Atypical GPI-Anchored T-Cadherin Stimulates Angiogenesis In Vitro and In Vivo

【摘要】  Objective- T-cadherin (T-cad) is an atypical GPI-anchored member of the cadherin superfamily. In vascular tissue, T-cad expression is increased during atherosclerosis, restenosis, and tumor neovascularization. In vitro, overexpression and/or homophilic ligation of T-cad on endothelial cells (ECs) facilitates migration, proliferation, and survival. This study investigated T-cad effects on angiogenesis.

Methods and Results- In vitro, T-cad homophilic ligation induced arrangement of ECs into a capillary-like network in a 2-dimensional model of EC differentiation and stimulated in-gel endothelial sprout outgrowth in an EC spheroid model and a modified Nicosia tissue assay. Sprouting from spheroids composed of adenoviral-infected T-cad overexpressing ECs or T-cad siRNA transfected ECs were significantly increased or reduced, respectively. In vivo, T-cad potentiated VEGF effects on neovascularization in a model of myoblast-mediated gene transfer to mouse skeletal muscle; vessel caliber after co-delivery of T-cad and VEGF was significantly greater than after delivery of VEGF alone.

Conclusions- We unequivocally identify T-cad as a novel modulator of angiogenesis and suggest that this molecule can be exploited as a target for modulation of therapeutic angiogenesis, as well as for prevention of pathological conditions associated with abnormal neovascularization.

This study demonstrates that GPI-anchored T-cadherin stimulates angiogenesis in 2-dimensional model of endothelial differentiation, in 3-dimensional endothelial spheroid, and Nicosia tissue assays in vitro. In vivo, T-cad potentiates VEGF effects on neovascularization in mouse skeletal muscle. We conclude that T-cad is a novel modulator of angiogenesis.

【关键词】  angiogenesis cadherin endothelial cell differentiation VEGF

Introduction

Angiogenesis, a structural and morphogenetic process by which new blood vessels are generated by sprouting from preexisting vessels, plays an important part in embryogenesis and in the adult for physiological repair and restoration of blood supply to damaged tissues. In the cardiovascular system, pathological neovascularization induced in response to inflammation and tissue ischemia during atherogenesis can predispose the atherosclerotic plaque to intramural hemorrhage and rupture causing thrombosis and subsequent facilitation of coronary artery stenoses, occlusion, and myocardial ischemia. 1 Deregulated angiogenesis can promote and aggravate pathological conditions such as tumor formation, diabetic retinopathy, and psoriasis. 2 However, therapeutic neovascularization based on supplementation of angiogenic growth factors with resulting enhancement of native angiogenesis is a promising strategy to treat myocardial and peripheral ischemia. Many experimental and preclinical trials reported that vascular endothelial growth factor (VEGF), the major endothelial-specific regulator of angiogenesis, significantly improves tissue perfusion and function when delivered as recombinant protein or by gene transfer. 3 Some concern, however, has been raised regarding adequacy in structure, function and stability of newly generated blood vessels as well as unwanted side effects such as formation of hemangiomas. 4-6 Consequently, there is compelling need for alternative strategies that would allow to finely regulate therapeutic angiogenesis rates and avoid unwanted side effects without impairing beneficial effects of growth factors on tissue perfusion.

Cadherins are transmembrane receptors mediating homophilic calcium-dependent intercellular adhesion. 7 T-cadherin (T-cad), an atypical member of the cadherin superfamily, shares the general molecular organization of cadherin extracellular domains, but lacks transmembrane and cytoplasmic domains and is attached to the plasma membrane via glycosylphosphatodylinositol (GPI) anchor. 8 A role for T-cad in the regulation of vascular cell function has been recognized only recently. T-cad is widely expressed in the cardiovascular system 9 and its expression in vascular cells is markedly increased during atherosclerosis, restenosis after balloon angioplasty, and tumor neovascularization. 9-11 In contrast to classical cadherins, T-cad is absent from adherens junctions, is located within lipid rafts of the plasma membrane, and is redistributed to the leading edge of migrating cells. 12,13 These characteristics invoke a function for T-cad as a signaling receptor involved in interpretation of extracellular cues rather than as a true adhesion molecule. In vitro, homophilic ligation of T-cad receptors on the endothelial cell (EC) surface with soluble recombinant T-cad protein or with agonistic antibody induces the motile phenotype via activation of Rho and Rac pathways and facilitates cell migration, 14,15 whereas T-cad overexpression increases cell cycle progression, proliferation, and survival via activation of PI3-kinase/Akt pathway. 16,17 Because EC phenotypic modulation, growth and migration are critical steps in the angiogenic process, we hypothesized a role for T-cad in neovascularization and examined its effects on EC angiogenic properties in vitro and in vivo.

Materials and Methods

The materials and methods used in this study are fully described in the online data supplement section (http://atvb.ahajournals.org).

Results

T-cad Ligation Induces Rearrangement of ECs Into Network Structures

Our previous studies demonstrated that T-cad can exert two types of effects on EC. Upregulation of T-cad expression level on the cell surface stimulates proliferation and survival. 14,17 Ligation of T-cad receptors with recombinant T-cad protein or agonistic antibody included into the matrix imitates homophilic T-cad-mediated intercellular interactions and promotes cell detachment and cell motility. 14 To analyze the effect of T-cad ligation on EC phenotype in the 2-dimensional monolayer system human umbilical vein endothelial cells (HUVEC) or human microvascular endothelial cell line (HMEC-1) cells were plated onto substrata containing recombinant T-cad protein which acts as an external ligand for T-cad molecules expressed on the surface of spreading cells. Cell morphology was examined following fixation of cultures ( Figure 1A and 1 B) and in living cells ( Figure 1 C). EC on control substratum displayed the typical cobblestone morphology of endothelial monolayers ( Figure 1 A, left panels), whereas ECs on T-cad-containing substratum formed a network-like pattern of interconnecting tubular structures ( Figure 1 A, middle and right panels). Total length of capillary-like tubes on T-cad substratum was 8-fold higher than on control substratum for HUVECs and 20-fold higher for HMEC-1 ( Figure 1 B). Similar data were obtained when T-cad ligation was induced by substratum inclusion of agonistic antibody against the first extracellular domain of T-cad. 14 Transfection of HUVECs with T-cad siRNA significantly decreased their tube-formation response to T-cad substratum; decreased T-cad protein level in HUVEC 72 hours after transfection with specific siRNA was confirmed by immunoblotting ( Figure 1 B). Time-lapse videomicroscopy showed that cells on T-cad substratum failed to spread and aggregated in small clusters from which elongated cells with highly motile lamellipodia protruded toward each other forming a network pattern within 5 hours ( Figure 1 C). Cells on control substratum spread fully within 1.5 to 2 hours after seeding (not shown). For videos showing EC behavior on control (video1-control.avi) and T-cad-containing substrata (video2-Tcad.avi) (please see http://atvb.ahajournals.org).

Figure 1. T-cad ligation induces EC rearrangement into tubular structures. A, HUVEC and HMEC-1 were plated onto dishes precoated with inclusion of BSA or recombinant T-cad protein. After 6 hours cells were stained with methylene blue (left and central panels) and with Hoechst for visualization of nuclei (right panels). B, Morphometric analysis showed that T-cad substratum significantly increased total tubular structure length (*** P <0.001). Silencing of T-cad using siRNA transfection decreased cell response to T-cad substratum (** P <0.001). Downregulation of T-cad protein in cells transfected by specific siRNA was proven by immunoblotting. C, Representative images of HMEC-1 on T-cad substratum recorded during time-lapse videomicroscopy. Note elongated cells with highly motile lamellipodia (white arrows). Bar=100 µm.

T-cad Ligation on ECs Stimulates Angiogenesis in the Spheroid In Vitro Model

The 2-dimensional models for cord-forming have been successfully used for studying effects of cell matrix interactions on EC morphology, but they do not reflect sprouting angiogenesis. Therefore, we implemented the EC-spheroid assay, which has proven to be useful for analysis of active outgrowth of newly formed capillary-like sprouts within 3-dimensional matrices. 18 The effect of homophilic T-cad ligation on sprouting was examined by embedding spheroids into gels prepared with inclusion of T-cad protein. Identical results were obtained with collagen and fibrin matrices ( Figure 2 presents data only for fibrin). We observed both VEGF-induced and a background "spontaneous" sprouting probably caused by autocrine cell activity and/or the presence of low concentrations of growth factors (eg, basic fibroblast growth factor) in the medium overlay. Inclusion of T-cad in gels significantly ( P <0.01) increased sprouting both in the absence and presence of VEGF (see photomicrographs in Figure 2 A and graphs in Figure 2B to 2 D). T-cad effect was apparently independent of exogenous VEGF since relative increases in sprout length in T-cad-containing gels ( 1.5- to 2-fold versus control) were similar without or with VEGF in medium overlay ( Figure 2C and 2 D). Use of the Hoechst stain to quantitate the total length of endothelial structures composed of 2 aligned cells revealed 2.5-fold increase ( P <0.01) in chain-like organization of sprouting EC within T-cad-containing gels ( Figure 2A, 2C, and 2 D). This effect of T-cad on EC alignment within outgrowing sprouts resembles its ligation-dependent stimulation of EC reorganization into chains in monolayer cultures.

Figure 2. T-cad ligation stimulates angiogenesis in the in vitro EC-spheroid model. HUVEC spheroids were embedded into fibrin gels without (control) or with inclusion of recombinant T-cad protein (80 µg/mL for A, C, and D; indicated concentrations for B) or BSA as negative control and incubated in the presence or absence of 50 ng/mL VEGF. A, EC spheroid-containing gels were stained with TRITC- phalloidin for visualization of cellular structures (red) and with Hoechst for visualization of nuclei (blue). Bar=100 µm. B to D, Morphometric analysis shows that sprout lengths were significantly greater for spheroids embedded within T-cad-containing gels (** P <0.01). Stimulation of angiogenesis in the presence of VEGF was dependent on T-cad concentration (B). Data for both the total length of all sprouts and the total length of sprouts composed of 2 and more cells are given as absolute units (B, C) and as % of the respective control (D).

T-cad Overexpression in ECs Stimulates Angiogenesis in the Spheroid In Vitro Model

To determine effects of T-cad upregulation per se on EC angiogenic behavior spheroids composed of T-cad-overexpressing (TC+) or control (parental, empty vector, or LacZ-infected) HUVECs were embedded into collagen and fibrin gels not containing T-cad protein as an external ligand. T-cad protein expression in infected HUVEC was monitored by Western-blotting ( Figure 3 C). Morphometric analysis (data expressed relative to sprouting in parental HUVEC spheroids) demonstrated that T-cad overexpression produced 2.5-fold ( P <0.01) increase in the total length and number of sprouts ( Figure 3A and 3 B). Sprouting behaviors of empty vector, LacZ, or parental HUVEC spheroids were comparable. As in the preceding experimental series on spheroids embedded within a T-cad containing gel, inclusion of VEGF in the medium overlay increased sprouting for all spheroids and the relative stimulatory effect of T-cad upregulation on sprouting was similar without or with inclusion of VEGF ( Figure 3 A). To confirm the importance of T-cad expression level for angiogenesis we studied the effect of T-cad silencing by siRNA transfection on HUVEC sprouting rates. Transfection with T-cad siRNA significantly ( P <0.01) decreased sprout outgrowth as compared with parental and control siRNA-transfected cells ( Figure 3 D). The lesser consequence of T-cad downregulation as compared with that of T-cad overexpression ( Figure 3A and 3 D) is likely caused by the lesser degree of change in protein expression achieved with siRNA as compared with adenovirus-dependent gene transfer, and to the fact that even low levels of protein can have significant functional activity.

Figure 3. T-cad overexpression in ECs stimulates angiogenesis in the in vitro spheroid model. Spheroids formed by parental HUVEC (control) or cells infected with empty, LacZ or T-cad cDNA-carrying adenoviral vectors (TC+) were embedded in collagen or fibrin gels and incubated in the presence or absence of 50 ng/mL VEGF. Total sprout length/spheroid (A, left graph) and number of sprouts/spheroid (A, right graph) were determined morphometrically and expressed as % of the respective control. Length and number of sprouts for TC+-spheroids were significantly greater (** P <0.01) than all controls (parental, empty and LacZ). Representative images of TRITC-phalloidin-stained spheroids are shown (B). Bar=100 µm. T-cad expression level in adenovirus-infected cells was monitored by Western blotting (C). Transfection of HUVEC with T-cad siRNA decreased sprout outgrowth into fibrin gels in the presence of VEGF (D). Sprouting assay from empty vector-infected and TC+ spheroids was performed in the absence or presence of 10 µmol/L VEGF receptor inhibitor (E). a indicates significant difference ( P <0.05) between empty-vector infected and TC+-spheroids; ***significant difference ( P <0.001) between control and VEGFR inhibitor-treated spheroids.

To further investigate whether T-cad-induced increase in angiogenesis is dependent on the presence of VEGF we analyzed the effect of specific VEGF receptor tyrosine kinase inhibitor on sprouting from T-cad-overexpressing spheroids. As expected, VEGF receptor inhibitor completely repressed VEGF-induced sprouting from control empty vector-infected spheroids to the basal spontaneous sprouting level ( Figure 3 E). However, the inhibitor neither blocked stimulation of sprouting from TC+-spheroids in the absence of exogenous VEGF nor completely eliminated the difference between TC+ and control spheroids in the presence of VEGF, but reduced sprouting to the same level as in its absence. These data indicate that T-cad stimulated EC sprouting is VEGF-independent.

Recombinant T-Cad Stimulates Angiogenesis in the Nicosia Heart Model

Of the currently available in vitro angiogenesis models, the Nicosia tissue assay is considered to more closely approximate the in vivo situation; it includes not only ECs but also surrounding nonendothelial cells in their microvascular environment, and the EC growing into fibrin gels from heart tissue fragments have not been preselected by passaging and thus are not in the proliferative state at the time of explantation, although tissue injury and in vitro conditions induce a certain level of spontaneous sprouting.

Morphometric quantification of total gel area invaded by endothelial sprouts and the total sprout length within fibrin gels ± recombinant T-cad demonstrated the stimulatory effect of T-cad on both spontaneous and growth factor (VEGF or bFGF)-induced sprouting ( P <0.01 versus controls) ( Figure 4 A). Representative photomicrographs are presented ( Figure 4B and 4 C). Note the presence of prominent cords and the fine interconnecting mesh of tubular structures within T-cad-containing gels ( Figure 4 C).

Figure 4. Angiogenesis in Nicosia model: influence of recombinant T-cad. Pieces of mouse hearts were embedded in fibrin gels without (control) or with inclusion of BSA or T-cad protein and incubated in the presence or absence of VEGF or bFGF (10 ng/mL). T-cad increases the area invaded by ECs and sprout length (** P <0.01). (A). Representative images of gels (± bFGF only) stained with AlexaFluor 488-labeled isolectin IB 4 are shown: low magnification pictures, bar=500 µm, invasion distances indicated with arrows (B) and high-magnification pictures illustrating formation of a capillary-like net, bar=200 µm (C).

T-Cad Modulates Dose-Dependent VEGF-Induced Angiogenesis in Mouse Skeletal Muscle

To demonstrate relevant in vivo angiogenic functions for T-cad we investigated effects of recombinant T-cad protein on VEGF-induced angiogenesis in skeletal muscle. In this model gene delivery is achieved by implantation of retrovirally transduced mouse myoblasts into the posterior auricular muscle of mice where the myoblasts differentiate, fuse with preexisting host myofibers and secrete the proteins of interest into the surrounding tissue. The following previously characterized 19 myoblast clones were used: clones VZ6 and VZ3 homogenously expressing distinct VEGF 164 levels ( 5 ng/10 6 and 70 ng/10 6 cells/d in culture, respectively) together with LacZ marker gene, and control myoblasts (Z) expressing LacZ only. For the purposes of estimating the ability of secreted T-cad protein to induce and/or modulate VEGF-stimulated angiogenesis, the myoblasts were overinfected with a retroviral vector encoding secreted form of recombinant T-cad linked to a truncated form of CD8a (VZ6/T, VZ3/T, Z/T) or with control vector expressing CD8a only (VZ6/C, VZ3/C, Z/C). Therefore, in each clonal population, every cell expressed the same amount of VEGF, ensuring its homogenous distribution in the myofiber microenvironment. This is crucial because it was previously shown that the phenotype of induced vessels depends strictly on VEGF dose. Moreover, different levels of expression do not average in vivo and remain highly localized in the microenvironment around each fiber. 19 Therefore in order to draw quantitative conclusions it is necessary to base comparisons on a given homogenous expression level of VEGF. The presence of secreted T-cad protein in the culture medium of T-cad/CD8a-transduced myoblasts was confirmed by Western blotting (data not shown). Representative images of the vessels formed at the sites of clone implantation are shown in Figure 5 A.

Figure 5. T-cad modulates VEGF-induced angiogenesis in mouse skeletal muscle. Mouse myoblast clones expressing LacZ only or LacZ and 2 distinct VEGF 164 levels were overinfected with a control retroviral vector (Z/C, VZ6/C, VZ3/C, respectively) or a vector expressing recombinant T-cad (Z/T, VZ6/T, VZ3/T) and implanted into the posterior auricular muscle of C.B.-17/SCID mice. Angiogenic effects were evaluated after 4 weeks. All examined vessels exhibited normal morphology, no aberrant structures or hemangiomas were observed (A, Bar=50 µm). Data for vessel diameter measurements are presented as scatter diagram (B) and vessel diameter distribution (C). Myoblasts co-expressing VEGF and T-cad induced formation of capillaries with larger diameters than myoblasts expressing only VEGF (** P <0.01).

Within the areas around control Z cells, only straight capillaries running parallel to the myofibers were observed. Implantation of VEGF-expressing myoblasts clones stimulated formation of new capillaries which were smaller (VZ6/C) or larger (VZ3/C) than pre-existing vessels depending on VEGF level ( Figure 5 B), as previously reported. 19 The absence of blue LacZ staining in the case of clones VZ3/C and VZ3/T indicates that implanted VZ3 myoblasts did not remain in the tissue for the entire 4 weeks; nevertheless, the morphological characteristics of the vessels attest to secretion of sufficient VEGF and the angiogenic response is as expected for VZ3 clone. 19 Clones VZ6/T and VZ3/T co-expressing VEGF and T-cad induced similarly normal capillaries, and no aberrant structures or hemangiomas were observed. However, morphometric analysis of vessel size and density revealed that T-cad-secreting clones stimulated formation of capillaries with larger diameters than myoblasts expressing only VEGF ( Figure 5B and 5 C). In the absence of VEGF, T-cad also induced a slight upward shift in vessel diameter distribution; this increase was not statistically significant. Vessel length density, calculated as total length of capillaries per unit area, was not influenced by T-cad in the absence of VEGF or together with high VEGF (not shown). In the presence of low VEGF T-cad caused a small increase of vessel density (119±14 mm/mm 2 for VZ6/T versus 109±6 mm/mm 2 for VZ6/C; P <0.05).

Discussion

Our previous studies demonstrated that T-cad can exert two types of effects on EC: T-cad overexpression stimulates cell cycle progression, proliferation, and survival, whereas homophilic ligation of T-cad receptors on the EC surface with recombinant extracellular domain of T-cad molecule included into the matrix causes cell detachment and promotes motility. 14-17 The main findings of this study are: (1) a further consequence of T-cad ligation-dependent modulation of cell adhesive properties in a 2-dimensional model of EC differentiation is induction of tubular structures and arrangement of EC into multicellular interconnecting chains which form a capillary-like/networked pattern closely resembling the initial response of EC to a pro-angiogenic environment; (2) T-cad stimulates in-gel outgrowth of endothelial sprouts in 3-dimensional in vitro EC-spheroid and heart tissue models of angiogenesis; and (3) T-cad facilitates VEGF-induced angiogenesis in mouse skeletal muscle in vivo.

It is recognized that EC of all origins share a common ability to organize in vitro into tubular networks that are very similar to vascular beds formed by vasculogenesis or angiogenesis in vivo. 20 Mechanical forces applied to cells from the matrix, dynamic control of cell matrix adhesion, and resulting alterations in cell shape are of crucial importance for determination of EC function. 21 The relationship between EC adhesivity and their ability to migrate or differentiate is not direct and unequivocal. Promotion of integrin-mediated attachment by ephrinB1/EphB4 tyrosine phosphorylation is necessary for EC migration and neovascularization, 22 and integrin blockade with anti-adhesive molecules decorin 23 or rhodostomin 24 elicits anti-angiogenic effects. However, dynamic inhibition of adhesion can be equally important for EC conversion toward differentiated angiogenic phenotype. Negative regulation of integrin function by class 3 semaphorins, ephrinA1/EphA2 and blocking anti- 2 ß 1 and v ß3 integrin antibodies promotes formation of cord structures on collagen I or fibrin in vitro 25 and stimulates neovascularization in vivo. 26,27 By analogy we suppose that induction of cord-like structures in response to T-cad ligation in monolayer EC might be caused by T-cad-dependent inhibition of adhesion to the matrix, 14 resulting in an "angiogenic switch" in the cell differentiation program and a facilitation of cell motility.

T-cad ability to influence neovascularization is supported by the 3-dimensional in vitro angiogenesis experiments. In both EC-spheroid and Nicosia tissue assays inclusion of T-cad ligand (recombinant T-cad protein) into gels promotes assembly of multicellular EC sprouts and increases total length and number of capillary-like structures invading the gels. Angiogenic behavior of ECs was also affected by modulation of cellular T-cad expression levels per se without inclusion of T-cad ligands in the gel. Adenovirus-mediated T-cad overexpression enhanced angiogenic rates in fibrin and collagen gels, whereas specific siRNA-driven downregulation of T-cad consistently reduced angiogenic sprouting, possibly caused by effects of T-cad expression levels per se on cell cycle progression and proliferation rates. 16 Furthermore, in multicellular in vitro systems and in vivo it is likely that T-cad ligation and overexpression represent 2 sides of the same process. Upregulation of surface T-cad levels would increase the chance of homophilic intercellular T-cad ligation on neighboring cells during vascular remodeling. One also cannot exclude the possibility of T-cad overexpression causing lateral cis contacts between T-cad molecules and a subsequent generation of intracellular signals caused by increased clustering and T-cad interactions with its putative membrane signaling partners. This hypothesis is supported by our observation that both overexpression and ligation of T-cad can activate the same signaling pathway and induce Akt phosphorylation (supplemental Figure I, available online at http://atvb.ahajournals.org). However, overexpression does not induce cell detachment and phenotypic modulation, and ligation results in phenotypic switch only when homophilic T-cad ligands are included in the matrix and not simply added to culture medium. Thus the functional consequences of T-cad activation might depend on the cell state (spreading cells versus already attached monolayer) and T-cad interactions with molecules participating in cell adhesion to the matrix.

In vivo effects of T-cad on neovascularization were studied using myoblast-mediated gene transfer to mouse skeletal muscle, a model that has specific advantages over other in vivo methods. Precise control of angiogenic stimulation in the target tissue is achieved by implantation of myoblast clones retrovirally transduced to homogenously express specific VEGF levels. Use of a single stimulatory growth factor at precise doses allows to standardize experimental conditions and to study even subtle effects of angiogenic regulators, these being more difficult in hind limb ischemia, Matrigel plaque, or tumor angiogenesis models where new vessel formation is induced by complex and multifactorial stimuli such as hypoxia or tumor environment. Our data demonstrate that implantation of myoblast clones co-expressing VEGF and secreted T-cad protein induces formation of capillaries of larger diameters compared with respective clones expressing the same dose of VEGF alone, unequivocally confirming the proangiogenic properties of T-cad. Importantly, whereas significantly increasing the caliber of the vessels at the sites of myoblast implantation, T-cad did not induce formation of any aberrant structures such as hemangiomas, often observed in response to delivery of higher VEGF concentrations. 5,19 Small increases in vessel size bring about large improvements in blood flow which is proportional to the fourth power of vessel radius, according to Poiseuille?s law. Therefore, these results suggest an interesting application for T-cad co-delivery to enhance the efficacy of VEGF-induced therapeutic angiogenesis at low and safe levels.

Mechanisms mediating T-cad-dependent stimulation of angiogenesis have yet to be clarified. T-cad-induced sprouting in vitro is not caused by induction of VEGF synthesis in EC. First, T-cad activation does not influence VEGF gene expression (see http://atvb.ahajournals.org). Second, VEGF receptor inhibition does not eliminate T-cad-induced increase in angiogenic rates. Stimulation of sprouting by T-cad in the spheroid and Nicosia models both in the absence and presence of VEGF further argues for VEGF-independence. However, in vivo T-cad failed to induce significant neovascularization in the absence of VEGF. The discrepancy between in vitro and in vivo findings with respect to VEGF-dependence could be explained by differences in EC activation states. In culture, cells exist in a state of activation caused by the presence of serum and growth factors in the culture medium, whereas in vivo EC are normally quiescent and in contact with pericytes and have a differentiated phenotype. Further, in vitro T-cad also increased VEGF-induced sprouting, whereas in vivo it did not affect vessel number, but only its diameter. These results are not contradictory if one considers that in vitro sprouting assays are really testing effects on proliferation and migration, but do not reproduce actual angiogenesis in vivo. The pro-migratory and proliferative effects of T-cad on EC in vitro translate into a modulation of VEGF-induced angiogenesis in vivo, affecting vessel diameter, but not length. Therefore, our data suggest that T-cad is not a primary angiogenic stimulus, but rather a modulator of angiogenesis that requires initial destabilization of the vessel by growth factors (eg, VEGF) to exert its facilitatory influence on EC phenotype conversion, proliferation and survival.

Identification of T-cad, a molecule that also participates in guidance of motor axons, as a regulator of angiogenesis further supports the surmise that nerves and blood vessels not only share architectural similarity and follow parallel trajectories during development but also often utilize identical mechanisms that regulate projection of neural and vascular structures and their organization into complex networks. 28 Interestingly, many regulators of neurogenesis and/or neovascularization localize in lipid rafts and caveolae, plasma membrane domains that function as assembling platforms for complexes between activated receptors and their signaling targets on the cell surface. Among them are guidance molecules ephrins, 29 semaphorins, 26 F3/contactin, 30 Rho family GTPases, regulatory components of PI3-kinase/Akt pathway 31 that also mediate T-cad effects in EC, as well as major caveolae protein caveolin-1 that mediates growth promotory influence of VEGF by eNOS and Flk-1/KDR translocation to the nucleus 32 and the knockout of which critically inhibits angiogenesis 33 in vivo, supporting a central role for these domains in neovascularization. T-cad is also located in caveolae-like domains. 12,13 Conceivably, lipid rafts may participate in T-cad effects on angiogenesis by permitting interaction of cell surface GPI-anchored T-cad with molecular adaptors that couple it to intracellular signaling systems.

In conclusion, this study reveals T-cad as a novel pro-angiogenic molecule. We propose a contribution of T-cad to the process of (patho)physiological neovascularization, given the present findings that T-cad increases endothelial sprouting in vitro and facilitates VEGF-induced angiogenesis in vivo, and the previous demonstrations of upregulation of T-cad on EC from newly formed tumor vasculature and very high expression levels of T-cad on vasa vasorum in the adventitial layer of human aorta. Silencing of T-cad in ECs may offer new possibilities for endothelial-targeted treatment of pathological conditions associated with excessive neovascularization. Alternatively, delivery of recombinant T-cad protein or modulation of T-cad expression on ECs using gene transfer might represent potential strategies to improve the outcome of growth factor-dependent angiogenic therapy of tissue ischemia.

Acknowledgments

We thank Drs Philippe Riou and Antoinette Lemoine for having provided the information about T-cad siRNA.

Sources of Funding

This study was supported by the Swiss National Science Foundation (grant no. 3100A0-105406), Herzkreislauf Stiftung, Swiss Cardiology Foundation, and EU-FP6 (grant no. 005206).

Disclosures

None.

【参考文献】
  Moulton KS. Plaque angiogenesis and atherosclerosis. Curr Atheroscler Rep. 2001; 3: 225-233.

Carmeliet P. Mechanisms of angiogenesis and arteriogenesis. Nat Med. 2000; 6: 389-395.

Yoon YS, Johnson IA, Park JS, Diaz L, Losordo DW. Therapeutic myocardial angiogenesis with vascular endothelial growth factors. Mol Cell Biochem. 2004; 264: 63-74.

Syed IS, Sanborn TA, Rosengart TK. Therapeutic angiogenesis: a biologic bypass. Cardiology. 2004; 101: 131-143.

Banfi A, von Degenfeld G, Blau HM. Critical role of microenvironmental factors in angiogenesis. Curr Atheroscler Rep. 2005; 7: 227-234.

Lee RJ, Springer ML, Blanco-Bose WE, Shaw R, Ursell PC, Blau HM. VEGF gene delivery to myocardium: deleterious effects of unregulated expression. Circulation. 2000; 102: 898-901.

Angst BD, Marcozzi C, Magee AI. The cadherin superfamily: diversity in form and function. J Cell Sci. 2001; 114: 629-641.

Vestal DJ, Ranscht B. Glycosyl phosphatidylinositol-anchored T-cadherin mediates calcium- dependent, homophilic cell adhesion. J Cell Biol. 1992; 119: 451-461.

Ivanov D, Philippova M, Antropova J, Gubaeva F, Iljinskaya O, Tararak E, Bochkov V, Erne P, Resink T, Tkachuk V. Expression of cell adhesion molecule T-cadherin in the human vasculature. Histochem Cell Biol. 2001; 115: 231-242.

Kudrjashova E, Bashtrikov P, Bochkov V, Parfyonova Y, Tkachuk V, Antropova J, Iljinskaya O, Tararak E, Erne P, Ivanov D, Philippova M, Resink TJ. Expression of adhesion molecule T-cadherin is increased during neointima formation in experimental restenosis. Histochem Cell Biol. 2002; 118: 281-290.

Wyder L, Vitaliti A, Schneider H, Hebbard LW, Moritz DR, Wittmer M, Ajmo M, Klemenz R. Increased expression of H/T-cadherin in tumor-penetrating blood vessels. Cancer Res. 2000; 60: 4682-4688.

Philippova MP, Bochkov VN, Stambolsky DV, Tkachuk VA, Resink TJ. T-cadherin and signal-transducing molecules co-localize in caveolin- rich membrane domains of vascular smooth muscle cells. FEBS Lett. 1998; 429: 207-210.

Philippova M, Ivanov D, Tkachuk V, Erne P, Resink TJ. Polarisation of T-cadherin to the leading edge of migrating vascular cells in vitro: a function in vascular cell motility? Histochem Cell Biol. 2003.

Ivanov D, Philippova M, Tkachuk V, Erne P, Resink T. Cell adhesion molecule T-cadherin regulates vascular cell adhesion, phenotype and motility. Exp Cell Res. 2004; 293: 207-218.

Philippova M, Ivanov D, Allenspach R, Takuwa Y, Erne P, Resink T. RhoA and Rac Mediate Endothelial Cell Polarization and Detachment Induced by T-Cadherin. Faseb J. 2005;Electronic publication ahead of print. http://www.fasebj.org/cgi/reprint/04-2430fjev1?ijkey=8gHmQWevo3t12.

Ivanov D, Philippova M, Allenspach R, Erne P, Resink T. T-cadherin upregulation correlates with cell-cycle progression and promotes proliferation of vascular cells. Cardiovasc Res. 2004; 64: 132-143.

Joshi MB, Philippova M, Ivanov D, Allenspach R, Erne P, Resink TJ. T-cadherin protects endothelial cells from oxidative stress-induced apoptosis. Faseb J. 2005; 19: 1737-1739.

Korff T, Augustin HG. Tensional forces in fibrillar extracellular matrices control directional capillary sprouting. J Cell Sci. 1999; 112: 3249-3258.

Ozawa CR, Banfi A, Glazer NL, Thurston G, Springer ML, Kraft PE, McDonald DM, Blau HM. Microenvironmental VEGF concentration, not total dose, determines a threshold between normal and aberrant angiogenesis. J Clin Invest. 2004; 113: 516-527.

Djonov V, Schmid M, Tschanz SA, Burri PH. Intussusceptive angiogenesis: its role in embryonic vascular network formation. Circ Res. 2000; 86: 286-292.

Ingber DE. Mechanical signaling and the cellular response to extracellular matrix in angiogenesis and cardiovascular physiology. Circ Res. 2002; 91: 877-887.

Huynh-Do U, Vindis C, Liu H, Cerretti DP, McGrew JT, Enriquez M, Chen J, Daniel TO. Ephrin-B1 transduces signals to activate integrin-mediated migration, attachment and angiogenesis. J Cell Sci. 2002; 115: 3073-3081.

Davies Cde L, Melder RJ, Munn LL, Mouta-Carreira C, Jain RK, Boucher Y. Decorin inhibits endothelial migration and tube-like structure formation: role of thrombospondin-1. Microvasc Res. 2001; 62: 26-42.

Yeh CH, Peng HC, Yang RS, Huang TF. Rhodostomin, a snake venom disintegrin, inhibits angiogenesis elicited by basic fibroblast growth factor and suppresses tumor growth by a selective alpha(v)beta(3) blockade of endothelial cells. Mol Pharmacol. 2001; 59: 1333-1342.

Gamble JR, Matthias LJ, Meyer G, Kaur P, Russ G, Faull R, Berndt MC, Vadas MA. Regulation of in vitro capillary tube formation by anti-integrin antibodies. J Cell Biol. 1993; 121: 931-943.

Serini G, Valdembri D, Zanivan S, Morterra G, Burkhardt C, Caccavari F, Zammataro L, Primo L, Tamagnone L, Logan M, Tessier-Lavigne M, Taniguchi M, Puschel AW, Bussolino F. Class 3 semaphorins control vascular morphogenesis by inhibiting integrin function. Nature. 2003; 424: 391-397.

Deroanne C, Vouret-Craviari V, Wang B, Pouyssegur J. EphrinA1 inactivates integrin-mediated vascular smooth muscle cell spreading via the Rac/PAK pathway. J Cell Sci. 2003; 116: 1367-1376.

Carmeliet P. Blood vessels and nerves: common signals, pathways and diseases. Nat Rev Genet. 2003; 4: 710-720.

Brantley-Sieders DM, Chen J. Eph receptor tyrosine kinases in angiogenesis: from development to disease. Angiogenesis. 2004; 7: 17-28.

Falk J, Bonnon C, Girault JA, Faivre-Sarrailh C. F3/contactin, a neuronal cell adhesion molecule implicated in axogenesis and myelination. Biol Cell. 2002; 94: 327-334.

Manes S, Ana Lacalle R, Gomez-Mouton C, Martinez AC. From rafts to crafts: membrane asymmetry in moving cells. Trends Immunol. 2003; 24: 320-326.

Griffoni C, Spisni E, Santi S, Riccio M, Guarnieri T, Tomasi V. Knockdown of caveolin-1 by antisense oligonucleotides impairs angiogenesis in vitro and in vivo. Biochem Biophys Res Commun. 2000; 276: 756-761.

Feng Y, Venema VJ, Venema RC, Tsai N, Caldwell RB. VEGF induces nuclear translocation of Flk-1/KDR, endothelial nitric oxide synthase, and caveolin-1 in vascular endothelial cells. Biochem Biophys Res Commun. 1999; 256: 192-197.

作者单位：Cardiovascular Signalling Group (M.P., D.I., R.A., T.R.), Department of Research, and Cell and Gene Therapy Group (A.B.), Departments of Surgery and of Research, Basel University Hospital, and Division of Cardiology (P.E.), Kantonsspital Luzern, Switzerland.