Journal of Applied Physiology
Vol. 83, No. 2, pp. 495-502, August 1997
PULMONARY CIRCULATION AND LUNG FLUID BALANCE

Flow-induced vasodilation in the ferret lung

Joseph H. Chammas, David. A. Rickaby, Margarita Guarin, John H. Linehan, Christopher C. Hanger, and Christopher A. Dawson

Departments of Surgery, Physiology, and Pediatrics, Medical College of Wisconsin, Milwaukee 53226; Biomedical Engineering Department Marquette University, Milwaukee 53233; and Zablocki Veterans Affairs Medical Center, Milwaukee, Wisconsin 53295

ABSTRACT

INTRODUCTION

MATERIALS AND METHODS

RESULTS

DISCUSSION

ACKNOWLEDGEMENTS

FOOTNOTES

REFERENCES

ABSTRACT

Chammas, Joseph H., David. A. Rickaby, Margarita Guarin, John H. Linehan, Christopher C. Hanger, and Christopher A. Dawson. Flow-induced vasodilation in the ferret lung. J. Appl. Physiol. 83(2): 495-502, 1997.To examine the possibility that shear stress may be a pulmonary vasodilator stimulus, we studied the effect of changing blood flow on the diameters of small pulmonary arteries in isolated perfused ferret lung lobes. The arteries studied were in the ~0.3- to 1.3-mm-diameter range, and the diameters were measured by using microfocal X-ray imaging. The diameters were measured at two flow rates, 10 and 40 ml/min, with the intravascular pressure in the measured vessels the same at the two flow rates as the result of venous pressure adjustment. The response to a change in flow was studied under both normoxic and hypoxic conditions. Hypoxia was used to elevate pulmonary arterial tone to increase the likelihood of detecting a vasodilator response. Under normoxic conditions, changing flow had little effect on the arterial diameters, but under hypoxic conditions the arteries were consistently larger at the higher flow than at the lower flow, even though the distending pressure was the same at the two flow rates. The results are consistent with the hypothesis that shear stress is a pulmonary vasodilator stimulus.

shear stress; hypoxic pulmonary vasoconstriction; pulmonary arterial diameters; nitric oxide

INTRODUCTION

INCREASING SHEAR STRESS can lead to vasorelaxation in various organs (6, 16, 17). Whether this phenomenon occurs in pulmonary arteries has not been clearly established. Examination of the possibility of shear stress-induced dilation in the pulmonary vascular bed presents different problems than in systemic arteries. Under normal conditions, pulmonary arteries have relatively little background tone, making detection of vasorelaxation problematic. Also, the distension that occurs with increased pressure usually associated with increased flow confounds any contribution of active flow-dependent relaxation to the pulmonary vascular pressure-flow relationship. Shear stress-induced dilation has been studied in isolated systemic vessels (24, 20), but it is not clear that intrapulmonary arteries isolated from the lung parenchyma retain all their in situ response patterns (21). Therefore, to examine the possibility that increased flow can be a vasodilator stimulus in pulmonary arteries within the intact lung, we took an approach that, to our knowledge, has not been utilized for this purpose before. We used microfocal X-ray imaging to measure the diameters of pulmonary arteries in the 300- to 1,300-µm-diameter range at two different flow rates but with the local intravascular pressure held constant. The experiments were carried out on left lower lobes from ferret lungs. To increase vascular tone, the PO₂ was decreased to produce hypoxic vasoconstriction. The ferret lung was chosen because, unlike lungs from some other species convenient for laboratory study, it has a stable and reproducible hypoxic response in isolated preparations required for the experimental design of these studies. Also the anatomy of left lower lobe of the ferret lung is particularly advantageous for isolated perfusion.

MATERIALS AND METHODS

Each of 11 ferrets [1.82 ± 0.28 (SD) kg body wt] was anesthetized (30 mg/kg) pentobarbital sodium, heparinized (3,000 IU), and exsanguinated via a catheter in the carotid artery. Thirty milliters of 10% dextran (Rheomacrodex; 40,000 mol wt) were infused in 10-ml increments during the exsanguination procedure. The autologous blood [~125 ml; hematocrit 24.8 ± 3.1 (SD)%] was added to the perfusion system. After exsanguination, the trachea was clamped, and the chest was opened through a midline sternotomy. Polyethylene cannulas (PE-240; 1.67 mm ID, 2.42 mm OD) 4-5 cm in length were placed in the left lower lobar artery, vein, and bronchus. The lobe was then removed and placed horizontally in a perfusion chamber. The perfusion system included a Masterflex roller pump (with 7014 head) that pumped the blood from a heated (37-38°C) reservoir, through a heat-exchange coil and an injection loop, and into the lobar artery. The blood then drained back into the reservoir. The reservoir height could be adjusted to set the venous pressure (Pv) to the desired level during the course of the experiment.

The lobes were ventilated via the lobar bronchus by a piston respirator with tidal volume of 6-8 ml at a frequency of 40 inflations/min, resulting in an end-inspiratory pressure of 6.9 ± 0.7 (SE) Torr and an end-expiratory pressure of 1.1 ± 0.3 Torr. The ventilating gas mixtures were ~16% O₂-6% CO₂-balance N₂ (normoxia) or 7% O₂-6% CO₂-balance N₂, (hypoxia). During hypoxia, the circulating blood volume was temporarily reduced to ~50 ml to shorten the time required for equilibration. When the ventilating gas was changed to the hypoxic mixture, it took ~5 min for the perfusion pressure to reach its new elevated steady level, and 15 min were sufficient for inlet and outlet PO₂ to nearly equilibrate. Fifteen minutes after the gas mixture was changed, arterial and venous blood samples were collected for blood-gas analysis. The objective was to maintain the PO₂ above a level that produced hypoxic vasoconstriction under "normoxic" conditions and to decrease the PO₂ to a level that produced an increase in perfusion pressure of ~5 Torr at a flow rate of 40 ml/min under hypoxic conditions. Also, under hypoxic conditions, it was considered important to equilibrate the entire blood volume of the perfusion system so that when the ventilation was stopped there would be no substantial change in the PO₂ within the lung. Blood-gas values are given in Table 1.

Table  1.   Blood gases Normoxia Hypoxia Arterial Venous Arterial Venous PO2, Torr 107 ± 2  132 ± 3  57 ± 1  55 ± 1  PCO2, Torr 37.9 ± 1.4  38.7 ± 1.2  35.7 ± 1.4  36.9 ± 1.2  pH 7.33 ± 0.01  7.32 ± 0.01  7.32 ± 0.02  7.31 ± 0.02 Values are means ± SE for 11 ferrets.

The perfusion chamber was placed between the X-ray source (40-µm focal spot) and the X-ray-sensitive video camera of a Nicolet NXR-200 X-ray imaging system so video images of the lobe could be recorded with an S-VHS videocassette recorder as a bolus of contrast medium passed through the lobar vasculature, as previously described (2, 3). An injection loop upstream from the arterial cannula allowed the introduction of the bolus of 0.15 ml of radiopaque contrast medium, 61% iopamidol (Isovue-300), into the lobar arterial inflow by the activation of a solenoid valve to redirect the inflow through the bolus-containing segment of the loop without changing the pressure or flow. The contrast medium was maintained at the same temperature as the perfusion system. Before each bolus injection, the ventilation was stopped at end expiration. The lobe thickness in the field of view, on the axis parallel to the vertical X-ray beam, was 0.5-1.0 cm.

To measure the arterial pressure (Pa), a small catheter was advanced either to a small artery in the field of view (in 5 of 11 experiments) or to the tip of the arterial cannula (in 6 of 11 experiments). The catheters were constructed by heating a length of polyethylene tubing (PE-50; 0.86 mm ID, 1.27 mm OD) and then pulling it while it cooled. This resulted in catheters with a 3- to 4-cm taper to an outside diameter of ~250 µm at the tip. A 90-µm guide wire was inserted into the catheter to aid in placement. To place the catheter in a small artery, it was negotiated through the major bifurcations so that the tip was in a vessel ~1 mm in diameter. Then, the guide wire was removed, and the catheter lumen was allowed to fill passively with blood. Then, the catheter was filled with contrast medium so that the final placement of the tip could be made into vessels with average diameter of 877 ± 191 (SD) µm. The catheter was then secured in place and connected to a pressure transducer. Images showing a portion of the arterial tree with a catheter in place are show in Fig. 1. Pressures were also measured at a site distal to the solenoid valve of the injector loop and proximal to the arterial cannula and at a site distal to the lobar vein cannula. The pressure drops between these pressure-measuring sites and the cannula tips were measured at the end of each experiment at each of the flow rates used in the study. Thus the cannula pressure drops could be subtracted from (arterial) or added (venous) to the pressures measured during the experiment to determine the lobar arterial inlet pressure and the venous outlet pressure (Pv), respectively. Pressures were measured by using Deseret transducers and recorded on a Grass model 7 recorder and CODAS data-acquisition system at 10 Hz. Any systematic differences between the lobar artery inlet pressure and the pressure measured with the catheter placed in a small pulmonary artery within the field of view were not detectable (Fig. 2). Therefore, in the presentation of the results, the Pa refers to the small-catheter tip pressure whether the catheter tip was placed in the field of view (Fig. 1) or at the lobar arterial inlet.

When the preparation was complete and equilibrated with the normoxic gas mixture, arterial inlet and venous outlet blood samples were collected for gas analysis, and the experiment to determine the influence of flow on vessel diameter at constant pressure was begun. Bracketing protocols were used, either with the flow set at 40 ml/min, then at 10 ml/min, and then returned to 40 ml/min (protocol I) or with the 10 ml/min bracketing 40 ml/min (protocol II). The Pa was set at approximately the same pressure (within 2 Torr) at the two flows by adjusting the height of the reservoir. The time after a change in flow before bolus injection was ~1 min. In each experiment, the first protocol was protocol I under normoxic conditions. Then, in the subsequent sequence, protocols I and/or II were carried out under hypoxic conditions, followed by normoxia. The normoxia and hypoxia sequence was repeated up to two more times, such that the average numbers of protocols I and II performed on each lung were 2.5 and 1.0, respectively, under normoxic conditions and 2.5 and 1.9, respectively, under hypoxic conditions.

To compare the vessel diameters under the same pressure and flow conditions during hypoxic and normoxic conditions, in six of the experiments, after a normoxia-hypoxia sequence, a bolus was injected under normoxic conditions and with the Pa increased to match the hypoxic Pa at the 10 ml/min flow by increasing the Pv.

To begin to address mechanisms that might contribute to flow-dependent effects, in four experiments, after these protocols had been carried out, N-nitro-L-arginine methyl ester (L-NAME) was added to the reservoir (0.22-3.7 µmol/ml blood). Protocol I was carried out in three of the four experiments during normoxia, and protocols I (4 of 4 experiments) and II (1 of 4) were carried out during hypoxia.

In addition, in each experiment for calibration of the imaging system, a remotely controlled micrometer was used to move the lung lobe a measured distance (~2 mm) while the lung image was being recorded.

The videotaped images were analyzed off-line to measure the internal diameters of three to seven vessels per experiment, the number depending on the number of measurable vessels in the field of view. A region of interest was placed over the image of the vessel with diameter that was to be measured. The videotape was advanced frame by frame until the maximum absorbance during passage of the contrast medium was attained. Then, the absorbance across the vessel cross section was measured, and the diameter was estimated by using the cylindrical model-based algorithm as previously described (1, 7). This procedure was carried out three times to provide triplicate estimates, which were averaged to obtain a single diameter estimate for each measured vessel for each bolus injection. To calibrate the imaging system for calculating vessel diameters in micrometers, the actual distance moved during the calibration displacement (µm) was divided by the distance moved by a vessel landmark on the image (pixels). For presentation of the data, the diameters or the differences in diameters between flows were averaged under each study condition such that each vessel provided one data entry under each study condition. The exception to this was when comparisons were made between the first and third boluses (the bracketing boluses) in each protocol to ascertain that the diameter changes were not time or previous-history dependent. For those comparisons, the diameters from the first and third boluses were averaged separately. Also, it was not possible to measure every vessel for every bolus because of variations in the image quality revealed by an inability of the diameter-measuring algorithm to converge. By this criterion, 1,096 of a possible 1,313 (~83.5%) diameter measurements were successful without systematic bias with regard to study conditions.

The reported pressures were measured at the time of the bolus injection and were averaged in the same manner as the diameters for the data presentations.

RESULTS

Figure 3 and Table 2 show Pa and Pv during normoxia and hypoxia in the protocol format, thus indicating the extent to which the objectives of maintaining similar Pa at the two flows and stability of the pressure response to hypoxia during each protocol were achieved. Also shown in Table 2 is the average fractional difference between the diameters at the two flows, indicating the extent to which the diameter differences between the first and second and between the second and third boluses in each protocol were reproducible.

Table  2.   Mean pressures and percent diameter changes in protocols I and II during normoxia and hypoxia Flow, ml/min 40 10 40 Normoxia protocol I  D, %  0.7 ± 0.4    0.6 ± 0.4  Pa, Torr 21.2 ± 0.9  21.3 ± 0.9  20.6 ± 0.9  Pv, Torr 11.0 ± 0.4  19.6 ± 0.6  11.2 ± 0.6  Pa  Pv, Torr 10.1 ± 0.8  1.7 ± 0.6  9.4 ± 0.7  Normoxia protocol II  D, %  1.2 ± 0.7    1.9 ± 0.5  Pa, Torr 22.3 ± 1.0  21.8 ± 1.1  22.5 ± 0.9  Pv, Torr 19.9 ± 0.6  11.0 ± 0.8  20.1 ± 0.6  Pa  Pv, Torr 2.4 ± 0.6  10.8 ± 1.0  2.7 ± 0.6  Hypoxia protocol I  D, %  14.1 ± 0.9    14.6 ± 1.0  Pa, Torr 27.5 ± 0.9  27.9 ± 0.9  28.3 ± 1.1  Pv, Torr 10.6 ± 0.4  24.7 ± 0.9  10.8 ± 0.7  Pa  Pv, Torr 16.9 ± 0.8  3.1 ± 0.7  17.5 ± 1.3  Hypoxia protocol II  D, %  13.2 ± 1.3    13.5 ± 1.3  Pa, Torr 29.1 ± 1.5  30.4 ± 1.0  30.1 ± 0.9  Pv, Torr 24.9 ± 1.3  10.0 ± 0.9  25.6 ± 0.9  Pa  Pv, Torr 4.2 ± 0.5  20.4 ± 0.7  4.4 ± 0.5 Values are means ± SE for 54 measured vessels. Pa, arterial pressure; Pv, venous pressure; change in artery diameter (D) is (D40/D10  1) × 100 on going from 40 to 10 ml/min flow or from 10 to 40 ml/min flow, where D40 is artery diameter at 40 ml/min flow and D10 is artery diameter at 10 ml/min flow.

The normoxic artery diameters at the flow of 40 ml/min (D_40N) as a fraction of their respective diameters at 10 ml/min (D_10N) vs. D_10N and the hypoxic diameters at a flow of 40 ml/min (D_40H) as a fraction of their respective diameters at 10 ml/min (D_10H) are plotted vs. D_10N on Fig. 4. During normoxia, any systematic differences in diameters between flows were very small, although D_40N averaged 0.85 ± 0.28% (SE) larger than D_10N, which was statistically significant (P < 0.005) as judged by paired t-test. When the vascular tone was increased by hypoxia, the Pa averaged 8.0 ± 1.1 (SE) Torr higher than during normoxic conditions, and the D_40H values were an average of 14.2 ± 0.9% (SE) larger (P < 0.001) than the D_10H values. The general trend was for a greater difference between D_40H and D_10H for smaller vessels in the range of vessel diameters studied.

In protocols I and II, the average of the vessel diameters was about the same during normoxia and hypoxia. Figure 5, which shows the effect of hypoxia on vessel diameters at 10 ml/min flow, indicates that this was the result of the fact that the small vessels tended to be narrowed by hypoxia, whereas the larger vessel diameters were usually increased during hypoxia. To determine whether the increased diameters resulted from vasorelaxation due to hypoxia or to distension due to the higher pressure during hypoxia, the experiments leading to the results shown in Fig. 6 were carried out. In these experiments, the Pa was raised during normoxia to the same level as during hypoxia by raising the Pv instead of by hypoxic vasoconstriction. At equal pressures, the hypoxic diameters were about equal to or smaller than the normoxic diameters, suggesting that the increases in diameter on Fig. 5 were due to distension rather than to vasorelaxation.

Table 3 provides the pressure data obtained after L-NAME was added to the blood. The format is the same as Table 2 except that only protocol I data are provided because protocol II was only carried out once with L-NAME. The addition of L-NAME did not significantly affect the diameters of the measured vessels. At 10 ml/min flow, the vessels diameters averaged 100.4 ± 5.0% (SD) of their pre-L-NAME diameters under normoxic conditions and 101.9 ± 11.8% (SD) of their pre-L-NAME diameters under hypoxic conditions. Figure 7 shows the effect of changing flow on diameter after L-NAME was added. The format is similar to Fig. 4, except that the effect of flow during hypoxia without L-NAME for those vessels in the L-NAME treatment group is also included for comparison. After the L-NAME was added, the diameter response to increased flow was not observed under either hypoxic or normoxic conditions even though the hypoxic vasoconstriction was not diminished by L-NAME.

Table  3.   Mean pressures and percent diameter changes in protocol I during normoxia and hypoxia after L-NAME treatment Flow, ml/min 40 10 40 Normoxia protocol I  D, %   1.4 ± 1.5    0.4 ± 0.8  Pa, Torr 21.1 ± 3.5  21.4 ± 3.2  19.4 ± 4.4  Pv, Torr 10.7 ± 0.7  19.9 ± 0.8  9.5 ± 1.0  Pa  Pv, Torr 10.4 ± 2.8  1.5 ± 2.6  9.9 ± 3.4  Hypoxia protocol I  D, %  0.4 ± 1.0    1.1 ± 1.5  Pa, Torr 29.5 ± 3.1  28.5 ± 1.9  29.2 ± 3.4  Pv, Torr 10.3 ± 4.0  24.4 ± 4.2  8.1 ± 1.9  Pa  Pv, Torr 19.2 ± 3.8  4.1 ± 1.2  21.1 ± 4.4 Values are means ± SE for 14 and 16 measured vessels for the normoxia and hypoxia, respectively. L-NAME, N -L-arginine methyl ester.

DISCUSSION

The results are consistent with the hypothesis that acutely increased shear stress can be a vasodilator stimulus in pulmonary arteries of the ferret lung. The increase in vessel diameter with flow was clearly observable only when the vessels were constricted by hypoxia. This is to be expected for pulmonary vessels having little tone under normoxic conditions. Hypoxia was chosen as the vasoconstrictor stimulus because it is the only stimulus, that we know of, that is both applicable to the isolated lung and has a stimulus strength that can be assumed to be flow independent. The use of vasoactive agents infused or injected into the blood is not a practical alternative because the rates of delivery to, and clearance from, the receptors, i.e., the strength of the stimulus, can vary with flow, and there is not an obvious method for adjusting the dosage according to the flow to achieve a constant stimulus strength. That problem might be minimized by using supramaximal doses of infused agents, but the high vascular resistances under such conditions would not be compatible with the experimental design of this study. The lack of an alternative stimulus for comparison with hypoxia leaves open the possibility that the flow-dependent vasodilator mechanism specifically interferes with the hypoxic vasoconstriction rather than being a general vasodilator response.

While these results apparently demonstrate that shear stress can be a pulmonary vasodilator stimulus, the overall implications will require further study. No attempt was made to provide a preparation that mimics in vivo conditions. Instead, the objective was to establish experimental conditions that led to clear separation of pressure and flow effects on the vessels under observation within the intact organ while not requiring pressures so high that they would be excessively injurious to the isolated lung. The left lower lobe is ~22% of the total ferret lung mass. Therefore, for the whole lung, the 10 and 40 ml/min flows would have been ~25 and 100 ml · min¹ · kg body wt¹, respectively. Measurements of cardiac output of anesthetized ferrets have ranged from ~168 (16) to 210 ml · min¹ · kg body wt¹ (29). Thus even the 40 ml/min flow would not be considered a high flow for the left lower lung lobe of this species. Raj et al. (29) measured an in vivo pulmonary arterial-venous pressure difference of ~9 Torr with the cardiac output of 210 ml · min¹ · kg body wt¹. When the lungs were isolated and perfused in that study, the 9-Torr pressure drop (Pa  Pv) was achieved with a flow rate of only ~48 ml · min¹ · kg body wt¹. In the present study, with an equivalent flow rate of ~100 ml · min¹ · kg body wt¹, the Pa  Pv was ~10 Torr.

There is evidence suggesting that shear stress-dependent vasodilator mechanisms do in fact play a role in normal pulmonary vascular physiology. For example, increased pulmonary blood flow has been associated with increased prostacyclin production in lungs (9, 31). Shear stress-induced prostacyclin production has been observed in pulmonary arterial endothelial cells in culture (22). Relaxation of the large pulmonary arteries in response to increased flow has been observed (10), and a lower pulmonary vascular resistance during pulsatile rather than steady flow has been attributed to flow-induced release of endothelium-derived relaxing factor (11). During exercise when the pulmonary vascular resistance and impedance are normally low, the pulmonary vessels are also subjected to vasoconstrictor stimuli (14), and it is conceivable that these stimuli are countered by locally mediated shear stress-sensitive vasodilator mechanisms. Kane et al. (13) found that combined -blockade and nitric oxide (NO) synthase (NOS) inhibition unmasked -adrenergic vasoconstriction in exercising sheep. During heavy exercise, the pulmonary arterial PO₂ can fall to levels that, according to the relationship established by Marshall et al. (23) for computing the hypoxic vasoconstrictor response as a function of both alveolar and mixed venous PO₂, should cause vasoconstriction. It is conceivable, therefore, that the effects of increased sympathetic activity and low pulmonary arterial PO₂ are normally superseded by increased shear stress-activated vasodilation resulting from increased cardiac output and, in some species, from increased blood viscosity. Interestingly, NOS inhibition did not alter the pulmonary vascular pressure-flow relationship in conscious resting dogs (26), but it resulted in higher resistance at a given cardiac output in exercising sheep (13). This difference would be consistent with the present observations, assuming that there is little vascular tone under resting conditions but that there is an increase in vasoconstrictor stimuli during exercise. The concept that shear stress-induced vasodilation might play a role during exercise does not imply a net increase or decrease in pulmonary vascular tone in exercise. Rather, the net result might be the absence of an increase.

Recently, in a study of normal subjects and subjects susceptible to high-altitude pulmonary edema, Eldridge et al. (8) concluded that the higher pulmonary vascular pressures in susceptible subjects, during exercise and during exercise and hypoxia, were consistent with augmented flow-dependent pulmonary vasoconstriction and/or a reduced vascular cross-sectional area in the susceptible subjects. Based on the present results, a possibility that might be considered is that the susceptible subjects had a diminished vasodilator response to increased shear stress resulting from increased flow and narrowing of vessels due to hypoxic vasoconstriction. Whether or not this is the case, the concept implies a possible role for disruption of shear dependent pulmonary vasodilator mechanisms in some forms of pulmonary hypertension.

A possible mechanism, consistent with the L-NAME result of the present study, is that endothelial NOS was more active at the higher flow. Prostacyclin and NO are pulmonary vasodilators (4), the production of which by vascular endothelium has been found to increase in response to increased shear stress (5, 9, 17, 19, 27, 30-32). Thus the effects of L-NAME are consistent with such observations. The results with a single inhibitor of NOS are not conclusive with regard to mechanism, and there are other potential explanations for the effects of arginine analogs (18). However, there are other observations consistent with a role for NO, including the impact of NOS inhibition during exercise (13) indicated above. The amount of exhaled NO has been found to increase during exercise, which might be considered consistent with a role for NO, but it is not clear what, if any, portion of the increase actually comes from the pulmonary endothelium because other sources can dominate the amount of exhaled NO in vivo (28). Thus the possibility that NO is involved deserves further study.

Because, initially, we were not certain that the arterial inflow pressure would be a sufficiently accurate reflection of the pressure in the arteries with the diameters that were to be measured, the strategy was to directly measure the pressure in vessels in the field of view by using the small catheter. The concept was that, although introducing the catheter might alter the fraction of the total flow perfusing the vessel of interest, a change in total flow would lead to a proportionate change in the flow through the observed vessel. Thus one would accept the possibility that local flow might be lower than it would be without the catheter to achieve the objective of knowing the local pressure. After we accumulated data from several preparations, it became clear that the pressure drop from arterial inlet to the catheter tip located in arteries in the field of view was not detectable even at the higher flow rate. This is what one would predict from the morphometry of the pulmonary arterial trees from other species (12), and it appears that the present study (Fig. 1) confirms this for the ferret pulmonary arterial tree.

Small-catheter measurements of vascular pressures have not always been consistent with the morphometric predictions of the longitudinal distribution of pulmonary Pa. This is presumably, at least in part, because if local flow is obstructed by an end-hole catheter, such as used in the present study, the pressure at the tip will be lower than it would have been in the absence of the catheter; i.e., the catheter-tip pressure will approach the wedge pressure as the ratio of vessel diameter to catheter diameter decreases (25). The catheters used in the present study were apparently small enough that the pressure drop between arterial inlet and the field of view that would have existed in the absence of the catheter was not substantially overestimated as the result of the catheter (Fig. 2). Because the pressure drop upstream from the vessels of interest was so small, we also carried out experiments with the small-catheter tip located at the arterial inlet so that the possibility that the catheter influenced local events was eliminated. The fact that the results with and without the catheter in the field of view were indistinguishable appears to rule out the possibility that local effects of the catheter had an important impact on the process under investigation. In the typical pressure-flow experiment with constant Pv, the intravascular pressures increase with increasing flow. In the protocol used in this study, the intravascular pressures decreased when the flow was increased. Thus the potential error in the estimated pressure, due to a pressure drop between the catheter tip and a downstream vessel within the field of view, would be in the direction of overestimation of the vessel pressure at the higher flow. Because the vessels had larger diameters at higher flow, if the local pressure had been over estimated, the general interpretation of the results would not be compromised. Instead, there would be a tendency toward underestimation of the magnitude of the dilation attributable to increased flow.

At a common flow in protocols I and II, the average of the diameters of the arteries studied was about the same during hypoxia and normoxia. Comparison of Figs. 5 and 6 leads to the conclusion that the higher pressure resulting from constriction of small vessels and a less-intense response to hypoxia in the larger vessels resulted in distension of the larger vessels. A similar observation has been made in dog lungs (2). Thus, while the inverse relationship between the fractional dilation and vessel diameter may involve a more effective shear stress response in smaller vessels, it appears that, in addition or instead, it reflects the greater hypoxic response in smaller vessels.

In conclusion, under normoxic conditions, changing flow had little effect on the arterial diameters, but under hypoxic conditions the arteries were consistently larger at the higher flow than at the lower flow, even though the distending pressure was the same at both flow rates. Thus the results are consistent with the hypothesis that shear stress is a pulmonary vasodilator stimulus.

ACKNOWLEDGEMENTS

This work was supported by National Heart, Lung, and Blood Institute Grant HL-19298 and by the Department of Veterans Affairs.

FOOTNOTES

Address for reprint requests: C. A. Dawson, Zablocki Veterans Affairs Medical Center, Research Service 151, 5000 W. National Ave., Milwaukee, WI 53295 (E-mail: dawsonc{at}vms.csd.mu.edu).

Received 17 December 1996; accepted in final form 21 April 1997.

REFERENCES

1.	Al-Tinawi, A., A. V. Clough, D. R. Harder, J. H. Linehan, D. A. Rickaby, and C. A. Dawson. Distensibility of small veins of the dog lung. J. Appl. Physiol. 73: 2158-2165, 1992[Abstract/Free Full Text].
2.	Al-Tinawi, A., G. S. Krenz, D. A. Rickaby, J. H. Linehan, and C. A. Dawson. Influence of hypoxia and serotonin on small pulmonary vessels. J. Appl. Physiol. 76: 56-64, 1994[Abstract/Free Full Text].
3.	Al-Tinawi, A., J. A. Madden, C. A. Dawson, J. H. Linehan, D. R. Harder, and D. A. Rickaby. Distensibility of small arteries of the dog lung. J. Appl. Physiol. 71: 1714-1722, 1991[Abstract/Free Full Text].
4.	Barnard, J. W., P. S. Wilson, T. M. Moore, W. J. Thompson, and A. E. Taylor. Effect of nitric oxide and cyclooxygenase products on vascular resistance in dog and rat lungs. J. Appl. Physiol. 74: 2940-2948, 1993[Abstract/Free Full Text].
5.	Berthiaume, F., and J. A. Frangos. Flow-induced prostacylin production is mediated by a pertussis toxin-sensitive G protein. FEBS Lett. 308: 277-279, 1992[Medline].
6.	Bevan, J. A., and I. Laher. Pressure- and flow-dependent vascular tone. FASEB J. 5: 2267-2273, 1991[Abstract].
7.	Clough, A. V., G. S. Krenz, M. Owens, A. Al-Tinawi, C. A. Dawson, and J. H. Linehan. An algorithm for X-ray angiographic estimation of blood vessel diameter (Abstract). FASEB J. 5: A529, 1991.
8.	Eldridge, M. W., A. Podolsky, R. S. Richardson, D. H. Johnson, D. R. Knight, E. C. Johnson, S. R. Hopkins, H. Michimata, B. Grassi, J. Feiner, S. S. Kurdak, P. E. Bickler, P. D. Wagner, and J. W. Severinghaus. Pulmonary hemodynamic response to exercise in subjects with prior high-altitude pulmonary edema. J. Appl. Physiol. 81: 911-921, 1996[Abstract/Free Full Text].
9.	Ellsworth, M. L., T. J. Gregory, and J. C. Newell. Pulmonary prostacyclin production with increased flow and sympathetic stimulation. J. Appl. Physiol. 55: 1225-1231, 1983[Abstract/Free Full Text].
10.	Grant, B. J. B., and S. M. Sherif. Flow-mediated vasodilatation of the main pulmonary artery. Respir. Physiol. 86: 77-90, 1991[Medline].
11.	Hakim, T. S. Flow-induced release of EDRF in the pulmonary vasculature: site of release and action. Am. J. Physiol. 267 (Heart Circ. Physiol. 36): H363-H369, 1994[Abstract/Free Full Text].
12.	Haworth, S. T., J. H. Linehan, T. A. Bronikowski, and C. A. Dawson. A hemodynamic model representation of the dog lung. J. Appl. Physiol. 70: 15-26, 1991[Abstract/Free Full Text].
13.	Kane, D. W., T. Tesauro, T. Koizumi, R. Gupta, and J. H. Newman. Exercise-induced pulmonary vasoconstriction during combined blockade of nitric oxide synthase and beta adrenergic receptors. J. Clin. Invest. 93: 677-683, 1994.
14.	Kane, D. W., T. Tesauro, and J. H. Newman. Adrenergic modulation of the pulmonary circulation during strenuous exercise in sheep. Am. Rev. Respir. Dis. 147: 1233-1238, 1993[Medline].
15.	Kempf, J. E., and H. T. Chang. The cardiac output and circulation time of ferrets. Proc. Soc. Exp. Biol. Med. 72: 711-714, 1949.
16.	Khayutin, V. M., A. M. Melkumyants, A. N. Rogoza, E. S. Veselova, S. A. Balashov, and V. P. Nikolsky. Flow induced control of arterial lumen. Acta Physiol. Hung. 68: 241-251, 1986[Medline].
17.	Koller, A., S. Dong, A. Huang, and G. Kaley. Corelease of nitric oxide and prostaglandins mediates flow-dependent dilation of rat gracilis muscle arterioles. Am. J. Physiol. 267 (Heart Circ. Physiol. 36): H326-H332, 1994[Abstract/Free Full Text].
18.	Kontos, H. A., and E. P. Wei. Arginine analogues inhibit responses mediated by ATP-sensitive K+ channels. Am. J. Physiol. 271 (Heart Circ. Physiol. 40): H1498-H1506, 1996[Abstract/Free Full Text].
19.	Kuchan, M. J., and J. A. Frangos. Role of calcium and calmodulin in flow-induced nitric oxide production in endothelial cells. Am. J. Physiol. 266 (Cell Physiol. 35): C628-C636, 1994[Abstract/Free Full Text].
20.	Kuo, L., W. M. Chilian, and M. J. Davis. Interaction of pressure- and flow-induced responses in porcine coronary resistance vessels. Am. J. Physiol. 261 (Heart Circ. Physiol. 30): H1706-H1715, 1991[Abstract/Free Full Text].
21.	Madden, J. A., A. Al-Tinawi, P. A. Keller, and C. A. Dawson. Intrinsic tone and distensibility of in vitro and in situ cat pulmonary arteries. Lung 174: 291-301, 1996[Medline].
22.	Marsh, J. A., K. A. Prichard, E. R. Jacobs, and E. K. Birks. Shear-induced prostacyclin production in cultured bovine pulmonary artery endothelial cells is attenuated by incubation with native LDL. Am. J. Respir. Crit. Care Med. 151: A630, 1995.
23.	Marshall, B. E., W. R. Clarke, A. T. Costarino, L. Chen, F. Miller, and C. Marshall. The dose-response relationship for hypoxic pulmonary vasoconstriction. Respir. Physiol. 96: 231-247, 1994[Medline].
24.	Melkumyants, A. M., S. A. Balashov, and V. M. Khayutin. Endothelium dependent control of arterial diameter by blood viscosity. Cardiovasc. Res. 23: 741-747, 1989[Medline].
25.	Nicolaysen, G., J. Shepard, T. Tanita, M. Onizuka, C. Hall, and N. C. Staub. Effect of catheter size on pressures recorded In small pulmonary veins in dog lung. J. Appl. Physiol. 71: 1296-1301, 1991[Abstract/Free Full Text].
26.	Nishiwaki, K., D. P. Nyhan, P. Rock, P. M. Desai, W. P. Peterson, C. G. Pribble, and P. A. Murray. N-nitro-L-arginine and pulmonary vascular pressure-flow relationship in conscious dogs. Am. J. Physiol. 262 (Heart Circ. Physiol. 31): H1331-H1337, 1992[Abstract/Free Full Text].
27.	Norris, M., M. Morigi, R. Donadelli, S. Aiello, M. Foppolo, M. Todeschini, S. Orisio, G. Remuzzi, and A. Remuzzi. Nitric oxide synthesis by cultured endothelial cells is modulated by flow conditions. Circ. Res. 76: 536-543, 1995[Abstract/Free Full Text].
28.	Persson, M. G., N. P. Wiklund, and L. E. Gustafsson. Endogenous nitric oxide in single exhalations and the change during exercise. Am. Rev. Respir. Dis. 148: 1210-1214, 1993[Medline].
29.	Raj, J. U., P. Kappa, R. Hillyard, and J. Anderson. Pulmonary vascular pressure profile in adult ferrets: measurements in vivo and in isolated lungs. Acta Physiol. Scand. 142: 41-48, 1991[Medline].
30.	Rubanyi, G. M., A. D. Freay, K. Kauser, A. Johns, and D. R. Harder. Mechanoreception by the endothelium: mediators and mechanisms of pressure and flow-induced vascular responses. Blood Vessels 27: 246-257, 1990[Medline].
31.	Rubanyi, G. M., J. C. Romero, and P. M. Vanhoutte. Flow-induced release of endothelium-derived relaxing factor. Am. J. Physiol. 250 (Heart Circ. Physiol. 19): H1145-H1149, 1986[Abstract/Free Full Text].
32.	VanGrondelle, A., G. S. Worthen, D. Ellis, M. M. Mathias, R. C. Murphy, R. J. Strife, J. T. Reeves, and N. F. Voelkel. Altering hydrodynamic variables influences PGI2 production by isolated lungs and endothelial cells. J. Appl. Physiol. 57: 388-395, 1984[Abstract/Free Full Text].