Fig 2: Murine endoglin in MAECs and MLEC is involved in F-actin polymerization and distribution. (a,b) Mouse aortic endothelial cells (MAEC) from a model of HHT1, stained for F-actin (green, Alexa 488) and DAPI (a), and its quantification (b). (c,d) Mouse lung endothelial cells (MLEC) from a model of HHT1 stained for F-actin (green, Alexa 488) and DAPI (c), and its quantification (d). A significant difference (** p < 0.01) between control and Eng+/− conditions was found in terms of actin distribution (b,d) by Image J considering fluorescence (IntDens). (e,f) Eng+/− MLEC stained for actin and tubulin (e) and their co-localization evaluated by Person’s coefficient (* p < 0.05) (f). (scale bar, 33 µm).

Fig 3: Endoglin silencing affects vessel sprout diameter and permeability. (a) Three-dimensional (3D) reconstruction of ECFC ctr-siRNA and ECFC Eng-siRNA to evaluate sprout diameter. In green, Actin F (Alexa 488), and in red, nuclei (TOPRO-3) are stained (scale bar, 150 µm). (b) When Eng is suppressed, sprouts are wider than in controls (*** p < 0.001). (c) Using a real-time impedance-based cell analyzer (iCELLigence system, ACEA Biosciences, San Diego, USA), ECFC ctr-siRNA and ECFC Eng-siRNA are analyzed in basal condition and under 20 ng/mL TNFα stimulation (gray and black lines, respectively). As shown by the quantification in (d), a significant difference between ECFC ctr-siRNA (white column) and ECFC Eng-siRNA (black column) is found 12 h after TNFα stimulation (* p < 0.05). (e) Immunofluorescence staining for VE-CAD in ECFC-ctr and ECFC Eng-siRNA with the absence of or in the presence of 20 ng/mL TNFα for 1 h (scale bar, 22 µm). (f) Quantification of the pictures in (c) does not show differences between control and Eng-siRNA in terms of VE-CAD membrane staining nor of fluorescence intensity in the presence of TNFα (* p < 0.05).

Fig 4: Effect of CXCL12 and TRAP6 on platelet adhesion to endothelial cells. a Calcein-labeled platelets were incubated for 10 min on HAEC monolayers in the absence or presence of CXCL12 or thrombin receptor activating peptide 6 (TRAP6) and washed twice with PBS. Adhesion of platelets to EC was visualized by fluorescence microscopy (×20 magnification) (top panels). Scale bars, 40 µm. Quantification of platelet adhesion repeated in quadruplicates with two different platelet preparations and analyzed by ImageJ software is shown in the histogram. Mean values ± SEM are displayed. **p < 0.01; *p < 0.05. b Calcein-labeled platelets were incubated for 10 min on HAEC monolayers in the absence (control) or presence of CXCL12, either alone or with the RGD peptide, the CXCR4 inhibitor AMD3100, or the recombinant extracellular part of endoglin (Sol.Eng), as indicated. Platelets were visualized by fluorescence microscopy and quantified (quadruplicates with two different platelet preparations) with ImageJ software. Mean values ± SEM are displayed. **p < 0.005; *p < 0.01

Fig 5: Endoglin silencing affects cofilin dynamics. (a,b) Western blot performed on n = 5 different clones of ECFC at passage 3–4, <30 days. A representative experiment is shown. Eng silencing in ECFC reduces P-Cofilin, suggesting an increased activity of cofilin when Eng is downregulated. Actin and tubulin were used as loading controls. The last row displays the reduced expression of Eng after siRNA silencing. (b) Quantification of five different ECFC clones shows a significant difference (*** p < 0.001) between ctr-siRNA and Eng-siRNA in the P-Cofilin/Total-Cofilin ratio. (c). Using a real-time impedance-based cell analyzer (iCELLigence system, ACEA Biosciences), ECFC controls (DMSO) and ECFC treated by LIMKi were analyzed in basal condition and under 20 ng/mL TNFα stimulation (gray and black lines, respectively). As shown by the quantification in (d), a significant difference between control ECFC (gray bar) and ECFC treated with LIMKi (black bar) 12 h after TNFα stimulation (** p < 0.01). (e,f) Immunofluorescence for F-actin in ECFC controls vs. ECFC Eng-siRNA +/−TNFα (e) and ECFC controls vs. ECFC LIMKi +/−TNFα (f). (scale bar, 22 µm). (g) The quantification of (e) and (f) confirms that ECFC Eng-siRNA and ECFC LIMKi display the same modifications in terms of F-actin (* p < 0.05).