Preservation of Vascular Endothelial Function in Late-Onset Postmenopausal Women.

The data supporting the findings of this study are available from the corresponding author upon reasonable request. All procedures were approved by the University of Colorado Boulder Institutional Review Board. Measurements were taken at the University of Colorado Boulder Main Campus at the Clinical Translational Research Center (CTRC). Written informed consent was obtained, and the intent, purpose, risks and benefits of the measures were explained to all participants in accordance with the Declaration of Helsinki. PMW enrolled in ongoing studies (2016–2024) in the Integrative Physiology of Aging Laboratory were identified through the laboratory database and selected based on prespecified criteria: late-onset PMW were identified with an age at menopause ≥55 years, and normal-onset PMW were identified with an age at menopause of 45–54 years 7 . All PMW were ≥55 years of age at the time of data collection and considered postmenopausal based on self-report of amenorrhea for ≥1 year in a screening questionnaire. Potential participants were excluded if they were on any form of estrogen, testosterone or progesterone hormone therapy or used hormone therapy in the 6 months prior to data collection, had an unknown age at menopause or menopausal status, prior history of an oophorectomy and/or hysterectomy, experienced premature/early menopause (before age 45), or were perimenopausal. A group of PRE were identified in the laboratory database based on age and self-reported premenopausal status and served as a reference group. Exclusion criteria for PRE included: irregular menstrual cycles, unwillingness to abstain or use approved contraception (i.e., hormonal contraception, intrauterine devices, barrier methods such as condoms with spermicide, and surgical sterilization), or were currently pregnant (confirmed with a urine human chorionic gonadotropin test). Menopausal status was confirmed in a subset of PRE and PMW with available serum samples to assess estradiol, follicle-stimulating hormone (FSH), luteinizing hormone (LH), estrone and progesterone. A questionnaire was administered to assess gynecological history including factors related to pre-menopause, menopause, parity and history of female-specific conditions 23 . Years since menopause was calculated based on current age and self-reported age at menopause using the Stages of Reproductive Aging Workshop (STRAW) +10 Guidelines 24 . History of breast cancer was determined by medical history. All PRE and PMW were free from overt CVD as determined by medical history, physical examination, carotid ultrasonography (to detect atherosclerotic plaques), blood chemistries and 12-lead electrocardiogram at rest. Participants were also excluded if they had alcohol dependence, uncontrolled thyroid disease or severe obesity (body mass index [BMI] >40 kg/m 2 ) or were not weight stable for at least 3 months (defined as 2.5 kg change in body mass). Any change in medication within the preceding 3 months was also a basis for exclusion. All measurements were performed following a ≥5 hour fast from food and caffeine, ≥24 hours abstention from alcohol, strenuous physical activity, and marijuana, and ≥48 hours abstention from over-the-counter medications and supplements. All measures in PRE were taken during the early-follicular phase (cycle days 2–5). Participants were advised to continue taking all prescribed medications routinely. Endothelial function was assessed in vivo as endothelium-dependent dilation and measured in PRE and PMW by brachial artery flow-mediated dilation (FMD BA ; expressed as % and absolute change in diameter from baseline) using high-resolution ultrasonography 28 , 29 . Endothelium-independent dilation was assessed as brachial artery dilation (% change) to sublingual nitroglycerin (0.4mg) for 10 minutes 30 . In a subset (n=27) of PMW, FMD BA was assessed before and 1-hour after oral administration of an acute, supratherapeutic dose of the mitochondria-targeted antioxidant MitoQ (160mg) as described previously 17 . The difference in FMD BA from before to 1-hour after acute MitoQ (ΔFMD BA, MTQ ) was taken as the tonic suppression of endothelial function by mitoROS-related oxidative stress 17 . Human aortic endothelial cells (HAECs; PromoCell) were cultured under standard conditions and grown in basal media (Endothelial Cell Growth Medium-2 BulletKit; PromoCell) supplemented with 10% serum 19 , 31 from PRE and PMW for 2 hours. Basal mitoROS activity was measured by fluorescent microscopy. Representative images were chosen based on the participant serum that induced levels of mitoROS bioactivity production closest to each group mean. Serum metabolomics, including lipidomics, non-lipid metabolomics, and oxylipins analyses, were conducted at the University of Colorado School of Medicine Metabolomics Core using ultra-high pressure liquid chromatography-mass spectrometry (UHPLC-MS). A one-way ANOVA was used to assess differences in continuous variables across three groups (with Tukey’s or Šídák post hoc test); a two-tailed Student’s t-test was used to assess differences in continuous variables across two groups. A two-way ANOVA was used to assess the effects of acute MitoQ on FMD BA . Multiple linear regression analyses were used to assess the effect of age at menopause on endothelial function independent of traditional CVD risk factors. Statistical significance was set a priori at α=0.05. P values noted in the text are for pairwise comparisons unless otherwise noted. Additional details on the methods can be found in the Supplemental Methods .

Ninety-two participants were identified for this analysis. Seventy-one PMW were divided into two age-matched groups based on age at menopause ( Table 1 ). Twenty-one PRE served as a reference group ( Table 1 ). There were no significant differences between late- and normal-onset PMW for traditional CVD risk factors ( Tables 1 , S1 ). By design, age at menopause was significantly different between postmenopausal groups (late-onset: 57±1 years; normal-onset: 51±1 years; P =1.1×10 −14 ) ( Table 1 ). In addition, because of the similar chronological age between the normal-onset and late-onset postmenopausal groups, time since menopause was significantly greater in late-onset PMW ( P =0.011). Both late-onset and normal-onset PMW exhibited differences in participant characteristics relative to PRE that are typical of aging and menopause ( Table 1 , S1 ). HDL and physical activity both differed between PRE and normal-onset PMW ( P =0.023, P =0.043, respectively) but not late-onset PMW ( P =0.11, P =0.29, respectively) ( Tables 1 , S1 ). Both late-onset and normal-onset PMW exhibited significantly lower circulating serum estradiol ( P =5.5×10 −7 , 3.7×10 −7 , respectively) and significantly higher FSH ( P =1.3×10 −8 , 3.9×10 −8 , respectively) and LH ( P =6.7×10 −5 , 7.0×10 −4 , respectively) compared with PRE ( Tables 2 , S1 ). Progesterone was not significantly different in late- ( P =0.052) or normal-onset ( P =0.13) compared with PRE ( Tables 2 , S1 ). Estrone was also not significantly different between the 2 groups of PMW compared with PRE (normal-onset PMW: P =0.54; late-onset PMW: P =0.86) ( Table 2 , S1 ). There were no differences in serum hormones between late-onset and normal-onset PMW ( Tables 2 , S1 ). Importantly, gynecological history, including age at menarche, past hormone therapy use (type and route), and prior prevalence of female-specific conditions (including endometriosis, fibroids, breast cancer, etc.) did not significantly differ between late-onset and normal-onset groups ( Tables 2 , S1 ). Number of pregnancies was significantly higher in both late- and normal-onset PMW compared with PRE ( P =6.2×10 −6 , 8.4×10 −4 , respectively; Tables 2 , S1 ). By design, reproductive lifespan was significantly greater in late-onset PMW compared with normal-onset PMW ( P =0.0029) ( Table 2 ). There were no significant differences in serum C-reactive protein ( P =0.84), interleukin-10 ( P =0.23) nor oxidized LDL ( P =0.85) between late-onset and normal-onset PMW ( Table 1 ). Interleukin-6 was significantly lower ( P =0.041) in late-onset versus normal-onset PMW suggesting lower pro-inflammatory signaling ( Table 1 ). FMD BA was significantly lower in both postmenopausal groups compared with PRE (PRE: 8.3±0.7%; late-onset: 6.3±0.6%; normal-onset: 4.1±0.4%; PRE vs. late-onset: P =0.032; PRE vs. normal-onset: P =5.4×10 −8 ) ( Figure 1A ), demonstrating the expected impairment in endothelial function with aging and after menopause. However, FMD BA was 54% higher in late-onset PMW compared with normal-onset PMW ( P =0.0047) such that FMD BA was only 24% lower than PRE in late-onset PMW compared with 51% lower than PRE in normal-onset PMW ( Figure 1A ). Significant group differences were still observed when FMD BA was expressed in absolute units ( P =2.1×10 −8 ) ( Table S2 ). Resting brachial artery diameter and other brachial artery parameters were not significantly different between groups ( Table S2 ). Age at menopause as a categorical variable was an independent, significant predictor of FMD BA ( P =0.0060) in a multiple linear regression analysis including traditional CVD risk factors such as age, blood pressure (BP), BMI, physical activity level, and circulating blood glucose and standard clinical lipids ( Table 3 , Regression 1). In PMW, FMD BA was also positively related to age at menopause as a continuous variable (r=0.32; P =0.0070) ( Figure 1B ), and age at menopause as a continuous variable remained an independent, significant predictor of FMD BA ( P =0.033) in our multiple linear regression analysis ( Table 3 , Regression 2). Taken together, these data indicate that late-onset PMW have higher endothelial function compared with normal-onset PMW independent of other CV- and metabolic-related factors. To determine if differences in time since menopause may be playing a role in the observed differences in endothelial function based on age at menopause, we performed an ANCOVA analysis to statistically control for time since menopause. FMD BA in late-onset PMW remained significantly greater than in normal-onset PMW after accounting for any effects of time since menopause (ANCOVA-adjusted FMD BA values: 6.0±0.6% versus 4.2±0.4%, P =0.011). To further confirm that differences in FMD BA based on age at menopause were independent of time since menopause and to assess if the observed differences were influenced by PMW in the late-onset group possibly being closer to menopause (and the associated acute changes in sex hormones and menopausal symptoms in the early period of post menopause 24 ), we performed additional analyses restricted to participants who were >5 years from the final menstrual cycle (i.e., “late” period of post-menopause, according to STRAW guidelines 24 ). In this subgroup, FMD BA was 44% higher in late-onset PMW (5.9±0.7%) compared with normal-onset PMW (4.1± 0.4%; P =0.046) ( Figure S1 ). Taken together, these data suggest the observed differences in FMD BA between late-onset and normal-onset PMW are independent of years since the menopause transition and not a result of a higher portion of late-onset PMW being closer to the menopause transition. There were no significant differences in brachial artery dilation to nitroglycerin between groups (PRE: 26.5±2.5%; late-onset: 28.5±2.9%; normal-onset: 27.8±2.1%; One-way ANOVA P =0.92), demonstrating the observed differences in FMD BA are endothelium-specific and not due to differences in smooth muscle sensitivity to nitric oxide. Twenty-seven PMW were administered an acute, supratherapeutic dose of MitoQ (160mg) to assess the contribution of mitoROS to the tonic suppression of endothelial function. Participant characteristics in the subset were similar to the larger cohort ( Table S3 ). Acute MitoQ increased FMD BA in both late-onset (pre-MitoQ: 7.2±1.0%; post-MitoQ: 8.6±1.2%; P =4.8×10 −4 ) and normal-onset (pre-MitoQ: 4.3±0.6%; post-MitoQ: 7.1±0.8%; P =4.4×10 −11 ) PMW such that the post MitoQ FMD BA in PMW was no longer significantly different from FMD BA in PRE ( Figure 2A ). However, a group (late-onset vs. normal-onset PMW) × time (pre-post MitoQ) interaction effect was present ( P =0.0033), and the absolute change in FMD BA with MitoQ (ΔFMD BA, MTQ ) was significantly smaller in late- (1.3±0.4%) compared with normal-onset PMW (2.8±0.3%; P =0.0015) ( Figure 2B ). Significant differences in the ΔFMD BA, MTQ persisted ( P =1.5×10 −10 ) when controlling for differences in baseline FMD BA (pre-acute MitoQ) between groups (ANCOVA-adjusted means±SEM: late-onset PMW: 1.9±0.3; normal-onset PMW: 2.9±0.3). These data demonstrate that tonic mitoROS-related suppression of endothelial function is lower in late-onset PMW and suggest reduced mitoROS-associated oxidative stress is a mechanism of higher endothelial function in this group. Similar group differences in FMD BA were observed in the subset of participants used for the in vitro studies as in the full cohort ( Figure S2 ). MitoROS bioactivity was significantly higher ( P =0.016) in HAECs exposed to serum from normal-onset PMW (1.12±0.02 AU) in comparison with serum from PRE (1.00±0.05 AU). MitoROS bioactivity in HAECs exposed to serum from late-onset PMW (0.99±0.05 AU) was significantly lower ( P =0.025) than in HAECs exposed to serum from normal-onset PMW and not significantly different from HAECs exposed to serum from PRE ( P =0.99). MitoROS bioactivity was negatively correlated with FMD BA (r=-0.35; P =0.015) and age at menopause (r=-0.49; P =0.0018) ( Figure 2D and E ). Dye specificity for mitoROS was confirmed ( Figure S3 ).These findings demonstrate the circulating milieu in late-onset PMW is a likely mechanism contributing to lower mitoROS compared with normal-onset PMW. To measure the circulating factors that contribute to age-at-menopause-associated differences in mitoROS bioactivity in serum-treated HAECs, a mass spectrometry-based -omics strategy was employed including polar non-lipid metabolomics, untargeted lipidomics, and semi-targeted oxylipins analyses in serum samples from a subset of PMW (N=21, 8 late-onset PMW, 13 normal-onset PMW). Glycolytic metabolite glyceraldehyde 3-phosphate (G3P, non-lipid metabolomics) was lower and arachidic acid (oxylipins) was higher in late-onset PMW compared with normal-onset PMW ( Figure S4 ). More robust differences were observed with the lipidomics analyses, highlighted by distinct circulating lipidome profiles between late- and normal-onset PMW as shown by separate clusters in a principal component analysis ( Figure 3A ). Analysis of the relative level of individual lipid species to determine the primary lipid metabolites responsible for the overall group differences revealed 15 (out of 496) individual lipids that significantly differed between late- and normal-onset PMW ( Figure 3B and C ). Three were significantly higher and 12 were significantly lower in late-onset PMW ( Figure 3B and C ). The most consistent class of lipid metabolites that significantly differed between groups were TG-related lipid metabolites, with lower levels of 7 TG species in late-onset vs. normal-onset PMW ( Figure 3C ). To identify the specific TG-related lipid species most responsible for differences in HAEC mitoROS bioactivity, the serum exposure experiments were repeated with serum from the subset of PMW in which the lipidomics analyses were performed. Again, lower mitoROS bioactivity in HAECs treated with serum from late-onset PMW compared with normal-onset PMW was observed ( P =0.019) ( Figure 3D ). Correlational analyses were then performed between the 7 TG lipid metabolites that significantly differed between late- and normal-onset PMW and mitoROS bioactivity. Three lipid metabolites were associated with mitoROS bioactivity ( Figure 3E ), with the strongest relation observed between TG(16:0) and mitoROS bioactivity (r=0.52, P =0.016) ( Figure 3E ). To determine the causal role of TG(16:0) as a mechanism of differences in mitoROS bioactivity, absolute concentrations of TG(16:0) were first determined using targeted lipidomics in late- and normal-onset PMW and in PRE. TG(16:0) concentration in serum was significantly higher in normal-onset PMW (135.88±18.49 ug/mL; n=13) compared with late-onset PMW (85.35±11.43 ug/mL; P =0.030; n=13) and higher compared with PRE (97.15±8.64 ug/mL; P =0.071; n=8) ( Figure 4A ). Levels of TG(16:0) were not significantly different between late-onset PMW and PRE ( P =0.62; Figure 4A ). Serum exposure experiments were then repeated with serum from normal-onset PMW, late-onset PMW and PRE in which the concentration of TG(16:0) was normalized across the three groups by adding TG(16:0) to serum from late-onset PMW and PRE at a concentration corresponding to the average difference from the normal-onset PMW. MitoROS bioactivity was not significantly different in HAECs exposed to serum from the three groups with similar levels of TG(16:0) (PRE: 1.00±0.02 AU, n=10; normal-onset PMW: 1.01±0.01 AU, n=13; late-onset PMW 1.05±0.02 AU, n=8; normal-onset PMW versus PRE+TG(16:0): P =0.99 and late-onset+TG(16:0): P =0.22) ( Figure 4B ). In combination, these data support a mechanistic role for circulating lipid metabolites, primarily TG-related metabolites, in modulating endothelial mitoROS bioactivity in PMW and identify attenuated TG(16:0)-related signaling as a mechanism by which late-onset PMW are protected from excess mitoROS-related oxidative stress.

Here, we show for the first time that impaired endothelial function in PMW is attenuated in women who had a natural, late age (≥55 years) of menopause compared with women who had a normal age (45–54 years) of menopause, independent of chronological age and traditional CVD risk factor profile. We also provide novel insight into the mechanisms of higher endothelial function in late-onset PMW, including in vivo and in vitro evidence for lower tonic mitoROS-related suppression of endothelial function. Moreover, we show that a more favorable circulating lipidome profile, particularly lower levels of select TG lipid metabolites, contribute to lower endothelial cell mitoROS in late-onset PMW. Collectively, our findings provide new evidence that age at menopause influences vascular aging in women, as well as biological insight into the mechanistic role of mitoROS-related oxidative stress. Age at menopause was recently recognized by the American College of Cardiology and the American Heart Association as a risk-modifying factor for CVD in PMW 32 , 33 , but the underlying physiological mechanisms are not known. Vascular endothelial dysfunction is a key antecedent to CVD 2 , but the influence of age at menopause on endothelial function has not been established. Here, we show that late-onset PMW are protected from the age- and menopause-associated decline in endothelial function compared with normal-onset PMW, supporting preserved endothelial function as one mechanism potentially contributing to age at menopause-related differences in CVD risk. Indeed, our findings indicate a 2.2 percentage-unit significantly higher FMD BA in late- versus normal-onset PMW – this difference is clinically meaningful, as meta-analyses show that each 1 percentage-unit higher FMD BA is associated with an ~15% lower risk of future CVD 34 . Higher endothelial function in late-onset PMW appears to be largely independent of differences in traditional CVD risk factors, as factors such as physical activity, BMI, BP, blood glucose and LDL cholesterol, were similar between late- and normal-onset PMW. Moreover, age at menopause remained an independent predictor of endothelial function when statistically controlling for these factors with multivariable analyses. Interestingly, our cohort of PMW appeared more healthy in comparison with the general U.S. population of PMW, exhibiting physical activity levels exceeding recommended guidelines for older adults 35 and lower mean BMI values. Although higher physical activity levels may influence the generalizability of our findings, the protective effect of a later age at menopause on endothelial function may be underestimated in our cohort given the largely beneficial effects of physical activity on aging in women. Therefore, PMW with a higher burden of CVD risk factors or comorbid conditions may be afforded even more protection from a late-onset menopause. A similar age at menarche in combination with a later age of menopause led to a longer reproductive lifespan in late-onset PMW. Longer reproductive lifespan due to a later age at menopause is associated with a lower risk for CV events 36 , 37 , which may be related to a longer lifetime exposure to vascular-protective sex hormones. Importantly, we did not observe meaningful and statistical differences in sex hormones, female-specific conditions nor pregnancy-related variables between late and normal-onset PMW. We did observe significantly lower levels of interleukin-6 in late-onset PMW, consistent with reduced pro-inflammatory signaling 38 , an established mechanism of endothelial dysfunction 38 , 39 . Taken together, our observations suggest an independent protective effect of a later age at menopause on endothelial function. As a result of carefully matching the chronological age of PMW in both groups, there was on average a modest (~5 year) difference in years since menopause between groups. Importantly, FMD BA remained higher in late-onset compared with normal-onset PMW after statistically controlling for time since menopause. In addition, although the majority of participants in both groups were in the “late” postmenopausal period (i.e., >5 years from the final menstrual period), we performed additional analyses restricted to PMW in the late postmenopausal period, which is considered a more stable period of post-menopause in which further changes in reproductive function and sex hormones are limited 24 . We observed a 1.8 unit-percentage significantly higher FMD BA in late-onset compared with normal-onset PMW in this subgroup, showing that higher endothelial function in late-onset PMW was not related to a higher portion late-onset PMW being in the more dynamic, early (within 5 years of the final menstrual period) postmenopausal period where symptoms and hormonal changes are still likely occuring 24 . Together, these findings demonstrate that age at menopause influences endothelial function independent from time since menopause. Moreover, the effect of age at menopause on endothelial function extends into the late postmenopausal period and is likely distinct from any acute effects of the menopause transition. We and others have shown that mitochondrial-targeted antioxidant administration improves endothelial function in small groups of older adults that included a mix of men and PMW 17 , 40 , providing some of the first clinical evidence supporting excess mitoROS is a mechanism of age-related endothelial dysfunction. However, these initial studies included only small cohorts of PMW and did not allow investigation of potential sex-specific effects of mitoROS nor factors such as age at menopause. Here we extend our previous observations and, first, demonstrate a role for excess mitoROS-related oxidative stress in suppressing endothelial function in PMW in vivo . Second, our findings indicate that reduced mitoROS-related suppression of endothelial function is one mechanism of higher endothelial function in late-onset PMW, independent of differences in baseline FMD BA . We did not perform our functional bioassay in the reference group of PRE because previous studies have shown that oxidative stress does not tonically inhibit endothelial function in PRE 13 , which is consistent with an absence of an effect of mitochondrial-targeted antioxidants on endothelial function in young adult mice and young healthy humans 41 . In combination, these observations suggest that age- and menopause-related increases in mitoROS contribute to endothelial dysfunction after menopause and that late-onset PMW are largely protected from those adverse effects. We have shown that interventions that reduce mitoROS change humoral factors in circulation to modulate endothelial cell mitoROS bioactivity 19 , 20 . As such, we sought to determine if differences in the circulating milieu played a role in age- and menopause-associated differences in mitoROS-related oxidative stress. Basal mitoROS bioactivity was higher in endothelial cells treated with serum from normal-onset PMW compared with serum from both PRE and late-onset PMW. To identify putative molecular transducers of the effects of age of menopause in serum from PMW, we performed metabolomics analyses as there are distinct changes in the circulating metabolome during and after menopause 42 , 43 . We observed the most robust differences in the serum lipidome between late- and normal-onset PMW, suggesting age at menopause more profoundly influences circulating lipid metabolites in comparison with non-lipid metabolites and oxylipins. Many of the lipid metabolites that significantly differed between groups are independent predictors of CVD risk beyond traditional CVD risk factors, including standard clinical lipid panels 44 . Lipids in circulation may induce mitochondrial dysfunction and promote excess mitoROS 21 , 45 . Accordingly, we focused on determining the mechanistic role of specific lipids in the differences in endothelial mitoROS bioactivity based on age at menopause. The differences in lipid species between late- and normal-onset PMW from our lipidomic analyses were primarily among TG-derived metabolites, which were generally lower in late-onset PMW. To identify the primary TG metabolites responsible for differences in mitoROS bioactivity, we focused on the metabolites that significantly differed most between groups. We then performed correlational analyses between the relative concentrations of those metabolites and mitoROS bioactivity and found the strongest relation with TG(16:0). Normalization of TG(16:0) levels across groups by adding TG(16:0) to serum from PRE and late-onset PMW to match the levels observed in normal-onset PMW abolished group differences in mitoROS bioactivity in serum-treated endothelial cells. These findings provide new insight into circulating lipid-derived metabolites as molecular inducers of endothelial cell mitoROS, which is consistent with other work suggesting dysregulated TG metabolism can induce endothelial dysfunction through upregulation of pro-inflammatory signaling and oxidative stress 46 , 47 . Importantly, our findings implicate the circulating lipidome as a mechanism of differences in endothelial mitoROS based on age at menopause and demonstrate a prominent role for TG(16:0) in mediating these effects. PMW who experienced premature, early, or surgically induced menopause were excluded so that we could isolate the effect of natural, late- versus normal-onset menopause. However, premature, early and surgical menopause are independently associated with an increased risk for CVD 48 – 52 . Thus, future studies should address whether these groups have reduced endothelial function as a potential factor contributing to their heightened CVD risk. Second, our participant population mostly identified as non-Hispanic White; however, PMW who identify as Black, Hispanic or Indigenous are more likely to experience premature/early menopause and the associated chronic disease risks 53 , 54 . Age at menopause-related differences in endothelial function should be assessed in these groups to determine if our observations extend to other PMW who are potentially more susceptible to adverse CV outcomes. Lastly, we did not have access to comprehensive data on menstrual cycle history 55 , which includes important factors that should be evaluated in future studies.

Cardiovascular diseases (CVD) are the leading cause of death in women in the U.S. 1 . In comparison with premenopausal women (PRE), postmenopausal women (PMW) are at markedly higher risk of CVD 2 . Women in the U.S. most commonly complete the menopausal transition between the ages of 45 and 54 years 3 . In contrast, ~10% of women reach the post-menopausal period at the age of 55 or later, which is considered late-onset menopause 4 . Epidemiological studies show that late-onset PMW are at a 10–20% lower risk of CVD compared with normal-onset PMW 5 – 8 , but the physiological mechanisms responsible are unknown. Vascular endothelial dysfunction is a key antecedent to overt CVD after menopause 9 , 10 , and clinical assessment of endothelial dysfunction as shown by reduced endothelium-dependent dilation is an independent risk factor for future CV events in PMW 11 . However, it is not fully understood if late-onset compared with normal-onset PMW are protected from age- and menopause-associated declines in endothelial function. Excessive reactive oxygen species (ROS)-related oxidative stress is a primary mechanism of endothelial dysfunction with aging and menopause 12 , 13 . An important source of vascular oxidative stress that contributes to reduced endothelial function is excess production of reactive oxygen species by mitochondria (mitoROS) 14 – 17 . We have shown in a randomized clinical trial that targeting excess mitoROS with the mitochondria-targeted antioxidant MitoQ improves endothelial function in older men and PMW; however, this trial was not designed to independently assess effects of MitoQ in PMW 17 . As such, whether excess mitoROS is a mechanism contributing to endothelial dysfunction in PMW, and if mitoROS-associated endothelial dysfunction is attenuated in late- vs. normal-onset PMW, remains to be established. An emerging driver of excess endothelial mitoROS is the circulating milieu – the collection of bioactive factors in the bloodstream 18 , 19 . Indeed, we have shown that improvements in endothelial function with interventions that reduce mitoROS-related oxidative stress are accompanied by changes in circulating factors that lower mitoROS bioactivity in endothelial cells in vitro 19 , 20 . However, whether aging and menopause lead to changes in the circulating milieu that drive excess endothelial mitoROS bioactivity and if late-onset menopause provides protection from these changes has not been assessed. Additionally, the identity of the circulating molecular factors that contribute to differences in mitoROS are not completely understood. Small molecule metabolites derived from lipids are drivers of mitochondrial-related oxidative stress as dysregulated lipid metabolism has been implicated in excess production of mitoROS 21 , 22 . The mechanistic role of the circulating lipidome and the specific lipid metabolite species mediating effects of aging-, menopause- and age at menopause-related differences in endothelial mitoROS has not been investigated. Here, we tested the primary hypothesis that late-onset PMW would demonstrate at least a partial protection from age- and menopause-related declines in endothelial function by comparing endothelium-dependent dilation (brachial artery flow-mediated dilation, FMD BA ) in late-onset PMW, normal-onset PMW and a reference group of premenopausal women. Mechanistically, we hypothesized that greater endothelial function in late-onset vs. normal onset PMW would be mediated by lower mitoROS-related suppression of endothelial function as assessed with a unique in vivo functional bioassay. Next, we tested the hypothesis that circulating factors in blood would contribute to differences in endothelial cell mitoROS by determining if 1) serum from normal-onset PMW induced greater endothelial cell mitoROS bioactivity in vitro compared with serum from PRE and late-onset PMW, and whether 2) these circulating milieu-induced levels of endothelial cell mitoROS were dependent on age at menopause in PMW. Lastly, we determined if lipid metabolites were potential circulating molecular transducers of the effects of serum, and if normalizing select lipid metabolites would abolish group differences in mitoROS.

In conclusion, we provide novel evidence that PMW who reached the post-menopausal period at a later age have an attenuation of endothelial dysfunction that is mediated by lower mitoROS-related oxidative stress. Lower endothelial mitoROS in late-onset menopause is in part driven by attenuated circulating lipid metabolites, particularly TG-derived species. Our work provides potentially important evidence supporting age at menopause as a female-specific characteristic that may influence CVD risk, at least in part, by modulating vascular endothelial function in the post-menopausal period. These findings also extend current understanding of how the timing of menopause may influence risk for future chronic diseases and, thus, encourages personalized clinical care of women to promote preservation of vascular function across the lifespan.