Induced microsomal PGE synthase-1 is involved in cyclooxygenase-2-dependent PGE2 production in gastric fibroblasts

We have previously shown that the cyclooxygenase (COX)-2/PGE2 pathway plays a key role in VEGF production in gastric fibroblasts. Recent studies have identified three PGE synthase (PGES) isozymes: cytosolic PGES (cPGES) and microsomal PGES (mPGES)-1 and -2, but little is known regarding the expression and roles of these enzymes in gastric fibroblasts. Thus we examined IL-1beta-stimulated mPGES-1 and cPGES mRNA and protein expression in gastric fibroblasts by quantitative PCR and Western blot analysis, respectively, and studied both their relationship to COX-1 and -2 and their roles in PGE2 and VEGF production in vitro. IL-1beta stimulated increases in both COX-2 and mPGES-1 mRNA and protein expression levels. However, COX-2 mRNA and protein expression were more rapidly induced than mPGES-1 mRNA and protein expression. Furthermore, MK-886, a nonselective mPGES-1 inhibitor, failed to inhibit IL-1beta-induced PGE2 release at the 8-h time point, while totally inhibiting PGE2 at the later stage. However, MK-886 did inhibit IL-1beta-stimulated PGES activity in vitro by 86.8%. N-(2-cyclohexyloxy-4-nitrophenyl)-methanesulfonamide (NS-398), a selective COX-2 inhibitor, totally inhibited PGE2 production at both the 8-h and 24-h time points, suggesting that COX-2-dependent PGE2 generation does not depend on mPGES-1 activity at the early stage. In contrast, NS-398 did not inhibit VEGF production at 8 h, and only partially at 24 h, whereas MK-886 totally inhibited VEGF production at each time point. These results suggest that IL-1beta-induced mPGES-1 protein expression preferentially coupled with COX-2 protein at late stages of PGE2 production and that IL-1beta-stimulated VEGF production was totally dependent on membrane-associated proteins involved in eicosanoid and glutathione metabolism (MAPEG) superfamily proteins, which includes mPGES-1, but was partially dependent on the COX-2/PGE2 pathway.     

p38 MAP kinase regulates IL-1 beta responses in cultured airway smooth muscle cells

We have previously reported that interleukin (IL)-1 beta causes beta-adrenergic hyporesponsiveness in cultured human airway smooth muscle (HASM) cells by increasing cyclooxygenase (COX)-2 expression. The purpose of this study was to determine whether p38 mitogen-activated protein (MAP) kinase is involved in these events. IL-1 beta (2 ng/ml for 15 min) increased p38 phosphorylation fourfold. The p38 inhibitor SB-203580 (3 microM) decreased IL-1 beta-induced COX-2 by 70 +/- 7% (P < 0.01). SB-203580 had no effect on PGE(2) release in control cells but caused a significant (70-80%) reduction in PGE(2) release in IL-1 beta-treated cells. IL-1 beta increased the binding of nuclear proteins to the oligonucleotides encoding the consensus sequences for activator protein (AP)-1 and nuclear factor (NF)-kappa B, but SB-203580 did not affect this binding, suggesting that the mechanism of action of p38 was not through AP-1 or NF-kappa B activation. The NF-kappa B inhibitor MG-132 did not alter IL-1 beta-induced COX-2 expression, indicating that NF-kappa B activation is not required for IL-1 beta-induced COX-2 expression in HASM cells. IL-1 beta attenuated isoproterenol-induced decreases in HASM stiffness as measured by magnetic twisting cytometry, and SB-203580 abolished this effect. These results are consistent with the hypothesis that p38 is involved in the signal transduction pathway through which IL-1 beta induces COX-2 expression, PGE(2) release, and beta-adrenergic hyporesponsiveness.     

Effects of sphondin,isolated from Heracleum laciniatum,on IL-1beta-induced cyclooxygenase-2 expression in human pulmonary epithelial cells

Recently, under large-scale screening experiments, we found that sphondin, a furanocoumarin derivative isolated from Heracleum laciniatum, possessed an inhibitory effect on IL-1beta-induced increase in the level of COX-2 protein and PGE(2) release in A549 cells. Accordingly, we examined in the present study the action mechanism of sphondin on the inhibition of IL-1beta-induced COX-2 protein expression and PGE(2) release in a human pulmonary epithelial cell line (A549). Pretreatment of cells with sphondin (10-50 microM) concentration-dependently attenuated IL-1beta-induced COX-2 protein expression and PGE(2) release. The IL-1beta-induced increase in COX-2 mRNA expression was also attenuated by sphondin (50 microM). The selective COX-2 inhibitor, NS-398 (0.01-1 microM), inhibited the activity of the COX-2 enzyme in a concentration-dependent manner, while sphondin (10-50 microM) had no effect. Sphondin (50 microM) did not affect the IL-1beta-induced activations of p44/42 MAPK, p38 MAPK, and JNK. Treatment of cells with sphondin (50 microM) or the NF-kappaB inhibitor, PDTC (50 microM) partially inhibited IL-1beta-induced degradation of IkappaB-alpha in the cytosol and translocation of p65 NF-kappaB from the cytosol to the nucleus. Furthermore, IL-1beta-induced NF-kappaB-specific DNA-protein complex formation in the nucleus was partially inhibited by sphondin (50 microM) or PDTC (50 microM). Taken together, we demonstrate that sphondin inhibits IL-1beta-induced PGE(2) release in A549 cells; this inhibition is mediated by suppressing of COX-2 expression, rather than by inhibiting COX-2 enzyme activity. The inhibitory mechanism of sphondin on IL-1beta-induced COX-2 expression may be, at least in part, through suppression of NF-kappaB activity. We conclude that sphondin may have the therapeutic potential as an anti-inflammatory drug on airway inflammation.     

The COX-2 inhibitor NS-398 causes T-cell developmental disruptions independent of COX-2 enzyme inhibition

Nonsteroidal anti-inflammatory drugs (NSAIDs) inhibit the function of cyclooxygenases, COX-1 and COX-2, which catalyze the first step in the synthesis of inflammatory mediators (PGE2). We sought to understand the roles of cyclooxygenases and NSAIDs in T-cell development. Our data show no significant defects in T-cell development in fetal thymic organ cultures of mice disrupted in both or either COX genes or in mice disrupted in either EP-1 or EP-2 receptor genes. On the other hand, NSAIDs reproducibly caused thymocyte developmental defects. However, the specific effects of the COX-2 inhibitors were not correlated with their potency for inhibition of COX-2 activity. We focused on the NS-398 COX-2 inhibitor and showed that its effects could not be reversed by exogenous PGE2. Furthermore, NS-398 was inhibitory even when its target, COX-2, was absent. These data show that the T-cell developmental effects of NS-398 are COX-2 and PGE2 independent.     

Contrasting effects of peroxisome-proliferator-activated receptor (PPAR)gamma agonists on membrane-associated prostaglandin E2 synthase-1 in IL-1beta-stimulated rat chondrocytes: evidence for PPARgamma-independent inhibition by 15-deoxy-Delta12,14prostaglandin J2

Microsomal prostaglandin E synthase (mPGES)-1 is a newly identified inducible enzyme of the arachidonic acid cascade with a key function in prostaglandin (PG)E2 synthesis. We investigated the kinetics of inducible cyclo-oxygenase (COX)-2 and mPGES-1 expression with respect to the production of 6-keto-PGF1alpha and PGE2 in rat chondrocytes stimulated with 10 ng/ml IL-1beta, and compared their modulation by peroxisome-proliferator-activated receptor (PPAR)gamma agonists. Real-time PCR analysis showed that IL-1beta induced COX-2 expression maximally (37-fold) at 12 hours and mPGES-1 expression maximally (68-fold) at 24 hours. Levels of 6-keto-PGF1alpha and PGE2 peaked 24 hours after stimulation with IL-1beta; the induction of PGE2 was greater (11-fold versus 70-fold, respectively). The cyclopentenone 15-deoxy-Delta12,14prostaglandin J2 (15d-PGJ2) decreased prostaglandin synthesis in a dose-dependent manner (0.1 to 10 microM), with more potency on PGE2 level than on 6-keto-PGF1alpha level (-90% versus -66% at 10 microM). A high dose of 15d-PGJ2 partly decreased COX-2 expression but decreased mPGES-1 expression almost completely at both the mRNA and protein levels. Rosiglitazone was poorly effective on these parameters even at 10 microM. Inhibitory effects of 10 microM 15d-PGJ2 were neither reduced by PPARgamma blockade with GW-9662 nor enhanced by PPARgamma overexpression, supporting a PPARgamma-independent mechanism. EMSA and TransAM analyses demonstrated that mutated IkappaBalpha almost completely suppressed the stimulating effect of IL-1beta on mPGES-1 expression and PGE2 production, whereas 15d-PGJ2 inhibited NF-kappaB transactivation. These data demonstrate the following in IL-1-stimulated rat chondrocytes: first, mPGES-1 is rate limiting for PGE2 synthesis; second, activation of the prostaglandin cascade requires NF-kappaB activation; third, 15d-PGJ2 strongly inhibits the synthesis of prostaglandins, in contrast with rosiglitazone; fourth, inhibition by 15d-PGJ2 occurs independently of PPARgamma through inhibition of the NF-kappaB pathway; fifth, mPGES-1 is the main target of 15d-PGJ2.     

Autoregulation of human monocyte-derived dendritic cell maturation and IL-12 production by cyclooxygenase-2-mediated prostanoid production

PG added to cell culture profoundly affect the in vitro maturation and function of monocyte-derived dendritic cells (MDC). Because unstimulated monocytes express cyclooxygenase (COX)-1, and COX-2 when activated, we examined whether MDC express these enzymes and produce prostanoids that autoregulate maturation and IL-12 production. Immature MDC (I-MDC) and mature MDC express COX-1, but, unlike monocytes, both MDC populations constitutively express COX-2. However, COX-2 regulation in both MDC populations differs from monocytes, as IL-4 does not suppress enzyme expression. COX-2 is functional in MDC as a specific inhibitor, NS-398, significantly reduces PGE(2) production. I-MDC undergoing maturation with soluble CD40 ligand (sCD40L) increase PGE(2) synthesis, but prostanoid synthesis is switched to COX-1. However, with IFN-gamma present, sCD40L-stimulated PG metabolism is redirected to COX-2, and PGE(2) synthesis increases severalfold. Endogenous PG production by MDC does not regulate CD40, CD80, CD86, or HLA DR expression; however, it does promote MDC maturation, as NS-398 significantly reduces CD83 expression in I-MDC matured with sCD40L/IFN-gamma. PG produced through COX-2 also autoregulate IL-12, but the effects are dependent on the MDC maturation state. Blocking COX-2 reduces I-MDC secretion of IL-12p40, whereas it increases IL-12p40 and p70 production by maturing MDC. COX-2-mediated PG production impacts MDC function as maturing these cells in the presence of NS-398 yields MDC that stimulate significantly more IFN-gamma in an allogeneic mixed lymphocyte response than MDC matured without this inhibitor. These studies demonstrate that MDC express both COX isoforms constitutively and produce prostanoids, which autoregulate their maturation and function.     

Interleukin-1beta induces interleukin-6 production through the production of prostaglandin E(2) in human osteoblasts, MG-63 cells.

This study was conducted to investigate the mechanism of interleukin-1beta (IL-1beta)-induced IL-6 production in human osteoblasts (MG-63 cells). Stimulation with IL-1beta resulted in the production of IL-6 and prostaglandin E(2) (PGE(2)). IL-6 production gradually increased and peaked 96 h after stimulation. IL-6 mRNA was detected between 4 and 72 h after IL-1beta stimulation. The patterns of PGE(2) production and the expression of cyclooxygenase-2 (COX-2) mRNA were biphasic after stimulation. Actinomycin D, cycloheximide, indomethacin, and NS-398 (COX-2 inhibitor) suppressed the production of IL-6 and PGE(2). Anti-PGE(2) antibody markedly reduced the production of IL-6. In addition, stimulation with 17-phenyl-PGE(2), a PGE receptor-1 (EP-1 receptor) agonist, led to the expression of IL-6 mRNA after pretreatment with IL-1beta. These findings indicate that IL-1beta-induced IL-6 production in MG-63 cells involves the following sequence of steps: IL-1beta-induced COX-2 activation, PGE(2) production, and EP-1 receptor signaling prior to IL-6 production.     

Interferon-gamma and interleukin 4 inhibit interleukin 1beta-induced delayed prostaglandin E(2)generation through suppression of cyclooxygenase-2 expression in human fibroblasts

Interleukin (IL-)1 stimulates prostaglandin E(2)(PGE(2)) generation in fibroblasts, and preferential couplings between particular phospholipase A(2)(PLA(2)) and cyclooxygenase (COX) isozymes are implicated with IL-1-induced delayed PGE(2)generation. The regulatory effects of interferon (IFN)-gamma and IL-4 on IL-1beta-induced COX, PLA(2)isoforms expression and terminal delayed PGE(2)generation were examined in three types of human fibroblasts. These human fibroblasts constitutively expressed cytosolic PLA(2)(cPLA(2)) and COX-1 enzymes, and exhibited delayed PGE(2)generation in response to IL-1beta. IL-1beta also stimulated expression of cPLA(2)and COX-2 only, while constitutive and IL-1beta-induced type IIA and type V secretory PLA(2)s (sPLA(2)s) expression could not be detected. A COX-2 inhibitor and cPLA(2)inhibitor markedly suppressed the IL-1beta-induced delayed PGE(2)generation, while a type IIA sPLA(2)inhibitor failed to affect it. IFN-gamma and IL-4 dramatically inhibited the IL-1beta-induced delayed PGE(2)generation; these cytokines apparently suppressed IL-1beta-stimulated COX-2 expression and only weakly suppressed cPLA(2)expression in response to IL-1beta. These results indicate that IL-1beta-induced delayed PGE(2)generation in these human fibroblasts mainly depends on de novo induction of COX-2 and cPLA(2), irrespective of the constitutive presence of COX-1, and that IFN-gamma and IL-4 inhibit IL-1beta-induced delayed PGE(2)generation by suppressing, predominantly, COX-2 expression.     

Regulation of bone morphogenetic protein-2 expression by endogenous prostaglandin E2 in human mesenchymal stem cells

Cyclooxygenase (COX)-2 is generally known as an inducible enzyme, and it produces arachidonic acid to prostaglandin E2 (PGE2), which modulates bone metabolism. Here, we investigated the expression and role of COX isomers in human mesenchymal stem cells. Human mesenchymal stem cells constitutively expressed COX-2 as well as COX-1, and secretion of PGE2 was completely inhibited by NS-398, a specific inhibitor of COX-2. Levels of secreted PGE2 were strikingly higher in human mesenchymal stem cells than in osteoblastic cells differentiated from the mesenchymal cells. This higher production of PGE2 in mesenchymal stem cells was due to higher expression of membrane-associated PGE synthase (mPGES) regulated by early growth response factor-1 (Egr-1). Treatment of human mesenchymal stem cells with NS-398 suppressed expression of bone morphogenetic protein-2 (BMP-2). The suppression of BMP-2 by NS-398 was abrogated by an EP4 receptor agonist as well as by PGE2. Moreover, BMP-2 expression was suppressed by an EP4 receptor antagonist. These data indicate that PGE2 produced by COX-2 increases BMP-2 expression via binding the EP4 receptor.     

15-Deoxy-delta(12,14)-prostaglandin J(2) inhibits IL-1beta-induced IKK enzymatic activity and IkappaBalpha degradation in rat chondrocytes through a PPARgamma-independent pathway

Peroxisome proliferator-activated receptor gamma (PPARgamma) ligands have been shown to inhibit the effects of proinflammatory cytokines such as interleukin-1beta (IL-1beta). This cytokine plays a key role in articular pathophysiologies by inducing the production of inflammatory mediators such as nitric oxide (NO) and prostaglandin E(2) (PGE(2)). We previously demonstrated that 15d-PGJ(2) was more potent than troglitazone to counteract IL-1beta effects on chondrocytes. Here, we studied the action of 15d-PGJ(2) on intracellular targets in nuclear factor-kappaB (NF-kappaB) signalling pathway in IL-1beta treated rat chondrocytes. We found that 15d-PGJ(2) decreased inhibitor kappaBalpha (IkappaBalpha) degradation but not its phosphorylation by specifically inhibiting IkappaB kinase beta (IKKbeta), but not IKKalpha, enzymatic activity. We further evaluated the involvement of PPARgamma in the anti-inflammatory action of its ligands. In chondrocytes overexpressing functional PPARgamma protein, 15d-PGJ(2) pre-treatment inhibited inducible NO synthase and COX-2 mRNA expression, nitrite and PGE(2) production, p65 translocation and NF-kappaB activation. Troglitazone or rosiglitazone pre-treatment had no effect. 15d-PGJ(2) exhibited the same effect in chondrocytes overexpressing mutated PPARgamma protein. These results suggest that 15d-PGJ(2) exerts its anti-inflammatory effect in rat chondrocytes by a PPARgamma-independent mechanism, which can be conferred to a partial inhibition of IkappaBalpha degradation.     

Oxidized low density lipoprotein activates peroxisome proliferator-activated receptor-alpha (PPARalpha) and PPARgamma through MAPK-dependent COX-2 expression in macrophages

It has been reported that oxidized low density lipoprotein (Ox-LDL) can activate both peroxisome proliferator-activated receptor-alpha (PPARalpha) and PPARgamma. However, the detailed mechanisms of Ox-LDL-induced PPARalpha and PPARgamma activation are not fully understood. In the present study, we investigated the effect of Ox-LDL on PPARalpha and PPARgamma activation in macrophages. Ox-LDL, but not LDL, induced PPARalpha and PPARgamma activation in a dose-dependent manner. Ox-LDL transiently induced cyclooxygenase-2 (COX-2) mRNA and protein expression, and COX-2 specific inhibition by NS-398 or meloxicam or small interference RNA of COX-2 suppressed Ox-LDL-induced PPARalpha and PPARgamma activation. Ox-LDL induced phosphorylation of ERK1/2 and p38 MAPK, and ERK1/2 specific inhibition abrogated Ox-LDL-induced COX-2 expression and PPARalpha and PPARgamma activation, whereas p38 MAPK-specific inhibition had no effect. Ox-LDL decreased the amounts of intracellular long chain fatty acids, such as arachidonic, linoleic, oleic, and docosahexaenoic acids. On the other hand, Ox-LDL increased intracellular 15-deoxy-Delta(12,14)-prostaglandin J(2) (15d-PGJ(2)) level through ERK1/2-dependent overexpression of COX-2. Moreover, 15d-PGJ(2) induced both PPARalpha and PPARgamma activation. Furthermore, COX-2 and 15d-PGJ(2) expression and PPAR activity were increased in atherosclerotic lesions of apoE-deficient mice. Finally, we investigated the involvement of PPARalpha and PPARgamma on Ox-LDL-induced mRNA expression of ATP-binding cassette transporter A1 and monocyte chemoattractant protein-1. Interestingly, specific inhibition of PPARalpha and PPARgamma suppressed Ox-LDL-induced ATP-binding cassette transporter A1 mRNA expression and enhanced Ox-LDL-induced monocyte chemoattractant protein-1 mRNA expression. In conclusion, Ox-LDL-induced increase in 15d-PGJ(2) level through ERK1/2-dependent COX-2 expression is one of the mechanisms of PPARalpha and PPARgamma activation in macrophages. These effects of Ox-LDL may control excess atherosclerotic progression.     

Anti-inflammatory effect of two isoforms of COX in H. pylori-induced gastritis in mice: possible involvement of PGE2

Neutrophil infiltration mediated by TNF-alpha is associated with various types of gastric injury, whereas PGs play a crucial role in gastric defense. We examined roles of two isoforms of cyclooxygenase (COX) and PGE2 in Helicobacter pylori-induced gastritis in mice. Mice infected with H. pylori were given selective COX-1 inhibitor SC-560 (10 mg/kg), selective COX-2 inhibitor NS-398 (10 mg/kg), or nonselective COX inhibitor indomethacin (2 mg/kg) with or without 16,16-dimethyl PGE2 for 1 wk. H. pylori infection increased levels of mRNA for COX-1 and -2 in gastric tissue by 1.2-fold and 3.3-fold, respectively, accompanied by a significant increase in PGE2 production by gastric tissue. H. pylori infection significantly elevated MPO activity, a marker of neutrophil infiltration, and epithelial cell apoptosis in the stomach. SC-560 augmented MPO activity and epithelial cell apoptosis with associated reduction in PGE2 production, whereas NS-398 had the same effects without affecting PGE2 production. Inhibition of both COX-1 and -2 by indomethacin or concurrent treatment with SC-560 and NS-398 resulted in a stronger increase in MPO activity and apoptosis than inhibition of either COX-1 or -2 alone. H. pylori infection elevated TNF-alpha mRNA expression in the stomach, which was further increased by indomethacin. Effects of COX inhibitors on neutrophil infiltration, apoptosis, and TNF-alpha expression in H. pylori-infected mice were abolished by exogenous 16,16-dimethyl PGE2. In conclusion, PGE2 derived from either COX-1 or -2 is involved in regulation of gastric mucosal inflammation and contributes to maintenance of mucosal integrity during H. pylori infection via inhibition of TNF-alpha expression.     

Role of cyclooxygenase-2 in Helicobacter pylori- induced gastritis in Mongolian gerbils

Cyclooxygenase (COX)-2 expression is induced in the gastric mucosa of Helicobacter pylori-infected patients, but its role remains unclear. We examined the effects of NS-398 and indomethacin on gastric pathology in H. pylori-infected Mongolian gerbils. COX-1 was detected in both normal and H. pylori-infected mucosa, whereas COX-2 was expressed only in the infected mucosa. PGE(2) production was elevated by H. pylori infection, and the increased production was reduced by NS-398, which did not affect PGE(2) production in normal mucosa. Indomethacin inhibited PGE(2) production in both normal and infected mucosa. Hemorrhagic erosions, neutrophil infiltration, lymphoid follicles, and epithelium damage were induced by H. pylori infection. NS-398 and indomethacin aggravated these pathological changes but did not increase viable H. pylori number. H. pylori-increased production of neutrophil chemokine and interferon-gamma was potentiated by NS-398 and indomethacin. Neither NS-398 nor indomethacin caused any pathological changes or cytokine production in normal animals. These results indicate that COX-2 as well as COX-1 might play anti-inflammatory roles in H. pylori-induced gastritis.     

Piroxicam and NS-398 rescue neurones from hypoxia/reoxygenation damage by a mechanism independent of cyclo-oxygenase inhibition

We studied whether NS-398, a selective cyclo-oxygenase-2 (COX-2) enzyme inhibitor, and piroxicam, an inhibitor of COX-2 and the constitutively expressed COX-1, protect neurones against hypoxia/reoxygenation injury. Rat spinal cord cultures were exposed to hypoxia for 20 h followed by reoxygenation. Hypoxia/reoxygenation increased lactate dehydrogenase (LDH) release, which was inhibited by piroxicam (180-270 microM) and NS-398 (30 microM). Cell counts confirmed the neuroprotection. Western blotting revealed no COX-1 or COX-2 proteins even after hypoxia/reoxygenation. Production of prostaglandin E2 (PGE2), a marker of COX activity, was barely measurable and piroxicam and NS-398 had no effect on the negligible PGE2 production. Hypoxia/reoxygenation increased nuclear factor-kappa B (NF-kappaB) binding activity, which was inhibited by piroxicam but not by NS-398. AP-1 binding activity after hypoxia/reoxygenation was inhibited by piroxicam but strongly enhanced by NS-398. However, both COX inhibitors induced activation of extracellular signal-regulated kinase (ERK) in neurones and phosphorylation of heavy molecular weight neurofilaments, cytoskeletal substrates of ERK. It is concluded that piroxicam and NS-398 protect neurones against hypoxia/reperfusion. The protection is independent of COX activity and not solely explained by modulation of NF-kappaB and AP-1 binding activity. Instead, piroxicam and NS-398-induced phosphorylation through ERK pathway may contribute to the increased neuronal survival.     

Cyclooxygenase-2-regulated vascular endothelial growth factor release in gastric fibroblasts

VEGF is a highly specific stimulator of endothelial cells and may play an important role in angiogenesis in the process of tissue regeneration. We previously showed that cyclooxygenase-2 (COX-2) expressed in mesenchymal cells of the ulcer bed is involved in the ulcer repair process. To clarify the role of COX-2 in angiogenesis during gastric ulcer healing, we investigated the relation between COX-2 expression and VEGF production in human gastric fibroblasts in vivo and in vitro. Gastric fibroblasts were cultured in RPMI 1640 with and without IL-1alpha or IL-1beta in the presence or absence of NS-398, a selective COX-2 inhibitor. Supernatant VEGF and PGE(2) concentrations were measured by enzyme-linked immunosorbent assay. COX-2 expression in fibroblasts was determined by Western blot analysis. VEGF and COX-2 expression in surgical resections of human gastric ulcer tissue was examined immunohistochemically. IL-1 dose dependently enhanced VEGF release in cultured gastric fibroblasts after a 24-h stimulation. IL-1 also stimulated PGE(2) production in gastric fibroblasts via COX-2 induction. NS-398 significantly suppressed VEGF and PGE(2) release from IL-1-stimulated gastric fibroblasts; concurrent addition of PGE(2) restored NS-398-inhibited VEGF release. COX-2 and VEGF immunoreactivity were colocalized in fibroblast-like cells in the ulcer bed of gastric tissues. These results suggest that COX-2 plays a key role in VEGF production in gastric fibroblasts stimulated by IL-1 in vitro and that angiogenesis induced by the COX-2-VEGF pathway might be involved in gastric ulcer healing.