- Email: [email protected]
Gastric prostacyclin (PGI2) prevents stress-induced gastric mucosal injury in rats primarily by inhibiting leukocyte activation
Prostaglandins & other Lipid Mediators 57 (1999) 291–303
Gastric prostacyclin (PGI2) prevents stress-induced gastric mucosal injury in rats primarily by inhibiting leukocyte activation Naoaki Harada, Kenji Okajima*, Kazunori Murakami, Hirotaka Isobe, Wenge Liu Department of Laboratory Medicine, Kumamoto University School of Medicine, Honjo 1-1-1, Kumamoto 860-0811, Japan Received 11 March 1998; received in revised form 6 July 1998; accepted 29 October 1998
Abstract We investigated whether, in rats, gastric prostacyclin (PGI2) prevented gastric mucosal injury that was induced by water-immersion restraint stress by inhibiting leukocyte activation. Gastric levels of 6-keto-PGF1␣, a stable metabolite of PGI2, increased transiently 30 min after stress, followed by a decrease to below the baseline 6 – 8 h after stress. Gastric mucosal blood flow decreased to ⬃40% of the baseline level 8 h after stress. Myeloperoxidase activity was significantly increased 8 h after stress. Treatment with indomethacin before stress inhibited the increase in 6-keto-PGF1␣ levels and markedly reduced mucosal blood flow. It also markedly increased leukocyte accumulation and mucosal lesion formation. Iloprost, a stable PGI2 analog, inhibited the indomethacin-induced decrease in mucosal blood flow, mucosal lesion exacerbation, and increase in leukocyte accumulation. Nitrogen mustard-induced leukocytopenia inhibited the indomethacin-associated lesion exacerbation and the increase in leukocyte accumulation, but not the decreases in mucosal blood flow. These observations indicate that gastric PGI2 decreases gastric mucosal lesion formation primarily by inhibiting leukocyte accumulation. © 1999 Elsevier Science Inc. All rights reserved. Keywords: Water-immersion restraint stress; Granulocyte elastase; Gastric mucosal lesion; Gastric mucosal blood flow; Prostacyclin
* Corresponding author. Tel.: ⫹81-96-373-5281; fax: ⫹81-96-373-5281. E-mail address: [email protected] (K. Okajima) 0090-6980/99/$ – see front matter © 1999 Elsevier Science Inc. All rights reserved. PII: S 0 0 9 0 - 6 9 8 0 ( 9 9 ) 0 0 0 7 7 - X
292
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
1. Introduction Prostaglandins (PG) play an important role in preventing gastric mucosal injury, referred to as gastric cytoprotection [1]. Although PGE1 and PGE2 inhibit gastric secretion [2,3], the cytoprotective effect is considered to be independent of the antisecretory mechanism [1]. In the gastric mucosa, PGE2 and PGI2 have been shown to prevent gastric mucosal injury induced by various noxious stimuli [4,5]. This protective effect is thought to involve increases in gastric mucus secretion, bicarbonate production, and mucosal blood flow [6], although its precise mechanism has not fully been elucidated. Activated leukocytes have been implicated in gastric mucosal injuries induced by nonsteroidal anti-inflammatory drugs [7], water-immersion restraint stress [8], hemorrhagic shock [9], and ethanol [10]. Activated leukocytes release a variety of inflammatory mediators, including granulocyte elastase and reactive oxygen species that damage the endothelial cells and other tissues [11–13]. PGI2, the most abundant PG in gastric mucosa, is synthesized in the endothelium and has various physiological actions at the interface between blood and tissue [14]. Prostaglandins, such as PGE1, PGE2, and PGI2, inhibit tumor necrosis factor-␣ production by monocytes through increasing intracellular cyclic adenosine 5⬘-monophosphate levels [15,16]. Both PGI2 and iloprost, a stable derivative of PGI2, inhibit neutrophil activation by increasing intracellular cyclic adenosine 5⬘-monophosphate levels in vitro [17,18]. Furthermore, PGI2 has been shown to inhibit neutrophil adhesion to endothelial cells in vitro [19]. These actions suggest that PGI2 and iloprost prevent various types of experimental tissue injury in vivo by inhibiting leukocyte activation [20]. These observations strongly suggest that the cytoprotective action of gastric PGI2 may also involve the inhibition of leukocyte activation. However, direct evidence for this hypothesis is lacking. Therefore, we investigated whether PGI2 prevented the stress-induced gastric mucosal injury in rats through inhibition of leukocyte activation.
2. Materials and methods 2.1. Materials Indomethacin, nitrogen mustard (NM), hexadecyl-trimethylammonium bromide, and odianisidine dihydrochloride were purchased from Sigma (St. Louis, MO, USA). Iloprost, a stable analog of PGI2, was kindly supplied by Eizai Pharmaceutical Co. (Tokyo, Japan). All other reagents used were of analytical grade. 2.2. Animals Adult male Wistar rats (Nihon SLC, Hamamatsu, Japan), weighing 280 –320 g, were used in each experiment. Care and handling of animals were in accordance with National Institutes of Health guidelines. All experimental procedures described below were approved by the Kumamoto University Animal Care and Use Committee.
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
293
2.3. Experimental design Animals were divided randomly into four groups as follows: Group 1, no treatment with water-immersion restraint stress (WIR) (n ⫽ 120); Group 2, pretreatment with indomethacin (IM) ⫹ WIR (n ⫽ 120); Group 3, IM ⫹ iloprost ⫹ WIR (n ⫽ 20); Group 4, NM ⫹ IM ⫹ WIR (n ⫽ 20). 2.4. Water immersion-restraint stress-induced gastric mucosal lesion formation Before each experiment, rats were deprived of food but not water for 24 h. Then the animals were placed in a restraint cage and immersed up to the xiphoid process in water at 22°C as described previously [21]. At the indicated times during water immersion-restraint stress, the animals were anesthetized by an intraperitoneal injection of sodium pentobarbital (50 mg/kg) and exsanguinated via the abdominal aorta. The stomachs were removed, filled with 10 ml of 1% formalin, and immersed in 1% formalin for 24 h, after which they were cut along the greater curvature and examined for mucosal lesions. Because most of the observed lesions were linear with widths of ⬍2 mm, the total length of each linear hemorrhagic erosion was recorded as the lesion index (mm) by an independent observer blinded to treatment, as previously described [22]. After gross examination, the stomachs were fixed in 10% buffered formalin, and 3-m sections were stained with hematoxylineosin for histologic evaluation. To assess polymorphonuclear leukocyte (PMN) accumulation in gastric mucosa, neutrophils were counted in 10 high-power fields (⫻400), and the mean ⫾ SD was calculated. 2.5. Administration of indomethacin Indomethacin (5 mg/kg) was suspended in bicarbonate-buffered saline and administered subcutaneously (s.c.) 30 min prior to WIR. The same dose of indomethacin, administered s.c. or orally, does not affect the gastric acid secretion in intact rats as previously described [23,24]. Control animals received the same volume of bicarbonate-buffered saline without indomethacin. 2.6. Administration of iloprost Iloprost was continuously infused (100 ng/kg/min) via the jugular vein for 8 h after placement of a polyethylene tube while the rats were under light ether anesthesia. Iloprost, at a dose of 100 ng/kg/min, inhibits dermal neutrophil accumulation in a canine model of skin inflammation and inhibits neutrophil accumulation in the infarcted dog myocardium [18]. A femoral arterial catheter was placed to measure the mean arterial pressure by using strain-gauge transducer. We previously reported that mean arterial pressure fell from 105.8 ⫾ 11.4 to 91.2 ⫾ 2 mmHg in rats 30 min after intravenous infusion of iloprost at a dosage of 100 ng/kg/min [25]. Control animals received normal saline instead of iloprost. Although whether iloprost inhibits gastric acid secretion in rats has not yet been studied,
294
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
PGI2 at a dosage of 100 ng/kg/min significantly inhibits gastric acid secretion in dogs by ⬃50% [26]. 2.7. Reduction in circulating leukocytes by NM Rats were made leukocytopenic by the intravenous NM injection at a dosage of 1 mg/kg for 2 days prior to the day of the experiment [27]. Leukocytopenia was confirmed by circulating leukocyte counts, for which blood was collected from the tail vein under light ether anesthesia, immediately before stress. After smearing on a glass slide and Wright– Giemsa staining, the slides were coded to avoid observer bias and examined under a light microscope with a ⫻100 objective. Control animals were injected over a similar time course with an equal volume of normal saline. 2.8. Determination of gastric 6-keto-PGF1␣ levels Gastric levels of 6-keto-PGF1␣, a stable metabolite of PGI2, were determined in animals before and during WIR, according to the methods described previously [28]. Briefly, the stomachs were weighed and then homogenized in 5 ml of 0.1 M phosphate buffer (pH 7.4) at 5°C. The homogenates were centrifuged at 2000 ⫻ g (KR-2000T, Kubota Co., Tokyo, Japan) for 10 min to remove minute amounts of solid tissue debris. The supernatant was then acidified with 1 M HCl. 6-keto-PGF1␣ was extracted from the supernatant by using columns packed with ethyl-bonded silica gel (ethyl C2, Amersham, Buckinghamshire, UK). Columns were prepared by washing with 2 ml of methanol, followed by 2 ml of water prior to use. The acidified supernatant was applied onto the column, followed by washing sequentially with 5 ml of water, 5 ml of 10% ethanol, and 5 ml of hexane. Elution of 6-keto-PGF1␣ was performed with 5 ml of methyl formate, and the solvent was evaporated under a stream of nitrogen gas. The concentration of 6-keto-PGF1␣ was assayed by using a specific enzyme immunoassay kit (Amersham). Results are expressed as g of 6-keto-PGF1␣ per g of tissue. 2.9. Measurement of gastric mucosal blood flow Gastric mucosal blood flow was measured by laser Doppler flowmeter (ALF21N, Advance, Tokyo, Japan) as described previously [29]. Rats were anesthetized with ether, and midline laparotomies were performed. The glass fiber, connected with the hemodynamic probe, was placed against the mucosa of the corpus along the greater curvature through a proximal incision. The incisions were sutured tightly to prevent water from entering the abdominal cavity during water-immersion stress loading. Results are expressed as a percentage of pre-WIR levels. 2.10. Measurement of gastric myeloperoxidase activity After animals were immersed for the indicated period of stress, all were immediately killed. Their stomachs were quickly removed and opened along the greater curvature. In some experiments, leukocyte infiltration in gastric mucosa was assessed by determining
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
295
tissue activity of myeloperoxidase (MPO), an enzyme used as a marker for leukocyte infiltration in a variety of tissues including rat gastric mucosa [30 –32]. MPO activity was determined by a modification of the method of Krawisz et al. [33]. Briefly, the stomachs were weighed and suspended in 6 ml of 50 mM phosphate buffer (pH 6.0) containing 1% hexadecyl-trimethylammonium bromide. The samples were homogenized, and the homogenate was sonicated, freeze-thawed, and then centrifuged (4500 ⫻ g for 15 min at 4°C). MPO activity was determined in the supernatant (0.1 ml) after the addition of 0.6 ml of phosphate buffer (pH 6.0) containing 0.05 ml of 1.25 mg/ml o-dianisidine dihydrochloride and 0.05 ml of 0.05% hydrogen peroxide. The change in absorbance at 460 nm over 6.5 min was measured in a spectrophotometer (DU-54; Beckman, Irvine, CA, USA). One unit of MPO activity was defined as the amount of enzyme able to reduce 1 mol peroxide/min. Results are expressed as units of MPO activity per g of tissue. 2.11. Statistical analysis All data are expressed as the mean ⫾ SD. Comparisons among different groups of data were performed by using analysis of variance with the Scheffe´’s (post hoc) test. A P value of ⱕ5% was considered significant.
3. Results 3.1. Changes in gastric 6-keto-PGF1␣ level, gastric MPO activity, gastric mucosal blood flow, and gastric lesion index To examine whether gastric PGI2 cytoprotection involves the inhibition of leukocyte activation, we investigated the relationship between gastric levels of 6-keto-PGF1␣, a stable metabolite of PGI2, and gastric accumulation of leukocytes in rats during WIR (Fig. 1). At the same time, we measured changes in the gastric mucosal blood flow and the gastric lesion index during WIR. Gastric levels of 6-keto-PGF1␣ were significantly increased 30 min after WIR compared to pre-WIR levels (Fig. 1A). These levels began to decrease 1 h after WIR and declined to below pre-WIR levels at 6 h and 8 h after WIR (Fig. 1A). Gastric accumulation of leukocytes as evaluated by gastric MPO activity, significantly increased 8 h after WIR compared to pre-WIR levels (Fig. 1B). Gastric mucosal blood flow significantly decreased 2 h after WIR, and decreased to 45% of the pre-WIR level 8 h after WIR (Fig. 1C). The gastric lesion index significantly increased 4 h after WIR and reached a maximum 8 h after WIR (Fig. 1D). 3.2. Effect of indomethacin pretreatment on gastric 6-keto-PGF1␣ level, gastric accumulation of leukocytes, gastric mucosal blood flow, and gastric lesion index Subcutaneous indomethacin (IM), 5 mg/kg, administered 30 min prior to WIR, completely inhibited the WIR-induced increase in gastric 6-keto-PGF1␣ levels (Fig. 1A). Furthermore, IM pretreatment decreased gastric 6-keto-PGF1␣ 2– 8 h after WIR to a level significantly
296
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
Fig. 1. Changes in gastric 6-keto-PGF1␣ levels, gastric mucosal blood flow, lesion index, and gastric MPO activity in rats subjected to WIR and the effect of indomethacin pretreatment on these changes. (A) Changes in gastric of 6-keto-PGF1␣ levels, (B) gastric mucosal blood flow, (C) lesion index, and (D) gastric MPO activity were determined before (pre-WIR) and with WIR for the time points indicated in the figure. Indomethacin (IM, 5 mg/kg) was administered s.c. to rats 30 min prior to WIR, as described in Section 2. Open bars, animals subjected to WIR alone; solid bars, animals pretreated with IM prior to WIR. Values are expressed as the mean ⫾ SD derived from five animal experiments. * P ⬍ 0.01 relative to pre-WIR values. § P ⬍ 0.05 relative to pre-WIR values. ‡ P ⬍ 0.01 relative to WIR values.
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
297
Fig. 2. Effects of iloprost and leukocyte depletion on the lesion index and gastric MPO activity in rats pretreated with indomethacin and subjected to WIR. Indomethacin (IM, 5 mg/kg) was administered s.c. to rats 30 min prior to WIR. Iloprost was infused continuously at 100 ng/kg/min. Control animals received normal saline instead of iloprost. Leukocytopenia was induced by i.v. administration of nitrogen mustard (NM, 1 mg/kg) 2 days prior to the experiments. The lesion index and gastric MPO activity were determined after 8 h of WIR. Each bar represents the mean ⫾ SD of six experiments. * P ⬍ 0.01 relative to pre-WIR values. ‡ P ⬍ 0.01 relative to saline ⫹ WIR values. † P ⬍ 0.01 relative to saline ⫹ IM ⫹ WIR values.
lower than that of the pre-WIR levels (Fig. 1A). In animals pretreated with IM, gastric 6-keto-PGF1␣ levels after 30 min, 2, 4, 6, and 8 h of WIR were significantly lower compared to those of animals subjected to WIR without IM pretreatment (Fig. 1A). In animals pretreated with IM, gastric MPO activity began to increase 2 h after WIR, reaching its maximum 8 h after WIR (Fig. 1B). In animals pretreated with IM, gastric MPO activity 2– 8 h after WIR was significantly higher than that of the control animals (Fig. 1B). Gastric mucosal blood flow was significantly lower in animals subjected to WIR with IM pretreatment compared to that of animals subjected to WIR alone (Fig. 1C). WIR-induced gastric mucosal lesion formation was significantly exacerbated in animals pretreated with IM compared to that in control animals (Fig. 1D). 3.3. Effects of iloprost and NM-induced leukocytopenia on IM-induced changes in gastric accumulation of leukocytes, gastric mucosal blood flow, and gastric mucosal lesion formation Continuous intravenous infusion (8 h) of iloprost (100 ng/kg/min), a stable analog of PGI2, inhibited IM-induced increases in both gastric MPO activity (Fig. 2A) and the gastric
298
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
Fig. 3. Effect of IM on WIR-induced changes of gastric mucosal blood flow in rats. Gastric mucosal blood flow was measured by a laser-Doppler flowmeter, as described in the Section 2. IM (5 mg/kg) was administered s.c. to rats 30 min prior to WIR. Iloprost was infused continuously at a rate of 100 ng/kg/min. NM (1 mg/kg) was administered i.v. to rats 2 days prior to the experiments. Values are expressed as the mean ⫾ SD deprived from five animal experiments. F, saline ⫹ WIR; f, saline ⫹ IM ⫹ WIR; ‚, iloprost ⫹ IM ⫹ WIR; 䡺, NM ⫹ IM ⫹ WIR. * P ⬍ 0.01 relative to pre-WIR values. ‡ P ⬍ 0.01 relative to saline ⫹ WIR values. † P ⬍ 0.01 relative to saline ⫹ IM ⫹ WIR values.
lesion index (Fig. 2B) 8 h after WIR. Iloprost infusion inhibited the decrease in gastric mucosal blood flow induced by IM pretreatment (Fig. 3A). Both the IM-induced increase in the gastric lesion index and the IM-induced increase in gastric MPO activity at 8 h after WIR were significantly inhibited in animals with NM-induced leukocytopenia (Fig. 2A and B). However, NM-induced leukocytopenia did not inhibit the IM-induced decrease in gastric mucosal blood flow (Fig. 3). 3.4. Effects of iloprost and leukocytopenia on the IM-induced gastric histological changes Rat gastric mucosa 8 h after WIR was examined under light microscopy (Fig. 4). These findings were compared to those seen in nonstressed animals (Figs. 4A and F). Necrotic erosions in the epithelial surface of the gastric mucosa and marked submucosal PMN infiltration were observed in rats subjected to WIR (Figs. 4B and G). Although IM pretreatment significantly exacerbated necrotic erosions and submucosal PMN infiltration compared to those in rats subjected to WIR alone (Figs. 4C and H), the administration of iloprost or NM-induced leukocytopenia markedly improved both erosions (Figs. 4D and E) and PMN infiltration (Figs. 4I and J) in stressed animals pretreated with IM. The number of PMNs infiltrating gastric tissue in stressed animals was significantly higher than in nonstressed animals (Table 1). IM pretreatment significantly increased the numbers of infiltrating PMN compared to the number seen in animals subjected to WIR alone (Table 1). Iloprost or NM-induced leukocytopenia significantly reduced the IM-induced increase in PMN infiltration of gastric mucosa in rats subjected to WIR (Table 1).
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
299
Fig. 4. Effects of iloprost and NM-induced leukocytopenia on histological findings in WIR-induced gastric mucosal lesions in rats pretreated with IM. Histopathological examination of the stomach was performed 8 h after WIR. The left panels (hematoxylin-eosin, original magnification ⫻100) of the figure represent the complete thickness of the gastric wall. The right panels (hematoxylin-eosin, original magnification ⫻400) show the lamina propria and submucosal layer. (B and G) Necrotic erosions in the epithelial surface of the gastric mucosa and marked submucosal polymorphonuclear leukocyte (PMN) infiltration were observed in rats subjected to WIR. (A and F) Such histological changes were not observed in the stomachs of animals not subjected to WIR. (C and H) IM pretreatment markedly increased the necrotic erosions and submucosal PMN infiltration. (D and I) Administration of iloprost markedly reduced both the erosions and PMN infiltration in stressed animals pretreated with IM. (E and J) The IM-induced histopathological changes were significantly attenuated in animals with nitrogen mustard-induced leukocytopenia. The photomicrograph show one typical result of six experiments.
300
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
Table 1 Numbers of polymorphonuclear leukocytes (PMN) in rat stomach wall Group
Number of PMNa (per high power field)
Nonstress WIR IM ⫹ WIR Iloprost ⫹ IM ⫹ WIR NM-induced leukocytopenia ⫹ IM ⫹ WIR
7.4 ⫾ 2.3 13.1 ⫾ 3.2* 21.5 ⫾ 4.1† 5.1 ⫾ 2.1‡ 3.3 ⫾ 2.3‡
a The number of PMN per high-power field was counted in stomach wall sections obtained 8 h after water-immersion restraint stress (WIR), as described in Section 2. Data are expressed as mean ⫾ SD. * P ⬍ 0.01 relative to nonstress values. † P ⬍ 0.01 relative to WIR values. ‡ P ⬍ 0.01 relative to indomethacin (IM) ⫹ WIR values.
4. Discussion WIR has been widely used as an experimental model of gastric mucosal injury because of its reliable reproducibility [21]. In the present study, by using this animal model, we have demonstrated that gastric PGI2 may play an important role in the prevention of the gastric mucosal injury by inhibiting leukocyte activation. Gastric PGI2 synthesis, evaluated by measuring gastric levels of 6-keto-PGF1␣, a stable metabolite of PGI2, was found to be significantly increased 30 min after WIR. This increase was followed by a decrease to levels significantly lower than pre-WIR levels 6 and 8 h after WIR. Gastric mucosal blood flow was significantly decreased 2 h after WIR and decreased to 45% of pre-WIR levels 8 h after WIR. Although gastric mucosal lesions were not observed during the first 2 h after WIR, they were significantly increased 4 h after WIR, peaking 8 h after WIR when gastric MPO activity was significantly increased compared to pre-WIR levels. Because PGI2 inhibits leukocyte activation and increases gastric mucosal blood flow [17,26], these results suggest that gastric PGI2 prevents WIR-induced gastric mucosal injury by inhibiting the decrease in gastric mucosal blood flow and the increase in gastric accumulation of leukocytes in rats subjected to WIR. Inhibition by IM of the stress-induced increase in gastric levels of 6-keto-PGF1␣ exacerbated the gastric mucosal lesion formation and increased the gastric accumulation of leukocytes. Administration of iloprost, a stable derivative of PGI2, reversed the IM-induced pathologic events in rats subjected to WIR. NM-induced leukocytopenia significantly inhibited both the gastric accumulation of leukocytes and the exacerbation of the gastric mucosal lesion formation induced by IM in rats subjected to WIR. This suggest that inhibition of gastric accumulation of leukocytes by iloprost can be an effector rather than an effect of lesion prevention by iloprost. These observations suggest that increased gastric PGI2 synthesis may act to prevent stress-induced gastric mucosal lesion formation by inhibiting the gastric leukocyte accumulation. Gastric mucosal blood flow decreased with WIR in rats. Pretreatment with IM further decreased gastric mucosal blood flow in animals subjected to WIR in the present study. Because iloprost administration significantly inhibited the IMinduced decrease in gastric mucosal blood flow as well as the exacerbation of gastric mucosal
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
301
lesion formation, it is possible that gastric PGI2 might prevent gastric mucosal injury by maintaining the gastric mucosal blood flow. However, IM-induced mucosal lesion exacerbation was markedly reduced in animals with NM-induced leukocytopenia, with no change in the IM-induced decrease in the gastric mucosal blood flow. This observation suggests that the maintenance of gastric mucosal blood flow may not be necessary for the reduction of stress-induced gastric mucosal injury. This notion is consistent with the observation that PGE1 and PGE2, gastric-cytoprotective prostaglandins, decrease gastric mucosal blood flow secondarily because of their inhibitory effect on parietal cells [1,34]. Furthermore, the fact that PGF2␣, a cytoprotective PG, is a vasoconstrictor also supports the notion that the mechanisms of PG cytoprotection need not depend on the maintenance of mucosal blood flow [1]. Thus, it is likely that gastric PGI2 may prevent stress-induced gastric mucosal injury primarily by inhibiting the leukocyte activation. This notion is also consistent with the report by Wallace et al. [7] demonstrating that gastric mucosal injury with nonsteroidal, antiinflammatory drugs is neutrophil-dependent. Because iloprost, at a dosage of 100 ng/kg/min, inhibits gastric acid secretion [35], it is possible that the preventive effects of iloprost on IM-induced exacerbation of gastric mucosal injury in rats subjected to WIR can be at least partly explained by its antisecretory activity. Gastric acid plays an important role in gastric mucosal injury induced by various noxious stimuli [36]. Because an increase in gastric acid secretion has been shown to contribute to gastric mucosal injury induced by WIR [37] and IM [23], gastric acid may play a role in the IM-induced exacerbation of gastric mucosal injury seen in this animal model. Back diffusion of gastric acid can be induced at the site of compromised gastric mucosa, leading to the gastric mucosal injury [38]. Because activated neutrophils release neutrophil proteases and oxygen free radicals, both of which are capable of producing tissue damage [11–13], the gastric mucosa may become vulnerable to gastric acid when marginated, activated neutrophils damage the gastric mucosa. Consistent with this hypothesis is the observation by Bozkurt et al. [39] that adenosine, which inhibits neutrophil activation, inhibits the IMinduced increase in the gastric mucosal permeability. These observations suggest that iloprost inhibits the IM-induced exacerbation of gastric mucosal injury mainly by the inhibition of leukocyte activation, but also partly by its antisecretory activity. Arakawa et al. [40] have demonstrated that the gastric level of PGE2 was decreased in rats subjected to WIR. Furthermore, administration of PGE2 has been shown to be effective in preventing stress-induced gastric mucosal injury. Because PGE2 has been demonstrated to inhibit leukocyte activation [41], it is likely that leukocyte-activation inhibition is involved in the gastric cytoprotective mechanism.
References [1] Robert A. Cytoprotection by prostaglandins. Gastroenterology 1979;77:761–7. [2] Robert A, Nylander B, Andersson S. Marked inhibition of gastric secretion by two prostaglandin analogs given orally to man. Life Sci 1974;14:533– 8. [3] Robert AR, Schultz JR, Nezamis JE, Lancaster C. Gastric antisecretory and antiulcer properties of PGE2, 15-methyl PGE2 and 16,16-dimethyl PGE2. i.v., oral and intrajejunal administration. Gastroenterology 1976;70:359 –70.
302
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
[4] Konturek SJ, Robert A. Cytoprotection of canine gastric mucosa by prostacyclin: Possible mediation by increased mucosal blood flow. Digestion 1982;25:155– 63. [5] Whittle BJR, Kaufmann GL, Boughton–Smith NK. Stimulation of gastric alkaline secretion by stable prostacyclin analogs in rat and dog. Eur J Pharmacol 1984;100:277– 83. [6] Fiske SC. Peptic ulcer disease, cytoprotection, and prostaglandins. Arch Intern Med 1988;148:2112–3. [7] Wallace JL, Keenan CM, Granger DN. Gastric ulceration induced by nonsteroidal antiinflammatory drugs is a neutrophil-dependent process. Am J Physiol 1990;259:G462–7. [8] Hamajima E, Sugiyama S, Hoshino H, Goto H, Tsukamoto Y, Ozawa T. Effect of FK506, an immunosuppressive agent, on genesis of water-immersion stress-induced gastric lesions in rats. Dig Dis Sci 1994;39:713–20. [9] Kushimoto S, Okajima K, Okabe H, Binder BR. Role of granulocyte elastase in the formation of hemorrhagic shock-induced gastric mucosal lesions in the rat. Crit Care Med 1996;24:1041– 6. [10] Teppermann BL, Vozzolo BL, Soper BD. Effect of neutropenia on gastric mucosal integrity and mucosal nitric oxide synthesis in the rat. Dig Dis Sci 1993;38:2056 – 61. [11] Starkey PM. Elastase and cathepsin G. The serine protease of human neutrophil leukocytes and spleen. In: Barret AJ, editor. Proteinases in mammalian cells and tissues. Amsterdam: Elsevier/North-Holland Biomedical Press, 1977. p. 57– 89. [12] Weiss SJ, LoBuglio AF. An oxygen-dependent mechanism of neutrophil-mediated cytotoxicity. Blood 1980;55:1020 – 4. [13] Weiss SJ, Young J, LoBuglio AF, Slivka A, Nimeh NF. Role of hydrogen peroxide in neutrophil-mediated destruction of cultured endothelial cells. J Clin Invest 1981;68:714 –21. [14] Granger DN, Kubes P. The microcirculation and inflammation: Modulation of leukocyte-endothelial cell adhesion. J Leukoc Biol 1994;55:662–75. [15] Eisenhut T, Sinha B, Grottrup–Wolfers E, Semmler J, Siess W, Endres S. Prostacyclin analogs suppress the synthesis of tumor necrosis factor (TNF)-␣ in LPS-stimulated human peripheral blood mononuclear cells. Immunopharmacology 1993;26:259 – 64. [16] Oh-ishi S, Utsunomiya I, Yamamoto T, Komuro Y, Hara Y. Effects of prostaglandins and cyclic AMP on cytokine production in rat leukocytes. Eur J Pharmacol 1996;300:255–9. [17] Kainoh M, Imai R, Umetsu T, Hattori M, Nishio S. Prostacyclin and beraprost sodium as suppressors of activated rat polymorphonuclear leukocytes. Biochem Pharmacol 1990;39:477– 84. [18] Simpson PJ, Mickelson J, Fantone JC, et al. Iloprost inhibits neutrophil function in vitro and in vivo and limits experimental infarct size in canine heart. Circ Res 1987;60:666 –73. [19] Boxer LA, Allen JM, Schmidt M, Yoder M, Baehner RL. Inhibition of polymorphonuclear leukocyte adherence by prostacyclin. J Lab Clin Med 1980;95:672– 8. [20] Riva CM, Morganroth ML, Ljungman AG, et al. Iloprost inhibits neutrophil-induced lung injury and neutrohil adherence to endothelial monolayers. Am J Respir Cell Mol Biol 1990;3:301–9. [21] Takagi K, Okabe S. The effect of drugs on the production and recovery processes of the stress ulcer. Jpn J Pharmacol 1968;18:9 –17. [22] Santucci L, Fiorucci S, Giansanti M, Brunori PM, Di Matteo FM, Morelli A. Pentoxifylline prevents indomethacin induced acute gastric mucosal damage in rats: Role of TNF alpha. Gut 1994;35:909 –15. [23] Takeuchi K, Ueki S, Okabe S. Importance of gastric motility in the pathogenesis of indomethacin-induced gastric lesions in rats. Dig Dis Sci 1986;31:1114 –22. [24] Arai I, Hamasaka Y, Futaki S, et al. Effect of NS-398, a new nonsteroidal anti-inflammatory agent, on gastric ulceration and acid secreation in rats. Res Commun Chem Path Pharmacol 1993;81:259 –70. [25] Taoka Y, Okajima K, Uchiba M, et al. Reduction of spinal cord injury by administration of iloprost, a stable prostacyclin analog. J Neurosurg 1997;86:1007–11. [26] Konturek SJ, Robert A, Hanchar AJ, Nezamis JE. Comparison of prostacyclin and prostaglandin E2 on gastric secretion, gastric release, and mucosal blood flow in dogs. Dig Dis Sci 1980;25:673–9. [27] Wedmore CV, Williams TJ. Control of vascular permeability by polymorphonuclear leukocytes in inflammation. Nature (London) 1981;289:646 –50.
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
303
[28] Redfern JS, Lee E, Feldman M. Effect of indomethacin on gastric mucosal prostaglandins in human. Correlation with mucosal damage. Gastroenterology 1987;92:969 –77. [29] Yamaguchi T. Relationship between gastric mucosal hemodynamics and gastric motility. Gastroenterol Jpn 1990;25:299 –305. [30] Cooper LC, Dial EJ, Lichtenberger LM. Effect of milk, prostaglandin and antacid in experimentally-induced duodenitis in the rat. Dig Dis Sci 1990;35:1211– 6. [31] Tepperman BL, Besco JM, Kiernan JA, Soper, BD. Relationship between myeloperoxidase activity and the ontogenic response of rat gastric mucosa to ethanol. Can J Physiol Pharmacol 1991;69:1882– 8. [32] Wallace JL. Glucocorticoid-induced gastric mucosal damage. Inhibition of leukotriene but not prostaglandin biosynthesis. Prostaglandins 1987;34:311–23. [33] Krawisz JE, Sharon P, Stenson WF. Quantitative assay for acute intestinal inflammation based on myeloperoxidase activity: Assessment of inflammation in rat and hamster. Gastroenterology 1984;87:1344 –50. [34] Jacobson ED, Chaudhury TK, Thompson WJ. Mechanism of gastric mucosal cytoprotection by prostaglandins. Gastroenterology 1976;70:897 (Abstr.). [35] Seidler U, Beinborn M, Sewing K-F. Inhibition of acid formation in rabbit parietal cells by prostaglandins is mediated by the prostaglandin E2 receptor. Gastroenterology 1989;96:314 –20. [36] Yamamoto O, Okada Y, Okabe S. Effects of a proton pump inhibitor, omeprazole, on gastric secretion and gastric and duodenal ulcers or erosions in rats. Dig Dis Sci 1984;29:394 – 401. [37] Kitagawa H, Fujiwara M, Osumi Y. Effects of water-immersion stress on gastric secretion and mucosal blood flow in rats. Gastroenterology 1979;77:298 –302. [38] Wallace JL. Gastric ulceration: Critical events at the neutrophil-endothelium interface. Can J Physiol Pharmacol 1993;71:98 –102. [39] Bozkurt A, Yu¨ksel M, Haklar G, et al. Adenosine protects against indomethacin-induced gastric damage in rats. Dig Dis Sci 1998;43:1258 – 63. [40] Arakawa T, Kobayashi K, Nakamura H, et al. Effect of water-immersion stress on prostaglandin E2 in rat gastric mucosa. Gastroenterol Jpn 1981;16:236 – 41. [41] Wong K, Freund K. Inhibition of the n-formylmethionyl-leucyl-phenylalanine induced respiratory burst in human neutrophils by adrenergic agonists and prostaglandins of the E series. Can J Physiol Pharmacol 1981;59:915–20.
Gastric prostacyclin (PGI2) prevents stress-induced gastric mucosal injury in rats primarily by inhibiting leukocyte activation Naoaki Harada, Kenji Okajima*, Kazunori Murakami, Hirotaka Isobe, Wenge Liu Department of Laboratory Medicine, Kumamoto University School of Medicine, Honjo 1-1-1, Kumamoto 860-0811, Japan Received 11 March 1998; received in revised form 6 July 1998; accepted 29 October 1998
Abstract We investigated whether, in rats, gastric prostacyclin (PGI2) prevented gastric mucosal injury that was induced by water-immersion restraint stress by inhibiting leukocyte activation. Gastric levels of 6-keto-PGF1␣, a stable metabolite of PGI2, increased transiently 30 min after stress, followed by a decrease to below the baseline 6 – 8 h after stress. Gastric mucosal blood flow decreased to ⬃40% of the baseline level 8 h after stress. Myeloperoxidase activity was significantly increased 8 h after stress. Treatment with indomethacin before stress inhibited the increase in 6-keto-PGF1␣ levels and markedly reduced mucosal blood flow. It also markedly increased leukocyte accumulation and mucosal lesion formation. Iloprost, a stable PGI2 analog, inhibited the indomethacin-induced decrease in mucosal blood flow, mucosal lesion exacerbation, and increase in leukocyte accumulation. Nitrogen mustard-induced leukocytopenia inhibited the indomethacin-associated lesion exacerbation and the increase in leukocyte accumulation, but not the decreases in mucosal blood flow. These observations indicate that gastric PGI2 decreases gastric mucosal lesion formation primarily by inhibiting leukocyte accumulation. © 1999 Elsevier Science Inc. All rights reserved. Keywords: Water-immersion restraint stress; Granulocyte elastase; Gastric mucosal lesion; Gastric mucosal blood flow; Prostacyclin
* Corresponding author. Tel.: ⫹81-96-373-5281; fax: ⫹81-96-373-5281. E-mail address: [email protected] (K. Okajima) 0090-6980/99/$ – see front matter © 1999 Elsevier Science Inc. All rights reserved. PII: S 0 0 9 0 - 6 9 8 0 ( 9 9 ) 0 0 0 7 7 - X
292
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
1. Introduction Prostaglandins (PG) play an important role in preventing gastric mucosal injury, referred to as gastric cytoprotection [1]. Although PGE1 and PGE2 inhibit gastric secretion [2,3], the cytoprotective effect is considered to be independent of the antisecretory mechanism [1]. In the gastric mucosa, PGE2 and PGI2 have been shown to prevent gastric mucosal injury induced by various noxious stimuli [4,5]. This protective effect is thought to involve increases in gastric mucus secretion, bicarbonate production, and mucosal blood flow [6], although its precise mechanism has not fully been elucidated. Activated leukocytes have been implicated in gastric mucosal injuries induced by nonsteroidal anti-inflammatory drugs [7], water-immersion restraint stress [8], hemorrhagic shock [9], and ethanol [10]. Activated leukocytes release a variety of inflammatory mediators, including granulocyte elastase and reactive oxygen species that damage the endothelial cells and other tissues [11–13]. PGI2, the most abundant PG in gastric mucosa, is synthesized in the endothelium and has various physiological actions at the interface between blood and tissue [14]. Prostaglandins, such as PGE1, PGE2, and PGI2, inhibit tumor necrosis factor-␣ production by monocytes through increasing intracellular cyclic adenosine 5⬘-monophosphate levels [15,16]. Both PGI2 and iloprost, a stable derivative of PGI2, inhibit neutrophil activation by increasing intracellular cyclic adenosine 5⬘-monophosphate levels in vitro [17,18]. Furthermore, PGI2 has been shown to inhibit neutrophil adhesion to endothelial cells in vitro [19]. These actions suggest that PGI2 and iloprost prevent various types of experimental tissue injury in vivo by inhibiting leukocyte activation [20]. These observations strongly suggest that the cytoprotective action of gastric PGI2 may also involve the inhibition of leukocyte activation. However, direct evidence for this hypothesis is lacking. Therefore, we investigated whether PGI2 prevented the stress-induced gastric mucosal injury in rats through inhibition of leukocyte activation.
2. Materials and methods 2.1. Materials Indomethacin, nitrogen mustard (NM), hexadecyl-trimethylammonium bromide, and odianisidine dihydrochloride were purchased from Sigma (St. Louis, MO, USA). Iloprost, a stable analog of PGI2, was kindly supplied by Eizai Pharmaceutical Co. (Tokyo, Japan). All other reagents used were of analytical grade. 2.2. Animals Adult male Wistar rats (Nihon SLC, Hamamatsu, Japan), weighing 280 –320 g, were used in each experiment. Care and handling of animals were in accordance with National Institutes of Health guidelines. All experimental procedures described below were approved by the Kumamoto University Animal Care and Use Committee.
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
293
2.3. Experimental design Animals were divided randomly into four groups as follows: Group 1, no treatment with water-immersion restraint stress (WIR) (n ⫽ 120); Group 2, pretreatment with indomethacin (IM) ⫹ WIR (n ⫽ 120); Group 3, IM ⫹ iloprost ⫹ WIR (n ⫽ 20); Group 4, NM ⫹ IM ⫹ WIR (n ⫽ 20). 2.4. Water immersion-restraint stress-induced gastric mucosal lesion formation Before each experiment, rats were deprived of food but not water for 24 h. Then the animals were placed in a restraint cage and immersed up to the xiphoid process in water at 22°C as described previously [21]. At the indicated times during water immersion-restraint stress, the animals were anesthetized by an intraperitoneal injection of sodium pentobarbital (50 mg/kg) and exsanguinated via the abdominal aorta. The stomachs were removed, filled with 10 ml of 1% formalin, and immersed in 1% formalin for 24 h, after which they were cut along the greater curvature and examined for mucosal lesions. Because most of the observed lesions were linear with widths of ⬍2 mm, the total length of each linear hemorrhagic erosion was recorded as the lesion index (mm) by an independent observer blinded to treatment, as previously described [22]. After gross examination, the stomachs were fixed in 10% buffered formalin, and 3-m sections were stained with hematoxylineosin for histologic evaluation. To assess polymorphonuclear leukocyte (PMN) accumulation in gastric mucosa, neutrophils were counted in 10 high-power fields (⫻400), and the mean ⫾ SD was calculated. 2.5. Administration of indomethacin Indomethacin (5 mg/kg) was suspended in bicarbonate-buffered saline and administered subcutaneously (s.c.) 30 min prior to WIR. The same dose of indomethacin, administered s.c. or orally, does not affect the gastric acid secretion in intact rats as previously described [23,24]. Control animals received the same volume of bicarbonate-buffered saline without indomethacin. 2.6. Administration of iloprost Iloprost was continuously infused (100 ng/kg/min) via the jugular vein for 8 h after placement of a polyethylene tube while the rats were under light ether anesthesia. Iloprost, at a dose of 100 ng/kg/min, inhibits dermal neutrophil accumulation in a canine model of skin inflammation and inhibits neutrophil accumulation in the infarcted dog myocardium [18]. A femoral arterial catheter was placed to measure the mean arterial pressure by using strain-gauge transducer. We previously reported that mean arterial pressure fell from 105.8 ⫾ 11.4 to 91.2 ⫾ 2 mmHg in rats 30 min after intravenous infusion of iloprost at a dosage of 100 ng/kg/min [25]. Control animals received normal saline instead of iloprost. Although whether iloprost inhibits gastric acid secretion in rats has not yet been studied,
294
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
PGI2 at a dosage of 100 ng/kg/min significantly inhibits gastric acid secretion in dogs by ⬃50% [26]. 2.7. Reduction in circulating leukocytes by NM Rats were made leukocytopenic by the intravenous NM injection at a dosage of 1 mg/kg for 2 days prior to the day of the experiment [27]. Leukocytopenia was confirmed by circulating leukocyte counts, for which blood was collected from the tail vein under light ether anesthesia, immediately before stress. After smearing on a glass slide and Wright– Giemsa staining, the slides were coded to avoid observer bias and examined under a light microscope with a ⫻100 objective. Control animals were injected over a similar time course with an equal volume of normal saline. 2.8. Determination of gastric 6-keto-PGF1␣ levels Gastric levels of 6-keto-PGF1␣, a stable metabolite of PGI2, were determined in animals before and during WIR, according to the methods described previously [28]. Briefly, the stomachs were weighed and then homogenized in 5 ml of 0.1 M phosphate buffer (pH 7.4) at 5°C. The homogenates were centrifuged at 2000 ⫻ g (KR-2000T, Kubota Co., Tokyo, Japan) for 10 min to remove minute amounts of solid tissue debris. The supernatant was then acidified with 1 M HCl. 6-keto-PGF1␣ was extracted from the supernatant by using columns packed with ethyl-bonded silica gel (ethyl C2, Amersham, Buckinghamshire, UK). Columns were prepared by washing with 2 ml of methanol, followed by 2 ml of water prior to use. The acidified supernatant was applied onto the column, followed by washing sequentially with 5 ml of water, 5 ml of 10% ethanol, and 5 ml of hexane. Elution of 6-keto-PGF1␣ was performed with 5 ml of methyl formate, and the solvent was evaporated under a stream of nitrogen gas. The concentration of 6-keto-PGF1␣ was assayed by using a specific enzyme immunoassay kit (Amersham). Results are expressed as g of 6-keto-PGF1␣ per g of tissue. 2.9. Measurement of gastric mucosal blood flow Gastric mucosal blood flow was measured by laser Doppler flowmeter (ALF21N, Advance, Tokyo, Japan) as described previously [29]. Rats were anesthetized with ether, and midline laparotomies were performed. The glass fiber, connected with the hemodynamic probe, was placed against the mucosa of the corpus along the greater curvature through a proximal incision. The incisions were sutured tightly to prevent water from entering the abdominal cavity during water-immersion stress loading. Results are expressed as a percentage of pre-WIR levels. 2.10. Measurement of gastric myeloperoxidase activity After animals were immersed for the indicated period of stress, all were immediately killed. Their stomachs were quickly removed and opened along the greater curvature. In some experiments, leukocyte infiltration in gastric mucosa was assessed by determining
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
295
tissue activity of myeloperoxidase (MPO), an enzyme used as a marker for leukocyte infiltration in a variety of tissues including rat gastric mucosa [30 –32]. MPO activity was determined by a modification of the method of Krawisz et al. [33]. Briefly, the stomachs were weighed and suspended in 6 ml of 50 mM phosphate buffer (pH 6.0) containing 1% hexadecyl-trimethylammonium bromide. The samples were homogenized, and the homogenate was sonicated, freeze-thawed, and then centrifuged (4500 ⫻ g for 15 min at 4°C). MPO activity was determined in the supernatant (0.1 ml) after the addition of 0.6 ml of phosphate buffer (pH 6.0) containing 0.05 ml of 1.25 mg/ml o-dianisidine dihydrochloride and 0.05 ml of 0.05% hydrogen peroxide. The change in absorbance at 460 nm over 6.5 min was measured in a spectrophotometer (DU-54; Beckman, Irvine, CA, USA). One unit of MPO activity was defined as the amount of enzyme able to reduce 1 mol peroxide/min. Results are expressed as units of MPO activity per g of tissue. 2.11. Statistical analysis All data are expressed as the mean ⫾ SD. Comparisons among different groups of data were performed by using analysis of variance with the Scheffe´’s (post hoc) test. A P value of ⱕ5% was considered significant.
3. Results 3.1. Changes in gastric 6-keto-PGF1␣ level, gastric MPO activity, gastric mucosal blood flow, and gastric lesion index To examine whether gastric PGI2 cytoprotection involves the inhibition of leukocyte activation, we investigated the relationship between gastric levels of 6-keto-PGF1␣, a stable metabolite of PGI2, and gastric accumulation of leukocytes in rats during WIR (Fig. 1). At the same time, we measured changes in the gastric mucosal blood flow and the gastric lesion index during WIR. Gastric levels of 6-keto-PGF1␣ were significantly increased 30 min after WIR compared to pre-WIR levels (Fig. 1A). These levels began to decrease 1 h after WIR and declined to below pre-WIR levels at 6 h and 8 h after WIR (Fig. 1A). Gastric accumulation of leukocytes as evaluated by gastric MPO activity, significantly increased 8 h after WIR compared to pre-WIR levels (Fig. 1B). Gastric mucosal blood flow significantly decreased 2 h after WIR, and decreased to 45% of the pre-WIR level 8 h after WIR (Fig. 1C). The gastric lesion index significantly increased 4 h after WIR and reached a maximum 8 h after WIR (Fig. 1D). 3.2. Effect of indomethacin pretreatment on gastric 6-keto-PGF1␣ level, gastric accumulation of leukocytes, gastric mucosal blood flow, and gastric lesion index Subcutaneous indomethacin (IM), 5 mg/kg, administered 30 min prior to WIR, completely inhibited the WIR-induced increase in gastric 6-keto-PGF1␣ levels (Fig. 1A). Furthermore, IM pretreatment decreased gastric 6-keto-PGF1␣ 2– 8 h after WIR to a level significantly
296
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
Fig. 1. Changes in gastric 6-keto-PGF1␣ levels, gastric mucosal blood flow, lesion index, and gastric MPO activity in rats subjected to WIR and the effect of indomethacin pretreatment on these changes. (A) Changes in gastric of 6-keto-PGF1␣ levels, (B) gastric mucosal blood flow, (C) lesion index, and (D) gastric MPO activity were determined before (pre-WIR) and with WIR for the time points indicated in the figure. Indomethacin (IM, 5 mg/kg) was administered s.c. to rats 30 min prior to WIR, as described in Section 2. Open bars, animals subjected to WIR alone; solid bars, animals pretreated with IM prior to WIR. Values are expressed as the mean ⫾ SD derived from five animal experiments. * P ⬍ 0.01 relative to pre-WIR values. § P ⬍ 0.05 relative to pre-WIR values. ‡ P ⬍ 0.01 relative to WIR values.
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
297
Fig. 2. Effects of iloprost and leukocyte depletion on the lesion index and gastric MPO activity in rats pretreated with indomethacin and subjected to WIR. Indomethacin (IM, 5 mg/kg) was administered s.c. to rats 30 min prior to WIR. Iloprost was infused continuously at 100 ng/kg/min. Control animals received normal saline instead of iloprost. Leukocytopenia was induced by i.v. administration of nitrogen mustard (NM, 1 mg/kg) 2 days prior to the experiments. The lesion index and gastric MPO activity were determined after 8 h of WIR. Each bar represents the mean ⫾ SD of six experiments. * P ⬍ 0.01 relative to pre-WIR values. ‡ P ⬍ 0.01 relative to saline ⫹ WIR values. † P ⬍ 0.01 relative to saline ⫹ IM ⫹ WIR values.
lower than that of the pre-WIR levels (Fig. 1A). In animals pretreated with IM, gastric 6-keto-PGF1␣ levels after 30 min, 2, 4, 6, and 8 h of WIR were significantly lower compared to those of animals subjected to WIR without IM pretreatment (Fig. 1A). In animals pretreated with IM, gastric MPO activity began to increase 2 h after WIR, reaching its maximum 8 h after WIR (Fig. 1B). In animals pretreated with IM, gastric MPO activity 2– 8 h after WIR was significantly higher than that of the control animals (Fig. 1B). Gastric mucosal blood flow was significantly lower in animals subjected to WIR with IM pretreatment compared to that of animals subjected to WIR alone (Fig. 1C). WIR-induced gastric mucosal lesion formation was significantly exacerbated in animals pretreated with IM compared to that in control animals (Fig. 1D). 3.3. Effects of iloprost and NM-induced leukocytopenia on IM-induced changes in gastric accumulation of leukocytes, gastric mucosal blood flow, and gastric mucosal lesion formation Continuous intravenous infusion (8 h) of iloprost (100 ng/kg/min), a stable analog of PGI2, inhibited IM-induced increases in both gastric MPO activity (Fig. 2A) and the gastric
298
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
Fig. 3. Effect of IM on WIR-induced changes of gastric mucosal blood flow in rats. Gastric mucosal blood flow was measured by a laser-Doppler flowmeter, as described in the Section 2. IM (5 mg/kg) was administered s.c. to rats 30 min prior to WIR. Iloprost was infused continuously at a rate of 100 ng/kg/min. NM (1 mg/kg) was administered i.v. to rats 2 days prior to the experiments. Values are expressed as the mean ⫾ SD deprived from five animal experiments. F, saline ⫹ WIR; f, saline ⫹ IM ⫹ WIR; ‚, iloprost ⫹ IM ⫹ WIR; 䡺, NM ⫹ IM ⫹ WIR. * P ⬍ 0.01 relative to pre-WIR values. ‡ P ⬍ 0.01 relative to saline ⫹ WIR values. † P ⬍ 0.01 relative to saline ⫹ IM ⫹ WIR values.
lesion index (Fig. 2B) 8 h after WIR. Iloprost infusion inhibited the decrease in gastric mucosal blood flow induced by IM pretreatment (Fig. 3A). Both the IM-induced increase in the gastric lesion index and the IM-induced increase in gastric MPO activity at 8 h after WIR were significantly inhibited in animals with NM-induced leukocytopenia (Fig. 2A and B). However, NM-induced leukocytopenia did not inhibit the IM-induced decrease in gastric mucosal blood flow (Fig. 3). 3.4. Effects of iloprost and leukocytopenia on the IM-induced gastric histological changes Rat gastric mucosa 8 h after WIR was examined under light microscopy (Fig. 4). These findings were compared to those seen in nonstressed animals (Figs. 4A and F). Necrotic erosions in the epithelial surface of the gastric mucosa and marked submucosal PMN infiltration were observed in rats subjected to WIR (Figs. 4B and G). Although IM pretreatment significantly exacerbated necrotic erosions and submucosal PMN infiltration compared to those in rats subjected to WIR alone (Figs. 4C and H), the administration of iloprost or NM-induced leukocytopenia markedly improved both erosions (Figs. 4D and E) and PMN infiltration (Figs. 4I and J) in stressed animals pretreated with IM. The number of PMNs infiltrating gastric tissue in stressed animals was significantly higher than in nonstressed animals (Table 1). IM pretreatment significantly increased the numbers of infiltrating PMN compared to the number seen in animals subjected to WIR alone (Table 1). Iloprost or NM-induced leukocytopenia significantly reduced the IM-induced increase in PMN infiltration of gastric mucosa in rats subjected to WIR (Table 1).
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
299
Fig. 4. Effects of iloprost and NM-induced leukocytopenia on histological findings in WIR-induced gastric mucosal lesions in rats pretreated with IM. Histopathological examination of the stomach was performed 8 h after WIR. The left panels (hematoxylin-eosin, original magnification ⫻100) of the figure represent the complete thickness of the gastric wall. The right panels (hematoxylin-eosin, original magnification ⫻400) show the lamina propria and submucosal layer. (B and G) Necrotic erosions in the epithelial surface of the gastric mucosa and marked submucosal polymorphonuclear leukocyte (PMN) infiltration were observed in rats subjected to WIR. (A and F) Such histological changes were not observed in the stomachs of animals not subjected to WIR. (C and H) IM pretreatment markedly increased the necrotic erosions and submucosal PMN infiltration. (D and I) Administration of iloprost markedly reduced both the erosions and PMN infiltration in stressed animals pretreated with IM. (E and J) The IM-induced histopathological changes were significantly attenuated in animals with nitrogen mustard-induced leukocytopenia. The photomicrograph show one typical result of six experiments.
300
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
Table 1 Numbers of polymorphonuclear leukocytes (PMN) in rat stomach wall Group
Number of PMNa (per high power field)
Nonstress WIR IM ⫹ WIR Iloprost ⫹ IM ⫹ WIR NM-induced leukocytopenia ⫹ IM ⫹ WIR
7.4 ⫾ 2.3 13.1 ⫾ 3.2* 21.5 ⫾ 4.1† 5.1 ⫾ 2.1‡ 3.3 ⫾ 2.3‡
a The number of PMN per high-power field was counted in stomach wall sections obtained 8 h after water-immersion restraint stress (WIR), as described in Section 2. Data are expressed as mean ⫾ SD. * P ⬍ 0.01 relative to nonstress values. † P ⬍ 0.01 relative to WIR values. ‡ P ⬍ 0.01 relative to indomethacin (IM) ⫹ WIR values.
4. Discussion WIR has been widely used as an experimental model of gastric mucosal injury because of its reliable reproducibility [21]. In the present study, by using this animal model, we have demonstrated that gastric PGI2 may play an important role in the prevention of the gastric mucosal injury by inhibiting leukocyte activation. Gastric PGI2 synthesis, evaluated by measuring gastric levels of 6-keto-PGF1␣, a stable metabolite of PGI2, was found to be significantly increased 30 min after WIR. This increase was followed by a decrease to levels significantly lower than pre-WIR levels 6 and 8 h after WIR. Gastric mucosal blood flow was significantly decreased 2 h after WIR and decreased to 45% of pre-WIR levels 8 h after WIR. Although gastric mucosal lesions were not observed during the first 2 h after WIR, they were significantly increased 4 h after WIR, peaking 8 h after WIR when gastric MPO activity was significantly increased compared to pre-WIR levels. Because PGI2 inhibits leukocyte activation and increases gastric mucosal blood flow [17,26], these results suggest that gastric PGI2 prevents WIR-induced gastric mucosal injury by inhibiting the decrease in gastric mucosal blood flow and the increase in gastric accumulation of leukocytes in rats subjected to WIR. Inhibition by IM of the stress-induced increase in gastric levels of 6-keto-PGF1␣ exacerbated the gastric mucosal lesion formation and increased the gastric accumulation of leukocytes. Administration of iloprost, a stable derivative of PGI2, reversed the IM-induced pathologic events in rats subjected to WIR. NM-induced leukocytopenia significantly inhibited both the gastric accumulation of leukocytes and the exacerbation of the gastric mucosal lesion formation induced by IM in rats subjected to WIR. This suggest that inhibition of gastric accumulation of leukocytes by iloprost can be an effector rather than an effect of lesion prevention by iloprost. These observations suggest that increased gastric PGI2 synthesis may act to prevent stress-induced gastric mucosal lesion formation by inhibiting the gastric leukocyte accumulation. Gastric mucosal blood flow decreased with WIR in rats. Pretreatment with IM further decreased gastric mucosal blood flow in animals subjected to WIR in the present study. Because iloprost administration significantly inhibited the IMinduced decrease in gastric mucosal blood flow as well as the exacerbation of gastric mucosal
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
301
lesion formation, it is possible that gastric PGI2 might prevent gastric mucosal injury by maintaining the gastric mucosal blood flow. However, IM-induced mucosal lesion exacerbation was markedly reduced in animals with NM-induced leukocytopenia, with no change in the IM-induced decrease in the gastric mucosal blood flow. This observation suggests that the maintenance of gastric mucosal blood flow may not be necessary for the reduction of stress-induced gastric mucosal injury. This notion is consistent with the observation that PGE1 and PGE2, gastric-cytoprotective prostaglandins, decrease gastric mucosal blood flow secondarily because of their inhibitory effect on parietal cells [1,34]. Furthermore, the fact that PGF2␣, a cytoprotective PG, is a vasoconstrictor also supports the notion that the mechanisms of PG cytoprotection need not depend on the maintenance of mucosal blood flow [1]. Thus, it is likely that gastric PGI2 may prevent stress-induced gastric mucosal injury primarily by inhibiting the leukocyte activation. This notion is also consistent with the report by Wallace et al. [7] demonstrating that gastric mucosal injury with nonsteroidal, antiinflammatory drugs is neutrophil-dependent. Because iloprost, at a dosage of 100 ng/kg/min, inhibits gastric acid secretion [35], it is possible that the preventive effects of iloprost on IM-induced exacerbation of gastric mucosal injury in rats subjected to WIR can be at least partly explained by its antisecretory activity. Gastric acid plays an important role in gastric mucosal injury induced by various noxious stimuli [36]. Because an increase in gastric acid secretion has been shown to contribute to gastric mucosal injury induced by WIR [37] and IM [23], gastric acid may play a role in the IM-induced exacerbation of gastric mucosal injury seen in this animal model. Back diffusion of gastric acid can be induced at the site of compromised gastric mucosa, leading to the gastric mucosal injury [38]. Because activated neutrophils release neutrophil proteases and oxygen free radicals, both of which are capable of producing tissue damage [11–13], the gastric mucosa may become vulnerable to gastric acid when marginated, activated neutrophils damage the gastric mucosa. Consistent with this hypothesis is the observation by Bozkurt et al. [39] that adenosine, which inhibits neutrophil activation, inhibits the IMinduced increase in the gastric mucosal permeability. These observations suggest that iloprost inhibits the IM-induced exacerbation of gastric mucosal injury mainly by the inhibition of leukocyte activation, but also partly by its antisecretory activity. Arakawa et al. [40] have demonstrated that the gastric level of PGE2 was decreased in rats subjected to WIR. Furthermore, administration of PGE2 has been shown to be effective in preventing stress-induced gastric mucosal injury. Because PGE2 has been demonstrated to inhibit leukocyte activation [41], it is likely that leukocyte-activation inhibition is involved in the gastric cytoprotective mechanism.
References [1] Robert A. Cytoprotection by prostaglandins. Gastroenterology 1979;77:761–7. [2] Robert A, Nylander B, Andersson S. Marked inhibition of gastric secretion by two prostaglandin analogs given orally to man. Life Sci 1974;14:533– 8. [3] Robert AR, Schultz JR, Nezamis JE, Lancaster C. Gastric antisecretory and antiulcer properties of PGE2, 15-methyl PGE2 and 16,16-dimethyl PGE2. i.v., oral and intrajejunal administration. Gastroenterology 1976;70:359 –70.
302
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
[4] Konturek SJ, Robert A. Cytoprotection of canine gastric mucosa by prostacyclin: Possible mediation by increased mucosal blood flow. Digestion 1982;25:155– 63. [5] Whittle BJR, Kaufmann GL, Boughton–Smith NK. Stimulation of gastric alkaline secretion by stable prostacyclin analogs in rat and dog. Eur J Pharmacol 1984;100:277– 83. [6] Fiske SC. Peptic ulcer disease, cytoprotection, and prostaglandins. Arch Intern Med 1988;148:2112–3. [7] Wallace JL, Keenan CM, Granger DN. Gastric ulceration induced by nonsteroidal antiinflammatory drugs is a neutrophil-dependent process. Am J Physiol 1990;259:G462–7. [8] Hamajima E, Sugiyama S, Hoshino H, Goto H, Tsukamoto Y, Ozawa T. Effect of FK506, an immunosuppressive agent, on genesis of water-immersion stress-induced gastric lesions in rats. Dig Dis Sci 1994;39:713–20. [9] Kushimoto S, Okajima K, Okabe H, Binder BR. Role of granulocyte elastase in the formation of hemorrhagic shock-induced gastric mucosal lesions in the rat. Crit Care Med 1996;24:1041– 6. [10] Teppermann BL, Vozzolo BL, Soper BD. Effect of neutropenia on gastric mucosal integrity and mucosal nitric oxide synthesis in the rat. Dig Dis Sci 1993;38:2056 – 61. [11] Starkey PM. Elastase and cathepsin G. The serine protease of human neutrophil leukocytes and spleen. In: Barret AJ, editor. Proteinases in mammalian cells and tissues. Amsterdam: Elsevier/North-Holland Biomedical Press, 1977. p. 57– 89. [12] Weiss SJ, LoBuglio AF. An oxygen-dependent mechanism of neutrophil-mediated cytotoxicity. Blood 1980;55:1020 – 4. [13] Weiss SJ, Young J, LoBuglio AF, Slivka A, Nimeh NF. Role of hydrogen peroxide in neutrophil-mediated destruction of cultured endothelial cells. J Clin Invest 1981;68:714 –21. [14] Granger DN, Kubes P. The microcirculation and inflammation: Modulation of leukocyte-endothelial cell adhesion. J Leukoc Biol 1994;55:662–75. [15] Eisenhut T, Sinha B, Grottrup–Wolfers E, Semmler J, Siess W, Endres S. Prostacyclin analogs suppress the synthesis of tumor necrosis factor (TNF)-␣ in LPS-stimulated human peripheral blood mononuclear cells. Immunopharmacology 1993;26:259 – 64. [16] Oh-ishi S, Utsunomiya I, Yamamoto T, Komuro Y, Hara Y. Effects of prostaglandins and cyclic AMP on cytokine production in rat leukocytes. Eur J Pharmacol 1996;300:255–9. [17] Kainoh M, Imai R, Umetsu T, Hattori M, Nishio S. Prostacyclin and beraprost sodium as suppressors of activated rat polymorphonuclear leukocytes. Biochem Pharmacol 1990;39:477– 84. [18] Simpson PJ, Mickelson J, Fantone JC, et al. Iloprost inhibits neutrophil function in vitro and in vivo and limits experimental infarct size in canine heart. Circ Res 1987;60:666 –73. [19] Boxer LA, Allen JM, Schmidt M, Yoder M, Baehner RL. Inhibition of polymorphonuclear leukocyte adherence by prostacyclin. J Lab Clin Med 1980;95:672– 8. [20] Riva CM, Morganroth ML, Ljungman AG, et al. Iloprost inhibits neutrophil-induced lung injury and neutrohil adherence to endothelial monolayers. Am J Respir Cell Mol Biol 1990;3:301–9. [21] Takagi K, Okabe S. The effect of drugs on the production and recovery processes of the stress ulcer. Jpn J Pharmacol 1968;18:9 –17. [22] Santucci L, Fiorucci S, Giansanti M, Brunori PM, Di Matteo FM, Morelli A. Pentoxifylline prevents indomethacin induced acute gastric mucosal damage in rats: Role of TNF alpha. Gut 1994;35:909 –15. [23] Takeuchi K, Ueki S, Okabe S. Importance of gastric motility in the pathogenesis of indomethacin-induced gastric lesions in rats. Dig Dis Sci 1986;31:1114 –22. [24] Arai I, Hamasaka Y, Futaki S, et al. Effect of NS-398, a new nonsteroidal anti-inflammatory agent, on gastric ulceration and acid secreation in rats. Res Commun Chem Path Pharmacol 1993;81:259 –70. [25] Taoka Y, Okajima K, Uchiba M, et al. Reduction of spinal cord injury by administration of iloprost, a stable prostacyclin analog. J Neurosurg 1997;86:1007–11. [26] Konturek SJ, Robert A, Hanchar AJ, Nezamis JE. Comparison of prostacyclin and prostaglandin E2 on gastric secretion, gastric release, and mucosal blood flow in dogs. Dig Dis Sci 1980;25:673–9. [27] Wedmore CV, Williams TJ. Control of vascular permeability by polymorphonuclear leukocytes in inflammation. Nature (London) 1981;289:646 –50.
N. Harada et al. / Prostaglandins & other Lipid Mediators 57 (1999) 291–303
303
[28] Redfern JS, Lee E, Feldman M. Effect of indomethacin on gastric mucosal prostaglandins in human. Correlation with mucosal damage. Gastroenterology 1987;92:969 –77. [29] Yamaguchi T. Relationship between gastric mucosal hemodynamics and gastric motility. Gastroenterol Jpn 1990;25:299 –305. [30] Cooper LC, Dial EJ, Lichtenberger LM. Effect of milk, prostaglandin and antacid in experimentally-induced duodenitis in the rat. Dig Dis Sci 1990;35:1211– 6. [31] Tepperman BL, Besco JM, Kiernan JA, Soper, BD. Relationship between myeloperoxidase activity and the ontogenic response of rat gastric mucosa to ethanol. Can J Physiol Pharmacol 1991;69:1882– 8. [32] Wallace JL. Glucocorticoid-induced gastric mucosal damage. Inhibition of leukotriene but not prostaglandin biosynthesis. Prostaglandins 1987;34:311–23. [33] Krawisz JE, Sharon P, Stenson WF. Quantitative assay for acute intestinal inflammation based on myeloperoxidase activity: Assessment of inflammation in rat and hamster. Gastroenterology 1984;87:1344 –50. [34] Jacobson ED, Chaudhury TK, Thompson WJ. Mechanism of gastric mucosal cytoprotection by prostaglandins. Gastroenterology 1976;70:897 (Abstr.). [35] Seidler U, Beinborn M, Sewing K-F. Inhibition of acid formation in rabbit parietal cells by prostaglandins is mediated by the prostaglandin E2 receptor. Gastroenterology 1989;96:314 –20. [36] Yamamoto O, Okada Y, Okabe S. Effects of a proton pump inhibitor, omeprazole, on gastric secretion and gastric and duodenal ulcers or erosions in rats. Dig Dis Sci 1984;29:394 – 401. [37] Kitagawa H, Fujiwara M, Osumi Y. Effects of water-immersion stress on gastric secretion and mucosal blood flow in rats. Gastroenterology 1979;77:298 –302. [38] Wallace JL. Gastric ulceration: Critical events at the neutrophil-endothelium interface. Can J Physiol Pharmacol 1993;71:98 –102. [39] Bozkurt A, Yu¨ksel M, Haklar G, et al. Adenosine protects against indomethacin-induced gastric damage in rats. Dig Dis Sci 1998;43:1258 – 63. [40] Arakawa T, Kobayashi K, Nakamura H, et al. Effect of water-immersion stress on prostaglandin E2 in rat gastric mucosa. Gastroenterol Jpn 1981;16:236 – 41. [41] Wong K, Freund K. Inhibition of the n-formylmethionyl-leucyl-phenylalanine induced respiratory burst in human neutrophils by adrenergic agonists and prostaglandins of the E series. Can J Physiol Pharmacol 1981;59:915–20.