Radical Scavengers Protect Murine Lungs from Endotoxin-induced Hyporesponsiveness to Inhaled Nitric Oxide

Nω -nitro-l-arginine-methyl-ester, indomethacin, bovine serum albumin, dextran, dimethylthiourea, polyethylene glycol (PEG), PEG-conjugated SOD (PEG-SOD), and PEG-conjugated catalase (PEG-catalase) were obtained from Sigma Chemical Co. (St. Louis, MO), Hanks’ balanced salt solution was obtained from GibcoBRL (Grand Island, NY), and U46619 was obtained from Cayman Chemicals (Ann Arbor, MI). EUK-8 19was generously provided by Eukarion Inc. (Bedford, MA). N -acetylcysteine was obtained from American Regent Laboratories Inc. (Shirley, NY), peroxynitrite was obtained from Upstate Biotechnology (Lake Placid, NY), E. Coli  0111:B4 lipopolysaccharide was obtained from Difco Laboratories (Detroit, MI), and NO gas was obtained from INO Therapeutics (Port Allen, LA).

The isolated mouse lung perfusion system was previously described. 8Briefly, after the mice were killed with pentobarbital sodium (200 mg/kg), they were ventilated via  an endotracheal tube, the pulmonary artery and left atrium were cannulated, and the pulmonary artery pressure (PAP) and left atrial pressure were measured and recorded continuously. Vascular perfusion was initiated at a constant flow of 50 ml · kg⁻¹· min⁻¹. The perfusate consisted of Hanks balanced salt solution with 5% bovine serum albumin and 5% dextran. Indomethacin (30 μm final concentration) and Nω -nitro-l-arginine-methyl-ester (1 mm final concentration) were added to the perfusate to inhibit endogenous prostaglandin and NO synthesis, respectively.

After stable pulmonary vasoconstriction was achieved with U46619, a thromboxane analog, a vasodilator dose–response curve to inhaled NO was obtained by sequentially ventilating the lungs with 0.4, 4, and 40 ppm NO. 8The vasorelaxation to inhaled NO (ΔPAP) was calculated as the reduction in PAP produced by inhaled NO (PAP pre-NO minus PAP after 5 min of NO inhalation) and expressed as a percentage of the increase in baseline PAP induced by infusing U46619.

RNA was extracted from lungs using the guanidine isothiocyanate-cesium chloride method, 22and 10 μg was fractionated in formaldehyde-agarose gels and transferred to nylon membranes. Membranes were hybridized with a ³²P-labeled 0.3-kb mouse NOS2  cDNA probe. 23Equal loading of RNA on gels was confirmed by staining 28S and 18S ribosomal RNA with ethidium bromide.

Nitrotyrosine concentrations in mouse lungs were estimated using semiquantitative immunoblotting. Lungs were homogenized in Laemmli buffer, 24boiled for 5 min, and centrifuged at 15,000 g  for 30 min. Soluble proteins (80 mg per lane) were fractionated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred onto nitrocellulose filters. Each gel was loaded with 0.5 μg of a nitrated bovine serum albumin standard (Upstate Biotechnology, Lake Placid, NY) to produce a positive control, with an additional positive control generated by pretreating the supernatant protein solution (1 mg/ml) with peroxynitrite (100 μl). Membranes were blocked at room temperature with 5% dry milk in phosphate-buffered saline for 30 min and incubated overnight at 4°C with rabbit polyclonal antinitrotyrosine antiserum (Biomol Research Laboratories, Plymouth Meeting, PA; diluted 1:500 in 5% milk–phosphate buffered saline with Tween 20). Bound antibody was detected using horseradish-peroxidase protein A (Boehringer Mannheim Biochemicals, Indianapolis, IN) and visualized using chemiluminescence (Renaissance Western Blot Chemiluminescence Reagent Plus; NEN Life Science Products, Boston, MA).

The investigations were approved by the Subcommittee for Research Animal Care of the Massachusetts General Hospital (Boston, MA). We studied wild-type SV129B6/F1 mice (F1-generation progeny of SV129 and C57BL/6 mice) of both sexes, weighing 16–30 g and aged 2–4 months (table 1). All drugs were administered intraperitoneally unless specified otherwise. Drugs were dissolved in saline so that the volume of each intraperitoneal injection was 0.01 ml/g body weight (except EUK-8, which was administered in 0.02 ml distilled water per gram). Endotoxin-challenged mice received 50 mg/kg body weight E. coli  0111:B4 lipopolysaccharide 16 h before lungs were isolated and perfused. 8 

Animals in group 6 were studied after prolonged exposure to inhaled NO. Mice were housed in chambers and breathed 20 ppm NO 8in air during the 16 h after intraperitoneal administration of saline, lipopolysaccharide, or lipopolysaccharide and 150 mg/kg N -acetylcysteine. Mice were studied 1 h after discontinuation of NO inhalation, allowing ample time for any vasodilator action of inhaled NO to dissipate.

After perfusion, both lungs, excluding hilar structures, were excised and weighed. Thereafter, the tissue was dried in a microwave oven for 60 min and reweighed as described previously. 8Lung wet/dry ratios were calculated.

All data are expressed as mean ± SEM. The overall responsiveness to the three doses of inhaled NO was compared between groups by analysis of variance for repeated measurements (Statistica for Windows; StatSoft, Inc., Tulsa, OK). A P  value < 0.05 indicated significance of the overall responsiveness to inhaled NO. Differences in wet/dry ratios between groups were determined using a two-tailed Student t  test.

In each murine lung perfusion, U46619 produced a stable increase of PAP that was reversible after discontinuing U46619. The dose of U46619 required to increase PAP by 5 or 6 mmHg did not differ in saline- or lipopolysaccharide-treated mice (data not shown).

Pulmonary vasorelaxation to 0.4, 4, and 40 ppm inhaled NO in saline-treated mice (table 1, group 1) was 34 ± 4%, 60 ± 5%, and 75 ± 4%, respectively (fig. 1). After endotoxin challenge, pulmonary vasorelaxation to 0.4, 4, and 40 ppm inhaled NO was 11 ± 2%, 26 ± 4%, and 45 ± 4%, respectively (P < 0.0001 vs.  saline;fig. 1).

To study the role of ROS in endotoxin-induced impairment of pulmonary vasodilation to inhaled NO, we administered in vivo  concurrent with the lipopolysaccharide challenge one of three chemically unrelated ROS scavengers, each of which possesses nonselective antioxidant activity. N -acetylcysteine is a virtually nontoxic scavenger possessing beneficial effects in ALI in humans. 25Mice were given a single intraperitoneal dose of 30, 150, or 500 mg/kg N -acetylcysteine immediately after lipopolysaccharide challenge (table 1, group 2), and lungs were isolated and perfused 16 h later. Administration of 30 mg/kg N -acetylcysteine did not alter the endotoxin-induced impairment of vasorelaxation to inhaled NO (data not shown). The pulmonary vasorelaxation produced by 0.4, 4, and 40 ppm inhaled NO in mice challenged with endotoxin and treated with 150 mg/kg N -acetylcysteine was 15 ± 2%, 47 ± 6%, and 58 ± 4%, respectively (P < 0.05 vs.  lipopolysaccharide alone). A single intraperitoneal dose of 500 mg/kg N -acetylcysteine (data not shown) did not confer greater preservation of the vasodilator response to inhaled NO than a single dose of 150 mg/kg. Similar preservation of the vasodilator response to inhaled NO was observed when 150 mg/kg N -acetylcysteine was administered subcutaneously (vs.  intraperitoneally) immediately after lipopolysaccharide challenge (data not shown).

Because N -acetylcysteine has a short plasma half-life as a result of rapid biotransformation and renal clearance, 25we also assessed pulmonary vasorelaxation after two doses of 150 mg/kg N -acetylcysteine were given 3.5 h apart (table 1, group 2). After two doses of N -acetylcysteine, pulmonary vasorelaxation to 0.4, 4, and 40 ppm inhaled NO was 25 ± 2%, 65 ± 5%, and 68 ± 4%, respectively (P < 0.01 vs.  lipopolysaccharide alone;fig. 1) and did not significantly differ from vasorelaxation in saline-challenged mice. Therefore, two doses of N -acetylcysteine resulted in greater protection than a single dose (P < 0.05 vs.  lipopolysaccharide and single N -acetylcysteine dose).

In lungs from mice treated with lipopolysaccharide followed by 330 mg/kg dimethylthiourea (table 1, group 3), the pulmonary vasorelaxation produced by 0.4, 4, and 40 ppm inhaled NO was 24 ± 3%, 55 ± 4%, and 71 ± 3%, respectively (P < 0.01 vs.  lipopolysaccharide alone, P = nonsignificant vs.  saline challenge;fig. 2). In mice treated with lipopolysaccharide followed by 30 mg/kg EUK-8, the pulmonary vasorelaxation produced by 0.4, 4, and 40 ppm inhaled NO was 19 ± 10%, 55 ± 6%, and 72 ± 3%, respectively (P < 0.01 vs.  lipopolysaccharide alone, P = nonsignificant vs.  saline challenge;fig. 2).

To examine whether fully developed impairment of inhaled NO responsiveness could be acutely reversed or ameliorated by late treatment with a ROS scavenger, mice were treated with 150 mg/kg intraperitoneal N -acetylcysteine at 15.25 h after lipopolysaccharide challenge and 45 min before they were killed (table 1, group 4). The impairment of pulmonary vasorelaxation to inhaled NO did not differ in mice challenged with lipopolysaccharide whether or not N -acetylcysteine was administered 15.25 h after lipopolysaccharide administration (data not shown).

The vasorelaxation to inhaled NO in lungs 16 h after treatment with saline followed immediately with either 150 mg/kg N -acetylcysteine, 330 mg/kg dimethylthiourea, or 30 mg/kg EUK-8 (table 1, groups 2 and 3) did not significantly differ from that of lungs of mice treated with saline alone (data not shown).

Pulmonary vasorelaxation to 0.4, 4, and 40 ppm inhaled NO in mice challenged with lipopolysaccharide followed by 20 U/g PEG-catalase administered intraperitoneally (table 1, group 5) was 12 ± 4%, 40 ± 4%, and 59 ± 4%, respectively (P < 0.05 vs.  lipopolysaccharide alone;fig. 3). In mice treated with lipopolysaccharide followed by 100 U/g PEG-catalase, the pulmonary vasorelaxation to 0.4, 4, and 40 ppm inhaled NO was 17 ± 3%, 54 ± 3%, and 71 ± 5%, respectively (P < 0.01 vs.  lipopolysaccharide alone and vs.  lipopolysaccharide plus 20 U/g PEG-catalase, P = nonsignificant vs.  saline challenge;fig. 3). This single intraperitoneal dose of PEG-catalase fully preserved the pulmonary vasodilator response to inhaled NO. Because PEG itself has antioxidant properties, 26we examined whether the PEG moiety alone was responsible for the protective effect of PEG-bound catalase observed in endotoxin-challenged lungs. PEG alone (table 1, group 5) did not protect endotoxin-challenged lungs against the impaired responsiveness to inhaled NO (data not shown).

The pulmonary vasorelaxation to 0.4, 4, and 40 ppm inhaled NO in endotoxin-challenged mice treated with 50 U/g PEG-SOD was 10 ± 3%, 35 ± 5%, and 49 ± 7%, respectively (P = nonsignificant vs.  lipopolysaccharide alone). PEG-catalase and PEG-SOD had no effect on the pulmonary vasodilator response to NO in saline-challenged control mice (table 1, group 5; data not shown).

N -acetylcysteine 27and other ROS scavengers 27,28may attenuate endotoxin-induced induction of NOS2  gene expression. A subsequent decrease of endogenous pulmonary NO concentrations could protect against endotoxin-induced impaired responsiveness to inhaled NO. 7,8To examine this possibility, murine pulmonary NOS2  mRNA concentrations were measured 6 h after intraperitoneal administration of saline, lipopolysaccharide, or lipopolysaccharide with 150 mg/kg N -acetylcysteine. This time point was chosen since murine pulmonary NOS2  mRNA concentrations are at their peak approximately 6 h after endotoxin challenge. 29,30After saline challenge, murine lungs had undetectable concentrations of NOS2  mRNA. In contrast, NOS2  gene expression was markedly elevated in lungs 6 h after challenge with lipopolysaccharide. Treatment with N -acetylcysteine did not reduce the ability of lipopolysaccharide to induce NOS2  gene expression (data not shown).

To confirm that the beneficial effects of N -acetylcysteine after endotoxin challenge are not attributable to diminution of pulmonary NO concentrations, responsiveness to inhaled NO was studied in mice that breathed 20 ppm NO for 16 h after challenge with saline, lipopolysaccharide, or lipopolysaccharide and N -acetylcysteine (table 1, group 6). As previously reported, 8prolonged NO inhalation did not alter the responsiveness to subsequent inhaled NO in either saline- or lipopolysaccharide-challenged lungs. The pulmonary vasorelaxation to 0.4, 4, and 40 ppm inhaled NO in mice that breathed 20 ppm NO for 16 h after receiving lipopolysaccharide and 150 mg/kg N -acetylcysteine was 18 ± 2%, 48 ± 6%, and 62 ± 8%, respectively (P < 0.05 vs.  lipopolysaccharide alone).

To determine whether endotoxin-induced impaired vasorelaxation to inhaled NO in mice was associated with increased formation of peroxynitrite, lungs concentrations of nitrotyrosine, a product of peroxynitrite, 31were assessed 7 h after lipopolysaccharide or saline challenge. We chose this time because previous reports suggested that nitrotyrosine could be detected in lungs within 3–12 h after lipopolysaccharide challenge 30,32and that maximum pulmonary nitrotyrosine concentrations coincided with maximal NOS2  expression. 30,33Nitrotyrosine was not detected in lungs obtained from either saline- or lipopolysaccharide-treated mice but was detected in a nitrated bovine serum albumin standard and in lung proteins pretreated with peroxynitrite (positive controls).

The lung wet/dry ratios did not differ in mice challenged with saline or lipopolysaccharide with or without treatment with a ROS scavenger. The lung wet/dry ratios were not affected by prolonged breathing of 20 ppm NO.

The novel finding of this study is that treatment with scavengers of ROS can protect murine lungs from endotoxin-induced hyporesponsiveness to inhaled NO. In accordance with several previous clinical 2and experimental studies, 6–8we found in an isolated mouse lung perfusion model that an lipopolysaccharide challenge attenuates the pulmonary vasodilatory effect of inhaled NO. When given concurrently with lipopolysaccharide, the nonselective ROS and reactive nitrogen species scavengers dimethylthiourea, N -acetylcysteine, EUK-8, or a selective H²O²scavenger, PEG-catalase, were able to preserve the pulmonary vasodilator response to inhaled NO when measured 16 h after lipopolysaccharide challenge.

We found that a single dose of EUK-8 or dimethylthiourea fully preserved murine pulmonary vasorelaxation to inhaled NO after lipopolysaccharide challenge, whereas a single dose of N -acetylcysteine was only partially protective (fig. 1). We considered the possibility that the incomplete protection of inhaled NO responsiveness achieved after a single intraperitoneal dose of N -acetylcysteine was attributable to the short half-life in mice (as has been observed in humans 25). We observed that the administration of two intraperitoneal doses of N -acetylcysteine, 3.5 h apart, fully preserved pulmonary vasorelaxation to inhaled NO 16 h after lipopolysaccharide challenge. This observation suggests that the process by which ROS impair pulmonary responsiveness to inhaled NO lasts more than a few hours after lipopolysaccharide challenge and that ROS scavengers are required over an extended time period in the early phase of endotoxemia to prevent the induction of hyporesponsiveness to inhaled NO.

We considered the possibility that ROS scavengers may accentuate the vasodilator response to inhaled NO by enhancing the activity of enzymes in the NO–cyclic guanosine monophosphate signal transduction pathway. 34This is unlikely, however, because no improvement in vasorelaxation was observed 16 h after administration of saline and ROS scavengers. We also considered the possibility that ROS scavengers potentiate the vasodilator response to inhaled NO by prolonging the biologic half-life of NO. N -acetylcysteine, for instance, may stabilize NO in the form of S -nitrosothiols, 35thereby preventing NO from reacting with O²⁻. However, no enhancement of pulmonary vasorelaxation was observed when intraperitoneal N -acetylcysteine was given 15.25 h after endotoxin-challenge and 45 min before lung isolation, although tissue concentrations of N -acetylcysteine were likely to have been elevated during perfusion. Our studies suggest that the protective effect of ROS scavengers originates from the early prevention or mitigation of lipopolysaccharide-induced functional alterations that impair NO vasorelaxation.

We recently reported that increased pulmonary NO concentrations produced by NOS2  are necessary to impair responsiveness to inhaled NO in this murine sepsis model. 8We considered the possibility that N -acetylcysteine preserved responsiveness to inhaled NO by preventing the induction of NOS2  gene expression by endotoxin. 27However, we observed that doses of N -acetylcysteine sufficient to preserve responsiveness to inhaled NO did not decrease NOS2  mRNA concentrations 6 h after lipopolysaccharide challenge. Moreover, breathing 20 ppm NO for 16 h after lipopolysaccharide challenge, which is sufficient to impair responsiveness in NOS2 -deficient mice, 8did not impair inhaled NO responsiveness in wild-type mice treated with N -acetylcysteine. Taken together, these results suggest that the protective effects of N -acetylcysteine against endotoxin-induced impairment of pulmonary inhaled NO responsiveness is not attributable to inhibition of lipopolysaccharide-stimulated increased pulmonary NO concentrations.

N -acetylcysteine, dimethylthiourea, and EUK-8 all scavenge a wide spectrum of ROS, 12,13,17,19peroxynitrite, 14,20and decrease nitrotyrosine formation by peroxynitrite. 15,18,20To further characterize the role of H²O²and O²⁻in endotoxin-induced impaired responsiveness to inhaled NO, the selective ROS scavengers polyethylene PEG-catalase and PEG-SOD, respectively, were given concomitantly with lipopolysaccharide. PEG-conjugated enzymes were studied because they have superior intracellular penetration over the native enzymes, 36,37a long plasma half-life (up to 40 h in rats), 37systemic efficacy after intraperitoneal administration, 38–40and protective effects in animal models of septic ALI. 41,42In our murine model of endotoxemia, PEG-catalase conferred a major protective effect, indicating an important role for H²O²in the impairment of inhaled NO responsiveness. PEG-SOD, in contrast, did not preserve responsiveness to inhaled NO. During the first few hours after administration, PEG-SOD had slower transmembrane penetration than PEG-catalase. 36The recommended dose of PEG-SOD in human is 1 U/g, 43and 2 U/g PEG-SOD was shown to effectively attenuate septic ALI in guinea pigs. 42We empirically chose a dose of 50 U/g PEG-SOD administered intraperitoneally. The dose we used may not have been enough to achieve a sufficiently high intracellular SOD concentration. After cellular uptake, PEG-conjugated enzymes are entrapped within endosomes or lysosomes. 36Unlike H²O², which crosses cellular membranes, O²⁻is membrane-impermeable, 10and endosomal or extracellular SOD cannot neutralize cytosolic or mitochondrial O²⁻. The overproduction of NO and O²⁻after lipopolysaccharide challenge may overwhelm the small increase in lung SOD activity conveyed by PEG-SOD, because the reaction of O²⁻with NO is much faster than its reaction with SOD. 31Augmented SOD activity may even paradoxically exacerbate tissue damage through generation of H²O²(and H²O²-derived oxidants) 10,44or by directly catalyzing tyrosine nitration. 9,31 

H²O²may induce ALI via  activation of the transcription factor NF-κB (nuclear factor-κB), which plays a central role in regulating the transcription of cytokines, adhesion molecules, and other mediators involved in endotoxemia and sepsis. 11H²O²is converted to hypochlorous acid by myeloperoxidase or decomposes through the Fenton reaction to yield hydroxyl radical (·HO). 10Recent investigations have demonstrated that H²O², hypochlorous acid and ·HO can directly activate various heme oxygenases (e.g. , myeloperoxidase) to generate peroxynitrite and other nitrating intermediates via  oxidation of nitrite and nitrate, the end products of NO metabolism. 9In the current study, nitrotyrosine was not detected by immunoblotting in either control murine lungs or 7 h after lipopolysaccharide administration. Our finding is in accordance those of with Baechtold et al. , 45who did not detect any change in nitrotyrosine content by immunoblotting in murine lungs 24 h after induction of polymicrobial sepsis. Kristof et al. , 30however, reported that lipopolysaccharide challenge in mice elicited a widespread increase in immunostaining for nitrotyrosine in lung slices (peaking at 12 h after challenge). It is possible that immunoblotting is less sensitive than immunohistochemistry, as immunostaining also detected nitrotyrosine in slices obtained from saline-challenged lungs. 30Alternatively, prolonged exposure to endotoxin may be required before detectable quantities of lung nitrotyrosine accumulates. 30Although nitrotyrosine-related oxidative modification of tyrosine, such as 3,3′-dityrosine or 3-chlorotyrosine, may alter cellular pathways in inflammatory diseases similar to nitrotyrosine, 9these molecules are not detectable by the common nitrotyrosine assays. 9,46However, our finding does not support a major role for protein tyrosine nitrosylation in impaired pulmonary vasorelaxation in endotoxemia.

Nitric oxide inhalation improves oxygenation in patients with acute respiratory distress syndrome by redistributing blood flow toward better-ventilated areas, thereby decreasing intrapulmonary shunting. 8Although hyporesponsiveness to inhaled NO importantly limits the clinical efficacy of inhaled NO in acute respiratory distress syndrome, means to treat it are unknown. Although extrapolation from a murine model to human disease requires many continuing research steps, this study suggests that treatment with scavengers of ROS might be used to preserve responsiveness to inhaled NO in ALI and thereby increase the number of patients who respond favorably to inhaled NO. Responsiveness to inhaled NO was preserved by N -acetylcysteine treatment in endotoxin-challenged mice breathing 20 ppm NO for 16 h, suggesting that when NO is inhaled by patients to treat pulmonary hypertension or to augment arterial oxygenation, antioxidants may be given concomitantly to scavenge ROS and preserve pulmonary vasodilator responsiveness to inhaled NO.

In summary, our study suggests that, when given to mice early after lipopolysaccharide challenge, nonselective ROS scavengers can protect the lungs from endotoxin-induced hyporesponsiveness to inhaled NO. The salutary effect of PEG-catalase suggests an important role for hydrogen peroxide in mediating the endotoxin-induced hyporeactivity to inhaled NO. This study suggests that a novel therapeutic approach, treatment with scavengers of ROS, may preserve responsiveness and prevent hyporesponsiveness to inhaled NO in ALI.

The authors thank Eukarion Inc., Bedford, Massachusetts, for donating EUK-8, and Dr. YuChiao Chang (Medical Practice Evaluation Center, Massachusetts General Hospital, Boston, MA) for statistical advice.