INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Although hypoxic pulmonary vasoconstriction (HPV) is an intrinsic property of the pulmonary vasculature (1), several studies indicate that red blood cells (RBCs) play a role in the mechanism of its occurrence. McMurtry and colleagues (2, 3) reported that isolated rat lungs perfused with plasma had rapidly decaying HPV when compared with lungs perfused with blood or plasma plus RBCs. Other investigators have reported a similar dependence of HPV on blood perfusate in isolated rat-, cat-, and rabbit-lung models (4). Using a whole-animal model, we studied the effects of isovolemic anemia on intrapulmonary shunt in rabbits with left lung atelectasis, and found that anemia resulted in increased blood flow through the atelectatic lung (8). This increased intrapulmonary shunt could not be explained by changes in cardiac output or blood viscosity, and we hypothesized that anemia resulted in inhibition of HPV by an as yet unelucidated mechanism.

McMurtry and associates conjectured that RBCs prevented accumulation of an inhibitor or released a factor that enhanced vascular responsiveness. In a series of experiments, they showed that RBC-mediated modulation of HPV is unlikely to be mediated by cyclooxygenase products or adenosine (3). Hakim and Malik provided evidence that potentiation of HPV by RBCs in both rat and cat lungs was partly explained by species differences in muscular pulmonary vessel size, and/or by changes in RBC deformability in response to hypoxia (4, 9). Other evidence suggests that this phenomenon is unrelated to the antioxidant capacity of RBCs (10). A convincing explanation of the mechanism by which RBCs modulate HPV is lacking.

Because RBCs are integral to the removal of both nitric oxide (NO) and adenosine from the pulmonary circulation (11, 12), and because both of these products are known to affect the intensity of HPV, we hypothesized that these mediators could explain the interaction between RBCs and HPV. More specifically, we believed that an excess of NO and/or adenosine in the presence of anemia would lead to relative inhibition of HPV. Although McMurtry and colleagues' work implied that adenosine is not involved in RBC-mediated modulation of HPV, we took a different approach in evaluating this issue by studying adenosine-receptor blockade. In addition, we hypothesized that the effect of RBCs on HPV is dose-dependent (varies with hematocrit [Hct]), a phenomenon suggested in our previous whole-animal model but not studied in previous isolated-lung models.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The protocol for our study was approved by the Animal Care Committee of the Seattle Veterans Affairs Medical Center.

Experimental Preparation

A total of 81 New Zealand White rabbits, weighing 3 to 3.5 kg, were anesthetized with ketamine (15 mg · kg¹) and xylazine (0.33 mg · kg¹). Mechanical ventilation was initiated after tracheal intubation with room air, using a ventilatory rate of approximately 30 breaths/ min and a VT of 12 ml · kg¹ (peak airway pressure of approximately 20 cm H₂O). A carotid arterial line was placed, intravenous heparin was administered at a dose of 300 U · kg¹, and the rabbits were rapidly exsanguinated. Following median sternotomy, the pulmonary artery and left atrium were cannulated in situ. The pulmonary circulation was then perfused through a closed circuit connected to a Masterflex pump (Cole-Palmer, Barrington, IL), with a total circuit volume of 150 ml at a constant flow of 150 ml · min¹ (perfusate composition described subsequently). The perfusate was warmed to 38° C with a countercurrent exchange heater. The ventilatory gas was switched to a mixture of 21% O₂/5% CO₂, with the balance made up by N₂. The rabbit's chest was covered with plastic wrap to maintain humidity and surface temperature. Pulmonary artery pressure (Ppa) and left atrial pressure (LAP) were continuously measured via pressure tubing placed inside the perfusion cannulae and positioned at the inflow and outflow sites. LAP was kept constant at 5 mm Hg by adjusting the height of the venous reservoir. FI_O2 and FE_O2 were measured near the tracheal tube.

Perfusate Preparation

Preparations were perfused with one of three RBC perfusates for each set of experiments. Perfusates consisting of RBCs in Krebs- Henseleit solution at Hcts of approximately 30% and 10% were prepared as follows: whole blood was collected from donor rabbits immediately before the experiment and heparinized to prevent clotting. The blood was centrifuged at 3,000 × g for 10 min. The plasma and buffy coat were discarded and the RBCs were resuspended in saline and centrifuged again for 10 min. The supernatant was again discarded, and the RBCs were resuspended in an appropriate volume of buffered Krebs-Henseleit solution containing 3% dextran and 1% albumin to achieve the desired Hct. The suspended RBCs were filtered through a Pall high-efficiency leukocyte filter (Pall Biomedical Inc., Fajardo, PR) to remove residual leukocytes, and were then added to the perfusion circuit. High-Hct perfusates consisted of washed and leukocyte-filtered rabbit RBCs diluted with buffered Krebs-Henseleit solution to achieve an Hct of approximately 30%; low-Hct perfusates were treated identically, but were further diluted to achieve an Hct of 10%. Some preparations were perfused with Krebs-Henseleit solution alone (Hct-0); in these experiments, the lungs were perfused for several minutes to flush residual RBCs from the lungs; this initial perfusate was discarded and replaced with fresh perfusate prior to recirculation. Hct was measured in all preparations with a microcapillary centrifuge (International Equipment Corporation, Needham Heights, MA), and WBC count was measured in some experiments using a Coulter Counter (Coulter Electronics, Hialeah, FL).

Experimental Set A: Effect of Hct on HPV

After a 30-min stabilization period, preparations perfused with 0% (n = 10), 10% (n = 20), or 30% (n = 16) Hct suspensions were subjected to repeated hypoxic challenges with a gas mixture of 5% O₂/5% Co₂, and balance consisting of N₂ for 5-min intervals, with a 10-min stabilization period between challenges. All preparations were stable through four hypoxic challenges. The response to hypoxic challenges (HPV) was quantitated as the absolute change in pressure from baseline. Exhaled NO and perfusate NO metabolites (NO_x) were measured in most experiments, using the methodology described subsequently. Samples for measurement of adenosine levels were obtained during normoxia and hypoxia.

Experimental Set B: Effect of NOS inhibition on HPV

We investigated the effect of inhibition of NO production by the L-arginine analog N-nitro-L-arginine methyl ester (L-NNA). Preparations were perfused with either 0% (n = 6) or 30% Hct (n = 7) suspension and were ventilated with normoxic gas for 30 min. The preparations were then subjected to three repeated 5-min hypoxic challenges separated by 10-min intervals of normoxia. After the third challenge, L-NNA was added to the perfusates the achieve a concentration of 0.1 mM. After 15 min of normoxic perfusion, an additional hypoxic challenge was performed. In two additional preparations perfused with 0% Hct suspension, excess L-arginine was added to the perfusates immediately before addition of L-NNA. Expired NO was measured in all experiments, as described subsequently. The data from the challenges immediately prior to and following administration of L-NNA were analyzed.

Experimental Set C: Effect of Adenosine-receptor Blockade on HPV

8-(p-Sulfophenyl)theophylline (8-PT; Research Biochemicals International, Natick, MA), a nonselective adenosine receptor antagonist, was added to 10% and 30% Hct perfusates to achieve a concentration of 5 ug · ml¹ (13) at the beginning of the stabilization period, yielding two groups (10 adenosine-receptor blockade [AB], n = 8; and 30AB, n = 5). The concentration of 8-PT used was verified to be effective in blocking the effects of adenosine in a subset of experiments. The hypoxic challenge protocol was followed as described earlier. Exhaled NO was measured. In an additional two preparations perfused with 0% Hct suspension, 8-PT was added to the perfusate after two hypoxic challenges. The pressor response to hypoxia was reevaluated 15 min after addition of 8-PT. The data from these two experiments are presented with those for groups 10AB and 30AB for simplicity.

NO Measurement Methods

Mixed expired NO was measured with a chemiluminescence detector (Sievers Instruments Inc., Boulder, CO) by continuous sampling from a 50-ml reservoir placed in the expired-gas line (14). The sample flow rate was 120 ml · min¹, of a fixed minute ventilation (E) of 900 ml · min¹. Calibration was done with gas from the normoxic and hypoxic tanks as a zero, and with a certified tank containing NO at 5.6 ppm (Air Liquide, Long Beach, CA), with sampling at 120 ml · min¹.

Intravascular NO release was quantitated in 22 experiments (Experimental Set A; Group 0: n = 8, Group 10: n = 6, Group 30: n = 8) by measurement of perfusate nitrite, peroxynitrite, and nitrate (NO_x) through the reduction of deproteinated perfusate samples (14). Perfusate samples were collected in 1.0-ml aliquots, which were placed on ice and then centrifuged to pack the RBCs. The perfusate was drawn off (0.3 ml), mixed with ice-cold ethanol (0.6 ml), placed on ice, and centrifuged to pellet the protein. The deproteinated perfusate was removed and frozen for later analysis. The samples were assayed for NO_x by reduction of these species to NO in an acidic vanadium (III) chloride solution at 85° C in a reaction vessel. The gas from the reaction was drawn into the NO detector at approximately 60 ml · min¹, and the peaks were recorded for later integration. Calibration was performed by reduction of known concentrations of NaNO₂ to NO in the reaction vessel. Peak areas were determined by summing the points making up the peak, with subtraction of any baseline.

Adenosine-Measurement Methods

Adenosine-level assays were performed as follows: 2-ml samples for adenosine assays were centrifuged at 12,000 × g for 15 s, and the supernatant was frozen in a methanol-dry ice bath within 30 s after the sample was obtained. Samples were lyophillized and stored at 70° C until analysis. Analysis of samples was done with high-pressure liquid chromatography (HPLC) (Gilson gradient HPLC system, Middleton, WI).

Statistical Analysis

Statistical analysis was done with the StatView software package (Abacus Concepts, Berkeley, CA). Factorial analysis of variance (ANOVA) was used to test for differences between groups in HPV and NO, and a paired Student's t test was used to test for within-group differences. A value of p < 0.05 was accepted as statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The typical Ppa profile in response to hypoxic ventilation in the preparation is shown in Figure 1. Ppa increased rapidly after initiation of hypoxia, and reached a plateau after approximately 2 min. Ppa returned to baseline within approximately 3 min after the resumption of normoxic ventilation.

View larger version (13K): [in this window] [in a new window]   Figure 1.   A representative Ppa tracing from a preparation perfused with a 10% Hct suspension. Ppa increases rapidly with the initiation of hypoxic ventilation and reaches a plateau after approximately 2 min. Ppa returns to baseline within approximately 3 min after resumption of normoxic ventilation.

Experimental Set A

Baseline (normoxic) pressure was higher in Group 30 than in Groups 10 or 0 (p < 0.05 versus Group 10, p < 0.01 versus Group 0; Figure 2). HPV during seven consecutive hypoxic challenges in groups perfused with 30%, 10%, and 0% Hct suspensions is represented in Figure 3. Because some preparations dropped out after the fourth challenge (developed intractable pulmonary hypertension or pulmonary edema), the data were tested statistically at this point. HPV was greater in Group 30 than in Groups 10 and 0 (p < 0.001), and was greater in Group 10 than in Group 0 (p < 0.01).

View larger version (46K): [in this window] [in a new window]   Figure 2.   Baseline (normoxic) Ppa in groups perfused with 30%, 10%, and 0% Hct suspensions. Baseline pressure is higher in Group 30 than in Group 10 (p < 0.05) or Group 0 (p < 0.01).

View larger version (15K): [in this window] [in a new window]   Figure 3.   HPV (change in Ppa from baseline during hypoxia) in groups perfused with suspensions having Hcts of 0%, 10%, and 30%. Statistical testing was performed at the fourth hypoxic challenge. *p = 0.01 versus Group 0; **p < 0.0001 versus Groups 0 and 10.

Expired NO data and perfusate NO_x levels at the fourth hypoxic challenge for Groups 30, 10, and 0 are shown in Figures 4 and 5. Expired NO levels varied inversely with Hct, and were significantly different among groups (p = 0.02 for Group 0 and versus Group 10, p < 0.001 Group 0 versus Group 30, and p = 0.02 for Groups 10 versus Group 30). Expired NO fell for with hypoxic challenges in all groups (p < 0.01, normoxia versus hypoxia, all groups). Perfusate NO_x levels were higher Groups 10 and 30 than in Group 0 during normoxia and hypoxia (p = NS); however, we were unable to detect significant changes in the rate of perfusate NO_x accumulation during hypoxic challenges.

View larger version (19K): [in this window] [in a new window]   Figure 4.   Expired NO during normoxia and hypoxia in groups perfused with suspensions having Hcts of 0%, 10%, and 30%. SE bars are included only at representative points (normoxia, peak hypoxia, and resumption of normoxia) to preserve the clarity of the figure. Expired NO falls in all groups with the initiation of hypoxia and returns to baseline upon resumption of normoxia. Expired NO varied inversely with Hct. *p = 0.02 versus Group 30; **p < 0.0001 versus Group 30 and p = 0.02 versus Group 10; p < 0.01, hypoxia versus normoxia, all groups.

View larger version (47K): [in this window] [in a new window]   Figure 5.   Perfusate NO metabolites (nitrite, peroxynitrite, and nitrate; NOx) in groups perfused with suspensions having Hcts of 0%, 10%, and 30% during normoxia and hypoxia. There was a trend toward increased perfusate NOx as Hct increased, but the differences were not statistically significant.

Adenosine was undetectable down to a detection level of approximately 10 pM in either Group 30 or Group 10 at any time during the experimental protocol.

Experimental Set B

Figure 6 illustrates the effect of NOS inhibition on HPV in groups perfused with 0% and 30% Hct suspensions. The baseline response to hypoxia was significantly greater in groups perfused with RBC-containing suspensions than in the 0% Hct group (p < 0.05). Administration of L-NNA led to marked augmentation of HPV in the 0% Hct group (p < 0.05 versus baseline), but produced no increase in the 30% Hct group. Addition of L-arginine to the 0% Hct perfusates just prior to administration of L-NNA prevented augmentation of HPV (group 0 + 1-arg). NOS inhibition caused a decrease in expired NO to less than 10 ppb in both the 0% and 30% Hct groups (p < 0.01 versus baseline) (Figure 7). Administration of L-arginine prior to L-NNA attenuated the decrease in expired NO.

View larger version (19K): [in this window] [in a new window]   Figure 6.   HPV in groups perfused with 0% and 30% Hct suspensions before and after administration of L-NNA. In group 0 + L-arg, L-arginine was added to the perfusates immediately before addition of L-NNA. HPV was higher in the groups perfused with RBCs prior to administration of L-NNA; after L-NNA, however, HPV was markedly augmented in the 0% Hct group. There was no baseline pressor response to hypoxia in the 0 + L-arginine group, and only a minimal response after administration of L-arginine plus L-NNA. Note that there was no initial pressor response to hypoxia in group 0 + L-arg. *p < 0.05 versus 10% and 30% Hct; **p < 0.05 versus baseline.

View larger version (15K): [in this window] [in a new window]   Figure 7.   Expired NO in groups perfused with 0% and 30% Hct suspensions before and after administration of L-NNA, and in a 0% Hct group in which L-arginine was added to the perfusates immediately before L-NNA. The concentration of expired NO was much higher in the 0% Hct group before administration of L-NNA; after L-NNA, expired NO fell to less than 10 ppb in both groups. L-arginine attenuated the L-NNA-induced decrease in expired NO. *p < 0.001 versus 10% and 30% Hct; **p < 0.01 versus baseline.

Experimental Set C

HPV in groups treated with the adenosine receptor antagonist 8-PT is shown in Figure 8. Adenosine-receptor blockade had no significant effect on the magnitude of HPV or on the augmentation of HPV by higher Hct, and HPV remained significantly higher in the 30% Hct than in the 10% Hct group (p < 0.05) (statistical testing versus the 0% group was not done because of the slightly different protocol followed for these studies). In similarity to the previous set of experiments, expired NO was greater at all times in the lowest Hct group (p = 0.02, 10% Hct versus 30% Hct). As in the previous groups, expired NO fell with hypoxic challenges. There were no significant differences in the absolute values of expired NO in the groups treated with adenosine-receptor blockade and the respective Hct groups without blockade.

View larger version (39K): [in this window] [in a new window]   Figure 8.   HPV after adenosine-receptor blockade with 8-(p-sulfophenyl)theophylline in groups perfused with suspensions with Hcts of 0% and 30%. The relationship between HPV and Hct was not affected by adenosine-receptor blockade. *p < 0.05 for groups versus one another.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Previous investigators have documented the potentiating effect of RBCs on HPV with isolated-lung models. Our data are consistent with previously published data; in addition, we have shown that augmentation of HPV varies with the number of RBCs, with HPV being stronger in the presence of a higher Hct. This "concentration-dependent" effect has not been previously demonstrated. We have also provided evidence that RBC-modulated enhancement of HPV: (1) may be related to NO metabolism, with anemia resulting in more available NO and thus in diminished HPV; and (2) does not result from modification of circulating adenosine levels by RBCs.

Although previous investigators have studied the effect of RBCs on HPV at single Hcts, (4, 9), our study is the first to use an isolated, perfused lung model to show that HPV varies in intensity across a broad range of Hcts. In a previous study using a whole-animal model of left-lung atelectasis, we made similar observations (8). That the number of RBCs, and not just the presence of RBCs, determines the strength of HPV is an important physiologic observation, which may have clinical relevance. Our findings are consistent with a dose-dependent interaction of RBCs with a mediator capable of affecting the intensity of HPV. Two potential mediators of this phenomenon are NO and adenosine.

NO and HPV

A potential mechanism by which anemia attenuates HPV involves the endothelium-derived relaxant factor NO or a related adduct, such as a nitrosothiol. NO acts as both a systemic and a pulmonary vasodilator, and changes in basal NO levels may act to modulate HPV (15). Administration of inhaled NO results in inhibition of HPV (16), and the observation that inhibition of NO synthesis markedly enhances the hypoxic pressor response suggests that basal release of NO acts to moderate vasoconstriction during hypoxia (17).

The physiology of NO after its production by pulmonary arterial endothelial cells is complex and incompletely understood. Our experiments have isolated only one aspect of the pulmonary biology of NO: the relative partitioning of NO between smooth-muscle cells, the blood stream, and expired gas as affected by alterations in hemoglobin (Hb) concentration, and the resulting vascular effects of that partitioning. NO is inactivated by binding to heme iron, for which it has great affinity (11, 21). RBCs modify the effect of circulating (and inhaled) NO (22, 23), potentially by scavenging NO and reducing the effective level, and also reduce the amount of NO appearing in expired gas (24). Anemia thus appears to cause a change in the concentration gradient for NO, resulting in the availability of more NO to vascular smooth-muscle, where it blunts the hypoxic pressor response, and to the plasma, where it is carried to the alveoli and subsequently appears in the expired gas. The idea is in part supported by our perfusate NO_x data, which showed that perfusate NO_x levels were lowest in the absence of an Hb "sink" to keep NO within the vascular space (Figure 5).

In our experiments, the concentration of expired NO was consistently higher in the presence of a lower Hct (Figure 4). The increased expired NO in the low-Hct groups in our study may at least partly explain the observed blunting of the pressor response to hypoxia seen in those groups. Administration of the NOS inhibitor L-NNA led to a marked reduction in expired NO and augmentation of HPV in a group perfused with a 0% Hct, suspension, but resulted in little to no change in a group perfused with a 30% Hct suspension (Figure 6). That augmentation of HPV by L-NNA is a specific result of NOS inhibition is illustrated by the prevention of this augmentation by L-arginine, a competitive inhibitor of L-NNA. Thus, our data show that HPV depends on the level of Hct, that NO flux is affected by levels of Hb (Hct) in a physiologically relevant range, and that NOS inhibition eliminates the Hct dependence of HPV.

L-NNA dramatically increased HPV in buffer-perfused lungs, to the extent that HPV was much stronger than in any similarly treated lungs perfused with RBCs (Figure 6). This interesting observation suggests that although RBCs augment HPV, their presence may in some way also limit the maximum intensity of HPV, particularly in the presence of NOS inhibition. Two potential mechanisms may account for this finding: (1) RBCs may cause production of other vasodilatory substances (e.g., prostacyclin) through shear-stress-related mechanisms (25); and (2) residual circulating NO equivalents, in the form of S-NOHb, S-NO glutathione, or other nitrosylthiols, may remain present after NOS inhibition, and continue to contribute to modulation of vascular tone for some time despite a measurable decrease in expired NO (26).

The concentration of expired NO decreased with individual hypoxic challenges (Figure 4), a finding consistent with that previously reported (29). Expired NO fell more with individual hypoxic challenges in the low-Hct groups, a result that is unexplained and counter to that expected on the basis of the relative vascular responsiveness of the groups. The actual decline in expired NO with hypoxia, and the differences in this decrement between groups, are, however, small, and may be of little physiologic importance, owing stores of buffered NO and NO derivatives.

Adenosine and HPV

Adenosine acts predominantly at pulmonary P₁-purinoreceptors (subtype A₂) to produce vasodilation (13, 32). Although there is no known direct link between adenosine and HPV, adenosine is released in response to hypoxia in many tissues, including lung (33), and inhibits HPV in human and rat lungs when exogenously administered (34, 35). In addition, adenosine is removed from the pulmonary circulation via cellular uptake by RBCs (12). This evidence suggests a potential modulating role for adenosine in HPV, with RBCs participating in this modulation by affecting circulating adenosine levels (3). We found that adenosine-receptor blockade did not significantly augment HPV or alter the Hct-dependent nature of HPV (Figure 8). In addition, adenosine levels were undetectable in any of the perfusates, and did not change with hypoxia or with time. These data make it unlikely that adenosine plays a significant role in RBC-mediated modulation of HPV.

In summary, we have shown that HPV in isolated rabbit lungs is potentiated by RBCs and is Hct dependent, being greater at a higher Hct. In addition, we have shown that expired NO levels are higher when Hct is lower, suggesting that a lower RBC number (and Hb concentration) results in diminished scavenging of NO. Most importantly, we have shown that inhibition of NO production reverses the observed differences in HPV between preparations perfused with buffer and those perfused with RBCs plus buffer. The RBC-dependence of HPV is thus explained by a relative excess of bioavailable NO in buffer-perfused preparations, which results in attenuation of HPV. This mechanism may also explain the difference in HPV between preparations perfused at differing Hcts. Buffer-perfused lungs may not be the ideal model for studying pulmonary physiologic processes, since they eliminate an important link in the modulation of pulmonary vascular tone in the form of RBCs. Given the potential importance in HPV of S-nitroso compounds (especially S-nitrosohemoglobin), future studies of RBC modulation of HPV and pulmonary vascular resistance should ideally employ erythrocytes whose S-nitrosohemoglobin concentrations are preserved.