Proteolytic inactivation of plasma C1- inhibitor in sepsis.

Activation of both the complement system and the contact system of intrinsic coagulation is implicated in the pathophysiology of sepsis. Because C1 inhibitor (C1-Inh) regulates the activation of both cascade systems, we studied the characteristics of plasma C1-Inh in 48 patients with severe sepsis on admission to the Intensive Care Unit at the Free University of Amsterdam. The ratio between the level of functional and antigenic C1-Inh (functional index) was significantly reduced in the patients with sepsis compared with healthy volunteers (P = 0.004). The assessment of modified (cleaved), inactive C1-Inh (iC1-Inh), and complexed forms of C1-Inh (nonfunctional C1-Inh species) revealed that the reduced functional index was mainly due to the presence of iC1-Inh. On SDS-PAGE, iC1-Inh species migrated with a lower apparent molecular weight (Mr 98,000, 91,000, and 86,000) than native C1-Inh (Mr 110,000). Elevated iC1-Inh levels (greater than or equal to 0.13 microM) were found in 81% of all patients, sometimes up to 1.6 microM. Levels of iC1-Inh on admission appeared to be of prognostic value: iC1-Inh was higher in 27 patients who died than in 21 patients who survived (P = 0.003). The mortality in 15 patients with iC1-Inh levels up to 0.2 microM was 27%, but in 12 patients with plasma iC1-Inh exceeding 0.44 microM, the mortality was 83%. The overall mortality in the patients with sepsis was 56%. We propose that the cleavage of C1-Inh in patients with sepsis reflects processes that play a major role in the development of fatal complications during sepsis.

Polyethylene glycol enhances the binding of C1q to circulating immune complexes.

By radioimmunoassay we measured the amount of endogenous C1q that was precipitated by polyethylene glycol (PEG) under the conditions of the 125I-C1q-binding test (C1q-BT). We found a linear correlation between the percentage endogenous C1q that was precipitated and the 125I-C1q-binding activity (C1q-BA). We concluded that the 125I-C1q behaves like the endogenous C1q. To detect circulating immune complexes (CIC) which had already bound C1q, human sera were added to tubes coated with anti-C1q. Under the conditions used, no C1q-bearing CIC were detected. In addition, 7 sera from patients with high C1q-BA were analyzed by sucrose-gradient ultracentrifugation. No C1q was found in the fast sedimenting fractions, although C1q-BA was detected in these fractions. With IgG-coated tubes we observed that PEG enhanced the binding of 125I-C1q as well as endogenous C1q to aggregated and monomeric IgG. PEG also enhanced the binding of CIC to C1q-coated tubes. The results suggest that CIC detected by the C1q-BT do not bear C1q in significant amounts in the circulation and that these CIC become detectable only in the presence of PEG.

Elevated plasma levels of the anaphylatoxins C3a and C4a are associated with a fatal outcome in sepsis.

PURPOSE AND PATIENTS AND METHODS: Both complement and contact system of coagulation have been implicated in the pathophysiology of sepsis. We therefore measured levels of the complement activation products C1-C1-inhibitor complexes and C3a in serial plasma samples (obtained every six hours) from 48 patients with clinically suspected sepsis, and related these levels to the clinical outcome. C4a was also measured in samples obtained on admission. RESULTS: C3a levels were elevated in 47 patients at least once during the observation period. These levels appeared to be considerably higher in patients who died than in patients who survived. This difference was found for the levels on admission (p = 0.0003), as well as for the highest (p = 0.0010) and the lowest (p less than 0.0001) levels encountered in each patient. The mortality in patients with plasma C3a levels of 13 nmol/liter or less on admission (27 patients) was 33 percent, compared with 86 percent in patients with levels of 14 nmol/liter or more. Patients with septic shock had significantly higher C3a levels than normotensive patients (p values between 0.046 and 0.004). No significant differences in C3a were found between patients who had respiratory distress syndrome and those who did not. C4a levels in plasma samples obtained on admission were elevated in 43 patients. These levels correlated very significantly with C3a levels (p less than 0.0001), and showed similar associations with a fatal outcome. C1-C1-inhibitor complexes were elevated in 23 patients at least once during the observation period. These patients had significantly higher levels of C4a and C3a than patients with normal amounts of C1-C1-inhibitor complexes. Patients who died had higher levels of C1-C1-inhibitor complexes than patients who survived. However, this difference was not significant. CONCLUSION: On the basis of our results, we propose that activation of the complement system via the classical pathway is involved in the development of fatal complications in sepsis.

Heparin-protamine complexes and C-reactive protein induce activation of the classical complement pathway: studies in patients undergoing cardiac surgery and in vitro.

The administration of protamine to patients undergoing cardiopulmonary bypass (CPB) to neutralize heparin and to reduce the risk of bleeding, induces activation of the classical complement pathway mainly by heparin-protamine complexes. We investigated whether C-reactive protein (CRP) contributes to protamine-induced complement activation. In 24 patients during myocardial revascularization, we measured complement, CRP, and complement-CRP complexes, reflecting CRP-mediated complement activation in vivo. We also incubated plasma from healthy volunteers and patients with heparin and protamine in vitro to study CRP-mediated complement activation. During CPB, CRP levels remained unchanged while C3 activation products increased. C4 activation occurred after protamine administration. CRP-complement complexes increased at the end of CPB and upon protamine administration. Incubation of plasma with heparin and protamine in vitro generated complement-CRP complexes, which was blocked by phosphorylcholine and stimulated by exogenous CRP. C4d-CRP complex formation after protamine administration correlated clinically with the incidence of postoperative arrhythmia. Protamine administration during cardiac surgery induces complement activation which in part is CRP-dependent, and correlates with postoperative arrhythmia.

Lack of activation of C1, despite circulating immune complexes detected by two C1q methods, in patients with rheumatoid arthritis.

The activation of C1 by circulating immune complexes in patients with rheumatoid arthritis was investigated. C1rC1s(C1-In)2 complexes in EDTA-plasma, reflecting C1 activation in vivo, were slightly raised in 35 of 57 patients with rheumatoid arthritis, though most patients had elevated levels of circulating immune complexes as measured with either the 125I-C1q binding test or the C1q solid phase assay. The activation of C1 by circulating immune complexes in vitro was investigated by measuring the generation of C1rC1s(C1-In)2 complexes during 60 minutes at 37 degrees C in diluted recalcified EDTA-plasma. In 16 of the 57 patients, a slightly increased C1 activation in vitro was observed. These patients tended to have high levels of circulating immune complexes. However, the majority of the patients with high levels of circulating immune complexes showed a normal C1 activation in vitro. Therefore, it was concluded that measurement of circulating immune complexes by either of the two C1q methods in patients with rheumatoid arthritis does not imply that these circulating immune complexes are able to activate C1.

A C1-inhibitor-complex assay (INCA): a method to detect C1 activation in vitro and in vivo.

A radioimmunoassay (the C1-inhibitor-complex assay, INCA) is described for the detection of complexes that are composed of at least C1s and C1-inhibitor. This INCA is based on demonstrating that C1s and C1-inhibitor (C1-In) are linked: after an incubation with anti-C1s-Sepharose, bound C1sC1-In complexes are detected by 125I-anti-C1-In. C1sC1-In complexes were prepared by the addition of a slight excess of C1s to normal human serum (NHS). As little as 2 ng C1-In bound to C1s was detected. Additional free C1s in serum hardly influenced the detection of C1sC1-In complexes. Complexes presumably composed of C1rC1s(C1-In)2 were generated by the addition of aggregated IgG to NHS. This generation was inhibited by lowering the temperature to 0 degrees C, and by EDTA, and depended on the concentration of aggregated IgG. These complexes had a sedimentation value of approximately 9S. Complexes of C1s and C1-In were also generated in NHS by the addition of DNP-albumin and protein A, but not by zymosan. The INCA was applied to blood samples from normal donors and patients. Sixteen out of 19 samples from patients with acute glomerulonephritis contained increased amounts of C1rC1s(C1-In)2 complexes as compared with the amounts in blood samples from normal donors. The INCA provides a useful tool to assess the activation of C1 in the presence of C1-In, both in vitro and in vivo.

SLE like syndrome and functional deficiency of C1q in members of a large family.

Two sisters and a brother from one family are described whose sera were deficient in haemolytic complement function. This defect was restored by addition of purified C1q. In their sera, C1q like material was found, whereas C1r and C1s were normal or increased in concentration, as were the other complement components tested. All three had suffered from glomerulonephritis during childhood. A renal biopsy in the brother recently disclosed a membranous glomerulopathy stage 1; otherwise, he is apparently healthy. In both sisters, a systemic lupus erythematosus like disease became manifest at the age of 20 and 23, respectively, resulting in the death of one of them. In the serum of these three family members, the C1q like material was antigenically deficient compared with normal C1q and had, on sucrose gradient analysis, a molecular weight of approximately 65,000 daltons. It did not bind to C1r and C1s. Binding of the dysfunctional C1q to aggregated human gammaglobulin could be demonstrated. On double immunodiffusion analysis, the abnormal C1q was identical with reduced and alkylated C1q. The possible structure of the abnormal C1q molecule is discussed.

Interleukin-2 induces activation of coagulation and fibrinolysis: resemblance to the changes seen during experimental endotoxaemia.

The administration of Interleukin-2 (IL-2) causes the release or generation of other cytokines such as tumour necrosis factor (TNF) which, by disturbing the anticoagulant properties of the endothelium, may induce a procoagulant state in patients receiving this drug. We therefore evaluated the effects of IL-2 on coagulation and fibrinolysis in 14 patients receiving 12 or 18 x 10(6) IU/m2/d of IL-2 given as a 15 min infusion for 5 d. Blood samples were drawn at short intervals after the first IL-2 infusion. The parameters were analysed by way of analysis for repeated measures (F tests rather than t tests). During the first day, thrombin-antithrombin (TAT) complexes started to increase 2 h after the IL-2 infusion, reaching peak levels at 4 h (n = 14; 11.2 +/- 6.4 micrograms/l v 49.8 +/- 49.2 micrograms/l, P < 0.01). Plasma alpha 2 antiplasmin (PAP) complexes showed a similar pattern rising from a mean baseline value of 17.5 +/- 7.6 nmol/l to 66.8 +/- 47.7 nmol at 4 h (P < 0.01). In four patients the peak of PAP preceeded that of TAT. Tissue plasminogen activator (tPA) rose from a mean baseline value of 4.9 +/- 3.7 micrograms/l to 26.3 +/- 13.5 micrograms/l at 4 h (P < 0.01). Plasminogen-activator-inhibitor-1 (PAI-1) levels increased from 59 +/- 35 micrograms/l to 113 +/- 39 micrograms/l at 6 h (P < 0.01). tPA PAI-1 complexes increased from 0.15 +/- 0.07 to 0.69 +/- 0.21 nmol/l at 6 h (P < 0.01). Our study indicates that IL-2 activates the coagulation and fibrinolytic systems in vivo. The changes resemble the perturbations observed after endotoxin/TNF administration. These abnormalities may play a role in the side-effects induced by IL-2 therapy.

The release of interleukin-8 during intravenous bolus treatment with interleukin-2.

OBJECTIVE: To study the role that interleukin-8 might play in the activation of polymorphonuclear neutrophils during interleukin-2 therapy and the relationship of these phenomena to interleukin-2 induced toxicity. DESIGN: A cohort study with measurements before and after the administration of interleukin-2. SETTING: Medical oncology department of a large teaching hospital. PATIENTS: Fourteen patients with metastatic renal cell carcinoma and 10 with metastatic melanoma being treated in a phase 2 study of the sequential combination of interferon-gamma and interleukin-2. MEASUREMENTS: Plasma levels of tumour necrosis factor-alpha, interleukins-6 and 8 and markers of neutrophil activation (neutrophil elastase and lactoferrin) were measured in patients receiving 5 daily injections of interferon-gamma (100 micrograms/m2/day) followed by 5 days of interleukin-2 (18 x 10(6) IU/m2/day). MAIN RESULTS: Tumour necrosis factor-alpha rose from baseline levels of 32 (range, 12 to 56) to 343 (103 to 787) pg/ml 3 hours after interleukin-2 administration returning to baseline values 21 hours later. Interleukins-6 and -8 rose from baseline levels of 6 (5 to 10) and 75 (35 to 100) to 2151 (152 to 7259) and 1283 (490 to 2500) pg/ml, respectively, at 4 hours after interleukin-2 with both returning to baseline values by 24 hours. Peak levels of neutrophil elastase and lactoferrin, both markers of neutrophil activation, occurred 6 hours after interleukin-2 administration. CONCLUSIONS: These data indicate that following administration of interleukin-2 tumour necrosis factor-alpha is released followed sequentially by rises in interleukins-6 and -8. It is hypothesised that these events result in activation of polymorphonuclear neutrophils. These activated neutrophils may play an important role in initiating endothelial cell damage leading to the haemodynamic toxicity and the capillary leak syndrome which is typically seen following the administration of interleukin-2.

Activation of the contact system of coagulation by a monoclonal antibody directed against a neodeterminant in the heavy chain region of human coagulation factor XII (Hageman factor).

We studied the characteristics of two monoclonal antibodies (mAbs), F1 and F3, against human coagulation factor XII (Hageman factor). Experiments with trypsin-digested 125I-factor XII revealed that the epitope for mAb F1 is located in the NH2-terminal Mr 40,100 portion of factor XII, whereas that for mAb F3 resides in the COOH-terminal Mr 30,000 portion of this protein. Factor XII in fresh plasma (single-chain factor XII) bound approximately 190 times less to mAb F1 than factor XII in dextran sulfate-activated plasma (cleaved factor XII). However, no difference in accessibility of the epitope for mAb F1 was observed between cleaved and single-chain factor XII when bound to glass. mAb F3 appeared to bind to both single-chain and cleaved factor XII in plasma as well as when bound to glass. Neither mAb F1, nor F3 affected the amidolytic activity of factor XIIa, whereas both mAb F1 and F3 inhibited factor XII-coagulant activity to about 15 and 70%, respectively, at a molar ratio of mAb to factor XII of 20 to 1. mAb F1, as well as F(ab')2 and F(ab') fragments of this antibody induced activation of the contact system in plasma, as reflected by the generation of factor XIIa. C1 inhibitor and kallikrein. C1 inhibitor complexes. Activation was induced neither upon incubation with mAb F3, nor with that of control mAbs. mAb F1-induced contact activation required the presence of factor XII, prekallikrein, and high molecular weight kininogen and, in contrast to activation by negatively charged surfaces, was not inhibited by the presence of Polybrene. Based on these results we propose that a conformational change in factor XII is a key event in the activation process of this molecule. This conformational change can be induced by binding of factor XII to a surface as well as by proteolytic cleavage. As mAb F1 can also induce this conformational change, this antibody may provide a unique tool in studies of the activation of factor XII.

Increased plasma levels of interleukin-6 in sepsis.

Interleukin-6 (IL-6) is likely to be an important mediator of the inflammatory response. We measured levels of this cytokine in plasma samples from 37 patients with sepsis or septic shock obtained at the time of admission to the intensive care unit and related these levels to hemodynamic and biochemical parameters as well as to clinical outcome. In 32 of the 37 patients, increased levels of IL-6 were found, occasionally up to 7,500 times the normal level. The highest IL-6 levels were encountered in patients who suffered from septic shock (P value of the difference between patients with and without shock less than .0001). In addition, IL-6 significantly correlated with plasma lactate (P less than .0001), heart rate (P = .05) and, inversely, with mean arterial pressure (P = .01) and platelet counts (P = .0002). Significant correlations of IL-6 with the anaphylatoxins C3a (P = .0001) and C4a (P = .0002) and with the main inhibitor of the classical pathway of complement, C1-inhibitor (inverse correlation, P = .05), were also observed. IL-6 on admission appeared to be of prognostic significance: levels were higher in septic patients who subsequently died than in those who survived (P = .0003), in particular when only patients with septic shock were considered (P less than .0001). All nine septic patients with levels of less than 40 U/mL on admission survived, whereas 89% of the nine patients with levels exceeding 7,500 U/mL died. These data provide evidence for a role of IL-6 in the pathophysiology of septic shock. Further studies are needed to reveal whether IL-6 in sepsis is directly involved in mediating lethal complications or whether it is to be considered as an "alarm hormone" that reflects endothelial cell injury probably mediated by the anaphylatoxines.

Activation of the complement system during immunotherapy with recombinant IL-2. Relation to the development of side effects.

Therapy with high doses of rIL-2 is complicated by the occurrence of hypotensive reactions and the development of a vascular leakage syndrome (VLS). In four patients, who together received seven cycles of high doses of IL-2 (up to 12 x 10(6) U per m2 per day), and who developed these side effects, we observed an unexpected increase in plasma levels of C3a, indicating activation of the complement system. C3a levels markedly increased during IL-2 therapy from 4 nmol/liter (mean level) before therapy to 23 nmol/liter at the end of the cycle. Activation of C3 occurred via the classical pathway inasmuch as C4a levels also increased during therapy. Mean daily C3a levels correlated with signs of the VLS, such as daily weight gain (p less than 0.001) and albumin levels (inverse correlation, p less than 0.001). In five additional patients, who together received seven cycles of lower doses of IL-2 (2 x 10(6) U per m2 per day) and who did not develop a VLS, only moderate increases in C3a levels (up to 13 nmol/liter) were observed. The highest levels at the first day of the regimen (mean: 7 nmol/liter) occurred 8 h after the IL-2 infusion. Thus, administration of IL-2 induces a dose-dependent activation of the complement system in vivo, which appeared to be related to the development of side effects of this therapy, such as the VLS.

Quantification of plasma factor XIIa-Cl(-)-inhibitor and kallikrein-Cl(-)-inhibitor complexes in sepsis.

Considerable evidence indicates that activation of the contact system of intrinsic coagulation plays a role in the pathogenesis of septic shock. To monitor contact activation in patients with sepsis, we developed highly sensitive radioimmunoassays (RIAs) for factor XIIa-Cl(-)-inhibitor (Cl(-)-Inh) and kallikrein-Cl(-)-Inh complexes using a monoclonal antibody (MoAb Kok 12) that binds to a neodeterminant exposed on both complexed and cleaved Cl(-)-Inh. Plasma samples were serially collected from 48 patients admitted to the intensive care unit because of severe sepsis. Forty percent of patients on at least one occasion had increased levels of plasma factor XIIa-Cl(-)-Inh (greater than 5 x 10(-4) U/mL) and kallikrein-Cl(-)-Inh (greater than 25 x 10(-4) U/mL), that correlated at a molar ratio of approximately 1:3. Levels of factor XII antigen in plasma and both the highest as well as the levels on admission of plasma factor XIIa-Cl(-)-Inh in 23 patients with septic shock were lower than in 25 normotensive patients (P = .015: factor XII on admission; P = .04: highest factor XIIa-Cl(-)-Inh; P = .01: factor XIIa-Cl(-)-Inh on admission). No significant differences in plasma kallikrein-Cl(-)-Inh or prekallikrein antigen were found between these patients' groups. Elevated Cl(-)-Inh complex levels were measured less frequently in serial samples from patients with septic shock than in those from patients without shock (P less than .0001). Based on these results, we conclude that plasma Cl(-)-Inh complex levels during sepsis may not properly reflect the extent of contact activation.

The activation of polymorphonuclear neutrophils and the complement system during immunotherapy with recombinant interleukin-2.

The toxicity due to interleukin-2 (IL-2) strongly resembles the clinical picture seen during septic shock. In septic shock activation of polymorphonuclear neutrophils (PMN) and the complement system contribute significantly to the pathophysiology of the condition. We therefore investigated whether similar events contributed to the toxicity observed with IL-2. Four patients received seven cycles of escalating dose IL-2 (18.0 to 72.0 X 10(6) IU m-2 day-1) and 16 were treated with 20 cycles of fixed dose IL-2 (12.0 or 18.0 X 10(6) IU m-2 day-1). Toxicity, as judged by hypotension (P = less than 0.005) and capillary leakage (fall in serum albumin 18.2 vs 4.0 gm l-1; P = less than 0.0005 and weight gain 4.0 vs 1.2 kg; P = less than 0.025) were worse with the esc. dose protocol. PMN became activated following IL-2 with mean peak elastase/alpha 1-antitrypsin (E alpha 1 A) and lactoferrin values of 212 (SEM = 37) and 534 (SEM = 92) ng ml-1 respectively occurring 6 h after the IL-2. Peak values for the esc. dose IL-2 group being generally higher than 500 ng ml-1. Activation of the complement cascade was evidenced by a dose dependent elevation of peak C3a values (fixed dose 9.1 (SEM = 0.6); esc. dose 25.7 (SEM = 6.33); P = less than 0.005) on day 5 of IL-2. There was a significant correlation between C3a levels and the degree of hypotention during the first 24 h after IL-2 (r = 0.91) and parameters of capillary leakage such as weight gain and fall in serum albumin (r = 0.71). These data suggest that activation of PMN initiates endothelial cell damage which subsequently leads to activation of the complement cascade. This latter system then contributes to the haemodynamic changes and capillary leakage seen in IL-2 treated patients.