INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Acute respiratory distress syndrome (ARDS) is a clinically and pathophysiologically complex syndrome of acute lung inflammation characterized by severe dyspnea, hypoxemia refractory to supplemental oxygen, decreased lung compliance, roentgenographic findings of diffuse pulmonary infiltrates, and normal artery occlusion pressure (1). The mortality rate in patients with ARDS is still in excess of 50% despite recent advances in intensive care (2, 3). Although the pathology of this syndrome is poorly defined, a variety of inflammatory mediators is thought to be involved in this complex disease (4).

Phospholipase A₂ (PLA₂) comprises a superfamily of enzymes that catalyze the hydrolysis of ester bonds at the sn-2 position of membrane phospholipids. The role of PLA₂ is well documented in the rate-limiting step in the production of lipid inflammatory mediators, including eicosanoids (e.g., prostaglandins, leukotrienes), lysophospholipids, and platelet activating factor (PAF) (5). The PLA₂ enzymes occur as cytosolic (cPLA₂) and secretory (sPLA₂) types. Although new sPLA₂s have been identified recently (5), well-characterized mammalian sPLA₂ enzymes comprise group I (pancreatic type, group IB) and group II (nonpancreatic type, group IIA) PLA₂. Group IB PLA₂ has been identified in the pancreas, spleen, and lung, whereas group IIA PLA₂ has been shown to be released from a variety of cells and tissues (6, 7).

Local and systemic levels of group IIA PLA₂ are elevated in numerous inflammatory conditions, including sepsis, septic shock, acute pancreatitis, rheumatoid arthritis, and pulmonary diseases such as ARDS (6, 7). Of particular interest, levels of group IIA PLA₂ often correlate positively with the severity of ARDS (8). In animals, intratracheal administration of sPLA₂ can induce lung injury with interstitial and alveolar edema, accumulation of inflammatory cells, and alveolar wall thickening, all of which are pathological features typical of those seen in the lung in ARDS (9, 10). These observations have suggested that sPLA₂, especially group IIA, plays an exacerbating role in the inflammatory cascade, making it an important target for the treatment of inflammatory disorders. Although some sPLA₂ inhibitors have been reported, none of them has yielded convincing evidence on the role of group IIA PLA₂ in inflammation. Very recently, a novel potent and specific inhibitor of group IIA PLA₂, S-5920/LY315920Na ([[3-(aminooxoacetyl)-2-ethyl-1-(phenylmethyl)-1 H-indol-4-yl]oxy] acetate) was codeveloped (by Shionogi & Co., Ltd, Osaka, Japan and Eli Lilly & Company, Indianapolis, IN) to address the complex issue of the role of sPLA₂ isotypes in various inflammatory conditions (11). S-5920/LY315920Na, which represents the first in a new class of anti-inflammatory agents carrying the designation SPI (secretary PLA₂ inhibitor), acts at the active site of the human group IIA PLA₂ similar to findings on structural analogues that were cocrystallized in the active site of the human group IIA PLA₂ (12). Selectivity of S-5920/LY315920Na was apparent by the 40-fold weaker activity against human, group IB, pancreatic sPLA₂ which has the same catalytic mechanism as the human group IIA PLA₂. Furthermore, S-5920/LY315920Na failed to inhibit the catalytic activity of cPLA₂ (12).

To investigate the role of group IIA PLA₂ in acute lung injury (ALI), we examined the efficacy of S-5920/LY315920Na in oleic acid (OA)-induced lung injury in rabbits (13). OA causes morphological and cellular changes similar to the findings seen in patients with ARDS, and thus has been extensively used to evaluate the efficacy of various treatment strategies in patients with ARDS (14). For this assessment, we compared the OA-induced changes of PLA₂ activity in bronchoalveolar lavage fluid (BALF) and of pulmonary function and histology in rabbits treated with or without human group IIA PLA₂ inhibitor.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

The current study was conducted according to the guidelines worked out by the Animal Care Review Board of Kobe University School of Medicine. Male Japanese white rabbits weighing 2.1 to 2.7 kg obtained from Charles River (Tokyo, Japan) were used in this study. All animals were caged at room temperature and allowed to eat and drink ad libitum.

Drugs and Chemicals

S-5920/LY315920Na (Lot No. 268SB4) was synthesized at Lilly Research Laboratories (Indianapolis, IN) (11). Ketamine hydrochloride injection and pancuronium bromide injection were purchased from Sankyo (Tokyo, Japan). Thiamylal sodium injection was from Yoshitomi (Osaka, Japan). S-5920/LY315920Na was dissolved in 5% glucose and injected as a bolus in a volume of 1 ml/kg and then infused in a volume of 1 ml/kg/h. OA was purchased from Wako Pure Chemicals (Osaka, Japan). All other reagents were made of analytical grade or better.

Experimental Groups

The animals were randomly assigned to each experimental group before the surgical procedure. Forty-two rabbits were randomly divided into five groups as follows: saline control group (n = 8), an OA control group (n = 10), and three groups given S-5920/LY315920Na (three different doses, n = 8, respectively). The saline control group received 0.1 ml/kg/h of saline for 2 h, the OA control group and S-5920/LY315920Na group received 0.1 ml/kg/h of OA for 2 h. The S-5920/LY315920Na groups were divided into high-, medium-, and low-dose groups: the high-dose group received bolus injection of 1.0 mg/kg intravenously of S-5920/LY315920Na at 30 min before the start of OA infusion, followed by a sustaining infusion of the drug at 0.5 mg/kg/h intravenously, the medium-dose group was given 0.2 mg/kg bolus and 0.1 mg/kg/h infusion, and the low-dose group was given 0.04 mg/kg bolus and 0.02 mg/kg/h infusion. Saline and OA control groups received bolus injection and infusion of similar volume of the 5% glucose solution instead of S-5920/LY315920Na.

Experimental Protocol

The preparation of the animals was described in detail elsewhere (15). Briefly, rabbits were initially anesthetized using ketamine hydrochloride (25 mg/kg, intravenously), the anesthesia was then maintained with continuous infusion of ketamine hydrochloride at a rate of 0.5 mg/kg/h. A tracheotomy was performed aseptically and a 3.5-mm uncuffed endotracheal tube was inserted and tied in place, with 2% lidocaine used for local anesthesia. Immediately after the injection of pancuronium bromide (4 mg/kg, intravenously) for neuromuscular blockade, animals were mechanically ventilated with a pressure-limited ventilator (Model IV-100B; Sechrist Co., Anaheim, CA) as follows: an inspired oxygen fraction (FI_O2) of 0.8, initial respiratory rate (RR) of 30 breaths/min with the ratio of inspiratory to expiratory time (I:E ratio) of 1:2, and supplemented with a positive end-expiratory pressure (PEEP) of 2 cm H₂O. During the experiment period, the RR was adjusted to maintain the arterial Pa_CO2 between 35 and 45 mm Hg. Also, the FI_O2 was changed to 1.0 when Pa_O2 was less than 50 mm Hg. Tidal volume was set to 10 ml/kg body weight measured by pneumotachograph immediately after the onset of ventilation, by adjusting peak inspiratory pressure values. The animals were placed on a heating pad under a radiant heat lamp to that the body temperature could be kept at 37.8 to 40.2° C. Anesthesia and muscle relaxation were maintained with vascular infusions of lactated Ringer's solution (containing 4.3% glucose, 0.5% NaHCO₃, 0.5 mg/ml ketamine hydrochloride, and 0.04 mg/ml pancuronium bromide) via a catheter inserted though the marginal ear vein in order to eliminate spontaneous breathing. Via a femoral cutdown, a catheter was placed in the distal aorta to monitor arterial pressure and to obtain blood samples. The central venous pressure (Pcv) was also monitored via a catheter inserted though the right internal jugular vein.

Immediately after baseline measurement of lung mechanics, hemodynamics, peripheral leukocyte count, and arterial blood gas pressure, 5% glucose or S-5920/LY315920Na was injected as a bolus and infused until the rabbits were killed. At 30 min after the start of treatment with 5% glucose or S-5920/LY315920Na, 0.1 ml/kg/h of OA was infused for 2 h in the OA control and S-5920/LY315290Na groups. All rabbits were killed 6 h after the start of OA infusion by injection of an overdose of thiamylal. In all groups, arterial blood samples were obtained at 0, 1, 2, 3, 4, 5, and 6 h after the start of OA to determine blood gas, blood cell counts, plasma concentrations of S-5920/LY315920Na, plasma PLA₂ activity, and plasma level of interleukin 8 (IL-8).

Estimation of Parameters of ALI

Arterial blood gas and cell counts. During the experimental period, the obtained arterial blood specimens were analyzed for Pa_O2, Pa_CO2, and pH using ABL2 (Radiometer, Copenhagen, Denmark), and the number of peripheral leukocytes and platelets were measured with a Coulter counter (Coulter Electronics, Harkenden, UK). The alveolar-arterial oxygen tension difference (AaDO₂) was calculated using the standard alveolar gas equation with an assumed respiratory quotient of 1.0.

Lung mechanics. Lung mechanics were measured by the passive expiratory flow-volume technique as described previously (15). The airflow was measured with a Fleisch 00 pneumotachograph and a differential pressure transducer (Model MP-45; Validyne Engineering Corp., Northridge, CA). Airway pressure was measured at the proximal end of the pneumotachometer with a semiconductor pressure transducer (Model P-300 501G; Copal Electronics Corp., Tokyo, Japan). The volume was determined for each breath by digital integration of airflow using a respiration monitor (Aivision, Tokyo, Japan) and a personal computer (PC9801 VM11; NEC, Tokyo, Japan). The lungs were inflated and the airflow was interrupted at 20 cm H₂O. The occlusion was rapidly released after airway pressure reached a plateau. Compliance and resistance of the total respiratory system were then calculated using a personal computer.

Lung wet-dry weight ratio. At the end of the experiment, the thorax was opened and the heart and lungs were removed en bloc by observers unaware of the nature of the experiment. The left upper lobe was weighed and then dried to constant weight at 60° C for 24 h in an oven. The ratio of wet weight to dry weight (W/D weight ratio) was calculated to assess tissue edema.

Preparation of BALF and Measurements

Through the right mainstem bronchus, 60 ml of saline with ethylenediaminetetraacetic acid disodium (EDTA-2Na) (final concentration 0.77 mM) at 4° C was slowly infused and withdrawn. Indomethacin and phenindione (final concentration 10 µg/ml and 30 µg/ml, respectively) were added to the BALF to inhibit further metabolism of arachidonic acid (AA) to prostaglandins and leukotrienes during analysis. The BALF was analyzed for cell count and cell differentiation. A cytocentrifuged preparation (Cytospin 2; Shandon Southern Products, Runcorn, UK) of the BALF was stained with Wright-Giemsa for cell differentiation. The cells present in the fluid were counted with a Coulter counter (Coulter Electronics) and the Bürker-Türk method.

The fluid was centrifuged at 250 × g at 4° C for 10 min to remove the cells. The cell-free supernatant was divided into several aliquots and stored at 80° C until assayed. The following substances, metabolites, and mediators in the BALF were then measured: (1) protein concentrations were determined by using protein assay reagent (Pierce, Rockford, IL); (2) the concentration of thromboxane A₂ (TXA₂) was quantified by enzyme immunoassay (EIA; Amersham, Little Chalfont, UK) as TXB₂,the stable metabolite of TXA₂; (3) the concentrations of leukotriene B₄ (LTB₄) and leukotriene C₄/D₄/E₄ (LTC₄/D₄/ E₄) were quantified by EIA (Amersham); and (4) the concentration of IL-8 was determined by ELISA (Amersham). The assay of IL-8 was done using the previously described human ELISA kit that cross-reacts to rabbit IL-8, using recombinant rabbit IL-8 for the standard concentration curve (16).

Surfactant Phospholipids Analysis

Lipids were extracted from BALF according to the method of Casals and colleagues (17). The individual phospholipids were separated by two-dimensional chromatography on precoated activated silica-gel type 60 G thin-layer plates. The plates were developed in chloroform/ methanol/H₂O (72:25:3, vol/vol) and chloroform/methanol/acetic acid/ H₂O (90:40:12:2 vol/vol), respectively (17). Lipid area of chromatograms was visually detected by exposure to I₂ vapor. The spots corresponding to individual phospholipids were scraped off for phosphorus determination.

PLA₂ Activity Assay

The standard assay mixture contained 1 mM 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol (POPG) (Avanti Polar Lipids, Inc., Alabaster, AL), 2 mM sodium cholate, 100 mM Tris-HCl buffer (pH 8.0), 150 mM NaCl, 10 mM CaCl₂, 1 mg/ml bovine serum albumin (BSA), and the enzyme sample in a final volume of 100 µl. The substrate was used in the form of mixed micelles of sodium cholate/POPG at a molar ratio 2:1, obtained by a combination of evaporation under a stream of N₂, with drying in vacuo and addition of an appropriate amount of buffer and vortex mixing until the solution became clarified. The enzyme reactions were initiated by addition of the enzyme sample to the substrate mixture. Enzyme content and reaction time were adjusted to ensure linear kinetics in all experiments. The reaction was carried out at 40° C for the defined time periods and was stopped by adding 400 µl of Dole's reagent (2-propanol:n-heptane:2 N sulfuric acid, 40:10:1, vol/vol/vol), with 6 nmol of margaric acid (Nu-Chek-Prep, Inc., Elysian, MN) added as an internal standard. Fatty acids were extracted according to Dole's extraction system followed by silicic acid treatment (18). PLA₂ activity was determined according to the method of Tojo and colleagues (18) measuring 9-anthryldiazomethane (ADAM)-labeled fatty acid with high-performance liquid chromatography (HPLC).

Histopathological Examination

Shortly after the rabbits were killed (< 5 min), the left lower lobe was fixed by instillation of 10% glutaraldehyde solution through the left lower bronchus at 20 cm H₂O. The specimens were embedded in paraffin wax, stained with hematoxylin and eosin, and examined under a light microscope. Lung injury was scored 0 (no damage) to 4+ (maximal damage) according to combined assessments of alveolar congestion, hemorrhage and edema, infiltration/aggregation of neutrophils in the airspace or vessel wall, thickness of the alveolar wall, and hyaline membrane formation by four blinded observers. For immunohistochemistry, the lung tissues were fixed in 4% paraformaldehyde, dehydrated, and embedded in paraffin. Four-µm-thick sections were created, and endogenous peroxidase was blocked with 0.3% H₂O₂ in methanol. The tissue sections were treated with 5% normal rabbit serum followed by 0.05% trypsin and incubated overnight with 5 µg/ml of polyclonal antibodies against IL-8 at 4° C. The sections were then incubated for 30 min with 5 µg/ml of biotinylated rabbit anti-goat IgG (Vector Laboratories, Burlingame, CA), rinsed and incubated for 30 min with avidin-biotin-peroxidase complex (Vector Laboratories, Burlingame, CA). As a chromogen, diaminobenzidine was used. Counterstaining was performed using hematoxylin. Preimmune IgG was used as a control.

Plasma Concentration of S-5920/LY315920Na

Plasma concentration of S-5920/LY315920Na was determined using reverse-phase HPLC. Plasma sample (50 µl) was added to 100 µl of acetonitrile containing 0.1% trifluoroacetic acid (TFA) and incubated at room temperature for 10 min. After centrifugation at 10,000 × g for 5 min, an aliquot (40 µl) of the supernatant was injected onto LiChroCART 250-4 Superspher 100 RP-18 column (Merck, Darmstadt, Germany). The mobile phase was a mixture of acetonitrile-water-TFA in the ratio of 40:60:0.1 (vol/vol/vol). The flow rate was constant at 1.0 ml/min and the column oven temperature was 26° C. The column effluent was monitored at 254 nm. A calibration curve was made by adding known amounts of S-5920/LY315920Na to plasma without drug, extracting it, and analyzing it using HPLC as previously described. The concentration of S-5920/LY315920Na was determined from the calibration curve.

Statistical Analysis

Data are expressed as mean ± SEM, except the lung injury score which is given as a median. Data were statistically analyzed using Student's nonpaired t test for comparison with the saline control group and data on S-5920/LY315920Na groups were compared with the OA control group using Dunnett's test. The cumulative ² test was used for the lung injury score. p < 0.05 was deemed significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Prophylactic Effect of S-5920/LY315920Na on Parameter of ALI

Oxygenation. As shown in Figure 1, the Pa_O2/FI_O2ratio in the saline control group remained at a level exceeding 500 mm Hg, whereas in the OA control group it gradually decreased from 1 h and was significant compared with the saline control group at 2 to 6 h after the OA infusion (Figure 1A). It satisfied the criteria of ALI ( 300 mm Hg) at 3 h (270.6 ± 40.8 mm Hg, p < 0.01) and further decreased to that of ARDS ( 200 mm Hg) at 6 h (182.0 ± 42.9 mm Hg, p < 0.01) (19, 20). High dose of S-5920/LY315920Na caused significant recovery of the OA-induced decrease of Pa_O2/FI_O2 ratio at 3 to 6 h (451.2 ± 27.7 mm Hg at 3 h, p < 0.01; 412.1 ± 43.9 mm Hg at 6 h, p < 0.01). Medium and low doses of S-5920/LY315920Na also significantly attenuated the fall of the Pa_O2/FI_O2 ratio at 3 to 6 h in a dose-dependent manner (454.6 ± 33.6 mm Hg at 3 h, p < 0.01; 395.7 ± 55.4 mm Hg at 6 h, p < 0.05, in the medium-dose group; 421.2 ± 37.1 mm Hg at 3 h, p < 0.05; 337.5 ± 55.3 mm Hg at 6 h in the low-dose group) (Figure 1A).

View larger version (23K): [in this window] [in a new window]   Figure 1.   Effect of S-5920/LY315920Na on changes of blood oxygenation and lung mechanics in OA-induced ALI in rabbits (A) PaO2/FIO2 ratio. (B) AaDO2. (C ) Lung compliance. Details of the experimental groups are presented in METHODS; open triangles, saline control group, n = 8; open circles, OA control group, n = 10; closed circles, high-dose S-5920/LY315920Na group, 1 mg/kg intravenous bolus plus 0.5 mg/kg/h intravenous infusion, n = 8; open squares, medium-dose S-5920/LY315920Na group, 0.2 mg/ kg intravenous bolus plus 0.1 mg/kg/h intravenous infusion, n = 8; closed squares, low-dose S-5920/LY315920Na group, 0.04 mg/ kg intravenous bolus plus 0.02 mg/kg/h intravenous infusion, n = 8. Treatment with S-5920 was commenced at 30 min before the start of OA infusion for 2 h.  p < 0.05,  p < 0.01 versus saline control group, ¶  p < 0.05, ¶¶ p < 0.01 versus time 0 (Student's t test). *p < 0.05, **p < 0.01 versus OA control group (Dunnett's test). Each point represents mean ± SEM.

In contrast to the PaO₂/FIO₂ ratio, AaDO₂ increased in the OA control group compared with the saline control group (362.3 ± 35.8 versus 106.5 ± 10.5 mm Hg at 6 h, p < 0.01) (Figure 1B). High dose of S-5920/LY315920Na significantly attenuated the OA-induced increase of AaDO₂ at 3 to 6 h (189.8 ± 30.9 mm Hg at 6 h, p < 0.01). Medium and low doses of S-5920/LY315920Na also significantly attenuated the increase of AaDO₂ at 3 to 6 h in a dose-dependent manner (200.7 ± 43.8 mm Hg at 6 h, p < 0.05, in the medium-dose group 227.0 ± 35.3 mm Hg at 6 h, p < 0.05, in the low-dose group) (Figure 1B). During the experiments, plasma S-5920/LY315920Na concentrations were maintained around 1,000, 200, and 100 ng/ml in high-, medium-, and low-dose groups, respectively, from 0 to 6 h after OA (Table 1).

Mean arterial pressure was not significantly affected at any point in any of the groups throughout the observation periods and there was no difference in Pcv among the five groups, indicating that all the observed changes of the lung functions were not cardiogenic (data not shown).

Lung compliance. Lung compliance immediately after the start of mechanical ventilation was not different among the five groups (Figure 1C). This variable was lower in the OA control group than in the saline control group at 2 h after OA infusion, and thereafter significantly decreased at 3 to 6 h after OA injection (0.78 ± 0.10 ml/cm H₂O at 6 h, p < 0.01). In the high-dose S-5920/LY315920Na group, this decrease of compliance was significantly attenuated compared with the OA control group (1.20 ± 0.07 ml/cm H₂O at 6 h, p < 0.05). Medium and low doses of S-5920/LY315920Na also attenuated the decrease of compliance at 3 to 6 h in a dose-dependent manner (1.11 ± 0.11 ml/cm H₂O at 6 h in the medium-dose group and 1.07 ± 0.11 ml/cm H₂O at 6 h in the low-dose group) (Figure 1C).

Accumulation of edema fluid and vascular permeability in the lung. The lung W/D weight ratio markedly rose in rabbits receiving OA compared with those receiving saline (6.49 ± 0.21 versus 4.77 ± 0.06, p < 0.01) (Figure 2A). The ratio in the lung lobe was attenuated in a dose-dependent manner with S-5920/LY315920Na pretreatment. Furthermore, a high dose of S-5920/LY315920Na significantly inhibited the increase in the W/D weight ratio (5.64 ± 0.16, p < 0.01) (Figure 2A). Protein concentrations in the supernatant of BALF were also higher in OA-treated rabbits than in the saline control rabbits (4.53 ± 0.57 versus 1.36 ± 0.18 mg/ml, p < 0.01) (Figure 2B). The high dose of S-5920/LY315920Na significantly decreased the protein concentrations in OA-treated rabbits (2.55 ± 0.33 mg/ml, p < 0.05) (Figure 2B).

View larger version (20K): [in this window] [in a new window]   Figure 2.   Effect of S-5920/LY315920Na on (A) accumulation of lung edema and (B) vascular permeability at 6 h after OA infusion. Lung lobe and BALF were collected at 6 h after the start of OA infusion. W/D ratio and BALF protein concentration were determined as described in METHODS. Results are expressed as mean ± SEM from eight rabbits except for OA control (n = 10).  p < 0.01 (Student's t test), *p < 0.05, **p < 0.01 (Dunnett's test).

Cells in peripheral blood and BALF. Peripheral leukocyte counts dramatically dropped with infusion of OA, reached the lowest level 3 h after the start of OA infusion, and remained low compared with the saline control group during the experiment (Table 2). The total number of leukocytes in BALF was significantly higher in the OA control group compared with that in the saline control group, suggesting that the neutrophils marginating to the pulmonary capillary endothelium had infiltrated into alveolar spaces from peripheral blood. S-5920/ LY315920Na led to a significant recovery of the decrease in peripheral leukocyte counts at 6 h. Total BALF leukocyte counts tended to decrease in the high-dose S-5920/LY315920Na group (Table 2). Differential counts of the leukocytes in BALF revealed that cells in the saline control group were mostly macrophages. In contrast, the polymorphonuclear leukocyte (PMN) total leukocyte ratio (%PMN) markedly increased in the OA control group. The highest dose of S-5920/LY315920Na treatment slightly reduced this ratio concomitant with the slight attenuation of increase in BALF PMN counts (Table 2). There was no significant change in peripheral platelet counts among all five treatment groups (Table 2).

Histopathology. Light microscopic findings in the OA control group included hemorrhage and edema, thickened alveolar septum, formation of hyaline membranes, and the existence of inflammatory cells in alveolar spaces (Figures 3A and 3B). In contrast, these changes were far less pronounced in their high-dose S-5920/LY315920Na group (Figure 3C).

View larger version (76K): [in this window] [in a new window]   View larger version (19K): [in this window] [in a new window]   Figure 3.   Effect of S-5920/LY315920Na on lung injury in rabbits at 6 h after OA infusion. Representative photomicrographs showing hematoxylin and eosin staining in (A) saline control group, (B) OA control group, and (C ) high-dose S-5920/LY315920Na group. (D) Lung injury score: lung injury was scored 0 (no damage) to 4+ (maximal damage) according to the criteria described in METHODS. Results are expressed as median (bars) from eight rabbits except for the OA control (n = 10).  p < 0.05,  p < 0.01 (cumulative 2-test).

The results of the grading of lung damage are summarized in Figure 3D. Infusion of OA caused extensive morphologic lung damage in the OA control group and less damage in all three dosage groups given S-5920/LY315920Na. Moreover, the ALI score for the high-dose S-5920/LY315920Na group was significantly lower than that for the OA control group (p < 0.05).

PLA₂ Activity in BALF

Plasma PLA₂ activity was not changed by OA infusion (data not shown). In contrast, PLA₂ activity in BALF obtained from OA-infused rabbits markedly increased compared with the saline control group (8.13 ± 1.07 nmol/min/ml versus 1.48 ± 0.10 nmol/min/ml, p < 0.01) (Figure 4A). This increase of PLA₂ activity was inhibited by S-5920/LY315920Na in a dose-dependent manner. The high dose of S5920/LY315920Na completely blocked the elevated PLA₂ activity to levels of the saline control group (1.74 ± 0.40 nmol/min/ml, p < 0.01). The medium and low dose of S-5920/LY315920Na also significantly attenuated the increase in BALF PLA₂ activity (3.74 ± 0.69 nmol/ min/ml, p < 0.01, for the medium dose; 4.67 ± 0.63 nmol/min/ ml, p < 0.01, for the low dose) (Figure 4A). The remainder of BALF PLA₂ activity in the medium- and low-dose S-5920/ LY315920Na-treated rabbits was blocked by S5920/LY315920Na to the saline control group level in vitro (data not shown). The BALF PLA₂ activity from the OA control group was strongly blocked by S-5920/LY315920Na with an inhibitory concentration of 50% (IC₅₀) of 14.1 ± 1.5 nM in vitro as well as by 5 mM EDTA, suggesting that PLA₂ activity in the BALF was mainly due to Ca⁺⁺-dependent 14-kD group IIA PLA₂ (Figure 4B).

View larger version (19K): [in this window] [in a new window]   Figure 4.   Effect of S-5920/LY315920Na on PLA2 activity in BALF in vivo (A) and in vitro (B) in OA-induced ALI in rabbits. (A) In vivo analysis: BALF was collected at 6 h after the start of OA infusion. PLA2 activity in BALF was determined as described in METHODS. Results are expressed as mean ± SEM from eight rabbits except for OA control (n = 10).  p < 0.01, Student's t test, **p < 0.01, Dunnett's test. (B) In vitro analysis: BALF was obtained from OA control animals and incubated with S-5920/LY315920Na in vitro. Results are expressed as mean ± SEM from seven rabbits in the OA control.

Correlation of PLA₂ Activity with Severity of OA-induced Lung Injury

The analyses of individual data from the five groups in terms of the severity of lung injury versus BALF PLA₂ activity revealed that the increase of PLA₂ activity in the BALF correlated well with the severity of lung injury at 6 h, as assessed by the fall of the Pa_O2/FI_O2 ratio (Figure 5A), decrease of lung compliance (Figure 5B), elevation of the W/D ratio (Figure 5C), and increase of protein concentration in BALF (Figure 5D).

View larger version (25K): [in this window] [in a new window]   Figure 5.   Significant correlation between PLA2 activity and parameters of ALI (A) PaO2/FIO2 ratio, (B) lung compliance, (C ) wet/dry weight ratio, (D) protein concentration in BALF. Experimental groups are the same as in Figure 1 (open triangles, saline control, n = 8; open circles, OA control, n = 10; closed circles, high-dose S-5920/LY315920Na, n = 8; open squares, medium-dose S-5920/LY315920Na, n = 8; closed squares, low-dose S-5920/LY315920Na, n = 8).

Analysis of BALF

Table 3 presents a detailed analysis of phospholipids in BALF obtained at 6 h. There was a significant decrease in the percentage of phosphatidylglycerol (PG) and a significant increase in the lysophosphatidylcholine (lyso-PC) in OA-infused rabbits compared with the saline control, whereas no change was observed in the percentage of phosphatidylcholine (PC), phosphatidylethanolamine (PE), and sphingomyelin (SM). These changes in BALF phospholipids were attenuated in a dose-dependent manner with S-5920/LY315920Na pretreatment (Table 3, Experiment 1). Furthermore, high dose of S-5920/LY315920Na significantly attenuated the decrease in PG content caused by OA infusion and tended to attenuate the increase in lyso-PC content caused by OA infusion (Table 3, Experiment 1). The high dose of S-5920/LY315920Na did not affect BALF phospholipids composition in normal saline control animals except for the significant decrease of lyso-PC (Table 3, Experiment 2). In addition, S-5920/LY315920Na had no effect on lung function such as Pa_O2/FI_O2 ratio, lung compliance, and W/D ratio of normal saline control animals (data not shown).

The TXB₂ level in BALF markedly increased in the OA control group compared with that in the saline control group (386.2 ± 70.9 versus 47.1 ± 9.5 pg/ml, p < 0.01). Treatment with S-5920/LY315920Na reduced TXA₂ generation in BALF in a dose-dependent manner with significant reduction to 145.4 ± 38.2 pg/ml (p < 0.05) with the high dose of S-5920/ LY315920Na (Figure 6A). In addition, LTB₄ concentration in BALF significantly increased in the OA control group compared with that in the saline control group (8.0 ± 1.2 versus 3.5 ± 0.4 pg/ml, p < 0.01). A high dose of S-5920/LY315920Na attenuated this increase of LTB₄ (4.3 ± 0.7 pg/ml, p < 0.05, Student's nonpaired t test) (Figure 6B). Peptide leukotrienes (LTC₄/ D₄/E₄) in BALF did not increase in the OA control group compared with the saline control group (data not shown).

View larger version (21K): [in this window] [in a new window]   Figure 6.   Effect of S-5920/LY315920Na on BALF eicosanoid concentration in OA-induced ALI in rabbits. BALF was collected at 6 h after the start of AO infusion. Eicosanoids were measured by EIA for TXB2 (A) and LTB4 (B). Results are expressed as mean ± SEM from eight rabbits except for OA control (n = 10).  p < 0.05,  p < 0.01 (Student's t test), *p < 0.05 (Dunnett's test).

The IL-8 concentration in BALF significantly increased to 12.7 ± 3.0 ng/ml in OA-treated rabbits (p < 0.01) (Figure 7A). The high and medium doses of S-5920/LY315920Na significantly suppressed this increase to 4.1 ± 1.4 ng/ml (p < 0.05) and to 4.5 ± 1.6 ng/ml (p < 0.05), respectively (Figure 7A). Furthermore, as shown in Figure 7B, the marked increase of the IL-8 concentration in plasma was also observed with a peak around 2 to 3 h after the OA infusion (14.8 ± 4.0 ng/ml to 14.8 ± 4.2 ng/ml, p < 0.05 versus saline control group). IL-8 decreased gradually thereafter, but a significant elevation of IL-8 persisted for 6 h (4.3 ± 1.2 ng/ml, p < 0.05) in comparison to the saline control group. Treatment with S-5920/LY315920Na tended to inhibit this increase in a dose-dependent manner at 1 to 6 h with a significant decrease to the saline control level at 6 h (0.4 ± 0.2 ng/ml, p < 0.05, versus OA control) with a high dose of S-5920/LY315920Na (Figure 7B).

View larger version (20K): [in this window] [in a new window]   Figure 7.   Effect of S-5920/LY315920Na on BALF (A) and plasma (B) IL-8 concentration in OA-induced ALI in rabbits. (A) BALF was collected at 6 h after the start of OA infusion. IL-8 concentration was measured by ELISA. Results are expressed as mean ± SEM from eight rabbits except for OA control (n = 10).  p < 0.01 (Student's t test), *p < 0.05 (Dunnett's test). (B) Plasma samples were collected at various times post-OA infusion and measured by ELISA. Experimental groups are the same as in Figure 1 (open triangles, saline control, n = 8; open circles, OA control, n = 10; closed circles, high-dose S-5920/LY315920Na, n = 8; open squares, medium-dose S-5920/LY315920Na, n = 8; closed squares, low-dose S-5920/LY315920Na, n = 8). p < 0.05 versus saline control group (Student's t test). *p < 0.05 versus OA control group (Dunnett's test). Each point represents mean ± SEM.

Immunohistochemistry

To estimate the tissue localization of IL-8 in the lung with OA, immunohistochemical staining was done for lungs obtained at the time of the peak plasma level of IL-8 (2 h after OA infusion). Staining specificity was assessed using preimmune IgG as a control antibody and staining was nil (Figure 8A). The results showed that IL-8-positive cells were both alveolar macrophages and infiltrating neutrophils. The majority of the cells producing IL-8 were neutrophils (Figures 8B and 8C).

View larger version (87K): [in this window] [in a new window]   Figure 8.   Immunohistochemical staining of the lung tissue for IL-8 at 2 h after OA infusion. Lungs from rabbits were prepared from 4% paraformaldehyde-fixed, paraffin-embedded lungs. Sections of the lung were incubated with anti-rabbit IL-8 IgG and detected as described in METHODS. (A) Staining was nil when pre-immune serum was used as the control. (B and C) IL-8-positive staining cells with brown dots were alveolar macrophage (M) and infiltrating neutrophils (N).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The key finding in this study is that prophylactic treatment with human group IIA PLA₂ inhibitor, S-5920/LY315920Na, significantly ameliorated OA-induced ALI in rabbits, as evidenced by significant improvement of hypoxemia and reduced compliance, the prevention of pulmonary edema, and a decrease in the leakage of proteins into alveolar airspaces, suggesting that group IIA PLA₂ is a crucial mediator in this model of ALI.

A recent study showed that the lung is the organ most vulnerable to the effects of OA, which is present in an amount nearly an order of magnitude greater than in other organs, and that both endothelial and alveolar epithelial cells were markedly more sensitive to OA injury than hepatocytes (21). These results suggest that lung injury caused by OA in the present study occurs in local (lung tissue) rather than systemic (other tissues) tissues. This assumption was further supported by our findings of an increase of PLA₂ activity in BALF but not in plasma. Also, there was a positive relationship between increased BALF PLA₂ activity and the pathology of the lung injury (Figure 5), which agrees well with a positive correlation between PLA₂ levels and the severity of lung disease in patients with ARDS (8).

PLA₂ activity is important for normal lung function as it is involved in the synthesis of lung surfactants. However, various structural and functional phospholipid constituents of the lung render this organ particularly sensitive to PLA₂, which may also be a potent mediator of ALI (6). Serious disturbances of surfactant composition and function, as well as reduced surfactant protein contents in BALF, have been described in animal models of ARDS (17) and in human ARDS patients (22). Surfactant dysfunction in vivo would lead to atelectasis, alteration of the permeability characteristics of the lung, fluid accumulation, decrease in lung compliance, and the progression and exacerbation of lung injury, eventually contributing to the morbidity and mortality occurring in ARDS (23). Several studies suggested that the phospholipid content and composition in BALF after OA-induced lung injury did not significantly differ from that in control animals despite increased protein content, but PG in BALF was markedly decreased in injured animals compared with control animals (17) as in ARDS patients (22). These findings agree with our present results. In terms of substrate specificity, the human group IIA PLA₂ preferentially hydrolyzes phospholipids in negatively charged membranes in the order of PG > PE > PC (6). Moreover, we found marked increase of PLA₂ activity in BALF (Figure 4A), suggested that degradation of surfactant in the lung is caused by group IIA PLA₂ and lysophospholipids are subsequently generated from the surfactant. Lysophospholipids are well known to increase airway and capillary permeability and alter alveolar epithelium barrier function (19). These results indicate that S-5920/LY315920Na caused recovery of the decrease of the Pa_O2/FI_O2 ratio induced by OA infusion, probably by inhibiting the degradation of PG and reducing the production of cytotoxic phospholipids such as lysophospholipids from lung surfactant by the enzyme action of group IIA PLA₂. Indeed, S-5920/LY315920Na attenuated both the decrease of PG and increase of lysoPC in BALF in the present study (Table 3). In addition, the importance of surfactant PG in alveolar stability is well recognized because lung surfactant contains an extremely high level of PG (5 to 10% of total phospholipids) in comparison to other tissues (23), and the lack of surfactant PG is closely related to respiratory distress syndrome of newborns (20), further supporting the previously mentioned efficacy of S-5920/LY315920Na.

In the present study, the levels of TXB₂ (the stable metabolite of TXA₂) and LTB₄ were increased in BALF, indicating the activation of PLA₂ in the lung in OA-induced lung injury. Furthermore, S-5920/LY315920Na inhibited the production of both TXB₂ and LTB₄. These results suggest that group IIA PLA₂ actually participates in the AA release in the present model. A number of studies have demonstrated the increase of TXA₂ production via the cyclooxygenase (COX) pathway in plasma and BALF after lung injury induced by OA (14). The role of TXA₂ has been inferred from study with COX inhibitor in OA-induced lung injury. Ibuprofen pretreatment blocked the release of TXA₂ after OA injection, and concomitantly prevented the increase in pulmonary artery pressure and pulmonary vascular resistance. Thus, most studies indicate that TXA₂ causes pulmonary venous hypertension and results in enhanced edema formation (14). In the present study, we found that S-5920/LY315920Na significantly inhibited the OA-induced TXA₂ generation and lung edema (increase in the W/D ratio) as observed with ibuprofen treatment. However, clinical trial of ibuprofen failed to elicit beneficial effects in patients with severe sepsis (24), and it is recently recommended that ibuprofen not be used in patients with ARDS (2). This might be explained by the increase of leukotrienes by COX inhibition, which is enzymatically derived from AA via the 5-lipoxygenase (5-LO) pathway, through the shunting of AA from the COX pathway to the 5-LO pathway in patients. Thus S-5920/LY315920Na has an advantage over COX inhibitors or 5-LO inhibitors as a prophylactic or therapeutic agent for ARDS, as it does not shunt AA from the COX pathway to the 5-LO pathway and vice versa.

Infusion of OA caused increase of neutrophil counts in BALF concomitantly with decrease of leukocyte counts in peripheral blood, suggesting that the neutrophils at the margin of the pulmonary capillary endothelium had migrated into alveolar spaces from peripheral blood (Table 2). Neutrophils have been suggested to play an important role in OA-induced ALI, because their depletion resulted in decreased permeability of the lung in rats receiving OA (25). Also, activated neutrophils can injure the capillary endothelium by releasing oxygen radicals and proteinases, and can generate AA metabolites and inflammatory cytokines (4). In the present study, S-5920/LY315920Na tended to attenuate the OA-induced infiltration of neutrophils from peripheral blood into alveolar spaces (Table 2). This could partly be explained by the decrease of LTB₄ and IL-8 in BALF by this compound, because both LTB₄ and IL-8 are potent chemotactic agents for neutrophils (26, 27) and are known to elevate vascular permeability (26, 28). Blocking the activity of LTB₄ or IL-8 attenuated the endotoxin-induced ALI in pigs or rabbits, respectively (29, 30). Furthermore, IL-8 concentrations in BALF correlate well with the development of ARDS, and its increased levels early in the disease process have been correlated with mortality (31, 32), strongly implicating an important role for this chemokine in the pathogenesis of ARDS. Although the mechanism by which S-5920/LY315920Na inhibits IL-8 production is not clear, a recent study showed that peripheral blood neutrophils can be stimulated by LTB₄ to synthesize and secrete biologically active IL-8 (33). Also IL-8 gene expression and protein production can be upregulated by PAF in human neutrophils and monocytes (34, 35). Although PAF concentrations were not measured in the present study, reduction of lysophospholipid concentrations by S-5920/LY315920Na strongly suggests that PAF concentrations should also be reduced in BALF because PAF is produced by the action of acetylhydrolase on lysophospholipids (6). As to the source of IL-8, immunohistochemical analysis in our study showed that the major cells producing IL-8 were infiltrating neutrophils. Taken together, these findings suggest that blockade of the PLA₂ activity by S-5920/LY315920Na indirectly inhibited IL-8 production through inhibiting LTB₄ and probably PAF production from infiltrating neutrophils. This would also contribute, in part, to the protection by this compound of OA-induced dysfunction of the lung.

In conclusion, PLA₂ activity was markedly elevated in BALF obtained 6 h after OA infusion, and this elevation correlated well with the severity of lung injury. Treatment with S-5920/LY315920Na was very efficacious, clearly indicating that group IIA PLA₂ is a crucial factor in OA-induced lung injury, contributing to the degradation of lung surfactant, production of AA metabolites, IL-8 production, and infiltration of inflammatory cells into the lung tissue. Although further studies are required to investigate the therapeutic (post-treatment) efficacy of the drug against this lung injury, the present study indicates that strong and specific inhibitors of human group IIA PLA₂, such as S-5920/LY315920Na, could be a novel strategy for treating ARDS.