Netrin-4 Acts as a Pro-angiogenic Factor during Zebrafish Development*

Background: The role of netrin-4 in angiogenesis is still unclear. Results: We show here that Netrin-4 is expressed in the zebrafish vascular system. We demonstrate that Netrin-4 expression is crucial for zebrafish vessel formation and this effect is mediated via activation of key protein kinases. Conclusion: Netrin-4 is a pro-angiogenic factor during development. Significance: We demonstrate the crucial role of Netrin-4 in blood vessel formation. Netrins form a heterogeneous family of laminin-related molecules with multifunctional activities. Netrin-4, the most distant member of this family, is related to the laminin β chain and has recently been proposed to play an important role in embryonic and pathological angiogenesis. However, the data reported so far lead to the apparently contradictory conclusions supporting Netrin-4 as either a pro- or an anti-angiogenic factor. To elucidate this controversy, Netrin-4 was analyzed for a vascular activity in both cell-based models (human umbilical vein endothelial cells and human umbilical artery endothelial cells) and two zebrafish models: the wild-type AB/Tü strain and the transgenic Tg(fli1a:EGFP)y1 strain. We show that Netrin-4 is expressed in endothelial cells and in the zebrafish vascular system. We also show evidence that Netrin-4 activates various kinases and induces various biological effects directly linked to angiogenesis in vitro. Using a morpholinos strategy, we demonstrate that Netrin-4 expression is crucial for zebrafish vessel formation and that a blood vessel formation defect induced by netrin-4 morpholinos can be partially rescued through drug delivery leading to protein kinase activation. Together these data underscore the crucial role of Netrin-4 in blood vessel formation and the involvement of protein kinases activation in Netrin-4-induced biological effects related to vascular development.

Netrins form a heterogeneous family of laminin-related molecules with multifunctional activities. Netrin-4, the most distant member of this family, is related to the laminin ␤ chain and has recently been proposed to play an important role in embryonic and pathological angiogenesis. However, the data reported so far lead to the apparently contradictory conclusions supporting Netrin-4 as either a pro-or an anti-angiogenic factor. To elucidate this controversy, Netrin-4 was analyzed for a vascular activity in both cell-based models (human umbilical vein endothelial cells and human umbilical artery endothelial cells) and two zebrafish models: the wild-type AB/Tü strain and the transgenic Tg(fli1a:EGFP) y1 strain. We show that Netrin-4 is expressed in endothelial cells and in the zebrafish vascular system. We also show evidence that Netrin-4 activates various kinases and induces various biological effects directly linked to angiogenesis in vitro. Using a morpholinos strategy, we demonstrate that Netrin-4 expression is crucial for zebrafish vessel formation and that a blood vessel formation defect induced by netrin-4 morpholinos can be partially rescued through drug delivery leading to protein kinase activation. Together these data underscore the crucial role of Netrin-4 in blood vessel formation and the involvement of protein kinases activation in Netrin-4-induced biological effects related to vascular development.
Netrins are a family of evolutionarily conserved extracellular proteins. The first member of this family, Netrin-1, has been extensively shown to be implicated in neuronal navigation embryogenesis but was more recently shown to have multiple roles beyond the brain. Netrin-1 has been shown to mediate its different effects through multiple receptors, including Deleted in Colorectal Cancer (DCC) and UNC5, UNC-5 homolog family: UNC5A, -B, -C, and -D (1). Although extensive literature exists on Netrin-1 and its multiple roles, much less is known on the other Netrins. In addition to Netrin-1, two other secreted Netrins (Netrin-3 and -4), and two glycosylphosphatidylinositol-anchored membrane proteins, Netrins G1 and G2, have been identified in mammals (2).
Netrin-4 was first described by Koch and colleagues and is also named ␤-Netrin, because, unlike other Netrins that display homology to the short arms of laminin ␥ chains, this Netrin is more related to the laminin ␤ chains. Netrin-4 is then described to be a basement membrane component present in the basement membranes of the vasculature, kidney, and ovaries. In addition, Netrin-4 is expressed in a limited set of fiber tracts within the brain, including the lateral olfactory tract and the vomeronasal nerve (3).
Similar to the neural network, the vascular network forms from central axial structures that send sprouts along predetermined trajectories to their distal destinations. The trajectories of nerves and blood vessels are often shared, leading to the hypothesis that tissues may use similar if not identical factors to instruct both their innervation and vascularization (4 -7). Angiogenesis, the formation of new blood vessels from preexisting endothelial vasculature, is an essential event involved in a wide variety of physiological processes, including the embryonic development and pathological processes such as the progression and metastasis of tumors (8,9). Blood vessels transport gases, nutrients, waste products, hormones, and circulating cells into every organ of vertebrate organisms. The main pro-angiogenic factors controlling this process are the VEGF 2 family members (10,11). The relevant VEGF receptors, the tyrosine kinases VEGFR2 and VEGFR3, are presented on the endothelial cell surface and, upon ligand binding, trigger downstream signaling, including activation of the mitogen-activated protein kinase (MAPK) pathway, phosphoinositide kinase-3 (PI3K) and Akt pathway, phospholipase C␥, and small GTPases such as Rac1 (12,13).
It was recently described that both Netrin-1 and Netrin-4 may be implicated in angiogenesis (14 -18). However, similarly to Netrin-1, Netrin-4 involvement in angiogenesis remains the subject of some controversial views. Indeed, on the one hand, some studies have reported the anti-angiogenic effect of Netrin-4 on VEGF-stimulated endothelial cells (17) or on human microvascular endothelial and human pancreatic carcinoma cells through Akt and JNK1/2 phosphorylation inhibition (18). On the other hand, other data reported its pro-angiogenic effect (16,19,20). Indeed, Wilson et al. (19) have reported that Netrin-4 stimulates proliferation, migration, and tube formation of endothelial cells. In the same group, Netrin-4 has been recently described to induce ERK, Akt, S6, FAK, and small GTPase activation and subsequent induction of lymphoangiogenesis (20). In addition, Netrin-4 has been described to enhance angiogenesis and neurological outcome after cerebral ischemia (16). However, these studies were either mainly in vitro or used pathological angiogenesis as the model. We thus decided to take advantage of the zebrafish model to analyze the role of Netrin-4 during embryonic vascular development. Indeed, much of the signaling controlling angiogenesis appears highly conserved among different vertebrate species, and the same components of the VEGF pathway control the growth of embryonic blood vessels in zebrafish (21,22). The transparency of the embryo, its development outside of the maternal organism, and high genetic and experimental accessibility make zebrafish particularly suitable for the observation of dynamic processes (23,24).
In this report, we demonstrate that Netrin-4 is expressed in endothelial cells in vitro and in vivo. Moreover, by using a morpholino strategy, we implicate Netrin-4 during blood vessel formation. We also show that Netrin-4-induced endothelial cell survival, proliferation, and migration are associated with ERK1/2, Akt, JNK1/2, and FAK activation and that Netrin-4induced angiogenic sprouting can be mimicked by protein kinases activators (okadaic acid (OA) and N-arachidonoyl-Lserine (ARA)) and inhibited by MAPK kinase inhibitor (U0126). We also provide evidence that OA and ARA delivery during zebrafish development partially rescues the blood vessel formation defect observed in netrin-4 morpholinos-injected Tg(fli1a:EGFP) y1 embryos. Together, these results underline, in vitro and in vivo, the crucial role of Netrin-4 in blood vessel formation and the involvement of protein kinases activation in Netrin-4-induced biological effects related to angiogenesis.

Fish Husbandry
Zebrafish strains of AB/Tü and Tg(fli1a:EGFP) y1 were reared and staged at 28.5°C according to Kimmel and collaborators (25).

Cell Proliferation Assay
HUVECs (1 ϫ 10 4 ) were allowed to adhere on a 96-well plate. After 16 h, the cells were washed with Hepes-BSS and cultured for 24 h in the presence of 0 -500 ng/ml Netrin-4 in ECBM2 at 37°C and 5% CO 2 . After 24 h, 15 l (i.e. 10% of culture media) of WST-1 solution was added per well, and cells were incubated for 4 additional hours at 37°C and 5% CO 2 . Absorbance was recorded at 450 nm with a 96-well plate reader as recommended by the manufacturer (Roche Applied Science).

Cell Viability Assays
HUVECs and HUAECs (4 ϫ 10 4 ) were allowed to adhere on a 12-well plate for 16 h. Cells were then washed with Hepes-BSS and cultured for 24 h in the absence of serum and in the presence of various concentrations of Netrin-4 at 37°C and 5% CO 2 . Cell survival was analyzed by using the Trypan Blue exclusion method. A ToxiLight assay was performed on serum-deprived HUVECs and HUAECs in 24-well plates after 24 h of incubation in the presence of various concentrations of human recombinant Netrin-4 (0 -500 ng/ml) according to the manufacturer's instructions (Lonza).

Cell Death Assays
Fluorometric Caspase-3 Activity Assay-HUVECs and HUAECs (1.5 ϫ 10 5 ) were allowed to adhere in a 6-well plate for 16 h. Cells were then washed with Hepes-BSS and cultured for 24 h in the presence or absence of 150 ng/ml recombinant human Netrin-4 in ECBM2 at 37°C and 5% CO 2 . After 24 h, cells were collected, and equal amounts of cell lysates were subjected to a caspase 3/CPP32 fluorometric assay as recommended by the manufacturer (Biovision). Results are given as the mean Ϯ S.E. of three separate experiments.
Cleaved Caspase-3 Immunostaining and TUNEL Assay-HUVECs and HUAECs were allowed to adhere on coverslips in complete endothelial medium. After 16 h, cells were washed with Hepes-BSS and cultured for 24 h in the presence or absence of 150 ng/ml Netrin-4 in ECBM2 at 37°C and 5% CO 2 . Cells were then fixed with 4% (w/v) PFA in PBS, and apoptotic cells were then detected by immunostaining using cleaved caspase-3 (Asp-175) antibody (1:1000, Cell Signaling) or through DNA fragmentation detection using a TUNEL assay kit following the manufacturer's instructions (Roche Diagnostics). Hoechst staining (1:1000) was used to detect cell nuclei.

Migration Assay
Cell migration was determined using a scratch assay. HUVECs were seeded on 6-well plates to reach 100% confluence within 24 h. Cells were serum-deprived for 4 h, and a similarly sized scratch was made across the center of each well. Wells were then washed with serum-free medium and immediately imaged at baseline. Cells were then treated with 150 ng/ml Netrin-4 or 25 ng/ml VEGF in serum-free medium for 8 h. The wells were not coated with gelatin or any other substrata.

In Vitro Angiogenesis Assays
Tube Formation Assay-Wells of a 96-well plate were precoated for 1 h at 37°C with Matrigel (BD Biosciences). HUVECs (3 ϫ 10 4 ) were then seeded in the presence of various concentrations of Netrin-4 (0 -500 ng/ml) or VEGF (25 ng/ml). The tubular formation was observed and photographed under a phase-contrast microscope after incubation at 37°C in a humidified chamber with 5% CO 2 for 8 h. Angiogenesis quantification was obtained by determining the number of branch points per view.
Three-dimensional Angiogenesis Assay-A three-dimensional angiogenesis assay was performed as recommended by the manufacturer (PromoCell). After 48 h of incubation in the presence or absence of 150 ng/ml human Netrin-4, 20 M ARA, 10 M U0126, or 25 ng/ml VEGF, the number and the length of the sprouts were counted for at least 10 spheroids per conditions and per assay. Angiogenesis quantification was obtained by multiplying the number of the sprouts per spheroid by their length.

mRNA Preparation and Rescue Experiments
Homo sapiens netrin-4 cDNA clone (SC314262) was used to synthesize mRNA (Origene). Sense transcripts from the cDNA clone were generated by using the mMESSAGE mMACHINE T7 kit (Ambion) and were purified with the RNeasy Mini kit (Qiagen). For the rescue experiments, 250 pg of human netrin-4 mRNA was co-injected with netrin-4 splice morpholino at the one-cell stage.

Whole Mount in Situ Hybridization
Whole mount in situ hybridization was performed as described previously (26). Genomic DNA was used as a template, and the following primers containing the T7 RNA polymerase promoter were used to amplify the probe template: For-NTN4-probe: TCGTGCTCCTACATGCAGTC, Rev-NTN4- FEBRUARY 3, 2012 • VOLUME 287 • NUMBER 6 probe: CTAATACGACTCACTATAGGGGGAGGGTTTCT-TCGGATCTC (the underlined sequence corresponds to the T7 RNA polymerase promoter). Probe was detected with either an anti-digoxigenin-AP antibody (1:5000) and subsequent NBT/BCIP exposure or an anti-digoxigenin-rhodamine antibody (1:200, Roche Applied Science).

Whole Mount Immunofluorescence Staining
AB/Tü embryos were sacrificed with an overdose of Tricaine (Sigma A5040) and were subsequently fixed overnight at 4°C in PBS, pH 7.4, with 4% (w/v) PFA. Permeabilization was performed by digestion with proteinase K (10 g.ml Ϫ1 ) at room temperature for the appropriate time (26). The proteinase K digestion was then stopped by incubating the embryos for 20 min in 4% (w/v) PFA in PBS. Embryos were incubated for 3 h at room temperature in blocking solution (PBS 0.1% Tween (PBST), 1ϫ Blocking solution (Roche Applied Science), and 2% BSA). Incubation with mouse monoclonal anti-human Netrin-4 antibodies (MAB1254, 1:100) was performed overnight at 4°C. After washing 6 ϫ 15 min with PBST, embryos were incubated with Alexa Fluor 488-conjugated anti-mouse secondary antibodies (1:400, Jackson Laboratory) for 4 h at room temperature. Finally, after washing 6 ϫ 15 min with PBST, the embryos were mounted for observation in fluorescent mounting medium (DakoCytomation S3023). Observations were made with an epifluorescence microscope (Axiovert 200M, Zeiss). In some experiments, immunofluorescence labeling was performed on Tg(fli1a:EGFP) y1 . GFP was then detected using rabbit polyclonal anti-GFP antibodies (1:400), and netrin-4 was detected with mouse monoclonal anti-human Netrin-4 antibodies (MAB1254). After washes, embryos were incubated with Alexa Fluor 488-conjugated anti-rabbit secondary antibodies and Cy3-conjugated anti-mouse antibodies. Note that human netrin-4 antibodies were validated through negative controls, i.e. 1) no secondary antibodies (in this case only autofluorescence of blood cells is detectable at a very high exposure) or 2) immunofluorescence labeling on MO-injected embryos.
For DNA fragmentation detection, whole embryos were fixed for 2 h in PBS, pH 7.4, with 4% (w/v) PFA at room temperature, and apoptotic cells were detected using the In Situ Cell Death Detection Kit, tetramethylrhodamine red following manufacturer's instructions (Roche Diagnostics).

Statistical Analysis
Results are given as the mean Ϯ S.E. The significance of differences between mean values was evaluated by Student t test (*, p Ͻ 0.05; **, p Ͻ 0.01; and ***, p Ͻ 0.001).

Netrin-4 Is Expressed in Zebrafish
Vasculature-The expression of netrin-4 during zebrafish development was first examined by quantitative RT-PCR and shows a low expression of netrin-4 at 8 hpf and 12 hpf, which rapidly increases from 24 hpf to 72 hpf and stabilizes at 96 hpf (Fig. 1A). The netrin-4 expression pattern was then analyzed by whole mount in situ hybridization using netrin-4 antisense RNA probes and revealed the spatial and temporal expression of netrin-4 during embryonic development of AB/Tü strain. Netrin-4 is ubiquitously expressed at 8 hpf. During the segmentation period (10.33-24 hpf), netrin-4 expression concentrates in the brain and in the tail along the yolk sac (at 18 hpf) and under the notochord (at 22 hpf). At 32 and 48 hpf, netrin-4 expression is expressed in dorsal aorta (DA) and posterior cardinal vein (PCV) of the trunk (Fig.  1B). A faint netrin-4 staining is also distinguished in the intersegmental vessels (ISV). According to the results obtained by Koch and colleagues in 2000 (3), in addition to netrin-4 mRNA localization to the retina and eye vasculature, netrin-4 mRNA is also localized in a limited set of fiber tracts within the brain, including the lateral olfactory tract and the vomeronasal nerve (supplemental Fig. S1). To validate netrin-4 expression in vascular system, we performed fluorescent in situ hybridization (FISH) using netrin-4 antisense probes coupled to GFP immunofluorescent staining on Tg(fli1a:EGFP) y1 embryos, and we showed that netrin-4 mRNA co-localized with GFP, and therefore with blood vessels, at 24 and 96 hpf (Fig. 1C). Netrin-4 protein expression in blood vessels of AB/Tü embryos was further investigated by whole mount immunofluorescence staining using human netrin-4 antibodies. The same expression profile was obtained for the protein compared with the spatio and temporal expression profile obtained by in situ hybridization; i.e. 1) ubiquitous expression of netrin-4 at 8hpf and 2) at 48 hpf, netrin-4 concentrates in the blood vessels of the trunk (PCV, DA, and ISV), the brain, and around the aortic arches (AA, Fig.  1D). Note that negative control was performed for each immunofluorescent staining at each developmental stage and that the only signal observed without primary antibodies (human netrin-4 antibodies) was the autofluorescence of blood cells at very high exposure time. Netrin-4 protein expression in blood vessels of Tg(fli1a:EGFP) y1 was then confirmed by double immunofluorescent staining for Netrin-4 using anti-human-Netrin-4 antibodies and anti-GFP antibodies at 48 hpf (Fig. 1E). All these results indicate that Netrin-4 is localized in zebrafish blood vessels.
Netrin-4 Is Crucial for Zebrafish Vascular Development-Based on in situ vascular expression pattern, we hypothesized that Netrin-4 has a role in ISV sprout and primary vessels formation in the trunk region of the developing zebrafish vasculature. To test this hypothesis, we investigated the function of netrin-4 gene during zebrafish embryonic development by using morpholinos antisense oligonucleotides directed against netrin-4 translation site (MO-tsl) or netrin-4 exon splice donor site to disrupt mRNA splicing (MO-spl) (27,28). The genomic structure of netrin-4 contains 10 exons of which splice MO targets the end of exon 2 ( Fig. 2A). The dose of both morpholinos was carefully titrated so that the minimal amount of morpholino (MO-tsl: 3 ng/embryo and MO-spl: 4 ng/embryo) gave nearly complete knockdown and leaded to the same phenotype observable from 22 hpf. No obvious phenotype was observed before this developmental stage. Note that in all studies, we excluded p53-mediated off-target effects on apoptosis and the observed netrin-4 morpholino-induced phenotypes, by co-injecting p53 morpholinos, as described (29) (data not shown). The efficiency of splice MO was confirmed by RT-PCR across the flanking exons (exons 2 and 3) of the targeted splice donor site ( Fig. 2A), which revealed, at 24 and Netrin-4 in Developmental Angiogenesis a, lanes 4 and 7, respectively, upper gel) when compared with MS- (Fig. 2B, panel a, lanes 2 and 5, respectively) and MOtsl-injected (Fig. 2B, panel a, lanes 3 and 6, respectively) embryos. Morpholino efficiency was also validated by Western blot, and we show that Netrin-4 protein expression was dramatically diminished in whole MO-tsl and MO-spl-in-jected embryos at 48 hpf when compared with MS-injected embryos (Fig. 2B, panel b).
The effect of netrin-4 knockdown was thus analyzed on Tg(fli1a:EGFP) y1 embryos. Morpholino-injected embryos were classified depending on the severity of the vascular defects. Control 32 hpf embryo displays normal vasculature in the trunk with normal DA and PCV with ISV extending dorsally (Fig. 2C,

Netrin-4 in Developmental Angiogenesis
upper panels). In netrin-4-MO-injected embryos with moderate defects (43.0% of total embryos), the majority of ISV were absent; remaining ISV did not extended dorsally, and were thinner and more diffuse (Fig. 2C, middle panels). Netrin-4 MO-injected embryos with more severe defects (39.5% of total embryos), lacked ISV outgrowth, and had gaps in the DA and PCV around the trunk; cranial vasculature and branchial aortic arches development was also largely affected (Fig. 2C, lower  panels). Note that only blood vessel formation (vasculogenesis and angiogenesis) was altered; i.e. 1) blood circulation was normal in the remaining vessels and started around 32 hpf; 2) the heartbeat has been determined and was around 130 -140 beats per minute for MO-injected embryos at 48 hpf as for MS-injected embryos; 3) no defect in somite formation was observed at 36 hpf. The more severe defects observed in MO-injected embryos were never compensated, and this strong blood vessel impairment finally resulted in death at around 3-5 days post-fertilization.
To definitely validate the specificity of netrin-4 MO, we produced human netrin-4 mRNA and co-injected it with spliceblocking netrin-4 MO in Tg(fli1a:EGFP) y1 embryos. In this experiment, human Netrin-4 expression following netrin-4 mRNA injection has been monitored in 32 hpf whole embryos and is shown in Fig. 2D (panel a). ISV development in the corresponding embryos was then analyzed, and we showed that human netrin-4 mRNA significantly rescued the blood vessel formation defect observed following splice-blocking netrin-4 MO injection (61.7% of embryos appeared with normal vasculature or with Ͼ5 ISV along the trunk in splice-blocking netrin-4 MO and human netrin-4 mRNA-injected embryos (MO-spl ϩ mRNA) compared with 28.3% of embryos with "normal" vasculature in splice-blocking netrin-4 MO-injected embryos (MO-spl)) ( Fig. 2D (panel b) and supplemental Fig.  S2). Note that, through this method, we could initially suppose that human netrin-4 mRNA injection, and thus Netrin-4 overexpression, would lead to an anarchical blood vessel formation in Tg(fli1a:EGFP) y1 embryos. However, no particular phenotype was observed even by injecting 1000 ng/embryo (data not shown). This could be explained by the view that the level of expression of the Netrin-4 receptor, even though unknown here, becomes the limiting factor. According to our previous data demonstrating that netrin1a morphants exhibit severe defects in blood vessel development and in parallel a strong apoptosis induction (15), we hypothesized that netrin-4 knockdown in zebrafish could also lead to apoptosis induction. To validate this hypothesis, we performed 1) TUNEL staining on netrin-4 MO-injected embryos and on MS-injected embryos and 2) caspase-3 activity detection on whole embryos (Fig. 2E). We thus showed that netrin-4 knockdown in zebrafish leads to a strong apoptosis induction (Fig. 2E). Altogether, these results support the requirement of Netrin-4 in zebrafish blood vessel formation.
Netrin-4 Induces Endothelial Cell Survival, Proliferation, and Migration and Is Pro-angiogenic in Vitro-To determine how Netrin-4 affects primary endothelial cells, we investigated the effect of Netrin-4 on apoptosis, proliferation, and migration of HUVECs and HUAECs. Because complete medium contains a large amount of Netrin-4 and because endothelial cells produce and secrete high levels of Netrin-4 in normal culture medium (supplemental Fig. S3 and Fig. 4 (A and B)), endothelial cells were serum-deprived for 16 h before in vitro experiments, and Netrin-4 effect was analyzed in the absence of serum. In the presence of Netrin-4, HUVECs and HUAECs presented a significantly reduced percentage of trypan blue-positive cells compared with the cells cultured in the absence of Netrin-4. The optimal effect of Netrin-4 was observed at 150 ng/ml (Fig.  3A). In the same manner, we demonstrated using ToxiLight assay that Netrin-4 decreased the cell death in both HUVECs and HUAECs (Fig. 3B). By analyzing more precisely the effect of Netrin-4 on endothelial cell survival, we showed that the exogenous Netrin-4 significantly protected HUVECs and HUAECs from serum deprivation-induced apoptosis as observed by caspase-3 activity assay (Fig. 3C). These results were confirmed by microscopic analysis of the cleaved caspase-3-positive cells and by DNA fragmentation detection using TUNEL assay kit. Quantification of cleaved caspase-3-positive cells and TUNELpositive cells demonstrated that Netrin-4 inhibited serum deprivation-induced cell death (5.72 Ϯ 0.76% compared with the control, 18.88 Ϯ 0.65%, for cleaved caspase-3 staining and 0.41 Ϯ 0.12% compared with the cells cultured in the absence of Netrin-4 (Ctrl ϭ index: 1) for TUNEL staining) (Fig. 3, D and E). Our results thus demonstrate that Netrin-4 supports endothe-  FEBRUARY 3, 2012 • VOLUME 287 • NUMBER 6 lial cell survival under conditions that had been shown to induce apoptosis. Netrin-4 effect on endothelial cell proliferation was thus examined using WST-1 reagent, and we showed that Netrin-4 induced HUVECs proliferation in a dose-depen-dent manner. The optimal effect was also observed at 150 ng/ml (1.28 Ϯ 0.09 compared with the cells incubated in the absence of Netrin-4: index ϭ 1) (Fig. 3F). Note that Netrin-4 had only a modest effect on cell proliferation, however, VEGF (25 ng/ml)

Netrin-4 in Developmental Angiogenesis
also led to cell proliferation at a low rate (1.20 Ϯ 0.03). Netrin-4 effect on endothelial cell migration was subsequently examined using a scratch assay and we showed that Netrin-4 stimulation led to a decrease of the distance between the two HUVEC monolayers and so induced HUVECs migration similarly to VEGF compared with cells incubated without any factors (60.2 Ϯ 13.7% with 150 ng/ml Netrin-4, 56.8 Ϯ 14.7% in the presence of 25 ng/ml VEGF compared with 92.1 Ϯ 13.3% for the Ctrl) (Fig. 3G).
HUVECs completely reorganize when cultured with complete medium or VEGF in three-dimensional Matrigel and finally form a network of tubular structures with multiple cellcell contacts. To investigate Netrin-4 effect on angiogenesis, we used two different methods: 1) the capillary-like tube formation assay (Fig. 4A) and 2) an in vitro three-dimensional angiogenesis assay enabling the detection of the entire angiogenesis process, i.e. proteolytic activity, migration, proliferation, and lumen formation (Fig. 4B). Through these methods, we showed that Netrin-4 significantly induced angiogenesis in a dose-dependent manner, and an optimal effect was observed at 150 ng/ml Netrin-4. All these results demonstrate that Netrin-4 is sufficient to promote endothelial cell survival, proliferation, migration, and angiogenesis in vitro.
Netrin-4 Biological Effects Are Mediated by Protein Kinase Activation-To determine the signaling pathway induced by Netrin-4 and leading to proliferation, survival, and migration of endothelial cells, we examined, in HUVECs, the activation of various classic signaling kinases known to be implicated in those processes. We showed by Western blot that Netrin-4 induced FAK, Akt, JNK1/2, and ERK1/2 phosphorylation in a dose-dependent manner, and the maximal effect was observed at 150 ng/ml, which was in accordance with the optimal concentration leading to Netrin-4-induced biological effects on endothelial cells (Fig. 5A).
As a second attempt to investigate whether protein kinase activation are implicated in Netrin-4 pro-angiogenic effect, we used two known MAPK/Akt pathway activators (OA and ARA). OA is a natural polyether compound isolated from a black sponge, Halichondria okadai (30) and is a potent inhibitor of protein phosphatase 1 and 2A (31). Through protein phosphatase 2A inhibition, OA notably enables PI3K/Akt and ERK1/2 activation (32). ARA is derived from the endocannabinoid N-arachidonoyl ethanolamine (anandamide), which is found both in the central nervous system and in the periphery and plays a role in numerous physiological systems. ARA has been reported to produce endothelium-dependent vasodilation of rat isolated mesenteric arteries and abdominal aorta and to stimulate phosphorylation of ERK1/2 (MAPK) and Akt in cultured endothelial cells (33,34). To validate the potential effect of OA and ARA on endothelial cells, serum-deprived HUVECs were treated for 30 min in the presence of these drugs and/or with Netrin-4 (150 ng/ml). We showed that OA (10 nM) and ARA (10 M) mimicked Netrin-4-induced ERK1/2, Akt, and SAPK/JNK phosphorylations, whereas, as expected these drugs had no effect on FAK phosphorylation (Fig. 5B). Moreover, Netrin-4-induced ERK1/2 phosphorylation was abolished by preincubating cells with U0126 (selective MAPK kinase inhibitor (10 M)) as expected, whereas no effect was observed on FAK, Akt, and SAPK/JNK activation induced by Netrin-4. These results thus suggest that Netrin-4 induces FAK, ERK1/2, Akt, and JNK activation in endothelial cells and that this effect is mimicked by drugs known to induce protein kinase phosphorylation. We then demonstrated that ARA mimicked Netrin-4induced pro-angiogenic effect in three-dimensional angiogenesis assay (Fig. 5, C and D). Indeed, a careful quantification of the number and the length of sprouts per spheroids showed a significant effect of Netrin-4 (2.89 Ϯ 0.35), which was mimicked by ARA (2.89 Ϯ 0.36) compared with the control (1.00). As expected, U0126 led to the inhibition of the Netrin-4 proangiogenic effect (1.24 Ϯ 0.08), and ARA significantly increased the pro-angiogenic effect induced by Netrin-4 (3.62 Ϯ 0.23) (Fig. 5, C and D). These results thus demonstrate that Netrin-4 induces protein kinase activation in endothelial cells and that this effect is mimicked by OA and ARA, two MAPK/Akt pathway activators.
Netrin-4 Morpholino-induced Blood Vessel Formation Defect Is Partially Rescued by OA and ARA-According to the data presented above, Netrin-4 is crucial for blood vessel formation, and its role is notably mediated through protein kinase activation. To verify this hypothesis in vivo, we analyzed the effect of OA and ARA on blood vessel formation of netrin-4-morpholino-injected embryos. We first showed, in a lateral view of index ϭ 1) and correspond to the mean Ϯ S.E. of three separate experiments. C-E, HUVECs and HUAECs were incubated in serum-free medium for 24 h to induce apoptosis and in the presence or absence of Netrin-4 (150 ng/ml). Caspase-3 activity in cell lysates was then detected using caspase-3 activity assay. Results are expressed as an index corresponding to the caspase-3 activity compared with the cells incubated in the absence of Netrin-4 (Ctrl; index ϭ 1) and correspond to the mean Ϯ S.E. of five separate experiments (C). HUVECs were allowed to adhere on coverslips for 16 h in complete medium. Cells were then incubated for 24 h in the presence or absence of Netrin-4 (150 ng/ml). Cells were then fixed, and apoptotic cells were detected by immunostaining using cleaved caspase-3 antibody (D) or through DNA fragmentation detection using TUNEL assay kit following manufacturer's instructions (E). Total cells were visualized through nuclei staining using Hoechst (blue). Representative fields were observed using an epifluorescence microscope,  FEBRUARY 3, 2012 • VOLUME 287 • NUMBER 6 embryos at 36 hpf, that the blood vessel formation defect observed in the trunk of netrin-4-MO-injected embryos from Tg(fli1a:EGFP) y1 zebrafish (MO-tsl) was partially rescued by incubation of the embryos from 6 hpf with both OA (10 nM) (MO-tsl ϩ OA) and ARA (10 M) (MO-tsl ϩ ARA). Indeed, ISV, PCV, and DA were observed even if their formations were not regular (Fig. 6A). Analysis of blood vessels demonstrated that the percentage of netrin-4 morpholino-injected embryos with Ͼ5 ISV increased at least 2-fold in the presence of OA and 1.8-fold in the presence of ARA (76.5% in MO-tsl ϩ OA and 66.7% in MO-tsl ϩ ARA compared with 35.7% in MO-tsl ϩ DMSO) (Fig. 6B). Note that treatment of MS-injected embryos with OA and ARA had absolutely no consequence on vascular architecture as quantified in Fig. 6B. These results thus underline the crucial role of protein kinases activation in Netrin-4induced blood vessel formation.

DISCUSSION
In this study, we identified netrin-4 as a gene expressed in endothelial cells and required for blood vessel embryonic development. In vitro data showed that Netrin-4 is necessary for survival, proliferation, and angiogenesis activities in endothelial cells and is mediated through common proteins involved in signaling leading to cell survival and proliferation (i.e. Akt and ERK1/2). Furthermore, Netrin-4 has a pro-angiogenic activity in vitro in three-dimensional angiogenesis assay and in vivo in zebrafish embryonic development and both effects are at least partially mediated through protein kinase activation.
We showed that Netrin-4 induces proliferation and survival of HUVECs and HUAECs in a time-and dose-dependent manner. These results could be interpreted as opposite results when compared with previous data reporting that Netrin-4 is an antiangiogenic factor (17,18). However, previous studies have analyzed Netrin-4 effect on endothelial cells incubated in the presence of complete endothelial growth medium (18), and we show here in supplemental Fig. S3 that Netrin-4 is expressed at a high level in this complete medium. We thus have performed all in vitro experiments on serum-deprived cells. Moreover, the optimal anti-angiogenic effect in previous studies has been observed at very high concentrations of Netrin-4; i.e. 400 ng/ml to 6 g/ml (17) and 10 -50 g/ml (18), whereas the optimal pro-angiogenic effect, that we reported, is obtained at 150 ng/ml, which is in accordance with the physiological concen- trations (50 -500 ng/ml) previously reported in Dean Li's laboratory (19,20). These results are thus consistent with the hypothesis that Netrin-4 stimulates in vitro endothelial cells at low concentrations but is inhibitory at higher doses (which is suggested by proliferation analysis (Fig. 3F) and phosphorylation analysis of various kinases (Fig. 5A) and the demonstration that HUVECs and HUAECs in vitro show a biphasic response to Netrin-1 (15, 17-20, 35)). Netrins, as axon guidance cues, have originally been described as bifunctional factors, promoting attraction or repulsion, dependent on their concentration, cell expression, or dimerization status of its receptors (1). It could then be suggested that either a combination of receptor expression or receptors affinity explains the biphasic activities of Netrin-4 in the endothelium in vitro.
Angiogenesis is notably known to be mediated through VEGF receptor leading to proliferation, vascular permeability, FIGURE 5. Netrin-4 induces FAK, MAPK, JNK1/2, and Akt activation and protein kinase activation is crucial for Netrin-4-induced angiogenesis. A, to study the effect of Netrin-4 on kinase phosphorylation, serum-starved HUVECs were stimulated for 30 min in the presence of various concentrations of recombinant human Netrin-4 (0 -500 ng/ml). Cell lysates were analyzed for FAK, Akt, JNK1/2, and ERK1/2 phosphorylation by Western blot. Total FAK, Akt, JNK1/2, ERK1/2, and ␤-actin served as loading control. B, Netrin-4-induced ERK1/2 phosphorylation is mimicked by MAPK/Akt pathway activators (okadaic acid (OA) and N-arachidonoyl-L-serine (ARA)). Serum-starved HUVECs were pretreated or not with 10 nM OA, 10 M ARA, or 10 M U0126 for 30 min and then stimulated for 30 min with 150 ng/ml Netrin-4. Cell lysates were analyzed for FAK, Akt, ERK1/2, and SAPK/JNK phosphorylation by Western blot. Total FAK, Akt, ERK1/2, SAPK/JNK, and ␤-actin were used as loading control. C, the effect of MAPK kinase inhibitor and protein kinase activators on Netrin-4-induced angiogenic sprouting was evaluated using three-dimensional angiogenesis assay. HUVEC spheroids embedded in collagen matrix were incubated in assay medium in the presence or absence of Netrin-4 (150 ng/ml) and/or ARA (20 M) or U0126 (10 M). After 48 h, at least 10 spheroids per well were photographed using a phase-contrast microscope. Representative pictures are shown (n ϭ 3 separate experiments). Scale bar: 50 m. D, quantification of angiogenesis induction was carried out by determining the cumulative sprout length accounting for both sprout number and sprout length (i.e. number of sprouts per spheroid x average sprout length per spheroid). Results are given as an index with arbitrary unit (0 h, Ctrl ϭ 1) corresponding to three separate experiments. FEBRUARY 3, 2012 • VOLUME 287 • NUMBER 6 JOURNAL OF BIOLOGICAL CHEMISTRY 3997 cell migration, and cell survival (36). These biological effects involve the activation of various signaling pathways such as the Ras/MEK/ERK pathway resulting in cell proliferation; the focal adhesion kinase pathway leading to focal adhesion turnover and cell migration; activation of the PI3K/Akt pathway maintaining cell survival and leading to vascular cell permeability; and activation of the MKK4/JNK1/2 pathway leading to cell survival (36,37). Regarding the involvement of these signaling pathways in angiogenesis induction, we analyzed in vitro the putative activation of FAK, Akt, JNK1/2, and ERK1/2 in Netrin-4-induced angiogenesis and we showed that all these proteins are phosphorylated and thus activated following Netrin-4 stimulation. The use of drugs inducing ERK1/2, Akt, and SAPK/JNK phosphorylation provides evidence that those pathways are involved in Netrin-4-mediated angiogenesis both in vitro and in vivo.

Netrin-4 in Developmental Angiogenesis
Using the zebrafish model we provide here strong evidence for a pro-angiogenic role of Netrin-4. We and others have previously reported using the same zebrafish model Tg(fli1a: EGFP) y1 that Netrin-1 is also a pro-angiogenic factor (15,19). However, in our previous work, we demonstrated that Netrin-1 acts as a pro-angiogenic factor by blocking the proapoptotic effect of the dependence receptor UNC5B (15). Although, UNC5B has been described to be strongly expressed in capillaries and endothelial tip cells, in their work, Wilson and colleagues were unable to validate the interaction of Netrin-4 with one of the known Netrin receptor (DCC, neogenin, UNC5A-D, and A2b) (19,38,39). Considering the very recent studies on Netrin-4 involvement in cell adhesion through integrin receptor, we could suppose that in our model, Netrin-4, could induce cell adhesion (40,41). However, in our conditions, despite an effect on FAK phosphorylation, Netrin-4 did not modulate HUVECs and HUAECs cell adhesion (data not shown).
Despite the fact that Netrin-4 elicits an endothelial response identical to that of Netrin-1 in our in vitro and in vivo assays, we were unable to identify the Netrin-4 receptor in endothelial cells through RNA interference strategy (data not shown). This suggests that Netrin-4 receptor might be an as yet unknown Netrin receptor, which remains to be identified.