Document Type : Original Research Article
1 Department of Nanotechnology and Advanced Materials, Materials and Energy Research Center, Karaj, Iran
2 Nanotechnology and Advanced Materials Department, Materials and Energy Research Center, Karaj, Iran
3 1. Hormoz Research Center, University of Hormozgan, Bandar Abbas, Iran 2. Department of Marine Biology, Faculty of Marine Science and Technology, University of Hormozgan, Bandar Abbas, Iran
4 Department of Clinical Pharmacy, Faculty of Pharmacy, Hormozgan University of Medical Sciences, Bandar Abbas, Iran
5 Department of Pharmaceutics, Faculty of Pharmacy, Hormozgan University of Medical Sciences, Bandar Abbas, Iran
In recent decades, cancer has become one of the most challenging global threats to human health and life. Among the leading strategies to overcome this disease, nanomedicine has opened promising horizons in terms of using effective delivery systems for promoting the efficacy and reducing the side effects of cancer diagnostic and therapeutic agents (1). The nanocarriers used in this field include different categories, among which albumin NPs have attracted prominent attention due to desirable characteristics, such as endogenous nature, biocompatibility and biodegradability, non-immunogenicity, non-toxicity, chemical stability and high aqueous solubility (2-5). Being the most abundant protein in the plasma with a long circulating half-life, albumin possesses different binding sites for both hydrophilic and hydrophobic compounds. The presence of amine, thiol and carboxylic acid functional groups confers the possibility of both covalent bonding and non-covalent interaction between this biological macromolecule and a variety of therapeutic agents (6). Accordingly, albumin NPs are able to transport pharmaceutical compounds, especially hydrophobic drugs in the physiological environment, and improve pharmacokinetic and biodistribution of their payloads through conserving them from rapid metabolism and elimination (7, 8). In addition, nanometer dimensions allow albumin NPs to target the tumor sites in a passive manner via the enhanced permeability and retention (EPR) mechanism. Furthermore, active targeting through gp60 receptors and SPARC pathway, both overexpressed in tumor cells, promote accumulation of anti-cancer agents in tumor sites (9, 10).
Various methods including emulsification, desolvation, thermal gelation, self-assembly, nano-spray drying and microfluidic mixing are used for preparation of albumin NPs. Amongst these techniques, desolvation is one of the common ones, widely used due to ease of process and reproducibility (7, 11). In this method, dropwise addition of a desolvating agent like ethanol or acetone to the albumin aqueous solution under stirring conditions, leads to the change in tertiary structure of protein due to disruption of the electrostatic and hydrophobic interaction’s balance. As the result, the water solubility of albumin decreases, it undergoes phase separation and the precipitate of NPs are formed (6, 12). Moreover, a crosslinker agent, often glutaraldehyde, is needed for stabilizing the obtained NPs. In crosslinking process, carbonyl moieties of glutaraldehyde are attacked by amine groups in lysine and arginine residues of protein which results in the formation of Schiff bases (13). In desolvation procedure, different parameters such as albumin concentration, pH, desolvating agent to albumin solution ratio, rate of desolvating agent addition, stirring rate, type and ratio of the crosslinker, need to be carefully controlled because of their significant role on the NPs size. Indeed, the size of NPs not only is involved in the loading amount of the drug, but also affect cellular uptake of NPs (14, 15).
1,7-bis-(4-hydroxy-3-methoxyphenyl)-1,6-heptadiene-3,5-dione or curcumin is a natural hydrophobic polyphenol extracted from the rhizome of a plant named curcuma longa. Besides pleiotropic features such as anti-inflammatory, antioxidant, antimicrobial and wound healing activities, curcumin has been illustrated to be effective in inhibition of tumor cells growth and proliferation in numerous studies. Despite the unique properties of curcumin, very low water solubility and bioavailability as well as physiological instability has limited its clinical applications (16-20). Curcumin is a symmetric molecule with a β-diketone skeleton and two aromatic groups (Fig. 1). In-vivo and in-vitro studies have shown that the presence of methylene active group and β-diketone moiety are involved in physiological instability and rapid metabolism of curcumin. On the other hand, efficacy of cyclohexanone derivatives of curcumin to increase cytotoxicity against estrogen receptor negative cancer cells has been reported (21, 22). Also, the hydroxyl groups of curcumin are altered in kidneys, liver, and intestinal mucosa, produce the curcumin glucuronide and curcumin sulfate, which have low bioactivity and efforts have been made to replace these groups as well (23, 24). 2,6-bis(3,4-methylenedioxybenzylidene)-1-cyclohexanone (BMC) is a cyclohexanone derivative of curcumin in which the methylenedioxy groups has replaced the hydroxyl groups (Fig. 1). According to the research carried out by Ramshini et al., this compound showed more stability and bioavailability than curcumin with an improved cytotoxicity against MCF-7 cancer cell line (25).
In the present study, the interaction of BMC with bovine serum albumin (BSA) was investigated for the first time with fluorescence spectroscopy and molecular docking techniques. Subsequently, BSA nanoparticles were used as a carrier for BMC hydrophobic compound for improving its solubility and anticancer properties. Two desolvation based procedures, adsorption and entrapment, were used to prepare BMC loaded BSANPs. Then, physicochemical properties of the two prepared formulations were compared and drug release behavior as well as kinetic of release were evaluated by mathematical models. Eventually, cytotoxic effects of the formulations in-vitro was investigated against MCF-7 cancer cell line.
MATERIALS AND METHODS
Bovine serum albumin (BSA > 99% w/w), 2,6-bis(3,4-methylenedioxybenzylidene)1-cyclohexanone (BMC > 99% w/w), glutaraldehyde (25% w/w solution in water), phosphate buffer saline tablet (PBS), dialysis tube (molecular weight cut off of 12 kDa), RPMI-1640 culture medium, MTT powder (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide), penicillin and streptomycin antibiotics and trypsin-EDTA solution (0.25% w/v) were purchased from Sigma-Aldrich (St. Louis, Missouri, United States). Fetal bovine serum (FBS) was prepared from Gibco (Waltham, Massachusetts, United States). Hydrochloric acid (37% w/w) and ethanol was purchased from Merck (Germany). All other reagents and solvents were analytical or HPLC grade. Ultra-pure deionized water was used for preparing all solutions.
Fluorescence spectroscopy studies
Fluorescence measurements for investigating the interaction of BMC compound and BSA protein were performed using a Cary eclipse fluorescence spectrophotometer. For this analysis, the excitation wavelength was set on 280 nm and the BSA emission spectra with a fixed concentration of 1.5 μM was obtained in the range of 300-500 nm in the absence and presence of the BMC at different concentrations of 0.5, 1, 1.5, 2, 2.5, 3, 4, 5, 7.5 and 10 μM. The excitation and emission slit widths both were adjusted on 5 nm and measurements were carried out at two temperatures of 298 and 310 K using a fluorescence cuvette with 1 cm path length.
The quenching mechanism was determined by applying the Stern-Volmer equation to fluorescence data as follow (26):
F0 ⁄ F=1+kq τ0 [Q]=1+KSV [Q] Equation (1)
Where, F0 and F are the fluorescence intensity in the absence and presence of BMC, respectively, kq is the biomolecular quenching rate constant, τ0 is the life time of the fluorophore in the absence of BMC as quencher and is 10-8 s for BSA, [Q] is the BMC concentration and KSV is the Stern–Volmer quenching constant.
Also, the binding parameters were determined using following equation:
log ∆F⁄F = log Kb+nlog[Q] Equation (2)
Where ∆F is the difference between F0 and F, Kb is the binding constant and n is the number of binding sites (27).
Finally, the thermodynamic parameters, including Gibbs free energy (∆G°), enthalpy (∆H°) and entropy (∆S°) changes were calculated from Van’t Hoff equation (28):
∆G°=-RT ln Kb=∆H°-T∆S° Equation (4)
R and T are the ideal gas constant (8.314 J/mol.K) and absolute temperature, respectively.
Molecular docking studies were performed using AutoDockTools version 1.5.7. BMC molecule was designed by ChemBio3D Pro 12.0 and geometry was optimized using MM2 force field method with the minimum RMS of 0.001. The crystal structure of BSA with PDB code of 3V03 was prepared from RCSB Protein Data Bank. BSA was considered as rigid molecule and BMC as a flexible one on BSA. Hydrogen atoms were added and non-functional water molecules were removed. Kollman and Gasteiger partial charges were also assigned to the protein and ligand. For global optimum binding position search, the Lamarckian genetic algorithm technique was used with following parameters:
Population size: 150, maximum number of generations: 27000, maximum number of energy evaluations: , mutation rate: 0.02 and crossover rate: 0.8.
The docking process was performed for 100 runs and the conformations were ranked according to the docked energy.
Synthesis of BSANPs and BMC loaded BSANPs
Desolvation method was used for synthesis of formulations with some modifications (11). For preparing of BMC@BSANPs-A, adsorption method, 20 mg of BSA was dissolved in 1 mL ultrapure water and stirred at 500 rpm for 20 min. Subsequently 4 mL ethanol as a desolvating agent was added dropwise to BSA solution at a rate of 1 mL/min and stirred for more 10 min. Then, 12 μL of 8% glutaraldehyde was added as crosslinker and the solution was stirred at 500 rpm for 24 h at 25 ℃. To remove any free glutaraldehyde and ethanol, the nanoparticle colloidal dispersion was purified by centrifugation at 12,000 rpm for 10 min at 25 ℃. Centrifugation was performed in 3 cycles and at every step the NPs were dispersed in 5 mL ultrapure water using ultrasonication for 5 min. Finally, the BSANPs were dispersed in 1 mL ultrapure water and stored at 4℃ in darkness or frozen overnight at -20 and lyophilized at -55℃ for further tests. For loading of BMC on BSANPs, BMC in ratios of 1:10, 1:20 and 1:40 to BSA (w/w) was dissolved in 4 mL ethanol and added to BSANPs aqueous solution and stirred at 500 rpm for 24 h in darkness. Then the colloidal dispersion was centrifuged at 12,000 rpm for 10 min to remove the ethanol and unbonded drug.
For BMC@BSANPs-E formulation preparation, BMC was loaded during the synthesis process of BSANPs. For this purpose, BMC (1:10, 1:20 and 1:40 (w/w) to BSA) was dissolved in 4 mL ethanol and added to 20 mg/mL BSA aqueous solution in desolvation step. Similar to adsorption method, after formation of NPs, 12 μL of 8% glutaraldehyde was added for crosslinking and after 24 h stirring at 500 rpm, suspension was centrifuged 3 times at 12,000 rpm for 10 min at 25 ℃ and final NPs were dispersed and stored in water or lyophilized for further tests.
Characterization of NPs
Hydrodynamic particle size, polydispersity index (PDI) and zeta potential were determined by dynamic light scattering (DLS) technique, using a Horiba SZ-100 nanoparticle analyzer. The NPs were diluted with ultrapure water (1:100 v/v) and measurements were performed at 25 ℃ using a scattering angle of 90°. The morphology of NPs was studied using field emission scanning electron microscope (FESEM), model TESCAN MIRA 3, and transmission electron microscope (TEM) model ZEISS LEO 906 E. To prepare FESEM samples, a drop of NPs colloidal dispersion was placed on a silicon slice and dried in air. The samples were coated by gold to make them conductive. The samples were imaged at 10 kV using the secondary electrons mode. TEM samples were prepared by drying the NPs colloidal dispersion on a copper grade. Also, the UV-visible absorption spectra of Pure BSA, BSANPs, BMC, BMC@BSANPs-A and BMC@BSANPs-E at the wavelength range of 200-600 nm were determined by spectrophotometer (UNICO 2150-UV) at ambient temperature and the wavelength of maximum intensity were determined for BSA and BMC. The chemical structure of samples as well as confirmation of drug loading in NPs were investigated by FTIR spectroscopy. For this test, 1 mg of the samples were mixed with 100 mg KBr powder and pressed in the form of disks. The FTIR spectra of the samples were obtained in the range of 4000-400 cm-1 using a Bruker ALPHA II FTIR spectrometer. Moreover, TGA analysis was used for the examination of BMC, BSA, BMC@BSANPs-A and BMC@BSANPs-E formulations weight loss using PL-STA 1640 instrument. The analysis was performed on 5 mg of each sample, up to 600℃ in air atmosphere. The heating rate set on 10 ℃/min.
Determination drug loading and entrapment efficiency
To calculate the drug loading percentage (DL%) and entrapment efficiency percentage (EE%), the supernatant of centrifugation step, in each of the loading methods, was collected and analyzed by UV-Vis spectrophotometer. The standard curves of BSA and BMC were plotted by measuring the absorption of maximum intensity at 283 and 389 nm respectively in different concentrations (Fig S1). DL% and EE% were calculated using the following equations:
Drug release study
Dynamic dialysis method was used to evaluate in-vitro release behavior of BMC from BMC@BSANPs-A and BMC@BSANPs-E formulations under two conditions with acidic and neutral pH of 5.5 and 7.2. For this purpose, 1 mL of each formulation (equivalent to 0.5 mg of drug) was poured and trapped inside a dialysis bag with 12 KDa molecular cut-off and immersed in 10 mL of PBS-ethanol release medium with a volume ratio of 6:4 v/v. HCl was added to the media to obtain an acidic pH. Drug release process was performed at 37 ℃ under shaking at 100 rpm in darkness. At predetermined intervals from 15 min to 120 h, 3 mL of each medium was withdrawn and replaced with equal amount of fresh one to maintain sink conditions. The amount of released BMC was obtained by analyzing the withdrawn samples by UV-Vis spectrophotometer at BMC characteristic peak (absorption at 389 nm) and applying the obtained equations from plotted standard curves of BMC in release media (Fig. S2). The cumulative release percentage was calculated using the below equation and the graph of release percentage was drawn versus time.
Drug release kinetics
The mechanism of BMC release from BSANPs was investigated using zero-order, first-order, Korsmeyer-Peppas, Higuchi and Hixon-Crowell kinetic models as follow:
Where Mt is the amount of BMC released at a given time t and M∞ is the total amount of BMC loaded into the NPs. Also, k0, k1, kH, kHC, kKP and n are constants for the zero-order, first-order, Higuchi, Hixon-Crowell and Korsmeyer-Peppas models, respectively. Nonlinear regression method was used to fit the models with the release experimental data.
MCF-7 human breast cancer cell line was prepared from the cell bank of Pasteur Institute (Tehran, Iran). The cells were grown in culture flasks containing RPMI 1640 media with 10% FBS and 1% Penicillin and Streptomycin (Pen-Strep) antibiotics. Then cells were placed in incubator with a humidified atmosphere by 5% CO2 at 37℃. The culture media was replaced every 2 days.
The in-vitro cytotoxicity studies were assessed against MCF-7 cells using MTT assay according to the manufacturer’s instructions. Briefly, the cells were seeded in 96 well microplate (8×103 cells per well) for 24 h. Then, the cells treatment was performed with 50-350 μg/mL of BSANPs or 5–60 μg/mL of free BMC and equivalent doses of BMC@BSANPs-A and BMC@BSANPs-E formulations. After incubation for 24 and 72 h, media were replaced by fresh one containing 0.25 mg/mL MTT reagent (without phenol red) and incubated for 3 h at 37ºC until purple precipitate was visible. Then, each medium was carefully removed and 150 μL DMSO per well was added and incubation was done for about 15 min at 25 ºC. Spectrophotometric absorbance of the formazan solution was obtained at 570 nm by Bio Tek ELx808 model microplate reader. Results were reported as percentage of cell viability and IC50 values. Percentage of cell viability was calculated according to equation (9) and untreated cells were considered as control samples.
% cell viability = 100 × (OD sample - OD control) / OD control. Equation (13)
SPSS 26 software was used for statistical analysis following one-way analysis of variance (ANOVA). All data were presented as the mean ± standard deviation of three independent tests. The p-value < 0.05 was considered significant in all tests.
RESULTS AND DISCUSSION
Molecular interaction study
One of the effective techniques for analyzing the small molecules and proteins such as albumin interactions, is fluorescence spectroscopy. Fluorescence emission of BSA is attributed to aromatic residues of Trp, Tyr and Phe (29, 30). Due to the low quantum yield of Tyr and Phe, mainly Trp is involved in the intrinsic fluorescence emission of BSA. Pursuant to the literature, fluorescence quenching, any process in which the fluorescence intensity of a protein is reduced, could be helpful for determining the interaction between a small molecule as a ligand and the protein (31). In this regard, the effect of BMC on the fluorescence intensity of BSA was determined in the absence and presence of different concentrations of BMC. The spectra in Fig. 2a shows a significant decrease in BSA fluorescence intensity with the increase of BMC concentration. This indicates that BMC has the ability to bind to BSA and quench the intrinsic fluorescence of Trp by changing the conformation of BSA or microenvironment around it. A variety of molecular interactions can lead this quenching, including excited state reactions, molecular rearrangements, ground state complex formation, energy transfer and collision quenching can lead this fluorescence quenching (32). In general, fluorescence quenching can occur through dynamic, static or both quenching mechanisms depending on the factors like temperature, viscosity, and diffusion coefficients (33). Accordingly, to determine the quenching mechanism, Stern–Volmer equation was employed. The Stern–Volmer plots based on F0/F versus [Q] at two studied temperatures are illustrated in Fig. 2b and the values of KSV and kq obtained from the slope of the fitted lines are shown in Table 1. The plots show good linearity, and according to the results, increasing the temperature from 298 to 310 K, the values of KSV decreased, indicating static mechanism for quenching. Also, higher values of kq compared to maximum scattering collision quenching constant () demonstrated the static mechanism of quenching as well. Moreover, the values of binding constant (Kb) and binding sites (n) calculated from Eq. 2 and Fig. 2c are shown in Table 1. The approximate equality of n to 1 values demonstrated that one binding site is available on the BSA for BMC molecule. As well, higher Kb value at 310 K than 298 K demonstrated the increasing of bonding affinity of BSA to BMC with temperature increment. Moreover, previous studies have shown that four non-covalent interaction forces including electrostatic force, hydrophobic interactions, hydrogen bonds, and van der Waals forces, may occur between small molecules and macromolecules (34). Interaction forces mode can be confirmed using thermodynamic parameters (∆H° and ∆S°), so that if ∆H° < 0 and ∆S° > 0, the electrostatic forces; ∆H° > 0 and ∆S° > 0, hydrophobic interactions; ∆H° < 0 and ∆S° < 0 van der Waals interactions and hydrogen bonds are more dominant binding forces (28). According to this, thermodynamic parameters were obtained using Van’t Hoff and standard Gibbs free energy (ΔG°) equations (Eq. 3 and Eq. 3) and presented in Table 1. Both ∆H° and ∆S° showed values above zero, indicating a predominant hydrophobic interaction between BSA and BMC. Also, negative values for ΔG° showed that the reaction of BSA and BMC was spontaneous.
Subsequently, molecular docking investigations were performed to supplement fluorescence spectroscopy studies and for further understanding of BSA protein and BMC interactions. Previously, subdomains IIA and IIIA are known as the main binding sites of albumin for small molecules. On the other hand, curcumin and its derivatives have shown an affinity to bind to hydrophobic cavity of albumin in subdomain IIA close to Trp-213 residue (35-38). Hence, our docking studies focused on subdomain IIA of BSA. The results showed that BMC successfully binds to the subdomain IIA in the interface of subdomain IIB and in the vicinity of Trp-213. This result also justified Trp fluorescence quenching of BSA in the presence of BMC. The best energy-ranked conformation with the lowest free energy (∆G°) of -6.59 kcal/mol and lowest RMSD value (0.0) is shown in Fig. 3. The ∆G° value in confirming the values obtained from thermodynamic parameters is negative and indicates spontaneous reaction. However, the difference with the experimental value (-11.28 kcal/mol) can be due to ignoring the presence of solvent and/or rigidity of some other residues of BSA in the docking studies (26). Moreover, BMC was surrounded by a number of amino acid with abundance of hydrophobic residues of Val-481, Leu-480, Trp-213, Ala-209, Ala-212, Lue-346, Ala-349, Leu-362, polar residues of Ser-453 and Ser-343, and charged residues of Arg-483, Arg-484, Arg-208 and Glu-353. Also, two hydrogen bonds were formed between BMC and Val-481 and Ser-453 residues of BSA (Fig. 3b). Therefore, the predominant binding between BSA and BMC is hydrophobic interaction followed by hydrogen bonds which is in consistent with thermodynamic parameters studies.
Preparation of BMC loaded BSANPs
Desolvation is a thermodynamically driven self-assembly technique used for preparing albumin NPs and loading both hydrophilic and hydrophobic anticancer drugs such as doxorubicin, docetaxel, curcumin, etc. (39-41). In various studies, drug loading has been performed in different approaches. Sun et al. after synthesizing the albumin NPs, loaded doxorubicin by adding the drug to the NPs solution and stirring it for a specified duration of time (42). Also, Fattahian Kalhor et al. with the aim of loading one of the curcumin derivatives (DVH) on albumin NPs by desolvation method, first synthesized NPs, then added the drug solution to the NPs solution and stirred it continuously for a period of time (43). On the other hand, Salehiabar et al. loaded curcumin in albumin NPs during the synthesis of NPs and in the step of adding acetone desolvating agent containing curcumin (44). Furthermore, loading the CDF derivative of curcumin on albumin NPs was done by Gawde et al. simultaneously as the NPs preparing (45). In the present study, two different loading approaches based on the desolvation method were used to investigate the effect of loading process on properties of drug-loaded NPs. In adsorption approach, first BSANPs were synthesized using ethanol as desolvating agent and glutaraldehyde as crosslinker, and in the next step by continuous stirring of the NPs aqueous solution and the ethanolic solution of BMC for a period of time the drug was loaded onto the NPs. In the entrapment method on the other hand, loading was performed simultaneously with the synthesis of BSANPs in the step of adding the ethanol containing the dissolved BMC. Considering the addition of drug molecules after complete formation of BSANPs in the adsorption method, it seemed that the BMC had attached to binding sites on the surface of BSANPs. However, in the entrapment method, due to the addition of the drug before the formation of BSANPs and its fixation with glutaraldehyde, BMC had entrapped within the protein network of NPs in addition to formation bonds to it. Fig. 4. illustrates the schematic of the two synthesis and loading procedures.
Physicochemical charactristics of BSANPs, BMC@BSANPs-A and BMC@BSANPs-E formulations with different BMC/BSA ratios are shown in Table 2. The size of NPs plays a major role in their biodistribution and delivery to cancereous cells. According to the results of this study, the hydrodynamic particle size of the formulations was in the range of 159.2 ± 3.45 to 201.7 ± 1.57 nm (Fig. S3). As previously reported, nanoparticles with a size range between 100 to 200 nm can circulate in the bloodstream for a longer period of time. Actually, this size range of NPs facilitates interlization into cells via receptor-mediated endocytosis, while larger ones undergo phagocytosis, and the smaller size of NPs may lead to their renal removal (46, 47). It has also been indicated that NPs of this size range can extravasate through the endothelial fenestration of tumors and accumulate in the tumor environment due to EPR effect (48). For both BMC@BSANPs-A and BMC@BSANPs-E, with increasing BMC/BSA ratio from 1:40 to 1:20, the average particle size did not change significantly (p > 0.05), but at the ratio of 1:10, average particle size increased significantly (p < 0.05). Likewise, in studies on the loading of paclitaxel and curcumin on albumin nanoparticles similar results have been reported (41, 49, 50). In addition, for all BMC/BSA ratios, BMC@BSANPs-E had smaller size than BMC@BSANPs-A. PDI values also indicated that the prepared formulations had a narrow size distribution. Meanwhile, zeta potential is an important parameter that indicates the surface charge and stability of NPs under colloidal conditions. NPs with zeta potential of more than ±30 show good colloidal stability due to surface charge repulsion (51). In addition, the negative surface charges of NPs lead to less toxicity as well as increase the circulation time due to the reduction of the chance of activating the immune system or opsonization and removal of NPs in the bloodstream (45, 52). According to the results of the present study, the zeta potential for all formulations had acceptable negative values, which indicated the good physical and electrostatic stability of prepared formulations (Fig. S4). The zeta potential values of BMC@BSANPs-E formulation were more negative than BMC@BSANPs-A, may be due to the smaller size of BMC@BSANPs-E which lead to higher specific surface and higher number of negative charged amino acids on NPs surface at neutral pH. In addition, increasing the BMC/BSA ratio enhanced the DL% in both formulations, which has been reported in similar studies on albumin NPs as nanocarrier (49, 53). Also, the DL% for the BMC@BSANPs-A in all ratios was slightly higher than that of BMC@BSANPs-E (p < 0.05), indicating that the amount of drug molecules attached to the surface of BSANPs was relatively higher than that of entrapped in the protein matrix. Since the NPs in the BMC@BSANPs-E formulation had smaller sizes, less drug may have been loaded in these NPs. On the other hand, EE% of BMC@BSANPs-E increased with rising BMC/BSA and reached to 91.93 ± 0.62 at 1:10 ratio, while for BMC@BSANPs-A, increasing the ratio from 1:40 to 1:20, EE% reached to 84.08 ± 0.35, but at 1:10, EE% remained almost constant. This illustrated that the BMC@BSANPs-E formulation had a greater capacity for encapsulating the drug molecules.
Fig. 5 shows the FESEM images of BSANPs, BMC@BSANPs-A and BMC@BSANPs-E at 50 kx magnification and the particle size distribution histogram obtained using ImageJ.JS. According to these data and TEM images in Fig. S5, the NPs had a spherical morphology with smooth surfaces and a proper size distribution. The mean particle sizes for BSANPs, BMC@BSANPs-E and BMC@BSANPs-A were obtained 112.03 ± 1.04 nm, 109.68 ± 1.72 nm and 133.97 ± 2.00 nm, respectively. Although the ratio of loaded drug in two formulations was the same, the BMC@BSANPs-A had a larger particle size than BMC@BSANPs-E which was consistent with the results of the DLS analysis. However, it should be noted that in DLS analysis, the hydrodynamic diameter of NPs is measured in aqueous medium and the presence of water molecules cause swelling and increasing of the size of NPs while FESEM imaging is performed on dried samples. Therefore, the size of NPs obtained from FESEM images are smaller than DLS analysis.
UV-Vis spectroscopy was used for structural studies. Absorption spectra of BSA, BMC and formulations are shown in Fig. 6a. The characteristic peak for BSA was obtained at 283 nm that was related to the amino acids with aromatic rings specifically tryptophan absorbance (54). This peak disappeared in NPs spectra, which could be due to the crosslinking of amino acids and/or entrapment of related functional groups into the NPs (13, 42). As well, in the absorption spectrum of BMC, two characteristic peaks were detected at 256 nm and 389 nm. However, peak at 389 nm was considered as the characteristic peak due to the higher intensity. This peak with lower intensity was also observed in the BMC@BSANPs-A and BMC@BSANPs-E spectra which could be related to BMC molecules adsorbed on the surface of nanoparticles.
The thermograms of TGA analysis, shown in Fig. 6b, illustrated that BSA, BMC@BSANPs-E and BMC@BSANPs-A had an initail weight loss of 9.13%, 8.77% and 5.95% up to about 160℃, respectively, which was attributed to water loss. For BMC, however, no significant initial weight loss was observed, which may be due to its hydrophobic nature. Thermal decomposition of BMC occurred with a relatively rapid rate beginning from 182℃ and reaches a total weight loss of 97.5% up to 570℃. Also the BSA showed total weight loss about 81.77%, with decomposition beginning from 160℃. For BMC@BSANPs-A and BMC@BSANPs-E decomposition began from about 162℃ and 173℃, respectively, which both were more than of BSA due to BMC loading in BSANPs. In addition, total weight loss up to 600℃ for BMC@BSANPs-A and BMC@BSANPs-E were obtaind 84.3% and 80.18%, respectively. As it turned out, both NPs had a slower decomposition rate than free BMC. However, BMC@BSANPs-A showed higher total weight loss, as well as higher onset of decomposition temperature, may be due to the more BMC loaded on the surface of NPs (Table 2).
Furthermore, FTIR spectra were obtained for more structural changes evaluations and confirmation of drug loading in NPs. As shown in Fig. 6c, the absorption peaks of the pure BSA at 3306 cm-1, 3068 cm-1, 2961 cm-1, 1654 cm-1, 1539 cm-1, 1395 cm-1 and 1242 cm-1 were related to O-H stretching vibrations, amide A (N-H stretching vibrations), amide B (N-H stretching vibrations of free ion NH3+), amide I (C=O stretching vibrations), amide II (C-N stretching and N-H bending vibrations ), CH2 bending vibrations, and amide III (C-N stretching vibrations and N-H bending vibrations) (40, 55, 56). On the other hand, according to Table 3, the BSANPs, BMC@BSANPs-E, and BMC@BSANPs-A spectra showed protein amide bands with slight shifts and a decrease in intensity of peaks compared to the pure BSA, indicating a change in CN or NH bands due to interactions of albumin different groups and changes in the secondary structure of protein during crosslinking with glutaraldehyde and interaction with BMC (54, 57). Also, in the FTIR spectrum of BMC, the absorbtion peak at 3366 cm-1 corresponded to the stretching vibrations of O-H bonds, the peaks in the range of 3049 cm-1 to 2613 cm-1 were due to aromatic and aliphatic C-H stretching vibrations and the peaks at 1640 cm-1, 1576 cm-1 and 1512 cm-1 were due to C=O and aromatic C=C stretching vibrations. In addition, absorbtion at 1420 cm-1 was due to olephinic bending C-H vibration and strong peaks in the range of 1255 cm-1 to 990 cm-1 were related to aromatic C-O strething vibrations and C-O-C strething vibrations of methylenedioxy groups. Peaks in the range below 900 cm-1 were also caused by aromatic C-H out of plane bending vibrations (58-61). In the BMC@BSANPs-E and BMC@BSANPs-A FTIR spectra, most of the pure BMC peaks were overlapped by BSA peaks, nevertheless the presence of 1512 cm-1, 1261 cm-1, 1160 cm-1 and 1031 cm-1 peaks of BMC indicated the successful loading of BMC into BSANPs.
In-vitro drug release study
The release of BMC from BMC@BSANPs-A and BMC@BSANPs-E was studied at pH 5.5 and 7.2. Fig. 7 shows the percentage of cumulative release of formulations over time. According to these results, BMC released from BSANPs in a biphasic pattern in all samples. The burst release was observed at the first 4 h which was subsequently followed by a controlled release process. Accordingly, in the first phase, the release percentage for BMC@BSANPs-A at 5.5 and 7.2 pH was 38% and 41% and for BMC@BSANPs-E was 9% and 8%, respectively. This initial rapid release was probably due to the desorption of drug molecules that were weakly attached to the NPs surfaces through hydrophobic interasctions. Sustained release in the next phase may be due to the desorption of the more strongly attached molecules or the ones have partially migrated and trapped into the protein matrix during the loading process of BMC@BSANPs-A or due to the difussion of the drug molecules entrapped into the NPs matrix during the preparation and loading process of BMC@BSANPs-E (43, 62). On the other hand, BMC release from NPs was highly dependent on pH and loading method. Accordingly, the maximum released drug from BMC@BSANPs-A after 120 h at pH 5.5 and 7.2 was 96% and 59%, while this value for BMC@BSANPs-E was 34% and 19%, respectively. It should be noted that crosslinking with glutaraldehyde in desolvation method plays an important role in degradation and subsequent release behavior of BSANPs (62). Actually for BMC@BSANPs-A, BMC was loaded after formation of NPs and stabilization by glutaraldehyde, therefore, the BMC molecules may have been mainly adsorbed on the NPs. Thus, during release, the adsorbed molecules are easily separated from the surface of the NPs. On the other hand, in BMC@BSANPs-E, the BMC was added at the NPs formation stage and then the NPs and drug molecules were fixed together by crosslinker. So, most of the BMC molecules were entrapped and only a small amounts were able to diffuse from the NPs matrix and release into the medium. Comparable results have been reported in a study by Merodio et al. and Sadeghi et al. on the loading of albumin NPs with ganciclovir and curcumin, respectively (63, 64). Despite the lower in-vitro release of BMC from BMC@BSANPs-E, it can be expected that in-vivo administration of this formulation prevents rapid metabolism of BMC and reduces drug leakage in the bloodstream which can lead to fewer side effects for normal cells. Moreover, the active and passive targeting ability of albumin NPs, leads to delivery of drug cargo to the tumor site where BMC@BSANPs-E can gradually release all the entrapped BMC due to enzymatic degradation of albumin. In addition, both formulations showed greater release rates in acidic pH than neutral condition which could be caused by degradation of the bonds between albumin and BMC as well as degradation of the protein matrix in acidic conditions (63). Compared to the neutral pH of the blood, the tumor sites as well as the endosomes and lysosomes of the cells have an acidic pH (4.5-6.0). Thus, greater release in acidic conditions can reduce the adverse effects of drugs on normal cells during circulation and aid in the targeted release of the drug in the tumor sites (41, 44, 65)
Relaese kinetic study
To determine the release mechanism of the BMC from BMC@BSANPs-A and BMC@BSANPs-E at pH 5.5 and 7.2, the cumulative release data were fitted to different mathmatical kinetic models (Fig. 8). Nonlinear regression method was used for this purpose. The results including the kinetic parameters and values of the coefficient of determination (R2) are shown in Table 4. R2 coefficient indicates the goodness of the model and when it is equal to 1, the obtained data are completely fitted to model. According to the results of Table 4, the first order kinetic model had the closest values of R2 to 1 for all the studied samples. Thus, the release of BMC from BSANPs followed a first order kinetic model. According to this model, the amount of
released drug depends on the concentration gradient of drug between the static fluid layer around the carrier surface and the release medium. In this model, it is assumed that the surface of the carrier remains constant and does not change its shape during release. This result meant that the release of BMC from BSANPs was not mainly due to the degradation of the protein network that could lead to its surface change or deformation. On the other hand, the value of n obtained from the Korsmeyer-Peppas model, which indicates the diffusion coefficient, is less than 0.5 for the samples. Conforming to this model, if n <0.5: the fickian diffusion, 0.5 <n < 1: the irregular diffusion, n = 1: type II transfer and for higher n values: super type II transfer controls the release mechanism (66-68). Therefore, the release mechanism of BMC@BSANPs-A and BMC@BSANPs-E formulations was controlled by fickian diffusion. This mechanism could occur because of the chemical potential gradient of BMC molecules. Similar release mechanism was reported for the release of curcumin from two different albumin containing carriers including egg albumin NPs and albumin shell of magnetic nanohybrid particles (cluster@silica@albumin) (69, 70).
The cytotoxicity of free BMC, BSANPs and BMC-containing BSANPs was investigated on MCF-7 cancer cell line using MTT assay. For this purpose, cells were incubated for 72 h with BSANPs samples at concentrations of 50, 150, 250, 300 and 350 µM. As shown in Fig. 9a, BSANPs did not show significant cytotoxicity against MCF-7 cells in the assessed concentrations (p > 0.05). Also, the cytotoxicity evaluation of free BMC as well as BMC@BSANPs-A and BMC@BSANPs-E with different concentrations of BMC (5, 10, 20, 40 and 60 µM) on MCF-7 cancer cells indicated the dependence of cell viability to BMC concentration and incubation time (Fig 9b and 9c). Increasing BMC concentration and incubation time from 24 to 72 h, significantly reduced the cell viability (p < 0.05). In addition, IC50 values, shown in Table 5, indicated the higher cytotoxicity and subsequently more efficient inhibition of cell growth and proliferation for BMC-containing BSANPs compared to free BMC. This further inhibitory effect of NPs could be due to increased endocytosis and cell uptake in the presence of albumin as an endogenous ligand, as well as sustained release of drug cargo by BSANPs (41, 61). The optical microscopy images in the Fig. 10 also showed that the cells treated with NPs had lower viability than those treated with free BMC (after 72 h incubation at the concentration of 20 µM), as well as the morphological changes and cell shrinkage were more evident in them. On the other hand, BMC@BSANPs-A showed more cytotoxicity than BMC@BSANPs-E after 24 h of incubation at concentration higher than 40 µM (p < 0.05). Considering the higher DL% (Table 2) and release behavior of BMC@BSANPs-A (Fig. 7), it can be concluded that in the first 24 h, more BMC release from the NPs surface resulted in higher cytoxicity. However, after 72 h of incubation, at concentrations equal to and greater than 20 µM, both formulations showed relatively similar cytoxicity (p > 0.05), although at lower concentrations BMC@BSANPs-A still showed higher toxicity compared to BMC@BSANPs-E (p < 0.05). This result could be due to the sustained release of BMC from BMC@BSANPs-E over a longer period of time as these NPs enters the cells and release their drug cargo via degradation of BSA.
In this study, fluorescence spectroscopy and molecular docking studies showed that the BMC has the affinity to bind to BSA and quench the Trp fluorescence, mainly through hydrophobic interactions. In addition, applying two adsorption and entrapment approaches, BMC loaded albumin nanoparticles were successfully prepared using desolvation method. Particle size and zeta potential of formulations were in the desirable range in terms of benefiting from EPR targeting and longer blood circulation time of drug. However, BMC@BSANPs-E had smaller particle size than BMC@BSANPs-A according to hydrodynamic particle size and microscopic measurements. Both formulations released more drug at acidic conditions. On the other hand, BMC@BSANPs-A released more drug at both pH which can be due to the ease of release of adsorbed molecules in this formulation in contrast to entrapping the drug molecules in the protein network in BMC@BSANPs-E. Also, the formulations follow the first-order kinetic model, and the release mechanism was controlled by fickian diffusion. The results of cytotoxicity against MCF-7 cancer line showed that despite the higher cytotoxicity of BMC@BSANPs-A in lower concentrations of BMC, after 72 h incubation, at higher concentrations, both formulations show relatively similar cytotoxicity and thus similar efficiency. Ultimately, according to the concluded results, loading of BMC in both formulations could improve its anti-cancer activity. However, due to its smaller size and less drug leakage probability, BMC@BSANPs-E formulation may be more useful for drug delivery applications.
This work was supported financially by the Department of Nanotechnology and Advanced Materials, Materials and Energy Research Center, Karaj, Iran, and Faculty of Pharmacy, Hormozgan University of Medical Sciences, Bandar Abbas, Iran. Authors are grateful for this support.
CONFLICT OF INTEREST
The authors declare no conflict of interest.