Neuropilin-1 promotes human glioma progression through potentiating the activity of the HGF/SF autocrine pathway

Abstract

Neuropilin-1 (NRP1) functions as a coreceptor through interaction with plexin A1 or vascular endothelial growth factor (VEGF) receptor during neuronal development and angiogenesis. NRP1 potentiates the signaling pathways stimulated by semaphorin 3A and VEGF-A in neuronal and endothelial cells, respectively. In this study, we investigate the role of tumor cell-expressed NRP1 in glioma progression. Analyses of human glioma specimens (WHO grade I-IV tumors) revealed a significant correlation of NRP1 expression with glioma progression. In tumor xenografts, overexpression of NRP1 by U87MG gliomas strongly promoted tumor growth and angiogenesis. Overexpression of NRP1 by U87MG cells stimulated cell survival through the enhancement of autocrine hepatocyte growth factor/scatter factor (HGF/SF)/c-Met signaling. NRP1 not only potentiated the activity of endogenous HGF/SF on glioma cell survival but also enhanced HGF/SF-promoted cell proliferation. Inhibition of HGF/SF, c-Met and NRP1 abrogated NRP1-potentiated autocrine HGF/SF stimulation. Furthermore, increased phosphorylation of c-Met correlated with glioma progression in human glioma biopsies in which NRP1 is upregulated and in U87MG NRP1-overexpressing tumors. Together, these data suggest that tumor cell-expressed NRP1 promotes glioma progression through potentiating the activity of the HGF/SF autocrine c-Met signaling pathway, in addition to enhancing angiogenesis, suggesting a novel mechanism of NRP1 in promoting human glioma progression.

Figure 1

NRP1 expression correlates with human glioma progression. (A) IHC of paraffin sections of normal human brain (panel a), P.A. (WHO grade I, panel b), diffuse astrocytoma (D.A., grade II, panel c), anaplastic astrocytoma (A.A., grade III, panel d) and GBM (grade IV, panel e) tissue. Insets in panels a–e are the isotype-matched IgG control staining of identical areas. Arrows indicate neurons (a) or tumor cells (b to e) that are positive for NRP1. Arrowheads indicate endothelial cells in tumor-associated vessels that express NRP1. Results are representative of three independent experiments. Original magnification: ×400. (B) Statistical analyses. A total of 92 individual primary tumor specimens (WHO grade I–IV) and four normal human brain biopsies were analysed.

Figure 2

Overexpression of NRP1 in U87MG glioma cells. IB analyses of HUVEC, NHA/ETR and various glioma cell lines (a) or U87MG and NRP1-no. 1 and NRP1-no. 8 cells (b) using a polyclonal anti-NRP1 antibody (C-18). β-Actin was used as a loading control. Similar results were also obtained using a polyclonal anti-NRP1 antibody (NP1ECD1A). Results are representative of three independent experiments.

Figure 3

Overexpression of NRP1 in U87MG cells promotes tumor growth and angiogenesis. (A) Growth kinetics of U87MG or NRP1 tumors at subcutaneous sites. Tumor volume was estimated (volume = (a² × b)/2, a<b) using a caliper at the indicated times. Data are shown as mean±s.d. (B) Tumorigenicity and angiogenesis of U87MG brain tumors. IHC analyses are shown for U87MG (panels a and b) or NRP1 (c and d) gliomas. Panels a and c are brain sections stained with hematoxylin and eosin (H&E). Panels b and d show CD31 staining for tumor vessels. Arrows in a and c, tumor mass. Arrows in b and d, blood vessels. Five to eight individual tumor samples of each group from each in vivo experiment were analysed. Original magnification: panels a and c, ×12.5; b and d, ×200. (C) and (D) Quantitative analyses of tumorigenesis and angiogenesis in various intracranial tumors. Data are means±s.d. Numbers in parentheses, the difference in fold between U87MG and NRP1 gliomas. (E) and (F) Cell proliferation of various intracranial gliomas. (E) IHC staining of U87MG brain tumors with a monoclonal anti-BrdUrd antibody (red) together with a polyclonal anti-vWF antibody (green). Blue arrows, proliferative nuclei in tumor cells. White arrows, proliferative nuclei of blood vessels. Blue arrowheads, vWF staining of vessels. Three to five serial sections from five to seven individual samples of each tumor type were analysed. Original magnification: ×400. (F) Quantitative analyses of cellular BrdUrd incorporation in U87MG and U87MG/NRP1 tumors. Numbers in parentheses, the difference in fold between U87MG/NRP1 and parental U87MG gliomas. Results in (A–E) are representative of three independent experiments.

Figure 4

NRP1 promotes survival of U87MG cells by enhancing autocrine HGF/SF/c-Met signaling. (A) Overexpression of NRP1 promotes U87MG cell survival. Data of cell survival assays are shown as mean±s.d. (B) Inhibition of HGF/SF, but not FGF-2 or VEGF, suppresses NRP1-promoted U87MG cell survival. Data of cell survival assays are shown as mean±s.d. (C) Inhibition of tumor cell-derived HGF/SF, but not VEGF or FGF-2, suppresses NRP1-potentiated activation of c-Met and Bad. IP and IB analyses of various U87MG cells treated with or without the HGF/SF neutralizing antibody. (D) Serum starvation did not alter NRP1 expression in U87MG, LacZ and NRP1 cells. IB analyses of various cell lysates under the same conditions as in (A). (E) and (F) Inhibition of endogenous c-Met but not Sema 3A attenuates NRP1-potentiated U87MG cell viability. (E) Suppression of endogenous c-Met and Sema 3A by siRNA. IB analyses of c-Met and Sema 3A proteins in various U87MG cells. (F) Cell survival assays of siRNA-transfected U87MG and NRP1 cells. Data are shown as mean±s.d. In (B–D), c-Met, Bad and β-actin were used as loading controls. Results in (A–F) are representative of three independent experiments.

Figure 5

NRP1 expression potentiates HGF/SF stimulation of cell proliferation mediated by c-Met signaling. (A) Expression of NRP1 in U87MG cells potentiates cell proliferation in response to a low dose of HGF/SF stimulation. Data of BrdUrd incorporation in U87MG, LacZ, NRP1-no. 1, and NRP1-no. 8 cells are shown as mean±s.d. (B) IB analyses of NRP1-potentiated HGF/SF stimulation of phosphorylation on c-Met, ERK1/2 and Bad in various U87MG cells. (C) U0126 attenuated NRP1-potentiated HGF/SF stimulation of phosphorylation on ERK1/2 and Bad (panel a, IB analyses) and cell proliferation in various U87MG cells (panel b, BrdUrd incorporation assays; data are shown as mean±s.d.). (D) Knockdown of Bad by siRNA (panel a, IB analyses) attenuates NRP1-potentiated HGF/SF stimulation of cell proliferation in various U87MG cells (panel b, data are shown as mean±s.d.). In (B–D), c-Met, ERK1/2, Bad and β-actin were used as loading controls. Results in (A) to (D) are representative of three independent experiments.

Figure 6

Inhibition of endogenous NRP1 attenuated HGF/SF stimulation of cell proliferation and HGF/SF/c-Met signaling in LNZ-308 glioma cells. (a) Inhibition of endogenous NRP1 by NRP1/siRNA in LNZ-308 cells detected by IB analyses. C, control siRNA. N1, N2 and N3, individual siRNAs for NRP1. NRP1, a pool of all three siRNAs of N1, N2 and N3. β-Actin was used as a loading control. (b) Inhibition of endogenous NRP1 in LNZ-308 cells using siRNA-attenuated rhHGF/SF stimulation of cell proliferation. Data of BrdUrd incorporation of siRNA-transfected LnZ-308 cells are shown as mean±s.d. (c) Inhibition of endogenous NRP1 in LNZ-308 cells using siRNA-suppressed HGF/SF stimulation of c-Met signaling detected by IB analyses. c-Met, ERK1/2 Bad and β-actin were used as loading controls. Results in (a–c) are representative of three independent experiments.

Figure 7

Upregulation of NRP1 correlates with the activation of c-Met in U87MG tumor xenografts and in high-grade human glioma specimens. (a) Overexpression of NRP1 in U87MG xenografts resulted in activation of c-Met in the established tumors. IB analyses of U87MG and NRP1 xenografted tumors. β-Actin was used as a loading control. (b) Upregulation of NRP1 expression correlates with activation of c-Met during glioma progression. IB analyses of 14 frozen primary human glioma specimens. c-Met and β-actin were used as loading controls. Results are representative of three independent experiments.