The role of the neutrophil and formed elements of the blood in an in vitro model of reperfusion injury

Using The globally ischaemic isolated guinea-pig heart we conducted studies to assess the role of activated neutrophils (PMNs) and the role of the endothelium in reperfusion injury. Reperfusion injury was induced by a 20 min period of global ischaemia followed by a 30 min reperfusion with Krebs' buffer supplemented with f-Met–Leu–Phe (fMLP) and heparinized blood. Ischaemia alone or blood alone resulted in a complete recovery in contractile function measured by developed pressure, fMLP (500 μM) and blood, administered to normoxic hearts did not affect contractile function. The combination of 100 μM fMLP and blood beginning at reperfusion and continuing for 30 min decreased the recovery in contractile function (max. 33 ± 6% reovery) while buffer and 100 pM fMLP resulted in a complete recovery in function. In hearts infused with buffer and neutropenic blood incubated with 100 μM fMLP a complete recovery in function was observed. Isolated peritoneal neutrophils, 7–70 × 105 PMN/ min, incubated with 100 μM fMLP and Krebs' solution decreased contractile function in a concentration-related manner (max. 44 ± 11% recovery). Platelets, plasma or red blood cells alone incubated with fMLP did not decrease recovery in developed pressure. Platelets and PMN incubated with 100 μM fMLP did not, while red blood cells and PMN did, elicit a reduction in recovery in contractile function (34 ± 4% recovery). A 20 min period of global ischaemia destroys the functional integrity of the endothelium (response to Ach). Pre-treatment of the heart with sufficient H2O2 to functionally damage the endothelium, followed by infusion of Krebs' solution supplemented with blood and 100 μM fMLP also elicited a reduction in recovery of contractile function (42 ± 15% recovery). In summary, partially activated neutrophils play a major role in reperfusion injury and there exists a cooperativity between the RBC and PMN in this model.


Introduction
The restoration of blood flow to the ischaemic myocardium via angioplasty, coronary bypass or fibrinolytic therapy has been shown to limit myocardial infarction and was thought to preserve ventricular function. However, at reperfusion the introduction of oxygen, cellular elements and, particularly, the neutrophil (PMN) into the ischaemic myocardium elicits a deleterious cascade of events which limit ventricular function and myocardial salvage. This phenomenon is called reperfusion injury. >4 Under normal conditions, the PMNs move along the vascular wall, but do not adhere. >7 However, the restoration of flow (reoxygenation) after an ischaemic period has been shown to cause an accumulation of PMNs in the ischaemic myocardium suggesting their role in reperfusion injury. 4'8 PMN contains a number of (C) 1993 Rapid Communications of Oxford Ltd substances capable of mediating vascular injury such as cationic proteins, acids, proteases which destroy vascular basement membranes, substances which induce release of PGI2 from the endothelium and promote PMN adhesion and PMN-derived elastase which lyses endothelial cells. The removal of PMNs via mechanical or physiological means [1][2][3][4][5][6][7][8][9][10][11][12] or inhibition of adhesion to the endothelium13'4 has been shown to decrease myocardial damage and preserve ventricular function. The present study was performed to assess the interaction of the formed elements of the blood and the role of the endothelium in an in vitro model of reperfusion injury.

Methods
Guinea-pigs of either sex (250-400g) were anaesthetized with 35 mg/kg pentobarbital i.p. and administered heparin (150U/kgi.p.) to inhibit microembolism in the coronary circulation. Hearts were quickly removed and mounted on a modified Langendorff apparatus and perfused with the Krebs-Henseleit buffer having a composition of 118.2 mM NaC1, 4.7 mM KC1, 2.2 mM CaC12, 1.2 mM HPO4, 1.2 mM MgSO4, 25 mM NaHCO3 and 11 mM D-glucose and oxygenated with O2-CO2 (95%:5%) at 37C. For assessment of developed pressure a fluid-filled balloon-tipped cannula was inserted into the left ventricle via the left atria and connected to a Gould pressure transducer (P23ID Gould, Oxnard, CA). After a 15 min equilibration period, end diastolic pressure was adjusted to 5 mmHg which was approximately 70-80% of the maximal stretch on the length pressure curve. Developed pressure was calculated as the difference between left ventricular systolic and diastolic pressures. Perfusion flow (Masterflex 7520, Barrington, IL) was adjusted to achieve a control mean perfusion pressure of 60mmHg (approximately 10ml/min). Therefore, a change in coronary resistance was detected by a change in perfusion pressure. Myocardial temperature was maintained at 37C by placing the heart in a heated chamber. The effects on developed pressure, heart rate and perfusion pressure were continuously displayed on a strip chart recorder (Grass model 7D).
Protocol: On completion of a 15 min stabilization period, the preparation was rendered globally anoxic for 20 min by shutting off flow to the system while maintaining the temperature at 37C. The hearts were then continuously reperfused with buffer (using the same flow as the control) and heparinized blood (0.1 ml/min) that had been pre-incubated with fMLP for 10 min at 37C. The effects on contractility, heart rate and perfusion pressure (coronary vascular resistance) were monitored for 30 min post-reperfusion. Measurement of oxidative burst activity: To determine the degree of neutrophil activation in response to fMLP stimulation, the following study was performed. Heparinized guinea-pig whole blood or blood diluted to a final concentration of 1% blood in 1 ml Krebs' buyer and warmed to 37C was studied. Luminol was added to a final concentration of 5/.tg/ml. Oxidative burst activity was measured in a continuously recording six-channel chemiluminometer (Biolumat LB9505, EG&G Berthold). After a 30 s stabilization period, fMLP was injected to induce the oxidative burst. The response was monitored for 5 min. The extent of the oxidative burst was quantified by integrating the area under the curve. Induction of neutropenia" To assess the role of the neutrophil in this model a group of guinea-pigs were rendered neutropenic by combining the methods of Bogman et al.lS and Henderson et al.16 Briefly, the animals were treated with 1.75 mg/kg mechlorethamine hydrochloride (Sigma) i.p. on day one followed by 1 mg/kg, i.p. on days two and three. To prevent infection, the animals were kept in a laminar flow unit under aseptic conditions. On the fourth day, the animals were anaesthetized and arterial blood was collected in a heparinized syringe (Sarstedt monovette; Li-heparin, Germany). Differential leukocyte determinations and neutrophil counts were performed using a Unopette kit (test 5856, Becton-Dickinson, Rutherford, N J) and a haemacytometer. Mechlorethamine treatment resuited in a marked reduction in leukocytes (9-15 x 106 cells/ml control vs. 1 x 106 0.09 x 106 cells/ml) with neutrophils preferentially suppressed by greater than 95% and mononuclear leukocytes by 50%.
Isolation of peritoneal neutrophils" To determine the role of the neutrophil and its interaction with the formed elements of the blood, the following isolations were performed. Guinea-pig peritoneal neutrophils were isolated as follows' briefly, 18 h prior to isolation, guinea-pigs were administered 6 ml of 6% Nacaseinate in saline i.p. (Sigma). Eighteen hours later, the guinea-pigs were euthanized and the cells recovered via peritoneal lavage. Viability (> 98%) of the preparation was determined via trypan blue exclusion.
Platelet isolation" Platelets were isolated from citrated guinea-pig concentrates using a gel filtration technique previously described by Ruggeri et al. 17 The final platelet plug was resuspended in Krebs' solution to a final concentration of 2 x 108 cells/ml. Red blood, cell isolation" Citrated guinea-pig blood (0.38% citrate) was centrifuged at 120 x g for 15 min. The plasma was removed and the red blood cell pellet was washed three times with normal saline and centrifuged at 950 x g for 15 min. The pellet was then resuspended in sufficient Krebs' solution to achieve a haematocrit of 42.
Endothelial function" Since myocardial ischaemia followed by reperfusion results in endothelial dysfunction characterized by a decreased release of endothelium-derived relaxant factor, we wished to determine the role of the endothelium in this model using the following protocol" on completion of a 15 min stabilization period, the hearts were paced at 4-5 V DC at 10% above the threshold rate. Acetylcholine (ACh) (0.5 tM) was infused for 1 rain to demonstrate endothelial functional integrity.
When the preparation stabilized, 0.5 #M sodium nitroprusside (SNP) was then administered for dilatory effects. Pacing was then halted and the heart rendered globally ischaemic for a period of 20 min. After 10 min of reperfusion (when all parameters returned to pre-ischaemic baselines) pacing was reinstituted and ACh and SNP administered as stated above.
HeOe induced injury: To determine the role of free radical-induced endothelial dysfunction, we performed the following protocol: on completion of a 15 min stabilization period, the hearts were paced at 4-5 V DC at 10% above the threshold rate. Acetylcholine (0.5/M) was infused for 1 min to demonstrate endothelial functional integrity. When the preparation stabilized, 0.5 #M SNP was administered for 1 min to assess endothelialindependent vasodilatory effects. Pacing was halted and the preparation returned to control values. H202 (0.1 mM) was infused into the Krebs' solution for 20 min. The hearts were then paced, as above, and the ACh treatment repeated. After pacing was halted and the preparation returned to control values, blood incubated with 100 #M fMLP for 10 min was infused into the Krebs' solution and the eflects on developed pressure, heart rate and perfusion pressure monitored.
Statistics" All data are expressed as the mean -+-S.E.
Data were analysed using analysis of variance and Dunnett's test or when appropriate Student's t-test.
Differences were considered significant atp < 0.05.

Results
Isolated hearts (n 6) subjected to a 20min period of ischaemia followed by a 30 min period of reperfusion did not have altered developed pressure, end diastolic pressure, heart rate or perfusion pressure values compared to the baseline control values of 59 -+-4 mmHg, 7 -+-2 mmHg, 218 _+ 8bpm and 61 -+-2mmHg, respectively.
Heparinized guinea-pig blood was infused, beginning at reperfusion and continuing for 30 min at 100/l/min and resulted in a complete recovery in developed pressure, while heart rate and perfusion pressure were unchanged, fMLP in Krebs' solution, administered to hearts not rendered ischaemic (n 5), at concentrations up to 500/M did not alter developed pressure, end diastolic pressure, heart rate and perfusion pressure from the baseline values of 57___5mmHg, 6_-q-2mmHg, 213+_5 bpm and 61 +__ 2 mmHg, respectively, fMLP (30 and 100 #M) (n-6 each) incubated in Krebs' solution for 10 min and infused at 100/l/min for 30min did not affect (98-+-13 and 123 + 22% recovery, respectively) recovery in contractile function. The combination of 3 and 10 M fMLP (n 5 each) incubated in blood for 10 min and administered at 100 #l/min, beginning at reperfusion and continuing for 30 min, did not affect (113 + 15 and 102 _--k 5% recovery, respectively) the recovery in contractile function, heart rate or perfusion pressure. Increasing concentrations of 30 and 100#M fMLP (n-7 each) significantly decreased the recovery in contractile function in a concentration-related manner (max. 33_ 6% recovery), increased perfusion pressure (52 q-8%) and did not affect heart rate. Increasing concentrations up to 300/M did not further decrease contractile function or increase perfusion pressure.
Heart rate remained unchanged (Table 1). respectively (Fig. 2). Heart rate was unchanged and perfusion pressure tended to increase (34 _+ 13%) from the baseline values of 203 6 bpm and 60-+-1 mmHg, respectively. Increasing the concentration of neutrophils above 70 105 cells/min caused precipitation in the syringe. Interaction offormed elements: To assess the role of the various blood components in this model, the following studies were conducted. In isolated hearts (n 5) subjected to a 20.min period of ischaemia followed by reperfusion with platelets (2 x 107 platelets/min) incubated with 100/M fMLP for 10min, and infused as a 1% platelet solution, recovery in developed pressure (84 _-q-8% recovery) from the baseline value of 70 q-5 mmHg was not altered. End diastolic pressure, heart rate and perfusion pressure were unchanged from the baseline values of 5 -+-2 mmHg, 221 + 25 bpm and 62 + 1 mmHg, respectively. In red blood cells, with the haematocrit adjusted to 42 with Krebs' solution and incubated with 100/M fMLP for 10 min prior to infusior, (n 6) recovery in developed pressure (109 q-23% recovery) was also not affected. Developed pressure, end diastolic pressure, heart rate and perfusion pressure were unchanged from the baseline values of 55 +/-3 mmHg, 5 +_ 2 mmHg, 229 10 bpm and 61 -+-1 mmHg, respectively.
Role of the endothelium" To assess the role of the endothelium the following two studies were conducted. In the first study, the functional integrity of the endothelium was assessed after a 20-min period of ischaemia. In the second study, the hearts were not rendered ischaemic, but the endothelium was functionally impaired via treatment with H202. In the first study, 0.5 #M ACh and 0.5 #M SNP administered to paced hearts (225 bpm) (n 6) prior to a 20 min period of global ischaemia, significantly decreased perfusion pressure from the baseline value of 60-+_ 1 mmHg by 21 + 6% and 25 _+ 4%, while developed pressure and end diastolic pressure were unchanged from the baseline values of 59---b_ 5 and 5-t-2mmHg, respectively. After a 10-min period of reperfusion (developed pressure returned to baseline 52-+-8 mmHg), 0.5 #M ACh failed to elicit the vasodilator response, while the response to SNP was unchanged from the baseline value of 64-+_ 4 mmHg.
In the second study, 0.5 zM ACh and 0.5/M SNP administered to paced hearts prior to H202 treatment significantly decreased perfusion pressure from the baseline value of 60-+-1 mmHg by 24 1% and 26 -+-3%, respectively, while developed pressure and end diastolic pressure were unchanged from the baseline values of 60 3 and 6 + 2mmHg, respectively. H202 (0.1 mM) for 20 min did not affect any of the parameters studied. Ach (0.5 zM) administered post-H20 2 failed to elicit the vasodilator response (61-+_ 3 to 64---F_ 5 mmHg) while the SNP response was unchanged (-21-+-4%). When the preparation stabilized (approximately 15 min) the hearts were assigned to one of the treatment groups. Blood alone or 100/zM fMLP alone (n 6 each), infused as a 1% solution for 30 min, did not affect any of the parameters studied. In contrast, 100/,M fMLP and blood significantly decreased developed pressure (42 15% recovery) (Fig. 4), while perfusion pressure increased by 30 11% and heart rate was unchanged. Baseline values for developed pressure, end diastolic pressure and perfusion pressure were 71 _ _ _ 7, 6 3 and 61 +__ 1 mmHg, respectively.

Discussion
The present study demonstrates that the neutrophil plays an essential role in an in vitro model of reperfusion injury. Secondly, it shows that neutrophils cooperate with other formed elements of the blood, to exacerbate reperfusion injury. The isolated guinea-pig model utilized in the present study meets all the necessary criteria required to study this phenomenon. Since the ischaemic insult by itself does not decrease recovery in myocardial contractility, the chemoactivator, fMLP (100/M), as well as blood alone, did not affect recovery in contractile function when administered to previously ischaemic hearts. However, after ischaemia the combination of fMLP and blood elicited a marked reduction in developed pressure indicating that some blood-borne element, when activated by fMLP, causes an extension of myocardial damage. Mullane and coworkers 18 using a bioassay technique demonstrated that previously infarcted rabbit hearts when stimulated with fMLP release LTD 4 and LTB 4. LTB 4 is a potent chemotactic agent that has been shown to amplify imflammatory responses and provoke PMN adhesion and degranulation. PMN degranulation results in further damage of the myocardium and increased PMN infiltration. 19'2 The neutrophii appears to be one of the major causative components of reperfusion injury. Numerous studies have shown that the PMN levels markedly rise in the heart when it is subjected to a period of coronary occlusion followed by reperfusion. 2'8'9 The PMN when fully activated (suflqcient to allow degranulation) releases a variety of substances such as proteolytic enzymes, oxygenderived free radicals, as well as substances which facilitate further activation and recruitment of circulating PMN to the injured tissue. 2'4'2 In addition, the removal of the PMN from the 90 Mediators of Inflammation. Vol 2.1993 preparation whether by antibody, 13'4 chemical 1 '2 or mechanical means has been shown to enhance recovery of function and lessen myocardial damage. In addition, inhibition of PMN adhesion to the damaged tissue has been shown to be effective. 22 '23 In the present study, neutropenic blood stimulated with fMLP also prevented reperfusion-induced myocardial dysfunction. However, many of these manoeuvres which reduce neutrophil count also decrease other formed elements of the blood such as platelets and monocytes. '15 To determine if the other formed elements of the blood were responsible for the reperfusion-induced damage we administered neutropenic blood supplemented with peritoneal PMN suflqcient to restore neutrophil count to normal. We found that the neutropenic blood supplemented with peritoneal PMN elicited a comparable reduction in contractile function as was observed when normal blood was used. These data therefore strongly suggest that the neutrophil is the major causative factor in reperfusion injury. Interestingly, when the concentration of PMN present in normal blood was incubated with fMLP and administered at reperfusion, a complete recovery in developed pressure was observed. These data strongly suggest that there exists a cooperation between the neutrophil and other formed elements of the blood. Similar  This extracellular transformation is sensitive to both heat and subject to proteolysis. 26 Platelets have also been shown to accumulate in the reperfused myocardium in a pattern similar to that of the neutrophil. In addition, PMN depletion using an antineutrophil serum has also been shown to reduce platelet deposition in the reperfused myocardium. 21'27 Several investigators have demonstrated an interaction between platelets and PMN in arachidonic acid metabolism. 25'28'29 They determined that platelet-derived arachidonic acid can be converted by the neutrophil to LTB 4. Mueller et al. 29 demonstrated that isolated rabbit peritoneal PMN incubated with A23187 synthesized and released PAF. We postulated that stimulation of the guinea-pig PMN with fMLP would, when the PMN degranulated, release PAF and stimulate platelets to produce mediators which would result in a synergy between the platelet and PMN. The present data does not support this hypothesis, since the combination of platelets and PMN incubated with fMLP did not appreciably decrease contractile function (80 _--k 9% recovery). However, these data do not preclude an interaction between the platelet and PMN and may reflect the use of fMLP as an agonist. Coeer et al. 25 studied the interaction of the PMN and platelet using zymosan as a chemoactivator and found a cooperation between the platelets and the PMN. Reperfusion injury, we feel, involves a sequential series of steps beginning with endothelial damage 3-33 followed by PMN rolling, activation firm adhesion and degranulation. To determine the role of endothelial damage in this model, we studied the effects of blood alone, fMLP alone and the combination of fMLP and blood in isolated hearts pre-treated with suflqcient H202 to functionally impair the endothelium. We found that blood alone or fMLP alone did not alter developed pressure while the administration of fMLP and blood markedly decreased contractile function. The magnitude of reducti,.on was similar to that observed when fMLP and blood was administered at reperfusion. In addition, a 20min period of ischaemia followed by reperfusion with Krebs' solution also functionally damaged the endothelium to the same magnitude as that observed with H202 treatment. It has previously been shown that H202 is a molecular species involved in endothelial injury and reperfusion injury. 34 HiOi-induced endothelial cell-dependent PMN adhesion was rapid having a short time to onset with the peak eect occurring within 20 min.
In summary, we have demonstrated that partially activated neutrophils play a major role in reperfusion injury. In addition, our data indicate that there exists a cooperativity between the RBC and PMN in this model. Lastly, these data support our working hypothesis that reperfusion injury is the result of an ischaemic insult causing a localized endothelial injury due to endothelial-derived oxygen-free radical production. This injury permits partially activated PMNs to adhere resulting in their release of proteolytic substances and chemoactivators causing exacerbation of the damage, as well as the recruitment of additional PMN to the already damaged tissue.