The FASEB Journal article fj.14-269258. Published online May 13, 2015.
The FASEB Journal • Research Communication
Increased endothelial cell permeability in endoglin-deficient cells Mirjana Jerkic*,†,‡,1 and Michelle Letarte*,† *Molecular Structure & Function Program, The Hospital for Sick Children, Toronto, Ontario, Canada; and †Department of Immunology and ‡Keenan Research Centre for Biomedical Science, Anesthesia Research, St. Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada ABSTRACT Endoglin (ENG) is a TGF-b superfamily coreceptor essential for vascular endothelium integrity. ENG mutations lead to a vascular dysplasia associated with frequent hemorrhages in multiple organs, whereas ENG null mouse embryos die at midgestation with impaired heart development and leaky vasculature. ENG interacts with several proteins involved in cell adhesion, and we postulated that it regulates vascular permeability. The current study assessed the permeability of ENG homozygous null (Eng2/2), heterozygous (Eng+/2), and normal (Eng+/+) mouse embryonic endothelial cell (EC) lines. Permeability, measured by passage of fluorescent dextran through EC monolayers, was increased 2.9- and 1.7-fold for Eng2/2 and Eng+/2 ECs, respectively, compared to control ECs and was not increased by TGF-b1 or VEGF. Prolonged starvation increased Eng2/2 EC permeability by 3.7-fold with no effect on control ECs; neutrophils transmigrated faster through Eng2/2 than Eng+/+ monolayers. Using a pull-down assay, we demonstrate that Ras homolog gene family (Rho) A is constitutively active in Eng2/2 and Eng+/2 ECs. We show that the endothelial barrier destabilizing factor thrombospondin-1 and its receptor-like protein tyrosine phosphatase are increased, whereas stabilizing factors VEGF receptor 2, vascular endothelial-cadherin, p21-activated kinase, and Ras-related C3 botulinum toxin substrate 2 are decreased in Eng2/2 cells. Our findings indicate that ENG deficiency leads to EC hyperpermeability through constitutive activation of RhoA and destabilization of endothelial barrier function.—Jerkic, M., Letarte, M. Increased endothelial cell permeability in endoglin-deficient cells. FASEB J. 29, 000–000 (2015). www.fasebj.org
Key Words: RhoA • TGF-b1 • VEGF • TSP-1 ENDOGLIN (ENG; CD105) IS A 180 kDa homodimeric transmembrane glycoprotein expressed mainly in endothelial cells (ECs) (1) and is known to modulate cellular responses to ligands of the TGF-b superfamily (2). ENG is essential for proper vascular development and function and has Abbreviations: ALK, activin receptor-like kinase; AVM, arteriovenous malformation; BBB, blood-brain barrier; BMP, bone morphogenetic protein; BSA, bovine serum albumin; CD148, receptor-like protein tyrosine phosphatase; E, embryonic day; EC, endothelial cell; ENG, endoglin; fMLP, formyl-methionineleucine-phenylalanine; GEF, guanine nucleotide exchange factor; (continued on next page)
0892-6638/15/0029-0001 © FASEB
a crucial role in maintaining EC homeostasis (3). ENG homozygous null embryos die at midgestation of impaired heart development and vascular abnormalities, including hemorrhages and a leaky vasculature (4). Mutations in the ENG gene lead to hereditary hemorrhagic telangiectasia (HHT) type 1 (HHT1), inherited as an autosomaldominant vascular dysplasia that affects 1 in 10,000 people worldwide and is characterized by frequent and abundant nosebleeds due to mucocutaneous telangiectasia and by arteriovenous malformations (AVMs) in multiple organs, potentially leading to severe hemorrhages and strokes (5). Haploinsufficiency is the underlying cause of HHT, indicating that reduced levels of ENG predispose to endothelial dysfunction and AVMs. ENG is a coreceptor for several members of the TGF-b superfamily (2). It was first shown to bind TGF-b1 and TGFb3 in association with the serine threonine kinase receptor type 2 (TGFBR2), leading to activation of type I receptors [activin receptor-like kinase (ALK)5 and ALK1] to initiate signaling. ALK1 (or ACVRL1) is a specialized receptor of ECs and like ENG, plays a critical role in angiogenesis and vasomotor function and is mutated in HHT2 (6). Previous studies reported that ENG could interact with several proteins critically involved in cell adhesion, migration, and vascular permeability, including integrins (7, 8), zyxin (9), and vascular endothelial-cadherin (VE-cadherin) (10). ENG interacts with a5b1 integrin and mediates the crosstalk between fibronectin/a5b1 and TGF-b pathways in ECs via the internalization of the a5b1 integrin/ENG complex (8). ENG-dependent cellular adhesion and transendothelial migration (TEM) were then ascribed to its role as a counterreceptor for the leukocyte a5b1 integrin (8). Analysis of integrin-rich sites of focal adhesion plaques showed that ENG interacted via its cytoplasmic domain with the LIM (zinc-binding domain present in Lin11, Isl-1, Mec-3) domain protein zyxin (9). Furthermore, VE-cadherin concentrated at cell-to-cell adherens junctions, interacts with many components of the TGF-b receptor complex, including ENG, TGFBR2, and ALK1 (10). Our previous study, using a high-throughput luminescencebased mammalian interactome assay (LUMIER), documented novel ENG, TGFBR2, and ALK1 interactions with 1 Correspondence: Keenan Research Centre for Biomedical Science, Anesthesia Research, St. Michael’s Hospital, 209 Victoria Street, Lab 653, University of Toronto, Toronto, ON, Canada M5B 1T8. E-mail:
[email protected] doi: 10.1096/fj.14-269258
1
several proteins including p21-activated kinase (PAK-1), GTP-binding proteins, and Ras-related C3 botulinum toxin substrate 2 (Rac-2) involved in endothelial barrier maintenance, cytoskeletal remodeling, and cell migration (11). Thus, ENG, directly, or as a component of the TGF-b receptor complex may regulate EC integrity; its absence could result in vascular hyperpermeability and contribute to the fetal lethality associated with the homozygous null genotype (9, 12). Three types of vascular permeability can be distinguished in vivo: basal vascular permeability of normal tissues; acute vascular hyperpermeability in response to a single, short exposure to VEGF-A or other permeabilizing factors; and chronic vascular hyperpermeability characteristic of pathologic angiogenesis associated with tumors, healing wounds, and chronic inflammation (13). Vascular lesions in HHT are also the consequence of pathologic angiogenesis, and we postulated that chronic vascular permeability would be affected. In vitro assays using confluent monolayers of ECs do provide a model for measuring chronic vascular hyperpermeability (13). In view of the lethal phenotype associated with the complete absence of ENG, we elected to assess the permeability of monolayers of EC lines derived from ENG homozygous null, heterozygous, and control mouse embryos and study candidate molecules implicated in cell-cell junctions and actin cytoskeleton organization. This should allow us to gain some understanding of the mechanisms whereby ENG regulates endothelial barrier maintenance and vascular permeability. MATERIALS AND METHODS
Canada) through the EC monolayer, as previously described (15). Briefly, 1 3 105 Eng+/+, Eng+/2, and Eng2/2 embryonic ECs were plated onto Transwell fibronectin-coated (10 mg/ml), 3 mm pore polyethylene membrane inserts in 24-well plates (6.5 mm insert size; Corning, Tewksbury, MA, USA) and left for 3 days to form mature monolayers. Confluence of the cell layer was confirmed in every experiment by testing the extent of bovine serum albumin (BSA) penetration of the monolayer using colorless media. BSA (1.4 mg/ml) was added to the medium in the upper chamber, incubated at 37°C for 45 min, and protein concentration was measured in media recovered from the Transwell plate lower chamber, using the Bio-Rad protein assay (Bio-Rad Laboratories, Mississauga, ON, Canada). The EC monolayer was considered confluent if ˂1% of the BSA added was detected in the lower chamber in comparison to values of 20–25% when confluence of the EC monolayer was not reached. Comparable results were obtained when either FITC-dextran 4000 or 40,000 was used under control conditions or after different lengths of cell starvation (4, 14, and 21 h). Permeability experiments were started by washing the upper and lower compartments with N-2hydroxyethylpiperazine-1-ethane sulfonic acid (HEPES)buffered salt solution, followed by adding FITC-dextran (2 mg/ml) in HEPES-buffered salt solution to the upper compartment. After a 2-h incubation at 37°C, the bottom well media were transferred into a 96-well black plate, and fluorescence was measured using a spectrofluorimeter (excitation, 490 nm; emission, 520 nm). A standard curve with FITC-dextran (0–2 mg/ml) was also prepared in each experiment and read at the same time. Data analysis was done using SoftMax Pro 5.3 software (Molecular Devices, Sunnyvale, CA, USA). Stimulation experiments (with VEGF and TGF-b1) were done using the 40,000 FITC-dextran. ECs were starved for 3.5 h and incubated in the presence of VEGF (50 ng/ml) for 30 min or starved for 3 h and incubated with TGF-b1 (10 ng/ml) for 11 h. All experiments were done in duplicate, and results are expressed as permeability relative to that of Eng+/+ control cells.
ECs Neutrophil transmigration assay Mouse ECs were derived from the yolk sac of embryonic day 8.5 (E8.5) ENG homozygous null (Eng2/2), heterozygous null (Eng+/2), and littermate control Eng+/+ embryos (on a C57Bl/6J background), as previously described (14). These cells were maintained in MCDB131 medium with 15% fetal bovine serum and 15 mg/ml endothelial mitogen (Biomedical Technologies, Villalbain, Spain), 100 IU/ml penicillin-streptomycin (Invitrogen Canada, Burlington, ON, Canada), and 1% glutamine (Invitrogen Canada). Permeability assay Permeability was assessed by the passage of FITC-conjugated dextran (Mr, 4000 and 40,000; Sigma-Aldrich, Oakville, ON,
(continued from previous page) GST, glutathione S-transferase; GTPase, guanosine triphosphatase; HEPES, N-2-hydroxyethylpiperazine-1-ethane sulfonic acid; HHT, hereditary hemorrhagic telangiectasia; IP, immunoprecipitation; LIM, zinc-binding domain present in Lin-11, Isl-1, Mec-3; LUMIER, luminescence-based mammalian interactome assay; MLC, myosin light chain; PAH, pulmonary arterial hypertension; PAK-1, p21-activated kinase; Rac-2, Ras-related C3 botulinum toxin substrate 2; Rho, Ras homolog gene family; ROS, reactive oxygen species; TEM, transendothelial migration; TGFBR2, TGF-b receptor type II; TSP-1, thrombospondin-1; VEcadherin, vascular endothelial-cadherin; VEGFR2, VEGF receptor 2; WB, Western blot
2
Vol. 29
September 2015
ECs (Eng+/+ or Eng2/2) were seeded on fibronectin-coated Transwell inserts (3 mm pores) until confluent monolayers were formed, as described above. Cells were then starved for 3 h and activated or not with TNF-a (20 ng/ml) for 4 hs. Normal C57Bl/ 6J mouse bone marrow neutrophils were isolated using Percoll gradients and labeled with 1.5 mM Calcein acetoxymethyl ester (Abcam Incorporated, Toronto, ON, Canada). Chemoattractant [1 mM formyl-methionine-leucine-phenylalanine (fMLP)] was added or not into the Transwell plate bottom wells, containing neutrophil media (a-minimum essential medium with 10 mM HEPES and 2.5% fetal calf serum). Cell culture inserts containing confluent monolayers of Eng+/+ or Eng2/2 ECs were transferred on top of the well, and 3 3 105 neutrophils were added into the insert and left for 2 h. A standard curve with known neutrophil numbers was also prepared. The bottom well medium containing migratory cells was lysed and fluorescence read at 494/517 nm. All experiments were done in duplicate, and results are expressed as transmigration relative to that of Eng+/+ cells.
Ras homolog gene family A and ENG coimmunoprecipitation experiments Ras homolog gene family (Rho) A-Flag (16) and ENG (2) expression constructs in the pCMV5 vector were generated and checked by restriction enzyme digestion and sequencing as described (2, 16). Human embryonic kidney-293T cells were seeded in 6-well plates at 2.5 3 105 cells per well and transfected with 1 mg
The FASEB Journal x www.fasebj.org
JERKIC AND LETARTE
total plasmid DNA (RhoA-Flag, with or without ENG) using FuGENE 6 (Hoffmann-La Roche Limited, Mississauga, ON, Canada). The preparation of cell lysates, immunoprecipitation (IP), and Western blot (WB) were carried out as described (11). The P4A4 mAb to ENG (1 mg) was used for IP and the mouse antiFlag (M2) antibody (Sigma-Aldrich, Oakville, ON, Canada) for WB.
Rho activity assay Eng+/+, Eng+/2, and Eng2/2 mouse embryonic ECs were grown on 100 mm plates to 70–75% confluence. Cells were starved for 3.5 h and treated with VEGF (50 ng/ml) for 15 min or with TGF-b1 (10 ng/ml) for 11 h (after 3-h starvation) because these conditions were found optimal. ECs were lysed as described above. After protein estimation, 700–800 mg total lysate protein was used for IP, whereas 20 mg was reserved for WB. Rhotekin-Rho-binding domain protein glutathione S-transferase (GST) beads (Cytoskeleton, Denver, CO, USA) were added to the precleared lysate according to the manufacturer’s instructions. Briefly, beads (60 mg) were added to each reaction tube and incubated for 1 h at 4°C with gentle rotation. Beads were pelleted by centrifugation and washed 3 times with 1 ml TNE buffer (50 mM Tris, 100 mM NaCl, and 1 mM EDTA) plus 0.1% Triton X-100. GST beads were eluted in NuPage LDS Sample Buffer (Invitrogen Canada) at 95°C for 5 min and analyzed by WB using a RhoA-specific mAb (mouse anti-RhoA, clone 26C4, 1:1000; Santa Cruz Biotechnology, Dallas, TX, USA).
WB procedure EC lysates were obtained as described above, and equal protein amounts were fractionated on 4–12% gradient NuPage gels (Invitrogen Canada) and transferred to a PVDF membrane (Immobilon-P; EMD Millipore, Billerica, MA, USA). After blocking with 5% milk in Tris-buffered saline with Tween, the blot was incubated with primary antibody for 2 h or overnight, followed by a secondary antibody conjugated with horseradish peroxidase for 1 h. Signals were detected using an ECL-Plus kit (Amersham Biosciences, GE Healthcare, Piscataway, NJ, USA). The following antibodies were used: 1:1000 mouse anti-TSP-1 (thrombospondin-1; Thermo Fisher Scientific, Fremont, CA, USA); 1:1000 goat anti-CD148 (receptor-like protein tyrosine phosphatase; R&D Systems, Minneapolis, MN, USA); 1:500 rat anti-ENG (clone MJ7/18; Southern Biotech, Birmingham, AL, USA); 1:500 mouse anti-VEGFR2 (VEGF receptor 2; R&D Systems); 1:10,000 G6-31 mAb (kindly provided by Genentech Inc., South San Francisco, CA, USA) and with high affinity and specificity for murine and human VEGF-A (17); 1:1000 mouse antiVE-cadherin (BioLegend, San Diego, CA, USA); 1:1000 rabbit anti PAK-1 (Santa Cruz Biotechnology); 1:1000 rabbit anti-RAC-2 (Santa Cruz Biotechnology); and 1:10,000 mouse anti-b-actin (Sigma-Aldrich, St. Louis, MO, USA). Appropriate secondary antibodies conjugated with horseradish peroxidase (1:10,000 dilution) and the ECL reagent (PerkinElmer, Shelton, CT, USA) were used for detection. Band intensities were quantified, expressed relative to that of b-actin (or total RhoA), and normalized to values corresponding to the mean of Eng+/+ cell control samples, for comparison between gels.
RESULTS ENG-deficient (homozygous and heterozygous null) EC monolayers are more permeable than control cells and constitutively activated We investigated the ability of ENG-deficient mouse embryonic ECs to allow FITC-dextran passage, including the Eng2/2 ECs associated with vascular leakage and lethality in vivo, and the Eng+/2 ECs, with predisposition to manifestations of the HHT vascular dysplasia. Figure 1A illustrates that more FITC-dextran passed through ENG-deficient than control EC monolayers; values were very similar for periods of 4 and 14 h. The measured permeability expressed relative to that of control EC monolayers was on average 2.9-fold for Eng2/2 cells and 1.7-fold for Eng+/2 cells. The relative permeability was significantly higher in Eng2/2 than Eng+/2 cells. A normal EC monolayer can be activated by several factors. We therefore tested the response of ENG-deficient cells to the most potent permeability factor, VEGF, and to TGFb1, whose effects are modulated by ENG and that also regulates permeability. Figure 1B demonstrates that normal ECs show a 1.9-fold increase in relative permeability in response to VEGF, whereas Eng+/2 and Eng2/2 ECs show no effect of VEGF. Control ECs were stimulated 1.7-fold by TGF-b1, whereas Eng+/2 and Eng2/2 ECs were not affected (Fig. 1C). Thus, ENG-deficient EC monolayers are constitutively activated in terms of permeability and leakier than control ECs. Higher permeability after long starvation stress and increased TEM in Eng2/2 EC monolayers The response of Eng2/2 ECs to increasing starvation periods was tested to assess the stability of the confluent monolayers relative to that of control cells. Figure 2A reveals that Eng2/2 cells are less able to sustain nutrient deprivation and that 21 h under such conditions lead to a 3.7-fold increase in relative permeability. Subsequent experiments were therefore done within 14 h of starvation, a time point that was no different from the 4-h starvation. More permeable EC monolayers are generally associated with increased cellular transmigration. Figure 2B shows that normal mouse bone marrow neutrophils transmigrated 1.3-fold faster through Eng2/2 than control EC monolayers, in the absence of stimulus (TNF-a) and chemoattractant (fMLP). Furthermore, whereas transmigration of neutrophils through the normal monolayer was increased 1.4-fold in the presence of TNF-a and fMLP, it was not significantly increased across Eng2/2 EC monolayers. Similar values were obtained for Eng2/2 EC monolayers when fMLP was used alone or together with TNF-a (data not shown), suggesting that ENG-deficient EC monolayers are in fact constitutively activated.
Statistical analysis
RhoA can associate with ENG and is constitutively active in ENG-deficient cells
Comparisons were performed by 1-way ANOVA, and significant overall differences were evaluated post hoc using the Bonferroni procedure. Results are expressed as the means 6 SEM, with P , 0.05 representing significance.
An increase in vascular permeability due to inflammatory mediators, or factors such as VEGF and TGF-b1, is generally mediated by formation of intercellular gaps between ECs of postcapillary venules, via opening of adherens and
ENDOGLIN AND ENDOTHELIAL CELL PERMEABILITY
3
4
*†
Eng+/3
2
Eng-/-
*†
*
*
4 hours
14 hours
1
0
B
Eng+/+
4
Eng+/+
Relative permeability
Eng+/3
2
*
Eng-/-
A
*
1
C VEGF 4
C VEGF
C VEGF
*
*
B
*
*
1
C TGF-β1
C TGF-β1
C TGF-β1
Figure 1. Eng+/2 and Eng2/2 ECs show increased permeability and are unresponsive to VEGF and TGF-b1. A) EC permeability was assessed in Eng+/+, Eng+/2, and Eng2/2 EC monolayers using 40 kDa FITC-dextran after 4 and 14 h of cell starvation. Results are expressed relative to Eng+/+ cells. There were 2–4 experiments done in duplicate (N = 9 for Eng+/+, 6 for Eng+/2, and 11 for Eng2/2 groups). *P , 0.05 vs. Eng+/+ cells; † P , 0.05 vs. Eng+/2 cells. B) VEGF or (C) TGF-b1 treatment increased permeability in control (C) cells but had no effect on Eng+/2 and Eng2/2 cells. ECs were starved for 3.5 h and treated with either VEGF (50 ng/ml for 30 min; 2–4 experiments in duplicate; N = 10 for Eng+/+ and Eng2/2 groups, and 6 for the Eng+/2 group) or TGF-b1 (10 ng/ml for 11 h; 2 experiments in duplicate; N = 5–6). *P , 0.05 vs. Eng+/+ control cells.
tight junctions to allow paracellular fluid passage. RhoA is a small guanosine triphosphatase (GTPase) that regulates cell adhesion by reorganization of the junction-associated cortical actin cytoskeleton. RhoA activation from a guanosine diphosphate-bound to a GTP-bound state due to nucleotide exchange factors and GTPase-activating factors is associated with increased permeability (18). We first tested whether an interaction between ENG and RhoA could be detected using transfectants overexpressing September 2015
4 hours
14 hours
21 hours
Eng+/+
2.0
Eng-/-
1
Vol. 29
2
*
0
4
3
0
Eng-/-
Relative transmigration
Relative permeability
2
*
Eng+/+
Eng+/+ Eng+/-
3
4
Eng-/-
0
C
constructs of these genes. Figure 3A shows that RhoA is immunoprecipitated with the mAb P4A4 to ENG, indicating that these 2 proteins form a molecular complex. The effect of ENG on RhoA activation state was determined using ENG-deficient ECs. Figure 3B demonstrates that Rho-GTP levels were 1.6- and 1.8-fold higher in Eng+/2 and Eng2/2 ECs, respectively, relative to control ECs. Figure 3B also shows that the potent permeability factor VEGF could trigger RhoA activation in control cells (1.6-fold) but not in Eng+/2 and Eng2/2 ECs. Similarly, Rho-GTP levels were increased 1.3-fold by TGFb1 in control but not Eng+/2 and Eng2/2 ECs, indicating constitutively active RhoA (Fig. 3C).
Relative permeability
Relative permeability
A
1.5
*
*
+ +
− −
1.0
0.5
0.0
TNFα fMLP
− −
+ +
Figure 2. Higher permeability after long starvation stress and increased TEM in Eng2/2 EC monolayers. A) EC permeability was assessed using 4 kDa FITC-dextran after 4, 14, and 21 h of cell starvation and is expressed relative to the value observed for Eng+/+ cells. Permeability was increased similarly at 4 and 14 h in Eng2/2 ECs compared to control ECs but was much higher in mutant vs. control cells at 21 h. There were 3 experiments done in duplicate (N = 5–6 for 4and 21-h points and 3–4 for 14-h point). *P , 0.05 vs. Eng+/+ cells. B) Freshly isolated normal bone marrow neutrophils from control Eng+/+ mice were labeled with Calcein and tested for transmigration through Eng2/2 and Eng+/+ EC monolayers, with and without stimulus (20 ng/ml TNF-a) and chemoattractant (1 mM fMLP). Migratory cells in the bottom wells were lysed, and their fluorescence was measured. The transmigration was higher through Eng2/2 monolayers, even in the absence of stimulus or chemoattractant. There were 3 experiments done in duplicate (N = 11–12 for Eng+/+ and 5–6 for Eng2/2 groups). *P , 0.05 vs. Eng+/+ cells without chemoattractant.
The FASEB Journal x www.fasebj.org
JERKIC AND LETARTE
A
WB: α-Flag
Lysate
Transfection:
IP: P4A4
RhoA-Flag
pCMV5
RhoA-Flag + Eng
RhoA
B C
C
C
Eng-/-
Eng+/+ V
V
C
C
V
V
Eng+/+ C
C
Eng-/-
T
T
C
C
T
T
RhoGTP Total RhoA Eng+/-
Eng+/+ C
V
C
C
Eng-/V
C
C
Eng+/+ V
T
C
Eng+/T
C
T
Eng-/T
C
T
T
RhoGTP Total RhoA
Eng+/+
Eng+/+
Eng+/-
Eng+/Eng-/-
2.0
*
* Rho-GTP/Total RhoA
Rho-GTP/Total RhoA
*
2.0
1.5
1.0
0.5
Eng-/-
*
*
C TGF-β1
C TGF-β1
1.5
* 1.0
0.5
0.0
0.0
C
VEGF
C VEGF
C
VEGF
C TGF-β1
Figure 3. RhoA is associated with Eng and constitutively active in Eng+/2 and Eng2/2 cells. A) Lysates of 293T cells overexpressing RhoA-Flag, pCMV5 control, or RhoA-Flag + ENG were analyzed directly or immunoprecipitated with P4A4 antibody to ENG. Western blotting with anti-Flag antibodies shows that RhoA is well expressed in the lysate and coimmunoprecipitates with ENG. WB from lysate and coimmunoprecipitation samples was done on the same gel, but intermediary bands were cut out because they were not relevant for this paper. B) Representative gels and graph show the results of RhoGST pull-down assays performed under basal conditions and after VEGF (V) activation, revealing increased Rho-GTP levels in Eng+/2 and Eng2/2 cells compared with control Eng+/+ cells. VEGF activation of RhoA is lost in ENG-deficient cells relative to Eng+/+ cells. Cells were stimulated with VEGF (50 ng/ml) or control (C) medium for 15 min, and RhoGST pull-down assays were performed. C) Representative gels and graph confirm increased Rho-GTP levels in Eng+/2 and Eng2/2 cells compared with control Eng+/+ cells. TGF-b1 (T) stimulation of RhoA does not occur in ENG-deficient cells but is observed in control cells. ECs were stimulated with TGF-b1 (10 ng/ml) for 11 h prior to the RhoGST pull-down assays. N = 3 for Eng+/2 and 5 for Eng+/+ and Eng2/2 per group for stimulation in both (B) and (C), and N = 5 for Eng+/2 and 10 for Eng+/+ and Eng2/2 per group for control samples. There were 2–3 experiments done. Minimal adjustment (color balance, contrast, and brightness) of representative WB images was equally applied to the entire gel. *P , 0.05 vs. Eng+/+ cells.
ENDOGLIN AND ENDOTHELIAL CELL PERMEABILITY
5
DISCUSSION In this study, we demonstrate that ENG-deficient mouse embryonic ECs are hyperpermeable and unresponsive to stimulation by VEGF and TGF-b1, factors that regulate vascular permeability (22). ENG null ECs allow increased leukocyte transmigration and respond to stress such as a 21-h starvation period by undergoing EC junction disassembly that may account for vascular leakage and midgestation lethality in Eng2/2 embryos. We show that RhoA interacts with ENG, is expressed at higher levels, and is constitutively 6
Vol. 29
September 2015
Eng+/+
Eng+/-
Eng-/-
kDa
TSP-1
180
β-Actin
42
*†
Eng+/+ Eng+/Eng-/-
6.0
Relative TSP-1 density
TSP-1 is an extracellular matrix protein produced following platelet activation under physiologic conditions. However, TSP-1 is secreted by several cell types including ECs in pathologic situations and in response to injury or growth factors. TSP-1 is a counteradhesive protein, increased in response to several injurious stimuli, and capable of inhibiting EC-substrate and cell-cell interactions (19). CD148 was recently described as a TSP-1 receptor (20) and is implicated in the regulation of vascular permeability. VEGF-dependent phosphorylation of CD148 was shown to be essential for proper regulation of EC permeability (21). We therefore assessed the expression of TSP-1 and its receptor, CD148 in ENG-deficient ECs. Figure 4A demonstrates a striking increase in TSP-1 in Eng+/2 (3-fold) and Eng2/2 (5.4-fold) cells relative to control cells. The levels of CD148 were significantly increased in Eng2/2 ECs but were not altered in Eng+/2 cells relative to control (Fig. 4B). The combined increase in TSP-1 and CD148 observed in Eng2/2 cells may contribute to the lethal phenotype of ENG null mouse embryos. Levels of proteins implicated in the stabilization of EC monolayers were measured in ENG-deficient cells. Figure 5 first demonstrates that ENG levels are indeed reduced in Eng+/2 and undetectable in Eng2/2 ECs. A similar pattern was observed for VEGFR2, whereas VEGF levels were unchanged between ENG-deficient and control cells. VE-cadherin, responsible for homophilic interaction at endothelial intercellular adherens junctions, was highly diminished in Eng2/2 but not Eng+/2 ECs. A similar pattern was observed for Rac-2, a Rho family GTPase that regulates cytoskeletal reorganization (Fig. 5). PAK-1, a p21 serine threonine kinase that links Rho GTPases to cytoskeletal remodeling, was reduced in Eng+/2 cells and significantly more so in Eng2/2 cells. Table 1 summarizes the changes observed in key protein expression in Eng+/2 and Eng2/2 ECs, as a percentage of their levels in Eng+/+ ECs, to facilitate comparison between the groups. The constitutive activation of RhoA is observed in both Eng+/2 and Eng2/2 ECs, as is the decrease in PAK-1. The changes in TSP-1 expression are the most noticeable, with levels at 305% in Eng+/2 ECs and 544% in Eng2/2 ECs, suggesting that ENG levels were inversely correlated to those of TSP-1. A rise in the TSP-1 receptor, CD148, and a decrease in VE-cadherin and Rac-2 are only observed in Eng2/2 ECs. Interestingly, the decrease in VEGFR2 levels parallels that of ENG, is half-reduced in Eng+/2 ECs, and is absent from Eng2/2 ECs.
A
4.5
* 3.0
1.5
0.0
B
Eng+/+
Eng+/-
Eng-/-
kDa
CD148
220
β-Actin
42
Relative CD148 density
Molecular signs of a destabilized endothelial barrier in ENG-deficient cells
1.5
Eng+/+ Eng+/Eng-/-
*†
1.0
0.5
0.0
Figure 4. The endothelial barrier destabilizing factor TSP-1 is increased in Eng+/2 and Eng2/2 cells, whereas CD148 is only increased in Eng2/2 cells. A) WB analysis revealed a large increase in TSP-1 expression in Eng+/2 and Eng2/2 cells compared with control Eng+/+ cells. B) Expression of CD148 was increased in Eng2/2 cells relative to control cells. N = 13–14 for both Eng+/+ and Eng2/2 groups and 8 for the Eng+/2 group for TSP-1, whereas for CD148, N = 20 for both Eng+/+ and Eng2/2 groups and 8 for the Eng+/2 group. Minimal adjustment (color balance, contrast, and brightness) of representative WB images was equally applied to the entire gel. *P , 0.05 vs. Eng+/+ cells; † P , 0.05 vs. Eng+/2 cells.
activated in ENG heterozygous and homozygous null ECs. Eng +/2 ECs also show a significant decrease in PAK-1 and a 3-fold increase in TSP-1, indicating that these changes may predispose to vascular lesions observed in HHT. Eng2/2 ECs display a further increase in TSP-1, suggesting that ENG regulates TSP-1, known to activate TGF-b1 and to act as a barrier destabilization factor. ENG null cells also show an increase in CD148, a receptor for TSP-1, and decreased expression of VE-cadherin and Rac-2, which could lead to improper assembly of EC junctions and barrier disruption. The expression of VEGFR2, a receptor critical for vascular development and function (23), parallels that of ENG. VEGFR2 is known to associate with VE-cadherin and may also contribute to stabilization of endothelial junctions.
The FASEB Journal x www.fasebj.org
JERKIC AND LETARTE
Eng+/+
Eng-/-
Eng+/-
kDa
Endoglin
170
VEGFR2
200
VEGF-A
38
β-Actin
42
VE-Cadherin
130
PAK-1
65
RAC-2
25
β-Actin
42
Eng+/+
Eng+/+
Eng-/-
Eng+/-
Eng+/-
Eng-/-
1.2
0.8
*
Relative density
Relative density
1.2
*
0.4
*
†
*
†
* 0.8
*
0.4
*
†
†
* 0.0
Eng
†
0.0
VEGFR2
VEGF-A
VE-Cadherin
PAK-1
Rac-2
Figure 5. Endothelial barrier stabilizing factors, VEGFR2 and PAK-1, are reduced in both Eng+/2 and Eng2/2 cells, whereas VEcadherin and Rac-2 are decreased only in Eng2/2 cells. WB analysis confirmed decreased levels of ENG in Eng+/2 ECs and no detectable levels in Eng2/2 ECs relative to control ECs, as expected. VEGFR2 expression was significantly reduced in Eng+/2 cells and barely detectable in Eng2/2 cells. VEGF-A expression was similar in all cell types. VE-cadherin and Rac-2 were highly reduced in Eng2/2 cells. PAK-1 expression was decreased proportionally in Eng+/2 and Eng2/2 cells, relative to control cells. N = 6–8 per group, except PAK-1 and VE-cadherin were N = 11–12 per group. Minimal adjustment (color balance, contrast, and brightness) of representative WB images was equally applied to the entire gel. *P , 0.05 vs. Eng+/+ cells; †P , 0.05 vs. Eng+/2 cells.
Our findings suggest that mouse ENG heterozygous cells, which serve as models for the HHT1 disease, do show significant changes in several proteins involved in maintaining barrier integrity of the vascular endothelium. Although the Eng2/2 phenotype is associated with early embryos (24), we previously studied large AVMs in brain and lungs of patients with HHT1. These very dilated and abnormal vessels have an extremely thin layer of endothelium and express lower levels of ENG than the expected 50% value due to haploinsufficiency (25). However, the relative ENG/CD31 density was unchanged, suggesting that many surface endothelial proteins are present at lower levels in these malformations. Therefore, ENG may reach a critically low threshold in AVMs, leading to the additional changes in reduced barrier stability reported here for the ENG null phenotype and ENDOGLIN AND ENDOTHELIAL CELL PERMEABILITY
accounting for the increased fragility of AVMs and their high risk of hemorrhage. The observations that the Eng2/2 ECs are constitutively activated in terms of RhoA and permeability, and allow for increased neutrophil transmigration, do support the notion of impaired barrier integrity. Increased permeability facilitates neutrophil TEM even though these 2 functions are not necessarily coupled and could be temporally separated (26). A weakened endothelial barrier and the improper assembly of EC-cell junctions could increase TEM via the paracellular pathway, which proceeds by sequential opening and reannealing of junctions (27). The involvement of ENG in transendothelial leukocyte trafficking was recently shown by Rossi et al. (7), who demonstrated that EC ENG interacts with leukocyte integrin a5b1. However, their data suggest that ENG 7
TABLE 1. Expression of proteins involved in endothelial barrier integrity Eng+/+ ECs (%)
Protein
ENG Rho-GTP TSP-1 CD148 VEGFR2 VEGF VE-cadherin PAK-1 Rac-2
100 100 100 100 100 100 100 100 100
6 6 6 6 6 6 6 6 6
Eng2/2 ECs (%)
Eng+/2 ECs (%)
5.5 7.7 3.8 3.0 14.6 7.3 1.6 3.2 18.1
65.4 160.7 305.1 117.9 56.2 96.6 90.2 81.9 128.9
6 6 6 6 6 6 6 6 6
4.5a 24.3a 48.3a 5.7 8.9a 5.0 3.5 5.0a 20.5
16.94 175.9 543.5 169.7 17.9 86.8 24.4 64.7 8.4
6 6 6 6 6 6 6 6 6
1.9a,† 7.7a 104.9a,b 9.5a,b 4a,b 4.0 2.2a,b 3.1a,b 1.3a,b
Protein expression, assessed by WB analysis, in Eng+/2 and Eng2/2 ECs relative to Eng+/+ cells is shown. Proteins with altered expression in the same direction or even mimicking the change in ENG levels are underlined. Levels of proteins inversely correlated with ENG expression in ECs are in bold. a P , 0.05 vs. Eng+/+ cells. bP , 0.05 vs. Eng+/2 cells.
promotes leukocyte transmigration and that inflammation triggered by carrageenan and LPS leads to lower leukocyte TEM to peritoneum or lungs in Eng+/2 vs. control mice. In contrast, our recent study showed increased gut inflammation and myeloid infiltration in colitic Eng+/2 vs. control mice (28). More experiments are needed to address these differences, but it should be noted that increased incidence of severe infections has been reported in patients with HHT, which could in part be associated with increased inflammation and leukocyte TEM (29, 30). Our results suggest that ENG is an important player in the activation of RhoA, the most important member of the small
Stabilization
Rho GTPases that orchestrate cytoskeleton rearrangements necessary for interendothelial gap formation. These processes are initiated by inflammation (31) or growth factors [VEGF (32) and TGF-b1 (33)], which induce Rho activation, leading to myosin light chain (MLC) phosphatase inhibition and prolonged MLC phosphorylation (Fig. 6). In consequence, stress fibers are formed, the contractile machinery is activated, interendothelial junctions are disassembled, and paracellular permeability is increased (34, 35). Our observations that RhoA is constitutively active and cannot be stimulated by VEGF-A or TGF-b1 in either Eng+/2 or Eng2/2 ECs indicate that the cells are already in an Destabilization
Tight junctions
Eng Cortical actin
Claudins & Occludins
ROS
eNOS
·
NO
RhoA
Cofilin - P
TGF-β receptors
LIM - P
PAK1 α
PAK Rac
MLC phosphatase
VE-Cadherin
Rho kinase
β
Stress fibre
Cortactin Adherens junctions
VEGFR2
TSP-1 TGF-β1
CD148 TSP -1
Figure 6. Proposed model of endothelial barrier disruption in ENG null ECs. The absence of Eng leads to increased expression and/or activity of proteins involved in endothelial barrier destabilization. ENG-deficient ECs exhibit increased ROS production due to eNOS uncoupling. Oxidative stress is a known cause of RhoA/Rho-kinase activity up-regulation. Constitutive RhoA activation leads to MLC phosphatase inhibition, prolonged MLC phosphorylation, and stress fiber formation. TSP-1 expression is markedly up-regulated in ENG null cells, causing activation of zonula adherens proteins (through its receptor, CD148) and barrier disruption. TSP-1 converts latent into active TGF-b1, and the large increase in TSP-1 may cause activation of a TGF-b1 autocrine loop in ENG null ECs. Proteins involved in barrier stabilization are decreased, including the key vascular stabilization factor VE-cadherin, Rac-2, and PAK-1, whereas VEGFR2 is absent. Rac/ PAK/LIM kinase coupling plays an important role in actin filament dynamics through cofilin inhibition and cortactin activation.
8
Vol. 29
September 2015
The FASEB Journal x www.fasebj.org
JERKIC AND LETARTE
inflammatory-like state and that ENG normally contributes to the regulation of RhoA activation and assembly of EC junction. RhoA activation has also been linked to oxidative stress (36), evidenced in our earlier studies as a major cause of EC dysfunction in Eng2/2 and Eng+/2 cells. We showed that ENG-deficient ECs exhibit increased reactive oxygen species (ROS) production due to eNOS uncoupling and impaired eNOS activity (37, 38). Increased endothelial oxidative stress, observed in tissues of Eng+/2 mice (39), is the major cause of pulmonary arterial hypertension (PAH) that develops with age in these mice because antioxidant therapy prevented the onset of PAH (38). Interestingly, over-activation of the RhoA/Rho-kinase signaling pathway is one of the main mechanisms involved in PAH pathogenesis (40). Impaired eNOS activity and increased ROS production have also been linked to RhoA/Rho-kinase activity up-regulation (41, 42). Blood-brain barrier (BBB) integrity alteration, induced by ROS, is mediated by signaling pathways that include RhoA (42); furthermore, specific RhoA inhibitors prevented ROSinduced monocyte migration across an in vitro BBB model. Increased ROS production caused RhoA activation in hyperoxic lungs (43), whereas antioxidant strategies alleviated bleomycin-induced lung fibrosis through RhoA/Rock pathway inhibition (44). Taken together, these findings suggest that eNOS-derived ROS overproduction in ENGdeficient cells leads to the constitutive activation of RhoA (Fig. 6). This mechanism could be in part responsible for cytoskeleton alterations and increased EC permeability and neutrophil TEM in ENG-deficient cells. Our recent study (11) demonstrated protein interactions between many Rho GTPases and TGF-b receptors. Pak-1 interacted with ENG, TGFBR2, ALK1, and ALK5 (11, 16), whereas Rac-2 and Rab38 interacted with the first 3. RhoH interacted with ALK1, RhoJ and Rasl with ALK1 and TGFBR2, and Arl8B with ENG and TGFBR2 (11). The Rho GTPase nucleotide exchange factor [guanine nucleotide exchange factor (GEF)] Arhgef25 was found to associate with ALK1 and TGFBR2, whereas Arhgef6 interacted with ALK1. These GEFs are activated by extracellular stimuli that work through GPCRs and stimulate Rho-dependent signals. Arhgef25 induces RhoA activation specifically and plays a key role in stress fiber formation and actin cytoskeleton reorganization in different tissues (45). Although the LUMIER screen, with its strict selection criteria, did not reveal a significant interaction between TGF-b receptors and RhoA, we detected its association with ENG by IP/WB (Fig. 3A). All of these findings strongly suggest that the TGF-b receptor complex is involved in RhoA activation. Our results also indicate that the presence of ENG in the receptor complex is crucial for proper RhoA cycling between inactive and active GTP-bound forms. We demonstrated previously that ENG binds TGF-b1 and TGF-b3 isoforms in association with TGFBR2 (46) and that Eng2/2 ECs show reduced binding to TGF-b1 as measured by affinity cross-linking. ENG alters TGF-b receptor levels in a very context-dependent manner and determines the balance between ALK5 (Smad2 and Smad3 mediated) or ALK1 (Smad1, Smad5, and Smad8 mediated) pathways (47). In a more recent study, we showed that ENG potentiates binding to bone morphogenetic protein (BMP)9 while reducing binding to BMP7, indicating that ENG can modulate the expression and activation of several receptors of the TGF-b superfamily (48). Of potential relevance to the current study is the report of ENDOGLIN AND ENDOTHELIAL CELL PERMEABILITY
an autocrine TSP-1/TGF-b1 loop in mesenchymal stem cells (49). In that study, activated TGF-b/Smad3 signaling up-regulated expression of TSP-1 via histone modification, and because TSP-1 converts latent into active TGF-b1, an autocrine feedback loop was proposed. We suggest that a related mechanism may operate in ECs, which produce and respond to TSP-1 (50). Such a loop may be turned on in ENG-deficient cells to compensate for reduced responses to TGF-b1 and in turn produce more TSP-1 to activate the TGF-b pathways. This would imply a critical role for ENG in the regulation of an autocrine TSP-1/TGF-b1 loop. It was recently recognized that TSP-1 opens the endothelial paracellular pathway by inducing tyrosine phosphorylation of proteins enriched at EC-EC boundaries (such as zonula adherens proteins and catenins), eventually causing endothelial barrier disruption (51). In a recent study, we documented a noticeable increase in TSP-1 expression in lungs of Eng+/2 mice (52), indicating that TSP-1 may play a role in endothelial barrier disruption and vessel destabilization, facilitating peripheral lung vasculature rarefaction and the onset of PAH observed in these mice (38). The pathway by which TSP-1 regulates permeability in the Eng+/2 vascular endothelium is not clear, but our current data suggest that its receptor, CD148 (DEP-1) is implicated in barrier destabilization in the Eng2/2 ECs. CD148 is required for permeability regulation (20). CD148 homozygous null mouse embryos show defects similar to those of ENG null embryos, displaying enlarged vessels impaired in vascular remodeling and branching, leading to death before E11.5. Interestingly, CD148 null ECs have a notable lack of ENG expression, confirming common regulatory elements for ENG and CD148 (53). In our case, Eng2/2 ECs showed increased CD148 expression, which likely follows the large increase in TSP-1, suggesting that the complete lack of ENG is associated with a substantial increase in the TSP-1/CD148 pathway, which might lead to the lethal phenotype of ENG null embryos (Fig. 6). Our findings also indicate that ENG-deficient ECs have decreased expression of proteins involved in barrier stability (i.e., PAK-1, Rac-2, and VE-cadherin). As mentioned above, the TGF-b receptor complex interacts strongly with the Rho GTPases PAK-1 and Rac-2 (11). The ability of PAK (p21activated) kinases to regulate cytoskeleton dynamics is underscored by the identification of a number of their important targets, including cortactin and MLC kinase (54). PAKs are downstream effectors of Rac (55), and their role in vascular permeability has been documented both in vitro and in vivo. In pulmonary ECs, increased endothelial leak could be prevented via a PAK-1-dependent mechanism and subsequent attenuation of Rho-mediated actin remodeling (56). The protective effects of atrial natriuretic peptide against LPS-induced vascular leak in murine lungs have been shown to be mediated at least in part by PAK-1-dependent signaling and EC barrier enhancement (57). Therefore, decreased PAK-1 expression in Eng+/2 ECs and more so in Eng2/2 ECs together with the lack of Rac-2 expression in ENG null cells could substantially contribute to barrier instability and increased endothelial permeability (Fig. 6). VE-cadherin is essential for the formation and regulation of endothelial adherens cell junctions. It associates with ENG and the TGF-b receptor complex, enhancing receptor complex formation and increasing TGF-b downstream signaling (10). Our data show that VE-cadherin is barely present in ENG-deficient ECs, which could cause EC junction 9
destabilization and further impairment of TGF-b signaling. The role of VE-cadherin in vascular permeability and leukocyte diapedesis was shown in vivo, in knockin mice expressing a VE-cadherin-catenin fusion construct (58). The fusion protein was most efficient at interacting with the actin cytoskeleton and led to highly stabilized skin EC contacts resistant to VEGF- or histamine-induced vascular leakiness and to reduced leukocyte recruitment into various tissues of these mice (57). These data corroborate our findings that ENG null cells show decreased VE-cadherin, increased permeability, and higher neutrophil transmigration. Hyperpermeable ECs with destabilized junctions could contribute substantially to vascular fragility and hemorrhage of the ENG null embryos. VE-cadherin also associates with VEGFR2 and CD148. VEGFR2 is internalized more rapidly when VEcadherin is absent or not engaged at junctions, whereas silencing the junction-associated CD148 restores VEGFR2 internalization and signaling (59). Therefore, the lack of VEcadherin expression and increase in CD148 in ENG null cells may account for the absence of VEGFR2 in these cells. ENG via its juxtamembrane and cytosolic domains interacts with many proteins involved in membrane stability and also serves as their scaffold (60). The absence of ENG may also contribute to the lack of expression of VE-cadherin and VEGFR2 at the cell surface (Fig. 6 and Table 1). Our current results support the spatial and functional association among VEcadherin, VEGFR2, and CD148 showing their critical involvement in EC barrier stability and cell permeability. In summary, our study suggests a novel and important role for ENG in regulating RhoA activation and vascular permeability that might explain why the complete absence of ENG leads to disrupted endothelial junctions, a leaky vasculature, and a fatal embryonic phenotype. Eng+/2 ECs also show constitutive activation of RhoA and permeability and a large increase in TSP-1 but reveal fewer changes in terms of barrier stabilizing and destabilizing molecules; however, the observed alterations are likely to contribute to the eventual generation of vascular malformations characteristic of HHT. The authors thank Albert Zhe Liang and Zobia Jawed for help with experiments. Special thanks are due to Dr. Katalin Szaszi for helpful discussion about design of the permeability experiments. M.J. made significant contributions to the concept and design of the experiments and was involved in acquisition of the data, their analysis, and interpretation. She has been critically involved in drafting and revising the manuscript for important intellectual content. M.L. made significant contributions to the concept of the experiments and was critically involved in data analysis and interpretation. She has been involved in drafting and especially revising the manuscript for important intellectual content. Both authors have given final approval of the submitted version of the manuscript and agreed to be accountable for all aspects of the work. This work was supported by a Grant T5598 from the Heart and Stroke Foundation of Canada (to M.L.). The authors declare no conflicts of interest.
REFERENCES 1. Gougos, A., and Letarte, M. (1990) Primary structure of endoglin, an RGD-containing glycoprotein of human endothelial cells. J. Biol. Chem. 265, 8361–8364 2. Barbara, N. P., Wrana, J. L., and Letarte, M. (1999) Endoglin is an accessory protein that interacts with the signaling receptor complex of multiple members of the transforming growth factor-beta superfamily. J. Biol. Chem. 274, 584–594
10
Vol. 29 September 2015
3. Arthur, H. M., Ure, J., Smith, A. J., Renforth, G., Wilson, D. I., Torsney, E., Charlton, R., Parums, D. V., Jowett, T., Marchuk, D. A., Burn, J., and Diamond, A. G. (2000) Endoglin, an ancillary TGFbeta receptor, is required for extraembryonic angiogenesis and plays a key role in heart development. Dev. Biol. 217, 42–53 4. Li, D. Y., Sorensen, L. K., Brooke, B. S., Urness, L. D., Davis, E. C., Taylor, D. G., Boak, B. B., and Wendel, D. P. (1999) Defective angiogenesis in mice lacking endoglin. Science 284, 1534–1537 5. Shovlin, C. L. (2010) Hereditary haemorrhagic telangiectasia: pathophysiology, diagnosis and treatment. Blood Rev. 24, 203–219 6. Oh, S. P., Seki, T., Goss, K. A., Imamura, T., Yi, Y., Donahoe, P. K., Li, L., Miyazono, K., ten Dijke, P., Kim, S., and Li, E. (2000) Activin receptorlike kinase 1 modulates transforming growth factor-beta 1 signaling in the regulation of angiogenesis. Proc. Natl. Acad. Sci. USA 97, 2626–2631 7. Rossi, E., Sanz-Rodriguez, F., Eleno, N., D¨uwell, A., Blanco, F. J., Langa, C., Botella, L. M., Cabañas, C., Lopez-Novoa, J. M., and Bernabeu, C. (2013) Endothelial endoglin is involved in inflammation: role in leukocyte adhesion and transmigration. Blood 121, 403–415 8. Tian, H., Mythreye, K., Golzio, C., Katsanis, N., and Blobe, G. C. (2012) Endoglin mediates fibronectin/a5b1 integrin and TGF-b pathway crosstalk in endothelial cells. EMBO J. 31, 3885–3900 9. Conley, B. A., Koleva, R., Smith, J. D., Kacer, D., Zhang, D., Bernab´eu, C., and Vary, C. P. (2004) Endoglin controls cell migration and composition of focal adhesions: function of the cytosolic domain. J. Biol. Chem. 279, 27440–27449 10. Rudini, N., Felici, A., Giampietro, C., Lampugnani, M., Corada, M., Swirsding, K., Garr`e, M., Liebner, S., Letarte, M., ten Dijke, P., and Dejana, E. (2008) VE-cadherin is a critical endothelial regulator of TGF-beta signalling. EMBO J. 27, 993–1004 11. Xu, G., Barrios-Rodiles, M., Jerkic, M., Turinsky, A. L., Nadon, R., Vera, S., Voulgaraki, D., Wrana, J. L., Toporsian, M., and Letarte, M. (2014) Novel protein interactions with endoglin and activin receptor-like kinase 1: potential role in vascular networks. Mol. Cell. Proteomics 13, 489–502 12. Karabegovic, A., Shinawi, M., Cymerman, U., and Letarte, M. (2004) No live individual homozygous for a novel endoglin mutation was found in a consanguineous Arab family with hereditary haemorrhagic telangiectasia. J. Med. Genet. 41, e119 13. Nagy, J. A., Benjamin, L., Zeng, H., Dvorak, A. M., and Dvorak, H. F. (2008) Vascular permeability, vascular hyperpermeability and angiogenesis. Angiogenesis 11, 109–119 14. Pece-Barbara, N., Vera, S., Kathirkamathamby, K., Liebner, S., Di Guglielmo, G. M., Dejana, E., Wrana, J. L., and Letarte, M. (2005) Endoglin null endothelial cells proliferate faster and are more responsive to transforming growth factor beta1 with higher affinity receptors and an activated Alk1 pathway. J. Biol. Chem. 280, 27800–27808 15. Hordijk, P. L., Anthony, E., Mul, F. P., Rientsma, R., Oomen, L. C., and Roos, D. (1999) Vascular-endothelial-cadherin modulates endothelial monolayer permeability. J. Cell Sci. 112, 1915–1923 16. Barrios-Rodiles, M., Brown, K. R., Ozdamar, B., Bose, R., Liu, Z., Donovan, R. S., Shinjo, F., Liu, Y., Dembowy, J., Taylor, I. W., Luga, V., Przulj, N., Robinson, M., Suzuki, H., Hayashizaki, Y., Jurisica, I., and Wrana, J. L. (2005) High-throughput mapping of a dynamic signaling network in mammalian cells. Science 307, 1621–1625 17. Liang, W. C., Wu, X., Peale, F. V., Lee, C. V., Meng, Y. G., Gutierrez, J., Fu, L., Malik, A. K., Gerber, H. P., Ferrara, N., and Fuh, G. (2006) Crossspecies vascular endothelial growth factor (VEGF)-blocking antibodies completely inhibit the growth of human tumor xenografts and measure the contribution of stromal VEGF. J. Biol. Chem. 281, 951–961 18. Bustelo, X. R., Sauzeau, V., and Berenjeno, I. M. (2007) GTP-binding proteins of the Rho/Rac family: regulation, effectors and functions in vivo. BioEssays 29, 356–370 19. Liu, A., Mosher, D. F., Murphy-Ullrich, J. E., and Goldblum, S. E. (2009) The counteradhesive proteins, thrombospondin 1 and SPARC/ osteonectin, open the tyrosine phosphorylation-responsive paracellular pathway in pulmonary vascular endothelia. Microvasc. Res. 77, 13–20 20. Takahashi, K., Mernaugh, R. L., Friedman, D. B., Weller, R., Tsuboi, N., Yamashita, H., Quaranta, V., and Takahashi, T. (2012) Thrombospondin-1 acts as a ligand for CD148 tyrosine phosphatase. Proc. Natl. Acad. Sci. USA 109, 1985–1990 21. Spring, K., Chabot, C., Langlois, S., Lapointe, L., Trinh, N. T., Caron, C., Hebda, J. K., Gavard, J., Elchebly, M., and Royal, I. (2012) Tyrosine phosphorylation of DEP-1/CD148 as a mechanism controlling Src kinase activation, endothelial cell permeability, invasion, and capillary formation. Blood 120, 2745–2756
The FASEB Journal x www.fasebj.org
JERKIC AND LETARTE
22. Seghers, L., de Vries, M. R., Pardali, E., Hoefer, I. E., Hierck, B. P., ten Dijke, P., Goumans, M. J., and Quax, P. H. (2012) Shear induced collateral artery growth modulated by endoglin but not by ALK1. J. Cell. Mol. Med. 16, 2440–2450 23. Holmes, K., Roberts, O. L., Thomas, A. M., and Cross, M. J. (2007) Vascular endothelial growth factor receptor-2: structure, function, intracellular signalling and therapeutic inhibition. Cell. Signal. 19, 2003–2012 24. Bourdeau, A., Dumont, D. J., and Letarte, M. (1999) A murine model of hereditary hemorrhagic telangiectasia. J. Clin. Invest. 104, 1343–1351 25. Bourdeau, A., Cymerman, U., Paquet, M. E., Meschino, W., McKinnon, W. C., Guttmacher, A. E., Becker, L., and Letarte, M. (2000) Endoglin expression is reduced in normal vessels but still detectable in arteriovenous malformations of patients with hereditary hemorrhagic telangiectasia type 1. Am. J. Pathol. 156, 911–923 26. Kim, M. H., Curry, F. R., and Simon, S. I. (2009) Dynamics of neutrophil extravasation and vascular permeability are uncoupled during aseptic cutaneous wounding. Am. J. Physiol. Cell Physiol. 296, C848–C856 27. Yuan, S. Y., Shen, Q., Rigor, R. R., and Wu, M. H. (2012) Neutrophil transmigration, focal adhesion kinase and endothelial barrier function. Microvasc. Res. 83, 82–88 28. Peter, M. R., Jerkic, M., Sotov, V., Douda, D. N., Ardelean, D. S., Ghamami, N., Lakschevitz, F., Khan, M. A., Robertson, S. J., Glogauer, M., Philpott, D. J., Palaniyar, N., and Letarte, M. (2014) Impaired resolution of inflammation in the Endoglin heterozygous mouse model of chronic colitis. [E-pub ahead of print] Mediators Inflamm. 10.1155/2014/767185 29. Dupuis-Girod, S., Giraud, S., Decullier, E., Lesca, G., Cottin, V., Faure, F., Merrot, O., Saurin, J. C., Cordier, J. F., and Plauchu, H. (2007) Hemorrhagic hereditary telangiectasia (Rendu-Osler disease) and infectiousdiseases:anunderestimatedassociation.Clin.Infect.Dis.44,841–845 30. Mathis, S., Dupuis-Girod, S., Plauchu, H., Giroud, M., Barroso, B., Ly, K. H., Ingrand, P., Gilbert, B., Goden`eche, G., and Neau, J. P. (2012) Cerebral abscesses in hereditary haemorrhagic telangiectasia: a clinical and microbiological evaluation. Clin. Neurol. Neurosurg. 114, 235–240 31. Tong, L., and Tergaonkar, V. (2014) Rho protein GTPases and their interactions with NFkB: crossroads of inflammation and matrix biology. Biosci. Rep. 34, e00115 32. Bryan, B. A., Dennstedt, E., Mitchell, D. C., Walshe, T. E., Noma, K., Loureiro, R., Saint-Geniez, M., Campaigniac, J. P., Liao, J. K., and D’Amore, P. A. (2010) RhoA/ROCK signaling is essential for multiple aspects of VEGF-mediated angiogenesis. FASEB J. 24, 3186–3195 33. Varon, C., Basoni, C., Reuzeau, E., Moreau, V., Kramer, I. J., and G´enot, E. (2006) TGFbeta1-induced aortic endothelial morphogenesis requires signaling by small GTPases Rac1 and RhoA. Exp. Cell Res. 312, 3604–3619 34. Vandenbroucke, E., Mehta, D., Minshall, R., and Malik, A. B. (2008) Regulation of endothelial junctional permeability. Ann. N. Y. Acad. Sci. 1123, 134–145 35. Hornburger, M. C., Mayer, B. A., Leonhardt, S., Willer, E. A., Zahler, S., Beyerle, A., Rajalingam, K., Vollmar, A. M., and Fu¨ rst, R. (2014) A novel role for inhibitor of apoptosis (IAP) proteins as regulators of endothelial barrier function by mediating RhoA activation. FASEB J. 28, 1938–1946 36. Aghajanian, A., Wittchen, E. S., Campbell, S. L., and Burridge, K. (2009) Direct activation of RhoA by reactive oxygen species requires a redox-sensitive motif. PLoS One 4, e8045 37. Toporsian, M., Gros, R., Kabir, M. G., Vera, S., Govindaraju, K., Eidelman, D. H., Husain, M., and Letarte, M. (2005) A role for endoglin in coupling eNOS activity and regulating vascular tone revealed in hereditary hemorrhagic telangiectasia. Circ. Res. 96, 684–692 38. Toporsian, M., Jerkic, M., Zhou, Y. Q., Kabir, M. G., Yu, L. X., McIntyre, B. A. S., Davies, A., Wang, Y. J., Stewart, D. J., Belik, J., Husain, M., Henkelman, M., and Letarte, M. (2010) Spontaneous adult-onset pulmonary arterial hypertension attributable to hemorrhagic telangiectasia increased endothelial oxidative stress in a murine model of hereditary hemorrhagic telangiectasia. Arterioscler. Thromb. Vasc. Biol. 30, 509–517 39. Jerkic, M., Sotov, V., and Letarte, M. (2012) Oxidative stress contributes to endothelial dysfunction in mouse models of hereditary hemorrhagic telangiectasia. Oxid. Med. Cell. Longev. 2012, 686972 40. Duong-Quy, S., Bei, Y., Liu, Z., and Dinh-Xuan, A. T. (2013) Role of Rho-kinase and its inhibitors in pulmonary hypertension. Pharmacol. Ther. 137, 352–364 41. Bei, Y., Duong-Quy, S., Hua-Huy, T., Dao, P., Le-Dong, N.-N., and Dinh-Xuan, A. T. (2013) Activation of RhoA/Rho-kinase pathway accounts for pulmonary endothelial dysfunction in patients with chronic obstructive pulmonary disease. Physiol. Rep. 1, e00105
ENDOGLIN AND ENDOTHELIAL CELL PERMEABILITY
42. Schreibelt, G., Kooij, G., Reijerkerk, A., van Doorn, R., Gringhuis, S. I., van der Pol, S., Weksler, B. B., Romero, I. A., Couraud, P. O., Piontek, J., Blasig, I. E., Dijkstra, C. D., Ronken, E., and de Vries, H. E. (2007) Reactive oxygen species alter brain endothelial tight junction dynamics via RhoA, PI3 kinase, and PKB signaling. FASEB J. 21, 3666–3676 43. Kondrikov, D., Caldwell, R. B., Dong, Z., and Su, Y. (2011) Reactive oxygen species-dependent RhoA activation mediates collagen synthesis in hyperoxic lung fibrosis. Free Radic. Biol. Med. 50, 1689–1698 44. Meng, Y., Li, T., Zhou, G. S., Chen, Y., Yu, C. H., Pang, M. X., Li, W., Li, Y., Zhang, W. Y., and Li, X. (2015) The angiotensin-converting enzyme 2/angiotensin (1-7)/Mas axis protects against lung fibroblast migration and lung fibrosis by inhibiting the NOX4-derived ROS-mediated RhoA/Rho kinase pathway. Antioxid. Redox Signal. 22, 241–258 45. Souchet, M., Portales-Casamar, E., Mazurais, D., Schmidt, S., L´eger, I., Javr´e, J. L., Robert, P., Berrebi-Bertrand, I., Bril, A., Gout, B., Debant, A., and Calmels, T. P. (2002) Human p63RhoGEF, a novel RhoA-specific guanine nucleotide exchange factor, is localized in cardiac sarcomere. J. Cell Sci. 115, 629–640 46. Cheifetz, S., Bell´on, T., Cal´es, C., Vera, S., Bernabeu, C., Massagu´e, J., and Letarte, M. (1992) Endoglin is a component of the transforming growth factor-beta receptor system in human endothelial cells. J. Biol. Chem. 267, 19027–19030 47. Lee, N. Y., Ray, B., How, T., and Blobe, G. C. (2008) Endoglin promotes transforming growth factor beta-mediated Smad 1/5/8 signaling and inhibits endothelial cell migration through its association with GIPC. J. Biol. Chem. 283, 32527–32533 48. Gregory, A. L., Xu, G., Sotov, V., and Letarte, M. (2014) Review: the enigmatic role of endoglin in the placenta. Placenta 35(Suppl), S93–S99 49. Qin, H., Qu, C., Yamaza, T., Yang, R., Lin, X., Duan, X. Y., Akiyama, K., Liu, Y., Zhang, Q., Chen, C., Chen, Y., Qi, H. H., Feng, X. H., Le, A. D., and Shi, S. (2013) Ossifying fibroma tumor stem cells are maintained by epigenetic regulation of a TSP1/TGF-b/SMAD3 autocrine loop. Cell Stem Cell 13, 577–589 50. Bornstein, P. (2009) Thrombospondins function as regulators of angiogenesis. J. Cell Commun. Signal. 3, 189–200 51. Garg, P., Yang, S., Liu, A., Pallero, M. A., Buchsbaum, D. J., Mosher, D. F., Murphy-Ullrich, J. E., and Goldblum, S. E. (2011) Thrombospondin-1 opens the paracellular pathway in pulmonary microvascular endothelia through EGFR/ErbB2 activation. Am. J. Physiol. Lung Cell. Mol. Physiol. 301, L79–L90 52. Ardelean, D. S., Jerkic, M., Yin, M., Peter, M., Ngan, B., Kerbel, R. S., Foster, F. S., and Letarte, M. (2014) Endoglin and activin receptor-like kinase 1 heterozygous mice have a distinct pulmonary and hepatic angiogenic profile and response to anti-VEGF treatment. Angiogenesis 17, 129–146 53. Takahashi, T., Takahashi, K., St John, P. L., Fleming, P. A., Tomemori, T., Watanabe, T., Abrahamson, D. R., Drake, C. J., Shirasawa, T., and Daniel, T. O. (2003) A mutant receptor tyrosine phosphatase, CD148, causes defects in vascular development. Mol. Cell. Biol. 23, 1817–1831 54. Kumar, R., Gururaj, A. E., and Barnes, C. J. (2006) p21-activated kinases in cancer. Nat. Rev. Cancer 6, 459–471 55. Maharjan, S., Kim, K., Agrawal, V., Choi, H. J., Kim, N. J., Kim, Y. M., Suh, Y. G., and Kwon, Y. G. (2013) Sac-1004, a novel vascular leakage blocker, enhances endothelial barrier through the cAMP/Rac/ cortactin pathway. Biochem. Biophys. Res. Commun. 435, 420–427 56. Tian, X., Tian, Y., Gawlak, G., Sarich, N., Wu, T., and Birukova, A. A. (2014) Control of vascular permeability by atrial natriuretic peptide via a GEF-H1-dependent mechanism. J. Biol. Chem. 289, 5168–5183 57. Birukova, A. A., Xing, J., Fu, P., Yakubov, B., Dubrovskyi, O., Fortune, J. A., Klibanov, A. M., and Birukov, K. G. (2010) Atrial natriuretic peptide attenuates LPS-induced lung vascular leak: role of PAK1. Am. J. Physiol. Lung Cell. Mol. Physiol. 299, L652–L663 58. Schulte, D., K¨uppers, V., Dartsch, N., Broermann, A., Li, H., Zarbock, A., Kamenyeva, O., Kiefer, F., Khandoga, A., Massberg, S., and Vestweber, D. (2011) Stabilizing the VE-cadherin-catenin complex blocks leukocyte extravasation and vascular permeability. EMBO J. 30, 4157–4170 59. Lampugnani, M. G., Orsenigo, F., Gagliani, M. C., Tacchetti, C., and Dejana, E. (2006) Vascular endothelial cadherin controls VEGFR-2 internalization and signaling from intracellular compartments. J. Cell Biol. 174, 593–604 60. Rossi, E., Lopez-Novoa, J. M., and Bernabeu, C. (2015) Endoglin involvement in integrin-mediated cell adhesion as a putative pathogenic mechanism in hereditary hemorrhagic telangiectasia type 1 (HHT1). Front. Genet. 5, 457 Received for publication February 6, 2015. Accepted for publication May 1, 2015.
11
