OUP user menu

Prostanoid signal transduction and gene expression in the endothelium: Role in cardiovascular diseases

Arántzazu Alfranca, Miguel A. Iñiguez, Manuel Fresno, Juan Miguel Redondo
DOI: http://dx.doi.org/10.1016/j.cardiores.2005.12.020 446-456 First published online: 1 June 2006


Endothelial cells play an active role in the maintenance of homeostasis. Endothelial injury can give rise to endothelial dysfunction in which the profile of mediators released by endothelial cells is altered. Among these mediators are factors that participate in the development of many cardiovascular disorders. Some of the most important are the prostanoids, which can modulate the progression of atherosclerosis, arterial hypertension, and angiogenesis. Prostanoids are produced by the sequential actions of cyclooxygenases and specific synthases and exert their actions through diverse cell-surface and nuclear receptors. The profile of prostanoids produced depends on cell type and the changing pathophysiological status, and these factors similarly affect the great array of biological responses elicited by these molecules. The resulting complexity enables extremely subtle and highly complex responses, and this provides opportunities for the development of targeted therapeutic approaches.

  • Angiogenesis
  • Atherosclerosis
  • Gene expression
  • Prostaglandins
  • Signal transduction

1. Introduction

The vascular endothelium is not an inert cellular barrier between circulating blood and the surrounding tissues; rather, it constitutes a dynamic interface that is central to the maintenance of homeostasis and to the onset of many pathological conditions. ECs provide a non-thrombotic vascular lumen; regulate the selective trafficking of molecules and cells between the lumen and surrounding tissues; and are essential for maintaining an adequate vascular tone. ECs moreover play key roles in the genesis and development of diverse cardiovascular disorders such as hypertension, thrombosis, atherosclerosis and other processes such as pathological angiogenesis.

These multiple actions are accomplished by a series of membrane-bound and soluble endothelial mediators, which can act in an autocrine or paracrine fashion. A variety of deleterious stimuli are able to induce phenotypic changes in ECs without evident physical cell damage, leading to an alteration of endothelial homeostasis and the occurrence of pathological events (a process known as endothelial dysfunction). For example, under physiological conditions ECs display a prominent anticoagulant activity mediated by diverse anti-thrombotic molecules, such as nitric oxide, prostacyclin, some ectonucleotidases, thrombomodulin, proteins C and S, and EPCR. However, many insults impair the secretion of some of these mediators, and induce the expression or release of a number of procoagulant factors such as TF, thromboxane, P selectin, or von Willebrand factor, thus giving rise to thrombotic events. Similarly, the development of arterial hypertension is related to impaired regulation of the secretion or activity of vasoactive substances that are normally produced by ECs to regulate basal vascular tone (e.g., nitric oxide, prostacyclin, thromboxane, endothelin-1, and endothelium-derived hyperpolarizing factor) [1]. In the early stages of atherosclerosis, ECs take part in the oxidative processing of LDLs, and are responsible for the recruitment of monocytes from peripheral blood to form fatty streaks–achieved through the injury-induced surface expression of adhesion molecules [2]. ECs also carry out the transition from the quiescent vasculature in the adult to the proliferating network of neovessels seen in angiogenesis: ECs are able to detach from adjacent cells, degrade and migrate through surrounding extracellular matrix, proliferate, and assemble to form new vessels. These events involve the expression by ECs–in response to diverse proangiogenic factors–of an array of mediators, including integrins αvβ3 and α5β1, and metalloproteases (MMP2, MMP9 and MT-MMP1) [3].

Endothelial dysfunction can be triggered by free radicals, elevated blood glucose, hypercholesterolaemia, or local factors such as hypoxia, inflammation and localised alterations to blood flow. These stimuli act either directly on ECs or through the regulation of additional molecules produced by other cells, which in turn modulate endothelial responses. Key roles in the development of and responses to endothelial dysfunction are played by prostanoids, a family of pleiotropic lipid mediators involved in many cellular processes, and which can be differentially synthesized and released in physiological and pathological conditions by diverse cell types. This review summarises current knowledge of the signal transduction pathways activated by prostanoids through binding to their receptors in endothelial cells, and discusses their potential role during the development and progression of cardiovascular disorders.

2 Prostanoid receptors and signal transduction

Prostanoids form part of a larger family of lipid mediators called eicosanoids. Eicosanoids are metabolic derivatives of arachidonic acid (AA), and the various eicosanoid sub-classes are generated by distinct metabolic pathways (detailed in Fig. 1). AA undergoes oxidation by one of three types of enzyme: lipoxygenases, to form leukotrienes and lipoxins; cyclooxygenases, to form prostanoids (i.e., prostaglandins and thromboxanes); and cytochrome P-450 monooxygenases, to form epoxides. This review will focus on signal transduction pathways activated by prostanoids, and their role on the development of a variety of cardiovascular disorders.

Fig. 1

Biosynthesis of eicosanoids. Eicosanoids are generated from arachidonic acid, which undergoes enzymatic oxidation by different metabolic pathways. The activity of lipooxygenases (LO) gives rise to leukotrienes and lipoxins. These enzymes are classified as 5-, 12- and 15-LO. Upon cell activation, 5-LO binds to 5-lipoxygenase-activating protein (FLAP), and catalyses the conversion of AA to 5-HPETE, which is modified to form leukotriene (LT)A4. This can be further metabolised either to LTB4 by LTA4 hydrolase, or, by LTC4 synthase, into so-called cysteinil leukotrienes (LTC4, LTD4 and LTE4). 12- and 15-LOs give rise to 12(S)- and 15(S)-HETE, and the latter is a precursor for Lipoxins A and B (LxA, B). COX-1, 2 catalyse the conversion of AA to the endoperoxide PGH2 by sequential reactions of cyclooxygenation and hydroperoxidation. PGH2 is then metabolised by specific synthases to produce prostanoids: thromboxane A2 (TXA2) and each of the diverse prostaglandins, the most important of which are PGE2, PGI2 (prostacyclin), PGD2, and PGF. These are usually non-enzymatically modified to produce a number of either inactive or active metabolites. Arachidonic acid can also be oxidized by cytochrome P450 monooxygenases to generate epoxides, or undergo non-enzymatic oxidation to form isoeicosanoids. TXAS, thromboxane A2 synthase; PGES, PGE2 synthase; PGIS, PGI2 synthase; PGDS, PGD2 synthase; PGER, prostaglandin endoperoxyde reductase.

Prostanoids normally act in an autocrine or a paracrine fashion, by binding to specific receptors in target cells. Most bind to and activate seven transmembrane domain, G-protein-coupled receptors: TP for TXA2; EP1-4 for PGE2; IP for PGI2; DP1 and CRTH2 for PGD2; and FP for PGF. Some also recognize nuclear receptors of the PPAR family. Transmembrane prostanoid receptors have been classically grouped into three categories, depending on the type of G-protein they bind to: the contractile receptors EP1, FP and TP are coupled to Gq and activate PLC, leading to intracellular calcium increases; the relaxant type, which comprises IP, DP1, EP2 and EP4, signal through Gs to induce adenylate cyclase activity; and the inhibitory receptor EP3 couples to Gi, leading to a decrease in intracellular cAMP content [4] (Fig. 2). This preferential coupling of each prostanoid receptor to a specific G-protein, with the subsequent activation of a definite signal transduction pathway, creates an additional regulatory level for the biological activities of these molecules in any given environment. This is particularly important for those prostanoids that can bind to multiple receptors or receptor isoforms, and are therefore able to activate diverse G-proteins and signal transduction pathways. The differential expression of these receptors in several cell types, or the modulation of their expression in a particular cell, can give rise to a wide array of complex biological responses to a single prostanoid.

Fig. 2

Prostanoid receptors. Prostanoids exert their biological functions through their binding to specific receptors. Most engage seven transmembrane domain, G-protein coupled receptors: TPα and β for TXA2; FP A and B for PGF; EP1, 2, 3 and 4 for PGE2; IP for PGI2; and DP and CRTH2 for PGD2. PGF can also bind some EP isoforms. TP, FP and EP1 receptors belong to the class of contractile receptors; EP3 is an inhibitory receptor; and EP2, EP4 and DP are relaxant type receptors. Additional actions of some prostaglandins can be mediated by nuclear receptors of the PPAR family: PGI2 activates PPAR β/δ; and both PGD2 and its metabolite 15d-PGJ2 bind to and activate PPARγ.

Signals triggered by prostanoids are mostly terminated by agonist-induced homologous desensitisation and internalisation of their receptors. Homologous desensitisation is achieved through receptor phosphorylation by Ser/Thr kinases activated by ligand binding, such as GRKs, PKC or PKA. This phosphorylation leads to the uncoupling of G protein subunits from the receptor, and in some cases the phosphorylated receptor binds to proteins of the arrestin family, and is then internalised via chlathrin-coated pits [5,6]. This process has been suggested to be a means by which receptors are dephosporylated and recycled to the membrane. In addition to homologous desensitisation, some prostanoid receptors can induce heterologous desensitisation of other receptors through the activation of second-messenger kinases such as PKA and PKC [7].

2.1 TXA2 receptors

Two isoforms of human TP receptor have been described: TPα and TPβ (Fig. 3A); and both are expressed in ECs [8]. They differ in the length of their COOH terminus, but this structural feature does not seem to differentially affect the association of TPα and β to Gq/G11. However, in platelets or ECs TPα can couple to Gs, and TPβ can couple to Gi, resulting in a differential modulation of cAMP content [9]. In platelets, TPα and TPβ also associate with G12/13 proteins, which would link TP receptors to the activation of the small GTPase Rho and the granule secretion process [10]. TXA2 analogues have been shown to upregulate ERK 1,2 and p38 MAP kinases, though the exact nature of the upstream activators involved is not clear [11]. Although TP ligand binding increases PI3K activity in some cells, in ECs TXA2 mimetics inhibit Akt phosphorylation and Akt-mediated survival signals induced by serum or VEGF [12].

Fig. 3

Signal transduction pathways activated by prostanoids. A) TP and IP-activated signal transduction pathways. TP receptors classically couple to Gq, which leads to an increase in calcium content and activation of PKC. The α isoform can also couple to Gs, and the β isoform to Gi, giving rise to opposing actions on cAMP synthesis. IP receptor mainly couples to Gs, though it can also activate Gq. Crosstalk has been proposed between the pathways activated by TP and IP at the level of ERK 1,2 activation. B) Diversity of transduction pathways activated by PGE2. PGE2 can bind and activate four different isoforms of EP receptor. EP2 and EP4 both couple to Gs and trigger common downstream events, such as PKA stimulation, with associated activation of CREB and impaired ERK 1,2 signalling. EP4 also induces the PI3K/Akt pathway, which activates ERK 1,2 and Tcf/Lef-mediated transcription. PGE2 also binds to two other receptors: EP3, which preferentially couples to Gi; and EP1, whose binding mediates an increase in intracellular calcium content.

2.2 PGE2 receptors

PGE2 receptors EP1–4 are encoded by separate genes, and are considerably heterogeneous in their sequences. EP2 and EP4 expression has been found in large vessel ECs in culture [13], and EP1, EP3 and EP4 can be detected in brain microvascular ECs in vitro and in vivo [14].

EP1 receptor binding triggers an increase in intracellular calcium and PKC activity, probably arising from its coupling to Gq and the subsequent activation of PLC. However, this idea has been seriously challenged, since the weak increase in IP3 mediated by EP1 cannot account for the strong increase in intracellular calcium concentrations observed [15] (Fig. 3B).

EP2 and EP4 are both functionally coupled to Gs, which activates adenylate cyclase to generate cAMP. Despite this similarity, these receptors display remarkable structural and functional differences in terms of ligand binding affinity and internalisation and desensitisation rates (see below). The generation of cAMP leads to the activation of PKA, which is central to the downstream events triggered by both receptors. In some cell types PKA phosphorylates and activates the transcription factor CREB [16], and PKA activity also impairs ERK 1, 2 signalling, which seems to be responsible for the decrease in MMP-1 release in synoviocytes [17]. In ECs, PKA activity is implicated in the activation of the small GTPase Rac by PGE2 [13]. PKA can also phosphorylate and inhibit GSK3α, thereby promoting transcription of Tcf-dependent genes such as COX-2. EP4 also specifically activates the PI3K/Akt pathway, which also phosphorylates and inhibits GSK3α [18]; in some cell types the activated PI3K augments ERK activity, which mediates expression of the transcription factor EGR-1 and subsequent transcription of genes such as mPGES-1 [19] (Fig. 3B).

EP3 receptors are more complex, in that several EP3 isoforms can be generated by mRNA splicing–four bovine isoforms, three mouse isoforms, and eight human variants have been reported. The diversity of EP3 receptors has interesting functional implications. In the case of bovine EP3, each isoform selectively couples to a distinct class of G-protein subunit (Gi for EP3A; Gs and G13 for EP3B; Gs for EP3C; and Gi, Gs and Gq for EP3D). No such pattern has been found for human EP3, but discrepancies in constitutive activities of the different isoforms have been reported [20]. Human EP3 mainly couples to Gi proteins, and its activation generally inhibits cAMP synthesis. A result of this is an increase in ERK activity, which has been reported in some cell types [21]. EP3 is also implicated in the activation of Src kinase, the transcription factor Stat3, and the upregulation of its target genes BclXL and cyclin D [22]. Similarly, as a consequence of its coupling to G13, EP3 activates the Rho pathway in PC12 and MDCK cells [23] (Fig. 3B).

PGE2 can regulate the JNK and p38 pathways in a number of cell types, such as synovial fibroblasts (where it induces COX-2 mRNA stabilization), and ECs [24,25]. It has also been shown to activate NFκB [26]; and is able to augment HIF-1α expression in a prostate cancer cell line [27].

Transactivation of tyrosine kinase receptors by PGE2 provides a further level of complexity to the signalling events triggered by this prostanoid. PGE2 transactivates the EGF receptor via activation of c-Src, which either activates EGFR directly [28] or indirectly, by inducing a matrix metalloproteinase activity that releases membrane-bound TGFα [29]. Recently a distinct transactivation mechanism has been identified, in which PGE2 modulates PPARδ nuclear receptor activity through PI3K/Akt pathway [30].

The possible coexpression of several EP variants raises the question of how a particular cell can process the mixed and conflicting set of signals generated by simultaneous binding of different EP receptors. This interesting issue has not been extensively investigated, but the final biological outcome will probably depend on many factors. Among these, ligand affinity for the receptor is likely to be important: in the case of PGE2 this is tenfold higher for EP3 and EP4 than for EP1; and fivefold higher for EP4 than for EP2 [20]. Other potential factors are crosstalk among the different signal transduction pathways triggered by the stimulus, and the internalisation and desensitisation rate of each variant. EP4 and EP2 are a clear example of this last factor: EP4 is more rapidly internalised and desensitised, resulting in comparatively lower cAMP production [31].

2.3 PGI2 receptor

PGI2 inhibits platelet aggregation and induces vasodilatation. Its receptor, IP, is found mainly in megakariocytes and smooth muscle cells, but is also expressed in ECs [32]. IP is preferentially coupled to Gs subunits; and PGI2 binding activates adenylate cyclase to increase cAMP levels. The subsequent activation of PKA could impair ERK signalling, and thus interfere with signalling by TXA2 [33] (Fig. 3A). In ECs, the activation of PKA by PGI2 is also involved in Rac activation [13]. IP can also mediate increases in intracellular calcium content via its association to Gq.

In addition to binding IP, PGI2 is a natural ligand of the nuclear receptor PPAR β/δ, and also binds PPAR γ with similar affinity [34].

2.4 PGD2 and 15d-PGJ2 receptors

PGD2 can bind to two unrelated receptors: DP and CRTH2, also known as DP2. Activation of DP leads to cAMP accumulation, thought to be mediated by Gs; however, some reports indicate that DP binding is also responsible for PLC activation and calcium mobilization, which imply functional coupling of DP to Gq [35].

PGD2 and its metabolite 15d-PGJ2 are both able to bind and activate nuclear receptor PPAR γ. However, the in vivo relevance of this activation has been questioned because of the high doses of 15d-PGJ2 needed to achieve it. The PPAR family of nuclear receptors comprises α, β/δ and γ isoforms, and all are found in ECs [36]. Upon ligand binding, PPARs heterodimerize with RXR, and these complexes recognize specific response elements (PPRE) in the regulatory regions of target genes. The transcriptional activity of these receptors is modulated by the recruitment of coactivators and corepressors, and also through their phosphorylation by several kinases [37]. PPAR γ also affects transcription through mechanisms independent of DNA binding: it inhibits the activities of transcription factors such as AP-1, STATs, NFAT and NFκB via a redox mechanism [38,39]. Finally, 15d-PGJ2 has effects not mediated by PPARγ, such as the activation of H-Ras, the enhancement of PKCζ activation by LPS in macrophages, and the blocking of LPS-induced Cot kinase activation [40–42].

2.5 PGF receptors

PGF binds to its specific receptor FP, and also with significant affinity to EP1 and EP3. FPA and FPB isoforms of the PGF receptor are generated by alternate mRNA splicing, and differ in their carboxyl-terminal regions. Both share common functional features, such as their coupling to Gq proteins, which activates PLC and PKC and triggers the Raf/MEK/ERK cascade [43]. The increased calcium influx provoked by PLC is implicated in the positive regulation of skeletal muscle growth through activation of the NFATc2 transcription factor [44]. Similarly, both FP receptors can couple with the Gi subunit, which could induce Ras activation [45]. Changes to the actin cytoskeleton, mediated by the small GTPase Rho pathway, have also been described [46]. A functional difference between the receptors is that FPB, but not FPA, can persistently induce nuclear translocation of β-catenin, with a prolonged increase in Tcf/Lef-mediated transcription [47].

3 Prostanoid signalling in ECs and cardiovascular disorders

Endothelial dysfunction involves the phenotypic modification of ECs, triggered by different kinds of injury. These injuries modulate the expression and/or activities of mediators (released by ECs or other cell types), which can act in an autocrine or a paracrine manner (Table 1). Prostanoids are among the most important of these mediators, and are major regulators of EC function. These molecules play key roles in the response to vascular injury and in the development and pathophysiology of the cardiovascular system, as confirmed by the phenotypes of mice deficient in prostanoid receptor genes (Table 2).

View this table:
Table 2

Cardiovascular phenotypes of mice deficient in prostanoid receptors

Prostanoid receptor knock outCardiovascular phenotypeRefs.
TP–Increased bleeding tendency[96]
–Reduced incidence of thromboembolic events[96]
EP2–Salt-sensitive hypertension[85]
–Impaired vasodepressor response to i.v. infusion of PGE2[85]
–Reduced angiogenic response in the polyps of Apc δ716 mice[87]
EP3–Enhanced vasodepressor response to i.v. infusion of PGE2[85]
–Impaired tumour-associated angiogenesis[88]
EP4–Patent ductus arteriosus.[97]
IP–Increased frequency of thromboembolic events, with no spontaneous arterial hypertension.[78]
PPARγ–Placental and cardiac malformations[98]
  • The table summarizes the cardiovascular defects described for mice deficient in prostanoid receptor genes.

View this table:
Table 1

Regulation of the expression of endothelial mediators by prostanoids

ProstanoidEffectMolecule/processCardiovascular disorderRefs.
PGI2 analoguesVCAM-1ATC[62]
PPARβ/δ agonistVCAM-1ATC[63]
PPARβ/δ agonistMCP-1ATC[63]
15d-PGJ2MIG, IP-10, ITACATC[65]
15d-PGJ2+PAI-1ATC, ANG[69]
15d-PGJ2− /+NOHT, ANG[80,81]
TPβ agonistNOHT, ANG[12]
15d-PGJ2ET-1HT, ATC[83]
15d-PGJ2TFATC, ANG[67]
15d-PGJ2KDR, Flt-1ANG[91]
15d-PGJ2+uPAATC, ANG[91]
TXA2 analogue+p53, TSP-1ATC, ANG[8]
15d-PGJ2+cIAP-1ATC, ANG[75]
15d-PGJ2+/ −ApoptosisATC, ANG[74,75]
TPβagonist+/ −ApoptosisATC, ANG[9]
PGE2+Spreading, migrationANG, ATC[13]
PGI2+Spreading, migrationANG, ATC[13]
  • The table summarizes the modulatory effects of different prostanoids on endothelial molecules or biological processes involved in the development of cardiovascular disorders. +, increased expression of the molecule or activation of the process; −, decreased expression of the molecule or inhibition of the process; +/ −, disputed effect. ATC, atherosclerosis; ANG, angiogenesis; HT, arterial hypertension.

3.1 Prostanoid signalling in ECs and atherosclerosis

Atherosclerosis is currently viewed as a chronic inflammatory disease [48]. Early stages of this disorder involve the formation of fatty streaks beneath the arterial endothelium. The generation of these lesions starts with the oxidation of plasma low density lipoproteins (oxLDLs), and their accumulation in the sub-endothelial space. COX-2 is considered an important source of oxygen radicals, and it is expressed in all cell types present in atherosclerotic lesions, where it may be induced by inflammatory mediators, growth factors, or CD40 ligation [49,50]. However, its expression is reduced in foam cells, probably due to the inhibitory effect of oxLDL on COX-2 synthesis in macrophages (but not ECs) [51]. The expression of COX-2 in atherosclerotic lesions is consistent with its known role in the development of atherosclerosis: non-selective and selective COX-2 inhibitors significantly reduce the formation of early atherosclerotic lesions in LDLR − / − and ApoE − / − mice [52,53]; and the formation of atherosclerotic plaques is inhibited by transplanting fetal liver cells from COX-2 − / − mice into LDLR − / − mice [53]. However, probably due to impaired fertility in COX − 2 − / − mice, efforts to obtain information from mice lacking both COX-2 and ApoE or LDLR genes have been unsuccessful.

COX-2 products thus have a significant influence on the atherosclerotic process, through their autocrine and paracrine effects on the different cell types present in the developing plaque. Formation of the initial lesion requires monocyte recruitment to the oxLDL deposits. Recruited monocytes migrate across the endothelial monolayer, and accumulate in the intima, where they are converted into foam cells, a hallmark of atherosclerosis. Monocyte recruitment is accomplished by means of endothelial expression of adhesion molecules and chemokines: the central role of endothelial E-selectin, P-selectin, VCAM-1, ICAM-1 and MCP-1 has been convincingly demonstrated by studies in knockout and transgenic mice [54–57]. The best characterized prostanoid modulators of this process are prostaglandins of the J2 series, which inhibit the cytokine-induced expression of several of these adhesion molecules. For example, 15d-PGJ2 inhibits E-selectin expression via the PPAR-γ-dependent transcriptional induction of ATF3, which is thought to displace ATF2/Jun dimers from their binding sites in the E-selectin promoter [58]. Furthermore, 15d-PGJ2 has a dual action on ICAM-1 in ECs: it directly induces ICAM-1 expression through an increase of AP-1 binding to its promoter; but it simultaneously blocks TNF-α-induced ICAM-1 expression, via the inhibition of NF-κB DNA binding. Both these actions are PPAR-γ-dependent [59]. PPAR-γ also mediates 15d-PGJ2 activation of diacylglycerol kinase (DGK) in ECs, which leads to inhibition of PKCβ. This action might underlie the observed decrease of E-selectin, ICAM-1 and VCAM-1 [60]. 15d-PGJ2 also acts on ECs via mechanisms independent of PPAR-γ. It impairs the TNF-α-induced expression of ICAM-1 and VCAM-1 through the activation of RORα1 [61]. TNF-α-induced VCAM-1 expression is also blocked by activation of the transmembrane IP receptor [62], and the activation of the PPAR-δ nuclear receptor inhibits cytokine-induced VCAM-1 and MCP-1 expression through the inhibition of NFκB activity [63].

In the progression of atherosclerosis to more advanced lesions, VSMCs migrate to the subendothelial space, where they proliferate and secrete diverse extracellular matrix proteins. This is partly mediated by endothelial production of PDGF, and 15d-PGJ2 has been reported to inhibit PDGF synthesis in ECs by decreasing the expression of Sp1 transcription factor [64]. Advanced lesions are also characterized by infiltration of activated T lymphocytes, which release an array of cytokines and leukocyte chemoattractants, including IFNγ, that amplify the inflammatory response. Thus at advanced stages of plaque formation virtually all the cell types present (ECs, VSMCs, and macrophages) become activated. This is evident from these cells– expression of MHC II and their production of cytokines, which perpetuate the inflammatory situation within the plaque. 15d-PGJ2 has been shown to impair IFNγ-induced expression of the chemokines MIG, IP-10, and I-TAC [65], thus potentially reducing the extent of lymphocyte infiltration in advanced atherosclerotic lesions. Another important effect of 15d-PGJ2 is its ability to block the IFN γ-induced expression of MHC II in ECs, through the inhibition of the transcription factor CIITA [66].

The presence of pro-coagulant factors in the mature plaque is the basis of the thrombotic events that take place in the later stages of the disease, and prostanoids are implicated in the regulation of these factors. Some studies indicate that 15d-PGJ2 impairs TNF-α-induced TF expression in ECs by inhibiting ERK 1, 2 and NF-κB activities [67]; but others describe the opposite effect [68]. 15d-PGJ2 has also been reported to induce the expression of PAI-1 [69]; however, no definite implication of this molecule in the pathogenesis of atherosclerosis has been established [70].

Some studies have reported increased expression of COX-2 and PGES in the macrophages present in the shoulders of symptomatic plaques [71]. Autocrine action of PGE2 synthesized by these macrophages would induce membrane metalloprotease activity, thereby contributing to plaque instability [72]. In addition, angiogenesis may participate in plaque growth and reduce plaque stability, as suggested by the effectiveness of some antiangiogenic compounds at inhibiting plaque progression in ApoE − / − mice [73]; PGE2 could trigger neovascularization in the lesion by acting directly on ECs (see below).

The formation of necrotic areas in the instable plaque involves apoptosis; and 15d-PGJ2 has been suggested to induce apoptosis in ECs in a PPAR-γ-dependent manner [74]. By contrast, 15d-PGJ2 also induces cIAP-1, which has been proposed to prevent endothelial apoptosis induced by bFGF withdrawal [75]. TXA2 acting through TPβ receptor, also induces apoptosis in ECs, by inhibiting serum- and VEGF-induced Akt phosphorylation [9].

COX-2, which is expressed in the atherosclerotic lesion throughout the progression of the disease, is potentially a source for a wide array of prostanoids with diverse biological activities. The profile of prostanoids synthesized at each stage will vary depending on the functional synthases present in each cell type in a particular environment, since many of them are inducible enzymes. The activity displayed by any given prostanoid will also be variable, as it will depend on the receptors expressed by different subsets of cells and also on the local concentration of the compound. As a result, a particular prostanoid receptor can be activated, with different affinities, by prostanoids other than its main ligand; and this can give rise to a variety of responses [76]; in addition, variations in the concentration of a given prostanoid can trigger different signals by binding to diverse isoforms of its specific receptor, or by inducing coupling to different G protein subunits [33]. These observations may go some way to explain the apparent discrepancy between the role of COX-2 in promoting the development of atherosclerosis, and the overall beneficial actions of its products in limiting endothelial dysfunction.

3.2 Prostanoid signalling in ECs and arterial hypertension

Several endothelial mediators participate in the maintenance of vascular tone; and among the most prominent are the vasodilators NO and PGI2, and the vasoconstrictor ET-1.

In contrast to PGI2, endothelial-derived NO is involved in the vasodilator response under homeostatic conditions; mice deficient in eNOS exhibit basal arterial hypertension [77], whereas IP-deficient mice do not show a spontaneous rise in arterial tension [78]. In addition, decreased NO activity and an associated impairment of endothelium-dependent vasodilatation have been reported in patients with essential hypertension, and also in their normotensive offspring [79]. Some prostanoids are able to regulate NO production by ECs, potentially modulating endothelium-dependent vasodilator responses. 15d-PGJ2 has been reported to stimulate NO release in ECs by increasing hsp90-eNOS interaction and eNOS Ser1177 phosphorylation [80]. In contrast, prolonged treatments with 15d-PGJ2 have been reported to decrease eNOS mRNA and protein expression, possibly indicating an indirect effect on eNOS expression [81]. TXA2, acting via the TPβ receptor, similarly decreases endothelial NO production in vitro, in this case by inhibiting Akt phosphorylation and impairing eNOS activation [12].

There is no a clear relationship between circulating levels of ET-1 and the development of arterial hypertension in humans, and it has been proposed that an altered sensitivity to this factor, rather than altered secretion, could be the basis of the disease [82]. However, some PPAR γ ligands have been reported to impair thrombin-induced ET-1 production in ECs via the inhibition of AP-1 binding to its promoter [83]; and some authors have also described inhibition of ET-1 release by PGE2 in ECs [84].

Intravenous infusion of PGE2 induces hypotension in mice, and this response is impaired mice lacking receptor EP2. This phenotype has been attributed to the predominance of signalling via EP1 and EP3 receptors in these animals. When fed a high-salt diet, EP2 − / − mice also show increased urinary excretion of PGE2, with a concomitant arterial hypertension, probably also due to the predominance of signalling via EP1 and EP3 [85]. The salt-sensitive hypertension developed by these mice suggests that this response may not depend only on vascular mechanisms, but also on altered renal excretion of sodium; and may reflect the impaired tubular response in EP2 knockout mice (Table 1).

3.3 Prostanoid signalling in ECs and angiogenesis

The development of supporting blood vessels is necessary for the onset and progression of many physiological and pathological processes, including cancer and chronic inflammatory diseases such as atherosclerosis. The formation of new vessels can be accomplished either by angiogenesis, which requires the existence of a previous vascular network, or by vasculogenesis, which involves the de novo recruitment of endothelial precursors. In the adult, the main mechanism through which new vessels are formed is angiogenesis. Angiogenesis requires the acquisition of an angiogenic phenotype by quiescent ECs, and is exquisitely regulated by the fine balance between pro- and antiangiogenic factors [3].

COX-2 plays a prominent role in carcinogenesis and angiogenesis. Its expression can be detected in neovessels, but not in quiescent vasculature, and is regulated in ECs by proangiogenic factors such as VEGF and bFGF. Moreover, tumoral growth and vascular density are both impaired in COX-2-deficient mice [86].

The relative contribution of the different prostanoids to angiogenesis is not fully defined, but PGE2 seems to be an important mediator of COX-2-regulated angiogenesis. This is consistent with the similar phenotypes of EP2 and COX-2 deficient mice in an Apcδ716 background: the number and size of intestinal polyps are reduced; and the vascularization and content of proangiogenic factors in these lesions are decreased [87]. Other authors have reported a striking reduction in tumour-associated angiogenesis in EP3 − / − mice, suggesting a tissue-dependent specificity for PGE2 activity [88]. PGE2 induces a proangiogenic response in several in vitro and in vivo models. PGE2 also induces the secretion of VEGF in many cell types–including ECs–through the activation of ERK2 [25]. The angiogenic response of microvascular ECs to PGE2 has also been suggested to be partly mediated by the PGE2-induced increase in the expression of the chemokine receptor CXCR4 [89]. An autocrine mechanism has been described for the PGE2-mediated regulation of EC spreading, in which the PGE2 released by ECs would act on these cells via EP2 and EP4 receptors, leading to cAMP synthesis and PKA activation. This in turn would activate Rac, and induce EC spreading, thereby promoting EC migration. A similar mechanism has been proposed for PGI2 [13].

Pro- and anti-angiogenic actions have been reported for 15d-PGJ2. The pro-angiogenic effect arises from its PPAR γ-mediated upregulation of VEGF expression in ECs [90]. However, there are several lines of evidence showing anti-angiogenic actions. 15d-PGJ2 inhibits mitogen-induced EC proliferation and tube formation by ECs in an in vitro model of angiogenesis. These inhibitory effects are probably mediated by the downregulation of VEGF receptors and uPA [91]. 15d-PGJ2 also inhibits VEGF-induced angiogenesis in a corneal assay, and PPAR γ ligands block choroidal neovascularization by inhibiting EC proliferation and migration [92]. Another possible anti-angiogenic action of 15d-PGJ2 is its PPAR-γ-mediated upregulation of CD36 expression. CD36 is a ligand for TSP-1, and mediates microvascular EC apoptosis [93]. 15d-PGJ2 thus may promote TSP-1-induced endothelial apoptosis, countering proliferation and angiogenesis [94]. Reports on the potential role of 15d-PGJ2 in EC survival during the development of atherosclerotic plaques have been discussed above. Further studies are needed to determine whether these conflicting data can be explained by differences in experimental conditions, or reflect a genuine dual role of 15d-PGJ2 in angiogenesis.

The evidence for the function of TxA2 in angiogenesis is also contradictory. TxA2 can regulate microvascular EC migration in vitro, and also mediates bFGF-induced angiogenesis in a corneal pocket assay [95]. However, several reports identify a negative regulatory role for TxA2 in angiogenesis. The activation of TPβ leads to an inhibition of tube formation in an in vitro assay, probably due to the triggering of endothelial apoptosis mediated by the inhibition of Akt phosphorylation [9]. This TP isoform has also been implicated in the inhibition of bFGF-induced vascularization of a matrigel plug in vivo, and in the impairment of EC migration in response to the same factor in vitro [8]. Finally, TPβ engagement also impairs the migration of ECs in response to VEGF, by inhibiting Src and the activation of focal adhesion kinase by KDR; this has the effect of reducing the number and size of focal adhesions [12].

As with atherosclerosis, a complex array of cells and stimuli participate in the regulation of angiogenesis. Although COX-2 seems mainly to promote angiogenesis, prostanoid action can regulate neovessel formation positively or negatively; the outcome at any given moment will depend on the balance of prostanoids and receptors present in any particular scenario.

4 Conclusion

The actions of prostanoids on ECs in pathological situations are varied and complex. Some appear to be beneficial, countering the worst effects of disease, whereas others contribute to its further development. This complexity derives from two factors: 1) the potentially immense variation in the profile of prostanoids released in any particular context; and 2) the similar variability in the responsiveness of ECs, based on the range of prostanoid receptors expressed and their coupling to different effector pathways. The outcome of COX-2 activity and subsequent prostanoid signalling will therefore be an exquisite and highly dynamic balance among these different and often conflicting signals. Disturbance of this delicate balance may determine whether prostanoid signals have beneficial or deleterious effects during the development of important cardiovascular pathologies; and this subtlety and complexity also provides opportunities for the development of novel therapeutic strategies to target specific prostanoid actions. But the achievement of this goal first requires a comprehensive understanding of prostanoid actions and functions during these disease processes.


The authors want to acknowledge Dr. S. Bartlett for his valuable advice and helpful discussion on the manuscript. Part of the work described in this review was supported by grant LSHM-CT-2004-005033 (EICOSANOX).


  • Time for primary review 16 days

  • Abbreviations:
    activator protein-1
    activating transcription factor
    basic fibroblast growth factor
    cyclic AMP
    cellular inhibitor of apoptosis protein-1
    cAMP response element binding factor
    chemoattractant receptor-homologous molecule expressed on T Helper type 2 cells
    receptor for CXC chemokine 4
    diacylglycerol kinase
    receptor for PGD2
    endothelial cell
    epidermal growth factor
    EGF receptor
    early growth response factor-1
    endothelial nitric oxide synthase
    receptor for PGE2
    endothelial protein C receptor
    extracellular regulated kinase
    receptor for PGF
    G-protein coupled receptor kinase
    glycogen synthase kinase-3 alpha
    hypoxia inducible factor-1 alpha
    heat shock protein 90
    intercellular adhesion molecule-1
    interferon gamma
    receptor for PGI2
    Interferon-Inducible Protein 10
    interferon inducible T-cell alpha chemoattractant
    Jun NH2-terminal kinase
    kinase insert domain receptor
    low density lipoprotein
    oxidized LDL
    LDL receptor
    monocyte chemoattractant protein-1
    MHC II
    major histocompatibility complex class II
    Monokine induced by interferon-gamma
    matrix metalloprotease
    membrane type MMP
    nuclear factor of activated T cells
    nuclear factor kappa B
    nitric oxide
    plasminogen activator inhibitor-1
    platelet derived growth factor
    PGE2 synthase
    microsomal PGES
    phosphatydil inositol 3 kinase
    protein kinase A
    protein kinase C
    phospholipase C
    peroxisome proliferator-activated receptor
    PPAR response element
    retinoic acid receptor-related orphan receptor α1
    retinoid X receptor
    tissue factor
    transforming growth factor alpha
    tumour necrosis factor alpha
    receptor for thromboxane A2
    urokinase type plasminogen activator
    vascular cell adhesion molecule
    vascular endothelial growth factor
    vascular smooth muscle cell


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
  51. [51]
  52. [52]
  53. [53]
  54. [54]
  55. [55]
  56. [56]
  57. [57]
  58. [58]
  59. [59]
  60. [60]
  61. [61]
  62. [62]
  63. [63]
  64. [64]
  65. [65]
  66. [66]
  67. [67]
  68. [68]
  69. [69]
  70. [70]
  71. [71]
  72. [72]
  73. [73]
  74. [74]
  75. [75]
  76. [76]
  77. [77]
  78. [78]
  79. [79]
  80. [80]
  81. [81]
  82. [82]
  83. [83]
  84. [84]
  85. [85]
  86. [86]
  87. [87]
  88. [88]
  89. [89]
  90. [90]
  91. [91]
  92. [92]
  93. [93]
  94. [94]
  95. [95]
  96. [96]
  97. [97]
  98. [98]
View Abstract