Vasculostatin inhibits intracranial glioma growth and negatively regulates in vivo angiogenesis through a CD36-dependent mechanism

Abstract

Angiogenesis is a critical physiologic process that is appropriated during tumorigenesis. Little is known about how this process is specifically regulated in the brain. Brain angiogenesis inhibitor-1 (BAI1) is a brain-predominant seven-transmembrane protein that contains five antiangiogenic thrombospondin type-1 repeats (TSR). We recently showed that BAI1 is cleaved at a conserved proteolytic cleavage site releasing a soluble, 120 kDa antiangiogenic factor called vasculostatin (Vstat120). Vstat120 has been shown to inhibit in vitro angiogenesis and suppress subcutaneous tumor growth. Here, we examine its effect on the intracranial growth of malignant gliomas and further study its antitumor mechanism. First, we show that expression of Vstat120 strongly suppresses the intracranial growth of malignant gliomas, even in the presence of the strong proangiogenic stimulus mediated by the oncoprotein epidermal growth factor receptor variant III (EGFRvIII). This tumor-suppressive effect is accompanied by a decrease in tumor vascular density, suggesting a potent antiangiogenic effect in the brain. Second, and consistent with this interpretation, we find that treatment with Vstat120 reduces the migration of cultured microvascular endothelial cells in vitro and inhibits corneal angiogenesis in vivo. Third, we show that these antivascular effects critically depend on the presence of the cell surface receptor CD36 on endothelial cells in vitro and in vivo, supporting the role of Vstat120 TSRs in mediating these effects. These results advance the understanding of brain-specific angiogenic regulation, and suggest that Vstat120 has therapeutic potential in the treatment of brain tumors and other intracerebral vasculopathies.

Figure 1. Expression of Vstat120 enhances survival of rats implanted with U87MGD glioma cells in the brain

A. Western blot analysis of cell lysates from U87MG parental cells, and U87MG-derived clones stably transfected with Vstat120 cDNA (U14 and U18). Note the expression of Vstat120 in the whole cell extract (WCE) and conditioned medium (CM) of the stably transfected clones. B. The in vitro proliferation rates of U87MG cells, and clones U14 and U18 was determined by the crystal violet assay. Note expression of Vstat120 did not alter the proliferation rate of the clones versus U87MG cells. C. CM from control U87MGD or Vstat120 expressing U14 cells were tested for their ability to inhibit the migration of HDMECs in a Transwell migration assay. The number of cells that migrated to the bottom of the chamber after 8 hrs was quantified as described in materials and methods. Random migration in response to medium alone was subtracted from the values. Expression levels of Vstat120 and thrombospondin-1 (TSP1) in the CM were assessed by Western blot. D. Intracranial tumorigenicity assay for U87MG and Vstat120 expressing clones, U14 and U18. 1 × 10⁶ cells were implanted stereotactically in the brain of athymic nude rats. Survival curves of rats implanted with cells expressing Vstat120 showed a significant improvement in their survival compared to the control parental U87MG cells (p<0.05).

Figure 2. Vstat120 expression suppresses subcutaneous and intracranial tumor growth of U87ΔEGFR cells despite the pro-angiogenic stimulus provided by EGFRvIII

A. Characterization of U87ΔEGFR and Vstat 120 expressing clones Δ19 and Δ22. Upper Panel shows western blot analysis of whole cell extract (WCE) and conditioned medium (CM) from U87ΔEGFR cells (U87Δ) and derived clones Δ19 and Δ22, which stably express Vstat120. Lower panel shows in vitro proliferation rates of U87ΔEGFR cells, and Vstat120 expressing clones Δ19 and Δ22, as measured using the crystal violet assay. Expression of Vstat120 does not alter the in vitro proliferation rates of these cells. B. Subcutaneous growth of U87ΔEGFR and Vstat120-expressing clones in nu/nu mice. U87ΔEGFR and derived clones stably expressing Vstat120 (Δ19 and Δ22) were injected subcutaneously into mice (n=6) and the tumor volume for the indicated clones was plotted as a function of time. Note the strongly decreased tumor growth of clones expressing Vstat120. C. Relative growth of U87MGF, U87ΔEGFR and Vstat120-expressing clones in rat brains. Upper panels show representative images of the MRI scans of individual rat brains 14 days after intracranial implantation of 10⁶ tumor cells. The presence of glioma is detected through the bright areas (white arrows) of contrast enhancement from the MRI contrast agent (Gd-DTPA). Note the small tumor in U87MGF cells, large tumor in U87ΔEGFR cells and barely detectable minimal tumors in clones Δ22 and Δ19. The lower panel shows corresponding histopathological brain sections stained with H&E. Tumor growth is visible as a darkly stained area (black arrow). D. Survival curves of rats implanted with U87MGF, U87ΔEGFR and Vstat120 expressing clones, Δ 19 and Δ22. 1 × 10⁶ cells were implanted stereotactically in the brain of athymic nu/nu rats. Rats implanted with U87ΔEGFR cells had the shortest survival time due to the very angiogenic and aggressive nature of these tumors. Vstat120 expressing clones Δ19 and Δ22, showed a significant improvement in their survival compared to the U87ΔEGFR and control parental U87MGF cells (p<0.05).

Figure 3. In vitro and In vivo Antiangiogenic effect of Vstat120 produced by glioma cells

A. Representative pictures of the immunohistochemistry for von Willebrand factor in tumor sections derived from U87ΔEGFR or Vstat120 expressing clone (Δ19) are shown. Brown staining indicates endothelial cells lining capillaries (arrows). B. Vessel densities in U87ΔEGFR and Vstat120 expressing clones (Δ19 and Δ22) were determined. Vstat120-expressing tumors showed significantly lower vessel density than parental tumors. Vessel densities are expressed as mean +/− SEM. * p<0.005

Figure 4. Vstat120 inhibits endothelial cell migration in a CD36-dependent fashion

A. Western blot analysis for expression of CD36 in U87ΔEGFR, HDMEC, and HUVEC cells, respectively. B: Production of secreted Vstat120 by transient transfection in 293 cells (Left). The cells (80% confluent) were left untreated (lane 1) or were transfected with either control pcDNA3.1lacZ vector (lane 2) or Vstat120 expression vector pcDNA3.1Vstat120-myc/his (lane 3). Vstat120 produced by cells transfected with full length BAI1 expression vector was utilized as a size control (lane 4). CM was collected in serum-free media for 48hrs and used in experiments B–D below. WCE, whole cell extract. Transwell migration assays (Right) examining the migration of HUVEC and HDMEC in the presence of CM from control or Vstat120 containing CM from 293 cells. The number of cells that migrated to the bottom of the chamber after 8 hrs was quantified as described in materials and methods. Note that Vstat120 containing CM reduces the migration of HDMEC, but not HUVEC. C. Scratch-wound migration assays. Confluent HDMECs were wounded, treated with CM prepared as in B. and the endothelial cells allowed to migrate for 8 hrs, then fixed and stained with crystal violet. Shown are representative pictures of migrated cells (Top panel). The black bars indicate initial wound width in micrometers. Distance of migration, percentage of wound closure, and speed of migration was quantified (Middle panel). The experiment was repeated twice with similar results. Data are expressed as mean +/− SEM; n=6 for each condition; * p<0.01 compared to Vstat120. CD36 function-blocking antibody prevents Vstat120 anti-angiogenic function (Bottom panel). HDMECs were wounded, then either left untreated or treated with anti-CD36 function-blocking antibody at 10 µg/mL for 30 min. The cells were next treated with CM (as above) for 30 min, followed by treatment with 10% serum to induce cell migration. Final wound width was measured after 8 h and the distance migrated was calculated. D. Transwell assay examining the migration of HBMEC in the presence of CM from U87MGD (Ctrl) and Vstat 120 expressing U14 cells (Vstat). Data is mean ± SEM. n=3 for each condition. * p<0.05 and n.s. not significant by Student’s T test.

Figure 5. Vstat120 binds to the purified CLESH domain of CD36

A: Schematic of CD36 structure with the CLESH domain. The two GST-CD36 constructs used (amino acids 5–143 and 67–157), both of which contain the CLESH domain (amino acids 93–120) are shown. B. Top: Coomassie stained gel showing purified GST, and GST tagged recombinant proteins encoding for amino acids 67–157 and 5–143 of CD36. Proteins purified to near homogeneity and migrated at their predicted molecular weight. Bottom: Western blot analysis of each fusion protein probed with anti-GST monoclonal antibody (MAB3317 Chemicon International). C. GST pull-down assay. GST alone or the two recombinant GST-CD36 peptides were bound to glutathione sepharose beads and CM from LN229 glioma cells stably expressing Vstat120 (+ lanes) or control cells (− lanes) was tested for protein interaction. A separate pull-down assay with CM from TSP1 expressing cells (LN229 clone C9) was used as a positive control. The bound proteins were eluted and analyzed for Vstat120 and TSP1 expression by western blot. Both the GST tagged CD36 containing recombinant peptides could pull down Vstat120 and TSP1 but not the purified GST. Positive control lanes are TCA precipitations of CM (collected serum-free after 96hrs) that express either Vstat120 or TSP1.

Figure 6. Vstat120 inhibits corneal angiogenesis in a CD36-dependent manner

Representative photographs (A) of mice cornea at 5 days post implantation of pellets containing 25ng of bFGF and CM of 293 cells (50ng total CM protein) transfected with Vstat120 or vector control (Ctrl). Photographs show FITC-dextran labeled capillaries (arrows) progressing toward the pellet, previously inserted in the mouse cornea. Angiogenic response was quantified by measuring the neovascular area in the cornea. Relative to control (A, upper left), CM collected from Vstat120-expressing cells (A, upper right) impairs capillary formation by 40%. This effect is totally negated in CD36 knockout mice (A, bottom pictures). The mean of neovascularized areas (mm²) in the corneas were quantified (B) as described in Materials and Methods. Each condition was carried out in at least 9 corneas. The values are expressed in means±SE. Statistical analysis was performed using the ANOVA test, *p<0.05.