The effect of transient sex hormone fluctuations on vascular endothelial function

Abstract

This review article summarizes the current literature investigating the effects of transient sex hormone fluctuations on large artery endothelial function, primarily concerning the menstrual cycle and the diurnal rhythm of testosterone secretion. Women and men experience acute fluctuations in circulating levels of sex hormones, and there is substantial variability in circulating levels of sex hormones in both sexes. These acute fluctuations in sex hormones generally coincide with alterations in endothelial cell function and in vivo endothelium-dependent vasodilation, and we see that the timing of these acute fluctuations match the timing of cardiovascular events in both women and men. It is important to improve our understanding of how acute fluctuations in sex hormones affect endothelial function in women and men, as clinical cardiovascular complications coincide with these changes. This would allow for the identification of novel therapeutic targets and aid in the prevention and treatment of cardiovascular disease.

INTRODUCTION

Cardiovascular disease is the leading cause of death in both women and men in the United States (1). Despite this similarity, women have been under-represented in clinical cardiovascular research (2–4). This is especially disparate for premenopausal women (4), because hormonal fluctuations associated with the menstrual cycle were thought to create excessive variability in data, resulting in the exclusion of women from clinical studies. Consequently, there are significant gaps in the current understanding of cardiovascular physiology in premenopausal women, which subsequently impacts our understanding of how this translates to cardiovascular pathophysiology in women (5, 6). This limitation is supported by recent observations that women of all ages have worsened cardiovascular outcomes than men (7), and that younger women burdened with cardiovascular disease now represent a particular and unique high-risk group being failed by the current understanding of clinical cardiovascular care (7).

Markers of vascular endothelial function are independent predictors of cardiovascular risk, events, and mortality (8–10). A commonly used method for measuring endothelial function is through flow-mediated dilation (FMD) of large arteries (e.g., brachial artery, popliteal artery), where blood flow is briefly interrupted via supra-systolic pressure applied to the artery with a cuff for a set duration (e.g., 5 min), followed by release of the cuff and an increase in endothelium-dependent dilation caused by an increase in blood flow and shear stress in the vessel (11). Decades of research have revealed that sex steroid hormones, including estrogens, progesterone, and testosterone, regulate molecular mechanisms underlying healthy endothelial function (12–15). Both women and men experience a reduction in endothelial function with advancing age that is associated with chronic reductions in circulating levels of sex steroid hormones (i.e., menopause and andropause, spanning years) (16–19). Interest in research addressing the effect of acute fluctuations (i.e., changes over days or hours) of these steroid hormones on vascular physiology has increased in recent years but has primarily only focused on the effect of the menstrual cycle in premenopausal women (recent reviews include (20–26)). However, it is equally important to recognize and discuss that men also experience transient fluctuations in sex hormones, such as the diurnal cycle of testosterone secretion (17), that may influence vascular physiology. Importantly, there are clinical observations of increased cardiovascular events that match the timing of these acute fluctuations in sex hormones in both women (e.g., angina, myocardial ischemia) and men (e.g., myocardial infarction) (27–29), but the underlying mechanisms of this pathophysiology remain incompletely understood.

Therefore, this review will summarize the current literature investigating the effects of transient sex hormone fluctuations on large artery endothelial function, primarily concerning (A) the menstrual cycle (Figure 1) and (B) the diurnal rhythm of testosterone secretion (Figure 2). This review will also briefly summarize relevant literature that has investigated the acute effects of sex steroid hormone exposure on endothelial cell function for mechanistic perspective. Finally, this review will further provide new opportunities for the discussion of how we approach experimental design to address research questions targeted to elucidate physiologic effects of acute sex hormone fluctuations.

Figure 1. Observed temporal associations between the menstrual cycle, endothelial function, and cardiovascular risk.

This is a representative image depicting observed temporal associations between hormonal changes (panel A), integrated conclusions from clinical and preclinical research regarding endothelial function (panel B), and the prevalence/incidence of cardiovascular (CV) events (panel C) across phases of the menstrual cycle. Phases of the menstrual cycle are represented as the following abbreviations relative to the day of ovulation: menstrual/early follicular (EF, d −15 to −6), late follicular (LF, d −5 to −1), ovulation (O, d0), early luteal (EL, d +1 to +4), mid-luteal (ML, d +5 to +9), and late luteal (LL, d +10 to +14). Dashed box indicates uncertain evidence. Additional abbreviations: ADMA, asymmetric dimethylarginine. cGMP, cyclic guanosine monophosphate. eNOS, endothelial nitric oxide synthase. GTPCH I, guanosine triphosphate cyclohydrolase I. H₂O₂, hydrogen peroxide. IL-6, interleukin-6. I/R, ischemia reperfusion injury. L-arg, L-arginine. L-cit, L-citrulline. NO, nitric oxide. PGI₂, prostacyclin.

Figure 2. Observed temporal associations between the diurnal cycle of testosterone secretion, endothelial function, and cardiovascular risk.

This is a representative image depicting observed temporal associations between hormonal changes (panel A), integrated conclusions from clinical and preclinical research regarding endothelial function (panel B), and the prevalence/incidence of cardiovascular (CV) events (panel C) across a 24-hr circadian rhythm. Dashed box indicates uncertain evidence. Abbreviations: BP, blood pressure. eNOS, endothelial nitric oxide synthase. ET-1, endothelin-1. K⁺, potassium ion. IL-1β, interleukin-1β. iNOS, inducible nitric oxide synthase. NLRP3, nucleotide-binding domain, leucine-rich repeat, and pyrin domain-containing protein 3. NO, nitric oxide. SNS, sympathetic nervous system.

The hypothalamic-pituitary-gonadal axis and vascular physiology

Sex steroid hormone synthesis and release is regulated by the hypothalamic-pituitary-gonadal (HPG) axis. The HPG axis is the regulated system by which the hypothalamus secretes hormones which interact with the pituitary gland, which in turn secretes hormones that interact with gonadal tissues (e.g., ovaries, testes) to secrete the primary sex steroid hormones estrogens, progesterone, and testosterone. The hypothalamus secretes gonadotropin-releasing hormone (GnRH), which stimulates the secretion of follicle-stimulating hormone (FSH) and luteinizing hormone (LH) from the anterior pituitary, which then interacts with gonadal tissues. In females, FSH induces the development of ovarian follicles and stimulates secretion of estrogens from follicular cells, while LH induces egg release from the ovary (i.e., ovulation) and corpus luteum formation, which secretes estrogens and progesterone. In males, FSH stimulates sperm production in the testes, and LH stimulates Leydig cells in the testes to secrete testosterone (30).

Importantly, both females and males secrete estrogens, progesterone, and testosterone and possess the relative receptors within the vasculature (reviewed in (12, 24)). Several endogenous estrogens circulate in women, but the primary active and most potent estrogen in premenopausal women is estradiol (31, 32). Therefore, the remainder of this review will discuss effects of estradiol specifically. The primary sex steroid hormones circulate through the blood and interact with receptors in various locations around the human body (30). This includes estrogen receptor alpha (ERα), estrogen receptor beta (ERβ), the G protein-coupled estrogen receptor (GPER), progesterone receptors (PR), and androgen receptors (AR), all of which are expressed in human vascular endothelial cells (EC) and smooth muscle cells (SMC) (33–39). This review will primarily focus on estradiol and progesterone effects in women that vary with the menstrual cycle and testosterone effects in men that vary with the circadian rhythm. Additional context for the impact of estrogens in men and testosterone in women on endothelial function has been previously reviewed (12, 24, 40–43).

Fluctuations in estradiol and progesterone and endothelial function

In healthy, young (≤40 y), cycling women, a range of circulating estradiol and progesterone levels are observed at any given time, such that estradiol varies from 58 – 1518 pM (16 – 414 pg/mL) and progesterone varies from 0.3 – 60.4 nM (0.1 – 24.2 ng/mL) across the typical menstrual cycle (18) (Figure 1A). Notably, circulating estradiol and progesterone can reach much greater concentrations during pregnancy (up to 17 times higher estradiol and 14 times higher progesterone concentrations (44)), demonstrating that there is a wide range of physiologically relevant circulating concentrations of these hormones in premenopausal women, though the range is much more narrow across the typical menstrual cycle. Chronic reductions in circulating estradiol and progesterone occur with menopause (18, 45–47). Postmenopausal women have estradiol concentrations that range from 7 – 80 pM (2 – 22 pg/mL) (18) and progesterone concentrations that average 1.2 nM (0.5 ng/mL) (48). It is experimentally difficult to separate the independent effects of menopause and chronological aging, as they occur simultaneously. However, young women who undergo premature ovarian failure (i.e., less than 40 years old, with 4+ months of amenorrhea) have significantly impaired brachial artery FMD compared with age-matched controls (49). In these women, endothelial function is restored with six months of hormone therapy (49). Together, this supports that circulating sex hormones alone contribute to in vivo endothelial function independent of aging. Therefore, it is well-supported that endothelial function varies with chronic changes in estradiol and progesterone. However, the literature is less conclusive on how endothelial function may be impacted by acute fluctuations in hormone concentrations across phases of the menstrual cycle.

The menstrual cycle and in vivo endothelial function

The menstrual cycle includes dynamic changes in estradiol and progesterone. Throughout the menstrual cycle, estradiol and progesterone vary across approximate hormonal phases estimated from the day of peak circulating LH (i.e., ovulation) (50). Parameters of phasic variation in circulating estradiol and progesterone for six approximate time frames across the menstrual cycle, as estimated from Stricker, et al. (50), are displayed and defined in Table 1 and Table 2, respectively, in typical units used in basic science (pM or nM) and clinical (pg/mL or ng/mL) settings. Despite known cyclic variation in circulating hormones across the menstrual cycle, there are also between- and within-individual variations in menstrual cycle parameters. For instance, the menstrual cycle can vary in duration (e.g., number of days) and magnitude (e.g., hormone concentrations) between individuals (recently reviewed in (24)) and within the same individual across cycles over time (51, 52). Thus, there is variability in estradiol and progesterone exposure among women across the menstrual cycle inherently, as well as between cycles (Figure 1A).

Table 1.

Phasic variation in circulating estradiol in young, healthy women across the menstrual cycle.

	Estradiol
	Common units for scientific reporting (pM)				Common units for clinical reporting (pg/mL)
	Absolute Range	Mean 5th percentile	Mean 95th percentile	Phase Mean	Absolute Range	Mean 5th percentile	Mean 95th percentile	Phase Mean
Menstrual/Early Follicular (d -15 to -6)	58 – 313	94	220	155	16 – 85	26	60	42
Late Follicular (d -5 to -1)	190 – 1518	292	863	533	52 – 414	80	235	145
Ovulatory (d 0)	482 – 1425	482	1425	781	131 – 388	131	388	213
Early Luteal (d +1 to +4)	155 – 665	198	527	344	42 – 181	54	144	94
Mid Luteal (d +5 to +9)	267 – 764	282	720	506	73 – 208	77	196	138
Late Luteal (d+10 to +14)	94 – 807	175	637	336	26 – 220	48	174	92

Estimated values from Stricker, et al. (50) in young healthy, non-pregnant women; n = 20. Values for each phase were calculated from time periods based on days from ovulation (defined as day of peak circulating luteinizing hormone). pM, picoMolar. pg/mL, picogram per milliliter.

Table 2.

Phasic variation in circulating progesterone in young, healthy women across the menstrual cycle.

	Progesterone
	Common units for scientific reporting (nM)				Common units for clinical reporting (ng/mL)
	Absolute Range	Mean 5th percentile	Mean 95th percentile	Phase Mean	Absolute Range	Mean 5th percentile	Mean 95th percentile	Phase Mean
Menstrual/Early Follicular (d -15 to -6)	0.3 – 3.8	0.3	1.9	0.9	0.1 – 1.5	0.1	0.8	0.4
Late Follicular (d -5 to -1)	0.3 – 1.6	0.4	1.2	3.3	0.1 – 0.6	0.2	0.5	1.3
Ovulatory (d 0)	1.2 – 4.1	1.2	4.1	2.7	0.5 – 1.6	0.5	1.6	1.1
Early Luteal (d +1 to +4)	2.2 – 45.5	7.4	25.5	16.9	0.9 – 18.2	3.0	10.2	6.8
Mid Luteal (d +5 to +9)	18.3 – 60.4	23.1	51.2	37.3	7.3 – 24.2	9.3	20.5	14.9
Late Luteal (d+10 to +14)	1.9 – 55.9	6.4	40.5	17.9	0.8 – 22.4	2.6	16.2	7.2

Estimated values from Stricker, et al. (50) in young healthy, non-pregnant women; n = 20. Values for each phase were calculated from time periods based on days from ovulation (defined as day of peak circulating luteinizing hormone). nM, nanoMolar. ng/mL, nanogram per milliliter.

The literature investigating the effect of the menstrual cycle on endothelial function has reported inconsistent findings. Several studies have found that brachial artery FMD does not change across phases of the menstrual cycle, including responses in the early follicular phase versus late follicular (52–57), early luteal (55), or late luteal phases (53, 55). Conversely, a number of studies support that endothelial function is reduced in the early follicular phase compared with the late follicular phase, as measured by FMD (58, 59) and internal carotid artery shear-mediated dilation (60). Further studies support that FMD is reduced in early follicular phase versus the mid- to late luteal phase, when measured in the brachial artery (59, 61, 62) and popliteal artery (63). Yet, additional findings support other phase-dependent differences, including that FMD is greater in the late luteal phase versus the early luteal phase (56, 64), is greatest in the late follicular phase compared with all other phases (58), and is lowest in the early luteal phase compared with all others (56). To date, the true nature for how endothelial function is affected by the menstrual cycle is still incompletely understood. The variation in menstrual cycle parameters, including between- and within-individual variation outlined above, coupled with relatively small sample sizes (n=9–26) and inconsistent methods used to experimentally define menstrual cycle phases, likely contribute to the inconclusive nature of findings within this field of study at this time.

A recent meta-analysis conferred that there are minimal changes in brachial artery FMD between phases of the menstrual cycle overall, and variability in the outcomes were attributable to a number of factors, most notably including methodological differences of FMD measurement and hormonal phase assessment (20). This meta-analysis found that large artery endothelial function was significantly increased during the late follicular phase compared with the early follicular phase (standard mean difference, SMD = 0.57, p = 0.0003), but that there was no change in large artery endothelial function between the early follicular phase and the early luteal (SMD = −0.27, p = 0.08), mid-luteal (SMD = 0.69, p = 0.08), or late luteal phase (SMD = 0.04, p = 0.38). Conclusions from this meta-analysis include that studies measuring FMD should prioritize continuous versus discrete diameter assessments, which provide more context to individual variability in responses, and that new strategies of confirming menstrual cycle phase are important for the reproducibility of findings. Additional insights from this meta-analysis include that endothelium-independent dilation does not significantly change across menstrual cycle phase, implicating endothelium-dependent mechanisms in the majority of phase-related differences.

Studies have also investigated whether menstrual cycle phase contributes to how endothelial function may be affected by acute insult or injury. A common experimental model for inducing temporary endothelial dysfunction is ischemia/reperfusion injury (I/R), which mimics clinical phenomena where blood flow is briefly interrupted (ischemia) followed by restoration of blood flow (reperfusion) (65), such as following an acute myocardial infarction or stroke. Experimentally, this results in temporary endothelial dysfunction that is absolved with time (65–67). In response to I/R, peak and total forearm blood flow, another measure of endothelium-dependent dilation, is significantly impaired during the early follicular phase; however, this is completely prevented during the late follicular and mid-luteal phases (66). A similar pattern was found when assessing FMD, such that I/R resulted in reduced FMD during the early follicular phase, but there was no change in FMD following I/R in the late follicular phase (67). This study found a significant inverse relationship between FMD after I/R and circulating estradiol levels, such that as estradiol levels rise, there is a lower likelihood of having blunted endothelial function after I/R (67). Similarly, popliteal FMD is lower during the early follicular phase compared with the late follicular phase after three hours of prolonged sitting; however, this appears to reflect an overall shift to greater popliteal FMD during the late follicular phase compared with the early follicular phase, as the relative change in FMD from pre-sitting to post-sitting was not different between the two phases (63). In contrast, not all acute insults appear to impair endothelial function in a phase-dependent manner in premenopausal women. Acute hyperglycemia blunts FMD after 90 minutes, but this effect is not different between the early and late follicular phases (57). As such, elevated female sex hormones, and likely estradiol specifically based on phasic analyses, appear to provide protection from some, but not all, acute vascular insults and injuries. This is likely related to the molecular pathways mediated by sex hormones and their receptors under different stressors, for which more preclinical and clinical research is needed to fully inform.

Circulating factors known to contribute to vascular physiology also change across menstrual cycle phases, and in some cases, correlate with endothelial function. This includes factors that are proinflammatory and are associated with the nitric oxide (NO) pathway (i.e., a primary pathway responsible for endothelium-dependent dilation), including interleukin-6 (IL-6), L-arginine, and asymmetric dimethylarginine (ADMA). The proinflammatory cytokine IL-6 can inhibit endothelial nitric oxide synthase (eNOS) phosphorylation at Ser1177 in human EC, though it does not appear to affect eNOS protein expression (68). There are greater serum levels of IL-6 during the early follicular phase versus the early luteal phase, but FMD did not vary by phase in this study (69). L-arginine is the biological precursor to NO and, as such, it’s accumulation suggests a mismatch in supply of L-arginine and activity of eNOS to produce NO (70). ADMA is an inhibitor of NO production, because it competitively binds to the active site of NOS, effectively preventing NOS from its normal utilization of the substrate L-arginine (71). Similar to IL-6, both L-arginine and ADMA (as measured by high performance liquid chromatography) are increased during the early follicular phase compared with the late follicular phase (55). Once again, there were no differences in FMD observed in this study across phase, but there was a significant negative correlation between ADMA and FMD overall and during the early follicular phase specifically (55). Prior research has demonstrated that estradiol negatively regulates circulating ADMA, where there was an observed reduction in serum ADMA in postmenopausal women 2 weeks after receiving a 100 mg subcutaneous estradiol implant (72). This finding was also replicated in human endothelial cells in a dose-dependent manner (72). This suggests that when sex steroid hormone concentrations are low during the early follicular phase, L-arginine levels may be elevated because elevations in ADMA are prohibiting it’s use by eNOS, thus affecting arterial endothelial function parameters through reduced NO production. Further, this could suggest potential targets to minimize endothelial dysfunction that accompany phases of low circulating hormones, which may also be applicable in menopause.

Perturbations of the menstrual cycle and in vivo endothelial function

While the literature is mixed regarding the effect of menstrual cycle phase on large artery endothelial function, there is strong evidence that perturbations of the menstrual cycle (e.g., amenorrhea, oligomenorrhea, polycystic ovary syndrome) are associated with large artery endothelial dysfunction (73–87). These perturbations include deviations from the typical hormonal fluctuations associated with the phases of the menstrual cycles, including reduced circulating estradiol or a disproportionate high level of circulating testosterone (88, 89). Still, the effect of other perturbations of the menstrual cycle on large artery endothelial function, such as dysmenorrhea or luteal progesterone insufficiency, has not been studied to date. Overall, as disruptions of typical hormonal fluctuations are well associated with reduced large artery endothelial function, this supports that the typical menstrual cycle promotes healthy endothelial function in premenopausal women.

Amenorrhea is often grouped by etiology, where primary amenorrhea is the condition in which women have no menstrual cycle by the age of 15–16 years old, and secondary amenorrhea is an absence of a menstrual cycle for 3+ consecutive months (88). Much of the literature to date assessing the impact of amenorrhea on FMD has been completed in women that participate in large quantities of exercise with insufficient caloric intake to balance high energy expenditure (79), and, thus this is a limitation to the current body of work. Compared with eumenorrheic women, amenorrheic women have significantly lower brachial artery FMD responses (73–79). In amenorrheic women, worsened FMD is associated with unfavorable lipid profiles and markers of inflammation (74). Additionally, amenorrhea-induced endothelial dysfunction can be improved with folic acid supplementation (76), initiation of oral contraceptive use (75), or the return of a regular menstrual cycle (90, 91). Oligomenorrhea is defined as having infrequent menstrual cycles, with extended cycle duration (35+ days) and reduced cycle frequency per year (e.g., 5–9/year). To date, less studies have directly assessed the impact of oligomenorrhea on FMD. Studies have shown a middling effect of oligomenorrhea, such that some samples of women have a FMD similar to that of eumenorrheic women (73) while others are more similar to amenorrheic women (74), consistent with the nature of this disorder. Small sample sizes remain a barrier to understanding the impact of oligomenorrhea on large artery endothelial function (sample size, n=9–11 (73, 74)). Overall, the literature supports an association between disruption of the typical menstrual cycle and large artery endothelial dysfunction.

Polycystic ovary syndrome (PCOS) is one of the most common endocrine and metabolic disorders in premenopausal women, defined as a combination of ovarian dysfunction, androgen excess, and metabolic dysfunction (89). The etiology of PCOS includes a number of unfavorable alterations in traditional cardiovascular risk factors (e.g., obesity, lipid profile, insulin resistance) alongside hormonal deviations and disruption of the typical menstrual cycle (89). Most of the literature supports that women with PCOS have worsened FMD responses compared with women who do not have PCOS (80–87). However, this finding does not replicate in some samples (92–96), particularly when overweight/obese status is an inclusion criteria for control participants as well or is a variable of study (92–94), highlighting a well-known detrimental impact of obesity on cardiovascular health in women (97). Inflammatory markers, including high sensitivity C-reactive protein (hsCRP) and high-mobility group box 1 (HMGB1) are associated with worsened FMD responses in women with PCOS (82, 98). Several treatments have been shown to improve FMD responses in women with PCOS, including metformin (98–100), folic acid (100), statin medications (85), heat therapy (101), and exercise training (96, 102). Additional research is needed to disentangle the independent effects of the unfavorable traditional risk factor profile and ovarian dysfunction/hormonal deviations both associated with PCOS in the resulting endothelial dysfunction.

Limitations of the current menstrual cycle and in vivo endothelial function literature

Research studies examining the acute influence of the menstrual cycle on cardiovascular physiology have recently increased in number and have provided a wealth of hypothesis-generating data, but some findings still require confirmation by more targeted, intentionally powered investigations. Many of the recent studies in this field have assessed several outcome variables, generally over multiple time points, within cohorts of small sample size (n=9–26). Most of these studies have not reported information to reflect the magnitude of effect (e.g., effect size), which is crucial to understand the clinical importance of findings. Further, many of these studies have either not reported justification for sample size (e.g., a priori power calculation) or are only powered for one of the many variables being experimentally tested. This creates instances where a non-statistically significant p-value fails to leave the reader with an objective interpretation of the presented findings (5, 103, 104).

Consider the following theoretical scenario: data has been collected for two groups of women (lean and obese) across three phases of the menstrual cycle (early follicular, late follicular, and mid-luteal). A sample size of 10 per group was used. The calculated p-value for 1 variable (e.g., brachial artery FMD) is p = 0.08, and 19 other variables were measured in this sample. The manuscript provides no detailed information regarding: (A) the magnitude of the mean difference between groups of women and/or phase of the menstrual cycle, (B) sample size determination, or (C) power calculation for any or all variables measured. Without this information, the reader is unable to determine if the study was appropriately powered to detect a statistical difference in FMD between two groups of women across three phases of the menstrual cycle. Therefore, the question remains if FMD is indeed non-different between groups of women and/or across the menstrual cycle, or if the comparison was under-powered to detect a statistical difference.

There are many challenges to carrying out large-scale menstrual cycle-based research (e.g., time, material resources, funding, participant attrition). Therefore, future investigations in this topical area should strongly consider intentionally planning experimental designs to achieve the appropriate power necessary to detect statistical differences in primary variables of interest. The consistent reporting of magnitude of effect (e.g., effect size) alongside p-value and sample size/power calculations would benefit the interpretation of completed research, the planning of future studies, and rigor and reproducibility within the field.

Effects of estradiol and/or progesterone on endothelial cell function

Estradiol

It is well-known that there is a relationship between estradiol, ERα, and eNOS that collectively contributes to NO production (105–109) and, thus, healthy endothelial function. A concentration of 0.05 – 2.5 nM of estradiol represents a physiologic concentration in premenopausal women across the menstrual cycle (44, 50). In human EC, concentrations of estradiol between 0.1 nM to 1000 nM have been shown to increase eNOS mRNA expression, eNOS protein expression, and NO production in an ER-dependent manner (110–113). One study specifically using more relevant concentrations to the menstrual cycle showed that 0.5 nM estradiol does not increase NO production in human EC at 20-, 40-, or 60-min exposure times, but estradiol does increase NO production at concentrations of 1.75 nM and 3.5 nM at all exposure times (113). Importantly, these effective estradiol concentrations are only relevant during the late follicular and ovulatory phases of the menstrual cycle. All three of these doses of estradiol (0.5, 1.75, and 3.5 nM) increase human EC viability during co-treatment with hydrogen peroxide (H₂O₂) though (113), suggesting estradiol yields protective, anti-oxidant effects in EC at even lower concentrations of estradiol, which could aid healthy endothelial function. Estradiol at 10 nM also increases guanosine triphosphate cyclohydrolase I (GTPCH I) mRNA expression (e.g., the enzyme responsible for tetrahydrobiopterin synthesis, another co-factor for NO synthesis) (110) and at 25 nM increases cyclic guanosine monophosphate (cGMP) release in an ER-dependent manner (111), which is responsible for the ultimate relaxation of vascular SMC in response to NO. Therefore, there is evidence that acute, concentrations of estradiol relevant across the menstrual cycle induce beneficial molecular responses in vascular EC, especially regarding the NO pathway and concentrations that align with those observed during the late follicular and ovulatory phases of the menstrual cycle. However, as briefly discussed here, many investigations have studied EC effects of estradiol at concentrations well above the physiologically-relevant range of estradiol across the human menstrual cycle (i.e., any estradiol concentration above ~2.5 nM) (44, 50). Some of these studies have used concentrations of estradiol relevant to other physiological situations, such as pregnancy, where circulating estradiol can reach up to 42.5 nM (44). This context is important to understand when interpreting the results of such in vitro studies with respect to in vivo EC biology across the menstrual cycle. Therefore, more preclinical work is needed to understand the effect of estradiol at concentrations relevant to the human menstrual cycle on EC biology.

Progesterone

A concentration of 0.3 – 60.0 nM of progesterone represents a physiologic concentration in premenopausal women across the menstrual cycle (50). Progesterone increases NO production in rat aortic EC at concentrations of 1, 10, and 100 nM, which is partially dependent on PR, cyclooxygenase (COX), mitogen-activated protein kinase kinase (MAPKK), and phosphatidylinositol 3-kinase (PI3k) pathways (114). In human EC, 20 nM of progesterone increases NO levels, eNOS activity, and phosphorylated eNOS (34). However, a recent review proposed that the impact of progesterone on NO bioavailability occurs in an inverted U-shaped curve, suggesting that there is an inflection point after which progesterone exposure may begin to yield less NO (115). Progesterone (10 nM) increases prostacyclin production (another endothelium-dependent dilator) and increases prostacyclin synthase mRNA and protein in a PR-dependent manner in human EC (116). Progesterone has also been shown to decrease thromboxane B2 production, which suggests reduced metabolism of the active form of the vasoconstrictor thromboxane A2, in human EC at concentrations of 10 and 100 nM but not 1 nM (116). Thus, acute, concentrations of progesterone relevant across the menstrual cycle induce beneficial molecular responses in vascular EC.

Estradiol + Progesterone

Progesterone is only elevated during the menstrual cycle in conjunction with elevated estradiol; therefore, investigations of the effects of co-treatment of progesterone and estradiol on EC are physiologically relevant and important for increased understanding of EC biology across the menstrual cycle. NO production is greater in rat EC when treated with 10 nM estradiol and 10 nM progesterone in combination versus either hormone alone (117). Similarly, 5 nM of each estradiol or progesterone alone yield no increase in NO production for human EC but do at this concentration when in combination (118), suggesting an additive effect of these hormones in combination. At 5 nM of each estradiol and progesterone in combination, there are also greater increases in COX I and prostacyclin synthase mRNA expression (118). These data suggest that estradiol with progesterone at matched concentrations may produce additive, beneficial results in vascular EC.

Though there is variation across phases of the menstrual cycle, progesterone reaches a much higher concentration than estradiol (e.g., highest concentrations achieved are 2.5 nM estradiol versus 60.0 nM progesterone). Combined 1 nM estradiol and 10 nM progesterone have an additive effect on prostacyclin production compared with progesterone alone (119). However, the same combined dose of estradiol and progesterone yield no difference in nitrite or citrulline production, which suggest the rate of NO synthesis, compared with estradiol alone in human EC (120). These comparisons begin to inform the physiological influence of the combination of estradiol and progesterone across certain phases of the menstrual cycle, but full understanding of the interaction of these hormones at relative concentrations relevant to the menstrual cycle currently remains unknown.

Still, investigations have also used concentrations of these hormones in combination that are not relevant to the menstrual cycle, and the results suggest that a combination over-representing the concentration of estradiol and/or progesterone reverts some beneficial effects of estradiol alone. In human EC, 10 nM estradiol combined with 100 nM progesterone reduces NO formation, eNOS mRNA expression, and GTPCH I mRNA expression versus estradiol alone (110). Combined estradiol (10 nM) and progesterone also results in reduced eNOS protein expression in human EC versus estradiol alone or control cells, but only when progesterone is at concentrations of 1, 10, or 100 nM (110). Therefore, it is important to consider that studies using combinations of estradiol and progesterone that are not relevant to concentrations seen during the menstrual cycle may not accurately reflect in vivo physiology but may begin to inform EC biology under certain perturbations of the typical menstrual cycle.

Integrative summary

Overall, there is low reproducibility between clinical studies investigating the effect of menstrual cycle phase on large artery endothelial function. When integrating clinical and preclinical findings, the majority of evidence supports a reduction in endothelial function during periods of low hormonal exposure (e.g., the early follicular phase) relative to periods of unchallenged increased concentrations of estradiol (e.g., the late follicular phase) (Figure 1B). There remains a low certainty of evidence that endothelial function is changed during the luteal phase relative to other phases of the menstrual cycle, though there is some evidence to support combined estradiol and progesterone exposure (as observed in the mid- to late luteal phases) may increase endothelial function relative to the early follicular or early luteal phases (Figure 1B).

Fluctuations in testosterone and endothelial function

Total testosterone measurements represent all testosterone in circulation, including the ‘inactive’ form bound to circulating proteins, such as albumin or sex hormone binding globulin (SHBG); however, free testosterone represents the amount of testosterone in circulation that is ‘active’ and free to bind to receptors to affect function (121). Notably, testosterone metabolites, such as dihydrotestosterone (DHT), are biologically active and may also influence of vascular physiology (122). In healthy, young (≤40 y) men, a range of circulating testosterone levels is observed, such that total testosterone varies from 9.8 – 45.9 nM (283 – 1324 ng/dL) and free testosterone varies from 0.2 – 0.9 nM (5.8 – 26.0 ng/dL) (123), with significant variability between individuals (Figure 2A). Chronic reductions in circulating testosterone occur with aging (17), but when adjusting for age, total and free testosterone are both positively associated with FMD (19). Neither total nor free testosterone correlate with nitroglycerin-mediated dilation (19), supporting a primary influence of circulating testosterone on EC versus SMC function with regard to functional vascular outcomes. Consequently, testosterone not only varies between individuals and with aging, but it also varies within individual, such as with the circadian rhythm.

The diurnal rhythm of testosterone secretion and in vivo endothelial function

Testosterone synthesis and release exhibit a diurnal rhythmicity, such that circulating levels of testosterone vary across a 24-hr period. In men, circulating testosterone levels peak around 08:00 hr on average, decline throughout the day, reaching a trough in the late evening (20:00 hr on average), at which point, synthesis then gradually increases during sleep (17, 124–127) (Figure 2A). Total and free testosterone demonstrate a 10–25% reduction between 08:00 hr and 16:00 hr, the magnitude of which corresponds with age, such that younger men display greater absolute reductions across this time period (17). This diurnal rhythm of testosterone secretion ensures that testosterone levels align with the body’s daily activity patterns.

Endothelial function has also been shown to change with circadian rhythm. Indeed, in healthy, mixed-sex cohorts, FMD is blunted in the early morning (06:00 hr and 10:00 hr) compared with FMD recorded during late morning, afternoon, or evening (128–130). When assessing even more timepoints, studies have shown greatest FMD responses at 00:00 hr compared with values taken at 4-hr increments throughout the day (129). Similarly, basal forearm vascular resistance is greater and forearm blood flow is lower in the morning (07:00 hr) compared with the afternoon (14:00 hr) and evening (21:00 hr), which is regulated in part by greater alpha-sympathetic vasoconstrictor activity in the morning (131). These findings support that endothelial function is blunted in the early morning and exhibits circadian-associated changes. Future studies might aim to investigate how additional factors may influence endothelial function across the circadian rhythm, such as circadian-related changes in blood flow and blood pressure (132, 133) or potential metabolic influences (134).

Less studies have addressed endothelial function across the circadian rhythm in all male cohorts, and no studies have simultaneously measured circulating testosterone. Studies that have been conducted in all male cohorts generally include smaller sample sizes (n=8–50), and results have varied. In agreement with literature from mixed-sex cohorts, FMD in 15 young men was reduced at 08:00 hr and 12:00 hr compared with FMD at 17:00 hr (135). Further, a study in 12 young men showed that circadian variation in 3,4-dihydroxyphenylglycol, a marker of sympathetic nervous system activation, increases from morning to afternoon and positively associates with FMD (136). However, other studies have shown no significant change in FMD in 2-hr increments between 07:00 hr and 13:00 hr (137) or at 07:00 hr versus 17:00 hr and 22:00 hr (138). However, one of these studies did reveal lower brachial artery diameter and greater endothelium-independent dilation at 07:00 hr versus 17:00 hr or 22:00 hr (138). These findings follow similar trends to suggest vascular impairments in the early morning, but it remains unclear whether this is a result of true functional differences of the vasculature mediated by circadian rhythm or if observed differences are secondary to structural changes in vessel, such as smaller baseline vessel diameter in the morning potentially from heightened sympathetic activity.

Limitations of the current circadian-testosterone and in vivo endothelial function literature

To date, no studies have directly investigated the relationship between circadian-mediated changes in circulating testosterone and endothelial function outcomes in men. However, the collective literature supports that both (A) circulating testosterone concentrations are highest in the early morning in men, and (B) that there is impaired endothelial function in the early morning. Importantly, testosterone is not the only hormone that has increased secretion in the early morning. For instance, cortisol (139) and endothelin-1 (140) also peak in the early morning, and both have also been shown to impair vascular function (recently reviewed in (141, 142)). However, there is also evidence that testosterone can increase the production of both cortisol (143) and endothelin-1 (144), as well as that cortisol can suppress testosterone synthesis (145, 146). Interestingly, in a clinical study, physical activity-induced increases in cortisol are negatively associated with total testosterone but positively associated with free testosterone (147). A relationship between cortisol and testosterone has been previously established in the study of behavior and other aspects of physiology (148, 149), and these findings further support potential dynamic interaction between the hormones in vascular physiology. Future studies might consider experimentally inhibiting cortisol or endothelin-1 synthesis to investigate inter-dependent effects on endothelial function.

Overall, the prior literature generates questions of (A) whether acute, intra-individual changes in testosterone across the circadian rhythm are associated with acute variations in endothelial function in men, and (B) whether that variability is similar to the variability in endothelial function in women across the menstrual cycle. Based on the literature summarized here, acute fluctuations in testosterone may generate different outcomes than chronic reductions in circulating testosterone in men, such that acute, morning-time surges in testosterone may temporally associate with decrements in endothelial function, but chronic reductions in testosterone overtime are well-supported to yield similar decrements.

Effects of testosterone on endothelial cell function

The effect of the diurnal variation in circulating testosterone on large artery endothelial function has not been directly investigated in healthy men, but findings from preclinical studies may provide some context on how varying concentrations of testosterone impact EC biology and function. Clinically, a circulating total testosterone level above 300 ng/dL (~10 nM) is considered normal (150), and levels both lower (151) and higher (such as with the use of anabolic-androgenic steroids, (152)) than that concentration are associated with reduced endothelial function. Relative to observations of the circadian variation in circulating testosterone in healthy men (123), a concentration of 10 – 50 nM represents physiologic concentrations of total testosterone, while 0.1 – 1 nM represents a physiologic range of free testosterone.

In human EC, acute exposure to 1 nM testosterone, increases eNOS activity and eNOS phosphorylation at Ser1177 within 30 min, and nitrite levels are increased within 12 hr of exposure (153). These patterns are similar with exposure to DHT (153), which is a non-aromatizable form of testosterone, suggesting that testosterone directly elicits these effects in EC, as opposed to indirectly via aromatization to estradiol. In rat EC, 5 min of exposure to 1 nM testosterone increases NO production (117, 154, 155) in a manner that is dependent on AR and 5α-reductase (155), as well as protein kinase C (PKC) and MAPKK pathways (154). Further, testosterone significantly impacts K⁺ efflux from EC, including through small conductance (SK_Ca) and large conductance calcium-activated potassium (BK_Ca) channels (156). This pathway is also AR-dependent, ERα-independent, and responses are similar with DHT (156), once again supporting a direct effect of testosterone on this mechanism of EC function. Thus, acute testosterone exposure that replicates healthy, physiologically-relevant concentrations of free testosterone induces beneficial molecular responses in vascular EC.

Conversely, literature supports that higher concentrations of testosterone can induce EC dysfunction. Young male mice with an average circulating testosterone level of 70 nM (via exogenous supplementation) have reduced endothelial function and increased expression of inflammatory proteins NLRP3 and IL-1β in the vasculature relative to control mice (157). In castrated rats treated with exogenous testosterone, rats with average peak circulating exposures of 75 nM and 225 nM testosterone have reduced endothelial function relative to control rats, as well as reduced vascular mRNA expression of beneficial proteins such as eNOS, sirtuin-1, and vascular endothelial growth factor and increased mRNA expression of the deleterious inducible NOS (iNOS) (158). Further, 100 nM testosterone increases ROS generation, decreases phosphorylation of eNOS, and counteracts bradykinin-induced NO release in human endothelial cells (159). Even higher concentrations of testosterone (1000 nM) reduce eNOS mRNA expression in human EC, which is restored with co-treatment of an antioxidant (160). Accordingly, this concentration of testosterone reduces the mRNA expression of antioxidant enzymes, including catalase, glutathione peroxidase, and superoxide dismutase in human EC (160). Therefore, supraphysiological concentrations of testosterone have a negative impact on EC biology. It is currently unclear how the transient, in vivo fluctuation of circulating testosterone directly impacts endothelial function, and additional clinical and preclinical research is necessary for further understanding.

Integrative summary

Overall, there is strong clinical evidence that large artery endothelial function is impaired in the early morning (between 06:00 and 12:00 hr) relative to other times throughout a 24-hr day (Figure 2B); however, no study to date has directly investigated the association with diurnal variations in circulating testosterone and large artery endothelial function. Preclinical findings support that mid-range levels of testosterone promote healthy endothelial function, while high to supraphysiologic levels have shown negative impacts on EC function (Figure 2B). Studies have yet to focus on how the higher range of physiologically-relevant concentrations of testosterone (~50 nM) affect EC function, as well as the potential interaction of temporally associated factors (e.g., circulating endothelin-1 and cortisol, blood pressure, vessel diameter, and sympathetic nervous system activation (Figure 2B)).

CLINICAL IMPLICATIONS AND FUTURE DIRECTIONS

The clinical importance of understanding the role of acute fluctuations in sex hormones in the cardiovascular system is apparent when we observe patterns in clinical cardiovascular data. In premenopausal women, variant angina and myocardial ischemia present quicker and more severely in response to stress during phases of the menstrual cycle when both estradiol and progesterone concentrations are low (i.e., the early follicular phase) (27, 28). Even without induced stress, premenopausal women with variant angina document a higher frequency of ischemic episodes during low hormone phases (28) (Figure 1C). Additionally, men experience more cardiac events compared with age-matched, premenopausal women, but there are circadian patterns of cardiovascular events, such that there is a 40% higher risk of myocardial infarction, 29% higher risk of cardiac death, and 49% higher risk of stroke between the hours of 06:00 hr and 12:00 hr (29) (Figure 2C). Therefore, acute fluctuations in sex hormones may not only affect vascular function in young, healthy individuals but may also have prominent, life-changing implications for cardiovascular health and what contributes to cardiovascular morbidity and mortality. Future studies should prioritize understanding the typical physiologic changes in vascular function that accompany acute fluctuations in sex hormones in women and men, including during the menstrual cycle and the diurnal cycle of testosterone secretion. This knowledge will help inform the mechanistic underpinnings of how acute fluctuations in sex hormones may contribute to cardiovascular events and complications in both women and men, in addition to providing novel therapeutic targets for the prevention and treatment of cardiovascular disease.

There are a number of additional considerations for interpretation of this literature review and for future research directions. First, testosterone also has physiological effects in women, and, likewise, estradiol in men (reviewed in (12, 40–43)), which should be considered and explored in future research. Second, new research has begun to indicate a specific role for FSH in vascular outcomes (47), highlighting an interesting new area for further investigation. Third, hormonal variations in women were discussed here with respect to the menstrual cycle, and likewise men with respect to the circadian rhythm. Research suggests minimal influence of the circadian rhythm on ovarian hormone release in women (161), and some data suggests testosterone may vary by month of the year (potentially related to climate aspects) (162, 163); therefore, the possibility remains that there may be variability in hormonal exposures across these alternate time frames as well. Finally, the present review focused on large artery endothelial function, primarily brachial artery FMD, but there is further need to understand how transient fluctuations in sex hormones via the menstrual cycle and circadian rhythm affect additional levels of the vascular tree, especially the microvasculature (5, 20, 164, 165), which can be more sensitive to earlier perturbations in endothelial function.

CONCLUSIONS

There are acute fluctuations in the circulating levels of sex hormones in both women and men, such as during the menstrual cycle and the diurnal cycle of testosterone secretion. Additionally, there is substantial inter- and intra-individual variability in the circulating levels of estradiol and progesterone in women and testosterone in men. These acute fluctuations in sex hormones generally coincide with alterations in endothelial cell function and in vivo endothelial function, and we see that the timing of cardiovascular events matches the patterns of these acute fluctuations in both women and men. Therefore, greater overall understanding of how typical acute fluctuations in sex hormones affect endothelial function in both women and men should be prioritized. This knowledge would inform the observations of clinical cardiovascular complications that coincide with acute hormonal fluctuations and allow for novel, precision medicine approaches for the prevention and treatment of cardiovascular disease.

DATA AVAILABILITY

This article contains no datasets generated or analyzed.