Impaired activity of adherens junctions contributes to endothelial dilator dysfunction in ageing rat arteries

Abstract

Key points

• Ageing‐induced endothelial dysfunction contributes to organ dysfunction and progression of cardiovascular disease. VE‐cadherin clustering at adherens junctions promotes protective endothelial functions, including endothelium‐dependent dilatation.

• Ageing increased internalization and degradation of VE‐cadherin, resulting in impaired activity of adherens junctions.

• Inhibition of VE‐cadherin clustering at adherens junctions (function‐blocking antibody; FBA) reduced endothelial dilatation in young arteries but did not affect the already impaired dilatation in old arteries. After junctional disruption with the FBA, dilatation was similar in young and old arteries.

• Src tyrosine kinase activity and tyrosine phosphorylation of VE‐cadherin were increased in old arteries. Src inhibition increased VE‐cadherin at adherens junctions and increased endothelial dilatation in old, but not young, arteries. Src inhibition did not increase dilatation in old arteries treated with the VE‐cadherin FBA.

• Ageing impairs the activity of adherens junctions, which contributes to endothelial dilator dysfunction. Restoring the activity of adherens junctions could be of therapeutic benefit in vascular ageing.

Abstract

Endothelial dilator dysfunction contributes to pathological vascular ageing. Experiments assessed whether altered activity of endothelial adherens junctions (AJs) might contribute to this dysfunction. Aortas and tail arteries were isolated from young (3–4 months) and old (22–24 months) F344 rats. VE‐cadherin immunofluorescent staining at endothelial AJs and AJ width were reduced in old compared to young arteries. A 140 kDa VE‐cadherin species was present on the cell surface and in TTX‐insoluble fractions, consistent with junctional localization. Levels of the 140 kDa VE‐cadherin were decreased, whereas levels of a TTX‐soluble 115 kDa VE‐cadherin species were increased in old compared to young arteries. Acetylcholine caused endothelium‐dependent dilatation that was decreased in old compared to young arteries. Disruption of VE‐cadherin clustering at AJs (function‐blocking antibody, FBA) inhibited dilatation to acetylcholine in young, but not old, arteries. After the FBA, there was no longer any difference in dilatation between old and young arteries. Src activity and tyrosine phosphorylation of VE‐cadherin were increased in old compared to young arteries. In old arteries, Src inhibition (saracatinib) increased: (i) 140 kDa VE‐cadherin in the TTX‐insoluble fraction, (ii) VE‐cadherin intensity at AJs, (iii) AJ width, and (iv) acetylcholine dilatation. In old arteries treated with the FBA, saracatinib no longer increased acetylcholine dilatation. Saracatinib did not affect dilatation in young arteries. Therefore, ageing impairs AJ activity, which appears to reflect Src‐induced phosphorylation, internalization and degradation of VE‐cadherin. Moreover, impaired AJ activity can account for the endothelial dilator dysfunction in old arteries. Restoring endothelial AJ activity may be a novel therapeutic approach to vascular ageing.

Key points

• Ageing‐induced endothelial dysfunction contributes to organ dysfunction and progression of cardiovascular disease. VE‐cadherin clustering at adherens junctions promotes protective endothelial functions, including endothelium‐dependent dilatation.

• Ageing increased internalization and degradation of VE‐cadherin, resulting in impaired activity of adherens junctions.

• Inhibition of VE‐cadherin clustering at adherens junctions (function‐blocking antibody; FBA) reduced endothelial dilatation in young arteries but did not affect the already impaired dilatation in old arteries. After junctional disruption with the FBA, dilatation was similar in young and old arteries.

• Src tyrosine kinase activity and tyrosine phosphorylation of VE‐cadherin were increased in old arteries. Src inhibition increased VE‐cadherin at adherens junctions and increased endothelial dilatation in old, but not young, arteries. Src inhibition did not increase dilatation in old arteries treated with the VE‐cadherin FBA.

• Ageing impairs the activity of adherens junctions, which contributes to endothelial dilator dysfunction. Restoring the activity of adherens junctions could be of therapeutic benefit in vascular ageing.

Abbreviations

AJ

adherens junctions

FBA

function‐blocking antibody

l‐NAME

N ^ω‐nitro‐l‐arginine methyl ester

ROCK

Rho‐associated protein kinase

TNFα

tumour necrosis factor‐α

Introduction

Ageing causes structural and functional deterioration of the arterial system that contributes to cardiovascular disease and organ dysfunction (Lakatta et al. 2009; Laurent, 2012). Endothelial dysfunction is a key contributor to this process of vascular ageing. Ageing arterial endothelial cells are compromised: their production of protective NO is reduced, their barrier function is impaired, they are highly susceptible to apoptosis, and they have a prominent pro‐inflammatory phenotype, including production of pathological mediators (Lakatta et al. 2009; Goel et al. 2010; Ungvari et al. 2010; Flavahan et al. 2016). Ageing endothelial cells are caught in a cycle of chronic stress, both producing and responding to pathological mediators, and their resulting frailty is considered a key instigator of vascular ageing (Ungvari et al. 2010; Wang et al. 2010a).

Maintenance of a normal protective endothelial cell layer is an active rather than a passive process (Murakami et al. 2008; Dejana et al. 2009; Flavahan et al. 2013; Flavahan, 2017). Protective features of arterial endothelial cells become fully engaged during the early postnatal period (Flavahan et al. 2013; Flavahan & Flavahan, 2014; Chang et al. 2016), and appear to be mediated by increased clustering of VE‐cadherin at adherens junctions, which promotes protective signalling including amplifying endothelial NO‐mediated dilatation (Flavahan et al. 2013). Indeed, endothelial adherens junctions promote endothelial and vascular stability, whereas their disruption compromises vascular integrity, increases vascular inflammation and fragility, and promotes the development of cardiovascular diseases (Murakami et al. 2008; Dejana et al. 2009; Giannotta et al. 2013; Flavahan, 2017). The present experiments were performed to determine whether ageing impairs the activity of endothelial adherens junctions and whether this could contribute to the ageing‐induced dysfunction in endothelial dilatation.

Methods

Ethical approval, animals

Young (3–4 months) and old (22–24 months) male F344 rats were obtained from Charles River and their National Institute on Aging colony, and killed by CO₂ asphyxiation. Aortas and tail arteries were rapidly removed and placed in cold Krebs–Ringer bicarbonate solution (Flavahan et al. 2016). Animal use was approved by the Institutional Animal Care and Use Committee and complied with the NIH Guide for the Care and Use of Laboratory Animals.

Functional responses

Tail arteries were cannulated with micropipettes, secured within a microvascular chamber (Living Systems, St Albans City, VT, USA), and maintained at a transmural pressure (P _TM) of 60 mmHg in the absence of flow (Flavahan et al. 2016). The arteries were superfused with Krebs–Ringer bicarbonate solution (37°C, pH 7.4, 16% O₂–5% CO₂, balance N₂), and the chamber was placed on the stage of an inverted microscope (Flavahan et al. 2016). Arterial diameter was continuously monitored and recorded (Biopac, Santa Barbara, CA, USA) (Flavahan et al. 2016). Concentration–effect curves to vasodilators were obtained during constriction to phenylephrine, and acetylcholine responses were determined in paired arteries with one segment serving as control. When analysing the effects of N ^ω‐nitro‐l‐arginine methyl ester (l‐NAME; 100 μm) to inhibit NO synthase or saracatinib (5 μm) to inhibit Src kinase (refers to the Src family of tyrosine kinases), arteries were incubated with these drugs for 60–90 min before and during exposure to acetylcholine. To determine the role of VE‐cadherin clustering at adherens junctions on vasodilator responses, the arteries were exposed to intraluminal administration of a function blocking antibody against VE‐cadherin (BV13, 50 μg ml⁻¹; eBioscience, San Diego, CA, USA) or a control antibody (50 μg ml⁻¹; eBioscience) for 2.5–3 h before analysing vasodilator responses (Flavahan et al. 2013). This antibody binds to an N‐terminal sequence in the most proximal extracellular domain (EC1) of VE‐cadherin (Corada et al. 2001; May et al. 2005).

Immunofluorescence

Tail arteries were mounted in specialized ‘flipper’ chambers (Living Systems) that enabled the blood vessel assembly to be rapidly (∼1 s) transferred from control solution to paraformaldehyde (3%, 4°C, 30 min) (Flavahan et al. 2013). Arteries were cut open longitudinally during fixation. Arteries were rinsed in phosphate‐buffered saline (PBS; 3 × 10 min), permeabilized (Triton X‐100, 0.5%, 15 min), rinsed again (PBS, 3 times) and then incubated in donkey serum (1.5%, 15 min) to reduce non‐specific binding. Arteries were incubated overnight with a goat polyclonal antibody to VE‐cadherin (C‐19, 1:500 dilution, Santa Cruz Biotechnology, Dallas, TX, USA). Arteries were rinsed (PBS, 3 × 15 min), then incubated (120 min) with Rhodamine or Alexa Fluor 488‐labelled donkey anti‐goat antibody (1:200, Jackson ImmunoResearch Laboratories, West Grove, PA, USA), and DRAQ5 (5 μm, nuclear stain, Biostatus, Loughborough, UK). Samples were viewed using a (Leica, Buffalo Grove, IL, USA), AOBS‐equipped SP5 laser scanning microscope. Images (1024 × 1024 pixels) were obtained using sequential acquisition, a pinhole of 1 Airy unit, scan speed of 400 Hz, six‐line averaging, an optical zoom of 3.0, and excitation/emission settings for Alexa Fluor 488 (488 nm/492–541 nm), Rhodamine (543 nm/555–620 nm) and DRAQ (633 nm/650–750 nm). For quantitative comparison of fluorescence, young and old arteries or control and treated arteries were processed at the same time using the same instrument settings. Z‐Stacks were obtained for the complete endothelial layer (0.25 micron steps). The width and intensity of adherens junctions was assessed using line profiles of VE‐cadherin fluorescent staining, as previously described (Flavahan et al. 2013). Maximal Z‐stack projections of line profiles were generated by selecting pixels with the maximal intensity from the z‐sections. Peak fluorescence intensities are expressed as absolute detector units (Zhao et al. 2015). The width of adherens junctions was determined as the distance in micrometres that VE‐cadherin intensity was above the immediate background level (Flavahan et al. 2013). This approach involved a non‐selective survey of endothelial junctions (Flavahan et al. 2013), although taking care to avoid gaps in the junctions. This is in contrast to a previous study that purposefully measured the widest part of VE‐cadherin junctional staining with the apparent goal of assessing gaps in the junctions (Huynh et al. 2011).

Arterial lysis

For sequential lysis, rat arteries were first lysed with TTX lysis buffer (50 mm Tris pH 7.6, 150 mm NaCl, 0.5% Triton X‐100, 0.5% deoxycholate, 2 mm EDTA, 20 μg ml⁻¹ leupeptin, 20 μg ml⁻¹ aprotinin, 10 mm NaF, 0.5 mm 4‐benzenesulfonyl fluoride hydrochloride (AEBSF)–HCl, 1 mm sodium orthovanadate), followed by centrifugation at 15 000 g (4°C, 10 min), to obtain the TTX‐soluble lysate (supernatant). The TTX‐insoluble pellets were subsequently lysed with sodium dodecyl sulfate (SDS) lysis buffer (TTX lysis buffer with 2% SDS), boiled for 5 min, and then clarified by centrifugation (15 000 g, room temperature, 10 min) to obtain the TTX‐insoluble SDS‐soluble lysates. Rat arteries were also lysed using a one‐step lysis method, whereby rat arteries were directly lysed with the SDS lysis buffer, boiled for 5 min before centrifugation at 15 000 g at room temperature to obtain the total lysates. Total protein concentrations were measured using a Pierce BCA protein assay kit (Thermo Fisher Scientific, Waltham, MA, USA).

Cell surface labelling

A Pierce Cell Surface Protein Isolation Kit (Thermo Fisher Scientific) was used to label VE‐cadherin on the endothelial cell surface. Briefly, aortas, which had been opened longitudinally, were incubated with Sulfo‐NHS‐SS‐biotin in cold Krebs–Ringer bicarbonate solution at 4°C for 30 min, according to the manufacturer's protocol. The reaction was stopped by the addition of quenching solution. Arterial lysates were then obtained using the one‐step lysis method. The biotinylated surface proteins were purified with immobilized NeutrAvidin gel followed by elution with sample buffer containing DTT, according to the manufacturer's protocol.

Immunoprecipitation and immunoblotting

The reliability of commercially available antibodies to tyrosine‐phosphorylated VE‐cadherin has been questioned (Orsenigo et al. 2012; Wessel et al. 2014). Therefore, immunoprecipitation was used to assess tyrosine phosphorylation of VE‐cadherin. Briefly, 100–300 μg of total protein in immunoprecipitation buffer (25 mm Tris‐Cl, pH 7.2, 150 mm NaCl, 2 mm EDTA, 20 μg ml⁻¹ leupeptin, 20 μg ml⁻¹ aprotinin, 10 mm NaF, 0.5 mm AEBSF–HCl, 1 mm sodium orthovanadate) was incubated with mouse monoclonal phosphotyrosine antibody (clone 4G10, Millipore, Billerica, MA, USA) (3 h, 4°C), followed by protein A/G agarose beads (Thermo Fisher Scientific) (18 h, 4°C). When using SDS‐solubilized proteins, the lysates was diluted with immunoprecipitation buffer to reduce the SDS concentration to 0.1% (this concentration was confirmed not to affect the immunoprecipitation procedure). Immunoprecipitated proteins were eluted by adding Laemmli sample buffer (Bio‐Rad Laboratories, Hercules, CA, USA) to the beads.

Protein samples were boiled in Laemmli sample buffer for 5 min. For improved visualization of different VE‐cadherin species, agarose (1%)‐acrylamide (4%) composite gels were employed, using a slight modification of a previously published approach (Tatsumi & Hattori, 1995). Briefly, agarose (0.3 g) was completely dissolved in 15 ml of deionized water by heating at 100°C before being cooled to 50°C in a water bath; 4 ml of acrylamide solution (30% acrylamide/Bis solution, 19:1, Bio‐Rad) was added to 7.5 ml of 1.5 m Tris buffer (pH 8.8) and 3.5 ml deionized water with 0.2% SDS, and warmed to 37°C in a water bath. The agarose and acrylamide solutions were then mixed, followed by addition of ammonium persulfate and tetramethylethylenediamine, then poured into gel casting cassettes and allowed to solidify at room temperature for 2 h. Proteins were transferred to nitrocellulose membranes (Bio‐Rad), and incubated with the primary antibodies (18 h, 4°C): rat monoclonal VE‐cadherin antibody (BV13, 1:400, eBioscience), mouse monoclonal claudin‐5 (1:2,000, Thermo Fisher Scientific), rabbit monoclonal Src (1:2000) or rabbit monoclonal phospho‐Src (Tyr⁴¹⁶, 1:2000, Cell Signaling Technology, Danvers, MA, USA). After extensive washing, the membranes were incubated with peroxidase‐conjugated secondary antibody (goat‐anti‐rabbit IgG, goat‐anti‐rat IgG, goat‐anti‐mouse IgG) (Jackson ImmunoResearch Laboratories) for 1 h, room temperature. Blots were then developed using enhanced chemiluminescence and quantified using ImageJ.

Drugs

Acetylcholine, l‐NAME and phenylephrine were from Sigma‐Aldrich (St Louis, MO, USA), DEA‐NONOate was from Enzo Life Sciences (Farmingdale, NY, USA), and saracatinib was from Cayman Chemical Co. (Ann Arbor, MI, USA)

Data analysis

Vasomotor responses were expressed as percentage change in baseline diameter. Agonist concentrations causing 50% dilatation of the phenylephrine constriction (EC₅₀) were calculated by regression analysis and compared as log EC₅₀ (Flavahan et al. 2016). Maximum responses were determined as the maximal observed dilatation of the constriction to phenylephrine (Flavahan et al. 2016). For functional and biochemical analyses, data are expressed as means ± SEM, for n animals from which arteries were studied. For image analysis of adherens junctions, data are expressed as means ± SEM for N junctions, assessed using blood vessels from at least three different animals. Statistical evaluation of the data was performed by Student's t test for paired or unpaired observations. When more than two means were compared, analysis of variance was used (Instat 3 software; GraphPad Software, La Jolla, CA, USA). If a significant F value was found, then the Tukey–Kramer test for multiple comparisons was employed to identify differences among groups. Values were considered to be statistically different when P < 0.05.

Results

VE‐cadherin, endothelial adherens junctions and ageing arteries

Biochemical analysis

Endothelial adherens junctions are formed by the clustering and engagement of VE‐cadherin on adjacent endothelial cells. The pattern of VE‐cadherin expression was different between young and old arteries. In young aortas, there was expression of a 140 kDa VE‐cadherin species and a lower molecular mass 115 kDa species (Fig. 1). These species were identified using an antibody directed to the most proximal extracellular domain (EC1) of VE‐cadherin (Fig. 1) (Corada et al. 2001; May et al. 2005) and confirmed using an antibody to EC3/EC4 domains (data not shown). When assessed using a single lysis approach, expression of the 140 kDa VE‐cadherin species was reduced in old compared to young aortas (to 13.4 ± 2.1% of levels in young aortas, n = 5, P < 0.001) and in old compared to young tail arteries (to 18.4 ± 2.2% of levels in young arteries, n = 3, P < 0.001) (Fig. 1 A). In contrast, expression of the 115 kDa VE‐cadherin species was reduced in young compared to old aortas (25.9 ± 9.2% of levels in old aortas, n = 5, P < 0.01) (Fig. 1 A). Expression of the 115 kDa species was reduced in tail arteries compared to aortas, and was generally below detectable levels (Fig. 1 A). Expression of claudin‐5, a transmembrane component of endothelial tight junctions (Harris & Nelson, 2010), was decreased in young compared to old arteries (to 55.1 ± 6.4% of levels in old aortas, n = 3, P < 0.05; to 36.5 ± 8.3% of levels in old tail arteries, n = 3, P < 0.05) (Fig. 1 A).

Figure 1. Immunoblot analysis of VE‐cadherin expression in young (yng) and old arteries.

A, aortas and tail arteries were processed using the single lysis approach with SDS‐containing lysis buffer and representative blots are presented. Upper panel, two prominent bands for VE‐cadherin were observed at 140 and 115 kDa: levels of the 140 kDa species were highest in young compared to old aortas and tail arteries, whereas levels of the 115 kDa species were highest in old compared to young aortas. Lower panel, expression of claudin‐5, a transmembrane protein involved in endothelial tight junctions, was higher in old compared to young arteries. B, aortas and tail arteries were processed using sequential lysis in TTX‐containing buffer (TTX‐soluble fraction) followed by SDS‐containing buffer (TTX‐insoluble fraction). Representative blots are presented. The influence of ageing on levels of the 140 and 115 kDa VE‐cadherin species were the same as described in A. In young and old arteries, the 140 kDa species was captured in the TTX‐insoluble fraction (SDS), whereas the 115 kDa species was present in the TTX‐soluble pool (TTX). C, cell surface proteins in young aortas were labelled using Sulfo‐NHS‐SS‐Biotin and captured with immobilized NeutrAvidin gel. A representative blot from three experiments is presented. The 140 kDa but not the 115 kDa species was localized to the cell surface.

In cultured cells, the clustering of VE‐cadherin at adherens junctions is associated with attachment of the junctional complex to the actin cytoskeleton and translocation of VE‐cadherin from the TTX‐soluble to the TTX‐insoluble pool (Lee et al. 1999). When arteries were processed using sequential lysis, the 140 kDa VE‐cadherin species was present predominantly in the TTX‐insoluble fraction, whereas the 115 kDa species was present predominantly in the TTX‐soluble fraction (Fig. 1 B). In young arteries, which had the highest levels of the 140 kDa VE‐cadherin protein, the level of the 140 kDa protein in TTX‐insoluble fractions was 26.8 ± 4.4‐fold higher (aortas, n = 6, P < 0.001) and 30.9 ± 7.6‐fold higher (tail arteries, n = 6, P < 0.001) than levels in TTX‐soluble fractions (Fig. 1 B). In contrast, in old aortas, which had the highest level of the 115 kDa protein, the level of the protein in the TTX‐insoluble fraction was 3.3 ± 0.8‐fold lower compared to the TTX‐soluble fraction (n = 6, P ≤ 0.01) (Fig. 1 B). As observed with the single lysis approach, expression of the 140 kDa VE‐cadherin species was reduced in old compared to young arteries (to 24.1 ± 4.3% of levels in young aortas, n = 6, P < 0.001; to 28.5 ± 4.8% of levels in young tail arteries, n = 6, P < 0.001) (Fig. 1 B)

When cell‐surface proteins were labelled with impermeant and cleavable Sulfo‐NHS‐SS‐Biotin, followed by NeutrAvidin purification and immunoblot analysis of VE‐cadherin, the 140 kDa but not the 115 kDa species was found to be present on the endothelial cell surface (Fig. 1 C).

Immunofluorescence analysis

In young tail arteries, VE‐cadherin immunofluorescence identified continuous adherens junctions delineating the borders of endothelial cells with a width of 1.40 ± 0.02 μm (N = 349) and peak intensity of 3298.7 ± 30.8 (N = 349) (Fig. 2). In ageing tail arteries, the width of adherens junctions was significantly decreased (1.24 ± 0.02 μm, P < 0.001, N = 419) as was the peak intensity (2607.7 ± 29.1, P < 0.001, N = 419), and there were numerous areas of severe thinning and disrupted junctions (Fig. 2).

Figure 2. Endothelial adherens junctions in young (left) and old (right) tail arteries, assessed by immunofluorescence staining of VE‐cadherin (red; DRAQ5 nuclear staining: light blue).

Images are maximal projections of Z‐stacks comprising the entire endothelial layer. The intensity of VE‐cadherin staining at adherens junctions and the width of the junctions were decreased in old compared to young arteries. Indeed, there are numerous areas in old endothelium with severe thinning or disruption of the junctions (^*). Fluorescence images are presented in their original unprocessed state. White scale bar is 10 μm.

Role of impaired junctional activity in ageing‐induced endothelial dilator dysfunction

Dilatation to acetylcholine was reduced in old compared to young tail arteries, which was characterized by a decrease in the maximal response and a rightward shift in the concentration–effect curve to the agonist (Fig. 3 and Table 1). Inhibition of NO synthase with l‐NAME (100 μm) suppressed responses to acetylcholine in young and old tail arteries (Fig. 3 and Table 1), and after l‐NAME, the residual dilatation to acetylcholine was greater in old arteries (Fig. 3 and Table 1). Dilatation to the NO donor NONOate was not significantly different between young and old tail arteries (Fig. 3 and Table 1).

Figure 3. Dilatation of young and old rat isolated tail arteries to the endothelial agonist acetylcholine (A) or the NO donor DEA NONOate (B).

Responses to acetylcholine (A) were assessed under control conditions (upper) and after inhibition of NO synthase with l‐NAME (lower). Dilator responses were analysed in pressurized arteries (P _TM of 60 mmHg) during stable constriction to phenylephrine (C). Data are expressed relative to baseline diameter (B) and presented as means ± SEM for A, n = 15 (control old), n = 15 (control young), or n = 8–9 (l‐NAME‐treated arteries); B, n = 6. Ageing decreased dilatation to acetylcholine under control conditions. After l‐NAME, the residual dilatation to acetylcholine (1 μm) was greater in old compared to young arteries. Ageing did not significantly affect dilatation to NONOate. See Table 1 for analyses.

Table 1.

Vasodilatation in young and old rat tail arteries to acetylcholine or DEA‐NONOate

Treatment	Maximal effect	Log EC50
Acetylcholine
Old, control (15)	91.6 ± 3.6 †††	−6.86 ± 0.09†††
Young, control (15)	108.1 ± 2.1	−7.33 ± 0.06
Old, l‐NAME (9)	56.1±10.1†	—
Young, l‐NAME (8)	29.3 ± 7.3	—
Old, control (6)	82.6 ± 7.6†	−6.63 ± 0.14††
Old, saracatinib (6)	100.7 ± 1.5*	−7.22 ± 0.11***
Young, control (6)	104.2 ± 2.9	−7.18 ± 0.08
Young, saracatinib (6)	102.4 ± 2.7	−7.11 ± 0.07
Old, control (6)	98.1 ± 1.0†††	−6.92 ± 0.10††
Old, VE‐cadherin FBA (6)	92.1 ± 5.1	−6.86 ± 0.10
Young, control (6)	111.1 ± 2.1	−7.28 ± 0.06
Young, VE‐cadherin FBA (6)	102.4 ± 3.1**	−6.92 ± 0.07***
Old, VE‐cadherin FBA (6)	90.1 ± 3.0	−6.73 ± 0.13
Old, VE‐cadherin FBA plus saracatanib (6)	82.1 ± 7.4	−6.63 ± 0.17
DEA‐NONOate
Old (6)	103.2 ± 2.9	−7.49 ± 0.15
Young (6)	106.8 ± 2.4	−7.54 ± 0.08

Data presented as means ± SEM, with number of replicates indicated in parentheses. ^†Statistically significant difference compared to the corresponding group of young arteries: ^† P < 0.05, ^†† P < 0.01, ^††† P < 0.001. ^*Statistically significant difference compared to the matching ‘control’ group: ^* P < 0.05, ^** P < 0.01, ^*** P < 0.001. For ‘Old VE‐cadherin FBA plus saracatinib’, the matching control group is ‘Old VE‐cadherin FBA’.

To evaluate the role of VE‐cadherin clustering at adherens junctions in the endothelium‐dependent dilatation to acetylcholine, arterial endothelial cells were exposed to a function‐blocking antibody (FBA) to VE‐cadherin (BV13, 50 μg ml⁻¹) or a control antibody (50 μg ml⁻¹) (Flavahan et al. 2013). The FBA inhibits VE‐cadherin clustering and disrupts endothelial adherens junctions (Corada et al. 2001; Flavahan et al. 2013). FBA‐mediated disruption of VE‐cadherin clustering inhibited the dilatation to acetylcholine in young arteries, causing a decrease in the maximal response and a rightward shift in the concentration–effect curve to the agonist, but had no significant effect on old arteries (Fig. 4 and Table 1). After inhibition of VE‐cadherin clustering, there was no longer any significant difference between the dilatation to acetylcholine in young and old arteries (Fig. 4 and Table 1).

Figure 4. Effects of a function blocking antibody to VE‐cadherin (VEC FBA) compared to a control antibody (Control) on dilatation to acetylcholine in young and old rat isolated tail arteries.

Dilatation was analysed in paired pressurized arteries (P _TM of 60 mmHg) during stable constriction to phenylephrine (C). One artery of each pair was treated with a VE‐cadherin function blocking antibody (BV13, 50 μg ml⁻¹) and the other with a control antibody (50 μg ml⁻¹). Data are expressed relative to baseline diameter (B) and presented as means ± SEM for n = 6. The VE‐cadherin FBA inhibited dilatation to acetylcholine in young but not old arteries. Indeed, after the FBA, there was no longer any significant difference in dilatation to acetylcholine between young and old arteries. See Table 1 for analyses.

Role of elevated Src activity in ageing‐induced impairment of endothelial junctions and endothelial dilatation

VE‐cadherin is continually internalized and degraded in cultured endothelial cells, and this process is markedly increased following exposure to inflammatory mediators resulting in impaired activity of adherens junctions (Orsenigo et al. 2012). Internalization of VE‐cadherin is initiated following tyrosine phosphorylation of the protein by Src kinases (Komarova & Malik, 2010; Orsenigo et al. 2012).

Src activity, measured as phosphorylation of Tyr⁴¹⁶ relative to total Src, was significantly increased by 2.2 ± 0.5‐fold in old compared to young arteries (n = 10, P < 0.05) (Fig. 5). This was associated with increased levels of tyrosine‐phosphorylated 115 kDa VE‐cadherin in old compared to young arteries (47.0 ± 9.1% and 5.7 ± 2.1% for old and young arteries, respectively, n = 3, P < 0.01, expressed relative to total 115 kDa in old lysates) (Fig. 5). In contrast, there was minimal to no tyrosine phosphorylation of the 140 kDa VE‐cadherin species detected in either old or young arteries (0.2 ± 0.2% and 2.1 ± 1.0% for old and young arteries, respectively, n = 3, NS, expressed relative to total 140 kDa in young lysates) (Fig. 5).

Figure 5. Tyrosine phosphorylation and ageing arteries.

A, Src kinase activity, assessed by immunoblot analysis of total compared to Tyr⁴¹⁶ phosphorylation, was increased in old compared to young (yng) tail arteries. Arteries were processed using the single lysis approach and a representative blot is presented. B, tyrosine phosphorylation of VE‐cadherin in old and young aortas was assessed by immunoprecipitation of tyrosine‐phosphorylated proteins followed by immunoblots for VE‐cadherin. This approach was performed on TTX‐soluble (TTX) and TTX‐insoluble (SDS) fractions, obtained using the sequential lysis approach. Representative blots for old and young arteries are presented. Tyrosine phosphorylation was detected in the 115 kDa but not 140 kDa VE‐cadherin species, and levels of tyrosine‐phosphorylated 115 kDa protein were increased in old compared to young aortas.

The selective Src inhibitor saracatinib (5 μm, 60 min) reduced Src activity in old rat tail arteries, measured as phosphorylation of Tyr⁴¹⁶ relative to total Src (to 9.6 ± 3.3% of control untreated arteries, n = 4, P < 0.001) (Fig. 6). Saracatinib (5 μm) concomitantly increased the level of the 140 kDa VE‐cadherin species by 61.6 ± 8.4% (n = 4, P < 0.001), increased the intensity of VE‐cadherin staining at adherens junctions (from 2586.3 ± 34.9 to 3255.0 ± 36.4 (P < 0.001), and increased junctional width (from 1.19 ± 0.02 μm to 1.56 ± 0.03 μm, P < 0.001) (n = 351 and 325 junctions in untreated and saracatinib‐treated arteries, respectively) (Figs 6 and 7).

Figure 6. Effect of the Src inhibitor saracatinib on levels of VE‐cadherin (upper panel) and on Src activity (lower panels) in old tail arteries.

Paired old tail arteries were incubated under control conditions (C) or in the presence of saracatinib (Sar, 5 μm, 90 min, at a P _TM of 60 mmHg), and then processed using the sequential lysis approach to obtain TTX‐soluble (TTX) and TTX‐insoluble (SDS) fractions. Representative blots are presented. Saracatinib reduced Src activity, as assessed by immunoblot analysis of Tyr⁴¹⁶‐phosphorylated compared to total Src. The Src inhibitor also increased levels of the 140 kDa VE‐cadherin species, which was associated with the TTX‐insoluble fraction.

Figure 7. Effect of the Src inhibitor saractinib on endothelial adherens junctions in old tail arteries, assessed by immunofluorescence staining of VE‐cadherin (green; DRAQ5 nuclear staining: light blue).

Paired old tail arteries were incubated under control conditions or in the presence of saracatinib (5 μm, 90 min, at a P _TM of 60 mmHg) before being processed for staining. Images are maximal projections of Z‐stacks comprising the entire endothelial layer. Saracatinib increased the intensity of VE‐cadherin staining at adherens junctions and the width of the junctions. Fluorescent images are presented in their original unprocessed state. White scale bar is 10 μm.

Saracatinib (5 μm) also increased dilatation to acetylcholine in old arteries, causing an increase in the maximal response and a leftward shift in the concentration–effect curve to the agonist (Fig. 8 and Table 1). However, saracatinib (5 μm) did not significantly affect the response to acetylcholine in young arteries (Fig. 8 and Table 1). Indeed, after saracatinib (5 μm), there was no longer any significant difference between the dilatation to acetylcholine in young and old arteries (Fig. 8 and Table 1). In old arteries treated with the function‐blocking antibody to VE‐cadherin, saracatinib (5 μm) no longer significantly affected the dilatation to acetylcholine (Fig. 8 and Table 1).

Figure 8. Effects of the Src inhibitor saracatinib on dilatation to acetylcholine in young (bottom) and old (top, middle) rat isolated tail arteries.

Dilatation was analysed in pressurized arteries (P _TM of 60 mmHg) during stable constriction to phenylephrine (C). Data are expressed relative to baseline diameter (B) and presented as means ± SEM for n = 6. Saracatinib (5 μm) increased dilatation to acetylcholine in old (top) but not young (bottom) arteries. After saracatinib, there was no longer any significant difference in dilatation to acetylcholine between young and old arteries. When old arteries were treated with VE‐cadherin FBA (BV13, 50 μg ml⁻¹), which did not affect dilatation in old arteries (see Fig. 4), the Src inhibitor no longer had any significant effect on dilatation to acetylcholine in old arteries (middle). See Table 1 for analyses.

Discussion

Endothelial dysfunction, including impaired endothelial dilatation, is considered a key contributor to the pathological process of vascular ageing (Ungvari et al. 2010; Hasegawa et al. 2012; Laurent, 2012). The present study demonstrates that ageing caused thinning and disruption of arterial endothelial adherens junctions, which was associated with increased phosphorylation of VE‐cadherin, loss of VE‐cadherin from adherens junctions and degradation of the protein in the endothelium of old arteries. Ageing was also associated with increased Src activity, and Src inhibition increased junctional levels of VE‐cadherin and restored the morphology of endothelial adherens junctions in old arteries. These molecular changes in adherens junctions were closely paralleled by changes in the contribution of endothelial adherens junctions to endothelium‐dependent dilatation, which was reduced in old compared to young arteries and was restored by Src inhibition. Impaired activity of arterial endothelial adherens junctions and interruption of VE‐cadherin‐dependent signalling is therefore an important mediator of endothelial dysfunction in ageing arteries, and likely to be a key mediator of vascular ageing.

Two prominent VE‐cadherin species were present in endothelium lining young arteries. The 140 kDa species was present on the cell surface and localized predominantly to the TTX‐insoluble fraction. The clustering of VE‐cadherin at adherens junctions in cultured cells is associated with attachment of the junctional complex to the actin cytoskeleton and movement of VE‐cadherin from the TTX‐soluble to the TTX‐insoluble fraction (Lee et al. 1999). The 140 kDa species is therefore the physiologically important VE‐cadherin species that is localized within adherens junctions of arterial endothelium. In contrast, the 115 kDa VE‐cadherin protein was not present on the cell surface and was localized predominantly to the TTX‐soluble fraction, which is consistent with this species being an internalized and degraded fragment. Although VE‐cadherin can be degraded by extracellular proteases that cleave the protein's extracellular domains (Hermant et al. 2003; Schulz et al. 2008; Sidibe et al. 2012), our immunoblot analyses employed an antibody that targets proximal amino acids of the first extracellular domain (EC1) of VE‐cadherin (Corada et al. 2001; May et al. 2005) and are proximal to cleavage sites of extracellular proteases (Hermant et al. 2003; Schulz et al. 2008; Sidibe et al. 2012). Therefore, the intracellular localization of a 115 kDa species containing this proximal N‐terminal sequence suggests that it was generated by intracellular and not extracellular cleavage of VE‐cadherin. In cultured endothelial cells, VE‐cadherin is internalized and degraded predominantly through a clathrin‐dependent endosomal–lysosomal pathway (Xiao et al. 2003b; Chiasson et al. 2009; Sawant et al. 2011; Orsenigo et al. 2012). Indeed, in cultured cells, the initial degradation of internalized VE‐cadherin generates an N‐terminal VE‐cadherin fragment that is ∼25 kDa smaller than the native protein (Su & Kowalczyk, 2016). This process and the subsequent degradation of VE‐cadherin into smaller fragments can occur rapidly (Xiao et al. 2003a; Orsenigo et al. 2012; Su & Kowalczyk, 2016). Therefore, the increased prominence of the 115 kDa VE‐cadherin species in aorta compared to tail arteries may reflect temporal differences in the degradative pathway of these endothelial cells. Because these VE‐cadherin species were observed in endothelial cells lining young arteries, it indicates that, as in cultured cells, the internalization and degradation of VE‐cadherin is a normal constitutive feature of native arterial endothelium (Xiao et al. 2003a; Orsenigo et al. 2012).

Ageing decreased the levels of the 140 kDa VE‐cadherin species in aortas and tail arteries, and concomitantly increased levels of the 115 kDa species. These results suggest that ageing increases the internalization and degradation of VE‐cadherin, resulting in loss of the physiologically important 140 kDa species from the cells and from adherens junctions. This was confirmed by immunofluorescence imaging of VE‐cadherin, demonstrating that ageing reduced the intensity of VE‐cadherin staining at adherens junctions and reduced the width of endothelial adherens junctions. A decreased junctional width is indicative of reduced clustering of VE‐cadherin and reduced stability and strength of the junctions (Sun et al. 2009; Liu et al. 2010; Flavahan et al. 2013; Breslin et al. 2015). Our results are consistent with previous studies demonstrating reduced endothelial junctional complexity and increased presence of thinning and open endothelial junctions (or gaps) in ageing aortas (Kao et al. 1994; Lee et al. 2001). The results are in apparent contrast to those of Hunyh et al. (2011) who proposed that the width of endothelial adherens junctions was increased in ageing mouse aortas. However, those authors purposefully assessed the widest part of endothelial junctions with the apparent goal of measuring open junctions or gaps in the endothelial monolayer (Huynh et al. 2011). Reduced proteostasis is considered one of the ‘hallmarks of ageing’, reflecting decreased activity of lysosomal and proteasomal pathways in ageing cells (Lopez‐Otin et al. 2013; De Meyer et al. 2015; Nussenzweig et al. 2015). Therefore, the increased internalization and degradation of VE‐cadherin in ageing endothelial cells is unlikely to reflect a generalized increase in proteolytic activity. Indeed, expression of claudin‐5, a transmembrane component of endothelial tight junctions (Harris & Nelson, 2010), was actually increased in ageing arteries, suggesting that VE‐cadherin may be specifically targeted by the ageing process.

The internalization and degradation of VE‐cadherin can be increased by inflammatory mediators and is preceded by a Src‐dependent increase in tyrosine phosphorylation of VE‐cadherin (Komarova & Malik, 2010; Orsenigo et al. 2012). Src phosphorylates VE‐cadherin at multiple sites including Tyr⁶⁵⁸ and Tyr⁷³¹, which inhibits the binding of p120 catenin and β‐catenin, respectively, resulting in destabilization of the junctional complex and internalization of VE‐cadherin (Komarova & Malik, 2010; Orsenigo et al. 2012). Indeed, the binding of p120 catenin to VE‐cadherin masks an endocytic motif in VE‐cadherin, preventing internalization and degradation of the protein (Xiao et al. 2003a; Nanes et al. 2012). Src activity was increased in old compared to young arteries, and this was associated with increased levels of tyrosine‐phosphorylated VE‐cadherin (115 kDa) in old arteries. Because of the low levels of the 115 kDa species in tail arteries, the analysis of tyrosine phosphorylation of this species was restricted to aortas. We were unable to detect significant tyrosine phosphorylation of the 140 kDa VE‐cadherin species in aortas or tail arteries, which is likely to reflect rapid internalization and processing of phosphorylated VE‐cadherin (Xiao et al. 2003a; Orsenigo et al. 2012; Su & Kowalczyk, 2016). The Src inhibitor saracatinib reduced the ageing‐induced increase in Src activity and increased the level of the physiologically important 140 kDa VE‐cadherin species in old arteries. Moreover, the increased level of VE‐cadherin following Src inhibition was localized to the TTX‐insoluble fraction suggesting that it increased clustering of VE‐cadherin and strengthened adherens junctions in ageing endothelium. This was confirmed by immunofluorescence imaging of VE‐cadherin, which demonstrated increased intensity and increased width of endothelial adherens junctions following Src inhibition. These results suggest that internalization and degradation of VE‐cadherin, and the impairment of endothelial adherens junctions in old arteries are mediated, at least in part, by heightened Src activity, and that Src inhibition can block these pathological changes and restore endothelial junctional activity.

The clustering of VE‐cadherin at adherens junctions stimulates assembly of a macromolecular signalling complex that regulates endothelial cell signalling, morphology and function, including amplifying phosphoinositide 3‐kinase (PI3K)/Akt signalling (Lampugnani et al. 2002; Lampugnani & Dejana, 2007; Taddei et al. 2008; Komarova & Malik, 2010). Indeed, VE‐cadherin‐dependent signalling at adherens junctions amplifies endothelium‐dependent dilatation to acetylcholine, which was mediated by PI3K/Akt‐dependent activation of endothelial NO synthase (Flavahan et al. 2013; Flavahan & Flavahan, 2014). Ageing is well known to impair endothelium‐dependent, NO‐mediated dilatation (Csiszar et al. 2002; Kim et al. 2009; Trott et al. 2011), which was confirmed in the present study in rat tail arteries. Endothelium‐dependent, l‐NAME‐sensitive dilatation to acetylcholine was reduced in old compared to young arteries, whereas dilatation to an exogenous NO donor was not affected. These results are consistent with ageing‐induced impairment in endothelial NO activity. In tail arteries, an FBA to VE‐cadherin, which inhibits VE‐cadherin clustering and disrupts endothelial adherens junctions (Corada et al. 2001; Flavahan et al. 2013), decreased dilatation to acetylcholine in young arteries, but had no effect in ageing arteries. This confirms that VE‐cadherin signalling at adherens junctions can contribute to endothelial dilatation, but only in young arteries. Indeed, the lack of effect of the FBA in old arteries suggests that the functional role of VE‐cadherin and adherens junctions in endothelial dilatation is impaired by ageing, consistent with impaired endothelial junctions and increased degradation of VE‐cadherin in old arteries. Importantly, the decrease in acetylcholine‐induced dilatation following disruption of adherens junctions by the FBA in young arteries was of the same magnitude as the decreased dilatation associated with ageing‐induced impairment of the junctions in old arteries. Indeed, after disrupting endothelial adherens junctions with the FBA to VE‐cadherin, there was no longer any significant difference in dilatation to acetylcholine between young and old arteries. These data suggest that an ageing‐induced impairment of endothelial adherens junctions can account for the decreased endothelial dilatation in old arteries.

Concomitant with restoring the morphology of ageing endothelial adherens junctions and increasing VE‐cadherin levels in the TTX‐insoluble junctional fraction of old arteries, Src inhibition with saracatinib increased endothelium‐dependent dilatation to acetylcholine in old arteries. Saracatinib had no effect on acetylcholine‐induced dilatation in young arteries, and following Src inhibition there was no longer any difference in acetylcholine‐induced dilatation between young and old arteries. Therefore, Src inhibition completely reversed the ageing‐induced impairment in endothelium‐dependent dilatation. Importantly, treatment of old arteries with an FBA to VE‐cadherin, which itself had no effect in old arteries, prevented the ability of saracatinib to increase endothelial dilatation to acetylcholine in old arteries. Thus, the protective effect of Src inhibition on endothelial dilatation in old arteries was dependent on the clustering of VE‐cadherin at endothelial adherens junctions. The biochemical, morphological and functional data are therefore in agreement and indicate that the endothelial dilator dysfunction in old arteries is mediated by increased Src activity resulting in phosphorylation, internalization and degradation of VE‐cadherin, impairment of endothelial adherens junctions, and loss of the amplifying effect of VE‐cadherin junctional signalling on endothelial dilatation.

Endothelial permeability is increased in aortas and in the microcirculation of ageing animals (Ezekowitz et al. 1980; Belmin et al. 1993; Chen et al. 2003), which would be consistent with impaired function of endothelial adherens junctions (Hordijk et al. 1999; Dejana et al. 2008; Schulte et al. 2011). Interestingly, despite the prominent role of the endosomal–lysosomal pathway in degrading internalized VE‐cadherin, a proteasomal inhibitor prevented the increase in endothelial permeability and disruption of adherens junctions caused by a VE‐cadherin decoy protein, whereas a lysosomal inhibitor was without effect (Sawant et al. 2011). It is unclear if this reflects a role of the proteasome in VE‐cadherin degradation or an indirect effect to reduce VE‐cadherin ubiquitination, a key step in clathrin‐mediated internalization of the protein (Xiao et al. 2003b; Sawant et al. 2011). Although inhibiting VE‐cadherin internalization or degradation pathways may be of benefit in ageing endothelial cells, the approach was not investigated in the present study. The phosphorylation and internalization of VE‐cadherin reduces the strength and stability of adherens junctions, but additional signalling may be required to cause complete dissolution of adherens junctions, leading to the opening of intercellular gaps and increased permeability (Dejana et al. 2008; Orsenigo et al. 2012; Adam et al. 2016). For example, inflammatory mediators may contribute to junctional disruption by increasing endothelial actin–myosin contractility following activation of endothelial myosin light chain kinase by Src or following inhibition of myosin light chain phosphatase by Rho/Rho‐associated protein kinase (ROCK) (Garcia et al. 1999; Komarova & Malik, 2010). Indeed, increased deposition of stiff matrix components such as fibronectin in ageing arterial intima could contribute to junctional disruption by increasing endothelial Rho/ROCK activity (Fujiwara et al. 2004; Huynh et al. 2011; Krishnan et al. 2011). However, junctional dissolution and increased endothelial permeability, e.g. in response to tumour necrosis factor‐α (TNFα), can occur independently of Rho/ROCK signalling or of actin–myosin contractility (Marcos‐Ramiro et al. 2014; Adam et al. 2016). The status of endothelial contractility, stress fibres or Rho/ROCK signalling in the endothelium of ageing arteries is currently unknown. Although tight junctions are present in peripheral endothelial cells, their structure and functional role in regulating permeability is greatly diminished compared to cerebral endothelium and the blood–brain barrier (BBB) (Kniesel & Wolburg, 2000; Harris & Nelson, 2010; Lampugnani, 2012; Bauer et al. 2014). Because of this differential activity and regulation, the maintained expression of claudin‐5 in ageing peripheral endothelial cells, as observed in the present study, does not provide insight into the influence of ageing on cerebral endothelial cells and the BBB. Indeed, as with peripheral endothelial cells, BBB permeability is increased in ageing and may contribute to ageing‐induced declines in cerebral function (Farrall & Wardlaw, 2009).

Impaired activity of endothelial adherens junctions likely contributes to multiple aspects of endothelial dysfunction and arterial deterioration in the ageing vascular system. Ageing arterial endothelial cells are severely compromised: in addition to impaired endothelial dilatation and increased permeability, they are highly susceptible to apoptosis and have a prominent pro‐inflammatory phenotype, including production of pathological mediators such as angiotensin II, TNFα and endothelin‐1 (Belmin et al. 1995; Asai et al. 2000; Csiszar et al. 2004; Lakatta et al. 2009; Goel et al. 2010; Ungvari et al. 2010). These processes are regulated by VE‐cadherin‐dependent signalling within adherens junctions. Indeed, VE‐cadherin activity at adherens junctions is essential for maintaining the protective structure and function of the endothelium, such as promoting signalling that amplifies NO activity, reduces apoptosis, and inhibits inflammatory activity including reducing endothelial permeability and leukocyte extravasation (Dejana et al. 2009; Flavahan et al. 2013; Giannotta et al. 2013; Flavahan, 2017). Internalization and degradation of VE‐cadherin would not only reduce the protective activity of endothelial adherens junctions, but would also release β‐catenin (Valenta et al. 2012). Although free β‐catenin is targeted by a destruction complex, numerous mechanisms enable the protein to escape degradation and translocate to the nucleus (Lampugnani & Dejana, 2007; Beckers et al. 2008; Valenta et al. 2012). β‐Catenin transcriptional activity can increase expression of mediators that are known to play an important local role in vascular ageing, including the renin–angiotensin system, endothelin‐1, TNFα, interleukin 6, inducible nitric oxide synthase, fibronectin, cyclooxygenase 2, matrix metalloproteinase (MMP) 2 and MMP9 (Levy et al. 2002; Kim et al. 2005; Sun et al. 2006; Gelfand et al. 2011; Vikram et al. 2014; Zhou et al. 2015; Zhou & Liu, 2016). Indeed, arterial stiffening, a prominent feature of ageing arteries, reflects the activity of a local angiotensin II signalling system within the endothelium and intimal smooth muscle cells, which promotes inflammatory and fibrotic activity resulting in intimal deposition of poorly distensible proteins, such as fibronectin, and MMP‐mediated degradation of highly distensible elastin fibres (Wang et al. 2005, 2006, 2007, 2010a,b, 2014; Lakatta et al. 2009). Ageing‐induced arterial stiffness, which is also amplified by the ageing‐induced decrease in endothelial NO activity (Santhanam et al. 2010; Jung et al. 2013), disrupts organ blood flow and is a significant predictor of cardiovascular complications and death. The ageing arterial wall is also more susceptible to disease progression, displaying a lower threshold, increased severity and poorer prognosis for vascular disease, including atherosclerosis (Ungvari et al. 2010; Wang et al. 2010b). Impaired activity of endothelial adherens junctions also plays a key role in the initiation and progression of the atherosclerotic process, contributing to a breakdown in the endothelial barrier and enabling the subendothelial accumulation of atherogenic lipids, increased inflammatory activity and increased monocyte transmigration, and contributing to formation and expansion of the atherosclerotic plaque (Flavahan, 2017). Therefore, an ageing‐induced impairment of endothelial adherens junctions would be expected to amplify these pathological processes and contribute importantly to ageing‐induced deterioration of cardiovascular and organ function.

A limitation of the present study is that the mechanisms responsible for the ageing‐induced increase in Src activity were not identified. Ageing endothelial cells are caught in a chronic cycle of inflammatory stress, both producing and responding to pathological mediators (Flavahan et al. 2016). Indeed, the TNFα, angiotensin II and reactive oxygen species (ROS) locally produced by endothelial and intimal smooth muscle cells are considered key mediators of the endothelial pathology of ageing, including endothelial dilator dysfunction (Arenas et al. 2005; Csiszar et al. 2007; Wang et al. 2010a; Flavahan et al. 2016). As demonstrated in the present study, impaired activity of adherens junctions and loss of VE‐cadherin‐dependent signalling contributes to endothelial dilator dysfunction in old arteries. Therefore, these same mediators are likely to be involved in the ageing‐induced impairment of endothelial adherens junctions. Indeed, TNFα, angiotensin II and ROS are each capable of activating Src, increasing tyrosine phosphorylation of VE‐cadherin and disrupting adherens junctions in endothelial cells (Nwariaku et al. 2004; Angelini et al. 2006; Sano et al. 2006; Monaghan‐Benson & Burridge, 2009; Birukova et al. 2013; Dikalov et al. 2014).

The present study demonstrates using biochemical, morphological and functional analyses that endothelial adherens junctions are impaired in ageing arteries, reflecting phosphorylation, internalization and degradation of VE‐cadherin. This pathological process appears to be mediated by increased Src kinase activity, with Src inhibition increasing junctional expression of VE‐cadherin and restoring the morphology of adherens junctions. The impairment in endothelial NO‐mediated dilatation occurring in old arteries can be accounted for by the dysfunction of adherens junctions, reflecting a selective loss in the contribution of adherens junctions to endothelial dilatation that can be restored completely by Src inhibition. Impaired activity of endothelial adherens junctions is likely to contribute to multiple aspects of vascular ageing, including promoting the progression of vascular disease and contributing to organ dysfunction. Interventions designed to augment endothelial adherens junctions should be of therapeutic value in countering vascular ageing.

Additional information

Competing interests

None

Author contributions

The experiments were all performed in the laboratory of N.A.F. at Johns Hopkins University School of Medicine. The authors contributed in the following manner: conception or design of the work (N.A.F., S.F., F.C.); acquisition, analysis or interpretation of data for the work (N.A.F., S.F., F.C.); drafting the work or revising it critically for important intellectual content (N.A.F., S.F., F.C.). All authors have approved the final version of the manuscript and agree to be accountable for all aspects of the work. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

This work was supported by an NIH award to N.A.F. (HD078639).