Biosafety of Mesoporous Silica Nanoparticles

Abstract

Careful analysis of any new nanomedicine device or disposal should be undertaken to comprehensively characterize the new product before application, so that any unintended side effect is minimized. Because of the increasing number of nanotechnology-based drugs, we can anticipate that regulatory authorities might adapt the approval process for nanomedicine products due to safety concerns, e.g., request a more rigorous testing of the potential toxicity of nanoparticles (NPs). Currently, the use of mesoporous silica nanoparticles (MSN) as drug delivery systems is challenged by a lack of data on the toxicological profile of coated or non-coated MSN. In this context, we have carried out an extensive study documenting the influence of different functionalized MSN on the cellular internalization and in vivo behaviour. In this article, a synthesis of these works is reviewed and the perspectives are drawn. The use of magnetic MSN (Fe₃O₄@MSN) allows an efficient separation of coated NPs from cell cultures with a simple magnet, leading to results regarding corona formation without experimental bias. Our interest is focused on the mechanism of interaction with model membranes, the adsorption of proteins in biological fluids, the quantification of uptake, and the effect of such NPs on the transcriptomic profile of hepatic cells that are known to be readily concerned by NPs' uptake in vivo, especially in the case of an intravenous injection.

Keywords: adverse outcome pathways; internalization; mesoporous silica; nanoparticles; protein corona; safety.

Figure 1

Schematic representation of the synthesis of bare magnetic mesoporous silica nanoparticles (Fe₃O₄@MSN). Firstly, Fe₃O₄ nanocrystals (NC) are obtained by thermal decomposition of FeO(OH). In another flask, cetyl trimethylammonium bromide (CTAB) micelles were obtained in alkaline water, at a temperature of 80 °C. Fe₃O₄ NC, after stabilization in oleylamine, were progressively added to the CTAB micelles, in 10 steps. After that, tetraethylorthosilicate (TEOS) has been added for sol–gel reaction and formation of Fe₃O₄MSN. Different washing steps were performed to extract CTAB surfactant from the pores.

Figure 2

(a) Different proteomic profiles of the protein corona after Fe₃O₄@MSN separation by magnetization (blue) or centrifugation (green). Sizes and colors of human protein clues are proportional to their relative percentage within the corona. The 35 highest abundant proteins are labelled. The color scale unit is Normalized Spectral Abundance Factor (%). Reproduced from Pisani et al. 2017 [17] with permission from the Royal Society of Chemistry. (b) Protein–protein interactions are represented in a network developed using the NetworkAnalyst software [18,19], based on InnateDB [20]. The target represents the time scale (0.5 min to 7 days). Each protein (represented by its gene symbol) is placed according to its time of appearance within the corona. The colors indicate the cluster membership. The grey lines represent the protein–protein interactions. Proteins that have a lot of interactions with other proteins are represented by a larger visual cue. Reproduced from Pisani et al. 2017 [5] with permission from the Royal Society of Chemistry.

Figure 2

(a) Different proteomic profiles of the protein corona after Fe₃O₄@MSN separation by magnetization (blue) or centrifugation (green). Sizes and colors of human protein clues are proportional to their relative percentage within the corona. The 35 highest abundant proteins are labelled. The color scale unit is Normalized Spectral Abundance Factor (%). Reproduced from Pisani et al. 2017 [17] with permission from the Royal Society of Chemistry. (b) Protein–protein interactions are represented in a network developed using the NetworkAnalyst software [18,19], based on InnateDB [20]. The target represents the time scale (0.5 min to 7 days). Each protein (represented by its gene symbol) is placed according to its time of appearance within the corona. The colors indicate the cluster membership. The grey lines represent the protein–protein interactions. Proteins that have a lot of interactions with other proteins are represented by a larger visual cue. Reproduced from Pisani et al. 2017 [5] with permission from the Royal Society of Chemistry.

Figure 3

Characterization of native, polyethylene glycol (PEG)-grafted, and lipid-coated Fe₃O₄@MSN. (a) Transmission electron microscopy (TEM) imaging of (1) native Fe₃O₄@MSN, (2) PEG-grafted Fe₃O₄@MSN, and (3) lipid-coated Fe₃O₄@MSN with dimyristoyl phosphatidylcholine (DMPC) lipids, showing a primary diameter of 100 nm, with very homogeneous shape, porosity, and coverage. Reproduced from Pisani et al. 2017 [6], published under the Creative Commons Attribution (CC BY-NC-ND 4.0) license (https://creativecommons.org/licenses/by-nc-nd/4.0/). (b) Characterization of Fe₃O₄@MSN PEG-grafting, with thermogravimetric analysis (TGA)/dynamic scanning calorimetry (DSC) spectra of (1) pristine Fe₃O₄@MSN and (2) PEG–Fe₃O₄@MSN. (c) Imaging of magnetic Fe₃O₄@MSN core–shell particles after incubation with DMPC small unilamellar vesicles (SUVs) (1). All the MSN are covered with a complete lipid bilayer, having a thickness of 5 nm. Three lipid-coated MSN are zoomed in on for a better observation of the lipid bilayer. (2) Scanning transmission electron microscopy (STEM) images of DMPC Fe₃O₄@MSN: DMPC Fe₃O₄@MSN overlay of TEM black field (BF), iron (Fe), silica (Si), and phosphorus (P) element cartography. (3) Each element is separately presented. The iron core localizes at the center of the silica nanoparticles and phosphorus is localized around the silica shell of the Fe₃O₄@MSN particles. (b) and (c) are reproduced and adapted from Nyalosaso et al. 2016 [3] with permission from the Royal Society of Chemistry.

Figure 4

Characterization of native and coated Fe₃O₄@MSN behavior in suspension in complex media, with or without proteins. (a) (1) and (2): Hydrodynamic diameter (HD) and polydispersity index (PDI), represented respectively by bars and dots, for native (blue), polyethylene glycol (PEG) (orange) and dimyristoyl phosphatidylcholine (DMPC) (red) Fe₃O₄@MSN in (x) HEPES buffered saline (HBS) 150 mM NaCl (pH 7.4) and (y) HBS 150 mM NaCl (pH 7.4) containing 10% fetal calf serum (FCS). (b) Quartz crystal microbalance with dissipation (QCM-D) frequency sensorgram following the interaction between nanoparticles and the egg phosphatidyl choline (EPC)-supported lipid bilayer (SLB). Native (blue), PEG (orange), and DMPC (red) Fe₃O₄@MSN were flowed into HBS 150 mM NaCl 10% SCF medium on the top of EPC SLB, at a concentration of 0.25 mg mL⁻¹ of nanoparticles. After adding Fe₃O₄@MSN into the medium on the top of the EPC SLB for 15 min, the flow was stopped for 10 h. The results on the variations of frequency are presented after the offset of the lipid bilayer formation. Reproduced and adapted from Rascol et al. 2017 [4], published under the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Figure 5

Biodistribution of Fe₃O₄@MSN in mice. (a) Quantification of silicon in different organs four days after injection. Inductively coupled plasma–mass spectrometry (ICP–MS) was used after acid digestion to quantify the silicon in the liver, lungs, spleen, kidneys, and urine four days after intravenous injection of native (blue), polyethylene glycol (PEG) (orange) and dimyristoyl phosphatidylcholine (DMPC) (red) Fe₃O₄@MSN at a concentration of 40 mg kg⁻¹ in comparison to control mice (white). (b) Nanoparticle level in blood. The silicon levels in blood were measured 2, 6, 24 h, and 4 days after intravenous injection of native (blue), PEG (orange), and DMPC (red) Fe₃O₄@MSN at a concentration of 40 mg kg⁻¹. The dashed line indicates the silicon level found in blood of control mice. For this experiment, 20 mice were divided into four groups of five animals. The values of the histograms represent the mean ± standard deviation (SD) of values of each animal of a group. * p < 0.05 indicates that a group is statistically different from all other groups treated with nanoparticles. Reproduced and adapted from Rascol et al. 2017 [4], published under the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Figure 6

Transmission electron microscopy imaging of HepG2 cells exposed for 3, 6, and 24 h at 50 μg mL⁻¹ for (a–c) native, (d–f) polyethylene glycol (PEG)-coated, or (g–i) dimyristoyl phosphatidylcholine (DMPC)-coated Fe₃O₄@MSN. The nanoparticles are indicated by arrows, near the cell membrane (M) or the nucleus (N). Reproduced from Rascol et al. 2017 [4], published under the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Figure 7

xCELLigence experiment. Real-time cell index (CI) monitoring of HepG2 cells (n = 3) exposed to 50 and 100 mg mL^-1 of pristine, polyethylene glycol (PEG)-coated, and dimyristoyl phosphatidylcholine (DMPC)-coated Fe₃O₄@MSN. (a) Pristine Fe₃O₄@MSN versus PEG-coated Fe₃O₄@MSN. (b) Pristine Fe₃O₄@MSN versus DMPC-coated Fe₃O₄@MSN. Reproduced and adapted from Nyalosaso et al. 2016 [3] with permission from the Royal Society of Chemistry.

Figure 8

Time- and dose-dependent effects of exposure to Fe₃O₄@MSN on the number of significantly differentially expressed genes. HepaRG cells were exposed to 1.6, 16, and 80 µg cm⁻² pristine, polyethylene glycol (PEG)-, and dimyristoyl phosphatidylcholine (DMPC)-coated Fe₃O₄@MSN for 24 (a) or 48 (b) h. After extraction and labeling, RNA was hybridized to a human oligonucleotide microarray (6 × 60 k Agilent V3 SurePrint). Bars represent the number of differentially expressed transcripts after statistical analysis, using Genespring GX13 software (Agilent), and with a p-value < 0.05 and a fold-change (FC) ≥ 2. Reproduced from Pisani et al. 2017 [6], published under the Creative Commons Attribution (CC BY-NC-ND 4.0) license (https://creativecommons.org/licenses/by-nc-nd/4.0/).

Figure 9

Canonical pathways elicited by each Fe₃O₄@MSN (80 mg cm⁻²). The percentage of modulated transcripts of our datasets belonging to six major altered canonical pathways after (a) 24 and (b) 48 h exposure to Fe₃O₄@MSN. These pathways were all significant according to a Fisher’s statistical test (p-value < 0.05), revealed with Ingenuity^® Pathway Analysis (IPA^®, QIAGEN). Reproduced from Pisani et al. 2017 [6], published under the Creative Commons Attribution (CC BY-NC-ND 4.0) license (https://creativecommons.org/licenses/by-nc-nd/4.0/).