The acute splanchnic and peripheral tissue metabolic response to endotoxin in humans

Bacterial translocation: the influence of dietary variables

Transmucosal passage of bacteria in critically ill patients may lead to a significant incidence of systemic sepsis. This has attracted much clinical interest, as it has been shown that malnutrition in itself, impairs various aspects of barrier function. Bacterial translocation is increased in animal models where nutrients are given by the parenteral route, while enteral feeding reverses this. Translocation is also considerably increased in response to a non-lethal endotoxin challenge, if there is pre-existing protein energy malnutrition. Similar results have been obtained where the insult is caused by the inflammatory agent, zymosan. Dietary fibre reduces the deleterious effects of either agent on translocation, although the type of fibre is important. Bulk forming but non-fermentable fibres are more effective than easily fermentable types (for example, pectin). Glutamine was not effective in preventing elemental diet induced bacterial translocation. Thus, although fermentable fibre and glutamine have positive effects on mucosal mass, they do not affect translocation. Enteral nutrition thus seems to be superior to parenteral nutrition in maintaining the functional barrier of the gut. A clearer understanding of the physiology of these effects may lead to use of specifically modified enteral diets in the critically ill patient.

Targets for sepsis therapies: tumor necrosis factor versus interleukin-1

Clinical testing of therapies for sepsis that target tumor necrosis factor and interleukin-1 is currently in progress. It is now clear from early clinical results that patients display a heterogeneous response to anti-cytokine therapies that may be related to the stage or severity of disease. Experimental results show that either cytokine may contribute to the metabolic derangements that lead to organ dysfunction in severe sepsis. They also provide a physiological basis for understanding the benefits of anti-cytokine therapy in the most severely ill patients.

Endotoxin increases parathyroid hormone-related protein mRNA levels in mouse spleen. Mediation by tumor necrosis factor

Parathyroid hormone-related protein (PTHrP) causes hypercalcemia in malignancy. However, the role and regulation of PTHrP in normal physiology is just beginning to be explored. PTHrP is found in the spleen and has several other features common to cytokines. Since endotoxin (LPS) causes many of its effects indirectly by inducing cytokines, studies were undertaken to determine whether LPS might also induce splenic PTHrP expression. LPS (100 ng/mouse) increased splenic PTHrP mRNA levels 3.6-fold in C3H/OuJ mice. This effect was maximal at 2 h and returned to baseline by 4 h. PTHrP peptide levels also increased 3.3-fold in splenic extracts in response to LPS (1 microgram/mouse). Murine TNF-alpha and human IL-1 beta, cytokines that mediate many of the effects of LPS, also increased splenic PTHrP mRNA levels. LPS-resistant C3H/HeJ mice, which produce minimal amounts of TNF and IL-1 in response to LPS, were resistant to LPS induction of splenic PTHrP mRNA, while TNF-alpha and IL-1 beta readily increased PTHrP mRNA levels in C3H/HeJ mice. Anti-TNF antibody blocked LPS induction of splenic PTHrP mRNA in C3H/OuJ mice by 68%, indicating that TNF is a mediator of the LPS induction of PTHrP levels. In contrast, an IL-1 receptor antagonist (IL-1ra) was ineffective. The increase in PTHrP in the spleen during the immune response suggests that PTHrP may play an important role in immune modulation, perhaps by mediating changes in lymphocyte proliferation and/or function.

Effect of anesthesia and surgery on plasma cytokine levels

Cytokines released in response to stress may have a profound impact on circulatory stability. There is no information on the effect of general anesthesia alone on plasma cytokine levels and little information on cytokine release following surgery. Plasma cytokine levels and hemodynamic parameters were measured during anesthesia and abdominal surgery under sterile and nonpyrogenic conditions in seven pigs anesthetized with ketamine and pentobarbital. Tumor necrosis factor (TNF) was measured by bioassay. Bioassays of low and high sensitivity were used to measure interleukin 6 (IL-6). Measurements were made sequentially during: (1) 4 hours observation with anesthesia alone; (2) 2 hours following laparotomy and traumatic intestinal manipulation (IM) sufficient to produce shock; and (3) after an intravenous bolus of 1 microgram/kg endotoxin as a positive control. Arterial blood pressure decreased following IM from 91.5 +/- 5.8 to 48.6 +/- 3.2 mm Hg, (mean +/- SE, P < .05), with no further change following endotoxin. Heart rate was unchanged during the experiment, and central venous pressure decreased after endotoxin (P < .05). There were no increases in TNF or IL-6 (using a low sensitivity assay) with anesthesia alone or following IM with shock, but both increased after endotoxin administration (P < .05); using a high sensitivity assay, IL-6 did not change during anesthesia alone but did increase fivefold following IM with shock (P < .05) and 50-fold following endotoxin administration (P < .05). We conclude that in a porcine model under sterile and nonpyrogenic conditions, prolonged anesthesia does not increase plasma cytokine levels.(ABSTRACT TRUNCATED AT 250 WORDS)

Elemental diet-induced bacterial translocation can be hormonally modulated

BACKGROUND

The authors have previously documented that feeding mice an elemental diet resulted in bacterial translocation (BT) that could be prevented by the provision of dietary fiber. To test whether the protective effect of fiber was related to the stimulation of trophic gut hormones, the effects of sandostatin and bombesin were tested.

METHODS

Mice fed either chow or the elemental diet were stratified into several groups and the ability of bombesin (10 micrograms/kg, tid) or sandostatin (100 micrograms/kg bid) to modulate BT was examined. After 14 days, mice were sacrificed and BT, cecal bacterial population levels, mucosal protein, and small bowel weight was measured. Segments of the ileum and jejunum were examined histologically.

RESULTS

Incidence of elemental diet-induced BT (75%) was reduced by fiber (9%) or the administration of bombesin (13%) (p < 0.01). Although sandostatin did not promote BT in chow-fed mice, it reversed the protective effect of fiber on BT (75%) (p < 0.01).

CONCLUSION

Elemental diet-induced bacterial translocation can be modulated hormonally and the beneficial effects of fiber on diet-induced BT appears to be hormonally mediated.

Response of man to endotoxin

Endotoxin, a cell wall component of Gram-negative bacteria, plays a central role in the pathogenesis of septic shock. By administering small doses of intravenous endotoxin to humans, a variety of acute inflammatory responses are induced which are qualitatively similar to those that occur during the early stages of septic shock. Within hours of the administration of intravenous endotoxin to human volunteers, changes occur in systemic hemodynamics, ventricular function, pulmonary gas exchange and permeability. In conjunction with these changes in organ function, a wide variety of inflammatory mediators are released which appear to contribute to these responses. These include the release of proinflammatory cytokines (e.g. tumor necrosis factor-alpha, IL-1 beta, IL-6, IL-8), activation of the fibrinolytic system, kallikrein-kinin generation and phospholipase A2 release. Phagocytic leukocytes are primed for enhanced inflammatory responses following endotoxin administration. Counter-regulatory responses are initiated in parallel and may serve to limit some of the end-organ responses by the inflammatory mediators. This human model provides a unique opportunity to extend previous concepts of acute inflammation and to evaluate the earliest responses activated after exposure to an important bacterial component. Defining the pathways and responses initiated during acute human endotoxemia may allow a better understanding of host responses that are critical to the development of organ dysfunction and shock due to severe infections.

Sequential changes in the splanchnic circulation during continuous endotoxin infusion in sedated sheep: evidence for a selective increase of hepatic artery blood flow and loss of the hepatic arterial buffer response

On-line recording of the sequential changes in systemic, pulmonary, mesenteric, hepatic and renal circulations during onset of endotoxaemia and at 24 h of established hyperdynamic sepsis were evaluated in seven chronically instrumented and sedated sheep receiving a continuous intravenous infusion of Escherichia coli endotoxin (20 ng min-1 kg-1). A transient and significant (P < 0.05) pulmonary arterial vaso-constriction was noted after 13 +/- 4 min, and was followed immediately by a simultaneous significant decrease of coeliac trunk, superior mesenteric artery, and portal vein blood flow to below 50% of baseline values. The superior mesenteric artery and portal vein blood flows partially recovered pre-endotoxin levels to 69 and 75% of baseline, respectively, after 70 min of endotoxin infusion. In contrast, the coeliac trunk blood flow remained reduced for a more prolonged period of time, but then completely recovered baseline values at 100 min. The response of the hepatic artery was biphasic, and consisted of a transient (5-10 min) vasoconstriction at 40 min followed by transitory increase of hepatic artery blood flow reaching a maximum of 921% of baseline values at 102 min. Contrasting with the early changes observed in mesenteric vascular resistances mostly unrelated to systemic haemodynamics, the response of the renal vasculature appeared to be more dependent on changes of renal perfusion pressure. A follow-up at 24 h revealed that the continuous intravenous infusion of endotoxin reproduced some of the most characteristic features of human sepsis with increased cardiac output and decreased vascular resistances of all vascular beds. We conclude that hepatic artery blood flow is selectively and considerably increased in early endotoxaemia in sheep independently of changes in portal vein blood flow, suggesting a disregulation of the physiologic hepatic arterial buffer response, most probably secondary to an increased liver oxygen demand required for phagocytosis, transport, and digestion of the the sudden overload of bacterial endotoxins.

Tumor necrosis factor stimulates amino acid transport in plasma membrane vesicles from rat liver

Severe infection is characterized by a translocation of amino acids from the periphery to the liver, an event that is mediated in part by cytokines such as tumor necrosis factor-alpha (TNF). We investigated the activities of Na(+)-dependent transport systems A, ASC, and N in hepatic plasma membrane vesicles (HPMVs) prepared from rats treated with TNF in vivo. TNF did not alter sodium uptake but resulted in time- and dose-dependent fivefold and 50% maximal increases in system A and system N activity, respectively, in HPMVs secondary to an increase in the transport Vmax. Maximal increases in transport were observed 4 h after exposure to TNF and had returned to basal levels within 24 h. Similarly, system ASC activity was stimulated 80% in HPMVs from rats treated with TNF. Incubation of HPMVs from normal rats in vitro with TNF did not alter transport activity. Pretreatment of animals with the glucocorticoid receptor antagonist RU 38486 attenuated the TNF-induced enhancement in transport activity by 50%. The marked increase in Na(+)-dependent amino acid transport activity by TNF is mediated in part by the glucocorticoid hormones and represents an important mechanism underlying the accelerated hepatic amino acid uptake that occurs during critical illness.

Increased intestinal protein synthesis during sepsis and following the administration of tumour necrosis factor alpha or interleukin-1 alpha

The influence of sepsis on intestinal protein synthesis was studied in rats. Sepsis was induced by caecal ligation and puncture (CLP); control rats were sham-operated. Protein synthesis was measured in vivo in the jejunum and ileum following a flooding dose of [14C]leucine. At 8 h after CLP the protein synthesis rate was increased by approx. 15% in jejunal mucosa, and at 16 h after CLP, the protein synthesis rate was increased by 50-60% in the mucosa and seromuscular layer of both jejunum and ileum. In a second series of experiments, rats were treated with recombinant tumour necrosis factor alpha (rTNF alpha) or recombinant interleukin-1 alpha (rIL-1 alpha) administered at a total dose of 300 micrograms/kg body weight over 16 h. Control rats received corresponding volumes of solvent. Treatment with rTNF alpha resulted in an approx. 25% increase in mucosal protein synthesis in jejunum. Following treatment with rIL-1 alpha, protein synthesis increased by 25% in jejunal mucosa and almost doubled in ileal mucosa. The results suggest that sepsis stimulates intestinal protein synthesis and that this response may, at least in part, be mediated by TNF and/or IL-1.

In vivo effects of the antiglucocorticoid RU 486 on glucocorticoid and cytokine responses to Escherichia coli endotoxin

The endogenous adrenocortical response to sepsis is critical for host survival. The in vivo interactions among the endogenous glucocorticoid response, the induction of cytokines, and host survival during endotoxemia were explored in this study by use of the glucocorticoid receptor antagonist RU 486. Male Lewis rats underwent sterile insertion of a right jugular venous catheter. After a 72-h recovery period, animals received a 50% lethal dose of Escherichia coli endotoxin (2.5 mg/kg) via the catheter after pretreatment for 30 min prior to lipopolysaccharide (LPS) treatment with (i) vehicle alone intravenously (i.v.) (-corticosterone [-Cort]/-RU 486/+LPS) (n = 10), (ii) the antiglucocorticoid RU 486 (10 mg/kg) i.v. (-Cort/+RU 486/+LPS) (n = 11), or (iii) RU 486 (10 mg/kg) i.v. in animals that had undergone subcutaneous implantation of a corticosterone pellet at the time of catheter insertion (+Cort/+RU 486/+LPS) (n = 10). Except in animals receiving corticosterone pretreatment, baseline plasma corticosterone levels were low in all groups. Plasma corticosterone levels increased significantly (P less than 0.001) above the baseline following LPS administration. Animals in the -Cort/+RU 486/+LPS-treated group exhibited significantly increased mortality (P less than 0.001), with only 9% of the animals surviving at 72 h, as well as significantly increased plasma interleukin-6 levels, compared with animals receiving the vehicle alone (-Cort/-RU 486/+LPS), which showed 50% mortality. Pretreatment with corticosterone and RU 486 (+Cort/+RU 486/+LPS) significantly (P less than 0.001) reversed the mortality observed with RU 486 pretreatment alone (-Cort/+RU 486/+LPS), with 70% of the animals surviving at 72 h, and significantly attenuated the peak plasma tumor necrosis factor and interleukin-6 responses to LPS, compared with those in the animals treated with vehicle alone. These data demonstrate that the blockade of glucocorticoid binding by RU 486 increases LPS-induced mortality. The reversal of this effect by the induction of hypercorticosteronemia prior to RU 486 and LPS exposure (+Cort/+RU 486/+LPS) improves survival and is further associated with significant attenuation of cytokine production. Therefore, these data suggest that the protective effect of the endogenous glucocorticoid response to acute endotoxemia may result from the down-regulation of a potentially lethal cytokine response.

Tumor necrosis factor soluble receptors circulate during experimental and clinical inflammation and can protect against excessive tumor necrosis factor alpha in vitro and in vivo

Tumor necrosis factor alpha (TNF alpha), a primary mediator of systemic responses to sepsis and infection, can be injurious to the organism when present in excessive quantities. Here we report that two types of naturally occurring soluble TNF receptors (sTNFR-I and sTNFR-II) circulate in human experimental endotoxemia and in critically ill patients and demonstrate that they neutralize TNF alpha-induced cytotoxicity and immunoreactivity in vitro. Utilizing immunoassays that discriminate between total sTNFR-I and sTNFR-I not bound to TNF alpha, we show that sTNFR-I-TNF alpha complexes may circulate even in the absence of detectable free TNF alpha. To investigate the therapeutic possibilities of sTNFR-I, recombinant protein was administered to nonhuman primates with lethal bacteremia and found to attenuate hemodynamic collapse and cytokine induction. We conclude that soluble receptors for TNF alpha are inducible in inflammation and circulate at levels sufficient to block the in vitro cytotoxicity associated with TNF alpha levels observed in nonlethal infection. Administration of sTNFR-I can prevent the adverse pathologic sequelae caused by the exaggerated TNF alpha production observed in lethal sepsis.

Interleukin-1 receptor blockade improves survival and hemodynamic performance in Escherichia coli septic shock, but fails to alter host responses to sublethal endotoxemia

The present study was undertaken to evaluate the extent to which an endogenous interleukin-1 (IL-1) response contributes to the hemodynamic and metabolic consequences of sublethal endotoxemia or lethal Gram-negative septic shock. Young, healthy baboons received either a sublethal dose of lipopolysaccharide (LPS) or an LD100 of live Escherichia coli bacteria, and one half of the animals in each group were continuously infused with IL-1 receptor antagonist (IL-1ra). Plasma IL-1 beta was not detected in this model of endotoxemia. Administration of IL-1ra had only minimal effects on the modest hemodynamic and metabolic responses to sublethal endotoxemia, and did not attenuate the plasma cytokine response. In contrast, high circulating levels of IL-1 beta (range 300-800 pg/ml) were seen during lethal E. coli septic shock. IL-1ra treatment significantly attenuated the decrease in mean arterial blood pressure (MAP) (from -72 +/- 8 to -43 +/- 6 mm Hg; P less than 0.05) and cardiac output (from -0.81 +/- 0.17 to -0.48 +/- 0.15 liter/min; P less than 0.05), and significantly improved survival from 43 to 100% at 24 h (P less than 0.05). The plasma IL-1 beta and IL-6 responses to lethal E. coli septic shock were also significantly diminished by IL-1ra treatment (P less than 0.05), whereas tumor necrosis factor-alpha (TNF alpha) concentrations were unaffected. We conclude that an exaggerated systemic IL-1 beta response is characteristic of lethal E. coli septic shock, and contributes significantly to the hemodynamic and metabolic consequences of E. coli septic shock. IL-1ra can significantly attenuate the cytokine cascade and improve survival.

Tumour necrosis factor alpha secretion in leporine endotoxaemia: role of the liver and effects of hepatic ischaemia

A leporine model to investigate tumour necrosis factor alpha (TNF alpha) secretion after peripheral vein or mesenteric vein lipopolysaccharide injection was devised. Mesenteric vein injection provoked lower arterial concentrations after 90 minutes (median (range), 2.81, (0.75-11.96) ng/ml) than peripheral vein injection (7.00 (4.27-14.95) ng/ml (p less than 0.05)). Mesenteric vein injection after 10 minutes' warm hepatic ischaemia, which impairs hepatic clearance, provoked higher median arterial TNF alpha values at 90 minutes (7.98 (2.85-21.48) ng/ml) than in normal animals (p less than 0.05). Portal vein endotoxaemia induced less TNF alpha production than systemic endotoxaemia unless hepatic clearance was impaired, thus the major source of TNF alpha in systemic endotoxaemia is probably extrahepatic.

Role for monokines in the metabolic effects of endotoxin. Interferon-gamma restores responsiveness of C3H/HeJ mice in vivo

To examine the role of cytokines in mediating the lipogenic effects of endotoxin (LPS), we studied the effects of LPS and cytokines on hepatic fatty acid synthesis in LPS-sensitive C3H/OuJ mice and in LPS-resistant C3H/HeJ mice, whose macrophages are defective in the ability to produce tumor necrosis factor (TNF) and IL-1 in response to LPS. HeJ mice were 16-fold less sensitive than OuJ mice to the lipogenic effect of LPS. In OuJ mice, 10 micrograms of LPS caused a maximal increase in hepatic lipogenesis (3.86 +/- 0.41-fold), whereas in HeJ mice the maximal increase was only 1.79 +/- 0.32-fold after 100 micrograms of LPS. This lipogenic response paralleled the decreased ability of LPS to increase hepatic and splenic levels of mRNAs for TNF and IL-1 and serum levels of TNF in HeJ mice. In contrast, the maximal effect of TNF on lipogenesis was greater and the sensitivity to TNF was increased 2.4-fold in HeJ mice compared to OuJ mice. Administration of IFN-gamma before LPS in HeJ mice had no effect on IL-1 mRNA, but partially restored the LPS-induced increase in hepatic and splenic mRNA for TNF and serum TNF levels, which may account for the partial restoration of sensitivity to the lipogenic effect of LPS after IFN-gamma treatment. These results indicate that cytokines produced by mononuclear leukocytes mediate the lipogenic effects of LPS.

Administration of anti-TNF antibody improves left ventricular function in septic shock patients. Results of a pilot study

In this pilot study, murine monoclonal anti-TNF antibody (2 mg/kg) was administered to ten patients within 24 h of septic shock which persisted after initial resuscitation with intravenous fluids and adrenergic agents. This treatment resulted in a reduction in heart rate (from 122 +/- 10 to 113 +/- 10 beats per minute at 4 h, p less than 0.01) associated with an increase in LVSWI (from 26.5 +/- 5.6 to 31.5 +/- 10.5 g.m2 at 2 h, p less than 0.05), indicating in the absence of change in cardiac filling pressures, an improvement in ventricular function. Arterial oxygenation improved concurrently in six patients. These changes, however, appeared transient. The improvement in cardiac function following anti-TNF antibody administration in patients is in keeping with recent experimental studies indicating the role of TNF in the myocardial depression characterizing septic shock.

Tumour location influences local cytokine production and host metabolism

To determine whether the location of tumour growth influences host cytokine and metabolic responses, experimental subcutaneous (SQ) and liver (LIV) tumours were compared in Buffalo rats. An LIV tumour that was only 1 +/- 1% (P < 0.05 versus SQ) of body weight produced similar anorexia, weight loss, acute phase response, and systemic cytokine responses as are SQ tumour that was 10 +/- 2% of body weight. Neither tumour-bearing group had abnormal liver function tests or evidence of obstructive biliary pathology. Tumour necrosis factor (TNF) was detected by western analysis in both tumour as well as histologically normal liver remote from the tumour in the LIV group but not in livers of animals in freely fed and SQ groups. The proximity of the tumour to competent tissue macrophage populations, such as hepatic Kupffer cells, may be sufficient to induce cachexia. Hence, tumour location may be as important as tumour burden in determining the host's response to cancer.

Tumor necrosis factor mimics the metabolic response to acute infection in healthy humans

To evaluate the role of tumor necrosis factor (TNF) in the initiation of the metabolic response to acute infection, we performed a crossover saline-controlled study in six healthy postabsorptive men, investigating the metabolic effects of a bolus intravenous injection of recombinant human TNF (50 micrograms/m2). TNF induced a transient stress hormone response, associated with an early and sustained rise in plasma glucose concentrations (percentage increase at 2 h 23 +/- 7; P less than 0.05). Glucose turnover, measured 7.5 h postinjection, was 10 +/- 3% higher after TNF administration (P less than 0.05). Plasma free fatty acids (FFA) and glycerol concentrations increased transiently after TNF injection, peaking after 4 h (percentage increase 363 +/- 83 and 67 +/- 14, respectively; both P less than 0.05). FFA turnover, determined 6.5 h postinjection, increased in five subjects to a variable extent (percentage increase 126 +/- 55; P less than 0.05). Finally, resting energy expenditure showed a transient rise after TNF injection (34 +/- 2% at 4 h; P less than 0.05). We conclude that intravenous TNF reproduces many of the metabolic changes observed in septicemia, suggesting that TNF may be an initiating factor in the development of the metabolic response to acute infection.

Elevated plasma interleukin-6 and increased severity and mortality in alcoholic hepatitis

Recent studies in alcoholic hepatitis have proposed a role for the cytokine tumour necrosis factor-alpha (TNF-alpha) a mediator of endotoxic shock in sepsis. In this study plasma levels of the closely related cytokine interleukin-6 (IL-6) were assayed in 96 samples from 58 patients with severe alcoholic hepatitis, and 69 patients in control groups (21 normal, 10 alcoholic without liver disease, 10 inactive alcoholic cirrhosis, 18 chronic liver disease, 10 chronic renal failure). Plasma IL-6 levels were markedly elevated in patients with alcoholic hepatitis when compared with all control groups (P less than 0.001). IL-6 levels were higher in patients who died (P = 0.04) and correlated with the features of severe disease including: increased grade of encephalopathy, increased neutrophil count, increased prothrombin ratio, hypotension, increased serum creatinine and increased serum bilirubin. Surprisingly, no correlation was found between levels of plasma IL-6 and plasma TNF-alpha or endotoxin, or the presence of infection; an inverse correlation was found between plasma IL-6 and serum globulins. These findings provide further evidence that the IL-6/TNF cytokine system is activated in severe alcoholic hepatitis and may mediate hepatic or extra-hepatic tissue damage.

Hyperdynamic circulation in cirrhosis: a role for nitric oxide?

Hypotension, low systemic vascular resistance, and a reduced sensitivity to vasoconstrictors are features of cirrhosis. These cardiovascular changes might be the result of increased synthesis of a vasodilator. Nitric oxide (NO), a potent vasodilator, is synthesised in and released from peripheral blood-vessels in man. Studies in animals indicate that bacterial endotoxin and cytokines induce NO synthase expression in vessel walls, with sustained NO release and consequent hypotension. Endotoxaemia is a common feature of cirrhosis; persistent induction of NO synthase may account for the associated haemodynamic changes.