Sex differences in vascular physiology and pathophysiology: estrogen and androgen signaling in health and disease

Abstract

Sex differences between women and men are often overlooked and underappreciated when studying the cardiovascular system. It has been long assumed that men and women are physiologically similar, and this notion has resulted in women being clinically evaluated and treated for cardiovascular pathophysiological complications as men. Currently, there is increased recognition of fundamental sex differences in cardiovascular function, anatomy, cell signaling, and pathophysiology. The National Institutes of Health have enacted guidelines expressly to gain knowledge about ways the sexes differ in both normal function and diseases at the various research levels (molecular, cellular, tissue, and organ system). Greater understanding of these sex differences will be used to steer future directions in the biomedical sciences and translational and clinical research. This review describes sex-based differences in the physiology and pathophysiology of the vasculature, with a special emphasis on sex steroid receptor (estrogen and androgen receptor) signaling and their potential impact on vascular function in health and diseases (e.g., atherosclerosis, hypertension, peripheral artery disease, abdominal aortic aneurysms, cerebral aneurysms, and stroke).

in recent years, greater attention has been placed on the impact of biological sex and hormonal status with regard to predisposition for cardiovascular disease (CVD) and response to therapy. Women generally have a lower risk for developing CVD compared with men of similar age (189), but this protection is lost during menopause (209), suggesting the importance of sex steroid hormone signaling. Although estrogens are viewed as female sex hormones and androgens are viewed as male sex hormones, estrogen and androgen signaling govern a multitude of physiological processes in both women and men. Current biomedical research focuses on the protective role of estrogen in the vasculature, with many fewer studies on androgen signaling. Here, we summarize current scientific literature of both estrogen [estradiol (E₂)] signaling and androgen [testosterone and dihydrotestosterone (DHT)] signaling in the pathophysiology of vascular disorders of public health significance, such as hypertension, atherosclerosis, peripheral artery disease (PAD), aneurysms, migraines, and ischemic stroke.

Estrogen Receptors

Estrogen, also referred to as E₂, is a classical female sex hormone synthesized primarily in the granulosa cells of ovaries in females and Sertoli cells in males. The physiological effects of E₂ are carried out through the activation of various estrogen receptors (ERs), of which there are at least three forms: estrogen receptor-α (ERα), estrogen receptor-β (ERβ) (36, 67), and membrane-bound G protein-coupled estrogen receptor (GPR30/GPER-1) (140, 185).

ERα and ERβ are well-studied nuclear steroid receptors associated with the plasma membrane, cytoplasm, and nucleus in cardiomyocytes, vascular smooth muscle cells (VSMCs), and vascular endothelial cells throughout the mammalian cardiovascular system (177). Both receptor types function as ligand-activated transcription factors and therefore exert long-term genomic effects by modifying gene expression through direct interaction between highly conserved DNA binding domains of nuclear ERs and estrogen response elements (ERE) located near the promoter of target genes (176). Distinct transcriptional activation functions (AF1 and AF2) recruit other regulatory proteins to DNA-bound estrogen receptors and can either operate individually or in unison to further modify gene transcription (176).

GPER-1 is expressed in the endothelium and smooth muscle of the vasculature (97), in intercalated and tubular cells of the kidney, and in high concentrations in the hypothalamic-pituitary-adrenal axis (79). Whether aldosterone is a ligand for GPER-1 remains controversial (79).

Besides genomic signaling, estrogen receptors also elicit cellular responses that occur too rapidly to be attributed to transcriptional regulation. These were originally referred to as “rapid, nongenomic” estrogen signals because they involve ion channels and enzymatic pathways after estrogen receptor activation (175, 178, 250). Estrogen receptors are associated with the cell membrane in various tissues. In some cell types, ERα is found in caveolae (invaginations of the cell membrane) and associates with large protein complexes (Fig. 1) (156). Other signaling factors are also located here and promote signaling cascades after E₂ binds to ERα. In addition, rapid responses mediated by estrogen receptors in the cytosol and membrane can still influence gene expression (207). Therefore, either the term “nonnuclear-initiated steroid signaling” or “membrane-initiated steroid signaling” may be more appropriate (102).

Fig. 1.

Rapid estrogen receptor (ER) and androgen receptor (AR) signaling pathways in vascular endothelial cells (ECs) and vascular smooth muscle cells (VSMCs). ERs and ARs interact with caveolin (Cav)-1 and Src. These signaling complexes facilitate rapid downstream signaling by phosphorylation. ECM, extracellular matrix; E₂, estradiol.

Androgen Receptors

Testosterone, the androgen synthesized in testicular Leydig cells and ovarian theca cells, is converted to the more potent androgen DHT by the enzyme 5α-reductase. Both androgens are ligands for androgen receptors (ARs), transcription factors that mediate androgen signaling (218). Testosterone is also converted to E₂ by the enzyme aromatase, demonstrating its importance as a steroid metabolite in estrogen receptor signaling.

Androgen receptors, like estrogen receptors, belong to the nuclear receptor superfamily (105). Two variants, AR-A and AR-B (281), are found in most tissues, with each showing varying expression levels depending on tissue type. For example, AR-B predominates in most tissue types where both receptor subtypes are present, although not much is known about how they differ in distribution throughout the body (281). Regardless, ARs are expressed in many vascular cell types, such as smooth muscle cells (SMCs), endothelial cells, and blood cells like macrophages and platelets (151).

In the absence of ligands, ARs are localized with heat shock proteins (87), cytoskeletal elements (154), and other chaperones (44) within the cytoplasm. Once bound to testosterone or DHT, androgen receptors undergo conformational changes and act as ligand-inducible transcription factors to regulate genomic actions via interactions with cis-acting androgen response elements (ARE) (124). Various coregulators also play a role in AR transcription, which influence their ligand and DNA binding capabilities (92). “Classical androgen signaling” refers to the genomic effects mediated by activated cytosolic ARs (240). Although androgens mediate male sexual differentiation, androgen signaling is also tied to cell proliferation, differentiation, metabolism, apoptosis, and protein secretion in many different tissues in both women and men (196).

Studies over the past two decades have also demonstrated that androgens can activate nongenomic rapid signaling pathways after binding to membrane-associated or cytosolic ARs (106, 240). This can then trigger the release of intracellular Ca²⁺ and activate various kinases such as MAPK (106), Akt (19), PKA, and PKC (82). Membrane-associated ARs in aortic endothelial cells specifically interact with Src kinase (c-Src) and the caveolae structural protein caveolin-1 (Fig. 1) (155, 299), possibly a necessary step before the initiation of rapid AR signaling originating from the plasma membrane. Other studies have suggested that membrane-associated AR is a G protein-coupled receptor that interacts with Gnα11 (240), but as this was observed in spermatogenic cells (240), it may not completely mirror cardiovascular cell types. The effects of AR rapid signaling mechanisms appear significant in aspects of vascular physiology and pathophysiology as well.

Hypertension

The development and progression of hypertension are undoubtedly multifactorial and may be explained, at least in part, by lifetime patterns of sex hormone signaling. Both clinical and epidemiological studies have confirmed sex differences in that the onset and progression of hypertension, a disorder that often precedes more serious cardiovascular complications (267), are more prevalent in men than women (233). Although in both sexes increased blood pressure often accompanies aging, postmenopausal women experience more rapid age-related elevations in blood pressure compared with men (209, 294). Estrogen has been thought to contribute to the lower prevalence of hypertension in women compared with age-matched men (204).

Estrogen signaling.

Estrogen is synthesized in the central nervous system (CNS) from cholesterols or converted from aromatizable androgens in presynaptic terminals (57). Additionally, estrogens are able to peripherally diffuse into the CNS (12). In men, testosterone is converted to E₂ in the brain via aromatase (235). Estrogen signaling in the brain is especially complex, and estrogen receptor subtype expression is cell type specific (119).

ERα and ERβ are expressed in the CNS in specific nuclei that regulate heart rate and blood pressure (246). They may be involved in regulating the renin-angiotensin system (RAAS), based on findings that renin, angiotensinogen, angiotensin-converting enzymes (ACE1 and ACE2), and angiotensin II receptors [ANG II type 1 and type 2 (AT₁ and AT₂) receptors] localize and act within these same brain regions (138, 174). Various animal studies have supported the hypothesis that E₂ regulates the RAAS within the CNS by showing that E₂ infusion effectively reduced both ANG II- and aldosterone-induced hypertension in male and ovariectomized female rodents (286, 289, 291). When estrogen was injected into the solitary tract nucleus (NTS), rostral ventrolateral medulla (RVLM), and parabrachial nucleus of ovariectomized female and male rats, sympathetic nervous system baroreflex control was enhanced with a subsequent decrease in resting blood pressure (230, 231). In contrast, when E₂ was injected into the paraventricular nucleus (PVN) in male rats, no effect was seen on heart rate or blood pressure (91). Consequently, specific nuclei in the forebrain and brain stem are sites where estrogen signaling interacts with the central RAAS differently, depending on biological sex and hormonal status.

ERα and ERβ regulate blood pressure differently (208, 289, 304). In female mice, when central ERα was activated, it protected against ANG II-induced hypertension (289). In female and male mice, activating central ERβ preserved resting blood pressure by regulating RVLM voltage-gated Ca²⁺ currents (239, 275). In female rats, ERβ knockdown in the PVN and RVLM using siRNA-ERβ increased aldosterone-induced hypertension (292), and ERα knockdown in female rats under the same conditions showed no effect on blood pressure (292). These results further support that, in females and males, specific ER subtypes mediate E₂’s protective effects on blood pressure within different CNS nuclei.

Overactivation of the RAAS is thought to promote hypertension onset and progression. When inappropriately activated, oxidative stress increases along with vascular constriction and remodeling (56). Recent evidence shows that RAAS constituents, such as ANG II and aldosterone, stimulate the production of ROS in the brain by activating NADPH oxidase (284). This may promote hypertension, since aldosterone-induced ROS in brain regions such as the subfornical organ (SFO) and PVN increases both sympathetic nerve activity and blood pressure (287, 288, 303).

Contrarily, there is evidence that female sex hormones protect against hypertension by directly regulating the RAAS in various organ systems (133).

In support of the antihypertensive effects of estrogen signaling, studies have shown that estrogen activity within the SFO can inhibit intracellular ROS production, thereby decreasing ANG II-induced hypertension (292, 293). Furthermore, intracellular ROS production increases after knockdown of ERα or ERβ expression with siRNA (292). These findings point to a novel estrogenic intracellular signaling mechanism through which estrogen receptor activation interferes with ROS production to mitigate hypertension.

Nitric oxide (NO) plays a significant role in the CNS signaling pathways that control blood pressure (17). Neuronal NO synthase (nNOS), which is responsible for NO production, was found at higher concentrations in the SFO and PVN of female mice compared with male mice (290). When nNOS inhibitors were administered, central ANG II’s hypertensive effects were exacerbated in female but not male mice (290). After ANG II infusion, nNOS levels increased in intact female mice but not in ovariectomized female or male mice (290).

Numerous studies have also confirmed that female sex hormones protect against hypertension via direct signaling in the vasculature. In both humans and murine models, estrogen promotes vasodilation (Fig. 2) by stimulating NO production in the endothelium as a means of regulating blood pressure (191). These rapid cellular responses suggest that estrogen’s vascular actions can occur through nongenomic mechanisms carried out by membrane-associated estrogen receptors and cellular signal transduction (Fig. 3).

Fig. 2.

Estrogens and androgens influence vascular tone. Estrogen enhances the effect of vasorelaxation in female and male vessels. The influence of androgens on vascular tone is more complex as androgen signaling has been demonstrated to promote vasorelaxation in clinical and in vitro studies but elevate blood pressure in vivo. NO, nitric oxide.

Fig. 3.

Schematic view of ER and AR signaling in vascular pathophysiology. ERs protect against hypertension and other vascular injury diseases such as atherosclerosis, stroke, cerebral aneurysms, and abdominal aortic aneurysms. AR signaling protects against atherosclerosis and possibly hypertension. However, increased androgen signaling is associated with elevated stroke risk; furthermore, androgen signaling experimentally exacerbates the formation of abdominal aortic aneurysms. GPR30, G protein-coupled estrogen receptor.

When ERα is activated, it promotes vasodilation by rapidly stimulating endothelial NO synthase (eNOS). For example, when eNOS was activated in pulmonary endothelial cells shortly after E₂ treatment, kinase inhibitors and ERα antagonist, but not gene transcription inhibitor, significantly interfered with eNOS stimulation (49), suggesting rapid eNOS activation by a nonnuclear ERα isoform. Furthermore, when extranuclear ERα was stimulated, it activated eNOS after binding with an estrogen conjugate excluded from the cell nucleus (46). Taken together, these findings implicate the necessity of activating membrane-bound ERα (Figs. 2 and 3). However, the antihypertensive potency of this mechanism is largely dependent on the type of artery being studied, since both ERα and eNOS localization and expression levels vary between different vascular beds (227).

The kidneys also show similar antihypertensive effects from estrogen signaling, since NO also participates in renal control of extracellular fluid homeostasis (173). Because females produce more NO than males due to naturally higher estrogen levels, free Na⁺ intake is higher in female versus male animals (205). Yet, clinical studies have demonstrated that the prevalence for salt sensitivity begins to increase once women transition into menopause and estrogen levels decline (280). This is likely because estrogen signaling increases NO bioavailability and decreases the AT₁-to-AT₂ receptor ratio, thereby preserving renal Na⁺ handling (108). Clinical studies have supported this hypothesis, where, in premenopausal women, salt loading during estrogen peaks resulted in a reduced filtration fraction and sustained renal vasodilation (206), whereas in postmenopausal women salt loading resulted in an increased filtration fraction (205).

Estrogen signaling may also contribute to blood pressure through renal-specific regulation of ACE2. This enzyme counteracts ANG II’s vasoconstrictor activity via conversion to ANG-(1–7) (118). Although these findings would seem to suggest that naturally higher amounts of estrogen in females contribute to hypertension via downregulation of renal ACE2, normotensive mice showed that basal ACE2 activity was lower in female compared with male mice (150). ERα is the receptor subtype that mediates estrogen regulation of renal ACE1/ACE2. For example, when ovariectomized apolipoprotein (Apo)E knockout (ApoE^−/−) mice possessing wild-type ERα (AAee) were chronically treated with estrogen for 3 mo, renal ACE mRNA was reduced by 1.7-fold, whereas the same treatment in ovariectomized ApoE^−/− ERα knockout mice (ααee) resulted in a 2.1-fold increase (30). Furthermore, in AAee mice, E₂ treatment decreased renal ACE2 mRNA and AT₁ receptor mRNA, but this was not observed in ααee mice. In AAee mice, E₂ treatment resulted in an 81-fold increase of AT₂ receptor mRNA, which was not observed in ααee mice (30).

Unfortunately, the rodent models used in the previously described studies did not model a hypertensive disorder, which is of note in that estrogen signaling in diseased states seems to regulate enzymes differently. Female rat models of renal hypertension demonstrate that estrogen signaling actually upregulates ACE2 activity in the kidneys to oppose ANG II-induced vasoconstriction. In female Sprague-Dawley rats with renal wrap on a phytoestrogen-free, high-Na⁺ diet (modeling renal hypertension), both ACE2 mRNA levels and protein activity were reduced after ovariectomy. This resulted in renal injury, which was dampened after ANG-(1–7) infusion into ovariectomized rats (123). Moreover, in both intact and ovariectomized female rats given E₂, there was decreased inhibition of ACE2 expression and progression of renal injury was significantly attenuated (123). Therefore, estrogen signaling appears to be important in upregulating ACE2 activity as a protective measure against renal hypertension.

GPER-1 is also found in endothelial cells and SMCs (21) and localizes in both arteries and veins of the human vasculature (99). When activated, GPER-1 has been shown to regulate blood pressure and protect against hypertension (Fig. 3). For example, incubation of porcine epicardial coronary artery rings with the selective GPER-1 agonist G1 promoted vasodilation and dampened endothelin-1-mediated vasoconstriction (180). In addition, GPER-1 agonists administered to both female and male rats resulted in endothelium-dependent vasodilation and reduced superoxide levels (31). Studies in knockout GPER-1 animals showed no effect of chronic GPER-1 deficiency on NO bioavailability but did show an increased VSMC response to endothelium-derived vasoconstrictor prostanoids (179). These findings point to a novel, NO-independent role of GPER-1 in regulating blood pressure and control of vascular hypertension (Fig. 3).

The vascular effects of GPER-1 have been studied in various rat models of hypertension, such as the mRen2.Lewis rat (276). When ovariectomized female mRen2.Lewis rats were infused with G1 for 2 wk, they showed reduced blood pressure (146). E₂ showed similar effects as G1 in stimulating this GPER-1-mediated effect (145). In contrast, male mRen2.Lewis rats treated with G1 showed no significant changes in blood pressure. When the same female rat model was infused with G15, a GPER-1 antagonist, vasodilation was significantly inhibited, even with E₂ administration (145). This points to a possible therapeutic application of agonizing GPER-1 in females to combat prostanoid- and ANG II-induced elevations in blood pressure.

In summary, in females, estrogen signaling has been shown to elicit different antihypertensive mechanisms through specific estrogen receptor subtypes, with ERα being the main mediator of NO production in the vascular endothelium and a regulator of renal ACE2 activity. GPER-1 may be primarily responsible for attenuating increased VSMC contraction to cyclooxygenase (COX)-derived prostanoid vasoconstrictors. A major limitation of many of these animal studies is their exclusion of hypertensive male animals since administering exogenous estrogen is likely not a feasible hypertension treatment in males. For males, future research is needed to examine the effects of selective estrogen receptor modulators.

Androgen signaling.

The effects of androgen signaling and hypertension are more complex and not as well understood as estrogen signaling. Studies using different rat models have suggested that androgens may promote vascular diseases such as hypertension, but clinical findings often demonstrated decreased androgen levels in males with cardiovascular-related diseases, including hypertension (162) (Fig. 2). The conflicting results between animal and clinical studies suggest an optimal physiological threshold of circulating androgens required to circumvent the onset of hypertension (Fig. 2).

Several studies have demonstrated that hypoandrogenism may be linked to hypertension in older men. In a nested case control study on 11,606 men aged 40−79 yr, surveyed from 1993 to 1997 and followed up to 2003, endogenous testosterone concentrations were found to be inversely related to cardiovascular disease mortality (127). Another study on older men found that total testosterone levels were inversely associated with systolic blood pressure and mortality risk over the following 20 yr of life after adjusting for preexisting conditions (137). Moreover, a short-term crossover study found that testosterone replacement therapy decreased diastolic blood pressure in obese men (167).

Chronic androgen administration exacerbated hypertension in rat animal studies. In a study using hypoandrogenic, obese male Zucker rats supplemented with exogenous testosterone for 10 wk improved both body weight and lipid profiles but increased blood pressure (63). These results conflict with human studies that have suggested that normal androgen levels are antihypertensive, which may be due to studies often using young instead of aged animals to model hypertension, a chronic disease that usually manifests in older human populations.

Additionally, inherent challenges exist in comparing physiological levels of circulating androgens between rodent models and humans. Androgen studies on animals are often acute due to short animal viability and high experimental costs, and responses may vary between animal strains. Therefore, results from these studies may not completely mirror clinical and epidemiological observations of androgen imbalance in humans. Even so, studies on rodents and cell culture have provided valuable information on the mechanisms of androgen signaling that raise blood pressure.

Conflicting results from animal studies may also be partially due to the genomic effects of androgen signaling in the kidneys. In one study, blood pressure, plasma renin activity, and ANG mRNA levels were higher in intact male rats compared with matched female rats (48). Orchiectomy of young male rats blunted hypertension development and lowered both plasma renin and renal ANG mRNA levels, whereas testosterone replacement restored these levels back to baseline (48). The results from this study indicate that androgen signaling elevates blood pressure and can promote hypertension by upregulating RAAS constituents in the kidneys. This may occur through the genomic effects of “classical” androgen signaling, since some of these studies observed elevated RAAS constituent mRNA levels in intact males. Another study observed that androgen signaling stimulated Na⁺ reabsorption in the proximal tubules of the kidneys (129). This may result from upregulation of intrarenal ANG II and subsequent expression of Na⁺/H⁺ exchanger 3, the primary transporter for Na⁺ reabsorption in the proximal tubule (217). These effects would be expected to shift pressure-natriuresis to increase extracellular volume and blood pressure.

Androgens have also been shown to contribute to ANG II-induced hypertension in male animals by increasing renal inflammation and immune cell infiltration. For example, in a study using male and female Rag-1^–/– mice, which lack both T cells and B cells, systolic blood pressure was similar in both sexes in response to ANG II (215). However, when male T cells were transferred into male Rag-1^–/– mice, significant increases in systolic blood pressure in response to ANG II were seen compared with female mice (215). This study also demonstrated that male mice had more renal lymphocyte infiltration than female mice in both the control and ANG II-treated groups (215). This study highlights an innate difference in the immune response to androgen signaling between males and females and implies that males are more susceptible to renal inflammation and ANG II-induced hypertension.

Very few studies have examined androgen signaling and hypertension in females. However, some studies have suggested that androgens are important in blood pressure regulation just as they are in men. Women with polycystic ovarian syndrome have been shown to exhibit excess androgen levels and are often hypertensive (221). For women with hyperandrogenemia, androgen-induced oxidative stress in the kidneys may account for an increased incidence of hypertension. For instance, previous studies on male rats have demonstrated that androgen signaling upregulates 20-HETE, a vasoconstricting metabolite of arachidonic acid. This increases renal microvascular reactivity, which can then activate NF-κB, increase superoxide production, and deplete bioavailable NO, thereby increasing blood pressure (192, 282). Very few studies have examined this mechanism in female rodents. However, one study demonstrated that chronic DHT administration in female and male rats increased systolic blood pressure and 20-HETE levels in both sexes (244). This suggests that increased blood pressure in women with hyperandrogenemia may be caused by androgen-mediated renal upregulation of vasoconstrictors such as 20-HETE (244). However, the doses of DHT used in this study were significantly higher than serum androgen levels commonly seen in women with hyperandrogenemia. Therefore, more studies that mimic serum androgen levels in women are needed to determine if this is a plausible explanation for increased blood pressure.

While in vivo studies have suggested that androgens contribute to hypertension through renal inflammation, upregulation of RAAS constituents, and modulation of pressure natriuresis, human and in vitro studies have suggested that androgen signaling in endothelial cells and VSMCs of arteries can exert antihypertensive effects (Figs. 2 and 3). For example, testosterone rapidly activates eNOS and increases NO release from vascular endothelial cells through the phosphatidylinositol 3-kinase/Akt signaling pathway (299). These results point to a membrane complex of ARs colocalized with caveolin-1 and c-Src, which mediate this mechanism (299). Endothelium-independent vasodilation has been demonstrated in other studies, where androgen signaling induced VSMC relaxation by inactivating voltage-gated Ca²⁺ channels (101, 188) and activating K⁺ channels (37, 236). This collective evidence demonstrates that direct androgen signaling in the vasculature can also elicit antihypertensive effects.

Further research needs to weigh the beneficial actions of androgen signaling in vascular beds against deleterious effects in the kidneys and other tissues to determine if physiological levels of androgens are truly antihypertensive. In females, the effects of hyperandrogenism need to be determined to elucidate the cellular mechanisms that cause hypertension.

Atherosclerosis

Atherosclerosis is characterized by the buildup of fatty plaques in arteries that can potentially rupture and impede blood flow to the heart and other organs. Vascular inflammation, lipid accumulation, intimal fibrosis, arterial stiffness, and plaque rupture are all key factors in the pathogenesis of atherosclerosis, which can eventually lead to more serious cardiovascular insults such as heart attack and stroke (103). The pathogenesis of atherosclerosis involves interactions between endothelial cells and SMCs of the arterial wall and infiltrating cells such as macrophages (144). These infiltrating cells require cellular adhesion molecules, such as VCAM-1, to bind to the endothelium (60). Significant sex disparities exist in the prevalence of atherosclerotic coronary artery disease (CAD). Compared with women, men have a significantly higher risk of developing CAD (47, 263) and dying from it at younger ages (189). Greater levels of adhesion molecules are expressed in the vascular tissue of males compared with females, which may contribute to greater male atherosclerotic predisposition (65, 172). However, the chances of women developing atherosclerosis greatly increase postmenopause (269), which suggests a role for steroid hormone receptor signaling in the sex differences for atherosclerosis.

Estrogen signaling.

The underlying physiological actions of estrogen are important determinants of cardiovascular physiology and development of vascular disease (271). ERα is considered to be the estrogen receptor subtype primarily responsible for mediating estrogen signaling and protecting against vascular injury, fibrosis, and atherosclerosis (114, 160). Elevated expression of ERα in the vasculature of premenopausal women correlates with low atherosclerosis incidence rates (153). However, other estrogen receptor subtypes may mediate protection against atherosclerosis as well. Platelets, which are heavily involved in thrombosis, express ERβ and AR and also respond to sex hormone signaling (90). In addition, the role of GPER-1 in the context of atherosclerosis has also been reported (45).

The vascular calcification process involved in CAD is lipid sensitive (62), and premenopausal female protection from atherosclerosis may be partly due to the beneficial effects of estrogen on lipid metabolism. High circulating levels of estrogens in females correlate with increased high-density lipoprotein (HDL) and decreased low-density lipoprotein (LDL) cholesterol levels, which have an overall beneficial effect on the lipid profile (131). Consistent with these findings, ovariectomy worsened lipid profiles by increasing LDL and reducing HDL cholesterol levels in female rat models of atherosclerosis (131, 203).

Estrogen exerts effects on LDLs and their receptors (LDL-Rs) to improve the lipid profile. Thus, the lack of LDL binding to its receptor results in hypercholesterolemia and subsequently increases the risk for atherosclerosis (93). E₂ enhances LDL-R expression and stimulates sterol-27-hydroxylase activity, which inhibits LDL production (55, 258), demonstrating that estrogen signaling regulates both the LDL ligand and its receptor in hindering preliminary atherogenesis. In addition, estrogen’s phenolic structure is effective at preventing the oxidation of LDL cholesterol (252, 273). E₂’s ability to regulate LDL is also effective for preventing disruption of the endothelial barrier initiated by vascular accumulation of minimally modified LDL (86).

Estrogens are also able to regulate lipoprotein activity and structure during atherosclerosis (184). For example, estrogen has been shown to inhibit macrophage uptake of acetylated LDL cholesterol, which implies that estrogen signaling may affect scavenger receptor function (253). Clinical studies on postmenopausal estrogen replacement therapy have demonstrated that estrogen upregulates ApoAI mRNA levels in the liver and also stimulates ApoE production (274). Furthermore, estrogen signaling facilitates reverse cholesterol transport by stimulating the production of the ABCA1 transporter (247). On the other hand, estrogen signaling hinders expression of scavenger receptor B1 (SR-B1), which prolongs the duration of circulating HDLs and improves the lipid profile (248). Taken together, this evidence also demonstrates mechanisms by which estrogen signaling can hinder initial plaque formation (Fig. 3).

Inflammation plays a large role in the pathogenesis of atherosclerosis (77), and multiple studies using cell cultures have demonstrated anti-inflammatory effects of estrogen signaling on immune cells. Indeed, many of the infiltrates found in atherosclerotic plaques, such as macrophages, T cells, B cells, and mast cells express sex hormone receptors (90, 269). This denotes a possible influence of estrogen on the complex inflammatory components of CAD pathogenesis. Macrophages are the primary infiltrates found in atherosclerotic plaques (54), and the general consensus is that estrogen exerts anti-inflammatory effects on macrophages.

For example, E₂ has been shown to reduce lipoprotein oxidation in healthy macrophages and push them toward the anti-inflammatory M2 phenotype (202) by inhibiting oxidation of LDL (224) and increasing ApoE expression (277). In vitro studies have also demonstrated that estrogens inhibit endothelial activation and adhesion of neutrophils and monocytes to the endothelium. For example, E₂ and various estrogen metabolites dampen expression of adhesion molecules (e.g., CD40, CD40L, E-selectin, P-selectin, VCAM-1, and ICAM-1) (Fig. 4) after being exposed to various factors that promote atherogenesis (88, 220, 243, 260). One possible mechanism behind this effect involves estrogen-mediated inhibition of NADPH oxidase activity, which decreases ROS production (272). Additionally, estrogen signaling blunts monocytic expression of factors such as integrins that facilitate binding to VCAM-1 and ICAM-1 (220). Taken together, this evidence indicates that estrogen signaling can protect against atherosclerotic plaque formation by pushing immune cells toward an anti-inflammatory phenotype and preventing their adhesion to the vascular endothelium.

Fig. 4.

Role of ER and AR signaling on adhesion molecule expression in ECs. ER signaling decreases the expression of adhesion molecules in ECs after exposure to atherogenic-promoting factors. However, VCAM-1 expression is stimulated by male-specific AR in ECs. Thus, male vessels may be more susceptible to inflammatory cell infiltration in atherosclerosis.

ERα was previously thought to mediate most of the atheroprotective effects of estrogen signaling, since early experiments demonstrated that ERα knockout in female ApoE^−/− mice was detrimental to atherosclerotic lesion size and increased serum cholesterol compared with controls, even when they were administered exogenous E₂ (114). However, studies in later years have demonstrated that estrogens are still able to exert antiatherogenic effects in females with full-length ERα knockout on high-cholesterol diets (270). In male rodents, ERα deletion is actually protective against atherosclerosis in some instances (268). This suggests that either different sex hormone receptors or isoforms of ERα are able to mediate estrogen’s antiatherogenic effects.

Recent studies using mouse models of atherosclerosis have indicated that GPER-1 activation may aid in the prevention of disease by exerting beneficial effects on lipid metabolism. As previously mentioned, dyslipidemia is a strong component of coronary atherosclerosis (20), and unfavorable lipid profiles, such as increases in triglycerides and decreases in HDL cholesterol, have been observed in studies using female GPER1-deficient mice (237). A study using female mice under atherogenic conditions found that genetic and pharmacologic knockout of GPER-1 resulted in a multitude of proatherogenic outcomes, such as increased LDL cholesterol levels, elevated inflammation from macrophage and T cell activation, and a reduction in bioavailable NO production (182). Although both human endothelial cells and VSMCs displayed intracellular expression patterns of GPER-1, this receptor specifically mediated eNOS activity in the vascular endothelium (182). This evidence not only confirms an atheroprotective role of GPER-1 in animals but in human vascular tissue as well.

GPER-1 activation reduces other major components of atherosclerotic pathogenesis such as VSMC proliferation and vascular inflammation. For example, activation of GPER-1 inhibits in vitro proliferation of cultured human VSMCs (98, 142), which is a key step in atherosclerotic plaque formation (20). Additionally, a genetic knockout study (237) on GPER-1 in aged male mice demonstrated that its absence is positively correlated with a proinflammatory state. Next, a study (182) in ovariectomized mice demonstrated that when the GPER-1 agonist G1 was administered, there was significantly reduced atherosclerosis and vascular inflammation. Furthermore, GPER-1’s anti-inflammatory effects have been affirmed through other in vitro studies using endothelial cells and immune cells such as macrophages (45, 223). A molecular genetic study (9) has demonstrated that GPER-1 can be cloned from human lymphoblast cDNA libraries. In another study (223), GPER-1 activation in macrophages resulted in decreased expression of Toll-like receptor-4. The biochemical mechanisms underlying GPER-1 modulation of inflammatory processes are not fully understood yet. However, results from some investigations suggest that the forkhead box P3 protein (FOXOp3) pathway, which is tied to immune cell secretion of IL-10, is involved (33, 34).

Thus far, the body of evidence from preclinical studies on GPER-1 has shown that activation of this receptor improves the lipid profile and dampens the inflammatory component of atherosclerotic pathogenesis. Therefore, GPER-1 has the potential to be a therapeutic target for preventing and treating CAD in men and postmenopausal women (181).

In summary, evidence shows that estrogen signaling can hinder the formation of initial fatty streaks germane to atherosclerosis by improving lipid profiles and dampening inflammation through the ER subtypes. Nevertheless, atherosclerosis is a multifactorial disorder that involves many cell types of both the vascular architecture and circulating blood. Although studies have identified some of the protective effects of ER activation in different cell types, more research is needed to fully understand the signaling pathways that estrogen receptors elicit in the context of atherosclerosis. Furthermore, few in vivo studies have examined the effects of estrogen signaling in male animals, and it is likely that males respond differently to estrogen signaling. Since administering exogenous E₂ to men is considered therapeutically impractical, estrogen receptor knockout mouse models and selective estrogen receptor modulators should be used in future research studies to delineate the mechanisms of vascular estrogen signaling in males.

Androgen signaling.

The effects of androgen signaling in the context of atherosclerosis still remain poorly understood. However, there is a general consensus that normal physiological levels of endogenous testosterone are atheroprotective. In fact, low androgen levels in men are associated with accelerated atherosclerosis. For example, clinical studies have shown that markers of atherosclerosis, such as intima-media thickness, are inversely correlated to endogenous androgen levels in men (161, 190, 266). Furthermore, evidence from the Rotterdam study population confirmed an inverse relationship between total bioavailable testosterone and the extent of abdominal aortic calcified deposits in men over 55 yr old (66). Other human studies have demonstrated that androgen deficiency is tied to endothelial dysfunction in men. For instance, clinical data demonstrate that flow-mediated brachial artery dilation in middle-age men was positively associated with circulating testosterone levels (7).

Some clinical studies have linked hyperandrogenemia in women to an increased risk of developing subclinical atherosclerosis. For example, the Coronary Artery Risk Development in Young Adults Women’s Study found that women with polycystic ovarian syndrome that manifested as both anovulation and hyperandrogenism had a 2.70 odds ratio for coronary artery calcification (38). Therefore, clinically, normal physiological levels of endogenous androgens may be protective against atherosclerosis (Fig. 3), since increased risk is associated with hypoandrogenism in men and hyperandrogenism in women.

Biomedical research on atherosclerosis using rabbits and rodents has proven to be difficult since some important characteristics of human atherogenesis are not recapitulated in animal models. Lipid metabolism in rabbits mirrors human physiology more closely than that of rodents, and, as a result, several studies have used rabbit models of atherosclerosis rather than mice to better understand androgen signaling in the context of human atherosclerosis. Nevertheless, research studies with both rabbit and rodent models have confirmed atheroprotective effects of normal androgen levels in males, but the underlying biochemical mechanisms still remain largely unclear.

In studies using male rabbits sustained on a high-cholesterol diet, reduced atherosclerosis in the aortic vessel as a result of exogenous testosterone administration was seen (9, 32). These effects were largely independent of plasma lipid levels (9). Another study using castrated LDL-R-deficient male mice demonstrated that the inhibitory effects of testosterone on fatty streak development were diminished subsequent to treatment with aromatase inhibitors (195). This suggests that atheroprotection from testosterone may be partially attributed to its conversion to E₂ by aromatase (195). On the contrary, some studies using castrated male rabbits have demonstrated that DHT, a nonaromatizable androgen, is also able to hinder progression of atherosclerosis by suppressing foam cell formation (216). These results point to a protective mechanism involving AR activation.

As mentioned previously, ARs are expressed in all cell types of the vasculature (132, 151). However, sex differences exist in AR concentrations expressed in vascular tissues. For example, in males, higher AR expression was seen in rat VSMCs (109), human mesenteric artery cells (65), and macrophages (171). Androgens may exert atheroprotective effects in males through multiple mechanisms, such as AR ligand activation (Fig. 3), conversion to E₂ via aromatase, or other unknown pathways. Some investigations have directly verified the atheroprotective effects of androgens through both AR-dependent and AR-independent mechanisms.

For example, administration of exogenous testosterone decreased atherosclerotic lesions in both AR knockout and wild-type mice that underwent gonadectomy (29). However, the atherosclerotic lesion sizes in AR knockout mice were still significantly larger than those seen in wild-type mice (29) and there were no traces of aromatase mRNA in the femoral artery or liver of the animals after quantitative PCR analysis (29), which ruled out protective mechanisms brought on by conversion of testosterone to estrogen. This clearly demonstrates that androgens are able to hinder atherosclerosis through mechanisms of both AR-dependent and AR-independent activation.

On the other hand, stimulation with DHT in cultured human umbilical vein endothelial cells from male donors results in increased expression of the VCAM-1 gene and increased levels of the cell surface protein (171). These results are abrogated in cells of female origin as a result of lower AR presence and inhibition of VCAM-1 mRNA and subsequent protein expression by E₂ (171). This androgenic stimulation of VCAM-1 in endothelial cells, which is amplified by male-specific AR levels (Fig. 4), may imply that males are more susceptible to immune cell infiltration in the pathogenesis of atherosclerosis.

In summary, normal physiological levels of androgens appear to be atheroprotective in both females and males. However, the cell signaling pathways and mechanisms underlying this androgen-mediated protection are poorly understood. Since current evidence suggests that androgen signaling affects atherosclerosis through multiple AR-dependent and AR-independent pathways, more in vitro studies are needed to better understand the different mechanisms behind androgen-mediated atheroprotection. Furthermore, there is very little research on androgen signaling and vascular disorders in females, creating a unique opportunity for future studies into these areas of vascular research. More research in female rodent models on androgen signaling and atherosclerosis is crucial for developing better therapies.

No animal model perfectly mirrors atherosclerotic development in humans so limitations exist whenever animals are used for translational research. Inducing atherosclerosis in rodents with short lifespans often requires special diets that result in serum cholesterol levels that greatly surpass those found in human disease states. Using these research models inherently limits our ability to translate what often takes decades to develop in humans.

It is important to note that wild-type mice are resistant to developing atherosclerotic lesions. Therefore, it is appropriate and often necessary to use both genetic and dietary modifications for murine models of atherosclerosis. In the animal studies discussed thus far, LDL-R^−/− and ApoE^−/− mice were used; they are well established and the most commonly used murine models of atherosclerosis. In fact, ApoE^−/− mice develop atherosclerosis even under normal dietary conditions. One clear advantage of this mouse model is that it develops atherosclerotic lesions in sites similar to humans such as the aortic root, aortic arch, carotid, and coronary arteries (24, 116). However, the model lacks the frequency of plaque rupture and thrombosis that often accompanies atherosclerosis in humans. One potential strategy to overcome this limitation in future research on sex differences in atherosclerosis would be to cross ApoE^−/− mice with mice deficient for SR-B1, as these hybrid mice would likely exhibit rapid coronary artery occlusion, myocardial infarction, and other characteristics that closely resemble advanced human CAD (262).

Observed atherosclerotic plaques in rabbits are highly similar to humans in terms of inflammation (1). Histological analysis of aortic plaques in rabbits reports macrophage accumulation, thin fibrous caps, and lipid-rich cores. These features are also characteristic to humans (1). Still, rabbit models of atherosclerosis are often not inbred or as easily genetically manipulated compared with mice, leading to more variations in response to high-cholesterol diets. Note that larger animals, such as pigs, are known to have cardiovascular anatomy and physiology quite comparable with humans, so this may be a model to consider. Moreover, studies on sex differences in atherosclerosis may benefit from the use of larger mammals.

PAD

PAD is a progressive circulatory disorder characterized by artery and vein blockage not in close proximity to the heart. PAD most often occurs in blood vessels of the legs and feet, and blockage can arise from thrombosis or narrowing of the arteries due to atherosclerosis. Much like CAD, PAD is more prevalent in men than women at younger ages, but the incidence rises in women after menopause (200). In addition, women generally display more severe PAD compared with men at the time of clinical diagnosis and experience greater complications (122). A point-prevalence study of 5,080 subjects in Sweden found that women had a higher prevalence of both asymptomatic PAD (12.6% vs. 9.4%) and PAD with limb ischemia (1.5% vs 0.8%) compared with men at clinical diagnosis (242).

Epidemiological studies have also demonstrated associations between sex hormone levels and PAD. For example, a cross-sectional investigation on 3,034 middle-aged participants from the Framingham Heart Study found that lower total testosterone and sex hormone-binding globulin concentrations in men were associated with prevalent PAD, but none of the analyses on female participants yielded significant results (104). Another investigation on 921 elderly participants from the InCHIANTI study found that low sex hormone-binding globulin levels in men and high testosterone levels in women were significantly associated with PAD, although E₂ was not significantly associated with PAD for either sex (159). These results agree with other epidemiological studies that demonstrated an inverse relationship between serum androgen levels and PAD symptoms in older men (295). Collectively, these studies suggest that androgens may protect men but not women from PAD.

Estrogen signaling.

Although many epidemiological and clinical studies have not found significant associations between serum estrogen levels and PAD, evidence from animal studies has shown that estrogen signaling is antithrombotic and can protect against PAD. For example, Abu-Fanne et al. (3) treated female ovariectomized mice with E₂, raloxifene, or placebo for 4 wk before inducing thrombosis and vascular occlusion. Their results demonstrated that ovariectomy significantly amplified thrombosis, whereas E₂ or raloxifene treatment was antithrombotic in both ovariectomized and intact female mice (3). The antithrombotic effects of both treatments were attributed to suppression of platelet adhesion and upregulation of the COX-2 pathway (3).

Estrogen signaling facilitates neovascularization and recovery from ischemic PAD. From one study investigating the effects of ovariectomy in female BALB/c mice followed by surgically induced hindlimb ischemia (HLI), ovariectomized mice had low plasma levels of E₂ coupled with a reduced blood perfusion index and capillary density compared with controls at 21 days postinduction of HLI (169). Furthermore, ovariectomized mice exhibited suppressed eNOS expression as early as 3 days postinduction of HLI (169). This implies that eNOS is an essential downstream target of the estrogen signaling pathway that promotes angiogenesis.

Proangiogenic effects of estrogen have been observed in male animals as well. In an investigation using a male mouse model of HLI, exogenous estradiol administration increased mobilization and incorporation of endothelial progenitor cells (EPCs) into sites of neovascularization in a dose-dependent manner (228). There was improved blood flow recovery and decreased limb necrosis in the E₂-treated group compared with the placebo-treated group (228). E₂-induced neovascularization was also attributed to eNOS stimulation (120) and subsequent matrix metalloproteinase (MMP)-9 activation in the bone marrow (121).

Androgen signaling.

Studies on animals and cell culture have indicated that androgens may aid in recovery from PAD-induced ischemic stress by promoting angiogenesis. For example, Yoshida et al. (297) demonstrated that AR knockout in both male and female murine models of HLI impaired angiogenesis, which resulted in increased skeletal muscle apoptosis, paired with an increased incidence of autoamputation in both sexes. AR knockout reduced phosphorylation of Akt and eNOS even though there was an increase in VEGF expression in the ischemic limb tissue. Furthermore, siRNA knockdown of AR in cultured vascular endothelial cells significantly decreased VEGF-mediated phosphorylation of Akt and stimulation of eNOS, thereby demonstrating cross-talk between androgen signaling and VEGF signaling pathways (297).

Although androgen signaling was found to be angiogenic in both males and females in the study, prior investigations found only angiogenic potency of androgen signaling in males. For instance, Sieveking et al. (241) demonstrated that in hypoxic conditions, castration in male mouse models of HLI blunted angiogenesis. Furthermore, administration of DHT in vitro augmented male, but not female, endothelial cell migration, proliferation, and tubulogenesis in a dose-dependent manner (241). These proangiogenic effects of androgen signaling in males were also dependent on increased expression of both VEGF and one of its receptors, KDR, which mediates most of the angiogenic actions of VEGF in the endothelium (241). In support of these findings, castration of male mice significantly blunted neovascularization after HLI was induced (241).

In summary, in males, endogenous androgen signaling facilitates neovascularization by activating the Akt signaling pathway and upregulating expression of VEGF and its receptors in endothelial cells (241, 297). Therefore, in males, androgens may be therapeutic in circumventing hypoxic conditions with occlusive PAD. However, there is conflicting evidence on the angiogenic potential of androgens in females. These differences may stem from using varying rodent models, since Sieveking et al. (241) used ovariectomized mice while Yoshida et al. (297) used AR knockout mice for their in vivo experiments. Estrogen signaling, however, has been shown to be angiogenic in both sexes by increasing both EPC mobility and incorporation into neovascularization sites via eNOS stimulation and MMP-9 activity in bone marrow (120, 121, 169, 228). Although PAD’s clinical aspects are well defined, greater emphasis needs to be placed on delineating the mechanisms and impact of AR signaling on PAD in females.

Abdominal Aortic Aneurysms

Aneurysms are characterized by the localized yet permanent dilation and weakening of the arterial wall. Both cerebral and abdominal aortic aneurysms (AAA) share multiple risk factors (e.g., smoking, atherosclerosis, and high blood pressure) (8, 35) but show sexual dimorphisms in their prevalence and age of onset (23, 64).

AAA is a localized ballooning of the abdominal aorta (113). Epidemiological studies of Western populations have confirmed that AAA is much more prevalent in men than women. For example, the Rotterdam Study, a large cross-sectional study on 5283 individuals in The Netherlands, found that AAA prevalence was 4.1% in men compared with 0.7% in women (214). The underlying causes of males being predisposed are still not completely clear because of the disease’s complex pathogenesis. However, key pathological characteristics of AAA development include the degradation of collagen and elastin in the vessel media and adventitia, increased release of chemokines and cytokines, and infiltration of immune cells into the aortic wall (13, 40).

Estrogen signaling.

High levels of circulating female sex hormones have been shown to protect against experimental AAA development by reducing inflammation and matrix protease activity in aortic tissue. Aortas of female rats in an elastase model of AAA displayed less MMP-9, fewer infiltrating macrophages, and overall decreased medial wall destruction compared with matched male rats (6). Furthermore, protection against AAA was abolished after female rat aortas were transplanted into male rats, whereas female rat aortas transplanted into female mouse recipients remained protected from developing AAA (6).

Similar results have been observed in other animal studies. Plasminogen activator inhibitor (PAI)-1 is involved in inhibiting plasmin and subsequent MMP-9 and MMP-2 production, as these MMPs contribute to the breakdown of the aortic wall. Thus, higher levels of PAI-1 in female animals show protection against AAA development (68). Similarly, another study using an elastase model of AAA found that rodents with higher levels of estrogen exhibited lower MMP-9 and MMP-2 levels and decreased immune cell infiltration in AAA tissue compared with rodents with lower estrogen levels due to ovariectomy (283). Significantly slower aneurysm dilatation rates in the rodent group with higher estrogen levels suggested that estrogen blunted AAA progression by reducing immune cell infiltration into the aortic wall while simultaneously dampening vascular tissue production of MMP-9 and MMP-2 (283).

Sex differences in MMP activity may originate in the phosphatidylinositol 3-kinase/Akt signaling pathway (89). For example, male rat aortic SMCs treated with elastase exhibit higher levels of phosphorylated Akt/total Akt than female cells (89). This translates to elevated pro-MMP-9 and active MMP-2 in male rat aortic SMCs. In addition, immunohistochemistry of male and female AAA patient tissue revealed increased p308/total Akt ratios in men compared with women (89). This study highlights the differences in cell signaling pathways that could explain sex differences in MMP activity and male predisposition to developing AAA.

NO is known to promote vasodilation and inhibit VSMC proliferation and leukocyte adhesion (130, 168). Increased NO production from estrogen signaling is also thought to protect against AAA; the incidence in ovariectomized NOS2^−/− female mice is 80% compared with wild-type ovariectomized mice at 40% (139).

ERα may mediate the protective effects of estrogen signaling for AAAs (Fig. 3). In a study using the elastase perfusion mouse model, smaller and less frequent aneurysms were observed in female mice. Mouse aortas from female mice showed increased ERα mRNA and protein levels, which correlated with lower MMP activity (136). Also, ERα levels were 80% higher in samples from female patients with AAA compared with men (136). These results indicate that ERα protects against AAA. More research on the downstream signaling mechanisms of estrogen receptor activation in vascular cells are needed to better understand the influence on AAA.

Androgen signaling.

Studies on male animals have shown that higher levels of androgen signaling exacerbate AAA formation (Fig. 3). Androgen signaling upregulates expression of RAAS constituents (80), and many studies use ANG II infusion in mice with hyperlipidemia to model AAA. In studies using ApoE^−/− and LDL-R^−/− mice infused with ANG II, a significantly greater incidence of AAAs was observed in male mice compared with female mice (163). Similar results were reported by Henriques et al. (107), where ANG II-induced atherosclerosis was similar between male and female mice, although male mice displayed a higher incidence of AAA in addition to greater aneurysm size when compared with matched female mice. Castration of male mice reduced ANG II-induced AAA formation to that of intact female mice but exacerbated atherosclerosis (107). Ovariectomy in female mice did not significantly affect the incidence of ANG II-induced AAA compared with female mice that underwent sham operations.

Thus, while androgens are primarily responsible for the sex differences in ANG II-induced AAA, the data also suggest that androgen signaling contributes to male AAA susceptibility through mechanisms outside of the RAAS. Henriques et al. (107) quantified plasma renin concentrations in the ANG II mouse model and reported that castration did not significantly change plasma renin concentration despite decreased AAA incidence but did result in an observable decrease in ANG II receptor density in the spleen. However, the decreased receptor densities in the spleen were similar in both ANG II-infused sham and castrated mice despite the significant differences in AAA incidence observed between the two groups (107). Still, these results come from analyzing spleen tissue and androgen signaling may have differential effects on RAAS constituents via direct signaling in the vasculature. Further studies are still needed to confirm any effects of androgen signaling on RAAS constituents in this context.

Other in vivo studies have suggested that androgen signaling may predispose males to AAA by upregulating MMP activity. For instance, castration of male ApoE-deficient mice blunted the progression of ANG II-induced AAAs (301). Removal of endogenous androgens halted the expansion of the vascular lumen but had no effect on the external dimensions of aneurysms compared with male mice that underwent sham operations (301). Halting of aortic lumen expansion postcastration might be mediated by upregulation of transforming growth factor-β and PAI-1 gene expression in aortic SMCs (301). A different study found that, in female mice, the increased expression of PAI-1 attenuated developing AAA via reduced serum plasmin levels (68). Downstream effects of reduced plasmin in female mice included reduced MMP activity and differential regulation of VSMC apoptosis (68). Thus, it is possible that increased androgen signaling in males may reduce PAI-1 expression and increase MMP activity associated with degradation and remodeling of the vascular wall.

In female rodent models, administration of exogenous testosterone seems to worsen AAA outcome as well. For example, when female neonatal mice were exposed to testosterone, they were no longer protected from developing AAA in adulthood (via increased AT_1a expression in abdominal aortic tissue) (302). Consequently, they suffered long-lasting, predisposing effects of androgen signaling in the context of AAA independent of serum androgens in adulthood (302). Furthermore, testosterone treatment in both female and male rats resulted in similar increased immune cell infiltration and aortic expansion, although male rats experienced more AAA rupture at earlier time points compared with female rats (52).

Although results from animal studies have increased our understanding into the pathogenesis of AAA in humans, the use of current animal models to study a largely multifactorial disease is challenging. In the elastase-induced mouse model, there is delayed medial elastin degradation with aneurysmal development apparent around 14 days postinfusion. Here, the majority of immune infiltrates consists of macrophages. The main limitation of elastase perfusion to model AAA is the involvement of mechanical stress at the aneurysm site during the perfusion process.

In contrast, ANG II infusion in ApoE^−/− mice results in large suprarenal aortic aneurysms (61) by inducing inflammation in the aortic wall through AT₁ receptor signaling (43), which results in accelerated macrophage recruitment and extracellular matrix degradation via MMPs (164, 234). A main advantage of this model is the RAAS involvement to induce aneurysm formation, since evidence already confirms a role of this physiological system in human aneurysms. However, a limitation to the ANG II-induced murine model of AAA is that aneurysms develop in suprarenal locations while humans usually experience infrarenal AAAs.

Large mammalian animal models, such as pigs, may be one strategy for overcoming some of these problems. In fact, there is an already established porcine model of AAA that combines balloon angioplasty and a solution of elastase/collagenase. These characteristics model those similar to humans: gradual expansion of an AAA by degrading elastic fibers, infiltrating immune cells, and VSMC loss (187). Yet, while using pigs may be more ideal for AAA studies, challenges include high costs, special requirements for housing and surgical equipment, and lower expected sample sizes. Nevertheless, large animals may be the most appropriate research models to use when translating findings to human biology.

Cerebral Aneurysms

Epidemiological studies have indicated that postmenopausal women are at greater risk than premenopausal women for subarachnoid hemorrhage (64). Furthermore, in postmenopausal women, hormone replacement therapy with estrogenic components appears to be protective against cerebral aneurysms (186). Thus, estrogen signaling may also protect women from the development and rupture of cerebral aneurysms (64, 186).

Various studies using rodent models of intracranial aneurysms support this hypothesis. In one study where female rats were subjected to common carotid artery ligation, estrogen deficiency coupled with hypertension significantly increased the risk for intracranial aneurysm (257). Here, hypertensive, ovariectomized rats displayed significant decreases in eNOS mRNA in the cerebral vascular wall (257). This resulted in endothelial dysfunction from disrupted NO bioavailability and increased NADPH oxidase subunit expression in cerebral aneurysm tissue (257). ERβ expression was increased while ERα expression was decreased in ovariectomized rats with intracranial aneurysms (257).

ERβ may mediate estrogen-dependent protection from cerebral aneurysms. For example, in an ovariectomized female mouse model of intracranial aneurysm, treatment of exogenous E₂ or a selective ERβ agonist was effective in preventing rupture (256). This indicated that ERβ is the receptor subtype mediating the protective effects of estrogen signaling via increased cerebral vascular NO production (256). Another comparable study yielded similar results when using selective ER agonists and ERβ knockout mice. Here, ovariectomized wild-type mice treated with the ERβ agonist showed a significantly reduced incidence of cerebral aneurysms (255). In contrast, other ovariectomized mouse groups (ERβ knockout, treated with ERα agonist, or given E₂) demonstrated no such reductions (255). Moreover, the protective effects of activated ERβ were eliminated after administration of NOS inhibitors (255). This supports that ERβ is the primary mediator of estrogenic protection from subarachnoid hemorrhage through production of bioavailable NO.

Cerebral aneurysms are more prevalent in females than males, and activating ERβ in cerebral arteries appears to be therapeutic in preventing hemorrhage. However, more studies are needed to thoroughly understand the effects of sex steroid hormone signaling on the pathogenesis of cerebral aneurysms and how this knowledge can be applied to manage this vascular disorder in patients. In particular, further research needs to be carried out on the effects of androgen signaling in cerebral aneurysms.

Migraines

Migraines and other headache-related events predominate in females (249, 251). The American Migraine Prevalence and Prevention study, which collected data on over 160,000 study participants in all age and racial groups, found a significantly higher prevalence of migraines in women compared with men (251). Although women and men reported similarities in headache pain and frequency, women reported more migraine-specific symptoms and greater disability (251). Migraines affect both sexes equally until puberty (147), when prevalence in women begins to rise. However, a reversal is usually seen once women complete menopausal transition (84). Therefore, differences in migraine prevalence between women and men after puberty may be related to the physiological effects of estrogen within the CNS.

Clinical studies mainly rely on observations of normal hormonal fluctuations and changes such as menstruation and menopause (278) as opposed to induced hormonal states using contraceptives or hormone replacement therapy (4, 158, 194). A disadvantage may be the varying pharmacological effects between different contraceptives and hormone replacement therapies, which result in inconsistent hormonal states and levels between study participants.

Sudden drops in estrogen have historically been suspected to trigger migraines (157, 245). For instance, from two studies evaluating the effects of estradiol valerate/dienogest on headaches (158, 194), the frequency of headaches and use of headache medicine were reduced in female subjects (158, 194).

Aura, the perceptual distortions that sometimes precede migraine onset (104a), is associated with higher estrogen concentrations (39). A retrospective study on 23 women demonstrated that extended cycle dosing of 15 μg ethinyl estradiol/0.120 mg etonogestrel contraceptive to maintain estrogen levels below that of a normal menstrual cycle significantly reduced aura and migraine frequency (39). Other studies where exogenous estrogenic hormones decreased migraine severity demonstrated that supplementing exogenous sex hormones can dramatically affect the frequency and severity of headaches and migraines in women.

A few genetic studies have considered the relative effects of estrogen receptor signaling on migraines. Rodriguez-Acevedo et al. (225) examined patient genetics by taking serum samples and surveys from 600 individuals (76 experiencing migraine) living on Norfolk Island, a genetically homogenous community with a high prevalence of migraines. In this study, researchers found an association between migraines and 10 markers related to the ESR1 gene that encodes for ERα (225), demonstrating that heritable differences in estrogen receptor signaling contribute to the prevalence of migraines in some populations.

Many migraine studies have used male animals as models for a predominantly female disorder (26, 211). While no animal model is able to accurately mirror pain and related symptoms that can be accurately measured in human studies, measuring vasospasm or neuronal activation in the different rat models does provide important insights into migraine pathophysiology. One study measured meningeal inflammation according to activation of dural mast cells (25), which promotes intracranial nociceptive activation. Mast cell density in female rats was remarkably higher compared with male rats except in the proestrus phase of peaked estrogen levels (25). Furthermore, dural mast cell density was greater in female rats during the estrus and diestrus phases, which is marked by decreased estrogen levels (25). Ovariectomized female rats exhibited decreased mast cell density, but increased mast cell density after exogenous estradiol was administered. This increase was time dependent, with effects diminishing after 72 h (25).

Another study evaluated estrogen’s effect on migraines using a rat model where nitroglycerin (NTG) was used to induce neuronal activation (96). During the proestrus phase of the estrus cycle where estrogen levels peak (75), female rats exhibited greater NTG-induced neuronal activation in several brain regions, such as the PVN and supraoptic nucleus of the hypothalamus, which resulted in greater pain sensitivity (96). Female rats also exhibited greater NTG-induced neuronal activation compared with intact males, implying that they experienced greater migraines (96). Ovariectomy reduced neuronal activation in the female rat brain, but when exogenous estrogen was administered, neuronal activation levels rose again to those of intact female rats (96).

Studies examining migraine in rats present conflicting results. Boes and Levy (25) found that decreases in migraine markers were associated with the proestrus phase, while Greco et al. (96), using NTG-induced neuronal activation, found that migraine in females was exacerbated during the proestrus phase. That vaginal smears were used to determine rat hormonal states instead of directly measuring serum levels of circulating sex hormones may account for the contradiction. This highlights some of the difficulties in translating animal migraine studies to humans. Unfortunately, little to no research has been carried out on the different estrogen receptors. Different contributions of the estrogen receptor subtypes in animals would certainly prove beneficial, as the varying effects of estrogen in the CNS and in regard to migraines are probably mediated by receptor-specific mechanisms.

In summary, even though migraine is a predominantly female disorder, and recent evidence suggests that female sex hormones play a large role in onset, the underlying mechanisms of how estrogen signaling exacerbates the situation remain unclear. Since expression of estrogen receptor subtypes is cell specific in the CNS (119), and estrogen signaling exerts a wide range of physiological effects throughout the body, the roles of sex hormones in migraines are probably multifactorial. Additional clinical and preclinical studies are needed to understand the cell signaling pathways and ER subtypes that contribute to the onset of migraines in women.

Stroke

Clinical and epidemiological evidence suggests that sex differences play a significant role in the physiological response to stroke in the young and elderly. Studies have shown that, compared with women, there is an ~33% higher incidence of stroke in men (14). Being male is a nonmodifiable risk factor for hemorrhagic and ischemic stroke observed not only in adulthood but also in neonatal, perinatal, and childhood populations (15, 27, 94, 95, 219). Sex disparities in ischemic stroke outcomes also exist, with women having better outcomes than men (112, 125, 238). However, the epidemiology of stroke changes for women over the age of 75 yr, when the stroke incidence increases and eventually surpasses that of age-matched men (212, 222). Elderly women also experience more severe strokes and greater disability compared with age-matched men (14, 85, 199, 226). The protective effects for premenopausal women suggest that gonadal sex hormones play significant roles in the physiological outcomes of ischemic injury.

Childhood stroke is rare but does exist. Prepubescent girls are more resistant to ischemic damage than boys (148, 166). Although androgen levels are relatively low in prepubescent children, neonatal and infant boys experience two postnatal testosterone surges (111), which may account for stroke being more common in boys than girls (152). Animal studies have yielded similar results in that, compared with age-matched female mice, neonatal male mice experience greater long-term disability such as brain volume loss and greater behavioral dysfunction after hypoxic-ischemic stroke (110, 135). Therefore, early organizational effects of sex hormones in the CNS have gained recent attention to better understand how females are protected from stroke during prenatal and early postnatal development (53).

In females, brain organization is influenced by the absence of E₂, since E₂ produced by the ovaries binds to α-fetoprotein and is unable to cross the blood-brain barrier (18). Sensitivity to ischemia may be at least partially molded early in life after the organizational effects of gonadal sex steroids (210). For example, neonatal male rats treated with testosterone propionate have reduced infarct sizes and increased E₂ levels poststroke (210). These effects were not observed in neonatal male rats treated with nonaromatizable androgens. This suggests that increased expression of aromatase and serum E₂ levels may protect against ischemic damage (210). However, these findings somewhat contradict the previous hypothesis that neonatal testosterone surges predispose males to pediatric stroke. Since conversion of testosterone to E₂ seems to be protective in the vasculature and detrimental in the CNS, more research is needed to fully assess how much early developmental effects of sex hormones in the CNS affect sex predisposition and response to pediatric ischemic stroke.

Stroke outcomes vary depending on the sex and age of the animals subjected to ischemic injury. Young female rodents are more resistant to both global (100) and focal (10) ischemic damage, but they become more susceptible to ischemic insult than male rodents after 15 mo of age (69, 141, 149). The loss of circulating estrogen and gonadal senescence is possibly responsible (141).

Various studies show that estradiol dampens inflammatory responses (210, 254, 259). TNF, a proinflammatory cytokine, acts on neuronal receptors (72, 78) as well as glial cells (70) and endothelial cells (22). Even though this cytokine is intended to aid in repair, levels are dramatically elevated in stroke patients (58), and an increasing body of evidence shows that increased TNF expression has neurotoxic effects (134, 261). Therefore, anti-TNF therapies are being investigated as treatment options. Postmenopausal women display higher levels of circulating TNF (126), suggesting that estrogen is able to reduce its expression. Female rats with low estrogen levels exhibit higher TNF levels compared with controls; furthermore, physiological levels are effective at inhibiting TNF secretion (143).

Estrogen is also able to modulate the inflammatory response after stroke by interfering with NF-κB, a transcription factor that mediates inflammatory signaling in neurons and other cell types (76, 170). The full relationship between NF-κB and stroke is still not fully understood since NF-κB has a dual role of both inducing proinflammatory cytokine expression (41, 201) and cell survival signaling (59, 81). However, hypoxic conditions (229) and oxidative stress (71), both characteristic of stroke, have been shown to increase NF-κB expression. In vitro experiments have confirmed that when E₂ activates ERβ, it promotes IκB-α, which is responsible for dampening NF-κB signaling (285). Administering estradiol increases the binding of E₂-ER complexes at the promoter sites of the vascular inflammatory genes monochemoattractant protein (MCP)-1 and cytokine-induced neutrophil chemoattractant-2β, which interfere with NF-κB signaling (285). Thus increased estrogen signaling here may benefit by dampening subsequent inflammatory cascades (Fig. 3).

Much like estrogens, androgens’ effects may contribute to sex disparities in stroke incidence. Several studies have shown androgens as having both beneficial and detrimental effects to the CNS after ischemic insult (50, 51, 193). Even though estrogen-mediated protection is thought to contribute largely to sex differences in stroke risk, high testosterone levels in young men may also contribute. The South London Stroke register shows that, compared with age-matched women, men between the ages of 14 and 44 yr have a higher incidence of stroke (279). To date, there has been limited clinical research on the relationship between androgens and stroke. However, some evidence has shown that anabolic steroid use increases vascular tone (73). Furthermore, a young man with a history of anabolic steroid abuse showed signs of stroke and other adverse cardiovascular events (298). Even though this evidence is not from large-scale studies, it still points to high androgen levels as a predisposing factor for stroke. Surprisingly, epidemiological data have shown that the female and male incidence of stroke begin to equalize in patients past 54 yr old (222). The gradual decline of testosterone levels in men as they age might be a contributor.

Although high androgen levels may contribute to stroke risk, clinical evidence links low levels of circulating androgens to increased stroke risk in men as well (296). For example, a prospective cohort study of 22,310 prostate cancer patients revealed a higher stroke risk in men who underwent androgen deprivation therapy (16), which may predispose men to ischemic brain injury. The data conflict with other studies that have linked high testosterone levels to increased stroke risk, suggesting a physiological threshold of circulating androgen levels protects against ischemic brain injury.

Rodent models of ischemic stroke have provided some insight into the influence of androgens in CNS injury that cannot be derived from clinical observations. As previously mentioned, male rodents often exhibit greater cerebral ischemic damage compared with female rodents of the same age (165). In studies using middle cerebral artery occlusion (MCAO) to model ischemic stroke, castrated young male rats and mice had decreased infarct areas (50, 265), suggesting that endogenous androgens exacerbated ischemic brain injury. Other studies have also demonstrated that very low doses of androgens complement the reduced infarct sizes observed in castrated male animals, while high doses worsen ischemic injury (265). Under these conditions, there was a dose-dependent response to androgen replacement.

There is ample evidence that AR activation is involved in multiple signaling pathways that mediate cell death or survival. Some experimental studies have demonstrated that AR signaling can be neuroprotective (197, 213). For example, low androgen doses result in activation of the MAPK/ERK pathway to encourage cell survival (197). Further evidence suggests that androgen signaling can increase the antioxidant catalase (5) and cAMP response element-binding protein activity (198) in the CNS. In contrast, androgen signaling can oppose cell survival and therefore be detrimental when faced with ischemic CNS injury (Fig. 3). For example, testosterone is able to induce all proapoptotic p53 genes in oligodendrocytes (42). Studies using cultured hippocampal cells and oligodendrocytes have also demonstrated that exogenous testosterone exacerbates excitotoxicity by increasing glutamate-induced Ca²⁺ influxes (42, 83, 115). In addition, some evidence has shown that androgen signaling activates target genes in the mouse brain after MCAO that may increase inflammation and disruption of the blood-brain barrier, thereby worsening CNS damage (50).

Conflicting bodies of evidence have demonstrated that androgen signaling modulates pathways that either promote or oppose cell survival in the event of ischemic brain injury. Nevertheless, there seems to be a dose-dependent relationship between androgens and stroke outcome, as observed in both clinical and animal studies. This makes it difficult to define a concrete role for androgens in the event of ischemic stroke.

Recent advances in preclinical stem cell transplantation research have shown promising results for future treatment of ischemic stroke. More specifically, selected studies have demonstrated that intracranial human neural stem cell transplantation improves ischemic stroke outcome through multiple anti-inflammatory and neurotrophic bystander effects (74, 117). Since stem cells can either be biologically male (XY) or female (XX), the sex of stem cells used in these newly developing therapies may also play a role in their expression of sex hormone receptors and response to the host’s hormonal environment.

Unfortunately, few studies have thoroughly investigated sex-specific stem cell effects in the context of neurovascular insults between females and males. However, a few investigations have demonstrated that significant differences exist between female and male stem cell behavior. For example, stromal stem cells found in menstrual blood are biologically female and display a high proliferative rate coupled with an immature phenotype containing similar markers as embryonic cells (11). In comparison, stem cells derived from Sertoli cells of the testis demonstrate anti-inflammatory properties most likely meant to prevent immune cell rejection of germ cells (28, 232). Although many of these studies examined the intrinsic properties of stem cells and their sex for transplantation, how the biological sex of grafted stem cells would ultimately affect behavior if transplanted into different hormonal environments between female and male hosts remains to be further studied. This poses a unique opportunity for future investigations on the interplay between stem cell sex and the host hormonal environment in cell therapies for neurovascular disorders.

Using rodents in experimental stroke studies creates inherent challenges when comparing results across this area of vascular research. It is well established that stroke occurs often in humans with preexisting disorders such as hypertension and diabetes (128, 264), yet experimental studies on ischemic stroke often use young animals without similar comorbidities (50, 265). Therefore, future research on sex differences in ischemic stroke may benefit from the use of aged hypertensive and/or diabetic rodents that are already stroke prone before MCAO or other techniques used to induce brain ischemia.

Conclusions

A large body of research indicates that estrogens and androgens have a multitude of complex effects on vascular physiology that often affect females and males differently. The physiological effects of androgen signaling in the vasculature appear to be even more complex and remain poorly understood. Many clinical studies have reported higher incidences of CVD with age-related androgen decline in men, but animal studies have often demonstrated that androgen signaling contributes to cardiovascular pathogenesis. Moreover, the relationship between androgen signaling and vascular physiology in women is even less understood and has not been as thoroughly investigated. This highlights the growing need for more research focusing on different cell signaling pathways and downstream targets that are regulated through AR and the different estrogen receptor subtypes. This will be a crucial step toward understanding sex differences in the pathogenesis of vascular disorders.

GRANTS

This work was supported by Tulane University Building Interdisciplinary Research Careers in Women's Health Grant 2K12HD043451-11, Louisiana Board of Regents Support Fund-Research and Development RCS Grant LEQSF(2016–2019)-RD-A-22, and startup funds from the Department of Pharmacology, Tulane University School of Medicine.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.C.B. and S.C.K. prepared figures; A.C.B., J.-P.L., and M.H.H. drafted manuscript; A.C.B., K.-J.Y., J.-P.L., and M.H.H. edited and revised manuscript; A.C.B., S.C.K., K.-J.Y., J.-P.L., and M.H.H. approved final version of manuscript.