Endothelial Mineralocorticoid Receptors Contribute to Vascular Inflammation in Atherosclerosis in a Sex-Specific Manner

Abstract

OBJECTIVE:

Mineralocorticoid receptor (MR) activation is associated with cardiovascular ischemia in humans. This study explores the role of the MR in atherosclerotic mice of both sexes and identifies a sex-specific role for endothelial cell (EC)-MR in vascular inflammation.

APPROACH AND RESULTS:

In the AAV-PCSK9 mouse atherosclerosis model, MR inhibition attenuated vascular inflammation in males but not females. Further studies comparing male and female littermates with intact MR or EC-MR deletion revealed that although EC-MR deletion did not affect plaque size in either sex, it reduced aortic arch inflammation specifically in male mice as measured by flow cytometry. Moreover, MR-intact females had larger plaques but were protected from vascular inflammation compared to males. Intravital microscopy of the mesenteric vasculature demonstrated that EC-MR deletion attenuated TNFα-induced leukocyte slow rolling and adhesion in males, while females exhibited fewer leukocyte-endothelial interactions with no additional effect of EC-MR deletion. These effects corresponded with decreased TNFα-induced expression of the endothelial adhesion molecules ICAM-1 and E-selectin in males with EC-MR deletion compared to MR-intact males and females of both genotypes. These observations were also consistent with MR and estrogen regulation of ICAM-1 transcription and E-selectin expression in primary cultured mouse ECs and human umbilical vein ECs.

CONCLUSIONS:

In male mice, EC-MR deletion attenuates leukocyte-endothelial interactions, plaque inflammation, and expression of E-selectin and ICAM-1, providing a potential mechanism by which the MR promotes vascular inflammation. In females, plaque inflammation and leukocyte-endothelial interactions are decreased relative to males and EC-MR deletion is not protective.

Introduction

Despite improvements in care and preventative strategies, cardiovascular complications of atherosclerosis are still the leading cause of mortality in men and women in the USA.¹ Increased activation of the aldosterone-binding mineralocorticoid receptor (MR), a hormone-activated transcription factor, is associated with a significantly increased risk of cardiovascular ischemic events in humans.^2–4 The majority of these events are caused by rupture and thrombosis of atherosclerotic plaques, resulting in ischemia of downstream tissues.⁵ Animal models reveal that MR activation promotes atherosclerosis, and conversely pharmacologic MR inhibition attenuates atherosclerosis, without changes in blood pressure.² However, one critical limitation of the preclinical literature is that most studies explored the role of MR in atherosclerosis only in male animals.² Scarce reports that include both sexes combine those data, obscuring any potential sex-specific actions of the MR in atherosclerosis.⁶ Similarly, clinical studies showing an association between MR activation and cardiovascular ischemia do not separate men and women.^3,4 Epidemiologic data have consistently shown that premenopausal women have a lower incidence of acute myocardial infarction than age-matched men.⁷ This difference disappears after menopause, implicating the sex steroid hormone estrogen presumably acting through its two estrogen receptor (ER) isoforms, ERα and ERβ.⁸ However, a potential role for the MR in this process has not been studied.

Human pathology data identify inflammation as a key driver of plaque rupture in humans, as ruptured atherosclerotic plaques are characterized by extensive inflammatory cell infiltration at the site of rupture.⁹ Therefore, substantial investigation has focused on understanding mechanisms driving atherosclerotic plaque inflammation. Inflammatory cells enter the vascular wall through a coordinated process of rolling, adhesion, and trans-endothelial migration through post-capillary venules. These interactions are mediated in part by adhesion molecules on endothelial cells (ECs), including E- and P-selectin and intercellular adhesion molecule (ICAM)-1.¹⁰ We previously found that low-dose aldosterone infusion increases plaque burden and inflammation in male atheroprone apolipoprotein E knockout (ApoE^−/−) mice, independent of changes in blood pressure.¹¹ Conversely, pharmacologic MR inhibition has been shown to reduce markers of inflammation in male atheroprone mice.² We also reported that the MR transcriptionally regulates ICAM-1 in cultured human coronary ECs¹² and that ICAM-1 is necessary for aldosterone-induced atherosclerosis in male ApoE^−/− mice.¹³ Further, we recently showed that endothelial-specific MR (EC-MR) differentially contributes to obesity-induced endothelial dysfunction in males and females.¹⁴ However, the role of the MR in vascular inflammation in females and the specific role of EC-MR in atherosclerosis have not been explored.

To address these gaps in knowledge, we used the mouse model of atherosclerosis induced by constitutively-active mutant PCSK9 to test the role of the MR in atherosclerosis in males and females and explored the specific role of EC-MR in atherosclerosis, vascular inflammation, and leukocyte recruitment in both sexes. By this strategy, we discovered a sex difference in the role of EC-MR in the vascular inflammation associated with atherosclerosis.

Materials and Methods

All animal models, antibodies, and cell lines used are detailed in the Major Resources Tables in the Supplemental Material.

Mouse models

Mice were handled in accordance with US National Institutes of Health standards. All procedures were approved by the Tufts University Institutional Animal Care and Use Committee. C57BL/6J mice were purchased from The Jackson Laboratory (Bar Harbor, ME). For survival surgeries and tissue harvests, mice were anesthetized by continuous inhalation of 2.5% isoflurane gas. Prior to subcutaneous pellet implantation surgery, mice were administered a subcutaneous injection of 0.05mg/kg buprenorphine (Reckitt‐Benckiser).

The animal studies reported herein were conducted according to the guidelines for experimental atherosclerosis studies described in the recent American Heart Association statement.¹⁵ Previous atherosclerosis studies in our lab indicated that 7–10 animals per group were needed to achieve sufficient power to detect differences between groups. Each experiment was carried out in at least 3 separate batches to ensure reproducibility, with all treatment groups represented in each batch. In the spironolactone study, mice were randomly assigned to either the spironolactone or placebo group after administration of AAV-PCSK9. Mice in the spironolactone study were excluded from analysis if they demonstrated signs of infection or abscess formation at the site of pellet implantation. At the end of all studies, mice were euthanized in a random order for tissue collection. In flow cytometry experiments, the entire aortic arch digest was acquired for quantification of the total number of inflammatory cells per aortic arch. Therefore, data points were excluded when there was evidence of technical errors with the flow cytometer that prohibited an accurate cell count, such as a clog or bubble in the fluidics line. No data was excluded from the histological analysis.

For the EC-MR deletion studies, we used our previously described mouse strain in which mice bearing two floxed MR (MR^fl/fl) alleles were crossed to mice with a Cre recombinase driven by the VE-Cadherin promoter (MR^fl/fl × VE-Cadherin-Cre^+/−). Previously, we found that these mice exhibited a 75% reduction in MR expression by qPCR specifically in isolated ECs, while MR expression in splenic leukocytes was unaffected.¹⁶ Further analysis of the tissues of these mice by endpoint PCR indicated no MR recombination in the splenic leukocytes or lymph nodes of these mice in multiple experiments.^16,17 We therefore deemed this mouse model to represent an endothelial-specific deletion of the MR, as we did not detect any recombination in hematopoietic cells isolated from those immune organs.

Notably, a triple genetic cross was first attempted between those MR^Fl/Fl × VE-Cadherin-Cre^+/− mice and the ApoE^−/− mouse (The Jackson Laboratory, Bar Harbor, ME) on the C57BL/6J background. While the genetic cross was successful, as confirmed by genotyping, there was no evidence of MR gene recombination in the lung, a tissue rich in ECs, in the MR^Fl/Fl × VE-Cadherin-Cre^+/− × ApoE^−/− mice (Supplemental Figure I–A). We therefore proceeded with the AAV-PCSK9 model of atherosclerosis that has previously been described.¹⁸ Analysis of DNA from the lungs from these mice after AAV-PCSK9 injection and 12 weeks of high fat diet (HFD) confirmed recombination of MR in lung tissue from both sexes specifically in Cre-positive animals (Supplemental Figure I–B).

Tail cuff blood pressure measurements, animal euthanasia and tissue harvest, aortic root histology, serum isolation, and serum glucose, cholesterol, and aldosterone measurements were performed as previously described.¹⁹ Mice were fasted for 4 hours prior to sacrifice to obtain serum for fasting glucose and cholesterol levels (Supplemental Table III). Tumor necrosis factor (TNF)α levels in these serum samples were measured using the R&D Systems Mouse TNFα Quantikine ELISA Kit (MTA00B).

Induction of atherosclerosis with constitutively-active AAV-PCSK9

Concentrated stocks of adeno-associated virus expressing mouse recombinant proprotein convertase subtilisin/kexin type-9 (PCSK9) with gain-of-function mutation D377Y (“AAV-PCSK9”)²⁰ were obtained from the Boston Children’s Hospital Viral Core and stored at −80°C. Mice were administered a single tail vein injection of 5×10¹¹ viral particles in a 100μL volume of sterile phosphate-buffered saline and immediately began 12 weeks of high fat diet (HFD, Envigo TD.88137).

In vivo systemic MR inhibition

Immediately after viral injection, 8-week-old male or female C57BL/6J mice were randomized to subcutaneous implantation with a slow-release 60-day pellet containing placebo or 20mg/kg/day of spironolactone (Innovative Research of America). After 60 days, a second 60-day pellet was implanted into all mice to extend the treatment period to 12 weeks.

Flow cytometry

At the time of animal sacrifice, aortic arches were harvested, carefully cleaned of adventitia, and digested to a single-cell suspension for flow cytometry as previously described.¹⁹ Gates were set using single-stained spleen controls acquired at the same time as experimental samples. Identical gates were applied to all samples in an experimental batch. Representative dot plots demonstrating the gating strategy are displayed in Supplemental Figure II.

Isolation of mouse lung endothelial cells (mLECs)

ECs were isolated from the lungs of 3-week-old mouse pups as previously described.¹⁶ Two male and two female pups were pooled for each of the EC-MR^+/+ and EC-MR^−/− cultures. To assess the purity of mLEC cultures, cells were analyzed using Di-I-Acyl-LDL reagent (Alfa Aesar J65597-&H) according to the manufacturer’s instructions (Supplemental Figure III–A). This indicated that the resulting cultures predominately contained ECs, and a 60% reduction in MR expression was observed by qPCR in EC-MR^−/− cultures compared to EC-MR^+/+ (Supplemental Figure III–B) using the following primers:

MR

• Forward: 5’-GAAAGGCGCTGGAGTCAAGT-3’

• Reverse: 5’-TGTTCGGAGTAGCACCGGAA-3’

GAPDH

• Forward: 5’-AGGTCGGTGTGAACGGATTTG-3’

• Reverse: 5’-TGTAGACCATGTAGTTGAGGTCA-3’

Cells, reagents, and in vitro assays

Human umbilical vein endothelial cells (HUVECs) were maintained as previously described.²¹ Spironolactone (Sigma) in dimethylsulfoxide (Sigma) and estrogen (Sigma) in 200-proof ethanol (Fisher) were diluted to 1μM and 10nM, respectively, in M199 media (Gibco) for application to cells. For stimulation of E-selectin, cells were pretreated with spironolactone, estrogen, or appropriate vehicle(s) for 1 hour and then stimulated with low-dose TNFα (80pg/mL, Biolegend) for 20 hours. Otherwise, spironolactone and estrogen were applied to cells for 20 hours. All samples contained the same concentration of DMSO and/or ethanol. Leukocyte adhesion and luciferase assays were performed as previously described.^12,21 Each cell treatment was carried out in triplicate, and each experiment was performed a minimum of three times.

Protein analysis of mesenteric veins and cultured cells

For mesenteric vein samples, mice were injected intraperitoneally with TNFα (0.3μg/mouse in sterile phosphate-buffered saline), and four hours later the mesenteric veins were harvested and dissected in in RNA Later (Sigma) on ice and snap frozen in liquid nitrogen. Vessels from two mice were pooled to obtain one sample for immunoblotting. The frozen mesenteric veins were pulverized on dry ice and lysed in sample buffer containing 100mM Tris-HCl (pH 6.8), 200mM 1,4-dithiothreitol, 4% sodium dodecyl sulfate, 0.2% bromophenol blue, and 20% glycerol.

Cultured mLECs and HUVECs were lysed using the same sample buffer recipe. Lysates were resolved by SDS-PAGE, transferred to PVDF membranes (Millipore), blocked in 5% milk solution, and probed with primary and secondary antibodies (see Major Resources Tables). Human GAPDH was used as a loading control for all HUVEC immunoblots and mouse β-tubulin was used for all mouse tissue and cell immunoblots. The MR positive control in Supplemental Figure VII was a lysate from a Pac1 cell line overexpressing human MR from HA-CMX plasmid.²² Recombinant ERα and ERβ proteins (Calbiochem) were also used for positive controls in the same figure. Protein bands were detected by enhanced chemiluminescent reagent (Fisher) and quantified in a blinded fashion.

Intravital microscopy of the mesenteric vasculature

The intravital microscopy procedure was adapted from previously published protocols.^23,24 Briefly, 5–6-week-old male and female EC-MR^+/+ and EC-MR^−/− littermates were injected intraperitoneally with recombinant mouse TNFα (Biolegend, 0.3μg/mouse in sterile phosphate-buffered saline). Four hours following TNFα injection, mice were anesthetized by intraperitoneal injection of ketamine (100mg/kg) and xylazine (10mg/kg) and administered 100μL of Rhodamine 6G (0.005–0.01% w/v in saline) via the tail vein to label all intravascular cells. The abdomen was opened via a vertical incision through the skin and muscle with sharp scissors and the peritoneal cavity was filled with warm CO₂-infused phosphate-buffered saline. The bowels were then carefully externalized over a large glass coverslip and fluorescent cells were visualized using a Nikon Eclipse Ti inverted microscope equipped with a warming lamp set to 37°C. The exposed bowels and vessels were kept moist with warm CO₂-infused phosphate-buffered saline. Second- and third-order mesenteric veins with lumen diameters of approximately 100μm were imaged. Thirty-second videos were recorded using a Samsung Galaxy S6 on a SnapZoom microscope stabilization attachment. The number and velocity of cells traveling through the center of the vessel were quantified by blinded investigators in 4–7 videos per mouse using the TrackMate v.3.6.0 plugin in ImageJ Fiji. The number of firmly adherent cells in the same analysis area was manually counted in each video in a blinded fashion.

PCR genotyping of mouse lung tissue

Genomic DNA was isolated from mouse tissues using the Qiagen DNeasy kit. PCR for recombined MR (454 base pair band) and intact floxed MR (364 base pair band) was performed with a combination of three primers as previously described (Supplemental Figure I):^16,19,25

• 5’-CCACTTGTATCGGCAATACAGTTTAGTGTC-3’

• 5’-CACATTGCATGGGGACAACTGACTTC-3’

• 5’-CTGTGATGCGCTCGGAAACGG-3’.

Statistics

All data were collected and analyzed in a blinded fashion. Statistical analyses were performed using SigmaPlot v.12.5. Means between two groups were compared with unpaired, two-tailed Student’s t test. Means between male and female EC-MR^+/+ and EC-MR^−/− littermates were compared using Two-way ANOVA with Holm-Sidak post-test. In vitro data were analyzed using One-way ANOVA with Holm-Sidak post-test. Data that did not satisfy either the normality or equal variance requirements of the above statistical tests were analyzed by nonparametric Rank Sum Test or Kruskal-Wallis ANOVA with Mann-Whitney post-test as indicated in the individual figure legends. Statistical significance was defined as p≤0.05. All data are presented as mean±SEM.

Results

MR inhibition with spironolactone reduces vascular inflammation in male mice in the AAV-PCSK9 atherosclerotic model

To investigate the contribution of the MR to the accumulation of inflammatory cells in atherosclerotic plaques, we subjected male C57BL/6J mice to AAV-PCSK9-induced atherosclerosis and randomized the mice to placebo or spironolactone pellets (20mg/kg/day) throughout 12 weeks of HFD feeding. This dose of spironolactone did not significantly affect blood pressure, weight gain, spleen weight, random glucose, or serum cholesterol levels in males. Serum aldosterone levels were significantly increased in spironolactone-treated animals, confirming effective MR inhibition (Supplemental Table I). Quantification of inflammatory cell content of the aortic arch by flow cytometry demonstrated that spironolactone significantly reduced the total number of CD45+ leukocytes within the aortic arch of these male mice (Figure 1A). Despite a trend towards a decrease in CD45+/CD11b-/CD3+ T cells (p=0.08, Figure 1C), this was not statistically significant. Although MR inhibition did not significantly alter the absolute number of CD45+/CD3-/CD11b+ myeloid cells in aortae (Figure 1B), a lower percentage of myeloid cells expressed Ly6C in spironolactone-treated animals, suggesting a shift away from the pro-inflammatory phenotype (Figure 1D). These data demonstrate that systemic MR inhibition attenuates inflammation of atherosclerotic plaques, with decreased total leukocytes and a shift towards anti-inflammatory myeloid cells in the aorta in males.

Figure 1: MR inhibition with spironolactone attenuates vascular inflammation in male but not female atherosclerotic mice.

Flow cytometry quantification of leukocyte populations in the aortic arches of male (A-D) and female (E-H) C57B/6J mice following AAV-PCSK9 injection and treatment with spironolactone (Spiro) or placebo during 12 weeks of HFD feeding. In males, the numbers of total leukocytes within the aortic arches were quantified (A) and sub-divided into myeloid (B) and T cell populations (C). The same analysis was performed in females (E-G). The percentage of myeloid cells positive for Ly6C was also determined (D, H). *p<0.05, **p<0.01, ns=not significant. (E), (F) and (H) analyzed by nonparametric Rank Sum Test, all other data by Student’s unpaired t-test.

MR inhibition does not affect vascular inflammation in female AAV-PCSK9-injected mice

The role of the MR in atherosclerosis and vascular inflammation has never been assessed specifically in female animals in any preclinical atherosclerosis model. Thus, we exposed female C57BL/6J mice to AAV-PCSK9 injection and 12 weeks of HFD and randomized to placebo or spironolactone. In females, MR inhibition did not alter blood pressure, weight gain, random glucose, or serum cholesterol and MR inhibition increased serum aldosterone levels, consistent with effective MR blockade (Supplemental Table II). Spleen weight was reduced in spironolactone-treated females compared to placebo, suggesting a decrease in overall inflammatory status with MR inhibition in females. Despite this, and in contrast to the results in male mice, there was no significant difference in the total number of leukocytes (Figure 1E), myeloid cells (Figure 1F), or T cells (Figure 1G) within the aortic arches of female spironolactone-treated mice compared to placebo-treated controls, nor did spironolactone treatment alter the percentage of myeloid cells expressing Ly6C in females (Figure 1H). Overall, these data indicate that systemic MR inhibition decreases vascular inflammation and alters myeloid cell phenotype in atherosclerotic males and reduces spleen weight in females, with no impact on atherosclerotic vascular inflammation in females.

EC-MR differentially contributes to vascular inflammation in males and females

As EC-MR has previously been shown to regulate the expression of inflammatory mediators,^12,21 we next tested the hypothesis that EC-MR mediates vascular inflammation in atherosclerosis and examined sex differences. Male and female MR-intact (EC-MR^+/+) and EC-MR knockout (EC-MR^−/−) littermates were injected with AAV-PCSK9 and fed HFD for 12 weeks to induce atherosclerosis. The number of CD45+ leukocytes within the aortic arches of male EC-MR^−/− mice was reduced by 38% compared to male EC-MR^+/+ controls (Figure 2A). This reduction was driven by a significant decrease (41%) in CD45+/CD3-/CD11b+ myeloid cells (Figure 2B) and a trend towards fewer CD45+/CD11b-/CD3+ T cells (36%) that did not achieve statistical significance (p=0.057, Figure 2C). The aortic arches of female EC-MR^+/+ mice contained 56% fewer total leukocytes, 62% fewer myeloid cells, and 52% fewer T cells than male EC-MR^+/+ littermates. EC-MR deletion did not further attenuate vascular inflammation in females; rather, the number of myeloid cells within the aortic arches of female mice was increased by EC-MR deletion (Figure 2B). The percentage of myeloid cells expressing Ly6C was reduced in females compared to males, independent of the presence of EC-MR (Figure 2D). Notably, although female mice weighed less and gained less weight than males, the effect of EC-MR on vascular inflammation was not due to any effect of genotype on blood pressure, weight gain, fasting glucose, fasting cholesterol, or serum aldosterone levels in males or females after AAV-PCSK9 injection and HFD (Supplemental Table III). To rule out the possibility of differences in baseline vascular inflammation between genotypes, we assessed inflammation in aortic arches from uninjected mice on normal laboratory diet. As expected, vascular inflammation was minimal without atherosclerosis and there were no differences in baseline inflammation between male and female mice regardless of EC-MR status (Supplemental Figure IV). Serum TNFα levels were also not different among non-atherosclerotic mice (Supplemental Figure V–A), and induction of atherosclerosis by AAV-PCSK9 with HFD induced an increase in serum TNFα in all four groups with no difference by sex or genotype (Supplemental Figure V–B). In summary, deletion of EC-MR attenuates vascular inflammation specifically in atherosclerotic male mice. MR-intact females have decreased inflammation compared to males and EC-MR deletion is not protective, but rather is associated with increased plaque myeloid cells.

Figure 2: EC-MR deletion attenuates vascular inflammation in male but not female atherosclerotic mice.

Aortic arches from AAV-PCSK9-injected male and female EC-MR^+/+ and EC-MR^−/− littermates were analyzed by flow cytometry following 12 weeks of HFD. Quantification of total leukocyte (A), myeloid cell (B), and T cell populations (C) are shown. (D) The proportion of myeloid cells positive for Ly6C was determined. *p<0.05, **p<0.01, ns=not significant. (B) analyzed by Kruskal-Wallis ANOVA with Mann-Whitney post-test, all other data by Two-way ANOVA with Holm-Sidak post-test.

EC-MR deletion does not alter atherosclerotic plaque size in males or females

The aortic roots from the mice in Figure 2 were assessed histologically and plaque size and composition were analyzed by methods demonstrated in Supplemental Figure VI. As expected, AAV-PCSK9 injection and 12 weeks of HFD feeding induced plaque formation in the aortic sinuses of male and female mice (Figure 3A), as was previously reported only in males.²⁰ In this model, females developed significantly larger aortic sinus plaques than males after 12 weeks of HFD (Figure 3C), a finding that is consistent with reports using genetic mouse atherosclerosis models.²⁶ However, there was no difference in plaque size between EC-MR^+/+ and EC-MR^−/− mice of the same sex. Quantification of plaque composition revealed that, regardless of the sex of the mice, the absence of EC-MR did not significantly influence lipid content (p=0.098 by genotype, Figures 3A, 3D) or the degree of plaque fibrosis (Figure 3B, F). Only necrotic core area as a percentage of total plaque was increased, specifically in male EC-MR^−/− mice compared to male EC-MR^+/+ and female EC-MR^−/− littermates (Figure 3E). Overall, these data suggest that EC-MR deletion does not significantly alter plaque size despite decreasing inflammation and increasing necrotic core in males and that MR-intact females have larger but less inflamed plaques.

Figure 3: EC-MR deletion does not affect plaque size in atherosclerotic mice of either sex.

Representative images of aortic root sections from AAV-PCSK9-injected male and female EC-MR^+/+ and EC-MR^−/− littermates stained with (A) Oil Red O (ORO) and (B) PicroSirius Red are shown. Scale bars=200μm. (C) Plaque area is quantified. Within the plaque, lipid content (D), necrotic core area (E), and collagen content (F) were quantified as a percentage of the total plaque area. *p<0.05, **p<0.01, ***p<0.001, ns=not significant. Two-way ANOVA with Holm-Sidak post-test.

EC-MR contributes to leukocyte adhesion and ICAM-1 expression in vivo in a sex-dependent manner.

We next examined the role of EC-MR in leukocyte adhesion to the mesenteric vasculature under flow conditions by intravital microscopy. As TNFα levels were increased in AAV-PCSK9-injected mice after 12 weeks of HFD feeding, we used acute intraperitoneal TNFα administration as a model to investigate the effect of EC-MR on leukocyte recruitment. The number of vessels with firmly adherent leukocytes was significantly reduced in male EC-MR^−/− and female EC-MR^+/+ mice compared to male EC-MR^+/+ littermates (Figure 4A). This corresponded to reduced expression of ICAM-1 in the mesenteric veins of TNFα-injected females and EC-MR^−/− males compared to EC-MR^+/+ males (Figure 4B).

Figure 4: EC-MR contributes to leukocyte adhesion and ICAM-1 expression in vivo in a sex-dependent manner.

(A) The number of firmly adherent leukocytes was quantified in the mesenteric veins of male and female EC-MR^+/+ and EC-MR^−/− littermates subjected to intravital microscopy 4 hours after intraperitoneal TNFα injection. The number of vessels per mouse exhibiting >4 adherent cells is quantified. (B) ICAM-1 and β-tubulin immunoblots of mesenteric veins from TNFα-injected male and female EC-MR^+/+ and EC-MR^−/− littermates. The data is quantified below (N=4). *p<0.05, **p<0.01, ns=not significant. (A) analyzed by Two-way ANOVA with Holm-Sidak post-test, (B) by One-way ANOVA with Holm-Sidak post-test.

EC-MR regulates ICAM-1 expression and promotes leukocyte adhesion in cultured ECs.

In mouse lung ECs (mLECs) isolated from EC-MR^+/+ and EC-MR^−/− littermates, EC-MR deletion attenuated ICAM-1 protein expression, as did estrogen administration to EC-MR^+/+ mLECs (Figure 5A). The adhesion of fluorescently-labeled U937 monocytes to static cultures of mLECs was likewise reduced with EC-MR deletion and with estrogen administration (Figure 5B). These results were recapitulated in HUVECs treated with the MR inhibitor spironolactone or with estrogen compared to vehicle (Figure 5C–D). These HUVECs express MR and ERα, but undetectable levels of ERβ (Supplemental Figure VII). When HUVECs were transfected with a luciferase reporter driven by the ICAM-1 proximal promoter, spironolactone or estrogen treatment significantly reduced ICAM-1 promoter activity (Figure 5E), supporting a transcriptional mechanism. In cultured EC studies, the co-administration of spironolactone and estrogen or the addition of estrogen to MR-deficient ECs had no additional effect beyond that of estrogen or MR blockade/deletion alone. One exception was the expression of ICAM-1 in mLECs: EC-MR^−/− cells treated with estrogen expressed lower levels of ICAM-1 than either vehicle-treated EC-MR^−/− cells or estrogen-treated EC-MR^+/+ cells (Figure 5A). Collectively, these results indicate that EC-MR regulates ICAM-1 expression and leukocyte firm adhesion to the mesenteric vein in vivo and to mouse ECs and HUVECs in culture. Estrogen also downregulates ICAM-1 expression in vitro, which corresponds to reduced expression of ICAM-1 in the mesenteric veins of female mice.

Figure 5: EC-MR regulates ICAM-1 expression and promotes leukocyte adhesion in cultured ECs.

(A) ICAM-1 and β-tubulin immunoblots of isolated mouse lung ECs (mLECs) from EC-MR^+/+ and EC-MR^−/− littermates in the presence or absence of estrogen (E2, 10nM) (N=4). (B) Adhesion of fluorescently-labeled U937 monocytes to static cultures of mLECs treated as in (A) (N=4). (C) ICAM-1 and GAPDH immunoblots of HUVECs treated with spironolactone (Spiro, 1μM) and/or E2 (N=4). (D) Adhesion of fluorescently-labeled U937 monocytes to static cultures of HUVECs treated as in (C) (N=3). (E) Luciferase reporter assay quantifying ICAM-1 proximal promoter activity in HUVECs with Spiro and/or E2 (N=3). **p<0.01, ***p<0.001, ns=not significant. (A) was analyzed by Kruskal-Wallis ANOVA with Mann-Whitney post-test, all other data by One-way ANOVA with Holm-Sidak post-test. Scale bars=50μm.

EC-MR contributes to leukocyte rolling interactions in a sex-dependent manner in vivo and to regulation of E-selectin in vivo and in cultured ECs.

TNFα-induced leukocyte fast and slow rolling on the venous endothelium was also quantified from intravital microscopy videos (see Supplemental Videos I–IV). Male EC-MR^+/+ mice exhibited significantly more intravascular cells rolling on the vessel surface compared to male EC-MR^−/− and female EC-MR^+/+ littermates (Figure 6A). This difference in total rolling cells was due predominantly to changes in the number of slow-rolling cells (velocity<10μm/s), as females and EC-MR^−/− males demonstrated significantly fewer cells rolling slowly along the venous endothelium than their male EC-MR^+/+ littermates (Figure 6B). By contrast, the number of fast-rolling cells (velocity≥10μm/s) was not significantly affected either by sex or genotype, though there was a non-significant trend towards a decrease in fast-rolling cells with EC-MR deletion (p=0.055 by genotype, Figure 6C). As slow rolling is dependent on endothelial E-selectin,¹⁰ we assessed whether EC-MR regulates E-selectin expression in the mesenteric veins of TNFα-injected mice. Indeed, the mesenteric veins of females and EC-MR^−/− males demonstrated reduced expression of E-selectin compared to EC-MR^+/+ males, with E-selectin expression being further decreased in females relative even to EC-MR^−/− males (Figure 6D). In cultured mLECs, E-selectin induction by low-dose TNFα was attenuated in cells lacking EC-MR or administered estrogen, with no further effect of estrogen in EC-MR^−/− cells (Figure 6E). TNFα-induced E-selectin expression in HUVECs was also blocked by spironolactone or estrogen treatment (Figure 6F). These data indicate that EC-MR deletion reduces leukocyte slow-rolling interactions with the endothelium in male mice and that female mice overall display protection from these interactions in this acute inflammation model. These findings are unaffected by EC-MR deletion in females and correlate with regulation of E-selectin by MR and estrogen in mLECs and HUVECs in vitro.

Figure 6: EC-MR contributes to leukocyte rolling interactions in a sex-dependent manner in vivo and to regulation of E-selectin in vivo and in cultured ECs.

(A) The total number of rolling leukocytes was quantified in intravital microscopy videos from male and female EC-MR^+/+ and EC-MR^−/− littermates. (B) The number of slow-rolling (velocity <10μm/s) and (C) fast-rolling (velocity ≥10μm/s) cells were calculated from the total rolling cell data. (D) E-selectin and β-tubulin immunoblots of mesenteric veins from TNFα-injected male and female EC-MR^+/+ and EC-MR^−/− littermates. The data is quantified below (N=4). (E) E-selectin and β-tubulin immunoblots of isolated mLECs from EC-MR^+/+ and EC-MR^−/− littermates in the presence or absence of estrogen (E2, 10nM) throughout 20 hours of stimulation with TNFα (80pg/mL) (N=3). (F) E-selectin and GAPDH immunoblots of HUVECs treated with spironolactone (Spiro, 1μM) and/or E2 throughout 20 hours of TNFα stimulation (N=4). *p<0.05, **p<0.01, ***p<0.001, ns=not significant. (A-C) analyzed by Two-way ANOVA with Holm-Sidak post-test, (D) by Kruskal-Wallis ANOVA with Mann-Whitney post-test, (E-F) by One-way ANOVA with Holm-Sidak post-test.

Discussion

This study is the first to compare the role of the MR in atherosclerosis and vascular inflammation between males and females. In males, we demonstrate that EC-MR contributes to: (1) inflammation of atherosclerotic plaques; (2) leukocyte-endothelial interactions in an acute inflammation model; and (3) expression of E-selectin and ICAM-1, endothelial adhesion molecules critical for leukocyte recruitment. By contrast, in females: (4) plaques are larger but less inflamed than in males in the PCSK9-induced mouse model; (5) neither MR inhibition nor EC-MR deletion reduces atherosclerotic inflammation; and (6) leukocyte recruitment to the vasculature and endothelial adhesion molecule expression are reduced compared to males, with no additional role for EC-MR. These data reveal a striking sex difference in atherosclerotic vascular inflammation and in the role of EC-MR in regulating inflammatory cell recruitment to the vasculature.

EC-MR as a mediator of vascular inflammatory signaling in males

Our studies in males are consistent with existing literature implicating the MR in atherosclerosis. Prior studies in male mouse models have shown that MR activation by aldosterone increases plaque size and inflammation,¹¹ while MR inhibition decreases plaque size.² In vitro, the MR has been shown to regulate ICAM-1 expression in human ECs.¹² Here we further demonstrate that EC-MR regulates ICAM-1 and also E-selectin expression in vivo in response to acute inflammatory stimuli in males. Importantly, we also demonstrate the EC-MR contributes to leukocyte firm adhesion and slow rolling in males by intravital microscopy.

ICAM-1 was recently found to be required for aldosterone to enhance plaque formation and inflammation in male ApoE^−/− mice,¹³ highlighting the importance of this molecule in the effects of the MR on the vasculature. Ample data supports that EC-MR regulates ICAM-1 transcription directly in its capacity as a hormone-activated transcription factor. This is evidenced by our luciferase reporter experiment in Figure 5E and by previously published studies demonstrating that deletion of a predicted MR binding site from the ICAM-1 proximal promoter prevents aldosterone-induced activation of ICAM-1 reporter activity in HUVECs.¹³ Far less is known regarding the mechanism by which the MR regulates E-selectin. While supraphysiologic concentrations of aldosterone have previously been shown to upregulate E-selectin expression in vitro,^27,28 our study is the first to demonstrate EC-MR regulation of E-selectin in vivo. Preliminary in silico analysis identifies potential MR binding sites in the E-selectin promoter, however further study is required to establish the functional significance of these putative sites. The MR is also known to modulate the activity of the pro-inflammatory transcription factor nuclear factor-κB,² which may represent an additional mechanism by which EC-MR regulates adhesion molecule expression.

While we largely found no effect of EC-MR deletion on the size or composition of atherosclerotic plaques, we did observe an increase in necrotic core specifically in EC-MR^−/− males compared to MR-intact littermates. These mice had decreased plaque inflammation, as evidenced by flow cytometry, but no change in overall plaque size. It is possible that EC-MR contributes to the survival of inflammatory cells within the atherosclerotic plaque, hence EC-MR^−/− male mice display decreased plaque inflammatory cells but a larger necrotic core that may be comprised of necrotic inflammatory cells. Further investigation is needed to fully explore the contribution of EC-MR to necrotic core formation and the potential ramifications of this finding for plaque stability.

A variety of non-endothelial cell types also contribute to atherosclerosis, with variable roles for the MR. We recently demonstrated in male ApoE^−/− mice that smooth muscle cell MR deletion does not impact atherosclerotic plaque size, inflammation, or composition.¹⁹ By contrast, myeloid cell MR has been implicated in lesion formation, macrophage cholesterol handling, and myeloid cell polarization.^29,30 Consistent with these findings, we observed that MR inhibition with spironolactone results in a reduction in the proportion of myeloid cells expressing the Ly6C inflammatory marker in the aortic arch in males. This finding is independent of EC-MR and thus may be mediated by the MR in myeloid cells or other non-ECs. Further study is warranted to fully understand the contribution of the MR in the various cell types within the atherosclerotic lesion to plaque formation, progression, and stability characteristics.

Reduced vascular inflammation and atherosclerosis in females, independent of the MR

Despite substantial data implicating the MR in atherosclerosis in males,² the present study provides the first evidence that MR signaling contributes to vascular inflammation differently in females, indicating that studies in males cannot be routinely extrapolated to both sexes. Whereas MR inhibition reduced atherosclerotic vascular inflammation in male mice via its activity in the endothelium, neither pharmacologic MR inhibition nor EC-MR deletion in females reduced vascular inflammation, leukocyte recruitment, or vascular adhesion molecule expression from a baseline that was already lower than that of males. The finding of less inflammation in females is consistent with a recent report that demonstrated attenuated inflammatory marker staining in the aortic root plaques of female mice compared to males³¹ and with human data indicating that premenopausal women have lower coronary calcium scores, a marker linked with plaque inflammation,³² than age-matched men.³³

We observed larger plaques in the aortic roots of females compared to males, a phenomenon that has been previously described in other mouse atherosclerosis models.³⁴ Perhaps this observation has led investigators to focus atherosclerosis studies predominantly on male mice, particularly since it is well known that premenopausal women experience fewer atherosclerosis-related adverse cardiovascular events than age-matched men.⁷ However, it is not clear that protection from cardiovascular events in premenopausal women is due to a decrement in plaque burden relative to men. Some studies suggest that women’s risk of subclinical atherosclerosis is underestimated by current clinical algorithms³⁵ and that women may in fact develop uncomplicated plaques at a similar rate as men.³⁶ Indeed, our data indicates that female mice in our model did develop substantial atherosclerosis and increased serum TNFα levels (similar to that of males), but that the plaques formed in the vasculature of females contained fewer inflammatory cells relative to their size. This decrease in plaque inflammation in female mice could indicate a more stable plaque phenotype and may represent a mechanism for the protection from cardiovascular ischemia observed in premenopausal women.⁷ Indeed, it is well established in humans that less inflamed plaques are less likely to rupture and cause subsequent adverse cardiovascular events.⁹ However, further investigation using animal models in which plaque rupture can be studied³⁷ will be needed to confirm this potential relationship between female sex, vascular inflammation, and plaque stability.

Overall, these data demonstrate that in females and in cultured ECs treated with estrogen, inflammatory measures were lower than that of males or cells without estrogen, and MR inhibition or EC-MR deletion did not significantly alter those findings. One exception is the increase in myeloid cell content in the aortic arches of females with EC-MR deletion compared to their MR-intact littermates (Figure 2B). This was opposite to the effect observed in males, in which EC-MR deletion significantly attenuated myeloid cell accumulation and total inflammation in the aortic arch. This effect in females was specific to the myeloid cell population, with no significant impact of EC-MR deletion on total aortic arch leukocytes or T cells in females. Additionally, the number of aortic myeloid cells in female EC-MR^−/− mice is not different from EC-MR^−/− males, despite a larger plaque area in females. These caveats notwithstanding, it is interesting to speculate on potential mechanisms for this sex difference in the regulation of atherosclerotic inflammation by EC-MR. While female sex hormones have been implicated in the protection from cardiovascular ischemia observed in premenopausal women,³⁸ hormone replacement therapy has not proven to be a successful strategy for preventing adverse events in women after menopause,³⁹ suggesting a more nuanced role for female sex hormones in the cardiovascular system. Indeed, in preclinical studies the actions of estrogen in the vasculature are controversial. While ERα has been found in some contexts to inhibit endothelial adhesion molecule expression²¹ and to mediate the atheroprotection elicited by estrogen in ovariectomized females,⁴⁰ other studies suggest that ERα may actually promote inflammation²³ and vascular stiffness⁴¹ in certain circumstances. Thus, it is possible that in females ERα could exert pro-inflammatory effects that may be blocked by EC-MR, thus providing a rationale for how EC-MR deletion in females could paradoxically increase the number of myeloid cells accumulating in the aortic arch. However, this model is highly speculative and will need substantial investigation to confirm the nature of EC-MR crosstalk with female sex hormones.

In a few instances, we did observe evidence of synergy between EC-MR and estrogen signaling or female sex. In isolated mouse lung ECs, the administration of estrogen to EC-MR^−/− cells reduced ICAM-1 protein levels relative to both vehicle-treated EC-MR^−/− and estrogen-treated EC-MR^+/+ cells (Figure 5A). This was the only instance of synergy between EC-MR and estrogen that we observed in our in vitro mechanistic studies. This did not translate to changes in the functional outcome of leukocyte adhesion to static cultures of these cells (Figure 5B) or to the pattern of ICAM-1 expression or promoter activity in HUVECs (Figure 5C, E), nor did it correlate with in vivo ICAM-1 expression in the mesenteric veins of female mice with intact MR versus EC-specific MR deletion (Figure 4B). A different pattern of potential synergy between female sex and EC-MR was observed in the expression of E-selectin in mesenteric vessels in vivo, as the vessels from females lacking EC-MR expressed significantly less E-selectin than vessels from males of either genotype (Figure 6D). Again, this did not correlate with changes in leukocyte slow rolling in the intravital microscopy studies (Figure 6B). As additive effects of estrogen/female sex and EC-MR deletion were not present under all circumstances and were not associated with functional changes in leukocyte-EC interactions in vitro or in vivo, the significance of these observations is unclear. Further study will be needed to fully characterize the complex nature of EC adhesion molecule regulation by MR and estrogen utilizing a variety of models both in vivo and in vitro.

Mechanisms for sex differences in vascular inflammation and the role of EC-MR

The mechanisms governing female cardioprotection have been extensively studied and are typically attributed to the actions of female sex hormones.³⁸ Consistent with this, we demonstrate herein that aortic plaque inflammation, mesenteric leukocyte-endothelial interactions, and expression of ICAM-1 and E-selectin in the mesenteric veins of TNFα-stimulated mice are attenuated in females compared to males. We further implicate estrogen in these effects, as estrogen treatment of mLECs and HUVECs also diminished ICAM-1 and E-selectin expression and reduced leukocyte adhesion in static culture. Estrogen-activated ER has been shown to inhibit the inflammatory transcription factor nuclear factor-κB, a known regulator of ICAM-1 and E-selectin, through a variety of genomic and non-genomic mechanisms.⁴² This MR-independent regulation may contribute to estrogen inhibition of inflammatory gene expression and leukocyte recruitment in this study. Clinical data also supports the potential for estrogen to regulate inflammatory signaling, as a small randomized clinical trial found that soluble ICAM-1 and E-selectin levels were reduced in postmenopausal women administered hormone therapy compared to untreated controls.⁴³ Together with our findings, these data suggest that estrogen downregulates pro-inflammatory signaling in ECs, thus reducing expression of these adhesion molecules, which are critical for leukocytes to interact with the endothelium and accumulate in atherosclerotic plaques.

In addition to these MR-independent mechanisms, the sex differences we observed in the role of EC-MR in vascular inflammation may involve interactions between estrogen signaling and the MR. We have previously demonstrated that in the presence of estrogen, ERα forms a complex with the MR in the nucleus of ECs and inhibits MR-mediated ICAM-1 promoter activity, protein expression, and leukocyte adhesion. This requires ERα to be able to translocate to the nucleus, but not to bind DNA or participate in rapid signaling.²¹ Limited data also suggests that ERβ can inhibit MR pro-inflammatory signaling,⁴⁴ and the interactions between MR signaling and the G-protein coupled ER are a topic of particular study and controversy.⁴⁵ Together with the results of the current study, these data are consistent with a model in which estrogen signaling inhibits the upregulation of endothelial adhesion molecules by EC-MR, thus reducing leukocyte recruitment and accumulation in the atherosclerotic plaques of female mice. Based on this interpretation, if the pro-inflammatory role of EC-MR is already inhibited by estrogen in females, deletion of EC-MR would not be expected to result in a further reduction of inflammation in gonad-intact females, as we observed. Further investigation into the role of the MR in ovariectomized mice or in mice lacking endothelial ERα, ERβ, or G-protein coupled ER could further clarify the nature of MR/estrogen crosstalk in the pathophysiology of vascular inflammation in females.

Conclusions, limitations, and clinical implications

Our data suggests a model in which EC-MR regulates the expression of endothelial adhesion molecules in males, leading to enhanced leukocyte recruitment to the vasculature and ultimately the accumulation of inflammatory cells in atherosclerotic plaques. In females, there are fewer plaque inflammatory cells than in males despite larger plaque size and equivalent systemic inflammatory biomarkers, leukocyte recruitment to the vasculature is diminished, and expression of endothelial adhesion molecules is lower, effects which are largely independent of MR signaling.

It is important to acknowledge the limitations of the current study. First, as atherosclerotic plaques in mice do not typically rupture, we did not specifically assess the role of the MR in plaque rupture, instead using established surrogate indicators of plaque stability, such as inflammation, lipid content, necrotic core, and collagen content. As new preclinical models of plaque rupture become available,³⁷ future studies could shed light on the contribution of the MR to plaque instability and rupture. We measured blood pressure by tail cuff plethysmography, rather than the more sensitive telemetry method, in order to reduce the potential for flow disturbance in the aorta by the telemetry catheter. Thus, we cannot rule out subtle alterations in blood pressure below the level of detection of this method. Finally, females in the EC-MR atherosclerosis study gained less weight than their male littermates, which may have impacted outcomes. However, these females developed the same degree of hypercholesterolemia and serum TNFα and even larger plaques than their male counterparts, therefore this difference in a single cardiovascular risk factor is unlikely to fully account for the sex difference in inflammation demonstrated in this study.

Despite these limitations, the current study has important clinical implications. Our finding that EC-MR promotes leukocyte recruitment, endothelial inflammatory marker expression, and atherosclerotic vascular inflammation in males suggests that MR inhibitor therapy could be a useful tool for the prevention of vascular inflammation in male patients with atherosclerosis. Conversely, we did not observe the same benefit of MR blockade or deletion in gonad-intact female mice, indicating that such therapy may not be beneficial for atheroprotection in premenopausal women. This may begin to be addressed with the MAGMA clinical trial, which is currently recruiting patients at high risk of cardiovascular ischemia for a one-year trial of spironolactone versus placebo, after which atheroma volume and inflammatory markers will be assessed.⁴⁶ Based on our data, we would hypothesize that the male patients in this study will exhibit a reduction in inflammatory markers, while premenopausal female patients may not experience this same benefit. Whether this leads to changes in the risk of cardiovascular events will not be tested in MAGMA. It is hoped that the current study may prompt further investigation into the mechanisms driving atherosclerotic vascular inflammation, particularly in females, and inform future clinical studies designed to prevent atherosclerotic plaque rupture and reduce the morbidity and mortality from atherosclerosis.

Highlights.

• Endothelial MR contributes to plaque inflammation in atherosclerosis differentially in male and female mice.

• EC-MR contributes to leukocyte rolling and adhesion to the endothelium and to adhesion molecule expression in vivo in males.

• In the AAV-PCSK9 model, female mice develop larger plaques with less vascular inflammation than males, a phenotype consistent with more stable plaques that are less likely to rupture in humans.

• EC-MR promotes, and estrogen inhibits, ICAM-1 and E-selectin expression in mouse mesenteric veins, mouse ECs, and HUVECs.

Abbreviations:

AAV

adeno-associated virus

ApoE^−/−

apolipoprotein E knockout

EC

endothelial cell

EC-MR

endothelial cell mineralocorticoid receptor

EC-MR^+/+

MR-intact

EC-MR^−/−

endothelial-specific MR knockout

ER

estrogen receptor

HFD

high fat diet

HUVECs

human umbilical vein endothelial cells

ICAM-1

intercellular adhesion molecule-1

mLECs

mouse lung endothelial cells

MR

mineralocorticoid receptor

PCSK9

proprotein convertase subtilisin/kexin type 9

Spiro

spironolactone