Circulating Humoral Factors and Endothelial Progenitor Cells in Patients With Differing Coronary Collateral Support
Background— The mechanisms underlying the variation in collateral formation between patients, even with similar patterns of coronary artery disease, remain unclear. This study investigates whether circulating humoral or cellular factors can provide an insight into this variation.
Methods and Results— Thirty patients with isolated left anterior descending coronary artery disease underwent percutaneous coronary intervention with collateral flow index (CFI) determined using a pressure wire. Patients with inadequate (CFI <0.25) compared with those with adequate (CFI ≥0.25) collateral support had, or tended to have, lower concentrations of coronary sinus growth factors and plasma exerting a weaker effect on endothelial cell migration and angiogenesis in vitro. However, there was an inverse correlation between serum mitogenicity and CFI (r=−0.61, P<0.01). No significant differences were detected between the 2 groups in plasma levels of total vascular endothelial growth factor, vascular endothelial growth factor165, or placental growth factor. There was a strong positive correlation between numbers of CD34/CD133-positive circulating hemopoietic precursor cells and CFI (r=0.75, P<0.001). In patients with inadequate, compared with those with adequate, CFI, the numbers of differentiated endothelial progenitor cells (EPCs) appearing in the circulation and in culture were significantly reduced by 75% (P<0.05) and 70% (P<0.05), respectively.
Conclusions— In this study, inadequate coronary collateral development is associated with reduced numbers of circulating EPCs and impaired chemotactic and proangiogenic but not mitogenic activity. These findings are consistent with current efforts to enhance collateral formation by augmentation of circulating EPCs.
Received January 6, 2004; revision received February 27, 2004; accepted March 4, 2004.
There is marked heterogeneity in the degree of coronary collateral development between patients, even when their patterns of coronary artery disease (CAD) are similar.1–3 However, the humoral and cellular mechanisms underlying this variability remain unclear. Clinical trials of angiogenic gene and peptide therapy are based on the paradigm that patients with inadequate collateral support are deficient in mitogenic growth factors.4 Emerging evidence indicates that bone marrow–derived circulating endothelial progenitor cells (EPCs) participate in postnatal neovascularization.5–7 These cells are reduced in both number and migratory activity in patients with risk factors for coronary disease or endothelial dysfunction.8,9 EPCs coexpress CD34, CD133 (hematological stem cell markers), and kinase-insert domain-containing receptor (KDR) (vascular endothelial growth factor [VEGF] receptor 2) and develop an endothelial phenotype in culture.10,11 In addition, they incorporate into developing collaterals in animal models of ischemia and thereby enhance neovascularization.12–14 These observations have resulted in the first clinical trial of EPC therapy after myocardial infarction.15
Despite this improved understanding of the circulating factors that may contribute to collateral growth, little is known of their relative contribution to collateral formation in patients. This study, in a carefully characterized population of patients with single-vessel disease of the left anterior descending coronary artery (LAD), was designed to investigate the humoral and cellular arms of the angiogenic response.
Patients were divided into the following 2 groups according to their pressure-derived collateral flow index (CFI): ↓Coll, CFI <0.25; and ↑Coll, CFI ≥0.25 (see below). Ischemic thresholds on exercise were assessed before percutaneous coronary intervention (PCI). Blood samples collected from the coronary sinus (CS) and aortic root before heparinization were used for determination of plasma growth factor concentration by ELISA and for in vitro biological assays to assess the effects of plasma on angiogenesis and chemotaxis and serum on mitogenicity. Peripheral venous samples taken within 4 weeks of PCI were analyzed by flow cytometry to determine the fraction of circulating EPCs and by cell culture to determine endothelial cell maturation.
Thirty patients (mean age, 63±4 years; 80% male) with a positive exercise ECG and single-vessel LAD disease awaiting PCI were identified prospectively. The protocol was approved by the St Thomas’ Hospital local research ethics committee, and all patients provided written informed consent.
Cardiac Catheterization and Sample Collection
Catheters were passed into the distal third of the CS, ensuring collection of venous blood from the LAD territory. Aortic root and CS samples were collected at baseline before heparinization. Blood collected in EDTA for plasma and SST tubes (Becton Dickinson) for serum was placed immediately on ice and within 30 minutes spun at 3000g at 4°C for 30 minutes before aliquoting and storage at −80°C. The severity of the LAD stenosis was measured from orthogonal views using a quantitative coronary angiography package (Quantim).
Measurement of Pressure-Derived CFI
CFI was calculated after at least 120 seconds of balloon inflation from coronary occlusion pressure (Poccl) using the physiological derivation of collateral flow according to Pijls and colleagues.16,17 Pharmacological vasodilators were not used.
Plasma and serum samples were analyzed using ELISAs and in vitro assays of mitogenicity, angiogenesis, and chemotactic potential. Details of these assays can be found in the accompanying online data supplement.
Isolation, Cultivation, and Characterization of Endothelial Progenitor Cells
The peripheral blood mononuclear cell (PBMNC) fraction isolated by density gradient centrifugation from 40 mL peripheral blood underwent both flow cytometric analysis and culture.5 PBMNCs were plated on chamber slides coated in fibronectin (4 μg/mL) and gelatin (50% vol/vol) at a density of 4×106 cells/well. Culture medium was M199 supplemented with endothelial growth medium (EBM-2, Sigma) and 20% FCS. After 72 hours in culture, nonadherent cells were removed by washing with PBS, and remaining cells were maintained in culture for a total of 7 days. They were then fixed in 2% paraformaldehyde and underwent staining for von Willebrand factor (vWF) (rabbit anti-human mAb-Sigma), CD31, VE cadherin, and KDR (mouse Ab, Santa Cruz) and were visualized by confocal microscopy using monoclonal secondary antibodies conjugated to Alexa 488 and 568 (Molecular Probes). Image acquisition and analyses were undertaken by observers blinded to the CFI of the patients. The mean number of EPCs identified per field from 3 random high-power images was recorded from each of 3 triplicate wells per patient or per control. Control PBMNCs were obtained by venesection of 5 volunteers without a history of heart disease.
Flow Cytometric Analysis
Aliquots of 106 PBMNCs were incubated with monoclonal antibodies against human KDR (Autogen-Bioclear), followed by a PerCP conjugated secondary antibody (DAKO), PE-conjugated anti-CD 133 (Miltenyi), and FITC-conjugated anti-CD34 monoclonal antibody (BD). Isotype-identical antibodies served as controls (BD, Sigma). After incubation, cells were washed with PBS, fixed in 2% paraformaldehyde, and analyzed on a Becton Dickinson Facstar II flow cytometer. Each analysis consisted of 200 000 events.
Data are expressed as mean±SEM. Continuous variables tested for normal distribution by the Kolmogorov-Smirnov test were compared by unpaired t test or 1-way ANOVA. In cases of nonnormal distribution, nonparametric tests were used (Mann-Whitney U test). Linear regression analysis was used to assess correlations in the humoral and cellular assays with CFI. Statistical significance was accepted if the null hypothesis could be rejected at P=0.05. All statistical analyses were performed using SPSS for Windows 7.0.
Patient Characteristics and Clinical Data
A CFI <0.25 is established on the basis of theory and clinical observation as the threshold below which collateral flow is inadequate to prevent rest ischemia during coronary occlusion.18 This dichotomy limit was validated in our study population, because subjects with ↓Coll developed 11.6-fold greater peak ST elevation during coronary occlusion than the ↑Coll group. These well-collateralized patients had a greater mean percentage diameter stenosis (92% versus 84%; P=0.04). No statistically significant differences existed between the two study groups in terms of age, number of cardiovascular risk factors, and use of statins. Importantly, diabetes and hypercholesterolemia, conditions associated with impaired collateral formation, were equally prevalent (Tables 1 and 2⇓).
Transmyocardial Growth Factor Concentrations
Growth factor levels measured before heparinization in aortic root and CS are presented in Table 3. A significant transmyocardial gradient for basic fibroblast growth factor (FGF) was identified in both groups, with the highest CS levels measured in the ↑Coll group. There was a tendency for CS levels of VEGF to exceed those in the aorta in the ↑Coll, but not the ↓Coll, group. The lack of statistical significance may be because the study was underpowered to detect such a difference. Placental growth factor did not seem to differ by group or sample site.
The rate of DNA synthesis was 80% greater in human umbilical vein endothelial cells (HUVECs) exposed to serum samples from ↓Coll versus ↑Coll patients (Figure 1A). There was a significant inverse correlation between CFI and relative mitogenic activity in HUVECs (Figure 1B) but no transmyocardial gradient in mitogenic activity (data not shown). Anti-VEGF and anti-FGF antibodies diminished mitogenicity to the same extent in CS samples from both groups (data not shown). However, anti-FGF Ab caused a mean 34% inhibition, whereas anti-VEGF antibody caused only 14% inhibition.
In Vitro Angiogenesis Assay
Four patients were excluded from the in vitro angiogenesis assay because of loss of viability of the HUVECs or cell clumping in the Matrigel wells. CS plasma 10% from ↑Coll patients stimulated significantly greater in vitro tube formation than plasma from ↓Coll patients. In addition, the resulting primitive tubular connections were more stable and had increased branching (Figure 2A).
Transwell Migration Assay
Six patients were excluded as a result of bacterial contamination. The chemotactic activity of patients’ plasma mirrored the in vitro angiogenesis assay activity. The weakest migration response to CS plasma occurred in the ↓Coll CS patients with the lowest tertile of CFI (<0.13) (Figure 2B). When the data were analyzed using the dichotomy limit of 0.25 for CFI, the ↓Coll group had a chemotaxis rate of 48% of control plasma and ↑Coll group 80% of control plasma (P=0.07).
The first 7 patients recruited into this study were not included in EPC characterization because of inadequate optimization of techniques. Circulating EPCs are characterized by coexpression of CD34, CD133, and KDR. There were no significant differences in the total number of CD34-positive cells between age-matched controls without known CAD and the patients with LAD stenoses (0.35%±0.05% versus 0.38%±0.06%). There was a strong positive correlation between CD34/133-positive hemopoietic stem cells and CFI, r=0.75, P<0.0001 (Figures 3A and 3B). We found that 8% of the CD34/CD133-positive cells also expressed KDR in the lymphocyte gate. All KDR-positive cells expressed CD133.
Circulating EPC numbers were reduced in the study patients compared with control subjects (CD34/KDR-positive, 0.016% versus 0.04%, respectively; P<0.05). ↑Coll patients had a 4-fold greater circulating EPC fraction than the ↓Coll group (Figure 4), approaching that seen in the control subjects. We did not identify any significant correlation between number of risk factors and CD34/KDR+ve or CD34/CD133+ve cell numbers. There were no differences in the numbers of circulating CD34/CD133+ve or CD34/KDR+ve cells in patients receiving statins versus those not receiving statins (0.23% versus 0.18% or 0.028% versus 0.012%, respectively).
EPC Culture Assay
Cells with an endothelial phenotype were derived from the PBMNC fraction and characterized after 7 days of culture. Figure 5A illustrates confocal microscopic images of vWF and KDR expression. A total of 99% of EPCs expressed vWF and KDR, and 30% coexpressed vWF and CD31 (not shown). This culture assay only partially reflects the higher circulating EPC pool identified by flow cytometry in the ↑Coll patients. The higher number of EPCs cultured from the ↑Coll patients could have arisen from either higher numbers of cells adhering initially or increased propagation of attached cells from days 3 through 7. As a group, these patients formed fewer endothelial cells in culture than healthy controls (Figures 5B and 5C). However, cultured endothelial cell numbers from ↑Coll patients approached control levels, with a positive correlation between endothelial cell density and CFI (Figure 5D).
This is the first study to investigate the relationship between circulating EPC populations and native coronary collateral development. These data reveal significant differences in the cellular and humoral components thought to contribute to the variability in collateral support in patients with flow-limiting LAD stenoses. In summary, patients with well-developed collaterals have more numerous circulating EPCs coupled with plasma that exerts a more powerful chemotactic and proangiogenic effect on endothelial cells in vitro.
Interrelationship Between Coronary Artery Stenosis Severity, EPCs, and Collaterals
Coronary artery stenosis severity is a long-recognized determinant of collateral development.2 In the current study, we also found that patients with well-developed collaterals had more severe LAD stenoses. This group of patients also had a greater number of circulating EPCs, and these cells more readily formed mature endothelial cells in culture. These observations can be interpreted in several ways. The most simplistic interpretation is that LAD stenosis severity is the key determinate of CFI and that the changes in EPCs are merely a reflection of the consequences of stenosis severity, such as myocardial ischemia. Within this scenario, the EPCs may either act as innocent bystanders, although this view is not consistent with our current understanding of the biological potential of these cells, or may act as envoys helping to establish the relationship between stenoses and collaterals. A more complex interpretation is that EPCs are a key determinate of collateral formation and of greater potential importance than stenosis severity. We believe this later interpretation is supported by our study and is also consistent with several other observations. For example, it is recognized that the considerable interindividual variation in collateral formation cannot be explained by variations in coronary artery stenosis severity.2 This is borne out in our current study, where percentage diameter stenosis only weakly determines CFI (r2=0.13). Even this determination coefficient may be overly optimistic, because in a much larger database of 450 patients, the value was even lower (r2=0.01) but nonetheless statistically significant.19 Lastly, the lack of a close relationship between collaterals and coronary artery stenosis severity has been most convincingly demonstrated by the great variability in CFI that exists between patients without flow-limiting coronary artery stenoses.3 Taken together, these observations suggest that factors other than stenosis severity mainly determine collateral flow. One such alternative factor may relate to the quantity and quality of circulating EPCs, a view supported by the current study, where the number of CD34/CD133+ve cells was more closely associated with CFI than stenosis severity.
Effects of Humoral Factors on Vascular Cell Behavior
The biological assays examined 3 specific processes relevant to collateral formation: endothelial cell proliferation, chemotaxis, and tube formation in Matrigel. The last of these is a complex assay integrating measures of cellular proliferation, chemotaxis, cell alignment, and apoptosis. The data from the Matrigel and migration experiments indicate that humoral factors in patients with a high CFI drive endothelial cells toward migration and cellular capillary-tube formation.
The inverse relationship between mitogenicity in HUVECs and CFI is counterintuitive but suggests differential effects on endothelial cell behavior, analogous to the differential modulation of migration and proliferation by FGF.20 This induction of a promigratory, rather than proliferative, endothelial phenotype in the ↑Coll patients may be an important determinant of in vivo angiogenesis. The finding of more mitogenic CS serum in ↓Coll patients could be explained by the fact that the myocardium in these patients is ischemic for more prolonged periods at the subendocardial level, illustrated by the longer ischemia recovery times after exercise. Additionally, there may be an intrinsic resistance in the responsiveness of the vascular elements involved in collateral formation to mitogenic stimuli, such that a persistent promitogenic stimulus is required to initiate collateral development. This concept is supported by Waltenberger et al,21 who showed that the migration response of monocytes is downregulated in diabetic patients, and observations by Metais et al22 of an impaired NO-dependent response to VEGF in atrial microvessels of CAD patients.
Although plasma levels of specific angiogens reflect the general cytokine milieu across the myocardial bed, they cannot necessarily be extrapolated to represent the myocardial interstitium given the effects that extracellular matrix proteins have on growth factor presentation and receptor binding. Indeed, there are an increasing number of reports that describe a complex system of collaboration between growth factors and extracellular matrix components in modulating cellular responses. For example, fibronectin, through its VEGF binding domains, enhances the migration and proliferation of endothelial cells in response to VEGF by promoting a specific physical association between VEGF receptor 2 and integrin α5β1, leading to a sustained activation of mitogen-activated protein kinase.23
Statin therapy restores EPC numbers24,25 and enhances angiogenesis in the ischemic hind-limb model.26 Thus, increasing the EPC population seems to play an important role in ischemic protection. The 4-fold increase in the circulating EPC population in patients with ↑Coll versus ↓Coll supports this contention. Our clinical findings directly support the suggestion that targeted therapeutic strategies to enhance circulating EPC numbers and functional activity either alone or in combination with delivery of proangiogenic cytokines may maximize functional collateral formation.
The Data Supplement is available online only at http://www.circulationaha.org.
Dr Meng is an employee of Genentech, Inc.
Levin DC. Pathways and functional significance of the coronary collateral circulation. Circulation. 1974; 50: 831–837.
Koerselman J, van der Graaf Y, de Jaegere PPTh, et al. Coronary collaterals: an important and underexposed aspect of coronary artery disease. Circulation. 2003; 107: 2507–2511.
Wustmann K, Zbinden S, Windecker S, et al. Is there functional collateral flow during vascular occlusion in angiographically normal coronary arteries? Circulation. 2003; 107: 2213–2220.
Asahara T, Murohara T, Sullivan A, et al. Isolation of putative progenitor endothelial cells for angiogenesis. Science. 1997; 275: 964–967.
Rivard A, Fabre JE, Silver M, et al. Age-dependent impairment of angiogenesis. Circulation. 1999; 99: 111–120.
Asahara T, Masuda H, Takahashi T, et al. Bone marrow origin of endothelial progenitor cells responsible for postnatal vasculogenesis in physiological and pathological neovascularization. Circ Res. 1999; 85: 221–228.
Peichev M, Naiyer AJ, Pereira D, et al. Expression of VEGFR-2 and AC133 by circulating human CD34+ cells identifies a population of functional endothelial precursors. Blood. 2000; 95: 952–958.
Gehling UM, Ergun S, Schumacher U, et al. In vitro differentiation of endothelial cells from AC133-positive progenitor cells. Blood. 2000; 95: 3106–3112.
Shintani S, Murohara T, Ikeda H, et al. Augmentation of postnatal neovascularization with autologous bone marrow transplantation. Circulation. 2001; 103: 897–903.
Assmus B, Schachinger V, Teupe C, et al. Transplantation of Progenitor Cells and Regeneration Enhancement in Acute Myocardial Infarction (TOPCARE-AMI). Circulation. 2002; 106: 3009–3017.
Pijls NH, van SJ, Kirkeeide RL, De BB, et al. Experimental basis of determining maximum coronary, myocardial, and collateral blood flow by pressure measurements for assessing functional stenosis severity before and after percutaneous transluminal coronary angioplasty. Circulation. 1993; 87: 1354–1367.
Seiler C, Pohl T, Wustmann K, et al. Promotion of collateral growth by granulocyte-macrophage colony-stimulating factor in patients with coronary artery disease: a randomized, double-blind, placebo-controlled study. Circulation. 2001; 104: 2012–2017.
Waltenberger J, Lange J, Kranz A. Vascular endothelial growth factor-A-induced chemotaxis of monocytes is attenuated in patients with diabetes mellitus: a potential predictor for the individual capacity to develop collaterals. Circulation. 2000; 102: 185–190.
Wijelath ES, Murray J, Rahman S, et al. Novel vascular endothelial growth factor binding domains of fibronectin enhance vascular endothelial growth factor biological activity. Circ Res. 2002; 91: 25–31.
Vasa M, Fichtlscherer S, Adler K, et al. Increase in circulating endothelial progenitor cells by statin therapy in patients with stable coronary artery disease. Circulation. 2001; 103: 2885–2890.