Differential regulation of prostaglandin E2 and thromboxane A2 production in human monocytes: implications for the use of cyclooxygenase inhibitors

There is an autocrine relationship between eicosanoid and cytokine synthesis, with the ratio of prostaglandin E2 (PGE2)/thromboxane A2 (TXA2) being one of the determinants of the level of cytokine synthesis. In monocytes, cyclooxygenase type 1 (COX-1) activity appears to favor TXA2 production and COX-2 activity appears to favor PGE2 production. This has led to speculation regarding possible linkage of COX isozymes with PGE and TXA synthase. We have studied the kinetics of PGE2 and TXA2 synthesis under conditions that rely on COX-1 or -2 activity. With small amounts of endogenously generated prostaglandin H2 (PGH2), TXA2 synthesis was greater than PGE2. With greater amounts of endogenously generated PGH2, PGE2 synthesis was greater than TXA2. Also, TXA synthase was saturated at lower substrate concentrations than PGE synthase. This pattern was observed irrespective of whether PGH2 was produced by COX-1 or COX-2 or whether it was added directly. Furthermore, the inhibition of eicosanoid production by the action of nonsteroidal anti-inflammatory drugs or by the prevention of COX-2 induction with the p38 mitogen-activated protein kinase inhibitor SKF86002 was greater for PGE2 than for TXA2. It is proposed that different kinetics of PGE synthase and TXA synthase account for the patterns of production of these eicosanoids in monocytes under a variety of experimental conditions. These properties provide an alternative explanation to notional linkage or compartmentalization of COX-1 or -2 with the respective terminal synthases and that therapeutically induced changes in eicosanoid ratios toward predominance of TXA2 may have unwanted effects in long-term anti-inflammatory and anti-arthritic therapy.     

Inhibition of phospholipase activity in human monocytes by IFN-gamma blocks endogenous prostaglandin E2-dependent collagenase production

Previous studies from our laboratory have demonstrated that exposure of human monocytes to a stimulant, such as Con A, results in the production of the enzyme collagenase through PGE2-dependent pathway. Inasmuch as rIFN-gamma has been shown to modulate monocyte/macrophage PG synthesis, we examined the effect of rIFN-gamma on the activation sequence leading to collagenase production. The addition of rIFN-gamma (10 to 1000 U/ml) to Con A-stimulated monocytes resulted in a dose-dependent inhibition of PGE2 and collagenase synthesis. The suppression of collagenase production by rIFN-gamma was related to its ability to reduce PGE2 levels as demonstrated by the restoration of collagenase activity by the addition of PGE2. HPLC analysis of the arachidonic acid (AA) metabolites released by monocytes showed that rIFN-gamma caused a reduction in the release of AA and products of the cyclooxygenase and lipoxygenase pathways. These data indicated that rIFN-gamma decreased eicosanoid production by inhibiting the release of AA from phospholipids. This conclusion was supported by the reduction in membrane bound phospholipase activity in rIFN-gamma-treated monocytes. Moreover, the inhibition by rIFN-gamma of PGE2 and collagenase was reversed by the addition of phospholipase A2. Our findings demonstrate that rIFN-gamma inhibits phospholipase activity in activated monocytes and as a result blocks PGE2-dependent collagenase synthesis.     

11,12-epoxyeicosatrienoic acid attenuates synthesis of prostaglandin E2 in rat monocytes stimulated with lipopolysaccharide

Cytochrome P-450 monooxygenase (epoxygenase)-derived arachidonic acid (AA) metabolites, including 11,12-epoxyeicosatrienoic acid (11,12-EET), possess anti-inflammatory and antipyretic properties. Prostaglandin E2 (PGE2), a cyclooxygenase (COX)-derived metabolite of AA, is a well-defined mediator of fever and inflammation. We have tested the hypothesis that 11,12-EET attenuates synthesis of PGE2 in monocytes, which are the cells that are indispensable for induction of fever and initiation of inflammation. Monocytes isolated from freshly collected rat blood were stimulated with lipopolysaccharide (LPS; 100 ng/2 x 10(5) cells) to induce COX-2 and stimulate generation of PGE2. SKF-525A, an inhibitor of epoxygenases, significantly augmented the lipopolysaccharide-provoked synthesis of PGE2 in cell culture in a concentration-dependent manner. It did not affect, however, elevation of the expression of COX-2 protein in monocytes stimulated with LPS. 11,12-EET also did not affect the induction of COX-2 in monocytes incubated with lipopolysaccharide. However, 11,12-EET suppressed, in a concentration-dependent fashion, the generation of PGE2 in incubates. Preincubation of a murine COX-2 preparation for 0-5 min with three concentrations of 11,12-EET (1, 5, and 10 microM) inhibited the oxygenation of [14C]-labeled AA by the enzyme. The inhibitory effect of 11,12-EET on COX-2 was time-and-concentration-dependent, suggesting a mechanism-based inhibition. Based on these data, we conclude that 11,12-EET suppresses generation of PGE2 in monocytes via modulating the activity of COX-2. These data support the hypothesis that epoxygenase-derived AA metabolites constitute a negative feedback on the enhanced synthesis of prostaglandins upon inflammation.     

Blocking tumor cell eicosanoid synthesis by GP x 4 impedes tumor growth and malignancy

Using tumor cell-restricted overexpression of glutathione peroxidase 4 (GP x 4), we investigated the contribution of tumor cell eicosanoids to solid tumor growth and malignant progression in two tumor models differing in tumorigenic potential. By lowering cellular lipid hydroperoxide levels, GP x 4 inhibits cyclooxygenase (COX) and lipoxygenase (LOX) activities. GP x 4 overexpression drastically impeded solid tumor growth of weakly tumorigenic L929 fibrosarcoma cells, whereas B16BL6 melanoma solid tumor growth was unaffected. Yet, GP x 4 overexpression did markedly increase the sensitivity of B16BL6 tumors to angio-destructive TNF-alpha therapy and abolished the metastatic lung colonizing capacity of B16BL6 cells. Furthermore, the GP x 4-mediated suppression of tumor cell prostaglandin E(2) (PGE(2)) production impeded the induction of COX-2 expression by the tumor stress conditions hypoxia and inflammation. Thus, our results reflect a PGE(2)-driven positive feedback loop for COX-2 expression in tumor cells. This was further supported by the restoration of COX-2 induction capacity of GP x 4-overexpressing L929 tumor cells when cultured in the presence of exogenous PGE(2). Thus, although COX-2 expression and eicosanoid production may be enabled by PGE(2) from the tumor microenvironment, our results demonstrate the predominant tumor cell origin of protumoral eicosanoids, promoting solid tumor growth of weakly tumorigenic tumors and malignant progression of strongly tumorigenic tumors.     

Identification of a Feed-Forward Loop Between 15(S)-HETE and PGE2 in Human Amnion at Parturition

Human parturition is associated with massive arachidonic acid (AA) mobilization in the amnion, indicating that large amounts of AA-derived eicosanoids are required for parturition. Prostaglandin E2 (PGE2) synthesized from the cyclooxygenase (COX) pathway is the best characterized AA-derived eicosanoid in the amnion which plays a pivotal role in parturition. The existence of any other pivotal AA-derived eicosanoids involved in parturition remains elusive. Here, we screened such eicosanoids in human amnion tissue with AA-targeted metabolomics and studied their role and synthesis in parturition by using human amnion fibroblasts and a mouse model. We found that lipoxygenase (ALOX) pathway-derived 15(S)-hydroxyeicosatetraenoic acid (15(S)-HETE) and its synthetic enzymes ALOX15 and ALOX15B were significantly increased in human amnion at parturition. Although 15(S)-HETE is ineffective on its own, it potently potentiated the activation of NF-κB by inflammatory mediators including lipopolysaccharide, interleukin-1β, and serum amyloid A1, resulting in the amplification of COX-2 expression and PGE2 production in amnion fibroblasts. In turn, we determined that PGE2 induced ALOX15/15B expression and 15(S)-HETE production through its EP2 receptor-coupled PKA pathway, thereby forming a feed-forward loop between 15(S)-HETE and PGE2 production in the amnion at parturition. Our studies in pregnant mice showed that 15(S)-HETE injection induced preterm birth with increased COX-2 and PGE2 abundance in the fetal membranes and placenta. Conclusively, 15(S)-HETE is identified as another crucial parturition-pertinent AA-derived eicosanoid in the amnion, which may form a feed-forward loop with PGE2 in parturition. Interruption of this feed-forward loop may be of therapeutic value for the treatment of preterm birth.     

PGE(2) is generated by specific COX-2 activity and increases VEGF production in COX-2-expressing human pancreatic cancer cells

In some cancers cyclooxygenase (COX) inhibition appears to be anti-mitogenic and anti-angiogenic, but the actions of COX-derived prostaglandins in pancreatic cancer (PaCa) are unknown. In this study COX-2 was detected in three of six PaCa cell lines while COX-1 was identified in all cell lines. COX-2 expression correlated with basal and arachidonic acid (AA) stimulated PGE(2) production. PGE(2) production was inhibited by the COX-2 inhibitor nimesulide. In COX-2 expressing cells, exogenous AA and PGE(2) increased VEGF synthesis via the EP(2) receptor. Whereas PGE(2) stimulated intracellular cAMP formation in COX-2 positive and negative cells, 8-bromo cAMP stimulated VEGF production only in COX-2 expressing cells. Stimulating COX-2 expressing PaCa cell lines with AA enhanced migration of endothelial cells, an effect which was inhibited by a COX-2 inhibitor and EP(2) receptor antagonist. These data identify a subset of human PaCa cell lines that express functional COX-2 enzyme. PGE(2) generated by specific COX-2 activity increases VEGF secretion in human PaCa cells through an autocrine mechanism.     

Synergistic activation of human monocytes by granulocyte-macrophage colony-stimulating factor and IFN-gamma. Increased TNF-alpha but not IL-1 activity

TNF-alpha and IL-1 activities and PGE2 levels were investigated in the supernatants of highly purified human monocytes cultured for 18 h with recombinant human granulocyte-macrophage CSF (GM-CSF). GM-CSF alone did not stimulate IL-1 or TNF-alpha activities or the production of PGE2. GM-CSF with IFN-gamma, but not with LPS, consistently activated the monocytes for TNF-alpha activity. In contrast, for increased IL-1 activity, GM-CSF synergized weakly and irregularly with LPS, but not at all with IFN-gamma. For the third monocyte product investigated, GM-CSF was a weak and inconsistent inducer of PGE2 and only in the co-presence of IFN-gamma. Thus, GM-CSF can elicit different responses in human monocytes depending both on the co-stimulus as well as the monocyte product being investigated.     

7beta-Hydroxy-epiandrosterone-mediated regulation of the prostaglandin synthesis pathway in human peripheral blood monocytes

7alpha-Hydroxy-DHEA, 7beta-hydroxy-DHEA and 7beta-hydroxy-EpiA are native metabolites of dehydroepiandrosterone (DHEA) and epiandrosterone (EpiA). Since numerous steroids are reported to interfere with inflammatory and immune processes, our objective was to test the effects of these hydroxysteroids on prostaglandin (PG) production and related enzyme gene expression. Human peripheral blood monocytes were cultured for 4 and 24 h in the presence of each of the steroids (1-100 nM), with and without addition of TNF-alpha (10 ng/mL). Levels of PGE(2), PGD(2) and 15-deoxy-Delta(12,14)-PGJ(2) (15d-PGJ(2)) were measured in the incubation medium, and cell content of cyclooxygenase (COX-2), and PGE and PGD synthases (m-PGES1, H-PGDS, L-PGDS), and peroxisome proliferator activated receptor (PPAR-gamma) was assessed by quantitative RT-PCR and Western blots. Addition of TNF-alpha resulted in elevated PG production and increased COX-2 and m-PGES1 levels. Among the three steroids tested, only 7beta-hydroxy-EpiA decreased COX-2, m-PGES1 and PPAR-gamma expression while markedly decreasing PGE(2) and increasing 15d-PGJ(2) production. These results suggest that 7beta-hydroxy-EpiA is a native trigger of cellular protection through simultaneous activation of 15d-PGJ(2) and depression of PGE(2) synthesis, and that these effects may be mediated by activation of a putative receptor, specific for 7beta-hydroxy-EpiA.     

Effects of TGF-beta, TNF-alpha, IL-beta and IL-6 alone or in combination, and tyrosine kinase inhibitor on cyclooxygenase expression, prostaglandin E2 production and bone resorption in mouse calvarial bone cells

Cyclooxygenase-2 (COX-2) and tyrosine kinase, which are involved in the biosynthesis of prostaglandin E(2) (PGE(2)) in mouse calvarial osteoblasts, are stimulated by cytokine interleukin-1beta (IL-1beta), tumor necrosis factor-alpha (TNF-alpha) and/or interleukin-6 (IL-6). IL-1beta and IL-6 and, to a lesser extent, TNF-alpha, enhances COX-2 mRNA levels in calvarial osteoblasts. Simultaneous treatment with IL-6 and IL-1beta and TNF-alpha resulted in enhanced COX-2 mRNA levels accompanied by the cooperative stimulation of PGE(2) biosynthesis compared to cells treated with IL-1beta or TNF-alpha or IL-6 alone. In contrast, the presence of TGF-beta reduced COX-2 mRNA level, PGE(2) biosynthesis and bone resorption induced by IL-1beta, TNF-alpha, IL-6 or a combination thereof. However, neither IL-1beta, TNF-alpha, IL-6 nor a combination of IL-1beta, TNF-alpha, IL-6 enhanced COX-1 mRNA levels in calvarial osteoblasts. A novel Src tyrosine kinase inhibitor, Herbimycin A (HERB), reduced COX-2 mRNA levels as well as PGE(2) production induced by IL-1beta, TNF-alpha and IL-6 or a combination of IL-1beta, TNF-alpha, IL-6, whereas COX-1 mRNA levels remained unaffected. Finally, HERB was found to inhibit in vitro bone resorption. These results indicate that the cooperative effects of IL-beta, TNF-alpha, IL-6 on PGE(2) production are due to the enhanced expression of the COX-2 gene and that tyrosine kinase(s) are involved in COX-2 signal transduction in mouse calvarial osteoblasts. Thus, the Src family of kinase inhibitors may be useful in treating diseases associated with elevated bone loss.     

Cyclooxygenases-1 and 2 couple to cytosolic but not group IIA phospholipase A2 in COS-1 cells

Phospholipases A2 (PLA2) and cyclooxygenases (COX) are important enzymes responsible for production of potent lipid mediators, including prostaglandins (PG) and thromboxane A2. We investigated coupling between PLA2 and COX isoforms by using transient transfection in COS-1 cells. Untransfected cells, incubated with or without phorbol ester + the Ca2+ ionophore ionomycin, generated trivial amounts of PGE2. In cells co-transfected with cytosolic PLA2 (cPLA2) and COX-1 or COX-2, phorbol ester + ionomycin markedly stimulated PGE2 production. There was no preferential coupling of cPLA2 to either of the COX isoforms. In contrast, group IIA secretory PLA2 (sPLA2) co-transfected with COX-1 or COX-2 did not lead to an increase in PGE2 production, despite high levels of sPLA2 enzymatic activity. Transfection of cPLA2 did not affect basal free arachidonic acid (AA) levels. Phorbol ester + ionomycin stimulated release of AA in cPLA2-transfected COS-1 cells, but not in untransfected cells, whereas sPLA2 transfection (without stimulation) led to high basal free AA. Thus, AA released by cPLA2 is accessible to both COX isoforms for metabolism to PG, whereas AA released by sPLA2 is not metabolized by COX.     

Novel regulatory mechanisms of CD40-induced prostanoid synthesis by IL-4 and IL-10 in human monocytes

Interleukins IL-4 and IL-10 are considered to be central regulators for the limitation and eventual termination of inflammatory responses in vivo, based on their potent anti-inflammatory effects toward LPS-stimulated monocytes/macrophages and neutrophils. However, their role in T cell-dependent inflammatory responses has not been fully elucidated. In this study, we investigated the effects of both cytokines on the production of PGE(2), a key molecule of various inflammatory conditions, in CD40-stimulated human peripheral blood monocytes. CD40 ligation of monocytes induced the synthesis of a significant amount of PGE(2) via inducible expression of the cyclooxygenase (COX)-2 gene. Both IL-10 and IL-4 significantly inhibited PGE(2) production and COX-2 expression in CD40-stimulated monocytes. Using specific inhibitors for extracellular signal-related kinase (ERK) and p38 mitogen-activated protein kinase (MAPK), we found that both kinase pathways are involved in CD40-induced COX-2 expression. CD40 ligation also resulted in the activation of NF-kappaB. Additional experiments exhibited that CD40 clearly induced the activation of the upstream kinases MAPK/ERK kinase 1/2, MAPK kinase 3/6, and I-kappaB in monocytes. IL-10 significantly inhibited CD40-induced activation of the ERK, p38 MAPK, and NF-kappaB pathways; however, inhibition by IL-4 was limited to the ERK pathway in monocytes. Neither IL-10 nor IL-4 affected the recruitment of TNFR-associated factors 2 and 3 to CD40 in monocytes. Collectively, IL-10 and IL-4 use novel regulatory mechanisms for CD40-induced prostanoid synthesis in monocytes, thus suggesting a potential role for these cytokines in regulating T cell-induced inflammatory responses, including autoimmune diseases.     

Activated human dendritic cells express inducible cyclo-oxygenase and synthesize prostaglandin E2 but not prostaglandin D2

Prostaglandins (PG) are well known lipid mediators with important immunoregulatory properties. While exogenous PGE2 has the ability to modulate the function and maturation of antigen presenting cells, such as dendritic cells (DC), it is not clear whether human DC have the capacity to synthesize PGE2 and other prostaglandins themselves. We therefore examined the expression of inducible cyclo-oxygenase (COX-2) by monocyte derived DC and the production of PGE2 and PGD2. Both monocyte derived DC and freshly isolated blood myeloid DC expressed little COX-2 constitutively, though COX-2 expression was rapidly but transiently upregulated in response to lipopolysaccharide stimulation. COX-2 mRNA was detectable within 1 h of LPS exposure, peaked at 4-6 h, and rapidly declined thereafter. COX-2 expression was accompanied by DC synthesis of PGE2, with peak levels present at 6-18 h post-stimulation. In contrast, PGD2 synthesis was not detected at any time point. When DC were activated with LPS in the presence of nimesulide, a COX-2 selective inhibitor, IL-10 synthesis was inhibited, indicating that endogenous prostaglandins regulate DC cytokine production. PGE2 production by DC may therefore modulate DC and T-cell function, thereby shaping the character of the immune response.     

Regulation of cyclooxygenase-2 expression by phosphatidate phosphohydrolase in human amnionic WISH cells.

Prostaglandins are known to play a key role in the initiation of labor in humans, but the mechanisms governing their synthesis in amnion are largely unknown. In this study, we have examined the regulatory pathways for prostaglandin E(2) (PGE(2)) production during protein kinase C-dependent activation of human WISH cells. In these cells, PGE(2) synthesis appears to be limited not by free arachidonic acid availability but by the expression levels of cyclooxygenase-2 (COX-2). Concomitant with the cells being able to synthesize and secrete PGE(2), we detected significant elevations of both COX-2 protein and mRNA levels. Specific inhibition of COX-2 by NS-398 totally ablated PGE(2) synthesis. All of these responses were found to be strikingly dependent on an active phosphatidate phosphohydrolase 1 (PAP-1). Inhibition of PAP-1 activity by three different strategies (i.e. use of bromoenol lactone, propranolol, and ethanol) resulted in inhibition of COX-2 expression and hence of PGE(2) production. These data unveil a novel signaling mechanism for the regulation of PGE(2) production via regulation of COX-2 expression and implicate phosphatidate phosphohydrolase 1 as a key regulatory component of eicosanoid metabolic pathways in the amnion.     

Prostaglandins regulate melanoma-induced cytokine production in macrophages

Tumor-secreted products can affect macrophage cytokine expression and in that way alter the immune response. Prostaglandins (PGs) are found in the tumor microenvironment and have been associated with local and regional immunosuppression. We investigated whether tumor-secreted factors could induce PG synthesis in macrophages and whether these PGs could alter macrophage production of immunoregulatory cytokines. In both murine and human models, melanoma conditioned medium (MCM) induced macrophage production of PGE(2), IL-6, and TNF-alpha. PGE(2) production increased over 24 h and was accompanied by an increase in cyclooxygenase-2 (COX-2) expression, while COX-1 expression remained unchanged. In the presence of 10 microM NS398, a selective COX-2 inhibitor, MCM-stimulated PGE(2) synthesis was almost completely suppressed, while production of IL-6 and TNF-alpha proteins and mRNA also was partially abrogated. In the murine model, 200 microM NS398 resulted in more significant inhibition of cytokine protein and mRNA production. Although MCM induced NFkappaB and NF-IL-6 activation, neither dose of NS398 altered this effect. We conclude that melanoma-secreted products stimulate COX-2 expression and PGE(2) synthesis in macrophages and that inhibition of COX-2-derived PG synthesis results in partial abrogation of macrophage cytokine production.     

Dual stimulation of phospholipase activity in human monocytes. Role of calcium-dependent and calcium-independent pathways in arachidonic acid release and eicosanoid formation

Human peripheral blood monocytes, prelabeled with [3H]arachidonic acid (AA), release labeled eicosanoids in response to soluble or particulate stimuli. Treatment with 12-O-tetradecanoate phorbol-13 acetate (20 nM), calcium ionophores, A23187 (2 microM) or ionomycin (1 microM), or serum-treated zymosan (300 micrograms) resulted in production of cyclooxygenase (CO) metabolites, 6-keto-PG-F1 alpha, thromboxane-B2, PGE2, PGF2 alpha, PGD2, PGB2, 12-L-hydroxy-5,8,10-heptadecatrienoic acid; 15-lipoxygenase products, including 15-hydroxyeicosatetraenoic acid (HETE); and unmetabolized AA. Labeled 5-lipoxygenase (LO) products, 5-HETE, and leukotriene-B4 were detected only after exposure to ionophore or serum-treated zymosan. The calcium dependence of 5-LO activation was confirmed in experiments where calcium was omitted from the incubation medium, and EGTA (0.5 mM) was added, as well as by direct measurement of increased intracellular calcium in phagocytosing monocytes. Combined or sequential treatment with two stimuli increased the release of unmetabolized AA without a commensurate augmentation of labeled metabolites, indicating that release of CO and LO metabolites does not necessarily reflect the extent of phospholipase activation. Quantitation of individual eicosanoids by RIA confirmed results by using radionuclides. These studies show the following. Activation of human monocyte phospholipase may be regulated by at least two pathways, one "12-O-tetradecanoate phorbol-13 acetate-like," which is largely independent of calcium, and another which is mediated by increased intracellular Ca2+ ("ionophore-like"). "Physiologic" stimulation of monocyte arachidonate release, such as that seen accompanying phagocytosis of opsonized particles, may occur via either a calcium-sensitive or calcium-insensitive pathway or both. Calcium may regulate eicosanoid formation at the level of phospholipase or 5-LO. Free AA, CO products, and 12- or 15-LO products are ordinarily released after phagocytosis, but leukotriene-B4, 5-HETE, or other 5-LO metabolites are produced only under conditions where calcium concentrations are optimal.     

Leukotriene D4-induced adhesion of Caco-2 cells is mediated by prostaglandin E2 and upregulation of alpha2beta1-integrin

Cell-cell and extracellular matrix adhesions play important roles in the progression of cancer. We investigated the involvement of the inflammatory mediator leukotriene D4 (LTD4) in the regulation of cell-matrix adhesion of colon cancer (Caco-2) cells. We observed that LTD4 acted via its CysLT1 receptor in these cells to induce increased adhesion to collagen I. LTD4 also enhanced the activation and expression of alpha2beta1-integrins on the cell surface, which we found to be responsible for mediating the increased adhesion to collagen I. LTD4 simultaneously augmented expression of the prostaglandin-generating enzyme cyclooxygenase-2 (COX-2) and increased prostaglandin E2 (PGE2) production in Caco-2 cells. The adhesive capacity of the Caco-2 cells was reduced by specific inhibition of COX-2 and was subsequently restored by PGE2, but not by LTD4. A selective PGE2 receptor antagonist abolished the increased adhesion and the augmented alpha2beta1-integrin expression induced by both PGE2 and LTD4. Summarizing, the inflammatory mediator LTD4 regulates the adhesive properties and migration of the Caco-2 cell line by upregulating COX-2 and stimulating PGE2-induced expression of alpha2beta1-integrins. This suggests that inflammatory mediators such as LTD4 can be involved in the dissemination and survival of colon cancer cells.