Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2014 Sep 4;9(9):e106346.
doi: 10.1371/journal.pone.0106346. eCollection 2014.

In vitro modeling of the neurovascular environment by coculturing adult human brain endothelial cells with human neural stem cells

Affiliations

In vitro modeling of the neurovascular environment by coculturing adult human brain endothelial cells with human neural stem cells

Chung-Hsing Chou et al. PLoS One. .

Erratum in

Abstract

Brain and vascular cells form a functionally integrated signalling network that is known as the neurovascular unit (NVU). The signalling (autocrine, paracrine and juxtacrine) between different elements of this unit, especially in humans, is difficult to disentangle in vivo. Developing representative in vitro models is therefore essential to better understand the cellular interactions that govern the neurovascular environment. We here describe a novel approach to assay these cellular interactions by combining a human adult cerebral microvascular endothelial cell line (hCMEC/D3) with a fetal ganglionic eminence-derived neural stem cell (hNSC) line. These cell lines provide abundant homogeneous populations of cells to produce a consistently reproducible in vitro model of endothelial morphogenesis and the ensuing NVU. Vasculature-like structures (VLS) interspersed with patches of differentiating neural cells only occurred when hNSCs were seeded onto a differentiated endothelium. These VLS emerged within 3 days of coculture and by day 6 were stabilizing. After 7 days of coculture, neuronal differentiation of hNSCs was increased 3-fold, but had no significant effect on astrocyte or oligodendrocyte differentiation. ZO1, a marker of adherens and tight junctions, was highly expressed in both undifferentiated and differentiated endothelial cells, but the adherens junction markers CD31 and VE-cadherin were significantly reduced in coculture by approximately 20%. A basement membrane, consisting of laminin, vitronectin, and collagen I and IV, separated the VLS from neural patches. This simple assay can assist in elucidating the cellular and molecular signaling involved in the formation of VLS, as well as the enhancement of neuronal differentiation through endothelial signaling.

PubMed Disclaimer

Conflict of interest statement

Competing Interests: Dr. John Sinden is an employee of ReNeuron Ltd. The cell lines described here are subject to patents, notably the D3 cell line is covered under patent WO200605679A1 and the STROC05 cell line is covered under patent US7666672. These cells are available through Material Transfer Agreement. This does not alter the authors’ adherence to all the PLOS ONE policies on sharing data and materials.

Figures

Figure 1
Figure 1. Establishing branching phenotype of endothelial cells and media for coculture with neural stem cells.
(A) Matrigel branching morphogenesis assay confirmed the potential for endothelial morphogenesis in D3 human cerebral microvasculature endothelial cells (hCMECs) and HMEC-1 dermal EC lines. (B) D3 hCMECs in EC medium formed a cobblestone-like monolayer on collagen-coated surfaces. D3 hCMECs in NSC medium started to detach from culture surfaces, and STROC05 human neural stem cells (hNSCs) in EC medium lost bipolar elongated morphology. In comparison, morphology of D3 hCMECs and STROC05 hNSCs was maintained in a 50∶50 mix of EC:NSC medium. Diamidino-2-phenylindole (DAPI, blue) serves as a nuclear counterstain. Scale bars represent 200 µm.
Figure 2
Figure 2. Protocols for EC/NSC coculture.
(A) Schematic description of the protocols for monoculture and coculture of D3 human cerebral microvascular endothelial cells (hCMECs) and STROC05 human neural stem cells (hNSCs). (B) 1. hCMECs (CD31+ cells) formed a dense monolayer on the collagen-coated surface, but did not form vasculature-like structures (VLS). hNSCS (polyclonal GFAP+ cells) did not form any VLS. 2. In a transwell coculture, only a dense monolayer of hCMECs without significant VLS could be found at the bottom of the lower chamber after 7 days of coculture. 3. No significant VLS was found when hCMECs were seeded in combination with hNSCs. 4. No VLS emerged when hCMECs were seeded on 7-day differentiated STROC05 cells. 5. A distinctive cytoarchitecture composed of CD31+ VLS and GFAP+ cells was observed when hNSCs were seeded on 7-day differentiated hCMECs for a further 7 days of coculture. Diamidino-2-phenylindole (DAPI, blue) serves as a nuclear counterstain. Scale bars represent 200 µm.
Figure 3
Figure 3. Immunohistochemistry of junctional markers on endothelial cells.
Localization of ZO1 expression/detection was heterogeneous, with it being only visible as part of the membrane (i.e. cellular junctions) on elongated and arranged endothelial cells (red arrows), but being mostly cytoplasmic in clumped and non-elongated cells (yellow arrows). As the cytoplasmic expression is decreased (grey arrow), it gradually shifts towards being exclusively present in the membrane.
Figure 4
Figure 4. Quantification of endothelial morphogenesis.
(A) A distinctive neurovascular cytoarchitecture emerged in which hCMECs (CD31+) formed vasculature-like structures (VLS) resembling a vascular network in between patches of hNSCs (polyclonal GFAP+). (B) The efficiency of VLS formation was quantified by measuring the length of segments between VLS branching points. (C) In a few samples, singular capillary-like structures comprising single layers of CD31+ cells between which a lumen-like space formation was observed. Diamidino-2-phenylindole (DAPI, blue) serves as a nuclear counterstain. Scale bars represent 50 µm.
Figure 5
Figure 5. Comparison of efficiency to form vasculature-like structures.
Total length of VLS in hCMEC/hNSC coculture was more efficient than any of the other culture conditions, including hCMEC monoculture (hCMEC only), the transwell coculture (hCMEC-hNSC), seeding of hCMECs and hNSCs simultaneously (hCMEC:hNSC), and seeding hCMECs on differentiated hNSCs (hNSC/hCMEC). Data points on the graph represent the median with bars reflecting the value range (post-hoc pairwise comparisons: * p<.05; **p<.01).
Figure 6
Figure 6. 3-dimensional cytoarchitecture of vasculature-like structures.
(A) A 3-dimensional cytoarchitecture composed of hCMECs (CD31+ cells in red) and hNSCs (polyclonal GFAP+ cells in green) was visualized using confocal microscopy. (B) At the base of the coculture, hNSCs were seen in a monolayer on which hCMECs formed VLS that was characterized by a multicellular organization that extended beyond the monolayer to form a unique structure. (C) Some hNSCs with an astrocytic phenotype were associated with these VLS by forming a layer of cells around the tubular walls in some cases with an elongated morphology and endfeet on the VLS (yellow arrows). The astrocytes were ensheathing the VLS. (D) A higher magnification 3D confocal image clearly demonstrates structural differences in astrocytes’ morphology upon interfacing with the VLS rather than being inside the neural patch. (E) A 3-dimensional cut through this area further indicates that astrocytes provide structural support (white arrows) to the VLS. The orange line indicates the monolayer formed by hNSCs onto which the VLS rests. Nevertheless, no hollow lumen is formed by the VLS. It hence here remains unclear if a 3-dimensional support structure is required to create a lumen and if they are inflatable in a flow system. Diamidino-2-phenylindole (DAPI, blue) serves as a nuclear counterstain. Scale bars represent 100 µm.
Figure 7
Figure 7. Cytoarchitectural characterization of vessel-like structures.
(A) Astrocytes (monoclonal GFAP+ cells) form a layer of support around endothelial cells that organize into tubular structures readily identified by phalloidin. (B) Tubular structures are tightly packed with endothelial cells expressing intercellular adhesion molecule-2 (ICAM-2), although there is a degree of inconsistency in along the VLS indicating different stages of development/maturity. (C) Endothelial cells organized into VLS are polarizing as indicated by the expression of the apical marker podocalyxin that influences astrocytes positioning of endfeet. (D) Endothelial cells within the VLS also express markers indicative of tight junctions, such as Claudin-5 and Occludin.
Figure 8
Figure 8. Deposition of a basement membrane and presence of tight junction molecules.
Characteristic extracellular matrix molecules, such as laminin, vitronectin, collagen I and IV, delineate the basement membrane which separates the vascular-like structures from the neural environment.
Figure 9
Figure 9. Time-lapse of vasculature-like structure formation.
Time-lapse images revealed the gradual development of a neurovascular cytoarchitecture in hCMEC/hNSC coculture. Arrowheads indicate the border of VLS that became prominent after coculture for 72 h. This solid cytoarchitecture was fixed after an observation period of 168 h. The contrast at borders between hCMECs (CD31+ in red) and hNSCs (polyclonal GFAP+ in green) areas was enhanced computationally. Diamidino-2-phenylindole (DAPI, blue) serves as a nuclear counterstain. Scale bar, 200 µm.
Figure 10
Figure 10. Time course of endothelial morphogenesis.
Immunocytochemistry using antibodies against CD31 (red) and polyclonal GFAP (green) revealed vascular morphogenesis in hCMEC/hNSC coculture with different culture durations of 1 to 7 days. The efficiency of endothelial morphogenesis that produced vasculature-like structures, where hCMECs (CD31+) align to form rods, was assessed by measuring the length of these rods (white lines), as well as the number of branching points between these for each image. Based on these quantifications, it was evident that total length of all individual rods and the number of branching points increased over 7 days in a linear fashion. A linear regression allowed the calculation of the slope of this progression and afforded a statistical comparison between both to indicate a significant difference in slope between VLS length and branching points. Diamidino-2-phenylindole (DAPI, blue) serves as a nuclear counterstain. Scale bar, 100 µm. Data points on the graph represent the median with bars reflecting the value range.
Figure 11
Figure 11. Differentiation status in hCMEC/hNSC coculture.
To determine the phenotypic effects of the coculture of hCMECs with hNSCs, specific markers relevant to the differentiation of hNSCs into neurons (MAP2+), astrocytes (monoclonal GFAP+) and oligodendrocytes (GalC+), as well as junction proteins in hCMECs (CD31, VE-cadherin, ZO1) were measured. There was a 3-fold increase in neuronal differentiation in coculture versus monoculture, but no significant effect on astrocytic or oligodendrocyte differentiation. Apart of ZO1, the percentage of hCMECs expressing CD31 and VE-cadherin was significantly reduced in coculture with hNSCs.

References

    1. Hawkins BT, Davis TP (2005) The blood-brain barrier/neurovascular unit in health and disease. Pharmacol Rev 57: 173–185. - PubMed
    1. Kleinfeld D, Blinder P, Drew PJ, Driscoll JD, Muller A, et al. (2011) A guide to delineate the logic of neurovascular signaling in the brain. Front Neuroenergetics 3: 1. - PMC - PubMed
    1. Moeller BJ, Cao Y, Vujaskovic Z, Li CY, Haroon ZA, et al. (2004) The relationship between hypoxia and angiogenesis. Semin Radiat Oncol 14: 215–221. - PubMed
    1. Jensen RL (2009) Brain tumor hypoxia: tumorigenesis, angiogenesis, imaging, pseudoprogression, and as a therapeutic target. J Neurooncol 92: 317–335. - PubMed
    1. Silvestre JS, Smadja DM, Levy BI (2013) Postischemic revascularization: from cellular and molecular mechanisms to clinical applications. Physiol Rev 93: 1743–1802. - PubMed

Publication types