Ovariectomy-Induced Arterial Stiffening Differs from Vascular Aging and is Reversed by GPER Activation

Abstract

Background:

Arterial stiffness is a cardiovascular risk factor and dramatically increases as women transition through menopause. The current study assessed whether a mouse model of menopause increases arterial stiffness in a similar manner to aging, and whether activation of the G protein-coupled estrogen receptor could reverse stiffness.

Methods:

Female C57Bl/6J mice were ovariectomized at 10 weeks of age or aged to 52 weeks, and some mice were treated with G protein-coupled estrogen receptor agonists.

Results:

Ovariectomy and aging increased pulse wave velocity to a similar extent independent of changes in blood pressure. Aging increased carotid wall thickness, while ovariectomy increased material stiffness without altering vascular geometry. RNA-Seq analysis revealed that ovariectomy downregulated smooth muscle contractile genes. The enantiomerically pure G protein-coupled estrogen receptor agonist, LNS8801, reversed stiffness in ovariectomy mice to a greater degree than the racemic agonist G-1. In summary, ovariectomy and aging induced arterial stiffening via potentially different mechanisms. Aging was associated with inward remodeling while ovariectomy induced material stiffness independent of geometry and a loss of the contractile phenotype.

Conclusions:

This study enhances our understanding of the impact of estrogen loss on vascular health in a murine model and warrants further studies to examine the ability of LNS8801 to improve vascular health in menopausal women.

Graphical Abstract

INTRODUCTION

Pulse wave velocity (PWV) is a well-documented measurement of arterial stiffness, and elevated PWV is an independent determinant of cardiovascular outcomes.^1,2 While aging-induced arterial stiffness is expected due to regeneration and remodeling across the lifespan, acute injury and chronic inflammatory conditions such as hypertension also contribute.^3,4 Arterial stiffening is often associated with extracellular matrix remodeling such as elastin fragmentation and increased collagen deposition,⁵ but these mechanisms do not fully explain the process behind vascular stiffening.^6,7 The contribution of vascular smooth muscle cells and cell-matrix interactions are less studied but play a significant role in models of aging and hypertension.^8–10

Since aging is a major contributor, the prevalence of arterial stiffening will increase as life expectancy continues to extend.¹¹ Despite this increased risk, arterial stiffness is not routinely measured or treated clinically.¹² Initial work in this field suggested that hypertension preceded arterial stiffness, but recent studies show that the relationship is bidirectional since older individuals with optimal blood pressure and no comorbidities exhibit high PWV.^13,14 Several mechanisms contribute to this phenomenon of “vascular aging”, including adiposity, inflammation, and endothelial dysfunction.^15,16 Because arterial compliance is similar across sexes at birth but diverges after puberty,¹⁷ sex and sex hormones also play a role in arterial stiffening. Recent data from the Study of Women Across the Nation (SWAN) shows arterial stiffness drastically increases in women within one year of the final menstrual period without any alterations in blood pressure, strongly indicating an essential role for estrogen in vascular homeostasis and remodeling.¹⁸

Our group previously showed that PWV is higher in adult male versus female mice, but this sex difference is lost with hypertension or aging.^19,20 Our previous studies also have explored the role of membrane-initiated estrogen signaling in female protection from cardiovascular disease.^21–23 We find that pharmacological activation of the G protein-coupled estrogen receptor (GPER) induces vasodilation, attenuates hypertension and salt-induced vascular remodeling, and promotes diastolic function.^24–30 In contrast, genetic deletion of this receptor increases arterial stiffness only in female mice.^31–33 These studies implicate a role for GPER in arterial stiffness, but the impact of estrogen loss was not assessed. Moreover, while the commercially available agonist G-1 contains multiple enantiomers, the novel GPER agonist LNS8801 is enantiomerically pure and is currently being tested in clinical trials for cancer.^34,35 Therefore, the current study hypothesized that estrogen loss via ovariectomy would recapitulate the impact of chronological aging on arterial stiffening, while pharmacological targeting of GPER would reverse the impact of estrogen loss.

METHODS

All data and materials have been made publicly available at the Harvard Dataverse and can be accessed at [link added upon publication].

Animals

Animal treatments and procedures were in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the Tulane University Institutional Animal Care and Use Committee. All mice were maintained and aged in the Tulane University animal facility. Adult (20 ± 2 weeks) and middle-aged (52 ± 2 weeks) female C57Bl/6J mice (N= 8-10 per group) were housed in a temperature-controlled vivarium with free access to drinking water and standard chow and a standard alternating 12-hour dark to light schedule. Mice were euthanized by exposure to isoflurane and secondary exsanguination.

Ovariectomy (OVX)

OVX was performed under isoflurane anesthesia following sterile technique. Mice were administered 2 mg/kg of meloxicam subcutaneously for pain management and monitored for successful recovery. Uterine horns were ligated with non-absorbable silk suture before removal of the ovaries. The internal muscle layer was sutured, and the outer layer of skin was closed with staples. Successful OVX was confirmed by uterine weight after euthanasia. Sham surgery was performed using the same procedure but without removal of the ovaries. Preliminary experiments tested the impact of sham surgery in both young and aged groups and found no impact on blood pressure or pulse wave velocity (Figure S1). Therefore, the intact group consists of both non-operated and sham-operated animals. Shapiro-Wilk test confirmed normal distribution of combined data, and t-tests found no differences in other parameters between non-operated and sham-operated animals.

Blood Pressure

Tail-cuff plethysmography was used to measure blood pressure in awake mice around using the Kent Scientific CODA^® Noninvasive Blood Pressure System. Measurements were taken for 5 days, including 2 days of training, at approximately the same time each day to avoid differences due to circadian rhythms. Another cohort of intact and OVX mice were implanted with radiotelemetry probes as previously described. as used to confirm tail cuff recordings.³³

Pulse Wave Velocity

Intracarotid PWV (icPWV) and aortic PWV (aPWV) were measured using the VisualSonics Vevo^® 1100 High-Frequency Ultrasound System as previously described.^19,20 Briefly, isoflurane anesthesia was induced and maintained via nose cone in the supine position on a heated EKG platform. The right carotid artery was imaged in Doppler mode at two points: distal to the aortic arch and proximal to the carotid bifurcation. Next, the abdominal aorta was imaged in Doppler mode. Aortic PWV (aPWV) was calculated as the distance measured externally with a caliper between the imaging locations for the carotid and the abdominal aorta divided by the difference in arrival time. For intracarotid PWV (icPWV), the distance between the two carotid locations was measured in VisualSonics Vevo LAB software (v.5.7.0) with PWV measurement package, which places the location of each Doppler measurement onto the B mode image. This precise distance measurement was then divided by the difference in transit time. PWV analysis was conducted by a different, blinded investigator.

Passive Biaxial Pressure Myography

Methods for biaxial pressure myography were previously described in detail. Briefly, carotid arteries were excised and cannulated on 27-gauge needles using 9-0 black mononylon for biaxial pressure myography in Hank’s balanced salt solution, which abolished the active properties of cellular components of the vascular wall.^19,20 Carotids underwent standardized testing to determine the estimated in vivo length which was confirmed where the axial force was nearly constant over a pressure of 10-150 mmHg. Next, vessels underwent three cycles of pressure diameter testing where pressure was applied to the lumen of the vessel and increased from 10-150 mmHg then decreased back to 10 mmHg while outer diameter and was recorded. Stress-strain curves were constructed by calculating stress (σ) as (pressure in dyns/cm²)(inner diameter)/(2*wall thickness) and strain (ε) as (D – Dmin)/Dmin. Beta stiffness (β) was calculated as ln(120/80)/(Dmax-Dmin)/Dmin. Incremental modulus was calculated Δσ/Δε.³⁶

Histology

Aortas from female mice were formalin-fixed, paraffin-embedded, and placed on microscope slides. Verhoeff-Van Gieson (VVG) stain was used to quantify elastin, Masson’s Trichrome (MTC) stained for collagen and smooth muscle, while Alcian blue stained for glycosaminoglycans as previously described.^28,31,37 Analysis was performed using Adobe Photoshop (v.24.5.0). For each tissue section, the entire cross section was imaged, and the number of pixels stained the selected color was counted and expressed as area fraction (percent of pixels with positive staining within the tissue area). Data analysis was performed by a different investigator who was blinded.

Bulk RNA-Sequencing

Bulk RNA-Sequencing was performed using whole aortas of adult female intact (n=4) and OVX (n=7) mice. Aortas were stored in RNAlater (ThermoFisher) at 4°C for 24 hours, then moved to −20°C for long term storage. Total RNA samples were isolated using RNeasy Plus Mini Kit (Qiagen). Nanodrop was used for RNA quantity followed by QA/QC using Agilent 2100 Bioanalyzer. Nextseq mid output (130M clusters) and single read 150 cycles kit was used. Sequences in paired-end fastq files, 35-76bp long, were assessed for quality using FASTQC (v.0.11.7). Reads were pseudoaligned to mouse reference transcriptome (mm10) and expression quantified using Kallisto (v0.46.0). Principal component analysis was used to exclude samples that were not clustering with their respective group. Differential expression was performed using DESeq2 (v.1.38.3). Expression profiles were identified using Gene Set Enrichment Analysis (GSEA, v.4.0.3). Differential splicing analysis was performed using rMATS-turbo (v.4.1.1).

Drug Treatment

A second cohort of mice were randomized to receive one of three drug treatments 8 weeks post-OVX to determine their ability to reverse arterial stiffness (Figure S2). This longitudinal study included PWV measurements at 3 time points: (1) At 10 weeks of age before OVX, (2) At 18 weeks of age, which was 8 weeks post-OVX, and (3) At 20 weeks of age, after two weeks of drug treatment. The three treatment groups were vehicle, the racemic GPER agonist G-1 (Cayman Chemical), or the purified active enantiomer LNS8801 (Linnaeus Therapeutics). The dose of G-1 and LNS8801 was 400 µg/kg/day based on our previous publication.²⁴ G-1 and LNS8801 were first dissolved in dimethyl sulfoxide then combined with an equal volume of 30% ethanol at 37°C to prevent precipitation. Drug or vehicle was added to osmotic minipumps (Alzet Model 1002), primed in sterile saline for 24 h at 37°C, and implanted subcutaneously under isoflurane anesthesia. Mice which did not experience arterial stiffening in response to OVX were excluded from the analysis (N=1-2 per group).

Droplet digital PCR (ddPCR)

RNA was extracted from aortas using the RNeasy Plus Mini Kit (Qiagen 74136), and ddPCR was conducted as previously described ^38,39 using the following validated mouse primers were obtained from Bio-Rad: Acta2 (dMmuCPE5117282), Tagln (dMmuCPE5113593), and Myh11 (dMmuCPE5116710).

Statistical Analysis

All data are expressed as mean ± SEM. Data were analyzed using GraphPad Prism (v.9.5.1). Individual replicates are represented as symbols on graphs and bars represent the mean values for each group. ANOVA P-values are presented in each figure legend, while the results of multiple comparisons tests are presented on each graph. Statistical differences were considered significant at P<0.05. Power analysis was performed with the data from Figure 1A. To detect a difference in PWV of 0.76 m/s with a standard deviation of 0.46 at 90% power with α=0.05, a sample size of 8.7 was predicted.

Figure 1.

Impact of OVX and aging on PWV and blood pressure. Values are mean ± SEM, N=10-11 per group. (A) Intracarotid PWV (icPWV) and (B) Aortic PWV (aPWV) were significantly increased in both OVX and middle-aged groups compared with intact controls, one-way ANOVA, P=0.007 and P=0.01. (C) Systolic blood pressure (SBP) was different by one-way ANOVA, P=0.01, due to a difference between the OVX and middle-aged groups. (D) Pulse pressure was not different between groups, one-way ANOVA, P=0.14. (E) Blood pressure recorded by radiotelemetry also shows no difference between intact and OVX, analyzed by multiple unpaired t-tests.

RESULTS

Neither OVX nor aging significantly impacted body weight, heart weight, or kidney weight (Figure S3A–C). The success of ovariectomy was confirmed by lower uterine weights (Figure S3D), while uterine weights were higher in middle-aged mice, indicating entrance into reproductive senescence.^31,39 Serum was sent to Oregon Health Sciences Endocrine Technologies Core for measurement of estradiol by LC-MS. Estradiol was not detected in OVX serum. The majority of values in intact mice were below the lower limit of detection of 2 pg/ml, as previously reported.^37,40 Including these values at the lower limit, we found that average estradiol levels were 13.2 ± 21.7 pg/ml in intact 20 week-old mice and 2.6 ± 1.1 in intact 52 week-old mice (N=7/group).

Both OVX and aging increased intracarotid PWV (icPWV) and aortic PWV (aPWV) to a similar extent (Figure 1A–1B). Systolic blood pressure as well as pulse pressure was similar across groups by tail cuff (Figure 1C–1D). Another cohort of mice was implanted with telemetry probes to confirm that OVX did not impact blood pressure (Figure 1E). Even when separating day/night data for systolic, diastolic and mean arterial pressures, no differences were found between the intact and OVX groups.

Wall thickness and wall/lumen ratio were significantly greater only in the middle-aged group (Figure 2A–B). This change in geometry was due to a decrease in luminal diameter, with no change in external diameter (Figure 2C–D). Pressure myography showed no differences in distensibility (Figure 3A). We next assessed wall stress generated by increasing pressures normalized to wall thickness to provide an assessment of stiffness that is independent of vessel size, also called material or intrinsic stiffness. The leftward shift in the stress-strain curve in OVX vessels indicated an increase in material stiffness, while this measure was similar in middle-aged versus young intact mice (Figure 3B). Beta stiffness index (Figure 3C) and incremental modulus (Figure 3D) were both calculated between the physiological range of 80-120 mmHg, and again show that OVX but not aging increased material stiffness.

Figure 2. Impact of OVX and aging on vascular geometry.

Values are mean ± SEM, N=10-11 per group. (A) Wall thickness and (B) wall-to-lumen ratio were significantly increased in middle-aged but not OVX mice, one-way ANOVA, both P<0.001. (C) Luminal diameter was significantly decreased in middle-aged mice, one-way ANOVA, P=0.009. (D) External diameter was not different between groups, one-way ANOVA, P=0.61. Significant findings from Bonferroni’s multiple comparisons tests are indicated on each graph.

Figure 3. Impact of OVX and aging on ex vivo arterial stiffness.

Values are mean ± SEM, N=10-11 per group. (A) There were no significant differences in carotid distensibility between groups over the range of physiological pressures, two-way ANOVA, P_group=0.38. (B) A leftward shift of the OVX stress-strain curve indicates increased material stiffness. (C) Beta stiffness index and (D) incremental modulus calculated between the physiological range of 80-120 mmHg shows that OVX but not aging increases material stiffness, one-way ANOVA, both P<0.01. Significant findings from Bonferroni’s multiple comparisons tests are indicated on each graph.

Analysis of the medial layer of aortic cross sections showed no significant differences in elastin content between groups (Figure S4A). Both smooth muscle and collagen content within the medial layer were increased only in the middle-aged group (Figure S4B–C). Alcian blue staining for glycosaminoglycans (GAGs) in the medial layer showed a significant decrease in GAG content in OVX but not middle-aged aortas (Figure S4D). While histological analysis first focused on analysis of the medial layer, we noted changes in the size of the adventitial layer and thus quantified this in all groups using stained sections. In OVX aortas, the medial layer was significantly decreased while the adventitial layer was significantly increased (Figure 4A–B), resulting in no change in total area and confirming the similar parameters obtained from pressure myography experiments (Figure 2). In contrast, aging increased both the medial and adventitial layers to contribute to the overall increase in wall thickness. Despite the fact that OVX and aging had divergent effects on layer composition, both induced a strikingly similar decrease when expressed as the ratio of media to adventitia (Figure 4C).

Figure 4. Impact of Aging and OVX on Aortic Wall Composition.

Values are mean ± SEM, N=10-11 per group. (A) The area of media and adventitia analyzed separately by one-way ANOVA were both P<0.001. Medial area was significantly decreased in OVX but increased with aging while adventitial area was increased by both OVX and aging. (B) Representative images stained with Verhoeff-Van Gieson stain are shown below the graph. (C) Summed data show the contribution of each layer to overall cross-sectional area. (D) The media to adventitia ratio is significantly reduced by both OVX and middle-age, one-way ANOVA, P<0.001. Significant findings from Bonferroni’s multiple comparisons tests are indicated on each graph.

To identify global changes in gene expression that may contribute to the stiffness in OVX aortas that occurred in the absence of vascular thickening, bulk RNA-sequencing was performed. Gene set enrichment analysis (GSEA) identified 272 gene sets that were significantly enriched in intact controls and 13 gene sets significantly enriched in OVX. The top 10 gene sets for each are shown in Figure 5A. Gene sets enriched in intact control aortas in the Reactome Pathway analysis included three pathways on Notch signaling, smooth muscle contraction, gap junctions, extracellular matrix factors, and interleukin signaling. Gene sets enriched in OVX aortas included multiple pathways related to cellular respiration and mitochondrial function, along with striated muscle contraction. A list of the top 50 genes enriched in each condition are shown as heatmaps (Figure 5B). The impact of OVX on smooth muscle contractile genes was of particular interest, as it is known that arterial stiffening can be due to alterations in vascular smooth muscle cell phenotype. We found that in addition to the Reactome smooth muscle contraction gene set, the much larger KEGG smooth muscle contraction gene set showed significant enrichment in intact aortas (Figure 5C). Based on what has been previously published, we looked at individual values for genes associated with the contractile and synthetic phenotypes. OVX induced a significant decrease in contractile genes (Figure 5D), but surprisingly this was not associated with a concomitant increase in genes associated with the synthetic or proliferative phenotype. In fact, OVX also had an overall effect to suppress proliferative genes (Figure 5E).

Figure 5. Gene set enrichment (GSEA) analysis of intact and OVX aortas.

Bulk RNA-Seq was performed on aortas collected from young adult female mice after OVX for 8 weeks compared to aortas from intact mice (control). (A) Reactome enrichment analysis showing the top 10 enriched gene sets in both conditions, all with nominal p value < 0.05. A positive normalized enrichment score (NES) indicates enrichment in intact controls (green), while a negative NES indicates enrichment in OVX (purple). (B) Heat map of the top 50 genes for each phenotype. Expression values are represented as a range from red (high expression) to blue (low expression). (C) Both Reactome and KEGG gene analysis for gene sets associated with smooth muscle cell contraction show enrichment in intact control aortas. (D) Selected contractile genes showed significant downregulation by OVX, two-way ANOVA, P_ovx=0.002. (E) OVX also decreased expression of synthetic genes, two-way ANOVA, P_ovx=0.049. Significant findings from Bonferroni’s multiple comparisons tests are indicated on each graph.

To determine whether GPER activation reverses OVX-induced arterial stiffening, a separate cohort of mice were OVX at 10 weeks and treated from weeks 18-20 with either vehicle, G-1, or LNS8801 (Figure S2). Body weight, heart weight, kidney weight, and uterine weight were not significantly different between any of the groups at the end of the study (Figure S5A–D). Additionally, no changes were detected in systolic blood pressure or pulse pressure due to treatment (Figure S5E–F). OVX significantly increased PWV in all groups by 8 weeks post-surgery, confirming previous data (Figure 6A). G-1 treatment tended to decrease PWV but did not reach statistical significance, while LNS8801 dramatically reduced PWV to levels observed before OVX. Neither G-1 nor LNS8801 impacted vascular geometry (Figure S6) which was not surprising considering that there was also no difference in these parameters between control and OVX. The impact of drug treatments on ex vivo material stiffness matched the trends in PWV, with LNS8801 producing a greater rightward shift than G-1 (Figure 6B). However, LNS8801 did not impact incremental modulus (Figure 6C). LNS8801 tended to show increased distensibility at higher pressures (Figure S7A) and induced a significant reduction in beta stiffness index (Figure S7B).

Figure 6. Impact of GPER activation on PWV.

Values are mean ± SEM, N=8-10 per group. (A) First data point in each group is PWV before OVX, second PWV is 8 weeks post-OVX, and third PWV was after two-week drug treatment. Repeated measures ANOVA showed a significant effect of time, P<0.001, and the results of Holm-Sidak multiple comparisons test are shown on the graph. (B) A rightward shift in the stress-strain curve was observed for both G-1 and LNS-8801 treatment groups. (C) Incremental modulus calculated between 80-120 mmHg was not altered, one-way ANOVA, P=0.60. (D) Contractile gene expression was different between groups, two-way ANOVA, P_gene<0.001 and P_group=0.002. (F) GPER mRNA in aortic tissue was unaffected by OVX or treatment, one-way ANOVA, P=0.27. Bonferroni’s multiple comparisons tests are indicated on each graph.

Histological analysis showed that G-1 but not LNS8801 increased medial area compared with vehicle and LNS8801 treatment, while adventitial area was unaffected (Figure S7C). The media to adventitia ratio was significantly increased in both GPER treatment groups (Figure S7D). Alcian blue staining for glycosaminoglycans showed no significant impact of treatment (Figure S7E). Since aortic wall composition was not changed, we investigated the impact of LNS8801 on the expression of contractile genes identified in RNAseq analysis. The three contractile genes with the highest expression in the aorta were smooth muscle actin (Acta2), transgelin (Tgln), and myosin heavy chain 11 (Myh11). These genes decreased with OVX and were partially restored by two weeks of LNS8801 treatment (Figure 6E). Aortic mRNA expression of GPER was not impacted by OVX or treatment with G-1 or LNS8801 (Figure 6F).

DISCUSSION

The current study found that PWV increased to a similar extent in response to either OVX at 10 weeks or aging to 52 weeks, yet the underlying mechanisms were different. Aging was associated with a significant increase in wall thickness due to inward remodeling and an increase in collagen content, while OVX increased material stiffness while decreasing the smooth muscle layer. The loss of the medial layer in OVX mice was associated with a decrease in smooth muscle contractile genes detected by bulk RNA-Seq. Administration of the enantiomerically pure GPER agonist, LNS8801, effectively reversed the impact of OVX on PWV and material stiffness and was associated with partial restoration of contractile genes despite no changes in wall composition. Taken together, these studies show that hormone decline in females is not merely early vascular aging but induces a different mechanism to promote arterial stiffening.

Intracarotid PWV increased in both OVX and middle-aged mice without a significant change in blood pressure in comparison with intact female mice. This data is consistent with our lab and other groups showing that PWV is higher in aged mice compared to young mice and occurs prior to measurable changes in blood pressure.^15,20 Our results also mimic recent clinical data showing that menopause is associated with significant increase in arterial stiffening without any change in blood pressure.¹⁸ However, blood pressure was significantly higher in middle-aged versus OVX mice, which may contribute to the differential changes in vascular wall properties that accompanied the arterial stiffening. We propose that implementing measurements of PWV may provide opportunities for early intervention against cardiovascular risk in women undergoing the menopausal transition.

The greater wall thickness in aging vessels provided more tissue for the distribution of stress, which explains the lack of change in material stiffness despite an increase in arterial stiffness measured in vivo by PWV. The increased wall thickness was due to a decreased luminal diameter and maintained external diameter, thus enhancing the wall to lumen ratio. Our previous study similarly found that material stiffness did not increase in one-year-old female mice, and while there was a trend for increased wall thickness and decreased lumen diameter it did not reach statistical significance.²⁰ Aging to 36 weeks in mice increases wall thickness in coronary arteries, but in this vascular bed an increase in stiffness was also present.⁴¹

In contrast to aging, OVX induced material stiffness that was independent of changes in vascular geometry. The mechanisms promoting arterial stiffening may differ between estrogen loss and aging, but it is also plausible that stiffening precedes changes in structure and a more prolonged estrogen deprivation would eventually promote increased wall thickness. Few animal studies have investigated the impact of OVX in the absence of other diseases. One study using C57BL/6J mice found decreased strain and Cauchy stress one-month after OVX along with increased intimal thickening.⁴² In rat coronary but not mesenteric arteries, neither OVX nor estrogen impacts vessel diameter but distensibility is greater in the presence of estrogen, similar to the current study.⁴³ These differing results indicate that each vascular bed most likely undergoes multiple remodeling processes after OVX until some level of homeostasis is obtained. Future studies should investigate estrogen loss across more timepoints to fully understand its impact on the vasculature.

Many studies implicate the extracellular matrix in arterial stiffening, yet only the aging group showed an increase in collagen content within the medial layer. However, both the OVX and aging groups displayed an increase in the amount of adventitia surrounding the aorta, which is primarily made of extracellular matrix proteins. We probed for glycosaminoglycans based on our previous finding that aortic remodeling in the rat aorta was associated with changes in this matrix factor and were modulated by GPER activation.²⁸ In contrast, OVX in the current study decreased glycosaminoglycan content within the medial layer. In vitro studies indicate that both 17β-estradiol (E2) and progesterone reduce collagen and enhance elastin deposition in aortic smooth muscle.² Similarly, in ovariectomized primates matrix changes in response to an atherosclerotic diet are ameliorated by estrogen treatment.⁴⁴ Therefore, the impact of estrogen on the extracellular matrix is most likely exacerbated under pathological conditions.

Vascular smooth muscle cells are unique by differentiating and dedifferentiating multiple times throughout their lifetime.⁴⁵ The dogma on this phenotypic diversity is that mature, healthy VSMC mostly exist in their differentiated or “contractile” state but then dedifferentiate into a “synthetic” phenotype to allow proliferation and damage repair.⁴⁶ During cardiovascular disease, this plasticity becomes aberrant and smooth muscle cells remain in the synthetic state, allowing proliferation and promoting arterial stiffening, atherosclerosis, and hypertension.⁴⁷ The switch to this phenotype is characterized by both downregulation of “contractile” proteins such as smooth muscle α-actin, myosin heavy chain, and calponin along with upregulation of “synthetic” markers which include cell cycle and extracellular matrix proteins such as calmodulin and collagen.⁴⁸ While previous studies indicate that estrogen slows the switch from contractile to synthetic phenotype in cultured VSMC,⁴⁹ our data are the first to assess the impact of in vivo estrogen loss on smooth muscle phenotype and do not support this yin and yang of contractile and synthetic markers. Recent single-cell RNAseq analysis indicates that vascular smooth muscle can exist in at least six different phenotypes,⁵⁰ and our preliminary data has perhaps uncovered another distinct VSMC phenotype that occurs in response to estrogen loss.

OVX-induced arterial stiffness was associated with a decrease in contractile genes including smooth muscle actin (Acta2), transgelin (Tagln/SM22α), and myosin heavy chain 11 (Myh11/smMHC). Other studies support that loss of these proteins important for smooth muscle cell contraction are associated with arterial stiffening. VSMC plated on a stiff substrate induces loss of Tagln and Myh11 and stiffening of the VSMC.⁵¹ In angiotensin II-induced hypertensive mice, lower levels of Acta2, smoothelin (Smtn), and calponin (Cnn1) are associated with greater arterial stiffness.⁵² Patients with reduced Myh11 function due to a familial mutation have higher pulse wave velocity despite no difference in aortic dimensions.⁵³ While these studies support the notion that loss of contractile gene expression promotes arterial stiffness, inhibition or knockdown of serum response factor reduces the transcription of contractile genes and is associated with decreased stiffness.^54,55 These differential findings may indicate that an optimal balance between VSMC phenotypes is needed to maintain arterial compliance.

Estrogen binds genomic estrogen receptors ERα and ERβ as well as the G protein-coupled estrogen receptor (GPER). Previous work shows that vascular remodeling is inhibited by decreasing ERα or increasing GPER expression in vivo.⁵⁶ The potential detrimental impact of ERα was also demonstrated in endothelial cells, where genetic deletion of this receptor is protective against high fat diet-induced arterial stiffening.³ Previous work from our group has demonstrated that selective GPER activation induces many protective cardiovascular effects including decreased blood pressure,²⁴ renal protection,²⁶ and beneficial effects on arterial stiffening and remodeling.^28,33 In the current study, we used an enantiomerically pure GPER agonist developed by Linnaeus Therapeutics (LNS8801), which contains only the enantiomer in G-1 that activates GPER and is currently in clinical trials for treatment of cancer (ClinicalTrials.gov Identifier: NCT04130516). We allowed arterial stiffness to develop after OVX for 8 weeks before initiating treatment with either LNS8801 or the commercially available racemic agonist G-1. We found that LNS8801 produced a robust decrease in PWV in OVX mice and was associated with upregulation of contractile genes within the aortic wall. Additional studies are needed to determine the molecular mechanisms by which GPER impacts contractile gene expression.

Our above experiments induce OVX at 10 weeks of age to exclude the interaction of estrogen loss with aging, since we know that aging itself induces arterial stiffness. However, since women approach menopause at the average age of 51, estrogen loss is intertwined with aging, and moreover only 10% of women enter menopause surgically. Additional work is needed to determine whether arterial stiffness is increased when estrogen loss is induced in aging mice and whether an ovary-intact mouse model of menopause using 4-vinylcyclohexene diepoxide (VCD) will produce a similar phenotype.^57,58 While the gold standard approach of telemetry was used in a subset of intact and OVX mice, future studies should also employ this method in aging mice to determine whether small changes in blood pressure parameters precede or follow arterial stiffening after estrogen loss. While both intracarotid and aortic PWV increased to a similar extent, ex vivo experiments such as pressure myography and histology were performed with either carotids or aortas with the assumption that changes occur in parallel. However, it must be noted that changes in function and structure may be different between these two vascular beds.

Lastly, we showed that intracarotid PWV is a reliable method for detecting arterial stiffness in mice and has the potential to be applied clinically. Since estrogen loss increased stiffness without altering vascular geometry, PWV may be superior to methods such as carotid intima-to-media thickness since it provides information on vascular compliance. Utilization of PWV as a cardiovascular risk factor in perimenopausal women, may allow for earlier intervention in this group and reduce mortality.⁵⁹ Additional studies are still needed, however, to determine treatment options for patients with significant arterial stiffening.

Perspectives.

This study demonstrates that estrogen loss induced a decrease in vascular compliance that presented differently than aging. Additionally, we found that selective GPER activation with LNS8801 reversed the effects of estrogen loss, indicating potential use as a therapy for decreased vascular compliance after menopause. Future studies are needed to determine whether this was a temporal change that will eventually promote increased wall thickness similar to aging. This work has translational relevance for women who undergo early menopause, removal of the ovaries, aromatase inhibition, or other estrogen-suppressing therapies.

NOVELTY AND RELEVANCE.

What Is New?

• We found using a mouse model of menopause that similar to chronological aging, estrogen loss induces arterial stiffening in the absence of increased blood pressure.

• The underlying vascular patterns associated with these two conditions were different, with aging inducing the expected increase in wall thickness but estrogen loss inducing material stiffness independent of geometry that was associated with downregulation of contractile gene expression.

• Treatment with a novel estrogen receptor agonist effectively reversed arterial stiffness after estrogen loss and upregulated contractile genes.

What Is Relevant?

• Arterial stiffening is known to occur in response to hypertension but occurs in a blood pressure-independent manner during chronological aging.

• Similar to aging, estrogen loss in young mice increases arterial stiffness independent of changes in blood pressure.

Clinical/Pathophysiological Implications

• Arterial stiffening may be an early indicator of vascular changes after menopause.

• Further studies to examine the ability of LNS8801 may provide a therapeutic option for improving vascular health in menopausal women.

NONSTANDARD ABBREVIATIONS AND ACRONYMS

aPWV

aortic PWV

GAG

glycosaminoglycan

G-1

commercially available GPER agonist

GPER

G protein-coupled estrogen receptor

icPWV

intracarotid PWV

LNS8801

GPER agonist currently in clinical trials

OVX

ovariectomy/ovariectomized

PWV

pulse wave velocity