Human umbilical vein endothelial cells (HUVEC) were isolated as described (Wilkens et al., 2024) and cultured in EGM-2 medium (Lonza) at 37°C and 5% CO2 and used for experiments between passages 3 and 6 (Ethics Committee of Muenster University Clinic Approval 2009-537-f-S). Each experiment was reproduced with cells from different donors.

The following antibodies were used for immunofluorescence and immunoblotting: Rabbit monoclonal anti-human FoxO1 (clone C29H4 #2880, Cell Signaling, 1:1,000 for immunoblotting and 1:200 for immunofluorescence staining); rabbit monoclonal anti-human LC3B (clone D11 #3868, Cell Signaling, 1:1,000 for immunoblotting); mouse monoclonal anti-human Tie-2 (clone Ab33, #05-584, Sigma, 1 μg/mL for immunoblotting or 5 μg/mL for immunoprecipitation); mouse monoclonal anti-human α-tubulin (clone B-5-1-2 #T6074, Sigma, 0.5 μg/mL for immunoblotting); mouse monoclonal anti-phosphotyrosine (clone 4G10, #05-321, Sigma, 0.5 μg/mL for immunoblotting); mouse monoclonal anti-human VE-cadherin (clone F-8 #sc-9989, Santa Cruz Biotechnology, 1:100 for immunofluorescence staining); rabbit polyclonal anti-human VEGFR2-pY1054/59 (#44-1047G, invitrogen, 1:250 for immunofluorescence staining) rabbit polyclonal anti-human VE-PTP (VE-PTPh1-8, homemade (Li G. et al., 2020), 10 μg/mL for immunofluorescence staining); rabbit polyclonal anti-human VE-PTP (VE-PTP-C, homemade (Nawroth et al., 2002), 1 μg/mL for immunoblotting). Secondary antibodies conjugated to Alexa Fluor 488 (1:1,000) or Alexa Fluor 568 (1:1,000) and Horseradish Peroxidase (1:5,000–10,000) were purchased from Invitrogen and Jackson ImmunoResearch, respectively. Hoechst 33342, trihydrochloride, trihydrate (H3570) was from Invitrogen.

To inhibit VE-PTP, HUVECs were treated with 5 μM of AKB-9778 for 4 h before starting and during shear stress exposure.

VE-PTP expression in HUVEC was silenced with a combination of two siRNA against VE-PTP (Hs_PTPRB_5, 5′-UAACUUGAUAAAGUCGACCGG-3′ and Hs_PTPRB_10 5′-UAUCGUUCCACAUUCCCAGAA-3′, Qiagen); FoxO1 was silenced with siRNA 5′AAGAGCTGCATCCATGGACAA-3′ (#SI04435144, Qiagen); Tie-2 was silenced with siRNA 5′-TCGGTGCTACTTAACAACTTA-3’ (#SI00604919, Qiagen); VEGFR2 was silenced with a combination of four siRNA (SMARTpool) against VEGFR2 (5′-GGGCAUGUACUGACGAUUA-3′, 5′-CUACAUUGUUCUUCCGAUA-3′, 5′-GGAAAUCUCUUGCAAGCUA-3′ and 5′-GCGAUGGCCUCUUCUGUAA-3′ Dharmacon, # L-003148-00-0005). Allstar negative siRNA (Qiagen) was used as a negative control. siRNA was transfected using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer’s protocol.

HUVECs were exposed to shear stress as we previously described (Shirakura et al., 2023). HUVECs were plated on fibronectin-coated ibiTreat surface µ-Slide VI 0.4 (80606, ibidi GmbH), µ-Slide I Luer 0.8 (80196, ibidi GmbH), µ-Slide I Luer 0.4 (80176, ibidi GmbH) or µ-Slide I Luer 0.2 (80166, ibidi GmbH) at a density of 1 × 10⁵ cells/cm² and cultured for 48 h in EGM-2 medium (Lonza). The culture medium was replaced either with EBM-2 medium (Lonza) containing 2% FBS and 1% penicillin/streptomycin, or complete M199 buffer (#11150-59, Gibco) containing 20% FBS, 50 μg/mL ECGS (#354006, Corning), 100 μg/mL heparin and 1% penicillin/streptomycin, or M199 containing 2% FBS and 1% penicillin/streptomycin 4 h before applying cells to shear stress. Shear stress was applied for 24 h (10902 Pump System, ibidi GmbH, Germany) at 37°C in the same humidified incubator. The magnitude of shear stress was calculated according to the manufacturer’s protocol.

HUVEC on μ-slide VI 0.4 were first exposed to shear stress for 30 min, followed by fixation in 4% PFA in PBS for 15 min, blocking in 1% BSA for 1 hour at room temperature and finally by incubation with 0.5 μg/mL of primary antibodies or isotype control IgG at 4°C overnight. Primary antibodies were detected with peroxidase-conjugated secondary antibody and o-phenylenediamine dihydrochloride (OPD) substrate (Thermo Fisher Scientific). The procedures were subjected to HUVECs attached on the μ-slide. Optical density (OD) at 490 nm was measured with a plate reader (Synergy 2, LabTek). Data in each experiment were normalized to OD490 values in empty lanes.

HUVECs on μ-slide I Luer 0.4 were harvested in lysis buffer containing 20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM CaCl₂, 1.5 mM MgCl₂, 1 mM Na₃VO₄, 1% Triton X-100, 0.04% NaN₃ and 2 × cOmplete EDTA-free Protease Inhibitor Cocktail (Roche) and lysed for 30 min at 4°C followed by centrifugation for 30 min, 200,00× g at 4°C. Lysate aliquots were either directly analyzed by immunoblotting (see below), or subjected to immunoprecipitations by incubating with Protein G Sepharose 4 Fast Flow (GE Healthcare) and the respective antibody for 2 h at 4°C. Precipitated immunocomplexes were washed five times with the respective lysis buffer and dissolved in sample buffer (200 mM Tris-HCl pH 6.8, 30% glycerol, 6% SDS, 0.1% bromophenol blue, and 150 mM dithiothreitol) for Western blot analysis. For blotting, either precipitated immunocomplexes or HUVEC cell lysates were separated on 8% SDS-gels and transferred to nitrocellulose (Schleicher & Schuell) by wet blotting. Target proteins were detected by mouse monoclonal Ab33 anti-human Tie-2, rabbit polyclonal VE-PTP-c anti-human/mouse VE-PTP, or mouse monoclonal 4G10 anti-phosphorylated tyrosine. Blocking buffer contained 2% BSA and 200 μM Na₃VO₄ instead of 3% skimmed milk when phosphorylated proteins were detected. Phosphorylated Tie-2 was standardized to the expression of total precipitated Tie-2.

HUVECs on μ-slide VI 0.4 or I Luer 0.2, 0.4 or 0.8 were fixed with 4% PFA in PBS for 15 min, then permeabilized with 0.3% Triton X-100 in PBS for 5 min, followed by blocking with 1% BSA in PBS for an hour at room temperature. After blocking, the cells were incubated overnight with primary antibodies at 4°C. Subsequently, the cells were washed with PBS and incubated with secondary antibodies and Hoechst for an hour at room temperature. Images were obtained using a Zeiss LSM780 or 880 confocal microscopies.

Cell alignment and elongation were determined after cells were stained for VE-cadherin to define the cell borders. Confocal microscopic images (1,024 × 1,024 pixels, Zeiss LSM 780 or 880 inverted confocal microscope with x20 NA 0.80) were imported into ImageJ/Fiji. To define cell shape, the VE-cadherin channel was applied to “Subtract Background, rolling = 20” and the filter “Gaussian Blur at sigma = 3” followed by converting images to binary images with “Auto threshold with Huang method”. To detect orientation and length of major and minor axis in each cell, we fitted ellipses to the cellular outline in the binary image with “Analyze particle”. The angle of long axis against shear stress direction was considered as cell orientation. We analyzed the orientation and cell elongation index of >500 cells in at least 5 images for each flow chamber lane. Average value in each flow chamber lane was calculated and plotted into graph.

Confocal microscopic images were analyzed with ImageJ Fiji. To prepare the nuclear mask, the Hoechst 33342 signal was applied to the filter “Gaussian Blur at sigma = 1.5”, and then the images were converted to binary images with “Auto threshold with Huang method”. Nuclear FoxO1 signals were extracted by applying the mask to the images of FoxO1 using the “image Calculator”.

Values are expressed as mean ± standard error of the mean (SEM). Differences between two groups were assessed by two-tailed Welch’s t-test or Mann-Whitney U-test and among more than two groups were assessed by ANOVA followed by Bonferroni correction test, Dunnett’s test, or Tukey test for multiple comparisons (Prism 10). The statistical significance of differences is shown in figure legends. P values < 0.05 were considered statistically significant.

To investigate the potential role of VE-PTP for cellular morphological responses to fluid shear stress, we analyzed cell alignment and elongation of HUVECs exposed to various magnitudes of shear stress. First, we applied 5 dyn/cm² of shear stress to HUVECs transfected with VE-PTP or control siRNA (Figures 1A–C). In control cells, 5 dyn/cm² of shear stress induced slight cell alignment, but no cell elongation along the direction of flow. VE-PTP silencing, however, promoted cell alignment and cell elongation. The same effects were observed when VE-PTP silencing was replaced by the treatment of cells with the pharmacological VE-PTP inhibitor AKB-9778 (Supplementary Figures S1A–C). These results indicate that VE-PTP suppresses cellular morphological responses to low 5 dyn/cm² shear stress levels, a function which depends on its phosphatase activity.

Next, we applied 15 dyn/cm² of shear stress to HUVECs transfected with VE-PTP or control siRNA (Figures 1D–F). In control cells, 15 dyn/cm² of shear stress induced well detectable cell alignment and elongation along to the direction of flow, whereas VE-PTP silencing suppressed cell alignment, but had no effect on cell elongation. Likewise, AKB-9778 treatment also suppressed cell alignment, but did not affect cell elongation (Supplementary Figures S1D–F). Thus, VE-PTP plays opposite roles for cellular morphological responses to low (5 dyn/cm²) and moderate (15 dyn/cm²) shear stress levels.

Analyzing high shear stress levels, we exposed HUVEC to 30 dyn/cm² of shear stress. This treatment induced cell alignment and cell elongation perpendicular to the direction of flow which occurred, interestingly, independent of whether VE-PTP was silenced or not (Figures 1G–I). Taken together, VE-PTP silencing increases the sensitivity of endothelial cells to low shear stress levels, reduces the sensitivity to moderate levels and is not relevant for the perpendicular alignment of cells at high levels of shear stress.

We have recently demonstrated that shear stress induces endocytosis of VE-PTP leading to reduced levels of VE-PTP on the cell surface, which results in increased phosphorylation of Tie-2 (Shirakura et al., 2023). In addition, we have reported that Tie-2 signaling is essential for cell alignment and elongation (Ghanbarpour Houshangi et al., 2025). Based on these studies, we now asked whether sensing of increasing magnitudes of shear stress levels would lead to inversely proportional levels of VE-PTP on the cell surface. To examine this, we exposed HUVECs to 5 and 15 dyn/cm² of shear stress for 30 min, followed by treatment with or without TritonX-100 and antibody staining for VE-PTP. Staining of detergent permeabilized cells allowed to detect the accumulation of VE-PTP in intracellular vesicles of cells exposed to 5 and 15 dyn/cm² (Figures 2A,B). Staining of non-permeabilized cells revealed that VE-PTP surface levels decreased gradually, with increasing levels of shear stress (Figures 2A,B). In agreement with these results, a cell surface ELISA assay showed that an increase of shear stress from 5 to 15 dyn/cm² led to a decrease of VE-PTP levels on the cell surface (Figure 2C). Such a negative correlation of VE-PTP surface levels and fluid shear stress levels was still observed after exposure of HUVEC to shear stress of different magnitude for 24 h (Supplementary Figures S2A,B). Collectively, these results suggest that fluid shear stress levels determine VE-PTP surface levels in an inversely proportional and magnitude-dependent manner.

We next investigated whether VE-PTP surface levels determine Tie-2 phosphorylation levels. To this end, we exposed HUVEC to static culture conditions or 5 and 15 dyn/cm² fluid shear stress followed by immunoblotting Tie-2 immunoprecipitants with the pan phospho-tyrosine antibody 4G10. As shown in Figure 2D, tyrosine phosphorylation of Tie-2 increased gradually with increased shear stress. The increase at 15 dyn/cm² was significant, while the increase at 5 dyn/cm² just missed being significant. We further investigated whether the less phosphorylation of Tie-2 at 5 dyn/cm² fluid shear stress is due to the less VE-PTP internalization. We exposed HUVEC transfected with control or VE-PTP siRNA to 5 dyn/cm² fluid shear stress and examined Tie-2 phosphorylation levels (Figure 2E). VE-PTP silencing significantly increased Tie-2 phosphorylation at 5 dyn/cm² fluid shear stress.

We have recently found that the signaling mechanism whereby Tie-2 controls cell alignment is mediated by the stimulation of FOXO1 translocation from the nucleus into the cytosol which leads to activation of autophagy (Ghanbarpour Houshangi et al., 2025). Based on this, we now tested whether these effects would also gradually increase with a gradual increase of fluid shear stress. Comparing static conditions and three different shear stress levels (4, 15 and 59 dyn/cm²) we found that the levels of FOXO1 in the nucleus decreased. In addition, we detected in immunoblots that the levels of activated LC3B, a central protein in autophagy (Tanida et al., 2008; Li X. et al., 2020), increased gradually with increased shear stress (Supplementary Figures S2C,D). Together, these results establish that the magnitude of shear stress is a key parameter to adjust VE-PTP internalization and the downstream signaling effects on the Tie-2-FOXO1-autophagy axis.

Although we have shown that Tie-2 is required for cell alignment and elongation upon exposure to fluid shear stress (Ghanbarpour Houshangi et al., 2025) and we have shown here that VE-PTP is also needed (Figure 1), we did not yet know, whether this function of VE-PTP is indeed mediated by Tie-2. This is especially important since VEGFR2 and VE-cadherin are also well described substrates of VE-PTP. Therefore, we tested the effect of VE-PTP inhibitor AKB-9778 on the morphological response of HUVEC cells to 5 dyn/cm² fluid shear stress after treating the cells with either Tie-2 or control siRNA. We found that AKB-9778 promoted the parallel cell alignment and elongation in HUVECs treated with control siRNA, whereas these effects were blocked for cells treated with Tie-2 siRNA (Figures 3A–C). We verified Tie-2 silencing efficiency by immunoblot (Figure 3D). In line with this, silencing FOXO1, a downstream target of Tie-2 signaling, blocked VE-PTP siRNA effects on cell alignment and elongation as well (Figures 3E–G). We verified FoxO1 and VE-PTP silencing efficiency by immunoblot (Figure 3H) Thus, VE-PTP regulates the endothelial cell morphological responses to fluid shear stress through Tie-2 activity.

Since we had found that VE-PTP silencing did not affect cell alignment and elongation perpendicular to the direction of flow when cells we exposed to 30 dyn/cm², we tested whether Tie-2 is relevant. Interestingly, we found that silencing of Tie-2 by siRNA did block these morphological cellular responses (Supplementary Figure S3). This result shows that Tie-2 is relevant for this cellular morphological response to high shear stress, but the effect cannot be further increased by interference with VE-PTP activity.

Our results so far suggested that VE-PTP inhibition increases the sensitivity of endothelial cells to fluid shear stress by increasing the activation level of Tie-2. To verify this conclusion, we tested whether the Tie-2 agonist Angpt1 would have similar effects as inhibition of VE-PTP. To this end, we exposed HUVEC to 5 dyn/cm² fluid shear stress in the presence of Angpt1 or vehicle alone. We found that Angpt1 significantly promoted parallel cell alignment and elongation (Figures 4A–C). Repeating similar experiments under higher fluid shear stress (15 dyn/cm²) resulted in an inhibitory effect of Angpt1 on parallel cell alignment and no effect on cell elongation (Figures 4D–F), similarly as we found for VE-PTP siRNA (Figures 1D–F). Taken together, these results further verify that the Tie-2 activation level determines the sensitivity to shear stress levels, and VE-PTP acts via Tie-2 in this process.

It is well known that the mechanosensory complex (PECAM-1, VE-cadherin, VEGFR2/3) controls the sensitivity to shear stress (Baeyens et al., 2015). Therefore, we asked how the VE-PTP-Tie-2 mediated control of shear stress sensitivity is related to the function of the mechanosensitive complex. First, we tested whether VE-PTP silencing would affect shear stress induced phosphorylation of VEGFR2. Exposing HUVEC for 1 min or for 5 min to 15 dyn/cm² fluid shear stress, we found that an antibody against pY1054/59 of VEGFR2 allowed specific staining at junctions after 1 min of flow (Figures 5A,B), which is in agreement with earlier findings (Coon et al., 2015). Silencing of VE-PTP by siRNA treatment did not affect this staining. (Figures 5A,B). We confirmed the specificity of the antibody by staining HUVECs after incubation under static conditions for 10 min with VEGF-A. Positive staining was seen with cells pretreated with control siRNA and was lost after pre-treatment with VEGFR2 siRNA (Supplementary Figures S4A,B). Likewise, staining of junctions with antibodies against VEGFR2-pY1054/59 after 1 min of fluid shear stress (15 dyn/cm²) was abrogated after pre-treatment with VEGFR2 siRNA (Supplementary Figures S4C,D).

Next, we analyzed the effects of VE-PTP inhibition to cell morphology in the absence of VEGFR2. Regardless of VEGFR2 expression, VE-PTP inhibition with AKB-9778 suppressed cell alignment and elongation induced by shear stress at 15 dyn/cm² (Figures 5C–E). We verified VEGFR2 silencing efficiency by immunoblot (Figure 5F). This indicates that VE-PTP-Tie-2 signaling controls cell morphological response to fluid shear stress independently of the mechanosensory complex. Surprisingly, however, VEGFR2 silencing did not inhibit cell alignment and elongation in our experimental conditions. Rather VEGFR2 silencing slightly but not significantly enhanced cell alignment and elongation. Together, these results show that the VE-PTP/Tie-2 signaling axis acts independently of the mechanosensory complex in controlling the cellular morphological responses to fluid shear stress. In addition, the medium conditions we used here may have an effect on how VEGFR2 influences cell alignment to flow.

It was unexpected and surprising that VEGFR2 silencing did not inhibit cell alignment and that it even increased cell alignment in the presence of VEGF-A. In previous reports (Coon et al., 2015; Chen et al., 2024; Tanaka et al., 2024), the roles of the mechanosensory complex have been mainly studied with HUVEC in M199 medium, whereas we performed our flow experiments throughout our study in EBM-2 medium (Lonza). Thus, we hypothesized that the different media compositions might influence the role of VEGFR2 in flow induced cell alignment.

To test this, we compared the cellular responses in EBM-2 medium with those in complete M199 medium. HUVEC were exposed to 12 dyn/cm² in each of the two media after pretreatment with either control, Tie-2 or VEGFR2 siRNA. We found that Tie-2 siRNA inhibited cell alignment and elongation regardless of the medium, whereas VEGFR2 siRNA suppressed cell alignment and elongation only in the complete M199 medium (Figures 6A–C). Complete M199 medium contains 20% FCS and several additional growth factors based on the addition of endothelial cell growth supplement (ECGS). In contrast, EBM-2 medium simply contained 2% FCS. To exclude differences in growth factors as a reason for the different cellular responses, we repeated our experiments with EBM-2 and M199 media each only containing 2% FCS without additional growth supplements added. We observed that a low percentage of HUVECs transfected with Tie-2 siRNA were detached in M199 media containing 2% FCS. Under these conditions, again, we found that Tie-2 siRNA suppressed cell alignment and elongation regardless of the medium, whereas VEGFR2 siRNA treatment suppressed cell alignment and elongation only in M199 medium, but not in EBM-2 medium (Figures 6D–F). These results indicate that the biochemical composition of the culture medium affects functioning of VEGFR2-based but not Tie-2-based sensing of fluid shear stress. Collectively, our results suggest that VE-PTP-Tie-2 is a novel pathway for the control of cell morphology, which responds in a quantitative way to different fluid shear stress magnitudes.