Combination of magnetic targeting with synergistic inhibition of NF-κB and glutathione via micellar drug nanomedicine enhances its anti-tumor efficacy
Abstract
Breast cancer is not only one of the most prevalent types of cancer, but also it is a prime cause of death in women aged between 20 and 59. Although chemotherapy is the most common therapy approach, multiple side effects can result from lack of specificity and the use of overdose as safe doses may not completely cure cancer. Therefore, we aimed in this study is to combine the merits of NF-κB inhibiting potential of celastrol (CST) with glutathione inhibitory effect of sulfasalazine (SFZ) which prevents CST inactivation and thus enhances its anti- tumor activity. Inspired by the CD44-mediated tumor targeting effect of the hydrophilic polysaccharide chon- droitin sulphate (ChS), we chemically synthesized amphiphilic zein-ChS micelles. While the water insoluble SFZ was chemically coupled to zein, CST was physically entrapped within the hydrophobic zein/SFZ micellar core. Moreover, physical encapsulation of oleic acid-capped SPIONs in the hydrophobic core of micelles enabled both magnetic tumor targeting as well as MRI theranostic capacity. Combining magnetic targeting to with the active targeting effect of ChS resulted in enhanced cellular internalization of the micelles in MCF-7 cancer cells and hence higher cytotoxic effect against MCF-7 and MDA-MB-231 breast cancer cells. In the in vivo experiments, magnetically-targeted micelles (154.4 nm) succeeded in achieving the lowest percentage increase in the tumor volume in tumor bearing mice, the highest percentage of tumor necrosis associated with significant reduction in the levels of TNF-α, Ki-67, NF-κB, VEGF, COX-2 markers compared to non-magnetically targeted micelles-, free drug-treated and positive control groups. Collectively, the developed magnetically targeted micelles pave the way for design of cancer nano-theranostic drug combinations.
1. Introduction
Cancer nanomedicine research has increasingly caught the scientific society attention due to its potential power to overcome the conven- tional challenges of chemotherapeutic agents, besides its capability to be hybridized with inorganic probes to act as cancer theranostics [1]. Although the initial results of nanotherapeutics is revolutionary in cancer therapy, they cannot be described as “faultless”. The low ther- apeutic outcome may be attributed to insufficient tumor penetration and hence low drug amount reaching the tumor [2]. Besides
nanoparticles have depended on the concept of passive targeting that rely on their intrinsic properties like size and surface properties, the introduction of the concept of “active tumor targeting” has successfully enhanced tumor penetration and increased drug accumulation within tumor cells whether through receptor-mediated or external stimuli like magnetic targeting [3]. Moreover, the therapeutic outcome can be greatly enhanced via the use of multiple therapeutic agents achieving co-delivery. On another avenue, the use of synergistic drug combination has been a cornerstone in cancer therapy to enable use of the optimal drug dose with minimal side effects [4]. Although thunder god vine is a poisonous plant, its root pulp is a source of multiple curative agents like alkaloids, steroids and terpenoids, e.g. celastrol (CST) and tripterine. Recently, CST has caught researchers’ attention due to its multiple biological efficacies. Its efficacy as anticancer agent has been revealed in multiple types of cancer like: hepatocellular carcinoma [5], lung adenocarcinoma [6], glioma [7] and breast cancer [8]. CST activity is mostly correlated to its ability to inhibit NF-κB system [8]. Regarding its chemical structure, CST belongs to triterpene quinine methides having carbons C-2 on ring-A and C-6 on ring-B that are prone to nu- cleophilic attack from thiol moieties of cysteine residues forming covalent Michael adducts [9]. This interaction can be a reason for di- minishing CST activity as a result of interaction with glutathione (GSH). Mainly, GSH consists of three amino acids: glutamate, cysteine, and glycine. Cysteine is the reduced form of cystine which enters the cell in exchange with glutamate via the Xc-cystine/glutamate antiporter, so inhibition of this antiporter will deprive the cell from the key amino acid which is pivotal for glutathione synthesis [10]. From the powerful agents that can be used as Xc- cystine/glutamate antiporter inhibitor is sulfasalazine (SFZ) [11]. Besides, it acts as an inhibitor of the pro-in- flammatory NF-κB which results in increasing the apoptosis of cancer cells [12]. Moreover, GSH is an antioxidant which plays a vital part in drug detoxification and elimination, and therefore, decreasing the GSH level can double the benefits by enhancing CST activity and increasing the oxidative stress. Therefore, we hypothesize that combining SFZ and CST in zein-ChS polymeric micelles could help increasing the ther- apeutic activity.
Among diverse nanoscale platforms that have been designed for targeted delivery of drugs to tumor tissue, polymeric micelles (PMs) can be ascribed as “core-shell” system that incorporates outer hydrophilic polymer to guarantee long circulation time and a hydrophobic poly- meric core that acts as a reservoir for hydrophobic drug [13,14]. Be- sides, PMs have the ability to deliver therapeutic agents either by physical drug encapsulation within their core or by chemical drug conjugation. As promising alternatives for synthetic polymers, the natural polysaccharide chondroitin sulphate (ChS) is a ligand of CD-44 receptors overexpressed by breast cancer cells; thus, providing drug targeting via receptor-mediated endocytosis. Moreover, ChS can be hydrophobically modified via chemical derivatization to the hydroxyl or carboxyl group [15]. On the other hand, zein is a prolamine hy- drophobic corn protein characterized by high percentage (50%) of nonpolar amino acids such as leucine, phenylalanine, alanine and proline [16]. Despite its biodegradability, zein-based nanoparticles can be rapidly captured by the macrophages, as a result of their hydro- phobicity and immunogenicity [17]. Therefore, in our current study, we developed amphiphilic zein-ChS nano-micelles for CD44-targeted combinatorial delivery of SFZ and CST to tumor cells.
In order to further enhance drug accumulation into tumor cells as well as enable tumor imaging, superparamagnetic iron oxide nano- particles (SPIONs) attracted much attention as cancer theranostics [18]. First, the magnetic field can be exploited to achieve magnetic drug targeting. Second, they can be utilized as MRI contrast agent by virtue of their high magnetization value. However, the magnetic property of SPIONs can be double-edged sword as it is also responsible for their high aggregation tendency. Therefore, they are usually coated with polymeric matrix to prevent their aggregation and permit functionali- zation [19].
Based on the abovementioned properties, we have developed a multi-functional theranostic green zein-ChS micelles dually targeted via magnetic and ChS-CD44 receptor mediated endocytosis for synergistic delivery of SFZ and CST to tumor cells. First, SFZ was chemically coupled to zein via tumor cleavable amide bond to form the hydro- phobic core. Second, the hydrophobic zein-SFZ core was chemically coupled to the hydrophilic ChS to form amphiphilic micelles. Third, CST was physically encapsulated within the hydrophobic reservoir. Finally, to enable magnetic targeting and MRI contrast imaging, we have synthesized oleic acid-coated SPIONs to be embedded in the hy- drophobic core of micelles. The developed nanomedicine was thor- oughly investigated in vitro and in vivo.
2. Materials and methods
2.1. Materials
Materials, solvents and reagents used in the current study are illu- strated in details in the supporting information.
2.2. Synthesis of oleic acid-coated SPIONs
SPIONs were prepared based on a previous protocol [20]. Briefly, 120 ml of double distilled water (DDW) was used to dissolve (5.84 g, 21.01 mmol) of ferrous sulphate heptahydrate FeSO4·7H2O at 50 °C under continuous stirring. Thereafter, 140 ml of alkaline 25% ammo- nium hydroxide solution was added. This was accompanied with the formation of black or dark green solution, then the temperature was raised to 90 °C and left to react for further 90 min. the reaction mixture was cooled to RT and an external magnet was then used to separate the synthesized SPIONs from the supernatant. SPIONs were washed with water, ethanol and acetone then left to dry. Oleic acid-coated SPIONs were prepared by addition of 0.8 ml of oleic acid to 1 gm of SPIONs on a magnetic stirrer overnight at room temperature. Excess oleic acid was removed by washing with ethanol and acetone, then SPIONs were left to dry.
2.3. Synthesis of zein-sulfasalazine conjugate (Zein-SFZ, C)
Different ratios of coupling reagents and polymers were tried to optimize the synthesis of zein-SFZ conjugate (C) as detailed in Table S1 (supporting information). First, 3 ml of DMSO was used to dissolve 100 mg of zein. Second, 2 ml of DMSO was used to dissolve SFZ, where its carboxylic groups were preactivated using either DIC/NHS for one h or DIC/Oxyma for 10 min. Finally, zein organic solution was added to the mixture and left to react under mild magnetic stirring for 48 h. The synthesized zein-SFZ conjugate was purified from unreacted drug and excess coupling reagents by dialysis against DMSO for 24 h. Reaction mixture was then dialyzed against DDW/DMSO mixture with gradual increase in water ratio overtime till dialysis against pure water for extra 24 h. Eventually, zein-sulfasalazine was obtained as orange powder after lyophilization (Cryodos-50, Telstar, Spain) at a pressure of 0.5 mbar and a temperature of −50 °C. The reacted SFZ was calculated indirectly from determining the concentration of free SFZ in the dia- lysate.
2.4. Synthesis of sulfasalazine/zein-chondroitin sulphate (SFZ/zein-ChS, M)
Different amounts of ChS were conjugated to the optimized zein-SFZ conjugate C4 and C5 (Table S2-The supporting information). First, ChS was dissolved in 1 ml DDW followed by activation with EDC.HCl/K- Oxyma for 10 min at a constant ratio between ChS and EDC.HCl/K- Oxyma. Second, zein-SFZ conjugate (C4 or C5) was dissolved in 3 ml DMSO and added dropwise to ChS solution under stirring and allowed to react for 48 h at RT. This was followed by purification of the reaction mixture through dialysis against DMSO/water mixture by gradual de- creasing of DMSO ratio for 72 h. Finally, polymeric SFZ/zein-ChS mi- celles were lyophilized to afford white powder.
2.5. Preparation of SPIONs/CST-loaded (SFZ/zein-ChS PMs, F)
Loading of oleic acid-coated SPIONs and CST in the core of PMs (F) was tried in consequent steps (Table 1). For consequent loading, SPIONs were first loaded using oil-in-water emulsion method to obtain the homogenous black suspension to which ethanolic solution of CST was added dropwise. 25 mg of SPIONs were suspended in 5 ml chloroform and added to 50 mg of ChS-zein-SFZ PMs (M5) in 25 ml DDW under sonication in probe-type sonicator (30 min, amplitude 85%, cycle 0.5 sec). Chloroform was evaporated using rotavapor after which centrifugation for 10 min at 5000 rpm was performed. Cen- trifugation was repeated 3 times to remove SPIONs large aggregates. 10 mg of CST were dissolved in 5 ml ethanol and added dropwise to SPIONs loaded PMs. The suspension was then stirred at RT for 24 h till evaporation of ethanol, followed by centrifugation for 10 min at 5000 rpm to remove unloaded CST. Finally, the SPIONs/CST-loaded SFZ-Zein-ChS PMs (F) were frozen at −80 °C and lyophilized to obtain PMs as black powder.
2.6. Characterization of SPIONs & SPIONs/CST-loaded SFZ-Zein-ChS PMs (F4)
2.6.1. Characterization of SPIONs
FTIR and XRD were used to confirm the chemical identity and purity of bare and oleic acid- coated SPIONs. TGA was used to quantify the amount of oleic acid coating SPIONs and TEM was used for de- termination of the particle size which was compared to the size ob- tained from XRD (Scherrar’s equation) as illustrated in the supporting information.
2.6.1.1. Characterization of Zein-SFZ (C5), blank PMs (M5) & SPIONs/ CST-loaded SFZ/Zein-ChS PMs (F3). 1H NMR and FTIR were used for chemical characterization of Zein-SFZ conjugate (C5) and SFZ/zein-ChS PMs (M5), to confirm successful conjugation [21]. The methodologies for determination of particle size, zeta potential, DSC thermograms, CMC measurement and TEM imaging, quantitation of SFZ, CST and SPIONs, drug release, magnetization, hemolytic potential and micellar stability in serum were illustrated in details in the supporting information [22–25].
2.7. In vitro cytotoxicity
The in vitro cytotoxicity of free CST, free SFZ, free combined CST/ SFZ solutions, non-magnetically & magnetically targeted PMs (F3) against MCF-7 breast cancer cells were investigated through MTT assay carried out as described [26] and demonstrated in the Supporting Information. CompuSyn software (version 1) was utilized to calculate Combination Index (CI) and Dose Reduction Index (DRI) to evaluate the efficacy of the designed PMs over free combined drugs.
2.8. Cellular uptake
2.8.1. Confocal laser scanning microscopy (CLSM)
MCF-7 breast cancer cells were used to evaluate the ability of ChS which is a part of the system as a targeting agent. Cellular inter- nalization of free RBITC and RBITC-labelled PMs in presence and in absence of free ChS was assessed using confocal microscopy (LEICA, DMi8, Mannheim/Wetzlar, Germany) as explained previously [27] and explained in the supporting information. The degree of cellular inter- nalization was quantified using flow cytometry analysis (BD FACSCa- libur, BD Biosciences, USA) as described in the supporting information [28].
2.8.1.1. Magnetic targeting using prussian blue staining. Staining with Prussian blue was used to study the effect of magnetic targeting on the cellular internalization of SPIONs-loaded PMs. MCF-7 cells were seeded in 6-well plates (5 × 104/well) for 24 h. Afterwards, media was changed and 1 ml of SPIONs-loaded PMs (F1) was added (Iron concentration 30 µg/ml) in presence and in absence of an external magnet (0.5 T) below the plate for 6 h. After washing and fixation of the cells, 4% potassium ferrocyanide (II) trihydrate acidified with 4% HCl was used for Prussian blue staining for 30 min. Thereafter, nuclear fast red staining was done for 5 min. Eventually, cells were visualized utilizing an optical microscope.
2.9. In vitro magnetic resonance imaging (MRI)
MCF-7 cells were cultured in two 6-well plates overnight (1.5 × 105/well), then media was changed and 1 ml of SPIONs-loaded PMs (F1) was added with various iron concentrations of 100, 300, 500, 700 µg/ml in absence and presence of an external magnetic field below the plates. After 6 h, PBS was used for washing the cells thrice followed by trypsinization and suspension in 0.5% agarose gel. Eventually, MRI scanner 1.5 GE SIGNA explorer (repetition time (TR) = 2564 ms and echo time (TE) = 85 ms) was used for imaging as illustrated in the supporting information [28].
2.10. In vivo studies
2.10.0.1. Animals
Tumor bearing mice were used for assessment of the anti-tumor efficacy of non-magnetically & magnetically targeted SPIONs/CST- loaded PMs (F3) through comparison to free SFZ, free CST and free combined SFZ/CST solution. The followed protocol is detailed in the supporting information.
2.10.0.2. In vivo anti-tumor activity
Ehrlich ascites tumor (EAT) cells were used for induction of EAT in mice as described in the supporting information. Mice were divided into 7 groups, 8 mice each. Groups were classified into negative control (healthy mice), positive control (EAT bearing mice which remain un- treated), free SFZ, free CST, free combined SFZ/CST solution (in a co- solvent of DMSO/PEG 400/saline 2.5:47.5:50 %v/v), non-magnetically & magnetically targeted SPIONs/CST-loaded PMs, where a permanent magnet was applied on the tumor side for 2 h after injection [29]. The
free drugs or PMs were intravenously (i.v.) injected into EAT bearing mice through the tail vein eq. to 0.32 mg/kg SFZ and 2 mg/kg CST two times weekly for three weeks [30]. According to the approved animal protocol, animals were terminated when tumor grew to be larger than 1000 mm3 or in case of tumor ulceration. After the last dose, cervical dislocation was used to sacrifice all the surviving animals. The excised tumors were washed by cold saline then split into two parts. The first part was used for histopathological examination. 10% neutral buffered formalin was used for its fixation followed by embedding in paraffin blocks. The second part was used for determination of tumor markers and oxidative stress parameters after homogenization using PBS.
2.11. Statistical analysis
Statistical analysis is explained in the supporting information.
3. Results and discussions
In the current work, we have rationally designed multi-functional drug delivery system in a trial to overcome the problem of drug un- desirable hydrophobicity, achieves tumor targeting using more than one mechanism and enables tumor imaging. We synthesized co-poly- meric conjugate (M5) to undergo self-assembly forming nano-micelles in physiological media with a hydrophobic core fabricated of zein-SFZ to encapsulate both SPIONs and CST and a hydrophilic shell of ChS to achieve tumor targeting and evade the immune system. Fig. 1 illustrates the steps of preparation of SPIONs/CST-loaded SFZ-zein-ChS PMs.
3.1. Oleic acid-coated SPIONs were synthesized via co-precipitation
SPIONs were prepared by the co-precipitation technique from single iron precursor followed by oleic acid coating. In order to confirm the successful synthesis of SPIONS, FTIR and X-ray diffraction were per- formed. The FTIR spectra showed the presence of bands at 578 and 3401 cm−1 attributed to Fe-O and OH vibrations, respectively, which confirmed its identity (Fig. 2a). Upon comparison with oleic acid- coated SPIONs spectrum, a band appeared at 3424 cm−1 which is due to OH stretching of the COOH group of oleic acid. The band at 2922 cm−1 is assigned to the CH3 stretching band, which confirms the presence of the oleic acid [31]. In addition, the XRD pattern revealed the synthesis of well-defined crystalline magnetite (Fig. 2b). All the diffraction peaks were matched with magnetite standard peaks in the X- ray diffraction atlas (JCPDS card 19-0629) [32]. Complete absence of any peaks of impurities revealed the purity of the synthesized SPIONs. Scherrer’s equation was used to calculate particle size where (311) peak was selected to input its (Full width at half Maximum) FWHM parameters and X-ray diffraction peak position [33]. Moreover, the morphology and diameter of bare and oleic acid-coated SPIONs were measured and confirmed via TEM images. Spherical bare and oleic acid- coated SPIONs were observed (Fig. 2c and 2d). Image analysis was performed using 200 particles to obtain the corresponding histograms (Fig. S1- supporting information). As clearly indicated in Table S3 in the supporting information, the mean diameter of bare SPIONs reached 50.05 ± 9.52 nm which was significantly reduced to
16.29 ± 3.25 nm after oleic acid coating. The high particle size of bare SPIONs could be explained by van der Waals forces and the magnetic dipolar forces between nanoparticles. Oleic acid has contributed sig- nificantly in decreasing the particle size by decreasing the forces be- tween the particles. The difference between the size obtained from TEM images and Scherrar’s equation is indicated in the supporting in- formation.
3.2. Amphiphilic Zein/SFZ-ChS PMs were synthesized via carbodiimide coupling
The zein-SFZ conjugate was synthesized by simple amide linkage between COOH of SFZ and NH2 groups of zein (Table S1-The supporting information). Fig. 3a and 3b shows the 1H NMR spectra of zein and SFZ- zein conjugate, respectively. The integration in the range 0.79–2.07 ppm is referred to the aliphatic protons, while the range 6.59–8.03 ppm corresponds to the aromatic protons. Comparing the integration of protons in both spectra, the zein-SFZ conjugate spectrum indicates an increase in the integration of protons in the aromatic re- gion. Further analysis is difficult to be achieved due to the complex structure of zein as a natural protein. On the other hand, the conjuga- tion between zein-SFZ conjugate and chondroitin sulphate (ChS) could be proved from the comparison between the aliphatic and aromatic protons (Fig. 3c). All protons in ChS spectrum were aliphatic; therefore, the 1H NMR spectrum of SFZ/zein-ChS showed an increase in the in- tegration ratio of aliphatic protons to aromatic protons from 2.31:1 to 2.44:1 (Fig. 3d) compared to 1H NMR spectrum of Zein-SFZ (Fig. 3b). The FTIR spectrum of SFZ revealed the presence of a peak at 1359 cm−1 referred to the asymmetric stretching of S]O group which remained present and characteristic in the spectrum of zein-SFZ conjugate (Fig. 4a). Furthermore, a broad peak was noticed at 3500–2500 cm−1 characteristic to the OH stretching of the carboxylic acid that dis- appeared after conjugation with zein while a predominant peak ap- peared at the range of 3600–3200 cm−1 which was correlated to the phenolic OH. The C]O group of the amide linkage in zein was ob- served at 1658 cm−1 [21]. Zein-SFZ spectrum showed an additional peak at 1640 cm−1, which may be due to the formation of new amide bonds. On the other hand, the IR spectrum of pure ChS was char- acterized by the presence of a peak at 1634 cm−1 due to the amide I band, which was overlapped by the slightly broad C]O band at 1658 cm−1 after conjugation with Zein-SFZ conjugate. In addition, the S]O stretching at 1259 cm−1 in pure ChS was shifted after conjugation to 1239 cm−1. The FTIR spectrum of SPIONs/CST-loaded PMs (Fig. 4a) in comparison to the previously discussed FTIR spectrum of ChS-zein- SFZ PMs showed additional bands at 578 and 1700 cm−1 corresponding to Fe-O vibration of SPIONs and C]O stretching vibration of celastrol carboxylic acid group, respectively [34].
Particle size is a very influential criteria that contributes significantly in the fate of NPs in the body. Therefore, ChS-SFZ-Zein PMs
(M5) was selected as the optimal formulation by showing the smallest size (226.3 ± 2.8 nm) among the micelles obtained using different ratios of ChS and SFZ-conjugated zein used in the conjugation reactions (Table S2-The supporting information). The micelles demonstrated a zeta potential of −25.2 ± 0.48 mV indicating good colloidal stability which explained by the presence of the anionic polysaccharide ChS [35].
3.3. The micelles displayed low critical micelle concentration (CMC)
CMC was assessed using pyrene which is a fluorescent hydrophobic probe whose fluorescence intensity is affected by the media polarity. Due to its hydrophobic character, pyrene had a little fluorescence in- tensity at low co-polymer concentration. Upon increasing the polymer concentration, the co-polymer tends to self-assemble in to micelles where pyrene molecules were incorporated into its core (Fig. 4b). From plotting the intensity ratio (I343/I333) against log concentration of the polymeric micelles, the CMC value was found to be 0.056 mg/ml (Fig. 4c). The low CMC value is a desirable criteria that gives an evi- dence of the system stability upon the in vivo dilution [36].
3.4. Spions & CST were efficiently loaded into SFZ/zein-ChS PMs
The SFZ/zein-ChS micelles were successfully synthesized and loaded with SPIONs using oil-in-water emulsification method and CST using solvent evaporation method (Table 1). Loading of SPIONs and CST into the micelles was tried separately to investigate the effect of loading on the particle size of micelles. Loading of SPIONs decreased the micelle size to 155.1 ± 4.5 nm with a PDI of 0.088 (F1) while CST loading decreased the micelle size to 153.5 ± 2.1 nm with a PDI of 0.28 (F2). Moreover, CST and SPIONs loading in the same formula has decreased the particle size to 154.4 ± 2.8 nm with a PDI of 0.202 zeta potential −34.8 mV (F3) (Fig. 4d and e). The decrease in the particle size of micelles after the entrapment of SPIONs and CST was expected due to the enhanced hydrophobicity and packing of the co-polymer core [37].
As demonstrated in Table 1, different feeding ratios and activating agents were tried to enhance the conjugation efficiency (%CE) of SFZ where C5 achieved the highest CE (13.5%). Loading of CST alone into the micelles resulted in 74.8% entrapment compared to 86.7% en- trapment when co-loaded with SPIONs into the micelles. This can be attributed to the entrapment of CST in the oleic acid layer surrounding SPIONs, which revealed the positive effect of SPIONs on drug loading efficiency. The decrease in %EE of CST and SFZ after loading of SPIONs was expected due to the increase in the total mass of the PMs.
In the DSC thermogram of SFZ, a strong endothermic melting peak at 258 °C was detected indicating the crystalline nature of SFZ (Fig. 5a) [38]. No phase transitions or peak fusions were detected in the ther- mogram of zein due to its amorphous nature [39]. Upon observation of the thermogram of zein-SFZ conjugate, a complete absence of the characteristic peaks of SFZ was observed which suggests its presence in the amorphous state [40]. ChS demonstrated an endothermic peak (80–100 °C) due to water evaporation in addition to the pyrolysis peak at about 240 °C [41]. The thermogram of the micelles revealed absence of the ChS characteristic peaks after conjugation.TGA was performed to oleic acid-coated SPIONs and SPIONs/CST-loaded PMs (Fig. 5b & 5c). The percentage of oleic acid was found to be 16.03% while the degradation of the PMs and oleic acid caused 88% weight loss. Therefore, the Fe3O4 content in the micelles was calculated to be 12.0% which agreed with the value obtained from ICP-OES. The magnetization value of bare SPIONs, oleic acid-coated SPIONs and SPIONs/CST-loaded micelles was determined. The hysteresis loop for the three samples had the distinctive sigmoidal shape of super- paramagnetic materials. As the field increased to 2 kOe, the magneti- zation increased in a non-linear pattern and when decreased, the magnetization falls to zero. The saturation magnetization (field H = 2 kOe) of bare SPIONs at room temperature was found to be 68.873 emu/ g, while it was 70.602 and 0.88594 emu/g for oleic acid-coated SPIONs and the SPIONS/CST-loaded PMs, respectively. The value of the sa- turation magnetization was corrected according to the percentage of SPIONs from the TGA analysis [42]. The values after correction to oleic acid-coated SPIONs and SPIONs/CST-loaded PMs was 84.05 and 7.41 emu/g, respectively (Fig. 6a). It was observed that the magneti- zation value for the oleic acid-coated SPIONs was higher than that of bare SPIONs which can be explained by the presence of magnetic dis- order on the bare SPIONs surface due to external media interactions. Therefore, coating with oleic acid decreases the interaction and the spins disorder on the surface of SPIONS, thus increasing the magneti- zation value. The apparent decrease in the value of magnetization of the micelles was attributed to the presence of a copolymer layer around SPIONs [43]. XRD analysis of the loaded micelles revealed the dis- appearance of most magnetite peaks due to the thickness of the copo- lymer (Fig. 2b).
3.5. The micelles exhibited controlled drug release
In vitro release of free CST, CST-loaded PMs and SPIONs/CST-loaded PMs was performed using dialysis bag method at pH of 7.4 to simulate the physiological conditions. As indicated in Fig. 6b, free CST showed burst release pattern with about 94.0% of drug was released within the first 4 h. On the other side, both CST-loaded PMs and SPIONs/CST- loaded PMs showed very slow release of CST without initial burst effect. About 8.30% of CST has released from the magnetically guided micelles within the first h, followed by a slow release over the entire period of time till reaching 38.46% after 72 h. Besides, 11.20% of CST was re- leased from SPIONs-free micelles after 4 h which was followed by slow release till reaching 44.86% after 72 h. These results showed that presence of SPIONs has significantly (p < 0.05) affected the release pattern of CST from the micelles which may result from CST entrap- ment in the oleic acid layer surrounding SPIONs. Regarding SFZ release, it was not detected all over the release test period which may be ex- plained by the stable amide bond between SFZ and zein [44]. This re- lease behavior will in turn prevent the premature drug release in sys- temic circulation while enable the site-specific release in tumor tissue upon bond cleavage by endolysosomal enzymes. 3.6. The micelles displayed good morphology & colloidal stability TEM micrographs showed that both micelles have spherical shape with no aggregation (Fig. 6c). The size SPIONs/CST-loaded PMs was about 82.96 nm. The size estimated by TEM was apparently much smaller than the DLS measured diameter which was explained by the shrinkage accompanied by dehydration during sample preparation for TEM [27]. Concerning their colloidal stability, the size of SPIONs/CST-loaded PMs was found to be 154.4 ± 2.8 nm which insignificantly increased to 159.2 ± 3.2 nm after 3 months of storage (Fig. 6d) in- dicating their excellent stability profile. At the same time, the zeta potential of micelles was almost not changed (-35.8 mV, Fig. 6d). 3.7. The micelles exhibited favorable hemolytic compatibility & serum stability Upon incubation with RBCs, the prepared SPIONs/CST-loaded PMs exerted 1.19, 2.3, 4.6 and 4.76% hemolytic rate at 0.25, 0.5, 1 and 2% concentration, respectively, (Fig. 6e). Consequently, the PMs have succeeded to prove their safety based on ASTM E2524-08 standard which requires hemolytic rate below 5% (Fig. S2-supporting informa- tion) [45]. The low hemolytic rate can be explained by the high ne- gative charge of ChS corona that diminishes the interaction with RBCs [46]. Moreover, after incubation with 10% FBS, there was a size in- crease of the micelles to about 177.4 ± 8.2 nm which decreased to 140.3 ± 3.1 nm after 6 h (Fig. S3- supporting information). The initial elevation of the size can be ascribed to the formation of protein corona around the micelles due to protein adsorption. Meanwhile, the latter size decrease was a result of the resultant osmotic pressure causing water escape from the micelles core [47]. 3.8. The micelles demonstrated enhanced cytotoxicity The cytotoxic effect of non-magnetically and magnetically targeted PMs was tested on MCF-7 and MDA-MB-231 breast cancer cells, com- pared to the cytotoxic effect of free SFZ, free CST and the free SFZ/CST combination (Fig. 7a & 7b). For both cell lines, blank PMs demonstrated negligible cytotoxicity to the cells for 24 h where the viability reached greater than 96% which evidenced their safety. Regarding MCF-7, The IC50 of free SFZ and free CST at 24 h was 84.57 and 5.16 μM, respectively. On the contrary, the IC50 of the combined free SFZ/CST solution at 24 h was 16.97- and 1.036- fold lower than that of SFZ and CST respectively, which proves to the synergistic cytotoxic effect of drug combination. Co-loading of both drugs in the PMs de- creased the IC50 to 4.35 μM. Furthermore, the application of external magnet to guide the PMs to the cells caused a supplemental reduction in the IC50 to 4.16 μM (Table 2). A supplementary statistical analysis was carried out utilizing CompuSyn software (version 1) designed by Chou and Talalay [48]. The combination index (CI) was analyzed to depict the synergism, additive effect or antagonism [49]. The CI of non-mag- netically targeted and magnetically targeted PMs was 0.73 and 0.69, respectively revealing the ability of PMs to enhance the synergistic ef- fect of the drug combination. Furthermore, the Dose Reduction Index (DRI) of SFZ in the non-magnetically and magnetically targeted PMs was 136.29 and 142.77, respectively. While the DRIs of CST were 1.38 and 1.45 in the non-magnetically and magnetically targeted PMs, re- spectively. For MDA-MB-231 cell line, The IC50 of free SFZ and free CST at 24 h was 63.931and 4.901 μM, respectively. The combination be- tween SFZ/CST has successfully decreased the IC50 to 13.157 and 1.036 fold lower than that of free free SFZ and free CST respectively. The PMs has successfully decreased IC50 to 4.35 μM. Magnetic targeting has additionally decreased the IC50 value to 4.3225 μM (Table 3). The same statistical analysis was performed utilizing CompuSyn software (version 1) which revealed that the CI of non-magnetically targeted and mag- netically targeted PMs was 0.81 and 0.76 respectively. Moreover, the Dose Reduction Index (DRI) of SFZ in the non-magnetically and mag- netically targeted PMs was 98.03and 103.75, respectively. While the DRIs of CST were 3.922 and 3.706 in the non-magnetically and mag- netically targeted PMs, respectively. The enhancement of cytotoxic ef- fect of the PMs could be associated with the promoted cellular uptake into breast cancer cells as a result of the have high affinity interaction of ChS with the overexpressed CD-44 receptors [50]. Accordingly, the cytotoxic effect of the magnetically guided PMs was further enhanced due to the combined effect of binding to CD44 receptors and the pre- sence of an external magnetic field which guide the PMs to their des- tination avoiding off-target distribution [51]. Schleich et al. have suc- cessfully proved the power of using magnetic targeting and active targeting and the influential effect of using dual targeting to increase cellular uptake and tumor penetration [3]. 3.9. The micelles exhibited enhanced cellular internalization To visualize the PMs inside the cells using confocal fluorescent microscopy, the micelles were rendered fluorescent through conjuga- tion to RBITC (Fig. 8a). Cellular uptake of RBITC-labeled PMs was compared to free RBITC after incubation with MCF-7 cells for 4 and 24 h. RBITC and RBITC-labeled PMs appeared as red dots inside the cells where the intensity of RBITC-labeled PMs exceeded that of free RBITC. Moreover, the fluorescence intensity has been enhanced with time revealing the time-dependency of the cellular uptake process. In a trial to prove the mechanism of PMs cellular uptake, cells were in- cubated with free ChS to block CD44 receptors before the addition of RBITC-labeled PMs [52]. As expected, the intensity of red fluorescence in case of free ChS incubated with RBITC-labeled PMs was lower compared to RBITC-labeled PMs without incubated free ChS. Flow cytometry was utilized for quantitative determination of the intensity of the fluorescence inside the cells (Fig. 8b). The values of Mean Fluorescence Intensity (MFI) indicated the presence of a sig- nificant difference (P < 0.05) between the internalization of RBITC- labeled PMs and free RBITC after 4 and 24 h (Fig. 8c). Moreover, the addition of free ChS caused a 3.8- and 11.52-fold significant decrease (P < 0.05) in MFI after 4 h and 24 h, respectively compared to RBITC- labeled PMs. Those findings proves the efficacy of ChS as a targeting moiety by virtue of its affinity to CD-44 receptors [53]. Furthermore, Prussian blue staining was performed to provide further confirmation of the micelle cellular uptake as it forms blue dots upon reaction with ferric ions (Fig. 9a). The control cells showed complete absence of blue dots whereas obvious blue color was observed upon incubation with SPIONs-loaded PMs. The intensity of the blue color was increased in the presence of external magnetic field which augmented that the cellular internalization of micelles was accounted to the double tumor-targeting mechanisms [54]. 3.10. In vitro magnetic resonance imaging (MRI) The capability of SPIONs-loaded PMs to be used as an imaging agent was evaluated using 1.5 T clinical MRI scanner. Different concentra- tions of SPIONs-loaded PMs (F1) were incubated with MCF-7 cells for 6 h in presence and absence of an external magnetic field. Cells treated with SPIONs-loaded PMs showed decreased intensity of the obtained signal with increasing the Fe concentration which was attributed to shortening T2 relaxation time by SPIONs taken up by the cells (Fig. 9b). Moreover, the exposure of the cells to the external magnetic field im- proved the cellular uptake of SPIONs and consequently caused further increase in the intensity of the signal with the same Fe concentrations. This test revealed the capability of our designed PMs to be used for drug delivery and monitoring them using MRI simultaneously [28,55]. 3.11. The micelles showed powerful in vivo anti-tumor efficacy Treatment of EAT-bearing mice lasted for three weeks with tumor size monitoring before termination. The untreated positive control showed the highest percentage increase of tumor volume reaching 201% which was significantly higher than those of free SFZ and free CST (170.5 and 146.12%, respectively) (Fig. 10a and 10b) [56]. The combination between free SFZ and free CST achieved synergistic anti- cancer effect which decreased the percentage increase of tumor volume to 110% revealing the efficiency of our combination rational. Meanwhile, the entrapment of this successful drug combination in the designed non-magnetically targeted PMs showed only 78.24% increase in tumor volume indicating successful drug accumulation within tumor via both passive and active targeting. Additionally, applying external magnetic field on the tumor to potentiate the tumor magnetic targeting has significantly decreased the percentage increase of tumor volume to 36.26% (p < 0.05). Furthermore, the average body weight of mice was monitored over the treatment period (Fig. 10c). Free CST and free CST/ SFZ combination caused a significant weight loss (p < 0.05) when compared to positive control, non-magnetically and magnetically tar- geted PMs. It is mainly due to the fact that treatment with CST has side effects that loading in PMs has succeeded to overcome [57]. Moreover, Kaplan-Meier survival curve of mice bearing Ehrlich tumor model was used to analyze the in vivo anti-cancer activity of our designed PMs (Fig. 10d). The application of magnetic targeting has enhanced the survival by 83.3% compared to the non-magnetically targeted PMs (75%). Collectively, the use of PMs has succeeded in diminishing the toxicity of free CST/SFZ drug combination (20% survival). 3.12. The micelles showed enhanced anti-tumor efficacy 3.12.1. The micelles suppressed the expression of the pro-inflammatory tumor necrosis Factor-alpha (TNF-α) CST was found to have an inhibitory effect on TNF-α level besides its ability to enhance the TNF-α induced apoptosis, in part, via NF-κB suppressing mechanism [8,58]. Meanwhile, the inhibitory effect of SFZ on TNF-α has been proved [59]. In the current study, the expression level of TNF-α, assessed using quantitative real time RT-PCR (qRT- PCR), was significantly reduced by both non-magnetically and mag- netically targeted PMs compared to free drugs (p < 0.05) (Fig. 10). The previous finding indicates the significant effect of our designed targeted PMs on the apoptotic effect of CST and STZ. 3.12.2. Cyclooxygenase-2 (COX-2) COX-2 is an inducible inflammatory enzyme whose overexpression is strongly related to cell proliferation besides tumor invasion in solid tumors including breast cancer [60]. Additionally, COX-2 is one of the responsive genes to the activated NF-κB whose prevention of its acti- vation under the effect of CST and SFZ would diminish its level too[87] [61,62]. Moreover, its expression is related to TNF-α expression level [63]. Accordingly, qRT-PCR was used to assess the expression level of COX-2 in the tumor tissues (Fig. 11). Non-magnetically and magneti- cally targeted PMs have successfully decreased COX-2 both mRNA and protein levels in tumor tissues when compared to free drugs and positive control (p < 0.05) which might contribute to the tumor proliferation inhibitory effect of non-magnetically and magnetically targeted PMs. 3.13. The micelles reduced the expression of proliferative, inflammatory and angiogenic markers Histopathological analysis revealed significant higher necrosis score of the tumor tissues treated with the magnetically targeted PMs more than that treated with non-magnetically targeted PMs which is in turn higher than positive control-treated group (p < 0.05) (Fig. 12a and b). Necrosis is considered a very efficient killing strategy to cancer cells by chemotherapy, as it will elicit the immune response opposing the im- munity suppression by chemotherapeutic agents [64].therefore, ne- crosis is considered a privilege in case of chemotherapy [65]. Tissue specimens from the positive control and free combined drugs-treated groups were found to be strongly stained for the proliferation marker Ki-67 when compared to the non-magnetically and magnetically tar- geted PMs (Fig. 12c). The analysis revealed significant high expression level (95.38%) of Ki-67 in tissue specimen of positive control group (Fig. 12d). On the other hand, the expression level of Ki-67 decreased to 52.93% and 34.67% in the groups treated with non-magnetically and magnetically targeted PMs, respectively. Those findings manifested the ability of the PMs to reduce cancer cells proliferation and therefore restrain the tumor growth [66]. NF-κB is a transcription factor that upon activation regulates sub- sequent signaling pathways that promote cancer development, pro- gression, metastasis, angiogenesis and proliferation [67]. CST and SFZ are reported to inhibit NF-κB activation through inhibition of IKK [12,68]. Upon performing IHC analysis for protein level determination, the positive control tissue of cancer specimen was strongly stained with high expression (91.59%) of NF-κB (Fig. 13a and b). The combination of CST and SFZ have significantly decreased its expression level to 69.27% (p < 0.05). Furthermore, the enhanced delivery of both drugs via the developed micelles has decreased the expression level of NF-κB to 49.18% versus positive control. In addition, the expression level of NF- κB in specimen treated with the magnetically targeted PMs decreased to 30.92% versus positive control sample. Vascular endothelial growth factor (VEGF) signals vascularization through tyrosine kinase receptor identified on the epithelial cells and plays an essential role in the regulation of tumor angiogenesis [69]. High expression of VEGF may be also related to poor prognosis [70]. CST showed considerable potential to be used as anti-angiogenic agent due to its ability to inhibit the VEGF expression [71]. Therefore, a high expression level (95.18%) of VEGF was revealed in the positive control specimen compared to 74.02% expression in the free combined drugs treated specimen (Fig. 13c and 13d). On the contrary, treatment with drugs-loaded PMs decreased its expression level to 63.91% while the application of magnetic targeting diminished its expression to 37.25%. This revealed the ability of the drug-loaded PMs and the magnetic targeting to potentiate the effect of CST as anti-angiogenic agent [72]. Additionally, COX-2 expression level was quantified through im- munostaining which agreed with results obtained from quantitative real time RT-PCR analysis. Its expression level reached 84.77% in the po- sitive control tissue specimen while treatment with the free combined drugs decreased its level to 62.94%. On the other side, the expression of COX-2 was reduced to 46.33% and 21.11% in the non-targeted and magnetically targeted PMs respectively (Fig. 13e and f). All the above- mentioned findings have proved the inhibitory effects of CST and SFZ on the inflammatory and angiogenic markers (NF-κB, VEGF and COX-2) and those inhibitory effects have been significantly enhanced by their incorporation into the dual CD44-receptor-mediated and magnetically targeted micelles. 3.14. Determination of oxidative stress parameters The extent of oxidative stress as indicative of tissue damage was estimated by measuring the levels of both MDA, a product of lipid peroxidation, and GSH in tumor tissues. SFZ caused a slight increase in the level of MDA that reached 4.07% while free CST caused a higher increase to 15.06% compared to the positive control (235.11 ± 22.52 nmol/g protein/g tissue) (Fig. 14a). Treatment with free drug combination resulted in 17.83% increase in MDA level, whereas their incorporation in the non-magnetically and magnetically targeted micelles reduced an increase in the MDA level to 28.58% and 79.85%, respectively which indicates high degree of oxidative stress on the cells and hence high degree of tissue damage [73] . The administration of free SFZ decreased the level of GSH in tumor tissues by 44.19% compared to the positive control (6 ± 0.56 nmol/g protein/g tissue) which is expected due to the inhibitory effect of SFZ on the Xc-cystine/glutamate antiporter (Fig. 14b). Furthermore, CST exerts an inhibitory effect to the level of GSH demonstrated as 37.07% reduction when compared to the positive control. Coadministration of CST and SFZ resulted in 50.30% decrease in the level of GSH. Conse- quently, when the two drugs were incorporated in non-magnetically and magnetically targeted PMs, the GSH level was decreased by 56.41% and 60.1%, respectively which potentiates the activity of CST and evidenced disruption of the mitochondrial membrane besides apoptosis [11,71]. 4. Conclusion In the current study, celastrol (CST) and sulfasalazine (SFZ) were formulated in polymeric micelles composed of the natural protein zein and polysaccharide chondroitin sulfate for breast cancer therapy. Physical loading of CST within the hydrophobic micellar core and chemical conjugation of SFZ to zein core have enhanced their poor solubility and enabled their injection into the circulation. The physical encapsulation of SPIONs in the core of PMs enables dual targeting via both active and magnetic targeting. Besides, it has added the capability to the micelles to be used for MRI imaging. The dual targeted micelles displayed higher cellular internalization and higher toxicity against MCF-7 and MDA-MB-231 breast cancer cells. Magnetically targeted SPIONs/CST-loaded PMs showed superior anti-tumor activity in vivo through inhibition of NF-κB activation, VEGF and COX-2. Histopathological examination and reduced expression level of Ki-67 have also confirmed the superior activity. Collectively, all the results have manifested the outstanding efficacy of CST and SFZ as a combi- nation therapy against breast cancer in the form of magnetically tar- geted zein-chondroitin PMs. The harmony between all the parts con- stituting the system was clear. Here we find, zein the hydrophobic core which carry CST, SFZ and SPIONs. Also, ChS which play a pivotal role in the targeted delivery of CST and SFZ together with SPIONs. Upon reaching the final destination of CST and SFZ inside the cell, they exerted their therapeutic effect which results in inhibition of the main factors which regulates tumor development, angiogenesis and pro- liferation.