Oestrogenic vascular effects are diminished by ageing

The beneficial role of oestrogen in the vascular system may be due, in part, through reduction of peripheral vascular resistance. The use of oestrogen therapy to prevent cardiovascular disease in post-menopausal women remains contentious. This study investigated the influence of the menopause and ageing on the acute vasodilatory effects of oestrogen in ex vivo uterine resistance arteries. Vessels were obtained from young (2.9 ± 0.1 months) and aged (24.2 ± 0.1 and 28.9 ± 0.3 months) female mice and pre-(42.3 ± 0.5 years) and post-menopausal (61.9 ± 0.9 years) women. Ageing was associated with profound structural alterations of murine uterine arteries, including the occurrence of outward hypertrophic remodelling and increased stiffness. Endothelial and smooth muscle function were diminished in uterine (and tail) arteries from aged mice and post-menopausal women. The acute vasodilatory effects of 17β-oestradiol (non-specific oestrogen receptor (ER) agonist), PPT (ERα-specific agonist) and DPN (ERβ-specific agonist) on resistance arteries were attenuated by ageing and the menopause. However, the impairment of oestrogenic relaxation was evident after the occurrence of age-related endothelial dysfunction and diminished distensibility. The data indicate, therefore, that adverse resistance arterial ageing is a dominant factor leading to weakened vasodilatory action of oestrogenic compounds.


Introduction
Cardiovascular diseases (CVDs) are the leading cause of death worldwide (Mozaffarian et al., 2015). Age is the biggest determinant of an individual's cardiovascular health and since the ageing population is growing substantially, there is a need to develop a greater understanding of the influence of age on the cardiovascular system (North and Sinclair, 2012). There are also marked differences in the prevalence of CVD in men and women; prior to 50 years of age, the average age of menopause onset, the risk is considerably lower in women but thereafter the incidence of CVDs is similar (Leening et al., 2014;Lerner and Kannel, 1986). Post-menopausal deprivation of female sex hormones, primarily oestrogen, may be related to the increased CVD risk in ageing women (Kuhl, 2005).
Oestrogen binds to two receptors, ERα and ERβ, which can then initiate ligand-activated transcription factor signalling (Miller and Duckles, 2008). However, oestrogen can also induce acute arterial vasodilation separate from genomic-directed actions (Al Zubair et al., 2005;Bolego et al., 2005;Corcoran et al., 2014;Montgomery et al., 2003;Patkar et al., 2011;Scott et al., 2007;Zhou et al., 2013). It has been proposed that oestrogen activates endothelial nitric oxide synthase (and thus NO production), endothelial prostacyclin and endothelium-derived hyperpolarizing (EDH) factor, which act on adjacent smooth muscle cells to induce relaxation (Al Zubair et al., 2005;Corcoran et al., 2014;Zhou et al., 2013). Indeed, it has been suggested that the contribution of EDH is greater in females due to the presence of oestrogen (Kong et al., 2015;Leung and Vanhoutte, 2017;Wong et al., 2014). Further to acting through the classical ERs, oestrogen has recently been demonstrated to act through the unrelated G-protein-coupled oestrogen receptor (GPER1) to induce vasodilation (Arefin et al., 2014;Broughton et al., 2010;Haas et al., 2009;Lindsey et al., 2009;Lindsey et al., 2013). In large arteries, the effects of oestrogen are considered to protect against inflammatory-related disorders such as atherosclerosis, whereas in the resistance vasculature, the acute vasodilatory actions may lower peripheral resistance and protect against hypertension (Kublickiene et al., 2005).
Consistent with the proposed cardiovascular protective role of oestrogen, early observational studies, such as the Nurses' Health Study, suggested oestrogen therapy might reduce the risk of CVD in post-menopausal women (Grodstein et al., 2000). However, randomised clinical trials largely failed to show any cardiovascular benefits of menopausal hormone therapy (MHT) in post-menopausal women, with guidelines concluding that MHT should not be used for CVD prevention (Duvernoy and Mosca, 2002;Grady et al., 2002;Rossouw et al., 2002). However, subsequent re-analysis of the Women's Health Initiative data revealed a trend for a lower risk of coronary heart disease in hormone-treated women from the youngest age group. This suggested the possibility of a 'timing' hypothesis wherein MHT initiated within 10 years of the menopause could be beneficial to CVD health (Grodstein et al., 2006;Rossouw et al., 2007;Valdiviezo et al., 2013). Indeed, a recent study suggested there to be some beneficial cardiovascular effects of oestradiol treatment in early (< 6 years) menopausal compared to late (≥ 10 years) menopausal women (Hodis et al., 2016). Recent evidence has also highlighted the role for ageing per se, independent of menopausal status, to cardiovascular risk in women (de Kat et al., 2017;Vaidya et al., 2011). Thus, it remains an important objective to elucidate the relative contributions of ageing and the menopause, and the influence of oestrogenic compounds, on vascular function. Therefore, in the present study we have sought to determine the contributions of age or the menopause on arterial function and oestrogenic responsiveness in vitro. To do so we employed two experimental models of arterial vascular function; uterine and tail arteries from young (3 months) and old (24 and 29 months) mice, to determine the influence of age, and uterine arteries from non-pregnant women ranging in age from 32-75 years, to assess the influence of menopause and/or age.

Results
Ageing negatively regulates murine uterine arterial function and passive structural characteristics.
The effect of ageing on passive structural characteristics was more complex. There was no difference in the passive structural characteristics of uterine arteries from 3-and 24-month-old mice. However, further ageing to 29 months increased the passive diameter of uterine arteries ( Figure 1C), which was associated with an increase in cross-sectional area (CSA) ( Figure 1E).
There was also a slight, but non-statistically significant, decrease in the wall-to-lumen ratio ( Figure 1D). Increases in passive diameter and CSA are consistent with outward hypertrophic remodelling. Ageing was also associated with a decrease in the strain: pressure relationship ( Figure 1G) with no change in the stress: pressure relationship ( Figure 1F). Consequently, there was an increased elastic modulus in arteries from aged mice ( Figure 1H), indicating a reduced arterial distensibility (i.e. increased stiffness). This is supported by a leftward shift in the stress: strain curve in uterine arteries from aged mice ( Figure 1H). For the vasoactive experiments, uterine arteries from 3-(n=26), 24-(n=10), and 29-month-old (n=5) mice were mounted on a wire myograph and exposed to the thromboxane agonist U4 (10 -6 M, A).
Following steady-state constriction, endothelium-dependent relaxation was assessed by the addition of acetylcholine (10 -5 M, B). * P < 0.05 from young mice (ordinary one-way ANOVA). For the assessment of passive structural characteristics, separate uterine arteries from 3-(n=5), 24-(n=7), and 29-month-old (n=4) mice were mounted on a pressure myograph and measurements of (C) vessel diameter and wall thickness were used to calculate (D) wall: lumen ratio, (E) CSA, (F) vessel stress and (G) strain at each pressure step (0 -120 mmHg) in calcium-free conditions. P < 0.05 from 24-month-old mice (∇) or 30month-old mice (∆) (repeated measures two-way ANOVA). Vessel stress was plotted against vessel strain to assess the stress: strain relationship, which is used to ascertain arterial distensibility (H). E = Elastic modulus. * P < 0.05 from young mice (unpaired t-test). Data are presented as mean ± SEM.

Oestrogenic responsiveness of murine uterine arteries is impaired in the oldest mice, after the occurrence of endothelial dysfunction and increased stiffness.
We examined the effect of ageing on the vasodilatory responses of resistance arteries to oestrogenic compounds. Figure 2 demonstrates the vasorelaxant effects of 17β-oestradiol (17β), PPT (ERα) and DPN (ERβ) on pre-constricted uterine arteries from 3-, 24-and 29-monthold mice. Ageing was associated with attenuation of the acute vasodilatory properties of all three oestrogenic compounds but the effects were agonist-specific. The acute vasodilatory effects of 17β and PPT, which induced the strongest responses in these vessels, were not different between 3 and 24 months but were reduced when comparing 3-and 30-month-old mice. Therefore, impaired relaxations to oestrogenic compounds occurred after altered functional and elastic arterial properties of these vessels were demonstrable (see Figure 1).

Human arterial function and oestrogenic responsiveness is weakened in post-menopausal women.
To determine whether the menopause influences vasoreactivity of resistance arteries from women, we studied the responses of uterine arteries from pre-and post-menopausal women from the ages of 32 to 75 years (Figure 4). The menopause was associated with impaired U4induced contractility (maximum contractions for pre-menopausal: 21.7 ± 0.9 kPa and postmenopausal: 10.1 ± 0.6 kPa, Fig 4A) and endothelium-dependent relaxations (maximum relaxations for pre-menopausal: 61.5 ± 2.4% and post-menopausal: 40.7 ± 2.6%, Figure 4B) to bradykinin.
Importantly, we found that the acute vasodilatory effects of 17β ( Figure 4C), PPT ( Figure 4D) and DPN ( Figure 4E) were attenuated in uterine arteries from post-menopausal women. Human uterine arterial function is impaired with increasing age.
The data presented in Figure 4 was re-plotted to examine to further examine the effect of ageing on vasoreactivity of uterine resistance arteries from women. Figure

Oestrogenic responsiveness of human uterine arteries is impaired at older ages than changes in endothelial function.
The effects of ageing on the acute responses of human uterine arteries to oestrogenic compounds were similarly assessed in female age groups (years): 30-39 (n = 5), 40-49 (n = 15), 50-59 (n = 8) and 60-75 (n = 6) (using data re-analysed from Figure 4). As shown in Figure 6A- influenced by increasing age. Interestingly, whereas the vasodilatory effect of 17β was impaired in women over the age of 50, a weakening of the effects of the ER-specific agonists appeared in women over the age of 60. Similar to murine arteries, the weakened responses of human arteries to these agonists were only apparent after the occurrence of endothelial dysfunction.
In addition, there was a negative trend between vasodilatory responses of human uterine arteries to oestrogenic compounds and ageing ( Figure 6D&F).

Discussion
The acute vasodilatory effects of oestrogen have previously been established but there is little information about how these arterial responses are influenced by the menopause and advancing age. This study investigated the influence of age and the menopause on the acute vasodilatory effects of 17β and ER-specific agonists. Our data demonstrated that both ageing and the menopause were associated with reduced vasodilatory effects of oestrogenic compounds in murine and human uterine arteries, respectively. Interestingly, however, ageing clearly altered endothelial function prior to oestrogenic responsiveness in human resistance arteries. Similarly, in mice, ageing was associated with reduced resistance arterial compliance, occurring before the altered effects of oestrogenic compounds became apparent. We therefore propose that age-related changes in arterial endothelial function and/or stiffness may be the predominating factor, rather than post-menopausal diminution of circulating oestrogen levels, which leads to impaired oestrogenic actions in arteries from post-menopausal women.
We report that the acute vasodilatory effects of oestrogenic compounds, 17β-oestradiol and ER agonists, are impaired in resistance arteries from post-menopausal women. An important consideration is what contribution ageing per se may make to these findings. Indeed, subgroup analysis of the data suggested that ageing negatively influenced the oestrogenic responsiveness of human uterine arteries. In particular, endothelial-dependent relaxations were greater in women aged 30-39 than those aged 40-49. Therefore, endothelial dysfunction of human resistance arteries occurred prior to the defect in oestrogenic responsiveness, and also the menopause. This is important when compared to the more controlled data from mice, which also demonstrated age-related impaired acute responses, in arteries from two different vascular beds, to 17β-oestradiol, PPT and DPN. Similar reductions in vascular relaxation to 17βoestradiol have been reported for rat vessels (Lindsey et al., 2013;Wynne et al., 2004). Of note, the most pronounced effects of ageing on 17β-and PPT-induced relaxations of murine uterine arteries were observed between 24 and 30 months rather than between 3 and 24 months. This suggests that the impairment of oestrogenic responsiveness occurred after the endothelial dysfunction and altered distensibility, each evident at 24 months, were established.
Furthermore, these changes were also evident before outward hypertrophic remodeling (occurring at 29 months), suggesting changes to wall structure may be a response to worsening arterial compliance of resistance vessels.
Alterations in arterial stiffness and endothelial dysfunction are often observed in unison and the former is a strong independent indicator for subsequent cardiovascular risk (Laurent and Boutouyrie, 2015). Endothelial dysfunction with ageing has been associated with impaired nitric oxide (NO)-and EDH-mediated relaxation, each of which is a potential target of modulation by oestrogenic compounds (Wynne et al., 2004, Chan et al., 2012, Kong et al., 2015. Either way, endothelial dysfunction, or deliberate removal of endothelial-dependent relaxation, resulted in a blunting of oestrogenic responsiveness.
There is currently much debate as to the role of the menopause, and altered oestrogen bioavailability, in determining risks for cardiovascular health and attenuation in oestrogenresponsiveness of the resistance vasculature. The finding reported herein, have implications for the interpretation of clinical studies examining cardiovascular outcomes in ageing females.

Early Oestrogen Prevention Study' (KEEPS) and 'Early versus Late Intervention Trial with
Oestradiol' (ELITE) trials, have been designed to assess if the length of time since menopause alters the effectiveness of oestrogen therapy in post-menopausal women. Early data from the KEEPS trial has been disappointing, with oestrogen therapy showing no significant improvement on cardiovascular mortality, endothelial function or carotid artery compliance in young postmenopausal women (Harman, 2014;Harman et al., 2014;Kling et al., 2015). However, clinical data from the ELITE trial demonstrated that oestradiol treatment had a beneficial effect on carotid artery intima-medial thickness in recently menopausal women (< 6 years) compared with distantly menopausal women (≥ 10 years), yet secondary outcome measures of coronary atherosclerosis risk were unaffected (Hodis et al., 2016).
These outcomes necessitate a continued appraisal of the likely impact of changes in cardiovascular function related to age in women. For example, although long-term risks of adverse cardiovascular outcomes may be similar for post-menopausal women and males there is evidence suggesting key gender differences including: first cardiovascular events occurring at older ages in females and the risk of coronary artery disease less than males (Leening et al., 2014). In addition, apparent age-related gender differences in ischemic heart disease mortality in England, Wales and the United States was attributed to a deceleration of associated events in older males, rather than an acceleration in post-menopausal women (Vaidya et al., 2011).
These scenarios do not fit seamlessly with a notion that the menopause per se imparts enhanced risk to female cardiovascular health.
Taking our in vitro experimental data in context with the above clinical studies, we, therefore, propose that adverse arterial ageing is a predominant factor in influencing the altered vasodilatory effectiveness of oestrogenic compounds. However, since the numbers recruited to the human study were small, we were not able to fully dissociate the effects of the menopause and ageing on the altered effects of oestrogenic compounds. Therefore, there is merit in performing a study, with larger numbers of human participants to explore further the relationship of age-related vascular dysfunction and altered oestrogenic responsiveness. This would require longer than the duration of the current study, and/or the co-operation of multiple sites. Access to other sources of resistance vasculature, such as from gluteal biopsies (Briet et al., 2013;Greenstein et al., 2010), would also extend the interpretations of findings to another human vascular bed. Such studies will be beneficial in determining the feasibility of developing new regimens of oestrogenic agents to treat vascular disease in post-menopausal women.

Experimental procedures Animals
Mouse strains from the C57BL background (C57BL/Ircfa) were housed in the Comparative Biology Centre, Newcastle University, UK. Female mice from three separate age groups (3-, 24-, and 29-month-old) were utilised in the study. Mice were anaesthetised in an induction chamber using a rising concentration of 3% isoflurane/3L oxygen per min. Once anaesthesia was achieved the animal was culled by cervical dislocation, thus confirming death as required by Animals (Scientific Procedures) Act 1986. The oestrous stage of the mice was determined by vaginal smear (Evans and Long, 1922). Vaginal smears from aged mice were characterised by persistent di-oestrous as described previously (Joshi et al., 1993). Small uterine and tail arteries were obtained from mice immediately after culling and placed in ice-cold tissue collection buffer (TCB) (in mM: 154 NaCl, 5.4 KCl, 1.2 MgSO 4 , 10 MOPS, 5.5 glucose, 1.6 CaCl 2 ; pH 7.4).

Human subjects
Ethical approval was obtained from Newcastle and North Tyneside Research Ethics Committee 1 (10/H0906/71) to perform research on samples collected as part of the Uteroplacental Tissue Bank. Women who accepted to take part in the study gave written informed consent in compliance with the Helsinki Declaration. Patient characteristics are detailed in Table 2. Those who had a last menstrual period (LMP) of more than one year before the operation were deemed to be post-menopausal. Uterine arteries (identified on the myometrial surface under a dissecting stereomicroscope, at 1.5 × magnification, Leica Microsystems, UK) from pre-and post-menopausal women were obtained from lower myometrial biopsies (~1 cm 3 ) taken at the time of total hysterectomy, and placed directly into ice-cold tissue TCB.

Isometric force measurements
Arteries were cut into 2-4mm segments and mounted on a small vessel wire myograph (610M; Danish Myotechnologies, Denmark) by inserting two tungsten wires (25μm for mouse arteries and 40μm for human arteries) through the lumen of the vessel. The first wire was attached to a micrometer (allowing adjustment of the lumen diameter during the normalization process) and the second wire was attached to a force transducer (which recorded changes in vessel wall tension). Vessels were normalised to a passive a passive diameter equivalent to 0.9 of L 13.3 kPa, as described previously (Davis and Gore, 1989;Mulvany and Halpern, 1977;Mulvany and Nyborg, 1980), and equilibrated in physiological salt solution (PSS) (in mM: 127 NaCl, 4.7 KCl, 2.4 MgSO 4 , 25 NaHCO 3 , 1.18 KH 2 PO 4 , 0.07 EDTA, 1.6 CaCl 2 , 6.05 glucose bubbled with 95% air/5% CO 2 ; pH 7.4) at 37°C for at least 20 minutes.

Passive structural characteristics
Murine arteries were mounted in a pressure myography system (Living Systems Instrumentation, USA) and allowed to equilibrate for 1 hour in calcium-free PSS (PSS with no added CaCl 2 plus 1mM EGTA) at 60 mmHg. Intravascular pressure was then reduced to 5 mmHg and subsequently increased to 10 mmHg, 20 mmHg and then in 20 mmHg steps to 120 mmHg.
Diameters were allowed to stabilise for 5 minutes before proceeding to the next pressure step.
Measurements of stable intraluminal diameter and left and right wall thickness (taken from 3 separate positions in the field of view, and averaged) were obtained at each pressure.

Statistics
All values were presented as mean ± SEM. Throughout the study, n refers to the numbers of women or mice. The passive structural characteristics of mouse arteries were compared using repeated measures two-way ANOVA ( Figure 1C-G). The parameters that showed significant differences with two-way ANOVA were then analysed with a Bonferroni test for the comparison between age groups. The elastic modulus (E) was compared between murine age groups using an unpaired t-test ( Figure 1H). The vasodilatory effects of increasing concentrations of oestrogenic compounds were compared using repeated measures two-way ANOVA. All maximum relaxations/contractions displayed in bar graphs were compared using an ordinary one-way ANOVA, with the exception of Figure 4, in which an unpaired t-test was used to compare arteries from pre-and post-menopausal women. For the human studies, the trend between ageing and the vasodilatory responses of oestrogenic compounds was also examined using a scatter plot. A linear regression line was plotted and significant deviation from zero slope was determined by an F-test. Analysis was carried out using the GraphPad Prism (6.0) software (La Jolla, CA, USA). Significance was assumed at P < 0.05.