- Email: [email protected]
Hypoxic vasoconstriction in blood and plasma perfused lungs
109
Respiration Physiology (1988) 72, 109-122 Elsevier
RSP 01384
Hypoxic vasoconstriction in blood and plasma perfused lungs T. S. Hakim and A.B. Malik Department of Physiology, McGill University, Montreal, Quebec, Canada and Albany Medical College, Albany, NY, U.S.A. (Accepted for publication 2 November 1987) Abstract. Isolated lungs of pigs, cats and rats were perfused in situ with blood and with plasma at constant flow rate. The experiments were specifically designed to compare the intensity of the hypoxic pressor response (HPR) during blood and plasma perfusion. Ventilation of the lungs with 6.0~o 02 during perfusion with blood, elevated the inflow pressure (Pa) from 18.9 + 1.5 to 35.8 +_2.0 mm Hg in pig lungs, from 12.9 + 0.5 to 17.9 + 1.1 mm Hg in rat lungs, and from 12.6 + 0.8 to 19.2 + 1.5 mm Hg in cat lungs. In comparison during perfusion of the lungs with plasma, the HPR was larger in pig lungs (Pa increased from 16.9 + 2.0 to 42.3 + 2.7 mm Hg), smaller in rat lungs (Pa increased from 10.2 _+ 0.9 to 11.4 + 1.2 mm Hg), and also smaller in cat lung (Pa increased from 9.9 + 1.2 to 11.9 + 1.2 mm Hg). The site of HPR in blood perfused cat lobes was primarily in the middle segment (arterial and venous technique) and remained in the middle segment (although blunted) during plasma perfusion. The vascular pressure-flow relationship in pig lungs showed that during blood perfusion, both the slope and intercept rose during hypoxia; these changes were similar but greater during plasma perfusion. Measurement of thromboxane A 2 and prostacyclin in the lung effluent revealed no relation to intensity of the HPR or to the difference in the HPR in blood and plasma perfused lungs. Thus the absence of red blood cells (RBCs) may have been responsible for the difference in the HPR during blood and plasma perfusion; in rats and cats, RBCs appear to be essential for the full expression of the HPR, in contrast, the HPR in pigs does not require the presence of the RBCs. Possible explanations for these differences are suggested.
Cat; Hypoxia; Pig; Prostaglandins; Rat; Red cell deformability
Although a hypoxic pressor response (HPR) can be elicited in cell-free perfused lungs from rats (Berkov, 1974; McMurtry et al., 1978; Kivity and Souhrada, 1981; Lingren et al., 1985), cats (Berkofsky and Holtzman, 1967; Said et al., 1974) and dogs (Gorsky and Lloyd, 1967), it is not certain whether the intensity, mechanisms and site of HPR in blood and in cell-free perfused lungs are similar. The HPR is a very labile response (Voelkel, 1986), and varies considerably among individuals and species (Tucker et al., Correspondence address: T. S. Hakim, Department of Physiology, McGill University, 3655 Drummond, Montreal, Quebec, Canada, H3G 1Y6. 0034-5687/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)
110
T.S. HAKIM AND A.B. MALIK
1975), strains (Ou and Smith, 1983), geographical regions (Peake et al., 1981; Bishop and Cheney, 1983), and the number ofhypoxic challenges to which the lung was exposed (McMurtry et al., 1978). The lability in the response makes comparisons of different results difficult even in blood-perfused lungs. In the absence of blood cells, HPR apparently becomes either not present (Hauge, 1968; Berkov, 1974), reduced (Gorsky and Lloyd, 1967) or cannot be maintained (McMurtry et al., 1978) unless plasma or a vasoconstrictor agent is added to the perfusate. Our intention in this study was to compare the HPR in lungs from three animal species, during perfusion with whole autologous blood and then during perfusion with plasma. Each lung served as its own control using the same mixture of normoxic and hypoxic gases and, to avoid the individual variability in HPR, each lung was perfused with blood, and with plasma. HPR was tested only once during each period of perfusion. The results show that HPR is diminished in rat and cat lungs when blood cells are absent from the perfusate but in contrast it becomes enhanced in pig lungs. The differences in HPR are discussed in terms of the site of the HPR. Moreover, the differences were not related to vasoconstrictor and vasodilator prostanoids released during blood or plasma perfusion.
Materials and methods
The animals used in this study were eight male Sprague-Dawley rats (311 + 7 g), nine miniature pigs of either sex (11.1 + 0.3 kg) and eight cats of either sex (3.6 + 0.1 kg). The rats and cats were anaesthetized with an intraperitoneal injection of sodium pentobarbital (25 mg/kg), while the pigs were premedicated with an intramuscular injection of ketamine (5 mg/kg) and atropine (2 mg/kg) followed (15 min later) by an intravenous injection of sodium pentobarbital, given slowly to induce anaesthesia (up to 25 mg/kg). A catheter was placed in the carotid artery and the animals were heparinized (700 units/kg). They were ventilated mechanically and exsanguinated through the carotid artery. Midway during the bleeding, 10-15 ml/kg body weight of plasma expander was injected into the animal. The plasma expander consisted of 6~o dextran 40 (Pharmacia) and Ringer's lactate (Abbott) mixed in equal volumes. A fraction of the blood was used to prime the perfusion system, and another fraction was saved in a plastic container and kept at 37°C in a shaker bath. The remaining third fraction was mixed with equal volume of plasma expander and centrifuged at 1750 x g for 10 min. The supernatant plasma was collected and saved in a plastic beaker in the shaker bath to be used later for perfusion of the lungs. Isolated whole lung preparation. The procedure of preparing the whole lung for perfusion in situ was identical in rats and pigs. A rapid mid-sternal incision was performed to expose the lungs. The main pulmonary artery was isolated and an appropriate size cannula was placed in it via an incision in the right ventricle. Another cannula was placed in the left atrium via an incision in the atrial appendage. The cannulas were secured in position with sutures, and heavy ties were placed around the heart and aorta.
PULMONARY HYPOXIC VASOCONSTRICTION
111
The pulmonary arterial cannula and left atrial cannula were connected to the primed perfusion system. A bypass connection was included between the arterial and venous cannulas. The perfusion system consisted of a reservoir open to room air, a roller pump (Masterflex 500) and a heat exchanger (39°C). The blood was pumped from the reservoir via the heat exchanger, into the pulmonary artery and the blood was allowed to drain back passively into the reservoir. During this time the lungs were ventilated with a gas mixture containing 5.5~ CO2, 35% 02 and nitrogen. End inspiratory pressure was kept below 10 mm Hg and end expiratory pressure was 1-2 mm Hg. The volume of blood in the perfusion system was approximately 20 ml in the rat lung preparations and about 400 ml in the pig lung preparations. Donor rats had to be bled to obtain adequate blood and plasma. The blood perfusion rate was increased gradually while keeping the pulmonary arterial pressure below 20 mm Hg. The rat lungs usually stabilized within 15 min while the pig lungs took up to 60 min to become stable. The pulmonary arterial pressure (Pa) and left atrial pressure were measured from side ports in the cannulas, and referenced to the top of the lungs. The left atrial pressure was set between 0 and 1.0 mm Hg and was kept constant in each experiment. During the period of stabilization, blood gases and pH were normalized by changing ventilation and by adding sodium bicarbonate into the reservoir. Once the Pa became stable (flow rate in pig and rat lungs was 485 + 32 and 10.4 + 0.4 ml/min respectively), the experimental protocol was initiated. Effluent blood samples were withdrawn for gases, pH, hematocrit (Hct) and prostaglandin (thromboxane and prostacyclin) analysis. The lung was ventilated with an isocapnic hypoxic gas mixture (5.5~o CO2, 6 ~ 02 and nitrogen) for a duration of 10 min while keeping flow rate constant. Another effluent hypoxic blood sample was withdrawn for gases, Hct, and PG analysis. Ventilation was returned to normoxic gas until Pa and blood gases returned to baseline values. During the normoxic and hypoxic periods in the pig lung, the Pa was measured at 3 or 4 different flow rates in random order, from which a pressure-flow relationship (slope and pressure intercept) was derived. The bypass connector was opened and two clamps were placed near the cannulas. The blood was removed out of the perfusion system rapidly and the system was washed with saline and refilled with plasma perfusate. It was circulated in the tubing for 2 min, and the perfusion was reestablished by removing the two clamps on the cannulas and closing the bypass. The initial bloody effluent from the lung was discarded. The perfusion rate was increased gradually until the lung reached a new steady state as suggested by a stable Pa. The flow rates were equal to those used during blood perfusion. During the time, the perfusion was switched from blood to plasma (3-5 min) the lungs continued to be ventilated with the normoxic gas mixture. The pH in the perfusate was adjusted by adding bicarbonate whenever necessary. Once stability was reestablished, a plasma sample was withdrawn for gases, Hct and PG analysis. The Hct was always less than 1% and could not be read in the usual manner as the blood. Ventilation was then switched to the hypoxic gas mixture for 10 min and again plasma samples were withdrawn for analysis as before. Ventilation with normoxic gas was resumed until Pa and gases returned to baseline level. The plasma was rapidly emptied out of the system
112
T.S. HAKIM AND A.B. MALIK
and replaced with fresh blood without allowing air bubbles to enter the lung. After reaching a steady state with blood perfusion using the same flow rates as before, blood samples were withdrawn during normoxic and hypoxic periods for measuring Hct, gases and pH, but P G analysis was not carried out during the second blood perfusion. The pressure-flow relationships in pig lungs during normoxic and hypoxic conditions were repeated during plasma perfusion and they were repeated in two lungs during the second period of blood perfusion.
Isolated left lower lobe (LLL) preparation in cats. The experimental protocol in the cat was identical to the one for rats and pigs, except that instead of perfusing the whole lung, only the left lower lobe was perfused in situ as described in detail previously for dog lungs (Hakim et al., 1983). Briefly, after the cats were exsanguinated, the LLL was exposed through a left thoracotomy. The main artery and vein of the LLL were isolated and cannulated. The cannulas were connected to a perfusion system as above. The volume of blood in the cat LLL perfusion system was about 30 ml. The LLL was initially perfused with autologous blood (flow rate = 49.7 + 2.2 ml/min), then with plasma, and again with fresh blood at the same flow rate. Replacement of the blood or plasma was accomplished as described above in the rat and pig lungs. One additional measurement was made in cat LLL, namely that during each steady state of normoxia and hypoxia, arterial and venous occlusion techniques were applied to determine if there was any difference in the site of H P R during blood and plasma perfusion. Cat LLL usually stabilized as rapidly as rat lungs. As before, PG samples were analyzed for the normoxic and hypoxic period during the first blood perfusion and during plasma perfusion only, but not during the second blood perfusion. As before only one H P R was tested during each period of perfusion. Preparation of the plasma sample for prostaglandin analysis. Effluent blood or plasma sample (2 ml) which was withdrawn for P G analysis was treated by adding 0.5 ml of 2.5 mg/ml indomethacin dissolved in distilled water, and centrifuged at 1750 × g for 10 min. The clear supernatant was stored in polypropylene vials, sealed and kept frozen at -70 °C until they were analyzed. Analysis for thromboxane B 2 (TxBz) and 6-ketoprostaglandin F j ~ (6-keto-PGF 1~) (stable metabolites of TxA 2 and prostacyclin) were carried out using a radioimmunoassay as described previously (Selig et al., 1986). Organic extraction to remove possible interfering proteins was not required, since plasma standards were included in this assay system. The cross reactivity of the 6-keto-PGFl~ antiserum was as follows: 3.3~o PGFI~: 0.1~o prostaglandin D 2 (PGD2); 0.24~o PGE2; 0.27~o PGF2~; 0.01~o with TxB2, AA, or linoleic acid. The cross reactivity of the TxB 2 antiserum was as follows: 0.06~o P G D 2 and 0.01 ~o for P G E e, PGF2~, PGF1~, 6-keto-PGFl~, AA, or linoleic acid. A linear working range for both TxB 2 and 6-keto P G F ~ antisera was established between 20 and 10000 pg for pig and cat samples, and between 20 to 2000 pg for rat samples. Statistical analysis of the results was done using a Student t-test and a P value of less than 0.05 was considered significant. Analysis of variance test was used to confirm the
PULMONARY HYPOXIC VASOCONSTRICTION
113
differences in the hypoxic pressor responses a m o n g the three animal species. All values are given as m e a n s + SE except in tables 1 a n d 2 in which _ SD are given instead.
Results A complete protocol consisted of one hypoxic challenge (10 min) and a return to baseline during each of the three perfusion periods; blood perfusion, plasma perfusion, a n d again blood perfusion. The total duration of each experiment was 2 - 3 h and usually depended on the time to reach a stable state. Because of the lack of adequate blood volume in some experiments, H P R was tested once with blood and once with p l a s m a in r a n d o m order. Table 1 summarizes the m e a n values of hematocrit (Hct), gases, and p H in the blood and plasma during the normoxic (N) and hypoxic (H) periods. Although the data during the second blood perfusion period are not shown in the table, the gases were similar to the first period, but the Hct were slightly lower (80-90~o) than during the first perfusion period.
TABLE 1 Hematocrits (Hct) and perfusate gases during normoxic (N) and hypoxic (H) conditions. Values are ~_+SD. Blood Hct ~ Pigs
Plasma pH
Pco2
Po2
pH
Pco~
Po2
7.37 + 0.05
43.3 _+4.9
201.8 +_23.4
7.37 + 0.03
44.4 + 6.3
198.2 + 30.7
H
7.40 +0.01
40.9 +2.5
46.2 +2.7
7.35 +0.05
46.1 _+2.9
54.1 _+11.1
N
7.38 +0.07
34.5 +2.6
169.1 +7.6
7.48 +0.05
35.3 +2.1
158.6 + 10.1
H
7.35 + 0.09
37.7 + 2.7
47.4 + 4.0
7.40 _+0.05
39.9 +_3.9
60.5 + 4.6
N
7.41 +0.03
33.6 +_1.4
181.3 +6.1
7.43 +0.09
35.5 +2.1
177.1 +7.8
7.43 + 0.04
33.0 _+1.6
40.5 + 3.7
7.43 _+0.08
36.4 + 1.6
53.7 _+3.7
N 26.6 + 1.6
Rats
22.4 +3.7
Cats
30.0 + 3.6 H
114
T.S. HAKIM AND A.B. MALIK
7_+SE
A
1- 40
/ /
E
g
Pigs
(lO)
~
3o (~..I
•~- 20 .......
< ~¢-
I""
¢...,,.,'~ o'1
i.~.!....i
Rats
...........~
Cats
~ ........... ~61
lO
0
E 5 o
I
I
N
H Blood
I
N Plasma
I
H
I
I
N
H Blood
Fig. 1. Pulmonary arterial pressure during ventilation with 35% 02 (N) and with 6% 0 2 (H) in isolated lungs of three animal species. The lungs were perfused with blood, then with plasma and again with blood. The number in brackets is the number of measurements in each group.
HPR in blood vsplasma-perfused lungs. In 12 out of 15 hypoxic challenges in pig lungs perfused with blood, the increase in Pa ranged between 17 to 24 mm Hg, whereas in 3 lungs Pa increased by 8 mm Hg. With plasma perfusion, the increase in Pa ranged between 20 and 37 in 9 out of 10 measurements and was 19 in one lung. The mean Pa rose from 18.9 + 1.5 to 35.8 + 2.0 and from 21.3 + 1.8 to 37.5 + 3.1 mm Hg during the 1st and the 3rd perfusion period (with blood) respectively, and from 16.9 + 2.0 to 42.3 + 2.7 mm Hg during perfusion with plasma (fig. 1). Thus, the H P R was unequivocally enhanced by about 10 mm Hg during plasma perfusion in spite of the higher Po2 (54.1 + 11.1 mm Hg) with plasma than with blood perfusion (46.2 + 2.7 mm Hg). The responses in the rat lungs were different. In 11 out of 12 measurements during blood perfusion, the increase in Pa in response to 6% 0 2 ranged from 3 to 9 mm Hg and in one lung there was only a 1 mm Hg rise in Pa. During plasma perfusion the increase in Pa was less than 2 mm Hg in 8 out of 9 measurements, and about 7 mm Hg in one lung. The mean Pa during blood perfusion rose from 12.9 + 0.5 to 17.9 + 1.1 mm Hg during the first period and from 15.8 + 1.7 to 20.9 + 2.1 mm Hg during the third period. With plasma perfusion, Pa rose from 10.2 + 0.9 to 11.4 + 1.2 mm Hg (fig. 1). Thus unlike the pig, the response in the rat to this level of hypoxia was attenuated in the absence of red blood cells. Note, however, that the Po2 in the plasma during hypoxia (60.5 + 4.8 mm Hg) was greater than the Po2 in the blood (47.4 + 4.0 mm Hg) even though the same hypoxic gas mixture was used in both cases. Whether this difference in Po~ is of physiological importance is not clear, in view of the evidence that blood gas Po2 is less important than alveolar gas Po2 in mediating H P R (Hauge, 1969).
PULMONARY HYPOXIC VASOCONSTRICTION
115
The response in the cat lung was similar to rat lung, but more variable. In 9 out o f 14 measurements, the increase in Pa ranged between 4 and 11 m m Hg, in 4 measurements the response was less than 4 m m Hg, and in 2 measurements it decreased by 0.5 m m Hg. With plasma perfusion the response usually ranged between 0 and 5 m m Hg. The mean Pa increased from 12.6 + 0.8 to 19.2 + 1.5 m m Hg during the first perfusion period, and from 12.3 + 1.6 to 17.3 + 2.0 m m H g during the third perfusion period. With plasma perfusion, the Pa rose from 9.9 + 1.2 to 11.9 + 1.2 m m H g (fig. 1). Thus, like the rats, but unlike the pigs, the H P R in cat was attenuated when blood cells were not present in the perfusate. In this case, the Po2 during hypoxia was also slightly higher in the plasma (54.1 + 11.1 m m Hg) than in the blood (46.2 + 2.7 m m Hg) in spite of using the same hypoxic gas mixture for both challenges.
Site of HPR during blood and plasma perfusion. To examine if the site of action is altered by the absence of the red blood cells from the perfusate, we performed the arterial and venous occlusion during normoxia and hypoxia in cat L L L perfused with blood and then with plasma. The results (table 2, fig. 2) revealed that during blood perfusion the H P R occurred primarily in the middle segment (P < 0.05), but was attenuated during plasma perfusion. In pig lung, we examined the site of action of hypoxia by studying the pressure-flow relationship. This allowed us to see if the effect on the slope and on the pressure intercept (critical closing pressure) is similar in blood and plasma perfused lungs. We succeeded in obtaining a complete set of pressure-flow data on 5 lungs during normoxia and hypoxia while perfusing with blood and with plasma (fig. 3). In two lungs we were able to repeat these measurements during the second blood perfusion period. These two showed changes similar to the changes observed during the first perfusion period. We were somewhat conservative in the number of data points, because on the basis of preliminary experiments we found that a combination of high flow rate, low hematocrit and hypoxia, there was a great tendency for the lungs to become edematous. During normoxia the mean slope and intercept during blood perfusion were TABLE 2 Segmental pressure gradients (mm Hg) in cat left lower lobes (n = 6) during normoxia and hypoxia, when perfused with blood, plasma and again with blood. Values are ~ + SE. Total
Arterial
Middle
Venous
Blood
Normoxia Hypoxia
9.8 + 0.8 16.2 + 2.2*
2.2 ± 0.6 3.6 +_0.7
3.6 + 0.5 8.7 ± 1.9"
3.5 +_0.2 3.8 + 0.4
Plasma
Normoxia Hypoxia
6.7 + 0.8 8.7 + 1.2
2.0 + 0.4 1.9 + 0.4
1.9 + 0.6 3.5 ± 0.9
2.9 + 0.2 3.4 + 0.3
Blood
Normoxia Hypoxia
8.5 + 0.9 12.7 + 2.2*
2.4 + 0.4 2.3 + 0.5
2.8 + 0.5 6.5 + 1.5"
3.3 + 0.4 3.9 _+0.6
* Significantly different from the value immediately above it.
Ot
116
T.S. HAKIM AND A.B. MAL1K n=6 x+SE
~rterial L.
lO
/ ~iddle
5 Cenous N
H
N
Blood
H
N
Plasma
H
Blood
Fig. 2. Distribution of vascular pressure gradients in cat left lower lobes during perfusion with blood and with plasma in normoxic (N) and hypoxic (H) conditions. Flow rate was kept constant throughout each experiment. The means and standard errors of the individual segments as well as their statistical significance are given in table 2.
25.7 + 4.6 m m H g . L - ~. min and 7.5 + 2.2 m m Hg, respectively, but both increased (P < 0.05) during hypoxia to 40.2 + 10.7 m m H g . L - ~ • min and 13.2 + 3.4 m m Hg, respectively. With plasma perfusion, the slope and intercept were 18.1 + 2.9 m m H g - L - ~ • min and 4.3 + 1.4 m m H g respectively, and both increased
/
Blood
//
Plasma
/
/
/ !
5O
/
-1-
E E
/
/
i/ /
40
/
z11 "II .~/11. /
/I/1
II
,,
g II
iI
~ if/
/ ~
09 tO
13_ 30
<
20
~
¢-
o
_E lO
~
-
+
Slope Normoxia Hypoxia I 0.2
25.7 -+ 4.6 40.2 _+ 10.7 I I 0.4 0.6
S
E Intercept 7.5 _+ 2.2
13.2 -+ 3.4 I I 0.8 1.0
Flow Rate
Normoxia
0
18.1 -+ 2.9
Hypoxia 40.1 -+ 7.5 I I I 0.4 0.6 0.2
4.3 +- 1.4 20.9 _+ 2.3 I I 0.8 1.0
(I/min)
Fig. 3. The pulmonary arterial pressure-flow rate relationship during normoxic (solid line) and hypoxic (broken line) condition in blood and plasma perfused pig lungs.
117
PULMONARY HYPOXIC V A S O C O N S T R I C T I O N TABLE 3
Effluent levels (ng/ml) of thromboxane B z (TxB2) and 6-ketoprostaglandin-Fl~ (6-keto-PGFl~) during normoxia (N) and hypoxia (H) in blood and plasma perfused lungs. Values are ~ + SD. Blood perfusion
Plasma perfusion
TxB2
6-keto-PGFl~
TxB2
6-keto-PGF~
Pigs (n = 4)
N H
1.8 + 1.9 1.3 + 0.8
1.4 + 0.8 1.0 + 0.7
1.9 + 0.3 1.2 + 1.3
1.6 _+ 0.6 1.8 _+ 0.7
Rats (n = 8)
N H
9.7 + 4.5 6.2 + 3.8
3.4 + 1,1 3.1 + 0.9
8.0 + 6.4 8.3 + 4.4
3.9 + 2.4 4.0 + 2.4
Cats (n = 6)
N H
0.8 + 0.5 0.5 + 0.2
1.8 + 1,0 1.8 +_ 1.1
0.5 + 0.2 0.5 + 0.2
1.4 + 0.9 2.1 + 1.5
(P < 0.05) during hypoxia to 40.1 + 7.5 mm Hg. L - ~ • rain and 20.9 + 2.3 mm Hg. Thus the increase in both the slope and the intercept were enhanced during plasma perfusion. These data indicate that the sites of H P R were similar but potentiated when the blood cells were not present. In an attempt to assess whether the release of mediators can account for the differences in the H P R among species, and more importantly between the blood- and plasma-perfused lungs, we measured the effluent perfusate concentration of TxB 2 and 6-keto-PGFl~: precursors of a potent pulmonary vasconstrictor (TxA2) and a pulmonary vasodilator prostaglandin (PGI2). The results are shown in table 3. As expected (Hwang, 1985) the levels of both mediators were different in the three animals. The concentrations of both mediators were not significantly different during blood and plasma perfusion, neither were the levels altered by changing from normoxic (35yo 02) to hypoxic (6% 02) conditions. These results suggest that TxA z and PGI 2 release could neither account for the differences in H P R in the three species, nor could it account for the differences in the H P R in blood vs plasma perfused lungs. Role ofprostaglandins in the HPR.
Discussion
Our main objective in this study was to determine the importance of red blood cells (RBCs) in the HPR. Although the H P R has been studied in lungs perfused with RBCs free perfusate, the difference in H P R in blood vs plasma perfused lungs has seldom been the focus of the study. In a recent preliminary study we found that RBCs became less flexible during hypoxia (Hakim and Macek, 1986) and postulated that RBCs may be involved during the HPR. Therefore, we set out to compare the H P R in lungs perfused
118
T.S. H A K I M A N D A.B. M A L I K
with whole blood and with cell free perfusate; we expected that because rat RBCs became relatively rigid during hypoxia (Hakim and Macek, 1986), they may contribute to the HPR by restricting flow in the capillaries (Braash, 1971). Thus, their removal from the perfusate would attenuate the HPR. In contrast, because pig RBCs were insensitive to hypoxia (Hakim and Macek, 1986), we expected that their removal from the perfusate would not alter the HPR. We also included the cat as a third test species. Because the HPR is a very labile response, the study was designed to minimize as much as possible the differences which may contribute to alterations in the hypoxic presser response (HPR) of the lung vasculature. The protocols and procedures were identical in all experiments, and the total duration of the experiment was kept between 2 and 3 h. Each lung served as its own control and to eliminate the time factor, the HPR was tested in most lungs before and after plasma perfusion. Each lung exposure to hypoxia was repeated three times with the same gas mixture, once during each perfusion period, with each time exposure lasting for 10 min. We avoided repeated challenges as this may alter the intensity of the HPR due to depletion or accumulation of metabolites (McMurtry et al., 1978). To reduce the effects of Pco~ and pH (Dawson, 1984), they were kept within a very narrow range during blood and plasma perfusion (table 1). The temperature of the blood and plasma was kept near 37 ° C. Thus, these experiments were designed not only to compare the response in different animals but also to compare the HPR in blood vs plasma perfused lungs. Our pig and cat lungs were as responsive as previously reported by other investigators (Dawson et al., 1974; Peake et al., 1981). Similarly, the HPR in rat lungs to 6~o 02 were also comparable to those found by others (McMurtry etal., 1978; Kivity and Souhrada, 1981; Walker etal., 1982). The differences in the acute HPR among species apparently cannot be explained by the medial thickness of smooth muscle in the muscular vessels (Kay, 1983). It is interesting to note that the muscular arteries in the pigs are limited to vessels 25-70/~m in diameter, while in cats or rats the muscular vessels range 20-500 #m, and 20-350 #m in diameter, respectively. Whether such differences contribute to the observed intensities in HPR is not clear. The HPR during plasma perfusion suggests that the absence of the RBCs from the perfusate affects the HPR differently in animal species. In some animals (rats, cats) the HPR was blunted, while in others (pigs) it was enhanced when blood cells were absent from the perfusate. Although leukocytes and platelets were also absent from the perfusate (2~o) their role is apparently minimal (McMurtry et aL, 1978). It is difficult to offer a single mechanism to explain the differences in HPR in plasmaand blood-perfused lungs of all three species. A number of mechanisms may be involved to different extents. These mechanisms may include the direct action of hypoxia on smooth muscle, alteration in vascular resistance by RBCs via changes in their deformability or their interaction with the vascular lumen, and the release of mediators (vasoconstrictors or vasodilators) from lung tissue sources. The release of mediators has indeed been the subject of many investigations (Dawson, 1984; Voelkel, 1986) and our study does not resolve this controversial subject; however, the data show that differences in prostanoid release could neither explain the differences among species nor the variability among individuals within the species, nor could it explain the difference in HPR in
PULMONARY HYPOXIC VASOCONSTRICTION
119
blood- vs plasma-perfused lungs. Therefore in our discussion, we will focus primarily on the influence of hypoxia on smooth muscle tone and on RBC deformability, because hypoxia has been shown in vitro to constrict isolated vessels (Madden et al., 1985) and to alter RBC deformability (Hakim and Macek, 1986). In the pig, the role of blood cells seems to be minimal, since the HPR was equally large during plasma perfusion and blood perfusion. This is consistent with in vitro measurement on RBC from the pig, which showed that hypoxia causes minimal alteration in RBC deformability (Hakim and Macek, 1986). It is also possible that in the pig, the RBCs could have in fact attenuated the HPR by either impeding the release of vasoconstrictors from the endothelial cells or from the platelets, or by taking up substances e.g. adenosine or prostacyclin (McMurtry et al., 1978). Thus by excluding the role of RBCs and the release of prostanoids, the direct effect of hypoxia on smooth muscle tone seems to be a likely mechanism in pig lungs. As previously observed (Rock et al., 1985), hypoxia caused the slope of the pressure-flow curve and its intercept to increase in pig lungs. Our results on the pressure-flow data in pig lungs (fig. 3) showed that the HPR became potentiated at all flow rates during plasma perfusion and suggest that the site of action is in both segments which contribute to the slope and the critical closing pressure of the vasculature. The results from rat lungs suggest entirely different mechanisms for the HPR. In the present study, removal of blood cells blunted the HPR in all but one rat. Thus the direct influence of hypoxia on smooth muscle tone (as suggested by a small HPR in plasmaperfused lungs), appears to be of little importance. Therefore factors other than the direct action of hypoxia or prostanoid release must be involved. On the basis of recent measurement of RBC deformability which revealed that rat RBC became less deformable during hypoxia (Hakim and Macek, 1986) we propose that RBC deformability may have played a significant role in the HPR in the rats. In contrast with our finding, McMurtry et al. (1978) reported that the HPR was not different in a group of lungs perfused with blood compared to another group perfused with plasma. The reason for the difference between the two studies is not clear: they compared two different groups, using less than 2~o 02 whereas we repeated the HPR in the same lungs using 6~o 02. Furthermore, the isolated lungs in our study were not pretreated with any substances to enhance the response. It is possible that the level of 02 is of critical importance because changes in RBC deformability and constriction of the smooth muscle may occur at different threshold. The response of cat lungs was unlike the rat or the pig. The HPR with blood varied considerably in different cats, and the response was usually diminished with plasma perfusion and independent of prostanoid release. In contrast Said et al. (1974) found that the HPR in cat lung was only slightly diminished with artificial perfusate. Although details of their preparation were not given, their lungs were perfused at extremely low flow rates. In their cat lung preparation they also identified the release of biologically active substances including prostanoids. We attempted to separate the increase in resistance due to obstruction of flow in the capillaries by less deformable erythrocytes from the increase in resistance due to
120
T.S. HAKIM AND A.B. MALIK
vasoconstriction. We postulated that there is a 'capillary' middle segment where the effects of erythrocyte deformability are likely to manifest, and two other segments (arterial and venous) where vasoconstriction occurs. Therefore during blood perfusion we expected hypoxia to cause a large increase in resistance of the middle segment (APm), while with plasma perfusion any increase in resistance would be primarily in the arterial or venous segments (APa or APv). Indeed, the arterial and venous occlusion data revealed that the H P R during blood perfusion occurred primarily in the middle segment with only little change in the arterial or venous segments. These findings are consistent with results on dog and pig lungs (Hakim et al., 1983; Rock et al., 1985). With plasma perfusion, the increase in the middle resistance during hypoxia was attenuated but still remained in the middle segment. These results do not contradict our prediction, neither do they support our contention fully, possibly because the middle segment is not entirely limited to capillaries, but also includes small muscular vessels on which hypoxia acts directly. The results on cat LLL could be interpreted in two ways. Firstly, that the hypoxic pressor response which occurs in the middle segment is partly due to contraction of smooth muscle in the arterioles (or venules) and partly due to obstruction of the capillaries by nondeformable erythrocytes. The effect of the latter is abolished when erythrocytes are not present, and thus the H P R becomes attenuated during plasma perfusion. The role of erythrocyte deformability during hypoxia is supported by in vitro studies in cat erythrocytes (Hakim and Macek, 1986). The second less likely explanation is that erythrocyte deformability is of no importance and that the response in the middle segment is entirely due to contraction of smooth muscle in the microvessels, which for unknown reasons becomes attenuated during plasma perfusion. The second explanation is less attractive because it cannot be uniformly applied to the three species. It should be emphasized that although our intention was to compare the hypoxic pressor response in the presence and absence of red cells, we may have altered inadvertently another component in the perfusate in spite of the identical protocols used in the three species. Whatever that component is, it apparently affected the response in pig lung in a different way than in cat or rat lungs. In summary, it appears that in some species the hypoxic pressor response is dependent on the presence of erythrocytes in the perfusate. A unified explanation for the H P R in different species is that hypoxia affects vascular smooth muscle tone and RBC deformability. The relative contribution varies among species: smooth muscle contraction is of primary importance in pig lungs, but RBC deformability may also be important in rat or cat lung in mediating the hypoxic pressor response in the pulmonary vasculature.
Acknowledgements.We wish to thank Dr. F. Blumenstockfor directingthe prostaglandinassays. This work was supportedby the MedicalResearch Councilof Canada and in part by grant HL 32418fromthe National Institute of Health. T. S.H. is a scholar of the Medical Research Council.
PULMONARY HYPOXIC VASOCONSTRICTION
121
References Bergofsky, E.H. and S. Holtzman (1967). A study of the mechanisms involved in the pulmonary arterial pressor response to hypoxia. Circ. Res. 20: 506--519. Berkov, S. (1974). Hypoxic pulmonary vasoconstriction in the rat. Circ. Res. 35: 256-261. Bishop, M.J. and F.W. Cheney (1983). Comparison of the effects of minoxidil and nifedipine on hypoxic pulmonary vasoconstriction in dogs. J. Cardiovasc. Pharrnacol. 5: 184-189. Braash, D. (1971). Red cell deformability and capillary blood flow. Physiol. Rev. 679-700. Dawson, C.A., F.A. Delano, L.H. Hamilton and W.J. Stekiel (1974). Histamine releasers and hypoxic vasoconstriction in isolated cat lungs. J. Appl. Physiol. 37: 670-674. Dawson, C.A. (1984). Role of pulmonary vasomotion in physiology of the lung. Physiol. Rev. 64:544-616. Gorsky, B.H. and T.C. Lloyd (1967). Effects of perfusate composition on hypoxic vasoconstriction in isolated lung lobes. J. Appl. Physiol. 23: 683-686. Hakim, T. S., T. P. Michel, H. Minami and H. K. Chang (1983). Site of pulmonary hypoxic vasoconstriction studied with arterial and venous occlusion. J. Appl. Physiol. 54: 1298-1302. Hakim, T. S. and A.S. Macek (1986). Erythrocyte deformability and hypoxic pulmonary vasoconstriction. Fed. Proc. 45: 161. Hauge, A. (1968). Conditions governing the pressor response to ventilation hypoxia in isolated perfused rat lungs. Acta Physiol. Scand. 72: 33-44. Hauge, A. (1969). Hypoxia and pulmonary vascular resistance. The relative effects of pulmonary arterial and alveolar PO2. Acta Physiol. Scand. 76: 121-130. Hwang, D.H. (1985). Species variation in platelet aggregation. In: The Platelets: Physiology and Pharmacology, edited by G.L. Longenecker. New York, Academic Press, pp. 289-303. Kay, J.M. (1983). Comparative morphologic features of the pulmonary vasculature in mammals. Am. Rev. Respir. Dis. 128: $53-$57. Kivity, S. and J.F. Souhrada (1981). Plasma and platelets potentiate a hypoxic vascular response of the isolated lung. J. Appl. Physiol. 51: 875-880. Lingren, L., C. Marshall and B. E. Marshall (1985). Hypoxic pulmonary vasoconstriction in isolated rat lungs perfused with perfluorocarbon emulsion. Acta Physiol. Scand. 123: 335-338. Madden, J.A., C.A. Dawson and D.R. Harder (1985). Hypoxia-induced activation in small isolated pulmonary arteries from the cat. J. Appl. Physiol. 59:113-118. McMurtry, I. F., B.W. Hookway and S.D. Roos (1978). Red blood cells but not platelets prolong vascular reactivity of isolated rat lungs. Am. J. Physiol. 234: H186--H191. Ou, L. C. and R. P. Smith (1983). Probable strain differences of rats in susceptibilities and cardiopulmonary responses to chronic hypoxia. Respiration 377. Peake, M.D., A.L. H arabin, N.J. Brennan and J.T. Sylvester (1981). Steady-state vascular responses to graded hypoxia in isolated lungs of five species. J. Appl. Physiol. 51: 1214-1219. Rock, P., G.A. Patterson, S. Permutt and J.T. Sylvester (1985). Nature and distribution of vascular resistance in hypoxic pig lungs. J. Appl. Physiol. 59: 1891-1901. Said, S.I., T. Yoshida, S. Kitamura and C. Vreim (1974). Pulmonary alveolar hypoxia: Release of prostaglandins and other humoral mediators. Science 185:1181-1183. Selig, W.M., T.C. Nooman, D.F. Kern and A.B. Malik (1986). Pulmonary microvascular responses to arachidonic acid in isolated perfused guinea pig lung. J. Appl. Physiol. 60: 1972-1979. Tucker, A., I. F. McMurtry, J.T. Reeves, A.F. Alexander, D. H. Will and R. F. Grover (1975). Lung vascular smooth muscle as a determinant of pulmonary hypertension at high altitude. Am. J. Physiol. 228: 762-767. Voelkel, N.F. (1986). Mechanisms of hypoxic pulmonary vasoconstriction. Am. Rev. Respir. Dis. 133: 1186-1195. Walker, B. R., N.F. Voelkel, I.F. McMurtry and E.M. Adams (1982). Evidence for diminished sensitivity of the hamster pulmonary vasculature to hypoxia. J. Appl. Physiol. 52: 1571-1574.
Respiration Physiology (1988) 72, 109-122 Elsevier
RSP 01384
Hypoxic vasoconstriction in blood and plasma perfused lungs T. S. Hakim and A.B. Malik Department of Physiology, McGill University, Montreal, Quebec, Canada and Albany Medical College, Albany, NY, U.S.A. (Accepted for publication 2 November 1987) Abstract. Isolated lungs of pigs, cats and rats were perfused in situ with blood and with plasma at constant flow rate. The experiments were specifically designed to compare the intensity of the hypoxic pressor response (HPR) during blood and plasma perfusion. Ventilation of the lungs with 6.0~o 02 during perfusion with blood, elevated the inflow pressure (Pa) from 18.9 + 1.5 to 35.8 +_2.0 mm Hg in pig lungs, from 12.9 + 0.5 to 17.9 + 1.1 mm Hg in rat lungs, and from 12.6 + 0.8 to 19.2 + 1.5 mm Hg in cat lungs. In comparison during perfusion of the lungs with plasma, the HPR was larger in pig lungs (Pa increased from 16.9 + 2.0 to 42.3 + 2.7 mm Hg), smaller in rat lungs (Pa increased from 10.2 _+ 0.9 to 11.4 + 1.2 mm Hg), and also smaller in cat lung (Pa increased from 9.9 + 1.2 to 11.9 + 1.2 mm Hg). The site of HPR in blood perfused cat lobes was primarily in the middle segment (arterial and venous technique) and remained in the middle segment (although blunted) during plasma perfusion. The vascular pressure-flow relationship in pig lungs showed that during blood perfusion, both the slope and intercept rose during hypoxia; these changes were similar but greater during plasma perfusion. Measurement of thromboxane A 2 and prostacyclin in the lung effluent revealed no relation to intensity of the HPR or to the difference in the HPR in blood and plasma perfused lungs. Thus the absence of red blood cells (RBCs) may have been responsible for the difference in the HPR during blood and plasma perfusion; in rats and cats, RBCs appear to be essential for the full expression of the HPR, in contrast, the HPR in pigs does not require the presence of the RBCs. Possible explanations for these differences are suggested.
Cat; Hypoxia; Pig; Prostaglandins; Rat; Red cell deformability
Although a hypoxic pressor response (HPR) can be elicited in cell-free perfused lungs from rats (Berkov, 1974; McMurtry et al., 1978; Kivity and Souhrada, 1981; Lingren et al., 1985), cats (Berkofsky and Holtzman, 1967; Said et al., 1974) and dogs (Gorsky and Lloyd, 1967), it is not certain whether the intensity, mechanisms and site of HPR in blood and in cell-free perfused lungs are similar. The HPR is a very labile response (Voelkel, 1986), and varies considerably among individuals and species (Tucker et al., Correspondence address: T. S. Hakim, Department of Physiology, McGill University, 3655 Drummond, Montreal, Quebec, Canada, H3G 1Y6. 0034-5687/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)
110
T.S. HAKIM AND A.B. MALIK
1975), strains (Ou and Smith, 1983), geographical regions (Peake et al., 1981; Bishop and Cheney, 1983), and the number ofhypoxic challenges to which the lung was exposed (McMurtry et al., 1978). The lability in the response makes comparisons of different results difficult even in blood-perfused lungs. In the absence of blood cells, HPR apparently becomes either not present (Hauge, 1968; Berkov, 1974), reduced (Gorsky and Lloyd, 1967) or cannot be maintained (McMurtry et al., 1978) unless plasma or a vasoconstrictor agent is added to the perfusate. Our intention in this study was to compare the HPR in lungs from three animal species, during perfusion with whole autologous blood and then during perfusion with plasma. Each lung served as its own control using the same mixture of normoxic and hypoxic gases and, to avoid the individual variability in HPR, each lung was perfused with blood, and with plasma. HPR was tested only once during each period of perfusion. The results show that HPR is diminished in rat and cat lungs when blood cells are absent from the perfusate but in contrast it becomes enhanced in pig lungs. The differences in HPR are discussed in terms of the site of the HPR. Moreover, the differences were not related to vasoconstrictor and vasodilator prostanoids released during blood or plasma perfusion.
Materials and methods
The animals used in this study were eight male Sprague-Dawley rats (311 + 7 g), nine miniature pigs of either sex (11.1 + 0.3 kg) and eight cats of either sex (3.6 + 0.1 kg). The rats and cats were anaesthetized with an intraperitoneal injection of sodium pentobarbital (25 mg/kg), while the pigs were premedicated with an intramuscular injection of ketamine (5 mg/kg) and atropine (2 mg/kg) followed (15 min later) by an intravenous injection of sodium pentobarbital, given slowly to induce anaesthesia (up to 25 mg/kg). A catheter was placed in the carotid artery and the animals were heparinized (700 units/kg). They were ventilated mechanically and exsanguinated through the carotid artery. Midway during the bleeding, 10-15 ml/kg body weight of plasma expander was injected into the animal. The plasma expander consisted of 6~o dextran 40 (Pharmacia) and Ringer's lactate (Abbott) mixed in equal volumes. A fraction of the blood was used to prime the perfusion system, and another fraction was saved in a plastic container and kept at 37°C in a shaker bath. The remaining third fraction was mixed with equal volume of plasma expander and centrifuged at 1750 x g for 10 min. The supernatant plasma was collected and saved in a plastic beaker in the shaker bath to be used later for perfusion of the lungs. Isolated whole lung preparation. The procedure of preparing the whole lung for perfusion in situ was identical in rats and pigs. A rapid mid-sternal incision was performed to expose the lungs. The main pulmonary artery was isolated and an appropriate size cannula was placed in it via an incision in the right ventricle. Another cannula was placed in the left atrium via an incision in the atrial appendage. The cannulas were secured in position with sutures, and heavy ties were placed around the heart and aorta.
PULMONARY HYPOXIC VASOCONSTRICTION
111
The pulmonary arterial cannula and left atrial cannula were connected to the primed perfusion system. A bypass connection was included between the arterial and venous cannulas. The perfusion system consisted of a reservoir open to room air, a roller pump (Masterflex 500) and a heat exchanger (39°C). The blood was pumped from the reservoir via the heat exchanger, into the pulmonary artery and the blood was allowed to drain back passively into the reservoir. During this time the lungs were ventilated with a gas mixture containing 5.5~ CO2, 35% 02 and nitrogen. End inspiratory pressure was kept below 10 mm Hg and end expiratory pressure was 1-2 mm Hg. The volume of blood in the perfusion system was approximately 20 ml in the rat lung preparations and about 400 ml in the pig lung preparations. Donor rats had to be bled to obtain adequate blood and plasma. The blood perfusion rate was increased gradually while keeping the pulmonary arterial pressure below 20 mm Hg. The rat lungs usually stabilized within 15 min while the pig lungs took up to 60 min to become stable. The pulmonary arterial pressure (Pa) and left atrial pressure were measured from side ports in the cannulas, and referenced to the top of the lungs. The left atrial pressure was set between 0 and 1.0 mm Hg and was kept constant in each experiment. During the period of stabilization, blood gases and pH were normalized by changing ventilation and by adding sodium bicarbonate into the reservoir. Once the Pa became stable (flow rate in pig and rat lungs was 485 + 32 and 10.4 + 0.4 ml/min respectively), the experimental protocol was initiated. Effluent blood samples were withdrawn for gases, pH, hematocrit (Hct) and prostaglandin (thromboxane and prostacyclin) analysis. The lung was ventilated with an isocapnic hypoxic gas mixture (5.5~o CO2, 6 ~ 02 and nitrogen) for a duration of 10 min while keeping flow rate constant. Another effluent hypoxic blood sample was withdrawn for gases, Hct, and PG analysis. Ventilation was returned to normoxic gas until Pa and blood gases returned to baseline values. During the normoxic and hypoxic periods in the pig lung, the Pa was measured at 3 or 4 different flow rates in random order, from which a pressure-flow relationship (slope and pressure intercept) was derived. The bypass connector was opened and two clamps were placed near the cannulas. The blood was removed out of the perfusion system rapidly and the system was washed with saline and refilled with plasma perfusate. It was circulated in the tubing for 2 min, and the perfusion was reestablished by removing the two clamps on the cannulas and closing the bypass. The initial bloody effluent from the lung was discarded. The perfusion rate was increased gradually until the lung reached a new steady state as suggested by a stable Pa. The flow rates were equal to those used during blood perfusion. During the time, the perfusion was switched from blood to plasma (3-5 min) the lungs continued to be ventilated with the normoxic gas mixture. The pH in the perfusate was adjusted by adding bicarbonate whenever necessary. Once stability was reestablished, a plasma sample was withdrawn for gases, Hct and PG analysis. The Hct was always less than 1% and could not be read in the usual manner as the blood. Ventilation was then switched to the hypoxic gas mixture for 10 min and again plasma samples were withdrawn for analysis as before. Ventilation with normoxic gas was resumed until Pa and gases returned to baseline level. The plasma was rapidly emptied out of the system
112
T.S. HAKIM AND A.B. MALIK
and replaced with fresh blood without allowing air bubbles to enter the lung. After reaching a steady state with blood perfusion using the same flow rates as before, blood samples were withdrawn during normoxic and hypoxic periods for measuring Hct, gases and pH, but P G analysis was not carried out during the second blood perfusion. The pressure-flow relationships in pig lungs during normoxic and hypoxic conditions were repeated during plasma perfusion and they were repeated in two lungs during the second period of blood perfusion.
Isolated left lower lobe (LLL) preparation in cats. The experimental protocol in the cat was identical to the one for rats and pigs, except that instead of perfusing the whole lung, only the left lower lobe was perfused in situ as described in detail previously for dog lungs (Hakim et al., 1983). Briefly, after the cats were exsanguinated, the LLL was exposed through a left thoracotomy. The main artery and vein of the LLL were isolated and cannulated. The cannulas were connected to a perfusion system as above. The volume of blood in the cat LLL perfusion system was about 30 ml. The LLL was initially perfused with autologous blood (flow rate = 49.7 + 2.2 ml/min), then with plasma, and again with fresh blood at the same flow rate. Replacement of the blood or plasma was accomplished as described above in the rat and pig lungs. One additional measurement was made in cat LLL, namely that during each steady state of normoxia and hypoxia, arterial and venous occlusion techniques were applied to determine if there was any difference in the site of H P R during blood and plasma perfusion. Cat LLL usually stabilized as rapidly as rat lungs. As before, PG samples were analyzed for the normoxic and hypoxic period during the first blood perfusion and during plasma perfusion only, but not during the second blood perfusion. As before only one H P R was tested during each period of perfusion. Preparation of the plasma sample for prostaglandin analysis. Effluent blood or plasma sample (2 ml) which was withdrawn for P G analysis was treated by adding 0.5 ml of 2.5 mg/ml indomethacin dissolved in distilled water, and centrifuged at 1750 × g for 10 min. The clear supernatant was stored in polypropylene vials, sealed and kept frozen at -70 °C until they were analyzed. Analysis for thromboxane B 2 (TxBz) and 6-ketoprostaglandin F j ~ (6-keto-PGF 1~) (stable metabolites of TxA 2 and prostacyclin) were carried out using a radioimmunoassay as described previously (Selig et al., 1986). Organic extraction to remove possible interfering proteins was not required, since plasma standards were included in this assay system. The cross reactivity of the 6-keto-PGFl~ antiserum was as follows: 3.3~o PGFI~: 0.1~o prostaglandin D 2 (PGD2); 0.24~o PGE2; 0.27~o PGF2~; 0.01~o with TxB2, AA, or linoleic acid. The cross reactivity of the TxB 2 antiserum was as follows: 0.06~o P G D 2 and 0.01 ~o for P G E e, PGF2~, PGF1~, 6-keto-PGFl~, AA, or linoleic acid. A linear working range for both TxB 2 and 6-keto P G F ~ antisera was established between 20 and 10000 pg for pig and cat samples, and between 20 to 2000 pg for rat samples. Statistical analysis of the results was done using a Student t-test and a P value of less than 0.05 was considered significant. Analysis of variance test was used to confirm the
PULMONARY HYPOXIC VASOCONSTRICTION
113
differences in the hypoxic pressor responses a m o n g the three animal species. All values are given as m e a n s + SE except in tables 1 a n d 2 in which _ SD are given instead.
Results A complete protocol consisted of one hypoxic challenge (10 min) and a return to baseline during each of the three perfusion periods; blood perfusion, plasma perfusion, a n d again blood perfusion. The total duration of each experiment was 2 - 3 h and usually depended on the time to reach a stable state. Because of the lack of adequate blood volume in some experiments, H P R was tested once with blood and once with p l a s m a in r a n d o m order. Table 1 summarizes the m e a n values of hematocrit (Hct), gases, and p H in the blood and plasma during the normoxic (N) and hypoxic (H) periods. Although the data during the second blood perfusion period are not shown in the table, the gases were similar to the first period, but the Hct were slightly lower (80-90~o) than during the first perfusion period.
TABLE 1 Hematocrits (Hct) and perfusate gases during normoxic (N) and hypoxic (H) conditions. Values are ~_+SD. Blood Hct ~ Pigs
Plasma pH
Pco2
Po2
pH
Pco~
Po2
7.37 + 0.05
43.3 _+4.9
201.8 +_23.4
7.37 + 0.03
44.4 + 6.3
198.2 + 30.7
H
7.40 +0.01
40.9 +2.5
46.2 +2.7
7.35 +0.05
46.1 _+2.9
54.1 _+11.1
N
7.38 +0.07
34.5 +2.6
169.1 +7.6
7.48 +0.05
35.3 +2.1
158.6 + 10.1
H
7.35 + 0.09
37.7 + 2.7
47.4 + 4.0
7.40 _+0.05
39.9 +_3.9
60.5 + 4.6
N
7.41 +0.03
33.6 +_1.4
181.3 +6.1
7.43 +0.09
35.5 +2.1
177.1 +7.8
7.43 + 0.04
33.0 _+1.6
40.5 + 3.7
7.43 _+0.08
36.4 + 1.6
53.7 _+3.7
N 26.6 + 1.6
Rats
22.4 +3.7
Cats
30.0 + 3.6 H
114
T.S. HAKIM AND A.B. MALIK
7_+SE
A
1- 40
/ /
E
g
Pigs
(lO)
~
3o (~..I
•~- 20 .......
< ~¢-
I""
¢...,,.,'~ o'1
i.~.!....i
Rats
...........~
Cats
~ ........... ~61
lO
0
E 5 o
I
I
N
H Blood
I
N Plasma
I
H
I
I
N
H Blood
Fig. 1. Pulmonary arterial pressure during ventilation with 35% 02 (N) and with 6% 0 2 (H) in isolated lungs of three animal species. The lungs were perfused with blood, then with plasma and again with blood. The number in brackets is the number of measurements in each group.
HPR in blood vsplasma-perfused lungs. In 12 out of 15 hypoxic challenges in pig lungs perfused with blood, the increase in Pa ranged between 17 to 24 mm Hg, whereas in 3 lungs Pa increased by 8 mm Hg. With plasma perfusion, the increase in Pa ranged between 20 and 37 in 9 out of 10 measurements and was 19 in one lung. The mean Pa rose from 18.9 + 1.5 to 35.8 + 2.0 and from 21.3 + 1.8 to 37.5 + 3.1 mm Hg during the 1st and the 3rd perfusion period (with blood) respectively, and from 16.9 + 2.0 to 42.3 + 2.7 mm Hg during perfusion with plasma (fig. 1). Thus, the H P R was unequivocally enhanced by about 10 mm Hg during plasma perfusion in spite of the higher Po2 (54.1 + 11.1 mm Hg) with plasma than with blood perfusion (46.2 + 2.7 mm Hg). The responses in the rat lungs were different. In 11 out of 12 measurements during blood perfusion, the increase in Pa in response to 6% 0 2 ranged from 3 to 9 mm Hg and in one lung there was only a 1 mm Hg rise in Pa. During plasma perfusion the increase in Pa was less than 2 mm Hg in 8 out of 9 measurements, and about 7 mm Hg in one lung. The mean Pa during blood perfusion rose from 12.9 + 0.5 to 17.9 + 1.1 mm Hg during the first period and from 15.8 + 1.7 to 20.9 + 2.1 mm Hg during the third period. With plasma perfusion, Pa rose from 10.2 + 0.9 to 11.4 + 1.2 mm Hg (fig. 1). Thus unlike the pig, the response in the rat to this level of hypoxia was attenuated in the absence of red blood cells. Note, however, that the Po2 in the plasma during hypoxia (60.5 + 4.8 mm Hg) was greater than the Po2 in the blood (47.4 + 4.0 mm Hg) even though the same hypoxic gas mixture was used in both cases. Whether this difference in Po~ is of physiological importance is not clear, in view of the evidence that blood gas Po2 is less important than alveolar gas Po2 in mediating H P R (Hauge, 1969).
PULMONARY HYPOXIC VASOCONSTRICTION
115
The response in the cat lung was similar to rat lung, but more variable. In 9 out o f 14 measurements, the increase in Pa ranged between 4 and 11 m m Hg, in 4 measurements the response was less than 4 m m Hg, and in 2 measurements it decreased by 0.5 m m Hg. With plasma perfusion the response usually ranged between 0 and 5 m m Hg. The mean Pa increased from 12.6 + 0.8 to 19.2 + 1.5 m m Hg during the first perfusion period, and from 12.3 + 1.6 to 17.3 + 2.0 m m H g during the third perfusion period. With plasma perfusion, the Pa rose from 9.9 + 1.2 to 11.9 + 1.2 m m H g (fig. 1). Thus, like the rats, but unlike the pigs, the H P R in cat was attenuated when blood cells were not present in the perfusate. In this case, the Po2 during hypoxia was also slightly higher in the plasma (54.1 + 11.1 m m Hg) than in the blood (46.2 + 2.7 m m Hg) in spite of using the same hypoxic gas mixture for both challenges.
Site of HPR during blood and plasma perfusion. To examine if the site of action is altered by the absence of the red blood cells from the perfusate, we performed the arterial and venous occlusion during normoxia and hypoxia in cat L L L perfused with blood and then with plasma. The results (table 2, fig. 2) revealed that during blood perfusion the H P R occurred primarily in the middle segment (P < 0.05), but was attenuated during plasma perfusion. In pig lung, we examined the site of action of hypoxia by studying the pressure-flow relationship. This allowed us to see if the effect on the slope and on the pressure intercept (critical closing pressure) is similar in blood and plasma perfused lungs. We succeeded in obtaining a complete set of pressure-flow data on 5 lungs during normoxia and hypoxia while perfusing with blood and with plasma (fig. 3). In two lungs we were able to repeat these measurements during the second blood perfusion period. These two showed changes similar to the changes observed during the first perfusion period. We were somewhat conservative in the number of data points, because on the basis of preliminary experiments we found that a combination of high flow rate, low hematocrit and hypoxia, there was a great tendency for the lungs to become edematous. During normoxia the mean slope and intercept during blood perfusion were TABLE 2 Segmental pressure gradients (mm Hg) in cat left lower lobes (n = 6) during normoxia and hypoxia, when perfused with blood, plasma and again with blood. Values are ~ + SE. Total
Arterial
Middle
Venous
Blood
Normoxia Hypoxia
9.8 + 0.8 16.2 + 2.2*
2.2 ± 0.6 3.6 +_0.7
3.6 + 0.5 8.7 ± 1.9"
3.5 +_0.2 3.8 + 0.4
Plasma
Normoxia Hypoxia
6.7 + 0.8 8.7 + 1.2
2.0 + 0.4 1.9 + 0.4
1.9 + 0.6 3.5 ± 0.9
2.9 + 0.2 3.4 + 0.3
Blood
Normoxia Hypoxia
8.5 + 0.9 12.7 + 2.2*
2.4 + 0.4 2.3 + 0.5
2.8 + 0.5 6.5 + 1.5"
3.3 + 0.4 3.9 _+0.6
* Significantly different from the value immediately above it.
Ot
116
T.S. HAKIM AND A.B. MAL1K n=6 x+SE
~rterial L.
lO
/ ~iddle
5 Cenous N
H
N
Blood
H
N
Plasma
H
Blood
Fig. 2. Distribution of vascular pressure gradients in cat left lower lobes during perfusion with blood and with plasma in normoxic (N) and hypoxic (H) conditions. Flow rate was kept constant throughout each experiment. The means and standard errors of the individual segments as well as their statistical significance are given in table 2.
25.7 + 4.6 m m H g . L - ~. min and 7.5 + 2.2 m m Hg, respectively, but both increased (P < 0.05) during hypoxia to 40.2 + 10.7 m m H g . L - ~ • min and 13.2 + 3.4 m m Hg, respectively. With plasma perfusion, the slope and intercept were 18.1 + 2.9 m m H g - L - ~ • min and 4.3 + 1.4 m m H g respectively, and both increased
/
Blood
//
Plasma
/
/
/ !
5O
/
-1-
E E
/
/
i/ /
40
/
z11 "II .~/11. /
/I/1
II
,,
g II
iI
~ if/
/ ~
09 tO
13_ 30
<
20
~
¢-
o
_E lO
~
-
+
Slope Normoxia Hypoxia I 0.2
25.7 -+ 4.6 40.2 _+ 10.7 I I 0.4 0.6
S
E Intercept 7.5 _+ 2.2
13.2 -+ 3.4 I I 0.8 1.0
Flow Rate
Normoxia
0
18.1 -+ 2.9
Hypoxia 40.1 -+ 7.5 I I I 0.4 0.6 0.2
4.3 +- 1.4 20.9 _+ 2.3 I I 0.8 1.0
(I/min)
Fig. 3. The pulmonary arterial pressure-flow rate relationship during normoxic (solid line) and hypoxic (broken line) condition in blood and plasma perfused pig lungs.
117
PULMONARY HYPOXIC V A S O C O N S T R I C T I O N TABLE 3
Effluent levels (ng/ml) of thromboxane B z (TxB2) and 6-ketoprostaglandin-Fl~ (6-keto-PGFl~) during normoxia (N) and hypoxia (H) in blood and plasma perfused lungs. Values are ~ + SD. Blood perfusion
Plasma perfusion
TxB2
6-keto-PGFl~
TxB2
6-keto-PGF~
Pigs (n = 4)
N H
1.8 + 1.9 1.3 + 0.8
1.4 + 0.8 1.0 + 0.7
1.9 + 0.3 1.2 + 1.3
1.6 _+ 0.6 1.8 _+ 0.7
Rats (n = 8)
N H
9.7 + 4.5 6.2 + 3.8
3.4 + 1,1 3.1 + 0.9
8.0 + 6.4 8.3 + 4.4
3.9 + 2.4 4.0 + 2.4
Cats (n = 6)
N H
0.8 + 0.5 0.5 + 0.2
1.8 + 1,0 1.8 +_ 1.1
0.5 + 0.2 0.5 + 0.2
1.4 + 0.9 2.1 + 1.5
(P < 0.05) during hypoxia to 40.1 + 7.5 mm Hg. L - ~ • rain and 20.9 + 2.3 mm Hg. Thus the increase in both the slope and the intercept were enhanced during plasma perfusion. These data indicate that the sites of H P R were similar but potentiated when the blood cells were not present. In an attempt to assess whether the release of mediators can account for the differences in the H P R among species, and more importantly between the blood- and plasma-perfused lungs, we measured the effluent perfusate concentration of TxB 2 and 6-keto-PGFl~: precursors of a potent pulmonary vasconstrictor (TxA2) and a pulmonary vasodilator prostaglandin (PGI2). The results are shown in table 3. As expected (Hwang, 1985) the levels of both mediators were different in the three animals. The concentrations of both mediators were not significantly different during blood and plasma perfusion, neither were the levels altered by changing from normoxic (35yo 02) to hypoxic (6% 02) conditions. These results suggest that TxA z and PGI 2 release could neither account for the differences in H P R in the three species, nor could it account for the differences in the H P R in blood vs plasma perfused lungs. Role ofprostaglandins in the HPR.
Discussion
Our main objective in this study was to determine the importance of red blood cells (RBCs) in the HPR. Although the H P R has been studied in lungs perfused with RBCs free perfusate, the difference in H P R in blood vs plasma perfused lungs has seldom been the focus of the study. In a recent preliminary study we found that RBCs became less flexible during hypoxia (Hakim and Macek, 1986) and postulated that RBCs may be involved during the HPR. Therefore, we set out to compare the H P R in lungs perfused
118
T.S. H A K I M A N D A.B. M A L I K
with whole blood and with cell free perfusate; we expected that because rat RBCs became relatively rigid during hypoxia (Hakim and Macek, 1986), they may contribute to the HPR by restricting flow in the capillaries (Braash, 1971). Thus, their removal from the perfusate would attenuate the HPR. In contrast, because pig RBCs were insensitive to hypoxia (Hakim and Macek, 1986), we expected that their removal from the perfusate would not alter the HPR. We also included the cat as a third test species. Because the HPR is a very labile response, the study was designed to minimize as much as possible the differences which may contribute to alterations in the hypoxic presser response (HPR) of the lung vasculature. The protocols and procedures were identical in all experiments, and the total duration of the experiment was kept between 2 and 3 h. Each lung served as its own control and to eliminate the time factor, the HPR was tested in most lungs before and after plasma perfusion. Each lung exposure to hypoxia was repeated three times with the same gas mixture, once during each perfusion period, with each time exposure lasting for 10 min. We avoided repeated challenges as this may alter the intensity of the HPR due to depletion or accumulation of metabolites (McMurtry et al., 1978). To reduce the effects of Pco~ and pH (Dawson, 1984), they were kept within a very narrow range during blood and plasma perfusion (table 1). The temperature of the blood and plasma was kept near 37 ° C. Thus, these experiments were designed not only to compare the response in different animals but also to compare the HPR in blood vs plasma perfused lungs. Our pig and cat lungs were as responsive as previously reported by other investigators (Dawson et al., 1974; Peake et al., 1981). Similarly, the HPR in rat lungs to 6~o 02 were also comparable to those found by others (McMurtry etal., 1978; Kivity and Souhrada, 1981; Walker etal., 1982). The differences in the acute HPR among species apparently cannot be explained by the medial thickness of smooth muscle in the muscular vessels (Kay, 1983). It is interesting to note that the muscular arteries in the pigs are limited to vessels 25-70/~m in diameter, while in cats or rats the muscular vessels range 20-500 #m, and 20-350 #m in diameter, respectively. Whether such differences contribute to the observed intensities in HPR is not clear. The HPR during plasma perfusion suggests that the absence of the RBCs from the perfusate affects the HPR differently in animal species. In some animals (rats, cats) the HPR was blunted, while in others (pigs) it was enhanced when blood cells were absent from the perfusate. Although leukocytes and platelets were also absent from the perfusate (2~o) their role is apparently minimal (McMurtry et aL, 1978). It is difficult to offer a single mechanism to explain the differences in HPR in plasmaand blood-perfused lungs of all three species. A number of mechanisms may be involved to different extents. These mechanisms may include the direct action of hypoxia on smooth muscle, alteration in vascular resistance by RBCs via changes in their deformability or their interaction with the vascular lumen, and the release of mediators (vasoconstrictors or vasodilators) from lung tissue sources. The release of mediators has indeed been the subject of many investigations (Dawson, 1984; Voelkel, 1986) and our study does not resolve this controversial subject; however, the data show that differences in prostanoid release could neither explain the differences among species nor the variability among individuals within the species, nor could it explain the difference in HPR in
PULMONARY HYPOXIC VASOCONSTRICTION
119
blood- vs plasma-perfused lungs. Therefore in our discussion, we will focus primarily on the influence of hypoxia on smooth muscle tone and on RBC deformability, because hypoxia has been shown in vitro to constrict isolated vessels (Madden et al., 1985) and to alter RBC deformability (Hakim and Macek, 1986). In the pig, the role of blood cells seems to be minimal, since the HPR was equally large during plasma perfusion and blood perfusion. This is consistent with in vitro measurement on RBC from the pig, which showed that hypoxia causes minimal alteration in RBC deformability (Hakim and Macek, 1986). It is also possible that in the pig, the RBCs could have in fact attenuated the HPR by either impeding the release of vasoconstrictors from the endothelial cells or from the platelets, or by taking up substances e.g. adenosine or prostacyclin (McMurtry et al., 1978). Thus by excluding the role of RBCs and the release of prostanoids, the direct effect of hypoxia on smooth muscle tone seems to be a likely mechanism in pig lungs. As previously observed (Rock et al., 1985), hypoxia caused the slope of the pressure-flow curve and its intercept to increase in pig lungs. Our results on the pressure-flow data in pig lungs (fig. 3) showed that the HPR became potentiated at all flow rates during plasma perfusion and suggest that the site of action is in both segments which contribute to the slope and the critical closing pressure of the vasculature. The results from rat lungs suggest entirely different mechanisms for the HPR. In the present study, removal of blood cells blunted the HPR in all but one rat. Thus the direct influence of hypoxia on smooth muscle tone (as suggested by a small HPR in plasmaperfused lungs), appears to be of little importance. Therefore factors other than the direct action of hypoxia or prostanoid release must be involved. On the basis of recent measurement of RBC deformability which revealed that rat RBC became less deformable during hypoxia (Hakim and Macek, 1986) we propose that RBC deformability may have played a significant role in the HPR in the rats. In contrast with our finding, McMurtry et al. (1978) reported that the HPR was not different in a group of lungs perfused with blood compared to another group perfused with plasma. The reason for the difference between the two studies is not clear: they compared two different groups, using less than 2~o 02 whereas we repeated the HPR in the same lungs using 6~o 02. Furthermore, the isolated lungs in our study were not pretreated with any substances to enhance the response. It is possible that the level of 02 is of critical importance because changes in RBC deformability and constriction of the smooth muscle may occur at different threshold. The response of cat lungs was unlike the rat or the pig. The HPR with blood varied considerably in different cats, and the response was usually diminished with plasma perfusion and independent of prostanoid release. In contrast Said et al. (1974) found that the HPR in cat lung was only slightly diminished with artificial perfusate. Although details of their preparation were not given, their lungs were perfused at extremely low flow rates. In their cat lung preparation they also identified the release of biologically active substances including prostanoids. We attempted to separate the increase in resistance due to obstruction of flow in the capillaries by less deformable erythrocytes from the increase in resistance due to
120
T.S. HAKIM AND A.B. MALIK
vasoconstriction. We postulated that there is a 'capillary' middle segment where the effects of erythrocyte deformability are likely to manifest, and two other segments (arterial and venous) where vasoconstriction occurs. Therefore during blood perfusion we expected hypoxia to cause a large increase in resistance of the middle segment (APm), while with plasma perfusion any increase in resistance would be primarily in the arterial or venous segments (APa or APv). Indeed, the arterial and venous occlusion data revealed that the H P R during blood perfusion occurred primarily in the middle segment with only little change in the arterial or venous segments. These findings are consistent with results on dog and pig lungs (Hakim et al., 1983; Rock et al., 1985). With plasma perfusion, the increase in the middle resistance during hypoxia was attenuated but still remained in the middle segment. These results do not contradict our prediction, neither do they support our contention fully, possibly because the middle segment is not entirely limited to capillaries, but also includes small muscular vessels on which hypoxia acts directly. The results on cat LLL could be interpreted in two ways. Firstly, that the hypoxic pressor response which occurs in the middle segment is partly due to contraction of smooth muscle in the arterioles (or venules) and partly due to obstruction of the capillaries by nondeformable erythrocytes. The effect of the latter is abolished when erythrocytes are not present, and thus the H P R becomes attenuated during plasma perfusion. The role of erythrocyte deformability during hypoxia is supported by in vitro studies in cat erythrocytes (Hakim and Macek, 1986). The second less likely explanation is that erythrocyte deformability is of no importance and that the response in the middle segment is entirely due to contraction of smooth muscle in the microvessels, which for unknown reasons becomes attenuated during plasma perfusion. The second explanation is less attractive because it cannot be uniformly applied to the three species. It should be emphasized that although our intention was to compare the hypoxic pressor response in the presence and absence of red cells, we may have altered inadvertently another component in the perfusate in spite of the identical protocols used in the three species. Whatever that component is, it apparently affected the response in pig lung in a different way than in cat or rat lungs. In summary, it appears that in some species the hypoxic pressor response is dependent on the presence of erythrocytes in the perfusate. A unified explanation for the H P R in different species is that hypoxia affects vascular smooth muscle tone and RBC deformability. The relative contribution varies among species: smooth muscle contraction is of primary importance in pig lungs, but RBC deformability may also be important in rat or cat lung in mediating the hypoxic pressor response in the pulmonary vasculature.
Acknowledgements.We wish to thank Dr. F. Blumenstockfor directingthe prostaglandinassays. This work was supportedby the MedicalResearch Councilof Canada and in part by grant HL 32418fromthe National Institute of Health. T. S.H. is a scholar of the Medical Research Council.
PULMONARY HYPOXIC VASOCONSTRICTION
121
References Bergofsky, E.H. and S. Holtzman (1967). A study of the mechanisms involved in the pulmonary arterial pressor response to hypoxia. Circ. Res. 20: 506--519. Berkov, S. (1974). Hypoxic pulmonary vasoconstriction in the rat. Circ. Res. 35: 256-261. Bishop, M.J. and F.W. Cheney (1983). Comparison of the effects of minoxidil and nifedipine on hypoxic pulmonary vasoconstriction in dogs. J. Cardiovasc. Pharrnacol. 5: 184-189. Braash, D. (1971). Red cell deformability and capillary blood flow. Physiol. Rev. 679-700. Dawson, C.A., F.A. Delano, L.H. Hamilton and W.J. Stekiel (1974). Histamine releasers and hypoxic vasoconstriction in isolated cat lungs. J. Appl. Physiol. 37: 670-674. Dawson, C.A. (1984). Role of pulmonary vasomotion in physiology of the lung. Physiol. Rev. 64:544-616. Gorsky, B.H. and T.C. Lloyd (1967). Effects of perfusate composition on hypoxic vasoconstriction in isolated lung lobes. J. Appl. Physiol. 23: 683-686. Hakim, T. S., T. P. Michel, H. Minami and H. K. Chang (1983). Site of pulmonary hypoxic vasoconstriction studied with arterial and venous occlusion. J. Appl. Physiol. 54: 1298-1302. Hakim, T. S. and A.S. Macek (1986). Erythrocyte deformability and hypoxic pulmonary vasoconstriction. Fed. Proc. 45: 161. Hauge, A. (1968). Conditions governing the pressor response to ventilation hypoxia in isolated perfused rat lungs. Acta Physiol. Scand. 72: 33-44. Hauge, A. (1969). Hypoxia and pulmonary vascular resistance. The relative effects of pulmonary arterial and alveolar PO2. Acta Physiol. Scand. 76: 121-130. Hwang, D.H. (1985). Species variation in platelet aggregation. In: The Platelets: Physiology and Pharmacology, edited by G.L. Longenecker. New York, Academic Press, pp. 289-303. Kay, J.M. (1983). Comparative morphologic features of the pulmonary vasculature in mammals. Am. Rev. Respir. Dis. 128: $53-$57. Kivity, S. and J.F. Souhrada (1981). Plasma and platelets potentiate a hypoxic vascular response of the isolated lung. J. Appl. Physiol. 51: 875-880. Lingren, L., C. Marshall and B. E. Marshall (1985). Hypoxic pulmonary vasoconstriction in isolated rat lungs perfused with perfluorocarbon emulsion. Acta Physiol. Scand. 123: 335-338. Madden, J.A., C.A. Dawson and D.R. Harder (1985). Hypoxia-induced activation in small isolated pulmonary arteries from the cat. J. Appl. Physiol. 59:113-118. McMurtry, I. F., B.W. Hookway and S.D. Roos (1978). Red blood cells but not platelets prolong vascular reactivity of isolated rat lungs. Am. J. Physiol. 234: H186--H191. Ou, L. C. and R. P. Smith (1983). Probable strain differences of rats in susceptibilities and cardiopulmonary responses to chronic hypoxia. Respiration 377. Peake, M.D., A.L. H arabin, N.J. Brennan and J.T. Sylvester (1981). Steady-state vascular responses to graded hypoxia in isolated lungs of five species. J. Appl. Physiol. 51: 1214-1219. Rock, P., G.A. Patterson, S. Permutt and J.T. Sylvester (1985). Nature and distribution of vascular resistance in hypoxic pig lungs. J. Appl. Physiol. 59: 1891-1901. Said, S.I., T. Yoshida, S. Kitamura and C. Vreim (1974). Pulmonary alveolar hypoxia: Release of prostaglandins and other humoral mediators. Science 185:1181-1183. Selig, W.M., T.C. Nooman, D.F. Kern and A.B. Malik (1986). Pulmonary microvascular responses to arachidonic acid in isolated perfused guinea pig lung. J. Appl. Physiol. 60: 1972-1979. Tucker, A., I. F. McMurtry, J.T. Reeves, A.F. Alexander, D. H. Will and R. F. Grover (1975). Lung vascular smooth muscle as a determinant of pulmonary hypertension at high altitude. Am. J. Physiol. 228: 762-767. Voelkel, N.F. (1986). Mechanisms of hypoxic pulmonary vasoconstriction. Am. Rev. Respir. Dis. 133: 1186-1195. Walker, B. R., N.F. Voelkel, I.F. McMurtry and E.M. Adams (1982). Evidence for diminished sensitivity of the hamster pulmonary vasculature to hypoxia. J. Appl. Physiol. 52: 1571-1574.