Human endotoxin administration as an experimental model in drug development.

Linking human physiology to inflammatory mechanisms discovered in vitro or in animal models is essential to determine their importance. Innate immunity underlies many of these inflammatory responses in health and disease. Bacterial endotoxin is the quintessential trigger of innate immune responses. When administered to humans, endotoxin has been an important means of demonstrating key inflammatory mechanisms in vivo. Furthermore, endotoxin challenges have provided opportunities to test the effects of novel inflammation-modifying agents in humans.

Human models of innate immunity: local and systemic inflammatory responses.

We review human studies where different body sites (e.g. systemic--intravenous and local--skin or lung) are exposed to small amounts of bacterial components as a means to study innate immunity in vivo. Intravenous endotoxin administration is widely used to assess systemic inflammatory responses, and these have many similarities to those seen in early sepsis. While blood levels of cytokines, activated inflammatory cells, and stress hormones rise acutely, the alveolar space remains relatively protected from these inflammatory responses. Skin blister windows provide a means to study local neutrophil exudation without systemic inflammatory responses, and has been used to characterize defects in neutrophil transmigration. Recently, skin blister windows have been adapted to study phagocytic cell function in response to bacterial antigens in patients with cirrhosis. Inhalation of endotoxin leads to pulmonary inflammation with increases in broncho-alveolar lavage neutrophils and cytokines and mild systemic responses. Whole lung exposure to endotoxin provides a means to study the pathogenesis of occupational lung disease. These three models are important methods to study innate immune responses and their regulatory mechanisms in normal and diseased states.

Effect of endotoxin on ventilation and breath variability: role of cyclooxygenase pathway.

To evaluate the effects of endotoxemia on respiratory controller function, 12 subjects were randomized to receive endotoxin or saline; six also received ibuprofen, a cyclooxygenase inhibitor, and six received placebo. Administration of endotoxin produced fever, increased respiratory frequency, decreased inspiratory time, and widened alveolar-arterial oxygen tension gradient (all p < or = 0.001); these responses were blocked by ibuprofen. Independent of ibuprofen, endotoxin produced dyspnea, and it increased fractional inspiratory time, minute ventilation, and mean inspiratory flow (all p < or = 0.025). Endotoxin altered the autocorrelative behavior of respiratory frequency by increasing its autocorrelation coefficient at a lag of one breath, the number of breath lags with significant serial correlations, and its correlated fraction (all p < 0.05); these responses were blocked by ibuprofen. Changes in correlated behavior of respiratory frequency were related to changes in arterial carbon dioxide tension (r = 0.86; p < 0.03). Endotoxin decreased the oscillatory fraction of inspiratory time in both the placebo (p < 0.05) and ibuprofen groups (p = 0.06). In conclusion, endotoxin produced increases in respiratory motor output and dyspnea independent of fever and symptoms, and it curtailed the freedom to vary respiratory timing-a response that appears to be mediated by the cyclooxygenase pathway.

Effects of anti-inflammatory agents on serum levels of calcitonin precursors during human experimental endotoxemia.

Calcitonin precursor (CTpr) levels are both markers and mediators of inflammation. The duration of their elevation after intravenous endotoxin challenge and the effects of anti-inflammatory therapies were studied in 52 subjects. CTpr levels maximized at 24 h in all subjects. At 7 days (n=4), after levels of acute-phase cytokines and C-reactive protein had normalized, CTpr levels remained 2-4-fold above baseline levels. The elimination half-life of CTpr levels ranged from 26.9 to 45.7 h. At 24 h, endotoxin and ibuprofen (compared with endotoxin alone) increased CTpr levels approximately 2-fold (P=.03), whereas soluble tumor necrosis factor receptor blunted the increase in CTpr levels by 2-3-fold (P=.0015). However, soluble interleukin-1 receptor failed to alter the increase in CTpr levels. Thus, the fact that anti-inflammatory agents may alter CTpr levels resulting from a single stimulus must be considered when CTpr is used as a clinical marker. Of importance, this study reveals that anti-inflammatory agents may modulate the CTpr level, which is a potential toxic mediator of inflammation.

Local inflammatory responses following bronchial endotoxin instillation in humans.

To study local lung inflammation, 34 subjects had endotoxin (1-4 ng/kg) instilled into a lung segment and saline instilled into a contralateral segment followed by bronchoalveolar lavage (BAL) at 2 h, 6 h, 24 h, or 48 h. Endotoxin instillation resulted in a focal inflammatory response with a distinct time course. An early phase (2 h to 6 h) revealed an increase in neutrophils (p = 0.0001) with elevated cytokines (tumor necrosis factor [TNF]-alpha, TNF receptors [TNFR], interleukin [IL]-1beta, IL-1 receptor antagonist, IL-6, granulocyte-colony-stimulating factor [G-CSF], all p < or = 0.002, but no change in IL-10) and chemokines (IL-8, epithelial neutrophil activating protein-78, monocyte chemotactic protein-1, macrophage inflammatory protein [MIP]-1alpha, MIP-1beta, all p < or = 0.001, but no change in growth-regulated peptide-alpha). A later phase (24 h to 48 h) showed increased neutrophils, macrophages, monocytes, and lymphocytes (all p < or = 0.02), and a return to basal levels of most mediators. Elevated levels of inflammatory markers (TNFR(1), TNFR(2), L-selectin, lactoferrin, and myeloperoxidase) persisted in the BAL at 48 h (p < or = 0.001). Increased permeability to albumin occurred throughout both phases (p = 0.001). Blood C-reactive protein, serum amyloid A, IL-6, IL-1ra, G-CSF, but not TNF-alpha increased by 8 h (all p < or = 0.008). The local pulmonary inflammatory response to endotoxin has a unique qualitative and temporal profile of inflammation compared with previous reports of intravenous endotoxin challenges. This model provides a means to investigate factors that initiate, amplify, and resolve local lung inflammation.

Plasma dehydroepiandrosterone levels during experimental endotoxemia and anti-inflammatory therapy in humans.

OBJECTIVE: To measure the effect of experimental endotoxemia and anti-inflammatory therapy on plasma dehydroepiandrosterone (DHEA) levels in humans. DESIGN: Controlled, randomized, single-blind, prospective clinical study. SETTING: Monitored unit in research hospital. SUBJECTS: Twelve healthy volunteers served as their own controls and were randomized to receive intravenous endotoxin (Escherichia coli) or saline separated by 1 wk. Six were randomized to receive ibuprofen, a cyclooxygenase inhibitor, and six were given placebo. INTERVENTIONS: Measurement of vital signs and hormones during a 24-hr period. MEASUREMENTS AND MAIN RESULTS: All subjects given endotoxin had a significant increase in plasma DHEA, cortisol, and adrenocorticotropic hormone (ACTH) levels (all p = .02). DHEA levels were maximum at 2 hrs and returned to baseline values by 6 hrs. Ibuprofen administration significantly blunted the endotoxin-induced increase in DHEA secretion (p = .001), whereas the increase in cortisol and ACTH was not affected. CONCLUSIONS: Acute endotoxemia leads to a rise in plasma DHEA levels in humans. Maximum levels of DHEA but not cortisol or ACTH were blunted by ibuprofen, suggesting a different regulation of these synthetic pathways in the adrenal cortex inner zone during acute inflammation.

High-dose corticosteroid therapy for diffuse alveolar hemorrhage in allogeneic bone marrow stem cell transplant recipients.

In a series of 74 patients with hematological malignancies undergoing allogeneic bone marrow or peri- pheral blood stem cell transplants from an HLA-identical sibling donor, four developed diffuse alveolar hemorrhage (DAH) between days 0 and 23 post transplant. Diagnosis was made by the radiographic finding of diffuse bilateral lung opacities, and bloody lavage fluid on bronchoscopy. Two patients required mechanical ventilatory support. They were treated with methylprednisolone 0.25-1.5 g/day for at least 4 days with slow tapering thereafter. All patients showed an immediate response and two became long-term survivors with normal respiratory function. Two had a relapse of DAH, developed acute respiratory distress syndrome (ARDS) and died with multi-organ failure. Risk factors for DAH were one or more courses of intensive chemotherapy pretransplant vs no treatment or low-dose chemotherapy (4/4 DAH vs 23/70 no DAH; P = 0.015), and second transplants (2/2 DAH vs 1/70 with no DAH; P = 0.006). These results indicate that DAH is life-threatening but is potentially reversible by prompt treatment with high doses of steroids.

New insights into the biology of the acute phase response.

Innate or natural immunity is a highly conserved defense mechanism against infection found in all multicellular organisms. The acute phase response is the set of immediate inflammatory responses initiated by pattern recognition molecules. These germ cell-encoded proteins recognize microbial pathogens based on shared molecular structures and induce host responses that localize the spread of infection and enhance systemic resistance to infection. Innate immunity also influences the initiation and type of adaptive immune response by regulating T cell costimulatory activity and antigen presentation by antigen presenting cells and by influencing mediator production, which affects lymphocyte function and trafficking. Acute phase protein concentrations rapidly increase after infection, and their production is controlled primarily by IL-6- and IL-1-type cytokines. The acute phase proteins provide enhanced protection against microorganisms and modify inflammatory responses by effects on cell trafficking and mediator release. For example, serum amyloid A has potent leukocyte activating functions including induction of chemotaxis, enhancement of leukocyte adhesion to endothelial cells, and increased phagocytosis. The constellation of inflammatory responses seen after endotoxin administration to humans represents an in vivo model of the acute phase response. Studies with inflammatory modifying agents, such as soluble dimeric TNF receptor and IL-10, show that these responses are not dependent on a single mediator but result from multiple overlapping inflammatory pathways. Understanding the factors that initiate and alter the magnitude and duration of the acute phase response represents an important step in the development of new therapies for infectious and inflammatory diseases.

Down-regulation of nitric oxide production by ibuprofen in human volunteers.

Ibuprofen has been shown in vitro to modulate production of nitric oxide (NO), a mediator of sepsis-induced hypotension. We sought to determine whether ibuprofen alters NO production and, thereby, vascular tone, in normal and endotoxin-challenged volunteers. Techniques for detecting NO were validated in 17 subjects infused with sodium nitroprusside, a NO donor. Then, endotoxin (4 ng/kg) or saline (vehicle alone) was administered in a single-blinded, crossover design to 12 other subjects randomized to receive either ibuprofen (2400 mg p.o.) or a placebo. Endotoxin decreased mean arterial pressure (MAP; P =.002) and increased alveolar NO flow rates (P =.04) and urinary excretion of nitrite and nitrate (P =.07). In both endotoxemic and normal subjects, ibuprofen blunted the small fall in MAP associated with bed rest (P =.005) and decreased alveolar NO flow rates (P =.03) and urinary excretion of nitrite and nitrate (P =.02). However, ibuprofen had no effect on the decrease in MAP caused by endotoxin, although it blocked NO production to the point of disrupting the normal relationship between increases in exhaled NO flow rate and decreases in MAP (P =.002). These are the first in vivo data to demonstrate that ibuprofen down-regulates NO in humans. Ibuprofen impaired the NO response to bed rest, producing a small rise in blood pressure. Although ibuprofen also interfered with the ability of endotoxin to induce NO production, it had no effect on the fall in blood pressure, suggesting that the hemodynamic response to endotoxin is not completely dependent on NO under these conditions.

Dose-related inflammatory effects of intravenous endotoxin in humans: evaluation of a new clinical lot of Escherichia coli O:113 endotoxin.

The administration of reference endotoxin (Escherichia coli O:113, Lot EC-5) to humans has been an important means to study inflammation in vivo; however, the supply of Lot EC-5 is depleted. A new lot of reference endotoxin (Clinical Center reference endotoxin [CCRE]), derived from the original bulk material extracted from E. coli O:113, was processed. The effects of 0-, 1-, 2-, and 4-ng/kg doses of intravenous CCRE and EC-5 were studied in 20 male subjects. CCRE resulted in dose-related increases in symptoms, temperature (P=. 016), total leukocyte count (P=.014), tumor necrosis factor-alpha (P=.004), interleukin (IL)-1 receptor antagonist (P=.004), IL-6 (P=. 005), IL-8 (P=.011), cortisol (P<.05), and C-reactive protein (P=. 04). These responses were attenuated (all P<.012) in subjects given Lot EC-5 (4 ng/kg) in comparison with those in subjects given CCRE, showing that, over several years, EC-5 had lost potency. Thus, in healthy subjects, the magnitude of exposure to CCRE results in a graded dose response of major components of innate immunity.

Human neutrophils secrete gelatinase B in vitro and in vivo in response to endotoxin and proinflammatory mediators.

Bacterial sepsis is characterized by a systemic inflammatory state, with activation of numerous cell types. Phagocytes participate in this phenomenon by secreting various proinflammatory cytokines and enzymes. Matrix metalloproteinases (MMPs) such as gelatinases are produced by phagocytes and are thought to play an important role in processes of cell transmigration and tissue remodeling. In this work, we show that endotoxin (lipopolysaccharide [LPS]) and other inflammatory mediators, such as tumor necrosis factor (TNF), interleukin-8, and granulocyte colony-stimulating factor, induce a rapid (within 20 min) release of gelatinase-B (MMP-9) zymogen in whole human blood, as determined by gelatin zymography. The polymorphonuclear neutrophil was identified as the cell responsible for this rapid secretion, as a result of the release of preformed enzymes stored in granules. Normal human subjects given LPS intravenously showed a similar pattern of proMMP-9 secretion, with maximum plasma levels reached 1.5 to 3 h after LPS administration (P = 0.0009). Prior administration of TNF receptor:Fc, a potent TNF antagonist, to subjects given LPS, only partially blunted the release of proMMP-9 (P = 0.033). Ibuprofen, a cyclooxygenase inhibitor, did not alter this pattern of release. Increased levels of proMMP-9 and proMMP-2, as well as activated forms of MMP-9, were found in plasma from two patients with gram-negative sepsis. The levels of MMPs paralleled the severity of clinical condition and a marker of the severity of sepsis, plasma procalcitonin. These data indicate that MMPs are released in whole blood in response to various inflammatory mediators and that they could serve as sensitive and early markers for cell activation during the course of bacterial sepsis.

Detection of macrophage inflammatory protein (MIP)-1alpha and MIP-1beta during experimental endotoxemia and human sepsis.

Macrophage inflammatory protein (MIP)-1alpha and MIP-1beta regulate leukocyte activation and trafficking. To assess the role of MIP-1alpha and MIP-1beta in human inflammation, healthy subjects were studied during experimental endotoxemia with prior administration of ibuprofen, a cyclooxygenase inhibitor, or dimeric p75 tumor necrosis factor (TNF)-alpha receptor, a TNF antagonist; septic patients were also studied. Following endotoxin, blood levels of both MIP-1 molecules rose acutely and fell to baseline by 6 h (P=. 001). While MIP-1 mediates fever in animals independent of cyclooxygenase blockade, in subjects given endotoxin and ibuprofen, MIP-1 levels increased and fever was suppressed. MIP-1 levels were not diminished by inhibiting circulating TNF-alpha in humans. In septic patients, elevated levels of MIP-1alpha and MIP-1beta were detected within 24 h of sepsis and fell in parallel with TNF-alpha and interleukin-6 (P<.01). MIP-1alpha and MIP-1beta increase during acute inflammation but are not associated with fever in endotoxemic humans during cyclooxygenase blockade.

Myocardial dysfunction in sepsis: clinical and experimental investigations.

OBJECTIVE: To review the clinical manifestations and mechanisms of cardiac dysfunction in septic shock. METHODS: Literature review of selected clinical studies and animal models. RESULTS: Depressed myocardial contractile function is a common consequence of severe infections. Bacterial factors, in conjunction with host inflammatory mediators, produce a profile of reversible cardiac dysfunction manifested by a decrease in ventricular ejection fraction, ventricular dilatation, and increased cardiac output. Global ischemia is not the major mechanism that mediates cardiac dysfunction during sepsis. Inflammatory mediators contribute to myocardial dysfunction by damaging the coronary microcirculation and contributing to myocardial edema and cardiocyte damage. However, trials of anti-inflammatory agents have not prevented or increased the rate of reversal of septic shock or improved survival. The link between nitric oxide and clinical myocardial depression remains unclear, as nonselective nitric oxide synthase inhibition does not block the development of ventricular dysfunction. CONCLUSIONS: Serious bacterial infections result in inflammatory injury to the heart manifested by a common profile of cardiac dysfunction. Therapy remains limited to treatment of the infection with antibiotics and supportive care, with fluid resuscitation and selective use of inotropes and vasopressors. Experimental models suggest that new anti-inflammatory strategies (e.g. tyrosine kinase inhibitors) may offer some advantages over those that target a single mediator; these agents remain to be clinically evaluated.

Circulating leptin levels during acute experimental endotoxemia and antiinflammatory therapy in humans.

Leptin, a newly discovered adipose tissue-derived weight-reducing hormone, is increased in acute inflammation and may be involved in the anorexia and wasting syndrome associated with infection. To determine whether this hormone responds to an acute inflammatory stimulus, plasma leptin concentrations were measured in 12 healthy subjects after intravenous administration of endotoxin. These subjects were randomized to receive concurrently ibuprofen or placebo normal saline (6 in each group). Endotoxin administration resulted in fever, leukocytosis, and an increase in plasma levels of the stress hormones adrenocorticotropic hormone (3.2 +/- 0.3 to 132.6 +/- 75.5 pmol/L, P = .001) and cortisol (431.6 +/- 44 to 796.9 +/- 99 mmol/L, P = .001). Plasma leptin levels, however, did not change significantly from baseline values after administration of endotoxin (0 h: 6.9 +/- 3.1 ng/mL; 6 h: 6.0 +/- 2.2; 24 h: 6.5 +/- 2.8). While ibuprofen suppressed fever and symptoms associated with endotoxemia, it had no effect on the plasma levels of leptin. In conclusion, acute experimental human endotoxinemia is not associated with acute changes in circulating leptin levels.

Recombinant tumor necrosis factor receptor p75 fusion protein (TNFR:Fc) alters endotoxin-induced activation of the kinin, fibrinolytic, and coagulation systems in normal humans.

The effects of inhibition of tumor necrosis factor (TNF) on cell and protease activation were evaluated in 18 normal volunteers given endotoxin (4 ng/kg, i.v.) after an infusion of low (10 mg/m2 i.v., n = 6) or high dose (60 mg/m2 i.v., n = 6) recombinant human dimeric TNF receptor protein (TNFR:Fc) or its vehicle (placebo n = 6). Activation of the coagulation system occurred by 2 h in the TNFR:Fc vehicle-placebo group manifested by decreased prekallikrein functional levels and increased levels of prothrombin F1+2 fragments (p < 0.0001). High or low dose TNFR:Fc delayed the fall in prekallikrein functional levels by 1 h and 4 h, respectively (p < 0.0002), but did not inhibit the increase in circulating levels of prothrombin F1+2 fragments. In contrast, endothelium activation, characterized by increased levels of tissue plasminogen activator, plasminogen activator inhibitor-1, and von Willebrand Factor antigen was blunted by both low and high dose TNFR:Fc (p < 0.001). While the endotoxin-associated decrease in platelet number was not altered, platelet-derived beta-thromboglobulin peak levels were blunted and delayed by TNFR:Fc (p < 0.02). Increased levels of neutrophil elastase were attenuated by low and high dose TNFR:Fc (p < 0.001). These results suggest that although TNF is functionally linked to the activation of endothelium, neutrophils, coagulation, and fibrinolysis, alternative pathways are present in vivo that result in activation of the kallikrein-kinin system after endotoxin-induced TNF release. These alternative pathways may limit some of the anti-inflammatory effects of TNFR:Fc.

Effects of recombinant soluble type I interleukin-1 receptor on human inflammatory responses to endotoxin.

Effects of soluble recombinant human type I interleukin-1 receptor (sIL-1RI) were evaluated in 18 volunteers given intravenous endotoxin and randomized to placebo (n = 6), low-dose (n = 6), or high-dose (n = 6) sIL-1RI. Soluble IL-1RI decreased IL-1 beta (P = .001), but decreased IL-1ra (P = .0001), and resulted in 10-fold and 43-fold dose-related increases in sIL-1RI-IL-1ra complexes compared with placebo (P < or = .001). High-dose sIL-1RI was associated with increased levels of immunoactive tumor necrosis factor-alpha (P = .02), IL-8 (P = .0001), and cell-associated IL-1 beta (P = .047). C-reactive protein levels were higher after sIL-1RI than placebo (P = .035). Soluble IL-1RI decreased the severity of chills (P = .03), but did not alter other symptoms, changes in temperature, systemic hemodynamic responses, or changes in leukocyte and platelet number. Thus, sIL-1RI had no discernable antiinflammatory effect following endotoxin administration due in part to low levels of circulating IL-1 beta and neutralization of IL-1ra inhibitory function. This latter interaction represents an indirect mechanism of agonist activity elicited by sIL-1RI and may contribute to increases in inflammatory mediators, limiting therapy with sIL-1RI during endotoxemia.

Experimental human endotoxemia increases cardiac regularity: results from a prospective, randomized, crossover trial.

OBJECTIVE: To determine whether human endotoxemia is associated with a loss of the physiologic beat-to-beat variability of heart rate. DESIGN: Prospective, randomized, crossover, single-blind study. SETTING: Clinical research center in a federal, nonuniversity hospital. SUBJECTS: Healthy volunteers. INTERVENTIONS: Intravenous administration of reference (Escherichia coli) endotoxin or saline placebo, with or without previous administration of oral ibuprofen. MEASUREMENTS AND MAIN RESULTS: Electrocardiograms were continuously recorded and digitized using series of 1000 beat epochs of R-R intervals over 8 hrs. Analyses included measures in the time domain (standard deviation), frequency domain (power spectra), and a measure of regularity (approximate entropy). Endotoxin administration was associated with loss of variability by all measures. This loss of variability remained significant even with administration of ibuprofen, which blocked the development of fever and endotoxin-related symptoms. CONCLUSIONS: Infusion of endotoxin into human volunteers causes loss of heart rate variability, as measured by standard deviation and power spectra, as well as an increase in heart rate regularity, as measured by approximate entropy. Changes in approximate entropy occurred earlier than changes in other heart rate variability measures and may be a useful means of detecting early sepsis. This reduction in regularity is consistent with a model in which the pathogenesis of multiple organ system dysfunction syndrome involves the physiologic uncoupling of vital organ systems.

Activation of the contact and fibrinolytic systems after intravenous administration of endotoxin to normal human volunteers: correlation with the cytokine profile.

Severe progressive failure of multiple organ systems has emerged during the past three decades as a common cause of death among patients in intensive care units. Sepsis is now better defined as systemic inflammatory response syndrome (SIRS), but its mortality rate has not changed and it continues to be a major health problem. Endotoxin interacts with plasma proteins and contributes to the pathophysiology of SIRS. Information is limited on the effect of endotoxin on human coagulation and fibrinolytic proteins in vivo, as well as on the cell response involved in the cytokine cascade. For this reason we performed quantitative assays to establish the sequence of events that occurs in vivo in the regulation of the contact and fibrinolytic pathways as well as in the cytokine cascade as a response to a single dose administration of endotoxin to normal non-smoking human volunteers.

Cofactors V and VIII after endotoxin administration to human volunteers.

Coagulation factor V (FV) and factor VIII (FVIII) are usually decreased in septicemic DIC. Low doses of endotoxin administered to healthy volunteers stimulate activation of the fibrinolytic, contact and coagulation systems, but not clinical DIC. Following the administration of endotoxin (4 ng/kg) to normal volunteers (n = 15), we applied new assays for FV antigens using monoclonal antibodies to the activation peptide (C1) and to the light chain of FV. At 5 hours, FV coagulant activity was significantly decreased (64 +/- 9%), as was the FV light chain antigen (74 +/- 6%), without a change in factor V C1 antigen or total protein C. In contrast, FVIII coagulant activity was greater than preinfusion levels at 2-5 hours. The decrease in FV activity may be due to APC cleavage of FV heavy chain, but the loss of light chain antigen suggests that plasmin and/or calpain also contribute. APC may not be the only enzyme responsible for cofactor inactivation. FV is one of the most sensitive markers, even reflecting subclinical activation of coagulation.

Effects of recombinant dimeric TNF receptor on human inflammatory responses following intravenous endotoxin administration.

Effects of dimeric TNF receptor (p80) Fc (TNFR:Fc) on acute phase responses were evaluated in 18 volunteers given endotoxin (4 ng/kg i.v.). Subjects were randomized to receive either placebo (n = 6), low dose TNFR:Fc (10 mg/m2 i.v., n = 6), or high dose TNFR:Fc (60 mg/m2 i.v., n = 6). TNFR:Fc blocked plasma TNF bioactivity (p = 0.001) and increased, in a dose-ordered fashion, TNF immunoactivity (p < 0.001). TNFR:Fc decreased secondary cytokine levels including IL-1 beta (p = 0.007), IL-8 (p < 0.001), IL-1 receptor antagonist (p < 0.001), granulocyte-CSF (p = 0.03), and growth regulated peptide-alpha (p = 0.001) but not macrophage inflammatory protein-1 alpha or IL-10. Low dose, but not high dose, TNFR:Fc blunted or delayed the release of epinephrine and cortisol (p < or = 0.026). Despite the absence of plasma TNF bioactivity, high dose TNFR:Fc was less immunosuppressive than low dose TNFR:Fc as measured by cytokine and stress hormone responses. Endotoxin-related symptoms were not altered by TNFR:Fc and the febrile response was delayed but not diminished (p = 0.004). Increases in cardiac index (72 +/- 19%) and heart rate (60 +/- 10%) and decreases in systemic vascular resistance index (47 +/- 7%) were unaltered by TNFR:Fc. These data suggest that the inflammatory response to endotoxin can escape from high levels of circulating TNF-blocking activity and redundant pathways, independent of circulating TNF, can sustain inflammation and clinical responses caused by acute endotoxemia.