Septic shock results from the dysregulation of the innate immune response following infection. Despite major advances in fundamental and clinical research, patients diagnosed with septic shock still have a poor prognostic outcome, with a mortality rate of up to 50%. Indeed, the reasons leading to septic shock are still poorly understood. First postulated 30 years ago, the general view of septic shock as an acute and overwhelming inflammatory response still prevails today. Recently, the fact that numerous clinical trials have failed to demonstrate any positive medical outcomes has caused us to question our fundamental understanding of this condition. New and sophisticated technologies now allow us to accurately profile the various stages and contributory components of the inflammatory response defining septic shock, and many studies now report a more complex inflammatory response, particularly during the early phase of sepsis. In addition, novel experimental approaches, using more clinically relevant animal models, to standardize and stratify research outcomes are now being argued for. In the present review, we discuss the most recent findings in relation to our understanding of the underlying mechanisms involved in septic shock, and highlight the attempts made to improve animal experimental models. We also review recent studies reporting promising results with two vastly different therapeutic approaches influencing the renin–angiotensin system and applying mesenchymal stem cells for clinical intervention.

Sepsis and the severe systemic response syndrome are very common illnesses that are responsible for a great amount of morbidity and death. These closely related conditions are characterized by a remarkable increase in the prohormone ProCT (procalcitonin). ProCT is both a marker of sepsis and a harmful mediator of the disease. In the present issue of Clinical Science , in a study in rats with endotoxin shock, Tavares and Miñano used an antibody to a segment of N-ProCT (aminoprocalcitonin) that is part of the ProCT molecule, and confirmed that immunoneutralization of ProCT saves the animals from this severe illness. Furthermore, they extensively studied the epiphenomena associated with this immunoneutralization.

Severe sepsis and septic shock are an important cause of mortality and morbidity. These illnesses can be triggered by the bacterial endotoxin LPS (lipopolysaccharide) and pro-inflammatory cytokines, particularly TNF-α (tumour necrosis factor-α) and IL (interleukin)-1β. Severity and mortality of sepsis have also been associated with high concentrations of N-PCT (aminoprocalcitonin), a 57-amino-acid neuroendocrine peptide derived from ProCT (procalcitonin). Previous studies in a lethal model of porcine polymicrobial sepsis have revealed that immunoneutralization with IgG that is reactive to porcine N-PCT significantly improves short-term survival. To explore further the pathophysiological role of N-PCT in sepsis, we developed an antibody raised against a highly conserved amino acid sequence of human N-PCT [N-PCT-(44–57)]. This sequence differs by only one amino acid from rat N-PCT. First, we demonstrated the specificity of this antibody in a well-proven model of anorexia induced in rats by central administration of human N-PCT-(1–57). Next we explored further the therapeutic potential of anti-N-PCT-(44–57) in a rat model of lethal endotoxaemia and determined how this immunoneutralization affected LPS-induced lethality and cytokine production. We show that this specific antibody inhibited the LPS-induced early release of TNF-α and IL-1β and increased survival, even if treatment began after the cytokine response had occurred. In addition, anti-N-PCT-(44–57) may increase long-term survival in LPS-treated rats by up-regulating the late production of counter-regulatory anti-inflammatory mediators such as ACTH (adrenocorticotropic hormone) and IL-10. In conclusion, these results support N-PCT as a pro-inflammatory factor in both the early and the late stages of lethal endotoxaemia, and suggest anti-N-PCT as a candidate for septic shock therapy.

Arginine has vasodilatory effects, via its conversion by NO synthase into NO, and immunomodulatory actions which play important roles in sepsis. Protein breakdown affects arginine availability and the release of asymmetric dimethylarginine, an inhibitor of NO synthase, may therefore affect NO synthesis in patients with sepsis. The objective of the present study was to investigate whole-body in vivo arginine and citrulline metabolism and NO synthesis rates, and their relationship to protein breakdown in patients with sepsis or septic shock and in healthy volunteers. Endogenous leucine flux, an index of whole-body protein breakdown rate, was measured in 13 critically ill patients with sepsis or septic shock and seven healthy controls using an intravenous infusion of [1- 13 C]leucine. Arginine flux, citrulline flux and the rate of conversion of arginine into citrulline (an index of NO synthesis) were measured with intravenous infusions of [ 15 N 2 ]guanidino-arginine and [5,5- 2 H 2 ]citrulline. Plasma concentrations of nitrite plus nitrate, arginine, citrulline and asymmetric dimethylarginine were measured. Compared with controls, patients had a higher leucine flux and higher NO metabolites, but arginine flux, plasma asymmetric dimethylarginine concentration and the rate of NO synthesis were not different. Citrulline flux and plasma arginine and citrulline were lower in patients than in controls. Arginine production was positively correlated with the protein breakdown rate. Whole-body arginine production and NO synthesis were similar in patients with sepsis and septic shock and healthy controls. Despite increased proteolysis in sepsis, there is a decreased arginine plasma concentration, suggesting inadequate de novo synthesis secondary to decreased citrulline production.

Thrombin is involved in various inflammatory responses. In sepsis, coagulation abnormalities are major complications. Acute lung injury is one of the most life-threatening problems that can result from sepsis. We hypothesized that high-dose heparin might be effective in attenuating acute lung injury in our sepsis model. Female sheep ( n = 16) were surgically prepared for the study. After a tracheotomy, 48 breaths of cotton smoke (<40°C) were insufflated into the airway. Afterwards, live Pseudomonas aeruginosa (5×10 11 colony-forming units) bacteria were instilled into the lung. All sheep were ventilated mechanically with 100% O 2 , and were divided into three groups: a heparin infusion group ( n = 6), a Ringer's lactate infusion group ( n = 6), and a sham-injury group ( n = 4; surgically prepared in the same fashion but receiving no inhalation injury or bacteria). The treatment was started 1h after the insult, and was continued thereafter for 24h. The dose of heparin was adjusted by monitoring to target an activated clotting time of between 300 and 400s (baseline = approx. 150s). Sheep exposed to lung injury presented with typical hyperdynamic cardiovascular changes, including an increased cardiac output and a fall in systemic vascular resistance. There was a decrease in the arterial partial pressure of O 2 . In conclusion, high-dose heparin did not prevent lung dysfunction in this model, in which acute lung injury was induced by combined smoke and septic challenge.

Increased production of nitric oxide (NO) is thought to be a factor in the pathogenesis of many human diseases - among them the hypotension that often accompanies sepsis. The supply of the cationic amino acid arginine is known to be rate-limiting for NO production. We hypothesized that cationic amino acid transport might be increased in cells producing excess NO from patients with septic shock. Peripheral blood mononuclear cells were isolated from patients with sepsis and from healthy control subjects. The rates of both NO production and cationic amino acid uptake were increased in cells from patients with septic shock. The increased transport was due almost entirely to an increase in the activity of one transporter, subtype y + . The activity of the other major cationic amino acid transporter (y + L) was unchanged. The expression of CAT2 mRNA, which encodes a y + transporter protein, was also increased in these cells. We suggest that CAT2 might be a therapeutic target to prevent excess NO production in sepsis and possibly other human disease states, while leaving basal production unchanged.

We hypothesized that cytokine production following delayed in vitro cell stimulation (to reproduce physiological cellular status at baseline) may be related to outcome in patients with septic shock. A total of 20 patients were included in a prospective clinical study, conducted in a medico-surgical intensive care unit in a university hospital. Blood samples were obtained at the onset of septic shock; these were treated to retain the cells, but to wash out autologous plasma (containing potential inflammatory stimuli such as cytokines, bacterial products and drugs) and replace it with foetal calf serum. Each treated sample was divided into two sets of four aliquots, to be stimulated either immediately or after an overnight period of resting incubation at 37°C. The rest period was to allow recovery from potentially reversible endogenous or pharmacologically induced alterations in cellular response, in order to reproduce a near physiological state at baseline. In vitro cellular challenges used low-dose (0.2ng/ml) or high-dose (1ng/ml) CD14-dependent lipopolysaccharide and CD14-independent pokeweed mitogen to induce the production of tumour necrosis factor-α (TNF-α), and interleukins-1β and -10. Levels of TNF-α, interleukin-1β and interleukin-10 were significantly higher ( P < 0.05) when cell stimulation was delayed for 16h, indicating a functional down-regulation of cells during septic shock. Moreover, TNF-α responses obtained with high-dose lipopolysaccharide were significantly greater in cells from patients who subsequently survived septic shock ( n = 13; median value 1392pg/ml; range 592-2048pg/ml) than in cells from non-survivors ( n = 7; median value 708 pg/ml; range 520-1344pg/ml). These observations support the existence of individual differences in the inflammatory response that could influence patient outcome following septic shock.

We investigated whether organ-specific differences exist in the role of inducible nitric oxide synthase (iNOS) in hyporeactivity to vasoconstrictors following 20h in vitro exposure of isolated superior mesenteric, renal, hepatic and coronary arteries from the rat to bacterial lipopolysaccharide (LPS). LPS attenuated contraction in response to depolarizing KCl in all arteries. Maximum contractile responses to noradrenaline were attenuated in superior mesenteric and hepatic arteries, and those to the thromboxane A 2 analogue U46619 were attenuated in coronary arteries. LPS shifted the concentration-response curve to noradrenaline in renal arteries to the right. Removal of extracellular l -arginine improved the response to noradrenaline in superior mesenteric and renal arteries only. Addition of the iNOS inhibitor aminoguanidine resulted in full recovery of the responses to noradrenaline in superior mesenteric, renal and hepatic arteries. Contractile responses in coronary arteries did not improve after inhibition of iNOS activity. Therefore the pattern of the LPS-induced changes in vascular reactivity, as well as the contribution of iNOS to impaired vascular constriction, differed among vascular beds. These differences are likely to represent a contributory factor in the sepsis-associated redistribution of cardiac output.

To obtain predictors of organ failure (OF), we studied markers of systemic inflammation [circulating levels of interleukin-6 (IL-6), IL-8, soluble IL-2 receptor (sIL-2R), soluble E-selectin and C-reactive protein, and neutrophil and monocyte CD11b expression] and routine blood cell counts in 20 patients with systemic inflammatory response syndrome and positive blood culture. Eight patients with shock due to community-acquired infection developed OF, whereas 11 normotensive patients and one patient with shock did not (NOF group). The first blood sample was collected within 48 h after taking the blood culture (T1). OF patients, as compared with NOF patients, had at T1 a lower monocyte count, a lower platelet count, higher levels of CD11b expression on both neutrophils and monocytes, and higher concentrations of IL-6, IL-8 and sIL-2R. C-reactive protein and soluble E-selectin concentrations did not differ between groups. No parameter alone identified all patients that subsequently developed OF. However, a sepsis-related inflammation severity score (SISS), developed on the basis of the presence or absence of shock and on the levels of markers at T1, identified each patient that developed OF. The maximum SISS value was 7. The range of SISS values in OF patients was 2–5, and that in NOF patients was 0–1. In conclusion, high levels of CD11b expression, depressed platelet and monocyte counts, and high concentrations of IL-6, IL-8 and sIL-2R predict OF in patients with community-acquired septic shock, and the combination of these markers may provide the means to identify sepsis patients who will develop OF.

1. We measured nitric oxide synthase activity in peripheral blood polymorphonuclear leucocytes from 10 patients with sepsis syndrome and 10 healthy subjects. 2. Synthase activity was significantly higher in patients with sepsis than in control subjects (1202 ± 579 compared with 595 ± 544 pmol of nitric oxide min −1 mg −1 of cell protein, P < 0.05). 3. Activity was greatest in those patients with the larger number of organ failures, although this failed to reach significance (1489 ± 560 in patients with three or more organ failures and 843 ± 404 pmol of nitric oxide min −1 mg −1 of cell protein in those with less than three, P = 0.11). 4. This study provides evidence for the role of overproduction of the vasodilator nitric oxide in sepsis syndrome.

1. Endotoxin induces a shock-like syndrome with increased nitric oxide synthesis. To clarify the cellular source of NO in endotoxic shock we used immunohistochemistry and in situ hybridization to localize inducible NO synthase in rats given lipopolysaccharide or Corynebacterium parvum and lipopolysaccharide. Immunohistochemistry was carried out with an antibody raised against a synthetic peptide of mouse macrophage NO synthase. In situ hybridization was performed with 35 S-labelled oligonucleotide probes corresponding to cDNA sequences common to mouse macrophage inducible NO synthase and rat vascular smooth inducible NO synthase. Monocytes and macrophages were identified by immunohistochemistry with the mouse monoclonal antibody ED1. 2. After lipopolysaccharide alone, the major site of NO synthase induction was monocytes and macrophages in multiple organs, principally liver and spleen. Bronchial, bile duct, intestinal and bladder epithelium and some hepatocytes also expressed inducible NO synthase. Expression peaked at 5 h and had returned to normal by 12 h except in spleen. 3. After priming with C. parvum , lipopolysaccharide led to a similar distribution of inducible NO synthase as lipopolysaccharide alone, but in addition there was more prominent hepatocyte staining, staining in macrophage granulomas in the liver and inducible NO synthase was present in some endothelial cells in the aorta. 4. These findings provide a direct demonstration of the cellular localization of inducible NO synthase after lipopolysaccharide.

1. Endothelin, a novel vasoconstrictor 21-residue peptide isolated from the supernatant of cultured porcine endothelial cells, has been shown to be increased in plasma in a variety of cardiovascular disease states, including acute myocardial infarction, acute renal failure and essential hypertension. We determined the time course of plasma and pulmonary lymph endothelin-like immunoreactivity in relation to the progressive deterioration of cardiopulmonary function in an ovine septic shock model leading to multi-organ failure syndrome and death within 42 h of a continuous intravenous infusion of Escherichia coli endotoxin (40 ng min −1 kg −1 ). 2. Plasma and pulmonary lymph endothelin-like immunoreactivity were measured by r.i.a. using a specific antiserum raised in rabbits against porcine endothelin-1. Endothelin-like immunoreactivity was further determined in lung tissue and the thoracic duct lymph of endotoxin-treated sheep by reversed-phase h.p.l.c. In control instrumented conscious sheep not infused with endotoxin, there were no significant changes in any of the measured cardiopulmonary and biochemical variables, with plasma and pulmonary lymph endothelin-like immunoreactivity remaining below the detection limit (< 1 pg/tube) throughout the 72 h study period. 3. Conscious sheep receiving endotoxin showed a major hypotensive septic syndrome, including persistently decreased systemic blood pressure, systemic vascular resistance, stroke volume, left ventricular stroke work, associated with sustained pulmonary vasoconstriction and protein-rich pulmonary oedema (> five-fold increase in pulmonary lymph flow and protein clearance), and marked lactic acidosis, leading to the death of animals within 14–42 h despite institution of mechanical ventilation and adequate intravascular volume replacement. 4. Appearance of endothelin-like immunoreactivity, as revealed by r.i.a., in arterial plasma and pulmonary lymph was simultaneous in both circulatory beds, with peak values measured between 4 and 12 h after the start of endotoxin infusion (plasma: 68 ± 8 pg/ml, pulmonary lymph: 88 ± 18 pg/ml, P < 0.05 compared with control sheep). After 12 h of endotoxaemia, endothelin-like immunoreactivity in both fluids progressively decreased up to the death of the animals, although remaining significantly above that measured in control sheep. The analysis of extracts of lung and thoracic duct by reversed-phase h.p.l.c. revealed that the r.i.a. method used in the present study mainly detected endothelin-1. 5. Our results demonstrate the presence of a marked and persistent increase in endothelin-like immunoreactivity in plasma and pulmonary lymph of sheep during lethal endotoxin shock with multi-organ failure, suggesting a continuous production and/or release of endothelin-1 into the pulmonary lymph and the systemic circulation upon continuous endotoxin infusion. These findings suggest that endothelin may contribute to the vasomotor disturbances observed during the development of septic shock, although studies using selective receptor antagonists or synthesis inhibitors are required to definitively confirm a potential pathophysiological role of endothelin during endotoxaemia.

1. The cardiovascular effects of intravenous injections of interleukin-1 (IL-1) and tumour necrosis factor (TNF) have been investigated in the conscious rabbit. They have been compared with the effects of bacterial lipopolysaccharide (LPS) because both IL-1 and TNF are released from macrophages by LPS. 2. IL-1, TNF and Escherichia coli J5-LPS all caused hypotension when given intravenously in a dose with low mortality. The time course of the hypotension caused by IL-1 and LPS was similar, although the maximal fall in mean blood pressure occurred earlier after IL-1. TNF produced a more sustained fall in blood pressure. Hypotension was not accompanied by a compensatory tachycardia after any of the test substances. Hypotension was associated with a fever after TNF, hypothermia after LPS and no significant change in temperature after IL-1. 3. The packed cell volume did not change during hypotension in any of the study groups, implying that the hypotension was not due to fluid loss resulting from increased capillary permeability. 4. IL-1 and TNF are candidates for the role of effectors of LPS-induced hypotension.