Gender differences in the cardiovascular effects of sex hormones

Oestrogen-mediated effects

Oestrogens affect vascular function through nongenomic and genomic mechanisms [9–11]. In fact, in vivo oestrogen administration induces vasodilation that occurs within 5–20 min from dosing and is not associated with changes in gene expression. The response appears gender-related, as intracoronary infusion of estradiol improves coronary blood flow in female but not in male patients with coronary artery disease (CAD) [12].

The acute vasodilating effects of oestrogens have been attributed to both endothelium-dependent and independent pathways. While the latter are mediated by ion-channels and related to the calcium-antagonist properties of oestrogens, the former depend on endothelial nitric oxide (NO) activity and release, through activation of a subpopulation of oestrogen receptors (ER) tethered to endothelial cell caveolae [11,13].

Furthermore, the acute effect of 17β-estradiol on endothelial NO-synthase activity and NO release has been suggested to occur via activation of ER alpha (α), while endogenous ER-beta (β) exert a prominent role in the nongenomic effects on endothelial NO-synthase activity [13,14].

The genomic effects of oestrogens involve changes in vascular-cell gene and protein expression, mediated by two distinct nuclear ERs (α and β ERs), and result in improvement of endothelial function and blood vessels response to injury and development of atherosclerosis [9–11].

A broad array of co-regulatory proteins interacts with ERs to alter transcription of target genes, including the expression of other sex steroid hormone receptors [15]. In addition, the receptor-dependent effects of oestrogens might also be mediated by trans-membrane-coupled receptors, as selective stimulation of the trans-membrane G protein-coupled ER has been occasionally associated with reduction in blood pressure [16].

Both ERα and ERβ mediate physiologically important effects in blood vessels. Animal studies demonstrate that ERα mediate protection against vascular injury and atherosclerosis [17,18], while ERβ control genes are involved in arterial tone and blood pressure regulation [19,20].

Because the expression of sex hormone receptors may vary depending on gender, gonadic function, and degree of vascular atherosclerosis, gender differences in vascular reactivity may be partly related to specific receptor density in vascular tissues. Although receptor density assessment and quantization in different vascular beds remain difficult, female arteries appear endowed with greater ERs populations than those of men [21]. In addition, ERs’ density is significantly reduced in human coronary arteries of women with CAD, and women who have been postmenopausal for several years have a reduced number and activity of vascular ERs [22,23].

This finding could account for the observation that the vascular protective effects of oestrogens are reduced in women who have been long oestrogen-depleted. Indeed, recent data from our group have shown that the effect of oestrogens on endothelial function is more dependent on time since menopause than on the actual age of PMW [23]. Overall, oestrogens induce systemic vasodilation [6,9], improve endothelial function of coronary and peripheral arteries of PMW, prevent coronary arterial spasm, reduce endothelin release in women with and without coronary atherosclerosis and affect the vascular response to injury and the development of atherosclerosis [9,10,24–26].

Oestrogens influence vascular function also in men. Previous investigations have shown that the presence of defective mutations in oestrogen synthesis or receptor expression is associated with reduced endothelial function and with premature development of atherosclerosis [27,28]. Oestrogen levels in men are mostly dependent on and derived from androgen production. Nearly 80% of plasma 17-estradiol in men derives from aromatization of testosterone and androstenedione into oestrogen. Furthermore, estradiol is generated directly in the male vasculature, where it can stimulate ERs of endothelial and vascular smooth muscle cells. Interestingly, in men with aromatase deficiency, endothelial dysfunction and diffuse atherosclerosis, low-dose oestrogens improve endothelial function [29,30]. Moreover, normal males treated with aromatase inhibitors develop reversible endothelial dysfunction [5,31].

Sex hormones may also influence vasomotor tone by modulating the response to different vasoactive substances, such as norepinephrine, angiotensin II (AII) or aldosterone [32]. Indeed, norepinephrine is known to induce less vasoconstriction in women than in men.

Testosterone-mediated effects

The vascular effects of testosterone are yet poorly understood [33,34]. The role of endothelium-derived NO in testosterone-induced vasodilation is not clear, as some studies using endothelium-denuded tissues and inhibitors of NO synthesis suggest a partial contribution, whereas others report no role for NO in testosterone-induced vasodilation [33,35–37]. Testosterone-induced, endothelium-independent relaxation is not gender-related nor mediated by hormonal receptors and seems to involve the opening of K+ channels and the blockade of the L-type calcium channel [33,38,39].

Although animal studies have been inconclusive, human investigations suggest that testosterone may exert a positive influence on vascular reactivity. Impaired endothelium-dependent vasodilation because of reduced NO availability has been documented in hypogonadal patients. Other studies have reported that increased endothelium-dependent dilation is significantly associated with low serum testosterone levels, and suggested that the withdrawal of male sex hormones may be associated with enhanced endothelial function in adult men [40]. In contrast, intracoronary infusion of testosterone induces relaxation of coronary arteries and augments blood flow in men with and without CAD [41]. Low-dose oral testosterone augments both endothelium-dependent (flow-mediated) and endothelium-independent vascular reactivity in the brachial artery of men with CAD [42]. In these patients, testosterone administration improves peripheral blood flow, enhances myocardial perfusion, and increases coronary vascular flow and vessel diameter at physiological (nanomolar) concentrations [33,41,43,44]. However, in men with low testosterone plasma levels, transdermal application of the hormone does not seem to affect endothelium-dependent dilation [45].

The conflicting results obtained by human studies assessing the effects of testosterone on vascular reactivity may be related to the baseline testosterone plasma levels, the underlying cardiovascular condition, the duration of treatment, patients’ age and type of vascular bed used for the investigation. Furthermore, composition and formulation of the hormones employed might also have affected results.

As for oestrogens, differences in androgen vascular receptors expression have been reported, with greater receptor densities in men, suggesting different, gender-related genomic responses to androgen exposure [5,6,33,46].

While in men the effect of testosterone is not completely understood, the effect of androgen in women is even less clear but seems related to the concomitant level of oestrogenization. In female to male trans-sexuals, receiving long-term testosterone resulting in plasma concentrations similar to those of men, brachial artery diameter is greater and nitrate-induced vasodilation is impaired. However, the pattern of endothelium-dependent relaxation is similar to that of age-matched female controls [47]. In oestrogen-treated PMW receiving parenteral testosterone, and exhibiting five-fold greater than normal concentrations of blood testosterone, improvement in flow-mediated vasodilation was also found [48].

Oestrogen-mediated effects

In cholesterol-fed female nonhuman primates, oophorectomy accelerates the development of atherosclerosis, while treatment with estradiol has an opposite effect. However, the timing of administration onset appears fundamental [49]. In animal models, oestrogens reduce progression of atherosclerosis only if started within a very short time from oophorectomy. No effect or even pro-inflammatory effects are observed, if therapy is initiated late [50,51].

Human findings support these observations. In women, oestrogen loss (after natural or surgical menopause) is associated with greater progression and extent of atherosclerosis. If started early after the menopause oestrogen/hormone replacement therapy (ERT/HRT) reduces atherosclerotic burden [52,53]. In the Women’s Health Initiative (WHI), HRT was associated with fewer cardiovascular events and reduced total mortality if initiated in younger women (50–59 years old) or within 10 years after menopause. In the WHI Coronary Artery Calcium Study, 50–59-year-old women receiving exogenous oestrogens had lesser coronary-artery calcification (a recognized marker of atherosclerosis) on multi-slice CT [54,55].

Therefore, the balance between the favourable and unfavourable vascular effects of exogenous oestrogens is time-dependent. The benefits of HRT in preventing atherosclerosis only occur if hormones are initiated before the development of advanced atherosclerosis, as the disease in women tends to develop in a time-dependent manner after the menopause [6,56,57]. In blood vessels that are normal or only exhibit early atherosclerotic lesions, physiological levels of exogenous oestrogens appear protective, as they improve endothelial function, vascular stiffness and exert anti-inflammatory effects [23,58,59]. Conversely, in advanced atherosclerosis, these hormones appear proinflammatory, and may facilitate plaque instability and thrombosis. In fact, oestrogen deficiency induces methylation of the promoter region of the α ER that becomes less sensitive to oestrogen stimulation [22].

Testosterone-mediated effects

As already pointed out, the effects of testosterone differ in male and females. When given to male mice and rabbits after castration, testosterone induces vascular effects that are similar to those of oestrogens in females [60]. In male rabbits, testosterone, and its combination with estradiol, reduces the progression of atherosclerosis, while in ovariectomized nonhuman primates, exogenous androgens increase atherosclerotic plaque size. These effects are partly prevented if androgens are associated with estrone [61]. Therefore, the estradiol/testosterone ratio, rather than the absolute levels of androgens, plays a key role in modulating the effect of androgens on atherosclerosis in females.

Conversion of testosterone to estradiol by aromatization appears fundamental for the vascular protective effects of this hormone to occur in males. Indeed, like castration, aromatase inhibition causes a substantial increase in fatty streaks [62] and the aromatizable androgen dehydroepiandrosterone (DHEA) inhibits atherosclerosis in intact cholesterol-fed rabbits. By contrast, nonaromatizable androgens have no effects [63].

Consistent with animal experiments, epidemiological studies indicate that low testosterone levels are independently associated with greater cardiovascular mortality in men [64–66]. An inverse correlation has been reported between free testosterone levels and the degree of CAD in male patients undergoing coronary angiography [44,67]. Moreover, lower androgens levels have been observed in men with increased intima–media thickness [68]. Finally, our group has reported an inverse correlation between plasma androgen levels and degree of coronary atherosclerosis and suggested an association between a low androgenic status and the development of CAD in men [69].

The hypothesis that androgens may exert an atherogenic effect in women was inferred from observations obtained in females with a state of relative hyperandrogenism, such as those with the polycystic ovary syndrome (PCOS). The finding that women with PCOS bear an unfavourable cardiovascular risk profile and develop more severe atherosclerosis than matched controls, suggested a causative role for the greater testosterone levels that are typically observed in these subjects. However, despite having a greater than seven-fold increase in predicted CV risk, women with PCOS have actually no excess CVD events or mortality [70]. Moreover, observational studies conducted in postmenopausal populations indicate that age-adjusted concentrations of androgens do not differ in women with or without heart disease and do not predict cardiovascular events and death [71]. The only available investigation assessing the effects of testosterone administration in females was conducted in female-to-male transgenders. In this study, genetic females receiving male testosterone doses, reported no excess CVD during 2418 patient-years of exposure [72].

Lipid metabolism

Sex hormones influence lipid profile by a direct effect on lipoprotein production, degradation and secretion and gender differences exist as regards lipoproteins, as total and LDL-cholesterol (LDL-c) appears more important in determining CVD in men, while high triglycerides, Lp(a) and low HDL-cholesterol (HDL-c) are more significant in women.

Unlike men, women develop a pro-atherogenic lipid profile later in their life, usually after the age of 50 or after the transition from the fertile state to the menopause. This is mostly because of impaired oestrogen production [80–82] as oestrogens reduce hepatic triglyceride lipase (which degrades HDL-c), stimulate synthesis of HDL-c and apolipoprotein A-I and improve the reverse cholesterol transport. As a result, these hormones reduce total and LDL-c, increase HDL-c and decrease LDL particle concentration as well as decrease triglycerides.

Testosterone may influence the lipid metabolism through both intrinsic androgenic effects and through its role as a substrate for aromatization to 17β-estradiol (E2).

High plasma testosterone levels have long been thought to worsen lipoprotein profile and to cause a greater susceptibility to CVD in both sexes. However, data gathered from observational studies in men suggest that plasma testosterone levels are positively correlated with HDL-c serum levels and inversely correlated with triglycerides, LDL and total cholesterol, fibrinogen and PAI-1 [33,83–86]. The possibility that the overall effect on lipid metabolism may be, in part, influenced by the confounding effect on adiposity and insulin resistance has been suggested by some studies [87], but clinical investigations suggested that testosterone therapy in ageing males exert a positive effect on lipid profile, without affecting HDL-c [86].

In women, the effect of testosterone therapy on lipid levels (mainly decrease in HDL-c) seems to depend on the degree of oestrogenization and, consequently, on the estradiol/testosterone ratio. However, an androgen-induced decline in circulating HDL-c should not automatically be assumed to be pro-atherogenic, as it may reflect a specific reduction in the HDL3 subfraction and therefore may not be associated with a significant reduction in the reverse cholesterol transport [88].

Obesity

Male/female differences in body fat distribution are partly related to sex hormones. Although androgens have been long implicated in the development of abdominal fat, it is now clear that the so called ‘android fat distribution’ is associated with a state of androgenic deficiency. Men with low levels of testosterone have an increased prevalence of android obesity, which correlates with an unfavourable cardiovascular risk profile and morbidity [89]. In women, both distribution and amount of body fat are linked with ovarian function. Indeed, while normo-menstruating women more frequently exhibit a gynoid fat distribution with peripheral adiposity, PMW tend to gain weight and to redistribute body fat from a peripheral (gluteus and hips) to a central (abdomen) location [90–92].

In both sexes, increased abdominal circumference or waist-to-hip ratio, both surrogate markers of android fat distribution, are independently associated with negative changes in insulin resistance, serum triglycerides, small dense LDL particles levels, blood pressure and with increased sympathetic drive. Apart from favouring the negative metabolic changes listed earlier, increased visceral fat, which occurs in both sexes in association with sex hormone deficiency, also interferes with hormone metabolism. Indeed, white fat is the major source of oestrogens in elderly women and men, because of the local conversion of androstenedione to estrone by aromatase activity [93,94].

That testosterone is anti-adipogenic is proven by the observation that, in men, testosterone supplementation reduces abdominal fat by stimulating lipolysis and thereby reducing fat storage in adipocytes. Reduction in body fat by testosterone translates into beneficial effects on several metabolic parameters, such as glucose and lipid metabolism [87,95–97]. Accordingly, in women, HRT reduces weight gain and prevents the postmenopausal shift from gynoid to android fat distribution. However, the exact mechanism of this effect is largely unknown.

Glucose metabolism

Although diabetes and altered glucose metabolism significantly increase the risk of CVD in both sexes, the probability of future cardiovascular events is greater in women with diabetes than in men, especially when the disease is associated with increased blood pressure [73,74]. Indeed, the relative risk for fatal CAD associated with diabetes is 50% greater in women [98], and premenopausal women with diabetes have approximately the same risk of men with diabetes of the same age [99].

Women with diabetes have lower oestrogen levels than their non-diabetic peers and, although oestrogen deficiency ‘per se’ does not appear to affect fasting glucose, it has been shown to cause a progressive decline in glucose-stimulated insulin secretion, and to increase insulin resistance by reducing insulin sensitivity. Therefore, it is conceivable that the changes in glucose metabolism that are frequently observed after the menopause are partly related to oestrogen withdrawal, though increased body weight and sympathetic activity, which are typical of this phase of a woman’s life, are probably also important [100].

Studies in men have shown that testosterone improves insulin sensitivity, possibly through an effect on muscle and fat mass [33,34]. Indeed, an association exists in men, between low levels of free and total testosterone and the risk of developing insulin resistance and type 2 diabetes [94]. The opposite occurs in women, where high testosterone levels are associated with a greater risk of acquiring type 2 diabetes [74]. However, as already discussed, hypertestosteronemia is often paralleled by low oestrogen levels and consequently, by low estradiol/testosterone ratios. Therefore, the inherent effect of testosterone on metabolic parameters is difficult to evaluate in women.

Blood pressure

The prevalence and complications of hypertension exhibit consistent gender differences. Blood pressure is higher in men than women since puberty. However, as from middle age, and above all, after cessation of ovarian function, both systolic and diastolic blood pressure start to increase in women, to eventually become greater than in men. It appears therefore that the menopause exerts a negative specific effect on the development of hypertension, independent of age-related changes.

The hypotensive effects of estradiol in women are well recognized. During the luteal phase of the menstrual cycle (when estradiol levels peak), blood pressure is lower than during the follicular phase [101,102]. Sex hormones regulate blood pressure not only through their direct vascular actions but also through their effects on the water/electrolyte homeostasis, on the Renin Angiotensin Aldosterone System (RAAS) and on the autonomic control of the cardiovascular system [102]. It is important to note that the ovary has a distinct RAAS, which is the major extra-renal source of prorenin and AII in young women. The increase in angiotensin-converting-enzyme activity and in salt sensitivity, occurring with oestrogen deficiency after the menopause, suggests a link between ovarian function and RAAS activation. However, oral oestrogens may increase the plasma levels of angiotensinogen and this effect may be exacerbated or fully antagonized by the concomitant administration of progestin with mineralcorticoid or anti-mineralcorticoid effects [103].

In men, the progressive decline of testosterone produces changes in blood pressure indistinguishable from those induced by ageing, and lower testosterone levels are associated with higher blood pressure levels in middle aged men. Nevertheless, acute and chronic administration of testosterone in physiological doses does not significantly affect blood pressure. The mechanisms by which testosterone affects blood pressure are not well understood. Conversion of testosterone into estradiol may have a role, especially in older men, and the integrity of the androgen receptor system appears to protect the vascular bed from AII-induced injury [104].

The metabolic syndrome, a complex condition characterized by overweight and hypertension, and by a cluster of alterations in lipid and glucose metabolism, is known to be associated with greater cardiovascular morbidity and mortality. The changes in CV risk factor profile ensuing from oestrogen and androgen deprivation may partly explain the greater incidence of this syndrome, which has been observed in PMW as well as in men with low testosterone levels.

ERT/HRT

Early observational studies in PMW suggested that HRT might be beneficial in primary cardiovascular prevention and hypothesized a similar effect also in women with overt CAD [105].

However, postanalysis of the results of the Framingham Study clearly showed that HRT is protective in younger women but can be harmful in those starting treatment late, and especially after the age of 60. The observation was confirmed by the Heart and Oestrogen/progestin Replacement Study (HERS), which failed to show significant protective effects of HRT in older PMW with CAD, and again suggested a potential initial harm when treatment was started late, unless concurrent statin therapy was also initiated [106].

More recently, the results of the WHI have added further confusion to the issue of whether or not HRT should be used for cardiovascular prevention [107,108]. The study showed that treatment does not reduce cardiovascular events in asymptomatic PMW, if ERT/HRT is started late (more than 10 years) after the menopause. Conversely, a reduction in total mortality was observed, in women starting therapy within 10 years since menopause. Moreover, ERT appeared to induce a trend towards reduction of coronary events of similar extent to that suggested by the observational studies.

This ‘timing hypothesis’ is supported by the evidence suggesting that besides the positive effect of ERT/HRT started early after the menopause, oestrogen have a greater vascular effect if started early after the menopause and that early PMW have a more favourable lipid profile (cholesterol) than late PMW this positively affecting the effect of oestrogens [23,109].

Although the results of the WHI indicate that time of HRT initiation since menopause is crucial to explain the widely divergent findings of observational and randomized controlled studies [110–112], the authors set a warning on a possible increase of CVD risk. As a result, current indications of ERT/HRT are only allowed for the relief of menopausal symptoms, even though available evidence indicate that the time of initiation and the choice of the right dose and type of oestrogen/progestin combination are both extremely important to maximize the cardioprotective effects [113,114].

Oestrogen replacement therapy has been considered harmful in men since the results of the ‘Coronary Drug Project’. This study showed that high oestrogen doses (eight-fold higher than those used in the WHI) were associated with an increased risk of myocardial infarction and death [115]. About 10 years later, the Framingham Heart Study suggested that men with high estradiol plasma levels run a lower risk of cardiovascular events [116]. However, circulating oestrogens levels in men are significantly lower than androgens. Furthermore, exogenous oestrogens may reduce free testosterone plasma levels and tissue bioavailability, by increasing sex hormone binding globulin, a protein that binds testosterone with a greater affinity than oestrogens. Therefore, the task of exploiting the supposedly protective effects of oestrogens on male vascular tissue would appear more realistically achievable by enhancing tissue oestrogen availability through aromatization of exogenously administered testosterone.

Androgen therapy

Although, in several small studies, therapeutic doses of testosterone reduced the severity and frequency of angina pectoris and improved electrocardiographic signs of myocardial ischaemia [44], the long-term effects of exogenous testosterone on cardiovascular events have not been yet investigated. Concerns regarding the effects of long-term testosterone administration might have contributed to discourage such trials, though life-time testosterone replacement therapy does not seem to significantly increase the risk of prostate cancer [117].

Testosterone has been recently proposed for treating Hypoactive Sexual Desire Disorder in women [118], and a recent meta-analysis has shown that androgen therapy, in isolation or on top of HRT, is safe and is not associated with increased cardiovascular risk [119]. A recent report also suggests that it may improve functional capacity in women with heart failure [120].