Adenovirus-Mediated Expression of Tissue Factor Pathway Inhibitor-2 Inhibits Endothelial Cell Migration and Angiogenesis
Objective— Extracellular matrix (ECM) remodeling during angiogenesis is accomplished through plasmin-dependent pericellular proteolysis and through the action of matrix metalloproteinases (MMPs). Because tissue factor pathway inhibitor-2 (TFPI-2), a Kunitz-type protease inhibitor with prominent ECM localization, inhibits plasmin and MMPs activity, we investigated the role of TFPI-2 in endothelial cell (EC) migration and angiogenesis.
Methods and Results— Real-time polymerase chain reaction and immunostaining showed that the expression of TFPI-2 mRNA and protein was upregulated in migrating ECs. The effect of TFPI-2 on angiogenesis was studied in mouse models of Matrigel and polyvinylalcohol sponge implants by overexpressing TFPI-2 through infection with a replication-deficient adenovirus (AdTFPI-2). Using (immuno)fluorescence and confocal microscopy we observed that TFPI-2 reduced neovascularization and promoted ECM deposition. Lateral cell migration and capillary tube formation in vitro also were impaired by TFPI-2, a process reversed by anti–TFPI-2 antibodies. Increased apoptosis occurred both in AdTFPI-2–treated ECs and in the mouse implants. Zymography and assays in the absence of plasminogen confirmed plasmin inhibition as a main mechanism through which TFPI-2 inhibits EC migration.
Conclusions— Our data suggest that TFPI-2 may be an important regulator of aberrant angiogenesis associated with tumor growth/metastasis, cardiovascular diseases, chronic inflammation, or diabetes.
Proteinase inhibitors are involved in blood coagulation, fibrinolysis, angiogenesis, wound healing, and tumor invasion.1 Tissue factor pathway inhibitor-2 (TFPI-2) is a Kunitz-type serine proteinase inhibitor2 synthesized by endothelial cells (ECs), smooth muscle cells (SMCs), and syncytiotrophoblasts,3,4 which associates through ionic interactions5 with the extracellular matrix (ECM),6 and inhibits trypsin, plasmin, and plasma kallikrein among others.7
The ECM is essential for the integrity of the cardiovascular system. ECs use the matrix as scaffolding during invasion, but they also degrade it to provide space for new capillaries. Serine proteases belonging to the plasminogen activator (PA)/plasmin system and (membrane-type)-matrix metalloproteinases [(MT)-MMPs] play major roles in all the steps of angiogenesis8: vessel sprouting, cell migration, tube formation, survival, generation of matrikines as angiogenesis inhibitors, and vessel stabilization/maturation and remodeling.9,10 Matrix metalloproteinase (MMP)-2, MMP-9, and MT1-MMP switch on neovascularization through degradation of the endothelial and interstitial matrix, and activation of growth factors (reviewed in10). Protease activities are controlled by specific activation mechanisms and inhibitors, of which tissue inhibitors of MMPs and serpins, like PA inhibitors (PAIs), represent major classes. MMPs are secreted as inactive zymogens that are activated by other proteinases, such as plasmin.
The involvement of urokinase-type PA (uPA) and plasmin in cell migration and invasion is well recognized.8,11 Plasmin generation through plasminogen activation by uPA is facilitated by uPA receptor (uPAR) and leads to sustained pericellular proteolysis. Cell-surface localization protects plasmin from inhibition by alpha2-antiplasmin12 and enhances its capability to degrade matrix components (laminin, collagens) and activate proMMPs.
The major role of TFPI-2 seems to be plasmin inhibition12 and thus of proMMP-1, proMMP-3, and proMMP-13 activation.13 As a direct inhibitor of MMP-2 and MMP-914 and potent inhibitor of both matrix-bound and cell-associated plasmin,12 TFPI-2 could regulate extracellular proteolysis and ECM remodeling, which are highly relevant both for normal development and for tumor invasion/metastasis, atherosclerosis, or chronic inflammation.
The few data concerning the effect of TFPI-2 on tumor growth and metastasis diverge because of inherent cancers’ variability.15–17 TFPI-2 was upregulated in tumors,18 atherosclerotic plaques,19 and angiogenic ECs.20
We sought to establish more firmly the mechanism of action of TFPI-2 in angiogenesis. We overexpressed TFPI-2 through adenovirus-mediated gene transduction or we used TFPI-2–containing medium from virus-infected cells, and we observed that TFPI-2 inhibits the formation of functional neovessels. Possible mechanisms include inhibition of protease-dependent cell recruitment and migration, inhibition of ECM degradation leading to excessive matrix deposition, and inhibition of vessel maturation.
Materials and Methods
Preparation of bicistronic adenoviruses,21,22 animal experimentation, Matrigel and sponge implant assays, real-time PCR, cell culture, antibodies and immunofluorescence, SDS-PAGE and Western blotting, cell migration, apoptosis assays, mouse aorta explant model, in vitro capillary tube formation, casein-overlay zymography, and statistical analysis are described in detail in the online supplement (available online at http://atvb.ahajournals.org).
TFPI-2 Is Upregulated in Migrating ECs
Multiple-“wounded”23 confluent EA.hy926 cells were immunostained for TFPI-2 at 24 and 48 hours post-wound and analyzed by confocal microscopy. ECs that migrated in the scratch displayed higher levels of TFPI-2 than the undisturbed cells, as well as extracellular accumulation of TFPI-2 (Figure 1a through 1c). TFPI-2 fluorescence intensity measured at the level of Golgi in individual cells on 30 images for each condition was expressed as median values and range of arbitrary units (AU) on an 8-bit scale with 0 to 256 levels. TFPI-2 fluorescence in migrating EC (M, 130 AU, range 67 to 163) was 2-fold enhanced over the level in quiescent cells (Q, 62 AU, range 41 to 88, P=0.00001; Figure 1d).
TFPI-2 mRNA expression in migrating EC (Figure 1e) was ≈2-fold increased over quiescent monolayer levels.
To study the role of TFPI-2 in EC migration, an anti–TFPI-2 IgG was included before wounding and during the recovery. The scratch was completely covered after 24 hours in the presence of the antibody, but not in native cells (Figure 1f and 1g). The antibody did neither affect the cell morphology, nor provoke cell detachment (not shown). Nonimmune goat serum had no effect.
TFPI-2 Expression After Adenovirus-Mediated Gene Transduction
COS-1, HEK293, and EA.hy926 cells were used to characterize the expression of TFPI-2 after infection with adenoviruses expressing TFPI-2 (AdTFPI-2), both TFPI-2 and green fluorescent protein (GFP) (AdTFPI-2GFP), or no protein (Ad-Control). Western blotting of cell lysates and conditioned medium (CM) from COS-1 cells showed the triplet of bands of Mr ≈27, 31, and 33 kDa typical for TFPI-26 (supplemental Figure IAa). There was no significant difference between TFPI-2 expression among the cells infected with AdTFPI-2 or AdTFPI-2GFP; neither did TFPI-2 reach detectable levels in non-infected COS-1 cells (compare lanes 1,1′ with 3,3′, and lanes 2 with 2′).
Western blot of CM from HEK293 infected with AdTFPI-2 (CM-TFPI-2, lane T) showed TFPI-2–immunoreactive triplet bands, whereas CM from Ad-Control–infected cells (CM-Control, lane C) displayed only a weak signal (supplemental Figure IAb). As reported for other ECs, including HUVECs,3 EA.hy926 cells also synthesize, secrete, and deposit TFPI-2 in the ECM (supplemental Figure IAb, EA.hy926 lanes C). TFPI-2 levels in AdTFPI-2-infected cells were increased by 2 to 4 times in the cells and medium and by ≈10 times in the ECM (lanes T; semi-quantitative estimation of pixel intensity).
The median fluorescence intensity for TFPI-2 in EC infected with AdTFPI-2 (193 AU, range 138 to 225) was significantly higher than for Ad-Control (72 AU, range 47 to 101; P<0.00001), and also higher than for noninfected migrating ECs in which TFPI-2 is naturally upregulated (see above).
High magnification confocal images (supplemental Figure IB) confirmed that both native TFPI-2 (a) and the inhibitor produced by AdTFPI-2GFP-infected cells (b) were largely deposited in the ECM.
Effect of TFPI-2 on Neovascularization In Vivo
Thick sections or whole-mount preparations of sponges harvested 2 and 3 weeks after subcutaneous implantation were examined by confocal microscopy24,25 after immunostaining for various markers.
Distribution of GFP, used as an indicator of infection efficiency, was quasi-uniform among the differently infected sponges (supplemental Figure IIa, IId, IIg, and IIj). Unlike the controls, AdTFPI-2 sponges displayed large deposits of TFPI-2 (supplemental Figure IIb, IIe, IIm, and IIp).
Most of the CD45-positive cells expressed GFP both in AdGFP- and AdTFPI-2GFP–infected sponges (supplemental Figure IIg through IIi and IIj through IIl), which indicates that leukocytes become infected during sponge invasion. Whereas few leukocytes were positive for TFPI-2 in controls (supplemental Figure IIm through IIo), the majority of CD45-positive cells in AdTFPI-2 sponges expressed large amounts of TFPI-2 (supplemental Figure IIp through IIs).
Immunostaining for CD34, a marker for neo-capillary ECs, showed strong neovascularization in controls, but severely impaired formation of neovessels in AdTFPI-2 sponges (Figure 2a through 2f). The high magnification insets illustrate that the defect did not reside with the level of expression of CD34, which was rather uniform among controls and AdTFPI-2 sponges, but with an apparent failure of clustered CD34-positive cells in the latter to get organized into regular 3D networks. Double immunostaining for CD34 and TFPI-2 in AdTFPI-2 sponges confirmed that the clusters of ECs expressed high levels of TFPI-2, which was also heavily deposited around the ECs (Figure 2g through 2l).
Similar differences appeared for laminin, which appeared alongside the networks of mature vessels in controls, but as only short stretches of non-organized material in the presence of excess TFPI-2 (Figure 2m and 2n).
Immunostaining for CD31 together with Dextran-fluorescein isothiocyanate (FITC) infusion (supplemental Figure III) revealed substantially impaired formation and functionality of small-diameter vessels in the presence of AdTFPI-2 as compared with the controls (a-f). 3D rendering of z-optical sections through whole-mount plugs25 demonstrate the microvessels paucity in AdTFPI-2 plugs (supplemental Movies I and II).
Detection of GFP proved that the infection was equally efficient among the tested conditions. Immunostaining for CD34 and NG2 (EC and pericytes) confirmed defects in both the assembly of EC and the stabilization of neovessels (supplemental Figure IIIg through IIIs).
Immunostaining for Ki67 (not shown) revealed 50% less proliferating cells in the presence of AdTFPI-2 (5.43±1.65 cells per field) than for Ad-Control (11.2±2.8 cells per field).
The content of hemoglobin decreased from 2.2±1.35 mg/mL in controls to 0.58±0.3 mg/mL in AdTFPI-2 plugs (P=0.034), thus confirming the decreased functionality of the neovessels in the latter.
TFPI-2 Enhances ECM Deposition
Histological staining of sponges (Figure 3A) revealed strong deposition of ECM components in the presence of TFPI-2. AdTFPI-2 sponges displayed massive collagen deposition (Sirius Red staining) in comparison with the controls (a through d, asterisks). Van Gieson staining confirmed the enhanced collagen deposition (e and f, pink areas), and also showed that the elastic fibers (intense blue) were less organized and fewer as occurrence in the AdTFPI-2 sponges than in controls.
Immunostaining for fibrin (Figure 3B) revealed that AdTFPI-2 sponges, but not the controls, contained extracellular deposits of fibrin, which did not overlap with TF-expressing cells (f). The expression of neither TF (Figure 3Bb and 3Be) nor TFPI (supplemental Figure IVb and IVe) was affected by TFPI-2 overexpression; therefore the increased fibrin deposition could not be attributed to either upregulation of TF or downregulation of TFPI. TFPI colocalized with CD34-positive cells regardless of the degree of neovessel organization (supplemental Figure IV).
TFPI-2 Inhibits Sprouting in Mouse Aorta Explants
Ad-control–infected aortic rings cultured in Matrigel displayed substantial sprouting from the adventitia, visible from the second day (Figure 4Aa and 4Ac). TFPI-2 overexpression almost completely inhibited the microvascular outgrowth (b and d). Outgrowth areas (mean±SD) after 7 days were: 1.2±0.33 mm2 (controls) and 0.035±0.01 mm2 (AdTFPI-2; P=0.0002; panel e).
Detection of GFP fluorescence proved that adenovirus infection was efficient for both conditions. Immunostaining for CD31 confirmed that EC sprouting was suppressed by TFPI-2 overexpression, as opposed to the normal neovessel outgrowth in controls (supplemental Figure V).
TFPI-2 Inhibits In Vitro Capillary Tube Formation
Substantial formation of vessel-like structures occurred in the presence of CM-Control (mean length 0.785±0.19 mm), whereas CM-TFPI-2 inhibited almost totally the tube formation (Figure 4Bb and 4Be). Confirming the role of TFPI-2, anti–TFPI-2 IgG partially reversed the effect of CM-TFPI-2 (mean length 0.5±0.07 mm; P=0.003; panel f). There was no significant effect of the anti–TFPI-2 IgG on the morphology of ECs or their adhesion, but the cells showed incipient tubulogenesis in the presence of the antibody earlier than in its absence (Figure 4Ba and 4Bc; arrows). Nonimmune goat serum had no effect. Similar inhibition of tubulogenesis was observed for AdTFPI-2 (not shown)
TFPI-2 Increases Apoptosis
By flow cytometry, the number of Annexin V–positive cells—indicative of cellular apoptosis—increased by 2-fold after infection with AdTFPI-2 (15.4% versus 8.9% in controls; supplemental Figure VIA).
Quantification of terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling (TUNEL) performed on Matrigel plugs revealed less than 1% apoptotic cells in controls and 15% in AdTFPI-2 samples (P=0.037; supplemental Figure VIB).
TFPI-2 Inhibits EC Migration In Vitro
In the Boyden chamber assay, EC migration decreased from 39.3±6 cells per field for Ad-Control to 25±4.5 cells/field for AdTFPI-2-infected cells (P=0.039). CM-TFPI-2–treated cells exhibited 23±5.5 cells per field versus 40±7.5 cells per field for CM-control (P=0.04). Anti–TFPI-2 IgG reversed the effect to control levels for all the conditions (not shown).
In the scratch assay, lateral cell migration measured over 48 hours was significantly reduced by excess TFPI-2. The wound coverage, expressed as percentage of the initial area, represents the index of wound healing. In contrast with the nearly complete recovery in ECs infected with Ad-Control (Figure 5a, 5b, and 5e), AdTFPI-2 cells recovered only ≈30% of the denuded area after 48 hours (c, d, and e). Similar degree of inhibition was observed for ECs incubated with CM-TFPI-2, where only 20% of the denuded area recovered after 48 hours (panel e). Comparable results were obtained in three different experiments for both HUVECs and EA.hy926 cells.
The relationship between the inhibition of cell migration by TFPI-2 and plasmin generation was studied either by depleting plasminogen in the FBS used in the cell medium, or by including inhibitory anti-plasminogen Ab in the assays. Confirming the absolute requirement of plasminogen for cell migration, neovascularization, and wound healing,26,27 we found that both Ad-control and Ad-TFPI2–infected cells failed to migrate in the absence of plasminogen even after 48 hours. No significant difference was observed between the healing index values among the time points or the conditions tested (Figure 5f through 5h). After 48 hours, addition of bovine plasminogen (100 μg/mL) almost completely restored the capability of Ad-control cells, but not of Ad-TFPI2 cells, to migrate and cover the wound during the next 24 hours. Similar results were obtained when using anti-plasminogen Ab (panel h).
TFPI-2 Inhibits Plasmin-Dependent Proteolysis
To confirm that inhibition of cell migration by TFPI-2 is related to plasmin inhibition we performed zymography on wounded ECs overlaid with agar-casein in the presence or absence of plasminogen. Significant areas of lysis appeared in controls: 0.025±0.008 mm2 for CM-Control and 0.031±0.01 mm2 for Ad-Control (Figure 6a and 6c). Caseinolysis is seen as dark areas against white background under dark-field illumination or as transparent zones in Nomarsky microscopy (Figure 6e through 6h). No lysis occurred in ECs incubated with CM-TFPI-2 or overexpressing TFPI-2 (Figure 6b and 6d), or in any of these conditions in the absence of plasminogen (insets). Anti–TFPI-2 IgG blocked the inhibition of plasmin-dependent caseinolysis by TFPI-2 for both CM-TFPI-2 and AdTFPI-2 (not shown).
The present data establish that endogenous overexpression of TFPI-2 or addition of excess TFPI-2 effectively inhibits angiogenesis in vivo, and EC migration and capillary tube formation in vitro.
To the best of our knowledge, we showed here for the first time that native TFPI-2 is upregulated in migrating ECs in culture, thus playing a role in the temporal and spatial control of cell migration. Paradoxical at first, this finding actually fits the idea that moderately increased levels of TFPI-2 are needed to balance the protease–anti-protease equilibrium that must be maintained during cell migration to confine extracellular proteolysis to the cell surface. Normal angiogenesis requires both enhanced proteolysis to facilitate cell invasion, and the means to prevent excessive ECM degradation that would ultimately inhibit angiogenesis, requirement that is achieved through physiologically controlled upregulation of both proteases and corresponding protease inhibitors at the same time. Migrating cells display increased functional urokinase plasminogen activator (uPA)/uPAR on the cell surface of their leading edge,28,29 which leads to localized plasmin generation, but also upregulates plasminogen activator inhibitor (PAI)-123 to inhibit uPA and equilibrate the balance. The upregulation of TFPI-2 as potent plasmin inhibitor may serve to limit excessive proteolysis, thus preventing the destruction of the scaffold required for invasion, or the formation of aberrant vascular structures.
Finding that ECs migrate faster when TFPI-2 is blocked with antibodies prompted us to investigate whether, similar to PAI-1,30 overexpressing TFPI-2 at levels significantly higher than those attained through physiological up-regulation will conversely inhibit cell migration and angiogenesis.
We first confirmed the efficacy of adenovirus-mediated gene transduction in cell cultures infected with various adenovirus preparations, where we found that TFPI-2 was overexpressed, on average, ≈5-times higher than in controls, and was, as reported,3 largely deposited in the surrounding ECM.
Data from the mouse in vivo models of neovascularization indicate that TFPI-2 strongly inhibits angiogenesis. Through adenoviral infection, the levels of expression of TFPI-2 are kept high throughout the angiogenic process, beginning with invasion and continuing with matrix remodeling, capillary network formation, and vessel maturation. Overexpression of TFPI-2 in leukocytes seemingly decreases both their invasiveness and subsequent proliferation, as demonstrated by the decrease in Ki67-positive cell numbers. Leukocytes contribute to the initiation and guidance of new blood vessels, with uPA/uPAR, MT1-MMP, and MMP-9 playing major roles in their recruitment (reviewed in10). The failure of clustered neocapillary ECs in the AdTFPI-2 sponges to get organized into a proper 3D network suggests a defect in the leukocyte-dependent guidance, most likely produced through TFPI-2 inhibition of plasmin and MMP-9.
Even when some neovessels do form, their density is low and they are functionally immature, as proven by the lack of access of Dextran-FITC and their reduced hemoglobin content. This is probably attributable to deficient assembly of both ECs and pericytes, the latter being a key regulator of the stabilization of newly formed vessels. Because pericyte recruitment is also dependent on MMP-9,31 this represents another step where overexpression of TFPI-2 could interfere.
EC adhesion to ECM, and their migration and proliferation, are critical for the progression of angiogenesis. Overexpression of TFPI-2 or addition of exogenous TFPI-2 reduced EC migration in vitro and inhibited both tube-like structures formation in cell culture and microvascular outgrowth in aorta explants. As reported,26,27 we confirmed that ECs fail to migrate in the absence of plasminogen, hence of plasmin, but do recover their migratory capability when plasminogen is replenished. Using this approach, we also proved that TFPI-2 suppresses cell migration only when plasminogen is present, therefore when plasmin is generated. Inhibition of cell migration in vitro correlates with inhibition of plasmin-dependent caseinolysis by TFPI-2. In vivo, overexpression of TFPI-2 promotes ECM deposition (collagen, laminin) and the formation of fibrin deposits unrelated to the levels of TF and TFPI. Taken together, these findings suggest that TFPI-2 acts primarily through inhibition of plasmin-dependent proteolysis and/or MMPs to regulate cell migration and ECM turnover. Interestingly, bikunin, another Kunitz-type protease inhibitor, also inhibits cell invasion, possibly through direct inhibition of cell-associated plasmin activity.32
Finding that overexpression of TFPI-2 increased apoptosis, both for ECs in vitro and for cells in Matrigel implants in vivo, indicates that TFPI-2 may also protect against aberrant neovessel formation through proteolysis-independent mechanisms. It remains to be established whether the process may also involve regulation of gene expression of proteins like VEGF-C, VEGFR1, or ECM components.16
Inhibition of neoangiogenesis by TFPI-2 is already a promising anti-cancer strategy.15 Because defective neovascularization appears in non-malignancies also, site-directed TFPI-2 overexpression might protect against uncontrolled proteolysis that leads to rupture of atherosclerotic plaques or aneurysms and to restenosis,19,33 neo-vascularization in chronic inflammation,34 or diabetic retinopathy.35 The ECM protection offered by TFPI-2, together with the inhibition of cell motility and vessel maturation, may render TFPI-2 a promising candidate for novel therapies to prevent or reduce unwanted angiogenesis.
We thank Julie Poirot for technical assistance with adenovirus production, and M. McDaniel for help with flow cytometry.
Sources of Funding
This work was supported by 5RO1GM037704–17 (L.I. and F.L.) and 5P20RR018758–02 (H.T. and C.L.) from the National Institutes of Health.
Original received June 20, 2006; final version accepted November 15, 2006.
Sprecher CA, Kisiel W, Mathewes S, Foster DC. Molecular cloning, expression, and partial characterization of a second human tissue-factor-pathway inhibitor. Proc Natl Acad Sci U S A. 1994; 91: 3353–3357.
Iino M, Foster DC, Kisiel W. Quantification and characterization of human endothelial cell-derived tissue factor pathway inhibitor-2. Arterioscler Thromb Vasc Biol. 1998; 18: 40–46.
Liu Y, Stack SM, Lakka SS, Khan AJ, Woodley DT, Rao JS, Rao CN. Matrix localization of tissue factor pathway inhibitor-2/matrix-associated serine protease inhibitor (TFPI-2/MSPI) involves arginine-mediated ionic interactions with heparin and dermatan sulfate: heparin accelerates the activity of TFPI-2/MSPI toward plasmin. Arch Biochem Biophys. 1999; 370: 112–118.
Rao CN, Reddy P, Liu Y, O’Toole E, Reeder D, Foster DC, Kisiel W, Woodley DT. Extracellular matrix-associated serine protease inhibitors (Mr 33,000, 31,000, and 27,000) are single-gene products with differential glycosylation: cDNA cloning of the 33-kDa inhibitor reveals its identity to tissue factor pathway inhibitor-2. Arch Biochem Biophys. 1996; 335: 82–92.
Pepper MS. Role of the matrix metalloproteinase and plasminogen activator-plasmin systems in angiogenesis. Arterioscler Thromb Vasc Biol. 2001; 21: 1104–1117.
van Hinsbergh VW, Engelse MA, Quax PH. Pericellular proteases in angiogenesis and vasculogenesis. Arterioscler Thromb Vasc Biol. 2006; 26: 716–728.
Heymans S, Luttun A, Nuyens D, Theilmeier G, Creemers E, Moons L, Dyspersin GD, Cleutjens JP, Shipley M, Angellilo A, Levi M, Nube O, Baker A, Keshet E, Lupu F, Herbert JM, Smits JF, Shapiro SD, Baes M, Borgers M, Collen D, Daemen MJ, Carmeliet P. Inhibition of plasminogen activators or matrix metalloproteinases prevents cardiac rupture but impairs therapeutic angiogenesis and causes cardiac failure. Nat Med. 1999; 5: 1135–1142.
Chand HS, Du X, Ma D, Inzunza HD, Kamei S, Foster D, Brodie S, Kisiel W. The effect of human tissue factor pathway inhibitor-2 on the growth and metastasis of fibrosarcoma tumors in athymic mice. Blood. 2004; 103: 1069–1077.
Neaud V, Hisaka T, Monvoisin A, Bedin C, Balabaud C, Foster DC, Desmouliere A, Kisiel W, Rosenbaum J. Paradoxical pro-invasive effect of the serine proteinase inhibitor tissue factor pathway inhibitor-2 on human hepatocellular carcinoma cells. J Biol Chem. 2000; 275: 35565–35569.
Rao CN, Lakka SS, Kin Y, Konduri SD, Fuller GN, Mohanam S, Rao JS. Expression of tissue factor pathway inhibitor 2 inversely correlates during the progression of human gliomas. Clin Cancer Res. 2001; 7: 570–576.
Crawley JT, Goulding DA, Ferreira V, Severs NJ, Lupu F. Expression and localization of tissue factor pathway inhibitor-2 in normal and atherosclerotic human vessels. Arterioscler Thromb Vasc Biol. 2002; 22: 218–224.
Bell SE, Mavila A, Salazar R, Bayless KJ, Kanagala S, Maxwell SA, Davis GE. Differential gene expression during capillary morphogenesis in 3D collagen matrices: regulated expression of genes involved in basement membrane matrix assembly, cell cycle progression, cellular differentiation and G-protein signaling. J Cell Sci. 2001; 114: 2755–2773.
Gerard R MR. Adenovirus vectors. In: Hames B GD, ed. DNA Cloning - A Practical Approach: Mammalian Systems. Oxford: Oxford University Press; 1995.
Lupu C, Hu X, Lupu F. Caveolin-1 enhances tissue factor pathway inhibitor exposure and function on the cell surface. J Biol Chem. 2005; 280: 22308–22317.
Rakic JM, Lambert V, Munaut C, Bajou K, Peyrollier K, Alvarez-Gonzalez ML, Carmeliet P, Foidart JM, Noel A. Mice without uPA, tPA, or plasminogen genes are resistant to experimental choroidal neovascularization. Invest Ophthalmol Vis Sci. 2003; 44: 1732–1739.
Pepper MS, Vassalli JD, Montesano R, Orci L. Urokinase-type plasminogen activator is induced in migrating capillary endothelial cells. J Cell Biol. 1987; 105: 2535–2541.
Pepper MS, Sappino AP, Stocklin R, Montesano R, Orci L, Vassalli JD. Upregulation of urokinase receptor expression on migrating endothelial cells. J Cell Biol. 1993; 122: 673–684.
Chantrain CF, Shimada H, Jodele S, Groshen S, Ye W, Shalinsky DR, Werb Z, Coussens LM, DeClerck YA. Stromal matrix metalloproteinase-9 regulates the vascular architecture in neuroblastoma by promoting pericyte recruitment. Cancer Res. 2004; 64: 1675–1686.
Moulton KS, Vakili K, Zurakowski D, Soliman M, Butterfield C, Sylvin E, Lo KM, Gillies S, Javaherian K, Folkman J. Inhibition of plaque neovascularization reduces macrophage accumulation and progression of advanced atherosclerosis. Proc Natl Acad Sci U S A. 2003; 100: 4736–4741.
Wilkinson-Berka JL. Angiotensin and diabetic retinopathy. Int J Biochem Cell Biol. 2005; 38: 752–765.