The protocol described here highlights the major steps in the differentiating induced pluripotent stem-cell derived brain-like endothelial cells, the preparation of Neisseria meningitidis for infection, and sample collection for other molecular analyses.
Meningococcal meningitis is a life-threatening infection that occurs when Neisseria meningitidis (meningococcus, Nm) can gain access to the central nervous system (CNS) by penetrating highly specialized brain endothelial cells (BECs). As Nm is a human-specific pathogen, the lack of robust in vivo model systems makes study of the host-pathogen interactions between Nm and BECs challenging and establishes a need for a human based model that mimics native BECs. BECs possess tighter barrier properties when compared to peripheral endothelial cells characterized by complex tight junctions and elevated trans-endothelial electrical resistance (TEER). However, many in vitro models, such as primary BECs and immortalized BECs, either lack or rapidly lose their barrier properties after removal from the native neural microenvironment. Recent advances in human stem-cell technologies have developed methods for deriving brain-like endothelial cells from induced pluripotent stem-cells (iPSCs) that better phenocopy BECs when compared to other in vitro human models. The use of iPSC-derived BECs (iPSC-BECs) to model Nm-BEC interaction has the benefit of using human cells that possess BEC barrier properties, and can be used to examine barrier destruction, innate immune activation, and bacterial interaction. Here we demonstrate how to derive iPSC-BECs from iPSCs in addition to bacterial preparation, infection, and sample collection for analysis.
The blood-brain barrier (BBB), and the meningeal blood-CSF barrier (mBCSFB) are extremely tight cellular barriers that separate the circulation from the central nervous system (CNS) and are primarily comprised of highly specialized brain endothelial cells (BECs)1,2. Together, BECs maintain proper brain homeostasis by regulating nutrients and waste products in and out of the brain, while excluding many toxins, drugs, and pathogens1,2. Bacterial meningitis occurs when blood-borne bacteria are able to interact with, and penetrate the barrier formed by BECs and cause inflammation. Neisseria meningitidis (Nm, meningococcus) is a Gram-negative bacterium that colonizes the nasopharaynx of 10‒40 % of healthy individuals, but in some cases can cause serious systemic disease3. In affected individuals, Nm can gain access to the blood stream where it can cause purpura fulminans as well as penetrate BECs gaining access to the CNS causing meningitis3. Nm is a leading cause of bacterial meningitis world-wide, and despite vaccination efforts, is still a primary cause of meningitis4. Modern medical intervention, such as antibiotic treatment, have made these conditions survivable, however those affected with meningitis often are left with permanent neurological damage5,6.
Previous studies have identified bacterial factors and host signaling that contribute to Nm-BEC interactions7,8,9,10,11. The identified adhesins and invasins such as the opacity protein Opc, and type-IV pili, as well as receptors such as CD147, have been conducted on various BEC models in vitro, however these models lack many defining BBB properties7,9,11,12. Complete understanding of Nm-BEC interactions remain elusive due partially to the inability to utilize in vivo models, incomplete vaccination protection, and lack of robust human BEC models in vitro.
Modeling hBECs in vitro has been challenging due to the unique properties of BECs. Compared with peripheral endothelial cells, BECs have a number of phenotypes that enhance their barrier properties such as high trans-endothelial electrical resistance (TEER) due to complex tight junctions12. Once removed from the brain microenvironment, BECs rapidly lose their barrier properties limiting the usefulness of primary or immortalized in vitro models that only form a weak barrier12,13. The combination of the human specificity of Nm infections, lack of robust in vivo models, and challenges modeling human BECs in vitro creates a need for better models to understand the complex host-pathogen interaction between Nm and BECs. Recently using model human induced pluripotent stem cell (iPSC) technologies BEC-like cells have been derived from iPSCs that better mimic BECs in vivo12,13,14,15. iPSC-BECs are of human origin, easily scalable, and possess expected BEC phenotypes compared to their primary or immortalized counterparts12,13,14,15. Additionally we and others have demonstrated that iPSC-BECs are useful for modeling various diseases of the CNS such as host-pathogen interaction, Huntington’s disease, and MCT8 deficiency that causes Allan-Hurndon-Dudley syndrome16,17,18,19,20,21. Here, we demonstrate how to derive iPSC-BECs from renewable iPSC sources and the infection of iPSC-BECs with Nm leading to activation of the innate immune response. We believe that this model is useful to interrogate host-pathogen interaction that is unable to be recapitulated in other in vitro models and is especially useful when examining interactions with human specific pathogens such as Nm.
NOTE: All media / reagent preparation, stem-cell maintenance, and differentiation steps are adapted from Stebbins et al.22.
1. Preparation of materials required for iPSC culture and BEC differentiation.
2. Maintenance IMR90-4 cell culture
NOTE: Here we use the IMR90-4 cell line as an example, however other induced pluripotent stem-cell lines such as CC3, CD10, CD12, DF19-9-11T, 83iCTR, 00iCTR, and CS03iCTRn2 have been successfully employed for differentiation into BECs13,14,15,16,17,23,27,28.
3. Differentiation of brain endothelial cells from human iPSCs
4. Transendothelial electrical resistance (TEER) as a measure of barrier tightness
NOTE: TEER is usually read on membrane inserts on days 9 and 10 of differentiation to confirm successful generation of barrier forming iPSC-BECs (Figure 1A).
5. Immunofluorescence (IF) staining to validate BEC phenotype
NOTE: To validate the quality of the fully differentiated and purified cells, iPSC-BEC monolayers are stained for the characteristic markers of brain endothelial cells on day 10 of the differentiation process as previously described (Figure 1B‒G) 13,14,15,16,17,19,22.
6. Preparation of bacteria and infection of iPSC-BECs
7. Innate immune activation by quantitative PCR
NOTE: Using a preferred RNA isolation, cDNA synthesis, and qPCR protocol, collect samples and run qPCR on selected cytokines.
The protocol described here is adapted from Stebbins et al. and highlights the process to differentiate iPSCs into brain-like endothelial cells that possess BBB properties, and how to utilize this model for infection studies using iPSC-BECs with Nm19,22. The iPSC-BECs, when differentiated properly, exhibit tight barrier properties measured by TEER that are often greater than 2000 Ω·cm2, and express endothelial markers such as VE-cadherin and CD31 (PECAM) (Figure 1A‒C). Additionally, they express and localize the tight junction markers Claudin-5, Occludin, and ZO-1 (Figure 1D‒F), and transporters such as Glut-1 (Figure 1G). Upon infection with Nm, iPSC-BECs respond to infection through the upregulation of neutrophilic proinflammatory cytokines as measured by qPCR such as IL-8 (CXCL8), CXCL1, CXCL2, CCL20, and IL6 (Figure 2A‒E). These representative results demonstrate how to ensure that iPSC-BECs are being differentiated reliably, and how to examine the response of iPSC-BECs to Nm infection.
Figure 1: Characterization of iPSC-BECs. (A) TEER of two separate, individual differentiations, read on days 9–12. Data presented as mean of triplicates. Error bars represent ± SD. (B‒G) Representative immunofluorescence data showing expression of endothelial cell markers VE-cadherin (B) and PECAM-1 (CD31; C), tight junction components Claudin-5 (D), Occludin (E), and ZO-1 (F), and glucose transporter GLUT-1 (G). Scale bar represents 50 μm. Panels B‒G of this figure have been used with permission from Kim et al. originally published in Fluids and Barriers of the CNS, a BMC journal17. Please click here to view a larger version of this figure.
Figure 2: Upregulation of cytokines by iPSC-BECs upon infection with Neisseria meningitidis. Representative qPCR data showing relative expression of CXCL8 (A), CXCL1 (B), CXCL2 (C), CCL20 (D), and IL6 (E) transcripts after 8 h of infection, comparing infected with uninfected iPSC-BEC monolayers. Data presented as mean of three independent experiments conducted in triplicate. Error bars represent ± S.D. Student’s t-test used to determine significance. *p < 0.05; **p < 0.01; ***p < 0.001. This figure has been modified and used with permission from Martins Gomes et al. originally published in Frontiers in Microbiology19. Please click here to view a larger version of this figure.
Supplementary Figure 1: IMR90-4 cells at different stages in the BEC differentiation process. Representative images of maintenance IMR90-4 culture ready to be passaged (A) and cells at different stages in the BEC differentiation process: (B) after seeding with ROCK inhibitor (day -2), (C) at the start of differentiation (day 0), (D) first day of confluence (day 3), (E) end of UM phase (day 6), (F) before BEC purification (day 8), and (G and H) after BEC purification (day 9 and day 10). Scale bar represents 500 µm. Please click here to view a larger version of this figure.
Modeling BECs and the BBB has had challenges, as primary and immortalized human BECs, in vitro, tend to lack robust barrier phenotypes. The advent of human stem cell technologies has allowed for the generation of iPSC derived BEC-like cells that retain expected hallmark BBB phenotypes such as endothelial markers, tight junction expression, barrier properties, response to other CNS cell types, and functional efflux transporters12,13,14,15, 22, 24, 25. This has enabled researchers to utilize BECs in vitro that closely mimic in vivo BECs and model various diseases with reported BBB dysfunction16,17,19,20,21,32. Nm is a leading cause of bacterial meningitis and is a human specific pathogen that lacks robust in vivo models19. This limitation has necessitated the use of better engineered models to drive the discovery of novel host-pathogen interaction between Nm and the BBB. Recently, we have demonstrated that iPSC-BECs are a viable model to interrogate Nm-BEC interaction19.
Here we describe a general method to derive iPSC-BECs and infect with Nm resulting in the upregulation of proinflammatory cytokines that are typically induced by bacterial infection19. For the derivation of iPSC-BECs we generally follow the protocol as described in Stebbins et al. for the generation of iPSC-BECs, with minor modifications22. In particular here we use StemFlex media instead of mTesR1, however either media can be used for the maintenance of the stem cell culture17. It has been established that this protocol works well with many iPSC lines, however it is important that the optimum seeding density is determined for each individual iPSC line15, 24. For this manuscript we used the IMR90-4 cell line and it was previously established that 1 x 105 cells/cm2 was the optimum initial seeding density24. Finally as a demonstration of the identity of BECs generated, iPSC-BECs express expected endothelial cell markers and tight junctions while exhibiting high TEER (Figure 1)13,14,15,24. These phenotypes, as well as being of human origin, make iPSC-BECs a powerful tool to interrogate Nm-BEC interaction.
The preparation of Nm for infection was adapted from previously published methods19,33. To ensure that the bacterial growth media was not introduced into the cell culture experiments, a washing step and resuspension in PBS was conducted as described in the methods. Finally, an MOI of 10 had been previously observed to result in the activation of iPSC-BECs through an upregulation of pro-inflammatory cytokines19. Activation of BECs in response to various bacteria have been observed namely through the upregulation of neutrophilic chemokines and cytokines6. It has been previously observed that iPSC-BECs upregulate the potent neutrophil chemoattractants IL-8, CXCL1, and CXCL2 after infection with Group B Streptococcus, and Nm16,19. This observed response of iPSC-BECs demonstrate that these cells are able to detect bacteria and activate an innate immune program resulting in the upregulation of cytokines. The methods to detect the upregulation of these cytokines by qPCR are well established and are briefly described above. However interestingly, while these pro-inflammatory cytokines are detected by qPCR, the coordinate protein products are either undetected or very low16,19. At present, it is unclear if this is an artifact of the iPSC-BEC models, or if the observed low abundance of cytokines is biologically relevant. Future research will be required to determine a mechanism behind the disconnect between expression and secretion.
A major strength of the iPSC-BEC model is the expression and localization of tight junctions that contribute to barrier function as read be TEER12,13,14,15,22. Previous work with Streptococcus agalactiae (Group B Streptococcus, GBS) has demonstrated that the upregulation of Snail1 contributes to the destruction of BBB tight junctions in vitro and in vivo34. More recently, this finding was confirmed in the iPSC-BEC model both with GBS and Nm suggesting a mechanism for how bacteria are able to disrupt BBB integrity during infection16,19. Additionally, it was demonstrated that Nm interacts with CD147 on endothelial cells that promote bacterial attachment, and ultimately reorganization of tight junctions leading to barrier dysfunction9. We have demonstrated that Nm colocalizes with CD147 in the iPSC-BECs potentially making this model ideal for the future elucidation of Nm-CD147 interactions as they pertain to BBB dysfunction19.
The method presented here demonstrates the differentiation of iPSC-BECs from a pluripotent stem cell source, and application with Nm infection. The iPSC-BECs are of human origin, express endothelial markers, and possess BBB specific phenotypes making them an ideal model for the examination of human specific pathogens such as Nm. Finally, we are able to demonstrate that the iPSC-BEC model respond to bacterial infection through the upregulation of a neutrophilic cytokine response. Taken together, the iPSC-BEC model has certain advantages over primary and immortalized model systems to examine the host-pathogen interactions at the BBB. Further work should be aimed at elucidating mechanisms of BBB destruction during bacterial meningitis.
The authors have nothing to disclose.
L.M.E. is supported by the DFG research training program GRK2157 entitled “3D Tissue Models for Studying Microbial Infections by Human Pathogens” awarded to A. S-U. B.J.K. is supported by a postdoctoral fellowship by the Alexander von Humboldt Foundation. Additionally, we acknowledge Lena Wolter for her technical assistance in the generation of the iPSC-BECs in culture.
Accutase (1x) | Sigma | A6964 | Enzymatic cell dissociation reagent |
Acetic acid | Sigma | A6283 | |
All-trans retinoic acid (RA) | Sigma | R2625 | |
Anti-CD31 (PECAM-1) | Thermo Scientific (Labvision) | RB-10333 | |
Anti-Claudin-5 | Invitrogen | 4C3C2 | |
Anti-Glut-1 | Thermo Scientific (Labvision) | SPM498 (MA5-11315) | |
Anti-Occludin | Invitrogen | 33-1500 | |
Anti-VE-cadherin | Santa Cruz | sc-52751 | |
Anti-ZO-1 | Invitrogen | 33-9100 | |
Bacto Proteose Peptone | BD | 211684 | |
b-Mercaptoethanol | Merck (Sigma-Aldrich) | 805740 | |
Cell culture plates and flasks | Sarstedt | ||
Centrifuge (Heraeus Megafuge 1.0R) | Thermo Scientific | ||
Class II biosafety cabinet | Nuaire | NU-437-400E | |
CO2 Incubator (DHD Autoflow CO2 Air-Jacketed Incubator) | Nuaire | ||
Collagen IV | Sigma | C5533 | |
Columbia ager + 5 % sheep blood | Biomerieux | 43049 | |
Costar Transwell polyester filters (12- or 24-well) | Corning | 3460, 3470 | |
D(+)-Glucose | Merck (Sigma-Aldrich) | G8270 | |
DAPI | Invitrogen | D1306 | |
DMEM/F12 | Gibco | 31330-038 | |
DMSO | ROTH | A994.1 | |
Dulbecco's phosphate-buffered saline (DPBS) | Gibco | 21600-069 | |
Epithelial Volt-Ohm Meter (Millicell ERS-2) with STX electrode | Merck (Millipore) | MERS00002 | |
Fe(NO3)3 | ROTH | 5632.1 | |
Fibronectin | Sigma | F1141 | |
Fluoresence microscope (Eclipse Ti) | Nikon | ||
Hemacytometer (Neubauer) | A. Hartenstein | ZK06 | |
Human basic fibroblast growth factor (bFGF) | PeproTech | 100-18B | |
Human Endothelial Serum Free Medium (hESFM) | Gibco | 11111-044 | |
Inverted microscope (Wilovert) | Hund (Will Wetzlar) | ||
iPS(IMR90)-4 cells | WiCell | ||
Kellogg's supplement | To prepare 110 ml of Kellogg's supplement, prepare 100 ml of 4 g/ml glucose, 0.1 g/ml glutamine, and 0.2 mg/ml thiamine pyrophosphate and 10 ml of 5 mg/ml Fe(NO3)3 and combine the solutions. Filter sterilize and store aliquoted at -20 °C. | ||
Knockout serum replacement (KOSR) | Gibco | 10828-028 | |
L-glutamine (GlutaMAX) | Invitrogen | 35050-038 | |
LunaScript RT SuperMix Kit | NEB | E3010L | cDNA synthesis kit |
Matrigel Matrix | Corning | 354230 | |
Methanol | ROTH | 4627.5 | |
MgCl2 | ROTH | KK36.1 | |
Micropipettes (Research Plus) | Eppendorf | ||
NaHCO3 | ROTH | 6329 | |
Nonessential amino acids (NEAA) | Gibco | 11140-035 | |
NucleoSpin RNA isolation kit | Machery-Nagel | 740955 | RNA isolation kit |
Pipette boy (Accu-Jet Pro) | Brand | ||
Platelet poor plasma-derived serum, bovine (PDS) | Fisher | 50-443-029 | |
PowerUp SYBR Green Master Mix | Applied Biosystems | A25742 | qPCR master mix |
qPCR film (MicroAmp Optical Adhesive Film) | Applied Biosystems | 4211971 | |
qPCR plates (MicroAmp Fast 96-well) | Applied Biosystems | 4346907 | |
ROCK inhibitor, Y27632 dihydrochloride | Tocris | 1254 | |
RT-PCR thermo cycler (StepOnePlus) | Applied Biosystems | 4376600 | |
Serological pipettes | Sarstedt | ||
StemFlex basal medium + 50x StemFlex supplement | Gibco | A3349401 | Stem-cell maintenance medium |
Swinging Bucket Rotor (Heraeus #2704) | Thermo Scientific | ||
Thiamine pyrophosphate | Sigma | C8754-5G | |
Trypan Blue Solution, 0.4% | Gibco | 15250061 | |
Versene | Gibco | 15040-033 | Non-enzymatic cell dissociation reagent (EDTA) |