SEARCH

SEARCH BY CITATION

Keywords:

  • Monocytes;
  • Endothelial progenitor cells;
  • Endothelial cells;
  • Circulating angiogenic cells;
  • Circulating endothelial progenitors;
  • Circulating endothelial cells

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. References

The generation of endothelial progenitor cells (EPCs) from blood monocytes has been propagated as a novel approach in the diagnosis and treatment of cardiovascular diseases. Low-density lipoprotein (LDL) uptake and lectin binding together with endothelial marker expression are commonly used to define these EPCs. Considerable controversy exists regarding their nature, in particular, because myelomonocytic cells share several properties with endothelial cells (ECs). This study was performed to elucidate whether the commonly used endothelial marker determination is sufficient to distinguish supposed EPCs from monocytes. We measured endothelial, hematopoietic, and progenitor cell marker expression of monocytes before and after angiogenic culture by fluorescence microscopy, flow cytometry, and real-time reverse transcription–polymerase chain reaction. The function of primary monocytes and monocyte-derived supposed EPCs was investigated during vascular network formation and EC colony-forming unit (CFU-EC) development. Monocytes cultured for 4 to 6 days under angiogenic conditions lost CD14/CD45 and displayed a commonly accepted EPC phenotype, including LDL uptake and lectin binding, CD31/CD105/CD144 reactivity, and formation of cord-like structures. Strikingly, primary monocytes already expressed most tested endothelial genes and proteins at even higher levels than their supposed EPC progeny. Neither fresh nor cultured monocytes formed vascular networks, but CFU-EC formation was strictly dependent on monocyte presence. LDL uptake, lectin binding, and CD31/CD105/CD144 expression are inherent features of monocytes, making them phenotypically indistinguishable from putative EPCs. Consequently, monocytes and their progeny can phenotypically mimic EPCs in various experimental models.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. References

The advent of cellular therapy for ischemic organ damage has introduced neoangiogenesis by bone marrow (BM)-derived stem cells (SCs) and progenitors as a reparative mechanism to improve perfusion after ischemia. Evidence has accumulated over the past few years that immigrating BM-derived cells differentiate along endothelial lineage but also provide paracrine help in the course of regenerative vascular remodeling [13].

BM-derived progenitors for therapeutic angiogenesis were first isolated from adult human peripheral blood by Asahara et al. [4] based on their expression of CD34, a molecule expressed by microvascular endothelial cells (ECs) and hematopoietic progenitor cells (HPCs) and hematopoietic SCs (HSCs). The same group provided the proof of the concept of BM-derived vasculogenesis by transplantation of genetically marked mouse BM into recipient mice. The mice were subsequently subjected to several distinct models of vascular remodeling, including myocardial or hind limb ischemia, endometrial neovascularization after induced ovulation, and cutaneous wound healing and vascular support of tumor growth, to show contribution of BM-derived cells to neovasculogenesis [5]. The exact origin and precise phenotype of the respective BM-derived progenitors have been a matter of debate at least in part due to the only 15.7% ± 3% purity of the CD34+ cells used in the initial study [4]. The minute population of vascular endothelial growth factor receptor (VEGFR)-2+/CD34+ HSCs has originally been considered to represent one interesting target for further investigations. These cells comprise 0.1%–0.4% of purified CD34+ cells in peripheral blood (PB), BM, adult G-CSF-mobilized PB (MPB), and umbilical cord blood (UCB) [6]. Peichev et al. [7] succeeded in defining functional circulating endothelial progenitors (CEPs) derived from VEGFR-2+/CD34+/133+ HSCs in UCB and adult MPB. This observation was extended by identifying a CD34+/VEGFR-2+ hemangioblast in UCB and adult human BM and CD34+/133+/VEGFR-3+ lymphangiogenic precursors in MPB and UCB [8, 9].

Previously, hemangioblasts were considered to exist only during embryonic development as SCs capable of generating both hematopoietic progeny and ECs [10]. The rarity of the hemangioblast population as well as the contamination of the CD34+CEP by a high proportion of CD34 cells in the seminal study by Asahara et al. [4] obviously led to the more practical use of unmanipulated PB mononuclear cells (MNCs) as the starting material for CEP culture in many laboratories. The resulting ex vivo–generated cell population has then been termed endothelial progenitor cells (EPCs) and documented, like ECs, to bind lectins and to integrate acetylated (Ac)-LDL together with several other marker molecules related to the EC lineage (CD31/CD34/CD144/KDR) [5, 11, 12]. The prototype Ac-LDL+/lectin+ cell type has been found to be reduced in patients with coronary artery disease, and this phenomenon was reversed by HMG-CoA reductase inhibitor (statin) treatment [1317].

These putative EPCs are meanwhile also considered to be derived from CD14+ monocytes, which have been shown to display several phenotypic (production of von Willebrand factor [vWF]), morphologic (Weibel-Palade bodies), and functional EC-like features after angiogenic culture (cord-like structures and vascular network formation) [18, 19]. The delineation of EPC ontogeny is complicated by the fact that CD34+ HPCs mobilized in patients with granulocyte-macrophage colony-stimulating factor (GM-CSF), just as PB-MNC-derived EPCs in certain studies coexpress EC molecules and the monocyte marker CD14 after angiogenic stimulation [11, 20].

The true progenitor cell character of monocyte-derived LDL+/lectin+ EPCs came into question, with data showing a clear lack of proliferation of these cells. The detection of potent angiogenic growth factor secretion further indicated a peculiar role during vascular remodeling that relies on cytokine production after appropriate stimulation [21]. In line with these observations, in a therapeutic nude mouse model for the regeneration of hind limb ischemia, human CD14+ monocytes differentiated into macrophages and dendritic cells were significantly less efficient than monocytes after angiogenic culture in improving neovascularization. This indicates that CD14+ cells primed along a nonendothelial pathway secrete factors that do not lead to an optimal support of angiogenesis [22].

Based on these variable results, we are currently faced with a considerable controversy about the nature and function of human blood EPCs. In light of the widespread experimental and initial clinical use of supposed EPCs, the discrimination of monocyte effects from that of progenitors and SCs may be of particular importance to select cells for regenerative therapy.

This study was performed to clarify whether common marker molecule determination is reliable to confirm an asserted transdifferentiation from monocytes into EPCs.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. References

Cell Isolation

PB-MNCs from healthy male volunteers (n = 9) were isolated from buffy coats by density-gradient centrifugation (Ficoll-Paque PLUS; StemCell Technologies, Vancouver, Canada, http://www.stemcell.com), washed twice in phosphate-buffered saline (PBS) containing 2 mM EDTA and 0.5% serum albumin, incubated with anti-CD14 MicroBeads (Miltenyi Biotec, Bergisch Gladbach, Germany, http://www.miltenyibiotec.com) for 15 minutes, and separated through a column of a MACS separator (Miltenyi Biotec). A purity of 99.2% ± 0.3% CD14+ cells could be obtained. To exclude the possible induction of phenotypic or functional changes by CD14-positive selection of monocytes, cells were incubated with the Monocyte Isolation Kit II (Miltenyi Biotec) and isolated by negative selection according to the manufacturer's instructions for comparative (flow cytometry and real-time reverse transcription–polymerase chain reaction [RT-PCR]) experiments (n = 4). Briefly, cells were incubated with a cocktail containing biotin-conjugated antibodies against CD3, CD7, CD16, CD19, CD56, CD123, and glycophorin A. After a washing step, a second incubation was performed with antibiotin microbeads, and cells were sorted using a MACS column. Positive isolated and untouched monocyte fractions were further termed CD14+SORT and MonoSORT, respectively. The MonoSORT fraction comprised 83.2% ± 3% of CD14+ monocytes.

DiI-Ac-LDL Uptake

Freshly isolated or 4- and 6-day cultured CD14+SORT and MonoSORT cells were incubated with 2.5 μg/ml 1,1′-dioctade-cyl-3,3,3′,3 β-tetramethylindo-carbocyanine perchlorate (Dil)-labeled acetylated LDL (Dil-Ac-LDL) (Biomedical Technologies, Stoughton, MA, http://www.btiinc.com) in EGM-2 medium (Cambrex, Walkersville, MD, http://www.cambrex.com) for 1 hour as published [14, 16, 22], washed, and analyzed by fluorescence microscopy and flow cytometry.

Cell Culture, Vascular Network Formation, and CFU-EC Assay

CD14+SORT and MonoSORT cells were seeded into six-well plates coated with human fibronectin (Sigma-Aldrich, St. Louis, http://sigmaaldrich.com) at a density of 2.5 × 106/cm2 in EGM-2 supplemented with single quots (containing hydrocortisone, human epidermal growth factor, 2% fetal bovine serum, 1 to 10 ng/ml VEGF, human fibroblast growth factor B, R3-insulin like growth factor-1, ascorbic acid, heparin, gentamycin, and amphotericin; Cambrex) and additionally supplemented with 20% fetal calf serum and 50 ng/ml human recombinant VEGF-165 (R&D Systems, Minneapolis, http://www.rndsystems.com) immediately after isolation [1417]. Vascular network formation of fresh (day-0), day-4, and day-6 cultured monocytes compared with vascular ECs from human umbilical vein (HUVEC) (Cambrex) and myelomonocytic leukemia cells (HL-60; ATCC CCL240) was tested in a standardized in vitro angiogenesis assay (Matrigel, Chemicon, Temecula, CA, http://www.chemicon.com). For this purpose, Dil-Ac-LDL-labeled cells (6,000 to 8,000 per well) were seeded in quadruplicates into 16-well chamber slides (Nalge Nunc International, Naperville, IL, http://www.nuncbrand.com) and investigated for their ability to form a vascular network. To test whether incubation with Dil-Ac-LDL would hamper network formation, Dil-Ac-LDL-labeled and unlabeled HUVECs were seeded onto Matrigel and network formation was examined 12 to 18 hours later. The capacity of monocytes and monocyte-derived cells to incorporate into a growing three-dimensional vascular network was assessed by coseeding of Dil-Ac-LDL-labeled myelomonocytic cells with vascular ECs, which were prestained with fluorescein isothiocyanate (FITC)-conjugated antibodies against the endothelial and melanoma cell adhesion molecule Muc18 (CD146, clone P1H12; Chemicon) [23] at a ratio of 1 to 1 (4,000 monocytes or monocyte-derived cells to 4,000 vascular ECs).

CFU-EC formation was studied in an established in vitro colony assay [24]. We used the commercially available standardized assay version (Endocult; StemCell Technologies). Briefly, cells were plated at a density of 5 × 105/cm2 in fibronectin-coated six-well plates in supplemented Endocult medium. After 2 days, nonadherent cells were transferred into fibronectin-coated 24-well plates at 5 × 105/cm2 in supplemented Endocult medium. This step intends to eliminate mature circulating ECs and is followed by a 3-day differentiation of the transferred cell fraction. At day 5, typical clusters comprising a central part of small round cells surrounded by radiating spindle-shaped cells were counted as CFU-EC as published [24]. We designed a subtractive approach to compare the CFU-EC outgrowth from PB-MNCs in the presence or absence of CD14+SORT cells. For this purpose, CFU-EC number in mock-sorted MNCs (cells just passed through the sorting column) was compared with that of CD14+SORT and CD14-depleted fractions remaining after MACS sorting. Purity of populations was confirmed by flow cytometry. Adding back CD14+SORT cells into their CD14-depleted derivative fractions was used to validate results.

Immunofluorescence and Immunohistochemistry

Dil-Ac-LDL-labeled fresh and cultured monocytes were counterstained with FITC-labeled lectins from either Bandeiraera simplicifolia (BS-1) (Sigma-Aldrich) or Ulex europaeus I (UEA-1) (Sigma-Aldrich), spun onto slides (Shandon Cytospin 3 Cytocentrifuge; Thermo Electron Corporation, Pittsburgh, http://www.thermo.com), and examined using a fluorescence microscope (Nikon Eclipse 600; Nikon Instruments, Melville, NY, http://www.nikonusa.com). Immunohistochemistry of acetone/methanol-fixed (3:2 parts vol/vol, 15 minutes, 0°C) adherent cells (cord-like structures) was performed using the Chem-Mate peroxidase detection kit to visualize binding of mouse anti-human vimentin, vWF, CD45, and Ki67 antibodies compared with isotype control. Staining was developed with diaminobenzidine following the manufacturer's instructions (all DakoCytomation Denmark A/S, Glostrup, Denmark, http://www.dakocytomation.com) and documented on a Diaphot 300 inverted microscope (Nikon). Pictures were taken with a Cool-pix 4500 digital camera (Nikon).

Multiparameter Flow Cytometry

The surface marker expression of freshly isolated or cultured monocytes was analyzed with a four-color FACSCalibur instrument equipped with a 488-nm argon ion laser and a 635-nm red diode laser (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bdbiosciences.com). Day-4 and day-6 cultured monocytes were detached by gently resuspending with 1 mmol/l EDTA in PBS (pH, 7.2 to 7.4), and cell count of adherent cells was determined using a hemocytometer. Fresh and cultured cells were washed, blocked with sheep serum, and labeled for 25 minutes at 4°C at concentrations according to individual titration with BS-1 and UEA-1 lectins and monoclonal antibodies against HLA-AB (Harlan Sera-Lab, Leicestershire, UK, http://www.harlanseralab.co.uk), HLA-DR, CD14, CD31, CD34, CD45, CD73 (BD), CD105 (Caltag Laboratories, Burlingame, CA, http://www.caltag.com), CD133 (Miltenyi Biotec), CD144 (Bender MedSystems, Burlingame, CA, http://www.bendermedsystems.com), CD146 (clone P1H12, Chemicon), and VEGFR-2/KDR (clone KDR-2, Sigma-Aldrich). Appropriate isotype-matched antibodies were used as negative controls (Becton, Dickinson and Company). Four-color measurements were performed, and data from at minimum 10,000 viable propidium iodine-excluding cells were stored. List mode files were analyzed with CellQuest Pro and Paint-A-Gate Pro software (Becton, Dickinson and Company). Change in mean fluorescence intensity is shown as mean ± standard error of mean (SEM) arbitrary units and results from the calculation of MFIsample − MFIIsotype control.

Real-Time RT-PCR

Freshly isolated and cultured monocytes and fresh and cultured PB-MNCs were harvested, and total RNA was isolated using TriZol (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) according to the manufacturer's instructions. HUVEC and human microvascular ECs (HMVECs) (Cambrex) were used as references. RNA concentration and purity were determined on SartSpec Plus Spectrophotometer (Bio-Rad Laboratories, Inc., Hercules, CA, http://bio-rad.com). Three micrograms of total RNA was reverse transcribed using High-Capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). For quantitative real-time PCR, the following components were prepared: 50 ng cDNA, 1.25 μl 20xTaqMan PCR assay, and 12.5 μl 2xTaqMan Universal Master Mix containing ROX (Applied Biosystems), and H2O was added to a final volume of 25 μl. Assays were run in triplicates under standard RT-PCR conditions (50°C, 2 minutes; 95°C, 10 minutes [95°C, 15 seconds; 60°C, 1 minutes] × 40 cycles). The possibility of amplification from contaminating DNA was excluded by use of template-free and RT- controls. mRNA expression levels of CD105 (assay ID, Hs00164438 m1), VEGFR-2 (Hs00176676 m1), CD34 (Hs00156373 m1), CD146 (Hs00174838 m1), CD31 (Hs00169777 m1), VEGFR-1 (Hs00176573 m1), VEGFR-3 (Hs00176607 m1), vWF (Hs00169795 m1), CD14 (Hs00169122 g1), and CD45 (Hs00236304 m1) were determined using the Applied Biosystems 7900HT Real-Time PCR System in 96-well microtiterplate format. All assays are inventoried TaqMan Gene Expression Assays (Applied Biosystems) and have a FAM reporter dye at the 5′ end of the TaqMan MGB probe and a nonfluorescent quencher at the 3′ end of the probe. Data were normalized to β-actin (forward primer: 5′-AgCCTCgCCTTTgCCgA; reverse primer: 5′-CTggTgCCTggggCg; TaqMan probe: 5′-FAM-CCgCCgCCCgTCCACACCCgCCT-Tamra) as endogenous control. Expression values from fresh monocytes (day 0) were taken as calibrator to determine relative changes in gene expression over time of angiogenic culture. To show the variation of mRNA expression between EC lines, MNCs, and sorted monocytes, data are expressed as Δ ΔCt (the ΔCt [target of interest] for HUVEC = cycle of threshold [β-actinHUVEC] − cycle of threshold [target of interestHUVEC], and ΔΔCt is ΔCt [targetcell type] − ΔCt [targetHUVEC]). For gene expression kinetics over time of culture, the ratio (R) was calculated [25]. In brief, real-time PCR efficiency (E) for all assays was deduced from the slopes obtained in cDNA standard-curve titration experiments, whereas the formula E = 10[−1/slope] was used to calculate the relative expression ratio:

  • equation image

Gene expression kinetics are calculated as ratios (R, log [2] transformed) of cell fractions after culture (= sample) versus uncultured, day-0 cells (= calibrator) and shown as mean ± standard deviation of three independent experiments.

Statistical Analysis

SPSS 11.5.1 software (SPSS Inc., Chicago, http://www.spss.com) was used for statistical analyses. Statistical differences in mean fluorescence intensity flow cytometry results from fresh and cultured monocytes and CFU-EC numbers were assessed using the paired Student's t-test. Significance was assumed at p < .05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. References

Monocyte-Derived Cells Resemble Putative Ac-LDL+/Lectin+EPCs

To validate our experimental model for the analysis of monocyte-to-EPC differentiation, we first confirmed that monocytes cultured for 4 to 6 days in EC medium perform Dil-Ac-LDL uptake and lectin binding. Cultured monocytes showed homogenous uptake of Dil-Ac-LDL and bound both BS-1 and UEA-1 lectins, resulting in the typical Dil-Ac-LDL+/lectin+ EPC phenotype (Figs. 1A, 1B). Within the monocyte-derived cultures, we regularly found cord-like structures of adherent cells arranged in rows of 2 to 10 cells, sometimes with a length of several millimeters (Fig. 1C). These cultured monocytes did not react with Ki67, indicating a lack of proliferation (data not shown). Freshly sorted monocytes also took up Dil-Ac-LDL (Fig. 1D). Within the remaining fraction of the monocyte-depleted PB-MNCs, we observed rare cells that visibly took up lower amounts of Dil-Ac-LDL compared with the purified monocytes (Fig. 1E).

Key Markers of Supposed EPC Differentiation Are Constitutively Expressed by Blood Monocytes

We used multiparameter flow cytometry to quantify the expression of typical markers of the supposed EPC differentiation of CD14+SORT and MonoSORT cells. Obviously, most tested attributes of the supposed EPC differentiation were already present in fresh blood monocytes analyzed immediately after blood drawing. Fresh monocytes showed strong uptake of Dil-Ac-LDL, bound FITC-labeled BS-1 lectin and expressed CD31, CD105, and CD144. Only weak reactivity above background was measured for CD34, KDR, and UEA-1. No significant difference was found between CD14+SORT and MonoSORT cells (Figs. 2A, 2C). High expression of the hematopoietic marker CD45, HLA molecules, and the monocyte marker CD14 was found on fresh monocytes (Figs. 2B, 2D).

Real-time RT-PCR was used to analyze marker expression at the mRNA level compared with unsorted PB-MNCs, HMVECs, and HUVECs. CD31 mRNA was the only message expressed at comparable levels in the endothelial as well as in the blood cell populations. CD105 mRNA was detected at ∼2−5 in the blood cells compared with EC lines (representing a 32-fold lower expression). Compared with HUVECs, the CD14 and CD45 mRNA expression in blood cell preparations was 212-fold and 219-fold (∼4 × 103-fold and 5 × 105-fold) higher, respectively. A 2−10- to 2−20-fold expression (103- to 106-fold lower) for VEGF-Rs, vWF, and CD146 was found in both sorted monocyte preparations in comparison with HUVECs. The lowest expression of CD34 mRNA was found in the highly purified CD14+SORT monocytes. A higher expression level for CD34 mRNA was detected in MNCs and negatively enriched MonoSORT cells, which contained 0.1% ± 0.01% and 0.38% ± 0.1% contaminating CD34+ cells, respectively (Fig. 3).

Monocyte-Derived Cells Do Not Upregulate Supposed Endothelial Markers upon Angiogenic Culture

A starting number of 2.5 × 107 seeded monocytes resulted in 1.33 ± 0.13 × 106 adherent cells after 6 days of culture. Using flow cytometry, we reproducibly found a nonsignificant increase of Dil-Ac-LDL uptake in MonoSORT cells only. For CD105 reactivity, a nonsignificant upregulation was found in both sorted monocyte populations upon angiogenic culture. In fact, we observed a significant drop of CD31, CD144, and BS-1 lectin binding and no increase in KDR, UEA-1, and CD34 reactivity. CD73, CD133, and CD146 could not be detected on fresh or cultured monocytes at the protein level (Figs. 2A, 2C). Comparative analysis of typical hematopoietic marker molecules revealed a significant loss of CD45, CD14, and major histocompatibility complex (MHC) class I and stable MHC class II reactivity for both CD14+SORT- and MonoSORT-derived cells (Figs. 2B, 2D).

To document the mRNA expression kinetics during culture, we calculated the relative mRNA expression ratio of cell populations derived from angiogenic culture versus freshly isolated cells. Unchanged or downregulated mRNA expression was found for VEGF-R1, VEGF-R3, CD34, CD146, CD31, vWF, CD14, and CD45 after angiogenic culture for both sorted monocyte populations. A fourfold to eightfold increase in CD105 and VEGF-R2 mRNA expression (ratio, 1.93 ± 0.86 and 3.07 ± 0.67, respectively) was detected in CD14+SORT-derived cells, and a fourfold increase in VEGF-R2 mRNA expression (ratio, 1.91 ± 0.42) was detected in MonoSORT-derived cells. However, the fourfold to eightfold upregulation of VEGF-R2 mRNA expression still resulted in a 215-fold lower mRNA expression compared with HUVECs (Fig. 3). This was accompanied by the above-described nonsignificant CD105 and VEGF-R2 protein expression, as determined by flow cytometry.

Monocyte-Derived Cells Do Not Form Vascular Networks but Contribute to CFU-EC Development

When CFU-EC development was tested, a starting number of 15.7 × 107 seeded PB-MNCs resulted in 1.8 ± 0.7 × 107 adherent cells after 5 days of culture, representing a recovery of 11.5% initial input cell number. mRNA expression of CFU-EC cultures (day 5) was compared with that of corresponding PB-MNCs before culture (day 0). We measured a fourfold to eightfold upregulation of CD105, VEGF-R2, and VEGF-R1 in adherent cells upon a 5-day culture in Endocult medium. Unchanged or downregulated expression was found for VEGF-R3, CD34, CD146, CD31, vWF, CD14, and CD45 (data not shown). We further analyzed whether monocytes play a role during this CFU-EC formation. A total of 14.5 ± 3 (mean ± SEM) CFU-ECs were obtained per 106 fibronectin nonadherent MNCs (n = 4). Pure CD14+SORT cells lacked the capacity to form CFU-EC, as did the CD14-depleted MNCs. A prominent effect of CD14 depletion was the complete loss of the spindle-shaped adherent cells in the monocyte-depleted CFU-EC cultures (Figs. 4A–4C). Adding back CD14+SORT cells to the CD14-depleted fraction restored CFU-EC formation to 12.9 ± 4.7 (p = .56) (Fig. 4D). Interestingly, the final CFU-EC product consisted of >99% CD45+ cells and did not upregulate vWF in our experiments (E.R. and D.S., manuscript in preparation).

We also tested whether monocytes or their progeny become involved in the formation of vascular networks in typical angiogenic function assays in vitro. Whereas we reproducibly observed the formation of cord-like structures in monocyte liquid cultures, neither fresh nor EGM-2-cultured monocytes showed a tendency to build vascular networks in quadruplicate Matrigel assays. To some extent, they distributed within the extracellular matrix equivalent without network creation. Negative control HL-60 myelomonocytic leukemia cells did not show any tendency to migrate and settled by gravity to the middle of the gel surface (Figs. 5A, 5B). Positive control vascular ECs seeded in this assay displayed the typical three-dimensional vascular network formation after 12 to 18 hours (Figs. 5C, 5D). Furthermore, we could not observe substantial network formation or integration of monocytes or monocyte-derived cultured cells into the tubular structures of coseeded vascular ECs (Figs. 5C, 5E, 5F). Dil-Ac-LDL-labeled vascular ECs were as active in network formation as unlabeled or lectin-labeled ones, thus excluding the possibility that the labeling procedure had hindered network formation.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. References

The results of this study demonstrate that Dil-Ac-LDL uptake, lectin binding, and CD31/CD105/CD144 expression are genuine features of blood monocytes. Primary monocytes already expressed most tested endothelial genes and proteins except VEGF-R2 and CD105 at even higher levels than their supposed EPC progeny. Compared with HUVECs and HMVECs, CD31 mRNA in the starting monocytes was expressed at comparable levels. For CD105 mRNA, an approximately 30-fold lower expression was measured by real-time RT-PCR. For other candidate EC marker genes, we found an up to 106-fold lower expression in monocyte preparations. Kinetic analysis of mRNA expression confirmed the flow cytometry data by showing a downregulation of all tested markers except VEGF-R2, which was expressed approximately 8,000-fold lower in monocyte-derived cells compared with control ECs and CD105, which was upregulated to the expression level found in HUVECs.

The common way of EPC identification is so far largely dependent on a phenotypic profile often focusing on Dil-Ac-LDL uptake, lectin binding, and CD31/CD105/CD144 expression. It is well established that circulating ECs (CECs) derived from vessel walls and a variety of BM-derived cells react with Dil-Ac-LDL and lectins and share the expression of endothelial-related markers, including CD34, CD31, CD105, CD144, CD146, and VEGF-Rs [2, 7, 2628]. Different protocols for the generation of cells that display this phenotype and are derived from various starting cell populations have been described so far [29]. These protocols can generally be divided into short-term differentiation of virtually nonproliferating cells (usually 4 to 10 days of culture) and late endothelial outgrowth of sometimes highly proliferating cells (usually more than 10 days and up to several months). Interestingly, both types of procedures have been considered to generate or contain progenitor cells. Recently, Ingram et al. [30] reintroduced the proliferative potential as a key defining aspect of EPCs, as it has been established for other types of progenitors. Using a (late) endothelial outgrowth assay, these authors succeeded in defining a hierarchy of high-and low-proliferative-potential endothelial lineage progenitors. Because these cells could be analyzed only after expansion in vitro, due to the rarity of the culture initiating cell population, the nature of the starting progenitors has not been discovered so far. It has been speculated that it should be a BM-derived progenitor because Hebbel and colleagues [27, 30] demonstrated that late endothelial outgrowth cells originate from BM and expanded 1,000-fold in vitro. These late endothelial outgrowth cells represent only a minority of blood CECs, with the majority being of recipient genotype, suggesting that most CECs in fresh blood are low proliferative potential cells originating from vessel walls [27]. In light of the discovery of high-and low-proliferative-potential cells also in normal HUVEC and HMVEC preparations [31], it is unclear whether these CD45-negative proliferating endothelial lineage cells originate from BM vasculature, BM-derived EC progenitors, or another transplantable cell population.

Another way to facilitate the understanding of the various putative vasculogenic cell types is to reconsider cell numbers attributable to the various cell populations. Approximately two total CECs per milliliter have been measured in whole blood of healthy individuals [32]. The CD34+/133+/KDR+ candidate hemangioblast, originally discovered as HSCs, has been estimated to circulate in the same low frequency of 3 to 5 cells/ml [7, 33]. The number of more than 2 × roliferating Ac-LDL+/lectin+ monocyte-derived cells obtained from 1 ml PB after only 4 to 6 days of angiogenic culture by far exceeds that of the above-described populations [14, 17, 34]. Monocyte-derived cells have been shown to express endothelial-specific surface markers (CD105/144/VEGFR-2/vWF/EC NO synthase), to take up Ac-LDL, to exhibit lamellar structures resembling Weibel-Palade bodies, and to form cord-like structures upon angiogenic stimulation [18, 19]. For the first time, our study shows by both protein and mRNA kinetic analysis that common markers previously thought to define EPCs are typically expressed by normal blood monocytes. Contrary to expectations in case of a monocyte-to-EPC transition, monocytes downregulate most of these markers during angiogenic culture. However, expression levels are still sufficient to allow for a mimicry of an EC or EPC phenotype by monocytes and monocyte-derived cells, respectively.

In this study we focused on well-established and frequently used in vitro assays, assessing the behavior of blood monocytes during their asserted priming toward the enigmatic EPC. The phenotypic profiling has been limited to 15 commonly used markers and gene expression analysis to 10 genes commonly used for EPC definition. The high variety of different LDL and lectin-binding structures avoided rational mRNA experiments so far. In additional functional experiments, monocyte-derived cells after angiogenic stimulation did not form vascular networks and did not integrate more than accidentally in a vascular network formed by control ECs in a three-dimensional extra-cellular matrix. Recent in vivo data show close adherence of BM-derived cells (including monocytes) to the lumen of new blood vessels but no integration within the endothelial layer during vascular regeneration [3537]. The lack of reproducible integration of monocytes or monocyte-derived cells into a growing network of vascular ECs in our experiments may actually resemble these in vivo observations. The obvious effects of monocyte-derived cells or of factors derived from these cells in terms of angiogenesis support are not challenged by these results [38, 39]. Studies of the effects of PB and BM-MNC implantations into ischemic myocardium in large animals revealed that the effects are not limited to angiogenesis and improved collateral perfusion but also included the supply of regulatory cytokines [40, 41]. At least some of these angiogenic growth factors have been proven to be derived from appropriately stimulated monocytes [21].

A peculiar role of monocytes or monocyte-derived regulatory factors may also be highlighted in the in vitro CFU-EC system. In this assay, which was developed to quantify EPCs [24], highly purified CD14+SORT cells reproducibly lacked any capacity to generate EPC progeny. We can show so far that the generation of the CFU-ECs is strictly dependent on the presence of monocytes. A detailed profiling of monocyte-derived cytokines operative in this system is currently underway. It is interesting to note that the exact origin of the EPCs in this assay is as unclear as their fate in vivo. CD34+/CD133+/VEGFR-2+ circulating HSCs are possible parental SC predecessors, but further studies are needed to eventually define candidate EPCs and to gain insight into the regulatory networks.

Based on our observations, together with the growing amount of published evidence, we fully agree with Rehman et al. [21] that the term EPC should be reserved for a cell population derived from a hemangioblast or a related type of SC [7, 42]. However, we cannot support the argument of these authors[21] that the Ac-LDL+/lectin+ cells may be referred to as circulating angiogenic cells because our data and data provided by other studies [18, 19, 21, 29, 43] give reason to continue to call them monocytes. The importance of monocytes and other BM-derived cells for therapeutic angiogenesis must not be restricted to generation of EPCs or soluble factors directly involved in EPC development. Additional hypotheses relate to the function of BM-derived cells as supporting cells during vascular repair [35, 44, 45]. Unraveling the mechanisms by which the various cell types contribute to vascular repair processes should therefore help to develop and optimize concepts for regenerative therapies. During this process, phenotypic results should be interpreted with adequate caution.

Disclosures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. References

The authors indicate no potential conflicts of interest.

thumbnail image

Figure Figure 1.. Monocytes cultured in EGM-2 plus 20% fetal calf serum and 50 ng/ml vascular endothelial growth factor-165 for 6 days were incubated with Dil-Ac-LDL (red) and counterstained with Bandeiraera simplicifolia-1 lectin-FITC (A) and Ulex europaeus I-FITC (green) (B). Cultured monocytes regularly formed strings or cord-like structures. Brown color is a result of vimentin reactivity visualized with diaminobenzidine (C). Freshly isolated CD14-purified (D) and CD14-depleted (E) cells were incubated with Dil-Ac-LDL, spun onto slides, and analyzed by fluorescence microscopy. Dil-Ac-LDL uptake could be demonstrated for CD14-positive monocytes. Pictures are representative of four experiments. Abbreviations: Dil-Ac-LDL, 1,1′-dioctadecyl-3,3,3′,3 β-tetramethylindo-carbocyanine perchlorate-labeled acetylated low-density lipoprotein; FITC, fluorescein isothiocyanate.

Download figure to PowerPoint

thumbnail image

Figure Figure 2.. Expression of markers commonly referred to as characteristic for endothelial cells (A) and hematopoietic markers on CD14+SORT monocytes (B) was measured by flow cytometry. Equal analyses were performed using monocytes purified by negative selection (MonoSORT) to question an influence of CD14 occupation (C, D). The Δ mean fluorescence intensity (Δ MFI) of freshly isolated, day-0 (equation image), and day-6 (□) cultured monocytes (EGM-2 plus 20% fetal calf serum and 50 ng/ml vascular endothelial growth factor-165) is shown as the mean ± standard error of mean of six (A, B) and four (C, D) experiments. ΔMFI was calculated by subtracting isotype control MFI from sample MFI. Significant mean ΔMFI differences in matched-pair analyses (p < .05) are marked by asterisks. Abbreviations: BS-1, Bandeiraera simplicifolia-1; UEA, Ulex europaeus.

Download figure to PowerPoint

thumbnail image

Figure Figure 3.. Comparison of mRNA expression levels between HUVECs, HMVECs, freshly isolated PB-MNCs, CD14-sorted monocytes (CD14+SORT) at day 0 and at day 6 after culture, negative enriched monocytes (MonoSORT) at day 0 and day 6 after culture, remaining MNCs after monocyte isolation (without CD14+ and without monocytes, respectively; n = 3) is shown as ΔΔCt ± SD for CD105, VEGFR-2, CD34, CD146, CD31, VEGFR-1, VEGFR-3, vWF, CD14, and CD45 (mean ± SD). mRNA expression levels of HUVEC serve as reference and are adjusted to 0. ΔCt is the β-actin normalized number of polymerase chain reaction cycles for each target. ΔΔCt was calculated as ΔCt [targetcell type] − ΔCt [targetHUVEC]. For example, a ΔΔCt of −5 represents 2 −5 = 32-fold lower, and a Δ ΔCt of 10 represents 210 = 1,024-fold higher expression of the target mRNA. Significant target mRNA expression differences compared with HUVECs are marked by asterisks (p < .05). Abbreviations: HMVEC, human microvascular endothelial cell; HUVEC, human umbilical vein endothelial cell; MNC, mononuclear cell; PB, peripheral blood; VEGFR, vascular endothelial growth factor receptor; vWF, von Willebrand factor.

Download figure to PowerPoint

thumbnail image

Figure Figure 4.. Mock-sorted peripheral blood mononuclear cells (MNCs) (A), CD14+SORT cells (B), CD14-depleted cells (C), and the reunited CD14+SORT/CD14-negative fractions (add backs) (D) were cultured in Endocult medium in fibronectin-coated plates. The nonadherent cells were transferred on day 2 into new culture dishes and evaluated on day 5. Typical clusters, described as endothelial cell colony-forming units, could be observed only in the MNC fraction and in cultures where isolated CD14-positive monocytes were added back to the CD14-depleted fraction before culture. Pictures are representative of four experiments.

Download figure to PowerPoint

thumbnail image

Figure Figure 5.. (A): Monocytes cultured under angiogenic conditions distributed within a Matrigel extracellular matrix equivalent without network formation. (B): Myelomonocytic leukemia cells (cell line HL-60) settled by gravity to the middle of the Matrigel surface. (C–F): Dil-Ac-LDL-labeled (red) 4-day cultured monocytes were seeded with CD146-FITC (green)-labeled human umbilical vein endothelial cells at a 1:1 ratio to build vascular structures within Matrigel. After 12 to 18 hours, photomicrographs were taken with light microscopy (C), fluorescent microscopy with a filter for FITC (D), a Texas red filter (E), and a dual bloc for red-green fluorescence (F). Neither substantial network formation nor more than accidental integration of cultured monocytes into the endothelial network of vascular endothelial cells was observed. Pictures are representative of four independent experiments. Similar results were obtained in experiments with fresh monocytes (data not shown). Abbreviations: Dil-Ac-LDL, 1,1′-dioctadecyl-3,3,3′,3 β-tetramethylindo-carbocyanine perchlorate-labeled acetylated low-density lipoprotein; FITC, fluorescein isothiocyanate.

Download figure to PowerPoint

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Disclosures
  8. References