SEARCH

SEARCH BY CITATION

Keywords:

  • Endothelial cells;
  • Endothelial progenitor cells;
  • Endothelial outgrowth cells;
  • Angiogenesis;
  • Vascular repair;
  • Bone marrow

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

A decade of research has sought to identify circulating endothelial progenitor cells (EPC) in order to harness their potential for cardiovascular regeneration. Endothelial outgrowth cells (EOC) most closely fulfil the criteria for an EPC, but their origin remains obscure. Our aim was to identify the source and precursor of EOC and to assess their regenerative potential compared to mature endothelial cells. EOC are readily isolated from umbilical cord blood (6/6 donors) and peripheral blood mononuclear cells (4/6 donors) but not from bone marrow (0/6) or peripheral blood following mobilization with granulocyte-colony stimulating factor (0/6 donors). Enrichment and depletion of blood mononuclear cells demonstrated that EOC are confined to the CD34+CD133CD146+ cell fraction. EOC derived from blood mononuclear cells are indistinguishable from mature human umbilical vein endothelial cells (HUVEC) by morphology, surface antigen expression, immunohistochemistry, real-time polymerase chain reaction, proliferation, and functional assessments. In a subcutaneous sponge model of angiogenesis, both EOC and HUVEC contribute to de novo blood vessel formation giving rise to a similar number of vessels (7.0 ± 2.7 vs. 6.6 ± 3.7 vessels, respectively, n = 9). Bone marrow-derived outgrowth cells isolated under the same conditions expressed mesenchymal markers rather than endothelial cell markers and did not contribute to blood vessels in vivo. In this article, we confirm that EOC arise from CD34+CD133CD146+ mononuclear cells and are similar, if not identical, to mature endothelial cells. Our findings suggest that EOC do not arise from bone marrow and challenge the concept of a bone marrow-derived circulating precursor for endothelial cells. STEM CELLS2013;31:338–348


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

The observation that peripheral blood contains a population of cells capable of forming endothelial cells in vitro and contributing to the formation of new blood vessels in vivo suggested the presence of a circulating endothelial progenitor cell (EPC) in the adult [1]. This concept challenged the traditional dogma that postnatal blood vessel growth occurred exclusively from the sprouting of existing mature blood vessels.

A decade of research has sought to identify the origin and phenotype of EPC and to harness their potential for cardiovascular regeneration. Ambiguity over the precise definition of an EPC continues to cause problems in the field [2]. The phenotypic and functional overlap between EPC, hematopoietic progenitor cells (HSC), and mature endothelial cells has contributed to this uncertainty. To date three methods have been used to isolate and culture EPC [2]. Circulating EPC have been identified and enumerated by flow cytometry by the expression of progenitor (CD34, CD133) and endothelial Vascular endothelial growth factor 2 (VEGFR2) cell markers [1]. However, there are disagreements surrounding this definition as it has been reported that CD34+CD133+VEGFR2+ cells do not give rise to endothelial cells but rather are a distinct subpopulation of HSC [3].

Alternatively, EPC have been identified and quantified by the culture of blood mononuclear cells (MNC) as early outgrowth EPCs (CFU-EPC) [4] or late outgrowth endothelial cells (EOC), also called endothelial colony forming cells [5]. EOC are thought to originate in the bone marrow together with CFU-EPC although EOC differ in that they demonstrate robust proliferative potential, express endothelial markers, but not hematopoietic or monocytic markers, and form de novo blood vessels when transplanted into immunodeficient mice [2].

While the literature points to the existence of more than one population of circulating cells supporting vascular repair and angiogenesis [6, 7], it has been suggested that only EOC are true EPC with clonogenic and proliferative potential [7, 8].

Unfortunately, two major points remain unresolved: (a) the cell markers used to identify EOC and their precursors are not able to distinguish them from circulating mature endothelial cells [9], and (b) while it has been suggested that the precursor of EOC arise from bone marrow and are mobilised in response to tissue ischemia or injury [10], the precise origin of these cells has not been defined.

Our aim was to identify the source and precursor of EOC. We sought to find a distinguishable phenotypic or functional characteristic, which was able to discriminate EOC from mature endothelial cells and to assess their regenerative potential compared to mature endothelial cells.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

Cell Sources

The study was performed with the approval of the local research ethics committee, in accordance with the Declaration of Helsinki and with the written informed consent of all participants. Venous blood samples were collected into heparin from healthy donors before and immediately following granulocyte-colony stimulating factor (G-CSF) mobilization of MNC into peripheral blood. Cord blood products were aspirated from the umbilical placental veins from normal caesarean deliveries. Bone marrow samples were obtained by aspiration from the posterior iliac crest of normal healthy donors.

Late Outgrowth and Mature Endothelial Cell Culture

EOC were cultured as described by Ingram et al. [5]. Briefly, MNC were isolated by buoyant density centrifugation over Ficoll-Paque Plus (GE Healthcare, Uppsala). MNC from bone marrow and peripheral blood (30 × 106 cells) or from cord blood (10 × 106 cells) were resuspended in endothelial growth medium (EGM-2 medium, Lonza) and plated onto type-1 rat-tail collagen-coated six-well tissue culture plates (BD Biosciences, U.K., http://www.bdbiosciences.com). The cells were incubated at 37°C, 5% CO2, 95% relative humidity for 3–4 weeks. Medium was changed every 2 days for 7 days and then twice a week until first passage. Colonies were counted after appearance. Human umbilical vein endothelial cells (HUVEC) (Lonza, U.K.) were cultured in EGM-2 medium on uncoated tissue culture flasks following the supplier's instructions.

Flow Cytometric Analysis

Cells were directly stained and analyzed for phenotypic expression of surface markers using preconjugated anti-human monoclonal antibodies (MAbs), as previously described [11]. The anti-human MAbs used included anti-CD45-FITC, anti-CD34-PerCP, anti-Tie2-APC, anti-CD102-APC, anti-CD29-APC, anti-CD49a-PE, anti-CD49b-FITC, anti-CD49c-PE, antiCD49d, anti-CD49e-PE, and anti-CD49f-FITC, anti-CD50-FITC, anti-CD54-PE, anti-CD51-FITC, anti-CD61-PE, anti-CD102-APC, anti-CD147-FITC (all from Becton Dickinson, U.K., http://www.bd.com), anti-CD133-APC (Miltenyi Biotech, U.K., http://www.miltenyibiotec.com), anti-CD31-FITC, anti-CD105-APC, anti-CD117-APC (eBioscience) anti-CD146-FITC, anti-VEGFR2-APC (R&D systems, Minneapolis, MN, http://www.rndsystems.com), fluorescein (FITC) phycoerythrin (PE) allophycocyanin (APC).

Cell Sorting

For the isolation of cells by fluorescence-activated cell sorting (FACS), cells were first enriched using anti-CD34 antibody-coupled magnetic MACS MicroBeads (Miltenyi Biotech) following manufacturer's instructions. Enriched samples were stained with fluorochrome-labeled antibodies to and finally resuspended in sterile phosphate buffered saline (PBS). Single fluorochrome stained cells were used as controls to set compensation, and appropriate fluorescence minus one stained samples were used to set selection gates for FACS sorting. Cells were sorted using a FACS Aria flow cytometer (Becton Dickinson) equipped with 488 nm and 633 nm lasers. Dead cells were excluded using an electronic side scatter (SCF) forward scatter (FSC) gate before applying sort gates to defined cell populations to be collected. A small sample of each collected fraction was run back through the cytometer to assess the success of sorting by measuring purity of the sample. Purification of subpopulations of CD34, CD133, and CD146 coexpressing cells was also performed using anti-CD34, anti-CD133, and anti-CD146 antibody-coupled magnetic MACS MicroBeads (Miltenyi Biotech) following the manufacturer's instructions. The purity of the enrichment was determined by flow cytometry.

Analysis of Aldehyde Dehydrogenase Activity

Activity of the stem cell marker aldehyde dehydrogenase (ALDH) was analyzed with Aldefluor reagent (StemCell Technologies, U.K., http://www.stemcell.com) according to the manufacturer's specifications. Aldefluor substrate (0.625 μg/ml) was added to 1 × 106 cells in Aldefluor assay buffer and incubated for 30 minutes at 37°C to induce the conversion of Aldefluor substrate to the fluorescent product. For each experiment, an aliquot of Aldefluor-stained cells was immediately quenched with 5 μl of 1.5 mM diethylaminobenzaldehyde, a specific ALDH inhibitor. After treatment cells were analyzed by flow cytometry.

Immunostaining

Cytospin preparations (5 × 104 cells) were dried and fixed in ice-cold methanol for 10 minutes. Slides were mounted in a staining rack (Shandon Sequenza). Cells on slides were washed with wash buffer (PBS/0.5% Tween-20/0.01% azide). Primary antibodies, anti-CD146 and anti-von-Willebrand factor were diluted (1:100) in wash buffer with 2% goat serum (Invitrogen, U.K., http://www.invitrogen.com) and incubated at 4°C overnight. Cells were washed followed by the addition of secondary antibodies: Alexa 488 goat anti-rabbit IgG and Alexa 555 goat anti-mouse IgG (1:200) at 37°C. After 2 hours, cells were washed and fixed with Prolong Gold anti-fade reagent with 4′,6-diamidino-2-phenylindole (DAPI) (Invitrogen). The staining was observed using Zeiss Axio-Observer A1 microscope equipped with appropriate filter combinations, and digital images were recorded on a high sensitivity Zeiss AxioCam MRm camera using Zeiss Axiovision software.

Metabolically Active Cells

The proliferative potential of EOC and HUVEC, and through different passages, was determined through a colorimetric assay to detect the number of metabolically active cells (Quick Cell Proliferation Assay Kit, Abcam, Cambridge, U.K., http://www.abcam.com). The assay is based on the cleavage of a water soluble tetrazolium salts (WST-1) to formazan by cellular mitochondrial dehydrogenases. HUVEC and EOC at different passages were plated at a density of 2 × 104 cells per well on a 96-well microtiter plate in EGM-2 medium in a final volume of 100 μl. Two wells of 100 μl of EGM-2 medium without cells were used as blanks. WST-1/ECS solution was added at 10 μl per well. All the wells with and without cells were incubated for 4 hours at 37°C. The formazan dye produced by viable cells was quantified by multiwell spectrophotometer by measuring the absorbance of the dye solution at 440 nm.

Cell Growth Kinetics

EOC and HUVEC were plated at a concentration of 1 × 104 in a 12-well collagen I-coated plate in 1 ml of EGM-2 media. At each subsequent passage, the cells were counted and plated again at the same cell concentration. Both cells were grown for nine passages, and population-doubling times (PDT) were calculated as previously described [5]. In brief, PDT was derived using the time interval between cell seeding and harvest by the number of population doublings between passages.

Single Cell Assays

Single EOC and HUVEC were plated in a 96-well flat bottom collagen I-coated plate in 200 μl of EGM-2 media. Cells were cultured as previously described [5] for 14 days changing media every 4 days. The number of cells per well were counted by visual inspection and classified in four different categories: 2–50 cells per well, 50–500 cells per well, 500–2,000 cells per well, 2,000–10,000 cells per well, and >10,000 cells per well.

Quantitative Polymerase Chain Reaction

Total RNA was extracted from HUVEC and EOC using Bio-RNA Xcell (II) (Biogene, Kimbolton, Cambridgeshire, U.K.), and DNase treated using TURBO DNase (Applied Biosystems, Warrington, U.K., http://www.appliedbiosystems.com). Reverse transcription was primed by means of random hexamers, and the polymerase was moloney murine leukemia virus (M-MLV) Reverse Transcriptase RNase H(-) Point Mutant (Promega UK Ltd, Southampton, Hampshire, http://www.promega.com), with a total volume of 50 μl per reaction. RT minus reactions were set up to act as controls for quantitative polymerase chain reaction (qPCR), and the initial cDNA reactions were tested for production of cDNA by standard PCR using primers against the Abl coding mRNA, these give a band of approximately 300 bp with cDNA, but approximately 800 bp from genomic DNA. One microliter of cDNA or RT minus control reaction was used for each qPCR reaction, and all reactions were carried out in triplicate. Detection of gene expression was by a qPCR using Taqman assay probe/primer sets for Wnt (2b,4,5a,5b,6,8a,10b,11), CD34, β catenin, DKK1, β-2-M, and GAPDH (supporting information Table 1). Control wells contained an equivalent amount of RNA as test wells without previous reverse transcription. qPCR reactions and analysis were carried out using Agilent Mx3005P (Agilent, Stockport, Cheshire, U.K., http://www.agilent.com).

Angiogenesis Array

The human angiogenesis Antibody Array G Series 1 (RayBiotech, Inc., Norcross, GA) http://www.raybiotech.com was used according to the manufacturer's instructions. Briefly, the glass array chip was incubated with 100 μl blocking buffer for 30 minutes, followed by 2 hour incubation with 100 μl of conditioned medium collected from both EOC and HUVEC cell cultures. After washing, the slides were incubated overnight at 4°C with 70 μl of biotin-conjugated antibody solution. The slides were then washed, incubated with 70 μl of IRDye-labeled streptavidin (1:2,000 dilution) for 45 minutes, and washed. The glass slide was scanned using the Odyssey Infrared Imaging System (LI-COR Biosciences, Lincoln).

In Vitro Vascular Tubule Formation Assay

HUVEC and EOC were resuspended at 1 × 105 per milliliter in EGM-2 (Lonza, U.K.). Matrigel Matrix (Becton Dickinson) solution was thawed overnight at 4°C and plastic ware was precooled. Wells were precoated with 250 μl Matrigel, which was allowed to solidify for 1 hour at 37°C before 500 μl of cell suspension was added in duplicate. Capillary structures and endothelial cell networks were examined after 24 hours of incubation at 37°C by phase contrast microscopy (40× lens) using an inverted microscope (Nikon Eclipse TS100-F). The endothelial cell network was quantified from the image fields by scoring the number of cell–cell connections.

Subcutaneous Sponge Implantation Assay for In Vivo Vascularization

Male non-obese diabetic (NOD) severe immunodeficiency genetic disorder (SCID)-IL-2gammaRnull mice aged 10–12 weeks were bred and maintained in the Biomedical Research Facility at Edinburgh University. Experimental procedures were approved by the local ethics committee and were authorized by the Home Office under the Animals (Scientific Procedures) Act 1986. Mice were anesthetized and a sterilized sponge cylinder (0.5 cm/1 cm) (Caligen Foam, Accrington, Lancashire, U.K) implanted subcutaneously on each flank. Each animal had a control vehicle-impregnated sponge (growth-factor-reduced [GFR]-Matrigel alone) implanted on one flank and cell-impregnated sponge (GFR-Matrigel and EOC or HUVEC) on the other flank. EOC or HUVECs were suspended in 200 μl PBS and 150 μl GFR-Matrigel (1 × 105 cells) and the sponges compressed into the solution. Mice were culled and the sponges excised 21 days following implantation. Sponges were fixed in 4% formalin before embedding in paraffin wax. Sections (5 μm) were stained with hematoxylin/eosin for identification of blood vessels, as described [12]. Vessel density within sponges was determined using the Chalkley count performed in triplicate in each of two sections per sponge. Sections for immunohistochemistry staining were mounted on slides and dewaxed and rehydrated using xylene and ethanol (100%–50%), respectively. The slides were then simmered in a microwave in EDTA solution pH 8.0 (Invitrogen) for 20 minutes for epitope retrieval. Slides were cooled and rinsed in tap water, mounted in a staining chamber (Shandon, Sequenza), and washed four times with wash buffer (PBS/0.5% Tween-20/0.01% azide). Image-iT FX (Invitrogen) blocking agent was added for 20 minutes at room temperature and removed by washing before the addition of 150 μl of primary antibodies diluted 1:200 in antibody diluent (wash buffer with 2% goat serum [Invitrogen]). After overnight incubation at 4°C, sections were washed four times and 150 μl of the fluorescent-labeled secondary antibody diluted 1:200 in antibody diluent were added and slides were incubated for 3 hours at 37°C. Paired anti-human specific and anti-mouse endothelial primary antibodies, including anti-CD31 and anti-CD146 were used with paired secondary antibodies; Alexa Fluor 488 goat anti-rabbit (A11070) and Alexa Fluor 488 goat anti-mouse (A11029) (Invitrogen). After incubation, slides were rinsed in tap water and mounted under coverslips using anti-fade mountant containing DAPI (Prolong Gold, Invitrogen) and stored at 4°C in the dark. Slides were examined under a fluorescent Zeiss Axio Observer A1 microscope equipped with appropriate filter combinations, and images were recorded on a Zeiss AxioCam MRm using Zeiss AxioVision 4.8 software, which was also used to process images to combine channels.

Hematopoietic, Bone, and Adipose Differentiation

Hematopoietic Colonies

MNC and purified CD34, CD133, and CD146 cells from magnetic bead separations were resuspended in 200 μl IMDM + 2% FCS and mixed with 400 μl methylcellulose (HSF-CFU complete with EPO, Miltenyi Biotech) in a 24-well plate at concentrations of 1 × 103, 1 × 104, and 1 × 105 cells per well. After 14 days, hematopoietic colonies (CFU-HSC) were quantified as per manufacturer's instructions.

Alizarin Red S Staining

Wells were rinsed twice with PBS and fixed in 70% ice-cold ethanol for 1 hour at room temperature. After incubation, cells were washed in distilled water and 0.25 ml of 10 mg/ml Alizarin red S was added per well. After 45 minutes incubation in the dark, Alizarin red S was removed, and 0.5 ml of PBS was added to each well for 30 minutes. Wells were then rinsed twice with distilled water and allowed to dry. Red staining reflecting bone formation was observed by microscopy.

Sudan Black Staining

Wells were fixed for 1 minute at 4°C in glutaraldehyde fixative solution (Sigma, U.K.) then rinsed thoroughly in deionized water before the addition of 0.3 ml of Sudan Black for 5 minutes with intermittent agitation. Wells were then rinsed three times with 70% ethanol and distilled water. Wells were counterstained with hematoxylin solution (Gill No3, Sigma, U.K.) for 3 minutes and washed thoroughly in tap water. Sudan black staining reflecting adipose tissue formation was observed by microscopy.

Statistical Analysis

Continuous variables are reported as mean ± SEM. Statistical analyses were performed with GraphPad Prism four (Graph Pad Software, USA). Two-way ANOVA, or two-tailed Student's t test, were used where appropriate. Statistical significance was taken at p < .05.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

Isolation of EOC Colonies from Different Cell Sources

Cord blood MNC (10 × 106 cells plated) generated EOC colonies from all donors (n = 6), and peripheral blood MNC (30 × 106 cells plated) generated EOC colonies from four donors of the six donors (n = 6; Fig. 1). In contrast, G-CSF mobilized peripheral blood (30 × 106 cells plated) and bone marrow-derived MNC (30 x106 cells plated) did not generate EOC colonies from any donor (n = 6 for both).

thumbnail image

Figure 1. Precursors of EOC Reside Within the CD34+CD133-CD146+ Fraction (A): Representative cell sort of cord blood mononuclear cells stained for coexpression of CD34, CD133, and CD146: characteristic scatter of all mononuclear cells with lymphocytes gated (gate one) (upper left panel); cells in gate one selected for CD34 expression (gate two) (upper right panel); cells in both gate 1 and 2 were sorted based on CD133 and CD146 expression to isolate CD34+CD133+CD146 (upper left gate), CD34+CD133CD146+(lower right gate) and CD34+CD133CD146 (lower left gate) cell fractions with percentage of cells in each gate given (lower left panel); phase contrast image of EOC isolated from CD34+CD133CD146+ cell fraction (magnification ×10) (lower right panel). (B): Association between the number of CD34+CD133CD146+ cells and the capacity to form EOC colonies (closed circles) or not (open circles) form peripheral blood, cord blood, bone marrow, and G-CSF mobilized peripheral blood. Abbreviations: EOC, endothelial outgrowth cells; FSC, forward scatter; SSC, side scatter.

Download figure to PowerPoint

Precursors of EOC Reside Within the CD34+CD133CD146+ Fraction

Enrichment of CD34+ cells from cord blood and adult peripheral blood by magnetic-bead cell sorting identified this fraction as the source of all EOC colonies (Fig. 1B; Table 1). EOC colonies were not isolated from CD34-depleted or CD133-enriched MNC, but EOC colonies were isolated from the CD133-depleted fraction. Following further fractionation of the CD133-depleted fraction into CD34+ and CD34-depleted populations, only the CD133CD34+ fraction gave rise to EOC colonies. When the CD133CD34+ fraction was further fractionated into CD146-enriched and CD146-depleted fractions, only the CD146+ fraction gave EOC colonies (Fig. 1B; Table 1).

Table 1. Endothelial outgrowth and haematopoietic stem cells from enriched and depleted blood mononuclear cells
  1. Values are number or mean ± SEM.

  2. Abbreviations: EOC, endothelial outgrowth cells; CFU-HSC, haematopoietic stem cell colonies.

inline image

The number of cells required to generate EOC from cord blood decreased with enrichment: from >5,000,000 cells per colony for unfractionated MNC to 2,000 cells per colony for CD34+CD133 cells, and 500 cells per colony for CD34+CD133-CD146+ cells (Table 1). Furthermore, EOC colonies were isolated only from those sources with detectable concentrations of CD34+CD133-CD146+ cells (>0.01%) (Fig. 1A). G-CSF-mobilized blood, bone marrow, and normal peripheral blood with negligible numbers of CD34+CD133CD146+ cells were not able to form EOC colonies even after more than 7 weeks in culture. The relationship between the number of CD34+CD133CD146+ cells and capacity to form EOC colonies implies that the precursor to EOC colonies is found within the CD34+CD133CD146+ cell fraction.

CD34+CD133CD146+ Cells Do Not Form Hematopoietic Colonies

Unfractionated MNC and CD34+ or CD133+ enriched subpopulations from all sources generated hematopoietic colonies (CFU-HSC) (Table 1). While both the CD34+CD133+ and CD34+CD133 fractions generated similar numbers of CFU-HSC, CD34+CD133 cells enriched for CD146 did not form hematopoietic colonies.

EOC Are Indistinguishable from Mature Endothelial Cells

Morphology

Phase-contrast images of typical EOC and HUVEC confirm that EOC have the typical cobblestone appearance of mature endothelial cells (Fig. 2A).

thumbnail image

Figure 2. EOC resemble Mature Endothelial Cells by morphology and phenotype. (A): Phase contrast microscopy images of EOC and HUVEC in culture (magnification ×5). (B): Percentage expression of endothelial, adhesion, and stem cell surface antigens in EOC (green) and HUVEC (red) (p > .05, Student's t-test for all). (C): Time in culture conditions does not alter phenotype of EOC or HUVEC (p = .90) with no significant change in antigen expression between early, mid, and late passage (p = .08) (two-way ANOVA). Data are expressed as mean ± SEM (n = 10 for each cell type). Abbreviations: EOC, endothelial outgrowth cell; HUVEC, human umbilical vein endothelial cell.

Download figure to PowerPoint

Phenotype

EOC strongly express endothelial surface antigens, such as Tie-2, KDR, CD31, CD34, CD105, and CD146 to the same extent as HUVEC on direct staining and flow cytometric analysis: CD146 (96 ± 5% vs. 93 ± 4%), CD31 (98 ± 3% vs. 98 ± 2%), and Tie-2 (73 ± 9% vs. 85 ± 4%) expression (n = 10, p > .05 for all (Fig. 2B). The expression of stem cell markers (CD34, CD133, CD45, CD117) and a wide range of integrin (CD29, CD49a, CD49b, CD49c, CD49d, CD49e, CD49f), adhesion (CD50, CD51, CD54, CD61, CD102, CD147) antigens believed to be important in angiogenesis was also indistinguishable between EOC and HUVEC (Fig. 2B; supporting information Figs. 1–4). No significant difference in protein expression was found between EOC and HUVEC across the entire panel (p > .05 for all antigens). Furthermore, surface expression did not differ between EOC and HUVEC through nine passages (Fig. 2C; p = .90 for cell type, p = .08 for time using two-way ANOVA). No difference was found in protein expression between cord blood and peripheral blood-derived EOC (supporting information Fig. 5). Data are expressed as mean ± SEM (n = 3).

Immunostaining

Direct immunostaining confirmed that both EOC and HUVEC strongly express endothelial markers with widespread expression of CD146 and localized granular organelle staining of von Willebrand factor morphologically corresponding to Weibel-Palade bodies (supporting information Fig. 6A).

mRNA Levels by qPCR

EOC and HUVEC both demonstrated low levels of gene expression (cycle threshold [Ct] 30–35) for Wnt genes (Wnt2b, Wnt4, Wnt5a, Wnt5b, Wnt6, Wnt8a, Wnt10b, and Wnt11), Wnt inhibitors such as β-catenin and Dkk1 had higher levels of gene expression (Ct <25 or Ct 25–30) (Table 2). Both EOC and HUVEC expressed stem cell marker CD34. There was no difference in mRNA levels of any of the genes analyzed between EOC and HUVEC indicating that both cell types are closely related for this highly conserved Wnt gene family.

Table 2. mRNA levels of the Wnt gene family and selected stem cell genes in endothelial outgrowth cells and mature endothelial cells
  1. Ct: cycle threshold, i.e., the number of PCR cycles after the fluorescent signal was detectable. The expression of all genes is represented as a comparison with the GAPDH gene. CD34 and β-catenin genes are expressed at a relatively high level, threshold cycle less than 25 on both cell lines. Wnt2b, Dkk1, and Oct-4 genes showed medium expression (Ct 25-30) while Wnt4, Wnt8A, Wnt9A, Wnt10, Wnt11 had little expression (Ct>30). No Ct was detected in Wnt5a, Wnt5b, and Wnt6 (Ct>35 consistent with no expression). B-2M and GAPDH were used as housekeeping genes. All experiments were performed in triplicate.

  2. Abbreviation: EOC, endothelial outgrowth cell.

inline image
ALDH Activity

Both EOC and HUVEC demonstrated ALDH activity. Expression of this stem cell marker was comparable between EOC and HUVEC (54 ± 8% vs. 46 ± 8%, respectively [p = .51, n = 3]) (supporting information Fig. 6B).

Cell Metabolism

Proliferative capacity was assessed by quantifying the number of metabolically active EOC and HUVEC through nine passages. There were no differences in the rate of proliferation between EOC and HUVEC (0.32 ± 0.01 vs. 0.25 ± 0.05 units, p = .12) or over time in culture (p = .52, two-way ANOVA, n = 5; supporting information Fig. 7A). Furthermore, we did not find any differences in PDT between EOC and HUVEC (p = .74) through the nine passages analyzed (p = .06, two-way ANOVA, n = 4) (supporting information Fig. 7B). We also report that both EOC and HUVEC have a similar hierarchy of proliferative cells (p = .96, two-way ANOVA, n = 4, supporting information Fig. 7C), although EOC (2.5% of cells), but no HUVEC, generated colonies with more than 10,000 cells.

Angiogenic Array

The angiogenesis protein array did not identify any difference in cytokine expression in the supernatant from EOC or HUVEC (supporting information Fig. 8). The array showed that both populations expressed the same cytokines growth-regulated oncogene (GRO) human monocyte chemoattractant protein-1 (MCP-1) tissue inhibitor of metalloproteinases-2 (TIMP-2) platelet-derived growth factor BB (PDGF-BB) interleukin (IL) human epidermal growth factor (EGF) basic fibroblast growth factor (bFGF) epithelial-derived neutrophil-activating peptide 78 (Ena78) (GRO, MCP-1, IL-8, IL-6, EGF, PDGF-BB, Ang, TIMP-2, bFGF, and Ena78) with no quantitative difference in the level of expression.

Angiogenic Potential of EOC In Vitro and In Vivo

Both EOC and HUVEC form tubules on Matrigel with no significant difference between the number of connections formed between cells (63.5 ± 6.3 vs. 57.5 ± 4.0 vessel connections per field) (p = .44) (Fig. 3A, 3B). Furthermore, coculture of EOC and HUVEC discriminated by prelabeling with different fluorescent dyes demonstrated colocalization of EOC and HUVEC within tubules and, therefore, direct incorporation of EOC into vascular networks formed by mature endothelial cells (Fig. 3C).

thumbnail image

Figure 3. Both EOC and HUVEC form tubules on Matrigel with no significant difference between the number of connections formed. (A): Phase contrast microscopy of tubule formation on Matrigel by EOC and HUVEC (magnification ×5). (B): The number of connections formed by EOC or HUVEC did not differ (p = .44, unpaired Student's t test). (C): Lipophilic fluorescent staining of each cell type before coculture: EOC (green DiO stain), HUVEC (red DiI stain) and composite image indicating that both cell types contribute to the formation of tubules. Nuclei are stained with DAPI (blue). Scale bar = 100 μm. Abbreviations: DAPI, 4′,6-diamidino-2-phenylindole; EOC, endothelial outgrowth cell; HUVEC, human umbilical vein endothelial cell.

Download figure to PowerPoint

In vivo, implantation of human cells in subcutaneous sponges did not increase the total number of blood vessels compared to vehicle-only GFR Matrigel-impregnated sponges for EOC (7.0 ± 2.7 vs. 8.2 ± 2.3 vessels, p = .06; n = 9) or HUVEC (6.6 ± 3.7 vs. 7.9 ± 3.6 vessels, p = .12; n = 9). (Fig. 4 A–4C). Detailed examination with species-specific MAbs demonstrated that human EOC are capable both of incorporation into murine vessels and formation of new vessels of human origin (Fig. 4D). Interestingly, first passage HUVEC were also able to form new blood vessels but lost this capacity following subsequent passages. Similarly, late passage EOC (>13 passage) lost their capacity to form new blood vessels (Fig. 4D).

thumbnail image

Figure 4. EOC are capable of the formation of new vessels of human origin Angiogenesis in subcutaneous sponge implants. (A): Male NOD/SCID-IL-2gammaRnull mice (NSG) with a sterilized sponge cylinder (0.5 cm × 1 cm) implanted subcutaneously on each flank. Each animal had a control vehicle-impregnated sponge (GFR-Matrigel alone) and cell-impregnated sponge (GFR-Matrigel and EOC or HUVEC) on the other flank. (B): Light microscopy of hematoxylin/eosin-stained sponge 5 μm sections from control vehicle-impregnated sponge (black arrows identify vessels). (C): Number of vessels determined by Chalkley counts in hematoxylin/eosin-stained sections of cell-impregnated sponges implanted for 21 days. Implantation of human cells did not significantly increase vessel number compared to the control sponge implants for EOC (p = .06; Student's paired t test, n = 9) or HUVEC (p = .12; Student's paired t test, n = 9). (D): Immunohistochemistry of sponges following implantation of early EOC (passage 1), late EOC (passage >13), early HUVEC (passage 1), and late HUVEC (passage >13) phase contrast (column 1); rabbit anti-mouse CD31 (green) showing only mouse-derived blood vessels (column 2); mouse anti-human CD146 antibody (red) showing only human-derived blood vessels (column 3); composite image of both antibodies (column 4). DAPI (blue) staining of cell nuclei. Data are expressed as mean ± SEM. Abbreviations: EOC, endothelial outgrowth cell; GFR, growth factor reduced (phase contrast); HUVEC, human umbilical vein endothelial cell; PC, Phase contrast.

Download figure to PowerPoint

Bone Marrow Outgrowth Cells Do Not Form EOC but Differentiate into Bone or Adipose Tissues

While G-CSF mobilized MNC did not survive in EGM-2 medium after 3 weeks of culture as previously [13], bone marrow-derived MNC were highly proliferative and spindle-shaped outgrowth cells (BM-OC) appeared after 1 week of culture (Fig. 5A). These cells did not express endothelial markers CD34 and CD31 and expressed high levels of mesenchymal markers CD166 and CD105 compared to cord blood or peripheral blood-derived EOC (Fig. 5B). While BM-OC generated bone and adipose tissues as demonstrated by Alizarin red S and Sudan Black staining, respectively, cord blood-derived EOC did not (Fig. 5A). BM-OC-derived cells did form some tubule-like structures. However, these were not morphologically comparable to endothelial-derived tubule networks (Fig. 5A). Furthermore, BM-OC-derived cells did not form human blood vessels when implanted into mice (Fig. 5C).

thumbnail image

Figure 5. Bone Marrow Outgrowth Cells Do Not Form EOC but Differentiate into Bone or Adipose Tissues. Comparison between bone marrow-derived outgrowth cells and cord blood-derived outgrowth cells (EOC) after 3 weeks culture on collagen in EGM-2 medium. (A): Characterization of outgrowth cells (OC): phase contrast (column 1); tubule formation in Matrigel (column 2); bone formation detected by Alizarin red staining (column 3); adipose formation detected by Sudan Black staining (column 4). (B): Flow cytometry histograms showing bone marrow-derived outgrowth cell expression of mesenchymal and integrin markers and the lack of expression of endothelial markers, such as CD146 or CD31. (C): Immunostaining of sections of bone marrow-derived outgrowth cell-impregnated sponges implanted for 21 days in NOD/SCID-IL-2gammaRnull mice (NSG) mice: phase contrast (column 1); rabbit anti-mouse CD31 (green) showing only mouse-derived blood vessels (column 2); mouse anti-human CD146 antibody (red) showing only human-derived blood vessels (column 3); composite image of both antibodies (column 4). DAPI staining of cell nuclei (blue). All vessels identified were of mouse origin with no human vessels identified in any of the mice implanted with bone marrow derived outgrowth cells. Abbreviations: EGM, Endothelial growth medium; EOC, endothelial outgrowth cell; PC, Phase contrast; OC, outgrowth cells.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

Endothelial outgrowth cells (EOC) are considered to fulfil the criteria of a true EPC. However, the exact origin, definition, and biology of EOC remain unclear. We confirm that EOC can be isolated reproducibly from blood MNC, and their precursor resides in the CD34+CD133CD146+ cell fraction, but EOC are not present in bone marrow or cells mobilized from the bone marrow. Furthermore, EOC are indistinguishable from mature endothelial cells by morphology, surface antigen expression, immunohistochemistry, gene and protein expression, proliferation, and functional assessments. EOC do not stimulate neovascularization per se but readily incorporate into vascular structures in vitro and in vivo. Together these findings suggest that EOC may be derived from a vascular source outside the bone marrow and challenge the concept of a bone marrow-derived circulating precursor for endothelial cells.

Precursors of EOC

There is not a clear antigen marker, or combination of markers, which could identify the precursor to EOC. Timmermans et al. [14] showed that EOC precursors are confined to a small fraction of CD34+CD45 MNC that coexpress VEGFR2 but not CD133. Case et al. [3] suggest that previous definitions of circulating EPC based on the expression of circulating CD34+CD133+VEGFR2+ and CD34+CD45+ cells in fact define a distinct hematopoietic stem cell population. Using enrichment and depletion strategies to isolate cell fractions based on the expression of stem cell and endothelial markers, we demonstrate that EOC were present only in the CD34+CD133 cell fraction. Following further enrichment using CD146, a transmembrane glycoprotein constitutively expressed on mature endothelial cells, we found that only the CD34+CD133CD146+ fraction gave rise to EOC colonies. These findings are consistent with a recent study by Mund et al. [15] who report that EOC precursors express CD146+.

We evaluated four potential clinical sources of endothelial progenitors and found that EOC could only be isolated from cord blood or peripheral blood MNC. EOC were not present in bone marrow or cells mobilized from the bone marrow. Interestingly, the number of cells required to produce EOC colonies was dependent on the presence and number of CD34+CD133CD146+ cells. Cord blood and peripheral blood samples which successfully generated EOC, contained CD34+CD133CD146+ cells at a concentration >0.01% of the gated CD34+ cells. In contrast, bone marrow and cells mobilized from the bone marrow by G-CSF had negligible numbers of CD34+CD133CD146+ cells and consistently failed to generate EOC colonies. This direct relationship between the number of circulating CD34+CD133CD146+ and the capacity to form EOC colonies reinforces the notion that the EOC precursor resides in the CD34+CD133CD146+ cell fraction. This combination of surface markers may represent an EOC precursor in circulation and quantification of this phenotype may be a useful biomarker in patients with cardiovascular disease or tumorgenesis. As CD34+CD133-CD146+ cells are rare in circulation, we agree with Estes et al. [16] that only accurate and carefully controlled flow cytometry protocols will be able to reliably quantify EOC precursors.

Interestingly, while all other CD34+ populations formed hematopoietic colonies on methylcellulose, CD34+CD133CD146+ cells did not. These findings are consistent with Mund et al. [15] and suggest no direct relationship or common precursor between EOC-forming CD34+CD133-CD146+ cells and hematopoiesis.

Are EOC Mature Endothelial Cells?

One of the challenges in identifying the precursor of EOC is that at present there are no unique markers known that reliably distinguish EOC from mature circulating endothelial cells (CEC) [8]. CD146 is constitutively expressed on mature endothelial cells, and CD146 immunobeads are routinely used to quantify blood CEC [17, 18]. While it has been suggested that EPC express CD146 and may be distinguished from CEC by the presence of the pan-leukocyte marker CD45 and hematopoietic stem cell marker CD133 [19], our own findings and those of from others [11, 14] suggest that EOC and hematopoietic stem cells do not share a common precursor and are negative for CD133 and CD45.

In the absence of a unique marker to differentiate between the precursor of EOC and CEC, we undertook a systematic comparison between EOC and mature endothelial cells in an attempt to identify phenotypic or functional differences that could be used to better define the progenitor population. Both EOC and HUVEC formed confluent cobblestone-shaped monolayers with an identical phenotype whether assessed by flow cytometry or immunohistochemistry. The expression of multiple endothelial, adhesion, integrin, and stem cell markers did not differ between cell types and did not change through sequential cell culture passages. In contrast to the results published by Van Beem et al. [20], we did not find any significant difference in expression of CD34, KDR, or Tie2 positive cells between EOC and HUVEC during in vitro culture. Furthermore, both cell types have identical levels of mRNA expression of a selected panel of Wnt and Wnt-inhibitor genes known to be involved in cell fate and cell growth [21]. Our findings are consistent with Medina et al. [22] and suggest that EOC and HUVEC share a common origin.

Stem cells express high levels of ALDH a widely accepted marker of cellular stemness that has previously been used to discriminate mature cells from progenitor cells [23]. Unexpectedly, mature endothelial cells had comparable ALDH fluorescence intensity to EOC suggesting that ALDH staining is not able to discriminate between progenitor and mature endothelial cells or EOC arise from or are essentially mature endothelial cells. Consistent with the latter interpretation, it is been reported that EOC have higher proliferation potential and are able to undergo more cycles of cell divisions than are HUVEC [24]. Consistent with the findings of Ingram et al. [25], we report EOC and HUVEC display similar clonogenic potential and growth kinetics during nine cell culture passages. Mature endothelial cells were once thought to have limited proliferative capacity. However, it has been recently demonstrated that HUVEC can divide for at least 40 PDT and contain a hierarchy of cells with differing proliferative capacity [25].

EOC behaved like mature endothelial cells in vitro with a similar capacity to form tubules on Matrigel compared to HUVEC. Furthermore, EOC and HUVEC readily combined to form vascular networks when cocultured. Following implantation in an immunodeficient mouse model of angiogenesis both EOC and HUVEC contributed to the formation of blood vessels in vivo without promoting angiogenesis per se. Interestingly, both EOC and HUVEC lost their capacity to form de novo vascular structures following multiple passages in vitro.

Although, we conclude that EOC do not differ from mature endothelial cells using a broad range of morphological, phenotypic, and functional assays, we cannot exclude the possibility of some differences between these cell populations detectable by any further alternative assays or techniques.

Origin of EOC

The precise site of origin of EOC is presently unknown, but it has widely been assumed that EOC like other putative EPC are derived from bone marrow [26, 27]. We demonstrate that bone marrow MNC had negligible numbers of CD34+CD133CD146+ cells and did not generate EOC colonies. In contrast, proliferative spindle-shaped bone marrow outgrowth cells appeared soon after plating in EGM-2. This outgrowth cells did not express endothelial markers and did not form blood vessels in vitro or in vivo but consistently expressed mesenchymal cell markers and, under the appropriate culture conditions, differentiated into bone and adipose tissues. These findings, taken together with our observation that EOC have a phenotype, gene expression profile, and proliferative and functional capacity that is identical to mature endothelial cells strongly supports the contention that EOC precursors are not bone marrow-derived EPC but in fact arise from the vasculature. We were not able to differentiate between EOC derived from cord blood or peripheral blood perhaps suggesting a common precursor. Vascular wall endothelial cells have generally been considered to be differentiated and postmitotic quiescent cells with limited capacity to proliferate. Evidence now clearly suggests the presence of resident precursors with proliferative capacity between the tunica media and tunica adventitia [28]. High proliferative potential-colony forming units (HPP-CFC) have been isolated from bovine and swine aorta, coronary artery, pulmonary artery, and corneal vascular cells [29]. These cells are capable of differentiation into mature endothelial cells and form capillary sprouts. It may be possible to harness vessel wall progenitor cells to promote vascular repair or isolate EOC for therapeutic purposes. While EOC may have similar functional properties to mature endothelial cells they potentially offer a therapeutic advantage in that they can be readily isolated from peripheral blood.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

We have identified a link between the number of CD34+CD133CD146+ cells and formation of EOC and suggest that this combination of markers may be used to define the precursor of late outgrowth endothelial cells. EOC do not differ from mature endothelial cells using a broad range of morphological, phenotypic, and functional assays and cannot be isolated from bone marrow derived MNC. Together these findings suggest that EOC may be derived from a vascular source outside the bone marrow and challenge the concept of bone marrow-derived circulating precursor for endothelial cells. Further studies are required to fully conclude whether EOC are true progenitor cells or mature endothelial cells with proliferative potential.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

Drs. Tura and Mills are supported by a Project Grant from the Chief Scientist Office, Scotland (CZB/4/812) and an Intermediate Clinical Research Fellowship from the British Heart Foundation (BHF) (FS/10/024/28266), respectively, and this research was supported by a BHF Project Grant (PG/06/051). We would like to acknowledge support from the BHF Centre of Research Excellence Award.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
sc-12-0320_sm_SupplFigure1.tif1521KSupplementary Figure 1. FACS analysis of endothelial markers (CD105, Tie-2, CD31, VEGFR2, CD34 and CD146) in EOC and HUVEC. Plots show representative profile of control (no antibody; black) and specific antibody staining (white).
sc-12-0320_sm_SupplFigure2.tif1521KSupplementary Figure 2. FACS analysis of adhesion molecules (CD29, CD49a, CD49b, CD49c, CD49d, CD49e and CD49f) in EOC and HUVEC. Plots show representative profile of control (no antibody; black) and specific antibody staining (white).
sc-12-0320_sm_SupplFigure3.tif1521KSupplementary Figure 3. FACS analysis of integrin markers (CD50, CD51, CD54, CD61, CD102 and CD147) in EOC and HUVEC. Plots show representative profile of control (no antibody; black) and specific antibody staining (white).
sc-12-0320_sm_SupplFigure4.tif1521KSupplementary Figure 4. FACS analysis of haematopoietic (CD45) and stem cell markers (CD133 and CD117) in EOC and HUVEC. Plots show representative profile of control (no antibody; black) and specific antibody staining (white).
sc-12-0320_sm_SupplFigure5.tif1521KSupplementary Figure 5. Percentage expression of endothelial, adhesion and stem cell surface antigens in peripheral blood derived EOC (black) and Cord blood derived EOC (grey) (P>0.05, student's t-test for all).
sc-12-0320_sm_SupplFigure6.tif1521KSupplementary Figure 6. A) Immunohistochemistry showing expression of CD146 and von Willebrand factor (vWF) on EOC and HUVECs. Bar = 100μm. Widespread expression of endothelial markers such as CD146 and localised staining of von Willebrand factor in storage granules in both EOC and HUVECs: phase contrast (column 1), A488 green fluorescence (column 2), A555 red fluorescence (column 3) and composite images (column 4). DAPI (blue) was used for detection of nuclear staining. B) Representative flow cytometric analysis of aldehyde dehydrogenase activity in HUVEC: cells are selected based on forward/side scatter properties; gates are defined using the Aldefluor negative control (substrate and inhibitor); Aldefluor positive cells with aldehyde dehydrogenase activity are quantified (substrate with no inhibitor).
sc-12-0320_sm_SupplFigure7.tif1521KSupplementary Figure 7. A) Quick proliferation assay (Abcam, UK) in HUVECs and EOC. There is no difference in the rate of proliferation as assessed by the quick proliferation assay between EOC and HUVEC (P=0.12) whether early (0-2), mid (3-5) or late (6-9) passage (P=0.5) (2-way ANOVA, n=5). Data are expressed as mean ± SEM. B) Growth kinetics of EOC and HUVEC. Population doubling time (PDT) is calculated through all the nine passages for both cell types. We did not find any differences in cell doubling times (PDT) between EOC and HUVEC (P=0.70) or through the nine passages analysed (P=0.06, 2-way ANOVA, n=4). C) Quantification of the clonogenic potential of single EOC and HUVEC. Number of cell progeny derived from a single EOC or HUVEC. EOC and HUVEC have similar hierarchy of proliferative cells (P=0.90, 2-way ANOVA, n=4).
sc-12-0320_sm_SupplFigure8.tif1521KSupplementary Figure 8 A) Conditioned medium from both EOC and HUVEC cell cultures were tested for the expression of angiogenic factors using the human angiogenesis antibody array G Series 1. Panel I represents 2 different angiogenic arrays using EGM-2 culture medium (no cells) as a control. Panel II represents n=3 EOC conditioned medium derived from 3 different cord blood samples. Panel III represents n=3 HUVEC conditioned medium from 3 independent samples. B) Expression levels of GRO (Growth regulator oncogene), MCP-1 (Monocyte chemoattractant 1), IL-8 (Interleukin 8), IL-6 (Interleukin 6), EGF (Epidermal growth factor), PDGF-BB (Platelet derived growth factor), Ang (Angiogenin), TIMP-2 (Tissue inhibitor of metalloproteinases), bFGF (basic Fibroblast growth factor), and Ena78 (epithelial-derived neutrophil-activating peptide 78) chemokines of EOC (black), HUVEC (dark grey) and EGM-2 medium only (clear grey). The array showed an up-regulation of these cytokines compared to EGM-2 medium only but could not detect any significant difference in the level of expression between EOC and HUVEC. Positive (POS) and negative (NEG) internal controls were also present in the array. Data are expressed as mean ± SD; n=3 independent samples for each cell type.
sc-12-0320_sm_SupplTable1.tif1521KSupplementary Table 1

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.