In animal models, endotoxin (lipopolysaccharide) challenge impairs the pulmonary vasodilator response to inhaled nitric oxide (NO). This impairment is prevented by treatment with inhibitors of NO synthase 2 (NOS2), including glucocorticoids and L-arginine analogs. However, because these inhibitors are not specific for NOS2, the role of this enzyme in the impairment of NO responsiveness by lipopolysaccharide remains incompletely defined.
To investigate the role of NOS2 in the development of lipopolysaccharide-induced impairment of NO responsiveness, the authors measured the vasodilator response to inhalation of 0.4, 4, and 40 ppm NO in isolated, perfused, and ventilated lungs obtained from lipopolysaccharide-pretreated (50 mg/kg intraperitoneally 16 h before lung perfusion) and untreated wild-type and NOS2-deficient mice. The authors also evaluated the effects of breathing NO for 16 h on pulmonary vascular responsiveness during subsequent ventilation with NO.
In wild-type mice, lipopolysaccharide challenge impaired the pulmonary vasodilator response to 0.4 and 4 ppm NO (reduced 79% and 45%, respectively, P < 0.001), but not to 40 ppm. In contrast, lipopolysaccharide administration did not impair the vasodilator response to inhaled NO in NOS2-deficient mice. Breathing 20 ppm NO for 16 h decreased the vasodilator response to subsequent ventilation with NO in lipopolysaccharide-pretreated NOS2-deficient mice, but not in lipopolysaccharide-pretreated wild-type, untreated NOS2-deficient or untreated wild-type mice.
In response to endotoxin challenge, NO, either endogenously produced by NOS2 in wild-type mice or added to the air inhaled by NOS2-deficient mice, is necessary to impair vascular responsiveness to inhaled NO. Prolonged NO breathing, without endotoxin, does not impair vasodilation in response to subsequent NO inhalation. These results suggest that NO, plus other lipopolysaccharide-induced products, are necessary to impair responsiveness to inhaled NO in a murine sepsis model.
INHALED nitric oxide (NO) is a selective vasodilator of human 1–3and animal pulmonary vessels. 4–7NO inhalation improves oxygenation in patients with acute respiratory distress syndrome (ARDS) by redistributing blood flow toward better ventilated lung areas, thereby reducing intrapulmonary shunting. 3However, 30–40% of ARDS patients do not respond to inhaled NO. 1,2Krafft et al. 8reported a high incidence of nonresponders (60%) in patients with sepsis-associated ARDS. Similarly, Manktelow et al. 9noted that ARDS patients with septic shock were less likely to respond to inhaled NO than were ARDS patients without septic shock. These reports suggest that endotoxemia or sepsis syndrome might impair responsiveness to inhaled NO.
Vasodilatation produced by NO is mediated primarily by stimulating soluble guanylate cyclase (sGC) to produce cyclic guanosine monophosphate (cGMP) from guanosine triphosphate (GTP). Cyclic GMP activates cGMP-dependent protein kinase, which phosphorylates several intracellular targets, resulting in smooth muscle relaxation (for review see Lincoln and Cornwell 10). Cyclic GMP is catabolized to GMP by phosphodiesterases.
Nitric oxide is produced by NOS through the conversion of L-arginine to L-citrulline in the presence of oxygen (reviewed in Knowles and Moncada 11and Moncada et al. 12). Three different NOS isoforms have been characterized. Neuronal (NOS1) and endothelial (NOS3) NOSs are expressed constitutively and produce NO in response to an increased intracellular calcium concentration. 12Transcription of the inducible NOS isoform (NOS2) is increased in response to endotoxin and cytokines, such as tumor necrosis factor α, interleukin-1β, and interleukin-6, and leads to accelerated production of NO. 11,12Excessive NO synthesis has been suggested as an important mechanism that causes systemic hypotension during septic shock. 13Moreover, NOS2 is capable of generating not only NO, but also superoxide radicals. 14NO and superoxides combine rapidly to form peroxynitrite, which can cause nitrosative injury. 15
Several authors have described alterations of NO–-cGMP-mediated vasorelaxation after endotoxin treatment. 16–18Holzmann et al. 18demonstrated impaired vasorelaxation to inhaled NO in isolated, perfused, and ventilated lungs obtained from lipopolysaccharide-treated rats and noted that inhibition of NOS enzyme activity by N G-nitro-L-arginine methyl ester (L-NAME) or aminoguanidine during the 16 h after lipopolysaccharide challenge maintained the pulmonary vasodilator responsiveness to inhaled NO. These data suggest a critical role for NOS2 in the development of endotoxin-induced hyporesponsiveness to inhaled NO.
However, these studies are limited because available NOS2 enzyme inhibitors are incompletely isoform specific. 19Furthermore, it is unknown whether NO, or other molecules produced by NOS2, such as superoxide, 15contributes to the lipopolysaccharide-induced hyporesponsiveness to inhaled NO. In a novel approach to understanding the effects of lipopolysaccharide-mediated NOS2 induction on pulmonary vascular responsiveness to inhaled NO, we studied mice with a congenital deficiency of the NOS2 gene. We report that lipopolysaccharide induces hyporesponsiveness to inhaled NO in wild-type mice, but does not produce hyporesponsiveness in NOS2-deficient mice. Furthermore, we provide evidence that NO, either endogenously produced by NOS2 or added to the ambient air inhaled by NOS2-deficient mice, is necessary to produce pulmonary vascular hyporesponsiveness to inhaled NO in the mouse undergoing endotoxin challenge.
Materials and Methods
These investigations were approved by the Subcommittee for Research Animal Care of the Massachusetts General Hospital. A total of 78 adult male mice weighing 20–35 g were studied, as listed in table 1and outlined herein. NOS2-deficient mice 20were generously provided by Dr. Carl Nathan. Mice of the same background (F1-generation of the parental strains SV129 and C57 Black/6) were used as wild-type mice. 21
Isolated, Perfused, and Ventilated Mouse Lung Model
Mice were killed by an intraperitoneal injection of pentobarbital sodium (200 mg/kg body weight) and placed in a 37°C water-jacketed chamber (Isolated Perfused Lung Size 1 Type 839; Hugo-Sachs Elektronik, March-Hugstetten, Germany). The trachea was isolated and intubated, and the lungs were ventilated with 21% O2, 6% CO2and 73% N2using a volume-controlled ventilator (model 687; Harvard Apparatus, South Natick, MA) at a ventilatory rate of 85 breaths/min and 2 cm H2O end-expiratory pressure. The tidal volume was adjusted to provide a peak inspiratory pressure of 10 cm H2O throughout each study. The lungs were exposed via a midline sternotomy, and a ligature was placed around the aorticopulmonary outflow tract. After injection of 10 IU heparin into the right ventricle, the pulmonary artery was cannulated with a stainless steel cannula (1 mm ID)via the right ventricle. The pulmonary venous effluent was drained via a stainless steel cannula (1 mm ID) placed through the apex of the left ventricle across the mitral valve and into the left atrium. Left atrial pressure was maintained at 2 mmHg. Lungs were perfused at a constant flow (50 ml · kg body weight−1· min−1; Ismatec Reglo-Analogue roller pump; Laboratoriumstechnik GmbH, Wertheim-Mondfeld, Germany) with a nonrecirculating system at 37°C. The perfusate used was Hanks’ Balanced Salt Solution (GibcoBRL, Grand Island, NY) containing 1.26 mM CaCl2, 5.33 mM KCl, 0.44 mM KH2PO4, 0.50 mM MgCl2, 0.41 mM MgSO4, 138.0 mM NaCl, 4.0 mM NaHCO3, 0.3 mM Na2HPO4, and 5.6 mM glucose. Bovine serum albumin, 5%, and dextran, 5%(both from Sigma Chemical Co., St. Louis, MO), were added to the perfusate to prevent pulmonary edema, as previously described in the isolated, perfused, and ventilated rat lung. 17Indomethacin, 30 mM (Sigma Chemical Co.), and 1 mM L-NAME (Sigma Chemical Co.) were added to the perfusate to inhibit endogenous prostaglandin and NO synthesis, respectively. Sodium bicarbonate was added to adjust the perfusate pH to 7.34–7.43. Lungs were included in this study if they had a homogenous white appearance without signs of hemostasis or atelectasis and showed a stable perfusion pressure less than 10 mmHg during the second 5 min of an initial 10-min baseline perfusion period. Using these two criteria, approximately 15% of lung preparations from each group were discarded before study.
Pulmonary artery pressure (PAP) and left atrial pressure were measured via saline-filled membrane pressure transducers (Argon, Athens, TX) connected to a side port of the inflow and outflow cannulae, respectively. Airway pressure (Paw) was measured using a differential pressure transducer (model MP-45–32–871; Validyne Engineering Corp., Northridge, CA) connected to the inspiratory limb just before the Y piece. Pressure transducers were connected to a biomedical amplifier (Hewlett Packard 7754B, Andover, MA), and data were recorded at 150 Hz on a personal computer using an analog-to-digital interface with a data acquisition system (DI-220; Dataq Instruments, Akron, OH). The system was calibrated before each experiment.
For NO inhalation, NO gas (800 or 80 ppm NO in nitrogen, Airco, Murray Hill, NJ) was blended (Oxygen Blender; Bird Corporation, Palm Springs, CA) with oxygen, carbon dioxide, and nitrogen to achieve a final concentration of 21% O2, 6% CO2, and the desired NO concentration. NO and higher oxidative states of NO (NOx; CLD 700 AL; Eco Physics, Dürnten, Switzerland), oxygen (Hudson Ventronics Division, Temecula, CA), and carbon dioxide (Datex CO2 monitor; Puritan-Bennett Corporation, Los Angeles, CA) concentrations were monitored continuously.
Pulmonary Vascular Response to Inhaled NO after Lipopolysaccharide Challenge
Wild-type and NOS2-deficient mice were injected intraperitoneally with 50 mg/kg body weight Escherichia coli 0111:B4 lipopolysaccharide (LPS; Difco Laboratories, Detroit, MI) dissolved in saline 16 h before isolated lung perfusion. This time point was chosen based on our previous studies in rats. 17Untreated wild-type and NOS2-deficient mice served as controls.
After an initial 10-min baseline perfusion period, pulmonary vasoconstriction was induced by continuous infusion of the thromboxane A2analog U-46619 (Cayman Chemicals, Ann Arbor, MI). The infusion rate was adjusted to provide a stable increase in PAP of 5 or 6 mmHg. Then, a dose–response curve to inhaled NO was obtained by sequentially ventilating the lungs with 0.4, 4, and 40 ppm NO for 5 min each. After each period of NO ventilation, the PAP was allowed to return to the pre-NO elevated baseline. U-46619 infusion was readjusted if the PAP was not within a range of ± 10% of the pre-NO value at 5 min after discontinuation of NO inhalation. The vasodilator response to inhaled NO (ΔPAP) was measured as the change in PAP produced by inhaled NO (PAP after 5 min of NO inhalation minus PAP pre-NO) as a percentage of the increase in PAP induced by U-46619 (PAP pre-NO minus PAP at initial baseline).
Effect of NO Exposure on Pulmonary Vascular Response to Inhaled NO
Four groups of mice breathed 20 ppm NO for 16 h. One group of wild-type mice and one group of NOS2-deficient mice were injected with 50 mg/kg lipopolysaccharide intraperitoneally immediately before NO exposure. Additional wild-type and NOS2-deficient mice groups were exposed to NO inhalation without receiving lipopolysaccharide. After 16 h of NO exposure, the lungs were isolated and perfused as described previously. Pulmonary vasoconstriction was induced by infusion of U46619, and the vasodilator response to 0.4, 4, and 40 ppm NO was measured.
During ambient-pressure NO exposure, animals were maintained in 40-l acrylic chambers. NO and NOxconcentrations were controlled carefully using soda lime 22at a high fresh gas flow rate of NO (10,000 ppm NO in nitrogen; Airco, Murray Hill, NJ), air, and oxygen, as previously described. 23
Two additional groups of NOS2-deficient mice were treated with lipopolysaccharide (50 mg/kg intraperitoneal) and then exposed to 0.2 and 2 ppm NO inhalation, respectively. Sixteen hours later, isolated lung perfusion studies measuring the degree of pulmonary vasodilatation produced by 0.4, 4, and 40 ppm inhaled NO were performed.
Pulmonary Vascular Response to 8-pCPT-cGMP
Nitric oxide synthase 2–deficient and wild-type mice were injected with lipopolysaccharide intraperitoneally, and, 16 h later, isolated lung perfusion was initiated as described previously. Other groups of wild-type and NOS2-deficient mice were studied without receiving lipopolysaccharide. U46619 was used to increase the baseline PAP by 5 or 6 mmHg. Lungs were then ventilated with 4 ppm NO for 5 min to evaluate vascular responsiveness to inhaled NO. After the PAP was allowed to increase to the baseline pressure, lungs were perfused sequentially with 2 and 20 mM 8-(4-chlorophenylthio)-guanosine-3′:5′-cyclic monophosphate (8-pCPT-cGMP; Biolog Life Science Institute, La Jolla, CA) for 10 min. 8-pCPT-cGMP was diluted with perfusate to reach the desired concentrations in two additional reservoirs before each experiment. This allowed immediate switching between perfusion with or without 8-pCPT-cGMP without discontinuing the flow of perfusate.
Wet-to-dry Lung Weight Ratio
At the end of each experiment, both lungs, excluding hilar structures, were excised and weighed (wet weight). Thereafter, the lungs were dried in a microwave oven for 60 min, as previously described, 17and then reweighed (dry weight). Wet-to-dry lung weight ratios were calculated by dividing the wet weight by the dry weight.
All data are expressed as the mean ± standard error (SE). To compare groups, a two-way analysis of variance was performed. When significant differences were detected by analysis of variance, a post hoc least significant difference test for planned comparisons was used (Statistica for Windows; StatSoft, Inc., Tulsa, OK). Statistical significance was assumed at a P value < 0.05.
Infusion of U46619 caused a stable increase of the PAP at a constant perfusate flow, which was reversible after discontinuing U46619 at the end of the experiment. The dose of U46619 necessary to increase the PAP by 5 or 6 mmHg did not differ in lipopolysaccharide-pretreated and untreated wild-type and NOS2-deficient mice.
Mice injected with intraperitoneal lipopolysaccharide had piloerection, diarrhea, and lethargy to a similar degree in both wild-type and NOS2-deficient mice. The mortality rate 16 h after lipopolysaccharide injection was approximately 15% and did not differ between the two mouse strains.
Pulmonary Vascular Response to Inhaled Nitric Oxide
Inhalation of NO decreased the PAP in a dose-dependent manner in all groups. A representative example of an original recording of PAP and left atrial pressure from an isolated–perfused mouse lung is provided in figure 1.
In the isolated–perfused lungs of wild-type mice that underwent lipopolysaccharide challenge, PAP decreased 79% and 45% less in response to 0.4 and 4 ppm inhaled NO, respectively, compared with untreated animals (P < 0.001;fig. 2A). The pulmonary vasodilator response to 40 ppm NO did not differ between these groups.
Response to inhaled NO in untreated NOS2-deficient mice did not differ from that of untreated wild-type mice. In contrast, lungs obtained from lipopolysaccharide-challenged NOS2-deficient mice showed greater vasodilatation to inhaled NO than the lungs of lipopolysaccharide-treated wild-type mice (P < 0.001 at each NO dose;fig. 2B). Moreover, NO-induced vasodilatation was enhanced in lipopolysaccharide-treated NOS2-deficient mice, compared with untreated NOS2-deficient or wild-type mice (P < 0.05, respectively, at each NO dose;fig. 2B).
Pulmonary Vascular Response to Inhaled NO after Inhaled NO Exposure
To investigate the role of molecular NO in the development of hyporesponsiveness to inhaled NO, we studied lipopolysaccharide-treated and untreated NOS2-deficient and wild-type mice that breathed air supplemented with 20 ppm NO for 16 h. Previous NO inhalation exposure did not alter the responsiveness to subsequently inhaled NO in perfused lungs obtained from untreated wild-type or NOS2-deficient mice or in lipopolysaccharide-pretreated wild-type mice. In contrast, the pulmonary vasodilator response to inhaled NO was decreased in lipopolysaccharide-pretreated NOS2-deficient mice exposed to ambient NO for 16 h, compared with non-NO–exposed lipopolysaccharide-pretreated NOS2-deficient mice. In isolated–perfused lungs from NOS2-deficient mice exposed to 20 ppm ambient NO for 16 h, the subsequent vasodilator responsiveness to inhaled NO was impaired after pretreatment with lipopolysaccharide (vs. untreated controls) at 0.4 (ΔPAP −24 ± 4%vs. −42 ± 4%;P < 0.05) and 4 ppm NO (ΔPAP −39 ± 5%vs. −58 ± 4%;P < 0.01), but not at 40 ppm NO (ΔPAP −55 ± 3%vs. −62 ± 5%;P = not significant;fig. 3). Similar to animals without previous NO inhalation exposure, NO-induced vasodilation was reduced in lipopolysaccharide-pretreated wild-type mice, compared to untreated wild-type mice that had breathed 20 ppm NO for 16 h before lung perfusion experiments (fig. 3).
To determine whether the inhalation of a lower level of NO for 16 h would impair vasoreactivity to short-term NO inhalation during lung perfusion, lipopolysaccharide-treated NOS2-deficient mice were exposed to 0.2 and 2 ppm NO inhalation. Breathing 0.2 ppm NO for 16 h after lipopolysaccharide administration did not cause subsequent hyporesponsiveness to short-term inhaled NO in NOS2-deficient mice. However, breathing 2 ppm NO for 16 h decreased the vasodilator response to 0.4 ppm NO inhalation, compared with the response in control mice (P < 0.05;fig. 4), but there was no reduction of vasodilator responsiveness with 4 and 40 ppm inhaled NO.
Pulmonary Vascular Response to 8-pCPT-cGMP
We investigated whether the altered pulmonary vasodilator response to inhaled NO after lipopolysaccharide challenge is associated with an impaired vasodilator response to cGMP. The vasodilator effect of the membrane-permeable, phosphodiesterase-resistant cGMP analog 8-pCPT-cGMP was studied in wild-type and NOS2-deficient mouse lungs with and without 16 h of previous lipopolysaccharide challenge. In preliminary studies, it was observed that 10 min of lung perfusion with 8-pCPT-cGMP was necessary to achieve stable vasodilation (data not shown).
The vasodilation produced by perfusing isolated lungs with 2 and 20 mM 8-pCPT-cGMP was reduced 63 and 32%, respectively, in lipopolysaccharide-pretreated wild-type mice as compared to untreated wild-type mice (P < 0.05;fig. 5A). In contrast, in NOS2-deficient mice, exposure to lipopolysaccharide did not alter 8-pCPT-cGMP–induced vasorelaxation (fig. 5B). Moreover, after lipopolysaccharide challenge, the pulmonary vasodilator response to 8-pCPT-cGMP was greater in NOS2-deficient mice (fig. 5B) than in wild-type mice (fig. 5A;P < 0.05 at both 2 and 20 mM).
Wet-to-dry Lung Weight Ratios
The absence of pulmonary edema was confirmed by unchanged wet-to-dry lung weight ratios after perfusion. There was no difference between lipopolysaccharide-pretreated wild-type (wet weight–dry weight: 4.3 ± 0.2) and NOS2-deficient (4.8 ± 0.1) mice, or untreated wild-type (4.6 ± 0.1) and untreated NOS2-deficient (4.6 ± 0.1) mice. Exposure to NO inhalation for 16 h did not alter the wet-to-dry lung weight ratios in lipopolysaccharide-pretreated wild-type (4.9 ± 0.1) and NOS2-deficient (5.1 ± 0.2) mice or in untreated wild-type (4.5 ± 0.3) and untreated NOS2-deficient (4.8 ± 0.1) mice, compared with unexposed mice. Wet-to-dry lung weight ratios did not correlate with the vasodilator response to inhaled NO.
The principal finding of this study is that the congenital absence of the gene encoding for NOS2 in mice completely prevents the development of lipopolysaccharide-induced hyporesponsiveness of the pulmonary vasculature to inhaled NO. In NOS2-deficient mice, the addition of 2 or 20 ppm inhaled NO to ambient air for 16 h after lipopolysaccharide challenge decreased pulmonary vasodilator responsiveness to short-term inhaled NO. Therefore, in mice, the NO molecule, either produced endogenously by NOS2 or added exogenously to ambient air, is necessary to produce the lipopolysaccharide-induced impairment of pulmonary vascular responsiveness to inhaled NO.
Hyporesponsiveness to Inhaled NO
Inhaled NO decreases pulmonary vascular resistance and improves oxygenation in patients with ARDS. Unfortunately, approximately 30–40% of patients with ARDS do not respond to inhaled NO therapy. 1–3The precise mechanisms responsible for this variability in the clinical therapeutic response to inhaled NO are unknown. Manktelow et al. 9reported that patients with ARDS associated with sepsis are less likely to respond to inhaled NO than are patients with ARDS associated with other disease processes. Studies of isolated rat pulmonary artery ring preparations 16and isolated–perfused rat lungs 17have showed that the administration of lipopolysaccharide can impair NO-mediated vasodilatation. In this study, we observed that pretreatment with lipopolysaccharide impairs vasodilator responsiveness to inhaled NO in an isolated–perfused mouse lung model (fig. 2A). In contrast to previous studies in rats, 18this impaired response is reflected by a rightward shift of the inhaled NO dose–pulmonary vasodilator response curve, rather than by a change in maximal effectiveness because the vasodilatory response to 40 ppm inhaled NO did not differ, with or without lipopolysaccharide-treatment.
Nitric Oxide Synthase 2 and Hyporesponsiveness to NO
Studies of isolated aortic rings 24and isolated–perfused lungs 18showed that lipopolysaccharide-mediated vasodilator hyporesponsiveness to NO can be partially prevented by agents that inhibit lipopolysaccharide-induced NOS2-synthesis, such as cycloheximide 24and dexamethasone, 18,24and by an inhibitor of NOS2 enzyme activity, aminoguanidine. 18These studies suggest a role for NOS2 in the development of hyporesponsiveness to inhaled NO. Definition of the precise mechanism is limited by the lack of specificity of the inhibitory drugs for NOS2. 19Therefore, we studied mice with a congenital absence of the NOS2 gene 20and developed an isolated, perfused, and ventilated mouse lung model to investigate the role of NOS2 in the development of lipopolysaccharide-induced hyporesponsiveness to inhaled NO.
Nitric oxide synthase 2 deficiency prevented lipopolysaccharide-induced hyporesponsiveness to inhaled NO (fig. 2A and B). Therefore, the expression of NOS2 is necessary for the production of lipopolysaccharide-mediated hyporesponsiveness to inhaled NO. Moreover, there was greater vasodilatation in response to inhaled NO in lipopolysaccharide-pretreated NOS2-deficient mice compared with untreated NOS2-deficient and wild-type mice (fig. 2). The mechanism responsible for the lipopolysaccharide-induced vasodilator hyper-responsiveness to inhaled NO in NOS2-deficient mice is unknown and needs further investigation.
Hinder et al. 25reported that 40 ppm NO inhalation was an effective pulmonary vasodilator in septic sheep administered L-NAME. This sheep study did not evaluate lower doses of inhaled NO and the effect of sepsis on NO responsiveness. Therefore, it cannot be directly contrasted with our murine studies.
Pulmonary Vascular Response to Inhaled NO after 16 h of Ambient NO Exposure
Because NOS2 was necessary for the development of lipopolysaccharide-induced hyporesponsiveness to inhaled NO, we considered the possibility that a product of NOS2, other than NO, could impair NO pulmonary vasodilator responsiveness. For example, superoxide generation by NOS2 has been reported in cells depleted of L-arginine. 15To learn whether NO itself is an essential factor in the mechanism leading to impaired responsiveness to inhaled NO, NOS2-deficient and wild-type mice, with and without lipopolysaccharide-pretreatment, breathed 20 ppm NO in air for 16 h before lung perfusion studies. Sixteen hours of NO inhalation did not affect the vasodilator response to short-term NO inhalation in untreated wild-type and NOS2-deficient mice, nor did it induce any further decrease of responsiveness to inhaled NO in lipopolysaccharide-pretreated wild-type mice (fig. 3). These results are consistent with the observations of other investigators, which showed that prolonged NO inhalation does not alter the pulmonary vasodilatation produced by administration of an NO donor compound 26,27or by inhalation of NO. 28In contrast, vasodilator response to short-term NO inhalation markedly decreased in lipopolysaccharide-pretreated NOS2-deficient mice breathing 20ppm NO in air for 16 h (fig. 3). This effect was dose-dependent: breathing 0.2 ppm NO for 16 h did not alter NO-induced vasodilatation, whereas breathing 2 ppm NO for 16 h decreased the short-term pulmonary vasodilator response to a subsequent challenge with 0.4 ppm NO, but not with 4 and 40 ppm (fig. 4). Therefore, it is clearly NO, and not another product of NOS2, that is a cofactor in pulmonary vasodilator hyporesponsiveness associated with lipopolysaccharide challenge. If we can extrapolate the known data from mice to humans, our observations suggest that, in some patients with septic ARDS, NO inhalation at 2 to 20 ppm may impair pulmonary vascular responsiveness to inhaled NO. The finding that a low concentration of inhaled NO did not decrease NO-responsiveness in NOS2-deficient mice supports the wide-spread clinical practice of using the lowest concentration of inhaled NO necessary to achieve the desired therapeutic effect. 29
Our data show that NO production by NOS2 is necessary for the development of lipopolysaccharide-induced hyporesponsiveness to NO. However, the observation that prolonged inhalation of NO does not alter the pulmonary vasodilator response to short-term NO inhalation in mice not treated with lipopolysaccharide suggests that NO alone does not account for the development of NO hyporesponsiveness in lipopolysaccharide-treated animals. Therefore, in addition to the induction of NO synthesis, other lipopolysaccharide-associated factors appear to be necessary for the development of lipopolysaccharide-mediated NO hyporesponsiveness. Ungureanu-Longrois et al. 30observed in isolated ventricular myocytes that induction of NOS2 was necessary, but was not sufficient, to cause cytokine-induced contractile dysfunction. NO-independent mechanisms that may contribute to lipopolysaccharide-mediated hyporesponsiveness to inhaled NO include the induction of cytokines and of enzymes responsible for superoxide radical production, such as xanthine oxidase or NADPH oxidase. 31It is possible that the reaction of NO with superoxide, by leading to the production of highly reactive peroxynitrite, may impair pulmonary vascular NO responsiveness. Future therapeutic approaches to improve pulmonary vascular responsiveness to inhaled NO should include the identification and modulation of these NOS2–NO-independent lipopolysaccharide-induced factors.
Mechanisms of Endotoxin-induced Hyporesponsiveness to Inhaled NO
Lipopolysaccharide-induced hyporesponsiveness to inhaled NO is correlated with an impaired pulmonary vasodilator responsiveness to the phosphodiesterase (PDE)–resistant cGMP analog 8-pCPT-cGMP (fig. 5). This suggests that the effects of lipopolysaccharide on NO signal transduction in isolated–perfused mouse lungs are, at least in part, independent of changes in cGMP synthesis or metabolism.
Our results differ from those of Fullerton et al., 16who found that, in pulmonary artery rings isolated from rats exposed to lipopolysaccharide for 6 h, the vasodilator response to the PDE-sensitive cGMP analog 8-Bromo-cGMP was preserved, whereas vasorelaxation in response to an NO donor compound was impaired. These investigators suggested that the lipopolysaccharide-induced impairment of NO-dependent vasodilatation was attributable to decreased NO-stimulated cGMP synthesis. Holzmann et al. , 17in an isolated–perfused rat lung model, reported that lipopolysaccharide pretreatment decreased the pulmonary vasodilator response to PDE-sensitive 8-Bromo-cGMP, but not to PDE-insensitive 8-pCPT-cGMP, suggesting that lipopolysaccharide-induced NO hyporesponsiveness was caused by increased pulmonary cGMP–PDE activity. Possible explanations for these differing results include differences in the species studied, the dose of lipopolysaccharide administered, the duration of exposure to lipopolysaccharide, and the experimental technique (i.e. , isolated vessel preparation vs. whole-organ perfusion).
It has been proposed that neutrophils mediate the endotoxin-induced impairment of the pulmonary vasodilator response to NO. 32Kristof et al. 33recently reported that endotoxin-induced pulmonary injury was reduced in NOS2-deficient mice compared with wild-type mice. They observed that, in wild-type mice, intraperitoneal injection of lipopolysaccharide (25 mg/kg) induced interstitial leukocyte infiltration, air-space cellularity, and exudation, associated with increased wet-to-dry lung weight ratios and increased nitrotyrosine immunostaining (reflecting peroxynitrite production). These changes were less marked or were absent in lipopolysaccharide-treated NOS2-deficient mice. In contrast, Hickey et al. 21found that, in response to a lower dose of lipopolysaccharide (30 μg/kg intravenous), recruitment of leukocytes into the lungs of NOS2-deficient mice was greater than into the lungs of wild-type mice. Differences in the findings reported in these two studies may be attributable to differences in the dose of lipopolysaccharide used. The finding that the administration of high doses of lipopolysaccharide induced interstitial swelling and exudation into air spaces raises the possibility that impaired responsiveness to inhaled NO was caused by decreased diffusion of NO from the alveoli to the pulmonary vasculature. Although we used a higher dose of lipopolysaccharide in our study (50 mg/kg intraperitoneal), wet-to-dry lung weight ratios did not differ in lipopolysaccharide-treated wild-type and NOS2-deficient mice from untreated mice of either genotype, suggesting the absence of lipopolysaccharide-induced pulmonary edema. Differences in the degree of endotoxin-induced pulmonary injury between our study and that of Kristof et al. 33may reflect differences in the strain of Escherichia coli lipopolysaccharide used. We observed that lipopolysaccharide from differing sources and lot numbers vary dramatically in the ability to induce pulmonary NOS2 gene expression (unpublished data). Moreover, the observation that vasodilation in response to the addition of 8-pCPT-cGMP to the perfusate was impaired in the lungs of lipopolysaccharide-treated wild-type mice provides additional evidence that hyporesponsiveness to inhaled NO is unlikely to be solely attributable to reduced diffusion of gaseous NO into the vasculature.
It is also unlikely that lipopolysaccharide-induced hyporesponsiveness to inhaled NO in wild-type mice is caused by nonspecific dysfunction of the pulmonary vascular contractile apparatus because lipopolysaccharide did not reduce the ability of U46619 to induce pulmonary vasoconstriction. Moreover, discontinuation of U46619 infusion resulted in prompt vasorelaxation of isolated–perfused lungs obtained from lipopolysaccharide-treated and untreated wild-type and NOS2-deficient mice.
Lipopolysaccharide-mediated development of pulmonary vascular hyporesponsiveness to inhaled NO was prevented in mice by targeted disruption of the NOS2 gene. Breathing 20 ppm NO in ambient air for 16 h in lipopolysaccharide-pretreated NOS2-deficient mice impaired NO-mediated vasodilation, suggesting that NO produced by NOS2 (or inhaled as a supplement in air) is essential for producing lipopolysaccharide-induced NO hyporesponsiveness. Prolonged NO inhalation alone did not produce NO hyporesponsiveness and must be accompanied by another lipopolysaccharide-mediated, NOS2–NO-independent inflammatory mediator or byproduct to produce hyporesponsiveness to inhaled NO.
The authors thank Jonathan Hromi for technical help, and YuChiao Chang, Ph.D., Medical Practice Evaluation Center, Massachusetts General Hospital, for statistical advice.