12
ORIGINAL PAPER The differential roles of Slit2-exon 15 splicing variants in angiogenesis and HUVEC permeability Yun-Chiu Yang 1 Pei-Ni Chen 2 Siou-Yu Wang 3 Chen-Yi Liao 3 Yu-Ying Lin 3 Shih-Rhong Sun 3 Chun-Ling Chiu 4 Yih-Shou Hsieh 2 Jia-Ching Shieh 5 Jinghua Tsai Chang 3,6 Received: 6 May 2014 / Accepted: 29 April 2015 Ó Springer Science+Business Media Dordrecht 2015 Abstract Slit2, a secreted glycoprotein, is down- regulated in many cancers. Slit2/Robo signaling pathway plays an important, but controversial, role in angiogenesis. We identified splicing variants of Slit2 at exon 15, Slit2- WT and Slit2-DE15, with differential effects on prolif- eration and invasive capability of lung cancer cells. The aim of this study was to elucidate the differential roles of these exon 15 splicing variants in angiogenesis. Our results revealed that both Slit2-WT and Slit2-DE15 inhibit moti- lity of human umbilical vein endothelial cells (HUVECs). The conditioned medium (CM) collected from CL1-5/VC or CL1-5/Slit2-WT lung adenocarcinoma cells blocked HUVEC tube formation and angiogenesis on chorioallan- toic membrane (CAM) assay when compared with untreated HUVECs and CAM, respectively. However, CM of CL1-5/Slit2-DE15 restored the quality of tubes and the size of vessels. Although both Slit2-WT and Slit2-DE15 inhibited permeability induced by CM of cancer cells, Sl- it2-DE15 exhibited stronger effect. These results suggested that Slit2-DE15 plays important roles in normalization of blood vessels by enhancing tube quality and tightening endothelial cells, while Slit2-WT only enhances tightening of endothelial cells. It appears that Robo4 is responsible for Slit2 isoform-mediated inhibition of permeability, while neither Robo1 nor Robo4 is required for Slit2-DE15-en- hanced tube quality. The results of this study suggest that Slit2-DE15 splicing form is a promising molecule for normalizing blood vessels around a tumor, which, in turn, may increase efficacy of chemotherapy and radiotherapy. Keywords Slit2 splicing variants Human umbilical vein endothelial cells Angiogenesis Chicken chorioallantoic membrane assay Tube formation Permeability Robo1 Robo4 Introduction Angiogenesis is one of the hallmarks of tumor progression, and development of a vascular supply is a pivotal step in tumor growth [1]. Anti-angiogenic therapy has been de- veloped to starve tumors by destabilizing tumor vascula- ture. However, the current anti-angiogenic agents that target vascular endothelial growth factor (VEGF) and its related pathways, the pro-angiogenic pathways, have yielded lim- ited clinic benefit. Intriguingly, when co-administered with chemotherapy, the result improved overall survival and progression-free survival in some cancers when compared with chemotherapy alone [2, 3]. Explanations for the Yun-Chiu Yang and Pei-Ni Chen have contributed equally to this work. & Jinghua Tsai Chang [email protected] 1 Department of Pulmonary Medicine, Tungs’ Taichung MetroHarbor Hospital, Taichung, Taiwan, ROC 2 Institute of Biochemistry, Microbiology and Immunology, Chung Shan Medical University, Taichung, Taiwan, ROC 3 Institute of Medicine, Chung Shan Medical University, No. 110, Sec. 1, Chien-Kuo N. Rd., Taichung 402, Taiwan, ROC 4 Institute of Medical and Molecular Toxicology, Chung Shan Medical University, Taichung, Taiwan, ROC 5 Department of Biomedical Sciences, Chung Shan Medical University, Taichung, Taiwan, ROC 6 Department of Medical Oncology and Chest Medicine, Chung Shan Medical University Hospital, Taichung, Taiwan, ROC 123 Angiogenesis DOI 10.1007/s10456-015-9467-4

The differential roles of Slit2-exon 15 splicing variants in angiogenesis and HUVEC permeability

Embed Size (px)

Citation preview

ORIGINAL PAPER

The differential roles of Slit2-exon 15 splicing variantsin angiogenesis and HUVEC permeability

Yun-Chiu Yang1 • Pei-Ni Chen2 • Siou-Yu Wang3 • Chen-Yi Liao3 •

Yu-Ying Lin3 • Shih-Rhong Sun3 • Chun-Ling Chiu4 • Yih-Shou Hsieh2 •

Jia-Ching Shieh5 • Jinghua Tsai Chang3,6

Received: 6 May 2014 / Accepted: 29 April 2015

� Springer Science+Business Media Dordrecht 2015

Abstract Slit2, a secreted glycoprotein, is down-

regulated in many cancers. Slit2/Robo signaling pathway

plays an important, but controversial, role in angiogenesis.

We identified splicing variants of Slit2 at exon 15, Slit2-

WT and Slit2-DE15, with differential effects on prolif-

eration and invasive capability of lung cancer cells. The

aim of this study was to elucidate the differential roles of

these exon 15 splicing variants in angiogenesis. Our results

revealed that both Slit2-WT and Slit2-DE15 inhibit moti-

lity of human umbilical vein endothelial cells (HUVECs).

The conditioned medium (CM) collected from CL1-5/VC

or CL1-5/Slit2-WT lung adenocarcinoma cells blocked

HUVEC tube formation and angiogenesis on chorioallan-

toic membrane (CAM) assay when compared with

untreated HUVECs and CAM, respectively. However, CM

of CL1-5/Slit2-DE15 restored the quality of tubes and the

size of vessels. Although both Slit2-WT and Slit2-DE15inhibited permeability induced by CM of cancer cells, Sl-

it2-DE15 exhibited stronger effect. These results suggested

that Slit2-DE15 plays important roles in normalization of

blood vessels by enhancing tube quality and tightening

endothelial cells, while Slit2-WT only enhances tightening

of endothelial cells. It appears that Robo4 is responsible for

Slit2 isoform-mediated inhibition of permeability, while

neither Robo1 nor Robo4 is required for Slit2-DE15-en-hanced tube quality. The results of this study suggest that

Slit2-DE15 splicing form is a promising molecule for

normalizing blood vessels around a tumor, which, in turn,

may increase efficacy of chemotherapy and radiotherapy.

Keywords Slit2 splicing variants � Human umbilical vein

endothelial cells � Angiogenesis � Chicken chorioallantoic

membrane assay � Tube formation � Permeability � Robo1 �Robo4

Introduction

Angiogenesis is one of the hallmarks of tumor progression,

and development of a vascular supply is a pivotal step in

tumor growth [1]. Anti-angiogenic therapy has been de-

veloped to starve tumors by destabilizing tumor vascula-

ture. However, the current anti-angiogenic agents that target

vascular endothelial growth factor (VEGF) and its related

pathways, the pro-angiogenic pathways, have yielded lim-

ited clinic benefit. Intriguingly, when co-administered with

chemotherapy, the result improved overall survival and

progression-free survival in some cancers when compared

with chemotherapy alone [2, 3]. Explanations for the

Yun-Chiu Yang and Pei-Ni Chen have contributed equally to this

work.

& Jinghua Tsai Chang

[email protected]

1 Department of Pulmonary Medicine, Tungs’ Taichung

MetroHarbor Hospital, Taichung, Taiwan, ROC

2 Institute of Biochemistry, Microbiology and Immunology,

Chung Shan Medical University, Taichung, Taiwan, ROC

3 Institute of Medicine, Chung Shan Medical University, No.

110, Sec. 1, Chien-Kuo N. Rd., Taichung 402, Taiwan, ROC

4 Institute of Medical and Molecular Toxicology, Chung Shan

Medical University, Taichung, Taiwan, ROC

5 Department of Biomedical Sciences, Chung Shan Medical

University, Taichung, Taiwan, ROC

6 Department of Medical Oncology and Chest Medicine,

Chung Shan Medical University Hospital, Taichung,

Taiwan, ROC

123

Angiogenesis

DOI 10.1007/s10456-015-9467-4

limited clinical benefit of anti-angiogenic agent treatment

alone are the existence of compensatory pathways that

cause intrinsic resistance to angiogenesis inhibitors and the

adaptive evasion of anti-angiogenic therapy by induction

of alternative angiogenic pathways [4]. Intrinsic and

adaptive pathways that cause resistance to anti-angiogenic

drugs may be contributed by tumor cells and/or their

microenvironment. The present focus of anti-VEGFR

improvement of the efficacy of co-administered che-

motherapy is normalization of tumor vessels by anti-an-

giogenic therapy, which alters tumor microenvironment to

achieve higher oxygenation and lower interstitial fluid

pressure, in turn increasing efficacy of chemotherapy and

radiotherapy [5].

Slit2 is a secreted glycoprotein that is down-regulated in

a variety of cancers and plays an important role in inhibi-

tion of growth, migration, and invasive capability of tumor

cells [6–11]. Emerging evidence has shown that Slit/Robo

signaling is involved in angiogenesis, however, with con-

troversial reports of both pro-angiogenic [12–16] and anti-

angiogenic [17–19] effects. We have identified two exon

15 splicing variants, Slit2-DE15 and Slit-WT. Exon 15 is

located at the end of the leucine-rich repeat 2 (LLR2) of

Slit2, a direct contact site of Slit2 to the first and second

immunoglobulin (IG) domains of Robo1 receptor [20–22].

We have demonstrated that Slit2-DE15 inhibits both

growth and invasive capability of CL1-5 lung cancer cells,

whereas Slit2-WT inhibits invasive capability only [23].

Recent studies have revealed that activation of Robo4 by

Slit2 stabilizes vascular structure [24, 25]. Thus, we were

prompted to evaluate whether exon 15 splicing variants of

Slit2 have different effects on proliferation, angiogenesis,

and permeability of human umbilical vein endothelial cells

(HUVECs).

Materials and methods

Materials

Endothelial cell growth supplement (ECGS) and heparin

were purchased from Sigma (St. Louis, MO, USA).

Medium 199, RPMI, and trypsin–EDTA were obtained

from Gibco (Gibco Invitrogen Corporation, Barcelona,

Spain). Protease inhibitors were supplied by Roche (Ger-

many). VEGF was purchased from Pepro Tech (NJ, USA).

Horseradish peroxidase (HRP) was supplied by Thermo

(IL, USA). Tetramethylbenzidine (TMB) was obtained

from Clinical Science Products Inc. (MA, USA). Matrigel

was supplied by BD Bioscience Pharmingen (MA, USA)

and enhanced chemiluminescence (ECL) kit by Perkin

Elmer (MA, USA).

Cell cultures

HUVECs were obtained from the Bioresource Collection

andResearchCenter (BCRC,Hsinchu, Taiwan) and cultured

on gelatin-coated culture dishes in medium 199 with 10 %

fetal bovine serum, 25 U/ml heparin, 30 lg/ml ECGS,

100 U/ml penicillin, and 0.1 mg/ml streptomycin. Subcul-

tures were performed with trypsin–EDTA. Cells from pas-

sages 5 to 9 were used. Media were refreshed every second

day. The CL1-5 lung adenocarcinoma cell line was kindly

provided by Dr. P.C. Yang (Department of Internal Medi-

cine, National Taiwan University, Taipei, Taiwan). CL1-5/

Slit2-WT or CL1-5/Slit2-DE15 stable clones were generatedby PCR cloning [23]. All cell cultures were maintained at

37 �C in a humidified atmosphere of 5 % CO2.

Preparation of conditioned medium (CM)

CL1-5 exogenous-expressing Slit2-WT-Myc/His or Slit2-

DE15-Myc/His cells were cultured in RPMI plus 10 %

FBS until confluence, then replaced with 7 ml of M199

medium. Incubation was carried out for 24 h. This was

followed by removal of the conditioned medium (CM) and

centrifugation at 1000 rpm for 5 min. Aspirate CM was

mixed with an equal volume of fresh M199 medium plus 1

or 10 % FBS for treatment. To evaluate the levels of se-

creted Slit2 isoforms, 50 ll of undiluted CM were used for

western blot analysis with anti-Myc antibody.

Migration assay

Inoculated 5 9 105HUVECswere plated onto 0.2 %gelatin-

coated 10-cm dish and incubated for 16 h. The next day,

treated HUVECs were placed in CM for 24 h and then tryp-

sinized, counted, and plated (3 9 104 cells) onto 24-well

Boyden chamber. The lower chamber was filled with M199

medium containing 30 ng/ml VEGF. After 12-h incubation at

37 �C, themembranewas fixedwith 100 % ice coldmethanol

for 30 min and stained with 20 % Giemsa for counting.

Matrigel tube formation assay

HUVECs were pretreated with/without CM (with 10 %

FBS) for 24 h. Ibidi u-slide plates (Applied Biophysics,

NY, USA) were coated with 10 ll Matrigel (10 mg/ml)

and incubated at 37 �C for 1 h. The pretreated HUVECs

were suspended in fresh CM (with 10 % FBS and ECGS)

and plated onto a layer of Matrigel at a density of

5.5 9 103 cells/well. The plates were then incubated at

37 �C for an additional 6 h, and capillary-like tube for-

mation was observed under microscope. For Slit2 deple-

tion, CM/Slit2-DE15 was incubated with 0.25 lg/ml of

Angiogenesis

123

anti-IgG (A9044, Sigma, St Louis, USA) and anti-Myc (05-

419,Millipore, CA,USA) at 4 �C for 16 h followed by protein

Gprecipitation.After centrifugation, theCMwasused for tube

formation assay. For blocking of Robo receptors, 0.25 lg/ml

of anti-IgG, anti-Robo1 (ab7279, abcam, UK), and/or anti-

Robo4 (ab10547, abcam, UK) were incubated with HUVECs

at 37 �C for 16 h followed by CM treatment for 24 h prior to

the tube formation assay. For RNAi, 50 pmol of si-Robo1

(50-GCAGACACGUGGCCUAAUATT-30), 100 pmol of

si-Robo4 (50-GCUUCUGGCUGUGCGAAUUTT-30), and

si-NC (50-UUCUCCGAACGUGUCACGUTT-30) were

transfected into HUVECs and incubated at 37 �C for 16 h,

respectively. Cells were then treated with CM of Slit2-DE15for 24 h prior to the tube formation assay. The quality of tube

formation was determined by tube length, cell-covered area

(capillary surface area), branching points, and loops using

WimTube of Wimasis image analysis system (Wimasis

GmbH, Munich, Germany).

Chicken chorioallantoic membrane (CAM) assay

Fertilized chicken eggs were transferred to an egg incu-

bator and maintained at 37 �C and 50 % humidity for

8 days. To separate the chicken chorioallantoic membrane

(CAM) from the shell membrane, two small holes were

drilled into the shell, one at the blunt end of the egg where

the air sac is located and the other at 90� halfway down the

length of the egg. Gentle suction was applied at the hole

located at the blunt end of the egg to create a false air sac

directly over the CAM. Then, a 1 cm2 window was re-

moved from the eggshell immediately over the second

hole. Holes were made in the CAM to facilitate penetration

of the 200 KDa Slit2 protein. CM of CL1-5/VC, CL1-5/

Slit2-WT or CL1-5/Slit2-DE15, or RPMI (control group)

was placed on the CAM, and the embryos were incubated

for an additional 48 h. The neovascular zones under the

disks were photographed, and the ratios of the diameters of

the peripheral blood vessels close to the CAM relative to

those far below the CAM were determined by Wimasis

software (Wimasis GmbH, Munich, Germany).

Permeability assay

Coating of 1.0-lm Transwell membrane (PIHT 30R 48;

Millipore, MA, USA) was carried out with 0.1 % gelatin at

37 �C for 1 h. Aspirated gelatin solution from the mem-

brane and 4 9 104 HUVECs in 200 ll of M199 with 10 %

FBS were added to the upper chamber. The same medium

was added to the lower chamber until the surface of the

medium was the same as in the upper chamber, followed

by incubation at 37 �C, 5 % CO2 for 3 days until cells

formed a tight monolayer. The insert of the transwell

membrane (upper chamber) was moved to a new well, and

the cells were treated with/without CM (1 % FBS), fol-

lowed by incubation at 37 �C, 5 % CO2 for 16 h. The next

day, the medium in the lower chamber was replaced with

new CM with 100 pg/ml of VEGF165 for 3.5 h. Then, the

medium in the upper chamber was replaced with 100 ll ofnew medium containing 10 lg/ml of HRP followed by

incubation for 1 h. The lower chamber medium was moved

to a new tube and mixed well. Then, 10 ll of the medium

were diluted to 50-fold and treated with TMB for 15 min.

The reaction was terminated with 50 ll of 2 N H2SO4. The

permeability of HRP in HUVECs was determined at

450 nm absorbance. Standard curve of HRP was estab-

lished with 1, 2, 4, 8, 16, and 32 ll of 100 ng/ml HRP in a

volume of 50 ll. For antibody treatment, HUVECs were

treated with 0.25 lg/ml of anti-IgG, anti-Robo1 or anti-

Robo4 after formation of tight monolayer, followed by CM

treatment for 24 h prior to the permeability assay. For

RNAi, 100 pmol of si-Robo1, si-Robo4, or si-NC was

transfected into HUVECs and incubated at 37 �C for 24 h.

Cells were seeded onto transwell membrane for perme-

ability assay after CM treatment.

Membrane protein extraction

HUVECs (1.5 9 106 cells) were plated onto 10-cm dish

and incubated at 37 �C, 5 % CO2 for 16 h. Medium was

replaced with fresh M199 medium with/without CM (with

10 % FBS) for 24 h. Then, the cells were treated with/

without VEGF (30 ng/ml) for 4 h. After removing the

medium, cells were washed with 1XPBS and frozen at

-80 �C for 1 h. Then, 300 ll of Tris Buffer (50 mM Tris–

Cl (pH 7.5), 4 mM EDTA, 2 mM EGTA, 19 proteinase

inhibitors) were added to the thawing dish with mixing and

scraping of the cells. The cells were subjected to another

freeze and thaw cycle, followed by observation under a

microscope to examine the stage of cell lysis (avoiding

small cell debris stage). Cells were centrifuged at

13,000 rpm at 4 �C for 30 min. The supernatant was col-

lected for cytosolic fraction, and the pellet was treated with

50 ll of RIPA buffer [50 mM Tris (pH 7.4), 150 mM

NaCl, 1 % NP-40, 0.25 % sodium deoxycholate, 2 mM

EDTA, 1 lg/ml aprotinin, 1 lg/ml leupeptin, 1 mM PMSF

and 2 lg/ml pipstatin A] and sonicated. The extract was

then centrifuged, and the supernatant was found to contain

proteins in membrane fraction.

Western blot analysis

CM (50 ll) was separated onto sodium dodecyl sulfate–

polyacrylamide gels (SDS-PAGE) and then transferred

from the gels onto polyvinylidene fluoride membranes

(Perkin Elmer). After blocking, the membranes were re-

acted with antibody at 4 �C overnight, followed by

Angiogenesis

123

incubation with horseradish peroxidase-conjugated sec-

ondary antibody for 1 h. The blots were visualized using an

ECL kit (Perkin Elmer). Then, 10 lg of membrane proteins

were loaded for western blot analysis. Antibodies used in

the membrane protein analyses were anti-Myc (Millipore,

Temecula, CA, USA), anti-VE-cadherin (Invitrogen,

Carlsbad, CA, USA), anti-Robo1 (ab7279, MA, USA) anti-

Robo4 (abcam, ab10547, MA, USA), and anti-ATPase

(Bioss Inc., Woburn, MA, USA).

Statistical analysis

All HUVEC assays were carried out in at least three in-

dependent experiments with three repeats for each ex-

periment with values presented as mean ± SD. Statistical

analysis was performed using the SPSS statistical software

program (version 13, SPSS, Inc.). One-way analysis of

variance (ANOVA) was used to analyze the significance

among groups followed by Scheffe test or Tukey’s HSD

test when appropriate for comparisons between two groups.

p\ 0.05 was considered statistically significant.

Results

The effects of Slit2-WT and Slit2-DE15 splicing

forms on growth of HUVECs

Although Slit2 has been shown to modulate angiogenesis,

its role in angiogenesis is still controversial. Our previous

study revealed that both Slit2-DE15 and Slit2-WT inhibit

invasive capability of lung cancer cells. However, only

Slit2-DE15, but not Slit2-WT, has the ability to inhibit

growth of lung cancer cells [23]. To assess whether dif-

ferent Slit2-exon 15 splice forms affect angiogenesis, we

first investigated the effects of Slit2-WT and Slit2-DE15 ongrowth of HUVECs. HUVECs treated with CM collected

from CL1-5 cells stably expressing Slit2-DE15 (CL1-5/

Slit2-DE15) or Slit2-WT (CL1-5/Slit2-WT) demonstrated

slightly slower growth when compared with those treated

with CM from vector control (CL1-5/VC) and untreated

cells (Fig. 1). However, no differences in growth inhibitive

effects were observed between Slit2-DE15 and Slit2-WT in

HUVECs, which is in contrast to our previous findings for

CL1-5 lung cancer cells [23].

Both Slit2-WT and Slit2-DE15 inhibit motility

of HUVECs

Next, we examined the effects of Slit2-WT and Slit2-DE15on the motility of HUVECs. Upon induction by VEGF165in bottom chamber of Boyden chamber, the motility of

HUVECs was so great that we were unable to differentiate

the effects of Slit2 on cell migration. Therefore, HUVECs

were pretreated with CM collected from CL1-5/Slit2-

DE15, CL1-5/Slit2-WT, or CL1-5/VC for 24 h prior to the

motility assay. This was followed by treatment with the

same CM during assay. In comparison with CL1-5/VC

CM, both Slit2-DE15 and Slit2-WT CM greatly inhibited

VEGF165 induced cell motility of HUVECs (Fig. 2c, d).

The effects of Slit2-WT and Slit2-DE15 on HUVEC

tube formation and angiogenesis on CAM assay

To examine whether Slit2 splicing forms differentially af-

fect tube formation by HUVECs, HUVECs were pretreated

with/without CM from CL1-5/Slit2-WT, CL1-5/Slit-DE15,or CL1-5/VC prior to the assay. Without CM, HUVECs

formed high-quality tubes within 6 h. Although the speed

of tube formation of HUVECs treated with VC CM was

similar to that of untreated cells, the tubes were thinner and

highly disconnected (Fig. 3a). Similar to VC CM, CL1-5/

Slit2-WT CM negatively affected tube formation by

HUVECs. Interestingly, following pretreatment with CL1-

5/Slit2-DE15 CM, tube size of HUVECs was restored to

that of untreated cells (Fig. 3a). Following quantification,

HUVECs treated with CM of CL1-5/VC and/Slit2-WT had

smaller cell-covered area, shorter tube length, fewer

branching points, and fewer loops when compared with

Fig. 1 Effects of Slit2-WT and Slit2-DE15 on growth of HUVECs.

a Expressions of Slit2-WT and Slit2-DE15 proteins in CM. b The

expression levels of Slit2-WT and Slit2-DE15 in 50 ll CM were

determined by densitometry. c Both Slit2-WT and Slit2-DE15 CM

slightly reduced growth of HUVECs

Angiogenesis

123

HUVECs treated with CM of CL1-5/Slit2-DE15 or un-

treated HUVECs (Fig. 3b). These results suggested that

Slit2-DE15 enhances the quality of tube formation by

HUVECs, while Slit2-WT does not. Next, we used chicken

embryos to examine the effect of CM containing various

Slit2-exon 15 splicing forms on angiogenesis via CAM

assay. Similar to the results of tube formation assay, CM of

CL1-5/VC greatly reduced the diameter of peripheral

vessels when compared with the control group (RPMI

medium). CM of CL1-5/Slit2-DE15 restored the diameter

of peripheral vessels, while CM of CL1-5/Slit2-WT did not

(Fig. 3c, d). These results suggested that the function of

Slit2-DE15 in angiogenesis is conserved between humans

and chickens. To demonstrate that Slit2-DE15, not Slit2-DE15-induced content in CM, is responsible for restoring

tube quality, Slit2-DE15 was depleted in CM prior to the

tube formation assay. Slit2-DE15-depleted CM resulted in

poor quality tubes in HUVECs when compared with anti-

IgG-treated CM (Fig. 3e), indicating that Slit2-DE15 di-

rectly enhances tube quality.

The effects of Slit2 splicing forms on permeability

of HUVECs

Slit2/Robo4 signaling promotes vascular stability and im-

pedes pathological angiogenesis by inhibiting endothelial

hyperpermeability [24, 25]. Since CM containing Slit2-

DE15 enhanced quality of tube formation while Slit2-WT

did not, we tested whether Slit2 splicing forms have dif-

ferent effects on the permeability of HUVECs. Monolay-

ered HUVECs were incubated with CM containing

different Slit2 splicing forms for 16 h with permeability

induced by VEGF165 for 3.5 h. The permeability was de-

termined by the diffusion of HRP from upper chamber to

lower chamber on transwell membrane. The results showed

that both CM of CL1-5/Slit2-WT and CM of CL1-5/Slit2-

DE15 reduce cell permeability of HUVECs when com-

pared with CM of CL1-5/VC. However, the reduction in

cell permeability was more significant with Slit2-DE15(Fig. 4).

Slit2-exon 15 splicing form enhances membrane

expression of Robo4

Vascular endothelial cadherin (VE-cadherin) is an en-

dothelial-specific adhesion molecule at adherens junctions

that modulates cell-to-cell interaction [26]. Disruption of

VE-cadherin membrane localization at cell junctions en-

hances permeability of HUVEC monolayer [27]. To de-

termine if CL1-5/Slit2-WT and CL1-5/Slit2-DE15 CM

restore CM-induced permeability via modulation of mem-

brane localization of VE-cadherin, the membrane fractions

of VE-cadherin were analyzed. The results showed that

membrane fractions of VE-cadherin were not altered under

treatment with different CM (Fig. 5a, b). It has been re-

ported that Slit2/Robo4 signaling reduces VEGF165-in-

duced retinal hyperpermeability [25]. Since HUVECs

express Robo1 and Robo4 receptors, we detected mem-

brane localization of Robo1 and Robo4 in HUVECs treated

with various CM. Interestingly, HUVECs treated with CM

of CL1-5/Slit2-WT or CM of CL1-5/Slit2-DE15 had highermembrane fractions of Robo4 than HUVECs treated with

CM of CL1-5/VC (Fig. 5c, d). However, the membrane

fraction of Robo1 remained unchanged (Fig. 5e, f). Thus,

binding of Slit2 may stabilize Robo4 on the membrane and

reduce permeability of HUVECs.

Fig. 2 Effects of Slit2-WT and Slit2-DE15 on migration of

HUVECs. a, b Expressions of Slit2 splicing forms in CM. The

designation ‘‘p’’ refers to positive control. c, d VEGF in the lower

transwell chamber greatly induced migration of HUVECs. Both Slit2-

WT and Slit2-DE15 CM strongly inhibited VEGF-induced cell

migration. One-way ANOVA and Tukey’s HSD test were used for

analysis F(4,10) = 399.59, p\ 0.001. ***p\ 0.001

Angiogenesis

123

Fig. 3 Effects of Slit2-WT and Slit2-DE15 on HUVEC tube

formation and angiogenesis on CAM assay. a Untreated HUVECs

formed good quality tubes on Matrigel. CM of CL1-5/VC and CM of

CL1-5/Slit2-WT inhibited tube formation by HUVECs when com-

pared with untreated HUVECs. However, CM of Slit2-DE15 greatly

enhanced the quality of tubes formed by HUVECs. b The cell-

covered area, tube length, branching points, and loops of tube

formation were determined using WimTube of Wimasis image

analysis system (n = 4). One-way ANOVA and Tukey’s HSD test

were used for analysis; cell-covered area, F(3,12) = 19.329,

p\ 0.001; tube length, F(3,12) = 17.994, p\ 0.001; branching

points, F(3,12) = 13.474, p\ 0.001; and loops, F(3,12) = 9.49,

p\ 0.01. *p\ 0.05; **p\ 0.01; ***p\ 0.001. c The results of

CAM assay also showed that CM of CL1-5/VC decreases the size of

peripheral vessels when compared with control group (RPMI

medium). CM of CL1-5/Slit2-DE15 restored the size of peripheral

vessels, while CM of CL1-5/Slit2-WT did not. Peripheral vessels are

designated by arrows, and underlying vessels are designated by

arrowheads. d The size of vessels was determined by Wimasis

software. The size of peripheral vessels is shown relative to

underlying vessels. One-way ANOVA and Scheffe test were used

for analysis F(3,8) = 18.761, p\ 0.001. **p\ 0.01. e Slit2-DE15was depleted in CM of Slit2-DE15 by anti-Myc antibody followed by

protein G precipitation. Slit2-DE15-depleted CM resulted in poor tube

formation when compared with anti-IgG-treated CM

Angiogenesis

123

Neither Robo1 nor Robo4 is involved in Slit2-DE15-mediated tube formation

It has been shown that Slit2/Robo4 pathway inhibits

VEGF-induced permeability. It is important to determine

whether Robo4 is also involved in Slit2-DE15-mediated

reduction in permeability. When the function of Robo4 was

blocked by anti-Robo4, the permeability of HUVECs in-

creased following treatment with CM of CL1-5/Slit2-WT

or CM of CL1-5/Slit2-DE15 in comparison with IgG

control (Fig. 6a, b). Neutralization of Robo1 did not in-

crease permeability of HUVECs. Similar results were ob-

tained when the expression of Robo4 was interfered with

by si-Robo4 (Fig. 6c, d), suggesting that Robo4 but not

Robo1 is required for Slit2-DE15-mediated inhibition of

permeability. To explore the roles of Robo1 and Robo4 in

Slit2-DE15-mediated tube formation, we neutralized

Robo1 or Robo4 in HUVECs, followed by treatment with

CM of CL1-5/Slit2-DE15. In comparison with anti-IgG

treatment, blocking of Robo1 or Robo4 alone did not re-

duce tube formation of HUVECs treated with CM of Slit2-

DE15. It is possible that Robo1 and Robo4 compensate for

each other in Slit2-DE15-mediated tube formation or that

these antibodies are unable to block the function of Robo

receptor in angiogenesis. Double blocking with anti-Robo1

and anti-Robo4 did not affect Slit2-DE15-mediated tube

formation (Fig. 7a). Similar results were obtained in cells

with si-Robo1 or si-Robo4 single knockdown and si-

Robo1/si-Robo4 double knockdown (Fig. 7b). These re-

sults are important because they point to the possibility of

yet to be elucidated Slit2 signaling pathway(s) in

angiogenesis.

Discussion

Tumor microenvironment plays an important role in

regulating cancer biology, including angiogenesis [28, 29].

The growth of new blood vessels is tightly regulated by

activators of angiogenesis [e.g., VEGF, fibroblast growth

factor 1 (FGF-1), and angiopoietin-2 (Ang-2)] and in-

hibitors of angiogenesis [e.g., thrombospondin (TSP-1) and

angiopoietin-1(Ang-1)] [30]. The net balance of pro- and

anti-angiogenic factors can be disturbed by the tumor and

its microenvironment. Moreover, newly formed blood

vessels in tumors are leaky, leading to high interstitial fluid

pressure, which deteriorates blood circulation and reduces

oxygenation of the tumors. Functionally impaired blood

vessels can lead to heterogeneous or inadequate drug

penetration and less effective radiation therapy in hypoxic

tumors [5].

Slit2 is a 200-kDa secreted glycoprotein that is highly

repressed in lung tumors. We identified two alternative

splicing forms of Slit2. Slit2-DE15 inhibits both growth

and invasive capability of CL1-5 lung cancer cells, while

Slit2-WT inhibits invasive capability only. Interestingly,

both lung tumor and adjacent normal lung tissues almost

exclusively express the Slit2-WT form, while lung tissues

from non-tumor patients and normal bronchial epithelial

cells (BEAS-2B) express high ratios of the Slit2-DE15form [23]. It is known that Slit/Robo signaling plays both

anti- and pro-angiogenic roles in angiogenesis. However,

how Slit2/Robo affects angiogenesis is still unclear. Robo4

is expressed in all endothelial cells, while Robo1 is ex-

pressed quite diversely among different endothelial cells

[31, 32]. Slit2 binds to Robo1 monomer or Robo1/Robo4

heterodimer promotes endothelial mobility [32], while Sl-

it2/Robo4 signaling counteracts VEGF-induced angio-

genesis and maintains vascular stabilization [19, 24, 25].

Interaction between LRR2 (leucine-rich repeat) domain of

Slit2 and Ig1 domain of Robo1 has been demonstrated in

crystalline structures [33]. Interaction between Slit2 and

Robo4 has been shown in co-immunoprecipitation [19, 34].

However, no direct interaction has been observed on Bia-

core assay [35]. Of note is that all Robo receptors except

for Robo4 contain the Ig domain, which binds to Slit2 [33].

Thus, it is possible that a co-receptor, such as Robo1,

syndecans [36], or heparan sulfate [37], is required for

Slit2/Robo4 interaction. A recent genetic study has clearly

Fig. 4 Effects of Slit2-WT and Slit2-DE15 on HUVEC permeability.

CM of CL1-5/VC enhanced permeability of HUVECs when com-

pared with VEGF165 treatment alone. Both CL1-5/WT and CL1-5/

Slit2-DE15 inhibited CM-enhanced permeability of HUVECs. The

designation ‘‘p’’ refers to positive control. One-way ANOVA and

Tukey’s HSD test were used for analysis F(4,10) = 11.71, p\ 0.001.

*p\ 0.05; **p\ 0.01

Angiogenesis

123

shown that Slit2 is involved in Robo4-mediated down-

regulation of VEGF angiogenesis in breast [38]. Although

Slit2 has been shown to stabilize blood vessels through

Robo4 receptor [24, 25], it is not known whether various

Slit2-exon 15 splicing forms differentially affect vessel

stability.

The integrity of vessels around a tumor is poor, leading

to high permeability. VEGF secreted by cancer cells is

largely responsible for this increased permeability of blood

vessels. Slit2-Robo4 signaling has the ability to stabilize

VEGF-treated blood vessels [24]. Consistent with those

findings, we observed that HUVECs treated with CM of

CL1-5 lung cancer cells are more permeable. Furthermore,

in HUVECs and CAM treated with CM of CL1-5/VC lung

cancer cells, there was formation of atrophic tubes and thin

blood vessels, indicating that CM of lung cancer cells in-

terferes with blood vessel quality. In HUVECs and CAM

treated with CM of CL1-5/Slit2-WT cells, there was also

formation of poor quality tubes and small vessels. How-

ever, in HUVECs and CAM treated with CM of CL1-5/

Fig. 5 Effects of various CM on membrane-associated VE-cadherin,

Robo1, and Robo4 expressions in HUVECs. a, b Membrane-associ-

ated VE-cadherin expression and cytosolic fraction of VE-cadherin

were not affected by various CM treatments. c, d CM of CL1-5/Slit2-

WT and CL1-5/Slit2-DE15 increased the expression levels of

membrane-associated Robo4 in HUVECs when compared with

treatment with CM of CL1-5/VC. One-way ANOVA and Tukey’s

HSD test were used for analysis F(4,10) = 11.71, p\ 0.001.

*p\ 0.05. e, f Membrane-associated Robo1 level was not affected

by various CM treatments

Angiogenesis

123

Slit2-DE15, thick tubes and normal-sized blood vessels

formed, suggesting that Slit2-DE15, but not Slit2-WT, has

the ability to overcome the effects of CL1-5 CM-induced

reduction in tube quality and blood vessel size (Fig. 8a).

Although both Slit2-WT CM and Slit2-DE15 CM can re-

duce CM-induced HUVEC permeability, Slit2-DE15 has

greater effect than Slit2-WT. Robo4 is required for Slit2

isoform-mediated reduction in HUVEC permeability,

while neither Robo1 nor Robo4 participates in Slit2-DE15-mediated formation of good quality tubes (Fig. 8b). These

results suggested that Slit2-exon 15 splicing forms act

through different pathways to affect tube formation and

permeability. Slit2-DE15 enhanced tube formation quality,

as well as tightened cell junctions, while Slit2-WT had no

effect on tube formation quality but had some effect on

tightening of cell junctions. These results suggested that

Slit2-DE15 plays a more important role in the normaliza-

tion of blood vessels than Slit2-WT.

It would be interesting to identify factor(s) in CL1-5 CM

that interfere(s) with tube formation and permeability of

HUVECs and to elucidate how Slit2-DE15 overcomes this/

these factor(s). Although reduced membrane-associated

VE-cadherin can increase the permeability of endothelial

cells [27], we did not observe altered membrane fraction of

VE-cadherin in HUVECs under various CM treatments.

Thus, VE-cadherin may not be involved in Slit2 CM-re-

duced cell permeability. Membrane-associated Robo4 sig-

nificantly increased in HUVECs treated with CM of Slit2-

WT or CM of Slit2-DE15. It is possible that Slit2/Robo4

interaction stabilizes membrane-associated Robo4 and

Fig. 6 Roles of Robo1 and Robo4 in Slit2-DE15-mediated inhibition

of permeability. a, b Blocking of Robo4 function by anti-Robo4

abolished both Slit2-WT and Slit2-DE15-mediated inhibition of

HUVEC permeability. One-way ANOVA and Tukey’s HSD test were

used for analysis. Slit2-DE15: F(3,8) = 313.19, p\ 0.001; Slit2-WT:

F(3,8) = 768.305, p\ 0.001. *p\ 0.05; **p\ 0.01; ***p\ 0.001.

c, d Knockdown of the expression of Robo4 also increased

permeability of HUVECs treated with both Slit2-isoforms when

compared with si-NC. One-way ANOVA and Tukey’s HSD test were

used for analysis. Slit2-DE15: F(2,4) = 1801.16, p\ 0.001; Slit2-

WT: F(2,6) = 657.141, p\ 0.001. *p\ 0.05; **p\ 0.01;

***p\ 0.001. e The effectiveness of si-Robo1 and si-Robo4

Angiogenesis

123

Angiogenesis

123

reduces permeability in HUVECs. However, increased

membrane-associated Robo4 level, reduced growth of

HUVECs, and inhibited HUVEC motility could not explain

the ability of Slit2-DE15 CM, but not Slit2-WT CM, to

enhance the quality of tube formation. In addition, antibody

neutralization and RNAi studies have clearly shown that

neither Robo1 nor Robo4 participates in Slit2-DE15-me-

diated tube formation. The results of the present study re-

vealed the presence of unknown receptor(s) of Slit2-DE15that transduce(s) the signal in the regulation of

angiogenesis.

Although Slit2-WT is the predominant splicing form

expressed in tumor microenvironment, normal human

bronchial epithelium (BEAS-2B) and lung specimens of

pneumothorax patients preferentially express the Slit2-

DE15 splicing form [23]. This implies that non-tumor lung

cells preferentially express Slit2-DE15, while lung tumors

and their normal lung counterparts express Slit2-WT. Since

Slit2-DE15, but not Slit2-WT, inhibits growth of tumor

cells, we suspect that during tumor development, the ex-

pression of Slit2 in tumor microenvironment switches from

Slit2-DE15 to Slit2-WT. Subsequently, the tumor mi-

croenvironment not only induces the loss of capability to

inhibit growth of tumor cells but also reduces blood vessel

quality in tumors. It is increasingly recognized that ab-

normal blood vessels not only create hypoxic and acidic

environment that enhances tumor malignancy, but also

impair treatments of tumor [39]. Thus, identifying the

pathway involved in Slit2-DE15-mediated normalization of

blood vessels would provide the opportunity to control the

growth of tumor while bypassing Slit2-DE15 to increase

treatment efficacies.

Acknowledgments Special thanks are owed to Dr. Pinpin Lin, Dr.

Jiunn-Liang Ko, Dr. Gwo-Tarng Sheu, Dr. Wen-Jun Wu, and Dr. Hui

Lee for their expert consultation and advice and Jingyao Chang for

financial support for this work. This project was supported by Na-

tional Science Council Grant NSC 102-2320-B-040-010-MY2 (Tai-

wan, ROC) and an inter-institutional Grant CSMU-TTM-099004

(Taiwan, ROC) to Jinghua Tsai Chang. This research was also sup-

ported by National Science Council Grant NSC 102-2320-B040-006-

MY3 to Pei-Ni Chen.

Conflict of interest The authors declare that they have no conflicts

of interest.

Ethical standard All procedures performed in this study comply

with Taiwan legislation and were approved by the Institutional and

Bioethical Use Committees (Chung Shan Medical University).

References

1. Hanahan D, Weinberg RA (2011) Hallmarks of cancer: the next

generation. Cell 144(5):646–674

2. Jayson GC, Hicklin DJ, Ellis LM (2012) Antiangiogenic ther-

apy—evolving view based on clinical trial results. Nat Rev Clin

Oncol 9(5):297–303

3. Carmeliet P, Jain RK (2011) Molecular mechanisms and clinical

applications of angiogenesis. Nature 473(7347):298–307

4. Bergers G, Hanahan D (2008) Modes of resistance to anti-an-

giogenic therapy. Nat Rev 8(8):592–603

5. Jain RK (2005) Normalization of tumor vasculature: an emerging

concept in antiangiogenic therapy. Science 307(5706):58–62

6. Dallol A, Da Silva NF, Viacava P, Minna JD, Bieche I, Maher

ER, Latif F (2002) SLIT2, a human homologue of the Drosophila

Slit2 gene, has tumor suppressor activity and is frequently inac-

tivated in lung and breast cancers. Cancer Res 62(20):5874–5880

7. Dallol A, Morton D, Maher ER, Latif F (2003) SLIT2 axon

guidance molecule is frequently inactivated in colorectal cancer

and suppresses growth of colorectal carcinoma cells. Cancer Res

63(5):1054–1058

8. Kim HK, Zhang H, Li H, Wu TT, Swisher S, He D, Wu L, Xu J,

Elmets CA, Athar M, Xu XC, Xu H (2008) Slit2 inhibits growth

and metastasis of fibrosarcoma and squamous cell carcinoma.

Neoplasia 10(12):1411–1420

bFig. 7 Roles of Robo1 and Robo4 in Slit2 splicing form-mediated

tube formation. a Functional blocking of Robo1 and/or Robo4 with

anti-Robo1 and/or anti-Robo4. There were no differences in tube

formation between neutralized Robo receptors and anti-IgG control.

b Single knockdown of the expression of Robo1 or Robo4 and double

knockdown of the expressions of Robo1 and Robo4 by RNAi

revealed that single or double knockdown of Robo receptors does not

affect Slit2-DE15-mediated tube formation when compared with anti-

IgG control

Fig. 8 a Schematic diagram of the differential effects of Slit2-WT

and Slit2-DE15 on the inhibition of lung cancer cell-induced

permeability and tube/vessel size. b Both Slit2-WT and Slit2-DE15decreased CM-induced permeability through Robo4, while neither

Robo1 nor Robo4 was required for Slit2-DE15-mediated normaliza-

tion of tube size

Angiogenesis

123

9. Mertsch S, Schmitz N, Jeibmann A, Geng JG, Paulus W, Senner

V (2008) Slit2 involvement in glioma cell migration is mediated

by Robo1 receptor. J Neurooncol 87(1):1–7

10. Werbowetski-Ogilvie TE, Seyed Sadr M, Jabado N, Angers-

Loustau A, Agar NY, Wu J, Bjerkvig R, Antel JP, Faury D, Rao

Y, Del Maestro RF (2006) Inhibition of medulloblastoma cell

invasion by slit. Oncogene 25(37):5103–5112

11. Yiin JJ, Hu B, Jarzynka MJ, Feng H, Liu KW, Wu JY, Ma HI,

Cheng SY (2009) Slit2 inhibits glioma cell invasion in the brain

by suppression of Cdc42 activity. Neuro-oncology 11(6):779–789

12. Kaur S, Castellone MD, Bedell VM, Konar M, Gutkind JS,

Ramchandran R (2006) Robo4 signaling in endothelial cells

implies attraction guidance mechanisms. J Biol Chem

281(16):11347–11356

13. Kaur S, Samant GV, Pramanik K, Loscombe PW, Pendrak ML,

Roberts DD, Ramchandran R (2008) Silencing of directional

migration in roundabout4 knockdown endothelial cells. BMC

Cell Biol 9:61

14. Sheldon H, Andre M, Legg JA, Heal P, Herbert JM, Sainson R,

Sharma AS, Kitajewski JK, Heath VL, Bicknell R (2009) Active

involvement of Robo1 and Robo4 in filopodia formation and

endothelial cell motility mediated via WASP and other actin

nucleation-promoting factors. Faseb J 23(2):513–522

15. Urbich C, Rossig L, Kaluza D, Potente M, Boeckel JN, Knau A,

Diehl F, Geng JG, Hofmann WK, Zeiher AM, Dimmeler S (2009)

HDAC5 is a repressor of angiogenesis and determines the an-

giogenic gene expression pattern of endothelial cells. Blood

113(22):5669–5679

16. Yang XM, Han HX, Sui F, Dai YM, Chen M, Geng JG (2010)

Slit–Robo signaling mediates lymphangiogenesis and promotes

tumor lymphatic metastasis. Biochem Biophys Res Commun

396(2):571–577

17. Han X, Zhang MC (2010) Potential anti-angiogenic role of Slit2

in corneal neovascularization. Exp Eye Res 90(6):742–749

18. Liu D, Hou J, Hu X, Wang X, Xiao Y, Mou Y, De Leon H (2006)

Neuronal chemorepellent Slit2 inhibits vascular smooth muscle

cell migration by suppressing small GTPase Rac1 activation. Circ

Res 98(4):480–489

19. Park KW, Morrison CM, Sorensen LK, Jones CA, Rao Y, Chien

CB, Wu JY, Urness LD, Li DY (2003) Robo4 is a vascular-

specific receptor that inhibits endothelial migration. Dev Biol

261(1):251–267

20. Howitt JA, Clout NJ, Hohenester E (2004) Binding site for Robo

receptors revealed by dissection of the leucine-rich repeat region

of Slit. EMBO J 23(22):4406–4412

21. Liu Z, Patel K, Schmidt H, Andrews W, Pini A, Sundaresan V

(2004) Extracellular Ig domains 1 and 2 of Robo are important

for ligand (Slit) binding. Mol Cell Neurosci 26(2):232–240

22. Morlot C, Hemrika W, Romijn RA, Gros P, Cusack S, McCarthy

AA (2007) Production of Slit2 LRR domains in mammalian cells

for structural studies and the structure of human Slit2 domain 3.

Acta Crystallogr A 63(Pt 9):961–968

23. Lin YY, Yang CH, Sheu GT, Huang CY, Wu YC, Chuang SM,

Fann MJ, Chang H, Lee H, Chang JT (2011) A novel exon

15-deleted, splicing variant of Slit2 shows potential for growth

inhibition in addition to invasion inhibition in lung cancer.

Cancer 117(15):3404–3415

24. Jones CA, London NR, Chen H, Park KW, Sauvaget D, Stockton

RA, Wythe JD, Suh W, Larrieu-Lahargue F, Mukouyama YS,

Lindblom P, Seth P, Frias A, Nishiya N, Ginsberg MH, Gerhardt

H, Zhang K, Li DY (2008) Robo4 stabilizes the vascular network

by inhibiting pathologic angiogenesis and endothelial hyperper-

meability. Nat Med 14(4):448–453

25. Jones CA, Nishiya N, London NR, Zhu W, Sorensen LK, Chan

AC, Lim CJ, Chen H, Zhang Q, Schultz PG, Hayallah AM,

Thomas KR, Famulok M, Zhang K, Ginsberg MH, Li DY (2009)

Slit2-Robo4 signalling promotes vascular stability by blocking

Arf6 activity. Nat Cell Biol 11(11):1325–1331

26. Dejana E (2004) Endothelial cell-cell junctions: happy together.

Nat Rev Mol Cell Biol 5(4):261–270. doi:10.1038/nrm1357

27. Alghisi GC, Ponsonnet L, Ruegg C (2009) The integrin an-

tagonist cilengitide activates alphaVbeta3, disrupts VE-cadherin

localization at cell junctions and enhances permeability in en-

dothelial cells. PLoS One 4(2):e4449. doi:10.1371/journal.pone.

0004449

28. Bissell MJ, Radisky D (2001) Putting tumours in context. Nat

Rev 1(1):46–54

29. Wiseman BS, Werb Z (2002) Stromal effects on mammary gland

development and breast cancer. Science 296(5570):1046–1049

30. Goel S, Duda DG, Xu L, Munn LL, Boucher Y, Fukumura D,

Jain RK (2011) Normalization of the vasculature for treatment of

cancer and other diseases. Physiol Rev 91(3):1071–1121

31. BallardMS,HinckL (2012)A roundaboutway to cancer.AdvCancer

Res 114:187–235. doi:10.1016/B978-0-12-386503-8.00005-3

32. Wang B, Xiao Y, Ding BB, Zhang N, Yuan X, Gui L, Qian KX,

Duan S, Chen Z, Rao Y, Geng JG (2003) Induction of tumor

angiogenesis by Slit–Robo signaling and inhibition of cancer

growth by blocking Robo activity. Cancer Cell 4(1):19–29

33. Morlot C, Thielens NM, Ravelli RB, HemrikaW, Romijn RA, Gros

P, Cusack S, McCarthy AA (2007) Structural insights into the Slit–

Robo complex. Proc Natl Acad Sci USA 104(38):14923–14928.

doi:10.1073/pnas.0705310104

34. Zhang B, Dietrich UM, Geng JG, Bicknell R, Esko JD, Wang L

(2009) Repulsive axon guidance molecule Slit3 is a novel an-

giogenic factor. Blood 114(19):4300–4309. doi:10.1182/blood-

2008-12-193326

35. Suchting S, Heal P, Tahtis K, Stewart LM, Bicknell R (2005)

Soluble Robo4 receptor inhibits in vivo angiogenesis and en-

dothelial cell migration. FASEB J 19(1):121–123. doi:10.1096/fj.

04-1991fje

36. Steigemann P, Molitor A, Fellert S, Jackle H, Vorbruggen G

(2004) Heparan sulfate proteoglycan syndecan promotes axonal

and myotube guidance by slit/robo signaling. Curr Biol

14(3):225–230. doi:10.1016/j.cub.2004.01.006

37. Hu H (2001) Cell-surface heparan sulfate is involved in the re-

pulsive guidance activities of Slit2 protein. Nat Neurosci

4(7):695–701. doi:10.1038/89482

38. Marlow R, Binnewies M, Sorensen LK, Monica SD, Strickland P,

Forsberg EC, Li DY, Hinck L (2010) Vascular Robo4 restricts

proangiogenic VEGF signaling in breast. Proc Natl Acad Sci

USA 107(23):10520–10525. doi:10.1073/pnas.1001896107

39. Shang B, Cao Z, Zhou Q (2012) Progress in tumor vascular

normalization for anticancer therapy: challenges and perspec-tives. Front Med 6(1):67–78. doi:10.1007/s11684-012-0176-8

Angiogenesis

123