Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Adenosine has been shown to modulate various physiological responses through activations of the four known receptor subtypes: A₁, A_2A, A_2B, and A₃ (Jacobson, 1998). The activation of A₁ receptors inhibits adenyl cyclase and reduces cAMP levels. In contrast, the activation of A₂ receptors stimulates adenyl cyclase and increases cAMP levels (Van Calker et al., 1979; Londos et al., 1980; Wolff et al., 1981; Stiles, 1992). Adenosine receptors exist on human monocytes and neutrophils; their activation modulates leukocyte functions as follows: 1) adenosine inhibits tumor necrosis factor (TNF)-, interleukin (IL)-6, and IL-8 productions from human monocytes stimulated by lipopolysaccharide (LPS) via A₂ receptors (Bouma et al., 1994); and 2) adenosine inhibits superoxide generation (Ward et al., 1988), neutrophil aggregation (Skubitz et al., 1988), and neutrophil adherence to endothelial cells (Cronstein et al., 1986).

Ischemia/reperfusion (I/R) is an important mechanism of tissue injury in which activated leukocytes are critically involved (Hernandez et al., 1987). I/R-induced hepatic injury is an important pathologic process leading to hepatic damage after circulatory shock, major hepatic trauma, major hepatic surgery, or hepatic transplantation (Atalla et al., 1985; Keller et al., 1985; Hasselgren, 1987; Harada et al., 1999b). Recent studies have demonstrated that the activation of adenosine A₂ receptors reduces I/R-induced myocardial injury (Jordan et al., 1997) and lung injury (Khimenko et al., 1995). Because leukocytes are implicated in the pathology of I/R-induced hepatic injury (Jaeschke et al., 1990; Farhood et al., 1995; Vollmar et al., 1995), it is possible that adenosine prevents I/R-induced hepatic injury by inhibiting leukocyte activation via its A₂ receptors.

To determine this possibility, we examined the effect of the novel A₂ receptor agonist 2-octynyladenosine (YT-146; Abiru et al., 1995) on I/R-induced hepatic injury in rats.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

In Vitro Experiments

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 fraction was collected. Neutrophils were isolated according to a Nycodenz gradient method using Nyco Prep 1.077 (Nycomed Pharma AS, Oslo, Norway) (Watabe et al., 1984). Contaminated erythrocytes were removed by hemolysis with 0.2% saline for 25 s. The resulting preparation, which contained more than 95% neutrophils, was washed twice with PBS. Cell viability of 95% or higher was confirmed by the trypan blue dye exclusion test. Cells were suspended in PBS in a volume of 5000/µl.

Release of Neutrophil Elastase from Neutrophils. The neutrophil suspension (5000/µl) in PBS was mixed with 5 µg/ml formyl-methionyl-leucyl-phenylalanine (fMLP), 5 µg/ml cytochalasin B, and 2 mM CaCl₂ in the presence or absence of adenosine, YT-146, CGS21680C, or ZM241385 (Smedly et al., 1986). After incubation for 30 min at 37°C, neutrophil suspensions were centrifuged at 5000g for 10 min at 4°C. Neutrophil elastase activity in supernatants was measured using a chromogenic substrate, S-2484, according to a previously described method (Tanaka et al., 1990).

Measurement of Intracellular Calcium Concentration ([Ca²⁺]_i) in Neutrophils. [Ca²⁺]_i was measured as previously described (Simon et al., 1992). Briefly, neutrophils isolated as described earlier were suspended at 5 to 10 × 10⁶/ml in RPMI 1640 with 10% fetal bovine serum and 2.5 µg/ml Indo-1 acetoxymethyl ester (AM) for 30 min at 37°C. Cells were then washed twice and resuspended at 1 × 10⁶/ml in RPMI 1640 with 10% fetal bovine serum. Immediately before the analysis of [Ca²⁺]_i, cells were washed in HEPES buffer (140 mM NaCl, 3 mM KCl, 1 mM MgCl₂, 10 mM glucose, 1 mM CaCl₂, and 20 mM HEPES; pH 7.23) and suspended (10⁶ cells in a volume of 1.7 ml) in a thermostatically controlled (37°C) cuvette. Dependence on extracellular Ca²⁺ was analyzed using neutrophils suspended in nominally Ca²⁺-free buffer A with the addition of 1 mM EDTA, a solution calculated to contain <10 nM free Ca²⁺. Fluorescence emission was measured in a spectrophotometer (Shimazu RF-5000; Shimazu Co., Kyoto, Japan) using an excitation wavelength of 355 nm and an emission wavelength of 400 or 500 nm. After equilibration of fluorescence to a stable baseline, the cells were stimulated with fMLP (1 µM), and fluorescence assessment was continued.

Isolation and Cultivation of Human Monocytes. Peripheral blood mononuclear cells were isolated from buffy coats obtained from healthy volunteer blood donors by isopyknic centrifugation on a Nyco Prep 1.077 according to a previously described method (Ford and Rickwood, 1982). The medium and buffer solutions were monitored for contamination with endotoxin using the Endotoxin Test-D (Seikagaku-kogyo, Tokyo, Japan). Mononuclear cells were cultured in RPMI 1640 plus 1% fetal bovine serum and then incubated in plastic dishes (1058; Falcon Plastics, Lincoln Park, NJ) for 2 h at 37°C in a humidified 5% CO₂ incubator. Lymphocytes were removed from adherent monocytes by repeated rinsing with serum-free RPMI 1640. Staining with Turk's solution and for nonspecific esterase activity confirmed that more than 90% of harvested cells were monocytes. Cells were adjusted to a volume of 5.0 × 10⁵/ml in RPMI 1640 and 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 adenosine, YT-146, CGS21680C, or ZM241385. After incubation, cell suspensions were centrifuged at 4500g for 10 min in an Eppendorf microcentrifuge (model 5412). Levels of TNF- in supernatant fractions were determined using an enzyme-linked immunosorbent assay (ELISA) kit for human TNF- (Otsuka Pharm. Co., Tokyo, Japan).

In Vivo Experiments

Animal Model of Hepatic I/R. Adult, pathogen-free, male Wistar rats (Nihon SLC, Hamamatsu, Japan) that weighed 220 to 280 g were used in each experiment. The care and handling of the animals were in accordance with the National Institute of Health guidelines. All experimental procedures described here were approved by the Fukuoka University Animal Care and Use Committee. All rats were deprived of food, but not water, for 24 h before each experiment. The hepatic I/R protocol was performed as described previously (Hayashi et al., 1986; Lee and Clemens, 1992). After the induction of anesthesia in the animals with ketamine hydrochloride (100 mg/kg i.p.; Parke-Davis, Morris Plains, NJ), the liver of each was exposed through a midline laparotomy. Silk ligatures were placed around the right and left branches of the portal vein and the hepatic artery. Complete ischemia of the median and left hepatic lobes was produced by clamping the left branches of the portal vein and the hepatic artery for 60 min. The right hepatic lobe was perfused to prevent intestinal congestion. During the period of hepatic ischemia, the animal's abdomen was covered with plastic wrap to prevent dehydration. After the period of ischemia, the ligatures around the left branches of the portal vein and hepatic artery were removed. To accurately evaluate blood flow of the median and left hepatic lobes after ischemia, the right branches of the portal vein and the hepatic artery were ligated to prevent shunting to the right lobe after reperfusion (Hayashi et al., 1986). The wound was closed with 3-0 silk. This procedure directed all portal and hepatic blood flow, except for a small amount of flow to the caudal hepatic lobe, through the lobes of the liver previously made ischemic. Sham-operated animals were similarly prepared except that no ligature was placed to obstruct the blood flow to the left and median hepatic lobes. Instead, the blood flow to the right lobe of the liver was occluded.

Measurement of Serum Liver Enzymes. Blood samples were taken 12 h after reperfusion to measure the level of serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) as previously described (Kushimoto et al., 1996). These blood samples were collected into test tubes from the anesthetized animals via withdrawal from the abdominal aorta with a 22-gauge needle. ALT and AST levels were measured by standard clinical automated analysis, and the results are expressed in international units per liter.

Measurement of Hepatic Tissue Blood Flow. Hepatic tissue blood flow was measured by laser-Doppler flowmeter (ALF21N; Advance, Tokyo, Japan) for 3 h after reperfusion, as described previously (Harada et al., 1999b). After anesthesia with 100 mg/kg i.p. ketamine hydrochloride, the right jugular veins of these animals were cannulated with a PE-10 catheter for the continuous infusion of normal saline or test drugs. The Doppler flowmeter probe was placed on the medial hepatic lobe. Hepatic tissue blood flow was measured from 30 min before ischemia until 3 h after reperfusion. The results are expressed as percentage of preischemia levels.

Determination of Hepatic Levels of TNF-. Hepatic levels of TNF- were determined by a modification of a previously described method (Shito et al., 1997). At the indicated times after reperfusion, the animals were anesthetized with an i.p. injection of ketamine hydrochloride (100 mg/kg) and exsanguinated via the abdominal aorta. In brief, the medial hepatic lobe was weighed and then homogenized in 5 ml of 0.1 M phosphate buffer (pH 7.4) containing 0.05% (v/w) of sodium azide at 5°C. The homogenate was first centrifuged at 2000g for 10 min to remove minute amounts of solid tissue debris. The supernatant was assayed using a rat TNF- ELISA system (Amersham, Buckinghamshire, UK). This ELISA is sensitive enough to detect 31 to 2500 pg/ml TNF-. The results are expressed as picograms of TNF- per gram of tissue.

Determination of Hepatic Levels of Cytokine-Induced Neutrophil Chemoattractant (CINC). CINC is an equivalent to human chemokine Gro (Watanabe et al., 1989). Hepatic levels of CINC were determined by a modification of a previously described method (Clark et al., 1991). In brief, the medial hepatic lobe in which the tissue blood flow had been measured was weighed and then homogenized in 5 ml of 0.1 M phosphate buffer (pH 7.4) containing 0.05% (v/w) of sodium azide at 5°C. The homogenate was first centrifuged at 2000g for 10 min to remove minute amounts of solid tissue debris. The supernatant was assayed using a CINC/gro ELISA system (Amersham). This ELISA was sensitive enough to detect 4.7 to 300 pg/ml CINC/gro. The results are expressed as picograms of CINC per gram of tissue.

Determination of Hepatic Myeloperoxidase (MPO) Activity. After the indicated period of reperfusion, the livers were quickly removed, and the accumulation of leukocytes was assessed by measuring MPO activity, which reflects the tissue accumulation of neutrophils, in the liver according to a previously described method (Yamaguchi et al., 1997). In brief, the livers were weighed and suspended in 6 ml of 50 mM phosphate buffer (pH 6.0) containing 1% hexadecyltrimethylammonium bromide. The samples were homogenized and the homogenate was sonicated, freeze-thawed, and then centrifuged (4500g for 15 min at 4°C). MPO activity in the supernatant (0.1 ml) was determined after the addition of 0.6 ml of phosphate buffer (pH 6.0) containing 0.167 mg/ml o-dianisidine dihydrochloride and 0.0005% hydrogen peroxide. The change in absorbance at 460 nm over 10 min was measured in a spectrophotometer (DU-54; Beckman, Irvine, CA). One unit of MPO activity was defined as the amount of enzyme able to reduce 1 µmol of peroxide/min. Results are expressed as units of MPO activity per gram of tissue.

Histological Examination

After 6 h of reperfusion, liver specimens were fixed in 10% buffered formalin and then embedded in paraffin. Samples were used to study the infiltration of polymorphonuclear leukocytes (PMNs). Sections (4 µm) were stained using the naphthol AS-D chloroacetate esterase technique to investigate the accumulation of PMNs in the liver (Moloney et al., 1960). PMNs were identified by positive staining and morphology and counted under 50 high-power fields of a light microscope.

Administration of Adenosine, YT-146, and CGS21680C

Adenosine and YT-146 were dissolved in 0.1% dimethyl sulfoxide (DMSO)-saline solution and continuously infused into the animals via the right jugular vein. CGS21680C was dissolved in normal saline and continuously infused into the animals via the right jugular vein. For those groups of animals in which the tissue blood flow was measured, these infusions began before the onset of hepatic ischemia and continued until the rats were sacrificed 3 h after the onset of reperfusion. Because serum levels of transaminases were measured 12 h after reperfusion, these agents should be given for 12 h after reperfusion. However, because anesthesia for 12 h suppressed respiration, we administered these agents continuously for 6 h after reperfusion. Control animals received a continuous infusion of the same volume of each vehicle instead of adenosine, YT-146, or CGS21680C.

Reagents

YT-146 [2-(1-octynyl)adenosine] was kindly provided by Toa Eiyo Ltd. (Tokyo, Japan). CGS21680C (2-[4-(2-carboxymethyl)phenethylamino]-5'-N-ethyl-1,3-dipropylxanthine) was kindly provided by Novartis Pharmaceuticals Co. (Summit, NJ). Adenosine, fMLP, cytochalasin B, HEPES, and naphthol AS-D chloroacetate esterase were obtained from Sigma Chemical Co. (St. Louis, MO). ZM241385, 4-(2-[7-amino-2-(2-furyl)[1,2,4]trizolo[2,3-a][1,3,5]triazin-5-ylamino]ethyl)phenol was obtained from Tocris Cookson Ltd. (Bristol, UK). Heparin was purchased from Novo Nordisk A/G (Genotofte, Denmark). S-2484 was obtained from Chromogenix AB (Stockholm, Sweden). LPS (endotoxin, Escherichia coli serotype O55:B5) was purchased from Difco (Detroit, MI). Indo-1 AM was obtained from Dojindo Laboratories (Kumamoto, Japan). Fetal bovine serum and RPMI 1640 were purchased from GIBCO (Gaithersburg, MD). All other reagents were of analytical grade.

Statistical Analysis

Data are expressed as mean ± S.D. The results were compared using either an ANOVA followed by Scheffé's post hoc test or an unpaired t test. A level of P < .05 was considered statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Effects of Adenosine, YT-146, CGS21680C, and ZM241385 on Neutrophil Elastase Release by Neutrophils In Vitro. Adenosine and selective A_2A receptor agonists YT-146 and CGS21680C significantly inhibited the fMLP-induced neutrophil elastase release in a concentration-dependent manner (Fig. 1). ZM241385, a selective adenosine A_2A receptor antagonist, markedly enhanced the neutrophil elastase release induced by fMLP (Fig. 1).

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Effects of adenosine, YT-146, CGS21608C, and ZM241385 on neutrophil elastase release from activated neutrophils of healthy human volunteers in vitro. The release of neutrophil elastase in the presence of various concentrations of these agents was determined. PBS containing 0.05% DMSO was used instead of the drug for the control experiments. Data are expressed as the mean ± S.D. of triplicate experiments. N.S., not significant versus control group. *P < .01 versus control group.

Effects of Adenosine, YT-146, CGS21680C, and ZM241385 on [Ca²⁺]_i. The stimulation of neutrophils by fMLP (at t = 0) in the presence of 1 mM CaCl₂ induced the elevation of [Ca²⁺]_i, as measured by Indo-1 AM fluorescence (Fig. 2). The marked decreases in the duration of [Ca²⁺]_i elevation were seen in absence of extracellular Ca²⁺, but there was only a slight decrease in amplitude compared with those in presence of extracellular Ca²⁺ (Fig. 2). Pretreatment of neutrophils with adenosine (Fig. 2, A and B), YT-146 (Fig. 2, C and D), and CGS21680C (Fig. 2, E and F) suppressed the fMLP-induced elevation of [Ca²⁺]_i in neutrophils in either the absence or presence of extracellular Ca²⁺. ZM241385 enhanced the elevation of [Ca²⁺]_i in fMLP-stimulated neutrophils, especially in the early phase (Fig. 2, G and H).

View larger version (40K): [in this window] [in a new window]   Fig. 2.   Effects of adenosine, YT-146, CGS21680C, and ZM241385 on [Ca2+]i. [Ca2+]i in neutrophils was monitored by fluorescence of Indo-1 AM. Neutrophils are stimulated by fMLP at time 0 in the absence (A, C, E, and G) or presence (B, D, F, and H) of extracellular Ca2+ concentration ([Ca2+]o). Neutrophils also are stimulated by fMLP in the presence (, 0.1 µM; , 1 µM; , 10 µM) or absence () of various concentrations of adenosine (A and B), YT-146 (C and D), CGS21680C (E and F), and ZM241385 (G and H). Data are expressed as the mean ± S.D. of triplicate experiments. *P < .01 versus control group.

Effects of Adenosine, YT-146, CGS21680C, and ZM241385 on TNF- Production by LPS-Stimulated Monocytes In Vitro. To determine whether adenosine, via A_2A receptor, inhibits the monocytic production of TNF-, which has been shown to be implicated in I/R-induced liver injury (Shito et al., 1997), the effects of adenosine, YT-146, CGS21680C, and ZM241385 on the production of TNF- by LPS-stimulated monocytes were examined in vitro. Adenosine, YT-146, and CGS21680C significantly inhibited the LPS-induced increase in TNF- production by monocytes in a concentration-dependent manner (Fig. 3). ZM241385 did not affect the LPS-induced increase in TNF- production by monocytes (Fig. 3).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Effects of adenosine, YT-146, CGS21680C, and ZM241385 on TNF- production in LPS-stimulated monocytes in vitro. Human monocytes were cultured in RPMI1640 with or without various concentrations of these agents and stimulated by LPS. PBS containing 0.05% DMSO was used instead of the drug for the control experiments. Data are expressed as the mean ± S.D. of triplicate experiments. §P < .01 versus vehicle group. *P < .01 versus control group.

Effects of Adenosine, YT-146, and CGS21680C on I/R-Induced Hepatic Injury. Serum levels of ALT and AST were significantly increased after 1 h of reperfusion compared with levels in sham-operated animals, peaking at 12 h after reperfusion (Kushimoto et al., 1996). ALT and AST levels in intact rats were 39.8 ± 6.3 and 61.6 ± 14.8 I.U./l (mean ± S.D.; n = 5), respectively.

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Effects of adenosine, YT-146, and CGS21680C on serum levels of aminotransferases in rats after hepatic I/R. Animals were subjected to 60 min of hepatic ischemia followed by reperfusion as described under Materials and Methods. Adenosine (0.1, 1, or 10 mg/kg/h), YT-146 (0.01, 0.1, or 1 mg/kg/h), and CGS21680C (0.01, 0.1, or 1 mg/kg/h) were infused continuously starting just before the onset of ischemia. Serum levels of AST and ALT were determined after 12 h after reperfusion. Each column represents the mean ± S.D.. *P < .01 versus control group.

Effects of Adenosine, YT-146, and CGS21680C on Changes in Hepatic Tissue Blood Flow in Rats Subjected to Hepatic I/R. During hepatic ischemia, the hepatic tissue blood flow decreased to approximately 30% of the preischemia level and then increased to 50% of the preischemia level 3 h after reperfusion (Fig. 5). The intravenous infusion of adenosine (1 mg/kg/h), YT-146 (0.1 mg/kg/h), and CGS21680C (0.1 mg/kg/h) significantly increased the hepatic tissue blood flow after 1 to 3 h of reperfusion.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Effects of adenosine, YT-146, and CGS21680C on I/R-induced changes in hepatic tissue blood flow in rats. Animals were subjected to 60 min of hepatic ischemia followed by reperfusion as described under Materials and Methods. Adenosine (1 mg/kg/h), YT-146 (0.1 mg/kg/h), and CGS21680C (0.1 mg/kg/h) were infused continuously starting just before the onset of ischemia. Each value represents the mean ± S.D. derived from 5 animals. , control; , adenosine; , YT-146; , CGS21680C. *P < .01 versus control group.

Effects of Adenosine, YT-146, and CGS21680C on I/R-Induced Changes in Hepatic Levels of TNF-, CINC, and MPO in Rats. Hepatic levels of TNF- or CINC, a member of the IL-8 family, were significantly increased after reperfusion compared with those of the sham-operated animals and peaked 1 or 2 h after reperfusion, respectively (Fig. 6, A or B). Because a slight insult associated with the surgical procedure might lead to the slight elevation of TNF- and CINC levels, the elevation of these cytokines in sham-operated animals may due to surgical procedures (Fig. 6, A and B). The intravenous infusion of adenosine (1 mg/kg/h), YT-146 (0.1 mg/kg/h), and CGS21680C (0.1 mg/kg/h) significantly inhibited these increases 1 or 2 h after reperfusion (Fig. 7, A or B). Hepatic MPO activity was increased significantly after reperfusion compared with that of the sham-operated animals and peaked 6 h after reperfusion (Fig. 6C). Adenosine, YT-146, and CGS21680C significantly inhibited this increase 6 h after reperfusion (Fig. 7C).

View larger version (22K): [in this window] [in a new window]   Fig. 6.   Changes in hepatic TNF- levels (A), CINC levels (B), and MPO activity (C) in rats subjected to hepatic I/R. Animals were subjected to 60 min of hepatic ischemia followed by reperfusion as described under Materials and Methods. Each value represents the mean ± S.D. derived from five animals. , sham-operated animals; , control animals. §P < .01 versus preischemia. P < .01 versus sham.

View larger version (22K): [in this window] [in a new window]   Fig. 7.   Effects of adenosine, YT-146, and CGS21680C on I/R-induced increase in hepatic TNF- levels (A), CINC levels (B), and MPO activity (C) in rats. Animals were subjected to 60 min of hepatic ischemia followed by reperfusion as described under Materials and Methods. Adenosine (1 mg/kg/h), YT-146 (0.1 mg/kg/h), and CGS21680C (0.1 mg/kg/h) were infused continuously starting just before the onset of ischemia. Each column represents the mean ± S.D. §P < .01 versus preischemia. *P < .01 versus control group.

Effects of Adenosine, YT-146, and CGS21680C on I/R-Induced PMN Accumulation in Liver in Rats. The number of PMNs in the liver was significantly increased in animals subjected to the hepatic I/R compared with that of sham-operated animals (Table 1). Adenosine, YT-146, and CGS21680C significantly inhibited the I/R-induced hepatic accumulation of PMNs (Table 1).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, adenosine and two A_2A receptor agonists (YT-146 and CGS21680C) inhibited fMLP-induced increases in both neutrophil elastase release and [Ca²⁺]_i. Furthermore, these agents inhibited the production of TNF- by monocytes in vitro. In vivo experiments, these agents inhibited the decrease in hepatic tissue blood flow and hepatic injury in rats subjected to hepatic I/R. The increases in hepatic tissue levels of TNF-, CINC, and MPO observed after reperfusion were also inhibited by these agents.

Activated leukocytes are considered to play a pivotal role in I/R-induced hepatic injury by releasing various inflammatory mediators that are capable of damaging endothelial cells (Colletti et al., 1990; Shito et al., 1997). We have previously demonstrated the involvement of neutrophil elastase in the pathologic process of this rat model of hepatic I/R (Kushimoto et al., 1996). Because neutrophil elastase increases endothelial permeability in vitro (Suzuki et al., 1994), neutrophil elastase might play an important role in the reduction of hepatic tissue blood flow by increasing the microvascular permeability in animals subjected to I/R. A neutrophil elastase-induced increase in microvascular permeability might lead to local hemoconcentration in the microcirculation at the site of endothelial damage, leading to tissue ischemia (Liu et al., 1998). Consistent with this hypothesis is our preliminary study demonstrating that ONO-5046, a specific inhibitor of neutrophil elastase, significantly inhibits the I/R-induced decrease in hepatic tissue blood flow.

Adenosine and two A_2A receptor agonists inhibited neutrophil elastase release at concentrations of 10⁷ to 10⁵ M in vitro, as shown in this study. Our preliminary study showed that the hepatic level of adenosine in rats 2 h after the continuous i.v. infusion of adenosine (1 mg/kg/h) was 7.0 × 10⁷ M. The expected plasma concentrations of YT-146 and CGS21680C in rats were estimated to be 1.2 × 10⁷ M (Dr. Kogi, Toa Eiyo Ltd., Tokyo, Japan, unpublished observations) and 1.9 × 10⁷ M (Mathôt et al., 1995), respectively, 2 h after the continuous i.v. infusion of these agents at a dosage of 0.1 mg/kg/h. These findings strongly suggested that adenosine and two A_2A receptor agonists could inhibit neutrophil elastase release in vivo. Thus, it is possible that adenosine, YT-146, and CGS21680C might inhibit the hepatic I/R-induced decrease in hepatic tissue blood flow, probably by inhibiting neutrophil elastase release.

Although the control mechanism or mechanisms for neutrophil elastase release are not well understood, an increase in [Ca²⁺]_i plays an important role in this process (Borregaard and Cowland, 1997). The elevation of [Ca²⁺]_i in activated neutrophils induced by fMLP apparently consists of two phases: a rapid transient elevation observed immediately after stimulation (early phase) and a subsequent gradual decline (late phase). The high value of [Ca²⁺]_i in the late phase is dependent on the influx of extracellular Ca²⁺, whereas the early phase is not. Therefore, the elevation of [Ca²⁺]_i in the early and late phases may be contributed to release from intracellular storage sites and influx from the extracellular space, respectively (Kainoh et al., 1990). Pretreatment of neutrophils with adenosine, YT-146, and CGS21680C inhibited the fMLP-induced increase in [Ca²⁺]_i in the absence or presence of extracellular Ca²⁺, suggesting that these agents could inhibit the release of Ca²⁺ mainly from the intracellular storage site, thereby inhibiting neutrophil elastase release. ZM241385, a selective A_2A receptor antagonist, significantly enhanced neutrophil elastase release and increase in [Ca²⁺]_i in neutrophils stimulated with fMLP, suggesting that adenosine A_2A receptor plays a role in inhibiting neutrophil elastase release.

TNF- plays a role in I/R-induced hepatic injury by activating both neutrophils and endothelial cells (Colletti et al., 1998). We have previously shown that gabexate mesilate, a synthetic serine protease inhibitor, reduces I/R-induced hepatic injury by inhibiting TNF- production (Harada et al., 1999 a). Consistent with this theory, hepatic levels of TNF- and MPO were significantly increased after reperfusion as shown in this study. Hepatic levels of CINC, a potent activator of neutrophils, of which the production is enhanced by TNF- (Baggiolini et al., 1989), were also increased after reperfusion. Hepatic levels of these variables were slightly increased in animals with sham surgery, probably due to the surgical procedure (i.e., laparotomy). Because liver has some peroxidases (Komatsu et al., 1992), MPO activity might reflect not only neutrophils but also other hepatic peroxidases. However, this possibility seems less likely, because prostacyclin, which inhibits endothelial adhesion of neutrophils, inhibits the increase in hepatic MPO activity (Harada et al., 1999b). Yamaguchi et al. (1997) have also shown that hepatic MPO activity reflects mainly the hepatic accumulation of neutrophils. Consistent with these observations, PMN accumulation in the liver was significantly increased after I/R in this study.

Both MPO activity and PMN accumulation in the liver were increased in animals subjected to the hepatic I/R by about 30 to 50% compared with those seen in animals with sham surgery in this study. Suzuki et al. (1993) reported that the hepatic accumulation of PMNs induced by hepatic I/R was increased by about 5- to 10-fold, which is much higher than that observed in this study. The difference could be explained by experimental conditions, because the hepatic MPO activities were increased by about 5- to 10-fold in rats subjected to 120-min I/R of liver by using our experimental protocol (our unpublished observation). In addition, PMNs were mainly observed in the extrasinusoidal space of the liver in animals subjected to 120-min I/R, which is in contrast to this observation showing that PMNs were mainly present in the sinusoidal space in rats subjected to 60-min I/R of the liver (our unpublished observation). Although neutrophil rolling is generally associated with the subsequent neutrophil infiltration, the two events are not always interdependent, as when the activation of neutrophils leading to rolling is insufficient to induce the neutrophil adhesion response (Kubes et al., 1995). Thus, the extent of hepatic accumulation of neutrophils after liver I/R may depend on the duration of ischemia leading to neutrophil activation.

Adenosine, YT-146, and CGS21680C, in the concentration of 10⁷ to 10⁵ M, significantly inhibited the LPS-induced TNF- production by monocytes to less than 50% of control in vitro. ZM241385, a selective A_2A receptor antagonist, did not affect the LPS-induced increase in TNF- production by monocytes. Thus, inhibition of TNF- production by adenosine, YT-146, and CGS21680C might contribute to the reduction in the hepatic injury by inhibiting neutrophil activation. Consistent with this hypothesis is the present observations that adenosine, YT-146, and CGS21680C inhibited the I/R-induced increases in hepatic levels of TNF-, CINC, and MPO as well as the increase in serum transaminase levels after I/R. Thus, these agents could inhibit neutrophil activation by inhibiting the production of these cytokines as well as by inhibiting neutrophil activation directly.

I/R of the liver has been shown to promote the TNF- production by both Kupffer cells (Wanner et al., 1999) and circulating monocytes (Okuaki et al., 1996). Although we measured the hepatic TNF- levels in this study, we could not completely separate TNF- levels in the circulation from that in the hepatic tissue. Thus, the elevation of hepatic TNF- levels after I/R might be a consequence of the production of TNF- by both Kupffer cells and circulating monocytes.

The precise mechanism by which adenosine reduces LPS-stimulated TNF- production by monocytes is not known. The nuclear factor (NF)-B/Rel family of transcription factors has been implicated in the inducible expression of many genes in monocytes, including TNF- (Ziegler-Heitbrock et al., 1993). Because elevated cAMP inhibits NF-B-mediated transcription in human monocytic cells (Ollivier et al., 1996), it is possible that adenosine may inhibit NF-B-mediated TNF- gene transcription in monocytes by increasing cAMP levels via adenosine A_2A receptor stimulation. This possibility is clarified by further investigation.