Article Text


Nitric oxide and infectious diseases
  1. David Burgner,
  2. Kirk Rockett,
  3. Dominic Kwiatkowski
  1. Molecular Infectious Diseases Group, Department of Paediatrics, University of Oxford, Level 4, John Radcliffe Hospital, Oxford OX3 9D, UK
  1. Dr Burgner. email: dburgner{at}

Statistics from

Nitric oxide (NO) has undergone something of an image change in recent years. Previously considered a pollutant from car exhausts, NO has now been implicated in many physiological processes. Here we review the dual roles of NO in infection—as a critical agent of host defence but also as a central mediator of pathogenesis—using septic shock, bacterial meningitis, and malaria to illustrate some current concepts and controversies. We have limited our discussion of the important physiological functions of NO in the cardiovascular and neurological systems to instances where these impinge on its role in infectious diseases.

NO is formed by the oxidative deamination of the amino acid L-arginine by nitric oxide synthases (NOS) (fig 1). Three isoforms of this enzyme are described (table 1).1 Neuronal NOS (nNOS or NOS1) is constitutively present in both the central and peripheral nervous systems, where NO acts as a neurotransmitter. Endothelial NOS (eNOS or NOS3) is constitutively expressed by endothelium and other cell types and is involved in cardiovascular homeostasis. In contrast, inducible NOS (iNOS or NOS2) is absent in resting cells, but the gene is rapidly expressed in response to stimuli such as proinflammatory cytokines. Once present, iNOS synthesises 100–1000 times more NO than the constitutive enzymes and does so for prolonged periods; the production of NO by eNOS and nNOS has been likened to a dripping tap, while that by iNOS to a fire hose. This high concentration of NO may inhibit a large variety of microbes, but may also potentially damage the host, thereby contributing to pathology.

Figure 1

Conversion of L-arginine to L-citrulline plus nitric oxide as catalysed by nitric oxide synthases. Two primary steps have been identified. The first step, a two electron oxidation, is a hydoxylation of one of the guanidino nitrogens of L-arginine requiring molecular oxygen and nicotinamide adenine diphosphate (NADPH) to form NG-hydroxy-L-arginine. The second step is a three electron oxidation, again requiring molecular oxygen and NADPH to perform an electron removal, oxygen insertion, and carbon-nitrogen bond scission to form L-citrulline and the free radical nitric oxide.

Table 1

Some characteristics of human nitric oxide synthases (NOS) (enzyme classification EC 1.14.23)

Antimicrobial effects of NO

NO inhibits the growth of many bacteria and parasites in vitro.2 The antimicrobial effect stems not from NO itself, but from reactive nitrogen intermediates formed by the oxidation of NO. For example, reaction between NO and the free radical superoxide (O2 ) results in the formation of the unstable molecule peroxynitrite (OONO), while that between NO and thiol groups produces nitrosothiols. These reactive nitrogen intermediates inactivate key microbial enzymes, such as ribonucleotide reductase and aconitase, by reacting with iron containing groups in these enzymes. Studies have shown that murine macrophages can produce sufficient NO to kill leishmania parasites in vitro.3 In vivo experiments using inhibitors of iNOS, and more recently knockouts of the iNOS gene, have clearly shown its role in the control of murine infections as diverse as malaria,4 leishmaniasis,5tuberculosis,6 and listeriosis.7

How much iNOS is produced by human leucocytes?

Although these observations undoubtedly indicate the potential importance of high concentration NO as an antimicrobial agent, a fundamental question remains as to the cellular source of iNOS derived NO in people. In mice it is well established that NO is produced by macrophages, but in humans this evidence has proved more elusive. The human iNOS gene can be expressed in various cell types (table 1), but it has been much more difficult to demonstrate high levels of iNOS expression by leucocytes,8 leading some to question the relevance of NO in human infection.9 However, there is a growing view that this may reflect the greater complexity of iNOS regulation, and possibly a greater degree of stimulus and tissue specificity in humans compared to mice.10 Whereas it is extremely difficult to stimulate leucocytes from healthy humans to produce iNOS in vitro, leucocytes from patients with inflammatory or infectious diseases have been shown both to express iNOS and to produce large amounts of NO. For example, NO producing leucocytes have been found in the peripheral blood of patients with septic shock11 and malaria,12 and in bronchoalveolar lavage fluid from patients with tuberculosis.13 Thus the issue is not whether human leucocytes can make NO, but what stimuli are required and whether they do so in sufficient quantities to have a significant antimicrobial and pathological impact.

Investigating NO production in disease

Most studies rely on reactive nitrogen intermediates (in practice measurement of nitrite and nitrate (NOx) in plasma or urine) as a surrogate for NO production. However NOxconcentrations are profoundly affected by diet, renal function, and hydration, all of which may be altered in severe illness.12 This has made interpretation of studies difficult and yielded conflicting results, especially when patients selected as controls have conditions which also result in iNOS expression. In addition, systemic concentrations of NOx may not reflect local production of NO in key organs, such as the spleen, bone marrow, or central nervous system. Measurements of other intermediates (such as nitrosothiols and nitrotyrosine, which reflect NO production and are not influenced by dietary nitrate), immunohistochemical investigation of iNOS expression, or investigation of genetic differences in the ability to produce NO may prove more useful in investigating the extent and site of NO production in inflammatory states.

NO and septic shock

Under physiological conditions eNOS derived NO activates soluble guanylate cyclase in vascular smooth muscle, increasing intracellular cyclic guanosine monophosphate and causing vasodilation. In sepsis, iNOS derived NO may mimic and exaggerate this process, causing chronic vasodilation, as well as interfering with and damaging other pathways. This makes iNOS an attractive candidate as a mediator of shock and multiorgan failure in sepsis. Animal models confirm widespread iNOS expression and NO production in response to Gram negative bacteria, endotoxin or inflammatory cytokines,14 resulting in hypotension, organ failure, and death. These effects are prevented by non-selective NOS inhibitors, such as NG-monomethyl-L-arginine (L-NMMA), which reverses systemic hypotension and improves survival in a number of animal models.15 iNOS knockout mice, in which the iNOS gene has been experimentally disrupted, are protected from endotoxin induced hypotension in some studies,7 although they are more susceptible to overwhelming infection with a number of intracellular organisms.10

There is good evidence for iNOS induction and increased NO production in human septic shock.11 Initial reports of the use of non-selective NOS inhibitors (such as L-NMMA) in those with refractory sepsis induced hypotension showed an increase in mean arterial pressure and systemic and pulmonary vascular resistance.16Enthusiasm waned when it emerged that non-selective NOS inhibition actually increased mortality in septic patients, possibly by its negative effects on cardiac output and organ perfusion.17This discrepancy between animal and human studies highlights the need to extrapolate animal studies cautiously. Intraperitoneal or intravenous infusion of a known amount of endotoxin or bacteria obviously differs from human infection, where the inoculum and stage of infection are unknown. Many animal studies pretreat with NOS inhibitors, before bacterial challenge, whereas in human trials pathological processes may be well advanced at presentation. More fundamentally, inhibition of a biological system with so many important physiological functions might be expected to have widespread and deleterious consequences. Current interest has focused on selective iNOS inhibitors, which might ameliorate the pathological effects of inappropriate NO production, but with less disruption of normal physiology. Although animal data support this approach,18a large trial of a selective iNOS inhibitor (546C88)17 has produced disappointing results.

NO and meningitis

In bacterial meningitis, meningeal inflammation is initiated by local production of proinflammatory cytokines, especially tumour necrosis factor (TNF) and interleukin 1β,19 which are both potent stimuli of iNOS expression in leucocytes, glia, and neurones. Cerebrospinal fluid NOx concentrations are increased in experimental20 and human meningitis, correlating with cerebrospinal fluid TNF concentrations in the latter.21 Immunohistochemical staining of postmortem tissue from experimental animals20 and humans (IA Clark, personal communication, 1999) show widespread iNOS expression and perimeningeal NO production. As before, NO may have a protective and pathological role; iNOS knockout mice are more susceptible to experimental Listeria monocytogenesinfection,7 and inhibition of NO worsens outcome in some models of bacterial meningitis, leading to focal cerebral ischaemia.20 Conversely many of the pathological changes seen in bacterial meningitis, such as disruption of the blood-brain barrier, cerebral oedema, changes in cerebral blood flow,22 and damage to cochlear cells,23 can be prevented experimentally by NOS inhibitors. Dexamethasone reduces iNOS transcription in vitro and this might contribute to its protective effect in reducing cochlear damage and thus residual deafness. Whether NO acts as a neurotoxic or neuroprotective molecule may be dependent on its redox state; in its oxidised form (NO+) inactivates glutamate receptors and reduces the neurotoxicity of excitatory amino acids, whereas the reduced form (NO•) can form peroxynitrate, a potent neurotoxin.24 The therapeutic use of NOS inhibitors in bacterial meningitis has been suggested, but as NO is likely to have profound effects on cerebral blood flow and other important physiological processes, our understanding of the role of NO in meningitis seems inadequate to justify this intervention at present.

NO and malaria

As in many parasitic infections, NO appears central to the host response in malaria, mediating the antiparasitic effects of proinflammatory cytokines at each stage of the parasite life cycle in both experimental and possibly human infection.25 In addition to its protective role, NO has been suggested as a mediator in cerebral malaria. Cerebral malaria is the most infamous of the severe manifestations of Plasmodium falciparuminfection, causing around 1 million deaths per year. The pathogenesis of cerebral malaria is debated, but derives from the sequestration of parasitised erythrocytes in the cerebral vasculature, where they may cause microvascular obstruction, local induction of inflammatory cytokines, or both. Clinically, there is often rapid recovery from profound coma without significant neurological sequelae, implying that a reversible process contributes to coma. It has been suggested that NO, produced by iNOS induced in vascular endothelium by proinflammatory cytokines, may cross the blood-brain barrier and affect local neuronal function, by mimicking and exaggerating the physiological effects of endogenous nNOS derived NO.25 Investigating this hypothesis has proved difficult, as plasma concentrations of NOx do not reflect local NO production within the brain. Some studies find that high concentrations of plasma NOxcorrelate with depth and length of coma in cerebral malaria,26 while others suggest that high plasma concentrations are protective against severe disease.12CSF NOx, which may be more reflective of local cerebral production and is less affected by diet and other confounders, is increased in cerebral malaria.27 Preliminary immunohistological studies suggest widespread iNOS induction in cerebral endothelium and resultant NO production (IA Clark, personal communication, 1999), although suitable control samples for such studies are difficult to obtain. Different genetic polymorphisms of the iNOS promoter region have been associated with increased risk of death from cerebral malaria in Gambian children28 and with protection from severe malaria in Gabonese children,29suggesting that complex genetic factors may determine iNOS production and thereby influence clinical outcome.


The L-arginine NO pathway is a vital antimicrobial defence mechanism. Because of the ubiquitous nature of this pathway in normal physiology, it is easy to see how high and continuous NO production in sepsis might contribute to pathology, either by interfering with NO dependent pathways or by causing direct tissue injury. Our knowledge of the role of NO in sepsis is incomplete. The regulation of iNOS expression is complex and may be cell type, differentiation stage, and stimulus dependent, making the extrapolation of animal studies to the clinical setting difficult. As our understanding increases, modulation of the NO L-arginine pathway in infection may become a valuable therapeutic option.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.