Simple and Highly Efficient Method for Production of Endothelial Cells from Human Embryonic Stem Cells

Abstract

Endothelial cells derived from human embryonic stem cells (hESC-ECs) hold much promise as a valuable tool for basic vascular research and for medical application such as cell transplantation or regenerative medicine. Here we have developed an efficient approach for the production of hESC-ECs. Using a differentiation method consisting of a stepwise combination of treatment with glycogen synthase kinase-3β (GSK-3β) inhibitor and culturing in vascular endothelial growth factor (VEGF)-supplemented medium, hESC-ECs are induced in 5 days with about 20% efficiency. These cells express vascular endothelial cadherin (VE-cadherin), VEGF receptor-2 (VEGFR-2), CD34, and platelet endothelial cell adhesion molecule-1 (PECAM-1). These hESC-ECs can then be isolated with 95% purity using a magnetic sorting system, and expanded to more than 100-fold within a month. The hESC-ECs thus produced exhibit the endothelial morphological characteristics and specific functions such as capillary tube formation and acetylated low-density lipoprotein uptake. We propose that our methodology is useful for efficient and large-scale production of hESC-ECs.

Introduction

Human embryonic stem cells (hESCs) are able to proliferate without limit and, potentially, to differentiate into cells of any type found in the human body (16,22). Their capacity for self-renewal and pluripotency makes them an attractive possible source for the large-scale production of endothelial cells (ECs).

Recently, cell transplantation of endothelial progenitor cells (EPCs) from peripheral blood (PB) (1,8), bone marrow (BM) (9,17), or cord blood (CB) (14,15) has been shown to be a promising strategy for therapeutic angiogenesis and/or vasculogenesis in the treatment of several vascular diseases (6,7,12). It is difficult, however, to obtain sufficient numbers of EPCs from PB, BM, or CB due to the limited cell population. In addition, EPCs thus derived are usually only weakly proliferative; their potential for expansion is therefore limited (4,12).

To date, many studies have succeeded in differentiating hESCs into ECs (3,11), by a variety of methods such as mediating embryoid body (EB) formation (10) or coculturing with animal stromal cells (18,23). Yet none of these methods produces hESC-derived ECs (hESC-ECs) with enough efficiency for clinical application. EB mediation approaches, for example, require time-consuming differentiation processes and/or are relatively inefficient, while stromal cell coculturing approaches involve the risk of contamination with animal cell components, which restricts their clinical usefulness (11,12). Yet hESC-ECs have important therapeutic potential: several animal experiments have demonstrated that transplantation of hESC-ECs promotes significant improvements in vascular disease models such as mouse hindlimb ischemia (12,25). Accordingly, it is necessary to develop a new, more efficient method for inducing hESC-ECs as well as for expanding them on the large scale that is required for basic and clinical research.

In the present study, we describe a more efficient methodology for the differentiation of hESCs into ECs, their isolation for purification, and their expansion for large-scale preparation. We present this as a new useful approach for the production of hESC-ECs.

Materials and Methods

hESC Culture

The hESC lines KhES-1, KhES-2, and KhES-3 were provided by Kyoto University (20). Undifferentiated hESCs were grown on mitotically inactive mouse embryonic fibroblast feeders in primate ES medium (Repro-CELL, Tokyo, Japan) supplemented with 4 ng/ml recombinant human basic fibroblast growth factor (bFGF) (Wako Pure Chemical Industries, Ltd., Osaka, Japan). Routine culture maintenance of hESCs was performed according to the protocol recommended by Kyoto University (20). All research on hESCs was conducted in conformity with “The Guidelines for Derivation and Utilization of Human Embryonic Stem Cells (2001)” published by the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Differentiation of hESCs Into ECs

Undifferentiated hESC colonies were detached with phosphate-buffered saline (PBS) containing 1 mg/ml collagenase, 0.25% trypsin, and 20% knockout serum replacement (Invitrogen, Carlsbad, CA, USA) and dissociated into small cell clumps (about 200–500 cells). The cell clumps were plated onto type I collagen-coated dishes and cultured for 1 day in primate ES medium without bFGF. On the next day, this medium was changed to Dulbecco's Mmdified Eagle's medium/F12 medium supplemented with N2 and B27 (N2/B27 medium) (Invitrogen) containing 5 μM BIO [(2′Z,3′E)-6-bromoindirubin-3′-oxime′ (Sigma-Aldrich, St. Louis, MO, USA), a glycogen synthetase kinase-3β (GSK-3?) inhibitor (13). The cells were cultured for 3 days. Then the medium was switched to StemPro-34 serum-free medium (SFM) (Invitrogen), which is specifically formulated to support the development of human hematopoietic dells, containing 50 ng/ml recombinant human vascular endothelial growth factor 165 (VEGF165; Peprotech, London, UK). The cells were cultured subsequently in the StemPro-34 SFM containing VEGF, with medium changes every other day.

Flow Cytometry Analysis

Single-cell suspensions were obtained by dissociating cells with 0.05% trypsin/0.2 mM EDTA. These were then washed with StemPro-34 SFM, incubated at 37°C for 1 h to recover cell surface molecules, and stained with the following monoclonal antibodies: R-phycoerythrin (R-PE)-conjugated anti-human vascular endothelial cadherin (VE-cadherin) antibody (Beckman Coulter, Fullerton, CA, USA), anti-human VEGF receptor-2 (VEGFR-2) antibody (this antibody KM1668 was a generous gift from M. Shibuya, Department of Cancer Biology, Institute of Medical Science, University of Tokyo, Japan), anti-human tumor-related antigen (TRA)-1–60 antibody (Chemicon/Millipore, Billerica, MA, USA), allophycocyanin (APC)-conjugated anti-human CD34 antibody (BD Biosciences, Franklin Lakes, NJ, USA), and fluorescein isothiocyanate (FITC)-conjugated anti-human platelet endothelial cell adhesion molecule-1 (PECAM-1) antibody (eBioscience, San Diego, CA, USA). The antibody for VEGFR-2 was coupled to APC using a Phycolink APC conjugation kit and that for TRA-1–60 was coupled to R-PE using a Phycolink R-PE conjugation kit (ProZyme, San Leandro, CA, USA). Control staining with appropriate isotype-matched control monoclonal antibodies (BD Biosciences Pharmingen, San Diego, CA, USA) was performed. 7-Aminoactinomycin D (7-AAD) (BD Biosciences Pharmingen) was used to identify nonviable cells. Fluorescent-stained cells were analyzed using a FACSAria flow cytometer (BD Biosciences).

Magnetic Cell Sorting (MACS)

Single-cell suspensions from differentiated hESCs were obtained by dissociating cells with 0.05% trypsin/0.2 mM EDTA. VE-cadherin, sup>+ cells were separated by positive selection using an automated magnetic cell separator (autoMACS software program Possel-S) (Miltenyi Biotec, Bergisch Gladbach, Germany) with PE-conjugated VE-cadherin monoclonal antibodies and Anti-PE MicroBeads (Miltenyi Biotec). The entire procedure was performed according to the manufacturer's instructions. After MACS separation, a portion of the positive cell reaction was analyzed with a FACSAria flow cytometer to determine the purity of VE-cadherin⁺ cells. All remaining isolated cells were then plated onto type I collagen-coated dishes, and cultured in StemPro-34 SFM containing 50 ng/ml VEGF.

Immunocytochemistry

Cells were fixed with 4% paraformaldehyde (PFA) in PBS for 5 min at 4°C and blocked with 1% bovine serum albumin and 0.1% Tween 20 in PBS for 1 h at room temperature. The cells were then incubated with the primary antibody against VE-cadherin (mouse monoclonal IgG, 1:100, Santa Cruz Biotechnology, Santa Cruz, CA, USA) overnight at 4°C followed by the secondary antibody, Alexa488-conjugated anti-mouse IgG (1:1000, Molecular Probes/Invitrogen), for 1 h at room temperature. 4′,6-Diamino-2-phenylindole (DAPI) was used to stain cell nuclei.

Tube Formation Assay

BD Matrigel Basement Membrane Matrix (250 jil, BD Biosciences) was dispensed into each well of a 24-well tissue culture plate and allowed to gel for 30 min at 37°C. The cells (5 × 10⁴ cells per well) were seeded onto the matrix and cultured for 24 h at 37°C.

Uptake of DiI-Ac-LDL

Cells were incubated in medium containing 2.5 μg/ml 1′,1-dioctadecyl-3,3,3′,3′-tetramethylindocarbocya-nine-labeled acetylated low-density lipoprotein (DiI-Ac-LDL) (Molecular Probes/Invitrogen) for 4 h at 37°C. After incubation, cells were washed three times with PBS and fixed with 4% PFA for 10 min. Incorporation of DiI-Ac-LDL was visualized with a fluorescent microscope.

Results and Discussion

The overall strategy for production of hESC-ECs is represented in Figure 1, and the details are described in the Materials and Methods section. The method consists of three steps: 1) differentiation to ECs, 2) isolation by means of magnetic cell sorting, and 3) expansion by means of subculture.

Figure 1.

Schematic representation of the three-step procedure for producing endothelial cells from human embryonic stem cells (hESC-ECs).

Differentiation of hESCs Into ECs

Briefly, prior to differentiation, undifferentiated hESCs were dissociated into small cell clumps and plated onto type I collagen-coated dishes (day 0: undifferentiated cells). Next, to induce endothelial differentiation, the cells were cultured in N2/B27 medium with 5 μM BIO for 3 days (days 1–3: BIO treatment). The cells were subsequently cultured in StemPro-34 SFM containing 50 ng/ml VEGF (day 4 and thereafter) (Fig. 1 and Materials and Methods).

During these steps toward differentiation, the expression levels of several endothelial cell surface molecules that are known markers for endothelial differentiation of hESCs were monitored by means of flow cytometry. On day 0, the undifferentiated hESCs were positive for the undifferentiated hESC marker TRA-1–60 (Fig. 2A). There was no expression of endothelial cell markers other than partial and weak expression of VEGFR-2. It has been reported that VEGFR-2 was expressed weakly in undifferentiated ESCs, but was downregulated once during differentiation into mesoderm and reexpressed subsequently (18,19). In our study, during the 3 days of BIO treatment the population of TRA-1–60⁺ (VEGFR-2⁺/TRA-1–60⁺) cells significantly decreased (Fig. 2B), and VEGFR-2⁺/TRA-1–60^- cells emerged. This result indicates that mesodermal differentiation process is proceeding. VE-cadherin⁺ cells had not yet been observed. After the medium was changed to StemPro-34 SFM containing VEGF, a population of VE-cadherin⁺/VEGFR-2⁺ cells was induced rapidly, increased to approximately 20% (range 15–26%) of the differentiated hESCs by day 5, and gradually decreased thereafter (Fig. 2C). It is thought that this is because VE-cadherin⁺/VEGFR-2⁺ cells were less proliferative than other cells and, over time, decreased in percentage relative to the total number of cells. These induced VE-cadherin⁺/VEGFR-2+ cells coexpressed CD34 and PECAM-1 (Fig. 2D). These findings indicate that a homogeneous population of ECs was generated from differentiated hESCs with a maximum efficiency of approximately 20% on day 5.

Figure 2.

Flow cytometry analysis of the differentiation of hESCs into ECs. (A) Expression of vascular endothelial growth factor receptor-2 (VEGFR-2), tumor-related antigen (TRA)-1–60, and vascular endothelial (VE)-cadherin on day 0. The undifferentiated hESCs were positive for TRA-1–60. There was no expression of VE-cadherin other than partial and weak expression of VEGFR-2. (B) Expression of VEGFR-2, TRA-1–60, and VE-cadherin on day 3. VEGFR-2⁺/TRA-1–60⁺ cells decreased during differentiation, and the VEGFR-2⁺/TRA-1–60^- cells emerged. (C) Expression of VE-cadherin and VEGFR-2 on days 5 to 9. A population of VE-cadherin⁺/VEGFR-2⁺ cells was induced on day 5 with an efficiency of 20%. (D) Expression of VE-cadherin, CD34, and platelet endothelial cell adhesion molecule-1 (PECAM-1) on day 5. The VE-cadherin⁺ cells coexpressed CD34 and PECAM-1. (E) Control experiment without (2′Z,3′E)-6-bromoindirubin-3′-oxime (BIO) treatment. (F) Control experiment in the absence of VEGF. No induction of VE-cadherin⁺/VEGFR-2⁺ cells was observed when either BIO treatment or VEGF-supplemented culture was omitted.

At this stage of the differentiation process, the sequential usage of BIO and VEGF is an important feature of our approach. We treated hESCs with the GSK-3β inhibitor BIO to achieve directional induction of the EC lineage. This decision was based on previous findings indicating that activation of the canonical Wnt/β-catenin signaling pathway using Wnt3A or an inhibitor of GSK-3β in ESCs results in mesodermal/endodermal-specific differentiation (2,21,24).

As we have shown, treatment with BIO for 3 days induced a VEGFR-2+ mesodermal progenitor population, and subsequent culturing with StemPro-34 SFM in the presence of VEGF directed these cells to become VE-cadherin⁺ ECs. In contrast, when either BIO treatment or VEGF was omitted, no VE-cadherin⁺/VEGFR-2⁺ endothelial population was induced (Fig. 2E, F). These results show that both BIO and VEGF are necessary to efficiently promote differentiation of hESCs to an EC lineage.

To examine the generality of our differentiation method, this differentiation protocol was applied to three different hESC lines, and the time courses and efficiency levels of EC differentiation were compared. Similar results were obtained in all three hESC lines: the maximum induction for VE-cadherin⁺/VEGFR-2⁺ cells was always observed on day 5 of differentiation. The efficiency levels achieved in KhES-1, KhES-2, and KhES-3 cell lines were 17%, 15%, and 26%, respectively (data not shown).

Isolation of hESC-ECs by Means of MACS

Clinical use requires the isolation and purification of the desired cells from a mixed cell population. For this purpose, the fluorescence-activated cell sorting (FACS) and magnetic-activated cell sorting (MACS) separation techniques have been widely used (5). Although FACS is a powerful system for the isolation of subpopulations out of complex cell mixtures, its application is often limited due to its relatively small separation capacity. FACS isolation of rare cells is time consuming and laborious. In many cases it is not suitable for processing large numbers of cells (e.g., for clinical application or molecular analysis). Furthermore, FACS-isolated cells often show decreased viability due to inevitable damage by the laser treatments in the sorting processes. MACS, on the other hand, is a simple system that can be used to process much larger numbers of cells within a short period of time. But MACS is not effective in case of the purification of rare cells, constituting a small subpopulation of less than a few percent. In the present study, we have succeeded in inducing VE-cadherin⁺/VEGFR-2⁺ cells from hESCs with enough efficiency (about 20%) to make the MACS system applicable. For these reasons, we chose MACS system to isolate VE-cadherin⁺ hESC-ECs. Actually, in the course of a single round of the auto-MACS separation, within less than 30 min, more than 2 × 10⁸ VE-cadherin⁺ cells could reliably be separated from the 1 × 10⁹ differentiated hESCs harvested from 20 dishes (10 cm in diameter) on day 5 cultured in STEMPro-34 SFM containing VEGF. As we had expected, due to the higher efficiency of hESC-EC induction, good separation was achieved: the VE-cadherin⁺ cells were isolated with more than 90% recovery and more than 95% purity (Fig. 3A). No further purification steps were required. Thus, we concluded that this MACS isolation process is useful as an easier and faster method of effective and reproducible purification of hESC-ECs (VE-cadherin⁺ cells).

Figure 3.

Representative results of isolation and expansion of hESC-ECs. (A) Flow cytometry analysis of magnetic automated cell sorting (MACS) isolation of VE-cadherin⁺ cells. Purity of VE-cadherin⁺ cells after isolation is compared with that before isolation. The VE-cadherin⁺ cells were isolated with 95% purity using MACS. (B) Cell Morphology of the VE-cadherin⁺ cells (hESC-EPCs) obtained after isolation and expansion. (C) Flow cytometry analysis of the cell surface marker expression. The isolated and expanded hESC-ECs maintained the expression of VE-cadherin, VEGFR-2, CD34, and PECAM-1. Scale bar: 100 μm (B).

Expansion and Characterization of hESC-ECs

The hESC-ECs (VE-cadherin⁺ cells) isolated through MACS exhibited a morphology typical of ECs (Fig. 3B), and proliferated actively when recultured in StemPro SFM containing VEGF. They could be subcultured in vitro using a standard procedure and several passages increased the number of cells more than 100-fold (4 to 5 passages) within a month, though their proliferation rate gradually decreased with increasing passage number (7 to 8 passages). Flow cytometry analysis revealed that, the isolated and subcultured hESC-ECs at passage 5 continued to maintain their initial purity and homogeneous marker expression pattern of VE-cadherin⁺/VEGFR-2⁺/CD34⁺/PECAM-1⁺ (Fig. 3C).

Next, we performed functional assays to verify the endothelial nature of the hESC-ECs by testing for the capillary tube formation and Ac-LDL uptake functions, which are specific to ECs. When the hESC-ECs were seeded on Matrigel, they formed capillary-like tube structures (Fig. 4A), showing their tube formation ability. When they were incubated with fluorescently labeled DiI-Ac-LDL, microscopic examination revealed DiI-Ac-LDL fluorescence in the cytoplasm of the cells with perinuclear distribution (Fig. 4B), demonstrating their ability to take up Ac-LDL. Previous reports have shown that hESC-ECs resemble human umbilical vein endothelial cells (HUVECs) in EC marker expression and angiogenic potential (10,11). We compared hESC-ECs with HUVECs. The expression of VEGFR-2, VE-cadherin, and PECAM-1 in hESC-ECs was similar to HUVECs, while the expression of CD34 in hESC-ECs was relatively stronger than HUVECs (data not shown). HUVECs moreover revealed robust uptake of DiI-Ac-LDL, whereas hESC-ECs incorporated it only weakly. These seem to represent that hESC-ECs obtained in the present study are still not fully matured as well as HUVECs are.

Figure 4.

Functional characterization of hESC-ECs. (A) Capillary-like tube formation. (B) Uptake of DiI-Ac-LDL (red). Immunostaining for VE-cadherin (green) and DAPI (blue) staining of cell nuclei are also shown. Scale bars: 100 μm (A), 25 μm (B).

We also confirmed that these hESC-ECs could be cryopreserved without losing their viability and functionality to any significant degree (data not shown), suggesting that these hESC-ECs ought to store stably.

Taken together, these results demonstrate the possibility of stable large-scale production of homogeneous populations of hESC-ECs that possess the functional properties specific to ECs.

In summary, this study presents a new efficient approach to the production of hESC-ECs. This approach is a three-step procedure consisting of 1) simple and highly efficient differentiation for EC induction, 2) easy and fast isolation through MACS, and 3) stable expansion through subculture. Among these steps, the differentiation step is especially crucial: sequential treatment with 5 μM BIO and VEGF-containing medium directionally promoted the differentiation of hESCs into ECs with high efficiency (approximately 20%) and in a short time (within 5 days). It thereby allows successful isolation of hESC-ECs with high purity (greater than 95%) and their expansion (more than 100-fold within a month). Compared to those presented in previous reports, our approach is more efficient. In conclusion, the three-step strategy presented here, as a total system, could be useful for large-scale production of fully purified hESC-ECs to be used not only in basic vascular study but in various clinical applications such as cell transplantation and regenerative medicine.