key: cord-015021-pol2qm74
authors: nan
title: Third International Congress on the Immune Consequences of Trauma, Shock and Sepsis —Mechanisms and Therapeutic Approaches
date: 1994
journal: Intensive Care Med
DOI: 10.1007/bf02258437
sha: 
doc_id: 15021
cord_uid: pol2qm74

nan

This issue of the journal contains the abstracts for the Third International Congress on the Immune Consequences of Trauma, Shock and Sepsis -Mechanisms and Therapeutic Approaches. We hope that the information contained in this special issue will stimulate you to participate in the congress, to contribute to the knowledge being developed in this field and to use this information to help you in providing better care for your patients. We thank the editors and the editorial board and publishers of the journal for their interest and support in preparation of this special issue. We also, on behalf of the scientific committee, welcome you to the Third International Congress in Munich on 2-5 March 1994. When, in the mid-1980s, we thought of having a worldwide congress, we hoped to bring together investigators to discuss this theme. The explosion of knowledge occurring around that time provided an excellent background against which the first conference in 1988 provided stateof-the-art information and consensus on factors involved in injury and sepsis. In 1991, the Second Congress was held at the time of another resurgence of research, study and information on injured and operated patients. It seemed then that there would be a lull in the development of new information and therapy, and that another state-of-the art conference might not be necessary until 1995 or 1996.

However, the explosion in molecular biology has continued. The wonderful world of cytokines has gone from ILl to IL-6 to IL-11, IL-12 and IL-13 and beyond. The vast amount of information about mediators and their importance in disease is impressive. This has all suggested a magic bullet that might be used to alter or block inflammatory responses. This has not happened, however, and the question is "Why not"? Our science is powerful, but our therapy is still weak. What are the issues, then, in 1994, to be dealt with at this symposium and congress? (1) Proposals for new terminology. There have been a number of proposals for new terminology and new classifications of injury, sepsis, inflammation and various other problems related to human illness. The question is whether this is the way to go. Will this contribute to better clinical trials, information basis and better research? The pros and cons of this development will be reviewed by those making the proposals and those questioning the need for and wisdom of this effort. (2) Magic bullets: the prospect of a magic bullet to deal with inflammation in injury and infection seemed highly promising earlier. Many preclinical trials and a lot of animal research suggested the possibility of a great breakthrough in clinical care. What has become, then, of all the expensive and extensive multi-institution randomized, placebo-controlled, double-blind clinical trials of agents that block mediators and endotoxin. Many such studies have yielded equivocal, marginal or negative results. The reasons for this and the future of clinical research will be the subject of presentations and discussions to set the stage for further work. (3) Should future clinical trials be based on new classifications of illness such as MODS, SIRS, APACHE III, SAP II, MRM, etc., or should trials be dedicated to specific diseases -urinary tract infections, pneumonia, trauma patients, cardiac surgery and other specific problems, rather than generalized problems of sepsis, the sepsis syndrome and other classifications? In other words, should we now begin to have clinical trials on specific diseases with causes that are known and can be attacked? The causality of disease becomes an important consideration in this regard. (4) A multitude of potential therapeutic agents has been proposed on the basis of animal studies. How should we decide which of them should be brought to clinical trial? The possibilities are endless as we develop new clinical information about the mechanisms and pathogenesis of human disease. (5) Information on the pathogenic mechanism of disease states and of injury continued to emerge in an explosive fashion, and in light of our gathering knowledge we can look forward to working out a cohesive system of response to injury. (6) Additional information will be provided in plenary sections, many symposia and free communication sessions and posters, which will update the participants on a variety of relevant topics presented by many of the leading in-IV vestigators in these fields. Topics will range from molecular mechanisms, such as signal transduction, through the explosive growth of information on the role of cytokines and pathophysiology, to practical considerations in the design of immunomodulatory therapeutic regimens. These merely touch on a few areas, from the basic to the clinical, which will be the subjects of those symposia. All this information will fit into the jigsaw of this exciting area and its stimulus to further research study.

This promises to provide an exciting, educational programme with experts and participants from all over the world. We hope it will set the stage for many years to come and will increase our understanding of trauma, shock and sepsis and help us to provide better therapy for those of our patients who are affected by such problems.

A. The clinical syndrome of MODS versus MOF will be reviewed in detail by those who have made these proposals.

B. An extensive review of the design and interpretation of clinical trials in patients with shock and injury will be provided. The reasons why so many clinical studies in the recent past have been negative will be reviewed.

The therapeutic strategies that are being developed for the treatment or prevention of MODS or MOF will be the subject of another panel discussion by experts who have been involved in and contributed to this area. A consensus conference or controversy conference will be presented about various aspects of MODS or MOF, including the benefits of supernormal oxygen delivery, bacterial translocation, parenteral nutrition, the immune response and other aspects. The successes and failures of completed clinical trials will be presented by those who are involved in these clinical trials, with a refreshing review of the problems related to that injury. There will be late news about studies just being completed at present or after the beginning of 1994 and where they stand.

C. The mechanisms and biochemical profiles of specific organ dysfunction or failure will be reviewed. What are the definitions? What are the mechanisms? How can organ dysfunction and/or failure be defined? An extensive review of the biological mechanisms involved in production of injury by mediators will be presented.

A session will be devoted to how future ongoing trials might be better designed and what can be done about the studies recently completed, many of which are negative. D. The immunological or inflammatory pathways resulting in organ injury will be reviewed in detail in presentations and a panel discussion.

We look forward to welcoming you to an exciting and rewarding conference, which undoubtedly possesses the potential to become a landmark event and major reference point for any scientific discussion about the complex of host defense dysfunctions following trauma, shock and sepsis.

Studies over the past 25 years have established that the contact system, which forms bradykin/~, is gax important mediator in hypotensive septicemia. In addition to hradyk{nln, another product of the contact system, kaIlikrein, can mediate inflammation by virtue of its chemotaetic mad neutrophJ/activating properties. Using functional and immunochemical tech~2ques, we have demonstrated activation of the contact system in the adult respiratory distress syndrome in typhoid fever and clin/cal sepsis. We have also been able to inhibit the hypotension but not the disseminated intravaseular coagulation in a model of primate sepsis by the use of a monoclonal antibody directed agsi~st factor XII, the initiating protein of the contact system, in volunteers given E. coil endotoxin, who did not develop hypotension, we were also able to demonstrate activation of the contact system with a rise of alpha-2macrogiebulin-kalllkrein complex. We have also examined, J~ an i~tensive care situation, patients with SIRS. We found that serial measuremezzts of the contact system were useful in eva~u~ting prognosis+ These studies suggest that inhibition of kalllkrein a~d l e r bradykinin actions might be useful i~ obviating many .of the features seen in sepsis and septic shock. Dextran sulfate (DXS) activates the contact system and, in vivo, produces transient hypotension. In order to better define the mechanisms underlying the DXS-induced hypotension, we investigated the effects of either the plasma kallikrein inhibitor, des-Pro2-iArg]5]aprotinin (BAY 4620) or the B2 kinin antagonist, Hoe 140 on the hypotensive response to DXS. In the first study, anesthetized miniature pigs (5 pigs/group, randomly assigned) were given one of the following treatment protocols: 1) DXS (5 mg/kg), 2-5) DXS plus BAY 4620 (45, 90, 180, or 360 rag), or 6) saline. DXS alone produced a profound but transient systemic arterial hypotension with a corresponding reduction in plasma kinin-containing kininogen. Circulating kinin levels, complement fragment C3adesArg and fibrin mom)mer were all increased. BAY 4620 produced a dose-dependent delay or attenuation in these effects with the highest dose completely blocking DXS-induced hypotension and elevations of kinin, C3adesArg and fibrin monomer levels. Thus, the effects of DXS are solely dependent on contact system activation and this activation is sensitive to BAY 4620. llowev~:r, contact system activation is known to produce changes in a variety of vasoactive mediators, all of which can affect blood pressure. In a second study, two groups of pigs (3/group) were given either DXS alone (2 mg/kg) or DXS 10 minutes after a bolus injection of Hoe 140 (30 #g/kg). DXS alone produced transient hypotenmon. This response was completely blocked by Hoe 140 pretreatment. Both groups had identical reductions in kinin-containing kininogen. We conclude that DXS-induced hypotension is produced by activation of the contact system which results in the production of bradykinin. Liberation of bradykinin is both necessary and sufficient to produce all of the hemodynamic changes observed.

Dr. Matthias Siebeck, Department of Surgery, University of Munich, Klinikum lnnenstadt, Nussbaumstrasse 20, D-80336 Munich, Germany

In experimental animals exposed to i.v. injection of endotoxin accumulation of leukocytes in various organs as lungs and the liver is a prominent feature. As a part of these morphological changes damages of endothelial ceils are regularly seen. This process, which is a part of endothelial-cellular interaction, leeds to exposure of the sub-endothelial basement membran. The basement membran is known f6r its capacity to activate the contact system of plasma. During this cascade activation, coagulation factor XII is converted to the active factor XII. This activation might produce increased plasma kallikrein activities and thereby give release of the vasoactive substance bradykinin. Using a porcine model we have noticed that endotoxin infusion (0,01 mg/kg) induces elevated plasma kailikrein activities within two hours after the start of the infusion. This enzyme activity remained increased during the next hours and reached value of up to 50 U/1.

In patients with sepsis we also have observed elevated plasma kallikrein activities with enzyme activities up to 200 U/1. In order to further elucidate the significance of these elevated enzyme activities, we prepared human plasma kallikrein and injected it intravenously in anaesthetized pigs (1). When very small plasma kailikrein activities (0,3 U/kg bodyweight) were given intravenously a 50% decrease in arterial blood pressure was seen in the animals.

In the patients with sepsis also decreases in prekallikrein values and functional plasma kallikrein inhibition are frequently seen. Furthermore, degradation of high molecular weight kioinogen is found in these patients indicating formation of bradykinin.

These experimental and clinical studies underline that contact activation in sepsis might results in the release of very powerful mediator substances which can be of pathophysiological importance in this disease. A number of pathological disorders as reperfusion injury, bone marrow transplantation, polytrauma and septic shock are associated with capillary leakage. As the activation of the complement system and the contact phase play a major role in these diseases we investigated whether Cl-lnhibitor (C1-INH), which inactivates Cl-esterase, kallikrein and clotting factors XII and Xl, could abolish vascular leakage. A capillary leakage was induced in rats by the administration of interleukin-2 (5 x 106 IU/kg). The increased vascular permeability was monitored for one hour as the extravasation of FITC marked rat serum albumin from a mesenterial vessel by a video-image processing system. CI-INH (Berinert®, Behringwerke) given as a single i.v. bolus in concentrations of 100, 250 or 500 U/kg dose-dependently prevented the capillary leakage. Carrageenaninduced inflammation in the rat leads to vascutar leakage and to edematous swelling of the paw. CI-INH in this model leads to a dose-dependent decrease in paw edema formation. Finally, we investigated the effect of CI-INH (infusion (100-125 U/kg x h) on a LPS-induced shock in the rat by combination therapy with the antithrombotlc agents antithrombin Ill (Kybernin®) or rec. Hirudin (both substances from Behringwerke). In this animal model mortality was 90 % in the untreated control. Both antithrombotic agents decreased mortality rates by inhibiting formation of DIC; a further significant improvement of survival was achieved by the treatment with CI-INH. Thus+ it could be concluded that C1-INH has a beneficial effect in diseases associated with a vascular leakage. ICLB and Laboratory for Experimental and Clinical Immunology, University of Amsterdam, The Netherlands; 2Thrombosis Research Center, Temple University, Penn., USA; 3Oklahoma Medical Research Foundation,. Ok. City, USA.

To evaluate the contribution of the contact system to activation of other mediator systems in an experimental model of sepsis, we investigated the effect of mAb C6B7 which inhibits activation of factor XlI, on activation of complement and fibrinolytic cascades and activation of neutrophils in baboons suffering from a lethal sepsis. Activation of the complement system was assessed by measuring circulating levels of C3b/c and C4b/c, and a significant reduction was observed in 5 animals that had received a lethal dose of E. coli together with mAb C687 (treatment group), compared to 4 animals that had received a lethal dose of E. coil only (control group). Activation of the fibrinolytic system as reflected by circulating plasmin-=2antiplasmin complexes and tissue plasminogen activator, and activation of neutrophils, assessed by measuring circulating elastase-=l-antitrypsin complexes, was also significantly less in the treatment group. We conclude that activation of the contact system protein factor Xll during the inflammatory response to a lethal dose of E. coil in this baboon model, modulates directly or indirectly activation of the complement and fibrinolytic systems and that of neutrophils.

In a prospective study, plasma levels of C3a, C3, and C5a were measured in 30 patients from an internal intensive care unit. 20 patients were clinically septic defined by the criteria of Bone et al.(l) . The remaining 10 patients were critically ill but didn't fulfill the clinical criteria of sepsis. From both groups of patients blood samples were taken over a l0 days period. During the first 3 days blood samples were drawn every 8 h, on day 4-6 every 12 h and the last 4 days once daily. Mean plasma concentrations of C3a within the first 32 h after clinical onset of sepsis were 1019 + 618 pg/ml, whereas non-septic-patients exhibited mean values of only 595 +_ 312 p_g/m/. C3 levels were lower for septic-patients (840 + 480 lag/ml) than for non-septic-patients (1340 _+ 770 lag/ml). The most profound difference between both groups was found, when the C3a/C3 ratio was compared (1.84 + 2.28 for septic-patients and 0.4 _+_ 0.18 for the control group). No significant differences between both patient groups were observed in C5a plasma levels (5.9 + 2.1 ng/ml in septic-patients vs. 5.6 _+ 1.5 ng/ml in control patients). In 13 of 20 cases of clinically defined sepsis causative organisms like bacteria, protozoa or fungi could be cultured from blood, bronchoalveolar lavages and/or section materials. Application of the complement parameters to survivors (n=9) and non-survivors (n=l 1) within the septic-group revealed, that the C3a/C3 ratio could also be used as a prognostic parameter for clinical outcome. The possibility of rapid and easy measurement of C3a and C3 in only 20-25 minutes (2) and the significant difference of the C3aJC3 ratio between the septic and non-septic group renders this parameter a good candidate for early diagnosis of sepsis in the intensive care unit. Hirudin, a single polypeptide chain composed of 65 amino acids with 6 cysteine residues (MR 7000 daitons), is the most potent and specific thrombin inhibitor, which is now available as a genetically engineered product (rec. Hirudin -HBW 023, Behringwerke; Marburg). The aim of our study was to establish a rabbit model of tissue factor (TF) induced activation of the extrinsic pathway of coagulation and to evaluate the therapeutic efficacy of rec. Hirudin. Coagulation was induced in female NZW rabbits by infusion of 0.5 p.g/kgxh thromboplastin for 7 hours. Development of disseminated clotting was manifested by a decrease of fibrinogen and platelets to 27.0 % and 33,0 % respectively, and by an increase of fibrin monomers from 13.2 to > 85.0 ~tg/ml. We administered rec. Hirudin to rabbits in 3 different concentrations (0.5, 1.0 and 2.0 mg/kg); treatment started simultaneously with the infusion as an i.v. bolus. Rec. Hirudin significantly prevented the decrease of fibrinogen, platelets and the increase of fibrin monomers. This effect was dose dependent and long lasting, even 7 hours after the administration of rec. Hirudin, clotting was still significantly reduced. As could be drawn from the plasma levels, rec. Hirudin had been cleared from plasma at this time. In a post-treatment study we administered rec. Hirudin (0.5, 1.0 and 2.0 mg/kg i.v. bolus) as late as 2 hours after the start of TF infusion. At this time there was already a prominent activation of coagulation. Even in this post-treatment regimen rec. Hirudin significantly prevented disseminated clotting. Hence, it was concluded, that rec. Hirudin by inkihiting thrombin could be effective in the prevention of coagulation disorders including disseminated intravascular clotting (DIC) induced by a septic disease.

Research Laboratories of Behringwerke AG, 35001 Marburg, Germany $3

Novel Protease Inhibitory Activities of the Second Domain of Urinary Trypsin Inhibitor (R-020) and its Effect on SepNs-lnduced Organ Injury in Rat Atsuo Murata 1, Hitoshi Toda 1, Ken'ichi Uda 1, Hidewaki Nakagawa 1 , Takesada Mori 1, Hideaki Morishita 2, Tom Yamakawa 2, Jiro Hirese 2, Atsushi Ni~ 2, Nariaki Matsuura 3 1Osaka University Medical School, Osaka, 2Mochida Pharmaceutical Co. Ltd.

Tokyo, 3Wakayama Medical Schoof, Wakayama, Japan Inhibitory-activities of the second KuNtz-type inhibitor domain of human urinary trypsin inhibitor (UTI) and its effect on sepsis-induced organ injury in rat were investigated by using the recombinant protein. UTI is a glycoprotein with a structure in which 2 Kunitz-type inhibitor domains are linked in a row. We isolated the gene encoding the second Kunitz-type inhibitor domain of UTI, and then constructed expression plasmids by ligating it to the E. coli PhoA signal peptide gene. These plasmids expressed the second domain in E. coil strain JE5505 which lacks the membrane lipoprotein. The recombinant second domain (R-020) inNb[ted trypsin, plasmin, neutrophil elastase and chymotrypsin. In addition it inhibited blood coagulation factor Xa and plasma kallikrein in a concentration dependent and competitive manner. The in vivo effect of the recombinant R-020 was investigated in a rat model of septic shock induced by cecal ligation and puncture. The administration of R-020 significantly improved the survival rate of the rats and attenuated the pathological changes of lung and iiver. We found out the novel protease inhibitory activities of the second domain of UTI and its protective effects on sepsis-induced organ injury. Macrophages are known to secrete lysosomal proteinases,mainly cathepsin B and cathepsin L, and also ~-proteinase inhibitor (PI),related to acute phase proteins.Disturbances of proteinases/ proteinase inhibitors correlates with inflammatory process,leading sometimes to noncontrol "pathGlogical" proteolysis (Jochum et ai.,1990) . The cathepsin L-like and cathepsin B-like activity were measured in serum of 42 patients with chronic bronchitis (14 -with obstructive, 28 -with nonobstructive bronchitis),acute bronchitis (15) and 35 healthy persons.Simultaneously the level of~PI was determined in the same groups.Cysteine proteinases were measured with help of fluorogenic substrates,as was presented earlier (Korolenko et ai.,1991) , ~PI with help of immune enzyme method. It was shown increase of cathepsin L-like and cathepsin B-like activities during aggravation of chronic bronchitis comparatively to the controls (2-4 fold) .After treatment there was a tendency to normalization of indices,but the increase was about 30-40% more than the control values.~PI level in this group was also increased (two-fold),in patients with acute bronchitis -2-4-times more comparatively to the control.It is possible to conclude that chronic bronchitis induced increased secretion both cysteine proteinases and d{PI into blood. Some peculiarities of ratio were noted in patients with emphysema. Endotoxins are microbial products derived from the outer cell membrane of Gram negative bacteria. The active component of endotoxin is lipopolysaccharide (LPS), a complex macromolecule consisting of polysaccharide covalently bound to a unique lipid, termed lipid A. Now recognized to embody the endotoxic principle of LPS, lipid A consists of a/31-6 diglucosamine backbone, both ester and amide linked fatty acids, some of which are acyloxyacylated, and charged constituents such as phosphate, phosphorylethanolamine and 4 amino arbinose LPS, exerts its biological effects in vivo by noncytotoxic interactions with a variety of host inflammatory mediator cells, primarily the mononuclear phagocyte and the endothelial cell, although other host cells also participate. These interactions are modulated by LPS-specific binding proteins found in plasma, including LPS-binding protein (LBP) sCD14 and perhaps other proteins as well. Specific receptors for LPS have been identified on mammalian cells which mediate signal transduction via multiple pathways. LPS-activated host cells are stimulated to secrete or express multiple proinflammatory mediators, including TNF-a, ILla, IL-1/3, IFN-a, IL-6, IL-8, IL-10, PAF, PGE, LTB 4 and procoagulant activity. The overproduction of these proinfiammatory mediators results in the manifestations of endotoxemia, observed experimentally as fever, hypotension, disseminated intravascular coagulation and death. Modulation of activity of these mediators protects animals against lethality. Similar pathways are thought to be operative in Gram negative sepsis, and control studies with human volunteers support such conclusions. Immunotherapeutic approaches in clinical Gram negative sepsis have, to date, been less successful. In vitro experiments and studies in animal models have recently shown that several proteinaceous bacterial exotoxins can evoke cytotoxic effects that ultimately lead to cardiovascular collapse and shock. Since the possible relevance of bacterial exotoxins in the pathogenesis of septic shock has received very little attention in the past, an attempt will be made here to provide a brief overview of this generaily neglected topic.

Protein toxins act intracellularly or they dz~nage the integrity and function of the plasma membrane. Major representatives of the former group are the adenosine diphosphate (ADP)-ribosylating toxins, e.g. cholera and cholera-like toxins, diphtheria toxin), and the neurotoxins. Most medically relevant toxins of this category have been studied in great detail. Although often responsible for severe and sometimes fatal disease, their association with septic shock is rare. In contrast, experimental evidence is accumulating for a role of membrane<lamaging toxins in the pathogenesis of inflammatory lesions. Formation of transmembrane pores is probably the most widespread mechanism via which such physical membrane perturbation can occur (1,2), the present discussion will be restricted to this category of toxins. The author shall first discuss basic properties of pore-forming proteins, and consider the factors that direct their attack to specific targets. Thereafter, attention will be drawn to diverse functional consequences that may follow membrane damage so that a general concept of how pore-forming toxins may contribute towards the development of shock will evolve. I. Bhakdi, S. and Tranum-Jensen. J. 1987 Gram positive bacteria produce exotoxins that, at least in part, exhibit the unusual properties of superantigens. Superantigens (SAg) activate V/3 selectively T cells to produce lymphokines including I1-2 and TNF/lymphotoxin. Systemic application of the SAg staphylococcal enterotoxin B (SEB) to D-galactosamine sensitized mice causes lethal shock within hours. SEB reactive V/38 + T cells accumulate within 90-120 min mRNA for the II-2, I1-4, TNFaI~ and IFN-3,. Parallel in time high concentrations of bloodborne I1-2 and TNF are noted. Anti TNF t~//3 neutralizing mAb protect mice against SEB induced T cell dependent lethal shock thus indicating a key role of TNF in effecting lethal toxicity. Within 12-18 h distinct levels of tolerance to SEB become discernable on ligand reactive V/38 T ceils, dependent of the dose of SEB used. These levels include anergy, T cell receptor downregulation, loss of CD2, CD4, CD8 accessory molecules and programmed cell death.

M.Freudenberg, Y.Yaegashi, P.Nielsen, C.Galanos. The sensitivity towards endotoxin (LPS) is genetically determined, however it may be increased under different experimental conditions. These include treatment with hepatotoxic agents such as D-galactosamine, and with live (infection) or killed bacteria. It is well established now that D-galactosamine sensitizes to LPS by increasing the susceptibility of the host to toxic activity of LPS-induced TNF~. Sensitization by bacteria, especially by Gram-negative once is of especial interest because it implies that infecting microorganisms not only produce LPS, but als0 sensitize the host to its lethal activity. Sensitization to LPS by bacteria proceeds in all LPS-sensitive (Lps n) mouse strains that have been investigated so far. It also proceeds in LPS-resistant (Lps d) C3H/HeJ mice. The absence of sensitization to LPS in Lps d C57BL/10ScCr mice was identified as being due to their inability to produce interferon-7 (IFN?). Administration of exogenous IFN? to these mice conferred sensitivity to LPS. Conversely, administration of anti-IFN? to Lpsn; mice inhibited the development of sensitization'by bacteria. It may be concluded therefore that IFN7 is the mediator of sensitization to LPS by bacterial infection.

The IFN? defect of C57BL/10ScCr mice concerns only the IFN7 response to bacteria, since the response to the T-cell mitogen Con A and to CD3 monoclonal antibodies was not impaired. The IFN?producing cells themselves were shown to be intact. On closer investigation the impaired IFN? response was identified as the result of a defective accessory function of macrophages. Macrophages of C57BL/10ScCr mice are incapable of producing interferon-~ (IFN~) which we could show to be obligatory for the production of IFN? in response to bacteria. Although antibiotics are required to clear bacteremia, they do not reverse evolving shock, because endotoxin free in the bloodstream or exposed on the surface of circulating bacteria continues to activate the production of inflammatory mediators, Furthermore, antibiotics have been shown to induce release of endotoxin from gram-negative bacteria during the treatment of severe infection. In an in-vitro study, using live Escherichia coil, we have observed the release of endotoxin upon treatment with several antibiotics like cefuroxim, imipenem, ceftazidime and aztreonam. Incubation with imipenem or ceftazidime induced an early lyeis and a mild rise in endotoxin levels, whereas incubation with cefuroxime or aztreonam resulted in the formation of filaments with a late but dramatic rise in endotoxin levels. In a second study we studied the influence of different antibiotics on the TNF-= production of E.coli stimulated human monocytes. We found again great differences in the production of this important cytokine among the different antibiotics according to their capacity to induce liberation of endotoxin. These differences in the amount of endotoxin release are probably caused by Penicillin-Binding-Proteins (PBP) specificities of the antibiotics with resultant differential morphological changes, In a rat sepsis model treatment with imipenem or ceftazidime resulted in a transient increase in IL-6 levels, whereas treatment with aztreonam resulted in progressively increasing levels of IL-6 and a significant increase in mortality. The increased mortality in PBP-3 specific aztreonam treated rats might have been the result of filament formation in the abdominal cavity of the septic rats. The endotoxin sequestered at local sites may account for the precipitation of gram-negative shock. Finally, we've also demonstrated that in patients under treatment for septic shock endotoxin release can be related to the administration of antibiotics.

Carbapenem-and Cephalosporin-lnduced Release of LPS from

Smooth and Rough Pseudomonas aeruginosa: In-Vivo Relevance J. Jackson and H. Kropp, Merck Research Laboratories, Rahway, NJ B-lactam (cell-wall active) antibiotics are generally deemed the class of antibiotics most responsible for the release of free endotoxin (LPS) from gram-negative bacteria due to their cell lytic actions although all antibiotic classes studied thus far induce release of endotoxin. We have recently shown in-vitro that antibiotics within the b-lactam class (i.e. imipenem , meropenem and ceftazidime , with equivalent MICs) differ in their propensities to release excessive amounts of LPS from both smooth-(s-) and rough-(r-) LPS laboratory and clinical and antibiotic-sensitive and -resistant P. aeruginosa isolates and that LPS release is independent of cell lysis. Additionally, variations in the induction of endotoxin release is antibiotic concentration dependant and correlate with antibiotic penicillin-binding protein (PBP) affinities which result in morphological changes. These studies also show that LPS differences measured via chromogenic LPS (C-LAL) assay correlate with LPS lethality in LPS-sensitive mice. Release of LPS was compared to changes in CFUs, cell mass, morphology and antibiotic PBP-specificity. Corresponding in-vivo studies in CD1 mice against s-LPS P. aeruginosa isolates show that, despite equivalent in-vitro protection, antibiotic efficacy in mice differ as much as 100-to lO00-fold. To establish a possible in-vivo role for antibiotic-induced release of free LPS we compared antibiotic efficacy results in mice challenged with virulent parent s-LPS and its relatively avirulent r-LPS mutant (lacking only its 0 side chain) P. aeruginosa isolates. Results show the relevance of quantity and physiochemical composition (bioreactivity) of free LPS to virulence and in-vivo antibiotic efficacy. In other in-vivo studies chemical agents that destroy cells (m~) which mediate LPS lethality in-vivo were used to pretreat mice prior to antibiotic therapy and bacterial challenge with wild type Pseudomonas. We conclude that circumstances in which antibiotic-induced filamentation occurs are also conditions that yield excessive LPS release, the in-vivo effects of which are shown through the requirement of larger amounts of antibiotic to afford equivalent protection. Bacterial endotoxins have been reported to play a significant role in the pathogenesis of Gram negative sepsis by their interaction with, and stimulation of, host inflammatory mediator cells to produce cytokines, biologically active lipid intermediates, and procoagulant activity (TFa). Although both microbe-associated and free endotoxin are capable of stimulating mediator production, studies from our laboratory suggest that free endotoxin may be the more potent stimulus for TNF and TFa. Further, our studies suggest that the majority of the proinflammatory activity released from antibiotictreated E. coli coisolates with LPS by two different fractionation techniques.

To assess whether actions of endotoxin directly contribute to lethality in Gram negative sepsis, an experimental model was developed in which mice were made hypersusceptihle to the lethal effects of endotoxin by treatment with D-galactosamine (DgalN).

Challenge of CF-1 DgalN-treated mice with E. coli Olll:B4 resulted in an LDs0 of approximately 2.5 x 104 CFU. Resolution of the peritonitis and bacteremia by treatment with cell wall-active antibiotics resulted in an increase in the LDso. Ceftazidime and imipenem, which differ in their mechanism of action and in their capacity to effect endotoxin release in vitro, also differed in their in vivo capacity to provide chemotherapeutic protection (3 fold vs 8 fold respectively, p< .01). Parallel studies with endotoxin hyporesponsive C3H/HeJ mice indicate significantly greater levels of protection with imipenem (> 80 fold vs saline controls). Collectively these data suggest that endotoxin may contribute directly to the pathogenesis of experimental Gram negative sepsis. Bacterial lipopolysaccharides (LPS) are the endotoxins of gram-negative bacteria and represent their major surface antigens. LPS is made up of three chemically, biologically and genetically disctinct regions, i.e, the O-chain, the core region and the lipid A moiety whereby the latter represents the endotoxic center.

It is our current understanding that LPS is responsible for many of the pathophysiological events observed during gramnegative infections and that one of the major mechanisms leading to shock and death is the LPS-induced activation of macrophages resulting in the production and release of lipid and peptide mediators, among which tumor necrosis factor seems to be the most important.

Therefore, in the fight against the lethal outcome of gram-negative infections, modern strategies, in addition to antibiotic treatment, aim at i) the neutralization of tumor necrosis factor, ii) the inhibition of the production of tumor necrosis factor or iii) the neutralization of the activation potential of LPS for macrophages by monoclonal, preferably human antibodies. The latter approach, to be effective against a broad spectrum of gram-negatives, must be directed against common structures of LPS (lipid A and core region). The molecular basis of this approach and the controversy in this field will be discussed.

Passive immunotherapy has been used since 1893, when von Behring described the administration of immune horse serum to treat a patient with diphteria infection. Even if this therapy was sometimes successful in bacterial infections, it has been largely replaced by antibiotics. However, antibiotics have their limitations, especially in critically-ill patients. To improve outcome, adjunctive therapies such as immunotherapy with polyclonal and monoclonal antibodies particularly against endotoxin are again considered. The role of humoral immunity in host defenses against bacterial infections is weU known. For instance, tile importance of antibodies in the defense against gramnegative infections has been established clinically by studies relating the outcome of patients with gram-negative bacteremia to tilers of antibodies directed at the offending pathogens at the onset ofbacteremia (McCabe 1972; Pollack 1979) . Ever since we know the role of endotoxins in the pathophysiology of sepsis, antibodies against the S-and R-LPS have also been detected in sepsis patients. The aim of the administration of IV/G to the sepsis patient is as follows: 1 ) Enhancing of opsonization and phagocytosis(antibactericidaI activity) 2) Synergistic effects with [3-1actam antibiotics 3) Neutralization of endotoxin, the main pathogenic mediator of gram-negative sepsis 4) Modulation and/or inhibition of cytokine release The enhancement of opsonic-and phagocytic-activity especially with IgG via Fc and C3 receptors has been well documented. Monoclonal antiendotoxin antibodies, proven in clinical studies, do not appear to neutralize endotoxin in vitro and are not reproducibly protective in animal models of sepsis. Also they can not suppress endotoxin-induced TNF-~, IL-6 release in mice (Baumgartner 1990, Corriveau and Danner 1993) . In conlrast, recent studies of a polyclonal immunoglobulin preparation, containing high levels of antibodies against gram-negative bacteria and their O-antigen of LPS in IgG, IgM and IgA classes (Pentaglobin®) provide evidence to neutralize endotoxin. This effect is demonstrated in vitro (Berger 1990 (Berger ,1993 , in animal models (Stephan 1985 , Berger 1993 and also in prospective, randomized, controlled clinical trials (Schedel 1991 , Poynton 1992 , Behre 1992 . Furthermore mortali b' was reduced statistically in patients with septic shock and endotoxemia by using this preparation, as has been demonstrated by Sehedel.

Anti-core LPS monoolonal antibodies: Binding specificity and biological properties F.E. Di Padova, R. Barclay, E.Th. Rietschel. Bacterial LPS and cytokines are responsible for the pathological processes of Gram-sepsis and are suitable targets for therapeutic interventions. Chemical characterization and structural analysis of different LPS have revealed common features. The inner core region of LPS shows a high degree of similarity among E. coli, Salmonella and Shigella. Among a large number of broadly cross-reactive murine anti-core LPS MAb one of these  IgG2ak) has been selected and chimerized into a human IgGlk (SDZ 219-800).

In ELISA and in immunoblots on purified LPS both SDZ 219-800 and WNI 222-5 show a strong reactivity with all smooth LPS from E. coli and Salmonella.

Reactivity with all the known complete core structures from E. coli and Salmonella (Ra) is evident. Reactivity with Re structures or free lipid A is not observed.

This MAb cressreacts with all clinical E. coli isolates from blood, urine and feces and with other enterobacteriaceae. SDZ 219-800 and WNI 222-5 have biological activity as they inhibit the LAL assay and the secretion of monokines (IL-6 and TNF) by mouse and human macrophages. Moreover, SDZ 219-800 and WNI 222-5 inhibit the release of IL-6 and TNF in vivo. In vivo SDZ 219-800 as well as WNI 222-5 neutralize the pyrogenic activity of E. coli LPS and protect mice from lethality in D-GaiN-sensitized mice. The possibility to use WNI 222-5 as a capture antibodies in the immunolimulus assay opens the possibility to differentiate the origin of the LPS in patients with endotoxemia.

Franco Di Padova, SANDOZ Pharma AG, CH4002 Basel, $chweiz $7

Presentation of LPS to CD14 by LPS Binding Protein Peter S. Tobias, Julie Gegner, Katrin Soldau, Lois Kline, Loren Hatlen, Douglas Mintz, and Richard J. Ulevitch.

The activation of myeloid cells by lipopolysaccharides (LPS) has been shown to require the serum glycoprotein LPS binding protein (LBP) and binding of LPS to membrane bound CD14 (mCD14). Other cells such as human umbilical vein endothelial cells (HUVEC), smooth muscle cells, and some epithelial cells, which do not express mCD14 but nevertheless respond to LPS in the presence of serum, have receptors for complexes of LPS with the soluble form of CD14 (sCD14). These complexes of LPS with sCD14 are only formed efficiently in the presence of LBP. We have begun to characterise the mechanisms by which LBP enables LPS to bind to CD14, either soluble or membrane bound. With the use of fluorophore and radiolabelled reagents we have developed procedures for quantitative measurement of the association of LPS with LBP and of LPS-LBP complexes with CD14. These results show that the delivery of LPS to sCD14 is catalysed by LBP, i.e., LBP is not included with the LPS-sCD14 complex. In contrast, on the surface of cells, LBP does not dissociate from the cells after LPS binds to mCD14. The kinetics, equilibria and stoichiometry of these reactions will be discussed in the context of models for cellular activation by LPS and cellular uptake of LPS. Supported by Nltt grants GM37696, AI32021, AI15136, GM08172, and assistance from the Pharmaceutical Research Institute of Johnson and Johnson.

The Scripps Research Institute, IMM-12, 10666 N. Torrey Pines Rd. La Jolla, CA USA 92037.

Modulation of endotoxin-induced cytokine production by LPS partial structures H.-D. Flad, H. Loppnow, T. Mattern, and A.J. Ulmer Department of Immunology and Cell Biology, Forschungsinstitut Borstel, D-23845 Borstel Lipid A constitutes the active moiety of endotoxin (LPS) of Gramnegative bacteria. It activates mononuclear phagocytes to produce cytokines, such as TNF, I1_-1, and IL-6, which are the major mediators of the endotoxic effect of LPS in vivo. Lipid A precursor la (synthetic compound 406) does not induce cytokines, but is able to specifically antagonize LPS-or lipid A-induced mediator production in human mononuclear cells, vascular endothelial cells, and smooth muscle cells. Furthermore, we present evidence for the first time that T-lymphocytes proliferate in response to LPS and express mRNA for interleukin-2 and interferon-~ and that these responses are also antagonized by synthetic lipid A precursor la. When comparing the agonistic and antagonistic activity of lipid A and different partial structures at the functional and binding level, the number and length of the fatty acids and the number of phosphoryl groups were Pound to be of crucial importance. Unexpectedly, lipid A precursor la, although biologically inactive, turned out to be both the most potent antagonist and competitor in inhibiting the binding of LPS. Taken together, our results provide evidence for a model in which lipid A partial structures compete with LPS for specific cell surface receptor(s). In this sense, biologically inactive lipid A analogues may be good candidates as therapeutic agents for the prevention of Gram-negative septic shock. Two mammalian lipid A-binding proteins have been identified that are believed to have important roles in mediating the host response to endotoxin: lipopolysaccharide-binding protein (LBP) and bactericidal/ permeability-increasing protein (BPI). Human LBP shares a 45% amino acid sequence identity with human BPI. Despite the sequence homology, the two lipid A-binding proteins have very different functional activities. LBP is an acute phase serum protein that markedly potentiates the proinfiammatory host response to gram-negative infection by a mechanism which involves binding of the LBP-LPS complex to CD14 receptors on monocytes, neutrophils and endothelial cells. In contrast, BPI is a neutrophil granule protein with potent bactericidal and LPS-neutralizing activities. The divergent functional properties of these two LPS-bindlng proteins can be explained by the inability of BPI-LPS complexes to bind to cell-surface CD14 receptors. A recombinant protein (rBPI23), corresponding to the amino terminal 23kD fragment of human BPI, has been shown to retain the potent biological activities of the hdlo protein and may represent a novel therapeutic agent for the treatment of gram-negative infections, sepsis and endotoxemia.

For therapeutic effectiveness in many clinical situations, rBPI23 will have to successfully compete with relatively high serum levels of LBP (5-60 ~g/mI) for binding to endotoxin and gram-negative bacteria. To evaluate this issue, experiments were conducted to compare the relative binding affinities of rBPI23 and human recombinant LBP (rLBP) for lipid A. The binding of both proteins to Iipid A was specific and saturable with apparent Kd'S of 2.6 nM for rBPI23 and 58 nM for rLBP. In a competition assay format rBPI23 was approximately 75-fold more potent than rLBP in inhibiting the binding of nSI-rLBP to lipid A. These results demonstrate that rBPI23 has a significantly higher affinity for endotoxin than does rLBP and may explain the potent inhibitory activity of low concentrations of rBPI23 in a variety of in vitro functional assays for LPS activation of cells despite the presence of high LBP levels. For example, rBPI23 at 0.2 ~tg/mi was able to totally inhibit LPS-induced TNF release from monocytes despite a 100-fold weight excess of rLBP over rBPI23. and for heparin binding. Three separate domains which inhibit the LAL reaction to LPS and bind to heparin were identified in amino acid regions 17-45, 65-99 and 142-169 . A single synthetic peptide (85-99) was bactericidal. These results suggest that rBPI23 contains three separate functional domains which may contribute to its high affmity interaction with Gram-negative bacteria and heparin. The individual activity of each domain and the cooperative interaction among domains provide the basis for developing rBPI23 analogues with increased biologic efficacy. A considerable body of experimental data has accumulated implicating tumour necrosis factor (TNF) as a principal mediator of the pathophysiological features of septic shock. These data prompted the development of clinical strategies designed to limit excess (inappropriate) TNF production. Monoclonoal antibodies (Mobs) were developed and a phase II dose escalation trial in 80 patients confirmed that the Mab was safe, and suggested that it was having a beneficial effect on certain parameters. Preliminary results of a large phase III study indicated that (a) the Mob was safe; (b) that it was of no discernible benefit in non-shocked patients; (c) that it reduced mortality in shocked patients, especially during the first 3 days. An alternative strategy was to take advantage of the high binding affinity of soluble receptors for TNF (sTNFR). sTNFR-IgGFc constructs were made for both the p55 and p75 receptors. Both were effective in animal models of LPS challenge, but when a clinical trial was done with the p75 sTNFR-Fc there was unexpected mortality in the treated arm. Using an animal model of live E.coli sepsis, we have shown that this may have been due to the release of bound TNF from the construct. Plasma enhances while BPI inhibits LPS-induced cytokine production from peripheral blood mononuclear cells (PBMC).

Pseudomonas species produce cytokine-inducing substances which are different from LPS as indicated by the fact that polymyxin B blocks only 40% of the cytokine-inducing activity of these pyrogens. We now tested the effect of plasma and BPI on the IL-1[3-inducing activity of Pseudomonas maltophilia -derived pyrogens (Pmp). Bacteria were cultured to the log phase and filtered (300kD) to obtain Prop. Dilutions of Pmp or LPS were added to PBMC alone or to PBMC in 5% plasma +/-BPI (500 ng/ml). PBMC were incubated for 24 hours at 37°C and total IL-I~ was measured by RIA. Results: IL-I[~ in ng/ml (n=7, mear~+SEM, *p<0.05 vs control). 0 control 0.1_+0 + BPI 0.1+0 5% plas. 0.2_+0 + BPI 0.2_+0 Pmp (ng/ml) LPS (ng/ml) 60 600 3000 1 10 3.2_+1 8.0_+1 16.3_+3 1.2_+. 2 5.2+.6 0.2_+.1 3.3_+.9" 13_+3" 0.1_+0" 0.1_+0" 2.2_+. 5 5.7+1 14_+2 5_+.6* 14_+2" 0.6+. 4" 5.3+1 15-+2 0.1-+0" 0.5_+.3"

CBA, C57BL/6, BALB/C, AKR, DBA, Swiss mice, guinea pigs, rabbits have been used in research work. The toxicity, immunogenicity, mitogenic and immunomodulating activity of LPS have been studied. The possibility of reduction of the toxic activity of LPS on macroorganism by bioglycansimmunomodulators obtained from sea invertebrates anymals (Crenomytilus grayanus, Stromhus gigas) have been investigated too.

LPS has been shown to induce specific antibody response of laboratory animals. CBA mice are high responsive to LPS. LPS stimulates humoral immune response of mice to Tdependent and T-independent antigens and suppresses intensity of the delayed hypersensitivity. The small doses of LPS stimulate functional activity of macrophages, the large doses of LPS -decrease one and show the cytotoxic effect. The bioglycans enhance the resistance of mice to the lethal effect of laDS and provide protection 59-75% of mice. One opens possibility to use of bioglicans for reduction of toxinemia in generalizated forms of pseudotuberculosis.

Thus, LPS from Y.pseudotuberculosis is immunogen and immunomodulator wich has influence on humoral and cellular factors of immunity and plays the important role in immunopathogenesis of infection. 

Endotoxaemia is implicated in the pathophysiology of obstructive jaundice. The Lirnulus lysate (LAL) assay is the gold standard method for measuring endotoxin concentrations, but inherent biochemical and technical problems limit the usefulness of this assay. The EndoCab ELISA is a novel assay which measures endogenous antibody (IgG) to the inner core region of circulating endotoxins (ACGA). Objectives We evaluated the significance of endotoxaemia in biliary obstruction using the EndoCab assay and subsequently the specificity of the humoral response to endotoxin compared with an exogenous antigenic challenge [Tetamls Toxoid (TT) ]. Materials and methods In Experiment I three groups of male Wistar rats (250-300g) were studied [No operation (n=39) , Sham operation (n=40), and Bile Duct Ligation for 21 days (BDL)(n=50)]. Plasma was collected and assayed for Bilirubin, Endntoxin(LAL) and ACGA(EndoCab). In Experiment II 30 rats were actively immunised with Tetanus Toxoid ('IT) and then randomised to have No op(n=8), Sham op(n=8) or BDL(n=I4). Blood was taken at this time (To) and 21 days later(T20 at sacrifice for ACGA concentrationslEndoCab] and IgG Produced to TT(TTab) [ELISA] . Antibody concentrations are expressed as % increase from control values.Results In BDL rats, ACGA concentrations were significantly increased compared with controlslP<0.001, Mann-Whitney]. Endotoxin concentrations were sporadically elevated in the jaundiced rats but the rise was not significant. In Experiment [I there was no difference between the ACGA or TTab concentrations in the fllree groups at To, BDL rats had a significant rise in ACGA concentrations by T21 [P<0,001,paired t-test] and humoral response to TT was significantly impaired in BDL rats compared with control groupslP<0.05, paired ttest

Data Plasma endotoxin was measured by means of an endotoxinspecific Endospecy test after pretreatment of the plasma with a new perchloric acid method that we developed. The normal value of plasma endotoxin is less than 9.8 pglml. Polymyxin B was administered at a dose of 12,500 U every 6 hours. Plasma endotoxin rapidly decreased to the normal range in 40 of the 42 patients. Body temperature fall significantly. APACHE II scores were also significantly improved. Tumor necrosis factor-o~ and interleukin 6 decreased in survivors, while in high values tended to persist in patients died. No side effects were observed in any of the patients. In conclusion, intramuscular injection of minute of Polymyxin B was useful in the treatment of endotoxemia.

19-1 Uchimaru, Morioka 020, Japan. l e v a n t G r a m n e g a t i v organisms. M e t h o d s : U n d e r general anesthesia, 12 N o r w e g i a n b r e d landrace pigs (17-30 kg) of either sex, 6 pr group, u n d e r w e n t t r a c h e o s t o m y a n d w e r e v e n t i l a t e d on a 50/50 air a n d o x y g e n m i x t u r e a i m e d at m a i n t a i n i n g a n o r m a l p H a n d a isocapnic level. Ventilation w a s not readjusted d u r i n g the observation period. The anesthesia w a s K e t a m i n e 0.6 m g / k g h a n d D i a z e p a m 2.4 m g / k g h i n t r a v e n o u s l y . H e m o d y n a m i c m o n i t o r i n g of m e a n aorta, p u l m o n a r y artery, central v e n o u s a n d p u l m o n a r y capillary w e d g e pressures w a s p e r f o r m e d w i t h a 5F S w a n -G a n z catheter a n d an aorta catheter.

A continous infusion of R i n g e r ' s acetate ( 1 0 m l / k g h ) w a s g i v e n intravenously. W h e n stabilised, the a n i m a l s w e r e g i v e n 2.5 x l09 CFU of E Colt intraperitoneally as a bolus in 100ml saline, the a n t i b o d y g r o u p received in a d d i t i o n 5 m g / k g E5 a n t i e n d o t o x i n i n t r a v e n o u s l y over 1 h o u r via a n infusion p u m p at the start of the observation period. The a n i m a l s w e r e observed for 6 hours. Results : A t 5 a n d 6 hours, the o x y g e n c o n s u m p t i o n increased by 30 % in the a n t i b o d y treated g r o u p w h e r e a s there w a s a significant fall of 30 % in the sepsis group. In the a n t i b o d y group, the arterial p H a n d the cardiac index were also significantly h i g h e r at the s a m e p o i n t s in time. There w a s no significant difference in arterial pO2. In severe bacterial infections it would be beneficial to neutralize the plasma endotoxin content with complex forming compounds.

The phenothiazines are able to form complexes with endoto×in and the existence of these complexes were already shown in differential speetrophotometry and animal experiments, however, the mechanism of partial neutralization was not clarified. Therefore some representative phenothiazines and structurally related compounds were tested for anti-endotoxin activity.

The endotoxin neutralizinB effects of several benzophenothiazines were investigated in differential speotrophotemetry, TNF induction and in the conventional Limulus test. In animal experiments some beneficial effect of complex forming compounds was found. The benzophenothiazines were not able to inactivate the biological effect of endotoxin in the Limulus test.

The recent findings indicates that a multifocal effect can be responsible for "anti-endotoxin action in vivo".

Effects of TNF inducing effect of endotoxin in leukocytes and bypotensiv action in experimental animals were reduced by some phenothiazine derivatives. Monophosphoril lipid A was without effect.

of Microbiology, Albert Szemt-GydrByi Medical University, Odm t~r lO, H-6720 Szeged~ HunBary 46 Involvement of Streptococcus pyogenes Erythrogenic Toxins in the Induction OflStreptococcal Toxic Shock Syndrome 3 Heide Mgller-Alou~* , Joseph E. Alouf , Die [er Gerlach , ~atherine Fitting., and Jean-Marc Ca~aillon . Unit6 des Toxines Microbiennes and "Unit6 d'Immuno-Allergie, Institut Pasteur, 28, rue du Docteur Roux -75015 PARIS (France) ; Institut f~r Experimentelle Mikrobiologie, JENA (Germany).

superantigen erythrogenic toxin A (ETA) is thought to be involved in toxic shock syndrome in humans by inducing massive release of cytokines by patient immune cells.

The cytokineinducing capacity of ETA w~:s £:ompa~ed to that of LPS, a gram-negative bacterial cell wall component. ETA elicited weak production of IL-1 d and ~, TNF ~ and IL-6 in purified human monocytes whereas LPS stimulated the production of high amounts of these cytokines.

In the presence of T cells, ETA elicited the production of significant amounts of IL-I~, IL-I~, IL-6 and IL-8. However, the most preponderant cytokine was TNF~, which peaked at i0 ng/ml after stimulation with i0 ~g ETA. Comparable amounts of TNFd (ca 4 ng) were induced by 0.i ~g ETA and 0.i ~g LP$. In contrast to LPS, ETA was a strong inducer of TNF~ which was produced only in marginal amounts by LPS. These results suggest that the septic shock induced by gramnegative bacteria (LPS) and by gram-positive bacteria {extracellular superantigens) follows different pathogenic pathways. LPS-induced shock is mainly mediated by monocytes and monocyte-produced cytokines (IL-I and TNF). The ETA-induced shock is mediated by T-cells or depends on T cell help for the production of monocyte-liberated cytokines. Production of T cell cytokines such as TNF~ and interferon 7 in addition to the other cytokines contribute very likely to the severity of the toxic-shock resulting from S. auzeus and S. pyogenes infections in humans. The present study was utidertakc~l to cvalu~tlc the effect of soluble chemically modified giucan during septic shock. Carboxylnethyl-b-i,3-glucan (ram 120000) was injected twice 48 and 24 h before the shock i.v. in a dose of 50 ing/kg. Shock was induced in u~?esthetizcd (sodikm~. l)mntobarbital) rats by i.v.

injection of endotoxin of Escherichia colli 0127 BS, 5 mg/kg. AIIofCMG pretreated ruts survived during first 24haher ¢ndotoxine, while in controi shock group the lethality was 90%. The concentration of ~col)terin in serum was significantly elevated 24 hafterthc second CMGinjection (appare~tly 80 % if compare with the control rats), but didu't chartged 60 rain and 1 S0 rain after endotoxin injectJom Cardiac output in CMOgroup was higher a* the i20 and 180 min after endotoxine onset ( 8I % trod 72~, respectively of initial level) than in the control shock group (64 % and 52% at the same time). Pretreatment of rals with soh~ble gIucan w~ts associated with beneficial effects o~ the hepatic c~ergy $Ia[tlS after 3 h after challenge of endotoxiae: the tissue level of lactale was ahnost twice lower than in the control ruts, me~mthne the tissue ATF in CMG pretreated group was higher at 40 %. Twice injected macrophage stimuhttor soluble glucan can prevent the endotoxic shock, and extremely ir~creased survival rate after endotoxine injection. The National Committee of Surgical Infections of the Spanish Association of Surgeons have produced a computer program for the collection and analysis of information on surgical infections. The program is suitable for IBM compatible hard disk personal computers and works through the MS-DOS system. The main menu is called up on the screen when the operating disk has been installed; it reads as follows: I. New record; 2. Modify records; 3. Erase records; 4. Searches; 5. Reports; 6. Configure; O. Ouit.

If you ask fdr a new record the screen will prompt you to enter the number of case, record number, hospital, age and sex. The next screen will come up and the words "Topographic diagnosis" will flash. A menu of 34 areas or organs will be displayed. Then, the words "Type of pathology" (inflammatory, neoplastic, traumatic and other). Days of postoperative period. Type of surgery (programmed and emergency). Type of operation (clean, clean contaminated, contaminated and dirty). Duration of surgery. This is followed by "Order of operation" and the "Type of anaesthesia (general, regional or local). You are then required to supply the "Diagnostic code of WHO" (ICD9) and the "Procedure code of WHO. Analytic and concurrent illnesses (total proteins, albumin, haemoglobin, haematocrit, leucocytes, red corpuscles, glucose and bilirubin). The next screen asks for "Risk factors" (obesity, uraemia, neoplasia, malnutrition, urinary catheter, distant infection, artificial valve, immunosuppressive drugs, over 65 years and anergy. This is followed by a screen headed "Postoperative complications". "Evolution" (the questions asked are drainage, systemic antibiotics, and on each ocasion a choice of 31 antibiotics is displayed), local antiseptics, reoperation, etc. Under "Microhiology" is a choice of 37 organisms and the chance of identifyin 3 organisms. Finally, "Sepsis score". Our recent work had shown that RenShen-Fuzi-ChaiHu mixture could increase the survival rate in experimented study. The purpose of this study was to determine the effect of combined administration of RenShen-Fuzi-ChaiHu mixtuer and Antibitics (SA) in patients with septic shock. The result showed that, in SA group (40 cases), the total effective rate was 87,5%, in the contral group (Combined administration of gentamycin and dexamethasone, 25 cases) the total effective rate was 84%. However the obviously effective rate in SA group 70% was significantly higher than in contral group 44% (P<O.OS). SA group appeared to be more obvious than the contral group in raising arterial pressure, recovering consciousness and pulse, improving cold lilmbs and reducing fever (P<O.O5). It was found that SA could inhibit the pathogeneses changes of septic shock such as the decreases of superotide dismutase (SOD) and glutathione peroxidase (GSH-PX) in erythrocytes and the increases of plasma free hemoglobin (PHb), plasma malondialdehyde (MDA) and 13-glucuronidase (13-GC) occured in septic shock. SA had stronger effect than GD in inhibiting the changes mentioned above. Hence, these results suggest that SA has beneficial effect in the treatment of septic shock. SA could diminish the decrease of the antioxdase activities and inhibit lipid peroxidization and stabilize cellmembrance. This may be one of the principal mechanisms of the beneficial effect of SA in the treatment of septic shock.

Donna L. Lehman, Heather A. Childers, Irshad H. Chaudry and Alfred Ayala.

The thymus is thought to be a privileged sight for Tcell generation. Here the T-lymphocytes are either selected for their potential to recognize and react competently to foreign antigens or de-selected, due to their potential to react to auto-antigens, de-selected.

Tcells with this latter capacity are thought to be deselected through a process referred to as programmed cell death or apoptosis.

While it is known that thymic apoptosis is elevated by a variety of stressers associated with infection and inflammation, little or no information is available concerning its presence in polymicrobial sepsis.

It was, therefore, the aim of this study to determine whether thymocytes exhibit increased apoptosis during sepsis.

To assess this, thymocytes were harvested from C3H/HeN mice at i and 24 h following cecal ligation and puncture (CLP; to induce sepsis) or Sham-CLP (Sham). Thymic yields were assessed and the extent of apetosis determined by staining the cells with propidium iodide (a DNA intercalating dye) for FACS analysis or by examining the extracted DNA electrophoretically for the endogenous endonuclease activity

The results (n=6/grp) indicate that at I h post-CLP there was no marked changes in any of these parameters.

However, at 24 h the thymic cell yield from CLP mice was significantly decreased (Sham:3.6!0.2 x l06 vs CLP:0.3i0.1 x 106* cells; *,p<0.05), the % of cells apoptotic by FACS analysis increased from 1.7i0.1% in Sham to 8.7±3.2%* in CLP, and the septic mouse genomic DNA exhibited DNA degradation These results indicate that CLP, but not simple laparotomy, induces marked thymic apetosis, which may contribute to the lymphocytic immune deficiency seen in these mice 

The purpose of this retrospective study was to evaluate effectiveness of irrigation to treat septic knee joints from longterm results. From 1982 to 1992 31 patients with septic knee joints have been treated. Treatment in acute cases started with asp#at[on of the effusion to evaluate for cultures. Without awaiting the result treatment was continued with immediate joint irrigation under arthroscopic control. Three redon tubes ( CH 16 ) were introduced for continuous irrigation and for intermittent distention of the joint cavity. In case of chronic joint infection additionally all foreign and synthetic material was removed. Systemic antibiotics were given. Patients could be re -examined with a mean follow-up of 6 years. Re-examination showed that immediate arthroscopic lavage started at the onset of septic sings had best results (less signs of osteoarthr[tis, less synovitis, less pain while loading and less range of motion loss ). 21 patients (80%) out of 26 patients with s tre_=.tsd acute se.~tic knee joint had excellent functional recovery, whereas all 5 patients with chronic septic knee joints showed poor results. Additionally, in 3 of these 5 patients with chronic septic knee joint after irrigation a further procedure was necessary due to recurrent septic knee joint. The concept that bacteria can translocate across the intestinal mucosa under a variety of circumstances is well established. However, due to the inherent limitations of in vivo studies, the cellular mechanisms underlying the process of bacterial translocation (BT) across the epithelial barrier are poorly understood. This is especially true in those circumstances in which there is no merphologic evidence of mucosal injury. Consequently, we have utilized an in vitro cell culture model system to investigate the cellular events involved in the translocation process. In this model system, a polarized monolayer of intestinal cells (Caco-2 cell line) is grown on a 3 micron filter in a two compartment system. After tight junction integrity is established, bacteria are placed in the apical surface chamber and BT across the Caco-2 monolayer is measured by culturing the basal chamber. The results of our studies utilizing the Caco-2 system indicates that; l) Several strains of non-pathogenic bacteria will translocate across the Caco-2 monolayer in a time-dependent and dose-dependent fashion. However, the incidence and magnitude of BT are highest for those strains of bacteria (gram-negative enterics) that are most commonly cultured in vivo.

2) Caco-2 cells are actively involved in the transcellular passage of bacteria across the Caco-2 monolayer, since inhibition of Caco-2 microtubule or microfilament function decreases the magnitude of ST.

3) Lectin but not integrin receptors are involved in bacterial translocation of E. col~ across the Caco-2 cell monolayer. 4) Caco-2 cells have the ability to ingest and kill certain strains of bacteria. The mechanisms involved in Caco-2 bactericidal activity appear similar to that of PMNs to the extent that both are oxidant but not nitric oxide mediated.

These in vitro studies suggest that Caco-2 cells can function as "nonprofessional" phagocytes and that both bacteria and enterocytes are involved in the translecation phenomenon. Bacterial translocation due to gut barrier dysfunction is considered to be a major cause of septic complications and multiple organ failure following trauma, burn injury, hypoxia, shock and shock-like states. The invasion of bacteria into the circulation from the gastrointestinal tract or even from other sources, however, should not necessarily result in systemic infection or organ cohinisation, as long as the humoral and cellular defense mechanisms are not impaired. Thus, a reduced RES clearance capacity and hepatic clearance function of bacteria and toxins can be observed under the same conditions which enable bacterial translocation from the gut. In order to evaluate the influence of circulatory, metabolic and humorul disorders on the elimination of bacteria out of the blood circulation, a series of experiments were performed in anesthetized and ventilated rabbits which received defined numbers (1.3 x 108 colony-furming units) of Escherichia coli as a bolus injection. In unaffected control animals the intravenously injected bacteria are eliminated about 99.9% within the first 15 minutes and are totaiy cleared within 90 minutes. 57% offue total CFU counted in the cultures of the organ homogenates were found in liver and spleen, whereas only very low numbers could be detected in the lungs and kidneys. Systemic injury of the animals by hemorrhagic shock, endotoxinemia, hypoxia, coagulation disorder, systemic vasoconstriction by norepinephrine and other types of injury reduced the elimination rate of bacteria and changed the pattern and degree of their dissemination and tissue distribution. The bacterial clearance was delayed mostly in rabbits subjected to hypoxemic hypoxia in which about 100 CFU of viable E.eoli per ml blood could still be counted 3 hours following their injection. The number of viable bacteria was some ten powers higher in organs of hypoxic rabbits in comparison to control ardmais, and an excessive colonisation of the lungs and kidneys with 25% respectively 11% of the total CFU of the cultared organ homoganates was observed. Unilateral iscbemia of the kidney prior to the intravenous injection of E.coli only moderately delayed the bacterial clearance, bowever, caused an intense eohinisation of the affected reperfused organ. In contrast to this, inhalation injury did not promote the accumulation of bacteria in the lungs. Conclusion: Shock, systemic and local affections reduce the clearance of bacteria from the blood circulation and enable their dissemination and accumulation in vital organs. The degree of this reduction and the organ pattern of the bacterial distribution varies with the type of injury. Thus, these factors seem to be essentially involved whether or not bacterial translocation and multiple organ failure will follow the invasion of bacteria into the circulation as in the ease of gut barrier dysfunction. Movement of enteric baaterla from gut to extralumeaal sites (bacterial tra~sloeation) m~y play a role ~. nraltipl~ orga~ failure. Traumatic stress (shock, hffectiort, and im~lammation) and severe illness are common alateeedents. Tnt~st~aal Lleus is e common proximme causal factor. We have exanlined the hypothesis that bacterial ttanslecation from the gut in expe~me~tal paaerestitis is due to bacterial overgrowth as a cuas~ucnee of a deere~ae in intestinal transit. Ne~rotiT_.~g pancreatitls fblIowlng bile duct iigation in the opossum is associated with colonization of the pa,se ~re, aa by gut derived organisms or salmonella of bilJary origin in infected animals (previously reported). To ti~rther define the mechanism of transloemdon in panereati* inflammation, pancteatitis was induced hy iigatlon of t~e lower bile duet distal to the pancreatic duets m the rat. Intestinal transit, enteric bacteriology, and tromsloeatien of bacteria to mesanterie lymph nodes, blood stream and splanelmie organs wen deletmhaed at 24 (n~ 7), 4g (n ~ 10), and 96 (n~ 10) hours with the fallowing results: lY~TESTINAI,..TRANSFI". Geometric Mean of Fluorescent Marker -25 % at 24 hour~ with return to 51% of gut leugth (similar to ~m) at 96 hour~.

BACTERIAL OVERGROWT_Hff -Increase in total bsctetia ~ad gram negative rod.~ at 2,; hours with return to sham control at 96 hours (9 Log CFU/g veraus 6 Log CFU/g ia ileum).

TRANSLOCATION TO MESENTERIC NODES -71, 90, and 30 percent at 24, 48, and 9fi hours compared to 49b in sham controls (n=l of 26).

CONCLUSION -Bacterial translocation following panetoatieobiliary duct lJgation induced panereatitls ha the rat is tighlly linked to intestinal trausit and bacterial overgrowth within the gut lumen. SPECLrJLATION -The ilens in in/~aRm~ation may be related to activation of eytokiaes and mediated by nitric oxide. Preliminary evidence for this notion will be discussed. In a series of studies we investigated: a) the time course of bacterial translocation (BT), b) the kinetics of LPS and cytokine appearance (TNF, IL-6, Soluble TNF receptor [sTNF-R] ) in the circulation, and c) the role of LPS and/or TNF with regard to haemorrhagic shock (HS) related pathophysiologie alterations and mortality in baboons and/or rats. Method: Baboons were subjected to femur fracture, soft tissue trauma (acute experiment), and HS (40 mmHg mean arterial pressure: MAP for 2-4 h terminated at an accumulated oxygen debt of 180-200 ml/kg followed by resuscitation). A chronic model of HS was set up by administration of zymosan activated plasma (ZAP) before and after HS but without trauma. HS was induced in rats by bleeding the animals to MAP of 30-35 mmHg 90-180 rain followed by resuscitation. Results and conclusion: In rats, LPS increased in the systemic circulation, plasma TNF levels were found to be significantly elevated at 90 min and were still markedly increased at 150 rain postshock. Plasma TNF, IL-6, and sTNF-R levels increased already during the shock period in baboons, while TNF was not detectable. Our data suggest that haemorrhagic shock may lead to early BT in the intestinal wall (at 30 min following resuscitation in rats subjected to HS for 90 min, similarly at 3 h after the end of oxygen debt period in baboons ) and transient access of gut-derived LPS into the circulation (levels became significant at 120 min, while in baboons higher levels up to 300 pg/ml were found at the end of shock and 1 h after reinfusion, partially accompanied by bacteremia). In addition, since the treatment of rats with anti LPS measures or anti TNF therapy significantly improved the 48-hour survival rate it can be assumed that endogenous LPS and LPS-induced mediator(s), in particular TNF, may lead to organ injury and mortality following haemorrhagic shock. Alteration of the mesenteric blood flow and the consequent ischemia / reperfusion injury to the intestine appear to be one of the earliest events in the systemic inflammatory response following severe thermal injuries. The causative relationship between the subsequent disruption of the intestinal integrity and the potentiation of the translocation of indigenous bacteria and/or endotoxins to remote body organs and the systemic circulation has been deeumented in several studies. Among other effects, endotoxemia solely or acting synergistically with thermal injuries has been shown to promote bacterial translocation. As these sequences continue without adequate intervention, multiple organ failure and eventually death may become inevitable. Hence, the crucial need of treatment modalities with the capacity of modulating the intestinal blood flow and preventing the manifestations of these pathological incidents. Although the underlying mechanisms of this process have not yet been fully elucidated, there is accumulating evidence that various interacting vasomediators are involved. The actions of these mediators can probably be modulated by either nonspecific or selective agents. Among the nonspecific agents, the composition, volume and timing of the resuscitation fluids appear to play an important role in this process. Nitropmsside, a nonspecific vasodilator, has been shown to prevent the decrease of the postbum mesentefic blood flow when selectively infused in the mesenteric artery. The nonspecificity of various vasodialators with their possible undesirable systemic effects emphasizes the importance of the identification and modulation of selective vasomNiiators. Thromboxane A2 (TxA2) and angiotensin II (A-II) appear to be the primary mediators of the postburn vasoconstriction. Both mediators were found to be elevated following thermal injuries. In a previous study, pretreatment with OKY-046, a thromboxane synthetese inhibitor was found to attenuate the postbarn mesenteric vasoconstriction. In a bum/sepsis porcine model the efficacy of DuP753, an A-11 receptor antagonist as a posthum treatment modality was evaluated. Treatment with DuP753 prevented the early posthum and the late posteodotoxin elevation in the mesenteric vascular resistance and subsequently abrogated the fall in the mesenteric blood flow. The incidence of burn & endotoxin induced bacterial translocation was significantly reduced by DuP753 from 86% to 29% (P<0.05, Fisher's exact test). The indigenous flora of the gastrointestinal tract exerts a potent influence on the developmental maturation of normal systemic immunologic responsiveness. Moreover, both oral and intraportal administration of experimental antigen is associated with a systemic state of immunologic hyporesponsiveness, known as oral tolerance. Since trauma and critical illness are associated with changes in the flora of the gut and with altered systemic immune reactivity, we have hypothesized that the interactions of an altered gut flora with immune tissues in the gut and liver play a role in the pathogenesis of the immunologic derangements of critical illness.

Prolonged feeding of two clinically relevant killed organisms-Pseudomonas and Candida-to a rat results in significant impairment of cutaneous delayed hypersensitivity (DH) reactivity. Conversely, measures directed against Gram negative bacterial overgrowth in experimental peritonitis result in partial restoration of impaired DH reactivity. To ascertain the role of the liver in the pathogenesis of these alterations, we studied the differential immunologic sequelae of portal versus systemic (IVC) infusion of killed bacteria. Experimental infusion of live or killed Gram negative organisms into the portal vein produces DH suppression in rive, and profound suppression of mitogen-driven lymphocyte proliferation in vitro; systemic administration of the same organism has no effect on these phenomena. Suppression is TGF3-dependent, and mediated by a soluble factor released by fixed tissue macrophages of the lung and spleen. The suppressive influence appears to be mediated at the level of interactions between IL-2 and its high affinity receptor. Release of this factor can be induced by a second circulating factor present in the plasma of portally-, but not systemically-infused animals two hours after bacterial administration.

These studies suggest that the release of inducible immunoregulatory factors from the liver, and possibly from the gut, may play a role in the pathogenesis of the profound derangements of systemic immune homeostasis that accompany trauma and critical illness.

Acad.Hospital St. Radboud,Postbus 9101,6500 lib Nijmegen Introduction. The central question being tested in this study was whether liposome encapsulated dichloromethyLene-diphosphonate (CI2MDP) mediated elimination of liver and splenic maorophages would influence zymosan induced systemic toxicity (ST) and bacterial lranslocation (BT). Materia|s and Methods. 90 male Swiss (ICR) mice were used weighing 21-25 g Elimination of liver and spleen macrophages was accomplished by intravenous injection of 200/11 suspension of CI2MDP-liposome suspension. Control mice received an i.v. injection with 200 pt phosphate-buffered saline (PBS) . Two days later all mice were challenged with an i.p. injection of zymosan suspended in paraffin (Z) at a dosage Of either 01 mg/g, 0.5 mg/g or 1.0 mg/g body weight (n=8 in each group), Signs of ST and bacterial translocation BT were measured 24 hours later The ST was assessed by blindly grading the degree of lethargy, conjunctivitis, diarrhea, and ruffled fur of each mouse on a four point scale. BT to the mesenteric lymph nodes (MLN) and systemic organs (liver, spleen, lung and blood) was tested by culturing on blood and MacConkey's agar At the time of sacrifice sections of the distal ileum were taken for histology. An additional two control groups (n=6 in each group) were tested to verify that neither paraffin nor PBS or CI2MDPqiposome suspension alone induced BT or ST. Furthermore survival over a 12-day period was assessed in a control and macrophage-depleted group with a dose of 1D mg/g Z. Results. Neither the paraffin nor CI2MDP-tiposome suspension induced BT or ST The incidence of bacterial translocation to the MLN was similar between macrophage depleted and control groups. However, macrophage depletion was associated with a significantly higher incidence of systemic spread of bacteria. Data expressed as positive tissues/total tissues tested: 11/32 vs. 1/32 at 0.1 mg/g Zymosan (p~O.01); 19132 vs. 10132 at 05 mg/g Zymosan (p~Q05); 23/32 vs. 11/32 st 1.0 mg/g Zymosan (p~O.O1).The magnitude of BT however did not show any differences. Although systemic BT was greater in the macrophage depleted groups, the systemic toxic response was significantly less at doses of 0.5 mg/g Zymosan (3.14 -+ 0.69 vs. 7.13 -+ 3, p~O.01) and 1.0 mg/g Zymosan (4.38 -+2.50 vs. 11.25 -+ 2.5, p<_.O.01) . The 12 day mortality after 1.0 mg/g Zymosan was 27% in the control group and 0% in the macrophage depleted group (p=O.05). Histology did not show any differences between the control and macrophage depleted groups. Conclusion. The lethal and toxic effects of zymosan in this model appear to be more related to the excessive activation of macrophages than to the systemic spread of bacteria. Bacterial translocation (BT) in hemorrhagic shock was studied using a porcine model. In this study three groups were randomized: one control (n=3) and two experimental (n=6 each). A portal vein catheter was placed and remained in situ 48 hrs. A femoral artery and vein catheter were in place 12 hrs. The arterial eaanula was used for monitoring mean arterial pressure (MAP) in all animals and to bleed the experimental animals and the venous cannula to repelfuse with Ringer's lactate (RL) or autologous blood (AB). Baseline samples and values were taken in this period. Control animals were observed for a sham shock period of 6.5 hrs., given 2 more hours of anesthesia, and studied for further 48 hrs. 40% of the estimated blood volume was withdrawn from each experimental animal over 30 min. These animals were kept in shock for 6.5 hrs. At the end of the shock one group was perfused with RL and the other with AB. Two hours later they were extubated and further studied for 48 hrs. Samples from portal blood for cultures and measurement of Leukotriene B4 (LTB4), Prostaglandin E2 (PGE2), and nitric oxide (NO) metabolites were taken from all animals at intervals beyond baseline up to 48 hrs. After 48 hrs. under general anesthesia a second laparotomy was performed and samples for culture from mesenteric lymph nodes (MIN), liver and spleen were taken; as well as samples for histologic study with light and scanning electron microscopy were taken from jejunum, ileum, and colon. Animals were then euthanized. RESULTS: significant shock was induced as indicated by MAP and arterial blood gases. Significant BT to MIN was observed in all groups. No significant BT was observed in cultures from liver and spleen in any of the groups. Increased values for LTB4, PGE2 and NO metabolites were found during the period of shock. These results suggest that the isehemic phenomenon triggers a significant inflammatory response, and that BT to MIN may be an epiphenomenon without apparent significant influence on the inflammatory response. Objective: To study the potential effects and mechanisms of action of systemic administration of monoclonal antibodies anti-endotoxin (HA-1A) in a animal model of gut-origin sepsis. Materials and Methods: Exoeriment A. BALB/c mice were transfused with allogeneic blood (from C3HB-IeJ mice). Five days post-transfusion the animals were gavaged with lxl09 viable Escherichia coil and randomized into 3 groups (n=22/group) to receive a sham burn (SB group) or a 20% thermal injury immediately followed by the systemic administration, via a tail vein, of monoclonal antibodies (6 mg/kg) (HA-1A group) or a 20% burn injury and administration of equal volume of sterile saline (C group). All 3 groups were resuscitated by intraperitoneal injection of 0.3 ml of saline. The animal survival rate was observed for 10 days post-burn. Experiment B, Transfusion and burn procedures were identical to experiment A but the mice (n=8/group) were gavaged with 109 viable E.coli labeled with mMndium oxine. Four hours after burn the mesenteric lymph nodes (MLN), liver, lungs and blood were harvested to determine plasma endotoxin levels and the magnitude of transtocation of labeled bacteria as measured by the residual radioactivity (dpm) in the organs. Circulating endotoxin levels were determined by limulus colorimetric assay. Diamine 0xidae(DAO) is an enzyme existing in the mucosa of small intestines, but its activity is low in the plasma. It is evoked in the blood with the intravenous injection of much heparin.

It is well known that DA0 declines under the chronic mueosal injury when the small intestines are cut off or the mucosae become atrophied. Aim of this paper is to confirm that DA0 goes up to the measureable level by the intravenous injection of a small amounts of low molecular heparin(LMH Objectives: The aims of this study is to investigate the inhibitor" effect of the prolease inhibitor, Urinastatin, on endotoxin induced bacterial transleeation in mice. Materials and methods:lCR mice were divided in six groups as the fotlows, Group I: Control mice receiving intraperitoneal injections (IP) of saline 24.5 and 24 hr prior to sacrifice, Group Ih Treated mice receiving IP of Utinastatin 24.5 hr and endotoxin (5mg/kg) 24hr prior to sacrifice,Group Ilk Sham4reated mice receiving IP of saline 24.5 hr and endotoxin (5mg/kg) 24hr prior to sacrifice,Group IV: saline control mice receiving IP twice during 30 rain intervals,Group V: Mice received IP of Utinasmtin, and 30 min later they were received IP of endotoxin (5mg/kg), Group VI: Mice received IP of saline,and 30 min later they received IP of endotoxin (5mg/kg).

In group l,II and Ill ,the mice were killed by cervical dislocation.Using stetile techniques,a middle-line incision was made and the mesenteric lymph nodes were harvested in order to culture on selective agars for quantitation of bacterial trauslocation. In group IV, V and VI, survival was recorded for 3 days. Unnastatin pretreatment could rednce mortality significantly.

First surgey, Shiga university of medical science, Seta ,Ohtsu 520-21,Japan

Increased gut permeability correlates with inflammatory response in multiple trauma aptients.

Pape, H.-C. Dwenger, A. Grotz, M. Regel, G.

PURPOSE: Disturbances of intestinal permeability have been advocated as a pathogenetic factor of posttranmatic multisystem organ failure (MOF). A close relationship with impairment of the stationary reticulocndothelial system (RES) and a systemic inflammatory response (SIR) was discussed. These tactors were investigated prospectively in patients with multiple trauma showing posttraumatie complications (multiple organ failure, MOF). METHODS: 31 polytranmatized patients were studied prospectively. According to a MOF-score (GORIS) those with posttrattmatic complications were selected (>7 points at 2 days), others were excluded. Every second day gut permeability according to the ratio of urine concentrations of lactulose and mannitol (L/M) was evaluated (enteral application). At parallel time points RES clearance capacity (K-value, invasion constant, normal range 0.6 -0.8 mind) was studied after i.v. injection of 100 mBq 99roTehuman albumin. Liver perfusion was calculated from these data, total serum bilirubin (/zmol/l) was documented. Serum elastase (#g/l) levels were determined enzymatieally. RESULTS 86.0 148+ 239+ Liver perfusion did not ehangu, bilirubin showed progressive worsening indicating MOF. A positive correlation was present between L/m and K (r=0.88) and between L/m and Ela (r=0.71). CONCLUSIONS: There is a positive correlation between the time pattern of intestinal permeability dysfunction and RES hyperactivity as well as between intestinal permeability and the systemic intlammatory response (elastase levels). The results speak in favor of an interaction between intestinal and extraintestinal inflammatory systems, which in eombiuation are likely to be responsible for post~anmafic complications. Endotoxemia, IL-6 release and consecutive acute phase reaction are observed as a host response to surgical trauma. As well vasodilative prostaglandins (PG) and thromboxane (TX) are released after abdominal meaenteric traction (MT). The following hypotension and acute hypoxeraJa are duo to prostacyelin (PGIz) arm can be avoided by perioperative cyclooxygenase inhibition. We therefore focused on the effect of PG and TX liberated following MT on the induction of endotoxemia. METHODS: In a prospective, randomized double-blinded protocol 50 patients, who were scheduled for major abdominal surgery (pancreatic or infrarenal abdominal surgery), were studied. Ibuprofen (400 mg i.v.) or a placebo equivalent was administered 15 minutes before skin incision. MT was applied in a uniform fashion. Baseline values were obtained before induction of anesthesia. Further measurements followed before the incision of the peri[onenm (tl) and 5, 30, 90, 180 min, . The plasma concentrations (,pc) of 6-keto-PGFt,, TXB: and-KI-12-PGF ~ (stable metabolites of PGI2, TXA2 and PGE~) were determined by RIA. We measured Endotoxin pc by Limulus-amoebocyte-lysate test and IL-6 levels by ELISA. Data are given as mean+SEM (* p< 0.05 placebo vs. [ibuprofen] ). RESULTS: Endotoxin plasma levels increased before incision of the peritoneum tl both in the ibuprofen pretreated and in the placebo group. Peak pc were observed 180 minutes after MT. Endotoxin pc were significantly higher in the ibuprufen treated group (t5 0.16+0.02e[0.32+0.06] EU/ml). IL-6 pc demonstrated an increase continuously from t4 to t7 (t7 115+12 [91 + 15] ng/l) in both groups. After intentional abdominal mesenteric traction we observed a marked increase of 6-keto-PGF~,, pc up to 6h after MT in untreated patients with a peak of 2450*[60] ng/1 at tl. Also TXB: and KH2PGE 2 pc showed a considerabe increase up to 6h after MT in the placebo group. In ibuprofen pretreated patients the PG and TX pc remained within the normal range. DISCUSSION: Our data clearly indicate a significant endotoxemia and IL-6 release following major surgical trauma which is not initiated either by prostaglandin or thromboxane release. Moreover endotoxemia is accentuated by ibuprofen pretreatment. Therefore we hypothesize that in major abdominal surgery prostacyclin release-after MT may play a crucial physiological role in maintaining splanclmic microcirculation and thus preserving gut mucosal barrier function. Objectives of the study It has been shown recently that parenteral and certain euteral diets promote the translocation of gut flora to the mesenteric lymph nodes (MLN) and systemic organs, a process termed bacterial translocation (BT). In chow fed rats BT usually does not occur without further promoting factors. The goals of the present study were to determine whether the provision of defined amounts of standard lab chow during IV-TPN administration wotfld redane the incidence of BT, Materials und methods Male SPF Spragnle-Dawley rats were divided into 6 groups. Group 1 received standard laboratory chow feeding ad lib. In group 2 a central venous catheter was placed, ligated and secured by a spring coil tether attached to a swivel allowing free movement in the housing cage and chow was fed ad lib. In group 3 50% of the calculated daily required calory intake (DRCI) (307/Kcal/kg) was given by IV-TPN (28% glucose, 4,25% amino acids) and 50% by limited chow administration. Groups 4 and 5 received 70% and 90% of the DRCI by i.v. TPN and 30% and 10% respectively by chow feeding. Group 6 received IV-TPN only. After 7 days the rats were sacrificed and the MLN, liver, spleen and cecum removed aseptically, homogenized and cultured for BT Samples of distal ileum were taken for light microscopy. The group with the least amount of chow shown to be protective against BT received the amount of non-fermentable fiber of that chow regimen during IV-TPN feeding and BT was studied. 4 5,5+0,9 4 12 23,9 -2,3 16 ,7 2/12 63+11 125~1 "7,7-+0,7 6,7-+1 5 12 24~6 -3,5 67 8/12 241+~243 170+163 8_+0,7 7+1,1 6 18 25, 5 -4 88,8 16118 2061-+3224 2211+4015 8,4~0,7 8,1-+1~1 Conclusions: The administration of 30% of DRCI by chow feeding during IV-TPN significantly reduced the incidence of BT and maintained gut barrier function. The addition of the respective amount of dietary fiber of this group did not prevent IV-TPN-indueed BT. Dr. med. M Naruhn., Dep. of General Surgery, Eberhard-Karls-University, Hoppe-Seyler-Str. 3 Previous experimental studies have suggested that a disturbed ecology of the enteric bacterial population might contribute to the development of bacterial translocation from the gut in acute liver failure (ALF). In the present study, the effect of oral administration of Lactobacillus reuteri R2LC and oat fiber on bacterial overgrowth and translocation was investigated in rats with acute liver failure induced by subtotal (90 %) liver resection.

The oatmeal soup base was anaerobically inoculated with lactobacillae and fermented for 15 hours, after which the animals were fed with either fermented or unfermented oatmeal or saline daily for 6 days prior to the operation. Bacterial translocation to mesenteric lymph nodes (MLN) and the systemic circulation was determined, as well as the intestinal bacterial flora and enterocyte protein content.

The incidence of bacterial translocstion to the systemic circulation was nit in rats subjected to sham operation and saline treatment and 17 % in animals subjected to 90 % bepatectomy and lreatment with fermented oatmeal, while 80-90 % and 34-50 %, respectively, in rats subjected to hepatectomy and treatment with either saline or unfermented oatmeal. Only one rat with fermented oatmeal demonstrated bacterial growth in MLN (p < 0.05 vs hepatectomy and treatment with saline or unfermented oatmeal). The enterocyte protein content significantly decreased (p < 0.01) in salinetreated animals following 90 % hepatectomy, while there was no significant difference between bepatectomized animals with oral administration of fermented or unfermented oatmeal. The number of anaerobic bacteria, Gram-negative anaerobes and Lactobacillus significantly decreased and the number of E.cnli increased in the distal small intestine and colon in hepatectomized animals with enteral saline or unfermented oatmeal as compared with animals subjected to sham operation or bepatectomy with fermented oatmeal.

Our results thus show that the occurrence of bacterial translocatiou from the gut in 90 % hepatectomy-induced ALF could be prevented by enteral administration of fermented oatmeal, maybe partly due to a positive effect on the enteric bacterial ecology. 190_+20" 16+_3" 11.7" Data=Mean_+SD, * stats ANOVA P<0.05 vs control. L+air and lap groups, both exposed to exogenous I.PS shnwm:t m significant increase (p<.05) in LPS gut translocation compared to control and L+CO2. This correlated with a significant increase in peritoneal inflammatory responses (O2-,TNF) above that of the control and L+CO2 groups, while Mac-1 and CR3 opsonized phagocytosis were significantly impaired. The absence of significant differences between L+air and lap groups indicates that LPS rather than wound factors is the principle mediator. Thus, LPS plays a significant role in regulating peritoneal responses in the early post-operative period Dept of Surgery, RCSI, Beaumont Hospital, Dublin 9, Ireland

Brlke E, Berger D, Staneseu A, ButtenschSn K, Vasilescu C, Seidelmann M, Beger HG In 32 patients undergoing a colonoscopy, endotoxin, endotoxin neutralizing capacity (ENC), thromboxane B o (stabile metabolite of tbmomboxane ~), 6-Keto-prostaglanSin, leueotriene C4, interleukin 6 and the incidence of bacteremia were determined before and then every five minutes during the procedure. Twenty-one of 32 patients showed a significant increase of endotoxin plasma levels during colonoscopy (p=0.003), whereas only one patient had a positive blood culture with bacteria obviously derived from the gastro-intestinal tract. The ENC decreased significantly five minutes after the beginning of eolonoscopy and was diminished further thereafter. The baseline values were reached after 24 hours. ~hromboxane B o levels also increased after five min. from 89 to 376 pgYml peaking at 30 min. with 1332 pg/ml. 6-Keto-prostaglandin,leucotriene C 4, Ii-6 and CRP remained unchanged. A control group of i0 volunteers who were not subjected to endoscopy, were prepared for eolonoscopy by orthograde lavage. The blood sampling procedure remained identical. No differences were seen in all described parameters for the controls. These data show that the gut barrier can be compromised by mininml invasive procedures, at least, concerning bacterial products. Living bacteria, on the contrary, do not pass the gastro-intestinal wall. Endotoxin, when determined by ENC, is more sensitive than the conventional limulus-amebocyte-lysate test. No acute-phase reaction was induceri by the observed endotoxin translocation. It can be speculated from the dramatically enhanced thromboxane B 2 levels, together with its hemodynamie effects, that the thromboxane release may support translocation of bacterial products. Sepsis is common after hemorrhagic shock. This study aims to demonstrate that hemorrhagic shock alone can promote translocation of gut bacteria from intestinal tract to its regional nodes and subsequently to blood. One hundred twenty mice, divided into 4 groups were subjected to 30, 60 and 120 minutes of 20%, 30% and 40% of hemorrhagic shock. On the specified time, blood cultures were taken and mice were sacrificed.

The intestinal tract were histologically examined for any changes which allows translocation and its regional nodes were quantitatively cultured for translocated bacteria.

There was a direct relationship between duration and degree of hemorrhagic shock and incidence of translocation (p 0.05). There was a high incidence of gut bacterial translocation to the mesenteric and mesocolic nodes in all degrees of shock (p 0.05). Bacterial growth in the regional intestinal nodes increased and blood cultures were positive in direct proportion to degree and duration of shock. Histologic evaluation of segments of GIT showed submucosal congestion to allow bacteria normally contained within the gut to cause systemic infections.

Translocation of gut bacteria in untreated hemorrhagic shock is clearly shown in this study on animal models. In this study, guotobiotic rats with 5 known species of bacteria were subjected to total parenteral nutrition(TPN) and subsequent hemorrhagic shock. The purpose of the study was to observe the impairment of gut barrier function following TPN and hemorrhagic shock and to study the mechanism of enterogenic infection induced by TPN and shock.The results were as follows:

1.Long term(8-12 days) TPN induced impairment of gut barrier function, evidenced by atrophy of intestinal mucosa, significant decrease in diamine oxidase activity of intestinal mucosa and blood, and marked microecologic imbalance of the intestinal mucosa flora with dorminant growth of aerobes and relative decrease in anaerobes. The degree of mucosal damage were proportional to the duration of TPN.

2.In TPN+shock groups, failure of gut barrier function was found.

ri,~ere were further damage in the mucosa, with a large number of gramnegative organisms invading mucosa and submucosa and a significant decrease in DAO activity as compared with each relative TPN groups. These changes were significantly correlated with enhanced bacterial translocation, elevation of LPS and MDA levels in the plasma.

These findings suggested that long term standard TPN impaired the gut barrier function, precipitating posttraumatic gut barrier failure. Thus infec.

fion following shock might be oi'iginated from the gut and it was obviously related to the impaired gut defence resulted from antecedent TPN. The determination of plasma DAO activity might provide a valuable tool for the ear.

ly diagnosis of gut injut;y during TPN and after trauma. In our earlier studies we have investigated the dynamics of granuloayte infiltration of the ischemic/reperfused s~all intestine (G. Illy~s, J. Hamar Int. J. Exp. 9athol. 73. 161. 1992.) . There was a increasing infiltration of the mucosa c111m~nating at the 3d to 4th hours of reperfusion. In the present series we have studied sc~e of the conseqn/ences and the possible role of this cellular reaction. 45 ~in isehemia was followed by a 4 hour reperfusion in the anesthetized rat. Arterial ~/ad mesenteric venous blood samples were collected at 5 m_in, I, 2~ 3, and 4 hours of reperfusion. Elastase and lactate concentrations were determined and hamoculture was carried out from the blood samples, and tissue pieces from the heart, lung, liver and kidney were collected for histological analyses at the above mentioned times of reperfusion. All blood samples were free of Cell bacteria. Staphylococci appeared only occasionally at the 4th hour in the arterial blood .and at the 3d and 4th hours in the venous blood, respectively. Arterial and venous elastase activities were high throughout the reperfusion, venous concentrations being higher at all times. Lactate concentrations of the arterial and mesenteric venous blood samples increased during shock. ~ranuloeyte infiltration of all organs studied appeared during the 2d hour and it increased at later times of reperfusion. It is concluded that heavy infiltration of the intestinal mucosa can block bacterial translocation in most of the cases during reperfusion. Granulocytes activated either by the reperfused area or by the released cytokines infiltrate other organs contributing by this way to the mesenteric shock s!rndrc~e. Intestinal motility plays an important role for maintaining nutrient transport and absorption and for balancing the enteric bacterial population. Disturbances of intestinal motility may be one of the earliest notable changes in intestinal function. In the present study, we aimed at determining early alterations in intestinal transit time following ischemia-reperfusion injury induced by occlusion of the superior mesenteric artery in the rat.

Intestinal ischemia was induced for 20 and 30 minutes by applying a microvascular clip on the superior mesenteric artery followed by reperfusion 2, 4 and 6 hours after clip removal. Intestinal transit time was measured by the propulsion of a radiolabelled solution (Cr51). Light microscopy was performed on intestinal samples.

Macroscopical pathological changes were not observed. However, microscopically, mucosal epithelial oedema, degeneration or slight ulceration occurred in rats 6 hours after reperfusion in ischemia-20 rain group and 4 and 6 hours after reperfusion in the ischemia-40 rain group. Delayed small intestinal transit time was seen from 2 hours and on after intestinal ischemia for both 20 and 40 rain ischemia followed by reperfusion. The distribution of radioactivity demonstrated that most radioactivity was accumulated in the first two segments following intestinal ischemia and reperfusion, significantly differing from what was seen in animals subjected to sham operation (p < 0.01). The distribution of radioactivity in segments 4 and 5 in the group with repeffusion 6 hours after intestinal iscbemia for 20 rain was significantly higher than that noted in the group with repeffusion 6 hours after intestinal ischemia for 40 min (p < 0.01).

q'he results indicate that a delayed intestinal transit time may be one of the earliest pathophysiological alterations noted, associated with duration of gut ischemia, and a potential factor for the development of bacterial overgrowth, gut barrier failure and bacterial translocation, in hypovolemic conditions. Bacterial infections still constitute a major cause of morbidity and mortality in patients with acute liver failure. The present study aimed at evaluating the effect of ethylhydroxyethyl cellulose (EHEC) on bacterial translocation following surgically induced acute liver failure.

Acute liver failure was induced by subtotal hepatectomy (90 %) in the rat. Water-soluble EHEC was administered orally 1 and 12 hours prior to hepatectomy. The incidence of bacterial translocation from the gut to mesenteric lymph nodes (MLNs) and systemic and portal circulation was evaluated and the number of isolated bacteria from these samples and from intestinal content were determined. Intestinal transit time, bacterial adherence onto the intestinal surface, intestinal mucosal mass, bacterial growth and DNA synthesis, bacterial surface characteristics (hydrobiology: hydrophobicity, hydrophilicity and neutrality; surface charges: positive, negative and neutral) were also determined.

Hepatectomized animals showed a 80-100 % translocation rate to MLNs or blood 2 and 4 hours after operation, while only 0-1% of rats subjected to sham operation or animals with 90 % hepatectomy and pre-treatment with EHEC (p < 0.01). Bacterial overgrowth, increased bacterial adherence onto the intestinal surface as well as decreased intestinal mucosal masses were observed in animals with subtotal liver resection alone, alterations that were prevented by enteral EHEC treatment. A delay in intestinal 2-hour transit time occurred in both groups with subtotal liver resection, with or without enteral EHEC. EHEC inhibited bacterial growth and DNA synthesis, and altered bacterial surface properties following 1hour incubation with bacteria.

In conclusion, the findings in the present study imply that EHEC alters enterobacterial capacities for metabolism, proliferation and invasion by effects on e.g. bacterial surface characteristics. Furthermore, EHEC seems to possess a trophic action on the intestine, rather than exerting its effect by enhancing intestinal motility.

Department of Surgery, Lund University Hospital, S-221 85 Lund, Sweden Disturbances in intracellular calcium signalling can potentially result in impairments of cellular responses vital to the functional integrity of both immune and non-immune cells, and thus contribute to a decrease in host resistance against infection and to multiple organ system failure during sepsis. Studies in our laboratory have focused on assessments of intracellular Ca 2÷ regulation and Ca~+-depended cellular responses in the liver, skeletal muscle and splenic Tlymphocytes harvested from rats subjected to gram-negative intraabdominal sepsis. Cytosolic Ca 2+ concentration, [Ca2*]i, and Ca 2+ fluxes were measured by the use of fluorescent Ca 2+ chelating dyes (Fura-2 or indo-1) and 45 Ca respectively. To assess sepsis-related changes in Ca 2+ dependent cellular responses, we measured the acute phase protein response in the liver, the regulation of protein and sugar metabolism in the skeletal muscle, and the proliferation response in the splenic Tlymphocytes.

Altered Ca2+ i signalling with sepsis was correlated with an exaggerated inappropriate acute phase protein response (100% ¢) in the liver, and a blunted insulin mediated sugar utilization (60% 4) and increased proteolysis (50% ~) in the skeletal muscle. In T-lymphocytes, a decrease in mitogen induced elevation of [Ca2+]i by 35-40% was correlated with a significant depression in their proliferative capacity. These studies clearly suggest that altered calcium signalling is correlated with disturbances in cellular responses in both immune and non-immune cells during sepsis. The altered cellular responses adversely effect the outcome of the septic injury. (Supported by NIH Grant GM 32288).

Alfred Ayala, Ping Wang and Irshad H. Chaudry.

Changes in macrophage capacity to respond to foreign pathogens are thought to be central to the developing immunosuppression associated with traumatic injury.

In this respect, the suppression seen in M~ functions following HEN (a common component of traumatic injury) may be mediated by the direct or indirect inhibition of their capacity to perceive external stimuli (e.g., opsonized & non-opsonized bacteria, and their cellular components, etc.} due to the breakdown of the receptormediated signal transduction system.

Results of a number of studies by our laboratory and others indicate that this inability to respond to external stimuli is in part due to the loss of cell surface receptors.

Decreases have been documented for not only la antigen, but also C3b, Fc, and TNF receptors following HEM in mice.

Furthermore, studies which have examined second messenger generation in these cells indicate that M~ derived from the peritoneum and spleen exhibit a decreased capacity to mobilize Ca +2 from intracellular stores.

This protein kinase dependent process of [Ca+2] i mobilization appears to be linked to the inability to synthesize inositol triphosphate.

Of interest, the depression in Ca +2 signal generation appears to be inversely related to presence of elevated levels of cAMP in M~ from HEN mice.

We have reported that M~ priming agents, such as IFN-7 (which exhibits salutary effects on M~ function following HEM), appear to restore cell signal transductive capacity while reducing the levels of cAMP.

Nonetheless, the extent to which depressed receptor signal transduction in HEM, is due to receptor loss~dysfunction or elevated antagonistic second messenger levels remains to be determined. Conclusions: Significant impairment of calcium signaling occurs at all time-points prior to and following PHA stimulation in trauma patients. Tcell activation failure can, in part, be explained by the inadequacy of this essential intracellular second messenger system. Restoration of immunocompetence following trauma will have to address strategies to better assess and restore this vital step in the activation sequence leading to proliferation during the antigen recognition process.

Patrick A. Bseuerle Institute 01 Biochemistry, Albert-Ludwigs-University, Hermann-Herder-Str. 7, D-79104 Freiburg, Germany

The active form of the transcriptional activator NF-~B is a heteredimer composed of a 50 and 65 kDa polypeptide. In this form, NF-'<B can bind with high affinity to decameric sequence motifs (consensus: 5"-GGGPuNNF'yPyCC-3") found in enhancers and promoters of many genes activated upon a wide variety of pathogenic conditions, including viral and bactedaI infections, acute phase response, oxidative stress and UV and gamma irradation. Clinically important target genes for NF-K:B encode inflammatory cytokines (IL-1B, TNF, [L-6, IL-8, GM-CSF etc), cell adhesion molecules (ICAM-1, ELAM-1, VCAM-1), acute phase proteins and immunoregulatory ceil surface receptors. In most cell types, NF-~:B is sequestered in the cytoplasm and, therefore, unable to initiate transcription of these genes. The cytoplasmic form of NF-~:B contains one additionaI subunit, referred to as h,;:B. I~B can reversibly suppress DNA binding and nuclear uptake of NF-~B by virtue of its association with p65. Upon stimulation of cells, IKB is proteolytically destroyed, allowing NF-KB to enter the nucleus and activate genes upon binding to transcriptional enhancers.

We have demonstrated that various antioxidative compounds inhibit the activation of NF-~B in response to many inducers. Moreover, NF-~:B was shown to be activated upon treatment of cell cultures with p.M-amounts of hydrogen peroxide. These data suggested that NF-~B prinicipaliy is an oxidative stress-responsive transcnption factor. The notion is supported by numerous reports on the induction of oxidative stress by many NF-nB-inducing agents. NF-~B activition can likewise be suppressed by protease inhibitors. These drugs apparently inhibit a yet unidentified protease responsible for the inducible decay of InB. We propose to use NF-nB inhibifors as novel drugs with a very broad application under acute phases el inflammation, injury and infection. Northern Blot experiments revealed that expression of the recently discovered Lipopolysaccharide Binding Protein (LBP), a 58/60 kD Glycoprotein, in the liver rises substantially during the acute phase response. Experiments with an in-vivo acute phase model using New-Zealand-White rabbits and also in-vitro experiments employing stimulated Hep-G2 cells showed that LBP transcripts could be induced 10 to 100 ford within 24 hours after induction with cytokines. By using a newly established nonradioactive nuclear run-on technique we show that induction of LBP during the acute phase is transcriptionally regulated. Furthermore we screened a human genomic library and were able to clone the LBP gone, containing the promoter 1.5 kB upstream. Several liver-specific and "acute-phase-typical" potential binding sites were found fn the promoter region and reporter gone assays were set up. Several CAAT-boxes, the IL-6 responsive elements HSTF-hsp 70.5 and H-APF-1 RS as well as the transcription factor HNF-5 and the glucocrticoid-responsive elements Oct-2-D and GCRE were found, which correlates with the induction pattern of LBP mRNA in Hep-G2 cells by IL-6, IL-1 and Dexamethasone. PCR-based analysis of the exon-intron pattern of the LBP gene revealed similarities with the genes of two other LPS binding proteins, namely BPI and CETP. Further analysis of this novel acute phase protein, that also has an important role in Endotoxin recognition and immunomodulation will help in understanding the complex acute phase mechanism and potentially lead the way to new therapeutic strategies.

LPS ACTIVATION OF NUCLE?/R FA-CTOR:~B " (NF:~BJ"IN  CD14.TRANSFECTED FI'BROBLASTS DOES NOT APPEAR  TO REQUIRE TYROSINE KINASE ACTIVITY  D. T. Golenbock, R. DeLude, R. Savedra, S. Vogel and M. J. Fenton Lipopolysaccharide (LPS) is the major constituent of the outer leaflet of Gram-negative bacteria. During the course of serious infection, shock may occur as a result of the interaction of LPS with macrophage LPS receptors. CD14, a 55 kDa glycosyl phosphatidylinositol (GPI)-linked protein, functions as an LPS receptor. A critical issue in LPS activation concerns the mechanism by which CD14, which has no transmernbrane domain, transduces a sig,nal after LPS binding. Evidence exists which suggests that CD14 may signal cells after LPS binding by interacting with an LPS signal transducer with tyrosine kinase (TK) activity.

Wild-type Chinese Hamster Ovary fibroblasts (CHO) normally do not respond to LPS, but can be activated following transfection with CD14 (CHO/CD14). Stimulation of CHO/CDI4 with as little as 1 ng of LPS/ml resulted in the release of 3H-arachidonic acid (3H-20:4) into the culture supernatant. In addition to 3H-20:4 release, we examined LPS-imiuced activation (transl0cation) of the transcription factor NF-~B. Similar to LPS-induced 3H-20:4 release, these responses were observed only in CHO cells lransfected with CD14 closely resembling the same response in the marine macrophage cell line RAW 264.7. Maximum NF-~cB expression occurred at 30 minutes. In contrast to LPS-induced 3H-20:4 release, dexamethasone, cycloheximide, and actinomycin D had no effect on activation of NF-r,.B in CHO/CD14, suggesting that NF-r,B activation begins early after LPS binding to CD14 and does not require transcription or translation. We then examined cells for LPS-induced TK activity by Western blotting cellular ly~ates with anti-phcsphot-'rosin:~ anl',bodie~" Although TK activity was seen in LPS-stimulated RAW cells and phorbol ester stimulated CHO/CD14, LPS-induced TK activity in CHO/CD14 was not observed. Although the TK inhibitor herbimycin inhibited the release of 3H-20:4 in CHO/CD14 and RAW cells, neither herbimycin nor genestein effectively blocked NF-r,B activation in either cell type.

These results suggest that TK activity is involved in a portion of the signaling pathway that is distal to initial signal transduction. The absence of observable LPS-induced phosphotyrosine residues in CHO/CD14 cells as well as the failure of TK antagonists to inhibit LPS-induced NF-~cB activation suggest that the proposed CD14-related LPS signal transducer in CHO/CD14 is not a tymsine ldnase. Dimished cardiac inotropic function and response to 13-adrenergic receptor (13-AR) stimulation are frequently seen in septic patients, These cardiac defects are also seen in animal models of entoxemia. To determine the characteristics of the decreased transmembrane signal associated with the decreased beta adrenergic response we measured the force and rate of contraction of atria harvested from Sprague-Dawley rats exposed to endotoxin (0:2 gm/kg). To stimulate various components of the transmembrane signal cascade of 15-AR was isoproterenol (I3-AR), acetylcholine (M2ACH receptor), NaF(Gs and Gi proteins), forskolin (adenylate cyclase), and calcium (intracellular contracttile signal) were used. To eliminate the transmembrane control of calcium and test the intrinsic contractile response, hypothermia (22°C) was used The results showed alterations in the inotropic function with no significant difference in chronotropic response. The inotropic reponse for the endotoxin exposed atria and controls is shown below These results show that there is a transmembrane signal defect both for the I~-AR signal and for Ca +. These data indicate that the level of the defect appears to be at the G proteins. Ischemia-reperfusion of the rat intestine produces an injmry both to the intestine and to distal organs, notably the lungs and liver. Prior studies have shown that 1 h of intestinal ischemia leads to a nemtrophil independent reperfusion injury of the gut. Thus, rats rendered neutropenic by Pl~l antiserum have been shown to express a severe local gut injury when subjected to ischemia and reperfusion. This gut injury is postulated to he mediated at least in part by the complement system. The soluble (s) CRI receptor which binds to C3b and Cdb was given to rats at a dose of 12 mg/kg by intravenous bolus, 15 min before reperfusion of the ischemic gut. A control group was given PBS buffer. At 4 h of rmperfmsion the animals were sacrificed. PBS treatment led to a significant activation of both classical and alternate complement pathways while sCRI inhibited both pathways (fall to zero of CHbo ). Intestinal mucosal injury score, assessed by histological grading was reduced by sCRI (2.49±0.22 to 1.73_+0.17,p<0.05) . Intestinal neutrophil sequestration estimated by assay of myeleperoxidase was also reduced (0.05 ± 0.01 to 0.02 ± 0.01 U/g). C3 was detected in the gut hy i~unohistochemistry. Remote organ injury was modified that is lung permeability (ration of concentration of radiolabelled albumin in broncho-alveolar lavage fluid to that in blood) was reduced (11.3±1.4 x i0 3 to 3.9±0.8 x 103, p<O.05). However, neutrophil sequestration by the lungs was unaffected. This remote protection is thougth to be related to prevention of iC3b deposition on Imng andotheli~ which activates adherent neutrophils. These findings support the hypothesis of complement mediated local and remote injury. We postulate that complement fragments may be deposited either as a result of the ischemia induced loss of membrane bound complement regulatory proteins, such as decay accelerating factor or membrane cofactor protein, which allow complement fragments to accumulate on the endothelial cell surface. Alternatively, increased expression of P-selectin on the endothelial cell surface may, by means of its complement binding domains, act as a lattice for complement deposition. Indeed, P-selectin has been detected on the surface of endothelial cells ans platelet-neutrophil aggregates recovered from the intestine of these animals. Ischemia (4hr) followed by reperfusion (dhr) of rat hind limbs resuits in local (muscle) injury as well as in remote (lung) injury, which is characterized by increased vascular permeability (leakage of J2Sl labeled albumin) and neutrophil accumulation (tissue content of myeloperoxidase, MPO). Within 60 minutes of reperfusion following ischemia, tumor necrosis factor-o~ (TNFc 0, interleukin-I (1L-I) and IL-6 plasma levels increased significantly, reaching maximum levels after 2 hours ofreperfusion. Polyclonal antibodies to TNFet and 1L-I provided significant protection against muscle and lung injury. Muscle and lung vascular permeability decreased in anti-TNFct treated rats by 41% and 74% respectively and MPO was reduced by 72% and 83%. AntML-1 yielded a protection in muscle and lung vascular permeability of 24.7°,4 and 52.4% respectively whereas muscle MPO was reduced by 46.4% and lung MPO by 25.9°,4. These results were confirmed by the use of soluble TNF~ receptor and IL-I receptor antagonist (IL-IRa). When ICAM-I expression in vivo was examined using 125Mabeled antMCAM-I, constititive expression of ICAM-1 was evident only in lung but not in muscle. During reperfusion ICAM-1 expression increased by 98.2%. Treatment with anti-TNFcx or IL-IRa reduced ICAM-I upregulation by 73.9°,4 and 74.8% respectively. Frozen sections of normal rat lung revealed no evidence of E-selectin expression, whereas lung sections obtained after ischemia and reperfusion immunhistochemically demonstrated expression of E-sdectin. Lungs from rats pretreated with anti-TNFct were negative for E-selectin staining. Our data indicate a role for cytokines in the upregulation of adhesion molecules in ischemia/reperfusion injury.

Adherence of neutrophils to the coronary endothelium is associated with significant myocardial injury following 90 min of ischemia (MI) and 4.5 h of reperfusion (R). P-selectin is expressed by activated endothelial ceils and platelets within 10 min and tethers polymorphonuclear leukocytes (PMNs) to the endothelium. We examined the cardioprotective effects of a monoclonal antibody (MAb) designated PB1.3 to P-selectin in a feline model of MI+R. PB 1.3 (1 mg/kg) administered 80 rain post-occlusion (10 min prior to reperfusion) significantly attenuated myocardial necrosis compared to a non-blocking MAb (NBP1.6) for P-selectin (15 + 3 vs 33 + 5% of area-at-risk, p<0.01). Endothelial dependent relaxation to acetylcholine was also significantly preserved in ischemic-reperfused coronary arteries isolated from cats treated with PB 1.3 compared to NBP1.6 (67 + 6 vs 11 + 3, p<0.01). Furthermore, PMN adherence to ischemic-reperfused coronary endothelium was significantly attenuated in PB1.3 treated cats (64 + 7 vs 136 + 12 PMNs/mm z, p<0.01). Also, normal coronary endothelium activated with thrombin (2 U/mt) had increased PMN adherence under conditions of shear stress, an effect which was significantly (p<0.01) blocked by PB1.3, but not by NBP1.6. Furthermore, P-selectin is markedly upregulated 10-20 min post-reperfusion in coronary arterial and venous endothelinm. Similar results were obtained in a rat model of mesenteric ischemia/reperfusion, including upregulation of P-selectin on mesenteric vascular endothelium. These results demonstrate that tethering of neutrophils to endethelium by P-selectin is an important early consequence of reperfusien injury, and a specific monoclonal antibody to P-selectin exerts significant endothelial preservation and cardioprotection in ischemia and reperfusion states. As recent interest has increased in the role of extracellular matrix proteins in inflammation, we examined the role and relationship of laminin and fibronectin to the processes of cell infiltration in a rat model of heart isografts. Changes occurring in this system are on a nonimmunological basis and may be an effect of the initial surgical manipulation and the warm and cold ischemic times of transplantation. Lewis (LEW) donor hearts were transplanted into LEW recipients, naive LEW acted as naive controls (NC). Grafts were harvested 2, 4, 8, 16, 24, 48, 72, 120 and 168h after transplantation (n=6/time point) . Snap frozen sections of organs were stained for CD-4 (W3/25), CD-8 (OX8), T-lymphecytes (TL) (OX-19), macrophages (ED1, ED2), neutrophils (PMN) (MOM), ICAM-I (IAP29), laminin and fibrinogen. Results were evaluated visually by counting cells/field of view (CF) for W3/25, OX8, OX19, ED1, ED2 and MOM, staining for ICAM1, Lain and Fib was assessed visually using a scale (VE=0-10). As early as 2h after transplantation, fibrinogen and laminin expression increased in I, peaking 8h after transplantation (VE= 4.5; NC: VE=I.5) . At 16 and 24h in I, on both vessels and interstitial cells laminin and fibrinogen expression had declined to baseline (VE=I.5) but rose again at 48h, peaking at 72h (VE=4) and remaining elevated thereafter (VE=3). The high expression at 8h correlated well with the maximal PMN infiltration at that time (CF=6) in I returning to baseline at 24h (CF=0.5) and thereafter. The second peak at 72h correlated with the onset of macrophage and T-lymphocyte infiltration. Thus, laminin and fibrinogen seem to guide leucocyte infiltration following graft ischemia during the time of reperfusion. Background. Reperfusion of ischemic kidneys with xenogeneic blood leads to activation of endothelial cells and circulating leucocytes with the final consequence of microvascular thrombosis. In concordant systems, the mechanisms of rejection can be studied due to cross-reactivity of mediators with anti-human monoclona~ antibodies. Aim of the study. To obtain information about the kinetics of proinflammatory cytokines and production of soluble adhesion molecules in the acute phase of reperfusion, eight kidneys from rhesus monkeys were perfused ex-vivo with human blood (group B/O) for one hour in a closed system. Blood levels of IL-lb, IL-6, TNF-a, soluble ICAM and E-Selectin were measured in ELISA-technique under steady-state conditions. Results. Cytokine levels rose significantly within the 60minute interval (IL-lb: 6,1__+2,6 to 161,1__+98,5 pg/ml; IL-6:30.24-7,7 to 274,2__+75,8 pg/ml; TNF-a: 544,2__+363,6 to 1651,0+25,7 pg/ml; p<0.05). Immediately after the beginning of reperfusion, soluble ICAM-1 and Selectin levels were abnormally high and constantly rising throughout the observation period, reaching significance at 60 minutes. Discussion. High levels of proinflammatory cytokines may lead to an induction of adhesion molecules, thus upregulating the leukocyte-endothelial interaction in an complement-independent mechanism. Specific pretreatment with monoclonal antibodies against ICAM-1, LFA-1 or other soluble mediators may be useful in downregulating hyperacute rejection in transspecies transplantation. Prolonged ischemia has been associated with later dysfunction of solid organs as it may cause upregulation of adhesion molecules, production of cytokines on vascular endothelium and migration of host cells into the graft. These early events may lead to irreversible organ injury. This study evaluates the effects of global ischemia on early immunohistological changes in cardiac grafts transplanted in rats. Isografted hearts (Lewis->Lewis) were were divided into ischemic and non-ischemic groups (n=30/group). All donor hearts were flushed immediately with cold saline. Non-ischemic hearts were then transplanted within 30 rain, ischemic hearts were stored in cold Ringer's solution for 3 hours before revascularization. Representative grafts were removed after 12. 24, 72hrs, 10 and 100 days, and evaluated immunohistologically (cells/field of view=C/F). Restitution of ventricular activity was significantly delayed in ischemic grafts (5 vs 1 rain). After 12 hrs, all ischemic grafts exhibited an extensive interstitial edema, declining slowly thereafter. At the same time, numbers of PMN peaked (3 vs 1 C/F in non-ischemic grafts), whereas EDl+macrophages (9 vs 2 C/F) and TNFe expression peaked by 24hrs. By 72 hrs T-lymphocytes began to enter ischemic myocardium and ICAM-1 was moderately increased. After 10 days cellular infiltration had returned to baseline, and no differences were seen among both groups after 100 days. Global myocardial ischemia inhibits initial graft function, and engenders a brisk inflammatory reponse, primarily PMN and macrophages, with increased MHC Class II and cytokine expression.

Leukocyte -endothelial interactions are the result of endothelial activation, leukocyte activation or combination of both, which are accompanied by nee-expression, upregulation or shedding of adhesion molecules (Selectins, Inlegrins). Such interactions differ with regard to the stimulus (e.g. thrombin or histamine for P-selectin, endotoxin or TNF/IL-1 for E-selectin), the time course of response (minutes versus hours) and the localisation in different organs. Recently assays are available for circulating soluble fragments of the cell bound adhesion molecules e.g. sE-seleetin was found to be increased in plasma concurrent with high circulating endoloxin and cytokine levels. The importance of adhesion molecules for the sepsis event is evident, while effectiveness of anti-adhesion inolecu]e therapy is controversial e.g. beneficial anti-E-selectin therapy in baboon bacleremia but deleterious effects of amti-CD18 treatment in the same model. In other species similar controversial results with anti-CD18 therapy in sepsis were reported.

Steven L. Kunkel,Theodore Standiford* and Robert M. Stricter. The migration of leukocytes to the lung during endotoxemia is dependent upon the coordinated expression of lung vascular adhesion molecules and the subsequent production of appropriate leukocyte chemotactic proteins. In experimental animals, neutrophils accumulate within the lung soon after the administration of endotoxin, while mononuclear cell infiltration occurs in a more distal manner. A kinetic analysis of lung leukocyte levels revealed a 3-fold increase in neutrophil numbers associated with dispersed lullg tissues 2 hours after LPS treatment, while macrophage levels increased by 4-fold at the 24 hour time point. Thus, the recruitment of different leukocyte populations to the lungs during endotoxemia is likely directed by different mechanisms. Recent studies have identified a supergene family of small inducible chemotactic cytokines (chemukines) which possesses chemotactic and activating properties for neutrophils. The prototype of this family is interleukin-8 (IL-8). Interestingly, a related supergene family has been identified which possesses activity for recruiting mononuclear cells. Examples of this group of inflammatory chemukines are monocyte chemotactic protein-i (MCP-I) and macrophage inflammatory protein-i alpha (MIP-I). In initial in viva studies we examined whether MIP-I was expressed systemically or in a compartmentalized fashion post LPS challenge. Assessment of plasma cytokine levels revealed maximal TNF levels occurred I hour post LPS administration, returning to baseline by 4 hours, while MIP-I levels were maximal at 2 hours (2,5 ng/ml), with a second peak at 24 hours after LPS challenge. Interestingly, aqueous extracts of liver homogenates from LPS treated animals demonstrated no MIP-I levels, while aqueous extracts of lung revealed a 4-fold increase in MIP-I levels over control lungs. Immunohistochemical analysis of the lungs from 24 hour LPS treated animals demonstrated the alveolar macrophage was a rich source of MIP-I protein. Cell-associated MIP-I was also expressed by blood monocytes adherent to the pulmonary vascular endotheliun, however the expression of monocyte-MIP-i was observed by 2 hours post LPS administration. Immunohistochemical analysis also demonstrated that MIP-I antigen is associated with the extracellular matrix on the interstitial side of the endothelium. This suggests that the extracellular matrix, which is produced during inflammation, can bind MIP-I and this may serve as a depot for the prolonged presence of NIP-1.

In additional studies we have demonstrated that the intratracheal instillation of rMuI [IP-l(lOOng) 

Activation of polymorphonuclear leukocytes by inflammatory stimuli may contribute to the development of multiple organ failure in septic patients. Thereby PMNL are proposed to avidly adhere to vascular endothelium causing damage by the subsequent release of toxic agents. As cellular adhesion is primarily mediated by 62-integrins and Lselectins, the present study compares the expression of these adhesionmolecules on PMNL in septic patients and healthy volunteers. Methods: Expression of 62-integrins and L-selectins on PMNL was measured in whole blood by flow cytometry using the monoclonal antibodies IB4 and Dreg 200, Baseline values were determined immediatley after drawing blood. In addition cells were incubated 45 min at 37°C to allow for spontaneous regulation of adhesion molecules. 55 blood specimens from 8 septic patients were obtained during the course of their illness. Control values were determined in 12 healthy volunteers. Results: Baseline expression of 62-integrins and L-selectins was not signifcantly different in septic and in healthy subjects. In contrast, there was a significant upregulation of g2-integrins and shedding of L-selectins of PMNL in septic patients (SP) compared to healthy volunteers (HV). The local or systemic production of inflammatory cytokines, such as tumor necrosis factor alpha (TNFc~), can serve to modulate multiple aspects of neutrophil function. The ability of neutrophils to leave the circulation and migrate to areas of infection is one essential component of host defense. L-selectin, a leucocyte-associated adhesion molecule, is responsible for the initial reversible contact between neutrophils and endothelium and the subsequent roiling action of neutrophils along the vessel wall. In contrast to other adhesion molecules, L-selectin expression is rapidly down-regulated after neutrophil activation. The loss of L-seleclin may thus be a critical determinant of how neutrophils become unbound from their endothelial attachments and enabled to proceed towards an underlying extravascular area of infection. We hypothesize that the shedding of L-selectin is a strictly controlled process, occurring primarily at localized sites of inflammation, which may be modulated by TNF~, A flow cytometric method of staining neutrophHs by monoclonal antibodies in whole blood is described whereby the kinetics of L-selectin shedding may be followed in real time. The dose response and time course of in-vitro L-selectin shedding by neutrophils from normal human subjects was assayed after exposure to N-formyl-methionylleucyl-phenylalanine (FMLP) and TNFc~. either singly or in combination, Our results show that L-selectin shedding can be reliably followed over time. A significant percentage of cells shed L-selectin after exposure to 2000 pg/ml TNFc~ or 1000nM FMLP (but not at 400pg/ml TNFc~ or 200nM FMLP). Greater numbers of cells were able to shed their L-selectin when FMLP and TNF~x were presented in combination rather than alone. High levels of TNFc~ did not appear to alter the threshold concentration of FMLP required to induce shedding, We conclude that the extent and rapidity of L-selectin shedding may be modified by different combinations of ligands and that shedding, by vidue of the high concentrations of cytokines or chemotactic factors required, is a process localized to sites of infection or inflammation. We prospectively studied 23 patients with severe sepsis syndrome; Group A : Septic shock with or without Adult Respiratory Distress Syndrome lARDS) (n =7, bacteremia = 6); Group B : Sepsis syndrome without septic shock (n = 16, bacteremia = 9). Serial plasma samples obtained on day 0, 1, 3, 6, 10 and 14, were assayed using ELISAs method (British Biotechnology), Normal control levels of soluble ICAM-1 and E-Selectin, obtained from 11 healthy volunteers, were respectively 205 ± 19.4 ng/ml and 45 ± 5.7 ng/ml (mean _+ SE), Acute lung injury was quantified dally on a tour-point score system (Murray, Am Rev Respir Dis,1989) . Compared to control mean values, initial levels of groups A and B were significantly higher for ICAM-1 (p < 10 -4) and E-Selectin (p < 10-3). Comparisons of group A and [] (* = p<0.05; ** = p<0.00t) Soluble ICAM-1 levels of group A enhanced significantly (p<0.05) during the first 24 hours, and a sustained high levels was of bad prognosis (25% of survivors at day 6). The evolution of soluble ICAM-1 and E-Selectin levels were significantly correlated with Murray's score (Spearman test : p <0.001). Conclusion: These results suggest that endothelial adhesion molecules are released into the plasma of patients with severe sepsis syndrome. Soluble ICAM-1 and E-Selectin are correlated with endothelial lung damage, and lOAM-1 seems to be a better indicator of the severity of endothelial injury. 

Introductory remarks to anti-adhesion molecule strategies as a therapeutic modality CH Wortel, Repligen Corporation, One Kendall Square, Building 700, Cambridge, MA 02139, USA. The development of antimicrobial therapy represented a major breakthrough in the struggle against disease. It strengthened the notion that disease could be overcome by eliminating foreign invaders threatening the host. This paradigm has proven to be very successful, the threat of many infectious diseases has significantly changed, some have even been eradicated. Nevertheless, sepsis has remained a severe condition, increasing in incidence while mortality remained very high. More recently, it has become increasingly clear that besides the nature and treatment of an exogenous agent, the reaction of the host defense itself plays a pivotal role in the outcome of the event. Endogenous mediators, such as TNF, IL-I, IL-6 and IL-8, govem many of the actions of the host defense system. While the expression of these cytokines more often than not benefit the host, (over)-expression can cause severe damage. Based on this hypothesis,anticytokine strategies, such as those targeted against TNF or IL-1, have been evaluated for the treatment of sepsis. Results of these early studies have not yet indicated success in improving the outcome of the disease. It has been difficult to define a patient population where a benefit could be reproducibly shown. Furthermore, it has been documented that synergy between cytokines occurs, but detailed knowledge of the cytokine network is not yet available. It is conceivable, that neutralization of one cytokine prompts the induction of another which will evoke the intended response in the host. Recent data obtained in human endotoxemic volunteer models seem to confirm this. If this turns out to be the case, neutralizing a single cytokine may not be a successful approach. Cytokines in tum, induce various adhesion molecules, such as ICAM-1. Such molecules regulate for instance the neutrophil-endothelial cell interactions, which are thought to play an important role in the pathogenesis of systemic organ injury. The potential for monoclonal antibodies to adhesion proteins to reduce vascular and tissue damage has been studied in a large number of experimental models. Protective effects have been observed in a wide variety of inflammatory, immune, and ischemia-reperfusion injuries. Thus, altering the host response by modulating the function of adhesion molecules may attenuate the inadvertent injury caused by inappropriate behavior of host defense cells. Targeting cellular surface interactions has been added to the efforts to change the outcome of disease. Modulation oftheseprocesses seems very promising, but may temporarily leave the host without effective defense mechanism. Great care therefore, must be exerted when studying this powerful two-edged sword in a clinical setting. Our knowledge of the role of adhesion molecules in the intlammatory response has increased rapidly due to the availability of new reagents and mice geneticly deficient in adhesion molecules. These molecules are important in interactions of leukocytes with endothelial cells, other leukocytes, platelets, and epithelial cells. When these molecules are engaged, they can also play a role in activating leukocytes and their effector functions. In the venules of the systemic circulation, adhesion often occurs through a series of sequential interactions. Initial interactions are mediated by members of the selectin family to loosely associate the leukocytes with the endothelium and are followed by firm adhesion requiring members of the integrin and immunoglobulin family. Later interactions with endothelium may require PECAM. Adhesion molecules are usually required for leukocyte emigration in response to extravascular stimuli and for neutrophil-mediated endothelial cell injury. They are critical for host response in many diseases including infections. However, when the inflammatory response results in damage to host tissues, patients may benefit from blocking the leukocyte response. Anti-adhesion molecule agents are an important potential antiinflammatory therapy. The focus of anti-adhesion therapy may be at any step of the sequence. Diseases where anti-adhesion molecule therapy may benefit patients include ischemia/reperfusion injury in many organs, ARDS and MOF, and transplantation, both to protect the donor organ from ischemia/reperfusion injury and to inhibit graft vs host disease. Many strategies have been considered and include: 1) blocking the ability of adhesion molecules to recognize their ligand using antibodies that have been humanized or soluble receptors linked to IgG to prolong their circulating halflife, 2) blocking the ligands for adhesion molecules using soluble adhesion molecules, peptide analogues, or oligosaccharides, and 3) blocking the production of the adhesion molecule using anti-sense oligonucleotides. Because the synthesis of adhesion molecules is usually regulated by cytokines, inhibiting the action of cytokines is another potential site for interrupting the adhesion process. Although important issues of safety must be evaluated, the potential for modulating the inflammatory response make this an exciting area of improvement in health care delivery. Claire M. Doerschuk, M.D.; Riley Hospital for Children, Room 2600; 702 Barnhill Drive; Indianapolis, IN 46202 USA.

Modulation of neutrophil-endothelial cell adhesion with anti-CDl I/CD18 monoclonal antibodies as a therapeutic modality.

CH Wortel, Repligen Corporation, One Kendall Square, Building 700, Cambridge, MA 02139, USA. The central role of inflammatory cells in the pathogenesis of lung and systemic organ injury is well recognized. Binding of neutrophils to endothelial cells and migration into the parenchyma are largely regulated by complementary adhesion molecules. The leukocyte integrins are glycoproteins expressed on the neutrophil surface and in the cytoplasmic granules. Integrins consist of a common beta2 or cluster differentiation (CD)18 chain covalently linked to one of three different alpha chains (CDlla, CDllb, CDIlc) and exist on the cell surface as three distinct heterodimers. CDlla/CD18 is expressed on all leukocytes, whereas CD 11 b/CD 18 and CD 11 c/CD 1.8 are restricted to cells of myeloid origin. CD i 1/ CD 18 interacts with intracellular adhesion molecule-1 (ICAM-I), its ligand on endothelial cells. The potential for monoclonal antibodies to adhesion proteins to reduce vascular and tissue damage has been studied in a large number of experimental models. Protective effects with anti-CD18 antibodies have been observed in a wide variety of inflammatory, immune, and isehemia-reperfusion injuries, such as arthritis, burns, endotoxic shock, bacterial meningitis, autoimmune diabetes, nerve degenemrion, allograft rejection, allergic asthma, acute lung inflammation, skin lesions, and ischemia-reperfusion models of the intestine, myocardium, lung, skeletal muscle, and central nervous system. Protective effects have also been observed in animals resuscitated following hemorrhagic shock. Blockage of CD18, however, would affect all leukocytes, as would antibodies to CDlla/CDI8. Targeting CDllb/CD18 would affect cells of the myeloid lineage only, which could prove to be beneficial. CD11b/CD18 is not only involved in transendothelial migration, but is also implicated in adherencedependent formation of reactive oxygen species. Blocking CD1 lb/CD18 may therefore not only reduce the numbe r of leukocytes accumulating in the tissue, but also attenuate the oxidant stress of infiltrated neutrophils. Anti-CD 11 b treatment has been used effectively to reduce tissue injury initiated by ischemia-reperfusion, complement activation and endotoxemia. Altering the host response by modulating the function of adhesion molecules may attenuate the inadvertent injury caused by leukocytes, but may also temporarily leave the host without effective defense machinery. Overall, animal studies suggest that it may be safe to inhibit neutrophil adhesion for a limited period of rime. These observations will have to be confirmed in carefully designed clinical trials. C, arbobydrams are ubiquizom constir~uts of cell sv.rfaees, and possess many c~xSsfies ttm~ m~,e ~em ide~. canaidates for r~ognifioa mole~ule& In m~y systems whe,~ ceR udhesioa ~lays a critical ro~ car~hydram l:~dtag ~otegas have been shown to b~ad tocell surfa~ earbohydzaxes ~nd pZrl~pate in cell-ceil Lumtaefion& Such sys,.ems include ~rti~za~io=, deveaopmeat, l~thoge~-hcet reeog--ition ~d i~zmmadon_ In particular, tb.z recent di%~ve~ of Lhe selec~ and th~ impo.~a~C~ in teukccy~udo~lelium adh~ion has -~f~M av.c~on OK l~in m~ted cell adhe~on. S~vere/poten~s/cs.rbohydr~ l~ga~S hRve ~e~l ~u~ilied for ~he s~lcc~ins. The,~ c~u be broadly di,,Sded La~o ~Wo m'oups -sibyl L~wis x m~ mh~.~l Oligo~chadd~s, ~d sf/~ ca~ohydmma, All ~:~ ~l~dns bind m siflyl L~wis x (sie$ o!igos~ccb.e.rkMs, zlthou~ w~ differing avi~Re~. 'We have i~¢n~ed the functional g~oups 0a s~ex ~n~ med/a~ ~he b~u ~di~g of ~h~ c~b0hydmm = E-Se/sedm We have used ~hat iv.formation to sya~esize sLe ~ '~mt0gs r.he, t focus on replacing sLslic ~Sd ~nd fuc0s¢ wi~ simpler, more stable strunt~es. A[~ou~a ~ proeer~ is ongoing, we hve been ~ucee,.~ful a~ rep~aein8 t.ke si~ic a~id. residue wi~ std.fzte. ~ce~ or la~c amd groupa We t'we ex2aninad &e ten, bunion of ezed~ hydroxyl group of the fizeose residue ~ billding of E-, L-~nd P-SeleeS..u. We have also found m~fi~fio~ of the reducing end ~¢.cha'i~ ~z increase mtagovSst activity. The, M¢ond. group of figs,rids a.R eontzin su~a~ 0u a ea.rbohydr~t¢ support,, und seem to bi~.d to t~e sele~ti~s wi~ dLf:ferem characteristics C0.an does sLe:, S=h compounds are m2ogniz~d by L-Selects. md p-Selectia, bur., in genera/, not E, Selecti~ These dam may mdicam r.hat L-and P-S~ ¢at~ h~d via o, second ~te thaz operates Lu~.ead of, or in conjunction with ~tc sLe" b~ding ~iite. Dam rela~&~g to ±e, se two types of ,ml~ liga~ds have beam t~ed to desig~ potential the2~peutics for i~fi~anmat0ry disease. Lr:rng maimaI models of acute lung Lu3ury we can demo~trate that eompmmds that inhibit seleetiu birding ~ ~i~o hzve ber~ficial effects when uc~d in rive. Progressive microvascular damage in the tissue adjacent to a cutaneous burn injury results in extension of burn size. The role of leukocytes in the pathogenesis of microvascular injury was investigated by inhibition of their adherence to the microvascular endothelium using monoclonal antibodies directed to leukocyte CDI 8 or its endothelial ligaud, intercellular adhesion molecule-1 (ICAM-1, CD54). A model of thermal injury was developed using New Zealand White rabbits. Two sets of three full-thickness burns separated by two 5 x 30-mm zones were produced by applying brass probes heated to 100°C to the animals' backs for 30 sec. Cutaneous blood flow determinations carried out with a laser doppler blood flowmeter were obtained for 72 hours. There were five experimental groups: controls given saline alone; animals given monoelonal antibody to the CD 18 R 15.7 prior to burn injury (pre-R 15.7); animals given R 15.7 30 min after burn injury (post-R 15.7); animals given a monoclonal antibody to ICAM-I, R 6.5 prior to burn (pre-R 6.5); and animals given the R 6.5 30 min postburn injury (post-R 6.5). Blood flow in the marginal "zone of stasis" between burn contact sites was significantly higher in the antibody-treated animals. Administration of the antibodies 30 min after injury was as effective as preburn administration in preserving blood flow. At 72 hr post-burn all antibody -treated animals had blood flow in the areas at risk for progression (i.e., the zone of stasis) at or above baseline levels while the control animals had levels equal to 34.7 _+ 12% of baseline (P < 0.05 by analysis of variance and Mann-Whitney U test). These results indicate that leukocytes play an important role in the pathogenesis of burn wound progression, and that this progression can be attenuated by moduIating adherence to endothelial cells. A wealth of information now supports the hypothesis that inhibition of cell adhesive mechanisms will Nter the course of immunologicand inflammatory processes. What remains unclear is whether inhibition of specific mechanisms wfl[ be of therapeutic benefit in any specific human disease. Current data derived from animal models are not inconsistent with the hope of therapeutic benefit, but techniques for inhibition (e.g., antibodies, antisense oligonucleotides, inhibitory peptides, inhibitory carbohydrates, smaII synthetic inhibitors, etc), tissue and species differences in the relative contributions of adhesion molecules to the inflammatory process, and the casCade model of adhesive interactions are all confounding issues, making predictions of therapeutic benefit in any specific human disease process very difficult.

Additional concerns involve the potential roles of adhesive mechanisms in host resistance to infection. As human therapeutictdals are initiated, more exact information on the roles qf specific adhesion molecules in human disease should emerge.

Inhibition of leukocyte adherence to endothelial cells can represent a novel therapeutic approach to septic shock. We performed a pilot study to evaluate the safety and tolerability to CY-1787, a monoclonal antibody against human E-selectin, in patients with septic shock. Septic shock was defined by clinical signs of sepsis, a documented source of infection, and fluid-resistant hypotension requiring the use of vasopressors. Eleven patients entered the study, but 2 patients who died during the first 24 hours were excluded, as this was part of the protocol. CY-1787 was administered as a single intravenous bolus of 0.1 mg/kg (N=3), 0.33 mg/kg (N=3) or i mg/kg (N=3) mg/kg. The antibody was well tolerated. None of the 9 patients died during the 28 day follow-up period. Organ failure was assessed for 5 organs (CNS, lungs, liver, kidneys and coagulation). The mean number of organs failing, which was initially 2.7 ± 1.2, decreased to 0.2 ± 0.4 at the end of the study (p<O.O01). These results are therefore encouraging. Administration of an antibody to E-selectin in patients with septic shock requires further study. The importance of cytotoxic T lymphocytes (CTL) in the host defense mechanism against viral as well as parasitic and bacterial diseases is becoming increasingly appreciated. The design of vaccines to generate cytotoxic T ceils from nonviable or subunit constructs, however, poses challenges due to the unique characteristics of antigen processing and presentation by class I major histocompatibility complex (MHC) molecules. Since the final antigen recognized by CTL are short peptides capable of high affinity binding to class I MHC molecules, we have designed strategies: a) to define the structural characteristics of peptides that are required for their high affinity binding to MHC; and b) to formulate these peptides into immunogenic vaccines that are capable of generating cytotoxic T lymphocytes. The strategy and results to accomplish these two goals will be discussed. Amnon Altman and Erich Gulbins Ras proteins represent essential intermediates in receptor-mediated growth and differentiation pathways. They couple signals initiated by receptor or nonreceptor protein tyrosine kinases (PTKs) at the cell surface to cytoplasmic targets which coordinate signal transduciion to the nucleus. Ras also participates in T lymphocyte activation initiated by the antigen-specific T cell receptor (TCR)/CD3 complex, which leads to lymphokine production and proliferation. Receptor ligation causes activation of associated PTKs of the Src and Syk families as the earliest identifiable event in this pathway. The importance of Ras in T cell activation is indicated by the findings that an oncogenic Ras protein activates the interleukin-2 (IL-2) gene promoter; conversely, a transdominant inhibitory ras mutant inhibits this event and the activation of MAP kinases. The rate-limiting step in Ras activation is the exchange of bound GDP for GTP, which is catalysed by guanine nucleotide exchange factors (GEFs). We identified the SH2/SH3 domain-containing Vav protein, which is selectively expressed in hematopoietic cells, as a Ras-specific GEF coupled to several hematopoietic receptors. Vav can be activated by two pathways, i.e., by PTK-mediated phosphorylation or by binding of diglyceride second messengers to a cysteine-rich, protein kinase C (PKC)homologous Vav domain. These pathways are independent as demonstrated by structure/function analysis and the use of pharmacological inhibitors. The two activation pathways may enable Vav to integrate signals from distinct hematopoietic cell receptors, and couple them to Ras. T cells express another, ubiquitous Ras GEF, i.e., Sos, which is known to be coupled to activated TPK receptors. T cell activation leads to membrane association of a signaling complex consisting of Sos and two other signaling proteins, Grb-2 and Shc, via direct binding of the Shc SH2 domain to the tyrosine-phosphorylated chain of the TCR/CD3 complex. Thus, multiple receptors, PTKs and Ras activators participate in signaling pathways that lead to hematopoieitc cell growth and differentiation. The interactions and function of these various signal transducers will be reviewed. phagocytes. Immunity is mediated by specific T lymphocytes, Cytokines produced by these specific T cells activate antimicrobial capacities in mononuclear phagocytes, thus contributing to protection. Other immune mechanisms also participate in the acquisition of resistance. On the other hand, T cells are also crucial for many pathologic sequelae of intracellular bacterial infections. Although IFN-y is central to macrophage activation, synergistic and antagonistic effects with other cytokines occur.

The cd/3 T cells, representing the major T cell population in experimental mice, are the main mediators of antibacterial immunity; an auxiliary role for y/~ T cells appears likely. The c~//3 T cell population further segregates into CD4 T cells which recognize microbial peptides in the context of gene products of the major histocompatibility complex (MHC) and CD8 T cells which see their peptide in the context of MHC class I.

The -y/~ T cells are generally double-negative. Knock-out mice in which the genes for various T cell receptor chains, for MHC class I or MHC class II molecules, for cytokines or cytokine receptors had been deleted, have provided extremely helpful tools for analyzing the role of T cell subsets and cytokines in antimicrobial immunity. Using these mutant mice, the role of distinct T cell subsets and cytokines in bacterial elimination and granuloma formation was analyzed in detail. These studies underline the central role of a//3 T cells and IFN-y in antibacterial immunity and, furthermore, show a distinct function of 7/$ T cells. Moreover, these studies show that the relative contribution of CD8 or CD4 T cells to antimicrobial immunity varies in response to different pathogens. Apoptosis is an active form of cell death in which the cell participates in its own demise, providing the biochemical pathways that trigger and mediate the complex events of the process. Although this phenomenon has been studied for many years, it is only recently that significant progress has been made towards the elucidation of the molecular mechanisms that control apoptosis. Once such mechanism came as something of a surprise: in a number of eases, apoptosis can be promoted by the action of the c-Myc protein, a protooncogene product that had previously been associated only with cell proliferation (and presumably, survival) . Expression of c.Myc in cells can cause them to enter an apoptotic pathway or proliferate, mad this "decision" depends upon additional signals, such as growth factors that inhibit apoptosis and allow the ceils to proliferate. Thus, c-Myc directs cells out of rest and the choice of proliferation or death depends upon the presence or absence of survival signals. Other survival signals are generated by oncogene productS that can cooperate with My¢, including Bcl-2 and AbL There are implications of this interplay between apoptotie and anti-apoptorlc signals in cell Iransformation and the resistance of some tumor ceils to therapeutic agents capable of inducing apoptosis. Oncogene interactions therefore control the life or death of transformed ceils but also have important roles in normal, development processes. In the immune system, developing T ceils are "screened" for potential self-reactivity by a process of negative selection, in which activation via the T cell receptor (TCR) induces apoptosis. This process can be mimicked in T cell hybridomas, and appears to depend upon the expression of the c-Mye protein.

Evidence that this represents a requirement for the Myc/Max heterodimer will be presented. As above, these cells are presented with the "choice" to proliferate or die and this depends not only upon Myc but also other signals. Thus, expression of functional v-AN kinase in these ceils can render them resistant to the activation of apoptosis via TCR ligation, Surprisingly, Bcl-2 does not appear to be capable of blocking this form of apoptosis, and the reasons for this will provide new insights into the development of the immune system and the process of apoptosis. Although the oncogene products discussed here probably do not directly participate in the biochemical process of apoptosis, they represent molecular controls on this complex mechanism. As such, they gave us new ways to look at the life and death of a cell Recent attention has focused on macrophage release of cytokines, such as IL-1, IL-6 and TNF as important mediators of septic shock. However, serum cytokine levels have correlated poorly with outcome of septic shock. The purpose of this study was to compare the functional status of macrophages to serum cytokine levels in early sepsis by an analysis of splenic macrophage protein synthesis during abdominal sepsis. Macrophage total protein distribution and biosynthetic activity was assessed by large format 2-D PAGE, autoradiography of 35S-labeled proteins and computer assisted densitometric analysis. Sepsis was induced in Sprague Dawley mts by cecal ligation and puncture (CLP). Untreated and sham operated rats served as controls. Splenic macrophages were harvested at 3 and 24 hrs. following CLP. Serum IL-6 levels were determined at 3,12, and 24 hours by the 7 TDI bioassay. By 12 hrs. 100% of the CLP mls exhibited multi-microbiaJ bacteremia. Control or sham operated rats did not show bacteremia. CLP resulted in a significant 286% elevation (p < 0.05) in serum IL-6 levels at 12 hrs. There was no difference in IL-6 at 3 and 24 hrs. when compared to the control groups. 2D PAGE studies examined splenic macrophage total protein distribution and biosynthetic activity at 3 and 24 hrs. post-CLP. At 3 hrs. macrophages from CLP rats showed a substantial decrease in total protein pattern (353 protein spots vs 503 protein spots) when compared to non-septic control macrophages. A decrease in the number of lower molecular weight proteins (< 40 kD) was observed. At 24 hrs. post-CLP, splenic macrophages from CLP rats showed a substantial increase in the synthesis of low molecular weight (< 40 kD) proteins, when compared to non-septic controls. We conclude that: l) fulminate abdominal sepsis induces an early (3 hrs.) Staining with CD14 and CD16 mAbs will identify two monocyte subpopulations in human blood: A major population of regular monocytes, which strongly expresses the CD14 antigen (CD14++) and a minor population with weak expression of CD14 and expression of the CD16 antigen lCD14+/ CD16+ cells). In healthy control donors(n=35) the latter cells account for 45+22 cells/mm 3 and 9%+5% of the monocytes. In sepsis patients, the CD]4+/CD16+ cells can become a major population with more than 50% of all monocytes in 3 of 18 patients and with more than 500 cells/mm 3 in 4 of 18 cases. There was no correlation of CD14+/ CD16+ cells to any clinical parameter except for CD14+/CD16+ percentage and body temperature (p= 0.013).

The CD14++ monocytes showed a substanttial decrease in CD14 antigen density in 9 of 11 patients. Three-color-IF shows that the CD14+/ CD16+ cells in sepsis patients when compared with the CD14++ monocytes exhibit a higher level of class II antigen and a lower level of CDIIb and CD33 antigens, consistent with a more mature nature of the CD14+/CD16+ cells. Levels of IL-6 were increased in sepsis patients: 3 of 5 patients with high numbers of CDI4+/CD16+ cells ( 200/mm 3) had high levels of IL-6 (250U/ml). These data suggest that septicemia may lead to substantial changes in blood monocyte composition and this may be related to elevated levels of cytokines such as IL-6. Human peritoneal macrophages were exposed to increasing doses of Lipopolysaccharide (LPS) or a synthetic Lipid A analogue (MRL 953) and production of the cytokines IL-lb, iL-6, TNF-a and G-CSF was assessed at the protein-and mRNAlevel. Cells were also stimulated with low doses of LPS and MRL 953 to study their "adaptation" to a secondary challenge with high doses of LPS. TNF-a production after stimulation with LPS could be almost completely relieved by preincubation of cells with low doses of LPS and similarly, IL-6 production was suppressed. Surprisingly, however, "adapted" cells were able to synthesize even larger quantities of of G-CSF and IL-lb when exposed to a secondary LPS challenge. The changes in TNF-a, IL-6 and G-CSF protein synthesis could a/so be detected on the mRNA level, whereas transcript levels for IL-lb remained unaltered as assessed by Northern blot analysis. High doses of the synthetic lipid A analogue MRL 953 were also able to "adapt" macrophages for a secondary LPS challenge by downregulating TNFa-and IL-6-production, while priming secretion of G-CSFand IL-lb as well. Taken together, here we describe for the first time the discordant adaptation of human peritoneal macrophages to a secondary LPS stimulus in vitro. These findings appear to bear ramifications for the in-vivo endotoxin response during inflammation and gramnegative septicemia. Shock states with inadequate cellular oxygen delivery may contribute to macrophage (MO) activation or dysregulation. In this study we compared the effect of transient hypoxia and endotoxin pretreatment (LPS 1) on MO mediator release with a second endotoxin stimulus (LPS2). Methods: In vitro cultures of marine peritoneal exudate MO were exposed to 2 hr. of hypoxic or normoxic conditions, then incubated 22 hr under identical normoxic conditions _+ 10 ng/ml of LPS 1 pretreatment. During the final 24 hours all M~l were exposed to a range of LPS 2 concentrations. M~i supernatants were assayed for TNF, IL-1, IL-6, PGE2, nitric oxide (NO), and superoxide release. Results: LPS 1 markedly inhibited MO TNF release by LPS 2 but hypoxia had no effect on LPS 2-triggered TNF release. Hypoxia increased MO IL-6 production in the absence of LPS 1, but inhibited LPS 1 augmentation seen under normoxic conditions. Pretreatment with LPS 1 increased NO production from LPS 2 under normoxic conditions, but bypoxiainhibited this effect. Pretreatment with LPS1 signifcantly augmented MO release of IL-i and PGE 2, but hypoxia had no significant effect (data not shown) Superoxide production was increased by LPS 1 under normoxic conditions, but hypoxia significantly inhibited superoxide release (not shown A. Lepape, C. Docile, F. Arpin, M. C. Gutowski, V. Banssillon and J. Bienvenu. I AIM OF THE STUDY. Increased levels of TNF are inconsistently correlated with the severity of sepsis. One possible explanation is a decrease ability of ceils to produce eytokines. The objectives of this study were m compare at different levels of of clinical severity the responsiveness of cytokine producing cells. This was evaluated by the response to a stimulation by LPS as well as to the inhibitory effect of PTX. The technic used is a whole blood ex-vivo model, as described by DESCH et al. (1) . H MATERIALS AND METHODS. Different groups of ICU patients, postoperative (n=12), sepsis (n=12), septick shock (n=13), and healthy control (n=10), were studied. Blood was drawn in endotoxio free heparinized robes. Stimulation was achieved by addition nf75 pl ofLPS (10ng/ml) to 675 Id of whole blood in presence of various concentrations of PTX (0, 10 "5, 10 -4, 10 "3, I0"2M). After 5 hrs incubation at 37°C, the tubes were centrifuged and supematants frozen at -80°C. TNFa and/L6 were measured by ELISA (Medgenix R and Immunotech R respectively). Monocytes were quantified on a Technicon H6000. The results are expressed as median values. Non parametric tests are used for testing differences between groups. HI RESULTS, TNFtx and 11,6, corrected for monocytes count, are given as % of their initial value, the baseline before stimulation or inhibition being 100%, 1) STIMULATION BY LPS. The control group is much more stimulated (>21000% for IL6, >8000% for TNFa). Blood samples taken postoperatively and in patients with simple sepsis are significantly less stimulated (>1500% for IL6, >600% for TNFa ). The lowest stimulation was observed in patients with septic shock (median = 168%), some patients being not stimulated at all. 2)EFFECTS OF PTX.The inhibitory effect of PTX on TNFtx production is effective in all groups at 10 -3 M (reduction to less than '¼ of the median values), and is almost complete at 10" 2 M. The septic shock group has a decreased sensitivity to PTX. IL6 production exhibits a lesser reduction at 10 -3 M (~ 'A to ½ of the median values), further increased at 10 -2 M. The septic shock group is again less sensitive to PTX. IV CONCLUSION: The reduced ability of circulating monocytes to produce cytokines during severe infections is confirmed here. PTX is able to reduce significantly TNFc~ at 10 -3 M and the inhibition is nearly complete at 10 -2 M. Surprisingly, there is a lesser, but significant suppressive action of PTX on IL6, not found in experiments using purified monocytes. One possible explination could be the interplay between cytokines production.

(1) Lymphokine research (1989) cDNA sequencing constitutes a powerful method of measuring steady-state mRNA levels for all genes transcribed in a given cell or tissue at a particular stage of differentiation. By comparing transcript abundance both prior to and following differentiation, individual genes can be identified whose transcription is regulated both positively and negatively. In order to examine monocyte activation, the human monocyte line THP-1 was induced with phorbol ester (48 h) and activated for 4 h with lipopolysaccharide (LPS) after which polyA + RNA was purified. The RNA from control and LPS-treated cells were each used to construct a cDNA library under identical conditions, and all resulting clones were selected for cDNA sequence analysis. Each clone sequence was evaluated by matching with both GenBank and our own gene databases. Very different patterns of gene expression were seen in the two libraries, the latter reflecting very high levels of known inflammatory mediators such as IL-1 and TNF. A second set of libraries were made from umbilical vein endothelial cells (HUVEC), both with and without LPS stimulation, and were analyzed in a similar fashion. The effects of LPS induction on specific gene transcription in both cell types will be discussed. T. Tadros, MD, Th Wobbes, ME) PhD, RJA Goris, MD PhD

To investigate whether the preactivation of regional macrophuges by liposomes containing muramyl tripeptide (MTP-PE) can counteract the detrimental effect of blood transfusions on both anastomotic repair and host susceptibility to infections. Methods Eighty Lewis rats received lmg/kg of either empty or MTP-PE encapsulated Liposomes, intraperitoneally (ip). Twenty-four hours thereafter, the animals underwent resection and anastomosis of both ileum and colon, and received 3 mL of either saline or blood from brown Norway donors,iv. The animals were killed 3 or 7 days after surgery and examined for septic complications and anastomotic repair.

The average anastomotic strength, as assessed by bursting pressure (+SD), was significantly diminished in the transfused animals, as compared to the non-transfused animals (ileum;day 3; 44-+8 vs 99+7, P<0.001). Transfused animals pretreated with MTP-PE encapsulated Liposomes showed a significant improvement of their anastomotic bursting pressure (93+9, P<0.001 vs transfusion). Pretreatment with MTP-PE encapsulated Liposomes decreased significantly the incidence of anastomotic abscesses in transfused animals ( from 90% in ileum on day 3 to 20%, P<0.001). Conclusions Preactivation of regional macrophges by intraperitoneal administration of MTP-PE encapsulated Liposomes prevents the detrimental effects of transfusions on anastomotic repair and reduces the incidence of intraabdominal sepsis.

Academic Hospital Nijmegen, Dept of General Surgery, PB 910 I, 6500 HB Nijmegen, The Netherlands.

LEUKEMIA CELL LINE, TEIP-1. Robin S. Wa, gner*, Perry V. Halushka "~, and James A. Cook*, Departments of Physiology , Pharmacology "l" and Medicine "t, Medical University of South Carolina, Charleston, S.C. 29425. Adherence of monoeytes to endothelium and extracella/ar matrix proteins is essential for accumulation at sites of inflammation. TXA2, an arachidonic acid metabolite, inhibits human monocyte chemotactic responses suggesting that TXA 2 may alter monocyte adhesiveness. We selected the THP-1 cell line, a human monocytic leukemia cell line to further investigate the effect of TXA 2 on adhesion. We tested the hypothesis that TXA 2 alters LPSinduced adhesion of THP-1 cells and that TXA 2 exerts its effect on adhesion via a cAMP dependent mechanism. THP-I cells were exposed to S. enteritidis endotoxin (lp.g/ml) _+ the cyelooxygenase inhibitor lndomethacin (IN), the TXA 2 mimetic I-BOP (0.11.tM,) or TXA 2 receptor antagonists BMS180291 and L657925 (10~M). Cells were allowed to adhere for 24 hours and adherent protein/well was determined. LPS-induced a significant (p<0.05;n=7) increase in adherence of THP-1 cells (Basal, 2.4+0.85gg protein/well; LPS, 20.4+_6.0p.g protein/well). The amino acid glutamine is an essential compound for synthesis of purine and pyrimidine basis and therefore necessary for RNA-and DNA synthesis. In human plasma the concentration of glutamine is between 0.5 -0.6 mM, and is reduced in septic patients up to 80 % (0.1 -0.2 mM). Monocytes play a central part in the inunune system and it was of interest, whether glutamine is involved in the modulation of cell surface markers and phagocytosis of these cells. Human peripheral blood mononuclear ceils were obtained from 500 ml heparinized blood of apparently healthy donors by Ficoll-Paque density gradient and isolated by counterflow elutriation. The puritiy was more than 90 %. Subsequently cells were cultured in phenolred-free RPMI 1640 medium with various concentrations of glutamine (0.05, 0.1, 0.2, 0.3, 0.6, 1, 2 mM) in teflon-fluorinated ethylene propylene bottles to exclude cell adhesion and possible cell activation. Aider seven days culture, cell viabilty was determined by trepan blue exclusion and varied between 70 and 90 %, independent of glutamine concentrations. Cell surface markers were detected by flow cytometry, noaspecifie phagoeytosis was measured with latex beads and specific phagocytosis with opsonizied E.eoli using a FACScan. Lower concentrations of glutamine decreased the expression of HLA-DR and ICAM-1/CD54 on monocytes in a dose-dependent manner. The receptor for Fc'/RUCD64 as well as the receptors for complement CR3/CDllb and CR4/CDllc were down-regulated. CR1/CD35 which is only slightly expressed on monocytes was not influenced. Furthermore, no effects on the expression of CDI4, the receptor for transferrin CD71 and Fc'fRIII/CD16 were seen. Our data indicate, that lower concentrations of glutamme influence the phenotype of monocytes. We are now interested to study whether glutmnine influences non-specific phagocytosis, or whether specific phagocytosis correlates with the decreased expression of Fc'/R and complement receptors. We investigated immunologically more than 300 patients who were admitted to ICU because septic syndrom during the last four years. Patients were immunologically followed up 2-3 times per week until release from ICU. The expression of HLA-DR antigen on monocytes turned out to be the best prognostic parameter. The persistence (> 5 days) of low HLA-DR expression (< 30%) predicts fatal outcome (mortality > 85%). The altered phenotype was associated with a functional deactivation of monocytes (diminished APC, ROS formation, cytokine secretion). We called this phenomenon "immunoparalysis". IFN-gamma and GM-CSF were able to restore the altered phenotype and function in vitro. However, addition of autologous plasma from septic patients with "immunoparalysis" to these cultures prevented the cytokine-induced restitution. The inhibitory activity could not be removed by dialysis. Therefore, we started a study to prove the therapeutic efficacy of plasmapheresis. Indeed, [7 of 33 patients recovered from "immunoparalysis" following repeated plasmapheres; 16 of them survived (46 %). 7 patients recovered temporarely and 9 patients did not respond (all 16 died). The survival rate in the control group of septic patients with persistent "immunoparalysis" was 4 of 38 (11%; p<0,01). In summary, plasmapheresis in association with immune monitoring may be an alternative strategy to improve survival rate in severe sepsis.

Taurolidine, a synthetic taurine-formaldehyde derivative has antiadherent, bactericidal and anti-LPS properties functioning primarily through binding of the lipid A region of the LPS molecule. The active derivative of Taurolidine, taurine, modulates calcium channel activity, critical to the initiation of a number of immunostimulatory pathways. We hypothesised that Taurolidine may have direct immunostimulatory activity. The aim of this study was to investigate the immune effects of Taurolidine on peritoneal macrophage (PMO) function and then determine the role of taurine in this response. Study 1: In vivo stimulation:CD-1 mice (n=35) were randomized to receive Taurolidine (200mg/kg BW/i.p.) or saline cor~trol. PeritoneaI cells were harvested after 24 hours and were assessed for PM0 function [Superoxide anion generation (O2-), Nitric oxide (NO), Tumor necrosis factor (TNF), Fc/CR3-mediated phagocytic function (Phago) Study 2: Control PM0 were harvested and cultured in vitro with taurine (1.0mg/ml for 10 hrs), after which time they were assayed for 0 2-and TNF release. In vivo stimulation with Taurolidine Taurolidine has specific immunological effects on M0. Release of the inflammatory mediators 0 2-and TNF, and Fc/CR3-mediated phagocytosis were significantly increased, while release of the endothelial relaxing factor NO was significantly reduced. In addition, the amino acid taurine, which is released as a byproduct of Taurolidines breakdown has an immunostimulatory effect on PMO and may be the active moeity of the compound Tanrolidine. In sepsis, a number of mediators which affect vasomotor tone and cardiovascular function are produced. Inasmuch as sepsis causes decrease in systemic vascular resistance (SVR), attention is usually focussed on vasodilators such as lactate, tumor necrosis factor, interleukin-i & 2, and nitric oxide. But injury and inflammation als0 cause production of several vasoconstrictors whose effect may not be evident in changed SVR, but may significantly affect organ blood flow or function in the paracrine environment. Endothelin (ET) is a 21 amino acid peptide vasoconstrictor produced by ischemic or injured endothelial cells (EC's). ET is also a potent constrictor for renal mesangial and coronary vessels, an endocrine regulator, and a negative cardiac inotrope. Systemic ET levels increase significantly in hypoperfusion and ischemia. While ET is principally produced by EC's, we asked if human monocytes might also produce ET and thereby regulate vasomotor tone in areas of inflammation. Monocytes from healthy donors were separated on Ficoll, resuspended in RPMI 1640 +5% fetal calf serum and stimulated with i0 ug/ml endotoxin (LPS). ET was measured by radioimmunoassay. LPS-stimulated monocytes produced 87 ! 6 fM of ET/106 cells (vs. unstimulated controls of <25). This calculates to 15-30% of the amount of ET observed in patients with low cardiac output, sepsis or ischemia. We conclude that ET is a cytokine produced by both EC's and monocytes with potent effects on numerous cells and organs in the critically ill. Wuppertal 5, Germany We and other authors showed that fatal outcome in septic disease is associated with a decreased capacity of peripheral blood monocytes for the in vitro production of proinflammatory cytokines, especially TNF-alpha. We found that this monocytic deactivation is completed by a persistent and marked decrease of HLA-DR expression on monocytes (< 30 % HLA-DR+ monocytes) and a diminished antigen presenting activity whereas the capacity to form the antiinflammatory IL-1 receptor antagonist remains high. In order to evaluate the in vivo situation and to determine at which level TNFproduction/secretion is altered we assessed the TNF-alpha mRNA expression in freshly isolated peripheral blood mononuclear cells (PBMNC) from septic patients. TNF-mRNA was onty rarely detected by semiqaantitative polymerase chain reaction in PBMNC's from septic patients with monocyte deactivation. Meanwhile, it was found in almost all PBMNCs from septic patients without monocytic deactivation. We wondered, whether IL-I0, which ,is known to depress monocytic proinflammatoly response and MHC class II expression, could be one of the mediators in fatal sepsis. In fact, we found that IL-10 message in PBMNCs of septic patients peaked in the beginning phase of monocytic deactivation. In further investigations we found that TNF-administration can induce monocytic deactivation in a murine model/n vivo and provoke IL-10 message in human PBMNCs in vitro. These results support our hypothesis that an excessive delivery of proinflammatory cytokines in a first phase can induce an overwheiming inhibitory feedback, mediated by immuninhibitory mediators like IL-l0, which leads to often fatal monocytic deactivation in a second phase. Interferon-gamma which is known to counteract IL-10 production and the effects of IL-10 on monocytes restores the function and phenotype of monocytes from septic patients with monoq, te deactivation in vitro and could be a possible therapeutic agent in otherwise fatal sepsis. Our laboratory previously reported that LPS dependent macrophagederived TNF-a production can be enhanced by pretreatment with LPS at substimulatory LPS priming doses coincident with a suppression of LPS dependent nitric oxide (NO) production (Zhang and Morrison, J. Exp. Med 17:511, 1993) . In order to extend the characterization of these LPS priming effects in mouse macrophages, we examined the capacity of substimulatory LPS to modify LPS dependent IL-6 production. Macrophages were obtained from peritoneal exudate of thioglycollate treated C3Heb/FeJ mice and cultured in RPMI 1640 medium containing 10% fetal bovine serum. Macrophages were pretreated with various subthreshold stimulatory concentrations of LPS (Olll:B4) for 6 hours, washed three times, and then stimulated with the effective stimulatory concentration of LPS for 18 hours. The amount of IL-6 in the supernatant was measured by IL-6 dependent cell line (B9 and 7TD1) proliferation assay. IL-6 was produced by macrophages at lower threshold doses of LPS than those required for TNF-o~ or NO production. Subthreshold doses of LPS modulated IL-6 production in a biphasic manner characterized by an initial suppression and then potentiation. Higher doses resulted in secretion of IL-6 during the initial incubation with LPS and subsequent desensitization.

IL-6, like TNF-~ and NO, is, therefore, also affected by LPS pretreatment. Moreover, TNF-a and IL-6 shared the similar potentiational pathway, but differed by the fact that only IL-6 was inhibited. (Supported by R37 AI23447 and PO1 A54474.) Department of Microbiology, Molecular Genetics and Immunology and the Cancer Center, 1000 Wahl East, University of Kansas Medical Center, Kansas City, KS 66160-7832.

Korolenko T.,Urazgaliev K.,and Arkhipov S. The role of macrophage (Mph) stimulation in mechanism of protective effect of new immunomodulators yeast polysaccharides -heteropolysaccharide Cryelan and homopolysaccharide mannan Rhodexman (both produced by Petersburg Chem.-Pharm. Inst.) was studied. In vitro according to NST test incubation of murine peritoneal Mphs with Cryelan or Rhodexman, 150 ~g/ml,30 min was followed by increase of potencial microbicidic activity of Mphs. In vivo Mph stimulation by immunomodulators studied included increase rate of carbon particles phagocytosis during single i.v. or i.p. mode of administration to mice 1-14 days after (peak at 2nd day for i.v. and 5th day for i.p. mode of administration of the same dose of 5 mg/100 g b.w.).The preliminary injection of Cryelan (5 mg/100 g, 24 or 48 h before) to mice with acute cold stress (-10 ° C, 1 h) revealed protective effect restorating the value of depressed phagocytosis up to the normal level;the positive effect on ultrastructure of hepatocytes was noted also.There was no changes of plasma corticosterone level between group with acute cold stress and mice with Cryelan + acute cold stress (several fold increase comparatively to the control mice).As was suggested, the mechanism of protection can include Mph stimulation and secretion of some acute phase proteins responsible for positive effect of immunomodulators. New yeast polysaccharides Cryelan and Rhodexman can be used for macrophage stimulation,especially in pathological states. Immunomodulators were shown to increase production and secretion of lysosomal enzymes (like zymosan). Secreted enzymes,especially cysteine proteinasescathepsins B and L -involve in the process of inflammation;however, excessive release of these enzymes may lead to noncontrolled proteolysis followed by tissue degradation (Assfalg-Machleidt et al.,1990) .The effect of zymosan,BCG and new immunomodulator carboxymethylglucan (CMG), second fraction on secretion of lysosomal enzymes by murine peritoneal macrophages was studied. Zymosan increased the secretion of N-acetyl-~-D-glucosaminidase and ~-galactosidase into the culture medium (3-4 fold); BCG possessed similar effect.CMG in the same concentrations (50 /~g/ml) increased release of these enzymes only sAightly (1.6 times).It's known that zymosan-induced secretion reflects the enzyme release from formed lysosomes (Warren,1989) .It was suggested that CMG activated macrophages via interaction with scavenger-receptors,followed by weak secretion of lysosomal enzymes and as a result decrease of tissue damage. In vivo zymosan induced stimulation of mononuclear system of phagocytes followed by increase of cysteine proteinases activity in liver at the 5th day. In the same time in blood N-acetyl-~-D-glucosaminidase and N-acetyl-~-D-galactosidase activity increased 2-5 fold. It was concluded that in drug design it's possible to select such immunomodulators,e.g. CMG,which can activate mononuclear system of phagocytes and do not damage tissue. Endothelin-i (ET-I) is produced by injured/ ischemic endothelium, mobilizes intracellular Ca ++ and is a potent vasoconstrictor. It is also a Ca ++ agonist for anterior pituitary or renal mesangial cells and monocytes. ET-I causes monocytes to produce interleukin-l, 6, 8, prostaglandin E2, and substances which trigger neutrophil superoxide production. ET-I levels increase in shock and ET may play a role in activating leukocytes post shock causing reperfusion injury. But blood flow experiments suggest splanchnic circulation changes more profoundly in shock than peripheral circulation. We therefore asked if ET-4 (or VIC), the ET which predominates in splanchnic vessels, had any effect on monocyte cytokine production. Human monocytes from health~ blood donors were separated on Ficoll. 0.5 x 10Ucells/ ml in RPMI 1640 + 10% FCS were incubated i0 min., 6 & 24 hrs. with 10 -8 M ET-I, 10 -8 M VIC or I0 ug/ml of LPS. Supernatants were assayed by ELISA. We have shown that low dose endotoxin pretreatment (LPS 1 ) for 24 hrs markedly inhibits the macrophage (MO) release of tumor necrosis factor (TNF) and increases interleukin-1 (IL-I) in response to a subsequent endotoxin stimulus (LPS2). In this study we examined the kinetics of LPS 1 inhibition of TNF and augmentation ofIL-1. Methods: Murine peritoneal exudate MO from Balbc mice were exposed in vitro to medium or 10 ng/ml of LPS1 for intervals of 0 to 24 hours. Culture medium was then replaced with 0, 10 or 1000 ng/ml of LPS2 for 24 hrs. TNF and IL-1 in MO supernatants were measured by specific bioassays. During sepsis endotoxin (LPS) activates macrophages (MO) to release mediators such as tumor necrosis factor (TNF), interleukin-6 (IL-6), interleukin-I (IL-I) and prostaglandin E2 (PGE2). We showed that preexposure to LPS (LPS 1) alters the response of murine M~i to subsequent LPS stimulation (LPS2). We hypothesized that in vitro cytokine release by LPS2 in human monocytes (MO) is also be altered by preexposure to LPSI. Methods: Human peripheral blood MO were obtained from healthy volunteers (N=8), cultured in vitro 24 hrs, then pretreated 24 hr _+ LPS 1-Cultures were then stimulated with LPS 2 and mediators in MO supernatant measured: TNF, IL-I, and IL-6 by specific bioassays, PGE2 by immunoassay kit. Serum cytokine levels (specific ELISA kits) were compared to in vitro supernatant levels. Data is expressed as % control_+SEM, LPS2= 1000 ng/mh The table shows that all mediators were increased, in the absence of LPS 1. Pretreatment with LPS 1 resulted in complete inhibition of LPS2-triggered TNF release. In contrast, LPS 1 significantly increased MO secretion of IL-1, IL-6 and PGE 2 (data not shown). Serum cytokine levels were as follows: TNF 407_+21, IL-I 241+208, and IL-6 8.0-+2.5 ng/ml. These serum levels were low, showed an extremely wide variation, and did not correlate with in vitro LPS2-triggered mediator production. Conclusion: Human monoeyte mediator production is differentially regulated by preexposure to LPS 1. Provocative in vitro testing of monocytes may ultimately be clinically useful to identify prior in vivo LPS exposure or MO Macrophages release numerous secretory products involved in host defense and inflammation. Activated macrophages with cytokines produced have been implicated in tissue damage in sepsis and multiple organ dysfunction. Aimed to elucidate the organ-association phenomena,this study is to compare peritoneal macrophage(PM),alveolar macrophage(AM), and Kupffer cells(KC) during sepsis in terms of cellular protein contents as symbol of activation by flow cytometry analysis. Sepsis were produced by cecal ligatien and perforation (CLP) in Wistar rats weighing 210-290g.PM were obtained by peritoneal lavage,AM by bronchial lavage and KC by incubating the collegenase digested liver with pronase-E. Leukocytes have been implicated as a mediator of the microvascular dysfunction associated with reperfasion of ischemic tissues. A role for Ieukocytes is largely based on observations that rendering animals anutropenic with anti-neutrophil serum or preventing leukocyte adhesion with monoclonal antibodies attenuates the increased fluid and protein leakage from the vaseulature that is normally observed in postischemic tissues. We have recently undertaken studies designed to determine the relationship between leukocyte-endothelial cell adhesion and albumin leakage ia rat mesenterlc venules exposed ~o ischemia-reperfusion (I/R). Leukocyte adherence and emigration as well as albumin extravasafion were monitored in single postcapillary venules using iatravital fluorescence microscopy, lschemia was induced by complete occ!usion of the superior mesenteric artery and ~dl parameters were monitored at various intervals following reperfusion. The magnitude of the leukocyte adherence and emigration, and albumin leakage elicited by I/R was positively con-elated with the duration of ischemia. The albumin leakage response was also highly correlated with the number of adherent and emigrated leukocytes. Monoclonal antibodies against the adhesion glycoproteins CD18, CDllb, ICAM-1 and L-selectin, but not P-or E-selecdn, reduced I/R-induced leukocyte adherence and emigration as well as albumin leakage. PhaUoidln, an F-aetin stabilizer, largely prevented the emigration (but not adherence) of leukocytes and greatly reduced, the raicrovascular protein leakage. Plateletleukocyte aggregates were formed in postischemic vemdes; the number of aggregates was reduced by antibodies against P-selecdh, CDIlb, CD18, and ICAM-1, but not E-selectin or Lselectin. A significant fraction of the mast ceils surrounding the posteapillary venules degranulated in response to ischemia/repeffusion, but mast cell stabilizers did not afford protection against the albumin leakage elicited by I/R. These results indicate that reperfusloninduced albumin leakage is tightly coupled to the adherence and emigration of leukocytes in posteapillary venules. This adhesiomdependent injury response is primarily mediated by CDllb/CDI8 on activated neutrophils and ICAM-1 on venular endothellum, and appears to require L-selecda dependent leukocyte rolling. Mast cell degranulation does not appear to conwibate to the vascular pathology associated with I/R.

M.D. Rod=iek, Boston, MA, USA The polymorphonuclear neutrophil (PMN) has long been known to pa~tlcipats in the inflammatory reBpons~ as a phagocyte and killer Of invading organisms, but little attention has been given to its potential as a participant in the in~une interaction of lymphocytes and macrophages.

We and others have shown that the PMN may have i~m~/nomcdulatory effects both in vitro and in vlvo. More recently it has been proven that the PMN can make mRNA for and secrete the proinflammatory oytokines ILla, IL-IB, TNFS, IL-6 and IL-8 as does the other major circulating phagocyte, the monocyte/macrophags. Furthermore it has been shown to make the potentially autoregulatory oytokines GCSF and GMCSF. These functional capabilities suggest that the PMN is not an Wend cell ~, but one which has a potential role in regulation cf ~he immune response and that this potential ~cle should no longer be ignored when considering the immune abnormalities existing in patients following maJo~ injury or surgery. We have investigated the proinflaznmatory oytokine secretion patter~ by PMN in patients following major ~hermal or tra~matic injury and in volunteers fellowinq endotoxemia. ?ollowing major injury there is variable PMN secretion of these cytokines when stimulated in vlero. Following endotoxemia in a group of human volunteers PMN showed a hypo=esponsivenesa to LPS 4 hrs following endotoxin infusion followed at 24 hre by an overshoot. Pretreatment with steroids modulated this overshoot phenomenon, suggesting that receptors for steroids are involved in the regulation of cytokin® secretlon by FMN. These results sugges~ that the PMN, the most numerous cell in the circulation and the first to respond to an ins~l~ may be a so~rce of the prolnflammatory cytokine cascade following injury that has been recognized as significant in the process which often leads to multiple o;gan failure, The immunosuppresslon which occurs following major thermal Injury may predispose these individuals to infection and sepsis, which remain a significant cause of morbidity and mortality.

Included among the many Immune aheratlons are the p2 Integrln (CDlla, b,c/CD18) dependent activities of adhesion, chemotaxls, diapodesls, and phagocytosls.

Our Investigations indicate that, following major thermal injuries, the expression of the [~2 Integrlns, but not CD14, is significantly decreased on neutrophlls (PMNs).

It remains unclear if PMNs from thermally Injured patients respond normally to LPS, The effects of treatment in vitro with LPS and f-met-leu-phe (fMLP) on the expression of CDtlb was examlned on PMNs from the peripheral blood of healthy volunteers and non-septic burn patients (>30~; total body surface area, >lS~ full thickness), The PMNs were Incubated with LPS (]ng-10p.g/ml) or f'MLP (10 "1 1 to 10"6M) et 37oc for 30 mln, In 1~; human AB serum, The expression of the 132 ]ntegrins was detected using monoclonat antibodies and flow cytometry. LPS and f'MLP resulted In a slight Increase (3 fold) In the expression of CD11b on PMNS from burned patients compared to an 8 and 14 fold increase, respectively, on PMNs from healthy individuals. This Inability of LPS or fMLP to increase CD11 b expression was not due to the amount of LPS bound to the two cell populations. Because the same defect Is seen after either LPS or fMLP stimulation, It Is speculated that the defect must be in the amount of preformed CD1 ] b or Its transport to the plasma membrane. Platelet-activating factor (PAF) and neutrophils have been implicated in the patbophysiology of ischemia-repeffusion injury, In addition, PAF stimulates neutrophi[ (PMN) oxidative metabolism in vitro. The present study examined the potential role of PAF in repeffusion injury in an in viva rabbit model. Eight anesthetized rabbi~s underwent retroperitoneal exposure of the infrarenal abdominal aorta after percutaneous insertion of a catheter through the jugular vein into the infrahepatic inferior vena cava. Doppler flow probes were placed around the abdominal aorta and the right common femoral artery to assess flow through these vessels. An occlusive ligature was placed around the abdominal aorta (superior to the flow probe) at t = 0 and total occlusion of blood flow to the lower extremities was maintained for g0 mins., after which the ligature was released allowing for reperfusion of the ischemic lower limbs. Effluent blood from the ischemic hind-limbs was collected through the IVC catheter at the times indicated below and assayed for PAF by a direct radioimmunoassay. In addition, neutrophil H202 production was determined by a previously described 2'7'-dichlorofluorescein flowcytametric assay. 185 _+ 53 amean _+ s.e.m, pg/ml blood; brelative fluoresenee (% of baseline); Caortic and femoral artery flow (% of baseline); *p < 0.01 vs. baseline; 1"p < 0.05 vs. baseline. A significant elevation of PAF was observed in ischemic hind-limb effluent blood at 15 min. after release of the aortic ligature during the repeffusion phase, as compared to baseline levels. In addition, PMN H20 2 production was increased by 2.3-fold above baseline values by 1 hour after ligature release during the reperfusion phase. Both of these elevations were transient and returned toward baseline by 2 hours post-isehemia. Tatar occlusion of hind-limb flow was achieved as evidenced by the absence of aortic or femorat flow at 90 rain. post-ischemia, however after release 01 the ligature a significant reactive hyperemia was observed by 15 mln. into the rapeffusion phase. Histolog[c examination of reper[used gastrocnemius muscle revealed moderate PMN infiltration into the interstitium. In conclusion, these data indicate that PAF is released into the circulation during repeffusion, and is likely involved as a mediator in the observed PMN oxidative burst activity, thereby contributing to reperfusien injury. Following thermal Injury and Infection granulocyte function ts abnormal. To elucidate the mechanism by which thermal injury and Infection affect the granulocyte's ability to polymerize and depolymedze actin, we serially measured F-Actin levels in granulocytes from 14 burned patients (mean age 42,1 +_ 17.7 years, mean burn size 39.8% _+ 22.2%) during the first S weeks post injury. Six of the patients had infections during the course of the study, (septicemia, wound Invasion and pneumonia). Actin levels In granulocytes from eleven healthy volunteers (mean age 34 years) were measured repeatedly and served as controls. Lysecl white blood cell preparations were brought to 370C and incubated with N-formyl-met-leu-phe (STIM) or with Dulbecco's phosphate unbuffered sellne (UNSTIM). The cells were concomitantly stained and fixed with formaldehyde, lysoleclthln and fiuoresceln phafioidin. Actin depolymedzation (DEPOL) was measured by Incubating stimulated cells at 0°C before the stain-fixative was added. Baseline (BASE) F-actln levels were assessed by adding stsln-fixatlve to Icecold unstimulated cells. Fluorescence was estimated In a FACSCAN and expressed as Ilnesr mean channel fluorescence_+ SEM (MCF). Figure 1 displays granulecyle Fectln levels In Infected and uninfected patients as compared to controls. F-actln levels were consistently lower In control cells than In those from burned or burn-infected patients under all measured conditions. Granulocytes from Infected burned patients demonstrated a significant decrement In their ability to depofymerlze F.actin compared to both uninfected burned patients and controls, while there were no significant differences between Infected and 1,~ uninfected patients In the baseline, unstlmuleted and stimulated conditions. Those results Indicate la5 that grsnulocytas from burned and bum-Infected patients contain higher levels of polymerized actln than ~ ,2S control cells. In order to study tumor necrosis factor (TNF) receptor sensitivity in septic critically ill patients we investigated 20 blood samples of such people in reaction of leucocyte migration inhibition. Migration of their polymorphonuclear leucocytes (PMNs) was studied with stimulation with human recombinant TNF in concentration of 166.7 U/ml (recommended by manufacturer is the range of 10-200 O/ml) and without such. Ten healthy blood donors formed control group.

The results obtained showed diminished PMN reactivity to TNF in patients (migration inhibition was absent) oscaring with significantly increased migration ability of their PMNs (107.2% of that in control group). At the same time normal PMNs in control group did show migration changes upon TNF stimulation. Considering all the above we come to a conclusion that externally added TNF fails to activate PMNs in critically ill patients more than they are by their endogenous TNF. Moreover, this TNF no longer serves a positive chemotactic factor for such PMNs. These findings may suggest that in critically ill septic patients reactivity of PMNs to TNF is deeply altered. TNF receptors of PMNs are either exhausted as such by excessive stimulation with endogenous TNF or further transmission of their message is impossible due to "fatigue" of the cell's activation mechanisms. We express our gratitude to REANAL Factory of Laboratory Chemicals for generously providing us with a TNF com~rcial sample.

~-SANGUIS Medical, Ekaterineburg Russia; S-Urals Med.lnst. Activated neutrophils infiltrating the local site of inflammation following trauma release high amounts of destructive lysosomal enzymes into the extracellular space. Cytokines were discussed to be involved in regulation of this early process. The task of this investigation was to evaluate the possible regulatory role of interleukin-6 (IL-6) and its potential immunosupressive opponent, the transforming growth factor-&, in regulation of neutrophil degranulation. We analysed the concentration of the al-proteinase-inhibitor complex of the lysosomal elastase as marker for the degranulation of neutrophils as well as the levels of IL-6 and TGF-61 in the plasma probes of patients undergoing multiple trauma and severe surgeries. The time courses of IL-6 and elastase were found to be highly correlated, wheras the concentrations of the cytokine TGF-E~ were found to be not significantly altered in comparison to the control group. This close temporal correlationship was confirmed by investigation of fluids derived from sites of inflammation. Interstingly, the inhibitory potential (~zcproteinase inhibitor, antithrombin III) was dramatically reduced in the early inflammatory phase.

To prove this in vivo findings, the effects of IL-6 and TGF-I~ 1 on the degranulation of isolated human neutrophils of healthy donors was investigated in vitro. Pathological high concentrations of rhlL-6 up to 104U/ml (as detected in fluids derived from local inflammatory site) were found to be capable to induce a significant release of lysosomal elastase in a concentration-dependent manner, whereas the degranulation of neutrophils was uneffected by TGF-61. In conclusion, these data suggest a contribution of IL-6 in regulation of neutrophil activation at sides of inflammation. The immunosuppressive cytokine TGF-I&~ seems to have no direct regulatory effect beside its described chemotactic function on neutrephils. Postirradiation chan~es of adhesive properties arid supercoiled nucleoid DNA structure of blood leukocytes were studied in Macaca nemestrina andrats. The dynamics of membrane chan~es after nonlethal irradiation of rats demonstrated the temporary increase of the leukocyte adherence at 24 h followed by return of this parameter to normal levels at 48 h.

After lethal irradiation of both animal species the increase in adhesive leukooytes fraction was detected as early as at 3 h. This hi~her index persisted until the end of experiments (5 days).

The early (3-5h) temporary loosin~ of supercoiled DNA structure was demonstrated in the leukocytes of nonlethally irradiated animals. This phenomenon seems to be connected with the lymphocyte fraction chan~es.

This process was not dependent on altered adhesive properties of leukocyte membranes.

The membrane chan~es of leukocytes preceded decondensation of supercoiled DNA after lethal irradiation of animals, In this case loosin~ of supercoiled DNA pro-~ressively increased at 24 h and at the later terms of postirradiation period. The systemic inflammatory response syndrome (SIRS) involves many inanunological reactions of the host including acfivatinn of inflammatory mediator cascades and depression of cellular reactivity in T-lymphecytes (1). There are reports of nentrophil dysfunction in inflammatory disorders of the skin (2), Are there dysfunctions concerning the unspecific host defense in SIRS, as well? In this study, we examined the reactivity of neutrophil granolocytes from patients suffering from SIRS. Twenty-one patients (APACHE II-score 21±5) with diagnosis of SIRS entered the study. Granulocytes were prepared as reported previously (2) . In parallel, granulocytes from 20 healthy individuals were tested. Two granulocyte functians were studied in vitro: 1. Migration of the ceils in a boyden chamber through a filter matrix following stimulation with 5 different receptor dependent stimuli (C5a, Intefleukin-8, Platelet-activating-factor, Leukotrien B4, FMLP). 2. Release of 13glucuronidase following stimulation with the aforementioned activators.

The results demonstrate, that the release of 13-glucuronidase in patients suffering from SIRS was comparable to the enzyme release of granulocytes prepared from healthy individuals. Each stimulant induced release of p-glucuronidase in a characteristic dose dependent fashion. All granulocyte preparations from the healthy donors showed a positive chemotaxis response in the migration-assay. In contrast, only ten out of twenty-one patients had granulocytes migrating after stimulation. The two groups of patients displaying reactive or non-reactive granulocytes differed clinically: The nonreactive group consisted of patients with multiple organ failure (10/11) and nonsurvivors (8/11), whereas 8/10 patients in the reactive group survived. Thus, the in vitro chemotaxis of granulocytes is impaired in a subgroup of patients with SIRS. This defect of the non-specific host defense may contribute to poor prognosis and outcome of these patients. Dermatol. 85: 194-198, 1985 Klinik ffir An~isthesiologie und Operative Intensivmedizin der CAU Kiel, Schwanenweg 21, 24105 Kiel, Germany.

Objectives of the study: Major emphasis has been given to the analysis of interactions of antibiotics with microorganisms. Effects of antibiotics on cells of primary host defense mechanisms, such as the neutrophils, are less well known. Therefore, attention has been focused on clindamycin, a member of the lincoseamide family. Materials and methods: The effect of clindamycin (i -i00 ~g/ml) on 8 granulocyte functions (healthy volunteers) such as random migration, chemotaxis (agarose method), ingestion (radiometric assay), superoxide (cytochrom c reduction) and hydrogen peroxide production (phenol red oxidation), lucigenin-and luminol-amplified chemiluminescence (luminometry) and degranulation (turbidometry with Micrococcus lysodeicticus)

were investigated in vitro. Results: Motility and degranulation were inhibited, ingestion of Saccharomyces cerevisiae, zymosan-induced lucigenin-and luminol-amplified chemiluminescence, superoxide and hydrogen peroxide production were stimulated in a dose dependent fashion. Conclusion: Clindamycin has granulocyte function modulating properties. Recognition of immunomodulating effects of antibiotics may have therapeutic significance, especially in patients with long-term antibiotic therapy or immune deficiencies. The intense muscle activity (EA) of rats resulted in increase of neutrophil influx in muscles during the recovery. We investigated neutrophil proteinases involvement in neutral proteinases balance of skeletal muscles by NA. The rats were submited to swim with the load (8% of body mass) till exhaustion. Immediately after NA the neutrophil antiserum was injected i.p. to rats of experimental group. Saline was injected to control animals° Injections were repeated in 12 h of the recovery and cytosol proteolytic activity (pH 7.4; FITC-casein) was determined.

Isolated soleus muscles were incubated also in vitro and proteolytic activity of incubation media was measured. It was found that there was 2-3 -fold proteinases activity increase in cytosols of all investigated muscles (soleus, white and red portions of quadriceps) of control animals by 6 h of the recovery (the comparison was done with the sedentary rats). In 12 h cytosol proteolytic activity decreased and then increased again by 24 h of the fast. Antiserum injections resulted in relible decrease of the proteolytic activities at every investigated time. When incubating m. soleus in vitro the activities of proteinases in incubation media turned out reliably less if soleus muscles were isolated from the animals to which antiserum was injected. The conclusion is that neutrophil proteinases can be involved in the balance of rat skeletal muscle neutral proteinases after ~A. A lot and new clinical problems complicating the outcome of polytrauma, burn and septic patients in Surgical Intensive Care Units, have arisen as the care improvement prolonged the patient's survival: a progressive degradation of organ and system functions often develops, usually making its first clinical appearance by ARDS, followed by the other organ failure (MOF) and sepsis symptoms. The clinical picture is polymorphic, the end result of a complex systemic pathophysiological reaction trigg~ed off by trauma consequences (tissues disruption, hypo~xygenatiun and necrosis). Nowadays there is not a preventi~ or specific therapy to lower the mortality rate (70-80%) and-'mdy-a~ early, aggressive surgical approach .-evacuating haematomas, stopping bleeding, toileting all septic, necrotic foci and restoring anatomic continuity-, seems to be of some help This complex clinical entity has not an univocal denomination yet. The proper labelling of an illness should come from the full understanding of its pathopysiology and suggest the proper treatment choice. Clinical and experimental studies demonstrated that pathophysiologic mechanisms involved in the past-traumatic illness, share the same anatomo-pathological elemem: the interstitial edema, due to a generalised endothelial micro circulatory injury. This alteration, as constantly seen in polytrauma patients, develops in a few hours after trauma as a consequence of the deregulation of the homoeostatic and immune mechanisms. In fact the overproduced oxygen free radicals and r~ombinam cytokines (IL1,TNF), together with the complement degradation fragments, the proteolytic enzymes and many other mediators are all strongly h~l ~ ,_he e,,j,yheha! ceils. Our~osect, atim~,-bnsed on examination of autopsical specimens from 85 polytraanm patients, showed that such endothelial damage, supporting the interstitial edema, is widely and simultaneensly distributed, ensues shortly aRer trauma and shows its effects in different organs at different times, only because each apparatus has different fimctienal reserves: the lung is the first organ to fail just because its ah, celocapillary membrane is one of the most delicate bodily structure, and its function is irroplace~le. We think it will be of a great help, in planning a preventive therapy, to chose a denomination focusing the physician's attention on the earl)" generalized endothelial injury and its effects, as in trauma patients it is present -even if latenflysince the first few hours. We would like to see the generalised endothelial microcircolatory injury properly highlighted when considering the best definition and the optimal nomenclature for the post-traumatic s3mdrome. The presence of interleukin (IL)-8 in bronchoalveolar lavage fluid of critically ill patients correlates clinically with the development of the adult respiratory distress syndrome lARDS), and inhibition of IL-8 in animal models can attenuate lung injury. Collectively, evidence to date suggests that IL-8 attracts and activates neutrophiis (PMN), which are then responsible for the capillary leak of ARDS. However, an alternative explanation is that IL-8 is directly toxic to the endothelial cell (EC). In this study, we have hypothesized that IL-8 can disrupt endothelial integrity independent of PMN. Meth ods: Human umbilical vein (HUV) EC monolayers were cultured to confluency on collagen-coated micropore filters. To assess EC integrity, 1251.albumi n leak was quantitated by measuring the 1251 counts which crossed the monolayer, using a gamma counter. IL-8 (lpg/ml) was incubated in the culture medium with 1251.albumi n for 21 hrs. The IL-8 dose was not cytotoxic. To determine the involvement of protein synthesis in this process, selected monolayers were pretreated with cycloheximide (CH) prior to 11.-8 addition. Statistical analysis was performed using ANOVMFisher PLSD. We have previously shown that Platelet Activating Factor (PAF) enhances CDt8 expression and primes PMN's for subsequent 02generation. Both are important steps in PMN mediated injury and are assumed to occur in concert. Following major trauma non-specific PMN inflammation is activated, however, unbridled systemic PMN activity needs to be minimized. Since circulating catecholamines are high early post-injury, we hypothesised that they downregu/ate CD 18 expression and PMN priming via the [3 adrenergic signal transduction pathway. Methods: Normal human PMNs were primed with PAF (10ng/ml for 5 min) or pre-treated with 10-5M of Isoproterenol (I) or Forskoklin (F) for 5 rain and then primed with PAF. CD18 expression was measured by flow cytometry (Fig.l) and 02-generation in response to 10-rM fMLP was determined as SOD inhibitable reduction of cytochrome C ( Fig.2 Holler** and Georg W. Bornkamm* Lymphocyte-endothelial interactions are crucial for various immune responses, including cytokine driven inflammatory processes. Protein kinase C (PKC)-inhibitors on the other hand are discussed as potential cytokine antagonists. In the present study we investigated the influence of the PKC-inhibitor GF109203X on cytokine-and endotoxin induced expression of intercellular adhesion molecule 1 (ICAM-1) and on adhesion of lymphocytes to cytokine activated endothelial cells. We found that Tumor necrosis factor alpha (TNFo0-and Lipopolysaccharide (LPS)-induced ICAM-1 expression on human endothelioma celts (EAhy926) were unaffected by the PKC-inhibitor and thus appeared to be independent of PKC activation. In contrast, GF109203X significantly reduced ICAM-1 expression induced by Interferon-y (IFN-?) and Interleukin-1 (IL-1). The functional relevance of these findings was evaluated in an adhesion assay using human umbilical vene endothelial cells (HUVEC) and peripheral blood mononuclear cells (PBMC). In fact, the IFN-? and IL-1 induced adhesion of PBMC to cytokine treated HUVEC could be downregulated by the PKC-inhibitor, whereas TNFc~-and LPS-mediated adhesion was not influenced. Additionally, the IL-1 driven ICAM-1 expression on EAhy926 cells as well as the IL-1 induced adhesion of PBMC to HUVEC was found to be TNF-dependent, since both effects could be inhibited by an anti-TNF monoclonal antibody (195F) . These in vitro data further support the idea of examining PKC-inhibitors, such as GF109203X, for their biological relevance in cytokine related dysregulations. Seiffge, D., Bissinger, T., Laux, V., During inflammation there are some key processes, which occur in the microcirculation: The release of mediators from various cell types, the migration of inflammatory cells towards a chemotactic stimulus in the tissue, the expression of adhesion molecules on different cells, and the extravasation of plasma proteins. The aim of the present study was to elucidate the mediator induced interaction of leukocyte adhesion and plasma leakage in postcapillary venules. Using an analogous video-image analysing system we have studied the effect of different mediators on leukocyte adhesion and macromolecular permeability in the mesentery of the rat. The increase in permeability was measured as changes in optical density. We found that topical administration of leneotriene B 4 (LTB4, 5x10 "6 tool/l) or intravenous injection of interleuldn-2 (IL-2, 5-106 IU/kg b.w.) and lipopolysaccharide (LPS, 15 mg/kg b.w.) resulted in a significant extravasation of FITC-labelled rat serum albumin (FITC-RSA) in venules but not in arterioles. We could correlate the changes in vascular permeability with a locally increased number of rolling and sticking leukocytes in venules. Both effects were dose dependently inhibited by different drugs. Pentoxlfylline inhibits LPS-indueed FITC-RSA extravasation and leukocyte adhesion at a dose of 3 mg/kg b.w., superoxid-dismutase (SOD, 10.000 IU/kg b.w.) was able to decrease the LTB 4 effect, and the immuumodulating drug leflunomide (HWA 486) exerted inhibitory effects on IL-2-induced permeability at a dose of 3 mg/kg b.w.i.v. The obtained results demonstrate that LPS, LTB 4 or IL-2 induced extravasation of FITC-RSA is mediated by activated leukocytes and can be deminished following administration of different drugs. Platelet-endothelial cell adhesion molecule-i (PECAM-I), a member of the immunoglobulin superfamily, is constitutively expressed at high levels on the endothelial cell surface. In vitro data have suggested that PECAM-I functions as a vascular adhesion molecule, specifically in neutrophil transmigration across the endothelium. This current work is the first demonstrating the in vivo role of PECAM-1 in neutrophil migration. Blocking antibodies to human PECAM-1, in which the antibodies are crossreactive with rat PECAM-1, were able to block the movement of neutrophils into the rat lungs after IgG immune complex deposition. Furthermore, when human foreskin was transplanted into mice with severe combined immunodeficiency and the site injected with TNF-alpha, anti-PECAM-I blocked neutrophil emigration into the dermal interstitium.

It has already been established that neutrophil recruitment is dependent upon selectin mediated rolling, followed by firm adherence that is ICAM-1/ integrin mediated. These data suggest, for the first time, that a third endothelial adhesion molecule (PECAM-I) is involved in the coordinated recruitment of neutrophils in vivo. To test whether trauma causes generalized activation or priming of PMNs, CDI8 adherence receptors were measured with iinmunomonitoring in whole blood after LPS stimulation ex vivo. Anesthetized (Fentanyl) mongrel pigs (15-25 kg) were subjected to 40% arterial hemorrhage + soft tissue injury and after liar, resuscitated with all the shed blood + supplemental fluid. Blood was collected at 24hr intervals from unanesthetized animals with indwelling catheters, PMNs were counted, and LPS was added (0,1,5,10 I.tg/ml) ex vivo. After 3hr incubation at 22-24°C, %CD18 (+) PMNs were determined with FITC-IB4 and flow cytometry from mean channel fluorescence histograms.

49±8 # p<0.05 vs BASELINE * p<0.05 vs sham $p<0.05 vs no anesthesia These observations provide direct evidence for time-dependent changes in PMN priming following major injury because CD18 expression was depressed for at ]east 24hr after trauma relative to sham but by 72hr, was enhanced, relative to sham, and because fentanyl anesthesia at 72hr had a greater effect on CD18 expression in trauma vs sham. Neutrophil (PMN) adhesion to vascular endothelial cells (•C) is a key element in the inflammatory response and tissue injury. Inflammatory mediators such as LPS (exogenous) and TNF (endogenous) can promote PMN-EC interaction which is believed to be responsible for capillary leakage and subsequent organ injury. However, the mechanism of this injury remains unclear.We hypothesised that the mechanism of tissue injury is due to EC necrosis with release of toxic products and that activated PMN are responsible. Human PMN were obtained from healthy donors, separated by density gradient, and activated with LPS (50ng/ml), TNF(10ng/ml), and LPS/TNF(25ng/10ng/ml). Cultures of the human EC tine(ECV-304) were used as surrogates of the microvasculature, were exposed to either LPS, TNF, LPS/TNF and PMN activated with LPS, TNF, LPS/TNF and incubated for 6, 12, 18, and 24hrs. EC necrosis was assessed by a 51Cr release cytotoxicity assay. PMN activation was assessed by CD1 lb receptor expression and respiratory burst activity

6hr 0_+0 1.7-+ 1 0-+0 3.2_+ 1 0_+0 4.4_+ 1 0_+0 2.3_+0.9 12hr 0+0 3.24_3 0_+0 9.3_+3 0_+0 11_+3" 0+_0 12+--2.1" lghr 04-0 0.9_+2 0+_0 284-12" O:fO 25,12" 0~0 30+-9.5* 24hr 0_+0 4.14-4 0-+0 33+_3* 0_+0 30_+9* 0_+0 32_+10" Data = EC % necrosis mean_+SD Stats: Student's t-test with significance (*) set at P<0.05 vs control. ( Our previous studies have indicated that despite the increased cardiac output and maintenance of tissue perfusion, hepatoceliular dysfunction occurs during early sepsis. Nonetheless, it remains unknown whether vascular endothelial cell function (i.e., the release of endothelium-derived relaxing factor/nitric oxide) is depressed under such conditions and, if so, whether endothelial cell dysfunction also occurs at the microcirculatory level. To determine this, rats were subjected to sepsis by cecal ligation and puncture (CLP), following which these and corresponding shams received 3 ml/100g BW normal saline. At 5 hr after CLP (hyperdynamic sepsis) or sham operation, the thoracic aorta was isolated, cut into rings, and placed in organ chambers. Norepinephrine (NE, 2xI0 -7 M) was used to achieve near-maximal contraction. Responses for an endothelium-dependeut vasodilator, acetylcholine (ACh, via nitric oxide), were determined. In additional studies, the small gut was isolated at 5 hr post-CLP. After pre-contraction of blood vessels in the isolated gut with 5xl04 M NE, vascular responses to ACh (5x10 6 M) and an endotheliumindependent vasodiiator, nitroglycerine (NTG, 5xl0 -7 M), were determined. Total vascular resistance (TVR, mmHg/mi/min/100g) was then calculated as pressure/ perfusinn rate. ACh-induced relaxation (%, n=6/group) in the aortic rings were:

ACh lxl0 i~s, sT-in ~ ~ significantly at 5 hr post-CLP (i.e., increased *P(O 05 vs. Sham; N-4 per group. TVR) in the absence of any changes in NTGinduced relaxation (Fig. A) . Thus, the vascular endothelial cell dysfunction observed in the aorta in early sepsis also occurs at the microcirculatory level. 

Introduction: The cytokine-mediated adherence of leulcooytes to vascular endothelium is considered as an early step in the cascade of pathologic reactions culminating in the "systemic inflammatory response syndrome" (SIRS); The purpose of this study was to evaluate the influence of interleakin-1 on leukooyteendothelial cell-interactions and microoirculation in the liver after hemorrhagic shock by means of intravital microscopy. Methods: In anesthetized female SPRDrats Co.w. 200-230g) shook was induced by fractionated withdrawl of arterial blood within 5 rain and maintained for 1 h (MAP at 40 mm Hg, Cardiac output 50 % of baseline). Rats were adequately resuscitated with 60% of shed blood and twice the volume in Ringer's solution additionally. Following 5 h of reperfusinn (MAP >100 mm Hg, CO >120% of baseline) the microcirculation in liver lobules was examined by intravital fluorescence microscopy after labelling of leukocytes. Continuous administration of IL-lra (Synergen, Boulder, Colorado, 5mg/kg/h) was started at different time points in a randomized and blinded manner. The animals in group P (n=6) received the IL-lra as pretreatment beginning 5 min prior to shock induction. In the group T (n=6) the application of IL-lm started at the beginning of the reperfusion period with a bolus injection of 5 mg/kg and was followed by continuons administration of 5 mg/kg/h. The control group C (n=4) received equal volumes in NaC10,9%, the sham-operated group S (n=4) was not exposed to shock. Results: Macrohemodynamics were comparable in all shook groups. The increased percentage of permanendy adherent leukocytes after hemorrhagic shook (S: 9,1% +1,1%; C: 42,1% _+5,4%) was significantly reduced by pretreatment or treatment with IL-lra (P: 16,9% -+1,9%; p<0.001, T: 28,8% -+3,6%, p<0.01, ANOVA). Temporary adhesion of leukocytes was unaffected by application of IL-lra. Liver microcirculation measured by volumetric blood flow in liver sinusoids and sinusoidal diameters was impaired after hemorrhagic shock in all groups and was not affected (C: 31993iam3/s +40971um3/s, P: 32768llm3/s +4445}am3/s, T: 326211amS/s -+2998 lam3/s, S: 390201am3/s -+39041am3/s). Di.seu~sinn: The results demonstrate that permanent adherence of leukocytes to endothelium is in part regulated by IL-1. Pathological adherence could be reduced by application of ILlra, even given at die time of resuscitation. The effect of lL-lm on permanent adhesion is a specific event and might be caused by reduced expression of specific receptors on sinusoidal endothelial cens and leukocytes. Objectives of the study. The adhesion of activated neutrophils (PMN) to endothelial ceils (EC) and the concomitant production of reactive oxygen metabolites (ROM) initiates organ damage after trauma, sepsis, shock and organ reperfusion. Aien of this study was to investigate the effect on adhesion and ROM production of the highly water-soluble, membrane-permeable and physiological ascorbic acid (ASC). Materials and methods. Adhesion of PMN to nylon fiber (cell count) and simultaneous ROM production (chemiluminescence-CL-response) were measured up to 6 retool/1 ASC as well as adhesion, ROM production and EC damage (lllln-release from labeled EC) of endotoxin-activated PMN to cultered EC moanlayers. In an in vivo animal model (sheep with lung lymph fistulas) the effect of ASC (1 g/kg bw bolus, followed by 0.2 g/ kg-h infusion) on the endotoxin-induced (0.5 Ixg/kg bw) neutropenia (cell count), lung capillary permeability damage (lung lymph protein clearance) and ROM production of neutrophils (zymosan-induced CL response) was measured. Results. ASC scavenged ROM dose-dependently during adhesion of PMN to nylon fiber (p<0.0005 at 6 mmol/l ASC), adhesion itself was unchanged. During the activated PMN/EC interaction ASC scavenged ROM (p<0.025 at 6 mmol/l ASC) and reduced the adhesion dose-dependently (p<0.0005 at 6 mmol/l ASC); EC damage was also reduced (p<0.0005 at 6 retool/1 ASC). In the in rive model ASC increased the endotoxin-induced blood PMN decrease (p<0.05), decreased the protein clearance (p<0.05) as well as the zymosan-induced ROM production (p<0.03), indicating the ASC-mediated reduction of adhesion, ROM production and lung tissue damage processes. Conclusions. By in vitro and in rive experiments ascorbic acid reduced the adhesion-and ROM production-initiated tissue damage. Therefore, i.v. administration of ascorbic acid is recommended for oxidative stress-associated states after trauma, sepsis, shock and organ reperfusion. for neut rophi l-accumulat ion and activation. We investigated the influence of OR to the activation and the expression of LECAM-I and CDIIb,CDI8 on neutrophils and lymphocytes. Methods:

From blood samples (n=5) all white blood cells (WBC) and neutrophils (NC) were isolated and cultured. OR were produced via the xanthine oxidase/hypoxanthine system. After 0,5, 15,30,60 and 120 minutes a Giemsa-staining to determine the granulation of neutrophils (n: normal, r : reduced ) and a FACS-analysis with monoclonal antibodies detecting CDIIb,CDI8 and LECAM-I was performed. Results: under the influence of OR a degranulation of neutrophils starting at 15min was observed in WBC-cultures (n/r: 0min 91/9, 15min 76/24, 30min 68/32, 60min 52/48, 120min 49/51). These data were confirmed in the dot-plots of FACS-analysis. Only in WBC-cultures OR induced a significant increase of LECAM-I expression on neutrophils up to 30min followed by a decrease to normal values at 60min. LECAM-I on lymphocytes disappeared totally during the observed period. CDllb,CDl8-expression was not altered. Conclusion:Increased LECAM-I expression on neutrophils due to OR could enhance the 'rolling' of neutrophils along the endothelium which is a prerequisite for neutrophil sticking and migration.

Further OR are able to activate neutrophils without endothelium. These changes seem to be mediated by other WBC. Introduction. Multiple Organ Failure (MOF) has been hypothesized to be the result of an excessive uncontrolled autedestructive inflammatory response. Since the complement system is an important mediator and initiator of the inflammatory response, interruption of this cascade could theoretically lead to an attenuation of MOF. In order to test this hypothesis we evaluated the response of C5-delicient mice in a model of zymesan indt~ed MOF. Materials and Methods. 25 C5-deficient B2D10/OId and 25 C5-sufficient B2D10/New mice were used in this study. On day 0 all mice received an intraperitoneal injection with zymosan suspended in paraffin in a dose of 1 mg/g body weight. Between day 0 and 12, biological parameters (temperature, body weight and clinical condition) were measured daily and mortality was monitored. Clinical condition was assessed by blindly grading the degree of lethargy, conjunctivitis, diarrhea, and ruffled fur of each mouse on a two point scale (maximum score=4). On day 12 all surviving mice were sacrificed and relative organ weights of lungs, liver, spleen and kidneys (relative organ weight= (organ weight/body weight)x100) wore calculated. Earlier experiments with our model have shown a good correlation between histological organ damage and relative organ weights. Statistical analysis of biological parameter was performed using the Koziol curve analysis. Analysis was divided in an acute phase (day 0-4) and a late phase (day 8-12). Relative organ weights were analyzed using Wilcoxon's test and mortality rate using Fischor's exact test. Results. All zymosan injected mice showed a typical triphesic illness. Deterioration of the clinical condition as indicated by the symptom score and the decrease in temperature and body weight in the acute phase were all significantly lass severe in C5 deficient mice (all p<0.005). In the late phase no differences could be noticed in the courses of biological parameters. Overall mortality was 2/25 (8%) in C5 deficient mice and 8/25 (32%) Jn C5 sufficient mice (p=0.049), a difference mainly due to a difference in the acute phase. Organ damage assessed as the relative organ weights did not show any statistical differences for any organ between both strains. Conclusion. Complement factor C5 appears to play an important role in the acute hyperdynamic septic response in this model but deficiency of C5 could not prevent organ damage in the late MOF phase. This suggests that other factors could be more important in the development of the inflammatory response leading to MOF. Proinflammatory cytokines are thought to play a critical role in the pathophysiology of Multiple Organ Failure (MOF). In mice, Zymosan-lnduced Generalized Inflammation (ZtGI) leads to MOF. Therefore we performed a sequential study into plasma levels of, and macrophage production capacity for, four cytokines during the development of MOF in the ZIGI model. Male young-adult C57BL/6 mice received zymosan (1 mg/g body weight) intraperitoneally. Groups of 14 animals were killed after 3, 6, 8, 18 and 24h and subsequently at each day until day 12. Plasma was collected and peritoneal macrophages were isolated and cultured overnight with or without lipopolysaccharide (LPS). Interleukin 1-ct, and -6 (IL-lc~,~,), and tumour necrosis factor-o~ (TNF-c 0 were measured in plasma and culture fluid by means of a RIA (detection limit 0.1 ng/ml). Interleukin-6 (IL~) levels were assayed using the B9 hybddoma cell proliferation assay.

Zymosan induces a three-phase disease in mice. After an acute phase the animals recover. Around day 7, they start to develop clinical signs which resemble MOF.

Plasma TNF-~ peaked within 3h after zymosan injection and disappeared within 8h. From day 8 onwards, TNF levels started to rise again. Plasma IL-6 behaved almost similarly in the acute phase, but in the MOF phase plasma IL-6 remained low. No circulating IL-16 could be detected at any time point. Macrophage LPS-stimulated production of IL-lcq IL-11~ and TNF--c~ was suppressed immediately after zymosan injection. Production of IL-16 and TNF-~ was normalized within 6h, while production of IL-lc~ remained lower than that in macrophages from untreated control mice. Only at day 6 did production of IL-I~ reach control values. IL-16 production was higher than control values from day 3 onwards. IL6 production was similar to that of ILI-IL The production of TNF-ct was strongly elevated between days 1 and 6 and again during days 10 to 12.

The development of MOF-like symptoms during ZlGI in mice is accompanied by increased plasma levels of TNF-ct without enhanced IL-16 or IL-6. Also, the ability of macrophages to produce excessive amounts of IL-16 and TNF--~, as well as the suppressed capacity to produce IL-lcq could be important mechanisms in the pathophysiology of MOF. When conjugated to an asialoglycoprotein, DNA and oligonucleotides are specifically taken up by the hepatocytes via the asialoglyccprotein receptor which is unique to the liver. Human asialoglycoprotein (~1-acid, ASGP) was derivatized with low molecular weight poly(L)lysine(PLL) and complexed with 5 antisense DNA's (AS) complementary to the 5' region of the IL-6 gpl30 receptor. The antisense were 5'-AGTTTAGGGATGAGG-3' (ASl), 5'-ATCTTCATCTTCTGAAT-3' (AS2), 5'-AAGTGAATGATTAAAACACT-3' (AS3), 5'-AAACCTTTATAGGCG-3' (AS4), and 5'-CGTTCTACAACTGCAACGT-3' (AS5). Using HepG2, the biological effects of these antisense complexes on the high affinity IL-6 receptor were evaluated by Scatchard analysis, cellular proliferation, and acute phase protein expression by radioimmunoprecipitation and two dimensional gel electrophoresis.

Scatchard analysis demonstrated that high affinity receptor expression was inhibited by incubation of cells with ASGP-PLL-ASI For 24 h. Underivatized ASl was less effective and the complex, ASGP-PLL-AS2, had minimal effects on high affinity binding. When the cells were treated with the conjugates and stimulated with IL-6 (i00 units) ASGP-PLL-ASI alone showed a dose dependent (0.i-0.4 ~M) inhibition of 3SS fibrinogen synthesis. Two dimensional gel electrophoresis showed that expression of other acute phase proteins was also blocked.

These results indicate that the targeted delivery of antisense molecules via conjugates recognized by the asialoglycoprotein receptor can block the cytokine stimulated acute phase protein response in hepatocytes, This approach may be relevant to the therapeutic management of patients with severe injury and sepsis. It has been established that immune cells are able to express neuropeptide genes and to release products that were considered to be of neuroendocrine origin. We have shown that Proenkephalin (PENK), a neuropeptide encoding gene, is expressed in lymphoid cells in culture. To study the physiological significance of these observations we have used the model of experimental endotoxemia. In this model, a disease state is induced by bacterial lipopolysaccharide (LPS), that activates the immune system, the adrenocortical axis and the nervous system. We found that the expression of PENKmRNA is markedly enhanced in vivo immediately after LPS injection both in the adrenal glands and in the lymph nodes. In situ hybridization analysis combined with immunohisto-chemistry indicated that the induced PENK expression is confined to macrephages within the lymph nodes and chromaffin cells in the adrenal medulla. Furthermore, this expression in lymph nodes is modulated by ligands of the adrenergic system.

Our results strongly support the notion that immune derived opioids participate in the bidirectional communication between the nervous and immune systems.

of Neurology Hadassah University Hospital, Jerusalem 91120, Israel. Objectives of the study: Multiple-organ-failure is recognized as the most severe, and often lethal, complication after multiple trauma. However there is no adeqate animal model available. Our goal was to develop an animal model, in which reproducable irreversible failure of parenchymal organs is achieved in the late phase after insults in the early phase (trauma).

Materials and methods: l0 female Merino-sheep were included (mean weight: 30kg). day 0: hemorrhagic shock (mean arterial pressure (MAP) 50mmHg for 2 hrs.), closed femoral nailing (AO-technique), day 1-5: bolusinjection of endotoxin (ET) (0,75~tg/kgBW) und zymosan-activated plasma (ZAP) (20ml) every 12hrs., day 6-10: observation. Bronchoalveolar lavage (BAL): day 1, 6, 10. The course of representative parameters of organ function was documented: COcardiac output (I/min), SVR -systemic vascular resistence (dyn ~ s cm-5), PAP -putm.art.pressure (mmHg), PaP2 -arterial oxygen pressure (mmHg), Bill -Bilirubin (;xmoV1), CrCI -Creatinin Clearence (ml/min) Statistics: Data as means+SEM, *significant from baseline (Wileoxon Test; p<0005) Results: baseline day 4 day 6 day 8 day 10 heart: CO 4,96_+0,32 5,33_+0,36 5,4_+0,2 7,30_+0,77* 7,98_+0,66* SVR1520_+134 1395+141 1403_+89 1119+_122" 1089+-91" lung: PAP 17,0_-20,7 19,6_+1,6" 22,0+-1,2"25,8+2,0" 28,8+-1,7'

PaP2 103,1+1,5101,5+-5,7 97,3_+4,7 91,9+-5,6 89,8+_3,9* liver: Bill 2, 94_+0, 344, 82_+1, 34' 5, 13_+0, 68'6, 13_+0, 7" 7, 19_+0, 91" kidney:CrCl 86, 5+_30, 9 109, [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] 4 74, 7_+10, 868, 8+14, 1 53, 1+17, 6 Histologic specimens showed all signs of fulminant MOF. Combination of hemorrhagic shock, femoral nailing, ET und ZAP (insults in the early phase) lead to an irreversible organ failure in the late phase. Prostaglandin E (PGE) levels are elevated by trauma, shock or sepsis and can profoundly affect the immune response. PGE2 is produced by many cell types including fibroblasts, macrophages, monoeytes, follicular dendritic cells, and epithelial cells and is induced by IL-I, bacterial LPS, components of the complement cascade, TNF, IL-6 and crosslinking of surface Fc receptors for IgG, IgA and IgE. Our research has shown that PGE inhibits B cell activation (specifically enlargement, class II ~C and Fc~ RII expression), proliferation, IgM and IgG3 responses, T cell proliferation, and IL-2 synthesis in the mouse model. In contrast, PGE greatly promotes class switching to IgE,the isotype responsible for Type I allergic hypersensitivity. Thus, our model mirrors th~ general immunosuppression and elevated IgE titers of the trauma or sepsis patient. PGE increases the number of cells secreting IgE and IgGl, acts on surface IgM positive B cells, synergizes with IL-4 and LP$ to induce preswitch germline transcripts, and induces more rapid expression of mature VDJ~ mP~A than in eontro~ PGE intracellular signalling occurs through cyclic adenosine monophosphate (cAMP) levels and can be mimicked by cAMP-inducing agents and blocl~ed by an inhibitor of cAMPdependent protein kinase A. PGE action requires de novo protein synthesis and candidate PGE-inducible regulatory proteins have been identified by 2D gel eleetrophoresis. Thus, PGE inhibits a number of immune mechanisms while promoting IgE production. A deeper understanding of PGE immune regulation may lead to more effective treatment of immune perturbations as sequelae of trauma, shock or sepsis. During infrarenai aortic surgery mesetueric traction (re.t.) results in prostacyclin (PGI:) release and consecutively in hemodynamic disturbances (decreased systemic vascular resisteace, mean arterial pressure; increased cardiac output, heartrate). These symptomes are bypassed by cyclooxygenase inhibition. Hemodynamic symptoms vanish after 30-45 rain even without cyclooxygenase inhibition although PGI2 levels remain elevated. To study the endocrine vasopressor system in a prospective double blinded protocol, we investigated 26 patients undergoing major abdominal surgery as compared to 26 ibuprofen (400 rag, i.v.) pretreated (ibu) patients. The surgeon applied m.t. in a uniform fashion. We chose a general anesthesia combined with a supplemental thoracic epidural anesthesia. At the points in time 5, 15, 30, 45, 90, 180, 360 , 1440 rain after and before (to) mesentzrie traction we determined the plasma concentrations (pc) of 6-keto-PGF~o~pr~, epinephrine, norepinephrine, dopamine, renin, aldosterone, ADH and cortisol. Pc of 6-k-PGF~,tp~, peaked 5 minutes after m.t. (2274_+284, ibu: 102_+12, to: 60+I ng/l) and declined monotonously over 6 h (203 +_42, ibu: 84_+ 12 ng/1). Catecholamine pc "s did not exceed the reference range during the observation period. Reninpc peaked after 45 rain (66_+23, ibu: 18+3, to: 16-+2/~U/ml); aldosteronc also presented a maximum after 45 rain (110 + 12, ibu: 66-+ 15, to: 56 +-7 pg/ml), whereas cortisol demonstrated irrespectively of circadian rhythms a maximum 6 h after m.t. (26+_1, ibu: 24-+3, to: 10+_1 ~g/1). ADH pc peaked 5 min after m.t. (71+7, ibu: 20-+6, to: 5+_1 pg/ml) and showed analogously to 6-k-PGFt~j~ pc a monotone decline over the observation period. Our data demonstrate a counteractive reaction to PGIz mediated vasodilation via ADH secretion. The second regulative is the renin-angiotensin-aldosterone system (RAAS), which is activated 45 min after m.t., the aldosterone pc does not paratlel the cortisol pc, which peaked post operafionem in both groups, probably due to the end of anaesthesia. A regulative release of catecholamines could not be documented. The activation of ADH and RAAS after MT is not a hormonal response primaryly related to surgical trauma and/or stress but a counterregulation to systemic vasoditafion induced by prostacyclin. Although ADH and RAAS support systemic circulation, angiotensin and vasopressin may compromise local organ blood flow (e.g. splancimic vascular bed).

Insfitut f. Klin. Chemic, Anaesthesiologie ~, Chirurgie l*, Univ. Ulm, 89070 Elm,

Expression of c-fos protein in rat brain following occlusion of superior mesenterie artery. Takanobu There is general agreement that neurologic abnormalities are seen in sepsis.

The aim of this study is to examine what effect does the brain receive in case of SMA occlusion by immunohistochemistry using antibody to c-fos, an immediate early gene, which is recently recognized as a genetic marker of activated neurons.

Moreover, we investigated the correlation between c-fos induction in the brain and plasma endotoxiu level. 20 rats of them received SMA clipping and others wee used as control.

Control and treated rats at 2, 4, 6, 8 hours were perfused and fixed. The brain were sectioned at 20 pm and stained by ABC method using c-fos antibody. Plasma endotoxin level of 5 rats were measured at 0, 2, 4, 6, 8 hours after the treatment by chromogenic limulus method.

Immunohistochemical study showed scarcely no immunoreactivity in control rat brain.

In treated rat brain, the significant expression of c-los was detected in specific nuclei including the habenula, some hypothalamie nuclei, amygdala, locus ceruleus and nucleus tractus solitarii.

Such immunoreactivities were increased in time curse, which well corresponded plasma endotoxin levels. The mean plasma endotoxin level of 0, 2,4, 6, 8 hours after the treatment were 6.48 ± 5.57, 15.0 _-4-4.37, 10.0 _+ 4.18, 14.4 ± 3.52 and 58.4 ± 28 .6 pg/ml, respectively.

The results indicate that limbic and hypothalamic-brainstem systems are involved in SMA occlusion, and suggest that such neuronal actival.jon may precede the elevation of plasma endotoxin Icy.el. systemic vascular resistance and increased cardiac output accompanied presumingly by a increased pulmonary shunt (Qs/Qt). This response is induced by prostacyclin (PGI2). We examined oxygen transport after traction on the mesentery root and the transpulmonary prostacyclin levels in a prospective placebo controlled study with intravenous ibuprofen. METHODS: With approval of the Human [nvestigadon Review Board we studied 52 patients in a prospective, randomized double-blinded protocol who were scheduled for major abdominal surgery. Ibuprofen (400 mg i.v.) or a placebo equivalent was administered 15 minutes before skin incision. Pulmonary artery thermodilution and radial artery catheters were placed after induction of anesthesia. MT was applied in a uniform fashion. Baseline values preceded the incision of the peritoneum (tO). Fulther assessments followed 5, 15, 30, 45, . Tile plasma concentrations (pc) of 6-keto-PGFt, (stable metabolite of PGI2) were determined in arterial and mixed venous blood by radioimmunoassay. At all points in time we measured arterial and mixed venous blood gases. Qs/Qt was calculated by standard formula. Data are given as median (p < 0.05 placebo vs. [ibuprofen] [117] mmHg (*p< 0.0i). These changes were accompanied by a marked increase of 6-keto-PGF~ pc up to 180 rain after MT in arterial and mixed venous blood of untreated patients with a peak of 2450*[60] ng/l tl (*p< 0.00Ol). There was no difference between arterial and mixed venous pc. Ibuprofen pretreated patients (n=zr) demonstrated stabile Qs/Qt and paO2 while 6-keto-PGF~ pc remained within the normal range. DISCUSSION: Our data clearly indicate that mesenteric traction response includes a critical rise in Qs/Qt followed by significant decrease of paOv Stable oxygen transport determinants following cyclooxygenase inhibition signify an action mediated by prostacyclin. An indicative transpulmonary gradient for 6-keto-PGFt~ was not detectable. A splanchnic vascular source for PGI2 release seems to be likely, but could not be proved by our current data. Department of Anesthesiology, Cliu. Chemistry * and Surgery*; University Clinics UIm, Prittwitzstral]e 43, 89075 Ulm, Germany It is unclear whether injuries like bums, in general, directly result in alterations of cell-mediated immunity that, in turn, promote endotoxic and bacterial translocation or, alternatively, whether these conditions allow increased bacterial invasion that, in turn, inhibits CMI. Aim: To determine whether infectious challenge, as CLP alone or combined with TI causes further immune abnormalities in the days following CLP. Study Plan: On day 1, two groups of n=36 8 week old AJ mice were subjected to either a 20% scold burn (TI), or were untreated (C) n=18. On day 10, 18 mice (TI+CLP) and 24 mice (CLP) were subjected to CLP. The two other groups (TI and C) were untreated. At days 11, 12 and 13 after thermal injury splenocytes (SP) were harvested and cultured with ConA for an assay of IL-2 and adherent splenocytes (AS) were cultured with LPS for IL-1, TNF, IL-6 and PGE2. Results: Either TI + CLP or CLP alone result in significantly decreased secretion of all cytokines tested. In the TI group almost every cytokine production determined was elevated in comparison to TI + CLP and Prosmcyclin (PGI2) has been implicated in the pathophysiology of septic shock. However, PGI~'s role in the inflammatory response to sepsis is not well-defined. The purpose of this study was to identify which acute septic events are mediated by PGI 2 during graded bacteremia. METHODS: Eleven ~nesrhetized, hemodynamically monitored adult swine were infused IV with Aeromonas h. (109/ml) at rates increased incrementally from 0.2 to 4.0 mI/kg/hr over 4 hours. Animals were studied in two groups: septic control (SC), graded bacteremia only (n=6); PGA (n=5), graded bacteremta plus anti-PGIz antibody, 50 ml/hr IV, beginning at 2 hours. Mean systemic (MAP) and pulmonary arterial (PAP) pressures and arterial pO2, mmHg, cardiac index (CI), L/min/M2, oxygen delivery index (DO2I) and consumption index (VOzI), ml/min/M2, and oxygen extraction (ER), %, )latelet aggregometry (PLT), %max., plasma PG6-keto F1 alpha ; In the first instance~ peak values of LT ~4 after i~ hrs post infarction were 6 times higher than in the controls and excess leucocyte infiltration was noted at the infarction zone. In second instance two levels of LT B4 led to weak infiltration of the infarction zone by leucocytes.

A. Mo~e~o, In~.~P~siolo~,D~t.E~.Cardiolo~,Bogotsolets4, ~ev252024, Ukrmne

Systemic lesion$of erythron in traumatic disease and possibilities of their regulation by opioid peptides.

Redkin Y. V., Fominih S. G.

Using clinical (121 patients) and experimental material(128 rats and 74 dogs) we revealed general regularities of erythron lesions after hard mechanical trauma of various genesis as well as some mechanisms of development of posttraumatic anemia and possibilities of its correction with preparations of opioid peptides. The condition of central and peripheral compartments of erythron was studied with unified morphologic, immunogematological, biochemical and radiological methods. It was revealed that irrespective of the experimental animal species (dogs, rats) or in clinical experiments (patients) and irrespective of the injuring factor type (skeletal trauma, craniocerebral trauma, loss of blood) in erythron can be observed one-directed unspecific reaction realized by the considerable lowering of hemoglobin concentration, erythrocytes number and hematocrit. In the initial period (1-7 days) in the system of erythron prevail processes of distraction and elimination of er~zthrocytes relatively to the general production of stimulated erythropoiesis. The primary alterating factor is the prolonged intensification of peroxydation of membrane iipids of erythrocytes with simultaneous lowering of reserves of reduced glutathione. The distraction of erythrocytes is supported by the developing phenomena of autoallergization of organism that becomes apparent by the appearance of sensitized T cells and antierythrocyte antibodies. The intensified production of erythropoietin rules to the realization of he program of fetal and terminal (reserved) erythropoiesis. Failure of erythropoiesis function is supported by disturbances of the processes of the injuring of cell metabolic apparatus. Using of dalargin (15 microgram per kilogram of body mass intrap'eritoneally within 5 days after the trauma) showed the precise pharmacotherapeutic effect revealed by the diminishing of anemia of experimental rats, more . Fiberbronohoscopic procedures are known to produce "PEEP-like" effects and to increase pulmonary artery (PA) resistance [2] . PEEP can affect RV function by reducing preload and ejection fraction (EF) [3] . Since changes of RV function during bronchoscopy in septic patients are not reported, we measured RV parameters before, during and after fiberoptic bronchoalveolar lavage (BAL). Method: This 34-year-old patient (APACHE-II:24) developed a hyperdynanlic septic state due to Staphylococcus aureus (blood culture). We inserted a "fast response" thermistor PA-catheter (Baxter-Edwards) to evaluate RV performance [4] . The therapeutic procedure included volume replacement, vasopressors (dopamine 5, dobutamine 3 gg/kg/min. IV) and analgosedatior/. Before bronchoscopy (Olympus BF-30, OD=6 mm) the patient received pancmonium for muscle relaxation. Ventilation was not changed during the procedure (endotracheal tube: ID=8 ram, Bennett 7200a, pressure controlled mode, Pm~x=25 mbar, PEEP=8 mbar, I:E=I:I, FiO2=1.0). We measured RV enddiastolic volume (EDV), stroke volume (SV), EF, heart rate (HR), cardiac index (CI) and mean PA pressure (MPAP Gerlach H, Gerlach M, Clauss M, Falke KJ Renal hypoxia and/or ischemia initiates the development of a deteriorated medullary perfusion based on fibrin deposition in the peritubular capillaries, vasoconstriction, and perivascular edema, which is followed by a swelling of the tubular epithelial ceils, intraluminal tubular obstruction, and a backleak of fluid through the injured tubules into the renal interstitium, finally leading to an acute tubular necrosis (ATN) [1 ], clinically diagnosed as acute renal failure (ARF). One important pathway for induction of enhanced vascular procoagulant activity and permeability is based on the synthesis and expression of macrophage-derived cytokines, which bind to specific endothelial cell surface receptors. We recently described the identification and purification .of a new 55,000 Dalton polypeptide, which is synthesized and expressed by murine macrophages after stimulation with lipopolysaccharide, and exerts procoagulant activity on cultured endothelial cells [2] . In the presented study, we demonstrate that the new polypeptid is also synthesized by macrophages under hypoxic conditions. The protein binds to specific receptors, which are expressed by endothelial cells dependent on the environmental oxygen tension. Animal studies were performed after approval by the local committee for animal safety; the animals were anesthetized, treated and supervised in accordance with the guidelines of this committee. In contrast to other authors, who performed long-term hypoxia experiments in awake animals, we preferred to implement the studies under anesthesia for ethical reasons, although regulatory functions for ventilation might be influenced. Animal studies demonstrated that the intravenous injection of the polypeptide initiates fibrin formation in the peritubular vessels. Keeping the animals under hypoxic conditions induces similar effects, which are reduced by a rabbit-antiserum against the new protein. In conclusion, the new polypeptide obviously contributes to the pathogenesis of acute renal failure by tubular necrosis during and after hypoxic events. The use of verapamil as cardioprotective agents for management of patients with acute ischemic/reperfused heart is based on the assumption that the increased intracellular Ca+ level is a key factor in causing cell death. Our in vitro study was designed to focus on effects of verapamil on the metabolic potential of cardiac slices after reversible ischemia in rats. The material consisted of two main groups : group A (non ischemia/reperfusion group) and group B (ischemia/reperfusion group), Each is subdivided into two subgroups (a and b). Each subgroup included 10 rat hearts. Group Aa is the control group, Group Ab is verapami] added group. Group Ba is ischemia group without verapamil. Group Bb is verapamil added group. Ischemic cardiac slices were obtained from rats subjected to 15 min. haemorrhage to induce reversible global ischemia. Both nonischemic and ischemic cardiac slices were placed in well oxygenated krebs Ringer phosphate buffer containing 150 mg% glucose & 5 gm% bovine albumin and incubated in Dubnoff shaking water bath for 60min at 37°C The results revealed that there was an enhancement in release of free fatty acids (FFA) (240%) and lactate (163%) and in glucose uptake (160%) in group Ba as compared with group Aa. These metabolic alternations produced by ischemic cardiac slices were reversed by verapamil addition (200 ml%) but in group Ab verpamil did not alter the release of FFA & lactate from non-ischemic cardiac slices, whereas it inhibited glucose uptake from these slices by 67%. The improvement of the metabolic intervention of ischemic myocardium indicates that verapamil may be of importance in reducing the extent and severity of acute myocardial ischemic injury in acute haemorrhage. Severe endothelial dysfunction occurs following injury to carotid arteries which is characterized by a decreased ability of these arteries to dilate when challenged with ACh or A23187, but not with a direct vasodilator (NaNO2). This failure to relax to ACh and A23187 reflects an inability of endothelium to generate EDRF, but relaxation recovers gradually to control values by 2 weeks. Exogenous NO donors (e.g., C87-3754 or SPM-5185), accelerate the recovery of the injured endothelium in rat carotid arteries. Intravenous infusion of an NO donor (30 p.g/day) with an implanted osmotic pump significantly accelerated the recovery of regenerated endothelium to produce EDRF at 7 days. Rat carotid artery rings relaxed only 17 + 5% and 15 + 5% to 10 gM ACh in vehicle treated rats and in inactive NO donor treated rats respectively 7 days following injury compared with 67 + 6% in NO donor rats (p<0.001). Relaxation to 100 gM NAN02 was normal in all groups indicating that the differences in relaxation were not the result of damage to vascular smooth muscle. Contraction to L-NAME (1 mM) was markedly reduced by injury, but was protected by NO donors (p<0.01). Thus, exogenous NO donors enhance the ability of the endothelium to regenerate and to release EDRF in response to endothelium-dependent vasodilators. This may be due to an anti-proliferative and anti-mitogenic effect of NO on vascular smooth muscle cells, allowing the endothelium to regenerate without intimal thickening. NO also has been shown to inhibit platelet aggregation, and to attenuate neutrophil adherence and activation. The superoxide scavenging effect of NO is not the basis for these effects since hSOD is inactive in preserving endothelial function in injured arteries. Thus, NO exerts a variety of cytoprotective effects which may be of importance in protecting against vascular injury. Much evidence has now accumulated to show that the excess production of the vasodilator nitric oxide (NO) in sepsis is an important contributor to the hypotension and multiorgan failure characteristic of this condition. Various cytokines play an important role in this process through their ability to induce the production of one of the enzymes responsible for NO synthesis, the inducible NO synthase (iNOS). We have studied the effects of cytokines on the induction of this enzyme both in vitro using vascular smooth muscle cells, and in a murine model of Gram-negative sepsis.

tn smooth muscle ceils, the cytokines IL-1, IFNq', and TNF-oc show strong synergy with one another in the production of iNOS. In order to define the molecular basis for this synergic effect, we have linked the promoter of the iNOS gene to a "reporter" gene, chloramphenicol acetyl transferase (CAT), and transfected these constructs into vascular smooth muscle cells. Assays of CAT activity reflect the activity of the promoter in this system, and by generating sets of deletion mutants of the promoter sequence we have been able to define the area within the promoter which mediates the synergic effect of these cytokines. In addition to stimufatory effects on iNOS production, certain cytokines are able to down-regulate the production of iNOS in vascular smooth muscle cells, and the effects of these counterregulatory cytokines will be discussed. The interaction of these cytokine effects in the whole organism has been studied in a murine model of Gramnegative sepsis. Widespread induction of iNOS occurs in this model as assayed by enzyme activity and through use of specific antisera to iNOS. Neutralizing antibodies to TNF-~ and tFN-y are both able to prevent death in this model, but it is only the anti-IFN-y which attenuates the induction of iNOS assayed in the liver. Clearly there is some redundancy in the effects of cytokines on the production of iNOS in sepsis, and greater understanding of the most important factors in iNOS production is required in order to target anti-cytokine therapy most appropriately. 

Effects of nitric oxide on hepatocyte metabolism in inflammation. J. Stadler, Department of Surgery, TU MQnchen, FRG Hepatocellular nitric oxide (NO) synthesis is induced by proinflammatory mediators such as tumor necrosis factor, interleukin-1 and interferon gamma or by bacterial toxins such as lipopolysaccharide. Stimulation of the hepatocytes (HC) with a combination of these agents leads to an output of NO in quantities which are not seen in any other celltype. It has been demonstrated by various investigators that important effects of these cytokines and bacterial toxins on HC metabolism can be attributed to the action of NO. In contrast to other celltypes HC seem to be relatively resistant to suppression of basic metabolic functions such as energy metabolism by NO. Therefore, cell damage has not been described to a significant extent following exposure to NO. However, NO does inhibit total protein synthesis. The exact biochemical mechanism of this phenomenon has not been uncovered yet, but it has been demonstrated for some specific proteins that their production is inhibited at a posttransscriptional level. As in many other celltypes cGMP generation is elevated in HC by NO through activation of the soluble guanylate cyclase. Cyclic GMP may possibly exert a plethora of metabolic functions, but it is interesting to note that most of the cGMP seems to be transported out of the cell. Some very specific effects of NO on HC metabolism include the inhibition of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and the cytochrome P450 (CYP) enzymes. Inhibition of GAPDH activity is mediated through nitrosylation of critical domains of the enzymes by NO which enhances auto-ADPribosylation. This effect on GAPDH activity might be responsible for the inhibition of gluconeogenesis by NO, which has been described recently. Finally, NO-mediated inhibition of CYPs may help to explain the suppression of hiotransformation processes which is a characteristic featur,'~ r ~ "~flamed liver.

Nitric oxide (NO) is an endogenous inhibitor of polymorphonuclear leukocyte (PMN) adhesion which limits PMN-endothelial cell interactions under normal conditions. We have previously demonstrated that following ischemia, NO production by the vascular endothelinm is dramatically reduced. Accordingly, we investigated the effects of NO-Donors on PMN accumulation and tissue injury following hemorrhagic shock and ischemia. Hemorrhagic shock was induced in anesthetized rats by bleeding to 35 mmHg for 3 hours followed by reperfusion. Segments of superior mesenteric artery (SMA) were isolated and suspended in organ baths. In rats receiving saline SMA relaxation to acetylcholine (ACh, 100 nM) was reduced by 80% compared to control SMA segments (p<0.01) while relaxation to sodium nitrite (100 gM) was unaffected. In addition, mesenteric tissue PMN accumulation as determined by myeloperoxidase (MPO) activity was significantly elevated compared to controls (p<0.0l). Interestingly, treatment with the NO-donating agent, S-nitroso-N-acetylpenicillamine (SNAP) significantly preserved SMA relaxation (p<0.01), attenuated mesenteric MPO (p<0.05) activity, and significantly improved survival compared to saline vehicle. In anesthetized, open-chest dogs we investigated the cardioprotective actions of a novel NO-Donor, SPM-5185 (Schwarz Pharma), following regional myocardial ischemia (1 hour) and reperfusion (4.5 hours) . Treatment with SPM-5185 (500 riM) significantly reduced myocardial necrosis by 70% (p<0.01) compared to an NO-deficient analog of SPM-5185, SPM-5267. Furthermore, MPO activity within the ischemic-reperfused zone was also significantly (p<0.01) reduced following treatment with SPM-5185 compared to SPM-5267 (1.6 + 0.5 vs. 3.8 + 0.3 U/100 mg tissue). These data strongly suggest that NO is a potent inhibitor of PMN-mediated tissue injury following hemorrhagic shock as well as in acute myocardial ischemia-reperfusion injury. Overproduction of nitdc oxide (NO') may contribute to sepsis-induced hypotension. During septic shock, excess NO" is produced by an isoform of nitric oxide synthase (NOS) which is induced by inflammatory mediators. Nonselective NOS inhibitors have been proposed as a new therapeutic approach to treating hypotension in septic shock. We studied the differential hemodynamic effects of N~-methyI-L-arginine (L-NMA), a NOS inhibitor, in normal canines versus those challenged with endotexin (LPS) and compared the activity of this drug across the venous, pulmonary and systemic vascular beds. Awake canines were challenged with LPS (0 mg/kg, n=15:8 mg/kg, n=21 ; or 16 mg/kg, n=30) and treated with L-NMA (0, 1,2, 4,10 mg/kg/hr) for 22 hours following a 0, 10, or40 mg/kg loading dose. Animals were resuscitated with IV Ringers solution (10 ml/kg/hr). Hemodynamic data were collected at 0, 2, 6, 10, 14, and 22 hours using intravascular catheters and radionuclide heart scans and analyzed by ANOVA. In both normal and endotoxemic animals, L-NMA at all doses studied similarly increased mean arterial pressure (p=0.07), and systemic vascular resistance index (p=0.Ol) and decreased cardiac index (p=0.05) and oxygen delivery index (p=0.01). In contrast, the effect of L-NMA on mean pulmonary artery pressure, central venous pressure, pulmonary capillary wedge pressure, and pulmonary vascular resistance index was greater in LPS-challenged canines compared to normal animals (p<0.05), but this differential effect on the venous and pulmonary circulation occurred, >6 hours after LPS challenge. L-NMA did not significantly increase survival rates or times at any of the doses studied (1, 2, 4, or 10 mg/kg/h) in either the low (8 mg/kg) or high dose (16 mg/kg) LPS-challenge groups. A nonsignificant (p>0.2) trend toward a beneficial effect on survival ol low dose L-NMA (1 mg/kg/h) in animals given the 8 mg/kg LPS-cha[lenge was not enhanced by increasing the lethality of the model or by administering higher L-NMA doses. At the highest L-NMA dose used in this study (10 mg/kg/h), survival time decreased significantly for both the low and high dose LPS-challenge animals (p<0.05). This increased mortality was not explained by changes in plasma concentrations of either LPS or TNFc~. Thus, L-NMA did not have a greater effect on the systemic arterial circulation in endotoxemic compared to normal canines. However, in the venous and pulmonary vascular beds, the effect of L-NMA increased with time after endotoxin-challenge These data suggest the induction of NOS activity by endotoxin in canines may be relatively greater in venous and pulmonary vessels compared to systernic arteries. L-NMA, a nonselective NOS inhibitor, did not decrease mortality in endoloxemic canines and the highest dose studied was harmful. Pulmonary hypertension (PH) and arterial hypoxemia are characteristic features of the adult respiratory distress syndrome (ARDS). Reducing pulmonary vascular pressures may promote the resolution of pulmonary edema. Intravenously infused vasodilators lower PH in ARDS, but, as a result of their general vasodilatatory effects, systemic mean arterial pressure may also decrease. Furthermore, blood flow may be increased to non-ventilated or poorly ventilated lung areas resulting in a rise of intrapulmonary shunt, thus causing a further fall in PaD 2. Recently, short term inhalation of low concentrations of the gas nitric oxide (NO), an endogenous endothelium derived relaxing factor, which is rapidly inactivated in blood by hemoglobin, was reported to decrease PH without causing systemic vasodilation in sheep [1]. Similar changes have been observed in patients with severe ARDS during repeated short term inhalation of NO (18 and 36 ppm), which rapidly and selectively decreased the mean pulmonary artery pressure (PAP) and, in contrast to intravenously infused prostacyclin, induced a remarkable increase of PaD2 [2] . This improvement in oxygenation was caused by a redistribution in blood flow away from intrapulmonary shunt areas to normal ventilated lung regions. Continuous NO inhalation (5-20 ppm) consistently lowered the PAP and augmented the PaO2/F.O 2 for up to 53 days. No negative side effects were observed during the whole time span examined. In particular methemoglobin levels always remained below 1.5%. Following these investigations, it could be shown that these effects may also occur using concentrations in the parts per billion range [3] , which may reduce possible toxic side effects. However, in the same study it was demonstrated that the dose-response curves for Pa02 and PAP have different patterns. Whereas PAP presented a continuous dose-dependent downward tendency with an EDs o of approximately 2-3 ppm, the improvement of oxygenation had a maximum at 10 ppm and, at higher doses, drifted back towards the baseline data. The ED~o was estimated at approximately 100 ppb, i.e. more than ten times lower than for the reduction of PAP.

In conclusion, inhalation of NO by patients with severe ARDS may result in persistent and reproducible decreases in PAP associated with an evident improvement in PAD2, thus allowing reduction of the F.O 2. NO inhalation should be performed using low concentrations which are less toxic, although any possible risks still have to be considered carefully. Dose-response studies for the individual patients are recommended urgently. Finally, controlled randomized studies are required to demonstrate that additional NO inhalation is able to reduce mortality of ARDS. Inhibition of the activity of glyceraldehyd-3-phosphate dehydrogenase (GAPDH), an enzyme of the glycolysis/gluconeogenetic pathway, through ADP-ribosylation is promoted by nitric oxide (NO). Since NO is produced in the septic liver and hypoglycemia is a major problem of late sepsis, it was investigated whether NO interferes with gluconeogenesis of hepatocytes. Hepatocytes (HC) were isolated from Sprague-Dawley rats using a collagenase perfusion technique and differential centrifugation. Exogenous NO was applied by incubation with the NO-donors S-nitrosyl-acetylpenicillamine and sodium-nitroprusside. Endogenous NO synthesis was induced by incubation with cytokines (TNFcq IL-1, IFNJ and lipopolysacchafide (LPS). 24 hrs later the incubation medium was changed to a solution containing lactate, ornithine, lysine, ammoniumchloride and glucagon for optimal conditions of gluconeogenesis. After 3 more hrs glucose and nitrite levels were determined spectrophotometrically. GAPDH activity was measured by the NADH-dependent conversion of 1,3-diphosphoglycerate to glyceraldehyde-3-phosphate. Incubation of HC with NO-donors led to a concentrationdependent inhibition of gluconeogenesis and GAPDH activity. However, GAPDH activity was about 5 times more sensitive to the inhibitory effect of exogenous NO. incubation of HC with cytokines and LPS induced NQ synthesis as measured by an increase in nitrite concentrations. Endogenously produced NO suppressed gluconeogenesis by 49_+3%. In contrast to exogenously applied NO, the effect of endogenous NO synthesis was less on GAPDH activity resulting in an inhibition of only 32_+12%.

In conclusion, exogenous and endogenous NO inhibited gluconeogenesis as well as GAPDH activity. However, there was no correlation between the extent of inhibition of these two parameters of hepatocellular glucose metabolism. We have shown that inhibition of hepatocyte (HEP) synthesis of nitric oxide (NO) potentiates cell injury in a model of acetaminopheninduced oxidative stress and the extent of damage was paralleled by depletion of reduced glutathione (GSH) stores. To clarify the role of NO in modulating the redox state of HEP, we studied the effect of inhibition of cytokine-mediated NO production on HEP GSH stores, in a system of isolated rat HEP in primary culture, NO synthesis was induced (STIM) by exposure to IL-1, TNF, IFN, and LPS for 16 hours. 20, 60, 120 and 200 ~M of N-monomethyI-L-arginine (NMMA), a specific inhibitor of NO synthesis, was added. Cells incubated in media alone served as controls (CONT). The NO metabolite (NO2); aspartate aminotransferase (AST), an indicator of cell injury; and GSH were assayed. (Data presented as mean + SEM; n=4.) GSH (nmoVma orotein) ..~ (nmol/ma orotein) CONT 14.1 + 0.3 32 + 4.0# STIM 13.4 + 0.6 139 + 12 STIM+2O tzM NMMA 12.7 + 1.1 88 + 5.0# STIM+60 ~M NMMA 9.9_..+ 0.4* 68 + 4.5# STIM+120 pM NMMA 7.6 + 0.6* 57 + 3.0# STIM+200 )lM NMMA 5.0 + 0.5* 37 + 2.5# ANOVA 0,0002 0.0001 (* p < 0.05 versus STIM, # p < 0.01 versus STIM; ANOVA with Neuman-Keuls) GSH in CONT+120 I~M L-NMMA was equivalent to that of CONT (13.3 vs.14.1). AST release was equivalent in all treatment groups. These data show that inhibition of HEP synthesis of NO depletes intracellular stores of reduced GSH. We conclude that hepatocyte NO production modulates cellular GSH homeostasis and as a result, may be hepatoprotective in oxidative injury. 

Nitric oxide (NO) is a modulator of immune response and may be involved in thE changes in immune reactivity after major trauma and operations. We investigated NO-generation in rat and mice spleen cells (SC) after partial hepatectomy (PH). C57BL/6 mice and LEW rats underwent a 50% and 70% PH, respectively. SC were prepared 1-6 days after PH and plated at 1 to 2 x 10ecells per well. After 20 h incubation at 37°C, NO-production was measured as nitrite levels (Griess reagent). Normal mouse SC did not produce NO, neither basal nor in response to LPS or Con A Starting at the second day after PH, we found a substantial production of NO. In rats, also SC from control animals were able to generate NO; both basal and stimulated NO-generation were further enhanced after PH (Table, values expressed as mean --SE). After shame operation, there was only a modest elevation of NOproduction in rat and mouse SC. In first experiments we could demonstrate NO-production also in phagocytes from a patient 3 days aider liver partial resection (1.4 nmol nitrite/106 cells)

Enhanced NO-production in macrophages may contribute to the changes of immune reactivity after partial hepatectomy. Nitric oxide (NO) is recognized as an important mediator in endotoxemia and sepsis. Increased synthesis of NO has been demonstrated in septic humans and animals, and NO inhibitors have been used in the treatment of septic shock. Recent reports have, however, suggested that this form of therapy may cause serious organ damage.

In the present investigation circulatory and metabolic changes in the liver were studied during treatment with the NO-synthase inhibitor N-nitro-L-arginine-methyl ester (L-NAME) in endotoxemia. Methods: Juvenile pigs were randomized to one of the following treatment groups: 1) Encletoxin and L-NAME, 2) Endotoxin, 3) NaCI and L-NAME, 4) NaCh Preliminary results from groups 1 (n=7) and 2 (n=8) are presented. Catheters for pressure measurement were introduced into the aorta, hepatic and portal veins and ultrasonic transit time flow probes were placed on the hepatic artery and portal vein. A catheter was introduced into the pulmonary artery. Endotoxin (1.7 gg/kg/h) was given as a continous portal infusion over the entire observation period of 8 hrs. L-NAME (25 mg/kg) was given as a bolus after 3 hrs. of endotoxemia. Results: Endotoxin transiently reduced portal vein flow (PVF) by 25%* and hepatic artery flow (HAl e) by 45%*, while L-NAME caused a further and lasting reduction in flow (PVF 78%, HAF 90%)*. Transhepatic (portal-hepatic vein) vascular resistance increased to 3 times baseline value during endotoxemia while L-NAME caused a further marked increase in resistance to 12 times initial value. Portal oxygen saturation (SO2) decreased by 60%* during endotoxemia. L-NAME caused a reduction in portal SO2 by 87%*. Arterial SO2 was unchanged in both groups. Hepatic oxygen uptake was not changed by endotoxin, but was markedly reduced after addition of L-NAME. Endotoxin caused a 27% reduction in cardiac output (CO). The addition of L-NAME reduced CO by a total of 71%*. *: p < 0.05. Conclusion: Is the present model of endotoxemia treatment with the nitric oxide synthase inhibitor L-NAME markedly reduced liver perfusion and portal oxygen supply. This might explain the increased liver damage reported in previous studies using NO-inhibitors. The increase in transhepatic resistance found after L-NAME treatment will tend to cause pooling of blood in the splanchnic veins, resulting in reduced filling of the heart and thus contribute to the observed reduction in cardiac output.

Institute for Surgical Research, Rikshospitalet, The National Hospital, University of Oslo, 0027 Oslo, Norway.

We have investigated the role of tumour necrosis factor (TNF) and interleukin-i (IL-I) in the induction of nitric oxide synthase (NOS) by bacterial endotoxin (lipopolysaccharide;

LPS; 2 mg kg -I i.v.) in vivo. In anaesthetized rats, pretreatment with a monoclonal antibody for TNF (TNFab; 20 mg kg -I s.c., at 16 h prior to LPS) or with an IL-I receptor antagonist (IL-Ira; 16 mg/kg bolus and 2.4 mg/kg/h infusion) ameliorated the fall in mean arterial blood pressure (MAP) at 90-180 min after LPS. For instance, endotoxaemia for 180 min resulted in a fall in MAP from 114-+6 (control) to 79-+4 mmHg (p<0.01; n=8). In contrast, animals pretreated with TNFab or IL-ira prior to LPS injection maintained significantly higher MAP at 180 min when compared to LPS-control:

118-+3 mmEg (n=5) and 103-+7 mmHg (n=5), respectively (p<0.01). Three hours of endotoxaemia significantly reduced the contractile effects of noradrenaline (NA) in the thoracic aorta ex vivo. The hyporeactivity to NA was partially restored by in vitro treatment of the vessels with NG-nitro-L-arginine methyl ester (L-NAME, 20 min, 3x10 -4 M). Pretreatment of rats with TNFab or IL-Ira significantly (p<0.05) prevented the LPS-induced hyporeactivity of rat aortic rings ex vivo. L-NAME did not alter or only slightly enhanced the contractions of aortic rings obtained from TNFab or IL-ira treated LPS-rats, respectively. At 180 min after LPS there was an induction of calcium-independent NOS activity in the lung (5.14-+0.57 pmol citrulline/mg/min, n=8), which was attenuated by TNFab and !L-ira by 37-+6% and 46-+6%, respectively (n=5; p<0.05). Thus, the production of both TNF and IL-I contributes to the induction of NOS by LPS in vivo. The protective effect of agents which inhibit the release or action of TNF or IL-I in shock may be, in part, due to inhibition of NOS induction. Neal Garrison, MD Objective: Sepsis is often accompanied by organ dysfunction, in part due to impaired microvascular perfusion. Recently, nitric oxide (NO) has been described as an important mediator of the hemodynamic changes of sepsis, and NO synthase (NO-S) inhibitors have been advocated for treatment of septic shock, but their visceral microcirculatory effects are inadequately characterized. We postulated that NO-S inhibition would EXaCErbatE the impaired organ pErfusion of sepsis. Methods: Six groups ofdecerebratE rats were studied. Bacteremia was induced with 109 live E. coli, which consistently increased cardiac output 15-20% above baseline (BL). The NO-S inhibitor Nm-nitro-Larginine methyl ester (L-NAME,1 mg/kg IV), prevented this increase and elevated MAP by 25-30%. In the first 3 groups, total hepatic blood flow (THBF, ml/min by time transit flowmetry) and microvascular perfusion (MI-IBF, ¼ BL by laser Doppler flux) were measured. In the other 3 groups, in vivo videomicroscopy was used to observe renal microvascular responses (ILA=interlobular artery, AFF=afferent arteriole, EFF=efferent arteriole; % BL for all). Results: Data are 60 rains after E. coB. N=5-7/group. * p<0.05 vs BL by REMANOVA and § p<0.05 vs E. coli alone by ANOVA.

EC+L-NAME 11-+2 -57_+10" § -45_+4* § -89_+3* § -33+5* -20+5* § Conclusions: L-NAME administration in controls decreased renal blood flow, indicating NO contributes to basal renal tone. Bacteremia decreased MtlBF but not THBF, and MI-IBF was further impaired by NO-S inhibition. E. coli caused renal preglomemlar, but not postglomerular constriction and reduced flow. L-NAME exacerbated these E. coli-induced alterations and caused EFF constriction. These data indicate that NO-S inhibition exacerbates bacteremia-induced impairment of renal and hepatic blood flow, suggesting that NO is an importam compensatory dilator mechanism in these organs during sepsis. IRF (iron responsive factor) is the central regulatory protein of intracellular iron metabolism able to bind to responsive RNA elements (IREs) present atthe 5'untranslated region (UTR) of ferritin mRNA and 3'UTR of transferrin receptor mRNA. Binding of IRF to IREs results in repression of ferritin mRNA translation and increased stability of transferrin receptor mRNA leading to enhancement of transferrin receptor translation. We describe here that either tetrahydrobiopterin dependent stimulation as well as cytokine (IFN-~)/lipopolysaccharidemediated induction of nitric oxide synthase activates IRF, which is due to direct interaction of nitric oxide with the iron-sulphur-cluster of IRF. This was shown by gene expression studies using a plasmid containing a ferritin IRE and a CAT indicator box which was transfected into K562 myelomonocytic cells, which were shown to have a constitutive form of nitric oxide synthase (NOS). Furthermore, the increased binding of 1RE to IRF due to IRF activation of IRF by nitric oxide was demonstrated by gel shift assays. IRF activity was much more increased in cellular extracts from murine macrophages (J774) where a cytokine inducible form of NOS has been characterized earlier as compared with IRF activity in K562 cells, where NOS was stimulated by increasing the availability of the essential NOS cofactor 5,6,7,8-tetrahydrobiopterin. We then demonstrated that activation of IRF by nitric oxide is accompanied by alterations in ferritin translation as checked by metabolic labeling and immunoprecipitation. These results suggest a reasonable mechanism for the regulation of iron disturbances under chronic inflammatory disorders, characterized by increased concentration of immune activation parameters like IFN-5or neopterin and low serum iron and hemoglobin concentrations. Taken Nitric oxide, NO, the putative endothelial derived relaxant factor, EDRF, has been shown to be a potent inhibitor ofplatelet aggregation in vitro. In vivo evidence however, is scarce. Accumulation of platelets in the lungs has been shown to occur during extracorporeal circulation. The aim of the present study was to investigate the effect of inhaled NO on this reaction. Materials and methods: The animals were divided into two groups, each consisting of 7 pigs. Platelets were selectively labelled with lUln-oxine. Dialysis was instituted via catheters in the femoral vessels.

In group 1, NO, 50 ppm, was added to the inhaled gas from the start of dialysis. In group 2 NO was not given. The activity over the lungs was followed dynamically with a gamma camera. Central hemodynamics was monitored via a Swan -Ganz catheter. Results: The activity was significantly lower in group 1, from 2 minutes after start of dialysis and onwards, indicating diminished accumulation of platelets in the lungs. Parallel to this the hemodynamic response in terms of increased pulmonary artery pressure and pulmonary vascular resistance was blunted in this group Conclusion: Inhaled NO in this model seems to affect pulmonary platelet sequestration. An associated attenuation of the changes in central hemodynamics was also seen. Previous studies from our laboratory have demonstrated that vascular contractility decreased in endothelium-intact blood vessel rings in early and late stages of sepsis. Although endothelium removal in early sepsis restored vascular contraction, the depressed smooth muscle contractility observed in late sepsis was not restored by endothelium removal. This indicates that impairment of smooth muscleper se may be responsible for such dysfunction in late sepsis. The aim of this study, therefore, was to determine whether or not smooth muscle-derived nitric oxide (NO) plays a role in producing vascular smooth muscle dysfunction during late stages of sepsis. To study this, rats (250-300g, n=4-5/group) were subjected to sepsis by cecal ligation and puncture (CLP). Septic and shamoperated rats then received 3 rrd/100g BW normal saline. The animals were killed at 10, 20, or 35 h post-CLP (10 h post-CLP=early sepsis; 20-35 h post-CLP=late sepsis), and thoracic aortic rings were prepared for contraction studies using organ chambers. The complete removal of endothelial cells was tested by the absence of any significant acetylcholine-induced vascular relaxation. Contractile responses to norepinephrine (NE, 10 9 to 10 -5 M) were determined in the aortic rings without intact endothelium. NG-monomethyl-L-arginine (L-NMMA, 300/~M, an inhibitor of NO synthase) was then added to the organ chamber and NE-induced peak contraction was determined before and after the addition of L-NMMA. The peak contraction (rag/rag tissue, mean_+SEM) is shown below:

The results indicate that the addition of L-NMMA did not significantly affect NE-lnduced peak contraction in endothelium-denuded vessel rings at 10 and 20 h after CLP. In contrast, L-NMMA administration produces an 18% increase (P<0.05) in peak contraction during late sepsis. Therefore, the vascular smooth muscle contractile dysfunction observed at 35 h post-CLP is partially due to smooth muscle-derived NO over-production. Thus, unlike macrophages in which inducible nitric oxide synthase (iNOS) is observed in early sepsis, the iNOS in vascular smooth muscle appears prominent only in the late stages of sepsis. In three cases of human septic shock in which NG-monomethyI-L-Arginine, (L-NMMA) a nitric-oxide-synthase-inhibitor was applied, we isolated three completely different types of pathogens: Candida, Pseudomonas aeruginose and multiresistant coagulase-negative staphylococci. This observation suggests that endotoxin alone is not the main factor triggering hypotension in septic shock by the nitric oxide pathway. In a 68-years-old woman in severe septic shock due to a Candida and Pseudomonas aeruginosa infection complicated by adult-respiratorydistress-syndrome conditions deteriorated despite adequate conventional therapy. In this trial, effects of L-NMMA on cytokin-levels were investigated. The study-protocol was approved by the ethical committee of the Department of Surgery. After two boll of 200 mg of L-NMMA, a continuous infusion was installed (0.05 mg/minute and kg body weight L-NMMA). As expected mean arterial blood pressure rose (62 to 134 mmHg}, heart rate stayed stable (124 + 4 b/rain), systemic vascular resistance increased (225 to 354 dyne.sec/cm5), cardiac output decreased (17 to 15.2 L/rain), and cardiac index declined (9.94 to 8.63 L/min/m2}. Before and after 90 minutes while the infusion of L-NMMA, blood samples for immunological measurements were taken and processed together. Pulmonary-shunt-volume was observed before the application of L-NMMA, after one hour and after 140 matutes. Neopterine increased from 16.51 to 32.55 ng/ml, tumour-necrosis-factor-a increased from 24.16 to 36.61 pg/ml and intedeukin-6 increased from 61.9 to 98.2 pg/ml. Immunoglobulines A, G, and M (3.2 to 2.9, 8.1 to 7.6, 1.1 to 0.9 g/I), Complement factor C-3c and C-4 (0.54 to 0.50, 0.22 to 0.23 g/I), alpha-l-antitrypsine (3.86 to 3.83 g/I), C-reactive-protein (111.1 to 105.6 rag/I), interleukin-1 (0 pg/ml) and soluble interleukin-2 (2128 to 1983 Units/ml) did not change significantly. Pulmonary-shuntvolume decreased from 54.1% to 35.4% within one hour and to 36.6% after 140 minutes. In septic shock blocking nitric oxide as an intervention at the end of a not ~,et ful!y understood cascade might have important influences on pulmonary-shunt-volume and inter-cell-communication.

Department of Surgery, Pharmacy* and Immunology**, University Hospital of Zurich, R~imistrasse 100, 8091 Zurich, Switzerland We previously reported that hypoferremic CBA mice had an increased resistance to Salmonella infection, and that injection of Ammonium Ferric Citrate (AFC) to these mice led to enhanced infection (Ganthier et at. 1986. Microbiol.Immuno130:425) . Because Nitric Oxide (NO) is involved in the antimicrobial activity of routine macmphages towards various inttacellular pathogens, we investigated the influence of iron on the bactericidal activity of CBA mouse macrophages towards S.typhimurium and on the production and activity of Reactive Nitrogen Intermediates (RNI).

Peritoneal macrophages hum CBA mice were cultured in the presence (or not) of AFC 50 ,uM, IFN-,/250 U/ml, LPS 1 fig/m/, NGMonomethyl-L--Arginine (MMLA) 2raM. Nitrite (NO2-) content of the supematants was determined by a standard Griess reaction, and H202 release was measured by the peroxidese dependant oxidation of phenol red. For intracellular killing, macrophages monolayers were infected, and, at various intervals, lysed by Triton X-100, and surviving bacteria enumerated by colony counting on agar. For in vivo experiments, mice were infected IP with 0.5 ml of a suspension of 5.10 ~" S.typhimurium, strain C5, and injected with Aminoguanidine (AG) 1 mg/ml in saline.

Our results show that the RN[ inhibitor AG strongly accelerates the mortality of infected mice, the survival rate decreasing from 60% in the control group to 20% in the treated group, 10 days after challenge. Correlatively the RNI inhibitor MMLA induces in vitro a decrease in the rate of bacterial killing, fxom 78% to 57%, in macrophages triggered with IFN-? + LPS. The cultivation of macrophages in the presence of AFC leads to a decreased NO 2-accumulation, 4.7 nmole/well v.s. 21 nmole/well. Conversely H202 production is enhanced from 09 nmole/well up to 2,33 nmole/well. Nevertheless, macrophages cultivated in the presence of AFC exhibit an increased tale of intracellular killing, 77% in iron exposed macrophages v.s, 68% in control macrophages. When triggered with IFN-~, alone, macrophages have a reduced antibacterial activity (57% v.s. 68%) whereas the addition of AFC to these macrophagas restores an elevated (72%) rate of killing.

In conclusion, the results show that bactericidal activity of CBA macrophages towards S.typhimurium depends on the production of NO by these macrophages ; but they also demonstrate that NO is not the only reactive species involved in the intracellular kil/ing of S.thyphimurium ; indeed AFC which strongly inhibits RNI production, stimulates H20 2 release by these macrophages and increase their bactericidal activity in vitro. Nevertheless AFC may promote bacterial growth in vivo.

CRSSA. Unit4 de Microbiologie. BP 87. 38702 La Tronche Cedex FRANCE.

Henning Jahr, Ulrike Noack, Karin Braun

The large amounts of NO produced by the inducible NO synthase in rat macrophages have direct antimicrobial effects, but inhibit the activation of the lymphocyte-dependent host defense system. The aim of this study was to investigate if complement activation influences NO-generation. Spleen cells from LEW rats were incubated at 37°in TCM-199/5% FCS, with or without additional rat serum. After 20 h, nitrite (end product from NO metabolism) was measured by Oriess reagent. In rat spleen cell preparations, most of the NO is produced by macrophages. Complement activation in vivo was carried out by i.v. injections of 240 U cobra venom factor/kg b.w. at days 0 and 2. Significantly higher (P<O.O1) LPS-stimulated NO-generation was found in spleen cells prepared at day 4 (13.0 4-1.3 nmol nitrite/106 cells, n=6) compared to spleen cells from non-treated animals (3.6 ± 1.3 nmol, n=6) Complement activation was effective also in vitro. When incubated in 25% zymosan-activated rat serum, both basal and LPS-stimulated NOproduction were enhanced in spleen cells (Table, values Increased production of Nitric Oxide (NO) is associated with sepsis, however, the effect of NO inhibition on survival during sepsis is unclear. We examined the effect of NO inhibition on host's ability to eliminate bacteria and survival in mice sepsis model.

Sixty female Balb/c mice were injected with E.coli (10qbody) into peritoneal cavity. The treated group received N-ultro-l-arginine-methyl-ester (L-NAME), an inhibitor of NOS, one hour before bacterial challenge. Thirty mice were observed for survival after E.coli challenge and the other mice were sacrificed at 4 hours. Blood was obtained by cardiac puncture and peritoneal lavage fluid (PLF), liver and lung were harvested. The number of peritoneal exudative cell (PEC) was counted and colony forming units (CFLD of bacteria in the tissues, blood, and PLF were also determined. In addition, PEC was cultured in vitro with or without lipopolysaecharide (LPS). TNF levels in the blood, PLF, and supematants of cultured PEC were measured by L929 bioassay.

Survival Rate (%) Qrql/,p n 8 hr 24 hr 4 dav control (normal saline 0.1ml/body i.p.) 10 100 50 40 N10

(L-NAME 10mg/kg i. Administration of L-NAME before E.coli injection increased the number of bacteria in PLF and TNF level in plasma, The result suggests that NOS inhibitor may depress the host's ability to kill the injected bacteria and that it may induce greater systemic TNF production. We conclude that NOS inhibitor is detrimental to animals in sepsis. The hypothesis that SOD prevents reperfusion injury to the liver by protecting the endogenous endothelium derived vascular relaxing factor (NO) from being inactivated by oxygen free radicals should be investigated. Male Wistar rats were subjected to 60 min of partial liver ischemia (70%) and 30 min of consecutive reperfusion in situ. Some rats were pretreated with the NO synthase inhibitor nitro-L-Arginin (NLA: I0 mg/kg). SOD, when used, was given as Mn-containing preparation at a dose of 6000 U. i.v. i0 min prior to ischemia. Results (SOD/ NLA+SOD/ NI~,, respectively): Bile flow (~i/g/min) : 0.93±0.11"#/ 0.52±0.08/ 0.48±0.05. ALT (U/2) • 171±40~#/ 477±173/ 506±182. GLDH (U/l): 5.2±1.2e#/ 10.7±4.2#/ 24.3±9.1. In absence of NLA, SOD enhanced postischemic bile flow and reduced leakage of ALT and GLDH into the plasma, while the effects of SOD disappeared, when endogenous 140~synthesis was inhibited by NLA. NLA had no ~ffect on untreated animals submitted to the same protocol. These results may suggest, that oxygen free radicals interfe~:e with the endogenous vasodilator NO and that the benefit of SOD can be attribuated in large part to a protection of NO during reperfusion. Nitric oxide (NO) reduces pulmonary vascular resistance and increases oxygenation by inhalation in ARDS patients [1] . Animal studies on the endogeneous NO production proved that NO is exhaled after synthesis in the respiratory tract [2] . With approval by the hospital ethics committee, we studied healthy volunteers and ventilated patients by measurement of inand exhaled NO by NO/NO2-chemiluminescence in segments of the upper respiratory tract. We found that NO is synthesized predominantly in the nose and in the upper nasopharynx, leading to inspiratory tracheal NO concentrations of 0.07-0.13 parts per million in non-smoking volunteers; data during mask ventilation per nose were as follows:

Non 0.089 0.0422 0.0721 0.0311,2 Parts per million (ppm) NO, mean, n = 10; ~ p < 0.05 (t-test) between smokers and non-smokers; 2 p < 0.05 (t-test) between in-and expiration. About 50% of NO is resorbe.d by the lung, and the exhaled NO -in contrast to former presumptions -is the remaining part of the autoinhaled NO. Intubated and ventilated patients are deprived of NO autoinhalation, and the exhaled NO in the trachea decreases more than 95% {O.043 ppm before intubation vs. 0.003 ppm after intubation, ~, < 0.01), whereas the nasal NO concentration increases by cumulation. Hence, low-dose NO inhalation in ARDS patients might be considered as a NO replacement, especially since the tracheal NO concentrations we measured in volunteers are similar to those which could be demonstrated to be effective for ARDS in former studies [3] . The data demonstrate that NO is synthesized in the nose, and inhaled and resorbed by the lung. This phenomenon of NO autoinhalation, which is reduced in smokers and in ventilated patients, does possibly contribute to the regulation of the ventilation-perfusion ratio in the lung. 

Kikuchi 1, Yoshihiro Inoue 1, Masao Yoshida 2 1 Critical Care and Emergency Center, and 2Department of Bacteriology, Iwate Medical University.

Nitric oxide (NO) is produced by macrophages and vascular endothelial cells. It is thought to be the true form of endotheliumderived relaxing factor, and to be involved in the fall of blood pressure during septic shock. We have reported previously that endotoxin (LPS), TNF-c~ and IL-2 contribute to the occurrence of septic shock (Circ Shock 38:264; . The present study investigated the in vitro effects of LPS, TNF-o~, IL-2 and IFN-'I on the production of NO by macrophages. Peritoneal macrophages were cultured with LPS, IL-2, TNF-~ and IFN-¥. In culture with LPS, NO was produced in a dose-dependent manner, and the production was suppressed by a NO production inhibitor, L-NMMA. IFN-Y, IL-2 and TNF-c~ used alone did not promote the production of NO, but helped LPS to facilitate NO production. A combination of IL-2 and TNF-c( enhanced LPS-facilitated NO production to a greater extent than IL-2 or TNF-c~ alone.

The above results suggest that these cytokines are involved in endotoxin shock through the mechanisms that facilitate NO production.

19-1 Uchimaru, Morioka 020. Japan.

T. Yolk 1,2, I. Ioannidis 1, W.J. Kox 2 and H. de Groot 1

Objectives: Nitric Oxide (NO.) is known to play an important role in neurotransmission, vasoregulation, and bactericidal as well as tumoricidal cellular defense systems. By means of NO-donating agents we studied the mechanism of NO-mediated cytotoxicity against rat liver microvascular endothelial cells. Materials and methods: 3-morphoiinosydnonimine-N-ethylcarbamide (SIN-l) and sodium nitroprusside (SNP) were used as NO. donators, LDH release and trypan blue exclusion were applied to indicate cell viability. Results: SIN-1 (5mM), which releases both NO. and O2-., caused a time-dependent cytoreduction of 60% and 80% after 3 and 6 hrs, respectively. This effect was not inhibited by superoxide dismutase (SOD), which removes 02--to form H202 and 02. In contrast, the addition of catalase (CAT), which removes H202 to form H20 and 02, diminished the cytotoxic effect ot SIN-1 to 30%, equivalent to high concentrations of SNP (20mM), which solely releases NO.. In addition, the amount of H202 formed by 5mM SIN-1 equaled the amount ot enzymaticany produced H202 by 5mU/L glucose oxidase (GOD). Whereas GOD in this concentration and SNP in low concentration (5mM) alone were not toxic to endothelial cells, the combination caused an 80% reduction of viability, comparable to the effect observed with SIN-1 (5mM) alone. Moreover, toxicity mediated by high concentrations of SNP (20mM and 50raM) was reduced by 20% under hypoxic conditions indicating synergism with NO-and 02. Conclusions: We conclude, that NO.-induced toxicity against endothelial cells is enhanced slightly in the presence of oxygen and is not due to an interaction with 02-. However, toxicity increases dramatically in the presence of H202. This may possibly occur either by a chemical formation of more potent reactive hydroxyl radicals or by independent actions on different cellular targets. 

Although it is becoming increasingly clear that Nitric Oxide (NO) is associated with otxan dysfunctions in septic patients, little is known as to die kinetics of NO production in these patients To clarify these kinetics, we have investegated serum arid monocyte associated NO in septic patients.

Five patients who had undergone gastrointestinal surgeries mid were complicated postoperatively with severe sepsis served as the subjects in this study (Sgroup), using twelve healthy volunteers as a control group (C group).

Serum NO2 and NO3 levels were measul*d sliectrophotometricallyMonocyte cultures were done for 24 Itrs with or without LPS+IFN-T stimulation mid NO2 and NO3 levels in the supernat,'utts were also measured spectrophotometrically. Furthermore,using an NO selective electrode,time course of the LPS+IFN-T induced NO production by monocytes was studied.

l) Serum NO2 levels in S group were slightly lfigher than dmse in C group,bnt there were no significoatt differences between the two. 2)Both NO2 levels in the supematmlts of monocytes cultured with LPS+IFN-T mid NO2, NO3 levels ill lhe supernatants of monocytes cultured without LPS+IFN-Y were significantly higher in S group. Further,the time required for the NO production by enltnred monoeytes was siglfificantly shorter also in S groap. Conclusions l)In severe septic patients, monocyte NO productions wei~ not only primed, but also activated concomitantly, suggesting that these monocyte activations m'e strongly associated with organ dysfunctions in sepsis.

2)Based on the restdts of serum NO nteasuremeuts, it can be deduced that NO p~'oduced by monocytes in septic patients affects tissues and org~s Ioeoregiomflly. Objectives of the study: Nitric Oxide (NO) is an important messenger which accounts for a wide spectrum of physiological and pathophysiological processes, e.g. mediating vasodilation in endoloxin shock Investigating the signal lransducfion pathways involved in the regulation of the inducible nitric oxide synthase (iNOS), we report the involvement of phosphatidylcholinespecific phospholipase C (PC-PLC) and proteinkinase C (PKC). Materials and Methods: NO-production was induced in the murine macrophage cell line J774 with lipopolysaccharide (LPS) [llJg/ml] and interferon ;f (IFNy) [200U/ml]. After 18 hours of incubation NO-release was determined as nitrite (NO2) by using a colorimetric assay based on the Griess-reaetion. Results: Preincubation of cells with the PKC-inhibitors staurosporine (STP), chelerythrine, bisindolylmaleimide (BIM) and D609, that recently has been identified as a selective inhibitor of PC-PLC, diminished significantly LPSand ]FNT-induced NO2-production (p<O,001). Although we observed no toxic effect on cell viability, NO2-production was related to protein concentrations to exclude any inhibitory effect on celt growth. Celts preincubated with STP [lO0nM] released 20% less NO 2 in response to LPS and IFN,[ compared to control cells cultured without inhibitor. Chelerythrine [30}JM] and BIM [IOIJM] reduced the NO2-formation by 50% and 85%, respectively. Interestingly, the most potent inhibitor of NO-production turned out to be the xanthate D609 Cells pretreated with D609 [301~g/mt] released 96% less NO 2 than stimulated controls The inhibitory effect on NO 2production decreased the later the inhibitore were added after stimulation; e.g. BIM added 6 and 12 hours after stimulation reduced NO2-production only by 50% and 12%, respectively. Accordingly, iNQS activity present in stimulated cell lysates supplemented with L-arginine and co-factors was not suppressed by the inhibitors tested. Conclusions: Our findings suggest, that the induction of iNOS but not its activity is a PKC and particularly PC-PLC dependent process.

Dr. K Tschaikowsky, Institut for Anaesthesiologie, Universital Erlangen--NQrnberg, Krankenhausstr 12, 91054 Erlangen Interleukin-1 (IL-1) affects nearly every cell type by increasing the expression of a variety of genes associated with the promotion of inflammatory processes.

These include upregulation of cyclooxygenase and nitric acid synthases.

The IL-1RI transmits a signal(s) through the cytoplasmic domain which is structurally related to the fruit fly Toll protein. We have recently cloned the promotor most proximal to the 5' UTR of the human IL-1RI gene.

The genomic sequences reveal a lack of TATA and CAAT boxes but rather a considerable (75%) GC rich region.

The translation of the type I IL-1R appears under a significant suppression and may limit the biological activities of IL-1 by low level of expression.

On the other hand, the type II IL-1R, lacking a significant cytoplasmic domain, does not transmit a signal and acts as a decoy receptor preventing the binding to the type I receptor.

There are several approaches to reducing IL-1 activity; inhibition of IL-1 converting enzyme which cleaves pro-IL-l[3, anti-IL-1RI, the IL-I receptor antagonist (IL-1Ra) and soluble (extracellular) IL-1RI and IL-1RII. IL-1Ra is structurally related to IL-1 and binds at 4°C to the IL-1RI with near equal affinity as that of bona fide IL-I; however, IL-1Ra does not transduce a signal.

IL-1Ra has been given to humans in pharmacologic levels (mean plasma levels of 30gg/ml) without evidence of agonist activity.

In addition, there has been no affect on normal homeostatic parameters, suggesting that IL-I does not play a role in health.

In healthy humans given intravenous endotoxin, IL-IRa reduced endotoxin-associated neutrophilia by 50% and completely reversed endoloxin-iuduced immunosuppression, IL-tRa is produced naturally and is elevated in the circulation in several diseases.

In a variety of diseases, plasma levels of IL-1Ra are in 100-fold molar excess to those of IL-I [3. It is unclear whether these endogenous levels of IL-1Ra are sufficient to block IL-1 activity in disease.

A single injection of ]L-6 or IFNa in humans does not induce IL-1 but rather high levels of IL-1Ra (10ng/ml).

Prof.dr.L.A. Aarden

Interleukin 6 (IL6) is a multifunctionaI cytokine with a central role in host defense mechanisms and consequently in inflammation. IL6 is thought to be involved in the pathogenesis of various diseases. Therefore, specific inhibitors could have therapeutic value. A human-mouse chimeric monoclonal antibody to IL6 has been applied in a phaseI clinical trial in patients with multiple myeloma and in primate models of sepsis. No toxicity was observed and the most remarkable outcome of this study was the build up in the circulation of IL6-anti-IL6 complexes, suggesting that plasma IL6 measurements represent a gross underestimation of IL6 production in the patient. Until now no natural antagonists of IL6 have been described. The biological activities of IL6 results from binding to a lowaffinity IL6-receptor (IL6R). The IL6/IL6R complex induces disulfide-linked homodimerization of the signal transducing receptor component, gp130. Studies on the structure-function relation of IL6 revealed that disruption of the gpl30-interaction site on IL6 gives rise to a mutant IL6 that, by virtue of its intact binding to the IL6R, acts as a potent antagonist on human hepatocytes and on human B cells. In vivo studies using this molecule will be initiated in the near future. The evolution of infla~ation involves a dynamic series of cellular events controlled by specific signals which are important to the initiation, maintenance, and final resolution of the inflammatory response. The various phases of inflammation are influenced via the expression and subsequent regulation of a number of key mediators which play a predominant role during each particular stage of the developing lesion. For example, the initial elicitation of blood born leukocytes to on inflamed area is a complex system that is dependent upon the expression of early response cytokineSo These proximal mediators, including both interleukin-i (IL-I) and tumor necroisis factor (TNF), are important in the initiation phase by activating both immune and non-immune cells and establishing a series of cytokine networks, Thus, the above proximal mediators can stimulate endothelial cells, epithelial cells, smooth muscle cells, and fibroblasts to generate more distal cytokines. These latter cytokines include interleukin-8 (IL-8), macrophage inflammatory protein-i (MIP-I), and monocyte cbemoattractant protein-i (MCP-I). The importance of the above chemukines can be found in the role they play in leukocyte elicitation, as well as wound repair. Numerous investigations have shown the diverse cellular sources of IL-8 and the ability of IL-8 to elicit neutrophils in vitro at pM to nM concentrations, however, recent studies have sho%~ an additional role for IL-8 during the resolution phase of the inflammatory process. Using corneal pocket animal models of angiogenesis/neovascularization, IL-8 was found to be a potent angiogenic factor at concentrations ranging from 2-40ng/ml. Sequential fluorescein angiograms demonstrated that IL-8 induced neovascularization by 14 days, which regressed by 6 weeks. In further analyses, both conditioned media recovered from either inflamed human heumatoid synovial tissue macrophages or LPS-stimulated peripheral blood monocytes were found to possess potent angiogenic activity, which was effectively blocked by neutralizing antibodies directed against human IL-8. In addition, an IL-8 antisense nligomunclentide blocked the expression of a functionally active angiogenic factor from LPS treated monocytes. The above data demonstrate that IL-8 has multifumctional activities during the initiation, maintenance, and resolution of a local inflammatory response.

inhibits specific T-cell responses to soluble protein antigens and alloantigens. The proliferative responses of Th0, Thl and Th2 CD4 ÷ T-cell clones and cytokine production by these T-cell subsets are equally well inhibited. The inhibitory effects of IL-10 are indirect and related to inhibition of antigen-presenting capacity and accessory function of the antigen-presenting cells (APC). However, IL-10 also directly inhibits T-cell proli/eration induced by cross]inked anti-CD3 or anti-CD2 mAbs (e.g. in the absence of professional APC), by specifically inhibiting IL-2 gene transcription and IL-2 protein production. IL-10 also inhibits the production of the pro-inflammatory cytokines IL-l-a, ~, IL-6, and TNF-a and the chemokines IL-8 and MIP-I-u, which are strong chemo attractants for neutrophils and macrophage, respectively. In contrast, IL-10 enhances the production of the IL-1 receptor antagonist, a cytokine with anti-inflammatory activity. IL-l0 produced by monocytes downregulates class II MHC expression and IL-10 production by these cells in an autoregulatory fashion.

Collectively, these in vitro data indicate that IL-10 is a powerful suppressor factor for immune-and inflammatory responses and therefore may have potential clinical utility as an anti-inflammatory agent. This notion is supported by recent studies which indicated that IL-10 also prevents lethal LPS-induced toxic shock in mice. Five of the chamekines (HIP-I~,eRA~TES. IP-IO and IL 8) also chemoattraet T lymphocytes, while others act on mast cells. We have studied the effaces of these chemokines on T cell subsets. IL 8 preferentially chemoattracts uns~imulated T Cells. RAISES ~ttracte resting as well as activated CD4 and CD8 memory T cells. IP-10 chemoattracts only preactivated (not resting) CD4 or CD8 T cells. HIP-l~ preferentially chemoattracts GD8 T calla, while HIP-18 more readily attracts the CD4 T call subset. The 8 chemoklnes and IP-IO were also sho~rn to promote the adherence of ~he same subsets of anti-CD3 activated T cells to ILI activated human umbilical vein endothelial cells (HITVEC). This was not associated with a~y i~crease in the number of lymphocyte adhesion molecules, bu~ could be inhibited by neutralizing monoclonal antibodies re LFA-I and VLA-4. These chemokines also rapidly increased ~ha adhesion of resting T l~phoeytes to plates coated wi~h 8 integrlns (I-CAM or V-CAM) or matrix proteins such as f~bronectln. This induction of adhesion began wiEhln 30', peaked by 1 hr and was completed in 3 hr. This suggests that chemoklnee rapldly increase the binding affinity of adhesion prote~ns on T cells. We also recently observed ~hat MCAF/MCPI and RANTES chemoattract unatimulated mast calls. Furthermore, IgS actlvared mast calls were chemoattraeted by MCAF, RA~TES, PF4 and MIP.I~. However. ~he chemoklnee did not induce mast cells ~o release histamines. Presumably, chemokines as regulators of the adhesion, migration and homing of all kinds of leukocytes participate in many types of inflammatory diseases. Natural killer cell stimulatory factor or interleukin-12 (IL-12) is an heterodimerie cytokine formed by two covalently linked glycosylated chains of 35kD and 40 kD. IL-12 was originally purified from the supermatant fluids of EBV-transformed human B ly~phobiastoid cell lines. However, in addition to B cells, phagocytic cells, i.e. monocytes, macrophages, and neutrophils, have been identified as the major physiological producers of IL-12. Phagocytic cells produce IL-12 in response to bacteria, bacterial products such as LPS, and intracellular parasites. The cell types on which IL-12 exert biological activity are T and NK cells. The major direct biological function of IL-12 on these cell types are the induction of cytokine production, primarily IFN-y, the enhancement of a generation of cytotoxic cells, and a mitogenic effect on antlgen-activated lymphocytes.

In part thromgh its ability to induce IFN-y production, IL-12, produced in response to infections, represents an important functional link between effector cells of innate resistance, phagocytic cells and NK cells, and effector cells of adaptiye resistance, T and B lymphocytes. The ability of IL-12 to induce IFN-y production from circulating. NK cells and at least a proportion of T cells determines a rapid, antigen-independet production of IFN-y during infections. IFN-y acts as a potent enhancer of phagocytic and bacteriocidal activity of phagocytic cells, participating early in infection to activate some of the inflammatory mechanisms that represent the first innate resistance against infection. This antigem-independet activation of phagocytic cells that involves, in addition to IL-12, other monocyte derived cytokines such as TNF and IL-l, has been shown to be important in experimental animals for the resistance against infection by microorganisms such as Limteria monocytogens and Toxoplasma gondii. In addition to this role, early in infection in the antigen-independet phagocytic cell activation, IL-12 has a major effect in the ensuing antigenspecific admptive immune response by facilitating the generation of T helper type 1 (TH-I) response and suppressing a Th-2 response. The presence of IL-12 during antigen stimulation in vitro using human peripheral blood lymphocytes, or both in vivo and in vitro in experimental animals, induces generation of Th-i cells producing IFN-y and IL-2, and inhibits the generation of Th-2 cells producing IL-4. IL-12 induces a direct differentiative effect on Th-cells, priming them for high IFN-y production. ~is priming effect is stable and maintained even when T cell clones are cultured for extended periods in the absence of IL-12, however, IL-12 needs to be present during the first few days after stimulation with antigen or mitogen and established Th clones are resistant to the priming effect of IFN-y. Unlike the generation of high IFN-y producing clones, the IL-12-mediated inhibition of IL-4 producing cells in probably due to selective mechanisms, which may include a selective mitegenic effect of IL-12 for Thl cells and an inhibitory effect of IFN-y on Th2 cell proliferation. In several experimental systems, it has been shown that the enhancing effect of IL-12 on Th-i responses is dependent on production of IF~-y and on the presence of NK cells, possibly needed for the early production of IFN-y. The importance of phagocytic cells and NK cells in determining the characteristics of the Th response during antigenic stimulation indicated that the mechanisms of innate resistance have a determining influence on the subsequent antigen-specific immune response: IL-12 appears to have a key role in mediating the interaction between phagocytic cells. NK cells, and T lympocytes in these proccessem.

Trinchieri, G. 1993, Interleukin-12 and its role in the generation of T E 1 cells, Imm~ol. On peripheral blood monocyte/macrophages, IL-13 behaves like IL-4 in a variety of assays, and shows both similar and contrasting, activties with either IL-10 or IFN-y. IL-13 reduces inflammatory monokine and IL-10 production. It reduces the pyrogen-induced expression of tissue factor (Herbert et el, 1993, FEBS Lett. 328, 268) . It mpregulates manmose receptor and ~C class II expression, while reducing CDI4 expression. IL-13 produces morphological changes (extensive cell processes, aggregation, giant cell formation) which have previously been observed with IL-4. It inhibits NIV-I production (Montaner et el, 1993, I. Exp. Mad. 178, 743) , and interferes with HIV-induced cell fusion.

The activities of IL-13 and IL-4 differ on T-ly•ocyte and large granular lymphocyte population. In particular IL-13 does not exhibit the suppressive effects of IL-4 on THl-type-helper functions (IFNy production) and cytotoxic functions (perforin production). Data will be presented showing that IL-4 and IL-l3 share a common receptor on a nu~er of cell types, but not on T-lympocytes (see also Zmrawski st el. 1993, EMBO J.7. 2663-2670), explaining both thai common and divergent activities. The levels of IL-13~A in activated peripheral blood lymphocytes are greater than or equivalent to those of IL-4 mP~JA and the two ml{NAs are expressed by different T lympocyte population: in particular, T8 cytotoxic lymphocytes express markedly higher levels of IL-13~RNA. In vivo, IL-13 may thus be contributing, significantly to some of the activites previously ascribed to IL-4

IL-10.

Recently, we cloned the human cDNA encoding Interleukin-13 (IL-13). Human IL-13 is a non-glycosylated protein of 132 aa with a molecular mass (Mr) of 10 kD and has biological activities on monocytes and B cells, IL-13, like IL-4, strongly enhanced the expression of class II MHC antigens on monocytes and induced expression of CD23 (Fc~RII). Furthermore, IL-13, like IL-4 and IL~10, inhibited the production of pro-inflammatory cytokines IL-1, ]L-6, TNFo~ and chemokines MIPlc~ and tL-8 by LPS activated monocytes. Both IL-13 and IL-4 enhanced the production of IL-lra by monocytes. In contrast, IL-13 lacked the T cell stimulating activities of IL-4. The responses of monocytes to IL-13 and IL-4 were not additive or synergistic, supporting the hypothesis that IL-13 and IL-4 receptors are complex and share a common component which is important for signal transduction, Taken together, these data indicate that IL-13, IL~4 and IL-10 have important anti-inflammatory activities on human monocytes. In addition, these results indicate that IL-13 is unique in that it can modify inflammatory reactions without stimulating (as does IL-4) or inhibiting (as does IL-10) T cell responses. Transforming growth factor beta I (TGF-fll) belongs to a family of proteins that have potent immunoregulatory properties ranging ftom effects on the growth and dfflerentiafion of primitive stem cells to the differentiated functions of immune system effector cells. TGF-I~I is synthesized as a large secretory precursor polypeptide of 390 amino acids that is inactive until processed to a disulfide linked homodimeric molecule of 1t2 amino acids each. Recently it has been found that TGF-[~I can have autocrine as well as paracrine effects on the immune system indicating that immune cells can activate the inactive secreted form of TGI~-~I. In addition, TGF-I~I has differential intraceUular effects on cell surface receptor modulation, tyrosine phosphorylation and cytokine gene transcription as well as cell mediated cytotoxicity. The systemic administration of TGF-131 has proven beneficial in a variety of animal models such as septic shock, allograft rejection and experimental forms of autoimmunity including collagen induced arthritis and experimental autoimmune encephalomyelitis. Moreover the increased levels of TGF-Ill and other TGF-13 isoforms found in several disease states associated with immunosuppression such as different forms of malignancy, chronic degenerative diseases and AIDS implicate the involvement of TGt~-I~ in the pathogenesis of some diseases. Hopefully, well designed clinical trials will define the potential roles of the TGF-[3 family of proteins in the treatment of diseases of man. Patient development of systemic inflammatory response syndrome (SIS) after severe trauma is accompanied by a wide range of immune aberrations and altered cytokine production. In particular, T lymphocyte proliferation and IFN, t production is suppressed while monocyte (MO) TNFc~, IL-6, PGE2 and TGF~ production is massively increased concomitant to decreased antigen presenting capacity. Recently, two new cytokines, IL-10 and IL-12, have been characterized as critical in regulation of MO TNFa and IL-6, as well as in induction of T lymphocyte proliferation and IFN 7 production. In this study, peripheral blood leukocytes (PBL) from severe trauma patients (Injury Severity Score >35) were analyzed for their IL-10 levels, their in vitro proliferation to mitogen (PHA) and their response after IL-12 addition. Since IL-10 produced either by MO or by T lymphocytes can depress M~ antigen presenting capacity, inhibit T cell IFN,/production and directly diminish T cell proliferation, it might be suggested that immunosuppressed patients' MO and/or T lymphocytes would have increased IL-10 levels. Increased patient IL-10 production might also be resulting from the high levels of TNFa a known stimulator of IL-10. Conversely, since IL-12 augments MO antigenpresenting capacity, Thl induction and proliferation, post-trauma leukocytes might be IL-12 deficient. PBL of trauma patients were compared to normals' PBL, either unstimulated or PttA induced, and their levels of IL-10 found to be dramatically and significantly reduced. Patients' isolated M~, either stimulated with the bacterial cell wall analogue, MDP, or unstimulated, also had depressed IL-10 production concomitant to elevated TNFa production when compared to normals' MO. Mechanisms for the depressed patients' MO IL-10 were explored. Increases in TGF[3 may have partially contributed to the patients' depressed IL-10 level, but elevated PGE2 had no effect. Addition of IL-12 to patients' PBL significantly increased their mitogen responses. These data imply that SIS is characterized by disruption in the interactions between MCi and T lymphocytes so that patients' M~i produce excesses of some mediators (TNFa, IL-6, PGE2) and a dearth of other monokines (IL-12, IL-IO). T lymphocytes are not activated and, therefore, unable to function in both immune defense and monocyte regulation. It is known that lgE receptor-mediated or Ca-ionophore-induced activation of mouse bone marrow-derived mast cells (13MMC) may result in the production of different cytokines including the interleukins (IL) 3, 4, 5, and 6 as well as GM-CSF and TNF-a. In the present study we analyzed the effects of exogeneously applied pro-inflammatory cytokines (IL-1, 1L-6, TNF-c 0 as well as various mast cell growth factors (IL-3, IL-4, IL-9, IL-10, NGF, KL (kit ligand)) on cytokine production in primary mouse BMMC using a standard activation protocol (lxl06 BMMC/ml; ll.uM ionomycin; 24-48h). The actixdties of BMMC supernatants were assessed in specific biological (IL-3, IL-4 IL-6, 1L-9) and/or ELISA assays (IL-5, IL-9). Here we show that homogeneous populations of BMMC (>97%Alcian blue+/Safranln-; in vitro age: 4 weeks) generated in the presence of recombinant (r) raiL-3 from normal BALB/c mice produced modest amounts of 1L-6 and low or undetectable levels of IL-3, -5, and -9 after induction with lp.M ionomycin only. However, a dramatic increase (5-to 10-fold) of these cytokine activities was noted, when in addition to ionomycin also human (11) rlL-la was provided during the induction period. This IL-1 effect was dose dependent with a maximgm at 2-10 U/ml hrlL-la and specific, as pre-incubation (lh) of BMMC with 20 ng/ml hrlL-1 receptor antagonist abolished the action of 2 U/ml hrlL-lcc Similar effects were noted with hrlL-lg or rurlL-lB (lng/ml, respectively), but not with rhlL-6 or rmTNF-~. Both mrlL-10 and hrlL-10 substantially enhanced ionomycin-induced 1L-6 production of BMMC in the absence or presence of IL-1. IL-10 significantly enhanced IL-6 and IL-9 production while decreasing IL-3 activities to abont 50-90% of control levels, when IL-I0 was provided in the presence of IL-l/ionomycin. A monoclonal anti-nfiL-t0 antibody (ascites 1:200) abrogated the effects of mrlL-10. Other mast cell-active cy~okines (]1,-3, IL-4, 1L-9, NGF, or KL) added to ionomycia-or 1L-1/ionomycin-treated BMMC had no major effects on cytokine production. IL-1 and IL-I0 did not induce significant cytokine release in the absence of ionomycin suggesting tlmt Cadependent signalling was required. At doses of 10"6M, dexamethasone, corticosterone, or hydrocortisone almost completely abolished ionomycin/IL-1/lL-10induced cytokine production. The inducer cocktails per se did not interfere with the cytokine bio-assays. In case of IL-9 inducibility of this cytokine in BMMC was confirmed at the mRNA level by Northern blot analysis. Hence our data show that activated mast cells are a source of IL-9 previously recognized as a product of TH2type lymphocytes only. Moreover, our study reveals novel functional roles for I-L-I, IL-10, and ghicecorticoids in the regulation of cytoldne production in mast ceils. Accumulating data suggests that cytokines, peptides involved in regulation of both physiological and pathological immunological responses, predominantly are produced at the local site of antigen stimulation. A new method was used to detect cytokine-producing cells in haman tissue at the protein level. Single-cell production of 17 different httman cytokines, ILia, ILl [3, ILlra, IL2, IL3, IL4, IL5, IL6, ILS, ILl0, GM-CSF, TNFa, IFN 7 and TGF[31.3, was identified by indirect immunohistochemical staining procedures and use of carefully selected cytokine-specific mAb's. Frozen sections were fixed with 4 % paraformaldehyde and permeabilized by 0.1% saponin treatment, eluting cholesterol from the membranes. The intracellular presence of all cytokines except ILl, ILlra (late) and TFG[31_3, could be demonstrated by a characteristic perinuclear configuration in producer cells. In addition, the immunoreactivity extended over a large extracellular area encompassing the producer cell. A localization of the cytokine to the Golgi-organelle was established by use of two culour staining including a haman Golgi complex specific mAb. This staining pattern was only evident in producer cells because injection of recombinant human cytgkines into the tissue caused a membraneous and extracellular staining pattern. Both the extra-and the intracellular types of staining reaction could, however, be blocked by preincubating the cytokine specific mAb with pure human interleukins. Oxygen radicals (OR) directly induce lipid peroxidation, indirectly they trigger adhesion and activation of PMN leukocytes. We investigated whether OR also lead to a release of acute-phase response cytokins such as TNF-alpha, IL-I beta or IL-6 in whole blood cultures to maintain the induced inflammatory reaction.

Methods: Blood samples from healthy volunteers (n=10) were incubated at 37°C. OR were produced by the xanthine oxidase (XO)/ hypoxanthine (HX) system. After 0,30,60,120,240 and 360 minutes plasma levels of TNF-alpha, IL-I beta and IL-6 were determined with ELISA kits.

Results: Under the influence of OR TNF-alpha plasma levels increased from 0,5 pg/ml at 0min to 14pg/ml, 236pg/ml, 343pg/ml after 60, 120 and 240 min. IL-ibeta (0,3pg/ml, 0,6pg/ml, 1,5pg/ml, 81pg/ml and 169pg/ml after 0,60,120,240 and 360min) and IL-6 (0,2pg/ml, l,lpg/ml, 4,1pg/ml, 58pg/ml and 72,9pg/ml after 0,60,120,240 and 360min) plasma levels were increased 60min later than TNF-alpha. Summary: These data suggest that OR do not only play an important role in initial accumulation and activation of PMN leukocytes but also lead to a stimulation of monocytes to produce the acute phase reaction cytokins TNF-alpha, IL-I beta and IL-6 to maintain and strengthen the inflammatory reaction.

Department of General Surgery, SteinhSvelstr.9, 89070 Ulm, Germany

Jan K. Horn MD, Greg A. Hamon MD, Robert H. Mulloy MD, Greg Chen BS, Rebecca Chow BS, and Christof Birkenmaier MD. Transforming growth factor-I~l (TGF-131) is released from inflammatory ceils following injury and in sepsis. In vitro experiments have confirmed that low concentrations of TGF-131 (0.1-1.0 ng/ml) are chemoattractive for monocytes, whereas higher levels of TGF-131 (>5.0 ng/ml) potentiate production of the immunedepressive prostaglandin E 2. Other investigators have shown that TGF-]31 can cause the appearance of CD16 (Fc immunoglobulin receptor) on monocytes exposed to 10 ng/ml of TGF-[~I for 24 hours. Monocytes also express on their surface a glycoprotein that binds complexes of lipopolysaceharide (LPS) and LPSbinding protein (LBP). Such binding is associated with generation of proinflammatory cytokines such as tumor necrosis factor alpha. We have shown that CD14 is depressed in septic patients and therefore we hypothesized that TGF-131 could account for the down-regulation of CD14 observed in these individuals. We incubated normal human monocytes with platelet-derived TGF-[31 for 2 and 24 hours at 37°C and examined ceils for CD14 and CD16 expression using flow cytometry after immunnfluoreseent staining with appropriate monoclonal antibodies. Monocytes were selected on the by usual criteria for size and granularity. Non-viable ceils were excluded with the use of propidium iodide. Two populations of monocytes could be found afcer incubation at 37°C alone. One displaying high density of CD14 had increased fluorescence over the homogeneous expression of CD14 in cells maintained at 4°C (baseline). The other population displayed decreased CD14 expression relative to the baseline cells. TGF-I~I (10-50 ng/ml) caused a shift of ceils from the high density into the low density CD14 population. This trend was observed within 2 hours of incubation and was complete by 24 hours. We observed a net decrease in CD14 expression 0f32% for all subjects studied (p<0.05 vs controls). Phorbol myristate acetate (100 ng/ml) also caused down-regulation of CD14 to a similar degree as TFG-I~I. We also confirmed that monocytes could be induced to express CD16 after incubation with TGF-131 (10 ng/ml) for 24 hours. These studies demonstrate that monocytes incubated with immunodepressive levels of regulation of CD14 by TGF-131 deplete their surface expression of CD14 while generating CD16. This down-regulation of CD14 by TGF-131 correlates with our clinical observations of lower CD14 expression on monocytes obtained from septic patients. For over 25 years, activated T lymphocytes have been considered to be the cellular source of MIF. We recently isolated and cloned the murine homolog of MIF after identifying the specific secretion of this protein by LPSstimulated pituitary cells in vitro and in vivo. However, further experiments showed that MIF protein is detectable both in T-cell deficient (nude) and hypophyseetomized mice, suggesting that yet additional cell types may produce MIF in vivo.

Since monocytes/macrophages are a major source of the cytokines that appear in response to LPS administration, we examined the possibility that MIF also is expressed in cells of the monocyte/macrophage lineage. We found that MIF is expressed constitutively in the murine macrophage-line RAW264.7 and in thioglycollate-elicited peritoneal macrophages. Significant amounts of MIF mRNA (RT-PCR) and protein (Western blotting) were observed in cell lysates. In RAW 264.7 cells, MIF secretion was induced by as little as 10 pg/ml of LPS (E.coli 011 l:B4), peaked at 1 ng/ml, but was not detectable at LPS concentrations >1 txg/ml. Similar data were obtained with elicited macrophages, but higher LPS concentrations were required, unless the cells had been preincubated with IFN 7. Production of MIF by LPS-stimulated (l ng/ml) macrophages peaked at 12 hr. Expression ofMIF mRNA and TNF mRNA by LPS-stimulated RAW 264.7 macrophages was investigated by RT-PCR. As expected TNF mRNA expression increased over the range of LPS concentrations (1 pg/ml to 1 p_g/ml). In contrast, levels of MIF mRNA correlated inversely with LPS concentration. By competitive PCR, MIF mRNA was observed to increase approximately 2-fold after LPS induction (100 pg/ml). MIF secretion also was induced by TNFoc (1 ng/ml) and IFN? (10 IU/ml), but not by IL-113 and IL-6 (up to 10 ng/ml). LPS and IFN 7 had additive effects in inducing MIF secretion.

In separate experiments, macrophages stimulated with recombinant mouse MIF (1 gg/ml) were found to secrete bioactive TNF~ (>750 pg/ml by L929 cytotoxicity).

We conclude that the macrophage is an important albeit overlooked cellular source of MIF in vivo. MIF secretion is induced by LPS, TNFc~ and IFN?. MIF also stimulates macrophages to secrete TNF. Taken together with previous observations that anti-MIF antibody protects against lethal endotoxemia, these data implicate MIF as a critical mediator of inflammation and septic shock. Inflammation is characterized by an exacerbation of proinflammatory cytokine production. Cytokines such as IL-4, IL-10, and TGF8, have been identified as anti-inflammatory mediators thanks to their ability to down regulate the production of IL-1, IL-6, IL-8, TNFc~ by activated monocytes / macrophages.

However, other cells, including polymorphonuclear cells (PMN) do contribute to the release of pro-inflammatory cytokines. We investigated the capacity of the so-called anti-inflammatory cytokines to control the release of IL-8 by activated neutrophils. Human PMN were purified following glucose-dextran sedimentation and FicolI-Hypaque centrifugation. The cells were cultured at 37°C for 24h in the absence or presence of lipopolysaccharide (LPS) or TNFa. IL-8 release was measured in the supernatants using a specific ELISA. Among tested cytokines, IL-10 was the most efficient inhibitor of IL-8 production by LPS-activated PMN. IL-4 was also active, whereas no down regulation was noticed with TGFP~I. When TNFa was used as a triggering agent, none of the cytokine could prevent IL-8 production. Northern analysis are under investigation to precise the level of the IL-4-and IL-10-induced inhibition of IL-8 production by PMN. Our data illustrate that IL-10 and IL-4 possess the capacity to down regulate the production of IL-8 by both monocytes and PMN, whereas TGFB has a more limited inhibitory activity. Ciliary neurotrophic factor (CNTF), a member of the IL-6 superfamily, has recently been shown to promote axonal growth and neuronal healing. CNTF production is also increased during neuronal and muscle damage, associated with soft tissue injury or trauma. We postulated that production of CNTF may explain the loss of skeletal muscM protein that occurs in inflammation. 40 female, Wistar (175-200 gm) rats received either 250 or 25 pg/kg BW s.c. injections of recombinant rat CNTF for seven days, or received sham injections and were freely-fed. Additional animals were pretreated with 10 mg/kg ibuprofen Lp prior to 250 pg/kg BW CNTF. Rats treated with 250 ,ug/kg BW CNTF lost 6.8_+0.4 gms BW as compared to freely-fed controls which gained 17.0_+3.1 gms (p<O.001 by ANOVA and LSD), and ibuprofen treatment had no effect on CNTF-induced weight loss (-5.8_+0.7). Animals pair fed to the high dose CNTF gained body weight (10.2-+3.5), although weight gain was less than seen in freely-fed controls. Rats treated with 25 pg/kg BW had reduced body weight gain (+7.8+_4.4 gin). Food intake was also reduced by 29% in 250 pg/kg CNTF (from 23.7+_0.7 gm/day to 16.8_+0.6 gm/day; p<O.05) and was also unaffected by prior ibuprofen treatment (18.4+-0.6 gm/day). Despite the loss in body weight noted in the high dose CNTF animals, there was no appreciable hepatic acute phase response, since liver weight (8.6+0,1 gms vs. 9.4_+0.3 gins), total protein (54.5+-2.9 gms/L vs 44.8+-3.6 gins/L) and albumin (29.0+1.5 gms/L vs. 26.0+_1.6 gms/L) were not different between CNTF and freely fed animals. We conclude that CNTF causes anorexia, and increases body weight loss in excess of the associated anorexia. Unlike the cachexia associated with IL-1 and TNF infusions, which is associated with an acute phase response, CNTF acts predominantly on food intake and carcass weight through a prostaglandin-independent mechanism and has only minimal acute phase inducing properties, N. Joseph Espat, MD. University of Florida, Department of Surgery, J-Box 100286, Gainesville, FL. 32610.

OF CYTOKINES BY NEUROTROPHIC FACTORS, THE NEUROENDOCRINE SYSTEM AND PHENOTIAZINES.

Cytokines, including TNF and IL-1, have a pathogenetic role in various diseases related to infection and inflammation. The action and/or production of cytokines can be regulated by drugs or endogenous mechanisms that involve the cetral nervous system (CNS)

We found that the antipsychotic chlorpromazine (CPZ) potently inhibits TNF production and protects mice against endotoxic shock. CPZ also inhibits brain TNF production in mice intracerebrally injected with endotoxin, whereas CPZ analogs that do not cross the blood-brain barrier inhibit only perypheral, not brain, TNF.

TNF production and susceptibility to the toxicity of endotoxin, TNF and IL-1 are increased in adrenalectomized mice, indicating that endogenous corticosterone (CS) controls cytokine production and toxicity, in a feedback fashion since endotoxin and cytokines increase serum CS. Increased TNF production was also observed in mice where the hypothalamus-pituitary-adrenal axis was blocked by chronic administration of dexamethasone, following a two-day washout.

Administration of cytokines acting through the gp130 signal transd,uction protein, including IL-6 and ciliary neurotrophic factor, CNTF potentiated IL-l-induced elevation of CS. IL-6 and CNTF also induced hepatic acute-phase proteins. Thus IL-6 and CNTF might have antiinflammatory activity, by potentiating the feedback responses of the neuroendocrine axis.

These data indicate how complex can be the interregulatory relationships between the brain and the immune system, how they can be used to pharmacomodulate cytokine action and how their disregulation might exacerbate IL-1-and TNF-mediated pathologies. Lipopolysaccharide (LPS) constitutes a well established activator of monocytes/ macrophages (MIZI) and murine B lymphocytes. The hydrophobic part of LPS, termed lipid A, represents the endotoxic principle of LPS. We now demonstrate that LPS is also capable of inducing proliferation of human T cells at a dose of 10 to 1000 pg/ml. The specificity of this stimulation was analysed by the following experiments: i) The synthetic hexaacyl Escherichia coiltype lipid A (compound 506) also stimulated human T cells; ii) the synthetic tetraacyl lipid A precursor (compound 406) or the 1phosphono-oxyethyl analogue PE-4, structures known to antagonize LPS-induced monokine production in human MZ, also inhibited LPSinduced T-cell proliferation. The T-cell proliferation expressed the following features: i) CD4 ÷-as welt as CD8 ÷ T cells proliferate in response to LPS, ii) limiting dilution analyses reveal that the frequency of responding cells was between 1/500 to 1/1000 cells; iii) purified T cells alone cannot be stimulated by LPS, but need the help of monocytes. This help cannot be replaced by B cells or fixed monocytes. Furthermore, for activation a direct cell-to-cell contact between T cells and monocyte is neccessary; iv) T cells stimulated with LPS express mRNA for IL-2 and IFNF, but not for IL-4 or IL-5 (determined by PCR). In supernatants, IFNF was detectable (ELISA); v) LPS-activated CD4 ÷ as well as CD8 + T cells express CD25, the high affinity IL-2 receptor. Our results indicate that in contrast to previous reports human T cells respond to LPS with IL-2 receptor expression, lymphokine production, and proliferation. The reactive cells appear to belong to the T,1 cell type, The finding that human T cells can be activated by LPS is of important relevance for the understanding of the pathomechanisms of infections by Gramnegative bacteria. This may lead to new strategies for the therapy of sepsis. We have recently shown that human eosinophils produce mRNA for interleukin (IL)-8 when stimulated with calcium ionophore (Eur. J. Immunol. 23 (1993), 956). We now present evidence that human eosinophils contain preformed IL-8 protein although they do not express mRNA for IL-8 as measured by RT-PCR. IL-8 protein expression was determined using specific anti-human IL-8 monoclonal antibodies by Western blotting, FACS analysis and immunohistochemistry. Moreover, preformed IL-8 was re/eased into supernatant by 99% pure eosinophils after priming with GM-CSF or IL-5 and subsequent 25-min stimulation with PAF, RANTES, ionomycin or PMA, however not with IL-1 or IL-2, as measured by ELISA. This observation implies that activation of protein kinase C and/or intracellular calcium mobilization are necessary events for IL-8 release in human eosinophils. The determined amounts of preformed IL-8 protein was higher in patients with asthma compared to normal individuals suggesting a role for eosinonhi! derived IL-8 in asthma. Since the eosinopnit is the ,,r~:=z~mmant cell in the asthmatic airways, we determined IL-8 concentrations in bronchoaIveolar lavage (BAL. ~ids from normal !i~.~iJvtduals and asthmatic patients. Indeed, IL-8 concentrations in BALs from both groups were similar to those seen in the supernatants released by activated eosinophils. The role for IL-8 in asthma remains to be determined. 

Based on a well established rat model named activity-induced ulceration (ASU) or activity based anorexia (ABA), we have developed a rat model for chronic stress. Male rats were exposed to a feeding schedule in order to reduce body weight by 20-30% within 14 days. While one group of rats was kept sedentary, the other had access to a running wheel. Food restricted rats allowed to exercise developed highly increased activity as compared to ad libitum fed rats in wheels (ca. 15 versus 2 kin/day). They developed elevated plasma corticosterone levels (30_+5gg/dl) at their circadian peak, increased adrenal weight and decreased thymus and spleen weight as compared to control rats and weight matched sedentary rats, establishing this paradigm as a model of chronic stress. No differences in plasma ACTH were seen among the various groups, suggesting increased sensitivity of the adrenal glands to ACTH. The effect of chronic stress accompanied by hypercorticism on baseline and endotoxin-induced cytokine levels was studied in this model. Under baseline conditions,splenic IL-1 a was suppressed in both food-restricted groups. This was more pronounced in rats with access to a running wheel. IL-6 and TNF activity in plasma was not affected. After administration of 1001.tg/kg lipopolysaccharide (LPS), TNF activity in plasma was suppressed by food restriction but there were no differences between sedentary or exercising rats. In addition, no differences were found among groups for corticosterone, spleen IL-lc~ and plasma IL-6 activity. We conclude, that the food restriction-induced hyperactive rat model is a useful model for studies on the effects of chronic stress. In this model, under basal conditions IL-le~ in tissues may be subject to downregulation by glucocorticoids, whereas after LPS-stimulation of cytokine production only TNF activity is regulated by fast acting feedback loop between cytokines and the hypothalamo-pituitary-adrenal axis (HPA-axis). This is supported by strong correlations between plasma TNF activity and corticosterone in stressed rats. Cytokine antagonists modulate cytokine actions and it appears that a balance between production of cytokines and cytokine antagonists is important to control of host responses. Recent observations show that during myocardial infarction there is an increase in circulating cytokines such as tumor necrosis factor (TNF) and interleukin 6 (IL-6). We tested the idea that compensatory increases in cytokine antagonists likewise occur and investigated changes in plasma concentrations of TNF and interleukin i (IL-1) and those of their specific antagonists interleukin-1 receptor antagonist (IL-1 ra) and soluble tumor necrosis factor receptor type I (s TNF r-I). We also measured plasma concentrations of ~-melanocyte-stimulating hormone (~-MSH), an endogenous neuropeptide that occurs in pituitary, brain, skin, gut and other tissues. When administered to laboratory animals, this peptide has potent antiinflammatory and antipyretic activity, perhaps by mimicking an endogenous modulatory influence on cytokine actions. Samples were obtained from 20 patients with acute myocardial infarction at presentation in the Coronary Unit, every three hours for 24 hours, and daily thereafter until discharge. We found elevations in the plasma cytokines IL-1 and TNF only in a proportion of subjects whereas cytokine antagonists c~-MSH, IL-lra, and sTNFr were elevated in all patients, o~-MSH was elevated at presentation but it decreased progressively in successive tests, reaching normal values within 24 hr whereas sTNFr and 1L-lra peaked at 3-6 hr or later. There were significant correlations between plasma concentrations of cytokine antagonists and enzymes released from injured myocardium, including creatine kinase (CK), aspartate serum transferase (AST), and lactate dehydrogenase (LDH). The data show that cytokine antagonists increase in the circulation of patients with acute myocardial infarction likely to modulate prninflammatory effects of cytokines. Objects of the study: Interleukin-6 (IL-6) is a multifunctionai cytokine known to be involved in important homeostatic processes. IL-6 levels are increased in many pathological situations, and correlates with disease activity and patient outcome. To exert its effect IL-6 binds to its receptor on the membrane of its target cell. This receptor is also present in a soluble form in plasma and in urine. In this study we quantitatively measured the availability of soluble IL-6 receptor (slL-6R) in biological fluids in healthy volunteers. Materials and methods: Blood, urine, cerebrospinal fluid (CSF), and synovial fluid (SF) are obtained from healthy volunteers, sIL-6R and IL-6 are measured using ELISA's. Both ELISA's are tested for interference of sIL-6R mid IL-6. The effect of slL-6R in the IL-6 bio-assay (B9) is tested. Results: Baseline levels are (mean + SD): 76.6 -+ 19.3 ng/ml in serum, 3.7 -+ 1.3 ng/ml in urine, 1.64 _+ 0.4 ng/ml in CSF and 11.6 + 3.3 ng/ml in SF. There is no interference of slL-6R in the IL-6 ELISA or bio-assay and no interference of 1L-6 in the sIL-6R ELISA. Conclusions: In all investigated biological fluids slL-6R can be found. These data provide baseline values for investigations concerning pathological situations. Both ELISA's described are useful tools to do these investigations. Trauma-associated immune aberrations coincide with augmented release of immunoregulatory and proinflammatory cytokines. The present study examines the effect of thermal trauma on the capacity for specific (receptormediated) response of lymphoid cells to interleukin 1 (IL 1) and to turnout necrosis factor ~x (TNFc0.

Methods. Bum patients (N=24, 25->75 % total body surface area) were studied weekly up to 42 days post-injury. The Limulus Amoebocyte Lysate (LAL) test was used to measure plasma endotoxin levels. The percentage of IL 1~-and TNFcz-binding T(CD 5) lymphocytes was assessed by flow cytometry analysis. Levels of IL 1 receptor antagonist (IL lra) in patients' plasma and cultures of peripheral blood ceils (PBC) were determined by immunoassay.

Results. Plasma endotoxin concentrations were significantly (P<0.05) increased up to 3 weeks post-bum (means 3.5+1 in non-surviving and 2.6+ 0.6 U/ml in surviving patients vs < 1U/ml in the control). Within 2 weeks of bum, the percentage ofT ceils expressing receptors for TNFa and IL 1[~ constitutively was elevated (by 10-15 fold). In contrast, the capacity for de novo receptor expression by activated PBC was reduced. Serum levels of IL Ira were significantly increased (range 2.5-40x10 J pg/ml vs <0.2x10 J pg/ml in the control). In all patients, high concentrations of IL lm were released spontaneously in unstimulated cultures of adherent ceils (range 20-30x10 -3 pg/ml vs 4-8x10J pg/ml in the control). However, its secretion was decreased in LPS-stimulated parallel preparations.

Conclusions. In the bum patient, susceptibility to the immunoregulatory effect of TNFcz and tL 1~ may be modulated by infection-related products. Alterations in the capacity for receptor expression and secretion of 1L lra may affect IL 1-regulated biological responses including specific immune reactions. While studies suggest that IL-10 is an important lymphokine involved in cell-mediated immunity, little is known about this mediator's role in HEM-induced immunesuppression.

Our aims, therefore, were to determine: I) if IL-10 contributes to depressed T-cell responses seen following HEM; and 2) how other agents, known to play a role in HEM, effect IL-10 release.

To study this, C3H/HeN mice were bled to and maintained at a MAP of 35 mmHg for 1 h and then adequately resuscitated.

Mice were killed 2 h post-HEM to obtain splenic T-cells (nylon-wool purified).

IL-10's immunosuppressant role was demonstrated by the ability of monoclenal antibody (Mab) to IL-10 to markedly improve the T-cell proliferative response [2.5 #g The marked increase in capacity of T-cells from HEM mice to produce IL-10 was significantly reduced by treatment with either IBU or Mabs.

Since IBU, TGF-~, as well as IL-6 are all reported to directly/indirectly influence prostanoid synthesis, this implies that eicosanoids play a major role in inducing IL-10 release by T-cells following HEM which depresses T-cell function. The mechanisms underlying immunosuppression induced by thermal injury and alcohol ingestion are in part due to cytokine dysregulatinn. IL-10 down-regulates production of eytokines by maerophages and may be an important regulator of the initiation of the immune response. IL-10 has also been demonstrated to inhibit the production of NO by macrophages. This study examined the alterations in eytokine production and effect of inhibition of NO production on immunologic function in a routine thermal injury model. METHODS: Balb/c mice (n=40) were randomized to 4 groups: Saline-Sham(NS-Sham), Alcohol-Sham(EtOH-Sham), NS-Bum, EtOH-Bum. Animals received 20% EtOH or NS daily for 14 days by gavage. A 30% full thickness bum was induced 4 hrs after the last dose of EtOH or NS. Animals were resuscitated, then sacrificed 4 days post bum. Splenic lymphocytes were cultured for 5 days with LPS, and LPS with two concentrations of N-Monomethyl-L-Arginine, a nitric oxide inhibitor (L-NMMA 2.5ug/ml, 10ug/ml). Splenocyte production of IL-10, Interferon-gamma, IL-2, PGE2 were measured, and lymphocyte proliferative response examined. RESULTS: IL-10 production was significantly suppressed in thermal injury. Exogenous L-NMMA normalized the suppression of 11.-10 in a dose-dependent manner, indicating nitric oxide may modulate IL-10 and Interferon-gamma production in thermal injury. IL-10 production is normal in EtOH-Burn animals.

CONCLUSION: IL-10 and Interferon-gamma production is altered in this murine thermal injury model, and may contribute to this injury-induced immunosuppression. Inhibition of NO synthesis normalizes IL-10 production and should be investigated further as an immanomodalator in thermal injury. Surgery, infection and inflammation results in the production of pro-inflammatory cytokines which mediate metabolic and immunologic host responses. The aim of this study was to characterise the elaboration of cytokine release following a variety of surgical procedures. Twenty one patients undergoing elective intermediate, hip, knee and major gastrointestinal surgery were studied. Levels of interleukin-1 (I1-1), interleukin-6 (I1-6), the interleukin-1 receptor antagonist (I1-1RA) and the acute phase C-reactive protein (CRP) were measured in bloods drawn 0, 1, 2, 4, 6, 12, 24 and 48 hours following operation. A portion of the results are shown (mean -+ sem).

135+37 1738-+843 145_+24 One and two factor ANOVA; *p<0.0002, #p<0.0001, §p<0.0009, ¶p<0.0014, for differences between groups I1-1 was not detected at any time point. Both II-IRA and I1-6 increased after surgery. Maximum responses occurred following major GIT and hip surgery, minimal responses were seen after intermediate and knee surgery. II-IRA levels increased within two hours and remained elevated for 24 hours; the B-IRA increase was a thousand fold greater than the rise in I1-6 levels. I1-6 levels increased up to 24 hours after surgery. CRP levels reflected maximum II-IRA and I1-6 levels (R2=.676, p<0.0001 and R2=.282, p<0.04 respectively).

High II-1RA and I1-6 levels reflect major surgery, however the II-IRA response is more rapid and of greater magnitude. The strong I1-1RA correlation with CRP may indicate that this regulatory cytokine is itself a mediator of host responses to surgery.

Dept. of Surgery, Meath/Adelaide Hospitals, Heytesbury St., Dublin 8, Ireland.

CHANGE OF IL-6 AND SOLUBLE IL-6 RECEPTOR LEVELS AFTER SURGERY S. Hisano, K. Sakamoto, S. Mita, T. Ishiko, M. Ogawa [Objectives] Under surgical stress, IL-6 plays a main role in producing acute phase proteins and contributes to host defense mechanism. Soluble IL-6 receptor (slL-6R) is considered to be agonistic to IL-6, unlike other soluble type receptors of cytokines. Here we measured IL-6 and slL-6R levels in the serum and drain fluid from surgical field in order to investigate the changes of IL-6 and slL-6R after surgery and their origins.

[Materials and Methods] Serum and drain fluid samples from 21 cases (6 of esophagectomy and 15 of gastrectomy ) were serially collected before and after surgery. IL-6 and slL-6R levels were measured by ELISA.

[Results] (1) Serum IL-6 : All cases reached the maximum level on POD-l, more precisely 3-6 hours after operation. (2) IL-6 in the drain : Maximal IL-6 levels in the drain were recognized 3-6 hours after operation, at almost the same time as serum IL-6. Furthermore the IL-6 values in the drain were much higher, about 100 times, than those in serum. (3) slL-6R in the serum : All cases reached minimum levels 3-12 hours after operation and recovered to the preoperative levels a few days later (decrease ratio : 43.9+5.6~,, range : 9-84~'). (4) slL-6R in the drain : slL-6R levels in the drain showed almost the same value and change as serum slL-6R.

[Conclusions] (1) IL-6 is produced from the cells gathering around operative fields whereas slL-6R is considered to be produced in the cells which do not gather around the operative fields. (2) There may be a mechanism that down-regulates slL-6R in the early stage of surgery. [Objectives] IL-6 plays an important role in host defense in the early stage after surgery. In the present study, we examined changes in IL-6 concentration after major thoracoabdominal surgery and elucidated the effect of surgical trauma and factors influencing postoperative elevation of serum IL-6.

[Materials and Methods] Thirty-eight patients undergoing elective surgery of the thoracoabdomen were classified into 6 groups according to the location of the operation. Bloods and drain fluids were serially obtained and samples were frozen until measured, keukocytes were simultaneously collected for Northern blot analysis. Concentration of IL-6 was measured by ELISA and IL-6 mRNA was detected by Northern blotting after total RNA was extracted by the Acid Guanidium Phenol Chloroform method.

[Results] (1) Serum IL-6 levels reached the maximum concentration on the 1st postoperative day in all patients. (2) The IL-6 peak was significantly correlated with surgical trauma as defined by the operation length and the volume of blood loss during operation (r=0.554, p<0.01, r=0.427, p<0.01, respectively). (3) The peak concentration of serum IL-6 in patients undergoing esophagectomy was significantly higher than in those undergoing pancreaticoduodenectomy (p<0.05), despite a similar degree of surgical trauma. (4) Peak 1L-6 concentration observed in a patient who underwent esophagectomy was about 100 fold greater in the drain fluid of thorax than in the peripheral blood. (5) IL-6 mRNA was demonstrated in leukocytes from thoracic and abdominal exudate at 6, 24 and 48 hours after surgery. In contrast, IL-6 mRNA could not be detected in leukocytes from the peripheral blood.

[Conclusion] IL-6 is mainly produced in the operative field and subsequently enter the peripheral blood to induce cytokinemia. The operation length, volume of blood loss and thoracotomy are factors influencing the concentration of cytokine in the blood. Zaragoza SPAIN Age may be an important factor influencing the function of immunocompeteut cells releasing cytokines after both accidental and surgical trauma The aim of the present paper is to ascertain if patients (PTS) over 60 years old show a different serum level cytokine pattern than PTS under 60 after a standard surgical procedure considered as a "medium strength trauma". PATIENTS AND METHODS: 33 PTS(22 females 11 males)with gallstone disease were perspectively studied, PTS were allotted in two groups: Gr.A:17 PTS under 60 years(mean age:49.5+-7)Gr.B:16 PTS over 60 years(mean age:65.5_+5). All PTS underwent cholecystectomy and cholangiography. 6 PTS in Gr.A and 7 PTS in Gr. B underwent common duct exploration. 1 spbintercctomy was performed in each group. On the day of surgery (PRE) and on the 1st and 7th postoperative day(lEO, 7PO) : percentages of CD19, CD5, CD4, CD8 and CD16 cells we measured by means of flow cytometry using MoAb. and levels of IL-1, IL-4, IL-6 and TNF "in vivo" by ELISA using MoAb. RESULTS: ERE: CD16% was 5.3_+3 in Gr.A and 7. 5 

Objectives of the study. After surgery for esophageal cancer multiple organ damage has been reported to be caused by polymorphonuclear leukocyte (PMN)-mediated injury. We measured serum granulocyte colony-stimulating factor (G-CSF) and interleukin 8 (IL-8) levels to determine a role of G-CSF and IL-8 in PMN function after surgery for esophageal cancer.

Materials and Methods. Peripheral PMN counts, peripheral PMN chemiluminescence, serum G-CSF levels, and serum IL-8 levels were measured before and after surgery in 12 patients with esophageal cancer (EC), and 12 patients of gastric cancer (GC). Esophagectomy with thoracotomy and laparotomy were performed for patients with EC, while subtotal gastrectomy with laparotomy were performed for patients with GC.

Results. Peripheral PMN counts (p<0.01) and peripheral PMN chemiluminescence (p<0.05) of patients with EC were significantly decreased compared to those of patients with GC at 4 and 8 hours after surgery. Serum G-CSF levels of patients with EC were significantly (p<0.01) increased compared to those of patients with GC at 4 and 8 hours after surgery. Serum IL-8 levels of patients with EC were significantly (p<0.01) increased compared to those of patients with GC at 4, 8 and 24 hours after surgery. Significant inverse correlations (p<0.0l) between peripheral PMN count and serum G-CSF and IL-8 levels were seen at 4 hours after surgery.

Conclusion. These results suggest that many circulating PMNs, which are excessively activated by G-CSF and IL-8, may adhere to the endotherial cells and then migrate into the tissues, and cause multiple organ damage after surgery for esophageal cancer.

ImmunNogical changes in patients with severe brain trauma receive increasing attention since morbidity and mortality ere still high. Interleukin-6 (IL-6) was previously detected in the cerebrospinal fluid (CSF) during different pathologies of the nervous system (1, 2, 3). in our study we monitored iL-6 and Nerve Growth Factor (NGF) production in the CSF after human brain trauma. Since astrocytes within the brain constitute one of the major cell type contributing to the inflammatory response through the release of cytokines and other factors after injury, we investigated the functional relationship of IL-6 and NGF on a single cell niveau using cultured astrocytes. Methods CSF was obtained from patients with severe brain injury (Glasgow Coma Score (GCS) <8 and CT abnormatities or GCS <8 over 6 hours) after implantation of intraventricular ICP monitoring device for therapeutic purpose and collected over 24 hours CSF and serum. IL-6 and NGF were assayed by ELISA. Astrocytes were isolated from neonatal mouse brain as described (4) . NGF production by cultured astrocytes was measured by ELISA in the presence of CSF, IL-6 and IL-6 antibody. Astrocyte migration was tested in a chemstaxis chamber. Results 17 head trauma patients were included in this study (approved by the University Hospital Medical Ethics Board) and the CSF was obtained through intraventricular catheters. High levels of IL-6 were detected in the CSF of these patients when compared to serum during the first days after brain trauma. Furthermore NGF could be found inside the intracerebral compartment. CSF containing high levels of IL-6 could stimulate NGF production in cultured astrocytes. This effect could be [nhibited partially by IL-6 antibodies, Purified IL-6 exposed to cultured astrocytes in vitro, stimulated the migratory activity of these cells in a dose response fashion.

IL-6 was found in the CSF of brain injured patients, suggesting a role for this cytokine in the pathophysiology of brain injury. Since astrocytes are involved in maintaining the homeostasis of the brain, we further investigated the possible role o1 IL-6 on astrocyte functions, IL-6 promoted NGF production in vivo and in vitro, thus contributing to neuronal cell survival and regeneration. Furthermore IL-6 stimulated astrocyte migration in a dose response fashion, potentially contributing to astrocytosis following brain injury and inflammation, These results show that IL-6 represents a key cytokine in traumatic human brain injury with possible systemic effects, which are at preserlt under investigation. We studied a) the role of TNF and b) the therapeutic effect of a mab to TNF with regard to haemorrhagic shock (HS) related ,pathophysiologic alterations and mortality in rats. Method: A prolonged HS was induced by bleeding to a blood pressure of 30-35 mmHg for 180 pin followed by reinfusion of shed blood (SB) and resuscitation with two times of SB volume of Ringer's lactate over 50 rain. 15 animals received a bolus dose (20 mg/kg) of TNF mab (Celltech, Berkshire, UK) at 15 min after resuscitation (TN3). The control group (n = 15) was treated similar to the TN3 group but received Ringer's lactate (CON). Results: At 180 min the prolonged HS resulted in a metabolic acidosis indicated by a significant decrease of blood pH (7.16 + 0.04), HCO3-(16.3 ___ 2.0 mM), and base excess (-12.5 + 1.9 raM) values with pCO2 (47.9 + 7.7 mmHg) and pO2 (136.2 + 20.1 mmHg) in the TN3 with no difference to the CON group. Immediately after resuscitation (230 min) plasma endotoxin levels were found to be increased in both groups (48.8 + 73.3 in TN3 vs 30.1 _ 25.7 pg/ml in CON group) . Prior to the treatment with TNF mab (230 min) there was also no difference between plasma TNF levels of the two groups (42.1 + 60.7 in TN3 vs 60 + 141.8 pg/ml in CON group). Treatment with the TNF mab at 65 rain post-HS improved the 48hour survival rate to 73.3 % as compared to 26.7 % in the control group. Macropathologic evaluations revealed frequency of intestinal bleeding in onIy 2 animals in the TN3 vs 11 in the CON group. No bleeding in the kidneys was found in the TN3 but in 4 rats in the CON group. The significant increase in lung wet weight observed in non-survivors in the CON (n = 11) was prevented in animals which died in the TN3 (n = 4) group ((9.3 +_ 2.4 vs 6.3 +_ 1.0 g/kg). Conclusion: Our data suggest that TNF formation induced by HS in rats is an important mediator for pathophysiologic alterations leading to multi organ failure and lethality. Antibodies to TNF might be a useful agent in the treatment of haemorrhagic shock related disorders. 55-+8 n=ll*$ 15-+5 n=7 78_+22 n=5* * p<0.05 vs BASELINE :~p<0.05 no anesthesia vs anesthesia Thus 1) TNF production increased 2-3 fold by 48-72 hrs following trauma in unstimulated blood, but was reduced or not changed after LPS stimulation, so circulating leukocytes are probably not an important source of TNF post trauma; 2) antiCD18 had no obvious effect on TNF production in unstimulated or LPS stimulated blood, relative to vehicle, which suggests that the protective mechanism of antiCD18 does not involve TNF suppression; 3) fentanyl anesthesia at 72 hrs following trauma unexpectedly decreased LPS-evoked TNF production, which suggests that anesthesia alone can influence an inflammatory response. Proinflamrnato~ cytokines have been shown to play a signific~t role in the pathogenesis of sepsis, which is a very common occurrence in born injury. TNFa is infrequently detected in the blood of burned patients, the ability to detect the shed receptors of sTNFg has not been determined. Serial serum Mmples from burn patients were collected from the time of admission until death from septic shock. These samples were analyzed using an enzyme-linked immunosorbent assay (ELISA) for sTNFr, L-Ira, TNF-a, and IL-Ib. The patients ranged in age from 26 to 72 yeas of age. The percentages of bum ranged from 40% -90%.

Cytokine concenlrntions vmled from patient to padent irrespective of bum size. TNFa levels were consistentiy in the range of30pgJml -150pg/ml. Peaks in the TNFa values were above 90pg/ml and were also associated with a peak in the sTNFr levels. These levels began at <10,000pghnl within the In,st 6 Ins of injury and gradually increased with time. Clinically. ti~ appearance of eytoklnes was independent of positive wound, blood, or respiratory cultures however peak values in TNFa and sTNFr were ~ialed with a fluid requirnmenL Levels of IL-I ra were also elevated independent of clinical findings as well as extent of injury. In PL5 there is a significant corresponding peak in IL-trn (>40~8/ml) at the same time as T/~:a and sTNFr levels. We aimed to characterise the pattern of secretion of interleukin-1 beta 0L-II3), intefleukin-6 (IL-6) and tumour necrosis factor alpha (TNFa) in multiply injured patients and to relate these results to their clinical condition and outcome. Two hourly blood samples were taken from ten patients from the time of injury until 48 hours. Cytokine levels were measured using sandwich enzyme-linked immunosorbent assays (ELISAs). Injury Severity Scores (ISS) were calculated and haemorrhage was assessed from the blood transfusion requirement over the 48 hours.

Patients' ages ranged from 17 to 86 years. ISS varied from 9 to 50 and transfusion requirement from 0 to 14 units. Five patients died after the study period. ]1,-6 was raised in 9/10 patients (max level 12,500 pg/ml) but was unrelated to condition or outcome. 5/8 showed a rise in IL-1B (max level 90 pg/ml) which was negatively correlated to ISS (I=-0.789, p<0.02). TNFa was raised in 2/10 (max level 95 pg/ml). Peak TNFc~ was positively correlated with ISS (1=0.676, p<0.05) and haemorrhage (I=0.602 but p<0.5). IL-IB and TNFa production was mutually exclusive. There was no common cytokine profile for these patients. Unlike elective surgery there was no correlation between peak 11,-6 and severity of injury: tissue damage may not be the stimulus for the cytokine response to multiple injury. Periods of ischemia or hypoxia produce endothelial damage in peripheral organs. Tumor necrosis factor-alpha (TNF) plays a central role for regulation of endothelial physiology during septic events, taking influence on vascular permeability and coagulant activity [1] . Animal experiments demonstrated a synergism between hypoxia and septic shock on letality, leading to the hypothesis that low oxygen tension leads to enhanced sensitivity of target cells for TNF [2] . Radioligand binding studies with ~261odid-TNF on cultured human endothelial cells were performed after incubation in several environmental oxygen tensions (PC2) for 12 hours. Data were achieved by nonlinear regression of an idealized saturation curve according to the equation: B = n " K./(1 + K,); B = totally bound TNF; K,: association constant (concentration for half-maximal binding); n: number of binding sites per cell. p_O0o (mm H¢I):

_K, (nM}: n (molecules/cell): 130 -150

1.83 ± 0.15 2600 _+ 500 55 -70

1.27 ± 0.15 3600 + 500 25 -35 0,41 ± 0.13 4800 -+ 500 10-15 0.18 + 0.07 5900 -+ 300 Presented are calculated values on the idealized curve + 95 % percentiles. Hypoxia induces enhanced binding of TNF to specific receptors on the endothelial cell surface in a time-and dose-dependent manner by a mechanism, which is not dependent on oxygen radicals, as shown by additional protocols with radical-scavenging drugs. With respect to former findings about a correlation between growth and TNF receptor affinity [3] , these data lead to the hypothesis that enhanced TNF binding during hypoxia is due to a biochemical conversion of the receptor protein from the low affinity to the high affinity state, possibly by posttranslational phosphorylation of the binding protein by intracel)ular kinases. The proposed involvement of TNF-dependent pathways in pathogenesis of organ dysfunction and multiple organ failure after hypoxia/ischemia may provide a basis for understanding the initiation of hypoxic vascular injury, as manifested by increased permeability and prothrombotic tendency, and, thus, merits further attention. The levels of activity of circulating cytokines (ILl, IL-6 and TNF-alpha) which are believed to play important regulatory role in response to trauma are determined (by hioassays and respective anti-cytokine antibodies) in mice and rats subjected to scald injury ion C,10 see, °0v BSA, LD 50) and (80 C, 30 see, 20 ~ B ~^)~ , respectively.

Biphasic increase of cytokine activity was noted in mice: initial increase of IL-I and IL-6, 1-3 hr following injury and of TRY activity 6hr after scald, followed by elevated levels of IL-I and IL-6 at 12hr, with tendency of decrease of activity at later time points. Increased activity of TNF was noted 24 hr following injury, In rats, initial, short-lived increase of IL-i and TNF activity was detected lhr following injury, folowed by increase on days I and 3

postburn. IL-6 increase peaked 3-12 hr after scalding and levels remained elevated 3-5 days following injury. Similar kinetics of appearance of proinflammatory cytokines (IL-I and TNF-alpha) both in lethal and ncnlethal injury concomitant with differential profile of circulating IL-6 activity (early,short-lived increase and later slow decrease of activity in lethal burn injury) with late persistent high levels of activity in nonlethai injury demonstrated in the present study highlight the need for investigation the relationship of these cytokines in burn-injury induced inflammation.

Zikica Jovicic,lnstitute for Medical Research, MMA,Crnotravska 17,11002 Belgrade~Yu.

Asadullah K (1), Woiciechowsky C (2), LiebenthaI C (1), Doecke WD (1), Volk HD (1), Vogel S (2), v. Baehr R (1); Depts. of Med. Immunology (1) and Neurosurgery (2) , Medical School (Char#d), Humboldt University Berlin, FRG In patients after polytrauma or major abdominal surgery a hyperinflammatory phase seems to be followed by the development of a phase of monocyte inactivation. The latter is charaeterised by a decrease of monocytic HLA-DR expression and a shift to anti-inflammatory cytokine production. As shown, by us and others, this phenomenon indicates severe immunodepression with a high risk of infection. However, the mechanisms leading to monocyte inactivation in the above mentioned syndromes may be multiple. To elucidate the influence of a selective, sterile trauma to the central nervous system (CNS) on immune reactivity the neurosurgieal patient is an interesting model. Initially, 30 patients who developed a systemic inflammatory response syndrome following neurosurgery were analysed. In all of them a marked decrease of monocytic HLA-DR expression was observed soon after the operation. These results suggest that neurosurgery alone can induce immunodepression and lead us to conduct a prospective study, in which we closely monitored l0 patients undergoing neurosurgery from the first preoperative day until at least day 6 after the operation. HLA-DR expression was decreased hi all patients to various extent only hours after surgery. In one patient only we found a persistently reduced HLA-DR expression and this was the only patient to develop sepsis syndrome. This suggests that a prolonged, postoperatively decreased HLA-DR expression is predictive of infection following CNS trauma. In order to assess, whether a decrease of HLA-DR expression was associated with a preceding inflammatory response, local cytokine release in the CNS was compared with systemic cytokine release. For this purpose, paired samples of earebrospinal fluid (CSF) from a vantricle drainage and peripheral blood plasma were obtained. In the CSF extremely elevated futerleakin (IL)-6 levels, peaking already a few hours after the operation were found. In plasma, by eontrast, IL-6( and TNF-alpha) was detectable not until days later and only if infection was present. The antiinflammatory ILI-RA, on the other hand, was also present in CSF but peaked after IL-6 and was detectable in peripheral plasma too. We believe there is an association between the inflammatory response in the CNS and the following depression of HLA-DR expression on peripheral blood monocytes. Our results suggest that even a sterile CNS-trauma by itself may contribute to general immunodepressinn leading to septic complications. The aim of this study was to evaluate the effect of haemorrhagic shock (HS) a) on total capacity of the host, and b) the circulating blood cells to produce TNF immediately after bleeding. In vivo studies: Baboons were subjected to a limited oxygen deficit (180-200 ml/kg) hypotension phase (mean arterial pressure = MAP of 35-40 mmHg for 2-4 hours followed by adequate resuscitation). Rats subjected to HS (MAP of 30-35 mmHg for 90 rain followed by reinfusion of shed blood and fluid resuscitation) were challenged with endotoxin (1 ~g/kg i.v.) at the end of shock (RHS group). The control group (RCO) received the same dose of endotoxin as RHS group but without prior bleeding. In Vitro studies: Whole blood (WB) obtained from both baboons and rats before and at the end of HS were incubated with endotoxin (100 ng/ml) for 2 hrs at 37 °C. Results: At 90 min post-LPS challenge we found significantly higher plasma TNF levels in rats that were subjected to HS prior to the endotoxin challenge as compared to the control group (956 _+ 585 vs 276 + 199 pg/ml) . After HS the TPC was significantly decreased in in vitro stimulated CBC of both rats (12 + 4 post-HS vs 164 + 89 ng TNF/ml pre-HS) and baboons (8 ± 16 post-HS vs 233 ± 188 pg TNF/ml pre-HS). In contrast, the IL-6 productive capacity was increased in baboons CBC (not yet analysed in rats) stimulated at the end of HS (224 ± 106 pre-vs 1545 ±_ 1254 pg IL-6/ml post-HS). Conclusion: From our data we suggest that despite of down regulation of the CBC to produce TNF the overall TPC is enhanced at the early stage of I-IS. With regard to the related literature (Chaudry's group) it can be assumed that among the macrophage/monocyte populations, as the main source only the Kupffer cells (KC) exhibit enhanced TNF production capacity following haemorrhage. The mechanisms of down/up regulation of cytokine response of CBC and/or KC following HS remain to be examined.

D. Eg~er, S. Geuenich °, C. Dertzlin~er °, E. Schmitt*, R. Mailhammer, H Ehrenreich #, P. Drrmer, and L. H01mer GSF-Instimt fox Experimentelle H~znatologie, °Medizinische Kliulk III, Klinikum Groghadern, Munich, *Institut for Immunologic, Johannes Gutenberg Universit/it, Malnz, and #Psychiatrische K/in& der Georg-Aagust-Universi~t, Grttingen, Germany.

It has been shown previously (Ehranreich et al., 1992, New Biol. 4: 147 ) that mouse bone marrow-derived mast cells (BMMC) synthesize and secrete endothelin-1 (ET-I) and express ETA-type endothelin receptors (ETA). So far, however, no functions of ET-1/ET A in BMMC have been described. In the present study we investigated the effect of exogeneously administered ET-1 on the release of histamine, serotonin, and leukotriene C 4 (LTC4) by primary mouse BMMC (in vitro age: 4 weeks) caltured with different recombinant mttrine cytokines (interleukin 3 (IL-3) and/or kit ligand (KL) in the presence or absence of IL4) for two weeks prior to activation. ET-1 (5x10 -8 to lxl0 -6 M) induCed an extremely rapid (_<lmin) and dose-dependent release of histamine and serotonin (up to 20% and 16% of total mediator content, respectively, comparable to IgE/Ag-indueed mediator release) from BMMC cultured with IL-3 and IL-4. Only when cultured in the additional presence of IL-4 rather than in medium containing KL or KL plus IL-3 alone, BMMC released histamine and serotoaln upon challenge with ET-1.Furthermore, ET-1 stimulated LTC 4 biosynthesis up to 4-fold in BMMC cultured in the presence of IL-4. In contrast, the peptide was without major effect on LTC4 biosynthesis in BMMC culturd without IL-4. The release of histamine and sereotonin was not altered if BMMC were preincubated for i h with IL-4 prior to activation with ET-I, suggesting that a differentiation process rather than a functional priming effect had been initiated by ]L-4. ET-1 (10"6M) -induced histamine and serotouln release from BMMC was inhibited by an ETA-specific antagonist (cyclic ID-Asp-Pro-D-Val-Leu-D-Tel ) in a dose-dependent manner, with a complete inhibition observed at an antagonist concentration of 10 -8 M. Crude peritoneal cells (containing 2-3% serosal mast cells) obtained from normal BALB/c-mice released 87% histamine upon challenge with ET-1 (10 -6 M; 20 rain). Taken together, these resu/ts demonstrate that ET-1 can directly function as a histamine and serotohin secretagogue and as a stimulator of LTC 4 production in mast cells. IL-4 appears to be involved in the differentiation of immature mast cell prec~trsors towards an ET-l-reactive functional phenotype. When inflammatory reactions are advanced, type II phospholipase A2 (type II PLA2) is produced in high levels by various cells at the site of inflammation. Cytokines such as TNF-a and IL-I activate type II PLA2 and promote the production of eicosanoid, which is one of the mechanisms by which they enhance the inflammatory reaction. In the present study, the blood levels of type II PEA2, endotoxin, TNF-c~, IL-6 and IL-8 were determined in patients with sepsis to examine the relationship between these levels and the patients' pathological conditions. The levels of type II PLA2 and TNF-a in these patients were 218.3 ± 179.9 ng/ml and 98.5 ± 92.1 pg/ml, respectively, the two parameters being significantly correlated (r = 0.6460, p = 0.0001 ). There were also a significant correlation between the level of type II PLA2 an IL-6 (r = 0.4262, p = 0.0188). There was no significant correlation between the level of type II PLA2 and IL-8. These in vivo findings suggest that TNF-a may be involved in the production of type II PLA2. 

A. Filzmaver, G. Reisbach, E. Schmitt °, R. Mailhammer, and L. Hiittner. GSF-Iustitut ff~r Experimentelle H~imatologie, Munich, and °Iustitut fttr Immunelone , Johannes Gutenberg Universit~t, Mainz, Germany The product of c-kit, a transmembrane tyrosin ldnase receptor, and a corresponding kit ligand (KL) are critically involved in the control of different hemopoietic cell lineages including the proliferation, maturation, migration, and function of mast cells. It has been reported previously that interleukin (IL)-4 down-regulates kit receptors in human primary myeloid leukemic cells and in a human mast celt line (Sinaber et al. 1991, J. Immunol. 147: 4224) . Transforming growth factor (TGF) gl, was also found to down-modulate c-kit mRNA and kit receptor proteins in human AML blasts, a mechanism probabely contributing to the anti-proliferative effect of TGFIH on KL-stimulated AML ceils in vitro (de Vos et at. 1993, Cancer Res. 53: 3638) . In porcine endothelial cells transfected with a retroviral construct carrying the human (hu) e-kit gena, exogenous hu KL transiently down-regulated kit receptors (Blume-Jeusen et al. 1991, EMBO J 10: 4121) . In the light of these previous findings we analyzed the effects of these exogenously applied cytokines (IL-4, TGFgl, KL) on kit receptor expression and KL-mediated proliferation and chemotactic migration of primary mouse bone marrow-derived mast cells (BMMC) (in vitro age: 4 weeks). Our results show that in BMMC cultures initiated with recombinant (r) marine (mu) IL-3 the additional presence of rmulL-4 for 3 or 14 days did not change the expression of kit protein (assessed by FACS analysis with the anti-mouse c-kit antibody ACK-2) nor KL-mediated proliferation or chemotaxis. In contrast, IL-4 syeergistically increased KL-stimulated growth of BMMC in a short-term proliferation assay (MTT-tes0. Pre-incubation of BMMC with rhuTGFl~ 1 (5.0, 1.0, or 0.1 ng/ml) for 24h had no effect on kit receptor expression, although TGFfil at concentrations of 0.5 -5.0 ng/ml completely suppressed the proliferative action of KL on these ceils. Similar anti-proliferative effects of TGFg 1 were observed in various factor-dependent mast cell lines stimulated with different mast cell growth factors (IL-3, 1I,-4, IL-9, and/or IL-10). Moreover, pre-incuhation (24h) of BMMC with TGF-gl (>500 pg/ml) significantly enhanced spontaneous undirected cell movement (chemokinesis) and synergistically increased IL-3-or KL-induced chemetaxis. When BMMC were preAncuhated with rmuKL (200 ng/ml) for 1, 3. or 5 days, a transient down-modulation of kit receptors with a maximum effect on day 1 was demonstrated by FACS analysis and correlated well with a decreased chemotactic response of these cells. In conclusion our results show that neither IL-4 nor TGFI31 affect expression of kit receptors in primary murine BMMC. It is reasonable to suggest that c-kit expression is controlled in a cell type-specific manner.Interestingly, TGFgl is obviously able to dissect the proliferative from the migrational signal transducted by KL in these cells. Objectives of the study: Antisense strategies using DNA-otigonucleofides (ODN) to modulate the cytokine response are presently under investigation. ODN are thought to act very specifically with little or no relevant negative side effects. We now report that ODN unspeeifically protect WEHI 164 cells from TNF-mediated cytolysis.

Material and Methods: WEHI 164 subclone 13 ceils (5 x 105), that are highly sensitive to the cytolytic activity of TNF, were grown on 96-well culture plates in RPM11640 medium. After 24 hours, phosphorothioate(PS)and partially PS-modified-ODN as well as phesphodiester-ODN (20-29bp) were added (0.1, 1 and 10 pM). Four hours after incubation with ODN, ce(I lysis was induced by recombinant murina TNF. After 18 hours the plates were washed and stained with crystal violet Cell lysis was determined by reading the absorbance (ABS) at 590 nm.

Results: WEHI 164 ceils incubated with TNF (1-8ng/ml) were completely lysed after 18 hours (0% ABS). Interestingly, WEHI cells incubated with TNF and ODN resisted complete lysis, eg cells incubated with 2.5ng/ml TNF and 11JM ODN showed still 30% of the absorbance observed in control ceils without TNF (100% ABS). The protective effect of ODN started at 0.1pM, reached a maximum at 1,uM, and diminished at 101JM. With increasing amounts of TNF the protective effect of QDN decreased and no protection was detectable at 8ng TNF per ml Conclusions: DNA-oligonucleotides were found to unspecifically inhibit TNF-induced cytolysis. We hypothesize, that this protective effect of QDN results from an inhibition of the binding of TNF to its receptor, or from interference of ODN with the subsequent signal transduction mechanisms. As a consequence, to discriminate the specific effect of ODN in biologic systems, several control ODN should be used. Secondly, whether DNA released by degradation of tumor cells or leukocytes can significantly impair tumor-and immune-defense mechanisms merits further investigation Dr. med. Michael Meisner, Institut for Anaesthesiologie der Universitat Erlangen-NQmberg, Krankenhausstral~e 12, D-91054 Erlangen.

In this study we investigated the involvement of serine protease and free radical generation in the systemic release of tumor necrosis factor-alpha (TNF) and interieukin I(IL-1), in the sepsis model of lipopolysaccharide (LPS, 5mg/kg i.p.) induced hepatitis in galactosamine (GAIN, 18rag/mouse, i.p.) sensitized mice.

Treatment of GAIN-sensitized mice with LPS (GAIN/LPS) led to dramatic increase in serum cytokine (TNF and IL-I) ievels and transaminase activity at 3 hr and 8 hr respectively. Pretreatment of serine protease inhibitor, c~jantitrypsin (a j-AT, 50mg/kg i.p.), 30 rains prior to GAIN/LPS treatment, fully protected the animals against the hepatotoxic challenge with significantly reduced serum TNF and IL-1 levels. In order to block and scavenge superoxide generation, the mice were pretreated with xanthine oxidase inhibitor, allopurinol (AL, 2 x 100mg/kg i.p.) and pyran polymer-conjugated superoxide dismutase (SOD, 2 x 200 unit/mouse i.v) r6spectively. Pretreatment with AL and SOD (24 and 1 hr prior to GAIN/LPS) prevented GAIN/LPS hepatitis and blocked LPS induced released of TNF and IL-1 into serum of the mice.

The protective agents like cq-AT or AL/SOD did not protect the mice against th~ hPp~totoxi£ ch~llPn-e indllee4 b'~ th~ recombinant mmlse TNF-o' (0.3 ~/rno~e J.p.) ~d oI~LPS 1~ CAlN-.~dlfa%aed mlce. It-l cett~aged la TNF (x/GAIN treated mJde was not detectable in animals pretreated with oq-AT or AL/SOD.

Our study suggests that a serine protease sensitive to cq-antitrypsin is responsible in regulating TNF release, possibly by proteolytic cleavage of a TNF-precursor or membrane bound TNF. In addition our evidence suggest that the balance of extracellular protease/antiprotease activity may be regulated by free radical generation, possible superoxide anion, resulting in inactivation of the antiprotease. IL-1 release may be subsequent to TNF release. Objective: During sepsis one can observe a dramatically impaired production of proinflammatory cytokines like the Tumor necrosis factor alpha (TNF-a), Interleukin I-alpha (IL-la), Intedeukin I-beta (IL-I&) and Interferon gamma (IF~) upon in vitro stimulation of circulating cells. However there is also evidence of a decreased ability to produce cytokines in other immuno-deficient states. In this study we compared the capacity to secrete proinflammatory cytokines upon in vitro stimulation of patients in severe sepsis and patients with malignant tumors. Methods: Heparinized blood samples of ten patients (62+ 14 years) in severe sepsis (sepsis score > 15 according to E}ebute and Stoner) were drawn at onset of disease, From fifteen patients with solid growing carcinoma (59+19 years) blood was drawn at diagnosis prior to any therapy. Controls were obtained from fifteen healthy volunteers. 50 pl of whole blood were incubated either with 450/4 of a standard medium or with 400 pl of a standard medium and 50 pl of phytohemagglutinin (PHA) a potent mitogen. After an incubation period of 24 hours plasma concentrations of TNF-a, IL-la, IL-16 and IF-~ were determined by ELISA. Comments: Our results suggest that down-regulation of cytokine secretion or of cell responsiveness to non-specific mitogens during sepsis has occurred. We observe a similar phenomenon for the group of carcinoma patients vs control significant for stimulated TNF-a and stimulated IF-T. Sustained immunological interactions between tumorcells and cytokine producing cells could effect responsiveness of the latter, A general increased immuno-tolerant state in patients with carcinoma has to be discussed. However we found significant differences between sepsis and cancer concerning the in vitro capacity of responsable cells to produce IL-la and IL-I#. The dramatically decrease of the ability to produce IL-I upon in vitro stimulation could be more sensitive for a septic state than stimulated TNF-a or IF-3,. Objective: Tumor necrosis factor alpha (TNF-a) has been implicated as a central mediator of sepsis and its sequelae. Increased systemic levels of this cytoklne seem to be correlated with severity of sepsis and outcome. However mechanism of action and metabolism of TNF-G are not fully understood. In most studies blood samples for TNF-a determinations are obtained either by peripheral venipuncture, a central venous catheter or by an indwelling arterial catheter. Very often blood samples are taken in different manners within the same study. In this study we measured circulating TNF-a and the amount of TNF-a released upon in vitro stimulation in arterial and central venous blood.

Methods: Heparlnized arterial and central venous blood samples of ten patients (6 males, 4 females, mean age 62+_14) with severe sepsis (sepsis score > 15, Elebute and Stoner} were drawn on day 1,3,6,8, and 14 of disease. Blood was immediately placed on ice and processed within 1 hour. 50 pl of whole blood were incubated with 450 pl RPMI-medium supplemented with antibiotics and L-Glutamlne or with 400 pl of RPMI-medium and 50 pl phytohemagglutinin (PHA) a potent mitogen. After an incubation period of 24 hours samples were centrifuged and plasma was harvested and stored at -30 ° celsius before assessment of TNF-a concentration by ELISA. Statistical analysis was performed with the paired Student-T-test.

Results: We found a significant difference (p < 0,05) for circulating mean arterial TNF-a concentration (237 pg/ml _+ 34 SEM} and central venous TNF-a (203 pg/ml +_ 29 SEM). Upon in vitro stimulation there was also a significant difference (p < 0,002) between released arterial TNF-~' {746 pg/ml _+ 88 SEM) and venous TNF-a (637 pg/ml +_ 76 SEMI. Conclusions: These results are difficult to interprete but could reflect the influence of paO 2 and SaO 2 on TNF a release. It could also be the result of different concentrations of TNF-o release influencing factors like for example endotoxin, interferon-F or prostaglandin. A possible pulmonary and/or a hepatic metabolism of TNF-n and TNF-a producing cells cannot be ruled out. However for better interpretations of TNF-a release in septic states it is necessary to use either arterial or venous blood samples. Early inflammatory processes following trauma and/or infections were found to be associated with the secretion of high amounts of proinflammatory cytokines. Besides intedeukin-t (IL-1), tumor necrosis factor-a (TNF-c 0 and interleukin-8 (IL-8) the multifunctional cytokine intedeukin-6 (IL-6) was described to be a central regulatory element of the primary cellular and humeral defence reaction. The previously described close temporal correlation of pathologically elevated IL-6-concentrations and the extracellulary release of lysosomal enzymes from activated pelymorphnuclear neutrophils suggests, that IL-6 may be a potential substrate of these preteases. The serine preteases elastase (EC 3.4.21.37 ) and cathepsin G (EC 3.4.21.20) derived from the azurophilic granules were assumed to be mainly involved in unspecific proteolysis at sites of inflammation by cleavage of structural as well as soluble proteins at random sites, if the inhibitory potential is decreased. The possible proteolytic activity of elastase and cathepsin G toward the proinflammatory cytokine interleukin-6 (IL-6) was investigated. The addition of purified neutrephil elastase and cathepsin G to recombinant human IL-6 leads to a rapid sequential degradation in vitro. At least two intermediate products could be detected by silver staining and western blotting following protein separation under reducing conditions. The serine protease inhibitor G-anitrypsin was shown to prevent the proteolytical degradation of intedeukin-6. Furthermore the loss of the biological activity of both, recombinant and natural human IL-6, was demonstrated by determination of the capacity of protease-treated IL-6 to stimulate hybddoma growth (7TD1 bioassay). These data suggest a possible downregulation of pathologically elevated IL-6 levels by proteolytic activity of extracellulary released enzymes at sites of inflammation. The aim of the study was to compare circulating levels of three cytokines -IL-1, IL-6, 11_-8 -between critically ill subjects who developed Gram-negative sepsis and who did not. MATERIALS AND METHODS: The patient population consisted of 39 patients admitted to an Intensive Cars Unit, with different underlying diseases. Sepsis diagnosis was given according to pre-estabilished cdteda. Nineteen cases were enrolled in sepsis group, twenty in control group. Serum sampling was collected in sterile tubes at study entry and every three days until study dismissal. Serum concentrations of IL-1, 11_-6 and IL-8 were measured using commercially available test kits, based on the dual immunometric sandwich principle. RESULTS: The causative patogens of sepsis were: Pseudomonas aeruginosa, Acinetobacter, Eseherichia co~i, Serratia marceseens, Proteus mirobilis and Citrobacter freundL The time of observation was equal to 12 days, for a total of four tests performed (to, tl, t2, t3). I1.-1 was not detected in any samples. The serological profiles of the two cytokines 11.-6 and 11_-8 were similar; augmented levels were found at study entry and throughout the observation period, peaking at t3 and decreasing at t4. However, in patients with sepsis, IL-6 and 11_-8 concentrations were significantly higher in respect to control group. CONCLUSION: Our observations shown that in ICU patients increased IL-6 and IL-8 release may be induced by cdtical illness; however, in subjects in which sepsis occurred, IL-6 and IL-8 production appears more significantly elevated, suggesting a role of IL-6 and 11_-8 in the pathophysiology of sepsis. The fact that II. Objective: To check whether continuous veno-venous haemofiltration (CVVH) could remove the cytokines, namely tumour necrosis factor alpha (TNFc 0 and interleukin 6 (IL-6) from the circulation of critically ill patients with sepsis ad multiple organ failure (MOF).

Setting: The intensive therapy unit of the medical school teaching hospital.

Patients: Nine critically ill patients with sepsis and MOF treated with CVVH.

Methods: Blood samples were collected before the CVVH had been started. Then, blood and Ultrafiltrate samples were collected simultaneously after 4 hours and every 24 hour. TNFct and IL-6 levels were measured using the bioassays with cell lines WEHI-164 ci13 and 7TD1, respectively. Other data were recorded from the patient notes and intensive therapy unit charts.

Results: No measurable concentrations of TNFct were detected in either blood or ultrafiltrate samples. IL-6 was found in all the patients' plasma samples and five patients' (55.5%) ultrafiltrate samples. The IL-6 blood level ranged from 17.4 to 3061.7 U/ml (mean 518.2, SD 686.8). The IL-6 level in positive ultrafiltrate samples ranged from 21.0 to 1150.2 U/ml (mean 252.2, SD 401.8).

Conclusions: Our preliminary results suggest that IL-6 is present in bloodstream of septic patients. We assume we coUld not detect TNFa in any sample because we usually started observations when septic state had developed. CVVH could extract cytokines from the circulating blood. It remains under discussion, whether that extraction may be beneficial to patients with MOF. The pattern of some significant cytokines TNF, IL-1 and IL-6 and their pharmacomodulation were evaluated in an experimental model of polimicrobial sepsis induced in CD-1 mice by cecal ligation and puncture (CLP) in order to understand their roles. This model of sepsis, which resembles the clinical situation of bowel perforation, was also compared with that induced by administration of pure endotoxin (LPS).

TNF was detectable in serum and tissues during the first 4h with a peak 2h after CLP at a significantly lower level than after LPS. IL-1 was measurable in serum only after 24h, significantly increased in spleen and liver after 4 and 8h and in mesenteric lymphonodes from 8 to 24h after CLP compared with shammice. IL-6 was significantly increased in serum throughout the first 16h after CLP.

Pretreatment with dexamethasone (DEX), ibuprofen (IBU) and nitro-L-arginine (N-ARG) significantly reduced the survival time while chlorpromazine (CPZ) and TNF did not affect it. Only the antibiotics and pentoxifylline (PTX) significantly increased the survival in CLP. However CPZ and DEX protected from LPS-mor~ality.

In conclusion, by inhibiting TNF with DEX, CPZ, PTX a reduced, unchanged and increased survival time was observed and by increasing TNF with IBU and TNF administration the survival was decreased or unchanged respectively suggesting that the modulation of this cytokine does not seem to play a significant role in CLP unlike LPS_ Moreover the negative effects of IBU and N-ARG suggest an important and protective role by prostaglandins and NO in CLP. To gain more insigths on the contribution of TNF~, IL-I~ and IF 7 to LPS toxicity, we explored the time-course of the cytokine production in Ealb/C mice given different doses, from the lethal (= 3 LD50) to the sublethal (= 1/3 LD50) of three different LPS (E.coli OIII:B4 and 026:B6; P.aeruginosa R261) endowed with different degree of toxicity Cytokines were measured in serum and organs with specific ELISAs up to I0 h after LPS administration. Results demonstrate that i) circulating and organ levels of TNF~ do not reflect LPS toxicity. In fact, the lethal dose of LPS 026:B6 induced as much TNF~ as the sublethal dose of LPS 0111:B4; furthermore, LPS R261, whose cytokine inducing capability is far lower than that of LPS from E.coli, induced higher TNF~ levels at the sublethal than at the lethal dose. In addition, policlonal anti TNF Ab, that were able to protect mice from E.coli LPS induced mortality, failed in mice treated with LPS R261 2) Circulating IL-I~ levels are generally low and increase significantly only in muribond animals. On the contrary, in spleen and lung very high levels of IL-I~ are persistent from I to 8 h post LPS administration Moreover, the treatment with 6 mgr of neutralizing policlonal anti IL-I~ Ab, did not modify survival in LPS challenged mice. 3) Circulating and organ levels of IF 7 are proportional to the dose and degree of toxicity of all the administered LPS even if LPS R261 was again a less efficient cytokine inducer than LPS from E.coli. CsA is an immunos~ppressive drug, able to inhibit gene expression for many cytokines, including IF 7. To study the effect of cytokines modulation on LPS toxicity, CsA was administered to mice twice at the oral dose of i00 mg/kg before the challenge with LPS. Mice were monitored in terms of mortality and TNF~, IL-I~ and IF 7 production. Together with the total ablation of IF7, the strong reduction of TNFu and unmodified IL-I~ levels, a significant increase of LPS toxicity was also observed. These results suggest the hypothesis that the numerous factors that jointly mediate LPS toxic effects, can also be protective, the final outcome depending on their relative ratio rather than on the absolute amount Interleukin-1 (IL-1) mediates the septic shock syndrome and affects intestinal secretion in vitro. We studied the intestinal production of IL-t and its effects on diarrhea during endotoxic shock. CD-1 mice were randomized to 15 mg/kg E.coli 0111:B4 LPS or saline infusion (I.P. or I.V.). Diarrhea invariably occurred following LPS infusion. Mice were sacrificed at 0, 30', lh, 2.5h, 4h, 6h, 12h, and 24h (3 mice/group/time-point). The small bowel was compressed and the intestinal contents were weighed and expressed per g SB weight. The small (SB) and large bowels (LB) were eventually frozen, weighed, and homogenized for either cytosolic protein or total RNA. IL-I~ (cell-associated agonist) was measured with a radioimmunoassay specific for mouse IL-l~ (detection limit 100 pg/mL) and expressed as ng/g weight + SEM (lowest detectable amount 1 ng/gWT). Northern analysis of total RNA and in sfu hybridization of paraformaldehyde-fixed frozen tissue were done with [0~-32p]-Iabeled mouse IL-lc~ cDNA probes. Only SB had IL-I~ constitutively present (6.2 + 1.6 ng/gWT). LPS I.P. or I.V. induced elevation of IL-lc¢ in both organs in a biphasic pattern; LPS I.V. induced 3-fold more IL-I~ than LPS I.P. Following LPS I.P., IL-I~ in SB was 19.2 + 0.6 ng/gWT at lh, reached maximal levels at 2.5h (31.8 -+ 0.5 ng/gW-I) and returned to baseline at 24h. Saline controls maintained their constitutive IL-I~ levels. SB had 2fold more IL-1 ¢ than LB and identical kinetics, but LB showed a clearer doseresponse. Northern analysis of SB-total RNA showed induction of IL-I~ mRNA by LPS in correlation with IL-lc¢ kinetics. IL-I~ mRNA producing cells were mononuclear cells in the lamina propda and epithelial cells at the bottom of the crypts of Ueberkuhn. Mucus and fluid were increased in the small bowel post-LPS in correlation with intestinal IL-lc~ kinetics (R2=0.9). Separate mice were pretreated with saline I.P. orthe IL-1 receptor antagonist (IRAP, 30 mg/kg bolus I.P.) and were challenged 20 rain later with 1.5 mg/kg LPS I.P. or saline I.P. Specific blockade of IL-1 by IRAP decreased intestinal secretion at 4h and 6h post-LPS challenge (p<_. 0.05, Student's-t-test).

These data indicate that local (intrinsic) intestinal IL-I~ mediates sepsis-induced intestinal changes. Inflammatory cytokines initiate the host response to endotoxemia, causing severe physiological and hemodynamic changes which may lead to septic shock. Among the regulatory systems that play an important rote in controlling host inflammatory responses is the pituitary. It has been known for many years for example, that hypophysectomized animals are extremely sensitive to LPS lethality. While investigating the possibility that protective, pituitary mediators might explain this phenomenon, we identified the cytoldne MIF to be a specific secretory product produced by pituitary cells in vitro and in vivo after LPS challenge. Analysis of serum MIF levels in control, T-cell deficient (nude), and hypophysectomized mice revealed that pituitary-derived MIF contributes significantly to the rise in serum MIF that occurs after LPS administration. Of note, pituitary MIF content (0.05% of total pituitary protein) and peak serum MIF levels (80-340 ng/ml) were determined to be within the range observed for other pituitary hormones that are released after pituitary stimulation.

To investigate a possible beneficial role for MIF in septic shock, we co-injected mice with purified, recombinant murine MIF (rMIF) together with LPS (15 mg/kg). Surprisingly, rMIF markedly potentiated LPS lethality compared to control mice that were injected with LPS alone (85% vs. 35%, P = 0.003). To confirm these results, mice were treated with anti-rMIF antibody prior to injection of a high dose of LPS (17.5 mg/kg). Anti-rMIF antibody fully protected mice against LPS lethality, increasing survival from 50% to 100% (P = 0.0004). Serum levels of TNF,~, the first cytokinc that appears in the circulation after LPS challenge, were reduced by 38.0 _+ 9.5% in anti-rMIF-treated mice.

We conclude that pituitary derived MIF contributes significantly to circulating MIF in the post-acute response in endotoxemia and may act in concert with other pituitary mediators to regulate both pro-and antiinflammatory effects. Moreover, MIF may play a critical regulatory role in the systemic host response in septic shock. Our results suggest that anti-rMIF antibody might be of potential therapeutic use in the treatment of septic shock. Although anti-Interleukin-1 (IL-1) antibodies and IL-1 receptor antagonist have been shown to improve survival in animal models of endotoxemia and abrogate the lethal effects of TNF, the presence of IL-1 in the serum does not correlate well with outcome. We hypothesized that this may be because IL-1 acts mainly in a paracrine fashion and is metabolized before it diffuses into the circulation. Methods: We measured the IL-I~ mRNA expression with the differential reverse transcription polymerase chain reaction (RT-PCR) using g-actin as internal standard in the peritoneal macrophages and lung tissue in normal controls and mice after cecal ligation and puncture (CLP). CLP resembles human intra-abdominal sepsis in that it is characterized by very slight elevations of serum IL-1 levels. Results: IL-lg mRNA levels after CLP are expressed as % of normal (mean+SEM, n=5 In several experimental models of infection exacerbation of disease was observed, when infected animals were depleted of tuaJor necrosis factor (TNF). After sublethal cecal ligation and puncture (CLP) leading to peritonitis and sepsis the survival of mice also critically depends on TNF as demonstrated in earlier studies, when CLP-treated mice injected with anti-TNF antibody died, whereas mice injected with a control antibody survived after CLP (Echtenacher et al. 1990, J. Inununol. 145: 3762) . From a panel of different cell types (macrophages, neutrophils, T lymphocytes, natural killer cells, mast cells) able to produce TNF upon activation~ the mast cell is apparantly the only one capable of storing in cytoplasmic granules preformed TNF-ct which is rapidly released following challenge. In the present study-we analyzed serum TNF after LPS injections as well as the outcome of CLP in severely mast cell deficient mutant mice (WAV v) as compared to syngeaeic wild-type littermates (+/+). We proposed that concentrations and/or kinetics of serum TNF should be different between WAVv mutants and wild-type mice, if mast cell-derived TNF significantly contributes to the rise in serum TNF levels following systemic stimulation with endotoxin. Although similar levels of increased TNF were detected in the sera of both genotypes after 1 and 2 hours of LPS injection (100 btg/ 0.5 ml / mouse i. p.), mast ceil-deficient mice indeed showed decreased serum TNF levels 30 iron after injection amounting to only 12 to 25% of the concentrations observed in the corresponding sera of normal wildtype mice. In the CLP model of septic peritonitis we found that mast celldeficient mutant mice were dramatically more sensitive to CLP than syngeneic normal mice resulting in 92% mortality in W/W v versus 8% mortality in +/+ mice 2.5 days after initiation of CLP. Further experiments with W/W v mutants selectively reconstituted with cultured bone marrow-derived mast cells from normal syngeneic wild-type mice and the use of an antibody specifically blocking the action of TNF tn vivo should clarify a potential protective function of mast cells in this model of septic peritonitis. Interleukin-10 (IL-10) inhibits cytokine production, including tumor necrosis factor (TNF), by lipopolysaccharide (LPS)-aetivated maerophages. We recently observed that LPS injection (E.coli 055:B5, 100 gg ip) into Balb/c mice induces the rapid release of circulating IL-10 (17±9 U/ml at 90 min). Blocking endogenous IL-10 using monocIonal antibody (JES5-2A5, 2 mg, 2h before LPS) resulted in a massive increase in TNF production (107±47 in LPS+anti-IL-10 treated mice vs 6±2 ng/ml in LPS alone, p<0.05, n=4 to 5 mice per group) and an enhanced LPS-induccd lethality (53% vs 7% in anti-IL-10+LPS or LPS alone respectively, p=0.01, n=15 mice per group). Irrelevant IgG1 rat monoclonal antibody (LO-DNP) did not influence neither TNF production nor lethality associated with endotoxin shock. This led us to study the production of IL-10 during human septicemia. Plasma samples were obtained from 65 patients with gramnegative (GNS, n=22) or gram-positive septicemia (GPS, n=43) and from 20 healthy volunteers. Among these patients, 17 suffered from septic shock at the time of sampling. IL-10 levels were measured by ELISA (detection limit: I 1 pghrd). We found that 35 patients (54%) had increased IL-10 plasma levels (range 12 to 2740 pg/nd). Patients with GPS had IL-10 levels similar to the ones observed in GNS (median: 12 vs 17.5 pg/M, respectively). Patients with septic shock had higher IL-10 values (median: 58 pg/ml) than septicemic patients without shock (11 pg/ml, p=0.002). No IL-10 was detected in plasma from healthy volunteers.

We conclude that IL-10 is produced daring human septicemia. Our experimental data suggest that IL-10 might be involved in the control of the inflammatory response induced by bacterial products.

Dr Arnand Marchant, Immunology Department, Hopital Erasme, 808 Route de Lennik, 1070 Brussels, Belgium. 

To provide information about the role of TNF in sepsis and MODS we measured TNF and sTNFR-I levels in septic patients and investigated if there is a relation between plasma concentration of these molecules and the severity of sepsis evaluated by two scores (APACHE 1I and SSS). Patients and Melhods: 20 septic patients fullfilling sepsis criteria of American College of Chest Physician and Society of Critical Care Medicine were studied. TNF-cc and STNFR-I (60kDa) were measured by enzyme immuneassays (norms1 values = 13+8 pg/mL and 1.06_+0A ng/mL respectively). Results: The mean TNF and sTNFR-I values for each patient (mean+SD) were 93+64 pg/mL and 11.1+5.5 ng/mL respectively. These values are approximately seven and ten times greater than those observed in normal healthy volunteers (P<0.001). Mean TNF concentrations for each patient were significantly greater in non survivors (112+52 vs 64_+74 pg/ml P<0.05); sTNFR-I levels also were greater in this group, but the difference was not statistically significant (12.6+5.2 vs 8.8_+5.5 ng/ml). Plasma TNF and sTNFR-I concentrations were significantly correlated (r = 0.65 P<0.005). Mean TNF levels were significantly correlated with APACHE II (r = 0.49 P<0.05) and SSS (r = 0.54 P<O.01); sTNFR-I levels were likewise correlated with both scores (r = 0.80 P<0.001 and r = 0.76 P<0.001 respectively). TNF and sTNFqGI levels increase with the number of dysfuctional organs; in particular TNF and STNFR-I increase when hemodynamics deteriorate and kidney failure ensues. Conclusions: The correlations between TNF and sTNFR-I levels and sepsis severity scores suggests that TNF is involved in sepsis and may influence the occurrence of MODS and finally clinical outcome. TNF and sTNFR-I are in fact especially increased when MODS occurs. In particular their concentrations are elevated when cardiovascular system and kidney failure ensue, suggesting that TNF can have a direct role in hemodynamic derangements caused by sepsis and that kidneys are involved in TNF and sTNFR-I excretion. LIF is a pleiotropic cytokine that has been implicated in the pathogenesis of sepsis in animal models. We conducted a prospective study in 33 septic patients to correlate LIF plasma levels with patient's clinical and biological data (among them IL-6). Plasma was drawn daily from day one to .ten and LIF presence was determined with a specific ELISA and with the DAI,a bioassay (sensitivity of 100 pg/ml and 1 ngml respectively). Risk factor associated wftt~ with elevated LIF plasma level was determined by an univariate analysis. A multivariate stepwise logistic regression analysis was subsequently performed with a BMDP statistical software. LIF was detected with the ELISA and with the bioassay in 11 and 2 out of the 33 patients, respectively. LIF peak plasma levels were statistically associated with high creatmin levels, temperature and IL-6. Multivariate regression analysis showed that high serum creatinin level was the major independent risk factor associated with elevated LIFplasma levels. This correlation was also found for IL-6. Peak plasma levels of both LIF and IL-6 correlated inversely with patients survival duration. Cox's analysis for plasma levels of LIF >480 pg/ml yelded a hazard ratio of [exp (1.86)=6.44]. Our study indicates that LIF levels were associated with clinical and biological parameters of illness severity and significantly increased (cut-off value 480 pg/mI) in patients with fatal outcome. Current consensus exists about the central role of tumor necrosis factor (TNF) alpha in initiating the systemic inflammatory response syndrome (SIRS). A correlation with SIRS has inconsistently been found. TNF effects its pleiotropic reactions upon two distinct cellular receptors. Soluble extracel]ular fragments of the human 55 kDa TNF receptor (sTNFRI) and the 75 kDa receptor (sTNFRII) are detectable in the circulation. The kinetics of these endogenously produced TNF-inhibitors were measured to evaluate their role in patients with SIRS. Fourteen patients of an operative ICU were included with the diagnossis of SIRS (mean APACHE II score: 20 points). Serial blood samples were obtained within 12 h after diagnosis of SIRS, every 6 hrs for the first 48 hrs and every 12 hrs thereafter until patients died or recovered. Soluble TNFRI and sTNFRII were assayed by an enzymed-linked immunological binding assay. Soluble TNFRI and II could be detected in all samples with a significantly higher level (p<O,O01) compared to healthy controls (normal value: TNFRh 1,5 ng/ml; TNFRII: 2,3 ng/ml). The serum concentration of TNFRI (14,1 ng/ml) as well as TNFRII (18,2 ng/ml) were significantly higher in nonsurvivors (n=6) compared to survivors (8,6 and 9,5 ng/ml; n=8) at the onset of SIRS and during the course of the inflammation response (p<0,05). A positive correlation exists between both receptors (r-0,57; p < 0,01 ). Increasing levels and maximal values of sTNFRI (59 ng/ml) and I1 (82 ng/mI) were detected only in patients who died. The serum concentrations of patients who recovered did not change and are remaining at the initial level. TNF alpha bioactivity was detected in 10 out of 14 patients. No significant correlation exists between the concentration of both receptors and TNF alpha as well as the APACHE II score at the onset of SIRS. Elevated serum concentration of sTNFRI and 11 are found in patients with SIRS. These naturally occuring antagonists reflect the inflammatory process. The measurement of soluble TNF receptor levels may be useful in monitorina SIRS and in the prediction of outcome. 11-2 is given to patients with advanced melanoma or renal carcinoma. However, this therapy is accompanied by severe side effects, including the Vascular Leak Syndrome (VLS) that closely resembles septic shock. To investigate the involvement of cytokines and phospholipase A z in the pathogenesis of the VLS we collected serial plasma samples from patients during the first 24 hours after administration of IL-2. In these patients we monitored temperature, blood pressure and heart rate and assessed levels of the cytokines TNF, IL-6 and IL-8 using immuno assays. In these plasma samples we also measured phospholipase A 2 activity using an enzyme activity assay, since this enzyme is beleived to play a major role in the generation of lipid mediators. We demonstrate that during IL-2 therapy the cytokines TNF, IL-6, and IL-8 are produced and that the release of these cytokines is followed by an induction of phospholipase Az activity in the plasma of these patients, We further discuss the role of these mediators in the development of the vascular leak syndrome observed in patients receiving IL-2. Objectives of the study: Sepsis caused by Candida albicans (CA) is no longer a clinical rarity but a common consequence of immunosuppression and surgery, and is associated with high mortality. Tumor necrosis faetor-c¢ (TNF-<~), interlcukin-lg (IL-lg) and interleukin-6 (IL-6) are thought to play an important role in the pathogenesis of the sepsis syndrome. We therefore studied the in vitro production of TNF-<x, IL-IB and IL-6 by human peripheral blood mononuclear cells (PBMC) stimulated by CA compared to stimulation by bacterial lipopolysaccharide (LPS). Materials and methods: PBMC were isolated from venous blood of 5 healthy volunteers and cultured at a concentration of 2.5 • 10 ~ eells/ml in culture medium with a dose-range of either heat-killed CA or LPS. Supernatants were harvested after 24 hours. Using ELISA, TNF-c~, IL-IB and IL-6 concentrations were measured. CA was isolated from a patient with CA sepsis and cultured under sterile conditions. Results: CA and LPS stimulate the production of TNF-~t, IL-11] and IL-6 by human PBMC. We observed a dose-dependent synthesis of these three cytokines in human PBMC. TNF-c¢, IL-IB and IL-6 in ng/ml as mean _+ standard error.

CA/ml 0.0 1. (2)

We constructed a single-chain Fv-fragment (sc-Fv) from the variable domains of a neutralizing antibody to human tumornecrosisfactor-alpha (TNF) because sc-Fv should have some advantages in vivo. The sc-Fv has a similar affinity (3,8x10-9M) as the Fab-fragment (2,7x10-9M) but bind the antigen with a eightfold lower avidity compared to the parental mAb (4,8x10-10). The parental mAb as well as its fragments can compete the binding of rhuTNF to both TNF-receptors. This was shown by competitive binding to the cell line HL-60 and to immobilized rTNF-receptors. In the nutralization assay on the cell line, L 929, the sc-Fv can neutralize TNF but in comparison to the mAb or its Fabfragment, the capacity was three-fold lower as expected from the molar ratio of the dissociation constants. The in rive-neutralizing capacity was tested in a mice receiving toxic dose of rhuTNF. The sc-Fv showed a 15-fold lower neutralizing capacity in comparison with the Fab-fragment. This could be explained in different manners: i) partial instability of scFv at 370C, ii) shorter half-life because sc-Fv but not Fab-fragments are eliminated via kidney, iii) elimination of TNF-immunocomplexes may be improved by opsonization mechanisms which requires some structures of the constant region (as in Fab or mAb). In summary, despite the encouraging in vitro experiments, the in vivo data suggest that sc-Fv are not the right strategy for neutralizing TNF in vivo.

TNF PRODUCTION IN ENDOTOXIN NON-RESPONDER MICE, EFFECT OF A GLUCOCORTICOID ANTAGONIST G. IAzgtr Jr, E. Duda, J. Ol~th, E. Husztik, G. IAzAr Glucocorticoids inhibit lipopolysaccharide (LPS)-induced TNF production and TNF can stimulate pituitary adrenocorticotropin secretion, resulting in the release of corticosterone. Earlier studies (1) reveales that the glucocorticoid antagonist RU 38486, nufepristone, sensitizes both endotoxin responder and nonresponder, C3H/HeJ, mice to endotoxin lethality. We have also shown that the glucocorticoid antagonist RU 38486 sensitizes endotoxin-tolerant (endotoxin-pretreated) and normal mice to the lethal effect of septic and endotoxin shock. On this basis, we set out to study the influence of RU 38486 on TNF production induced by endotoxin in endotoxin responder and non-responder mice. One and 2 hrs following LPS injection, RU 38486 significantly increased the TNF concentration in the serum, liver and spleen of treated animals as compared with the control and methylprednisolonetreated groups. In tissue cultures, RU 38486 induced the TNF synthesis of myeloid cells and increased the TNF production of genetically modified HeLa cells, which synthesize TNF constitutively. Normal and tumor cell cultures exhibited an increased sensitivity toward TNF in the presence of RU 38486. However, the same dose of LPS did not induce TNF production in the serum, liver and spleen of endotoxin non-responder mice and this was not influenced by concomitant RU 38486 treatment. These studies suggest that RU 38486 aggravates LPS lethality via factors other than TNF in endotoxin non-responder, C3H/HeJ mice. The cytokine response to a novel exnacellular mitogenic factor, MF, produced by the group A streptococci (GAS), and to the streptococcal pyrogenic exotoxins (Spe) A and B, was determined by in vitro stimulation of peripheral blood mononuclear cells (PBMC) obtained from healthy blood donors. The induction and kinetics of 14 different cytokines was studied at single-cell level using cytokine specific monoclonal antibodies and intracellular immunofluorescent staining. All three mitogens induced a massive IFNy and TNFI3 response in 10-16% of the PBMC after 48-96 hours of cell stimulation. In contrast, IL2 and TNFc~ containing cells was only detected in 1-3% of the PBMC. The induction ofIL3, IL4, IL5, and ILl0, was weak or completely absent. Thus, the response indicates activation of TH1 cells. However, ILlc~, ILll3, ILlra and IL8, were consistently found and gradually produced, peaking at 24 hems in approximately 5-8% of the PBMC. MF induced an equally extensive cytokine as well as proliferative inducing capacity, as did SpeA and SpeB, which suggests that also MF is a superantigen. A marked inter-individual variation could be noted between lymphocytes isolated from different individuals, both in the toxin induced proliferation and cytokine induction. This may be one explanation for the varying clinical severity noted in patients after infection with clonal GAS strains. Introduction -Tumour necrosis factor (TNF) is released by mononuclear cells in response to inflammation and sepsis. It has been implicated as a possible mediator in inflammatory bowel disease (IBD) but its pathogenic role remains uncertain. This study assessed the relationship between TNFct and disease activity by measuring plasma and faecal TNF concelnmtions in an experimcntaI model of IBD. Methodologo' -Colitis was induced in male Wistar rats by intmcolonic instillation of 30rag 2,4,6-trinitrobenzenesulphunic acid in 0.5mL 50% ethanol (n=12). Control animals received an isovohlmetric instillation of normal saline (n=12). Faecal pellets were collected before induction of colitis (Day 0) and throughout the experiment. After 8 days the colon was excised, ueighed and the colon macroscopic score (CMS) assessed Plasma TNF (pTNF) samples were assayed by bioassay and ELISA. Faecal samples were vortcxed in water and the supernatant removed for estimation of protein and faecal TNF (iTNF) content (ELISA). There is a significant increase in ITNF on day 4 -#p=0 02 alld the total fTNF measured during the study correlatcs with the CMS oll day 8 -p=0.04 There is no correlation between fTNF and pTNF. Conclusions -Therc is evidence of coloific TNF excretion by macroscopically normal animals and following induction of colitis there is increased faecal TNF which correlates with disease actMty. There is no significant elevation in bioactive or imnmnoreactive plasma TNF in colitie animals This data would suggest that faecal TNF is a marker of colonic inflamrnatiort and serial TNF assessment inay be useful as an indicator &severity in IBD. 

To study the effect of in-vivo hypoxia/reoxygenation on cytokine elaboration by routine peritoneal macrophages. Materials and Methods: Adult male CBA mice (20-25g) were primed intraperitoneaUy with either lipopolysaccharide (LPS) 10ug]anlmal or saline. Five days later, the animals were subjected to hypoxia (8%02) for 160, followed by 20 of normoxia. Harvested peritoneal macrophages were plated in-vitro, culture supernatants were collected at 20, 40, and 60 and assayed for tumor necrosis factor (TNF), prostaglandin E2 (PGE2) and nitric oxide (NO). Control animals underwent the same procedures, but were maintained in normoxia (21%O2 Disease 5, Berlin, Germany The precise mechanisms of reactivation of latent human cytomegalovirus (CMV) infection are still unknown. Apart from the permissive effect of diminished cellular immunity there is evidence of a cytokine mediated CMV reactivation as it is known for other systemic virus infections (e.g. HIV). We found that tumor necrosis factor-alpha (TNF) -in contrast to all other cytokines tested -can stimulate the activity of the CMV-IE-enhancer/promoter region in the human monocytic cell line, HL 60.

We wondered whether our in vitro results were actually reflected by the incidence of CMV reactivation in diseases which go together with high TNFalpha levels. Cellular immunodeficiency as well as at least temporarily increased TNF levels are known to occur in septic disease. We tested for the presence of CMV infection in peripheral blood mononuclear cells (PBMNC) by means of 24h centrifugation culture, immunocytological detection of different CMV antigens (APAAP-technique), and PCR technique (CMV-IE DNA). In striking contrast to healthy controls almost all examined septic patients were positive, irrespective of the method for CMV-ddtection applied, including the less sensitive immunocytology. Follow-up studies showed that CMV-antigen positive PBMNC ceased to be detectable in patients with good outcome once septicaemia had been overcome. To further back up our hypothesis, we looked for a coincidence of enhanced TNF-alpha serum levels and CMV antigenaemia (immunocytology) in some other diseases which are known to have either enhanced TNF serum levels or increased incidence of active CMV infection and found the majority of patients with B-cell chronic lymphocytic leukemia, liver cirrhosis, and common variable immunodeficiency to be positive for these two parameters. We now wondered whether, apart from CMV reactivation, TNF-alpha could also cause cellular immunodeficiency this way promoting CMV replication and spreading and eventually causing CMV disease. In summary, we were able to show that a diminished cellular immune response results from/n vitro or in vivo Methylprednisolone (MPS) is used to suppress both inflammatory and immune responses. IL-1 receptor antagonist (IL-lra) and TNF binding protein (TNFbp) are naturally occuring anti-cytokine proteins, possibly modulating inflammatory and immunological responses. To define MPS modulation of cytokine and anticytokine responses, we examined effects of MPS on IL-lra production and TNFbp generation as well as IL-I~ and TNFa production by human peripheral blood mononuclear cells (PBMC) stimulated with lipopolysaccharide (LPS). PBMC were purified from the blood of healthy volunteers, then incubated with LPS (10ng/ml) supplemented with different concentrations of MPS (0, 500pg/ml, 50mg/ml). After 24hrs incubation, the supernatants were collected for RIAs of secreted cytokines. The cell lysates obtained by 3 cycles of freeze-thaw, were also collected for R IAs of cell-associated cytokines. Both the supernatant for secreted cytokines and the cell lysates for the cell-associated cytokines were subjected to RIAs for IL-I~, TNFa, IL-1 ra, and TNFbp-I (TNFsRp55). IL-la, IL-1 ra, and TNFbp were produced mainly as cell-associated forms in response to LPS, whereas most TNFa was secreted into the supernatant of LPS-stimulated PBMC. MPS supplement resulted in reduction of IL-la, IL-lra, and TNFa production. IL-la production was reduced up to 5.4+1.6%(50mg/ml) by MPS in a dose dependent fashion, whereas about 50% decrease in TNFc, and IL-lra production was observed at both high and low dose MPS supplement. However, total TNFbp generation was not reduced significantly by MPS supplement. Furthermore, TNFbp secretion was increased 12 times more in the supernatant of LPS-stimulated PBMC with 500gg/ml of MPS. These data suggest that MPS could effectively block TNF activity by both reducing its production and up-regulating TNFbp generation and that MPS works as an anti-inflammatory drug as well as an immunosuppressant by modulating cytokine and anticytokine responses. In severe gram-negative sepsis, excess of proinflammatory cytokines is associated with increased morbidity and the mortality. We have found that immunocytes possess functional 13-adrenergic receptors (~-AR) which, when activated, down regulate the lipopolysaccharide (LPS) induced gene expression and subsequent secretion of TNF-~ in vitro. In the present study, we used a lethal endotoxic model in mice to test the hypothesis that 13-AR stimulation will down regulate the gene expression thus decreasing the circulatory levels of proinftammatory cytokines and improve the survivals in severe endotoxicosis. The animals (ISO group) were injected 3 times with !3-AR agohist isoproterenol (0.5 mg/kg i.m. and 0.5 mg/kg s.c.) 4, 2 and 0 hours prior to a lethal dose of LPS injection (25 mg/kg i.p.). Control group (CTL) received same volume of saline at the same times before the LPS insult. At the appropriate time points after LPS injection animals were sacrificed and venous blood was collected from the inferior vena cava. The plasma levels of TNF-cc and IL-loc were determined by ELISA. In the mortality study, each grou 

The data showed that isoproterenol significantly decreased the mortality and the circulatory levels of TNF-c~ and IL-lcc (Fig. 1) . These data suggest that 13-AR activation of immunocytes down regulates cytokine gene expression, decreasing circulatory eytokine levels thus reducing endotoxicosis mortality. These results provide a new pharmacological approach to reduce the excess of proinflammatory cytokines and mortality in sepsis. Introduction: Cytokines released from monocytic cells are thought to play an important role in the pathophysiology of sepsis. During the last few years several investigations have been performed aimed at modulating this mediator response. In the present investigation using a full blood model we have studied the effect of a standard immune globulin and an antitipopolysaccharide (LPS) immune globulin preparation and a monoclonal anti-LPS antibody in modulating the endotoxin induced cytokine response.

Methods: Citrated blood from healthy human donors was incubated at 37 °C in rotating polystyrene tubes. Samples were taken before addition of 1 x 103 ng/l E. col± endotoxin and after 2 hours. Plasma was assayed for the cytokines minor necrosis factor alpha (TNF-~) and interleukin-6 (IL-6) using ELISA methods. The eodotoxin concentration was assayed by LAL and end-point chromogenic substrate technique. The blood was preincubated with standard immune globuline (human IgG from pooled plasma, GammagardR, Baxter) or with an anti-LPS immune glubuline (human anti-LPS IgG obtained by screening of blood donors, Novo Nordisk) in the concentrations of 4mg/ml or 8 mg/ml or with the monoclonal lipid A IgM antibody HA-1A for ten minutes before the endotoxin was added.

Results: Two hours after endotoxin addition markedly elevated TNF-~ and IL-6 values were found.in the plasma. Cytokine values at 2 hours: TNF-ec 1689 pg/ml (mean) and 1L-6 1123 pg/ml (mean). Preincubalion with 8 mg/ml standard immune globulins reduced IL-6 values by 49 % (mean) while there were no effects on TNF-c~ values at 2 hours. A slight reduction in IL-6 values was also found with the lower concentration, 4mg/ml, but this was not statistically significant. Preincubation with anti-LPS immune globulins had no effect on IL-6 nor on TNF-cc values. Preincubation with the monoclottal lipid A IgM antibody HA-1A in variable concentrations had no effect on cytokine release in this model.

In this full blood in vitro model standard immune globuline reduced endotoxin induced 1L-6 release while TNF-cc values were unchanged.The anti-LPS immune glubuline preparation and the monoclonal lipid A IgM antibody HA-1A had no effect on the IL-6 or TNF-ct release. The effect of anti-TNF treatment in severe haemorrhagic/traumatic shock was investigated in a conscious rabbit model. The jugular vein (for administration of substances and blood withdrawal) and right femoral artery (for measurement of BP) were cannulated and an aortic thermistor probe inserted into the left; femoral artery for the measurement of cardiac output (CO). The following day, upto 40% of the estimated blood volume was withdrawn over 80-40 rain, to reduce BP to 40-50mmttg and CO by 2/8 mad withheld for 60 rain. Shed blood (60%) was then reinfused followed by lactate solution to give a total potential dreulat%ag fluid volume of 120%. Zymoeen activated plasma (0,36gg/ml) was given 10 rain prior to haemorrhage and during blood r~n%eion, All ~,~ir, ak were killed at 72h and o~gane removed for histology. Rabbita received either monoclonal antirabbit TNF (R6, rat IgG1, 8mgkg i.e. n=7) or eMine (nffi6), 60 rain prior to haemorrhage.

The cardiovascular and biochemical changes during the haemorrhage and hypotensive phase (is. haemorrhagic stimulus) were similar in both groups. Upon reinfusion, haematccrit and white cell count remained sigutfia~utly (P<0.05) higher in saline compared to R6-treated animals, indicating a relative haemoconcentration (i,e. reduced circulating volume) in saline animals. The lettkocytosis persisted in the saline animals upto 72h. Plasma glucose, lactate and asparteto .m~-o transferase all rose similarly in both gorups but plasma ¢rea~!~+-e levels were markedly (P<0.05) higher in ealine ve R6 animals at 48 and 72h indicative of kidney damage, Using a macro and micro. pathological scoring system, major organ damage was seen in the lung, liver and kidney which was significantly (P<0.05) attenuated in R6 treated animals. In the hmg and liver, this was primarily due to a reduced bleeding and PMN infiltrate and to an additional maintenance of tubular integri~:y in the kidney. Hence, in tlais model anti-TNF treatment markedly reduced the biochemical and histological indices of severe ,, ~gan damage during liaemorrhage and rAnerfn~ion. Thermal injuries and systemic bacterial sepsis produce a wasting diathesis with anorexia and marked loss of lean tissue and body fat. Endogenously produced cytokines, including interleukin-1 (IL-1 ), are thought to mediate many beneficial as well as pathologic host changes that occur during burn injury and infection. To evaluate whether bluckade of IL-1 in vivo would affect survival and attenuate the anorexia and cachex[a associated with a thermal injury and chronic fulminant gram negative infection, 24 Sprague-Dawley rats were studied. Anesthetized rats were divided in 3 groups, implanted with Alzet R minipumps and randomized to receive the following treatment for 14 days:i: 25 gg/kgBW/min of IL-lcc; II: 25 ~tg/kgBW/min of IL-1 receptor antagonist (IL-Ira); III: buffered vehicle. The animals were subjected to a 30% BSA, full thickness scald burn and the wounds then infected with 10 + cfu/mL Pseudomonas aeruginosa. Eight additional rats were anesthetized, but were not burned, and served as healthy controls. . IL-hx reduced food intake transiently but was otherwise without effect However, IL-lra significantly attenuated weight loss over the first 8 days and improved food intake between days 5 and 8. We conclude that blockade of IL-1 after a thermal injury with superimposed infection does not adversly affect the progression of the disease, but does attenuate the anorexia and weight loss associated with this model. These data indicate that IL-I receptor blockade may offer a means to minimize the morbidity associated with catabolic illness. Tumor necrosis factor (TNF) and its downstream induced mediators have pivotal roles in the pathogenesis of sepsis and septic shock. The suppressive effect of pentoxyfillin (PTX) on TNF production may be of clinical importance. When peripheral white blood ceils were stimulated with either LPS or Staphylococcus aureus, pentoxyfillin inhibited TNF production. In contrast, no inhibitory effect was observed on the induction of IL-6. PTX inhibited not only the TNF production of monocytes but also the TNF secretion of both granulocytes, and unseparated whole blood. The FACS analysis revealed that the expression of CD14 antigen on monocytes was not influenced by pentoxyfillin. The in vitro TNF and IL-6 producing capacity in septic patients were higher than in the control group, but decreased on subsequent days especially in fatal cases. Administration of PTX (200 mg/day) in septic patients resulted in TNF and IL-6 productions similar to those found in control donors. It also subsequently led to an improvement of the clinical status. A further improvement in APACHE score could be achieved by administration of Pentaglobin ° (Biotest)( 5ml/kg/day ). The prevention of LPSinduced in vitro TNF and IL-6 production by Pentaglobin ° could be demonstrated in in vitro experiments involving the use of whole blood rather than purified lymphocytes, very probably because of the presence of LPS binding protein in the plasma. The level of soluble ICAM-1 in sera of septic patients was significantly higher (800-1200 ng/ml) than in normal individuals (50-150 ng/ml), but it decreased following the PTX and Pentaglobin ° therapy. It is presumed that PTX and Pentaglobin ° can antagonize eytokine production at different levels, resulting synergistic action being beneficial in the treatment of sepsis.

Albert Szent-Gy6rgyi Medical University, Institute of Microbiology, Szeged, Hungary H-6720 Ddm t. 10.

E~I~IM~/TII, SEPTIC SH0~ '. TNF alfa/PGE, and Somatostatln Lands JI., Alvarez J., Gram M., Llanos K., Alcalde J., Sanchez JA., Balibr~d JL. Taurine, a semi-essential sulphur-containing t:~-amian acid, promotes neutrophil phagocytic activit3' against yeasts and enhances intracellular killing of E.coli. Paradoxically it protects against hypochlorous acidmediated biomembmne oxidatio~ and abrogates the pyrogenic effects of intracerebral endotoxin. It is therefore possible that taurine mediates some of these effects by modulating the cytokine cascade.

Aim To assess the cytekine responses in a rat model of intraperitoneal sepsis, pretreated with supplenmntary taurine, by measuring TNF and IL-6 concentrations. Methodology Female Wistar rats (wt 160-200g) (n-6 per group) were prefed with 2% tanrine, 3.33% glyciuc ( in tap water) or tap water (TW) for 15 days prior to caecal ligation aud puncture (CLP). Normothermia was maintained intraoperatively and the animals received 0.9% saline subcutaneously (3 mls/100g) post-procedure. Animals were venosected by direct cardiac puncture at 6 or 18 hours post-operation and the blood placed in endotoxin-free tubes. Centrifilgation and storage of plasma at -70°C was completed within 2 hours. TNF and IL-6 were measured using WEHI and B9 bioassays respectively. Results (Means ± SEM) Bioactivc TNF concentrations were not significantly different between CLP or sham groups. However plasma IL-6 estimation revealed a statistically significant reduction at 6 hours in tanrine-treated animals compared to glycino and TW controls ( Objective: To evaluate the effects of allogeneic blood transfusion, thermal injury and bacterial garage on interteukin 4 (IL-4), tumor necrosis factor alpha (TNF) production and host mortality and to study if the administration of thymopentth (THY) could affect these events. Materials and Methods: BALB/c mice were transfused with allogeneic blood (from C3H/HeJ mice) and treated daily with THY (I mg/kg)(T+THY) . Control animals were transfused but not treated (T). After 5 days systemic blood was harvested to determine the circulating levels of IL4 and TNF. A second group was treated with THY for 5 days and then received a gavage wide lxl09 Escherichia coli and a 20% burn injury (BG+THY). One hour post-bum, blood was collected to measure cytokins. Control animals received the same treatment but THY (BG). A Third group was transfused, treated with THY for 5 days and then received gavage and burn (TBG+THY). One hour post-burn, blood was collected to measure cytokins. Control animals received the same treatment but THY (TBG). All treated and control groups had 8 mice/group. In separate groups (n=20/group), receiving the same treatments, mortality was observed for 10 days post-burn or transfusion. The production and release of oxygen radicals by phagocytic leukomytee is one of the most relevant and important functions of these substances in biology.

When neutrophilic granulocytes or certain macrophages phagocytize they produce super oxide and hydrogen peroxide and form secondary products of these activated species. All of these substances are released in to the phagosome. It has been appreciated since 1977 that the amount of oxygen consumed by phagocytes in producing free radicals is prodigious. What has not been totally appreciated is the significant contribution this 02 consumption contributes to total body oxygen consumption.

Our understanding of increased oxygen consumption following sepsis is compounded by a paradox of increased cardiac index and a narrow AV02 difference.

It was McLean in a study reported in 1967 that showed an increase in cardiac index and a narrowing of the AVO z difference in humans following sepsis. Other studies documented similar changes in humans and it has recently been speculated that increasing O~ delivery might have a favorable impact on outcome. At least three different explanations have been put forward as to why there are possible alterations in energy metabolism during sepsis.

These include decreases in oxygen supply, peripheral shunts and direct action o~ cellular mediators on O~ delivery, A provocative review by Hodgkiss and Nark in 1992 shewed that in experimental animals sepsis did not cause any evidence of bioenergatic failure in muscle I liver, heart or brain tissue. They further showed that the increase in serum lactate is mot necessarily due to cellular hypoxia. They conclude that cellular hypoxia and defects in energy producing metabolites are not the cause of metabolic abnormalities in sepsis. We set out on a series of experiments to try to explain the paradox in cellular metabolism and whether or not white cells might contribute to total body oxygen consumption during sepsis.

The first set of experiments involved growing rat cardiac myocytes on tissue culture.

Pyruvate was added to the medium as substrata and physiologic levels of peroxide were also added simultaneously, carbon 14 labeled CO~ production was measured as was nucleotides and degradation products from the cell.

The peroxide induced a significant inhibition of pyruvate dehydrogenase.

In addition, there was a loss of cellular ATP and a corresponding rise in the release of inosine and adenine from the cell.

We concluded from this first sst of experiments that physiologic levels of peroxide are a potent inhibitor of pyruvate dehydrogenase in isolated cardiac mitochendria as well as the cultured myocyte.

We further showed that pyruvate dehydrogenase inhibition blocks the aerobic oxidation of glucose and leads te adenine nucleotide metabolism with adenosine and inosine release from the cell. other enzymes within the TCA cycle did not appear to be specifically blocked by peroxide.

This would explain the increase in lactate and the ability of the cell to continue to make high energy phosphate by entry of metabolites in to other entry points of the TCA cycle,

The second set of experimBnts was designed to estimate the magnitude of whole body reactive oxygen generation via NADPH oxidase following granuloeyte activation invivo.

Using a guinea pig model baseline oxygen consumption was determined. ~iaphenyl iodinium (DPI) was used as a specific inhibitor of granulocyte NADPH oxidase.

Animals were divided into two groups; those that received DPI and those that served as control, All animals were then injected with Phorbcl Myristate Acetate (pMA) intravenously. pMA is a potent stimulus for granulocyte activation and subsequent reactive oxygen generation.

Whole body oxygen consumption measurements were then repeated after the P~LA injection.

In addition, arterial blood gas amalysis was performed, circulating neutrophils were collected and assayed for their ability to generate hydrogen peroxide and the r~ght lung was removed for wet and dry weight measurement.

Total body oxygen consumption increased by 25 percent after PMA injection.

Arterial oxygen tension fell significantly in the control animal.

Neutrophil hydrogen peroxide generation rate doubled and lung water in the untreated animals increased by 25 percent. In this animal model we could show that granulocyte activation substantially increases whole body oxygen consumption to approximately 120 percent of control values.

The NADPH oxidase inhibitor, DPI, decreases granulocyte hydrogen peroxide generation invivo and prevents the pulmonary injury induced by PMA.

using this animal modal it is possible to transpose this to the human septic state.

If we assume there are 5,000 neutrophils per cubic millimeter of blood, a 70kg man would have ~1.75XI0 u circulating neutrophils. When activated ixl0 ~ neutrophils generate one nanemole of hydrogen peroxide per minute.

T~erefore, the circulating neutrophil pool could generate -1.75xi0 nanomoles of hydrogen peroxide per minute.

This corresponds to an oxygen consumption of -4.2ml/OJmin by the circulating neutrophil pool alone. This does not take into account the production of oxygen by macrophages or the amount of oxygen consumed to produce nitric oxide. Manipulation of oxygen delivery in human sepsis may or may not be beneficial.

Irshad H. Chaudry, Ping Wang, and Alfred Ayala

Life threatening blood loss, due to soft tissue and bony injuries, is a frequent complication in trauma victims. Although numerous organs and cells are adversely affected by hemorrhage, the focus here will be to discuss whether or not there are any differential alterations in splenic and hepatic immune functions. In particular, M~ antigen presentation (AP) and associated processes will be described since one of the important functions of the M6 is to activate resting T-cell lymphocytes by processing and presenting foreign antigens in a MHC Class II-restricted fashion. Studies have shown that splenic M6 and Kupffer cell (KC) AP was significantly depressed following hemorrhagic shock and remained so for at least 96 h alter resuscitation. The presentation of antigenic fragments associated with MHC Class II antigen itself was not decreased following hemorrhage, but the catabolism of antigens in antigen-presenting cells was decreased. This is likely due to a significant loss of cellular ATP.

Although M6 AP was depressed in splenic, as well as KC, only KC showed an enhanced capacity to produce inflammatory cytokines, IL-1, IL-6, and TNF. This difference in response between KC and splenic Md may be due to differences in the microenvironment in which these cell populations are found, as well as potential differences in the effects that hemorrhage has on the environment from which these cells are obtained. Splenic M6 from hemorrhage-resuscitated mice exhibited a significantly reduced cytotoxic response, whereas KC showed an enhanced cytotuxic capacity. The increased KC cytotoxicity is probably mediated through the increased expression of cell-associated TNF and the release of reactive nitrogen. The enhanced production of reactive nitrogen may play an important role in the clearance of microbes translocated from the gut following hemorrhage. However, excessive release of reactive nitrogen by KC may have deleterious effects on the adjacent hepatncytes.

Such an effect could, in part, explain the reduction in active hepatocellular function, which is seen following hemorrhage-resuscitation. Thus, hemorrhage induces differential effects on splenic and KC M6 populations; it decreases splenic M6 but increases KC capacity to mount a cytotoxic response. The depressed splenic Md and KC AP was, however, significantly improved by posttreatment of animals with ATP-MgCIa, chloroquine, diltiazem or interferon-'/. These agents also decreased the susceptibility to sepsis following hemorrhage. Thus, the use of ATP-MgCI~, dfltiazem, chloroquine or interferon-7 offers new therapeutic modalities in the treatment of immunosuppression and for decreasing the susceptibility to sepsis, which is observed following severe blood loss. The production of macrophages from bone marrow precursor cells is a dynamic and highly regulated process. The differentiation of myeloid lineage-restricted progenitor cells is initially controlled by interleukin-3, which drives the progenitors along a maturation pathway that leads to their response to specific lineagerestricted colony-stimulating factors(CSFs)

We have previously hypothesized that thermal injury can initiate a self perpetuating cycle of production of defective immune cells: Thermal injury can initially affect circulating immune cells. Then these cells, by unbalanced production of colony stimulating factors and other mediators, can cause a derangement of hematopoiesis resulting in an abnormal production profile of TNF, IL-I, and IL-6 and cytotoxic activity of bone marrow derived cells.

The results of our studies so far have supported parts of this hypothesis.

IL-3 and IL-3+M-CSF treatments increased the production of TNF in burn 4 day macrophages and progenitor cells compared to normal cells. IL-3+M-CSF+IL-10 increased the production of IL-6 by burn progenitor cells compared to normal macrophagee.

Preliminary results for the ratios of TNF/~ actin mRNA indicate that IL-3+M-CSF increased the mRNA for TNF in the 4-day macrophages derived from burned animals compared to macrophages derived from normal animals.

Ratios of IL-6/~ actin mRNA indicated that IL-3, IL-3+M-CSF, and M-CSF increased the IL-6 ~RNA in progenitor cells from burn animals compared to progenitor cells from normal animals.

These preliminary results support our hypothesis that bone marrow derived macrophages and progenitor cells from burned animals act differently compared to cells from unburned animals regarding the production of inflanunatory cytokines. More specifically, the results suggest that the different cell types are regulated in different ways by cytokines, and cells from burned animals are regulated differently than cells from unburned animals. It i,~ to ~ssume that the translneatinn is due to a Ixx~H,v functioning ~astroiutestlnal surface pro|ection system, which leads to an unacceptably high load ¢ln the Imrsunc dcfcncc system, partlcul~ly the local system in the gastrointestinal wall, leading tu exh~ustiou Of Ihi~ System. Tlae g&~lrointesliual I]'ac( can be regal'deal a.s a t~2servoJf of appr 200 m ~ separating 10 I~ eucayotic ceils of the hosl ttom l0 ~4 bacterial culls. In the human body the the myriad of cndngan¢nts micrix)rganistns, exogenous itl,~ttdittg micro~ganisms altd taxies m-'e sepia"areal li'otll Lhe rest of the body by a simple epithclhd layer. The barrier function is heavily tlcpcnding tm =~ proleclive layer cmtsisting ie. surl~:Emts, nluCltS, probiotic bacteria, and ,~ub~tances with strong antioxid~lt, antipfotease al~d other i~rotective eft'ecru, The surfaet~ml htycr consists at tile besl if1 I(.) bimolecular layers, fl is thus very Iltil~, The epilhelial layer is renewed every 24 hours , and tile tanuwer of the mucus is ~so fa,~t, All attDphy of die epithelial layei induced by g&sttointestinal st~'vation does also include the mu~usprndocing cells, Gnhlet cells, which leaduhg lu slmnage aad luwer quality of GI mucus, altered viscoelas,Jc properties and reduced protection, ]~'obioiic bactcrla keeps the PPM's low, produce nutrients for the intestinal mllcosa, eliminates tuxia~ and stimulates the local immune system, The Gl surface is conslanlly under hc wv wearing by ingested bacteria and their enzymes, ingested substaucas, chefftie,'ds at~d toxins. 'll~e N~tective flora is reduCed in disease, by stress and by cm'clcss antibiotic therapy. Wc have made attempts tn recondition the I,,.astroiniesli!l~d. tnucoSa by rcstlpply of surfactants or sttrfactanfliko glyc~lipids, gels with pseudomuein [until.oil, and probiotl¢ laetobaeilli. By supply of glycolipids we have bt:cn able m prevent and heal intlammatory and tlleerallve injuries (and Io p)'cveut microbe transhteatkln) in uppe~' and lower (.31 u'acl, This c~al be fu~'ther augmentented by shnultaueos supply el' a p.~uedomucin,likc gel like pectin. ()at contain~ one hundred times more ot membrane lipid$ th~a =nay other food, much of fiber, ~spceially watcrsnlnhlc betaglucaal~, at~d has an close to ideal amino acid cornpo~ition, <')at, termented by species-specific lactobacilli, has a strOl~ ability to prevent local inflammatory and ulcerative chunges,transh)cadnn and sepsis in experimeut~d attimals mid to pl~ven~ revert MOF in ba,nan.~.lt seems us, that it the mucos~lnucus/probiotic bacterial filnclinns cat'= 1%' restated by mucosa reconditioning, even in severely sick iudividuals) the hx:~d iuuuuue ~y,~teme will be capable tu pr~)tec! the Ix~y ag~lls~ sepsis ~nd MOF. The impaired immune response associated with multiple system organ failure (MSOF) has been most widely studied following surgery for inflammatory disease and trauma. The majority of late deaths following trauma are due to infection and MSOF. Not all patients with multiple organ failure have concomitant infection present, yet most have been septic at some time in their hospital course. Their generalized inflammatory condition often persists whether or not organisms have been recovered. Immune failure probably precedes MSOF with or without overt infection and indeed may perpetuate such widespread sepsis. Macrophage tumor necrosis factor-alpha (TNF) production is thought to represent an important pathogenic mechanism by which this is mediated. Cecal ligation and puncture (CLP) is a clinically relevant murine model of intra-abdominal infection and sepsis. Serum TNF and endotoxin levels, and peritoneal macrophage TNF production are only slightly elevated in this model even in endotoxin sensitive animals. Mortality in this model, therefore, does not appear to be attributed to overwhelming endotoxemia and/or TNF production. It has therefore been postulated that TNF and other cytokines may act in a paracrine fashion to Cause local injury. TNF messenger RNA (mRNA) expression was therefore measured and found to be increased in the lung and liver following CLP. A different pattern of expression was seen after endotoxin administration, characterized by a more rapid rise and fall, and enhanced TNF mRNA expression in the spleen as well. Route of spread of infection as well as the type of stimulus may determine the organ specific cytokine response. These data support the concept of locally produced TNF as a contributing factor in tissue damage and multiple organ failure during sepsis. The estimation of histamine's role in SIRS changed over the years based on increased knowledge in shock pathogenesis, evidence later found to be faulty, and the discovery of other, more fashionable mediators. Over the decades, the prevailing view has been that histamine is a noxious mediator contributing to the fatal outcome of shock.

The analysis of experimental data on exogenously administered, endogenously released and/or newly formed histamine in septic/endotoxic shock suggests that histamine contributes to the early cardiovascular changes via HI-receptor vasoconstriction thus excerbating (directly and indirectly) the effects of catecholamines. In the later phase of septic shock (low out-put, high resistance states), however, it can be assumed that histamine exerts strong vasodilator and positive inotropic effects via H2-receptors. These effects are considered beneficial in counteracting or attenuating the effects of sympathetic responses of catecholamines and other mediators. Thus, a blockade of the early reaction with Hi-antagonists and a stimulation of the H2-receptors by H2-agonists are worthwhile therapeutic approaches. Shock produc~s ~IJ alterations, an encrgy crisis and eventual c~ll damage, However, shock in itse|f do~ not lead freqoenfly to r~mote organ damage. In animal ~xpcrhnents and clinical expariez:ces in Korea and Vietnam, shock I~r s¢ was not associated with lung damage, renal failure or ARDS. The same is true with G.L bleeding. However, when shock is due to trauma or associated with tlssus injmy, then organ malfunction m~d damage is frequent. Shock is bclleved to increase the frequency of infection, but clinical evidcnc~ for this is scant. Hemonhsgio shock akme is an unusual o~orrenee dinica[ly. Immune dysfolletion dons occur witl~ experhnental hypovolemic shock and in pailcnts with hypotcnsion or after receiving blood ~ansfosions. Tbc most significant factor affecting reoorrcnce of resented ca,lorectal liver mclastasis was hypotensivc episodes during operation. Blood transfosions depreSs immune function, improve transplant ~raft survival and increase ttll'rmr rec~R'Tcnos ill co[eli and head and nc~k cancers. In experimental hemorrhagic shock, we found decreased response of [ymphocytes to mitogens, el:clad mixed lymphocyte reactilln and I[,-2 decreased. Othe~.x hsvc found decreased macrophage and T and B ¢¢lt function, deci~ased re,ease of/L~I, increased POE v depressed macrophagc antigen presanlatiota, increased IEN y, decreased IL-2, 3 and 5, and shared mscrophags funclion with loss nr F and C 3B r~ccpto~s. This is inhibited by chlornqoine, ibuprofen, dihiazem, and ATP. Shock may activate ncutrophils contributing to ilappitlg in cspillarics, no refluw, increased adhesion, cytotoxicity and organ damage. This may be related to decreased clearance of bacteria after shock. T~lls shock, hypotcnsion and decreased microcirculatory blued flow initiate or set the stage for immune dysfmtction. They, along with tissue injury trigger inflammatory cascades. This also contributes to immunologic dysfonctitm, which is associated with increased Lufoction. Thus altered bh,od flow and mediator cascades are bolh involved, The question now is can wc rapidly improve microcirculatory blood flow, and dg'masa Inflsrmnatory Fosponss, which is also necessary for survival The effe~ of ~nterferon ~amma @n wour~ healing was ewaluated ~n a routine mo~el. ~mma ~n~erferon we8 g~ven in doses of ~37.5 to ~2,500 U~%$ synchronous with ereatio~ of a pa~splnal woumd then daily for IS ~lays. ~mteET~on ~amma impaired tensile strength a% all ~oses relative to control.

M0zphology suggested ~reas~ ¢oll~gen deposition and r~du~-~ neovag~ulaEit¥ ~n t~eat~4 animals, Interferon ~amma was a2so ~valuated in a murine mo~el of cecal llgatlon and punneure. One hundred tO 22,S00 uni~ vere ~iv.n following ~nju~ an~ than daily. Interferon qamma was associated with inco:ea~ed mortalit~ and earlier death a~/a~n ~n a dose dependant manner (P< 0,05).

Afmlnistra~ion The proinfiammatory cytokines tumor necrosis factor-c~ (TNF-c0, interleukin-16 (IL-II3), interleukin-6 (IL-6), and interleukin-8 (IL-8) have been implicated in the pathogenesis of the multiple organ dysfunction syndrome (MODS) following shock and trauma. These mediators which are released in high concentrations by activated macrophages, endothelial cells, and connective tissue cells cause a systemic inflammatory response syndrome (SIRS), thus leading to a shock-like state and tissue destruction. Recently, a number of proteins have been characterized in in vitro and in vivo studies demonstrating potent anti-inflammatory properties. These latter cytokines include interleukin-4 (IL-4), interleukin-10 (IL=10), transforming growth factor-6 (TGF-J3), and the newly sequenced interleukin-13 (IL-I3). They are considered antiinflammatory, since in vitro studies using purified macrophage cultures or whole blood assays revealed an inhibitory effect of these antMnfiammatory cytokines on ~he synthesis and release of TNF-cL and IL-U3. In addition, they reduced the severity of disease and reduce plasma levels of TNF-c~ and IL-II3 when administered to animals with infection or inflammation. Furthermore, naturally occurring substances have been described that specifically neutralize the bioactivity of IL-1B and TNF-a. The IL-1 receptor antagonist (IL-I ra) is related structurally to IL-I and binds to the same cell-surface receptors, but does not stimulate the cell. In animal models orE. coB-induced shock, inflammatory bowel disease, and peritonitis, injection of IL-lra significantly reduced the severity of disease. The cytotoxicity of TNF-c~ can be inhibited through two types of soluble TNF receptors (sTNFR) type I (p55) and type II (p75). These sTNFRs are proteolytic cleavage products of the cell-bound TNFreceptors. When given to baboons to combat E. coil-induced shock, soluble receptm's to the TNF type [ receptor decreased the severi~ of the shock-like state. It has been well accepted that shock and severe injury result in an increased release of proinflammatory cytokines with a high incidence of SIRS and MODS. Our studies investigating plasma levels of anti-inflammatory mediators (IL-4, IL-lra, sTNFRs) in patients after severe injury further suggest that shock and injury lead to an activation of anti-inflammatory processes with altered plasma levels of these mediators. However, while plasma levels of IL-4 and 1L-lra were significantly increased in trauma patients compared to healthy volunteers, sTNFRs did not differ from control samples. Thus, shock and severe injury result in different activation patterns of anti-inflammatory processes.

M,L. Rodrick, Boston, b~USA Following severe thermal injury significant suppression of the immune response has been demonstrsfed.

These changes are associated with increased susceptibility not only to co,on human pathogens, but to organisms not normally pathogenic in the uncompromlsed host.

Early observations included prolonged graft rejection and skin test anergy in patients following burn injury. Work from our laboratory and that of numerous others has shown that the~e is a profound £mmunoeuppresslon following severe thermal injury.

I~mune deficiency has been identified in these patients hyz II lack cf T nell responses to mitogens, 2) early decreases in total T and helper T cells in the peripheral bloo~ 3} lack of response to Tetanus toxoid i~uniza~io~ in spite of increased non-speoifi~ i~k~unoglobulin praduction and 4) severe and prolongei deficiency of Interleukin-2 (IL-2) prcductlon by peripheral blood mononuclear cells (PBMO). This IL-2 deficiency could be an underlyin~ cause of all the afo~ementloned immune defects by failing to activate both helper and suppressor T cell functions. Mo~a recent work has focused on the molecular mechanisms of this IL-2 deficiency end shown that the defect lies post-transcrlptionally since mRNA for IL-2 is also depressed following thermal injury. We have also investigated the potential of a defect in Ii-2 receptor status on peripheral blood mononuclear uolls from patients fQllowing burn injury and in a murine model. In both oases no decrease in IL-2R was found either by phenotypic markers or by funational assays. Another hallmark of thermal injury is hyperao~ivatlon of proinflammatory ~ytok~ne secretion, which may play a significant role in multiple organ failure following burn injury. Therapy O~ major burn injury may reasonably be aimed, therefore, at up-regulating T cell ~unction and down-regulating hyperactive secretion Of proieflaam~ntory cytokines. The majority of patients dying of bums succumb to sepsis which coincides with dysfunction in the specific and non-specific host defences. Generally, concepts relating to the mechanism(s) of specific immune failure in thermal trauma imply that inhibition of T (CD4+) cell responses is imposed by the injury or infection-related factors. However, the same factors elicit strong biological reactions within the lymphoid compartment. Recent evidence indicates that systemic activation is inherent in the burn patient. To establish the relation of molecular and cellular events to the development of post-bum anergy we examined the state of and the capacity for T cell activation with regard to: 1) occurrence of activation-induced cell death (AICD), 2) activation-related phenotypic and functional_ diversity in the pool of peripheral CD4+ T cells. Initially, (up to 7 days post-bum), peripheral blood lympliocytes from severely injured (> 35% total body surface area) patients exhibited high levels of constitutive expression of surface receptor for ]L 2 (CD 25) and spontaneous blastogenesis. The presence of activation-related T cellproducts in bum plasma was also apparent. Subsequent impairment of the T cell receptor (TCR)-regulated T cell responses in vitro was accompanied by significantly increased DNA fragmentation that is associated with cell death by the mode of apoptosis. Using molecular markers we established that flesh peripheral blood ceils from immunosuppressed patients also contain large numbers of apoptotic cells. Fluctuations in the number of viable (PI-) peripheral blood lymphocytes involved primarily CD4+/CD45RO+ (memory) subset of T ceils. The above observations suggest that thermal trauma-associated T cell anergy develops through AICD, a phenomenon commonly associated with the tolerogenic activity of bacterial superantigens. Persistence of staphylococcal infections in the burn patient may support this assumption.

RESPONSE FOLLOWING TRAUMA JANE SHELBY, Ph.D. The immune system is integrated with other physiologic systems, and is exquisitely sensitive to changes in nervous and endocrine systems changes following traumatic stress challenge. The immune, nervous and endocrine systems interact via both direct and indirect pathways which utilize neuro and endocrine hormones, neurotransmitters, neurepeptides and immune cell products. It is now known that the immune system may be affected by all of the neuroendocrine products produced during a stress response, with evidence for innervation of iymphoid organs, lymphoid cell receptors for neuroendocdne products, and leukocyte production of chemicals which are virtually identical to certain neuroendocdne peptides (ACTH, endorphins). Trauma induced alterations in the equilibrium of various neuropeptides and neuroendocdne hormones have a significant impact on immune response potential, affecting control of proliferation, differentiation and function of immune cells. For example, the neurohormone melatonin is thought to be a natural antagonist to counteract glucocorticeid associated immunosuppression resulting from stressful challenges, such as surgery and trauma, Plasma melatonin levels are known to be significantly reduced in burn patients. The administration of exogenous me[atonin improved cellular immune response following burn injury in an animal model. Melatonin was also shown to have in vivo cytokine regulatory activity, increasing the potential for IL-2 secretion and downregulating excessive IL-6 and IFN~ in burn injured, stress susceptible mice.

The regulatory interactions between the immune, nervous and endocrine systems provide mechanistic pathways for trauma associated immune dysfunction. Increased knowledge of these interactions will enhance the potential for the design of novei clinical interventions to improve immune response and decrease the risk for infection in trauma and surgical patients. . Animals receiving E were given a single dose daily of either 2.4 g/kg of E in a 20% solution by garage (GE), or 1.25 g/kg of sterile IVE in saline. Four hours following the last dose, bum animals were subjected to a 30% body surface area bum injury to their dorsum. Twentyfour hours following injury, the animals were sacrificed and spleen cells were harvested for assessment of lymphocyte function. Splenocytes were prepared by mincing the spleen, followed by incubation on glass petri dishes to remove adherent macrophages. Non-adherent cells were then tested for proliferative response to T-cell mitogen Concanavalin A (Con A) and B-cell mitogen lipopolysaccharide (LPS). Data were analyzed by ANOVA. RESULTS: Chronic alcohol exposure and burn injury independently inhibit lymphocyte response to Con A but not to LPS. The combination of E plus bum injury, however, pmfouedly decreases this response to both Con A and LPS as outlined in the This data clearly identifies the synergistic impairment of immune function produced by ethanol and bum injury. It is furthermore apparent that Ibis effect is gut mediated and that gastrointestinal exposure to alcohol is necessary to produce this effect. Further studies will work to identify cellular and subcellular mechanisms to explain this effect. In experimental animal studies and investigations on human volunteers endotoxin infusion is mgulary accompanied by the release of the cytokine tumor necrosis factor a (TNF-~) determined by ELISA technique. In patients with menigococcal sepsis also elevated TNF-a values have been found using a functional assay.

We have studied the role of TNF-et in surgical ICU patients with sepsis. Using functional technique, we were not able to detect TNF-~ activities in the patient plasmas. When this cytokine, however, was determined by immunochemicaI technique (EL1SA) elevated TNF-e~ values where frequently oberserved. In order to further elucidate these observations, we studied shedding of TNF receptors in the patients. In these studies, we noticed that shedding of TNF receptors oecured regulary in the patients. At the time of diagnosis, soluble TNF receptor p55 and p75 were both 2-3 fold higher than values found in plasma samples obtained prior to die diagnosis of sepsis. We also observed that the sepsis patients revealed higher maximum values of p55 and p75 during the ICU stay compared to values found in surgical ICU patients without sepsis.

These observations indicate that soluble TNF receptors are available in sufficient amounts to bind TNF-ot which is released in surgical patients developing sepsis. This mechanism may explain why functional TNF-c~ was not detected in the patients.

Institute for Surgical Research, Rikshospitalet, The National Hospital, University of Oslo, 0027 Oslo, Norway.

Decker, D., Sch6ndorf, M., Bidlingrnaier, F., Hirner, A., yon Rfcker, A.

The advantage oflaparoscopic cholecystectomy over conventional open surgical approaches in the treatment of symptomatic cholelithiasis has been shown convincingly by clinical studies. In order to facilitate comparisons of different surgical approaches, we evaluated the cell biological characteristics of tissue trauma by measuring changes in various cell surface markers on leukocytes and eytokines in plasma as a possible means to assess tissue trauma in choleeystectomy.

Patients recruited into our study had experienced at least one typical bifiary colic, had ultrasound-proven cholelithiasis (stages 0-II according to Me Sherry), were 35-55 years old, and presented for elective choleeysteetomy. Patients could choose between laparoscopic and conventional eholeeystectomy after being informed about the advantages and disadvantages of each procedure. Cell surface markers on leukoeytes were determined using whole blood techniques with the help of commercially available fluorescent monocloml antibodies and flow cytometry. Shed cell surface markers in plasma and cytoldnes were measured with the help of sandwich-ELISA kits. Blood samples were drawn 24 h before surgery, immediately before incision (after anaesthesia), 2 h and 24 h after incision.

Seventeen cell surface markers were examined on different cell populations and cellular subsets in 18 laparoscopic and 15 open-surgery patients. Three soluble cell surface markers and six cytokines were monitored. By statistical analyses (multivariate regression analysis, Student's t test, WilcoxomMann-Whituey's rank sum test) the six markers/cytekines that best distinguished open surgical from laparoscopic procedurea were determined. These were 1. the interleuldn-2 receptor and im soluble form (CD25/sCD25); 2. the activation antigen Fd-1 and its soluble form (CD30/sCD30), a member of the nerve-growth-factor receptor family; 3. the CD45RO epitope which characterizes T memory ceils; 4. the trausferrin receptor CD71; 5. the soluble adhesion molecule ICAM-1; and 6. the cytokines interieukin-6 and interleuldn-8. On the basis of these results, a tissue trauma activation (TTA) index was calculated by combining the marker/cytoldne measurements by simple multiplication.

Anaesthesia and pre-ineision maneuvers did not significantly change cell marker or cytokine levels in either surgical approach as compared to 24 h before surgery. 2 h after incision the TrA index in open cholecystectomy showed a distinct 22-37 fold increase, whereas in laparoseopic surgery a mere 2-3 fold increase was noted. 24 h after incision, the TrA-index returned to near pre-surgery levels.

In conclusion, our results demonstrate that changes in cell surface markers and cytokines can help evaluate the magnitude of tissue trauma in diffei'ent surgical approaches. The relationship between lymphocyte subpopulation changes after thermal injury and the increased susceptibility of burned patients to infection is unclear. In this study, we have attempted to correlate such subpopulation changes with the presence of infection in burned patients. Peripberal blood from 19 patients was monitored for lymphocyte subpopulation changes three times weekly for three weeks postburn and weekly thereafter for three additional weeks. Mean bum size was 44.7% (range 24%-84%) of total body surface and mean age was 38 years. Infection was diagnosed by carefully defined clinical and laboratory criteria and its presence or absence noted each time blood was drawn. Samples taken when patients had wound infection, bacteremia, or pneumonia were compared with samples taken in the absence of systemic infection. Whole blood samples were stained with four monoclonal antibodies, the red blood cells lysed and the leukocytes fixed and analyzed by flow cytometry. For each patient sample, the proportion of lymphocytes falling within the light scatter gates was determined as the percentage of cells negative for CD14 and most strongly positive for CD45. This percentage was used to correct each sample for the presence of debris or nonlymphocytic cells. The proportion of CD4 and CD8 positive cells was slightly greatc~ in the samples from infected patients, while the proportion of B cells (CD20+) was unchanged and NK (CD16+) cells were decreased by ahnos[ 50% compared to sampie~ li'om uuiuleclcd patients. The percentage of cells positive for CDIlb (c~ integrin) decreased sharply and CD4RO (memory cells) decreased slightly in samples from infected patients while the expression of the lymphocyte homing receptor and CD29 were unchanged. CD25 (IL2 receptor) and CD69 (early activation marker) were significantly increased in the samples from the infected patients while HLADR was unchanged. These changes in lymphocyte phenotype correlate with the presence of infection. If they closely precede or occur during the early development of infection they may be valuable clues to the mechanism of susceptibility following thermal injury. Trauma patients are subjected to an immediate massive impact on their host defense integrity due to the combined effect of tissue trauma, shock and endotoxemia. Cytoldnes are playing a crucial role within the course of an impaired cell mediated immune response (CMI) resulting from a disruption of intact M%/Tcell interaction. The current study was undertaken to further elucidate the mechanisms of dysfimctional CMI following major burn and mechanical trauma -via comparative analysis of mRNA expression and protein release. The major regulatory levels for different cytokines were determined in mitogen, respectively LPS stimulated peripheral blood mononuclear cell (PBMC) cultures of trauma patients on consecutive days (13) t, 3, 5, 7 and 10 post injury. We analyzed the cumulative data for Interleukin-1 beta (IL-I[3), IL-2, IL-8 as well as tumor necrosis factor alpha (TNF-~) and saw a considerable impairment of the protein release in the stimulated PBMC cultures until D5 post-trauma and recovery thereafter. *p < 0.05, ** p < 0.01 vs control Comparing the autoradiographies of the specific cytokine mRNA expression with the protein release in the supernatants, we saw a good correlation between mRNA signal intensity and protein synthesis for IL-8 and 11,-2, suggesting that for these cytokines the main regulatory mechanisms are located at the pre-/transcriptional level. For the other cytokines investigated one has to suppose posttranseriptional mechanisms.

The analysis of our data clearly indicates a severe impairment of forward regulatory immune mechanisms following trauma. Most likely the regulatory mechanisms, that are involved are greatly different among the cytokines investigated. It may be concluded, that depressed CMI responses post-trauma are partly due to an impaired pro-inflammatory cytokine production. The severity of the injury (ISS) correlated with the development at multiple organ failure (MOF-score; r=0.612). The levels of mediators and markers of the inflammatory response were generally higher in the more severely injured group (ISS>30, n=16). I1-6, 11-8, G-CSF, FPA, and C3a -levels differed significantly (p<0.005) between the ISS-groups (>-< ISS 30) at the time of admission, whereas on day 3 TNFa, C3a, 11-6, and EalPI showed significant differences. Beyond the first week, major differences were restricted to PGE 2 and C3a. The formation of two groups with respect to later multiple organ failure (MOF < 3; MOF > 2 n= 14) yielded similar results. Leukocyte-FACS analysis revealed significant differences mainly in the CD14 (monocytes), CD3/CD25 (I1-2R + T-cells), and CD29/CD4 (TH 2calls) populations. Summarizing our findings we were able to detect some alterations in the surface antigens of immunocompetent cells. The inflammato D response, however, seemed to be more pronounced and correlates wi~ the further clinical course. Using an experimental bum model in rodents, we have demonstrated that administration of a full thickness, scald burn involving 20% or more of the total body surface area (TBSA) elicits systemic responses which are characterized by numerous alterations in T-ceU function (i.e., lymphokine production and contact hypersensitivity (CH) responses) plus an enhanced susceptibility to bacterial infection. In the present study we questioned whether the apparent systemic effects mediated by large burns would be elicited as site-specific alterations in immune function following administration of small area burn trauma (5% TBSA). Following a 20% TBSA burn, CH responses to contact sensitizing antigens were found to be altered. The depression in CH responses could be induced independent of the site used for topical skin sensitization. Following a 5% TBSA thermal injury, development of CH responses were affected in a site-specific manner. Immunization of 5% TBSA thermally injured mice in a site near the position of the burn resulted in depressed responsiveness, whereas immunization through a contralateral site resulted in responses that displayed both the intensity and kinetics of a CH response equivalent to sham-bumed mice. Similar systemic and site-limited changes in lymphokine production were observed with 20% and 5% TBSA thermal injuries, respectively. A 20% TBSA injury affected the lymphokine producing potential of all cells regardless of which lymphoid tissue the cells were isolated from. The effect of a 5% TBSA burn was significant but site-specific. Thus, ceils from lymph nodes receiving drainage from thermally injured tissue were specifically affected, whereas lymphokine production by cells from lymphoid organs receiving drainage from unaffected skin was normal. It was concluded that modulation of lymphokine production and cellular immune responses may be a normal consequence of burntrauma regardless of the size of the burn. Changes in immune competence can be mediated either regionally or systemically in direct proportion to the area of skin exposed to the burn injury. This work is supported by PHS grant GM46899 and the Office of Navy Research N00014-92-J-1612.

Division of Cell Biology and Immunology, Department of Pathology, University of Utah School of Medicine, Salt Lake City, UT 84132. Post spleneetomy septic sequelae may Be fatal, but the mechanisms remain unclear. The objectives ef this study were to assess the mortality from concomitant splen-'etomy and ]~eritoneal bacterial challenge and to elucidate the local cetkdar responses. 50 CD-1 mice were randomised to receive laparotomy and sham splenectomy (L) or splenectomy (S) with simultaneous ca'-cal ligation and "):mcture and the survival patterns assessed. Subsequently, 150 CD-1 mice were randomised into control (C), L or S groups and peritoneal cells studied at 24 hours for bacterial phagocytosis and killi:~g, superoxide (02-) and tumour necrosis factor (TNF) production and macrophage activation vsing MAC-I(CD-11b) receptor in~.ensity expressed es mean channel of fluorescence (MCF). These resides indicate that SF!enectomy predisposes to nrortal~ty from bacterial sepsis ia the early pos~ operative period compared to sham operated animals. Failure ~f P'.acrophages to kill bacteria in the splenectomv group '~:cured in t?~e absence of impairment of oxygen freeradical or TNF pred:~ctien. The macrovh~ge ac!ivotion marker MAC-1 was significantly reduced in both L and S groups and impaired phagocytosis of bacteria oceured in both operative groups compared to controls. Laparotomy a!one reduces macrophage activity in terms of surface re:eptor MAC-1 expression and !ingestive capacity. Splenectomy however s~gnificantiy ~mpairs r-acrophage-wediated l~,acterial killing and this qefect rTtav co~tribut~ sig~ifJcav'ly to th-~ dissemination of local infection and to n':ortalit).

Depts of Haem~ tology & Surgery, Beaumont Hosoital, Dub!in 9,Eire. INTRODUCTION: Loss of cell membrane integrity appears to be a common pathway of injury to tissues subjected to high-voltage electrical shock. The cell membrane is the most heat labile structure in the cell, and is also the most vulnerable to externally-imposed electrical forces. Skeletal muscle and nerve cells are particularly susceptible to electroporation by clinically relevant electric fields. Restoration of membrane integrity is essential for cell survival in victims of electrical shock.

We have studied the effect of non-ionic triblock copolymers ( poloxamer class) on the transport properties of isolated rat skeletal muscle cells following electroporation-induced membrane disruption. 1-3 mm long adult skeletal muscle fibers were isolated by enzymatic digestion from the rat Flexor Digitorium Brevus and maintained under standard culture conditions. They were loaded with the Calcein-AM dye and placed in a 37,C chamber for recording by real-time video confocal microscopy. The cells were subjected to 4 mSec, 150 V/era, 1 A field pulses with a low duty cycle to allow thermal relaxation. Peak temperature rise was 0,2.C. The uye content of the cell was monitored in real time. Experiments were carried out in calcium-free phosphate buffered saline, with 1 mM Mg%. Experiments were repeated with 4 mM neutral dextran ( The aim of the present paper is to ascertain if thuracotomy induces a different pattern of variations of cytokines, immunocompetent cells and antibodies from laparotomy in the early postoperative period. 60 patients (34 males 26 females,mean age: 59.5_+5) 30 with gallstone disease and 30 with non neoplastic pulmonary disease were studied. None of these patients received blood transfusion, biological response modifiers, radiotherapy or surgery for at least 6 months before being included in our study. Anaesthetic procedures were similar in all patients and none were matnourished. On the day of surgery and on the 1st and 7th postoperative days (Pre, lPO, 7PO) percentages of CD19, CD5, CD4, CDS, CDI6 were measured by means of flow cytometry using MoAb., and levels of Ig A, lgG, IgM, IgE. by nephelometry Cytokine levels in peripheral blood(IL-1, IL-2, IL-4, IL-6, TNF) were measured in 10 pts. of each group by means of ELISA using MoAb. _R. ESULTS:Variations of iL-2 and IL-4 were not s.s.. IL-6 increased but differences between groups were not statistically significant (s.s). IL-I decreased on 1PO and increased on 7PO in both groups but were only s.s. in the TH.g., and therefore, the differences between groups were s.s (P<0.05).TNF decreased in the L.g. and increased in the Th.g. on the 1PO, the difference was s.s(P<0.05); on 7PO, TNF decreased in the L.g. and decreased in the TH.g. but these variations were not s.s. Cell percentages decreased an lPO and increased on 7PO, except for %CD19 cell that increased on lPO and decreased on 7PO ,in both groups of pts. Differences were not s.s. Ig A, IgM decreased and IgE increased in both groups (P< 0.0i), but differences between them were not s.s. In contrast, IgG decreased on 1PO (P<0.01) and increased on 7PO in both groups, but the decrease iu the TH.g. was greater than in the L.g. Twenty male children,aged from six months to 3 years,admitted for elective inguinal operation were studied. The operations were performed under balanced combined anaesthesia (fentanyl,thiopemtone,vecuronium, 70% nitrous oxide in oxygen) and blood samples were collected before flunitrazepam premedication,after anaesthesia,4 and 24 hours after anaesthesia. Cells from the wound were collected with Cellstick sponge which was removed from the wound 4 or 24 hours after anaesthesia. The study was approved by the local Ethical Committee.

The percentage of neutrophils was increased and that of lymphocytes was decreased in perpheral blood after the operation.The values in the wound were close to the values found in peripheral blood. The percentage of T-lymphocytes (CD3) and helper-T-cells (CD4) decreased in peripheral blood being lower in the wound than in peripheral blood after the operation. The percentage of T-eytotoxic cells (CD8) also decreased in peripheral blood and was similar to that in the wound. B-lymphocyte (CD19) percentage was increased in pe~pheral blood after the operation and was higher than in the wound. The percentage of activated T-cells (CD3+HLA-Dr-positive cells) in peripheral blood increased while that of natural killer cells (CD3+CD16+Leu19-pos) was increased just after anaesthesia being decreased at g and 24 hours after the operation. Spontaneous lymphocyte proliferative responses didn't change while phytohemagglutinin A and concavalin A induced responses were decreased in peripheral blood samples 4 hours after the operation with recovery at 24 hours.Pokeweed mitogen induced lymphocyte proliferative responses were decreased at 4 hours (p<O.01) and 24 hours after the operation We previously reported that plasma IgE increased after traumatic injury. In this study, we measured plasma IgE Ievels in 40 trauma patients on hospital admission and 3x weekly in the intensive care unit to evaluate IgE kinetics. Patient characteristics were: mean age 36-+ 16 years (33 m, 9 f), mean Injury Severity Score (ISS) 26+ 10, sepsis (19 patients), sepsis syndrome (9 patients), ARDS (4 patients), renal failure (5 patients).

IgE increased in 28 patients (70 %). The mean increase was 527 ng/ml (range 0 to 7362 ng/ml; 95% CI was 160 to 895 ng/ml). Increase in plasma IgE was significantly greater in patients with sepsis syndrome (p=0.021) and renal failure (p<0.001). Although mean plasma IgE was higher with ARDS, pneumonia, hepatic dysfunction, and shock, these differences were not significant (p>0.05). Plasma IgE increase was not related to severity of injury by ISS score (p =0.42). The mean day to highest IgE was 9.1-+6.5. The day sepsis was first observed preceded the day of highest IgE by 4.2+6.6 days. There was a significant association between the day of sepsis onset and the day of highest IgE (p=0.018). Eight of nine patients with sepsis syndrome had > 100% increase in plasma IgE from admission. One patient's IgE levels were normal (20-100 ng/ml) for 24 days and then increased to 7300 ng/ml over the next 12 days, after onset of sepsis syndrome. Changes in IgE plasma levels may reflect the action of cytokines, such as IL-4, which concurrently regulate production of IgE and IL-1 receptor antagonist in a response to sepsis. Sepsis remains a leading cause of late mortality in trauma and HS. Although HS-induced bacterial translocation is supposed to be the major cause of sepsis and MOF, depression of the RES increases susceptibility to infection after injury. The purposes of this study were: a) to evaluate the RES in the lung, spleen and liver after HS and subsequent Hypertonic Saline (HSL) treatment, and b) to document the patterns of phagocytic activity in these organs during 24 hrs. Adult male Wistar rats (250+_23 gin) were submitted to HS (SBP 40 tort) and after t hr (Shock I hr) and 24 hrs (Shock 24 hrs) HSL (NaC17.5%, 3.5 ml/Kg) treatment, 1131 E. coli (I 131) was injected into the portal vein 10 ~tCi (n_>12). Twenty minutes later, the lungs, spleen and liver were harvested and scintilographic counts obtained. Data is depicted as mean_%+SEM * p<0.05, ~" p<0.001 and statistical analysis was performed by Analysis of Variance and Wilcoxon tests. One hr after treatment, lung 1131 uptake was increased and liver and spleen 1131 uptake were reduced compared to sham. Twenty four hrs after treatment, all organs, except lung uptake, returned to normal values. Radioautographic histological analysis revealed radiolabeled particles inside phagocytic cells of all organs. We conclude that pulmonary phagocytic activity increases after 1 hr of HS HSL reatment, diminishing by 24 hrs although still above normal values. In contrast, RES suppression occurs in liver and spleen after 1 hr HS HSL treatment, returning to normal values by 24 hrs. These results may explain lung complications and immunosuppression after trauma. Infusion of endotoxin as well as major surgery is followed by lymphopenia in peripheral blood. The purpose of this study was to investigate to which tissues the lymphocytes are redistributed in response to endotoxaemia and major surgery. In addition changes in lymphocyte subpopulations and expression of MECII was measured.

Lymphocytes were isolated from peripheral blood of 30 rabbits, labelled with 111Indium-tropolene and reinjected intravenously into the rabbits, i0 rabbits received an infusion of Escherichia coli endotoxin 2 ~g/kg, while i0 rabbits were subjected to a major sham operation and i0 rabbits served as a control group. The redistribution of lymphocytes were imaged with af gamma camera, and calculated with an interfaces computer before, and 2, 4 and 6 hours after major surgery or infusion of endotoxin or saline. Interleukin-l~ and serum cortisol were measured. In addition we followed CD2, CD3, CDlla/b, CDIS, CD20, CD44, MHCII and CD4/CD8 ratio. Following endotoxaemia Interleukin-lf~ increased significantly, following endotoxaemia as well as major surgery serum cortisol increased significantly. Following major surgery as well as endotoxaemia there was significant lomphocytepenia in peripheral blood with a decreased CD4/CD8 ratio while the CD3 positive subpopulation increased. In addition there was a decrease in the expression of MHCII on the lymphocytes peripheral blood. The radioactivity of the lymphatic tissue in and around the intestine increased to 129% of initial values following endotoxaemia and to 125% following major surgery. The results indicate that endotoxaemia as well as major surgery induces redistribution of lymphocytes from peripheral blood to lymphatic tissue. Among the lymphocytes staying in peripheral blood there was a decreased expression of MHCII and a relative decrease in CD4 cells compared to CD8 positive lymphocytes. In order to analyze the effects of immune suppressive substances on expression of mRNA of interleukin-2(IL-2) and interleukin-2 reeeptor(IL-2R), this study was carried out. Twenty male rabbits with comminuted fracture were used in the study.

Ten ml blood were taken at 0, i, 2, 4, 6 days after injury. The sera were tested for the effects on lymphocyte blastogenesis and induction of IL-2 stimulated by concanavalin A(Con A): The sera from the rabbits 2 days after injury were analyzed with SDS-PAGE gel eleetrophoresis, and divided into three groups by ultrafiltration (UFPI TTK, 30KD,Milipore; Centricon-10,10KD,Amicon), that are less than 10KD, between i0 and 30KD, and more than 30KD.

Each group of the substances also was tested for the expression of IL-2 and IL-2R by the dot blot hybridization. The results showed that: i) All sera from the rabbits after injury had significant suppression on lymphocyte proliferation and secretion of IL-2 by the Con A-stimulated splenocyte in mice; 2) The sera from the rabbits 2 days after injury had more profound suppression than other injured sera; 3) There was a marked band at about 10KD in sera from the rabbits 2 days after injury, but nothing at the same position in normal sera analyzed with electrophoresis; 4) The substance with molecular weight of about IOKD had more obvious suppressive action on expression of mRNA of IL-2 and IL-2R than other groups substances, of which molecular weights are more than 10KD. It is concluded that:

I) The sera from the injured rabbits can reduce immune response; 2) There is kind of substance, of which molecular weight is about 10KD, it is probable the main factor involved in the pathogenesie of postinjury suppression immune; 3} The substance can depress the expression of mRNA of both IL-2 and IL-2R.

Research Institute of Surgery Daping, Chongqing, 630042 P. R. China

Acute ethanol uptake prior to injury modulates monocyte TNFo~, production and mononuclear cell apoptosis. G. Szabo, B. Verma, P. Mandrekar, D. Catalano Monocytes (MO) have been shown to contribute to immunosuppression after both major injury and alcohol consumption.

We reported that acute ethanol exposure of M(3 results in decreased antigen presentation, induces TGF-13 and PGE2 while inhibiting inflammatory monokine production. We also showed that post-trauma immunosuppression is mediated by hyper-elevated MO TNFc~ and IL-6. Consequently, here we investigated rnonokine production in trauma patients (n=10) who had elevated (>O.lmg/dl) or had no blood alcohol level (n=t0) at the time of emergency room admission. None of the patients had chronic alcohol use history. Met TNFc~ production from trauma patients with prior alcohol uptake was undetectable during days 0-3 post-injury in contrast to patients without alcohol exposure.

Furthermore, decreased TNF~x levels were found in alcoholic patients' MCi after MDP or IFNy + MDP induction. However, MCl TNFc~ levels during the 4-8 days post injury period became higher in alcoholic trauma patients. Furthermore, over 9 days post-injury, alcoholic trauma patients showed significantly elevated MCi TNFo~ production after adherence isolation, MDP, or IFN+MDP stimulation compared to patients without alcohol. These results suggest that acute ethanol uptake prior to injury decreases TNF(x inducibility in the early post-trauma period, but these patients' MO produce hyper-elevated TNFa levels later post-injury, thereby prolonging their cytokine shock risk. TNF ng/ml 0-3 days post-injury 9 days post injury Stimulus Ale. Pt.

Pt 9.3 11.9 5.6 4.0 Immunosuppression might also be increased by the elevated apoptotic activity found in trauma patients' mononuclear ceils, which was even greater in alcoholic trauma patients' cells.

In non-alcoholic trauma patients' preactivated MO, in vitro acute ethanol (25-150mM) exposure resulted in a significant down-regulation of TNFc~ (p<0.001) and IL-6 (p<0.01) production. In contrast, in vitro ethanol exposure increased the production of inhibitory monokine, TGFI]. These results provide both in vivo and in vitro evidence for the effect of acute ethanol exposure increasing immunosuppression and cytokine shock. The 'systemic inflammatory response syndrome' (SIRS) with consecutive septic multi-organ dysfunction represents the major cause of late death following major mechanical and burn trauma. Systemic hyperinflammation and concurrent depression of cell mediated immune response (CMI) render the traumatized host anergic, resulting in profound susceptibility to opportunistic infection. Monooytes/macrophages (MO) play a central role within the host defense system in developing and manifesting states of injury, shock and sepsis. The mechanistic scrutiny of the synthesis patterns of crucial cccytokines appears to be a helpful tool to further analyse MO behaviour in the compromised individual. The objective of this study was to further dissect the characteristics of cytokine regulation in PBMC under stressful conditions, via analysis of the expression of CD14+ receptor, the proinflammatory mediator IL-6, the macrophage activating factor IFN-5,, and Neopterin (NPT) a metabolite of activated MO. We investigated PBMC's on consecutive days 1, 3, 5, 7 and 10 after mechanical trauma of 9 and after bum trauma of 5 patients (mean age 38~17 years; mean ISS 34±2 pts). In trauma patients we saw a massive increase of PHA induced neopterin synthesis compared to controls. However, when discriminating the NPT levels in the supernatants for the amount of MO stimulated, the NPT output of the individual cell was lower compared to MO of nontraumatized individuals. Interestingly there was a contrary coarse in the cumulative protein release patterns of IL-6 and IFN-7 in mechanical versus burn trauma patients. Wheras in burn patients IFN-y was decreased significantly (16+5 U/ml) compared to controls (58+6 U/ml) as well as mechanical trauma (59+12 U/ml). IL-6 showed a significant suppression following mechanical trauma (662+58 U/ml) vs control (1266+91 U/ml) and bum patients. The rt~,NA signal intensity for beth eytokines was in concurrence with the protein release in more than 85% of the individual patients investigated. From these data we can conclude that the inadequate low NPT synthesis predominantly in bum patients appears to be a sign of cellular immaturity and is probably partly due to low T-cell IFNo T signals. In addition we could state that the quality of trauma is apparently responsible for the different synthesis patterns of ]L-6 and IFN-q,. It has been postulated that bacterial invasion or endotoxemia are necessary for cytokine production following burn injury. We studied the organ distribution and kinetics pattern of IL-fc~ (cell-associated IL-1 agonist) in eutrophic rats subjected to either 20% TBSA cutaneous scald injury (BI), muscle scald injury of equivalent 20% TBSA (MBI), sham muscle bum (resection of skin only, up to 20% TBSA) (SMBI), and sham cutaneous burn (SBI), followed by saline resuscitation (15 mUkg I.P.). Separate rats were infused with 15 mg/kg E.coli 0111:B4 LPS or saline LV. Unmanipulated rats were baseline normal controls. Liver, lung, spleen, ileum, thymus, kidney, skin, and plasma were harvested at various time-points within the first 24h. Tissues were frozen, weighed, homogenized, the homogenates centrifuged and the supernates assayed with a radioimmunoassay specific for rat IL-l(z (detection limit 150 pg/rnL). IL-lc~ was expressed as ng/g weight + SEM (lowest detectable amount 1.5 ng/gWT). IL-lo~ was constitutively present only in the skin (102 + 16.4 ng/gWT). Cutaneous burn and sham cutaneous bum induced biphasic elevations of IL-lcc in the liver and lung only, with maximal levels at 2.5h (in the liver, BI = 16.5 _+ 6.2 ng/gWT, SBI = 1.7 + 0.1 ng/gWT, p _< 0.05; in the lung, BI = 10.3 + 1.3 ng/gWT, SBI = 1.9 + 0.8 ng/gWT, p -< 0.002). Of note, both BI and SBI rats had detectable IL-I~ in the liver at timepoint 0 already (2 min real-time). These levels increased in parallel until 30 min and became eventually different by 1 log at 1 -2.5h. All other organs as well as plasma were below detection limits. Muscle burn injury and sham muscle burn (skin resection) induced similar elevations of IL-10~ in the liver at lh, indistinguishable from each other and from cutaneous burn. In contrast, LPS challenge induced dramatic elevation of IL-t~ in all organs tested except for the kidney; the spleen was the most responsive organ to LPS-induced IL-lo~ production. These data indicate that thermal or mechanical injuries induce very early and organ specific production of IL-1 c~ in vivo by mechanisms other than endotoxemia. Injury-induced complement and platelet activation may be involved as well as the neuro-endocrine axis, which may explain the low levels of IL-lo~ induction observed in all rats at the very early time-points.

Trauma Services, Massachusetts General Hospital, and Department of Surgery, Harvard Medical School. Fruit, St, Boston, MA 02114.

J. F. Schmand *#, A. Ayala* and I. H. Chaudry* Studies indicate that i.v. infusion of the colloid HES in normal animals does not adversely affect non-specific immunity. It remains unknown, however, if lIES affects cell mediated, specific immune functions after trauma and hemorrhage (hem). To study this, non-heparinized C3H/HeN mice underwent midline laparotomy to induce trauma and were then bled to and maintained at a BP of 40 mmI-Ig for 60 rain. The animals were then resuscitated with either 4 times (x) the shed blood vohune as lactated Ringer's solution (LRS) or 2x LRS + lx 6% lIES. Sham mice were neither hemorrhaged nor resuscitated. At 2 or 24 hours post hem serum, peritoneal (PM~) and splenic macrophages (SM~) were obtained. Bioassayes were employed to assess the levels of II-l, IL-6 ( Alternatively PMqb showed no differences in IL-1 release between all groups at 2 and 24h, while SM~ from the LRS + HEN group showed a depression at 24h. TNF production by PM~ was depressed in all groups at 2h and remained so in the LRS + HES group at 24h. SM~b showed decreased TNF release values in both hem groups at 2 and 24h. In summary, the levels of inflammatory cytokines (particularly the values of circulating IL-6) after trauma/hem are positively influenced by the administration of HES. This might be due to a protective effect on PMqb and SM~, but also on other cytokine producing cells, e.g. Kupffer ceils. We conclude that HES is not only a safe, but also beneficial agent in the resuscitation of patients atler trauma/bemorrhagic shock. This study investigated endotoxemia and consecutlve immune response in 39 patients with multiple trauma (median Injury Severity Score = 20,5). Blood samples.were collected shortly after injury and after 0,5, 1, 3, S and l0 days. Endotoxin was measured with limulus-amebocyte lysate test and the specific antibody content (SAC) against endotoxins of the classes IgG, IgM and lgA by ELISA-technique. Five antigens were used: Lipopolysaccaride (LPS) of E.coli (EC), Lipid A of E.coli (LA), LPS of Pseudomonas aerog. (PA), LPS of Vibrin cholerae (VC) and cX-Hemolysin of Staphylococcus anreus (OtH). A nephelometer indicated the total concentrations of IgG, IgM and IgA. Differences were checked with Wilcoxon-test and p<0,0S was considered significant. Cross-reactivity was calculated with rank correlation coefficients.

Results: Endotoxemia peaked shortly after injury (0-3h) at 0,425 EkI/ml (median), decreased thereafter to 0,04 EH/ml at day S and remained on this level. SAC oflgMclass increased to all endotoxins and peaked at day 3 revealing the lfighest level to LA followed by PA (= 80% of LA-SAC), EC (= 65% of LA-SAC) and VC (= 40% of LA-SAC). lgA antibodies increased as well but only slightly and not significant (exception: SAC to LA was elevated significantly at day 5). IgG antibodies increased similar to IgA class only slightly and again only SAC to LA was significantly higher at day 5 and 10. However SAC to (XH of all Ig-classes remained continuously on the same level troughout the observation time. Correlation analysis revealed strong cross-reactivity (r>0,5; p<0,01) most often between antibodies of IgM-elass (73%) followed by IgAclass (40%) and lgG class (2%].

Conclusions: Multiple trauma is associated with temporary endotoxemia. Endotoxins probably translocated from the gut cause specific increase of anti endotoxin antibodies in blood of the IgM-class. Endotoxins cause no increase of antibodies to gramposilave bacteria. IgM antibodies are most unspecific. During cardio-pulmonary bypass, as well as postoperatively, high levels of endotoxin, interleukin-6 (II-6) and C-reactive protein (CRP) were measured in 30 patients. I0 female and 20 male, ageing from 30 to 73 with a median age of 59. Blood sampling was done preoperatively, immediately after induction of anaesthesia, after thoracotomy, after cannulation of the aorta and right atrium after the first half of the reperfusion phase, after closure of the thorax, 2 and 4 hours after the operation and then every morning until the 5th postoperative day. Blood was drawn into heparinized tubes (I0 IU/ml) which were free of endotoxin. CRP levels were determined through the use of the Behring Nephelometer. 11-6 levels were measured by using commercially-available ELISA test. The endotoxin level was determined by a chromogenic modification of the limulus amebocyte test. The statistical analysis was done using the Wilcoxon ranks test and correlation analysis. A significant increase {p 0.01) in endotoxin plasma occurred during surgery, culminating in a peak (median value of 0.795 EU/m!) during reperfusicn. Plasma levels of endotoxin continued to be slightly raised till the 5th day after surgery, whereas those of interleukin-6 rose at the end of the operation and were at their highest 4 hours later (median value of 217.5 pg/ml). CRP levels were also high postoperatively with a median value of 114 mg/l, and were markedly raised on day 2 (191 mg/l). A definite, statistically significant correlation between the plasma levels of endotoxin and 11-6 during the operation was establisthed (p 0.05), leading us to conclude that the endotoxin liberated during cardiac surgery acts as the main trigger in the releasing of 11-6, and thus induces the postoperative acute phase reaction. There was no evidence of a correlation between CRP and endotoxin or 11-6 plasma levels. Impaired immune function is well described following trauma and hemorrhagic shock (HS). Prior studies have utilized peripheral blood or spleen cells to index immune function following HS. However, changes in mucosal immunity are not weII characterized in this setting. Gut origin sepsis is thought to be an important cause of organ failure and death following trauma. A rodent model was utilized to allow comparison of mucosal-associated immune function vs, systemic compartments after HS.

Fischer rates underwent HS (MAP 35±5mm Hg) for 60 minutes followed by resuscitation with shed blood and LR. Sham animals were instrumented only. Rat tears were collected at 24 and 72 hours following HS for quantitation of slgA by RIA. Animals were sacrificed at 24 hours and spleen (SPL), peripheral lymph nodes (PLN), and mesenteric lymph nodes (MLN) harvested for cell population analysis using flow cytometry and mitogen stimulation analysis.

Cell marker expression analysis revealed no changes in T or B ceil populations following HS.

Mitogen Mucosal immune function appears relatively spared following HS. The mechanism(s) for this variability in immune function requires further investigation. We have found that transplantation of bone marrow in a hind-limb graft to syngeneic lethally irradiated recipient is followed not only by rapid repopulafion but also overpopulation of bone marrow cavities. The question arises whether this unexpected phenomenon could be the result of stimulation of stem cells by factors (cytokines) released from surgical wound at the site of anastomosis of graft with recipient. AIM OF THE STUDY was to investigate which tissues damaged during the procedure of limb transplantation may be a potential source of humoral factors accelerating in vivo bone marrow proliferation. METHODS. Experiments were carried out on LEW rats in 5 groups. In group I, the hind limb was transplanted orthotopically to a syngeneic recipient; in group II, sham operation was performed; in group III, a four-cm long cutaneous wound was made on the dorsum; in group IV, limb skin was harvested, fragmented and implanted into peritoneal cavity; in group V, BM from femur and tibia was implanted intraperitoneally. BM, lymphoid tissues and blood were sampled 10 and 30 days later for cell concentration and phenotype evaluation. RESULTS. The yield of nucleated cells from tibia was on day 10 in the control 29.3 + 3.2, in group 1 58.1 +2.0, in group II 47.1 + 1.6, in group III 50.8+2.0, in group IV 48.5_+2.0, in group V 37.4_+4.9x10(6). The evident increase in BMC yield in all groups continued until day 30. Increase in weight and total cell count of spleen and mesenteric lymph nodes in all but group III was also found. No differences in percentage of maturing erythroid cells, but higher of mature myeloid cells and lower of lymphocytes were observed. CONCLUSIONS. Trauma of skin, muscles, and bone brought about an increase in bone marrow cellularity and acceleration of maturation of myeloid lineage. Transplantation of BM ceils alone did not produce this effect. Transplantation of BM in limb graft is a good model for studies of natural factors reaulatin~ BM hemormesis. This study sought to determine a relationship, if any, between the degree of hypochclesterolemia upon trauma patients' admission and their subsequent outcome.

All blunt and penetrating trauma patients admitted to a Level I facility from 1987 through 1991, and who had serum cholesterol assayed during the first 24 hrs were retrospectively studied for development of death or significant organ dysfunction.

The Mantel-Kaenzel chisquared test was used to determine significance of data at the p< 0.05 level.

RESULTS: 2909 trauma patients were admitted during the four-year period who had serum cholesterol assays performed in the first 24 hrs. 24 patients had cholesterol levels less than 50 mg/dl; 8 of these (33.3%) died, 3 (12.5%) developed ARDS, 5 (20.8%) developed acute renal failure, and 8 (33.3%) developed multisystem organ dysfunction; hypocholesterolemia in these patients was not due to liver injury or massive fluid administration.

The risk of death was 5 times greater and risk of multi-organ failure 3 times greater in this group than in those with a normal serum cholesterol (>if0 mg/dl; 1820 patients; p<0.05). CONCLUSIONS:

Admission serum cholesterol level in the trauma patient serves as a powerful marker for those at risk of subsequent organ failure or death. Hypocholesterolemia in this setting may result from organ hypoperfusion and humeral mediator release. Lung tissue contains many immunocompetent cells. Resection, therefore, is expected to activate extensively inflammatory mediators such as PMN-elastase, pmstanoids and pteridines. In a prospective clinical study we compared 35 patients (pts) undergoing either thomcotomy with or without lung tissue msectioh and tboracoscopic lung resection concerning activation of inflammatory response.

Material & Methods: Group A pts (n=8) had thoraantomy but no lung tissue injury; group B pts (n=lS) had thoracotomy and lung tissue resection due to benign diseases; group C (n=9) represents group B tissue resection but using a thomcoscopic procedure. The following parameters were determined pre-, peri-, and postoperatively: Elastase and CRP as indicators of activation of PMN-leukocytes and injury severity; prostacyclin (PGI 2) and thromboxane (TXA~) as parameters of lung endothelial response; prostaglandin F2~ (PGF~) and PGM representing pulmonaly metabolic activity; PGE a and neopterin as proof of macmphage activation. Statistics were performed using analysis of variance for repeated measures. Results: Group B pts revealed postoperatively an increase in CRP (p<0.001) indicating a higher injury severity in comparison to the thoracoscopic procedure (C). Both, controls (A) and group C pts did not show PMN-activation, whereas group B demonstrated a reversible increase in elastase. Surgical trauma caused in all groups a release of PGI z and TXA 2 which was more pronounced in C (p<0.05) and most in B (p<0.01). Similar results were found for PGE~ and PGF2=. There was no activation of maerophages since neopterin did not increase. Apparently, metabolic lung function was not impaired because there was no marked rise in PGM except in B (p<0.05 vs. C). Discussion: Our results demonstrate that lung tissue injury aggravates the mediator release induced by thoracic traum. These mediators among others are able to increase capillary pressure and hence lung edema formation. Impairment of lung function, however, seems dependent on the extent of the liberation. Therefore, the maximal release reactions occured in group B and C after lung tissue resection, whereas the controls showed the highest levels immediately after the incision. We conclude that thoracoscopic procedures are superior in reducing the resection trauma per se and hence might prevent severe mediamr-induced (pulmonary/systemic) sequelae. In a prospective study we investigated 45 patients using radiochemical method according to Sch~dlich (S) and photometric method according to Hoffmann (H). Serum of severly traumatized patients was withdrawn directly after admission at our emergency room and in narrow time intervals during first 24 hours after trauma. Follow up control samples were taken daily until day ten. Whereas no elevated PLA-CA was found during first 24 hours, a peak was regularly observed around day four. There was high correlation between PLA-CA and ISS (r=0.71, p<O.O001) except patients suffering from thoracic trauma, Thoracic trauma {TT) provoked similar gons-score-and PLA-CA-values compared to multiple trauma (MT) despite significantly lower ISS: PLA-CA (H) MT 32.9 + 6.4

TT 28,7 +_ 9,5 n.s. PLA-CA(S) MT 11.7 +_ 2.6 TT8.3 _+ 2.6 n.s. Goris MT 3.5 4-0.7 TT3.6 _+ 1.3 n.s. ISS MT 35 4-3 TT21 4-3 p<0.01 We found delayed elevation of PLA-CA after severe trauma, However detection of delayed PLA-CA in the serum of traumatized patients cannot give reliable information about the exact role of PLA in the inflammatory reaction after trauma as latest results show that PLA is secreted early but bound at lipophilic sites of vascular membranes.

Hern~mdez M J, Amiguet JA, Monreal JI, Vid~n JR, Jim6nez F J, L6pez-Vivanco G Acute phase reactant proteins increment in serum of patients with traumatisms and inflammation has been previously reported. Alpha-1 antitrypsin (AAT), Alpha-2 macroglobulin (AMG) and Ceruloplasmin (Cp) are evaluated in the following of 33 patients, 27 males and 6 females (mean age: 43 years), with compound fractures. 57 % of them were located in tibia and fibula. Intercurrent infection developed in 4 patients and internal fixation rejection in one. Measurements were made by radial immunodiffusion on days 0, 3, 10, 17, 24, 31 and 38. A control group of 30 healthy volunteers was also evaluated. AAT and Cp showed increased values from day 3 which kept elevated untill the end of the study. AMG elevated from day 24. When infection developed, AAT and Cp values were even higher whereas AMG values decreased. Fixation rejection caused AAT, AMG and Cp increment. AAT and Cp increment in non complicated fractures might be due to accompanying inflamation, by means of cytokines release and APRPs hepatic synthesis induction. Therefore, increment is higher when infection or rejection develop. AMG behaviour at early stages of evolution is secondary to hiperconsumption, considering mean life shortening after proteases ligation. AMG modifications in fixation rejection remain obscure.

APRPs increment modulates inflammation and bony callus formation by means their antyprotease and antioxidant properties. Clinically, APRPs might be useful parameters in determining complications onset in fractures evolution. Hemorrhagic shock (HS) is a common complication of trauma, as well as some major surgical procedures. The alteration of the immune system by HS is thought to contribute to the increased septic morbidity and mortality seen in these patients. Multiple mediators are released following HS. IL-1, IL-6, TNF, PGE 2 and TGF-B have all been implicated in these immune system changes. In vitro, PGE 2 causes many of the immune system changes following HS, yet the time course of PGE 2 release has not been well deliniated. The aim of this study was to determine the time course of PGE 2 release by peritoneal and splenic macrophages following HS. C3H/HEN mice were bled to a mean BP of 35+5 mm Hg for 60 minutes. The animals were resuscitated with the shed blood and normal saline. Control animals were cannalated, but blood was not withdrawn. The animals were sacrificed at 2, 12, 24, 48, and 72 hours following HS. Peritoneal and splenic macrophages were harvested from each animal and cultured with LPS for 24 hours. The supernatants were collected and frozen until assayed. PGE 2 was assayed using an ELISA (Cayman Chemical). Results are shown below for the peritoneal macrophages: PGE 2 release by peritoneal macrophages was significantly elevated at 24, 48 and 72 hours over control. It was maximal at 48 hours, but by 72 hours it had started to decline. Splenic macrophage PGE2 release was much more variable. (Data not shown.) There was less consistency between time points and no clear trend was seen. In conclusion, PGE 2 is significantly elevated at 24 hours following HS, and reaches a peak at 48 hours. PGE 2 release may be responsible for the immune system changes seen in the first few days following HS. Splenic and peritoneal macrophages resPOnd differently in their release of PGE 2 following HS; which may be due to differences in their milieu. There is an obvious association between the altered immunocompetence of patients following traumatic injury of various types and the increased riskto develop complications. The present investigation was undertaken in order to analyze the value of both Neopterin (N) and C-reactive protein (CRP) in monitoring disease course in patients with multiple trauma. We were interested to compare Nconcentrations as marker of T-cell/Macrophage activation with CRP levels as indicator of the inflammatory acute-phase response Daily serum concentrations(N and CRP) were measured in 12 patients (9 male, 3 female, mean age 31.42), admitted to our ICU following serious trauma (mean ISS=34) and in 10 control subjects. The clinical course of each patient was evaluated with APACHEII score according to Knaus. Mean stay in ICU was 9.25 (range 5-13). 5 patients didn't develop organ failure(NOF-group), 4 patients survived, developing MOF (S,MOF-group), 3 patients died with MOF (NS, MOFgroup). Three patients(one S and two NS) showed signs of septicemia and septic shock.

The serum samples were measured for Neopterin (Immu-Test-RIA, Henning-Berlin) and for CRP (immunoturbidimetric method) -at the University Hospital for Internal Medicine, Innsbruck, Austria.

The highest CRP levels were measured 48h after trauma, but they were not accompanied by significant increase in N-values. We found significantly elevated N-concentrations from the days 5-6. N and CRP levels showed a parallel release peaks on the days 6-7 and the values discriminated significantly among the three outcome groups(the second CRP peak is lower in comparison to the first posttrauma peak). By the NS-group we found constant and parallel increasing CRP and N levels, reaching extremly high values 24-48h and preceding clinical signs of MOF and sepsis. The values increased dramatically before death.

The serum levels by all 10 control subjects were upper the normal limit for N and CRP..

The parallel occured peak changes in N and CRP concentrations showed a significant correlation with clinical status, using a disease activity score (APAC H Eli).

Our findings suggest that both-cellular and humoral mediated immune mechanisms are activated after multiple trauma and that simultaneous determination of N-CRP may be of advantage as diagnostic and prognostic parameter in polytrauma-patient monitoring panel. It is apparent, that measurement of very high levels may be of value as a srong indicator in developing of MOF or septic complications and thus offer the possibility for earlier therapeutic intervention.

The modification of T-cell/Macrophage and acute-phase responses may be potentially useful in the treatment of ICU patients, resuscitated after severe trauma, and suggest that N-CRP may be relevant indicator for testing experimental immunomodulating therapies. Although several in vitro studies suggested the catalytic action of iron in oxygen free radical generation, few data are available concerning iron-metabolism following ischemia-reperfusion injury and hemorrhage in vivo. Aim of the present studies was to evaluate iron metabolism following mild and severe hemorrhage in the presence and absence of intraperituneal hematoma. Methods. Male Lewis rats (300 g) divided in 4 groups (observation time 120 hrs): A: mild hemorrhage by sequential blood sampling (0.9 ml each)(0, 2, 4, 6, 24, 48, 72, 96, 120 hrs). B: A with hemoperitoneum (HP)(2 ml/100g), C: hemorrhagic shock: blood withdrawal of 3 ml/100g over 30 min and resuscitation with Ringer's lactate (2 x shed blood) and isogenic donor blood; D: C with HP. Parameters: serum iron ([g/dl], ferrozin-method); 59Fe labeling of erythrocytas (application of 1 Ci 59Fe i.a. into a donor rat, 5 days later bleeding of the rat and injection intraperitoneally in experimental rat; plasma transferrin ([g/ml], RIA). T-tests and Mann Whitney. Data are expressed as mean + SEM. Results. Compared to baseline, mild and severe hemorrhage caused a significant increase in serum iron at 24 and 48 hrs, respectively (212+30 vs. 248+35; 185+_28 vs. 255+_22). HP reduced this increase (209+_17 vs. 177+21; 194+25 vs. 197+-48, p<.05). I-IP significantly increased hemoglobin between 24 to 96 hrs in mild (12.1+_.8 vs. 14.2+.7) and severe hemorrhage (10.3+.6 vs. 13.1+.6). In mild and severe hemorrhage comparable maximal resorption (59Fe labeled erythrocytes in circulating blood) of intraperitoneal hematoma occured at 24 hrs (mild: 318+_71 vs. severe: 364+-114 counts/rain). HP increased survival-rate following hemorrhagic shock (no HP: 10/17; HP: 10/13). In mild hemorrhage no alterations of plasma transferrin concentrations. In shock group without HP, there was at 96 hrs a significant higher transferrin level than with HP alone (3.2+_.5 vs. 2.23 + .2). In conclusion, hemoperitoneum with hemorrhage prevents increase in plasma iron levels by increasing hemoglobin through resorption of erythrocytes and thereby suppressing iron mobilization from endogenous iron resources, e.g. liver. This suggests that intraperitoneal hematoma not necessarily promotes iron-mediated oxygen free radical production, especially since survival in these groups was improved. Increased transfarrin levels in group C at the end of the experiment suggest increased iron turnover which was not observed in shock groups associated with hemoperitoneum. Primary immune response to thymus-dependent (SRBC) and thymus-independent (Vi-polysaccharide) antigens was determined during 30 days after CCT. Antigenic challenge was made on day i, 5, 15 or 30. Number of spleen plaque forming cells (PFC) and serum haemagglutinin (HA) titre were studied on day 5 after immunization. CCT stimulated immune response to T-dependent, but not to T-independent antigen.

Enhanced response was noted when SRBS were given in posttraumatic day i, it was higher than preceding one after injection on day 5 and reached a maximum when antigen was presented on day 15 after trauma (PFC number 2.4 times exceeded the level of immunized, but nontraumatized rats). The rise of PFC number in spleen correlated with increased HA level. Thymectomy had been performed 30 days before CCT completely abolished posttraumatio stimulation of immune response to T-dependent antigen. Standard thoracotomy also enhanced humoral response when SRBC were administered on day 15 after the operation.

Thymus trauma modeled by pincett squeezing of its right lobe and performed at the same time with thoracotomy fully abrogated as well as thymectomy stimulation of immune reaction.

Thus, different experimental chest traumas are able to increase antibody production via thymus-dependent mechanisms. Endothelin is a 21 amino acid peptide produced by ischemic or injured endothelial cells. In vitro and in animal studies, ET is also a potent constrictor for renal mesangial and coronary vessels, a negative cardiac inotrope and an endocrine regulator. Our lab has shown that ET is an agonist for monocyte production of interleukins i, 6 & 8, prostaglandin E 2 (PGE2) , and substances which trigger superoxide production. Systemic ET levels increase in hypoperfused and septic patients, and ET may be yet another important cytokine in critical illness. But while ET has been shown to be a potent vasoconstrictor in healthy dogs, systemic vascular resistance does not increase in critically ill patients with high ET levels. (We hypothesize that this might be due to PGEz-mediated vasodilation predominating over ET-mediated vasoconstriction.) We next asked what happened to systemic ET levels in patients with major burns (>20%.) Ten hemodynamically stable patients resuscitated by a modified Parkland formula to a urine output >30 cc's per hour had ET levels drawn on admission, at i, 6 12, 24, and 48 hrs. ET levels were measured by radioimmunoassay. Mean levels were elevated at 205±28 pg/ml at all time points versus levels in healthy controls of 39±9. In summary, systemic ET levels increase significantly in patients with major burns. ET may be yet another cytokine playing a significant role in the immune, inflammatory and multiorgan dysfunction observed with major burns. Restoration processes in an organism after ischemic damage are realized through ~n~lammatory mechanisms~ the intensity of which is significantly defined by blood levels of neuropeptides. Myocardial infarction (MI) was chosen for studyin 9 these processes since it eradicates the influence of infectious factc~rs. 32 dogs~ in whom MI underwent different forms o¢ healer, g; bhn~ed ~h~t during the acute phase of the disease there was a characteristic rise of ne!~ropeptides in the blood. These neuropeptides had nociceptive and antinociceptive effects. Particularly substance P and -endorphins triggered off the development of compensatory and adaptive mechanisms and defined the intensity of inflammatory reaction at the zone of ischem~t: damage-Notable fall in substance P levels after an ~nitial increase, while The ~-endorphins stayed high was an important condition for non complicated healing of MI. On the other hand high levels of substance P with low ~-endorphin concentrations lead to increased infiltration o~ neutrophils into the infarction zone and weakened the activity of synthetic processes~ thereby leading to left ventricular aneurysm. At the same time low intitial levels of substance P slowed down the development of necrotic processes which lead to delay in refunctioning of the heart and complicated the healing process. Thus, regulation of the levels of neuropeptides in the blood in trauma forms a perspective method of its treatment.

OF LAPARASCOPIC VERSUS OPEN CHOLEOCYSTECTOMY C. Schinkel, S. Zimmer, V. Lange, D. Fuchs, E. Faist The impairment of immune function due to surgical trauma may be followed by deleterious septic sequelae. Compared to open abdominal surgical procedures (LAP), laparaseopic surgery (LSC) is associated with a decrease in hospital stay and in accelerated patient recover. The aim of the study was to evaluate the sensitivity of the immune sermn parameters of IL-6, SAA and neopterin, the percentage of CD14+ cells, the in-vitro IL-2 synthesis after mitogen stimulation and lymphocyte proliferation, in order to purposefully discriminate differences in the severity of trauma. We investigated the blood of 27 patients with cholecystolithiasis undergoing either laparascopic (16) or open (I 1) cholecystectomy on consecutive perioperative days -1, 1, 2 and 3. There was no significant difference between the two groups concerning age and sex. Patients with clinical signs of acute cholecystitis were excluded from the study. Operation time and hospital stay were obviously longer in LAP patients (67 versus 145 minutes, 5 versus 20 days) compared to the LSC group. Concerning the unspecific acute phase reaction we could show no difference in the increment of senun amyoid A (SAA) synthesis in the LSC group (D-I 40 + 1 lng/ml, D1 362 + 43ng/ml) versus LAP group (D-1 61 + 19ng/ml, D1 457 + 45ng/ml), while in serum IL-6 levels we saw a less steep increment in the LSC group (7-fold from D-1 to D 1) compared to the LAP group (10-fold from D-1 to D 1). The analysis of CD14+ receptor expression and serum neopterin did not reveal any difference between the groups. Lymphocyte function showed an impairment of proliferation to antigen stimulation in LAP (D -1:24.000 + 3.330 cpm, D 1:17.300 + 3.000 cpm) compared to the LSC group (D -h 22.500 + 4.200 cpm, D h 26.000 + 2.900 cpm). In both groups IL-2 synthesis was decreased post-operatively. Our data indicate that laparascopic cholecystectomy reusults in a less distinct unspecific acute phase reaction post-trauma compared to that following LAP. Neopterin serum levels and CD14 receptor expression show that these parameters apparently are less useful markers to detect differences of surgical trauma severity while it appears that the impact of LAP is reflected most impressively on the lymphocyte compartment. 

Trauma alters the host resistance of organism and is accompained by appearence of excgenic and endogenic proteins in the body.

To understand the molecular mechanisms of host resistans disorders in trauma, as a first step, the genetic regulatory mechanisms of immune response after antigen injection has been studed.

The appearence of specific protein factors (13-19 and 25 kDa), in the nucleus of rat splenic and brain cells, accordingly, was shown after immunization with sheep erythrocytes. The stimulatory effect of these factors on the IL-2 mRNA and IL-2 production was detected. The nucleotide sequences of the human IL-2 gene regulatory region bounding by the splenic nuclear proteins were determined between +39-141 b.p. The IL-2 trans-factors shows the affinity to splenic and thymic lymphocytes in vitro.

Thus, the antigen causes the appearence of specific protein factors in the cells,which act on the gene level,stimulate IL-2 production and the host resistance. These results cause the next step of experiments using the same model, but after trauma. These investigations will let us verify the hypothesis that the protein IL-2 gene trans-factors may play a definite role in the decrease of the cell immune responce after trauma. Confronted with the routine procedure of prophylactic treatment of candidates for surgery in a rural African hospital, we initiated studies on the fre'Quency of post-surgical malaria. In Tanzania 195 non-pregnant patients from rural areas were followed. Of preoperative patients 10% had a parasitaemia and those maintaining it showed no increase or complaints. Nine percent of patients without detectable parasitaemia before surgery came down afterwards and one-third had malaria-like complaints. Spinal and general anaesthesia were equally applied in these last patients. In Burkina Faso we studied 120 patients of which 16% had a parasitaemia on admission and 6% had postoperative malaria. Half of the surgical patients came from rural areas, whilst only 14% of those with malaria lived in the city (with much less exposure and immunity). 57% underwent major surgery and 43% minor. Bloodtransfusions (4% with parasites) never evoked a parasitaemia in recipients. Post-surgical malaria is thus a reality in about 10% of the adult cases, both in East and West Africa. Surgery evokes a cascade of factors, varying from cortison to interleukines and acute phase proteins; immune responses may temporarily be suppressed. Clinical attacks of malaria in otherwise immunes could be evoked by one of these factors. Though malaria can easily be cured, the differential diagnosis is difficult because of post-surgery fevers; we found that 16% was treated without justified indication. The involvement of "student-doctors" A. This study examines glucose uptake and hexose monophosphate (I~IP) shunt activity in normal human peripheral lymphocytes and polymorphonuclear leukocytes (PMN). Glucose uptake was determined by measurir,g the uptake of tritiated deoxyglucose, a non-metabolized glucose analogue. Adsorption of CO2 derived from [I-14C] glucose was used to determine kNP shunt activity. In vitro assays were carried out in hormone concentrations approximating normal and elevated trauma blood levels. (Normal -cortisol 0.12 ~g/ml, glucagon 80 #g/m1, epinephrine 20 ~g/ml, insulin 18 t~U/ml; Traumaeortisol 0.40 ~g/ml, glucagon 500 /*g/ml, epinephrine 420~g/ml, insulin 36 ~IJ/ml. Analysis of twenty subjects showed a reduction of 207° ~MP shunt activity by lymphoeytes and a ]3% reduction in glucose uptake by P~N in normal vs. trauma honTc,nes p < 0.05. Lymphocyte glucose uptake was also reduced by trauma hormones p~0.05. It ha~ be.ea~ suggested thgt idiopatNo pulmonary fibrous (Y.PF) [s a consequence of severe alveolar epithelial injury and is associated with an Nveolar irNammamry reactio~ and the presence 0f.neutr0phils. There~bre, neutr0pk~ chemoattra~ant~ are probably important in the genegs oft.he iNfial lesions of IPF. The obse,"wSon that stimulated macrophages are oR~n histologically promin~t in fibmfio [-~gs ~.nd am capable of p~oducmg a v~dery 0f flbrogenic pep'ides also a~gues for their role ~n the pathogenic prc~e~ oflPF. The observation that stimuMe~ maerophages ere often histologica[Iy prominent in fibrotio lungs and ~re ~pable of producing a varie~, offibroge.~e peptide~ also argues for tkek role in the pathogenic process, Therefore, we ha-~e tested the potentN for iater!eukln-8 (I1..-8) and mo~tocyte chemotacde pop, de 1 (X¢CP-1) to induce neutro~hil ~d mononuclear phagocyte accumuhdon in lungs of pafient~ with pulmonary .~r~idosis and I~F. Brenet~o.alveolar lavabo (BAL) fluids from 12 IPF and 15 sar~Qidosis patient were conexntrateA by reversed-phase chromatography, ~d II.4 arid MCP-I asso.~ed by ELLS& ehemotaxis mad enzyme-reieasing ~ssas's on msnocyte~ and neatrophiis. ELISA revealed significenfly elevated B AL-eoneentrations o£MCP-I (20.1 ng]mg aibumm) in purisms with p~monary sarcoidodis artd in IPF (41.8 ng!mg) in comparises to 1.1 normal individuals (4.24 ng/mg) and to 15 patients w~th obreic broneNtis (CB) (~, 16 rig/rag). Similarly, chemota*dc ac~A~' for monocles (MCP-1 e.qu/va]ent) was strongly increased in sareoidosis (86.03 ngJmg) as well as ~n 12F pag,nts (54.47 ng/mg). Norra.al indlvidu~s and CB patiants hzd a 7. Or 3-fold lower ~cn%i~y, re~peefively. Patients with IPF and sarcoidosi~ also h~l eIevated IL-8 IeveI~ (15.5 and 26.0 rig/rag, respe~veIy; nomzls: 2.14 rig/rag; CB: 4.23 ng/mg) mad nvatropMI ohemotax~ (60.25 ~'~d 49.68 nNmg, res!z~ztiveiy; aormals: 0.25 ng,'mg; CB: 1L06 ngmg). These data suggest that increased Ievels of borN MCP.1 ~d IL-8 may be oharacted~tie for ~arcoidosis or IPF_ It appears Iikely that both ehernoattraetants ~ontribute to the influx ofmonocytes and neutrophils into the pulmonary alveoIus and interStit~um in these dlsea~es.

We have recently shown that the combined administration of noninjurious doses of LPS and PAF in the rat produce ARDS-like lung injury characterized by neutrophil adhesion to lung capillary venules, neutrophil accumulation in lung parenchyma, pulmonary edema, and increased protein and neutrophil count in BAL fluid. This new paradigm of lung injury was associated with elevated serum TNFc~ and pretreatment with anti TNFa mAb dose-dependently prevented these responses. Also, the combined administration of LPS and PAF induced lung mRNA levels of TNFe~ (5 fold vs. LPS or PAF alone), lL-lg (80 fold), KC (60 fold) and IL-6. Taken together, these data suggest that this new paradigm of lung injury is cytokinemediated and that LPS/PAF in vivo can functionally couple to the activation of gone expression of a multi-cytokine network system, all of which may be involved in the pathogenesis of ARDS. Materials and methods. The sheep model included hemorrhagic shock and closed femoral nailing at day 0, 12hourly injections of E. coli endotoxin and zymosan-activated autologous plasma at clays 1-5 and further observation and measurements at days 6-10. From venous blood and bronchoalveolar lavage(BAL)fluid of ten merino sheep (mean weight 30 kg) neutrophil counts (10e6 PMN/ml blood or epithelial lining fluid-ELF-), the ELF/ plasma ratio of albumin (R), and the zymosan-induced (stim) and non-induced (spont) chemiluminescence response (CL) of blood (10e6 cpm/250,000 PMN), and of blood-and BAL-isolated PMN (10e6 cpm/25,000 PMN) were measured. For statistical calculations the Wilcoxon test was used. Data of the Changes in polymorphonucleur leukocyte (PIVINL) metabolism have been suggested to play a pivotal part in the post-traumatic systemic inflammatory response syndrome. The underlying cellular mechanisms which control this response are not yet completely understood. Since the 'Ca 2+ second messenger'-system has been shown to be involved in regulation of PMNL-'respiratory burst', we investigated changes in PMNL-Ca z÷ regulation in relation to oxygen free radical mediated injury. Methods. In 12 polytranmatized patients (mean injury severity score = 32) arterial and venous blood samples during 15 days. Daily evaluation of Horowitz-quotiant (PO2/FiO2), plasma lactate (mg/dl) and body temperature ( Results. Body temperature peaked at day 6 and 7 (day 0: 36+.4; day 7: 39.8+.4). Plasma lactate was significantly increased at day l (36+2) and day 7 (16.2+2). Hurowitz-quotient (day 0: 344+43) was low at day 4 (199+19) and day 7 to 10 (178+25)(p<.05). At day 7 a substantial rise in venous PMNL-superoxide production (day 5: 1.9+_.45, day 7: 3.3+.7, day 9: 1.72+_.8), oecured with significant increase in plasma lipid peroxidation (day 1:0.6+0.2 ; day 7: 1.3+0.2). PIVIN~-myeloperoxidase activity was high at day 2 (2.3+--.5) and then continuously declined (day 7: 1.3+.3). Plasma antiexidant activity (glutathione pemxidase) was reduced by 34% at day 7 (day 1: 5.3+.3; day 4: 4.4+_.2; day 7: 3.5+.2). Whereas basal Ca 2+ concentration remained unchanged (day 4: 68+_17, day 7: 68+_15), FMLP-stimulated cytosolic Ca 2+ mobilization increased at day 7 (day 4: 457+89, day 7: 11084410, day 9: 670+262). Conclusion. The present study in polytraumatized patients shows, that seven days after injury the agonist-induced PMNL Ca 2+ mobilization is significantly enhanced. At the same time, PMNL-oxygen free radical release and phagocytotic activity, systemic fever response and lactate concentrations were maximal. These observations were accompanied by post-tranmatic respiratory failure and in some patients by clinical signs of multiple organ failure. Preliminary data from an ongoing study using HES-and dextran-infusions in these patients show attenuation of this inflammatory response.

Stefan Rose, M.D., Trauma Surgery, Univ. of Saarland, 66421 Homburg/Saar

1Donnelly SC, 1Haslett C, 1Dransfield I, 2Robertson CE, 3Grant IS, 4Carter C, 4Ross JA, 5Tedder TF. 1Dept's of Respiratory Medicine, 2Accident & Emergency, 3Intensive Care, 4Surgery, University of Edinburgh, Scotland and 5Dept. Tumor Immunology, Dana Farber Cancer Institute, Boston.

The selectins are a family of adhesion molecules (L-selectin, E-selectin, Pselectin), all of whom are implicated in inflammatory cell transendothelial migration. They, as a family can be proteolytieally cleaved from their parent cell and exist in a soluble form within the circulation. ARDS is a disease state in whic neutrophils and neutrophil transendotheliat migration have been implicated. In this study we wished to investigate whether the levels of these circulating soluble receptors from patients at-risk of ARDS at initial hospital presentation, correlated with subsequent ARDS progression. Eighty-two patients were enrolled (pancreatitis (n=19), perforated bowel (n=12), and multiple trauma (n=51)), of whom 14 progressed to ARDS. Assays for soluble L,P & E-selectin were performed on collected plasma samples via a sandwich ELISA. (NS = Not significant, **** = p<O.0001) We have found a highly significant inverse correlation between a low circulating soluble L-selectin (and not soluble E & P-selectin) and subsequent ARDS. These results further our understanding of neutrophilendothelial interactions in the early stages of ARDS pathogenesis and offer the promise of sL-selectin in combination with other markers of inflammatory events being used to identify individuals at particularly high risk of ARDS progression on initial hospital presentation. It has been suggested that polymorphonuclear leukocytes aggregate in the lung capillaries of patients developing the adult respiratory distress syndrome (ARDS) and account for the endothelial damage while releasing toxic metabo-Iites such as o.a. free oxygen radicals. Among many antioxidants which have been investigated in in vitro systems and in animal models, N-acetylcysteine (NAC) has been established to be useful as a free radical scavenger in vivo. To assess the influence of N-acetylcysteine (NAC) on the activation of neutrophils in patients undergoing cardiopulmonary bypass (CPB) surgery with extracorporeal circulation (ECC), we evaluated the plasma levels of elastase and myeloperoxidase (MPO) throughout the operation and up to 24 hours after surgery. Four hours after surgery also a bronchoalveolar lavage was performed. A spectrophotometric method was used for the detection of plasma elastase activity and MPO levels were measured using a radioimmunoassay. We found that pretreatment with intravenous NAC prevented the increase in elastase activity (632 _+ 231 pmol/I vs 1666 _+ 187; P -0.027), but not the release of neutrophil related mediators (2.20 _+ 0.56 ng/I vs 1.80 _+ 0.17; P=0.63). Besides, NAC had the capability to prevent the enhancement of neutrophil numbers in lavage fluid as is normally seen during CPB (1.4 + 0.8 % vs 25.0 _+ 14.3; P =0.04). Although not significant, NAC prevented also the increase in elastase levels in lavage fluid (19.0 + 18.5 pmol/I vs 172.7 + 101.4; P = 0.12). It is concluded that NAC is able to protect against the inactivation of a 1-proteinase inhibitor, the main inhibitor of elastase activity, and that NAC interferes with adhesion and migration of neutrophils. Therefore NAC may be useful in the prevention of tissue damage. This study is supported by Inpharzam Belgium. Sodium nitroprnsside (SNP) has previously been shown to attenuate the neutrophil induced lung injury associated with limb ischaemia and repeffusion. Glyceryl Trinitrate (GTN), already widely used in aortic surgery, also increases nitric oxide but has a less marked splanchnic "steal" effect. We examined the effect of GTN on lung injury, neutrophil infiltration and activation in a model of aortic occlusion (30 rain) and repeffusion (120 min). Sprague Dawley rats were randomized into: control (N=I 1); ischaemia repeffusion (IR) (N=12); and IR treated with GTN (2ug/kg/min) during repeffusion (N=10). Myeloperoxidase activity (MPO) measured pulmonary neutrophil influx. Pulmonary endothelial permeability was measured by wet to dry weight ratio (W/D), broncboalveolar lavage protein (BAL) and neutrophil counts (BAL pmn). Neutrophil superoxide release (mean channel fluorescence MCF) was measured by flow cytometry in a further IR v GTN experiment (N=6 per group).

Control 573_+t02"* 992_+125 579_+54** BAL pmn/mm 3 36l+-126 3"219+1087# 820_~221 *p<0.05, #p<0.02 v Control/GTN;**p<0.01 v IR. Students t test The significant increase in MPO activity produced by IR was attenuated by GTN. The increase in pulmonary microvascular leakage after reperfusion was also prevented by GTN. Neutrophil superoxide release increased significantly from 9.4+0.76 to 13.9_+1.92 MCF after 40 minutes repeffusion (p<.01). This increase in superoxide was prevented in the GTN treated group. These results indicate that GTN inhibits neutropbil superoxide release during limb reperfusion, thus preventing pulmonary vascular leakage and neutrophil infiltration. These data suggest that the beneficial effects of GTN in aortic surgery may be due in part to a protective effect onpulmonary microvasculature.

Department of Surgery, Beaumont Hospital, Dublin 9, Ireland. LIPTON Adult respiratory distress syndrome (ARDS) is marked by increased vascular permeability of the lung to white blood ceils (WBC). Indeed, influx of WBC into the lung is believed to be the central mechanism in acute lung injury of ARDS. Evidence is developing that the neuropeptide c~-melanocyte stimulating hormone (c~-MSH), a proopiomelanocortin derivative, inhibits inflammatory reactions in animaI models of inflammatory responses in humans. To test the influence of a-MSH on experimental ARDS, the peptide was administered to rats after intratraeheal infusion of endotoxin (S. typhosa, 500 #g in 0.25 ml saline). Three groups (n=10 each) received: intratracheai infusion of endotoxin plus ip injections of o~-MSH (100 ~tg in 0.2 ml saline) at 0, 2, and 4 hr post-endotoxin treatment; intratracheal endotoxin infusion plus saline injections; control intratracheal saline infusion plus saline injections. The BAL data showed overall differences among groups (F2,29=6.2, p<.003), o~-MSH treatment inhibited endotoxin-induced WBC migration into the pulmonary tree (p<.05). Circulating WBC counts were markedly lower in both groups given endoroxin than in the control group (p<.01), but there was no difference in WBC count between these two endotoxin-treated groups. The results indicate that c~-MSH can reduce WBC migration in experimental lung injury. In further experiments we tested the hypothesis that c~-MSH, administered alone or in combination with an inhibitor of gram-negative bacterial growth, increases survival in a model of peritonitis/endotoxemia/septie shock. Cecal ligation and puncture were performed under sodium pemobarbital anesthesia (75 mg/kg, i.p.) in female BALB/c mice (16-20g) that were then randomly assigned (15-17 mice per group) to receive at time 0 and 3 hr later: control saline thjectioas (i.p.), injections of ~-MSH (100 ~tg), gentamicin sulfate (I0 mg/kg), or both ~-MSH and gentamicin. One ml of normal saline was given s.c. to each animal for fluid replacement. Both Ix-MSH and gentamicin treatments administered singly improved survival; when given in combination survival was even greater These findings suggest that the anticytokine ot -MSH molecule might be useful for treatment of systemic inflammation, perhaps particularly when used in combination with agents that inhibit bacterial growth. Previous studies showed a close relationship between xanthine oxidase activation, membrane damage and postischemic cardio-respiratory failure following aortic clamping and reperfusion. Aim of the present study was to evaluate alterations of polymorphonuclear leukocyte (PMNL)-metabolism and signal lransduction systems during ischemia/reperfusion induced by aortic clamping in man. Methods, In 14 Patients (3 female, 59+4 years) with elective reconstruction of infrarenal aortic anearysms, arterial (a) and venous (v) blood samples were obtained before clamping and declamping, and 60 min after declamping. Arterial PLA 2 concentrations were higher than venous (113+32 vs. 71+30). While cytosolic basal PMNL Ca 2+ concenlxation was not altered at the end of ischemia (v: 52.3+8; a: 42_+6) and during reperfusion (v: 60+8; a: 69+12), FMLP-stimulated cytosolic Ca 2+ mobilization showed a significant arterial increase as compared to venous (1652_+601 vs. 638+85, p < .05). These presented data indicate that the reperfusion syndrome after infrarenal aortic clamping is characterized by 1) pulmonary membrane damage possibly due to PMNL-degranulation, 2) activation of PMNL-superoxide radical production with release of activated PMNL in arterial circulation, and 3) a significant arterial increase of PMNL cytosolic Ca 2+ mobilization after FMLP-stimulation parallel to PMNL-activation. In conclusion, ischemia/reperfusion in man induces pulmonary damage and activation of PMNL superoxide production possibly due to an altered PMNL-Ca 2+ messenger system by inflammatory mediators (e,g. II-8, TNF, PAF). The septic lung diseases (i.e. ARDS) generally show distinct alveolar damage and surfactant disturbances. It is thought that alveolar macrophages are involved in specific release products like H202 and cytokines like TNF. In this pathway the type II alveolar epithelial cell (P2 cell) is not only an important target, but as recently shown it is also capable of producing H202. The aim of this study was to investigate the influence of endotoxin on H202 release of P2 cells. We obtained P2 cells from lungs of female wistar rats and investigated the effect of lipopolysaccharid (LPS) on H202 release of fresh isolated cells and on cells cultured for 2 days in a fibronectin matrix. Furthermore we conditioned isolated alveolar macrophages for 16 hours with 1 p.g/ml LPS and tested the conditioned medium (MCM) on cultured P2 cells. P2 cells 5 h ex vivo were _> 70 % pure, _> 95 % vital and had an basal H202 release of 64.5 _+ 1.8 nmol H202 per hour and million cells. Adding 1 p.g/ml LPS to the incubation medium the H202 release decreases slightly but significantly to 54.2 _+ 2.3 nmol. Adding 0.3 p.g/ml phorbol myristate acetate (PMA) to the basal incubation medium the H202 release increased 3-fold to 148.3_+ 26 nmol. PMA induced H202 release decreased to 128.3 + 41.5 nmol after addition of 1 p.g/ml LPS. After 2 culture days the P2 cells were _> 98 % pure and showed a PMA inducible H202 release of 174.8 _+ 9.9 nmoL Addition of 1 p.g/ml LPS had the inverse effect as on freshly isolated cells as it increased the H202 release up to 216.2 _+ 2.4 nmol. Addition of MCM to cultured P2 cells increases PMA-stimulated H202 release to 223.3 +_ 10.7 nmol. The release decreased to 191.9 _+ 2.4 nmol when an murine anti-TNF-alpha antibody was added. Vitality of cultured cells was > 96 % in all experiments. The results show that LPS has an direct effect on P2 cells cultured on fibronectin. We conclude that the observed additional stimulatory effects of MCM seems to depend on TNF-alpha. The induction of H202 release of P2 cells could be important for generating internal oxidative stress in P2 cells before external oxygen radicals exceed. The produced H202 did not necessarily damage P2 ceils, but it can effect surfactant metabolism, especially when extracellular H202 release of alveolar macrophages following an immune response is increasing. INTRODUCTION: Primary stabilization of femoral shaft fractures in patients with multiple trauma is beneficial. However, in patients with associated lung contusion we have found an increased incidence of ARDS, apparently associated with primary Reamed Femnral Nailing (RFN). Previous animal studies revealed, that perioperative disturbances of lung ftmetion appear to be related to the reaming procedure, IX~ssibly due to pulmonary embolizafion of bone marrow fat. In a prospective clinical analysis we compared effects of intrameduUary nailing with and withont reaming on parameters known to be related to ARDS-pathoganesis. In order to gain further insight into the role of endotoxin and cytokines in the pathogenesis of the adult respiratory distress syndrome (ARDS), we enrolled 23 patients with severe lung injury after sepsis (17) or polytrauma (6) and obtained multiple blood samples (14 days) for endotoxin, tumor necrosis factor e (TNFa), interleukin 18 (IL-18) and interleukin 6 (IL-6) determination.

To evaluate the cytokine releasing capacity of the blood, plasma concentrations of TNFe, IL-l8 and IL-6 were also determined after the "in vitro" stimulation of the whole blood samples with lipopolysaccharide (LPS, 0.1 ng/ml) for 2 hours at 370 C (stimulated values). The difference among stimulated cytokines levels and the basal plasma concentrations were defined as "delta values", an expression of the cytokine releasing capacity of the blood. The pao2/fiao2 quotient was used as an index of the severity of lung injury (SLI). The endotoxin plasma level was significantly higher in patients with SLI < 200 (0.288 ± 0.038 EU/ml, mean values ± SEM) versus the patients with a SLI > 200 (0.175 ± 0.023 EU/ml, p<O.01). The basal plasma concentration of ll-6 also correlates with the SLI index (p<O.O01). The delta TNFe and delta IL-6 values were s~gnificantly decreased in patients with a severe lung injury. Compared with the survivors (n:11), the deceased patients (n=12) showed higher endotoxin level, a reduced TNFa and IL-6 releasing capacity of the blood and higher basal IL-6 levels. We can conclude that endotoxemia, high IL-6 plasma level and a decreased capacity of the blood to release TNFa an IL-6 under endotoxin stimulation associates with the severity of lung injury. [Objectives] Experimental and clinical findings implicate the involvement of inflammatory cytokines in the pathogenesis of the decreased oxygenation in the lung after major surgery, though the mechanism remains unclear. In the present study, we elucidated the involvement of IL-8 in the pathogenesis of lung injury defined by deterioration of pulmonary oxygenation after esophagectomy.

[Materials and Methods] Seventeen patients underwent subtotal esophagectomy through a right thoracotomy. In all patients, tbe alveolar-arterial oxygen tension difference (AaDO2) was measured everyday and bronchoalveolar lavage fluid (BALF) samples were obtained from a single segment ($4 or $5) of the right and left lung through a bronchoscope on days l, 3, and 5 days after surgery. They were then examined for cell analysis, measurement of cytokines by ELISA, and measurement of neutrnphil elastase (NE) by ELISA.

[Results] AaDO2 unexceptionally deteriorated postoperatively and the mean AaDO2 in 17 patients reached the maximum (mean ± SEM; 372 _+ 23 mmHg) on the 3rd postoperative day. The mean concentrations of IL-8, NE and neutrophil count in BALF also increased postoperatively and reached their maximum level (2137 + 618 pg/ml, 1603 -+ 493 gg/l, and 1561 ± 477 x 103 cells/ml, respectively) on the 3rd postoperative day. These increases in BALF were recognized in both the right and left lung. A highly significant correlation was observed between IL-8 and NE levels in BALF (r=0.82, p=0.0001). Also, significant correlations between AaDO2 and the neutropbil count, and the concentration of NE or 1L-8 were observed. In contrast to IL-8, neither IL-6, IL-113 nor TNFa was detected in BALF.

[Conclusion] Major surgical trauma such as esophagectomy may induce recruitment of neutrophils to the lung and cause lung injury. IL-8 increased in the lung is an important mediator of neutrophil recruitment and activation in the lung. Phorbol myristate acetate (PMA) is commonly used to cause edema and other tissue damages in the lung. Lung injuries induced by the administration of PMA has been shown to be mediated mainly by neutrophils. Neutrophils recruited to the lower respiratory tract may damage lung tissues by releasing reactive oxygen species, neutral proteases and lysosomal enzymes. The present study was conducted to investigate whether c~tocopherol, entrapped in dipalmitoylphosphatidylcholine liposomes and delivered directly to the lung, could counteract some of the PMA-induced lung injuries. Plain liposomes or c~tocopherol-containing liposomes (8 mg c~-tocopherol/kgbody wt.) were instilled into the lungs of rats 24 h prior to PMA exposure (25 pg/kg, intratracheally) and treated rats were killed 3 h later. Lungs of control animals exposed to PMA developed increases in lung weight and lipid peroxidation as well as decreases in lung angiotensin converting enzyme (ACE) and alkaline phosphatase (AKP) activities. PMA treatment also caused an increase in myeloperoxidase (MPO) activity in the lung, suggestive of neutrophi] infiltration. Pretreatment of PMA-treated rats with plain liposomes had no effect on PMA-induced injuries. In contrast, pretreatment of rats with liposomal c~-tocopherol, 24 h prior to PMA administration, resulted in a significant elevation of pulmonary a-tocopherol concentration, accompanied by a concomitant reduction in MPO activity and reversal of PMAinduced changes in lung edema, lipid peroxidation, ACE and AKP activities.

These results appear to demonstrate that the intratracheal administration of a liposome-associated lipophilic antioxidant, such as a-tocopherol, can significantly ameliorate the toxic effects of reactive oxygen species, released from PMAstimulated pulmonary target cells and infiltrating neutrophils.

JK Wojcik, G Palasiewicz, W Roszkowski Immunology Department,Institute of Tuberculosis and Lung Diseases~ Warszawa, Poland, M~larhospital, Eskilstuna, Sweden. Introduction:The critical point in neutrophil migration to the alveolar space in the course of ARDS is the attachment between giyeoproteins of neotrophil membranes(sialyl Lewis x carbohydrate epitope containing u-l-fucose) and lectin domains of endothelial P and E selectins (GMP-140,ELAM-I).

A potential beneficial therapeutic strategy may be designeP to suppress neutrophil recruitment to the lungs by preventing their rolling and attachment to the pulmonary endothelial cells. ObjecLive:To investigate if tz-l-fueose prevent experiments pulmonary hypertension in ARDS as effective as metylpredniselone. Methods:Only healthy rabbits (N-24) weighing 5.D-5.5 kg OaselinePaO>11 kPa and mean pulmonary arterial pressure (MPAP)<I.5 kPa were included in this study. Anaesthesia was induced with Thiopental 20-25 mg/kg bw and "blind" Jntubation was performed. A 5F Swan Ganz catheter was introduced via the left jugular vein into the pulmonary artery. PMA, as was used in our experiments activates neutrophils and produces acute respiratory failure with pulmonary hypertension and pulmonary edema. Eight rabbits serving as control animals received PMA 25 ug/kg bw iv only. Sixteen other animals were divided into two groups receiving either ~-l-fucose 100 mg iv or metylprednisolone 150 mg iv 15 min before PMA admininstration. Results:After PMA administration notable physiological changes including marked increase in MPAP(2.0-+O.~6 kPe3 ..lere found. The increase of NPAP was observed to be completely blocked in both groups of animals which received ~-l-fucose or metylprednisolone. Conclusion:In aur rabbit model ARDS CZ-l-fucose pretreatment prevents pulmonary hypertension after PMA administration. The effect is comparable with high dose metylprednisolone. The possible explanation is that G-l-fueose molecules block iectin domains of GNP-140 or ELAM-I preventing the adhesion of the neutrophils to the pulmonary endotheliom.

Bartens C, Elmadfa I, Steltzer H, Fischer M, Druml W Introduction: Various micronutrients and vitamins are particulary effective in modulating immune functions and host defense. The nutritive ant±oxidants vitamin C, E and B-carotene, as well as selenium lower the burden of reactive free radicals and protect the structural integrity of cells and tissues of the immune system, as well as other systems in the body. Certain cells of the immune system like polymorphonuclear leucocytes (PMN's) use free radicals as a weapon to destroy invading pathogens. These reactive molecules, when released into the surrounding area, mediate lipid peroxidation and tissue injury. There is growing evidence that PMN's are activated in vivo by complement or immune complexes in disorders such as ARDS. The respiratory host defense mechanisms of patients with adult respiratory distress syndrome (ARDS) may be defective. The aim of this study was ta evaluate the status of the free radical scavengers and their activity in ARDS and to investigate the respiratory burst of ARDS patients. Methods: Seven ARDS patients (3 sepsis, 4 trauma) with an Fie 2 of 0.59+0.(36, a PEEP of 8.50±0.99, and a Murray score of 2.83 ±0.18 were included in this study. 15 healthy adults served as controls. Superoxide anion production in granulocytes was measured photometrically, and the hydrogen peroxides by fluorimetric assay. Vitamin A, E, and g-carotene concentrations were assessed by HPLC, and plasma vitamin C photometrically. Plasma selenium was determined by dectrothermal AAS. Plasma lipid peroxidation pruducts, expressed as malondialdehyde (MDA), were determined by thiobarbituric acid assay and HPLC. Results: MDA-plasma (/xmol/l) 0.74~: 0.46* 0.37 ± 0.16 *= <0.01, **=p<0.001 Condusion: ARDS patients showed significantly decreased plasma concentrations of vitamin C, E and g-carotene, as well as selenium. These nutritive ant±oxidants are consumed during increased oxidative stress. MDA, a final product of lipid peroxidation, is increased. However, a difference in rates of release of hydrogen peroxides and superoxide-anion of PMN could not be detected. Therefore, oxygen radical accumulation is more likely a result of excessive inspiratory oxygen concentrations (F.iO2), pro-oxidant drugs, and a high settlement of PMN in the lungs, and not an increased release of free radicals of the single PMN.

kD glycoprotein secreted by the alveolar type II cells. Recent study showed that SP-A exists in the sera from normal healthy adults and that the significantly elevated serum SP-A levels were observed in the patients with interstitial pneumonia and pulmonary alveolar proteinosis. These findings means there may be a mechanism that the SP-A produced in the alveoli escapes into the bloodstream. Here we examined the serum SP-A levels in the patients with critical illness including sepsis and ARDS. ]Clinical Subjects & Methods] A total of 99 serum samples were collected from 39 patients hospitalized in the Division of Critical Care Medicine(ICU), Sappom Medical College Hospital. Serum SP-A levels were measured by an enzyme-linked immunosorbent assay using monoclonal antibodies to human SP-A.

]Results] There was no significant difference between the septic (n=12) and non-septic (n=27) patients. Patients with ARDS (n=10, 52.0+4.85ng/ml) and with respiratory disease other than ARDS (n=12, 49.6+-5.35ng/ml) showed significantly higher levels of serum SP-A than those with non-respiratory disease (n=17, 27.4_+4.86ng/ml). As to the ARDS patients, it was likely that the serum SP-A levels were getting higher in their clinical courses. However, there was no significant correlation between the serum SP-A levels and the Pa%/FiO 2 ratio. ]Discussion] At the present time, the exact mechanism of the escape of SP-A into the bloodstream remain to be unclear. Our results may suggested that there was some relation between this leakage of SP-A and the alveolar-capillary permeability because of the higher SP-A levels observed in the ARDS patients. However it is also likely that higher serum SP-A level represents the proliferation of alveolar type II cells as the regeneration of damaged lung. Further studies are now being done. Neutrophils contain a number of enzymes within intracellular granules,potentially damagingto lung tissue.Release of these enzymes into the lung tissue from the sequestred activated neutrophils is felt to play significant role in lung injury in the course of AROS.

Using 6-hours acute animal model of AROS the activity of cathepsins,beta-glucuronidase and acid phosphatase was determined in the lung tissue. Results of this research were compared with neutrophil proteases activity in serum and broncho-alveolar lavage.Also hemodynamic,respiratory and biochemical investigations were performed before beginning of experiment,and in a two-hours time intervals of its duration.After 6 hours was determined total and free activity of cathepsins,beta-glucuronidase and acid phosphatase in full homogenate of the lung tissue,mitochondriallysosomal fraction and the final supernatant.

In the course of our experiment we observed an increase (30.i-49.6%) both total and free activity of neutrophil proteases in all fractions of the lung tissue.Generally accepted current pharmacological treatment of AROS only attenuated this increase but never caused reversion to the level before beginning of our investigations.

The presence of AROS was histologically proved.An activity of acid phosphatase in serum and the lung tissue fractions was compared with histochemical pictures of the lung.

Results of our research indicate that neutrophil proteases are a very important mediators in AROS and they are a cause of the lung injury. In addition neutrophil-derived proteases can amplify the inflamatory process by enzymatitally activating of many other mediators and they also may to increase an oxidant induced injury by imparing of antioxidant defenses.

Oepartment of General Surgery, Sk~odowskiej 24a 15-276 ~ia~ystok, Poland, tel/fax +48 85 619742.

In vitro studies demonstrated that PAF activates polymorplionuclear leukocyte (PMN) 5-1ipoxygenase, but the generation of leukotrienes (LTs) is critically dependent on exogenous arachidunic acid (AA) disposal (F. Grimminger et at., Mol. Pharmacol. 41:757, 1992) . We investigated the features of PAF-evoked LT synthesis in a model of pulmonary microvascular leukostasis, in the presence and absence of intravascular n-3-and n-6 -prescursor fatty acid supply. Human neutrophils were infused into isolated, ventilated and perfused rabbit lungs to achieve microvascular sequestration of ligand-responsive leukecytes. Leukotrienes (LT) and hydroxyeicosatetra(penta)enoic acids (HET(P)E) were quantified by ItPLC techniques. Intravascular application of PAF (5 uM) or arachidonic acid (AA;IOuM) provoked the generation of limited quantities of 4-series LTs and 5-HETE (total sum of 5-1ipoxygenade products rd. 7 and rd. 27 pmol/ml.; both in PMN-preloaded and non-preloaded lungs). Combined administration of PAt r and AA caused amplification of 5-1ipoxygenase product formation with predominance of cysteinyl-LT synthesis both in lungs without (total sum rd. 67 pmol/ml) and, much more strikingly, with pro-application of neutrophils (total sum rd. 308 pmoFml). Eicosapentaeooic acid (10 uM) elicited exclusive generation of 5-series LTs and 5-HEPE (total sum rd. 82 pmol/mi). Dual stimulation with PAF and EPA provoked amplification of EPA-derived 5-1ipoxygenase product formation, again with predominance of cysteinyl-LTs, in lungs without (total sum rd. 224 pmul/ml) and in particular in those with preceding microvascular PMN entrapment (total sum rd. 545 pmol/ml). We conclude that the PAF-evoked 5-1ipoxygenase product formation in the neutrophil-harbouring lung capillary bed is critically dependent on intravascular precursor fatty acid supply, with EPA representing the preferred substrate as compared to AA. PMN-related transcellular eicosanoid synthesis is suggested to underly the predominant generation of cysteinyl-LTs. These findings may be relevant for lipid mediator profiles provoked by infusion of n-3-versus n-6enriched lipid emulsions in diseases with PMN-related inflammators events. Adult respiratory distress syndrome (ARDS) involves damage to the alveolar epithelium and microvascular endothelium. Products released from neutrophils (PMN) are thought to be among the chief causes of this damage, while the role of mononuclear cells (MNC) is uncertain.

We studied patients at risk of or with acute ARDS. We measured the concentration of myeloperoxidase (MPO) and elastase (EL) within circulating PMN as an index of PMN activation, elastin degradation products (EDP) in the plasma as an index of tissue damage, as well as the cytotoxici~y of PMN and MNC to endothelial cells (EC) in vitro.

Cells and plasma were prepared from venous blood of healthy controls; or systemic arterial blood of patients on ventilators in Intensive Care but not at risk of ARDS; at risk of ARDS because of sepsis; at risk because of multiple transfusion/major trauma; or with acute ARDS. Lung injury, multi-organ failure and APACHE H scores were recorded.

To measure cytotoxicity, human umbilical vein EC were labelled with Cr-51, incubated with either PMN or MNC for 18 hours, and the supernatant counted. The adherence of the PMN and the MNC to the EC was also measured. Standard assays were used to measure MPO and EL in the PMN, mad EDP in the plasma.

The MNC cytotoxicity did not differ between the groups. The PMN cytotoxicity was higher in the transfusion/trauma patients compared to each other group, which did not differ from each other. There was no correlation between cytotoxicity and adherence for any group.

The MPO and the EL concentrations were significantly decreased in the PMN from each of the patient groups, including the patient controls, compared to healthy controls. The EDP were only increased in patients at risk of or with ARDS. Thus PMN degrannlation occurred in patients not at risk of ARDS, but tissue damage as measured by increased EDP only occurred in patients with or at risk of ARDS, suggesting that factors additional to PMN activation and degranulation are required for the progression to ARDS.

There was no correlation between the parameters measured and the indices of disease activity, or outcome. These measures of PMN and MNC activation and function are not of value as prognostic indicators in patients with or at risk of ARDS. 

The liver is made up of several different cell types which subserve distinct functions in vivo. In addition to the different functions of the distinct cell types, it is now evident that even within the population of hepatocytes substantial functional heterogeneity exists. This heterogeneity occurs in a very organized fashion in the normal liver based upon the position of the hepatocyte in the acinus. Although it has been proposed that the acinar heterogeneity of hepatocytes arises from migration of immature hepatocytes from the periportal region to the perivenous region where they become senescent and die, many of the heterogeneous responses can be accutely reversed by reversal of the oxygen gradient via retrograde perfusion. Under normal conditions in which the acinus is exposed to unidirectional flow, the periportal hepatocytes are exposed to relatively high levels of oxygen, substrates and hormones. As these substances are extracted along the aeinus their concentration decreases leaving relatively low concentrations for the perivenous hepatocytes. In the normal liver, the heterogeneity of response is attenuated by extensive communication via gap junctions. The major liver gap junctions are made up of hexamers of connexin 32 (Cx32) which forms channels between cells capable of pas.~ing any molecule smaller than 1000 D. Coupling is sufficient to allow diffusion of molecules such as ATP or cAMP from one hepatocyte into > 20 adjacent hepatocytes within 60 seconds. During stress conditions such as endotoxemia or zymozan inflammation, expression of Cx32 is markedly decreased while the secondary gap junction protein Cx26 is either unchanged or even increased. While Cx 26 readily effects electrical coupling, molecules > 350 D pass only very slowly. This would result in restriciton of transmission of moecules the size of ATP or cAMP. Since inhibition of gap junctions also attentuates metabolic response to hormone or nerve stimulation, it is evident that modulation of hepatocyte hetereogeneity by gap junctions must be considered in determining the mechanisms of metabolic alterations during stress. 

Already minor haemorrhage decreases portal venous blood supply to the fiver and the reduction in portal blood flow becomes more pronounced with more profound btood loss.

Severe hacmorrhagic hypovolemia also reduces hepatic arterial blood supply which, however, is maintained over a vide range of haemorthage. The net effect of blood loss is a reduction in liver oxygee supply and this reduction is in proportion to the vulume Iossed. However, oxygen supply to the liver exceeds the demands of the normal liver and this is the ca~ stilt following reduction of 30 % of blood volume.

The situation in sepsis is more complicated. Po~l venous supply to the liver is redur.~i fairly early following normovolemic sepsis while hepatic arterial blood supply is maintained at le,~t initialIy, Oxygen saturation might be maintained in arterial blood but may also be slightly reduced during sepsis, Oxygen saturation of portal venous blood is significantly reduced during sepsis due to increased extraction of the intestines. Therefore oxygea delivery to the liver during sepsis becomes sigalfkzntly reduced. At the s,~ne time and for mai.v.ly unknown reasons the need for oxygen becomes significantly increased in the ~-~ptic liver. As a consequence liver oxygen consumption becomes flow dependent and the liver is likely to suffer from ischemia during septic conditions. $94

Although liver failure is well recognized in sepsis, it is generally thought to be a late complication following pulmonary and renal failure. Jaundice, hypoglycemia, encephalopathy and bleeding secondary to low levels of liver-synthesizing clotting factors are, however, signs of rather severe end-stage hepatic failure. Furthermore, elevated liver enzymes (SGOT and SGPT) represent hepatucyte damage and not hepatocellular dysfunction. In view of this, a more sensitive indicator of hepatic function is desirable in order to detect early hepatic abnormality. In this respect, indocyanine green (ICG) is a tricarbocyanine dye that possesses several properties which makes it particularly valuable inthe assessment ofhepatic function. This dye is bound m albumin and is cleared exclusively by the liver through an energydependent membrane transport process and is nontoxic at lower doses. We propose that maximal velocity (Vm~,) of ICG clearance is a valuable measure of active hepatocellular function, since the total concentration of functioning receptors is directly proportional to Vm~. We have utilized a fiber optic catheter and an in vivo hemoreflectometar to continuously measure the administered ICG in vivo and consequently determine its clearance without the need of blood sampling. Using this technique, we have found that in the early stages of sepsis (i.e., 2 and 5 h following cecal ligation and puncture), the Vm~ and kinetic constant (K=) of ICG clearance was significantly depressed. It should be noted that at this stage of sepsis, there was no elevation in serum enzyme levels. Furthermore, hepatic blood flow and cardiac output increased at the above mentioned time points. Thus, the extremely early depression in active hepatocellular function in sepsis, despite the increased hepatic blood flow and cardiac output, may form the basis for cellular dysfunctions leading to multiple organ failure during sepsis. Additional studies indicated that following hemorrhage, active hepatocellular function was markedly depressed. This returned to prehemorrhage levels after Ringers lactate resuscitation, however, this function was not maintained and decreased significantly after fluid resuscitation. Nevertheless, the depressed active hepatocelinlar function following hemorrhage was markedly improved by post-treatment of animals with either ATP-MgCI2, peutoxifylline or diltiazem. Thus, the use of ICG clearance provides an early sensitive indicator of hepatic abnormality during sepsis and following hemorrhage and this method should be used, not only experimentally, but also in the clinical arena for the early detection of hepatocellular abnormality. Although multiple organ dysfunction syndrome (MODS) remains a major cause of mortality and morbidity in intensive care units, very little is known about the mechanisms that precipitate its development. Since an episode of inadequate tissue oxygenation is considered to be the trigger for MODS, we have proposed that a primary localized injury such as ischemia/reperfusion may be sufficient to cause a change of gene expression of remote and apparently unaffected organs. Such modulation of remote organ gene expression may decrease the organ's tolerance to a subsequent stress contributing to the development of MOFS.

To test this hypothesis, rats were subjected to hepatic regional ischemia by clamping the blood flow (hepatic artery and portal venous inflow) of the left and median liver lobes. Intestinal congestion was prevented by allowing flow through the smaller right and caudate lobes. After 60 minutes of ischemia, the clamp was removed and the blood flow restored. The animals were allowed to recover for 1, 4 and 24 hours. Kidneys were removed, total RNA was isolated and poly(A) ÷ selected by affinity chromatography on oligo(dT) columns. Message was in vitro translated using rabbit reticulocyte Iysates in the presence of radioactive amino acids. The gene products (radiolabeled polypeptides) were fractionated by two dimensional gel electrophoresis, and visualized by fluorography.

Analyses of the two dimensional fluorograms indicate that there is a dramatic change in the electrophoretic pattern of in vitro translated products in samples corresponding to kidneys obtained after 60 minutes of hepatic ischemia and 24 hours of reperfusion with respect to kidney samples obtained after sham operation or from control rats. The latter were not subjected to any surgical manipulation. These studies suggest that the gene expression of the kidneys is specifically modified after a remote organ injury (hepatic ischemia/reperfusion). We speculate that this change of gene expression in kidneys after an indirect injury may be part of the early events leading to the development of MODS. A priming event, e.g. local ischemia, in combination with a second insult, e.g. sepsis, may amplify a host's response and lead to multiple organ failure. To better understand the mechanisms involved in the pathophysiology, male Fischer rats were subjected to 20 min of hepatic ischemia followed by reperfusion (RP) and injection of 0.5 mg/kg Salmonella enteritidis endotoxin (ET) at 30 min of RP. ET injection potentiated the postischemic liver injury as indicated by histopathology and an increase of plasma ALT activities from 870+122 U/L (I/RP only) to 3900+470 U/L at 4h RP. Inhibition of Kupffer cells (KC) with gadolinium chloride (7 mg/kg) attenuated liver injury in this model by 70%, however, monoclonal antibodies (CL26, WT3) directed against adhesion molecules (132 integrins, CD18) on neutrophils had no effect on the injury despite the substantial accumulation of neutrophils in the liver at that time (467+52 PMNs/50 HPF; baseline: 13+3). Isolation of KC and neutrophils from the postischemic liver indicated a 10-fold increase of the spontaneous superoxide formation only in the KC fractions [3.85+0.68 nmol O2-/h/10%elts (KC2); 6.82_+0.92 (KCa) ] at 4h RP compared to control cells. In addition, stimulation with phorbol ester or opsonized zymosan revealed a substantial priming of KC for reactive oxygen formation. In contrast to the short-term experiments (4h), the antibody WT3 (4 mg/kg) attenuated liver injury by 90% at 24 h of RP and improved survival. Conclusion: Liver injury during the early RP phase is mediated mainly by KC generating excessive amounts of reactive oxygen while neutrophils are primarily responsible for organ damage during the later RP period. (ES-06091 and GM-42957) 

Tumor necrosis factors (TNF) are cytokines which are cytotoxic towards some tumors in vivo and certain tumor lines in vitro. Moreover, these polypeptides are powerful immunomodulators and have been found to be distal mediators in several models of septic shock and septic organ failure. One of the best-characterized experimental systems is the hepatitis caused by LPS or TNF in galactosamine (GalN)-sensitized mice.

Here we describe a cell culture system, in which the direct toxicity of TNF towards mouse hepatocytes was examined. The toxicity of TNF, as determined by LDH-release or formazan-formation, was dose-and time-dependent. The threshold of toxicity was 10 ng/ml, which corresponds to serum concentrations found in mice after LPSinjection. Toxicity was only observed in hepatocytes sensitized with transcriptional inhibiters such as GalN, actinomycin D (ActD) or cxamanitin. Sensitization was neither observed with different translational inhibitors nor with various other metabolic inlaibitors or toxins. Inhibitors of protein synthesis or protein processing such as cycloheximide, puromycin, tunicamycin and ricin protected ActDsensitized hepatocytes from TNF-induced cytotoxicity. TNF induced apoptotic changes and DNA-fragmentation in sensitized hepatocytes which is in line with the above findings that cell death is dependent on protein synthesis. Thus TNF may be a trigger of programmed cell death during inflammatory organ damage. With the purpose of studying the role of complement activation in tissue injury after ischaemia and reperfusion we blocked the complement cascade in a model of rat liver isehaemia and reperfusion, either by administration of Soluble Human Complement Receptor Type 1 (sCRI), 25 mg/Kg IV after vascular occlusion (n=8) or by depleting the complement system using Cobra Venom Factor (CVF), 0.5 mg IM, 24 and 6 hours before ischaemia (n=10). Non-ischaemic rats (n=8) and ischaemic non-treated rats (n=15) were used as controls. The experimental procedure consists of the temporary interruption of arterial and portal blood flow to the left lateral and medial lobes of the liver during 45 minutes, followed by reperfusion, recording the liver blood flow and haemoglobin saturation with a Laser Doppler Flowmeter and Photometer during one hour after declamping; ALT levels were assayed and immunoperoxidase stainings for C3 and C9 were performed. There were statistically significant differences between the experimental ~roups and the untreated ischaemic control group in terms of post-isehaemic blood flow (p<0.001) and haemoglobin saturation (p<0.001). C3 and C9 were present in the endothelium of the ischaemic control group. No deposits of C3 or C9 were found in the CVF group. Few C3 and no C9 were found in sCRI treated rats. These results show that the effect of reperfusion injury in the rat liver is ameliorated either by depleting complement with CVF or by regulating complement activation with sCRI. Hepatic dysfunction, a major cause of mortality following hemorrhagic shock, has not yet been well characterized. The present study was designed to assess the effects of liver blood flow and cytokine levels on hepatic function following resuscitation from severe hemorrhagic shock in normal and cin-hotic rats. METHODS: Aftor pentobarbltal anesthesia, 23 control and 38 cirrhotic Sprague-Dawley rats were subjected to severe hemorrhage to reduce their systolic blood pressure to 50 + 5 mm Hg. This level of hypotension was maintained until the skeletal muscle transmembrane potential (Era) depolarized by 25%.; the animals were then resuscitated with Ringer's Lactate solution in three times the volume of the shed blood. Serial blood samples for tumor necrosis factor (TNF) determination (a modified flow-cytomeUic WEHI cell bioassay) were obtained at baseline, during hemorrhage and following resuscitation. Liver blood flow measurements by low dose galactose clearance (GLC) and functional bepatocyte mass (FHM; defared as galactose elimination capacity [GEC] from the zero order portion of the plasma disappearance curve following an intravenous galactose bolus [500 mg/kg], divided by liver weight) were measured before shock and after resuscitation. RESULTS: Higher survival rates (p < .05) were observed in control as compared with cirrhotic rats. Shock produced a significant reduction in GEC (to < 0.01); FHM (19 < .01); and liver blood flow (p < 0.01) in normal and cirrhotic rats. Decreases in GEC and FI-IM were greater (p < 0.05) in cirrhotic rots. TNF levels were higher (p < 0.05) in cirrhotic rats at baseline and during induction of shock.

Pre Gap junctions provide pathways for metabolic signals between cells. In the liver, the majority of gap junctions are composed of connexin 32 (Cx32) polypeptide subunits, and are regulated by gluconeogenic hormones. Since sepsis and other inflammatory states alter hepatic glucoregulatory control, we have evaluated the contribution of gap junctional conductance to the metabolic dysregulation in the liver. An acute inflammation was induced in rats by injection with E. coli endotoxin (LPS lmg/kg). Northern blot/hybridization analysis of total RNA isolated from livers after endotoxin injection show a decrease in the steady state transcript levels of Cx32 to 18% of sham controls. Immunostaining of liver sections using anti-Cx32 revealed punctate fluorescent staining on the plasma membrane at regions of call-cell contact in saline injected animals, whereas, staining was only observed in cytoplasmic vesicles 6 hrs after animals were treated with LPS, suggesting the internalization of Cx32 without replacement on the cell surface. The staining was quantitated and expressed as % of pixels above threshold. At 6hr post injection 3.6% ofpixels exceeded threshold, compared to 16.2% in sham controls. Functional gap junctional communication was assessed by dye coupling using lucifer yellow in an isolated perfused liver under intravital fluorescence microscopy. Dye diffusion was markedly decreased 6hr after endotoxin injection. This suggests that decreased metabolic coupling after LPS injection results from decreased gap junction abundance. The present data suggest that metabolic dysregulation during sepsis may arise in part from changes in intercellular communication caused by a decrease in gap junctional expression and communication. Given the marked metabolic heterogeneity of hepatocytes with respect to acinar location, metabolic signaling via gap junctions most likely serves to moderate this heterogeneity, contributing to a coordinated metabolic response. Altered cellular Ca 2÷ regulation might be a critical step in organ dysfunction during sepsis and ischemia/reperfusion events. The aim of the present study was to evaluate hepato-ceUular Ca 2÷ regulation in isehemiah'eperfusion after hemorrhage and to assess effectiveness of TNFc~-monoclonal antibody (TNFo~-MoAB). Methods. Male Sprague-Dawley rats (250g, n>_6/group; pentobarbital 50 mg/kg) with hemorrhage for 60 rain at 40 mm Hg. Reperfusion by Ringer's lactate (2 x maximal bleed out/60 min) and 60% of citrated shed blood. TNFcz-MoAB (TN3, CeUtech, 20 mg/kg in 0.9% NaC1) infused during fLrst 5 min of reperfusion. At baseline, end of ischemia and 60 min of reperfusion, hepatecyte isolation by liver collagenase perfusion.

" 45 Hepatocyte incubation (30 mg w.w./ml) with CaCI 2 (0.05 2+ 2+ 2+ MBq/ml) for 60 rain (Ca influx [slope, 1/mini; Ca uptake [nmol Ca /mg protein]) w/ and w/o epinephrine (EPI, 100 nM). Hepatecyte resuspension in radioisotope-free medium and farther incubation (exchangeable Ca 2+ (Ca2+ex) [nmol Ca2+/mg protein]; Ca 2+ membrane flux [nmol Ca2+/mg protein'min]). During incubation, aliquots (100 ~tl) were centrifuged through oil/lanthanum gradient and acivity measured by scintillation counting. Statistics: ANOVA. Mean + SEM. Results. Hepatocyte Ca2+ex and membrane Ca 2+ flux were significantly increased at both, the end of ischemia (14.3+.9; 0.41+.05) and reperfusion (16.4+.9; 0.50+.2), as compared to sham-operated animals (7.6+_.5; 0.09+.02)(19<.05 ). TNFc~-MoAB treatment significantly prevented reperfusion-induced increase of Ca2+ex (10+.8) and membrane Ca 2+ flux (0.17+.03)(p<.01). Fast Ca 2+ influx was significantly increased by epinephrine in hepatecytes from sham-operated rats (0.23+.03 vs. EPI: 0.52+.1, p< .05). This hormone effect was not observed in isehemia (0.47+.05, EPI: 0.6!-_.2) or reperfusion (untreated: 0.13+.06, EPI: 0.07+.01; TNFt~-MoAB: 0.26_+.04, EPI: 0.33+.07). Conclusion. The present study clearly demonstrated hepato-cellular Ca 2+ overload in ischemia and reperfusion as a result of hemorrhagic shock. Analysis of membrane Ca 2+ fluxes and hormone Ca 2+ mobilization suggests disturbances of membrane Ca 2+ transport mechanisms, e.g. through Ca2+-ATPases. Reperfusion-induced oxygen free radical generation which affect exchange kinetics of cellular Ca 2+ buffering compartments might also be operative. Prevention by TNFct-MoAB indicates the pivotal role of TNF as an early inflammatory mediator of hepatocellular alterations in signal transduetion mechanisms and cellular homeostasis. Although the precise mechanism has not yet been elucidated, bacterial translocation and endotoxin absorption have been frequently shown after burn, and have been postulated to be one of the underlying processes of sepsis. The purpose of the current study is to define the hemodynamic response of the liver to endotoxin release in burns, in correlation to bacterial translocation. Twelve female minipigs, weighing 20-28 Kg, underwent a laparotomy & transition time ultrasonic flow probes were positioned on the portal vein, the common hepatic artery, and the superior mesenteric artery. 6.5 Fr catheters were inserted in the superior mesenteric vein and the left hepatic vein. A jejunostomy was also performed. After five days all animals were anaesthetized and randomized to receive 40% of TBS A third degree burn. Eighteen hours after burn. 100 gg/Kg E. Coli LPS was intravenously administered over 30 rain. alI animals were studied for additional 24 hours and then sacrificed. Several recent data suggest that in severe injuries, such as shock state, the gradual activation of Kupffer cells and the excessive release of destructive and immunosuppresive products from macrophages may contribute to the development of "multiple organ failure". In in vivo experiments in mice, the effect of Kupffer cell phagocytosis blockade on the correlation between the tissue distribution of LPS, endotoxin sensitivity and LPS-induced TNF production was investigated. To depress the activity of the Kupffer cells, gadolinium chloride (GdC13) or carrageenan was used. Th~e studies indicate the dissociation of tissue localisation of Cr Jllabelled endotoxin and endotoxin lethalithy. Both GdC13 and carrageenan depressed Kupffer cell activity, but endotoxin sensitivity was enhanced only by carragenan treatment. However, there was a close correlation between the sensitivity to LPS and LPS-induced TNF production as measured in the serum, since LPSinduced TNF production was enhanced only by carrageenan treatment. On the other hand, GdC13 pretreatment significantly increased TNF production in the spleen. These results support our earlier findings that GdC13-indueed Kupffer cell phagocytosis blockade leads to activation of the spleen, and may explain some of the immunological effects of GdC13. Inositol(l, 4, 5) triphosphate (IP3) has been proposed as a second messenger for calcium mobilization. The addition of IP3 at low concentration has been shown to cause calcium release from intracellular microsomal store in rat hepatocytes. The effects of sepsis on the IP3 binding from microsomal fraction of rat hepatocytes during sepsis were investigated. Sepsis was induced by cecal ligation & puncture (CLP). Control rats were sham-operated. Three microsomal fractions (rough, intermediate and smooth) were isolated from rat liver. Study of IP3 receptor binding was performed with tridium label IP3. The results shewed that the IP3 binding was significantly depressed by 40-47% (p<0.05) during late sepsis (18 hrs after CLP), but not in early sepsis (9 hrs after CLP). The IP3 binding depression during late sepsis was most significant on rough and intermediate endoplasmic reticulum (p<0.05), but not on smooth subfraction. Since IP3 binding plays an important role in the regulation of intracellular calcium homeostasis in hepatocytes, an impairment in the calcium release due to depressed IP3 binding on smooth and intermediate endoplasmic reticulum during late sepsis may have a pathophysiological significance in contributing to the development of altered hepatic metabolism during septic shock.

Septic organ failure is currently recognized as an overactivation of the nonspecific immune system by bacterial stimuli giving rise to proinflammatory mediators. Little is known about the mechanisms of the resulting cellular injury. Here, a synergism is described between TNF as a major mediator of septic organ injury released by macrophages and hydrogen peroxide (H202) as a representative of reactive oxygen species as formed by e.g. neutrophils.

Rat hepatocytes are only slightly sensitive to either agent alone. When treated with a conbination of TNF and H#09 a stronq synergistic toxicity was found, especially w~e~ TNF-treatment preceeded challenge with H~O~. We have recently described a coculture model bfZrat liver macrophaqes and hepatocytes where LPS induces hepatocyte cell death partially mediated by macrophage TNF release. When H202 was also employed in fhis more complex cellular system a similar synergism was found: The ECc~ of LPS was 62 consecutive patients with liver cirrhosis admitted to the Department of Surgery over a 1 year period from January to December 1992 were studied for their complement profiles in relation to other parameters of liver function, The aim of the study was to determine if a direct correlation existed between low complement levels and end stage liver cirrhosis. Cirrhotic patients were divided into Child's A, B and C categories using Child's classification. Complement levels (C3, C4) were measured and functional assay for complement (CH50) were performed in each of these groupings in addition to normal blood donor controls. These results show that the qualitative C3, C4 and the functional CHS0 complement assays have good predictive values in assessing deteriorating liver function• in particular, the functional assay for complement (CH50) showed marked impairment in Child's C patients (p<0.000001) confirming the impaired immunological status of these patients. Sera from this group of patients (Child's C) were titrated with pig red blood cells (RBCs) in a haemolytic assay. The results showed that there were significantly less haemolysis of pig RBCs in these patients (p=0.0177) as compared to the controls. This findings strongly support an impaired immunological status in Child's C liver cirrhosis and may explain the high incidence of sepsis as a terminal event in these patients. Aim:Kupffer cells(KC) have an importamt play to cause hepatocellular injury in sepsis, because these cells release many kinds of substances. We reported that oxygem radicals released by KCs stimulated by lipopolysaccharide (LPS) caused hepatocellular injury. Aim of this study is to investigate the relationship between imtracellmlar calcium(Ca) concentration of cultured rat KCs stimulated by LPS and release of oxygen radicals, and effect of prostaglandin E~ (PGE~) on imtracellular Ca concentration. Production of acute phase proteins (C-reactive protein, CRP, transferrin, Tf) and £erritin (F) in rat hepatocytes (Hps) and its dependence on extracellular matrix components were studied. Hps isolated from the liver by collagenase perfusion were cultured at ~O5 per 0.5 ml medium FI2+DMEM (1:1) with 10% fetal calf serum for 2 days on uncoated or type I collagen coated plastic surface or in the presence of dextrane sulphate in the medium. Hps were stimulated by conditioned medium (GM) from I~IA-P or E. coli LPS preineubated human blood mononuclear cells. Production of CRP, Tf and F by Hps was detected by ELiSA. It was found that both CMs decreased Tf synthesis in Hps by 15-50% (p<O.OS).

The level of F synthesis didn't change or only slightly decreased. The addition of CMs led to the enhancement of CRP synthesis more than 2.5 times. The most reproducible results were obtained when plastics had been coated with collagen. Dextran sulphate markedly increased CRP synthesis. Thus using this model of an acute phase reaction in vitro we found the enhancement of CRP synthesis and the decrease of Tf synthesis. Our data correspond with nublished materials on the acute phase in vivo and point to relevance of the proposed model. Metabolism of hepatocytes under traumatic illness has not been sufficiently investigated. Therefore we have applied absorption and fluorescent cytophotometry techniques to study the content of RNA, glycogen and its fractions in hepatocytes of patients with severe mechanical trauma, both with and without endointoxication at various stages. For these purposes slides were prepared from the repeated aspiration biopsy material according to special techniques (Kudryavtseva et al., 1983) . Slides were stained by gallocyanin-chromalum to measure RNA or underwent fluorescent PAS-reaction to measure glycogen and its fractions (Kudryavtseva et al., 1970 (Kudryavtseva et al., ,1974 

The ability of 02 metabolites derived from the xanthine-xanthine oxidase system to inhibit mitochondrial function was examined using freshly isolated rat liver mitochondria. Under uncoupled conditions, mitochondria exposed to free radicals exhibited a significant decrease in 02 consumption supported by NAD+-Iinked substrates, but showed almost no change in 02 consumption in the presence of succinate and ascorbate. The activity of electron transfer complex I (NADH oxidase and NADH-cytochrome c oxidoreductase) and the energy-dependent reduction of NAD+ by succinate were unaltered by oxidative stress. Exposure to free radicals also had an uncoupling effect at all three coupling sites. The degree of mitochondrial swelling was closely correlated with the inhibition of state 3 oxidation of site I substrates and with the increase in state 4 oxidation of succinate. The immunosuppressive agent cyclosporin A (CyA) completely prevented the mitochondrial damage induced by oxygen free radicals (swelling, Ca2+ release, sucrose trapping, uncoupling, and selective inhibition of the mitochondrial respiration of site I substrates). The same protective effect was found when Ca2+ cycling was prevented, either by chelating Ca 2+ with EGTA or by inhibiting Ca2+ re-uptake with ruthenium red. These findings suggest that the deleterious effect of free radicals on mitochondria in the present experimental system was triggered by the CyA-sensitive and Ca2÷-dependent membrane transition, and not by direct impairment of the mitochondrial inner membrane enzymes.

Correspondence should be addressed to: Naoshi Takeyama The aim of this study was to clarify the critical role of reduced glutathione (GSH) on the progression of lipopolysaccharide (LPS)induced liver damage of mouse. The conditions of GSH-depletion and -richness were produced by intraperitoneal administration of buthionine sulfoximine (BSO), a specific inhibitor of GSH synthetase, and of GSH-monoethyl ester (GSH-ME), respectively. GSH-ME, which was esterified the caboxyl group of the glycine residue, is readily transported into cells and is hydrolyzed intracellularly; this leads to greatly increased the cytosolic and mitochondrial levels of GSH. BSO and GSH-ME were administered intraperitoneally 3 mmol/kg and 10 mmollkg, respectively, and LPS (2 mg/kg) given 1 h later. Hepatocytes were isolated by in situ perfusion of the liver with collagenase 6 h after LPS injection. The degree of hydrogen peroxide (H202) synthesis and mitochondrial membrane potential were measured by flow cytometry using fluorescence probe dichrolofluorescein diacetate and tetrametyl rhodamine ethyl ester, respectively. The degree of mitochondria membrane permeability transition (MMP) was assessed by [014]sucrose entrapment. We found that LPS treatment results in a decrease in hepatic GSH level and mitochondrial membrane potential, and an increase in H202 synthesis and MMP. LPS administration to BSQ-pretreated mouse worsened at[ these parameters. In contrast, GSH-ME pretreatment ameliorates. All together, GSH may acts an important role in cellular defence against LPS-mediated liver damage, and MMP may result in the decrease in mitochondrial membrane potential. It has been shown, up to now, that disturbances of enzymatic processes in the cell organella are the basic factor of the changes taking place during endotoxin shock It has been observed that lysosomes are biochemically changed as result of the effect of lipopotysacharides of gram negative endotoxic bacteria. The purpose of the experiment was: 1.evaluation nfthe course of the EES due to the biochemic changes 2.determination of the influence of the EES on the activity of lysosomal enzymes in liver cell 3.deterrmnatian of the influence of H and D on the activity of some lysosomal enzymes in liver cell during EES Material and method: Examinations ware carried out on 30 dogs. The shock was evoked by i.v. administration of endotoxin E.coli. The dogs were divided into 5 groups: Icontrols, I[ -dogs with untreated shock, III-dogs in shock treated with H, IVdogs in shock treated with D, V -dogs in shock treated with H and D. The following parameters were determined: 1.Activity of acid phosphatase in the venous blood serum. 2-6.Total and free activity of acid phosphatase and catepsins in the full liver cell homogenate, in mitochondrial-and-lysosomal fraction, and in the superuatunt (all after 4 hours of the experiment). Conclusions 1. Essential increase of activity of lysosomal enzymes was observed in the EES. 2. Increased activity of lysosomal hydrolases in the liver cell homogenate corresponds, as a result of the endotoxin effect, with increased enzymatic activity in the peripherial blood. Extensive investigations from our laboratory of the pathophysiology of hepatic no-flow ischemia (45 min) -reperfusion injury demonstrated substantial Kupffer cell activation with reactive oxygen formation during the first few hours of reperfusion as indicated by increases in plasma glutathione disulfide (GSSG) levels in vivo (Am. J. Physiol. 260: G355, 1991) and enhanced superoxide formation by Kupffer cells (KC) isolated from the postischemic livers (Free Radic. Res. Commun. 15:277, 1991) . The early KC-induced injury initiated the later, primarily neutrophilmediated reperfusion injury (FASEB J. 4: 3355, 1990 ). To test whether or not similar mechanisms are operative during hepatic ischemia induced by hemorrhagic shock, male Fischer rats (230-250 g) were subjected to 60 rain of hemorrhage (mean arterial pressure 40 mm Hg) followed by resuscitation (RS) (reinfusion of shed blood). Hepatic ATP levels declined from 2.3+0.2 txmol/g to 0.24_+0.08 (60 min shock) indicating severe ischemia, and recovered to 1.04--0.2 at 90 min RS. To test for potential oxidant stress during RS, plasma GSSG conc. were measured. Although GSSG levels increased by 120% compared to baseline levels (0.69!-_0.31 I.tM), there was no significant difference to shamoperated controls. Spontaneous superoxide formation of isolated KC was 0.32+0.05 nmol O2-/min/10ecells (KC2) and 0.51+0.11 (KC3) in controls and did not change significantly after shock and 90 rain RS. Addition of phorbol ester or opsonized zymosan to isolated KC did not indicate a priming of these cells. Conclusion: In striking contrast to our previous findings with hepatic no-flow ischemia, there is no evidence for a significant KC activation after 60 min of hemorrhagic shock and up to 90 rain of resuscitation. Objectives-This study investigates morphometric changes in KC nltrastmcture which might account for this diminution in phagacytosis.

Materials and methods -Three groups of Wistar rats were studied: Control, Sham-operated [Sham] and Bile duct ligated [EDL] . After 3 weeks animals were sacrificed and livers were "perfusion fixed" with 3% glntaraldehyde and processed for transmission electron microscopy and image analysis.

Results -In the the BDL group Kupffer cell numbers were greatly increased.

Biliary ductular proliferation was evident and sinusoidal spaces were compromised. Eleven nuclear and cytoplasmic parameters were measured and statistically significant results are summarised in the RESULTS: Ascorbic acid levels were slightly elevated after the HS and returned to basal values after 120 rain. Glutathione concentrations were increased significantly after the HS and the GSH levels kept rising continuously to as much as 4-fold of basal levels at the end of 120 min. MDA showed no significant variation before and after the hemorrhagic shock. Plasma concentrations of ratine, teucine, isoleucine, phenylalanine, proline, threonine and serine showed significant increase over basal levels during the whole ischemic period. Glutamic acid was slightly elevated at the onset of HS and remained the same for the rest of ischemic period. Hydroxyproline was significantly below the basal value at the mid-point of ischemic period. Alanine, glycine, histidine and aspartate did not change significantly throughout experimental time course. CONCLUSIONS: (1) Oxidative stress might not be severe in the systemic hemorrhagic shock for pigs since no significant variations were observed for ascorbic acid and MDA (2) the increase of GSH might be that it was released from hepatocytes when the animal was under systemic ischemia. (3) Celzain amino acids might be acting as transmitters in the event of ischemia. However, fm"lher investigations are needed to clarify the detailed mechanism in regard with antioxidants and amino acids.

Dr. Fang-Ku P'eng Taichung Veterans General Hospital Talchung, TAIWAN 407, R.O.C.

Jia Shi Yong, Huang Zhi Oiang and Men xian Jun. In patients underi~ent major hepatic surgery,obstructive jaundice has been implicated with fnoreased susceptibility to sepsis and liver failure. Jaundice was said to he associated with derangement of iamune function and result in expression of pr,:.inflar...aatory cytoki~es that cause hepatooellular damage. This study is ~.o investigate the effects ,.,t lip,~polysaccharide(LPS)rats. Jaundice were produced in healthy ~istar rats of either sex weighing 180-220 m~ by common hi l~ duet l igation(Bl)L).Seven days post l igation, rats were given LPS(236.gg.10)0.7rag/kg intraperitoneally. At the end ,:,f l,gand 5 hours after LP$ ehanllenge, liver were removed and prepared for forzen sections immediately. I)emonstration of INF in liver parenchyna were performed by method of ABU i~munohistochemistry using r~onoolonal antibody against TNF, Leve of mRNA for TNF ~ere measured by in-situ hybridization using biotin-labeled eDNA probe.

Results sho'~ed that TNF stained granules appeared in kup~'fer eell~ and polymorphonuclear leukocytes(PNNl,but not in hepatoeytes nor endothelial eel Is. They ~'ere located within cytoplasms and peaked at 3-5 hours after LPS ohal lenge. There was vocomitant TNFmRNA expression but peaked earlier 1 hour post LPS challenge. TNFmRNA vcere detected notably in necrotic area and barely in BDL rats ~ithout LPS challenge. It is therefore postulated that production of TNF by inflammatory effeetor cells-kupffer cell and PHN in site, in the obstructive jaundice rats during sepsis may atribute to the hepatocellular damage. It also provide evidence for the beneficial effects of early release of biliary obstruction. In the present study, an animal model of A0C in Wlstar rats was developed by tigatlng the common bite duct and injecting D.3mt solution of E. goli % B, (g×10' efm/mt) into the bite duct. The mortality rate was 40~ at 48h after AOC. At 6, 12, 24, 48h after AOC, KCs were isotated and cultured atone or cocuttured with hepatocytes to evaluate the modulation effect of Kgs on hepatoeyte protein synthesis.

The results showed that tactodehydrogenase leakage was increased progressively as the injured k~ function and structure developed. TNF and IL-I tn the supernatant were detected with the peak values at 12h after AOC. At 48h after AOC, ~-teueine incorporation was obviousty decreased in the normal hepatocytes eocuttured with stimulated KCs compared to the uormai hepatocyte cultured group (P<8.01), but it was slgnificantiy increased in the group cocuLtured with normal KCs. In the A0C group, 'H-teuclne ineorporatien was decreased progressively in the injured hepatoeytes cocuttured with stimuLatedKgs and it was markedly elevated when cocuttured wlthnormat KCs at 48h after A0C (P<0. Og).

It is concluded that the impaired hepatocyte proterin synthesis was directly mediated hy the stimulated KC function during AOC. Such a mechanism possibly may be involved in the patho0ensis of hepatic dysfunction and M0F following h06.

" The project supported by NSFC. In the past few years it has become evident that cytokines are implicated in various pathophysiological mechanisms. Therefore measurement of cytokines and their potential inhibitors in biological fluids may be helpful in diagnosing and monitoring of diseases. However, dependent on the type of assay used investigations performed in different laboratories have often yielded conflicting results. In order to assess reliability and reproducibility of cytokine measurements two comparative studies for the determination of cytokines in biological fluids were launched by several investigators interested in this field: one national Austrian ~tudy and one European study in the context of the European Workshop for Rheumatnlogy Research. Serum and synovial fluid samples were distributed by the organisers, and participants had to measure one or several of the following cytokines in all samples: IL-1, IL-2, 1L-6, I58, TNF-alpha, IFN-gamma, GM-CSF, and the soluble receptors for IL-2 and TNF; both commercially available immunoassays and home-made assays were allowed. So far, the main conclusions emerging from these preliminary studies are: (1) the same immurzoassay (ELISA) yielded comparable results in different laboratories; (2) immunoassays from different manufactureres usually measured the highest concentrations in the same samples, yielded similar relative values but disparate absolute values; (3) assays for soluble receptors yielded less discrepant results; (4) some assays seem to be more suitable for measuring cytokines in biological fluids than others. The mean retrieval of exogenous recombinant human cytokines was approximately 50% for most cytokines tested. However, it must be borne in mind that natural and recombinant cytokines may behave differently in immunoassays. This raises the question as to the necessity of setting up international standards for all cytokines. Possible interferences by serum components such as (lipo)proteins, complement, rheumatoid factors, etc., which may either decrease assay sensitivity or yield false positive results, must also be considered. In addition, natural cytokine binding proteins (e.g. soluble receptors) might interfere with cytokine determination in immunoassays. The problem of immunoassays versus bioassays has not yet been approached, but is without any doubt worth indepth investigation. Thus, these studies are a promising first approach to deal with the difficulties of cytokine analysis in biological samples, and it is hoped that they will help to overcome some of the major methodological problems in the near future. Tumor necrosis factor (TNF) is a pleiotropic cytokine which plays a key role in a number of immunologically mediated disorders like septicemia or cachexia. TNF receptors are found on the surface of a variety of cells, where they provide molecular signals for the cytokine effect. There exist two types of soluble TNF receptors (sTNF-Rs), TNF-R p55 and TNF-R p75. These sTNF-Rs can compete with the cell surface receptors and block their activity. The expression on and shedding from the cell surface of sTNF-Rs is enhanced by interferon gamma.

Increased concentrations of sTNF-Rs can be found in patients with infectious as well as malignant disorders. In patients undergoing immune stimulatory therapy with interleukin-2 and interferon alpha a simultaneous increase of TNF and its soluble receptors can be seen. In patients with HIV infection a strong correlation exists between the levels of sTNF-Rs and markers of immune activation like neopterin or beta-2-microglobulin. Likewise patients with hematological disorders show elevated sTNF-R levels. Patients with weight loss have higher concentrations than patients with stable weight.

The final value of sTNF-Rs within the complex network of cytokines is not yet clear. However, there exist striking parallels between infectious and malignant disorders pointing to a key role of endogenous immune activation. Biochemicatly, D-erythro-neopterin derives from guanosine triphosphate. In vitro studies show that large amounts of neopterin are produced by human monocytes/macrophages stimulated with interferon-j (IFN-J). Neopterin is biologically stable, and its halflife in the blood seems to solely depend on renal excretion. Specimen for neopterin determination can be stored without special precautions. Increased concentrations of neopterin have been described in body fluids of patients suffering from diseases which are apparently associated with an activated cellular immune response and with enhanced endogenous formation of IFN-~', e.g,, 1 ) increasing neopterin levels predict rejection episodes in allograft recipients 2) virus infections induce increased neopterin concentrations, and the degree of neopterin elevation predicts prognosis in these patients which is particularly true in HIV-1 infection 3) in autoimmune disorders such as systemic lupus erythematosus or rheumatoid arthritis neopterin levels reflect activity of the disease 4) increased neopterin concentrations in patients suffering from certain types of cancer indicate poor prognosis. Significant associations between changes of neopterin and decrease of hemoglobin as well as the development of weight loss and cachexia, e.g., in cancer patients and in patients with HIV-1 infection, support the concept that endogenously formed cytokines such as IFN-~" and tumor necrosis factor-~( are involved in the pathogenesis of such symptoms. In conclusion, neopterin monitoring is a sensitive tool to detect the activation status of the cell-mediated immune system in vivo. Repair and healing or perpetuation of acute inflammation is characterized by a complex network of interacting cellular and humoral defence mechanisms. Of the numerous inflammatory mediators/effectors investigated hitherto, the proteolydc lysosomal enzyme PMN elastase has turned out to be highly destructive when released extracellularly thus contributing considm:ably to organ dysfunctions in severe posttraumatic and postoperative courses.

Besides various cytokines, elastase in complex with its main antagonist c¢ 1-proteinase inhibitor (a 1PI) is also supposed to induce the shedding of intercellular adhesion molecules from inflammatory cells (PMNs, monocytes/macrophages, endothelial cells). These soluble adhesion molecules may modulate the inflammatory process in many different ways, e.g. via acting as chemotaxins, blocking neutrophil activation or competing with the membrane-bound forms in cell-cell adhesion.

Thus, measuring circulating or local levels of elastase-a 1PI and soluble adhesion molecules (ICAM, E-selectin,-Pselectin, L-selectin) may be a helpful tool in judging the severity of an acute inflammatory process. In several clinical studies on surgical patients suffering from multiple trauma and/or sepsis we could demonstrate that the measurement of these parameters in consecutive plasma samples and local body fluids was highly valuable for early diagnosis and prognosis of multiple organ failure.

Prof. Dr. M./ochum, Institut Fdr klinischeChemie und Biochemie, LMU Miinchen, Nufibaumstrage 20, 80336 M0nchen, Germany

Procalaitonin (ProCT) a 116 AA peptide is dramatically increased in the blood of patients with various sepsis and infections. The increase begins as soon as 3 hours after the andoto:edn release and readied maximal level with a 200 to 80B-fold increase 24 h after LI~ edrmrdstration. Prcc, alcitorfin response was temporally different from these of TNF~ and II-6, that reached peak levels at 2 h and 3 h respectively. At the opposite of eytokine (TNFcz, ILd, J, Is), ProCT is a very stable molecule. The half Iife of ProCT in ~he blood is surprising if we consider the half llfe (about nine mirtute) of mature ¢alcitonin (CO. We have found that it is due to the binding of ProCT with a protein on the N-termktal part. Thin complex of about 80 000 daltons is unable to cross the kidney and the C_N9 barriers and the half life is at least elghteen hours. This long half llfe is very useful in the evaluation of a sepsis. A very sensitive and specific sandwich in'ununoassay has been developped. The results cart be delivered in tWO hours.

At time, we know that in the healthy adults, the values are less thau 0.1ng/ml. It is also clear that ProCT is not increased ~n patients wlth inflammatory dleeasea without lnfectlon. For instance, ProCT is not signLffcat3vely increased in purpura rhumatoM, arthritis, Crohn's disease, rectocolitis, also in various cancer patients with inflammatory components. Am other interesting finding, is the absence of increase or a very low increase in the viral infections. At the opp0site, in all the patients with bacterial sepsis or in the malaria pat/ants, ProGT was dramatically increased, from I00 to 10 000 fold. P~oCT can be useful to the follow up of patients with sepsis. Actually, we don't know the origin of this ProCT production, in the United States, will also be discussed.

The SepTest portion of this study will eventually enroll over 500 patients tentatively diagnosed as septic. The results of SepTest will be compared with commonly accepted symptoms and criteria associated with sepsis in order to determine the clinical utility of this endotoxin assay. Preclinical results of patients and normals be presented as well as data on the time course of endotoxemia and clinical outcome of selected septic patients.

Preparation of DNA Libraries Clifford W. Schweinfest, Ph.D.

A fundamental tool of the molecular biologist today is the ability to work with purified DNAs representing genes of interest with respect to the investigator's field of study. Originally applied almost 20 years ago to the genetic dissection of simple organisms such as viruses and bacteria, molecfllar cloning was rapidly applied to study the more complex genetics of the mammalian cell. Today, virtually any study of cell physiology which depends upon specific gene expression is approachable using the tools of the molecular biologist. Therefore, stress induced gene expression (such as the heat shock gene, hsp70) or the regulation of the nitric oxide synthetase gene or the induction of cytokines and growth factors as a result of trauma or injury lend themselves to investigation at the level of the gene.

During these workshops, we will discuss the means by which cDNAs and genes are isolated, amplified and identified. We will discuss the strategies and methods for cloning cDNAs and genomic DNAs, for organizing such cloned libraries, for finding specific genes and for characterizing those genes. This speaker will focus on the techniques and considerations involved in the synthesis of cDNA libraries. Topics include RNA isolation, mRNA quality, reverse transcription, choice of cloning vectors and library quality. Genomic DNA library synthesis will also be discussed with respect to the vector/host systems available and applications which are specific to genomic DNA. Finally, polymerase chain reaction (PCR) will be discussed briefly with regard to its impact on DNA cloning. The identification of the genes or gene products that have a role in cellular processes is fundamental to our understanding of genetic control and methodologies to achieve this aim are currently available. All levels of the various signal transduction pathways can b~ molecularly analyzed to define the mechanism by which critical steps in each pathway are achieved.

The questions that are unresolved in each area of investigation serve to direct the scientist towards analyses of gene products or gene regulation of the gene itself. Thus, selection of the type of library (genomic or cDNA) to be used is dependent upon the specific goals of each investigator. To facilitate resolution of this question, an overview of the uses for the different types of libraries will be presented.

At least four methodologies are currently available for screening a library, which are dependent upon specific interaction between nucleic acids, nucleic acids and proteins, antibodies and antigens or between different proteins. The use and advantages of each of these methodologies will be discussed along with a description of probe preparation.

Sequence methodologies required to begin characterization of gene clones will be presented.

Hopefully, the participants in the workshop will help define areas of emphasis and allow time for directed interaction to discuss specific applications based upon their own experimental systems. 

Alterations of physiological conditions, such as those occurring after shock and sepsis, induce changes in gene expression of cells composing the major organ systems. The challenge is to detect and characterize these changes in geae expression and relate them to the injury. The development of cloning techniques has emerged as the methodology of choice. Changes in gene expression are usually characterized by variations in the concentration of cellular messages because synthesis of the final product "the polypoptide" is usually proportional to the concentration of mRNA. Due to the rapid regulation of gene expression the cellular concentration of messages changes very quickly. This is commonly referred to as steady-state levels, a balance between transcription and degradation. Steady-state levels of mRNA can be detected in a single cell (in situ hybridization) or in total RNA isolated from cells or organs and separated in agarose gels (Northern blots). Although analysis of mRNA steady-state levels are very useful, they do not provide information about the mechanisms that regulate gene expression. Changes in gene expression primarily occur at the level of transcription; characterization of the RNA species being synthesized at a given time is performed by the nuclear run-off technique. When there is no a priori information about the gene that may be regulated after a particular situation such as sepsis, the total pool of messages present in cells at given time can be characterized by a combination of in vitro translation of the cellular messages and fractionation of the in vitro translated products by two dimensional gel electrophoresis. Such approach permits the visualization of the general pattern of gene expression, a "finger print", of the physiologic state. In summary, researchers now have access to a great variety of techniques to identify and characterize genes expressed in response to a physiological condition that may be utilized as "molecular tools" to study the organism's responses to shock and sepsis. The acute phase proteins are a set of plasma proteins with greatly increased plasma concentrations during acute inflammations and after tissue trauma, e.g. after major surgery. The proteins counteract the source of injury, facilitate clearance of infectious particles by phagocytes, assist immune responses and initiate wound healing. The corresponding genes are expressed in liver hepatocytes and regulated by the inflammatory mediators interleukin 1 and interleukin 6 (ILl, IL6). Class1 acute phase genes are mainly controlled by ILl and by combinations of ILl plus IL6; class2 genes mainly by ii,6. The human C-reactive protein(CRP), serum amyloid A(SAA) and complement C3 genes will be discussed as examples of class1 genes, and rat c~2 macroglobulin as a prototypical class2 gene. Both classes of genes use different types of ¢ytokine response elements in their promoter-proximal transcription control regions; class1 genes use the transcription factor NF-IL6 or C/EBP as the key factor to mediate cytokine responsiveness; class2 genes use a different, so far unidentified factor to achieve responsiveness to IL6. Ongoing efforts to purify this factor, called the IL6-Response Factor (IL6-RF), from rat liver nuclear protein extracts will be described. The IL6-RF apparently mediates the effects of IL6 in a broad range of cell types, including B lymphocytes and other hematopoietic cells. It is therefore likely to be of equal general importance for gene regulation by IL6 as NF-IL6.

The importance of transcription factors as potential targets for novel bioactive substances and possibly therapeutic agents will be discussed with the help of several recent examples. New methods for altering the germ lines of mice provide powerful tools for understanding the roles of individual genes in normal and pathological processes, and for reproducing specific genetic mutations that result in congenital disease.

Three approaches will be discussed. Conventional transgeulc mice are frequently constructed using artificial transgenes that express genes from a promoter other than the normal gene promoter. This paradigm is used to produce very high levels of a gene product, for example, with a viral promoter, or to attempt to regulate gene expression using an inducible promoter. Alternatively, normal promoter sequences can be hooked to a gene whose expression is lethal to the cell (e.g., diptheria toxin A chain) to ablate all cells expressing that gene. Another approach uses an intact gene with its normal regulatory sequences. Here, an elevated level of the gene product is delivered from the correct cells in the correct tissues at the correct developmental stages or in response to natural signals. This approach has been refered to as "genetic pharmacology," and provides a precise delivery of the gene product to the target. However, to be most effective, the transgene should contain arl regulatory sequences as welt as the entire genomic segment containing gene coding regions.

The introduction of conventional transgenes is limited to roughly 50 kilobases, while genes containing all regulatory sequences frequently stretch over dozens or even hundreds of kilobases. To overcome this problem, procedures have been developed recently to introduce yeast artificial chromosomes (YACs) into mouse embryonic stem (ES) cells. YACs can include over 2 Mb of human DNA, large enough to encompass the largest known human genes. ES cells are also useful for targeted gene deletions and mutations.

Homologous recombination can be targeted to any known gene in ES cells. When these modified cells are injected into a mouse blastocyst, they can be incorporated into all tissues of the resulting mouse, including germ cells, so subsequent generations will pass the genetic modification as an inherited trait. Assessment of the effects of a null allele on development can provide valuable insights into its normal role. Ultimately, these technologies may be adapted to human cells -not for embryonic intervention, but to provide modified stem cells for various tissues (e.g., gut, eye, hematopoietic system) which could then be applied to somatic therapies. with major illness/injury that respirswy faihne fi'equently followed sepsis, Tilney el al nse, d the term "soqueutial syste~ failure" for patients with renal failure after Solmir of ruptured abdominal aortic ~aeurysms, I then reviewed our experiences after inju~ and oporstion and suggested that multiple, progsesslve or sequential systems or organ failure was a syndrome to be reckoned with in the, 70's. Eiseman et al then wrote about "multiple organ failure", especially with liver failure. Polk el M found that renmte orgau failure often signalled occult anita-abdominal infection. Fry et al ampbasized the role of uncontrolled infection in organ failure, Border et aI described metabolic alterations with MOF and used the term multiplesystems organ feilmo (MSOF). Other early conltibutots to our knowledge of MOF include Faist, Trunkey and Miller, Marshall and Dimic&, Cassone, Catlleo, Meakins and Marshall, (}otis et at., Mater and many others. MOF or MSOF were used eommoifly. Cerra coined the terms "post-trsumatic septic sFndromc" and "hyper metabolism otgau failure complex", and Border et al, used lhe ex3~ression "gut origin seplic states" to illdioate important eurrem aspects of abe gt.'nerlc problem of organ failure and death, Other terms include acute organ-system failure, multiple organ injury syndrome, post-traumatic multi-system organ failure nod post-~aumatic organ system infection s~dmmc (FI~SIS), I bdieve the latin "maltiple organ failure" is clean, neat and says it all, Now there is a proposal for a new set of torminolugios -MODS and SIRS. Will they be helpful in the study and esro of patients? Our speakers in this session will review thcsa proposals and the evidence as we strlVO for eoasensns. It has been well demonstrated that the administration of pro-inflammatory cytokines, and especially tumor necrosis factor (TNF) can result in a picture similar to septic shock with secondary multiple organ failure. It is also clear that tissue hypoxia can lead to organ damage. We believe that it is the interaction between the two phenomenons that can lead to MOF. On the one hand, the inflammatory response is amplified in the presence of hypoxia. On the other hand, the release of the mediators of sepsis can increase tbe oxygen demand ef the tissues, alter their oxygen extraction and decrease myocardial contractility, and the combination of these elements accounts for the development of septic shock. This hypothesis is supported by two lines of evidence. First, MOF is usually preceded by an episode of acute circulatory failure (even though frank circulatory shock may not be evident). Second, recent experimental and clinical studies indicated that immunotherapy (especially antibodies to TNF) is particularly effective in the most severe forms of sepsis, associated with circulatory shock. Hence, an effective way to prevent MOF may be a combination of aggressive cardiovascular management and immunotherapy. Development of the multiple organ dysfunction syndrome (MODS) in the critically ill follows an heterogeneous group of stimuli including infection, sterile inflammmion, extensive tissue injury, ischemia, intoxication with a variety of substances, and immune system activation. Whether the resulting physiologic abnormalities reflect a common pathologic process, or are simply a non-specific manifestation of tissue injury is unknown. Moreover, the lack of clearly-defined functional criteria for the characterization of the syndrome on the one hand, and of reliable biochemical markers of its evolution on the other, has made attempts to define its epidemiology and natural history difficult.

Using a numeric scoring system to quanfimte the clinical severity of MODS, we demonstrated that, although MODS is more common in patients who have significant infection at the time of ICU admission, a much stronger correlation is evident between the severity of MODS and the subsequent development of nosocomial ICU-acquired infection. Furthermore the severity of MODS correlates strongly with the severity of the clinical septic response, and with the degree of immunologic compromise evident on delayed type hypersensitivity skin testing.

To determine the relative importance of the initial stimulus, and the host response to that stimulus, to the subsequent emergence of MODS, we undertook a matched cohort study of 109 critically ill patients admitted to the ICU with infection, matched by age, sex, and APACHE II score to 109 uninfected controls. Organ dysfunction scores were higher in patients admitted with infection, but were also significantly associated with the expression of a septic response at the time of ICU admission, independent of the presence of infection. By stepwise regression analysis, the magnitude of the septic response was the most powerful predictor of the severity of MODS in both infected and uninfected patients.

These data suggest a model of MODS in which a stimulus (eg infection) and the host response to that stimulus (eg the septic response) result in a state of altered systemic homeostasis (manifested in this example by immunologic dysfunction). Moreover they suggest that the principle determinant of the emergence of MODS is the response of the host, rather than the nature of the stimulus, and are consistent with the hypothesis that a stereotypie systemic response to an heterogeneous group of injurious stimuli underlies the pathogenesis of MODS.

Arthur E. Baue, M.D, Ever since the attorapt 10 build the Tower ef Babel, as desclibcd in the Old Testament of the Biblc, it has been difficult to got agreement o;x common definitions or language, Although MOF has served well, there will always he now classifications proposed with good reasons for them, Thus MODS and SIRS are not the final oxpressinns in the lexico~aphic sweepstakes. Attcmpts to develop standard animal shock and scpsis models (hemorrhage, endotoxin, foes[ pellets, cecal ligation and puncture, E. colt hlf0sioI0 to evaluate therapy have net been suoca.ssf~. Clinical syndromes of MOF, sepsis, sepsis syndrome and athers have not been aeacpte.d by all. Our problem is that we arc tryittg tu d¢ffino a non-emily. MOF or MODS or SIRS is net a disease or oven a syndrome. It is a series of complications of injury, disease and intensiva care which loads to death for many patients in intensive care units, It is not a fixed entity, but an ewilving picture. We have lumped (ugethet for therapy and definition a number of diseases and problems according to clinical manifestatieos rather than the eause of Ihe problem, The search for unified mechanisms has not bean successful. Will we benefit Oleo from a more precise classification and definition of a non-outjty, a hodgu-podge of human disorders which have charsmeristic.s and manifestations of tissue injury, inflammation and infactian? Re.hi clinical trials of potential "magle ballets" suggest that we should go in a different direction--instead of lumpors, we should be splitters, To seek effective therapy, we must fucus on specific disorders such as trauma, specific infections, inflammatory disease and chronic or acute organ damage (COPD -renal failure, etc.), I will say more aboet this later ha the me.ethag. Thcre is cue potentially important pert of the MODS proposal, and that would be do~mcntation of organ dysfunction before fsiiure occurs and before tile need for external supporl dovelope. This could lead to hatter methods of preventing organ failure. Preveution is ultimately the only answer to organ failure, m MOF, MODS and SIRS. Have wo reached a consensus?--Only that we arc dealing wilh a diffl~x:lt probtolR looking for solutlons. Acute lung injury (ALI) and its most severe form, acute respiratory distress syndrome (ARDS) characterized by severe hypoxemia, diffuse infiltration of both lungs, a reduction of compliance and a wedgepresure lower than 18 mm Hare related to direct or indirect injury. When ALIresults from direct injury (toxic agents, lung contusion, pneumonia), the lesions of epithelial barrier and the activation of lung macrophages are the first events leading to the release of inflammatory mediators wh:ch are responsible for alteration of the membranar permeability and for invasion of the alveoli by fluids, proteins and cells. An inflammatory reaction develops, with injury to endothelial cells, adhesion of polymorpbonuclear leucocytes (PMN) and the release in blood stream of intlammatory mediators dispersing the inflammatory reaction to other organs and tissues. The failure of lung function also results into hypoxia in other organs, leading to tissular ischemia, triggering an inflammatory reaction leading to organ dysfunction. Frequent complications of direct lung injury are septic pneumonia with endotoxin release, phagocytes activation and cytokine production disseminating inflammation. By this way, direct lung injury can be at the origin of the multiple organ dysfunction syndrome (MODS), the frequence of which having been estimated to 40% and is increased when ALl has occured in patients alreadypresenting an organ dysfunction (immunodepresslon, cancer...). When ALI results from redirect or extrathoraclc injury (trauma, infections, hypovolemic shock, acute pancreatitis...), inflammatory mediators and micro-aggregates released in particular organs or tissues reach the lung via the blood and lymph streams, are filtered in the organ or trapped in the lung capillaries, and are responsible for a local inflammatory reactaon (endothelial cell activation, PMN trapping and adhesion, platelet aggregation, release of mediators). Sepsis acts by the way of endotoxiu and sequential eytokines release (TNF, ILl,...) while trauma with important blood loss acts especially by the way of hypoperfusionreperfusion phenomena leading to a disseminatedinflammatory reaction and to a possible bacterial transloeation when intestinal hypopeffnsion is present. In these conditions of indirect injury, ALl (or ARDS) is a Iocal early (often the first) manifestation of a general phenomenon of altered membrane permeability and inflammatory reaction, easily and rapidly recognized by its clin~cai signs. The iung is one of the numerous organs participating to the MODS, In these conditions, the mortality rate is high, reaching more than 80% when sepsis is present or when at least 3 organs are implicated in MODS. By direct or redirect injury, lung is so a target organ for MODS and injured lung can be the cause of MODS or simply a participant to this syndrome.

Jorge Cervts-Navarro Main determinants for brain dysfunction are breakdown of the blood-brainbarrier and raise of intracranial pressure (ICP), in second place decrease of systemic blood pressure, disorders of blood coagulation and respiratory function. The blood-brain-barrier of cerebral blood vessels can be opened by stroke, cerebral trauma, inflammation, convulsions, acute hypertension and changes in the blood osmolarity. The consequence is brain oedema, increase of lOP and decrease of the perfusion pressure. A close correlation between intraventricular pressure and the fate of patients with severe brain injuries has been established. Regional CBF values of 16 to 18 ml/100g/min are reflected in isoelectric EEG, and values about t5ml/100g/min, result in loss of somatosensory evoked potentials. For ionic pump failure the threshold has been determined by values of about 10 ml/100g/min and is reflected in massive release of intracellular potassium. In stroke and in brain trauma when the oedematogenous condition is initially localized the tissue pressure variance within and between the hemispheres lead to tentorial and falciform herniations are the common cause of death even before the critical threshold of intracranial pressure is reached. The increase of the ICP over the level of the perfusion pressure leads to the arrest of the CBF. If this persists {or periods exceeding 6 seconds of global brain ischemia loss of conciousness and developing seizures appear. If it exceeds 10 min irreversible injuries of the brain appear. Following cerebral ischemia and after severe head injury a tendency to increased coagulation can be detected. Vascular occlusion may develop either as a result of local thrombus formation or from emboli caused by circulating platelet aggregates. The formation of microthrombi is triggered by the plateletactivating factor a mediator in central nervous injury also responsible for aggravation of posttraumatic brain edema. Acute hemorrhagic leucoencephalitis and brain purpura are known to arise as complications of acute viral infections and following Gram-negative septicaemia.

Institute of Neuropathology, Free University of Berlin, Hindenburgdamm 30, 12200 Berlin, Germany

Undoubtedly, attempts to quantify as objectively as possible the degree of dysfunction of the various organs is very valuable. However, cardiovascular failure has a peculiar place in the context of sepsis. The development of acute circulatory failure (shock) is of paramount importance as a cause rather than as a consequence of organ failure.

It is therefore essential to clearly separate patients with and without circulatory failure. Once this has been done, attempts to list the cardiovascular system among the other systems have been unconvincing.

In a number of studies, cardiovascular complications are assimilated to a low systemic vascular resistance (SVR), but since SVR is basically the ratio of blood pressure over cardiac output, and since hypotension is already included in the definition of shock, then a low SVR is basically equivalent to a high cardiac output, which is a characteristic rather than an ominous complication of severe sepsis. Development of a low cardiac output unresponsive to fluids is usually a terminal event. It is not advisable to list other problems such as severe arrhythmias, myocardial infarction or tamponade as complications of severe sepsis.

HOST DEFENSE DYSFUNCTION FOLLOWING TRAUMA, SHOCK AND SEPSIS E. Faist, G. F. Babcock * Major accidental, burn and operative injury often precipitates a profound multicentric immunologic depression that is believed to predispose patients to become anergic and thus to develop towards a higb probability of infection. Anergy or a lack of immunologic reactiVity stands for a total elimination of skin test reactivity and recall antigen responses in the inununocompromised patient. Anergy following overwhelming trauma or major septic conditions results from a massive impairment of cell mediated immunity (CMI) together with a whole body malignant inflammationiike state. We and others have demonstrated clearly that the skeration of CMI following trauma is mainly due to the disruption of intact monocyte (MOO) / T-cell interaction. Within this phenomenon we see a shift of the cell ratio in the compartment of peripheral blood mononuclear cells (PBMC) with a considerable increase of prostaglandin E 2 synthesizing M~ and a simultaneous decrease of functionally competent CD3+ and CD4+ lymphocytes. T-cell dysfunction in states of profound stress is characterized by impaired synthesis of two crucial lymphokines -Interleukin-2 (IL-2) and gamma-Interferon (y-IFN). The inability to produce adequate amounts of IL-2, most likely attributable to alterations in the transmission of signals from the cell membrane to the nucleus, results in incomplete proliferative T-cell responses to antigenic stimuli, while a lack of ?-IFN results in inadequate M~ antigen processing and presentation. The role of the two functionally distinct T-helper subsets, T-helper-1 (Thl) secreting principally IL-2 and ?-IFN and T-helper-2 (Th2) acting principally in inducing antibody formation with a secretion of IL-4, IL-5, IL-10 and TNF-[3 has still to be established within the context of major trauma. Antibody formation to common protein antigens is impaired, the predominant finding is a shift from IgM to IgG synthesis. Although excessive secretion of prninflammatory cytokines (IL-U3, TNF-c~, IL-6, IL-8) has been considered to be deleterious for the host in states of major inflammation and infection, in-vitro and whole blood investigation of Me function has clearly revealed that there is initially a marked suppression of MOO from septic patients to synthesize and secrete proinflammatory cytokines to endotoxin. This probably will result in immunodeficiency of traumatized as well as septic patients, since these cytokines in low concentrations are involved in the upregulation of the essential humoral and cellular immune functions. Abnormalities of PMN function noted after serious injury have been interpreted by some investigators as signs of unresponsiveness of this cell population, others have demonstrated signs of PMN byperactivation. Latest findings revealed that there is a clear PMN dysfunction in circulating blood: 70% of all PMN's in this compartment show downregulated or non existent ~-integrins within 72 hours posttrauma -these cells do not phagocytose, have a reduced respiratory burst and appear to be completely degranulated. Restoration of these cell functions within 7 days post-trauma is significantly correlated to survival of traumatized patients. Other reports however stress increased expression of CR3+ receptors (CDll/CD18 complex) on circulating PMN's, increasing the ability of these cells to adhere to intercellular adhesion molecules, thus contributing to play a major role in inducing the pulmonary capillary leakage. Our knowledge about the perturbation of host defenses associated with serious injury and sepsis is continuously increasing and represents the precondition for the design of effective therapeutic measures to prevent and counteract the resistance to invading microorganisms.

Dept. of Surgery, Klinikum Grosshadern, Ludwig-Maximilians-University, Munich, Germany. * Dept. of Surgery, University of Cincinnati, Cincinnati, Ohio, USA

LG Thijs~ CE N~ck Sepsis is the most common cause of acute disseminated intravascular coagulation (DIC) and coagulation disorders as well as abnormalities of fibrinolysis are common in septic patients. Some clinical and experimental studies indicate that DIC is a pathogenetic factor for the development of multiple organ failure. Endotoxin and various mediators (TNF, IL-I) enhance tissue factor and decrease expression of thrombomodulin on endothelial cells in vitro. Also, tPA activity is suppressed and release of PAI-I is stimulated. In such a way the endothelial surface becomes procoagulant whereas fibrinolysis is inhibited. Following administration of endotoxin to normal volunteers levels of prothrombin fragments (FI+2) and thrombin-antithrombin (TAT) complexes increase, indicating thrombin generation. Also the fibrinolytic system is activated as evidenced by a rise of levels of tPA and plasmin-antiplasmin (PAP) complexes but this is counteracted by a subsequent release of PAI-I. In severely ill septic patients levels of TAT complexes are increased and concentrations of the natural inhibitors of coagulation (AT III, protein C) are usually decreased. Also, tPA levels snd PAP complexes as well as concentrations of PAI-I are elevated. The severity of many of these alterations is related to mortality. Thus, both coagulation and fibrinolysis are activated in sepsis, but not in a balanced way, thereby promoting coagulation. One of the hallmsxk pathological alterations associated with sepsis is the development of microvascular thrombosis, an entity ~lmre~erizod by microvas~tlar plugBing with leukoeyies (predominately nentrophils) and fibrin deposits.

PreRanably, ischemia remlt~g fibm ve~-'ttlac ow,.luwion contributes sisni.ficantly to end orgen darnaje and consequent dysfatlctlon. P, er~t sRldies haw focused pmdominately on the role of adhesion molecules in the pathogenesis of neu~ophil sequestration during infection and s~sis. This preparation will review the mechanisms underlying intravak-'v.ler i~brin deposition and its contribution to organ injury.

The normal endothelial cell sure is generally conddered 1o be anticos~am. This state is maintained by two major anticoag~lam molecules on the cell so,ace: hepax'an su,lfale and thmmbomodulin. Studies performed over the past decade have do~mentod a general shii~ towards a procoa~qtlant staw during gram negative hffeodon, aft e~| mediated mainly by changes in the endothalial cell surface. antt-TF antibodies have been shown to be pro~ective during l~'ml endotexemia.

In summary, ~erova~'~ler fibrin deposition, in ~rt mediated by ~rcgulation of cellular TF, is a consistemt patholo~cal flndin~ during sepsis and contributes to organ injury. Strategies designed to abrogate this event may n~nimJze the magnitude of erg~ dysftm~on.

N, V, Christou MD PhD, Department of Surgery, MeGill University, Montreal, Quebec, CANADA Interactions between Immune calls and the endothelium vie adhesion molecules serve to modulate immune cell traffic between the blood strum and the extrmvascular space, These families of molecules acting in a cascade and closely integrated with the oytoklna network, blood coagulation and the complement system, are responsible for the homing of leukocytes to areas of inflammation and the realrculatlon of tymphocytes. There are three types of immune ceils that must exit the blood stream into the extravascular space: lymphocytes ; polymorphonuclear leukocytes; and non-lymphocyta mononuclaar cells (monooytes, basophlls, aoslncphils). B lymphocytes exit the blood stream and recircu}ate mostly via the high endothelial venule (HEV) in lymphatic tissue, Rarely, in chronic }nfectlon=, they may recirQulate vie HEV in non-lymphoi:t tissue. T-lymphocyte recircutation on the other hand can occur via HEV in lymphoid tissue, as well as, the endothalium of the skin, returning to peripheral lymph nodes via the lymphatic channels. Lymphocyte recirculation is the mechanism which allows ¢ontaot between the immune repertoire and pathogens, and homing delivers these cells to the appropriate environment for activation and for effeotor function, PMNs exit the blood stream through endotbelium in close proximity to the site of inflammation after appropriate endothelial membrane receptor expression. They must reach the site of pathogen tnveslon quickly in order to be effective. Their exudation to the inflammatory site can be modulated by their intravascular pdmlng, which results from appropriate receptor expression. Because endotoxin leak from the gut has been postulated as an important trigger mechanism for the systemic inflammatory response syndrome seen after serious injury, the immunologic and metabolic effects of bacterial endotoxin infusion into normal volunteers can be expected to shed considerable light on the validity of this hypothesis. We and ethers have used tbis model to show that endotoxin causes a temporary paralysis in the ability of circulating monocytes to prudce the proinflammatory cytokines Interleukin-i (IL-I) and TNF-a. Endotoxin, however, causes increased production of PGE2 by the same cell population. Endotoxin infusion causes significant suppression of circulating T-cell numbers and a profound suppression in the ability of the circulating T-cell population to proliferate and to produce Interle~in-2 (IL-2) upon in vitro stimulation. After endotoxin circulating T-cells are also more sensitive to the inhibitory action of exogenous PGE2. Administration of cyclooxygenase inhibitors at the time of endotoxin infusion largely abrogates the effect of endetexin on T-Cell proliferation and IL-2 production. Endotoxin infusion also causes the release of increased levels of circulating catabolic hormones including cortisol. Administration of cortisol alone or cortisol along with endetowin to volunteers has allowed dissection of cortisol effects from those of endotoxin itself. Cortisol infusion alone causes a diminution in the circulating T-cell population but the remaining ceils continue to produce IL-2 after appropriate stimulation. Cortisol infused along with endotoxin blocks the overshoot in monocyte proinflammatory cytokine production seen 24 -48 hours after endetoxin infusion. At present, one may conclude that administration of sublethal doses of bacterial endetoxin to human volusteers produces alterations in cytokime production and Tlymphocyte activation which are both similar and dissimilar to those seen following serious injury. Hemodynamic support is based first on intravenous fluids and second on vasoactive agents. Colloids are usually preferred to crystalloids, for their more rapid hemodynamic effects and their lesser passage in the interstitial space. In the presence of hypotension, the pharmacological profile of dopamine makes it the first agent of choice. It is only when high doses of dopamine are insufficient to restore a sufficient tissue perfusion pressure that norepinephrine should be added. Even when cardiac output is not decreased, a dobutamine infusion is often indicated to counteract the myocardial depression, prevent vasoconstriction, and increase oxygen delivery to the tissues. The benefit derived from the administration of "renal" doses of dopamine is doubtful, but stimulation of dopaminergic receptors may increase blood flow also to the gut. In any case, none of these catecholamines has been shown to significantly improve the oxygen extraction capabilities of the tissues. Agents with antiinflammatory properties but also with some vasodilating properties, such as N-acetylcysteine, prostaglandin El or pentoxifylline represent more attractive options to increase oxygen extraction. Cardiovascular support should aim not only at the restoration of a sufficient arterial pressure but also at the maintenance of an optimal oxygen delivery to the cells. It should thus be guided also by measurements of cardiac output, mixed venous oxygen saturation and oxygen extraction, and indexes of cellular hypoxia, such as blood lactate levels, gastric intramucosal pH or vanearterial PC02 gradients. 

Supranorma[ 02 delivery contributes to the prevention of tissue hypoxia

Tissue hypoxia is considered to be the common final pathway in the pathogenesis of multipte systems organ failure. It may directly contribute to cellular injury and organ dysfunction as well as enhance further mediator activation and release by a positive feed back mechanism. Prevention of tissue hypoxia, is therefore, one of the most important aims in the supportive therapy of critically ill patients, especially in those with sepsis and septic shock. Mismatch of cellular O= supply and demand in these patients may resuit from the impairment of the cardiocircu[atory system on all levels. 1. Myocardial depression with a drop in 02 delivery (DO~) 2. Redistribution of blood flow between the different organ systems 3. inadequate nutritive blood flow due to tissue edema and endothelial damage (gIC, leucocyte swelling etc.)

These pathological changes may go along with increased metabolic needs. Cellular and organ 02 supply may therefore be inadequately matched with cellular 02 needs in patients with normal and even supral normal DOz. Hyperdynamic circulation is often present in adequately volume resusucitated patients. It is most likely that one of the bodys main compensatory mechanisms would maintain cellular O= supply in the face of increased metabolic needs and impaired Oz extraction capabillities. This pathophysiology may explain the findings in several clinical investigations that patients with normal or even supranormal DOz can have signs of inadequate regional tissue oxygenation. It is also consistent with the results of different studies which demonstrated that the achievement of supranormal DOz levels, either spontanuously or due to adequate supportive therapy, resuits in better patient outcome. Any attempt to prevent or treat tissue hypoxia in these patients, however, has to take into account the effects of therapy not only on global Dgz but especially on regional and microcirculatory blood flow.

R. Rossaint, D. Pappert, K. Lewandowski, H. Gedach, K. Slama, K.J. Falke Klinik for Anaesthesiologie und operative Intensivmedizin, UKRV, Freie Universit §t Berlin, Germany Important contributory factors to the high mortality of severe acute respiratory distress syndrome lARDS) are the aggressive therapies required, such as mechanical ventilation with high inspiratory oxygen concentrations and airway pressures. As specific therapies for reducing or preventing the general inflammatory reaction of the lung resulting in severe hypoxemia is unknown, today's therapy is limited to procedures which predominantly support or maintain pulmonary function, i.e. pressure limited ventilation with PEEP and permissive hypercapnea, avoiding or treatment of fluid overloading, and positional maneuvers. Extracorporeal lung assist combined with low frequency positive pressure ventilation has been recommended, when conventional therapy fails. Today, extracorporeal gas exchange may be carried out using a system internally coated with covalently bound heparin. This allows for a lower level of systemic heparinzation reducing the risk of severe hemorrhagic complications of extracorporeal respiratory support. From April 1989 to August 1993 89 patients, aged from 15 months to 60 years, were transferred to our intensive care unit (ICU) for treatment of severe ARDS. All fulfilled at least the slow entry criteria of the U.S. National ECMO Study. Due to hypoxemia 13 patients required veno-venous extracorporeal membrane oxygenation (ECMO) immediately after transfer to our ICU. The other 76 patients were first treated with pressure controlled mechanical ventilation, permissive hypercapnea, prone position, and dehydration, if necessary. In 23 out of these 76 patients, in which the gas exchange did not improve in the following days, veno-venous ECMO with heparin-coated systems was additionally performed. Heparin was given to achieve a partial thrombin time between 40 -55 s and an activated clotting time between 120 -150 s. If surgery was performed during ECMO, heparin infusion was interrupted. No severe bleeding complications related to anticoagulation for the extracorporeal bypass could be observed, however, in some patients, after three weeks of ECMO thrombi in the drainage cannula were observed. A problem of of membrane leakage shortened the life span of the membrane lungs to a mean of about four 4 days. In the conventionally treated group 87% survived and in the additionally with ECMO treated group 58% survived, representing a 75% survival rate in patients whose risk of death is considered more than 50%. On the basis of these experiences, we conclude that the described strategy, including veno-venous ECMO with heparin-coeted systems, may improve the survival in patients suffering from severe ARDS.

Abnormally high skeletal muscle pO 2 in septic patients and its normalization during recovery: Evidence against tissue hypoxia but for impaired oxygen metabolism of skeletal muscle? P. Boekstegers, K. Werdan Department of Medicine I, Gro6hadern University Hospital, Munich, Germany A pathological oxygen uptake supply dependence was suggested to indicate tissue hypoxia in sepsis-a hypothesis which is discussed controversially. Because the detection of reduced oxygen transport to tissue and &tissue hypoxia would have important implications for therapy, skeletal muscle oxygenation was studied in patients with sepsis. Intermittent and continuous determination of skeletal muscle pO 2 by polarographic needle or catheter probes provided no evidence for skeletal muscle hypoxia even in severe stage of sepsis (Boekstegers et al., Crit Care Med 1994) . In contrast, mean skeletal muscle pO 2 was found to be abnormally high in patients with sepsis (48 _+ 10,4 mmHg, n=39) compared to patients with limited infection (28, 4 + 4, 7 mmHg, p<0.001, n=16) , to patients with cardiogenic shock (22,3 + 6,2 mmHg, p<0.001, n=lS) and to healthy subjects (33,3 _+ 8,1 mmHg, p<0.001, n=10). Furthermore, serial measurements in patients with sepsis showed that skeletal muscle pO 2 was higher (44,1 _+ 10,3 mmHg) on days with severe sepsis than on days with intermediate state (30,1 _+ 8 mmHg) and than on days without sepsis (26, 8 _+ 5, I mmHg) . Thus, a decrease of abnormally high skeletal muscle pO 2 was related to improvement of sepsis. This was also demonstrated by comparing the decrease of abnormally high skeletal muscle pO 2 with the decrease of APACHE II score or Elebute score as well as Interleukin-6 serum levels in a separate study (Boekstegers et el., Shock, in press). In summary, our data provide no evidence for skeletal muscle hypoxia in patients with sepsis but support the assumption &impaired oxygen metabolism in severe sepsis. NeutropNlic emigration from the vascular space into the extracellular matrix is important for the response to tissue injury or contamination by providing a large recruitable pool of tissue-based phagocytes. The consequences of neutrophil exposure to intravaseular chemoattractants, likely relevant for IL-8 and C5a in sepsis, trauma, and in burn injury, are suppression of neutrophil attachment to endothelial cells and matrix proteins. It is probable that proteoglycan-bound attractants are of considerable biologic relevance, and may result in enhancement of responses to subsequently encountered cytokines. In the presence of connective tissue proteins expressing integrin-specific binding ligands, neutrophils respond to TNF-cq GM-CSF and other cytokines by producing large amounts of hydrogen peroxide. This response is likely important in the digestion of contaminated or injured tissue by facilitating activity of neutrnphil granular proteases and inactivating plasma anti-proteases. This matrix-dependent oxidative response is important in defining potential novel targets for therapeutic intervention. The response appears to involve signalling by integrin-linked tyrosine kinases. The net effect of matrix protein injury through this mechanism is enhancement of the fibrotic response which underlies much of the prolonged ventilator dependence of patients with the adult respiratory distress syndrome. This adherence dependent response has also been implicated in direct hepatocyte injury, and may represent an early component of the syndrome of multisystem organ failure. Study purpose: In this patient (pat.) population with a considerable sepsis morbidity and mortality, parameters proposed for sepsis assessment (Elebute sepsis score [Ele], SIRS) were investigated with regard to their use in the early classification of pop. pat. at risk for developing sepsis. Patients and Methods: In a previous observational study on 110 consecutive pat. after elective open-heart surgery, we had determined Ele daily for at least 3 pop. days (in pat. with a longer stay: until ICU discharge). After publication of the newly proposed SIRS definition, these criteria were retrospectively calculated and compared to Ele. Results: On the total of n =515 patient-days, both Ele and SIRS showed significant (p<0.0001) correlations with parameters reflecting the general and sepsis-related severity of the pop. clinical course. Compared to SIRS, the Ele correlations were better (ICU mortality: 0.59 vs. 0.32; ICU treatment days: 0,51 vs. 0.32; days on mechanical ventilation: 0,54 vs. 0.32; days with vasopressor requirement: 0.59 vs. 0.33; number of microbiological findings: 0.44 vs. 0.29), The optimal classification criteria for the mainly sepsis-related outcome gave maximal Youden indices of 0.87 for Ele (Ele >_12 on >_2 days, accuracy: 94%) compared to 0.42 for SIRS (SIRS present on >3 days, accuracy: 67%). Accordingly, Ele ROC curve areas for both overall (0.92) as well as sepsis-related prognostic evaluation (0.99) were significantly (p<0,05) larger compared to SIRS (0.79 and 0.86, resp.), This higher overall accuracy of the Ele criterion was primary due to a more valid assessment already on the first and second pop. day, where SIRS still had a high false positive classification rate (46% and 53%, compared to 7% and 1%, resp.). Conclusion: In the early postoperative course after cardiac surgery, the SIRS definition displayed a high false-positive classification rate (low specificity) for subsequent sepsis-related mortality compared to better classification results obtained by the Elebute sepsis score.

From the Departments of Medicine I and of "Cardiac Surgery, Grosshadern University Hospital, Marchioninistr. 15, D-81377 Munich, FRG.

Correlation between physiological and immunological parameters in critically ill septic patients. MA Rogy, H Oldenburg, R Trousdale, S Coyle, L Moldawer, SF Lowry A relationship between physiological parameters of severe sepsis and immunological function has not been established.

In an effort to assess such a relationship we prospectively evaluated nine severely ill septic patients. Physiological risk was assessed by the APACHE III score 1, while one component of immunologic function was evaluated by peripheral blood mononuclear cells (PBMC) eytokine production after in vitro LPS stimulation 2. Four of the nine patients died. APACHE III scores at 72 h were lower in survivors (S) than in non-survivors (NS), (51-+13 vs 111-+1 5 p<0.05), while APACHE III scores at admission were not significant different between S and NS (82-+13 vs 95-+13). Down regulation of cytokine production by PBMC upon LPS stimulation was a transient event in S. While S demonstrated an 3 fold increase of TNF-a bioactivity with[r~ 72 hours, NS did not demonstrate any increase at all. A similar pattern was demonstrated for IL-1[3 and IL-6 immunoactivity. TNF was measured by WEHI bioactivity, IL-1[~ and IL-6 immunoactivity were determined by ELISA. The sensitivity was 35 pg/ml for TNF, 30 pg/ml for IL-ll3 and 35 pg/ml for IL-6, respectively. In conclusion, both physiological as well as immunological functions of severe critically ill septic patients demonstrate predictive value for ultimate survival. While patients biological status seems to be more predictable by APACHE III at day 3, p<0.05, the pattern of cytokine production by PBMC upon LPS stimulation over the first 72h might be a reliable predictor as well. Introduction: Therapy of sepsis and its sequelae depends largely on its early recognition. Many studies have investigated the change of certain mediators during sepsis and their potential to predict multiple organ failure and outcome. It was the objective of this study to investigate whether the onset of sepsis can be predicted by alterations of levels of interleukin-6 (IL-6), tumour-necrosis-factor (TNF), PMN-elastase and C-reactive protein (CRP). Materials and methods: Over a one year period, 40 polytraumatized patients were prospectively studied (mean age 42y, 71% male, ISS 28). Serum and EDTA-plasma samples were taken in 12h intervalls until the patient left the ICU. IL-6, TNF, elastase, and CRP were determined immunologically. Sepsis was defined according to the criteria of 'systemic sepsis' (Veterans" Administration Study, 1987) with at least 4 of 7 clinical signs: (1) tachycar-dia> 100/min, (2) temperature >38,9°C, (3) blood pressure < 100mmHg, (4) mechanical ventilation, (5) leukocytosis > 15.000 /ml, (6) thrombocytopenia < 100.000/ml and (7) presence of an obvious septic focus. Clinical parameters, sepsis severity and serum levels were documented on a daily basis, beginning on day 2 after trauma. Results: 21 of 40 patients developed a systemic sepsis (52.5%), and 5 died. All mediator levels were elevated under septic conditions. The clinical severity of sepsis correlated well with the respective levels of mediators. In patients, who developed a sepsis the following day, IL-6 (510 vs. 195 ng/l; p=0.029), CRP (160 vs. 121 mg/l; p=0.035) and TNF (32 vs. 15 ng/l; p=0.045) were significantly increased as compared to those patients who remained non-septic. Elastase levels were considerably elevated but did not reach the level of significance. We conclude that IL-6, TNF and CRP appear to be sensitive markers for prediction of septic complications in polytraumatized patients. Objectives of the study: The assessment of liver function in polytraumatized patients who are at risk of developing MOF is too inaccurate and late by using conventional biochemical parameters.

Methats: The injury severity of the patients (n=26) was determined by the Injury Severity Score (ISS). Lidocaine is given at a dose of 1 mg/kgBW over 2 rain. i.v. and is metabolized in the liver by a cytochrome P-450 mechanism to monoethylglycinexylidide (MEGX). The metabolite is measured by a fluorescence polarization immunoassay. Serial determinations of the test were performed between the 1 ~t and the 11 ~ day after trauma and were compared with other liver function tests (bilimbin, GLDH, ALT, AST). The systemic inflammatory response syndrome (SIRS) is still a challenge concerning early diagnosis, therapy and prognosis. Therefore, evaluation of inflammatory and disease activity becomes more important. C-reactive protein (CRP) is a well established acute phase protein in chronic inflammatory diseases. Recent reports suggest an induction of CRP by interteukin-6 (IL-6), a cytokine involved in the mediator cascade of SIRS. On the other hand, tumornecmsisfactor alpha (TNFcx) is a very early released mediator in SIRS removed very rapidly from circulation. In addition, soluble TNF receptors (sTNFR~5, sTNFR75) are released into circulation in the acute phase response. This study examines the kinetics of five acute phase proteins (CRP, IL-6, TNFot, sTNFR55 , sTNFR75 ) in patients suffering from SIRS. Eighteen patients entered the study after diagnosis of SIRS. Blood samples were drawn every six hours during the first two days and every twelve hours thereafter. CRP was measured in an routine turbimetric assay. IL-6 was detected in an biological assay using the/L-6 dependent 13-cell line 1313/29. Detection of TNFc~ was performed in an ELISA system using a monoclonal antibody" for TNFo~. Soluble TNF receptors were also measured by ELISA. CRP levels were elevated (>10 mg/L) in all patients and at all time points. CRP values did neither differ significantly in patients with (143±33 mg/L) or without (129a:39) multiple organ failure (MOF) nor in survivors (133±41) or non-survivors (147:t:44). In contrast, 1L-6 was elevated in patients wilh MOF (mean 740 pg/ml, range 0-8900 pg/ml). IL-6 levels correlated especially with lung dysfunction. TNF(x levels were consistently elevated in patients with MOF. CRP, IL-6 and TNFoc did not correlate with each other. In contrast, levels for both sTNFR showed a positive correlation (r=0.7). Patients could be divided into two groups by values for sTNFR~5 and sTNFR75: The group with higher soluble TNF receptor levels showed increasing values combined with a poor prognosis. The group with lower levels of soluble TNF receptor consisted of patients surviving MOF or without MOF.

In conclusion, CRP does not monitor the course of SIRS adequately. In contrast, IL-6 correlates with MOF and episodes of high disease activity. High sTNFR levels may indicate poor prognosis.

Klinik F0r An/isthesiologie and Operative Intensivmedizin der CAU Kiel, Schwanenweg 21, 24105 Kiei, Germany.

Ch. Waydhas, MD; D. Nast-Kolb, ivID; M. Jochum, Phi); L. Schweiberer, MI)

Objective: To evaluate the irfflarranatory response after different types of orthopedic operations and compare them with the systemic effects of accidental trauma of varying severity. Patients: In 135 consecutive patients with multiple injuries (ISS 40.6) the inflammatory response to trauma was prospectively studied. The patients were divided into 4 groups according to their ISS points. Additionally, the alterations after 79 secondary operations (>24 hr) were determined (msteosynthesis of the femur (n=28), pelvic girdle (n=ll) and spine (n=8), facial reconstruction (n=13), smaller osteosynthesis (n=13) and others (n=6)). Methods: Specific and unspecific parameters of the inflammatory response were determined in the trauma patients every 6 h, beginning on admission of the patient to the emergency room for a period of 48 hr, and in the operative patients on the morning of the operation, at the end of the procedure and every 6 hr during the first two days. Results: Lactate, neutrophil elastase, heart rate, pO2/FiO2-ratio, and other parameters discriminated significantly between the 4 injury severity groups during the first 48 hr (Kruskal-Wallis-test, p<.05). The degree of postoperative changes differed significantly (Kmskal-Wallis-test, p<.05) between the 5 types of operations for lactate, heart rate, pO2/FiO2-ratio, nitrogen excretion and showed a strong discriminating tendency for neutrophil elastase and C-reactive protein. The extent of changes were highest after operations of the pelvic girdle, followed by procedures on the femur, spine, smaller bones, and the facial region. The postoperative changes after osteosynthesis of the femur or pelvis were comparable to the alterations noticed after smaller (ISS 1 to 24) or moderate (ISS 25 to 40) accidental trauma for neutrophil elastuse, heart rate, pO2/FiO2-ratio and parameters of the coagulation system. Conclusions: There is a considerable inflammatory response to operative procedures that varies with the type of surgery. Large operations cause changes in the body homeostasis that resemble those after multiple injuries. It remains to be established whether the inflammatory sequelae of surgical trauma are additive to the changes caused by accidental trauma. Objective of the study: We retrospectively compared characteristics of elderly patients (~65 years) and yeunger patients admitted to a surgical {SICU) and a medical intensive care unit (MICU). We further studied the relations between advancing age, chronic disease, sepsis, organ system failure (OSF) and mortality in the elderly group. Material and methods: During a 1-year period, 538 patients were consecutively admitted into the 8ICU; and during a 2-year period, 487 patients were consecutively admitted to t~MICH. Criteria for chronic disease, sepsis, OSFsi.e. cardiovascular (CF), pulmonary (PF), renal (RF), neurological (NF), haematological (HF), hepatic (LF), and gastrointestinal failure (GF)-were derived from the literature. Results: Patients from the SICU and~CU were similar in age, number of OSF, and length of stay. However, when compared to SICU patients, MICU patients had more CF (p<O.O1), sepsis (p<O.O01), and mortality rate (27% vs. ii~, p<O.OOl). In contrast, SICU patients had more HP and NF, and prior chronic disease (all, p<O.O01). Of the total group of 1,025 patients, there were 406 (40%) patients 65 years of age or older. Elderly and younger patients had comparable length of stay in the ICJ, but elderly patients had significantly more chronic disease, sepsis, and number of OSF (all, p<O.O02). All OSFs, except HF and NF, were more ~o~mon i~ mdeLjl, eompard to you-ger patients. ~he mortality is also higher in elderly patients (29% vs. 11% in younger patients). After multiple logistic regression, only number of OSF increased mortality by a factor of 2.48 (95% confidence limits, 1.97-3.12), age did not increased [odds ratio 1.05 (i.01i.i0)], while admission to the SICU reduced risk of death by a factor of 0.15 (0.08-0.30).

Conclusions: Based on these results, we conclude that age does not have any impact on ICU outcome in the elderly. The only prognostic factor is the extent of acute disease, as measured by the number of OSF. Hence, prevention of OaF may influence outcome in elderly patients with critical illness. The lower risk of death in SICUpatients may be explained by the large proportion of postoperative elective patients with relative mild chronic disease, and the lower incidence of sepsis.

Department of Surgery, Free University Hospital, De Boelelaan 1117, 1081 ~V Amsterdam, The Netherlands.

SEPTIC SHOCK AND LOW OUTPUT SYNDROME H.Miyakawa, T.Noguchi, S.Hattori, A. Mizutani, M. Mori and N.Honda

Objectives: Cytckines are thought to cause the acute phase reactions under several critical conditions. We had investigated the relationships between cytokines network which included IL-6,1L-8,ACTH and cortisol,and the outcome of the patients with septic shock or low output syndrome (LOS). Materials and methods: In the septic group and LOS group,nine and four patients were enrolled respectively. Furthermore, the patients in the septic group were divided into two groups,survival (n=5) and non-survival group (n=4).tn these patients,lL-6,1L-8,ACTH and cortisol had been measured every day after diagnosis of septic shock or LOS.The total number of measurements were 12 points in survival septic group, l l points in non-survival septic group and 10 points in LOS group.Statistical analysis was used student's t-test to compare the mean of each group,and the changes were reported as clinicat significant if P<0.05. Results: IL-8 levels in non-survival septic group were higher than in both survival septic and LOS group.lL-6 levels in non-survival septic group were more significant than in both survival and LOS group.Otherwise,there were no differences with respect to ACTH and cortisol among three groups. Conclusions: We assumed that LOS would make several organs dysfunction as well as septic shock. But our results showed septic shock, especially non-survival,had different cytokins network mechanism for organs failure from LOS. Lasson ,~, G/3ransson J, JiJnsson K.

Eady diagnosis of complications after trauma is imperative to decrease morbidity and mortality. For this purpose an easy, cheap and fast diagnostic method is needed. The aim of this study was to analyse if complications after surgical trauma of various extent could be diagnosed earlier using preoperative or daily postoperative measurements of various plasma proteins than by using climcal signs of complications. Material and Methods: Two acute phase proteins (C-reactive protein and antichymotrypsin), lbur main plasma protease inhibitors (alpha-2-maeroglobulin, C 1estemse inhibitor, antithrombin Ill and alpha-2-antiplasmin) and one collagen precursor (procollagen III peptide) were measured preoperatively and then once dally for 7 days postoperatively. For the plasma protease inhibitors immunorcactive as well as functional levels were measured. Haemoglbin, albumin and the white blood cell count were also measured. Measurements were done preoperatively in 256 patients and continued postoperatively in 78 patients, 70 % of the patients were operated on for malignancy. Resulls: Preoperative plasma C-reactive protein levels were higher in patients developing postoperative complications. This discrimination increased when plasma was sequentially analysed postoperatively, when a remaining high level ora second increase in plasma levels depicted a complication 1-3 days earlier than clinical signs. High levels were seen both in patients developing local as well as in patients developing general complications. There was no clear correlation between plasma protease inhibitory levels and complications, neither pre-nor postoperatively. Procollagen III paptide levels were slightly (but not significantly) higher postoperatively in patients developing complications. Contusions: The acute phase response, especially concerning C-reactive protein, predicts postoperative complications better than both plasma protease inhibitors or collagen precursors. Preoperative plasma C-reactive protein levels indicate the risk of postoperative complications, while daily postoperative measurements more readily depicts a complication, usually preceeding the clinical signs of complication by 1-3 days.

Dept. of Surgery M'alm6 General Hospital, S-2 I4 01 Malmr, Sweden. MOF is a major threat to the extensive burned patients and associated with a high mortality rate. The role of endotoxemia in the pathogenesis of MOF in burned patients remains controversial. In this study, We examined the relationship of plasma endotoxin levels to development of MOF, and outcome in patients with thermal injury. Methods: A prospective cohort study of 17 patients admitted with burns covering more than 70 per cent of body surface area was undertaken. Circulating endotoxin concentrations were measured by modified limulus amebocyte lysate assay in serial samples of plasma. Results: 7 out of 17 burned patients developed MOF according to multiple criteria. The plasma endotoxin concentrations of patients with MOF were 0.512--1.127 EU/ml, significantly higher than that of 10 patients without MOF (0.216-0.553 EU/ml) on 3, 14, 21 and 28 days postburn (p< 0.05--0.01). Significantly more incidence of positive endotoxin test (>_ 0.120 EU/ml) was found in patients who developed MOF as compared to that of Non-MOF during the observation period (p<0.05). As the mean endotoxin levels increased, the prevalence of MOF and death also increased (see table below), persistent endotoxemia carried a poor prognosis. Conclusions: The present investigation provide further evidence that endotoxemia in severely burned patients commonly occur. Cimulating endotoxin has also been found to be strongly associated with development of MOF and mortality following major burn injury. Multiple hemostatic changes occur in sepsis mad multiple organ failure (MOF). To evaluate the role of platelcts in patients with sepsis and MOF, we examined changes in surface glyeoproteins on circulating platelets of t4 patients with suspected sepsis and MOF. The severity of sepsis and MOF was assessed by Eiebute and APACHE 1I scoring system, respectively.Using flow cytometric techniques and platelets specific monoclonal antibodies, platelet surface expression of fibrinogen receptor on GPIIb-IIIa, ofvon Willebrand receptor GPIb, and of granula glycoproteins (thrombospondin, GMP-140, and GP53) was measured. Receptor density of GPIIb-Illa mad GPIb on circulating platelets was not affected by sepsis or MOF. In septic patients surface expression of activated fibrinogen receptor (LIBS1 expression) was significantly elevated (p<0.05) and correlated well with severity of disease (F0.597). NO significant change in surface expression ofthrombospondin, GMP-140 or GP53 was noted in septic patients. In contrast, degranulation ofgraanle glycoproteins was significantly elevated in MOF (!0<0.05) that correlated well with severity of MOF (GMP-140, r=0.611; Thrombospondin, r=0.643).We speculate, that platelets in sepsis circulate in a hyperaggregable (fibrinogen receptor activation ) but still reversible state that results in increased risk of microthrombotic events. In the course of the disease, irreversible platelet degranulation might occur and may play an important role in development of MOF. Abdominal sepsis is still associated with high morbidity and mortality. The present study aimed at evaluating patients with abdominal sepsis treated at our surgical intensive care unit during a 10-year period with the aim of identifying potential prognostic factors, bacteriological cultures, diagnostic procedures, treatment and outcome. During the period 1983-1992 I66 patients with abdominal sepsis were treated at the ICU at our university hospital. 60 patients were women and 106 men with a mean age of 64 (17-101) years. In 74 cases, the abdominal sepsis occurred as a postoperative complication. The patients were scored according to APACHE II and bacteriological cultures and the occurrence of organ failure were noted.

The patients were hospitalized in median for 40 (-273) days out of which 7 (-178) in the intensive care unit. 46 out of 166 patients (28 %) died in median after 18 (1-194) days. The primary cause of mortality was multiple organ failure (38/46; 83 %). APACHE II scoring could not predict a fatal outcome. Abdominal bacterial cultures were dominated by bacteria of enteric origin (83 %) and in 60 % cultures grew multiple bacteria. 134 patients bad organ failure and 64 multiple organ failure. 29/166 patients (17 %) had abdominal sepsis due to diffuse peritonitis despite a morphologically intact gastrointestinal tract and the absence of localized abscess formation. Mortality in this group was significantly higher as was the percentage of positive blood cultures and the occurrence of multiple organ failure.

Abdominal sepsis is still associated with a high mortality, predominantly caused by multiple organ failure. Abdominal culture findings are dominated by bacteria of enteric origin. In about 1/5 of patients with severe abdominal sepsis a diffuse peritonitis with intact gastrointestinal tract without localized abscess formation was found. In this group the mortality was increased as well as the risk of developing multiple organ failure. During the period from January 1985 to September 1993 56 patients, mean age 24+4 years were referred to our department of Resuscitologywith the diagnosis of eclampsia. All the patients were delivered by cesarian section and were mechanically ventilated for 7.3_+5.7 days. Diagnosis of sepsis was confirmed in 6 cases by clinical and microbiological methods. Patients were divided in two groups: lnon septic patients, 2-patients with sepsis, the control group consisted of 20 patients after cesarian section without symptoms of eclampsia or infection. We determined plasma concentrations of immunoglobulins A,G,M(A,G,M), complement factors (C3,C4), alphal-antitrypsin (AAT), trausferrin (TRF) and albumin (ALB) using BECKMAN (USA) analyzer,protein concentration, using KONE (Finland) analyzer.

A(mg/dL) G(mg/dL) M(mg/dL) C3(mg/dL) C4(mg/dL) K 203+-31 971+47 205_+21 134+12 30+-3 1 154-+29" 568-+48* 142_+24" 81-+14'

17_+4" 2 102+_21'* 420-+30** 96-+21"* 67-+10"* 11_+3" In a prospective study we investigated serum of 12 severly traumatized patients withdrawn directly after admission at our hospital (TR I). Follow up controls were taken daily until day ten after trauma (TR II). Two control groups were performed: Serum of healthy volunteers (CO, n =24) was investigated as. well as serum of patients undergoing elective herniotomy (n=12) 24 hours before (OP I) and 24 hours after operation (OP II). Serum bactericidal index (SBI) was determined using a hemolytic E.coli strain 08:K87:H19. 200/11 suspension with a final concentration of 250 -400 CFU were incubated with l OOpl serum. After overnight incubation SBI was calculated according a special formula. Results: CO 74.8 _+ 6.3 OPI 77.8 _+ 6.4 OPII 68.2 _+ 7.7* TRI 38.4 _+ 9.1"* TRII 57.6 + 8.9** (*:p< 0.05; **:p<O.01) SBI represents a simple and reproducible method to detect immunobiological alterations after trauma and elective surgery. Patients undergoing elective surgery showed slight but significant diminuation of SBI 24h after operation. Deterioration of SBI was observed in serum of traumatized patients directly withdrawn after admission at our hospital. Ten days after trauma significant improvement of SBI was found. Further investigations have to be done to proof sensitivity and prospectivity of this simple but reliable test. A sepsis like syndrome (SLS) occurs in i0 % of our patients following cardiac surgery. SLS is characterized by a severe decrease in systemic vascular resistance requiring the use of vasopressors to maintain adequate hemodynamics in the presence of negative blood cultures. In order to determine whether preoperative factors predict the occurrence SLS we retrospectively studied 62 patients with SLS (group i: age 61.1 ~ 10.8 years); they were compared to 52 control patients (group 2: age 58 ± 14.3 years). Treatment of SLS consisted of aggressive fluid replacement and norepinephrine infusion. Rydrocortisone (200 mg/24 hrs) was given on the first day. Antibiotics were switched prophylactic to cover gramnegative bacteria in 48 pts. Preoperatively pts in group 1 revealed a significantly lower cardiac index (2.5 ± 1 i/min/m 2) compared to group 2 (3.1 ± 1.2 i/min/m2; p < .02) higher left atrial pressure (group i: 17.8 ~ 8.1 mmHg; group 2: 13.1 ~ 6.1 mmHg; p < .005) and lower ejection fraction (group i: 57.9 Z 16.3 %; group 2 65.2 ± 10.9 %; p < .03). Bypass time, minimum and mean blood pressure on bypass and aortic clamp time were not different between the groups. Immediately postoperatively (pop), the white blood count was elevated (group i: 15,700 ± 8,200/ui; group 2: 10,200 ~ 4,400/ui; p < .0005) and core temperature 12 hours pop was higher (group i: 38.3 7.6 °C; group 2: 37.6 ~ 8.4 °C; p < .004). Serum lactate was already elevated on arrival on the ICU pop (group I: 5,0 Z 1.9 mmol/l, group 2: 2.2 ~ 0.9 mmol/l) but dropped within 24 hours (p < .0005). ICU stay (group i: 3.7 ! 4.3; group 2: 1.6 ± 1.9 days; p < 0.001) and mortality over 3 months (group I: 19.1 %; group 2: 3.1 %; p < .005) were significant higher in group i. SLS is observed more frequently in patients with preoperative cardiac congestion and results in longer ICU stay and higher mortality. Although intraoperative hemodynamics are similar in both groups, the white blood count and serum lactate levels suggest an intraoperative cause for SLS. Further investigations will focus on possible intraoperative causes of SLS. The objectives of the study were to estimate the problem of stress ulcers in critically ill patients and to compare Ranitidine and Omeprazole to determine their efficacy in the prophylaxis of stress ulcers. 33 patients who were admitted to the Intensive Care Units between January and July '92 with polytrauma, sepsis or those requiring ventilatory support for more than 48 hours were selected for the study. The patients were randomized into 3 groups to receive Ranitidine, Omeprazole or a placebo. The changes in gastric pH were monitored 4 hourly and the patients underwent an upper GI endoscopy 48 hours after the start of the study to look for stress ulcers and hemorrhage. The drugs were withdrawn one week after the start of the study. Of the 33 patients selected for the study, 14 had undergone cardiac surgery, 6 had polytrauma and 13 had sepsis. The change in average pre and post drug gastric pH was +1.1 in the Omeprazole group, +0.5 in the Ranitidine group and -).9 in the placebo group (p < 0.05). Endoscopic evidence of stress ulcers was seen in 62% of patients not on prophylaxis, 18.1% of those on Omeprazole and 54.5% of those on Ranitidine. While all the patients in the placebo group who had stress ulcers showed endoscopic evidence of hemorrhage, 82% of the patients with ulcers in the Ranitidine group and none of the patients with ulcers in the Omeprazole group showed hemorrhage (p < 0.01). In critically ill patients the incidence of stress ulceration is high. An alkaline gastric pH is protective against stress ulcers. Omeprazole when used for prophylaxis against stress ulcers in critically ill patients or those on short term ventilation will reduce the incidence of formation and hemorrhage from stress ulcers. The early and accurate diagnosis of sepsis is of key importance for critically ill patients survival. Many studies have shown that tumor necrosis factor ~x (TNF ~) and interleukin-6 (IL-6) are mediators of the inflammatory response in sepsis. Bacterial antigens interact also with antigen specific T cell receptors and the activation of T ceils takes place. Activated T cells release soluble interteuldn-2 receptors (sIL-2R). The aim of this study was to evaluate: the significance ofTNF c~, IL-6 and slL-2R in the diagnosis of bacterial sepsis and to evaluate the correlation of clinical and laboratory parameters with serum levels of TNF ~x, IL-6 and slL-2R. We examined 19 patients with clinical signs of sepsis treated in the intensive care units of University Medical Centre Ljubljana. Venous blood was drawn from each patient within 24 hours after the first clinical signs of sepsis. Clinical data (body temperature, blood pressure), laboratory dam (peripheral white blood cell count with differential, platelet count, C-reactive protein), and quantitative blood cultures were recorded. Serum levels of TNF c~, 11,-6 and sIL-2R were measured in 19 septic patients and in 20 adult healthy controls. TNF ct, IL-6 and sIL-2R serum levels were measured by commercially available RIA and ELISA (Amersharn and Eurogenetics) kits. The differences in serttm levels of TNF ~, IL-6 and sIL-2R between healthy control and patients and correlation between clinical and laboratory parameters were calculated. We found that only slL-2R serum levels were significantly (p<0.005, Student t test) increased in patients in comparison with healthy controls. Of all indicators of sepsis that we compared, only hypotension was significantly correlated with TNF ~x levels and leukocytosis with ]1,-6 levels. We conclude that only the increased serum sIL-2R levels were predictive of bacterial sepsis within 24 hours after the first septic signs. TNF c~ and IL-6 levels were not significantly increased at that time in lifts small group of patients. In order to create high precision prognostic system for patients with sepsis and septic shock we have made prospective and retrospective analis of 200 cases of sepsis.

Patients were included in this analis according Bone's criterions of sepsis. Comparing with APACHE-H, we have excluded such phisiologic variables as:temperature,Na+ ,K+ ,}It, pH,which had a little influence on prognos of illness. We have added the other, more important variables: biilirubin, plateles,PTI. We took into account: the seat arrangement, methods of respiratory support, comorbid conditions as: caneer, diabets,obersity.

In general new score has 5 headings: -physiologic variables, -age, -seat arrangement, -chronic comorbid conditions, -methods of respiratory support. ROC -analis have showed a greater precision of our score, compared with APACHE-II score. There is significant difference between the ROC-curve of APACHE-II scare.

The k.P.Tit~v.,I.E.Gurmanchuk.

Objectives of the study. Sepsis is a severe complication of the local infection, Th e fever and the systemic inflammatory response syndrome may take place in such a case as manifistations of the inflammation induced by bacteria. The systemic fever observed in infection is known to develop due to the action of certain cytoldnes.

Other systemic manffistation of inflammation are the production of acute phase reactants, acute phase proteins, the activation of the complement system.

IVratertals and methods. We observed 35 children with light and 45 children with severe course(l-3 class of serionuess) of the sepsis. The ftmetional activity of the complement system (CH50, level of C 1-C5, factors B told D, C3a), level of C-reactive protein (CRP} and tumor necrosis factor (TNF) were investigated, Results. The level of CRP was increased in the children with more severe class of the septic process in both group (60%-1 st and iOO%-2 nd}( with light and hard eours of disese) of patients.

Parameters of the complement system -the activity of Cl, C4, C3 components, activity of B and D factors-were increased in the seram of children with light course of sepsis. In children with severe course of sepsis these parameters were signffieantty decreased, especially in the group of chiidrellwith higer level of CRP.

We found the negative correlation of the TNP-alfa concentration with the functional activity of classical pathway components (Cl~3) and the positive onewith the funfional activity of the alternative pathway factors B-and D of the complement system.

Conclusions. Thus, in children with different course of sepsis certain changes in levels and functional activity of an acute phase o[ inflammation proteins -CRP and the complement system ones -&Ire characteristic. The relationship between the TNF-alfa level and the activity of classical and alternative pathways proteins indicates on its role in the complement proteins genes expression. 

Schr6der J, Braun J, Rennekampff Ha, Schr6der DW Various alterations of the immune response are known in trauma, but the course will not be complicated by multiple organ dysfunction syndrome (MODS) in all patients. Phenotyping of cells offers a rapid system of evaluation certain aspects of the immune response. Different subsets of lymphocytes and neutrophils were determined to evaluate their prognostic role in developement of MODS. In 16 trauma patients with an injury severity score (ISS) > 20 (mean ISS = 32; mean age 41 years) lymphocyte and neutrophil phenotypes CD 3 (T-cells), CD 4 (T-helper cells), CD 8 (T-suppressor cells), ratio CD 4/CD 8, CD 11b (receptor for CR 3) and CD 16 (FcRIII) were measured on day 1,2,3,5,7 and 10 post trauma. The expression of class II histocompatibility antigen (HLADR) on monocytes (HLADR+ CD 14) and IL 2-receptors on T-helper cells (CD 25/CD 41 were determined as well. The percentage of cells was monitored by immunofluorescence using monoclonal antibodies and three color cytometry. The percentage of HLADR+ CD 14 were significantly lower an day 2,5,7 and 10 in patients who developed MODS (p<0,05) compared to patients without MODS and a healthy control (p<O,O01). The expression of IL 2receptors on CD 4 was significantly different only on day 2. No differences could be measured in lymphocyte phenotypes CD 3, 4, 8 and CD 4/CD 8-ratio as well as in expression of CD 11b and CD 16 on neutrophils. Class II histocompatibility antigen (HLADR) on monocytes offers the best ability to reflect cellular immunosuppression in trauma. This parameter allows the best differentiation of patients with and without MODS. We retrospectively studied the relative importance of age, prior chronic disease, sepsis and organ system failure (OSF) to determine outcome in critically ill patients with acute renal failure (ARF). ARF was defined as serum creatinine > 280/zmol/I, a twofold creatinine rise in prior renal insufficiency or the need of acute renal replacement therapy. Definitions for prior chronic disease and other OSFs -i.e. cardiovascular (CF), pulmonary (PF), neurological (NF), haematological (HF), hepatic (LF), and gastrointestinal failure (GF)-were derived from the literature and described previously.

Of the 1025 consecutively admitted patients to a surgical and a medical intensive care unit during 1 -ye0r period, 136 (13%) had ARF. ARF mortality was 52%. Ninety-eight percent had other OSF. Overall, CF, PF, GF, and NF was significantly more common in nonsurvivors than in survivors (all, p<O,02). Although both the number of OSF before and after ARF was higher in nonsurvivors than in survivors (p<0.02), the number of OSF before was higher compared to after ARF in both groups of patients (p<0.001). Furthermore, age, CF and PF before the ontset of ARF, NF thereafter, and renal replacement therapy were related to mortality (all, p<0.03). However, prior chronic disease was not related to mortality and chronic renal insufficiency was inversely related to outcome (p<0.002). After multiple logistic regression, advancing age, cardiopulrnonary failure and sepsis before ARF, and NF after ARF remained significant.

These results show the high prevalence of ARF in critically ill patients and that ARF rarely occurs alone. The prognosis of ARF in critically ill patients is poor, especially if associated with advancing age, additional OSF, particularly cardiopulmonary failure and sepsis before ARF, and NF after ARF. Hence, prevention of cardiopulmonary and sepsis before the outset of ARF may influence outcome in ARF in patients with critical illness. The search cominnes for a magic bullet ta help critically ill petienta, bat recent elini ',.ml 1rials of agents that showed Ixomise in animal studies raise questions about such trials. ExpcrlmemaI uvidan~ :in anlmats and human volanteet~ for concepts, mechanisms and treatment of inju~ ur illness can be substentlal, l~rauasive and exciting but the positive effects of potential therapy suggested by such animal and ctinieal research may be difficult to prove in sick patients. Efficacy of a new txeatment can be difficult to prove became elinie.al situations, hi spi~e of important biologic phenomena, are v~iable and conrplex. There is redundancy and overlap of mediators ~d problems of timing of therapy. A majt~r cause of this dilemtns is not with the trials or hypotheses, but rather the proMem of the cause or the causability of the human dise~.se that is trebled. This is/hu "causa nets" or '*specific cause" as described by Stehbens, When prevention or tre.atmcnt is based em the sauna nero, extinetiem of the disease is Ix~ssible as with saul/ pox. Only elimination of the cause wl12 cure the disease.

Careful animal studit~s evaluate single pcrturbmiuns for survival or other changes. An LDso preparation may be infiuancad aignificantly by a ~mall clangs which may be insignific~lt ht pstiants, Most human experiments ate carried out in normal volunleers with a single porlurbation, such as a low-level, non-lethal dose of a substance. Such studies are huportant in defining mechanisms attd mediators of disease, In our world of injm'ed, operated, infected and inflslned patients, there are many diseases, MOF is not one of lhem. it isn't even a syndrume, It is the final pathway for a number of diseaaes~ each of which has a cause. MeCormock believes thal a "multi-causal" etiology is a euphemism for ignoranc# or a synonym fo~ unknown etiology. Admission Iv sn ICU, a high APACHE seer% the sepsis syndrome, MOF, MUDS, or SIRS and having predictors indicating potential mortality, are not diseases. The inmpors are getting negative results in trials of agents on eiek or septic ICU patients and aru now becoming splitters, dafining subgaoups at higher risk for death. Such an approach may produce positive results if the diseases are identified and the treatmeul is specific for them, A major challenge in deterraintng the efficacy of new therapies for sepsis nnd shock Is Identifying patients whose e#temtc Inflammatory response Is appropriate from those in whom the mediator| are contributing to end organ system damage end death, Traditional and recently revised categorical patient descriptions such as sepsis syndrome, sepsis syndrome with shock, multiple organ system failure, or the recently proposed systemic inflammatory response syndrome (SIRS) may not be sufilcient since they Identify individuals at varying levels of risk for short term mortality, The efficacy of new lmmunotherapy may be directly related to the patient's severity end risk of mortality at entry to the study. We recently reported (JAMA 1993;270:1Z33-1~41) a comprehensive risk assessment approach for patients with sepsis syndrome, a common eategarlcal description used as an entry eriterlon for many trials. By screening s database of 58,737 recent admissions to 107 hospitals in the United States and Western Europe, we identified 1,195 patients who met sepsis syndrome criteria but were not enrolled tn clinical trinh, We then developed a survival model end severlty assessment method to estimate their 2g.day mortality risk. k..ente physiologic abnormalities were the most important prognostic factors (82% of total chl square) influencing outcome. The specific disease resulting in Ice admission and the days the patient spent In the hospital and ICU before qualification were independent risks, as were age and a clinical history of cirrhosis. The predictive model was cross-validated within the original 1,195 patients with a Receiver D_perator Characteristics (ROC) area of 0.76 and In two Independent databases, The first independent database contained 295 hospitalized patients with gram-negative infection meeting criteria for sepsis syndrome (ROC=0.80); in the second, the 302 placebo patients withtn the recently completed Phase Ill e~aluatlan of IL-1Ra (ROCffi0.76). In all databases the risk model was able to accurately risk stratify patients throughout the range of risks that varled from a very low 1% to a very hlglt 99%.

By drawing on the ready avatlabillty of large clinically aeuurate, current databases and using the power of modern statistical techniques to model the bodfs response to sepsis, we are able to produce very aeettrate pre-treatment risk estimates, These prospectively defined and continuous risk estimates reduce reliance on multiple subgroups and data dredging that can compromise a stady's integrity. They cart provide a nnlform method for patient description across clinical trials of different attempts at immmlotherapy. Risk calculations can also be useful in developing entry criteria for Phase II evaluations in order to initially test efficacy on a limited number of high risk patients, Fonowing sueeessinl completion of a definitive trial, the risk estimates can be used to develop specific |ndlcattons for s drug's use and can monitor the impact of both new and multiple compounds on patient outrome, When dealing with therapeutic trials for the treatment of sepsis, there are 2 ways of checking the comparability of the groups (placebo and treated groups) -To use several description items such as age, sex, preexisting disease, number and nature of organ failures, physiology score, such as SAPS IL (1) -To use a model for risk of death either from a score (APACHE II, lII, (2) SAPS II) or directly (MPM II system (3)) Before using the model for risk of death in septic patients, is it mandatory to validate it (by calibration, i.e. goodness of fit, and discrin'dnation, i.e., ROC curves) first on a data base using the same definition for sepsis as in the trial and secondly on the placebo group of the trial ? or is it better to use a specific model for each trial ?

The general models usually do uot fit for septic patients. There are 2 methods to make them fit : either to add variables to the original score (model for sepsis using the APACHE III system (4)) or to customize the model. For SAPS II the customization is applied to the equation regression, using the same SAPS II score. For MPM 1I each component of the model can be customized. Several Remarks :

Only one model is published to be applied at the ICU entry ( MPM II 0) before any therapy.

2 -Most of the published scores are calculated from the lsl ICU day. Applying these scores to the let day of sepsis could have a very different significance.

3 -Is the ACCP classification of sepsis useful at the bedside to select oatients ?

( The pathophysiology of sepsis involves a dynamic interaction between the host and the infecting microorganism(s). A multitude of biochemical and hemodynamic interactions occur which function in concert to determine the ultimate outcome of the septic episode. The complexities of the septic process preclude outcome analysis by in vitro systems or computer models, Animal studies have become indispensable in evaluating new therapeutic agents in the treatment of sepsis. These studies provide valuable information on safety, pharmacokinetics, relative efficacy and dosing strategies for potential therapeutic agents in the treatment of sepsis.

There are three basic types of animal models: (1) toxicity models with injection of bacterial endotoxin or exotoxins; (2) live or killed bacterial challenge models; (3) actual infection models. All models employ some form of external manipulation under controlled conditions to induce sepsis and to analyze potential therapeutic efficacy. These experimental requirements may limit the ability of animal models to accurately predict the clinical value of new agents in human sepsis.

Four major end points are analyzed in the various animal models: (1) hemodynamic parameters;

(2) modulation of host-derived inflammatory mediators; (3) resolution of end organ injury; or (4) lethality. Each animal model has certain advantages and limitations. Species-specific differences in physiologic and immunologic features make it difficult to directly extrapolate the results of animal experiments into clinical medicine. While no ideal animal model exists, consistent benefit of a new immunotherapeutic agent in multiple animal systems provides the most compelling preclinical evidence of its therapeutic potential in septic patients.

Consensus-assisted development of a study protocol on sepsis: an important difference from previous randomized trials

There have been several, and perhaps too many metaanalyses of cliuical trials on surgical infections, but their results have only rarely been incorporated into the design of new randomized trials in sepsis. Metaanalysis is retrospective. It seems more important to analyse and to criticize the design of tria/s before they are ongoing. Incorporation of that into development of a study protocol is the aim of the following procedure:

The time-schedule of about two years should include further cell culture and animal experiments after the first successful studies showing effectiveness. Exhaustive evidence for a dose-dependent cost-banefit and near complete explanations of the pathomechanism should not be awaited for development of the protocol of the clinical study. However, a discussion forum of experts should be performed in a recently published, almost perfect trial in order to see the frontiers of the present research in cell and molecular biology, clinical pharmacology, statistics and decision making. This should be followed by a metaanalysis of at least 10 study designs on sepsis with methods of formalized medical decision making (decision tree). Clinical algorithms -developed by objective methods including nominal group processes -should be developed to standardize any concomitant treatment in the centers considered for a multinentre trial. The latter is usually necessary in the field of sepsis.

Finally, a pilot study should be performed to demonstrate not only feasability, but to learn about all possible types of interventions which modify endpoints as response variables. Especially mortality is not free from such effects.

The time for the individual phases of this consensus-assisted process of protocol development have to be shortened by their temporal hiterloeldug. Due to the fast discovery of new chemical factors in SIRS it wilt otherwise be impossible to come up with results of randomized trials which are fascinating enough to be incorporated in daily routine treatment. Severity scoring systems are intended to provide a population-based estimate for the risk of an unwanted outcome resulting from some disease process. In the area of infections and the "septic" response, this has routinely been defined as death. For anti-infective trials, there has been a reasonable correlation of severity (APACHE ll) with outcome measured as persistent or recurrent infection, but this is not as robust as the association with mortality. It is expected that non-lethal failure would not immediately correlate with disturbed physiology: the basis for antiinfective failure depends on a variety of factors not measured in severity scoring, such as type of infecting organisms, density and susceptibility to administered antimicrobia[s, and anatomic features of infection (e.g., pancreatic). Severity scoring is important in differentiating these "categorical" variables from continuous (physiologic) variables. There are, additionally, a series of problems related to the cause of death in relation to severity scoring. In those tdals in which cause of death is analyzed, high severity scores did not routinely predict death as an immediate sequelae of sepsis/septic shock. Many of the predicted deaths occurred remotely of progressive background diseases. Other important problems with severity scoring have to do with lead-time bias: hew long patients were ill before entry into the study. Statistical techniques for condensing various categorical and continuous variables, as well as factoring in information requiring the pharmacodynamics of agents which affect outcome, are being developed to further refine deviation from predictive equations as an outcome measure. Stepwise logistic regression analysis has identified the presence of shock, intm-abdominal or unknown source of infection, hospital acquired infection, age greater than 70 years, neutropenia and inappropriate antimicmbial therapy as independent variables associated with increased mortality in sepsis. Maldistribution or" these variables is a concern in targe randomized clinical studies. It should be noted that only the selection of antimicrobiaI therapy could be addre.~'sed and altered at the time of enrollment in a sepsis clinics] trial. Inappropriate antimicrobial therapy has been noted in 41%-66% of septic patients. Well designed, randomized sepsis trials do not assure the absence of imbalance between treatment and control groups with regard to inappropriate antibiotic therapy, especially with subgroup analysis. Statistical manipulation to correct for any imbalance is appropriate but avoiding imbalance is ideal. Selection of antibiotics in septic patients requires such considerations as clinical setting (history, physical examination and underlying disease), appropriate laboratory tests (such as gram stain), identification of source and associated bacterial flora, antibiotic susceptibility patterns for that hospital and likelihood of resistant organisms. Determination of inappropriate antibiotic therapy is made retrospectively after culture results are available. In some circunmtances the organism and antibiotic sensitivities cannot be predicted, while in others it can. Leibovici et al identified hospital acquired infection, antibiotic treatment before a bacteremic episode, endotracheal intubetion and thermal trauma as independent variables which predicted isolation of a multiresistant organism. The same authors developed a predictive index for resistant organisms, which when compared to prescriptions of attending physicians, improved antibiotic selection in critically ill patients. Including a standard antibiotic regimen for all septic patients in a clinical trial is impractical. There is too much interhopsital variability in flora and antibiotic susceptibility. In addition, the aggressive broad spectrum coverage necessary for some patients is inappropriate for others. A protocol which guides a prescribing physician through a logical chain of reasoning might improve antibiotic selection in critically ill patients and could be included in the design of a clinical trial in sepsis. Although this approach might minimize potential for maldistribution of inappropriate antimierobial therapy, it would likely create multiple problem areas. Will the patient's physicians agree to the protocol choice? What is the availability of speciftc antibiotics at a participating hospital? What is the aplpiieability of study results in typical clinical situations? If there is agreement that thin is an appropriate antibiotic protocol, should it not be used in all seriously ill septic patients in the hospital? In other conditions such as ARDS, the success of protocol therapy is dependent upon measurable goals such as Pao 2 and highest plateau pressure, while measurable goals are not as clear in choosing antibiotics. Based on the above confounding issues, a complex and extensive algorithm covering many potential possibilities of error and specific antibiotics seems impractical; however, an algorithm focusing on three to four major common errors in antibiotic selection and dealing with classes of antibiotics is plausible for inclusion in sepsis clinical trials. In this session we will attempt to outline the methodology of such a future study, emphasizing the immense difficulties involved in its proper conduction including: design, selection of patients, surgical considerations, supportive treatment, the "human factor" and ~nd points.

Only a close cooperation between a few dedicated surgeons, obsessively studying a large number of well selected and stratified patients over years, could make such a study possible.

Algorithmis have frequently been mistaken for the graphic format in which they are presented. However, an algorithm is neither a flow chart nor a list of instructions; it is a step-by-step approach to solving a problem. Likewise, a clinical algorithm is a step-by-step approach to solving a clinical problem. CAs are deterministic, which does not mean that they are rigid.

They are practical rather than mathematical algorithms, that tan be used only with clinical judgment, and must be tailored to each patient. There are a number of types of, or uses for, CAs that can improve sepsis management, including: diagnostic or severity scorm CAs, a management CA for managing sepsis constructed according to recently published standards for use in teaching research protocol algorithms that must be used to collect data or guide implementation of a clinical trial aimed at validating one or more dicisions in the management CA; finally, protocol-style CAs drawn from the management CA and should be validated using matching outcome studies. 

Organisational problems peculiar to multicentre trials

The organisation of any randomised clinical trial is fraught with difficulties, and that of a multicentre trial particularly so. There are problems associated not only with the principles on which the investigation is based but also with the detailed organisation and analysis. Are all the participants truly unbiased about the relative merits of the regimens being studied? Is like being compared with like -are all the end points and the standards for measuring them clearly defined so that they cannot be misunderstood? Are all eligible patients being studied, and their results reported? Is truly informed consent being sought that will satisfy not only the patient's right to choose what happens to him, but also the law of the land? And, finally, how can the participants be sure that the trial will be stopped at the right time, to ensure that no treatment is given to any patient that might be harmful? These questions may seem insoluble, but until we tackle them we shall never be certain.

Dr.M. Ev~s, Scar~mu~ Hospital, Scarborough, NonhYo~s~ Y0126QL, U.K.

Increasing knowledge of the pathogenesis of sepsis and septic shock has led to the development of many new therapies and technologies capable of preventing, blocking or even reversing the deleterious cycle of events. Recent results of subsequent multicenter trials testing some novel therapeutic drugs, however, did not confirm the promising sometimes overwhelming effects obtained experimentally. It is postulated that this is largely due to inappropriate design and conduct of multicenter trials as currently performed. One of the shortcomings emphasized in this paper is the "treatment mix" problem.

Multicenter trials in sepsis are performed because no single center is able to recruit the necessary number of eligible subjects within a reasonable time. Each single center (ICU) has its own individual policy which cannot be standardized for the trial (treatment mix). Even if we ascribe that each ICU did the "best" for their patients, it is highly probable that the outcome differs between the ICUs. This translates to the effect of the new therapy under investigation. ]t is therefore worthwhile to consider that ICUs must first demonstrate a certain standard based on the patients' outcome (audit) before becoming accepted as a participating center. This prerequisite ensures comparable quality between participating centers, led to the reduction of "noise level" and increases the probability of positive study results. The RIYADH-ICU program based on standard mortality ratios is recommended as an audit instrument.

During the last decade most clinical trials of potential therapeutic agents in systemic sepsis have failed to demonstrate benefit from the drug under study. Although in many instances it is entirely possible that the agent in question does not have clinical efficacy, an equally plausible explanation for the multitude of negative results is that the studies were improperly designed and thus could not show therapeutic benefit, if it existed. Problems which have been identified in these various trials include inappropriate or unrealistic definitions of response to the drug; inadequate sample size, with attendant insufficient statistical power to demonstrate a difference, if it existed; insufficient standardization of customary therapeutic maneuvers in septic patients; imprecise criteria for patient entry into the study, which creates unacceptable inhomogeneity between control and treatment groups and invites unjustified post-study subgroup analysis; the lack of a formal sequential monitoring procedure for interim data analysis, which could potentially affect patient safety; and a primary analysis of study results that excludes protocol deviants and is not according to intent-to-treat. These and other problems have impacted upon the validity of the various trials conducted to date and may well have prevented the resolution of controversies surrounding the use of certain agents in the treatment of septic patients. The complexity of bacterially-induced septic shock involves a cascade of events that evolve over time, beginning with the proliferation of offending organisms and followed by the release of toxic mediators from the bacteria. Endotoxins [e.g. lipopolysaccharides) from gram-negative bacteria initiate septic shock by precipitating a systemic inflammatory response syndrome (SIRS) through release of a broad spectrum of hostderived mediators. Over the last century, hundreds of these endogenous mediators have been proposed to serve as pathophysiological substances in this "alphabet soup" of cytokines, eieosanoids, biogenic amines, kinins, peptides, ions and hormones. An evaluation of the literature on septic shock leaves the investigator with a sense that, although many pieces of the septic shock puzzle have been presented, no clue was provided to indicate how these pieces fit together, In an attempt to assess objectively the causal role of individual mediators, we recently completed a collective meta-analysis of 24 mediators that were individually reviewed and subjected to Koch-Dale analysis of causality by distinguished authorities in the field (1). In summary, our attempt analyze available evidence to support a cause-and-effect relationship for putative mediators of septic shock revealed that only eight of the 24 mediators analyzed could be strongly inferred as playing a causal role in septic shock. Evidence supported the involvement of endotoxln, endorphins, complement, interleukins, tumor necrosis factor, eieosanoids, platelet activating factor and calcium. Other mediators either lacked adequate supportive evidence or were not conclusively involved. Posttraumatic outcome may be reduced by an amplified stress response mediated by neural and humoral stimuli. The classical hormonal catabolic response is predominantly mediated directly or indirectly by neural stimuli, while most changes in inflammatory responses such as leukocytosis, acute phase proteins and changes immunofunction are mediated by humoral mediators. Clinical experience from controlled studies suggest afferent neural blockade and modification of stress response to improve organ function and postoperative outcome, and limited experimental studies also suggest neural blockade to be of beneficial value in shock states. Of special value may be the improved gastro-intestinal motility after neural (visceral) blockade. No outcome studies are available from patients with major truma, multiple organ failure or sepsis, conditions where humorai mediators may be more important to modify than neurally mediated responses. Future studies should investigate the clinical outcome effect of combined neural and humoral mediated blockade, as well as the potential role of an initial neural blockade in elective trauma (first hit) to improve the metabolic scene and outcome if a postoperative complication (second hit) should occur. Neutrophil elastase is the predominant protease of the human neutrophil. This enzyme, and a variety of other enzymes, including neutrophil collagenase, are released by activated neutrophils in the setting of high concentrations of reactive oxygen metabolites. In addition, in this extracellular environment, the normal antiprotease defense mechanisms of the organism may have been severely inactivated oxidatively especially along capillary endothelial cells. Accordingly, release of neutrophil elastase which mediates inflammation and degrades most extracellular matrix proteins, including elastin, fibronectin and many forms of collagen, may contribute to tissue injury and organ failure in a variety of diseases ranging in severity and acuity from pulmonary emphysema to the Adult Respiratory Distress syndrome (ARDS). As a result, significant efforts by the pharmaceutical industry have been directed not only toward the development of human neutrophil elastase inhibitors, but also toward their characterization in a variety of animal models of neutrophil mediated diseases, and their evaluation as potential therapeutics for the treatment of a variety of human clinical conditions. Data from animal models of organ injury and failure along with data collected from a series of patients at risk for ARDS and with ARDS will be presented and discussed as a rationale for inhibition of neutrophil elastase. Parameters that need to be monitored to obtain success in future clinical studies will also be presented in this context. have turned out to initiate and maintain organ dysfunctions in severe posttraumatic and postoperative courses. Such proteolysis-induced disorders are especially rendered possible by the concurrently arising consumption of inhibitory regulators (e.g. a Vproteinase inhibitor=a 1PI, antithrombin III=AT III) via cofnplex formation and/of proteolytic/oxidative inactivation.

Besides the well-known destructive potency against protein substrates, proteinases such as elastase or thrombin (active or in complex with the serpin inhibitors ~ 1PI or AT III) are also supposed to increase the inflammatory reaction by inducing cytokine release or shedding of soIuble adhesion molecules from various inflammatory cells (PMNs, monocytes/macrophages, endothelial cells). Those cell-derived mediators may provoke further cell activation through an autocrine/paracrine loop thus increasing significantly the proteinase burden at the inflammatory focus. Therefore, this noxious circle might be interrupted by a convenient proteinase inhibitor therapy to ameliorate the inflammatory process.

To verify this hypothesis, high dosis of AT III (mean plasma levels up to 140 %) were applied in first controlled randomized studies on surgical patients suffering from multiple trauma or severe sepsis. In control patients we could clearly demonstrate the sequential appearance of proteinase/serpin-complexes (Fa 1PI, TAT), cytokines (IL-8, IL-6), and soluble intercellular adh6sion molecules (ICAM, ELAN, P-selectin) in the circulation. These factors turned out to be a helpful tool for early diagnosis and prognosis of forthcoming organ dysfunctions. Interestingly, m AT III-treated patients a significantly reduced release/production of inflammatory mediators became obvious concomitant with the improved clinical course in comparision to an increasing deterioration in control patients. Soluble mediators of inflammation are cytokines (e.g. IL-1, IL-6 and TNFe) as well as breakdown products of complement proteins (e.g. C5a and terminal complement complex). Therefore, attenuation of both, release and generation of these synergistically functioning mediators together with stimulation of release of antagonists including soluble forms of receptors are potentially beneficial in all kind of inflammatory diseases. Intravenous immunoglobulins (IVIGs) are known to have an anti-inflammatory effect in humans (NIH Consensus Conference on WIG, 1990) . The mechanism of action becomes more and more clear. In single cells stimulated by anti-CD3 mAb WIG was able to down-regulate production of IL-2, IL-10, IFN% and TNFJ~ significantly. IgG coated solid-phase stimulates release from monocytes of interleukin-1 receptor antagonist (IL-lra) but not of IL-1; this observation is very much in contrast to the effect of endotoxin in the same in vitro system. Furthermore, naturally occurring anti-cytokine antibodies can be demonstrated in apparently healthy individuals and in WIG prepared from plasma pools. Some of these antibodies can be detected by antigen/antibody reactions in fluid phase (anti-lL-le, anti-IL-6), some only by solid-phase detection systems (anti-lL-8, anti-IFN% anti-TNFc 0. Infusion of WIG to patients suffering from inflammatory disorders have resulted in a decrease in IL-6 in circulation. During infusion WIG might induce an acute but transient rise of TNFo~, IL-6, and IL-8. Self limitiation of this process might be due to demonstrable instant release of IL-1 ra and soluble TNFo~ receptors. In vitro IVIG has also been shown to attenuate complement activation via the classical pathway. Furthermore, acid haemolysis of erythrocytes in paroxysmal nocturnal haemoglobinuria could partially be prevented by IVIG. IVIG was able to attenuate Forssman shock in guinea pigs. We have seen a patient with congenitally elevated turnover of alternative pathway of complement who repeatedly showed an increase of haemolytic titres after administration of IVIG. Thus, WIGs apparently have multiple potencies including attenuation of soluble mediators of inflammation and might therefore be of benefit in trauma, shock, and sepsis.

Significance of sinusoidal liuing cells and of humoral factors on hepatic function following sepsis and major surgery Hirata K, Katsuramaki T, Yamaguchi H, Mukaiya M, Yamashiro K. Regeneration of the liver after hepatic necrosis or partial removal has been extensively studied, but uncontrolled mechanisms to irreversible hepatic failure following sepsis or major surgery of liver or with hepatic dysfunction have so far not been well documented. We suggested the importance of the intrasinusoidal aggregation between sinusoidal lining cells and intrasinusoidal running ceils in the mechanisms of hepatocyte dysfunction. Thus, the purposes of this study was to investigate the changes in each sinusoidal lining cell during this process and the effect of cytokine on hepatocyte under various oxygenation.

[ Materials and Methods ] ( 1 ) Clinical study : Thirty two patients with histologically proved malignant tumor underwent major hepatectomy were devided in two groups, i_e. the cirrhotic liver group and the non-cirrhotic [Aver group. Most of the former were the patients with hepatocellular carcinoma and most of the latter were the patients with metastatic carcinoma of colorectal cancer. Five patients were complicated septic conditions with hepatic failure. Following operation in each patient, a verous blood sample was taken for measurement of serum IL-1, IL-6, TNF and C-reactive protein concentration. And hepatic biopsies were sometimes undergone in patients with anomalous hepatic function.

(2) Experimental study : Kupffer ceils, sinsoidal endothelial cells and hepatocytes were separated from mouse ( C3HHeN ) liver. And also polymorphonuclear cells and platelets were purified from blood and peritoneal macrophages were from peritoneal cavity of mouse. Co-cultures between or among these cells with Iipopolysaccharide were performed. Cytokine concentrations in media and hepatocytic function were compared.

[ Results and Conclusion ] Our findings in above experiments demonstrated the cleanly different concept for the mechanism of hepatocyte dysfunction following sepsis or major surgery. Namely, the effect of hepatoeytie function by cytokines or superioxide were significantly affected under low oxygenation, but not significant under good oxygenation. Hirata M.D.,Nishi 17,Hokkaido,Japan $120 470 immune Complexes as .Mediators of Sepsis and Immune Dyafumcffon laa K Horn, Chxistof Birkenmaier, and Greg H2mon Departmen~ of Surgery, San Frandsco General Hospital, UrdversR 3, of Calif(~rrda, San Frandsco.

Immune complexes (IC) a;:e commonly found circulating in parians during infection and sepsis; and following traumm, bums, and massive l~ensfusion. Clearance of 1C occurs by phago~'Tmsis of esll-bou_nd or opsonized p~tlcles. Monocyte activation du~Jng phmgocytos/s could lead ~o release of cytokilIes ieletexiot~ to the hosh To examine tb/s phenomenon, we formed insoluble IC by precipitating Iow-endotoxin bovine serun~ albumin (BSA) with rabbit ant/-BSA IgG.

We have ~town that these IC are ingested by monocytes and concomitantly stimv/ate re~pizatory bu~t activity measuxed by assay of in%racellul~ peroxldaffon. We have shown that IC ingesffon produces signLqcant el,era,lore in the monoey~e phenotype. Spedacally, CDI4 is down-regu/ated, along wl~ CD 32 and CD64. Thus responsiveness to lipopolysacchmdde (128) a~xd i=m~unoglobulin Fe st~mu/mtlon is reduced. In cert,+rest, the complement rote.p tot CDC42 is u~-regulamd, indicaling Mcremsed re~ponalx=~ness to opsonized pat,rotes, we examined monocyte superna~n~s for production of hzmor necrosis factor alpha (TNFg) following IC stimulation and showed that IC are capable of provoking ~elaase of hrmaunl)reac~ve TNI~. J[C also increase mono=yte producHon of prnstaglandin E2.

in summary, ~/~ese results demonstrate Lhmt IC are capable of indudng production of ¢ytokines and the imrau-nosuppressive prostaglandin PGE2. IC also allez ~he surface expression of important receptors involved in actlvaffon by LP~ and phagocyIosis. Thus, IC aze competent for indud/on of the mepl/e s!0otLse. By examining m0n0cyte phenotypms, it may be possible tO d~mrmkne extent oHn'u~u~e complex activation and predict immune depression iu criffcally ill pa~ents. A great deal of experimental work has focused on preventing and/or treating sepsis and organ failure following injury. Many of these strategies have tried to attack mediator substances released during the systemic inflammatory response following injury. Sugermann has demonstrated the improvement of pulmonary function, but not eappillary leak using [buprofen in the porcine model. In a clinical trial, Faist demonstrated the effectiveness of Indomethacine in amelierating the post trauma of cellular amenity using in vitro parameters. Morbidity and mortality, however, were not different between control and treatment groups. Baue und Chaudry have suggested that ATP magnesium chloride may have protective effect in renal failure uM Kupfer cell dysfunction seen after shock. Pentoxiphyllin has been found to be effective in improving tissue oxygenation and oxygen consumption following hemorrhage, and there may be promise for this drug in the ischemia reperfusion injury. Monoclonal antibodies against endothxin and inflammatory mediators seemed promising at first glance; however, preliminiary clinical data of the treatment of sepsis with either HA-IA or E5 monoelonoal antibodies against bacterial cell wall have been disappointing. Initial studies with IL-I receptor antagonist also appeared encouraging, but no difference was seen in the most recent trial.

Currently studies are ongoing with monoclonal antibodies to tumor necrosis factor which may have more efficacy given the fact that it is a macrophage mediator released by endotoxin. Although we seem closer to understanding and preventing the problems of sepsis and organg failure after trauma, we must continue to appreciate the complex interrelationships and delicate balances inherent in the host defense system. Overzealous attempts at restoring immunity in the absence of understanding pathophysiology may be just as dangerous at post-traumatic immlmospremsion. Severe acute pancreatitis was caused by different etiologic factors, but once pancreatic tissues were infected, patients show a similar pattern of aggravation and develop organ failure. Remarkable elevation of serum IL-6 was unexceptionally observed in patients with severe acute pancreatitis. The serum levels were more than 500 pg/ml (normal: less than 4 pg/ml) during the first one or two days after onset. There was a significant correlation between the peak serum IL-6 level and the number of organs showing failure (r=0.883).

TNF-a production by isolated peritoneal macrophages following lipopolysaccharide (LPS) stimulation was significantly increased in cerulein-induced pancreatitis rats. Serum TNF-a activity was elevated following intraperitoneal administration of LPS as the septic challenge both in pancreatitis rats and in control rats, being significantly higher in the former (p<0.05). Histological findings and liver function tests revealed that LPS induced more severe liver damage in pancreatitis rats than in control rats. These results indicate that increased TNF-a production by peritoneal macrophages in acute pancreatitis augmented LPS-induced liver injury.

Organ failure in severe acute pancreatitis could be caused, at least in part, by increased cytokine production.

It is clear now, that in MODS, organ injury is not directly due to exogenous factors, such as bacteria or toxins, but instead is largely a consequence of the host's own endogenously produced mediators. There is an extensive body of literature on the inhibition of mediator production, release and action in different cascade systems by glucocorticoids. They can serve as a host protection against overactive defense reactions if used adequately.

A new understanding of the regulation of cytokine synthesis, especially TNF, may lead to an insight of the conflicting findings between the benefits of glucocorticoid therapy in animals and its current failure in recent clinical trials. Glucocorticoid shares its role with other novel therapeutic approaches also shown to be successful only in experimental settings. A mete-analysis of steroids in sepsis, however, revealed a beneficial effect in the subgroup of patients with gram-negative infections. There is now a series of arguments to reevaluate the indication for steroid thera-py and/or prophylaxis in MODS and sepsis.

Honjo i-1-1, Kumamoto 860, Japan

IN SEPTIC SHOCK J. Briegel, M. Halter, H. Forst, W. Kellermaun, M. Bittl, K. Peter Imrodnetlea and Methods: Hydrocortisone (HC) at an infusion rate similar to the maximal secretory rate of cortisol in man improves hemodynamics in human septic shock (1). We hypothesized that the HC-induced hypercortisolemia prevents a harmful overshoot of immune reactions. To test this hypothesis we prospectively investigated 12 patients admitted to the ICU in refractory septic shock. After resuscitation (mechanical ventilation, fluid resuscitation, titration of vasopressors, and initiation of antimicrobial therapy) HC Was started with a loading dose of 100 mg/ 30 rain, followed by a continuous infusion (10 mggh). With hemodynamic improvement (dopamine < 2 gg/kg/min) the infusion rate nfHC was tapered over a period of 3 days. Phospbolipase A 2 (PLA2), C-reactive protein (CRP), and elastase-proteinase inhibitor complex (E-PI) were deierminated daily. Results: According to our previous results hemodynamics improved during HC infusien. After 4 days dopamine requirements decreased to less than 2 gg&g/min in all patients. This corresponded to an attenuation of the systemic inflammatory response syndrome (SIRS) indicated by the abrogation of fever and the decrease in heart rate and inflammatory mediators. PLA 2 activity and E-PI concentrations decreased to near normal values and CRP concentrations decreased as well. After cessation of HC infusion (between day 5 and 7) a reinforcement nfthe SIRS was observed despite cortisol levels within the normal range. This was first indicated by E-PI, followed by fever, increases in heart rate, CRP, PLA2, and finally by the relapse of septic shock in four patients on days 7,8 9, and 1O. A second infusion of HC again resulted in shock reversal and in a decrease in PLA2, CRP, and E-PI in the patients under study. Discussion: There is growing evidence that the endogenous steroid HC and proinflammatory cytokines integrate a complex immunoregulatory feec~ack circuit. The elaboration of TNF, IL-16, IL-6, and IL-8 is attenuated by-100 rag HC prior to endotoxin challenge in humans (2, 3) . Cytokines like TNF, IL-16, IL-6, and IL-8 are potent proinflammatory signals for acute reactants (--, CLIP), neutrophil degranulation (--, E-PI), and PLA 2 synthesis and secretion. Our data provide evidence that HC at an infusion rate similar to the maximal secretory rate of cortisol in man attenuates the systemic inflammatory response in human septic shock. To evaluate the feasibility of a nigh-dose regimen of antithrombin III (AT III) treatment in patients with multiple injuries regarding blood levels of antithrombin III and undesired side effcts and to analyse effects of the initibitortherapy on the systemic inflammatory response. Patients and methods: 20 patients with an ISS nf more than 28 points who could be treated with AT III administration within 180 rain after trauma were randomized to receive either AT III or placebo. AT III was given to reach an antithrombin III inhibitory capacity in plasma of 140% by using the following calculation: AT III to administer = (140% -AT III measured) x body weight. AT III was given every six hours for a period nf48 hours and on the 4th day. Patients were followed up throughout their stay in the ICU but at least 14 days. In this abstract the data of the first 10 patients are included. Results: The patients' median injury severity score was 41 points with a median age of 42 years. The patients were equally distributed between verum and placebo groups with respact to age , ISS, prehospital treatment, blood pressure, hemoglobin and AT nMevel on admission, and time from the accident until the test solution was given. The patients of the AT III group received 18,100 units of the inhibitor on average during the first 4 days. With our regimen, the AT III levels could be kept between 125% and 140% of normal, whereas control patients an AT HI concentration of 40% to 60% was observed. Patients with AT nI treatment required less red blood cells bur received the same amount of FFP and fluids. There were no differences with respect to platelet count, partial thromboplastin time, l~rothrombin time and prothrombim. We did not observe bleeding disorders or any other side effect with peak concetrations of up to 206% of normal. White blood count, PMN--elastase, Interleukin 6 and 8 and C-reactive protein tended to be lower in the AT III group. There were no differences with respect to renal and hepatic function parameters but the duration od mechanical ventilation was shorter in the AT II1 group (8.5 vs. 15.8 days) Conclusions: We conclude that our treatment protocol (a) was able to restore AT IIl levels to the desired range of 140%, (b) did not lead to coagulation disorders, (c) did not increase transfusion requirements, (d) did not reveal other undesired side effects, and (e) showed a tendency towards reduced posttranmatic inflammation and mechanical ventilation requirements Department of Trauma Surgery, 45122 Essen, Germany Antithrombin-lll (AT-Ill) acts as an inhibitor of thrombin, further proteases and the alternative pathway of the complement system. Inhibiting thrombin, AT-Ill works as a functional antagonist of PAF~ Therefore, in state of hypevolemic-traumatic shock, continuous supranormal serum -and tissue -concentrations of AT-Ill may be efficient to reduce the local posttraumatic reactions resulting from complement and PMN-activation. 20 patients with severe multiple trauma (7 treatment [AT-Ill] --13 controls [C]) were investigated.

Study cdteda were age > 16 and < 65 years, injury severity (ISS) > 30 points and Glasgow-Coma-Scale > 8 points; Randomization and treatment has to be started within 6 hours after trauma. Permission for the clinical study was given by the local ethic committee. Bradykinin (BK) and related kinins are potent inflammatory peptides which possess the ability to induce, vasodilation, increased vascular permeability and hyperalgesia. CP-0127, a novel homodimer BK antagonist has previously been shown to increase survival in rat and rabbit models of lethal endotoxin shock and is now in clinical trials for sepsis. We have now evaluated the effect of CP-0127 in other models of inflammation. Male rats were precannulated with a catheter in the carotid artery. 24h later BK was injected ia and the pain score ranked from 0 (no responses) to 5 (vocalization). CP-0127 at 3.6 umoles/kg completely inhibited the pain responses for a period of 2.5 -3h. CP-0127 at 3.6 umoles/kg s.c. was also found to inhibit the increase in paw volume and hyperalgesia induced in rats over a 3-4h period by an intraplantar injection of 0.1% carrageenan. The abdominal constriction response 1o an intraperitoneal injection of kaolin was inhibited in a dose-dependent manner by CP-0127. When 50 ul of 0.5% formalin was injected into the paw of a mouse a characteristic licking response was observed which was biphasic in nature. CP-0127 significantly inhibited both the first (0-5min) and second (15-30 min) phase responses. ]n a rat burn model, where the hind paw is immersed in water at 55°C for 15 sec the increase in paw volume was significantly reduced by pretreatment with CP-0127, 3.6 umoles/kg s.c. Finally cerebrai edema was induced in rats by applying cold (-78°C for 10 sec) to the dural surface following a craniectomy. CP-0127 at 3.6 umoles/kg s.c. produced a significant reduction in the amount of edema compared with sham controls 24 h later. These data suggest that BK is an important mediator of inflammation and hyperalgesia and that the bradykinin antagonist, CP-0127, may be useful in the treatment of such inflammatory, hyperalgesic disorders. Partial hepatectomy in humans is associated with a considerable morbidity due to hemodynamic and metabolic derangements, which increase the risk for organ failure and mortality. We hypothesized that endotoxemia may play a pivotal role in these complications. We therefore, investigated whether peri-operative infusion of rBPI23, a recombinant protein of the human neutrophil BPI with bactericidal and endotoxin-binding capacity, could prevent postoperative derangements following partial hepatectomy. Male Wistar rats (230-250 g.) received a 70% liver resection (phx) or a sham operation (sh), and a continuous intravenous infusion of either 0.2 mg/kg/hr rBPI23 (phx-bpi, n=8; sh-bpi, n=7) or the (iso-electric, iso-kD) control protein thaumatin (phx-con, n=8; sh-con, n-8). Various parameters were measured 4 h after the resection or sham operation. Mean arterial pressure, cardiac output and heart rate were significantly decreased in phx-con rats compared with sh rats, which effects were not observed in phx rats treated with rBPI23. Blood pH was significantly decreased in the phx-eon group, whereas the leucocyte count, hematocrite and IL-6 levels were significantly increased compared to sham levels. In the phx-bpi group, these parameters were restored to near sham levels. In vitro experiments with rat plasma and human mononuclear cells (MNCs) revealed that plasma of phx-con rats is highly capable of activating MNCs, accompanied by the release of cytokines. This activation is attenuated with phx-bpi plasma. In vitro added aCD14 or polymyxin B was able to reduce the activation by phx-con rat plasma to the levels of phx-bpi rats Thus, these data suggest that systemic endctoxemia, possibly of gut origin, is a major cause of postoperative hemodynamic and metabolic derangements following phx and that rBPIzz can prevent these changes. More recently we reported a transient appearance of both endotoxin and TNF in the circulation of rats subjected to the haemorrhagic shock (HS) already at 90 -120 rain. Similar to BPI, recombinant BPI was found to bind LPS and inhibit TNF formation in vitro. The aim of this study was to investigate the effects of rBPI21 (kindly provided by Xoma corporation, Berkeley, CA) against haemorrhage related endotoxemia and mortality in rats. Method: A prolonged HS was induced by blood withdrawal to a mean arterial pressure of 30-35 mmHg for 180 rain followed by reinfusion of shed blood (SB) and resuscitation with two times of SB volume of Ringer's lactate over 50 rain. rBPlg. 1 was administered at a total dose of 5 mg/kg i.v. (2.0 mg/kg at the16-eginning followed by two doses of 1.5 mg/kg each at end of shock and the end of resuscitation). The control group was treated similar to the BPI group but received thaumatin as a protein control preparation at the same dose as rBPI21. Results: Imrffe?diately after resuscitation (230 min) the detected plasma endotoxin levels in the control group (mean = 61, range = 0-120 pg/ml) were almost neutralized by rBPI21 treatment (mean = 13, range = 0-53 pg/ml) . Plasma TNF levYIs were not significantly influenced by rBPI treatment at the two time points 230 and 320 min of experiment (means: 65 and 51 in BPI vs 49, 65 pg/ml in the control group). The 48-hour survival rate was improved from 6/16 (37.5 %) in the control to 11/16 (69 %). Conclusion: These data suggest that haemorrhagic shock may lead to bacterial translocation and/or transient endotoxemia with concomitant cytokine formation that may play an important role in the pathogenesis after shock and trauma, rBPI21 might be a useful therapeutic agent against endogenous bacterFal/endotoxin related disorders in hemorrhagic shock. Morbidity and mortality after hypoxia of the vital organs had been correlated to the production of oxygen radicle which is mediated by xanthine oxidase activity,

In this study we have evaluated the survival rate after allopurinol. 42 rabbits weighed 900 + 200 grams divided into two groups. Group I included 22 Tabbits were treated with allopurinol 80 mg/kg for seven days before induction of haemorrhage. Group II as a control included 20 rabbits. All rabbits were subjected to 25% arterial blood loss through the central ear artery for one hour then resusciatation was done by the heparinized withdrawn blood through a marginal ear vein. During the experiment blood pressure and heart rate were monitored through the central ear artery. ALso uric acid, lactic acid, glutathione activity were estimated. Animal survival was followed for 10 days. Postmortem vital organ histochemistry and histopathology examinations were done. In group I the survival after three days was 10 out of 22 while in group II it was two out of 20. Our conc|usion, allopurinol had increased the survival in aiiopurinol pretreated rabbits which may indicate the value of allopurinol premedication for patient prepared for elective bloody surgical intervention . H2 receptor antagonists are commonly used for stress ulcer prophylaxis, but their actions on the septic response are largely unknown, In an experimental model, 26 pigs were first anesthetized, then injured with 100 joules of energy to the posterior thigh, then hemorrhaged 30-40% of their blood volume. After I hr of shock, all the shed blood plus 2x the hemorrhage volume as lactated ringers was infused. Following resuscitation, ranitidine (1.5 mg/kg iv twice daily) or saline placebo was begun. The treatment group was randomly assigned in a blinded fashion. After 72 hrs, a septic challenge was administered (15 Bg/kg of E. coil endotoxin (LPS)). Serial gastroscopy, gastric pH, hemodynamics, ABG's, physiologic dead space ventilation, leukocyte counts, and tumor necrosis factor (TNF) levels were recorded for 180 min. Baseline values and units were cardiac index 102 _+ 9 ml/min/kg (CI), arterial PO2 228 + 11 mmHg(PaO2), base excess 6.5 -+ 1 meq (BE), physiologic dead space fraction 23 +_ 4% (PDS), and TNF 0.4 + 0.1 units/ml. Baseline gastric pH was 3.9 -+ 0.4 and 4.8 _+ 0.5 in the placebo and ranitidine groups, respectively. The gastritis following hemorrhage was marginally attenuated in the ranitidine group. Following LPS infusion the following were obtained: CI PaO2* BE* Gastric* PDS* Peak* 120 rain 60 rain 120 rain pH 120 min TNF Ranitidine 167_+17 118_+15 -4.4±1. Bum injury results in hypermetabolism, fever and nitrogen wasting. Endotoxin (LPS) has been proposed to mediate these effects, either directly or via activation of macrophages to produce cytokines such as interleukin-6 (II-6). This study was designed to clarify the role of LPS and 11-6 in the metabolic response to bum injury. Twenty-five burn patients (49 -+ 17%; 16 + 15% FT BSA burn; 32 _+ 16 years old) were studied serially for three weeks post bum. Patients underwent partitional calorimetry to assess metabolic rate and compartmented heat loss. Nitrogen was assayed using chemiluminescence.

LPS and I1-6 were measured with Limulus Amebocyte Lysate assay and ELISA. Patients were excluded if they suffered smoke inhalation, showed any sign of sepsis or failed to rapidly meet their nutritional needs via the enteral route. Ten patients received intravenous Polymixin B (5,000 U/kg/day to bind LPS). These patients did not differ for the remainder. All patients were hypermetabolic and febrile in proportion to the size of their bum wound but were not endotoxemic (3.1 +_ 5.6 pg/ml; normal <40 pg/ml). I1-6 did demonstrate a significant correlation with cole temperature (Tr~ = 37.6 + 0.21ogi1-6, P=0.04) and with nitrogen excretion (Nou t = -5.2 4-0.091ogi1-6 + 0.14TR, P=0.01). Administration of polymixin B had no effect on metabolic rate, temperature or I1-6 levels but did reduce nitrogen excretion resulting in more positive nitrogen balance (0.t grn/day vs. -0.1 gm/day, P=0.04). Although bum injury does not produce an obligatory endotoxemia, I1-6 does appear to play a role in the fever and nitrogen wasting seen with such injuries. The effect ofPolymixin B on nitrogen excretion suggests that LPS may play a role either locally or in the portal system. Introduction: There is substantial evidence that release of inflammatory mediators by activated Kupffer cells contribute to the course of a systemic inflammatory process, e.g. after shock or lrauma. Besides the systemic effects of mediators such as TNF, PAF or interleukines, local actions on hepatic microvasculature and hepatic inflammatory response have to be considered. Our aim was to assess the role of TNF and PAF by blocking their effects using anti-TNF monoclonal antibody, pentoxifylline and a PAF antagonist. Methnds: In anesthetized SPRD-rats, hemorrhagic shock was induced by withdrawl of arterial blood within 5 rain and shock state was hold for 1 h at a MAP of 40 mm Hg (cardiac output of 50 %). Following adequate resuscitation with 60% of shed blood and twice of this volume as Ringer's solntion, animals recovered to MAP >1130 mm Hg and CO >120%. Hepatic microcirculation and sinusoidal leukocyte-endothelium interactions were examined by intravital epi-fluorescence microscopy at 1, 3, or 5 hours after resuscitation. In a blinded fashion, a rat-specific monoclonal anti-TNF antibody [2 mg/kg, Celltech, UK) , pentoxffylline (PTX, 50 mg/kg, Hoechst, D), and a PAF antagonist (WEB 2086, Boehringer, Ingh., D) were given either as pretreatment or at the time of resuscitation (n=6-8 group Bolla. K*., Duchateau, J., HajOs, Gy., Mbzes, T., Hern~di, F.

Prevention of temporary/secondary immune deficiencies or reduction of their severity and/or duration as well as the reduction of the perifocal inflammatory processes belong to the rational targets of posttraumatic/pedsurgical medication. Such a targeted medication can result in less frequently occurring nosocomial infections, and in reducing the duration of the intensive care and convalescence period. The results of in vitro studies performed with the amino acid sequence 32-35 of thymopoietin, i.e., with thymocartin in whole blood and peripheral mono-nuclear celI(PBNC) cultures clearly show some characteristic effects of this immunomodulator. Preincubation with the tetrapeptide significantly (p<O.Ol) and concentration-dependent reduced the endotoxin(LPS)-induced TNFalfa production from 95,7 to about 20 pglml, but did not abolish it. The inhibition of the LPS(0.1 ug/ml)-induced TNF production occured already in presence of 0,1 pg peptide/ml and peaked at the concentration of 0,01 ng/ml. Higher concentrations did not increase this effect. In contrast to the other two small thymic peptides (TP-3 and TP-5), thymocartin did not enhanced the PWM-induced PGE 2 production in PBMC cultures up to a concentration of 1 ng/ml, inclusive. The selection of thymocartin (TP-4) for perisurgical medication and/or for treatment of trauma-induced temporary immune depression has been strongly supported also by targeted animal experiments. Thus, mortality of about 90% registered in mice infected with Pseudomonas aeruginosa after bum injury could be reduced with the tetrapeptide treatment to a level which did not differ significantly from that of the infected non-burned controls. Moreover, observations gathered first of all with the inflammatory model of air poach/croton oil grenuloma (SELYE) clearly showed, that a very definite, antiexudative effect can be achieved with this thymic peptide. As to this effect, thymocartin has been confirmed, e.g., to be equipotent with locally administered fluocinolone acetonide and/or s.c. injected prednisolone.

The interim results of two explorative double-blind clinical studies (involving more than 260 patients after polytrauma and hip-fracture untill now, respectively) correspond to the expectations. The treatment significantly reduced the occurence of nosocomial infections, the days spend in intensive care and on antibiotics as well as it shortened the bed-out time and reduced the additional medication. Objective of this prospective randomized study was to further delineate the mechanisms leading to these alterations and to investigate the effects of immunomodulatory intervention. Patients and Methods: 20 patients (pts) formed control group A. 20 group B pts received 3x50 mg Indomethacin (Indo) from the first (dl) to the fourth postoperative day (d5) intravenously to block prostaglandin E2induced suppression of CMI response as well as 50 mg Thymopentin (TP-5) subcutaneously preop., on d2 and d4 to augment CMI reactivity. In vitro investigations preoperatively, on dl and d7 included measurements of the percentage of CD4+ T-helper (TH) cells, pbytohemagglutinin (PHA)induced synthesis of interleukin (IL)-2 as one cytokine primarily produced by THl-cells, and PHA-induced synthesis of IL-6 as one cytokine primarily produced by TH2-cells. Delayed type hypersensitivity (DTH) skin response to an antigen skin test battery and specific antibody (AB) production following tetanus vaccination preoperatively and on d7 were used as in vivo tests for THl-and TH2-induced immune response respectively. Results: Postoperatively, group A pts showed a persistent significant reduction of CD4+ TH cells, IL-2/IL-6 ratio, and DTH skin response as compared to baseline values (BL), while AB production increased (p<0.05 vs.BL). In group B pts no change in CD4+ TH cells, IL-2/IL-6 ratio, or DTH skin response was observed (p<0.05 vs.group A), and postoperative AB production increased significantly (n.s. vs. group A). Conclusions: 1.These results clearly indicate a significant suppression of TH1 induced CMI response, while TH2 induced response remains normal following OHS. 2.These alterations may gain clinical significance whenever a postoperative infection requires a TH1 response and may explain the susceptibility to opportunistic microorganisms observed in pts after OHS. 3 The aim of this study was to find out whether the establishment of administration of Thymostimulin (TP-I) in jaundiced and anergic rats would return the rats to the state of immunocompetence. Material and Methods. Male Wistar rats weighing 250-300 g were injected with haemocyanin (KLH). All rats with a positive response were included in the study and we evalu ated the T cell subpopulations and IL-2. They underwent laparotomy and the common duct was ligated and divided. Those rats whic become jaundiced and anergic one week after the operation, were divided in two groups: control group ( Objective of study. The goal of this study was to prove the necessity of thymus derived immunomodulators treatment in acute period of infantine sepsis, Materials and methods. Immune status was investigated in 11O infants. Clinical conditions of children were defined by syndrome of multiple polyorgan]c insufficiency.

The heaviness nf condition was estimated as the sum of points according to efforts needed to restore of basic living functions -respiration, hemodynamJcs, hemocoagulatlon, metabolism, functions of liver and kidneys. It was expressed in form of clinic condition classes from 1 to 4.

Results.It was found that in 100% of infants with 3-4 classes of heaviness immunode[iciency took place simultaneously with II and III stages of hemocoagulattve syndrome. Amounts of T cells and T helpers were decreased more then 2 times, ratio Th/Ts was reduced to 1.5 or lower. Concentration of lgG and IgAwas reduced almost a hull in comparison to control. Phagocytic and metabolic activities nf neutrophils were depressed. Complement system state was changed: the functional activity of the alternative pathway factors B, D and the level or C3a subcomponent were increased but the level of C2, C3,, C4 and C5 components of the classic pathway were decreased. During the development of coagnlopathy the direct correlation between CHSO, levels of plasminogen and antithrombine III was found. In greater extent the function of immune system was disodered in accordance with 2 point metabolic violations and hemocoagulative disturbances considered as disseminated inmtravaseular clotting syndrome on II or IlI stages.

Conclusions. The rehabilitation of the immune system functions was promoted by taetivine and thymaline in doses of I-3 mcg/kg per day and 2 mg/kg per day accordingly during a week in children with 3 -4 classes el heaviness. AIter 11-14 days tactivine had advantage. When eoagulopathy took place thynaalJne promoted restoration of (1-3)-~-D-glucan (BGC) is a major cell wall compornent of pathogenic fungi and shows many immunopharmacological activities on experimental animals.

LPS is also derived from gram-negative bacteria and have in~ttnomodulating activities on immune systems positively or negatively. Because of difficulty to cure contaminating LPS from BGC preparations, it is very hard to evaluate the exact activities of BGC.

In this study, we compared the immunological activities of LPS ans highly purified BGC on murine system. Materials and Methods: C3H/HeN or C3H/HeJ mice, 6 to 12 wkolds of both sexes were used through experiments.

Highly purified (1-3 )-{5-D-glucan (Gurdran; BGC)was courteous gift from Seikagaku Kogyo Co. Tokyo, Japan and resolved in endotoxin-free ultra-pure water. Escherichia cell LPS was purchased from Sigma Chemicals, USA.

Mitogenic activities of BGC or LPS on splenic cells were measured by MIT dye assay. In vitro activation of peritoneal exudate macrophage (PEN) have moniotored by phagocytosis and intraphagocytic killing of Pseudomonas aeruginosa (Z~B-139.

Induction of cytokine productions from PEM or spleen cells induced by BGC or LPS were detected by cytotoxic assay. Results and (bnclnsions: I ) In vitro cultures of PEM with BGC have induced enhanced phagocytic and bactecidal activities in C3H/HeN and C3H/HeJ mice, whereas LPS only induced such activities in C3H/HeN mice.

Most effective dose of BGC was 10 ug/ml.

2) BGC-dependent proliferative response of C3H/HeN splenic cells was not detected under the culture condition used.

3) Induction of TNF~ prodction was apparently observed in BGC-stimulated C3H/HeJ PD4 cultures, but not in LPS stimulated one.

These findings indicate that activation mechanisms of Pr94 and splenic cells by BGC are different from that by LPS. To investigate the BSC-induced P~4 activation, intracellular signaling pathways of P~4 by BGC-stimulation are under consideration.

Yamagami K, Uedono Y, Kawakami K, Kitazawa Y and Tanaka T According to the latest reports of use of IL-2 in a burned-sepsis model, the dose was too high to adjust for human therapy. Adoptive transfer of L~ymphokine-activated killer (LAK) cells has been demonstrated as a means of enhancing therapy in malignant tumors. We therefore attempt to use LAK therapy on burned-infected mice instead of IL-2. Under anesthesia, 8-weekold male C3H-Hej mice were subjected to a 20% scald or sham burn and then resuscitated. Sham burned mice served as controls. Scald burn mice were subjected to peritonitis by intraperitoneal innoculation of 105 organisms of P. aeruginosa 24 hr after burn. Burned-infected-mice were divided into two groups. One group had no further treatment, and the other group received LAK cells (2x105) intraperitonelly after septic challenge. The mice were scored for survival daily. On day 7 after burn, Using Moabs dual-color flow cytometry, we studied lymphocyte expression in mice with use of blood and spleen. Simultaneousely we studied the alteration of IL-1 and IL-6 in their serum using immunoassy kits. Sham burned animals had no mortality. The mortality of the LAK-treated mice was significantly reduced when compared with that of the burned-infected mice. The most consistent changes observed were depressions in percentages of Thyl, 2 positive cells and a remarkable depression of NK cells in spleen. After LAK cell treatment, those two percentages of cells significantly increased. All the data of assay of ILl and [L 6 were not detected on sham burn mice serum. Due to burn and septic challenge the levels of ILland IL 6 (n=6) were raised to 45.9_+17.4 and 3866_+1045 pg/ml, while the levels in LAK treated mice were reduced to 4.23_+2.37 and 437_+110 pg/ml, respectively. The results reported here demonstrate that LAK therapy is able to improve cell-mediated immunity in burned-infected mice. This is associated with a significantly improved survival rate, since ILl has been demonstrated to mediate immune and inflammatory responses. Also a cause effect relationship between elevated endetoxin levels and increases of [L6 has been recently established. The aim of our present study was to investigate the effect of parenteral indomethacin on the immune system of patients under major operations.

Our study included 20 operated patients, 10 of which received 100mg indomethacin before and 1,2 and 3 days after surgery (group A) and 10 patients received no antiinflammatory therapy (group B).

We measured total T-cells helper (CD4), suppressor (CD8), NK-cells and interleukin-1 and TNF production by PHA or endodoxin (LPS) stimulated peripheral blood mononuclear cells at day 0 and 1, 3 and 7 days after operation.

In group A we detected a significant (p<0.05) increase in CD4/CD8 ratio on day 3 while no differences were ~oted in NK-cells or cytokine production in both groups (Table 1) . It seems, therefore, that parenterai indomethacin may have a benefitial effect on the immune system of patients under major operations by increasing the T-helper /T-suppressor cell ratio while having no effect on the production of cytokines by peripheral blood mononuclear cells. This study was undertaken tD dete~nnine the role of prostaglandin synthesis inhibitor on survival post traulna sepsis.

Analysis of the effects of prostaglandin synthesis inhibitor on survival of four groups of 15 mice each were made after laparatomy and intravenous administration of live E.coli.

Post trauma sepsis, Group A received no treatment; Group B received antibiotics IM; Group C received prostaglandin synthesis inhibitor IM and Group D received both antibiotics and prostaglandin synthesis inhibitor IM.

Survival time was longest in group treated with prostaglandin synthesis inhibitor and antibiotics post trauma sepsis.

The singular use of either antibiotic or prostaglandin synthesis inhibitor did not alter the outcome on survival.

Mortality rate was lowest in the group treated with combined prostaglandin synthesis inhibitor and antibiotics. Use of this combination showed a reduction in bacterial growth in peritoneal and blood samples, mild morphologic changes secondary to sepsis in the heart, lungs, liver, kidney and stomach and marked enhancement of mean level of activity.

The study indicates that addition of prostaglandin synthesis inhibitor in treatment of trauma-sepsis has clear beneficial effects. Interferon-? (IFN-y) is a potent immunostimulant which has been shown to be detrimental when administered during septic shock. Blockade of this cytokine however is beneficial during the acute septic response. The aim of this study was to evaluate the cellular effects of this cytokine and its blocking monoclonal antibody (MoAb) in lethal sepsis following injury. 100 female CD-1 mice were randomized into two groups: Septic=LPS vs injury=hind limb amputation (HLA vs IgG. nStats= ANOVA=p<0.05 vs lgG IFN-y was detrimental when given to control septic animals. However its blocking MoAb was protective (p<0.05). Conversely IFN-y was proteetive in injured septic animals compared with anti-IFN-? (p<0.05). These findings demonstrate that the immune stimulant IFN-? has therapeutic potential in the immunosuppressed injured host predisposed to sepsis, while blockade of this cytokine is required for protection in the septic host with a normal immune response.

Dept of Surgery, RCSI, Beaumont Hospital, Dublin 9, Ireland.

Oxidants play a role in physiology. The potential noxious oxidant biochemistry is restrained by chemically distinct antioxidants. These antioxidants, which may reside in different cellular compartments, can act synergistically. In normal physiology an oxidant/antioxidant bai&nce exists. Unbalance in favour of the oxidants is called oxidative stress. Oxidative stress has been implicated in many pathologies including lung diseases(e.g, doxorubicin induced cardiotoxicity), ischemia/reperfnsion damage and central nervous system trauma. Transition metals such as iron are involved in the early events leading to oxidative damage in these pathologies. This has been underlined by the finding that postischemic CNS pathology closely resembles that caused by a chemical oxidative insult (e.g. iron microinjection). Moreover, a protective effect of oxygen-radical scavening agents or compounds that inhibit lipid peroxidation (antioxidants) in brain ischemia/trauma is apparent. Some known drugs (e.g. anti-arrhythmics or histamine H2-antagonists may act as non-specific antioxidants. Howe~er, recently, interesting new membrancespecific antioxidants (e.g. 21-aminosterdids) have been designed. Subt&e effects on iron redox chemistry contributes to the potent antioxidant activity of these 21aminosteroids. Since a few years, one area that greeted enthusimastlcally in clinical medicine is that of reoxygenatlon injury or ischemia-repeffusion, a phenomenon accepted to make a majo:" contribution to organ dysfunction in trauma, shock and sepsis through the formation o( oxygenated free radical species. However, thei detection in vivo always remains a difficult challenge. Efforts have been made recently to directly establish the formation of free radicals in biological samples as complex as blood, plasma or tissue with the use of electron spin resonance (esr) spectroscopy. Using spin trapping agents, the presence of reactive carbon-and oxygen-centered radicals has been demonstrated in animals blood submitted to heart, renal and intestinal ischemia-repeffusion or to liver transplantation. Recently, the in vivo formation of free radicals in the cerebra; extracellnlar fluid during cerebral isehemia-repeffusion has been assessed by the spin trapping technique coupled to brain microdialysis. Esr investigations have also been conducted to detect endotoxin-induced free radical generation in rat or baboon liver and 

Oxygen free radicals (OFR), produced both at intra-and extra-cellular levels, seem to play a major role in the pathophysiology of tissue injury and organ dysfunction in sepsis, through induction of lipid pemxidation in cell membranes. Oxidative damage can result both to an OFR-overpmduction or to a depletion of endogenous antioxidant defenses, which are represented by scavenger enzymes (e.g. SOD), hydrosoluble and liposoluble antioxidants (e.g. ascorbate, vitamin E), and other mechanisms such as metal-ion sequestration. In animal models, oxidative damage has been proved by detection of increased levels of lipoperoxidative products, and a depletion of antioxidant defenses was found both in plasma and tissues. In the few clinical studies a similar results have been reported, and oxidative damage appears to correlate with DIC, with organ dysfunction and with red blood cell deformability. In a group of critically ill septic patients we found increased levels of conjugated dienes (CD) (bioproducts of lipoperoxidation) and decreased plasma levels of alpha-tocopherol and coenzyme Qlo (CoQI0)(endogenous liposoluble antioxidants). A relationship between CD and uric acid (p<0.004) was found, may be related to xantine-dehydmgenase to xantine-oxidase conversion. Antioxidant depletion is due both to overconsumption and to a nutritional deficiency, even if a reduced synthesis can also be present. In fact one more relationship between CoQI0 and plasma cholesterol (p<0.0005) demonstrates the commun hepatic biosynthetic defect. Is there any role for antioxidative therapy in sepsis? Can it achives a significant improvement in the septic course, organ dysfunction, and final outcome? Several antioxidant drugs, have been evaluated in experimental and clinical sepsis: scavenger enzymes, natural antioxidants, alloporinol, 21-aminosteroids or lazaroids, have been proved to reduce lipopemxidative damage, to protect organ dysfunction, and in some cases to improve survival. Several questions must be addressed in evaluating safety and utility of antioxidative therapy. More specific and standardized methods must be applied to detect and quantify the lipoperoxidative damage. Antioxidant drags must act selectively on OFR-production, without interference with other than that are beneficial for the organism. Antioxidants must be able to achive in adequate amount the tissue or cellular compartment where their action is required. Many antioxidants have also a pro-oxidative effect which can he detrimental in some conditions. The timing of administration, the dosage used and the association wih others antioxidant drugs must be defined. Nutrition plays an important role as antioxidative therapy, since it supplies all compounds needed to maintain adequately levels of endogenous antioxidant or to support their synthesis. Thiobarbituric acid reactive substances (TBARS) concentration in serum has been determined in a large population of healthy subjects and in patients suffering acute hepatitis and chronic cases of hepatitis C, as well as several other processes. The correlation with different physiological and clinical parameters in these populations is also presented. Treatment with interferon of the chronic active hepatitis C patients, 5x10 U three times a week during two months, led in those patients whose SGPT activity normalized in serum, to a concomitant decrease in serum TBARS content. The possible theoretical involvement of peroxidation and antioxidants in this beneficial effect of i~terferon in hepatitis C patients is discussed. The results presented confirm the value of TBARS as laboratory test in the management of disease and as a useful tool for the study of pathogenic and/or therapeutic mechanisms. 4-Hydroxyalkenals are among the different substances measured by the TBARS assay. These compounds show several biological effects that have been demonstrated mainly in different experimental models. We have studied the capacity of different tissues to conjugate these compounds 

Patients surviving the initial subarachnoid hemorrhage (SAH) from a ruptured in~raeranial aneurysm are prone to develop cerebra] ischemia from vasospasm which is associated with significant morbidity and mortality. Among the many treatments that have been prol~osed to treat this condition, none has been shown to be satisfactorily effective. Recently, a new drug, namely the 21-aminosteroid Tirilazadmesylats, has been studied in a large, prospective, multicentor trial for its effectiveness to improve the outcome of patients with aneurysma! SAId. The study was conducted in Europe and Australia in 41 neurosurgical centers and included 1023 patients. All patients received presently accepted standard treatment, including nimodipine. Patients were randomized to four different groups: placebo and 8 groups of tirilazad in escalating dosage (0.6 -6 mg/kg/day). The clinical outcome was compared for these 4 groups and cerebral a~giography was performed in 75 % of patients on day 7 -11 follovang SAH, to study the effect 0f'the dflf~ off Cerebral vasospasm. The study reached its planned endpoint 6 months ago. Preliminary results ~how a significant improvement with regardto mortality in those patients receiving the highest tirilazad dose as compared to placebo and the two lower dose groups. Additional beneficial effects were seen in the 6 rag/day groups with regard to the occurence of symptomatic vasospasm and good clinical outcome. Although the final analysis of all data is still pending, these results suggest that tirilazad-mesylate could represent a major improvement for the treatment of patients with SAIl. Experimentally lipidperoxidation products (LPO) are increased in acute pancreatitis (AP) due to oxygen radicals (OR). The purpose of this clinical study was to determine the antioxidant status and LPO in patients suffering from AP and tO evaluate the effect of antioxidant treatment with sodium selenite (SE) thus improving the function of the SE dependant glutathione peroxidase which is the major detoxifying system for OR. Materials and Methods: In Z0 patients sufferin s from severe AP (C-reactlve protein >120 mE/l) we determined on day 1 and day 8 after admission the LPO ma!ondialdehyd (MDA), conjugated dishes (CD), reduced (GSR) and oxidized (GSSG) glutathione, the vitamins A,C,E and SE. Moreover the patients were evaluated clinically using the RANSON and the APACHE II score. I0 patients were randomly treated with 600 ug/day of SE for 8 days. Results: All patients suffered from a severe depletion of antioxidants,especially a low concentration of SE (only 1/3 of normal). Thereby the increase in LPO correlated with the clinical course. During SE treatment LPO decreased and the levels of antioxidant vitamins improved. SE had no influence on leth-slity the lenl or the chan in RS or AP II. Background: Since reperfusion injury occurs when oxygen is reintroduced into ischemic tissue, the ideal timing for administration of therapeutic compounds aimed at ameliorating oxygen radical mediated injury is at the time of initial fluid resuscitation. Currently used colloid or crystalloid preparations do not provide optimal, or even significant, anti-oxidant protection. Systemic iron chelation affords protection against the iron catalyzed components of oxygen and lipid radical mediated tissue injury. The conjugate resulting from chemical attachment of the clinically approved iron chelator, deferoxamine (DFO, Desferal ®, Ciba), to hydroxyethyl starch (HES) represents a novel approach to colloid based fluid resuscitation. HES-DFO contains 85% HES and 15% chemically bound DFO. The polymer-drug conjugate has a lower molecular weight than that of HES in order to allow more rapid excretion. Results: Preclinical and initial clinical trials indicate that HES-DFO is well tolerated, even at high doses. In animal studies, fluid resuscitation with HES-DFO does not significantly improve central hemodynamic recovery beyond that observed with HES, but HES-DFO seems to afford better protection of microcirculation in organs at risk (lung, liver and gut), possibly by decreasing neutrophil sequestration. In a burn model, total fluid requirements are lower and oxygen utilization higher in HES-DFO treated animals compared to HES controls, suggesting decreased vascular leak and improved tissue perfusion. Conclusion: HES-DFO represents a means by which potent antioxidant protection can be administered at resuscitation. Iron has been suggested to play a pivotal role in oxygen flee radical mediated tissue injury. In vitro experiments indicated its critical role as a katalyst in hydroxyl free radical generation 0Fenton-reaction). Since iron chelator deferoxamine administered in shock alone demonstrated severe side effects, a hydroxyethylstarch (HES)daferoxamine (DFO)-conjugute was used to modulate oxygen free radical injury during the ischemia/reperfi~ion syndrome induced by hemorrhagic shock. Methods. Female LEWIS rats (190-215 g, n>6; pentobarbital anesthesia 50 mgjkg), in hemorrhagic shock ( The aim of the study was to elucidate (1) whether the generation of OR would affect lung and kidneys as primary shock organs in the very early phase of sepsis and (2) whether DFO-HES could prevent this tissue damage. Methods: In 67 rats sepsis was induced by cecal ligation puncture (CLP) peritonitis. The animals were randomly assessed to 2 groups: one group was treated with 3 ml DFO-HES (100 mg/kg iv), the other rats received solely 3 ml of the carrier starch solution. 30, 60, 120, and 240 min after induction of sepsis respectively, the animals were sacrificed, the organs collected, and tissue contents of glutathione (GSH), malondialdehyde (MDA), myeloperoxidase (MPO) and conjugated dienes (CD) determined. Plasma samples were obtained for analyses of endotoxin (chromogenic LAL test). Blood pressure (MAP) was measured via a carotid artery catheter. Results: CLP caused sepsis with high (> 1.84 EU/ml) endotoxin levels. MAP in both groups decreased slightly but significantly during sepsis regardless any treatment. In the lungs MPO concentration was increased (p<0.05) in the lIES group already 30 min after sepsis induction. Concomitantly, tissue GSH level decreased and lipid peroxidation was pronounced as shown by elevated MDA and CD levels. DFO-HES diminished tissue PMN accumulation and MPO concentration. Moreover, at each time point lung MDA and CD levels were lower (p<0.02). Histomorphological examination showed marked micro-atelectases, destruction of the alveolar septa, and splicing of the basal membranes in the lIES group. In contrast, in DFO-HES treated rats the alveoli remained well-ventiIated and only some enlarged reticular fibers without splicing were observed. Almost similar results were found for the kidneys. MPO levels differed neither within nor between both groups. The slight decrease in GSH levels seen after 60 min in the DFO-HES group seems to demonstrate an oxidative stress to a lesser degree. The most impressive effect of iron chelation, however, was revealed by the lipid peroxidation products. At each time point, MDA and CD levels were lower (p<0.02) compared to the HES group. Light and electron microscopic examination disclosed tubulotoxic and mitochondriat damages while DFO-HES lxeatment prevented that alterations. Conclusion: Both the biochemical and histological results of this study reveal an early and remarkable generation of OR in peritonitis-induced sepsis. Thereby, these OR obviously cause pulmonary and renal tissue damages, intravenous application of DFO-HES may, however, benefit by preventing early lipid peroxidation of the tissue. The proteolytic irreversible COnversion of xanthine dehydrogenase (XD) to xanthine Oxidase (XO) is triggered by calcium flux. The aim of our study is to clarify ~he link between intracellular Ca2+ levels and XO activity determined by uric acid release, and to evaluate the efficacy of verapamil, on the generation of hydrogen peroxide associated with reperfusion by assaying lactate & pyruva~e release and the levels of cytosolic free NAD /NADH ratio. Experimental protocol consisted of :(A) Non ischemic/reperfused experiment in which normal cardiac slices of rats were perfusated with oxygenated kreb's Ringer phosphate buffer containing glucose (150mg%) and bovine albumine (5gm%) for 60min at 37°C.It composed of 3 groups, group Aa (control group), and groups Ab & Ac (perfusate supplemented with verapamil in the dose of lOO&200 mi% respectively).

(B) Ischemic reperfused experiment in which ischemic cardiac slices were obtained from rats subjected to 15 min ~aemorrhage.lt was also divided into two groups; group Ba and Bb (verapam~/ 200mi% added to perfusate}. Verapamil stimulated uric acid release from normal rat cardiac slices were 267% in group Ab and 767% in group Ac(dose related). Rates of uric acid release is enhanced by verapamil in group Bb. Moreover, rates of uric acid release in groups Ac & Bb are insignificant. In verapmil added groups (group Ab, Ac & Bb), increase uric acid release is associated with an enhancement in pyrurate release and with increase levels of cytosolic free NAD+/NADH ratio, although it is not evident ~ ischemic group (group Ba).It is concluded that the conversion of XD to XO is calcium independent. Eicosanoids like thromboxane A2, leukotriene B4 and leukotriene C4 are known as promoters of initial inflammatory reactions. We investigated whether oxygen radicals (OR) are able to induce a release of these eicosanoids in whole blood.

Blood from healthy volunteers was incubated with xanthine oxidase/hypoxanthine to generate oxygen radicals. After 0,5,15,30 and 60 minutes plasma levels of thromboxane B2 (TxB2), leukotriene B4 (LTB4) and leukotriene C4 (LTC4) were determined via ELISA technique. Another volunteer had taken 1000mg Aspirin one day before taking the blood sample (no7). Results:

TxB2 plasma levels increased from 14pg/ml at 0min to 144pg/ml, 350pg/ml, 370pg/ml and 660pg/ml at 5,15,30 and 60min (p<0,01) . LTB4 and LTC4 plasma levels showed an increase during the first few minutes (LTB4: 0min: llpg/ml,5min: 87pg/ml; LTC4: 0min: 23pg/ml, 5min: 97pg/ml (p<0,05)) followed by a decrease to normal values at 15min. In the sample no7 the cyclooxigenase-pathway was completely inhibited, the TxB2 plasma-levels did not alter at all, whereas LTB4 and LTC4-plasma levels weren't affected. 

OPALLOGENEIC BLOOD TRANSFUSION Jane Shelby, Ph.D., and Edward W, Nelson, M.D, There have been numerous investigations dudng the last two decades examining the effect of surgery, anesthesia, blood loss and transfusion on vadous immune parameters in humans and animal models. There appears to be concurrence among several well controlled studies that transfusion of whole blood (containing leukocytes), has regulatory effects on immune ceil function which include decreased cell mediated immune response, and inhibition of IL-2 secretion. These effects occur following transfusion alone and in con.cart with the distinct immune effects of surgery, trauma and anesthesla,

The clinical consequences of this immune modulation by transfusion include decreased allogeneic response to transplanted organs, which has been exploited clinicelly in renal transplant patients. Additionally, there is evidence for a strong association with increased risk for infection in transfused patients following surgical procedures. AIIogeneio blood transfusions have been shown to inhibit cellular anti.bacterial mechanisms, causing increased susceptibility to bacterial pathogens, in humans and in animal models. There is also concern that allog~neiC transfusion may adversely affect cancer patients, resulting in decreased disease-free survival.

Several stategies have been proposed to minimize the adverse effects of blood transfusion. There is evidence that the risk of immune mediated infectious complications associated with transfusion may be greatly minimized wlth the use of autologous blood and leukocyte free allogeneic blood.products in surgical and trauma patients, it also appears that the inhibition of cellular immune response and IL-2 productiorl following atlogeneic blood transfusion may be mediated by increased prostaglandin E2 secretion, and that immune response may be preserved in allogeneio whole blood transfused subjects receiving c3Lc~oxygenase inhibitors such as ibuprofen. Among these are various alterations in immune function. Efforts have therefore been made to utilize alternatives to homologous transfusions. These include the use of autologous predonation, supplemental iron therapy, and recombinant human erythropoietin.

Although initially considered innocuous, these therapies are now recognized to have potential deliterious immune sequelae. Erythropoietin, by its ability to lower serum iron levels, can impair both lymphocyte and NK cell activity.

Autologous donation impairs NK cell function.

Finally, supplemental iron therapy can stimulate bacterial growth and increase the rate of infectious complications.

This talk will present a discussion of these factors as well as a weighting of their importance.

R.L Rutan, RN;BSN, Shriners Burns Institute and The University of Texas Medical Branch, Galveston TX, USA The serious sequelae of homologous blood transfusions have resulted in vigorous efforts at identifying alternate therapies for correcting red blood cell (RBC) deficits. Erythropoietin (EPO) was hypothesized to exist in the early 20th century, however the protein was not isolaled until 1977. The human gene was identified and cloned in 1983, which permitted the production of EPO through recombinant techniques. The earliest clinical trials were performed in anemic end-stage renal failure palients on hemodialysis. Treated patients experienced increases in erythropoiesis with normalization of hematocrit and hemoglobin levels, cessation of lrans-fusion requirements and improvement in general wellbeing. These studies, however, identified side effects of EPO treatment such as hypertension, seizures and Ee deficiency. Volunteer trials have established that the hypertension is not a direct pressor effect but rather the result of abnormally rapid increases in red cell mass in the face of the incompetent volume-controlling mechanisms of the end stage renal failure patient. Lower doses of EPO and the subsequent gradual increases in red cell mass are associated with significantly lower incidences of hypertensive complications of EPO therapy. Likewise, seizure activity is not the result of a direct epileptogenie effect but parallels the incidence of hyper-tensive-related sequelae during high.dose EPO treatment. In cross-over designed studies, pre-existing iron deficiency has been demonstrated to decrease or negate stimulated erythropoiesis but effective-hess can be restored with appropriate Fe supplementation. Exogenous EPO is effective whether given by IV or SQ routes and dose response curves do not vary with route of administration. Increases in RBC mass are directly related to the dose of EPO, both in amount and frequency of administration although there is a 3-5 day time lag between the first EPO dose and laboratory indications of its action (i.e. increase in the number of reticulceytes in peripheral Wood). EPO is currently labelled for use in the treatment of anemias associated with end-stage renal disease and AIDS. However, its use in the surgical population has been explored because of its unique direct dose-response, EPO has been used to effectively increase the blood harvest amounls in autologous pre-donation, significantly increase hematocrils in children following thermal trauma and successfully increase red blood cell mass following essential surgical procedures in patients with religious aversion to transfusion.

BY BLOOD TRANSFUSION IN COLORECTAL CANCER SURGERY MM Heiss MD, Ch Delanoff MD, R Stets MD, J Hofinann, E Faist MD, KW Jauch MD, FW Schildberg MD Allogeneic blood transfusions are associated with an increased risk for postoperative infections in colorectal surgery when compared with autologous blood transfusions. Attribution of this effect to immunomodulation was suspected in our previous study (Lancet 1993; 342:1328-33) . Task of the recent investigations was to analyze which specific effector systems were affected in-vivo by this transfusion-associated modulation.

For global in-viva assessment of cell-mediated immunity (CMI) multiple recall skin-reactions were applied prior and post-operative. The specific humoral immune mechanisms were investigated by applying Tetanus-Toxoid one day preoperatively and deterimnating the quantitative IgG-response. For indication of macrophage stimulation in-vivo TNF-levels were determinated by bioassay.

DTH-responses were significantly suppressed (p<0.05) in patients receiving allogeneic blood (n=36) or operated without blood transfusions (n=17). DTHresponses were not suppressed and tendentiously increased in patients with autologous blood transfusions (n=24). In contrast, specific IgG-levels increased sigmficantly (p<0.05) in patients receiving allogeneie blood (from 1.6 + 4.4 to 9.7 _+ 18.4 IE/ml) whereas in patients receiving autologous blood a smaller increase (from 1.9 + 4.1 to 4.0 + 6.7; p=0.068) was observed. TNFlevels demonstrated a similar pattern with a higher increase in patients receiving allogeneic transfusions (l 1.4 + 12.0 to 28.5 + 11.4 u/ml) compared to those patients with autologous blood (8.7 + 12.5 to 17.1 + 23.0).

In conclusion these data indicate that allogeneic blood transfusions lead to a remarkable macrophage/RHS stimulation. This is corroborated by the boostered humoral IgG-response which was initiated before onset of surgical trauma and blood transfusion. Concerning CMI this caused a substancial suppression probably due to a stimulated secretion of immunosuppressive monokines. Objective: Firstly, to analyse the concentrations of the cytokines tumor necrosis factor (TNc), interleukin-1 (IL-I), interleukin-6 (IL-6) and coagulatioo/fibrinolysis parameters in postoperatively retrieved blood from a surgical area, secondly to characterize the correspanding cytokine patters in the patients and thirdly to study cytokine concentrations in the initial portion of drainage blood from a surgical area. Materials and Methods: Blood retrieval was performed in a closed-loop system without anticoagulant during 4-6 hours after surgery in 10 patients undergoing arthroplasty (9 hips and 1 knee). 7KF, IL-1, It-6, thrembin-antithrombin complexes (TAC) and antithrombin (AT) ~ere determined in shed blood. Patient plasma TN v, IL-I and IL-6 concentrations ~ere analysed at the beginnLqg and end of the 4-6 hour blood retrieval period. In a separate study (5 hip arthroplasties) f~F, IL-I and IL-6 ~ere determined in the initial portion of drainage blood. Cytekine analyses ~re performed usiog ipmuooassays. An omidolytic method was used for AT determinaf.ion and TAC was analysed by ELISA. N~n-poram~tric tests was used for the statistical comparison. Results: The patient plasma IL-6 coocemtratiems rose from a median value of 0 to 116 pg/ml, p<O.01, during the blood ret.rieval period. IN c was detectable in 4 patients, range:15-50 pg/ml. IL-I was not detectable in the patients. In retrieved blood TAG was >200 mg/ml in all samples (ref:<4.1 mg/ml) and AT was 0. 16-0.51 units/ml (ref:O.80-1.20) . The IL-6 concentrations in retrieved blood was >1000 pg/ml in all samples. TN v or IL-I was not detectable. In the separate study, (n=5), characterlzing eytokine content in the initial portiere of drainage blood, IN= (range: 11-277 pg/ml) and IL-I (range:5-125 pg/ml) ~re present in all samples but II-6 (range:O-25 pg/ml) was detectable in o.qly one semple. Conclusion: Theses findings indicate that hypereoagulability and hic~ ccrcentratioos are present in retrieved blood. The cytokine pattern in the initial portion of blood from a surgical area differed from these observed in retrieved blood and in the systemic circulation. To identify the role of both autOlogous and homologous blood on postoperative infections in elective cancer surgery. Materials and methods: 159 patients with colo-rectal cancer submitted to curative elective surgery were prospectively studied. On hospital admission the following nutritional measurements were assessed: serum level of albumin, cholinesterase, delayed hypersensivity response , total lymphocyte count and weight loss, as were age and sex, Duration of operation , operative blood loss, amount and type of blood given, pathological Dukes' stage of the disease and the attending surgeon were also recorded. Results : Eighty-four patients (52.8%) were perioperatively transfused. Thirty-six (42.8%) patients were given autologous blood , while 48 (57.2%) received homologous blood. No patients received both autologous and homologous blood. Twenty eight (17.9%) patients developed postoperative infections. Non transfused patients had a 9.3 % infection rate , those receiving autologous blood had a 13.9 % infection rate, whi]e in the homologous blood group the infection rate was 33.3% (p < 0.001). Univariate analysis showed that infections were significantly related to operative blood loss (p<0.01), length of operation (p<0.05) blood transfusion (p<0.05) and attending surgeon (p<0.05) . Multivariate analysis identified homologous blood transfusion as the only variable related to the occurrence of postoperative infections , while the other variables failed to reach statistical significance. Blood transfusion (BT) remains an essential life-saving treatment for surgical patients. However, besides the beneficial short-term impacts, negative longer-term effects are observed, which include various alterations in the immune responsiveness. In surgical patients these alterations may contribute to the increased risk for infections and cancer recurrence. Since relatively few data demonstrate immunologic changes occurring in other lymphoid compartments than blood after BT, we studied the effect of ET on the frequency and responsiveness of immune cells in bone marrow (BM), spleen (SPL) and blood (B) in a rat model. Normovalemic, 2 month old rats were transfused intravenously with syngeneic heparinized venous blood (3x2ml, every other day), and 3, 7 and 14 days after the last transfusion BM cells ( LEH is an experimental oxygen-carrying resuscitation fluid. Since LEH is cleared from the circulation primarily by the MPS, its effect on the development of sepsis and the nature of its relationship with the MPS remain a major concern. Preliminary in vivo data from our laboratory failed to show any LEH effect on the hemodynamic and hematologic responses to endotoxin lipopolysaccharide (LPS) in the rat. In contrast, LEH exacerbated the LPS-induced TNFa production and early mortality. The exacerbation of early mortality by LEH was attenuated by pretreatment with the TNFu synthesis inhibitor rolipram. Ex vivo, peritoneal macrophages from rats treated with LEH and LPS have shown increased IL-lg mRNA signal as compared to LPS alone. Also, LEH increased TNFtx production by peritoneal macrophages in response to LPS stimulation in vitro. Additionally, recent pilot studies indicate that LEH attenuates PMA-induced superoxide production from rat peritoneal macrophages and that LEH augments fMLP-induced migration of human monocytes. Taken together, these data strongly support possible interactions of LEH with the MPS and therefore the nature of such interactions should be further explored. Over the last decade, we have developed liposome encapsulated hemoglobin (LEH) as an artificial oxygen carrying fluid, or blood substitute. Our efforts have focused on studies to define the safety and efficacy of this resuscitative solutions. LEH consists of distearoyl phosphatidylcholine, cholesterol, dimyristoyl phosphatidylglyeerol, and alpha tocopherol in a 10:9:0.9:0.1 mole ratio and can encapsulate hemoglobins of different origin (bovine, human, recombinant human). LEH is fabricated using hydrodynamic shear to create an average particle size of 0.3 microns. LEH can be lyophilized using disaccharides and stabilized in the dry state and easily reconstituted before administration. Histopathology and clinical chemistries indicate that LEH rapidly accumulates in tissue resident macrophages in small animals injected in the tail vein, principaI1y in the liver and spleen. The consequences of accumulation in the reticuloendothelial system are manifest by transient increases in liver transaminases (AST, ALT), bilirubin, and BUN over 24-48 hours with no change in biliary function (GGT, aP) . Clearance through the liver and spleen is observed over the course of 1-2-weeks. More recent attention has been focused on secondary consequences of LEH administration especially with regard to inflammatory eytokines. LEH does not elicit expression of tumor necrosis factor in vivo and in isolated macrophage cultures, but does result in a transient increase in serum IL-6. We have also examined the interaction of LEH with LPS in vitro macrophage culture to further understand how this blood substitute may effect the immune system. We have labeled LEH with technetium-99m (99mTc) to study the biodistribution of LEH non-invasively in anesthetized rabbits. Rabbits were infused with a 25% topload of LEH (200 mg of phospholipid, 2.5 g of hemoglobin per kg of body weight) and imaged continuously with a gamma camera. At 20 hours, images were again acquired. Animals were then sacrificed and tissue counts obtained, images revealed an initial rapid uptake bythe liver, 8% at 30 minutes and 15% by 2 hours. The spleen accumulated activity at a slower rate, 3% at 30 minutes and 7% at 2 hours. At 20 hours, autopsy biodistribution studies revealed that approximately 42.6% of the dose is in the blood pool, 15.4% in liver, 18.1% in spleen, 3.2% in lungs, 2.4% in muscle and 1.6% in urine, with trace levels in kidney, brain and heart (<1 °/o). In a hypovolemic model, rats were 10% or 50% exchange transfused with 99mTc-LEH. In the 10% exchange model, 99mTc-LEH was rapidly taken up by the liver and spleen with minimal activity in the circulation at 20 hours. With the 50% exchange, 50% of the LEH was in circulation at 20 hours.

The interaction of LEH with platelets labeled with indium-111 was also studied. After infusion of LEH, the labeled platelets rapidly moved from the circulation to the lungs and liver. Over the next 30 minutes, the platelets gradually returned to circulation. This effect was not seen with Iiposomes of the same lipid composition but containing no hemoglobin.

Non-invasive imaging is proving to be a very useful tool for the investigation of LEH. The need for a safe, efficacious and commercially viable blood substitute is unequivocal. Of the several strategies pursued to invent an adequate blood substitute, Liposome Entrapped Hemoglobin (LEH) has been already established as a leading possibility. Major advances in liposome technology have already resulted in liposome preparations compatible with clinical use for drug delivery.

Recent technological advances made by the U.S. Naval Research Laboratories resulted in the capacity to entrap hemoglobin into liposomes in a way which secludes hemoglobin from interacting freely with biological systems. The LEH produced has already been tested in in vivo systems and was foun.d to be well tolerated. Moreover, the LEH originally produced as a solution can be transformed into a lyophilized form which can be reconstituted and delivered as a fresh solution. While important milestones in LEH development for a practical blood substitute have been achieved, several issues remain to be explored. Most notably, the long term consequences of LEH on host defense mechanisms and, in particular, immune cell function. In addition, it is important to understand more fully the metabolic fate and repercussions of LEH delivered at clinically relevant dose/schedule regimens. Finally, while LEH is a highly promising strategy for a blood substitute, the present formulations consist of human hemoglobin derived from human blood, To improve the safety profile, a recombinant preparation for liposome entrapment will be much desired, Aa-ginine, a semi-essendai dietary amino acid, possesses several unique and potentially pharmacologic properties. Argirdun is a potent secretagogue for pituitary growth hormone and prolacfin and for pancreatic insulin and glueagon; it modulates host protein metabolism by increasing nkmgen retention and enhancing wound collagen synthesis. It also is a potent T call function regulator. AIt of these effects coupled with its relative lack of toxicity and safety make it an a~antive nulritionai pharmacologic agem (t).

Rodents fed supplemeutal arginine exhibit increased thymSc weight which is due to increased numbers of thymic lymphocytes present in the gland. Thymic lymphocytes from animals fed supplemental ar~e demonstrate increased blastogenesis in response to Coma. and PHA (2) . Peripheral blood lymphocytes from humans given supplemental arginine also have heightened mitogunic responses to mitogen or antigens (3) . In postsurgery padents supplemental arginine abrogates or diminishes the deleterious effects of trauma on lymphocyte responsiveness and restores peripheral blood lymphocyte responses much faster than observed in controls.

Overall host immunity is also enhanced by arginine. Allograft rejection is enhanced and septic animals survive longer when given supplemental arginine (4) . Tumor bearing urginine-supplemented animals have decreased tumor growth and enhanced survival (i). Lastly, asgmine can induce T cell maturation and T cell mediated responses in athyrnic nude mice.

Arginine also has remarkable effects on host nitrogen metabolism post-injury. In increases nitrogen retention in healthy human volunteers and in surgical patients. This beneficial effect on overall nitrogen metabolism is accompanied by a unique effect on the healing wound. Supp]emental arginine increases wound collagen synthesis which also translates into increased wound breaking strength (5) . Arginine has no effect ou epithelialization.

Douglas W. Wilmom, M.D.

Boston, MA

Gintamine is the most abundant amino acid in the body, but it has long been considered a nonessential amino aeid because it is synthesized in many tissues. FoHov~g st,~'vation~ injury or infection, skeletal muscle pmteln inoresses its net tale of degradation and releases amino acids into the blunds~mm at an aocelerared rate. App~o)~mately one-third of the amino nitmgea is ghitamine, which is metabolized by the kidney where it parth:~pates in acid-base homeostasis, is the primly ~ for lymphocytes, mac~optmgcs and untexocyms, and contm'butcs to the synthesis of giumth~une.

Olmamine degrades slowly while in ~olu~ou, especially at usual room teml~mtums. Because giulamine was considered nonessential, it has beer absent R'om nil intravenous and most Gluts.mine should be considered a cendittona]ly essential nutrient for individuals with serious ilinesses, uspccially those confoanded by infcctinn and inflammation. Over the uc~:t 5-7 years, glutamine will be incorgorated into most feeding formulas designed for patients with critical illness.

526 O]~GA-3 PUFA There continues to much interest in the application of the 0mega-3 PUFA in clinical nutrition. The basic principle has been that the 0mega-3 PUFA will displace arachidunic acid and result in a decrease in eic0san0id production. In addition these changes in PUFA will after the physical characteristics of the membrane including flujdity, receptor function and transmembrane signals. Animal studies have shown that there is omega-3 incorporation with continuou~ enteral feeding both in control and endotoxic animals within 3 days. This includes the liver, spleen, circulating and alveolar marc0phages and the lung. This incorporation resuls in significant changes in the eicosan0id production including PGF 2 and ket0-PGFlalpha. There is improvement in the cardio-vascular reep0nse of these animals with ~ecreamed lactic acidosis and improved cardiac contractility. As well there is improved immune function with improved T cell response to mit0gens. The ~ of a mumber of pharmacological agents blocking cicosanoid production can enhance the cell effects of 0mega-3 PUFA. Clinical studies using short term entsral nutrition with 0mega-3 either alone or with other enteral supplements in a number of clinical settings have shown significant 0mesa-3 incorporation and decreased eicosan0id production. These positive results must be discussed with the additional evidence that long term omega-3 supplementation decrease eic0san0id production but als0 induce a state of immune suppression that is capable of increasing transplant sunvival. These 10ng te~ inune effects may benefit clinical conditions including rheumatoid arthritis and Cr0hn' disease Early enteral nutrition instituted i~mediately afte~ injury will decrease the entry of bacteria into the intestinal wall and decrease the number of bacteria that translocate into the portal blood. These reductions are associated with & decreased catabolic response, decreased plasma cortisnl levels, end decreased VMA excretion in the urine and prevention of mueosal atrophy. SDecific nutrients also affect the transloeation process. Addition of arginlne to the diet significantly improves the ability to kill translocated organisms.

However. translooetion across the gastrointestinal barrier is not affected. In contrast, glutamine diminishes the rate of translooation across the Imtestinal barrier and also improves killing of the beetarla that do translooate. The omega 3 fatty acids in the form of fish oil slightly decrease the rate of translocation but more significantly increase the ability of the animal to kill translo~ated organisms, All three dietary additives, i.e. argini~e, glu=amine and fish nil. significantly improve survival, hut adding glyoine or medium chain triglyeeridem do not, Combinations of srginine and glutamlns, glutamine and fish oil, and fish ell end arginine each improve survival, and to a greater degree than a combination of all three. These studies add further evidence that translocation is an important determinant of survival after injury, Early feeding with immunonutrlent enriched dices will improve survival and dsarease transloeation to varying degrees, depending upon the nutrients provided. OBJECTIVES: We studied effects of supplementing a commercial enteral diet, Impact R (IMP, Sander Nutr lnc), with fiber (IMP/FIB) or alanyl-glutamine (IMP/AG, exogenous glutamine (GLN) 14 gms/L) on influencing the incidence of BT to mesenteric lymph nodes (MLN) in burned mice. Fiber has been shown to improve GI integrity under certain stress/treatment conditions. The dipeptide AG is a water-stable source of GLN, which is a specific fuel for many cells including enterocytes. TraumaCal (TRCAL), a high-protein, high-fat enteral diet (Mead Johnson Iuc), was also studied, as well as Rodent Chow (Harlan Teklad Inc), which contains very high protein & fiber. METHODS: Anesthetized CF-1 mice aged 8-12 wks received 32% TBSA fullthickness dorsal burns & were resuscitated with 2 cc ip saline. Diets were allowed ad lib; caloric intakes were comparable in all Gps except fasted Gp (fast 24 hrs, Chow 24 hrs). At 48 hrs postburn MLN were sterily removed, homogenized and plated on heart brain infusion agar; CFU/g MLN tissue were determined. BT was analyzed by Fishers Exact Test, CFU/g by ANOVA-Bonferroni. * P<0.01, ** P<0.05 compared to IMP and Burn-Fast Gps. Background. Infectious complications following trauma, major operation, or critical illness adversely affect hospital cost and length of stay (LOS). Some key nutrients have been shown to possess immune enhancing properties. This multicenter trial was conducted to determine if early administration of an enteral formula supplemented with arginine, dietary nucleotides and fish oil can decrease LOS and infectious complications in ICU patients. Methods. This was a prospective, randomized, double-blind study of 326 adult ICU patients who required enteral feeding for > 7 days. Patients entered the study within 48 hr of the event, were stratified by age and disease, and were randomized to receive either the supplemented formula (IMPACT®) or the conventional formula (Osmolite ® HN). Feedings were initiated at full strength and advanced to at least 60 ml/hr by 96 hr after event.

Results. Both groups tolerated administration of formula well. For patients fed > 7 days, the median LOS was 25% shorter (p=O.OL) for the--supplemented group (21 days) compared to the conventional group (28 days). The incidence of most infectious complications was lower in the supplemented group, but this difference reached significance only for urinary tract infections (p=O.O05). The supplemented group had a significantly shorter LOS from onset of infectious complication until discharge for patients with pneumonia (19 vs. 27 days) and skin/soft tissue infection (19 vs. 37 days). Conclusions. Administration of the supplemented formula was safe and well tolerated. When fed > 7 days, it reduced the incidence of most infectious complications, and significantly reduced LOS. Materials and Methods: Twenty-seven patients were randomised into 3 groups ( n= 9 each) to receive either a standard enteral formula, the same formula enriched with arginine, RNA and omega 3 fatty acids (enriched group) or isonitrogen, isocaloric parenteral nutrition. Early enteral nutrition was started within 12 hours following surgery (10 ml/hour). It was progressively increased reaching a full regimen on day 4. On hospital admission and on post-operative day 1 and 8, the following parameters were assessed: serum level of transferrin , albumin , prealbumin, retiool binding protein (RBP), cholinesterase. Delayed hypersensitivity response, IgG, IgM, IgA, lymphocyte subsets and monocyte phagocytosis ability were evaluated on admission and on post-operative day 1, 4, 8.

The three groups were comparable for sex, age, cancer stage, type and duration of surgery, intra-operative blood loss and amount of blood transfused . In all groups a significant drop in all the nutritional and immunological parameters was observed on postoperative day 1. Comparing post-operative day 1 versus day 8 a significant increase of prealbumin (p<0.02) and RBP (p<0.005) was found only in the enriched group. With respect to immunological variables an increased phagocytosis ability (p<0.001) and a significant recovery in delayed hypersensitivity response (p< 0.005) was observed only in the enriched group. Conclusions : These data are suggestive for a more effective post-operative recovery of both. nutritional and immunological status in cancer patients fed with enriched enteral formula. Gastrointestinal intolerance was equivalent (18% in each group) and laboratory screening confirmed that both diets were safe. When analyzing clinical outcome for all patients, there were no significant differences in septic complications (Immun-Aid = 41% vs Vivonex TEN = 35%), mean MOF score (Immun-Aid = l.B vs Vivonex TEN = 3.6), or mortality (Immun-Aid 0% vs Vivonex TEN = 6%) .

Kowever, when analyzing the subgroup of patients with severe injury (ISS or ATI _> 25), patients receiving Immun-Aid appeared to have fewer septic complications (33% vs 50%) and their mean MOF was significantly lower (1.8 _+ 0.7 vs 5.8 + 1.7, p = 0.05, Student's t-test) . These preliminary data indicate that Immun-Aid is tolerated well when aggressively delivered immediately postinjury.

The ultimate affect on clinical outcome appears ~avorable for Immun-Aid, but needs to be confirmed in larger patient groups.

Kemp?n, M., Neumann, H.A., He I[michh B: As both increased, normal and reduced phagocytic capabilities of polymorphonuclear leukocytes (PMN) and monocytes in acute batterial infections have been reported, the role of phagocytes in patients with severe sepsis is less clear.We examined PMN and monocytes from 10 patients in septic shock and 17 heaiLhy votunteers for phagocytic function. Phagocytosis was determined by flow cytometry (FACScan) and was measured by the ability of PMN and monocytes to phagocytose E.Coli marked with fluorescent antibodies. A septic shock was defined by the presence of a ~ource of i, nfoctiQn with a known bacteriology, distinct signs of a systemic response and defined minimum scores in 3 ICU scoring systems indicating the presence of a multiple organ failure. Additionally we examined how phagocytosis is influenced when a new enteral diet formulation containing substrates suggested to improve immune function or Arginine, one of its major compononts, is added in vitro in defined concentrations and incubated for 10 minutes. PMN (p{O,05) and monocytes (p<O,01) of the septic patients showed a significantly enhanced phagocytosis compared toLhe phagocytes of the healthy group. The incubation with the enteral diet in vitro was followed by a higher dose-related rate of phagocytosis, especially in the healthy group, that was not statistically significant. After addition of L-Argintne we atso observed an iqcrease in phagocytosis, that was lower than in the group with the complete diet. Using a modern immunefluorescence-cytometric essay we founfl an improved phagocytic function in septic shosk. There are signs for an influence of nutrient substrates on phagocytosis which seems to be complex and should be further investigated. ARG and GLN were lower in the 49%-BCAA vs 16%-BCAA group; ARG was related directly to ARG dose and inversely to BCAA dose (p<O.O01 for all).

Clearances of BCAA increased with increasing BCAA dose (p<O.O01); ARG and GLN clearances were directly related to the BCAA clearances (rZ=O.65, p<O.O0i). Relationships between ARG, GLN and other AA seem to be ruled by patterns involving specific intracellular AA transport systems. Regulation of ARG and GLN clearances through BCAA administration may be related to an increased flux of AA into synthetic processes. Intestinal segments from rats immunized by infection with the nematode Trichinella spiruUs undergo intestinal anaphylaxis or type I hypersensitivity when challenged in vitro with parasite antigen. Immunized rats exposed to T-radiation (7Gy) and sustained on rat chow,fail to fully express type I hypersensitivity up to 14 Days Post Irradiation (DPI). We hypothesized that enteral nutritional support immediately following irradiation may be effective in preserving mucosal immune responsiveness or in reducing the recovery time after radiation-induced injury. Rats were infected with 3x103 T.spiralis larvae. Thirty to fifty days later rats received 7Gy T-radiation from a Co source as total abdominal irradiation (TAI). Immediately following TAI, rats were fed an elemental amino acid diet (EAAD) and at various DPI sacrificed and tested in Ussing chambers for local anaphylaxis, expressed as antigen-induced CI secretion. The normal CI secretory response to antigenic challenge which is suppressed after irradiation,was restored by 3 DPI when rats were placed on the EAAD. Antigen-induced C[ secretion is dependent on the interaction o~ antigen with specific IgE antibodies on the surface o5 mucosal mast cells and their subsequent degranulation. Mast cell-derived 5-HT,histamine and PG, together effect CI secretion.In irradiated rats on rat chow,the failure to respond to antigenic challenge appears to be associated with the destruction of mast cells. They are undetectable with standard staining procedures and mucosal histamine levels are drastically reduced. Also,el secretion can be induced by direct stimulation of epithelium with cr secrctagogues and circulating IgE levels are normal. Despite a more rapid recovery of immune responsiveness in irradiated rats receiving EAAD,mast cells were not restored. We conclude that the EAAD contributes to an adaptation that supports the expression of local anaphylaxis in the absence of mast cells. Total parenteral nutrition (PN) and bowel rest may influence the host response to infection. This study examined the different host response to infection in parenterally and enterally fed animals. Forty-three rats were divided into PN group and total enteral nutrition (EN) group. They received identically constituted isocaloric isonitxogenous diets for seven days. Animals were then challenged intraperitoneaUy with 3xl(Y Escherichia Coli and survival were observed. In other experiments, they were sacrificed before and two hours after bacterial challenge. Peritoneal cavity was lavaged with 10ml saline and peritoneal exudative cells (PEC) were harvested and cultured in vitro for 24 hours with and without lipopolysaccharide (LPS). Colony f(m, ning units of bacteria (CFU-b) in the peritoncal lavaged fluid (PLF) were determined and TNF in the PLF, serum and PEC culture supernal,ants were measured by L929 bioassay.

Twenty-fonr hour survival rate in EN group was significantly greater than that in PN group (60% vs 0%). The data suggest that local immune responses of PN-treated animals may be depressed with consequent disturbance in the clearance of bacteria. This may lead to a greater systemic cytokine response and resulting in a poor survival after bacterial challenge in PN group. The effects of intragastfic tube feeding of diets enriched with m-6 polyunsaturated fatty acids (PUFA) from safflower oil (SO) or (o-3 PUFA from menhaden oil (MO) on eicosanoid production, and onthe membrane phospholipid fatty acid composition were investigated in a severe acute septic shock model. The percent calories from fat maintained at 30% in the diets was adjusted to provide 10% each of saturated, monounsaturated, and PUFA using olive oil and palm oil as filler fats. After feeding for 5 days, lipopolysaccharide (LPS, 9mg/100g) was injected through the tail vein. The percent of animals surviving in the MO group was 67% (14/21), and did not differ significantly compared to 57% (12/21) in the SO group. The plasma concentrations of TNF ct for the groups of animals fed SO (1.2+3.0 ng/ml) and those fed MO (1.8+1.1) did not differ significantly. The levels of prostaglandin(PG)E2, 6-Keto-PGFlc~ and tumor necrosis factor-ct (TNF-c 0 were determined 90 min after LPS injection. The fatty composition (relative mole%) of the liver was determined before (LPS-) and 3h after the LPS administration (LPS+). The data (mean_+SD; n=6) are as follows.

Group These data indicate that a marked reduction in the plasma levels of POE2 and 6-Keto-PGFlct for rats fed MO diet was associated with a reduction in AA which was displaced by EPA in the membrane phospholipids. Further, in these rats, we observed that there was a 2.8% reduction in the percent of EPA following LPS challenge compared to the basal levels. However, for the group of rats maintained on SO diet, the liver membrane phospholipid fatty acid composition before and after LPS administration was unaltered. These finding suggest that although a reduction in AA was associated with a decrease in the production of pro inflammatory eicosanoids after 5 days of continuous MO feeding, there was no marked effect on the survival following LPS challenge. Animal models have been used to examine the effects of various nutrients and nutrient combinations upon the immune system. We studied the effects of standard and specialized enteral formulations on the response of mice to infectious challenge with Listeria monocytogenes, Influenza A or Candida albicans in order to test the efficacy of specialized ingredients. CF-1 outhred female mice (12-15 g) were fed Purina #5002 chow, commercially available Osmolite HN ®, Perative ® or Impact*. Mortality and tissue burden of pathogen were examined during the 21 d period following induced infection. The results indicate that there were no differences between the groups fed the liquid formulas with regards to mean survival time or % survivors in these models of infection. Examination of spleens in the groups challenged with L. monocytogenes and kidneys in the groups challenged with C. albicans revealed no differences in tissue burden of pathogenic organisms. Alterations in the length of pre-feeding from 0 to 8 days prior to infection did not affect these results.

These data demonstrate that, contrary to previous reports, the use of specialized nutrients did not improve resistance to disease in mice challenged with L. monocytogenese, Influenza A or C. albicans as compared with mice fed Osmolite HN. Animals fed standard chow tended to be more resistant to infectious disease than those fed the liquid formulas perhaps due to the form of diet or complexity of ingredients.

The results indicate that these animal models of infectious challenge may be of limited use to distinguish effects of modified nutrient composition of enteral formulas.

Laboratories The beneficial effects of sesame oil (SES) have been attributed to the presence ofsesamin, a non-fat constituent which is claimed to iultibit A-5 desaturase activity. We investigated the effects of ad libidum feeding of SES enriched diet on fatty acid composition of phospholipids from plasma, liver, on eicosanoid production, and on the protective effects in mice after cecal ligation and puncture (CLP) and compared with safflower oil (SO) diets. Olive and palm otis were added as filler fats to the SO diet to maintain the ratios of saturated, and unsaturated fatty acids similar to SES diet. Thus, while the total fat remained at 15wt%, the only difference between the two diets is the presence ofsesamin in SES diet. The animals were fed for 3 weeks.

The incorporation of dihomo-y-linolealc acid (DGLA: 20:3 o)-6) in the phospholipids from plasma and from the liver for the group of mice fed SO diet ranged between 1-3% compared to the undetectable levels for those fed SO diet. There were no significant differences in the incorporation of other fatty acids either in plasma or in the cell membranes between the two groups. Following an intraperitoneal administration ofLPS (10p.g/kg body wt), the plasma levels of both PGE2, and TxB2, were significantly decreased (P<0.02) by nearly 50% for the group of animals maintained on SES diet compared to those maintained on SO diet.

These findings suggest that sesamin inhibited the release of arachidonic acid (AA) which upon accumulation is retroconverted to DGLA which could reflect a modest increase in the content of DGLA. Failure to decrease the formation of AA could mean that the effects of sesamin on chain elongation process may be after AA is formed, and that sesamin may not inhibit A-5 desaturase activity. On the other hand, our data suggest that sesamin present in sesame oil inhibited the activity of cyclooxygenase which converts AA to its metabolites), or alternatively could block the activity of phosphalipase A 2 (PLA2) (which releases AA from membrane phospholipids) as is evident from a marked decrease in eicosanoid production.

The percentage of animals surviving after cecal ligation and puncture in rite group of mice fed SES diet was 65% (13/20) which was markedly higher (p<0.01) compared to only 20% (4/20) for the SO group. Increase in survival in the group of mice fed SES diet was associated with nearly 50% reduction in the production of TNF in response to LPS i.p. challenge, for SES group compared to SO fed animals.

Our data suggest that sesamin, present in sesame oil, decreased the production TNF~, inhibited the activity ofPLA 2 and consequently offered a significant protection against CLP. Thus sesamin or sesame oil appear to be novel antiinflammatory agents which are present naturally in many edible products. [Dept. Surgery, ~E. Deasoness Hosp.,Harvard Medical School, Boston, USA] Stress causes alterations in the homeostasis of an organism with major changes in various organs, lifespans of cells, protein turnovers, metabolic pathways of activation or stimulation, energy mobilization, etc. Punctual changes have been described in several organs and various cell types of the nervous, endocrine and immune systems. Many stored or newly synthesized proteins are released to the bloodstream to be transported to target organs or to be disposed of. As stress alters protein synthesis and requirements in the targets, the bloodstream is a useful system for monitoring the relevant changes. We present here the pattern of newly synthesized or released serum proteins in C57BL/6 mice that have been stressed by immobilization. The proteins have been labelled with 3sSmethionine in vivo and the study was performed by two dimensional gel electrophoresis based on the method originally described by O'Farrel. The electrophoretic run was carried out in an ISODALT apparatus. A Molecular Dynamics laser scanner and the Kepler Image Analysis System (LSB, Rockville, USA) were used for modelling the radiofluorographic images. The radiofluorographic images of the gels from serum samples obtained from mice injected with I mCi of 3SS-methionine reveal about 440 protein spots, from which a small fraction -about 20are altered in both qualitative and quantitative manner. The major changes are in the range of 25 to 60.000 d. The analysis displays a highly reproducible pattern, where clear differences between stressed and non-stressed conditions are shown. Differences between patterns attributed to chronic vs. acute stress will be presented. Patients with severe trauma or major surgery are known to acquire alterations in neutrophil functions which contribute to their enhanced susceptibility towards microbial infections. We analyzed neutrophil functions from fourty patients admitted for major surgery 6 days and 1 days preoperatively and on the ist and 6th postoperative days (study-days i, 2, 3, and 4) in a randomized placebo-controlled double-blind study. Patients were divided into two groups, one ~eceived 6 days preoperatively an enteral nutrition enriched with omega-3 fatty acids, arginiee, and RNA (IMPACT), the other an isocaloric, not enriched control nutrition (placebo). Meutrophils were isolated from the peripheral blood and stimulated with the Ca-ionophor A23187 (7.3 pM). The capacity of the respective neutrophils to generate leukotrienes was analyzed by RP-HPLC. Neutrophils from IMPACT group generated significantly higher amounts of LTB5 (mean values 6.6 ng/l x i~ cells) compared to the placebo group (2.6 ng) at study-day 2 [p, 0.01). In contrast, the capacity of neutrophils to produce LTB4 was not significantly different in both patients populations at study-days 2 and 3. The omega-oxidation of LTB4 to its biologically less active metabolites OH-LTB4 and COOH-LTB4 was not significantly different in both groups. However, neutrophils from group F patients generate more LTC4 at study-day 2 (p(0/5). The capacity of the patients'neutrophils to generate reactive oxygen radicals upon stimulation with fMLP, PMA or the Caionopbore A23187 exhibited patient dependent differences but no correlations to the repective nutritional supplementation as was measured by luminol dependent ohemiluminescense= These data indicate that an enteral nutrition enriched in omega-3 fatty acids lead to alterations in distinct parameters of neutrophil functions. Clinical patient characteristics and mean caloric intake were similar between the two groups. Both formula were tolerated well and the incidence of diarrhea was low. The number of infectious and wound complications as well as the APACHE II score was evaluated for the two study groups.

Although the number of complications in the group supplemented with arginine, RNA and omega-3 fatty acids was lower no significant difference in the occurance of infectious and wound complications has been observed between the two groups. However, the APACHE II score indicated a significantly improved clinical outcome in patients with supplemented diet.

In conclusion, early enteral nutrition with an arginine, omega-3 fatty acids and RNA supplemented diet helps to recover more rapidly after surgical Total parenteral nutrition (TPN} is associated with intestinal atrophy and dysfunction attributed to the absence of the non-essential amino acid glutamine in current parenteral nutrition solutions. In this animal study with 18 male Wistar-rats we tested the hypothesis that glutamine-dipeptide supplementation during TPN for 10 days would restore small bowel atrophy and TPN induced dysfunction. A conventional TPN solution (250 Kcai/kg bw) was compared to an isocaloric and isonitrogenous TPN (0,5 g N/animal/d) supplemented with alanin-glutamin (Ala-GIn) dipeptide (0,15 g N derived from Ala-GIn =30% GIn-N). An enteral fed cqntrol group was included. Mucosal thickness, villous height and architecture, absorption of water and glucose as well as dissacharidase activity of maltase and alkaline phosphatase were evaluated. Total parenteral nutrition in rats causes villous atrophy, a significantly reduced absorption rate and a decreased activity of villous enzymes compared to an enteral fed control (p<O,01). Addition of Ala-GIn to parenteral nutrition solutions prevents intestinal atrophy and restores functional alterations with a significantly increased rate of water and glucose absorption as well as a higher dissacharidase activity compared to the standard TPN group (p<0,01). There was no significant difference found between the enteral fed control and the dipeptide supplemented animals concerning intestinal structure and absorption, The enzyme activity was significantly decreased in the Ala-GIn supplemented group compared to the enteral fed control (p<0,01). In this rat model supplementation of parentera] nutrition solutions with Ala-GIn dipeptide restores TPN induced intestinal atrophy and dysfunction. Boston. ~ USA Injury, infection and major o1~'ations s~nulate a variety of factors whi~ initiate the body's response to th~ st~..sses. 'i"hese reactions are m~liated by a cbmage ia the neta'al hormonal, eaC.ranmEmt, by the elaboration of eytokines and tl~ougb the stimu/ation of other inflammatory mediators. This eatabolio setting zesulis in body protein loss, which g-taduaI1y erodes both fimetianal and stm~ tissue. Wiu3_B the normally nom'ished individual can withslzod "d~ response for a brief period of time, Ircolonged eatabolLsm L~ associated with ~,nmunosuppmssion, poor wound healing, and proioagcd convalescence.

Surgeons have reJ2ed on intraveaous or eater~ feedings to ~everse this process. A va~ety of data now indicates that it is extremely di~edt to replete protein-containing tissue d~g the eataholie state of illness; the umutt response to hypercaloiSe feedings in this setting is the incxeased accretion of body fat md water.

Administration of gxow~ ho~moae ~cesults i.a n shift of the body's oxidative machinery to fat rather than eazbohydmte utiq~afic~; ¢oncomitaatiy, protein synthesis is stimulated. This metabolic state may have clinical utility in p~e.n.ts who need to heal large wotmds, in those who need to gain strength in order to be w~ed 5ore veatilatory support, or those patients who ~ve mulfiorgan failure and nee~ to ealaaace protein syatSesis to in.suze recovery. 235als are now in progress to dcterm~e the benefit of growth hormone in these and other disease state.s, rn the fature, the increased use of growth hormone, will occur;, this and other growth factors will serve to su~2po~ the host and shorten convalesceace following catabolic diseases. Our objective was to study effects of pre-trauma growth hormone (GH) treatment on liver and skeletal muscle metabolism in a trauma model in piglets.

Twenty-four piglets were randomized to three treatment groups; one group was pretreated with GH 24 IU daily three days before the experiment (GH-3), another group treated with GH 24 IE at the start of the surgical procedure (GH-1) and one group served as untreated controls (CON). They underwent a standardized surgical procedure to place sampling cathethers in the portal, hepatic and femoral veins and doppler flow probes around the portal vein, the hepatic and the femoral arteries. All pigs recieved a primed constant infusion of U-14C-glutamine. Substrate fluxes were measured one (lh) and five (5b) hours after termination of the surgery.

Nitrogen excretion was significantly decreased in the GH treated animals (GH-3:30.8+4.2 mg/kg, GH-I: 37.6+_.5.8 mg/kg, CON: 55.4+-3.1 mg/kg, p=0.004). In the hindieg, there was reduced net glutamine efflux due to decreased absolute glutamine release (GH-3:0.9 + 0.8gmol/min at lh and -2.4+1.4 In.tool/rain at 5h, GH-I: -3.4+1.5 I.tmol/min at lh and -1.4+1.2 pmol/min at 5h, CON: -8.7+_2.8g mol/min at lh and -11.6+3.4/amol/min at 5h, p=0.005, p=0.006), whereas the absolute uptake of glutamine in hindleg was unchanged (p=0.84). Glucose uptake in the hindleg was decreased (p=0.0006). Net glutamine uptake in liver was attenuated (p=0.03 at 5h), whereas net glutamate release from liver was increased (p=0.005).

Thus, pre-trauma treatment with GH improved nitrogen economy, attenuated net glutamine loss from hindieg due to decreased absolute release, and attenuated hindleg glucose uptake. Liver net glutanaine uptake was decerased and net glutamate release increased. Patients with major abdominal surgery exhibit a considerable catabolic response with negative nitrogen balance and protein breakdown, which may have detrimental effects on outcome. We investigated the effects recombinant human growth hormone on substrate metabolism in parenterally fed patients. 48 metabolic healthy patients were randomized to receive either placebo or HGH in a dosage of 0,075; 0,15 or 0,3 I.U. per kg BW. Substrate oxidation rate, energy expenditure as well as short life proteins were measured daily. Six patients were excluded from analysis as drop outs, due to surgical complications. We could find a dose-dependend effect of growth hormone on resting energy expenditure, carbohydrate oxidation, fat and proteine oxidation. With increasing growth hormone resting energy expenditure increased from 20.9 calories per kg body weight to 26.6 calories. This was mainly due to an increased carbohydrate and fat oxidation, while proteine oxidation decreased. This was in accordance with a decreased urea.production rate and an improvement in cumulative nitrogene balance, which increased from minus 12.7 to plus 8.5 g N / 7 days. In consequence short life proteins were also increased. The only negative side effect was an increased blood glucose concentration in some of the patients, making insuline administration mandatory in 3 patients. Our results are in accordance with other studies showing improvement of nitrogen balance, maintainance of lean body mass and muscular strength. If growth hormone administration may have any impact on morbidity, mortality and length of hospital stay in these patients it should be evaluated in a larger number of patients. In 30 patient after liver transplantation (OLTx) the effect of recombinant human growth hormone (rhGH) on protein metabolism and hormonal changes was studied.

Patients and methods: the underlying disease in all patients was end stage liver cirrhosis, non of the patients included in the study was suffering from malignancies. The patients were randomly allocated to receive either 81U of rhGH bid (sc) or no alternative medication. The study period lasted 14 days. From daily 24 hrs udne collection urinary nitrogen loss and daily loss of urea were determined. Body compostion analysis (BCA, bioimpedance, Akern-ltaly) was performed preoperativly and at the postoperative days 7 and 14 to evaluate changes in body water and body cell mass (BCM). The effect on resting energy expenditure (REE) was analysed using indirekt calorimetry wRhin the same intervalls (Metabolic monitor, Datex). Blood levels of GH, IGF1 and IGF2, IGFBP3 were measured daily during the study period, on day 3 a 24 hour profile of GH was.performed. Samples were collected each 20 minutes.

Results: cumulative udnary nitrogen and urea loss were not signifcantly different for the whole study period but for the first 4 study days. BIA indicated for the rhGH-group a loss of BCM to a lesser extent than for the controls (n.s.), changes in bodywater where similar for both groups. REE changes (kcal/kg BCM) were not significantly different for the two groups. Blood levels of GH and IGF1 differed significantly between the groups for the whole study period, but not IGF2 and IGFBP-3. The circadian rhythm of endogenous GH-excretion had not been altered by GH, but absolute levels of GH were increased.

Conclusion: During the first 4 days after OLTx we could proof the protein sparing effect of GH treatment. Although the differences between the two groups were not significantly different for the whole study period the results showed on all days a less catabolic situation for the rhGH treated patients. To possibly improve protein balance after major abdominal surgery we studied the effects of GH on protein metabolism after LTX. Excluding malignant diseases, 29 candidates for LTX were randomized to either postoperatively receive the usual treatment (CO=control group, N=14, age range 25-59 yr, BMI 23.9+5.7 kg/m 2) or for 14 days additional GH (2x8 I.U.s.c.) (GH group, N=14, 21-59 yr, 21.8+2.2 kg/m2). Metabolic studies (calorimetry and lsC-leucine infusion) were performed 5 (test A) and 11 days (test B) after surgery. Tracer analysis was done by gas chromatography-mass spectrometry. During the 2-week study clinical parameters did not differ significantly between groups; however, there was a tendency for increased glucose levels, insulin need, and energy expenditure at the time of test B in the GH group. Leucine oxidation (X+SE) during test A;B was 117+10; 111+15 (CO) and 110+7; 100±12 (GH) pMol/kg FFM/h (n.s.). Protein breakdown was 134±16; 131+30 (CO) and 156+14; 178+30 (GH) pMol/kg FFM/h (n.s.). Nonoxidative-leuine disposal was 119±9; 127+16 (CO) and 139±11; 177+20 (GH) pMol/kg FFM/h (p<0.05 for the rise in the GH group). Tracer data were supported by cumulative urea nitrogen excretion (days 1-4: GH 52+3 vs CO 66±7 g; p<0.05). We conclude that after major abdominal surgery GH lowers nitrogen excretion by increasing total body protein synthesis. It is not known which organs contribute to this effect. In several investigations it has been shown that the protein saving effect of recombinant human Growth Hormone (rhGH) in the postoperative state is accompanied by raised IGF-1 levels in plasma. It was concluded that the anabolic effects of rhGH were probably, at least in part, mediated by IGF-1. This study was aimed at detecting the efficacy of IGF-I on modulation of the protein metabolism in the catabolic state.

Patients and Methodes: 38 patients (both sexes, age: 40-75 years, BMI 17-30 kg/m 2) with major upper gastrointestinal surgery (gastrectomy or partial gastrectomy) received 80 ug rhlGF-1/kg body weight/twice daily or placebo during 5 treatment days beginning on the first postoperative day in a double-blind, randomized study. All patients received hypocaloric and hyponitrogenous total parenteral nutrition (3g CHNg, O.6g AA/kg) troughout the whole study period. Cumulated nitrogen balance, 3-methyl-histidin exretion in urine as well as Serum-IGF-1 levels have been determined. The two tailed Wilcoxon test was used for statistical analysis Results: first results will be presented Previous studies have suggested that growth hormone (GH) potentiates various activities of leukecytes. However, it is not clear whether GH can improve survival of animals with gram-negative sepsis. We investigated the effects of GH on host response to infection in septic mice model. METHODS BALB]c mice were randomized to 3 groups (n=10/gronp): GH-high (4+8mg/kg/day), GH-Iow (0A8mg/kg/day) or control (saline). Following subcutaneous treatment (every 8 hours for 6 days) with GH or saline, mice were challenged i.p. with E.coli (lxl08/body). Survival rate was recorded every 2 hours. In the next experiment, GH-high and control animals were sacrificed 4 hours after bacterial challenge. Peritoneal cavity was lavaged with 3ml saline and the peritoneal exudative cells (PEC) were harvested and counted. Colony forming units (CFU) of bacteria in the peritoneal lavaged fluid (PLF), liver, lung and blood were determined. PEC were cultured in vitro for 24 hours with lipopolymccharide (10p.g/ml In normotensive adults growth hormone (GH) rises when clonidine challenge is performed. Recently evidence was shown for GH to accelerate wound healing. While GH secretion patterns are inconstant the GH-dependant insulin like growth factor (IGF-1) and the insulin like growth factor binding protein (IGFBP-3) reflect the integrated GH secretion over days. Since we administer clonidine to patients with acute alcohol abuse we determined plasma levels of IGF-1 and IGFBP-3. Materials and methods: 21 patients sheduled for esophagus resection were investigated once they had given written informed consent. 8 patients showed acute alcohol abuse and were treated with clonidine postoperatively• 13 Patients with no history of acute alcohol abuse served as controls. Besides liver enzymes IGF-1 and IGFBP-3 were determined.

Results: Both groups had a comparable hepatic capacity of synthesis with decreased cholinesterase levels as to be expected after a major operation. Regression analysis of IGF-1 and IGFBP-3 levels showed no differences between the groups.

Discussion: IGF-1 and IGFBP-3 plasma levels are not increased in normotensive patients who are treated with clonidine for prevention of postoperative alcohol withdrawl syndrom. Further studies are required to discriminate whether our results are due to impaired postoperative liver function or if clonidine has not any impact on the GH-IGF-axis in patients with alcohol abuse. Recent studies have revealed that IGF-I has several immanoregulatory roles. However, little is known about its effects on septic animals. We tested whether IGF-I could increase the survival O f mice challenged intraperitoneally (i.p.) with E. coli. M~THQD$ BALB/c mice received either high dose IGF-I (n=10, 24mg/kg/day), low dose IGF-I (n=10, 2.4mg/kg/day) or saline (11=10) every eight hours subcutaneously for six days. Animals were then challenged i.p. with lx108 E. coli. Survival was observed every two hours. In the other experiment, mice that received high dose IGF-I or saline were sacrificed four hours after bacterial challenge. Peritoneal cavity was lavaged with 3ml saline and the peritoneal exudative ceils (PEC) were harvested and counted. Colony forming units (CFU) of bacteria in the peritoneal lavaged fluid (PLF), liver, lung and blood were determined. In addition, PEC were cultured/n vitro for 24 hours with lipopolysaccharide (10).tg/ml). Tumor necrosis factor (TNF) in the supernatants of PEC culture, PLF and serum were measured by L929 bioassay. RESULTS IGF-I improved the survival rate at a dose of 24mg/kg/day. Severe chronic neutropenia (SCN) is a disorder characterized by severe neutropenia secondary to either a maturational arrest of myelopoiesis at the level of promyelocytes (Kostmann-Syndrome; SCN) or regular cyclic fluctuations in the number of blood neutrophils with a median absolute neutrophil count (ANC) of less than 500/~tl (cyclic neutropenia). Patients suffering from SCN have an impaired acute inflammatory response to injury and an increased susceptibility to infections, i.e. bacterial or fungal infections, resulting in chronic oral inflammation, severe pneumonia, abscess formation and bacteraemia. We have treated 36 patients (32 SCN, 4 cyclic neutropenia) with r-metHuG-CSF (filgrastim). Thirty of 32 patients with SCN and all 4 cyclic neutropenia patients responded to this treatment with an increase of the median ANC to greater than 1000/~tl. The dose of r-metHuG-CSF needed to achieve and maintain the response varied between 0.8 and 120 ~tg/kg/d. Long-term treatment did not exhaust myelopoiesis: the ANC remained stable after up to 6 years of treatment and antibodies to r-metHuG-CSF were not detected. The increase of ANC was associated with dramatic clinical responses: significant reduction of severe bacterial infections, reduction of intravenous antibiotic treatment, and reduction of hospitalizations. No severe bacterial infections occurred in any of the r-rnetHuG-CSF responders during longterm treatment. Severe adverse events, most likely due to the underlying disease, included the development of MDS/Leukemia in two patients, and osteopenia/osteoporosis in 12 patients. These results demonstrate the beneficial effects of r-metHuG-CSF treatment in severe chronic neutropenia patients. The newborn is predisposed to mortality during infections due in large part to developmentally immature host defenses. Recently cytokines have been identified that regulate the development of these defenses. One such cytokine is specific for neutrophils and is termed granulocyte colony stimulating factor (G-CSF). In the studies to be discussed, we have attempted to impact hematopoiesis in the fetus using exogenous rhG-CSF delivered through the pregnant dam. Our rationale was that an enhanced myeloid system may result from such in utero treatment leading to, perhaps, enhanced protection to microbial insult. RhG-CSF crossed the placenta and reached peak fetal serum concentrations 4 hours after administration. These levels were 1000-fold lower than the levels detected in the adult dams serum. White blood cell counts were elevated approximately 2-4-fold over control newborn blood. This increase was due to circulating numbers of neutrophils. The marrows from these pups were hyperplastic consisting of predominately immature and mature neutrophilic cells. No changes were seen in the thymus or livers. Pups born to mothers treated with rhG-CSF since day 15 of gestation (parturition in rodents occurs on approximately day 21) survived a lethal challenge of group B-13 hemolytic Streptococcus. In contrast their untreated yet infected counterparts did not survive. Recent clinical developments have led to trials of rhG-CSF in certain neutropenic states including cyclic and chronic neutropenia. We have identified several of these patients who were pregnant during treatment with rhG-CSF. Biologically and anfigenically identifiable rhG-CSF was detected at increased levels in the cord serum of these neonates. We will discuss our findings with these patients in light of our animal model of neonatal myeloid development. We have previously demonstrated decreased G-CSF production and G-CSF mRNA expression from activated mononuclear cells from newboms compared to adults (Cairo, Pediatr Res 31:574, 1992) . We have also recently observed that administration of a single dose of rhG-CSE to one-day-old newborn rats induced significant neatrophilia, while administration for 7 days induced neulrophilia and increased the bone marrow (BM) neutrophil storage and progenitor pools (Cairo, Blood 76:1788 , 1990 Cairo, Pediatr Res 27:612, 1990) . We embarked on a phase I trial exploring the safety, pharmacokinetics, and biological efficacy of G-CSF in newborns with presumed sepsis. 42 infants <72 hrs old with a diagnosis of presumed sepsis were stratified into three groups: very preterm (26-30 wk), preterm (31-35 wk), and term (>36 wk) and randomized to receive either a placebo or 1, 5, and 10 gg/kg/qd or 5 and 10 p.g/kg/bid of rhG-CSF infused by pump over 1 hour for 3 consecutive days. CBCs were obtained at 2, 6, 24, 48, 72 and 96 hrs. Tibial bone marrow aspirations were performed 72 hrs after study entry and differential counts and CFU-GM pools were determined. C3bi expression was determined at 0 and 24 hrs after rhG-CSF, and G-CSF pharmacokinetics were performed after the first dose of rhG-CSF utilizing a sandwich ELISA. A significant increase in the ANC was observed at 48, 72 and 96 hrs following administration of both 5 and 10 ~tg/kg/d of rhG-CSF. The maximum increase in the ANC occurred 72 hrs after 5 and 10 ~tg/kg/d (397 -116%) (p<0.01) and (621% -+ 200%) (p<0.001), respectively. There was a significant dose-dapendeat increase in the BM neutrophil storage pool (18 _+ 4% vs. 62 + 6%) (p<0.01) (placebo vs. 10 ~tg/kg/d). There was no significant difference in the nantrophil proliferative pool. An increase in CFU-GM and CFU-GEMM was seen at all doses tested, compared to placebo (53.5 _+ 8.6 vs. 87 -+ 15) (colonies/l(P cells/plate). C3bi expression was significantly increased 24 hrs after 10 Bg/kg/d of rhG-CSF (233 + 81% vs. 83 +-26%) (p<0.012). Peak serum G-CSF levels occurred at 2 hrs and were dosedependent. The half-life of rhG-CSE was 4.44 + 0.4 hrs. Most importantly, there was no observed toxicity from G-CSF in all patients studied. 31 of 42 patients were on ventilators prior to administration of rhG-CSF and there was no increase in pulmonary toxicity. These preliminary data suggest that rhG-CSF is well tolerated at all gestational ages in newborns with presumed sepsis. A multi-center phase II/III randomized double-blindad placebo controlled trial is required to determine the efficacy of rhG-CSF in this clinical setting. 

We investigated the effects of recombinant canine granulocyte-colony stimulating factor (G-CSF) on survival, cardiopulmonary function, serum endotoxin levels and tumor necrosis factor (TNF) levels in a canine model of lethal bacterial septic shock (Clinical Research.41:240, 1993) .

Methods: Awake 2 ylo beagles had serial cardiopulmonary and laboratory studies before and for up to 10 days after intraperitoneal placement of an E. celi infected clot. Nine days before and daily until 3 days after clot placement, animals received high (n=17) or low dose (n=17) G-CSF or protein control (n=20) subcutaneously.

Results: Survival in high dose G-CSF animals (14/17) was significantly improved compared to low dose (10117) and controls (12120) (p<0.04 Wilcoxon). High dose G-CSF also improved cardiovascular function evidenced by a higher mean left ventricular ejection fraction (day 1 after clot, p<0.001) and mean arterial pressure (day 2, p<0,02) compared to low dose and controls. High dose rcG-CSF increased (p<0.001) peripheral neutrophil numbers both before and after clot implantation (2 hours to 4 days) compared to low dose and controls. In addition, high dose rcG-CSF produced a more rapid (p<0.0001) rise (day 2) and fall (day 4) in alveolar neutrophils determined by bronchoalveolar lavage compared to low dose and controls. Lastly, high dose rcG-CSF decreased serum endotoxin (2 to 8h, p<0.002) and tumor necrosis factor (TNF, 2h, p<0.02) levels compared to low dose and controls.

Discussion: These data suggest that therapy with G-CSF sufficient to increase peripheral neutrophil numbers during peritonitis and septic shock may augment host defense and endotoxin clearance, reduce cytokine levels (TNF) and improve cardiovascular function and survival. The use of G-CSF in sepsis prophylaxis in neutropenic patients is well established and has been ascribed to accelerated recovery in granulccyte counts. Here, an additional sepsis-prophylactic property could be demonstrated in healthy volunteers: Eleven volunteers were employed in a sinqle-bTind, controlled study and were given 480 uq G-CSF or saline placebo via subcutaneous injection. Blood was withdrawn immediately before and 20 or 44 hours later. LPS-inducible TNF, IL-1, sTNF-R p75 and IL-lra were assessed in the supernatant of whole blood incubations stimulated with 10 ug/ml LPS from Salmonella abortus equi. Similarly to previous animal studies, LPS-inducible TNF was attenuated by about 50% 20 hrs. after treatment. The same was true of IL-lb. In contrast, LPS-inducible sTNF-R p75 which was indetectable in blood incubations from untreated donors increased dramatically 44 hrs. after G-CSF treatment. IL-lra found after LPS challenge was increased tenfold by G-CSF treatment. It is concluded that G-CSF treatment switches peripheral leukocytes to an antiinflammatery state characterized by an attenuation of IL-I and TNF releasing capacity and an augmentation of the release of cytokine antagonists. This findinq minht offer a novel concept in septic shock prophylaxis. Objective.The aim of the study was to investigate the effect of recombinant human G-CSF (rhG-CSF) on survival, bone marrow neutrophil myelopoiesis, neutrophil counts, levels of bacteria and some important sepsis mediators in a model of rat abdominal sepsis.

Lethal peritonitis was induced with a 4 mm coecal perforation (CP) in male Wistar rats. rhG-CSF was administered as 10 /.tg/kg iv every 12 h, first dose at sepsis induction. Bone marrow neutrophi] progenitors were determined as blast colonies, CFU-GM and CFU-G. Neutrophils and bacteria were determined in peripheral blood and peritoneal fluid. LPS, TNF, Endothelin 21 and Lactate were measured in blood from femoral vein. Mortality rates were registered with G-CSF treatment starting either 1 or 7 days before or 4 hours after CP. Results. Mortality was reduced from 90% to about 50% with rhG-CSF intervention and there was no difference between the pretreatment and treatment groups. Bone marrow blast colonies were not influenced while neutrophil myelopoiesis was augmented at the stages of CFU-GM and CFU-G. Neutrophils in blood and peritoneal cavity were enhanced and numbers of bacteria in the same compartments were substantially reduced. Circulating LPS, TNF, Endothelin 21 and Lactate were attenuated the first 24 hours after CP.

Neutrophil myelopoiesis is augmented with increased number of neutrophils in blood and peritoneal cavity, resulting in enhanced clearance of pathogens. LPS, TNF, Endothelin 21 and Lactate are suppressed the first 24 hours during sepsis course. A. Wendel, J. Barsig, G. Tiegs GM-CSF stimulates the proliferation and differentiation of granulocytic and monocytic progenitor cells. In addition the hemopoietic cytokine activates the inflammatory response in mature leukocytes. The priming effect of GM-CSF towards lipopolysaccharide (LPS)-induced cytokine production in vitro has been described, but little is known about proinflammatory GM-CSF effects in vivo.

We detected GM-CSF in plasma of LPS-challenged mice with kinetics similar to TNF, reaching peak levels 2h after LPS administration.

GM-CSF pretreatment (50 ~tg/kg i.v.) enhanced mortality in mice challenged by a sublethal dose of LPS. Plasma levels of Tumor Necrosis Factor (TNF) and Interleukin-6 (IL-6) were significantly enhanced. A monoclonal antibody, which neutralizes GM-CSF bioactivity, rendered mice less sensitive towards lethal LPS-challenge. TNF-and IL-6-tevels were reduced in these mice compared to control animals without antibody treatment.

In addition, severalfold potentiation of LPS-induced cytokine release by GM-CSF was observed in vitro in murine bone marrow cell cultures.

These data demonstrate the proinflammatory capacity of GM-CSF and suggest that the hemopoietic cytokine plays also a role as an endogenous modulator of LPS toxicity. Immune dysfunction, developing in the wake of multiple trauma, overwhelming infection and other forms of critical surgical illnes% is associated with increased infections, morbidity and mortality. The mechanisms responsible for alterations in immune regulation are incompletely understood but monocyte appear to play a central role. Polymorphonuclear leukocytes (PMN) are known to play a central role in the inflammatory response of the host toward invading microrganisms. Reports of defects in all the aspeots of PMN function have been accumulated in recent years. The possible role of GM-CSF in modifing the state of immuno suppression detected in severe intraabdominal infected pt~. inspite of surgical appropriate procedures and in reducing the expected mortality is investigated. The safety of rh-GM-CSF administration in sepsis is also evaluated. A double blind randomized study is proposed. This study include 60 ICU patients who do not exhibit signs of shock and/or ARDS, with clinical signs and symptoms of abdominal infection. Immunodepressed patients-AIDS, Chronic chemotherapy or chronic steroid administration do not partecipate to the study. Patients will receive rGM-CSF (l~g/Kg/day) or placebo in 24 hs. continuous infusion for 15 days. SafetyandefYieacy will be assessed till to Day 30. The APACHE II Score is adopted for risk stratification of patients because it is reliable and validated, objective and composed of information that is indipendent of diagnostic criteria. Patient's entry criteria is APACHE II > 16 (Score 16 corresponds to expected mortality rate of 38%).In this protocol the surgeons report the judgement of the efficacy of surgical procedure to remove or not the focus of infection. Objectives: Infections and subsequent septic responses remain the leading cause of death among surgical intensive care (SICU) patients despite tmprovetaunts in supportive care and brond-epectrum antibiotics. Usually invading bacteria are efficiently cleared by neutrophil granulocytes. However, during sepsis various neatrophil dysfunctions have been demonstrated, leading to impaired host defense. Granulocyte colony-stimulating factor (G-CSF) induces a sustained increase in circulating neutrophils and enhances various noutrophil functions. It was the purpose of the present study, to evaluate the safety and efficacy of G-CSF (filgrastim) in SICU patients at risk of sepsis. Materiel a.d Methods: The study was designed as an open-label phase-ll study of filgrastim. Ten consecutive SlCU patients, with a Therapeutic Interveotion Score greater than 30, were included in the study. Filgrastim was given by daily continuous intravenous infusion for 7 days or discharge from the SICU. APACHE ll-score, multiple-organ-failure (MOF) score, definitions of infections, sepsis, systemic inflammatory response syndrome (SIRS), and acute respiratory failure were applied daily. A response to filgrastinl th_erapy was defined as an improvement in disease severity quantified by a decrease of > 4 APACHE 1I score points on day 4 after onset of treatment. Results: None of the 10 patients developed a sepsis or MOF later on and no patient died during hospitalization. Specific postoperative complications occured in one patient ~Jth a leekage of the oesophagou-gastric anastomosis after oesophageus resection. At study entry the leucocytes amounted to 10.110 + 2.000/~tl (mean + SEM) and reached a level of 29.280 +_ 5.810/tal at day 6 after onset offilgrastim therapy. The APACHE II score initally was 14 + 1.57 (mean + SEM) and as an indicator of filgrastim response a decrease of 5 points ~dthin 4 days oceured in 7 out ot 10 patients. Filgrastim was well tolerated, side effects were not noted. Growth of solid tumors might be modulated by the activity of inflammatory and/or immune effector cells of undefined specificity. In this study patients undergoing surgical treatment for gastric (n= 41) or colorectal (n= 47) cancers were evaluated for endogenous serum levels of granulocyte colony-stimulatingfactor (G-CSF) during a pre-and postoperative time period. From the same blood specimens mononuelcar cells (MNC) were prepared. The release of IFN-%, and IL-2, which are secreted by Thl cells, were stimulated in vitro by PHA during a cell culture period up to 48 hours. The patients were further classified for their immunreactivity by responses in DTH skin testing to seven different antigens (e.g. tetanus toxoid, PPD, diphtheria toxin, trichophyton, streptococcus, candida and proteus antigens). DTH testing has been repeated in each patient Two remarkable results were obtained. The serum levels of endogenous G-CSE showed a biphasic increase with maximum values of 900 pg/ml (preoperative < 70 pg/ml) on day 1 and day 3 to 5 after surgical treatment. Similar patterns of G-CSF production were found in both groups of patients with gastric or colorectal cancers. High serum levels of G-CSF were significantly (p < 0,01) correlated with infectious complications in patients whh gastric cancer (n= 17/22). Secondly patients could be arranged into two groups according to an anergic (n= 25) or normergi¢ (n = 63) responsiveness in DTH testing. The frequency of anergi¢ responsiveness was similar in both patients with gastric (n= 11/41) or colorectal (n= 14/47) cancers. Interestingly we found a significant correlation (p < 0,01) between low serum levels of G-CSF and anergy during the postoperative period in both groups. Stimulation of MNCs from anergic patients (n= 30) within the pre-and postoperative period resulted in reduced mean values (about 50%) for IFN-ff release (preoperative means llO0 pg/nfl), if compared to patients with normergic DTH (n= 63, preoperative means 1960 pg/ml). Similar, but less significant results were obtained for IL-2 secretion. Our results confirm a correlation between infectious complications and G-CSF in the postoperative period, however elevated levels were also found in some patients without any signs of infections. More interestingly there might be an association between cytokine (C~CSF, IFN-% and IL-2) release and DTH, which is known to be mediated by activated Thl calls. To recognize anergic DTH as a possible higher risk in the postoperative outcome of cancer patients extended periods of observation are needed. Objectives of the study Effects of recombinant huraan granulocyte colony-stimulating factor(rhC-CSF)a galnst severe septic infections were investigated by its single use or by its corn b{nation with cephera antibiotlcs.We examined its effects on the mortality,and circulating blood neutrophyis counts and functlons,such as phagocytic activity and H 2 02 production using the rat severe septic model.

Rats were subcutaneously administsrd rhC~CSF(S0orl0O ~ g/k~ body wt)after on set of peritonitis brought about by cecal ]igation and one puncture withe2-gaug e needle once a day for three days.in addJtlon,cefmetazol Na(CMZ)(50m$/k9 bo dy wt)was injected intrarnustularly to the rats tv~ce a day for three days. Cirehlatlng blood neutrophyls counts were determoned electronically with a hem ocytometer,and blood smears stained with May~runwaldM.qlemsa~taln. Neutrophyls functions in vltro,such as phagocytic activity and H 20 2 producti on using the rat severe septic model was analyzvd by automated flow cytometri c single cell-analysis methods.

The reortallty rate after 2 weeks was significantly decreased by administratlon of rh~-CSF(p<0,01).ln addJtion,a combination therapy of rhG-CSF wlte cephern ant~biotics(CMZ)showed a significantly survive] advantage and the rate had b een reached 57.1%. Nextly,treatn%ent wlth rhG-CSF(S0 ~ $/k9 body wt)increased the nuzaber of the peripheral blood neutrophJls slgn[fieantly(p<0.01). iV~oreover,functions of neutrophlis which were phagocytic activity and H2 02 p roduction were remarkably enhanced by admlnlstratlon of rhG-CS~(50 ~ 9/ks b ody wt) (p<0.(15).

These findings suggest that combination therapy of rhCrCSF with cephern antib iotlcs(CMZ)is an efficient regime against severe infectlons.and the increased ne utrophils counts and enhanced neutrophiis functions were played a important ro le about the survival advantage. Granulocyte macrophage colony-stimulating factor (GM-CSF) is a haematopoietic growth factor active on neutrophils and macrophages.

Leukopenia often occurs following renal transplantation and can be associated with infection and/or the myelosuppressive effect of azathioprine. Aim: We report the use of GM-CSF in 17 renal allograft recipients with leukopenia. Nonglycosylated recombinant GM-CSF was obtained from E. coli transvected by human GM-CSF gene. M~terial ~,nd Methods : Written informed consent was obtained from all patients. Patients were suffering from toxic neutropenia (neutrophils < 500/mm 3) with medullar hypocellularity on bone marrow aspiration, or leukopenia (neutrophils < 1000/ram 3) with cytomegalovirus infection requiring ganciclovir administTation. GM-CSF was given subcutaneously at a dally dose of 2 to 5 mcg/kg/day, according to renal function. Results : In all cases, neutrophil counts returned to normal levels within 1 to 4 days. In most of them, spectacular correction was observed within 24 hours, with a single injection. Adverse events due to GM-CSF at this dose were mild and easily managed (2 cases of bone pain treated with paracetamol). One acute rejection episode was observed after correction of leukopenia. Conclusion : On the basis of this study, it appears that GM-CSF at a dose below 5 mcg/kg/day is an effective treatment for renal transplant recipients with leukopenia associated with CMV infection or toxic neutropenia.

Department of Nephrology, 161, rue de S~vres, HOPITAL NECKER, 75743 PARIS, FRANCE.

Changes in serum G-CSF and IL-6 after surgical intervention Hitoshi Toda 1, Atsuo Murata 1, Hidewaki Nakagawa 1 , Takesada Mori 1, Nariaki Matsuura 2 1Osaka University Medical School, Osaka, 2Wakayama Medical School, Wakayama, Japan

We measured serum immunoreactive interleukin 6 (IL-6) and granulocyte colony-stimulating factor (G-CSF) levels of the patients undergoing major thoraco-abdominal surgery for esophageal cancer. Serum samples were collected from eight patients on the day before surgery, at the time of operation, and thereafter at suitable intervals for one week. IL-6 and G-CSF were measured by means of enzyme linked immunoassay. The normal range of serum ]L-6 was less than 2 pg/ml and G-CSF less than 4 pg/ml. Values between groups were compared with linear regression analysis. Both serum G-CSF and IL-6 levels reached their maximal levels at the first postoperative day and decreased thereafter. The correlation between G-CSF (y) and IL-6 (x) was y=3.28x+6.16 (r=0.786, n=86, p<0.01 ), showing a significant correlation. In the case who suffered from aspiration pneumonia and ARDS at the second postoperative day, the peak level of IL-6 was 50 pg/ml and G-CSF 69 pg/ml respectively. The estimated value of G-CSF was 171 pg/mi by the regression equation. This means the real G-CSE level was less than half of the estimated value. It suggests that low responsiveness of G-CSF is one of the reason of immunodeficient state after the major surgery, Neutrophils from injured patients ingest and kill bacteria less efficiently as compared to those of healthy individuals, probably reflecting the suppression in respiratoly burst which occurs after severe trauma. One of the main mechanisms of killing bacteria by neutrophil granulocytes is production of oxygen radicals (respiratory burst). Granulocyte colony-stimulating factor (G-CSF), a 20kilodalton cytokine, leads to a sustained, dose-dependent increase in circulating neutrophils. Thus, it was investigated whether filgrastim (recombinant human granulocyte colony-stimulating factor, rhG-CSF) therapy fits for prophylaxis of sepsis in postoperative/posttraumatic patients, and whether, besides an expected increase in neutrophil count, filgrastim would also augment neutrophil function. MATERIAL AND METHODS: This study was designed as an open label, prospective phase II study of filgrastim and performed in a surgical intensive care unit (SICU) (university hospital). 10 postoperative/post-traumatic patients with a Therapeutic Intervention Scoring System (TISS) score greater than 30 were treated with filgrastim (0.5 -1 l.tg/kg/day) for prophylaxis of sepsis on 7 days or until discharge from the SICU. Production of oxygen radicals can be quantified by analysis of FMLP-and zymosan-induced chemiluminescence. Neutrophil oxygen radical production was tested by FMLP-and zymosan-induced chemiluminescence by the polymorphonuclear cells (PMN) of these patients in multiple blood samples over a period of up to 14 days. RESULTS: None of the patients treated with filgrastim for prophylaxis of sepsis developed sepsis. In vitro FMLP-induced (10 -7 reel/l) neutrophil oxygen radical production was significantly increased under therapy with filgrastim by a maximum of 797% +-570% (217% -1826%) compared to pretreatment values of 100%. Tapering of filgrastim resulted in a reduction of FMLP-induced neutrophil oxygen radical production within 48 hours. In contrast, zymosan-induced neutrophil oxygen radical production was not affected by filgrastim treatment. CONCLUSIONS: Besides its quantitative effect on neutrophil counts enhanced neutrophil function, documented here as increased FMLP-induced oxygen radical production, may account for the beneficial effect of filgrastim for prophylaxis of sepsis in posttraumatic/post-operative patients. Granulocyte colony stimulating factor (G-CSF) and granulocytemacrophage colony stimulating factor (GM-CSF) have been recently introduced in the treatment of chemotherapy-induced neutropenia. Effects of these CSFs on cellular immune system were evaluated in 38 neutropenic gynecological cancer patients during chemotherapy. G-CSF and GM-CSF were equally able to induce a rapid recovery of white cell count within one or two days. G-CSF treatment resulted in a significantly higher concentration of leukocytes measured in the peripheral blood although by GM-CSF a sufficient effect was achieved (p<0.05). Before initiation of CSF treatment urinary neopterin was similar in both groups of patients (283+/-38 and 267+/-40 lamol/mol creatinine for GM-CSF and G-CSF respectively expressed as mean +/-one SD). In G-CSF treated patient only a marginal induction of neopterin was observed. On day 4 the mean value was about 40% above the basal level (p<0.05). On the other hand GM-CSF treated patients were characterized by a pronounced increase in urinary neopterin levels. In comparison with the basal level a more than 2 fold induction was noted and the difference between G-CSF and GM-CSF was highly significant (p<0.01). This effect was confirmed in vitro by investigating the effects of these CSFs on interferon-gamma mediated pathways in THP-1 human myelomonocytic cells. Results suggest activation of immune effector cells by GM-CSF which may help the organism to overcome infections. However, activated macrophages produce several growth factors which may increase malignant proliferation, and augmented neopterin production as sign of macrophage activation has also been associated with poor prognosis m several malignancies. More data are therefore necessary to clarify whether CSF mediated immune activation is beneficial or deleterious for cancer patients but considering our results caution in applying CSFs in oncology seems advised. From a historical perspective, the development of humoral immunity to bacterial endotoxin has assumed a prominent position in the spectrum of therapeutic approaches which have been explored for the treatment of Gram negative septic shock. Predicated upon the fact that rough strains of bacteria manifest LPS containing exclusively conserved structural features common to LPS from all Gram negatives, specific antibodies were elicited which conveyed cross protective immunity in experimental models of bacteremia and endotoxemia.

Such studies culminated in a well-conducted, randomized, double-blind placebo-controlled clinical trial using passively administered human polyclonal antiserum to treat patients with suspected Gram negative sepsis. The efficacy of treatment established in that trial spurred efforts to develop monoclonai reagents which, to date, have not been uniformly successful in reproducing those earlier studies with polyclonai antibodies. Nevertheless, the numerous successes which have been documented in experimental models of endotoxemia continue to foster promise for this immunotherapeutie approach. Several recent studies with human polyclonalimrnunoglobulin preparations containing antibodies reactive with LPS and lipid A have yielded promising results in treatment of patients with sepsis. In addition, the recent development of an antiidiotypic monoclonal antibody which reflects an internal image of a kDO specific monoclonal antibody has provided an alternative experimental approach to generate anti-LPS antibody. Immunization of mice with the antiidiotype provides significant protection against subsequent LPS lethality consistent with the development of circulating immunoglobulin specific for LPS. Thus, the use of polyclonal immunoglobulins contrives to provide an alternative and potentially cost effective method for the treatment of endotoxin shock. Supported by R37 A123447 and POt CA54474.

John Holaday, Anne Fortier, Shawn Green, Glenn Swartz, John Madsen, Carol Naey, and Jan DiJkstra EntreMed, Inc.. Rockville, MD, 20850.

At the time of diagnosis, the signs and symptoms of septic shock are an indication that the systemic inflammatory response is well underway; thus, it has been argued that the endotoxin "cat is out of the bag", and that subsequent passive immunization may be too late to achieve therapeutic benefit. Our approach has been to evaluate active immunization as a prophylax~s against sepsis. Mice were inoculated twice (two weeks apart) with liposomes containing DMPC[I.8], DMPG[0.2], cholesterol [1.5] , and monophosphoryl lipid A [20-200 gg/txmole phospholipid] by several routes (i.p., i.m.), and serum was collected 10-11 days after each inoculation. After a single injection, highest tilers of Ab were produced in mice inoculated i.p., but mice inoculated by all routes produced anti-lipid A Ab. Following the second injection. Ab levels were roughly equivalent in mice inoculated by all routes, regardless of lipid A concentration. Mice vaccinated i.p. with liposomes containing 0, 20 or 200 gg lipid A were treated with cyclophosphamide to produce neutroperda and then challenged with E. cole in an infection model of gram negative sepsis. The LDS0 for control (liposomes with no lipid A) mice was 3x10 bacteria; LD50 for mice vaccinated with 20p.g was 24x103 (8-fold increase in resistance) and with 200~tg was 48x103 bacteria (16-laid increase in resistance). Mice vaccinated as before were also treated with Actinomyein D to increase sensitivity to LPS (Salmonella minnesota) challenge in an endotoxemia model of grain negative sepsis. The LD50 for control (liposomes with no lipid A) mice was 30 ng LPS; the LD50 for 20gg lipid A was 70rig LPS (2-fold increase in resistance) and for 2001xg was 570ng LPS (19-fold increase in resistance). Mice were similarly vaccinated and challenged with an aggressive gram negative pathogen, Francfsella tularensis. The LD50 of FranciseUa in normal mice or mice inoculated with liposomes without lipid A was 20-40 bacteria. In contrast, mice vaccinated with liposomal lipid A (200ggl survived challenges as high as 30,000 bacteria, (3 logs of protection). The impressive protective capacity of this vaccine did not correlate with Ab Liter in any of the sepsis models, nor did it correlate with classic nonspeeific events, such as macrophage activation. Maerophages harvested from the peritoneum of mice vaccinated and protected against sequelae of gram negative infections did not spontaneously kill the bacteria in vitro, but could be activated by IFN-y for antimicrobial activity equivalent to that of macrophages from unt#eated mice. Research is underway to defme the protective mechanism(s) activated by this liposomal-Lipid A vaccine.

INTERVENTION BY MONOPHOSPHORYL LIPID A IN SEPTIC SHOCK Jon A. Rudbach, Ribi ImmunoChem Research, Inc., Hamilton, Montana, USA Monophosphoryl Lipid A (MLA), the clinical form of which is called MPL®-immunostimulant, has been tested extensively as an intervenient material in septic shock. MLA is protective when given to experimental animals prior to a live microbial challenge or challenge with lethal doses of microbial products or certain cytokines. This is shown with Gram negative and Gram positive bacteria, Gram negative bacterial endotoxins, and Gram positive bacterial exotoxins. Furthermore, animals treated with a regimen of MLA which results in a refractory state to a lethal dose of Gram negative bacterial endotoxin concomitantly display increased resistance to a live bacterial challenge. Thus, both endotoxin tolerance and nonspeciflc resistance to infection can be manifested simultaneously. Also, prophylactic doses of MLA do not interfere with other therapies given subsequently; an additive or a synergistic protective effect can be demonstrated with certain combinatorial treatment regimens, such as MLA followed by antiendotoxin monoclonal antibodies. The preclinical studies were extended to human trials wherein the safety of agonistic doses of MLA was verified. Furthermore, when MLA was administered to human volunteers 24 hr before challenge with a pharmacologically active dose of reference endotoxin, febrile, cardiac, TNF, IL-6, and IL-8 responses were all decreased significantly as compared with the responses of subjects pretreated with a control solution and challenged with endotoxin. Human trials with MLA are being extended into patient cohorts which have high probabilities of developing septic shock; this will expand the safety base and establish clinical efficacy for MPL®-immunostimulant. A considerable body of in vitro evidence supports the concept that the effects of LPS on cells of the immune/inflammatory systems are controlled by interactions of LPS with CD14. To evaluate if blocking LPS-CD14 interactions has potential as a therapeutic in septic shock we have evaluated the effect of anti-CDI4 monoclonal antibody (mAb) on LPS-induced cytokine production and physiologic changes in an experimental model of endotoxin shock performed in cynomolgus monkeys. A novel model has been established where animals were treated with interferongamma for three days prior to infusion of highly purified LPS over an eight hour period. In this model LPS challenge resulted in marked release of eytokines in the blood, substantial hemodynamic changes, release of liver enzymes and alteration in lung permeability observed over a 24 hour period. To evaluate the effect of treatment with anti-CD14 mAb, animals were given either nothing, an isotype control or anti-CD14 mAb (5mg/kg) 30 rains, prior to the beginning of the LPS infusion. Evaluation of physiologic changes including mean arterial blood pressure and cardiac output, quantitative analysis of eytoldne levels including TNFct, IL-113,1I,-6, IL-8 and IL-10, and liver enzymes during a 24 hour period revealed that treatment with anti-CD14 mAb markedly attenuated all parameters of injury including decreased mean arterial blood pressure, increased cytnkine levels and the release of liver enzymes observed in animals given the isotype control mAb or those not treated. Administration of anti-CD14 mAb to interferon-gamma treated animals NOT challenged with LPS did not induce any detectable physiologic changes or increases in cytoldnes. These studies suggest that strategies to block LPS-CD14 interactions will have utility in diseases such as septic shock or ARDS where LPS plays a central role in initiating injury.

Preclinical Studies with Recombinant Bactericidal/Permeability Increasing Proteins (rBPI and rBPI23). P.W. "Frown, Dept. of Preclinical Science, XOMA Corporation, Berkeley, California, USA. Bactericidal/permeability increasing protein (BPI), from neutrophils, binds to and neutralizes lipopolysaccharide (LPS); it also specifically kills gram-negative bacteria (GNB). These properties, which reside in the N-terminal half of the molecule, indicate potential therapeutic application in the treatment of gram-negative sepsis. The gene for human BPI has been cloned and recombinant holoprotein (rBPI) and a 23 Kd N-terminal fragment (rBPI;3) have been produced in sufficient quantities for preclinical studies. Both rBPI and rBPI23 bind to lipid A and neutralize the biological activities of LPS derived from a variety of organisms, rBPI23 has equivalent antibacterial activity to BPI against rough GNB but is up to 30X more potent than BPI vs. serum-resistant and smooth GNB. rBPI and rBPI23 compete with LPS-binding protein (LBP) for binding to LPS under physiological conditions. Consequently, both rBPI and rBPI23 block the CD14-dependent LPSinduced synthesis of the cytokines TNF, IL-1, EL-6 and IL-8 in vitro. rBPI23 has also been shown to inhibit the LPS-induced synthesis of reactive 02 metabolites, endothelial adhesion molecules and the procoagulant molecule tissue factor. In animals, rBPI has been reported to increase survival of endotoxin-challenged rats and mice, to inhibit the dermal Schwartzman reaction in rabbits and to increase survival of neutropenic rats with Pseudomonas bacteremia, rBPI23 increases survival and decreases cytokine production in endotoxin challenged mice and rats. It normalizes LPS-induced changes in hemodynamic, pulmonary and/or metabolic parameters in LPS-induced rats, rabbits and pigs. Treatment with rBPI23 also increases survival and decreases cytokine production in bacterial challenge models in rats and mice. rBPI23 was not toxic to rats after 10 daily consecutive i.v. doses of 10 mg/kg. This combination of properties indicate that recombinant BPI may be useful in the treatment of sepsis. Phase I/II clinical trials of rBPI23 have begun. The discovery of LPS binding protein (LBP) and subsequent identification of CD14 as a receptor for LPS or LPS-LBP complexes has resulted in a new understanding o£ how LPS responsive ceils are stimulated. CD14 is found either as a glycosylphosphatidyl-inositol (GPI)-anehored membrane glycoprotein (mCD14) of myeloid cells or as a soluble serum protein (sCD14) lacking the GPI-anchor. Binding of LPS to mCD 14 triggers cell activation while binding of LPS-sCD14 complexes to cells such as endothelial or epithelial cells that normally do not express mCD14 activates these cells. These pathways are shown in schematic form below. 4 ~DI mCD14 plays a crucial role in presentation of LPS to additional membrane components that make up a functional LPS receptor. An immediate consequence of engagement of this functional receptor is protein tyrosine phosphorylation. The molecular mechanisms leading to these events will be discussed. Understanding of these pathways will lead to the development of new therapeutic approaches to controlling host responses to LPS. Pretreatmen t Posttreatment (before or after TNF peak) D) with different antibody dosages: 15 mg/kg ---0.1 mg/kg Pretreatment with anti-TNFab prevented death in most model situations (except peritonitis), but also posttreatment up to 4h after sepsis induction was successful in the few studies performed. There is additional evidence that low-dose TNFab is partially effective. Especially baboon anti-TNFab studies provided many insights into the pathophysiological sequences of sepsis induction, due to crossreactivity with human reagents. Those events include the cytokine sequence with TNF-dependent IL-I, IL-6, or IL-8, but also IL-lra or sTNF receptor release. Granulocyte as well as endothelial cell activation were shown to be partly TNF related, and the procoagulatory response was influenced by anti-TNF treatment. From many animal studies the concept that TNF plays a pivotal role in sepsis is clearly evident and therefore anti-TNF therapy is a major candidate tbr clinical studies. The beneficial or harmful effects of TNF-mediated inflammatory responses depend on the clinical context. Decreasing exaggerated TNF-mediated inflammatory responses may be useful in some patients with organ failure. TNFR:Fc (Immunex, Seattle, WA) is a recombinant human protein composed of two identical extracellular p80 TNF receptors linked by the Fc region of IgGl. It neutralizes TNF with an affinity for TNF<z of 10-10 M and has a T1/2 of • 63h in normal humans. Methods: 18 normal volunteers were randomized to receive either TNFR:Fc vehicle (n=6), or low dose (LD) TNFR:Fc (10 mg/m 2 iv, n=6), or high dose (HD) TNFR:Fc (60 mg/m 2 iv, n=6) infused over 30 rain prior to endotoxin (E) (Escherichia coil 0:113; 4 ng/kg iv) administration. Plasma cytokines were measured using ELISA and TNF bioactiv'dy. Systemic hemodynamics were evaluated with indwelling arterial and pulmonary artery catheters and radionuclide heart scans and compared before and after fluid loading in subjects given vehicle with HD TNFR:Fc. Results: Endotoxin-related symptoms were unchanged after LD or HD. Peak temperature occurred at a later time point (4-5h) in subjects given TNFR:Fc compared to vehicle (3h) (p=.004). At 3h, 5h (post fluid load ), and 8h, subjects given vehicle developed increased cardiac index and heart rate and decreased systemic vascular resistance index (all p=.0001 ). HD TNFR:Fc did not alter this hyperdynamic cardiovascular response. LD and HD inhibited the E-induced teukopenta at I h post endotoxin (p<.03) and later leukocytosis (p<0.05). Peak TNF bioactivity in subjects given vehicle was 154 + 44 pg/mL and <10 pg/mL in the LD or HD TNFR:Fc groups (p<.001). Two immunoassays for TNF were performed because the detection of receptor bound TNF varies with ELISA. TNF (Biosource) was decreased in LD vs vehicle or HD (p=.0001) whereas TNF (R&D Systems) was increased after HD or LD vs vehicle (p<.001). Decreased levels of IL-8 (p=.00t), granulocyte colony stimulating factor (p=.03), and IL-1 receptor antagonist (p=.001) were found after LD or HD vs vehicle. IL-6 levels were lower after LD vs vehicle or HD (p=.006). IL-10 levels were similar among subjects given vehicle, LD, and HD TNFR:Fc. Conclusions: LD and HD TNFR:Fc delayed the febrile response to E but did not alter the early hyperdynamic cardiac response. This suggests that mediators other than TNF play an important role in the initial cardiovascular response to E in humans. TNFR:Fc attenuated the release of some E and TNF-induced cytokines. Ti~e antiinflammatory responses of TNFR:Fc may be useful in some patients with disease processes resulting from excessive TNF-mediated inflammatory responses. The biosynthesis of TNF within macrophages is regulated both at the transcriptional and the translational levels. The 5' flanking region of the TNF gene contains several transcriptional control elements, including two KB enhancers similar to those of immunoglobulin genes and a "Y box" similar to that of the MHC class II promoter. These elements are critical for the enhancement of transcription noted after stimulation of macrophages with LPS. Despite these and other elements, transcriptional regulation alone accounts neither for the absence of TNF protein during normal homeostasis nor the profound increasein TNF production following stimulation with LPS or other secretagogues.

The translation of TNF mRNA into protein is also tightly regulated via mechanisms involving the octameric "TTATTTAT" motif present in the 3'untranslated region of TNF, human inducible NO synthase, and several other cytokine genes. This region not only destabilizes mRNA, but also confers a translational blockade so that TNF protein is not normally synthesized despite leaky basal transcription. LPS stimulation releases translational blockade and, together with enhanced trancriptional rate, accounts for significantly increased TNF secretion. Opportunities for inhibition of TNF production are available at each of these two levels. Agents which increase intracellular cAMP (pentoxif¥1ine, amrinone) inhibit accumulation of TNF mRNA; in contrast, corticosteroids act primarily at the level of translation, inhibiting LPS induced translational activation. Understanding the regulation of TNF biosynthesis may assist in clinical strategies designed to regulate TNF secretion. The serine protease thrombin is formed by enzymatic conversion from its proenzyme form after coagulation activation and may enhance the inflammatory response in various ways. In the last step of the coagulation cascade, thrombin cleaves its substrate fibrinogen, thus forming fibrin. The fibrin clot may prevent phagocytosis of bacteria and promote local bacterial growth. Thrombin can also stimulate a specific th.rombin receptor that exists on a variety of cells. For example, minute amounts of thrombin ma), stimulate platelets to develop tissue factor and boost the activation of more thrombin, thus initiating a positive feedback loop. Other cellular effects of thrombin receptor stimulation include increase in intracellular calcium of platelets and monocytes, contraction of smooth muscle cells, increase in endothelial permeability, thromboxane generation by neutrophils and lymphocytes, pulmonary vasoconstriction, and enhanced cytokine production.

Animal experiments have been conducted with a variety of thrombin inhibitors including heparin, antithrombin IlI, hiradin, and hirudin-like peptides. An alternative approach is the inhibition of the coagulation cascade at an earlier point in the cascades, or administration of the anticoagulant enzyme protein C Work from our own group in a porcine model of lipopolysaccharide-(LPS)-induced shock with disseminated intravascular coagulation has shown that prelreatment with recombinant desulfatohirudin (r-H) or with a preformed complex of human donor antithrombin III with heparin or post treatment with r-H are able to block the degradation of fibrinogen and the formation of fibrin monomer. In addition, after pretreatment with r-H it was found that LPS-induced lung injury was reduced. Since both natural hirudin as well as r-H were well tolerated by healthy volunteers, adminislration of r-H may be a promising therapeutic approach in patients with sepsis and SIRS.

Corticosteroids have been used in the management of various shock syndroms including sepsis. The data concerning the effects of this therapy is, however, conflicting and in large prospective studies, the use of early administration of high doses of corticosteroids have been found to give no beneficial effects in patients with sepsis. We have particularly been engaged in studies on possible effects of corticosteroids on the plasma cascade systems and on the mediator release from leukocytes. In in vitro studies, high doses of methylprednisolone were found to facilitate endotoxin induced generation of plasma kallikrein. Furthermore, plasma prekallikrein and kallikrein inhibitor values decreased together with formation of kallikrein-C1 inhibitor complexes (1). In this in vitro plasma model we also found that methylpredoisolone alone in doses of 1 mg/ml induced contact activation with the formation of kallikrein-C1 inhibitor complexes. Methylprednisolone was also found to have an additive effect on endotoxine induced decreases in kallikrein inhibitor functions. In the same experimental model on endotoxemia methylprednisoione was found to give an additive effect on activation of the initial part of the complement cascade (2) . Activation of the terminal part of complement, however, was inhibited by the same dose of methylprednisolone. Addition of methylprednisolone alone to plasma was also found to activate the initial part of complement. Using a full blood model, we studied the effect of methylpredoisolone in modulating the endotoxin induced cytokine response. When 103 ng/l E. coli endotoxin was added to citrated blood from healthy human donors preinkubation with 20 gg/ml methylprednisolone significantly reduced the release of tumor necrosis factor et (79 %) and interleukin-6 (63 % ) at two hours after exposure to endotoxin.

In conclusion, these studies show that corticosteroids might facilitate contact activation of plasma and reduce the function of major protease inhibitors. Furthermore, we also observed that this drag in in vitro conditions facilitated activation of complement. On the other hand, corticosteroids were found to reduce endotoxin induced cytokine release from leukocytes in whole blood. These studies also showed a dose dependence of the effects of corticosteroids on endotoxin induced cellular and cascade system activation. Different predisposing conditions like extensive trauma, infection or pancreatitis result in a remarkably similar inflammatory response. Although thls inflammatory response primarily aims at the removal of devitalized tissues, destruction of bacteria and reestablishment of the integrity of host tissues, it may potentially become unregulated and generalized and thereby producing deleterious effects to vital organs or the body" as a whole. Cytokines play an important role in the development of the systemic inflammatory response. Most of their biological effects, however, are not directly mediated but exerted through other mediator systems. When the inflammatory response results in the formation of capillary leak syndrome and refractory hypotension, the unregulated activation of complement and contact systems appears to be the cause. Both systems are almost exclusively regulated by C1-1nhibitor. Am unbalanced degradation of the inhibitor protein was found to be associated with the onset of capillary leakage in various conditions such as bone marrow transplantation, severe burns, and septic shock. These observations suggest a relative deficiency of biologically active C3-1nhibitor in these septic shock like syndromes. Therapeutic intervention studies were initiated using high doses of C1-1nhibitor concentrate.

First results are promising, as in most patients the administration of the drug was followed by a quick and marked improvement of the patient's hemodynamics. Whether the administration of C1-1nhibitor alone or in combination with other drugs is also sufficient to improve long-time survival has to be investigated in double-blind, placebocontrolled studies.

SALUTARY EFFECT OF HUMAN SOLUBLE COMPLEMENT RECEPTOR TYPE-1 IN MODELS OF ADULT RESPIRATORY DISTRESS SYNDROME G. Feuerstein, M. DiMartino, and *R. Rabinovici, SmithKline Beecham Pharmaceuticals, King of Prussia, PA, USA, and *Jefferson Medical College, Philadelphia, PA, USA Adult Respiratory Distress Syndrome (ARDS) is a severe pulmonary insufficiency syndrome associated with diverse diseases and injuries. ARDS is believed to be the consequence of a massive acute inflammatory reaction consisting of activated neutrophils infiltration into the lung. Neutrophil activation is likely the result of multiple factors of which complement is believed to play a major role. A potent complement regulatory system in the form of the Complement Receptor Type-1 (CR-1) exists on many cells where protection against complement injury is provided by inhibition of both the alternative and classical pathways for C3/C5 convertase formation. We have recently purified the human recombinant CR-1 in a truncated form (BRL55730, TP-IO, sCR-1) and tested this soluble protein for protective effects in several models of ARDS. Specifically, we have used the following rat models: 1) endotoxin-induced ARDS in PAF primed animals; 2) 11.-2induced ARDS; 3) thermal injury (30% body surface)-induced ARDS; 4) acid aspiration-induced ARDS. ARDS-like features included edema (increase in lung wet weight and water content), neutrophil accumulation (histological inspection and myeloperoxidase assay) and increase in cells and protein in the bronchoalveolar lavage (BAD. sCR-1, 1-25 mg/kg, i,v., given as prophylactic (and in some cases, therapeutic) treatment significantly inhibited edema formation and leukocyte infiltration and, in some cases (e.g., endotoxin/PAF or IL-2), virtually completely. These comprehensive studies strongly support the therapeutic potential of sCR-1 in ARDS due to diverse injuries. The xanthine derivative pentoxifylline (PTX) and several analogs of PTX have been evaluated for their protective effect in various animal models of septic shock. Data from these in vivo studies demonstrate that (1) PTX suppresses LPS-induced TNF release from activated macrophages; (2) PTX prevents TNF-induced PMN activation; (3) PTX attenuates LPS-or TNFinduced acute lung injury as evidenced by prevention of increased pulmonary vascular permeability and detoriation of gas exchange; (4) PTX exerts its .protective effects even when administered following initiation of sepsis. The sum of these actions suggest that administration of PTX in sepsis may have therapeutic potential.

Hans Hoffmann MD, Dept. of Surgery, Klinikum Grosshadern, Ludwig-Maximilians-Universit~.t, MarchioninJstrasse 15, D-81377 M0nchen, Germany.

Pentoxifylline [POF] and related xanthins are drugs of known haemorrheologieal activity and widely used clinically. Recently, new pharmacological aspects of these established drugs could be demonstrated. It was found that POF selectively inhibits the formation of TNF but not IL-6 most likely by causing accumulation of intraceliular cAMP via inhibiting phosphodiestersse activity, and, secondly, by inducing prostacyclin in different cell types. Furthermore, POF is able to counteract TNF-mediated adherence of neutrophils to vascular endothelium and to inhibit FMLP-indueed respiratory burst in neutrophils. We and others have, therefore, studied its effect in different animal models. It was found that POF protected from increased pulmonary vascular permeability and sequestration of neutrophils into the lung in different animal models of acute lung injury. Moreover, in pigs with induced faecal peritonitis the drug improved haemodyrmmle variables as well as pulmonary compliance and reduced the number of pulmonary neutrophila and lymphocytes. Consequently, as a general outcome, POF could be found to improve survival in different models of haemorrhagie and endotoxie shock. The protective effect of POF was dose-dependent and observed even when POF was given up to 4 hours after endotoxin. In order to evaluate these pharmacological effects of POF in humans, we studied POF under controlled conditions of endotoxlnemia in volunteers. We found that POF selectively inhibits the LPS-indueed endogenous formation of TNF without affecting IL-6 and IL-8. Moreover, POF counteracts the LPS-indueed initial leukocytopenia by interfering with the adherence of neutrophils in the microvasculature. Meanwhile the inhibition of endogenous formation of TNF by POF could be demonstrated under different clinical conditions of acute and chronic cytokine release syndromes, {OKT3 first-dose reaction, cancer-or tuberculosis-induced cachexla}. It appears worthwile and promising to investigate wether POF exhibits beneficial effects also in septic patients. Objectives: To determine the specific role of PAF in both lethal primate bacteremia and nonlethal human endotoxemia. Materials and methods: Two double-blinded placebo controlled studies were performed. First, ten baboons (14.5 +.7 kg), were randomized to a control group receiving 101° cfu/kg of intravenously administered E. co/i (n=5) and a treatment group (n =5), pretreated with .1 mg/kg of PAF receptor antagonist Ro24-4736 two hours prior to E. co/i infusion. In a second study, ten healthy male volunteers were randomized to a control group receiving 4 ng/kg of intravenously administered endotoxin (LPS) (n=5) and a treatment group (n=5) pretreated with 10 mg of Ro24-4736 eighteen hours prior to LPS administration. Physiologic parameters and cytokine levels were measured continuously in both studies for 24 hours. Results: Baboons pretreated with the PAF antagonist had improved oxygenation (96 + 11 mm Hg v 59 -+ 6 in controls, p < 0.05) and creatinine clearance 12 hours post E.coli (39.8 + 23.4 ml/min v 0.1 _+ 0.1 in controls, p < 0.05). Similarly, volunteers pretreated with the PAF antagonist experienced fewer symptoms, including rigors and myalgias at 1 hour post LPS. This was in concert with a diminished peak cortisol (668 +_ 107 mmot/L v 959 +-159 mmol/L in controls, p < 0.05), epinephrine secretion (1057 + 165 nmol/L v 2029 + 431 nmol/L in controls, p < 0.05). Hemodynamics and circulating TNFa levels (data not shown) were unaffected in either study by prior PAF antagonism. Concluslons: PAF influences some components of the multi-organ failure syndrome in lethal gram negative sepsis and the counter-regulatory hormonal system in human endotoxemia through TNF-independent mechanisms. PAF antagonists may serve as adjuncts to cytokine blockade in the treatment of gram negative sepsis. The mediator Role of Platelet-Activating Factor in Gramnegative Septic Shock Pierre G. Braquet, David Hosford and Matyas Koltai Institut Henri Beaufour, Le Plessis Robinson, France Considerable evidence has been accumulated to support the concept that a PAF/cytokine autogenerated feedback network is responsible for priming and amplification of inflammatory mediator release in septic shock. Cytokines, such as TNF~x, IL-1 and other interleukins interact with PAF which then amplifies cytokine production, therefore, PAF may be the principle mediator in endotoxin shock. A single factor identified as TNF is suggested to play a pivotal role in the fatal phase of septic shock. Presumably excess TNF release is the result of amplification process primarily triggered by PAF. One may assume that the high TNF concentration in the blood of 'non survivors' is the consequence rather than the cause of cell death. The dubious results of clinical trials with anti-TNF antibodies and IL-1Ra, a naturally occurring IL-1 antagonist protein, have risen debate around the exceptional role of cytokines in the pathomechanism of septic shock or systemic inflammatory response syndrome (SIRS) induced by Gramnegative microorganisms. There are similar problems with the interpretation of the results obtained in clinical trials of monoclonal IgM antibodies against E. coli endotoxin, HA-1A and E5. Future studies will answer the question as to the effectivity of inefficiency of this treatment. Perhaps when the plasma concentrations of LPS and cytokines exceed a critical level, treatment fails to save the patients life. With regard the causal role of LPS in septic shock, the putative role of bacterial porins may have to be taken into consideration. The marked release of PAF induced by endotoxin leading to excess TXA2 production may be responsible for the microvascular failure and death in septic shock. Antagonists of PAF markedly protect against the release of TXA2, and may interrupt the vicious circle of amplification process. NO produced by the inducible NO synthase plays a pivotal role in causing vascular paralysis in LPS-induced SIRS. To explore the relationship of PAF to NO synthesis will provide better understanding of the pathomechanism of septic shock. Several PAF antagonists, including BN 52021, have undergone phase II and phase III clinical trials. The strategy in the treatment of SIRS, however, remains multifactorial, since little is known about the SIRS caused by microorganisms other than Gram-negative bacteria. Evidence has begun to accumulate which suggests that in sepsis excess production of LPS and cytokines can lead to uncontrolled production of iNOS and that this might in part explain the severe unresponsive hypotension seen in some septic patients. A number of strategies have been explored in an attempt to limit excess NO production. A favourite target has been competitive inhibition of NOS by arginine analogues such as L-NMMA. In animals, some investigators have shown that the haemodynamic effects of LPS challenge can be attenuated by L-NMMA, but there is a narrow toxic-therapeutic ratio. In a model of live E. cell sepsis, we have been unable to reproduce these findings. Localisation of NO production by immunohistochemistry suggests that NO production is compartmentalised, and more subtle methods of regulation may be required if this approach is to have clinical valuc.

Anti-adhesion molecule strategies CH Wortel, Repligen Corporation, One Kendall Square, Building 700, Cambridge, MA 02139, USA. The neutrophil has been implicated as a pivotal player in the pathogenesis of multiple organ dysfunction. An important first step in the process of neutrophilmediated organ injury involves the binding of neutrophils to endothelial ceils. This process is largely regulated by complementary adhesion molecules, some of which are present constitutively on the cell surface and others that can be upregulated in response to chemotactic and pro-inflammatory stimuli. Several different adhesion molecules have been described. L-selec tin is expressed on the plasma membrane of neutrophils and is shed from the surface early in the inflammatory process. It is therefore thought to mediate the initial interactions with endothelial cells. E-selectin and P-selectin are endothelial adhesion molecules. E-selectin is not constitutively expressed but is upregulated by inflammatory mediators, particularly IL-1 and TNF. P-selectin is present in the Weibel-Palade bodies within the endothelial cytoplasm and can be upregulated rapidly in response to thrombin, histamine, and oxidants. All selectins recognize oligosaccharides, particularly sialylated Lewis X antigen. This carbohydrate moiety is expressed on many proteins, including the selectins themselves. The leukocyte integrins are glycoproteins expressed on the neutrophil surface and in the cytoplasmic granules. Integrins consist of a common beta2 or cluster differentiation (CD)I 8 chain covalently linked to one of three different alpha chains (CD11 a, CD1 lb, CD1 lc) and exist on the cell surface as three distinct heterodimers. CD1 l a/CDI 8 is expressed on all leukocytes, whereas CD1 l b/ CD18 and CD1 lc/CD18 are restricted to cells of myeloid origin. CD1 I/CD18 interacts with intracellular adhesion molecule-1 (ICAM-1), its ligand on endothelial cells. ICAM-1 is a member of the immunoglobulin family of adhesion molecules. It is constitutively expressed on endothelial cells and can be upregulated by cytokines. The potential for monoclonal antibodies to adhesion proteins to reduce vascular and tissue damage has been studied in alarge number of experimental models. Protective effects have been observed in a wide variety of inflammatory, immune, and ischemia-reperfusion injuries such as arthritis, bums, endotoxic shock, bacterial meningitis, autoimmune diabetes, nerve degeneration, allograft rejection, allergic asthma, acute lung inflammation, skin lesions, and ischemia-reperfusion models of the intestine, myocardium, lung, skeletal muscle, and central nervous system. Protective effects have also been observed in animals resuscitated following hemorrhagic shock. Since modulating the function of adhesion molecules may temporarily leave the host without effective defense machinery, safety evaluations are of utmost importance in evaluating this therapeutic approach.

Treatment of Gram-negative septic shock with an immunoglobulin preparation: A prospective, randomized clinical trial INGOLF SCHEDEL, URSULA DREIKHAUSEN; BIRGIT NENTWIG; MARION HOCKENSCHNtEDER, DIRK RAUTHMANN, SALIM BALIKCIOGLU, ROLF COLDEWEy, HELMUTH DEICHER, MD Endotoxin may play a major role in the pathogenesls of muitiorgan failure in the context of the toxic process in septicemia induced by gram-negative pathogens. The hypothesis which has led to a prospective clinical study consisted in the idea of inhibiting the biological effects ofendotoxin during the early phase of septic process, and thereby attampting to attain a positive effect on the clinical course of the disease.

-Objecave: To evaluate the effectiveness ofa polyclonal immunoglobulln (Ig) preparation containing IgG, lgM, and IgA as an adjunctive therapy for septic shock.

-Desigru Prospective, randomized clinical trial. Patients: Fifty-five patients w~th septic shock were randomly allocated to two groups according to criteria of septic shock.

-Intervention: One group of patients (n = 27) received a commercially available immunoglobulin preparation (containing high titers of antibodies specific for determinants to bacteria1 endotoxin) during the first 3 days after incIosion in the study. The other randomized group (n = 28) did not receive any immunogiobulin preparation.

-Measuremems and Main ResMts: During the period of <6 wks after the beginning of clinically apparent septic shock, death related to the septic process occurred in one (4%) of 27 patients who received immunoglobulln. By comparison, nine (32%) of 28 control group patients died during this period (p <. 01 ). Within the first 48 hrs after onset of the clinically apparent septie process, significaptly increased activity of circulating endotoxin and simultaneously decreased specifie IgG serum titers to lipid A were detected in the group of nonsurvivors. The rationale for the use of polyvalent intravenous immunogiobulins (IVIG) to treat severe infections stems from in vitro and animal studies documenting that high-dose IgG enhance opsonization and phagocytosis of pathogenic bacteria. Moreover, recent clinical studies have shown that in septic patients the phagocytic function of circulating polymorphonuclear neutrophils (PMN) is defective; the degree of such impairment correlates to the severity of the septic state, The beneficial effect of administering high dose IgO in septic patients can be two-fold: I) IVIG provides large quantities of antibodies against invading pathogens; the IgG bound to the pathogens can opsonize their phagocytosis; 2) the opsonization of phagocytosis by exogenously administred polyvalent IgO would be of special importance in those patients who present a significant defect of their phagocytic function, when they are challanged by a large bacterial invasion. A prospective, randomized, double-blind study was carried out comparing the administration of IVIG (Sandoglobulin@) vs, paeebo (human albumin) in severely septic surgical patients, Only patients with gepis score >_20 (meaning a mortality risk >70%) were accepted into this protocol. Patients were randomized to receive 0.4 g/Kg of IVIG or placebo on day 0 (when they reached sepsis score>20), repeated on day +1 and +5. At the beginning of ICU treatment, the two groups of patients were similar for severity of sepsis, age, concomitant disease, type of surgical procedures, antra and perioperative procedures, antibiotic administration. The results of the study indicated a significantly reduced mortality in patients with severe surgical sepsis treated with IVIG as compared to placebo control patients (mortality: 38% vs, 67% respectively; p< 0,05). In conclusion, the results of our study in patients with severe surgical sepsis were the following: 1) IVIG plus multimodal treatment of sepsis, including antibiotics, reduce mortality significantly', 2) the reduction of mortality seems to be due to a decreased incidence of lethal septic shock.

Despite substantial clinical research, the avallable data regarding the effectiveness of supplemental immunoglobulin (Ig) treatment in sepsis in adult patients do not yet allow definitive conclusions. In view of the persistently high sepsis mortality there is a need to continue clinical investxqations regarding supplemental sepsis treatmen~ in general, as well as concerning Ig administration in particular. We present and discuss the protocol of the ongoing ,,Score-Based-Immuneglobulin Therapy of Sepsis (SBITS)" study. The protocol (Theoret Surg 8(1993)61-83) of this multicenter, randomized, prospective and double-blind trfal relies on the results of an observational trial on i.v. IgG treatment in 163 patients with sepsis and septic shock (Infection 19~1991)216-227), carried out as a prerequisite for the present trial. Using microcomputer-based bedside routine score monitoring, we regard quantitative measures of severity of disease and sepsis: only patients with a certain degree of both severity of disease (APACHE II score 20-35) and severity of sepsis (Elebute sepsis score 12-27) will be included.

By observing these previously validated inclusion criteria, this trial snould iQentify a priori and include patients with potentially optimal response to therapy, consisting o~ either placebo (0.i % albumin) or Polyglobin N" -12 ml (0.6g)/kg on day 0 and 6ml (0.3 g)/kg on day i. With an anticipatedpopulation size of 800 patients the study should comply with the statlstical requirements (estimated mortality: 30%, with a 33 % reduction in 28-day mortality in the treatment groupl to prove or disprove the question of IgG effectiveness in sepsis in terms of improved prognosis. Up to November 1993, more than 460 patients had been included; patient enrollment will be finished in 1994. Previous studies have demonstrated rhlL-I ra, a naturally occurring antagonist of IL-1, increases survival in animal models of andotoxemia and EscheHchia coli bacteremia and attenuates the decrease in mean arterial pressure resulting from challenge with both gram-negative and gram-positive bacteria. Previously, in 99 patients, rhlL-lra was demonstrated to increase survival in patients with sepsis syndrome and septic shock in a dose-dependent manner. Methods: A randomized, double-blind, placebo-controlled, malticenter, clinical trial enrolled 893 patients at 63 academic medical centers in Europe aad North America. Eligible patients received either placebo (vehicle) or rhIL-lra (anakinra) 1.0 or 2.0 mg/kg/hr by continuous intravenous infusion for 72 hours. The presence of organ dysfunction (i.e., ARDS, DIC, renal, and hepatic) at study entry was determined prospectively by a Clinical Evaluation Committee using definitions which were developed a-priori. Survival time was evaluated over 28 days utilizing a linear dose-response model, assuming a log-normal distribution. Results: 563 patients had one or more sepsis-induced organ dysfunction(s) at study entry. A dose-related increase in survival time was observed with rhlL-lra compared to placebo in patients with ARDS, DIC, and renal dysfunction (P --< 0.04 Endotoxin infusion releases platelet-activating factor (PAF), a potent phospholipid mediator which leads to an autocatalytic amplification of cytokine release. BN 52021 (Ginkgolide B), a natural PAF receptor antagonist, has provided significant protection against sepsis in different animal models• A randomized, placebo-controlled, double blind, multicenter trial on efficacy (mortality at D28) and tolerance of BN 52021 (2 IV infusion of 120 mg x 2/day over 4 days) in severe sepsis has enrolled 262 pts. The 28day mortality rate was 51% for the placebo group and 42% for the BN 52021 group (P = .17). The efficacy of BN 52021 was greater in pts with gram-negative sepsis: the 28-day mortality rate was 57% for the placebo group and 33% for the BN 52021 group (P = .01). BN 52021 also reduced mortality among pts with gram-negative septic shock (mortality was 65% for placebo vs 37% for BN 52021; P = .01). Using statistical adjusments for pronostic factors, the relative risk of death of the BN 52021 group was 0.61 (0.34 -1.08, 95% confidence interval; P = .09). This risk corresponds to an adjusted reduction in mortality of 39% for pts receiving BN 52021. No differences in mortality rates were found between the placebo and the BN 52021 groups in the absence of gram-negative sepsis• There were no differences in adverse events between the placebo and the BN 52021 groups. BN 52021 is a safe and promising treatment for patients with severe gram-negative sepsis. A confirming study, focused on gram negative sepsis, is In progress.

V~3lliam A. Kanus M.D. and The rhlL-lra

It has been traditional within the field of infection and sepsis to think in terms of specific indications for drugs based on the type of infecting organisms, Advances in antibiotic therapy now control or ltnflt the growth of bacteria. The majority of deaths are now caused by either an initial overwhelming response to infection or subsequent multiple organ system failure attributed, in part, to the effects of intrinsic biologic responses of the host. Type of organism, therefore, may not be as critical as determining the exact severity of the host's severity or risk of death from infection. We also know that both the relative benefit of a new treatment across groups and its absolute benefit for an Individual patient will vary with their risk in a predictable fashion.

We recently iuve~iguted the relationship between one measure of host response, the acute risk of death as prospectively estimated by u comprehensive risk mode[ for 28-day mortality (JAMb. 1993; 270:12,33-1241) , by its retrospective application to the results from the Phase In evaluation of recombinant human intcrlenkin-1 receptor antagonist (rhlL. Ira). We found that there was a significant interaction between the patient's predicted risk of mortality at the time of entry to the study and the ability of rhIL-lra to prolong survival time (X2 = 7.6, p [] 0.02, log.normal) for all 893 patients in the trial• Survival benefit began st approximately 24% baseline risk of 28-day mortality. For the $80 patients with a predicted risk > 24%, there was a 22% reduction (P=0,00$ log normal). When we examined the variation in patients above and below the 24% risk level with hazard functions, i.e., their daily risk of death during the study period, we found that placebo patients With < 24% risk had lltile acute daffy risk during the hlltial two days follawh~g study entry and this risk was little affected by rhIL-lra, In contrast, patients with > 24% risk had high daily mortality risks during the tuttlal two days that high dose rhtL-lro substantially reduced.

These results are compatible with our current understanding of outcome from sepsis and the proposed mechanism of action o£ immunotherapy, The earliest deaths from sop sis are secondary to an immediate inflammatory response followed closely by deaths secondary to multiple organ system failure, Later deaths (after 14 days) are not as closely related to the acute effeete of the inflammatory cascade. Because of the timing and action of most proposed tmmunotherapy, they may be capable of preventing mortality primarily in these initial two phases. In this study, an independent predicted risk of mortality reflected this mortality pattern ned illustrated the potential benefit of immtmotherapy. Use of a predicted risk of mortality in the design and analysis of clinical trials could improve our understanding of the clinical benefit of these new therapeutic approaches. The systemic inflammatory response syndrome (SIRS) is a term recently proposed to describe patients with systemic inflammatory responses to insults such as infections (sepsis), trauma, burns, pancreatitis, and other initiating events. Patients with SIRS may have similar activation of inflammatory mediators and similar outcomes independent of the initiating event. These outcomes include organ dysfunction and failure, shock, and death. Challenges to the successful conduct of clinical trials in SIRS include the complexity of illness in these patients and the important--but limited--clinical benefits of novel compounds that may be limited to selected patient subsets. Addressing these challenges will require new tools and approaches. These will include more sensitive and appropriate endpoints, and the use of methods such as baseline risk adjustment, to allow detection of drug risk interactions not captured adequately by categorical definitions, such as sepsis syndrome.

On the basis of supportive preclinical and Phase I safety studies, we have initiated Phase II clinical trials of a novel bradykinin antagonist, CP-0127, in four SIRS subcategofies: sepsis, multiple trauma, burns, and pancreatitis. Each of these studies is designed to measure the effect of CP-0127 on mortality, organ dysfunction and failure, and activation of mediators. In addition to investigating rates of organ failure using standard definitions--a new endpoint--a continuous summary measure of organ dysfunction (the Acute Physiology Score of APACHE TM III) is being used to quantify the degree of organ dysfunction and the speed and pattern of recovery of physiologic stability.

In the sepsis study, another new approach--a study specific risk model based on the APACHE Ill database--has been developed which will be used to assign a pre-treatment baseline risk to each patient enrolled. The primary outcome variable will be risk adjusted survival time to 28 days. This type of risk-adjusted analysis may allow for more efficient and powerful trials and more accurate and useful indications for use. Study PurPose: In post-cardiac surgical patients (pat.) at risk for sepsis, the efficacy of early i.v. immunoglobulin (Ig) treatment was compared to a matching historical control (con.) population. Postoperative risk assessment: Using APACHE II scores LAP) (first postoperative [pop.] day) in a pilot study phase, we were able to differentiate between the large population (95.2%) of pop. low-risk pat. (AP< 19; mortality: 1%) and the small groups of pop. pat. at risk LAP= 19-23) and high risk LAP_24) with a significantly higher mortality (14% and 76%, mainly due to sepsis). Subsequently, among 1341 consecutive pop. pat. we prospectively identified and treated these pat. Iq treatment reqimens: First study period (n =41): (gG (Psomaglobin N a, Tropon Biologische Pr~parate, Cologne, FRG, day 1:8 ml/kg, day 2:4 ml/kg). Second study period (n=25): IgGMA (Pentaglobin R, Biotest, Dreieich, FRG, 5 ml/kg on days 1 to 3). Results: Ig pat. and con. were comparable in demographic data, operation characteristics and baseline disease severity LAP and Elebute sepsis scores). In contrast to con. (risk: n=21, high-risk: n-21), the Ig pat. showed a marked improvement in disease severity (fall in AP), especially in the high-risk group (IgG, n=26: p<O.05; IgGMA, n=13: p: 0.08). In this group, Ig led to higher (p<0.05) response rates, defined as an AP score decrease ->7 within four days (IgG: 54%, IgGMA: 62%; con.: 19%), and reduction in mortality (IgG: 46%, IgGMA: 46%; con.: 76%), statistically significant (p<0.05) for Ig treatment as a whole (IgG and IgGMA). Conclusion: Given the good comparability of the study groups, our results indicate, despite the non-randomized design, that early supplemental Ig treatment can improve disease severity and may improve prognosis in prospectively APACHE II score-identified high-risk patients after cardiac surgery. Objective. Elevated plasma levels of Endothelin (ET) have been demonstrated in both experimental and human sepsis. ET has been proposed as a sepsis mediator leading to vasoconstriction with tissue hypoperfusion and organ failure. The aim of the study was to determine the effects of sepsis treatment with volume resuscitation, antibiotics and the anti-LPS monoclonal antibody ES® on big ET and active, 21 aminoacids ET (ET 21) in rat abdominal sepsis. Methods. Lethal peritonitis was induced with a 4 mm coecal perforation (CP) in male Wistar rats. Plasma levels of big ET and ET 21 were determined with Amersham TM Endothelin RIAs 4, 8 and 12 h after sepsis induction. Experimental groups: 1. CP control, 2. Volume replacement (VR); 0,9% saline 10 ml/kg/h continous iv infusion started after 2 h, 3. Antibiotic; imipenem 40 mg/kg iv after 2 h, 4. E5®; 10 mg/kg iv after 2 h, 5. VR + imipenem + ES® after 2 h. Results. High concentrations of both big ET and ET 21 could be demonstrated after 2 h and lasting for 12 h after CP. Neither volume replacement nor imipenem did influence the elevated plasma ET. E5® significantly reduced ET 21 both 4, 8 and 12 h after sepsis induction, but did not reduce big ET. When ES® was combined with VR and imipenem, reduction of ET 21 was the same as for E5® alone. These results strongly suggest that bacteria and hypovolemia per se are not decisive stimuli for ET production during sepsis. E5® reduces circulating LPS and TNF which is the probable mechanism of the suppressed ET 21 synthesis. The unaltered big ET fraction after E5® treatment indicates conversion of big ET to ET 21 as the site of action responsible for reduced ET 21.

Conclusion. Lethal peritonitis in the rat is followed by elevated plasma levels of big ET and ET 21. E5® anti-LPS antibody significantly reduces plasma ET 21 while volume resuscitation and antibiotics failed to do the same. ES® did not reduce plasma big ET. PMX treatment on severe endotoxemia with multiple organ failure was safety and effect in prognosis, and sepsis related parameters. It was certified that reduction of plasma endotoxin was effective in severe endotoxemia.

A. Lechleuthner,S. Aymaz, G. Grass, C. Stosch, S. Dimmeler, M. Nagelschmidt, E. Neugebauer. II. Dept. Surgery, University of Cologne, Germany. Introduction: The cardiovascular therapy of hypodynarnic shock states is a challenging problem. In clinical as well as experimental studies beneficial functions of a new Hg-agonist BU-E-75 in congestive heart failure has been demonstrated 03aumann, 1989). Therefore, we investigated the effect of BU-E-75 in hypodynamic shock in pigs. Materials and Methods: Pigs (Deutsches Hausschwein, Pitrain, [20] [21] [22] [23] [24] [25] were anesthesized with Fentanyl/Dormicum, ventilated (N20:O 2 = 2:1) and cardiovascular parameters were monitored with a complete ICU-eqnipment. The hypodynamic model was established in a pilot study (4 animals) to evaluate the effective concentration of BUE-75 in healthy and endotoxin (LPS)-treated animals. Endotoxic shock was induced by continous infusion of 10 ~g LPS/kgKG/h (055:B5, Fa. Difco). The hypodynamic state was defined as a decrease of cardiac output by 30 % of steady state levels. A wedge pressure of 10-12 mmHg was kept constant by volume resucitation during the experiment. In a subsequent randomized controlled trial (RTC) 3 groups with 6 animals per group were studied. The groups were treated as follows: group I, LPS and 0,9 % NaC1; group II, LPS and BU-E-75 (100 #g/kgKG/h); group III, Famotidine (H2-blocker) pretreatment (1 mg/kgKG), LPS and BU-E-75. Results: The pilot study in healthy pigs revealed, that BU-E-75 had positive inotropic effects. These effects were inhibited by the H 2antagonist Famotidin. BU-E-75 however had no beneficial effects in the hypodynamic phase of endotoxic shock in the RCT. Cardiac Index (CI) and the oxygen delivery (DO2) were not significantly influenced by BU-E-75 application (group I versus group II). BU-E-75 did not ameliorate the negative inotropic effect measuring left ventricular stroke work (LVSW) in hypodynamic shock phases. On the contrary, BU-E-75 led to a further significant decrease of LVSW (p < 0,05). Famotidin pretreatment did not affect the response (group III versus group II). Conclusion: In hypodynamic shock states the H2-agonism seemed to have no beneficial effect under these experimental conditions. Receptor down regulation or changes of signal transduction under septic conditions may be responsible. Cellular studies may help to identify these mechanisms. Objectives. Antithrombin III inactivation of proccagulant proteases is so far the only inhibitory therapeutic approach to disseminated intravascutar coagulation (DIC). We therefore set out to investigate whether Cll substitution reduces coagulation activation in an endotoxin induced rabbit DIC model. Materials and Methods. 29 male rabbits Chbb:HM(SPF) were randomty assigned to one of the following groups. Group K : NaCI 0.9% (control without endotoxin, n=10). Group E : Endotoxin 120 tJg kg "1 bolus i.v. + NaCI 0.9% (control with endotoxin, n=10). Group C : Endotoxin 120 pg kg -1 bolus i.v. + Cll 400 U kg -1 bolus + 400 U kg "1 4h "~ i,v. (treatment group, n=9). All animals were anesthetized and mechanically ventilated. Blood samples were drawn prior to endotoxin administration (M1) and after 120 (M2) and 240 rain. (M3). Thereafter, lung and liver tissue samples were taken intravitatly in a standardized fashion for H&E microscopic fibrin quantification using a triple score (FIBS). From all blood samples the prothrombin time (PT), activated partial thromboplastin time (aPTT), fibrin monomers (FM), and d-dimers (DD) were measured. For statistical significance of differences between the groups ANOVAs and the Wilcoxon test (FIBS) were performed. Results. FIBS for lung/liver were significantly different (p<0.05) between group E (lung 180, liver 159) and C (lung 89, liver 0) (Group K : lung 11, liver 0). , a synthetic serine proteinase inhibitor, has an anticoagulant activity in the absence of" antithrobim III. Gabexate has been reported to be useful in the treatment of disseminated intravascular coaguiation due to neoplastic diseases. In this study, we investigated gabexate therapy for the treatment of DIC due to sepsis in the postoperative critical patients. MATERIALS AND METHODS: From July 1992 to June 1993, 15 patients in the surgical intensive care unit met the criteria of DIC or pre-DIC. Eleven were male and four were female with the mean age of 66.7 years. All these patients suffered from some complication of operations which led to the development of sepsis. FOY was administered at the rate of 2mg/kg/hr untiI the coagulation profile retumed to normal or the patient died. The coagulation parameters were monitored before and on the 1st, 3rd, 5th and 7th day. RESULTS: Fourteen of these fifteen patients died despite transient improvement of the coagulation parameters in five patients. These patients suffered from sepsis resulting from surgical complications which could not be well controlled. The only survival was a case of recurrent intrahepatic duct stone with biliary tract infection complicated with sepsis and DIC. After choledocholithotomy and the use of FOY, the patient recovered gradually. CONCLUSION: DIC is a late manifestation of sepsis in the critical surgical patients. The most important thing is to eradicate the cause of sepsis. If the underlying septic focus cannot be controlled, DIC will persist despite the use of gabexate mesilate.

Emergency Surgery, Taipei Veterans General Hospital, Taipei, Taiwan. There are 2 main types of bradykinin (BK) receptor, namely BK~ and BK z. The BK 2 receptor is constitutive. The BK 1 receptor is also constitutive but in the majority of cases is inducible and involved in chronic inflammatory syndromes such as sepsis, hyperalgesia and airways hyperreactivty in animals. The mechanism(s) involved in the upregulation of the BK 1 receptor is unclear, however a variety of agents including LPS, E Coil and ILl are particularly efficacious in vitro and in vivo. ILl and Bradykinin acting at their respective receptors are believed to be involved in SIRS/Sepsis. We have investigated the effect of antagonists at ILl (Antril), BK 2 (Bradycor [CP-0127]),BK~ (CP-0298) and BKz/BK 1 (CP-0364) receptors on the de novo generation of BK~ receptors (reflected by hypotensive responses to a BK 1 agonist) in the LPS-treated (10ug iv) rabbit.

In LPS treated rabbits hypotensive responses to BK~ but not BK 2 agonists increased with time and at time 210min appeared maximally induced. Constant iv infusions of CP-0127 blocked BK 2 but not BK~ and CP-0298 BK~ but not BK 2 responses. CP-0364,CP-0127+CP-0298 and Antril+CP-0364 blocked both BK 2 and BK~ responses. Antril alone had no effect on BK 2 or BK~ responses. Within 30-60 min after stopping the infusions of antagonists the responses to BK~ and BK z agonists were the same as those in nonantagonist infused rabbits. These results indicate, at least in the LPS-treated rabbit, that neither BK2,BK ~ or ILl receptors alone or in combination, are involved in the de novo generation of BK 1 receptors. In vitro studies demonstrated that beth Bradycor and CP-0364 (but not Antril) were antagonists at both BK z and BK~ receptors. If both BK z and BK 1 receptors are significantly involved in chronic inflammatory situations in man such as SIRS/sepsis then the rationale for the use of compounds such as Bradycor or CP-0364 is clear. Infection is a major cause of or contributor for morbidity and mortality in liver transplant recipients. Effectiveness of prophylactic and therapeutic protocols is important for the success of liver transplantation ( OLT ). SDD is used as prophylaxis for reduction of infection caused by Gram negative or fungal microorganisms. Between September 1988 and July 1993 400 OLT's in 368 patients were performed at our department. The actuarial 1-year patient survival is 90 %. Infection prophylaxis is started with SDD and ciprofloxacin once the patient is accepted as an OLT candidate. Perioperatively metronidazol, tobramycin and cefotaxim, postoperatively cotrimoxazol are prescribed additionally. The table shows Pneumonia, peritonitis, major wound and urinary tract infection are common nosocomial infections following severe injury. In a series of severely injured patients from the University of Louisville Hospital, pneumonia was the most common infection followed by peritonitis, intra-abdominal abscess formation and burn wound infection. Pneumonia is actually the leading cause of death from nosocomial infection. These are defined as occurring from 48 to 72 hours after hospital admission. This definition has important implications for antibiotic therapy because the likely pathogens and their respective sensitivities are different for community acquired pneumonia. The diagnosis of nosocomial pneumonia is difficult following major injury as many patients will have pre-existing fever, leukocytosis, tachypnea, and chest X-ray changes. Reliance on sputum gram stain and culture is important and best obtained by a bronchoalveolar lavage or protected specimen brush during bronchoscopy. Predisposing risk factors include severe head injury, emergent intubation and shock, and such patients have been shown to benefit by early tracheostomy. Staph aureus has been the most common pathogen isolated from the sputum and the remainder gram-negative organisms with Pseudomonas aeruginosa, and Klebsiella pneumonia predominating. Bacteria recovered by site as well as by intensive care unit is published in the six month antibiogram which also includes recent antibiotic sensitivities. This aids in empiric antibiotic selection against such nosocomial organisms. In a series of severely injured patients (ISS -30), mean temp. was 101.4F, leukocytosis was 16k, Pan2 was 87, Fin2 was .47, and PEEP was 5.1 at the time of diagnosis (ARDS excluded). There was marked reduction in class II histocompatibility antigen (HLA-DR) density on peripheral and BAL monocyte/macrophages which recovered over time with resolution of pneumonia. Immune suppression occurred prior to development of pneumonia, was especially localized to the infected tissue, but recovered with clinical improvement. Specific immune modulation targeted to pulmonary white cells may hasten clinical recovery and minimize pulmonary dysfunction.

-CLINICAL EXPERIENCE J. Tnllemar Amphntericin B remains the drug of choice for many systemic fungal infections. Its advantages include a broad spectrum of activity and intravenous administration. The major disadvantages of amphoterlcin B is its severe side-effects, especially the nephrotoxicity. To decrease the toxic side..cffccts various liposomal amphoteficin B formulations have been produced. It was found that these liposemal formulations were as effective as amphotericin B but in contrast had a low incidence of toxicity. At present there are three ~different variations of lipid formulations under assessment: Amphotericin B lipid complex (ABLC), amphotericin B coloidal dispersion (ABCD) or true liposomes. The ABLC has a ribbon like structure. It has been shown to have a reduced toxicity and an efficacy ranging from being as effective to four times less effective that conventional amphotericin B. Regarding ABCD the particles have a disk-like structure with a diameter of around t00 am and a thickness of 4nm. The ami-fungal efficacy is 2-4 times less than that of conventional amphotedcin B. Both ABLC and ABCD are presently investigated in phase II/III studies in the US. AmBiseme is currently the only commefieally available true lipesome. AmBiseme is a spherical small unilamellar lipesome with a diameter less than 100 nm with a mutina LD50 of > 175 mg/kg. It has been used in dosages up to 6 mg/kg/day in compassionate based studies with good tolerability. The mycological efficacy range from a 83% response rate for invasive Candida infections to 41% response rate for Aspergillosis. AmBisomc have been evaluated as anti-fungal prophylaxis in randomized trials in bone marrow (BMT) and liver transplant (LTX) recipients. It was well tolerated. In BMT recipients the incidence of proven fungal infections was 8% among placebo treated patients compared to 3% for the AmBisome treated patients (ns). In LTX recipients AmBisome prophylaxis was effective, significantly reducing the incidence of deep fungal infections from 16% to 0% ill placebo and AmBisome treated patients respectively (p<0.01). Prospective randomized trials comparing these various amphotericin B preparations with conventional amphotericin B is needed to determine their future place in the therapeutical arsenal. Two patlentgroups ere particularly at risk to develop serious CMV disease: CMV seronegative transplant recipients of seroposltlva donors and those patlants treated for rejection with anti T-ceil preparations, We have evaluated the value of prophylactic anti-CMV immunoglobulin (Cytotect", Biotest Pbarma GmbH, Dreieich, FRG) administration in high risk heart and kidney transplant recipients, In a double blind placebo Controlled study 40 kidney transplant recipients, treated for biopsy proved re)action with rabbit ATG, received globullntplacebo Infusions. The preparatlons were given I,v, In a dose of 100 mg/kg at day 0,7,14,35,56 and 77 after the Initiation of anti = rejection therapy, Passive immunization completely prevented CMV related death, although it did not reduce th~ incidence of CMV isolation, viraemia or disease, This effect was mainly observed in CMV saronegativa recipients of a serop0sitive donorktdney. Seroposltive recipients did not benefit from treatment and seronegatlve recipients of a seronegetlye donor were not et risk for CMV infection at e!l. In a open study the incidence of CMV Infection and disease was evaluated In 143 consecutive I~eart sllograft recipients. Sixty-five patients were CMV seronagatlve and they all received passive immunlzation according to the dosage schedule used in the kidney patients, but starting on the day of transplantation, This scheme resulted in median sntI-CMV IgG titers of 1200 ELISA units during 3 months. CMV infection occurred in 21/65 ~eronegetlve and in 40/81 seropositive recipients (n,s,), In ssronegetive donor-recipients pairs the incidence was significantly lower (3/34] , The passively Immunized seronegstive recipients of e seroposltlve donorheart showed comparable incidence of CMV infection f18t29) vs the seropositive recipients. Primary infection more often resulted in disease than secondary infection (11121 v8 10/40), but no difference in incidence of disease (11165 vs 10/81 ) or severity in symptoms was noted between the Immunoglobulln treated serone(]ative patients and the seropositiva recipients.

Apparently passive immunization induces anti-CMV immunity which crossly resembles naturally acquired resistance.

Abdulkadirov K.,Chebotkevich V., Moiseev S.

The incidence of infection is still high in patients underwent BMT. This complication is the major cause of mortality if it is not recognized and treated promptly and properly. Our data showed that from 21 patients with different types of leucemia after autologous and allogenzc BMT 19 had the episodes of fever. In the ma i ority of these episodes the bacterial etiolog$ gram negative bacflli and gram positive cocci) can be proved. On the other hand, in 62% of the fever cases we detected also viral respiratory (corona-, adeno-, RS-and other) infection. Our previous investigations showed that even in healthy persons the viral infection has influence on antibacterial immunity, in the cases of model experimental reaction in volunteers we found the decrease of delayed hypersensitivity 10-14 days after intranasal inoculation of influenza virus A (H3N2-3) to bacterial (Staphylococcal, Streptococcal and Pneumococcal) and ~iycoplasma pneumoniae antigens in the leucocyte migration inhibition test. These results showed that respiratory viruses may be the important pathogenic factor in the development of bacterial infection in posttransplanted period. We consider the constant control of latent and visual respiratory viral infection in BMT patients to be very important. FIccB the ~ter£~Li of the Nation~l Institute of TraD/~atoloqy in Budapest 1.00 consecutive cases of revision hip grafting were carried out arthroplasties wlth hemoloquous bone between the years 1990 and 1993. In the same period of time 732 pri~ total hlp replacen~nts were performed under i4entieal technical conditions. The average septic rate for the 'total hip althroplasties was less than 1%. In the selected i00 cases the septic rate was 4% indicating the role of bone grafting° Homografts were prepared by deep freezing~ It .is recognized that the cells of the hL~grafts become destroyed by the ium~unological, response of the host~ and the patients develop ~ti-HL~, ar~tib'o~ies. The dead ~trix, however, has a bone-inducing capacity that stimulates host osteoblasts to recolonize the *I~/trix Which serves as scaffolding. The sequence of events favours the infections. For this reason, beside preventive perioperative systemic ant/biotic treatment, local ~ntibioties were also applied in the form of antibiotic-//npregnated cement. The role of age and the .immune status of the patients .is discussed.. The purpose of this study is to evaluate the rate of toxemia in patients with acute panereatitis and to find this coudition to the activation of cascade systems that are encountered in the subsequent complications of the disease. We studied a series of 45 patients with acute pancreatitis, the severeness of which was evaluated by the Ranson's criteria and the Apach-II scoring system. All of them were considered to have severe acute puncreatitis. The determination of toxemia was made using the Limulus test (LAL test). We also determined the levels of the third (C3) and fourth (C4) Complement components as weU as the coagulation factors, Iibrinolysis faeters and Kimns by serial measurements. The severity of the disease was serially determined by the Apach-II scoring system. It was found that Complement activation ( which was also assessed using a graphically illustrated method by a aggregometer ) was followed by an increase of morbitity and mortality .We also detected that toxemia (positive LAL-test) was closely correlated with complement activation and more of the Ranson's criteria. A clear relation existed between the number of Ranson's signs and the enmplieations' rate (1"= -0.766, p < 0.001). The documentation of toxemia and the complement activation cannot predict the kind and the severity of complications.

The study of coagulation, fibrinolysis and kinms systems didn't reveal any results with statistical significance. Necrotizing pancreatitis still represents a life-threatenthg disease. Infectious complications dominate among the causes of death. Differences in the individual immune response could possibly explain different clinical courses even in patients with comparable pancreatic morphology. To explore the inflammatory response in acute pancreatitis, the following investigation was performed. Methods: Peripheral-venous blood was withdrawn on admission and furthermore twice weekly in as yet 36 patients with acute pancreatitis and tested for the parameters mentioned below. In parallel, polymorphounciear granaIocytes were isolated using density gradient centrifugation and assessed for superoxide anion and hydroxyl radical producing capacity using electron spin resonance techniques. Results: Total leukocyte cotmt and total lymphocyte count did neither reflect the clinical course nor predict complications. This comes tree also for serum IgG, IgM, IgA, C3, C4, CRP, alpha-l-antitrypsin and neopterth as well as for plasma IL-la, IL-IB, IL-1RA, IL-2, IL-2R, IL-6R, TNF-ct, TNF-~R (p75) and ICAM-1. In contrast, PMN-elastase, IL-6 and IL-8 closely correlated to the clinical course. Isolated PMN's in vitro capacity to produce oxygen radicals depended on the respective radical species and was slightly elevated (superoxide anions) or decreased (hydroxyl radicals), respectively. Patients with a CD4+/CD8+ ratio below i were seen at risk of developing septic complications. In contrast, a percentage of monocytes of 30 % or more among total mononuclear cells indicated an uncomplicated course, in general.

Conclusions: The immune status of the individual patient may significantly influence the course of acute pancreatitis. The cytokine pattern in peripheral blood is very complex and most parameters are of little use for the clinician. The PMN-elastase, IL-6 and IL-8, however, closely correlate to the clinical course and may prove valuable for follow-up. The CD4+/CD8+ ratio was found the best predictor of septic complications, but it failed in non-septic patients. A percentage of 30 % or more of monocytes among total mononuclear ceils indicated a rather mild course. The reduced ability of the PMNs to produce hydroxyl radicals may help to explain the frequent development of septic complications in severe necmtizing pancreatitis. Peroxidation of membrane lipids contributes to ceil injury in pancreatitis.

Overwhelming release of toxic metabolites by infiltrating neutrophils is regarded a major pathogenetic factor, too. As yet little is known about the mechanisms by which oxidative stress and leukocytes damage pancreatic cells. The present study examines (I) the susceptibility of pancreatic acinar cells to attacks by oxidants and leukocytes and (2) the potential of antioxidants to prevent such damage in order to better understand the cellular mechanisms of pancreatic injury in inflammatory states. Methods: Freshly isolated rat pancreatic acinar ceils were exposed to a model system of oxidative stress consisting of 20 mU/ml xanthine oxidase (XOD), 500 mM hypoxanthine (HX), 50 mM FeC13 and 75 mM EDTA. In a second set of experiments, acinar cells were exposed to excess autologous neutrophils or neutrophils obtained from patients with acute pancreatitis. Neutrophils were stimulated by zymosan A, PMA, and IL-8. Cell viability was assessed by both cellular uptake of trypan blue (TB) and by release of LDH. Results: The XOD/HX system caused a time-dependent acinar cell injury. This injury was effectively prevented by catalase (CAT) and gfutathione peroxidase (GPx). In comrast, superoxide dismutase (SOD) enhanced cell injury. Addition of both SOD and CAT abolished the damage seen with SOD alone. The non-enzymatic scavengers mannitol, DMSO, DMTU and the iron chelator deferoxamine were not protective and at a higher concentration even accelerated cell decline. The newly developed antioxidants of the lazaroid type effectively prevented oxidative acinar cell damage. Stimulated neutrophils, both autologous and heterologous, did not damage healthy acinar cells but had even protective effects. Conclusion: Pancreatic acinar ceils are very susceptible to oxidative injury. A combination of catalase and SOD prevented cell damage effectively. SOD when given alone may rather damage than protect aelnar cells when H202 is generated in concentrations overwhelming the capacity of endogenous catalase. Therapeutic approaches to pancreatic disease using antioxidants should, therefore, include combinations of protective substances. The lazaroids seem to be candidates for clinical use as antioxidants in pancreatitis. The results argue against direct toxic effects of stimulated neutrophils to pancreatic acinar cells. are ch~act~z~ by the presence of a polymicrobial flora, The pmtotyI~ cffthese inf~ons is secend~,y bacterial pedtonitlw, whereby a pathololoeal process in the ~trointesfimd tract r~ful~ in tim disrup~on ofi~ inteffrlty and ¢ollseqtlent sptl]nge of inte~.i,o~.l gontents into the peritoneal c~iry. The ensuing infection invariably contains a mixtm~ of gt~m negative enteric bacilli, gram positive b~eria and anaerobe& Experimental and clinical =t~ies have de~ed the eantrlbution of each of th¢~ components to ti~ ovemU virulence of these in~ons, Gram negative enteri~ such as F.veher~chla coil ere endowed with a virulent l~l~x~lyse~haride ptill~ly t~sponsible for lethality, By contrast, Bacteroldes sl~cles, which rarely c~se death, prornot~ abscess fOnllation, A uniqm~ capsul~ polyseccluu'ide, particularly on B.j~ogilJs slrai~, oontributes to tJtis erect, Several mecltanims have bccn pml~ed whereby or~ microorganism mi~t interact with its microbial ~net to augment the overall virulence of a r~xed im~edan. These include: l) provision of nutrients by one apexes which stimulates the growth of its ~opathoge& 2) inhibition of host deletes by one of the migroorganisms so that the other microbes might persist and exert their virulence, 3) the trant~ of viM.©n~e traits between ~renr~a.,dsms and 4) the ~.mizatian d the mi~oe~vironmental con~tion$ by one d the baetez'isl pa#, so that the other might persist. Exampl~ for each of these m~banisms Imv~ been provided by experimental ttudies i~stigating E.co!l-B.p~flls synergistic in~ra~ons. Byproducts ofg.coli metabolim l~¢ovide essential short ebath fatty acids £~ optimal B,frosili~ ga'owth. Fm-ther, oxygen ¢ons~tmption by Kcelt lowers oxygen tension end redox potantial to levels eomlucive to B#a#lts gro~h. Coawr~ely, B,~agtlis rolea~s proteases and fatty acids wl~¢h impair pl'tsgocy~¢ ~lt Rmctlon tnd permit F-..¢oli proliferation and expression of its intrinsic virulent.

In summaxy, interactions among the separate microbial cemponents of mixed infections heighten the overall virttienee of these lafectiot~, This knowledge provides ~r rationale for targetting of antibiotic therapy against the knowa eantributors of these synergistic pro~¢sses,

Intraabdominal abscess formation and the macrophage William G. Cheadle, M.D., Department of Surgery, University of Louisville School of Medicine, Louisville, KY 40292

Inflammation of the peritoneal cavity following bacterial contamination has been classified into primary, secondary and tertiary, the last two relating to bacteria originating from the gastrointestinal lumen. The natural history of such infection is either resolution without clinical sequelae, which is uncommon, abscess formation, or generalized peritonitis, which occurs as a result of failure of peritoneal host defenses. Early clearance of microorganisms by peritoneal fluid circulation and filtration througti subdiaphragmatic lymphatics into the thoracic duct and systemic circulation occurs as well. Simultaneously peritoneal macrophages and the omentum approach the area of inflammation and lead to neutrophil influx and abscess formation adjacent to the affected viscus. We have found a shift in peritoneal macrophage function from antigen presentation to proinflarnmatory cytokine production that occurs early after experimental peritonitis produced by cecal ligation and puncture. This is also reflected by reduced class II histocompatibility antigen expression on peripheral blood mononuclear cells and peritoneal macrophages. This is accempauied by an influx of both neutrophils and macrophages into the peritoneum and subsequent abscess formation. Interestingly, there is little serum endotoxin or TNF seen in this model despite TNF mRNA expression in peritoneal macrophages. We believe this model is more clinically relevant than other models of endotoxemia or bacteremia in which different patterns of cytokine expression are seen. Newer agents aimed at reduction of systemic manifestations of sepsis originating from intra-abdominal infection such as monoclonal antibodies against cytokines or IL-1 receptor antagonists may need to be directed against remote organ macrophage populations while preserving peritoneal macrophage function.

Inflammation is a complex process involving microcirculatory changes, extravasation of fluid and a cellular influx in the affected body area.

In our communication, we will only consider the regulation of the cellular infiltrate which plays a major role in the defense of the peritoneum against microbial invasion. Until recently, it was thought that the influx of leukocytes in the abdomen was induced by bacterial products, local humeral factors and secretions of resident macrophages.

There is now increasing evidence that this view is too simplistic. Many other cell types present in the abdominal cavity or composing the peritoneal membrane (mast-cells, mesothelial cells, fibroblasts) are able to release or secrete vasoactive or chemotactic substances such as histamine, prostagtandines, or cytokines. They are most likely to play a role in the regulation of intraperitoneal inflammatory reactions.

The emigration of leukocytes towards the abdominal cavity is also modulated by a previous contact with Gram negative bacteria.

In the rat, this intriguing phenomenon is long lasting, cannot be transferred by serum and seems independent from T lymphocytes. The clinical relevance of these various regulating mechanisms has still to be determined.

Kinnaert Paul, H6pital Erasme, Route de Lennik 808, 1070 Bruxelles Belgium

Generalized response in secondary peritonitis

The clinical course of an intraabdominal infection may depend on a variety of variables including the capacity of host defense mechanisms and the degree of the inflammatory response. If local defense mechanisms fail to restrict the inflammation to the abdominal cavity a generalized inflammatory reponse will result. in a first stage generalized signs of a local inflammation become detectable whereas the second stage comprises the overwhelming systemic inflammatory response. The extent of this systemic response determines the outcome. Sometimes it may appear to be unrelated to the severity of the intraperitoneal findings. The activation of plasma systems and cellular elements leads to a fast release of cytokines, inflammatory mediators and other substances. These parameters precisely reflect the degree of the generalized response. Inflammation of the peritoneum causes significant morbidity.

OBJEKTIVES: To test the hypothesis that peritoneal mesothelial cells play a role in regulating inflammatory responses within the peritoneal cavity, we examined neutrophil-chemotactic activity (Interleukin 8) and monocyte-chemotactic cytokine (MCP) release by sytokine-etimulated mesothelial cells.

Confluent human peritoneal mesothelial cells were exposed to varying concentrations of phorbolmyristate-acetate (PMA) and the cytokines tumorneerosis factor a (TNF a) and Interleukin i~ (IL-I~). The supernatant was examined for IL-8 by ELISA and for MCP by investigating the ehemotactic activity for isolated human monocytes.

Mesothelial cells express low levels of IL 8 and monocyte chemotactic activity when cultured. These activies were significantly increased (2-fold) after stimulation with either TNF a or IL-I~. Additionally macrophage inflammatory protein was detected.

These observations provide a probably important mechanism whereby peritoneal mesothelial cells respond to imflammatory stimuli released during peritonitis and how leucocyte recruitment by liberation of chemotactic cytokines is regulated. The perioperative course of LPS, TNFa and IL-6 in 19 patients with bacteriologic proven abdominal infection (intraabdominal abscess 10, diffuse peritonitis 4, pancreatic necrosis 2, pancreatic abscess 3) was followed prospectively and evaluated for possible correlation with septic state and organ function. Methods: Patients were studied in a 6 to 10 hours period during their first surgical intervention because of intraabdominal infection. All were monitored for their cardiovascular, respiratory, hepatic and renal function. Plasma samples for LPS. TNFa and IL-6 determination were drawn preoperatively, intraoperatively, and until 2h postoperatively in regular intervals (min 7/pat), Results: Preoperative APACHE II was 12 in median (rain 4, max 25). 10 patients fulfilled the criteria of SIRS. 4 of them were in septic shock.There was a significant correlation between preoperative TNFa and APACHE II (p=0,000 I, Spearman coefficient). Preoperative cardiovascular (systol. RR<90 mmHg) and respiratory (PaO2<75mm Hg) dysfunction were associated with significantly elevated TNFa (cardial: p=0,000 I, Wilcoxon; pulmonal: p=0,0003) and IL-6 (Cardial: p=0,2; pulmonal: p=0.0001) Overall, LPS, TNFa and IL-6 values varied considerably during the observation period. However, TNFa was markedly higher in patients with SIRS and septic shock (group A: n= I 0, mean 182 pg/ml) than in those who did not fulfill these criteria (group B; n=9, mean 24 pg/ml; p=0,00 I, Wilcoxon). IL-6 was significantly higher in group A (mean 2169 pg/ml) than in group B (mean 191 pg/ml; p=0,0O I Wilcoxon). Conclusion: Perioperative TNFa and IL-6 were shown to correlate significantly with preoperative organ function, APACHE II and the severity of sepsis. These results could help to define patients that might benefit from further therapeutic strategies, e.g. antibody administration.

Department of Surgery, University Vienna, AKH Wien, Wahringer Gurtel 18-20, 1090 Wien.

Aim of the study: The purpose of this pilot study was to establish and to prove a standardized reproducible animal model of intraperitoneal sepsis induced by E.coli-endotoxinaemia in LEW.lW-rats in order to investigate early immunoserological responses to find a mediator based evaluating system of peritonitis sepsis. Materials and methods: In 30 LEW. lW-rats, diffuse peritonitis was induced by intraperitoneal injection of a mixture of E.coli (KhU +) and autogenous haemoglobin solution. In the control animal group (n= 12) an intraperitoneally injection of physiological saline solution was done. Blood samples were obtained by heart puncture after 6 hours. Stastistieal calculations were performed on a personal computer with the SPSS programm Vers. 4.01 (correlation with Pearson's R, Mann-Whitney-U-Test, descriptives statistics, discriminant analysis). Results: In contrast to the sham treated rats, the peritonitis animals showed significant differences in the concentrations of endotoxin, interferon-gamma (WN-y), the pteridin derivate biopterin and serum PLA2-activities [endotoxin range from 0.064 EU/I, SD=0.19 to 102.38 EU/1, SD-145.4 (p < 0 0001), IFN-¥ levels, range from 147.9 pg/ml, SD-141.6, to 4591 pg/ml, SD=6541 (p < 0.0001), circulating PLA2-activities range from 116.2, SD=45.4 to 185.4 U/1, SD=105.4 (p < 0.01) and biopterin range from 54.4 nmol/l SD=17.2 to 127.8 nmol/l, SD=76.8 (p < 0.001)]. For the peritonitis group we found strong correlations between the degree of endotoxinaemia to elevated levels of IFN-'~ (rp = 0.72, p < 0.0001) and bioptefin synthesis (rv= 0.82, p < 0.0001). The increase of IFN-T levels was correlated to the regulatory synthesis of biopterin (r = 0 73 p < .

.. P • , .

0.0001) and to the PLA2-actwtUes (rp = 0.50, p < 0.005). The biopterin synthes~s correlates slightly with the PLA2-actn,ities (rp= :0.38; p < 0.05). Using the para, meters of endotoxin, IFN-y levels, biopterin and the PLA~ -activities only, the statistical procedure of the linear discriminant analysis makes it possible, to distinguish between non-septic animals and septic animals correctly at a rate of 87 %. Anaerobes were found in 53.4%, anaerobes were isolated in 22.8%. There were aerobic and anaerobic associations in 12.1% and microflora was not found in 11.6% of the cases. Express method of anaerobes discovering let to receive information on 1 -3 days early than in generally accepted nethods.

Intraaotal transfusion of oxygenate blood and laser irradiation of blood reduces the duration of anaerobic sow, disminishes intoxication and accelerate the patients recovery.

Patients with abdominal sepsis are subject to long periods of hospitalization and high associated morbidity and mortality rates. This category of patients is thus consuming extensive facilities and costs. As the age-related outcome of abdominal sepsis is not fully known, the aim of the present study was to investigate abdominal sepsis in the elderly. Out of 166 patients with abdominal sepsis treated at the surgical intensive care unit during a 10-year period, 72 (43 %) had an age of 70 years or more. 28 were women and 44 were men, a sex distribution not differing with patients younger than 70 years. The patients were scored according to APACHE II and septic severity score (SSS) upon arrival to the intensive care unit. Bacterial cultures, the occurrence of organ failure, hospitalization and outcome was noted.

In median two operations were performed for both "younger and elderly" patients. The median time of hospitalization in the elderly was 29 (-183) days including in median 10 days in the ICU. Figures in patients less than 70 years of age were comparable (34 (-293) days out of which in median 10 days in the ICU). APACHE II and SSS-scores did not significantly differ (14.2 vs 13 and 23.3 vs 26.5, respectively), between the groups. Neither did the incidence of organ failure differ (57/73 vs 77/94). However, the incidence of multiple organ failure was significantly lower in elderly patients (21/72 vs 42/94 (p < 0.001)). The mortality rate, however, did not differ between the groups (21/72 vs 26/94).

In conclusion, severe abdominal sepsis in the elderly was not associated with an increase in mortality, incidence of organ failure or hospital stay. With the help of light transmissional scanning electron microscopy morphology of erythrosytes of peripheric blood was studied in patients with different stages of diffuse peritonitis before and after intravascu!ar irradiation of blood with heliun-neon laser.

Peritoneal morphology was investigated in patients who died from peritonitis, It was established that in all phases of peritonitis occured stomatocytoric and echinocytoric transformation of erythrocytes which progressed simultaneously with increase of intoxication.

It combined with strongly pronounced vessels variability of microcirculatory peritoneal bed which displaied by erythrocytes aggregation, stasis and microtrombogenesis.

In intravascular laser irradiation of blood number of erythrocytes which underwent to stomatocytoric and echinooytorie transformation was lower than in patients without laser irradiation.

It indicated that the intravascular irradiation of blood with helium-neon laser can prevent development of severe alterations of rheological property of blood and consequently variability of microcirlatory peritoneal bed in patients with diffuse peritonitis.

Abdominal sepsis is still associated with high morbidity and mortality rates, frequenfly caused by multiple organ failure. It has been reported that changes in capillary permeability play a role in the pathogenesis of multiple organ failure. The present study aimed at evaluating the influence of intraabdominal sepsis induced by cekal ligation and puncture on capillary permeability in multiple organs and tissues.

96 adult male Sprague-Dawley rats were subjected to laparotomy with separation of the cekum (sham operation) or induction of intraabdominal sepsis by cekal ligation and puneatre (n--48 in each group). At 3, 6, 12, 24, 48 and 72 hours (n=6/timepoint), the animals were evaluated concerning mortality and capillary permeability as determined by the passage of :25I-labelled albumin from capillaries to the peritoneum, the proximal and distal small intestine, cekum, colon, spleen, kidneys, lungs.

The mortality rate in rats with intraabdominal sepsis was 33 % both at 48 and 72 hours. Capillary permeability in the peritoneum, cekum, colon and kidneys significantly increased from 6 hours and on in rats with intraabdominal sepsis. In septic animals, capillary permeability in the lungs and spleen increased from 12 hours and on and in the proximal and distal small intestine from 24 hours and on. Different types of alterations in capillary permeability seem to appear: 1) a temporary short increase e.g. in the proximal small intestine and spleen; 2) a temporary longer increase e.g. in the colon and kidneys; 3) a persisting increase e.g. in the peritoneum, cekum, distal small intestine and lungs.

We conclude that experimentally induced intraabdominal sepsis induces early alterations in capillary permeability in multiple organs and tissues. Such changes may contribute to explain the development of sepsis-induced multiple organ failure. Despite a number of significant advances in the care of burn and non-burn traumatic injury, infection and sepsis remain major causes of morbidity and mortality. The severe immunosuppresslon often seen in patients with severe trauma or large burns may predispose these patients to life threatening infections. Included among the many Immune alterations are changes In the Functional capabilities of neutrophlls (PMNs). We have examined the expression of the p 2 integrins (CD1 l a, b,c/CD18), and the Fc'?R (CD16, CD32, and CD64), as well as several functional parameters, on PMNs from thermal and non-thermal traumatic injury, PMNs were obtained from patients sustaining severe trauma (initial APACHE II score >10) or thermal injury (>30~ total body surface area, 15% full thickness), and healthy controls. The expression of CD11 b and c and to a lesser degree CDI 1 a was significantly reduced on PMNs. The expression of CD16 and CD32 but not CD64 was also significantly reduced. PMNs displaying this reduction In receptor expression have a significantly reduced ability to phagocytose bacteria and undergo the oxidative metabolic burst response. Thermal and traumatic injury result In global reduction in the expression of 132 integrins and FoR which may lead to decreased functional capabilities, These abnormalities may In turn account at least In part for the Increased rate of Infection in these patlems,

Institute, Dept. of Surgery, 2~1 8ethesda Ave, Cincinnalt, OH, USA, 45267-0S5B,

Antibiotic-phagocytic cell interactions: their effect on endotoxin release. C G C-emmet1, Dep[Baeteriolog.z, Univer_sitv of Glasgow, Scotlan~_d Increasingly it is recognised that pathogenic bacteria are capable of surviving intracellularly within phagocytic cells in addition to their capacity to produce disease whilst in the extracellular milieu. As well as providing protection from certain antibiotics which fail to penetrate the phagocyte, such intraceltular bacteria may be transported from the initial site of infection to a distant more vulnerable body site wherein they may proliferate.

It is also known that some antibiotics are capable of becoming concentrated within phagocytic cells mid displaying bioactivity therein. Such bioactivity might be responsible for the release of endotoxia #orn gram-negative bacteria which when liberated from the celt could ~gger the cytokine cascade. Anfib,.'otic-induced damage to the ultrastructure of bacteria can also occur when the target bacteria are exposed to low (sub-MIC) concentrations of certain drugs. Such bacteria may present quite altered surface components m host-defense cells as well as releasing biologically active ceil wall components such as endotoxin. The nature of these interactions at the cellular level as well as the consequences for the host will be discussed. New Jersey Medical School: UMD, Newark, NJ 07107 A technique of physiologic state classification has been developed based on the m~itlvariable analysis of patient derived data sets of seventeen physiologic variables.

These multivariable data sets obtained from critically ill patients requiring Intensive care, were aormallsed by the mean and the standard deviation of recoverin~ Trauma patients who were not critically Ill, The resulting normalized seventeen variable sets were then clustered. Seven independent data groupings were developed.

The normal stress response hyperdynamic state seen post-Trauma and in compensated sepsis (A Stets)/ metabolic insufficiency seen in septic decompsnsation (B StSte}; early (C,) and late (e2) respiratory insufficiency associated with ARDS; cardlogenlc dscompensation (n State); post-Trauma hyvolemla without shock (R Stats).

The Stats closest to a new patient's values allows patient classifi0atlon with regard to his previous physiologic state.

Classifying observations f~om patients who lived or died who fell into these physiologic states enables a probability of death (P death) to be obtalned. Utilizing this criteria for the staging of severity in 44 recent Trauma patients the physiologic States accurately and significantly predicted the likelihood that the patient had an increased circulating level of the eytoklnes TNF and IL-1. The probability of death (P death) as well as the cytoklne levels appear to be a function of the physiologic B State with the highest levels being seen in the B State of metabolic insufficiency and the C~ State of oombined respiratory and metabolic insqffioienoy characteristic of septlc ARDS. The increase in the magnltude of metabolic abnormalities associated with the transition from non-sepsls to septic A, septic B, or septic C z States WaS associated with an increasing probability of death (P denth)(mean A state =.28, mean B State = .57, mean ~ State = .61).

The accuraay of this estimate was prospectively analyzed in this group of 44 m~Itlple patients of whom 72% had sepsis and 36% had ssptlo ARDS.

The 22 survivors had a mean P death of 0.39 and the 22 deaths had a mean P death of 0.63. The severity of post-Trauma sepsis can be quantified by probability analysis and stra~Ifie~ by physiologic State. Serologic tests have not been extensively tes'~ed in surgical patients but seem to be of limited value. We use nystatin as the main form of chemoprophyhxis. Patients "~'ith signs of infection who do not rapidly improve with antibacterial therapy are candidates for anti-funsal therapy, Amphoteradn B remains the first llne of therapy although combination therapy '~'ith flueonazole is use;l with increasing freque~;c)', The recovery of C~dida from an antra-abdominal site represents a challenging problem, Anti~ngal therapy in such patients depends on the underlying disease, the nature of the infected material and overall patient risk.

Role of neural stimuli and pain

Principles and Practice of Anesthesiology

Effect of combined prednisolone, epidural analgesia and indomethacin on the systemic response after colonic surgery

Arginine: Biochemistry, physiology and therapeutic irnplications

Immunosfimulatory effects of arginine in normal and injured rats

Arginine stimulates lymphocyte immune response in heahhy humans

Rote of arginine in trauma, sepsis and immunity

Arginine enhances wound healing in humans

IF LaBrecque t, GV Campion t, and the rhlL-lra Phase I//Sepsis Syndrome Study Group The Cleveland Clinic Foundation

A murine-anti-human TNF-monoclonal antibody known as CB6 was the first ANTI-tnf MaB which was studied in a phase II multinational trial in the treatment of patients with severe sepsis.This was an open-label, dose-escalation trial consisting of 80 patients who were enrolled into one of four treatment groups: (1) 0.1 mg/kg of anti-TNF mAb, (2) 1.0 mg/kg, (3) 10 mg/kg or (4) 1.0 mg/kg at study entry and the second dose 24 hours later. The small sample size in each group (n=20) precludes detailed statistical inference in this study. Nonetheless, a considerable amount of useful information was obtained from this investigation. irst, this study demonstrated the clinical feasibility of specific anticytoldne therapy in septic patients. Second, the measurement systemic levels of TNF proved to be an elusive target; Interleukin-6 may prove to be a more useful indicator of cytokine activation. Third, immunologic reactions including TNF: anti-TNF mAb immune complexes and human anti-routine antibodies were frequently found in these patients. Despite their apparent lack of overt toxicity in this study, these immunologic reactions may complicate this form of anticytokine therapy. Additionally, the potential benefits of anti-TNF mAb therapy occur within the first 48 hours of therapeutic administration in these septic patients. Infecting organisms differ in their potential to induce TNF in vitro and these differences correlate with circulating TNF levels observed in septic patients. Rapid methods to define those patients most likely to respond to anticytokine therapy are needed to determine the ultimate therapeutic potential of these agents in clinical medicine. Wherry, J., Abraham E., Wunderink R., Silverman H., Perl T., Nasraway S., Levy H., Bone R., Wenzel R., Balk R., Allred R., Pennington J. and the TNFa MAb Sepsis Study Group.TNFa MAb (Bay x1351) is a murine monoclonal antibody raised against human tumor necrosis factor. TNF~ MAb has been shown to reduce morbidity and mortality in animal models of septic shock and has been safely administered to septic and non septic patients.To evaluate the efficacy and safety of TNF~ MAb in patients with sepsis syndrome, a prospective, multicentered, double-blind, placebo-controlled trial was conducted in 31 hospitals in North America. Patients were prospectively stratified into shock or nonshock groups and then randomized to receive a single intravenous infusion either of 15 mg/kg TNF~ MAb, 7.5 mg/kg TNF~ MAb or placebo (0.25% human albumin).Patients received standard aggressive medical/surgical care during the 28 day post dosing period.The three treatment arms were well balanced with respect to demographics, APACHE II score and other parameters. For all infused sepsis syndrome patients, those who received TNF~ MAb had slightly reduced 28 day all cause mortality compared to placebo. Among shock patients there was a more pronounced trend towards efficacy at day 28 post dosing with lower mortality rates in both active treatment arms. Among nonshock patients TN~ MAb did not appear beneficial. The initial clinical experience with a chimeric anti-TNF monoclonal antibody, cA2, was undertaken in septic patients. The objectives of the study were to determine the safety, pharmacokinetics and effects on cytokine levels of cA2. as a single infusion or in combination with HA-1A in septic patients. The study was conducted with the intent to progress to an efficacy trial based on the information collected.The trial was conducted in three stages. Stage 1 was an open label trial in which 4 groups of 5 patients each with the clinical diagnosis of sepsis received ascending doses of cA2 (0.1, 1, 5, 10 mg/kg). Stage 2 was a randomized, double blind study in which patients received a single dose of HA-1A (100 mg) and placebo or one of 3 doses of cA2 (1,5,10 mg/kg). Stage 3 was a randomized, double blind study in which patients received a single dose of placebo or one of 3 doses of cA2 (0.1, 1, 10 mg/kg). In addition to usual laboratory tests, the following assays were performed: chimeric anti-TNF concentration, anti-chimeric antibody, endotoxin, TNF, IL-1, and IL-6 levels.A total of 141 patients were enrolled from 13 clinical sites (20 in stage 1, 60 in stage 2 and 61 in stage 3). Primary analyses were performed on patients in stage 1 and 3. There were 65 patients who received cA2 exclusively and 16 patients received placebo. Administration of cA2 was well tolerated at doses up to 10 mg/kg. No patient discontinued treatment due to adverse events. Human anti-chimeric antibody responses were positive in 61% (30/49) of evaluated patients. Mean Cma × and AUC increased proportionally with increasing doses of cA2. The mean half-life was -70 hrs (58-96 hrs). A dose related decrease in TNF concentration was observed 1 hr post infusion of cA2. TNF is considered to be one of the central endogenous mediators for the inili'ation of the pathophysiological changes in patients with sepsis and septic shock. High TNF levels were demonstrated to correlate with patient outcome. Blocking or neutralising TNF with specific antibodies was effective in preventing death in some animal modets of sepsis. In a placebo controlled prospective randomized study we tested the mur~ne derived antibody MAK 195F. It is a F(ab')2 fragment. The fragment rather the complete antibody was selected in order to reduce the potential immunogenicity and to facilitate tissue penetration. 122 patients with severe sepsis or septic shdck were enrolied in the study, three different doses of MAK 195F or placebo were administered (0,1; 0,3 and I mg/kg) over a perid of 72 hours in random order. The patients were evaluated for side effects, hemodynamics, organ dysfunction, cytokines (IL 6, IL 8 and TNF), and outcome. At this time only an interim analysis of 60 patients is available I indicating that MAK 195F in all 3 dosage groups resulted in a decrease in IL 6. This contrasted to a further in crease of IL 6 in the placebo patients. No serious side effects have been reported so far. A more detailed analysis on all 122 patients in the study will be presented and discussed.$152 S154 Staubach,K.H., Otto, V., Kooistra,A,, RosenfeId,J.A., Bruch, H.P., Univ. Lfibeek, Germany Once endotoxinemia occurs in sepsis a vieieus cycle with translocation of ET can be established. Increasing the clearance capacity for ET would therapeutically be the ulimate aim. We developed a new ET on-line adsorption (AD) system in whole blood by means of Polymyxin B (PB) coupled eovalently to a matrix (acrylic particles) via a 13 atom-chain spacer. The detoxification capacity was 150 ug[ET/ml column material. The biocompatbility resulted in 90~ platelet recovery. The column contained 7 ml of ADmaterial and was sterilized by high steam autoclave, Anticoagulation was achieved by heparine 1.000 IU/h in the inflowline after bolus injection of 5.000 IU. HP was performed on 8 pigs at a rate of 50 ml/min by means of a roller-pump until the animals succumbed (H). 8 animals served as controls (C). Serum ET levels rose from 3.40 pg/ml to 74,9 pg/ml after 5 hours in the C and from 2.77 pg/ml only to 28 pg/ml in the H group after 7 hours whieh was highly significant. Survival time could be extrended from 216 to 313 min. Results are listed in the following L. Blinzler, P. Zaar, M. Leier, R. B(2rger, D. Heuser Clinic of Anaesthesiology , City Hospital Nuremberg, Germany Sepsis and multiple organ failure (MOF) are still related with poor prognosis inspire of pharmacological and technical progress. Impressed by revealing reports about blood purification the continuous veno-venous hemofiltration (CVVH) was used as supporting treatment beside the critical cam basic therapy of MOF.

From 1989 to 1991 78 consecutive patients were treated by CWH. MOF was caused by hemolrhagic-traumatic noxa in 21°,4 and by septic-toxic event in 79%. All patients required mechanical ventilation (FiO2>0, 5) . 86°4 showed hyperdynamic shock. 85% had renal and 53% hepatic failure. Medium APPACHE II score amounted to 29,8 points. CVVH was performed in postdilution mode with a polyamide membrane (FH 66) and high volume exchange (50 l/die). Anticoagulation was done with heparin. Hemofiltration in MOF was installed, when critical cam basic therapy including adequate respiratory and hemodynamic management, pamnteral nutrition, antibiotic treatment, etc., failed to stabilize organ functions.

During consequent application of CVVH most of these patients showed improvement of their clinical course. Pulmonary stabilization was seen in 71%, hemodynamic in 68% and renal in 66% of the cases. 42% of the patients survived and were discharged from hospital. 10 of 45 non-survivors (58%) died because of fatal MOF within 24h after admission to ICU. Patients with early application of CVVH in MOF showed a better survival rate.Mediators of MOF, i.e. products of the complement cascade measured in blood and nitrafiltrate by ELISA, were partially removed by CVVH. The testing ultrafiltrate by HPLC demonstrated decreasing spikes ofpolypeptides during hemofiltration.

MOF seems to be generated by cascade-activation of immune competent cells and plasmatic mediators (e.g. bmdykinin, eicosanoides, cytokines, anaphylatoxins, etc.). Therapeutic approaches aim to inactivate or eliminate single substances. CWH with high-flux membranes in combination with high-volume exchange allows elimination of many mediators with different molecular weight and therefore may contribute to improve the prognosis of MOF. Other significant advantages of this teqalnique like adequate nutrition, optimized fluid balance and control of body temperature should not be negicctod. Introductioni Pseudomonas (P) aeruginosa has to be considered an important pathogen of nosocomial pneumonia and septic organ failure. The lung seems to be the predominant target organ for the pore-forming P. aeruginosa cytotoxin, thus inducing microvascular injury. With respect to therapeutical consequences, the potential protective effects of PAF-antagonist (WEB 2086), cyelooxygenase inhibitor (diclofenac) and specific and unspecific antibodies on cytotoxin-induced pulmonary vascular reaction and mediator release were studied in the isolated perfused rabbit lung. Methods: Cytotoxin (6p_g/ml) was administered into the perfusion fluid in all groups, either in the absence of inhibitors (n=6), or after pretreatment with WEB 2086 (5xl0-gM, n=6), or diclofenac (10#g/ml, n-6). Furthermore, the application of specific antitoxin (mg/ml, n=6) was tested in comparison with the unspecific immunoglobulins (Venimmun®, Behring, 7.5 mg/ml) (n=6) and the combination of immunogiobulins, WEB 2086 and diclofenac (n=6). Six experiments without toxin served as controls. The arterial pressure mad the weight gain as an indicator of edema formation were continuously monitored during the three hour peffusion phase. Arachidonic-ucid metabolites, as well as lactate dehydrogenase (LDH) and K + concentrations were determined at 30 rain intervals. Results: Cytotoxin caused a gradual increase in pulmonary arterial pressure, reaching a maximum value of 2.5 times higher than the control, starting after 1 min and a delayed onset of edema formation resulting in a mean weight gain of 20 g after 120 min. This was paralleled by a significant increase in prostacyclin generation and a continuous release of K + and LDH. Thromboxane synthesis exceeded about 10 times that of controls in the toxin treated lungs. Pretreatment with WEB 2086 or diclofenac significantly attenuated the pressure response and edema formation evoked by cytotoxin. The addition of the unspecific immunognbulin preparation alone induced a transient pressure increase within the first minutes, but mean values remained below those of the cytotoxin group in the continuing observation period. Mmost complete inhibition of the pressure reaction, the edema formation and the metabolic alterations was achieved mainly by the combination of immunoglobulin, WEB 2086 and diclofenac and to lesser extend by the specific toxin antibody. Conclusion: The current results point towards the crucial role of PAF and AA-metabolites as mediators of cytotoxin induced microvascular injury. The systemic or local application of cytotoxin antibodies or even unspecific immunoglobolins in combination with PAF-antagonist and diclofenac appears to be a promising therapeutic approach in the case of infection with cytotoxin-preducing strains. Cytokines have long been shown to be of particular importance in the metabolic derangements occurring in LPS-induced shock. Recent studies strongly imply the involvement of Platelet Aggregating Factor (PAF) in the pathogenesis of Gram-negative bacterial sepsis. An autocatalytic feedback network has been postulated to exist between PAF and Tumor Necrosis Factor (TNF), a key cytokine involved in septic metabolic cascade, leading to an uncontrolled amplification of inflammatory mediator release. We have previously shown that ST 899 (4-N,N,N trimethylammonium-(R)-3-isovaleroyloxy-butanoic acid Z-3-(5-chlorphtalidiliden) ethyl ester bromide) was quite effective in inhibiting the "in vitro" binding of 3H-PAF (Ki=7.5x10 -8 M) to rabbit platelets. The present study shows that pretreatment of C57BL/6 mice with ST 899, administered by different routes, dose-dependently and significantly reduces the lethality induced by endotoxin (E.coli 055:B5 injected at 30 mg/kg intraperitoneally). Very interestingly, ST 899 administered at the same doses as above (i.e. 1.25, 2.5, and 5 mg/kg body weight) results to be significantly effective in reducing the endotoxin-induced release of serum TNF. The reported dual activity of ST 899 (i.e. PAF antagonism and decreased circulating TNF levels) may turn out to be greatly beneficial, in combination with current therapies, in the treatment of diseases that involve overproduction of TNF and PAF such as septic shock. Introduction: Recently, we reported that prophylactic whole body hyperthermia (42.5°C) induces heat shock protein ('ASP) 72 and increases smvival 2-4 fold in a mouse endotoxin model (Am. J. Physiol. in press). Other investigators reported that prophylactic pharmacologic induction of HSP-72 by sodium arsenite improves survival in a rat sepsis model (abstract A95 Am. Rev. Resp. Dis. Vol. 147, 1993) . The effects of heat are complex and in addition to formation of liSP-72 include release of cytokines, changes in cellular pH etc. Thus, the protective mechanisms of heat may differ from those due to pharmacologically induced . The purpose of this study was to compare the protection of heat vs the protection of pharmacologically induced HSP-72 in a mouse endotoxin model to determine if different protective mechanisms were likely to be involved.. i%'lethods: Both sodium arsenite (10 mg/kg) and ethanol (400 ~1 of 39% ethanol) caused marked induction of HSP-72 in lung, gut, kidney, and liver, which was comparable to heat-induced HSP-72. Female ND4 mice weighing 22-25 gms were pretreated with arsenite or alcohol 8 hours prior to challenge with Escherichia coli endotoxin (-LD 80) and survival was compared to control mice.

Results: Survival at 24 hrs. for arsenite treated and alcohol treated mice was 96% and 88% respectively and was statistically different from the 40% survival for control mice. (p<0.01) (n=25 mice per group). However, at 7 days post endotoxin, there were no differences in survival in the 3 groups, i.e., ~ 4% survival for all 3 groups. In contrast, the protective effect of hyperthermia remains present at 7 days, i.e., ~ 70% survival vs 20% survival control. Conclusion: The protective effect of heat is probably due to other factors such as the effect of hyperthermia to release IL-lc~ and is not due solely to HSP-72 formation. It was the aim of the study to examine whether bacteria play a causative role in the pathogenesis of anastomotic insufficiency following gastrectomy in man.The study was carried out in form of a prospective, randemised, double-blind, multicenter trial. Primary endpoints were the rate of anastomotic insufficiencies, infectious-and uncomplicated postoperative courses. All pat. received a periop, i.v. prophylaxis with Cefotaxim. Identical numbered vial either contained Placebo or Polymyxin B, Tobramycin, Vancomycin and Amphotericin B . The vials were administered 4X per day from the day be ~ fore the operation until the 7th postop, day. Insufficiencies were detected by gastrographin swallow and recorded by X-ray on day 7 postop.. Evaluation was carried out on an "intention to treat'basis. Statistical analysis was done with the Pearson's Chi Square and Fisher's Exact Tests~ Results: Interim analysis was carried out in 3/93 after 137 pat. had been recruited. Along with a significant reduction of S.aureus and enterobacteria there was a reduction in the rate of anastomotic insufficiency of the esophago-jejunostomy from 10.6 % in the Placebo-group to 3.5 % in the treatment group. The difference was not yet significant. The rate of nosocomial infections (e.g. respiratory tract infection and UTI) were significantly reduced from 33.3 % in the Placebo-group to 16.4 % in the Treatment-group (p ~ 0.0281;Fisher's Exact Test). In march 1994 final results with more than 200 patients will be presented for the first time. (= pO2 <80 mm Hg, b s-creatinin >2 mg%). Respiratory insufficiency was the most frequent systemic complication followed by sepsis and respiratory insufficiency. Etiology of pancreatitis and initial serum increase of pancreatic enzymes predicted neither complications nor outcome. Only 2 of 14 deaths occurred during the 1st week, all other deaths occurred late (after 4-12 weeks), generally as the consequence of septic complications and multi-organ failure. High levels of CRP were correlated with a compliacted course and a fatal outcome. Although same cytokines (e.g. 11--6) were found increased in severe disease, the predictive value of these markers was not better than the combination of ctinical scores (Ranson, Imrie, APACHE II) with GT or CRP. CONCLUSIONS: Intensive care medicine can often control the inital shock situation in severe pancreatitis. Thus. only 15% of deaths today occur eady in the course of the disease, whereas this percentage varied between 40-70% just 10 years ago. Nowadays, most deaths are caused by late septic complications and multi-organ failure. Ranson-and CT-scores as well as serum CRP predict a course with systemic complications; they are less helpful for prediction of sepsis and late mortality. It is doubtful whether measurements of cytokines will help to better predict the late outcome. As yet, only careful and continuous monitoring of patients (e.g. by APACHE scores) may help to early identify those who develop septic complications and multi-organ failure.

The classic description of severe acute pancreatitis has hinged upon the release of large volumes of activated enzymes into the peritoneal cavity and thertce the lymphatics and blood stream. These activated enzymes escape from the pancreas due to disruption of cells with associated ischaemia and occasional infarction of tissue. For 20 to 25 years it has been postulated that the bocly's defence system to activated pancreatic enzymes required supplementation iu the form of anti-protease support either in the vascular space or in the peritoneal cavity. All controlled studies have shown that this is either impracftcal or unnecessary.Hore recently release of a large number of cytokines from monocytes, macrophages and neutrophils have been considered to be harmful to the body and various agent~ which oppose the action of TNF alpha, PAF and similar cytokines are being examined in experimental anim~is and certain clinical trials, It has clearly been shown that higher levels of cytokines are released in the patients with objectively graded severe acute pancreatitis than in those with milder disease. We now seem to be moving into an exciting phase of potentially beneficial therapy in acute pancreatitis which has had no specific effective therapy through studies utilising aprotinin, gabexate mesilate and fresh frozen plasma. Inflammation cascades may play a role in the pathogenesis of acute pancreatitis. To evaluate the status of the cellular immune system we examined serum concentrations of immune activation markers in 24 patients with acute pancreatitis (14 males, 10 females; median age: 56 years, range: 32-81 years). Concentrations of neopterin, serum soluble tumor necrosis factor receptor (sTNF-R) and serum soluble intercellular adhesion molecule type 1 (slCAM-1) were determined using immunoassays (Henning, Bender, T Cell Sciences). 13/14 had increased concentrations of sTNF-R compared to the 75th percentile obtained in healthy controls (>4.2 ng/ml), and 9/24 patients had increased neopterin (> 8.7 nmol/I), 9/24 presented with elevated slCAM-1 (>440 U/I). All patients with increased neopterin also had increased sTNF-R, 4 patients had concentrations of all three markers outside the normal range. There existed a significant correlation between neopterin and sTNF-R (rs = 0.728, p < 0.001 ). Weak associations between age and sTNF-R (rs=0.435, p=O.05) or neopterin (rs=0.456, p = 0.044) were also found. Our results demonstrate activation of the cell-mediated immune system taking place in a sub-group of patients with acute pancreatitis. The finding of increased neopterin and sTNF-R levels implies that activated monocytes/macrophages are involved in the pathogenesis of the disease. Further data are necessary to evaluate potential associations between changes of marker concent-rations and the course of the disease. Pancreatic injury after heart surgery was reported as soon as 1970 ( 1,2) and characterized by increased serum or urine amylase levels (in about 33% of patients) in the 5 fi~t postoperafi.'ve days. This pancreatic injury, which sometimes led to acute pancreatitis, was atreaay at~buted to inappropriate perfusion of this organ. In the 198ffs, 3 studies were published dealing with pancreatic suffering alter heart surgery, in large series of patients, concluding ~n~at panc~a~c injury (with a low incidence of pancreatifis) is more common than previously recognized and is a potential source of complication after camliac surgery (3, 4, 5) . In a recent study (6), evidence of pancreatic cellular injury was found in 80 out of 300 patients undergoing cardiac surgery, with 22 out of these 80 patients presenting abdominal signs or symptoms and 3 developing severe pancreafitis. This injury was associated w~th preoperative renal insufficiency, valve surgery, ~..stoperalive hytxXension, calcium administered periopuratively and length of bypass. We studied 300 patients submitted to cardiopulmunary bypass (CPB) for heart surgery and used the measurement of uN:~sin, pancreatic iso-amylase and lipase in plasma for biochemical characterization of pancreatic cellular injury. Blood samples were obtained before surgery, directly aller surgery (return to inte~ve care unit), 8 hours alter surgery and in the folfowing days alter surgery (days 1, 2, 3, 4 and 8). Computed tomography scan of pancreas was performed in patients presenting hi~ levels of amylase on day 1. We measured abnormal levels of trypsin and pancTeatic iso-amylase in 30% of patients and observed 2 simultaneous releases of these 2 enzymes, the fi,'st one in the 24 hours after surgery and the second more intense from day 4 and pa~icularly on day 8 after smgery. This second release was concomitant with abnormal levels of llpase. These biochemical observations were accompanied by radiological and clinical signs of pancreatic injury in about 2% of our patients : pancrealic abnormalities were revealed by scan in 7 patients and acute pancreatitis in I patient. More pronounced pancreatic suffering was observed in patients undergoing valve replacement than in patients undergoing coronam-anrtic bypass grafm~g. Analysis of trypsin and pare're, tic 1so-amylase are sw.cific of pancreatic cellular injury and their simultaneous ir~rease in plasma alter CPB in our padents confirms the presence of an exocrine pancreatic injury. The presence of a simultaneous peak of lipase mcaezse~ the specificity of overt pancreatic injtu 7 diagnosis. The precise cause of th/s injury could he related to hypoperfnsion leading to ischemic injury of foe splancbnic area, pancreas being largely sensible to hypoperfnsion (7). This hypoperfosion could he responsible for the ftmt release of pancrealac enzymes observed in our patients and would contribute to the deterioration of other organs leading to an inflammatory reaction developing in the following days and responsible for the second release of pancreatic enzymes observed in our patients. Patients with necrotizing pancreatitis show a heigh rate of pulmonary, renal and septic complications, whereas the course in acute interstitial pancreatitis is generally very mild. We have prospectively analysed the value of Endotoxin, Interleukin-6(IL-6) and Transferrin in compare with c-reactive protein(CRP) for the early assessment of the severity of acute pancreatitis. Patients aud methods: The values of Endotoxin(measured by limulus-lysate-test), II-6(ELISA), transferrin and CRP (nephelometry) were analysed daily along the first i0 days of hospitalisation by 28 patients with acute pancreatitis admitted to our Hospital from 4/1991 to 4/1993 . It was judged whether the patients have either interstitial (AIP) (n=9) or necrotizing (ANP) (n=lg) pancreatitis. 5 patients with ANP have died during the course of pancreatitis (Mortality=17.5%). Results: 1-Severity o~ Pancreatitis: Signifcant differences (P<O.05, Mann-Whitney test) could be calculated between AIP and ANP for endotoxin, IL-6, transferrin and CRP during the I0 days of monitoring. 2-Mortality: Except of IL-6 on days 2, 3, 4 and 5 after ad-mission there were no significant differences between the values of analysed parameters in patients who survived and those values in patients who died. 3-Organ failure: The patients with pulmonary, renal or multiorgan failure have significantly heigher values of endotoxin, IL-6 and CRP along the I0 days of monitoring in compare with those patients without an organ failure. Conclusions: Endotoxin, IL-6 and Transferrin, an important endotoxin carrier, are promising parameters to differentiate between the severe and mild form of pancreatitis comparable with CRP. This parameters may be helpfull in the early clinical assessment of patients with acute panereatitis. The determination of cytokines may elucidate the pathophysiologic mechanisms that produce early systemic complications in acute interstitial (i) or necrotizing (n) pancreatitis (AP). The appearence of respiratory insufficiency (RI) is used to appraise the early systemic complications and is compared to the results of cytokine blood levels.

In a prospective clinical trial from 10/92 -8/93, 23 patients with AP were recruited and blood samples taken for cytokine determination with commercial ELISA-kits. The differentiation between i (n=12) and n (n=l 1) AP was made through Ct-scan in all and laparotomy for drainage and necrosectomy in 13 patients. The etiology of AP was gallstones in 9, alcohol abuse in 7, hyperlipidemia in 2 and other or unknown causes in 5 patients.Five of 11 patients with nAP died early (n=l) or of late septic complications. None died of iAP.The peak of cytnkine level in the first three days of hospitalisation was used for calculation. RI was diagnosed in the first week of hospitalisation, if oxygenmask or iutubation for at least 3 days was nescessary to treat arterial hypoxemia. For statistical analysis U-Test of Wilcoxon, Mann and Whitney was applicated.The IL-6 concentration in blood reached up to 2600pg/ml in the first days depending on the severity of AP and dropped to almost zero in the next days, independently of the clinical course. The differentiation of i-versus nAP, using a cutoff-line of 600 pg/ml was correct in 20 patients (87%, sensitivity (SE): 82%= specificity (SP): 91%, (z:0,001). High levels of IL-6 over 500 pg/ml correlated well with the prognosis ifRI in the first week (Accuracy: 87%, SE: 80%, SP: 100%, c~: 0,001). The blood levels of IL-8 reached a maximum of 1381 pg/ml in the first days, depending on the severity of AP, and showed a correlation to the clinical course in the following days. The peak of 1l,-8 blood levels indicated correctly the severity of AP in 18 out of 23 patients using a cut offtevel of 200 pglml (Accuracy: 78%, SE: 82%, SP: 75%, ~: 0,01). There was no correlation between cytokine blood levels and mortality, Also there seemed to be no correlation between the levels of IL-6 and IL-8. In the bloodsamples of five patients with i-or nAP no c~-TNF was detectable.The blood levels of IL-6, and to a lesser extent of [L-8, can predict the severity and early systemic complications of AP in men. The excessive rise of cytokines can be explained by the stimulation of immunological cells ( macrophages, lymphocytes and endothelial cells) in the course of AP. Objectives: In an opossum model, ligation of the sphincter of Oddi or separate ligation of the bile and the pancreatic duct together leads to severe haemorrhagic pancreatffis while single ligation of the pancreatic duct causes only an exocdne atrophy. Since BO has been discussed to be a contdbutor t¢, RES dysfunction, this may depdve the organism of a potent defensive mechanism thus promoting the course of pancreatitis. The present study wa,,; carded out tc develop a new method for measuring dynamic liver RES func.. tion and to test the hypothesis that BO causes a RES dysfunction. Material & Methods: 18 opossums were randomly placed in three groups. All received a central venous line. After midline laparotomy, a specially designe0 tourniquet was placed around the common hepatic duct above its junction with the pancreatic duct (group A, n=6), around the pancreatic duct (group B, n=6) or around both ducts (group C, n=6). After one week, the tourniquets were closed. Using a scintillation camera, the liver uptake of Nanocoll, a 99mTcalbumin colloid with a diameter of 25 nm, was measured before, 2,5 and 6 days after the obstruction. The curves were analysed by a computer and two parameters (R, S) were calculated using natural logarithmic regression. As a common measure for RES function, the corrected granulopexic index (a) wa,,; determined, After day 6, the pancreas of each opossum was searched for necrosis by morphometdc methods. Results: All animals survived till day 6. The opossums in groups A & 8 had a macro-and microscopic normal pancreas, while those in group C showed a severe hemorrhagic pancreatitis. The granuinpexic index showed a significan~ change to the worse starting at day 5 in group A & B (p<O,05) and at day 2 it~ group C (p<0,Ol). At day 6, RES function in group C was significantly worse than that in group A & B (p<0,01). The regression parameter R showed no significant change in groups A & B, but a highly significant decrease in group C starting at day 2 (p<0.005), thus showing a dysfunction of the hepatic RES. Conclusions: Obstruction of the hepatic or pancreatic duct alone does no{ result in immediate dysfunction of the hepatic RES, while obstruction of beth ducts does. Because only ligation of both ducts is followed by a severe hereof.. rhagic pancreatitis, RES dysfunction could be an important factor in the pathogenesis of pancreatitis and resulting organ failure.Logarithmic regressinrL has proven to be a valuable tool to analyse dynamic hepatic RES function. (B6) in Lewis rats weighing 220-250g. There were five groups: group A (n=14) were untreated controls; group B (n=7) were given isotonic saline intraperitoneally and observed for 48 hours; group C (n=7) 2 ml x 108 E. coli intraperitoneally and observed for 48 hours; group D (n=7) were given 4 ml x 108 E. coli and observed for 48 hours; group E (n=7) 4 ml x i08 E. coli and observed for 24 hours, The rats were than killed, the spleens were removed and heparinised blood taken for immunological tests. Total T-ceils, helper T-ceils, suppressor T-cells, monocytes, and the interleukin 2 receptor were determined by monoclonal antibodies, indirect immunfluorescence, and flow cytometry (FACS analyser 11). Statistical analysis was by the t test on rank transformed data.In group B (isotonic saline) the total T-ceils increased in the blood (p=0,003) and spleen (p=0,0001 )com pared with the untreated rats (group A), In the groups with peritonitis (C, D, and E) we found highly significant rises in suppressor T-cells in the blood and spleen (both p<0,0001 ) compared with the non-infected control groups A and B. This increase was significantly higher in group D than in groups C and E. As well as other alterations, we found a uniform increase in the number of T suppressor ceils in our model of acute peritonitis that was both time and dose dependent, Department of Surgery, University Vienna, AKH Wien, W~hringer Gurtel 18-20, 1090 Wien.

D~'na~'aics of Som~ l~mnunologic Indices in Children with Abdo;~tinal Shock V.Z.i~ioskalenko, S.F.Vesiol$,T.V.P~bina Serum (IgA,:~I,G, circulating imaune co!~plexes, co:apleaentary serum activity, antimesothelialvisceral and parietal antibodies) and cellular (i-l~q~hocytes, ~s/Th,B-lyaphoeytes,i nLaunoleukosis to aesothelial allergens; spontaneous blast cell transformation to phytohemaaglutinin and .nesoth<-~lial ~,togen; phagooFtic neutrophil activity) im:~une indices v.'ere studied in 60 children operated v:ith symi~to'as of abdo~ainal shock. ,'~!-~ of them had appendicnlar(26-1ocalized JO-diffuse, 9-~eneralized) and 5-hemoperitoni-~is (dull abdominal trau;na with injayy of 'the spleen-4, the spleen and liver-q), flO patients tJere operated on for co~Iplete iieus (~-intussnsc.sption, 5-com:lissural ileus). 2he follo~'~ing serum factors: cireula:;ing im~aune complexes and ~).esothelial antibodies are the aost i~portant ~ar~ers of the abdominal shock course. :Yith the progress of the process the indices increase irresponsive of t.L~e cha±.aeter of pathology. In case of ileus regress and hemoperitonotis an abrupt drop in ~he titer of anti~aesothelial antibodies is noted. In bacterial peritonitis a high ~iter of anti:aesothelial parietal antibodies in great dilutions ("+++","++++"-q ; 720) persisted even in the period of clinical recovery. Cellular factors cham'acterize dFnaaics of the abdominal shock coulees less specifically Llast cell transfor~aation and phagocyue neut~ophil activit 2 can inderectly characterize transition of shock f~orn reversible into irreversible phases. The past SO years have brought a number of major advances in burn care. The initial advance was the discovery that burn victims required large volumes of resuscitation fluid in the initial 24 hours to maintain hsmodynamic stability, This was followed hy the dsvelopmsnt of topical antimicrohlal agents to prevent end treat infections. Next, was the reporting that early excision of burn wounds, with autografting of the excised wounds was possible.Finally, was the identification of the importance of aggrsssive nutritional support for the hypermstabolic burn patient.These advances have markedly increaesd survival rates for given burn mimes, s~ch that burns which would previously have been considered to be aniformly fetal, are now readily survivable.Thsre remain two unsolved problems in burn care, which are limiting survivability cf burn patients.The first ie ths treatment of inhalation injury. Ths sscond is how to obtain Coverage of ths maaaivsly burned patient, who has limited skin graft donor sites available.This ssssion will focus on the new treatments being developed to obtain permanent coverage of burn wounds.These include the use of various tcplcal growth £aotore which can Speed the rate of reepithelialization of wounds, Both the beneficial and dstrimental effects of using cultured karatinooyte preparations, to obtain in excess of one squats meter of a pseudoskin from a single fifteen square centimeter biopsy of unburned skin, will be discussed, Finally, the used for artificlal dermal preparations will be discussed. integumentary wound healing consists of simultaneous epithelializalion and contraction. Not long ago, it was thought that healing proceeds at an optimum fixed rate and could only be delayed. The recent explosion in molecular biology and recombinant gene techniques have produced information on the biological activity of compounds previously believed to be biologically inert. Many current attempts to modulate the rate of wound healing center on the topical application of various growth factors produced by integumentary cells or agents designed to modulate growth factor expression or response. Epidermal growth factor (EGF), plateiet derived growth factor (PDGF), transforming growth factor (TGF) and fibroblast growth factor (FGF) have been demonstrated, both in vitro and in vivo, to stimulate the proliferation and diferentiation of their designated cells types. EGF and PDGF have been clearly demonstraled to increase the rate of wound closure in partial and full thickness wounds. PDGF also has demonstrated efficacy in the closure of recalcitrant pressure sores, whereas TGF increases the tensile strength of incisiona[ wounds and FGF stimulates angiogenesis. It appears that epidermal keratinocyte migration during wound healing depends on integrin-mediated interactions with compounds and ceils extracellular matrix. Additionally, the activities of extracellular glycoproteins such as fibronectin, fibrinogen, laminin and thrombospondin are coordinated by multiple integrins with specific distributions and binding affinities. Currently, agents which incorporate specific protein sequences essential for the interaction of the glycoproteins with the cells of the extracellular matrix are being investigated. It is has been clearly established that the rate and nature of wound healing can be influenced by the application of various agents and that sufficient quantities of these agents can be manufactured. The future challenge is to develop techniques to Now the objective evaluation of the clinical efficacy of these various agents and to employ the appropriate agents in a cost-effective manner. 

transplantation; and 4) the immune status/manipulation of the host.Multiple interactive variables will determine the effect of employing allogeneic cells and tissue in wound coverage design and a batter understanding of the dynamics of the wound-graft microenvironment will facilitate the dove opment of clinicaUy beneficial graft composites.

Cadaver allograft skin (CAS) is the "gold standard" for wound coverage, but has limitations including varied availability & quality, immunogenicity leading to rejection, & potential for disease transmission. Our development of a temporary living skin replacement (TLSR) addresses these issues; the TLSR consists of highly screened human neonatal fibroblasts (HF) cultured in Biobrane a, a bilayer dressing consisting of nylon mesh laminated to a thin silicone membrane which controls water-vapor transmission. Studies in our lab indicated that "fresW ~ TLSR grafts reproducibly close full-thickness wounds in mice & perform better than BB controls. We studied wound adherence & structural/molecular properties of fresh & cryopreserved (CRYO) TLSR. METHODS: TLSRs were prepared by seeding 4000/co 2 HF in BB nylon mesh in DMEM. HF quickly attached to the nylon mesh & were allowed to proliferate 3-6 wks. Fibroblast mitochondrial activity was assayed by MTT assay. Matrix proteins were identified by immunostaining, mRNAs for specific growth factors were identified by reverse-transcription polymerase chain reaction amplification, & quantitated by gel scans. BB structure was studied by scanning electron microscopy & stretching measurements pro/post CRYO. 80 grafts were CRYO'd in 10% DMSO, quick-thawed & sutured onto 2X2 cm excised full-thickness dorsal wounds on athymic mice. 40 "fresh" grafts were similarly placed. Controls grafts were CAS and aeellular BB. Adherence was quantitated by a device which peeled grafts from wounds at intervals following placement, using a serve motor and a strain gauge. RESULTS: PostCRYO, storage and subsequent thawing tile TLSRs maintained >60% cell viability by the MTT assay, indicating continued mitochondrial activity, and BB structure & stretchability were maintained. Multiple matrix proteins secreted and deposited in the BB nylon mesh (types l/III collagen, decorin, fibroneetin) were identified by specific immunostaining. Growth factor mRNAs in the TLSRs (aFGF, bFGF, KGF, TGF~,p~,) were present in 10-10,000X higher levels in fresh/CRYO TLSRs than in adult HCS. Grafts adhered to wounds on mice through 40 days of followup. Histologic exams on days 6-40 showed excellent vascular ingrowth and minimal inflammation. Adherence of TLSRs to wounds was >CAS adherence. Burn wound coverage in the massively burned patient remains a difficult problem. Although cultured keratinocytes have been utilized for burn wound coverage, their impact on the patient with burns greater than 80% total body surface area has not been spectacular, with poor graft take and unstable epithelium.Current investigations have been directed toward dermal replacement beneath either very thin split-thickness autografts (STAG) or utilizing cultured keratinocytes. Current products include:

Collagen dermal replacement with thin STAG (Burke, et al).

Collagen dermal replacement with cultured keratinocytes and fibroblasts (Boyce, et aI).

Allograft dermis with cultured keratinocytes (Cnno, et al).

Allograft dermis with thin STAG (Life Cell).

Polyglactin acid mesh and neonatal human fibroblasts with thin STAG (Hansbrnngh, et al).Investigations regarding culture media, use of growth factors, topical nutrients and antibiotics, and melanocytes for pigmentation as well as safety and efficacy are needed before any of the current products become viable options for coverage of the massively burned patient. The~ is a growing world-wide problem with the uJc Of cadaver tissues and ocgans bae, au~ of the tren~m~s~km of dilemma such a; Cmutzfeldt.Jukob disease and IIIV as we1] as ready availability of urdform liS~ue~. On Dec~mt~r 14, 1993, the FDA assumed control of aS1 tissue bar~s in the Uldtod St=tea In an attempt to bflng ~s difficult problem of dise~s~ transmission under ¢onlrol. In Europe, ~om¢ of the governments are consldofll~ a c~mplcte bat) on the use of cadaverlc fissu~S such as ddn, 'This |ncroam in regulation of cadavefle ~s,quct will incmar¢ the difficulty of obtain~g and dlslflbuLmg them. However, thc nc~ for these tissues contlnue~ m incrcaso, We will discuss ~'L¢ solulion to this Important pmbl~n: Tissue Engineering. Tlssu~ Engineering is an in~rdisdpllnary field that applies pdnclplc~ of angin~edng and die life sclcnce~ reward the development of ~olok~¢al sub~dtute,~ Ih= mslom, maintain, or improve tissue function, "13ssuc ongln~cdng can provide ~ho nccassary tlssuoa for wound repair ~d Ibe assuranoe fl'~t the lissuos are d.ls¢~¢ free. In addition, a ds~uo-cng~ne~n~l wound covering will bo u~lvemally acceptable and evntlublc as "off G~o shell", consis~t products, Them are several approaches to restating thls function in a large wound, 'l'nosc i~elud~ tmmcdiete long term coverage, short t=nn coverage, uandtl~el coverage and compost= dssu¢ coverage, "flssuo onglncrcd wound coverings that meet those vaflous ne,.cds will he r~vlowod.Cllni~:sl and experimental d~la in venous ulcer, dlabctl¢ ulcers, prossur~ ulcers and bum wounds wgJ be mvlcw~, a~ welt as new approacl~s U~ csrtilag¢, bone, liver and bone marrow It~Suos.

C Oomplon, K Nadirs, W Press, G Wetland, J Fallen IV, Shrtners Burns Institute and Massachusetts General Hospital, Boston, Ma~schusetts, USA The clinical "take" rate o? cultured epithelial autografts (CEA) has been observed to Increase with transplantation to allodermls, but the reasons for the improved clinical performance have not yet been defined. The aim of this study was to determine the biological impact of normal human dermis on CEA differentiation and maturation, Biopsies of CEA transplanted to engrafted and de-opldermlzed human homograft dermis have been compared to Nopsles of CEA transplanted to granulation tissue In tullthickness burn wound beds on the same patient, each patient serving as hls or her own control. Paired test and control biopstes from six patients have acquired from as early as one week postgrafting to as late as 3 years postgrafting (one patient) and analyzed histopathologlcally, ultrastructurally and immunoh[stochemloally, Results demonstrate more rapid normalization of differentiation markers (e,g., Involucfln, fllaggrln, cytokeratln profiles) in the CEA transplanted to allodermls compared to their corresponding controls by in all patients, The proliferation rate within the basal layer ot the epidermis as determined by Ki-67 (proliferation-associated antigen) Is seen to norh~altze more quickly In the CEA transplanted to allodermls in every case, Persistence of allodermal matrix can be doOumented In all patients by elastic tlssue-trichrome stain, allowing visualization of the dermal elastin network. The popu;atlon densities ot Intraepldarmal Langerhans cells are conslstently and signlflcantly higher In CEA transplanted to ,allodermls, possibly reflectlng an Immunologlcal reaction to the underlying allogenlc tissue. Overall, these preliminary results indicate that transplantation to a normal human dermal matrix accelerates the maturation of CEA-deflved epidermis,

Wound closure continues to be a major problem in patients who have sustained a major thermal InJury, Cultured epidermal autografts (CEA) have been utilized extensively since 1984 when Galllco et el reported theh'use In two brothers with greater than 90% total body surface area burn. Unfortunately, CEA take rate varies widely and the resultant skin coverage Is often fragile and the cosmetic results are less than optimal However the overall take rate and durability of the coveraSe can be markedly Improved by using nn allodermls base as the recipient bed. A review of CEA applications performed by physicians using cultured outologens epithelium obtained from Blusurfaoe Teclmology, Inc. shows a marked discrepancy in the results obtained utilizing different methods of wound bed preparation. TGF-B is an important modulator coordinating complex physiological events associated with growth and development. It is assumed that TGF-B is also involved in the well-coordinated process of cutaneous wound healing by regulating proliferation, differentiation, chemotaxis and matrix deposition. The purpose of our study was to analyze the spatial and temporal pattern of TGF-B expression during granulation tissue formation in patients with accidanutl surgical trauma (monotraumata mid polytraumata) and bum wounds. After debridement (day 0), the full thickness wounds were covered with Epigard, a synthetic dressing until day 10. After this time the granulated wounds were closed by transplantation of mesh graft. Biopsies of the wound center were taken from 30 patients at the beginning of surgical treatment (day 0) and after 3, 6, and 10 days. Cryosections were stained with antibodies against TGF-fi s using the APAAP technique and -for standard histology -with Hematoxylin-Eosin. For identification of the cell type expressing TGF-6, double staining immunofluorescence experiments were conducted using antibodies specific for monocytes/macrophages, polymorphoanclear neutropkils and fibroblasts. The results showed a characteristic pattern of TGF-t~ distribution during wound development. TGF-fi appearence was mainly cell-associated znd the absolute and relative number of cells that were positive increased with lime. Infiltrating cells and developing blood vessels were most prominently stained; epithelial and T-cells showed no immuno-reactivity. A delay of emergence for TGF-B during the time course could be seen in one patient group. This might reflect various regulation patterns depending on the type and severity of injury.(1) PharmaTec GmbH, Frankfurt (2) Institut fiir Immonologie and Serologic, Heidelberg ( Immune cells extravasating specifically in skin recognize and eliminate the invading antigens (bacteria, viruses, etc.) either in situ or transport them to regional lymph nodes. They also participate in the process of skin wound healing. Cells which traffic through the skin can be harvested from efferent lymph drained from a given area of skin. The type of migrating cells changes after trauma, heating and infection. We have developed a method for collection of human afferent lymph in lower limbs. The method allows obtaining immune cells from normal and injured skin and their characterization. AIM OF THE STUDY was to characterize skin immune cells in situ and in skin lymph with use of immunohistological methods (staining, FACS). RESULTS. Group 1, cells migrating through skin: 75+4% T lymphocytes (CD3), 6+3% Langerhans and dendritic cells (CDla, HLA DR, S100), 41+9% CD4, 18+6% CD8, no B cells (CD22, 23), 74% CD45R0 (memory cells), 6+3% IL2R. Approximately 90% cells possessed CDlla and 18 antigens. CD1 lc was expressed only on large cells. The frequency of all phenotypes was different from the blood populations. Group 2, cells in skin: Langerhans cells were found only in epidermis, CD3,4 and 8, CD45R0, RB, ila/18 cells around venules, CD68 (macrophages) uniformly dispersed, no IL2R and B cells. HLA DR positive were endothelial and some dispersed mononuclear cells. Group 3, one, three and thirty days after surgical wound (simple varicous vein extirpation): high density of epidermal Langerhans cells, HLA DR positive keratinocytes and all endothelial ceils, few IL2R cells, perivenular infiltrates of CD68, 45R0 but less CD3 cells, high density of CDlla/18 cells. Classic staining of isolated and in situ located ccl!s with MGG or HE did not allow to follow kinetics of changes. CONCLUSIONS. This study presents the first in the literature quantitative data of immune cell traffic through normal and injured human skin. IN THE  CONTROLLED  RELEASE  OF  BIOLOGICAL  RESPONSE  MODIFIERS FOR SOFT TISSUE REGENERATION. Alan S. Rudolph, Helmut Speilberg, Mariam Monshipouri, and Florence Rollwagen, and Barry J. Spargo.

We have employed lipid microstructures as controlled release vehicles for the delivery of growth factors in wound repair. Traditional liposomes as well as novel lipid based microcylinders have been examined for their in vitro kinetics of the release of transforming growth factor beta (TGF-b). In vitro reiease has been examined by setting up models with examine the physical release of iodinated TGF-b as well as a cell based bioassay (based on the HT3 bioassay). The hollow lipid microcylinders (50 microns in length and I micron in diameter) show an initial burst (5-10 ng) followed be zero order kinetics which result in the release of approximately I ng TGF/day. This release behavior can be modified by temperature based on the phase behavior of the lipid bilayer which comprises the microcylinder.We have also examined the cellular response to lipid microcylinders applied in vivo. The lipid microcylinders are mixed in agarose and implanted as a composite hydrogel block under the flank of a mouse. The blocks are removed 1, 3, and 5 days following implant and the cells analyzed by FACs sorter analysis. The observed pattern of ceil recruitment to the blocks mimics that seen in a local inflammatory response. Cell surface phenotype studies included the determination of CD3 and CD4, Mac-l, and Ig bearing cells. We have also begun to examine the change in cell surface phenotype and kinetics of recruitment following the inclusion of TGF-beta in the lipid microcylinders.Center for Biomolecular Science and Engineering, Code 6900, Naval Research Laboratory, Washington, DC. 20375-5348.

Expression pattern of heat shock proteins in acute, good healing and chronic human wound tissue.

Abstract: Wound healing is a complex biologic process that is well characterized at the histological level, but its molecular regulation is poorly understood. After clot formation, inflammatory cells are rapidly drawn into the wound, followed by migration of fibroblasts and epithelial cells that divide and repopulate the wound area. During the last decade peptide growth factors and cytokine are thought to play a key role in initiating and sustaining the phase of tissue repair. These factors which are released from different cells appear to initiate the cascade of events that lead to healing. Different studys described the rapid activation of a family of proteins,named heat shock proteins (HSP) in differnt tissue that were exposed to various forms of stress (heat, toxic agents, mechanical). In this context HSP's have the ability to regulate protein folding and assembly, to transport proteins across cytoplasm and membranes, to disrupt protein complexes, to stabilize, degrade and regulate the synthesis of proteins and to take part in DNA replication and repair. We now attempted to find out if HSP-gene activation is also involved in injury and wound healing, which likewise resemble a stress situation for cells. Therefore we collected tissue samples during operation and single biopsies from chronic wounds (Decubitus for example) and granulation tissue. After RNA preparation from these samples we used RNA-PCR and Nothern analysis to study the expression of Objectives of the study Chronic, non-healing cutaneous tflcers are a challenging clinical and socioeconomic problem. Several animal studies have shown that cytukines (e.g. EGF, PDGF, FGF, TGFB) accelerate the healing process and tissue repair in general. Results from first clinical trials indicate a promising value of cytokines in the treatment of chronic non-healing diabetic and venous ulcers. Recent reports in the literature indicate that the biological activity of the solution of platlet derived wound healing formula (PDWt~) released from c~-granules (mainly PDGF & TGFI~) is greater than the activity of the recombiant single factors like e.g. PDGF-BB (Robson, Lancet 1992) . The aim of our study was to determine whether a correlation exits between the concentration of TGFI~ & PDGF and the time course of wound healing. Materials and methods PDWHF was prepared from 200ml of auto]ogous patient blood and diluted with a special buffer to a final concentration of 30 ng/ml g-thromboglobulin. The concentrations of PDGF and TGFg were determined by ELISA-tests developed in our laboratory. 22 patients with chronic non-healing ulcers have been evaluated alter treatment by topical application of PDWHF. PDFG and TGFf~ concentrations of the topical solution were measured and two patient groups formed for analysis The time course of wound healing was regularly and meticulously documented and evaluated by photography and casting. The time from initiation of treatment instil o wound volume reduction to 50 go of the origional size (T50%) was noted• Results: Healing of extensive burn wounds can be accelerated by grafting cultured autologous or allogeneic keratinocytes. The stimulation of granulation tissue formation and reepithelialization is presumably based on growth factors and cytokines released by keratinocytes. We wanted to prove this hypothesis by investigating the bFGF expression during wound development, bFGF is mainly described as an angiogenic protein with mitogenic activity on various mesodermal and ectodermal cell types pointing to its stimulating potential in wound heating. In the present study we compared the pattern of human bFGF m-RNA expression and the localization of bFGF protein during the first days of wound healing. Biopsies were taken from 5 juvenile human bum patients, immediately after wound debridemerit mad on day 8 after transplantation of cultured allografts. Biopsies were snap frozen and cryosected. The pattern of bFGF expression was assessed by in situ hybridization of the bFGF m-RNA with a digoxigenin-labelled antisense-RNA and the parallel detection of the mature protein with an anfi-bFGF monoclonal antibody. Our study revealed typical patterns of bFGF-m-RNA-expression and intense bFGFprotein deposition during granulation tissue formation and reepithelialJzation of healing bum wounds. 'it, is known that major thermal injuries cause early impairment of wound healing followed by decreased influx of granuiocytes St. the site of injury. The role of granuiocytes in the process of wound healing is not ~"~ "" elucidated, it is now assumed that they are not merely phagocytic cells but active participants in ~n~*'1~.,.,a+~o~: processes secreting_ a number of various cvt-;kines, in order to investigate the effect of There is accumulating evidence that neuropeptides could be involved in the pathogenesis of several inflammatory reactions. Vasocactive Intestinal Polypeptide (VIP) and Substance P (SP) have been detected by immunohistochemistry in normal as well as inflammed skin mostly in perivascular and periglandular location. Both VIP and SP are involved in vasodilatation, mast cell degranulation and irnmunomodulation.We determined the influence of SP and VIP on the proliferation of lymphocytes in patients with psoriasis and healthy individuals. Peripheral blood T-lymphocytes of 6 psoriatics and 6 healthy controls were isolated by density gradient centrifugation and passage over nylon wool. Cell enrichment was controlled by FACS analysis, lx105 T-lymphocytes were then incubated alone or in coculture with 2x104 irradiated autologous lymphocytes in culture medium containing 10 -7 mol/I SP or VIP. Cell proliferation was measured semiquanfitatively by TdR uptake in a betacounter. Significance was tested by the Wilcoxon signed-rank test.Our results show that SP and VIP exert only an effect on unstirnulated T-cells. In healthy individuals but not in patients with psoriasis SP increases significantly proliferation of T-cells. VIP, however stimulates significantly the blastogenesis of T-lymphocytes only in psoriatics.Our results confirm the psychoneuroimmunologic component in inflammatory reactions and VIP and SP could be partially implicated in their pathogenetic mechanisms. Moreover psoriatic lymphocytes show an altered reaction to SP and VIP. This might be due to a preexisting (genetic?) or more likely to an epiphenomenal receptor defect. The adhesive interactions between endothelial cells and circulating ~enkocytes in shock and innammatory vondltions is mediated by several distinct families of ce11-surface determinants. Of particular importance are the leukocyte integrins CDIB / CDlla-c. In this study monoclonal antibodies to two of the u chains (CDlla & CDIIb) and the common [~ chain (CDIB) have been used to investigate leukocyte-dependent and leukocyte-independent plasma leakage in tee skin of rabbite. Plasma leakage was measured as the local accumulation of t2sI-HSA over a 30 rain period, The chemotac~c peptide IMLP (0.1. 30ng) and bradykinin were used to induce cell.dependent and cell-5ndependent leakage respectively, The antibodies used were 6.5E (CDIS), NRI85 (CDlla) and antibody 198 (CDllb). ]ntradermal in~ections of bradyklnin and ~dLP both caused a dose dependent increase in plasma extravasatien (15.~. ffi 2.1p.l to 54.4 z B.Bttl and 16.4 ,-1.8~ to 59.8 z 6.71d respectively. 6.5E (0.0072-2.4mF,/k~ iv) caused a dose dependent inhibition of IMLP-induced but not bradyldnin.inducecl plasma exudation. At 0.24mk/kg, the plasma leakage was completely inhibited, Antibody NR185 produced similar results, Treatment with antibody 198 did not cause inhibition o£ plasma leakage due to either Tnedi~tor. In vitro, The irmnune system ex~nination in persons with bone, chest and abdominal traumatic injury (I group . 22 patients without infectious coz~lications and 2 group -53 patients with wound infections development) was carried out. To restore found immunity disorders and host defense to infection 23 patients of the 2 group were treated with Thymalin-the biologically active peptides prepared from bovine thymus. The examination on t~e I-3 days after injury revealed a considerable decrease of lymphocytes, eD3",$D4 ~ and CD8 cells amo~it in the blood, CD4 /CD8 ratio and indexes of let~ocyte migration inhibition test in both groups of patients. The imm~lity disorders recovered to norm on the 7-10 days in pateents of+the I group. But stable ~eple$ion of CD3 and CD4 cells amount, lower CD4 /CD8 ratio and indexes of leukocyte migration inhibition test in patients of the 2 group were observed~ Besides that, these persons showed higher CD19cells amount and Ig level in the blood. After Thymalin therapy valid ii~rovement of inun~e status was discovered. Also good clinical effect of immunotherapy and best wo~id healing observed in 80% of cases. These results allow us to propose that the thymus involution and the reduction of cell-mediated immunity responsiveness with disturbances of immu_uoregulatio~ on the level of restriction of activated CD4 ThO cells play the most important role in the pathogenesis of wound infections development in persons with traumatic injury.Dept. of Immunology, Military-Nedical Academy, Lebedeva Str. 6, 194175, St.Petersburg, Russia

A severe Impairment of neutrophil (PMN) function often occurs following severe thermal or non-thermal traumatic Injury. Our laboratory has previously reported that following severe burn or non-burn traumatic injury the expression of the p2 Integrlns (CD11a,b,c/CD18) and the FW receptors (CD16, and CD32) were significantly decreased on PMNs, In this study, the effects of GM and G-CSF on the expression of the F~ R and the ~2 Integrln family on PMNs were examined, PMNs were obtained from severe trauma (initial APACHE II score ;z 10) or thermal injury (>30~; total body surface area, >15~ full thickness) and Incubated /n v/tro with GM or G-CSF. The J32 integrins or FcyR were detected with monoclonal antibodies and flow cytometry. GM end G-CSF Induced a sllght increase In the percentage of PMNs expressing CD1 lb, CD16, and CD18 while GM bur not C-CSF Induced an Increase in the percentage expressing CDI 1 a, CD1 lc, and CD32, GM-CSF and to a lesser extent G-CSF Induced an increase in the density (1,5 fold) of the ~2 integrlns on PMNs from normal, burn, and trauma patients, These data suggest that cytoklne modulation with CSFs could have a role clinically in certain situations.

Institute, Dept. of Surgery, 231 Bethesda Ave, Cincinnati, OH, USA, 45267-0558.

Funl~al Infections after Solid Organ Transplantatlon(SOT) Lewis Flint, MD and Ed,~-afd E. Etheredge, ME) Dept. of Surgery Tullrte Univ. School of Medicine New Orleans. Louisiana Infections contribute to increased gra5 loss and mortaliw following SOT. Pr~isposing facton include diabetes, hepatitis, leukopenia, cc.¢xistem infection, and intense, especially triple drug, immunosuppression. Funga] infections occur ~s isolated conditions in 6% and in association with bacterial infection(l 1%), viral infection(3*/.), and combined infections(It%), Candida sp. is the most common fungus recovered but Aspecgillus, Coccidiodies, Cryptococcus, Histoplasma, Mueor~ ghizopus, Tinea, and ToruIop~is s?. also are pathogens. Clinical syndromes vary among orga.aizms or may be variable with a single p~tthogen, For ~ample, with aggressive immunosuppression, Candlda my be localized esophagitis or cystitis or systemically iavaslve with an associated high mortality. AspergilIus presents ~ a diffuse pneumonia while Cryptococcus causes pulmonary and centrad nervons sy'Stem infection, Clinical examination, CT scanning and aggressive sampling for c'ultures a.s wall as serologic tests contribute to diagnosis. Empiric the~py is ind',cated where there is a high level of suspicion. Preventlon of Ca.adlda izfection is ~ci~itated by early remov-a.1 of central }ants, ca~hetess and stents as well as by the use of oral nystatin. Amphotericin ]~ remains the drug of choice for treatment of in.save fungd infection, Surgical resection of infectious loci in the lung and brain is indicated in selected patientS. The main problems of diagnosis in lower respirator-), tract infection are the differentation of infection from colonization or contamination, and the isolation of a reliable and true pathogen. Expectorated sputum may be unreliable in pneumonia, because of contamination by oropharyngeal flora. Although blood cultures may be negative, they provide a precise diagnosis and should be obtained in all pneumonias. Other more invasive procedures are transtracheal needle aspiration, fibrobronchoscopic techniques including protected specimen brush and bronchoalveolar lavage with quantitative culturing and cytological analysis, transthoracic needle aspiration, thoracoscopy -guided biopsy and open lung biopsy. Recently M. E1-Ebiary, A. Torres et al, 1993 reported quantitative cultures of endotracheal aspirates for the diagnosis of ventilator-associated pneumonia offering reliable results in these patients and should be further investigated. Any invasive procedure in a severely ill patient should be carefully directed weighing the risks as well as the benefits, whilst taking the underlying diseases and expected survival into consideration. -Current therapeutic approach is based mainly on monotherapy with broad spectrum antibiotics. Combination therapy is apparently indicated only in P. aeruginosa infections and severe S. aureus pneumonia. 

Graft infection can lead to fulminant graft failure or rapid progressive cirrhosis. For prevention of graft infection immunoprophylaxis, i,e. administration of human polyclonal anti HBs hypedmmunoglobuTin (HIG), starting in the anhepatic phase during operation, has proved to be at least partially succesful when performed on a long term basis.From a total of 415 OLT in adult patients 100 OLT were performed for HBsAg positive liver disease (cirrhosis n=86, fulminant liver failure n=8, retransplantation n=6) in 94 pat. All pat. received 10.000 U HIG in the anhepatic phase and 2.000 U/per day for the first week. A small group of 9 pat. received HIG only for I week (short term immunoprophylaxis), in all other pat. HIG is administered on a long term basis to keep anti HBs serum levels above 100 UII or until graft infection occurs (long term immunoprophylaxis);One-year survival rates are 100% in pat. who were transplanted for fulminant hepatitis, 96% in pat. with cirrhosis and long term prophylaxis, and 78% ir~ pat. with short term prophylaxis. All fatalities were related to HBV graft infection. The total rate of graft infection was 100% under short term prophylaxis and was independent from preoperative HBV DNA status, Under long term prophylaxis graft infection occurad in 28% in pat, negative for HBV DNA. In HBV DNA positive pat. infection rate was 56%, the total rate of reinfection for all pat. with long term prophylaxis was 36%The results of liver transplantation in HBsAg positive pat. are comparable to other indications, Graft infection with hepatitis B virus ist the major risk factor for these patients. Under long term therapy with HIG the rate of graft infection can be significantly reduced.

THE CRUCIAL CELLULAR ELEMENT FOR MODS-MOF: MONOCYI'F_./M ACROPHAOE Ronald V. Meier, M,D., F.A,C,S. The severely :injured or crldcally ill surgical patient is at high risk for immune dysfunction. A major consequence of this immune dysfunction is multiple organ dysfunction and failure leading to death, The underlying etiology is now recognized to be an uncontrolled, unfocused, disseminated activation of the host normally protective inflammatory. ,, cascades.. The resultant "mahgnant' systemic" inflan'a'natlon produces d~ffuso multiple organ bystander injury !eading to progressive organ dysfunction and failure. Systemic malignant inflammation involves diffuse actlvatton of all components of the humoral and cellular inflammatory host response. Of these various components, the macropha~e is the crucial central cellular element. The tissue fixed macrophage Is ideally located diffusely throughout the various organs injured to orchestrate the inflammatory process. The macrophage is long-lived and highly metabolic, The macrophage regulates both the extent and the dissemination of the inflammatory processes. The macrophage is an exu'emely active c¢11 capable of producing and releasing not only directly eytotoxlc agents, s irnil~, to the neutrophil, including oxidants and numerous proteases out also the multitude of other cytokines and initiators of the interacting inflammatory cascades. The macrophage is the central source for ehemotactic agents (IL-8, LTB4, C5a) for neutrophils and other inflammatory cells, production of vasoaetive arachidonie acid metabolites (TX, PGI2, POE, LT's), complement components (C3a, CSa), thrombotic agents (PCA, TX), metabolic and physiologic modulators (IL,1, IL-6 or TNF), and immunosuppressivc agents (POE2, IL-10). These products of the macrophage are highly effective in enhancing and augmenting the inflammatory response. Disseminated activation otthe macrophage Is critical to the induction of the long-term diffuse activation of inflammation necessary to induce multiple organ injury and failure. Our ability to elucidate the molecular mechanisms that control the macrophage will lead to our ability to conu'ol the maerophage response and prevent MODS-MOF.Flarborview Medical Center, 325-9th Ave ZA-16, Seattle, WA 98104 USA