Estrogens and development of pulmonary hypertension: interaction of estradiol metabolism and pulmonary vascular disease.

The female preponderance of PAH is well documented, yet our understanding of the role of estrogens in development and progression of PAH are quite limited. Further studies of estradiol metabolism and in-depth investigation of the vascular effects of the more than dozen biologically active metabolites of E2 in PAH are required. Preclinical studies should include newer animal models of PAH with significant pulmonary endothelial injury and with genetically or pharmacologically altered E2 metabolism. The highly selective and sensitive liquid chromatography-mass spectrometry methods rapidly becoming available for simultaneous quantification of multiple estradiol metabolites should foster further investigation and elucidation of the roles of E2 and its metabolites in PAH. The effects of 2ME in patients with PAH are unknown. However, in several clinical studies investigating the potential antitumor effects of 2ME, high doses of 2ME were well tolerated in patients with solid malignancies ( 156 ). Likewise, in a phase I clinical trial in healthy volunteers, we recently detected no estrogenic or other adverse effects from a subcutaneously injected long-acting formulation of 2ME in doses up to 10 mg/kg ( 157 ). Both agents with strong anti-proliferative, anti-angiogenic and pro-apoptotic effects, such as 2-methoxyestradiol, and interventions that increase the bioavailability of endogenous 2ME potentially could be effective in the treatment of severe PAH. In this regard, a dozen 2ME analogs and inhibitors of 17 β -HSD (which may improve 2ME disposition) have been synthesized and are available for pharmacological evaluation, including investigation of potential therapeutic effects in PAH. Finally, our knowledge about the effects of progestins and androgens in PAH are scant at best. Progestins and androgens may also have significant influence on vascular pathobiology; because of their significant multilevel interactions with estrogens, further studies in both animals and humans are needed to fully understand the effects of sex steroid interplay on the development and progression of PAH.

The effects of progestins in experimental PAH have been poorly studied and the effects of progesterone in patients with PAH are unknown. This is surprising, as progesterone receptors are expressed in intact human endothelial cells (ECs) and in modified ECs within plexiform lesions from patients with PAH ( 121 , 122 ); natural progesterone inhibits proliferation of ECs and VSMCs ( 123 - 126 ); and progesterone exerts vasodilatatory properties in various vascular beds ( 127 , 128 ). With this mind, we recently examined the effects of natural progesterone and synthetic progestins medroxyprogesterone and tibolone on the development of MCT-induced PAH in ovariectomized female rats. Progesterone attenuated development of PAH and right ventricular hypertrophy and inhibited pulmonary vascular remodeling ( 129 ). In a separate study, medroxyprogesterone showed effects similar to progesterone, and notably, tibolone, a combined progestin/estrogen compound, prevented development of PAH and RV hypertrophy and eliminated MCT-induced late (35 days post-administration) mortality ( 130 ). These studies are limited by the fact that they were conducted in estradiol (E2) deficient animals, and therefore, it is unclear whether progesterone would demonstrate the same effects on PAH development in the presence of E2, given the significant and complex receptor-level interaction between the two compounds. Further studies are warranted to investigate more fully the role of progesterone in PAH and how its interaction with E2 influences its effects. There are no data on the effects of testosterone in PAH and limited data concerning its effects on pulmonary circulation. Testosterone induces vasodilation in the isolated rat pulmonary vasculature most likely via calcium antagonistic effects ( 131 ), and it has been confirmed that in isolated human pulmonary arteries, these vasodilatory effects are independent of gender ( 132 ). In rodents, testosterone is an even more potent pulmonary vasodilator than estradiol ( 133 ). Interestingly, recent studies suggest that dehydroepiandrostendione (DHEA), which serves as a circulating precursor for both estradiol and testosterone, has significant effects in experimental PAH. DHEA protects against development of PAH and reverses PAH associated with chronic hypoxia ( 134 - 136 ), and these effects are associated with increased expression and function of pulmonary artery Ca++ activated K+ channels ( 135 ) and upregulated soluble guanylate cyclase ( 136 ). Furthermore, DHEA reverses PAH in male rats with PAH induced by MCT and unilateral pneumectomy (which exhibits neointimal proliferation) and also in a model of occlusive/angioproliferative PAH (SU5416+hypoxia; 137 , 138 ); treatment with DHEA is associated with increased plasma levels of both E2 and testosterone. The fact that DHEA and its sulfated ester DHEAS have the highest level of all circulating steroids indicates that interconversion of circulating steroid hormones ( Figure 1 ) may serve as the mediator of the female predominance of human PAH, but data regarding the effects of DHEA in female animal models and patients with PAH are lacking.

A severe form of PAH in humans is characterized by clustered proliferation of endothelial cells (ECs) in the lumina of small size pulmonary arteries resulting in concentric obliteration of the lumina and the formation of complex vascular structures known as plexiform lesions ( 79 ). Compared with normal ECs, the endothelial cells in affected vessels show reduced prostacyclin and nitric oxide synthesis ( 80 , 81 ) and over-expression of endothelin-1 ( 82 ). Three-dimensional analysis of vascular lesions reveals the existence of two EC phenotypes in severe PAH: (i) normal quiescent, apoptosis-sensitive ECs located in the peripheral areas that are negative for phosphorylated MAPK and have a high expression of cell cycle inhibitory protein p27 kip1 (a marker of low growth potential), and (ii) highly proliferative, apoptosis-resistant cells in the central core of the vascular lesion that have elevated MAPK activity and increased expression of HIF1-α, VEGF protein and VEGF-2 receptor and low expression of p27 kip1 ( 4 , 83 ). Thus, the tumor-like proliferation of ECs may be the underlying process in severe PAH ( 84 ). Therefore, it is conceivable that agents that reverse the above alterations and have strong anti-proliferative, anti-angiogenic, and pro-apoptotic effects would be effective in the treatment of severe PAH. The published data suggest that this may be the case with 2-methoxyestradiol. Cumulating evidence suggests that E2 and its downstream metabolites may differ significantly in regard to their antimitogenic effects in cell lines involved in pulmonary vascular remodeling and fibrosis. Both the antimitogenic effects of E2 on proliferation of various types of animal and human vascular smooth muscle cells (VSMCs) and the beneficial effects of E2 on vascular remodeling in systemic circulation are well documented ( 85 , 86 ). However, data on the effects of E2 on pulmonary artery vascular smooth muscle cells (PASMCs) are limited, and it seems that the effects of E2 on PASMCs may differ from those in systemic VSMCs. In this regard, Farhat et al. ( 87 ) have reported that E2 dose-dependently stimulates thymidine incorporation in rat PASMCs; furthermore, in canine pulmonary arterial segments, E2 tends to inhibit PASMCs proliferation in segments with intact endothelium, but significantly stimulates thymidine incorporation in segments stripped of endothelium ( 87 ). It is possible that E2 may have different anti-remodeling effects on systemic and pulmonary blood vessels of different phylogenetic origin. Additionally, our own recent study in human PASMCs indicates that 2ME exhibits concentration-dependent antigrowth effects whereas E2 has only mild antimitogenic effects and only at micro molar (pharmacological) concentrations ( 88 ). Therefore, it is plausible that, in contrast to the well described inhibitory effects of E2 on pulmonary vascular remodeling in classical models of PAH (infra vide; Table 3 ), in PAH patients with severely altered endothelium, E2 may have no effect on media remodeling or, as described below, may even adversely affect endothelial remodeling. In human pulmonary artery endothelial cells (hPAECs), E2 stimulates proliferation at physiological concentrations (1–10 nM), has no effects at high physiological concentrations (100 nM), and inhibits EC growth at pharmacological concentrations (1 μM and 10 μM; 89 ). This is consistent with the previous report of biphasic effects of E2 on growth of human umbilical vein ECs ( 90 ). E2, via estrogen receptors, promotes the phosphorylation of p42/44 and p38 MAPK and downregulates cell cycle inhibitor p27 Kip1 , stimulates migration and proliferation of ECs, induces the synthesis of VEGF and HIF-1α expression, and protects ECs against apoptosis ( 91 - 94 ). The antigrowth effects of E2 observed at higher concentrations may be explained by E2 conversion to 2ME, as both enzymes critical to this conversion (CYP1A1 and COMT) are present in ECs. Further studies should test this possibility. In contrast to E2, 2ME at physiological concentrations (1– 10 nM) had no effects on hPAEC growth ( 88 , 110 ), but significant though mild stimulatory effects on human umbilical vein ECs ( 90 ). It is not clear whether this discrepancy is due to different experimental conditions or to the different EC type examined. In this regard, at 100 nM concentrations, 2ME stimulates the growth of porcine vascular ECs, but has antimitogenic effects in rabbit vascular endothelial cells ( 95 ). Previous studies of cellular growth in fibroblasts from various origins reveal both stimulatory and inhibitory properties of E2 ( 96 , 97 ). The variable effects of E2 may reflect the phenotypic heterogeneity among fibroblasts, including their differentiation into myofibroblasts. In fetal lung fibroblasts, E2 produces an antimitogenic effect only at micro molar concentrations ( 98 ), whereas in our recent study in human lung fibroblasts (hLFs), E2 had no effects on hLFs growth in concentrations up to 10 μM, while 2ME concentration-dependently (10nM-10uM) inhibited proliferation ( 88 ). Importantly, in all three types of cells involved in pulmonary vascular remodeling (hPAECs, hPASMCs and hLFs), the synthetic estradiol analog 2-ethoxyestradiol (2EE) is ten times more potent than 2ME in inhibiting growth ( 89 ). 2EE also inhibits vascular remodeling in MCT-induced PAH ( 89 ), suggesting that anti-proliferative agents, including synthetic analogs of estradiol metabolites, may be protective in PAH. 2ME is extensively metabolized by 17βHSD-2 to 2ME1, a metabolite largely considered to be biologically inactive. Indeed, recently we confirmed that in hPASMCs and human lung fibroblast only at high pharmacological concentration (10μM) 2ME1 has mild (−10%) anti-mitogenic effects ( 58 ). However, in the presence of retinoic acid, a type-2 17βHSD-2 inducer, 2ME1 strongly and concentration-dependently (10nM-10μM) inhibits growth in both types of cells ( 58 ), indicating that 2ME– 2ME1 inter-conversion takes place in cells involved in vascular remodeling in PAH. The fact that very fast cellular uptake and high (micromolar) intracellular plasma concentrations of 2ME have been reported ( 99 ), suggests that 2ME disposition, i.e., 2ME2-2ME1 inter-conversion may play critical role in the biological and pharmacological effects of 2ME in PAH. The first important aspect of the cellular effects of 2ME is that this major non-estrogenic metabolite of estradiol may modify many of the previously described alterations seen in humans with severe PAH, including levels of prostacyclin, endothelin and nitric oxide, three targets for currently approved therapies for PAH. In this regard, 2ME is markedly more potent, and 2HE somewhat more potent, than E2 itself in increasing prostacyclin synthesis and release ( 100 , 101 ) and in inhibiting endothelin synthesis and MAPK activity in endothelial cells ( 102 ), and 2ME is more potent than E2 in stimulating both basal and ionophore induced-NO release from aortic endothelium ( 103 ). The order of potency (2ME>2HE>E2) for inhibition of endothelin synthesis and MAPK activity is opposite the order of potency for binding to or activation of estrogen receptors, suggesting the involvement of estrogen receptor-independent mechanisms. At the present in is not clear whether stimulatory effects of E2 on prostacyclin and NO synthesis/release and inhibitory effects of E2 on endothelin system are mediated in part by its downstream metabolite 2ME. Importantly, 2ME inhibits cell growth and induces apoptosis in actively growing, but not in quiescent, endothelial cells ( 104 , 105 ). Furthermore, in a rat model of endothelial injury, 2-ME up-regulates p27 Kip1 , a key cell cycle inhibitor and a marker of low growth potential ( 106 ). 2ME also inhibits the synthesis of hypoxia–inducible factor-1α (HIF-1α 107 , 108 ), a transcription factor which regulates more than forty genes and their respective protein products, including those that play a key role in vascular reactivity and remodeling, angiogenesis, and cell proliferation ( 109 ). The significance of HIF-1α is evident (i) in severe PAH in humans where HIF-1α is over-expressed in obliterative endothelial lesions ( 4 ); (ii) in MCT- and chronic hypoxia-induced PAH, where a similar time dependent increase in HIF-1α levels correlates with the development of PAH and vascular remodeling ( 110 ; and (iii) in mice where heterozygous deficiency of HIF-1α and HIF-2α protects against development of PAH ( 111 , 112 ). The second important aspect of the cellular effects of 2ME is that at the endothelium level, 2ME may behave as a biological antagonist to E2. First, contrary to the case in other cardiovascular cells, E2 and 2ME have opposing effects on endothelial cells, a pivotal cell type in the pathobiology of PAH. Second, at physiological concentrations, 2ME interferes with non-genomic, ERα-mediated action of E2. Thus, in human leiomyoma cells, over-expression of COMT inhibits E2-induced proliferation, ERα signaling and HIF-1α expression ( 113 ), and importantly, the same effects are produced by 2ME. Similarly, in breast cancer cells, physiological concentrations (10-50nM) of 2ME inhibit the E2-induced growth and prevent E2-induced Akt phosphorylation ( 114 ). Therefore, it is plausible that in endothelium 2ME antagonizes the effects of E2. Based on its cellular effects, as a unique endogenous molecule, 2ME may be capable of correcting key alterations in pulmonary vasculature seen in severe PAH, and therefore should be expected to provide protection in severe PAH. Indeed, our previous studies unequivocally confirm that 2ME and its metabolic precursors and analogs are effective in treatment and prevention of experimental PAH in rats. Thus, in male rats, 2ME, its metabolic precursor 2HE, and/or its synthetic analog 2-ethoxyestradiol attenuate MCT-, hypoxia-, and α-naphthylthiourea-induced PAH ( 89 , 115 - 119 ). Furthermore, in studies in ovariectomized female rats with MCT-induced PH and bleomycin-induced pulmonary fibrosis and PAH, 2ME mediates the protective effects of E2 ( 89 , 120 ). Also, in a model of occlusive angioproliferative PAH, 2ME, but not E2, has preventive and therapeutic effects in intact and OVX female rats ( 30 ).

Estrogens exhibit a multitude of cardiovascular effects through genomic pathway by transcriptional activation of their cytosolic or membrane-bound receptors (ER) α and β. However, building evidence indicates an alternative rapid nongenomic pathway that involves membrane G-protein-coupled ER ( 8 , 9 ). The effects of estrogens on pulmonary vasculature are well defined and are mediated through both non-genomic and genomic mechanisms. Thus estradiol, via rapid, non-genomic mechanisms, increases prostacyclin release and production of nitric oxide ( 10 , 11 ), and through estrogen receptor-dependent mechanisms increases endothelial cells’ eNOS mRNA levels and eNOS activity ( 12 ). Furthermore, ovariectomy augments hypoxia-induced increase in endothelin-1 (ET-1) and prepro-ET1 mRNA levels, and E2 replacement reduces hypoxia-induced increase in preproET-1 mRNA and ET-1 peptide expression ( 13 ). The later effect of E2 is mediated via inhibition of hypoxia-induced increases in ET-1 promoter and estrogen response element-mediated reporter gene activity and involves a functional interaction between ER and HIF-dependent pathways. Notably, several reports suggest that 2ME (a major non-estrogenic metabolite of E2 with little or no affinity for estradiol receptors) is more potent that E2 itself in increasing prostacyclin and NO synthesis and in inhibiting endothelin synthesis ( infra vide ). In vitro, estradiol causes vasorelaxation of pulmonary artery under normoxic conditions ( 14 ) and inhibits hypoxic pulmonary vasoconstriction ( 15 ). Importantly, exogenous E2 rapidly attenuates pulmonary artery vasoreactivity and acute hypoxic vasoconstriction, suggesting non-genomic mechanisms of E2-induced pulmonary vasorelaxation ( 16 ). In experimental PAH induced by pneumotoxin monocrotaline (MCT) or by exposure to chronic hypoxia, 17β-estradiol (E2) has protective effects. When exposed to chronic hypoxia, female rats, mice and swine develop less severe PAH than male animals and female rats with intact ovaries develop more severe PAH than ovariectomized (OVX) rats, with estradiol attenuating the severity of disease in OVX rats ( 17 - 20 ). Also, estradiol improves pulmonary hemodynamics and vascular remodeling in perinatal PAH in lambs ( 21 ), suggesting that estradiol may be protective in PAH. Similarly, in rats with monocrotaline-induced PAH females develop less severe disease than males ( 22 , 23 ); pretreatment of male rats with E2 attenuates development of PAH and prevents against pulmonary vascular remodeling and right ventricular hypertrophy ( 24 ); and ovariectomy (OVX) exacerbates disease, whereas treatment of OVX rats with E2 attenuates it ( 25 ). Finally, in female and OVX rats with MCT-induced PAH, both selective estrogen receptor modulator raloxifene and phytoestrogens attenuate PAH ( 26 , 27 ). The apparent beneficial effects of estrogens in MCT- and hypoxia-induced PAH paradoxically oppose the female preponderance of the disease. In both chronic hypoxia and MCT models, the hallmark vascular change is medial thickening with little or no alteration of endothelium ( Table 2 ). This opposes the marked endothelial alterations seen in severe PAH in humans. As described below, in pulmonary artery for E2 to inhibit medial remodeling may require intact and quiescent endothelium, and therefore hypoxia- and MCT-induced PAH may not be appropriate model for studying the effects of estrogens in PAH. Our recent findings support this notion. In a rat model of occlusive/angioproliferative PAH which shares many features of the human form of the disease including endothelial disruption ( 28 , 29 ; Table 2 ), female rats develop more severe PAH than males, ovariectomy delays the development of PAH, and E2 has no preventive or therapeutic effects on elevated pulmonary artery pressure in OVX rats ( 30 ). However, in this model (similarly to humans) females develop less severe right ventricular hypertrophy than males. Finally, in recently developed models of PAH in mice in which disease is induced by consumption of dexfenfluramine ( 31 ) or over-expression of serotonin transporter ( 32 ) female gender does not provide protection, but rather is permissive in the development of PAH. These recent findings ( 30 - 32 ) contradict earlier studies of protective effects of female gender and estrogens in classical models of experimental PAH ( 17 - 27 ), and suggest that the classical models of PAH (i.e., chronic hypoxia and monocrotaline) may not be appropriate for evaluation of the effects of estrogens in PAH. Human studies, like animal studies, are inconclusive and even contradictory. That is, PAH mainly develops in young women and females are at higher risk of developing PAH ( 33 , 34 ), yet observational studies suggest estradiol may be protective against development of PAH in high altitude natives ( 35 ) and among patients with systemic sclerosis ( 36 , 37 ). In contrast, a recent study suggests that higher penetration of disease in women with a familial form of PAH is due to altered E2 metabolism ( 38 ). Thus, women diagnosed with familial PAH who have a bone morphogenic protein receptor type 2 mutation (BMPR2) have lower 2-hydroxyestradiol /16α–hydroxyestrone urine level ratios than healthy women with the BMPR2 mutation and never diagnosed with familial PAH ( 39 ). Furthermore, female gender is associated with increased risk of portopulmonary hypertension in patients with advanced liver disease ( 39 ). Also, independently of gender, in patients with advanced liver disease, genetic variation in aromatase (the rate-limiting enzyme in the conversion of androgens to estrogens), which leads to elevated plasma estradiol levels, is associated with increased risk of portopulmonary hypertension ( 40 ). Finally, the high prevalence of prolonged exposure to exogenous estrogens in both premenopausal and postmenopausal women with PAH suggests that estrogens may contribute to disease pathogenesis or susceptibility in PAH ( 41 ). The above discussion underscores the need for further investigation of the role of estradiol metabolism in PAH.

We previously proposed ( 115 ) that the apparent contradictions posed by the overall effects of estrogens in both experimental and human PAH—the estrogen paradox in PAH—may be explained by the complexity of estradiol metabolism and the influential balance between estradiol and its metabolites on pulmonary vascular homeostasis ( Figure 1 ). We further propose that in PAH, 2ME present in injured endothelium acts as biological antagonist of E2, and inhibits the mitogenic, angiogenic, and anti-apoptotic effects of E2 ( effects otherwise desirable in normal, quiescent endothelial cells). The adverse vascular effects of E2 in PAH may be even more significant if they are not opposed by 2ME (e.g., in the event of reduced E2 conversion to 2ME due to hypoxia, inflammation, drugs, environmental factors or genetic polymorphism of metabolizing enzymes). Both local conversion of estradiol to 2-methoxyestradiol and in situ 2ME disposition (i.e., 2ME -2ME1 inter-conversion by 17β-HSD) may have a beneficial impact on vascular pathobiology in PAH. Therefore, we hypothesize that unbalanced estradiol metabolism (i.e., a shift toward the 16-hydroxylation pathway and/or reduced activity of the 2-hydroxylation pathway, corresponding to elevated E2 and decreased 2ME levels) increases the risk of PAH and/or exacerbates the progression of disease.