Introduction

Since the 1940s, it has been recognized that sex steroids have important roles in the cardiovascular system.1, 2 A number of epidemiological studies have shown that sex differences are apparent in the incidence of atherosclerotic disease. The incidence of cardiovascular diseases (CVDs), such as hypertension and coronary artery disease, is lower in younger women than in men of the same age.3, 4, 5 However, it rises after menopause and, with age, catches up to that among men. These phenomena have been explained by the atheroprotective action of endogenous estrogen and its deprivation in postmenopausal women. In the past 20–30 years, many studies have suggested the efficacy of hormone replacement therapy (HRT) in postmenopausal women for the prevention of CVD and the putative vasoprotective effects of estrogen. However, reports from the Heart and Estrogen-Progestin Replacement Study (HERS)6 and the Women’s Health Initiative (WHI)7 denied the efficacy of estrogen therapy in CVD.

By contrast, the actions of androgens on the cardiovascular system remain unclear. In the process of atherosclerosis, androgens may exert complex effects on vessel walls. Both beneficial and detrimental effects have been reported. For many years, it was widely believed that androgens have unfavorable roles in the development of atherosclerosis. Recently, however, the link between androgen deficiency and atherosclerosis has been demonstrated in a number of studies.8, 9, 10 Various epidemiological and experimental studies have also demonstrated that androgens exert beneficial influences on CVD via the direct and indirect action of androgens on the blood vessels.

As the effects of estrogen on the cardiovascular system have been extensively studied and reviewed,11, 12, 13, 14 we allocated a small portion of our research to estrogen, highlighting recent developments. A larger part of this review focuses on androgens, particularly testosterone, to discuss the biological role of testosterone in vascular physiology and pathology in aging men.

Action of estrogen on the cardiovascular system

Effects of estrogen on cardiovascular risk factors

A number of studies have reported that estrogen therapy in postmenopausal women decreases the serum levels of both total and low-density lipoprotein cholesterol while raising high-density cholesterol and triglycerides, primarily by influencing the expression of hepatic apoprotein genes.11, 15 Also, estrogen inhibits the lipid peroxidation of low-density lipoprotein in vitro and in vivo.16, 17 Furthermore, estrogen can modulate glucose metabolism and prevent other risk factors for CVD, such as obesity (Table 1).18, 19

Table 1 Anti-atherosclerotic effects of estrogen

Direct vascular action of estrogen

Two estrogen receptor (ER) subtypes, ERα and ERβ, have been identified and are expressed in the vasculature, and experimental studies have demonstrated the vasodilator effects of estrogen/ER through their action on the endothelium, smooth muscle and extracellular matrix. Estrogen enhances endothelium-dependent vasorelaxation via increased release of nitric oxide (NO),20, 21, 22 endothelium-derived hyperpolarizing factor23 and PGI224, 25 and decreased production of endothelin-1 (Table 1).26 Several studies have demonstrated that estrogen inhibits calcium influx27, 28 and stimulates calcium efflux29 in vascular smooth muscle cells (VSMCs), leading to endothelium-independent vasodilation. Moreover, estrogen inhibits neointima formation in response to balloon injury30, 31 and perivascular cuff placement.32 Endothelial regeneration,33 inhibition of endothelial apoptosis34 and inhibition of VSMC migration and proliferation32 may account for the inhibitory effects of estrogen on neointima formation. Analyses of knockout mice for ERα and ERβ have provided more information regarding the molecular mechanism of estrogen’s action on the blood vessels.5 Recent progress in nuclear receptor research has also clarified the non-genomic action of estrogen on the vasculature,14 such as the direct interaction of ERα with the regulatory subunit of phosphatidylinositol-3-OH kinase.35

Role of the novel ER G protein coupled receptor 30 (GPR30) in the cardiovascular system

In addition to the two classical ER subtypes, ERα and ERβ, a third membrane-bound and G-protein-coupled ER, GPR30, has been identified in human vascular endothelial cells (ECs) and smooth muscle cells.36, 37, 38 Haas et al.37 reported that G-1, a selective stimulator of GPR30, acutely blocked vasoconstrictor-induced changes in intracellular calcium concentrations and vascular tone, resulting in lowering of blood pressure in normotensive rats. Similar vasodilator effects of GPR30 have been confirmed in other studies.39, 40, 41 It has also been reported that stimulation of GPR30 blocks VSMC proliferation.37, 42

The vasodilator action of G-1 may be mediated by NO-independent40 and NO-dependent37, 39, 40 pathways; the latter involves GPR30-induced endothelial NO synthase (eNOS) phosphorylation.43 Also, G-1 decreases nicotinamide adenine dinucleotide phosphate-stimulated superoxide production by the carotid and intracranial arteries, indicating the scavenging effects of GPR30 on superoxide anions.39 In the heart, G-1 reduces ischemia/reperfusion injury and preserves cardiac function through the phosphatidylinositol 3-kinase/Akt and extracellular signal-regulated kinase pathways and by eNOS phosphorylation.44, 45 Treatment with G-1 for 2 weeks reduced the expression of angiotensin II type 1 receptor and angiotensin-converting enzyme.40 The non-selective ER antagonist ICI 182780 and selective ER modulators, such as tamoxifen and raloxifene, have been shown to act as GPR30 ligands.46 Moreover, both GPR30 and ER are required for estrogen action in some situations, whereas GPR30 can act alone in the absence of ER,46, 47 suggesting a complex network between GPR30 and ER.

HRT and CVD

Observational studies have suggested that HRT decreases the risk of CVD in postmenopausal women.48, 49 However, large-scale, placebo-controlled, randomized trials, such as the HERS6 and the WHI,7 did not confirm the findings of the observational studies. In the WHI, HRT with conjugated equine estrogen plus medroxyprogesterone acetate increased the incidence of CVD instead, particularly in women older than 60 years of age, although women who started HRT soon after menopause tended to have a decreased risk for coronary heart disease.50

Additional data from other studies have supported the concept that the vasoprotective effects of estrogen are evident only when hormone therapy is initiated soon after the onset of menopause and before the development of atherosclerosis. In a meta-analysis of hormone therapy, CVD mortality was lower in younger women on hormone therapy (mean age of 55 years old) than in age-matched controls.51 Women aged 50–59 years who were enrolled in the conjugated equine estrogen trial of the WHI had significantly lower scores for coronary artery calcification 8.7 years after randomization than with placebo.52

Two ongoing clinical trials, the Kronos Early Estrogen Prevention Study53 and the Early Versus Late Intervention Trial with Estradiol Study (available at http://clinicaltrials.gov/ct2/show/NCT00114517; accessed 16 November 2011), were designed to examine the timing, dosage, route and limited duration of administration on patients’ cardiovascular outcomes and to prove the benefits of HRT in atherosclerosis when HRT is initiated soon after menopause. In the near future, these trials will provide additional insight into HRT and cardiovascular health in younger postmenopausal women.

Association of low testosterone levels with CVD

Plasma testosterone levels decrease with aging, and >20% of healthy men older than 60 years of age have testosterone levels below the standard range in young men aged 20–30 years.54, 55 Lower testosterone levels are associated with cognitive dysfunction, muscle weakness, anemia, osteoporosis, mood disturbances and impaired general and sexual health in aging men.56, 57 Recently, many studies have demonstrated the relationship of testosterone with CVD, indicating a consistent inverse relationship between endogenous testosterone and adverse cardiovascular events.

A case–control study among 117 Indian men aged 30–60 years with old myocardial infarction showed that testosterone concentrations were significantly lower in the patients with myocardial infarction than in the control subjects.58 Similar results were reported in men with acute myocardial infarction.59 Cross-sectional results from the Massachusetts Male Aging Study (1709 men aged 40–70 years) showed that serum total and free testosterone levels bear an inverse relationship with CVD, independent of cardiovascular risk factors.60 Recently, epidemiological studies have found that low testosterone levels are a predictor of all-cause and cardiovascular mortality in elderly men.61, 62 These findings were followed by studies investigating the incidence of CVD and testosterone levels.63, 64 According to these observations, endogenous testosterone appears to exert beneficial effects on the cardiovascular system.

Association of low testosterone with surrogate markers of atherosclerosis

The mechanisms underlying the epidemiological associations of low testosterone with CVD are complex and poorly understood. However, it is assumed that endogenous testosterone has physiological effects on the blood vessels and exerts atheroprotective effects. Actually, an increasing body of evidence has shown that low levels of endogenous androgens are associated with atherosclerosis progression in elderly men. Carotid artery intima-media thickness, a common marker of clinical and subclinical atherosclerosis, has been shown to be correlated inversely with testosterone levels.65, 66, 67 Demirbag et al.68 reported a similar finding by examining the intima thickness of the thoracic aorta in older men. Similarly, in the Rotterdam Study population, Hak et al.69 demonstrated that both bioavailable and total testosterone levels were negatively associated with calcified deposits in the abdominal aorta in men older than 55 years of age.

Arterial stiffness, measured as pulse wave velocity or augmentation index, is a predictor of cardiovascular events.70 Yaron et al.71 reported that age- and blood pressure-adjusted pulse wave velocity was significantly higher in hypogonadal men. Similarly, low testosterone levels in male hemodialysis patients were associated with increases in pulse wave velocity and CVD mortality.72 Clinical and preclinical evidence exists linking endothelial dysfunction to androgen deficiency. In 187 Japanese men aged 47±15 (s.d.) years, flow-mediated dilatation of the brachial artery, a reliable marker of endothelial function, was positively correlated with plasma testosterone levels, independent of other atherosclerosis risk factors.73 Comparable results were reported from Europe74 and specifically from Turkey.75

Clinical effects of androgen replacement therapy

As early as the 1940s, Lesser2 demonstrated that testosterone administration alleviates symptoms and ECG abnormalities in men with angina. Subsequent studies have shown that short-term testosterone administration in men with coronary artery disease results in coronary artery dilation and resistance to ischemia. Indeed, testosterone infusion into the coronary arteries induces vasodilation,76 and intravenous administration of testosterone reduces the exercise-induced ischemic response in men with stable angina.77, 78 Furthermore, acute administration of testosterone in men with chronic heart failure reduces peripheral vascular resistance and cardiac afterload, resulting in an increased cardiac index.79 Chronic administration of testosterone also improves functional capacity and symptoms in heart failure patients.80

Several reports have shown that testosterone administration improves arterial stiffness and endothelial vasomotor function in men. Testosterone replacement in hypogonadal men results in acute (48 h) and chronic (3 months) decreases in pulse wave velocity.71 It was also reported that testosterone replacement in men with coronary heart disease and low plasma testosterone decreased radial and aortic augmentation indices.81 Acute intravenous infusion82 and 8-week oral administration of testosterone83 improved flow-mediated vasodilation of the brachial artery.

Testosterone therapy in hypogonadal men with type 2 diabetes mellitus suppressed the production of inflammatory cytokines by circulating monocytes.84 A randomized, placebo-controlled, double-blind trial of 184 men with hypogonadism and metabolic syndrome showed that intramuscular administration of testosterone undecanoate decreased plasma levels of interleukin-1β, tumor necrosis factor-α and C-reactive protein in association with reductions in body mass index and waist circumference, while interleukin-6 and interleukin-10 did not change significantly.85

Taken together, testosterone administration, at least in hypogonadal men, may have a favorable vascular effect, including endothelium-dependent or -independent vasodilation and reduction of arterial stiffness and inflammatory markers. In contrast, the effects of testosterone replacement on the progression of carotid intima-media thickness or other atherosclerotic lesions, as well as on CVD risk,86 are unknown.

Direct effects of testosterone on vascular walls

Risk factors, such as metabolic syndrome, may partly explain the association of low testosterone with CVD. As the relationship between testosterone and metabolic syndrome has been extensively reviewed,87, 88 this section focuses on the direct effects of testosterone on the vascular wall and the underlying molecular mechanism.

As mentioned above, testosterone therapy can improve vascular function and several markers of atherosclerosis in men. Therefore, vascular ECs, VSMCs and macrophages may be targets of androgen’s actions. Indeed, androgen receptor (AR) has been shown to be expressed in these cells.89, 90, 91

Effects of testosterone on animal models of atherosclerosis and neointima formation

It has been demonstrated that the administration of testosterone in castrated male rabbits that were fed a high-cholesterol diet reduced aortic atherosclerosis, largely independent of plasma lipids.92, 93 In addition, neointima formation after coronary balloon injury in swine was increased by castration and was reversed by testosterone replacement.94 Regarding the role of AR, conflicting findings have been reported. Nathan et al.95 demonstrated the inhibitory effects of testosterone on fatty streak formation in castrated low-density lipoprotein receptor-deficient male mice, but the effects of testosterone were abrogated by treatment with an aromatase inhibitor, suggesting that estradiol converted from testosterone had a major role. Conversely, Qiu et al.91 showed that nonaromatizable dihydrotestosterone suppressed atherosclerosis formation in castrated male rabbits, indicating a role for AR. Exaggerated vascular remodeling in AR-deficient mice, in response to angiotensin II infusion, also suggests an important role for AR.96 A recent study by Bourghardt et al.97 may provide a hint in addressing this issue. They administered testosterone in AR-deficient mice with apolipoprotein E-deficient backgrounds and showed that testosterone reduced atherosclerotic lesions, both in AR-deficient and castrated wild-type male mice, but testosterone was less effective in AR-deficient mice, suggesting AR-dependent and -independent mechanisms.

Effects of testosterone on ECs

Several reports have implicated the effects of testosterone on endothelial regeneration. Cai et al.98 demonstrated that testosterone induced time- and dose-dependent proliferation of human aortic ECs via an AR-dependent pathway. In young hypogonadal men, low testosterone levels were associated with a small number of endothelial progenitor cells,99 and testosterone replacement was able to increase the number of progenitor cells.100 The synthesis and release of vasoactive substances by EC may have a role in these effects. Of the substances synthesized by EC, NO is a critical molecule that regulates vascular tone and atherosclerosis, and it is a major target of testosterone. It has been reported that testosterone-induced endothelium-dependent vasodilation is mediated in part by NO.101 We recently demonstrated that testosterone rapidly induces NO production via AR-mediated activation of eNOS in human aortic ECs.89 Furthermore, we showed that AR directly interacts with the p85 subunit of phosphatidylinositol 3-kinase, resulting in phosphorylation/activation of Akt/eNOS signaling. Taking together with our preliminary observation about the involvement of extracellular signal-regulated kinase 1/2 signaling and [Ca2+]i in AR-dependent eNOS activation, quite similar signaling pathways to those for estrogen can be proposed for testosterone (Figure 1), although some of these pathways should be verified in further studies. The genomic action of testosterone in ECs has not been studied extensively.

Figure 1
figure 1

Signal transduction pathways of eNOS activation by estradiol and testosterone in vascular endothelial cells. AR, androgen receptor; E2, estradiol; eNOS, endothelial NO synthase; ERα, estrogen receptor α; ERK, extracellular signal-regulated kinase; NO, nitric oxide; PI3K, phosphatidylinositol 3-kinase; T, testosterone. Dotted curves indicate the plasma membrane. Dotted arrows indicate probable but undetermined pathways.

It has been reported that testosterone increases the number of ECs secreting endothelin-1,102 although its contribution to the modulation of vascular tone and of CVD is unknown. Testosterone at physiological concentrations seems to have a beneficial influence on the hemostatic system through tissue plasminogen activator expression and inhibition of plasminogen activator inhibitor type 1 secretion by human umbilical vein ECs.103

Effects of testosterone on VSMCs

Most of the rapid vasodilator effects of testosterone are endothelium independent and thus are attributable to its action on VSMCs. In particular, vasodilator responses to pharmacological concentrations of testosterone seem to be AR independent. Yue et al.104 reported that the relaxing response of rabbit coronary arteries to testosterone was significantly inhibited by the potassium-channel inhibitor barium chloride but not by the inhibition of NO synthesis or by removal of the endothelium. Several groups have shown that testosterone inhibits the agonist-induced rise of [Ca2+]i in VSMCs, as has been documented for estrogen. Crews and Khalil28 reported that testosterone at supra-physiological doses (10–100 pmol l–1) significantly suppresses the vasoconstriction of porcine coronary artery strips induced by prostaglandin F2α or by KCl, in parallel with the inhibition of Ca2+ entry. Hall et al.105 demonstrated, using the A7r5 VSMC cell line, that testosterone and dihydrotestosterone selectively suppressed Ca2+ entry via L-type Ca2+ channels. Similar results have been reported in different experimental conditions by other groups.106, 107, 108

The involvement of potassium channels in testosterone-induced vasodilatation has also been studied by many researchers.109, 110, 111 Cairrao et al.112 reported that an AR antagonist, flutamide, and an adenosine triphosphate-sensitive potassium-channel inhibitor, glibenclamide, had no influence on the testosterone relaxant effect, whereas a voltage-sensitive potassium-channel inhibitor, 4-aminopyridine, decreased this effect of testosterone. Opening of voltage-sensitive potassium channels induces hyperpolarization of the plasma membrane, which in turn may lead to the closing of L-type Ca2+ channels. These pharmacological studies, most of which used chemical inhibitors, may be strengthened by studies employing molecular-targeting strategies.

Accumulation of VSMCs in damaged vascular layers is a critical process in the development of atherosclerosis and is closely related to hypertension and its complications. Many, but not all, of the previous studies indicated that testosterone might inhibit VSMC growth. Hanke et al.113 reported, using an ex vivo organ culture system, that testosterone at 10–100 ng ml–1 significantly inhibited neointima formation in association with increased expression of AR in endothelium-denuded rabbit aortic rings after 21 days of incubation. Somjen et al.114 demonstrated the dose-dependent inhibitory effects of dihydrotestosterone and membrane-impermeable testosterone on DNA synthesis in cultured VSMCs derived from the human umbilical artery, suggesting a role for membrane AR. The above-mentioned study by Tharp et al.94 showed that the expressions of protein kinase C delta and p27 (kip1) were increased in coronary artery sections of testosterone-treated swine.

Androgen-responsive genes directly regulated by AR in VSMCs have not been determined, except for AR itself. However, we recently found that growth arrest-specific gene 6 was transactivated by testosterone in human VSMCs via binding of AR to the promoter region of the growth arrest-specific gene 6.90 In this study, testosterone inhibited inorganic phosphate-induced VSMC apoptosis, leading to the suppression of VSMC calcification. To further elucidate the mechanism underlying the effects of testosterone on the cardiovascular system, identification of androgen-responsive genes in VSMCs, as well as in ECs, is required in future studies.

Natoli et al.115 investigated, using human aortic VMSCs, and found that testosterone significantly reduced collagen and fibrillin-1 deposition, while it had no effect on elastin. They also found that testosterone increased the expression of matrix metalloproteinase-3, which has an important role in vascular remodeling.

Possible harmful effects of testosterone on blood vessels

Although many studies have shown the beneficial effects of testosterone on the blood vessels, as mentioned above, other studies have suggested that long-term administration of testosterone may elicit harmful effects, especially vasoconstriction via upregulation of thromboxane A2,116 norepinephrine synthesis,117 angiotensin II118 and endothelin-1.102 It has been also reported that testosterone accelerates vascular remodeling119 and stimulates renal prohypertensive processes, including the renin–angiotensin–aldosterone system.120 Recent meta-analyses have revealed that CVD events were not different between testosterone and placebo groups,86, 121 indicating the complexity of testosterone therapy, as was shown for estrogen therapy in women.

Testosterone deficiency and CVD in women

An age-related reduction in circulating levels of androgens occurs in women as well.122 However, it is unclear whether this decline adversely affects vascular health in women. Higher serum testosterone concentrations, within the physiological range, have been associated with lower carotid intima-media thickness,123 suggesting potential protective effects of endogenous testosterone on cardiovascular health in pre- and postmenopausal women. Conversely, it is well known that women with polycystic ovary syndrome, who exhibit high androgen levels, are at a higher risk for CVD. Some studies have reported that high testosterone is associated with an adverse CVD risk factor profile in postmenopausal women, irrespective of polycystic ovary syndrome.3, 124 Polymorphism of the (CAG)n repeat of the AR gene was associated with CVD and risk factor profiles in postmenopausal women.125 Thus far, evidence is lacking for an association of testosterone with CVD events in women, and it is uncertain whether testosterone could be used as a postmenopausal hormone therapy.

Conclusion

In this review, we illustrated the sex hormones’ effects on the cardiovascular system, focusing on the action of testosterone on the blood vessels. Endogenous androgens, as well as estrogen, may display favorable effects on the vasculature, but whether HRT protects aging men and women from CVD is still unknown. Although testosterone administration seems to have diverse or contradictory effects in younger men and women, androgen therapy may provide hope for elderly hypogonadal men. This issue will remain unclear unless clinical trials of testosterone therapy are conducted. Also, progress in basic research on hormonal effects on blood vessels is essential to understanding the role of sex hormones in the development of CVD.