INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Tissue factor pathway inhibitor (TFPI) is a multivalent, Kunitz-type plasma protease inhibitor known to inhibit Factor Xa and Factor VIIa bound to tissue factor (1). Because depletion of endogenous TFPI has been found to sensitize rabbits to disseminated intravascular coagulation (DIC) induced by tissue factor or endotoxin (2, 3), TFPI plays a role in preventing DIC. In addition, Carr and colleagues (4) have shown that infusion of recombinant-TFPI (rTFPI) reduces the mortality, as well as the coagulation abnormalities in baboons injected with lethal doses of Escherichia coli. Since the lethal effect of E. coli is not reduced by attenuation of coagulopathic responses (5), anticoagulant effects of TFPI appear not to contribute to its reduction of lethal effects in the septic baboon model.

Acute respiratory distress syndrome (ARDS) is a serious complication in septic patients (6). Activated leukocytes have been shown to play a central role in tissue injury in ARDS by releasing various inflammatory mediators, such as cytokines and neutrophil proteases, that are capable of damaging endothelial cells (7, 8). Such activated leukocyte-induced pulmonary vascular injury is implicated in the pathogenesis of ARDS (9). Because ARDS is a critical pathologic condition in the development of multiple organ failure, and therefore adversely affects the outcome of patients with sepsis (6), prevention of ARDS might contribute to improving such patients' outcome.

Since rTFPI has been shown to inhibit interleukin (IL)-8 synthesis in the human whole-blood culture system (10), it is possible that rTFPI reduces lipopolysaccharide (LPS)-induced pulmonary vascular injury by inhibiting leukocyte activation.

In the present study we examined whether rTFPI reduces pulmonary vascular injury by inhibiting leukocyte activation, as well as the coagulation abnormalities in rats challenged with LPS. We further investigated the mechanism(s) by which rTFPI inhibits leukocyte activation in vitro.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Materials

Human rTFPI was obtained from the Chemo-Sero-Therapeutic Research Institute, Kumamoto, Japan. rTFPI was expressed in Chinese hamster ovary cells and purified according to the method described previously (11). rTFPI used in the present study was carboxyl-terminus truncated TFPI showing a single band at 42.5 kD on sodium dodecylsufate-polyacrylamide gel electrophoresis in the presence of 2-mercaptoethanol. The amino acid sequences of rTFPI were confirmed by sequence analyses of the peptides produced by digestion of rTFPI with lysyl endopeptidase. Inhibition of Factor Xa (F. Xa) and of tissue factor-bound Factor VIIa (TF-F.VIIa) by rTFPI was measured with Z-Pyr-Gly-Arg-MCA and Boc-Leu-Thr-Arg-MCA, respectively (12). The Ki values of rTFPI for the inhibition of F. Xa and TF-F.VIIa were 5.6 nM and 17.4 nM, respectively (12). Recombinant Factor VIIa treated with dansyl-glutamylglycylarginyl-chloromethyl ketone (DEGR-F.VIIa) was kindly provided by Dr. Bregengaard (Novo Nordisk, Gentofte, Denmark) (13). DX-9065a, a synthetic, potent anticoagulant and selective inhibitor of Factor Xa, was a generous gift from the Daiichi Pharmaceutical Co. Ltd. (Tokyo, Japan) (14). LPS (E. coli serotype 055:B5) was obtained from Difco (Detroit, MI). All other reagents used were of analytical grade.

In Vivo Experiments

Animal model of LPS-induced pulmonary vascular injury and coagulopathy. The study protocol was approved by the Kumamoto University Animal Care and Use Committee, and the care and handling of the animals used in the study were in accordance with U.S. National Institutes of Health guidelines. Adult, pathogen-free male Wistar rats (body mass 200 to 220 g) (Kyudo, Kumamoto, Japan) were given an intravenous injection of LPS (5 mg/kg) via the tail vein. rTFPI was injected 30 min before or after the injection of LPS. DX-9065a (3 mg/ kg, subcutaneously) and DEGR-F.VIIa (3 mg/kg, intravenously) were injected 30 min before the injection of LPS. Animals were anesthetized by intraperitoneal injection of pentobarbital (50 mg/kg) and were exsanguinated via the abdominal aorta. Blood and lung tissue samples were obtained at various times after LPS injection. Blood samples were collected in tubes containing 1:10 (vol/vol) of 3.8% (wt/ vol) sodium citrate and were centrifuged at 3,000 × g for 15 min. The lung vasculature was perfused through the right cardiac ventricle with 10 ml of saline. Acute lung injury (ALI) was evaluated with the pulmonary vascular permeability index and through the lung wet-to-dry weight ratio, and coagulopathy was estimated by measuring serum concentrations of fibrin and fibrinogen degradation products E (FDP[E]). Control animals received saline instead of the study drugs.

Determination of pulmonary vascular permeability index. ¹²⁵I-labeled bovine serum albumin (¹²⁵I-BSA) was prepared with Bolton-Hunter reagent (Amersham International plc., Little Chalfont, UK) and administered intravenously (2.0 × 10⁵ cpm/kg body mass) to rats at 5 min before the intravenous administration of a bolus dose (5 mg/kg) of LPS via the tail vein. rTFPI (1 mg/kg) was administered intravenously to animals 30 min before LPS administration (pretreatment), and was again (1 or 2 mg/kg) administered intravenously at 30 min after LPS administration (posttreatment). Saline, DX-9065a (3 mg/kg, intravenously), or DEGR-F.VIIa (3 mg/kg, intravenously) was administered 30 min before LPS injection. Blood and lung samples were obtained as described earlier for the animal model of LPS-induced ALI and coagulopathy. Amounts of radioactivity remaining within the tissue and blood were measured with a gamma scintillation counter (Model 5130; Packard Instrument, Downers Grove, IL). LPS-induced pulmonary vascular injury was assessed in terms of the increase in vascular permeability and was expressed as the permeability index (the ratio of the amount of radioactivity present in lung tissue to that in 1 ml of blood) (15).

To assess lung hemorrhage or pulmonary congestion induced by the administration of LPS, we measured the accumulation of ⁵¹Cr-labeled red blood cells in rats given LPS and in controls. Red blood cells were labeled with ⁵¹Cr (Amersham) as previously described (15). Animals were injected intravenously with ⁵¹Cr-labeled red blood cells (50 µl, containing 8.0 × 10⁴ cpm) 30 min before the injection of LPS. Rats were killed 6 h later, and the radioactivity in their lungs was measured.

Measurement of serum FDP(E). For measurement of serum FDP(E), blood samples were withdrawn from the aorta 6 h after the administration of LPS. The serum concentration of FDP(E) was determined by latex agglutination assay as previously described (15).

Measurement of lung wet-to-dry weight ratio. To determine the water content of the lung, the wet-to-dry weight ratio of the lungs was estimated at 6 h after administration of LPS. The lungs were dissected free of nonpulmonary tissue, weighed, and then dried to constant weight in an oven at 130° C. The wet-to-dry weight ratio was obtained by dividing the wet weight by the final weight of the dried lungs (15).

Measurement of lung TNF- and cytokine-induced neutrophil chemoattractant concentrations. Saline, rTFPI, DX-9065a, or DEGR-F.VIIa was administered 30 min before the injection of LPS (5 mg/kg). Lung samples were obtained at various times after injection of LPS. Lung samples were homogenized in 0.1 M phosphate buffer (pH 7.4) with a 0.05% solution of sodium azide. Lung levels of TNF- were determined with an enzyme-linked immunosorbent assay (ELISA) kit for murine TNF- (Genzyme, Cambridge, MA). Lung cytokine-induced neutrophil chemoattractant (CINC) levels were measured with an ELISA for rat granulocyte chemoattractant (GRO)/CINC-1 (Amersham).

Measurement of lung myeloperoxidase activity. At various times after administration of LPS, the lung vasculature was perfused through the right cardiac ventricle with 10 ml of cold saline. The lungs were then removed, and accumulation of leukocytes in the lung was evaluated by assessing myeloperoxidase (MPO) activity as previously described (15). Briefly, lung samples were homogenized in 6 ml of homogenization buffer containing 0.05 M phosphate buffer and 0.5% hexadecyltrimethylammonium bromide (pH 6.0). The homogenate was then sonicated and centrifuged at 4,500 × g for 30 min at 4° C. MPO activity in the supernatant was then determined. The test samples (0.1 ml) were mixed with 0.6 ml of 0.05 M phosphate buffer containing 0.167 mg/ml o-dianisidine dihydrochloride and 0.0005% hydrogen peroxide (pH 6.0). The change in absorbance at 460 nm was measured over a period of 1 min at 25° C in a spectrophotometer (DU-54; Beckman Instruments, Irvine, CA).

Histopathologic studies of the lungs. Histopathologic examination of the lungs was done 6 h after the administration of LPS. Saline or rTFPI was injected intravenously 30 min before the infusion of LPS, as described earlier. Samples were fixed with 10% formalin, embedded in paraffin, sectioned into 6-µm-thick slices, and stained with hematoxylin-eosin (H&E). Samples were analyzed by a pathologist who did not know whether the animals belonged to the experimental or control groups. The number of neutrophils was counted in 10 randomly selected fields per slide under oil at a magnification of ×1,000 by a pathologist who did not have knowledge of the animal grouping. Fields containing large vessels or bronchi were excluded. The number of neutrophils per field was counted and normalized to the number of alveoli per field to control for lung infiltration.

RNA isolation and northern blotting. For measurement of TNF- messenger RNA (mRNA), samples of lung tissue were taken from rats treated with saline or rTFPI (1 mg/kg, intravenously) at 30 min before administration of 5 mg/kg LPS, and were frozen at 80° C. Total RNA was isolated with the technique described by Chomczynski and Sacchi (16). Twenty micrograms of RNA per sample were loaded onto 1% agarose/formaldehyde denaturing gels, electrophoresed at a constant voltage, and transferred on to nylon filters. The filters used to assess TNF- mRNA were hybridized with ³²P-labeled DNA probe, using the random primed labeling technique described by Feinberg and Vogelstein (17). After hybridization at 42° C for 16 h, the filters were washed at room temperature in a solution containing 1% SDS/ 2× standard saline-citrate (SSC) and 1% SDS/0.2× SSC. Hybridization was assessed by radioisotope counting and autoradiography.

Plasma anticoagulant activities of animals given rTFPI, DX-9065a, and DEGR-F.VIIa. A dilute thromboplastin clotting assay was used to measure plasma anticoagulant activities of animals given LPS and anticoagulants (18). Plasma samples were obtained from animals pretreated with various anticoagulants at 90 min after LPS administration. In brief, 50 µl of normal human plasma was mixed with 10 µl of the rat plasma sample. Fifty microliters of 4,000-fold-diluted thromboplastin (Neoplastin Plus; Boehringer Mannheim, Mannheim, Germany) was added to the mixture. After incubation for 1 min at 37° C, clotting time was measured in the sample with a KC-1A coagulometer (Amelung, Lemgo, Germany) after addition of 50 µl of CaCl₂ (20 mM).

Determination of plasma rTFPI levels. Plasma levels of rTFPI were measured in rat plasma samples obtained 90 min after LPS administration through use of an ELISA method, as described previously (12).

In Vitro Experiments

Isolation and cultivation of human monocytes. Peripheral blood mononuclear cells obtained from healthy volunteer blood donors were isolated from their buffy coats as described previously (19). Monocytes were adjusted to a volume of 5.0 × 10⁵ /ml in RPMI-1640, and were then stimulated with LPS (20 ng/ml) for 16 h at 37° C in a humidified 5% CO₂ incubator in the presence or absence of various concentrations of rTFPI. After incubation, cell suspensions were centrifuged at 12,000 × g for 10 min. Concentrations of TNF- in supernatant fractions were determined with an ELISA kit for human TNF- (Otsuka Pharmaceutical, Tokyo, Japan).

Preparation of neutrophils from normal human blood. Heparinized venous blood obtained from normal volunteers was mixed with an equal volume of 2% Dextran solution and allowed to stand for 30 min to permit erythrocyte sedimentation. The supernatant was centrifuged, and the precipitate was collected. Neutrophils were isolated as described previously (19). Contaminating erythrocytes were removed by hemolysis with 0.2% NaCl for 25 s. The resulting preparation, which contained > 95% neutrophils, was washed twice with phosphate-buffered saline (PBS). Cell viability of > 95% was confirmed with the trypan blue dye exclusion test.

Release of neutrophil elastase. The neutrophil suspension (5,000 cells/µl) in PBS was mixed with formyl-Met-Leu-Phe (fMLP 1.0 µM) (Sigma Chemical Co., St. Louis, MO), 5 µg/ml of cytochalasin B (Sigma), and 2 mM calcium chloride in the presence or absence of rTFPI (19). Neutrophil elastase activity in supernatants was measured by using the chromogenic substrate S-2484 (Chromogenix AB, Stockholm, Sweden).

Measurement of superoxide radical production by neutrophils. Neutrophil production of superoxide radical (O₂) was measured with a chemiluminescence assay, using a luminescence reader (BLR-201; Aloka, Tokyo, Japan), as previously described (19). Neutrophil suspensions (5.0 × 10⁶ /ml in PBS) were mixed with 10 mg/ml of luminol and various concentrations of rTFPI for 5 min at 37° C. The mixture was stimulated with fMLP (1.0 µM), and changes in chemiluminescence activity were monitored continuously.

Measurement of intracellular Ca²⁺ concentration. The intracellular ionized calcium concentration ([Ca²⁺]i) was measured as previously described (20). Briefly, neutrophils isolated as described earlier were suspended at 1.0 × 10⁶ /ml in RPMI-1640 with 10% fetal calf serum and 2.5 µg/ml indo-1-acetoxymethyl ester (Dojindo Laboratories, Kumamoto, Japan) for 30 min at 37° C. Fluorescence emission was measured in a spectrophotometer (Hitachi 850; Hitachi Ltd., Tokyo, Japan), using an excitation wavelength of 331 nm and an emission wavelength of 410 nm. After equilibration of fluorescence to a stable baseline state, the cells were stimulated with fMLP (0.3 µM) and fluorescence assessment was continued.

Measurement of CD11b and CD18. Neutrophils were isolated from peripheral blood of healthy human volunteers as described earlier. The neutrophils were suspended in PBS (pH 7.4), and various concentrations of rTFPI were added. The samples were stimulated with fMLP (1.0 µM) for 30 min at 37° C. Anti-CD11b and anti-CD18 antibody binding to neutrophils was measured with a FACScan flow cytometer (Becton Dickinson, Sandy, UT), using the channel number (log scale) representing the mean fluorescence intensity per 1.0 × 10⁴ cells (21).

Statistical Analysis

Data are presented as mean ± SD. Results were compared through analysis of variance and Scheffe's post hoc test or the unpaired t test. A value of p < 0.05 was accepted as statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effects of rTFPI, DX-9065a, and DEGR-F.VIIa on Pulmonary Vascular Permeability, Lung Wet-to-Dry Weight Ratio, and Serum FDP(E) Levels in Rats Given LPS

To determine whether rTFPI reduces LPS-induced pulmonary vascular injury, we investigated its effect on the LPS-induced increase in pulmonary vascular permeability in rats. Pulmonary vascular permeability has been shown to increase significantly after administration of LPS, with the peak effect occurring from 4 to 8 h after its injection (15). The lung wet-to-dry weight ratio was significantly increased at 4 h after LPS administration, and remained increased throughout the 8-h observation period in LPS-treated animals (15). No significant increase was observed in the number of ⁵¹Cr-labeled red blood cells during the 8-h observation period after LPS injection in either perfused or nonperfused lungs (data not shown).

Intravenous administration of rTFPI (1 mg/kg) at 30 min before LPS administration prevented the increase in LPS- induced pulmonary vascular permeability that occurred at 6 h after LPS administration (Figure 1A). The increase in lung wet-to-dry weight ratio at 6 h after LPS administration was also attenuated in animals that received rTFPI (Figure 1B).

View larger version (20K): [in this window] [in a new window]   Figure 1.   Effects of rTFPI, DX-9065a and DEGR-F.VIIa on increases in pulmonary vascular permeability index (A), lung wet-to-dry weight ratio (B), and serum FDP(E) level (C ) in rats given lipopolysaccharide (LPS). rTFPI was administered intravenously 30 min before or after intravenous injection of LPS (5 mg/kg). Saline, DX-9065a (3 mg/kg, subcutaneously) or DEGR-F.VIIa (3 mg/kg, intravenously) was administered 30 min before intravenous injection of LPS. Pulmonary vascular permeability index, lung wet-to-dry weight ratio, and serum concentration of FDP(E) were determined 6 h after LPS administration. Control animals received saline instead of LPS. Data are expressed as mean ± SD. Numbers in parentheses indicate number of animals in each experiment. *p < 0.01 versus control; p < 0.01 versus LPS; p < 0.01 versus LPS + rTFPI (before LPS).

When rats were given LPS (5 mg/kg), intravenously, serum levels of FDP(E) increased over time and peaked 6 h after LPS administration (15). rTFPI significantly inhibited the increase in serum level of FDP(E) at 6 h after LPS administration (Figure 1C). Although intravenous administration of DX-9065a (3 mg/kg), a selective inhibitor of Factor Xa, and DEGR-F.VIIa (3 mg/kg) inhibited the increases in FDP(E) levels at 6 h after LPS administration, neither of these anticoagulants had any effect on LPS-induced pulmonary vascular injury (Figure 1).

Both permeability index and lung wet-to-dry weight ratio observed in animals given LPS and rTFPI were significantly lower than those of animals given LPS and DX-9065a and those of animals given LPS and DEGR-F.VIIa (Figure 1).

Although neither increases in pulmonary vascular permeability nor lung wet-to-dry weight ratio were prevented by posttreatment with 1 mg/kg of rTFPI, they were prevented by posttreatment with 2 mg/kg of rTFPI (Figures 1A and 1B). Serum levels of FDP(E) at 6 h after LPS administration were significantly decreased by posttreatment with 2 mg/kg of rTFPI (Figure 1C).

Effects of rTFPI on Changes in Pulmonary Histology Induced by LPS

Light-microscopic examination of lung tissue at 6 h after administration of LPS revealed interstitial edema (Figure 2B) that was not present in animals treated with saline alone (Figure 2A). The amount of LPS-induced pulmonary interstitial edema was less in animals pretreated with rTFPI (1.0 mg/kg, intravenously) (Figure 2C) than in those not so pretreated. Neither DX-9065a nor DEGR-F.VIIa had any effects on the LPS-induced histologic changes (data not shown).

View larger version (77K): [in this window] [in a new window]   Figure 2.   Effect of rTFPI on changes in the pulmonary histological findings induced by LPS in rats. Microscopic observations of lung tissue from animals treated with saline (A; original magnification: ×400), LPS (B; original magnification: ×400), and LPS plus 1 mg/kg rTFPI (C; original magnification: ×400). Saline or rTFPI was administered intravenously 30 min before LPS administration (5 mg/kg). Histologic examination of the lungs was performed 6 h after LPS administration. Interstitial edema and neutrophil infiltration were not observed in saline-treated animals (A), while they were observed in LPS-treated animals (B). Intravenous administration of rTFPI markedly reduced the LPS-induced pathologic changes (C ). Typical results are shown for five animals examined in each group.

Light-microscopic examination of fixed lung tissue revealed different numbers of neutrophils per alveolus at 6 h after LPS administration, whereas control lung tissue did not (Table 1). Neutrophil infiltration was mainly observed in the interstitial space of the lung in this animal model (15). Pretreatment with rTFPI (1 mg/kg, intravenously) significantly reduced the neutrophil infiltration at 6 h after LPS administration. Neither DX-9065a nor DEGR-F.VIIa reduced the number of infiltrating neutrophils (Table 1).

Plasma Anticoagulant Activities in Animals Given LPS and Various Anticoagulants

We determined plasma anticoagulant activities in animals given LPS and various anticoagulants at 90 min after they were given LPS (Figure 3). Plasma anticoagulant activities were significantly increased in animals given rTFPI, DX-9065a, and DEGR-F.VIIa over those of control animals (Figure 3). The plasma anticoagulant activities of animals given DX-9065a and DEGR-F.VIIa were significantly higher than those of animals given rTFPI (Figure 3).

View larger version (23K): [in this window] [in a new window]   Figure 3.   Plasma anticoagulant activities in rats given LPS and various anticoagulants. rTFPI (1 mg/kg), DX-9065a (3 mg/kg), and DEGR- F.VIIa (3 mg/kg) were injected 30 min before LPS administration. Plasma samples were obtained 90 min after LPS administration. Control animals received saline instead of anticoagulants. Anticoagulant activities of the plasma samples were measured according to the technique described in METHODS. Data are expressed as mean ± SD of triplicate experiments. Numbers in parentheses indicate the number of animals examined in each experiment. *p < 0.01 versus experiments without anticoagulants; p < 0.01 versus LPS + rTFPI.

Plasma Levels of rTFPI in Animals Challenged With LPS but Pretreated with rTFPI

Plasma levels of rTFPI were 2.54 ± 0.56 (mean ± SD) µg/ml (58.8 ± 9.3 nM) (n = 5) in animals pretreated with rTFPI (1 mg/ kg, intravenously) at 90 min after LPS administration.

Effects of rTFPI, DX-9065a, and DEGR-F.VIIa on LPS-Induced Increases in TNF-, CINC, and MPO Activity

Lung tissue levels of TNF- were increased after LPS administration, peaking at 1.5 h after LPS administration. Lung tissue levels of CINC were also increased with time after LPS administration, and were increased at up to 4 h after LPS administration. Lung MPO activities were increased after LPS administration, peaking at 1.5 h after LPS administration (data not shown). Intravenous administration of rTFPI (1 mg/kg) at 30 min before LPS challenge significantly inhibited the increases in lung tissue levels of TNF- (Figure 4A) and CINC (Figure 4B) observed 1.5 h and 4 h, respectively, after LPS administration. The LPS-induced increases in lung MPO activities (Figure 4C) at 1.5 h after LPS administration were also inhibited by pretreatment with rTFPI. However, neither DX-9065a nor DEGR-F.VIIa had any effects on these variables (Figure 4).

View larger version (20K): [in this window] [in a new window]   Figure 4.   Effects of rTFPI, DX-9065a, and DEGR-F.VIIa on LPS-induced increases in lung TNF- levels (A), lung CINC levels (B), and lung MPO activity (C ). Saline, rTFPI (1 mg/kg, intravenously), DX-9065a (3 mg/kg, subcutaneously), or DEGR-F.VIIa (3 mg/kg, intravenously) was administered 30 min before injection of LPS (5 mg/kg). Control animals received saline instead of LPS. Lung tissue levels of TNF- and MPO activity, an index of accumulation of pulmonary leukocytes, were measured 90 min after intravenous administration of LPS (5 mg/kg). The concentration of lung CINC was determined 4 h after LPS administration. Data are expressed as mean ± SD. Numbers in parentheses indicate the number of animals in each experiment. *p < 0.01 versus saline-treated animals; p < 0.01 versus LPS-treated animals; p < 0.01 versus LPS + rTFPI.

Effect of rTFPI on the Increased Expression of TNF- mRNA in the Lung In Vivo

The levels of expression of TNF- mRNA in lungs of animals increased over time after LPS administration. These increases in the expression of TNF- mRNA in the lungs were significantly inhibited in animals pretreated with rTFPI (Figure 5).

View larger version (28K): [in this window] [in a new window]   Figure 5.   Effect of rTFPI on LPS-induced increase in expression of lung TNF- mRNA. Saline (A) or rTFPI (1 mg/kg) (B) was administered 30 min before injection of LPS (5 mg/kg, intravenously). Lungs were removed 30, 60, or 90 min after LPS administration. Lung mRNA was detected by Northern blotting analysis. Five identical experiments performed independently gave similar results, and typical results are shown.

Effect of rTFPI on TNF- Production in LPS-Stimulated Monocytes In Vitro

To determine whether rTFPI directly inhibits TNF- production in vitro, we examined the effect of rTFPI on the production of TNF- by LPS-stimulated monocytes. rTFPI significantly inhibited the production of TNF- by LPS-stimulated monocytes in a concentration-dependent manner (Figure 6).

View larger version (16K): [in this window] [in a new window]   Figure 6.   Effect of rTFPI on TNF- production by LPS-stimulated monocytes in vitro. Monocytes isolated from normal human blood were incubated with LPS (20 ng/ml) for 16 h at 37° C in a humidified 5% CO2 incubator in the presence or absence of rTFPI. The TNF- concentration in the medium was determined by ELISA. Data are expressed as the mean ± SD of triplicate experiments. *p < 0.01 versus experiments without rTFPI.

Effect of rTFPI on the Functions of Activated Neutrophils In Vitro

To determine whether rTFPI inhibits neutrophil activation, we examined the effect of rTFPI on the release of neutrophil elastase and the production of O₂ by neutrophils. rTFPI inhibited the release of neutrophil elastase (Figure 7A) and O ₂ production (Figure 7B) from fMLP-stimulated neutrophils in a concentration-dependent manner.

View larger version (21K): [in this window] [in a new window]   Figure 7.   Effect of rTFPI on release of neutrophil elastase (A) and superoxide radical production (B) by neutrophils. Neutrophil elastase release from fMLP-stimulated neutrophils in the presence of various concentrations of rTFPI was determined with the chromogenic substrate S-2484. (A). The release of oxygen free radicals from fMLP-stimulated neutrophils in the presence of various concentrations of rTFPI was determined with a chemiluminescence assay. Control experiments were performed in the absence of rTFPI. Data are expressed as the mean ± SD of four determinations in one experiment. *p < 0.01 versus experiments without rTFPI.

Effect of rTFPI on the Expression of CD11b and CD18 by Activated Neutrophils

Flow-cytometric studies have shown increased expression of CD11b and CD18 adhesion molecules on the surface of neutrophils treated with fMLP. Nonspecific binding of monoclonal antibody IgG ₁ to these adhesion molecules was not observed in the present study. rTFPI at a concentration of 0.05 µg/ml significantly inhibited neutrophil expression of CD11b (Figure 8A) and CD18 (Figure 8B).

View larger version (19K): [in this window] [in a new window]   Figure 8.   Effect of rTFPI on expression of CD11b and CD18 on neutrophils in vitro. Expression of CD11b and CD18 on neutrophils was analyzed with flow cytometry. Neutrophils were preincubated with various concentrations of rTFPI followed by incubation with fMLP (1 µM). Mean fluorescence intensity/1.0 × 104 cells was calculated according to the procedure described in METHODS. Data are expressed as the mean ± SD of triplicate experiments. *p < 0.01 versus experiments without rTFPI.

Effect of rTFPI on Changes in [Ca²⁺] after Stimulation of Neutrophils with fMLP

Intracellular calcium is an important second messenger in the metabolic responses of activated neutrophils. [Ca²⁺]i increased rapidly after stimulation with fMLP, and then gradually decreased (Figure 9). rTFPI inhibited the increases in [Ca²⁺]i in a concentration-dependent manner (Figure 9).

View larger version (19K): [in this window] [in a new window]   Figure 9.   Effect of rTFPI on [Ca2+] in fMLP-stimulated neutrophils in vitro. [Ca2+]i in neutrophils was determined by monitoring the fluorescence of indo-1-acetoxymethyl ester. Neutrophils were stimulated with fMLP (at Time = 0) in the presence or absence of various concentrations of rTFPI. Open circles: no rTFPI; open squares: rTFPI at 1 µg/ml; open triangles: rTFPI at 10 µg/ml. Typical results of three independent experiments are shown.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study, rTFPI was shown to reduce LPS-induced pulmonary vascular injury. Since the amount of ¹²⁵I-BSA injected intravenously was not increased in bronchoalveolar lavage fluid in rats challenged with LPS (15), the LPS-induced lung injury in this animal model of lung injury might be limited to the endothelial cells. Thus, rTFPI may have reduced the LPS-induced lung injury mainly by inhibiting endothelial injury in the lung.

Plasma levels of rTFPI in animals pretreated with rTFPI (1 mg/kg, intravenously) at 90 min after LPS administration were 2.54 ± 0.56 µg/ml (58.8 ± 9.3 nM [mean ± SD]; n = 5). These values were higher than Ki values for the inhibition of TF-Factor VIIa and Factor Xa by rTFPI (12), suggesting that rTFPI might inhibit the LPS-induced coagulation abnormalities in this animal model. In accord with this hypothesis, the LPS-induced increase in serum level of FDP(E) was markedly attenuated by rTFPI administration. Since development of ARDS is closely related to disseminated intravascular coagulation (DIC) in the clinical setting (22), rTFPI might reduce LPS-induced pulmonary vascular injury by inhibiting abnormalities in coagulation. However, this is unlikely, because in the present study, neither DX-9065a, a selective inhibitor of Factor Xa, nor DEGR-F.VIIa reduced the LPS-induced pulmonary vascular injury despite their potent anticoagulant activities. Plasma anticoagulant activities of animals given rTFPI, DX-9065a, and DEGR-F.VIIa were significantly increased as compared with those of control animals and were inversely proportional to serum levels of FDP(E). These observations strongly suggest that rTFPI could reduce LPS-induced pulmonary vascular injury independent of its anticoagulant effects.

We have reported that activated leukocytes are importantly involved in this animal model of LPS-induced pulmonary vascular injury (15), suggesting that rTFPI may reduce this injury by inhibiting leukocyte activation. Recently, Senden and coworkers (23) showed that Factor Xa exerts proinflammatory effects by stimulating the endothelial production of IL-6, IL-8, and endothelial leukocyte adhesion molecules such as E-selectin in cultured human umbilical vein endothelial cells. Thus, rTFPI might prevent LPS-induced lung injury by inhibiting leukocyte activation through inhibition of Factor Xa activity. However, neither inhibition of Factor Xa generation by DEGR-F.VIIa nor inhibition of Factor Xa activity by DX-9065a reduced the pulmonary vascular injury in our study, suggesting that Factor Xa could not be implicated in LPS- induced pulmonary vascular injury in the animal model we used.

rTFPI significantly inhibited the LPS-induced increases in lung tissue levels of TNF-, CINC, and MPO in our animal model. Furthermore, rTFPI significantly reduced the LPS- induced expression of TNF- mRNA in the lungs of animals given LPS, indicating that rTFPI inhibited TNF- production by inhibiting its transcription. These observations suggest that rTFPI might reduce pulmonary vascular injury by inhibiting leukocyte activation through inhibition of TNF- production. Since rTFPI inhibited the production of TNF- by LPS-stimulated monocytes in vitro, rTFPI might directly inhibit TNF- production by monocytes.

rTFPI inhibited the pulmonary accumulation of leukocytes in animals given LPS. CINC is a member of the IL-8 family that promotes the accumulation of neutrophils (24). Since the increase in the pulmonary tissue level of CINC in rats given LPS in our study was significantly inhibited by rTFPI, rTFPI at least partly inhibited the pulmonary accumulation of leukocytes by inhibiting CINC production in vivo. Since TNF- enhances CINC production (24), rTFPI might inhibit CINC production secondarily, owing to inhibition of TNF- production.

rTFPI directly inhibited the metabolic responses of activated neutrophils, such as neutrophil elastase release and O₂ production in rats given LPS in our study. Since neutrophil elastase and oxygen free radicals play important roles in producing LPS-induced pulmonary vascular injury (25, 26), rTFPI, by inhibiting neutrophil activation, protects against such injury. Furthermore, neutrophil elastase plays an important role in neutrophil infiltration (26), suggesting that rTFPI may partly reduce the pulmonary infiltration of neutrophils by inhibiting release of neutrophil elastase. Since TNF-  activates endothelial cells to increase their expression of endothelial leukocyte adhesion molecules such as E-selectin (27), inhibition of TNF- production by rTFPI could also contribute in this way to reducing the pulmonary infiltration of neutrophils in rats given LPS.

Concentrations of rTFPI required to inhibit leukocyte functions in vitro (0.1 to 1.0 µg/ml) were lower than the plasma levels of rTFPI (2.54 ± 0.56 µg/ml) measured 90 min after LPS administration in the present study. It is therefore possible that rTFPI inhibits leukocyte activation in vivo, as well as in vitro.

In the present study, we examined the effect of rTFPI on fMLP-induced leukocyte activation under the experimental condition in which neither Factor X nor Factor VII were available. Our observations suggested that the anticoagulant activity of rTFPI might not be critical for the inhibition of leukocyte activation in vitro. This hypothesis is consistent with the observation that rTFPI inhibited leukocyte activation independent of its anticoagulant effects in vivo. Callender and coworkers (28) showed that rTFPI became bound to OC-2008 cells even in the absence of Factors Xa and VIIa. Furthermore, they showed that this binding was independent of the presence of both tissue factor and calcium ion, and was reduced by heparin. Thus, rTFPI might bind to the cell surface of the leukocyte by interacting with heparinlike substances.

The mechanisms by which rTFPI inhibits neutrophil activation are not well understood. Neutrophils are activated via a protein kinase C (PKC)-mediated mechanism that leads to a number of cellular responses, such as the release of neutrophil elastase and the production of reactive oxygen species (29). Increases in [Ca²⁺]i have been shown to activate PKC (30). In the present study, rTFPI was shown to inhibit the increase in [Ca²⁺]i in fMLP-stimulated neutrophils. This may at least partly represent the mechanism by which rTFPI inhibits neutrophil activation. It is widely recognized that the increase in [Ca²⁺]i induced by fMLP apparently consists of two phases: a rapid transient increase observed immediately after stimulation (early phase) and a subsequent gradual decline (late phase). The increase in [Ca²⁺]i in the early and late phases may be attributable to release of Ca²⁺ from intracellular storage sites and to its influx from the extracellular space, respectively (30). Pretreatment of neutrophils with rTFPI significantly inhibited the fMLP-induced early-phase increase in [Ca²⁺]i, suggesting that rTFPI inhibits the release of calcium from intracellular storage.

We found that rTFPI also inhibited the increase in expression of CD11b and CD18 on the cell surface of activated neutrophils. Since the secretory vesicles constitute the most important reservoir of both CD11b and CD18, which together constitute the macrophage antigen-1 (Mac-1) complex, and [Ca²⁺]i plays a critical role in release of the contents of secretory vesicles (31), rTFPI may reduce the expression of CD11b and CD18 by inhibiting the increase in [Ca²⁺]i in activated neutrophils. The Mac-1 complex of CD11b and CD18 interacts with intercellular adhesion molecule-1 on the endothelial cell surface, thereby permitting neutrophil infiltration of the extravascular space (32). The inhibition of expression of these adhesion molecules by rTFPI might therefore at least partly contribute to the inhibition of pulmonary accumulation of neutrophils.

Taken together, these observations suggest that rTFPI could prevent LPS-induced pulmonary endothelial cell injury by inhibiting leukocyte activation. Creasey and coworkers (33) demonstrated that rTFPI also inhibits increases in serum IL-6 levels in baboons challenged with a lethal dose of E. coli. Johnson and associates (10) have shown that rTFPI inhibits IL-8 synthesis induced by coagulation and endotoxin in the human whole-blood culture system. These reports are consistent with our present observations. Collectively, these findings strongly suggest that rTFPI may be useful for treating sepsis complicating DIC and ARDS.