Author + information
- Received September 2, 2008
- Revision received January 6, 2009
- Accepted February 16, 2009
- Published online June 9, 2009.
- Stephan H. Schirmer, MD, PhD⁎,§,
- Niels van Royen, MD, PhD⁎,⁎ (, )
- Perry D. Moerland, PhD†,
- Joost O. Fledderus, PhD‡,
- José P. Henriques, MD, PhD⁎,
- René J. van der Schaaf, MD⁎,
- Marije M. Vis, MD⁎,
- Jan Baan Jr, MD, PhD⁎,
- Karel T. Koch, MD, PhD⁎,
- Anton J.G. Horrevoets, PhD‡,
- Imo E. Hoefer, MD, PhD∥ and
- Jan J. Piek, MD, PhD⁎
- ↵⁎Reprint requests and correspondence:
Dr. Niels van Royen, Department of Cardiology, Academic Medical Center, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands
Objectives Our aim was to determine cytokine and oxygen gradients over the collateral circulation in humans.
Background The molecular background of the maturation of the collateral circulation in response to coronary narrowing is poorly understood in humans, partly because of difficulties in obtaining local samples from the human collateral circulation.
Methods Coronary collateral blood was sampled in 60 patients with nontotal (n = 25) or total coronary occlusions (n = 35) using a special wide-lumen catheter. Coronary collateral flow index (CFI) was assessed by intracoronary pressure measurements. Oxygenation and lactate content was measured as well as 30 cytokines potentially involved in collateral artery growth, using a custom-made multiplex assay.
Results No rise in lactate or change in pH was found in collateral blood. Oxygen gradient between coronary and collateral arterial blood correlated inversely with CFI (r = −0.61, p < 0.001). Locally increased plasma levels were found for basic fibroblast growth factor, eotaxin, macrophage migration inflammatory factor, monocyte chemoattractant protein-1, and transforming growth factor-beta, while stem cell factor and stem cell growth factor-beta were significantly decreased. The highest cytokine gradients were found in patients with the least developed collateral circulation. The majority of cytokines correlated more strongly with the pO2gradient across the collateral bed than with CFI.
Conclusions Intravascular ischemia during brief balloon coronary occlusion is absent in human coronary collateral arteries. The oxygen gradient found over the collateral circulation is increased in patients with a less matured collateral circulation and relates to local levels of several known pro-arteriogenic cytokines. In case of a more developed collateral circulation, relatively low levels of cytokines are present, suggesting that growth factor therapy might be beneficial at this stage.
Although mechanisms of arteriogenesis have been well elucidated in experimental models, knowledge of cytokines and growth factors that mediate collateral artery growth in humans is still limited (1). Also, the role of ischemia in human collateral artery growth is presently unknown. Local plasma sampling directly from the human coronary collateral circulation can provide such data. This is cumbersome, however, because of the poor accessibility of the collateral circulation for blood sampling. Therefore, previous studies have investigated systemic cytokine levels in patients with different degrees of coronary collateralization. These studies have provided conflicting results, probably due to the fact that systemic levels of growth factors do not reliably reflect the local process of collateral artery growth (2,3). Direct sampling from the collateral circulation, as performed in previous studies, yielded only a small volume of blood in a limited percentage of patients. Such small blood volumes have nevertheless allowed the analysis of a few cytokines showing increased levels of basic fibroblast growth factor (bFGF) and transforming growth factor (TGF)-beta in the coronary collateral circulation (4,5).
In the present study, to allow analysis of a wide array of cytokines potentially involved in human arteriogenesis, we have used the Proxis catheter (St. Jude Medical, St. Paul, Minnesota), a wide-lumen suction catheter originally developed as a proximal embolic protection device (6), to obtain larger amounts of blood from the collateral circulation. In addition to plasma levels of cytokines, the use of this catheter facilitated additional blood gas and lactate analyses of locally sampled collateral blood, which have not been performed previously in human coronary collateral arteries. Together with the analysis of plasma cytokines, these metabolic parameters can provide valuable insights into the growth and functional status of the collateral circulation in response to epicardial obstructions.
This study was approved by the medical ethics committee of the Academic Medical Center, Amsterdam, the Netherlands. Sixty Caucasian patients scheduled for elective percutaneous coronary intervention for stable coronary artery disease (CAD) with symptoms of angina pectoris for ≥4 weeks were included after giving informed consent. Patients with a subtotal stenosis (≥70%) were selected if they had single-vessel CAD; patients with chronic total occlusions (CTOs) had single-vessel or multivessel disease. Exclusion criteria were previous myocardial infarction in the area of collateralization, previous cardiac surgery, severely depressed left ventricular function, diabetes mellitus, neoplastic disease, and signs of inflammatory illness.
After pre-dilation of the stenosis, a 7-F (inner diameter 0.051 inches) proximal embolic protection device (Proxis catheter, St. Jude Medical) was advanced through the guiding catheter and into the treated coronary artery, which received collateral blood. Inflating a low-pressure (0.67 atm) balloon at the end of the catheter temporarily stopped antegrade blood flow in the artery. Complete obstruction of antegrade blood flow was ensured in a pilot study of 5 patients showing no contrast dye entering the epicardial artery after balloon occlusion (data not shown). Subsequently, gentle suction was applied manually to the proximal end of the catheter. The first 3 ml aspirated was discarded to prevent contamination with contrast medium or with blood already in the epicardial vessel before balloon occlusion (and thus not from the collateral circulation). Approximately 5 to 15 ml blood (depending on collateral flow) was then aspirated within 60 s. After deflation of the balloon, the catheter was withdrawn into the guiding catheter to obtain a control sample from the coronary circulation. Blood was transferred into citrate tubes and centrifuged at 1,550 gfor 30 min. Plasma was taken off carefully, aliquoted, snap frozen in liquid nitrogen, and stored at −80°C. Collateral flow index (CFI) was measured as previously published (7).
Oxygen, pH, and lactate measurements
Immediately after aspiration, partial oxygen (pO2) and carbon dioxide pressure (pCO2), pH, and saturation were measured from 1 sample per site per patient on a Rapidlab 865 (Siemens, Germany). Lactate concentrations were measured using a standard clinical chemistry test.
Multiplex assay for the measurement of cytokine concentrations
A custom-made panel of 30 cytokines was measured using the Bio-Plex suspension array system (Bio-Rad Laboratories, Hercules, California) according to the manufacturer's instructions as previously described (8). The panel consisted of interferon-alpha-2, hepatocyte growth factor, macrophage colony-stimulating factor (MCSF), monocyte inflammatory factor (MIF), monokine induced by gamma-interferon, nerve growth factor-beta, stem cell factor, stem-cell growth factor-b, stromal cell-derived factor-1a, tumor necrosis factor (TNF)-alpha, TNF-beta, TNF-related apoptosis inducing ligand, interleukin (IL)-1b, IL-4, IL-6, IL-8, IL-10, IL-16, eotaxin, bFGF, granulocyte-macrophage colony-stimulating factor (GM-CSF), interferon-gamma, interferon-gamma-inducible protein-10, monocyte chemoattractant protein-1 (MCP-1), macrophage inflammatory protein 1-alpha (MIP-1α), MIP-1β, platelet-derived growth factor, beta chain, regulated on activation, normal T expressed and secreted (RANTES), and vascular endothelial growth factor. Plasma samples were thawed at room temperature, diluted 1:1 in the manufacturer's sample diluent, and incubated with bead-bound antibodies to previously mentioned cytokines for 60 min on a 96-well filter plate. After repeated washing, biotinylated detection antibody was added, and the plate was incubated for another 60 min. Finally, streptavidin-PE (phycoerythrin) was pipetted onto each well. After a short incubation, the constituents of each well were drawn up into the flow-based Bio-Plex suspension array system, which identifies and quantifies each cytokine concentration based on bead color and fluorescence. Data were analyzed using Bio-Plex Manager software, version 3.0 (Bio-Rad Laboratories).
Enzyme-linked immunosorbent assays (ELISAs)
Because human interferon-beta and TGF-beta were not represented on our multiplex assay, we tested concentrations of these cytokines using commercially available ELISA kits (Quantikine ELISA, R&D Systems, Minneapolis, Minnesota, and PBL Interferonsource, Piscataway, New Jersey) according to the manufacturer's instructions. Since the concentrations of bFGF were often found to be too low to be properly detected with the multiplex system, a high sensitivity ELISA assay was performed (HS Quantikine ELISA, R&D Systems) according to the manufacturer's instructions. Briefly, samples were incubated on a 96-well plate coated with the appropriate antibody. After the necessary washing steps, a biotinylated antibody and subsequently streptavidin-conjugated horseradish-peroxidase were linked to the adherent probes. Using tetramethyl-benzidine as a substrate, absorption was measured at 450 nm in an EL808 spectrophotometer (BioTek, Winooski, Vermont).
Clinical characteristics are presented as mean ± SD or median and interquartile range for quantitative variables and as observed numbers (%) for nominal variables. Fisher exact test was used for testing association in 2 × 2 contingency tables. From pO2, oxygen gradient was calculated as: pO2 gradient= pO2 coronary− pO2 collateral. pCO2 gradient was calculated as: pCO2 gradient= pCO2 coronary− pCO2 collateral. Cytokine gradients were calculated as: concentrationcollateral− concentrationcoronary. Oxygen and cytokine gradients were calculated on a logarithmic scale to correct for skewed distributions. Oxygen and cytokine concentrations are shown as median (interquartile range). Comparisons among 2 groups were performed by Student ttest for normally distributed parameters and Wilcoxon test for non-normally distributed parameters. Statistics on cytokine concentrations were corrected for multiple testing using Benjamini-Hochberg correction (9). False discovery rates were calculated and expressed as percentage. Comparisons between 3 groups were performed by analysis of variance testing. Correlations were calculated using a Pearson correlation.
Sixty patients were included in the study. Patients were age 61.0 ± 11.4 years, and CFI ranged from 0.09 to 0.65 (mean value 0.31 ± 0.12). Thirty-five patients had a CTO of a coronary artery, while 25 had a subtotal stenosis (ranging from 60.1% to 91.5% diameter stenosis in quantitative coronary angiography). All lesions undergoing intervention were proximal lesions, resulting in arterial diameters of 3.0 to 4.0 mm. All CTO patients had sufficient collateralization to prevent ischemia during balloon occlusion as indicated by absence of ST-segment deviation, and no symptoms of angina pectoris at rest. Patients with a nontotal occlusion were grouped into collateral responders (CFI >0.21, n = 13) and nonresponders (CFI ≤0.21, n = 12). ST-segment elevation during balloon occlusion (shown as median [interquartile range]) was significantly (p < 0.001) different between collateral responders (0.0 [1.0] mm) and nonresponders (2.0 [2.0] mm). Clinical baseline characteristics did not correlate with CFI and were not different between the 3 groups (nonresponders, responders, CTOs) (Table 1).
Metabolic parameters in the coronary collateral circulation
Oxygen pressure measurements were performed in 45 patients in total in this study. In the complete patient group, average pO2was found significantly lower in the collateral sample than in the coronary sample (10.2 [2.75] kPa vs. 11.3 [2.4] kPa; p < 0.001). In contrast, lactate concentrations were equal in coronary and collateral blood (0.94 mmol/l vs. 0.91 mmol/l, p = NS), and also pH did not differ significantly (coronary: 7.42, collateral: 7.41, p = NS). Lactate and pH remained unchanged in collateral versus coronary blood in collateral nonresponders, responders, and patients with CTOs. There was no significant correlation between CFI and lactate (r = 0.19, p = 0.29) or CFI and pH (r = 0.12, p = 0.44). Also, pCO2was not found different between proximal and distal sampling sites (5.18 vs. 5.17 kP, p = NS).
Comparing the 3 patient groups, nonresponders showed a significantly higher oxygen gradient than responders or CTOs (2.24 [2.0] kPa vs. 1.63 [1.0] and 0.95 [0.77] kPa, p = 0.003). In the patient group as a whole, a significant inverse correlation was found between pO2gradient and CFI (r = −0.61, p < 0.001) (Fig. 1).Similarly, an inverse correlation was found between oxygen saturation ratio and CFI (r = −0.46, p = 0.002). Conversely, the pCO2gradient did not correlate with CFI or oxygen gradient (r = 0.17, p = NS) (Fig. 2).
Local cytokine concentrations in the collateral circulation
In the complete patient population, eotaxin, MCP-1, MIF, bFGF, and TGF-beta were found in significantly higher plasma concentrations in the collateral specific samples as compared with the coronary samples, while stem cell factor and stem cell growth factor-beta were significantly down-regulated (Table 2).Other cytokines measured did not show significant differences between collateral and coronary samples. Interferon-beta was not detectable in either the coronary or the systemic sample in the majority of patients (35 of 60 patients).
Dividing patients into the 3 groups (nonresponders, responders, CTOs) resulted in the highest collateral-coronary gradients for all of the previously mentioned cytokines in the nonresponders, followed by the responders, while patients with CTOs displayed smallest concentration differences (Fig. 3).
Group-specific cytokine profiles in relation to CFI and oxygen gradient
In the patient group with nontotal occlusions, a positive correlation with oxygen gradient between coronary and collateral arterial blood was found for coronary-collateral gradients of MCP-1, eotaxin, MIF, MIP-1β, and hepatocyte growth factor. Conversely, a negative correlation with oxygen gradient was found for gradients of nerve growth factor-beta and TNF-related apoptosis-inducing ligand (TRIAL) (Table 3).No cytokine gradient correlated significantly with CFI (data not shown) in this group. After correction for multiple testing, none of the coronary-collateral cytokine gradients correlated with oxygen gradient (false discovery rate >25%) or CFI in patients with CTO.
In this study on the molecular background of the maturation of the coronary collateral circulation in man, no changes in levels of pH or lactate were found in the collateral circulation, showing that no intravascular ischemia is present in the human collateral circulation. While pCO2and pH were unchanged in collateral versus systemic blood, pO2was significantly lower in collateral as compared with coronary blood samples, suggesting local oxygen extraction in the collateral circulation. Oxygen gradient between coronary and collateral arterial blood correlated with CFI. Eotaxin, bFGF, MCP-1, MIF, and TGF-beta were found selectively increased in collateral blood plasma, and cytokine levels in the collateral circulation correlated well with oxygen gradient. Collateral cytokine gradients were strongest in patients with insufficient collateral arteries and almost absent in patients with CTOs.
For a comprehensive analysis of human arteriogenesis, we selected patients with single-vessel CAD and a severe coronary narrowing (nonresponders and responders) as well as patients with CTOs. This approach allows an analysis of the collateral vascular development in 2 distinct clinical phases (i.e., during an accelerated phase in 1-vessel disease and a more quiescent phase in chronic coronary occlusion). Moreover, the analysis of cytokine concentrations was performed in a uniform fashion allowing, for the first time, an interpretation of the adaptation of collateral vessels in a wide range of coronary narrowing. Patients with previous myocardial infarction, as shown by echocardiography or pathological Q waves on electrocardiogram, were excluded from the study.
Collateral-specific blood sampling
The Proxis embolic protection device has been shown earlier to safely occlude coronary vessels and provide retrograde blood aspiration (10). For the first time, we describe the use of the catheter for the selective collection of coronary collateral blood. Previous studies using an over-the-wire catheter or a multifunctional probing catheter for coronary-collateral-specific blood sampling resulted in limited amounts of blood and possible activation of the sample because of the small lumen of the catheter (4,5). Using pericardial fluid or coronary sinus blood results in sufficient sample volumes that are, however, not selective for the collateral circulation (11,12).
In the clinical setting, catheter-assisted aspiration of collateral blood cannot exclude potential admixture of capillary/venous blood, especially in patients with low CFI. The fact that in the present study the fall in pO2is not accompanied by a rise in pCO2or a decrease in pH in the retrogradely drawn samples, however, makes significant admixture of venous blood less likely although contamination of a small amount of distal coronary capillary blood cannot be fully excluded.
We postulate that the gradient in arterial oxygen pressure between coronary and collateral arterial blood found in this study rather relates to actual oxygen extraction along the collateral anastomoses. This corroborates experimental studies also reporting significant oxygen extraction along arteriolar systems (13,14). Local production of cytokines by the activated endothelium of the collateral circulation is an energy-consuming process. Indeed, in patients with a high oxygen gradient we also found the highest levels of cytokines.
Cytokines in human arteriogenesis
The collateral-specific up-regulation of MCP-1, bFGF, and TGF-beta confirms earlier data on the involvement of these cytokines in arteriogenesis (4,5). Other previously described pro-arteriogenic cytokines, such as GM-CSF, granulocyte colony-stimulating factor, and TNF-alpha, were not found differential between the 2 sampling sites. Next to CCL2 (MCP-1), we here also first describe enhanced expression of another CC-chemokine, CCL11 (eotaxin) in collateral arteries. Its up-regulation in collateral arterial blood might reflect a role in the fine-tuning of mononuclear cell homing, partly counteracting the effects of MCP-1 (15). MIF has earlier been described to stimulate angiogenesis, but a link to collateral artery growth has not been reported previously (16). MIF acts as a major positive regulator of inflammatory cell recruitment (17) and macrophage response to lipopolysaccharide (18), both of which play an important role in arteriogenesis.
Cytokine expression in different phases of collateral artery growth
Since no intravascular ischemia could be detected in the collateral arteries, collateral-specific expression of cytokines is not regulated by intravascular hypoxemia or tissue ischemia directly adjacent to the collateral circulation. However, ischemia in tissue distant from the major collateral arterioles still occurs, particularly in patients with a nonsufficient collateral circulation, as shown by ST-segment elevation during balloon occlusion. Our findings are in agreement with earlier results from experimental studies showing that arteriogenesis takes place proximal to the site of tissue ischemia (19). Low oxygen gradient between coronary blood and blood in well-developed collateral arteries demonstrated the physiological significance of these bypass vessels, which are capable of supplying myocardial tissue with normal oxygen concentrations.
Interestingly, cytokine concentrations in the collateral circulation could be correlated to the oxygen gradient between coronary and collateral blood. We show that the degree of collateral maturation is reflected by oxygen pressure in the collateral artery. Cytokine gradients between collateral and coronary blood were greatest in nonresponders and almost absent in CTO patients. The correlation of cytokine concentrations with the oxygen extraction shows strong arteriogenic activity where collateral vessels are not yet capable of providing sufficient oxygen levels. Those cytokines that most significantly correlated with oxygen gradients were the same that were most strongly up-regulated in collateral compared with coronary plasma. Interestingly, a number of cytokine gradients were found to inversely correlate with the oxygen gradient, suggesting a potential antiarteriogenic role. In the chronic phase of arteriogenesis (i.e., in CTOs), a reduction of cytokine gradients indicates that collateral artery growth has reached a plateau phase where less cytokines and growth factors are secreted. This is in line with the results of a study by Werner et al. (5) showing higher bFGF concentrations in more recent occlusions, where arteriogenesis was presumably still ongoing, as well as increased concentrations of MCP-1, placental growth factor, and TGF-beta in smaller collateral arteries with higher shear stress. Even if CFI does not reach high levels in all patients, collateral artery growth may have come to a halt. This is consistent with earlier findings showing a genetic predisposition of individual patients to develop an adequate or inadequate collateral circulation (7) and supports strategies in which local delivery of growth factors/cytokines in the coronary circulation is striven for. In that earlier study, increased monocyte inflammatory signaling was found in patients with weak collateralization, which is supported by plasma measurements in the present study showing increased cytokine gradients in insufficient collateral development. Potentially, inhibition of antiarteriogenic signaling selectively at the site of interest is an alternative treatment strategy for the stimulation of collateral artery growth.
Assessment of collateral flow index
Central venous pressure (CVP) was assumed to be 5 mm Hg instead of measuring CVP in every patient individually because of the already extensive study protocol. This assumption, however, reduces accuracy of the CFI measurements especially in patients with low CFI (20,21). To evaluate if differences in CVP would significantly influence our study results, we recalculated correlations of CFI with pO2and pCO2gradients as well as cytokine gradients using a CVP of 0, 5, and 10 mm Hg. Correlation of CFI with pO2gradient was almost unchanged (R = 0.60, p < 0.001). Similarly, changing the CFI cutoff to 0.25 did not significantly influence the results, still showing significantly different oxygen gradients (1.4 [1.8] vs. 2.1 [1.6]) in collateral responders versus nonresponders. Correlations with cytokine gradients and increased gradients in nonresponders compared with responders were also unchanged with CVP varying from 0 to 10.
A limitation to pressure-derived CFI measurements is reduced accuracy in low collateral flow, when back pressure increases due to increased wall stress, which may influence CFI measurements (22). No data on left ventricular end-diastolic pressure are available from the present study population as it requires additional intracavitary instrumentation. However, patients with depressed left ventricular function on echocardiography were excluded from the study, which limits the possibility of variation in back pressure as a confounding factor in the present study.
Finally, applying negative pressure during intracoronary blood withdrawal might have affected collateral flow rate, which potentially has led to an actual underestimation of the pO2and cytokine gradients in patients with a low CFI.
CFI can directly be related to the gradient of oxygen pressure over the collateral vascular bed. In patients with a nonsufficient collateral circulation, the increased oxygen gradient is accompanied by higher local levels of cytokines, pointing to an active phase of arteriogenic remodeling. Less pronounced cytokine gradients in collateral arteries in CTOs indicate a quiescent phase of arteriogenic activity, which supports strategies of local growth factor therapy in these patients.
The authors thank the nursing staff of the cardiac catheterization laboratory for their excellent assistance.
This study was supported by a VENI-grant by the Netherlands Organisation for Scientific Research (NWO) to Dr. van Royen (No. 916-66-019) and a grant by the Deutsche Forschungsgemeinschaft (DFG) to Dr. Schirmer (SCHI 1012/1-1). Drs. Schirmer and van Royen contributed equally to this work.
- Abbreviations and Acronyms
- basic fibroblast growth factor
- coronary artery disease
- collateral flow index
- chronic total occlusion
- central venous pressure
- enzyme-linked immunosorbent assay
- granulocyte-macrophage colony-stimulating factor
- hepatocyte growth factor
- monocyte chemoattractant protein
- macrophage migration inflammatory factor
- macrophage inflammatory protein
- transforming growth factor
- tumor necrosis factor
- Received September 2, 2008.
- Revision received January 6, 2009.
- Accepted February 16, 2009.
- American College of Cardiology Foundation
- Schirmer S.H.,
- van Nooijen F.C.,
- Piek J.J.,
- van Royen N.
- Fleisch M.,
- Billinger M.,
- Eberli F.R.,
- Garachemani A.R.,
- Meier B.,
- Seiler C.
- Werner G.S.,
- Jandt E.,
- Krack A.,
- et al.
- Schirmer S.H.,
- Fledderus J.O.,
- Bot P.T.G.,
- et al.
- Kraaijeveld A.O.,
- de Jager S.C.A.,
- de Jager W.J.,
- et al.
- Benjamini Y.,
- Hochberg Y.
- Lambiase P.D.,
- Edwards R.J.,
- Anthopoulos P.,
- et al.
- Panchal V.R.,
- Rehman J.,
- Nguyen A.T.,
- et al.
- Cabrales P.,
- Tsai A.G.,
- Johnson P.C.,
- Intaglietta M.
- Duling B.R.,
- Berne R.M.
- Ogilvie P.,
- Bardi G.,
- Clark-Lewis I.,
- Baggiolini M.,
- Uguccioni M.
- Amin M.A.,
- Volpert O.V.,
- Woods J.M.,
- Kumar P.,
- Harlow L.A.,
- Koch A.E.
- Ito W.D.,
- Arras M.,
- Scholz D.,
- Winkler B.,
- Htun P.,
- Schaper W.
- Perera D.,
- Biggart S.,
- Postema P.,
- et al.
- Seiler C.,
- Fleisch M.,
- Garachemani A.,
- Meier B.
- Spaan J.A.E.,
- Piek J.J.,
- Hoffman J.I.E.,
- Siebes M.