Optimized Apamin-Mediated Nano-Lipidic Carrier Potentially Enhances the Cytotoxicity of Ellagic Acid against Human Breast Cancer Cells

Ellagic acid has recently attracted increasing attention regarding its role in the prevention and treatment of cancer. Surface functionalized nanocarriers have been recently studied for enhancing cancer cells’ penetration and achieving better tumor-targeted delivery of active ingredients. Therefore, the present work aimed at investigating the potential of APA-functionalized emulsomes (EGA-EML-APA) for enhancing cytototoxic activity of EGA against human breast cancer cells. Phospholipon® 90 G: cholesterol molar ratio (PC: CH; X1, mole/mole), Phospholipon® 90 G: Tristearin weight ratio (PC: TS; X2, w/w) and apamin molar concentration (APA conc.; X3, mM) were considered as independent variables, while vesicle size (VS, Y1, nm) and zeta potential (ZP, Y2, mV) were studied as responses. The optimized formulation with minimized vs. and maximized absolute ZP was predicted successfully utilizing a numerical technique. EGA-EML-APA exhibited a significant cytotoxic effect with an IC50 value of 5.472 ± 0.21 µg/mL compared to the obtained value from the free drug 9.09 ± 0.34 µg/mL. Cell cycle profile showed that the optimized formulation arrested MCF-7 cells at G2/M and S phases. In addition, it showed a significant apoptotic activity against MCF-7 cells by upregulating the expression of p53, bax and casp3 and downregulating bcl2. Furthermore, NF-κB activity was abolished while the expression of TNfα was increased confirming the significant apoptotic effect of EGA-EML-APA. In conclusion, apamin-functionalized emulsomes have been successfully proposed as a potential anti-breast cancer formulation.


Introduction
Breast cancer is considered the most frequent form of malignancy in women and the leading cause of cancer mortality among women in the globe, claiming the lives of 685,000 women in 2020 [1,2]. Breast cancer was diagnosed at an estimated rate of 14 new cases per 100,000 people in women over 20 years old. These data were obtained from the National Institute of Statistics and Geography in 2020 [2]. The treatment of breast cancer is usually a multimodal approach that could possibly include surgery, radiation, chemotherapy and hormonal therapy [3]. Furthermore, the efficacy of the aforementioned applied in targeted drug delivery [33]. It is worth noting that cytotoxic activity has been reported for bee venom components [31].
Therefore, the present work aimed at exploring the potential of APA-functionalized emulsomes (EGA-EML-APA) as a peptide-mediated vesicular delivery approach for enhancing cytotoxic activity of EGA against human breast cancer cells. Central composite response surface design was implemented for optimization of the proposed emulsomal formulation with minimized vesicle size (VS) and maximized absolute zeta potential (ZP). The optimized formulation was tested for cytotoxic and proapoptotic activity against human breast cancer cells.

Results and Discussion
Several nano-sized formulations such as liposomes, nanoemulsions, niosomes, proniosomes, nanoparticles and ethosomes have gained a lot of attention in the field of drug delivery. Nevertheless, emulsomes possess unique characteristics that impart potential advantages to this system compared to others. The most unique characteristic of emulsomes, compared to other vesicular systems including liposomes and niosomes, is the solid fat core enclosed by phospholipid. This allows for enhancing the solubility of poorly soluble drugs and entrapping high amounts of such drugs within the core. The enclosed drug exhibits prolonged drug release and consequently extended efficacy. Second, an emulsomes-based system showed adequate potential for targeting by virtue of their nano-size. Another reported advantage is surpassing the development of multi drug resistance, often associated with over expression of a cell membrane glycoprotein, which causes efflux of the drug from the cytoplasm and results in an ineffective drug concentration inside the cellular compartment [34]. Accordingly, emulsomes were chosen for investigation in this work. In addition, apamin was used for surface-functionalization of the optimized mulsomal formulation to provide additional advantage of enhancing uptake of EGA by cancerous cells [35]. It is worth noting that the reported cytotoxic activity of bee venom components including apamin could also augment the cytotoxicity of EGA [31].

Model Fit Statistics
Fit statistical analysis results for the responses, namely, vs. and ZP are presented in Table 1. Based on the highest R 2 and the least PRESS, the vs. data fitted the quadratic model, while the ZP fitted the two-factor interaction (2FI) model. The adjusted R 2 and the predicted R 2 for each response exhibited appropriate coincidence with a difference of less than the permissible limit of 0.2 verifying the model validity. Moreover, the selected model for each response exhibited adequate precision value higher than the desirable value of 4 indicating appropriate signal-to-noise ratio. According to the previously computed parameters, the selected models could be adequately utilized for the exploration of the experimental design space. Abbreviations: EGA, ellagic acid; EML, emulsomes; APA, apamin; VS, particle size; ZP, zeta potential.
Furthermore, verification of the goodness of fit of selected models was carried out via developing diagnostic plots illustrated in Figures 1 and 2. Box-Cox plot for power transforms, seen in Figures 1A and 2A, show the best lambda (λ) value of 1.52 and 0.87 (represented by the green line) for vs. and absolute ZP, respectively. The computed confidence intervals (represented by the red lines) around these lambdas comprise the current λ value of 1 (represented by the blue line); therein, no specific transformation for observed vs.
is suggested [28]. The computed maximum to minimum measured vs. and absolute ZP ratio of 2.28 and 2.44, respectively, supports the absence for necessity of transformation, where a ratio exceeds 10 calls for a transformation requirement. Studentized residual is a good criterion for identifying potential outliers that could influence the regression model. In our study, the colored dots that symbolize the measured vs. and absolute ZP in the externally studentized residuals vs. predicted response plot, Figures 1B and 2B, shows randomly scattered points within the boundaries implying the absence of constant error. In addition, the colored points in the externally studentized residuals vs. run plots, and Figures 1C and 2C show that no lurking variable could influence the determined responses as evidenced by random scatter of points and absence of trends that could possibly indicate a time-related variable lurking in the background. Additionally, the predicted versus actual plots, illustrated in Figures 1D and 2D showed a highly linear pattern, indicating that the observed responses are analogous to the corresponding predicted values [36,37].

Influence on vs. (Y1)
Invading cancerous tissues is considered a significant problem in the development of formulations for moieties with anticancer activity. This needs ongoing research to alter the features of drug delivery systems in order to improve tumor accessibility [38]. Because of their preferential distribution within solid malignant masses, nanoparticulate formulations of average sizes being less than 400 nm have lately gained an enormous amount of attention in the arena of malignant tumors therapy [39,40]. Despite this discovery, it has been reported that ineffective tumor tissue penetration, resulting from the pathological situation generated by malignancy progression, could overcome the preferential accumulation of nano-sized delivery systems within the malignant tissues and their related therapeutic efficacy [41]. Accordingly, it is possible to verify tumor penetration enhancement by lowering the size to the smallest achievable value, which in turn leads to a higher surface area available for penetrating the tissues [42]. Accordingly, minimized size was set as a goal to increase surface area available for permeation, and thus to ensure effective tumor penetration. Invading cancerous tissues is considered a significant problem in the development of formulations for moieties with anticancer activity. This needs ongoing research to alter the features of drug delivery systems in order to improve tumor accessibility [38]. Because of their preferential distribution within solid malignant masses, nanoparticulate formulations of average sizes being less than 400 nm have lately gained an enormous amount of attention in the arena of malignant tumors therapy [39,40]. Despite this discovery, it has been reported that ineffective tumor tissue penetration, resulting from the pathological situation generated by malignancy progression, could overcome the preferential accumulation of nano-sized delivery systems within the malignant tissues and their related therapeutic efficacy [41]. Accordingly, it is possible to verify tumor penetration enhancement by lowering the size to the smallest achievable value, which in turn leads to a higher surface area available for penetrating the tissues [42]. Accordingly, minimized size was set as a goal to increase surface area available for permeation, and thus to ensure effective tumor penetration.
In our study, the VS of the prepared EML ranged between 263.7 ± 7.9 and 601.8 ± 18.9 nm ( Table 2). On the basis of analysis of variance (ANOVA) provision for VS, the significance of the quadratic model was confirmed as evidenced by the F-value of 265.35 (p = 0.0001). The lack of fit F-value of 6.32 (p = 0.1448) shows a non-significant lack of fit; thus, fitting of the observed vs. to the recommended model is ensured. The equation demonstrating the quadratic model in terms of coded factor was generated by the software is given as: Abbreviations: EGA, ellagic acid; EML, emulsomes; APA, apamin; PC: Phospholipon 90 G; Cholesterol: CH; TS; Tristearin; VS, particle size; ZP, zeta potential; SD, standard deviation. * Data are expressed as mean ± SD of five determinations.
As per ANOVA results, all the linear terms belonging to the analyzed factors had a significant effect on the emulsomal vs. (p < 0.0001 for all terms). Furthermore, the interaction terms as well as the quadratic terms were also significant at the 95% level of significance. The perturbation graph, displayed in Figure 3A, demonstrates the influence of the investigated factors on the VS, while the 3D-response and the 2D-contour plots, displayed in Figure 4, demonstrate the interaction between them. The illustrations show that the EML vs. significantly increases at higher PC: CH molar concentration, PC: TS weight ratio and APA molar concentration. Clearly, the positive sign associated with the linear term coefficients; X 1 , X 2 and X 3 supports this finding. The most influential factor was PC: CH molar ratio as indicated by the highest coefficient of its linear term X 1 in the developed equation, followed by PC: CH molar ratio, then the APA molar concentration.
The effect of PC: CH ratio could be attributed to the possible creation of numerous bilayers, which results in greater vs. [25]. Nonetheless, a comparable increase in vs. of zidovudine EML was previously reported upon increasing the phospholipid relative to the solid lipid and cholesterol. On the other hand, the VS was inversely proportional to the level of cholesterol. In fact, given that, at increased cholesterol levels, the EML acquires a stronger lipophilic property, which encourages inhibition of water uptake across the lipid bilayer; this explains the observed reduction in VS. This observation is consistent with Sudhakar and Chaitanya's [43] work, who found an inverse association between vs. and cholesterol level of ritonavir liposomes.
ables. The illustrations reveal higher ZP absolute values, with considerable increase in the negativity of the emulsomal surface, which were observed when the PC: CH molar ratio was increased. On the other hand, the absolute ZP decreases at higher APA concentrations. The positive sign of X1 and the negative sign of X3 coefficients support this observation. The observed interlink between increased negativity with the concomitant increase in PC: CH molar ratio could probably be a result of an increase in the proportion of a negatively charged phospholipid in the outer layers [3,21]. On the contrary, the observed reduced negativity at higher APA molar concentration could be credited to the possible interaction between the positively charged lysine moieties present in APA structure with the negatively charged phospholipid of the emulsomal vesicles.

Influence on Zeta Potential (Y 2 )
The surface charge of the vesicles is described by ZP, which reflects their physical stability, where higher surface charge results in electrostatic repulsive forces that prevent particles coalescence and aggregation. Generally, nano-formulations with ZP values of about or greater than ±30 mV are considered stable [23]. The prepared EGA-EML-APA exhibited negative ZP values ranging from −18.9 ± 0.6 to −46.1 ± 1.9 mV, Table 2. The negative charge could be credited to the negatively charged phospholipid that forms the lipid bilayer surrounding emulsomal core. It is worth noting that, although a negatively charged surface might slightly affect the cellular internalization, the stability of nano-systems against aggregation is crucial for benefiting from the nano-size. In addition, many studies have reported enhanced anticancer activity for negatively charged nanoformulation [44][45][46][47]. Accordingly, the study aimed at maximizing the absolute ZP value.
Based on Analysis of variance (ANOVA) for absolute ZP, the significance of the twofactor interaction model was confirmed, as depicted by the corresponding F-value of 44.69 (p < 0.0001). The lack of fit F-value of 4.37 (p = 0.1997) reflects non-significant lack of fit; thus, the fitting of absolute ZP values to the proposed model is confirmed. The equation generated by the software demonstrates the suggested model for the ZP in terms of coded factor as given below : Y 2 (zeta potential) = 29.03 + 1.72 X 1 + 0.84 X 2 − 6.70 X 3 + 1.75 The statistical analysis demonstrated that the linear terms X 1 and X 3 corresponding to PC: CH molar ratio and APA molar concentration exhibited a significant effect on absolute ZP (p = 0.0057 for X 1 and p < 0.0001 for X 3 ). Additionally, all the interaction terms were significant at p < 0.05. Figure 3B depicts the perturbation graph that demonstrates the main effects of the investigated variables on the absolute ZP, while Figure 5 illustrates the 3Dsurface plots and the 2D contour plots that show the interaction between the variables. The illustrations reveal higher ZP absolute values, with considerable increase in the negativity of the emulsomal surface, which were observed when the PC: CH molar ratio was increased. On the other hand, the absolute ZP decreases at higher APA concentrations. The positive sign of X 1 and the negative sign of X 3 coefficients support this observation. The observed interlink between increased negativity with the concomitant increase in PC: CH molar ratio could probably be a result of an increase in the proportion of a negatively charged phospholipid in the outer layers [3,21]. On the contrary, the observed reduced negativity at higher APA molar concentration could be credited to the possible interaction between the positively charged lysine moieties present in APA structure with the negatively charged phospholipid of the emulsomal vesicles.

Optimization of EGA-EML-APA
Numerical optimization and the desirability techniques were implemented for predicting the optimized variables' levels that, upon combination, could result in minimized vs. and maximized absolute ZP. The ramp graphs presented in Figure 6A depict the optimized levels and the predicted responses, while the desirability for each response and the overall desirability are graphically illustrated in Figure 6B. The measured vs. of 267.6 nm and absolute ZP of 32.1 mV coincide well with the predicted ones showing relative percentage error of 1.47% and 4.17%, respectively. Furthermore, the measured vs. is less than 400 nm indicating adequate preferential distribution within solid cancerous tissues [39,40], and the absolute ZP is greater than 30 mV, indicating adequate repulsive forces that guard against aggregation [23]. The relatively small computed percentage errors prove the reliability of the optimization process. It is worth noting that the PS and ZP values of the optimized formulation are in good agreement with those reported previously for other nanocarriers with enhanced cytotoxicity against various cancer cells. Barani et al. reported enhanced anticancer activity for paclitaxel loaded niosomes with an average size of 240 nm [48]. The reported size was sufficient for tumor-specific accumulation. Abbas et al. reported an enhanced cytotoxic effect for optimized curcumin bilosomes with maximized negative absolute zeta potential of −27.05 mV [46]. On the other hand, previous studies that involve vesicles with higher size in the micro range exhibited low cytotoxicity profiles on different cell lines. Lai et al. developed oleic acid containing vesicles with an average size of 2.35 ± 0.9 µm for transdermal delivery [49]. Cytotoxicity studies of such vesicles in various skin cell lines demonstrated no loss of cell viability in all concentrations indicating high safety profile and appropriateness for transdermal use with no side effects on the skin. However, the diminished cytotoxicity of the micro-sized vesicles could indicate poor ability to accumulate within the cells, thus highlighting the benefit of the nano-size in our study. This indicates the importance of developing a carrier to improve delivery of loaded active ingredients [50].

Characterization of Optimized EGA-EML-APA
The shape of the optimized EGA-EML-APA was visualized using TEM as illustrated in Figure 7. The TEM micrograph reveals spherical vesicles that possess even rounded contours and adequate uniform distribution. The average size of the vesicles is in good agreement with that measured by the dynamic light scattering technique. Zhou and Chen [23] reported a similar spherical shape for silybin nanoemulsomes.
The optimized formulation showed high entrapment efficiency of 93.7 ± 4.1%. The high entrapment ability of the proposed formulation could be due to the solid lipid core enclosed by the lipid bilayer which aids in increasing the entrapping of poorly soluble active ingredients [34]. It is worth noting that this feature contributes to the superiority of emulsomes over other vesicular carriers including niosomes; for example, pH-responsive niosomes showed entrapment of less than 80% for paclitaxel in a previous study [48].

EGA-EML-APA Inhibited the Viability of MCF-7 Cells
Breast cancer is the most commonly occurring cancer in women and the second most common cancer overall, specifically estrogen receptor positive (ER) breast cancer. Ellagic acid has been demonstrated to have antitumor and anti-apoptotic effects in several cancer cells [8]. However, in MCF-7, the anti-proliferative effect of ellagic acid was observed in only high concentrations [6,51,52]. Therefore, MCF-7 cells were selected to examine the anti-proliferative effect of the optimized EGA-EML-APA formula. EGA-EML-APA was found to significantly inhibit MCF-7 cell viability, and in a manner that seems to be dose-dependent, more than EGA. As shown in Figure 8, the IC 50 value for EGA-EML-APA in MCF-7 was 5.472 ± 0.21 µg/mL compared to 9.09 ± 0.34 µg/mL for EGA. Furthermore, there was no effective cytotoxic effect of blank EML-APA when compared with the EGA and EGA-EML-APA in MCF-7 cells. The observed cytotoxic activity of the EGA-EML-APA is in line with the previously reported cytotoxicity of EGA against breast cancer cells including MCF-7 cells [6,53]. However, EGA-EML-APA showed significantly improved cytotoxic activity against MCF-7 cells compared to EGA. This promising effect may be to the result of the enhanced cellular penetration and accumulation of EGA in MCF-7 cells by virtue of the optimized formulation. The ability of EML to enhance cellular uptake of actives has been previously reported [54][55][56].

Characterization of Optimized EGA-EML-APA
The shape of the optimized EGA-EML-APA was visualized using TEM as illustrated in Figure 7. The TEM micrograph reveals spherical vesicles that possess even rounded contours and adequate uniform distribution. The average size of the vesicles is in good agreement with that measured by the dynamic light scattering technique. Zhou and Chen [23] reported a similar spherical shape for silybin nanoemulsomes.
The optimized formulation showed high entrapment efficiency of 93.7 ± 4.1%. The high entrapment ability of the proposed formulation could be due to the solid lipid core enclosed by the lipid bilayer which aids in increasing the entrapping of poorly soluble active ingredients [34]. It is worth noting that this feature contributes to the superiority of emulsomes over other vesicular carriers including niosomes; for example, pH-responsive niosomes showed entrapment of less than 80% for paclitaxel in a previous study [48].

EGA-EML-APA Inhibited the Viability of MCF-7 Cells
Breast cancer is the most commonly occurring cancer in women and the second most common cancer overall, specifically estrogen receptor positive (ER) breast cancer. Ellagic acid has been demonstrated to have antitumor and anti-apoptotic effects in several cancer cells [8]. However, in MCF-7, the anti-proliferative effect of ellagic acid was observed in only high concentrations [6,51,52]. Therefore, MCF-7 cells were selected to examine the anti-proliferative effect of the optimized EGA-EML-APA formula. EGA-EML-APA was found to significantly inhibit MCF-7 cell viability, and in a manner that seems to be dosedependent, more than EGA. As shown in Figure 8, the IC50 value for EGA-EML-APA in MCF-7 was 5.472 ± 0.21 µg/mL compared to 9.09 ± 0.34 µg/mL for EGA. Furthermore, there was no effective cytotoxic effect of blank EML-APA when compared with the EGA and EGA-EML-APA in MCF-7 cells. The observed cytotoxic activity of the EGA-EML-APA is in line with the previously reported cytotoxicity of EGA against breast cancer cells including MCF-7 cells [6,53]. However, EGA-EML-APA showed significantly improved cytotoxic activity against MCF-7 cells compared to EGA. This promising effect may be to the result of the enhanced cellular penetration and accumulation of EGA in MCF-7 cells by virtue of the optimized formulation. The ability of EML to enhance cellular uptake of actives has been previously reported [54][55][56].

EGA-EML-APA Modulated MCF-7 Cell Cycle
To determine the mechanism by which EGA-EML-APA exerts its cytotoxic influence against MCF-7 cells, the effect on cell cycle phases was investigated. While untreated MCF-7 cells exhibited quick growth properties, treatment of the cells with the EGA, and EGA-EML-APA resulted in altered cell cycle progression and increased cell fraction in the pre-G1 phase ( Figure 9C-E). The percentages of cells accumulated in the pre-G1 phase were 21.57 ± 0.69% and 28.42 ± 0.94% when cells were treated with EGA and EGA-EML-APA, respectively. As can be seen from Figure 9E, the most significant increase in cell fraction of the pre-G1 phase was associated with EGA-PHM-APA. Additionally, EGA-EML-APA caused a marked reduction in cell population in G0/G1 phase in comparison with other groups (p < 0.05) ( Figure 9C,E). Furthermore, the fractions of cell in G2/M and S phases significantly increased when compared to the control and EGA (p < 0.05). These

EGA-EML-APA Modulated MCF-7 Cell Cycle
To determine the mechanism by which EGA-EML-APA exerts its cytotoxic influence against MCF-7 cells, the effect on cell cycle phases was investigated. While untreated MCF-7 cells exhibited quick growth properties, treatment of the cells with the EGA, and EGA-EML-APA resulted in altered cell cycle progression and increased cell fraction in the pre-G1 phase ( Figure 9C-E). The percentages of cells accumulated in the pre-G1 phase were 21.57 ± 0.69% and 28.42 ± 0.94% when cells were treated with EGA and EGA-EML-APA, respectively. As can be seen from Figure 9E, the most significant increase in cell fraction of the pre-G1 phase was associated with EGA-PHM-APA. Additionally, EGA-EML-APA caused a marked reduction in cell population in G0/G1 phase in comparison with other groups (p < 0.05) ( Figure 9C,E). Furthermore, the fractions of cell in G2/M and S phases significantly increased when compared to the control and EGA (p < 0.05). These data suggest that EGA-EML-APA caused sequential cell cycle arrest at G2/M and S phases, followed by apoptosis, which is represented by the increase in apoptotic pre-G1 population. A previous report demonstrated a comparable ability of EGA to interfere with the cell cycle especially by increasing the population of cells at the pre-G1 phase [57]. However, treatment with EGA-EML-APA resulted in a significant increase in the S, G2-M and pre-G1 phases compared to EGA. Hence, the abovementioned results highlight the improved cytotoxic activity of the optimized formulation.

EGA-EML-APA Induced Apoptosis in MCF-7 Cells
The increased cellular fraction in the pre-G1 phase strongly indicates that the optimized EGA-EML-APA exhibited a significant apoptotic activity. Thus, annexin-V staining was utilized to determine the percentages of apoptotic cells associated with blank EML-APA, EGA and EGA-EML-APA. As can be seen from Figure 10B,C,E, treating MCF-7 cells with EGA-EML-APA formula resulted in the most significant increase in early, late and

EGA-EML-APA Induced Apoptosis in MCF-7 Cells
The increased cellular fraction in the pre-G1 phase strongly indicates that the optimized EGA-EML-APA exhibited a significant apoptotic activity. Thus, annexin-V staining was utilized to determine the percentages of apoptotic cells associated with blank EML-APA, EGA and EGA-EML-APA. As can be seen from Figure 10B,C,E, treating MCF-7 cells with EGA-EML-APA formula resulted in the most significant increase in early, late and total apoptotic cell death when compared to EGA (p < 0.05). Necrotic cell death showed a similar pattern in which the most significant increase was associated with EGA-EML-APA ( Figure 10C,E). This apoptotic effect produced by EGA-EML-APA could be due to the ability of EGA to alter the expression of cellular components involved in apoptosis and cellcycle regulation. Therefore, the expression of apoptosis markers involved in the intrinsic mitochondrial mediated apoptosis pathway such as bax, bcl-2, p53 and casp-3 is investigated. total apoptotic cell death when compared to EGA (p <0.05). Necrotic cell death showed similar pattern in which the most significant increase was associated with EGA-EML-AP ( Figure 10C,E). This apoptotic effect produced by EGA-EML-APA could be due to th ability of EGA to alter the expression of cellular components involved in apoptosis an cell-cycle regulation. Therefore, the expression of apoptosis markers involved in the i trinsic mitochondrial mediated apoptosis pathway such as bax, bcl-2, p53 and casp-3 is i vestigated.

EGA-EML-APA Apoptotic Effect Evidenced by MMP Loss and Apoptotic Markers
Upon exposure to apoptotic stimuli, damage to the mitochondrial membrane might occur, resulting in the loss of MMP [58]. Hence, the change in MMP was measured in MCF-7 control cells, and after adding blank EML-APA, EGA or EGA-EML-APA. As shown in Figure 11A, MMP was significantly compromised only when the cells were treated with the optimized EGA-EML-APA (p < 0.05), leading to a significant loss of MMP in comparison to raw EGA. Thus, an enhanced efficacy with EGA-EML-APA could be obviously shown and can be attributed to the lipophilic nature of the delivery system, which plays a major role in enhancing the delivery of EGA incorporated into this novel nanocarrier platform [59].
The effect of ellagic acid through the intrinsic apoptotic pathway in MCF-7 cells was successfully reported in several works [51,52,60]. The expression of multiple markers of this pathway including Bax, Bcl2 and casp3 was also examined to confirm the apoptotic activity of EGA-EML-APA. As depicted in Figure 11B, Bax mRNA was significantly higher in EGA-EML-APA-treated cells relative to EGA. In contrast, EGA-EML-APA was associated with the lowest expression level of Bcl2, Figure 11C. Additionally, MCF-7 cells treated with EGA-EML-APA exhibited the most significant increase in casp3 expression compared to EGA treatment (p < 0.05), Figure 11D. These findings highlight the superior activity of the EGA-EML-APA in apoptosis induction relative to EGA, considering that increased Bax levels and decreased Bcl-2 levels favor apoptosis [61]. It was reported that EGA exhibits potent apoptotic activities by enhancing the expression of casp3, which is an apoptotic transcription factor [62]. In this regard, EGA-EML-APA treatment increased the expression of p53 by 8-fold compared to 5-fold increases associated with EGA ( Figure  11E). This further confirms that EGA-EML-APA significantly improved the cytotoxicity and apoptotic activity EGA.

EGA-PHM-APA Induced Changes in the Expression of tnf-α and Nf-Kb
TNF-α has been utilized as an inflammatory marker in wound healing studies [63]. It has also been explored as an apoptotic marker; it is reported that TNF-α can induce apoptosis in breast cancer cells, and this process can be inhibited upon NF-кB activation [64,65]. As shown in Figure 12A, greater levels of TNF-α were detected with EGA-EML- The effect of ellagic acid through the intrinsic apoptotic pathway in MCF-7 cells was successfully reported in several works [51,52,60]. The expression of multiple markers of this pathway including Bax, Bcl2 and casp3 was also examined to confirm the apoptotic activity of EGA-EML-APA. As depicted in Figure 11B, Bax mRNA was significantly higher in EGA-EML-APA-treated cells relative to EGA. In contrast, EGA-EML-APA was associated with the lowest expression level of Bcl2, Figure 11C. Additionally, MCF-7 cells treated with EGA-EML-APA exhibited the most significant increase in casp3 expression compared to EGA treatment (p < 0.05), Figure 11D. These findings highlight the superior activity of the EGA-EML-APA in apoptosis induction relative to EGA, considering that increased Bax levels and decreased Bcl-2 levels favor apoptosis [61]. It was reported that EGA exhibits potent apoptotic activities by enhancing the expression of casp3, which is an apoptotic transcription factor [62]. In this regard, EGA-EML-APA treatment increased the expression of p53 by 8-fold compared to 5-fold increases associated with EGA ( Figure 11E). This further confirms that EGA-EML-APA significantly improved the cytotoxicity and apoptotic activity EGA.

EGA-PHM-APA Induced Changes in the Expression of tnf-α and Nf-κb
TNF-α has been utilized as an inflammatory marker in wound healing studies [63]. It has also been explored as an apoptotic marker; it is reported that TNF-α can induce apoptosis in breast cancer cells, and this process can be inhibited upon NF-κB activation [64,65]. As shown in Figure 12A, greater levels of TNF-α were detected with EGA-EML-APA treatment. In addition, EGA-EML-APA produced the most significant decrease in NF-κB activation (p < 0.05) ( Figure 12B). Therefore, these findings further substantiate the superior cytotoxic activity of EGA-EML-APA compared to the blank EML and EGA. APA treatment. In addition, EGA-EML-APA produced the most significant decrease in NF-кB activation (p < 0.05) ( Figure 12B). Therefore, these findings further substantiate the superior cytotoxic activity of EGA-EML-APA compared to the blank EML and EGA.

Response Surface Methodology for Formulation of EGA-EML-APA
Response surface methodology was implemented for the formulation and optimization of EGA-EML-APA. Specifically, three-level, three-factor, face-centred central composite experimental design (α = 1) was utilized in this study. Phospholipon ® 90 G: cholesterol molar ratio (PC: CH; X1, mole/mole), Phospholipon ® 90 G: tristearin weight ratio (PC: TS; X2, w/w) and APA molar concentration (APA conc.; X3, mM) were considered as numerical independent variables. Vesicle size (VS, Y1, nm) and zeta potential (ZP, Y2, mV) were investigated as responses. The coded levels of the independent variables, represented as -1, 0, +1 analogous to lower, middle and upper values respectively, along with the corresponding actual values for each variable are listed in Table 3. Design-Expert software (Version 12; Stat-Ease Inc., Minneapolis, MN, USA) was used to generate the experimental runs. As per the selected design, 17 runs were generated including factorial points, axial points and three replicated centre points to give appropriate prediction ability close to the centre of the variable space; the variables' levels represented in each run are compiled in Table 2. The sequential model that best fits the data of each response was determined according to the predicted and adjusted determination coefficients (R 2 ), as well as the predicted residual error sum of squares (PRESS) statistics. The equations expressing the optimal model for the responses were then developed by the software. The significance of the variables on the measured responses was analyzed using analysis of variance (ANOVA) at p < 0.05. Perturbation and response surface plots were developed to display the influence of the explored factors as well as their interactions.

Response Surface Methodology for Formulation of EGA-EML-APA
Response surface methodology was implemented for the formulation and optimization of EGA-EML-APA. Specifically, three-level, three-factor, face-centred central composite experimental design (α = 1) was utilized in this study. Phospholipon ® 90 G: cholesterol molar ratio (PC: CH; X 1 , mole/mole), Phospholipon ® 90 G: tristearin weight ratio (PC: TS; X 2 , w/w) and APA molar concentration (APA conc.; X 3 , mM) were considered as numerical independent variables. Vesicle size (VS, Y 1 , nm) and zeta potential (ZP, Y 2 , mV) were investigated as responses. The coded levels of the independent variables, represented as -1, 0, +1 analogous to lower, middle and upper values respectively, along with the corresponding actual values for each variable are listed in Table 3. Design-Expert software (Version 12; Stat-Ease Inc., Minneapolis, MN, USA) was used to generate the experimental runs. As per the selected design, 17 runs were generated including factorial points, axial points and three replicated centre points to give appropriate prediction ability close to the centre of the variable space; the variables' levels represented in each run are compiled in Table 2. The sequential model that best fits the data of each response was determined according to the predicted and adjusted determination coefficients (R 2 ), as well as the predicted residual error sum of squares (PRESS) statistics. The equations expressing the optimal model for the responses were then developed by the software. The significance of the variables on the measured responses was analyzed using analysis of variance (ANOVA) at p < 0.05. Perturbation and response surface plots were developed to display the influence of the explored factors as well as their interactions.

Optimization of EGA-EML-APA
A numerical optimization technique following a desirability approach was used for prediction of the optimized EGA-EML-APA formulation; the desirability function amalgamates both responses aiming at anticipating the optimized levels of investigated factors [66,67]. The objective of the study was achieving minimized vs. and maximized ZP absolute value as shown in Table 1.

Preparation of EGA-EML-APA
A previously reported method was utilized for the preparation of EGA-EML-APA [26]. Accurately weighed quantities of EGA (50 mg), Phospholipon ® 90 G, cholesterol and tristearin were solubilized in 15 mL of chloroform/methanol blend (2:1, v/v). Evaporation of the solvent blend was done under reduced pressure using Rotavapor (BÜCHI Labortechnik AG, Flawil, Switzerland) at 40 • C. By placing the formed film in a vacuum oven for 24 h, any organic solvent residual was eliminated. Hydration of the films was then achieved by mild agitation with 10 mL distilled water containing APA for 3 h at room temperature. The resultant dispersion was ultra-sonicated for two cycles, each of 45 s with a time gap of two minutes for vs. reduction [25,68].

Vesicle Size and Zeta Potential Determination
Mean vs. (z-average) as well as ZP of the proposed EGA-EML-APA were determined by light scattering and electrophoretic techniques, respectively, using Nano ZSP (Malvern Panalytical, UK) at 25 ± 1 • C. Adequate dilution with distilled water was done prior to measurement. Data were expressed as the mean of five determinations.

Characterization of Optimized EGA-EML-APA
The optimized EGA-EML-APA were visualized using a JEOL GEM-1010 (JEOL Ltd., Akishima, Tokyo, Japan) transmission electron microscope (TEM) at 80 kV. One drop of diluted formulation was placed on a carbon-coated grid, which was then left to dry at temperature of 25 ± 0.5 • C. Furthermore, the sample was negatively stained with 1% phosphotungstic acid and then dried for 20 min at room temperature before being visualized.
Entrapment efficiency of the optimized formulation was performed by an indirect method. In addition, 1 mL of emulsomes dispersion was ultracentrifuged at 100,000 rpm for 1 h at 4 • C. The residue was washed twice with phosphate buffer (pH 6.8) and subjected to re-centrifugation for 1 h. The combined supernatant was diluted with phosphate buffer (pH 6.8), and analyzed using the reported HPLC method of assay with UV detection at 254 nm. The procedure was repeated thrice [15].
The entrapment efficiency was calculated using the following equation: where EGA t represents the amount of total ellagic acid; and EGA u represents the amount of unentrapped ellagic acid.

Cytotoxicity Assay
As directed by the manufacturer, the cytotoxic effect of EGA, blank EML-APA and the optimized EGA-EML-APA on MCF-7 cells was evaluated using the MTT viability test (ABCAM, Cambridge, UK). The cells were grown overnight at a density of 5 × 10 3 cells per well in 96-well plates and further treated for 24 h with each treatment in a concentration range of 0.4 to 100 µg/mL. EGA was dissolved in DMSO (final concentration 0.1% w/v). The culture media were then aspirated and replaced with MTT solution at a final concentration of 2 mg/mL. The cells were subsequently incubated for 3 h at 37 • C. Thereafter, removal of the MTT solution was performed, then 100 µL of DMSO was added in order to dissolve the formazan crystals. Measurement of the absorbance was then carried out at a 570 nm wavelength using an Absorbance-based Spark ® multimode microplate reader (Tecan Group Ltd., Seestrasse, Maennedorf, Switzerland). The results were represented as a percentage of cell viability relative to the control. Fitting of the dose response curves followed by calculating IC50 values was undertaken using GraphPad prism software (GraphPad, Inc., La Jolla, CA, USA).

Cell Cycle Analysis
Flow cytometry analysis was applied to analyze the effects of EGA, blank EML-APA, and the optimized EGA-EML-APA on MCF-7 cell cycle phases using propidium iodide (PI) Flow Cytometry Kit (ab139418, Abcam, Cambridge, UK). In addition, 3 × 10 5 cells/well were seeded in a 6-well plate prior to their incubation for 48 h with each treatment. Subsequently, the cells were rinsed, trypsinized and then centrifuged for 15 min at 1000× g. The cells were then fixed using 70% ethanol and incubated at 4 • C for 2 h. The fixed cells were then centrifuged at 1000× g for 15 min prior to staining with PI (10 g/mL) and RNase treatment. A flow cytometer (FACScalibur, BD Bioscience, San Jose, CA, USA) was used to assess cell DNA content, with a minimum of 20,000 events obtained for each treatment. Analysis of the data was done using the CellQuest Software (Becton-Dickinson, BD, Erembodegem, Belgium).

Apoptosis Assay
The Annexin V-FITC Apoptosis Detection Kit (BD Bioscience, San Jose, CA, USA) was used to investigate the effects of EGA, blank EML-APA and the optimized EGA-EML-APA on the apoptosis profile of MCF-7 cells. Briefly, in a 6-well plate, cells were plated at a density of 1 × 10 6 cells per well and incubated for 48 h with each treatment. After that, the cells were trypsinized and subjected to centrifugation at 10,000× g for 5 min. The supernatant was then discarded; and the cells were rinsed in PBS and stained in dark conditions for 30 min using annexin V-FITC/PI dyes, as directed by the manufacturer. FACScalibur (BD Bioscience, USA) was used to evaluate stained cells, and a minimum of 20,000 events were obtained for each treatment.

Mitochondrial Membrane Potential (MMP)
The determination of MMP was carried out as per the manufacturer's directions using the MitoProbeTM TMRM Assay Kit for Flow Cytometry (Thermo Fisher Scientific, CA, USA). MCF-7 cells culture was performed at a density of 1 × 10 5 cells per well in 96-well plates. Cells were stained using 20 M tetramethylrhodamine, methyl ester (TMRM) following 48 h of incubation with EGA, blank EML-APA, or optimized EGA-EML-APA and then incubated for a period of 30 min at 37 • C. TMRM is a fluorescent, cell-permeant dye that creates a strong bright signal upon its accumulation in active healthy mitochondria of cells. TMRM buildup ends and the signal disappearing indicates that the mitochondrial membrane potential is lost during apoptosis. The cells were subsequently washed in PBS, loaded with live-cell imaging buffer and then analysed using flow cytometry (FACSCalibur, BD Bioscience, USA).

Real-Time Polymerase Chain Reaction (RT-qPCR) RNA Extraction
Qiagen's RNeasy mini kit (Qiagen, UK) was used to extract RNA from MCF-7 cells as per the manufacturer's instructions. A Nanodrop spectrophotometer (ND-2000C, Thermoscientific) was used to confirm the concentration and purity of RNA. A ratio of A260 nm/A230 nm, not smaller than 1.8, and a A260 nm/A280 nm ratio, not smaller than 1.9, were detected in all RNA samples.
cDNA Synthesis and PCR Amplification RNA was normalized between samples and reverse transcribed to complementary DNA (cDNA) as per the manufacturer's instructions using the iScriptTM One-Step RT-PCR Kit With SYBR ® Green kit (BioRad, Hercules, CA, USA). Relative expression patterns of p53, bcl2, bax casp3, tnfα and nf-κb were done utilizing 10 ng of RNA template in a 50-mL reaction mixture of iScript one-step RT-PCR kit with SYBR ® Green mix using a 7500 Fast real-time PCR system (Applied Biosystems, Waltham, MA, USA).

Conclusions
Face-centred central composite design has been successfully employed for optimizing EGA-EML-APA with minimized vs. and maximized absolute ZP. The use of APA in formulating the emulsomes was proposed to provide additional advantage of enhancing uptake of EGA by cancerous cells, in addition to the cytotoxic activity of bee venom components including apamin. The measured responses of the optimized formulation were 267.6 nm for vs. and −32.1 mV for ZP. The measured responses coincide well with the predicted ones, confirming the validity of the numerical optimization adopted in this study. The present work confirmed the increased percentage of cytotoxic activity and apoptosis of EGA against the human breast cancer cell line (MCF-7) after the formulation EML-APA. This was confirmed by observations in cell viability (MTT assay) as well as the changes in the expressions of apoptotic markers and inhibiting NF-κB activity. Arresting G2/M and S cell cycle phases by EGA-EML-APA is a superior finding with respect to the mechanism by which the new formulation is inhibiting MCF-7 cell proliferation. These data demonstrate that the proposed formulation could be a successful delivery platform to enhance the cytotoxicity of EGA against MCF-7 cells. Thus, our study established the in vitro efficacy of optimized EGA-EML-APA as a potential anti-breast cancer formulation. Further in vivo investigation of such formulation following administration via non-invasive para-enteral routes compared to intravenous and oral ones will be considered in future work.