1 Nuffield Department of Anaesthetics, University of Oxford, Radcliffe Infirmary,Woodstock Road, Oxford OX2 6HE, UK
*Corresponding author: Department of Human Anatomy and Genetics, University of Oxford, South Parks Road, Oxford OX1 3QX, UK, E-mail: michael.reade@anaesthetics.oxford.ac.uk
Clinical studies investigating the pathogenesis of septic shock are fraught with problems. The focus of infection, the time between septic insult and presentation, and the varying treatments work together to produce a heterogeneous group of patients. There are ethical obstacles to observational studies on critically ill patients unable to give consent, as well as practical problems in obtaining specimens at a consistent time point in a rapidly evolving disease process. Compounding these problems, true septic shock is a relatively rare disease. In contrast, laboratory studies using cell cultures or rodent models are limited only by the available finance and expertise. As a result, the bulk of the literature investigating the pathogenesis of septic shock uses rodent models, or isolated or cultured cells. Therapeutic advances based on this research may well promise new hope for owners of rats requiring intensive care, but there is increasing evidence that human sepsis may be a fundamentally different disease. Conclusions based on animal models may not be applicable to humans.
There are important differences in the clinical picture produced by endotoxin infusion (or bacterial sepsis) between species. Dogs, rats and pigs are more resistant to endotoxin than humans, requiring a dose 10100 times greater to produce similar symptoms. These animals also typically have a reduced cardiac output and increased systemic vascular resistance in response to endotoxin; the reverse of human sepsis.1 This suggests that there are fundamental differences in the regulation of the inflammatory and cardiovascular response to endotoxin in different species.
Nitric oxide is considered central to the pathogenesis of septic shock. As well as acting as a vasodilator, it can stimulate inflammation,2 interfere with cellular oxygen utilization, form cytotoxic free radicals, and is a negative inotrope (reviewed by Titheradge3). There is little doubt that nitric oxide production increases in human sepsis. Plasma nitric oxide metabolite concentrations are increased in these patients, even when the confounding effect of renal failure is removed.4 The highest reported mean concentration of nitric oxide metabolites in a group of septic patients is 124 µmol litre1 (range 20187 µmol litre1),5 compared with 36 (1460) µmol litre1 in healthy subjects. Other studies report more modest increases to between 52,6 and 82 µmol litre1.7 However, animal studies show much more marked changes. All rodent studies of lipopolysaccharide (LPS) models of sepsis show nitric oxide metabolites rise to high concentrations: 430 µmol litre1 in rats,8 and 804 µmol litre1 in mice.9 This suggests that rodents may be more resistant to the pathological effects of nitric oxide, as might be expected of animals living in an unhygienic environment. Humans may also have evolved counter-regulatory mechanisms limiting nitric oxide production. Additionally, the lower concentrations of nitric oxide in humans might signify that there are other pathogenic mechanisms present to a greater degree in our species than in animals.
Abnormalities of the nitric oxide pathway have been identified in specific cells from patients with more chronic human diseases such as rheumatoid arthritis,10 and ulcerative colitis.11 However, the cells responsible for the increased nitric oxide production in either human or animal sepsis are yet to be determined. Studies of nitric oxide production in sepsis demonstrate tissue and species heterogeneity, along with differences in the response to different stimuli. Consensus views have been formed based on results of the most easily repeatable experiments, but that this variability exists suggests these models must be validated using tissue from humans with clinical disease.
A very large number of studies have demonstrated upregulation of the nitric oxide synthase (NOS) pathway in rodent macrophages stimulated with LPS or cytokines.12 13 Yet Schneemann and colleagues14 could not activate the nitric oxide synthetic pathway in human macrophages despite testing a very wide range of stimuli, all of which increased nitric oxide production in similarly treated murine cells. In contrast to results from murine macrophages studied under identical conditions, human peripheral blood mononuclear cells (PBMCs) had no change in their inducible NOS (iNOS) mRNA after incubation with LPS,12 and no LPS stimulated increase in nitric oxide production or NOS activity.15 PBMCs from patients with clinical sepsis did, however, produce more nitric oxide than controls,16 suggesting that human cells require some extra stimulus (not needed by animal cells) that was not present in any of the in vitro models studied to date. Cells of the immune system thus display both species differences, and differences in their in vitro and in vivo response.
Nitric oxide produced by vascular smooth muscle cells is potentially of greater pathophysiological significance than that from inflammatory cells, as it may mediate the vasoplegia seen in septic shock. It is now recognized that in the laboratory, vascular smooth muscle cells can produce nitric oxide, as well as acting as its target. Rat aortic smooth muscle cells in culture are easily stimulated to produce nitric oxide by a variety of substances.17 This nitric oxide is functionally significant; arterial rings denuded of endothelium from rats infused with LPS in vivo were hyporesponsive to norepinephrine, and this was reversed by an inhibitor of NOS.18 The arteries from rats with bacterial peritonitis induced by caecal ligation and puncture also had increased NOS activity. However, the contractile hyporesponsiveness of these arteries was not attributable to nitric oxide as, in contrast to the LPS model described, it was not reversed by an inhibitor of NOS.19 For rodent vascular smooth muscle cells too, there is evidence that disease induced by bacteria differs from experimental LPS infusion and in vitro activation.
Differences in the nitric oxide pathway between in vitro models and clinical disease are also present in human vascular smooth muscle. Human isolated aortic smooth muscle cells,20 or the media of human isolated internal mammary artery,21 when stimulated with LPS and cytokines, both express new iNOS mRNA. In vivo experimentation in humans is obviously difficult. The one volunteer study of human vessels activated with cytokines in vivo showed induced hyporesponsiveness because of nitric oxide, though surprisingly, there was no change in mRNA for any NOS isoform in these veins,22 in contrast to the findings of the human and animal in vitro studies. The same authors also performed the experiment with LPS, which induced a similar, but now surprisingly, non-nitric oxide-dependent hyporesponsiveness.23 Recall that LPS in the rodent model induced a hyporesponsiveness which, in contrast, was nitric oxide dependent.18 The response of vascular smooth muscle thus also appears highly species and stimulus specific.
The only studies of vessels taken from patients with clinical sepsis do implicate nitric oxide in the reduced contractility (compared to appropriate controls) observed in mesenteric,24 and omental25 arteries. N-nitro-L-arginine methyl ester (L-NAME; a NOS inhibitor) appeared to increase the contractile responsiveness of the septic omental arteries, while having no effect on controls.24 However, these studies involved a very small number of patients, neither sufficient to allow the identification of statistically significant differences. L-NAME also improved the responsiveness of the septic mesenteric vessels mentioned above,25 but unfortunately the effect of L-NAME on control tissue was not reported. This is also a common fault in clinical studies: NOS inhibitors do increase the systemic vascular resistance in patients with septic shock,26 but healthy subjects also have a marked pressor response to these compounds, invalidating the conclusion (based on inhibitor studies at least) that it is nitric oxide which is responsible for the hypotension of sepsis.
Arginine is the substrate for all of the isoforms of NOS, and transport of arginine across the plasma membrane is rate limiting for the production of nitric oxide,27 and the vasodilation so produced.28 Plasma membrane arginine transport may be of interest as a potential therapeutic target to limit nitric oxide production. On the other hand, maintenance of arginine transport may be helpful in conditions where NOS activity is increased, as iNOS produces highly toxic superoxide (and subsequently peroxynitrite) when deprived of adequate substrate.29 For these reasons, changes in plasma membrane arginine transport in various models of sepsis have been extensively studied. Most studies of arginine transport in animal models of sepsis have found an increase in the function of one type of arginine transporter, y+, of the subtype CAT2B. This is true of rat,30 and mouse,13 27 macrophages, as well as rat vascular smooth muscle cells,17 stimulated in vitro. Human endothelial cells,31 and T-cells,32 show similar effects after in vitro activation. As with all the factors affecting nitric oxide production, there is some variability depending on tissue, species and stimulus. For example, cytokine activated rat aortic smooth muscle cells upregulated an arginine transporter with kinetics more akin to another subtype, CAT1.33 In the same cells, CAT1 mRNA underwent a more long-lasting change when stimulated with angiotensin II.34 In contrast, it was CAT2A rather than CAT2B mRNA which was upregulated in mouse skeletal muscle after surgical trauma.35 To date, there is only one report of changes in arginine transporter expression or function in PBMCs from patients with septic shock; we found an increase in y+ transport and new expression of CAT2 mRNA.16 There are no similar studies of vascular smooth muscle from either clinically septic animals or humans.
The regulation of nitric oxide production in clinical sepsis is likely to result from an interplay between pro- and anti-inflammatory mediators. Nitric oxide feeds back to reduce its own production directly,36 and a number of anti-inflammatory cytokines present in clinical sepsis are known to reduce nitric oxide production.3 Another possible modifier of nitric oxide production is the haem oxygenase (HO)/carbon monoxide system.
HO catalyses the conversion of haem to biliverdin and carbon monoxide. There is one inducible (HO-1) isoform and two constitutive (HO-2 and HO-3) isoforms. In cell lines and rodent models, the amount and activity of inducible HO (HO-1) is increased in conditions causing oxidative stress such as LPS exposure or bacterial sepsis.37 The exhaled breath of critically ill patients contains increased concentrations of carbon monoxide,38 and their plasma carbon monoxide concentrations are also increased.39 To date there have been no studies of the cellular origin of this increased carbon monoxide production in human sepsis. Carbon monoxide interacts with nitric oxide in many ways. Nitric oxide is a strong stimulus to the induction of HO-1, acting at multiple levels (as summarized by Hartsfield40). Conversely, there are several mechanisms whereby HO-1 directly, or through carbon monoxide, can reduce nitric oxide production.40 The relative amounts of HO activity in human and animal sepsis are unknown. If HO is upregulated to a greater degree in humans than animals, this could explain both the lesser nitric oxide production by septic humans (because of carbon monoxide inhibition of nitric oxide), and also the non-nitric oxide-dependent link between sepsis and shock, which has been hypothesized to explain the findings reported by Bhagat and colleagues.23
Increasing knowledge of the complex dynamic mixture of pro- and anti-inflammatory cytokines present in clinical sepsis, along with the rapid turnover of cells (particularly inflammatory cells), makes the observed shortcomings of using in vitro cell models and in vivo LPS/cytokine infusions perhaps not so surprising. However, it is much easier to experiment on readily available cells, which behave in a predictable manner, than it is to collect clinical samples from the relatively heterogeneous group of patients diagnosed with sepsis. This certainly explains the very large number of studies investigating nitric oxide biology performed using rat and mouse macrophages, and rat aortic smooth muscle cells. Much knowledge has been accumulated on the nitric oxide biology of these cells, and, to a lesser extent, on human cells activated in vitro. There is a strong temptation to generalize these findings to human clinical disease. However, this brief review has highlighted a number of cell, species and stimulus differences which suggest that, at the very least, in vitro and animal models must be validated by comparison with the human clinical disease. While accepting that there are ethical and practical limitations on the scope of human studies, to not do this runs the risk of developing excellent treatments for rat sepsis that are of no use to humans at all.
Acknowledgement
We thank Dr C.A.R. Boyd for his helpful suggestions during the preparation of the manuscript.
References
1 Spitzer J. Animal models of endotoxaemia and sepsis. In: Schlag G, Redl H, eds. Pathophysiology of Shock, Sepsis and Organ Failure. Berlin: Springer-Verlag, 1993
2 Lander HM, Sehajpal P, Levine DM, Novogrodsky A. Activation of human peripheral blood mononuclear cells by nitric oxide generating compounds. J Immunol 1993; 150: 150916
3 Titheradge MA. Nitric oxide in septic shock. Biochim Biophys Acta 1999; 1411: 43755[ISI][Medline]
4 MacKenzie IM, Garrard CS, Young JD. Indices of nitric oxide synthesis and outcome in critically ill patients. Anaesthesia 2001; 56: 32630[CrossRef][ISI][Medline]
5 Evans T, Carpenter A, Kinderman H, Cohen J. Evidence of increased nitric oxide production in patients with the sepsis syndrome. Circ Shock 1993; 41: 7781[ISI][Medline]
6 Groeneveld PH, Kwappenberg KM, Langermans JA, Nibbering PH, Curtis L. Nitric oxide (NO) production correlates with renal insufficiency and multiple organ dysfunction syndrome in severe sepsis. Intensive Care Med 1996; 22: 1197202[CrossRef][ISI][Medline]
7 Krafte-Jacobs B, Brilli R, Szabo C, Denenberg A, Moore L, Salzman AL. Circulating methemoglobin and nitrite/nitrate concentrations as indicators of nitric oxide overproduction in critically ill children with septic shock. Crit Care Med 1997; 25: 158893[ISI][Medline]
8 Sakemi K, Ohno Y, Tsuda M. NO2-/NO3-levels in blood and principal organs in rats treated with lipopolysaccharide. Kokuritsu Iyakuhin Shokuhin Eisei Kenkyusho Hokoku 1998; 116: 1016[Medline]
9 Harbrecht BG, Di Silvio M, Demetris AJ, Simmons RL, Billiar TR. Tumor necrosis factor-alpha regulates in vivo nitric oxide synthesis and induces liver injury during endotoxemia. Hepatology 1994; 20: 105560[ISI][Medline]
10 Sakurai H, Kohsaka H, Liu MF, et al. Nitric oxide production and inducible nitric oxide synthase expression in inflammatory arthritides. J Clin Invest 1995; 96: 235763[ISI][Medline]
11 Godkin AJ, De Belder AJ, Villa L, et al. Expression of nitric oxide synthase in ulcerative colitis. Eur J Clin Invest 1996; 26: 86772[ISI][Medline]
12 Chesrown SE, Monnier J, Visner G, Nick HS. Regulation of inducible nitric oxide synthase mRNA levels by LPS, INF-gamma, TGF-beta, and IL-10 in murine macrophage cell lines and rat peritoneal macrophages. Biochem Biophys Res Commun 1994; 200: 12634[CrossRef][ISI][Medline]
13 Caivano M. Role of MAP kinase cascades in inducing arginine transporters and nitric oxide synthetase in RAW264 macrophages. FEBS Lett 1998; 429: 24953[CrossRef][ISI][Medline]
14 Schneemann M, Schoedon G, Hofer S, Blau N, Guerrero L, Schaffner A. Nitric oxide synthase is not a constituent of the antimicrobial armature of human mononuclear phagocytes. J Infect Dis 1993; 167: 135863
15 Weinberg JB, Misukonis MA, Shami PJ, et al. Human mononuclear phagocyte inducible nitric oxide synthase (iNOS): analysis of iNOS mRNA, iNOS protein, biopterin, and nitric oxide production by blood monocytes and peritoneal macrophages. Blood 1995; 86: 118495
16 Reade MC, Clark MF, Young JD, Boyd CAR. Increased cationic amino acid flux through a newly expressed transporter in cells overproducing nitric oxide from patients with septic shock. Clin Sci (Lond) 2002; 102: 64550[ISI][Medline]
17 Gill DJ, Low BC, Grigor MR. Interleukin-1 beta and tumor necrosis factor-alpha stimulate the CAT2 gene of the L-arginine transporter in cultured vascular smooth muscle cells. J Biol Chem 1996; 271: 112803
18 Julou-Schaeffer G, Gray GA, Fleming I, Schott C, Parratt JR, Stoclet JC. Loss of vascular responsiveness induced by endotoxin involves L-arginine pathway. Am J Physiol 1990; 259: H103843
19 Vromen A, Arkovitz MS, Zingarelli B, Salzman AL, Garcia VF, Szabo C. Low-level expression and limited role for the inducible isoform of nitric oxide synthase in the vascular hyporeactivity and mortality associated with cecal ligation and puncture in the rat. Shock 1996; 6: 24853[ISI][Medline]
20 MacNaul KL, Hutchinson NI. Differential expression of iNOS and cNOS mRNA in human vascular smooth muscle cells and endothelial cells under normal and inflammatory conditions. Biochem Biophys Res Commun 1993; 196: 13304[CrossRef][ISI][Medline]
21 Chester AH, Borland JA, Buttery LD, et al. Induction of nitric oxide synthase in human vascular smooth muscle: interactions between proinflammatory cytokines. Cardiovasc Res 1998; 38: 81421[CrossRef][ISI][Medline]
22 Bhagat K, Hingorani AD, Palacios M, Charles IG, Vallance P. Cytokine-induced venodilatation in humans in vivo: eNOS masquerading as iNOS. Cardiovasc Res 1999; 41: 75464[CrossRef][ISI][Medline]
23 Bhagat K, Collier J, Vallance P. Local venous responses to endotoxin in humans. Circulation 1996; 94: 4907
24 Stoclet JC, Martinez MC, Ohlmann P, et al. Induction of nitric oxide synthase and dual effects of nitric oxide and cyclooxygenase products in regulation of arterial contraction in human septic shock. Circulation 1999; 100: 10712
25 Tsuneyoshi I, Kanmura Y, Yoshimura N. Nitric oxide as a mediator of reduced arterial responsiveness in septic patients. Crit Care Med 1996; 24: 10836[CrossRef][ISI][Medline]
26 Petros A, Lamb G, Leone A, Moncada S, Bennett D, Vallance P. Effects of a nitric oxide synthase inhibitor in humans with septic shock. Cardiovasc Res 1994; 28: 349[ISI][Medline]
27 Closs EI, Scheld JS, Sharafi M, Forstermann U. Substrate supply for nitric-oxide synthase in macrophages and endothelial cells: role of cationic amino acid transporters. Mol Pharmacol 2000; 57: 6874
28 Hishikawa K, Nakaki T, Tsuda M, et al. Effect of systemic L-arginine administration on hemodynamics and nitric oxide release in man. Jpn Heart J 1992; 33: 418[ISI][Medline]
29 Xia Y, Zweier JL. Superoxide and peroxynitrite generation from inducible nitric oxide synthase in macrophages. Proc Natl Acad Sci USA 1997; 94: 69548
30 Hammermann R, Dreissig MD, Mossner J, et al. Nuclear factor-kappa B mediates simultaneous induction of inducible nitric oxide synthase and Up-regulation of the cationic amino acid transporter CAT-2B in rat alveolar macrophages. Mol Pharmacol 2000; 58: 1294302
31 Irie K, Tsukahara F, Fujii E, et al. Cationic amino acid transporter-2 mRNA induction by tumor necrosis factor-alpha in vascular endothelial cells. Eur J Pharmacol 1997; 339: 28993[CrossRef][ISI][Medline]
32 Boyd CAR, Crawford DH. Activation of cationic amino acid transport through system y+ correlates with expression of the T-cell early antigen gene in human lymphocytes. Pflügers Arch 1992; 422: 879[ISI][Medline]
33 Durante W, Liao L, Schafer AI. Differential regulation of L-arginine transport and inducible NOS in cultured vascular smooth muscle cells. Am J Physiol 1995; 268: H115864
34 Low BC, Grigor MR. Angiotensin II stimulates system y+ and cationic amino acid transporter gene expression in cultured vascular smooth muscle cells. J Biol Chem 1995; 270: 2757783
35 Kakuda DK, Finley KD, Maruyama M, MacLeod CL. Stress differentially induces cationic amino acid transporter gene expression. Biochim Biophys Acta 1998; 1414: 7584[ISI][Medline]
36 Buga GM, Griscavage JM, Rogers NE, Ignarro LJ. Negative feedback regulation of endothelial cell function by nitric oxide. Circ Res 1993; 73: 80812[Abstract]
37 Camhi SL, Alam J, Otterbein L, Sylvester SL, Choi AM. Induction of heme oxygenase-1 gene expression by lipopolysaccharide is mediated by AP-1 activation. Am J Respir Cell Mol Biol 1995; 13: 38798
38 Scharte M, Bone HG, Van Aken H, Meyer J. Increased carbon monoxide in exhaled air of critically ill patients. Biochem Biophys Res Commun 2000; 267: 4236[CrossRef][ISI][Medline]
39 Shi Y, Pan F, Li H, et al. Plasma carbon monoxide levels in term newborn infants with sepsis. Biol Neonate 2000; 78: 2302[CrossRef][ISI][Medline]
40 Hartsfield CL. Cross talk between carbon monoxide and nitric oxide. Antioxid Redox Signal 2002; 4: 3017[CrossRef][ISI][Medline]