Fluorouracil (5-FU), Span^® 60 (sorbitan monostearate) and cholesterol were purchased from Millipore Sigma (Burlington, MA United States) and used without further purification. Chloroform, methanol, polyethylene glycol 2000 (PEG 200), hyaluronic acid (HA), FA-derivatized mPEG2000-DSPE (FA-PEG2000-DSPE), dimethyl sulfoxide (DMSO), dialysis membrane (MWCO 12,000 Da), sodium dodecyl sulfate (SDS), phosphate-buffered saline (PBS) and Amicon (Ultra-15-Membrane, MWCO 30,000 Da) were acquired from Merck Chemical Co. (Darmstadt, Germany). The MCF10A, MCF7 and 4T1 cell lines were purchased from Pasteur Institute Cell Bank (Tehran, Iran). Medium RPMI-1640 (Dulbecco’s Adjusted Eagle Medium), formaldehyde, trypsin-EDTA, Trypan blue, fetal bovine serum (FBS), phosphate-buffer, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) and penicillin/streptomycin (PS) 100 X were acquired from Gibco, ThermoFisher Scientific (Waltham, MA, United States). 1X binding buffer and annexin V-FITC flow cytometry kit was obtained from Affymetrix Biosciences, ThermoFisher Scientific. Dialysis membrane (MWCO 12,000 Da), Nile red, and Coumarin 6 were purchased from Millipore Sigma. DCFDA/H2DCFDA—Cellular Reactive Oxygen Species (ROS) Assay Kit was obtained from Abcam Company (Cambridge, United Kingdom).

The response surface methodology (RSM) was used for the optimization of niosomal formulations via a central composite design (CCD) approach. To investigate the relationship between a set of independent variables and the dependent variables by fitting the data using a polynomial equation, two numerical parameters (Span^® 60 and cholesterol) were selected to study the impact of their concentration on niosomal particle size (nm), polydispersity index (PDI), entrapment efficacy percentage (EE %) and percentage drug release in 24 h. The polynomial equation was obtained using Design-Expert software (Version 10, Stat-Ease, Inc., Minneapolis, MN, United States). Comparisons were made between the experimental data and the predicted responses. Using the point prediction method, the optimal formulation was chosen for further study (Khanal et al., 2021). Span^® 60 and cholesterol were chosen based on their low, medium and high levels (Supplementary Table S1).

Niosomes loaded with 5-fluorouracil (Nio/5-FU) were prepared using the thin-film hydration method (Akbarzadeh et al., 2020b; Moghtaderi et al., 2021). Briefly, cholesterol and Span^® 60 were dissolved in 9 ml of chloroform/methanol (2:1; v/v). The solution was transferred to a 50 ml round bottom flask. The organic solvent was evaporated under vacuum using a rotary evaporator (Heidolph Instruments, Schwalbach, Germany) at 60°C and 150 rpm for 30 min, until a thin dried film was formed in the bottom of the flask. The 5-FU-loaded niosome was formed by hydration of the thin film at 60°C using PBS containing 10 mg of 5-FU (1X, 10 ml, pH 7.4). Subsequently, the reactants were dried at 60°C and 150 rpm for 30 min to obtain dried specimens. Homogenized samples were prepared using an ultrasonic processor (UP50H compact laboratory homogenizer, Hielscher Ultrasonics, Teltow, Germany) to ensure that the Nio/5-FU had an optimal size distribution (Figure 1). The non-entrapped 5-FU was separated from the entrapped 5-FU with the ultrafiltration method. The samples were kept at 4 C to characterize the different niosomal formulations listed in Supplementary Table S2.

The above prepared Nio-5-FU suspension was added into PEG solution at 0.5 mg/ml concentration, stirred for 15 min at room temperature, and left overnight. The sonicated solution was homogenized for 10 min at 12,000 rpm (Alemi et al., 2018). Niosomes coated with HA (Nio/5-FU/HA) were synthesized by titrating 20 ml of PBS containing 0.1% (w/v) HA into 0.2 g of 5-FU loaded niosomes (Nio/5-FU). The optimum Nio-5-FU was then centrifuged (40,000 rpm, 60 min), and the pellet continued under stirring were added dropwise into the HA suspension. The resulting preparations were stirred up to 24 h at room temperature. The hyaluronic acid solution was prepared by stirring the weighed HA powder in deionized water and adding it to the hydration process. (Zeng et al., 2016). Niosomes coated with folic acid (Nio/5-FU/FA) were synthesized by dissolving Span^® 60, cholesterol and FA-PEG2000-DSPE in 9 ml of chloroform/methanol (2:1; v/v), subsequently, the solvent was removed by rotary evaporation at 160 rpm, at 60°C for 30 min, forming a dried thin film residue. Nio/FA loaded with 5-FU was formed by hydration of the thin film at 60°C using PBS containing 10 mg of 5-FU (1X, 10 ml, pH 7.4). Subsequently, the reactants were dried at 60°C and 150 rpm for 30 min to obtain dried specimens. The samples were sonicated for 5 min to obtain a uniform size distribution, and finally, they were stored at 4°C until further use (Figure 1).

The particle size, Zeta Potential, and polydispersity index (PDI) of the niosomal formulations were determined at 25°C using a benchtop dynamic light scattering/electrotrophoretic light scattering system (Zetasizer Nano S90, Malvern Panalytical Ltd., Malvern, United Kingdom). The morphology of the specimens was observed using transmission electron microscopy (TEM). A drop of niosomal formulation was placed on a carbon-coated copper grid and stained with a 1% phosphotungstic acid. The stained niosomes were imaged with TEM (Model EM900, Zeiss Microscopy, Jena Germany) at 100 kV (Muzzalupo et al., 2005; Sohrabi et al., 2016).

The niosomal formulations were ultra-filtered (Eppendorf^® 580R centrifuge, Hamburg, Germany) with an Amicon Ultra-15-membrane at 4,000 rpm for 20 min at 4°C. The non-entrapped 5-FU was separated from the entrapped 5-FU to determine the percentage of entrapment efficacy (EE%). The free drug concentration was calculated at 266 nm using UV-Visible light spectrophotometer (UV-1700 PharmaSpec, Shimadzu, Kyoto, Japan). The following equation was used to calculate EE% (Rochani et al., 2016).  EE% = [ ( A − B ) / A ] × 100 % where (A) is the initial 5-FU concentration for the niosomal preparation and (B) is the concentration of non-entrapped 5-FU after centrifugation.

For evaluation of drug release, 2 ml of each sample was placed in a semipermeable acetate cellulose dialysis bag (MWCO 12 kDa). The latter was immersed in 50 ml of PBS-SDS (0.5% w/v; release medium). The assembly was agitated at 50 rpm using a magnetic stirrer in various pH conditions (7.4 and 5.4) at 37°C for 72 h. At a specific time (1, 2, 4, 8, 24, 48, and 72 h), 1 ml of the release medium was withdrawn and replenished with the same volume of fresh PBS-SDS (Akbari et al., 2020). The amount of drug released at predetermined intervals was estimated at 238 nm using an ultraviolet light spectrophotometer. The test was repeated using a free drug as control, in which the drug concentration was equivalent inside and outside the dialysis bag.

Different mathematical models were used to evaluate the release kinetics of 5-FU from the samples: the Korsmeyer-Peppa model (log cumulative % drug release vs. log time), the Higuchi model (cumulative % drug release vs. square root of time), first-order model (cumulative % drug remaining vs. time) and zero-order model (cumulative % drug release vs. time) (Dash et al., 2010a; Kamboj et al., 2014; Shaker et al., 2015; Alemi et al., 2018). The correlation coefficient values (r) obtained by regression of the plots derived from the above models were used to calculate the linear curve. The zero-order model is dependent on drug dissolution, which describes the system where the drug release rate is independent of the concentration. The first-order rate equation describes drug release where the rate of drug release depends on its concentration (Dash et al., 2010a). There is a direct relationship between the amount of drug released from a matrix system and the square root of time in the Higuchi and the Korsmeyer-Peppa models (Higuchi, 1963; Korsmeyer et al., 1983). An initial release of 60% was adequate to determine the most suited model for drug release (Bettini et al., 2001).

A perusal of molecular interaction between ciprofloxacin and niosomes, Fourier Transform Infrared Spectroscopy (FTIR) (Spectrum Two, U.S.A.) was used. For this test, lyophilized samples were mixed separately in KBr and the pellets formed by placing the samples in a hydraulic press. FTIR analyses were accomplished in the scanning range of 4000 to 400 cm⁻¹ in a constant resolution of 4 cm⁻¹ and at room temperature.

The Nio/5-FU, Nio/5-FU/PEG, Nio/5-FU/FA, and Nio/5-FU/HA samples were stored at 25 ± 1°C or 4 ± 1°C, at 60 ± 5 % relative humidity for 1 month. After storage, the samples were analyzed for particle size, PDI, and percentage of drug remaining for evaluation of the stability of the formulations.

The human mammary carcinoma cell line MCF7 and the mouse mammary carcinoma cell line 4T1 were cultured at 37°C in an atmospheric supplemented with 5% CO₂. The culture medium consisted of RPMI-1640 fresh medium supplemented with 10% FBS and 1% penicillin/streptomycin (complete growth medium). The medium was aspirated after the cells reached 85-95% confluence. The cell monolayer was detached using 0.25% (w/v) trypsin-EDTA. The detached cells were resuspended in a complete growth medium, labeled trypan blue and counted with a hemocytometer.

Different concentrations of 5-FU, Nio/5-FU, Nio/5-FU/PEG, Nio/5-FU/HA, or Nio/5-FU/FA (0.25, 0.5, 1, 2, 4 and 8 μg/ml) were added to the cultured MCF7 and 4T1 cells and incubated for 48 h. For control, different dilutions of empty niosomes (Nio) were added to a non-tumorigenic epithelial cell line (MCF10A) and incubated for 48 h. To evaluate cell proliferation, the three types of cells were individually incubated with 0.5 mg/ml of MTT for 4 h to reduce the colorless tetrazolium dye MTT to insoluble formazan, which has a purple color. Formazan was dissolved in 100 μL of DMSO for colorimetric determination of the oxidoreductase enzymatic activity. Quantification was performed using the formula:

Percentage cell viability (%) = Optical Density 570-630 treatment/Optical Density 570-630 control × 100%

The flow cytometry assay was used to evaluate the percentage of apoptotic MCF7 and 4T1 cells. Cultured cells were treated with Nio, 5-FU, Nio/5-FU, Nio/5-FU/PEG, Nio/5-FU/HA or Nio/5-FU/FA for 48 h. The Apoptosis and Necrosis Quantitation Kit was used to stain apoptotic cells with green fluorescence and necrotic cells with red fluorescence for examination by flow cytometry. The cells were rinsed twice with PBS and suspended in 1X binding buffer (5 × 10⁵ cells/ml). Cells were stained sequentially in annexin V-FITC (green fluorescence) and propidium iodide (red fluorescence) and analyzed with a benchtop flow cytometer (FACSCalibur, D Biosciences, Franklin Lakes, NJ, United States).

The culture of both types of cancer cells was performed using coverslips placed inside a 4-well plate containing RPMI-1640 medium. After culturing, the cells were incubated with 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCFDA) at 37°C for 30 min. The cells were washed with buffered saline and subsequently treated with Nio, 5-FU, Nio/5-FU, Nio/5-FU/PEG, Nio/5-FU/HA or Nio/5-FU/FA for 24 h. The cells were washed in PBS before being incubated for 30 min at 37°C with 80 mM H2DCFDA. A microplate reader was used to quantify the fluorescence intensity at 530 nm.

1 × 10⁵ cells were seeded for 24 h in plates containing RPMI-1640 medium supplemented with 10% FBS. Nile red was used as a model molecule (500 μL) and loaded into the niosomes. The unloaded stain was removed by dialysis (MWCO 12 kDa). The Nile red-loaded niosomes (50 μg/ml) were incubated with cells for 3 h. The cancer cells were then washed with PBS, fixed with 4% formaldehyde, stained with coumarin-6 (green fluorescence) for 15 min, and examined with a confocal laser scanning microscope (TCS SP5, Leica Microsystems, Wetzlar, Germany).

Statistical analysis and curve fitting were performed using GraphPad Prism software version 8 (GraphPad Software, Inc., San Diego, CA, United States). Data from three independent experiments were expressed as means ± standard deviations. Statistical significance was determined with a one-way analysis of variance after validating the normality and homoscedasticity of the data sets. For all analyses, statistical significance was pre-set at α = 0.05. Central composite design (CCD) was performed using Design-Expert software version 10 (Stat-Ease Inc., Minneapolis, MN, United States).

Zeta Potential is attributed to the stability of the carrier (Manconi et al., 2017). The Zeta Potential graph is shown in Figure 3E. All niosome formulations have a negative Zeta Potential, which shows that they have good stability due to the weak electrostatic repulsive force in the niosomal bilayer. The zeta value of empty Nio was −27.54 ± 1.34 mV, and the Zeta Potential of Nio/5-FU was −13.45 ± 2.55. After HA-coated (Nio/5-FU/HA), the Zeta Potential changed from −13.45 ± 2.55 mV to −34.73 ± 1.50 mV. Nio/5-FU/PEG obtained in this experiment had a Zeta Potential of −19.54 ± 1.49. As seen in Figure 3E, Nio/5-FU/FA had a negative Zeta Potential (−23.64 ± 1.22) than Nio/5-FU.

The morphology of niosome-based carriers was examined by transmission electron microscopy. The size distribution of undecorated and decorated Nio/5-FU formulations was determined by dynamic light scattering (DLS) with a particle size analyzer (Figures 4A–D). The particle size of synthesized niosomes was measured using ImageJ software (National Institute of Health, Bethesda, MD, United States). According to the morphology of prepared niosomes formulations, it can be seen from transmission electron micrograph (TEM) that all niosomes were spherical and homogeneous. Furthermore, no drug crystal was observed in any of the images (Figures 4E–H). Since the morphology of prepared noisome formulations showed that Nio’s vesicle scales were less than 100 nm, all formulations have excellent size for cancer applications. Among coated and uncoated niosomal formulations, Nio-5-FU-PEG (Figure 4F) sample shows more homogeneous and spherical than Nio-5-FU-FA (Figure 4G) and Nio-5-FU-HA (Figure 4H). Also, this niosomes coated with PEG have the lowest average particle size (15.14 nm) in compared with Nio-5-FU-FA and Nio-5-FU-HA. Furthermore, the aggregates in Nio/5-FU/PEG (Figure 4F) were less than other noisome formulations, which shows that the coated formulation with PEG is more stable. The size of the noisome loaded with FU was wider when the niosomes were coated with HA (55.9 nm) (Figure 4H) in compared to two other niosomal formulations. Furthermore, the diameters of the vesicles increased with FA addition compared to uncoated niosomes 40.6 nm (Figure 4G) and 34.9 nm (Figure 4E), respectively. A possible reason for this result is that adding FAPEG2000 causes an increase in the surface tension of the membrane bilayer and higher surface tension helps to decrease the fluidity of the bilayer and increase the particle size of niosomes (Liu and Guo, 2005).

Release of 5-FU from the niosomes was measured for 72 h in phosphate-buffered saline-sodium dodecyl sulfate at physiological pH (7.4) and pathological (cancerous) pH (∼5.4). There was an initial burst release of 5-FU during the first 2 h. The release rates of 5-FU from Nio, Nio/FA, Nio/HA, and Nio/PEG in physiological pH (Figure 4I) were significantly lower than that of undecorated and decorated niosomes at pathological (cancerous) pH (Figure 4J). It is because of the typical behavior of the niosomes at acidic condition, which result in swelling and breaking niosome in this condition. Furthermore the reason for this can be attributed to the electrostatic interaction between the drug and the surfactant, and there is an ionization state at physiological pH (Dash et al., 2010b). After the initial burst release attributed to diffusion of 5-FU from the outer layer of the niosomes, all formulations exhibited the second phase of slower release up to 72 h. The decorated niosomes Nio/5-FU/FA, Nio/5-FU/HA and Nio/5-FU/PEG had lower release rates than the uncoated niosomes. The release rate of Nio/5-FU/HA was relatively lower compared to other undecorated and decorated formulations.

The different release kinetics models obtained for optimizing niosomal formulations are shown in Supplementary Table S6. The best-fitting model for 5-FU, as determined by the highest R² value for curve-fitting, was the first-order model because the value R² for the first-order model in 5-FU was 0.9695. In this model, there is a correlation between drug release and concentration. The best model for optimized Nio/5-FU, in two different pH scenarios, was the Higuchi, because its value was 0.9649 and 0.9754, respectively. The best-fitting model for Nio/5-FU/PEG, and Nio/5-FU/FA in two different pH scenarios was the Higuchi, drug release model (Nio/5-FU/PEG: 0.9340 at pH 7.4 and 0.9408 at pH 5.4, Nio/5-FU/FA: 0.9452 at pH 7.4 and 0.9571 at pH 5.4). In contrast, the best model for Nio/5-FU/HA was the Korsmeyer-Peppas model. the value R² for the Korsmeyer-Peppas model in Nio/5-FU/HA in pH 7.4 and pH 5.4 was 0.9470 and 0.9529, respectively. Nio/5-FU/PEG, and Nio/5-FU/FA in two different pH scenarios was the Higuchi, drug release model because. The classical Higuchi equation is represented by: Q = A [D (2Co-Cs)Cst]½, where Q is the cumulative amount of drug released in time t per unit area, CO is the initial drug concentration, CS is the drug solubility in the matrix and D is the diffusion coefficient of the drug molecule in the matrix.

The empty noisome exhibits stretching peaks for C-O, C=O, and C-H at 1125, 1747, and 2900 cm⁻¹, respectively. Additionally, it exhibits a carbonyl bond at 1625 cm⁻¹ and a -NH stretching vibration at 3100–3400 cm⁻¹, indicating that noisome was successfully formed. Following drug loading, the carbonyl group shifted to 1614 cm⁻¹ and the stretching amide group to 3095 cm⁻¹, indicating the presence of 5-fu in the niosome structure. Following PEGylation of the structure of the drug-loaded niosome, the C-H group corresponds to the PEG structure visible in the final formulation’s 1300 cm⁻¹ region, confirming the structure’s PEGylation. After adding hyaluronic acid to the drug-loaded niosome, a peak at 1655 cm⁻¹ corresponding to the amide group appears in the region, confirming the successful incorporation of HA into the final structure. The final stage of the work involved adding folic acid to the final structure (drug-loaded niosome), which resulted in the appearance of the C-N stretching group in the 1015 cm⁻¹ and 1279 cm⁻¹ regions, confirming the presence of folic acid in the final formulation (Figure 4K).

The stability of undecorated and decorated niosomes and the percentage of residual 5-FU, was examined after storage at 4 and 25°C for 30 days. The Nio/5-FU/PEG nanocarriers were more stable than the other niosomes. Significant differences in PDI and EE were observed for the Nio/5-FU and Nio/5-FU/PEG formulations after storage at 4 and 25°C. The increase in size and PDI and the decrease in EE at 4°C were lower than 25°C after storage (Supplementary Table S8). Although the Nio/5-FU/PEG niosomes were physically stable during storage, leakage of 5-FU occurred for the other niosomal formulations. A more severe leakage was observed for these formulations when stored at 25°C. Changes in particle size and PDI were also more prominent at 25°C. This is because the niosomal nanocarriers have lower mobility and permeability at 4°C. These results suggest that the Nio/5-FU/PEG niosomes are more stable than other niosomes and are more stable at 4°C than 25°C. Supplementary Table S7 indicates that similar to other niosomal formulations, the size, PDI and the EE% of Nio/5-FU/FA increased with increased storage temperature. However, the EE of Nio/5-FU/FA was more sensitive to temperature when compared with the other niosomal formulations. The Nio/5-FU/FA could be stored at 25°C for up to 1 month with only minor changes in size and drug content.

Apart from the 1:1 dilution, there were no statistically significant changes in the viability of MCF10A cells (non-malignant breast epithelial cells) after they were exposed to different concentrations of niosomes obtained by serial dilutions (p < 0.05; Figure 5A). The cell toxicity results for each formulation at different times are reported in Figures 5B,C. The effect of the different niosome formulations on the survival of MCF7 breast cancer cells is summarized in Figure 5D. The half-maximal inhibitory concentration (IC50) values of the different formulations were: 5-FU (1.29 ± 0.06 μg/ml), Nio/5-FU (0.94 ± 0.03 μg/ml), Nio/5-FU/PEG (0.61 ± 0.04 μg/ml), Nio/5-FU/HA (0.32 ± 0.04 μg/ml) and Nio/5-FU/FA (0.19 ± 0.03 μg/ml). Compared to 5-FU, there were also significant decreases in the IC50 values of Nio/5-FU, Nio/5-FU/PEG, Nio/5-FU/HA and Nio/5-FU/FA (p < 0.001, p < 0.001, p < 0.001, and p < 0.001, respectively). The Nio/5-FU/PEG, Nio/5-FU/HA and Nio/5-FU/FA also showed statistically significant decreases in IC50 compared with Nio/5-FU (p < 0.001).

The effect of the niosome groups on the survival of malignant 4T1 cells (a breast cancer cell line derived from the mammary gland of a BALB/c mouse) is summarized in Figure 5E. The IC50 values for the different formulations were 5-FU (1.69 ± 0.13 μg/ml), Nio/5-FU (0.94 ± 0.03 μg/ml), Nio/5-FU/PEG (0.91 ± 0.04 μg/ml), Nio/5-FU/HA (0.64 ± 0.05 μg/ml) and Nio/5-FU/FA (0.27 ± 0.05 μg/ml). Compared to 5-FU, there were significant decreases in the IC50 values of Nio/5-FU, Nio/5-FU/PEG, Nio/5-FU/HA and Nio/5-FU/FA (p < 0.001 for all comparisons). The IC50 values for Nio/5-FU/PEG, Nio/5-FU/HA and Nio/5-FU/FA were significantly lower than those of Nio/5-FU (p < 0.01, p < 0.001, and p < 0.001, respectively). There is an interact between niosomal formulation and folate receptors on the surface of these breast cancer cell lines. As a result, a folate receptor-mediated endocytosis can enhance the internalization process of folic acid functionalized niosomes. This was further conformed by comparing Fu loaded to niosomal formulation with FA and other niosomal modifications (Nio/5-FU/PEG and Nio/5-FU/HA) where the functionalized niosomal formulation with folic acid resulted in greater toxicity in cancer cells after 48 h of treatment compared to two other niosomal (Akbarzadeh et al., 2020a). Moreover, Nio/5-FU/FA exhibited a significant increase in cell inhibition activity when compared with Nio/5-FU/PEG and Nio/5-FU/HA. It is possible that Nio/5-FU/FA further the uptake into cell in turn enhanced cellular toxicity. This study approved earlier results, which showed that FA-modified formulations exhibited the greatest cytotoxicity as a result more effective inhibited cell growth (Lei et al., 2017; Lin et al., 2020).‏

Apoptosis of breast cancer cells was measured quantitatively with flow cytometry after the cells were double-stained with fluorescein thiocyanate-labeled annexin V (annexin V-FITC) and propidium iodide (PI). Administration of Nio, 5-FU, Nio/5-FU, Nio/5-FU/PEG, Nio/5-FU/HA, and Nio/5-FU/FA to the MCF7 cells (Figures 6A,C) and the 4T1 cells (Figures 6B,D) induced apoptosis of both types of breast cancer cells. The percentages of apoptotic cells in both cell lines treated with Nio were not significantly different from the control cells that were not exposed to any of the niosomal formulations (p > 0.05). There were significant increases in the percentage of apoptotic MCF7 cells over the control after the breast cancer cells were exposed to 5-FU, Nio/5-FU, Nio/5-FU/PEG, Nio/5-FU/HA and Nio/5-FU/FA (p < 0.001). The percentages of apoptotic MCF7 cells after treatment with Nio/5-FU/PEG, Nio/5-FU/HA and Nio/5-FU/FA increased significantly, compared to those treated with Nio/5-FU (p < 0.001; Figure 6A). On the contrary, the percentage of apoptotic cells decreased when 5-FU was used instead of the undecorated niosome nanocarriers (p < 0.001). Similar trends were identified for the 4T1 murine breast cancer cells (Figure 6B). Pairwise comparisons indicated that the anticancer effect of Nio/5-FU/FA on the two cell types was higher than Nio/5-FU/PEG or Nio/5-FU/HA.

An important causative factor for cell death of tumor cells is oxidative stress (Joo and Jetten, 2010). Although a small increase in the level of reactive oxygen species (ROS) helps promote cell growth, extreme levels of ROS stimulate cell death. Compared to the control, 5-FU, Nio, Nio/5-FU, Nio/5-FU/PEG, Nio/5-FU/HA, and Nio/5-FU/FA enhanced the fluorescence of 2′,7′-dichlorodihydrofluorescein (DCF), a fluorescent probe for measuring ROS, in both MCF7 cells (Figure 7A) and 4T1 cells (Figure 7B). The 4T1 cells exposed to Nio did not show any significant changes compared to the control group. In contrast, the MCF7 cells exposed to 5-FU, Nio/5-FU, Nio/5-FU/PEG, Nio/5-FU/HA and Nio/5-FU/FA showed significant increases in DCF fluorescence compared to the control (p < 0.001). DCF fluorescence of MCF7 human breast cancer cells treated with decorated niosomes was significantly increased compared to cells exposed to Nio/5-FU (p < 0.001) (Figure 7A). On the contrary, a significant reduction in fluorescence occurred for cells exposed to 5-FU (p < 0.001). Similar results were observed with the 4T1 murine breast cancer cells (Figure 7B).

Confocal laser scanning microscopy (CLSM) was used to study the uptake of niosomes into cancer cells. Nile red was used as a hydrophobic model molecule. As shown in Figure 8A, Nile red-containing niosomes were readily uptaken by MCF7 cells and localized in the cytoplasm. Compared with cells exposed to 5-FU loaded pristine niosomes (Nio/5-FU) and the control cells that were not exposed to anything niosome formulation, there was more profuse penetration of the 5-FU-located, decorated niosomes into the cancerous cells. As expected, the targeted niosomes were beneficial in promoting cellular uptake in comparison with free 5-FU. In the case of Nio/5-FU/HA and Nio/5-FU/PEG, Nile red was significantly localized probably outside the cells instead of a distribution in the endosomes, indicating that the Nio/5-FU/PEG HA and Nio/5-FU/PEG had less cell uptake than Nio/5-FU/FA. The Nio/5-FU/FA shows a significant profuse uptake and better anticancer effect compared to the free 5-FU drugs, Nio/5-FU/HA and Nio/5-FU/PEG. The differences were attributed to the chemical reaction between the folate drug-loaded conjugates and the folate receptor cells in the membrane of cancer cells, which may be explained by the fact that cancer cells overexpress the folate receptor since folic acid is a high-affinity ligand for the folate receptor. As a result, modified niosomes with FA facilitated efficient cellular uptake of the Nio/5-FU/FA into the tumor cells leading to better cytotoxicity in the process than the free 5-FU anticancer drugs, Nio/5-FU/HA and Nio/5-FU/PEG (Lin et al., 2020). The release rate of 5-FU from the FA, HA or PEG-decorated niosomes at pH 5.4 was higher than the free drug and the release rate at pH 7.4. In addition, endocytosis was the most prominent in the cells exposed to Nio/5-FU/FA, compared with cells exposed to Nio/5-FU/HA or Nio/5-FU/PEG (Figure 8B). The variation of the pH in the endosome cause to dissociation of folate from the receptors, which can be recycled back to the membrane. As a result, Nio/5-FU/FA efficiently releases drugs into cancer cells compared to Nio/5-FU/HA and Nio/5-FU/PEG (Tavano et al., 2016).