Markers of nitric oxide are associated with sepsis severity: an observational study

Sepsis affects millions of people worldwide and its incidence is increasing [1]. A dysregulated immune response to an infection is the main feature of sepsis pathogenesis, which, together with hemodynamic and microcirculatory changes, may lead to insufficient tissue oxygenation, organ dysfunction, and septic shock [2, 3]. Nitric oxide (NO) is an important regulator of physiological processes in the immune system and circulation [4, 5]. NO is generated by a family of nitric-oxide synthases (NOS) within cells [6]. There are three tissue-specific isoforms termed endothelial (eNOS), neural (nNOS). and inducible NOS (iNOS), which are expressed in immune tissue. All NOS catalyze the production of NO from the substrates L-arginine (lArg) and homoarginine (hArg) [7]. NO levels are reduced when NOS is inhibited by asymmetric dimethylarginine (ADMA) in a competitive manner. ADMA levels again are controlled by dimethylarginase-dimethylalaminohydrolase-1 and 2 (DDAH1 and 2) activity, which inactivates ADMA by cleavage. DDAH2 is predominantly expressed in immune cells [8]. NO is a labile compound with a short half-life in blood making its direct detection and measurement unfeasible. Therefore, concentrations of substrates and endogenous inhibitors of NOS are considered surrogate markers for NO bioavailability [6].

NO levels in sepsis are critical and of particular interest for two reasons. First, NO regulates vascular function. Endothelial-derived NO dilates blood vessels by relaxing vascular smooth muscle (VSM). NO activates guanylate cyclase, which induces VSM relaxation by increasing intracellular 3,5-cyclic guanosine monophosphate (cGMP) concentration. In sepsis, NO can be thought of as a Janus-faced signaling molecule. On the one hand, excessive production of NO leads to severe hypotension and may cause signs of shock. On the other hand, NO is essential in maintaining microvascular function by regulating the supply and distribution of oxygen and nutrients throughout all tissues and organs [9]. In this context, NO maintains microvascular homeostasis by dilating and regulating vascular tone, red blood cell deformability, and leukocyte and platelet adhesion to endothelial cells [9]. Second, NO is essential in the immunological response to pathogens and has been extensively studied in macrophages [4]. NO is a free radical and has immediate antimicrobial effects including disruption of bacterial target structures and inhibition of bacterial metabolism (e.g. inactivation of the Krebs cycle or of virulence factors such as Clostridium difficile toxins A and B) [10‐12]. Antimicrobial effects of NO play an important role in the response to infections. In mice infected with Leishmania major, NO was shown to suppress the metabolic activity of L. major without directly killing the pathogen but facilitated resolution of the disease by the immune response [13]. Recently, DDAH2, the ADMA-degrading enzyme in immune cells, has been studied in knockout mice. Global knockout of DDAH2 was associated with 80% lethality compared to wild-type animals when polymicrobial sepsis was induced, underlining the role of NO bioavailability in sepsis [14].

Considering the known effects of NO on vascular function and the immune system, we sought to investigate whether plasma concentrations of NOS substrates (lArg and hArg) and of the NOS inhibitor ADMA, and whether DDAH2 expression in immune cells are altered in patients with sepsis and whether the magnitude of any such alterations is related to disease severity.

We found that plasma hArg levels were significantly reduced whereas lArg levels remained unchanged in patients with sepsis. Plasma ADMA levels were increased in patients with sepsis; and both hArg and ADMA were associated with sepsis severity. Taken together the plasma hArg:ADMA ratio was increasing, and was closely associated with septic shock.

Sepsis is a systemic inflammatory response to an infection [2, 3]. The two most important factors affecting sepsis outcome are circulatory failure and immune suppression. NO is a pivotal signaling molecule with regulatory functions in both the circulation and the immune response. In this context, we hypothesized that decreased NO bioavailability in sepsis may contribute to disease progression. Physiologically NO regulates blood pressure. In vascular disease associated with hypertension such as atherosclerosis, coronary heart disease, or chronic renal failure low substrate levels but high inhibitor levels of NOS are markers of disease severity and mortality [22]. This is supported by experimental observations in animals and humans; supplementation of lArg in rabbits can restore vascular relaxation and endothelial function, whereas in humans infusion of ADMA results in increased systemic vascular resistance (SVR) and mean arterial blood pressure (MAP) [23‐26]. In contrast, the hallmark of sepsis and septic shock is reduced oxygen delivery depending on alteration in cardiac output (CO), the product of SVR and MAP [27]. This leads to the hypothesis that excessive NO production may contribute to reduced SVR and MAP [28]. Therefore it was appealing to test NOS inhibitors in sepsis to restore oxygen delivery. A randomized clinical trial compared intravenous administration of the non-selective NOS inhibitor L-N^G-methyl-L-arginine hydrochloride (546C88) with placebo. Indeed, 546C88 directly increased SVR but cardiac output and oxygen delivery were subsequently blunted [29]. Nonetheless, in the long term, administration of 546C88 was associated with higher mortality [30]. This can be explained by the fact that systemic inhibition of NOS in all tissues might be responsible for reducing signs of hypotension but may have an impact on capillary exchange or the immune response [6]. An alternative therapeutic concept might be the inhibition of the downstream signaling of NO, e.g. by methylene blue [31‐33].

We found lower levels of hArg in patients with sepsis than in controls but unchanged levels of lArg, which is remarkable. K_m values of hArg are much higher than those for lArg [34]. One may argue that hArg is less relevant than lArg and most studies in sepsis have focused on lArg. However, the usefulness of lArg as a sepsis marker is questionable as results of various clinical studies are inconsistent. In a group of 44 patients with cardiogenic or septic shock, lArg levels were shown to remain unchanged and in a randomized trial including 267 sepsis patients no changes in lArg levels were reported [35, 36]. In addition, a longitudinal study of 60 patients with septic shock observed no changes in lArg levels within the first 24 h but increased lArg levels at day 4 [37]. This suggests that the timing of measuring lArg in blood might be critical and usefulness as an early sepsis marker is questionable. To investigate hArg levels was one purpose of our study as data for hArg levels in sepsis are limited. Recent studies indicate that hArg is also involved in vascular homeostasis. According to a population-based cohort of 746 elderly participants, hArg and lArg are independently and antagonistically associated with blood pressure [38]. Of note, in interventional studies in mice, hArg supplementation has been shown to improve neurological outcome and cardiac function [39, 40]. Interestingly in kinetic experiments, when healthy humans were injected with lipopolysaccharide (LPS), NO-dependent vasodilatation increased but no changes in the concentrations of lArg in the blood were observed [41]. Therefore, systemic NO levels in blood during sepsis may not be lArg but hArg dependent. Unfortunately, the study cited did not report hArg levels. However, if hArg has a higher K_m value than lArg, we do not know if the catalytic efficiency of NOS is altered or if different NOS isoforms are preferentially metabolizing hArg during systemic inflammation and sepsis, which would explain why low hArg but not lArg levels correlated with disease severity in our study.

Despite NO effects in the circulation, reduced levels of NO in sepsis might be harmful for another reason. Cecal ligation reduces survival in NOS knockout mice compared to wild-type animals [31]; and bone marrow transplantation of wild-type to NOS knockout mice increases release of cytokines such as TNF-α and was found to improve survival in a model of idiopathic pneumonia [32]. Together with the discovery that mouse macrophages produce large amounts of nitrite (NO2-) and nitrate (NO3-) upon LPS stimulation, it has been suggested that decreased NO levels may impair the innate immune response [33], which is characteristic of sepsis progression in humans. Human monocytes show diminished defense mechanisms. For instance, the reduced ability to release pro-inflammatory cytokines after endotoxin stimulation has been referred to as monocyte “anergy” [42]. In this context it has been shown that hArg can serve exclusively as a substrate for NOS in macrophages [43]. Interestingly, the pharmacological transformation of hArg to the NOS inhibitor NH2-homoarginine results in significantly decreased NO production by macrophages [44]. One may speculate that decreased hArg levels may indicate monocyte “anergy” and this may explain why the hArg:ADMA ratio may be better than the lArg:ADMA ratio for diagnosing sepsis severity.

In addition to substrate availability, NOS activity is also regulated by the presence of NOS inhibitors. ADMA is an endogenous NOS inhibitor and increased levels are associated with decreased NO bioavailability. Other authors have hypothesized that increasing ADMA levels may contribute to poor sepsis outcome. For instance, in a cohort of patients with sepsis caused by the malaria parasite Plasmodium falciparum, ADMA was increased and associated with mortality [45]. Recently, findings from another sepsis study have suggested that ADMA may provide a non-invasive measurement of microvascular function. ADMA levels were not only increased and associated with the SOFA score, but pathologic values of the reactive-hyperemia index to estimate microvascular function were associated with a decreasing lArg to ADMA ratio [46], and in the aforementioned randomized controlled trial increased ADMA levels were independently associated with 90-day mortality [36].

However, it is unknown exactly how ADMA levels are regulated in sepsis. ADMA is metabolized to citrulline and dimethylamine by DDAH and enters cells through cationic amino-acid transporters (CAT). Two isoforms of DDAH are expressed in humans. DDAH1 predominates and is extensively expressed in the liver, lungs, and kidneys [47]. DDAH2 is primarily expressed in endothelial and immune cells [8, 48]. Circulating ADMA levels are genetically determined by promoter polymorphism in a regulatory gene encoding DDAH2 polymorphism and have been investigated by other research groups. Interestingly, in a cohort of 236 patients with postoperative inflammation after elective cardiac surgery, the DDAH2 -449G allele was identified as a polymorphism associated with an increased requirement of vasopressor to maintain organ perfusion [49]. In another series of 47 patients with severe sepsis, high ADMA concentration was also associated with the DDAH2-449G polymorphism [50]. In contrast, in pediatric sepsis, the DDAH2-449G polymorphism was associated with low ADMA concentration but with an increased likelihood of “cold” shock [51]. However, children with sepsis have much more variable hemodynamic profiles often with increased incidence of low cardiac output and elevated vascular resistance, which makes pediatric sepsis different from adult sepsis [52]. Interestingly, we found an association between decreasing DDAH2 mRNA expression in PBMC and disease severity. This is intriguing as animal experiments in global DDAH2-knockout mice have shown unchanged systemic ADMA and NO concentrations compared to wild-type mice. However, knockout mice injured by cecal ligation had 120-h survival of only 12% compared to 53% in wild-type animals [14]. The authors attributed this phenotype to impaired macrophage function. Monocyte-specific deletion of DDAH2 results in a similar pattern of increased severity to that seen in globally DDAH2-deficient animals. DDAH2 knockout in macrophages was associated with a significantly higher bacterial load in plasma and the peritoneum. Lambden and colleagues have shown that NO production in activated mouse macrophages is DDAH2-dependent with reduced intracellular NO levels within the cell. Moreover, DDAH2 knockout also impaired motility and phagocytosis in these cells.

With data from our small cohort we cannot conclude that decreased DDAH2 expression is monocyte or macrophage specific, as PBMC contain not just blood monocytes, but also B cells, dendritic cells, and activated T cells, which is a limitation of our expression findings. Limitations of our study are that it was carried out at a single center and involved relatively small numbers of patients, and larger cohorts should eliminate controls and potential selection bias. However, we observed associations between disease severity and systemic inflammation and sepsis, with surrogate markers of NO metabolism indicating decreased NO bioavailability in sepsis. To explain why hArg is decreased and ADMA is increased in sepsis requires further experimental studies. Nevertheless, we believe that our observations warrant follow-up studies with larger patient groups to confirm the power of the hArg:ADMA ratio to predict septic shock and sepsis severity.

Taken together, the close association between decreased hArg:ADMA ratio and sepsis severity in our cohort may hint at perturbation of endothelial function and pathogen defense. Impaired endothelial function and defense mechanisms due to decreased NO production in endothelial and immune cells may be a mechanism linking hArg, ADMA, and DDAH2 expression with organ dysfunction and impaired immune response in sepsis. The DDAH2-ADMA axis is a potential target and may be important in individual tailoring of therapy. Agents that compete with ADMA for NOS (such as hArg) or that potentiate DDAH2 activity should be further investigated in sepsis.