OUP user menu

Interaction between Tie receptors modulates angiogenic activity of angiopoietin2 in endothelial progenitor cells

Koung Li Kim, In-Soon Shin, Jeong-Min Kim, Jin-Ho Choi, Jonghoe Byun, Eun-Seok Jeon, Wonhee Suh, Duk-Kyung Kim
DOI: http://dx.doi.org/10.1016/j.cardiores.2006.08.002 394-402 First published online: 1 December 2006


Objective: Ischemia-dependent upregulation of angiopoietin2 (Ang2) led us to hypothesize the potentially proangiogenic Ang2-Tie2 signaling in endothelial progenitor cells (EPCs). Given the well-known vascular destabilizing action of Ang2 in mature endothelium, we investigated the yet unidentified mechanism behind cell-dependent differential activity of Ang2.

Methods and results: Both in vitro and in vivo experiments showed that Ang2 promoted angiogenicity of human cord blood-derived EPCs, where Ang2 directly activated Tie2 and its related downstream signaling molecules. However, Ang2 had no such effect in fully differentiated human umbilical vein endothelial cells (HUVECs) under the same condition. Such a cell-dependent Tie2 activation by Ang2 was explained by comparing EPCs and HUVECs, where most Tie2 receptors in EPCs were found to be present unbound to Tie1, whereas those in HUVECs existed as heterocomplexes with Tie1. When Tie2 in HUVECs was prevented from forming heterocomplexes by silencing Tie1 expression, they underwent rapid phosphorylation upon Ang2 treatment, as shown in EPCs.

Conclusions: In contrast with its roles in mature endothelial cells, Ang2 has proangiogenic activities in EPC directly through Tie2 signaling pathway. Such a cell-dependent differential reactivity of Ang2 was for the first time found to be modulated by physical association between Tie1 and Tie2, which inhibited Ang2-mediated Tie2 activation.

  • Angiogenesis
  • Angiopoietin2
  • Endothelial progenitor cell
  • Tie1
  • Tie2

This article is referred to in the Editorial by K.-G. Shyu (pages 359–360) in this issue.

1. Introduction

Tie2, one of the major endothelial-specific receptor tyrosine kinases, plays a central role in the formation of blood vessels [1]. Among Tie2 ligands, angiopoietin2 (Ang2) is strongly upregulated at sites of ischemia, whereas angiopoietin1 (Ang1) gene expression is not significantly altered by hypoxia. In animal studies and clinical reports, Ang2 has been found to be profoundly elevated in various ischemic tissues, and in patients with acute coronary syndromes or chronic heart failure [2–5]. This hypoxia-dependent upregulation of Ang2 implies that Ang2 might be a potentially important factor in neovascularization occurring at regions of injured vasculature. In recent years, it has become clear that postnatal neovascularization not only depends on pre-existing endothelial cells, but also involves the bone marrow-derived endothelial progenitor cells (EPCs) [6–8]. This EPC-mediated neovascularization is regulated by a variety of hypoxia-inducible genes that play primary roles in promoting EPC mobilization and homing, and in enhancing EPC survival and differentiation in local hypoxic milieus [9,10]. These findings circumstantially led us to speculate that Ang2 might play proangiogenic roles in EPC-mediated neovascularization at ischemic tissues, as other hypoxia-inducible cytokines do.

Ang2 was first identified as a natural antagonist of Ang1 in the interaction with the Tie2 receptor, in that Ang2 transgenic mice exhibit similar cardiovascular defects observed in Ang1- and Tie2-deficient mice during embryonic vascular development [11]. However, recent studies indicate that Ang2 does not simply act as an inhibitor of Ang1, but that it also acts agonistically in a context-dependent manner. In an ocular neovascularization, Ang2 overexpression was found to increase the retinal capillary density at early postnatal stage when VEGF levels remained low [12]. However, as vessels matured, they appeared to lose the sensitivity to Ang2, which reduced the capillary density as low as that in control mice at later stages. Although this study did not elucidate the mechanism of altered sensitivity to Ang2, it interestingly claimed that there must be something to modulate the sensitivity of vessels to Ang2 in different vascular milieu. In addition, lymphatic vessel defects shown in Ang2-deficient mice and their genetic rescue with Ang1, indicate that Ang2 acts as a Tie2 agonist in lymphatic endothelial cells, being requisite for lymphatic patterning and maturation during development [13]. Despite growing evidences to exhibit two opposing properties of Ang2, it has yet been clearly explained what factors regulate such a differential activity of Ang2 under varying conditions. Therefore, the present study has aimed to investigate (1) the potentially proangiogenic Ang2-Tie2 signaling in EPCs and, if any, (2) the mechanism behind context-dependent different effect of Ang2 on Tie2 activation, by comparing with well-known Tie2 antagonistic activity of Ang2 in mature endothelial cells.

2. Methods

2.1. Cell culture

Human umbilical cord bloods were obtained from healthy newborns. The Institutional Review Board at Samsung Medical Center approved all protocols, and informed consent was obtained from all donors. The investigation conformed with the principles outlined in the Declaration of Helsinki. After mononuclear cells have been isolated from umbilical cord bloods by Ficoll density centrifugation, CD34+ cells were selected using immunomagnetic beads (Miltenyi Biotec, Gladbach, Germany), and cultured in endothelial growth medium (EGM)-2 MV Singlequot (Clonetics, Walkersville, MD, USA). EPCs usually appeared by two weeks of culture and were passaged up to eight times while being maintained at 50–60% confluence. HUVECs were cultured in EGM to passages four to seven. HUVECs, culture media, and fetal bovine serum (FBS) were purchased from Clonetics.

2.2. In vitro angiogenic assays

For the analysis of angiogenic responses toward angiopoietins (Alexis Biochemicals, Lausen, Switzerland), or VEGF (50 ng/ml; R&D Systems, Minneapolis, MN, USA), tube formation, boyden migration, and cell survival assays were performed as follows. For the analysis of tube formation, cells were seeded onto Matrigel (BD Bioscience, Bedford, MA, USA)-coated plates, and incubated in M199+1% FBS media or supplemented with angiopoietins (200 ng/ml) or VEGF (50 ng/ml). After overnight incubation, tube networks were quantified by measuring the tubule length in four random microscope fields. Migration assays were performed using a modified Boyden chamber (Costar, Cambridge, MA, USA), where cells were placed in the upper chamber and the lower chamber was filled with M199+1% FBS media, or supplemented with angiopoietins or VEGF. After incubation of 24 h, migrated cells that attached to the lower side of the filter were stained with Giemsa solution and counted. For cell survival assays, confluent cells were incubated overnight with angiopoietins or VEGF in M199+1% FBS media. Cells were then washed with phosphate buffered saline (PBS) and stained with Annexin V (Molecular Probes, Eugene, OR, USA) according to the manufacturer's protocol. The adherent Annexin V-negative cells were counted in four random microscope fields. In Tie2 receptor blocking experiments, EPCs were pretreated with anti-human Tie2 (R&D Systems) or -human CD34 (DAKO, Carpinteria, CA, USA) antibodies for 30 min at 4 °C. After washing cells with PBS several times, in vitro angiogenic experiments were carried out under the same condition. When blocking the Akt and eNOS signaling pathway in EPCs, cells were pretreated with PBS, Akt inhibitor (1–4 μM; Calbiochem, Darmstadt, Germany), or Nw-nitro-l-arginine methyl ester (l-NAME, 6 mM; Sigma-Aldrich, St. Louis, MO, USA) 2 h before addition of PBS or Ang2 (200 ng/ml). Then, in vitro angiogenic assays were performed as described above.

2.3. Immunoprecipitation and western blot analysis

For the immunoprecipitation with Tie2 antibody, cells were pre-crosslinked with 0.5 M of dithio-bis(succinimidylpropionate) (DTSSP) for 30 min, quenched by addition of 100 mM of Tris (pH 7.5) and lysed in lysis buffer (50 mM Tris, 50 mM NaCl, 1% Triton X-100, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 1 mM EGTA, and protease inhibitor). Then, an equal amount of cell lysate protein was incubated with Tie2 antibody (R&D Systems, AF313), followed by addition of protein A-Sepharose (Santa Cruz Biotechnology Inc. Santa Cruz, CA, USA). The immunocomplexes captured by protein A-Sepharose were separated using 7.5% SDS-polyacrylamide gel electrophoresis (PAGE) and blotted with primary antibodies against Tie1 or Tie2 antibodies (antibodies from Santa Cruz Biotechnology Inc. for immunoprecipitated samples; SC-342 and SC-9026 for Tie1 and Tie2 respectively, antibodies from R&D Systems for normal cell lysates; AF619 and AF313 for Tie1 and Tie2 respectively). Blots were incubated with horseradish peroxidase-conjugated secondary antibody and developed using a chemiluminescence kit.

For the phosphorylation analysis of Tie2, Akt, and eNOS, cells were starved overnight and then stimulated with Ang2. After separation of cell lysate through SDS-PAGE, the membrane was transferred and probed with primary antibodies against phospho-Tie2, phospho-Akt, phospho-eNOS, Tie2, Akt, or eNOS. Blots were incubated and developed as described above. Densitometric analysis of western blots was performed with the use of the Image-Lab (MCM design, Birker⊘d, Denmark). Antibodies to Akt, phospho-Akt, and phospho-Tie2 were purchased from Cell Signaling Technology (Beverly, MA, USA). Antibodies against eNOS and phospho-eNOS were obtained from Zymed Laboratories (South San Francisco, CA, USA) and Upstate Cell Signaling Solutions (Lake Placid, NY, USA) respectively.

2.4. RNA interference and reverse transcriptase-polymerase chain reaction (RT-PCR)

To silence the Tie1 expression in HUVECs, subconfluent cells were transfected with Tie1-specific small interfering RNA (siRNA) (catalog no; 139494 (Tie1-I), 139495 (Tie1-II), Ambion, Austin, TX, USA) or an unrelated siRNA (Ambion) as a control, by using RNAi-Shuttle siRNA transfection reagent (Orbigen, San Diego, CA, USA). The gene expression of Tie1, Tie2, and β-actin was determined by RT-PCR, where PCR primers were designed as follows: Tie1 forward, 5′-atggctgctcttgtggatct-3′, reverse, 5′-gggcactttcacattgacct-3′; Tie2 forward, 5′-ctgcctaaaagtcagaccac-3′, reverse 5′-gtgttgactctagctcggac-3′; β-actin, forward, 5′-aagacattttcgggctcac-3′, reverse, 5′-ggcactttagtagttctcc-3′.

2.5. In vivo angiogenesis assay

EPCs were suspended in extracellular matrix (ECM) solution composed of rat tail type 1 collagen (BD Bioscience) and human plasma fibronectin (Sigma-Aldrich), as reported by Schechner et al. [14]. After incubation at 37 °C, the solidified gel was uniformly cut into circular disk-shaped gels using a skin biopsy punch (Acuderm Inc., Fort Lauderdale, FL, USA). For the gel implantation, the dorsal skin of male athymic nude mice (Charles River Laboratories, Yokohama, Japan) was incised and infected with adenoviral LacZ or Ang2 (4×108 plaque forming units in 20 μl). Then, circular disk-shaped gels were implanted underneath the adenovirus-transduced skin and confined using a simple intersecting suture. All procedures were performed in accordance with the Guidelines for the Care and Use of Laboratory Animals published by the US National Institute of Health (NIH publication No. 85-23, revised 1996). For surgical procedures, mice were anesthetized with an intraperitoneal injection of 100 μl of solution containing 2.215 mg of ketamine and 0.175 mg of xylazine.

2.6. Immunohistochemistry

The gels, harvested from the mice after 2 months, were fixed and embedded for the analysis of capillary density. After quenching the endogenous peroxidase activity and blocking with 10% normal goat serum, slides were incubated with anti-human CD31 antibody (DAKO) and then with biotinylated anti-sheep IgG (Jackson ImmunoResearch, West Grove, PA, USA). Positive immunoreactivity was visualized using ABC-peroxidase kits (Elite kit, Vector Laboratories, Burlingame, CA, USA) and 3,3′-diaminobenzidine tetrahydrochloride (Vector Laboratories). Controls of immunostaining preparations were prepared using incubations with irrelevant class-and species-matched IgGs.

2.7. Statistical analysis

All data are presented as mean±SEMs. One-way analysis of variance was used to determine the significance of differences between groups, where appropriate, with post hoc Student's t-tests for unpaired observations and Bonferroni correction for multiple comparisons. p<0.05 was assumed statistically significant and the number of samples examined is indicated by n.

3. Results

3.1. EPC characterization

EPCs were obtained from endothelial cultivation of human cord blood CD34+ cells (as described in the Supplementary information) and their endothelial phenotypes were characterized as reported [15,16]. Human cord blood-derived EPCs displayed the cobblestone morphology similar to that of endothelial cells and formed tubular networks on Matrigel (Supplementary Fig. IA). In fluorescence-activated cell sorting analysis, EPCs were strongly positive for endothelial-specific markers such as CD31 (>99%), KDR (72%), and Tie2 (>99%) (Supplementary Fig. IB). The contamination of EPCs with hematopoietic cells would be negligible due to several subcultures for 2 weeks.

3.2. Ang2 enhances angiogenicity of EPCs

Since binding of Ang1 to the Tie2 receptor is known to stimulate endothelial tube formation, migration, and cell survival, we performed in vitro angiogenesis assays to examine whether Ang2 as well as Ang1 affected the angiogenicity of EPCs. As shown in Fig. 1A, Ang2 significantly promoted tube formation, chemotactic migration, and cell survival activities in EPCs, and these proangiogenic effects of Ang2 on EPCs were observed to the same extent as those of Ang1. In this experiment, a VEGF-treated group was included as a positive control to confirm that in vitro angiogenic assays were appropriately performed with EPCs. When the same experiments were carried out with mature endothelial cells, Ang1 and VEGF elicited angiogenic responses in HUVECs, whereas Ang2 did not exhibit the proangiogenic property unlike in EPCs (Fig. 1B).

Fig. 1

Ang2 has angiogenic activities in EPCs, but not in HUVECs. A. Ang2 (200 ng/ml) promotes tubule formation, migration, and cell survival in EPCs (*p<0.01 vs. PBS control, n=6). B. Angiogenic activity of Ang2 shown in EPCs was not observed in mature endothelial cells. In vitro angiogenic assays were performed with HUVECs (gray) under the same condition as described for EPC (black) in A. Their responsiveness to angiopoietins or VEGF was compared by normalizing their values relative to those of the corresponding untreated PBS control. Asterisks indicate a significant difference from the values for EPCs and HUVECs (*p<0.01, n=6).

3.3. Angiogenic activity of Ang2 in EPCs is mediated via Tie2-related signaling pathway

To elucidate whether this Ang2-mediated angiogenic response of EPCs would occur through Tie2 receptor and/or other signaling pathways, we performed the same in vitro angiogenesis assays with EPCs whose Tie2 receptors were blocked by pretreatment with anti-Tie2 antibody. As shown in Fig. 2A, blocking Tie2 receptors abolished the Ang2-induced angiogenic responses in EPCs, indicating that Ang2 stimulates tube formation, chemotactic migration, and cell survival in EPCs via Tie2 receptors. To further confirm the interaction between Ang2 and the Tie2 receptor in EPCs, western blotting was performed to determine whether Tie2 and related downstream signaling molecules were activated upon Ang2 treatment. Fig. 2B shows that Tie2 receptors were rapidly phosphorylated after Ang2 exposure, and Akt and eNOS were also subsequently phosphorylated within 60 min. Indeed, Fig. 2C showed that activation of Akt and eNOS is required for Ang2-mediated angiogenic signaling in EPCs, in that the Ang2-induced angiogenic responses in EPCs were completely abrogated in the presence of selective Akt inhibitor or l-NAME (NOS inhibitor). These results indicate that Ang2 truly acts as a Tie2 agonist in EPCs, and that Akt and eNOS, well-known downstream molecules regulating the Ang1-Tie2 signaling pathway, were also involved in the Ang2-mediated cellular responses of EPCs. However, treatment of HUVECs with Ang2 at the same concentration (200 ng/ml) did not induce phosphorylation of Tie2, Akt, or eNOS, whereas treatment with a tenfold greater dose of Ang2 (2 μg/ml) moderately activated Tie2 and Akt. This result suggests that EPCs possess higher responsiveness to Ang2 than do HUVECs, in that EPCs activate the Tie2 signaling pathway and exhibit angiogenic responses at a low concentration of Ang2 where HUVECs do not.

Fig. 2

Angiogenic activity of Ang2 in EPCs is mediated via Tie2/Akt/eNOS signaling pathway. A. Angiogenic activity of Ang2 occurs through the Tie2 receptor in EPCs. With EPCs pretreated with anti-Tie2 or anti-CD34 antibody, in vitro angiogenic assays were performed as described above (*p<0.01 vs PBS control; #p<0.01 vs Ang2, n=6). B. Ang2 (200 ng/ml) activates Tie2/Akt/eNOS signaling pathway in EPCs, but not in HUVECs. After serum deprivation, cells were exposed to Ang2 for 15 min (Tie2) or 60 min (Akt, eNOS). Cell lysates were subjected to western blotting using antibodies against phospho-Tie2, phospho-Akt, and phospho-eNOS. After being stripped, the membranes were reprobed with antibody against Tie2, Akt, and eNOS for the estimation of total protein loaded. Similar results were obtained in three additional experiments. C. Angiogenic activity of Ang2 requires the activation of Akt and eNOS. With EPCs pretreated with PBS, Akt inhibitor (1 μM for tube formation, 4 μM for apoptosis and migration), or l-NAME (6 mM) 2 h before addition of PBS or Ang2 (200 ng/ml), in vitro angiogenic assays were performed as described above (*p<0.05 vs PBS control; #p<0.05 vs Ang2, n=5).

3.4. EPCs are different from HUVECs in the physical association between Tie1 and Tie2

In order to understand the differential responsiveness to Ang2, we examined the expression levels of the Tie2 receptor in EPCs and HUVECs. As shown in Fig. 3A, the Tie2 receptor expression in EPCs was found to be similar to that in HUVECs. Furthermore, the expression levels of Tie1, an orphan receptor, but with high homology to Tie2, were also similar between EPCs and HUVECs. Although it was unexpectedly found that there was no difference in the total amount of both Tie receptors, interaction between Tie receptors was presumed to be responsible for the differential responsiveness to Ang2 in EPCs. When assayed for their physical associations by immunoprecipitation with anti-Tie2 antibody and immunoblotting (Fig. 3B), a substantial amount of Tie2 receptors in HUVECs was present as heterocomplexes with Tie1, whereas most Tie2 receptors in EPCs were found to be present unbound to Tie1.

Fig. 3

EPCs are different from HUVECs in the physical association between Tie1 and Tie2 receptors. A. There are no differences in the expression levels of Tie receptors between EPCs (E) and HUVECs (H). Both Tie1 and Tie2 receptor levels were analyzed using western blotting of whole cell lysates from EPCs and HUVECs. B. Most Tie2 receptors in HUVECs (H) exist as heterocomplexes with Tie1, whereas those in EPCs (E) are present unbound to Tie1. After crosslinking with DTSSP, cells were lysed and immunoprecipitated with anti-Tie2 antibody. The immunoprecipitates were subjected to western blotting with antibodies against Tie1 and Tie2. For A and B, data are representative of five independent experiments.

3.5. Tie1/Tie2 heterocomplex inhibits Ang2-mediated Tie2 activation

To further substantiate that Tie1/Tie2 heterocomplexes are associated with Ang2-mediated Tie2 activation, HUVECs were transfected with Tie1-specific siRNA in order to knock down the Tie1 expression, which substantially diminished the extent of Tie1/Tie2 heterocomplexes, and left most Tie2 receptors in HUVECs unbound to Tie1, just like those in EPCs (Fig. 4A,B). When these Tie1-specific siRNA treated HUVECs were exposed to a low dose of Ang2 (200 ng/ml), Ang2 promoted a significantly high degree of phosphorylation for Tie2 and Akt in Tie1-silenced HUVECs, as observed in EPCs (Fig. 4C). These data indicate that the physical association of Tie1 with Tie2 receptor plays an inhibitory role in Ang2-mediated Tie2 activation in endothelial cells.

Fig. 4

Silencing Tie1 induces the Ang2-mediated Tie2 activation in HUVECs. A and B. Knock down of the Tie1 receptor substantially diminished the extent of Tie1/Tie2 heterocomplex, leaving most Tie2 receptors unbound to Tie1. HUVECs were transfected with two different Tie1-specific siRNAs (Tie1-I, Tie1-II) or unrelated control siRNA (control). They were subjected to RT-PCR (A) or immunoprecipitation/western blotting (B) in order to determine the expression level of Tie1 and the extent of Tie1/Tie2 heterocomplexes respectively. C. Ang2 phosphorylates the Tie2 signaling pathway in Tie1-silenced endothelial cells, as observed in EPCs. HUVECs were transfected with unrelated control siRNA (control) or Tie1-specific siRNA (Tie1-I). After the overnight starvation and subsequent stimulation with 200 ng/ml of Ang2 for 15 min or 60 min, cell lysates were subjected to western blotting using anti-phospho-Tie2 and -phospho-Akt antibodies. After being stripped, the membranes were reprobed with antibody against Tie2 and Akt for the estimation of total protein loaded. Data are representative of three independent experiments.

3.6. Ang2 enhances EPC-mediated neovascularization in vivo

To investigate the proangiogenic effect of Ang2 on EPC-mediated neovascularization in vivo, ECM gels containing EPCs were implanted underneath adenoviral Ang2 or LacZ-injected skin layers of immunodeficient nude mice. The enzyme-linked immunosorbent assay showed that adenoviral-Ang2 transduced tissues produced high levels of Ang2 (3.97 ng/mg of protein on the third day; 16.68 ng/mg of protein on the seventh day after injection), presumably supplying Ang2 into adjacent EPCs-containing ECM gels. After two months, ECM gels were harvested and analyzed using hematoxylin and eosin (H & E) staining and anti-human CD31 immunohistochemistry (Fig. 5A,B). As shown in Fig. 5C, EPCs in ECM gels formed higher numbers of vessel-like structures in the adenoviral-Ang2 transduced tissues than in the adenoviral-LacZ transduced tissues. When examined for the presence of erythrocyte-filled blood vessels, not all of the human EPC-incorporated vessel-like structures contained intraluminal erythrocytes. This implies that overexpression of Ang2 alone might be enough for EPCs to form vessel-like structures, but not sufficient for mature host vasculature to undergo vascular remodeling and interconnect with EPCs in ECM gels. Indeed, murine VEGF levels in ECM-implanted sites (20.66±6.87 pg/mg of protein on the third day; 17.68±2.81 pg/mg of protein on the seventh day after implantation) were similar to its basal level found under normoxia (16.49±5.68 pg/mg of protein). This was because the ECM gels were implanted under the skin without damaging murine vasculature, which allowed us to rule out the influence from various angiogenic factors known to be upregulated under hypoxic conditions, thereby avoiding any mis-interpretation in determining the Ang2 effect on EPCs. As a negative control, ECM gels without EPCs were implanted into adenoviral-Ang2 injected tissues, showing only a few vessel-like structures lined with murine endothelial cells (0–2/high power field).

Fig. 5

Overexpression of Ang2 enhances EPC-mediated neovascularization in vivo. H & E (A) and anti-human CD31 (B) staining of collagen/fibronectin (ECM) gel without EPCs, ECM gel with EPCs in adenoviral LacZ-infected tissues, and ECM gel with EPCs in adenoviral Ang2-infected tissues. Scale bar, 50 μm. C. The numbers of EPC-incorporated vessel-like structures were increased in adenoviral Ang2-infected tissues, compared with adenoviral LacZ-infected tissues. Paraffin sections of ECM gels were sampled at 100 μm intervals over a 1 mm range and scored for the presence of human CD31-positive endothelial cell-lined vessel-like structures. Asterisk indicates significant difference between the values for the adenoviral LacZ and adenoviral-Ang2 groups (*p<0.05, n=11).

4. Discussion

The first aim of the present study was to investigate the Ang2-Tie2 signaling pathway in EPCs, where Ang2 was found to promote their angiogenic properties through Tie2 phosphorylation and the related Akt/eNOS downstream signaling, as have been defined for Ang1. This finding is distinct from the previously proposed role of Ang2 in the vasculature, i.e., an Ang1 inhibitor by preoccupying the Tie2 receptor, but not directly stimulating receptor phosphorylation by itself [11,17]. However, recent findings demonstrate that roles of Ang2 in vascular biology are more complex than previously recognized. When present for prolonged period or at high concentration, Ang2 alone induces Tie2 phosphorylation in endothelial cells and enhances their angiogenicity [18–20]. In addition, Ang2 was recently found to act as an agonist for Tie2 receptors during lymphatic development, and its overexpression was shown to stimulate the retinal angiogenesis in early postnatal ocular neovascularization process [13]. Such a dose- and cell-dependent Tie2 activation property of Ang2 was also observed in our study, wherein EPCs required a relatively low dose of Ang2 (200 ng/ml) for Tie2 phosphorylation and subsequent increase in angiogenicity, whereas HUVECs needed approximately a tenfold greater dose of Ang2 (2 μg/ml), even for modest Tie2 activation. However, the mechanism by which Ang2 displays such a differential activity in Tie2 activation remains poorly understood. In this regard, the second aim of this study was set to investigate what modulates the Tie2 phosphorylation activity of Ang2 by comparing EPCs with HUVECs in terms of their expression levels of Tie2 and its interaction with other molecules.

The cell-specific activity of Ang2 might be controlled by differential responsiveness of the target cells, probably because of the difference in the Tie2 receptor. In particular, it is presumed that the Tie2 receptor should be multimerized for its activation, as with other receptor tyrosine kinases, so that the higher expression level of Tie2 is more likely to facilitate receptor multimerization and subsequent phosphorylation at the same concentration of ligand [21,22]. In this regard, we performed western blotting assays, but found unexpectedly that there was no significant difference in the expression levels of Tie2 receptors between EPCs and HUVECs. One possible alternative mechanism is that interaction of Tie2 with other co-receptors or co-regulators might somehow modulate the Ang2-mediated Tie2 activation. This hypothesis is supported by the observation that Ang2 induces Tie2 phosphorylation in Tie2-transfected nonendothelial cells, while antagonizing native Tie2 receptors on cultured endothelial cells [23–25]. It implies that an endothelial-specific receptor interacting with Tie2 may be responsible for the agonistic or antagonistic functions of Ang2. In this context, recent findings regarding the existence of Tie1/Tie2 heterocomplexes in endothelial cells, led us to speculate that interactions between Tie1 and Tie2 receptors might explain the differential responsiveness to Ang2 in between EPCs and HUVECs [26,27]. When assayed for the physical association between Tie receptors, a substantial amount of Tie2 receptors in HUVECs existed as heterocomplexes with Tie1, whereas most Tie2 receptors in EPCs were present unbound to Tie1. Given that Ang2 phosphorylates the Tie2 receptor on Tie2-transfected, but Tie1-deficient, nonendothelial cells, our results suggest that the endothelial-specific Tie1 receptor might play an inhibitory role in Ang2-mediated Tie2 activation by forming heterocomplexes with Tie2 receptors. This assumption was attempted to verify via preventing the formation of Tie1/Tie2 heterocomplexes in HUVECs by knocking down the Tie1 expression using siRNA. In this Tie1 silencing experiment, it was found that reduced physical association between Tie1 and Tie2 receptors resulted in a robust phosphorylation of Tie2 receptors and a subsequent activation of Tie2 downstream signaling molecules even at a low concentration of Ang2 where native Tie2 receptor in HUVECs won't be activated. These data confirm that the formation of Tie1/Tie2 heterocomplex prevents Tie2 receptors from being activated upon Ang2 exposure, and suggest that unbound Tie2 might be the receptor undergoing Ang2-mediated autophosphorylation. In other words, the extent of unbound Tie2 receptor might determine the degree of Ang2-mediated Tie2 activation, so that EPCs with more unbound Tie2 receptors might require lower concentration of Ang2 in activating Tie2 receptor, when compared with HUVECs with fewer unbound Tie2 receptors.

In conclusion, we believe our result showing proangiogenic Ang2-Tie2 signaling in EPCs will contribute to a better understanding of the mechanisms of EPC-mediated neovascularization, and have implications for the pathophysiology of increased Ang2 levels in ischemic tissues. From a therapeutic point of view, such a functional characterization of EPCs has gained increasing importance, as we have limited source and amount of EPCs for clinical application. To maximize the therapeutic effects in injured tissues with minimal number of EPCs, it is important to understand how EPC-mediated neovascularization is regulated at sites of ischemia. Furthermore, this study provides the first explanation for the mechanism behind context-dependent activity of Ang2 in endothelial cells, which is that the Tie1 receptor bound to Tie2 plays an inhibitory role in the Ang2-mediated Tie2 signaling pathway. Although Tie1 shares significant homology with Tie2 and is required for the normal embryonic vascular development, its function in postnatal neovascularization has been very much a matter of debate [1]. In this regard, the present result is very suggestive for the role of Tie1 receptor in regulating the vascular sensitivity toward Ang2 during postnatal neovascularization. However, there are still a lot of unresolved questions, such as the mechanism by which Tie1 receptor inhibits the Ang2-induced Tie2 activation when forming an heterocomplex. Given that Ang2 acts as a competitive inhibitor of Ang1 in HUVECs, we suspect that Tie1/Tie2 heterocomplex may not influence on the binding between Ang2 and Tie2 at extracellular levels, but could affect the subsequent conformational change of Ang2-bound Tie2 receptors, somehow preventing the Tie2 phosphorylation and subsequent activation of its related signaling molecules at cytoplasmic levels.


This work was supported by a National Research Laboratory Grant from the Korea Institute of Science and Technology Evaluation and Planning (M1-0412-00-0048), the Samsung Biomedical Research Institute Grant (SBRI; C-A5-104-2), and a Science Research Center Grant (Molecular Therapy Research Center).

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.cardiores.2006.08.002.


  • 1 These corresponding authors equally contributed to this work.

  • Time for primary review 24 days


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
View Abstract