Sex differences in abdominal aortic aneurysms

Abstract

Abdominal aortic aneurysm (AAA) is a vascular disorder with a high case fatality rate in the instance of rupture. AAA is a multifactorial disease, and the etiology is still not fully understood. AAA is more likely to occur in men, but women have a greater risk of rupture and worse prognosis. Women are reportedly protected against AAA possibly by premenopausal levels of estrogen and are, on average, diagnosed at older ages than men. Here, we review the present body of research on AAA pathophysiology in humans, animal models, and cultured cells, with an emphasis on sex differences and sex steroid hormone signaling.

INTRODUCTION

Abdominal aortic aneurysms (AAAs) are characterized by the localized but permanent expansion and weakening of the abdominal aorta (173). Although AAA pathogenesis is complex and not yet fully understood, key pathological characteristics include vascular inflammation, oxidative stress, destruction of the aortic extracellular matrix (ECM), and thinning of the aortic wall from loss of vascular smooth muscle cells (VSMCs) (63, 91, 123).

Clinical diagnosis is usually made when the abdominal aortic diameter increases by at least 50% (77). However, AAAs are often asymptomatic until rupture and thus are usually discovered by chance during unrelated clinical examinations (77). With an ~80% fatality rate, AAA poses a unique public health concern because ruptures are a major cause of mortality (77). To date, there are no pharmacological interventions, and the only treatment option is to monitor the size of the aneurysm until surgery is appropriate (131). Risk factors include smoking, age, and complex atherosclerosis (100), with male sex being a major predisposing factor (~4–6 times more prevalent in men compared with women) (80, 100, 165). Men 50 yr and older have a higher risk of AAA. The prevalence of AAA in the United States ranges from 1.3% in men 45–54 yr of age to 12.5% in men 75–84 yr of age (55).

Several epidemiological studies that have highlighted sexual dimorphism in the development and progression of AAAs have shown that women are at greater risk for aneurysm rupture and morbidity after surgical repair (80, 167). Data also suggest women are no longer protected from developing AAA postmenopause (185). Studies on rodents have demonstrated that male animals are more susceptible to experimental AAA formation compared with age-matched female animals (1). These observations suggest that endogenous sex hormone signaling contributes to sex differences in AAA (Fig. 1).

Fig. 1.

Diagram of sex steroid hormones on abdominal aortic aneurysm (AAA) development. Estrogen signaling blunts AAA development by dampening inflammation, oxidative stress, and proteolysis. Androgen signaling may facilitate AAA development by promoting inflammation, oxidative stress, and proteolysis.

SEX STEROID HORMONES IN AAA

Endogenous estrogen, or estradiol (E₂), is generally believed to be vasoprotective in premenopausal females in the context of cardiovascular diseases, such as hypertension and atherosclerosis (206). Evidence also suggests that E₂ may protect women from AAA development because larger AAAs in women are linked to earlier onset of menopause (185).

Furthermore, the protective activity of estrogen was also indicated by data from Lederle et al. (101) from the Women’s Health Initiative, which showed a lower incidence of AAA events in women taking hormone replacement therapy (HRT).

However, some clinical studies on HRT and AAA in postmenopausal women have presented contradictory results. For example, postmenopausal women with prior hysterectomy in the Estrogen Alone Trial were at greater risk for AAA events if they received conjugated equine estrogen HRT compared with a placebo (71). However, these contradictory findings may be explained by the hormone timing hypothesis, which asserts that protective effects of HRT on cardiovascular health are lost as time between menopause and the start of therapy increases (67, 68). Evidence from animal studies also supports that vasoprotection from estrogen signaling is age dependent and reliant on proper estrogen receptor (ER) expression (9). Therefore, E₂ may indeed contribute to female protection against AAA.

Male subjects have higher levels of circulating androgens and less estrogen compared with female subjects. Clinical studies have often indicated that endogenous androgen levels are often inversely related to risks of aortic aneurysms and other cardiovascular diseases in men (193, 207). For example, a cross-sectional study on 3,620 men ages 70–88 yr found that lower free testosterone levels were associated with AAA in older men (207). On the other hand, the majority of animal studies have indicated the association of male gonadal sex hormones with AAA progression. For example, male apolipoprotein E-deficient (ApoE^−/−) and low-density lipoprotein receptor-deficient (LDLR^−/−) mice infused with angiotensin II (ANG II) display a greater incidence and size of AAA compared with age-matched female mice (115). However, castration of male ApoE^−/− mice halts expansion of the vascular lumen and dampens the progression of ANG II-induced AAA (213). These conflicting findings in clinical and animal studies suggest that in male subjects, there may be an optimal physiological threshold of circulating androgens that protects against AAA (Fig. 2).

Fig. 2.

Mortality rates for abdominal aortic aneurysms (AAA) in the United States for the years of 2015 and 2016 stratified by age and sex. Mortality rates for AAA are higher for men in all age groups. Data correspond to the following ICD-10 Codes: I71.3 (AAA, ruptured), I71.4 (AAA, without mention of rupture), I71.5 [thoracoabdominal aortic aneurysm (TAAA), ruptured], I71.6 (TAAA, without mention of rupture). These data are from Centers for Disease Control and Prevention, National Center for Health Statistics, Compressed Mortality File 1999–2016 on the CDC WONDER Online Database, released December 2017. Data are from the Compressed Mortality File 1999–2016, Series 20, No. 2V, 2017, as compiled from data provided by the 57 vital statistics jurisdictions through the Vital Statistics Cooperative Program. Accessed at https://wonder.cdc.gov/cmf-icd10.html.

SEX DIFFERENCES IN ANATOMY AND BIOMECHANICS

Male and female subjects have anatomic differences in the abdominal aorta that may attribute to sex-specific AAA susceptibility and prognosis. AAA rupture depends on multiple factors, including blood pressure, tissue properties, and vascular geometry (96). As stated above, men are more susceptible to developing AAA, but women are at greater risk for progression and rupture (40, 129). The increased risk of rupture in women may be a result of unfavorable biomechanical properties of the female aorta. For example, uniaxial tensile strength measurements on tissue specimens from patients undergoing open repair for AAA showed a trend toward decreased strength in female aortas compared with male aortas (181). Female aortas require significantly lower energy for dissection compared with male aortas (176). Furthermore, finite-element analysis demonstrated that women with AAA had a higher peak wall rupture risk compared with men with AAA (96).

Thickening of arterial walls is believed to be a compensatory mechanism for preventing increased wall stress, as aortic diameter increases with age (8). Some clinical findings have indicated that this response is insufficient in men, leading to increased stress on the abdominal aorta with age. For instance, a study on healthy subjects using B-mode ultrasonography demonstrated that men had greater abdominal aortic wall stress than women, and, in the male aorta, wall stress increased with age while remaining consistent in women (6). These findings might partially explain why men are more susceptible to initial AAA formation as they age, whereas women remain resistant until menopause (Fig. 2).

Biomechanical differences between male and female aortas may stem from sex-specific expression of ECM components that make up the structural integrity of the vessel walls. Most research on aortic ECM has focused on collagen, which confers arterial strength, and elastin, which is important for proper mechanical recoil. For example, it has been reported that intima-media composites of male AAA walls have lower elastin content, whereas female AAA walls have lower percentages of collagen (176). However, in a different clinical investigation using biopsies from patients with AAA undergoing open repair, elastin expression was greater in the nonthrombus-covered aneurysm wall in men compared with women, but there were no differences in elastin or elastolytic enzymes between the sexes in thrombus-covered aneurysm tissue (184). On the contrary, analysis of AAA biopsies from both men and women found no significant differences in aneurysm wall thickness or collagen content, but there was a sex difference in collagen cross linking, with women having a lower hydroxyl pyridinoline-to-lysyl pyridinoline ratio (HP/LP) than men (183). Although this study conflicts with previous reports identifying differences in collagen content between male and female AAA tissue, it indicates a possible sex difference in posttranscriptional modifications of collagen that affect intermolecular cross linking. This may be significant in the context of AAA because cross linking is responsible for much of the mechanical stability of collagen (140), and higher HP/LP is believed to confer greater stability (7).

Discrepancies among the findings stated above may stem from small sample sizes and different methods used to quantify structural components of aortic ECM in biopsies from patients with AAA. Furthermore, it is inherently difficult to compare AAA biopsies from men and women because not all patients with AAA can be closely matched for age and stage of disease. For example, AAA biopsies in one study were from men that were younger than their female counterparts (176), whereas another study used biopsies from age-matched men and women (183). In addition, the lack of control tissue from elderly men and women does not allow for comparison of collagen and elastin deposition in healthy versus aneurysmal aortas to analyze possible sex differences in AAA susceptibility or risk of rupture. Thus, future clinical research on sex differences in aortic anatomy and AAA needs to properly adjust for age and disease stage.

In vivo studies have demonstrated sex-specific aortic ECM differences (38, 133). One study on young and old monkeys demonstrated that aortic stiffness increased more in old male monkeys compared with old female monkeys and that thoracic aortic elastin density decreased significantly with age only in male monkeys (151). Sex differences in ECM composition have also been documented in rodent models of aortic aneurysm. The elastase-infused rat model of AAA was first described in the early 1990s and involves surgical exposure of the infrarenal aorta with subsequent application of pancreatic elastase (3). This AAA model, which involves enzymatic degradation of medial layer elastin and changes in collagen levels, elicits both an inflammatory response and gradual expansion of the abdominal aorta (3). One investigation using the elastase infusion rat model of AAA demonstrated a decrease in both type I and type III collagen in male subjects compared with female subjects at 7 days postperfusion (16). These studies proposed that alterations in collagen and elastin turnover may contribute to sex differences in AAA development.

SEX DIFFERENCES IN PROTEOLYSIS PROFILES

The destruction of the aortic ECM and loss of vessel integrity are defining characteristics of AAA pathology. In recent years, proteolytic enzymes secreted from the aneurysm wall have gained interest as potential biomarkers for profiling aneurysm progression. Human AAA tissue studies have suggested that matrix metalloproteinases (MMPs) activate endothelial cells (ECs) and aortic smooth muscle cells and degrade the aortic ECM and contribute to aneurysm development (27, 148). More specifically, mounting clinical evidence indicates that MMP-2 and MMP-9 are highly associated with AAA pathogenesis (4, 32, 58, 148, 187). Both MMP-2 and MMP-9 are capable of cleaving elastin in addition to type IV and type V collagen, which are all important structural components of the abdominal aorta (70, 180). Interestingly, MMP-8 is increased at the site of AAA rupture in aneurysm wall biopsies, but sex differences were not analyzed (195).

MMP-9, a proteolytic enzyme highly associated with AAAs (63, 171), may serve as a biomarker for sex differences in the development and pathophysiology of AAA (186). For instance, a prospective case-control study demonstrated that men with AAAs have lower plasma levels of MMP-9 compared with women with similarly large AAAs (186). This study marks an important sex difference with respect to circulating MMP-9 profiles in AAA.

Animal experiments and cell culture studies have suggested that differences in MMP profiles between men and women with AAAs may result from estrogen signaling. A study using elastase perfusion to model AAA in Wistar rats demonstrated that ovariectomized (OVX) female rats had lower estrogen levels, accelerated aneurysm dilation, and higher MMP-9 and MMP-2 expression compared with female rats that underwent sham operations (198). When administered exogenous E₂, male rats displayed higher circulating estrogen levels, decreased expression of MMP-9 and MMP-2, and slower aneurysm progression compared with male rats given saline injections (198). In another study, both MMP-2 gene expression and protein levels were three times greater in male rat aortic smooth muscle cells (RASMCs) compared with female RASMCs after stimulation with IL-1β levels that are similar to those found in human AAA tissue (197). In vitro, when stimulated with exogenous E₂, there was no significant effect on MMP-2 activity in male or female RASMCs. However, in vivo treatment with exogenous E₂ significantly reduced MMP-2 production in male rat aortas to levels seen in female rat aortas (197). This indicates that increased estrogen signaling in the female vasculature may regulate MMP expression and activity in a way that protects against AAA development.

Increased ER-α expression and activation in female subjects may be responsible for mediating estrogen inhibition of MMPs and protection from aneurysm development. Aortic ER-α protein levels were 100% greater in female mice compared with male mice at 14 days after elastase infusion (97). Both MMP-2 and MMP-9 activity were decreased in female mouse aortas compared with male mouse aortas at day 3. MMP-2 activity remained decreased in female aortas at 14 days (97). Western blot analysis of human AAA samples showed 80% greater ER-α protein levels in female compared with male subjects (97). This demonstrates that increased ER-α in female subjects is inversely associated with both MMP activity and AAA development.

Members of the MAPK family are thought to contribute to sex differences in proteolytic activity that contribute to AAA. ERK and JNK have been implicated in the aortic regulation of MMP-2 and MMP-9 and thus may be important in AAA pathogenesis (30, 34, 50). One study demonstrated that phosphorylated ERK levels were higher in male RASMCs compared with female RASMCs, which resulted in elevated pro-MMP-2 levels in male RASMCs (34). In addition, AAA tissue from C57BL/6 mice displayed significantly higher JNK levels than female tissue, which resulted in increased levels of pro-MMP-2, active MMP-2, and pro-MMP-9 in male mice (30).

The phosphatidylinositol 3-kinase (PI3K)/Akt pathway is also regulated differently in male and female subjects. This may also contribute to sex differences in MMP activity and AAA susceptibility. A recent study has demonstrated that elastase administration in vitro induced Akt phosphorylation in both male and female RASMCs, but male cells contained higher levels of phosphorylated Akt compared with female cells (51). Pharmacological inhibition of Akt resulted in less pro-MMP-9 in response to elastase in male and female RASMCs. Also, Akt knockdown decreased MMP-2 activation in male cells (51). Levels of phosphorylated Akt were also significantly higher in tissue from male patients with AAA (51). These findings suggest that male RASMCs may be more susceptible to Akt phosphorylation and downstream activation of MMPs that facilitate AAA development.

In brief, sex differences in MMP profiles may explain why female subjects are resistant to AAA development, whereas male subjects are more susceptible. Higher estrogen levels in female compared with male subjects appear to inhibit MMP expression and activity through mechanisms mediated by ER-α. Although male aortic cells have higher levels of ERK (34), JNK (30), and Akt (51) compared with female cells, it is unclear whether this is a result of sex hormone signaling. This underlies the lack of research surrounding androgen signaling and potential regulation of proteolytic enzymes implicated in AAA. Further studies need to examine the effects of androgen deprivation and androgen receptor (AR) antagonists on MMP-9 and MMP-2 in murine models of AAA.

SEX DIFFERENCES IN THE RENIN-ANGIOTENSIN-ALDOSTERONE SYSTEM

The renin-angiotensin-aldosterone system (RAAS) is a major regulator of extracellular fluid homeostasis and blood pressure. Multiple studies have indicated that the RAAS contributes to human AAA pathogenesis, as the enzymes involved in ANG II formation are upregulated in human AAA tissue (25, 62, 137, 178). Furthermore, the RAAS is involved in vascular remodeling and inflammation seen in animal models of AAA pathogenesis (25, 56). Thus, many in vivo studies have implemented chronic infusion of ANG II in ApoE^−/− mice to model human AAA pathology (23, 24, 158).

Major sex differences exist in the expression levels of RAAS constituents and the balance between pressor and depressor arms of the system (10, 87), with female subjects shifting toward the depressor arm (10, 109, 156). Male and female subjects respond differently to RAAS stimulation and inhibition under normal and pathophysiological conditions (157, 203). For instance, ANG II increases blood pressure to a greater extent in men than women (33), even though clinical findings indicate that normotensive premenopausal women often have plasma ANG II levels similar to men of the same age (127). Therefore, it is possible that regulation of the RAAS through sex hormone signaling may contribute to male susceptibility and female resistance to AAA.

Activation of the ANG II type 1 receptor (AT₁R) induces vasoconstriction, Na⁺ reabsorption, fibrosis, and inflammation (179). OVX resulted in increased AT₁R density, whereas estrogen replacement reversed this effect in female rat aortic tissue (28). Moreover, in spontaneously hypertensive rats (SHRs), aortic AT₁R mRNA levels are lower in female compared with male rats, but OVX increases AT₁R mRNA levels to those observed in male rats (163). Estrogen is able to downregulate the expression of AT₁Rs through a posttranscriptional mechanism by modulating cytosolic RNA-binding proteins that recognize the 5′-leader sequence of AT₁R mRNA before translation (89). In support of these findings, an ex vivo study on VSMCs demonstrated that the absence of estrogen resulted in the upregulation of AT₁Rs, whereas the presence of estrogen reversed this effect (136). On the basis of these findings showing a role of ANG II activation of AT₁R in AAA pathogenesis (191), estrogen downregulation of these RAAS constituents likely protects against AAA.

Many animal studies have indicated that intact male subjects experience accelerated experimental AAA formation after chronic ANG II infusion. This may result from androgen-mediated upregulation of vasoconstricting RAAS constituents, including angiotensinogen gene expression, AT₁R expression, and renin activity (13, 64). For instance, it has been hypothesized that higher renin levels in male subjects results from increased proximal tubular Na⁺ reabsorption in the presence of testosterone (152). Hyperandrogenism in female subjects has been shown to remove protection from developing an aneurysm by upregulating RAAS constituents as well. For example, neonatal exposure of female mice to testosterone increased AT₁R mRNA abundance in abdominal aortic tissue and obviated protection against ANG II-induced atherosclerosis and AAA despite low serum androgen levels in adulthood (214). On the contrary, neonatal testosterone exposure in male mice had no significant effect on ANG II-induced vascular insults (214). These findings suggest the developmental effects of sex hormone signaling as significant long-term determinants of sex-specific AAA pathogenesis.

Castration of male ApoE^−/− mice reduces experimental AAA development to the level seen in female mice in response to chronic infusion of ANG II (65, 213). One possible explanation may be that naturally higher levels of androgen signaling in male rodents upregulate aortic AT₁R expression in vascular regions susceptible to aneurysm development (64). For example, the AT₁R mRNA expression level in the abdominal aortas of male mice was eight times greater compared with thoracic aortas, and this expression decreased postcastration in abdominal but not thoracic aortic vessels (64). Furthermore, administration of dihydrotestosterone (DHT), a nonaromatizable androgen, to castrated male mice restored AT₁R mRNA abundance in abdominal but not thoracic aortas and even increased AT₁R mRNA expression in abdominal aortas of female mice (64). Interestingly, administration of DHT did not alter serum cholesterol or lipoprotein distributions, nor did it affect areas prone to atherosclerosis in either male or female mice; it did, however, increase AAA incidence in both male and female mice (64). Therefore, this study accentuates the physiological significance of increased AT₁R expression in male abdominal aortas as a result of endogenous androgen signaling, which may partly explain the male disadvantage in AAA susceptibility.

In addition to the cellular effects of sex hormone signaling, there is a growing body of evidence indicating that sex chromosomes contribute to sex differences in RAAS. For example, the Sry gene on the Y chromosome enhances the promoter activity of angiotensinogen, renin, and angiotensin-converting enzyme (ACE) while decreasing promoter activity of ACE2 (128). This would result in the upregulation of the vasoconstricting arm and downregulation of the depressor arm of the RAAS. Indeed, overexpression of Sry genes increases renal sympathetic nerve activity, plasma renin activity, and renal ANG II content, thereby elevating arterial pressure (35). Thus, male subjects are more susceptible to shifts toward the classical vasoconstricting arm of the RAAS. On the contrary, AT₂R and ACE2 genes are on the X chromosome (86, 174), which supports a greater role of the depressor arm of RAAS in female subjects.

Sex-specific genetic polymorphisms influence the RAAS. For example, ACE and AT₁R polymorphisms are significantly associated with hypertension in men but not women (138, 153). In contrast, angiotensinogen and ACE2 gene polymorphisms are associated with hypertension only in women (112, 130). A meta-analysis indicated the G8790A ACE2 variant was highly associated with hypertension in women but not men (112). Taken together, these studies highlight explicit genetic differences between men and women that influence the RAAS and potential susceptibility to AAA.

In brief, a large body of research confirms sex differences in regulating the RAAS, a hormone system that is heavily involved in AAA pathophysiology. Naturally higher levels of estrogen signaling in female subjects appear to be protective against AAA formation by downregulating AT₁R expression (28, 89, 136, 163, 210). In contrast, higher levels of androgen signaling in male subjects seem to upregulate aortic AT₁R expression and contribute to aneurysm susceptibility (13, 64, 214). Additional evidence suggests that the Y chromosome drives the RAAS toward an aneurysm-favorable state through epistasis (128). AT₂R and ACE2 genes are located on the X chromosome (86, 174), and further studies need to examine whether these genes avoid inactivation and contribute to protecting females against AAA.

OXIDATIVE STRESS AND SEX HORMONE SIGNALING

Oxidative stress is often defined as the imbalance between antioxidants and reactive oxygen species (ROS), with the latter in excess. Oxidative stress has a pathophysiological significance in the vasculature because an abundance of ROS can potentially result in cellular damage and death (162). Furthermore, a substantial amount of evidence indicates that oxidative stress leads to increased inflammation and creates a positive feedback loop, resulting in further tissue injury (126, 204). The human body is able to regulate redox states in normal physiological conditions to avoid injury, but studies have indicated that this ability may be severely hindered in pathophysiological conditions, such as AAA (121, 162).

As previously stated, inflammation of the aortic wall and accelerated production of ROS have been implicated in AAA pathogenesis (36). Both clinical and experimental studies have indicated that oxidative stress contributes to vascular degeneration, leading to AAA (126, 159). For example, biopsies of human aneurysmal aortas had greater superoxide ($O_2^{-}$) levels and NADPH oxidase (NOX) activity compared with nonaneurysmal aortas (126), whereas higher inducible nitric oxide synthase (iNOS) expression levels were reported in the medial and adventitial layers of human AAA tissue compared with control tissue (211). Markers of oxidative stress, such as 8-oxo-2′-deoxyguanosine (8-OHdG) and glutathione peroxidase, are also higher in both ANG II- and CaCl₂-induced mouse models of AAA (75, 159).

All vascular cell types generate ROS, such as $O_2^{-}$, often through membrane-associated enzymes such as NOS (41). Both human and animal studies have reported elevated levels of ROS and associated prooxidative enzymatic activity with AAA (42, 118, 126). ROS upregulate proteolytic enzymes such as MMPs (110), activate transcription factors such as NF-κB (107, 145), and promote VSMC apoptosis (92). Therefore, oxidative stress contributes to many of the pathological mechanisms implicated in the development of AAA.

The NOX family is a major source of RAAS-induced ROS generation in ECs, VSMCs, and macrophages through its catalytic and regulatory subunits, Nox1–Nox5, Duox1 and Duox2, p22^phox, p47^phox, Noxo1, p67^phox, Noxa1, p40^phox, and Rac1 (59, 98, 143). Biomolecules involved in AAA pathogenesis, such as ANG II and TNF-α, act through multiple pathways that increase NOX activity in vascular cells (104). For example, ANG II is able to cause mitochondrial dysfunction through the PKC pathway that activates NOX and generates ROS (31). Animal studies have indicated that elevated NOX activity may have direct implications in the pathogenesis of AAA. For instance, the ANG II-induced mouse model of AAA demonstrated that deletion of the cytosolic p47^phox subunit of NOX blunted the activity of the enzyme, reduced aortic ROS production, and significantly reduced AAA formation (172). This evidence suggests that deregulated NOX activity contributes to AAA development.

Estrogen Signaling

Preclinical evidence suggests that female subjects are more resistant to oxidative stress in vascular disease compared with male subjects. Animal studies have demonstrated that OVX increases $O_2^{-}$ generation by NOX, but treatment with E₂ lowers $O_2^{-}$ levels to those seen in intact female subjects (22, 177, 212). Similarly, in vitro estrogen administration attenuates both NOX- and ANG II-induced ROS production in RASMCs (99, 177).

Estrogen signaling likely decreases NOX activity by downregulating expression of its subunits. OVX in SHRs leads to increased expression of p22^phox, p67^phox, and Rac1, which was prevented with estrogen replacement (99). Similarly, one in vitro study demonstrated that administration of E₂ blunts both mRNA and protein expression of the Nox2 and p22^phox subunits in human umbilical vein ECs (189). Another study reported that arteries of female SHRs generate lower amounts of basal $O_2^{-}$ compared with male SHRs as a result of lower expression of NOX subunits such as Nox2, p22^phox, and p47^phox (21). This sex difference in NOX expression may result from estrogen regulation of AT₁R expression and signaling because the administration of losartan, an AT₁R antagonist, abolished sex-specific differences in ROS production and NOX subunit expression in SHRs (21). This highlights a potential sex difference in AT₁R signaling or expression that may influence NOX activity.

Several other studies have affirmed that estrogen signaling regulates NOX subunit expression. For instance, chronic administration of exogenous E₂ to OVX rats reduces the level of the membrane-associated NOX subunit p47^phox and blunts production of vascular $O_2^{-}$ (39). Estrogen administration dose dependently decreases expression of Nox2 in human ECs, which was reversed in ECs that were preincubated with ER antagonists, such as tamoxifen and ICI-182,780 (189). These results clearly suggest an ER-dependent mechanism, whereby estrogen signaling inhibits Nox2 expression. However, the ER subtype responsible for mediating these effects has yet to be determined.

Androgen Signaling

In opposition to estrogen signaling, androgen signaling increases NOX activity and expression. For instance, administration of exogenous testosterone increased NOX activity and expression in adult Wistar rats as well as leukocyte migration to the adventitia (15). These effects were inhibited with the addition of flutamide (AR antagonist) or apocynin (NOX inhibitor) (15). This suggests that androgen signaling is able to induce leukocyte migration through a NOX-dependent mechanism, which may contribute to both oxidative stress and immune cell infiltration in AAA. Another study using OVX SHRs demonstrated that administration of exogenous testosterone abolished the protective effects of conjugated equine estrogen replacement (18). In this study, estrogen replacement restored endothelium-dependent relaxation, which was impaired in OVX rats. However, the addition of testosterone with conjugated equine estrogens impaired the benefits of estrogen on OVX-induced endothelial dysfunction and ROS generation in rat aortas through a mechanism that involved phosphorylation of the p47^phox subunit of NOX (18). This evidence indicates that higher levels of androgen signaling may predispose men to AAA development by upregulating NOX activity, ROS generation, and oxidative stress.

NOS AND SEX HORMONE SIGNALING

NO plays important roles in cardiovascular physiology, but, under pathological circumstances, excessive production of NO and generation of reactive nitrogen species can lead to cardiovascular damage (46, 60, 169). Of the NOS that catalyze the generation of NO, both iNOS and endothelial NOS (eNOS) are expressed in the vasculature and have been implicated in AAA pathogenesis (2, 42, 48, 211). For example, studies using the CaCl₂ mouse model of AAA have indicated that iNOS deficiency confers partial resistance to aneurysm formation in association with decreased aortic MMPs (201).

Increased production of ROS and the resulting oxidative stress in the vascular wall are able to inactivate eNOS and uncouple it from NO production, which generates more oxygen free radicals and abolishes NO-mediated vasoprotection (76, 182). For example, mice deficient in the eNOS cofactor tetrahydrobiopterin (BH₄) display increased uncoupling of eNOS and greater susceptibility to ANG II-induced AAA (48). Moreover, other studies using the ANG II-infused ApoE^−/− mouse model of AAA have indicated that treatment with folic acid abrogates uncoupling of eNOS from its cofactors and decreases aortic macrophage infiltration, ECM degradation, and AAA progression (166). Although these enzymes are undoubtedly complex and the mechanisms behind their involvement in AAA pathogenesis remain to be elucidated, studies have nevertheless indicated that sex hormones may influence their activity and therefore contribute to sex differences in vascular oxidative stress and AAA development.

Estrogen Signaling

Female subjects are thought to be more resistant to vascular oxidative stress compared with male subjects as a result of higher bioavailability of NO. In fact, clinical data have shown that total NO production is higher in premenopausal women than in men (43). In vivo aortas from female rats release significantly more NO compared with those of male rats (81), and OVX attenuates vascular NO production and increases oxidative stress (66).

One of the mechanisms underlying the antioxidant properties of estrogen involves the upregulation of eNOS expression in both ECs and VSMCs. For instance, administration of the selective ER modulator raloxifene increased production of bioavailable NO by upregulating eNOS gene expression and protein activity in aortic tissue of SHRs (192). Furthermore, treatment with raloxifene reduced $O_2^{-}$ production in the aortic wall (192). Other studies have indicated that rapid estrogen signaling activates eNOS through the PI3K/Akt signaling pathway (105, 196).

Estrogen signaling prevents the uncoupling of eNOS and subsequent ROS production by enhancing intracellular availability of BH₄. In fact, clinical data have shown that administration of BH₄ to postmenopausal women improves endothelial dysfunction and decreases arterial stiffness (132). In support of this finding, estrogen treatment increased the availability of aortic BH₄, attenuated $O_2^{-}$ production, restored total antioxidant capacity, and improved NO-mediated vasodilation in OVX Sprague-Dawley rats (95). Therefore, higher levels of estrogen signaling in female subjects may protect against AAA by preventing ROS generation from eNOS uncoupling.

Androgen Signaling

Several in vitro studies have examined the effects of rapid androgen signaling on eNOS activation. For instance, one study using human aortic ECs (HAECs) reported that administration of testosterone at physiological levels resulted in the phosphorylation of eNOS and rapid increases in NO generation (209). This effect was abolished when HAECs were treated with AR antagonists or AR siRNA (209). Activation of eNOS after treatment with testosterone was unaltered in HAECs when ER-α was knocked down with siRNA or cells were pretreated with aromatase inhibitors, thereby ruling out the possibility that increased NO production was a result of estrogen signaling in this system (209). This was further confirmed when DHT also activated eNOS in HAECs (209).

Androgen signaling through ARs likely induces rapid, nongenomic eNOS activation through the PI3K/Akt pathway. Treatment of HAECs with testosterone results in a rapid, dose-dependent phosphorylation of Akt (209). This effect was attenuated in HAECs that were pretreated with the PI3K/Akt inhibitors wortmannin and SH-5 (209). Furthermore, activation of eNOS in HAECs with testosterone treatment was also unaffected by pretreatment with transcriptional inhibitors, such as actinomycin D (208). Testosterone rapidly increases phosphorylation of c-Src, but AR antagonists and AR siRNA inhibit this effect (208). Src kinase inhibitors, such as PP2, are able to inhibit testosterone-mediated phosphorylation of Akt and eNOS activation (208). These findings suggest that androgens induce rapid assembly of a membrane-bound AR signaling complex that induces the c-Src/PI3K/Akt cascade and activates eNOS (208). Although these studies have clearly indicated that androgen signaling is involved in activating vascular eNOS, more research on androgenic regulation of eNOS cofactors is needed before this mechanism can be deemed protective against AAA.

SEX DIFFERENCES IN CYCLOOXYGENASES

Cyclooxygenases (COX) catalyze the conversion of arachidonic acid to prostaglandins (114), both of which are reported to be highly expressed in aneurysm tissue (11, 14, 69). Specifically, COX-2 catalyzes the production of unstable endoperoxides from arachidonic acid, which are then converted to PGE₂ via PGE₂ synthase (144). PGE₂ is an important mediator of the inflammatory response and can stimulate MMP-9 expression in cells that reside within the circulatory system, such as macrophages (146). Therefore, inhibition of COX-2-dependent inflammation and MMP-9 expression may blunt AAA development. In agreement, the nonsteroidal anti-inflammatory drug and selective COX-2 inhibitor Celecoxib has proven effective at reducing the incidence and severity of AAA in the ANG II-induced mouse model (83). Genetic knockout of COX-2 resulted in decreased expression of monocyte chemoattractant protein (MCP)-1 and macrophage inflammatory protein-1α and inhibition of AAA development in mice infused with ANG II (54). In addition, administration of COX-2 inhibitors, such as MF-tricyclic, downregulates MMP-9 expression and blunts AAA progression in the elastase infusion rat AAA model (82).

Estrogen Signaling

Estrogen signaling increases the expression of COX-2 in the aorta. One study using female rats demonstrated that OVX significantly reduced the release of thromboxane A₂ and PGI₂, whereas estrogen replacement therapy reversed these effects (106). Expression of COX-2 mRNA was significantly higher in both ECs and VSMCs of intact female rats and OVX rats treated with E₂ compared with male or OVX female rats (106). Immunohistochemical staining also indicated higher levels of COX-2 protein levels in ECs and VSMCs from intact and E₂-treated OVX female mice compared with the other treatment groups (106). These results are in agreement with another in vitro study on human umbilical vein ECs, which reported that COX-2 expression was significantly increased via estrogen signaling through the PI3K/Akt pathway (147). Therefore, estrogen signaling may regulate prostanoid function in the aorta by upregulating COX-2 expression in ECs and VSMCs. Although female subjects are resistant to AAA development before menopause, this particular estrogen signaling mechanism may contribute to aneurysm development.

Androgen Signaling

Research on androgen signaling in COX regulation has yielded mixed results. An in vivo study on rats demonstrated that administration of DHT upregulated COX-2 expression and resulted in increased middle cerebral artery constriction (57). In another study using human coronary smooth muscle cells, administration of DHT increased COX-2 expression (141). However, an opposite effect was also observed in the same study whereby administration of DHT reduced IL-1 and lipopolysaccharide (LPS)-induced COX-2 expression (141). Therefore, it is possible that androgen signaling regulates COX-2 expression differently in the presence of inflammation (141). Still, aortic COX-2 expression was increased in castrated male rats (119). Although these studies presented conflicting results, it is likely that androgen signaling differentially regulates COX-2 expression depending on multiple factors such as physiological serum androgen levels or inflammatory conditions. Because COX-2 is a major source of ROS production in the vasculature (15), further studies are warranted to investigate the cardiovascular implications of androgen signaling and ROS generation from COX-2 in AAA.

SEX DIFFERENCES IN INFLAMMATION AND THE IMMUNE RESPONSE

Inflammation, which involves monocytes, lymphocytes, and blood plasma cells, is a key component of AAA pathogenesis (26, 37, 124, 160). Although the initial stimuli that initiate the inflammatory processes in AAA are still not completely understood, experimental studies have indicated that damaged ECM components, such as elastin fragments, recruit more inflammatory cells to the aneurysm wall (85, 161). Strong sex differences exist in the regulation of the human immune system from an early age (175, 194), as human immune cells bear sex hormone receptors (19, 61, 102, 149). Therefore, sex hormone signaling may influence the inflammatory environment of AAA pathogenesis differently in men and women.

Multiple studies have confirmed sex differences in ER expression in various human tissue types. One study reported that monocytes from men express a higher ratio of ER-α to ER-β compared with monocytes from premenopausal women (149). However, monocytes from men had similar ER expression profiles to those from postmenopausal women (149). Furthermore, the same study reported no differences in T or B cell ER expression between men and premenopausal women (149). Nevertheless, this study still indicated that inflammatory cells in men and postmenopausal women may respond differently to estrogen signaling than cells in premenopausal women. ARs are also expressed in human immune cell lines, such as mast cells, macrophages, and B cells and T cells (12, 122, 190). This implies that androgen signaling is also involved in the development and regulation of the human immune system.

Lymphocytes (T Cells)

Lymphocytes are present in AAA tissue, where the majority are CD4⁺ T cells (139). The presence of T lymphocytes indicates that an adaptive autoimmune response may contribute to AAA development and progression. Indeed, an in vivo study has indicated that the absence of CD4⁺ T cells prevents AAA formation (202). Despite their presence in AAA tissue, the exact role of lymphocytes in the development of AAA is still not completely understood. CD4⁺ T cells can be classified into phenotypic subsets, such as T helper (Th)1 and Th2, based on their cytokine profiles and immunological roles. The Th1 phenotype secretes type 1 cytokines, such as interferon (IFN)-γ, that favor strong cellular immune responses by activating macrophages and enhancing inflammatory cell recruitment (113, 215). In contrast, type 2 cytokines from the Th2 phenotype are more closely associated with humoral immunity.

Th1 phenotype.

The Th1 cell phenotype contributes to AAA progression. Th1 cells are activated by IL-12, which triggers STAT4 and the subsequent expression of IFN-γ (216). Several studies have indicated that both STAT4 and IFN-γ are upregulated in human blood serum and aneurysm tissue (47, 73, 108). This indicates an elevated presence of Th1 cells with limited Th2 lymphocyte activity in the later stages of AAA. Similar results implicating the Th1 phenotype have also been demonstrated in studies using the CaCl₂ murine model of AAA. For instance, mice deficient in CD4⁺ T cells displayed reduced MMP expression and no significant aneurysm formation in the CaCl₂ model of AAA, whereas IFN-γ treatment abolished this protective effect and led to aneurysm development (202).

Estrogen signaling.

To date, it is unclear whether estrogen signaling protects against aneurysm formation through regulation of the T cell phenotype. Clinical data have indicated that higher estrogen levels correlate with a stronger Th2 response, whereas menopause results in elevated Th1 cytokines such as IFN-γ (29, 74, 188). Furthermore, the higher estrogen levels observed during pregnancy are reported to skew the immune response from Th1 to Th2 (120). Men have higher Th1-to-Th2 cytokine ratios (53). On the other hand, women are more susceptible to Th2-associated autoimmune disorders, such as systemic lupus erythematosus, compared with men (134).

Although clinical data have indicated that estrogen dampens Th1-associated immune responses, preclinical research on estrogen modulation of adaptive immunity has yielded conflicting results. Studies have demonstrated that estrogen modulates IFN-γ-secreting Th1 lymphocytes by enhancing IFN-γ expression in both humans and mice (44, 79). This is potentially mediated by direct interaction of ERs with estrogen response elements in the promoter region of the Ifn-γ gene (44) or upregulation of the Th1-specific transcription factor T-bet (78, 79). ER-α-deficient mice have decreased IFN-γ-secreting cells in lymph nodes, suggesting that estrogen-driven Th1 cell responsiveness is dependent on ER-α-mediated signaling (116). Nevertheless, there is a need for future studies to examine the effects of estrogen regulation of T cell phenotypes in rodent models of AAA.

Androgen signaling.

Androgen signaling significantly impacts peripheral T cell differentiation and function (84, 94). Higher levels of androgen signaling elicit anti-inflammatory effects by inhibiting Th1 polarization (84). For instance, a study evaluating the effects of androgen signaling on murine CD4⁺ T cells demonstrated that castration of mice resulted in significantly altered gene expression patterns important to IFN signaling and Th cell differentiation (84). Castrated mice showed upregulated expression of genes related to Th1 polarization, such as T-bet and IFN-γ, as well as increased Th1 surface markers, such as chemokine (C-X-C motif) receptor 3 (84). Testosterone treatment inhibited Th1 differentiation of CD4⁺ cells from male mice by inhibiting IL-12-induced Stat4 phosphorylation (84). Castration significantly increased T cell infiltration in the lungs and prostate, suggesting that androgen signaling dampens T cell autoimmunity (84). These findings may nevertheless be applicable to immune cell infiltration in AAA. Future studies need to examine whether androgen deprivation or treatment affects the extent of T cell infiltration in both male and female animal models of AAA.

Macrophages

Early aortic infiltration of macrophages is a key characteristic of AAA pathogenesis in animal models (103). In fact, proinflammatory cytokines associated with macrophages such as IL-6, IL-1β, IFN-γ, and TNF-α are now being studied as biomarkers of AAA progression because they are elevated in patients with AAA (73). Macrophages are capable of producing cytokines and therefore contribute to the differentiation and recruitment of other immune cells (88). In addition, certain macrophage phenotypes are known to secrete proteases, such as MMP-9 and MMP-12, that degrade the aortic ECM and contribute to aneurysm development (111). However, macrophage activity depends on polarization to the proinflammatory (M1) phenotype or the inflammation-resolving (M2) phenotype.

M1 macrophages secrete MMPs, promote inflammation, and have been implicated in the pathogenesis of AAA. Macrophages from patients with AAA display higher levels of M1 cell surface markers compared with macrophages from healthy control subjects (49, 155). M1 macrophages can be activated by certain proinflammatory cytokines. IFN-γ potentiates macrophage activation but cannot induce the M1 phenotype by itself (170). An additional signaling molecule, such as TNF-α, is required for classical M1 macrophage polarization via Toll-like receptor 4 (90, 135). Alternative activation of M2 macrophages is mediated by IL-4 or IL-13, which oppose the actions of IFN-γ (168). Knowledge of these signaling mechanisms is important because an imbalance in M1/M2 macrophages is implicated in AAA development (20, 150).

Estrogen signaling.

Increased estrogen signaling prevents the infiltration of macrophages into the vascular endothelium. For example, male rat aortas have greater macrophage infiltration compared with female rat aortas after elastase perfusion (1). Male-to-male aortic transplants develop aneurysms after elastase perfusion, but female-to-female aortic transplants remain resistant to aneurysm formation (1). However, female aortas lost their aneurysm-resistant phenotype after transplantation into male recipients. Interestingly, male rats treated with exogenous E₂ exhibited smaller aneurysms and less macrophage infiltrates compared with control rats (1). These observations clearly suggest that female resistance to AAA is, at least in part, due to estrogen regulation of macrophage activity.

As previously stated, TNF-α is a proinflammatory cytokine involved in the activation of M1 macrophages (142). However, TNF-α also induces the expression of inflammatory molecules in ECs through the NF-κB signal transduction pathway (17, 154). Growing evidence indicates that estrogen signaling may reduce infiltration of macrophages into the vascular wall by regulating TNF-α and dampening downstream NF-κB-induced expression of proinflammatory genes in aortic ECs. For instance, an in vivo study using the ANG II-induced AAA model demonstrated that E₂ decreased the expression of ICAM-1, VCAM-1, E-selectin, MCP-1, and macrophage-colony stimulating factor (117). Furthermore, E₂ was shown to decrease VCAM-1 mRNA and protein expression in human ECs through inhibition of NF-κB, activator protein 1, and GATA (164). In vitro estrogen signaling attenuated the expression of MCP-1 (45, 125, 205). These findings hold great significance because monocyte infiltration is strongly associated with AAA pathology (37).

Active VSMCs are characteristic of AAA pathology and also display increased secretion of proinflammatory cytokines and cellular adhesion molecules. Estrogen signaling can inhibit the inflammatory phenotype of aortic smooth muscle cells by enhancing Iκ-Bα synthesis and inhibiting the binding of NF-κΒ to promoters of inflammatory genes that code for molecules such as cytokine-induced neutrophil chemoattractant-2β and MCP-1 (199, 200). For example, estrogen decreased LPS-induced MCP-1 expression in cultured RASMCs in a dose-dependent manner by inhibiting the p38 MAPK/NF-κB cascade (72). RASMCs pretreated with cytokine-induced neutrophil chemoattractant-2 (CINC-2) had blunted TNF-α-stimulated expression of P-selectin, ICAM-1, VCAM-1, MCP-1, and CINC-2β mRNA through the ER-β subtype (199).

Androgen signaling.

As previously stated, monocytes and macrophages express ARs (5). Therefore, differential androgen signaling between male and female subjects may regulate macrophage populations in a way that contributes to the sex difference in AAA. Castration of male rodents suppressed macrophage recruitment to wounds and promoted healing (5, 52). In one study that used a mouse model of cutaneous wound healing, both castration and administration of flutamide, an AR antagonist, resulted in decreased expression of proinflammatory cytokines, such as TNF-α, at the site of the injury. In vitro administration of testosterone enhanced LPS-induced TNF-α expression in macrophages (5). Another study using myeloid-specific AR knockout mice demonstrated that androgen signaling suppressed wound healing through upregulation of local TNF-α and chemokine (C-C motif) receptor 2 expression in infiltrating macrophages, which promote chemotaxis and recruitment of additional macrophages to the site of injury (93). Together, these studies suggest that androgen signaling increases macrophage activity and exacerbates inflammation, thereby contributing to AAA pathogenesis.

CONCLUSIONS

Although research indicates that estrogen protects female subjects from aortic aneurysms by reducing oxidative stress and macrophage-associated inflammation, more studies are needed to determine how estrogen signaling influences T cell activity in the context of AAA. Moreover, prior clinical evidence has indicated that estrogen replacement therapy actually increases the risk of aneurysmal development in women as time increases after the onset of menopause. This highlights a need for more research on age-dependent estrogen effects on AAA development with respect to the expression and functional roles of different ER subtypes.

The preclinical evidence thus far suggests that it is androgen signaling that contributes to aneurysmal development, although clinical studies have suggested that endogenous levels of androgens are inversely related with cardiovascular disease risk in men. It is possible that there is a physiological threshold of circulating androgens in males that is cardioprotective. Thus, larger-scale studies with greater numbers of patients are necessary for studying the impact of androgens on AAA. Also, future clinical studies on AAA need to better adjust for patient age and stage of disease. In addition, further preclinical research should consider 1) using chronic administration of sex hormones at physiologically relevant doses and 2) using pigs or other larger, aged animals to model AAA because these animals have cardiovascular systems that are more comparable to that of humans. Ultimately, to better understand AAA sexual dimorphism and identify therapeutic targets, we need more research into cell signaling pathways that are regulated through different ER subtypes and ARs.

GRANTS

This work was supported by the Tulane University Building Interdisciplinary Research Careers in Women's Health (National Institutes of Health Grant 2-K12-HD-043451-11), 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.B. prepared figures; A.B., J.-P.L., and M.H.H. drafted manuscript; A.B., L.C., K.-J.Y., Y.E.C., J.-P.L., and M.H.H. edited and revised manuscript; A.B., L.C., K.-J.Y., Y.E.C., J.-P.L., and M.H.H. approved final version of manuscript.