Skip to main content
NCBI home page
Annals of Surgery logoLink to Annals of Surgery
. 2001 Aug;234(2):233–244. doi: 10.1097/00000658-200108000-00014

Bactericidal/Permeability-Increasing Protein Attenuates Systemic Inflammation and Acute Lung Injury in Porcine Lower Limb Ischemia–Reperfusion Injury

Denis W Harkin *†, Aires A B Barros D’Sa *, Kevin McCallion †, Margaret Hoper †, M Isla Halliday † and, F Charles Campbell †
PMCID: PMC1422011  PMID: 11505070

Abstract

Objective

To investigate the role of recombinant bactericidal/permeability-increasing protein (rBPI21) in the attenuation of the sepsis syndrome and acute lung injury associated with lower limb ischemia–reperfusion (I/R) injury.

Summary Background Data

Gut-derived endotoxin has been implicated in the conversion of the sterile inflammatory response to a lethal sepsis syndrome after lower torso I/R injury. rBPI21 is a novel antiendotoxin therapy with proven benefit in sepsis.

Methods

Anesthetized ventilated swine underwent midline laparotomy and bilateral external iliac artery occlusion for 2 hours followed by 2.5 hours of reperfusion. Two groups (n = 6 per group) were randomized to receive, by intravenous infusion over 30 minutes, at the start of reperfusion, either thaumatin, a control-protein preparation, at 2 mg/kg body weight, or rBPI21 at 2 mg/kg body weight. A control group (n = 6) underwent laparotomy without further treatment and was administered thaumatin at 2 mg/kg body weight after 2 hours of anesthesia. Blood from a carotid artery cannula was taken every half-hour for arterial blood gas analysis. Plasma was separated and stored at −70°C for later determination of plasma tumor necrosis factor (TNF)-α, interleukin (IL)-6 by bioassay, and IL-8 by enzyme-linked immunosorbent assay (ELISA), as a markers of systemic inflammation. Plasma endotoxin concentration was measured using ELISA. Lung tissue wet-to-dry weight ratio and myeloperoxidase concentration were used as markers of edema and neutrophil sequestration, respectively. Bronchoalveolar lavage protein concentration was measured by the bicinclinoic acid method as a measure of capillary-alveolar protein leak. The alveolar–arterial gradient was measured; a large gradient indicated impaired oxygen transport and hence lung injury.

Results

Bilateral hind limb I/R injury increased significantly intestinal mucosal acidosis, intestinal permeability, portal endotoxemia, plasma IL-6 concentrations, circulating phagocytic cell priming and pulmonary leukosequestration, edema, capillary-alveolar protein leak, and impaired gas exchange. Conversely, pigs treated with rBPI21 2 mg/kg at the onset of reperfusion had significantly reduced intestinal mucosal acidosis, portal endotoxin concentrations, and circulating phagocytic cell priming and had significantly less pulmonary edema, leukosequestration, and respiratory failure.

Conclusions

Endotoxin transmigration across a hyperpermeable gut barrier, phagocytic cell priming, and cytokinemia are key events of I/R injury, sepsis, and pulmonary dysfunction. This study shows that rBPI21 ameliorates these adverse effects and may provide a novel therapeutic approach for prevention of I/R-associated sepsis syndrome.

Despite successful surgery and intensive care support, ruptured abdominal aortic aneurysm repair carries a death rate of 50% to 75%. 1,2 Adult respiratory distress syndrome with or without associated multiple organ failure is the primary cause of 20% such deaths and a contributory cause of the remainder. 2–4 Previous work has shown the central role for ischemia–reperfusion (I/R) injury in pathogenesis.

Reperfusion of the ischemic lower torso initiates a systemic inflammatory response syndrome characterized by proinflammatory cytokine 3,5 and increased circulating polymorphonuclear leukocyte (PMN) activation. 6 Pulmonary 4 sequestration of activated neutrophils is followed by acute pulmonary microvascular injury, 7,8 adult respiratory distress syndrome, 4 and a high subsequent death rate. 2 High circulating levels of proinflammatory cytokines responsible for leukocyte activation (e.g., tumor necrosis factor [TNF]-α, interleukin [IL]-6, IL-8) 6 as well as endotoxin 9 have been shown after abdominal aortic aneurysm repair. 5,10

We and others have previously shown that lower limb I/R injury is associated with increased intestinal permeability and endotoxemia. 5,11,12 Endotoxin is a lipopolysaccharide component of the gram-negative bacterial cell wall that provides potent stimulation of cytokine production and leukocyte activation. 13,14 Its precise role in systemic sepsis syndrome after vascular surgery remains controversial. 5

Bactericidal/permeability-increasing protein (BPI) is an antibacterial protein isolated from human neutrophils 15 that binds lipopolysaccharide, inhibits leukocyte activation, lyses or kills bacteria, and ameliorates endotoxin effects. 15–17 The recombinant N-terminal fragment of this protein (rBPI21) retains similar activity. 17 BPI inhibits lipopolysaccharide-induced neutrophil activation and cytokine secretion 18 and decreases tissue PMN deposition. 19 Recent work has shown that recombinant amino-terminal fragments and analogs of BPI (rBPI21 and rBPIn) increased survival, reduced cytokine production, and decreased the hemodynamic and metabolic derangement associated with lethal endotoxin challenge, 20 although human benefits are unclear. 21

We hypothesize that transmigration of endotoxin across a hyperpermeable intestinal barrier, proinflammatory cytokine levels, and remote PMN-mediated acute lung injury after lower limb I/R injury may be ameliorated by rBPI21 pretreatment. In this study, we have tested this hypothesis in a porcine model of lower limb I/R against endpoints of endotoxemia, cytokinemia, and parameters relevant to acute lung injury.

METHODS

Experimental Protocol

These studies were performed in accordance with the Animals (Scientific Procedures) Act of 1986 and with the approval of our institution’s Animals Research Committee. Immature male large white Landrace pigs (28–35 kg; 6–8 weeks old) were deprived of food other than water overnight before surgery.

Anesthesia and Ventilation

Pigs were sedated with intramuscular Stresnal (azaperone BP; 2–3 mg/kg) and anesthesia was induced with intravenous Sagatal (sodium pentobarbital; 20 mg/kg) and maintained with a continuous intravenous infusion of Sagatal (8 mg/kg/h), titrated to maintain apnea. After creation of a tracheostomy with insertion of a 7.5F endotracheal tube, mechanical ventilation (tidal volume 15 mL/kg) was instituted with an anesthesia machine. The ventilatory rate was adjusted to maintain the arterial partial pressure of CO2 (PaCO2) at 30 to 42 mmHg. End-expiratory pressure was zero. The fraction inspired O2 (FiO2) concentration was adjusted according to the experimental protocol. A continuous infusion of compound sodium lactate (Hartmann’s solution; Baxter Healthcare Ltd., London, UK; 15 mL/kg/h) was given from induction of anesthesia throughout the experiments. Heating blankets were used to ensure body temperature of 38.2 ± 0.8°C, monitored by rectal probe and a Model 7835A Hewlett-Packard Patient Monitor (Andover, MA).

Physiologic Measurements

The left carotid artery was exposed surgically and vascular polyethylene catheters (PE-190; Thomas, Philadelphia, PA) were positioned in the left external carotid artery. Continuous arterial blood pressure was monitored by electronic transduction. Intermittent arterial blood gas sampling and systemic blood sampling were carried out. A balloon-tip pulmonary artery catheter was advanced through the left external jugular vein and guided into the pulmonary artery by pressure waveform analysis for monitoring of central venous pressure, mean pulmonary artery pressure, and intermittent pulmonary arterial wedge pressure. Mean pressures were determined using calibrated physiologic pressure transducers (Model 1280C) driving an amplifier monitor (Model 7835A; Hewlett-Packard). The left cephalic vein was exposed surgically and a vascular polyethylene catheter (PE-160; Thomas) was inserted for continuous intravenous infusions of sodium pentobarbital (0.06 mg/kg/min) and compound sodium lactate solution.

Surgical Procedures

After a midline laparotomy incision was made, a 14-FG silicon urinary catheter was placed in the bladder through a cystostomy and secured in place with a purse-string suture for measurement of hourly urine output. A 7F catheter was advanced through the splenic vein to the portal vein for blood sampling. Its position was confirmed by palpating the tip of the catheter through the wall of the portal vein. Tonometers (Tonometrics Inc., Bethesda, MD) were placed in the gastric antrum and sigmoid colon by the endoluminal route, and in the terminal ileum through an enterotomy situated approximately 20 cm from the ileocecal junction. After hemostasis was ensured, the laparotomy was closed. The surgical procedures took approximately 60 minutes. Thereafter, pigs were allowed to stabilize for 30 minutes, as defined by constant heart rate, arterial pressure, and end-tidal CO2.

Pressure Measurements

Intravascular pressure readings were obtained every 15 minutes for core temperature, mean arterial blood pressure, mean pulmonary artery pressure, pulmonary capillary wedge pressure, and central venous pressure.

Blood Gas Measurements

Blood was obtained every 30 minutes from the carotid artery for measurements of arterial partial pressure of O2 (PaO2), PaCO2, and O2 content. The gas analysis of the arterial sample were performed in an automatic blood gas analyzer (Model 1304 pH/blood gas analyzer, Instrumentation Laboratory, Warrington, UK) without delay. FiO2 was measured in a ventilatory circuit using a 4700 Oxycap Monitor (Ohmeda, Louisville, CO). The arterial–alveolar (A-a) gradient was measured using the formula [(A-a) gradient = FiO2 × 710 -(arterial pCO2/0.8) - arterial Po2]. A large (A-a) gradient indicates impaired oxygen transport and hence lung injury.

Tonometry

Saline solution (5 mL at room temperature) was injected into the tonometer balloons and left for 30 minutes. At the end of this equilibration period, saline was removed according to the manufacturer’s instructions and saline PCO2 was measured immediately with the automatic blood gas analyzer. With the conversion table supplied by the manufacturer, measured PCO2 was transposed to a steady-state PCO2 depending on the exact duration of equilibration (correction factor = 1.13; equilibrium time 30 minutes). Gut mucosal pH was calculated substituting the PCO2 and the simultaneously measured arterial bicarbonate in the Henderson-Hasselbalch equation, pHi = 6.1 + log10([HCO3]/(0.03 × PCO2ss)), where 6.1 accounts for HCO3− pK value at 37°C, [HCO3−] is the measured arterial blood bicarbonate concentration (mmol/L), PCO2ss is the steady-state time-adjusted PCO2 (mmHg) of tonometer saline, and 0.03 is the CO2 solubility coefficient in plasma (mmol/L/mmHg) at 37°C. 22

Intestinal Permeability

Intestinal permeability was assessed by measuring the urinary excretion of lactulose and mannitol. After an 8-hour fast, the test solution (5 g lactulose and 10 g mannitol dissolved in 50 mL distilled water) was instilled via a previously inserted gastric tube. The bladder was emptied and subsequently all urine was collected for a 6-hour period. The volume of urine was measured and aliquots of this and the pretest sample were stored at −20°C until assayed. The concentrations of lactulose and mannitol were measured enzymatically by the reduced nicotinamide adenine dinucleotide (phosphate)-linked enzymatic assays, modified from techniques described by Lunn et al, 23 using a Cobras Fara centrifugal analyser (Roche Diagnostics, Welwyn Garden City, UK). The amount of lactulose and mannitol excreted was calculated as a percentage of the administered dose, and intestinal permeability was expressed as the lactulose-to-mannitol ratio.

Experimental Design

Basal values were recorded after the 30-minute recovery period. Animals were randomized into three groups (n = 6 per group). Two groups underwent bilateral external iliac artery occlusion by application of bulldog arterial clamps for 120 minutes followed by 150 minutes of reperfusion on release of bulldog clamps. Absence of flow, and restoration when appropriate, was confirmed using an ultrasonic flow probe (Transonic Systems Inc., Ithaca, NY) applied to the vessel wall. Six pigs were allocated as sham controls and kept for 270 minutes with recordings as for the experimental animals (see below).

Treatment

Animals undergoing bilateral hind limb ischemia were randomized to receive treatment with either rBPI21 (Xoma [US] LLC, Berkeley, CA; 2 mg/kg body weight by intravenous infusion over 30 minutes from the beginning of reperfusion) or thaumatin, a control-protein preparation 24 (2 mg/kg body weight administered by intravenous infusion over 30 minutes from the beginning of reperfusion). The control group also received thaumatin (2 mg/kg body weight) at the same time points.

Blood and Tissue Sampling

Blood samples and measurements were taken every half-hour during the experimental period. Blood samples were collected in heparinized (20 units per milliliter blood), sterile pyrogen-free tubes. Samples were immediately transferred on ice to be centrifuged at 400 g (at 4°C) for 10 minutes. Plasma was then aliquoted into sterile cryotubes (Nunc, Intermed, Roskidle, Denmark) and stored at −70°C until the time of assay. Immediately after final blood samples were taken, each animal underwent a midline thoracotomy and samples of right lung were excised from predetermined sites. Separate samples were either fixed in 4% formalin or placed immediately in a sterile pyrogen-free cryotube (Nunc), snap-frozen in liquid nitrogen, and stored at −70°C until the time of histologic assessment or assay, respectively.

Bronchoalveolar Lavage

Bronchoalveolar lavage was performed through the endotracheal tube using a Silastic pulmonary suction catheter at 240 minutes, into the right lung. The distal end was wedged into third- or fourth-order bronchi of the middle and lower lobes. All lavages were performed by the same person, and catheter placement was confirmed on cadaver studies. Each lobe was washed with 50 mL sterile 0.9% sodium chloride; retrieval volumes consistently exceeded 50% of instilled volume. Lavage fluid was centrifuged at 400 g at 4°C for 10 minutes, and the supernatant was stored at −20°C. The lavage protein content was measured in the noncellular fraction by the bicinclinoic acid method.

Luminol-Enhanced Chemiluminescence Assay

Chemiluminescence was assayed with an LKB Wallac Luminometer 1251 (Wallac Oy, Tusku, Finland). This instrument detects and records the number of photons released by luminol deoxygenation. Under the conditions used, this luminol-dependent chemiluminescence assay represents primarily PMN respiratory burst activity when applied to whole blood. To trigger a respiratory burst, assay tubes contained 100 μL phorbol 12-myristate 13-acetate. Reactions were initiated by the automated injection of 600 μL Luminol Complete medium into each reaction tube and 100 μL blood (prediluted 1:40 in blood diluting medium). Some reaction tubes contained 47 pmol/L recombinant human TNF-α receptor fusion protein. All chemiluminescence assays were performed in duplicate at 37°C.

Activity is reported as the integral of the chemiluminescence events registered during the initial 10 minutes of activation. The degree of respiratory burst activity is dependent on prior priming in vivo; further, by maximally priming some samples with addition of TNF-α to the diluted whole blood, a ratio could be determined between whole blood priming and whole blood + TNF-α, thus comparing in vivo priming with maximal priming.

Total White Cell Counts

Arterial blood samples were drawn into sterile glass tubes containing 0.15% acetic acid and kept at 4°C (Vacutainer, Becton Dickinson, Rutherford, NJ). Small aliquots of blood were set aside for white blood cell counts and blood smear differential cell counts that were performed as previously described. The remainder of the samples were centrifuged at 500 g at 4°C for 20 minutes, and the resulting plasma was stored at −20°C.

Tumor Necrosis Factor-Alpha Assay

Biologically active TNF-α was bioassayed by an MTT tetrazolium-based cytotoxicity assay based on the TNF-α-sensitive WEHI 164 clone, as previously described. 25 Absorbance was then read at 570 nm, and the amount of TNF-α in each sample was computed from the standard curve. Interassay and intraassay coefficients of variation were less than 10%. The detection limit of the assay is 0.02 pg/mL TNF-α.

Interleukin Assays

Biologically active IL-6 was measured using a bioassay based on the proliferation of IL-6-dependent B9 hybridoma cells (a generous gift of L. Aarden, Amsterdam), as previously described. 25 Absorbance was read at 570 nm, and the amount of IL-6 in each sample was computed from the standard curve. Interassay and intraassay coefficients of variation were less than 10%. Plasma levels of IL-8 and IL-10 were analyzed by sandwich enzyme-linked immunosorbent assays (ELISAs) according to the instructions of the manufacturer (Biosource International Inc., Camarillo, CA).

Endotoxin

Endotoxin concentration was determined using a quantitative Limulus amebocyte lysate (LAL) chromogenic assay (Quadratech, Epsom, UK). Detection limits of the assay were 0.1 to 100 pg/mL, the interassay coefficient of variation was less than 10%, and the intraassay coefficient of variation was less than 10%.

Measurement of Lung Tissue Myeloperoxidase

The tissue specimen (200–400 mg) was homogenized in HTAB buffer on ice. The pooled homogenate were sonicated for 10 seconds, freeze-thawed thrice, and centrifuged at 40,000 g for 15 minutes. In the supernatant, myeloperoxidase activity was measured spectrophotometrically: 0.1 mL supernatant was combined with 2.9 mL of 50 mmol/L phosphate buffer (pH 6.0) containing 0.167 mg/mL O-dianisidine hydrochloride and 0.0005% hydrogen peroxide. The change in absorbance at 460 nm was measured with a Beckmann DU-2 spectrophotometer (Beckmann Instruments, Inc., Cedar Grove, NJ). One unit of myeloperoxidase activity is defined as that degrading 1 micromol of peroxide per minute at 25°C.

Lung Tissue Wet-to-Dry Weight Ratios

Wet-to-dry ratios of lung and skeletal muscle tissue samples (200–400 mg) were calculated. Each specimen was blotted dry, weighed, and then placed in a vacuum freeze-dryer at −70°C for 48 hours. Specimens were then reweighed and the wet-to-dry tissue weight ratio was calculated.

Statistical Analysis

Summary values are expressed as mean ± standard error of the mean or median (interquartile range). One-way repeated measures analysis of variance was used to compare sequential measurements for parametric data. The Dunnett test was used to make further comparisons if analysis of variance revealed significant differences. The control value for the Dunnett test was designated as the last measurement obtained at the end of the baseline period (time 0). The Student t test was used for independent comparisons for parametric data. Kruskall-Wallis H analysis of variance was used in conjunction with the Mann-Whitney test for nonparametric comparisons. A two-tailed P < .05 was considered statistically significant.

To assess the degree of association between the variables, the Spearman rank correlation coefficient was calculated. Significance was taken at the 5% level. Statistical calculations were performed using SPSS (Version 9.0, Microsoft, Redmond, WA) software.

RESULTS

Physiologic and Hemodynamic Variables

Central venous pressure and pulmonary capillary wedge pressure did not change significantly from baseline throughout the experimental period. Mean arterial pressure was significantly greater in both I/R groups (P < .01) and the rBPI21-treated I/R group (P < .001) versus the control group during the ischemic period (t = 0–120) and remained significantly elevated in the rBPI21-treated I/R group during the reperfusion period versus the control group (P < .001) and the untreated I/R group (P < .01).

Arterial partial pressure of oxygen decreased significantly during the reperfusion period in the I/R group versus the control group (P < .05) and the rBPI21-treated I/R group (P < .01) (Fig. 1). Pulmonary artery pressure was significantly greater during reperfusion in the I/R group versus the control group (P < .001) and the rBPI21-treated I/R group (P < .001). However, the pressure in the treated group remained significantly higher than in the control group (P < .02).

graphic file with name 14FF1.jpg

Open in a new tab

Figure 1. (A) Arterial PO2 (mmHg) in control, ischemia–reperfusion (I/R), and rBPI21 + I/R groups. *P < .05 vs. control and rBPI21 + I/R; †P < .01 vs. control; ‡P < .005 vs. rBPI21 + I/R. (B) Pulmonary artery pressure (mmHg) in control, I/R, and rBPI21 + I/R groups. *P < .05 vs. control; †P < .001 vs. control; ‡P < .001 vs. rBPI21 + I/R. (C) Alveolar–arterial gradient in control, I/R, and rBPI21 + I/R groups. *P < .05 vs. control; †P < .03 vs. rBPI21 + I/R. Data represent mean ± SEM. Repeated measures two-way analysis of variance (time and group), *P < .0001, †P < .0001, ‡P < .034. Post hoc Dunnett T3 test, *P < .05, †P < .05, ‡P < .01.

Alveolar-arterial O2 gradient values were similar during the ischemia period but were significantly higher after reperfusion in the I/R group versus the control group (P < .05). This finding suggests impaired alveolar oxygen transport. This impairment in oxygen transport was prevented in the rBPI21-treated I/R group (P < .04).

Tonometry Measurements

Temporal changes in gastric mucosal pH, ileal mucosal pH, and sigmoid mucosal pH at baseline and for each time point during the experiment are shown in Figure 2. A significant decrease was seen in ileal pHi during the reperfusion period (t = 120–270) in the untreated I/R group versus both the control group (P < .05) and the rBPI21-treated I/R group (P < .05). Sigmoid pHi decreased during the reperfusion period (t = 120–270) in the untreated I/R group versus the control group (P < .005) and the rBPI21-treated I/R group (P < .02).

graphic file with name 14FF2.jpg

Open in a new tab

Figure 2. (A) Gastric pHi in control, ischemia–reperfusion (I/R), and rBPI21 + I/R groups. *P < .05 vs. control; †P < .01 vs. control. (B) Ileal pHi in control, I/R, and rBPI21 + I/R groups. *P < .05 vs. control; †P < .05 vs. rBPI21 + I/R. (C) Sigmoid pHi in control, I/R, and rBPI21 + I/R groups. *P < .01 vs. control; †P < .05 vs. rBPI21 + I/R. Data represent mean ± SEM. Repeated measures two-way analysis of variance (time and group), *P < .05, †P < .05, ‡P < .05.

Plasma Interleukin-6

Plasma IL-6 levels were significantly greater after 1 hour of reperfusion (t = 180) in the I/R group, 1726.00 (1130.00–2130.00) pg/mL, versus the control group, 303.50(274.35–374.25) pg/mL (P < .003), or the rBPI21-treated I/R group, 362.5 (304.48–452.58) (P < .001) (Fig. 3). There was no significant difference in the plasma IL-6 level during reperfusion between the systemic, portal sampling sites.

graphic file with name 14FF3.jpg

Open in a new tab

Figure 3. (A) Systemic interleukin (IL)-6 levels (pg/mL) in control, ischemia–reperfusion (I/R), and rBPI21 + I/R groups. *P < .003 vs. control; †P < .01 vs. control; ‡P < .001 vs. rBPI21 + I/R. (B) Portal IL-6 levels (pg/mL) in control, I/R, and rBPI21 + I/R groups. *P < .003 vs. control; †P < .02 vs. control; ‡P < .001 vs. rBPI21 + I/R. Data represent median (interquartile range).

Plasma Tumor Necrosis Factor-Alpha

Plasma TNF levels were significantly greater after 1 hour of reperfusion (t = 180) in the I/R group, 86.83 (48.71–182.02) pg/mL, versus the control group, 32.67(0.95–42.82) pg/mL, and the rBPI21-treated group, 5.21 (1.88–15.46) (P < .05). There was no significant difference in the plasma TNF level during reperfusion between the systemic, portal sampling sites (Fig. 4).

graphic file with name 14FF4.jpg

Open in a new tab

Figure 4. (A) Systemic tumor necrosis factor (TNF)-alpha levels (pg/mL) in control, ischemia–reperfusion (I/R), and rBPI21 + I/R groups. *P < .05 vs. control and rBPI21 + I/R. (B) Portal TNF-α levels (pg/mL) in control, I/R, and rBPI21 + I/R groups. No significant difference between groups, P < .2. Data represent median (interquartile range).

Endotoxin Concentration, Portal Vein

Endotoxin concentrations were comparable in the I/R and control groups in the reperfusion period. In the I/R group, the endotoxin concentration in the systemic circulation decreased during reperfusion time points t = 180, 11.30 (0.00–61.00), and t = 270, 16.20 (0.00–49.60), versus the portal vein at t = 180, 500 (28.00–514.20) (P < .01), and t = 270, 505.70 (92.00–530.70) (P < .001), respectively (Fig. 5).

graphic file with name 14FF5.jpg

Open in a new tab

Figure 5. (A) Systemic endotoxin levels (pg/mL) in control, ischemia–reperfusion (I/R), and rBPI21 + I/R groups. No significant difference between groups, P < .3. (B) Portal endotoxin levels (pg/mL) in control, I/R, and rBPI21 + I/R groups. *P < .01 vs. control; †P < .009 vs. control; ‡P < .05 vs. rBPI21 + I/R. Data represent median (interquartile range).

Portal vein endotoxin concentration increased during the reperfusion period in the I/R group, 505.70 (92.00–530.70) pg/mL, versus the control group, 18.92 (8.65–78.67) (P < .009). This effect was significantly attenuated in the rBPI21-treated I/R group, 27.41 (8.55–95.65) (P < .05).

Lung Tissue Ratio

The lung wet-to-dry tissue weight ratio was significantly greater in the I/R group, 10.13 (7.66–12.66), versus the control group, 5.57 (4.54–7.37) (P < .011). This increase was significantly attenuated in the rBPI21-treated I/R group, 6.46 (5.06–7.28), versus the untreated I/R group (P < .02) (Fig. 6).

graphic file with name 14FF6.jpg

Open in a new tab

Figure 6. (A) Lung wet-to-dry ratio in control, ischemia–reperfusion (I/R), and rBPI21 + I/R groups. *P < .01 vs. control; †P < .02 vs. rBPI21 + I/R. (B) Lung myeloperoxidase level (U/g) in control, I/R, and rBPI21 + I/R groups. *P < .0001 vs. control. (C) Bronchoalveolar lavage protein content (mg/L) in control, I/R, and rBPI21 + I/R groups. *P < .05 vs. control; †P < .04 vs. rBPI21 + I/R. Data represent median (interquartile range).

Lung Tissue Myeloperoxidase

The lung tissue myeloperoxidase concentration was significantly greater in the I/R group, 33.81 (15.65–52.66) units of absorbance per gram dry tissue, versus the control group, 7.54 (0.97–8.63) (P < .005). This increase in the lung tissue myeloperoxidase concentration was significantly attenuated in the rBPI21-treated I/R group, 24.47 (13.16–25.87), versus the untreated I/R group (P < .02).

Lavage Noncellular Fraction Protein Concentration

The lavage protein concentration was significantly greater after 2 hours of reperfusion (t = 240) in the I/R group, 0.73 (0.31–1.60), versus the control group, 0.23 (0.18–0.31) (P < .05). This increase was significantly attenuated in the rBPI21-treated I/R group, 0.21 (0.20–0.21), versus the untreated I/R group (P < .042).

The Spearman rank correlation for the myeloperoxidase and wet-to-dry ratio was P < .0001, r = 0.57; the correlation for lavage protein and the wet-to-dry ratio was P < .05, r = 0.389.

Lactulose/Mannitol Urinary Excretion Ratio

The lactulose/mannitol urinary excretion ratio was significantly greater in the I/R group, 0.191 ± 0.03, versus the control group, 0.092 ± 0.03 (P < .024) and versus the rBPI21-treated I/R group, 0.06 ± 0.02 (P < .001). Data represents mean (standard error of the mean), comparisons with analysis of variance and the Student t test (Fig. 7). There was a significant (P < .008), positive correlation (r = 0.604) between plasma portal endotoxin concentration and the lactulose/mannitol urinary excretion ratio by Spearman’s rank correlation.

graphic file with name 14FF7.jpg

Open in a new tab

Figure 7. (A) Lactulose/mannitol ratio in control, ischemia–reperfusion (I/R), and rBPI21 + I/R groups. *P < .02 vs. control; †P < .001 vs. rBPI21 + I/R. (B) Spearman rank correlation between lactulose/mannitol ratio and portal endotoxin (pg/mL) at t = 270;P < .008, r = 0.604.

Whole-Blood Chemiluminescence

There was a significantly greater whole blood phagocytic priming ratio after 30 minutes of reperfusion (t = 150) in the I/R group, 1.01 ± 0.08, versus the control group, 0.56 ± 0.11 (P < .033) and versus the rBPI21-treated I/R group, 0.62 ± 0.06 (P < .009). After 2 hours of reperfusion (t = 240), the priming ratio remained significantly greater in the I/R group, 0.86 ± 0.14, versus the control group, 0.26 ± 0.07 (P < .021) and the rBPI21 -treated group, 0.54 ± 0.07 (P < .05). The priming ratio decreased significantly in the control group from the start (t = 0), 0.65 ± 0.11, to the end (t = 240), 0.26 ± 0.07 (P < .005). A significant increase in priming occurred in the I/R group from t = 0 to t = 240, 0.58 ± 0.06 versus 0.86 ± 0.14 (P < .05) (Fig. 8 A).

graphic file with name 14FF8.jpg

Open in a new tab

Figure 8. (A) Circulating phagocytic cell priming (% maximum) in control, ischemia–reperfusion (I/R), and rBPI21 + I/R groups. *P < .05 vs. control; †P < .01 vs. rBPI21 + I/R group; ‡P < .05 vs. control. (B) Systemic interleukin-8 levels (pg/L) in control, I/R, and rBPI21 + I/R groups. *P < .02 vs. control; †P < .05 vs. control. (C) Systemic interleukin-10 levels (pg/L) in control, I/R, and rBPI21 + I/R groups. *P < .001 vs. control; †P < .05 vs. control; ‡P < .005 vs. control. Data represent mean ± SEM.

Whole Blood Leukocyte Counts

No difference in total whole blood leukocyte count at time zero and after 240 minutes was found between the control group, 15.55 (10.50–24.97) × 109/L and 23.90 (20.91–25.75) × 109/L, or the I/R group, 13.56 (12.27–21.78) × 109/L and 20.94 (14.45–26.91) × 109/L. Nor was there any difference in the neutrophil fraction between the control group and the I/R group at the start or the end of the experiment.

Plasma Interleukin-8

Figure 8 B shows the temporal changes in the plasma IL-8 concentration in the control group, the untreated I/R group, and the rBPI21-treated I/R group. After 1 hour of reperfusion (t = 180), plasma IL-8 levels were elevated in the control group, 198.10 (116.10–373.90) pg/mL, versus the I/R group, 42.42 (25.21–107.10) (P < .026), and the rBPI21-treated I/R group, 68.69 (64.68–98.61) (P < .008).

Plasma Interleukin-10

After 1 hour of reperfusion (t = 180) compared with the control group, 11.19 (6.75–14.11) pg/mL, plasma IL-10 levels were significantly elevated in the I/R group, 30.76 (19.56–77.70) (P < .001), and the rBPI21-treated I/R group, 32.67 (24.63–42.10) (P < .001).

DISCUSSION

Our results show that a novel recombinant antilipopolysaccharide protein ameliorates the entire cell and molecular cascade of I/R-initiated sepsis syndrome. Invasive hemodynamic measurements and maintenance intravenous volume supplementation ensured that the large animal model remained normotensive throughout the experimental period. We used the dual sugar (lactulose and mannitol) absorption test, intestinal tonometry, and the LAL ELISA as indices of intestinal permeability and endotoxemia. We used plasma cytokinemia and luminol-enhanced whole blood chemiluminescence as indices of systemic inflammation and PMN activation. Lung tissue wet-to-dry weight, myeloperoxidase concentration, bronchoalveolar lavage protein concentration, and pulmonary physiologic measurements were used as indices of acute lung injury.

Ruptured abdominal aortic aneurysm repair is accompanied by I/R and carries a 50% to 75% death rate. 1 Increased levels of endotoxin, a lipopolysaccharide component of the wall of gram-negative bacteria, have been shown in the portal and systemic circulation of patients after aneurysm repair. 9,26 The conversion of an essentially sterile I/R injury to a septic syndrome has led many to question the role of endotoxin and sepsis in the high death rate after ruptured aneurysm. 5,27 Intestinal injury in this model was associated with a significant increase in gastric, ileal, and sigmoid mucosal acidosis during the reperfusion period. Further, intestinal permeability, assessed with the dual sugar absorption test, was significantly increased. Treatment with rBPI21 at the start of the reperfusion period significantly attenuated the increase in intestinal permeability and intestinal mucosal acidosis and completely attenuated the concentration of portal endotoxin. It is not possible to tell whether the reduction in portal endotoxin concentration was due to decreased transmigration or increased leukocyte-mediated clearance of endotoxin. The gut has a large resident population of gram-negative bacteria, 28 and in health some endotoxin does enter the portal circulation without any major systemic inflammation. 29 The relevance of endotoxemia to the inflammatory response in vascular surgery has been questioned. 5,26 Despite this, endotoxin is undoubtedly a potent stimulus to systemic inflammation and activation of PMNs. 13,30 Woodcock et al 31 recently showed bacterial translocation in patients with abdominal aortic aneurysm and concluded that it increased septic complications. Treatment with rBPI21 attenuated the intestinal mucosal acidosis, a reliable indicator of mucosal blood flow, in this model. Impaired splanchnic blood flow has been suggested as a cause of extended gut injury in sepsis. 32 Perhaps increased clearance or decreased transmigration of endotoxin into the microvasculature may decrease vasoconstriction. 26 Another possible explanation is that rBPI21 treatment attenuates the IL-6 response, reduces systemic inflammation, and therefore reduces secondary hemodynamic and mediator-induced intestinal injury.

The sepsis syndrome secondary to endotoxemia remains a leading cause of death and complications. Initial optimism about the benefits of antiendotoxin immunotherapy 33 has given way to doubts about the clinical efficacy and indeed the endotoxin-binding properties of these antibodies. 34 BPI and its N-terminal fragments and analogs have certain advantages, being neither serotype-specific nor inhibited by the O-specific side chain of lipopolysaccharide. 16,17 Their short plasma half-life makes the timing of drug delivery crucial. In vitro analogs of BPI have been shown to inhibit lipopolysaccharide-induced PMN activation and cytokine secretion. 18 Hansbrough et al 19 showed decreased tissue PMN deposition in a sterile acute burn injury model using recombinant BPI, suggesting that this effect may be mediated by inhibition of PMN activation. This finding may suggest that BPI and its analogs have beneficial effects in the absence of endotoxin. Recent reports on BPI (rBPI21 and rBPI23) have shown increased survival, reduced cytokine production, and decreased inflammatory response to lethal endotoxin challenge in animals. 20 Its beneficial effects in humans are still under debate. 35

High levels of TNF-α, IL-1β, IL-6, and IL-8 have been shown after abdominal aortic artery repair, 36–38 and persistently high levels correlate with death. 39 We have shown small but significant elevations in plasma TNF-α, a potent early proinflammatory mediator, in response to hind limb I/R. Treatment after ischemia with rBPI21 attenuates the TNF-α response after 1 hour of reperfusion. Reperfusion of ischemic skeletal muscle increases TNF-α production in the absence of any septic stimulus. TNF-α has been implicated in the development of acute lung injury after lower limb I/R via induction of nitric oxide synthesis; injury can be prevented by TNF antagonists. 40 We have shown highly significant elevations in plasma IL-6 levels, most markedly throughout the reperfusion period. Treatment with rBPI21 at the start of reperfusion completely abrogated this increased production of IL-6, a secondary mediator with both pro- and antiinflammatory activities, suggesting that prevention of endotoxin-induced inflammatory activation prevents progression to uncontrolled inflammation. Cytokines such as IL-6 have also been shown to inhibit PMN apoptosis in vitro and therefore may increase the functional longevity of such leukocytes and their ability to induce tissue injury. 36 IL-8 is a powerful chemoattractant cytokine, and its presence has been implicated in the development of acute lung injury after abdominal aortic artery repair. 37 We have shown significantly lower levels of IL-8 after I/R injury in this model versus control, unaffected by treatment with rBPI21. This is an unexpected finding, because IL-8 has been shown be a chemoattractant to PMNs, and levels are elevated in a variety of inflammatory states. Of interest, the levels of IL-10, a primarily antiinflammatory cytokine, were significantly elevated in the two I/R groups versus the control group. This may suggest a shift in the balance toward compensatory antiinflammatory response syndrome. IL-10 has been shown to decrease the amount of circulating proinflammatory mediators such as IL-1β, TNF-α, and IL-8. The administration of exogenous IL-10 has been shown to reduce local limb injury and to attenuate remote lung injury. 38,41 From these data, it is not possible to confirm the source of the cytokines observed. It has been suggested that cytokine production by patients undergoing abdominal aortic artery surgery is a response to occult ischemia of the colon. 22,26,42 Lower limb I/R is associated with systemic endotoxemia; hence, a possible second source of cytokines is remote activation of enterocytes and gastrointestinal macrophages. 43,44 In this study, ischemia was confined to the lower limb, indicating that lower limb I/R injury itself is the stimulus for cytokine production. Further, IL-6 concentrations were no greater in the portal circulation than in the systemic circulation. The continuous increase in IL-6 concentrations throughout the reperfusion period supports the finding of others that IL-6 production is proportional to the extent of reperfusion injury. 5,39

Increased circulating PMN priming and activation have been shown after abdominal aortic artery repair. 6 The process whereby exposure to a plasma factor increases the capacity of PMNs to produce reactive oxidant species, in response to a secondary stimulus, is called priming. 45 In reperfusion injury, a variety of cytokines and mediators may be responsible for priming neutrophils. 45,46 Luminol-dependent chemiluminescence of whole blood was developed to estimate the phagocytic function of granulocytes and the serum opsonic activity simultaneously. It measures reactive oxygen species production and neutrophil degranulation, which is specific and azurophilic. 46 We found in our model increased circulating whole blood phagocytic cell priming, and we also showed that treatment with rBPI21 can attenuate this increase. This decreased priming and potential for respiratory burst-induced injury may explain the decreased intestinal and pulmonary injury.

In this study, the development of remote pulmonary dysfunction was observed only after reperfusion, with development of hypoxemia, pulmonary hypertension, and impaired oxygen transport. Humoral and/or cellular mediators may be implicated. 47 Neutrophil-mediated acute lung injury is suggested by elevated levels of lung tissue myeloperoxidase, lavage protein concentration, and wet-to-dry weight ratio. The prevention of this acute pulmonary dysfunction in the rBPI21-treated group suggests a role for endotoxin-induced circulating neutrophil activation in the development of this lung injury. Of interest, the myeloperoxidase concentration in the rBPI21-treated I/R group was elevated compared with the control group, although neither one was significantly different from the control or untreated I/R group. However, there was significantly less edema and protein leakage in the treated group, perhaps suggesting that sequestered neutrophils are less primed and induce less tissue injury. Our findings are in agreement with previous reports of a varying degree of pulmonary dysfunction, characterized by increased microvascular permeability and leukocytic sequestration within the lung. 4 Early signs of adult respiratory distress syndrome have been reported within hours of aortic declamping in patients undergoing abdominal aortic artery repair 4 and animal models of reperfusion injury. 4,48,49 We witnessed a significant degree of pulmonary hypertension in our I/R group associated with the impaired alveolar oxygen transport, which has been shown in patients with adult respiratory distress syndrome. 28,50 The importance of PMNs in acute lung injury after limb I/R has been confirmed in leukopenic models. 51 Although acute lung injury has been shown in PMN-depleted animals, 52 it is less severe, probably caused by vasoactive mediators such as thromboxane A2 and leukotriene B4. 53,54 The mechanism whereby lung injury is primarily due to neutrophil adherence, sequestration, and subsequent respiratory burst-induced oxidative injury is supported by the observation that lung injury can be attenuated with anti-CD18 monoclonal antibodies 49 or by antioxidant preloading. 48

A growing body of evidence supports a role for gut hyperpermeability 55 and translocation of endotoxin in the propagation of the inflammatory response to abdominal aortic aneurysm surgery. If gut endotoxin converts a recoverable inflammatory response into a lethal endotoxemia in a certain subset of patients, then rBPI21 may represent an important advance in the treatment of these patients.

In conclusion, we have shown that rBPI21 can reduce endotoxemia and the production of TNF-α and IL-6. It significantly attenuates the remote acute pulmonary dysfunction seen after hind limb I/R injury. This is associated with a reduction in circulating neutrophil priming. We therefore conclude that rBPI21 attenuates the systemic inflammatory response syndrome and lung injury associated with limb I/R by preventing a secondary hit septic (lipopolysaccharide) insult.

Acknowledgment

The authors thank Xoma Corp. (US) LLC (Berkeley, CA) for generous provision of rBPI21.

Footnotes

Presented in part to the Surgical Research Society, first-place David Patey Prize, Cardiff, Wales, May 2000.

Dr. Harkin is the recipient of a project grant from the Wellcome Trust.

Correspondence: Denis W. Harkin, MD, Department of Surgery, The Queen’s University of Belfast, Institute of Clinical Sciences, Grosvenor Road, Belfast BT12 6BJ, Northern Ireland.

E-mail: d.w.harkin@qub.ac.uk

Accepted for publication January 8, 2000.

References

  • 1.Adam DJ, Mohan IV, Stuart WP, et al. Community and hospital outcome from ruptured abdominal aortic aneurysm within the catchment area of a regional vascular surgical service. J Vasc Surg 1999; 30: 922–928. [DOI] [PubMed] [Google Scholar]
  • 2.Huber TS, Harward TR, Flynn TC, et al. Operative mortality rates after elective infrarenal aortic reconstructions. J Vasc Surg 1995; 22: 287–293. [DOI] [PubMed] [Google Scholar]
  • 3.Groeneveld AB, Raijmakers PG, Rauwerda JA, et al. The inflammatory response to vascular surgery-associated ischaemia and reperfusion in man: effect on postoperative pulmonary function. Eur J Vasc Endovasc Surg 1997; 14: 351–359. [DOI] [PubMed] [Google Scholar]
  • 4.Paterson IS, Klausner JM, Pugatch R, et al. Noncardiogenic pulmonary edema after abdominal aortic aneurysm surgery. Ann Surg 1989; 209: 231–236. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Roumen RM, Hendriks T, Nieuwenhuijzen GA, et al. Cytokine patterns in patients after major vascular surgery, hemorrhagic shock, and severe blunt trauma. Relation with subsequent adult respiratory distress syndrome and multiple organ failure. Ann Surg 1993; 218: 769–776. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Barry MC, Kelly C, Burke P, et al. Immunological and physiological responses to aortic surgery: effect of reperfusion on neutrophil and monocyte activation and pulmonary function. Br J Surg 1997; 84: 513–519. [PubMed] [Google Scholar]
  • 7.Anner H, Kaufman RP Jr, Valeri CR, et al. Reperfusion of ischemic lower limbs increases pulmonary microvascular permeability. J Trauma 1988; 28: 607–610. [DOI] [PubMed] [Google Scholar]
  • 8.Welbourn CR, Goldman G, Paterson IS, et al. Pathophysiology of ischaemia reperfusion injury: central role of the neutrophil. Br J Surg 1991; 78: 651–655. [DOI] [PubMed] [Google Scholar]
  • 9.Baigrie RJ, Lamont PM, Whiting S, et al. Portal endotoxin and cytokine responses during abdominal aortic surgery. Am J Surg 1993; 166: 248–251. [DOI] [PubMed] [Google Scholar]
  • 10.Caty MG, Guice KS, Oldham KT, et al. Evidence for tumor necrosis factor-induced pulmonary microvascular injury after intestinal ischemia-reperfusion injury. Ann Surg 1990; 212: 694–700. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Corson RJ, Paterson IS, O’Dwyer ST, et al. Lower limb ischaemia and reperfusion alters gut permeability. Eur J Vasc Surg 1992; 6: 158–163. [DOI] [PubMed] [Google Scholar]
  • 12.Yassin MM, Barros D’Sa AA, Parks TG, et al. Lower limb ischaemia-reperfusion injury alters gastrointestinal structure and function. Br J Surg 1997; 84: 1425–1429. [PubMed] [Google Scholar]
  • 13.van Deventer SJ, Buller HR, ten Cate JW, et al. Experimental endotoxemia in humans: analysis of cytokine release and coagulation, fibrinolytic, and complement pathways. Blood 1990; 76: 2520–2526. [PubMed] [Google Scholar]
  • 14.van Deventer SJ, Hack CE, Wolbink CE, et al. Endotoxin-induced neutrophil activation–the role of complement revisited. Prog Clin Biol Res 1991; 367: 101–109. [PubMed] [Google Scholar]
  • 15.Weiss J, Olsson I. Cellular and subcellular localization of the bactericidal/permeability-increasing protein of neutrophils. Blood 1987; 69: 652–659. [PubMed] [Google Scholar]
  • 16.Marra MN, Wilde CG, Griffith JE, et al. Bactericidal/permeability-increasing protein has endotoxin-neutralizing activity. J Immunol 1990; 144: 662–666. [PubMed] [Google Scholar]
  • 17.Weiss J, Elsbach P, Shu C, et al. Human bactericidal/permeability-increasing protein and a recombinant NH2-terminal fragment cause killing of serum-resistant gram-negative bacteria in whole blood and inhibit tumor necrosis factor release induced by the bacteria. J Clin Invest 1992; 90: 1122–1130. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Meszaros K, Parent JB, Gazzano-Santoro H, et al. A recombinant amino terminal fragment of bactericidal/permeability-increasing protein inhibits the induction of leukocyte responses by LPS. J Leukoc Biol 1993; 54: 558–563. [DOI] [PubMed] [Google Scholar]
  • 19.Hansbrough J, Tenenhaus M, Wikstrom T, et al. Effects of recombinant bactericidal/permeability-increasing protein (rBPI23) on neutrophil activity in burned rats. J Trauma 1996; 40: 886–892. [DOI] [PubMed] [Google Scholar]
  • 20.Vandermeer TJ, Menconi MJ, O’Sullivan BP, et al. Bactericidal/permeability-increasing protein ameliorates acute lung injury in porcine endotoxemia. J Appl Physiol 1994; 76: 2006–2014. [DOI] [PubMed] [Google Scholar]
  • 21.Elsbach P, Weiss J. Prospects for use of recombinant BPI in the treatment of gram-negative bacterial infections. Infect Agents Dis 1995; 4: 102–109. [PubMed] [Google Scholar]
  • 22.Fiddian-Green RG, Amelin PM, Herrmann JB, et al. Prediction of the development of sigmoid ischemia on the day of aortic operations. Indirect measurements of intramural pH in the colon. Arch Surg 1986; 121: 654–660. [DOI] [PubMed] [Google Scholar]
  • 23.Lunn PG, Northrop CA, Northrop AJ. Automated enzymatic assays for the determination of intestinal permeability probes in urine, 2. Mannitol. Clin Chim Acta 1989; 183: 163–170. [DOI] [PubMed] [Google Scholar]
  • 24.Bauer RJ, Wedel N, Havrilla N, et al. Pharmacokinetics of a recombinant modified amino terminal fragment of bactericidal/permeability-increasing protein (rBPI21) in healthy volunteers. J Clin Pharmacol 1999; 39: 1169–1176. [PubMed] [Google Scholar]
  • 25.Yassin MM, Barros D’Sa AA, Parks G, et al. Mortality following lower limb ischemia-reperfusion: a systemic inflammatory response? World J Surg 1996; 20: 961–966. [DOI] [PubMed] [Google Scholar]
  • 26.Soong CV, Blair PH, Halliday MI, et al. Endotoxaemia, the generation of the cytokines and their relationship to intramucosal acidosis of the sigmoid colon in elective abdominal aortic aneurysm repair. Eur J Vasc Surg 1993; 7: 534–539. [DOI] [PubMed] [Google Scholar]
  • 27.Sugita T, Watarida S, Katsuyama K, et al. Endotoxemia after elective surgery for abdominal aortic aneurysm and the effect of early oral feeding. J Cardiovasc Surg (Torino) 1998; 39: 547–549. [PubMed] [Google Scholar]
  • 28.van Deventer SJ, ten Cate JW, Tytgat GN. Intestinal endotoxemia. Clinical significance. Gastroenterology 1988; 94: 825–831. [DOI] [PubMed] [Google Scholar]
  • 29.van Deventer SJ, Knepper A, Landsman J, et al. Endotoxins in portal blood. Hepato-Gastroenterology 1988; 35: 223–225. [PubMed] [Google Scholar]
  • 30.Wilde CG, Seilhamer JJ, McGrogan M, et al. Bactericidal/permeability-increasing protein and lipopolysaccharide (LPS)-binding protein. LPS binding properties and effects on LPS-mediated cell activation. J Biol Chem 1994; 269: 17411–17416. [PubMed] [Google Scholar]
  • 31.Woodcock NP, Sudheer V, el Barghouti N, et al. Bacterial translocation in patients undergoing abdominal aortic aneurysm repair. Br J Surg 2000; 87: 439–442. [DOI] [PubMed] [Google Scholar]
  • 32.Fiddian-Green RG. Splanchnic ischaemia and multiple organ failure in the critically ill. Ann R Coll Surg Engl 1988; 70: 128–134. [PMC free article] [PubMed] [Google Scholar]
  • 33.Ziegler EJ, Fisher CJ Jr, Sprung CL, et al. Treatment of gram-negative bacteremia and septic shock with HA-1A human monoclonal antibody against endotoxin. A randomized, double-blind, placebo-controlled trial. The HA-1A Sepsis Study Group. N Engl J Med 1991; 324: 429–436. [DOI] [PubMed] [Google Scholar]
  • 34.Baumgartner JD, Glauser MP, McCutchan JA, et al. Prevention of gram-negative shock and death in surgical patients by antibody to endotoxin core glycolipid. Lancet 1985; 2: 59–63. [DOI] [PubMed] [Google Scholar]
  • 35.Elsbach P, Weiss J. Role of the bactericidal/permeability-increasing protein in host defence. Curr Opin Immunol 1998; 10: 45–49. [DOI] [PubMed] [Google Scholar]
  • 36.Biffl WL, Moore EE, Moore FA, et al. Interleukin-6 delays neutrophil apoptosis. Arch Surg 1996; 131: 24–29. [DOI] [PubMed] [Google Scholar]
  • 37.Froon AH, Greve JW, Van der Linden CJ, et al. Increased concentrations of cytokines and adhesion molecules in patients after repair of abdominal aortic aneurysm. Eur J Surg 1996; 162: 287–296. [PubMed] [Google Scholar]
  • 38.Engles RE, Huber TS, Zander DS, et al. Exogenous human recombinant interleukin-10 attenuates hindlimb ischemia-reperfusion injury. J Surg Res 1997; 69: 425–428. [DOI] [PubMed] [Google Scholar]
  • 39.Svoboda P, Kantorova I, Ochmann J. Dynamics of interleukin 1, 2, and 6 and tumor necrosis factor alpha in multiple trauma patients. J Trauma 1994; 36: 336–340. [DOI] [PubMed] [Google Scholar]
  • 40.Tassiopoulos AK, Carlin RE, Gao Y, et al. Role of nitric oxide and tumor necrosis factor on lung injury caused by ischemia/reperfusion of the lower extremities. J Vasc Surg 1997; 26: 647–656. [DOI] [PubMed] [Google Scholar]
  • 41.Hess PJ, Seeger JM, Huber TS, et al. Exogenously administered interleukin-10 decreases pulmonary neutrophil infiltration in a tumor necrosis factor-dependent murine model of acute visceral ischemia. J Vasc Surg 1997; 26: 113–118. [DOI] [PubMed] [Google Scholar]
  • 42.Boyd AJ, Rubin BB, Walker PM, et al. A CD18 monoclonal antibody reduces multiple organ injury in a model of ruptured abdominal aortic aneurysm. Am J Physiol 1999; 277: H172–182. [DOI] [PubMed] [Google Scholar]
  • 43.Welbourn R, Goldman G, O’Riordain M, et al. Role for tumor necrosis factor as mediator of lung injury following lower torso ischemia. J Appl Physiol 1991; 70: 2645–2649. [DOI] [PubMed] [Google Scholar]
  • 44.Adam DJ, Ludlam CA, Ruckley CV, et al. Coagulation and fibrinolysis in patients undergoing operation for ruptured and nonruptured infrarenal abdominal aortic aneurysms. J Vasc Surg 1999; 30: 641–650. [DOI] [PubMed] [Google Scholar]
  • 45.Brown GE, Reiff J, Allen RC, et al. Maintenance and down-regulation of primed neutrophil chemiluminescence activity in human whole blood. J Leukoc Biol 1997; 62: 837–844. [DOI] [PubMed] [Google Scholar]
  • 46.Brown GE, Silver GM, Reiff J, et al. Polymorphonuclear neutrophil chemiluminescence in whole blood from blunt trauma patients with multiple injuries. J Trauma 1999; 46: 297–305. [DOI] [PubMed] [Google Scholar]
  • 47.Sasaki S, Takigami K, Kunihara T, et al. Abdominal aortic aneurysms in aged patients: analysis of risk factors in non-ruptured cases. J Cardiovasc Surg (Torino) 1999; 40: 1–5. [PubMed] [Google Scholar]
  • 48.Punch J, Rees R, Cashmer B, et al. Acute lung injury following reperfusion after ischemia in the hind limbs of rats. J Trauma 1991; 31: 760–765. [DOI] [PubMed] [Google Scholar]
  • 49.Welbourn R, Goldman G, Kobzik L, et al. Role of neutrophil adherence receptors (CD 18) in lung permeability following lower torso ischemia. Circ Res 1992; 71: 82–86. [DOI] [PubMed] [Google Scholar]
  • 50.Leeman M. Pulmonary hypertension in acute respiratory distress syndrome. Monaldi Arch Chest Dis 1999; 54: 146–149. [PubMed] [Google Scholar]
  • 51.Klausner JM, Paterson IS, Valeri CR, et al. Limb ischemia-induced increase in permeability is mediated by leukocytes and leukotrienes. Ann Surg 1988; 208: 755–760. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Deeb GM, Grum CM, Lynch MJ, et al. Neutrophils are not necessary for induction of ischemia-reperfusion lung injury. J Appl Physiol 1990; 68: 374–381. [DOI] [PubMed] [Google Scholar]
  • 53.Klausner JM, Paterson IS, Goldman G, et al. Thromboxane A2 mediates increased pulmonary microvascular permeability following limb ischemia. Circ Res 1989; 64: 1178–1189. [DOI] [PubMed] [Google Scholar]
  • 54.Klausner JM, Paterson IS, Kobzik L, et al. Leukotrienes but not complement mediate limb ischemia-induced lung injury. Ann Surg 1989; 209: 462–470. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55.Goldman G, Welbourn R, Klausner JM, et al. Thromboxane mediates diapedesis after ischemia by activation of neutrophil adhesion receptors interacting with basally expressed intercellular adhesion molecule-1. Circ Res 1991; 68: 1013–1019. [DOI] [PubMed] [Google Scholar]

Articles from Annals of Surgery are provided here courtesy of Lippincott, Williams, and Wilkins

RESOURCES