Disordered Purinergic Signaling Inhibits Pathological Angiogenesis in Cd/Entpd-Null Mice

【摘要】  CD39/ecto-nucleoside triphosphate diphosphohydrolase-type-1 (ENTPD1) is the dominant vascular ecto-nucleotidase that catalyzes the phosphohydrolysis of extracellular nucleotides in the blood and extracellular space. This ecto-enzymatic process modulates endothelial cell, leukocyte, and platelet purinergic receptor-mediated responses to extracellular nucleotides in the setting of thrombosis and vascular inflammation. We show here that deletion of Cd39/Entpd1 results in abrogation of angiogenesis, causing decreased growth of implanted tumors and inhibiting development of pulmonary metastases. Qualitative abnormalities of Cd39-null endothelial cell adhesion and integrin dysfunction were demonstrated in vitro. These changes were associated with decreased activation of focal adhesion kinase and extracellular signaling-regulated kinase-1 and -2 in endothelial cells. Our data indicate novel links between CD39/ENTPD1, extracellular nucleotide-mediated signaling, and vascular endothelial cell integrin function that impact on angiogenesis and tumor growth.
--------------------------------------------------------------------------------
CD39/ecto-nucleoside triphosphate diphosphohydrolase-1 (ENTPD1) catalyzes phosphohydrolysis of extracellular nucleoside di- and triphosphates to the monophosphate derivatives.1 Extracellular nucleotides released by platelets, leukocytes, and vascular cells signal endothelial cells through type-2 purinergic (P2) receptors termed P2Y (G protein-coupled receptors) or P2X (ligand-gated channels).2 Deletion of Cd39 has been shown to perturb P2 receptor signaling in Cd39-null mice, resulting in platelet dysfunction and disordered thromboregulation.3 Activation of P2 receptors appears to influence endothelial cell chemotactic and mitogenic responses in vitro.4 These data suggest roles for CD39 in modulating such responses during angiogenesis, and Matrigel plug-induced vascularization can be shown to be absent in Cd39-null mice.5
Angiogenesis is important in cancer progression and crucial for tumor growth and metastasis.6,7 The process of neovascularization is influenced by integrins,8 including endothelial Vß3.9 P2Y2 co-localizes with vß3 to modulate endothelial integrin function and migration.10 Ligand binding to integrins induces phosphorylation of focal adhesion kinase (FAK), activating extracellular signaling-regulated kinase-1 and -2 (ERK1/2).11,12 Activation of mitogen-activated protein kinase ERK1/2 promotes angiogenesis by blocking apoptosis and promoting proliferation.12
These observations suggest potential functional links between CD39, P2Y receptors, and endothelial integrins. We have therefore investigated the effects of Cd39 deletion on tumor angiogenesis and metastasis in vivo and studied underlying pathogenic mechanisms in vitro.

【关键词】  disordered purinergic signaling inhibits pathological angiogenesis cd/entpd-null

Materials and Methods

Antibodies and Other Reagents

Antibodies against PECAM, pan-endothelial antigen (MECA-32), CD3, CD19, CD41, LY-6G, and LY-6C were from BD Biosciences (San Jose, CA). Antibodies against perlecan and pericyte marker (NG2) were from Chemicon (Temecula, CA). Antibodies against platelet-derived growth factor-ß, FLK-1, and V-integrin were from Santa Cruz Biotechnology (Santa Cruz, CA). Antibodies to the macrophage marker (F4/80) were from Serotec (Raleigh, NC). Antibodies against Fab2 and fibrinogen/fibrin were from DAKO (Carpinteria, CA). Antibodies against reticular fibroblasts (clone ER-TR7) were from Cedarlane Laboratories (Burlington, ON, Canada). The fluorescein-labeled anti-smooth muscle -actin and MMP-9 antibodies were from Sigma-Aldrich (St. Louis, MO). Antibodies against FAK and phosphorylated-FAK (Tyr397) were purchased from BD Transduction Laboratories (Franklin Lakes, NJ). Antibodies against ERK1/2 and phosphorylated ERK 1/2 (Thr202/Tyr204) were from Cell Signaling Technology (Beverly, MA). Secondary antibodies were from Vector Laboratories (Burlingame, CA). The luciferase assay kit was from Promega (Madison, WI). Grade VII potato apyrase (a soluble NTPDase) and vitronectin from rat plasma were from Sigma-Aldrich.

Mouse Tumor Inoculation

Animals were housed in accordance with the guidelines from the American Association for Laboratory Animal Care. The Beth Israel Deaconess Medical Center Institutional animal care and use committees approved all research protocols. C57BL/6X129svj or C57BL6 backcrossed (for more than six generations) strains of wild-type and type-matched Cd39-null mice were inoculated with 3.0 x 106 cells B16-F10 melanoma (n = 8 and 6), B16-CG (stable ß-HCG expression vector transfected) melanoma13 (n = 4), or Lewis lung cancer (LLC) cells (n = 4) into the dorsal subcutaneous space. Tumor size was measured daily, and the tumor volume was calculated using the formula V = L.W2/6, where L is the longer diameter and W the short diameter of tumor. Wild-type and Cd39-null mice were also injected with 5.0 x 105 B16-F10 melanoma into the subperitoneal plane of an abdominal wall flap (n = 6). The subperitoneal tumors were examined at 2-day intervals by limited laparotomy, the tumor size measured, and new blood vessels growing into the tumor were counted by microscopy. At 15 days, mice were sacrificed and the tumors analyzed by immunohistochemistry.

Measurement of Mouse ß-HCG Excretion

The urine of mice inoculated with B16-CG melanoma cells was collected as described.13 Urine ß-HCG was measured with ß-HCG enzyme-linked immunosorbent assay kits (Alpha Diagnostic, San Antonio, TX) and urine creatinine with the Creatinine assay kit (Sigma Diagnostics, Inc., St. Louis, MO).

Pulmonary Metastasis Model

B16-F10 cells (1.5 x 105) were injected into the inferior vena cava of C57BL/6X129svj strain wild-type (n = 6) and type-matched Cd39-null mice (n = 4). After 15 days, the mice were euthanized and the pulmonary metastases analyzed for size and number.

Measurement of Tumor Cell Sequestration in the Pulmonary Vasculature

Luciferase-expressing B16/F10 cells (lucB16/F10) were established as described.14 luc-B16/F10 cells (5 x 105) were injected into penile veins of age- and sex-matched C57/BL6 wild-type (n = 3) and Cd39-null mice (n = 3). One hour later, the mice were sacrificed, the lungs snap-frozen in liquid nitrogen, and the tissue homogenized in luciferase lysis buffer (Promega). The 20 µl of lung homogenate was added to 100 µl of luciferase assay buffer (Promega) and luciferase activity was measured using a luminometer. Luciferase activity was expressed as relative light units (RLUs) normalized for protein content and background of control lungs.

Immunohistochemical Staining

Frozen sections of implanted B16-F10 tumors were analyzed by immunohistochemistry and immunofluorescence as described.5

Primary Lung Endothelial Cell Isolation

Wild-type and Cd39-null mouse lungs were minced, collagenase-digested (Life Technologies, Inc., Grand Island, NY), and strained and the resulting cell suspension plated on flasks coated with 0.2% gelatin (Sigma). Endothelial cells were purified by a single negative (FC-RII/III antibody; Pharmingen, La Jolla, CA) and two positive (ICAM-2; Pharmingen) sorts using anti-rat IgG-conjugated magnetic beads (Dynal; Invitrogen, Carlsbad, CA) producing a >98% pure population.

Endothelial Cell Proliferation and Growth Factor Signaling

Wild-type and Cd39-null C57BL6 endothelial cells were treated for 24 hours with vascular endothelial growth factor (VEGF) (10 ng/ml), insulin (10 nmol/L), hepatocyte growth factor (50 ng/ml), fibroblast growth factor (10 ng/ml), or insulin-like growth factor-1 (10 ng/ml). Cells were then pulsed with thymidine for 8 hours and uptake measured as described.15 VEGF-induced activation of ERK1/2 was measured after stimulation of wild-type and Cd39-null endothelial cells with VEGF (10 ng/ml) for 5 minutes. ERK1/2 activation was expressed as the ratio of phosphorylated to unphosphorylated ERK1/2 after Western blot analysis.

Adhesion Assays

Adhesion assays were performed as previously described.16 For the pretreatment experiments to deplete extracellular nucleotides, cells were preincubated with a soluble NTPDase, grade VII apyrase (5 U/ml), in serum-free conditions for 6 hours before assay. For integrin activation experiments, manganese chloride (MnCl2, 500 µmol/L) was added to cells during the adhesion process.

v Integrin Expression

Wild-type and Cd39-null endothelial cells were stimulated with 100 µmol/L UTP for 0, 2, 4, or 6 hours. Cells were lysed in radioimmunoprecipitation assay buffer, and 250 µg of protein immunoprecipitated using V integrin antibody. Total V expression was determined in a semiquantitative manner by Western blot.

FAK and ERK1/2 Activation Assay

Wild-type and Cd39-null endothelial cells were serum-starved for 30 hours, trypsinized, and added to vitronectin-precoated dishes. Total cell extracts were prepared, and sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blotting with antibodies to FAK, phosphorylated FAK (Tyr397), ERK1/2, and phosphorylated ERK1/2 (Thr202/Tyr204) were performed. For apyrase pretreatments, cells were incubated with grade VII apyrase (5 U/ml) for 6 hours, trypsinized, washed, and plated onto vitronectin.

Statistical Analysis

The Student??s t-test with one-tailed distribution was used to analyze data (Excel; Microsoft, Redmond, WA).

Results

Transplanted Tumor Growth Is Decreased in Cd39-Null Mice

Matched wild-type and Cd39-null mice were injected subcutaneously with murine melanoma (B16-F10, American Type Culture Collection: CRL-6457; Manassas, VA) or lung carcinoma (LLC, American Type Culture Collection: CRL-1642) cells. The growth of both tumors was significantly attenuated in Cd39-null mice (Figure 1a) . Wild-type mice had significantly larger B16-F10 (Figure 1b) and LLC (not shown) tumor volumes at 1 week. Tumor growth then entered the exponential phase in wild-type mice but in Cd39-null mice was severely retarded up to 15 days (Figure 1, a and b) . Release of ß-HCG after implantation of B16-CG melanoma cells was also used to estimate tumor volume.13 ß-HCG excretion correlates well with tumor volume in this model (R2 = 0.96, data not shown). Urine ß-HCG excretion was significantly decreased in Cd39-null mice (Figure 1c) .

Figure 1. Deletion of CD39 is associated with decreased tumor growth. a: Representative examples of tumor growth in wild-type (left) and Cd39-null (right) mice. Upper tumor is LLC, and lower is B16-F10 melanoma. b: Measurement of murine B16-F10 and LLC (not shown, comparable data) subcutaneous tumor growth in wild-type () and Cd39-null () mice. Bars show mean tumor volume in mm3 ?? SD. *P < 0.05. c: ß-HCG excretion after subcutaneous inoculation with B16-CG tumors. Wild type (), Cd39-null (). Bars show urinary ß-HCG/creatinine (mIU/mg) ?? SD. *P < 0.05. d: Subperitoneal B16-F10 tumor examined for 8 days after implantation in wild-type and Cd39-null mice. Vessel counts in subperitoneal B16-F10 tumors in wild-type () and Cd39-null () mice. Bars show mean number of vessels ?? SEM. *P < 0.05.

Tumor Angiogenesis Is Decreased in Cd39-Null Mice

Direct visualization of B16-F10 tumors in wild-type and Cd39-null mice 7 days after implantation confirmed poor tumor growth in Cd39-null mice (not shown). In addition, this model allowed direct assessments of neovascularization of the tumor by dissection microscopy. By 2 days after tumor injection, wild-type mice had evidence of new vessels infiltrating the tumor that increased steadily in number (Figure 1d) . This neovascularization was abrogated in Cd39-null mice.

Tumor Metastases Are Decreased in Cd39-Null Mice

We next studied development of pulmonary metastases in wild-type and Cd39-null mice after intravascular inoculation of 1.5 x 105 B16-F10, as described.17 Wild-type lungs became infiltrated with tumor cells and, at 15 days after intravascular injection, tumor metastases had increased in size and replaced the entire lung (Figure 2a) . In contrast, Cd39-null lungs had limited tumor numbers that were also of smaller size. Wild-type lung weights were also significantly increased more than Cd39-null levels (1738-mg wild type versus 604-mg Cd39-null, P < 0.01; Figure 2b ).

Figure 2. Decreased growth of inoculated tumor metastases, in Cd39-null mice. a: Representative examples of wild-type (right) and Cd39-null (left) mouse lungs at 15 days after tumor cell injection intravenously. b: Wild-type () and Cd39-null () lung weights 15 days after intravenous tumor injection. a and b: Mice were injected with 1.5 x 105 B16-F10 cells per 25 g of body weight into the inferior vena cava and lungs were examined at 15 days. Bars represent mean weight ?? SEM. *P < 0.006. c: Retention of cell-associated luciferase activity in wild-type () and Cd39-null () mice 1 hour after injection of 5 x 105 luciferase-expressing B16-F10 cells (lucB16/F10). Bars represent relative light units (RLUs) per µg of lung ?? SD.

We next examined sequestration of tumor cells within the lung vasculature by determining short-term retention of injected tumor cells expressing luciferase. These experiments were done in wild-type and Cd39-null mice and assays conducted at 1 hour after injection of 5 x 105 luciferase-expressing B16/F10 cells (lucB16/F10). There was no significant difference in luciferase activity in wild-type and Cd39-null lungs (wild-type 11.1 ?? 3.1 RLU/µg versus Cd39-null 13.9 ?? 4.6 RLU/µg, P = 0.1). The decreased development of lung metastases in Cd39-null mice did not appear secondary to decreased seeding of circulating tumor cells within the pulmonary vasculature (Figure 2c) . These data suggest that the decreased development of pulmonary metastases in Cd39-null animals might be secondary to subsequent failure of growth and potentially defective angiogenesis with an endothelial cell defect.

Immunopathology

Staining with the endothelial marker PECAM-1 showed growth of blood vessels into the implanted B16-F10 tumor mass in wild-type mice (Figure 3A) . In contrast, tumors in Cd39-null mice were devoid of vessels, which were confined to the interface between the tumor and adjacent normal tissue (Figure 3B) .

Figure 3. Immunohistochemistry of B16-F10 tumors. Tumors were implanted and grown in the subcutaneous space in wild-type (A, C, E, G, and I) and Cd39-null (B, D, F, H, and J) mice. A and B: CD31-positive endothelial cells can be seen infiltrating tumor (arrow) in wild-type (A), but not Cd39-null animals (B). Endothelial cells are confined to the interface between tumor and surrounding tissue in Cd39-null mice (arrow). C and D: There was marked infiltration of F4/80-positive macrophages (arrows) in tumors grown in wild-type animals (C), in particular in the necrotic central zone (*). In contrast, macrophages were mostly absent within tumors grown in Cd39-null mice (D). E and F: Platelets (integrin IIb-positive) were associated with vessels within the tumor mass in wild-type mice (E) but were confined to the tumor interface in Cd39-null animals (F). G and H: In tumors grown in wild-type mice (G), perlecan staining showed basal lamina associated with tumor vessels (arrow) but not tumor cell nests. In contrast, in tumors grown in Cd39-null mice (H), which were primarily devoid of vasculature, basal lamina divided individual tumor cell nests into gland-like structures (*). I and J: Trichrome staining showed tumors cells (arrow) in wild-type mice (I) invading deep into the underlying muscle layer (*). In contrast, in tumors grown in Cd39-null mice (J), tumor cell invasion was absent, and the tumor mass and muscle layer (*) was clearly separated. Scale bars = 20 µm.

Significant numbers of macrophages had migrated to the interface between tumor tissue and adjacent normal tissue and into the viable and necrotic areas within the tumors in wild-type animals (Figure 3C) . In contrast, few macrophages were noted in Cd39-null sections (Figure 3D) . Tumors and surrounding tissues in wild-type and Cd39-null animals contained similar amounts of B and T lymphocytes as well as granulocytes (Table 1) . Platelet deposition was noted within vessels in the tumors in wild-type mice (Figure 3E) but was confined to the tumor interface in Cd39-null animals (Figure 3F) .

Table 1. Distribution of Cell-Type-Specific Markers, Cellular Activation Markers, and Extracellular Matrix Components in Wild-Type and Cd39-Null Mice after B16-F10 Tumor Implantation

Basal lamina deposition, namely anti-perlecan staining, was noted around tumor vessels and muscle bundles in wild-type mice (Figure 3G) . In Cd39-null mice, this was present in adjacent normal vessels encompassing tumor mass (Figure 3H) . Tumor cell islands were enveloped in basal lamina, forming glandular-like structures in Cd39-null mice (Figure 3H) . This was absent in tumors in wild-type mice (Figure 3G) potentially indicating a higher level of tumor cell differentiation in tumors in Cd39-null mice. In addition to enhanced growth, tumors in wild-type mice were further capable of invading the underlying dermal muscle fascia (Figure 3I) , whereas those in Cd39-null mice did not display invasive behavior (Figure 3J) .

In tumors implanted into wild-type mice, vessels were partially enveloped by pericytes and surrounded by interstitial fibroblasts (Figure 4D) . In contrast, tumors in Cd39-null mice were devoid of endothelial cells, pericytes, and fibroblasts (Figure 4H) , indicating pluripotent cellular migration defects in Cd39-null mice. Markers of blood vessel activation (namely dissolution of the enveloping basal lamina, increased pericyte expression of NG2, platelet-derived growth factor-ß receptors, decreased pericyte expression of smooth muscle -actin, and increased endothelial expression of VEGFR2 receptors) were apparent at the interface between tumor and normal adjacent tissue in wild-type mice. These changes were not seen in Cd39-null mice (Table 1) . Platelet and fibrin deposition were noted in and around infiltrating vessels in wild-type tumors (Figure 4L) but in Cd39-null mice were confined to the rim of tissue surrounding the tumor (Figure 4P) . Thus, despite deposition of a provisional matrix, blood vessels failed to appropriately activate and migrate into the tumor mass in Cd39-null mice (Table 1) .

Figure 4. Triple immunofluorescent staining patterns of B16-F10 tumors. B16-F10 tumors were grown in the subcutaneous space in wild-type (ACD, ICL) and Cd39-null (ECH, MCP) mice and stained with antibodies recognizing: pericytes and fibroblasts (pan-reticular fibroblast marker) (A and E); pericytes and smooth muscle cells (smooth muscle -actin) (B and F); fibrin (I and M); platelets (integrin IIb) (J and N); and endothelial cells (CD31) (C, G, K, and O) and their composites (D, H, L, and P). The dotted white line delineates the interface between the tumor mass and adjacent normal tissue. The arrow indicates the direction of the tumor. ACD: In tumors grown in wild-type mice the endothelial lining (blue) of vessels were only partially covered by mature smooth muscle -actin-expressing pericytes (green) and interspersed in a network of interstitial fibroblasts (red). Note that the interstitial fibroblasts do not express smooth muscle -actin. ECH: In Cd39-null mice, the endothelial lining (blue) of vessels in normal tissue adjacent to the tumor were to a large extent covered by smooth muscle -actin-expressing pericytes (green) and surrounded by interstitial fibroblasts (red). Similar structures were absent within the actual tumor mass. ICL: In tumors in wild-type mice, a provisional matrix consisting of fibrin (red) and platelets (green) were associated with a large portion of CD31-positive (blue) tumor vessels (arrow). MCP: In Cd39-null mice fibrin (red) and platelets (green) were concentrated to a compacted rim of CD31 (blue)-positive vessels (arrow) at the interface between the tumor and normal adjacent tissue. Note nonspecific fluorescent staining in the central necrotic zone. Scale bars = 20 µm.

Endothelial Cell Proliferation in Vitro Does Not Correlate with Defective Angiogenesis in Cd39-Null Mice in Vivo

Purinergic signals have been implicated in promoting endothelial cell proliferation, as well as migration, during angiogenesis. However, we found no difference in endothelial proliferation by thymidine uptake in wild-type and Cd39-null cells at baseline (100 ?? 39% wild-type versus 97 ?? 44% Cd39-null, P = 0.47) and after stimulation with VEGF (10 ng/ml) (135 ?? 19% wild-type versus 133 ?? 37% Cd39-null, P = 0.47) in vitro (not shown). Similarly, we showed no difference in wild-type and Cd39-null endothelial cell proliferation after stimulation with a number of endothelial mitogens, including insulin, insulin-like growth factor-1, hepatocyte growth factor, and fibroblast growth factor (data also not shown). In addition, VEGF-induced activation of ERK1/2 was not attenuated in Cd39-null endothelial cells (Figure 5a) .

Figure 5. Cd39 deletion and associated integrin vß3 dysfunction. a: Phosphorylation of ERK1/2 (Thr202/Tyr204) in wild-type () and Cd39-null () endothelial cells at baseline and after stimulation with VEGF (10 ng/ml) for 5 minutes. Bar chart represents densitometry results (means ?? SD) of phosphorylated ERK1/2 corrected for total ERK1/2. *P < 0.05. b: Wild-type () and Cd39-null () endothelial adhesion to collagen, fibronectin, and vitronectin. Results expressed as a percentage of wild-type adhesion to collagen ?? SEM. *P < 0.05. c: Wild-type () and Cd39-null () endothelial cell adhesion to vitronectin after a 6-hour pretreatment with soluble apyrase (5 U/ml). Results expressed as mean percentage of wild-type adhesion to vitronectin ?? SEM. *P < 0.05. d: The v-integrin subunit expression in wild-type and Cd39-null endothelial cells was determined by immunoprecipitation and Western blot analysis. Cells were treated with UTP (100 µmol/L) for 0, 2, 4, and 6 hours. No differences were evident (not shown). Bar chart represents wild-type () and Cd39-null () endothelial cell adhesion to vitronectin after treatment with MnCl2 (500 µmol/L). Results expressed as a percentage of wild-type adhesion to vitronectin ?? SEM. *P < 0.05.

Deletion of Cd39 Is Associated with Endothelial Cell-Selective Integrin Dysfunction

Purinergic signaling has been suggested to modulate the function of a number of integrins, including Vß3, the vitronectin receptor.10 Therefore, we measured the adhesion of wild-type and Cd39-null endothelial cells to a number of basement membrane substrates, including vitronectin. When contrasted to wild-type cells, Cd39-null endothelial cells exhibited normal, or near-normal, attachment to gelatin and fibronectin but had markedly decreased adhesion to vitronectin (Figure 5b) . This defect was corrected by pretreatment of endothelial cells with soluble NTPDases (Figure 5c) .

UTP has previously been shown to up-regulate the expression of the V integrin subunit in human umbilical vein endothelial cells.18 However, we found no substantial difference between total V integrin levels in wild-type and Cd39-null endothelial cells. Nor did UTP increase the levels of V expression in either cell type by Western blot analysis (not shown). In addition, the surface expression of V was determined to be equal in wild-type and Cd39-null endothelial cells by immunofluorescence (data not shown). These data imply that qualitative differences in integrin activation are responsible, at least in part, for the defective adhesion phenotype.

MnCl2 is a potent effector of a number of integrin-mediated adhesion events and is thought to activate integrins by binding to specific cation-binding sites and inducing conformational changes in the ligand-binding site.19 In keeping with the hypothesis that Vß3 activation is perturbed in Cd39-null endothelial cells, treatment with exogenous MnCl2 was able to correct the defective adhesion of Cd39-null endothelial cells to vitronectin in vitro (Figure 5d) .

Integrin-Associated Signaling Pathways in Cd39-Null Endothelial Cells

We also studied the effect of Cd39 deletion on the downstream signaling pathways post-integrin ligation. In keeping with the defective adhesion of Cd39-null endothelial cells to vitronectin, deletion of Cd39 was also associated with poor activation of FAK after Vß3 ligation. Wild-type endothelial cells exhibited robust phosphorylation of FAK 15 and 30 minutes after plating onto vitronectin, whereas Cd39-null cells had significantly attenuated FAK activation (Figure 6a) . FAK phosphorylation in Cd39-null cells was 60% of the wild-type levels at 15 minutes (P < 0.05) and approximated half of wild-type levels at 30 minutes (P < 0.05). FAK activation reached wild-type levels in Cd39-null cells after 1 hour of adhesion (data not shown). In addition, the downstream activation of MAPK after integrin ligation was defective in Cd39-null cells. Phosphorylation of downstream ERK1/2 was half that of wild-type levels in Cd39-null endothelial cells at 30 minutes after vitronectin plating (Figure 6c , P < 0.015).

Figure 6. Deletion of Cd39 is associated with defective vß3-mediated intracellular signaling. a and c: Graphs represent phosphorylation of FAK (Tyr 397) (a) and ERK1/2 (Thr202/Tyr204) (c) in wild-type () and Cd39-null () endothelial cells at 0, 15, and 30 minutes after plating onto vitronectin. b and d: Phosphorylation of FAK (b) and ERK1/2 (d) in wild-type () and Cd39-null () endothelial cells pretreated for 6 hours with soluble Apyrase (5 U/ml). Bar charts represent densitometry results (means ?? SEM) of phosphorylated FAK or ERK1/2 divided by total FAK or ERK1/2, respectively. *P < 0.05. Representative blots are shown below graphs.

Decreased Activation of Vß3 in Cd39-Null Endothelial Cells Is Associated with P2 Receptor Dysfunction

UTP has been shown to be chemoattractant to endothelial cells in vitro, implying a role for purinergic signaling in integrin activation.4,18 We hypothesized that, analogous to platelets, Cd39-null endothelial cells undergo P2 receptor desensitization and associated integrin inactivation. To test this, we pretreated wild-type and Cd39-null endothelial cells with soluble NTPDases, to reconstitute ecto-enzymatic functions of CD39 and restore P2 receptor functions, and then washed cells to remove residual apyrase activity before the adhesion assays.3

Apyrase pretreatment corrected the defective vitronectin adhesion of Cd39-null cells, as shown above (Figure 5c) , and normalized the recruitment of FAK and downstream signaling to ERK1/2 to wild-type levels (Figure 6, b and d) . Apyrase pretreatment had minimal effects on adhesion and on FAK and ERK1/2 activation in wild-type endothelial cells.

Discussion

We have shown that abnormal purinergic signaling in the Cd39-null vasculature results in profound defects in tumor growth and angiogenesis, indicating novel mechanisms of tumor angiogenesis. These data are consistent with our previous finding of defective angiogenesis studying Matrigel plug models in these mutant mice.5

Cd39-null mice exhibit the normal initial phases of angiogenesis, including mother vessel formation, increased vascular permeability, and fibrin deposition. However, there is a profound defect in the migration of endothelial and inflammatory cells into tumors in Cd39-null animals. Our in vitro data suggests that this defect is not caused by defects in endothelial proliferation. Although there are substantial defects in Cd39-null cellular adhesion noted in vitro, sequestration of injected tumor cells by the pulmonary vascular bed of Cd39-null mice does not contribute in a major way to the failure of metastatic tumor growth in these mutant mice. Importantly, there is markedly abnormal angiogenesis in Cd39-null mice characterized by poor recruitment of surrounding pericytes and smooth muscle cells to new blood vessels. Because embryonic vascular development is normal in Cd39-null mice,3 our data suggest that purinergic signaling has differential roles in physiological blood vessel development as opposed to pathological angiogenesis.

In vitro analyses indicate that the cellular migration defect in Cd39-null mice is likely to be secondary to abnormal adhesion and vitronectin receptor function in Cd39-null endothelial cells. The reconstitution of Cd39-null integrin function by apyrase pretreatment and the hydrolysis of extracellular nucleotides or by directed integrin effects by MnCl2 implies that P2 receptor desensitization (and secondary integrin inactivation) is responsible for these findings (Figure 5) .

Although several members of the integrin family have been implicated in angiogenesis, the largest body of data has linked the vitronectin receptor Vß3 with the promotion of neovascularization. Numerous Vß3 antagonists, including blocking antibodies directed at the extracellular domain of vß3 (LM609), inhibit angiogenesis in experimental animal models9 and are currently undergoing human clinical trials.20

Despite the efficacy of these Vß3 inhibitors, genetic deletion of Vß3 paradoxically results in increased tumor angiogenesis.21 It has been proposed that integrins might provide positive and/or negative signals to cells in the presence of appropriate extracellular ligands. The absence of these interactions could result in loss of putative negative signal(s) and enhanced angiogenesis.22 These observations imply that Vß3 could be involved in the fine-tuning of angiogenesis responses in vivo.

We have focused on elements of nucleotide-mediated signaling and shown that normal P2-type purinergic responses are required for Vß3 function and for the process of angiogenesis. Cellular signaling by G protein-coupled receptors (GPCRs), such as P2Y receptors, involves not only the coupling of the receptors to G proteins but also the formation of large protein complexes that assist in transmitting an extracellular signal to an intracellular response.23,24 P2Y2 contains an integrin-binding, arginine-glycine-aspartic acid (RGD) domain in its first extracellular loop, suggesting interactions with integrin Vß3. Mutation of the RGD sequence to RGE decreased the co-localization of v-integrins with P2Y2 10-fold and greatly impaired UTP-induced phosphorylation of FAK.10

Direct association between P2Y2 and v-integrins seems to also be necessary for UTP-induced endothelial cell migration, implying an important link between purinergic signaling, integrin function, and proangiogenic cellular behaviors.18 It has previously been shown that ATP and UTP activate FAK by calcium-dependent signaling mechanisms.18 The Vß3integrin-mediated activation of FAK can be shown to be defective in Cd39-null endothelial cells. Deletion of Cd39 results in differential P2 receptor desensitization, which subsequently modulates the activity of Vß3. In keeping with this supposition, apyrase pretreatment of Cd39-null endothelial cells, which corrects P2 receptor desensitization, normalizes Vß3 functions.

We have recently noted that the N terminus of CD39 interacts with the multiadaptor scaffolding membrane phosphoprotein RanBPM.25 RanBPM binds the hepatocyte growth factor receptor cMET and modulates the Ras-ERK-SRE pathway.26 In addition, RanBPM interacts with integrins such as LFA-1 (lymphocyte function-associated antigen-1) and may affect signaling pathways directly.27 These results suggest that CD39 may regulate cellular signal transduction pathways and integrin function by two separate mechanisms: by first modulating P2 receptor function after hydrolysis of extracellular nucleotides, and secondly, by interacting with other signaling proteins directly, possibly using RanBPM as a bridge.

There are other potential factors that could be also impacted on by the NTPDase1 biochemical functions. In addition to modulating P2 receptor signaling by nucleotide hydrolysis and scavenging, CD39 in association with CD73 ultimately generates adenosine, which also has important proangiogenic functions. For example, adenosine, acting via the A2A receptor, promotes wound healing by stimulating angiogenesis.28 Second, the A2A receptor has recently been implicated in the inhibition of antitumor T cells within the adenosine-rich tumor microenvironment.29

In these currently submitted angiogenesis studies, we have preferentially detailed the role of endothelial cell integrin dysfunction in Cd39-null mice. Some aspects of the Cd39-null antitumor phenotype could further be explained by decreased adenosine generation. In this regard, we have evidence that CD39 forms an important part of the immunosuppressive apparatus of regulatory T cells (Treg), at least in part via the generation of adenosine.30 We do not address the role of T cells in immune regulation of tumors here.

In conclusion, we have provided evidence for the importance of extracellular nucleotides in modulating endothelial cell responses in models of pathological angiogenesis. We have also described novel links between purinergic signaling and integrin activation that are modulated by CD39. These findings raise the possibility of new anti-angiogenesis therapies based on modulating extracellular nucleotide-mediated signals. These goals could be achieved by either blocking the action of CD39 directly or by targeting those nucleotide (P2) receptors that are crucial in the process of angiogenesis.

Acknowledgements

We thank Dr. Julie Lively, Medicine, Beth Israel Deaconess Medical Center, for help with endothelial cellular cultures; and Dr. Takashi Murakami, Jichi Medical Center, for the kind gift of luciferase-expressing B16/F10 cells (lucB16/F10).

【参考文献】
  Kaczmarek E, Koziak K, Sevigny J, Siegel JB, Anrather J, Beaudoin AR, Bach FH, Robson SC: Identification and characterization of CD39/vascular ATP diphosphohydrolase. J Biol Chem 1996, 271:33116-33122

Fredholm BB, Abbracchio MP, Burnstock G, Daly JW, Harden TK, Jacobson KA, Leff P, Williams M: Nomenclature and classification of purinoreceptors. Pharmacol Rev 1994, 46:143-152

Enjyoji K, Sevigny J, Lin Y, Frenette PS, Christie PD, Esch JS, Imai M, Edelberg JM, Rayburn H, Lech M, Beeler DL, Csizmadia E, Wagner DD, Robson SC, Rosenberg RD: Targeted disruption of Cd39/ATP diphosphohydrolase results in disordered hemostasis and thromboregulation. Nat Med 1999, 5:1010-1017

Satterwhite CM, Farrelly AM, Bradley ME: Chemotactic, mitogenic, and angiogenic actions of UTP on vascular endothelial cells. Am J Physiol 1999, 45:H1091-H1097

Goepfert C, Sundberg C, Sevigny J, Enjyoji K, Hoshi T, Csizmadia E, Robson SC: Disordered cellular migration and angiogenesis in Cd39-null mice. Circulation 2001, 104:3109-3115

Hanahan D, Folkman J: Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 1996, 86:353-364

Carmeliet P, Jain RK: Angiogenesis in cancer and other diseases. Nature 2000, 407:249-257

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

Hodivala-Dilke KM, Reynolds AR, Reynolds LE: Integrins in angiogenesis: multitalented molecules in a balancing act. Cell Tissue Res 2003, 314:131-144

Erb L, Liu J, Ockerhausen J, Kong Q, Garrad RC, Griffin K, Neal C, Krugh B, Santiago-Perez LI, Gonzalez FA, Gresham HD, Turner JT, Weisman GA: An RGD sequence in the P2Y2 receptor interacts with Vß3 integrins and is required for Go-mediated signal transduction. J Cell Biol 2001, 153:491-501

Parsons JT: Focal adhesion kinase: the first ten years. J Cell Sci 2003, 116:1409-1416

Berra E, Milanini J, Richard DE, Le Gall M, Vinals F, Gothie E, Roux D, Pages G, Pouyssegur J: Signaling angiogenesis via p42/p44 MAP kinase and hypoxia. Biochem Pharmacol 2000, 60:1171-1178

Shih IM, Torrance C, Sokoll LJ, Chan DW, Kinzler KW, Vogelstein B: Assessing tumors in living animals through measurement of urinary ß-human chorionic gonadotropin. Nat Med 2000, 6:711-714

Sato A, Ohtsuki M, Hata M, Kobayashi E, Murakami T: Antitumor activity of IFN- in murine tumor models. J Immunol 2006, 176:7686-7694

Maeshima Y, Colorado PC, Torre A, Holthaus KA, Grunkemeyer JA, Ericksen MB, Hopfer H, Xiao Y, Stillman IE, Kalluri R: Distinct antitumor properties of a type IV collagen domain derived from basement membrane. J Biol Chem 2000, 275:21340-21348

Hodivala-Dilke KM, McHugh KP, Tsakiris DA, Rayburn H, Crowley D, Ullman-Cullere M, Ross FP, Coller BS, Teitelbaum S, Hynes RO: ß3-Integrin-deficient mice are a model for Glanzmann thrombasthenia showing placental defects and reduced survival. J Clin Invest 1999, 103:229-238

Palumbo JS, Kombrinck KW, Drew AF, Grimes TS, Kiser JH, Degen JL, Bugge TH: Fibrinogen is an important determinant of the metastatic potential of circulating tumor cells. Blood 2000, 96:3302-3309

Kaczmarek E, Erb L, Koziak K, Jarzyna R, Wink MR, Guckelberger O, Blusztajn JK, Trinkaus-Randall V, Weisman GA, Robson SC: Modulation of endothelial cell migration by extracellular nucleotides: involvement of focal adhesion kinase and phosphatidylinositol 3-kinase-mediated pathways. Thromb Haemost 2005, 93:735-742

Smith JW, Piotrowicz RS, Mathis D: A mechanism for divalent cation regulation of ß3-integrins. J Biol Chem 1994, 269:960-967

Gutheil JC, Campbell TN, Pierce PR, Watkins JD, Huse WD, Bodkin DJ, Cheresh DA: Targeted antiangiogenic therapy for cancer using Vitaxin: a humanized monoclonal antibody to the integrin Vß3. Clin Cancer Res 2000, 6:3056-3061

Reynolds LE, Wyder L, Lively JC, Taverna D, Robinson SD, Huang X, Sheppard D, Hynes RO, Hodivala-Dilke KM: Enhanced pathological angiogenesis in mice lacking ß3 integrin or ß3 and ß5 integrins. Nat Med 2002, 8:27-34

Carmeliet P: Integrin indecision. Nat Med 2002, 8:14-16

Pierce KL, Premont RT, Lefkowitz RJ: Signalling: seven-transmembrane receptors. Nat Rev Mol Cell Biol 2002, 3:639-650

Hall RA, Lefkowitz RJ: Regulation of G protein-coupled receptor signaling by scaffold proteins. Circ Res 2002, 91:672-680

Wu Y, Sun X, Kaczmarek E, Dwyer KM, Bianchi E, Usheva A, Robson SC: RanBPM associates with CD39 and modulates ecto-nucleotidase activity. Biochem J 2006, 396:23-30

Wang D, Li Z, Messing EM, Wu G: Activation of Ras/Erk pathway by a novel MET-interacting protein RanBPM. J Biol Chem 2002, 277:36216-36222

Denti S, Sirri A, Cheli A, Rogge L, Innamorati G, Putignano S, Fabbri M, Pardi R, Bianchi E: RanBPM is a phosphoprotein that associates with the plasma membrane and interacts with the integrin LFA-1. J Biol Chem 2004, 279:13027-13034

Montesinos MC, Shaw JP, Yee H, Shamamian P, Cronstein BN: Adenosine A2A receptor activation promotes wound neovascularization by stimulating angiogenesis and vasculogenesis. Am J Pathol 2004, 164:1887-1892

Ohta A, Gorelik E, Prasad SJ, Ronchese F, Lukashev D, Wong MKK, Huang X, Caldwell S, Liu K, Smith P, Chen J, Jackson EK, Apasov S, Abrams S, Sitkovsky M: A2A adenosine receptor protects tumors from antitumor T cells. Proc Natl Acad Sci USA 2006, 103:13132-13137

Deaglio S, Dwyer KM, Gao W, Friedman D, Usheva A, Erat A, Chen JF, Enjyoji K, Linden J, Oukka M, Kuchroo VK, Strom TB, Robson SC: Adenosine generation catalyzed by CD39 and CD73 expressed on regulatory T cells mediates immune suppression. J Exp Med 2007, 204:1257-1265

作者单位：From the Department of Medicine,* Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts; the Second Department of Surgery, Asahikawa Medical College, Asahikawa, Hokkaido, Japan; and the Department of Medical Biochemistry and Microbiology, Biomedical Center and Children?