uPA and uPAR shRNA inhibit angiogenesis via enhanced secretion of SVEGFR1 independent of GM-CSF but dependent on TIMP-1 in endothelial and glioblastoma cells

Abstract

The uPA/uPAR system is known to play a critical role in angiogenesis of glioblastoma. Previously, we have shown that shRNA against uPA and uPAR attenuates angiogenesis by blocking nuclear translocation of angiogenin, inhibition of angiopoietin/Tie2 signaling, and regulating several other pro-angiogenic, angiostatic and anti-angiogenic molecules. Further analysis revealed that GM-CSF, a pleiotropic cytokine, was significantly inhibited in U87MG and 4910 co-cultures with endothelial cells transfected with shRNA against uPA and uPAR. The role of the uPA/uPAR system in this process is not completely understood. Analysis of tumor conditioned medium of U87MG, 4910 and HMECs transfected with shRNA against uPA or uPAR alone or in combination (pU2) revealed inhibition of GM-CSF-enhanced secretion of SVEGFR1 as shown by Western blotting and ELISA. Moreover, phosphorylation of JAK2 and STAT5, the downstream effectors of GM-CSF signaling, was also inhibited in all three cell lines. Phosphorylation at Tyr 166 position of the GM-CSFRβ subunit, the signal activating subunit of the GM-CSF receptor, was inhibited in HMEC, U87MG and 4910 cells. Further analysis revealed that shRNA against uPA and/or uPAR increased secretion of TIMP-1, which is known to enhance SVEGFR1 secretion in endothelial cells. Moreover, addition of purified uPA (with and without GM-CSF) activated JAK2/STAT5 signaling in HMEC. Exogenous addition of SVEGFR1 to pU2 tumor conditioned medium enhanced inhibition of VEGF-induced endothelial capillary tube formation as assessed by an in vitro angiogenesis assay. To determine the significance of these events in vivo, nude mice with pre-established tumors treated with shRNA against uPA and/or uPAR showed decreased levels of GM-CSF and increased levels of SVEGFR1 and TIMP-1 when compared with controls. Enhanced secretion of SVEGFR1 by puPA, puPAR and pU2 in endothelial and GBM cells was mediated indirectly by MMP-7 and augmented by ectodomain shedding of VEGFr1 by tyrosine phosphorylation at the 1213 position. Taken together, these results suggest that the uPA/uPAR system could prove beneficial as an indirect target for inhibition of angiogenesis in glioblastoma.

Figure 1

shRNA against uPA and uPAR inhibits secreted levels of GM‐CSF in cancer cells and endothelial cells. (A–C) 1.5×105 HMEC, U87MG and 4910 cells were transfected as described earlier. Conditioned medium was collected after 72h and assayed for GM‐CSF levels by ELISA as per the manufacturer's instructions. Data represented were the average of three independent experiments. ∗p<0.01; ∗∗p<0.001. (D) HMEC, U87MG and 4910 cells were transfected with shRNA constructs, and conditioned medium was collected. After protein normalization, equal quantity of conditioned medium was loaded and immunoblotted with primary antibody for GM‐CSF and corresponding secondary antibodies. A representative blot of three independent experiments is shown. (E) Quantitative analysis was done by optical densitometry as per standard protocols. Data represented were the average of three independent experiments.

Figure 2

shRNA against uPA and uPAR enhances SVEGFR1 in U87MG, 4910 and endothelial cells. (A–C) 1.5×105 HMEC, U87MG and 4910 were transfected as described earlier. Conditioned medium was collected after 72h and assayed for SVEGFR1 levels by ELISA as per the manufacturer's instructions. Data represented are the average of three independent experiments. ∗p<0.01; ∗∗p<0.001. (D) HMEC, U87MG and 4910 cells were transfected with shRNA constructs and conditioned medium was collected. After protein normalization, equal quantity of conditioned medium was loaded and immunoblotted with primary antibody for SVEGFR1 and corresponding secondary antibodies. (E) Quantitative analysis was done by optical densitometry as per standard protocols. Data represented were the average of three independent experiments.

Figure 3

shRNA against uPA and uPAR down regulates phosphorylation of JAK2/STAT5 in glioblastoma and endothelial cells and phosphorylation of GM‐CSFrβ at Tyr 766 position in HMEC, U87MG and 4910 cells. (A) Western blot analysis of shRNA‐transfected HMEC, U87MG, and 4910 cells was carried out with primary antibodies against pJAK2 (Tyr1007/1008) and pSTAT5 (pTyr 695/699) and with total forms of JAK2 and STAT5. A representative blot of three independent experiments is shown. (B and C) Quantitative analysis was done for pJAK2 and pSTAT5 by optical densitometry as per standard protocols. (D) Western blot analysis of shRNA‐transfected HMEC, U87MG, and 4910 cells was carried out with primary antibody against pTyr766 position of the GM‐CSFrβ unit and appropriate secondary antibody. GM‐CSFRβ antibody was also immunoblotted to demonstrate equal loading of lysates. A representative blot of three independent experiments is shown. (E) Quantitative analysis was done by optical densitometry as per standard protocols.

Figure 4

shRNA against uPA and uPAR enhances secretion of TIMP‐1 from HMEC, U87MG and 4910 cells by Western Blot and ELISA. (A) (Upper panel) 1.5×105 HMEC, U87MG and 4910 cells were transfected as described earlier. Conditioned medium was collected after 72h and TIMP‐1 levels determined by Western blotting. Equal amount of conditioned medium was loaded after normalization of protein content. (B) Graph represents quantification of Western blotting results using NIH Image J software. (C) 1.5×105 HMEC, U87MG and 4910 cells were transfected as described earlier. Conditioned medium was collected after 72h and TIMP‐1 levels determined by ELISA as per the manufacturer's instructions. Data represented are the average of three independent experiments. ∗p<0.01; ∗∗p<0.001.

Figure 5

uPA activates JAK2/STAT5 independent of GM‐CSF in HMEC. (A) Lysates were made from HMEC untreated or treated with pSV, puPA, purified uPA, siGM‐CSF, rhGM‐CSF, siGM‐CSF or uPA and immunoblotted for pJAK2, pSTA5, and total forms of JAK2 and STAT5 and with corresponding secondary antibodies. A representative blot of three independent experiments is shown. (B) Conditioned medium from untreated, control si, and siGM‐CSF treated HMEC was immunoblotted with primary antibody to GM‐CSF and corresponding secondary antibody. A representative blot of three independent experiments is shown. (C) Graph represents quantitative representation of the Western blotting results for three independent experiments, p<0.05. (D) Lysates were made from HMEC left untreated or treated with purified uPA, with rh GM‐CSF or both and immunoblotted for phospho and total forms of JAK2 and STAT5. A representative blot of three independent experiments is shown. (E) Graph represents quantitative expression of the Western blotting, p<0.05.

Figure 6

Enhanced secretion of SVEGFR1 by uPA/uPAR downregulation in GBM and HMEC cells is mediated by MMP‐7 and augmented by tyrosine phosphorylation of VEGFr1 at 1213 position. (A) HMEC, U87MG and 4910 cells were left untreated or treated with different concentrations of TIMP‐1 and lysates were made at 24 and 48h and immunoblotted for MMP‐7 with appropriate primary and secondary antibodies. Representative blots from each cell line are shown. The blots were probed for GAPDH to verify equal loading of lysates. (B–D) Graphs represent quantitative representation of the Western blotting results in HMEC, 4910 and U87MG, respectively. (E) Conditioned medium from above‐mentioned conditions were immunoblotted for SVEGFR1. A representative blot from three independent experiments is shown. (F–H) Graphs represent quantitative representation of the Western blotting results in HMEC, U87MG and 4910 cells, respectively. (I) Lysates made from HMEC, U87MG and 4910 cells, which were left untreated or treated with pSV, puPA and pU2 were immunoblotted for pVEGFr1 pTyr 1213 position primary antibody and with appropriate secondary antibodies. A representative blot from three independent experiments is shown. Equal loading of lysates is shown by reprobing the blots and blotting for GAPDH. (J) Graphical representation of quantitative estimation of the Western blotting results. (K) Lysates made from HMEC, U87MG and 4910 cells which were left untreated or treated with pSV, puPA and pU2 were immunoblotted for MMP‐7 primary antibody and with appropriate secondary antibodies. A representative blot from three independent experiments is shown. Equal loading of lysates is shown by reprobing the blots and blotting for GAPDH. (L) Graphical representation of quantitative estimation of the Western blotting results is shown.

Figure 7

Exogenous addition of SVEGFR1 enhances inhibition of capillary tube formation as shown by in vitro angiogenesis assay. (A) HMECs treated with conditioned media from U87MG cells, which were left untreated or treated with pU2 and/or supplemented with sVEGFR1 and VEGF alone were cultured with the collected conditioned medium in 48‐well plates for 24h. After the incubation period, the medium was removed, and the cells stained with Hema‐3 stain and examined under a microscope. (B) Quantification of angiogenesis in endothelial cells that were left untreated, treated with pU2 conditioned medium, or treated with pU2 conditioned medium supplemented with SVEGFR1. Values are mean±S.D. from three different experiments. (C–E) shRNA against uPA/uPAR inhibits secretion of GM‐CSF and enhances SVEGFR1 and TIMP‐1 secretion in pre‐established intracranial tumors in nude mice. U87MG cells in suspension (2×106 in 10μL serum‐free medium) were injected intracranially in nude mice. One week later, the mice were injected with either pSV or shRNA‐expressing vectors (puPA, puPAR and pU2) using an Alzet mini‐osmotic pump (constructs diluted to 1.5μg/mL in PBS and injected at 0.25μg/h, with 5 mice in each group). After a five‐week follow‐up period, mice were sacrificed, and blood was collected and serum was separated to assay for the levels of mSVEGFR1 (C) and msGM‐CSF (D) and mTIMP‐1 (E) by ELISA according to the manufacturer's instructions.

Figure 8

Schematic diagram showing the mechanism by which uPA/uPAR shRNA inhibits angiogenesis in glioblastoma cells and xenografts. Panel (A) Binding of uPA to uPAR co‐ordinates with GM‐CSF receptor to activate JAK2/STAT5 mediated angiogenic signaling pathway. VEGF/VEGFR signaling is also well known to initiate angiogenesis. Panel (B) Knockdown of uPA and uPAR inhibits the angiogenesis signaling induced by both GM‐CSF and VEGF. uPA/uPAR down regulation decreased GM‐CSF secretion resulting in inhibition of JAK2/STAT5 activation. Further knock down of uPA/uPAR increased sVEGFR1 secretion which binds to free bound VEGF and acts as a scavenger inhibiting VEGF‐mediated signaling. sVEGFR1 secretion was mediated by TIMP‐1 and ectodomain shedding of VEGFR1 regulated by phosphorylation at the Tyr1213 of the VEGFR1 receptor.