Aging-Shifted Prostaglandin Profile in Endothelium as a Factor in Cardiovascular Disorders

Age-associated endothelium dysfunction is a major risk factor for the development of cardiovascular diseases. Endothelium-synthesized prostaglandins and thromboxane are local hormones, which mediate vasodilation and vasoconstriction and critically maintain vascular homeostasis. Accumulating evidence indicates that the age-related changes in endothelial eicosanoids contribute to decline in endothelium function and are associated with pathological dysfunction. In this review we summarize currently available information on aging-shifted prostaglandin profiles in endothelium and how these shifts are associated with cardiovascular disorders, providing one molecular mechanism of age-associated endothelium dysfunction and cardiovascular diseases.


Introduction
Cardiovascular disorders, including atherosclerosis, coronary artery disease, heart failure, and hypertension, remain the leading cause of death worldwide [1]. These diseases are among several pathological conditions that are associated with aging [2][3][4], and age is a primary risk factor for their development [5,6]. Endothelium is a thin layer of epithelial cells which line the interior of lymph and blood vessels and is a major component of the vascular wall. One important contributor to the development of cardiovascular diseases is a dysfunctional endothelium. Endothelial dysfunction is considered a fair predictor of cardiovascular diseases [4,[7][8][9][10][11].
Normal endothelial function is regulated by a controlled balance between endothelium-dependent relaxing factors and endothelium-dependent contracting factors. The main vasoactive factors released by endothelial cells are nitric oxide (NO) and cyclooxygenase-(COX-) derived eicosanoids [4, 40,41]. NO production has been shown to be reduced with aging [42][43][44][45]. There is less information on how eicosanoids change in the endothelium with age. It is also not well understood how changes in eicosanoid profile might contribute to endothelium dysfunction. Nevertheless, accumulating evidence indicates that the age-related changes in endothelial eicosanoids contribute to endothelium dysfunction and to the development of age-associated cardiovascular diseases.
There is limited data on how eicosanoids change in humans [4], and most experiments have been conducted in animal models and most commonly in rat [2]. Rats of 1.5-2 months or less are considered immature, rats of 3-6 months are considered young adult, and rats of approximately 24 months or more are considered aged, though there are differences between strains [2].

Cyclooxygenases and PGH 2
There are two isoforms of the cyclooxygenases (COX1 and COX2) encoded by two different genes. Both COX1 and COX2 are expressed in the endothelial and vascular smooth muscle cells, and the expression levels are 20-fold higher in endothelial cells than in smooth muscle cells [78]. In endothelium, both of the COX enzymes are constitutively expressed [79,80]. However, they are also inducible, for instance, by shear stress [79][80][81]. Endothelial cells express COX1 preferentially over COX2 [82,83].
In human mesenteric microvessels of individuals greater than 80 years of age, COX1 levels are 50% increased, while COX2 levels are slightly decreased [21]. In normotensive rats, both COX1 and COX2, in either whole vascular tissue or endothelial cells from vasculatures, are increased with aging from 1-fold to 5-fold [29,42,63,[84][85][86]. Comparable effects of aging on COX1 and COX2 expression levels have been observed in mice [33,34]. At similar ages, COX1 or COX2 expression, measured at the mRNA or protein levels, is almost doubled in the aorta of spontaneous hypertensive rats (SHRs) as compared to normotensive control Wistar-Kyoto (WKY) rats [63,84,87,88]. Similar increases in COX1 and COX2 were observed in N ω -nitro-L-arginine methyl ester-(L-NAME-) induced hypertensive rats as compared to control Sprague-Dawley rats [89]. Increased COX2 was also reported in the renal artery of hypertensive patients [89]. These data indicate that there are age-associated increases in COX1 and COX2 levels, as well as an association between elevated COX1/COX2 levels, in both animal models and human studies, and clinical cardiovascular disorders.

PGI 2
PGI 2 (prostacyclin) is the first described metabolite of arachidonic acid, and endothelium is the major site of its biosynthesis [51,57]. In endothelium, both COX1 and Journal of Aging Research 3 COX2 are the upstream contributors of PGI 2 synthesis [80,[102][103][104]. PGI 2 is synthesized by its terminal specific PGI 2 synthase (PGIS) [105, Figure 1]. PGIS colocalizes with COX1 in endothelial cells [106]. In endothelium, PGIS is by far the most abundant PG terminal synthase, with its expression level 5-100-fold higher than the other PG terminal synthases [54,64,65,84]. Accordingly, PGI 2 is the most abundant endothelial eicosanoid, with expression levels 10-100-fold higher than that of the other eicosanoids in humans [107,108] and in animals [54,97,109,110]. PGI 2 triggers potent vasodilation [51,57] by interacting with the PGI 2 receptor (IP) (Figure 1), which located in smooth muscle cells [108,111]. The vasodilation effect of PGI 2 has also been shown in pig coronary arteries at low concentrations [58]. At higher concentrations PGI 2 may induce vasoconstriction [32, 54,64]. PGI 2 cannot cause vasoconstriction until its concentration reached 1 μM or higher. 1 μM is 1000-fold higher than the endogenous concentration of PGI 2 , which is in the 0.2-1 nM range [112]. Even at elevated concentrations, PGI 2 is a weak vasoconstrictor and induces modest tension in the rat aorta [32, 54,64]. Modest vasoconstrictive effects of PGI 2 may emanate from weak cross-activation of TP, which can induce vasoconstriction [49]. At lower concentrations, PGI 2 , especially endogenous PGI 2 , is a vasodilator. In addition, PGI 2 is the most potent endogenous anticoagulation agent [113]. The vasodilation and anticoagulation effects of PGI 2 have been confirmed by a recent report showing that IP deletion in mice results in hypertension and reduced anticoagulation activity [114].
In human blood, PGI 2 , measured as PGF 1α , is 400 pg/mL in new born infants, 230 pg/mL in infants, 150 pg/mL in adolescents, and 85 pg/mL in adults [112]. Age-associated PGI 2 decline is also observed in urine of humans [115,116]. The endothelium is the main site for PGI 2 synthesis [50,51]. Although there has been no report on PGI 2 production in isolated human vessels, PGI 2 levels were reported to decline in cultured human vascular endothelial cells during serial passage [117][118][119]. Based on these reports, one would expect that PGIS in endothelium decreases with age. Yet there have been no reports evaluating age-associated PGIS changes in the human endothelium. In the endothelial cells from rat aorta, there is a slight and insignificant age-associated decrease in PGIS mRNA [84]. However, additional evidence shows that mRNA or protein of PGIS is 2-4-fold higher in aorta or coronary arteries of aged normotensive rats [85,86,110,120] suggesting that lower PGI levels may be caused by increased PGI 2 degradationwith age, rather than the change in PGI 2 synthesis. In fact, there is no apparent correlation between circulating PGI 2 level with level of endothelial PGIS, suggesting the necessity of investigation of the effects of age on the metabolism/degradation of PGI 2 . More work is needed to determine whether circulating PGI 2 correlates to endothelial PGI 2 and to clarify the effects of age on PGI 2 in the endothelium and in the circulation. Age-associated reduction in IP level has been consistently reported in rats [84,85]. The reduced IP is expected to lead to reduced sensitivity to PGI 2 effects. Consistently, dilation in response to PGI 2 is significantly blunted in aged humans as determined by forearm blood flow measurements [121].
Reports on the change in PGI 2 or PGIS under pathological conditions, such as hypertension, are contradictory. While one group reported a 50% reduction in PGI 2 in SHR aorta as compared to WKY aorta [96], another group reported insignificant differences in PGI 2 levels in SHR and WYK rats [64,65]. In addition, Tang and Vanhoutte reported that PGI 2 mRNA is 4-fold higher in the endothelial cells of SHR aorta than in WKY aorta [84]. These limited and inconsistent reports indicate a need for more complete and thorough investigations into how aging affects PGI 2 , its synthase, receptor, and metabolism. Moreover, clarifying PGI 2 effects in the development of cardiovascular disorders in animal models and in humans could be of potential therapeutic significance.

PGE 2
Prostaglandin E 2 (PGE 2 ) is the most abundant prostaglandin in the human body. In endothelium, however, its level is lower than that of PGI 2 , in line with a lower expression level of the corresponding synthases, which are 5-100-fold lower than PGIS [54,64,65,84]. There are three types of known PGE 2 synthases (PGESs), the cytosolic PGES (cPGES) and two forms of membrane PGES, mPGES1 and mPGES2 [122, Figure 1]. cPGES is constitutively expressed and functionally coupled to COX1 [122,123]. mPGES1 is inducible and functionally coupled with COX2 [124] and is the major PGE 2 synthase responsible for PGE 2 production [123]. In endothelium, the expression levels of the PGESs are comparable to other PG synthases [54,64,65,84]. Consistently, the amount of PGE 2 in endothelium is comparable to other PGs, but lower than the amount of PGI 2 [54,97,[107][108][109][110]125]. In further accord, the contribution of PGE 2 to endothelium-dependent vasoaction is marginal [84,125]. Chen et al. showed that deletion of mPGES1 in mice resulted in abolished production of PGE 2 but did not affect blood pressure [114]. Yang, on the other hand, showed that mPGES1 deletion in mice resulted in exaggerated hypertensive in response to high salt and angiotensin II infusion [126], suggesting that mPGES1 may be an important physiological regulator of blood pressure. While the role of mPGES1 in blood pressure regulation is debatable, mPGES1 is implicated in atherosclerosis. Deletion of mPGES1 in mice retards atherosclerosis development [127]. PGE 2 acts through four PGE 2 receptors (EP1, EP2, EP3, and EP4), which are mainly located in the smooth muscle cells in the vessels [125, 128, Figure 1]. Activation of EP1 and EP3 receptors induces calcium mobilization/release and inhibits adenylyl cyclase release, which triggers vasoconstrictions [111,129]. In contrast, activation of EP2 and EP4 receptors stimulates adenylyl cyclase and induces cyclic adenosine monophosphate release, which triggers vasorelaxation [111,129]. The vascular actions of PGE 2 are complex due to the opposing vasoactions triggered by the binding of PGE 2 to the variant PGE 2 receptors. Depending on the circumstances, PGE 2 may be vasodilating [47,[67][68][69][70] or vasoconstricting [53,54,[71][72][73]. In addition to the distributions 4 Journal of Aging Research of different PGE 2 receptors expressed in the vascular system, PGE 2 concentration is also important. This complexity likely explains the reported inconsistent effects of mPGES1 deletion on blood pressure [114,126]. PGE 2 has a biphasic effect on human blood platelet aggregation. At low concentrations (0.01-1 μM), it potentiates platelet aggregation, and, at higher concentrations (10 μM), it inhibits ADP-and collagen-induced aggregation in platelet rich plasma [71,[130][131][132]. The endogenous PGE 2 concentration is below 1 μM [133], making PGE 2 a stimulator of atherosclerosis. Thus, reduced PGE 2 level by mPGES1 deletion retards atherosclerosis development [127].
There is little information available on age-related changes in any of the PGESs, PGE 2 , or EPs. A recent report by Tang and Vanhoutte revealed that while cPGES and mPGES1 in the aorta endothelial cells are insignificantly higher in aged rats, mRNA of mPGES2 is 5-fold higher [84], which can presumably result in higher level of PGE 2 . PGE 2 secreted from coronary arteries is increased in aged rat as compared to young rats [120]. Expression of EP1-4 increased with age, with EP4 elevated 2-fold in endothelial cells from rats of 72 weeks as compared with rats of 36 weeks [84]. Since vasoaction depends on the ligand and the type of receptors, ageincreased PGE 2 and EP4 are assumed to predispose to increased vasodilation. Further investigation is required to determine the effect of age-related changes in PGE 2 and its synthases and receptors in different vascular beds and on relaxation/constriction of vasculatures.
Information on the effects of aging on PGF 2α is limited. PGFS mRNA was doubled in the endothelial cells from aged rat aorta as compared to that from young rat aorta [84]. Consistently, PGF 2α is 2-fold higher in the aorta of aged rats versus young rats [110,148]. Change in FP mRNA in the endothelial cells of rat aorta with age, however, is insignificant [84]. Basal PGF 2α is slightly higher in the aorta of SHRs than that of WKY rats, but the difference is increased upon acetylcholine stimulation [54]. Research needs to be conducted to obtain more complete information on ageassociated changes in PGF 2α in humans and the effects of those changes on the development of cardiovascular disorders.
There is only one report on the effect of aging on PGDS and DP. While aging had no effect on L-PGDS, it caused a 5fold increase in H-PGDS mRNA in aged rat aorta endothelial cells [84]. Age had no apparent effect on DP [84]. H-PGDS is 3-fold higher in aorta endothelial cells from SHRs versus WKY rats, whereas L-PGDS is decreased in these cells in SHRs versus WKY rats [84]. In the smooth muscle cells from the same aorta preparations, DP mRNA was measured to be 3-fold higher in SHRs as compared with WKY rats [84].
TxA 2 elicits diverse physiological/pathophysiological reactions, including platelet aggregation and vascular smooth muscle contraction [49]. Activation of platelet aggregation is thought to be the dominant biological function of TxA 2 . TxA 2 causes platelet shape change, aggregation, and secretion, which promotes thrombus formation and thrombosis [168][169][170][171]. Thrombosis can cause acute myocardial infarction and atherogenesis [166,[171][172][173][174]. TxA 2 -induced contraction effects are variable, depending on the specific vascular beds examined and the agent used to induce contraction [116,175,176]. The majority of reports coincide with the view that the contraction induced by endotheliumderived TxA 2 is weak, because inhibitors of TXS do not induce relaxation [91,92,96,176]. Contraction effects are likely mediated by TP activated by PGH 2 because inhibitors of PGHSs and TP induce relaxation [91,92,96,175,176]. Several publications reported a 2-5-fold increase in TxA 2 in aorta or mesenteric arteries of aged rats as compared to that of young rats [42,86,172]. Consistently, Tang and Vanhoutte reported a 4-fold increase in TXS mRNA [84]. In contrast, a single investigation of age-dependence of TxA 2 did not find any significant difference in TxA 2 between young and aged rat aortas [110]. Aging did not show any significant effect on rat aorta TP mRNA [84].
An increased production of TxA 2 has been found in patients and animal models of several cardiovascular diseases including unstable angina [177], experimental myocardial ischemia and infarction [178], cerebral vasospasm, pregnancy induced hypertension [179,180], and congenital heart disease [116]. TxA 2 levels reported in those studies are systemic, rather than endothelial. In endothelium, there is no difference in aorta TxA 2 between SHRs and WKY rats [54,64,65,87]. However, TXS mRNA is doubled in the aorta endothelium of SHRs versus WKY rats [84]. Age-related changes in TP have not been found [84,181].
In summary (Table 1), aging has been consistently shown to cause severalfold increase in COXs, that is, the synthesis of PGH 2 [29, 42,63,[84][85][86]. Aging probably reduces PGI 2 , the predominant PG in the endothelium [112,[115][116][117][118]182], though it is not certain and requires more work. Aging has been shown, or has the potential, to change other PGs in the endothelium. However, because the level of PGI 2 is 10-100fold higher than that of the rest of PGs, the shift of PG profile in the endothelium during aging will be predominantly determined by PGI 2 and untransformed PGH 2 . PGI 2 and PGH 2 have opposing effects on vessels and platelets. The net result of the effects of aging will be a shift toward a proconstrictive mediator profile, as shown in Figure 2.

Therapeutics That Modulate Prostaglandins in Cardiovascular Disorders
Because prostaglandins and thromboxane are such important factors in endothelium functions and therefore in the physiology and pathology of the vascular system, numerous pharmacological agents that target these factors have been developed to mitigate cardiovascular diseases. As listed in Table 2, prostacyclin (PGI 2 ) and analogues are used clinically to treat hypertension, especially pulmonary hypertension [75,[200][201][202]. They are also used to inhibit arterial thrombosis and ameliorate myocardial ischemia [203][204][205][206][207]. Although the vascular actions of PGE 2 are complex, PGE 2 and analogues are used to reduce blood pressure and to alleviate congestive heart failure [208][209][210], owing to their ability to stimulate renin release and natriuresis and diuresis [211][212][213]. PGE 2 , PGE 1 , and their analogues are more often used to maintain the patency of the ductus arteriosus in infants with congenital heart disease [214][215][216][217]. Antagonists of TXS and TP are potent antithrombosis agents and used to treat atherosclerosis, myocardial ischemia, and stroke [218][219][220][221][222][223][224][225][226][227]. The underlying principle of the design of these drugs is to selectively increase the effects of vasodilators and anticoagulators and to selectively reduce the effects of vasoconstrictors and coagulators by modulating the amount of ligands, synthases, or receptors of a specific eicosanoid. Because prostaglandins and thromboxane A 2 are from the same precursor but elicit opposing effects, selectivity is crucial in the design of these therapeutics. Nonselective inhibition of the upstream synthases, COX1 and COX2, can result in undesirable side effects including hypertension, manifestation of myocardial ischemia, and increased incidents of acute myocardial infarction and stroke, which occur more often in the elderly [104,[228][229][230].
Intriguingly, low dose of aspirin, an inhibitor of COX1, is popularly used in the prevention of cardiovascular diseases [231][232][233]. Aspirin covalently acetylates a specific serine moiety (serine 530 of COX-1 and serine 516 of COX-2) [234,235], and its binding to COX1 is about 170-fold stronger than that to COX-2 [236]. Thus, aspirin is a covalent inhibitor of COX1 inactivating it irreversibly. TxA 2 is mainly produced in platelets [100,101], whereas PGI 2 is mainly produced by endothelial cells [51,57]. Different from most other cell types, platelets do not possess nuclei, which are required for protein synthesis. While COX1 can be regenerated in other cells, such as endothelial cells, COX1 cannot be regenerated in platelets. Nor can COX1 activity be recovered after inactivation by aspirin. Therefore, low dose of aspirin irreversibly and selectively inhibits TxA 2 production in platelets.
However, new platelets are constantly formed, and TxA 2 is persistently produced [237], which leads to a need for continuous dosing to constantly inhibit COX1. Aspirin resistance is a common clinical phenomenon [238] and has been observed for more than twenty five years [239]. Aspirin resistant patients, partially due to inherited polymorphisms in COX1 [240,241], have a nearly 4-fold increase in risk of suffering a vascular event compared with aspirin responders [242][243][244]. As an alternative to aspirin therapy, antagonists of TXS and TP, which can also be combined with aspirin, have been applied to ameliorate thrombosis and prevent cardiovascular diseases [226].

Conclusion and Perspective
The incidence and prevalence of cardiovascular diseases increase with advancing age, to the extent that age has been identified as the dominant risk factor for these pathologies [2,[4][5][6]. It is well established that PGs are powerful endogenous vasodilators and vasoconstrictors and platelet aggregators, playing important roles in regulating homeostasis in vascular systems. Although limited, the current analysis of the literature suggests that there is a modified PG profile associated with age and indicates that age has significant effects on the abundance of PGs, their synthesis, as well as their signaling transduction pathways. Aging-modulated PG profile offers a potentially important molecular mechanism underlying age-dependent endothelial dysfunction and ageassociated cardiovascular diseases. Knowledge of age-associated PGs profile changes can be important for designing new pharmacological interventions to prevent or slow down age-associated cardiovascular diseases. Given their biological roles, improved investigation of age-associated changes in PG synthesis, metabolism, and signaling in all major vascular beds is needed.
It is clearly difficult to obtain human vascular tissues to determine age associated changes. Surrogate tissues and fluids such as human blood or urine are plentiful but are of limited value for assessing tissue-specific effects. Defining the relationship between PGs, particularly PGI 2 and PGH 2 , in vascular tissues and the amounts in blood or urine in animal models could be helpful to interpret PG profiles in humans. Technical challenges exist due to metabolite instability. For example, PGH 2 is transformed to other PGs and is biologically important in its own right, but untransformed PGH 2 is difficult to measure [98,245]. The development of user-friendly methods could facilitate acquiring these measurements [91,98,245]. For example, PGH 2 can be instantly reduced to 12-heptadecatrienoic acid (12-HHT) by FeCl 2 [91,98,245]. 12-HHT is stable and inactive and measurable [91,98,245]. Therefore, total PGH 2 can be measured as 12-HHT. A relatively mild reducing agent, SnCl 2 , can reduce untransformed PGH 2 to PGF 2α . Untransformed PGH 2 can be calculated by subtracting the estimate of PGF 2α in samples without SnCl 2 from the corresponding estimate in samples with SnCl 2 [91,98,125]. Alternatively, epidemiological approaches could avoid these technical difficulties and offer valuable genetic information. Haplotype analyses have revealed that several polymorphisms in COX, PGIS, and IP are associated with age and cardiovascular diseases [246][247][248][249][250].
Research on an important aspect of age-associated changes in PGs is largely absent in the literature; that of age-associated effects on PG metabolism. One of the most important features of PGs is rapid clearance. Most PGs are metabolized to inactive forms within 1-3 minutes [119,251], and consequently their signaling is terminated within that time frame. This is due to an effective and efficient metabolism system mainly composed of prostaglandin transporter (PGT) and 15-hydroxyprostaglandin dehydrogenase (15-PGDH) [252]. Both PGT and 15-PGDH have been shown to regulate PG degradation [245,253,254]. Thus far, there have been no reports on the influence of age on PG metabolism.
In conclusion, PGs and TxA 2 play critical roles in many important events involved in the normal functions of vascular system, including vasodilation, vasoconstriction, platelet aggregation, and inflammation. Although these eicosanoids were discovered in the 1970s, the research into ageassociated shifts of the PG profile has just begun. Age-associated alterations in PG profiles are not only interesting, but also important in defining the molecular mechanisms of ageassociated cardiovascular pathological conditions and informing strategic and personalized prevention and cure of those diseases.