Differentiated control of deranged nitric oxide metabolism: a therapeutic option in sepsis?

Derangement of nitric oxide (NO) metabolism represents one of the key mechanisms contributing to macro- and microcirculatory failure in sepsis. Sepsis-related therapy combining fluid resuscitation with administration of vasopressor and inotropic agents, however, does not guarantee correction of maldistributed nutritive perfusion between and within organs. Therefore, the differentiated and selective pharmacologic modulation of NO-mediated vascular function could play a useful role in hemodynamic management of patients with sepsis. This viewpoint carefully evaluates the potential role of intentionally using partially opposing effects of NO donors and NO synthase inhibitors to complement current therapy of hemodynamic stabilization in patients with sepsis.


Microcirculation
An increased level of the endogenous NOS inhibitor asymmetric dimethylarginine (ADMA) was associated with reduced endothelial NO production and microvascular dysfunction and correlated with the severity of organ failure in patients with sepsis [14]. In a human sepsis model, NO-mediated vasodilation in the cutaneous microcirculation was inhibited during acute endotoxemia [15].
Hand-held microscopes introduced the possibility of visualizing the microcirculatory blood fl ow of patients at the bedside. In patients with severe sepsis, sublingual microcirculation measured by the orthogonal polari zation spectral imaging technique showed a signifi cantly reduced amount of perfused small vessels compared with humans without sepsis and a reduced proportion of perfused small vessels in non-survivors compared with survivors, suggesting a functional interrelationship between impaired sublingual microcirculation and negative outcome in patients with sepsis [16]. In patients with septic shock, sidestream dark fi eld imaging demonstrated a decreased perfused capillary density in non-survivors compared with survivors [17]. Th e development and correct use of measures for improving microcirculatory blood fl ow may be a promising strategy to further reduce sepsis-associated mortality (Figure 1). In this context, it is indispensable to determine causation or association of microcirculatory changes and the patient's condition. Given the anatomical and practical limitations preventing accessibility of the target microcirculation in the organ of interest, clinicians usually have to rely on easy-to-access areas, such as the sublingual region. Th e nature of sepsisrelated systemic changes, in turn, favors the use of the sublingual region to indirectly assess eff ects of pharmaco logic modulation. Th is, however, must be combined with safety parameters to prevent missing pharmacologically induced vasoconstriction when modulating functionally heterogeneous regulatory pathways.

Active manipulation of nitric oxide pathways as a therapeutic option in the management of sepsis
Sepsis-related alterations in NO metabolism have prompted pharmacologic studies aimed at reducing as well as increasing NO.
In a placebo-controlled multi-center phase III clinical study, the eff ects of 546C88, a non-selective NOS inhibitor, were studied in patients with septic shock. However, increased mortality in the verum group resulted in a premature termination of the study [21] (Table 1). Likewise, the Phoenix trial, investigating the eff ects of the NO scavenger pyridoxalated hemoglobin polyoxyethylene in patients with catecholamine-resistant distributive shock, has been stopped prematurely because of the absence of benefi cial eff ects [22]. NOS inhibition has provided heterogeneous results. It is important to consider, among other reasons, the type of NOS inhibitors -that is, selective versus non-selective, the species treated (that is, rodent versus large animal versus humans (volunteers versus patients)), and the time point of administration (that is, early versus late phase). While non-specifi c NOS inhibition might exert negative eff ects, the development of highly selective iNOS inhibitors appears promising. Lange and colleagues [23] reported that pulmonary eNOS protein expression at 8 and 12 hours and iNOS protein expression from 8 to 24 hours after injury in septic sheep were signifi cantly increased with unchanged nNOS protein expression at any investigated time point. More studies are required to further characterize the mechanisms and infl uence of diff erent NOS isoforms and NO concentrations during the course of the disease in the face of concomitant therapeutic infl uences.

Nitric oxide donor
One experimental strategy aimed at attenuating deranged microcirculation during sepsis is to increase NO availability via administration of NO donors. In this context, attenuation of LPS-induced increase in platelet adhesion to endothelial cells of murine intestinal venules by the NO donor diethylenetriamine-nitric oxide (DETA-NO) [24] indicates an NO-mediated protective infl uence on microvascular blood fl ow. Endotoxemia-induced inhibition of eNOS-derived NO [25] can be positively infl uenced by infusing, for example, the NO donor sodium nitroprusside. In fact, this approach signifi cantly improved LPS-induced impaired hepatic microcirculation at 2 and 3 hours after endotoxemia-induction in rats [26], and the NO donor 3-morpholinosydnonimine (SIN-1) increased cardiac index and mesenteric blood fl ow and improved myocardial function in the early phase of endotoxic shock in fl uid-resuscitated dogs [27]. Th ese data strongly suggest that the window of opportunity to replace insuffi ciently released eNOS-derived NO by an NO donor is during the early endotoxemic phase (Table 1).
Following pressure-guided fl uid resuscitation in patients with septic shock, intravenous infusion of the NO donor nitroglycerin at a loading dose of 0.5 mg and a subsequent continuous infusion of 2 mg/hour improved the disturbed sublingual microcirculation within 2 minutes after administration of the initial nitroglycerin dose. Th is fi nding strongly supports the notion that impaired microvascular fl ow can be improved by cautiously infusing nitroglycerin following previous restoration of intravascular volume [28] (Table 1). In a double-blind randomized placebo-controlled trial in which nitroglycerin was administered at 2 mg during the fi rst 30 minutes and 2 mg/hour during the subsequent 23.5 hours following systemic resuscitation of patients with severe sepsis, no additional nitroglycerin-mediated improvement in the microvascular fl ow index was observed after 24 hours. However, mortality in the nitroglycerin group exceeded mortality in the placebo group: 34.3% versus 14.2% (n = 35 per group) [29] ( Table 1). In this context, it is important to consider that the use of nitroglycerin to manipulate NO levels relies on the availability of specifi c enzymes. As such, aldehyde dehydrogenase type 2 (ALDH2) is essential to generate NO [30]. Under septic conditions, however, this enzyme might be depleted [31].
A target-oriented increase in NO availability, at least within the fi rst few hours after onset of severe sepsis, appears to allow successful modulation of microvascular dysfunction. Further studies need to determine under which circumstances titration of an NO donor may be effi cient and safe to correct a deranged microcirculatory blood fl ow. Given that certain pathophysiologic processes are limited to the early phase, in which patients are rarely accessible to pharmacologic modulation, timing will play Table 1

Non-clinical trials
Pseudomonas aeruginosa instilled into airways of sheep after inhalation of cotton smoke [18] BBS-2 (highly selective iNOS inhibitor) Pulmonary gas exchange improved and airway obstruction was partially reduced.
Intravenous infusion of P. aeruginosa in pigs [19] Combined L-NIL (selective iNOS inhibitor) and Tempol (superoxide scavenger) Promoted protection against cardiovascular, hepatosplanchnic metabolic, renal, and coagulation abnormalities Intravenous infusion of P. aeruginosa in pigs [20] L-NIL (selective iNOS inhibitor) Protective eff ect on, for example, ileal mucosal microcirculation LPS-induced endotoxemia in rats [26] Sodium nitroprusside (NO donor) Inhibition of increase of non-perfused sinusoids LPS-induced endotoxic shock in dogs [27] SIN-1 (NO donor) Low and moderate doses increased mesenteric blood fl ow.

Clinical trials
Humans with septic shock [21] 546C88 (non-selective NOS inhibitor) Increased mortality Patients with septic shock [28] Nitroglycerin (NO donor) Temporary decrease of MAP and central venous pressure but also strong increase in microvascular fl ow after 2 minutes Patients with severe sepsis [29] Nitroglycerin (NO donor) No improvement of the microcirculation compared with control group after 24 hours; higher mortality in the nitroglycerin group a crucial role. In daily routine, patients are seen mainly during their later phases of severe sepsis with prevailing extensive vascular hyporeactivity. Nevertheless, patients developing sepsis during their hospital stay could be amenable to early approaches as studied under experimental conditions. In this context, studies in clinically relevant animal models considering not only the very early stage of (severe) sepsis but also later stages might be useful.

The impact of conventional therapeutic strategies on nitric oxide metabolism in sepsis
Before introducing a complex concept of simultaneously or sequentially activating and attenuating NO pathways, it is also important to characterize the potential confl icting infl uence of standard measures of treating sepsis on the targeted NO pathways. Th e following paragraphs provide a brief overview of the interrelation between conventional therapeutic strategies in the treatment of severe sepsis/septic shock and NO metabolism.

Antibiotics
Under in vitro conditions, the antibiotic polymyxin B reduced LPS-induced NO production in rat hepatic Kupff er cells refl ected by decreased NO 2 − levels [32]. Ceftazidime signifi cantly decreased the concentrations of pulmonary 3-nitrotyrosine (a biomarker of nitrotyrosine formation) in sheep with acute lung injury and sepsis [33], indicating reduced nitrosative stress. Whether this eff ect is antibiotic-specifi c, occurs dose-dependently, and can be excessively increased by the co-administration of several antibiotics and whether the same mechanisms persist under conditions of antibiotic resistance remain to be determined.

Fluid resuscitation
In an endotoxemic rat model, fl uid resuscitation with albumin reduced LPS-induced increase of cardiac iNOS mRNA and protein expression [34]. Under in vitro conditions, increased fl ow seemed to cause endothelial shear stress, resulting in the release of NO and vasorelaxation [35]. However, fl uid resuscitation may infl uence NO synthesis, and this should be taken into account when establishing a new therapeutic concept aimed at controlling deranged NO levels.

Vasopressor agents and inotropic therapy
In isolated rat hepatocytes exposed to pro-infl ammatory cytokines, epinephrine and norepinephrine reduced NO production refl ected by signifi cantly decreased NO 2 − [36]. In incubated macrophages (RAW264.7 cells), however, epinephrine or norepinephrine enhanced LPSinduced increase in NO production [37]. In endotoxic rabbits, dobutamine administration was associated with reduced NO synthesis refl ected by decreased NO 3 − /NO 2 − concentrations in serum and the gut luminal perfusate [38]. More studies are required to characterize organspecifi c and dose-and duration-dependent eff ects of vaso pressors and inotropics and the infl uence of their continuous, prolonged, or short infusion on NO synthesis, metabolism, and downstream pathways. In a sheep model of septic shock, iNOS inhibition with BYK191023 was superior to nor epi nephrine in improving microcirculation; interestingly, combined administration of both substances versus single dosing markedly increased the maintenance of arterial pressure and kidney blood fl ow [39].

Steroids
Th e available data suggest a dose-dependent eff ect of hydrocortisone on NO synthesis. Whereas a loading dose of 100 mg over 30 minutes followed by a continuous infusion of 10 mg/hour for 3 days signifi cantly decreased plasma NO 3 − /NO 2 − levels [40], administration of hydrocortisone 50 mg/6 hours did not aff ect plasma NO 3 − /NO 2 − levels in patients with septic shock [41]. Overall, the impacts of these standard measures in the treatment of severe sepsis/septic shock on NO meta bolism have not been defi nitely clarifi ed. Further studies are needed to uncover the infl uence of standard therapy on the release of NO, degradation, and downstream pathways. Th e development of a therapy to adequately regulate NO synthesis by controlled increase as well as decrease, including organ-specifi c changes, might be a promising step forward in the treatment of severe sepsis/ septic shock.

Combining volume therapy with nitric oxide synthase inhibition and nitric oxide donors: hurdles to overcome in transferring this innovative approach to clinical routine
Th e concept of a well-controlled combination of fl uid infusion together with the administration of an NO donor and an iNOS inhibitor might improve locally deranged microcirculation under conditions of regionally and temporally heterogeneous NOS expression. In addition to NO-related issues, the impact of quantity and quality of fl uid therapy on NO production and degradation as well as phase-specifi c requirements must be characterized. Perhaps pharmacologic modulation may close the gap when fl uid administration is not necessarily appropriate to correct microcirculatory disturbance in patients (for example, during the late phase of severe sepsis) [42]. New therapeutic strategies targeting both macro-and microcirculatory dysfunction should be investi gated to assess whether such approaches are eff ective to prevent or treat organ dysfunction and failure.
Fluids are directed predominantly to the iNOSexpressing mucosa during fl uid resuscitation, thereby shunting the serosa microcirculation [43]. Combining fl uid resuscitation with either the iNOS inhibitor 1400 W [44] or the NO donor SIN-1 normalized serosa microvascular oxygenation [45]. Th is strongly suggests that iNOS inhibition may help to redirect blood fl ow to the serosa, thereby promoting correction of sepsis-associated heterogeneous blood fl ow while the NO donor promotes and supports intended blood fl ow redistribution through its vasodilatory eff ect [43]. In logical consequence, combining fl uid resuscita tion with iNOS inhibition, plus NO synthesis to regulate NO-mediated pathways spiraled out of control during sepsis, encourages further studies (Figure 2).
On the other hand, it needs to be considered that reduced NO synthesis might promote disease progres sion as NO is decisive in the killing of microorganisms [3]. Th is is also refl ected by signifi cantly delayed bacterial clearance and higher mortality [46] following non-selective NOS inhibition by N ω -nitro L-arginine methyl ester (L-NAME) administration in septic rabbits. In regard to host defense, NO level manipulation in septic animals revealed controversial results which may be due to diff erences in, among other things, the timing of NOS inhibition [47]. In addition, NO-driven sustained peroxy nitrite synthesis impairing vaso reactivity must be considered.
Given that results from animal studies might not necessarily be completely transferable to humans, several issues have to be clarifi ed before establishing the concept of improving the microcirculation by combining volume therapy with specifi c iNOS inhibition and an NO donor. In addition, the question of whether a disturbed microcirculation is just epiphenomenal or indeed the cause for higher mortality in sepsis has to be addressed. Th e outlined therapeutic concept must consider the temporally and regionally (organ-specifi c) heterogeneous changes in NOS expres sion/function, NO levels, and NO-mediated eff ects. Th us, co-targeting a reduction as well as an increase in NO levels might also involve a selected intra-arterial approach to reduce undesired systemic eff ects. Th e time window for administration and the dose of each com pound are key aspects that need to be defi ned. Th e approach described above also must consider that the current standard therapy for patients with sepsis consists of interventions that by themselves already infl uence NO-mediated pathways. Given the heterogeneity of the patient population, the plethora of diff erent important infl uences determined by components of routine inter ventions (antibiotics, crystalloids, colloids, transfusions, steroids, vasopressors, and inotropics), patient-specifi c features (age, gender, ethnicity, and comorbidities), time-point of treatment, phase-specifi c requirements, severity of illness, and other factors will probably require a well-diff erentiated and personalized treatment concept.
Since the eff ects of (simultaneous) NO donors and inhi bi tors are not easy to anticipate, these agents should be carefully titrated and used only on the basis of an adequate intravascular volume status to prevent them from contributing to hemodynamic collapse as well as excessive ischemia-promoting vaso constriction. Appropri ate monitoring for a safe and effi cient treatment is mandatory and should include bedside visualization of the microcirculation together with further global and regional surrogate markers of altered perfusion, such as lactate, SvO 2 (venous oxygen satura tion), cardiac output, organ function, and organ-specifi c biomarkers. Future clinical studies need to determine whether such diff erentiated manipulation of the NO pathways reduces overall morbidity and increases the likelihood of survival.