EGFL6 promotes endothelial cell migration and angiogenesis through the activation of extracellular signal-regulated kinase

Abstract

Angiogenesis is required for bone development, growth, and repair. It is influenced by the local bone environment that involves cross-talks between endothelial cells and adjacent bone cells. However, data regarding factors that directly contribute to angiogenesis by bone cells remain poorly understood. Here, we report that EGFL6, a member of the epidermal growth factor (EGF) repeat superfamily proteins, induces angiogenesis by a paracrine mechanism in which EGFL6 is expressed in osteoblastic-like cells but promotes migration and angiogenesis of endothelial cells. Co-immunoprecipitation assays revealed that EGFL6 is secreted in culture medium as a homodimer protein. Using scratch wound healing and transwell assays, we found that conditioned medium containing EGFL6 potentiates SVEC (a simian virus 40-transformed mouse microvascular endothelial cell line) endothelial cell migration. In addition, EGFL6 promotes the endothelial cell tube-like structure formation in Matrigel assays and angiogenesis in a chick embryo chorioallantoic membrane. Furthermore, we show that EGFL6 recombinant protein induces phosphorylation of ERK in SVEC endothelial cells. Inhibition of ERK impaired EGFL6-induced ERK activation and endothelial cell migration. Together, these results demonstrate, for the first time, that osteoblastic-like cells express EGFL6 that is capable of promoting endothelial cell migration and angiogenesis via ERK activation. Thus, the EGLF6 mediates a paracrine mechanism of cross-talk between vascular endothelial cells and osteoblasts and might offer an important new target for the potential treatment of bone diseases, including osteonecrosis, osteoporosis, and fracture healing.

FIGURE 1.

Differentially expression of EGF-like family in osteoclasts and osteoblasts. A, schematic representations of domain structures of EGF-like family members. EGFL3, EGFL6, EGFL7, and EGFL8 are predicted as secreted protein, whereas EGFL2, EGFL5, and EGFL9 are predicted as membrane-bound protein. EGF-like family members vary in amino acid length and contain differing number of EGF domains. B, RT-PCR amplification of EGF-like family members in differentiating osteoclasts and calvaria bone. RAW_264.7 cells derived osteoclasts were cultured in the presence of RANKL for up to 7 days. mRNA was extracted and subjected to RT-PCR using primers specific for EGF-like family members and 36B4. C, gene expressions of EGF-like family members up-regulated during osteoblast differentiation. RNA was extracted from primary osteoblast at 0 (mock), 7, 14, and 21 days after stimulation with β-glycerophosphate and ascorbic acid. RNA was then subjected to RT-PCR amplification using primers specific for EGF-like family members, osteoblast-specific genes, and 36B4. D, -fold change of EGFL6 gene expression during osteoblast differentiation, with its expression peaking at day 14. E, tissue expression profile of EGFL6 determined by RT-PCR analysis.

FIGURE 2.

EGFL6 exists as a secreted homomeric complex. A, two expression constructs encoding mouse full-length EGFL6, pcDNA3.1-EGFL6-HA, and pcDNA3.1-EGFL6-c-myc were generated. pGEX-3X-GST-EGFL6 encoding EGFL6 (264–393) was generated to produce GST-EGFL6 fusion protein for antibody production. B, COS-7 cells transfected with empty vector or expression vector encoding HA-tagged EGFL6. Medium and cell lysate were collected and subjected to Western blotting using an anti-HA antibody. EGFL6 was detected in medium and cell lysates. C, COS-7 cells were singly or co-transfected with the pcDNA3.1-EGFL6-HA and pcDNA3.1-EGFL6-c-myc constructs. Two days after transfection, supernatant was collected and subjected to immunoprecipitation (IP) using an anti-c-myc antibody, followed by Western blotting (WB) using an anti-HA antibody. Immunoprecipitation results indicate that EGFL6-c-myc and EGFL6-HA proteins form a homomeric complex. D, purified GST-EGFL6 proteins shown by Coomassie Blue staining. GST-EGFL6 proteins were detected by Western blotting using an anti-GST antibody. GST alone was use as a control. E, COS-7 cells transfected with HA-tagged EGFL6 vector or an empty vector. Culture supernatants were harvested after 24 h. EGFL6 proteins in culture supernatants were detected using immunoblotting with anti-HA antibody and anti-EGFL6 antibody. F, supernatant and cell lysates harvested during primary osteoblast differentiation and subjected to Western blot analysis using the anti-EGFL6 antibody. The protein levels of EGFL6 were up-regulated during osteoblast differentiation. G, proteins extracted from limb buds (E14.5) and long bones (E18.5, week 1, and week 16) of mice and subjected to Western blot analysis. EGFL6 proteins were expressed during bone development. Cell lysate from day 21 differentiated primary osteoblast was used as a positive control. Black arrows indicate the predicted size of EGFL6 proteins.

FIGURE 3.

EGFL6 promotes endothelial cell migration. A, scratch wound healing assays performed in SVEC cells treated with conditioned medium containing EGFL6 or vehicle control for 24 h. Representative microscopic views at 0 h, 12 h, and 24 h are shown. B, quantitative analysis of cell migration area after 12 h. C, transwell migration assay. SVEC cells were seeded in the upper chamber, conditioned medium containing EGFL6 or vehicle control was placed at the bottom side. After 24 h, migrated cells were quantified by fluorescence dye. PBS and bFGF were used as a negative and positive control, respectively. D, RT-PCR amplification of EGFL6 and EGFL7 in calvaria bone (OB), RAW_264.7 cell-derived osteoclasts (OC), and SVEC endothelial cells (EC). E, RT-PCR amplification of angiogenic genes in endothelial cells treated with conditioned medium containing EGFL6 or vehicle control for 24 h. **, p < 0.01. Error bars, S.D.

FIGURE 4.

EGFL6 promotes tube formation on Matrigel and induces angiogenesis in a CAM assay. A, SVEC cells previously starved overnight seeded on Geltrex^TM matrix. Cells were cultured in the conditioned medium containing EGFL6 or vehicle control for 24 h. PBS and bFGF were used as a negative and positive control, respectively. Representative photographs of each sample are shown. B, quantitative analysis of tube length. C, photographs of the chick CAMs after 48 h containing the gelatin sponge carrying EGFL6 proteins and vehicle control. Representative images are shown. D, average scores of the angiogenesis response intensity. **, p < 0.01. Error bars, S.D.

FIGURE 5.

EGFL6 promotes endothelial cell migration through activation of ERK. A, cellular phosphorylation levels of ERK1/2 determined by Western blot analysis after SVEC cells were stimulated with conditioned medium containing EGFL6 or vehicle control for the indicated times. B, EGFL6-induced ERK phosphorylation inhibited by MEK inhibitor, U0126. C, effect of U0126 on SVEC cell migration. A scratch wound healing assay was performed with the conditioned medium containing EGFL6 or vehicle control in the presence or absence of U0126 for 24 h. D, quantitative analysis of the scratch wound healing assay with U0126 after 12 h. E, tube formation assay performed with the conditioned medium containing EGFL6 or vehicle control in the presence or absence of U0126 for 24 h. PBS and bFGF were used as a negative and positive control, respectively. *, p < 0.05; **, p < 0.01. Error bars, S.D.

FIGURE 6.

EGFL6-induced endothelial cell migration was blocked by RGD peptides. A, scratch wound healing assays performed with the conditioned medium containing EGFL6 or vehicle control in the presence or absence of RGD peptides for 12 h. B, quantitative analysis of cell migration area after 12 h. **, p < 0.01. Error bars, S.D.

FIGURE 7.

EGFL6 does not induce proliferation and mineralization of osteoblastic cells. A, freshly isolated mouse calvaria cells were cultured in the presence of 2% FBS with conditioned medium containing EGFL6 or vehicle control. Cell viability was measured at 24 h, 48 h, and 72 h using the CellTiter 96® aqueous nonradioactive cell proliferation assay. 2% FBS and 10% FBS were used as a negative and positive control, respectively. B, KusaO cells were cultured in differentiation medium with conditioned medium containing EGFL6 or vehicle control for 14 days and stained with alizarin red. KusaO cells cultured without differentiation medium were used as a negative control. BMP-2 was used as a positive control. (C) The mineralization areas were quantified and presented in percentages comparing to control. *, p < 0.05; **, p < 0.01. Error bars, S.D.