Constrictor prostanoids and uridine adenosine tetraphosphate: vascular mediators and therapeutic targets in hypertension and diabetes

Abstract

Vascular dysfunction plays a pivotal role in the development of systemic complications associated with arterial hypertension and diabetes. The endothelium, or more specifically, various factors derived from endothelial cells tightly regulate vascular function, including vascular tone. In physiological conditions, there is a balance between endothelium-derived factors, that is, relaxing factors (endothelium-derived relaxing factors; EDRFs) and contracting factors (endothelium-derived contracting factors; EDCFs), which mediate vascular homeostasis. However, in disease states, such as diabetes and arterial hypertension, there is an imbalance between EDRF and EDCF, with a reduction of EDRF signalling and an increase of EDCF signalling. Among EDCFs, COX-derived vasoconstrictor prostanoids play an important role in the development of vascular dysfunction associated with hypertension and diabetes. Moreover, uridine adenosine tetraphosphate (Up₄A), identified as an EDCF in 2005, also modulates vascular function. However, the role of Up₄A in hypertension- and diabetes-associated vascular dysfunction is unclear. In the present review, we focused on experimental and clinical evidence that implicate these two EDCFs (vasoconstrictor prostanoids and Up₄A) in vascular dysfunction associated with hypertension and diabetes.

Table of Links

TARGETS
Enzymesa	GPCRsc
Akt	AT1 receptor
AMPK, AMP-activated protein kinase	EP receptor
CDK2, cyclin-dependent kinase 2	DP receptor
COX-1	GPER (GPR30)
COX-2	IP receptor
cPLA2, cytosolic PLA2	P2Y2 receptor
ERK1/2	P2Y4 receptor
HO-1, haem oxygenase-1	P2Y6 receptor
L-PGDS, lipocalin-type PGD synthase	TP receptor
NOS	Catalytic receptorsd
PGIS, prostacyclin synthase (CYP8A1)	PDGFR, PDGF receptor
TxS, thromboxane synthase (CYP5A1)	Nuclear hormone receptorse
ROCK, Rho kinase	Erα oestrogen receptor (NR3A1)
Ligand-gated ion channelsb	Erβ oestrogen receptor (NR3A2)
P2X1 receptor

LIGANDS
20-HETE	LY294002
AA, arachidonic acid	NO
ADP	Metformin
Angiotensin II	PD98059
Diclofenac	PGD2
Clonidine	PGE2
DOCA, deoxycorticosterone acetate	PGF2α
E2, 17β-oestradiol	PGI2
EPA, eicosapentaenoic acid	Rapamycin
ET-1, endothelin-1	Suramin
L-arginine	Testosterone
Ip5I	TxA2, thromboxane A2
L-NAME	U46619
Losartan	VP, vasopressin

These Tables list key protein targets and ligands in this article which are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Pawson et al., 2014) and are permanently archived in the Concise Guide to PHARMACOLOGY 2013/14 (^a,b,c,d,eAlexander et al., 2013a,b,c,d,e,,,,).

The endothelium plays a pivotal role in the regulation of vascular tone (Vapaatalo and Mervaala, 2001; Pries and Kuebler, 2006; Flammer and Luscher, 2010; Toda et al., 2010; Flammer et al., 2012; Favero et al., 2014). In response to mechanical forces (e.g. shear stress) and endogenous ligands, endothelial cells (ECs) release a diversity of factors that mediate or directly induce vascular smooth muscle contraction or relaxation (Vapaatalo and Mervaala, 2001; Flammer and Luscher, 2010; Flammer et al., 2012). Accordingly, these factors are largely divided into relaxing and contracting factors [namely, endothelium-derived relaxing factors (EDRFs) and endothelium-derived contracting factors (EDCFs) respectively]. Nitric oxide (NO), endothelium-derived hyperpolarizing factor (EDHF) and prostacyclin (PGI₂) are well-established EDRFs (Furchgott and Zawadzki, 1980; Chen et al., 1988; Feletou and Vanhoutte, 2004; Feletou, 2009; Garland et al., 2011; Flammer et al., 2012). On the other hand, vasoconstrictor prostanoids, endothelin-1 (ET-1), angiotensin II and more recently uridine adenosine tetraphosphate (Up₄A) have been described as major EDCFs (Yanagisawa et al., 1988; Jankowski et al., 2005; Vanhoutte et al., 2005; 2009,; Vanhoutte and Tang, 2008; Matsumoto et al., 2014a).

Endothelial dysfunction is a hallmark of the pathophysiology of diabetes and arterial hypertension and contributes to macrovascular and microvascular complications associated with these disease states (De Vriese et al., 2000; Kobayashi et al., 2000; Vanhoutte et al., 2005; 2009,; Versari et al., 2009; Ding and Triggle, 2010; Barton et al., 2012; Bazi et al., 2012; Campia et al., 2012). An imbalance between EDRFs and EDCFs plays a key role in the development of vascular dysfunction in hypertension and diabetes (Vanhoutte et al., 2005; 2009,; Michel et al., 2008a; Vanhoutte and Tang, 2008; Ding and Triggle, 2010; Tang and Vanhoutte, 2010; Mukohda et al., 2012). EDRF production, bioavailability and signalling in smooth muscle cells (SMCs) are decreased in vessels under diabetic and hypertensive conditions (Kamata et al., 1989; De Vriese et al., 2000; Vanhoutte et al., 2005; 2009,; Vanhoutte and Tang, 2008; Ding and Triggle, 2010). On the other hand, EDCF production and/or signalling are increased in vessels in the presence of those diseases (Vanhoutte et al., 2005; 2009,; Vanhoutte and Tang, 2008; Feletou et al., 2009; 2010a; 2011,,; Tang and Vanhoutte, 2010; Matsumoto et al., 2014a). Therefore, the fine regulation of this balance may be important in order to prevent the development of diabetes- and hypertension-associated vasculopathies.

EDRFs have attracted great attention and several excellent reviews have been written on EDRF-mediated vascular signalling in metabolic and cardiovascular diseases including diabetes and hypertension (De Vriese et al., 2000; Griffith, 2004; Triggle et al., 2005; Feletou et al., 2008; 2009,; Matsumoto et al., 2008a; Grgic et al., 2009; Barton, 2010; Ding and Triggle, 2010; Edwards et al., 2010; Forstermann and Li, 2011; Campia et al., 2012; Tousoulis et al., 2012; Sena et al., 2013; Toda et al., 2013; Bruder-Nascimento et al., 2014; Manrique et al., 2014; Goulopoulou and Davidge, 2015). The role of EDCF in vascular dysfunction associated with these disease states, however, is not well understood. Constrictor prostanoids are EDCFs with potent and complex effects on vascular smooth muscle function and have been implicated in both diabetes and hypertension (Vanhoutte et al., 2005; Feletou et al., 2009; 2011,; Versari et al., 2009; Barton et al., 2012). Up₄A, identified as an EDCF in 2005 (Jankowski et al., 2005), also modulates vascular function (Matsumoto et al., 2011a); however, the pathophysiological role of Up₄A is unclear.

In this review, we focus on the contribution of constrictor prostanoids and Up₄A to vascular dysfunction in diabetes and arterial hypertension. We address the mechanisms associated with the production and signalling of prostanoids and Up₄A in vascular SMCs and the interactions between EDCFs with EDRFs. Further, we discuss experimental evidence that implicates prostanoids and Up₄A in diabetes- and hypertension-associated vascular pathology. Potential therapeutic strategies to target the overproduction and actions of EDCFs and to decrease EDCF-mediated responses are also reviewed. The influence of sex and sex steroid hormones on vascular function have been previously described (Tostes et al., 2003; 2008,; Orshal and Khalil, 2004; Khalil, 2005; Arnal et al., 2010; Huxley and Wang, 2010; Hart et al., 2011; Nilsson et al., 2011; Bubb et al., 2012; Miao and Li, 2012; Kittikulsuth et al., 2013) and considerable progress has been made in our understanding of the cellular and molecular mechanisms underlying the effects of sex hormones and their receptors in the vascular system (Tostes et al., 2003; 2008,; Orshal and Khalil, 2004; Khalil, 2005; Arnal et al., 2010; Huxley and Wang, 2010; Hart et al., 2011; Nilsson et al., 2011; Bubb et al., 2012; Miao and Li, 2012; Kittikulsuth et al., 2013). In the present review, therefore, the effect of sex on EDCF vascular signalling in the context of diabetes and arterial hypertension is also discussed.

COX-mediated production of vasoconstrictor prostanoids

The production and signalling of vasoconstrictor prostanoids are increased in arterial SMCs of subjects with hypertension and diabetes (De Vriese et al., 2000; Vanhoutte et al., 2005; 2009,; Vanhoutte and Tang, 2008; Feletou et al., 2009; 2010a,b,; 2011; Versari et al., 2009; Ding and Triggle, 2010; Tang and Vanhoutte, 2010; Wong and Vanhoutte, 2010; Barton et al., 2012; Figure 1). Increased constrictor prostanoid production and activity are attributable to the enhanced expression and/or activity of COXs (COX-1, COX-2) and to increased production of reactive oxygen species (ROS). ROS are considered EDCFs themselves as well as enhancers/modulators of EDCF responses (Tang et al., 2007; Tang and Vanhoutte, 2010) in hypertensive arteries (Vanhoutte et al., 2005; 2009; Vanhoutte and Tang, 2008; Feletou et al., 2009; 2010a; 2011,,; Tang and Vanhoutte, 2010).

Figure 1.

Major EDCF-mediated signalling pathways in hypertension and diabetes. When intracellular Ca²⁺ levels are increased via the activation of GPCR by, for example, ACh, ADP or Ca²⁺ ionophore (A23187), AA is produced from membrane phospholipids by PLA₂. AA is metabolized to endoperoxides [PGG₂ and PGH₂] through activation of COXs (COX-1 and COX-2). PGH₂ is further metabolized into PGD₂, PGE₂, PGF₂α, PGI₂ and TxA₂ by specific synthases including PGDS, PGES, PGFS, PGIS and TxS. All these prostanoids and endoperoxides can activate TP receptors in vascular SMC leading to contraction via multiple pathways. Although PGI₂ usually leads to vasodilatation through activation of the IP receptor/adenylate cyclase (AC)/cAMP pathway, PGI₂ responses are impaired in some conditions, including hypertension and diabetes. Instead of abnormalities in the IP receptor/AC/cAMP pathway, other prostanoids such as PGE₂, PGF₂α and TxA₂, and endoperoxides may be increased and then stimulate TP receptors. In hypertensive arteries, mainly PGI₂, TxA₂ and PGH₂ mediate EDCF responses (in red). In addition to these prostanoids, PGE₂ and PGF₂α can also behave as EDCF in arteries from diabetic subjects (in blue). PGD₂ another EDCF has less significant effects compared with other PGs. COX generates ROS. ROS are considered EDCFs themselves as well as enhancers/modulators of EDCF responses, since ROS (1) further activate endothelial and smooth muscle COX and (2) increase production of isoprostanes via AA, leading to TP receptor activation. COX activity is regulated and further enhanced by several amplifier/inducers in hypertension and diabetes. Details are described in the text.

Stimulation of ECs (e.g. by ACh and ADP) leads to an increase in intracellular concentration of calcium (Ca²⁺) and subsequent release of arachidonic acid (AA), which is then metabolized by COXs (Vanhoutte et al., 2005; 2009,; Vanhoutte and Tang, 2008; Feletou et al., 2009; 2010a,b,; 2011; Tang and Vanhoutte, 2009; 2010,). PLA₂ significantly contributes to this process via Ca²⁺-dependent and -independent mechanisms (Tang and Vanhoutte, 2009; Feletou et al., 2010a,b,). Elevated cytosolic concentration of Ca²⁺ is a trigger for endothelium-dependent contractions, probably by activating Ca²⁺-dependent PLA₂ in ECs (Tang and Vanhoutte, 2009; 2010,; Vanhoutte, 2009; Vanhoutte et al., 2009; Feletou et al., 2010a,b,). Indeed, inhibition of Ca²⁺-dependent PLA₂ suppresses EDCF-mediated contractions (Tang et al., 2007; Ishida et al., 2011). Endothelial ligands such as ACh, ET-1 and nucleotides lead to increased intracellular Ca²⁺ and then release of COX-derived EDCFs (Feletou et al., 2010a,b,; Tang and Vanhoutte, 2010). The calcium ionophore A23187 directly increases intracellular Ca²⁺ in ECs, and therefore, A23187 may act as an inducer of EDCF production (Shi et al., 2007; Tang et al., 2007). Excessive accumulation of intracellular Ca²⁺ in ECs is critical and triggers the production of COX-derived EDCFs. This abnormality has been reported in arteries from hypertensive and diabetic subjects (Shi et al., 2007; Tang et al., 2007; Tang and Vanhoutte, 2010). Ca²⁺-independent PLA₂ inhibition also suppresses ACh-induced EDCF production and contraction in aorta of spontaneously hypertensive rats (SHR), a genetic rat model of hypertension (Wong et al., 2010a). Thus, both Ca²⁺-dependent and -independent mechanisms are involved in the vascular production of EDCF in hypertensive and diabetic conditions.

COXs (COX-1 and COX-2) are rate-limiting enzymes in the AA cascade (Rouzer and Marnett, 2009; Feletou et al., 2011; Nakano, 2015). AA is enzymically cyclized and oxygenated to produce the endoperoxide PGG₂. COX, which catalyses this cyclooxygenation reaction, also reduces a hydroperoxyl in PGG₂ to a hydroxyl to form PGH₂ via a separate peroxidase site on the enzyme (Simmons et al., 2004; Rouzer and Marnett, 2009; Salvemini et al., 2013). COX-1 is constitutively expressed and is usually abundant in ECs, whereas endothelial COX-2 is mainly induced by inflammatory stimuli (Feletou et al., 2009; 2010a,b,; 2011; Tang and Vanhoutte, 2009; 2010,; Vanhoutte, 2009; Grosser et al., 2010). A growing body of evidence suggests that COX-1 is the primary isoform related to endothelium-dependent contractions (Tang et al., 2005b; Tang and Vanhoutte, 2008; Vanhoutte, 2009; Wong et al., 2009; Feletou et al., 2010a,b,; Rovati et al., 2010). For instance, endothelium-dependent contractions are abolished by specific COX-1 inhibitors, but are relatively insensitive to specific COX-2 inhibitors (Yang et al., 2003a; 2004a,). In rat aorta, both COX-1 and COX-2 are detected; however, the amount of COX-2 transcripts in ECs or SMCs is markedly less than that of COX-1 (Tang and Vanhoutte, 2008). Moreover, endothelium-dependent contractions are seen in aortae isolated from wild-type and COX-2 deficient mice, but not from COX-1-deficient mice, suggesting that COX-1 is necessary for endothelium-dependent contractions (Tang et al., 2005b).

Although in healthy conditions the contribution of COX-2 to endothelium-dependent contractions is negligible, COX-2 is implicated in such contractions in hypertension, diabetes and aging (Matsumoto et al., 2007a; Shi et al., 2008). In arteries of SHR or diabetic rats, COX-1 expression is up-regulated, and COX-1 inhibitors block the augmented endothelium-dependent contractions (Yang et al., 2002; Gluais et al., 2006; Shi et al., 2007). COX-1-derived PGI₂, thromboxane A₂ (TxA₂) or endoperoxides all contribute to endothelium-dependent contractions in hypertension (Gluais et al., 2006; 2007,). In hamster aortae, endothelium-dependent, TxA₂/endoperoxide (TP) receptor-mediated contraction induced by ACh is mediated by PGF₂α. In this study, PGF₂α derived from COX-2 and not from COX-1, enhanced the COX-2/PGF₂α/TP pathway (Wong et al., 2009). Human renal arteries also exhibit TP receptor-mediated ACh- or PGF₂α-induced contractions and COX-2-dependent release of PGF₂α (Wong et al., 2009). Endothelial COX-2 rather than COX-1 seems to be responsible for enhanced generation of EDCF in aortae from rats under chronic treatment with the NOS inhibitor L-NAME (Qu et al., 2010).

In diabetic animal models, COX-2-derived prostanoids have been suggested to induce abnormal vasoconstrictor responses or to account for the development of endothelium-derived vasoconstrictor activity (Quilley and Chen, 2003; Bagi et al., 2005; Guo et al., 2005; Nacci et al., 2009; Lopez-Lopez et al., 2011; Ramos-Alves et al., 2012a,b,; Vessieres et al., 2013). Indeed, COX-2 expression and activity are increased in diabetic arteries (Bagi et al., 2006; Sanchez et al., 2010; Kassan et al., 2013; Martinez et al., 2014). Moreover, at the cellular level, high glucose, which is a characteristic of diabetes (De Vriese et al., 2000; Rask-Madsen and King, 2007; 2013,; Forbes and Cooper, 2013), increases COX-2 expression and activity in ECs (Cosentino et al., 2003; Sheu et al., 2005). Others, however, have reported up-regulation of COX-1 and increased production of COX-1-derived prostanoids in arteries of diabetic subjects (Matsumoto et al., 2007a; 2009b,; Shi et al., 2007; Shi and Vanhoutte, 2008; Feletou et al., 2011). Using COX-1-deficient mice fed with a high-fat diet and treated with streptozotocin (STZ; animal model of type 1 diabetes), Zhu et al. (2014) demonstrated that COX-1 remains the major contributor to the synthesis of endothelial PGI₂ that leads to vasoconstrictor activity in these conditions (Zhu et al., 2014). Discrepancies about the contribution of COX-1 and COX-2 to endothelium-mediated contractions in hypertension and diabetes may be attributed to different species, vascular beds, disease models and disease duration.

COX is involved in the endothelial generation of ROS (Tang and Vanhoutte, 2009; Feletou et al., 2010a; Sena et al., 2013; Hernanz et al., 2014). Increased COX-derived endothelial production of ROS contributes to the development of endothelial dysfunction in SHR aorta (Tang et al., 2007). ROS affects NO bioavailability (Hattori et al., 1991; Kamata and Kobayashi, 1996; Matsumoto et al., 2007b) and COX activation (Korbecki et al., 2013; Hernanz et al., 2014) and thus, it may also affect the balance between EDCF and EDRF signalling. Indeed, excessive formation of superoxide eliminates the vasodilatory, anti-thrombotic and anti-adhesive effects of NO and PGI₂ and stimulates the potent vasoconstrictor, pro-thrombotic and pro-inflammatory actions of PGH₂ and TxA₂ (Zou, 2007). Upon diffusion into vascular SMCs, ROS activate COX and further stimulates the production of contractile prostanoids (Katusic and Vanhoutte, 1989; Yang et al., 2002; Feletou et al., 2010a,b,; Tang and Vanhoutte, 2010). Moreover, ROS induce the production of isoprostanes such as 8-iso-PGF_2α and 8-iso-PGE₂, which can activate TP receptors on vascular SMC, leading to contraction (Bauer et al., 2014). Thus, ROS are key players in EDCF-mediated signalling (Tang et al., 2007; Tang and Vanhoutte, 2010) in hypertensive arteries (Vanhoutte et al., 2005; 2009,; Vanhoutte and Tang, 2008; Feletou et al., 2009; 2010a; 2011; Tang and Vanhoutte, 2010).

Prostanoids: endothelium-derived contracting factors

In the prostanoid pathway, both COXs oxidize AA to form unstable PGH₂ via the intermediate PGG₂. PGH₂ is further metabolized into PGD₂, PGE₂, PGF₂α, PGI₂ and TxA₂ by specific synthases including PGD synthase (PGDS), PGE synthase (PGES), PGF synthase (PGFS), PGI synthase (PGIS) and thromboxane synthase (TxS) respectively (Simmons et al., 2004; Tang and Vanhoutte, 2009; Figure 1). Among these prostanoids, PGI₂ and TxA₂ are primarily implicated in EDCF-mediated responses; however, PGD₂, PGE₂ and PGF₂α may also contribute to EDCF responses (Vanhoutte et al., 2009; Wong and Vanhoutte, 2010; Figure 1). Although endoperoxides have a very short life, they have contractile actions (Feletou et al., 2009). Stimulus (i.e. ligand), type of vessel and disease state are all determinants of which prostanoids and what amount of each prostanoid (and combination) would contribute to endothelium-dependent contraction. For example, in SHR aorta, ACh-induced endothelium-dependent contractions are largely due to the release of PGI₂ rather than other PGs (Gluais et al., 2005). However, TxA₂ contributes to EDCF-mediated contraction induced by other ligands such as ADP, A23187 and ET-1 (Gluais et al., 2006; 2007; Vanhoutte et al., 2009; Wong and Vanhoutte, 2010). ACh-induced increased production of TxA₂ and PGE₂, but not PGF₂α or PGI₂ has been reported in superior mesenteric arteries from type 2 diabetic Otsuka Long-Evans Tokushima Fatty (OLETF) rats (Matsumoto et al., 2007a; 2008b).

Prostacyclin is the major COX-derived metabolite of AA via PGIS in ECs (Tang and Vanhoutte, 2009; Wong and Vanhoutte, 2010; 2011). Gene expression of PGIS is by far the most abundant and averages more than 90% of the total expression of prostanoid synthases in EC of SHR aorta (Tang and Vanhoutte, 2008). Moreover, the extent of co-localization of PGIS with COX-1 in EC is greater than that with COX-2 (Kawka et al., 2007) and PGIS expression is increased in the aortae of SHR [vs. control Wistar-Kyoto rats (WKY)] (Numaguchi et al., 1999; Tang and Vanhoutte, 2008). Thus, the majority of the endothelial COX-1-derived endoperoxides are transformed into PGI₂ in hypertensive arteries (Gluais et al., 2005; 2006,; Tang and Vanhoutte, 2008; 2010,; Feletou et al., 2009). Prostacyclin-induced IP receptor activation usually results in endothelium-dependent vasodilation via the activation of the adenylyl cyclase/cAMP pathway (Alfranca et al., 2006; Feletou et al., 2010b). However, in experimental hypertension, PGI₂ binds to TP receptors to evoke vasoconstriction (Gluais et al., 2005; 2006,). This may be due to dysfunction of IP receptors and the AC pathway (Gomez et al., 2008). Indeed, the relaxation induced by PGI₂ or its stable analog iloprost is lost in aortae from SHR and aged WKY (Rapoport and Williams, 1996; Gluais et al., 2005). Although high concentrations of PGI₂ led to vasoconstriction via the activation of TP receptors, PGI₂ may be only weakly effective at TP receptors, as PGI₂ is rapidly degraded into its inactive metabolite 6-keto-PGF₁α (Gluais et al., 2005). In disease states including hypertension and diabetes, PGI₂ abnormalities refer to its interaction with the IP receptor and its downstream signalling as well as its synthesis by PGIS. In hypertension and diabetes, ROS interact with NO to form peroxynitrite (Zou et al., 2004; Szabo, 2009; Forstermann and Li, 2011). Peroxynitrite inhibits PGIS activity via nitration of tyrosine residues (Zou et al., 2002; Nie et al., 2006; Zou, 2007). Thus, the regulation of PGIS activity (e.g. suppression of nitrosative stress) and of IP receptors are important to improve PGI₂ signalling in hypertension- and diabetes-associated vasculopathy (Szabo, 2009; Forstermann and Li, 2011; Vanhoutte, 2011).

TxA₂, a metabolite of AA through TxS enzymic activity, mediates a number of cellular responses including vasoconstriction and platelet aggregation (Narumiya et al., 1999; Nakahata, 2008). Alterations in TxA₂ biosynthesis and actions have been investigated in a series of pathophysiological conditions such as atherosclerosis, myocardial ischaemia, asthma, diabetes, hypertension and pregnancy-induced hypertension (Meagher and FitzGerald, 1993; Hopkins, 2013; Taguchi et al., 2014). The contribution of TxA₂ to endothelium-dependent contraction may be ligand, vessel and disease specific. ACh-induced endothelium-dependent contractions are not affected by TxS inhibitors in SHR aortae (Auch-Schwelk et al., 1990; Tang and Vanhoutte, 2009) or canine basilar arteries (Katusic et al., 1988). On the other hand, inhibitors of TxS suppressed ADP-, A23187- and AA-induced endothelium-dependent contraction in the same preparations (Katusic et al., 1988; Gluais et al., 2006; 2007,). Treatment of type 2 diabetic OLETF rats with a TxS inhibitor suppressed ACh-induced TxA₂ production, partly improved ACh-induced endothelium-dependent, NO-dependent and EDHF-dependent relaxations, and inhibited ACh-induced endothelium-dependent contractions in superior mesenteric arteries. These data suggest that TxS inhibition normalizes endothelial dysfunction in type 2 diabetes (Matsumoto et al., 2009c).

As mentioned above, PGD₂, PGE₂ and PGF₂α also induce endothelium-dependent contractions; however, the contribution of these prostanoids to EDCF-mediated contractions is considered marginal in most cases, including disease states (Vanhoutte and Tang, 2008; Tang and Vanhoutte, 2009).

Expression of PGDS is increased in response to fluid shear stress (Taba et al., 2000). Studies using lipocalin-type PGDS (L-PGDS)-deficient mice indicate that L-PGDS plays an important role in regulating insulin sensitivity and the development of atherosclerosis in type 2 diabetes (Ragolia et al., 2005). PGD₂ may act as an EDCF via activation of TP receptors rather than DP receptors (Gluais et al., 2005; Feletou et al., 2011).

Expression of PGES and PGFS has been detected in aortae from the SHR (Tang and Vanhoutte, 2008; Tang et al., 2008) but their expression was lower than the expression of PGIS. Further, there was no difference in production of either PG between WKY and SHR aortae (Gluais et al., 2005; 2006; 2007,,). Under normal physiological conditions, PGE₂ and PGF₂α are natural agonists of the EP and FP receptors, respectively (Alfranca et al., 2006; Feletou et al., 2011). In hypertension, however, PGE₂ and PGF₂α activate the TP receptor (Gluais et al., 2005; Feletou et al., 2011). PGE₂ and PGF₂α can act as EDCFs when PGIS is inhibited or the metabolism of PGH₂ is diverted (Gluais et al., 2005), a phenomenon that may occur when severe oxidative stress leads to tyrosine nitration of PGIS (Zou et al., 2002). Therefore, in the SHR aortae, TxA₂, PGH₂, PGI₂ and, depending on the situation, PGE₂ and PGF₂α can all act as EDCFs (Feletou et al., 2010a).

Interactions between vasoconstrictor prostanoids and other endothelium-derived factors

In healthy ECs, the production and/or release of EDCFs is tempered by the presence of NO (Tang et al., 2005a; Vanhoutte and Tang, 2008; Tang and Vanhoutte, 2009; Vanhoutte et al., 2009) and EDHFs (Michel et al., 2008a). Endothelium-dependent contractions induced by ACh and nucleotides were relatively weak under basal conditions, but NOS inhibition amplified these contractions (Yang et al., 2003b; 2004a,b,; Matsumoto et al., 2007a; Ishida et al., 2011). Accordingly, in cases of endothelial dysfunction (viz. impaired bioavailability of EDRFs), the production of EDCF predominates. Treatment with L-arginine, a substrate for NOS (Forstermann and Li, 2011), inhibited contractions and normalized relaxation responses to ACh in renal arteries from hypertensive Dahl salt rats, whereas U46619-induced contraction and nitroprusside-induced endothelium-independent relaxation were not affected by L-arginine treatment (Zhou et al., 2001). Endothelium-dependent contractions were suppressed in arteries exposed to endogenous or exogenous NO (Tang et al., 2005a; Feletou et al., 2008). EDCF, on the other hand, leads to reduced NO bioavailability, suggesting a reciprocal relationship between NO and EDCF. ROS derived from stimulation of EC or TP receptor activation (Zhang et al., 2008; Del Turco et al., 2014) may reduce NO bioavailability. Activation of TP receptors in ECs has inhibitory effects on NO production, suppressing endothelium-dependent vasodilation (Liu et al., 2009). EDCF can also suppress EDHF signalling. Michel et al. (2008a) found that EDCF-mediated contraction in rat renal arteries was amplified not only by NOS inhibition, but also EDHF inhibition. In contrast, TxA₂ can interact with signalling pathways associated with EDHF (Ellinsworth et al., 2014). TP receptor stimulation is associated with loss of small-conductance Ca²⁺-activated potassium (SK_Ca) channel activity and with decreased EDHF-mediated responses in the rat mesenteric artery (Crane and Garland, 2004).

ET-1 is an EDCF with an important role in the pathogenesis of hypertension and diabetes (Maguire and Davenport, 2014; Matsumoto et al., 2014a; Sandoval et al., 2014). Interactions between ET-1 and vasoconstrictor prostanoids have been previously reported. Earlier studies showed that incubation of human aortic ECs with ET-1 led to production of the PGI₂ metabolite 6-keto-PGF₁α, the TxA₂ metabolite TxB₂ and PGE₂ (Hollenberg et al., 1994). In rat aorta, ET-1 induced the release of PGH₂ from ECs and this effect was greater in the hypertensive state (Asano et al., 1994). Basal and ET-1-stimulated release of TxB₂ was found in endothelium-intact aortae from SHR, but not WKY rats. ET-1-induced contraction was reduced by denudation of the endothelial layer, inhibition of TxS, or antagonism of TP receptors in aortae from SHR, but not WKY (Taddei and Vanhoutte, 1993). Furthermore, exogenous ET-1-induced vasoconstriction was largely blocked by intrabrachial administration of indomethacin (Taddei et al., 2000) in patients with essential hypertension, but not in normotensive subjects. Thus, ET-1/prostanoid interactions may be essential in the regulation of vascular tone.

Several studies in humans also suggest interactions between EDCF and other endothelium-derived factors. In healthy subjects, infusion of ACh into the brachial artery led to a concentration-dependent increase in flow-mediated dilation, which was diminished by NOS, but not by COX inhibition (Taddei and Vanhoutte, 1993; Taddei et al., 1993; 1998,). In contrast, ACh-induced vasodilatation was blunted in patients with essential hypertension and was resistant to NOS inhibition (Dohi et al., 1990). COX inhibition enhanced ACh-induced vasodilatation in these patients, suggesting that COX-derived vasoconstrictors are largely responsible for the abnormal response to endothelium-dependent vasodilators in essential human hypertension (Vanhoutte et al., 2005; Versari et al., 2009; Virdis et al., 2010).

Vasoconstrictor prostanoid-mediated signalling in vascular SMCs: focus on TP receptors

Prostanoids signal their constrictor effects on vascular SMC primarily via activation of the TP receptor, a member of the seven transmembrane GPCR superfamily (Nakahata, 2008; Woodward et al., 2011; Alexander et al., 2013,,,,). In humans, two distinct TP receptor isoforms have been found, namely TPα and TPβ, but these are not present in other species, such as non-human primates and rodents (Hopkins, 2013). TP receptors are expressed in endothelial and vascular SMC. Those located in vascular SMC are the primary contributors to the hypertensive phenotype (Sparks et al., 2013) and to COX-dependent, endothelium-dependent vasoconstriction (Yang et al., 2003a).

TxA₂-induced contractions in isolated bovine aortic SMC and isolated vascular tissue from various species involve both Ca²⁺-dependent and -independent mechanisms. Various mouse arteries exhibit contractile responses to ACh (carotid > abdominal aorta > femoral) and these responses are abolished by TP receptor antagonism (Zhou et al., 2005), suggesting that endothelium-dependent contractile responses in healthy vascular tissues are evoked through activation of TP receptors. Although TxA₂ is the preferred TP receptor ligand, high concentrations of endoperoxides, other PGs, as well as isoprostanes and hydroxyeicosatetraenoic acids (HETEs; Roman, 2002; Miyata and Roman, 2005) also activate the TP receptor (Feletou et al., 2010a; 2011,). 20-HETE has been shown to induce arterial contraction via TP receptor activation. The rate-limiting step of this action is the conversion of 20-HETE to 20-endoperoxides (20-OH-PGH₂, 20-OH-PGG₂) by COX (Escalante et al., 1989; Schwartzman et al., 1989). 20-HETE-induced contraction is partly dependent on the presence of the endothelium (Schwartzman et al., 1989; Randriamboavonjy et al., 2003) and is abolished by COX inhibition (Escalante et al., 1989; Schwartzman et al., 1989) and by TP receptor antagonists (Schwartzman et al., 1989). In cerebral arteries, flow-mediated constriction involves increased production of ROS and COX activation, and is mediated by the interaction between 20-HETE and TP receptors (Toth et al., 2011). These data suggest that TP receptors may be an integral part of all eicosanoid pathways.

Activation of TP receptors mainly involves stimulation of Gq-dependent phospholipase C (PLC) β, leading to Ca²⁺-dependent activation of myosin light chain (MLC) kinase and 20 kDa MLC (MLC₂₀) phosphorylation (Fukata et al., 2001; Wilson et al., 2005). Increases in Ca²⁺ influx due to opening of both receptor-operated and voltage-gated Ca²⁺channels contribute to TP activation-induced Ca²⁺-dependent effects (Okon et al., 2002). One of the signalling molecules activated by TP receptors in smooth muscle is Rho kinase (Somlyo and Somlyo, 2003). The activation of Rho kinase leads to inhibition of myosin light chain phosphatase, which reduces MLC₂₀ dephosphorylation (Somlyo and Somlyo, 2000; Nunes et al., 2010). Activation of TP receptors involves receptor co-coupling to G₁₂ and activation of RhoA/Rho kinase (ROCK) signalling, engaging Ca²⁺-independent increases in overall levels of phosphorylated MLC₂₀, which leads to vascular contraction (Wilson et al., 2005). To study the relationship between ROCK and EDCF signalling, Chan et al. (2009) investigated the expression of ROCK and the effect of Rho kinase inhibitors on endothelium-dependent contraction in aorta isolated from 1-year old SHR and WKY. They found that in the presence of an NOS inhibitor, ROCK inhibitors reduced endothelium-dependent contractions induced by ACh and the Ca²⁺ ionophore A23187 in aortae from both SHR and WKY. ROCK inhibitors did not affect the production of 6-keto PG F₁α (PGI₂ metabolite), but suppressed U46619- and PGF₂α-induced contraction and ROCK expression in both groups (Chan et al., 2009). These findings suggest that the suppression of EDCF-mediated contraction by ROCK inhibition is mainly due to direct suppression of EDCF-mediated signalling in vascular SMCs and not due to an effect on EDCF release (Chan et al., 2009). Denniss et al. (2010) reported that increased ACh-induced contractions in SHR (vs. WKY) were abolished by endothelial denudation or COX-1 inhibition, and nearly eliminated by TP receptor antagonists or by ROCK inhibition (Denniss et al., 2010). ACh-induced PGI₂ production was greater in carotid arteries from SHR (vs. WKY), but ROCK inhibition did not affect PGI₂ production. Protein expression of RhoA, but not ROCK-II, as well as Rho activation was increased in carotid arteries of SHR (vs. WKY artery). In addition, RhoA activation was increased by ACh stimulation. Quenching superoxide with tiron or NAD(P)H oxidase inhibition by apocynin reduced ACh-induced contraction in SHR, whereas the superoxide dismutase mimetic tempol amplified the response. Exogenous H₂O₂-induced contractions in carotid arteries were greater in SHR than in WKY and these responses were abolished by COX-1 inhibition, and greatly reduced by TP receptor antagonism or ROCK inhibition (Denniss et al., 2010). These results suggest that the RhoA/ROCK pathway may be a molecular switch, transducing signals from endothelium-derived PGs and ROS, to turn on vascular SMC contractile pathways (Denniss et al., 2010). This evidence suggests that RhoA/ROCK is a key mediator of abnormal EDCF signalling in hypertension. In rat isolated mesenteric resistance and uterine arteries, inhibition of PKC, ERK1/2 and p38 MAPK (significantly reduced TP receptor-dependent contractions (Goulopoulou et al., 2012). These data indicate that in addition to RhoA/ROCK, PKC and MAPKs are also important downstream effectors of activated TP receptors (Bolla et al., 2002; Nakahata, 2008).

Activation of TP receptors significantly contributes to the pathogenesis of hypertension (Keen et al., 1997; Francois et al., 2004). TP receptor-deficient mice do not develop hypertension in response to chronic angiotensin II infusion and have a reduced BP response to chronic treatment with L-NAME (Keen et al., 1997; Francois et al., 2004; 2008,). Spontaneously hypertensive rats have increased endothelium-dependent contractions in response to ACh and endoperoxides and these responses appear after the development of the hypertensive phenotype (Ge et al., 1995; Yang et al., 2002; 2003a,). TP receptor antagonism reversed ACh-induced vasoconstriction to vasodilatory responses in cerebral arterioles and abolished isometric contractions in aortic rings from hypertensive rats (Okon et al., 2002; Yang et al., 2002). Moreover, contractions induced by the TP receptor agonist U46619 were increased in various arteries from SHR (vs. WKY; Gluais et al., 2005; Chan et al., 2007; Garcia-Redondo et al., 2015). Thus, TP receptors are critical mediators of endothelium-dependent contractions in hypertension.

As described above, ROS, which are enhancers/modulators of EDCF responses, modulate TP receptor function, density and stability. Hydrogen peroxide prevents the translocation and degradation of TP receptors, increasing their density at the cell membrane (Valentin et al., 2004), and TP receptor activation increases the stability of the receptors through a ROS-dependent post-transcriptional mechanism (Wilson et al., 2009). Thus, it is reasonable to speculate that in hypertension, a condition characterized by increased production of ROS, an increase in TP density and stability contributes to augmented endothelium-dependent contractions. Interestingly, there are no differences in the mRNA and protein expression of TP receptors between aortae of WKY and SHR (Tang and Vanhoutte, 2008; Tang et al., 2008), suggesting that an increase in TP receptor density is not the reason for the enhanced endothelium-dependent contractions found in the SHR aorta. Thus, an increase in EDCF production as well as alterations in signalling pathways downstream of TP receptors are more likely to contribute to the hyper-responsiveness to EDCF in hypertensive vessels.

Enhanced activation of TP receptors and augmented endothelium- and TP receptor-dependent vascular contractions are also observed in various metabolic diseases and experimental models of diabetes and hypercholesterolemia (Traupe et al., 2002; Jerez et al., 2008; Michel et al., 2008b). Increased U46619-induced contractions are seen in arteries from obese and diabetic animals (Nobe et al., 2008; Yang et al., 2008; Baretella et al., 2014; Matsumoto et al., 2014b). Obesity and diabetes augment the responses to EDCF, possibly due to an increase in TP receptor gene expression (Traupe et al., 2002). PG receptors other than TP are also involved in diabetes-associated increases in EDCF-induced contractions. In a rat model of type 1 diabetes (STZ-treated rats), PGE₂ promotes vascular contractions via EP receptor activation (Shi et al., 2007). RhoA/ROCK signalling is central in diabetes-associated EDCF-induced contractions (Ishida et al., 2012). However, TP receptor-induced augmented vasoconstriction in penile arteries from pre-diabetic obese Zucker rats is coupled to enhanced Ca²⁺ influx rather than ROCK-mediated augmentation of myofilament Ca²⁺ sensitization (Villalba et al., 2011).

In summary, in hypertension and diabetes, endothelium-dependent contractions are mainly TP receptor-dependent. The molecular mechanisms downstream of TP receptors involve Ca²⁺- dependent and -independent pathways, various kinases (ERK1/2, PKC, Rho kinase, p38 MAPK) and their predominance may be vascular bed and disease specific (Figure 1). TP receptors are a promising pharmacological target because their antagonism may suppress EDCF-induced contractions and restore the balance between EDRF and EDCF vasoactive actions (Belhassen et al., 2003; Zuccollo et al., 2005; Gelosa et al., 2010).

Pharmacological targets for the control of EDCF signalling in diabetes and hypertension

Therapies aimed to correct the imbalance in the synthesis/release and signalling of vasoconstrictor and vasodilator prostanoids may be important in the treatment of hypertensive and diabetic subjects. Accordingly, putative targets for the control of EDCF signalling in hypertensive arteries have been suggested.

The regulation of COX activity plays an important role in the control of EDCF signalling. Non-steroidal anti-inflammatory drugs (NSAIDs) inhibit production of PGs by acting on COX-1 and/or COX-2 (Amer et al., 2010). Non-selective NSAIDs, such as naproxen and ibuprofen, inhibit both COX-1 and COX-2 whereas selective NSAIDs act on COX-1 (aspirin) or COX-2 (celecoxib) isoenzymes (Amer et al., 2010). Although COX inhibition seems a possible strategy to prevent COX-associated vascular complications, the incidence of serious adverse cardiovascular effects with COX-2 selective inhibitors (rofecoxib or Vioxx®) has greatly antagonized this concept (Bunimov and Laneuville, 2008; Wong et al., 2010b). There is consensus, however, that low-dose aspirin exerts protective vascular effects. Further investigation of COX inhibitors is required, especially towards their specificity and/or direct inhibition of PGIS activity (Wong et al., 2010b). Meanwhile, antagonism of TP receptors may emerge as a therapeutic alternative to reverse prostanoid-mediated vascular dysregulations (Chamorro, 2009; Jones et al., 2009; Giannarelli et al., 2010; Siller-Matula et al., 2010; Wong et al., 2010b; Davi et al., 2012). Since the actions of prostanoids are complex, the regulation of prostanoid signalling may need to be carefully planned, on a disease-by-disease basis.

Other molecular targets for the suppression of EDCF responses in hypertension and diabetes have been reported (Table 2006). Ford and Rush reported that 5-aminoimidazole-4-carboxamide-1-β-D-ribofuranoside (AICAR), an activator of AMP-activated protein kinase (AMPK; Lempiäinen et al., 2012), suppresses EDCF-mediated responses in aortas from SHR and this was reversed in the presence of the AMPK inhibitor Compound C (Ford and Rush, 2011). Spitler et al. recently found that endoplasmic reticulum (ER) stress contributed to increased COX-1 expression and prostanoid release (Spitler et al., 2013). In addition, ER stress increased phosphorylation of ERK1/2 and cytosolic PLA₂ (cPLA₂), contributing to enhanced EDCF-mediated responses. Treatment with chemical chaperones [ER stress inhibitor, taurourosodeoxycholic acid or 4-phenylbutyric acid] decreased ACh- and AA-induced contraction, COX-1 expression and phosphorylation of cPLA₂ and ERK1/2 in SHR aorta. Accordingly, pharmacological inhibition of ER stress suppressed EDCF-mediated responses in aortas and lowered BP in SHR (Spitler et al., 2013). Li et al. (2011) found that treatment with haemin, an inducer of haem oxygenase-1 (HO-1; Ndisang et al., 2001), decreased EDCF-mediated contractions induced by ACh or A23187 via suppression of COX-1, but not COX-2 in aortae from SHR (Li et al., 2011). Moreover, induction of HO-1 with haemin improved endothelium-dependent vascular relaxation through suppression of ROS production and inhibition of COX-2 upregulation induced by diabetes (Wang et al., 2014). Therefore, drugs aimed to increase AMPK or HO-1 activity to inhibit ER stress may represent therapeutic strategies for the suppression of EDCF-induced contractions and the improvement of treatment of vascular function in hypertension.

Table 1.

Potential pharmacological tools to suppress enhanced vascular EDCF-mediated signalling in hypertensive and diabetic models

Compound name	General action	Main outcome (animal model)	References
PDTC	Low-molecular-weight thiol antioxidant and potent inhibitor of NF-κB (Cau et al., 2011)	Suppresses EDCF- and AA (a source of EDCF)-mediated contraction, as well as the production of prostanoids (OLETF rat)	Matsumoto et al., 2009a
Metformin	A biguanide derivative and one of the most commonly used drugs for the treatment of type 2 diabetes (Almabrouk et al., 2014)	Reduces EDCF production and contraction induced by ACh and reduces oxidative stress (OLETF rat)	Matsumoto et al., 2008b
AICAR	An activator of AMPK (AMP-activated protein kinase), a putative regulator of metformin signalling (Lempiäinen et al., 2012)	Suppresses ACh-induced EDCF- mediated contraction and the production of vasoconstrictor prostanoids (TxA2 and PGE2; OLETF rat and SHR)	Matsumoto et al., 2008b; Ford and Rush, 2011
Pravastatin	One of the statins which inhibit the enzyme HMG-CoA reductase and the production of cholesterol (Kobayashi et al., 2000)	Pravastatin reduces EDCF-mediated responses by suppressing Rho kinase activity and by stimulating antioxidant activity (OLETF rat)	Ishida et al., 2012
Losartan	Angiotensin AT1 receptor antagonist (Kobayashi et al., 2008)	Reduces ACh-induced production of prostanoids and EDCF(PGE2)-mediated contraction in SMCs and normalizes oxidative stress (OLETF rat)	Matsumoto et al., 2010; Ishida et al., 2011
		Suppresses EDCF-mediated responses induced by purinergic receptor stimulation by suppressing COX-2 expression and cPLA2 phosphorylation in arteries (GK rat)
TUDCA PBA	Chemical chaperones (Kraskiewicz and FitzGerald, 2012)	Suppresses ACh- and AA-induced vascular contractions and decreases expression of COX-1 and activities of cPLA2 and ERK (SHR)	Spitler et al., 2013
Haemin	HO-1 inducer (Ndisang et al., 2001)	Suppresses ACh- and A23187-induced vascular contraction via a decrease in COX-1, but not COX-2 (SHR)	Li et al., 2011
EPA	An omega-3 PUFA that modulates cell membrane fatty acid composition (Calder, 2012).	Suppresses EDCF- and AA-induced contractions and improves EDRFs (NO and EDHF)-mediated relaxation – beneficial effects that are associated with reduced COX-2 expression and activities of ERK and NF-κB (OLETF)	Matsumoto et al., 2009b

PBA, 4-phenylbutyric acid; PDTC, pyrrolidine dithiocarbamate; PUFA polyunsaturated fatty acid; TUDCA, tauroursodeoxycholic acid.

In OLETF rats, a model of type 2 diabetes associated with the metabolic syndrome, we found increased production and signalling of COX-derived vasoconstrictor prostanoids in mesenteric arteries (Matsumoto et al., 2007a). Although the causal factors of enhanced vasoconstrictor prostanoids in OLETF rats remain unclear, potential pharmacological tools leading to suppression of EDCF-mediated signalling have been reported (Table 2006). AICAR, an AMPK activator (Lempiäinen et al., 2012), inhibited ACh-induced contraction and production of vasoconstrictor prostanoids (Matsumoto et al., 2008b). Likewise, metformin (Almabrouk et al., 2014) decreased ACh-induced contractions, as well as prostanoid superoxide production (Matsumoto et al., 2008b). Treatment with pyrrolidine dithiocarbamate, an inhibitor of NF-κB (Cau et al., 2011), suppressed ACh- and AA-induced contraction and ACh-induced prostanoid release (Matsumoto et al., 2009a). Treatment with pravastatin [an hydroxymethylglutaryl-CoA (HMG-CoA) reductase inhibitor] attenuated endothelium-dependent contractions by inhibiting ROCK activity and up-regulating antioxidant activity (Ishida et al., 2012). The angiotensin II AT₁ receptor antagonist, losartan (Kobayashi et al., 2008), suppressed ACh-stimulated prostanoid production, ACh- and AA-induced contraction and reduced superoxide production (Matsumoto et al., 2010). Treatment with losartan suppressed endothelium-dependent contraction by nucleotides and cPLA₂ phosphorylation in superior mesenteric arteries from type 2 diabetic Goto-Kakizaki (GK) rats at the chronic stage of the disease (Ishida et al., 2011). The fish oil, eicosapentaenoic acid (EPA), an omega-3 polyunsaturated fatty acid that modulates cell membrane fatty acid composition (Calder, 2012) has several beneficial effects against vascular dysfunction. EPA has anti-inflammatory, anti-platelet, anti-fibrotic, anti-oxidant and anti-atherogenic effects and also increases NO bioavailability (Balakumar and Taneja, 2012; Calder, 2012; Yates et al., 2014). Treatment of OLETF rats with EPA suppressed ACh- and AA-induced contraction, improved ACh-, NO- and EDHF-mediated relaxation, reduced ACh-stimulated prostanoid release, COX-2 expression and activities of ERK1/2 and NF-κB (Matsumoto et al., 2009b). A summary of the actions of these compounds on EDCF signalling is presented in Table 2006. Collectively, these studies indicate that the beneficial effects of these drugs or dietary supplements may be partly attributable to their capacity to normalize EDCF-mediated signalling. In other words, the regulation of EDCF signalling may be a therapeutic target for diabetic and/or hypertensive vasculopathy.

Effects of sex hormones on prostanoid-induced vascular response and signalling

Sex hormones influence EC function and, consequently, EDRFs and EDCF synthesis and their signalling pathways in physiological and pathological conditions (Orshal and Khalil, 2004; Khalil, 2005; Miller and Mulvagh, 2007; Duckles and Miller, 2010). Oestrogen stimulates the release of various endothelium-derived factors (Tostes et al., 2003; Hermenegildo et al., 2006; Miller and Mulvagh, 2007). 17β-oestradiol (E2) stimulates NOS in human and animal ECs – ECs from human umbilical vein (HUVEC), bovine aortae and human aortae (Hayashi et al., 1995; Hishikawa et al., 1995). E2 also induces the production of PGI₂ in HUVEC (Mikkola et al., 1995; Sobrino et al., 2010) and ovine fetal pulmonary artery ECs (Sherman et al., 2002). Moreover, oestrogen leads to increased EDHF-mediated signalling (Burger et al., 2009). In cerebral arteries, E2 reduced vascular tone by shifting the primary end product of the endothelial COX-1 pathway from the constrictor PGH₂ to dilator PGI₂ (Ospina et al., 2003).

Effects of oestrogens are mediated by multiple receptors such as the classic oestrogen receptors ERα and ERβ and the novel G protein-coupled oestrogen receptor (GPER, previously termed GPR30; Revankar et al., 2005; Prossnitz and Barton, 2011) cloned from human ECs (Takada et al., 1997). Natural oestrogens such as E2, a non-specific agonist of ERα, ERβ and GPER (Prossnitz and Barton, 2011), modulate vasoconstrictor prostanoid signalling (Miller and Vanhoutte, 1990; Dantas et al., 1999; Zhang and Kosaka, 2002; Li and Stallone, 2005; Hermenegildo et al., 2006; Li et al., 2008). Moreover, inhibitory effects of E2 on COX-dependent responses to vasoconstrictors have suggested a role for oestrogen receptors (Meyer et al., 1997) on endothelium-dependent contraction. However, the specific receptor(s) involved in this mechanism have not been described.

GPER activation is associated with beneficial and protective effects in the vasculature (Meyer et al., 2011; Han et al., 2013; Ferreira et al., 2015). Meyer et al. (2012) showed that acute GPER antagonism enhanced endothelium-dependent contractions and reduced endothelial NO bioactivity. Chronic GPER deficiency was associated with increased endothelial prostanoid-induced vasoconstriction but had no effect on endothelial NO bioactivity, eNOS and TP receptor gene expression or vascular structure. GPER deletion also increased TP receptor-mediated contraction (Meyer et al., 2012). Therefore, the regulation of GPER activity may have therapeutic potential for vascular complications associated with vasoconstrictor prostanoid signalling in hypertension and diabetes.

Testosterone also influences the release of endothelium-derived factors by both inhibiting EDRF release and increasing EDCF expression (Farhat et al., 1995; Hutchison et al., 1997). Testosterone has been shown to stimulate TxS as well as COX-1 and COX-2 in rat thoracic aorta and mesenteric arteries (Cheuk et al., 2000; Song et al., 2004). Moreover, in vitro treatment with testosterone increased the number of functional TP receptors both in cultured rat aorta (Masuda et al., 1991) and in vascular SMCs of the guinea pig coronary artery (Schror et al., 1994). In addition to its effects on vascular tone under physiological conditions, testosterone also has pro-inflammatory effects. Accordingly, long-term treatment with testosterone, or with its non-aromatizable androgen receptor agonist dihydrotestosterone, exacerbates endotoxin-induced inflammation in the cerebral circulation by mechanisms that involve increased nuclear NF-κB activation and increased levels of COX-2 and inducible NOS (Gonzales et al., 2009). Reinforcing a role for testosterone in vascular inflammation, data from our laboratory demonstrate that testosterone induces leukocyte migration by COX-2-dependent mechanisms (Chignalia et al., 2015).

The modulatory effects of sex hormones on vascular tone may also be due to their influence on the production of vasoconstricor prostanoids. Miller and Vanhoutte (1990) found that AA-induced endothelium-dependent contractions of aortic rings were enhanced by treatment of ovariectomized rabbits with oestrogen for 2 weeks (vs. placebo treatment). In addition, oestrogen treatment augmented contractions induced by PGI₂ but not PGE₂, PGF₂α or U46619, whereas indomethacin suppressed noradrenaline-induced contraction in endothelium-intact aortae from estrogen-treated rabbits. These results suggested that chronic treatment with oestrogens could affect noradrenaline-induced contraction via an endothelium-dependent mechanism that may involve the metabolism of AA by COX, and that altered sensitivity of the SMC to PGI₂ may contribute in part to the enhanced contractions to AA upon oestrogen treatment (Miller and Vanhoutte, 1990). Oestrogen also potentiates vascular reactivity to vasopressin (VP), which releases TxA₂ and PGI₂ from both male and female rat aortae (Li et al., 2008). Whereas ovariectomy attenuated, oestrogen therapy restored VP-stimulated release of TxA₂ and PGI₂, an effect mediated by upregulation of COX-2 and TxS expression in both ECs and vascular SMCs and up-regulation of TP expression in vascular SMCs (Li et al., 2008). Moreover, sex differences in the endothelial regulation of vasoconstrictor responses due to modulatory effects on vasoconstrictor prostanoids have been described. Whereas the endothelium negatively modulates clonidine (α₂-adrenoceptor agonist)-induced contraction entirely via NO in female rats, an endothelial vasoconstrictor prostanoid contributes to clonidine responses in male animals (Tejera et al., 1999).

Important sex differences in endothelial (dys)function have been reported in hypertensive and diabetic subjects (Kauser and Rubanyi, 1995; Hermenegildo et al., 2006; Aloysius et al., 2012). In hypertensive rats, E2 affects the release and/or action of endothelium-derived NO (Huang et al., 1997; Costa et al., 1998) and enhances endothelium-dependent relaxation in aortae of female SHR (Williams et al., 1988). Further, it antagonizes the increased tone in renal arteries of female Dahl salt-sensitive rats by enhancing NO-dependent relaxation and suppressing EDCF-mediated responses via NO-independent mechanisms (Zhang and Kosaka, 2002). Regarding the effects of female sex hormones on the synthesis and the effects of EDCFs in the vasculature, Kahonen et al. demonstrated that diclofenac, a COX inhibitor, abolished sex differences in ACh vascular responses in SHR (Kahonen et al., 1998). In addition, the removal of ovarian steroid hormones increased the generation of COX-derived vasoconstrictors, such as PGH₂/PGF_2α (Davidge and Zhang, 1998; Dantas et al., 1999). Sex differences in renal prostanoid production have been reported in arterial hypertension, with female SHR exhibiting enhanced urinary excretion of PGE₂ and TxA₂ metabolites along with enhanced renal microsomal PGES and COX-2 expression, compared with male SHR (Sullivan et al., 2005). In rabbit isolated carotid arteries, testosterone induces a concentration-dependent relaxation, which is increased in diabetic conditions by mechanisms that involve increased release of NO and COX-2-derived PGI₂ rather than the absence of COX-1-derived TxA₂ (Marrachelli et al., 2010). An imbalance of prostanoid synthesis, with overproduction of vasoconstrictor prostanoids and reduced PGI₂ production has been observed in diabetes-associated vascular dysfunction in males (Bolego et al., 2006; Du et al., 2006; Nie et al., 2006; Matsumoto et al., 2007a; 2008b; 2009b; 2010,,,; Ishida et al., 2011). Using a model of diabetes in female rats, Akamine et al. (2006) reported that ACh-induced relaxation, which was decreased in arterioles of diabetic female rats, was ameliorated by diclofenac. In addition, improved production of PGF_2α and 6-keto PGF_1α, but not TxB₂, as well as superoxide generation were observed in ACh-stimulated arterioles from diabetic rats, and diclofenac normalized these alterations. These data suggest that increased release of constrictor prostanoids, most likely PGF_2α, is involved in the reduced endothelium-dependent vasodilation in diabetic females and that enhanced COX activity may be a source of superoxide generation in this model (Akamine et al., 2006).

The above-mentioned reports indicate that sex and sex hormones influence EDCF synthesis and EDCF-mediated signalling in physiological conditions and that EDCF-mediated responses are disrupted in diabetic and hypertensive conditions. While this adds further complexity to the regulation of EDCF signalling, it raises the possibility of using different therapeutic treatments to target EDCF-mediated responses in male and female patients.

Uridine adenosine tetraphosphate

Up₄A, a dinucleotide with purine and pyrimidine moieties, was identified by Jankowski et al. as a novel potent EDCF (Jankowski et al., 2005). Up₄A is released from ECs in response to various stimuli such as mechanical stress, endogenous ligands (ACh, ET-1, ATP and UTP) and Ca²⁺ ionophore (A23187; Jankowski et al., 2005). Although the molecular mechanisms underlying the production/release of Up₄A remain unclear, Jankowski et al. recently demonstrated that the intrinsic enzymic activity of the VEGF receptor 2 leads to the production of Up₄A in ECs (Jankowski et al., 2013). Only two papers have been published on circulating Up₄A levels in disease states. Jankowski et al. found that circulating levels of Up₄A were increased in juvenile hypertensive patients compared with normotensives (Jankowski et al., 2007) and Schuchardt et al. found that patients with chronic kidney disease had a higher plasma Up₄A concentration compared with healthy subjects (Schuchardt et al., 2012).

The effects of Up₄A on vascular functions have been explored in various animal models (Matsumoto et al., 2011a; Figure 2). Up₄A induces vascular calcification (Schuchardt et al., 2012), proliferation and migration of vascular SMC (Gui et al., 2011; Wiedon et al., 2012). Because these events play important roles in the development of vascular dysfunction in diabetes and hypertension (Touyz and Schiffrin, 2004; Chen and Moe, 2012), the regulation of Up₄A-induced signalling in vascular SMCs may be a potential therapeutic target. Furthermore, Up₄A has been shown to modulate vascular tone in various arteries (Matsumoto et al., 2011a). For example, Up₄A induces relaxation of rat aorta (Linder et al., 2008) and swine coronary artery (Zhou et al., 2013) and stimulates contraction in the rat pulmonary artery (Gui et al., 2008), rat aorta (Linder et al., 2008), mouse aorta (Hansen et al., 2010) and rat perfused kidney (Jankowski et al., 2005). Although evidence from vascular functional studies were all derived from non-disease models, we recently found that Up₄A-induced contraction is altered in various arteries isolated from deoxycorticosterone-acetate (DOCA)-salt hypertensive rats (Matsumoto et al., 2011b; 2012) and type 2 diabetic GK rats (Matsumoto et al., 2014b). In DOCA-salt hypertensive rats, Up₄A-induced contraction is increased in basilar, femoral and renal arteries and is decreased in small mesenteric arteries (vs. control uninephrectomized rats; Matsumoto et al., 2011b; 2012). In renal arteries from hypertensive rats, increased contraction to Up₄A may be attributable to increased ERK1/2 activity in SMCs rather than alteration of P2Y receptors, which are putative Up₄A receptors (Jankowski et al., 2005; Matsumoto et al., 2011b). Moreover, we recently found that the Up₄A-induced contraction is increased in renal arteries from GK rats (vs. control Wistar rats), and this might be due to the enhanced activation of COXs/TP receptor signalling (Matsumoto et al., 2014b). These findings provide the possibility of cross-talk between Up₄A and vasoconstrictor prostanoids in the presence of diabetes.

Figure 2.

Effects of Up₄A on vascular smooth muscle functions. When EC are stimulated with, for example, ACh, ATP, uridine triphosphate (UTP), ET-1, mechanical stress or Ca²⁺ ionophore (A23187), Up₄A is generated. Released Up₄A acts in both ECs and vascular SMC. In ECs, Up₄A binds to P2Y1/2 or P1 receptor and activates endothelial nitric oxide synthase (eNOS); releasing NO and producing relaxation. In vascular SMCs, Up₄A leads to calcification, proliferation, migration, inflammation and contraction. Details are shown in text.

Uridine adenosine tetraphosphate-mediated signalling in vascular SMCs

Since Up₄A has purine and pyrimidine moieties, Up₄A may also exert its effects via activation of purinergic receptors in vascular SMCs. Purinergic receptors have been classified into two subtypes including P1 (i.e. adenosine receptor) and P2 [P2X (ionotropic) and P2Y (metabotropic)] receptors (Abbracchio et al., 2006; Burnstock, 2007; Erlinge and Burnstock, 2008; Alexander et al., 2013b,c,; Burnstock and Ralevic, 2013). There are reports showing a link between Up₄A and purinoceptors in the vasculature. For instance, Jankowski's group demonstrated that Up₄A-induced vasoconstriction in perfused kidneys results from the activation of P2X1 receptors and probably also of P2Y₂ and P2Y₄ receptors (Jankowski et al., 2005). More recently, the same group found that in addition to smooth muscle P2X1 receptor-mediated vasoconstriction in the rat perfused kidney, Up₄A also induces concentration-dependent P2Y₂ receptor-mediated, long- lasting vasoconstriction (Tolle et al., 2010). Moreover, Up₄A-induced vasoconstriction is followed by vasodilation due to P2Y₁ and P2Y₂ receptor activation on ECs, which leads to NO release (Tolle et al., 2010). Using rat pulmonary arteries, Gui et al. (2008) found that Up₄A induces concentration-dependent contraction that is inhibited by suramin, a non-selective P2 receptor antagonist, but not by Ip₅I, an antagonist of P2X receptors, or by desensitization of P2X receptors with α,β-methylene-ATP (Gui et al., 2008). We also confirmed that Up₄A-induced rat renal arterial contraction is inhibited by suramin but not by Ip₅I (Matsumoto et al., 2011b). Linder et al. found that Up₄A-induced contraction in rat aorta is blocked by P1 and P2X receptor antagonists (Linder et al., 2008).

The downstream pathways of Up₄A-induced responses in vascular SMCs have been investigated (Matsumoto et al., 2011a; Figure 2). Linder et al. found that Up₄A-induced rat aortic contraction is suppressed by L-type Ca²⁺ channel blockade or by ROCK inhibition (Linder et al., 2008). On the other hand, Gui et al. suggested that Up₄A-induced contraction in rat pulmonary artery involves extracellular Ca²⁺ influx and Ca²⁺ release from intracellular stores, but not Ca²⁺ sensitization via the ROCK pathway (Gui et al., 2008). In rat renal arteries, we found that Up₄A-induced contraction was inhibited by the ERK1/2 pathway inhibitor PD98059, whereas the ERK1/2 activity (determined by phosphorylated ERK1/2 levels) was increased upon Up₄A stimulation in renal arteries from DOCA-salt hypertensive rats (vs. control uninephrectomized rat; Matsumoto et al., 2011b). These data suggest that Up₄A-induced responses rely on intracellular calcium-dependent and -independent (e.g. kinases) pathways in vascular SMCs.

Up₄A-induced vascular SMC migration dramatically depends on secretion of osteopontin (OPN), a multifunctional molecule associated with cell migration, adhesion, survival and tissue remodeling (Scatena et al., 2007). Simultaneous incubation with Up₄A and an OPN-blocking antibody suppressed vascular SMC migration (Wiedon et al., 2012). Using specific and non-specific purinoceptor antagonists and inhibitors of MEK/ERK pathway, Wiedon and colleagues demonstrated that Up₄A-induced vascular SMC migration was mediated by MEK/ERK pathway upon P2Y₂ receptor activation. Furthermore, Up₄A-induced migration was reduced by a platelet-derived growth factor receptor (PDGFR) antagonist and by PDGFR-β siRNA treatment, suggesting that transactivation of the PDGFR plays a role in vascular SMC migration mediated by Up₄A (Wiedon et al., 2012). Schuchardt et al. found that exogenous application of Up₄A increased mineral deposition in mouse and rat aortae and in rat vascular SMCs. In addition, Up₄A increased the expression of different genes specific for osteochondrogenic vascular SMCs such as Cbfa1 (Steitz et al., 2001; Naik et al., 2012), while decreasing the expression of SM22α, a specific marker for vascular SMCs (Steitz et al., 2001; Dong et al., 2012). The influence of different P2Y receptor antagonists on Up₄A actions indicated that P2Y_2/6 receptors might be involved. Mechanisms downstream of P2Y receptor signalling involved activation of the MEK/ERK1/2 pathway. Therefore, Up₄A activation of P2Y receptor could affect phenotypic transdifferentiation of vascular SMCs to osteochondrogenic cells, suggesting that purinergic signalling upon Up₄A stimulation may be involved in vascular calcification (Schuchardt et al., 2012). Gui et al. (2011) found that Up₄A-induced increase in bromodeoxyuridine incorporation was blocked by the mammalian target of rapamycin and the MEK/ERK1/2 inhibitor, PD98059, in human vascular SMCs. In addition, Up₄A-induced phosphorylation and activation of S6 kinase (S6K) and ERK1/2 were inhibited by PD98059, whereas S6K but not ERK1/2 activity was inhibited by rapamycin. Up₄A also increased Akt phosphorylation, which was inhibited by the PI3K inhibitor, LY294002. Up₄A-induced activation of S6K, but not ERK1/2, was also prevented by LY294002, whereas a P2 receptor antagonist, suramin, but not a P2X receptor antagonist, Ip₅I, inhibited Up₄A-induced phosphorylation and kinase activity of S6K and ERK1/2. Finally, Up₄A increased protein expression of cyclin-dependent kinase 2 (CDK2), which was prevented by rapamycin, PD98059 and suramin. These results demonstrate that the signalling mechanisms underlying Up₄A-induced proliferation of vascular SMCs are mediated by P2Y receptors and involve the PI3K/Akt and ERK1/2 pathways, leading to the independent activation of S6K and increased in CDK2 expression (Gui et al., 2011). In addition, Up₄A promotes the production of the chemokine CCL2 in vascular SMC (Schuchardt et al., 2011), suggesting a role for this EDCF in vascular inflammation. Up₄A induced ROS generation via NAD(P)H oxidase activation, which further stimulated CCL2 formation through ERK1/2 and p38 MAPK activation in vascular SMCs. This process by Up₄A requires the activation of P2Y₂ receptors (Schuchardt et al., 2011).

Intracellular signalling of Up₄A in the vasculature of hypertensive and diabetic subjects has not been fully characterized. The effects of sex and sex steroid hormones of Up₄A production and activity is also an unexplored area of study.

Conclusions and perspectives

In conclusion, vasoconstrictor prostanoids and Up₄A play an important role in vascular dysfunction associated with diabetes and arterial hypertension. The signalling pathways activated by these EDCFs may differ based on vessel type, stages of disease (e.g. early or chronic stage) and sex. The complexity of the intracellular signalling and the interactions between vasoconstrictor prostanoids and Up₄A provide an exciting area of investigation for the pursuit of new pharmacological targets for the management of vascular dysfunction in hypertension and diabetes. Sex and sex steroid hormones are essential mediators of these pathways and should be considered in the design of experimental studies and the development of therapeutic compounds.

Glossary

AA

arachidonic acid

AICAR

5-aminoimidazole-4-carboxamide-1-β-D-ribofuranoside

AMPK

AMP-activated protein kinase

CDK2

cyclin-dependent kinase 2

cPLA₂

cytosolic PLA₂

DOCA

deoxycorticosterone-acetate

E2

17β-oestradiol

ECs

endothelial cells

EDCF

endothelium-derived contracting factor

EDHF

endothelium-derived hyperpolarizing factor

EDRF

endothelium-derived relaxing factor

EPA

eicosapentaenoic acid

ER

endoplasmic reticulum

ET-1

endothelin-1

GK

Goto-Kakizaki

GPER

G protein-coupled oestrogen receptor

HETEs

hydroxyeicosatetraenoic acid

HO-1

haem oxygenase-1

HUVEC

human umbilical vein endothelial cells

L-NAME

N^G-nitro-l-arginine methyl ester

L-PGDS

lipocalin-type PGD synthase

MLC₂₀

myosin light chain 20

NO

nitric oxide

NSAIDs

non-steroidal anti-inflammatory drugs

OLETF

Otsuka Long-Evans Tokushima Fatty

OPN

osteopontin

PDGFR

platelet-derived growth factor receptor

PGI₂

prostacyclin

PGIS

prostacyclin synthase

ROCK

Rho kinase

ROS

reactive oxygen species

S6K

S6 kinase

SHR

spontaneously hypertensive rats

SMCs

smooth muscle cells

STZ

streptozotocin

TP

TxA₂/endoperoxide receptor

TxA₂

thromboxane A₂

TxS

thromboxane synthase

Up₄A

uridine adenosine tetraphosphate

VP

vasopressin

WKY

Wistar-Kyoto rats

Conflict of interest

There are no potential conflicts of interest among the authors regarding the publication of this manuscript.