Evidence of Human Thrombomodulin Domain as a Novel Angiogenic Factor
Background— Thrombomodulin is an anticoagulant, endothelial-cell-membrane glycoprotein. A recombinant thrombomodulin domain containing 6 epidermal growth factor–like structures exhibits mitogenic activity. This study explored the novel angiogenic effects of the recombinant domain using in vitro and in vivo models.
Methods and Results— Human recombinant thrombomodulin containing 6 epidermal growth factor–like structures (TMD2) and TMD2 plus a serine and threonine-rich domain (TMD23) were prepared using the Pichia pastoris expression system. Combined with purified TMD2 or TMD23, thrombin effectively activated protein C. TMD23 had higher activity than TMD2 in stimulating DNA synthesis in cultured human umbilical vein endothelial cells. Additionally, TMD23 stimulated chemotactic motility and capillarylike tube formation in human umbilical vein endothelial cells, an effect mediated through phosphorylation of extracellular signal–regulated kinase 1/2 and p38 mitogen-activated protein kinase and the phosphatidylinositol-3 kinase/Akt/endothelial nitric oxide synthase pathway. TMD23 also stimulated endothelial cell expression of matrix metalloproteinases and plasminogen activators, which mediated extracellular proteolysis, leading to endothelial cell invasion and migration during angiogenesis. Furthermore, TMD23-containing implants in rat cornea induced ingrowth of new blood vessels from the limbus. With the murine angiogenesis assay, TMD23 not only induced neovascularization coinjected with Matrigel and heparin but also enhanced angiogenesis in Matrigel containing melanoma A2058 cells in nude mice.
Conclusions— The recombinant thrombomodulin domain TMD23 enhanced the angiogenic response in vitro and in vivo, suggesting that thrombomodulin fragments may play a role in the formation of new vessels. These findings may provide a new therapeutic option for treating ischemic diseases.
Received July 20, 2004; revision received December 20, 2004; accepted December 21, 2004.
Thrombomodulin (TM) is an endothelial-cell-membrane glycoprotein that forms a complex with thrombin, converting the latter from procoagulant to an anticoagulant enzyme.1 Formation of the thrombin-TM complex inhibits thrombin from catalyzing conversion of fibrinogen to fibrin while allowing thrombin-mediated conversion of protein C into activated protein C. The latter, in turn, inactivates factors Va and VIIIa, slowing the blood coagulation cascade.1,2
Structurally, human TM consists of a single polypeptide chain with 5 distinct domains: an NH2-terminal lectinlike region designated D1, which comprises Ala1 through Asp226; a domain with 6x epidermal growth factor (EGF)–like structures (Cys227 through Cys462) designated D2; an O-glycosylation site–rich domain (Asp463 through Ser497) designated D3; a transmembrane domain consisting of (Gly498 through Leu521) designated D4; and a cytoplasmic tail domain (Arg522 through Leu557) designated D5.2 The fourth through sixth EGF-like structures of D2, in which the binding sites for protein C and thrombin are located, are responsible for the anticoagulant activity of TM.2
Although TM is well recognized as a physiologically important natural anticoagulant in vascular endothelial cells, wider physiological functions are implied because of its wide extravascular expression.3 Ablation of the TM gene in mouse embryos causes early postimplantation death before a functional cardiovascular system is established.4 Soluble forms of TM derived from extracellular TM domains found in plasma and urine5,6 have been recognized as markers of endothelial cell injury.7 Indeed, clinical studies have demonstrated an inverse correlation between plasma levels of soluble TM and coronary artery disease and atherosclerosis.8 A recombinant D2 domain containing 6 TM EGF-like structures exhibits mitogenic activity for Swiss 3T3 cells.9 In addition, the lectinlike domain endows TM with antiinflammatory properties in addition to roles in coagulation and fibrinolysis.10
In previous work, we demonstrated that the lectinlike domain of both TM-transfected melanoma A2058 cells and the endogenous TM-expressing keratinocyte HaCaT cell line mediates cell-to-cell adhesion by Ca2+-dependent binding to specific ligands on neighboring cells.11 We also observed that the lectinlike domain–deleted form of TM-transfected A2058 cells had a significantly higher growth rate than full-length TM-transfected cells. We suggested that an increase in cell-to-cell adhesion on TM expression may slow tumor growth in vivo.11 Therefore, TM may have multifunctional potential in modulating cellular responses; further investigation is warranted. In the present work, recombinant TMD2 and TMD23 were prepared, and possible roles in cell proliferation, migration, and promotion of angiogenesis were investigated.
Expression and Purification of Recombinant TMD Proteins
The pPICZαA vector (Invitrogen Corp) was used for expression and secretion of recombinant human TMD2 and TMD23 in the Pichia pastoris protein expression system. Briefly, DNA fragments coding for TMD2 (residues Ala224 through Cys462) and TMD23 (residues Ala224 through Ser497) were obtained by a polymerase chain reaction of human umbilical vein endothelial cell (HUVEC) cDNA using primers as described previously.12 Fermentation medium containing expressed TMD2 or TMD23 was applied to a nickel-chelating Sepharose column (Amersham Pharmacia Biotech AB), and TMD-containing fractions were eluted with imidazole.
TM Activity Assay
Cofactor activity of TMD2 and TMD23 for thrombin-dependent protein C activation was measured as described previously.11
SDS-PAGE was performed with Laemmli’s procedure13 using a 12.5% separating gel under reduced conditions.
Quantification of Recombinant TMD Proteins
TMD protein concentration was determined using the BCA kit (Pierce) with BSA as standard.
Amino acid sequence determinations were performed automatically by Edman degradation with a model 477A sequencer (Applied Biosystems).
Endothelial cells were isolated from human umbilical cord veins with previously described methods,14 and cells from the second passage were used in all experiments.
Cell Proliferation Assay
To evaluate the effect of TMD proteins on DNA synthesis in HUVECs, we performed a 5-bromo-2′-deoxyuridine (BrdU) incorporation assay using a commercial quantification kit (Roche Diagnostic GmbH) and following manufacturer’s protocol. Mitogenic effects induced by TMD proteins were also determined in the presence of polyclonal rabbit anti-TMD23 antibody (prepared by GlycoNex, Inc) purified by protein G–Sepharose column (Amersham Pharmacia Biotech AB).
The chemotactic motility of HUVECs was assayed using Transwell (Costar) with 6.5-mm-diameter polycarbonate filters (8-μm pore size). The lower filter surface was coated with 10 μg gelatin (Sigma-Aldrich). Various concentrations of TMD23 in the absence or presence of 10 μg/mL polyclonal rabbit anti-TMD23 antibody or 10 ng/mL vascular endothelial growth factor (VEGF) (R&D Systems, Inc.) in M199 were placed in the lower wells. Cell suspension (100 μL) containing 1×105 cells was loaded into each upper well. The chamber was incubated at 37°C for 4 hours. Cells were fixed with methanol and stained with 4′, 6′-diamidino-2-phenylindole (DAPI). Chemotaxis was quantified with optical microscopy (Leica) at ×100 magnification using MetaMorph imaging software (Universal Imaging Corp) by counting the cells that migrated to the lower side of the filter. Five random fields in each well were counted. Each sample was assayed in triplicate, and assays were repeated 3 times. To determine the role of extracellular signal–regulated kinase 1/2 (ERK1/2), p38 mitogen-activated protein kinase (MAPK), and phosphatidylinositol 3-kinase (PI3 kinase)–protein kinase B/Akt (Akt) in TMD23-mediated cell migration, assays were performed in the presence of U0126, a MAPK/ERK kinase (MEK) inhibitor; SB 203580, a p38 MAPK inhibitor; and LY 294002, a PI3 kinase inhibitor. Cell viability determined by trypan blue staining was not altered by treatment. U0126 was purchased from Promega, with SB 203580 and LY 294002 from Calbiochem.
Assay of ERK1/2, p38 MAPK, Akt, and Endothelial Nitric Oxide Synthase Phosphorylation
HUVECs were cultured to confluence in a 6-cm-diameter dish and incubated in M199 containing 1% FBS for 18 hours. Cells were washed and incubated with serum-free M199 for 6 hours and treated with the indicated concentration of TMD23. Cell lysates were separated by SDS-PAGE, and the levels of phospho-ERK1/2 (Tyr 204) and total ERK1/2, phospho-p38 (Thr-180/Tyr-182) and p38, phosphor-Akt (Ser-473) and Akt, and phospho–endothelial nitric oxide synthase (eNOS) (Ser-1173) and eNOS were detected by Western blotting with specific antibodies (Cell Signaling Technology).
Gelatin and Casein Zymographies and Plasminogen Activator Inhibitor Activity Assay
HUVECs were seeded at 80% confluence on 100-mm dishes in M199 medium. After 24 hours, cells were rinsed twice with serum-free M199 and incubated in serum-free M199 with various concentrations of TMD23, VEGF (10 ng/mL), or PMA (70 ng/mL) for 12 hours. The conditioned medium containing 5 μg secreted proteins was analyzed by gelatin- and casein-based zymographies.15 The digested area appeared clear on a blue background, indicating the location of matrix metalloproteinase (MMP) activity. To further verify MMP caseinolytic activity, the casein-containing gel was incubated with incubation buffer containing 10 mmol/L EDTA, an inhibitor of MMPs, to observe disappearance of the clear band. Plasminogen activator inhibitor (PAI) activity was measured by titrating samples with increasing amounts of tissue-type plasminogen activator (Boehringer Ingelheim) into a fixed volume of endothelial cell–conditioned media.14
In Vitro Matrigel Angiogenesis Assay
Growth factor–reduced Matrigel (BD Biosciences) was thawed at 4°C. A total of 200 μL Matrigel was added to wells of a 48-well plate and polymerized for 30 minutes at 37°C. Cell suspensions (200 μL) containing various concentrations of VEGF, TMD23, or TMD23 in the presence of polyclonal rabbit anti-TMD23 antibody (10 μg/mL) or various signaling pathway inhibitors were plated on Matrigel-coated wells at a density of 2.4×104 cells per well in M199 containing 10% FBS for 24 hours at 37°C in a 5% CO2 humidified atmosphere. Each concentration was tested in triplicate in the same plate, and wells were photographed with a Leica camera (×40 magnification). Tube length was quantified by measuring tubes in 5 randomly chosen fields from each well with MetaMorph software. Experiments were repeated ≥3 times.
Rat Corneal Angiogenesis Assay
The rat corneal assay was performed as described elsewhere.16 Uniformly sized pellets of Hydron (polyhydroxyethylmethacrylate; Sigma-Aldrich) containing either TMD23 or basic fibroblast growth factor (bFGF) and sucralfate (Sigma-Aldrich) were implanted into the rat corneal stroma adjacent to the temporal limbus. Briefly, sterile saline suspensions containing 5 to 40 μg TMD23 or bFGF plus 10 mg sucralfate were made and speed vacuumed for 5 minutes; 12% Hydron in ethanol was added as described.16 The suspension was deposited onto an autoclaved, sterilized, 15×15-mm nylon mesh. In some preparations, polyclonal anti-TMD23 IgG (10 μg/mL) was included during pellet preparation to specifically inhibit TMD23 activity. The eyes of Sprague Dawley rats (250 to 300 g) were topically anesthetized, and a single pellet was inserted into a surgically created pocket in the corneal stroma. Corneas were examined daily with a dissecting microscope for up to 24 days for capillary growth. Maximal vessel length, clock hour, and vascular areas of corneas of all rats were measured with a slit-lamp stereomicroscope.
Murine Angiogenesis Assay
To assess angiogenic effects in vivo,17 growth factor–reduced liquid Matrigel (0.5 mL) containing heparin (60 U/mL) and various concentrations of TMD23, bFGF, or VEGF was subcutaneously injected into FVB mice near the abdominal midline. Matrigel with PBS plus heparin served as the negative control, with bFGF and VEGF plus heparin as the positive control. Four days after injection, mice were euthanized, and the Matrigel plugs were surgically removed. For macroscopic analysis of angiogenesis, hemoglobin content in Matrigel was measured with Drabkin reagent 525 (Sigma-Aldrich). For histological analysis, rat anti-mouse CD31 (platelet endothelial cell adhesion molecule-1 [PECAM-1]) antibody (Pharmingen) was used. The same specimens were used for hematoxylin and eosin staining.
Effect of TMD23 on Tumor Neoangiogenesis
To assess TMD23 effects on tumor neoangiogenesis, 1×106 human melanoma A2058 cells (ATCC CRL-11147) were mixed with Matrigel (0.5 mL) containing heparin (60 U/mL) and various concentrations of TMD23 and subcutaneously injected into athymic nude mice. Matrigel plugs were harvested after 18 days and frozen in liquid nitrogen. Blood vessels were detected by immunohistochemistry using anti-CD31 antibody as described above. Microvessel density was calculated by counting 6 highly vascularized fields in each section (×200 magnification; MetaMorph software).
Animal care conditions and design of experiments were approved by the Institutional Animal Care and Use Committee of the National Cheng Kung University (Tainan, Taiwan).
Data are expressed as mean±SD. Statistical significance was analyzed by unpaired Student t test. Differences between >2 groups was compared by 1-way ANOVA, followed by Bonferroni post hoc test, with values of P<0.05 considered statistically significant.
Expression and Purification of Recombinant TMD2 and TMD23
Recombinant TMD2 (Ala224 through Cys462) containing 6 EGF-like structures (D2) and TMD23 (Ala224 through Ser497) containing D2 and the serine and threonine (Ser/Thr)–rich domain D3 (molecular masses of 47 and 50 kDa, respectively) were prepared with the Pichia pastoris expression system. TMD2 or TMD23 purified from fermentation medium was homogeneous as judged by SDS-PAGE and Western blotting with monoclonal anti-TM antibody (Figure 1). NH2-terminal sequence analysis demonstrated that TMD2 and TMD23 had an expected single sequence starting at the fusion peptide, Glu-Phe, followed by the sequence TM Ala224. Comparison with the published sequence of human TM18 revealed a T to C change at nucleotide 1418 and a corresponding amino acid change at residue 455 (amino acid 234 in TMD2 and TMD23) from Val to Ala. Cofactor activity of TMD23 for thrombin-dependent protein C activation was 2650 pmol of APC formed per minute per 1 mg TMD23, which was 25% higher than that of TMD2 on an equal-molar basis and higher than the value previously reported for recombinant TMD2.9
Effect of TMD2 and TMD23 on HUVEC Proliferation
We showed that TMD2 and TMD23 stimulated DNA synthesis in HUVECs and that TMD23 had higher mitogenic activity than TMD2 (Figure 2A). TMD23 at a concentration of 4.5 nmol/L (150 ng/mL) stimulated HUVEC proliferation by ≈2-fold (Figure 2A). Treatment with polyclonal anti-TMD23 IgG markedly attenuated both TMD2- and TMD23-induced cell proliferation (Figure 2A), consistent with a TM-specific mitogenic effect.
TMD23 Induces HUVEC Migration
The effect of TMD23 on migration of HUVECs was evaluated with the Transwell assay. As shown in Figure 2B, TMD23 markedly induced a dose-dependent chemotactic response in HUVECs. The migratory activity at 10 ng/mL TMD23 was increased by 80% over the control, with the effect of TMD23 comparable with that of VEGF (10 ng/mL), a known stimulator of HUVEC migration. The stimulatory effect of TMD23 on HUVEC migration was blocked by U0126, SB 203580, LY 294002, and a polyclonal antibody against TMD23 (Figure 2B). This result demonstrated that TMD23 promoted endothelial cell migration, perhaps through ERK1/2, p38, and PI3 kinase/Akt signaling pathways.
TMD23 Induces ERK1/2, p38 MAPK, Akt, and eNOS Phosphorylation
The ERK pathway, which is activated by many growth factors, may mediate endothelial cell proliferation and migration. It has also been reported that p38 MAPK activation by VEGF mediates actin reorganization and cell migration in HUVECs and thus may be an important regulator of angiogenesis.19 Furthermore, it has been demonstrated that the PI3 kinase/Akt/eNOS signaling axis plays critical roles in angiogenesis.20 To validate whether ERK1/2, p38, Akt, and eNOS activity is involved in TM-induced HUVEC proliferation and migration, we studied the effect of TMD23 on ERK1/2, p38, Akt, and eNOS phosphorylation by Western blotting. As shown in Figure 3A, TMD23 dose dependently induced ERK1/2, p38, Akt, and eNOS phosphorylation in HUVECs, whereas total ERK, p38, Akt, and eNOS proteins remained constant. Induction of phosphorylation by TMD23 was time dependent, with maximal phosphorylation occurring at ≈30 minutes (ERK, p38, and Akt) and 120 minutes (eNOS) (Figure 3B).
TMD23 Induces MMP Expression and Reduces PAI Activity in HUVECs
MMPs secreted by various cells, including endothelial cells, are responsible for degradation of extracellular matrix components, thus facilitating cell migration. Because endothelial cell migration was stimulated by TMD23, we examined whether TMD23 could induce production of MMPs. Gelatin and casein zymographies were performed to examine the activities of MMP-1, MMP-2, MMP-3, and MMP-9. TMD23 had no significant effect on the expression of MMP-2 and MMP-9 but stimulated expression of MMP-1/3 as shown by gelatin zymography (Figure 4A). Increased MMP-3 expression was further confirmed by caseinolytic zymography (Figure 4B), which was blocked specifically in the presence of EDTA (Figure 4C).
Because MMP and plasminogen activator–plasmin systems play important roles in angiogenesis,21 we measured the PAI activity of conditioned TMD23–treated cell medium. PAI activity in conditioned media of HUVECs treated with 100 or 1000 ng/mL TMD23, 70 ng/mL PMA, or 10 ng/mL VEGF for 12 hours decreased significantly (Figure 4D), suggesting that both angiogenic factors, TMD23 and VEGF, induced plasminogen activator release into the conditioned media. This result was consistent with previous observations that VEGF induces urokinase-type and tissue-type plasminogen activator in endothelial cells.22,23
TMD23 Promotes Vascular Tube Formation on Matrigel
The effect of TMD23 on morphological differentiation of HUVECs on Matrigel was evaluated. Control culture developed small amounts of tube formation and incomplete networks within 24 hours (Figure 5A). HUVECs stimulated with VEGF (10 ng/mL) as positive control or TMD23 (5 to 250 ng/mL) formed elongated capillarylike structures, with complete networks observed by 24 hours (Figure 5B through 5E). The stimulatory effect of TMD23 on HUVEC tube formation was blocked by polyclonal TMD23 IgG (10 μg/mL), 10 μmol/L U0126, 10 μmol/L SB 203580, 10 μmol/L LY 294002, and 10 μmol/L NG-nitro-l-arginine methyl ester (an inhibitor of NO production) (Figure 5F). Counting tubule length per well showed that TMD23 clearly stimulated tube formation in dose-dependent manner (Figure 5G). The result of this in vitro assay indicates that TMD23 has a novel vasculogenic or angiogenic activity mediated through ERK1/2, p38, and PI3 kinase/Akt/eNOS signaling pathways.
TMD23-Induced Angiogenesis in Rat Corneas
To further investigate TMD23 angiogenic activities in vivo, TMD23-induced rat corneal neovascularization was studied. Samples were implanted in avascular rat cornea to allow blood vessel ingrowth from the limbus. Control pellets without growth factors did not induce an angiogenic response (Figure 6A and 6C). Conversely, implants containing TMD23 induced dose-dependent and time-dependent corneal neovascularization (Figures 6B and 6C). Maximal response to TMD23 was observed with a 90-ng dose on day 10, and this was slightly greater than the response to 90 ng bFGF (Figure 6C). TMD23-induced angiogenic response regressed gradually after day 10, with complete abrogation of response by day 24. To test whether the angiogenic response was dependent on recombinant protein, heat-inactivated TMD23 (100°C for 30 minutes) was assayed for angiogenic efficacy. Residual protein C activity of heat-inactivated TMD23 was 40% that of control without heat treatment, suggesting that the recombinant TMD23 protein was as stable as the natural TM isolated from human placenta or bovine lung.24,25 The magnitude of neovascularization response stimulated by 90 ng of heated TMD23 sample was about the same as that obtained with 45 ng of unheated TMD23 (Figure 6C). In addition, implants containing 90 ng TMD23 plus polyclonal anti-TMD23 IgG gave no positive responses (n=6), suggesting that the angiogenic response is TMD23 specific.
TMD23 Induces Angiogenesis in the Matrigel Plug Assay In Vivo
To further evaluate TMD23 angiogenesis effects in vivo, liquid Matrigel containing various amounts of TMD23 (10, 100, or 1000 ng), bFGF (250 ng), VEGF (50 ng) plus heparin (30 U), or heparin alone was injected into FVB mice. Figure 7A shows the appearances of plugs recovered on day 4. The control plugs were pale, whereas plugs containing TMD23, bFGF, and VEGF plus heparin were bright red (Figure 7A). Intense vascularization was observed at all TMD23 concentrations and with the positive controls (Figure 7A). The angiogenic response was apparent in gels 2 to 3 days after injection, reaching a maximum at 3 to 5 days. To clarify the heparin effect on angiogenic factor–induced angiogenesis, Matrigel plugs containing TMD23, bFGF, or VEGF without heparin were assessed. Plugs containing TMD23, bFGF, and VEGF without heparin were pale (Figure 7A). The observations shown in Figure 7A were consistent with a heparin-dependent enhancement of the angiogenesis induced by TMD23, bFGF, and VEGF.
Histological examinations revealed cellular invasion. Capillarylike structures were found in Matrigel plugs with TMD23 and heparin (Figure 7B). Moreover, use of anti-mouse CD31 (PECAM-1) antibody revealed that capillarylike structures had been formed by mouse endothelial cells (Figure 7B). This observation was consistent with the idea that TMD23 could induce endothelial cell invasion and neovessel formation in vivo. The degree of angiogenesis was also determined by measuring the hemoglobin content of recovered Matrigel plugs. Matrigel plugs containing VEGF or TMD23 plus heparin had significantly increased hemoglobin content compared with controls (Figure 7C).
TMD23 Stimulates Neoangiogenesis in Tumors
Because angiogenesis is an essential step for many physiological and pathological processes, we assessed TMD23 effects on tumor neoangiogenesis. A2058 cells were mixed in Matrigel supplemented with TMD23 and heparin and injected subcutaneously into athymic nude mice. The tumor developed by Matrigel plugs containing A2058 cells and TMD23 (or bFGF) was bigger than that for A2058 cells alone (Figure 8A). Immunohistochemical staining of CD31 demonstrated significantly more blood vessels in Matrigel plugs containing TMD23 (or bFGF) (Figure 8B and 8C).
Neovascularization encompasses both angiogenesis and vasculogenesis. Vasculogenesis is defined as creation of blood vessels by differentiation of endothelial precursor cells, whereas angiogenesis represents formation of new capillaries from preexisting mature endothelial cells, a process including migration and proliferation of endothelial cells, extracellular matrix degradation, and capillary tube formation.26 Many molecules have been identified that induce vessel formation both in vitro and in vivo, including acidic FGF, bFGF, transforming growth factor-α and -β, hepatocyte growth factor, tumor necrosis factor-α, angiogenin, interleukin-8, and angiopoietins.27,28 Conversely, thrombospondin, platelet factor-4, angiostatin, arresten, restin, endostatin, and tumstatin have been identified as important endogenous antiangiogenic molecules.29–31 Protease cleavage of collagen types IV, XV, and XVIII generates arresten (and tumstatin), restin, and endostatin, respectively, whereas plasminogen proteolysis yields angiostatin. Several other molecules and their proteolytic cleavage products have the potential to exert opposing effects on angiogenesis. For example, certain parental proteins—including prolactin, growth hormone, and placental lactogen—can promote angiogenesis. After proteolytic processing, however, the same proteins generate peptide fragments that have antiangiogenic properties.32 Here, we report a novel angiogenic-promoting function of TM domains 2 and 3. Recombinant TM fragment TMD23 can stimulate endothelial migration, proliferation, and tube formation in vitro, a process that may be mediated via phosphorylation of ERK1/2, p38, Akt, and eNOS. We also found that TMD23 stimulates endothelial cell expression of MMPs and plasminogen activators that mediate extracellular proteolysis, leading to endothelial cell invasion and migration during the early stages of angiogenesis.21 Using a murine Matrigel model and rat corneal neovascularization assays, we documented a TMD23-dependent induction of angiogenesis in vivo. It may be worthwhile to examine whether TMD23 could become a novel therapeutic agent for some ischemia-related diseases.
The D3 of TM is rich in Ser and Thr, which can support posttranslational attachment of a chondroitin sulfate moiety.33 In the present study, we demonstrated that TMD23 was more effective than TMD2 in stimulating proliferation of HUVECs (Figure 2A) and in inducing in vivo angiogenesis using the Matrigel plug assay (data not shown). These results are in line with previous observations that a novel chondroitin sulfate/dermatan sulfate proteoglycan promotes hepatocyte growth factor mitogenic activity34 and that a chondroitin sulfate containing glycosaminoglycan enhances in vitro FGF-2–dependent angiogenesis.35
During the course of the murine angiogenesis assays, we noted that Matrigel supplemented with TMD23, bFGF, or VEGF produced a variable angiogenic reaction but that the magnitude of the reaction was considerably greater in gels that were supplemented with both angiogenic factors and heparin. These results are consistent with previous findings.17 Matrigel, a solution of basement membrane proteins isolated from the Engelbreth-Holm-Swarm tumor, can serve as a vehicle for slow release of angiogenic factors, with heparin promoting the binding, modulation, and sustained release of these factors.17,36 Still, heparin is not necessary to obtain the angiogenic response in the rat corneal neovascularization assay because the combination of angiogenic factors with sucralfate and incorporation into Hydron achieve stabilization of angiogenic factors and sustain release over a prolonged time.16
It has been demonstrated that cell surface proteoglycan syndecan-4 participates in mediating the effects of FGF-2 on cell function.37 It is possible that the biological activities of TM domains may be mediated by binding of the peptide to its specific site on the cell surface9; however, TM domains may interact with other membrane molecules such as syndecan or synectin, which can mediate or trigger the effects of TM on cell function, as in the case of FGF-2.37 Therefore, TM domains may act indirectly.
Expression of TM proteins has been documented in myriad tumors and cell lines.38 An inverse relationship of TM expression and rate of tumor-cell proliferation has been reported in hepatocellular carcinoma, ovarian carcinoma, and esophageal squamous carcinoma.39,40 It is possible that various soluble TM fragments released from the cell surface by in situ proteolytic enzymes—including plasminogen activators-plasmin, elastase, and MMPs—may function as paracrine substances to modulate angiogenesis in tumors. Soluble forms of TM derived from the extracellular domain of the molecule have been found in the plasma and urine of healthy subjects, suggesting that it is cleaved under physiological conditions.5,41 The increase in plasma TM fragments in patients with various diseases suggests that TM release from endothelial cells is accelerated by proteolytic activity generated on the surface of the endothelium.5 The proteases responsible for TM shedding are still unknown, although neutrophil-derived enzymes—including elastase, proteinase-3, and cathepsin G—have been implicated.42 Recently, it has been reported that TM is efficiently and specifically cleaved by RHBDL2, a mammalian rhomboid, suggesting that this cleavage may be physiologically significant.43 Several soluble TM subspecies that maintain the structure of the fourth to sixth EGF-like repeats have protein C–activating cofactor activity, although the physiological significance is obscure.41 Whether these natural soluble TM subspecies mimic the mitogenic or angiogenic properties noted with recombinant TMD23 remains to be studied. In the present study, we demonstrated that TMD23 can promote angiogenesis in vitro and in vivo. In the in vitro Matrigel assay, we also noticed that TMD23 could promote the morphological differentiation of HUVECs to form tubelike structures with well-connected network. Therefore, TMD23 may also have the potential to induce vasculogenesis. Nevertheless, whether TM domains can function as a vasculogenesis factor by promoting proliferation and mobilization of endothelial precursor cells from bone marrow to neovascularization sites needs to be further explored.
This work was supported by the Ministry of Education Program for Promoting Academic Excellence of Universities under grant 91-B-FA09–2-4 and by the National Science Council, grant NSC 92-2311-B-006-007, Taiwan, Republic of China. We thank Fong-Yi Chou, I-Hui Wu, Bi-Ing Chang, and Wan-Ting Tsai for expert technical assistance.
Esmon NL, Owen WG, Esmon CT. Isolation of a membrane-bound cofactor for thrombin-catalyzed activation of protein C. J Biol Chem. 1982; 257: 859–864.
Boffa MC, Burke B, Haudenschild CC. Preservation of thrombomodulin antigen on vascular and extravascular surfaces. J Histochem Cytochem. 1987; 35: 1267–1276.
Healy AM, Rayburn HB, Rosenberg RD, Weiler H. Absence of the blood-clotting regulator thrombomodulin causes embryonic lethality in mice before development of a functional cardiovascular system. Proc Natl Acad Sci U S A. 1995; 92: 850–854.
Takano S, Kimura S, Ohdama S, Aoki N. Plasma thrombomodulin in health and diseases. Blood. 1990; 76: 2024–2029.
Yamamoto S, MIizoguchi T, Tamaki T, Ohkuchi M, Kimura S, Aoki N. Urinary thrombomodulin, its isolation and characterization. J Biochem (Tokyo). 1993; 113: 433–440.
Salomaa V, Matei C, Aleksic N, Sansores-Garcia L, Folsom AR, Juneja H, Chambless LE, Wu KK. Soluble thrombomodulin as a predictor of incident coronary heart disease and symptom less carotid artery atherosclerosis in the Atherosclerosis Risk in Communities (ARIC) study: a case-cohort study. Lancet. 1999; 353: 1729–1734.
Hamada H, Ishii H, Sakyo K, Horie S, Nishiki K, Kazama M. The epidermal growth factor–like domain of recombinant human thrombomodulin exhibits mitogenic activity for Swiss 3T3 cells. Blood. 1995; 86: 225–233.
Conway EM, Van de Wouwer M, Pollefeyt S, Jurk K, Van Aken H, De Vriese A, Weitz JI, Weiler H, Hellings PW, Schaeffer P, Herbert JM, Collen D, Theilmeier G. The lectin-like domain of thrombomodulin confers protection from neutrophil-mediated tissue damage by suppressing adhesion molecule expression via nuclear factor kB and mitogen-activated protein kinase pathways. J Exp Med. 2002; 196: 565–577.
Huang HC, Shi GY, Jiang SJ, Shi CS, Wu CM, Yang HY, Wu HL. Thrombomodulin-mediated cell adhesion: involvement of its lectin-like domain. J Biol Chem. 2003; 278: 46750–46759.
Shi GY, Hau JS, Wang SJ, Wu IS, Chang BI, Lin MT, Chow YH, Chang WC, Wing LYC, Jen CJ, Wu HL. Plasmin and the regulation of tissue-type plasminogen activator biosynthesis in human endothelial cells. J Biol Chem. 1992; 267: 19363–19368.
Kenyon BM, Voest EE, Chen CC, Flynn E, Folkman J, D’Amato RJ. A model of angiogenesis in the mouse cornea. Invest Ophthalmol Vis Sci. 1996; 37: 1625–1632.
Shirai T, Shiojiri S, Ito H, Yamamoto S, Kusumoto H, Deyashiki Y, Maruyama I, Suzuki K. Gene structure of human thrombomodulin, a cofactor for thrombin-catalyzed activation of protein C. J Biochem. 1988; 103: 281–285.
Kawasaki K, Smith RS, Hsieh CM, Sun J, Chao J, Liao JK. Activation of the phosphatidylinositol 3-kinase/protein kinase Akt pathway mediates nitric oxide–induced endothelial cell migration and angiogenesis. Mol Cell Biol. 2003; 23: 5726–5737.
Pepper MS. Role of the matrix metalloproteinase and plasminogen activator-plasmin systems in angiogenesis. Arterioscler Thromb Vasc Biol. 2001; 21: 1104–1117.
Folkman J, Shing Y. Angiogenesis. J Biol Chem. 1992; 267: 10931–10934.
Cao R, Wu HL, Veitonmaki N, Linden P, Farnebo J, Shi GY, Cao Y. Suppression of angiogenesis and tumor growth by the inhibitor K1–5 generated by plasmin-mediated proteolysis. Proc Natl Acad Sci U S A. 1999; 96: 5728–5733.
Corbacho AM, Martinez de la Escalera G, Clapp C. Roles of prolactin and related members of the prolactin/growth hormone/placental lactogen family in angiogenesis. J Endocrinol. 2002; 173: 219–238.
Bourin MC, Lundgren AE, Lindahl U. Isolation and characterization of the glycosaminoglycan component of rabbit thrombomodulin proteoglycan. J Biol Chem. 1990; 265: 15424–15431.
Bechard D, Gentina T, Delehedde M, Scherpereel A, Lyon M, Aumercier M, Vazeux R, Richet C, Degand P, Jude B, Janin A, Fernig DG, Tonnel AB, Lassalle P. Endocan is a novel chondroitin sulfate/dermatan sulfate proteoglycan that promotes hepatocyte growth factor/scatter factor mitogenic activity. J Biol Chem. 2001; 276: 48341–48349.
Tapon-Bretaudiere J, Chabut D, Zierer M, Matou S, Helley D, Bros A, Mourao PAS, Fischer AM. A fucosylated chondroitin sulfate from echinoderm modulates in vitro fibroblast growth factor 2–dependent angiogenesis. Mol Cancer Res. 2002; 1: 96–102.
Horowitz A, Tkachenko E, Simons M. Fibroblast growth factor–specific modulation of cellular response by syndecan-4. J Cell Biol. 2004; 157: 715–725.
Miettinen M, Sarlomo-Rikala M. Expression of calretinin, thrombomodulin, keratin 5, and mesothelin in lung carcinomas of different types: an immunohistochemical analysis of 596 tumors in comparison with epithelioid mesotheliomas of the pleura. Am J Surg Pathol. 2003; 27: 150–158.
Tezuka Y, Yonezswa S, Maruyama I, Matsushita Y, Shimizu T, Obama H, Sagara M, Shirao K, Kusano C, Natsugoe S, Yoshinaka H, Baba M, Fukumoto T, Aikou T, Sato E. Expression of thrombomodulin in esophageal squamous cell carcinoma and its relationship to lymph node metastasis. Cancer Res. 1995; 55: 4196–4200.
Wilhelm S, Schmitt M, Parkinson J, Kuhn W, Graeff H, Wilhelm OG. Thrombomodulin, a receptor for the serine protease thrombin, is decreased in primary tumors and metastases but increased in ascitic fluids of patients with advanced ovarian cancer FIGO IIIc. Int J Oncol. 1998; 13: 645–651.