Endothelial RhoGEF Trio is involved in angiogenesis

Abstract

Endothelial cells lining the vessel wall are instrumental in angiogenesis. They initiate this process and remodel their actin cytoskeleton to facilitate proliferation and migration from pre-existing vessels. However, the mechanisms that coordinate the remodeling of the endothelial actin cytoskeleton to promote angiogenesis are not fully understood. Here, we show that the RhoGEF Trio is involved in angiogenic sprouting in vitro and in vivo. Inhibition of Trio activity reduces sprout formation of intersomitic vessels in embryonic zebrafish. Moreover, a strong reduction in the length of developing sprouts is observed in a murine retinal explant assay. In addition, embryonic lethally Trio^−/− mice show impaired disposition of the vasculature. Mechanistically, we show through rescue experiments that the N-terminal part of Trio, including the GEF1 domain, supported endothelial sprouting. Together, we conclude that Trio is involved in angiogenesis.

Introduction

The process of blood vessel formation by sprouting from pre-existing vessels is known as angiogenesis and is involved in many physiological and pathological conditions, including embryonic and post-natal development, wound healing, tumor growth and the formation of collaterals during atherosclerotic plaque formation (1, 2). One of the key modulators of sprouting angiogenesis is the actin cytoskeleton (3). The actin cytoskeleton is regulated through small GTPases that cycle between an active, GTP-bound and inactive GDP-bound state (4). In particular, the small GTPase Rac1 is instrumental in vascular development since endothelial-specific deletion of Rac1 induces embryonic lethality at E9.5, due to hemorrhages and reduced vascular networks (5). Defects were found in the development of the major vessels as well as a complete lack of branching. Postnatal Rac1 depletion led to delayed and aberrant retinal angiogenesis at P8. Together, these observations point to a crucial role for Rac1 in angiogenesis (5). One of the major angiogenic factors, VEGF/VEGFR2, triggers a pathway in endothelial cells (ECs) linking VEGF to the remodeling of the actin cytoskeleton through Rac1 (6). In line with these findings, blocking Rac1 activity or depleting Rac1 from ECs suppressed sprout formation (7). However, how Rac1 activity is regulated during angiogenic sprouting conditions to potentially stabilize cell–cell junctions is not clear.

Cortical actin is required to support endothelial sprout outgrowth by maintaining the stability of VE-cadherin-based cell–cell contacts (8). Our earlier work additionally demonstrated that the RhoGEF Trio stabilizes VE-cadherin-based cell–cell junctions (9, 10). However, whether Trio is the responsible GEF for angiogenic sprouting is not known. In this study, we show that reducing Trio expression using shRNA or blocking Trio or Rac1/RhoG activity perturbs the outgrowth of sprouts using in vitro sprouting assays, using retinal explants in an ex vivo setting and in developing zebrafish in vivo. Analysis of Trio^−/− embryos shows deformation of PECAM-1-positive vessels. Mechanistically, we show that the N-terminal domain of Trio, but not the TrioGEF1 domain only, is sufficient to rescue impaired sprouting phenotype. Together, our data show an important role for Trio in efficient angiogenesis.

Materials and methods

Antibodies and reagents

Polyclonal antibody (pAb) to Trio (clone D-20) and monoclonal antibody (mAb) VE-cadherin (clone F8) were from Santa Cruz Biotechnologies (Germany). Actin (clone AC-40) mAb was purchased from Sigma-Aldrich (the Netherlands). mAb GFP (clone JL8) was purchased from Clontech/Westburg (the Netherlands). Secondary horseradish peroxidase (HRP)-conjugated goat-anti-mouse, goat-anti-rabbit and rabbit-anti-goat antibodies were purchased from Dako (Belgium).

Adenovirus production

GFP-TrioD1 and GFP-TrioN were generated as previously described (11). Adenovirus expressing GFP-TrioD1 and GFP-TrioN was produced by transfecting PacI-digested (Westburg, Leiden) constructs into HEK293T cells.

Cell culture and transfections

Primary human umbilical vein endothelial cells (HUVECs) were purchased from Lonza (USA) and were cultured on fibronectin-coated (30 μg/mL) tissue-culture-treated culture dishes (TPP, Switzerland) in EGM2-containing SingleQuots (Lonza). For optimal cell culturing, 25,000 cells per 1 cm² were plated and cultured for 48 h before harvesting for the experiments. ECs were cultured up to no more than six passages. VE-cadherin null ECs were a kind gift from Prof. Dr E. Dejana (IFO, Italy). To inhibit TrioGEF1 activity, cells were treated with ITX3 (50 μM), purchased from Merck Millipore. Rac1 activity was blocked by treating cells with 12.5 μM EHT 1864 (Selleckchem, USA). shRNA constructs targeting Trio (shTrio#1; TRC864 and shTrio#2; TRC10561) and a non-specific control shRNA (shCtrl; shc002) were cloned in a tag-RFP vector (pLKO.1-PURO-CMV-TagRFP construct from Sigma) and were used to produce lentivirus in HEK293T cells by using the third-generation packaging plasmids (11). The lentivirus-containing supernatant was harvested after two and three days after transfection, filtered and concentrated by using Lenti-X Concentrator (Clontech, France) according to the manufacturer’s protocol. Lentivirally transduced cells were used three days after transduction for further processing. Adenovirally transduced cells were used one day after transduction and sorted.

Western blotting

Cells were washed three times with ice-cold phosphate-buffered saline (PBS) containing 1 mM CaCl2 and 0.5 mM MgCl2 and boiled in SDS sample buffer containing 4% mercaptoethanol. Samples were analyzed by SDS-PAGE or 3–8% Tris–acetate gradient gels (Invitrogen) and subsequently transferred to a 0.2 μA nitrocellulose membrane (Whatman, Germany) and blocked with blocking buffer containing 5% (w/v) non-fat dry milk in Tris-buffered saline with Tween-20 (TBST). Samples were loaded with typically 25 μg protein per lane. Subsequently, the membrane was incubated with primary antibodies for 1 h at room temperature, followed by incubation with secondary HRP-conjugated antibodies. In between the incubation steps, the blots were washed at least three times with TBST for 5 min. Stainings were visualized by using an enhanced chemiluminescence detection system (Pierce, USA).

Confocal laser scanning microscopy

Cells were cultures on FN-coated glass coverslips and transfected as indicated. After treatment, cells were washed with room temperature PBS, containing 1 mM CaCl₂ and 0.5 mM MgCl₂ and fixed in 3.7% (w/v) formaldehyde for 15 min. After fixation, wells were permeabilized in PBS supplemented with 0.5% (v/v) Triton X-100 for 2 min. Next, cells were incubated with primary and secondary antibodies and were washed in between steps. Histological sections of Trio-deficient embryos were fixed in acetone (Merck Millipore, USA) at 4°C for 1 h. After fixation, cells were incubated with primary and secondary antibodies and, after each step, washed with PBS. Fluorescent imaging was performed with a confocal laser scanning microscope (LSM510/Meta; Carl Zeiss MicroImaging, Germany).

Sprouting angiogenesis assay

In vitro endothelial sprouting was studied using a collagen gel-based EC spheroid model, as described by Korff et al. (12). Prior to forming spheroids, VE-cadherin null ECs transfected with VE-cadherin-GFP or VE-cadherin-Δβ-catenin-binding domain-GFP were sorted for GFP expression using the BD FacsAriaIII (BD, the Netherlands). Furthermore, ECs transduced with either shRNA control-tagRFP, shRNA Trio-tagRFP and GFP-TrioD1 or GFP-TrioN were sorted for both RFP and GFP expression prior to seeding. To generate multicellular spheroids, 7,500 HUVECs/mL were suspended in EGM2 medium supplemented with 0.1% (w/v) methyl cellulose (4000 cP viscosity; Sigma-Aldrich, USA) and 100 μL/well were seeded in a non-adherent round-bottom 96-well plate (Greiner, Germany). Under these conditions, suspended HUVECs aggregate spontaneously within 16 h to form a single endothelial spheroid. These spheroids were embedded in a type I collagen matrix consisting of EGM2 medium supplemented with 20% fetal calf serum (FCS), 25 ng/mL recombinant human VEGF₁₆₅ (R&D systems, USA) and 1 mg/mL collagen (type I; Sigma-Aldrich, USA) and seeded on top of a collagen coat in a 96-well plate. After 1 h of polymerization at 37°C, 100 μL EGM2 with or without 25 ng/mL and 220 ng/mL angiopoietin-1 (Peprotech, UK) was pipetted on the spheroid-containing collagen gel, and the gel was incubated at 37°C for 2 days. To quantify sprouting angiogenesis, the cumulative length of all capillary-like sprouts originating from the central plain of an individual spheroid was measured at 10× magnification using ImageJ software (plugin NeuronJ). For mosaic sprouting assays, HUVECs were labeled with CellTracker Green or Red (Invitrogen). At least 10 spheroids per experimental group and experiment were analyzed.

Ex vivo retinal explant sprouting

The isolation and sprouting was performed according to Sawamiphak and colleagues (13). Animal experiments were complied with all relevant guidelines and regulations that occur in the Netherlands. In brief, newborn C57BL/6 pups of the age of 4–5 days were anesthetized on ice and sacrificed. Extract the lens and incubate the retinas at 35°C for 4 h and fix and analyze using microscopy.

Tube formation assay

In vitro tube formation was determined with three-dimensional fibrin matrices and HMVECs, as described before (14). Fibrin matrices were prepared by the addition of thrombin (0.05 U/mL) to a 2 mg/mL fibrinogen solution in M199 medium, and 100 μL were added to the wells of a 96-well plate. After 2 h polymerization, thrombin was inactivated by incubating the matrices 2 h with M199 supplemented with 10% HSi and 10% NBCSi. HMVECs were seeded in confluent density on the fibrin matrices. After 24 h, and subsequently at 48 h intervals, the HMVECs were stimulated with M199, 10% HSi, 10% NBCSi, 10 ng/mL tumor necrosis factor-α (TNFα, Sigma, USA) and 25 ng/mL vascular endothelial growth factor (VEGF, Invitrogen, USA) or 10 ng/mL fibroblast growth factor-2 (FGF-2, Preprotech, UK). The formation of tube-like structures from HMVECs into the fibrin matrices was analyzed by phase contrast microscopy and Optimas image analysis software. Tube formation of three HMVEC donors was determined in quadruplicate wells.

Sprouting of intersegmental vessels in zebrafish

Zebrafish were maintained as previously described (15). All experiments were performed in accordance with federal guidelines and were approved by the Kantonales Veterinäramt of Kanton Basel-Stadt. Adult zebrafish (Danio rerio) of transgenic strain Tg (kdrl:EGFP^s843) were kept using standard procedures (8) and were cultured at 28°C in E3-medium supplemented with 0.2 mM phenylthiourea (Sigma-Aldrich) to prevent pigmentation. Embryos were obtained by natural spawning of adult zebrafish in mesh-bottomed breeding tanks and kept under standard conditions (8). To inhibit Trio or Rac1, we treated embryos for 6 h with 100 μM ITX3 (for Trio; Sigma-Aldrich) or 94 or 300 μM NSC23766 (for Rac1; Sigma-Aldrich) from 21 to 22 h post-fertilization (hpf) in E3 medium, then fixed with 3.7% PFA and imaged with a Leica SP5 confocal microscope with 10x (NA 0.3) and 20x (NA 0.7) air objectives. In addition, controls were fixed at the 21–22 hpf time point to analyze the starting situation. To combine data from separate experiments, the raw fluorescence intensity values of the samples within the experiment were normalized with the mean value of the control group in the same experiment. Statistical analyses were performed using SPSS and Microsoft Excel.

Statistical analysis

Statistical comparisons between experimental groups were performed using Student’s t-test. A two-tailed P-value of ≤0.05 was significant. Unless stated otherwise, a representative experiment out of at least three experiments is shown.

Results

Trio activity is required for sprouting angiogenesis

The small GTPases Rac1 and RhoG are involved in angiogenesis, as is shown by us (Fig. 1A) and others (16). Silencing either Rac1 or RhoG resulted in a significant reduction in sprout outgrowth (Fig. 1A). The RhoGEF Trio can activate both GTPases through its N-terminal GEF domain (17, 18, 19). To investigate the role of Trio in angiogenesis, we used a 3D in vitro angiogenesis assay based on the sprouting of ECs from collagen-embedded aggregates. EC spheroids were seeded in a collagen gel and stimulated with VEGF₁₆₅ and angiopoietin-1. After 48 h, control ECs originating from the embedded spheroids invaded the gel to form sprout-like structures (Fig. 1B). The total length of sprouts from a spheroid of Trio-depleted cells was significantly reduced to 759 μm as compared to 1,395 μm in controls (Fig. 1B). Using mosaic sprouting assays in which Trio shRNA-transfected ECs were labeled with a green fluorescent dye and mixed with red-labeled control shRNA-transfected cells, we found that the sprouts consisted mainly of control ECs. Trio-deficient ECs remained in the cell aggregate and did not participate in sprout formation, indicating that Trio defects are not restricted to the tip cell only (Fig. 1C). In addition, we used an alternative 3D in vitro angiogenesis assay, where human microvascular endothelial cells (HMVECs) were seeded on top of a fibrin matrix, and tube formation was induced by simultaneous stimulation with VEGF and TNFα. Trio-deficient HMVECs failed to adhere to the fibrin matrix (Fig. 1D). Comparable results were obtained after stimulation of HMVECs with the combination of fibroblast growth factor-2 (FGF-2) and TNFα (Fig. 1D). Using Matrigel assay to study tube formation, we found that Trio-depleted cells showed disorganized formation of tubular structures, as well as more single cells (Fig. 1E). It is important to note that these reductionist assays do not recapitulate all steps of sprouting angiogenesis but are used as a screening model to dissect the involvement of Trio. From these assays, we conclude that loss of Trio expression perturbs angiogenesis in vitro.

Figure 1.

Trio is involved in sprouting angiogenesis. (A) Capillary sprouting from collagen gel-embedded HUVECs transduced with shRNA ctrl or Rac1 or RhoG was treated with VEGF. The red lines show the circumference of the angiogenic sprouts. Bar: 100 μm. Western blotting shows efficient silencing of Rac1 and RhoG. The bar graph shows the total length of sprouts originating from one spheroid. Rac1- or RhoG-deficient ECs show reduced sprout formation. Data are mean ± SEM of at least three independent experiments. *P < 0.05. (B) Capillary sprouting from collagen gel-embedded ctrl or Trio shRNA-transduced spheroids. The bar graph shows the total length of sprouts originating from one spheroid. Data are mean ± SEM of at least three independent experiments. Bar: 100 μm. (C) HUVECs transduced with control or Trio shRNA were labeled with green (shCtrl) or red (shTrio) fluorescent dyes, mixed, and spheroids were generated and embedded in collagen. Bar: 100 μm. (D) In vitro angiogenesis of ctrl or Trio shRNA-transduced HMVECs on fibrin matrices. Tube formation upon stimulation with VEGF and TNFα or FGF-2 and TNFα was reduced in Trio-depleted HMVECs, since fewer cells adhered to the fibrin matrix. (E) 12 h tube formation assay on Matrigel shows the formation of capillaries in vitro after VEGF stimulation. Impaired capillaries and more single cells are observed in Trio-deficient ECs. Bar: 1 mm.

As the RhoGEF Trio is well known for its capacity to activate Rac1 and RhoG (17, 18, 19), we first confirmed the involvement of Rac1 and RhoG activity for the formation of capillary sprouts using the known Rac family inhibitor EHT1864 (20, 21). The results showed that EHT1864 blocked Trio-induced activation of both Rac1 and RhoG (Fig. 2A). In line with previous reports (5, 7), EHT1864 significantly inhibited the outgrowth of sprouts (Fig. 2B). We next assessed whether blocking Trio activity by the specific chemical compound inhibitor ITX3 would perturb sprout formation (22). Like Trio silencing, ITX3 significantly reduced sprout formation (Fig. 2B). Together, these data indicate that Trio activity is required for angiogenic sprout formation.

Figure 2.

Trio inhibition impaired angiogenic sprouting. (A) The inhibitor EHT1864 blocked TrioD1-induced Rac1 and RhoG activation. Left panels show classical pull-down assays with GST-ELMO1 for active RhoG and GST-Pak-binding domain for active Rac1. Right panels show input. (B) Outgrowth of sprouts from HUVEC spheroids treated with DMSO, the Rac1 inhibitor EHT1864 or the TrioGEF1 inhibitor ITX3. The bar graph shows the total length of sprouts originating from one spheroid. Data are mean ± SEM of three independent experiments. Bar: 100 μm. (C) Zebrafish embryos at 21–22 h post-fertilization were treated as indicated, and intersegmental vessel length from the dorsal edge of dorsal aorta to the most distal tip of sprout was measured. Trio or Rac1 inhibition significantly impaired sprout length in vivo. Number of fish: pre-treatment group n = 11, DMSO n = 39, ITX3 n = 37, NSC 94 μM n = 30 and NSC 300 μM n = 35. In total, 883 sprouts were analyzed from 152 embryos. *P = 0.012; ***P < 0.001. Right panel shows a representative image of GFP-positive vessels after fixation. (D) Trio and Rac1 inhibition significantly reduced sprout speed. **P = 0.008; ***P < 0.001. (E) Isolated retinas from Tie2-GFP KI-mice were treated with DMSO or ITX3, and sprouting was induced by VEGF. Trio blockage reduced sprout length in this retinal explant model. ***P < 0.001. (F) Zooms (area indicated by the square) show PECAM-1-positive vasculature in the forelimbs of Trio^−/− and Trio^+/− embryos at E13.5. Transverse sections are stained for PECAM-1 (green), F-actin (red) and nuclei (blue). Upper bar graph, mean fluorescent intensity of PECAM-1 staining per μm². Lower bar graph, the number of PECAM-1-positive vessels. Data are mean ± SEM of three animals per group. *P < 0.01.

To test the physiological relevance of Trio for angiogenic sprout formation in vivo, we used GFP-transgenic zebrafish to measure angiogenic sprouts from the developing segmented vessels. Inhibiting Trio with ITX3 or perturbing Rac1 signaling using NSC23766 (23) significantly reduced the average spout length (Fig. 2C) as well as the average speed of the developing sprout (Fig. 2D). To further establish the crucial role of Trio in angiogenesis, we used retinal explant cultures from Tie2-GFP mice that were ex vivo treated with ITX3. We measured a strong reduction in the length of the developing sprouts in these explants (Fig. 2E). Moreover, analyzing the vasculature of Trio knockout animals at E13.5 revealed a significant reduction in the number of PECAM-1/CD31-positive vessels in the forelimb of the Trio homozygous knockout mice compared to their heterozygote littermates (Fig. 2F). The hindlimb may have been an interesting area to investigate for vessel development; however, we did not have access to the hindlimb slides of these animals. Nevertheless, these data further support a role for Trio in angiogenesis.

The N-terminal part of Trio is required for sprouting

Previous research by us and others showed that transfecting cells with only the TrioGEF1 domain resulted in the activation of endogenous Rac1 and RhoG (17, 18). Expression of the N-terminus of Trio, including the GEF1 domain, also resulted in the activation of endogenous Rac1 and RhoG (24). An overview of the different Trio constructs is shown in Fig. 3A. To show that angiogenic sprouting was regulated by the first GEF1 domain of Trio (TrioD1), we set up an experiment in which TrioGEF1 activity was rescued in Trio-deficient ECs by expressing the TrioGEF1 domain only. Trio was silenced using RFP-tagged shRNA constructs as previously described (9). Based on the fluorescent tags and using flow cytometry cell sorting, specific Trio-deficient and rescued ECs were isolated and analyzed for sprouting. As expected, the sprouting of Trio-deficient ECs (RFP-positive) is not restored after the expression of GFP alone (Fig. 3B). Surprisingly, the rescue of only the first GEF domain (TrioD1) resulted in only a few limited numbers of sprouts (Fig. 3B). Interestingly, rescue of Trio expression with TrioN, comprising the Sec14 domain, spectrin repeats and TrioGEF1 domain, showed a full rescue of the Trio-defect in sprout formation (Fig. 3B). Quantification of the outgrowth of sprouts underscored that TrioN significantly rescued the Trio-deficiency phenotype in sprouting (Fig. 3B), suggesting that Trio required its full N-terminus to promote angiogenic sprouts.

Figure 3.

N-terminus of Trio drives angiogenic sprouting. (A) Full-length (FL) Trio comprises a Sec14 domain, spectrin repeats, the GEF1 domain (TrioD1, able to activate Rac1 and RhoG), the GEF2 domain (able to activate RhoA), an Ig-like domain and a serine/threonine (S/T) kinase domain. GFP-TrioD1 consists of the GEF1 domain only, whereas GFP-TrioN consists of the N-terminus of Trio, including the GEF1 domain. (B) HUVECs transduced with shRNA ctrl or shRNA expressing an RFP-tag targeting Trio were subsequently transduced with GFP, GFP-TrioD1 or GFP-TrioN, sorted for both RFP and GFP using flow cytometry to ensure the purity of the knockdown and rescue cell population, embedded in collagen and treated with VEGF. The red lines show the circumference of the angiogenic sprouts. Bar: 100 μm. The bar graph on the right shows the total length of sprouts originating from one spheroid. Sprout formation of Trio-deficient ECs is rescued with TrioN. Data are mean ± SEM of at least three independent experiments. *P < 0.05. ***P < 0.001.

VE-cadherin-induced sprouting requires parts of the intracellular tail and Trio activity

The regulation of VE-cadherin-based cell–cell junctions, i.e. the assembly and disassembly of the individual ECs, is pivotal for angiogenesis (7, 8, 25). We previously found Trio to interact with VE-cadherin (9). To interfere with this interaction, we used a VE-cadherin mutant construct that lacks the intracellular binding site for β-catenin but still encodes the p120-catenin-binding site (Fig. 4A) (26). Immunoprecipitation studies using GFP antibodies showed that VE-cadherin-wt-GFP was associated with Myc-Trio-wt, but the cytoplasmic deletion mutant of VE-cadherin did not (Fig. 4B). Interestingly, the expression of the VE-cadherin-ΔC mutant in VE-cadherin-null ECs did not rescue sprouting, whereas VE-cadherin-wt did (Fig. 4C). These data show that ECs need the full-length VE-cadherin to efficiently sprout.

Figure 4.

Trio–VE-cadherin interaction is required for sprout formation. (A) Schematic overview of VE-cadherin constructs used tagged to GFP. VE-wt shows full-length VE-cadherin, and VE-ΔC shows deletion of β-catenin-binding site but including the p120-catenin site. (B) HUVECs were transfected with Myc-tagged full-length wt-Trio (myc-TrioFL) and VE-cadherin-GFP (VE-wt) or VE-cadherin-GFP lacking the intracellular β-catenin-binding site (VE-ΔC). IP for Myc showed the interaction between Trio with VE-cadherin-wt and reduced interaction with VE-ΔC. (B) VE-cadherin-null ECs were transfected with VE-cadherin-GFP (VE-wt) or VE-cadherin-GFP lacking the intracellular β-catenin-binding site (VE-ΔC) and showed that wt but not ΔC mutant rescued sprouting. Quantification is shown on the right. Experiment is carried out three times independently from each other. Data are mean ± SEM. *P < 0.05.

Discussion

In this study, we report that the RhoGEF Trio is involved in the induction of angiogenic sprouting. In combination with our finding that Trio can associate with VE-cadherin (9), it is tempting to speculate that Trio/VE-cadherin interaction is needed for angiogenic sprouting. We show here that the RhoGEF Trio promotes in vitro formation of endothelial sprouts through its N-terminal GEF1 domain. Trio is a unique GEF as it has two separate GEF domains. The N-terminal GEF1 domain can activate Rac1 and RhoG, whereas the C-terminal GEF2 domain can exchange GDP for GTP on RhoA (17, 18, 27). Using inhibitors and RNAi silencing techniques for Rac1 or RhoG, we underscore data by others that show the involvement of active Rac1 or RhoG in efficient sprouting angiogenesis (28, 29, 30). In addition, others have reported that Rac1 is required for the generation of small vessels by controlling cell migration, adhesion and tubulogenesis (11, 31). From these data, it is tempting to link a role for Trio directly to Rac1 and RhoG activation.

In this study, we have used the inhibitor EHT1864. This inhibitor does not only block Rac1 activity but also blocks RhoG activity. This is also true for the Trio ITX3 inhibitor: ITX3 blocks TrioGEF1-induced Rac1 and RhoG activation independently from each other (18, 32). We and others showed that RhoG silencing as well as Rac1 depletion also perturbs sprouting (16, 30). These data indicate that there may be a redundant role for Trio in Rac1- and/or RhoG-mediated angiogenic sprouting. To fully assess whether the N-terminal part of Trio, including the GEF1 domain, is needed for sprouting, one would need a GEF1-catalytic dead TrioN mutant. These experiments may provide definitive answers as to whether TrioN, including the GEF1 domain, is required for efficient sprouting.

Our data furthermore indicate that the GEF1 domain of Trio alone is not sufficient to promote sprouting. Surprisingly, the N-terminus of Trio is also required. The N-terminus of Trio encodes, next to the GEF1 domain, a Sec14 domain, and several spectrin repeats (33). We previously showed that the spectrin domains 5–6 bind directly to the intracellular tail of VE-cadherin (9). This is in line with our biochemical studies showing that TrioN interacts with VE-cadherin.

In vivo experiments in zebrafish showed that endothelial sprouting required stable VE-cadherin-based connections (8, 34). On the other hand, Abraham and colleagues showed that abrogating VE-cadherin function in angiogenesis assay and in zebrafish promotes sprouting (7). These data show the complexity of the regulation of VE-cadherin for EC sprouting. Indeed, blocking antibodies that inhibit VE-cadherin function showed concentration-dependent effects on sprouting (7). Interestingly, Sauteur and co-workers showed that angiogenic sprouting required VE-cadherin to organize junctional and cortical actin cytoskeleton remodeling (8). Furthermore, junctional remodeling required the β-catenin-binding site in VE-cadherin (15). These data may implicate a role for Trio in regulating sprouting by mediating junctional actin cytoskeleton remodeling by binding to VE-cadherin. As our in vitro tube formation assays show a role for Trio in cell adhesion and we previously found Trio to promote endothelial enlargement (35), a unique role for Trio independently of VE-cadherin cannot be excluded.

Our data obtained with the Trio-deficient embryos show that Trio is required at least for the development of a full vascular network. According to O’Brien and co-workers, the disorganized vasculature phenotype in the knockout mice appears at a relatively late development stage as the embryos die between E15.5 and birth (36). At this stage of development, the major vascular network is already developed, indicating that Trio does not control the development of major vessels. However, Trio may be essential to drive angiogenic sprouting in distal parts of the growing vascular tree.

In conclusion, our data suggest that Trio is involved in angiogenesis. Our experiments point to a model where Trio requires VE-cadherin to locally remodel the actin cytoskeleton at cell–cell contacts to promote angiogenic sprouting.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the work reported.

Funding

JK was supported by a Dutch Heart Foundation Dekker fellowship (2005T3901) and a Senior Scientist Dekker grant (03-004-2021-T045) and was funded by the European Union (ERC, ENDOMET-STEER, 101076407). IT was supported by a KiKa grant (No. 374).

Author contribution statement

JK, IT, SAvdP, PK, HGB and JDvB conceived and designed the experiments. JK, IT, IP, MH and EMW performed the experiments. JK, IT, IP and EMW analyzed the data. JK, IT and JDvB wrote the paper.