The soybean oil NE (SONE) oil/water (O/W) was prepared using the pseudo-ternary phase diagram, using a low energy method with phase inversion by changing the composition [14], with the following products: Liza^® commercial soybean oil (Cargill, Rio Verde, Goiás, Brazil) (Hydrophile-Lipophile Balance = 7.0), surfactants—Sorbitan Sesquioleate (Span^® 83) (Hydrophile-Lipophile Balance = 3.7) (Lipo do Brasil, São Bernardo do Campo, São Paulo, Brazil) e Polysorbate 80 (Tween^® 80) (Hydrophile-Lipophile Balance = 15.0) (Oxiteno, São Paulo, São Paulo, Brazil), Purified water by reverse osmosis (OS lOLX, Gehaka, São Paulo, São Paulo, Brazil), Phenova^® microbiological preservative (Croda do Brasil, Campinas, São Paulo, Brazil). Initially, the aqueous phase and the oil phase were heated separately to 75˚C ± 2˚C. Then, the oil phase was added slowly to the aqueous phase under constant stirring (IKA RW 20, Staufen, Baden-Württemberg, Germany) until the mixture reaches room temperature (25˚C ± 2˚C). Finally, the microbiological preservative was added [15].

The SONE was evaluated macroscopically 24 hours after preparation, being subjected to centrifugation in triplicate, at the speed of 3000 rpm for 30 minutes to identify possible instability [16]. Particle size and polydispersity index were determined by Dynamic Light Scattering Analyser (Dynamic Light Scattering-DLS) (Zetasizer Nano ZS90, Malvern, United Kingdom) [17].

Subsequently, in order to optimize SONE production, three different agitation speeds were tested on the agitator arm: 600 rpm, 900 rpm and 1200 rpm (IKA RW 20, Staufen, Baden-Württemberg, Germany) and two rods: cog (H1) and helix (H2) (Figure 1) [18]. pH was measured by a phmeter (PG 1800 Gehaka,

São Paulo, São Paulo, Brazil). The index of refraction (IR) was obtained by a refractometer (Abbe, Baush and Lomb Optical Company, New York, United States of America). Electrical conductivity was assessed using the Digimed conductivity meter (DM32, Digimed, São Paulo, São Paulo, Brazil). Finally, the determination of absolute and relative density was performed using the pycnometer (SpLabor, Presidente Prudente, São Paulo, Brazil). All tests were performed in triplicate [19].

Still, the thermal stress was accomplished by heating in a thermostatic bath (NT 281 Nova Técnica^® Ltda, Santa Rosa, Piracicaba, São Paulo, Brazil) at a temperature of 45˚C ± 2˚C to 80˚C ± 2˚C, gradually increasing the temperature every 4˚C ± 2˚C and each value was maintained for 30 minutes, with macroscopic analysis to observe the phase separation at each temperature change. The SONE freeze-thaw cycle was submitted to a temperature of 45˚C ± 5˚C for 24 hours and then to a temperature variation of 4˚C ± 2˚C also for 24 hours, completing a cycle. Macroscopic evaluation and analysis of physical-chemical parameters were performed in the sixth and twelfth cycle [20].

Chicken embryonated eggs (Gallus domesticus) were incubated in an automatic oven at 37˚C and 60% - 70% relative humidity for sixteen days. On the fifth day of incubation, a circular opening in the eggshell was performed in a laminar flow chamber using a micro-rectifier (Dremel MultiPro, Ouro Preto, Minas Gerais, Brazil). Soon after, a drop of 0.9% NaCl was added over the already vascularized chorioallantoic membrane (CAM). The opening was sealed with adhesive tape and a new incubation was carried out.

At the end of the thirteenth day of incubation, filter paper discs were placed directly on the CAM soaked with the test substance: SONE produced, negative control: sterile distilled water (Samtec Biotecnologia, Ribeirão Preto, São Paulo, Brazil), induction control: Regederm^® (Pele Nova Biotecnologia, São Paulo, São Paulo, Brazil) and inhibition control: injectable dexamethasone 4mg/mL (Aché Laboratórios Farmacêuticos S.A., Guarulhos, São Paulo, Brazil). The eggs were returned to the incubation until the sixteenth day [21].

On the sixteenth day of incubation the CAMs were removed, fixed with formaldehyde solution 3.7% for 5 minutes and cut with blunt curved scissors and kept in Petri dishes in the presence of 10% formaldehyde solution to obtain photo registration (640 × 480 pixels; RGB 24 bits) for analysis and quantification of a newly-formed vascular [22].

Acute embryotoxicity tests on zebrafish embryos followed OECD 236 recommendations. Experimentation was performed with embryos from adult fish placed under ideal conditions, 60 males and 20 females in aquariums with water recirculation system (28.5˚C ± 2˚C, 80% humidity) and the photoperiod was adjusted to a 14 hours light/10 hours dark cycle, fed four times a day with commercial floccular feed (FlakesFood^®) and Artemia salina. For experiments, embryos from reproduction were collected, transferred to Petri dishes containing E3 solution and classified aided by stereomicroscope (OLYMPUS CX 31, Shinjuku, Tokyo, Japan) as good, intermediate, bad and non-fertilized, using as standard: cell coloration, disposition and proliferation, in addition to egg fertilization and malformation [23].

After selection, the embryos classified exclusively as good were placed in polypropylene plates, white in color and transparent bottom with 96 wells. Twelve different serial dilutions of SONE were tested at the initial concentration of 124 mg/mL and negative control (E3 solution) in 5 replicates. The solutions were changed daily and kept at room temperature. Embryonic developmental stages were evaluated for: decreased heart rate, yolk sac edema, pericardial edema, spinal cord alteration, hatching rate inhibition, lethal concentration (LC₅₀) and mortality rate through photos and/or videos were obtained in the periods of 0, 48, 72, 96, 120, 144 and 168 hours passed after fertilization (hpf) using the light microscope (LEICA DM750, Leica Microsystems, Rio de Janeiro, Rio de Janeiro, Brazil) coupled to the ICC50 HD digital camera and the LAS® EZ3.0.0 software [24].

For analysis of the preliminary stability tests of SONE, the results were presented as mean (±) standard deviation. The analysis of the variance values was analyzed by repeated (ANOVA) measurements followed by post hoc analysis using a Tukey test with a p < 0.05 significance, using GraphPad Prism, version 7.0 (Graph Pad, San Diego, California, United States of America).

The data on angiogenic activity and embryotoxicity were analyzed using SPSS version 24.0 (IBM SPSS Statistics for Windows, Version 24.0. Armonk, New York, United States of America) and GraphPadPrism version 7.0. Initially, the normality variables of the study were verified by the Shapiro-Wilk test [25]. The parameters measured in angiogenic activity were presented for each group as median and interquartile range (QII). To compare the values found, the Kruskal-Wallis nonparametric test for independent samples was performed, followed by post hoc analysis by Dunn’s test for multiple comparison in case of statistical significance [26] [27]. The parameters evaluated for embryotoxicity were presented as mean and standard error of the mean (SEM). The comparison of the values found in relation to the control group was performed by the Mann-Whitney test for independent samples, p-values (p < 0.05) were considered statistically significant. To determine the LC₅₀, the Probit method was used.

This study was approved by the Research Ethics Committee of the Pontifícia Universidade Católica de Goiás, PUC-Goiás, Brazil (Consubstantiated Report No. 8235150816/2018).

The pseudo-ternary phase diagram guided the preparation of SONE, numbered from one to seven, with each phase described in Table 1.

The seven SONEs produced, initially with an agitation speed of 600 rpm and with a H2 helix, were submitted to preliminary stability assessment by macroscopic analysis and centrifugation. In the macroscopic evaluation after 24 hours (Figure 2), the color was milky for all formulations, there was no bluish reflex and those identified as number three, four and five maintained normal aspects (Table 2). The SONE identified as number five was the most stable after centrifugation, being the only one in which no phase separation was observed and it obtained the smallest particle size (550.2 nm).

Abbreviations: SONE = Soybean oil nanoemulsion, mL= milliliters, Surfactants (Span 83 + Tween 80) used in equal proportions.

Abbreviations: SONE = Soybean oil nanoemulsion, T = Transparent, M = Milky, No = Not observed, N = Normal, SM = Slightly Modified, Mod = Modified.

In order to optimize the preparation process, SONE number five was tested with different speeds in the agitator arm of 600 rpm, 900 rpm and 1200 rpm. A difference in particle size was observed with the use of different speeds, with the lowest speed (600 rpm) inducing the smallest size. A small difference in particle size was observed using different rods in the preparation and, in general, the H2 rods obtained better performance. The polydispersity index of all formulations was equal to or close to 1000 (Figure 3).

The formulation with the smallest particle size, shown in Figure 3(d) for SONE (550.2 nm), was chosen for follow-up and subjected to stability tests and to the freeze-thaw cycle (triplicate) (Table 3), which showed no statistically significant difference in any of the evaluated parameters. In the thermal stress evaluation, the appearance of creaming was observed at 75˚C and after centrifugation, SONE did not present phase separation.

Abbreviations: SONE = Soy oil nanoemulsion, Pt = Parameter evaluated, IR = index of refraction, Absol Dens = Absolute Density, Rel Density = Relative Density, Cond = Electrical conductivity, PS = Particle Size, SD = Standard Deviation, *Comparison of stability between times by Analysis of Variance (ANOVA): pH (p-value = 0.002), absolute density (p-value = 0.065), relative density (p-value = 0.063), conductivity (p-value = 0.069) and PS (p-value = 0.065).

The digital images obtained for the experiment (Figure 4) show, in the CAM treated with inhibition control (injectable dexamethasone), a reduction in the

quantity of vessels while, in the negative control (sterile distilled water), maintenance of the expected quantity of vessels, in the induction control (Regederm^®), an increase in the newly formed vascular network and, for the SONE test substance, a pattern similar to sterile distilled water is noted.

Table 4 and Figure 5 summarize the descriptive and comparative analysis between groups to assess the potential of the SONE test substance for angiogenesis. There was a statistically significant difference between groups for all the parameters evaluated (p-value < 0.001). Length, caliber, junctions, and number of blood vessel complexes were higher in SONE when compared to the inhibition control (Dexamethasone), but lower than in the induction control (Regederm^®), with a statistically significant difference. Still, there was no statistically significant difference between the SONE substance and distilled water for all parameters.

Figure 6 summarizes the assessment of the mortality rate, LC₅₀, heart rate and hatching rate of SONE. There were statistically higher mortality rates in zebrafish embryos exposed to SONE when compared to the control group from the concentration of 0.0193 mg/mL at 24 hpf and 48 hpf and from the concentration of 0.096 mg/mL at 72 hpf. From the concentration of 0.0775 mg/mL, mortality rates of 100% have been verified in the vessels exposed to SONE in all periods of exposure. Thus, it was observed that exposure to SONE from zebrafish embryos caused an increase in the time-concentration dependent mortality rate (Figure 6(a)).

Abbreviations: d.f. = degrees of freedom, *Data presented as median (IQR), **Kruskal-Wallis test for independent samples.

The SONE LC₅₀ varied from 0.0180 mg/mL to 0.0050 mg/mL between 24 hpf and 168 hpf (Δ% = −72.2), with LC₅₀ decreasing statistically with increasing exposure (p-value = 0.046) (Figure 6(b)).

Heart rate decreased significantly with increasing concentration at all exposure times (p-value < 0.05), a result of progressive embryo mortality. The heartbeat of the larvae of the zebrafish embryos exposed to differences in concentrations and periods showed no significant difference with the control group up to the concentration of 0.0048 mg/mL at 96 hpf. From that concentration and persisting until the final concentration (1.24 mg/mL), it was found that the heart rate of the control group was statistically higher than that of SONE at 96 hpf (p-value < 0.05). At 72 hpf and 48 hpf, significant differences in heart rate were verified from the concentrations of 0.0096 mg/mL and 0.0193 mg/mL, respectively (Figure 6(c)).

The analysis of the hatching rate of zebrafish embryos showed that embryos exposed to SONE showed late hatchability until the concentration of 0.0193 mg/mL, while no hatching rate was verified from that concentration, resulting from the mortality of the embryos. At 24 hpf and 48 hpf, no or low hatch rate was observed in all SONE concentrations, which differed statistically from the hatch rate observed in the control group (p-value < 0.05) (Figure 6(d)).

Exposure of zebrafish embryos to different concentrations of SONE induced malformations such as spinal cord alteration, edema of the pericardium and edema of the yolk sac at a rate of less than 10% in all concentrations. However, there was no significant difference in the rate of malformations of embryos exposed to SONE when compared to the control group: rate of pericardial edema, yolk sac edema, and spinal cord alteration.