De novo designed lactoferrin–oleic acid-loaded chitosan nanoparticles with improved activity and selectivity toward four human cancer cells as compared to conventional complexes

1 Protective and Therapeutic Proteins Laboratory, Protein Research Department, Genetic Engineering and Biotechnology Research Institute, City of Scientific Research and Technological Applications, Alexandria, Egypt. 2 Medical Biotechnology Department, Genetic Engineering and Biotechnology Research Institute, City of Scientific Research and Technological Applications, Alexandria, Egypt. 3Biological Sciences Department, Faculty of Science, King Abdulaziz University, Jeddah, Saudi Arabia.


INTRODUCTION
The glycoprotein of lactoferrin (Lf) represents a milkderived multipurpose iron chelator that also exists in other mammalian cell secretions, for instance, saliva and tears, in addition to vaginal and seminal fluids (Ward et al., 2002). Some clinical studies confirmed the presence of proteins that prevent cancer in milk (Tsuda et al., 2002). Lactoferrin has been assigned with its suppressing effect on cancer cells propagation in addition to its anticarcinogenic, antioxidant, and anti-inflammatory capabilities in vitro and in vivo (Matsuda et al., 2006;Pan et al., 2007).
More remarkably, it has been found that the Lf gene downregulation might be linked with elevated breast tumors' incidence rates (Furmanski et al., 1989). Also, the exogenous Lf as well as its variants supply could proficiently suppress the growth of cancer in vitro and in vivo (Xu et al., 2010;Yamada et al., 2008). Bovine lactoferrin was proven to have an in vitro anticancer effect on many breast cancer cells and gastric tumor cells (Amiri et al., 2015;Zhang et al., 2015b). Nevertheless, the exact mechanisms by which lactoferrin performs its cytotoxicity activity against cancer cells are still quite unknown.
Recently, Fang et al. (2014) formed an antitumor complex from Lf and OA that was similar to the complex consisting of human α-lactalbumin and OA (HAMLET). They demonstrated that the bLf-OA complex prompts apoptosis in tumor cell lines, thus inducing high antitumor activity. The interaction of both lactoferrin and α-lactalbumin with oleic acid was found to be via van der Waals forces in addition to hydrogen bonding, but Lf showed about twice the binding capability of α-lactalbumin with OA. They reported that both Lf and OA were responsible for the antitumor activity of their complex, which was not the case with HAMLET that had an OA-dependent antitumor activity. In another study, two antitumor lactoferrin-OA and lactoferrin-linoleic acid complexes were successfully prepared, which displayed a perceptive prospect of cancer treatment (Zhang et al., 2015a).
In view of the established and significant association between lactoferrin and cancer and the previous studies confirming high OA/Lf-dependent antitumor activity of bLf-OA complex besides considering the problem of OA adverse effects on both cancer and normal cells, the present work studied the antitumor activity of formulated camel lactoferrin-OA (cLf-OA) and human lactoferrin-OA (hLf-OA) nanocomplexes.

Anticancer agents
Used milk was collected and processed for camel lactoferrin (cLf) and human lactoferrin (hLf) purification. Both lactoferrins were purified from processed milk at our lab as formerly reported (Almahdy et al., 2011;El-Fakharany et al., 2013). The purified preparations of camel and human lactoferrins were filter sterilized via a syringe filter of 0.22 µm pore size from TPP (St. Louis, MO) and then preserved in the freezer at −20°C. Cell culture-tested OA with purity of ≥99.0% was purchased from Sigma-Aldrich (St. Louis, MO).

Determination of cLf and hLf activity
The activity of lactoferrins was estimated as reported by Ye et al. (2000), with slight modifications. About 50 µl of each lactoferrin were added to a mixture comprising 15 µl Tris-HCl buffer at a concentration of 50 mM and pH 8.0 with 75 µl dihydronicotinamide-adenine-dinucleotide-phosphate at a concentration of 300 µM, nitroblue tetrazolium (NBT) at a concentration of 300 µM, and phenazinmethosulfate at a concentration of 30 µM. The absorbance at 580 nm was recorded initially and after 5 minutes of reaction. The used positive control was L-ascorbic acid. The activity of lactoferrins was then calculated from a standard curve made by the use of a series of NBT standards. Lactoferrins activity was reported as international units per milligram of protein.

Preparation of conventional lactoferrin-OA complexes
Different concentrations of camel and human lactoferrins were prepared by dissolving inphosphate buffered saline (PBS) buffer (10 mM, pH 8.0). OA was mixed with each protein solution at 50 molar equivalents of OA to lactoferrin, then vortexed for 30 seconds using FINEVORTEX (FINEPCR, Gunposi, Korea) and incubated for 20 minutes at 45°C in MultiTemp III water bath (Pharmacia Biotech, Uppsala, Sweden). After removal of excess OA by centrifugation at 4°C, the prepared complexes were ultrafiltrated by Centricon with a membrane of 3 kDa cut-off (Merck, Darmstadt, Germany) (Uversky et al., 2017).

Evaluation of OA concentration in the prepared lactoferrin-OA complexes
OA in the prepared lactoferrin-OA complexes was evaluated by Duncombe colorimetric method (Duncombe, 1963). In brief, the prepared protein-OA complexes were shaken with a solution of chloroform containing copper. According to this method, the estimation of copper in chloroform by use of a color developer (sodium diethylthiocarbamate) is equivalent to OA in the protein-OA complexes, since fatty acids like OA can form a complex with copper ions which is soluble in chloroform. A copper reagent was prepared by mixing 1 M triethanolamine (9 volumes), 1 N acetic acid (1 volume), and 10% (w/v) copper sulfate (10 volumes). The prepared protein-OA complexes (500 µl) or standard OA (500 µl) were vigorously shaken with about 2.5 ml of the copper reagent using vortex mixer. Afterward, chloroform (5 ml) was added and vigorously shaken for an additional 1 minute. Subsequently, 3 ml of the lower (chloroform) layer was separated and mixed with 0.1% sodium diethylthiocarbamate dissolved in butanol at a volume of 500 µl. Finally, the absorbance of samples was measured at 440 nm.

Preparation of nanocomplexes
The cLf/hLf-OA complex-loaded chitosan nanoparticles were formulated according to the ionic gelation protocol, as formerly reported by Anitha et al. (2011). Chitosan at a concentration of 2 mg/ml was prepared by dissolving in 0.1% aqueous acetic acid with stirring continuously. pH adjustment to 5.5 was carried out by 1 N NaOH using Jenco pH meter (Jenco, San Bernardino, CA). The prepared cLf-OA and hLf-OA complexes at 1 mg/ml were added dropwise over the course of 1 hour to the chitosan solution. Dextran sodium sulfate at a concentration of 0.5 mg/ml was added dropwise to the above-mentioned solution as a cross-linker to obtain nanocomplexes. The nanoformulated cLf/hLf-OA complexes were precipitated by centrifugation twice at 12,000 rpm and 4°C for 40 minutes and the precipitates were suspended in PBS then freeze-dried.

Characterization of nanoformulated cLf/hLf-OA
The protein content in the supernatant was determined as absorbance at 280 nm. Loading capacity for cLf-or hLf-OA and encapsulation efficiency of nanoformulations were evaluated. The mean hydrodynamic size of the nanoformulations was assessed by the particle size analyzer Zetasizer Nano ZS (Malvern, UK).

Detecting cytotoxic effect on normal and cancer cells
The normal human lung fibroblast WI-38 cell line CCL-75 (ATCC, Manassas, VA) was used to detect cytotoxicity of cLf, hLf, cLf-OA, hLf-OA, and their nanocomplexes, as well as OA. WI-38 cell line at a density of 1 × 10 4 cells per well was maintained in the Dulbecco's modified Eagle medium (DMEM) growth medium (Lonza, Verviers, Belgium) supplemented with 10% fetal bovine serum (FBS, Lonza) in a 96-well cell culture flat-bottom microplate (Greiner, Frickenhausen, Germany) and incubated at 37°C in 5% CO 2 incubator (BINDER, Tuttlingen, Germany). After cell attachment for 24 hours, WI-38 cells were treated with OA, cLf, hLf, their complexes, and nanocomplexes at serial concentrations and then incubated for 72 hours. The viability of cells was assessed by the MTT technique Mosmann, 1983). Twenty microliters of MTT (Sigma) at a concentration of 5 mg/ml were added to each well, followed by incubation of the microplate at 37°C for 3 hours. The MTT solution was discarded, 100 µl of dimethyl sulfoxide (DMSO) (Sigma) was then added, and the absorbance of the samples was measured at 570 nm using a microplate reader (BMG LabTech, Germany). The values of inhibitory concentrations (IC 50 ) (concentration achieving 50% cell viability) and concentration achieving 100% cell viability (EC 100 ) of OA, cLf, hLf, their complexes, and nanocomplexes were estimated with the Graphpad Instat software (Graphpad Software, San Diego, CA). This assay was repeated thrice, each in triplicate, and then the obtained results were expressed as mean ± standard error of the mean (SEM).
The anticancer potential of OA, cLf, hLf, cLf-OA, hLf-OA, and their nanocomplexes was evaluated against four human tumor cells; liver tumor cell line HepG-2 HB-8065, cervical tumor cell line HelaCCL-2, colon tumor cell line Caco-2 HTB-37, and breast tumor cell line MCF-7 HTB-22 were obtained from ATCC (Manassas, VA). Hela, HepG-2, and MCF-7 were maintained in RPMI-1640 supplemented medium (Lonza), while Caco-2 cell line was cultured in a DMEM supplemented medium. Both media were supplemented with 10% FBS. All tested cancer cells were seeded into 96-well cell culture microplates at a density of 5 × 10 3 cells/well, and then left to attach for 24 hours. The four cancer cell lines were treated with OA, cLf, hLf, their complexes, and nanocomplexes at serial concentrations and then incubated at 37°C for 72 hours in 5% CO 2 incubator. MTT method was performed and IC 50 values were calculated as described above.
Moreover, the morphological alterations in the test normal and cancer cells after 72 hours exposure to the tested samples were checked by a phase-contrast microscope (Olympus, Japan). These changes were compared to untreated cells.

Apoptotic investigation for treated cancer cells using nuclear staining analysis
cLf-OA and hLf-OA and their nanocomplexes were incubated for 72 hours with the cancer cell lines at their safe doses (EC 100 ). After that, cancer cells before and after treatment were stained with double nuclear stains of ethidium bromide and acridine orange (Sigma). Then, apoptosis was investigated in all tested cancer cells using the fluorescence phase-contrast microscope (Olympus) with an excitation filter (480/30 nm) and a dichromatic mirror, 505 nm cut-on.
Additionally, cell apoptosis was assessed by PI and annexin V staining (Sigma) and flow cytometry (PARTEC, Görlitz, Germany); cLf-OA, hLf-OA, and their nanocomplexes were incubated for 72 hours with the cancer cell lines at their EC 100 . After trypsinization, the cancer cells before and after treatment were stained for 15 minutes with annexin V and PI nuclear staining. Afterward, cells were fixed and incubated with 5 µg/ml of streptavidine-fluorescein (Sigma) for 15 minutes. Flow cytometry was used to estimate cell death via quantification of both PI-stained necrotic cells and annexin-stained apoptotic cells.
These assays were repeated thrice in triplicate, and all the obtained results are presented as mean ± SEM.

Determining the increment in the reactive oxygen species (ROS) production in cancer cells after treatment
For the detection of intracellular ROS level, 5 µM of 2′, 7′-dichlorodihydrofluorescein diacetate (H 2 DCFDA, Sigma) was incubated at 37°C for 30 minutes in the dark with cancer cells after and before treatment with either conventional or nanocomplexes. Then, cells were harvested with the Trypsin-ethylenediaminetetraacetic acid (EDTA) solution and suspended in fresh PBS. The percentage of oxidized fluorescence DCFDA was analyzed with a flow cytometer with excitation and emission settings at 488 nm and 530 nm, respectively. This test was repeated thrice in triplicate and the obtained results are presented as mean ± SEM.

Immunocytochemical nuclear staining protocol for Ki-67, Bcl-2, and p53 IN HepG-2
Following trypsinization, untreated in addition to treated HepG-2 were centrifuged and washed with PBS. Cell pellets were suspended in 10% formalin in PBS. The fixed cell specimens were dehydrated in ascending grades of alcohol. After immersing for 1 hour in xylene thrice, cells were impregnated in melted paraffin to produce solid paraffin blocks. Each paraffin block was cut into 3-5 μm thick sections using a rotator microtome; then, sections were transferred to positively charged slides. Slides were dried for 1-2 hours at 60°C-70°C then dewaxed by immersion in xylene thrice. Subsequently, slides were rehydrated in descending grading ethanol and incubated in 3% H 2 O 2 for 10 minutes. After washing twice in PBS for 3 minutes, the slides were placed in 10 mM citrate buffer, followed by heating for 10-20 minutes. After cooling and washing in PBS, the slides were separately soaked overnight in primary antibodies against Ki-67, Bcl-2, or p53. Slides were washed in PBS, covered with biotinylated goat anti-polyvalent secondary antibody for 10 minutes and then streptavidin-peroxidase was added for 10 minutes. The substrate 3, 3′-diaminobenzidine was added, followed by washing in PBS and placing for 1-4 minutes in hematoxylin bath then washing in PBS for 1 minutes and water for 3 minutes. The imaging analysis software CellSens of a phase-contrast microscope (Olympus) was used to determine the percentage of immunostained cells.

Lactoferrins activity evaluation
In the current study, the activity of both cLf and hLf was estimated by their ability to inhibit superoxide radical generation. Their mean activity was 6.79 ± 2.31 U/mg of protein for hLf and 9.84 ± 1.18 U/mg of protein for cLf.

Estimation of OA concentration in the prepared lactoferrin-OA complexes
OA concentrations in the prepared hLf-OA and cLf-OA complexes were presented as mean ± SEM; for cLf-OA it was 0.241 ± 0.02 mM and for hLf-OA it was 0.376 ± 0.017 mM.

Characterization of nanoparticles' formulations for cLf-OA and hLf-OA
The nanoparticles formulations for cLf-OA and hLf-OA were characterized with regard to loading capacity for cLf-or hLf-OA, encapsulation efficiency of nanoparticles, and the mean hydrodynamic size of the nanoparticles as presented in Table 1. Results presented in Table 1 reveal that the size of cLf-OA and hLf-OA complexes had decreased from 16.04 ± 0.85 and 18.15 ± 0.4 µm, respectively, to the nanoscale (197.5 ± 1 and 199.6 ± 1.4 nm, respectively) to achieve improvement in the release and safety of these complexes.
The IC 50 values of cLf and hLf before and after combination with OA and nanoformulation against human cancer cell lines demonstrated in Table 3 reveal that cLf-OA and hLf-OA complexes or nanocomplexes showed significantly (p < 0.05) higher anticancer activity than cLf and hLf before combination with OA. The antitumor activity of OA is shown in Table 3, where the IC 50 values of OA are 1.103 ± 0, 0.599 ± 0.11, 1.983 ± 0.02, and 0.445 ± 0.04 mM for HepG-2, Caco-2, Hela, and MCF-7, respectively. The actual concentration values of OA in the prepared complexes were 0.241 ± 0.02 mM for cLf-OA and 0.376 ± 0.017 mM for hLf-OA, which were significantly (p < 0.05) lower concentrations than the IC 50 values against test cancer cells, thus the improvement in antitumor effect of cLf-OA and hLf-OA complexes or their nanoformulations was not a result of an elevated OA concentration in the prepared lactoferrin-OA conventional or nanocomplexes, but it established the fact that lactoferrin proteins participate in this improved antitumor effect of conventional and nanocomplexes.
The problem was that IC 50 values of cLf-OA and hLf-OA complexes against Caco-2 and MCF-7 cells were significantly (p < 0.05) higher than their EC 100 against normal WI-38 cells, All values are expressed as mean ± SEM. Where A represents the total protein amount, B is the free protein amount in the supernatant, and C is the nanoformulation weight.   indicating that they could not be selective in killing tumor cells. On the contrary, all IC 50 values of cLf-OA and hLf-OA nanocomplexes against HepG-2, Caco-2, Hela, and MCF-7 cells were significantly (p < 0.05) lower than their EC 100 against normal WI-38 cells, indicating their selectively in killing tumor cells.
The morphological alterations in the test normal and cancer cells after 72 hours exposure to the tested samples under a phase-contrast microscope are shown in Figures 1 and 2.

Apoptotic investigation for treated cancer cells using nuclear staining analysis
The results of the acridine orange and ethidium bromide staining for detecting cancer cell apoptosis by a fluorescence phasecontrast microscope are shown in Figure 3. The nanoparticles formulations for cLf-OA and hLf-OA induced apoptosis in test tumor cells after 72 hours exposure to their EC 100 , in a more notably effective manner than cLf-OA and hLf-OA complexes.

DISCUSSION
OA has antitumor activity against various carcinoma cells because of its ability to disturb the structure of biological membranes and thus membrane-bound proteins' function (Amara, 2013(Amara, , 2015. Many studies confirmed in vitro anticancer effect of bLf on various tumor cells (Amiri et al., 2015;Zhang et al., 2015b).
Previous studies found that bLf could bind OA to kill cancer cells and that the resulting complex has a much higher anticancer activity than HAMLET (Fang et al., 2014;Zhang et al., 2015a). We previously focused on evaluating antibacterial (Almehdar et al., 2019(Almehdar et al., , 2020Redwan et al., 2016) and anti-hepatitis C virus (El-Baky et al., 2017;El-Fakharany et al., 2008;El-Fakharany et al., 2013;Liao et al., 2012;Redwan and Tabll, 2007) activities of cLf and confirmed that cLf was the most active lactoferrin among the different species studied. Moreover, the complex consisting of OA and camel α-lactalbumin was previously proven to be an effective and promising anticancer entity, and formulated OA-camel protein complexes were considered a potential approach for cancer therapy  (Uversky et al., 2017), decreasing the size of a drug to nanoscale results in improvement in the drug release, delivery, and safety (Alishahi and Elsabee, 2011;Khan et al., 2019). Nanoparticles formulations of lactoferrin were previously prepared and were proved to have enhanced therapeutic efficacy (Kanwar et al., 2012;Kumari and Kondapi, 2017;Tomitaka et al., 2015).
In this work, we formulated cLf-OA and hLf-OA complex-loaded chitosan nanoparticles that achieved elevated cLf/hLf-and OA-dependent cytotoxicity to different cancer cells while sparing normal healthy cells. It was observed that the IC 50 values of OA were 1.103 ± 0, 0.599 ± 0.11, 1.983 ± 0.02, and 0.445 ± 0.04 mM for HepG-2, Caco-2, Hela, and MCF-7, respectively. On the other hand, the IC 50 and EC 100 values of OA were 0.0307 ± 0 and 0.0109 ± 0 mM, respectively against WI-38 cells. Based on these data and the concentration values of OA in the prepared complexes measured as 0.241 ± 0.02 mM for cLf-OA and 0.376 ± 0.017 mM for hLf-OA, it could be concluded that OA caused cLf-OA complex (IC 50 of 88.18 ± 1.22 µg/ml) and hLf-OA complex (IC 50 of 91.52 ± 1.17 µg/ml) to have the highest inhibitory effect against normal cells. Moreover, IC 50 values of cLf-OA and hLf-OA complexes against Caco-2 and MCF-7 cells were significantly (p < 0.05) higher than their EC 100 against normal WI-38 cells, indicating that they could not be selective in killing tumor cells. The nanoformulations for the two complexes achieved enhanced anticancer efficacy against test cancer cells and improved safety to normal cells. The concentrations of OA in the prepared lactoferrin-OA complexes were found to be significantly (p < 0.05) lower than all of its IC 50 values against test cancer cells. These results are in agreement with those described by Fang et al. (2014), who demonstrated that the improvement in antitumor effect of cLf-OA and hLf-OA complexes or their nanoformulations was not a result of an elevated OA concentration in the prepared lactoferrin-OA conventional or nanocomplexes but established the fact that lactoferrin proteins also participate in this improved antitumor effect of conventional and nanocomplexes.
The results of the acridine orange, ethidium bromide, annexin V, and PI staining protocols and flow cytometry revealed that cLf-OA and hLf-OA nanocomplexes induced apoptosis in test tumor cells in a more notably effective manner than cLf-OA and hLf-OA conventional complexes. Additionally, a significant (p < 0.05) increase was found in ROS of test cancer cells after treatment with cLf-OA and hLf-OA and their nanocomplexes, but this increase was higher in case of nanocomplexes. Recently, a study reported that lactoferrin causes an increasing in the level of ROS of cancer cells (Zalutski et al., 2017) in addition to the increase in ROS production caused by OA (Hatanaka et al., 2013). Our results show that cLf-OA and hLf-OA and their nanocomplexes induced apoptotic cell death in all tested cancer cells via the intrinsic pathway by the generation of ROS.
Alterations in the levels of Ki-67, Bcl-2, and p53 in response to treatment of HepG-2 cells with cLf-OA and hLf-OA nanocomplexes were immunocytochemically analyzed using antibodies which recognized these proteins and found a significant decrease in levels of Ki-67 and Bcl-2, while the level of p53 was elevated. Thus, the cell apoptosis induced by cLf-OA and hLf-OA nanocomplexes involved the activation of the p53/Bcl2 pathway. These results present our nanoformulations as a promising therapy for cancer.
The nanoformulation of lactoferrin was previously confirmed to possess stable activities when compared with those of free protein at 4°C for 9 weeks (Abu-Serie and El-Fakharany, 2017).