Nitric oxide: physiology and pharmacology.

Abstract

The relevance of the nitrogen oxides for clinical medicine has been recognized for over a century. Nitrous oxide (N2 O) has been used safely and effectively for analgesia and anesthesia since the 18th century. Nitric oxide (NO) has emerged as a chemical messenger in a multitude of biologic systems having homeostatic activity in maintenance of cardiovascular tone, platelet regulation, and central nervous system signaling, as well as a role in gastrointestinal smooth muscle relaxation, immune regulation, and as a possible effector molecule for the volatile anesthetics. This review will discuss the biochemical basis of NO activity and examine the current clinical applications of NO, focusing primarily on the practice of anesthesia and critical care medicine. NO Synthesis and Transport NO is produced by a five-electron oxidation of the guanidino nitrogen of L-arginine in the presence of molecular oxygen to yield citrulline and the free radical, NO. This reaction is catalyzed by NO synthase (NOS), and requires several cofactors including a heme moiety. NO is a stable, colorless gas that undergoes rapid oxidation to nitrite and nitrate in solution where estimated half-life is less than 4 min. In biologic systems, half-life is estimated to be 3-30 s, being inactivated by superoxide anion or oxyhemoglobin. Molecular functions of NO include activation of guanylyl cyclase, and scavenging of superoxide anion [1-3]. NO was originally thought to act as a protective free radical scavenger by binding superoxide [4]. The product, however, is peroxynitrite (ONOO--), a relatively long-lived, potent oxidant [5]. Peroxynitrite release has been associated with immunologic activation of macrophages [6-9]. What remains controversial is the physiologic relevance of in vivo peroxynitrite production, given a half-life of less than 1 s at physiologic pH [10]. Interactions of NO with other oxygen derived free radicals to form nitrogen oxide species such as nitrogen dioxide (NO2) have been studied, although the actual quantities produced in biologic systems and their effects remain highly controversial [1,3,11]. Molecular NO as a free radical is an unstable species, and thus may also be oxidized, reduced, or complexed with other biomolecules, depending on the surrounding microenvironment. As mentioned above, NO2 is one such product of practical importance in the clinical setting. In addition, NO can react with transition metals, again to result in formation of toxic species. NO2 is known to be toxic to the pulmonary system, at times causing fatal injury. Situations in which this may occur clinically include storage of NO in high-pressure metal tanks, and in combination with oxygen in ventilator tubing and in the circuitry of the anesthesia machine. NO is synthesized by cell-specific isoforms of NOS, and has been identified in cells of endothelium, vascular smooth muscle, platelets, hepatocytes, Kupffer cells, glia, and neurons. NOS is broadly classified into subtypes: constitutive (cNOS) and inducible (iNOS). The constitutive isoform is calcium- and calmodulin-dependent, continuously expressed, and produces NO in picomolar concentrations. iNOS, in contrast, requires inducers such as specific cytokines (IL-1, tumor necrosis factor, interferon-gamma) or endotoxin for expression and is calcium-independent. After induction, iNOS remains active for 4-24 h, yielding nanomolar concentrations of NO, 100-fold greater than those of cNOS. In addition, iNOS is transcriptionally regulated, a process inhibited by glucocorticoids, transforming growth factor-beta, IL-4, and IL-10 [1,3,11]. No physiologic agents have yet been identified that specifically inhibit either cNOS or iNOS enzyme activity. Human forms of cNOS and iNOS have been cloned and localized to chromosomes 7 and 17, respectively [12]. The mechanism of NO transport to its molecular targets is not yet known. Simple diffusion of NO is an implausible mechanism to account for its paracrine effects. Under physiologic conditions, nitrogen oxides readily combine with protein-bound thiol groups to form stable, biologically active S-nitroso compounds. Indeed, NO may circulate in mammalian plasma as an S-nitroso-adduct of albumin [13,14]. S-Nitroso compounds exhibit in vivo endothelium-derived relaxing factor (EDRF)-like properties mediated by a cyclic guanosine monophosphate (cGMP)-dependent process [14]. Dinitrosyl iron complexes have also been proposed as carrier molecules for NO [15]. In precontracted deendothelialized rat aorta segments, dinitrosyl iron complex vasodilator responses more closely resemble acetylcholine-induced relaxation (i.e., EDRF) than that of true NO [16]. Some investigators contend that the true identity of EDRF is not NO but an NO-carrier compound with iron- or thiolcontaining ligands [15]. However, it is agreed that NO remains the final biochemical mediator for the actions of EDRF. Pharmacologic Regulators of NO Synthesis Many pharmacologic regulators of NO synthesis have been identified. Promoters of NO bioactivity include organic nitrates such as nitroglycerin, sodium nitroprus-side (SNP), and S-nitroso thiols, while inhibitors include flavoprotein, calmodulin and heme binders, substrate analogs, and tetrahydrobiopterin-depleting agents. Substrate analogs have become the most commonly used inhibitors (NG-nitro-L-arginine, NG-methyl-L-arginine [L-NMMA], and NG-nitro-L-arginine methyl ester [L-NAME]), acting as both competitive substrate inhibitors and modulators of membrane transport of L-arginine [1,3]. They exhibit variable affinities for cNOS and iNOS, although none are truly specific. Pulmonary System NO activity has been implicated in the neural regulation of bronchial and pulmonary vascular smooth muscle. NOS has been localized to the nonadrenergic, noncholinergic (NANC) neural system, and inhibition of NO synthesis by L-NAME and L-NMMA blocks NANC-stimulated relaxation of airways in guinea pigs. Within the pulmonary vasculature, NO has been shown to mediate acetylcholine-associated vasodilation in multiple human and animal models. Pulmonary vasodilation may result from NO activity in the NANC neural system [17-20]. Despite its role in systemic vascular tone, however, there is no clear evidence that NO is the major determinant of basal pulmonary vascular tone. Inhibition of NO synthesis results in exaggerated vasoconstrictor responses in a variety of species and in pulmonary vessels of all sizes [20]. Whether continuous NO release maintains basal vascular tone is unclear: NO may be immediately released only when a vasoconstrictor stimulus is detected, perhaps to prevent increased pulmonary vascular tone. NO activity does appear to be pivotal in acute and chronic hypoxic pulmonary vasoconstriction. During an acute hypoxic challenge, NO production consistently enhances the pressor response. In chronic pulmonary hypertension, endothelium-dependent relaxation to acetylcholine is impaired while relaxation to SNP (a direct NO donor) is maintained, indicating NO deficit as the cause. NO has a well documented relaxant effect on constricted pulmonary vessels, but minimal effect in nonconstricted vessels. How NO mediation of hypoxic pulmonary vasoconstriction progresses to a state of NO deficiency in chronic pulmonary hypertension is unknown [19,20]. NO effects in pulmonary hypertension suggest therapeutic uses for inhaled NO as a selective pulmonary vasodilator. Its short biologic half-life (3-30 s) suggests minimal effects on other body systems. Investigators have found pulmonary vasodilation without systemic changes in healthy human volunteers given inhaled NO, and no associated impairment of oxygenation or change in systemic blood pressure [21-23]. One explanation is that, due to its short half-life, NO dilates only the vessels immediately surrounding ventilated alveoli, thereby decreasing mean pulmonary artery pressure (PAP), pulmonary vascular resistance (PVR), and shunt fraction, and improving oxygenation. This, in turn, enhances vasoconstriction around the still-unventilated alveoli, further improving shunt fraction. Little effect is seen in nonconstricted vessels [23]. The necessity of providing NO via inhalation combined with the complexities of dealing with a potentially toxic and unstable radical species limits its use. Currently, the most accepted indication is persistent pulmonary hypertension of the newborn. Short-term treatment with inhaled NO allows time for closure of a patent foramen ovale or patent ductus arteriosus, initiation of extracorporeal membrane oxygenation or surgery [23]. Similarly, complex congenital heart defects, left-to-right shunts, and pulmonary hypertension can be stabilized to permit transfer of patients to appropriate referral centers [24]. Applications for inhaled NO in adults are more diverse. In adult respiratory distress syndrome (ARDS), NO decreases PVR, improves oxygenation, and may attenuate bronchoconstriction. Inhaled NO at 40 ppm increases conductance but has no effect on compliance in asthmatics or on airway mechanics in patients with chronic obstructive pulmonary disease suggesting that NO effects may be limited to the proximal airways [25]. In a porcine oleic acid model of ARDS, inhaled NO reversed pulmonary hypertension, improved oxygenation, and decreased shunt fraction in a concentration-dependent manner [26]. Clinical trials in patients with severe ARDS demonstrated that inhaled NO (18-36 ppm) consistently improved PVR, shunt fraction, and oxygenation. Survival rate for these patients was high (86%, 6/7), but confounded by concurrent use of extracorporeal membrane oxygenation, making it difficult to distinguish the benefits of NO [27]. Whether inhaled NO therapy can improve outcome of ARDS patients is unclear. Mortality most often is related to sepsis or multiorgan failure [23]. For example, use of inhaled NO, 25 ppm for 48 h, to treat right heart failure and pulmonary hypertension after heart transplantation improved both mixed venous oxygen saturations (47%-63%) and PAP (41-31 mm Hg), but the patient died of multiorgan failure 48 h later [28]. Studies of cardiac surgical patients have demonstrated pulmonary vasodilatory responses to NO after cardiopulmonary bypass to be proportional to PVR, and point to the beneficial effects of inhaled NO (20 ppm) in supine cardiac surgical patients with moderate pulmonary hypertension undergoing onelung ventilation. PVR was reduced by 23% and shunt fraction improved. However, no benefit was obtained in the absence of preexisting pulmonary hypertension [29,30]. Toxicity of inhaled NO and its contaminants remains a difficult issue. Animal studies have shown no adverse effects of inhaled NO, 10-40 ppm, after 6 days to 6 mo. Occupational Safety and Health Administration standards for healthy industrial workers allow 25 ppm for up to 8 h per day [23]. However, these standards fail to consider effects on newborn or previously damaged lungs. In the presence of oxygen, NO rapidly forms NO2, a highly toxic species. NO2 is thought to be the causative agent for silo-filler's disease and other related lung syndromes involving pulmonary edema, hemorrhage, and bronchiolitis obliterans. NO2, delivered at 2 ppm for 24 h, causes cilia depletion and epithelial hyperplasia, but these changes are reversible, even after 21 days of administration [31]. Interestingly, NO is found in cigarette smoke at levels of 100 ppm or higher and may be an etiologic agent in development of lung disease in smokers. At 50 ppm, NO impairs performance of learned tasks and prolongs brainstemevoked potential responses in rats. Anecdotal evidence suggests that inhaled NO may cause a bleeding diathesis, possibly by inhibiting platelet function [32]. Finally, there is risk of methemoglobinemia: in one study, 48 h of exposure to NO at 25 ppm resulted in a methemoglobin level of 1.6%; in another, 4/123 patients developed levels over 5% in response to exposures ranging from 6 to 96 h. Of note, the patient who developed an increased level in only 6 h was a female Native American child, an ethnic group with an increased incidence of methemoglobin reductase deficiency [28,33,34]. Techniques of administering a stable concentration of NO via a mechanical ventilator, anesthesia machine or face mask have been described in detail by several authors. The goals of such a system have been defined as: 1) minimizing time of gas in the delivery circuit to avoid NO2 formation; 2) allowing precise control of the NO dose administered; 3) allowing ease of on-line analysis of NO, O2, and NO2; and, 4) incorporating stringent controls for exhaust gas treatment [34,35]. NO is available as an investigational new drug of medical grade in concentrations of 800 ppm or 2200 ppm with nitrogen as the balance gas. The gas is "scrubbed" of oxygen and is free from conversion to NO2 for 18 mo. Industrial chemiluminescence devices have been used to monitor for NO and NO2, and should be connected as close to the patient as possible. Utmost vigilance is required to ensure that significant amounts of NO2 do not accumulate, especially when using higher doses of NO (e.g., >20 ppm) [34,35]. Cardiovascular System The earliest identified biologic role for NO was in cardiovascular homeostasis. cNOS is now recognized as a major physiologic regulator of basal systemic blood vessel tone [36]. Enhanced NO release can be stimulated by pulsatile flow and shear stress, and cytokine-rich states. NO is active in controlling renin release and sodium and water homeostasis, indicating its importance in regulating intravascular volume as well as vascular tone [37]. NO originating from platelets and other sources plays a role in modulating platelet adhesion and aggregation, as well as leukocyte adhesion, endothelin generation, plasminogen activator enzyme function, and vascular smooth muscle proliferation [1]. Recent studies implicate deranged NO production in many pathologic states. In primates fed an atherogenic diet, NO-mediated vasodilator response to acetylcholine is reduced, while vasoconstrictor response is enhanced. With regression of the atherosclerotic lesions, basal vasodilation and vasoconstrictor responses are restored [32]. In hypertensive animals, similar derangements in vascular response are reversed with antihypertensive therapy [38]. These and other studies imply that various "vasculopathic" states may arise either from impaired release of NO or alteration in smooth muscle signal response characteristics. NO may even be protective. Prolonged endothelial exposure to oxidized low-density lipoprotein causes irreversible inhibition of NO-dependent vasodilation. NO inhibits oxidation of lipoproteins, preventing free radical-mediated oxidative membrane injury [39,40]. Some investigators have suggested a relationship between NO and the atherosclerosis-resistant state associated with estradiol synthesis in premenopausal females. NO release from aortic rings of female rabbits was found to be greater than that of males, an effect ablated by oophorectomy [40]. Modulations in NO production have shown promise as therapeutic interventions. Diets rich in L-arginine have been shown to prevent development of hypertension in animals at risk, and L-arginine infusion causes rapid reduction of blood pressure in humans with essential hypertension [41-43]. Also, L-arginine, whether dietary or parenteral, has been shown to decrease vascular smooth muscle proliferation after balloon catheter-induced arterial injury in rabbits, indicating a potential role in prevention of restenosis after angioplasty [44]. IL-1 beta-induced NO release by vascular smooth muscle has been shown to be enhanced by plasmin and attenuated by thrombin [45]. Thus, endogenous NO is intimately involved in the fibrinolytic and thrombolytic consequences of endothelial injury and may prove a new method of pharmacologic "anticoagulation" in the setting of arterial injury. Actions of NO in regulation of platelet activity have been of great interest. Basal NO release leads, via several guanylyl cyclase-dependent steps, to suppression of intraplatelet calcium release and decreased platelet activation. Additionally, NO acts both alone and synergistically with prostacyclin to inhibit aggregation and actively disaggregate platelets. Responsible sources of NO are multiple, and include the platelet itself [46]. NO production in the cytokine-rich state of septic shock may result from generalized iNOS activation. Sources of NO in this setting include endothelium, vascular smooth muscle, and macrophages. Studies using hamster papillary muscle have shown that cytokine-induced myocardial depression is blocked by L-NMMA and exacerbated by L-arginine, the true NOS substrate [47]. These results suggest that NO may be the long-hypothesized myocardial depressant factor in sepsis. NO also is involved in development of hypotension accompanying septic shock. In animals, hypotension can be alleviated by administration of L-NMMA and reproduced by L-arginine [48,49]. On the other hand, NOS inhibitor administration resulted in prolonged intense systemic vasoconstriction with an associated increase in mortality in one animal study, while, in another, large doses of L-NMMA resulted in accelerated and augmented hypotension [48-50]. L-NAME infusion in pigs was found to potentiate lipopolysaccharide-induced pulmonary hypertension (PAP, 16.8 to 36.0 mm Hg); in mice, L-NAME infusion after lipopolysaccharide infusion resulted in 0% survival rage (compared 87.5% in controls) [51,52]. These conflicting results suggest that inhibition of NO production is not a panacea for patients in septic shock. Case reports of NO inhibition in patients with pressor-resistant shock or multisystem organ failure have been published, and clinical trials are beginning to appear. In a randomized prospective double-blind placebo trial, 12 patients in septic shock received L-NAME as a bolus followed by an infusion for 6 h [53]. Mean arterial blood pressure and systemic vascular resistance increased but cardiac output decreased by as much as 30%. These changes persisted for up to 16 h after termination of the infusion. Survival was not reported. These results raise concerns regarding correction of blood pressure at the expense of vital organ perfusion. Obviously, more investigation is required to determine if there is a role for systemic NOS inhibition in septic shock. Cytokine-induced shock is seen after IL-2 immunotherapy for selected solid tumors. L-NMMA, when given concurrently with IL-2, prevented changes in blood pressure, hepatic, renal and hematologic variables, and dramatically improved survival in dogs [54]. In humans receiving IL-2/lymphocyte activated killer cell therapy, NO metabolites were found to be ninefold greater than in controls [55]. Clinical trials have been proposed to determine whether NOS inhibition might decrease toxicity associated with IL-2 immunotherapy, thereby increasing its therapeutic index. NO plays a major role in vascular homeostasis with relevant functions in vascular tone, coagulation, fibrinolysis, and overall endothelial and vascular smooth muscle function. However, evidence suggests that aberrant NO synthesis, whether too little or too great, is linked to vascular derangements observed in septic shock, hypertension, and atherosclerosis. Thus, the protective or toxic effects of NO may be concentration-dependent. Although it is known that small amounts of NO are necessary for homeostasis, large amounts, such as those released after iNOS activation, may be deadly [3]. This paradoxical activity may be a teleologic line of defense against cellular invaders or tumor cells. Inhibition of only the inducible isoform of NOS may prevent the consequences of massive overproduction while allowing basal NO release to continue. NO Interactions with Anesthetics NOs implications in anesthetic actions have been reviewed by several authors [56,57]. Halogenated anesthetics attenuate in vitro endothelium-dependent relaxation [58-60]. Halothane reportedly attenuates the vasodilatory response to bradykinin and acetylcholine in isolated vessels with intact endothelium, but permits responses to nitroglycerin, an NO donor [61]. Exogenous NO can induce a halothane-mediated decrease in arteriolar relaxation [62]. Halogenated drugs interfere with the arginine-NO-cGMP pathway and suppress the resultant vasodilation, contradicting their own observed vasodilating properties [57,63]. How the inhaled drugs disrupt the NO-cGMP pathway remains controversial. Halogens have high electron affinities and are strongly attracted to electron donating sites, such as ferrous-heme proteins, suggesting a pathway by which halogenated agents may suppress the bioactivity of NO [57]. NO-stimulated cGMP content in rat aorta rings is attenuated by halothane, suggesting the site of action to be the vascular smooth muscle cell, perhaps by interference with activation of guanylyl cyclase (GC) [64]. GC is a hemoprotein enzyme activated by interaction of NO with its ferrous iron. In vitro studies using hepatic tissue exposed to volatile anesthetics support NO-mediated inhibition of GC [65]. In contrast, studies performed in bovine aortic endothelial cells indicate that halothane does not interfere with endothelial cell release of EDRF/NO or its smooth muscle relaxant properties, not the relaxant effects of SNP, a direct NO donor. The latter results suggest that halothane-derived free radicals scavenge NO and that GC is not the site of action [66]. Several reports have shown that significant pressor responses to NOS inhibition may be attenuated by halogenated drugs. Coadministration of L-NMMA with halothane or isoflurane increased mean arterial blood pressure and systemic vascular resistance, with isoflurane exhibiting greater increases in resistance in several regional vascular beds [67]. Conversely, infusion of 1 minimum alveolar anesthetic concentration (MAC) isoflurane into the coronary vessels of openchest dogs produced an initial 4.5-fold increase in coronary flow followed by a slow decline toward baseline in both the presence and absence of L-NAME [68]. Integrating these results with those demonstrating suppression of the NO pathway by halogenated drugs is difficult. Evidence implicating NO in the systemic circulation suggests that NO may mediate the cerebral hyperemia associated with halogenated anesthetics. In dogs anesthetized with isoflurane or pentobarbital, responses to a cholinergic agonist indicated cerebral hyperemia in only the isoflurane group and L-NAME reversed this response [69]. Similarly, L-NAME prevented cerebral hyperemia resulting from isoflurane, halothane, and nitrous oxide, indicating that the increased cerebral blood flow (CBF) may be due to NO [70]. In awake rats, L-NAME significantly decreased global CBF, but in those exposed to isoflurane, reduced only flows to the medulla and pons [71]. In another study, L-NAME decreased cerebral hyperemia in both awake and anesthetized rats [72]. In the presence of inhaled anesthetics, NOS inhibition prevents or attenuates hyperemia in the cerebral circulation. NO is involved in basal regulation of CBF, as elsewhere, and it is unlikely that vascular endothelium would differ fundamentally in various parts of the body. Therefore, if inhalation anesthetics suppress the relaxant effects of the NO pathway elsewhere, how would they augment its activity in the central nervous system? One possible explanation may be production of increased amounts of NO by neuronal and glial cells. NO also may mediate anesthetic effects on awareness. It was originally reported that NOS inhibition with L-NAME decreased the MAC of halothane in a dose-dependent, reversible manner in rats, an effect not seen with the stereoisomer, D-NAME, and reversed by infusion of L-arginine [73]. Recently, administration of L-NAME in similar doses in rats showed no decrease in MAC, using similar experimental methods (i.e., tail clamp response testing) [74]. No explanation for these opposite findings was offered. One investigator has likened general anesthesia to several other disorders of consciousness including heat stroke, cerebral malaria, ethanol intoxication, and opioid narcosis. These pathologic conditions all inhibit glutamate-dependent neuroexcitatory transmission in the brain causing decreased NO activation and a NO deficit [75]. Although it has never been shown clinically, it is intriguing to hypothesize that interference with NO activity as a neurotransmitter might result in "anesthesia" as produced by halogenated drugs. Studies examining the role of NO in pain perception are conflicting, although NO has been identified as a NANC neurotransmitter and this may be important in the processing of nociceptive stimuli [76,77]. In fact, the method of determining MAC in the studies reported above (tail pinch) maybe a measure of NO's role in nociception rather than one of maintenance of consciousness. Also, NO activity with respect to intravenous anesthetics is poorly defined, although it is agreed that most clinically useful intravenous agents have little, if any, effect on the EDRF/NO pathway [57]. Additional Considerations NO probably acts as a NANC neurotransmitter in regulation of smooth muscle relaxation in both large and small bowel [78]. It is possible that deranged NO synthesis or release is involved in such neuromuscular disorders as Hirschsprung's disease, achalasia, and "nutcracker" esophagus. Gastroesophageal reflux may be the result of increased NO production or endorgan sensitivity [79]. NO also has been implicated as a mediator of mucosal blood flow and mucosal integrity throughout the gastrointestinal tract [80,81]. NO also may be protective of mucosal integrity after endotoxin-induced shock [78]. Chronically increased NO production may be involved in the hyperdynamic circulation of cirrhosis [82,83]. And NO has been investigated as a mediator of hepatocellular dysfunction in multiorgan system failure, and implicated as a mediator of pancreatic function [84-86]. NO is toxic to pancreatic islet cells, hastening development of type 1 diabetes mellitus, and is involved in control of insulin production [86,87]. Again, although low levels are regulatory, high levels of NO are toxic. A great deal of attention has been focused on the roles of NO in the central nervous system. The gene for neuronal cNOS, isolated and identified in the cerebellum, differs from endothelial cNOS and has been localized to chromosome 12. Recent evidence indicates that iNOS expression does not occur in neurons but does occur in glial cells [88]. In the CNS, NO is thought to help control CBF, memory formation, and neurotoxicity, and has been implicated in the pathogenesis of hypertension, subarachnoid hemorrhage, Huntington's chorea, Alzheimer's disease, and hypercapnia-mediated increases in CBF [39]. Unlike NO activity in other vascular beds, changes in CBF accompanying hypoxia are not NO-related, although NO is central to the hyperemia that accompanies hypercapnia [88,89]. The effect of NOS inhibition on cerebral hyperemia is maximal at PCO2 = 50-60 mm Hg, and disappears at PCO2 > 100 mm Hg [90] NO levels are high only during initial periods of cerebral ischemia, increasing again during reperfusion [91]. Whether these effects are ultimately protective or toxic is unclear. Animal studies examining infarct size and neuronal survival report mixed results [88,90]. NO may maintain CBF acutely, decreasing platelet and leukocyte adhesion and blocking excitatory neurotransmitter receptor activity, but becomes toxic after a period of hours to days as a result of oxidative injury. Conclusion NO is deceptively simple in structure but highly complex in activity. We are just beginning to understand its role as a chemical messenger in multiple physiologic systems in the body. Clinically, it offers promise as an inhaled drug for the treatment of various pulmonary disorders involving derangements of vascular function or oxygenation. Similarly, manipulation of the NO pathway may prove therapeutic in selected cardiovascular disorders and cancer immunotherapy. However, NO's broad spectrum of activity and effects on other organ systems make large-scale experimental and clinical trials of potential applications essential before it can be used safely and effectively as systemic therapy for specific conditions. Thanks to Winifred von Ehrenburg for editorial assistance.