The short-chain fatty acid butyrate exerts a specific effect on VE-cadherin phosphorylation and alters the integrity of aortic endothelial cells

Abstract

Short-chain fatty acids (SCFAs) like butyrate (BUT) largely influence vascular integrity and are closely associated with the onset and progression of cardiovascular diseases. However, their impact on vascular endothelial cadherin (VEC), a major vascular adhesion and signaling molecule, is largely unknown. Here, we explored the effect of the SCFA BUT on the phosphorylation of specific tyrosine residues of VEC (Y731, Y685, and Y658), which are reported to be critical for VEC regulation and vascular integrity. Moreover, we shed light on the signaling pathway engaged by BUT to affect the phosphorylation of VEC. Thereby, we used phospho-specific antibodies to evaluate the phosphorylation of VEC in response to the SCFA sodium butyrate in human aortic endothelial cells (HAOECs) and performed dextran assays to analyze the permeability of the EC monolayer. The role of c-Src and SCFA receptors FFAR2 and FFAR3 in the induction of VEC phosphorylation was analyzed using inhibitors and antagonists for c-Src family kinases and FFAR2/3, respectively, as well as by RNAi-mediated knockdown. Localization of VEC in response to BUT was assessed by fluorescence microscopy. BUT treatment of HAOEC resulted in the specific phosphorylation of Y731 at VEC with minor effects on Y685 and Y658. Thereby, BUT engages FFAR3, FFAR2, and c-Src kinase to induce phosphorylation of VEC. VEC phosphorylation correlated with enhanced endothelial permeability and c-Src-dependent remodeling of junctional VEC. Our data suggest that BUT, an SCFA and gut microbiota-derived metabolite, impacts vascular integrity by targeting VEC phosphorylation with potential impact on the pathophysiology and therapy of vascular diseases.

Keywords: VE-cadherin; aortic endothelial cell; butyrate; c-src; free fatty acid receptor (FFAR); permeability; phosphorylation; short-chain fatty acid (SCFA).

FIGURE 1

BUT increases permeability in a VEC-dependent manner in HAOECs. (A) Permeability of the HAOEC monolayer was checked by stimulating HAOECs with different concentrations of BUT for 1 h. Fb/Fa: ratio of fluorescence intensity in the basal chamber (Fb) and fluorescence intensity apical chamber (Fa). (B) KD of VEC expression in HAOEC with a decrease of 65.99% was confirmed by western blotting. (C) Permeability of the HAOEC monolayer as described in (A) in the control (CTRL) and VEC KD cells. Cells were treated with 1 mM BUT for 1 h. (A, C) Data represent mean values of n = 6 experiments each measured in biological duplicate (C) or triplicate (A). Statistical significance was considered for p ≤ 0.05 (*), p ≤ 0.005 (**), p ≤ 0.001 (***), and p ≤ 0.0001 (****). (D) Immunofluorescence microscopy images of non-permeabilized HAOEC stained for junctional VEC (antibody clone E6N7A) after 1 mM BUT or medium-only (CTRL) treatment for 1 h. Images 1–4 represent zoomed-in areas indicated by white rectangles with arrowheads pointing to specific remodeled junctional regions. Images show representative data of n = 3 independent experiments. Scale bar (100 μm). (E) Quantification of the junctional VEC pattern for VEC coverage, interface linearity, and VEC interface occupancy using the program Junction Mapper (see Materials and Methods). Numbers of analyzed junctions of n = 3 experiments are given underneath the bars.

FIGURE 2

BUT induces specific tyrosine phosphorylation of VEC. (A) Phosphorylation of VEC at tyrosine-731, -685, and -658 (phospho-VECY731, Y685, and Y658) in response to various concentrations of BUT analyzed by western blotting. HAOECs were treated for 1 h with indicated concentrations of BUT. Phospho-VEC signals where normalized to the total VEC. Relative phospho-VEC signals were normalized to the level in control cells (dashed red line). β-Actin served as a loading control. (B) Kinetic of BUT-induced phospho-VECY731 level analyzed by western blotting. Data display densiometric quantification of relative phospho-VEC and VEC signals normalized to the level in control cells (dashed red line) of n ≥ 4 experiments. For better illustration, the identical dataset for 1 h treatment at 1 mM from panel (A) (phospho-VECY731) is shown. (C) Phospho-VECY731 level after 1 h treatment of HAOEC with BUT (1 mM) in the presence or absence of Ang-II (100 ng/mL). Data display densitometric quantification of relative phospho-signals as described in (B).

FIGURE 3

Perturbation of FFAR2/3 expression and function diminishes BUT-induced VEC phosphorylation. (A) Detection of FFAR2/3 expression by RT-PCR and immunohistochemistry (IHC). IHC staining was performed on fixed and agarose-embedded HAOEC (left panel) and of FFPE samples of human thoracic aorta (right panel). Images below represent zoomed-in areas indicated by black rectangles with red arrowheads pointing to specific receptor staining. Scale bar (100 μm). (B) BUT-induced VEC phosphorylation analyzed as shown in Figure 2 upon pre-incubation overnight with 0.1 µM of FFAR3 antagonist GLPG0974 (left panel) and co-incubation with FFAR2 antagonist ß-HB (0.1mM, right panel). For the GLPG0974 approach, cells were treated either with a vehicle alone (DMSO) or in combination with BUT and GLPG0974. Data represent n ≥ 5 experiments. (C) Validation of FFAR3 KD by RT-qPCR. Relative expression of FFAR3 normalized to GAPDH between CTRL (=1) and FFAR3 siRNA-transfected HAOEC of n = 3 independent transfection experiments. (D) BUT-induced VEC phosphorylation analyzed as shown in Figure 2 in the presence and absence of FFAR3 KD of n = 6 experiments. In FFAR3 KD cells, Y731 phosphorylation in response to butyrate treatment was decreased by 40.91% compared to scrambled siRNA. (E) Immunofluorescence microscopy images of non-permeabilized HAOEC stained for junctional VEC (antibody clone BV6) after 1 mM BUT or medium-only (-BUT) treatment for 1 h. Cells were transfected with control or FFAR3 siRNA under the same conditions used in (C, D). Small images represent enlarged areas of the sections marked by white rectangles. Images show representative data of n = 3 independent experiments. Scale bar (100 μm).

FIGURE 4

BUT engages Src kinases for specific tyrosine phosphorylation at VEC. (A) HAOECs were treated for 15 min before and during BUT incubation with PP2 (10 μM) or vehicle (DMSO). Subsequently, phospho-VECY731 levels were analyzed by western blotting. Phospho-signals of n = 4 experiments were quantified by densitometry. (B) Immune fluorescent microscopy of HAOECs treated with the vehicle (DMSO) or PP2 and BUT (5 mM) for 1 h and subsequently stained for VEC and general phospho-tyrosine residues. Scale bar (10 μm). (C) Validation of Src KD by RNAi. HAOECs were transfected with Src-specific siRNA or control-siRNA. Cells were lysed and analyzed for Src expression 12 h later by western blotting. (D) Stimulation of Src KD and control cells with BUT (1 mM) for 1 h and analysis of phosho-VECY731 expression as shown in Figure 2. Quantification corresponds to n = 6 experiments. (E) Phospho-SrcY416 western blotting of BUT-treated or -untreated cells, transfected with Ctrl or FFAR3 siRNA. Data display densiometric quantification of relative phospho-Src signals normalized to the level in BUT-untreated Ctrl siRNA-transfected cells (dashed red line) of n = 3 experiments.