Nitrovasodilators such as nitroglycerin (NTG) and sodium nitroprusside are frequently used in anesthesia and critical care. Since patients receiving nitrovasodilators often have pulmonary edema, it is important to know whether nitrovasodilators directly alter pulmonary vascular permeability. These compounds produce their vasodilator effects by releasing nitric oxide (NO). Although several studies suggest that NO (either endogenously produced or exogenously administered) may alter vascular permeability, the direction of the effect is unclear, with studies demonstrating that NO results in increased [1-7], decreased [7-19], and unchanged [8,20] permeability. These contradictory results may be due to differences in the species, tissues, and models studied interacting with several different effects of NO, such as vasodilation, modulation of neutrophil function, and alterations in free radicals. Although the in vivo effects of nitrovasodilators are important, understanding the direct effects of nitrovasodilators on pulmonary vascular permeability requires the use of a model that confines the effects to the lung and allows direct measurement of vascular permeability. The current study, therefore, examined the effect of NTG on pulmonary vascular permeability in both injured and noninjured isolated buffer-perfused rabbit lungs.
The protocol was approved by the Stanford Administrative Panel on Laboratory Animal Care. Our previously described isolated perfused lung preparation was used [17,21]. Male New Zealand White rabbits weighing 2.2-3.0 kg were anesthetized with ketamine (85 mg/kg intramuscularly), followed by sodium pentobarbital (12.5-25 mg/kg intravenously). A tracheostomy was performed, and the lungs were ventilated with oxygen at a frequency of 20 breaths per min, a peak inspiratory pressure of 10-12 mm Hg, and 2.5 cm H2 O positive end-expiratory pressure. A midline sternotomy was performed, heparin (300 U/kg intravenously) was administered, and the left atrium and pulmonary artery were cannulated via left and right ventriculostomies, respectively. The lungs were then ventilated with 95% air, 5% CO2 using the same ventilatory variables. The pulmonary circulation was perfused with Krebs-Henseleit solution containing 3% bovine serum albumin at 37 degrees C and pH 7.4 using a Masterflex pump (Cole-Parmer, Barrington, IL). The pulmonary circulation was initially flushed with 700 ml of nonrecirculated perfusate at a rate of 20-40 mL/min. The heart and lungs were then excised en bloc and suspended by the trachea from a counterbalanced force-displacement transducer (model FT-03, Grass, Quincy, MA) to continuously measure lung weight (LW). The preparation was enclosed in a warmed, humidified chamber. The lungs were then perfused in a recirculating manner at a flow rate of 150 mL/min using a total circuit volume of 450 mL. Pulmonary artery (PPA) and left atrial (PLA) pressures were continuously monitored via side ports in the respective cannulae. PPA and PLA were referenced to the height of the left atrium, and PLA was maintained at 2 mm Hg by adjustment of the reservoir height.
Pulmonary vascular permeability was assessed by measurement of the pulmonary capillary filtration coefficient (Kf,c) from the Starling equation. This technique begins with measurement of the isogravimetric pressure (PISO). PISO is the pulmonary capillary pressure (PPC) at which the Starling forces are balanced so that the lung neither gains nor loses weight . In order to measure PISO, a shunt between the pulmonary arterial and left atrial cannulae was opened and perfusion was discontinued so that PPA, PLA, and PPC were all equal. PPC was then increased in increments of 1 mm Hg by elevating the perfusate reservoir, and the effect on LW was recorded. PISO was defined as the maximum vascular pressure at which the lungs did not gain weight over an interval of 3 min. Kf,c was then measured using a modification of the method of Drake et al. , as previously described . This method of measurement of Kf,c is based on the fact that when PPC equals PISO, the Starling forces are balanced so that there is no net transfer of fluid across the pulmonary capillary membrane. According to the Starling equation, when PPC equals PISO, then JV = Kf,c ([PISO - PIS] - sigma [Pi PC - Pi IS]) = 0 where JV is the net fluid flux, Kf,c is the pulmonary capillary filtration coefficient, PISO is the pulmonary capillary hydrostatic pressure, PIS is the interstitial hydrostatic pressure, sigma is the osmotic reflection coefficient, Pi PC is the intravascular oncotic pressure, and Pi IS is the interstitial oncotic pressure. When the venous reservoir height is then raised so that PPC is abruptly increased from PISO to PISO + 7 mm Hg, the other Starling forces initially remain unchanged so that edema occurs at a rate of 7 x Kf,c. However, lung weight gain after the increase in PPC is due to both intravascular volume expansion and pulmonary edema formation. The intravascular volume expansion is rapid and essentially complete within several minutes. Therefore, the rate of LW gain every minute from 3 to 10 min was recorded on a semilogarithmic plot and extrapolated to Time 0 by linear regression. The resulting value of Jv was then used to calculate Kf,c, which is expressed in mL [center dot] min-1 [center dot] mm Hg-1 [centered dot] 100 g lung-1.
Isolated lungs (n = 13) were perfused for 30 min as described above, and baseline PISO and Kf,c were measured. After resumption of perfusion for 15 min, NTG was added to the perfusate to produce a concentration of 12.5 micro g/mL. Perfusion was continued for an additional 2 h, and final PISO and Kf,c were measured. The control group (n = 7) underwent the identical protocol but without the addition of NTG.
Isolated lungs (n = 8) were perfused and baseline measurements obtained as above. After an additional 15 min of perfusion, NTG was added to produce a perfusate concentration of 12.5 micro g/mL. Fifteen min later, oxidant lung injury was produced as previously described . Purine (0.5 mM; Sigma Chemical Co, St. Louis, MO) was added to the perfusate followed 5 min later by xanthine oxidase (0.001 U/mL Type IV, from milk; Sigma Chemical Co). The combination of purine (substrate) and xanthine oxidase (enzyme) results in the generation of superoxide radicals. This technique produces moderate lung injury over the subsequent 3 h, at which time final Kf,c and PISO measurements were obtained. The control group (n = 10) underwent the identical protocol but without the addition of NTG.
Data are presented as mean values +/- SE. Statistical analysis used analysis of variance followed by Student-Newman-Keuls tests. P < 0.05 was considered to be statistically significant.
There were no statistical differences in baseline PPA, PISO, or Kf,c between the control and NTG groups with noninjured lungs (Table 1). There was no significant change in any of these parameters between the baseline and final measurements, and there were no differences between the two groups at the final measurements. Weight gain during the final hour of perfusion did not significantly differ between the two groups.
There were no statistical differences in baseline PPA, PISO, or Kf,c between the control and NTG groups with injured lungs (Table 2). There was no significant change in PPA or PISO between the baseline and final measurements, and there were no differences between the two groups at the final measurements. Kf,c increased in both groups with no significant differences between the two groups. Weight gain during the final hour of perfusion did not significantly differ between the two groups.
In the current study, NTG did not alter vascular permeability in either the normal or the injured buffer-perfused rabbit lung. The lack of effect of NTG is not due to an insufficient dose, because this dose produces maximal pulmonary vasodilation during thromboxane-mediated pulmonary hypertension in the perfused rabbit lung . The effects of NO on vascular permeability are controversial. Similar to the current study, Addicks et al.  demonstrated that NTG and nitroprusside did not affect permeability in the isolated perfused rat mesentery. Kubes and others [8-10] have extensively studied the effects of NO in the feline small intestine. Similar to the current study, NO did not alter permeability under normal conditions. However, during injury from ischemia-reperfusion or platelet-activating factor, NO attenuated the injury-induced increase in permeability. Similar protective effects of NO have been demonstrated in endotoxin-induced intestinal injury in the rat , in revascularization-induced lung injury in the rat , and after platelet-activating factor injury in the rat . Similarly, NO decreases permeability in endothelial cell monolayers under basal conditions  and after thrombin  and hydrogen peroxide injury . Inhaled NO decreases pulmonary vascular permeability after oxidant-induced injury [17,18] and neutrophil-mediated injury  and decreases pulmonary transvascular albumin flux in patients with acute lung injury . Eppinger et al.  noted that inhaled NO decreased ischemia-reperfusion lung injury but that NO could initially exacerbate the early phase of lung injury as a result of increased oxidant injury from reactions between NO and superoxide.
However, detrimental effects of NO have been found by other investigators. Meyer and Huxley  reported that NO and drugs that increase cyclic guanosine monophosphate increase hydraulic conductance in the frog mesenteric capillary. NO appears to increase permeability in the hamster cheek pouch after pharmacologic injury [2,3], in the leukotoxin-injured rat lung , in the rat skin and paw after carrageenin , and in the guinea pig conjunctiva after pharmacologic injury .
These contradictory results may be due to differences in the species, tissues, and models studied interacting with several effects of NO. NO may directly alter vascular permeability, as seen in studies of endothelial cell monolayers in which NO relaxes the intracellular cytoskeleton and thereby decreases the size of the intercellular gaps [14,15]. As a vasodilator, NO may decrease capillary pressure, thereby decreasing edema formation without directly altering permeability. The effects of NO may differ depending upon whether studies examine the effects of endogenous basal NO (assessed by inhibition of NO synthesis), exogenous NO (studied by the use of NO donors), or endogenous stimulated NO (assessed by inhibition of NO synthesis after stimulation of inducible NO synthase during injury). In contrast to directly altering permeability, NO may directly affect the injury produced in a given model. NO modulates the function of neutrophils, which are frequently involved in in vivo injury. As a free radical, NO may promote or attenuate free radical-mediated injury. NO may react with superoxide anion and other substances to form the highly reactive peroxynitrite radical and other toxic free radicals, such as hydroxyl radicals. These compounds are toxic to endothelial cells, leading to direct damage to the basement membrane and increased permeability. Work by Mulligan et al.  suggests that NO and other biochemical products of L-arginine, may be responsible for immune complex-induced vasculature injury. On the other hand, NO may also act as a free radical scavenger, producing less toxic compounds . It is therefore not surprising that studies have produced contradictory results in the setting of lung injury.
In conclusion, this is the first study to examine the effects of NTG on vascular permeability in the perfused lung. Our results demonstrate that NTG in maximal doses does not alter permeability, either in the normal state or after oxidant-induced lung injury.
1. Meyer DJ Jr, Huxley VH. Capillary hydraulic conductivity is elevated by cGMP-dependent vasodilators. Circ Res 1992;70:382-91.
2. Ramirez MM, Quardt SM, Kim D, et al. Platelet activating factor modulates microvascular permeability through nitric oxide synthesis. Microvasc Res 1995;50:223-34.
3. Mayhan WG. Nitric oxide accounts for histamine-induced increases in macromolecular extravasation. Am J Physiol 1994;266:H2369-73.
4. Ishizaki T, Shigemori K, Nakai T, et al. Leukotoxin 9,10-epoxy-12-octadecenoate causes edematous lung injury via activation of vascular nitric oxide synthase. Am J Physiol 1995;269:L65-70.
5. Ialenti A, Ianaro A, Moncada S, Di Rosa M. Modulation of acute inflammation by endogenous nitric oxide. Eur J Pharmacol 1992;211:177-82.
6. Meijer F, Ruijter JM, Van Delft JL, Van Haeringen NJ. Nitric oxide induces vascular permeability changes in the guinea pig conjunctiva. Eur J Pharmacol 1995;284:61-7.
7. Eppinger MJ, Ward PA, Jones ML, et al. Disparate effects of nitric oxide on lung ischemia-reperfusion injury. Ann Thorac Surg 1995;60:1169-75.
8. Kubes P, Reinhardt PH, Payne D, Woodman RC. Excess nitric oxide does not cause cellular, vascular, or mucosal dysfunction in the cat small intestine. Am J Physiol 1995;269:G34-41.
9. Payne D, Kubes P. Nitric oxide donors reduce the rise in reperfusion-induced intestinal mucosal permeability. Am J Physiol 1993;265:G189-95.
10. Kurose I, Wolf R, Grisham MB, Granger DN. Modulation of ischemia/reperfusion-induced microvascular dysfunction by nitric oxide. Circ Res 1994;74:376-82.
11. Laszlo F, Whittle BJ, Moncada S. Attenuation by nitrosothiol NO donors of acute intestinal microvascular dysfunction in the rat. Br J Pharmacol 1995;115:498-502.
12. Abdih H, Kelly CJ, Bouchier-Hayes D, et al. Nitric oxide (endothelium-derived relaxing factor) attenuates revascularization-induced lung injury. J Surg Res 1994;57:39-43.
13. Filep JG, Foldes-Filep E. Modulation by nitric oxide of platelet-activating factor-induced albumin extravasation in the conscious rat. Br J Pharmacol 1993;110:1347-52.
14. Oliver JA. Endothelium-derived relaxing factor contributes to the regulation of endothelial permeability. J Cell Physiol 1992;151:506-11.
15. Westendorp RG, Draijer R, Meinders AE, van Hinsbergh VW. Cyclic-GMP-mediated decrease in permeability of human umbilical and pulmonary artery endothelial cell monolayers. J Vasc Res 1994;31:42-51.
16. Suttorp N, Hippenstiel S, Fuhrmann M, et al. Role of nitric oxide and phosphodiesterase isoenzyme II for reduction of endothelial hyperpermeability. Am J Physiol 1996;270:C778-85.
17. Kavanagh BP, Mouchawar A, Goldsmith J, Pearl RG. Effects of inhaled NO and inhibition of endogenous NO synthesis in oxidant-induced acute lung injury. J Appl Physiol 1994;767:1324-9.
18. Poss WB, Timmons OD, Farrukh IS, et al. Inhaled nitric oxide prevents the increase in pulmonary vascular permeability caused by hydrogen peroxide. J Appl Physiol 1995;79:886-91.
19. Guidot DM, Repine MJ, Hybertson BM, Repine JE. Inhaled nitric oxide prevents neutrophil-mediated, oxygen radical-dependent leak in isolated rat lungs. Am J Physiol 1995;269:L2-5.
20. Addicks K, Arnold S, Fasching P, et al. Nitric oxide mediates microvascular permeability in the isolated perfused rat mesentery? Agents Actions 1995;45:139-44.
21. Wakerlin GE JR, Benson GV, Pearl RG. A thromboxane analog increases pulmonary capillary pressure but not permeability in the perfused rabbit lung. Anesthesiol 1991;75:475-80.
22. Drake R, Gaar KA, Taylor AE. Estimation of the filtration coefficient of pulmonary exchange vessels. Am J Physiol 1978;234:H266-74.
23. Kavanagh BP, Thompson JS, Pearl RG. Inhibition of endogenous nitric oxide synthase potentiates nitrovasodilators in experimental pulmonary hypertension. Anesthesiology. In press.
24. Benzing A, Brautigam P, Geiger K, et al. Inhaled nitric oxide reduces pulmonary transvascular albumin flux in patients with acute lung injury. Anesthesiology 1995;83:1153-61.
25. Mulligan MS, Hevel JM, Marletta MA, Ward PA. Tissue injury caused by deposition of immune complexes is L-arginine dependent. Proc Natl Acad Sci USA 1991;88:6338-42.
26. Feigl EO. EDRF-a protective factor? Nature 1988;331:490-1.