Assessment of Cellular Uptake Efficiency According to Multiple Inhibitors of Fe3O4-Au Core-Shell Nanoparticles: Possibility to Control Specific Endocytosis in Colorectal Cancer Cells

Magnetite (Fe3O4)-gold (Au) core-shell nanoparticles (NPs) have unique magnetic and optical properties. When combined with biological moieties, these NPs can offer new strategies for biomedical applications, such as drug delivery and cancer targeting. Here, we present an effective method for the controllable cellular uptake of magnetic core-shell NP systems combined with biological moieties. Vimentin, which is the structural protein, has been biochemically confirmed to affect phagocytosis potently. In addition, vimentin affects exogenic materials internalization into cells even though under multiple inhibitions of biological moieties. In this study, we demonstrate the cellular internalization performance of Fe3O4-Au core-shell NPs with surface modification using a combination of biological moieties. The photofluorescence of vimentin-tagged NPs remained unaffected under multiple inhibition tests, indicating that the NPs were minimally influenced by nystatin, dynasore, cytochalasin D, and even the Muc1 antibody (Ab). Consequently, this result indicates that the Muc1 Ab can target specific molecules and can control specific endocytosis. Besides, we show the possibility of controlling specific endocytosis in colorectal cancer cells.


Introduction
Nanomaterials have opened new avenues for clinical diagnostics and therapeutics. Especially, nanoparticles (NPs) are one of the most important tools and have been used in applications such as biosensors [1,2], diagnostics [3,4], and targeted drug delivery systems [5,6]. For biomedical applications, NPs generally composed of organic materials surrounding surface on core materials [7][8][9].
The core materials, which is consisted of magnetic materials, semiconductor materials, or other types of materials, have useful physicochemical properties, and the outer organic surface provides chemical stability and functionality to the NPs. For applications in biological targeting systems, not only the physicochemical properties but also the outer organic surface, which are biofunctionalized for targeting, are critical parameters. Examples of targeting moieties for functionalization are antibodies or ligands that are specific for a target. Depending on the biofunctionalized materials on the outer surface, the endocytosis mechanisms of NPs are determined. The mechanism that allows NPs to enter cells has been the subject of many recent research works because of their importance in nanomedicine applications [10][11][12][13][14][15].
In particular, magnetic NPs have been widely used in many specific site targeting applications, including cell sorting [16,17], MRI [18], DNA isolation [19], drug delivery [20], hyperthermia treatment [21], and cancer targeting [22]. Among various magnetic NPs, magnetite nanocrystals have been most widely used in biomedical applications because of their biocompatibility and chemical stability. Although many efforts have been devoted to biomedical applications using magnetic NPs, there still are some critical issues such as good dispersibility in aqueous solution, functionality, and biocompatibility. To overcome these problems, many studies have focused on the surface modification of NPs using a variety of functional groups (e.g., carboxyl and amine groups) [23]. However, the attachment of the functional groups to the surface of magnetite NPs is a time-consuming and laborious process. Given this fact, core-shell-type Au-coated magnetic NPs are attractive because the Au surface can link easily to biomolecules and organic materials.
Especially, the magnetic properties of the magnetic core-Au shell NPs enable the magnetic separation, increase resolution in MRI imaging, and can be applied to hyperthermia therapy. Moreover, the superior chemical binding properties of gold are advantageous for building receptor-mediated delivery systems for specific cancertargeting [24][25][26].
Over the past few decades, many researchers have reported receptor-mediated delivery systems for cancer targeting [27][28][29].
Receptor-mediated targeting of cancer cells is a form of active targeting. The choice of target is the key for effective active targeting, and the targets must be overexpressed on the extracellular membrane. Most researchers have used monoclonal antibodies for cancer treatment, and the therapeutic effect could be greatly increased when monoclonal antibody therapies are combined with conventional chemotherapy [30]. Despite the success of monoclonal antibody therapy, monoclonal antibodies present several limitations in cancer targeting. Their large size (approximately 150 kDa) is a major obstacle for tumor penetration [31,32], and their low stability and low solubility hinder their widespread use [33]. The inhomogeneous direction of their attachment on the targeting carrier is also considered an obstacle to nonspecific binding. To produce antibodies with improved tumor penetration, a wide range of antibody formats have been engineered and tested [34]. Apart from classical antibodies, a unique antibody format is present in species from the family Camelidae. The so-called heavy-chain antibodies (HCAbs) occur naturally in the peripheral blood and milk of these species. The antigen-binding fragments of such HCAbs are composed of one single domain, the heavy-chain variable domain (VH) of the camelid HCAb (VHH). The VHH, recombinantly obtained after cloning and expression in bacteria or fungi, is called a nanobody. It has a molecular weight of 11-15 kDa and is the smallest antibody among all mAbs [35][36][37]. Not only their small size makes them potentially suitable as targeting probes against antigens in isolated locations, but also their easily modifiable terminal end is attractive for application in cancer targeting.
The efficient delivery of NPs with suitable targeting and internalization of cells are also important factors in the delivery system. It has been reported that vimentin acts an important role as a component of pathogen attachment and intracellular entry pathways. Silencing of vimentin gene expression inhibits phagocytosis [38], whereas cleaved vimentin is a signal that significantly increases phagocytosis [39]. Therefore, neutralizing cell phagocytosis resistance caused by vimentin on the cell surface is important for efficient nanoparticle delivery.
In this study, we investigate the endocytosis pathways of nanobody-tagged Fe 3 O 4 -Au core-shell NPs modified with PEG (polyethylene glycol) spacers with different lengths. Vimentin, which is known to have a strong effect on phagocytosis by biochemical experiments [39], was compared as a control, and it was confirmed that it effectively acts on cell internalization of NPs. Besides, the Muc1, which is a cell surface glycoprotein and overexpressed in various cancer, such as pancreatic, breast, lung, and stomach cancer, is utilized as a cancertargeting biomarker. We confirmed the efficient internalization of Fe 3 O 4 Au core-shell NPs and the methods of controllable targeting to cancer cells through the Muc1 receptor-mediated endocytosis pathway in colonic cells.

Synthesis of Fe 3 O 4 -Au Core-Shell NPs
The Fe 3 O 4 -Au core-shell NPs were synthesized via a nanoemulsion method. The synthetic process for coreshell NPs consists of two steps: (1) formation of the Fe 3 O 4 core NPs and (2) coating of the Au shell on the magnetic NPs. In the first step, the Fe 3 O 4 NPs were prepared from a mixed solution of Fe(acac) 3 (0.1766 g or 0.5 mmol), 1,2-hexadecandiol (0.6468 g or 2.5 mmol), and block copolymer (poly(ethylene oxide)-poly(propylene oxide)-poly(ethylene oxide); PEO-PPO-PEO) (0.41 .2 g) in octyl ether. The mixed solution was heated at 300°C to reduce the Fe precursor. The formation of Fe 3 O 4 core NPs was completed by cooling the heated solution. The second process was continuously conducted without any purification process after the formation of the magnetic core. Au precursors (0.2338 g or 0.62 mmol) and 1,2-hexadecandiol (0.88 g, 3.4 mmol) were added into the emulsion consisting of Fe 3 O 4 NPs, and then the mixed solution was heated at 230°C. After cooling down to room temperature, the emulsion was precipitated by centrifugation and the core-shell NPs were separated.

Modification of Core-Shell Fe 3 O 4 -Au NPs
OPSS-PEG-NHS at various lengths (2, 5, and 10 K) was dissolved in 0.1 M sodium bicarbonate for activation of the thiol groups. Activated OPSS-PEG-NHS was added to the solution of synthesized core-shell Fe 3 O 4 -Au NPs and agitated for 12 h at 4°C. The thiol groups of activated OPSS-PEG-NHS were covalently linked to the Au surface of the core-shell NPs. Then, a nanobody solution (0.25 mg/mL) was added to the PEGylated Fe 3 O 4 -Au core-shell NPs for 12 h at 4°C. The amine groups of the ten lysine (K) tails at the terminal were covalently linked to the NHS groups of OPSS-PEG-NHS at pH 8.3. Cy3 and Cy7.5 were tagged to the residual amine groups of the nanobody.

Internalization Curve
CT26 cells were seeded at 5 × 10 3 cells per well in a clear-bottom 96-well plate and incubated in 250 μL of culture medium for 24 h at 37°C in 5% CO 2 in the dark. The medium was removed, and 250 μL of fresh culture medium containing 50 μg/mL Cy3-labeled Fe 3 O 4 -Au NPs and PEG-Cy3 or PEG-nanobody-Cy3-labeled NPs were added to each well. The cells were further incubated for different periods (0, 10, 20, 30, 60, 120, and 360 min). The cells were then washed three times with PBS to remove free NPs, and the fluorescence of each well was measured with trypan blue as a membraneimpermeable fluorescence quencher by SpectraMAX GEMINI (Molecular Devices, CA, USA). Each experiment was carried out with equal amounts of NPs (50 μg/ mL) and repeated three times [40].
After 1 h at 37°C and 5% CO 2 , the cells were washed three times with PBS to remove free NPs, and the fluorescence was measured with trypan blue as a membrane-impermeable fluorescence quencher by Spec-traMAX GEMINI (Molecular Devices, CA, USA). Each experiment was carried out with equal amounts of NPs (50 μg/mL) and repeated four times.

Results and Discussion
The core-shell NPs were synthesized by a published method [16,17]. Transmission electron microscopy (TEM) observations in Fig. 1a, b show that the Fe 3 O 4 -Au core-shell NPs were spherical with an average diameter of 13.5 nm and narrow size distribution.
The increase from the~8.5 nm of the core NP (Fe 3 O 4 ) stems from the coating of~2.5-nm-thick Au shell on the core surface, resulting in a core-shell NP. A high-resolution TEM image with the fast Fourier transform (FFT) analysis of the Fe 3 O 4 -Au core-shell NP is included in Supplementary Information Fig. S1.
The product manufactured in organic solvent was purified using magnetic separation and transferred into the water.
The core-shell NPs were well dispersed and stable in water without any surface modification, owing to residual block co-polymers that were present on the NPs. Figure 1c and d show the NPs in aqueous solution before and after applying an external magnetic field. Under an external magnetic field, the core-shell NPs rapidly changed from a homogeneous dispersion (Fig. 1c) to a clear and transparent solution (Fig. 1d).
The absorption band of the core-shell NPs was investigated using UV-Vis spectrometry. As shown in Fig. 1e, an absorbance peak appeared at~530 nm, indicating the presence of Au on the surface of the NPs (Supplementary Information Fig. S2 includes the result of EDX data for Fe 3 O 4 -Au core-shell NPs). As the sample had been purified, the optical results demonstrated the formation of the core-shell structure.
Magnetic hysteresis loops were obtained from vibrating sample measurements to investigate the magnetic properties of the Fe 3 O 4 core and the core-shell NPs. Both NPs showed superparamagnetic behavior with a coercivity of near 0 Oe at room temperature (Fig. 1f).
As reported in previous works, the susceptibility of core-shell NPs was higher than that of the magnetite NPs, which could be partly owing to proximity effects and unique spatial configurations [41,42]. Besides, the saturation magnetizations of the core NPs and the coreshell NPs are~37 emu/g and~21 emu/g at 10 kOe, respectively. The difference in the Ms stems from the existence of a nonmagnetic component (Au) in the coreshell NPs.
The VHH 5-24 K10 gene was cloned in-frame to produce pET-23a-anti-Muc1-VHH 5-24 K10 after PCR amplification (Fig. 2a). The recombinant protein was expressed in E. coli BL21 that was transformed with pET-23a-anti-Muc1-VHH 5-24 K10 after induction with IPTG and purified by Ni-NTA His·Bind Resin. Recombinant anti-Muc1-VHH 5-24 K10 was readily expressed in E. coli as a soluble 18-kDa protein. From a 1-L Fig. 1 Characterization of the synthesized Fe 3 O 4 -Au core-shell NPs. a, b TEM observations of the synthesized Fe 3 O 4 -Au core-shell NPs. c, d NPs in aqueous solution, before and after applying an external magnetic field. e UV absorbance peak of synthesized core-shell NPs appears at~530 nm. f Magnetic hysteresis loops of Fe 3 O 4 core culture, we obtained 1 ± 0.5 mg of purified recombinant anti-Muc1-VHH 5-24 K10.
The purified protein was verified by 15% SDS-PAGE gel. Coomassie blue staining of the purified protein revealed that it was > 95% pure (Fig. 2b). The synthesized Fe 3 O 4 -Au NPs were modified in three steps, namely PEGylation, antibody tagging, and dye labeling (Fig. 2c). After each modification step, the zeta potential was measured to confirm the successful modification. Table 1 shows the effect of modification on the corresponding zeta potentials. The zeta potential of bare core-shell NPs was − 19.8 ± 6.68 mV. PEGylation of NPs was carried out using OPSS-PEG-NHS. To produce a series of nanocomplexes with different sizes, OPSS-PEG-NHS with various lengths (2 K, 5 K, and 10 K) was used.
The zeta potential of bare Fe 3 O 4 -Au core-shell NPs was − 19.8 ± 6.68 mV. After PEGylation, the zeta potential was reduced to near − 40 mV. These results indicate that the PEG molecules were well bonded covalently to the Au shell of the core-shell NPs because PEG molecules have negatively charged N-hydroxysuccinimide functional groups. Meanwhile, the zeta potential increased after dye tagging to the nanobody (− 38.5 ± 5.61 mV for NP-PEG2 K-nanobody and − 12.5 ± 7.25 mV for NP-PEG2 K-nanobody-dye). This outcome is reasonable because the recombinant nanobody has ten lysine tails at the terminal end. Each type of nanobody was categorized by zeta potential measurement, and to determine antibody binding on the nanobody, we measured the fluorescence for each nanobody type.
As shown in Fig. 3, we confirmed that all types of nanoparticles and nanobodies have well cellular uptake and internalization in the absence of inhibitor restrictions. A cellular internalization curve was obtained from cells incubated in the presence of 50 μg/mL Cy3-labeled Fe 3 O 4 -Au NPs, PEG-Cy3-labeled NPs, and PEGnanobody-Cy3-labeled NPs for different periods (between 0 and 360 min) after removing media, washing out free NPs, and finally measuring the total fluorescence of the cells with trypan blue (Fig. 3).
According to the result of fluorescence intensity measurement, we can determine the NPs were internalized in the cells within 1 h (Fig. 3a). The fluorescence intensity of NP reached a maximum within 1 h, and the fluorescence intensity gradually decreased after reaching a steady state. Even though there is a slight time difference depending on the presence of Ab, the fluorescence intensity per cultivation time was not significantly different  from the result of bare NPs (Fig. 3b). As shown in Fig. 3c, it was confirmed that the effect of Ab, by using heterogeneous Ab of the Muc1 and vimentin, was negligible in the cellular uptake and internalization of NP.
With WST-1 assay, the viability (%) of CT26 mucin cells depending on the concentration and surface modification of Fe 3 O 4 -Au core-shell NPs was estimated following various exposure times (Fig. 4a, b). The viability of CT26 cells did not show any significant differences following 24 h and 48 h of exposure, either with varying doses or after surface modification of the NPs. Cell viability was greater than 90% on both the bare Fe 3 O 4 -Au NPs (Fig. 4a) and the surface-modified NPs (Fig. 4b). Figure 5 indicates that NPs entered the CT26 mucin cells via various endocytosis pathways (clathrin-mediated, caveolae-mediated, and phagocytosis/macropinocytosis pathways). Interestingly, Fig. 5b shows that anti-Muc1 Ab also mainly affected the endocytosis of PEGnanobody-Cy3-labeled NPs. To understand the pathway of NP internalization, we tried to inhibit endocytosis pathways with specific chemical inhibitors (Fig. 5). The pathways of endocytosis were well known to be divided into three types: clathrin-mediated, caveolae-mediated, and macropinocytosis/phagocytosis.
The targeting moiety is the key to the success of cancer-targeting, which is exceptionally important in cancer therapy. For targeting, effective surface modification is very important to increase therapeutic efficiency and limit side effects. Nanobody-tagged Fe 3 O 4 -Au coreshell NPs were successfully made from the synthesized NPs and recombinant nanobody. Table 1 clearly shows that each modification step was carried out successfully.
Cell viability is one of the essential elements for the biological application of nanomaterials. The cell viabilities were greater than 90% on the bare core-shell NPs and modified NPs (Fig. 4a, b). These results imply that bare Fe 3 O 4 -Au NPs and modified NPs did not cause significant concentration-and modification-dependent cytotoxicity and that the modified NPs were suitable for biological application.
The studies of internalization efficiency and the inhibitor effect of NPs provide important information to understand the mechanisms through which NPs enter cells. PEGylated NPs were relatively slow internalized into cells compared with bare NPs, but nanobody-tagged NPs were internalized into cells slightly faster than bare NPs (Fig. 3a, b). Because PEGylation is a well-known surface modification Fig. 3 Normalized photofluorescence of CT26 mucin cells after incubation with 50 μg/mL a Fe 3 O 4 -Au NPs, b PEGylated NPs and PEG-nanobody-tagged NPs, and c Fe 3 O 4 -Au NPs, nanobody-tagged NPs, and vimentin-tagged NPs at 37°C in 5% CO 2 for different periods (10, 20, 30, 60, 120, and 360 min) method to prevent the internalization of NPs, the internalization tendency of PEGylated NPs is easily explained. Moreover, the nanobody induced the endocytosis of NPs. To check the specificity of the nanobody, we confirmed the internalization rate of vimentin Ab-tagged NPs. Interestingly, vimentin Abtagged NPs did not promote cellular internalization (Fig. 3c).
These results indicate that the nanobody can effectively induce the internalization of nanomaterials into CT26 mucin cells and imply that the internalization tendency can be controlled with specific modification of the outer membrane of NPs. Additionally, the mechanisms of the endocytosis of nanobody-tagged NPs were clearly shown via inhibition tests and confocal microscopy imaging. The photofluorescence of both the nanobody-tagged NPs and the non-tagged NPs showed similar decreasing values when cultured with CPZ, nystatin, or dynasore (Fig. 5a, b). CPZ, nystatin, and dynasore play a role in inhibiting, respectively, clathrin-mediated endocytosis, caveolaemediated endocytosis, and dynamin, which is a large GTPase implicated in the budding and scission of nascent vesicles from parent membranes. Thereby, the photofluorescence values in both cases decreased rapidly because dynamin is closely related to the production of vesicles for clathrin-mediated and caveolaemediated endocytosis. As shown in Fig. 5a, both nontagged (Fig. 5a) and nanobody-tagged NPs (Fig. 5b) showed rapid decreases in photofluorescence.
In particular, the nanobody-tagged NPs displayed significantly lower photofluorescence values in CPZ, nystatin, and dynasore. Moreover, we confirmed that the non-tagged NPs were more strongly affected than nanobody-tagged NPs were when applied to cytochalasin D, which is a cell-permeable toxin that blocks polymerization of actin filaments for phagocytosis [43]. These results imply that non-tagged NPs were internalized through multiple mechanisms such as clathrinmediated endocytosis, caveolae-mediated endocytosis,  Normalized photofluorescence from CT26 mucin cells treated with chemical endocytosis inhibitors for 1 h and incubated with 50 μg/mL a bare core-shell NPs and b nanobody-tagged NPs at 37°C in 5% CO 2 for 30 min. Inhibitors with a statistical effect on the internalization (Student's t test, p (*) < 0.05 and p (**) < 0.01) are marked with black asterisks and phagocytosis. Consequently, the cellular internalization of nanobody-tagged NPs depends on clathrinmediated and caveolae-mediated endocytosis. Besides, the amount of cellular uptake of nanobody-tagged NPs was reduced considerably when cells were cultured with the Muc1 antibody (Fig. 5b). This result indicates that the free Muc1 antibody plays a role as a competitor of the nanobody on the modified NPs in attaching to the CT26 cell membrane and that the Muc1 antibody plays an important part in the cell internalization of the modified NPs. Peculiarly, vimentin Ab-tagged NPs showed obvious differences compared with nanobody-tagged NPs in terms of inhibition ability. The photofluorescence of vimentin-tagged NPs remained unaffected under multiple inhibition tests, indicating that the NPs were minimally influenced by nystatin, dynasore, cytochalasin D, and even the Muc1 Ab. This phenomenon could be evidence of the efficacy of vimentin, which has been biochemically confirmed to potently affect phagocytosis [39]. Consequently, this result indicates that the Muc1 Ab can target specific molecules and can control specific endocytosis.
As shown in Fig. 6, analogous results were obtained when the cells were treated with Cy7.5-labeled bare coreshell NPs and PEG-nanobody-tagged NPs, indicating similar cellular uptake in both cases in the absence of dynasore inhibition. Dynasore inhibition induced distinctly lower cellular internalization of the PEG-nanobody-tagged NPs compared with bare NPs (Fig. 6b, bottom row). These results imply that there are two endocytosis mechanisms, which are non-specific endocytosis of bare NPs and restricted endocytosis of nanobody-tagged NPs via the dynamin molecule. Once the nanobody attaches to the external cell membrane, the nanobody-tagged NPs could easily pass through the cell membrane because of the simultaneous activation of the clathrin-and caveolaemediated mechanisms. Consequently, we could suppose that the main mechanisms are both clathrin-and caveolae-mediated endocytosis for the internalization of nanobody-tagged NPs in CT26 mucin cells.

Conclusions
Nanomaterials for cancer targeting and controllable insertion of exogenic materials such as drugs, genes, and peptides are critical advances in biomedical applications. These familiar but creative concepts can offer strategies for new therapeutic methods. In this paper, we demonstrated enhanced cellular uptake of Fe 3 O 4 -Au core-shell NPs after PEGylation with the Muc1 antibody. The main endocytosis mechanisms of nanobody-tagged NPs were demonstrated, showing the possibility of controllable specific endocytosis in colorectal cancer cells. These findings provide insight into the targeting between nanobody-tagged NPs and colorectal cancer cells to aid the design of high-efficiency targeting carriers.