INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

PROGRESSIVE MYOCARDIAL dysfunction, characterized by dilatation and reduced ejection fractions of both ventricles (33, 34), contributes to the cardiocirculatory abnormalities of septic shock. Despite the clinical importance of this disease (35), the pathogenetic mechanisms leading to the loss of myocardial function in sepsis are still not fully elucidated. Cardiodepressant effects of cytokines, predominantly tumor necrosis factor- and interleukin-1, have been implicated in the development of cardiac depression in sepsis (26, 27). In addition, experimental data suggest thatdespite preserved coronary blood flow in sepsis (10)impairment of coronary vasoregulation and myocardial regional perfusion may also depress cardiac performance (7, 13, 22, 40). Moreover, there is recent experimental evidence that recruitment and activation of polymorphonuclear neutrophils (PMN) may also contribute to the development of septic myocardial failure. In endotoxemic animals PMN have been shown to accumulate in the myocardium (3, 14), and very recently it was demonstrated that leukocytes originating from endotoxemic rabbits are retained in the coronary circulation of isolated hearts and cause contractile dysfunction (17).

In general, adhesion of PMN to the vascular endothelium requires the upregulation of endothelial adhesion molecules [e.g., E-selectin, intercellular adhesion molecule (ICAM)-1], and abundant expression of these ligands may occur under inflammatory conditions (1). Subsequent transmigration and activation of PMN, resulting in the release of a variety of inflammatory mediators, may well contribute to organ failure in sepsis (42). Next to the liberation of toxic oxygen species and proteolytic enzymes, the formation of 5-lipoxygenase (5-LO) products of arachidonic acid (AA), in particular leukotrienes (LTs), may be relevant under these circumstances. In addition to the release of LTB₄ and 5-hydroxyeicosatetraenoic acid (5-HETE) by activated neutrophils, the generation of vasoactive cysteinyl (cys)-LTs (i.e., LTC₄, LTD₄, LTE₄) may contribute to maldistribution of regional perfusion and thereby organ dysfunction. Cys-LTs were, indeed, shown to impair coronary and contractile function in vitro (37), and in vivo cys-LTs were found to contribute to the tissue destruction in myocardial infarction (28, 36). Although PMN are unable to synthesize cys-LTs by themselves, the unstable intermediate LTA₄ may be released by PMN and metabolized by vicinal endothelial cells to cys-LTs, a concept termed transcellular or cooperative biosynthesis of cys-LTs (11, 30).

Proteinaceous exotoxins of clinically relevant bacteria, among others the -toxin of Staphylococcus aureus, have been shown to promote neutrophil adhesion to endothelial cells in vitro (25). Moreover, very recent studies from our laboratory (16, 40) demonstrated that staphylococcal -toxin provokes coronary vasoconstriction and severe myocardial depression in isolated rat hearts, with toxin-elicited thromboxane generation turning out to be the main contributor to both events (40). Although quantification of exotoxins in biological samples is a problem not fully resolved because of the very rapid membrane incorporation of this agent, experimental data suggest that exotoxins elicit distinct biological effects over a wide concentration range and even a few numbers of exotoxins may suffice to activate a variety of target cells (5).

Against this background, the aim of the present study was to evaluate whether low doses of staphylococcal -toxin cause the accumulation of neutrophils in isolated rat hearts and, if so, to study the functional consequences of coronary PMN retention. An ICAM-1-dependent sequestration of PMN was, indeed, found in the -toxin-treated hearts. Subsequent challenge with the bacterial pathogenicity factor N-formyl-methionyl-leucyl-phenylalanine (fMLP) and supply with the LT precursor AA then resulted in a marked generation of LTs, associated with an increase in coronary perfusion pressure (CPP), and a loss of contractile performance. Interestingly, pharmacological interventions and mediator analysis suggested that the cardiac abnormalities were largely attributable to transcellular synthesis of cys-LTs in -toxin-treated hearts.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Materials. Purified -toxin from S. aureus was kindly provided by S. Bhakdi (Institute of Medical Microbiology, Johannes Gutenberg University, Mainz, Germany). Aliquots of the lyophilized toxin, proven to be endotoxin free (6), were dissolved in phosphate buffer solution and stored at 80°C until experimental use. The murine monoclonal anti-rat ICAM-1 antibody (anti-ICAM-1 Ab 1A29) was obtained from Pharmingen (Hamburg, Germany) and the lipoxygenase inhibitor MK-886 from Calbiochem (Bad Soden, Germany). Ficoll-Paque was purchased from Amersham Pharmacia Biotech (Uppsala, Sweden) and polyvinyl alcohol from Merck-Schuchardt (Hohenbrunn, Germany). The nonspecific murine antibody MOPC-21, HEPES, fMLP, AA, o-dianisidine dihydrochloride, and Gill's hematoxylin 3 were obtained from Sigma (Deisenhofen, Germany). Other materials used were as described for specific procedures.

Preparation of human PMN. For neutrophil isolation, EDTA-anticoagulated blood obtained from healthy donors was centrifuged in a Ficoll-Paque gradient. Erythrocytes were sedimented in polyvinyl alcohol, and residual erythrocytes were removed by hypotonic lysis as previously described (19). Cells were washed twice (150 g, 10 min, 4°C), resuspended in Hanks' balanced salt solution (HBSS)-HEPES, and added to the perfusate at a final concentration of 10⁶ PMN/ml. Cell purity was >98% and cell viability >96% as assessed by trypan blue dye exclusion. Surface expression of CD11/CD18 was verified by using a FACStar^Plus flow cytometer (Becton Dickinson, Mountain View, CA) according to standard techniques.

Isolated heart perfusion and experimental protocol. The heart perfusion technique was previously described in detail (40). Briefly, male Wistar rats (Charles River) were heparinized (heparin 1,000 IU/kg) and anesthetized (pentobarbital 60 mg/kg) by intraperitoneal injection. The hearts were rapidly excised, attached to a Langendorff perfusion apparatus, and perfused at constant flow (10 ml · min¹ · g¹) with a modified Krebs-Henseleit buffer solution (KHBS) containing (in mM) 125 NaCl, 4.3 KCl, 1.1 KH₂PO₄, 1.3 MgCl₂ · 6H₂O, 2.4 CaCl₂ · 2H₂O, 25 NaHCO₃, and 13.32 glucose. The perfusate was gassed with carbogen (5% CO₂-95% O₂). PO₂ was 500 ± 45 Torr (66.7 ± 6.0 kPa) and PCO₂ was 35 ± 5 Torr (4.7 ± 0.7 kPa) at 37°C. All hearts were initially rinsed with 150 ml of KHBS in a nonrecirculating mode before switching to recirculation (total volume 50 ml).

Determination of myocardial myeloperoxidase activity. Myeloperoxidase (MPO) was determined as described previously (8). The supernatants of the homogenized left ventricular free wall were reacted with 0.167 mg/ml of o-dianisidine dihydrochloride and 0.0005% H₂O₂ in 50 mM phosphate buffer at pH 6.0. The change in absorbance was measured photometrically at 460 nm. One unit of MPO is defined as the quantity of enzyme hydrolyzing 1 mmol of peroxide/min at 25°C. MPO activity is given in units per gram of wet tissue.

Immunohistochemical and histological analysis of ICAM-1 expression in myocardium. For immunohistochemical analysis hearts were perfused either in the presence or in the absence of -toxin (0.125 µg/ml) for 210 min without PMN, fMLP, and AA. Specimens of the left ventricle were prepared for immunohistochemical analysis as plastic sections as described previously (40). Immunohistochemical procedures were performed as described by Beckstead et al. (4) with the avidin-biotin immunoperoxidase technique (Vectastain ABC reagent; Vector Laboratories, Burlingame, CA). Immunohistochemical analysis was performed with anti-ICAM-1 Ab 1A29. Incubation of the primary antibody was carried out overnight at different dilutions, of which 1:50 gave the highest degree of immunolocalization and the least nonspecific background staining. The sections were lightly counterstained with Gill's hematoxylin 3, and examined using a Zeiss light microscope (Zeiss, Göttingen, Germany) at ×400. To determine the presence of adhering and infiltrating neutrophils, sections stained with hematoxylin and eosin were examined at ×400.

Relative mRNA quantitation. Relative mRNA quantitation was performed with the Sequence Detection System 7700 (PE Applied Biosystems, Foster City, CA) and real-time PCR. Applying comparative quantitation threshold cycle of target gene was normalized to an internal standard gene as previously described in detail (12).

cDNA synthesis and real-time PCR. For cDNA synthesis and real-time PCR, reagents as well as primers and probes were applied as previously described (15). Two microliters of cDNA were applied to each sample. Primers were added to a final concentration of 300 nM each and hybridization probes to a final concentration of 200 nM in a final volume of 50 µl (Table 1). Cycling conditions were 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 60 s.

Analysis of LTs. After the experiments were terminated, the perfusate was collected in total and stored at 20°C. Determination of LTs, i.e., LTB₄, 5-HETE, and the cys-LTs LTC₄, LTD₄, and LTE₄, was performed as previously described in detail (24). Identification and quantification of LTs were performed using a sequence of solid-phase extraction, isocratic reversed-phase HPLC separation, on-line photodiode array detection, and spectrum analysis for identification and measurement of all compounds within one run as described (24).

Statistical analysis. All data are given as means ± SE and were analyzed by one-way analysis of variance, followed by Tukey's honestly significant difference test. P < 0.05 was considered to be significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Immunohistochemical localization and mRNA expression of ICAM-1 in -toxin-perfused hearts. Immunohistochemical analysis of hearts perfused with staphylococcal -toxin (0.125 µg/ml) for 210 min revealed a positive staining for ICAM-1 distributed on the coronary vascular endothelium, whereas no immunostaining was found in time-matched control hearts (Fig. 1). When either the primary or the secondary antibody was replaced by nonimmune serum, no staining was observed (not shown). Moreover, relative expression of ICAM-1 mRNA was increased in toxin-perfused hearts after 180 min by ~33% (control hearts 0.798 ± 0.108 ICAM-1 mRNA copies per 1 copy PBGD mRNA, -toxin-perfused hearts 1.062 ± 0.277 ICAM-1 mRNA copies per 1 copy PBGD mRNA).

View larger version (77K): [in this window] [in a new window]   Fig. 1.   Intercellular adhesion molecule (ICAM)-1 expression within the myocardium of -toxin-perfused hearts (0.125 µg/ml) after 210 min. Photomicrographs of heart tissue incubated with antibodies directed against ICAM-1 and labeled with peroxidase substrate solution are shown. Whereas ICAM-1 was not detected in the myocardium of control hearts (a), ICAM-1 was located on the vascular endothelium of -toxin-perfused hearts (b). Brown reaction product is present at sites of antigen localization. Magnification ×400.

Application of PMN, fMLP, and AA depresses cardiac performance of -toxin-perfused hearts: impact of an anti-ICAM-1 Ab and lipoxygenase inhibitor. Neither perfusion of -toxin (0.125 µg/ml) before application of PMN nor addition of PMN to the perfusate significantly altered CPP, LVDP, and the maximum rate of left ventricular pressure rise (dP/dt) compared with control hearts: after 180 min (before application of PMN) CPP was 72 ± 5 (-toxin) vs. 75 ± 10 (control) mmHg, LVDP was 81 ± 8 (-toxin) vs. 79 ± 5 (control) mmHg, and dP/dt was 2,780 ± 198 (-toxin) vs. 2,800 ± 235 (control) mmHg/s. Likewise, after PMN had been added to the perfusate of control or -toxin-perfused hearts, no significant alterations of cardiac performance were observed: CPP was 64 ± 5 (-toxin) vs. 68 ± 7 (control) mmHg, LVDP was 79 ± 5 (-toxin) vs. 78 ± 4 (control) mmHg, and dP/dt was 2,700 ± 155 (-toxin) vs. 2,725 ± 214 (control) mmHg/s. However, when fMLP (2 µM) and AA (25 µM) were subsequently coadministered (designated as fMLP-AA), a significant depression of both LVDP and dP/dt and a significant increase in CPP occurred in -toxin-treated hearts at 210 min. In contrast, no changes in these parameters were noted in -toxin-free control hearts challenged with PMN and fMLP-AA in a corresponding manner (Fig. 2).

View larger version (48K): [in this window] [in a new window]   Fig. 2.   Alterations of cardiac performance in response to polymorphonuclear neutrophils (PMN), N-formyl-methionyl-leucyl-phenylalanine (fMLP), and arachidonic acid (AA) in -toxin-perfused hearts: impact of an anti-ICAM-1 antibody (Ab) and a lipoxygenase inhibitor. Hearts were perfused with -toxin (0.125 µg/ml) for 180 min in the absence or presence of an anti-ICAM-1 Ab (2 µg/ml), a nonspecific antibody (MOPC-21, 2 µg/ml), or MK-886 (7.5 µM). PMN (106/ml) were added to the perfusate (185 min) and subsequently (200 min) stimulated with fMLP (2 µM) and AA (25 µM). Control hearts were perfused in the absence of -toxin according to the same protocol. Bar graphs represent alterations of physiological parameters 10 min after application of fMLP and AA for left ventricular developed pressure (LVDP; A), maximum rate of left ventricular pressure rise (dP/dt; B), and coronary perfusion pressure (CPP; C). Means ± SE of at least 3 independent experiments are given. *P < 0.05 vs. all other groups.

Quantification of PMN accumulation in -toxin-perfused hearts. Histological analysis showed marked retention of PMN in toxin-perfused hearts (Fig. 3A). To determine the accumulation of infused neutrophils in -toxin-treated hearts MPO activity was measured at the end of the experiments. Baseline MPO activity contained in -toxin-perfused hearts without application of PMN before fMLP-AA exposure was 0.166 ± 0.021 U/g wet wt. Although neutrophil addition in control hearts slightly increased MPO activity, markedly higher MPO activity was found in -toxin-challenged hearts perfused with PMN, indicating accumulation of neutrophils in the cardiac vasculature in response to -toxin. When the anti-ICAM-1 Ab (2 µg/ml) was applied to toxin-perfused hearts before PMN application the increase in MPO activity was completely suppressed. In the presence of the 5-LO inhibitor MK-886, the increase in MPO activity in toxin-perfused hearts was not prevented (Fig. 3B).

View larger version (100K): [in this window] [in a new window]   Fig. 3.   Neutrophil accumulation in -toxin-perfused hearts. A: to determine the presence of adhering and infiltrating neutrophils in control (a) and -toxin-perfused (b) hearts, myocardial sections were stained with hematoxylin and eosin and examined at ×400. PMN accumulation was determined by quantifying cardiac myeloperoxidase (MPO) activity after terminating the experiments depicted in Fig. 2. B: bars represent means ± SE of at least 3 independent experiments. *P < 0.05 vs. all other groups.

Generation of cys-LTs, LTB₄, and 5-HETE in -toxin-perfused hearts. Without prior addition of PMN, application of fMLP-AA to either control or toxin-perfused hearts caused no detectable baseline formation of cys-LTs (LTC₄, LTD₄, LTE₄; not shown) except for a minor increase in LTD₄ (9.1 ± 2.7 pg/ml) in control hearts. In control hearts perfusion of PMN with subsequent fMLP-AA exposure resulted in an increase in LTD₄ and LTE₄ but not LTC₄. However, infusion of PMN before application of fMLP-AA in -toxin-perfused hearts caused a marked increase in perfusate levels of all cys-LTs (Fig. 4). In contrast, when the anti-ICAM-1 Ab was applied to -toxin-treated hearts before PMN, generation of cys-LTs in response to fMLP-AA was significantly suppressed. An even more pronounced inhibition of cys-LTs synthesis was noted in the presence of MK-886.

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Release of cysteinyl (cys)-LTs into the perfusate after application of PMN, fMLP, and AA in -toxin-perfused hearts: impact of an anti-ICAM-1 Ab and a lipoxygenase inhibitor. LTC4 (A), LTD4 (B), and LTE4 (C) in the organ perfusate were quantified by using solid-phase extraction and isocratic RP-HPLC with on-line photodiode array detection. Depicted are perfusate levels of cys-LTs after terminating the experiments designated in Fig. 2. Bars are means ± SE of at least 3 independent experiments. *P < 0.05 vs. all other groups.

View larger version (40K): [in this window] [in a new window]   Fig. 5.   Release of LTB4 and 5-hydroxyeicosatetraenoic acid (5-HETE) into the perfusate after application of PMN, fMLP, and AA in -toxin-perfused hearts: impact of an anti-ICAM-1 Ab and the lipoxygenase inhibitor MK-886. LTB4 (A) and 5-HETE (B) in the organ perfusate were quantified by solid-phase extraction and isocratic RP-HPLC with online photodiode array detection. Bars represent levels of LTB4 and 5-HETE at the end of the experiments shown in Fig. 2. Means ± SE of at least 3 independent experiments are given. *P < 0.05 vs. all other groups.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present study demonstrates that low doses of -toxin, a prominent pathogenicity factor in local and systemic staphylococcal infections, induce upregulation of the endothelial ligand ICAM-1 with subsequent accumulation of PMN in isolated, perfused rat hearts. In the presence of the neutrophil ligand fMLP and the leukotriene precursor AA, substantial cys-LT synthesis is then noted, linked with coronary vasoconstriction and loss of myocardial performance. Transcellular cys-LT synthesis in the toxin-challenged coronary vasculature is suggested as the underlying mechanism. Because bacterial exotoxins are synthesized by a large number of clinically relevant bacteria (5), these observations may be relevant for myocardial dysfunction in sepsis and septic shock.

In previous reports from our laboratory (16, 40), we demonstrated that -toxin may provoke a marked coronary vasoconstrictor response accompanied by severe depression of myocardial performance in the absence of any further agent. In addition to this, we now demonstrate that a fourfold lower dose of -toxin, which fails to provoke significant alterations of cardiac performance per se, induces upregulation of the endothelial ligand ICAM-1 in the coronary vasculature as detected by immunohistochemical imaging and quantification of mRNA. Because no endotoxin was detectable in the perfusate by a Limulus-based LPS-assay and no expression of ICAM-1 was found in sham-perfused hearts, the ICAM-1 upregulation was clearly attributable to the staphylococcal toxin and not to putatively contaminating LPS. This view is further supported by the fact that enhanced PMN accumulation, an indicator of adhesion molecule expression, was only noted in the presence of -toxin. Although these findings clearly suggest staphylococcal -toxin to be a potent inducer of coronary ICAM-1 expression, the underlying signaling events are largely unknown. Transmembrane pore formation resulting from either binding of the toxin to receptors or from nonspecific absorption to the cell membrane of its target cells has been established as the basic toxin mechanism in the toxin concentrations used in the present study, and subsequent intracellular events may largely be linked to pore-related transmembrane electrolyte fluxes (5). In a previous study, hydrolysis of phosphatidylinositol was demonstrated in endothelial cells challenged with -toxin (21), and inositol phosphates may initiate a variety of downstream signaling events such as activation of protein kinase C and translocation of nuclear factor-B. However, it remains to be elucidated whether such sequelae are also operative in the upregulation of ICAM-1-expression by -toxin in our experimental setup.

Firm adhesion of PMN to activated endothelium is believed to proceed largely via binding of neutrophil CD11/CD18 to ICAM-1. In animal models of cardiac ischemia and reperfusion ICAM-1-dependent adherence of PMN greatly contributes to coronary endothelial dysfunction and myocardial tissue necrosis (29, 36). Therefore, to evaluate the functional consequences of toxin-induced ICAM-1 expression, we enriched the perfusate with PMN. Histologically detectable accumulation of neutrophils and enhanced myocardial MPO activity was then noted in -toxin-treated hearts, indicating retention of PMN in the coronary vasculature. The fact that in the presence of the anti-ICAM-1 Ab the increase in MPO activity was completely abrogated whereas the increase was not affected by a nonspecific antibody strongly suggests the functional significance of toxin-induced expression of ICAM-1 for neutrophil retention in our model. To some minor extent PMN retention was also noted in sham-perfused hearts and in toxin-exposed hearts treated with the anti-ICAM-1 Ab. These observations may be explained by 1) the fact that ICAM-1 may be constitutively expressed on quiescent endothelium to a small extent (2) and this basal expression may have escaped immunohistochemical detection, and 2) the contribution of endothelial ligands other than ICAM-1 being involved in PMN-endothelial cell interaction in the coronary vasculature.

The pathophysiological significance of neutrophil accumulation in -toxin-treated hearts became evident when the perfusate was enriched with the eicosanoid precursor fatty acid AA and challenged with the neutrophil ligand fMLP. fMLP is a bacterial pathogenicity factor representative of a variety of formylated peptides synthesized by all bacteria (31, 39) and is known to stimulate different neutrophil functions, among others, the 5-LO pathway. However, although fMLP sufficiently activates neutrophil 5-LO, exogenous supply of its substrate AA is mandatory because of the inability of fMLP to activate neutrophil phospholipases for supply with endogenous AA for leukotriene synthesis (e.g., LTA₄, LTB₄, and 5-HETE; Refs. 9, 18). It is noteworthy that the concentrations of AA currently used do not exceed those known to arise at sites of inflammatory events in vivo (23, 41), and recent studies in patients have shown that free plasmic AA concentrations may even exceed 100 µM under conditions of severe sepsis and septic shock, thus surpassing the currently used concentration of this precursor fatty acid by more than one order of magnitude (32). In -toxin-treated hearts, but not in control hearts, a marked increase in CPP, accompanied by a decrease of left ventricular contractile function, was noted on admixture of PMN, fMLP, and AA. Most likely, these functional abnormalities were attributable to the generation of LTs, in particular cys-LTs, under these conditions. First, next to the presence of -toxin, the admixture of neutrophils was a prerequisite for the appearance of the functional abnormalities. Second, when cardiac neutrophil accumulation was blocked by an anti-ICAM-1 Ab, neither synthesis of cys-LTs nor coronary vasoconstriction and loss of contractile function occurred. Third, the functional abnormalities were fully prevented when the 5-LO metabolism was blocked by the specific inhibitor MK-886. These findings are in line with previous studies demonstrating corresponding physiological effects (i.e., rise in CPP, impairment of contractile function) on infusion or generation of cys-LTs in the coronary vasculature of isolated hearts (36-38), and in models of cardiac ischemia and reperfusion a significant cardioprotection was demonstrated when LT bioactivity was blocked (28, 36).

Cumulative evidence suggests transcellular (cooperative) eicosanoid synthesis as the most likely metabolic pathway of cys-LT formation in our experimental setup (30). Because PMN are devoid of glutathione-S-transferase, they may cooperate with neighboring cells to produce cys-LTs; by releasing the very unstable intermediate LTA₄ into their extracellular microenvironment, this intermediate becomes available to adjacent acceptor cells, e.g., endothelial cells (11, 20), which convert LTA₄ to cys-LTs. Our finding that synthesis of cys-LTs was exclusively found when PMN were retained in the coronary vasculature is in favor of this assumption, because only adherent PMN are in sufficiently close contact to endothelial cells to permit transfer of the very short-lived intermediate LTA₄ from the donor to the acceptor cell. In the absence of PMN, no release of cys-LTs was detected, excluding the myocardium itself as major source of cys-LTs under the current experimental conditions. Another noteworthy aspect of the current study refers to the synthesis of LTB₄ and 5-HETE. Because both metabolites are well-known chemoattractants for neutrophils, their liberation may maintain PMN-mediated cardiac dysfunction by triggering further cycles of neutrophil adhesion and activation. Although neutrophils are per se capable of generating LTB₄ and 5-HETE when appropriately activated, intercellular cooperation may further enhance this type of lipoxygenase product formation, e.g., by endothelium-to-neutrophil AA transfer.

Interestingly, the appearance of LTs in the toxin- and neutrophil-challenged coronary vasculature was linked with both vasoconstrictor response and loss of myocardial performance. Because LTs apparently do not directly suppress heart muscle contractile function (37), the most reasonable explanation is a maldistribution of regional perfusion caused by the vasoconstrictor potency of the cys-LTs. Such a phenomenon, characterized by the coexistence of under- and overperfused capillaries in close vicinity, has long been considered as a pathogenetic mechanism in septic heart failure (7, 12, 22). Interestingly, similar abnormalities were also observed in rat hearts undergoing high-dose -toxin challenge, with thromboxane being implicated as the predominantly responsible vasoconstrictor agent under these conditions. Alternatively, the generation of reactive oxygen species by neutrophils retained in the coronary microcirculation might account for a portion of the observed cardiodepression. The fact that this cardiodepression was abrogated by the 5-LO inhibitor MK-886 does not refute this latter assumption, as neutrophil LTB₄ synthesis may activate the generation of oxygen species in a paracrine fashion (19).

It is a matter of speculation whether the presently described pathogenetic sequelae may be relevant for the appearance of cardiac dysfunction in human sepsis. There are, to our knowledge, no studies in humans directly addressing the contribution of activated neutrophils to this type of organ failure. Experimental data in endotoxemic animals do, however, support the view that neutrophils may be linked with the loss of heart contractile performance in sepsis (3, 14, 17), a finding well in line with the present observations.

In conclusion, low doses of staphylococcal -toxin may promote cardiac accumulation of PMN by upregulating the endothelial ligand ICAM-1. Under appropriate conditions, this may result in coronary vasoconstriction and depression of left ventricular contractile function. LT generation, in particular transcellular synthesis of cys-LTs, is suggested as the predominant underlying mechanism. A role of both bacterial exotoxins and activated PMN in septic heart failure must thus be considered.