Vascular endothelium is covered with an extensive mesh of glycocalyx constituents, which acts like an effective barrier up to several micrometers thick that shields the luminal surface of the vasculature from direct exposure to flowing blood. Many studies report that various enzymatic and pharmaceutical challenges are able to increase glycocalyx porosity, resulting in farther permeation of plasma macromolecules and greater access of red blood cells into glycocalyx domain. Attenuation of glycocalyx barrier properties therefore potentially increases the amount of blood that effectively occupies available microvascular volume. We tested in the present study whether attenuation of coronary glycocalyx barrier properties actually increases coronary blood volume and whether such changes would be noticeable during measurements of coronary flow reserve using adenosine. In anesthetized goats (n = 6) with cannulated left main coronary artery that were perfused under controlled pressure, coronary blood volume was measured via the indicator-dilution technique using high-molecular-weight (2,000 kDa) dextrans as plasma tracer and labeled red blood cells as red blood cell tracer. Coronary blood volume was determined at baseline and during intracoronary infusion of adenosine causing maximal vasodilation (0.2–0.6 mg·kg−1·h−1) before and after intracoronary hyaluronidase treatment (170,000 units) of the glycocalyx. With an intact glycocalyx, coronary blood volume was 18.9 ± 1.1 ml/100 g heart tissue at baseline, which increased to 26.3 ± 2.7 ml/100 g after hyaluronidase treatment of the coronary glycocalyx. Maximal vasodilation by administration of adenosine further increased coronary blood volume to 33.9 ± 6.8 ml/100 g, a value not different from the maximal coronary blood volume of 33.2 ± 5.3 ml/100 g obtained by administration of adenosine in the absence of hyaluronidase treatment. Adenosine-induced increases in coronary conductance were not affected by hyaluronidase treatment. We conclude that acute attenuation of glycocalyx barrier properties increases coronary blood volume by ∼40%, which is of similar magnitude as additional changes in coronary blood volume during subsequent maximal vasodilation with adenosine. Furthermore, maximal coronary blood volume following administration of adenosine was similar with and without prior hyaluronidase degradation of the glycocalyx, suggesting that adenosine and hyaluronidase potentially increase glycocalyx porosity to a similar extent. Hyaluronidase-mediated changes in coronary blood volume did not affect baseline and adenosine-induced increases in coronary conductance, demonstrating that measurements of coronary flow reserve are insufficient to detect impairment of coronary blood volume recruitment in conditions of damaged glycocalyx.
- blood flow
- coronary circulation
our previous study demonstrated the presence of a thick glycocalyx in the coronary circulation (35). The glycocalyx thickness has been estimated to be 0.2–0.9 μm in capillaries (11, 27, 31–33, 38, 42, 43). Because of the limited access of flowing blood to an intact glycocalyx, microvascular blood volume might therefore in effect be reduced by as much as 50% of available anatomical microvascular volume due to glycocalyx presence. Others have demonstrated that enzymatic treatment of the glycocalyx with, for example, heparinase or hyaluronidase (9, 11, 35) and also exposure of the glycocalyx to the vasodilators adenosine, bradykinin and sodium nitroprusside (21, 31, 38, 41) increases glycocalyx porosity for plasma macromolecules and red blood cells. However, measurements of the effect of attenuated glycocalyx barrier properties on potential increases in coronary blood volume have never been performed.
Since it was established for hyaluronidase and adenosine that the porosity of the glycocalyx increases, we hypothesize that attenuation of glycocalyx barrier properties by hyaluronidase treatment and adenosine administration increases coronary blood volume. Furthermore, in the view of the fact that adenosine is the preferred vasodilator in the clinic for evaluation of coronary flow reserve, we tested whether changes in glycocalyx porosity would be noticeable during measurement of coronary flow reserve using adenosine. In anesthetized goats, coronary blood volume was measured under control conditions and after hyaluronidase treatment at baseline and during the administration of adenosine. Coronary blood volumes were determined by applying the tracer-dilution technique, using dextrans with a molecular weight of 2,000 kDa (dextran-2000) as plasma tracer and labeled red blood cells as red blood cell tracer.
General Surgery and Anesthesia
All procedures and protocols were approved by the animal care and use committee of Maastricht University (DEC no. 2007-061). Studies were conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Experiments were performed on female adult goats of 18–27 kg (n = 6). At the beginning of an experiment, the goats were anesthetized with an intramuscular injection of Nimatek (15 mg/kg; Eurovet Animal Health) and Dormicum (0.75 mg/kg; Roche). Next, goats were intubated and ventilated with a 1:2 O2-air mixture, and anesthesia was maintained by intravenous administration of Sufentanil-hameln (9.375 μg·kg−1·h−1; Hameln Pharmaceuticals), Dormicum (0.625 mg·kg−1·h−1) and propofol (10 mg·kg−1·h−1; Fresenius Kabi Nederland). Depth of anesthesia was adjusted according to the stability of femoral artery blood pressure (Pfem) and heart rate (HR). Arterial and coronary venous blood gasses and pH were measured every 30 min. When necessary, ventilation was adjusted to maintain oxygen and CO2 pressures within physiological limits, and sodium bicarbonate was administered to avoid acidosis.
A left thoracotomy was performed in the fourth intercostal space, and one of the ribs was removed. The great cardiac vein was cannulated via the azygos vein to obtain coronary venous blood samples. Next, a double purse string was made in the outer wall of the aorta. Subsequently, the heart was exposed and suspended in a pericardial cradle. The left main coronary artery was dissected free, and a ligature was placed around the vessel. A catheter (7-F; Sentron) connected to a Sentron interface (type no. 811-000) was inserted through the left auricle into the left ventricle to measure pressure (Plv). After administration of a 3-ml heparin bolus (5,000 IE/ml; LEO Pharma), anticoagulation was maintained by continuous infusion of heparin (1,000 IE/h). Next, the left carotid artery was cannulated and a stainless steel Gregg cannula was inserted in the aorta via the purse string. Blood collected from the carotid artery was circulated through a perfusion system and back via the Gregg cannula into the aorta for 10 min. Hereafter, the Gregg cannula was ligated into the left main coronary artery, without disrupting the coronary flow. Coronary perfusion was controlled using a roller pump perfusion system (34). Blood collected from the left carotid artery was heated and filtered and was circulated with a roller pump via a reservoir (70 ml) into the left main coronary artery. Blood level and pressure in the reservoir were kept constant. Perfusion pressure (Pperf) was measured at the tip of the Gregg cannula. An in-line flow probe (6-mm; Transonic Systems) was interpositioned into the perfusion system to measure coronary blood flow (Qcor). Qcor, depending on coronary resistance, was pulsatile as a result of cardiac contraction.
Pfem, Pperf, Plv,sys (left ventricular systolic blood pressure determined from maximal Plv per beat), Qcor, and HR (determined from Plv) were stored for off-line analysis (IDEEQ 250 Hz; IDEE). At the end of the experimental procedure, a battery was placed on the heart to induce ventricular fibrillation. The heart was taken out, and total heart weight as well as the weight of the perfused area was measured. Perfusion area was determined by injection of white paint into the left main coronary artery and dissection and weighing of the demarcated area.
After surgery, the preparation was allowed to equilibrate for 30 min. In each animal, distribution volumes of two tracers, the plasma tracer FITC-labeled dextran-2000 (1.3 mg/ml; Sigma-Aldrich) and 5-(and 6)-carboxyfluorescein diacetate, succinimidyl ester [5(6)-CFDA,SE; Invitrogen] labeled red blood cells were measured using the indicator-dilution technique. The day before the experiment, 10 ml of blood were taken from the goat and centrifuged at 1,200 g. Plasma was removed, and 30 ml of 5(6)-CFDA,SE tracer (33.3 ng/ml) were added to the red blood cells. The solution was kept in the dark for 5 min before it was centrifuged and the supernatant removed. Red blood cells were washed three times with PBS, after which 1 ml of plasma was added to the red blood cells. Labeled red blood cells were kept at 4°C in the dark overnight. At the experimental day, 0.75 ml of dextran-2000 was mixed with 0.75 ml of labeled red blood cells for each volume measurement. Before the first injection of tracers, a single bolus of 10 ml of dextrans with a molecular mass of 1 kDa (Promiten; NPBI) was injected intracoronary to attenuate the risk for anaphylactic reactions. Within 15 min (22), a bolus injection of 1 ml of the tracer mix was given by hand into the left main coronary artery via the perfusion pressure catheter at the tip of the Gregg cannula. Blood samples were collected from the great cardiac vein at a rate of 24.2 ml/min using a roller pump (205S Watson Marlow). Blood was sampled in consecutive 1.5-ml tubes at intervals of 0.5–1 s for 50 s. A total of 70 samples per coronary volume measurement was obtained.
Before each volume measurement, an arterial and coronary venous blood sample was taken for determination of blood gasses. Fractional myocardial oxygen extraction was defined as the arteriovenous oxygen difference as a percentage of the arterial oxygen content, with the oxygen content of the blood sample being computed as 1.34 × plasma hemoglobin × oxygen saturation × 0.01 + 0.0225 × oxygen pressure. Myocardial oxygen consumption was defined as coronary blood flow × arteriovenous oxygen difference. In each animal, a similar order of coronary blood volume measurements was done; first, a baseline measurement was taken at a Pperf of 127.1 ± 5.7 mmHg. Next, adenosine (0.2–0.6 mg·kg−1·h−1) was infused via the perfusion system into the left main coronary artery at a concentration inducing maximal dilation (as verified by lack of reactive hyperemia after the perfusion line had been clamped for 15 s) without significant changes in HR. During adenosine infusion, Pperf was reduced to 95.2 ± 3.6 mmHg. Pperf was corrected during adenosine infusion to avoid a mismatch between inflow (carotid artery) and outflow (coronary vasculature) of the perfusion system. The tracer mixture was injected at least 10 min after the start of adenosine infusion to allow for a new hemodynamic steady state. Infusion of adenosine was stopped after the last coronary venous sample was collected. When Qcor had returned to baseline (5–10 min), Pperf was increased to its initial value (118.6 ± 8.2 mmHg) and infusion of 100 ml hyaluronidase was started (1,700 U/ml, type IV-S, intracoronary; Sigma-Aldrich). After 50 min of hyaluronidase infusion, a second baseline measurement was performed. The fourth measurement was performed during infusion of adenosine at Pperf of 92.3 ± 6.1 mmHg. For reference, three additional experiments were performed in which measurements were done using saline instead of hyaluronidase to determine reproducibility of the baseline coronary blood volume measurement after adenosine washout. Collected blood was analyzed using a FACScalibur flow cytometer (BD Biosciences) for labeled red blood cell dilution and subsequently centrifuged. Plasma dextran-2000 concentration was analyzed using fluorometry (Victor3; Perkin Elmer). The remaining amount of tracer mix after injection (0.5 ml) was used for making calibration curves.
From n = 6 experiments, the hyaluronan levels in the plasma determined for baseline, during adenosine, and after hyaluronidase (n = 3) or saline (n = 3) were given using a hyaluronan ELISA (Echelon). Removal of the glycocalyx by hyaluronidase was confirmed by Alcian blue 8GX staining of the glycocalyx in two additional goat hearts (1 control and 1 hyaluronidase). In essence, staining was performed as described for the isolated rat heart (35), except for the left main coronary perfusion (using the perfusion system). Cardiac tissue analysis was performed in an identical manner as described for the rats.
Determination of tracer mean transit time.
Per coronary blood volume measurement, 70 coronary venous plasma dextran concentrations and labeled red blood cell fractions in time were obtained. Dextran-2000 concentration and labeled red blood cell fraction curves vs. time were fitted with a probability density function, the local density random walk (LDRW) (2) (see Eq. 1). To overcome contamination of the output signal by recirculating tracer, information from the curves was used up to 15% of the peak concentration on the descending slope of the curve. The selection of data up to such a low concentration of the peak concentration was allowed because of the use of the perfusion system (volume ∼110 ml), delaying the moment at which recirculation of tracers starts to interfere with the measured data. Fitting of the probability density function was performed with a nonlinear least-squares fitting procedure. (1) where m is the mass of the injected tracer, μ is the mean residence time of the tracer, Φ is the flow of the carrier, λ is a parameter related to the skewness or asymmetry of the curve, and t0 represents the zero time of the distribution. Figure 1 shows an example of obtained concentration curves and LDRW fits of dextran-2000 vs. time (left) and labeled red blood cell fraction curves vs. time (right) for baseline (solid) and adenosine (shaded) before (top) and after hyaluronidase (bottom).
Using data from the fitted curve, we calculated the mean transit time (MTT) of each indicator according to Eq. 2, in which C(t) is the fitted dextran-2000 outflow concentration or labeled red blood cell fraction. (2) Tracer MTT was corrected for sampling catheter volume by subtraction of sampling transit time (i.e., 9.9 s).
Coronary blood volume and conductance.
Multiplying the corrected MTT with the tracer carrier flow gives the distribution volume of each tracer (24). Plasma and red blood cell flow were determined from coronary venous hematocrit, determined before each volume measurement, and Qcor. Coronary blood volume (V) was defined as the sum of calculated dextran-2000 and red blood cell volume. To correct for an influence of perfusion pressure on coronary flow, we calculated coronary conductance (C) as the ratio of Qcor to Pperf during the 50 s when samples were taken after tracer mixture injection.
The effect of time on blood pressure, HR, and coronary venous hematocrit was analyzed using a Friedman test with Dunn's multiple comparison as post hoc test. Effects of adenosine and hyaluronidase on hemodynamics, myocardial oxygen extraction, MTT, volume, and conductance were analyzed with a paired Wilcoxon signed rank test. The effect of hyaluronidase and saline on changes in myocardial oxygen extraction and myocardial oxygen consumption from baseline to adenosine was tested using a t-test. Results were considered statistically significant with P < 0.05. Summary data are means ± SE.
During the infusion of adenosine, hyaluronan levels increased from 3,407 ± 629 to 5,944 ± 1,432 ng/ml, an increase of 89 ± 39% compared with baseline. After subsequent hyaluronidase (n = 3), an additional increase in hyaluronan of 3,871 ± 2,271 ng/ml (92 ± 57%) was observed compared with the concentration of these samples during adenosine. In contrast, when saline was given, there was no additional increase (7 ± 19%) in hyaluronan levels.
Degradation of glycocalyx structures after hyaluronidase treatment is illustrated by the electron microscopic images in Fig. 2. Whereas in the control goat heart (left), abundant polysaccharide structures of the glycocalyx are present on the luminal side of the endothelium, these structures appear to be lost in the hyaluronidase-treated heart (right). These observations confirm previous observations of severe glycocalyx degradation with hyaluronidase in isolated rat hearts (35).
Baseline hemodynamic parameters are presented in Table 1. In six animals, volume measurements at baseline and during adenosine were repeated after hyaluronidase treatment, whereas in three additional animals, saline was used instead of hyaluronidase. No difference in hemodynamic parameters was found when comparing the saline group with animals treated with hyaluronidase. Femoral artery blood pressure was reduced in both groups (P < 0.05 for hyaluronidase and P = 0.07 for saline) in the second adenosine measurement, whereas Plv,sys and HR were unaffected. During administration of adenosine, increases in blood flow were matched by proportional decreases in myocardial oxygen extraction (as a percentage of that during corresponding baseline), resulting in similar amounts of oxygen delivered to the myocardial tissue (103.7 ± 16.8% of baseline). In contrast, after hyaluronidase-mediated degradation of the glycocalyx, oxygen consumption during administration of adenosine was significantly reduced to 68.3 ± 4.3% of corresponding baseline due to disproportional reduction of oxygen extraction (Table 1). In the saline group, oxygen consumption during adenosine was not different before or after the saline infusion.
MTTs of tracer outflow curves were derived from fitting of the coronary venous concentration-time curve for both tracers with a LDRW fit (see methods, Fig. 1). After hyaluronidase, there was a trend for an increase in MTT for both tracers (from 4.8 ± 0.4 to 6.1 ± 0.7 s for dextran-2000 and from 4.6 ± 0.5 to 5.9 ± 0.5 s for red blood cells, both P = 0.06). MTTs tended to decrease (P = 0.06) when adenosine was subsequently administered (4.3 ± 0.5 and 4.3 ± 0.7 s for dextran-2000 and red blood cells, respectively). In contrast, with an intact glycocalyx, the MTT of both dextrans and red blood cells did not change upon infusion of adenosine (4.5 ± 0.5 and 4.6 ± 0.6 s for dextran-2000 and red blood cells, respectively).
The sum of red blood cell and dextran-2000 distribution volumes (Fig. 3) yielded total coronary blood volume during each condition. Total coronary blood volume at baseline was 18.9 ± 1.1 ml/100 g heart tissue and increased to 26.3 ± 2.7 ml/100 g (P < 0.05) after hyaluronidase treatment. Adenosine infusion after hyaluronidase resulted in coronary blood volume of 33.9 ± 6.8 ml/100 g. Coronary blood volume during adenosine was not different before and after hyaluronidase treatment (33.2 ± 5.3 ml/100 g before hyaluronidase degradation of the glycocalyx).
To verify that the increase in coronary blood volume after hyaluronidase was not the result of incomplete recovery after adenosine, we measured baseline coronary blood volume after infusion of saline instead of hyaluronidase in three additional experiments. No difference in baseline blood volume was observed before and after saline (14.4 ± 4.1 and 14.9 ± 4.5 ml/100 g, respectively; relative change 104.1 ± 8.3%), indicating that the effect of adenosine (23.1 ± 5.9 and 21.7 ± 6.5 ml/100 g; no significant difference) on coronary blood volume was transient.
Adenosine infusion resulted in an increase in coronary conductance (Fig. 4) from a baseline of 2.1 ± 0.2 to 4.7 ± 0.4 ml·min−1·100 g−1·mmHg−1 (P < 0.05). After hyaluronidase treatment, baseline coronary conductance (2.4 ± 0.3 ml·min−1·100 g−1·mmHg−1) and maximal adenosine-induced coronary conductance (5.3 ± 0.6 ml·min−1·100 g−1·mmHg−1) were not changed compared with their corresponding measurements before hyaluronidase treatment.
In the current study in anesthetized goats, we demonstrated that intracoronary hyaluronidase infusion increased baseline coronary blood volume by 40% (see Fig. 3, bottom). Further increases in coronary blood volume of ∼35% by adenosine on top of maximal glycocalyx porosity by hyaluronidase might then represent increases in vascular volume as a result of adenosine's well-known vasodilating capacity, suggesting that increases in coronary blood volume by increased glycocalyx porosity and vasodilation are of similar magnitude. Maximal coronary blood volume following administration of adenosine was identical with and without prior glycocalyx degradation by hyaluronidase, indicating that adenosine might increase glycocalyx porosity in a hyaluronidase-like manner. In line with this possibility, we report that administration of adenosine resulted in a twofold increase in the concentration of plasma hyaluronan in the coronary effluent. Plasma hyaluronan levels during hyaluronidase still further increased compared with those during adenosine administration, indicating that the shedding of hyaluronan from the glycocalyx was less sensitive to adenosine compared with enzymatic degradation. Nevertheless, the adenosine-induced shedding of hyaluronan during adenosine appeared to result in similar loss of blood-excluding barrier properties of the glycocalyx compared with hyaluronidase treatment, suggesting that the remaining hyaluronan structures during adenosine do not constitute a serious glycocalyx barrier for the exclusion of flowing blood.
However, although available data demonstrate that enzymatic glycocalyx degradation greatly impairs the potential for adenosine to increase coronary blood volume, we report that adenosine's potential to increase coronary conductance was not affected by glycocalyx degradation, demonstrating that glycocalyx-mediated modulation of microvascular blood volume has very little impact on coronary conductance during administration of adenosine and that assessment of coronary flow reserve capacity is therefore insufficient to detect glycocalyx damage and subsequent impairment of coronary blood volume reserve.
As a consequence of the instrumentation, atrial natriuretic peptide (ANP) might have been released into the coronary system and consequently induced shedding of the endothelial glycocalyx (4). We assume that the instrumentation phase mainly might have contributed to a potential release of ANP in our experiments and that the levels were more or less constant during the successive coronary blood volume measurements. Furthermore, the presence of heparin to prevent coagulation has been associated with impaired glycocalyx barrier properties in mice (39). Both ANP release and heparin infusion might therefore have caused a reduced baseline coronary glycocalyx volume in our study. As a result, the recruited coronary glycocalyx volume during adenosine and after hyaluronidase might have been underestimated.
In the current study, we measured coronary blood volumes by applying the tracer-dilution technique in open-thorax goat hearts using a perfusion system. The perfusion system enables injection of tracers and control of perfusion pressure independently of cardiac function. Measuring distribution volumes in a beating heart in vivo makes it impossible to fulfill all requirements associated with the indicator-dilution technique. Flow and volume in a beating heart are not stationary, for example. However, flow and volume variations at cardiac frequencies have been indicated to produce little error as long as the cycle length is less than a quarter of the dye curve passage time (1, 24). In our measurements, there were ∼10 heart beats during the passage of the indicator from the left main to the great cardiac vein; therefore, we expect a minimal error due to fluctuation in flow and volume. On the other hand, the respiration frequency is much lower, and although the respiration frequency was held constant, 13–15 breaths/min, it may have introduced an error in the measurement. To overcome the problem of contamination of the outflow by recirculating tracer, we used a local density random walk fit that was based on the curve up to 15% of the peak concentration on the descending slope of the outflow concentration and labeled red blood cell fraction curves (2). In the current study we used a perfusion system with a relative large volume (∼110 ml), delaying the moment recirculation started to interfere with the first-pass outflow curve [with a minimum of 13.3 s at the highest flow measured in the current study (495.1 ml/min)], and as a result, data up to a relatively low percentage of the maximum peak value could be used. The fitting of the curves provided results with average correlation coefficients of R2 = 0.968 (standard deviation equal to 0.03) for the plasma tracer and R2 = 0.973 (standard deviation equal to 0.04) for the red blood cell tracer. To minimize and standardize the influence of injection on MTT, the tracers in this study were injected in ∼1 s, always by the same person. Since we injected the tracers at the entrance of the left main coronary artery, complete mixture of the tracers with the flowing blood may not have occurred. However, our concentration outflow curves showed little to no noise (see Fig. 1), suggesting a thorough mixing of the tracers with the flowing blood.
We used two tracers to obtain coronary blood volume measures, a red blood cell tracer and a plasma tracer. We used labeled red blood cells and flow cytometry for measuring red blood cell distribution volume (28). The red blood cells, labeled the day before the experiment, were kept at 4°C until the experiment. No differences in size were found among labeled cells, normal unlabeled cells, and cells that were kept overnight at 4°C (data not shown). Furthermore, no effect in labeling efficiency was found between labeled cells that were kept overnight at 4°C and freshly collected and labeled cells. As a result, the distribution of labeled red blood cells can be considered representative to that of the normal, unlabeled red blood cells in the body.
As intravascular plasma tracer, we used FITC-labeled dextrans with a molecular mass of 2,000 kDa. Intravital microscopic studies have demonstrated the intraluminal disposition and significant exclusion by the glycocalyx of dextrans with a molecular mass of 70 kDa and larger in microvessels of striated muscle (11, 42), indicating their usefulness as intravascular plasma tracers. In those studies, a systemic bolus infusion [140 (11) or 100 units (5)] of hyaluronidase was shown to increase glycocalyx porosity for dextrans of 70 and 145 kDa, but not for the larger size dextrans. We observed a 41 ± 16% increase in dextran-2000 distribution volume without a change in coronary conductance after hyaluronidase, suggesting a robust increase in glycocalyx porosity for these dextrans. Moreover, RBC distribution volume increased 34 ± 6%, indicating that the increase in glycocalyx porosity was associated with an increase in functionally perfused blood volume. The present effect of hyaluronidase on large dextran distribution might be the result of the prolonged exposure of the coronary vascular bed to a higher dose of hyaluronidase. Thus, in the current study, hyaluronidase was administered as a continuous intracoronary infusion during 50 min at a much higher dose (170,000 units in total). Enzymatic glycocalyx degradation using heparinase and pronase was demonstrated in isolated coronary arterioles to result in a manifold increase in apparent vascular permeability for the plasma proteins α-lactalbumin and albumin, which have a size of 14 and 65 kDa, respectively (14). We anticipate, however, that the first-pass outflow curve of the 30-fold larger 2,000-kDa dextrans measured after hyaluronidase treatment represented their intravascular distribution in circulating plasma. The relative increase in dextran-2000 distribution volume after hyaluronidase was not different from the relative increase in red blood cell volume, confirming that after hyaluronidase, the distribution volume of dextran-2000 indeed likely represents the circulating plasma volume.
Our reported coronary baseline blood volume of 14–19 ml/100 g is in the high range of coronary blood volumes found in literature using the indicator dilution technique (from 7.3 to 17.8 ml/100 g) (8, 12, 26, 45). These relatively high baseline blood volumes coincide with relatively high baseline coronary blood flows in the current study, 190–250 ml·min−1100 g−1 (Table 1). Whereas baseline coronary blood flows reported in anesthetized large animals in the literature are typically ∼100 ml·min−1100 g−1 (10), the higher flows in the current study should be explained by a reduction in basal vasomotor tone of the resistance vessels due to the thorough instrumentation of the heart, including the use of a perfusion system, and the relatively high heart rates and perfusion pressures. Nevertheless, increases in coronary conductance during maximal adenosine were ∼2.3-fold, illustrating that there was sufficient potential for vasodilation.
Induction of anesthesia and the very initial surgery (until opening of the chest) were associated with a large decrease in large vessel hematocrit. Hematocrit of blood taken the day before the experiment was 32.6 ± 1.6%, corresponding nicely with the normal value of 34% that is reported in goats (41a), but this decreased immediately after the start of surgery to 24 ± 1.7%. Subsequent surgery with open thorax further decreased the hematocrit to ∼21% (see Table 1), and this small decrease might be explained by the addition of ∼110 ml of Gelofusine (B. Braun Melsungen), used for filling the perfusion system. The ∼2% decrease in hematocrit during the following blood volume measurements are likely the result of the blood sampling (∼20 ml per measurement) and subsequent replacement of lost blood with infusion fluid.
Acute Attenuation of Glycocalyx Barrier Properties Increases Coronary Blood Volume
The increase in coronary blood volume by adenosine found in this study, ∼75% of baseline volume, is consistent with the effect of this vasodilator in other studies, i.e., increases of 30–100% have been reported (8, 13, 36, 44) using different techniques: magnetic resonance imaging, measurements of oxygen content, X-ray contrast enhancement, and myocardial radioactivity. Hyaluronidase induced degradation of the glycocalyx constituent hyaluronan increases the glycocalyx porosity. As a result, coronary blood volume increased ∼40%. Subsequent to the degradation of glycocalyx by hyaluronidase, the additional 35% increase in coronary blood volume by administration of adenosine is most likely due to the vasodilator effects of adenosine on the resistance vessels.
Adenosine-induced blood volume recruitment.
Our data support the hypothesis that adenosine can increase microvascular blood volume by modulating the barrier properties of the endothelial glycocalyx and that this potential for recruitment is lost upon glycocalyx degradation (40, 41), attenuating the effect of adenosine on coronary blood volume increases. Our observation is in line with the intravital microscopic studies in cremaster tissue by Desjardins and Duling (9) who showed that whereas in control tissue, adenosine superfusion increased capillary tube hematocrit two- to threefold, adenosine did not increase the already elevated capillary tube hematocrit after heparinase treatment of the glycocalyx. Duling and coworkers (9, 21) were the first to suggest the possibility that the increase in capillary blood volume during adenosine administration could be the result of a reduction in glycocalyx exclusion volume, and, more recently, their group showed an increased porosity for 70-kDa dextrans into the glycocalyx domain (31). Recent observations from our group in mouse cremaster muscle showed that, similar to the effect of adenosine, the vasodilators bradykinin and sodium nitroprusside also increased capillary tube hematocrit and that the increase was paralleled by an increased porosity of the glycocalyx for 70-kDa dextrans (38). These effects were absent in hypercholesterolemic mice, which had an elevated capillary tube hematocrit and glycocalyx porosity under baseline conditions already, indicating glycocalyx degradation. Whereas these previous studies assessed changes in tube hematocrit and glycocalyx porosity in microvessels of cremaster tissue, we extended these observations to the coronary circulation of a large animal and determined the contribution of the glycocalyx to baseline and adenosine-induced total coronary blood volume in the heart.
Measurement of coronary flow reserve using adenosine.
Adenosine is the preferred vasodilator in the clinic for assessment of the relevance of coronary artery stenosis and perfusion abnormalities. During adenosine administration in the current study, there was an equal increase in flow and coronary conductance before and after hyaluronidase. Although hyaluronidase treatment has been associated with a reduction in shear-induced endothelium-derived nitric oxide (NO) release (25, 29), the equal increase in coronary conductance during adenosine infusion before and after hyaluronidase treatment implies that a potential hyaluronidase-associated depression of NO production did not affect the flow increase during adenosine. The coronary flow reserve appeared, therefore, unlike the volume that could be recruited, not influenced by enzymatic treatment of the glycocalyx, demonstrating that measurement of coronary flow reserve is insufficient to detect impairment of coronary blood volume recruitment in conditions of damaged glycocalyx.
The mechanism of coronary blood volume regulation appears to be, to some extent, uncoupled from that of blood flow regulation. It can be anticipated that the increase in coronary blood flow during adenosine is primarily a result of relaxation of the resistance vessels, whereas the coronary blood volume increase during adenosine is due to both recruitment of glycocalyx volume and vasodilation. Assuming complete removal of the glycocalyx after hyaluronidase, the remaining volume increase during adenosine infusion after enzymatic glycocalyx treatment (i.e., from 26.3 ± 2.7 to 33.9 ± 6.8 ml/100 g) will mainly represent the volume increase due to vasodilation only.
The endothelial glycocalyx is situated at the luminal side of endothelial cells throughout the entire vascular bed, and its thickness has been estimated to vary between different vessel types, from 0.2–0.9 μm in capillaries (11, 27, 31–33, 38, 42, 43), to 2–3 μm in small arteries with a diameter of ∼150 μm (37), to 4–5 μm in carotid arteries (23). Since by far the majority of the endothelial surface is in the microcirculation [of which ∼90% is capillaries (17)], these numbers indicate that the glycocalyx occupies a large part of the anatomic vascular volume particularly in the microcirculation. Therefore, we suggest that the reported effects on glycocalyx recruitment are to a large extent occurring in the microvascular compartment, particularly in the capillaries. In Fig. 5, we have summarized schematically the hypothesized effects of hyaluronidase and adenosine on capillary (top) and resistance vessel (bottom) blood volume and glycocalyx porosity. At baseline there is a blood-excluding glycocalyx present on the endothelial cells reducing functionally perfused capillary volume for flowing blood, as shown with intravital microscopy (11, 27, 31–33, 38, 42, 43). During adenosine infusion, recruitment of glycocalyx volume causes a robust increase in microvascular blood volume as reflected by the increase in capillary red blood cell content (and tube hematocrit), whereas the concurrent dilation of resistance vessels accounts for the coronary conductance increase as well as part of the increase in microvascular blood volume. Hyaluronidase induces a loss of blood-excluding glycocalyx volume, causing a significant increase in microvascular perfused volume at baseline. The increase in coronary blood volume during adenosine after hyaluronidase, only by vasodilation of the resistance vessels, is, as a result, reduced compared with the increase in coronary blood volume during adenosine with an intact glycocalyx. The impaired potential for glycocalyx volume recruitment can occur without a change in adenosine-induced increases in coronary conductance because of an unaltered dilator response of the resistance vessels to adenosine.
Physiological and Clinical Relevance
We found under control conditions, i.e., in the presence of an intact glycocalyx, that the transit time for red blood cells and dextrans through the coronary system was not reduced during adenosine despite a 1.7-fold increase in blood flow due to a concomitant increase in coronary blood volume. Consistent with this finding, direct observations of epicardial coronary capillary hemodynamics in the canine heart by Kiyooka et al. (20) showed that a 4.2-fold increase in blood flow with adenosine was associated with only modest increases in capillary red blood cell velocity in the epicardial capillaries. After hyaluronidase treatment of the glycocalyx, red blood cell transit time decreased during adenosine infusion compared with baseline transit times. Loss of glycocalyx volume recruitment capacity therefore could possibly impair the available time for red blood cells to offload oxygen when blood flow is increased to meet increases in metabolic demand. In line with this, we found in the current study a 20–30% decrease in myocardial oxygen consumption during administration of adenosine after glycocalyx degradation by hyaluronidase. Since in addition to regulation of blood transit time during conditions of increased blood flow, recruitment of glycocalyx volume in the capillaries also augments the surface area for exchange in these vessels, agonist-induced modulation of glycocalyx volume might provide a mechanism for the heart to match exchange capacity to increases in flow, e.g., during exercise.
Angina during exercise has been generally related to a diseased macrocirculation, for example, the presence of flow-limiting stenosis in a large coronary artery. A considerable number of patients with symptoms of chest pain during exercise, however, do not show a prominent stenosis in their coronary arteries or any coronary artery disease at all (16). Furthermore, a large number of the patients with a detectable lesion have a hemodynamically nonsignificant intermediate coronary artery stenosis, and these patients do not benefit from a percutaneous coronary intervention (30). Rather, in all these cases, the angina may relate to coronary microvascular dysfunction (7), of which loss of glycocalyx may be an early trigger (3, 39, 40). The current findings suggest that, rather than coronary flow reserve, the measurement of coronary blood volume recruitment may discriminate loss of glycocalyx. As such, MRI (15, 44) or contrast-enhanced ultrasound using microbubbles (18, 19) may be appropriate noninvasive methods for early detection of coronary vulnerability due to glycocalyx loss in patients with angina but without overt coronary artery disease.
This work was supported by Netherlands Heart Foundation Grants 2005T073 and 2003B181.
We thank Marion Kuiper, Anniek Lampert, Arne van Hunnik, Bart Eskens, and Hanneke Cobelens for assistance.
- Copyright © 2010 the American Physiological Society