key: cord-0006105-k7vx9vly
authors: Vadas, P.; Pruzanski, W.; Stefanski, E.
title: Extracellular phospholipase A(2): causative agent in circulatory collapse of septic shock?
date: 1988
journal: Agents Actions
DOI: 10.1007/bf02028289
sha: 93f759546b5344122523b07fcc0e69ce3066bffa
doc_id: 6105
cord_uid: k7vx9vly

nan

The syndrome of septic shock is still a common cause of death in spite of recent advances in antimicrobial therapy. This is, paradoxically, a reflection of medical progress in that the factors predisposing to septic shock are those of old age, debilitating underlying disease, prior antibiotic therapy, use of immunosuppressive agents and increasing reliance on invasive diagnostic and therapeutic interventions [1] . There are an estimated 100 000 to 300 000 cases of gram-negative sepsis in the United States per year [2] with mortality rates ranging from 30-80%. These high mortality rates have remained virtually unchanged for decades [3] . In attempting to understand the pathogenesis of the septic shock syndrome, various circulating mediators of cardiovascular collapse have been sought. Among these putative mediators are included tumour necrosis factor [4] , histamine, kinins, endorphins, myocardial depressant factors and various eicosanoids [5] . So far, none of these candidates has fulfilled the requisite criteria for mediators in the setting of septic shock [6] , although products of the arachidonic acid cascade are certainly amongst the frontrunners. In spite of the lack of success in the search for such a mediator, the concept of a circulating factor, released into the systemic circulation in response to bacte-Supported by The Arthritis Society and The Wellesley Hospital Research Institute. remia or endotoxemia remains an attractive notion. On the basis of early studies in our laboratory describing the mediation of regional blood flow in inflamed sites by extracellular phospholipase A2 (PLA2) [7-9], we postulated that the intravascular release of a soluble PLA2 initiates a sequence of events culminating in cardiovascular collapse in the setting of septicemia. The experimental and clinical studies described below suggest that circulating PLA2 may well play a central role in the pathogenesis of circulatory collapse in septic shock.

Phospholipase A2 has been found in association with localized sites of inflammation in both experimental animals and in man. High levels of PLA2 activity have been identified in glycogen-induced ascitic fluid in rabbits [10] and in lymph plasma in sheep subsequent to intradermal challenge with vaccinia virus or purified protein derivative of Bacille Calmette-Guerin [8] . A soluble PLA2 has been described [9, 11] , characterized [12] and purified [13] from synovial fluids of patients with rheumatoid arthritis. The levels of PLA2 were found to correlate significantly (p<0.0001) with disease activity in rheumatoid arthritis [ 14] . Soluble PLA2 is secreted extracellularly in lymph by activated macrophages [8, 15] , whereas the cellular sources of the synovial fluid enzyme include osteoblasts [16] and chondrocytes [17] . Extracellular PLA2 is both pro-inflammatory and vasoactive. In view of the association of PLA2 to inflamed sites, the development of inflammatory changes in rabbit skin were studied subsequent to intradermal injection of both active and inactivated venom and pancreatic PLA2. PLA2, at concentrations encountered in human disease, caused acute inflammatory changes characterized grossly by erythema and induration, and histologically by inflammatory cell infiltration, vascular and tissue damage, and microabscess formation [18] . The severity of the resultant lesions was both time-and dose-dependent. The development of inflammatory lesions subsequent to intradermal injection was also associated with increased regional blood flow or hyperernia [19] . Using the microsphere method for quantitation of regional blood flow [20] , PLA2 was found to induce a profound and sustained increase in dermal blood flow after intradermal injection [7] . This hyperemia was abrogated by pretreatment of local injection sites with indomethacin, suggesting that at least part of the vasoactive effect of PLA2 was mediated by cyclooxygenase products. Thus, a soluble PLA~ is released extracellularly in response to local inflammatory stimuli and exogenous PLA2 is locally vasoactive causing increased regional blood flow.

From these earlier data we postulated that vasoactive changes in a systemic inflammatory response may well involve fundamentally similar pathways as for the local inflammatory response, but differing in magnitude. Endotoxemia, the experimental counterpart to clinical gram-negative septic shock, has been regarded as a systemic inflammatory response [21, 22] , and was therefore chosen as our experimental model. Endotoxin shock was induced in rabbits by i.v. challenge with E. coli endotoxin 0127:B8. Changes in mean arterial blood pressure (MABP) and serum PLA2 activity were followed over a 5-hour period. Animals challenged with an LDso dose of E. coli endotoxin became hypotensive over the period of observation with a progressive fall in MABP. Serum PLA2 activity rose concomitantly by a mean of l l.2-fold over baseline levels. The correlation of the rise in serum PLA2 activity with the fall in MABP yielded a statistically significant monotonic relationship [23] . In contrast, there was no change in either MABP or serum PLA2 activity in control animals injected with non-pyrogenic saline. Since we had previously shown that mitogenstimulated macrophages secreted PLA2 extracellularly whereas those coincubated with dexamethasone sodium phosphate did not [15] , we investigated the effect of dexamethasone pretreatment in the setting of endotoxin shock. Animals pretreated with dexamethasone remained normotensive and serum PLA2 levels remained unchanged suggesting that steroids administered prior to systemic endotoxemia may prevent the liberation of active PLA2 into the systemic circulation [23] . Finally, the PLA2-containing fraction of septic shock serum (MW 10000-13 000) obtained by gel filtration chromatography, was re-infused into normal rabbits. Infusion of the exogenous PLA2containing fraction to levels encountered in endotoxemia, reproduced the hypotensive effect of endotoxin itself. Pretreatment of the PLA2 fraction with the PLA2 inhibitor p-bromophenacyl bromide (pBPB) prior to re-infusion resulted in inhibition of PLA2 activity, and as well, abrogated the hypotensive effect of exogenous PLA2. Linear regression analyses of MABP responses of those groups of rabbits treated with (a) saline, (b) endotoxin, (c) PLA2-active fraction of shock serum, and (d) pBPB-treated PLA2-active fraction of shock serum revealed that the rate of fall in MABP in the PLA2-active fraction-treated group exactly paralleled that of the endotoxin-treated group consistent with the postulated common pathway of action. These data demonstrated that PLA2 is 1) released in massive amounts into the systemic circulation following endotoxin challenge in a model of experimental endotoxin shock, 2) the rise in serum PLA2 activity is directly proportional to the fall in MABP, 3) dexamethasone pretreatment prior to endotoxin challenge protects against hypotension and prevents release of active PLA2 into the systemic circulation, and 4) exogenous PLA2 isolated from septic shock plasma causes systemic hypotension.

Studies of the role of circulating PLA2 in the pathogenesis of hypotension were extended to the clinical counterpart of experimental endotoxin shock, i.e. gram-negative septic shock (GNSS). Sera taken from patients during the acute hypotensive phase of GNSS were assayed for PLA2 activity and correlated to the presence (or absence) of adult respiratory distress syndrome (ARDS). Thirty-four patients with septic shock were subdivided into two groups: those with GNSS with ARDS vs those with GNSS without ARDS, as diagnosed by Hallman's criteria [241. During the acute, hypotensive phase of septic shock, serum PLA2 was elevated in 34 of 34 patients (100%) by a mean of 16-fold over that of serum PLA2 activity of 21 in-hospital controls matched for underlying disease (p < 0.0001) [25] . Patients with GNSS and ARDS had a mean serum PLA2 activity of 20.2-fold greater than controls (p < 0.0001) whereas those with GNSS without ARDS had a mean 14-fold increase in PLA2 activity over controls (p < 0.0001). The difference in serum PLA2 activity in patients with confirmed GNSS with and without ARDS was statistically significant (p = 0.01) [25] . Thus in this cross-sectional study of 34 patients with GNSS, 1) serum PLA2 activity was increased in all patients during the acute hypotensive phase of GNSS and 2) septic patients with concurrent ARDS had a mean serum PLAz activity 70% greater than septic patients without respiratory insufficiency.

Twelve consecutive patients with the diagnosis of GNSS were entered into a prospective study of serum PLA2 profiles and concomitant hemodynamic changes from diagnosis to death or discharge. The magnitude of circulatory instability or collapse was quantitated by the hemodynamic instability score (HIS) [26[ . In all 12 patients, serum PLA2 activity correlated with the hemodynamic status of each patient as reflected by HIS scores. Peak serum PLA2 activity consistently cor-responded to periods of maximal circulatory collapse with a return towards normal levels in the convalescent phase. Peak serum PLA2 activities as high as 291-fotd over controls were observed. The correlation of PLA2 and HIS achieved nominal levels of significance of p< 0.001 using a logistic regression model for ordinal data. Correlations were positive in 12 of 12 patients, consistent with a highly statistically significant correlation of PLA~ activities and HIS scores (p < 0.02) [26] . This was the first prospective study of serum PLA2 profiles in patients with septic shock. This study demonstrated that endogenous serum PLA2 activity correlated with the magnitude of circulatory collapse in clinical septic shock, as it did in rabbits with experimental endotoxic shock [23] .

Since we had previously shown that dexamethasone prevents the intravascular release of active PLA2 following endotoxin challenge [23] , and since others have shown that steroids induce the synthesis of specific phospholipase inhibitory proteins known as lipocortins (reviewed in [27, 28] ), it was of interest to examine the serum profiles of endogenous cortisol, ACTH and PLAz in patients during the course of septic shock. Eight patients with septic shock were studied prospectively. Serum cortisol and ACTH levels were quantitated by radioimmunoassay simultaneously with PLA2 determination. In 4 survivors, peak cortisol levels ranged from 662-2480nmol/1 (reference range 50-600 nmol/1). Cortisol levels in 4 non-survivors ranged from 450-2320 nmol/l. In contrast, there was little fluctuation in serum ACTH levels in all patients studied, with ACTH levels consistently less than 5 pmol/l. The correlation coefficient of simultaneous serum PLA2 and cortisol levels in survivors of septic shock showed a consistent statistical correlation of r>0.833 (p<0.01) whereas the coefficient of correlation of PLA2 and cortisol in non-survivors was consistently less than r--0.505. There was no correlation in any patient of serum ACTH with cortisol or PLA2 [29] . Using McCullagh's regression model for ordinal data [30] , a statistically significant association between PLA2 and cortisol was detected amongst survivors (p<0.0001) as well as for the total sample (p<0.001). In pa-tients who died, the assocciation between PLA2 and cortisol achieved only borderline significance (p=0.07). Thus, in the first prospective study of cortisol levels in man during septic shock we noted a strong statistical correlation of cortisol and PLA2 in survivors of septic shock, while such a correlation was absent in all non-survivors. The absence of a correlation of cortisol and ACTH in patients with septic shock suggested that the stimulus for cortisol release arises from outside the hypothalamic-pituitary axis.

Septic shock serum was fractionated by gel filtration column chromatography as described [7] . Fractions were assayed for PLA2 activity using radiolabeled E. coli substrate. Human shock serum PLA2 was eluted as a single peak of molecular weight 14000 [7]. Human septic shock serum PLA2 activity was assayed [31] using the substrates 2-[l*C]-palmitoyll-palmitoyl-3-phosphatidylcholine or 1-[~*C]-palmitoyl-sn-glycero-3-phosphocholine in micellar dispersion in aqueous buffer. Assays were performed in substrate excess, in the linear range of the assay with respect to incubation times and enzyme concentration. Incubation of human septic shock serum phospholipase with synthetic 2-[1-~*C]-palmitoyl -1-palmitoyl-3-phosphatidylcholine resulted in hydrolysis of 23.2% of substrate. Of the total products formed, 96.6% of radiolabel was associated with free fatty acid, and 3.4% was associated with lysophosphatide. In comparison, incubation of radiolabeled substrate with authentic PLA2 from Naja Naja venom resulted in 97.1% of radiolabeled product comigrating with free fatty acid and 2.9% with lysophosphatide. These data were consistent with a 2-acyl positional specificity of septic shock serum phospholipase [32] . The pH optimum of HSS-PLA2 activity was determined in the presence of either 7 mM CaCI2 or 5 mM EDTA, since PLA2 may be absolutely calcium-dependent or optimally active in the absence of exogenous calcium [27] . In the presence of calcium, HSS-PLA2 displayed a monophasic pH optimum of 7.5. Chelation of calcium by EDTA significantly reduced PLA2 activity at pH 7.5 (by 97%) and did not reveal a calcium-independent species of PLA2, particularly in the acidic range [32] . The calcium dependence of the enzyme was examined over the range of 0-10 mM CaC12. PLA2 was optimally active in the presence of 2 mM calcium with a progressive loss of enzyme activity at higher concentrations. The effect of temperature on the rate of hydrolysis of radiolabeled E. coli phospholipids was examined. Enzyme activity increased linearly with temperature over the range of 30-40 ~ [32] . Septic shock serum PLA2 was tested for immunoreactivity against anti-human pancreatic PLA2 by radioimmunoassay [26] . Sera with PLA2 activities as high as 260-fold normal did not crossreact with anti-human pancreatic PLA2.

We had postulated that the presence of bacteria or bacterial products in the systemic circulation causes the intravascular release of large amounts of a soluble PLA2 and by generation of vasoactive products of PLA2 hydrolysis, contributes to circulatory collapse in septic shock. Exogenous PLA2 is vasoactive both locally and systemically. Intravenous infusion of snake venom PLA2 caused a rapid and profound drop in MABP in dogs [33] , cats [34] and rats [35] . Similarly, infusion of septic shock serum PLA2 produced hypotension in rabbits at a rate parallel to that induced by endotoxin [23] . Serum levels of endogenous PLA~ correlated significantly with the fall in MABP in both experimental endotoxin shock as well as in patients with septic shock. Hyperphospholipasemia has been confirmed in endotoxin shock in rats [36] as well as in patients with severe septicemia [37] . The source of circulating PLA2 is, as yet, undefined. The septic shock PLA2 appears to be of extra-pancreatic origin based on the absence of cross-reactivity with anti-human pancreatic PLA2 antiserum. Nor does the enzyme appear to originate from spleen or circulating leukocytes in that comparable elevations of serum PLA2 have been noted in septic patients post-splenectomy or with profound leukopenia post-cytotoxic chemotherapy (unpublished observations).

While the mechanism of PLA2-induced hypotension is not known, the release of PLA2 into the systemic circulation may lead to the generation of vasoactive eicosanoids and lysophosphatides. Perfusion of rat liver ex vivo with exogenous porcine pancreatic PLA2 resulted in the release of 6-ketoprostaglandin F~a and thromboxane B2 from intact cells and tissues [38] . Intravenous infusion of snake venom PLA2 in rats caused a significant fall in MABP which correlated with the rise in the ratio of 6-keto-PGFla to TXB2 [39, 40] . PLA2 was also shown to cause an increase in lung perfusion pressure [41] , which correlated linearly with the ratios of TXB2 to 6-keto-PGF a. These data suggest that circulating PLA2 attacks phospholipid substrate of intact cells and tissues leading to the generation of a number of vasoactive lipids. Glucocorticoids appear to exert at least part of their anti-inflammatory effect through the synthesis and release of endogenous phospholipase inhibitory proteins known as lipocortins (reviewed in [27, 28) ). Several lines of evidence have been advanced to support this concept. Exposure of a number of cell types to steroids in vitro resulted in the extracellular release of lipocortin [42, 43]. As well, lipocortin was recovered in body fluids after parenteral administration of steroids [34] . Significantly, the anti-inflammatory effects of steroids were abrogated in vivo by coadministration of a monoclonal anti-lipocortin antibody [28]. Experimentally, we have shown that pretreatment of rabbits with steroids prior to endotoxin challenge protected against endotoxin-induced hypotension and prevented PLA2 release into the systemic circulation. These observations suggest an antiphospholipase effect of exogenous or endogenous corticosteroids. In patients with septic shock, cortisol levels were concordant with serum PLA2 levels in all survivors of septic shock but discordant in all non-survivors. Interestingly, there was evidence of complete dissociation of the pituitary-adrenal axis in these patients, with presumptive evidence of extrapituitary regulation of adrenocortical function. Thus, postulated mechanisms of pro-inflammatory and anti-inflammatory pathways involving phospholipase A2, phospholipase inhibitory proteins (PLIP) and corticosteroids, will require further investigation. These data demonstrate that PLA2 fulfills several requisite criteria for mediators of circulatory shock as proposed by Lefer [6] . First, there is a demonstrable increase in circulating PLA2 in response to bacteria or bacterial products. Second, exogenous PLA2 was shown to cause hypotension in experimental animals and endogenous PLA2 levels correlated with the severity of hypotension in both animals and man. The third and final criterion, amelioration of circulatory function by inhibiting the formation or antagonizing the action of PLA2 awaits the development of a selective and non-toxic inhibitor of circulating PLA2 in man.

Gram-negative bacteremia. Reassessment of etiology, epidemiology and ecology in 612 patients

Current perspectives on septic shock

The effects of high.dose corticosteroids in patients with septic shock

Shock and tissue injury induced by recombinant human cachectin

Release ofvasoactive hormones and circulatory changes in shock

Eicosanoids as mediators of ischemia and shock

Extracellular phospholipase A2 mediates inflammatory hyperemia

The appearance and significance of phospholipase A2 in lymph draining tuberculin reactions

Role ofextracelhdarphospholipase A2 in inflammation

Isolation and characterization of a phospholipase A: from an inflammatory exudate

Phospholipase A2 activity in sera and synovial fluids in rheumatoid arthritis and osteoarthritis: its possible role as a proinflammatory enzyme

Characterization of extracellular phospholipase A2 in rheumatoid synovial fluid

Purification of a soluble phospholipase A2 from synovial fluid in rheumatoid arthritis

Serum phospholipase A2 levels correlate with disease activity in rheumatoid arthritis

The efficacy of anti-inflammatory agents with respect to extracellular phospholipase A2 activity

Synthesis and secretion of soluble phospholipase As from cultured osteoblasts

lnterleukin 1 activates phospholipase A 2 in rabbit chondrocytes: a possible signal for IL-1 action

Inflammatory effect of intradermal administration of soluble phospholipase A2 in rabbits

Secretion of a hyperemia-inducing moiety by mitogen or glycogen-stimulated mononuclear inflammatory cells of sheep and rabbit

Cutaneous bloodflow measurements: a standardization of the microsphere assay for yasocative agents

Bacterial shock

The inflammatory response ofendotoxins

lnvovlement of circulating phospholi. pase A2 in the pathogenesis of hemodynamic changes in en. dotoxin shock. Can

Evidence of lung surfactant abnormality in respiratory failure

Elevated plasma phospholipase A2 levels: correlation with the hemodynamie and pulmonary changes in gram-negative septic shock

The pathogenesis of hypotension in septic shock: correlation of circulating phospholipase A2 levels with circulatory collapse

Role of secretory phospholipases A2 in the pathobiolog), of disease

Macrocortin and the antiphospholipase proteins

Concordance ofendogenous cortisol and phospholipase A2 levels in gram-negative septic shock, A prospective study

Regression models for orindal data

Comparative analysis of assays for phospholipase A~ activity using biomembrane associated and micellar substrates. The influence of albumin

Potential therapeutic efficacy of inhibitors of human phospholipase A2 in septic shock

Pharmacological studies of the major fractions of bee venom

Beitrage zur pharmaeologie yon phospholipase

The effects of honey bee venom and two of its constituents, melittin and phospholipase A2, on the cardiovascular system of the rat

Enhanced phospholipase A2 activity in rat plasma, liver and intestinal mucosa following endotoxin treatment: possible explanation for the protective effect of indomethacin in endotoxic shock

Activity ofphospholipase A compared in serum of patients with pancreatic and nonpancreatic diseases

Phospholipase A2 stimulated release of prostanoids from the isolated perfused rabbit liver: implications in regional cellular injury. Can

Effects of phospholipases A2 from Vipera russelli snake venom on blood pressure, plasma prostacyclin level, and renin activity in rats

Cardiovascular effects ofphospholipases A2 purified from various snake venoms

Release of slow-reacting substance from the guinea-pig lung by phospholipases A2 of Vipera russelli snake venom

A phospholipase A2 inhibitory protein in rabbit neutrophils induced by glucocorticoids

Glucocorticoids induce the formation and release of anti-inflammatory and anti-phospholipase proteins into the peritoneal cavity of the rat

The authors thank Ms. Susan Mueller for skilled secretarial assistance. Received 20 October 1987; accepted 4 December 1987