The role of survivin in angiogenesis during zebrafish embryonic development

Background Survivin is the smallest member of the inhibitor of apoptosis (IAP) gene family. Recently, the zebrafish survivin-1 gene has been cloned, showing remarkable sequence identity and similarity over the BIR domain compared with human and mouse survivin gene. Here we investigated the role of survivin in angiogenesis during zebrafish development. Morpholinos (MOs) targeting the 5' untranslated region (UTR) (SurUTR) and sequences flanking the initiation codon (SurATG) of zebrafish survivin-1 gene were injected into embryos at 1–4 cell stage. Vasculature was examined by microangiography and GFP expression in Tg(fli1:EGFP)y1 embryos. Results: In embryos co-injected with SurUTR and SurATG-MOs, vasculogenesis was intact but angiogenesis was markedly perturbed, especially in the inter-segmental vessels (ISV) and dorsal longitudinal anastomotic vessels (DLAV) of the trunk, the inner optic circle and optic veins of developing eyes and the sub-intestinal vessels. Apoptosis was increased, as shown by TUNEL staining and increase in caspase-3 activity. Efficacy of SurUTR and SurATG-MOs was demonstrated by translation inhibition of co-injected 5'UTR survivin:GFP plasmids. The phenotypes could be recapitulated by splice-site MO targeting the exon2-intron junction of survivin gene and rescued by survivin mRNA. Injection of human vascular endothelial growth factor (VEGF) protein induced ectopic angiogenesis and increased survivin expression, whereas treatment with a VEGF receptor inhibitor markedly reduced angiogenesis and suppressed survivin expression. Conclusion: Survivin is involved in angiogenesis during zebrafish development and may be under VEGF regulation.


Background
Survivin is the smallest member of the inhibitor of apoptosis (IAP) gene family containing a single Baculovirus IAP Repeat (BIR) domain and an extended -COOH termi-nal α-helical coiled coil [1]. Survivin is not expressed in most normal adult tissues but is highly expressed in solid and hematological malignancies, where it has been linked to increased angiogenesis and tumorigenesis [2,3]. During human and murine embryonic development, survivin is ubiquitously expressed [4]. However, homozygous knock-out of survivin in mouse ES cells results in disrupted microtubule formation and polyploidy as well as early embryonic fatality, precluding characterization of its functions during murine development [5]. As a result, the role of survivin during embryonic development remains unclear.
Recently, the zebrafish survivin-1 gene (abbreviated survivin) has been cloned, showing remarkable sequence identity and similarity over the BIR domain compared with human and mouse survivin gene [6]. Microarray analysis showed that survivin is significantly up-regulated in a zebrafish chordin morphant in which the intermediate cell mass (ICM, where vascular and primitive hematopoietic tissues arise) was expanded [7]. Here, we investigated if survivin plays a role in vascular formation during zebrafish embryonic development.

Expression of survivin in zebrafish embryos
Whole-mount in-situ hybridization was performed to examine survivin mRNA expression in zebrafish embryos at 26 hpf. Survivin was detected diffusely throughout the developing brain and neural tube. It was also expressed at the vicinity of the axial vasculature from which the intersegmental vessels arise (Figure 1a-b). This was further confirmed in histological sectioning in which the areas corresponding to the developing axial vasculature and neural tube showed increased expression relative to the adjacent tissues (Figure 1b, insert). Furthermore, double in-situ hybridization showed that survivin was expressed in the developing axial vasculature dorsal to the intermediate cell mass (ICM), where gene encoding for embryonic hemoglobin α was expressed. The pattern was remarkably similar to that of flk1, a VEGF receptor tyrosine kinase (Figure 1c-d).

Survivin morphants
The role of survivin during embryonic development was investigated by knocking-down its function using MOs. The phenotypic penetrance of survivin MOs was dose-and time-dependent. At 22 hpf, when injected with either 3 ng Sur UTR or 3 ng Sur ATG -MOs (referred as Sur UTR mo and Sur AT-G mo embryos), most embryos had a relatively normal morphology (Figure 2a,c). However, at 48 hpf, 74.8 ± 7.3% and 72.0 ± 4.0% of embryos manifested "characteristic phenotypes" with reduced eye and head sizes and a mildly curved tail (Figure 2b,d). There was no overt tissue necrosis in these embryos. At 6 ng of either MOs, increasing numbers of embryos became severely deformed and died shortly after 48 hpf (Figure 2d, insert). Co-injecting Sur ATG + Sur UTR -MOs (3 ng each) resulted in specific phenotypes in 79.4 ± 7.2% embryos without increase in toxicity or mortality as compared with 3 ng of either MO alone (Figure 2e). The combination regimen remained significantly less toxic than that of Sur UTR -MO at 6 ng. In all subsequent experiments, Sur ATG and Sur UTR -MOs were co-injected at 3 ng each (referred as Sur UTR+ATG mo embryos). Only embryos with characteristic phenotypes were investigated while those which were severely deformed were excluded from analysis.

Effects of survivin knock-down on angiogenesis
We have previously shown that survivin is significantly upregulated in a zebrafish chordin morphant in which the ICM was expanded [7]. Therefore, we first examined the effects of survivin knock-down on vascular formation in Tg(fli1:EGFP) y1 embryos. In uninjected embryos, the axial circulation (AC), inter-segmental vessels (ISV), dorsal longitudinal anastomotic vessels (DLAV), vertebral and subintestinal vessels (SIV) were readily observable ( Figure  3a,c). In Sur UTR+ATG mo embryos, the dorsal aorta and posterior cardinal vein were patent, indicative of intact vasculogenesis (see additional file 1: Wild-type embryos and file 2: Survivin morphants). However, the development of vertebral and ISV was perturbed with defective or total absence of sprouting as well as failure to form the DLAV and SIV (Figure 3b,d). These defects were seen in all 54 Sur UTR+ATG mo embryos observed (n = 3 separate experiments) with the characteristic phenotypes. The results were confirmed using microangiography in which defective ISV sprouting and failure to form the DLAV, as well as defective inner optic circle (IOC) and optic veins (OV) of the developing eyes were seen in the Sur UTR+ATG mo embryos (Figure 3e-h). Similar patterns of angiogenesis defects were observed when either Sur UTR or Sur ATG morpholinos were injected (data not shown).

Effects of survivin MOs on apoptosis as shown by TUNEL and caspase-3 activity
As a member of the IAP family, survivin has been shown to inhibit apoptosis by regulating caspase activity [1,2]. Therefore; we investigated if there was increased apoptosis in the Sur UTR+ATG mo embryos as measured by TUNEL assay. At both 24 and 48 hpf, increased TUNEL staining was detected in the developing neural tube and the brain (not shown), with significant, albeit weaker, staining at the vicinity of the axial vasculature (Figure 3i,j). The increased apoptosis was further confirmed by specific caspase-3 activity which was significantly increased in 48 hpf Sur UTR+ATG mo embryos (299.1 ± 8.3 arbitrary units) compared with control embryos injected with a random sequence MO at 6 ng (103.0 ± 2.3 arbitrary units, n = 3 experiments using 240 embryos, p < 0.05).

Specificity of survivin knock-down
To further demonstrate the efficacy of Sur UTR and Sur ATG MO binding to survivin mRNA, embryos were co-injected with a 5'UTR survivin:GFP plasmid (50 pg) and Sur UTR+ATG -MOs (3 ng each). Injecting the plasmid alone lead to GFP expression in 79.7 ± 9.4% embryos ( Figure  4a Embryos injected with a combination of Sur ATG (3 ng) and Sur UTR morpholinos (3 ng) (Sur ATG+UTR ) at 1-4 cell stage. Noted that while there was no significant morphological changes at 22 hpf, most of the embryos injected with Sur ATG+UTR morpholinos at 48 hpf showed a "characteristic phenotype" with reduced head and eye size (circled) and a mildly curved tail (arrowheads). Similar phenotypes were also seen in embryos injected with either Sur ATG or Sur UTR morpholinos at various doses but not in WT embryos injected with random sequence. Insert (d) showed a severe phenotype at 48 hpf characterized by severely deformed embryos which did not survive beyond 48 hpf. These embryos were not included in the analysis. Each picture is representative of at least three experiments. (e): The dosedependence of either Sur ATG , Sur UTR or Sur ATG+UTR morpholinos. Optimal response was observed when embryos were co-injected with 3 ng of each MO (Sur UTR+ATG ). Results were expressed as mean ± S.E.M. In each experiment, MOs at different doses were injected into the same batch of embryos and were scored at the same time. More than 40 embryos have been injected at each dosage.
Whole-mount in-situ hybridization showing the expression of survivin in zebrafish embryos

Effects of VEGF on survivin expression
VEGF plays an important role in angiogenesis during zebrafish embryonic development [8]. In-vitro studies have shown that survivin mediates the proliferative and anti-apoptotic effects of VEGF in endothelial cells [9]. Therefore; we investigated if survivin expression during embryogenesis is regulated by VEGF. Exogenous human VEGF protein (2 ng) was injected into zebrafish embryos at one-cell stage [10]. Angiogenesis was examined in the sub-intestinal vessels at 96 hpf, where the vasculature was well-developed and any ectopic structures could be readily detectable. In 78 out of 110 embryos (70%) (from three separate experiments), VEGF induces ectopic angiogenesis which was associated with a significant up-regulation of survivin mRNA expression (Figure 5a-b,f). We also incubated embryos with a VEGF receptor inhibitor (VEGFTKR) at one-cell stage. VEGFTKR (25 μmol/L) induced defective angiogenesis in all treated embryos at 48 hpf (Figure 5c-d) and could not be rescued by survivin mRNA injection (30 pg) (Figure 5e). Survivin mRNA expression was significantly down-regulated in these embryos ( Figure 5f).

Discussion
In this study, we demonstrated that knock-down of survivin with MOs gives rise to embryos with reduced eye and head sizes and a mildly curved tail. Similar phenotypic changes have been described in a survivin mutant generated in a retrovirus insertional mutagenesis screen [11].
The survivin morphants had defective angiogenesis but vasculogenesis, i.e. formation of axial vasculature, was not affected at the doses of MOs tested. Development delay in these morphants was not observed, as shown by the normal onset and pattern of pigmentation and heart beat (data not shown). Our results corroborate with in-vitro studies showing that survivin is important for the maintenance of proliferation and survival on endothelial cells [9,12]. In addition, our data provided new information on the role of survivin during embryonic development.
In-vitro and tumorigenesis studies have shown that survivin mediates the angiogenic effects of VEGF [9,12,13].
In zebrafish embryos, VEGF signaling is important for angiogenesis. In particular, mutants defective in a Effects of survivin knock-down on angiogenesis and circulation zebrafish orthologue of flk1 (a VEGF-receptor tyrosine kinase), the schwentine [14], and in phospholipase Cγ (a tyrosine kinase mediating effects of VEGFR), the y10 [15], exhibit specific defects in angiogenesis. MO targeting of VEGF results in defective circulation in the head, axial and inter-segmental vasculature in a dose-dependent fashion [8]. In this study, VEGF induces ectopic angiogenesis and up-regulates survivin mRNA expression, suggesting that survivin may mediate the angiogenic effect of VEGF. The link between VEGF and survivin during zebrafish angiogenesis has not been examined but may involve PKB/Akt signaling as reported in human endothelial cell lines [16]. Intriguingly, co-injecting embryos with survivin mRNA could only rescue the vascular defects seen in Sur UTR+ATG MO embryos but not in embryos treated with a VEGF receptor inhibitor. Therefore, additional downstream mediators may be involved in the angiogenic effects of VEGF [17]. Reversely, whether VEGF can rescue the angiogenesis defects in Sur UTR+ATG MO embryos has not been examined. Perturbation of VEGF signaling may also result in changes in blood vessel synthesis and the observed changes in survivin mRNA may reflect changes in endothelial cell number rather than a direct mechanistic link to VEGF signaling. This issue would have to be evaluated in future study.
Both human and murine studies have demonstrated that survivin is involved in haematopoietic stem and progenitor cell proliferation [18,19]. However, in the present study, early specification of hematopoietic progenitors in the Sur UTR+ATG MO embryos was not affected, as shown by the normal expression of genes encoding for hematopoietic transcription factors and embryonic hemoglobins, as well as the normal distribution of gata1 + population in Tg(gata1:GFP) embryos at 18 hpf, before the onset of functional circulation (data not shown).
That the targeting of the survivin MOs was specific was shown using several control studies. First, the phenotypic changes seen in Sur UTR+ATG MO embryos were similar to those observed in survivin mutants generated by retrovirus insertional mutagenesis screening [11]. Indeed, it would be valuable to examine this mutant for similar defects in angiogenesis. Second, co-injecting Sur UTR+ATG -MO with a 5'UTR-survivin:GFP plasmid inhibited translation and hence green fluorescence induced by the latter in all embryos, proving efficacious binding of Sur UTR+ATG -MO to the 5'UTR-survivin mRNA. Third, the angiogenesis defects of ISV induced by Sur UTR+ATG -MO could be rescued by survivin mRNA. Whether the defects in SIV and OV/IOC were similarly reversed and whether injection of survivin mRNA alone would induce angiogenesis defects would have to be further examined. Finally a splice-site morpholino recapitulated the phenotypes seen with Sur UTR+ATG -MO. Therefore, the angiogenesis defects in Sur UTR+ATG MO embryos represent a specific phenotype due to knock-down of survivin function in zebrafish embryos.
In human, murine [4] as well as Xenopus embryos [20], survivin is ubiquitously expressed. These observations, together with the fact that the developing head and eye of the Sur UTR+ATG MO embryos were reduced in size have Effect of survivin knock-down was gene-specific raised a concern whether the angiogenesis defects in these embryos might be caused by a general developmental defect and cell death, rather than a specific requirement for survivin function. However, in zebrafish embryos, survivin is more robustly expressed in the axial vasculature as well as the developing brain and neural tube. This observation has been confirmed by histological sectioning as well as double in-situ hybridization in which the pattern of survivin expression showed remarkable similarity to that of flk1, a marker of vascular endothelium. We cannot exclude a low level but diffuse expression of survivin in adjacent tissues but this should not negate the specific role of survivin during angiogenesis. First, we have confined our examination to the Sur UTR+ATG MO embryos with characteristic phenotypes and none of them showed overt tis-sue necrosis at 48 hpf. The 20% embryos with severe deformity have been excluded from analysis. Second, we demonstrated that embryonic development is not overtly delayed in the Sur UTR+ATG MO embryos as shown by the normal onset and pattern of pigmentation and heart beat (data not shown). Finally, in the present study, survivin expression appeared to be under VEGF regulation, providing a possible link between VEGF and embryonic angiogenesis. Our findings were also consistent with those by Pasquier et al. [20] in which over-expression of survivin in Xenopus embryos induces endothelial cell proliferation in-vivo.
Several observations in this study have remained unexplained. For instance, we did not observe a direct causal link between increased apoptosis and the angiogenesis defect in the Sur UTR+ATG MO embryos. Apoptosis was detectable not only in the axial vasculature, but also in the developing brain and neural tube of the Sur UTR+ATG MO embryos. Both in-vivo and in-vitro studies have demonstrated that in addition to its anti-apoptotic function, survivin plays an important role in the regulation of cellular proliferation and cytokinesis [1,2]. Recent study in Xenopus embryos also showed that survivin expression induces endothelial cell proliferation independent of apoptosis [20]. Therefore, the relative modest TUNEL staining in the axial vasculature did not preclude the role of survivin in angiogenesis. It is also possible that survivin plays a noncell autonomous role in the angiogenesis process. Childs et al. (2002) [21] demonstrated in zebrafish embryos the migration of angioblasts from the aorta to the dorsal aspect of the neural tube and to the inter-phase between notochord and the somites, where they develop into DLAV and ISV. Therefore, vascular patterning may depend on signaling cues that direct the site of angiogenesis sprouts. Whether the occurrence of apoptosis in the developing neural tube might have perturbed these signals hence the formation of DLAV and ISV would have to be carefully examined. The proposition may also explain the lack of robust expression at the site of ISV and DLAV in wild-type embryos. Furthermore, although survivin is expressed robustly in the axial vasculature, concomitant expression was noted in the developing central nervous system. The expression of survivin within these structures needs to be defined in future study. Moreover, the developing eye and head structures in the survivin morphants are generally smaller. Whether this reflected changes secondary to defective angiogenesis or alternative functions of survivin during development have not been elucidated. Finally, survivin gene in zebrafish has undergone duplication during evolution [22] and the function of the duplicated gene would have to be further investigated. Notwithstanding these limitations, our data still supported the proposition that survivin is involved in the regulation of angiogenesis during zebrafish development. When the three groups of data were compared using Kruskal-Wallis Test, p-value = 0.016. When the data of uninjected vs VEGFTKRI treated embryos were compared using Mann-Whitney U Test, p-value = 0.037.
Survivin is strongly expressed in both solid organ and hematological malignancies where it is associated with treatment failure and a poor prognosis [2,3]. Loss of function studies have also demonstrated that survivin expression is linked to angiogenesis and tumorigenesis in gastric and colonic cancers and has become a potential target for anti-cancer therapy [23,24]. Our observation that survivin regulates angiogenesis in zebrafish embryos highlights the relevance of using zebrafish embryos in the screening for survivin-based anti-cancer agents.
In summary, we demonstrate that survivin plays an important role in angiogenesis during embryonic development and may be one of the down-stream effectors of VEGF signaling. Early hematopoiesis was not affected but the role of survivin during late hematopoiesis remains to be determined.

Zebrafish and morpholinos
Wild-type zebrafish (Danio rerio) were obtained from local aquarium and were maintained and raised under standard conditions at 28°C. Transgenic Tg(fli1:EGFP) y1 embryos were used to track endothelial cell populations. Anti-sense morpholinos (MO) (Gene-Tools, OR, USA) were designed to target the 5'untranslated region (UTR) or sequences flanking and including the initiation codon (ATG) of the zebrafish survivin gene. A splice-site (SS) MO was designed to target the exon2-intron junction of the survivin gene (Sur SS -MO). A random sequence MO was used as a control as described previously (Table 1). Procedures for micro-injection, whole mount in-situ hybridization, microangiography, TUNEL and caspase-3 activity assays have been described previously [7,25,26].

Double in-situ hybridization
Wild-type (WT) embryos at 24 hpf were fixed with 4% paraformaldehyde (PFA) and dehydrated. After stepwise re-hydration, the embryos were incubated in pre-hybridization buffer (50% formamide, 5 × SSC, 50 μg/ml heparin, 0.1% Tween20, 5 mg/ml rRNA in phosphatebuffered saline, PBS) at 65°C followed by overnight incubation with digoxigenin (DIG)-(either flk-1 or survivin-1) and fluorescein-labeled riboprobe (α-embryonic globin) at 65°C. The embryos were washed and incubated with alkaline phosphatase (AP) conjugated anti-DIG antibody (Roche Molecular Biochemicals, Mannheim, Germany) for overnight at 4°C. Blue color was developed using NBT/BCIP (Roche Molecular Biochemicals, Mannheim, Germany) as substrate and the reaction was stopped with 0.5 mM EDTA in PBT. AP was destroyed by washing the stained embryos with 0.1 M glycine-HCl, pH 2.2 in PBT for 10 min twice. Background staining was removed by washing the embryos in absolute ethanol with continuous monitoring. After re-hydration to PBT, embryos were incubated with AP conjugated anti-fluorescein antibody