Synthesis, decoration, and cellular effects of magnetic mesoporous silica nanoparticles

Abstract

Mesoporous Silica Nanoparticles (MSN) are now considered as multifunctional platforms for pharmaceutical development. The goal of this study was to optimize a synthesis procedure to obtain reproducible monodisperse magnetic core@shell Fe₃O₄@MSN with different coatings and study their uptake by cells. 100 nm core@shell nanoparticles with a unique 18 nm magnetic core were synthesized and covered with PEG groups or coated with a lipid bilayer in a controlled manner and their cellular fate was investigated. Both PEG and lipidic coated nanoparticles exhibit a low toxicity when incubated with Hep-G2 cells compared to pristine ones. Furthermore, the different real-time impedance cellular profiles that were observed and the particles uptake by the cells investigated by TEM suggest different internalization mechanisms or uptake kinetics depending on MSN coverage. This study is a first essential step to ensuring the preparation of well-defined nanomaterials for medical applications; it is considered as a crucial step to be able to perform detailed research about cellular trafficking and signaling pathways.

---

1 Introduction

The emerging field of nanomedicine has been extensively developed over the past two decades taking advantage of the dual benefits of materials at the nanoscale. Firstly, due to the size reduction, new physicochemical properties such as optical, magnetic or thermal properties appear. Second, nanodrugs are able to overcome certain biological barriers and would therefore have a specific biological fate in terms of biodistribution and biokinetics that can be exploited for biomedical purposes. A major field of application of nanomedicines is that of cancer for diagnosis, imaging, drug delivery and cell targeting and many research studies are directed towards this broad area of application.¹ As part of this studies, it was shown that the biological behaviour of nanoparticles (NPs), i.e. the biodistribution, cellular uptake, or cytotoxicity is very dependent on their physicochemical properties.^2–4 Among the most important physicochemical parameters were identified the particle size, shape and surface chemistry.² As a general rule, in terms of size, it has been shown that the smaller nanoparticles are more toxic than their larger counterparts, any other parameter (chemical composition, coating, etc.) equivalent otherwise.^5,6 Also, more specifically to the core@shell nanoparticles, the fact that in a given sample be present NPs without a core (underestimation) or with multiple cores (overestimation) false the real dose of used NPs. It is therefore necessary to establish standards for the use of nanoparticles and indeed the recent development of nanomedicine is now associated to official recommendations depending on countries. In France, the required characterizations are listed by the norm ISO/TR 13014. This norm argues the careful measured knowledge of NPs' physicochemical properties for comprehension of their biological fate. The European Medicinal Agency published guidelines for the evaluation of the safety of nanomaterials where it is notified that the physicochemical and pharmacokinetic comparison of generic NPs to a reference is not sufficient to approve the safety of those generic NPs (EMA/CHMP/SWP/100094/2011).⁷

The FDA created a section dedicated to the evaluation of nanotechnology since the first Nanotechnology Task Force Report published in 2007.⁸ This section elaborated guidance for industry considering whether an FDA-regulated product involves the application of nanotechnology in 2014.⁸ Each of these official recommendations refers to the major influence of the fundamental parameters that are the physicochemical properties, in particular size, chemical composition and surface properties on the NPs biological behaviour. It is therefore essential to have nano-objects that meet these standards in prior to any detailed research about the cellular trafficking and signaling pathways.

Among nanomedicines, the mesoporous silica nanoparticles (MSN) have been considered as a multifunctional platform for pharmaceutical development⁹ and found particularly advantageous for theranostic applications^10,11 MSN surface exhibits silanols groups, which could be functionalized by covalent ligands or by electrostatic coupling paving the way for an easy modification of their surface properties which could be one way to control the biodistribution of particles.¹² Concerning cancer applications, accumulation of the NPs (size range 30–200 nm) in the tumour has been explored due to the enhanced permeation and retention (EPR) effect.¹² The rapid neovascularisation (angiogenesis) occurring in the solid tumour presents imperfections, leading to fenestration; and the association of the blood pressure effect with fenestration is favourable for NPs accumulation into the tumour. Nevertheless, NPs could also accumulate into spleen and liver, due to important macrophages uptake of NPs by phagocytosis. To avoid the latter, surface functionalization with polyethylene glycol (PEG) moieties is now a well-known strategy for the production of furtive MSN.^13,14 An alternative strategy modifying the fate of NPs consists in wrapping the nano-objects by a lipid bilayer, the latter including PEG grafted molecules and/or targeting probes.¹⁴ Among numerous properties, MSN pores could be loaded with cytotoxic molecules.^15,16 Otherwise, MSN having a magnetic core allows following nanoparticles (NPs) using magnetic resonance imaging (MRI),¹⁷ or for drug release induced by the magnetic field.¹⁵ Indeed, drug release could be enhanced by heating through magnetic field application to a temperature of 50 °C. Plus, after MSN cell internalization, heating could enhance the intra-cellular temperature from 37 °C to 50 °C, leading to hyperthermia.¹⁸ Cell death generally occurred for a temperature upper to 42 °C, depending of cell type as cancer cells seemed more sensitive to heat stress than normal cells.

Sol–gel chemistry was developed since 2008 as a popular soft chemistry route to core@shell nanoparticles synthesis with well-ordered mesoporosity and tuneable pores.¹⁹ Since this date, numerous publications reported this synthesis; however few reached the initial required quality for nanomedecine. Most synthesis resulted in non-homogeneous shaped NPs^15,20–23 or provide NPs presenting several magnetic cores per NPs.^24–26 The goal of this work was to synthesize Fe₃O₄@MSN core–shell nanoparticles, monodisperse, homogeneous, and containing one magnetic core per particle. Even if MSN are generally considered as biocompatible, as previously mentioned their surface properties are often modified for biomedical applications²⁷ in order to enhance their colloidal stability in biologically relevant conditions¹² and then to improve their bio-stability and biocompatibility (reduced protein adsorption and resistance to nonspecific uptake).¹³ Plus, comparison of particles with sizes comprised between 150 and 500 nm with zeta potentials between −40 mV and +35 mV showed that NPs with slight negative charge (zeta potential of −15 mV) and 150 nm accumulates more efficiently in tumour than the other ones.²⁸ In the present work, a MSN size of about 100 nm was chosen for optimal penetration in tumors,²⁸ and the surface properties of NPs were modified by grafting PEG or coating with lipid bilayers. The cellular toxicity of the three types of particles, pristine, PEGylated or lipid coated, was measured on Human hepatocyte carcinoma cell line (Hep-G2 cells). Using the same cell line, real-time cellular effect of pristine, PEGylated or lipid coated NPs were compared by cellular impedance measurements. The cellular fate of NPs was analysed in relation with the nature of NPs coverage; the uptake of NPs was monitored using transmission electronic microscopy.

2 Results and discussion

2.1 Synthesis and characterization of pristine Fe₃O₄@MSN, PEG-Fe₃O₄@MSN and phospholipid coated Fe₃O₄@MSN

2.1.1 Pristine Fe₃O₄@MSN. In this work, Fe₃O₄@MSN synthesis process was optimized to obtain monodisperse, 100 nm sized, mesoporous nanoparticles with one magnetic core per MSN and key parameters were evidenced to obtain reproducible results. Purity, concentration and hydrophobicity of MIONs (magnetic iron oxide nanocrystals) are crucial parameters for obtaining monodisperse magnetic silica core–shell nanoparticles and preventing the formation of core-free silica particles. MIONs with a size of 18 ± 3 nm were prepared according to the thermal decomposition of iron oxide (II).²⁹ In this procedure, n-docosane was used as the solvent (boiling point 368 °C) and oleic acid was used as both reducing agent and nanocrystals stabilizer. The size of MIONs is related to the amount of oleic acid used in the synthesis. The more amount of oleic acid, the larger are the nanoparticles.³⁰ The synthesized and washed MIONs were better dispersed in non-polar solvents such as chloroform and were stabilized by addition of oleylamine. The resulting hydrophobic properties conferred to the MIONs ensure an optimal distribution of the magnetic nanoparticles into the micellar phase through the self-assembling between oleylamine and cetyltrimethylammonium bromide (CTAB). The concentration of MIONs in chloroform was estimated at 9.8 g L⁻¹ (determined by thermogravimetric analysis (TGA) measurements, see ESI†).

It is important for MIONs to be well dispersed in chloroform in a way that each single nanocrystal compound ideally occupies one micelle of CTAB so that to obtain single core Fe₃O₄@MSN. Thus, in order to reduce the formation of multi-core MSN, MIONs were dispersed in chloroform during 3 hours using ultrasound bath and nanocrystals were injected by 10 steps into the CTAB micellar reaction medium, to enhance their distribution in micelles. However, the presence of chloroform in the core–shell reaction mixture interferes with the sol–gel reaction and does not promote the formation of uniform and regular mesoporous silica particles. Therefore, it is essential to eliminate the chloroform from the reaction mixture by applying a brief vacuum at 70 °C after each injection of the MIONs suspension in the reaction mixture. Before silica precursor injection, the reaction medium temperature was stabilized at 80 °C during one hour. After reaction completion, differential centrifugation steps were performed. Firstly, 5200g centrifugation was accomplished to pellet multicore MSNs. After that, 35 700g centrifugation step was performed. MSN without magnetic core were recovered in the supernatant. Finally, Fe₃O₄@MSN were extensively washed in an alcoholic solution of ammonium nitrate by reflux during 12 hours. The as-prepared Fe₃O₄@MSN core–shell sample is a black powder. TEM images showed that the particles are monodisperse and regular silica spheres with radial mesostructures encapsulating 18 nm spherical Fe₃O₄ nanocrystals are observed (Fig. 1). It should be noted that no core-free silica particles were observed. The core–shell NPs have an overall diameter of 100 ± 5 nm. The size and monodispersity of native Fe₃O₄@MSN were confirmed by dynamic light scattering (DLS). Hydrodynamic diameter (HD) and zeta potential (ZP) were measured in low ionic strength buffered medium (20 mM Hepes, 5 mM NaCl, pH = 7.4) to prevent NPs aggregation. The HD was measured at 125.47 ± 1.61 nm (see Table 2). The surface properties have been studied using a zeta potential (ZP) analyzer. The measurements revealed that the sample exhibits negative ZP values around −32.09 ± 1.30 mV (Table 2). This relatively high negative value confers to the nanoparticles a colloidal stability in aqueous solution and prevents them from aggregation in agreement with previous studies which argued that above the isoelectric point (pH > 2), the surface of silica is negatively charged due to the presence of deprotonated silanol groups (SiO⁻).^25,31

Fig. 1 TEM images of (a) magnetic iron oxide nanocrystals and (b) monodisperse magnetic mesoporous silica core–shell particles (Fe₃O₄@MSN) prepared according to the optimized synthesis protocol.

Textural properties of native Fe₃O₄@MSN have been characterized by N₂ sorption and small-angles X-ray diffraction (sa-XRD). The textural parameters are given in Table 1. The material exhibits a type IV isotherm according to the IUPAC classification. The isotherm profile is characteristic of MCM-41 type materials with a surface area of 675 ± 7 m² g⁻¹ and uniform mesopores of 2.84 ± 0.2 nm of diameter, according to the pore size distribution. The hysteresis observed at P/P₀ > 0.85 is likely related to inter-particles porosity and packing defects.³² The small-angle XRD pattern revealed that only the first-order scattering peak (d₁₀₀) appears well resolved at 2θ = 1.69°. This may be due to the spherical morphology of the core–shell nanoparticles implying a constraint on the pore organisation. The wide-angle XRD pattern (ESI Fig. 3†) evidenced the presence of Fe₃O₄ nanocrystals with the related indexed diffraction peaks planes (311, 400 and 440) embedded in the characteristic broad signal of amorphous SiO₂.^29,33 The observed diffraction lines confirmed the magnetite crystals structure of magnetic iron oxide nanocrystals (Fig. 2).

Table 1 Textural parameters of native Fe₃O₄–MSN and PEG Fe₃O₄–MSN nanoparticles

	SBET (m^2 g^−1)	Sext^a (m^2 g^−1)	Vp^b (cm^3 g^−1)	DBJH^c (nm)
a Sext: external surface as inferred from the t-plot analysis applied to N2 adsorption isotherms.b Vp: pore volume calculated at the saturation point of the N2 sorption isotherms at the relative pressure P/P0 = 0.85.c DBJH: mesopore volume, as determined from the BJH method (desorption branch).
Fe3O4@MSN	675 ± 7	58 ± 1	0.97 ± 0.02	2.84 ± 0.20
PEG Fe3O4@MSN	475.0 ± 3.6	55 ± 1	0.64 ± 0.02	2.75 ± 0.20

---

Fig. 2 Textural and structural properties of pristine Fe₃O₄@MSN core–shell particles: (a) N₂ sorption isotherms with ● adsorption ○ desorption (inset is the pore size distribution), (b) small-angle X-ray diffraction pattern with the fitted peak.

2.1.2 PEG Fe₃O₄@MSN. The grafting of polyethylene glycol at the surface of the NPs, PEG Fe₃O₄@MSN, was achieved in situ.^13,34 PEG molecules were covalently bound to the silanol groups at the end of the Fe₃O₄@MSN synthesis. After the condensation process induced by the injection of TEOS, the mixture was slowly cooled to 50 °C with continuous stirring. Then, 1 mL ethanol containing 100 mg of silanized PEG 2000 was slowly added and the reaction mixture was stirred overnight. After completion of the reaction, the mixture was cooled to room temperature before applying the same washing steps as those described above for the pristine Fe₃O₄@MSN.

The PEG Fe₃O₄@MSN core–shell sample is a black powder. TEM images (ESI Fig. 7†) showed that the particles are monodisperse and regular silica spheres with radial mesostructures were observed. There is no difference on TEM images between pristine Fe₃O₄@MSN and PEG Fe₃O₄@MSN. The PEGylation induced an HD increased to 141.76 ± 13.96 nm due to PEG groups and an increase of ZP from −32.03 to −26.30 mV (see Table 2). The presence of the PEG moieties at the particle surface was evidenced using Raman spectroscopy (Fig. 3c). Here, the Raman active vibrational modes of the PEG grafted at the MSN particle surface are highly decreased in intensity, twisting vibrations of methylene group of PEG at 1221–1366 cm⁻¹ and CH₂–CH₂ bending vibration at 1458–1479 cm⁻¹. This could be due to hydrogen bonding/van der Waals' interaction with the silanol groups of the silica surface.³⁵

Table 2 Physicochemical characterization of pristine, PEG-grafted and DMPC coated Fe₃O₄@MSN in low ionic strength medium (Hepes 20 mM, NaCl 5 mM, pH 7.4)

NPs	Hydrodynamic diameter (nm)	Polydispersity index	Zeta potential (mV)
Fe3O4@MSN	125.47 ± 1.61	0.14 ± 0.04	−32.09 ± 1.30
PEG Fe3O4@MSN	141.76 ± 13.96	0.14 ± 0.05	−26.30 ± 1.35
DMPC Fe3O4@MSN	179.40 ± 2.40	0.20 ± 0.02	−10.80 ± 0.25

---

Fig. 3 Characterization of Fe₃O₄@MSN PEG-grafting. TGA/DSC spectra of: (a) pristine Fe₃O₄@MSN and (b) PEG–Fe₃O₄@MSN. Quantification of PEG grafted on Fe₃O₄@MSN was calculated from TGA analysis, by comparison of the weight lost from PEG–Fe₃O₄@MSN to bare Fe₃O₄@MSN. (c) Raman spectra is the proof that Fe₃O₄@MSN are associated to PEG groups with characteristic bands at 1221–1366 cm⁻¹ and 1458–1479 cm⁻¹.

TGA analysis of pristine particles (Fig. 3a) exhibits a gradual change of mass loss of 4.4% in the temperature range 200–800 °C ascribed to a progressive dehydroxylation of the silica shell. It is known that a slow condensation of germinals, vicinals, and isolated silanols occurred upon heating of amorphous silica.³⁶ The grafted amount of PEG onto Fe₃O₄@MSN was determined based on TGA analysis (Fig. 3b). The thermal decomposition of PEG is observed to start at 251 °C. The measured mass loss of 12% arisen from a grafted PEG amount of 2.36 mmol g⁻¹ of particles taking into account the dehydroxylation of silica.

2.1.3 Phospholipid coated Fe₃O₄@MSN. The coating of Fe₃O₄@MSN with phospholipid was achieved after the synthesis and extensive washing of pristine Fe₃O₄@MSN. The bilayer wrapping of Fe₃O₄@MSN was achieved by addition of a suspension of DMPC small unilamellar vesicles (SUV) to the Fe₃O₄@MSN mixture. Then, SUV rupture in contact with the MSN surface ensured the coating of the particles.^37,38 DMPC coating of Fe₃O₄@MSN have been achieved in a high ionic strength medium (Hepes 20 mM, NaCl 150 mM, pH = 7.4). This process has been described as a one-step event in a high ionic strength medium, by van der Waals attractions between zwitterionic SUV and MSN surface.³⁸

DMPC coated structures were firstly characterized using HD and ZP measurements. After lipid coating, the HD was increased to 179.40 ± 2.40 nm and ZP was increased to −10.80 ± 0.25 mV as compared to 125.47 ± 1.61 nm and −32.09 ± 1.30 mV for naked nanoparticles. The effective coating of Fe₃O₄@MSN by the DMPC bilayer was confirmed by cryoTEM observations. CryoTEM observations show that no free SUV was present in the preparation; a unilamellar lipid bilayer is effectively wrapping the surface of the NPs (Fig. 4). Moreover, most of the NPs observed were covered by a complete DMPC lipid bilayer around the NPs in our experimental conditions. After counting 250 particles 84% were found completely wrapped. Scanning transmission electron microscopy (STEM) was used to confirm the presence of phospholipids around Fe₃O₄@MSN. After TEM observation of Fe₃O₄@MSN samples, topochemical analysis was performed by scanning with energy dispersive X-ray spectrometer (EDS). The STEM images (Fig. 4c and d) and the corresponding elemental analysis (ESI Fig. 8†) clearly distinguished from the inside to the outside part of the particles, the iron oxide core, the silica shell and in case of lipid covered particles the presence of a sphere mapping the presence of phosphorus due to the presence of DMPC.

Fig. 4 Imaging of magnetic Fe₃O₄@MSN core–shell particles after incubation with DMPC SUVs (a). All the MSN are covered with a complete lipid bilayer having a thickness of 5 nm. Three lipid-coated MSN are zoomed for a better observation of the lipid bilayer (b). STEM images DMPC Fe₃O₄@MSN: (c) DMPC–Fe₃O₄@MSN overlay of TEM black field (BF), iron (Fe), silica (Si) and phosphorus (P) element cartography. Each element is separately presented (d). The iron core at the centre of the silica NPs and phosphorus is located around the silica shell of Fe₃O₄@MSN particles.

2.2 Cell viability and cell impedance measurements

We investigated the biocompatibility of the 3 types of nanoparticles. For this, the cytotoxicity of MSNs (ranging from 1 to 150 μg mL⁻¹) on human hepatocyte carcinoma (Hep-G2) cells was measured with the MTT assay. The percentage of living cells was up to 90% for 100 μg mL⁻¹ of PEG Fe₃O₄@MSN, and 80% and 90% for 50 μg mL⁻¹ of DMPC Fe₃O₄@MSN and pristine ones, respectively (Fig. 5). This result demonstrates the absence of a significant toxicity of these batches of nanoparticles for the dose of 50 μg mL⁻¹ and a higher biocompatibility for PEG Fe₃O₄@MSN.

Fig. 5 Cytotoxicity analysis: Hep-G2 cells were incubated with increasing concentrations of MSNs (from 1 to 100 μg mL⁻¹). After 72 h treatment, a MTT assay was performed and data are mean ± SD of 3 experiments.

Following this cell viability assay, the two doses 50 and 100 μg mL⁻¹ were chosen for real-time cell impedance analyses. The impact of nanoparticles on cells was performed using the xCELLigence System (ACEA Biosciences), a cell-based microelectronic biosensor which provides real-time and label-free cellular analyses, represented by a Cell Index (CI). This assay allows exceeding the limits of endpoint analysis by capturing data throughout the entire time course of an experiment and obtaining more physiologically relevant data. CI reflects both modifications of cell morphology and cell viability. As control, all these nanoparticles were tested in acellular conditions and no interference on impedance measurements was recorded (data not shown). Impedance was recorded for 60 h of exposure. Pristine Fe₃O₄@MSN (blue curves in Fig. 6) induced a reduction of the CI in a dose-dependent manner, compared to control cells (black curve in Fig. 6). At both doses, i.e. 100 and 50 μg mL⁻¹, a biphasic CI response was observed with an increase followed by a decrease from 10 h of exposure until the end of the experiment. PEG–Fe₃O₄@MSN (purple curves in Fig. 6) at 100 μg mL⁻¹ induced a dose-dependent reduction of the CI with a biphasic pattern, similar to that observed with pristine Fe₃O₄@MSN. Interestingly, PEG Fe₃O₄@MSN at 50 μg mL⁻¹ had no effect on the CI of Hep-G2 cells. DMPC Fe₃O₄@MSN (red curves in Fig. 6) induced an early biphasic CI response (after 5 h of exposure) at both doses. After 60 h of exposure, DMPC Fe₃O₄@MSN at 100 μg mL⁻¹ had the same CI than pristine Fe₃O₄@MSN at 50 μg mL⁻¹. To summarize, pristine Fe₃O₄@MSN are the most deleterious for Hep-G2 cells, followed by DMPC Fe₃O₄@MSN and PEG Fe₃O₄@MSN.

Fig. 6 xCELLigence experiment. Real-time cell index (CI) monitoring of Hep-G2 cells (n = 3) exposed to 50 and 100 μg mL⁻¹ of pristine, PEG- and DMPC Fe₃O₄@MSN. (a) Pristine Fe₃O₄@MSN versus PEG Fe₃O₄@MSN. (b) Pristine Fe₃O₄@MSN versus DMPC Fe₃O₄@MSN.

Finally, the uptake of the Fe₃O₄@MSN, decorated or not with PEG or DMPC was assessed after 6 h of exposure at 50 μg mL⁻¹ via TEM. A high amount of Fe₃O₄@MSN was observed inside the Hep-G2 cells, as showed in Fig. 7b. Pristine Fe₃O₄@MSN were shown as individual dispersed NPs around the nucleus (Fig. 7b). In contrast, in the same experimental conditions, we observed a lower uptake of PEG Fe₃O₄@MSN and DMPC Fe₃O₄@MSN by Hep-G2 cells (Fig. 7c and d).

Fig. 7 TEM imaging of Hep-G2 cells (a) unexposed (control) and exposed at 50 μg mL⁻¹ for 6 h to (b) Fe₃O₄@MSN, (c) PEG Fe₃O₄@MSN, (d) DMPC Fe₃O₄@MSN. Hep-G2 cells were then fixed for TEM observation. Scale bar 500 nm, N indicates the nucleus and arrows evidence the NPs.

Results obtained in cellular assays shown, firstly, that pristine Fe₃O₄@MSN, PEG grafted Fe₃O₄@MSN, and lipid coated Fe₃O₄@MSN, induced a dose-dependent response on liver Hep-G2 cells. Fe₃O₄@MSN grafted with PEG did not impact the normal behavior of Hep-G2 cells at a dose of 50 μg mL⁻¹. After the cells were exposed to NPs; changes in the cell impedance profiles were observed earlier for DMPC Fe₃O₄@MSN than for pristine and PEG Fe₃O₄@MSN. The use of PEG is well known to produce a steric hindrance around nanoparticles which prevents the adsorption of serum proteins and can counteract the opsonization process.¹⁴ Similar observations were also applicable to DMPC Fe₃O₄@MSN, but to a lesser extent. Indeed, DMPC Fe₃O₄@MSN (i.e. 50 or 100 μg mL⁻¹), induced a decrease of CI values all along the experiment, reflecting earlier and more deleterious effects on Hep-G2 cells than those observed with PEG Fe₃O₄@MSN. Nevertheless, these two coverages were found to diminish the toxic potential of the pristine Fe₃O₄@MSN.

Previous studies described a longer half-life in vivo (longer circulation time)³⁹ and better accumulation in tumors¹⁶ for MSN coated with lipid bilayers, associated to PEG groups. However, the individual effect of the lipid bilayer or PEG groups is not clear. Similarly, grafting of PEG onto commercially available liposomal formulations is currently used to reduce the amount of proteins adsorbed on liposome. However, the molecular consistency of the corona is not significantly different in the absence or in the presence of PEG coating.^40,41 Plus, corona formation on liposomes inhibits their internalization and their pegylation reduces also the internalization of liposomes.

3 Conclusion

In this work, the synthesis procedure of magnetic NPs was fully optimized to obtain Fe₃O₄@MSN core@shell nanoparticles, monodisperse, homogeneous, and containing one magnetic core per particle. These criteria were defined to obtain reproducible biological results after interactions of NPs with cells. Then, we performed a preliminary study of the impact on cell fate of the pristine, PEG-grafted and lipid-coated Fe₃O₄@MSN.

The Fe₃O₄@MSN were synthesized according to an adapted protocol that allows obtaining spherical NPs presenting one magnetic core per MSN and a primary diameter of 100 ± 5 nm. The PEG grafting was achieved in situ, after the last step of the Fe₃O₄@MSN synthesis; whereas SLB formation was achieved by vesicle fusion and rupture upon contact with the MSN surface. The polydispersity index for the hydrodynamic diameters measured for pristine Fe₃O₄@MSN, PEG grafted ones and for lipid coated nanoparticles are always inferior to 0.2 indicating a monodispersity, and the absence of aggregates when NPs were dispersed in 20 mM Hepes pH 7.4 buffer containing 5 mM NaCl. It is noteworthy that the lipid coating ascertained using cryoTEM imaging was confirmed by appropriate STEM analyses.

The NPs behaviors with three different surfaces were compared using cellular tests. The results show differences in cell responses upon the exposure with NPs having three different surface properties. PEG Fe₃O₄@MSN appear less toxic than lipid coated and pristine ones. Only the PEG Fe₃O₄@MSN exhibits no toxic effect at a dose of 50 μg mL⁻¹ which was not the case for DMPC and pristine Fe₃O₄@MSN. DMPC Fe₃O₄@MSN (i.e. 50 or 100 μg mL⁻¹), induced a decrease of CI values all along the experiment, reflecting earlier and more deleterious effects on Hep-G2 cells than those observed with PEG Fe₃O₄@MSN. Nevertheless, these two coverages were found to diminish the toxic potential of the pristine Fe₃O₄@MSN.

To the best of our knowledge, this is the first study that compares the cellular effect of PEG grafting or lipid coating as individual strategies for enhancing the stealth effect of MSN. This is particularly interesting because different internalization mechanisms are involved for the MSN presenting different surface coatings. In a previous study, in which MSN coated with lipids were compared to pristine MSN for their interactions towards cells, the data suggested internalization of large aggregates of pristine MSN inside vesicles whereas lipid-coated MSN enter in the cells individually.⁴² The optimized synthesis protocol of Fe₃O₄@MSN developed in this work allows the comparison of MSN that were fully characterized in terms of physicochemical properties. This study is an essential step to ensure the quality of NPs material for medical applications. As a lower toxicity was observed for the PEG grafted or lipid coated Fe₃O₄@MSN compared to the pristine ones, as the real-time impedance profiles suggested different internalization processes; future work should be orientated towards intracellular trafficking of NPs and in depth analysis of cellular responses. Moreover, in vitro studies using model membranes will be needed to clarify the interactions between pristine, PEG and DMPC Fe₃O₄@MSN with membranes.

4 Experimental

4.1 Synthesis

4.1.1 Materials. All reagents were commercially available and used without any further purification. Hydrated iron oxide [FeO(OH), catalyst grade 30–50 mesh], oleic acid (90%), oleylamine (99%), ether (≥99.9%), anhydrous ethanol (≥99.8%), anhydrous pentane (99+%), anhydrous chloroform (99+%), tetraethoxysilane (TEOS, ≥99.9%), cetyltrimethylammonium bromide (CTAB), ammonium nitrate (NH₄NO₃), dimethyl sulfoxide (DMSO) and (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) (Hepes) were purchased from Sigma-Aldrich. Sodium hydroxide (NaOH) and n-docosane (99%) were purchased from Acros. 1,2-Dimyristoyl-sn-glycero-3-phosphocholine (DMPC) was purchased from Avanti polar lipids. DPBS buffer (KCl 2.66 mM, KH₂PO₄ 1.47 mM, NaCl 137.93 mM, Na₂HPO₄–7H₂O 8.05 mM) was provided from Gibco (ThermoFisher scientific). Silanized PEG (CH₃O–PEG₂₀₀₀–Si(OCH₃)₃) was purchased from Rapp polymer. RPMI 1640 (Roswell Park Memorial Institute medium) and fetal calf serum (FCS) were obtained from Invitrogen. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was provided from Promega.

4.1.2 Preparation of magnetic Fe₃O₄ nanocrystals (MIONs). The synthesis of spherical magnetic Fe₃O₄ nanocrystals described in ESI† was achieved according to the procedure previously reported in the literature.⁴³

4.1.3 Synthesis of monodisperse mesoporous silica shell with magnetic cores (Fe₃O₄@MSN). The synthesis of Fe₃O₄@MSN followed the procedure initially reported by Rho et al.³³ This procedure was slightly modified in order to avoid efficiently the heterogeneous nucleation as depicted in ESI.†. PEG Fe₃O₄@MSN nanoparticles have been synthesized with the protocol described in ESI.† After the condensation process induced by the injection of TEOS, the mixture was slowly cooled to 50 °C with continuous stirring. Then, 1 mL ethanol containing 100 mg of silanized PEG 2000 was slowly added and stirred overnight. The mixture was cooled to room temperature before applying the same washing steps as described in ESI† for pristine Fe₃O₄@MSN. Lipid-coating of Fe₃O₄@MSN core–shell nanoparticles was performed as follows. Small unilamellar vesicles (SUVs) were prepared following Bangham's method.⁴⁴ Briefly, a dry lipid film was prepared from a DMPC lipid stock solution that was stored at 5 mg mL⁻¹ in chloroform at −20 °C. The dry lipid film at the bottom of a roundish glass tube was obtained by solvent evaporation under nitrogen flow and was kept under vacuum overnight. The lipid film was resuspended at 10 mg mL⁻¹ of lipids in PBS buffer and vortexed for 5 minutes. SUVs were obtained by extrusion through polycarbonate membranes. Lipid suspension was passed back and forth approximately 15 times through 100 nm pores and 25 times through 50 nm pores. The resulting SUVs have a hydrodynamic diameter of 67.0 ± 1.3 nm and a zeta potential of −4.45 ± 0.66 mV. For lipid coating, Fe₃O₄@MSN were dispersed ultrasonically for 2 minutes at 20 mg mL⁻¹ in absolute ethanol. Fe₃O₄@MSN dispersion was added to 6.4 mL of 5 mg mL⁻¹ SUV suspension (corresponding to a surface area ratio 8/1 SUVs/NPs). This mixture was vortexed for 1 minute and sonicated in ultrasound bath for 1 minute. Rotating agitation of the sample was performed for 3 hours at 37 °C. Lipid coated DMPC Fe₃O₄@MSN were then isolated by 4 centrifugation steps (4000g, 20 minutes), to remove exceeding SUVs.

4.2 Methods of characterization

Transmission electron micrographs (TEM) were obtained with a JEOL 1200 EX II microscope. Nitrogen adsorption and desorption isotherms were measured using a TriStar 3000 (V6.06 A) automated gas adsorption system with nitrogen as adsorbate at −196 °C. Prior to the sorption experiment, the sample air was evacuated under vacuum at 110 °C for 15 h. The specific surface area (S_BET) was calculated according to the Brunauer–Emmett–Teller (BET) method, taking the value of 0.162 nm² as a cross-sectional area per nitrogen molecule in the BET calculation. The total pore volume (V_p) was taken at P/P₀ = 0.85. Pore size distribution (PSD), used to measure the pore diameter (D_p), was obtained from the desorption branch of the nitrogen isotherm using the Barrett–Joyner–Halenda (BJH) equation. X-ray diffraction patterns (XRD) were collected on a Bruker AXS D8 Advance with Solid 1D Lynx Eyes Detector, using Ni-filtered Cu Kα radiation. XRD techniques were applied to determine the pore ordering and the presence of crystalline phases of Fe₃O₄. BJH parameters were employed to calculate the effective pore diameter.²⁰ Thermogravimetric analysis (TGA) was performed with DSC 7 (Perkin Elmer), from 30 to 780 °C.

Hydrodynamic diameter and zeta potential were determined using a Nano ZS apparatus (Malvern). Each measurement was performed at 20 μg mL⁻¹ dilution NPs from a stock suspension at 10 mg ml⁻¹ in EtOH 95%. The dilution required a 30 s bath sonication step of the stock suspension and 1 min bath sonication step of the dilution. The suspensions were prepared by dilution in Hepes 20 mM, NaCl 5 mM buffered at pH 7.4 and 25 °C. Data were collected from the He–Ne laser light source (633 nm) at 173° from the transmitted beam. Results are presented as Z-average obtained in intensity mode, associated to the polydispersity index.

CryoTEM experiments have been performed using a JEOL 220FS transmission electron microscope (Japan) with a 4k × 4k slow-scan CCD camera (Gatan, USA). Samples were prepared on copper quantifoil grids R 2/2 (Quantifoil, Germany). DMPC Fe₃O₄ MSN suspensions were prepared at a concentration of 3 mg mL⁻¹ of particles. A 3 μL drop of suspension was filed on the copper grid, dried 2 seconds and frozen in ethane for approximately 6 seconds. Grids were kept in liquid nitrogen until microscope observation. Effective NPs coverage was characterized with cell counter (Image J). Scanning transmission electron microscopy (STEM) was performed using a JEOL 2100F transmission electron microscope (Japan) with 200 kV field emission and energy dispersive X-ray spectrometer. Pristine and DMPC Fe₃O₄@MSN samples were analyzed on copper grids from suspensions containing a concentration of 3 mg mL⁻¹ of particles. After TEM observation of Fe₃O₄@MSN samples, element analysis have been performed by scanning with a counting rate of 398 and 422 counts per sec for pristine Fe₃O₄@MSN and DMPC Fe₃O₄@MSN, respectively. The scanning was performed using a 0.7 nm EDX probe with a probe current of 1 nA.

4.3 Biological materials and methods

4.3.1 Cell culture. The Human hepatocyte carcinoma (Hep-G2) cell line was obtained from Sigma-Aldrich. Cells were cultured in RPMI 1640 (Invitrogen) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (100 U mL⁻¹, 100 μg mL⁻¹) and incubated in a humidified incubator at 37 °C and 5% CO₂. Cells were used between passages 20 to 40. Cells were passed once a week and the medium was changed twice a week, keeping confluence below 80%.

4.3.2 Cytotoxic study. Hep-G2 cells were seeded into 96-well plates at 500 cells per well in 200 μL culture medium and allowed to grow for 24 h. Then, increasing concentrations of nanoparticles (from 1 to 150 μg mL⁻¹) were incubated in culture medium of Hep-G2 cells during 72 h. To evaluate the toxicity of these treatments, a MTT assay was performed. Briefly, cells were incubated for 4 h with 0.5 mg mL⁻¹ of MTT in media. The MTT/media solution was then removed and the precipitated crystals were dissolved in EtOH/DMSO (1 : 1). The solution absorbance was read at 540 nm.

4.3.3 Real-time impedance measurement. The experiments were performed using an xCELLigence System from ACEA Biosciences. A background resistance of the E-plates (ACEA) was determined with 100 μL culture medium. Hep-G2 cells were seeded in E-plates at 10 000 cells in 200 μL total volume. E-plates were placed into the Real-Time Cell Analyzer (RTCA) station (ACEA) and incubated at 37 °C and 5% CO₂. Cells were grown for 48 hours corresponding to the adhesion and growth phases. After 48 hours, cells were exposed (n = 3) to three types of nanoparticles: pristine Fe₃O₄@MSN, PEG grafted and lipid-bilayer coated Fe₃O₄@MSN, exposed at 50 and 100 μg mL⁻¹. Real-time impedance measurements of control cells and nanoparticles alone in culture cell media were also recorded as controls. Cell index (CI) raw data values were calculated by the RTCA software 2.0. Normalized cell indexes were also calculated at a selected normalization time point, which was chosen at the time of addition of nanoparticles.

4.3.4 Following NPs by TEM. Hep-G2 cells were seeded on glass coverslips for 24 h. After controlling their adherence and growing, cells were exposed to 50 μg mL⁻¹ of NPs in RPMI cell culture medium for 6 h. The medium was removed and cells were rinsed twice with PBS. Cells were fixed by incubation with 2.5% (v/v) glutaraldehyde in PBS buffer, for 1 h at room temperature (RT). After that, cells were extensively washed with PBS. The staining of samples was obtained upon incubation with 1% osmium tetroxide. Samples were dehydrated by ascending grades of EtOH; for impregnation, the samples were firstly treated with a mix EtOH/EPON resin (1 : 1, v/v) for 1 h, and twice in EPON for 2 h. The polymerization is performed by embedding cells in EPON resin for 12 h at 60 °C, plunged in liquid nitrogen at −195 °C to detach the coverslip, and placed for two days at 60 °C for polymerization. The ultrathin sections (70 nm) were obtained using an ultramicrotome and disposed on the copper grids. The grids were incubated in uranyl acetate for 2 min, rinsed in water, and then incubated in lead citrate for 2 min, and finally rinsed in water.

Abbreviations

BET	Brunauer Emmett Teller
CI	Cell index
CTAB	Cetyltrimethylammonium bromide
CryoTEM	Cryogenic transmission electron microscopy
DMPC	1,2-Dimyristoyl-sn-glycero-3-phosphocholine
EDS	Energy dispersive X-ray spectrometer
EDX	Energy dispersive X-ray
EPR	Enhanced permeation and retention
EtOH	Ethanol
FCS	Fetal calf serum
HD	Hydrodynamic diameter
Hepes	(4-(2-Hydroxyethyl)-1-piperazineethanesulfonic acid)
Hep-G2 cells	Human hepatocyte carcinoma cells
MIONs	Magnetic iron oxide nanocrystals
MRI	Magnetic resonance imaging
MSN	Mesoporous silica nanoparticles
MTT	3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide
NPs	Nanoparticles
PEG	Polyethylene glycol
RPMI	Roswell Park Memorial Institute medium
STEM	Scanning transmission electron microscopy
SUV	Small unilamellar vesicles
TEM	Transmission electron microscopy
TGA	Thermogravimetric analysis
TEOS	Tetraethoxysilane
XRD	X-ray diffraction
ZP	Zeta potential

Acknowledgements

This work was supported by the French national research agency (ANR-13-NANO-0007, BioSiPharm project). We thank technological platform of microscopy of Montpellier University for the use of their electron microscope and their technical support, in particular Franck Godiard and Véronique Richard. We thank Gilles Renou from SIMaP of Grenoble for STEM analysis.

References

1. S. Nie, Y. Xing, G. J. Kim and J. W. Simons, Annu. Rev. Biomed. Eng., 2007, 9, 257–288.
2. A. E. Nel, L. Madler, D. Velegol, T. Xia, E. M. Hoek, P. Somasundaran, F. Klaessig, V. Castranova and M. Thompson, Nat. Mater., 2009, 8, 543–557.
3. H. Yang, C. Liu, D. Yang, H. Zhang and Z. Xi, J. Appl. Toxicol., 2009, 29, 69–78.
4. N. Ma, C. Ma, C. Li, T. Wang, Y. Tang, H. Wang, X. Moul, Z. Chen and N. Hel, J. Nanosci. Nanotechnol., 2013, 13, 6485–6498.
5. B. Fadeel and A. E. Garcia-Bennett, Adv. Drug Delivery Rev., 2010, 62, 362–374.
6. T. Yu, K. Greish, L. D. McGill, A. Ray and H. Ghandehari, ACS Nano, 2012, 6, 2289–2301.
7. EMA, Reflection paper on non-clinical studies for generic nanoparticle iron medicinal product applications, 2011.
8. FDA, Nanotechnology task force report, 2007.
9. D. Tarn, C. E. Ashley, M. Xue, E. C. Carnes, J. I. Zink and C. J. Brinker, Acc. Chem. Res., 2013, 46, 792–801.
10. M. Liong, J. Lu, M. Kovochich, T. Xia, S. G. Ruehm, A. E. Nel, F. Tamanoi and J. I. Zink, ACS Nano, 2008, 2, 889–896.
11. Z. Li, J. C. Barnes, A. Bosoy, J. F. Stoddart and J. I. Zink, Chem. Soc. Rev., 2012, 41, 2590–2605.
12. K. Pombo Garcia, K. Zarschler, L. Barbaro, J. A. Barreto, W. O'Malley, L. Spiccia, H. Stephan and B. Graham, Small, 2014, 10, 2516–2529.
13. V. Cauda, C. Argyo and T. Bein, J. Mater. Chem., 2010, 20, 8693–8699.
14. C. Graf, Q. Gao, I. Schutz, C. N. Noufele, W. Ruan, U. Posselt, E. Korotianskiy, D. Nordmeyer, F. Rancan, S. Hadam, A. Vogt, J. Lademann, V. Haucke and E. Ruhl, Langmuir, 2012, 28, 7598–7613.
15. J. Liu, C. Detrembleur, D. Pauw-Gillet, S. Mornet, L. Vander Elst, S. Laurent, C. Jérôme and E. Duguet, J. Mater. Chem. B, 2014, 2, 59–70.
16. H. Meng, M. Wang, H. Liu, X. Liu, A. Situ, B. Wu, Z. Ji, C. H. Chang and A. E. Nel, ACS Nano, 2015, 9, 3540–3557.
17. J. Liu, S. Z. Qiao, Q. H. Hu and G. Q. Lu, Small, 2011, 7, 425–443.
18. F. M. Martin-Saavedra, E. Ruiz-Hernandez, A. Bore, D. Arcos, M. Vallet-Regi and N. Vilaboa, Acta Biomater., 2010, 6, 4522–4531.
19. J. Kim, H. S. Kim, N. Lee, T. Kim, H. Kim, T. Yu, I. C. Song, W. K. Moon and T. Hyeon, Angew. Chem., Int. Ed. Engl., 2008, 47, 8438–8441.
20. M. Kruk, M. Jaroniec and A. Sayari, J. Phys. Chem. B, 1997, 101, 583–589.
21. Y.-S. Lin and C. Haynes, Chem. Mater., 2009, 21, 3979–3986.
22. K. R. Hurley, Y.-S. Lin, J. Zhang, S. M. Egger and C. L. Haynes, Chem. Mater., 2013, 25, 1968–1978.
23. R. G. Digigow, J.-F. Dechézelles, H. Dietsches, I. Geissbühler, D. Vanhecke, C. Geers, A. M. Hirt, B. Rothen-Rutishauer and A. Petri-Fink, J. Magn. Magn. Mater., 2014, 362, 72–79.
24. V. Maurice, T. Georgelin, J.-M. Siaugue and V. Cabuil, J. Magn. Magn. Mater., 2009, 321, 1408–1413.
25. E. Bringas, O. Koysuren, D. V. Quach, M. Mahmoudi, E. Aznar, J. D. Roehling, M. D. Marcos, R. Martinez-Manez and P. Stroeve, Chem. Commun., 2012, 48, 5647–5649.
26. M. Perrier, M. Gary-Bobo, L. Lartigue, D. Brevet, A. Morère, M. Garcia, P. Maillard, L. Raehm, Y. Guari, J. Larionova, J.-O. Durand, O. Mongin and M. Blanchard-Desce, J. Nanopart. Res., 2013, 15, 1–17.
27. F. Tang, L. Li and D. Chen, Adv, Mater., 2012, 24, 1504–1534.
28. C. He, Y. Hu, L. Yin, C. Tang and C. Yin, Biomaterials, 2010, 31, 3657–3666.
29. W. W. Yu, J. C. Falkner, C. T. Yavuz and V. L. Colvin, Chem. Commun., 2004, 2306–2307.
30. J. Park, K. An, Y. Hwang, J.-G. Park, H.-J. Noh, J.-Y. Kim, J.-H. Park, N.-M. Hwang and T. Hyeon, Nat. Mater., 2004, 3, 891–895.
31. J.-F. Dechezelles, V. Malik, J. J. Crassous and P. Schurtenberger, Soft Matter, 2013, 9, 2798–2802.
32. J. L. Nyalosaso, G. Derrien, C. Charnay, L.-C. de Menorval and J. Zajac, J. Mater. Chem., 2012, 22, 1459–1468.
33. W.-Y. Rho, H.-M. Kim, S. Kyeong, Y.-L. Kang, D.-H. Kim, H. Kang, C. Jeong, D.-E. Kim, Y.-S. Lee and B.-H. Jun, J. Ind. Eng. Chem., 2014, 20, 2646–2649.
34. M. Catauro, F. Bollino, F. Papale, M. Gallicchio and S. Pacifico, Mater. Sci. Eng., C, 2015, 48, 548–555.
35. H. Y. Jung, R. K. Gupta, E. O. Oh, Y. H. W. Kim and C. M. Whang, J. Non-Cryst. Solids, 2005, 351, 372–379.
36. S. Ek, A. Root, M. Peussa and L. Niinista, Thermochim. Acta, 2001, 379, 201–212.
37. R. Veneziano, G. Derrien, S. Tan, A. Brisson, J. M. Devoisselle, J. Chopineau and C. Charnay, Small, 2012, 8, 3674–3682.
38. H. Wang, J. Drazenovic, Z. Luo, J. Zhang, H. Zhou and S. L. Wunder, RSC Adv., 2012, 2, 11336–11348.
39. M. M. van Schooneveld, E. Vucic, R. Koole, Y. Zhou, J. Stocks, D. P. Cormode, C. Y. Tang, R. E. Gordon, K. Nicolay, A. Meijerink, Z. A. Fayad and W. J. Mulder, Nano Lett., 2008, 8, 2517–2525.
40. M. A. Dobrovolskaia, B. W. Neun, S. Man, X. Ye, M. Hansen, A. K. Patri, R. M. Crist and S. E. McNeil, Nanomedicine, 2014, 10, 1453–1463.
41. M. Hadjidemetriou, Z. Al-Ahmady, M. Mazza, R. F. Collins, K. Dawson and K. Kostarelos, ACS Nano, 2015, 9, 8142–8156.
42. D. B. Tada, E. Suraniti, L. M. Rossi, C. A. Leite, C. S. Oliveira, T. C. Tumolo, R. Calemczuk, T. Livache and M. S. Baptista, J. Biomed. Nanotechnol., 2014, 10, 519–528.
43. S. Dib, M. Boufatit, S. Chelouaou, F. Sadi-Hassaine, J. Croissant, J. Long, L. Raehm, C. Charnay and J. O. Durand, RSC Adv., 2014, 4, 24838–24841.
44. A. D. Bangham, M. M. Standish and J. C. Watkins, J. Mol. Biol., 1965, 13, 238–252.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra09017f

---