Abcam | ab255811 | 1:500 | |
Astrocytes | S100β | Sigma-Aldrich | S2532-100uL | 1:500 |
CD44 | Cell Signaling Technology | 3570S | 1:400 | |
AQP-4 | Invitrogen | PA5-53234 | 1:300 | |
GFAP | ||||
ALDH1L1 | ||||
EAAT1 | ||||
EAAT2 | ||||
Amyloid-β | 6e10 | Biolegend | SIG-39320 | 1:1,000 |
Thioflavin T | Chem Impex | 22870 | 25 µM |
Table 1: Recommended cell markers for differentiation quality checks. Cell markers for the different cell types of the BBB that can be used to check the quality of the differentiations and to identify the cells in the formed iBBB. The markers used in this paper are bolded.
BBB dysfunction is a co-morbidity, and potentially, a cause or exacerbating factor in numerous neurological diseases7,40,41. However, it is nearly impossible to study the molecular and cell biology underlying the dysfunction and breakdown of the BBB in humans with neurovascular disease. The inducible-BBB (iBBB) presented in this protocol provides an in vitro system that recapitulates important cell interactions of the BBB, including vascular tube formation and localization of astrocyte end-feet with vasculature. The iBBB can be used to study the molecular pathways involved at any stage of BBB dysfunction and can model neurovascular disease phenotypes, such as amyloid-β aggregation, as seen in cerebral amyloid angiopathy.
The assembly of the iBBB is straightforward, although the quality of the iBBBs is highly dependent on the quality of the iPSC-derived cells that are used. While the iBBB provides a multi-cellular niche that can promote the maturation of the component cells, each cell type needs to be properly patterned before being encapsulated. It is critical to perform quality checks on each differentiation by performing immunofluorescence staining for cell-type specific markers on the individual monocultures (Table 1). Every iPSC line behaves differently and some conditions, such as seeding density or the number of days in each patterning medium, might need to be determined empirically and adjusted to optimize differentiation efficiency.
The described protocol suggests a 50 µL size, but the iBBB can be scaled down to as small as 5 µL, depending on cell availability, the number of desired iBBBs, and the downstream application. Larger iBBBs contain more cells, making them ideal for protein, lipid, or nucleic acid collection. Smaller iBBBs can facilitate drug screenings or other scalable assays.
The iBBB is a very versatile tool for studying the BBB in vitro. Because each cell type is differentiated independently, iBBBs can be assembled from different genetic backgrounds, allowing us to study how genetic risk factors affect specific cell types to contribute to BBB dysfunction. This strategy was applied to show that APOE4, the most common risk factor for Alzheimer's disease and cerebral amyloid angiopathy, acts in part through a pathogenic mechanism specifically in pericytes24. Further use of this tool would allow us to dissect the individual contributions of ECs, PCs, and astrocytes in maintaining BBB integrity and how each cell type falters during disease development.
Currently, the most common method of modeling the BBB in vitro is by using a transwell system, seeded with ECs, and sometimes co-cultured with pericytes and/or astrocytes, to form an impermeable monolayer42,43,44. The 3D structure of the iBBB enables the self-assembly of the vascular unit and allows the formation of tubular structures that more resemble physiological vasculature. A drawback of seeding iBBBs in a well-plate, as described in this protocol, is they do not experience the sheer stresses caused by constant flow vasculature in an in vivo system. To overcome this, the iBBB can be seeded into a microfluidic chip that can generate a dynamic flow45,46,47. This system can also be used to test vasculature permeability and perfusion of small molecules.
In conclusion, this method provides a flexible, 3D, iPSC-derived model of the BBB that can be used as a platform to study countless aspects of the BBB on a cellular level, including the ability to recapitulate neurovascular disease phenotypes and the potential to screen drug BBB permeability for a wide range of applications.
The authors have nothing to disclose.
This work is supported by NIH 3-UG3-NS115064-01, R01NS14239, Cure Alzheimer's Fund, NASA 80ARCO22CA004, Chan-Zuckerberg Initiative, MJFF/ASAP Foundation, and Brain Injury Association of America. C.G. is supported by NIH F31NS130909. Figure 1A was created with BioRender.com.
6e10 amyloid-β antibody | Biolegend | SIG-39320 | Used at 1:1000 |
Accutase | Innovative Cell Technologies | AT104 | |
Activin A | Peprotech | 20-14E | |
Alexa Fluor 488, 555, 647 secondary antibodies | Invitrogen | Various | Used at 1:1000 |
Amyloid-beta 40 fibril | AnaSpec | AS-24235 | |
Amyloid-beta 42 fibril | AnaSpec | AS-20276 | |
Aquaporin-4 antibody | Invitrogen | PA5-53234 | Used at 1:300 |
Astrocyte basal media and supplements | ScienCell | 1801 | |
B-27 serum-free supplement | Gibco | 17504044 | |
BMP4 | Peprotech | 120-05ET | |
CHIR99021 | Cyamn Chemical | 13112 | |
DMEM/F12 with GlutaMAX medium | Gibco | 10565018 | |
Doxycycline | Millipore-Sigma | D3072-1ML | |
FGF-basic | Peprotech | 100-18B | |
Fluoromount-G slide mounting medium | VWR | 100502-406 | |
Forskolin | R&D Systems | 1099/10 | |
GeltrexTM LDEV-Free hESC-qualified Reduced Growth Factor Basement | Gibco | A1413302 | |
Glass Bottom 48-well Culture Dishes | Mattek Corporation | P48G-1.5-6-F | |
GlutaMAX supplement | Gibco | 35050061 | |
Hoechst 33342 | Invitrogen | H3570 | |
Human Endothelial Serum-free medium | Gibco | 11111044 | |
LDN193189 | Tocris | 6053 | |
Minimum Essential Medium Non-essential Amino Acid Solution (MEM-NEAA) | Gibco | 11140050 | |
N-2 supplement | Gibco | 17502048 | |
Neurobasal medium | Gibco | 21103049 | |
Normal Donkey Serum | Millipore-Sigma | S30-100mL | Use serum to match secondary antibody host |
Paraformaldehyde (PFA) | ThermoFisher | 28908 | |
PDGF-BB | Peprotech | 100-14B | |
PDGFRB (Platelet-derived growth factor receptor beta) antibody | R&D Systems | AF385 | Used at 1:500 |
Phosphate Buffered Saline (PBS), pH 7.4 | Gibco | 10010031 | |
Pecam1 (Platelet endothelial cell adhesion molecule 1) antibody | R&D Systems | AF806 | Used at 1:500 |
Penicillin-Streptomycin | Gibco | 15140122 | |
PiggyBac plasmid (PB_iETV2_P2A_GFP_Puro) | AddGene | Catalog #168805 | |
S100B antibody | Sigma-Aldrich | S2532-100uL | Used at 1:500 |
SB43152 | Reprocell | 04-0010 | |
Thioflavin T | Chem Impex | 22870 | Used at 25uM |
Triton X-100 | Sigma-Aldrich | T8787-250mL | |
VE-cadherin (CD144) antibody | R&D systems | AF938 | Used at 1:500 |
VEGF-A | Peprotech | 100-20 | |
Y27632 | R&D Systems | 1254/10 | |
ZO-1 | Invitrogen | MA3-39100-A488 | Dilution = 1:500 |