Rapid shear stress-dependent ENaC membrane insertion is mediated by the endothelial glycocalyx and the mineralocorticoid receptor

Abstract

The contribution of the shear stress-sensitive epithelial Na⁺ channel (ENaC) to the mechanical properties of the endothelial cell surface under (patho)physiological conditions is unclear. This issue was addressed in in vivo and in vitro models for endothelial dysfunction. Cultured human umbilical vein endothelial cells (HUVEC) were exposed to laminar (LSS) or non-laminar shear stress (NLSS). ENaC membrane insertion was quantified using Quantum-dot-based immunofluorescence staining and the mechanical properties of the cell surface were probed with the Atomic Force Microscope (AFM) in vitro and ex vivo in isolated aortae of C57BL/6 and ApoE/LDLR^-/- mice. Flow- and acetylcholine-mediated vasodilation was measured in vivo using magnetic resonance imaging. Acute LSS led to a rapid mineralocorticoid receptor (MR)-dependent membrane insertion of ENaC and subsequent stiffening of the endothelial cortex caused by actin polymerization. Of note, NLSS stress further augmented the cortical stiffness of the cells. These effects strongly depend on the presence of the endothelial glycocalyx (eGC) and could be prevented by functional inhibition of ENaC and MR in vitro endothelial cells and ex vivo endothelial cells derived from C57BL/6, but not ApoE/LDLR^-/- vessel. In vivo In C57BL/6 vessels, ENaC- and MR inhibition blunted flow- and acetylcholine-mediated vasodilation, while in the dysfunctional ApoE/LDLR^-/- vessels, this effect was absent. In conclusion, under physiological conditions, endothelial ENaC, together with the glycocalyx, was identified as an important shear stress sensor and mediator of endothelium-dependent vasodilation. In contrast, in pathophysiological conditions, ENaC-mediated mechanotransduction and endothelium-dependent vasodilation were lost, contributing to sustained endothelial stiffening and dysfunction.

Keywords: ENaC; Endothelial dysfunction; Glycocalyx; Mineralocorticoid receptor; Shear stress.

Fig. 1

Shear stress-induced ENaC membrane abundance within minutes. A ENaC membrane abundance was quantified under chronic laminar shear stress (48 h, 8 dyne/cm²) and showed a significant increase compared to non-flow controls by 48.6 ± 3.6%. B Application of acute (15 min) laminar shear stress (8 dyne/cm²) results in an increase of ENaC expression by 58.5 ± 4.5%. 1 nM aldosterone was present in both conditions (N = 3; n = 195−198; *** = significant difference (p ≤ 0.001)). Scale bar: 25 μm

Fig. 2

Analysis of αENaC membrane expression. A Under acute (15 min) LSS (8 dyne/cm²), application of Brefeldin A (5 μg/ml) reduced the membrane abundance of ENaC by 32.3 ± 5.2% (N = 3, n = 142). These observations indicate a non-genomic membrane insertion of ENaC. B Application of CA (100 nM), the active metabolite of the MR antagonist spironolactone and/or acute (15 min) application of LSS (8 dyne/cm²) prevents the LSS effect under flow conditions by 20.7 ± 4.0% compared to static control conditions (N = 3, n = 157). C Application of Benzamil (1 µM) prevents the LSS effect under flow conditions by 20.7 ± 4.0% compared to static control conditions (N = 3, n = 117). D Enzymatic removal of the glycocalyx with heparinase I (1 SU/ml) reduced the ENaC membrane abundance only under LSS conditions 13 ± 3.0% (N = 3, n = 95) and was still augmented compared to static conditions by 21.1 ± 3.2% (N = 3, n = 195), whereas under static conditions no effect could be observed (* = significant difference (p ≤ 0.05), *** = significant difference (p ≤ 0.001))

Fig. 3

Impact of non-laminar shear stress. A ENaC membrane abundance was quantified and showed a significant increase in LSS (8 dyne/cm²) conditions compared to non-flow controls by 21.4 ± 4.7%. NLSS (8 dyne/cm²) increased ENaC membrane abundance as well (+ 25.0 ± 4.3%), but it did not matter which form of shear stress was applied (N = 3, n = 120). B LSS lead to stiffening of the cell cortex by 18.9 ± 5.5% compared to static controls. NLSS further increased the cortical stiffness by 67.9 ± 8.9% compared to non-flow conditions (N = 3, n = 37) (* = significant difference (p ≤ 0.05, *** = significant difference (p ≤ 0.001)). C Cortical F-actin was quantified by confocal microscopy and showed a significant increase by 19.2 ± 4.0% in LSS conditions compared to non-flow controls. NLSS further increased F-actin by 65.5 ± 5.4% compared to static controls (N = 3, n = 39). D Scheme and representative images of the analysis of cortical actin. Static, LSS and NLSS conditions are shown

Fig. 4

Changed ENaC regulation and endothelial mechanics in ApoE/LDLR^-/- ex vivo endothelial cells. A In ApoE/LDLR^-/- ex vivo endothelial cells, the ENaC membrane expression is reduced compared to WT by − 23%. In WT, amiloride and spironolactone significantly decreased the ENaC membrane expression by − 43% and − 41% (N = 3, n = 28), respectively. In the ApoE/LDLR^-/- mouse model, this effect is abolished, indicated an alternative ENaC regulation under dysfunctional conditions. Compared to WT, in ApoE/LDLR^-/- ex vivo endothelial cells the ENaC membrane expression is not reduced after application of amiloride and spironolactone (N = 3, n = 32–44, **significant difference = p ≤ 0.01, ***p ≤ 0.001). B The cortical stiffness of endothelial cells ex vivo derived from ApoE/LDLR^-/- aortae is increased compared to WT by 6% (N = 3; WT: 0.98 ± 0.021 pN/nm; ApoE/LDLR^-/-: 1.04 ± 0,025 pN/nm). In WT, amiloride and spironolactone softens the endothelial cortex by − 17% (N = 3, n = 28) and − 20% (N = 3, n = 56, ** significant difference = p ≤ 0.01, ***p ≤ 0.001), respectively. This effect is absent in the ApoE/LDLR^-/- model. C Phalloidin stainings revealed the amount of F-actin in ApoE/LDLR^-/- endothelial cells is significantly increased to WT by 14% (N = 4, n = 109–125, ** significant difference = p ≤ 0.01). D In in vitro endothelial cells, actin is depolymerized by CyD (− 28%), amiloride (− 41%), benzamil (− 49%) and spironolactone (− 18%). (N = 3; n = 60, * significant difference = p ≤ 0.05)

Fig. 5

Effects of benzamil and canrenone on endothelial function in vivo in C57BL/6 and ApoE/LDLR^-/- mice. A Flow-induced (FMD) vasodilation in the femoral artery (FA), after 5 min of vessel occlusion, and acetylcholine(Ach)-induced vasodilation in the abdominal aorta (B, AA) 30 min after Ach administration are shown. Results from 4–5-month-old ApoE/LDLR^-/- (n = 5) and C57BL/6 (n = 5) mice pre-treated with benzamil (n = 5, both groups) or canrenone (n = 5, both groups) in comparison to age-matched, untreated ApoE/LDLR^-/- (n = 5) and C57BL/6 mice (n = 5), respectively. FMD response was measured 30 min after benzamil (1 mg/kg, i.v) or canrenone (10 mg/kg, i.v) administration. Ach-dependent response was measured 1 h after benzamil or canrenone administration. Statistics: two-way ANOVA followed by Tukey’s post hoc test (normality was assessed using the Shapiro–Wilk test): *** p ≤ 0.001

Fig. 6

Maladaptive loss of ENaC-dependent regulation of endothelial response to flow contributes to endothelial dysfunction Left side: in a healthy condition, the endothelial cells can flexible adapt to changes in their environment, e.g., different rates of shear stress. This includes an appropriate ENaC-dependent mechanotransduction and the ability of the vessel to react to flow with dilation. Right side: in case of a damaged and atheroprone endothelium, the endothelial cells lost their plasticity. This includes disturbed regulation and mechanotransduction via ENaC, chronic stiffening of the cortex and as a result a loss of normal functional properties of the vessel. It could be debated if such conditions further impair ENaC function, supposing a vicious circle of vascular dysfunction