© 2002 American Thoracic Society
Aerosolized Linear Polyethylenimine-Nitric Oxide/Nucleophile Adduct Attenuates Endotoxin-induced Lung Injury in SheepDepartment of Anesthesiology, Faculty of Medicine, University of Tromsø, Tromsø, Norway; and Inotek Corporation, Beverly, Massachusetts Correspondence and requests for reprints should be addressed to Lars J. Bjertnaes, M.D., Ph.D., Department of Anesthesiology, Faculty of Medicine, University of Tromsø, N-9037 Tromsø, Norway. E-mail: lars.bjertnaes{at}unn.no
Pulmonary hypertension and edema are mainstays of acute lung injury (ALI). We synthesized linear polyethylenimine-nitric oxide/nucleophile adduct (DS-1), a water-soluble nitric oxide donor, and demonstrated that it is a potent relaxant of precontracted rat aortic rings without inducing desensitization. Moreover, DS-1 does not suppress the viability of human pulmonary epithelial cells in vitro. We also tested whether DS-1 counteracts ALI in endotoxemic sheep. Animals were instrumented for a chronic study. In 16 awake, spontaneously breathing sheep, Escherichia coli endotoxin (10 ng/kg/minute) was infused for 8 hours. From 2 hours of endotoxemia, sheep received either nebulized DS-1 (1 mg/kg/hour) or isotonic saline. DS-1 reduced endotoxin-induced rises in pulmonary arterial and microwedge pressures and vascular resistance index by 40–70%. In parallel, DS-1 decreased the accumulation of extravascular lung water by 60–70% and reduced the increment in right ventricle stroke work index and the falls in right ventricle ejection fraction, stroke volume, and left ventricle stroke work indices. Furthermore, DS-1 reduced venous admixture and improved arterial oxygen saturation. In four healthy animals, DS-1 alone slightly increased arterial oxygenation but had no other effects. Thus, aerosolized DS-1 attenuates endotoxin-induced ALI in sheep by reducing pulmonary hypertension and edema and improving myocardial function and gas exchange.
Key Words: acute lung injury • pulmonary circulation • extravascular lung water • nitric oxide donor • endotoxin
Activation of inflammatory mediators, pulmonary hypertension, lung edema, and deteriorated gas exchange are hallmarks of acute lung injury (ALI) (1). Nitric oxide (NO), a key regulator molecule of vascular tone and integrity, may modify these responses (2). Inhalation of gaseous NO alleviates ALI by inhibiting inflammation, selectively dilating the pulmonary vasculature, reducing lung fluid filtration, increasing perfusion of properly ventilated lung areas, and subsequently, improving gas exchange (3–6). However, despite these favorable effects, it has not been unequivocally established that inhaled gaseous NO improves the outcome of ALI (3, 7). Conversely, NO may exert detrimental effects. During inhalation, NO may react with oxygen in the airways to form nitrites and nitrates (NOX), and with deoxygenated hemoglobin to produce methemoglobin. When administered at high concentrations and for a prolonged period of time, NO may increase the generation of peroxynitrite and other toxic oxygen species that cause intra-alveolar activation of coagulation system, induces pulmonary fibrosis, and results in lung damage (2, 3, 8, 9). In addition, abrupt withdrawal of NO gas is associated with cardiopulmonary deterioration and worsening of gas exchange (10). To prevent these adverse effects, administration of gaseous NO requires special delivery and monitoring systems restricting its use to intensive care units only (11). An alternative to inhalation of gaseous NO is to administer an agent belonging to a new family of NO-releasing prodrugs. Compounds known as NO/nucleophile adducts (NONOates) are formed when NO reacts with various nucleophiles (12). These agents can be instilled into the airways or inhaled as an aerosol. Reaching the airway mucosa, NONOates release small amounts of NO, which diffuses into the pulmonary vasculature. Thereby, NONOates may reduce pulmonary hypertension and improve arterial oxygenation (13–18); however, if absorbed by the bronchial circulation, which is a part of the systemic circulation, NONOates may, in addition, produce arterial hypotension, like other nitrosylated vasodilators (15, 19, 20). Whether NONOates possesses adverse effects other than gaseous NO remains unsettled. Moreover, no data are available regarding the action of NO donors on extravascular lung water (EVLW) accumulation and myocardial performance in ALI. Here we describe the synthesis of a water-soluble and biodegradable NO-releasing compound, designated as linear polyethylenimine-NONOate (L-PEI-NONO or DS-1). We also evaluated its vasodilatory and cytotoxic properties in vitro, as well as its cardiopulmonary effects, when administered in a continuous aerosolization regimen in healthy sheep and in sheep with endotoxin-induced ALI.
In Vitro Studies We synthesized L-PEI-NONO and evaluated its effects in isolated aortic rings of rats and cultured human epithelial cells as described in the online data supplement.
In Vivo Studies
Measurements and Samples
Experimental Protocol After euthanasia at 8 hours, lung samples were taken for determination of myeloperoxidase (MPO) and nitrotyrosine activities (22, 23). Postmortem EVLW and wet-to-dry lung weight ratio were determined by the gravimetric technique (24, 25).
Statistical Analysis
Vascular and Cellular Actions of DS-1 We first evaluated the vascular effects of DS-1 in isolated rat aortic rings in vitro (see online data supplement for Tables E1–E3). After the equilibration period, rings were precontracted with epinephrine 3 x 10-6 M, and relaxation responses to DS-1 or decomposed DS-1, both at concentrations of 10-4 to 3 x 10-1 mg/ml, were measured. DS-1 caused a dose-dependent relaxation of precontracted aortic rings in contrast to decomposed compound, which was not effective (Figure 1A). In a separate set of experiments in precontracted aortic rings, DS-1 was added up to 3 x 10-3 mg/ml to achieve approximately half-maximal relaxation followed by washout for 15 to 20 minutes. This cycle was repeated three times to establish whether there was any desensitization to the vasorelaxant effect of the compound. There was no desensitization observed to the vasorelaxant effect of DS-1 (Figure 1B). In another set of experiments, we evaluated the recovery time from the relaxation after DS-1 addition up to 3 x 10-3 or 6 x 10-3 mg/ml. In approximately 150–180 minutes, a complete recovery was observed from the vasorelaxant effect of DS-1 (Figure 1C).
In addition, we assessed the effect of DS-1 on the viability of human pulmonary epithelial cells. We found that DS-1 or its carrier did not exert toxic effects to the pulmonary epithelial cells in vitro, as evaluated in the concentration range of 1 to 500 µg/ml (data not shown).
Effects of DS-1 in Endotoxemic and Healthy Sheep
Table E1 and Figure 4 display that DS-1 reduced the endotoxin-induced tachycardia and falls in stroke volume index, left ventricular stroke work index, and right ventricle ejection fraction (p < 0.05). In addition, DS-1 attenuated the rise in right ventricular stroke work index (p < 0.05). At 1 hour, cardiac index declined in both endotoxemic groups (p < 0.01). In the ET group, the cardiac index also decreased transiently at 4 hours in comparison with intragroup baseline (p = 0.04). In both endotoxemic groups, mean systemic arterial pressure and systemic vascular resistance index declined toward the end of experiment, whereas right atrial pressure and left atrial pressure rose without intergroup differences.
As depicted in Table E2, DS-1 attenuated moderate hyperventilation and changes in pH and PaCO2 after endotoxin (p < 0.05). From 3 to 4 hours, DS-1 increased SaO2 and reduced the increment in venous admixture by 60 to 70%, as compared with the ET group (p < 0.04; Figure 5). In parallel, DS-1 tended to increase PaO2 (p = 0.06) and attenuated the rise in alveolar–arterial oxygen tension difference (p = 0.039) in comparison with the ET group (Table E2). At 4 hours, DS-1 also increased oxygen venous saturation (SvO2) (p < 0.05). The body temperature and hemoglobin concentration rose throughout the experiment, demonstrating no intergroup differences (data not shown). Methemoglobin remained unchanged from baseline in both endotoxemic groups. In contrast, plasma NOX increased in the ET group and even to a greater extent in the ET + DS-1 group (p < 0.05).
As compared with the ET group, DS-1 prevented the rise in the level of MPO and decreased the nitrotyrosine score after endotoxin (p < 0.04; Figure 6). When given alone to healthy sheep, DS-1 slightly decreased venous admixture and increased SaO2 and plasma NOX (p < 0.05) but had no further effects on hemodynamics and gas exchange (Table E3).
This study demonstrates that the novel NO donor, compound DS-1, produces a dose-dependent relaxation of precontracted rat aortic rings without inducing desensitization and does not suppress the viability of human pulmonary epithelial cells in vitro. In endotoxemic sheep studied awake, inhaled aerosolized DS-1 counteracts ALI and improves myocardial function. Moreover, inhalation of DS-1 alone to healthy sheep does not cause any evident cardiopulmonary effects. Before investigating the cardiopulmonary effects of DS-1 during endotoxemia in sheep to model the changes occurring in human ALI, we assessed its effects on vascular reactivity in vitro. This part of the study demonstrates that the vasodilation of precontracted rat aortic rings by the active form of DS-1 resides in the 1–3 µg/ml range, has a dose-dependent manner, and is likely more potent in comparison with other NONOates (18). The vasodilatory effect lasts 150–180 minutes and is believed to result from relaxation of vascular smooth muscle on slow local release of NO and, subsequently, of cyclic guanosine 3'-5' monophosphate (12). The repeated exposition to DS-1 did not induce desensitization of its vasorelaxant action. In sheep, endotoxin triggers the release of arachidonic acid metabolites, cytokines, and reactive oxygen species from endothelial cells, macrophages, and various immune competent blood cells, causing a two-phase pattern of lung responses. During the early phase, thromboxane A2 and later endothelin contribute to an increase in pulmonary vascular resistance. In the late phase, the increase declines gradually secondary to reduced generation of vasoconstrictors and increased endogenous production of NO (4, 21, 26, 27). The pulmonary hypertension, the increase in EVLW, and the deterioration of gas exchange as observed in this investigation are all characteristic features of ALI in humans (1). Furthermore, these changes were accompanied by endotoxin-induced myocardial dysfunction and systemic vasodilation toward the end of the experiment. Recent investigations have shown that short-term nebulization or tracheal instillation of NO prodrugs may bring about a modest and transient reduction of increments in pulmonary vascular resistance (13–18). In this study, DS-1 was administered continuously as an inhaled nebulized aerosol to mimic inhalation of gaseous NO that is often used as an adjuvant treatment of ALI in clinical practice (3, 7). As a result, continuous inhalation of aerosolized DS-1 markedly alleviated pulmonary hypertension, as evidenced by a decline in pulmonary arterial pressure, pulmonary microwedge pressure, and PVRI for the entire administration period. The latter effects are similar to those obtained with gaseous NO (3, 4, 20) and confirm the vasodilation properties of DS-1 that was observed during the in vitro part of our study. Thus, continuous administration may significantly increase the efficacy of nebulized NONOates for control of pulmonary hypertension.
In addition to inducing pulmonary hypertension, endotoxin damages endothelium and increases pulmonary capillary permeability, thereby facilitating the development of lung edema (28). In this study, DS-1 diminished the severity of pulmonary edema, as assessed with EVLW and wet-to-dry lung weight ratio, primarily by reducing pulmonary microvascular pressure. The latter finding is consistent with the action of inhaled NO (3–5). Moreover, recent investigation on endotoxemic sheep has revealed that the inhalation of NO may also decline the enhanced microvascular permeability of pulmonary endothelium (4). The latter effect may be caused by inhibition of nuclear factor- The endotoxin-induced pulmonary vasoconstriction and myocardial depression are often accompanied by right heart failure (28). In this study, myocardial dysfunction is manifested by falls in right ventricle ejection fraction, stroke volume index, and left ventricular stroke work index occurring in parallel with increased right ventricular stroke work index and tachycardia. Inhalation of nebulized DS-1 attenuated these changes. We interpret the favorable effect of DS-1 on heart rate and cardiac performance as a result of a decreased right ventricle afterload secondary to pulmonary vasodilation after the release of NO. The latter explanation is also in agreement with several reports on inhaled NO (33–37). In this investigation, endotoxin caused an increase in intrapulmonary shunting and arterial hypoxemia, thereby triggering a moderate hyperventilation. DS-1 reduced hyperventilation, venous admixture, and alveolar–arterial oxygen tension difference and enhanced SaO2 and SvO2. These results are consistent with some previous investigations of aerosolized NO donors (16, 38) and of inhaled gaseous NO (4, 5, 20, 37) in ALI. Selective pulmonary vasodilation and increased blood flow through well-ventilated areas are believed to improve the ventilation–perfusion relationship of the lungs and be responsible for the favorable effects of NO (3). In addition, NO may improve gas exchange by inhibiting lung inflammation, attenuating endothelial dysfunction, and reducing pulmonary edema (5, 6, 29, 32). In the group of sham-operated sheep, DS-1 also transiently improved arterial oxygenation in comparison to baseline. The latter effect might be caused by redistribution of pulmonary blood flow from atelectatic areas. Such areas, probably arising from the surgical preparation, have been found at postmortem morphologic examination of the sheep. We designed DS-1 specifically to target the lung via the inhalational route. DS-1 is a 35 kD polymer, in which the carrier, L-PEI, is covalently bonded to NO. We reasoned that the intact DS-1 polymer would be sterically hindered from transmucosal flux and maintained within the airways. After NO release from the macromolecule, however, the resultant compound, L-PEI (molecular weight of 22 kD), could permeate the lung. This scenario would provide for selective vasodilation of the pulmonary vasculature, as the prodrug molecule would not pass into the systemic circulation until such time as NO had been released. Furthermore, it would ensure that the spent DS-1 would not accumulate in the lung but would be systemically absorbed and excreted. The linear polyamine structure of L-PEI has been widely used as a gene transfer vector in investigations of in vivo gene therapy (39–41). Animal studies have shown that even high doses of L-PEI given during inhalation for 21 days cause no pulmonary toxicity or immune response (39). Our results from the in vitro part of this study also demonstrate that L-PEI, or decomposed DS-1, neither alter vascular tone nor exert toxic effects to pulmonary epithelial cells. Taken together, these data suggest that chronic exposure to L-PEI is safe, and the compound can be used as a carrier of NO. The levels of inhalational exposure to L-PEI reported by Ferrari and colleagues (39) are many folds greater than the maximum dose that would be experienced in the clinical situation. Moreover, the dose of L-PEI delivering to the pulmonary circulation is less than the aerosolized dose. Based on existing literature on aerosol deposition (42–44), we estimate that only 10–30% of aerosolized DS-1 would reach the lungs. In humans, we anticipate an aerosolized L-PEI exposure of 8 mg/kg/day, based on an aerosolized dose of 0.5 mg/kg/hour of DS-1 given that L-PEI represents two-thirds of the mass of DS-1. This dose of DS-1 corresponds to an airway NO concentration of 14 parts per million. The dose of DS-1 chosen for nebulization in sheep is based on the results of our previous pilot study on conscious rats and dogs (G. Southan and colleagues, unpublished observations). In these experiments, aerosolized DS-1 was administered continuously for 3 hours at doses of 1, 3, and 8 mg/kg/hour. Based on in vitro studies measuring NO release by chemiluminescence, we estimate that these rates correspond to airway NO concentrations of 28, 85, and 239 parts per million, respectively. After 3 days, no toxicity was observed at the rates of 1 and 3 mg/kg/hour, but traces of inflammation in the lungs were noticed at the highest dose (8 mg/kg/hour). Thus, these tests may indicate a good toxicity profile of DS-1 and its favorable therapeutic ratio. The demonstration of toxic effects at the highest dose was not unexpected, given the extensive data in the literature showing that doses of inhaled NO in excess of 40 parts per million may be associated with pulmonary toxicity (3, 8, 9). Most likely, NO donors have the same adverse effects as gaseous NO (15). Hitherto, no toxic effects have been described specifically for NONOates. In this study, we did not observe any toxicity of DS-1 on human alveolar epithelial cells even at concentrations as high as 500 µg/ml, whereas the vasodilatory effects of the compound resided in the 1–3 µg/ml range. During experiments in vivo, we did not find any evident toxic effects of DS-1. Circulating methemoglobin concentrations did not increase during the nebulization of DS-1 to endotoxemic sheep, despite the fact that we used a relatively high dose. In accordance, screening for biologic markers of oxidant-induced ALI, such as MPO and the nitrotyrosine score, did not reveal any signs of toxicity of DS-1. The increase in the plasma concentration of NOX after nebulization of DS-1 is consistent with results of previous reports on other NO donors (13, 18). In parallel with the rise in NOX, high concentrations of inhaled NONOates and gaseous NO have been shown to jeopardize systemic hemodynamics (3, 15). However, the dose of nebulized DS-1 used in this study did not have any evident effect on systemic circulation, as mean systemic arterial pressure and systemic vascular resistance index demonstrated no intergroup differences. If given intravenously, DS-1 would lose its lung selectivity, cause systemic hypotension, worsen ventilation/perfusion matching, and result in arterial hypoxemia like other nitrosylated vasodilators (19, 20). The pharmacologic characteristics of NO donors could differ from those of gaseous NO. According to the results of our pilot experiments, the DS-1–induced pulmonary vasodilation is developing within 10 minutes after the start of nebulization that is slower than for inhalation of gaseous NO but comparable with other NONOates (14, 15). The washout period of NO effects after withdrawal of the nebulized DS-1 is approximately half an hour, which markedly exceeds the washouts of inhaled NO gas and other NO donors (15, 45). Summarizing the findings of other authors and our own results, the advantages of DS-1 and other NO donors over inhaled NO gas may include a longer half-life, a simplified delivery system, an ability to provide both intermittent and continuous inhalation, and decreased environmental release of NO or NO2 (3, 11, 14, 15, 45). Because physiologic effects of DS-1 are similar to gaseous NO, we suggest that it should have the same indications for use, including diseases associated with pulmonary hypertension and ALI (3). In addition, we speculate that in contrast to NO gas, DS-1 or its derivatives could be used for treatment of chronic pulmonary hypertension on an outpatient basis. In that case, a metered dose inhaler would be a useful tool for delivery of NONOates to the airways. We conclude that the novel NONOate, compound DS-1, counteracts epinephrine-induced vasoconstriction of rat aortic rings without any desensitization to the vasorelaxant effect. In endotoxin-induced ALI in sheep, continuous inhalation of the aerosolized DS-1 attenuates pulmonary hypertension and edema and improves myocardial function and arterial oxygenation. We interpret these effects as a result of pulmonary vasodilation, reducing right ventricular afterload and improving the ventilation/perfusion relationship of the lungs subsequent to local release of NO. There are no evident signs of toxicity of DS-1, neither on human pulmonary epithelial cells in vitro nor on endotoxemic and healthy sheep; however, further studies are warranted to evaluate pharmacokinetic and pharmacodynamic characteristics of nebulized DS-1, to find the optimal therapeutic regimen, and to determine whether it can be used for treatment of ALI and pulmonary hypertension in humans.
The authors appreciate the advice of Leif Jørgensen, M.D., Ph.D., the technical assistance of Randi Olsen and Helga Marie Bye, and the help of Rod Wolstenholme for preparing the figures.
Supported in part by the Research Council of Norway (grant 120473/730); the Department of Anesthesiology, University Hospital of Northern Norway; the Laerdal Foundation for Acute Medicine; the Blix family fund; and departmental funds. This article has an online data supplement, which is accessible from this issue's table of contents online at www.atsjournals.org. Received in original form February 11, 2002; accepted in final form August 15, 2002
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
p < 0.05 versus baseline in ET + DS-1 group; 
p < 0.05 versus baseline in ET group.
B, reduced cell activation, scavenging of reactive oxygen species, and modulation of cyclic guanosine 3'-5' monophosphate–dependent signaling mechanisms (
