Article Text

Interventional cardiology
Dissociation of phenotypic and functional endothelial progenitor cells in patients undergoing percutaneous coronary intervention
  1. N L Mills1,
  2. O Tura2,
  3. G J Padfield1,
  4. C Millar3,
  5. N N Lang1,
  6. D Stirling3,
  7. C Ludlam3,
  8. M L Turner2,
  9. G R Barclay2,
  10. D E Newby1
  1. 1
    Centre for Cardiovascular Science, Edinburgh University, UK
  2. 2
    Centre for Regenerative Medicine, Edinburgh University, UK
  3. 3
    Department of Haematology, Royal Infirmary of Edinburgh, UK
  1. Correspondence to Dr Nicholas L Mills, Centre for Cardiovascular Science, University of Edinburgh, Chancellor’s Building, Edinburgh, EH16 4SU, UK; nick.mills{at}


Objectives: Endothelial progenitor cells (EPCs) are circulating mononuclear cells with the capacity to mature into endothelial cells and contribute to vascular repair. We assessed the effect of local vascular injury during percutaneous coronary intervention (PCI) on circulating EPCs in patients with coronary artery disease.

Design and setting: Prospective case-control study in a university teaching hospital.

Patients: 54 patients undergoing elective coronary angiography.

Interventions and main outcome measures: EPCs were quantified by flow cytometry (CD34+KDR+ phenotype) complemented by real-time polymerase chain reaction (PCR), and the colony forming unit (CFU-EC) functional assay, before and during the first 24 hours after diagnostic angiography (n = 27) or PCI (n = 27).

Results: Coronary intervention, but not diagnostic angiography, resulted in an increase in blood neutrophil count (p<0.001) and C-reactive protein concentrations (p = 0.001) in the absence of significant myocardial necrosis. Twenty-four hours after PCI, CFU-ECs increased threefold (median [IQR], 4.4 [1.3–13.8] vs 16.0 [2.1–35.0], p = 0.01), although circulating CD34+KDR+ cells (0.019% (SEM 0.004%) vs 0.016% (0.003%) of leucocytes, p = 0.62) and leucocyte CD34 mRNA (relative quantity 2.3 (0.5) vs 2.1 (0.4), p = 0.21) did not. There was no correlation between CFU-ECs and CD34+KDR+ cells.

Conclusions: Local vascular injury following PCI results in a systemic inflammatory response and increases functional CFU-ECs. This increase was not associated with an early mobilisation of CD34+KDR+ cells, suggesting these cells are not the primary source of EPCs involved in the immediate response to vascular injury.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Ischaemic heart disease is a major cause of morbidity and mortality worldwide. Despite advances in percutaneous coronary intervention (PCI), major adverse cardiac events occur in up to 30% of patients following balloon angioplasty and in 20% following stenting.1 Vascular trauma, induced by percutaneous intervention, initiates a sequence of events in which the release of cytokines and growth factors results in the proliferation of smooth muscle and deposition of platelets and leucocytes at the site of injury, accelerating vascular repair. Endothelialisation is necessary to prevent mural thrombus formation and neointimal hyperplasia that may otherwise lead to ischaemic complications and restenosis.

The traditional paradigm of vascular repair is based on the proliferation and migration of pre-existing mature endothelial cells from the adjacent vasculature.2 The discovery by Asahara and colleagues that mononuclear cells in peripheral blood have the potential to differentiate into endothelial cells has launched a new field of cardiovascular research.3 Endothelial progenitor cells (EPCs) have been characterised by their expression of both haematopoietic (CD34+) and endothelial cell antigens (KDR+), and by their ability to proliferate, migrate and differentiate into mature cell types. These putative EPCs form vascular structures in vitro and are incorporated into the vessel wall in experimental models of neovascularisation.4 These cells may have an important role in the maintenance and repair of the vascular endothelium, and in the pathogenesis of atherosclerotic plaque formation and its consequences.

Endothelial progenitor cells can be isolated and cultured from a variety of cell populations in peripheral blood and bone marrow, but as yet no definitive phenotype has been ascribed to EPCs. Comparisons between clinical studies have been limited by the use of a variety of phenotypic markers to discriminate EPCs and by the lack of comparable functional assays. In the face of an uncertain phenotype, the EPC colony forming unit assay (CFU-EC) has emerged as an alternative specific enumeration system for EPCs.5 Although groups are increasingly quantifying either phenotypic EPCs (CD34+KDR+) or functional CFU-ECs, few clinical studies report both or comment on the relation between phenotype and function.

Endothelial progenitor cells are infrequent in peripheral blood, but numbers increase rapidly in response to myocardial ischaemia and acute myocardial infarction.6 7 8 Reduced numbers of EPCs have been demonstrated in cigarette smokers,9 patients with diabetes mellitus10 and in those with evidence of endothelial dysfunction.11 These patients are at high risk of complications following PCI. Furthermore, patients with diffuse in-stent restenosis have reduced EPC number and function in comparison with matched controls at the time of presentation.12 Inadequate EPC number and function before angioplasty, as well as inadequate early and sustained EPC recruitment, may favour a maladaptive response to arterial injury and result in an increased incidence of in-stent thrombosis, restenosis and ischaemic complications. The immediate effects of local vascular injury during angioplasty and stenting on the mobilisation of EPCs are not known. The aim of the present study was therefore to measure circulating phenotypic EPCs and functional CFU-ECs following PCI in patients with stable coronary heart disease.



Fifty-four patients undergoing elective coronary angiography participated in this study, which was performed with the approval of the local research ethics committee, in accordance with the Declaration of Helsinki, and the written informed consent of all volunteers. All patients were recruited following referral for angiography to investigate symptoms suggestive of stable angina. Patients with a recent acute coronary syndrome or coronary intervention (<3 months), renal or hepatic failure, or a systemic inflammatory disorder or malignancy were excluded from the study. Twenty-seven patients underwent diagnostic coronary angiography alone, and 27 required balloon angioplasty and stenting because of flow limiting luminal stenosis of a major epicardial vessel (table 1).

Table 1

Clinical characteristics and angiographic findings of patients undergoing diagnostic angiography or percutaneous coronary intervention

Coronary angiography and PCI

All patients were treated for 2 weeks with 75 mg clopidogrel before angiography or PCI. Coronary angiography was performed via right femoral or radial artery with 6F arterial catheters. Elective PCI was performed in all patients after 5000 IU intravenous heparin administration and in one patient after intravenous glycoprotein IIb/IIIa inhibitor. Coronary stents (Liberté, Boston Scientific) were implanted in all patients after balloon predilatation of the lesion without apparent procedural complications.

Blood sampling and assays

A venous cannula (17-gauge) was inserted into a large subcutaneous vein of the antecubital fossae for blood sampling before, immediately after and at 6 hours and 24 hours following angiography. EDTA anti-coagulated blood (Sarstedt-Monovette, Germany) was collected for flow cytometry, and for preparation of plasma for storage in all subjects. Whole blood was analysed for total cells, differential count and platelets using an autoanalyser (Sysmex, UK). Plasma troponin I concentrations were measured using an automated immunometric assay (Ortho-clinical Diagnostics, High Wycombe, UK). Serum C-reactive protein (CRP) concentrations were measured using an immunonephelometric assay (Behring BN II nephelometer, Marburg, Germany). In 40 subjects (20 diagnostic angiograms and 20 PCI), 20 ml of whole blood was drawn at baseline and at 24 hours for mononuclear cell preparation, cell culture and real-time polymerase chain reaction (PCR).

Flow cytometry

Whole blood cells were phenotyped by flow cytometry. Cells were directly stained and analysed for phenotypic expression of surface proteins using pre-conjugated anti-human monoclonal antibodies including anti-CD34-FITC, anti-CD45-PerCP (Becton Dickinson, Oxford, UK) and anti-KDR-PE (R&D systems, Minneapolis, USA). Appropriate negative controls (isotype and/or no antibody) were used to establish positive stain boundaries. Undiluted samples (100 μl) were stained with antibodies for 30 minutes in the dark. Erythrocytes were lysed (lysing solution, Becton Dickinson), and samples were centrifuged at 200 g for 10 minutes, washed with phosphate buffered saline (PBS), and fixed (Cell Fix solution, Becton Dickinson). For each sample, 50 000 events were acquired in the lymphocyte region (as determined by characteristic forward and side scatter profile) using a FACS Calibur flow cytometer (Becton Dickinson) and data were analysed using FCS Express (DeNovo Software). The number of leucocytes per ml of blood was measured using an automated cell counter in our regional haematology laboratory. EPCs were quantified based on the percentage of CD34+KDR+ double positive leucocytes and expressed as number of cells per ml of blood.

Endothelial progenitor colonies

Mononuclear cells were isolated by Ficoll density gradient separation, washed twice with PBS and resuspended at 2.5×106 cells/ml in complete endothelial culture medium (CECM, Stem Cell Technologies, UK). A volume of 2 ml of cell suspension was placed into each of six-well Fibronectin-coated plates (Becton Dickinson) and incubated at 37°C, 5% CO2 with 95% humidity. After 2 days the non-adherent cells were transferred to a fibronectin-coated 24-well plate (Becton Dickinson) at 1×106/well for 3 days. Colonies (CFU-EC, early outgrowth colony forming unit endothelial cells; Stem Cell Technologies) were defined as a central core of “round” cells with elongated “sprouting” cells at the periphery5 and were counted on day 5 in a minimum of four wells by observers unaware of the subjects’ clinical profiles. In order to confirm endothelial-cell lineage, direct staining was performed on colonies using acetylated low-density lipoprotein (LDL) and co-stained with lectin (Ulex europaeus I agglutinin). The endothelial phenotype was further characterised by immunohistochemistry using endothelial surface markers.

RNA extraction and quantitative real-time PCR

Total leucocyte RNA extraction from 1 ml of whole blood was performed using Qiagen’s RNeasy Mini Kit (Qiagen Ltd, Crawley, UK). A volume of 1 μg of total RNA was transcribed into cDNA in each reverse transcription reaction with 200 units of M-MLV reverse transcriptase for 60 minutes at 37°C in 20 μl reactions containing 1 μl (0.5 μg/μl) of random hexamer primers, with 0.625 μl (40 units/μl) of RNAse inhibitor, 5 μl of dNTP mix and 5 μl of 5X RT reaction buffer.

Real-time PCR was carried out using the ABI Prism 7900HT system (Applied Biosystems, Warrington, UK) to determine the relative quantity of mRNA. PCR primers and probes for amplification of cDNA derived from CD34 and CD14 transcripts were obtained from Applied Biosystems (Foster City, CA, USA). Each assay contained forward and reverse PCR primers and one Taqman MGB probe. A volume of 4 μl of the reverse transcription reaction was analysed in each PCR reaction. The PCR reactions were run in triplicate. Analysis was performed using ABI 7900HT SDS software in order to obtain the relative quantities of mRNA compared to a calibrator.

Data analysis and statistics

Analysis was performed by an observer blinded to whether patients had undergone angiography or PCI. Statistical analyses were performed with GraphPad Prism (Graph Pad Software, USA) using a two-tailed Student t test, repeated measures ANOVA, Mann-Whitney U or Wilcoxon paired tests where appropriate. D’Agostino and Pearson omnibus normality test was used to assess whether parameters were normally distributed with continuous variables are reported as mean (SEM) or median [IQR]. Statistical significance was taken at p<0.05.


There were no complications arising from angiography or PCI and all patients were discharged home 24 hours after the procedure. Flow limiting luminal stenosis of a major epicardial vessel was identified in 41 patients. In 27 patients, PCI was performed to treat 36 lesions with an average vessel diameter of 2.9 (0.1) mm at the point of stent deployment.

Inflammation and myocyte necrosis

Diagnostic angiography did not increase peripheral blood leucocyte count or serum C-reactive protein concentrations (table 2). Coronary intervention increased neutrophil count (Δ 0.9 (0.3) × 109/l, p<0.001) and serum C-reactive protein concentrations (Δ 1.5 [0.4–2.0] mg/l, p = 0.001) at 24 hours. There was a transient reduction in monocyte count immediately following PCI (Δ −0.07 (0.02)×109/l, p = 0.004) and cardiac catheterisation alone (Δ −0.10 (0.03) × 109/l, p = 0.006). Monocyte count returned to pre-procedure levels by 24 hours. There was little evidence of myocyte necrosis following diagnostic angiography or PCI, with plasma cardiac troponin I at 24 hours increased in two patients following angiography and four patients following PCI: median concentration 0.2 [0.2–0.69] ng/ml and 0.2 [0.2–1.6] ng/ml respectively (p = 0.30).

Table 2

Markers of inflammation following diagnostic angiography or percutaneous coronary intervention

Endothelial progenitor cells

Direct staining confirmed that CFU-ECs, like mature endothelial cells, bind lectin and integrate acetylated-LDL and express a range of endothelial surface proteins. Although unaffected by diagnostic angiography, the number of CFU-ECs increased threefold 24 hours after PCI (median [IQR], 4.4 [1.3–13.8] vs 16.0 [2.1–35.0], p = 0.01; fig 1).

Figure 1

(Upper panel) The number of endothelial progenitor cell colony forming units (CFU-ECs) increased following percutaneous coronary intervention (PCI) compared to baseline (median [IQR], 4.4 [1.3–13.8] vs 16.0 [2.1–35.0], p = 0.01), but were not changed following diagnostic angiography alone (median [IQR], 4.7 [0–21.5] vs 3.2 [1.1–9.7], p = 0.70). (Lower panel) (A) Phase contrast microscopy of a typical CFU-EC (magnification ×20). Cultured EPCs were counterstained with a nuclear dye DAPI (blue) and stained for a range of markers used to characterise a mature endothelial phenotype. (B) Acetylated low-density lipoprotein (red) and with Ulex Europaeus (green). (C) Angiopoietin receptor Tie-2 (red) and transmembrane glycoprotein endoglin (CD105; green). (D) Melanoma cell adhesion molecule MCAM (CD146; red) and vascular endothelial growth factor receptor-2 (KDR; green). (E) CD34 a stem cell marker expressed on capillary endothelium. (F) Vascular endothelium adhesion molecule VE-cadherin (CD144). (G) Platelet endothelial cell adhesion molecule-1 PECAM-1 (CD31). (H) Endothelial nitric oxide synthase (eNOS).

The increase in CFU-ECs or functional EPCs was not associated with an increase in the number of circulating CD34+KDR+ cells (table 3). Interestingly there was no correlation between EPCs identified by phenotype (CD34+KDR+ cells) and the number of functional EPCs quantified using cell culture (CFU-ECs) either before (r = −0.24, p = 0.10) or 24 hours after (r = −0.09, p = 0.61) angiography. Similarly, there was no increase in leucocyte CD34 mRNA following angiography or PCI. There was, however, a reduction in leucocyte CD14 mRNA immediately after catheterisation in both diagnostic and interventional studies that coincided with the reduction in circulating monocytes (see Heart online for data supplement). Relative quantities of leucocyte CD14 mRNA increased significantly 24 hours following PCI (p<0.05).

Table 3

Circulating CD34+KDR+ cells following diagnostic angiography or percutaneous coronary intervention


We have demonstrated that PCI, but not diagnostic angiography, is associated with a systemic inflammatory response and an increase in functional CFU-ECs in peripheral blood. However, these effects were not associated with increases in leucocyte CD34 mRNA or circulating CD34+KDR+ phenotypic EPCs. In this clinical model of acute local vascular injury, we suggest that the immediate response to injury is mediated by CFU-EC rather than the primitive CD34+KDR+ cells that have hitherto been considered as the major endothelial progenitor cell source in blood and bone marrow. Circulating CD34+KDR+ cells may ultimately be capable of endothelial cell differentiation, but are rare in peripheral blood and are not mobilised rapidly in the early response to acute local vascular injury.

Functional EPCs (CFU-EC) are reduced in people with cardiovascular risk factors and evidence of vascular impairment, raising the possibility that limited EPC reserves may contribute to a maladaptive response to vascular injury and predispose to atheroma formation.5 Previous studies have demonstrated early mobilisation of EPCs following vascular injury in patients with acute myocardial infarction,7 major burns and following coronary artery bypass graft (CABG) surgery.13 These clinical events involve extensive damage to a number of tissues in addition to the vasculature that may contribute to the mobilisation of progenitors. Our findings suggest that local selective vascular injury can influence the number of circulating progenitor cells, and provide a rationale for the therapeutic mobilisation of EPCs at the time of PCI to influence outcome.

Current strategies to reduce the incidence of complications following percutaneous intervention are based on suppressing cellular proliferation rather than enhancing vascular repair. Vessel injury during PCI exposes underlying collagen and tissue factor, activating platelets and the coagulation cascade, and may result in acute or sub-acute stent thrombosis.14 In the absence of an intact endothelium, local platelet/platelet and platelet/leucocyte complexes form, and persistent inflammation encourages smooth muscle hypertrophy and in-stent restenosis. Drug-eluting stents have dramatically reduced the incidence of early in-stent restenosis,15 but local anti-proliferative therapy may interfere with vascular healing and prevent formation of a functional endothelial layer.16 In experimental studies, transfusion of EPCs following vascular injury prevents both thrombus formation and neointimal proliferation.17 Indeed, patients with diffuse in-stent restenosis have reduced numbers of circulating EPCs in comparison with matched controls.12 Stents coated with antibody to CD34 may encourage seeding of EPCs to facilitate the early formation of an endothelial layer. In early feasibility studies, such stents have been safely deployed in humans.18

Coronary intervention increased the number of CFU-ECs but was not associated with a mobilisation of CD34+KDR+ cells. These findings initially appear discordant and require further discussion. The original description of the putative endothelial progenitor cell was based on cell culture and adhesion techniques. Asahara et al described a population of adult human circulating CD34+ cells that could differentiate into cells with endothelial-like characteristics in vitro.3 The exact origin and phenotype of these progenitors remains a matter of debate, in part because the purity of CD34+ cells used in this initial study was only 15%.3 Subsequently the co-expression of transmembrane glycoproteins CD34 and vascular endothelium growth factor receptor-2 (VEGFR-2 or KDR) has been used in an increasing number of clinical studies to phenotype and to quantify circulating EPCs. As the field has developed, an increasing number of methods have emerged to define vascular progenitors and quantify regenerative capacity. The CFU-EC assay has been used widely since it was first described by Hill et al.5 While quantification of CFU-EC provides an accurate measure of the capacity of circulating mononuclear cells to form endothelial cells, it is doubtful whether these colonies arise directly from the circulating CD34+ stem cells. Consistent with our previous findings19 and those of George et al,20 we found no correlation between the number of peripheral blood CD34+ cells and the number of CFU-ECs.

Studies addressing the origin of endothelial progenitor lineage in adult peripheral blood have demonstrated that monocytes also express endothelial lineage markers such as VEGFR-2 and can differentiate into mature endothelial cells.21 Rehman et al found that the majority of CFU-ECs expressed monocyte markers such as CD14, Mac-1 and CD11c, suggesting that peripheral blood EPCs are derived from monocyte-like cells.22 These findings are confirmed in the elegant studies by Rohde et al in which they demonstrate formation of colony forming units by mononuclear cells depleted of CD34+ cells and report that the CFU-ECs are primarily composed of monocytic and lymphocytic cells.23

The concept that functional endothelial cells may originate from a CD14 expressing progenitor is supported by reports that mature endothelial cells isolated from human umbilical vein express CD14.24 Furthermore, Urbich et al demonstrate that isolated CD14+ cells also have the capacity to improve neovascularisation after hindlimb ischaemia.25 CD14+ monocyte-like cells are more abundant in normal peripheral blood than in bone marrow and are therefore capable of rapidly homing to sites of vascular injury. While CD34+KDR+ cells may ultimately give rise to endothelial cells, they are much less prevalent in peripheral blood than in bone marrow, and are not mobilised following PCI. Our findings are consisent with those recently published by Thomas et al,26 and together these studies do not suggest an important role for CD34+KDR+ cells in the early response to vascular injury.

While we did not specifically measure CD14+ cells by flow cytometry, we did quantify monocyte numbers and total leucocyte CD14 mRNA. We identified a reproducible decrease in both variables immediately after cardiac catheterisation. It is possible that these cells immediately localise to the site of vessel damage: both at the site of arterial puncture and at the site of coronary angioplasty and stenting. The number of peripheral blood monocytes was restored to pre-procedural levels by 24 hours, and CD14 mRNA levels increased 24 hours after PCI. Mobilisation of CD14+ monocyte-like cells or a change in the function of these cells through upregulation of surface protein expression may explain the increase in CFU-ECs in peripheral blood observed 24 hours after arterial injury. These cells may contribute to vascular repair either through formation of mature endothelial cells and incorporation into the vessel wall, or through the release of angiogenic growth factors at the site of vessel injury.

It may take up to 3 months for a complete functional endothelial layer to form following percutaneous coronary intervention,27 and we cannot discount a role for CD34-derived progenitors, perhaps released in response to secondary angiogenic factors, in the later stages of vascular repair. It is also possible that other putative EPC populations not measured in the present study, such as those expressing the stem cell marker CD133, may have a role in the response to vascular injury. Further studies are required to explore specifically the role of CD14+ and CD133+ subpopulations, to define the time course of this response to vascular injury and to assess the impact of EPC mobilisation on restenosis and clinical outcomes.


Acute local vascular injury following angioplasty and stenting results in a systemic inflammatory response and increases the number of functional CFU-ECs. Circulating CD34+KDR+ cells are rare in peripheral blood, and are not mobilised rapidly in the early response to vascular injury. We suggest that the acute response to injury is mediated by CFU-EC derived from circulating monocytic cells rather than the primitive CD34+KDR+ cells that have hitherto been considered as the major endothelial progenitor cell source in blood and bone marrow. A better understanding of the cellular response to vascular injury is necessary to allow a more sophisticated approach to reducing the complications of percutaneous coronary intervention with new strategies designed to enhance vessel repair.


We thank Pamela Dawson, Marion Rae, Sharon Cameron, Tom McColm and all the staff at the Wellcome Trust Clinical Research Facility, Edinburgh, for their assistance with the studies.


Supplementary materials


  • See Editorial, p 1971

  • ▸ Additional data supplement is published online only at

  • NLM and OT contributed equally to first authorship.

  • Funding Michael Davies British Cardiovascular Society Research Fellowship (NM); British Heart Foundation Project Grant (PG/07/017/22405); National Health Service Research and Development Fund (SPG2005/27).

  • Competing interests None declared.

  • Provenance and Peer review Not commissioned; externally peer reviewed.

Linked Articles