Human umbilical vein endothelial cells (HUVECs, PromoCell) were cultured in endothelial cell medium (PromoCell) added with Growth Medium Supplement Pack (PromoCell) in a humidified atmosphere at 37°C and 5% CO₂. Human aortic endothelial cells (TeloHAEC) were immortalized by stably expressing human telomerase catalytic subunit hTERT (ATCC). In addition, TeloHAECs were also stably transduced with the IncuCyte^TM NucLight^TM red Fluorescent Protein Lentivirus Reagents (Essen BioScience) and with Tead Luciferase lentiviral construct (Vectalys) containing wild type or 7× mutated (inactive form) TEAD-responsive synthetic element/region (8×GTIIC region followed by a minimal chicken TNNT2 promoter, before the luciferase gene, called TEADwt-Luc and TEADmut-Luc, respectively) driving luciferase expression (Kim and Gumbiner, 2015). To mimic the diabetes environment, HUVECs or TeloHAECs were cultured with high glucose (15 or 25 mM) for 24 h under static conditions. In order to replicate the in vivo-like flow conditions of an artery, HUVECs were additionally exposed to fluidic conditions. Briefly, HUVECs (2,5 × 10⁵ cells) were seeded onto glasses 0.4 mm Luer μ-slides I (Ibidi; channel length, 50 mm; channel width, 5 mm; and channel height, 0.4 mm) coated with rat tail collagen I (Gibco; 100 μg/ml in PBS) and grown until confluence. The slides were assembled into flow chambers and connected to an Ibidi Pump System (Ibidi). HUVECs were cultured with high glucose (25 mM) and exposed to steady laminar shear stress (12 dyn/cm²) or oscillatory shear stress (0.5 ± 6 dyn/cm²; 1 Hz) for 72 h without renewal of the culture medium. Glucose concentration at baseline (0 h) or after 72 h was not assessed: given the volume of medium used in the Ibidi Pump System (12 ml) and based on previous reports (Barcelos et al., 2009; Mangialardi et al., 2013; Zhang J. et al., 2019), glucose concentration is not expected to change in these experimental settings. The flow system was maintained at 37°C, and the circulating medium was equilibrated with a humidified atmosphere of 5% CO₂. For all experiments, HUVECs were used between P2 and P5. Human monocytic cell line THP-1 (Leibniz Institute DSMZ-German Collection, GmbH) were transfected with the IncuCyte^TM NucLight^TM Green Fluorescent Protein (GFP) Lentivirus Reagents (Essen BioScience). THP-1-GFP were cultured in RPMI 1640 medium (Gibco) and 10% FBS (Gibco).

Inhibition of the YAP/TAZ signaling pathway was obtained by using either siRNAs against YAP and TAZ or the small molecule TEAD inhibitor K-975 (Kaneda et al., 2020). For siRNA-mediated inhibition of YAP/TAZ, HUVECs were transfected with siRNAs using Lipofectamine^TM RNAiMAX reagent (Thermo Fischer) in Opti-MEM and L-Glutamine (Gibco) according to the manufacturer’s instruction. Endogenous YAP/TAZ expression was knocked down by transfecting cells with 100 nM of siRNA against YAP (Silencer-select siRNA s20366, Ambion) and TAZ (ON-TargetPlus siRNA, Horizon discovery). Control cells were transfected with negative control siRNA (Ambion). Cells were incubated with transfection complexes at 37°C in 5% CO2 for 4 h and media was replaced with complete endothelial growth medium for 20 h. Then, HUVEC were cultured for additional 24 h on either normal (5 mM) or high glucose (25 mM) condition in endothelial cell medium added with Growth Medium Supplement Pack (PromoCell). For the purpose of TEAD inhibition, K-975 was synthesized internally based on Kaneda et al. (2020). K-975 at 200 nM was added to HUVECs cultured on either normal (5 mM) or high glucose (25 mM) for 24 h, and under either static conditions or subjected to fluidic conditions. DMSO (0.01%, v/v) was used as control.

TeloHAECs were seeded at the density of 15 × 10³ cells/well in a 96-well plate format and maintained in a complete medium (CM) composed of Vascular Cell basal medium (BM), 5.5 mM glucose supplemented with Vascular Endothelial Cell Growth Kit-VEGF with 2% FBS and cytokines (ATCC), and incubated at 37°C in 5% CO₂ for 24 h. Then, the medium was removed, and cells were kept in BM supplemented with 0.1% BSA with or without K-975, or under high glucose conditions (25 mM). After 24 h of incubation, the fluorescence intensity was detected by a fluorescence plate reader (Clariostar, BMG Labtech). Then, cells were lysed and the Luc-Screen^TM Extended-Glow Luciferase Reporter Gene Assay System (Invitrogen) was used to detect luciferase activity (according to the manufacturer’s instruction). This reporter assay allows a direct investigation of the interaction between YAP/TAZ and TEAD: if TEAD interacts with YAP/TAZ the luciferase activity increases; if their interaction is prevented (i.e., by a TEAD inhibitor like K-975) the luciferase activity decreases.

The adhesion assay was performed by perfusing the THP-1-GFP positive cells on HUVECs cultured under either normal or high glucose conditions (with or without K-975, 200 nM), and already subjected to laminar flow (12 dyn/cm2) for 72 h. THP-1-GFP were resuspended at the concentration of 5 × 10⁵ cells per mL in RPMI 1640 (supplemented with 10% serum), then perfused on the HUVECs monolayer for 60 min at 37°C. To allow THP-1 cells attachment to HUVECs, the laminar flow conditions were set at 5 dyn/cm². After removing the non-attached THP-1 cells (by washing with RPMI), fluorescent images were taken using an IN Cell Analyser 2200 (GE Healthcare). To distinguish THP-1-GFP cells from HUVECs, HUVECs were prelabelled with Hoechst (Invitrogen). IN Cell Analyser software was used to quantify the number of adherent THP-1 cells.

Human umbilical vein endothelial cells were lysed in ice-cold cell extraction Buffer (Invitrogen) supplemented with protease/phosphatase inhibitor mixture (Sigma). Protein concentration was determined with the Pierce BCA Protein Assay Kit (Thermo Fischer). Equal amounts of protein (10 μg) were loaded and resolved by SDS/PAGE. After electrophoresis, proteins were transferred to PVDF membranes, followed by blocking with 10% blotting-Grade Blocker (Bio-Rad). The membranes were incubated with primary antibodies against phospho-YAP (Ser-127) (1:1,000), YAP (1:1,000), phospho-TAZ (Ser-89) (1:1,000), TAZ (1:2,500), VCAM-1 (1:1,000), ICAM-1 (1:1,000), cytokines connective tissue growth factor (CTGF; 1:1,000), cysteine-rich angiogenic inducer 61 (CYR61; 1:1,000) (all from Cell Signaling), GAPDH (Santa Cruz Biotechnologies, 1:10,000), β-actin (Sigma, 1:10,000) with gentle agitation overnight at 4°C. The membranes were washed three times for 10 min each with TBS-T (containing 0.1% Tween-20) and incubated with horseradish-peroxidase (HRP)-conjugated anti-mouse-IgG (Jackson ImmunoResearch, 1:10,000) or HRP-conjugated anti-rabbit-IgG (Jackson ImmunoResearch, 1:10,000) followed by chemiluminescence detection using GE Healthcare ECL western blotting substrate according to the manufacturer’s protocol. The blot image was acquired using G:BOX Chemi XL 1.4 (Syngene) imaging system and the band density was quantified by using GENETOOLS V4.03 (Syngene) software.

Total RNA was isolated from HUVECs by using the Maxwell RSC simply RNA Tissue kit (Promega). The concentration of RNA was determined by using a NanoDrop spectrophotometer (Thermo Fischer). One microgram of total RNA was reversely transcribed into cDNA by using SuperScript VILO cDNA Synthesis Kit (Invitrogen). The real-time PCR was carried out with TaqMan Universal PCR Master Mix (Applied Biosystems) on Stratagene Mx3000P (Agilent Technologies) detection system. Primer sets used in our study (predesigned by Thermo Fischer) were as follows: CTGF (Hs00170014_m1); CYR61 (Hs00155479_m1). The relative gene expression was determined by 2^–ΔΔCt method.

Human umbilical vein endothelial cells were briefly washed with sterile HBSS (Gibco), detached from the plate using versene with 5% of trypsin-EDTA (0.025%, v/v), then washed with staining buffer (BD Pharmingen). Cells were then transferred to a FACS plate, washed and resuspended in 10% PE-conjugated mouse anti-human ICAM-1 and FITC-conjugated mouse anti-human VCAM-1 (both from BD Pharmingen) in staining buffer for 30 min on ice and in the dark. Then cells were washed, resuspended in HBSS and analyzed using a Guava EasyCyte flow cytometer (Merck-Millipore). Data analysis was performed using the Guava software (Merck-Millipore). Mouse IgG1κ-FITC and PE were used as isotype controls.

Diabetic male db/db mice and sex- and age-matched non-diabetic db/ + control mice (on the C57BL/6J background; Jackson Laboratories) were used to conduct animal studies. Mice were housed with free access to standard chow diet (Harlan) and water. Blood glucose levels were measured before tissue collection with Accu-Chek^® glucose meter (Roche). All procedures involving animals were performed in agreement with the European Community standard on the care and use of laboratory animals (2010/63/UE) and were approved by the IACUC of Sanofi R&D. All procedures were performed in AAALAC-accredited facilities in full compliance with the recommendations of the French Ministry of Research.

To induce changes in shear stress, a perivascular shear stress modifying cast was tied around the right common carotid artery of 18-week-old mice for 4 weeks. The cast consists of 2 longitudinal halves which, when put together, form a cylinder with an inner diameter of 500 μm (non-constrictive) that gradually declines to 250 μm, thus producing a gradual stenosis and gradual increase in shear stress (Cheng et al., 2005) (Sepran; Sepra, Delft, Netherlands). Briefly, mice were anesthetized with isoflurane, and the right anterior cervical triangle was accessed by incision. The right common carotid artery was dissected from connective tissue and both halves were placed and tied with a suture. Wounds were closed and animals were allowed to recover. The left carotid artery was left untouched and served as control.

Twenty-two-week-old mice were euthanized by exsanguination under profound anesthesia and immediately perfusion-fixed with heparinized PBS followed by 4% formalin. Both left (control) and right carotid arteries were removed and embedded in paraffin. All samples were cut into transverse sections (5 μm in thickness) and prepared for subsequent immunohistochemical analyses. For cast-instrumented arteries, sections were cut approximately 250 μm downstream of the cast, in the region of low and oscillatory shear stress (Cheng et al., 2005). Immunostaining were performed using the Ventana Discovery XT or BenschMark Ultra automated Systems (Ventana Medical System, Inc., Roche) with standard DabMab or UltraMab detection systems. Paraffin-embedded sections were incubated with primary antibodies against phospho-YAP (Ser127) (D9W2I, Cell Signaling, 1:1,000), YAP (D8H1X, Cell Signaling, 1:50) and VCAM-1 (ab134047, Abcam, 1:500). Optimal concentrations for each antibody were determined by individual pilot studies. Samples were counterstained with hematoxylin II and bluing reagent (Ventana) and coverslipped with Pertex. For all immunostaining, images of the entire slides were acquired on an Olympus Virtual Slide scanner (VS120) and quantitatively analyzed on whole tissue sections using Area Quantification Brightfield v1.0 module of Indica Labs HALO software (based on the intensity of the immunolabeling). Briefly, the intima-media area of the analyzed sections was manually delimited on images and positive stained area including weak (in yellow), moderate (in orange) and strong (in red) positive pixels was automatically quantified for each animal (as illustrated in Supplementary Figure 1).

Values are presented as mean ± standard error of mean (SEM). Non-parametric two-tailed Mann-Whitney test or parametric two-tailed Student’s t test was used for comparisons between two groups, and non-parametric Kruskal-Wallis test was used to compare more groups, respectively. Statistical analysis was performed using GraphPad Prism software 8.0.2 A p value < 0.05 was interpreted to denote statistical significance.

A series of in vitro experiments using two endothelial cell models, HUVECs and TeloHAECs, were carried out to investigate the consequence of high glucose culture conditions on YAP signaling. As shown in Figure 1A, phospho-YAP was decreased in HUVECs exposed to high glucose, as compared to normal glucose. By contrast, no differences were observed on total YAP, phospho-TAZ and total TAZ (Figures 1A,B). In addition, Western blot and qRT-PCR analysis showed that the expression of both CTGF and CYR61 (two transcriptional target genes of YAP/TAZ/TEAD) (Zhao et al., 2008; Shome et al., 2020) were increased by high glucose conditions (Figure 1C and Supplementary Figure 2, respectively). Accordingly, the signal response was increased by high glucose in TeloHAECs transduced with TEAD-luciferase reporter (Figure 2A). The specificity of the model was previously validated by using inactive TEAD-mutant or inhibition with TEAD inhibitor K-975 (Supplementary Figure 3). Taken together, these results suggest that high glucose activates YAP but not TAZ in endothelial cells.

The consequences of YAP activation in endothelial cells in response to high glucose were characterized in HUVECs by hampering YAP activity by using either siRNA-mediated YAP/TAZ deletion or the TEAD inhibitor K-975. VCAM-1 expression was monitored as a marker of endothelial activation known to be highly sensitive to YAP/TAZ deletion (Choi et al., 2018) and CTGF was chosen as a downstream signal of TEAD activity (Zhao et al., 2008). The expression of VCAM-1 and ICAM-1 was induced by high glucose as assessed by Western blot (Figures 2B,C,E for VCAM-1 and Figure 2E for ICAM-1) and flow cytometry (Figure 2F). siRNA-mediated deletion of YAP/TAZ prevented the increase in VCAM-1 (Figures 2B,C) and CTGF expression (Figures 2B,D) as observed under high glucose exposure. Accordingly, K-975, as a blocker of the YAP/TEAD activation cascade, blunted glucose-mediated alteration of VCAM-1 and ICAM-1 expression and cell surface expression (Figures 2E,F). These findings suggest that YAP/TAZ signaling plays a critical role in high glucose-induced endothelial activation. Confirmation of siRNA-mediated deletion of YAP/TAZ is shown in Supplementary Figure 4. The residual expression of TAZ observed after the siRNA approach is in line with what has been already published by others (Yu et al., 2012; Rausch et al., 2019). Importantly, combination of YAP and TAZ siRNAs was preferred to siRNA YAP alone to efficiently block the pathway and avoid any possible compensation due to the residual expression of TAZ, as already shown (Nallet-Staub et al., 2014; Choi et al., 2018). Supplementary Figure 5 shows that K-975-mediated inhibition of YAP/TEAD prevented the increased expression of the YAP/TEAD downstream effectors CTGF and CYR61 otherwise observed in HUVECs under high glucose, thus confirming the inhibitory effect of K-975 on YAP/TEAD signaling.

Since vascular mechanotransduction was shown to activate YAP/TAZ signaling, we next evaluated the effects of high glucose on YAP phosphorylation in HUVECs exposed to oscillatory flow, compared to laminar flow conditions. Whereas the expression of total YAP, phospho-TAZ and total TAZ was not altered by high glucose under either laminar or oscillatory flow, phospho-YAP significantly declined under laminar flow and high glucose exposure (Figure 3A). YAP phosphorylation was reduced by oscillatory flow conditions, but high glucose additional effect was not significant (Figure 3A). The expression of VCAM-1 (Figure 3B) was increased by high glucose under laminar flow, as compared to normal glucose conditions. Similarly, VCAM-1 expression increased under oscillatory flow conditions, but high glucose in combination with oscillatory flow conditions did not exert any additional effects on VCAM-1 expression (Figure 3B). Interestingly, a similar response was also observed for the expression of CYR61 (Figure 3B and Supplementary Figure 6). Subsequently, the rate of monocyte-endothelial cells attachment was increased by high glucose under laminar flow, as compared to normal glucose conditions (Figure 3C). Conversely K-975 blunted the glucose-induced THP1 monocyte attachments to HUVECs (Figure 3C). The rate of monocyte attachment to HUVECs subjected to oscillatory flow was significantly increased compared to laminar flow exposure (Figure 3D). However, no additional effects were observed when cells were cultured in high glucose conditions (Figure 3D).

Next, we aimed to evaluate the effects of high glucose and shear stress on YAP in the vasculature of type 2 diabetic mice. First, high blood glucose levels in db/db mice (658 ± 115 mg/dL) compared to db/ + mice (255 ± 22 mg/dL) were confirmed before tissue collection. Then, we analyzed the vascular wall of carotid arteries without casts in db/db and db/ + mice. As shown in Figures 4A,C, compared to db/ + mice, an enlargement of the intima-media thickness was observed in control carotid arteries of db/db mice. Phospho-YAP was detected in the vascular wall of db/ + mice and significantly decreased in db/db mice (Figures 4A,D). No difference in the expression of total YAP was observed (Figures 4A,E), thus resulting in a decrease in phospho/total YAP ratio and YAP activation in the vascular wall of diabetic mice (Figure 4F). Additional immunohistochemical analysis showed that the expression of VCAM-1 increased in the carotid artery vascular wall of diabetic mice (Figures 4A,G). Finally, with the exception of VCAM-1 expression, no differences were observed for all other parameters in cast-instrumented carotid arteries compared to carotid arteries without cast (Figures 4B–G), and no differences were observed in cast-instrumented carotid arteries between diabetic and non-diabetic mice (Figures 4A–G).