Elastin regulation of vasoreactivity in resistance arteries

Abstract

Background:

Endothelial cells (ECs) are the primary producer of elastin in the internal elastic lamina (IEL) of resistance arteries. These arteries have distinct gaps in their IEL where ECs facilitate heterocellular communication with smooth muscle in a signaling microdomain termed the myoendothelial junction (MEJ). However, the contribution of the IEL to vasodilation and blood pressure in resistance arteries is not well understood.

Methods:

An endothelial-specific elastin knockout mouse (EC-specific Eln^fl/fl/Cre⁺) was utilized to alter the IEL and MEJs. MEJ-resident proteins were localized by en face, pressure myography assessed the effect of elastin depletion on vessel dilation and blood pressure was measured using radiotelemetry.

Results:

Using scRNA-seq, we found Eln mRNA enriched in arterial endothelium. In EC-specific Eln^fl/fl/Cre⁺ mice, the localization of the MEJ-resident protein alpha hemoglobin (Hbα), becomes diffuse and disorganized. Normally, Hbα regulates endothelial nitric oxide synthase (eNOS) by sequestering NO, promoting endothelial-derived hyperpolarization as the predominant vasodilation mechanism. However, in EC-specific Eln^fl/fl/Cre⁺ mice, Hbα expression and interaction with eNOS is significantly reduced, corresponding to increased NO signaling via acetylcholine dilation. Intact arteries also exhibit decreased smooth muscle contractility with the diminished IEL. These vascular deficiencies suggested a hypotensive phenotype, but EC-specific Eln^fl/fl/Cre⁺ mice blood pressure was not different from controls.

Conclusions:

Our findings suggest elastin deficiency in resistance arteries alters their vasoreactive properties, resulting in poor contraction and dilation. Furthermore, the absence of the HIEL mis-localizes Hbα and eNOS in resistance arteries, switching the vasodilatory mechanism from endothelial-derived hyperpolarization to NO signaling, mimicking larger conduit arteries.

Graphical Abstract

INTRODUCTION

Resistance arteries are the location of the largest pressure drop in the vascular system, such that changes in vasodilation or vasoconstriction of these vessels are impactful on total peripheral resistance and thus blood pressure. How endothelial cells (ECs) and smooth muscle cells (SMCs) of resistance arteries can control resistance artery diameter and thus total peripheral resistance has been studied extensively (e.g.,¹). However, the ways in which the extracellular matrix layer separating these two cell types—termed the internal elastic lamina (IEL)—may contribute to regulation of vasodilation and vasoconstriction is less understood.

It has been previously shown that elastin is critical to cardiovascular development and function, especially in the heart and conduit arteries, because it provides vessels elasticity to reversibly expand and contract.² Mice with a global elastin (Eln) deletion exhibit stiff, stenotic and tortuous vessels and are not viable, dying shortly after birth due to arterial obstruction by SMC over proliferation into the vessel lumen.^3–5 Mice heterozygous for a global Eln deletion are viable, but exhibit moderate cardiac hypertrophy, hypertension, and smaller, thinner arteries with seemingly compensatory increases in the amount of lamellar units in vessel walls.⁶ In humans, elastin haploinsufficiency through mutations or deletions in ELN causes supravalvular aortic stenosis, a disease which also exhibits hypertension and elevated numbers of lamellar units.^3,6,7 Furthermore, reduced elastin and increased arterial stiffening and compliance, most notable in conduit arteries, is associated with aging and increases the risk of developing cardiovascular diseases such as hypertension.^8,9

Much less is known about the effects of diminished elastin content in resistance arteries on vasodilation, vasoconstriction, and blood pressure. Previous work using an Eln floxed mouse model demonstrated an interesting dichotomy with elastin and the IEL that is dependent upon arterial location (resistance arteries vs conduit arteries) and cell type (EC vs SMC).¹⁰ SMC Eln deletion significantly impairs the IEL structure in conduit arteries and results in death days after birth, but the resistance arteries are for the most part anatomically normal. Whereas with EC Eln deletion, mice are viable and show no overt phenotypes, but there is significant impairment in the IEL of muscular and resistance arteries, with the IEL in conduit arteries resembling controls.^10,11 Despite the lack of IEL in mice with EC-specific Eln deletion, EC and SMC function remain intact as evidenced by preserved NS309-mediated dilation and KCl constriction.¹¹

In resistance arteries, the IEL (composed primarily of elastin and microfibrils¹²) has a high density of circular shaped holes (HIEL), where elastin and microfibrils are absent.^13–17 At each of these HIEL,¹¹ ECs extend through and directly contact SMC to form highly specific heterocellular signaling domains termed myoendothelial junctions (MEJs).^{15,16,18–27} The unique protein localization at MEJs coordinates endothelial-derived hyperpolarization (EDH), which is the dominant vasodilation mechanism in resistance arteries.^{16,24,28–32} However, how the vasoreactive properties of resistance arteries may change, particularly by affecting EC dilation and/or SMC contraction, in the absence of elastin in the IEL is unknown. This is an especially important question considering the anatomical differences when Eln is deleted from EC, which results in a significant decrease in the number of HIEL of resistance arteries.¹¹

The relative contribution of EDH- and nitric oxide (NO)-mediated vasodilation is correlated with vessel size, the presence of HIEL and the number of MEJs—large conduit arteries with few to no MEJs primarily dilate via NO, and resistance arteries with a high incidence of MEJs dilate via EDH.^31,33–36 In HIEL of resistance arteries, alpha hemoglobin (Hbα) localizes to MEJs.^16,36,37 Hbα complexes with endothelial nitric oxide synthase (eNOS) and scavenges NO, preventing NO diffusion to SMCs and reducing the overall contribution of NO to relaxation. This provides a mechanism whereby EDH can predominate. Conversely, NO-mediated relaxation in conduit arteries is facilitated by the absence of Hbα in endothelium.³⁸ Thus, if vasodilation is altered in resistance arteries in the absence of elastin, it may be due to disruptions of Hbα localization.

Based on the above, we hypothesized EC-specific elastin knockout mice (Eln^fl/fl/Cre+) may have disrupted vasoreactive properties. In line with this hypothesis, we demonstrate a loss of Hbα in endothelium when elastin is absent that alters the degree of dilation as well as the method of dilation (from EDH to NO). We also found EC-specific Eln^fl/fl/Cre+ mice exhibited reduced SMC contractility of resistance arteries compared to controls. Although this data may hint at an overall hypotensive phenotype, EC-specific Eln^fl/fl/Cre+ mice do not show any changes in blood pressure compared to controls.

METHODS

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Ethical Approval

All experiments were approved by the University of Virginia Animal Care and Use Committee (Protocol #3648 and #4051) and conducted in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals.

Mice

Extensive details of the mice used in the single cell RNA-sequencing (scRNA-seq) experiments were previously published.³⁹ For all other experiments, elastin floxed (Eln^fl/fl) mice with or without a VE-cadherin (Cdh5) Cre (Eln^fl/fl/Cre⁺ and Eln^fl/fl/Cre⁻, respectively) were used. Eln^fl/fl mice were created by Drs. Wagenseil and Mecham (Washington University) by inserting LoxP sites into the Eln gene using CRISPR/Cas9, and microinjecting Cas9 pairs and donor vectors into C57BL/6J embryos.¹⁰ Eln^fl/fl mice were bred to a Cdh5Cre⁴⁰ background at the University of Virginia for EC specificity.

Statistics

All statistical analyses were performed using Graphpad Prism version 10. Unless otherwise specified, an unpaired Student’s t-test was used to determine statistical significance (P<0.050). All analyses with N=3 were tested for normality and passed a Shapiro-Wilk normality test (P>0.050). Data is presented as mean ± standard deviation.

RESULTS

Using a scRNA-seq dataset of isolated mesenteric ECs recently published by our group, we first investigated the expression of various extracellular matrix genes along the vascular tree including ECs previously identified as arterial, capillary, venous and lymphatic.³⁹ This dataset contains transcriptomic profiles from ECs isolated from both adipose and mesentery of ROSA26-eYFP^+/+ Cdh5-CreER^T2+ mice fed either a normal chow or high fat diet after the induction of YFP with tamoxifen. However, we focused our analyses only on ECs derived from the mesentery of normal chow mice to analyze matrix proteins in resistance arteries under homeostatic conditions. Of the extracellular matrix genes examined, which included elastin (Eln), collagens (Col1a1, Col1a2, Col4a1, Col4a2, Col4a3, Col4a4, Col4a5, Col4a6), fibronectin (Fn1), and hyaluronic acid synthases (Has1, Has2), we found that both elastin and fibronectin were the most highly expressed in arterial ECs (Figure 1A). A feature plot^41,42 of the data further showed elastin to be primarily expressed in the arterial endothelium and secondarily in the venous endothelium (Figure 1B–C).

Figure 1. Elastin is highly expressed in the arterial endothelium.

Data mining from a publicly available scRNA-seq dataset (GEO accession no. GSE235192)³⁹. Male ROSA26-eYFP^+/+ Cdh5-CreER^T2+ mice were fed a tamoxifen diet to induce endothelial YFP expression, followed by normal chow for 12–13 weeks. Endothelial cells (ECs) were then isolated from the mesentery, with tissue from six mice pooled for two separate scRNA-seq experiments. A, Dot plot of extracellular matrix (ECM) genes elastin (Eln), collagens I and IV (Col1a1, Col1a2, Col4a1, Col4a2, Col4a3, Col4a4, Col4a5, Col4a6), fibronectin (Fn1), and hyaluronic acid synthases (Has1, Has2). B, Uniform Manifold Approximation and Projection (UMAP) of mesenteric ECs. C, Feature plot of elastin expression in the mesenteric endothelium.

Next, we investigated the function of elastin in the arterial endothelium with endothelial-specific elastin knockout mice (EC-specific Eln^fl/fl/Cre⁺) in comparison to Eln^fl/fl/Cre⁻ controls (Figure S1).¹⁰ En face preparations of EC-specific Eln^fl/fl/Cre⁺ mesenteric arteries demonstrated an abnormal IEL morphology marked by the absence of distinct fenestrations in an otherwise solid matrix layer (Figure 2A). Localization of Hbα, an MEJ-resident protein in resistance artery ECs, is typically characterized by distinct, high intensity puncta that correspond to HIEL³⁸ in the en face view of control mice (Figure 2B). This distinct localization pattern was lost in EC-specific Eln^fl/fl/Cre⁺ arteries (Figure 2C), where Hbα staining was more diffuse, covering more area with increased number of individual puncta (Figure 2D–E; P=0.0480 and 0.0144, respectively), and less intense (Figure 2F; P=0.0107).⁴³ These results align with our previous observation of another MEJ component phosphatidylserine in EC-specific Eln^fl/fl/Cre⁺ resistance arteries, where the lack of MEJs also compromised phosphatidylserine localization and abundance.¹¹

Figure 2. Loss of internal elastic lamina disrupts Hbα distribution in the endothelium.

A, Representative en face view of third-order mesenteric arteries from Eln^fl/fl/Cre⁻ control and EC-specific Eln^fl/fl/Cre⁺ mice with nuclei in blue and the internal elastic lamina (IEL) in grey. Scale bar 30 μm. En face view of third-order mesenteric arteries from Eln^fl/fl/Cre⁻ control (B) and EC-specific Eln^fl/fl/Cre⁺ (C) mice. Arteries are stained for Hbα (magenta) and IEL (hydrazide, grey). White arrows indicate the presence of holes in the IEL (B) or residual IEL staining (C). Fiji was used to produce a heatmap (right column), which displays the signal intensity of Hbα ranging from low (blue) to high (white). Area with Hbα positive staining per image (D), total puncta count (E), and percentage of high intensity staining (F) in Eln^fl/fl/Cre⁻ control versus EC-specific Eln^fl/fl/Cre⁺ mesenteric arteries. Scale bar 10 μm. N=3 mice per group. Student’s t-test.

Additionally, we found that both Hbα mRNA and protein levels were markedly reduced in endothelial-specific Eln^fl/fl/Cre⁺ mice as determined by examining expression in mesenteric resistance arteries (Figure 3A–B; P=0.0015 and P=0.0802, respectively), thoracodorsal resistance arteries (Figure S2A), and the highly vascularized lung tissue which is rich in ECs (Figure S2B; P=0.0145). This indicates the reduction in Hbα levels with EC-specific Eln deletion occurs across vascular beds. Since the function of Hbα in the resistance artery endothelium is to bind to and regulate the function of eNOS,¹⁵ we evaluated the extent of Hbα interactions with eNOS by proximity ligation assay (PLA) in intact third-order mesenteric arteries prepared en face. Each punctum in a PLA image represents a protein-protein interaction where the two proteins being evaluated are within 40 nm of each other. The number of Hbα-eNOS interactions showed a significant reduction in EC-specific Eln^fl/fl/Cre⁺, where Eln^fl/fl/Cre⁻ control ECs had 111.7 puncta per EC nuclei, whereas Eln^fl/fl/Cre⁺ mice had only 75.23 puncta per EC nuclei on average (Figure 3C–D; P=0.0202). The average puncta size also followed this trend with a significant decrease in puncta size in EC-specific Eln^fl/fl/Cre⁺ mice (P=0.0305; 0.3668 μm vs 0.1798 μm respectively). Negative controls with either eNOS or Hbα primary antibody confirmed the specificity of the PLA assay with no detectable signal (Figure 3C). Furthermore, en face staining of eNOS in EC-specific Eln^fl/fl/Cre⁺ third-order mesenteric arteries demonstrated increased eNOS diffusivity corresponding to a significant increase in signal area (P=0.0128) compared to a more compact eNOS localization to the perinuclear region in Eln^fl/fl/Cre⁻ control ECs (Figure 3E–F). There was also a nonsignificant trend to an increase in the number of eNOS puncta present in EC-specific Eln^fl/fl/Cre⁺ mesenteric arteries (Figure 3F; P=0.1318). Taken together, both the reduction in Hbα levels as well as its mis-localization with Eln deletion in the endothelium contributes to reduced Hbα-eNOS interactions and altered eNOS localization.

Figure 3. Loss of elastin decreases Hbα interaction with eNOS as well as total Hbα.

A, RT-qPCR of mesenteric arteries to evaluate the relative mRNA expression of Hba1/Hba2. N=5. B, Western blot showing Hbα levels in mesenteric arteries from Eln^fl/fl/Cre⁻ control and EC-specific Eln^fl/fl/Cre⁺ mice. Protein sizes in kDa. Quantification of Hbα protein levels normalized to total protein. N=5. C, Hbα and eNOS interactions shown in red via PLA in en face preparations of third-order mesenteric arteries from Eln^fl/fl/Cre⁻ control and EC-specific Eln^fl/fl/Cre⁺ mice. Representative images of negative controls on Eln^fl/fl/Cre⁻ control arteries where primary antibodies for either eNOS or Hbα were used independently. D, PLA puncta per EC nuclei as well as the average puncta size were quantified via automatic thresholding in Eln^fl/fl/Cre⁻ control and EC-specific Eln^fl/fl/Cre⁺ arteries. N=4 arteries from 3 mice. Scale bar 10 μm. E, Representative en face views of eNOS staining (yellow) in third-order mesenteric arteries from Eln^fl/fl/Cre⁻ control and EC-specific Eln^fl/fl/Cre⁺ mice. Scale bar 20 μm. F, Area with eNOS positive staining per image (signal area) and total eNOS puncta count (number of puncta) in Eln^fl/fl/Cre⁻ control versus EC-specific Eln^fl/fl/Cre⁺ arteries. N=3. Student’s t-test.

To test if this reduced interaction had a functional effect on the vasodilation of intact arteries, we assessed the dilatory capacity of isolated, intact third-order mesenteric arteries to ACh in the presence or absence of the NOS inhibitor N(ω)-nitro-L-arginine methyl ester hydrochloride (L-NAME). The Eln^fl/fl/Cre⁻ control arteries showed no differences in percent dilation with the addition of L-NAME, a common observation in pressurized resistance arteries where the contribution of NO is minimal. EC-specific Eln^fl/fl/Cre⁺ arteries had a lower dilation overall in response to acetylcholine (ACh), indicating reduced dilatory potential (Figure 4A). Representative traces are demonstrated in Figure S3A. However, what was particularly striking was that the presence of L-NAME had an effect on ACh-induced dilation, with a significantly reduced dilation in EC-specific Eln^fl/fl/Cre⁺ arteries that was not seen with L-NAME treatment of Eln^fl/fl/Cre⁻ control arteries (Figure 4A). This phenotype was specific to smaller, resistance arteries since pressure myography of the larger, conduit carotid arteries from both genotypes showed no differences in ACh-mediated relaxation with or without L-NAME treatment (Figure S4). Thus, reinforcing that ECs are the major contributor to the IEL in resistance arteries, but not large elastic arteries since the IEL remains intact with EC-specific Eln ablation.^10,11 The NO component to vasodilation was calculated as the percent difference of vasodilation when L-NAME was included compared to vehicle controls at each dose. EC-specific Eln^fl/fl/Cre⁺ arteries had a 61.59% contribution of NO to the dilation elicited by ACh, whereas dilation in Eln^fl/fl/Cre⁻ control arteries were not inhibited by L-NAME (Figure 4B). However, when the expression levels of eNOS and p-eNOS were assessed in Eln^fl/fl/Cre⁻ control and EC-specific Eln^fl/fl/Cre⁺ mesenteric vasculature via Western blot⁴⁴ (Figure 4C), the levels of phosphorylation of eNOS at Ser1177 (P=0.2291) remained unchanged despite an upregulation of total eNOS levels in EC-specific Eln^fl/fl/Cre⁺ mesenteric arteries (P=0.0056). This suggests that although the abundance of total eNOS is increased with EC-specific Eln deletion, the signaling pathways responsible for activating eNOS through its phosphorylation at Ser1177 are not compromised. The mRNA expression of EDH-associated endothelial potassium channels Kcnn4 (IK_Ca) was also unchanged in Eln^fl/fl/Cre⁺ mesenteric arteries compared to controls (Figure 4D). The mRNA for Kcnn1 (SK_Ca) and Kcnj2 (Kir2.1) trended downwards in EC-specific Eln^fl/fl/Cre⁺ mice but were not statistically significant (Figure 4D). These findings are in line with previous results demonstrating no change in Kir2.1 levels or function in EC-specific Eln^fl/fl/Cre⁺ mesenteric vasculature compared to controls based on Kir2.1 immunoblotting and NS309 dose-responses.¹¹ Therefore, the increased NO component of dilation in EC-specific Eln^fl/fl/Cre⁺ mesenteric arteries is unlikely due to increased activation of eNOS or reduction of the presence of key ion channels in the canonical EDH pathway.

Figure 4. Loss of internal elastic lamina favors NO signaling.

A, Acetylcholine (ACh) dose-response curves of third-order mesenteric arteries from Eln^fl/fl/Cre⁻ controls and EC-specific Eln^fl/fl/Cre⁺ mice with or without 30 μM of L-NAME. N=5. Two-way ANOVA. Following cannulation, the artery was equilibrated by increasing the pressure to 60 mmHg and temperature to 37°C before beginning treatments. Arteries were then pre-constricted with 1 μM phenylephrine (PE), and the first ACh dose was applied after the arterial diameter plateaued for 5 min following PE constriction. Percent vasodilation was calculated as the percent increase in diameter from PE pre-constriction. B, Quantification of the NO component of dilation in Eln^fl/fl/Cre⁻ controls and EC-specific Eln^fl/fl/Cre⁺ third-order mesenteric arteries, where the difference between the average dilation of vehicle and L-NAME was calculated for each dose. Each data point represents the %NO contribution at one ACh concentration ranging from −5 to −1 log M. N=5. C, Western blot showing expression of eNOS and p-eNOS in Eln^fl/fl/Cre⁻ controls and EC-specific Eln^fl/fl/Cre⁺ mesenteric vasculature. Protein sizes in kDa. Western blot quantification, where the expression of p-eNOS was normalized to eNOS and total eNOS was normalized to total protein. N=5. D, RT-qPCR results on mesenteric artery lysates to evaluate the relative mRNA expression of Kcnn4 (KCa3.1), Kcnn1 (SKCa), or Kcnj2 (Kir2.1). N=8–10 mice per group. Student’s t-test.

We further investigated the contractile ability of EC-specific Eln^fl/fl/Cre⁺ third-order mesenteric arteries to determine if vasoconstriction was also impaired. EC-specific Eln^fl/fl/Cre⁺ arteries exhibited a reduced capacity to constrict when stimulated with both phenylephrine (PE) and KCl (Figure 5A; P≤0.0001; representative traces in Figure S2B). These findings corresponded to a trend to increased maximum lumen diameter and vascular cross-sectional area in EC-specific Eln^fl/fl/Cre⁺ mesenteric arteries compared to Eln^fl/fl/Cre⁻ controls (Figure 5B; P=0.0969 and P=0.1031, respectively). Diameter measurements were taken in Ca²⁺-free conditions after stabilization at a pressure of 60 mmHg. Interestingly, RT-qPCR results indicated this decreased contraction was not due to reductions in α smooth muscle actin (Acta2; αSMA) or integrin (Itgav, Itgb3) expression (Figure 5C; P=0.1835, 0.2445, and 0.5215, respectively). Since the same mesenteric artery lysates were used to assess Eln expression (significantly reduced in Eln^fl/fl/Cre⁺ lysates),¹¹ the expression of Acta2, Itgav, or Itgb3 could be plotted relative to the elastin mRNA content in each sample. Although there were no significant reductions in the levels of key SMC contractile machinery, positive correlations were identified with Eln expression for Acta2, Itgav, and Itgb3 in EC-specific Eln^fl/fl/Cre⁺ but not Eln^fl/fl/Cre⁻ control mesenteric arteries (Figure 5D), suggesting elastin may drive expression of these genes. However, at the protein level, tissue immunofluorescence of thoracodorsal arteries demonstrated αSMA had similar levels and localization between genotypes, matching the Acta2 mRNA expression (Figure 5E). Furthermore, despite a thinner IEL in EC-specific Eln^fl/fl/Cre⁺ mesenteric arteries,¹¹ collagen architecture appeared similar between genotypes by Picrosirius Red staining, and the circumference, normalized wall thickness and normalized externa thickness were not significantly different from Eln^fl/fl/Cre⁻ control arteries in histological cross-sections stained with Masson’s Trichrome⁴⁵ (Figure S5; P=0.5604, 0.2935, and 0.8232, respectively). The thickness of the tunica media was also unchanged from control mice, demonstrating no reduction in SMC cross-sectional area (Figure S5B; P=0.8232). Unlike the mesenteric arteries, there were minor, but significant differences in some structural parameters between Eln^fl/fl/Cre⁻ control and EC-specific Eln^fl/fl/Cre⁺ large arteries (Figure S6, S7). Eln^fl/fl/Cre⁺ aortas had significantly increased externa thickness and significantly reduced media thickness compared to controls (P<0.05), but no differences in normalized wall thickness or circumference. Eln^fl/fl/Cre⁺ carotids had significantly decreased wall thickness (P<0.05), but all other parameters showed no differences. Despite these subtle variations in collagen abundance and organization, our previous findings showed that the elastin area of both carotid and aortic arteries were similar between genotypes.¹¹ Taken together, this demonstrates that the reduced contractile ability of EC-specific Eln^fl/fl/Cre⁺ mesenteric arteries may not be due to reduced SMC thickness in the vascular wall or changes to the α-SMA contractile protein present in the SMCs.

Figure 5. Loss of resistance artery EC elastin inhibits contractility.

A, Third-order mesenteric arteries were constricted with 10 μM PE or 30 mM KCl. N=4. Percent constriction induced by either PE or KCl treatment was calculated as the percent decrease in diameter from 10 μM ACh pre-dilation. B, Maximum diameters and vascular cross-sectional areas of third-order mesenteric arteries from Eln^fl/fl/Cre⁻ controls and EC-specific Eln^fl/fl/Cre⁺ mice. Maximum diameter measurements were taken in Ca²⁺-free conditions after vessel stabilization at a pressure of 60 mmHg and used to calculate vascular cross-sectional area. C, RT-qPCR results on mesenteric artery lysate to evaluate the relative mRNA expression of Acta2 (αSMA), Itgav (integrin subunit αV), or Itgb3 (integrin subunit β3). Student’s t-test. D, mRNA expression of Acta2, Itgav, or Itgb3 was correlated with respect to Eln mRNA expression within the same mesenteric artery lysates from Eln^fl/fl/Cre⁻ controls and EC-specific Eln^fl/fl/Cre⁺ mice. N=6–9 mice per group. E, Representative tissue immunofluorescence of αSMA (magenta) in Eln^fl/fl/Cre⁻ control and EC-specific Eln^fl/fl/Cre⁺ thoracodorsal arteries. Internal elastic lamina (IEL) in grey, nuclei in blue. Scale bar 20 μm. N=3.

With enhanced NO signaling and impaired contractility in EC-specific Eln^fl/fl/Cre⁺ mesenteric arteries, as well as decreased myogenic tone in resistance arteries of EC-specific Eln^fl/fl/Cre⁺ mice reported previously,¹¹ we assessed whether these phenotypes were reflected physiologically in these mice. Surprisingly, we found no differences in daily systolic, diastolic and mean arterial blood pressure or heart rate between genotypes (Figure 6). This indicates EC-specific Eln^fl/fl/Cre⁺ mice are normotensive despite the compromised vasodilatory and contractile properties observed in their resistance arteries. We further investigated plasma sodium or blood urea nitrogen (BUN) concentrations obtained from anesthetized mice to assess kidney function but saw no differences in (Figure S8A–B; P=0.2422 and 0.2080, respectively). However, we did observe a significant increase in plasma renin levels, suggesting a renal compensation may contribute to maintaining baseline blood pressure in Eln^fl/fl/Cre⁺ mice (Figure S8C; P=0.0005), but further research is needed to confirm this hypothesis.

Figure 6. Mice lacking an internal elastic lamina have normalized blood pressure.

Systolic pressure (A), diastolic pressure (B), mean arterial pressure (C) and heart rate (D) from radiotelemetry blood pressure measurements in Eln^fl/fl/Cre⁻ control and EC-specific Eln^fl/fl/Cre⁺ mice. Measurements were taken every minute for 5 days and values corresponding to each time frame (day, night or 24 hr) were averaged. N=5–6 mice per group.

DISCUSSION

The genetic deletion of elastin in ECs has previously been shown to be unique in compromising the IEL of the resistance arteries, and not conduit arteries.¹⁰ However, the functional outcomes of endothelial-specific Eln^fl/fl/Cre⁺ mesenteric arteries was yet to be described. Here we demonstrate mesenteric resistance arteries in EC-specific Eln^fl/fl/Cre⁺ mice have significantly decreased vasoreactive properties, including endothelial-mediated dilation and smooth muscle constriction that we posit are due to changes in Hbα-eNOS interaction.

Resistance arteries in EC-specific Eln^fl/fl/Cre⁺ mice have thinner IELs and lack the canonical circular gaps absent of protein termed HIEL.^10,11 Without designated anatomical sites for heterocellular contact (e.g., HIEL), the key vasodilation protein Hbα loses its canonical organization of high intensity puncta and becomes more diffuse and unorganized (i.e., loss of binding to eNOS). The overall Hbα protein is also decreased in these mice, suggesting Hbα protein expression may be in part be driven by its organization into distinct signaling domains via EC-elastin deposition. The decreased interaction between eNOS and Hbα has been shown to significantly increase the amount of NO available^15,16,36, which may explain the increased NO-mediated dilation and the decreased SMC contraction.

Functionally, the decrease in Hbα expression and more diffuse localization in the arterial wall corresponded to reduced interactions with eNOS in ECs of elastin-deficient arteries. Given the negative regulation of eNOS by Hbα, their decreased interactions should result in increased endothelial NO signaling.^15,16,36 We demonstrated this is the case via pressure myography on intact resistance arteries from EC-specific Eln^fl/fl/Cre⁻ controls and Eln^fl/fl/Cre⁺ mice, where application of the eNOS inhibitor L-NAME significantly decreased ACh dilation in arteries lacking canonical HIEL, whereas control arteries maintained dilation as expected. Without the ECs depositing elastin and contributing to the unique morphology of the HIEL, EC-specific Eln^fl/fl/Cre⁺ arteries undergo a vasodilatory profile switch to rely on NO more heavily than EDH-mediated vasodilation. This increase in NO signaling was not due to increased activation of eNOS or decreased expression of key channels in EDH-based signaling, further confirming the expected functional consequence for the reduced Hbα-eNOS interaction.^15,36 mRNA levels of EDH-associated potassium channels were also unchanged in EC-specific Eln^fl/fl/Cre⁺ compared to Eln^fl/fl/Cre⁻ control vessels, and we found previously that the dilation to EDH-associated agonist NS309 was not significantly different in EC-specific Eln^fl/fl/Cre⁺ arteries.¹¹ This result implies EC deposition of elastin is a key regulator of protein localization in the endothelium, and consequentially the molecular pathways of vasodilation. These findings are complementary to the data demonstrating the formation of MEJs in a conduit artery shifts the vasodilatory mechanisms in the artery.¹⁶ That is, when MEJs are induced in conduit arteries, vasodilation switches from being NO-mediated to primarily EDH-mediated, and when canonical MEJ sites are disrupted in resistance arteries there is a switch to NO-mediated vasodilation.

Besides the L-NAME inhibition of NO, the exact molecular mechanisms responsible for the changes in vasodilation and vasoconstriction when vessels lack EC elastin remain unresolved. We initially interpret that the changes in ACh-mediated vasodilation with L-NAME are due to more free NO. This is because of data from previous reports demonstrating a lack of eNOS/Hbα interaction causes increased NO responses,^46–49 which is consistent with our findings. However, the changes in L-NAME responses could also be due to decreased sensitivity to cGMP in SMCs,⁵⁰ or changes in oxidative stress,⁵¹ prostaglandins,⁵² or epoxygenase-derived epoxyeicosatrienoic acids.⁵³ One potential pathway that could be probed to examine the changes in the mechanisms of vasodilation/vasoconstriction is integrin signaling due to the centralized role of integrins in communicating changes in the extracellular matrix to promote changes in cellular function. In this study, we showed no statistical differences in αv and β3 integrins between genotypes at the mRNA level, but instead a tight relationship between elastin and integrin expression where the amount of αv and β3 integrin appears to correlate with the amount of elastin present in resistance arteries. This suggests integrin interaction may in part drive Hbα expression and these associations may be a potential area to explore for understanding the changes in vasodilation/vasoconstriction. However, others have shown in aortic stenosis a negative correlation with the amount of β3, showing a clear increase in β3 as elastin decreases.⁵⁴ Therefore, there may not be a simple elastin-integrin interaction in resistance arteries. Besides integrins, this phenomenon could also be explained by the fact that the endothelial plasma membrane is disorganized upon loss of elastin. Matrix regulation of cellular protein organization (e.g., cell-cell junctions) is observed in many cell types and can be driven by integrins, as well as a host of other factors (e.g., ^55,56). More work on this observation is required to elucidate the interaction between elastin, integrins and Hbα at the cellular level, but whatever the molecular mechanism is that alters the vasoconstriction/vasodilation of EC-specific Eln^fl/fl/Cre⁺ arteries, it is clear elastin plays an important and unexpected role.

Despite these significant alterations in EC and SMC function, blood pressure measurements taken from freely moving, non-anesthetized mice show no significant differences in baseline blood pressure in EC-specific Eln^fl/fl/Cre⁺ mice. It is possible that the normotensive phenotype we observed in EC-specific Eln^fl/fl/Cre⁺ mice may have resulted from a compensatory increase in renin-angiotensin-aldosterone system activity to ensure normal blood pressure, which has some similarities to global Eln^+/− mice, but also some important differences. Eln^+/− mice were shown to have a significant increase in plasma renin, but were correspondingly hypertensive,⁶ although the plasma renin levels in SMC-specific Eln knockout mice is not known. More work into the blood pressure phenotype such as characterizing the renin-angiotensin-aldosterone system and kidney function of Eln^fl/fl/Cre⁺ mice is warranted based on these findings.

In the spontaneous hypertensive rat model, the IEL of mesenteric resistance arteries exhibited stiffer elastin and smaller fenestrae which preceded arterial narrowing.^57,58 However, in our EC-specific Eln^fl/fl/Cre⁺ mice there was no change in media thickness or diameter measured from histological cross-sections of mesenteric arteries. In addition, there was a decrease in resistance artery contractility of EC-specific Eln^fl/fl/Cre⁺ observed at both the level of SMC function via depolarization with KCl, and agonist-induced constriction with PE. An RT-qPCR screen of contractile machinery and integrins demonstrated that mRNA trended to a decrease in EC-specific Eln^fl/fl/Cre⁺ mesenteric arteries but was statistically unchanged. Similarly, αSMA appeared similar in protein abundance and localization between genotypes. As noted above, the change in integrin αv and β3 corresponding with Eln transcript levels may provide clues to changes in contractility responses, such that changes in integrin expression may alter expression of calcium channels in SMC. For example, there is evidence elastin can regulate L-type Ca²⁺ channel opening,⁵⁹ which may provide the most logical explanation for the significant decrease in contraction we detected in EC-specific Eln^fl/fl/Cre⁺ resistance arteries. It is also possible the increase in NO signaling observed in EC-specific Eln^fl/fl/Cre⁺ arteries may be responsible for the decrease in SMC contractility. Investigating the mechanisms of elastin-dependent contractility responses and crosstalk with integrin signaling may be an especially interesting area for future investigation.

We recognize that there are two limitations in our study regarding the potential effects of aging and sex. Firstly, we did not examine an aging phenotype, analyzing mice from 10–20 weeks of age. Elastin haploinsufficiency in humans and in global Eln^+/− mice both show progressive alterations in vascular function as aging progresses,^8,60 meaning it is possible the properties of EC-specific Eln^fl/fl/Cre⁺ resistance arteries may change over time. However, it is unlikely that resistance artery function would improve as time progresses but rather become more severe with age. The blood pressure phenotype may be altered, and the mice may become hypertensive since renin levels are unlikely to decrease with age. Thus, we do not suspect there would be a dramatic aging phenotype in these mice, especially since there is a lack of early death observed in EC-specific Eln^fl/fl/Cre⁺ mice at least out to 6 months of age (data not shown). The second limitation is that only males were used in this study. However, previous experiments using global Eln^+/− did not exhibit any differences in vascular phenotypes between sexes.⁶¹ As such, more research in these two areas would be helpful to fully elucidate how elastin deletion from ECs alters hemodynamic properties.

PERSPECTIVES

In this study, we show EC deletion of elastin results in a resistance artery-specific vasoreactive phenotype. This includes a significant reduction in the number of HIEL and the levels of the MEJ protein Hbα as well as Hbα dispersion in the endothelium. The functional effects of the change in Hbα were a NO-dominant, EC-specific mechanism of vasodilation, which is akin to large artery vasodilatory mechanisms that are normally devoid of MEJs and Hbα. In addition, we found SMC have a significant reduction in the capacity to contract to KCl and PE, but the overall structure of the resistance arteries was not changed. Regardless of the significant decrease in contractile capacity and increased NO-dilatory potential, there was no change in blood pressure. Our work showcases the novel interaction of the structural component elastin with resistance artery function and indicates a potential avenue for developing hypertensive therapeutics. Future work should further investigate the interplay between elastin levels and the renin-angiotensin-aldosterone system which may prevent a hypotensive phenotype which would be expected in the absence of EC elastin.

Supplementary Material

SUPPLEMENTAL MATERIAL

Expanded Materials and Methods

Figures S1–S8

Major Resources Table

Supplemental References

Uncropped Blots and Gels

NOVELTY AND RELEVANCE.

What Is New?

• Genetic ablation of endothelial elastin results in impairment of multiple vasoreactive properties of resistance arteries, including fundamental dilation and constriction mechanisms.

• The loss of elastin in resistance arteries strongly correlates with a loss of Hbα in endothelium and interaction with eNOS resulting in a NO-mediated dilation rather than EDH.

What is Relevant?

• The loss of elastin in resistance arteries is not reflected in changes to blood pressure.

Clinical/Pathophysiological Implications?

• It may be important to focus on treating resistant artery function in patients with elastin deficiency disorders, regardless of their blood pressure.

Nonstandard Abbreviations and Acronyms

ACh

acetylcholine

EC

endothelial cell

EDH

endothelial-derived hyperpolarization

Eln

elastin

eNOS

endothelial nitric oxide synthase

Hbα

alpha hemoglobin

HIEL

holes in the internal elastic lamina

IEL

internal elastic lamina

L-NAME

N(ω)-nitro-L-arginine methyl ester hydrochloride

MEJ

myoendothelial junction

PE

phenylephrine

SMC

smooth muscle cell