EGFL7 is expressed in bone microenvironment and promotes angiogenesis via ERK,

STAT3 and integrin signaling cascades

Shek Man Chim1, Vincent Kuek

1, Siu To Chow

1, Bay Sie Lim

1, Jennifer Tickner

1, Jinmin Zhao

2,

Rosa Chung3, Yu-Wen Su

3, Ge Zhang

4, Wendy Erber

1, Cory J Xian

3, Vicki Rosen

5*, Jiake Xu

1,2*

1School of Pathology and Laboratory Medicine, University of Western Australia, Nedlands WA

6009, Australia;

2Research Centre for Regenerative Medicine, Department of Orthopaedic Surgery, The First

Affiliated Hospital of Guangxi Medical University, Guangxi, 530021, China;

3Sansom Institute for Health Research, School of Pharmacy and Medical Sciences, University of

South Australia, Adelaide, SA 5001, Australia;

4Institute for Advancing Translational Medicine in Bone & Joint Diseases, School of Chinese

Medicine, Hong Kong Baptist University, Hong Kong;

5Developmental Biology, Harvard School of Dental Medicine, Boston, MA 02115, USA.

Running title: EGFL7 regulates endothelial cell activities

*Co-corresponding Author:

Prof. Jiake Xu, School of Pathology and Laboratory Medicine , The University of Western

Australia M508, 35 Stirling Hwy, Crawley, WA, 6009, Australia, Tel: +61 8 9346 2739; Fax: +61

8 9346 2891; E-mail: jiake.xu@uwa.edu.au

Prof. Vicki Rosen, Developmental Biology, Harvard School of Dental Medicine, Boston, MA

02115, USA, Tel: +617 432 5910; Fax: +617 432 3246 ; E-mail: vicki_rosen@hsdm.harvard.edu

Keywords: EGFL7, Bone, Osteoblast, Osteoclast, Endothelial cell, Angiogenesis

This article contains 8 figures.

Abstract

Angiogenesis plays a pivotal role in bone formation, remodeling and fracture healing. The

regulation of angiogenesis in the bone microenvironment is highly complex and orchestrated by

intercellular communication between bone cells and endothelial cells. Here, we report that EGF-

like domain 7 (EGFL7), a member of the epidermal growth factor (EGF) repeat protein

superfamily is expressed in both the osteoclast and osteoblast lineages, and promotes endothelial

cell activities. Addition of exogenous recombinant EGFL7 potentiates SVEC (simian virus 40-

transformed mouse microvascular endothelial cell line) cell migration and tube-like structure

formation in vitro. Moreover, recombinant EGFL7 promotes angiogenesis featuring web-like

structures in ex vivo fetal mouse metatarsal angiogenesis assay. We show that recombinant

EGFL7 induces phosphorylation of extracellular signal-regulated kinase 1/2 (ERK1/2), signal

transducer and activator of transcription 3 (STAT3), and focal adhesion kinase (FAK) in SVEC

cells. Inhibition of ERK1/2 and STAT3 signaling impairs EGFL7-induced endothelial cell

migration, and angiogenesis in fetal mouse metatarsal explants. Bioinformatic analyses indicate

that EGFL7 contains a conserved RGD/QGD motif and EGFL7-induced endothelial cell

migration is significantly reduced in the presence of RGD peptides. Moreover, EGFL7 gene

expression is significantly upregulated during growth plate injury repair. Together, these results

demonstrate that EGFL7 expressed by bone cells regulates endothelial cell activities through

integrin-mediated signaling. This study highlights the important role that EGFL7 expressed in

bone microenvironment plays in the regulation of angiogenesis in bone.

Introduction

The vasculature in bone is pivotal for skeletal development during the embryonic stage, postnatal

growth and bone remodeling. During endochondral bone formation, the vascular supply provides

critical input allowing for the establishment of ossification centers within mesenchymal

condensation. Vascular invasion allows recruitment of osteoblast precursors and deposition of

bone within the cartilaginous diaphysis to form the primary ossification center. In postnatal life,

secondary ossification centers are established within cartilaginous epiphyses by a similar

mechanism. Once formed, bone is continuously undergoing remodeling to maintain mechanical

integrity (Burkus et al., 1993; Carlevaro et al., 2000; Gerber and Ferrara, 2000). Communication

between vascular endothelial cells, bone building osteoblasts and bone-resorbing osteoclasts is

central to remodeling. This could be facilitated by multiple regulatory proteins that incorporate

autocrine and/or paracrine modes of signaling to mediate the recruitment, survival, proliferation

and differentiation of bone cells and vascular endothelial cells (Brandi and Collin-Osdoby, 2006;

Chim et al., 2013; Eriksen et al., 2007). It has been demonstrated that osteoclasts, osteoblasts and

osteocytes (terminally differentiated osteoblasts embedded within bone matrix) produce

angiogenic factors such as VEGF, bFGF and EGFL6 (Chim et al., 2011; Thi et al., 2010; Utting

et al., 2010). In contrast, endothelial cells produce regulatory factors that influence bone cells

such as macrophage colony-stimulating factor (M-CSF), receptor activator of nuclear factor

kappa-B ligand (RANKL) and chemokines (Brandi and Collin-Osdoby, 2006; Collin-Osdoby et

al., 2001).

Bone remodeling takes place in a specialized vascularised structure called the bone remodeling

compartment (BRC), which is closely linked with capillaries (Andersen et al., 2009; Hauge et al.,

2001). The canopy of the BRC provides a closed microenvironment for bone remodeling by

separating the BRC from open bone marrow spaces. This allows multiple regulatory factors to

reach a critical concentration and avoids them dispersing within the marrow (Eriksen et al.,

2007). Both EGF and EGF-like proteins have been implicated in angiogenesis (Kim et al., 2003;

Leker et al., 2009; Mehta and Besner, 2007; Penta et al., 1999; Schmidt et al., 2007). Recently,

we identified several EGF-like family members, EGFL2, EGFL3, EGFL5, EGFL6, EGFL7,

EGFL8 and EGFL9, that are differentially expressed in osteoclasts and osteoblasts. In addition,

we demonstrated that EGFL6 is secreted by osteoblasts and mediates angiogenesis via a

paracrine mode of action (Chim et al., 2011). However, the specific roles of the other individual

EGF-like family members in mediating angiogenesis in the bone microenvironment remains to

be elucidated.

EGF-like family members mediate a wide range of biological activities such as migration,

adhesion, proliferation and differentiation (Singh and Harris, 2005). The human and mouse

EGFL7 genes map to chromosomes 9 and 2 respectively. EGFL7 is a secreted protein that

contains an N-terminal cysteine-rich domain (EMI domain), followed by two EGF repeat

domains. EGFL7 gene expression has been previously reported in endothelial progenitors,

endothelial cells and neurons (Fitch et al., 2004; Parker et al., 2004; Schmidt et al., 2009).

EGFL7 regulates tubulogenesis during embryogenesis, migration of endothelial cells, and spatial

organization in angiogenic sprouts (Campagnolo et al., 2005; De Maziere et al., 2008; Parker et

al., 2004; Schmidt et al., 2007). It has been suggested that EGFL7 acts as an antagonist to Notch

and regulates ligand-mediated Notch receptor signaling (Nichol et al., 2010; Schmidt et al.,

2009). EGFL7 can also regulate human hepatocellular carcinoma cell motility through EGFR-

mediated FAK signaling (Wu et al., 2009). However, the role of EGFL7 in bone

microenvironment is hitherto unknown.

In this study, we show that EGFL7 is expressed in osteoclasts, osteoblasts and bone tissues.

Moreover, exogenous recombinant EGFL7 protein is capable of promoting endothelial cell

migration, angiogenesis and integrin-mediated signaling including the phosphorylation of

ERK1/2, STAT3 and FAK. EGFL7-induced endothelial cell migration was significantly blocked

in the presence of RGD peptides. Furthermore, we observed up-regulation of EGFL7 gene

expression during the angiogenic and repair phase in a rat model of injured growth plate bony

repair. We propose that EGFL7 mediates endothelial cell activities in the bone local environment.

Materials and Methods

Cells and Cell Culture

Primary mouse bone marrow macrophages (BMM) were isolated from C57BL/6J mice as

previously described (Wu et al., 2014). Briefly, bone marrow cells were collected from femora

and tibia of mice by flushing the bone with α-modified Eagle’s medium (α-MEM; Gibco),

supplemented with 10% (v/v) fetal bovine serum (FBS; Gibco), 1X penicillin-streptomycin

(Gibco)(complete α-MEM), followed by filtration through a 100µm mesh. BMM were

maintained in complete α-MEM in a humidified incubator at 37°C and 5% CO2. BMM were

cultured in the presence of 50ng/ml RANKL and 10ng/ml MCSF (R&D Systems) for 5 days to

form mature osteoclasts. RANKL recombinant proteins were expressed and purified as

previously described (Xu et al., 2000). The cells were fixed and stained for Tartrate-resistant acid

phosphate activity (TRACP) to identify osteoclasts using TRACP staining kit (Sigma Aldrich).

Primary mouse osteoblasts were prepared from the calvariae of neonatal C57BL/6J mice by

enzymatic digestion according to published protocols (Bakker and Klein-Nulend, 2012). For

osteoblast differentiation, primary calvarial osteoblastic cells were seeded onto 12-well plates at

a cell density of 4x104 cells/well and grown in Dulbecco’s Modified Eagle Medium (DMEM)

supplemented with 10% FBS, 1X penicillin-streptomycin (Gibco)(complete DMEM) until

reaching confluence. Cell medium was then replaced with osteogenic differentiation medium,

complete DMEM containing 50μg/ml ascorbic acid (Sigma Aldrich), 10mM β-glycerophosphate

(Sigma Aldrich), and 10nM dexamethasone (Sigma Aldrich). The medium was replaced every

two to three days with fresh differentiation medium up to 21 days. Osteoblast mineralization was

confirmed by alizarin red staining as previously described (Guo et al., 2012). COS-7 and SVEC

cells were maintained in complete DMEM.

Semi-quantitative Reverse Transcription (RT)-PCR

Total cellular RNA was isolated from cultured cells using RNeasy Mini Kit (Qiagen) in

accordance with the manufacturer’s protocol. For RT-PCR, single-stranded cDNA was reverse

transcribed from 2μg total RNA using reverse transcriptase with an oligo-dT primer. All PCR

was carried out using 1μl of each cDNA using cycling parameters 94°C, 40secs; 55°C, 40secs;

72°C, 40secs for 30 cycles with primers designed against the following mouse sequences:

EGFL7 (forward:5’-AGTGGGGATCCCCACTTACAATGCAGAC-3’; reverse:5’-

TCTTCTAGAAGATCTTTTTTGCAGGAGCAG-3’), CTR (forward:5’-CGGACTTTGACA

CAGCAGAA-3’; reverse:5’-CAGCAATCGACAAGGAGTGA-3’), TRAP (forward:5’-

CAGCAGCCAAGGAGGACTAC-3’; reverse:5’- ACATAGCCCACACCGTTCTC-3’), DC-

STAMP (forward:5’- CTTGCAACCTAAGGGCAAAG-3’; reverse:5’-

TCAACAGCTCTGTCGTGACC-3’), EGFL6 (forward:5’-AAGCTTGGATCCGAAT

TCAGTATGCAGCCGCCCTGG-3’; reverse:5’-

CTCGAGTCTAGAAGATCTACCTTCTACAGATAAAAAGT-3’), ALP (forward:5’-

AACTGCTGGCCCTTGACCCCT-3’; reverse:5’- TCCTGCCTCCTTCCACCAGCA-3’), OCN

(forward:5’-GCGCTCTGTCTCT CGTGACCT-3’; reverse:5’-ATAGATGCGTTT GTAGGCGG-

3’), 18s (forward:5’-ACCATAAACGAT GCCGACT-3’; reverse:5’-TGTCAATCCTGTC

CGTGTC-3’). PCR samples were analyzed by DNA agarose gel electrophoresis.

Western Blotting

Total cellular proteins were extracted from cultured cells using RIPA lysis buffer as previous

described (Chim et al., 2011). Briefly, lysates were cleared by centrifugation at 16,000g for 20

mins at 4°C and post-nuclear supernatants were collected. For immunoblotting, equivalent

amounts of extracted proteins diluted in SDS-sampling buffer with 5% β-mercaptoethanol and

incubated at 95 ºC for 5 mins. Samples were resolved by SDS-PAGE gels and then electroblotted

onto nitrocellulose membranes. Following transfer, membranes were blocked with 5% skim milk

for 1hr and then probed with primary antibodies for 2hrs. Primary antibodies used were mouse

monoclonal anti-c-myc (Sigma Aldrich), goat polyclonal anti-EGFL7 (R&D Systems), mouse

monoclonal anti-p-ERK, rabbit polyclonal anti-p-ERK5 (Santa Cruz Biotechnology), rabbit

polyclonal anti-ERK (Promega), rabbit monoclonal anti-p-STAT1, rabbit polyclonal anti-STAT1,

rabbit polyclonal anti-p-STAT3, rabbit monoclonal anti-STAT3, rabbit polyclonal anti-p-FAK

and rabbit polyclonal anti-pEGFR (Cell Signaling Technology), mouse monoclonal anti-β-actin

(JLA20; DSHB University of Iowa). All antibodies were used at the concentrations

recommended by the suppliers. Membranes were washed and then incubated with HRP-

conjugated secondary antibodies (Sigma Aldrich) for 1 hr. Immunoreactivity was visualized

using the Western Lighting Ultra (PerkinElmer) and FujiFilm LAS-4000 Gel Documentation

System (FujiFilm). Signal intensities were quantified by NIH imageJ software. To examine

EGFL7 protein expression during bone development, proteins were extracted from long bones at

different time points of mouse development (E18.5, Week 1, Week 7 and Week 16). The

dissected samples were snap frozen in liquid nitrogen and ground using a mortar and pestle.

Proteins were then extracted using RIPA lysis buffer as above.

Preparation of Conditioned Medium Containing EGFL7

Conditioned medium containing EGFL7 was produced according to protocol described

previously (Chim et al., 2011). Briefly, COS-7 cells were cultured overnight in a 6-well plate in

complete DMEM at a density of 4x105 cells/well. The following day, the culture medium was

transfected with EGFL7 expression vector pcDNA3.1-EGFL7-c-myc/His or empty pcDNA3.1

vector using lipofectamine 2000 (Invitrogen). After 6 hrs, cells were washed twice and incubated

with Opti-MEM reduced serum medium (Gibco). EGFL7-enriched supernatant was harvested at

24 hrs, centrifuged at 2000 rpm for 10 mins to remove cell debris and snap frozen in aliquots at -

80ºC. The presence of EGFL7 in the supernatant was examined by Western blotting using anti-c-

myc and anti-EGFL7 antibodies. The concentration of EGFL7 in conditioned medium was

estimated to be ~211ng/ml (supplementary Fig. S1). Recombinant human EGFL7 was purchased

from OriGene Technologies.

Scratch-Wound Healing Assay

The migration assay was performed as previously described (Chim et al., 2011). Briefly, the

confluent SVEC cell monolayer was serum starved overnight prior to wounding. Cells were

incubated at 37ºC for 16 hrs with the conditioned medium containing EGFL7, or vehicle control.

Human bFGF (PeproTech Inc) was used as a positive control. Time-lapse images were captured

at 0 and 16 hr time points in the same position using a Nikon Eclipse TE2000-5 microscope. Five

selected fields of view were captured in each sample and the wound areas are estimated by

Nikon NIS-Elements computer software. For the scratch-wound healing assay with MEK1/2

inhibitor U0126 (Promega), STAT3 inhibitor Stattic (Sigma Aldrich) and RGD peptide (Sigma

Aldrich), SVEC monolayers were treated with inhibitors for 1 hr prior to wounding.

Tube Formation Assay

The migration assay was performed as previously described (Chim et al., 2011). Briefly, cells

were seeded onto the layer of Geltrex™ matrix (Invitrogen) and cultured with conditioned

medium containing EGFL7 or vehicle control for 24 hrs. Five random selected fields of view

were captured using a Nikon Eclipse TE2000-5 microscope. Tube formation was quantified by

measuring the length of tube-like structures by Nikon NIS-Elements computer software.

STAT3 Luciferase Reporter Assay

STAT3-responsive luciferase kit was purchased from Qiagen and used to measure the EGFL7 -

induced activation of STAT3. Briefly, SVEC were transfected with STAT3 Firefly luciferase

reporter plasmid and control Renilla luciferase plasmid using Lipofectamine 2000. Twenty-four

hours after transfection, cells were treated with conditioned medium containing EGFL7 or

recombinant human EGFL7 (OriGene) for an additional 24 hours. Luciferase activity was

determined using a Dual-Luciferase Reporter Assay kit (Promega). Firefly luciferase activity was

normalized to Renilla luciferase activity. The STAT3-responsive activity was presented as fold

change against controls.

Mouse Fetal Metatarsal Angiogenesis assay

The metatarsal angiogenesis assay was performed as previously described (Deckers et al., 2001).

Briefly, metatarsals were dissected from E17.5 embryos. The isolated metatarsals were cultured

in 24-well plates in complete α-MEM for 72 hours. The explants were cultured with human

EGFL7 (200ng/ml) in the presence or absence of inhibitors for 14 days, then fixed and stained

for CD31 (rabbit polyclonal, Abcam). Human VEGF (PeproTech Inc) (50ng/ml) was used as a

positive control. Images were captured using a Nikon Eclipse TE2000-5 microscope and

quantified by NIH imageJ software. Cultures were performed in sextuplicate.

Growth Plate Injury Repair Time Course

Eight-week-old male Sprague Dawley rats were subjected to experimental growth plate injury in

the proximal tibia of both hind legs as previously described (Xian et al., 2004). All protocols

followed the Australian code of practice for the care and use of animals, and were approved by

the Animal Ethics Committee of the SA Pathology, South Australia. Under anaesthesia, an

incision was made to expose the anterior-medial aspect of the proximal tibial bone of both hind-

limbs. A 2-mm surgical drill was then used to make a cortical window in the metaphyseal bone

on the medial side. A central disruption of the growth plate (about 30% area) was then induced

by inserting the drill through the cortical window and perpendicular through the growth plate

cartilage. Rats (n=8 per group and per time point) were sacrificed for specimen collection at days

6, 14 and 28 post-surgery, three time-points found suitable to observe the fibrogenic, osteogenic

and remodeling repair phases respectively (Xian et al., 2004). A normal group of rats (n=8) were

sacrificed on day 14 and used as non-injured controls. Both tibias were dissected, cleared of soft

tissue and samples collected, with the left tibia fixed, decalcified and paraffin-embedded for

collecting 4-μm sections for histological studies and the right tibia used to collect the repair

tissue within the growth plate injury site only for gene expression studies as described (Chung et

al., 2013).

Isolectin-B4 Labelling of Endothelial Cells

To investigate vascularization during growth plate injury repair, Isolectin-B4 labelling – (a

known marker for endothelial cells) was conducted (Grossmann et al., 2002). Isolated from

Griffonia simplicifolia (Bandeiraea), isolectin-B4 has previously been used to identify

neovascular structures in tumor networks (Niethammer et al., 2002). Briefly, sections were

deparaffinised and endogenous peroxidase quenched using 0.3% H2O2 in methanol. Sections

were blocked for 30 mins with 1% BSA/PBS solution and then incubated with biotinylated

isolectin-B4 (Vector Labs) (1:100 in PBS) for 30 mins. After washes, sections were incubated

with avidin-biotin complex reagent (Dako) for 30 mins and then liquid DAB Plus (Dako) for

colour development. Replacement of biotinylated isolectin-B4 with 1% BSA in PBS was used as

a negative control.

Gene Expression of EGFL7, OCN and VEGF at the Injured Growth Plate

Real-time quantitative RT-PCR assays were carried out to examine expression of EGFL7,

osteocalcin (OCN) and VEGFA in injured growth plate in rat, respectively. Total RNA from

growth plate injury site samples was extracted using the Aurum Total-RNA mini-kit (Bio-Rad)

and purified through a column (with on-column DNase treatment). cDNA was then synthesized

from 1μg RNA using iScript cDNA synthesis kit (Bio-Rad). Relative real-time PCR was carried

out using gene specific primer pairs for EGFL7 (forward 5’- TGTGGGATGGCAGGGAGATA-

3’ and reverse 5’-GGCACCAGTAACTTCCCACA-3’), Osteocalcin (forward 5’-

ATTGTGACGAGCTAGCGGAC-3’ and reverse 5’- TCGAGTCCTGGAGAGTAGCC -3’),

VEGFA (forward 5’- ATCTTCAAGCCGTCCTGTGTG-3’ and reverse 5’-

ATCTTCAAGCCGTCCTGTGTG-3’) and Cyclophilin-A (CycA) (forward 5’-

GAGCTGTTTGCAGACAAAGTTC-3’ and reverse 5’- CCCTGGCACATGAATCCTGG-3’).

CycA was used as an internal control (Zhou et al., 2004). Gene expression analysis was

expressed as fold change against non-injured controls.

Statistical Analysis and Data Presentation

All in vitro data shown represent one of at least three independent experiments. Data are

expressed as means ± SD and statistics were performed using Students’ t-test with significance

taken at p<0.05, except for real-time quantitative RT-PCR assays (mean + SEM). For

comparisons between time points, data were analysed using one-way ANOVA, and, when

significance levels (p<0.05) were achieved, a post hoc analysis of groups was done using a

Tukey’s test.

Results

Gene and protein expression of EGFL7 in bone microenvironment

In a previous study, we detected EGFL7 gene expression in both osteoclast and osteoblast

lineages (Chim et al., 2011). We have now confirmed the gene expression of EGFL7 during

primary osteoclast and osteoblast differentiation from BMM and calvariae; respectively by RT-

PCR. Osteoclast differentiation was followed using the osteoclast marker genes calcitonin

receptor (CTR), TRACP and dendritic cell-specific transmembrane protein (DC-STAMP), which

were upregulated during osteoclast differentiation (Fig. 1A). Gene expression of EGFL7 was

detected throughout osteoclast differentiation. Osteoblast differentiation was monitored by

following the expression of osteoblast associated marker genes alkaline phosphatase (ALP) and

OCN, and EGFL6 (known to be upregulated during osteoblast differentiation) (Chim et al.,

2011). As shown in Figure 1A, unlike the osteogenesis markers whose expression was induced,

EGFL7 gene levels were expressed at similar levels before and throughout osteoblast

differentiation (Fig. 1A). Furthermore, Western blot analyses revealed that EGFL7 protein

expression is present in developing and mature long bones (Fig. 1B). We also examined the

tissue distribution of EGFL7 from selected mouse organs by semi-quantitative RT-PCR. It was

revealed that EGFL7 was highly expressed in heart, kidney and lung (Fig. 1C).

Production of conditioned medium containing recombinant EGFL7

It has been reported that EGFL7 encodes a secreted protein (Campagnolo et al., 2005; Parker et

al., 2004). We sought to produce conditioned medium containing exogenous recombinant

EGFL7 proteins for further functional studies, as mouse recombinant EGFL7 protein expressed

in a mammalian system is not commercially available. To aid this, we generated an expression

construct of mouse full length EGFL7, pcDNA3.1-EGFL7-c-myc/His (Fig. 2A), and performed

transient transfection in COS-7 cells. As shown in Figure 2B, the c-myc antibody specifically

reacted to a major 34KDa and a minor 68KDa protein in EGFL7 transfected samples but not in

vehicle transfected samples by Western blot analysis (Fig. 2B). The 68 KDa EGFL7 protein has

not been previously described and might represent a homodimer or post-modified form of the

major 34 KDa EGFL7 protein. The identity of these proteins was further confirmed by Western

blot analysis using an anti-EGFL7 antibody (Fig. 2C).

Induction of cell migration and tube formation of SVEC endothelial cells and angiogenesis in

fetal mouse metatarsal explants by EGFL7

EGFL7 is a chemoattractant for endothelial cells (Campagnolo et al., 2005) and regulates

vascular development (Parker et al., 2004; Schmidt et al., 2007). As a first step to investigate the

role of EGFL7 in angiogenesis, we examined the effects of recombinant EGFL7 protein on

SVEC migration by a scratch wound healing assay. As shown in Figure 3A-B, conditioned

medium containing recombinant EGFL7 proteins significantly enhanced endothelial cell

migration. PBS and bFGF were used as negative and positive controls; respectively. This result

was consistent with the chemotactic effects of EGFL7 on endothelial cells by a transwell

migration assay (Campagnolo et al., 2005). As EGFL7-containing conditioned medium was

shown to be active in angiogenic assays it was used for subsequent functional studies.

The organization of endothelial cells into a three-dimensional tube-like structure is important in

angiogenesis. Next, we examined the effects of EGFL7 on angiogenesis in vitro by tube

formation assays. As shown in Figure 3C-D, EGFL7-containing conditioned medium

significantly enhanced tube-like structure formation after 24 hrs of culture compared with

vehicle and PBS controls. bFGF was used as a positive control. We further investigated the

effects of EGFL7 on angiogenesis in bone environment using an ex vivo fetal mouse metatarsal

angiogenesis assay. As shown in Figure 3E-F, human recombinant EGFL7 significantly induced

angiogenesis featuring web-like structures in cultured metatarsals.

Activation of ERK1/2, STAT3 and FAK signaling pathways by EGFL7

It has been suggested that EGFL7 mediates endothelial cell and neural stem cell activities

through modulating Notch signaling (Nichol et al., 2010; Schmidt et al., 2009) and promotes

hepatocellular carcinoma cell migration by activation of FAK through EGFR (Wu et al., 2009).

In an effort to address the mechanism of action of EGFL7 on endothelial cells, SVEC cells were

stimulated with conditioned medium containing recombinant EGFL7 or empty vehicle for 0, 5,

10, 20, 30 and 60 mins, and the cell lysates were analyzed by Western blotting for levels of

phosphorylation of key signaling molecules involved in cell migration. As shown in Figure 4,

EGFL7 (compared to vehicle control) induced the phosphorylation of ERK1/2 more than 2.5 fold

at 5 and 10 mins, STAT3 more than 1.5 fold at 30 and 60 mins, and FAK more than 2 fold at 30

and 60 mins. However, EGFL7 did not affect phosphorylation of ERK5, STAT1 and EGFR. In

addition, we examined the effects of EGFL7 on STAT3 transcriptional activity by STAT3-

responsive luciferase assay in which a luciferase reporter was driven by a promoter containing

STAT3 DNA-binding sequence. Luciferase reporter assays revealed that conditioned medium

containing EGFL7 induced STAT3-responsive activity ~2 fold (Fig. 4B). STAT3-responsive

luciferase activity was also induced by human EGFL7 in a dose dependent manner (Fig. 4C).

The induction level of conditioned medium was similar to that of 200ng/ml human EGFL7 (Fig.

4B, 4C), which is consistent with the estimated concentration of EGFL7 conditioned medium. To

further confirm that ERK1/2 and STAT3 pathways are involved in EGFL7-mediated endothelial

cell activities, we examined the effects of U0126 (MEK1/2 inhibitor) and Stattic (STAT3

inhibitor) on EGFL7-induced endothelial cell migration. As shown in Figure 5A-B, U0126 and

Stattic significantly inhibited EGFL7-induced endothelial cell migration.

EGFL7-induced angiogenesis ex vivo is mediated through the activation of ERK1/2 and STAT3

Having demonstrated that EGFL7 influences endothelial cell migration through ERK1/2 and

STAT3 signalings, we next investigated whether EGFL7 regulates angiogenesis via the similar

mechanisms by using ex vivo fetal mouse metatarsal angiogenesis assay. As shown in Figure 6A-

B, human EGFL7 significantly enhanced blood vessel growth from metatarsals. Interestingly,

EGFL7-induced angiogenesis was significantly inhibited by U0126 and Stattic. The inhibitors

alone had no significant effect on angiogenesis as compared to PBS control (Figure 6).

RGD Peptides Impaired EGFL7-induced Endothelial Cell Migration

Bioinformatic analyses indicate that mouse and rat EGFL7 proteins contain a conserved QGD

motif, and human EGFL7 an RGD motif, indicative of involvement in integrin signaling

(Supplementary Fig. S2). It has been reported that the change of RGD to QGD maintains integrin

binding ability (Elphick et al., 2009; Erb et al., 2001; Gresham et al., 1992). To determine

whether EGFL7 signaling is mediated through its RGD/QGD functional domain, we examined

the effect of RGD peptides on the EGFL7-induced endothelial cell migration. Cells were

pretreated with RGD peptides for 1 hr prior to performing scratch wound healing assays. As

shown in Figure 7, EGFL7-induced endothelial cell migration was significantly reduced in the

presence of RGD peptides at concentrations of 250ng/ml and 500ng/ml. bFGF-induced

endothelial migration was not inhibited by RGD peptides. These results suggest that EGFL7

regulates angiogenic activities through its RGD/QGD domain and integrin receptor complex.

Upregulation of EGFL7 gene expression during growth plate injury repair- To further confirm

the involvement of EGFL7 in the bone microenvironment in vivo, its expression during the bony

repair time course was examined in a rat model of tibial growth plate injury repair, which is

known to involve mesenchymal cell infiltration, osteoblastic differentiation and bone formation,

angiogenesis, and osteoclast recruitment and remodeling (Chung et al., 2011; Chung et al.,

2009). Haematoxylin and eosin (H&E) and alcian blue staining revealed an infiltration of

mesenchymal-like tissue and other cells (Fig. 8A) at day 6 following injury. By day 14 post-

injury, the cells within the injury site had undergone differentiation and tissues such as bone

trabeculae were evident. Signs of bone remodeling were also evident with the presence of

osteoblasts and osteoclasts surrounding the newly formed bony repair tissue (Fig. 8B). To

investigate vascularization during growth plate injury repair and bony repair tissue formation,

expression of isolectin-B4 was used to label endothelial cells during growth plate injury repair.

Within the growth plate injury site itself, there was an increase in vessel number and size at day

14 compared to day 6 (Fig. 8C-D).

Quantitative RT-PCR gene expression analysis of the growth plate injury site showed increased

EGFL7 expression post-injury peaking at day 14 during the repairing process (Fig. 8E). By day

28, however, levels of EGFL7 at the injury site had decreased to near or below the basal level.

No significant changes were seen in OCN expression at days 6 and 14 (despite a trend of

increase on day 14) until a significant 30-fold increase at day 28 in comparison to non-injured

controls (Fig. 8F). Interestingly, compared to the normal growth plate cartilage, levels of VEGFA

were significantly lower on days 6 and 14 followed by an increase back up to basal levels on day

28 (Fig. 8G). VEGFA levels at day 28 were more than 1.5 fold greater than at day 6 and 14.

Discussion

The crosstalk between bone cells and endothelial cells is important for normal bone growth,

remodeling and fracture healing (Brandi and Collin-Osdoby, 2006; Chim et al., 2013; Eriksen et

al., 2007; Fang et al., 2005). Identification of angiogenic factors produced by bone cells could

enhance our understanding of intercellular communication in the BRC, which might help to

develop new targets for the treatment of bone fracture and bone disorders such as osteoporosis

and osteonecrosis. The present study demonstrates that EGFL7 is expressed in osteoclasts,

osteoblasts and bone tissues, and that it regulates endothelial cell migration, angiogenesis and

integrin-mediated signaling of ERK1/2, STAT3, and FAK.

EGFL7 is a secreted protein that contains an EMI domain at the N-terminus, followed by two

EGF repeat domains. We have detected a 68 KDa EGFL7 protein which might represent a

homodimer or post-modified form of EGFL7 protein. Future studies will be required to

investigate the protein modification of EGFL7 and functions of the isoforms. Fitch et al. has

suggested that EGFL7 is expressed in highly vascularized tissues including brain, heart, lung,

kidney and ovary (Fitch et al., 2004). EGFL7 is also expressed in endothelial progenitors and

endothelial cells (Fitch et al., 2004; Parker et al., 2004). Loss of EGFL7 function in zebrafish

embryos results in a vascular tubulogenesis defect (Parker et al., 2004). Moreover, vasculature

development in organs is reported to be delayed in EGFL7 knockout mice (Schmidt et al., 2007).

These results suggest that EGFL7 regulates angiogenesis by an autocrine mechanism. Here we

have demonstrated that EGFL7 is expressed by osteoclasts and osteoblasts allowing us to

hypothesize that EGFL7, as an extracellular protein, regulates capillaries near the BRC by a

paracrine mechanism.

Initially, the mechanism of EGFL7-mediated cellular activities was reported to involve the

binding of EGFL7 to the extracellular domain of Notch, allowing it to act as an antagonist of

Notch signaling in neural stem cells (Schmidt et al., 2009). A further study demonstrated that

EGFL7 modulates endothelial cell activities by a similar mechanism (Nichol et al., 2010). Other

studies have proposed that EGFL7 promotes human hepatocellular carcinoma cell migration by

EGFR-mediated activation of FAK (Wu et al., 2009). Surprisingly, we could not detect any

phosphorylation of EGFR by EGFL7 in endothelial cells when adding as an exogenous factor. It

is important to note that human EGFL7 protein contains an RGD motif whereas mouse and rat

EGFL7 proteins consist of a QGD motif, raising an intriguing possibility that EGFL7 might

interact with integrins. It has been reported the change of RGD to QGD maintains integrin

binding ability (Elphick et al., 2009; Erb et al., 2001; Gresham et al., 1992). Consistent with this

notion, we found that EGFL7-induced endothelial cell migration was significantly reduced in the

presence of RGD peptides, attesting that EGFL7 could function via a RGD/QGD motif.

Interestingly, a recent study demonstrated that EGFL7 interacts with integrin αvβ3 and regulates

endothelial cell motility (Nikolic et al., 2013). The binding of ligand to integrin triggers

intracellular signaling cascades (Kumar, 1998; Leveille et al., 2007; Shattil and Newman, 2004),

including MAPK, STAT and FAK (Bauvois, 2012; Behera et al., 2010; Guo and Giancotti, 2004;

Parsons, 2003), which are important in cell migration (Chim et al., 2011; Long et al., 2010;

Mehta and Besner, 2007; Teng et al., 2009). In this study, we demonstrated that recombinant

EGFL7 induced the phosphorylation of ERK1/2, STAT3 and FAK in endothelial cells for the first

time. Moreover, inhibition of ERK1/2 and STAT3 signaling pathways blocks EGFL7-induced

endothelial cell migration and angiogenesis. Taken together, these results suggest that EGFL7 as

an exogenous factor regulates endothelial cell activities through its RGD/QGD functional

domain and integrin-mediated signaling.

As a step to demonstrate the involvement of EGFL7 in bone remodeling in skeletal repair,

EGFL7 was found upregulated during growth plate injury repair. The growth plate injury model

involves an initial inflammatory response (day 1-3) followed by a fibrogenic response involving

mesenchymal infiltration (day 3-7), trabecular bone formation (day8-14) and remodeling (day 14

onwards) (Chung et al., 2011). Angiogenesis at the growth plate injury site takes place starting

from the mesenchymal cell repair phase and is more prominent during the bone formation phase

(day 14). It has been suggested that growth plate injury responses and bony repair mechanism are

similar to the process of bone fracture healing (Chung et al., 2011). However, the processes of

revascularization during growth plate injury healing and fracture healing remain unclear. Kilarski

et al suggested that translocation of pre-existing vascular network is responsible for the rapid

formation of vasculature in granulation tissues at early stages of tissue healing. The vasculature

is then further remodeled by angiogenic splitting and sprouting during later stages of healing to

fine-tune vascular networks (Kilarski et al., 2009). It has been demonstrated that neovessel

formation in granulation tissue is mediated by intussusceptive angiogenesis, a mechanism at

relatively low energy costs as compared to vessel sprouting (Burri and Djonov, 2002; Kilarski et

al., 2009). The current study showed that EGFL7 gene expression was markedly upregulated

during the growth plate injury repair process at the period of time when VEGF expression was

found relatively low. VEGFA/VEGFR2 signaling has been postulated to regulate angiogenic

sprouting (Gerhardt et al., 2003). Consistently, Kilarski et al demonstrated that early steps of

vascularization in tissue repair are independent of angiogenic sprouting (Kilarski et al., 2009).

Recently, it has been demonstrated that inhibition of Notch signaling induces intussusceptive

angiogenesis (Dill et al., 2012; Dimova et al., 2013). Given that EGFL7 is an antagonist of Notch

signaling (Schmidt et al., 2009), EGFL7 could mediate intussusceptive angiogenesis in

granulation tissues during the early stages of growth plate repair. Furthermore, EGFL7 might

regulate sprouting angiogenesis during osteogenesis at the later stages of healing through

integrin-mediated signaling of ERK1/2, STAT3, and FAK.

Recently, we have reviewed the expression of angiogenic factors in the bone microenvironment,

and the complex interrelationship between angiogenesis and osteogenesis (Chim et al., 2013).

The spatiotemporal regulation of molecules mediates different stages and types of angiogenesis

in both physiological and pathological conditions (Arima et al., 2011; Chim et al., 2013; Clapp et

al., 2009). EGF-like family members have been implicated in angiogenesis (Kim et al., 2003;

Leker et al., 2009; Mehta and Besner, 2007; Schmidt et al., 2007). In the bone

microenvironment, EGF, heparin binding-EGF, amphiregulin, betacellulin, and transforming

growth factor-α are expressed in osteoclasts and osteoblasts and regulate angiogenesis

(Nakamura et al., 2010; Qin et al., 2005; Yi et al., 2008). Our previous study and the current

work have identified that EGFL6 and EGFL7 are both expressed in the bone local environment

and have a regulatory role in angiogenesis (Chim et al., 2011). EGFL2, EGFL3, EGFL5, EGFL8

and EGFL9 are relatively newly identified EGF-like family members, and their cellular functions

are still not fully understood. It is possible that EGF family members are expressed by different

cell types in bone local environment, and regulate angiogenesis during fracture healing and bone

remodeling. Future studies will be needed to address the specific roles of EGF family members

in angiogenesis and their signaling cascades.

In summary, we have shown that EGFL7 is expressed in the bone microenvironment and has a

role in promoting endothelial cell migration and tube-like structure formation through integrin-

mediated pathways. This study highlights the important role that EGFL7 plays in the regulation

of endothelial cell activities in the bone microenvironment, and might help to develop novel

therapeutic approaches for bone fracture and bone disorders such osteonecrosis.

Acknowledgments:

This work was funded in part by National Health and Medical Research Council of Australia,

Western Australia Medical & Health Research Infrastructure Fund, and a grant from the National

Natural Science Foundation of China (NSFC, no. 81228013). Prof. Jiake Xu is a visiting

Professor to the Research Centre for Regenerative Medicine, Guangxi Medical University in

2013. Professor Vicki Rosen was an Australia-Harvard Fellow to the School of Pathology and

Laboratory Medicine, the University of Western Australia in 2012.

References

Andersen TL, Sondergaard TE, Skorzynska KE, Dagnaes-Hansen F, Plesner TL, Hauge EM, Plesner T, Delaisse JM.

2009. A physical mechanism for coupling bone resorption and formation in adult human bone. Am J Pathol

174(1):239-247.

Arima S, Nishiyama K, Ko T, Arima Y, Hakozaki Y, Sugihara K, Koseki H, Uchijima Y, Kurihara Y, Kurihara H.

2011. Angiogenic morphogenesis driven by dynamic and heterogeneous collective endothelial cell

movement. Development 138(21):4763-4776.

Bakker AD, Klein-Nulend J. 2012. Osteoblast isolation from murine calvaria and long bones. Methods Mol Biol

816:19-29.

Bauvois B. 2012. New facets of matrix metalloproteinases MMP-2 and MMP-9 as cell surface transducers: outside-

in signaling and relationship to tumor progression. Biochim Biophys Acta 1825(1):29-36.

Behera R, Kumar V, Lohite K, Karnik S, Kundu GC. 2010. Activation of JAK2/STAT3 signaling by osteopontin

promotes tumor growth in human breast cancer cells. Carcinogenesis 31(2):192-200.

Brandi ML, Collin-Osdoby P. 2006. Vascular biology and the skeleton. J Bone Miner Res 21(2):183-192.

Burkus JK, Ganey TM, Ogden JA. 1993. Development of the cartilage canals and the secondary center of

ossification in the distal chondroepiphysis of the prenatal human femur. Yale J Biol Med 66(3):193-202.

Burri PH, Djonov V. 2002. Intussusceptive angiogenesis--the alternative to capillary sprouting. Mol Aspects Med

23(6S):S1-27.

Campagnolo L, Leahy A, Chitnis S, Koschnick S, Fitch MJ, Fallon JT, Loskutoff D, Taubman MB, Stuhlmann H.

2005. EGFL7 is a chemoattractant for endothelial cells and is up-regulated in angiogenesis and arterial

injury. Am J Pathol 167(1):275-284.

Carlevaro MF, Cermelli S, Cancedda R, Descalzi Cancedda F. 2000. Vascular endothelial growth factor (VEGF) in

cartilage neovascularization and chondrocyte differentiation: auto-paracrine role during endochondral bone

formation. J Cell Sci 113 ( Pt 1):59-69.

Chim SM, Qin A, Tickner J, Pavlos N, Davey T, Wang H, Guo Y, Zheng MH, Xu J. 2011. EGFL6 promotes

endothelial cell migration and angiogenesis through the activation of extracellular signal-regulated kinase. J

Biol Chem 286(25):22035-22046.

Chim SM, Tickner J, Chow ST, Kuek V, Guo B, Zhang G, Rosen V, Erber W, Xu J. 2013. Angiogenic factors in bone

local environment. Cytokine & growth factor reviews 24(3):297-310

Chung R, Foster BK, Xian CJ. 2011. Injury responses and repair mechanisms of the injured growth plate. Front

Biosci (Schol Ed) 3:117-125.

Chung R, Foster BK, Zannettino AC, Xian CJ. 2009. Potential roles of growth factor PDGF-BB in the bony repair of

injured growth plate. Bone 44(5):878-885.

Chung R, Wong D, Macsai C, Piergentili A, Del Bello F, Quaglia W, Xian CJ. 2013. Roles of Wnt/beta-catenin

signalling pathway in the bony repair of injured growth plate cartilage in young rats. Bone 52(2):651-658.

Clapp C, Thebault S, Jeziorski MC, Martinez De La Escalera G. 2009. Peptide hormone regulation of angiogenesis.

Physiol Rev 89(4):1177-1215.

Collin-Osdoby P, Rothe L, Anderson F, Nelson M, Maloney W, Osdoby P. 2001. Receptor activator of NF-kappa B

and osteoprotegerin expression by human microvascular endothelial cells, regulation by inflammatory

cytokines, and role in human osteoclastogenesis. J Biol Chem 276(23):20659-20672.

De Maziere A, Parker L, Van Dijk S, Ye W, Klumperman J. 2008. Egfl7 knockdown causes defects in the extension

and junctional arrangements of endothelial cells during zebrafish vasculogenesis. Dev Dyn 237(3):580-591.

Deckers M, van der Pluijm G, Dooijewaard S, Kroon M, van Hinsbergh V, Papapoulos S, Lowik C. 2001. Effect of

angiogenic and antiangiogenic compounds on the outgrowth of capillary structures from fetal mouse bone

explants. Lab Invest 81(1):5-15.

Dill MT, Rothweiler S, Djonov V, Hlushchuk R, Tornillo L, Terracciano L, Meili-Butz S, Radtke F, Heim MH,

Semela D. 2012. Disruption of Notch1 induces vascular remodeling, intussusceptive angiogenesis, and

angiosarcomas in livers of mice. Gastroenterology 142(4):967-977 e962.

Dimova I, Hlushchuk R, Makanya A, Styp-Rekowska B, Ceausu A, Flueckiger S, Lang S, Semela D, Le Noble F,

Chatterjee S, Djonov V. 2013. Inhibition of Notch signaling induces extensive intussusceptive neo-

angiogenesis by recruitment of mononuclear cells. Angiogenesis 16(4):921-937.

Elphick GF, Sarangi PP, Hyun YM, Hollenbaugh JA, Ayala A, Biffl WL, Chung HL, Rezaie AR, McGrath JL,

Topham DJ, Reichner JS, Kim M. 2009. Recombinant human activated protein C inhibits integrin-mediated

neutrophil migration. Blood 113(17):4078-4085.

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. 2001. An RGD sequence in the P2Y(2) receptor interacts with

alpha(V)beta(3) integrins and is required for G(o)-mediated signal transduction. J Cell Biol 153(3):491-

501.

Eriksen EF, Eghbali-Fatourechi GZ, Khosla S. 2007. Remodeling and vascular spaces in bone. J Bone Miner Res

22(1):1-6.

Fang TD, Salim A, Xia W, Nacamuli RP, Guccione S, Song HM, Carano RA, Filvaroff EH, Bednarski MD, Giaccia

AJ, Longaker MT. 2005. Angiogenesis is required for successful bone induction during distraction

osteogenesis. J Bone Miner Res 20(7):1114-1124.

Fitch MJ, Campagnolo L, Kuhnert F, Stuhlmann H. 2004. Egfl7, a novel epidermal growth factor-domain gene

expressed in endothelial cells. Dev Dyn 230(2):316-324.

Gerber HP, Ferrara N. 2000. Angiogenesis and bone growth. Trends Cardiovasc Med 10(5):223-228.

Gerhardt H, Golding M, Fruttiger M, Ruhrberg C, Lundkvist A, Abramsson A, Jeltsch M, Mitchell C, Alitalo K,

Shima D, Betsholtz C. 2003. VEGF guides angiogenic sprouting utilizing endothelial tip cell filopodia. J

Cell Biol 161(6):1163-1177.

Gresham HD, Adams SP, Brown EJ. 1992. Ligand binding specificity of the leukocyte response integrin expressed

by human neutrophils. J Biol Chem 267(20):13895-13902.

Grossmann R, Stence N, Carr J, Fuller L, Waite M, Dailey ME. 2002. Juxtavascular microglia migrate along brain

microvessels following activation during early postnatal development. Glia 37(3):229-240.

Guo J, Ren F, Wang Y, Li S, Gao Z, Wang X, Ning H, Wu J, Li Y, Wang Z, Chim SM, Xu J, Chang Z. 2012. miR-

764-5p promotes osteoblast differentiation through inhibition of CHIP/STUB1 expression. J Bone Miner

Res 27(7):1607-1618.

Guo W, Giancotti FG. 2004. Integrin signalling during tumour progression. Nat Rev Mol Cell Biol 5(10):816-826.

Hauge EM, Qvesel D, Eriksen EF, Mosekilde L, Melsen F. 2001. Cancellous bone remodeling occurs in specialized

compartments lined by cells expressing osteoblastic markers. J Bone Miner Res 16(9):1575-1582.

Kilarski WW, Samolov B, Petersson L, Kvanta A, Gerwins P. 2009. Biomechanical regulation of blood vessel

growth during tissue vascularization. Nat Med 15(6):657-664.

Kim HS, Shin HS, Kwak HJ, Cho CH, Lee CO, Koh GY. 2003. Betacellulin induces angiogenesis through activation

of mitogen-activated protein kinase and phosphatidylinositol 3'-kinase in endothelial cell. FASEB J

17(2):318-320.

Kumar CC. 1998. Signaling by integrin receptors. Oncogene 17(11 Reviews):1365-1373.

Leker RR, Toth ZE, Shahar T, Cassiani-Ingoni R, Szalayova I, Key S, Bratincsak A, Mezey E. 2009. Transforming

growth factor alpha induces angiogenesis and neurogenesis following stroke. Neuroscience 163(1):233-

243.

Leveille C, Bouillon M, Guo W, Bolduc J, Sharif-Askari E, El-Fakhry Y, Reyes-Moreno C, Lapointe R, Merhi Y,

Wilkins JA, Mourad W. 2007. CD40 ligand binds to alpha5beta1 integrin and triggers cell signaling. J Biol

Chem 282(8):5143-5151.

Long W, Yi P, Amazit L, LaMarca HL, Ashcroft F, Kumar R, Mancini MA, Tsai SY, Tsai MJ, O'Malley BW. 2010.

SRC-3Delta4 mediates the interaction of EGFR with FAK to promote cell migration. Mol Cell 37(3):321-

332.

Mehta VB, Besner GE. 2007. HB-EGF promotes angiogenesis in endothelial cells via PI3-kinase and MAPK

signaling pathways. Growth Factors 25(4):253-263.

Nakamura T, Toita H, Yoshimoto A, Nishimura D, Takagi T, Ogawa T, Takeya T, Ishida-Kitagawa N. 2010. Potential

involvement of Twist2 and Erk in the regulation of osteoblastogenesis by HB-EGF-EGFR signaling. Cell

Struct Funct 35(1):53-61.

Nichol D, Shawber C, Fitch MJ, Bambino K, Sharma A, Kitajewski J, Stuhlmann H. 2010. Impaired angiogenesis

and altered Notch signaling in mice overexpressing endothelial Egfl7. Blood 116(26):6133-6143.

Niethammer AG, Xiang R, Becker JC, Wodrich H, Pertl U, Karsten G, Eliceiri BP, Reisfeld RA. 2002. A DNA

vaccine against VEGF receptor 2 prevents effective angiogenesis and inhibits tumor growth. Nat Med

8(12):1369-1375.

Nikolic I, Dudvarski Stankovic N, Bicker F, Meister J, Braun H, Awwad K, Baumgart J, Simon K, Thal SC, Patra C,

Harter PN, Plate KH, Engel FB, Dimmeler S, Eble JA, Mittelbronn M, Schafer MK, Jungblut B, Chavakis

E, Fleming I, Schmidt MH. 2013. EGFL7 ligates alphavbeta3 integrin to enhance vessel formation. Blood.

Parker LH, Schmidt M, Jin SW, Gray AM, Beis D, Pham T, Frantz G, Palmieri S, Hillan K, Stainier DY, De Sauvage

FJ, Ye W. 2004. The endothelial-cell-derived secreted factor Egfl7 regulates vascular tube formation.

Nature 428(6984):754-758.

Parsons JT. 2003. Focal adhesion kinase: the first ten years. J Cell Sci 116(Pt 8):1409-1416.

Penta K, Varner JA, Liaw L, Hidai C, Schatzman R, Quertermous T. 1999. Del1 induces integrin signaling and

angiogenesis by ligation of alphaVbeta3. J Biol Chem 274(16):11101-11109.

Qin L, Tamasi J, Raggatt L, Li X, Feyen JH, Lee DC, Dicicco-Bloom E, Partridge NC. 2005. Amphiregulin is a

novel growth factor involved in normal bone development and in the cellular response to parathyroid

hormone stimulation. J Biol Chem 280(5):3974-3981.

Schmidt M, Paes K, De Maziere A, Smyczek T, Yang S, Gray A, French D, Kasman I, Klumperman J, Rice DS, Ye

W. 2007. EGFL7 regulates the collective migration of endothelial cells by restricting their spatial

distribution. Development 134(16):2913-2923.

Schmidt MH, Bicker F, Nikolic I, Meister J, Babuke T, Picuric S, Muller-Esterl W, Plate KH, Dikic I. 2009.

Epidermal growth factor-like domain 7 (EGFL7) modulates Notch signalling and affects neural stem cell

renewal. Nat Cell Biol 11(7):873-880.

Shattil SJ, Newman PJ. 2004. Integrins: dynamic scaffolds for adhesion and signaling in platelets. Blood

104(6):1606-1615.

Singh AB, Harris RC. 2005. Autocrine, paracrine and juxtacrine signaling by EGFR ligands. Cell Signal

17(10):1183-1193.

Teng TS, Lin B, Manser E, Ng DC, Cao X. 2009. Stat3 promotes directional cell migration by regulating Rac1

activity via its activator betaPIX. J Cell Sci 122(Pt 22):4150-4159.

Thi MM, Suadicani SO, Spray DC. 2010. Fluid flow-induced soluble vascular endothelial growth factor isoforms

regulate actin adaptation in osteoblasts. J Biol Chem 285(40):30931-30941.

Utting JC, Flanagan AM, Brandao-Burch A, Orriss IR, Arnett TR. 2010. Hypoxia stimulates osteoclast formation

from human peripheral blood. Cell Biochem Funct 28(5):374-380.

Wu F, Yang LY, Li YF, Ou DP, Chen DP, Fan C. 2009. Novel role for epidermal growth factor-like domain 7 in

metastasis of human hepatocellular carcinoma. Hepatology 50(6):1839-1850.

Wu X, Chim SM, Kuek V, Lim BS, Chow ST, Zhao J, Yang S, Rosen V, Tickner J, Xu J. 2014. HtrA1 is upregulated

during RANKL-induced osteoclastogenesis, and negatively regulates osteoblast differentiation and BMP2-

induced Smad1/5/8, ERK and p38 phosphorylation. FEBS Lett 588(1):143-150.

Xian CJ, Zhou FH, McCarty RC, Foster BK. 2004. Intramembranous ossification mechanism for bone bridge

formation at the growth plate cartilage injury site. J Orthop Res 22(2):417-426.

Xu J, Tan JW, Huang L, Gao XH, Laird R, Liu D, Wysocki S, Zheng MH. 2000. Cloning, sequencing, and

functional characterization of the rat homologue of receptor activator of NF-kappaB ligand. J Bone Miner

Res 15(11):2178-2186.

Yi T, Lee HL, Cha JH, Ko SI, Kim HJ, Shin HI, Woo KM, Ryoo HM, Kim GS, Baek JH. 2008. Epidermal growth

factor receptor regulates osteoclast differentiation and survival through cross-talking with RANK signaling.

J Cell Physiol 217(2):409-422.

Zhou FH, Foster BK, Sander G, Xian CJ. 2004. Expression of proinflammatory cytokines and growth factors at the

injured growth plate cartilage in young rats. Bone 35(6):1307-1315.

Figure Legends

Figure 1. The gene and protein expression of EGFL7 in bone microenvironment. (A) RT-

PCR amplification of EGFL7 showing that EGFL7 was expressed during osteoclast and

osteoblast differentiation. Primers specific for osteoclast marker genes CTR, TRACP and DC-

STAMP were included as markers for osteoclast differentiation. EGFL6, ALP and OCN were

used as markers for osteoblast differentiation. TRACP staining for osteoclasts and alizarin red

staining for osteoblast mineralization were also included in parallel experiments. Scale bar, 100

µm. (B) Western blot analyses showing that EGFL7 protein expression was present in

developing and mature long bones. (C) Tissue expression profile of EGFL7 determined by RT-

PCR analysis.

Figure 2. Production of conditioned medium containing EGFL7 protein. (A) An expression

construct encoding mouse full length EGFL7, pcDNA3.1-EGFL7-c-myc/His, was generated. A

conserved QGD motif is located between EGF repeats. (B-C) Detection of EGFL7 proteins,

using anti-c-myc and anti-EGFL7 antibodies, in COS-7 cells transfected with empty vector or

expression vector encoding c-myc/His-tagged EGFL7. EGFL7 was detected in both medium and

cell lysates. Black arrows indicate the EGFL7 proteins at the positions of 34KDa and 68KDa;

respectively.

Figure 3. EGFL7 promotes endothelial cell migration, tube-like structure formation and

angiogenesis. (A) Representative microscopic views of scratch wound healing assays performed

in SVEC cells treated with conditioned medium containing EGFL7 or vehicle control for 0 or 16

hrs. Scale bar, 100 µm. (B) Quantitative analysis of cell migration area after 16 hrs showing

EGFL7 enhanced endothelial cell migration. (C) Representative photographs showing tube

formation by SVEC cells cultured for 24 hrs in conditioned medium containing EGFL7 or

vehicle control. Scale bar, 100 µm. (D) Quantitative analysis of tube length showing EGFL7

promoted tube-like structure formation. PBS and bFGF being used as a negative and positive

control, respectively. (E) Mouse metatarsals were dissected from E17.5 embryos for an ex vivo

metatarsal angiogenesis assay. Representative microscopic images showing that human EGFL7

(200 ng/ml) induced blood vessel growth from metatarsals. PBS and VEGF (50 ng/ml) were

used as a negative and positive control; respectively. Scale bar, 250 µm. (F) Quantitative analysis

of vessel sprouting. **p<0.01, ***p<0.001.

Figure 4. EGFL7 induced phosphorylation of ERK1/2, STAT3 and FAK. (A) Conditioned

medium containing EGFL7 induced the cellular phosphorylation levels of ERK1/2 at 5 and 10

mins, STAT3 and FAK at 30 and 60 mins compared to vehicle control. EGFL7 did not induce

phosphorylation of ERK5, STAT1 and EGFR. β-actin was used as loading control. Signal

intensities were quantified by imageJ software. Induction ratios at each time point were

compared to 0 min, with p-ERK1/2 normalized to ERK1/2, p-STAT3 normalized to STAT3, and

p-FAK normalized to β-actin. (B-C) Determination of STAT3 transcription activity in SVEC by

STAT3-responsive luciferase reporter assays. EGFL7 conditioned medium and recombinant

human EGFL7 significantly induced STAT3-responsive luciferase activities. Luciferase activity

was compared to vehicle or PBS. *p<0.05, **p<0.01.

Figure 5. EGFL7-induced endothelial cell migration was blocked by U0126 and Stattic. (A)

Representative microscopic views of scratch wound healing assays showing EGFL7-induced

endothelial cell migration was inhibited in the presence of U0126 (5µM) and Stattic (1µM).

Scale bar, 100 µm. (B) Quantitative analysis of cell migration area after 16 hrs showing U0126

and Stattic significantly inhibited EGFL7-induced endothelial cell migration. PBS and bFGF

being used as a negative and positive control, respectively. *p<0.05.

Figure 6. EGFL7-induced angiogenesis was blocked by U0126 and Stattic. (A) Mouse

metatarsals were dissected from E17.5 embryos for an ex vivo metatarsal angiogenesis assay.

Representative microscopic images showing that EGFL7-induced angiogenesis was inhibited in

the presence of U0126 (5µM) and Stattic (1µM). Inhibitors alone had no effects on angiogenesis

as compared to PBS control. Scale bar, 250 µm. (B) Quantitative analysis of vessel sprouting.

***p<0.001.

Figure 7. EGFL7-induced endothelial cell migration was blocked by RGD peptides. (A)

Scratch wound healing assays performed with the conditioned medium containing EGFL7 or

vehicle control in the presence (250 or 500ng/ml) or absence of RGD peptides for 16 hrs

showing that EGFL7-induced endothelial cell migration was reduced by RGD peptides. EGFL7-

induced endothelial cell migration was not inhibited by RGD peptides. Scale bar, 100 µm. (B)

Quantitative analysis of cell migration area after 16 hrs showing that RGD EGFL7-induced

endothelial cell migration was significantly inhibited by RGD peptides at the concentration of

250ng/ml and 500ng/ml. PBS and bFGF being used as a negative and positive control;

respectively. *p<0.05.

Figure 8. EGFL7 is upregulated during growth plate injury repair. (A) Histology revealed a

majority of mesenchymal repair tissue (Mes) at the injury site on day 6 post- growth plate injury.

(B) By day 14, bone trabeculae repair tissue (Bt) surrounded by osteoblasts (OB) and osteoclasts

(OC). (C-D) Isolectin-B4 positive endothelial cells at the growth plate injury site (C) 6 days and

(D) 14 days post- injury. (E-G) Quantitative real-time RT-PCR expression data for (E) EGFL7,

(F) OCN and (G) VEGFA are expressed as fold change to non-injured controls. Original scale

bar= 125µm (applies to A-D). **p<0.01.

B

Figure 1

C

A

EGFL7

0 1 3 5

Osteoclast Differentiation

Day

CTR

18S

TRACP

DC-STAMP

Day0 Day5

EGFL7

0 7 14 Day

Osteoblast Differentiation

EGFL6

ALP

OCN

18S

21

Day0 Day21

β-actin

EGFL7

0

0.5

1

1.5

2

Ra

tio

(C

om

pa

red

to

18

s)

EGFL7

18s

Figure 2

B C

A

37

25

KDa

50

75

Medium Lysate

WB: EGFL7

37

25

KDa

50

75

WB: c-myc

Medium Lysate

c-myc/

His EGFL7-c-myc/His

QGD motif

1 131 275 aa

A

C

B

Ce

ll M

igra

ted

Are

a

0%

50%

100%

150%

200% *** **

***

PBS bFGF

Vehicle EGFL7

0%

50%

100%

150%

200%

250%

300%

Tu

be

le

ng

th

***

**

** D

Figure 3

PBS bFGF

0h

r 16h

r

Vehicle EGFL7

PBS VEGF EGFL7

E F

0%

50%

100%

150%

200%

250%

Va

sc

ula

r D

en

sit

y (

%)

*** ***

Figure 4

Vehicle

0 5 10 20 30 60 mins

p-ERK1/2

STAT1

p-STAT1

p-STAT3

STAT3

EGFL7

5 10 20 30 60

β-actin

ERK1/2

p-ERK5

p-FAK

p-EGFR

1.0 1.6 1.6 1.7 1.9 0.9 4.6 4.3 2.9 2.5 1.6

1.0 1.6 2.0 1.3 2.1 1.9 1.3 1.8 2.5 3.2 3.5

1.0 0.4 0.5 0.7 0.3 0.9 1.0 0.8 0.8 1.9 1.8

B

A

0.0

0.5

1.0

1.5

2.0

2.5

ST

AT

3-r

es

po

ns

ive

lucif

era

se a

cti

vit

y (

Fo

ld) *

C

0.0

0.5

1.0

1.5

2.0

2.5

hEGFL7(ng/ml)

50 100 200

ST

AT

3-r

es

po

ns

ive

lucif

era

se a

cti

vit

y (

Fo

ld)

** *

Vehicle+Stattic EGFL7+Stattic

0h

r 1

6h

r

Vehicle+U0126 EGFL7+U0126

Figure 5

0h

r 1

6h

r Vehicle EGFL7 PBS bFGF

0%

50%

100%

150%

200%

250%

Cell

Mig

rate

d A

rea

*

*

*

*

B

A

PBS EGFL7

EGFL7+U0126 EGFL7+Stattic

Stattic U0126

A B

0%

50%

100%

150%

200%

250%

Va

sc

ula

r D

en

sit

y (

%)

***

***

***

Figure 6

Figure 7

Vehicle EGFL7

0h

r 1

6h

r

PBS bFGF

0h

r 1

6h

r 0

hr

16

hr

RG

D(0

ng

/ml)

R

GD

(25

0n

g/m

l)

RG

D(5

00

ng

/ml)

A

B

RGD(0ng/ml) RGD(250ng/ml) RGD(500ng/ml)

0%

50%

100%

150%

200% *

*

Ce

ll M

igra

tio

n A

rea

* * *

* *

D

D14, isolectin-B4, injury site

B

D14, H&E, injury site

A

D6, H&E, injury site

Bt

Bt

* * OC

OB Mes

0.0

1.0

2.0

3.0

4.0

Control D6 D14 D28

EGFL7 E

**

** **

0.0

10.0

20.0

30.0

40.0

Control D6 D14 D28

OCN F

**

** **

0.0

0.5

1.0

1.5

2.0

Control D6 D14 D28

VEGFA G

** **

** **

Figure 8

Fo

ld c

ha

ng

e

Fo

ld c

ha

ng

e

Fo

ld c

ha

ng

e

D6, isolectin-B4, injury site

Mes

C