INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

IT IS WELL DOCUMENTED that endotoxin, nontoxic derivatives of endotoxin, and endotoxin-induced cytokines can protect against ischemia-reperfusion (I/R) injury in the heart (5, 6, 10, 28). Administration of endotoxin (28) or cytokines (6) or induction of sepsis with gram-negative bacteria (25) has been shown to protect the heart from I/R injury 24 h later. Endotoxin, or lipopolysaccharide (LPS), causes an increased myocardial synthesis of proteins, e.g., heat shock proteins and the antioxidant catalase, that may be involved in protection (5, 9). Sepsis causes an increase in nitric oxide (NO) production (via the inducible NO synthase), which can inhibit polymorphonuclear neutrophil (PMN) adhesion and infiltration (16). All of these mechanisms may play a part in ultimately protecting the endothelium and, in turn, the myocardium from injury as a result of a secondary ischemic insult. The focus of past studies on the protective mechanisms of these factors has been limited to intracardiac mechanisms (5, 9). The objective of this study was to determine whether extracardiac mechanisms could also be manipulated to attenuate I/R injury.

Several studies have demonstrated that LPS causes a rapid reduction in leukocyte-endothelium interaction (13, 24, 31). This interaction is the first step required for PMN infiltration into tissues from the vascular space. Both integrins and L-selectin on the PMNs, as well as their endothelial counter receptors (intercellular adhesion molecule-1 and E-selectin, respectively), are important for PMN adhesion to the endothelium. Endotoxin has been shown to cause profound alterations in PMN surface expression of adhesion molecules, which directly affect binding kinetics. This suggests that LPS may modulate PMN-mediated I/R injury (8, 18, 24, 40). Therefore, we determined whether PMNs obtained from a rat administered a single nonlethal dose of LPS have an impaired ability to induce PMN-mediated I/R injury in the isolated perfused heart compared with PMNs from saline-treated control rats. This is based on the hypothesis that PMNs from LPS-treated rats induce less PMN-mediated I/R injury in the isolated perfused heart as a result of reduced PMN adhesion to the endothelium and extravasation into the myocardium.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

To study PMN-mediated injury, PMNs were obtained from one animal and infused through the coronary circulation of an isolated heart obtained from another animal. The present study had two objectives: to determine the effects of PMNs from LPS-treated rats on PMN-mediated IR injury and to correlate them with the in vivo effects of LPS on PMNs.

Animal Preparation

Rats used for PMN isolation and for characterization of a nonlethal model of endotoxemia were anesthetized with ketamine-xylazine (6:0.6 mg/100 g body wt im). Under aseptic conditions, polyethylene catheters (PE-50) were implanted in the carotid artery and the jugular vein. Rats were allowed 24 h to recover. Rats were given either a nonlethal bolus (100 µg/kg) of LPS (Seratype 0111:B4; Sigma) or sterile saline. PMNs were isolated from these rats 7 h later.

After LPS administration, the number of PMNs, blood pressure (BP), heart rate (HR), and body temperature were measured every hour for 8 h in a subset of seven animals. Total white blood cell (WBC) and differential counts were determined on arterial blood. BP and HR were measured with a P23 XL pressure transducer connected to a Grass physiograph. At 1 h after LPS treatment, neutropenia (0.97 ± 0.27 × 10⁶ PMNs/ml blood) was observed followed by neutrophilia (10.48 ± 1.21 × 10⁶ PMNs/ml blood) by 7 h. The cardiovascular response to LPS included a decline in mean BP after the first hour (58 ± 5 mmHg) compared with the initial level (80 ± 7 mmHg). From 2 to 8 h, BP was similar to control values, and HR, after an initial increase, recovered to pre-LPS values (442 ± 21 beats/min) by 7 h.

Neutrophil isolation. Rats were anesthetized with pentobarbital sodium (45.4 mg/kg iv). Blood was collected with EDTA (15 mg EDTA, Becton Dickinson), and PMNs were separated from mononuclear cells and red blood cells (RBCs) with a One-Step Polymorph Nycoprep density gradient (Accurate Chemical & Scientific) as described by Bautista (4). Briefly, 5 ml of blood were layered over 4 ml of the Nycoprep gradient in sterile conical tubes and centrifuged at 1,600 rpm for 25 min at 18-20°C. The PMNs were suspended in 50 ml of Krebs-Henseleit bicarbonate (KHB) buffer (pH 7.4) and centrifuged at 200 rpm for 10 min; the pellet was resuspended in 50 ml of 0.14 M NH₄Cl for 10 min to lyse RBCs. After centrifugation, the final PMN pellet was resuspended in 1.0 ml of KHB buffer (pH 7.4). Cell viability was assessed by Trypan blue exclusion, and percent purity of PMNs was assessed by visualizing the nucleus with 3% acetic acid. Only preparations yielding greater than 19 × 10⁶ PMNs/rat with a viability of >97% were used in these studies. The purity was ~85% and the success rate was ~90%.

Myocardial Performance

Isolated heart preparation. Naive animals were anesthetized with pentobarbital sodium (64.8 mg/kg ip). Hearts were excised and placed in ice-cold KHB. Hearts were attached to the metal cannula of a water-jacketed Langendorff perfusion apparatus via the aorta, and perfusion was initiated with KHB buffer (pH 7.4) gassed with 95% O₂-5% CO₂ and warmed to 37°C. A compliant latex balloon tied to the end of a size 4-Fr dual lumen catheter was directed through the atrium, into the ventricle, and through the apex of the myocardial wall. A small piece of the balloon was tied to the outer surface of the heart, and the catheter was tied securely to the left atrial appendage. A Grass stimulator was used to pace the heart. A heated water jacket was placed around the heart, and an insulating foam cover was set on top of the chamber to maintain the chamber temperature at 37°C. A TXD-310 pressure transducer (Micro-Med) was attached to one lumen of the dual lumen ventricular catheter, and maximal ventricular pressure, left ventricular end-diastolic pressure (LVEDP), and the measurement of maximum rate of left ventricular pressure rise and fall during a cardiac cycle (+ and dP/dt_max) were recorded by a heart performance analyzer (HPA model 100, Micro-Med). Left ventricular developed pressure (LVDP) was calculated as the difference between systolic and diastolic pressure. The balloon was filled with warm water (~100 µl) until ventricular diastolic pressure was 5-10 mmHg. The height of the buffer in the Langendorff column was maintained at a level to deliver coronary flow (CF) at a perfusion pressure of 80-83 mmHg. CF was measured by collecting coronary effluent for 30 s as it dripped from the heart (25).

I/R protocols. Hearts stabilized for 15 min at an initial LVEDP of 5-10 mmHg and CF of ~14 ml/min. After control measurements were taken, a 1-ml KHB suspension of the treatment (if PMNs, complement, and/or plasma were used, then one-tenth of posttreatment amounts were used) was injected into the heart over a 10- to 15-s period. The heart was immediately made ischemic, and the water was removed from the balloon to maximize myocardial damage upon reperfusion (15). After 20 min of ischemia, the same volume of water was reintroduced into the balloon, and the heart was reperfused. During the first 5 min of reperfusion, treatment was infused into the heart at a rate of 1.1 ml/min. CF was measured at 30-s intervals, left ventricular pressures were recorded continuously, and the heart was paced at the preischemic rate. Reperfusion was maintained for another 60 min with KHB alone (total reperfusion = 65 min). Throughout the last 60 min of KHB reperfusion, ventricular pressure and CF were measured every 5 min for the first 40 min and every 10 min for the last 20 min. In some experiments, samples of CF were taken at 20, 40, and 60 min of KHB reperfusion for measurement of lactate dehydrogenase (LDH). At the end of reperfusion, water was removed from the balloon. Half of the hearts (n = 4) per group were frozen with metal tongs precooled in liquid nitrogen and used for analysis of malondialdehyde (MDA), creatine phosphate (CrP), and ATP. The other hearts (n = 4 per group) were dropped into Z-fix (Anatech, Battle Creek, MI) for histology.

Sham ischemia protocols. Hearts stabilized for 15 min at an initial LVEDP of 5-10 mmHg and CF of ~14 ml/min. Control measurements were taken, and then a 1-ml KHB treatment suspension was injected into the heart over a 10- to 15-s period. The water was removed from the balloon, and perfusion of the heart was maintained for 20 min (sham ischemia). The same volume of water was reintroduced into the balloon, and, over a 5-min interval, the treatment was infused into the heart, and CF was measured at 30-s intervals. Left ventricular pressure was recorded continuously, and the heart was paced at the preischemic rate. Perfusion of the heart was maintained for an additional 60 min with KHB buffer alone. Throughout the last 60 min of KHB perfusion, measurements of myocardial performance were taken every 5 min for the first 40 min and every 10 min for the last 20 min.

Plasma isolation. To determine whether guinea pig complement had the same effect on PMNs as did rat plasma, plasma was prepared. Anesthetized rats were catheterized and rapidly exsanguinated by drawing out 14-16 ml of blood into heparinized syringes. Blood was centrifuged at 3,000 rpm for 25 min at 4°C; the plasma was centrifuged again at 3,000 rpm for 15 min and then filtered to remove any aspirated platelets. Fresh plasma was obtained on the day of the experiment and stored on ice until use.

Biochemical Analyses

Lactate dehydrogenase. One of the early markers for sarcolemmal membrane disruption and cell death is the release of the soluble cytosolic protein LDH. Coronary effluent samples obtained at 20, 40, and 60 min of KHB reperfusion were assayed for LDH activity using an automated digital Seralyzer instrument (AMES).

Heart analysis. Frozen hearts were pulverized with a liquid nitrogen-cooled mortar and pestle. Aliquots of tissue were used for wet-to-dry ratio and for ATP, CrP, and MDA assays. Lipid peroxide formation was determined by the thiobarbituric acid method for estimation of MDA content (30). ATP and CrP were determined on a neutralized perchloric acid extract of heart tissue by fluorometric methods (22). The remaining frozen powdered tissue was weighed, dried overnight in a 60°C oven, and weighed again for calculation of the wet-to-dry weight ratio. This ratio was used to correct ATP, CrP, and MDA levels to the dry weight of the heart.

Histology

Characterization of Neutrophils

Basal and PMA-induced production of superoxide anion by isolated rat PMNs were determined by super oxide dismutase-inhibitable reduction of cytochrome c (expressed as nmol/ml × 10⁶ cells) (1). Intracellular production of H₂O₂ was measured by the NBT test (2, 3). At least 100 PMNs were examined. Cells with stippled cytoplasmic deposits of blue-black formazan or dense clumps of formazan in the cytoplasm were counted as NBT positive (activated) and were expressed as a percentage of the total PMNs examined (10).

L-selectin homing receptors were measured on PMNs suspended in KHB buffer containing 0.5% bovine serum albumin and 1% paraformaldehyde (pH 7.4). Neutrophils were incubated with FITC-labeled mouse anti-rat L-selectin (CD62L) monoclonal antibody (PharMingen). Both the monoclonal antibody-labeled cells and the unlabeled PMNs were analyzed by flow cytometry using a fluorescence-activated cell sorter (Coulten Elite, Hialeah F1).

Statistical Analysis

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Myocardial Performance

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Left ventricular end-diastolic pressure (LVEDP) and left ventricular developed pressure (LVDP) in isolated hearts before ischemia (, n = 24) and during 60-min reperfusion after 20-min ischemia and 5-min reflow with Krebs only (no PMN/Comp; , n = 8) or complement and polymorphonuclear neutrophils (PMNs) from lipopolysaccharide (LPS)-treated (, n = 8) or saline-treated (, n = 8) rats. Time 0 represents discontinuation of treatment and beginning of reperfusion with Krebs alone. *P < 0.05, different from saline-treated controls. +Not different from the respective preischemic values.

At the end of the 60-min KHB reperfusion period, LVEDP was significantly increased from the preischemic level in all groups (Fig. 1). However, LVEDP was similar in the LPS-PMN and no PMN-Comp groups and was significantly elevated in the saline-PMN group throughout reperfusion. The left ventricular rate-pressure product (RPP) showed a similar trend to LVDP. RPP was lowest in the saline-PMN-treated hearts (50 ± 4% of preischemic, n = 8) and not different between LPS-PMN-treated hearts (79 ± 2% of preischemic, n = 8) and hearts in the no PMN-Comp group (83 ± 4% of preischemic, n = 8).

Both +dP/dt and dP/dt (Fig. 2), as percentage of the preischemic values, were lowest during KHB reperfusion in the saline-PMN-treated hearts. As with the other determinants of myocardial performance, there were no significant differences in either +dP/dt or dP/dt between hearts infused with LPS-PMNs and hearts in the no PMN-Comp group.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Positive (+) and negative () left ventricular pressure change over time during a cardiac cycle (dP/dt in mmHg/s) as a percentage of preischemic ±dP/dt during Krebs-Henseleit bicarbonate buffer (KHB) reperfusion after ischemia and 5-min reflow with Krebs only (no PMN/Comp; , n = 8) or complement and PMNs from LPS-treated (, n = 8) or saline-treated (, n = 8) rats. *P < 0.05, different from saline-treated controls. All values during KHB reperfusion were statistically different from their respective preischemic values in all groups.

Differences in CF were observed during the 5 min of PMN (or vehicle) infusion among the three groups (from 5 to 0 min on Fig. 3, expanded in inset). The CF in the no PMN-Comp group rapidly returned to preischemic levels (14.9 ± 0.4 ml/min, n = 24) by the end of the vehicle infusion. The CF of hearts in the saline-PMN group remained approximately one-third of the preischemic value throughout the 5 min of PMN infusion (5 to 0 min on Fig. 3). During the 5 min of PMN infusion in LPS-PMN hearts, CF gradually increased to approximately two-thirds of the preischemic value (9.7 ± 1.2 ml/min). Within 30 s of the completed PMN infusion (t = 0), CF rose to 12.1 ± 1.1 ml/min. CF decreased from the initial rebound during the following 60 min in the no PMN-Comp group and was significantly less than the preischemic value by 10 min. At the onset of posttreatment KHB reperfusion in the saline-PMN hearts, CF remained depressed, but by 5 min of KHB perfusion CF had increased to approximately two-thirds of the preischemic value and remained at this level throughout reperfusion. CF was different from the preischemic value throughout the KHB reperfusion in the LPS-PMN group but was not different from CF in the no PMN-Comp group. At 60-min reperfusion, CF was 57.3 ± 3.0% of preischemic CF in the saline-PMN group and 74.2 ± 3.2% and 82.2 ± 1.8% in the LPS-PMN group and the no PMN-Comp group, respectively.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Coronary flow (CF) in isolated hearts before ischemia (, n = 24) and during 5-min reflow (5 to 0 min) with Krebs only (no PMN/Comp; , n = 8) or complement and PMNs from LPS-treated (, n = 8) or saline-treated (, n = 8) rats and during 60 min with KHB reperfusion. Inset: CF during the PMN (or vehicle) infusion, i.e., from 5 to 0 min, and the first 5 min of KHB reperfusion (0-5 min) in an expanded form. *P < 0.05, different from saline-treated controls. +P < 0.05, different from both saline-treated controls and LPS-treated. ++Values not different from the preischemic values.

LDH was detected in the coronary effluent of five out of eight saline-PMN control hearts at 20, 40, or 60 min of post-PMN reperfusion (Fig. 4). In contrast, LDH was not detected in the coronary effluent of any LPS-PMN rats (n = 5).

View larger version (7K): [in this window] [in a new window]   Fig. 4.   Lactate dehydrogenase (LDH) content in coronary effluent from hearts infused with either saline-PMNs or LPS-PMNs. LDH levels were measured in coronary effluent collected at 20, 40, and 60 min KHB reperfusion, and the highest value measured at any of those time points is plotted.

Concentrations of ATP and CrP (which normally are ~18 and 24 µmol/g dry wt, respectively, in nonischemic perfused hearts) were not different between the saline-PMN and LPS-PMN hearts at the two time points measured during reperfusion (Table 1). At 10 and 65 min of reperfusion, MDA content was similar in both groups (Table 1). MDA content was significantly reduced at 65 min of reperfusion compared with the 10-min level, but again there was no difference between the groups.

Three additional studies were performed to validate the PMN-mediated I/R injury: 1) hearts were infused with plasma instead of complement (in the presence or absence of PMNs) during reperfusion; 2) ischemic hearts were infused with PMNs only during the first 5 min of reperfusion after ischemia; and 3) time-matched sham-ischemic hearts were infused with complement and PMN or complement alone.

Because the preparation of plasma would have required the used of another animal for each heart and PMN preparation, we tested the ability of the guinea pig complement to activate PMNs. When rat plasma was used instead of complement to activate PMNs, LVDP was depressed to the same extent at the end of reperfusion as in hearts in the complement saline-PMN group (Table 2). Infusing plasma alone caused no adverse effects on myocardial performance after ischemia compared with infusing hearts with KHB alone (81 ± 7 and 79 ± 3 mmHg, respectively). At the end of reperfusion, CF, as a percentage of preischemic flow, was depressed to the same extent in hearts infused with plasma and PMNs as with complement and PMNs (60.4 ± 4.4%, n = 4, and 57.3 ± 3.0%, n = 8, respectively). In hearts infused with plasma alone, the recovery of CF was 82.1 ± 6.1%.

The infusion of complement and saline-PMNs subsequent to, but not prior to, 20 min of ischemia induced a significant depression in LVDP compared with ischemic hearts infused with KHB alone (68 ± 3 vs. 79 ± 3 mmHg, respectively). However, throughout 60 min of posttreatment KHB reperfusion, LVDP in hearts that received PMNs only after ischemia was less severely depressed than was LVDP in hearts infused with PMNs and complement both prior to and subsequent to ischemia (Table 2). Moreover, at the end of the 5-min PMN infusion period, both LVDP and CF in hearts that received PMNs only after ischemia were less severely depressed than in hearts infused with PMNs before and after ischemia.

Time-matched sham ischemic hearts were infused with four different treatment regimes to determine whether PMNs and/or complement could induce damage to nonischemic myocardium. Hearts were infused with 1) KHB alone, 2) complement alone, 3) complement and saline-PMNs, or 4) complement and LPS-PMNs. In all four groups myocardial performance and CF were similar to the initial values after 60 min of posttreatment KHB perfusion (Table 3). However, during the 5-min infusion period, LVDP rapidly fell in all groups except the KHB-alone group. The decline in LVDP was primarily due to a fall in LVPSP because LVEDP did not change. Immediately after treatment, LVDP rebounded toward the initial values with full recovery by 5 min of KHB posttreatment perfusion. Changes in ±dP/dt and RPP reflected the changes in LVDP (data not shown).

The CF did not change in the KHB-alone group (Table 3). CF rapidly fell in hearts treated with complement alone, with saline-PMNs, or with LPS-PMNs during the 5-min infusion period. Within 5 min of KHB posttreatment perfusion, CF returned to initial values in all groups. By 60 min, there were no significant differences among the groups.

PMN Infiltration

To histologically quantitate the PMNs retained within hearts, PMNs were counted at five levels of the hearts from three to four serial sections per heart level. The number of PMNs counted within a heart perfused with complement and LPS-PMNs (46 ± 6 PMNs/level of the heart; n = 5) was significantly lower than the number obtained from a heart perfused with complement and saline-PMNs (224 ± 16 PMNs/level of the heart; n = 5) (Fig. 5).

View larger version (18K): [in this window] [in a new window]   Fig. 5.   The average number of PMNs per cross-sectional area in the vascular space, the tissue space, and total in hearts from LPS or saline-treated animals.

Effects of LPS on Neutrophils

View larger version (31K): [in this window] [in a new window]   Fig. 6.   Basal and phorbol 12-myristate 13-acetate (PMA)-induced superoxide anion production by PMNs isolated from saline-treated (n = 6) and LPS-treated (n = 5) rats.

Neutrophils obtained 7 h post-LPS showed a significant increase in the percentage of activated PMNs as assessed by oxidation of formazan crystals by endogenous H₂O₂ production. The LPS-PMN hearts were NBT positive at 70.5 ± 3.6% (n = 6), whereas 28.9 ± 3.5% (n = 5) of the saline-PMN hearts were NBT positive.

To investigate the potential role of LPS-induced alterations of L-selectin homing receptors, cell surface expression of this adhesion molecule was assessed by FITC-labeled monoclonal rat anti-L-selectin (FITC-CD62L monoclonal antibody) binding to PMNs. The percentage of PMNs positive for FITC-CD62L monoclonal antibody decreased by half in PMNs isolated 7 h after LPS was administered (Fig. 7A). Detectable fluorescence intensity of FITC-CD62L monoclonal antibody binding to saline-PMN hearts was 53.1 ± 8.6% compared with 27.9 ± 4.1% in the LPS-PMN hearts (Fig. 6A). A representative flow cytometric analysis of the FITC-CD62L monoclonal antibody binding to LPS-PMN hearts and saline-PMN hearts is shown in Fig. 7B.

View larger version (20K): [in this window] [in a new window]   Fig. 7.   Cell surface expression of the L-selectin Homing Receptor (CD62L) adhesion molecule detected by a CD62L-specific FITC-monoclonal antibody on isolated PMNs from LPS- or saline vehicle-treated rats. A: percentage of PMNs positive for FITC anti-L-selectin binding. *P < 0.05, different from saline-treated control. B: representative flow cytometric analysis of CD62L-specific FITC-monoclonal antibody binding to isolated PMNs. Representative histograms of CD62L-specific FITC-monoclonal antibody binding (B) to PMNs from an LPS-treated rat (L) and a saline control-treated rat (S) are shown. Approximately 5,000 cells were gated. The y-axis is the cell count × 100. The x-axis is fluorescence units.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Neutrophils obtained from LPS-treated rats did not exacerbate I/R injury as did PMNs obtained from the saline-treated control rats. The PMN-mediated injury was a function of both a decrease in LVPSP and an elevation in LVEDP. In addition, dP/dt_max, an index of rate of relaxation, was significantly lower in saline-PMN hearts, possibly due to impaired removal of calcium from the cytosol (14). Postischemic LVEDP was similar to levels seen in other ischemic models in which myocardial depression is linked with calcium loading and associated cytologic abnormalities (19). In contrast, when ischemic hearts were infused with LPS-PMNs, there was no evidence of additional damage due to PMNs. Postischemic LVEDP and dP/dt of hearts infused with LPS-PMNs or without PMNs or complement was similar. Moreover, LDH was detected in the coronary effluent of saline-PMN hearts but not in LPS-PMN hearts. LDH is an early indicator of cell damage, and in hearts losing LDH, persistent damage is occurring well into reperfusion. The PMNs from LPS-treated rats apparently did not exacerbate reperfusion injury of the heart.

The greater depression of postischemic function in saline-PMN hearts did not appear to be linked with high-energy phosphate levels (ATP and CrP). ATP and CrP levels were similar in the two PMN-treated groups at 10 and 65 min of reperfusion, suggesting that high-energy phosphates were not limiting ventricular recovery more in the saline-PMN group than in the LPS-PMN group.

There were significant differences in recovery of CF. In the first 5 min of reperfusion (concurrent with PMN and complement infusion), CF was most severely depressed in the saline-PMN group. Throughout posttreatment reperfusion, CF remained depressed (~60% of the preischemic value) in saline-PMN hearts but recovered to ~80% in LPS-PMN hearts and in the no PMN-Comp hearts. The similarities in recovery between these two groups and the lack of LDH release from hearts perfused with PMNs from LPS-treated rats suggest that depressed myocardial performance was a result of ischemic injury itself and/or PMN-independent reperfusion injury.

The depressed postischemic performance of hearts infused with saline-PMNs may have occurred, in part, as a result of a decreased CF, which may have resulted from a depressed endothelium-dependent relaxation (34-36). Neutrophils produce toxic oxygen products (e.g., hydroxyl radicals, superoxide anions, hypochlorous acid) and release cytotoxic proteases (e.g., elastase) and lipases, which can damage myocytes and depress function (23). Early PMN-mediated damage to the endothelium and PMN aggregates may have caused areas of vascular plugging during reperfusion (13, 14).

To date, most of the in vitro studies used to investigate PMN-mediated myocardial injury have used hearts from small animals (e.g., rats and rabbits) and PMNs from humans (32, 35). To avoid potential problems inherent in mixing tissues from different species, we applied an improved method for the isolation of PMNs to use syngeneic PMNs in an in vitro model of PMN-mediated I/R injury. The use of rat PMNs in the rat heart optimized PMN-endothelium interactions by decreasing the variability that might occur as a consequence of using tissues from different species and allowed us to study the effects of in vivo administration of LPS on PMN-mediated I/R injury of the myocardium.

To reproduce the in vivo situation, we injected the complement and PMNs before ischemia so that they would be trapped within the coronary circulation during the ischemic period. The reperfusion injury was less (but still present) in hearts treated with PMNs and complement only subsequent to ischemia compared with hearts treated with PMNs both before and after ischemia. During ischemia, PMNs trapped within the coronary circulation with complement may have become activated and, under static or no-flow conditions, may have adhered, extravasated, and induced injury. Because PMNs interact with the vascular endothelium by way of a CD11b/CD18-intercellular adhesion molecule-1 interaction, which has been shown to generate an endothelium-dependent vasoconstriction (26), trapped PMNs may have enhanced "no reflow" upon reperfusion or may have increased transit time for the postischemically infused PMNs.

To elucidate the mechanism(s) that contributed to the change in PMNs from LPS-treated rats, several facets of PMN function were assessed. Superoxide production by LPS-PMNs was not different from saline-PMNs under basal conditions or when stimulated with PMA. Therefore, LPS did not alter the intracellular pathways by which PMA activates protein kinase C and stimulates superoxide production.

We next tested the degree of activation of the PMNs. The NBT test specifically assesses the ability of PMNs to endocytose foreign particles and produce superoxide anion. A significantly greater percentage of LPS-PMNs were primed as determined by H₂O₂ (product of the superoxide dismutase reaction) reduction of NBT. Studies using NBT and LPS show that NBT is a good indicator of phagocytic activity (2). Our findings agree with those of Eising et al. (10) that despite the greater degree of "activation" or "priming," LPS-PMNs were still less likely to exacerbate I/R injury.

Cellular mechanisms associated with the ingestion of foreign particles are distinct from the receptor binding interactions required for PMNs to adhere to the endothelium, diapedese, and infiltrate the tissue (24). In the present model for PMN aggregation and adhesion, both lectin and integrin adhesion molecules are required sequentially (7). L-selectin receptors, are maximally expressed on quiescent or inactive PMNs and are fully active (18). The number of L-selectin receptors declines within 3-10 min of LPS administration in vivo (18). The CD11/CD18 adhesion molecules are in an inactive conformation until PMNs are activated, and then only a small subpopulation will change to an active state (8).

Inhibition of PMN aggregation after exposure to LPS suggests that LPS may affect a balance between aggregation and adhesion by causing PMNs to shed L-selectin receptors (24). As we had hypothesized, the L-selectin density on PMNs from LPS-treated rats was lower (by 51%) than on PMNs from saline-treated control rats. Studies that have compared CD11b/CD18 and L-selectin densities versus PMN aggregation have shown that PMN aggregation decreases as L-selectin surface expression decreases irrespective of the CD11b/CD18 density (13, 24, 31). To test this, we determined the percentage of infused PMNs that were retained within the reperfused heart, and we assessed PMN infiltration in tissue slices from hearts. There was a threefold greater number of saline-PMNs (75% of infused) retained within the I/R hearts compared with LPS-PMNs (25% of infused). Histology confirmed this observation suggesting that LPS-treatment inhibited PMN-mediated I/R injury by inhibiting adherence of PMNs to the endothelium, resulting in reduced infiltration.

There are several possible explanations for the changes in L-selectin receptors on PMNs isolated during the neutrophilic stage of endotoxemia. These PMNs, which have been released en masse from the bone marrow, may be premature and have low numbers of L-selectin receptors or the receptors may have a reduced binding affinity. However, Jutila et al. (17), using the MEL-14 monoclonal antibody, have demonstrated that L-selectin density on PMNs from bone marrow is comparable to that of PMNs from peripheral blood. Bone marrow PMNs were fully capable of binding as determined by an in vitro adhesion assay of PMN binding to high endothelium in inflamed lymph node venules (19, 16).

Tumor necrosis factor-, leukotriene B₄, and C5a (an anaphylatoxin and PMN activator/chemoattractant), all expressed following in vivo endotoxin exposure, are potent inducers of L-selectin shedding (18). In addition, LPS complexes with LPS binding protein (LBP) to activate PMNs via the CD14 receptor, resulting in both upregulation of CDllb/CD18 and shedding of L-selectin by a receptor-mediated process (39, 38, 24).

There is potential clinical relevance in using endotoxin or, more likely, a nontoxic derivative of endotoxin, as a treatment to attenuate PMN-induced tissue reperfusion injury. Under in vivo conditions, the ability of LPS to attenuate PMN-mediated I/R injury may be potentiated. One of the primary counter receptors for L-selectin is the sialyl-Lewis^X-specific E-selectin (epithelial-leukocyte adhesion molecule-1) on the endothelium (27). This receptor is initially upregulated by cytokine stimulation in response to sepsis or endotoxin, and then subsequently there is a decline in its surface activity (29). From 4 to 24 h after in vivo LPS administration, there is a linear increase in plasma levels of soluble E-selectin, which parallels the decline in its surface activity on the endothelium (22). Moreover, it has been shown that intravenous administration of soluble E-selectin inhibits PMN emigration by 64% at 6 h after an LPS challenge, suggesting that endogenous soluble E-selectin shed from activated endothelium may further deter PMNs from adhering in the coronary circulation and inducing PMN-mediated I/R injury (37). Finally, as with the soluble E-selectin, serum-soluble L-selectin with functional activity has been isolated (33). Thus under in vivo conditions soluble L-selectin from PMNs and soluble E-selectin from endothelium are shed in response to LPS. These soluble receptors may act as competitive inhibitors to their respective membrane-bound counter receptors on PMNs and endothelium and inhibit adherence.

In summary, we have shown that in vivo endotoxin treatment attenuated PMN-induced reperfusion injury in the isolated heart. Hearts infused with LPS-PMNs had a higher postischemic LVDP and CF but no difference in ATP and CrP or MDA levels compared with hearts infused with saline-PMNs. Fewer PMNs from LPS-treated rats were retained within the myocardium. Finally, only half the numbers of PMNs from LPS-treated rats expressed L-selectin adhesion molecules compared with PMNs from saline-treated control rats. These data collectively suggest that the lack of PMN-mediated injury when PMNs from LPS-treated rats were infused into ischemic hearts may be due to reduced adherence of PMNs to coronary endothelium, in part, as a result of a reduced surface expression of the L-selectin receptor.