Long-term effects of somatostatin analogues in rat GH-secreting pituitary tumor cell lines

Purpose First-generation somatostatin analogs, octreotide (OCT) and lanreotide, are the cornerstone for the medical treatment of growth hormone (GH)-secreting pituitary tumors. A new multireceptor analog, such as pasireotide (PAS), showed better activity than OCT in long-term treatment of patients with acromegaly, but modulation of intracellular key processes is still unclear in vitro. In this study, we evaluated the antitumor activity of OCT and PAS in two GH-secreting pituitary tumor cell lines, GH3 and GH4C1, after a long-term incubation. Methods The effects of PAS and OCT on the cell viability, cell cycle, apoptosis, GH secretion, and tumor-induced angiogenesis have been evaluated through a colorimetric method (MTS Assay), DNA flow cytometry with propidium iodide, and Annexin V-FITC/propidium iodide staining, ELISA assay and zebrafish platform, respectively. Results PAS showed a more potent antitumor activity compared to OCT in GH3 cell line exerted through inhibition of cell viability, perturbation of cell cycle progression, and induction of apoptosis after 6 days of incubation. A concomitant decrease in GH secretion has been observed after 2 days of incubation only with PAS. No effect on tumor-induced angiogenesis has been reported after treatment with OCT or PAS in zebrafish/tumor xenograft model. Conclusion Long-term incubation with PAS showed a more potent antitumor activity than that reported after OCT in GH3 cells, mainly modulated by a cell cycle perturbation and a relevant induction in apoptosis.


Introduction
Growth hormone (GH)-secreting pituitary tumors account for about 30% of all functioning pituitary tumors. The excess of GH and insulin-like growth factor 1 (IGF-1) results in a disease known as acromegaly, that is associated with increased morbidity and mortality [1]. First line management for these patients is aimed at normalizing GH and IGF-1 levels, to ameliorate signs and symptoms of this disease and to reduce mortality [2].
Medical therapy is recommended for acromegalic patients who fail to achieve remission after surgery, and for patients who refuse or have contraindications to surgery. GH-secreting pituitary tumors predominantly express somatostatin receptor (SST) -2 and -5 [3]. Somatostatin receptor ligands (SRLs) selective for SST 2 , such as octreotide (OCT) and lanreotide, are the cornerstone for the medical therapy of these tumors [4,5]. Long-term treatment of acromegaly with OCT and lanreotide has been widely studied and showed normalization of GH and IGF-1 levels in about 20-70% and tumor shrinkage in 36-75% of patients [6][7][8][9]. Therefore, a relevant group of patients showed partial or total resistance to SRLs [10]. This phenomenon is probably due to the absence, reduced density, genetic aberration or desensitization of SSTs [11][12][13]. Pasireotide (PAS), a novel SRL with multireceptor-binding profile, has been recently used in the therapy of acromegaly [14]. When compared with OCT, PAS has a higher binding affinity to SST 5 , SST 1 and SST 3 and results in rapid recycling of SST 2 to the plasma membrane after endocytosis [15]. PAS-long-acting release (LAR) showed a better biochemical control rate than OCT or lanreotide in naïve patients with acromegaly or resistant to conventional SRLs [16][17][18][19]. Despite the clinical efficacy of PAS in acromegaly, the antitumor activity of this compound has been studied in vitro on short-term with contradictory effects.
On this basis, we evaluated the antiproliferative, antisecretory, and antiangiogenic activities of OCT and PAS in rat GH-secreting pituitary tumor cell lines (GH3 and GH4C1) after long-term incubation.

Drug preparation and cell line cultures
OCT Acetate and PAS Pamoate were kindly provided by Novartis and diluted in DMSO at a concentration of 10 −3 M. Rat GH-secreting pituitary tumor cell lines, GH3 and GH4C1 were provided by ATCC. GH3 cells were grown at 37 °C in F12 with Kaighn's Modification medium, while GH4C1 in DMEM/F-12 medium, both containing 10% fetal bovine serum, 2 mM glutamine and 10 5 U/l penicillin-streptomycin and maintained in a humidified atmosphere of 5% CO 2 . The cells were grown in 75 cm 2 flasks and passed once every 4-7 days on a 1:2 split. They are characterized to be loosely adherent cells with floating clusters.

RNA isolation
Total RNA was extracted from GH3 and GH4C1 cells with tryzol (Invitrogen, California, USA) according to the manufacturer's instructions. RNA samples were stored at − 80 °C. In each reaction 500 ng of the total RNA was reverse-transcribed into complementary DNA (cDNA) with oligo(dT) primers using GoScript™ Reverse Transcription System (cat. A5000, Promega Corporation, Madison, USA) following the manufacturer's instructions.

Touchdown-polymerase chain reaction (TD-PCR)
TD-PCR was performed for evaluating the expression of SST 1 , SST 2, SST 3 , SST 4 and SST 5 in GH3 and GH4C1 cells. Touchdown PCR conditions for SST 1 and SST 5 consisted in 94 °C for 5 min, a first stage of 10 cycles consisting of a denaturation step of 94 °C for 30 s, an annealing step of 30 s that began at 65 °C and decreased by 0.5 °C per cycle until it reached 60 °C and an elongation step of 72 °C for 30 s, then a second stage of 35 cycles with an annealing temperature of 60 °C followed by a final extension of 72 °C for 7 min. For SST 2 the first stage consisted of an annealing temperature of 65 °C (decreasing by 0.5 °C per cycle until 57 °C) for 16 cycles, followed by the second stage of 25 cycles at 57 °C of annealing. For SST 3 , the first stage consisted of an annealing temperature of 62 °C (decreasing by 0.5 °C per cycle until 54 °C) for 16 cycles, followed by the second stage of 24 cycles at 54 °C of annealing. For SST 4 , finally, the first stage consisted of 6 cycles to decrease the annealing temperature from 53 to 50 °C, while the second stage was composed of 39 cycle at 50 °C. PCR reactions were carried out in a total volume of 25 μL containing 5 µl of 5X reaction buffer with MgCl 2 , 1 µl of 10 mM of dNTPs, 1 µl of 10 pmol/µL of primer forward and reverse each, 1 µL of cDNA sample and 0.25 µl of 5 u/µL GoTaq ® G2 DNA Polymerase (M784B, Promega Corporation, Madison, USA). For SST 1 and SST 4 reaction, 10% of DMSO was also added to the volume. A reaction lacking template was used as negative control. As positive control, PCRs were conducted using genomic DNA extracted from GH3 and GH4C1 using QIAamp DNA Mini Kit (according to manufacturer's instructions), to confirm that the reaction has been set up correctly. PCR products were visualized after 2% agarose gel electrophoresis and Midori Green Advanced (MG04, Nippon Genetics Europe) staining. The sequences of SST 1 , SST 2, SST 3 , SST 4 and SST 5 specific primers and the length of each amplified fragment were as follows: SST 1

Cell viability assay
GH3 and GH4C1 cells were seeded in 96 well plates at a density of 1.5 × 10 4 cells/well. The plates were then placed in a 37 °C, 5% CO 2 incubator. Cell culture medium of both cell lines was replaced the day after with medium containing different concentrations of OCT and PAS (ranging from 10 -11 to 10 −4 M) or the vehicle Dimethyl Sulfoxide (DMSO) as control (CTR) for 3 days. For the experiment of long-term incubation, the medium was replaced with a new one containing drugs or vehicle at the same different concentrations for further 3 days, at the end of which cells were analyzed by a cell viability assay, the CellTiter 96 ® AQueous One Solution Cell Proliferation Assay (MTS, Promega, cat. G3580), according to the manufacturer's instructions.

Analysis of cell cycle and apoptosis by flow cytometry
GH3 and GH4C1 cells were plated in duplicates in six-well plates at a density of 1.5 × 10 5 cells/well. The following day, cell culture medium was replaced with medium containing OCT and PAS or vehicle for 3 days as CTR. Then, the medium was replaced with a new one containing drugs or vehicle at the same different concentrations for further 3 days, at the end of which cells were harvested by gentle trypsinization, washed three times with cold phosphatebuffered saline (PBS), calcium and magnesium-free, and collected by centrifugation at 1200 × g for 5 min.

GH level assay
GH3 cells were plated in duplicates in six-well plates at a density of 1.5 × 10 5 cells/well. The following day and after 24 h from the first treatment, cell culture medium was replaced with medium containing OCT and PAS or vehicle as CTR. After 24 and 48 h from the first treatment, cell culture media were collected and stored at − 80° C until analyzed. Rat GH was measured by a rat/mouse GH ELISA (EMD Millipore, Billerica, Massachusetts, cat. #EZRMGH-45K) according to the manufacturer's procedure.

In vivo zebrafish assay for tumor-induced angiogenesis
Adult zebrafish (Danio rerio) were maintained, according to European laws (2010/63/EU and 86/609/EEC). 48 h postfertilization (hpf) Tg(fli1a:EGFP) y1 transgenic embryos were anesthetized with tricaine (Sigma-Aldrich) and implanted with GH-3 and GH4C1 cells, using a procedure previously described for neuroendocrine tumors [21][22][23]. Briefly, tumor cells were labeled with a red fluorescent viable dye (Cell-TrackerTM CM-DiI, Invitrogen), resuspended with PBS, and grafted into the subperidermal space of Tg(fli1a:EGFP) y1 embryos, close to the sub-intestinal vessels (SIV) plexus. As control of the implantation, we considered embryos injected with only PBS, the cell resuspension solution. This transplantable platform was used to test the effects of SRLs effects on tumor-induced angiogenesis. Before the implantation, tumor cells were pretreated with DMSO vehicle, as CTR, and with 2 × 10 -5 M OCT and PAS for 6 days. After the implantation, DMSO vehicle and SRLs (10 -4 M) were injected into the Cuvier Duct, as previously described [24]. Assays were performed 3 times, considering about 20 embryos in each experimental group. As arbitrary unit (A.U.) of tumor-induced angiogenesis. We calculated by Fiji software the total cumulative length of vessels sprouting from the plexus of subintestinal vessels (SIV) and the common cardinal vein (CCV) in each embryo at 24 and 48 h post implantation (hpi). The average ± S.E.M was statistically compared between the experimental groups with GraphPad Prism 5.0 (GraphPad Software, San Diego, CA).

Statistical analyses
All experiments were carried out at least 3 times and gave comparable results. For statistical analysis, GraphPad Prism 5.0 (GraphPad Software, San Diego, CA) was used for cell viability assay, cell cycle and apoptosis. Half maximal effective concentration (EC 50 ), as an indicator of drug potency, was calculated using nonlinear regression curve-fitting program. The comparative statistical evaluation among groups was first done by Analysis of variance (ANOVA). Statistical comparisons of the logEC50 and maximal inhibitory effect (as an indicator of drug efficacy) were performed with the extra sum-of-squares F test approach (cutoff at p = 0.05). When significant differences were found, a comparison between groups was made using the Newman-Keuls test. The unpaired Student's t test was chosen to analyze the effects of OCT and PAS on GH concentration. In all analyses, values of p < 0.05 were considered statistically significant. The values reported in the figures are the mean ± Standard Error of the Mean (S.E.M).

Expression of SSTs in GH3 and GH4C1 cells
We evaluated the mRNA expression of SST 1 , SST 2 , SST 3 , SST 4 and SST 5 in GH3 and GH4C1 cells by TD-PCR ( Fig. 1). In both cell lines, we observed a strong expression of SST 2 , a moderate expression of SST 1 and SST 3 and a very weak expression of SST 4 subtype transcript, while SST 5 was not detected.

Long-term SRLs treatment decreased viability of rat GH-secreting pituitary tumor cell lines
Dose-response curves showed that both OCT and PAS significantly inhibited the viability of GH3 and GH4C1 cells in a dose-dependent manner (Fig. 2).
In GH4C1 cells, mild and comparable inhibitory effects on cell viability have been observed with both drugs after 3 days (Fig. 2b

Long-term effect of SRLs on cell cycle phases of rat GH-secreting pituitary tumor cell lines
After 6 days of incubation both drugs significantly decreased the percentage of GH3 cells in S phase, (OCT: − 33%, vs control, p < 0.01; PAS: − 42%, vs control, p < 0.01) and increased the number of cells in G 2 /M phase (OCT: + 30%, vs control, p < 0.05; PAS: + 21%, vs control, p < 0.05) (Fig. 3a-c). No statistically significant effect on cell cycle distribution was observed after incubation with both SRLs in GH4C1 cells (Fig. 3d-f).

Long-term effect of SRLs on apoptosis of rat GH-secreting pituitary tumor cell lines
OCT induced a statistically significant increase of GH3 cells in early apoptosis (+ 151% vs untreated cells, p < 0.05) (Fig. 4a). PAS significantly induced a prominent increase of GH3 cells in both early (+ 378% vs untreated cells, p < 0.01) and late apoptosis phase (+ 28% vs untreated cells, p < 0.05) after 6 days of incubation (Fig. 4a, b). Both treatments did not significantly affect necrosis (Fig. 4c). In GH4C1 cells both drugs did not significantly modify the fractions of cells in early apoptosis, late apoptosis, and necrosis compared to controls (Fig. 4d-f).

Modulation of GH release after SRLs exposure
We evaluated the antisecretory activity of OCT and PAS. In GH3 cells, no GH release modulation was observed after 24 h of exposure with both SRLs (Fig. 5a). After 48 h of incubation, only PAS significantly inhibited GH secretion (− 30%, compared to untreated cells, p < 0.05) (Fig. 5b),

SRLs effect on GH3 cell line-induced angiogenesis
To analyze the antiangiogenic potential of OCT and PAS on GH3 and GH4C1 cell lines, we used an innovative in vivo platform, that we have recently developed implanting neuroendocrine tumors cells in Tg(fli1:EGFP) y1 zebrafish embryos [22]. Before the implantation, GH3 or GH4C1 cells were pre-treated in vitro with DMSO (CTR), OCT and PAS for 6 days. These cells were then implanted in 48 h post fertilization (hpf) Tg(fli1:EGFP) y1 embryos into the subperidermal space. After the implantation, DMSO, OCT and PAS were injected into the Cuvier duct. Afterwards, we evaluated the density of tumor-induced endothelial structures around the tumor graft. In our in vivo assays, we did not observe any significant change of tumor-induced angiogenesis after the treatment with OCT and PAS in a temporary window of 24 and 48 hpi for GH3 (Fig. 6) and GH4C1 (Fig. 7) cells.

Discussion
This study evaluated the long-term effects of different SRLs on GH-secreting pituitary tumor cell lines, supporting a more potent anti-tumor effect of PAS than OCT.
SSTs, especially SST 2 and SST 5 , are the main classic targets to inhibit excessive hormone release and cell growth in GH secreting pituitary tumors [25]. The anti-proliferative effects of SRLs in tumors are directly exerted through the induction of apoptosis and cell cycle inhibition, and indirectly through inhibition of angiogenesis and secretion of several growth factors [26]. Although several clinical trials revealed that PAS has a superior efficacy over OCT in patients with acromegaly [17][18][19], there are several contradictory data concerning the antitumor activity and related mechanisms [27][28][29]. In addition, most of the in vitro studies are related to a short-term incubation of GH-secreting pituitary tumor cells with SRLs.
OCT (10 -8 M) exerted a significant, but transient, inhibition of GH3 cell growth with a maximum effect at 24 h, no longer detectable after 48 h [27]. Hubina and coworkers demonstrated that both OCT and PAS decreased GH3 cell proliferation after 72 h incubation time through inhibition of ERK-pathway and an increase in p27 expression at 10 min of exposure [28]. Both SRLs (10 -8 M) showed in vitro comparable inhibition of cell viability after incubation for 24-72 h in primary GH-secreting pituitary tumor cells [29]. These discrepancies between clinical trials and in vitro studies are probably related to differences in both receptor expression pattern and activity of SSTs after interactions with SRLs [30]. The expression of these receptors has been already described in rat GH-secreting pituitary tumor cell lines. SST 1 and SST 2 were the most expressed subtypes in native GH3 cells [31][32][33][34]. Wild-type GH4C1 showed mRNA abundance for SST 1 , SST 2 , SST 3 [30,35]. The high SST 2 expression in rat GH3 cells [31] may explain the receptor desensitization after stimulation [36]. Indeed, PAS modulates SSTs trafficking in a clearly distinct manner from OCT. Lesche and coworkers reported that PAS caused a significantly lower internalization and rapidly recycling to the plasma membrane of SST 2 compared to OCT after endocytosis in HEK 293 cells [15]. Indeed, PAS stimulated only phosphorylation of Ser341 and Ser343 residues of human SST 2 , which is followed by a partial receptor internalization compared to OCT [15,37]. Another study confirmed that the degree of SST 2 internalization by PAS was smaller compared to OCT [38]. In human pancreatic neuroendocrine tumor primary cultures PAS resulted in a rapid and transient internalization of SST 2 followed by persistent recycling of the receptor at the cell surface [39]. While, in GH4C1 cells it has been recently observed that both OCT and PAS (10 −8 M) resulted in a robust internalization of SST 2 and a comparable inhibition of cell proliferation after 48 h [40]. Therefore, a cell and tissue type variability of SST functions and intracellular trafficking may have a role to explain such divergent responses in several studies.
In vitro experiments with long-term incubation should better evaluate the antitumor activity of SRLs. Indeed, this experimental condition is closer to the clinical reality. In the current work, we found only a mild and comparable inhibition of cell viability in GH3 and GH4C1 cells after 3 days of incubation with OCT or PAS and in GH4C1 cells after 6 days. While, in GH3 cells the antitumor activity of PAS was more potent than that of OCT after 6 days. These data were also confirmed after 9 days of incubation (data not shown). We observed a similar SSTs profile in both cell lines, with a strong expression of SST 2 , a moderate expression of SST 1 and SST 3 and a very weak expression of SST 4 subtype transcript. Therefore, we cannot exclude that the differences in the inhibitory effects of SRLs observed between GH3 and GH4C1 cells are probably due to different postreceptor mechanisms. While the stronger inhibition of cell viability observed after 6 days with PAS than OCT in GH3 cells could be related to the differential SST downregulation stimulated by the two drugs. However, to our knowledge, there are no data currently reporting a differential modulation of SST 2 expression after long-term treatment with SRLs.
Direct antitumor effects of SRLs are modulated by the induction of cell cycle lock and apoptosis [41]. It has been already demonstrated that in GH3 cells, OCT had a cytostatic effect by blocking cells in G 0 /G 1 phase after 24 h of incubation [42], through the inhibition of the early response gene c-fos or DNA binding of the heterodimeric transcription factor complex [43]. However, unless OCT was replenished, cell cycle block was transient and overcome by 36-48 h [42]. In addition, both somatostatin-14 and OCT were unable to induce apoptosis in GH3 cells after short-term incubation [42]. On the light of this experimental background, modulation of cell cycle and apoptosis after PAS and after a long-term treatment with SRLs has not been exhaustively documented in GH-secreting tumor cells. After 6 days of incubation, only in GH3 cells, we found that both OCT and PAS induced a comparable decrease of cells in S phase and an increase in G 2 /M phase. Interestingly, after a long-term incubation both SRLs induced apoptosis in only GH3 cells, with a more potent proapoptotic activity after PAS compared to OCT.
The anti-proliferative effects are independent of antisecretory actions of SRLs both in vivo and in vitro [44,45]. Indeed, each SST can have a different effect on the modulation of cell proliferation and GH secretion [46]. OCT (10 -6 and 10 -7 M) reduced GH production after 24 h of incubation of GH3 cell line stimulated by forskolin [47] and after 72 h (10 -8 and 10 -7 M) [48]. GH suppression by OCT (10 -8 M) ranged from 8.5 to 73.7% in GH-secreting primary cells of 24 pituitary tumors from acromegalic patients after 72 h of treatment [49]. A recent critical analysis of preclinical studies comparing the antisecretory activity of PAS vs OCT in somatotroph tumor primary cultures, showed comparable inhibitory effects on GH secretion (incubation time from 4 to 72 h) [50]. An in vitro long-term study on human primary GH secreting pituitary tumor cells found a dose-dependent inhibition of GH release after incubation with OCT for periods ranging from 4 days up to 3 weeks, and a parallel increase in the intracellular GH levels and GH mRNA expression [51]. Due to the low GH production of GH4C1 cells, we evaluated the effects of OCT and PAS on GH release in only GH3 cells conditioned media. For these experiments, we selected a short incubation time, in order to avoid any interference on GH concentrations related to the antiproliferative activity of SRLs. We found a significant decrease in GH secretion after 48 h of incubation only with PAS. At this time, we did not observe any effect on the viability of GH3 cells after PAS or OCT.
Somatostatin and its analogs are also able of inhibiting angiogenesis. SST 1 is highly expressed in vessels, where it inhibits endothelial proliferation, migration, and neovascularization [52]. OCT (10 -10 -10 -6 M) and PAS (10 -9 -10 -6 M) inhibited proliferation of HUVECs, preferentially expressing SST 2 and SST 5 during proliferation, in a dose-dependent manner [53]. SST 3 has been shown to downregulate the transcription of vascular endothelial growth factor (VEGF), which drives the development of new vessels in the growing tumor during hypoxia. The inhibition of endothelial nitric oxide synthase by SST 1 , SST 2 and SST 3 may contribute to the anti-angiogenic activity of SRLs [54]. Vidal and coworkers showed a lower microvascular density in GH-producing tumors treated with OCT than those untreated, although the differences did not reach statistical significance [55]. However, the role of SRLs in modulating tumor-induced angiogenesis is poorly understood. We have recently developed an innovative angiogenesis assay based on the injection of human neuroendocrine tumor cells in transgenic zebrafish embryos [22]. Inoculation of tumor cells in zebrafish embryos can induce a potent angiogenic response through the secretion of several growth factors [22]. VEGF/fibroblast growth factor (FGF) gradient produced by the tumor is able to guide the sprouting of new blood vessels from the close vascular network (SIV and CCV). In our model, implantation of GH3 and GH41C cells in zebrafish embryo significantly stimulated angiogenesis within 24-48 h from engraftment, while long-term pre-incubation with OCT or PAS showed no significant effect on the migration and growth of sprouting vessels toward both tumor implants.
The main limitation of this study is the use of only two cell lines. However, only a few preclinical models of acromegaly are available. GH3 and GH4C1 represent the  (f and g). The red channel was omitted in panels b, b′, d, d′, f and f′ to highlight the tumorinduced microvascular network. Digital magnifications of graft region are shown in white boxed regions b′, d′ and f′. The peritumoral density of endothelial structures, that sprouted from the SIV and CCV and reached the GH-3 tumor mass, did not result in difference in SRL-treated embryos compared to CTR. Here we show the quantification of tumor-induced endothelial structures at both 24 and 48 hpi (h). All images are oriented so that rostral is to the left and dorsal is at the top. Scale bar in a, 100 µm ◂ most widely used GH-secreting pituitary tumor cell lines for the studies of the somatostatin network.
In conclusion, we found that a long-term incubation of GH3 cells with PAS showed a more potent antitumor activity compared to that reported after OCT, while no significant impact has been observed on tumor-induced angiogenesis. This effect is modulated by a cell cycle perturbation and a relevant pro-apoptotic activity.
Acknowledgements Novartis Farma (Origgio, Italy) supplied freely octreotide and pasireotide. We thank Dr. Maurizio Spinello (Novartis Farma) for major contribution to all the administrative and operational aspects.
Funding Open access funding provided by Università degli Studi di Milano within the CRUI-CARE Agreement. This study was supported with an unconditional research grant from Novartis.

Conflict of interest
The authors declare that they have no conflict of interest.
Ethical approval The study was approved by the Institutional Ethics Committee (Approval number: 2017_09_27_04).
Informed Consent For this type of study Informed consent is not required.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. Fig. 7 Effect of treatment with SRLs on GH4C1 cells-induced angiogenesis. Representative epifluorescence images of 48 hpi Tg(fli1:EGFP) y1 zebrafish embryos injected with only PBS (a) or implanted with GH4C1 cells (b-g) and subsequently treated with DMSO vehicle (b and c), OCT (d and e) and PAS (f and g). The red channel was omitted in panels b, b′, d, d′, f and f′ to highlight the tumor-induced microvascular network. Digital magnifications of graft region are showed in white boxed regions b′, d′ and f′. The treatment with SRLs did not reduce the network density of endothelial structures, that sprouted from the SIV and CCV and reached the GH4C1 tumor mass, compared to vehicle-treated CTR embryos. Here we show the quantification of tumor-induced endothelial structures at both 24 and 48 hpi (h). All images are oriented so that the rostral is to the left and dorsal is at the top. Scale bar in a, 100 µm ◂