Since acetic acid gastric and duodenal ulcer models in rats have been established more than 30 years ago (1, 2), such models have been used to both study the mechanisms underlying the healing of chronic ulcers and screen for anti-ulcer drugs that enhance ulcer healing (3-10). Although Wang et al. (11, 12) and Ogihara and Okabe (13) have reported that a 4-wk course of indomethacin clearly delays ulcer healing in rats, the ultimate outcome of the delayed ulcers was not evaluated. We have previously reported that a 4-wk indomethacin treatment results in production of "unhealed gastric ulcers" that persist for up to 12 wks after treatment cessation (14). Such "unhealed gastric ulcers" appear to be quite sensitive to mucosal protective drugs, such as sucralfate, but considerably resistant to anti-secretory drugs, which includes acid pump inhibitors. Nonetheless, the reason why such ulcers do not heal after cessation of indomethacin treatment remains unclear. Accordingly, the present study was performed to better characterize such an "unhealed gastric ulcers" model by measuring parameters such as prostaglandin (PG) E2 biosynthesis, inflammation, and angiogenesis in the ulcer base.


MATERIALS AND METHODS

Animals
Male Donryu rats (260-280 g; Nihon SLC, Shizuoka) were used for the study. The animals were kept in a room with regulated temperature (approximately 20-22°C), humidity (approximately 55%), and light (12/12-hr light/dark cycle). To induce ulcers, animals were deprived of food for 5 hrs prior to operation to allow for easy injection of an acetic acid solution into the gastric wall. The animals were kept in mesh-bottom cages to prevent coprophagy. To determine gastric acid secretion, the ulcerated animals were deprived of food for 18 hrs and water for 2 hrs prior to experimentation. Animal maintenance and experimental procedures were carried out in accordance with the guidelines of the Ethics Committee of Kyoto Pharmaceutical University. In addition, the Board members of the Ethics Committee saw the protocol and agreed for this study.

Ulcer induction
Standardized gastric ulcers were produced according to a previously described method (1) with a slight modification. In brief, under ether anesthesia, gastric ulcers were induced by submucosal injection of 0.03 ml of 20% acetic acid (v/v) into the border between the antrum and fundus along the anterior gastric wall. The acid was injected using a 0.25 ml microsyringe (Terumo, Tokyo). After closure of the abdomen, the animals were routinely maintained with food and water. After sacrificing the animals under ether anesthesia at specified intervals, the stomachs were removed, opened along the greater curvature, and flattened with pins on a corkboard. The area (mm2) of ulceration was determined under a dissecting microscope (10x; Olympus, Tokyo) with a square grid. The person who determined the size of the ulcers was blinded as to which treatment had been administered to any given animal. Since deep, well-defined ulcers were consistently observed 5 days following acid injection, the 5th day after injection was defined as the initial day of ulceration (day 0). "Unhealed gastric ulcers" were produced by twice daily indomethacin treatment delivered as a subcutaneous (s.c.) dose of 1 mg/kg for 4 wks. Indomethacin was suspended in Tween-Saline solution and the vehicle alone was administered for the same period as the control at the volume of 1 ml/200g body weight.

Determination of PGE2 biosynthesis in gastric tissue
Mucosal PGE2 production in normal rats was determined by the method established by Lee and Feldman (15, 16). In brief, gastric specimens were removed from the stomachs and weighed 0, 3, 9, 12, and 24 hrs following the first indomethacin treatment, as well as 3 and 15 hrs after the second indomethacin treatment. The stomachs were subsequently removed and the corpus mucosa was punched out with a cork borer (9-mm internal diameter). The stomachs were then placed in 50-mM Tris-HCl (pH 8.4) buffer, packed with ice, and finely minced for approximately 15 sec with scissors. After washing and re-suspending the tissue samples in 1 ml of buffer, samples were subjected to vortex mixing at room temperature for 1 min to stimulate PGE2 production. The samples were then centrifuged at 10,000 x g for 15 sec. PGE2 levels in the resulting supernatants were determined by an enzyme immunoassay (PGE2 EIA kit; Cayman Chemicals, Ann Arbor, MI). PGE2 production levels were expressed as pg PGE2/ mg tissue/min. In another series of experiments, the effects of indomethacin on PGE2 production in the ulcerated area were determined 2 and 4 wks after treatment onset, as well as 2 and 4 wks after treatment cessation.

Determination of myeloperoxidase activity levels
Myeloperoxidase (MPO) activity levels were determined by the method established by Krawisz et al. (17). Gastric tissue samples (40 mg) from ulcerated areas were punched out with a cork borer (9-mm internal diameter), homogenized with a Polytron in 1 ml of 50 mmol/l phosphate buffer (pH 6.0) containing 0.5% hexadecyltrimethylammonium bromide (Sigma Chemical Co., St. Louis, Mo), and subsequently subjected to freeze-thawing sessions. The homogenates were then centrifuged at 1,600 g for 10 minutes. After an aliquot (5 µl) of each supernatant was mixed with 145 µl of phosphate buffer containing 0.167 mg/ml o-dianisidine dihydrochloride (Sigma) and 0.0005% H2O2, the change in the rate of absorbance at 450 nm was measured with a microplate reader (Thermo Max; Molecular Devices, Sunnyvale, CA). MPO activity levels were expressed as degradation of H2O2 µM/min/g tissue. Horseradish peroxidase (Sigma) was used as a standard.

Histological studies
At the time of autopsy, 10-µm frozen sections were prepared. Peroxidase staining was performed by washing sections in phosphate-buffered saline containing 3% Triton-X100, followed by staining with 3,3'-diaminobenzidin (DAB; Dojindo Laboratories, Kumamoto, Japan) in the presence of 0.005% H2O2. The number of peroxidase-positive cells in the ulcer edge and base was determined in three randomly chosen 1-mm2 fields. Counts were expressed as the number of peroxidase-positive cells per mm2 of ulcerated area. For angiogenesis studies, sections were incubated with an antibody for von Willebrand factor (factor VIII-related endothelial antigen; DAKO, Glostrop, Denmark) after deactivation of endogeneous peroxidase with 0.3% H2O2 and blockage of nonspecific binding sites was performed. Microvasculature was visualized by the avidin-biotin-peroxidase complex method using a Vectastain ABC-peroxidase kit (Vector, Burlingame, CA). The sections were successively stained with hematoxylin. The degree of microvasculature in the ulcer base granulation tissue was determined in three randomly chosen 1-mm2 fields. Microvasculature density was expressed as number of vessels per mm2 of ulcer base. In another series of experiments, small pieces of ulcer tissue were embedded in paraffin, sectioned at 4 µm, and then subjected to Azan.

Statistical analysis
All data are presented as means ± SEM. Statistical analysis was performed with the Student's t-test; a P< 0.05 was regarded as significant.


RESULTS

Effects of indomethacin on PGE2 biosynthesis in normal rat gastric mucosa
The effects of a single and double (9 hrs between doses) indomethacin treatment on gastric mucosal PGE2 biosynthesis were examined in normal rats. Three hrs after a single indomethacin treatment, a significant reduction in PGE2 biosynthesis was observed; the inhibition rate was 75.1% (Fig. 1). Nine and 12 hrs after treatment, such a significant reduction persisted, although to a lesser degree than the inhibition that was observed 3 hrs after treatment. The reduction was insignificant 24 hrs after treatment. When an additional indomethacin dose was administered 9 hrs after the first treatment, PGE2 biosynthesis was found to be reduced 12 and 24 hrs after the first treatment, with inhibitory rates of 71.1% and 63.1%, respectively.

Fig. 1. Time course changes for gastric mucosal prostaglandin E2 bio-synthesis in normal rats. Indomethacin was admi-nistered s.c. twice at a dose of 1 mg/ kg (0, 9 hr). Data are presented as means ± SEM for 6 animals. * Significantly different from levels in non-treated animals; P<0.05 was regarded as significant.

Effects of indomethacin on ulcer healing
The ulcerated area was 37.8 ± 4.0 mm2 (n = 6) on the day of ulceration. Indomethacin significantly prevented gastric ulcer healing after both 2- and 4-wk treatments (Fig. 2). As healing was delayed following indomethacin treatment, the ulcers maintained larger average areas after cessation of indomethacin treatment. The ulcerated area 2 and 4 wks after cessation of indomethacin treatment was 12.3 ± 3.3 mm2 and 13.3 ± 3.0 mm2, respectively, which contrasts with values of 4.4 ± 2.3 mm2 and 3.5 ± 1.3 mm2 in control animals.

Fig. 2. Time course changes in the ulcerated area of "unhealed gastric ulcers" in rats. Indomethacin was administered s.c. at a dose of 1 mg/kg twice daily for 2 and 4 wks. After cessation of indomethacin treatment, changes in the ulcerated area were monitored. Data are presented as means ± SEM for 5-6 animals. * Significantly different from the control group, with P<0.05 regarded as significant.

Effects of indomethacin on mucosal PGE2 biosynthesis in ulcerated rats
The degree of PGE2 biosynthesis in the ulcerated area was approximately 4 times that observed in normal stomachs, i.e. 109.4 ± 5.0 vs. 27.2 ± 3.3 pg/mg tissue/min (Fig. 3). Nonetheless, such increased PGE2 biosynthesis gradually decreased 6 wks after ulceration. Compared to the respective controls, the degree of PGE2 biosynthesis in indomethacin-treated rats was significantly reduced by 72.7% and 40.4% at 2 and 4 wks after indomethacin treatment, respectively. It should be noted that 2 and 4 wks after cessation of indomethacin treatment, the degree of PGE2 biosynthesis tended to be much greater than levels measured in control animals. Four wks after cessation of indomethacin treatment, the degree of PGE2 biosynthesis was similar to the level measured on ulcer day 0 (104.8 ± 17.4 vs. 109.4 ± 5.0 pg/mg tissue/min).

Fig. 3. Changes in prostaglandin E2 bio-synthesis in ulcerated tissue in rat stomachs. Indomethacin was admi-nistered 4 wks beginning the day of ulceration. Data are presented as means ± SEM for 5-6 animals. * Significantly different from the non-treated group, with P<0.05 regarded as significant. N.S.: not significant compared with the corresponding control.

MPO activity in the ulcerated area
Similar to PGE2 biosynthesis, MPO activity was extensively increased on the day of ulceration compared with normal stomachs, i.e. 486.4 ± 41.4 vs. 13.6 ± 1.7 µM H2O2/min g tissue (Fig. 4). Such enhanced activity, however, gradually returned to normal levels by 8 wks after ulceration. Indomethacin treatment for 2 or 4 wks did not affect MPO activity, resulting in activity levels similar to those observed in vehicle-treated animals. Nonetheless, MPO activity levels 2 and 4 wks after cessation of the 4-wk indomethacin treatment, i.e., 154.0 ± 21.3 and 282.8 ± 40.3 µM H2O2/min g tissue, respectively, were significantly higher than those measured in the respective controls, i.e., 68.0 ± 19.3 µM and 49.2 ± 13.3 µM H2O2/min g tissue.

Fig. 4. Myeloperoxidase activity in ulcerated tissue in rat stomachs. Note that higher activity levels were maintained in the "unhealed gastric ulcers." Data are presented as means ± SEM for 6 animals. * Significantly different from the non-treated group, with P<0.05 regarded as significant.

Histological analysis
A large number of peroxidase-positive cells were observed in the periphery of the ulcerated area after cessation of the 4-wk indomethacin treatment (Fig. 5). The number of peroxidase-positive cells in the indomethacin-treated group was approximately 2.2 times that observed in vehicle-treated animals, i.e. 288 ± 43 vs. 129 ± 44 positive cells/mm2. Indomethacin treatment significantly prevented angiogenesis in the ulcer base compared with control animals; respective microvasculature counts of 34.7 ± 1.8 and 50.0 ± 4.0 microvessels/mm2 were obtained (Fig. 6). Furthermore, severe and irregular fibrosis was observed in the base of unhealed gastric ulcers (Fig. 7).

Fig. 5. Microscopic observation of peroxidase-positive cells in the ulcer bases of normal healing ulcers (A) and "unhealed gastric ulcers" (B). Frozen sections were prepared and peroxidase staining was performed. The number of peroxidase-positive cells were counted. Data are presented as means ± SEM for 6 animals. * Significantly different from the non-treated group, with P<0.05 regarded as significant.

Fig. 6. Microscopic observation of Factor VIII-positive cells in the ulcer bases of normal healing ulcers (A) and "unhealed gastric ulcers" (B). Frozen sections were prepared and immunostaining with anti-Factor VIII antibody was performed. Factor VIII-positive cells were considered to represent newly formed microvasculature. Data are presented as means ± SEM for 6 animals. * Significantly different from the non-treated group, with P<0.05 regarded as significant.

Fig. 7. Microscopic observation of ulcerated areas in rats 8 wks after ulceration. Note the dense fibrosis in the ulcer base. Compared to rats with normal healing ulcers (A), fibrosis in "unhealed gastric ulcers" (B) was irregular.

DISCUSSION

The present study confirmed the well-established fact that repeatedly administered indomethacin markedly prevents spontaneous healing of acetic acid ulcers. In addition, the present study confirmed that gastric ulcers treated with indomethacin for 4 wks maintained their size even 4 wks after cessation.

Hayles et al. (18) reported that the healing of acetic acid ulcers was not delayed by short acting non-steroidal anti-inflammatory drugs (NSAIDs). Based on Hayles' findings, it is possible that the duration of PGE2 synthesis inhibition is related to the fact that long-acting NSAIDs prevent ulcer healing. The present study first examined the inhibitory effect of indomethacin on mucosal PGE2 synthesis in normal rats. It was found that a single indomethacin dose inhibited such production for 12 hrs, while an additional indomethacin dose resulted in continuous inhibition for at least 24 hrs after the first treatment. Accordingly, twice daily indomethacin dosing was selected to obtain a severe delay of ulcer healing.

Wang et al. (11) reported both that ulcer margin mucosal PGE2 levels were decreased following indomethacin treatment and that exogenously administered PGE2 prevented the typical delay in ulcer healing induced by indomethacin in acetic acid ulcers. Given such findings, it was postulated that the mechanism underlying indomethacin-induced delayed healing is related to a reduction in mucosal PGE2 levels. The present study confirmed that indomethacin treatment persistently reduced PGE2 synthesis in rats with gastric ulcers compared with rats without ulcers. As expected, PGE2 synthesis gradually increased after cessation of indomethacin treatment. It was surprising, however, that PGE2 levels eventually returned to levels that were similar to both levels observed prior to indomethacin treatment and levels measured in vehicle-treated animals. Given the fact that such ulcers did not heal after cessation of indomethacin treatment, it is likely that the mechanism underlying production of "unhealed gastric ulcers" remains unrelated to the degree of gastric mucosal PGE2 biosynthesis. Although PGs are generally known to represent both strong cytoprotective medications and ulcer healing drugs, increased endogenous PGE2 levels observed in the present study could not either prevent development of "unhealed gastric ulcers" or enhance ulcer healing. Such a finding implies that the gastric ulcers were altered such that indomethacin treatment reached levels whereby PGE2 was not effective. Above data strongly indicate that besides PG deficiency, other factor might be involved in "unhealed gastric ulcers". Brzozowski et al. (19) reported that indomethacin-induced delayed healing was due to suppression of endogenous PG and excessive cytokine expression and release. In addition, they reported (20) that COX-2 derived prostaglandins might play an important role in the acceleration of ulcer healing by various growth factors (bFGF, HGF, EGF). Taken together, the increase in proinflammatory cytokines such as TNF-alpha, interleukin-1ß, and/or the impairment of growth factors biosynthesis such as bFGF, HGF, EGF might contribute for mechanism underlying "unhealed gastric ulcers. "

Similar to PGE2 levels, MPO activity in the ulcerated tissue was found to be significantly higher than that measured in normal tissue, suggesting that the ulcers were initially severely inflamed on ulcer day 0. Interestingly, MPO activity in the ulcerated tissue remained elevated after cessation of indomethacin treatment, suggesting that the inflammatory reaction persisted in the ulcerated area. Indeed, histological studies confirmed marked neutrophil infiltration in the ulcer bases of "unhealed gastric ulcers." Consequently, increased MPO activity was postulated to result from neutrophil infiltration in the ulcer bases. Arakawa et al. (21) demonstrated that indomethacin treatment during the initial healing period of acetic acid ulcers both promotes persistent polymorphonuclear cell infiltration and increases the likelihood of ulcer reccurrence. Based on Arakawa's report and our findings, it would appear that continuous neutrophil infiltration might be involved in the mechanism underlying production of "unhealed gastric ulcers".

Ulcer base angiogenesis is known to represent an important factor for appropriate ulcer healing. Indeed, several reports have demonstrated that certain growth factors promote angiogenesis in ulcer bases, resulting in accelerated ulcer healing (22-26). Factor VIII immunohistochemical studies demonstrated poor angiogenesis in "unhealed gastric ulcers" compared with the non-treated control group. Given such a finding, it remains possible that angiogenesis inhibition is involved in the mechanism underlying production of "unhealed gastric ulcers".

To summarize, "unhealed gastric ulcers" are characterized by increased PGE2 synthesis, persistent neutrophil infiltration, and reduced angiogenesis in the ulcer base. The question then becomes why unhealed gastric ulcers possess such characteristics? Wallace et al. (27) reported that indomethacin administration (>20-30mg/kg) induces gastrointestinal mucosal lesions via neutrophil activation. Tarnawski et al. (28) reported that indomethacin treatment inhibited angiogenesis in acetic acid ulcers produced in rats. The present study demonstrated decreased microvasculature in the ulcer bases of "unhealed gastric ulcers," even after cessation of indomethacin treatment. Furthermore, it was also found that suppressed PGE2 synthesis levels in "unhealed gastric ulcers" eventually returned to levels observed in the non-treated control group. Accordingly, the direct action of indomethacin appears to be unrelated to continuous neutrophil infiltration and decreased ulcer base angiogenesis that is observed in "unhealed gastric ulcers." Nonetheless, there remains little doubt that indomethacin leads to production of such changes during the initial period of ulcer healing. In other words, alterations in ulcer character resulting from indomethacin treatment during the initial period of ulcer healing is thought to result in production of an "unhealed gastric ulcer." Azan staining demonstrated severe, irregular fibrosis in the ulcer base of unhealed gastric ulcers. Tsukimi et al. (29) suggested that indomethacin-induced heat shock protein 47 overexpression in ulcer bases is involved in production of ulcer base fibrosis, which results in delayed gastric ulcer healing in rats. In addition, it is of note that Ogihara and Okabe (30) previously reported that ulcer base connective tissue possesses the ability to contract, such that severe fibrosis might interfere with healing.

In conclusion, "unhealed gastric ulcers" that developed after cessation of indomethacin treatment exhibited extensively elevated PGE2 levels. Severe fibrosis, continuous neutrophil infiltration, and poor angiogenesis in the ulcer base appear to be involved in the mechanism underlying production of such ulcers.

Acknowledgments: The authors wish to thank C. J. Hurt (John Hopkins University, School of Medicine, U.S.A.) for a critical reading of the manuscript and A. Shimogai, N. Fujiwara, and S. Mori for technical assistance.


REFERENCES

1. Takagi K, Okabe S, Saziki R. A new method for the production of chronic gastric ulcer in rats and the effect of several drugs on its healing. Jpn J Pharmacol 1969; 19: 418-426.
2. Okabe S, Roth JLA, Pfeiffer CJ. A method for experimental gastric and duodenal ulcers in rats. Dig Dis Sci 1971; 27: 277-284.
3. Kobayashi N, Kataoka T, Ono A, Tsukimi Y, Okabe S. Role of p38 mitogen-activated protein kinase in the healing of gastric ulcers in rats. J Physiol Pharmacol 2001; 52: 195-210.
4. Ma L, del Soldato P, Wallace JL. Divergent effects of new cyclooxygenase inhibitors on gastric ulcer healing: Shifting the angiogenic balance. Proc Natl Acad Sci U S A. 2002; 99:13243-13247.
5. Konturek PC, Brzozowski T, Kwiecien S, et al. Effect of Helicobacter pylori on delay in ulcer healing induced by aspirin in rats. Eur J Pharmacol. 2002;451:191-202.
6. Chai J, Tarnawski AS. Serum response factor: Discovery, biochemistry, biological roles and implications for tissue injury healing. J Physiol Pharmacol 2002; 53: 147-157.
7. Slomiany BL, Slomiany A. Activation of peroxisome proliferator-activated receptor g suppreses inducible cyclooxygenase and nitric oxide synthase during oral mucosal ulcer healing. J Physiol Pharmacol 2002; 53: 159-169.
8. Sairam K, Priyambada S, Aryya NC, Goel RK. Gastroduodenal ulcer protective activity of Asparagus racemosus: an experimental, biochemical and histological study. J Ethnopharmacol. 2003; 86: 1-10.
9. Wen CY, Ito M, Wang H, et al. IL-11 up-regulates Tie-2 expression during the healing of gastric ulcers in rats. World J Gastroenterol 2003; 9:788-790.
10. Takeuchi Y, Kitano S, Bandoh T, Matsumoto T, Baatar D, Kai S. Acceleration of gastric ulcer healing by omeprazole in portal hypertensive rats. Is its action mediated by gastrin release and the stimulation of epithelial proliferation? Eur Surg Res 2003; 35:75-80.
11. Wang JY, Yamasaki S, Takeuchi K, Okabe S. Delayed healing of acetic acid-induced gastric ulcers in rats by indomethacin. Gastroenterology 1989; 96: 393-402.
12. Wang JY, Nagai H, Okabe S. Effect of omeprazole on delayed healing of acetic acid-induced gastric ulcers in rats. Jpn J Phrmacol 1990; 54: 82-85.
13. Ogihara Y, Okabe S. Effects of indomethacin and prednisolone on the different stages of healing of acetic acid-induced gastric ulcers in the rat. Jpn J Pharmacol 1993; 61: 133-136.
14. Amagase K, Okabe S. A new model, "unhealed gastric ulcers", induced by chronic treatment with indomethacin in rats with acetic acid ulcers. J Physiol Pharmacol 1999; 50: 169-181.
15. Lee M, Feldman M. Nonessential role of leukotrienes as mediators of acute gastric mucosal injury induced by aspirin in rats. Dig Dis Sci 1992; 37: 1282-1287.
16. Lee M, Feldman M. Age-related reductions in gastric mucosal prostaglandin levels increase susceptibility to aspirin-induced injury in rats. Gastroenterology 1994; 107: 1746-1750.
17. Krawisz JE, Sharon P, Stenson WF. Quantitative assay for acute intestinal inflammation based on myeloperoxidase activity. Gastroenterology 1984; 87: 1344-1350.
18. Hayles CD, Andrews FJ, O'Brien PE. Tiaprofenic acid inhibits mucosal prostaglandin E2 synthesis without delaying experimental gastric ulcer healing. J Gastroenterol Hepatol 1998; 13: 572-578.
19. Brzozowski T, Konturek PC, Konturek SJ, et al. Classic NSAID and selective cyclooxygenase (COX)-1 and COX-2 inhibitors in healing of chronic gastric ulcers. Microsc Res Tech 2001; 53: 343-353.
20. Brzozowski T, Konturek PC, Konturek SJ, et al. Effect of local application of growth factors on gastric ulcer healing and mucosal expression of cyclooxygenase-1 and -2. Digestion 2001; 64: 15-29.
21. Arakawa T, Watanabe T, Fukuda T, et al. Indomethacin treatment during initial period of acetic acid-induced rat gastric ulcer healing promotes persistent polymorphonuclear cell-infiltration and increases future ulcer recurrence. Possible mediation of prostaglandins. Dig Dis Sci 1996; 41: 2055-2061.
22. Hase S, Nakazawa S, Tsukamoto Y, Segawa K. Effects of prednisolone and human epidermal growth factor on angiogenesis in granulation tissue of gastric ulcer induced by acetic acid. Digestion 1989; 42: 135-142.
23. Tsuchida T, Tsukamoto Y, Segawa K, Goto H, Hase S. Effects of cimetidine and omeprazole on angiogenesis in granulation tissue of acetic acid-induced gastric ulcers in rats. Digestion 1990; 47: 8-14.
24. Satoh H, Shino A, Sato F, et al. Role of endogeneous basic fibroblast growth factor in the healing of gastric ulcers in rats. Jpn J Pharmacol 1997; 73, 59-71.
25. Suzuki N, Takahashi S, Okabe S. Relationship between vascular endothelial growth factor and angiogenesis in spontaneous and indomethacin-delayed healing of acetic acid-induced gastric ulcers in rats. J Physiol Pharmacol 1998; 49: 515-527.
26. Ernst H, Konturek PC, Hahn EG, Stosiek HP, Brzozowski T, Konturek SJ. Effect of local injection with basic fibroblast growth factor (BFGF) and neutralizing antibody to BFGF on gastric ulcer healing, gastric secretion, angiogenesis and gastric blood flow. J Physiol Pharmacol 2001; 52: 377-390.
27. Wallace JL, Keenan CM, Granger DN. Gastric ulceration induced by non-steroidal anti-inflammatory drugs is a neutrophil-dependent process. Am J Physiol 1990; 259: G462-467.
28. Tarnawski A, Stachura J, Douglass TG, Krause WJ, Gergely H, Sarfeh IJ. Indomethacin impairs quality of experimental gastric ulcer healing: A quantitative histological and ultrastructural analysis. In Mechanisms of Injury, Protection and Repair of the Upper Gastrointestinal Tract, A Garner, PE O'Brien (eds). Chichester, Wiley Publishers, 1991, pp. 521-531.
29. Tsukimi Y, Nakai H, Itoh S, Amagase K, Okabe S. Involvement of heat shock proteins in the healing of acetic acid-induced gastric ulcers in rats. J Physiol Pharmacol 2001; 52: 391-406.
30. Ogihara Y, Okabe S. Mechanism by which indomethacin delays gastric ulcer healing in the rat: inhibited contraction of the ulcer base. Jpn J Pharmacol 1993; 61: 123-131.