This article is distributed under the terms of the Creative Commons Attribution 4.0 International License ( http://creativecommons.org/licenses/by/4.0/ ), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
If you want to learn more, please visit our website.
Large-scale production and use of amorphous silica nanoparticles (SiNPs) have increased the risk of human exposure to SiNPs, while their health effects remain unclear. In this review, scientific papers from to were systematically selected and sorted based on in vitro and in vivo studies: to provide an update on SiNPs toxicity and to address the knowledge gaps indicated in the review of Napierska (Part Fibre Toxicol 7:39, ). Toxicity of SiNPs in vitro is size, dose, and cell type dependent. SiNPs synthesized by wet route exhibited noticeably different biological effects compared to thermal route-based SiNPs. Amorphous SiNPs (particularly colloidal and stöber) induced toxicity via mechanisms similar to crystalline silica. In vivo, route of administration and physico-chemical properties of SiNPs influences the toxicokinetics. Adverse effects were mainly observed in acutely exposed animals, while no significant signs of toxicity were noted in chronically dosed animals. The correlation between in vitro and in vivo toxicity remains less well established mainly due to improperunrealisticdosing both in vitro and in vivo. In conclusion, notwithstanding the multiple studies published in recent years, unambiguous linking of physico-chemical properties of SiNPs types to toxicity, bioavailability, or human health effects is not yet possible.
The comprehensive review of Napierska et al. ( ) suggested that exposure to SiNPs (1100 nm) induced toxic effects in vitro (immortalized mammalian cell lines) and in vivo (rats and mice). Physico-chemical properties such as size, surface area, and surface features were found to play a key role in the toxicity of SiNPs. Importantly, Napierska concluded that physico-chemical properties of SiNPs differ based on their production method and, therefore, may cause different biological effects. However, no definite conclusions were made due to insufficient or no data available for,
Human health effects associated with silica exposure, especially crystalline silica (0.510 µm), have widely been studied. Occupational exposure to crystalline silica induces silicosis in workers (a fibrotic lung disease) and is also associated with lung cancer, emphysema, and pulmonary tuberculosis (Leung et al. ). Conversely, natural amorphous silica is generally considered as less harmful, since the toxicological potential of silica has so far been linked to its crystallinity. Recent studies have revealed that amorphous SiNPs can be as reactive as crystalline particles (Turci et al. ). In vivo, amorphous SiNPs are, however, cleared more rapidly from the lung, which may contribute to explain their lower pathogenic potential (Arts et al. ). The human health effects of nanosilica remain to be clarified and toxicologists believe that exposure to SiNPs, due to their small size, may bring different adverse effects compared to micron-sized silica (Napierska et al. ).
In recent years, large-scale industrial production and global commercialization of SiNPs have resulted in increased risk of human exposures at workplaces (Kim et al. a ; Oh et al. ). Food additive silica (E551) is also in the nano size range (Dekkers et al. ), indicating that the general population is probably more exposed than initially anticipated. Moreover, in view of the efforts to use NM in medical applications, SiNPs could also be intentionally introduced into the human body for disease diagnosis and treatments (Croissant et al. ). Such growing potentials for exposure raised a global concern regarding the safety and potential adverse health effects of SiNPs.
Nanosilica, also known as the nanoform (<100 nm) of silicon dioxide or silica nanoparticles (SiNPs), possesses distinct physico-chemical characteristics compared to its bulk form; smaller size materials have an increased surface-to-volume ratio and a higher surface reactivity (Oberdörster ; Napierska et al. ). Due to their appealing properties, SiNPs are now extensively used in agriculture, food, and consumer products including cosmetics (Napierska et al. ; Khot et al. ; Kasaai ; Brinch et al. ). Until , nearly 1.5 million tons of SiNPs had already been placed in the global market (Liljenström et al. ) and SiNPs became one of the three most produced nanomaterials (NMs) worldwide in . Among the 846 nano-based products listed in a consumer products inventory, approximately 100 claim to contain SiNPs (Vance et al. ). Moreover, amorphous SiNPs are being synthesized with highly tunable biocompatibility and stability, and considered as a very promising candidate for various bio-medical applications such as gene carrier, drug delivery, and molecular imaging (Tang and Cheng ; Bitar et al. ).
Induction of oxidative stress is considered as the major mechanism involved in SiNPs toxicity (Wang et al. ; Ye et al. a , b ) and, therefore, the in vitro section of the review was structured according to toxic endpoints and its association with oxidative stress. For in vivo, studies were sorted based on exposure route and modalities, since they can significantly influence the toxicokinetics of SiNPs. Throughout the review, the following abbreviations were used to indicate the different types of silica nanoparticles: SiNPs when it is not clear which type was used, C-SiNPs for colloidal silica; S-SiNP for stöber silica; M-SiNPs for mesoporous silica; Pr-SiNPs for precipitated silica; and Py-SiNPs for pyrogenic silica.
We searched two databases for papers published from June (after Napierska et al. ) to December . In the PubMed http://www.ncbi.nlm.nih.gov/PubMed and EMBASE https://www.embase.com databases, the following keyword combinations were used: silica nanoparticle OR silica NP AND toxicity NOT review. We retrieved a list of 859 and 405 articles in English, respectively. In a second step, duplicates were removed and the titles were screened to identify studies that best matched with our search terms, leaving 611 relevant papers. In a third step, we excluded papers that met the exclusion criteria and left 128 most relevant papers. Finally, 82 papers reporting a minimum set of physico-chemical characterization and toxic effects were selected for the main content of the review. Notably, only one study was found on the toxicity of crystalline nanosilica (Chu et al. ), but it was finally excluded due to insufficient data on the size of particles. As a result, the review is dealing only with amorphous SiNPs.
In vitro chronic Py-SiNPs (12 nm) exposure of intestinal epithelial cell line (C2BBe1) was examined in a recent study. The cells were exposed to 10-µg/cm 2 SiNPs for 24 h. After 24 h, the medium was replaced (without NPs) and cells were allowed to grow for 46 days. At the end of incubation, cells were passaged and again exposed for 24 h and grown for 46 days; this cycle was repeated for 29 passages (total life span). The cells and supernatants were collected at the end of each passage for analysis. Though the particles were internalized (only in a fraction of cells), no significant induction of necrosis, apoptosis, and LDH release and decrease in cell viability was observed for any of these conditions (McCracken et al. ).
Sub-merged (co) cultures and ALI systems (Lenz et al. ; Panas et al. ) have been claimed to more closely mimicking the in vivo exposure scenarios compared to monocultures. In these systems, the toxicity and pro-inflammatory responses are significantly modulated by SiNPs, which represent the complexity of in vivo systems and need for the establishment of physiologically relevant in vitro cultures. However, at this moment, it is difficult to know whether these biological responses were influenced by SiNPs physico-chemical properties.
Napierska et al. ( b ) tested SiNPs with primary size 2, 16, 60, and 104 nm (dosed at 10 µg/cm 2 ) in a co-culture (A549 at the apical and EA. hy926 at the basolateral compartment) and observed increase in cytokines such as IL-6, IL-8, TNF-α, and MIP-1α only for 60 and 2 nm (except IL-8). When THP-1 were added to the co-culture, a significant increase in IL-8 and decrease in TNF-α were observed only for 2 nm. The expression of cytokines was also differentially regulated for 16 and 104 nm before and after THP-1 added, but the effects were stronger for 60 and 2 nm, particularly 60-nm NPs.
Co-cultures of lung cells are usually made with epithelial cells on the apical and endothelial cells on the basal compartment of a transwell membrane, with or without monocytes on top of the epithelial cells. In a co-culture (A 549 at the apical and ISO-HAS-1 at the basolateral compartment) exposed to 100-µg/ml 30-nm C-SiNPs (coated with or without surfactant), nearly a fivefold increase of IL-8 release for both forms of SiNPs was observed in both compartments (Kasper et al. ). When other epithelial cells were usedH441 cellsat the apical together with ISO-HAS-1 cells at the basolateral compartment, these C-SiNPs induced IL-8 were expressed in both compartments, while SiCAM-1 (6600 µg/ml) and IL-6 (at 6 and 60 µg/ml) were observed only in the apical part (Kasper et al. ). The same co-culture model was exposed to 100 µg/ml of S-SiNPs (15, 35, and 80 nm) and the authors noticed an increase of IL-8, TNF-α, and surfactant protein (SP-A1 and SP-A2) expression compared to the control. In addition, less IL-8 and surfactant protein expression, and more TNF-α were observed in the co-culture added with THP-1, notably the effect was the highest for 35 nm (Farcal et al. ).
A dose- (80640 µg/ml) dependent decrease in the viability and increase of apoptosis were observed in HepG2 cells in presence of SiNPs (7 and 20 nm), but a significant reduction was observed in normal human liver cells (LC-02) only at the unrealistic dose level of 640 µg/ml (Lu et al. ). In another study, SiNPs (1050 nm) induced a dose-dependent (100600 µg/ml) increase in LDH release in Caco-2 cells, but a significant release of LDH was observed only at the unrealistic dose of 600 µg/ml in human gastric epithelial cells (GES-1). Furthermore, exposure to these SiNPs (200 µg/ml for 48 h) also induced cell cycle arrest in S phase for GES cells and G2/M in Caco-2 cells (Yang et al. b ). In a porcine kidney cell line (LLC PK1) exposed to 20-nm SiNPs, a dose- (550 µg/ml) dependent increase in DCF fluorescence and MDA formation was observed, whereas human kidney cells (HK-2) showed little effects at 50 µg/ml (Passagne et al. ).
In vitro studies used cell lines of CNS-based cells, mainly neuron like cells. Although the data on the neurotoxic effects of SiNPs are very limited, studies above suggest that SiNPs can induce adverse effects including the markers of Alzheimers disease, when in direct contact with neuroblastoma cells.
Rat medulla tumor cells (PC12 cell line) incubated with the supernatant of 20 nm SiNPs-treated-microglial cells (250 µg/ml and 500 µg/ml for 24 h) did not show any effects compared to the control. Earlier in this study, no secretion of bio-mediators was observed in SiNPs-treated-microglial cells (Xue et al. ). In contrast to this study, PC12 cells exposed directly to SiNPs (25 nm; 25200 µg/ml for 24 h) showed increased uptake and a dose-dependent increase in the induction of autophagy (increase in LC-II and Beclin 1) and inhibition of PI3 K-Akt-mTOR signaling (Xie and Wu ). Yang et al. ( a ) showed that exposure to SiNPs (15 nm; 10 µg/ml for 24 h) induced pathological signs of Alzheimers disease such as altered expression of amyloid precursor protein (APP) and neprilysin, enhanced phosphorylation of tau at Ser262 and Ser396, and activation of glycogen syntheses kinase (GSK)-3β in human SK-N-SH and mouse neuro2a neuroblastoma cells.
Endothelial cells and platelets together play a key role in maintaining the vascular homeostasis (Rajendran et al. ). C-SiNPs induced oxidative stress and disturbed NO/NOO ratio, which resulted in the aggregation of platelets and endothelial dysfunction. This information is not available for other types of SiNPs. Furthermore, C-SiNPs mediated endothelial dysfunction resulted in pro-inflammatory signals via the secretion of cytokines and adhesion molecules. Together, these results suggest the potential of SiNPs to cause vascular thrombosis and atherosclerosis (Radomski et al. ). Furthermore, SiNPs caused hemolysis of RBCs in a size-, charge-, and porosity-dependent manner.
Exposure to different concentrations (12.5100 µg/ml) of S-SiNPs (58 nm) resulted in a dose-dependent increase in inflammatory mediators such as IL-1β, IL-8, and TNFα, intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule (VCAM-1), and monocyte chemoattractant protein-1 (MCP-1) in HUVEC cells (Guo et al. ). In the study of Corbalan et al. ( ), 10 µg/ml of 10-nm C-SiNPs induced extremely low NO/NOO ratio (~0.1) in HUVEC cells. Furthermore, free radical production, pro-inflammatory cytokines (IL-6 and IL-8), and NF-κB-binding activity were significantly increased in treated cells. In these two studies, increase in ROS was observed at all tested concentrations.
Maurer-jones et al. ( ) investigated the role of porosity of SiNPs on blood cell toxicity. M-SiNPs (25 nm) reduced the cell viability of red blood cells (RBCs) to 50% at the concentration of 270 µg/ml, while non-porous S-SiNPs of similar size required only 20 µg/ml to reach this level of cytotoxicity. In another study, 10% hemolysis (LC 10 ) of RBCs was observed at 36 µg/ml for S-SiNPs (115 nm), while M-SiNPs required 154 µg/ml to induce the same effects. For amine-coated counterparts, LC 10 were 97 and 30 µg/ml for S- and M-SiNPs, respectively (Yu et al. ).
Nemmar et al. ( ) showed that mouse blood platelets could aggregate after 3 min of incubation with 5 or 25 µg/ml of 50-nm C-SiNPs, while in the study of Jose Corbalan et al. ( ), such aggregation was observed in 15 min (10 µg/ml of 10 nm C-SiNPs). The latter study also showed fourfold reduction of the nitric oxide (NO)/peroxynitrite (ONOO ) ratio compared to non-treated platelets.
Several studies suggest that NPs, when inhaled or ingested, can translocate across barriers (such as airblood) of the body, enter the circulation, and interact with the cardiovascular system. In this section, we summarized the studies that report the effects of SiNPs on blood and endothelial cells.
SiNPs, particularly S-SiNPs induced autophagy mainly via oxidative stress-mediated upregulation of autophagy-related genes and differential regulation of Akt/mTOR signaling. Similar to cytotoxicity, 25 µg/ml appeared to be the lowest exposure concentration at which SiNPs exhibited significant effects. Furthermore, induced autophagy is correlated to cytotoxicity, suggesting that exposure to SiNPs caused irreversible (serious) cellular damage and resulted in autophagic cell death. Besides autophagy induction, lysosomal and autophagy dysfunction could be a potential mechanism of NPs toxicity (Stern et al. ), which has, however, not been investigated for SiNPs.
SiNPs (413 nm, 62.5 µg/ml) induced a time-dependent (24, 48, and 72 h) reduction in cell viability and increase in oxidative stress (DCF fluorescence and GSH depletion) in the lung fibroblast cell line MRC-5. Compared to control, a significant increase of autophagic vacuoles and LC-3 II/LC3-I ratio was also observed in a time-dependent manner (Voicu et al. ). A549 cells, when exposed to 100 and µg/ml of 20-nm SiNPs, showed threefold and fivefold increase in MDC fluorescence, respectively. In addition, autophagy genes such as ATG-12 and BECN were significantly upregulated (30- and 50-fold, respectively) along with increased production of ROS in cells dosed with µg/ml (Nowak et al. ).
The same S-SiNPs (62 nm) were used in three studies to investigate the induction of autophagy. Along with the dose-dependent (25100 µg/ml) increase in ROS production, increase in autophagy bio-marker-microtubule-associated protein 1A/1B-light chain 3 (LC3) and monodansylcadaverine (MDC) labelled autophagic vacuoles were detected in HepG2 cells treated with 62 nm S-SiNPs. In addition, transmission electron microscopy (TEM) images revealed that autophagosomes and autolysosomes induced in the presence of SiNPs (Yu et al. ). The same S-SiNPs (62 nm) induced increase of LC3-II/LC3-I ratio and decrease of p-mTOR/mTOR, p-P13 K/P13 K and p-Akt/Akt in HUVEC cells in a dose-dependent (25100 µg/ml) manner (Duan et al. b , a , b ). The results of Guo et al. ( ) also suggest that 50 µg/ml of S-SiNPs (58 nm) could induce autophagy via MAPK/Bcl-2 and PI3K/Akt/mTOR signaling in HUVECs.
The main cells used to study immune responses to SiNPs were innate cells such as monocytes and macrophages. Therefore, the identified in vitro studies only address a very limited part of the immune system, essentially pro-inflammatory responses and potential phagocytosis. Furthermore, the data on immune responses and oxidative stress are very limited and, therefore, no firm conclusions can be made. SiNPs, not only induced stronger pro-inflammatory responses compared to sub-micron and micron sized particles but also size-specific effects within the nano-range in immune cells are observed. Besides size, shape and porosity seem to influence the phagocytosis of SiNPs.
Uemura et al. ( ) showed that SiNPs (10 and 50 nm) caused dose-dependent (6.25100 µg/ml) increase in the production of TNF-α and decrease of IL-6 in RAW.264.7 macrophages, while their amine surface-modified counterparts did not. Furthermore, 300 and nm micron-sized particles (both bare and amine modified) also showed a dose-dependent decrease of IL-6. Notably, the effects were stronger for 50 nm compared to other particles. The same cell line was utilized by Yu et al. ( ) to investigate phagocytosis using inductively coupled plasma mass spectroscopy (ICP-MS), and they found that S-SiNPs (25 nm) were phagocytosed at least ten times more than M-SiNPs of same size and high aspect ratio SiNPs (AR 2, 4, and 8). In the study by Napierska et al. ( b ), THP-1 cells dosed with 5 µg/cm 2 S-SiNPs (2 nm) showed a significant increase of IL-8, TNF-α, and macrophage inflammatory protein (MIP)-1α, while only a non-significant increase in MIP-1α expression was observed for 16 and 104 nm S-SiNPs.
Hara et al. ( ) exposed THP-1-derived macrophages to 100 µg/ml of SiNPs (30 nm) for 6 h and found a significant increase in interleukin-1-beta (IL-1β), ROS production, and SiNP uptake via phagocytosis. In the study of Choi et al. ( ), larger sized SiNPs (150200 nm) were effectively phagocytosed by primary rat microglial cells after 24 h of exposure to different concentrations (0.7.28 µg/ml). A significant increase in ROS, reactive nitrogen species (RNS) and IL-1β was detected at all concentrations.
NPs entering the body will most probably interact with immune cells, as they are the first line of defence in human body. In this section, we presented the immune responses induced by SiNPs in different cell lines.
C-SiNPs and S-SiNPs induced genotoxicity in human tumor cell lines (lung, kidney, skin, and gastro-intestinal systems) and the amplitude of the effect negatively correlated with the size of the NPs. DNA strand breaks were observed at low concentrations (2.510 µg/ml), particularly in skin-derived cell lines. The genotoxic effects of C- and S-SiNPs were mainly associated with the induction of oxidative stress, while such information is very limited for other types (Py- and Pr-SiNPs). One study indicated that Py-SiNPs induce DNA damage without the generation of ROS, suggesting that other mechanisms such as direct DNA damage might be involved (Magdolenova et al. ). However, it is very difficult to judge whether such genotoxic effect is direct or indirect, since the cellular uptake and subcellular localization of SiNPs are not often reported. Furthermore, several factors such as SiNP properties, cell type, and exposure scenarios (such as concentrations, assays, and endpoints) may influence the outcomes (Magdolenova et al. ), making the comparison difficult between studies and indicating an urgent need for the standardization of genotoxicity studies.
At any tested concentrations (1100 µg/ml for 72 h), SiNPs (1025, 530, 35, 15, 80, and 90 nm) neither induced cytotoxicity nor micronuclei in immortalized Balb/3T3 fibroblasts (Uboldi et al. ). In another study, Pr-SiNPs (NM-200 and NM-201) and Py-SiNPs (NM-202 and NM-203) with primary size between 14.5 and 16 nm did not induce any micronuclei (cytokinesis block micronucleus assay) in human peripheral lymphocytes exposed to different concentrations (200 µg/ml) over 24 h (Tavares et al. ). It is also worthy to note that, in the latter study, the positive control used did not differ from control conditions.
Exposure to C-SiNPs (~7 nm) resulted in positive genotoxic effects (Lymphoma assay) in mouse lymphoma cells treated with 100 and 150 µg/ml for 4 h (Demir and Castranova ). SiNPs (12 nm) induced DNA strand breaks in RAW 264.7 at 200 and 400 µg/ml, but the induction of micronuclei was noticed only at 400 µg/ml. The particles were internalized in vesicles and in the nucleus (Hashimoto and Imazato ).
M-SiNPs (100 nm) induced a significant increase in phosphorylated-ɣH2AX-foci in HT-29 cells treated with a dose of 10 µg/ml for 24 h (Sergent et al. ). In the human embryonic kidney cell line (HEK293), 579 genes were upregulated and genes were downregulated after 24 h of exposure (100 µg/ml) to 100 nm M-SiNPs (Zhang et al. ). In another study, 15-nm C-SiNPs induced a significant increase in DNA strand breaks (comet assay) in chinese hamster cells (V79) and A549 cells at 100 µg/ml (24 h), but, for 55 nm, this effect was observed only in A549 cells (Maser et al. ). A significant increase in DNA tail length (comet assay) was observed in HaCaT cells treated with 30 µg/ml (24 h) of 70 nm SiNPs (Nabeshi et al. b ).
Genotoxicity of Py-SNP (20 and 2570 nm), Pr-SNP (20 nm), and C-SNP (15 and 4080 nm) SiNPs were studied in Chinese hamster lung fibroblasts. Py-SiNPs (20 nm) induced a significant increase in DNA strand breaks at 66 µg/ml (24 h), while C-SiNPs (15 nm) showed a similar effect only at 252 µg/ml. Neither of these SiNPs did induce ROS. SiNPs in the size range 2580 nm exerted no or little genotoxicity (Guichard et al. ).
S-SiNPs (62 nm) induced increase in DCF fluorescence and decrease in superoxide dismutase (SOD) and glutathione peroxidase (GSH-Px) in HUVECs in a dose-dependent manner (25100 µg/ml for 24 h). Oxidative stress was linked to cell cycle arrest at G2/M checkpoint (upregulation of chk 1 and downregulation of Cdc25c, Cyclin B1, and Cdc2) and increase in apoptosis (Duan et al. a ). In the study by Li et al. ( ), a size-dependent (19, 43, and 68 nm) increase in oxidative stress (DCF fluorescence) and cell cycle arrest in S and G2/M was observed in HepG2 cells exposed to 100 µg/ml of C-SiNPs. Cell cycle arrest in G2/M phase along with the increase in ROS was also noticed in human hepatic cell line (LC-02) treated with S-SiNPs (50 nm) in a dose-dependent manner (50200 µg/ml for 24 h) (Wang et al. ).
The viability of human Caucasian colon adenocarcinoma (Caco-2) cells dropped to 40% when exposed to 15 nm C-SiNPs (64 µg/ml for 24 h), and, at this same concentration, nearly a threefold increase in micronuclei formation, fivefold increase in histone phosphorylation (ɣH2AX), and a significant increase in DCF fluorescence were observed. The particles were localized within lysosomes and endocytic compartments, but not in the nucleus. 55 nm C-SiNPs did not induce any of these effects at the same concentration (Tarantini et al. b ).
Exposure to SiNPs (15, 30, and 100 nm) resulted in a size- and dose- (2.510 µg/ml for 24 h) dependent increase in 8-hydroxy-2-deoxyguanosine levels (8-OH-dG), phosphorylation of histone on serine-139 (ɣH2AX), and DNA strand breaks (comet) in human keratinocytes (HaCaT) (Gong et al. ). Nabeshi et al. ( a ) also demostrated that exposure to SiNPs (70 nm; 1090 µg/ml for 24 h) resulted in the increase of oxidative DNA damage (8-OH-dG levels) in HaCaT cells. SiNPs were taken up via actin-mediated endocytosis. Micron-sized particles used in these studies showed no or little effects.
In this section, we presented studies on genotoxic effects of SiNPs as it is used as another major endpoint to characterize hazard of NMs. Direct interaction with DNA, oxidative DNA damage, depletion of anti-oxidants, cell cycle arrest, and abnormal expression of genes have been identified as potential mechanisms of NP mediated (geno)toxicity (Donaldson et al. ).
Furthermore, some authors used very high concentrations that may cause overloading of cells and modify the nature of NPcell interactions (Wittmaack ). In these cases, it is difficult to evaluate whether the observed effects are physiologically relevant. Although it is challenging, we consider a dose of 384 µg/cm 2 or higher as irrelevant to human inhalation exposure for amorphous silica, based on the estimation that can be derived from the occupational exposure levels (OELs) (Fig. ).
Cytotoxicity of SiNPs was investigated using different cell lines and incubation times, making the comparison between studies difficult. However, from Table , it is clear that all types of SiNPs induced cytotoxicity. Significant (compared to untreated cells) cytotoxic effects were observed only at or above the concentration of 25 µg/ml. Furthermore, it can be clearly seen that SiNPs induced oxidative stress and mediated apoptosis mainly via the intrinsic or mitochondrial pathway (caspase-dependent pathway) in a size- and dose-dependent manner. ROS-mediated toxicity is believed to be an important mechanism of NP toxicity including SiNPs (Manke et al. ). Nevertheless, Py- and S-SiNPs caused cytotoxicity without measurable levels of ROS production. It was demonstrated that the disturbance of membrane integrity due to direct cell-membrane interaction might be another possible mechanism of NP cytotoxicity (Fröhlich et al. ; Thomassen et al. ). However, neither of these studies did substantiate these observations and, therefore, SiNPs cytotoxic effects in the absence of oxidative stress remain poorly understood.
Py-SiNPs (12 and 40 nm) induced a significant size and dose- (31.3, 93.8, and 156.3 µg/cm 2 culture well) dependent cytotoxicity (LDH, Sulphorodhamine B assay (SRB) and water-soluble tetrazolium-1(WST-1)) in human colon carcinoma cell line (HT29), while no induction of ROS (DCFH-DA) was observed (Gehrke et al. ). In the study by Napierska et al. ( a ), 50 μg/ml (24 h) of 16 nm iron-doped S-SiNPs and pure S-SiNPs induced strong cytotoxicity (MTT and LDH) in a human endothelial cell line (EA.hy926), but a significant increase in oxidative stress markers [GSH depletion, malondialdehyde (MDA formation), induction of heme oxygenase-1, glutathione reductase, and NADPH oxidase-1] was observed only for iron-doped SiNPs.
SiNPs (20 and 80 nm) induced P53-mediated apoptosis in human fetal lung fibroblasts (HFL-1). At the dose of 500 µg/ml, 20 nm SiNPs induced a threefold increase in DCF fluorescence compared to 80 nm. In addition, increased expression of P53, upregulation of cytochrome C (CytC) and caspase 9, and downregulation of anti-apoptotic protein B-cell lymphoma 2 (bcl2) was observed in cells treated with µg/ml for 48 h (Xu et al. ). Another study with lung fibroblasts also showed that SiNPs (20 nm) could reduce cell viability (MTT) by inducing apoptotic cell death (fluorescence microscopy) in a dose-dependent manner (250 µg/ml for 48 h) (Zhang et al. ). Athinarayanan et al. ( ) isolated SiNPs (1050 nm) from commercial food products processed with food additive silica (E551) and exposed human lung fibroblasts (WI-38 cell line) with increasing doses (25400 µg/ml). After 24 h, they observed cytotoxicity (MTT) in a dose-dependent manner and ROS production (DCFH-DA) at 50 µg/ml.
Duan et al. ( a ) showed that S-SiNPs (62 nm) induced time- (6, 12, and 24 h) and dose-dependent (25100 µg/ml) reduction in cell viability (assessed by 3-(4,5-dimethylthiazol-2-Yl)-2,5-diphenyltetrazolium bromide, i.e., MTT), loss of membrane integrity (lactate dehydrogenase (LDH) release) and apoptosis (Annexin V/PI staining) in human umbilical vein endothelial cells (HUVECs). Apoptosis was also induced in lung (A549) and skin epithelial cells (A431) treated with Pr-SiNPs (15 nm). A dose-dependent increase (25200 µg/ml for 72 h) in cytotoxicity (MTT and LDH), ROS production (assessed by dichlorodihydrofluorescein assay, i.e., DCFH-DA), lipid peroxidation (measurement of malondialdehyde, i.e., MDA), and apoptosis (caspase 3 and 9 activity) was observed in both cell lines. The lung cells showed, in general, a slightly higher toxic response compared to skin cells (Ahamed ).
The in vivo toxicity studies have been carried out using rats and mice, and exposure through various routes of administration. In general, short-term exposure to SiNPs induced adverse effects in the lungs, kidneys, liver, and brain. SiNPs administered mainly via inhalation, ingestion, and intravenous routes were majorly distributed in the liver, lungs, spleen, and kidneys, and in the brain of the intranasally exposed rats. Most of the administered SiNPs were excreted via feces and to a lesser extent via urine, in a size- and shape-dependent manner. However, in most cases, the administered doses were very high compared to relevant human inhalation (Barsan ) and ingestion (Winkler et al. ; Dekkers et al. ) exposures to amorphous silica. Some studies showed accumulation of SiNPs in organs such as the liver, but such accumulation was not associated with any major effects. However, long-term effects of accumulated SiNPs were not studied. It is surprising that in contrast to acute studies, no toxicity (local or systemic) was observed in chronic oral and dermal exposure studies, regardless of the size of SiNPs and the high doses used. Moreover, surface-modified SiNPs showed a significant increase in absorption by the GI tract compared to bare SiNPs. Dosing SiNPs via i.v. showed in all studies some sign of damage/toxicity. No clear type-or size-dependent effect can be identified from the set of reviewed studies. Furthermore, where genotoxicity was studied, it was not clear whether SiNPs were taken up by the target cells or were able to reach the tissue examined, which is obviously needed to exert a direct genotoxic effect.
Zhuravskii et al. ( ) demonstrated that intravenously injected (70 mg/kg bw) Py-SiNPs (13 ± 5 nm) distributed and persisted in the liver of Wistar rats at 60-day post-administration. The authors also indicated that the administered SiNPs induced fibrosis and liver tissue remodeling by noticing the increase in blood ALT, and presence of mast cells, connective tissues, and foreign body-type granulomas in the liver.
A significant increase in the levels of kidney injury bio-markers such as BUN and CREA and morphological changes associated with renal interstitial fibrosis was noticed in BALB/c mice intraperitoneally treated with 198-nm M-SiNPs (150, 300, or 600 mg/kg bw for 2 or 12 days). There was also a significant increase in the levels of fibronectin (FN), TGF-β and ICAM-1, and nuclear translocation of p65 in 300- or 600-mg/kg treated mice (Chen et al. ).
At 24 h of post-intravenous administration (30 mg/kg bw), TEM imaging of the BALB/c mice tissues showed that 70-nm SiNPs were localized in the regional lymph nodes, cerebral cortex, hippocampus, and the cytoplasm, and nucleus of hepatocytes (liver). Using in vivo imaging, the authors also found that the SiNPs were distributed near the liver immediately after the administration (20 min) and moved near the intestinal tract over time (6 h). Micron-sized silica particles (300 and nm) used in these studies were accumulated mainly around the gall bladder (Nabeshi et al. b ).
Yu et al. ( ) estimated the i.v. median lethal dose (LD 50 ) for 64 nm C-SiNPs as 262 mg/kg bw in ICR mice using the Dixons Up and Down method. After 14-day post-administration, intravenously injected C-SiNPs (64 nm, 177.5 mg/kg bw) induced a significant increase in liver injury bio-markers (LDH, AST, and ALT). Histopathological analysis of target organs showed that the number of megakaryocytes in the spleen was significantly increased, and pulmonary hyperemia and interstitial thickening were observed in the lungs. Nearly 35% of the injected SiNPs were distributed in the spleen, 12.5% in the liver, and 2% in the lung of SNP-treated mice.
In the study of Duan et al. ( a , b ), ICR mice were intravenously injected with 29.5, 103.5, or 177.5 mg/kg bw of 62 nm S-SiNPs. After 14 days, induction of autophagic vacuoles and mitochondrial rupturing in the heart tissues was observed using TEM. In addition, increase in LC3 positive staining in the heart tissues was detected at 103.5 and 177.5 mg/kg bw, and the expression of ICAM-1 and VCAM-1 was significantly decreased at 177.5 mg/kg. Huang et al. ( ) injected the mice intravenously with mesoporous rods (pure or PEGylated) at a dose of 20 mg/kg bw and found a dysfunction in biliary excretion and glomerular filtration. Blood analysis showed that there was a significant increase in total bilirubin (TBIL), blood urea nitrogen (BUN), and creatinine (CREA). The mesoporous rods mainly distributed in the lung, liver, and spleen, but PEGylation significantly reduced their distribution in these organs. In addition, short rods cleared via urine and feces more rapidly than long rods.
In another study, 50 nm C-SiNPs (0.25 mg/kg bw) induced a significant increase in blood parameters such as leukocyte number, creatinine kinase (CK), ALT, AST, and LDH release in male tuck ordinary mice at 24-h post-intraperitoneal injection. Oxidative stress bio-markers such as MDA formation, SOD, and catalase were significantly increased in the lungs, liver, kidney, and brain of SiNPs-treated mice. Cytokines such as IL-6, IL-1β, and TNF-α were significantly increased in the lung and differentially expressed in other organs. Furthermore, all organs showed a significant DNA damage compared to saline treated mice and the extent of damage was in the order heart > kidney > lung > liver > brain (Nemmar et al. ).
In a 48-h study, exposure to 15-nm SiNPs (dose of 50 mg/kg bw by intravenous injection) resulted in a significant increase of CD68-positive Kupffer cells (KCs), WBCs, lymphocytes, monocytes, neutrophils, and TNF-α in the serum of male SD rats. There was a decrease in GSH activity and elevation in MDA levels in the liver of treated mice. Furthermore, bio-markers of liver dysfunction such as lactate, phosphorylcholine, sn-glycero-3-phosphocholine, tyrosine, phenylalanine, and lysine were increased in blood, while the levels of succinate, glucose, and glycine were significantly decreased in mice exposed to SiNPs (Chen et al. ).
The skin of SD rats was repeatedly exposed to different doses (500, mg, or mg/kg bw) of l-arginine coated C-SiNPs (20 nm) for 90 days (6-h exposure/day). The repeated exposure neither induced gross changes in the skin nor in any organs. In addition, hematological and blood biochemical parameters did not change in SiNPs-treated mice compared to controls (Ryu et al. ). In another study, Shim et al. ( ) showed that C-SINPS (20 and 100 nm) did not induce toxicity in the right and left brain or distribution in the cerebellum, hippocampus, or striatum of the dermally exposed rats ( or mg/kgbw daily for 90 days). In contrast to these studies, a significant increase in apoptosis (TUNEL positive cells) was observed in the skin of the BALB/c mice topically exposed to 70-nm SiNPs (250 mg/ear/day for 28 days). The author reported that the SiNPs were found not only in the skin but also in the regional lymph nodes, cerebral cortex, hippocampus, and in the liver (Nabeshi et al. b ).
Intratracheally administered (2 mg/kg bw; 15 times, once every 3 days) S-SiNPs (58 nm) resulted in reproductive toxicity in C57 mice via enhanced ROS production and cell cycle arrest (G0/G1 phase) in the testicular tissues, and by decreasing the number of mature sperms and primary spermatocytes (Zhang et al. ). In another study, repeated intracheal instillation of S-SiNPs (58 nm; 0, 7, 21, and 35 mg/kg bw; 5 times, once every 3 days) induced local (macrophage activation in lung, liver, and spleen) and systemic inflammation (Increase in serum IL-8, TNF-α, and IL-6) in BALB/c mice. In addition, SiNPs were also detected in the lysosomes of macrophages in lung, liver, and heart tissues (Yang et al. ).
Male Wistar rats were intranasally instilled with 150 µg SiNPs (10 and 80 nm) repeatedly for 30 days. After 30 days, a size-dependent increase in the levels of hydrogen peroxide (H 2 O 2 ), MDA formation, TNF-α, IL-1β, MCP-1, and NF-ĸβ was observed in the frontal cortex, corpus striatum, and hippocampus of the brain. A similar quantity of Si was detected in all these three regions. In addition, a significant decrease in GSH levels was observed in these tissues (Parveen et al. ). The latter study shows the potential translocation of (mainly small) SiNP from the nose to the brain, but it has to be noted that the mice received a high dose during 30 days.
A study compared the toxicity and biodistribution of pristine SiNPs (19 nm) and aged paints containing SiNPs in BALB/c mice. The suspensions (20 µg/aspiration) were oropharyngeally aspirated once a week for 5 weeks, and mice were sacrificed either at 2- or at 28-day post-final aspiration treatment. Pristine SiNPs were significantly distributed in the lungs and liver, while only a low amount of Si was detected in the liver of paint-exposed mice. No signs of toxicity were observed, except a slight inflammation (a slight increase in macrophages and neutrophils together with an increase of IL-1β) with pristine SiNPs (Smulders et al. ).
A size- and dose-dependent increase in the distribution of SiNPs was observed in the serum and heart of male Wistar rats intratracheally instilled (2, 5, or 10 mg/kg bw/day for 16 days) with SiNPs (30, 60, 90, and 300 nm). Blood parameters (WBCs and platelets), inflammatory bio-markers (TNF-α, IL-1β, and IL-6), and oxidative stress bio-markers (ROS and MDA formation) were significant increased, while NO, NOS, and eNOS were significantly decreased in the serum of the treated mice (Du et al. ).
In the study of Morris et al. ( ), C57BL/6 mice were intratracheally instilled with 4- or 20-mg S-SiNPs/kg bw and significant effects were observed only at the dose of 20 mg/kg. Twenty hours after instillation, approximately 20- and 10-fold higher cell number was observed in the bronchoalveolar lavage (BAL) of mice treated with the bare and amine-coated SiNPs, respectively, compared to control mice; neutrophils were also increased about 30- and 20-fold, respectively.
In male Wistar rats that exposed to 1015 nm SiNPs (oral gavage; 333.3 mg/kg bw/day for 5 days), histopathological analysis revealed gross tissue damage in kidney (cell swelling and necrosis), lung (interstitial pneumonia and bronchopneumonia), and in the testis (congestion, reduction of spermatogenesis and edema). In addition, blood biochemical parameters such as albumin, cholesterol, triglycerides, total proteins, urea, high-density lipoprotein (HDL), and low-density lipoprotein (LDL), as well as alkaline phosphatase (ALP) and aspartate aminotransferase (AST) activities were significantly increased in treated mice (Hassankhani et al. ).
Yoshida et al. ( ) found an increased intestinal absorption of carboxyl (twofold) and amine-coated SiNPs (1.5-fold) compared to uncoated SiNPs (70 nm) in orally exposed BALB/c mice (2.5 mg/mouse/day for 28 days). In this study, the whole small intestine was processed using the everted gut sac method and a significant amount of Si was detected by ICP-MS. However, no signs of systemic toxicity were observed for any of these SiNPs.
Zande et al. ( ) investigated the sub-acute and sub-chronic toxicity of Py-SiNPs (7 nm) and NM-202 (1025 nm). In the sub-acute part, male SD rats were orally (via food) exposed to 100, , or mg Py-SiNPs/kg bw/day for 28 days, while in the sub-chronic part, the rats were exposed only to the highest dose ( mg/kg bw/day) repeatedly for 84 days. ICP-MS analysis of target organs showed that NM-202 was significantly distributed in the lung, kidney, and spleen. Although no Si was detected in the liver, histopathological analysis and gene expression studies revealed that fibrosis was induced in the liver after 84 days of exposure. Si was found in the spleen of Py-SiNPs exposed rats only at the end of chronic exposure. The examined parameters such as blood biochemistry, antibody levels (IgG and IgM), and cytokines indicated no systemic toxicity in any of the Py- or NM-202 treated rats.
SD rats were administered (via oral gavage) with low doses (5, 10 or 20 mg/kgbw) of Pr-SiNPs (NM 200 and 201) and Py-SiNPs (NM 202 and 203) at 0, 24 and 45 h and sacrificed at 3 h after the last administration. No significant genotoxicity was observed in cells extracted from different organs (duodenum, colon, blood, kidney, liver, and spleen), and no pathological conditions were recorded in any of these organs (Tarantini et al. a ).
C-SiNPs (12 nm) neither induce abnormal changes in blood biochemical and hematological parameters nor accumulated in any organs of the acutely (489.8, 979.5, or mg/kg bw/day, during 14 days) or sub-chronically (244.9, 489.8, or 975.9 mg/kg bw/day, during 13 weeks) exposed (via gavage) SD rats (Yun et al. ). In the study by Kim et al. ( c ), C-SiNPs (20 and 100 nm) did not induce any significant changes (compared to control) in clinical signs, blood biochemical, hematological, and histopathological analysis in sub-chronically exposed Crl: CD (SD) specific pathogen-free rats (500, , or mg/kg bw/day for 90 days).
Kim et al. ( b ) exposed C57BL/6 mice (oral gavage; 750 mg/kg bw) to C-SiNPs (20 and 100 nm) that had been modified with or without l-arginine to determine the influence of surface charge on immunotoxicity in vivo. After 14 days, the author noticed a size-dependent decrease in WBCs cell count and cytokines (in blood), and reduced proliferation of B cells and T cells (from spleen) only for uncoated SiNPs.
You will get efficient and thoughtful service from Hebei Silicon Research Electronic Materials Co., L.
In the study of Li et al. ( ), M-SiNPs with different aspect ratios (AR, spherical 83 nm with AR 1, short rods with AR 1.75 and long rods with AR 5) were administered in mice (40 mg/kg bw via gavage). After 7 days, a very high amount of Si was detected only in the liver of spherical SiNPs exposed mice. Urinary excretion, intestinal absorption, and organ distribution of SiNPs were decreased with increasing aspect ratio. Furthermore, increase in renal damage such as hemorrhage, vascular congestion, and renal tubular necrosis with increasing aspect ratio was observed after 14 days of administration.
Lee et al. ( ) investigated the tissue distribution and excretion profiles in SpragueDawley (SD) rats orally administered with a very high dose of 500 or -mg/kg bw of C-SiNPs (20 and 100 nm). Silicon (Si) levels were significantly elevated in liver, kidney, lung, and spleen at 6-h post-administration, while no such increase was noticed in brain, ovaries or testes, esophagus, stomach, and intestine even after 7 days of post-administration. Nearly 7580% of administered SiNPs were excreted via urine and 78% via feces. The author also noticed that 20 nm SiNPs excreted faster than the 100 nm SiNPs. In another study, C-SiNPs (12 nm) did not induce any toxicity after 14 days in SD-rats administered with a single oral dose ( or mg/kg bw) (Yun et al. ).
Growing production and use of SiNPs increase the risk of human exposures. Available toxicity studies mainly focused on effects after exposure via inhalation or skin (occupational exposure), ingestion (food additive), or nano-therapeutics (parenteral exposure). In vitro and in vivo studies demonstrated that SiNPs can induce adverse effects, but multiple inconsistencies were found in the collected toxicity data set. As small differences in the physico-chemical properties of SiNPs could contribute to significant variation in the toxicity (Napierska et al. ), it is critical to discriminate how these variations influence toxicity. It is also worthy to note that the choice of cell type, culture system, assay conditions (Fede et al. ; Geys et al. ; Hayashi et al. ), and exposure route influence the toxic responses to SiNPs. However, the nature of these variations remains unclear, which continues to limit our understanding of SiNPs toxicity and hampers their hazard assessment.
In this review, a basic set of physico-chemical characteristics of primary SiNPs such as primary size, shape, crystallinity, and chemical composition (or) purity was set as inclusion criteria. Most studies associated the toxic endpoints to SiNPs size, while only a few related the toxicity to SiNPs porosity, shape, surface charge, and surface chemistry. Furthermore, hydrodynamic diameter is another widely reported property in these studies and the results suggest that, in most cases, SiNPs are often aggregated/agglomerated (AA) in cell culture media. Conversely, AA formation is also very likely in real-world matrices such as in air (Kim et al. a), water, and in commercial products such as food (Dekkers et al. ). Such AA formation not only substantially alters the overall characteristics (such as size, shape and surface topology) but also potentially influences the biological outcomes (Luyts et al. ; Drescher et al. ). Although some progress has been made in the characterization of SiNPs AA in recent years (De Temmerman et al. ), their biological effects remain poorly understood.
Synthetic amorphous SiNPs are produced via different methods such as thermal (pyrogenic) or wet route (colloidal, precipitated and gel). The physico-chemical characteristics of SiNPs produced by these methods differ (Fruijtier-Pölloth ; Napierska et al. ) and may influence the biological outcomes. In this review, nearly 70% of the papers clearly reported the synthesis method. Among those 70%, about 80% reported the toxicity of wet method based SiNPs and only approximately 20% on the other types (such as pyrogenic and mesoporous). A general overview of toxic effects induced by different types of SiNPs is presented in Table . Cytotoxicity and genotoxicity induced by S- and C-SiNPs are strongly correlated with the induction of oxidative stress. For Pr-SiNPs, oxidative stress was associated with cytotoxicity but not genotoxicity. Interestingly, Py-SiNPs caused cytotoxicity, mostly without the generation of oxidative stress. In addition, recent studies showed that Py-SiNPs are biologically more reactive than C-SiNPs (Zhang et al. ) and Pr-SiNPs (Di Cristo et al. ) of same composition and size. It is known that C-, S-, and Pr-SiNPs are hydrophilic in nature, while Py-SiNPs are hydrophobic due to the de-hydroxylation of surface OH groups during the production process (Napierska et al. ), and such difference in surface chemistries might contribute to different biological activities. However, more systematic studies are required to verify these differences depending on production processes.
A fundamental question was raised by Napierska et al. (): do amorphous SiNPs induce biological responses similar to crystalline silica? Oxidative DNA and membrane damage have been reported as the major toxic mechanisms involved in the health effects of micron-sized crystalline silica, which is also observed for C- and Pr-SiNPs. The latter two possess silanols on the surface while Py-SiNPs mostly contain siloxanes (Napierska et al. ). In addition, recent studies showed that the presence of surface moieties such as silanols is more correlated with crystalline silica toxicity than crystallinity (Zhang et al. ; Turci et al. ). Thus, silanols appear to be a common surface feature in C-SiNPs, Pr-SiNPs, and crystalline silica, which might contribute to a similar toxic activity. Furthermore, toxicity elicited by Py-SiNPs appears to be oxidative stress-independent, indicating that Py-SiNPs induce adverse effects via other mechanisms (Gehrke et al. ; Napierska et al. a). More studies are required to verify these hypotheses.
The exposure route obviously influences the in vivo absorption, biodistribution, and toxicity of SiNPs. For instance, after oropharyngeal aspiration or intratracheal instillation, the lungs are clearly the main target (Smulders et al. ), while after ingestion (Zande et al. ) or iv injection (Yu et al. ), the liver and/or the spleen were the targets. SiNPs accumulated in the brain of intranasally exposed rats (Parveen et al. ) but did not after oral exposure (Shim et al. ). In addition to the exposure route, physico-chemical properties such as size and shape clearly influence the clearance, distribution, and toxicity (Huang et al. ; Li et al. ).
In repeated (oral and dermal) dose studies (28 days), SiNPs did not induce any local or systemic toxicity even in (very) highly dosed rats. A recent study demonstrated that mild and highly dosed rats excreted most of the orally administered SiNPs via the feces (Yun et al. ). In addition, Van der Zande et al. () demonstrated the in vitro gelation of SiNPs with increasing concentrations, which might reduce the gastro-intestinal absorption in highly dosed animals and increase excretion via feces. Therefore, the use of (low) realistic exposure doses appears more appropriate, since the toxicokinetics may depend on the level of exposure (Paek et al. ). Moreover, information on the physico-chemical properties of ingested and digested SiNPs is lacking, representing a huge knowledge gap in the risk assessment of SiNPs in food (Dekkers et al. ).
Correlation between in vitro and in vivo effects is an indication that specific cells or tissues are potential targets for SiNPs toxicity. The results of in vitro and in vivo experiments (included within the same study) suggest that exposure to SiNPs could induce Kupffer cell mediated liver injury (Chen et al. ), kidney injury via the activation of NF-kB signaling pathways (Chen et al. ), and endothelial dysfunction via autophagy (Duan et al. a). Several in vitro studies showed that SiNPs caused DNA double-strand breaks in a wide range of immortalized cell lines at low doses, but no such genotoxic effects were observed, even in animals at high doses. In addition, SiNPs did not induce micronuclei either in vitro or in vivo.
Accurate in vitro dosimetry is an important yet complex aspect of nanotoxicology (Lison et al. ). In many studies reviewed here, authors often tested high in vitro exposure doses. Considering the exposure dose as nominal dose may be appropriate only for well dispersed and stable SiNP suspensions (Lison et al. ). However, Py-SiNPs, for instance, often re-agglomerate in culture media with an effective density lower than the material density (Deloid et al. ), thereby potentially affecting the dose reaching the cells (Cohen et al. ). Thus, a realistic estimation of the delivered dose is necessary to compare the biological effects between studies and to establish good in vitro and in vivo correlations (Pal et al. ).
Immune responses after the administration of SiNPs are very crucial, as they can induce cascades of events by the secretion of cytokines, which may be harmful or beneficial. The data collected in this review show that the smaller the size, the stronger the pro-inflammatory effect. In addition to the size, the surface charge appears to play a role, since less negative charges seem to suppress the immune response. Porosity is another crucial factor influencing blood biocompatibility, i.e., the more porous the SiNPs, the less hemolysis of RBCs (Maurer-jones et al. ). Furthermore, SiNPs aspect ratio was shown to closely relate to in vivo organ retention and clearance (Huang et al. ). Therefore, the size, surface area, porosity, and geometry (shape) appear to be the key parameters for designing less-toxic and less-inflammagenic SiNPs, for bio-medical applications for instance.
In conclusion, SiNPs generally exhibit acute toxic effects in vitro and in vivo. The data on chronic effects of SiNPs exposure are rather conflicting with the acute effects and are still insufficient to draw firm conclusions. No concrete data were found to conclude whether amorphous SiNPs induce fibrosis like micrometric crystalline silica. Moreover, translation to human health effects is impossible at this moment due to the lack of realistic exposure and epidemiological data. Surface moieties (such as silanols, silanolates, and siloxanes) of SiNPs are found to be production process-specific and seem to be the key determinants of SiNPs toxicity. AA formation for some SiNPs is very dynamic in physiological media, but it is still unclear how it contributes to the hazard characterization. Therefore, the design of safe(r) SiNPs for food, medical, and other applications will only be possible when physico-chemical characteristics can be unambiguously linked to toxicity. Furthermore, detailed investigations on the SNP bioaccumulation/bioavailability and their long-term consequences in vivo are required for a safer use.
Until recently, toxicological research into silica particles focused mainly on "natural" crystalline silica particles of 0.5 to 10 μm (coarse or fine particles). This research was/is fed by the clear association of occupational inhalation exposure and severe health effects, mainly on the respiratory system. The typical lung reaction induced by chronic inhalation of crystalline silica is silicosis, a generally progressive fibrotic lung disease (pneumoconiosis), exemplified by the development of silicotic nodules composed of silica particles surrounded by whorled collagen in concentric layers, with macrophages, lymphocytes, and fibroblasts in the periphery. Epidemiologic studies have found that silicosis may develop or progress even after occupational exposure has ended; therefore, above a given lung burden of particles, silicosis was suggested to progress without further exposure [3234]. Calvert et al. [35] recently reported an association of crystalline silica (mainly quartz) exposure and silicosis, as well as lung cancer, chronic obstructive pulmonary disease (COPD), and pulmonary tuberculosis. The carcinogenicity of quartz and cristobalite has been shown in several epidemiological studies [3638]. In , the International Agency for Research on Cancer (IARC) classified some crystalline silica polymorphs (quartz and cristobalite) in group 1 (sufficient evidence for the carcinogenicity to experimental animals and to humans), whereas amorphous silica (silicon dioxide without crystalline structure) was classified in group 3 (inadequate evidence for carcinogenicity) [39]. This classification has recently been confirmed [40]. Checkoway and Franzblau [41] reviewed occupational epidemiologic literature on the interrelations among silica exposure, silicosis and lung cancer and concluded that the appearance of silicosis is not necessarily required for the development of silica-associated lung cancer. Hnizdo and Vallyathan [42] suggested that chronic exposure to levels of crystalline silica dust, which does not cause disabling silicosis, may cause chronic bronchitis, emphysema, and/or small airway disease leading to airflow obstruction, even in the absence of radiological evidence of silicosis. Evidence has linked silica exposure to various autoimmune diseases (systemic sclerosis, rheumatoid arthritis, lupus, chronic renal disease), as reviewed by Steenland and Goldsmith [43]. A study by Haustein et al. [44] reported on silica-induced (silica dust) scleroderma.
Amorphous silica has been far less studied than has the crystalline form [39]. Warheit [45] briefly described the inhalation toxicity data related to amorphous silica particulates and concluded that some forms of amorphous silica are more potent in producing pulmonary effects as compared to others. He also emphasized the great need for adequate toxicological testing of many of these amorphous silicates given their importance in commerce and widespread potential for exposure. Workers exposed to precipitated or fumed silica did not exhibit pneumoconiosis [46, 47], but evidence of pulmonary fibrosis was reported in workers exposed to amorphous silica dust produced as a byproduct of silicon metal production [48]. Merget et al. [49] reviewed the current knowledge of the health effects of a wide range of amorphous forms of silica in humans. The major problem in the assessment of health effects of biogenic amorphous silica is its contamination with crystalline silica. This problem applies particularly to the well-documented pneumoconiosis among diatomaceous-earth workers. Although the data are limited, a risk of chronic obstructive bronchitis disease, COPD or emphysema cannot be excluded [49]. Animal inhalation studies involving synthetic amorphous silica (colloidal silica, fumed silica and precipitated silica) showed at least partially reversible inflammation [50, 51], granuloma formation and emphysema, but no progressive fibrosis of the lungs [52, 53]. However, high doses of amorphous silica may result in acute pulmonary inflammatory responses, which could conceivably trigger long-term effects, despite a low biopersistence of the particles [54]. The debate on the health effects of micron-sized crystalline or amorphous silica is beyond the scope of this article. Readers are referred to other publications [3538, 41, 5557].
As mentioned, most of the toxicological research into silica has focused on crystalline silica particles of 0.5 to 10 μm (coarse or fine particles). Despite the relatively large amount of available studies, the mechanisms of crystalline silica toxicity at the cellular and molecular levels are still unclear, and whether any single mechanism underlies all the above-mentioned diseases induced by these particles is uncertain [43]. However, severe inflammation following exposure to silica particles appears to be a common initiating step [58, 59].
The crucial role of reactive oxygen species (ROS) in the inflammatory, fibrogenic and carcinogenic activity of quartz is well established [60, 61]. Oxidative membrane and DNA damage are considered the most important mechanisms involved in the health effects of micron-sized crystalline silica. A few of the numerous reports clearly demonstrate these findings: ROS generated by the silica surface can induce cell membrane damage via lipid peroxidation that may subsequently lead to increased cellular permeability [62], perturbation of intracellular calcium homeostasis [63] and alterations in signaling pathways. Schins et al. and Fanizza et al. [64, 65] demonstrated that respirable quartz particles induce oxidative DNA damage in human lung epithelial cells. Li et al. [66, 67] demonstrated that micron-sized quartz particles induce .OH generation through an iron-dependent mechanism. A close association of .OH and iron ion concentration has been reported for amorphous silica particles [66, 67]. The study of Ghiazza et al. [30] indicates that crystallinity might not be a necessary prerequisite to make a silica particle toxic; both quartz and vitreous silica showed stable surface radicals and sustained release of HO. radicals. When tested on macrophages, vitreous silica and pure quartz showed a remarkable potency in cytotoxicity, release of nitrite and tumor necrosis factor α (TNF-α) production, suggesting a common behavior in inducing of oxidative stress [30]. Ding et al. [68] discuss the molecular mechanisms of silica-induced lung injuries with a focus on NF-kB activation, generation of cyclooxygenase II and tumor necrosis factor α (TNF-α). The review of Castranova [69] summarizes evidence that in vitro and in vivo exposure to crystalline silica results in activation of NF-kB and AP-1 signaling pathways. In vitro and in vivo animal studies, as well as investigations in humans, strongly support the role of macrophage products in the development and progression of silicosis [70]. Such products include a large panel of cytokines [71], with TNF-α seeming to determine the development of silica-induced pulmonary fibrosis [72]. In addition, recent evidence implicates interleukin 1β (IL-1β) and its activation by the NALP-3 inflammasome [73].
A large body of experimental work in the past 20 years has shown that 2 main factors seem to govern the hazardous nature of crystalline silica: particle surface reactivity and the form of silica [74]. Fenoglio et al. [75] evaluated these factors systematically, studying synthetic quartz samples differing only in size and shape. Cytotoxicity appeared to be primarily governed by the form of the particles and the extent of the exposed surface. Several studies indicate that the surface silanol groups are directly involved both in membranolysis [7678] and in toxicity to alveolar cells [79, 80]. Therefore, the distribution and abundance of silanols determines the degree of hydrophilicity (see "Physico-chemical properties of synthetic silica materials related to toxicity" described above) and seems to modulate cell toxicity [80, 81]. Experimental work with respirable silica particles and the survey of published data by Bagchi [82] suggest that the toxicity of these particles is caused by the large amount of positive charges they carry. Ghiazza et al. [83] reported that formation of a vitreous phase at the surface of some commercial diatomaceous earth prevents the onset of oxidative stress effects. Donaldson and Borm [84] emphasized that the ability of quartz to generate ROS could be modified by a range of substances that affect the quartz surface, such as substances originating from other minerals. The authors concluded that the toxicity of quartz is not a constant entity and may vary greatly depending on the origin/constitution of the sample.
The origin/synthesis of SNPs plays a crucial role in determining the physico-chemical properties of these particles and, consequently, their potential interactions with biological systems. Surface area, surface morphology, surface energy, dissolution layer properties, adsorption and aggregation properties are relevant parameters. Depending on the manufacturing process, amorphous silica has a wide range of physico-chemical properties that determine its industrial application. Bye et al. [85] showed that the cytotoxic activity of different forms of amorphous silica does not depend on a crystalline silica component but, rather, is caused by surface charges and the morphologic features of particles. Synthetic amorphous silica has been the subject of dissolution testing with a simulated biological medium, and the silica dissolution rate was reported as being more rapid than the reverse precipitation rate [86]. Solubility has been defined as a key driver in the clearance mechanisms involved in amorphous silica removal from lung [87]. Warheit [45] reviewed pulmonary responses to different forms of silica and reported that cristobalite produced the greatest lung injury, quartz produced intermediate effects, and amorphous silica produced minimal effects. In terms of analytical technique, small differences in dissolution exist among these different forms of silica, and dissolution, in turn, influences pulmonary effects through the concept of persistence. In addition, components from the biological system may react with the surface of the particle. A systematic investigation of iron-containing SNPs as used in industrial fine-chemical synthesis demonstrated the presence of catalytic activity that could strongly alter the toxic action of nanoparticles [88].
On the whole, considering the great variety of silica forms, degree of crystallinity, surface state and the presence of contaminants, there is a critical need for carefully characterized standard silica samples to unravel the relationships between physico-chemical factors and toxicity, both micron- and nano-sized. The main goal of this review is to focus on the toxicity of nanosilica, which has never been properly reviewed. Moreover, nanosilica occurs mainly in amorphous forms, and the potential hazard posed by these nanomaterials cannot be simply related to, as has already been reviewed many times, studies of micron-sized crystalline materials.
The growing abundance and industrial applications of nanotechnology has resulted in a recent shift of toxicological research towards nanoparticles [8994]. Ultrafine particles (< 0.1 μm) have been demonstrated to cause greater inflammatory responses and particle-mediated lung diseases than have fine particles (< 2.5 μm) per given mass [9597]. Also, experiments involving silica have shown that nanoparticles, both ultrafine colloidal silica [98, 99] and crystalline silica [99], have a greater ability to cause lung injury as compared with fine particles. Thus, the unique properties (i.e., small size and corresponding large specific surface area; cell penetrating ability) of nano-sized SiO2 are likely to impose biological effects that differ greatly from micron-scale counterparts.
A structured summary of in vitro studies of the toxicity of SNPs can be found in Table 2.
Table 2 In vitro studies on nanosilica particles (SNPs) toxicityFull size table
Chen and von Mikecz [100] investigated the effects of nanoparticles on structure, function, and proteasomal proteolysis in the cell nucleus by incubating different cell lines with unlabeled and fluorescently labeled amorphous silica particles of different sizes [100]. SiO2 particles between 40 nm and 5 μm were applied to epithelial cells in culture and observed on confocal laser scanning microscopy with differential interference contrast. Particles of all tested sizes penetrated the cytoplasm; however, nuclear localization was observed exclusively in cells treated with SiO2 nanoparticles between 40 and 70 nm. Fine and coarse SiO2 particles (0.2-5 μm) were exclusively located in the cytoplasm and accumulated around the nucleus, forming nuclear indentations. The uptake of SNPs in the nucleus induced aberrant clusters of topoisomerase I and protein aggregates in the nucleoplasm -- the former inhibiting replication, transcription, and cell proliferation -- without altering cell viability. Cells treated with fine (0.5 μm) or coarse (5 μm) SiO2 particles had the same replication and transcription activity as that of untreated control cells [100].
Jin et al. [101] investigated the potential toxicity of luminescent amorphous SNPs (50 nm) in freshly isolated rat alveolar macrophage cells and human lung epithelial cells (A549 cells). The SNPs penetrated the cells but were not detected in the nuclear region and did not cause significant toxic effects at the molecular and cellular levels below a concentration of 0.1 mg/ml.
Lin et al. [102] investigated the cytotoxicity of amorphous (colloidal) SNPs (15 and 46 nm) in cultured human alveolar epithelial cells (A549 cells). Cell viability decreased in a time- and dose-dependent manner (down to 100 μg/ml), and nanoparticles of both sizes were more cytotoxic than were fine quartz particles (Min-U-Sil 5). Exposure to 15-nm SNPs generated oxidative stress in A549 cells as reflected by reduced glutathione (GSH) levels, elevated production of malondialdehyde (MDA) and lactate dehydrogenase (LDH) leakage, which is indicative of lipid peroxidation and membrane damage, respectively [102]. In the study by Wottrich et al. [103], A549 cells and macrophages (THP-1, Mono Mac 6) exposed to 60 nm amorphous SNPs showed distinctly higher mortality than did larger silica particles (diameter 100 nm). Another study by Choi et al. [104], involving A549 cells and amorphous SNPs (14 nm), showed a pro-inflammatory response triggered by nanoparticles without blocking cell proliferation or causing cell death to any great extent. A recent work by Akhtar et al. [105] examined cytotoxicity (by MTT and LDH assay) and oxidative stress (ROS levels, membrane lipid peroxidation, GSH level and activity of GSH metabolizing enzymes) in A549 cells exposed for 48 h to amorphous SNPs of 10 and 80 nm. The SNPs were cytotoxic to studied cells through oxidant generation (ROS and membrane lipid peroxidation) rather than depletion of GSH. Eom and Choi [106] studied oxidative stress caused by amorphous SNPs (7 and 5-15 nm) in human bronchial epithelial cells (BEAS-2B) and observed the formation of ROS and induction of antioxidant enzymes.
Shi et al. [107] exposed A549 cells to amorphous SNPs (10-20 nm) at concentrations up to 200 μg/ml and observed low cytotoxicity as measured by MTT and LDH assays. However, co-treatment with the same nanoparticles and lipopolysaccharide, a bacterial product that may contaminate (nano)materials, significantly enhanced the cytotoxicity.
Yu et al. [108] examined the cytotoxic activity (by MTT and LDH assay) of well-dispersed amorphous silica particles (30-535 nm) in mouse keratinocytes. All sizes of particles were taken up into the cell cytoplasm; nuclear uptake was not studied. The toxicity was dose and size dependent, with 30- and 48-nm particles being more cytotoxic than 118- and 535-nm particles. The reduced GSH level significantly decreased only after exposure to 30-nm nanoparticles [108]. Nabeshi et al. [109] showed the size-dependent cytotoxic effects of amorphous silica particles (70, 300 and nm) on mouse epidermal Langerhans cells. The smallest particles induced greater cytotoxicity (by LDH assay) and inhibited cellular proliferation (by [3H]-thymidine incorporation). The observed effects were associated with the quantity of particle uptake into the cells.
Yang et al. [110] evaluated the effects of amorphous SNPs (15 and 30 nm) and micron-sized silica particles on cellular viability, cell cycle, apoptosis and protein expression in the human epidermal keratinocyte cell line HaCaT. Microscopy examination revealed morphological changes after 24-h exposure; cell growth also appeared to be significantly inhibited. The cellular viability of HaCaT cells was significantly decreased, and the amount of apoptotic cells was increased in a dose-dependent manner after treatment with nano- and micron-sized SiO2 particles. Furthermore, smaller silica particles were more cytotoxic and induced a higher apoptotic rate. Proteomic analysis revealed differential induction of expression of 16 proteins by SiO2 exposure; proteins were classified into 5 categories according to their functions: oxidative stress-associated proteins, cytoskeleton-associated proteins, molecular chaperones, energy metabolism-associated proteins, and apoptosis and tumor-associated proteins. The expression levels of the differentially expressed proteins were associated with particle size [110]. In a recently published study [111], the same research group used these SNPs to study the global DNA methylation profiles in HaCaT cells; the authors reported that nanosilica treatment can induce epigenetic changes.
Cousins et al. [112] exposed murine fibroblasts to small amorphous (colloidal) silica particles (7, 14 and 21 nm) over a long incubation period (1, 3 and 7 days and up to 7 weeks) and observed a distinctive cellular response affecting the morphologic features, adhesion and proliferation of the fibroblasts but not cell viability. Chang et al. [113] exposed selected human fibroblast and cancer cell lines for 48 h to amorphous SNPs and assessed cellular viability by MTT and LDH assays. Cytotoxicity was seen at concentrations > 138 μg/ml and depended on the metabolic activity of the cell line. However, the average primary size of tested silica particles was 21 and 80 nm, but their average hydrodynamic particle size was 188 and 236 nm, respectively, so in media, aggregates/agglomerates were formed.
In the study of Yang et al. [114], cell membrane injury induced by 20-nm amorphous silica nanoparticles in mouse macrophages was closely associated with increased intracellular oxidative stress, decreased membrane fluidity, and perturbation of intracellular calcium homeostasis.
Besides inhalation, ingestion is considered a major uptake route of nanoparticles into the human body [3]; however, the possible harmful effects of engineered nanoparticles in the gastrointestinal tract are still largely unknown. Recently, Gerloff et al. [115] investigated the cytotoxic and DNA damaging properties of amorphous fumed SiO2 nanoparticles (14 nm) in the human colon epithelial cell-line Caco-2. Exposure to SNPs for up to 24 h caused cell mortality, significant DNA damage and total glutathione depletion. The results of an in vivo study of mice fed nanosized silica are discussed in section 3.2.2.
Ye et al. [116] reported on induced apoptosis in a human hepatic cell line after exposure to amorphous (colloidal) SNPs (21, 48 and 86 nm). The viability of cells was assessed by LDH and MTT assay; oxidative stress was studied by measurement of ROS, lipid peroxidation and GSH concentration; and apoptosis was quantified by annexin V/propidium iodide staining and DNA ladder assays. Nano-SiO2 caused cytotoxicity in a size-, dose- and time-dependent manner.
Because nanoparticles are probably distributed by the blood stream (e.g., with medical applications), endothelial cells would also come in direct contact with these particles, for pathogenic particle-endothelial interactions. Peters et al. [117] evaluated the effects of 4- to 40-nm amorphous SiO 2 particles in vitro on human dermal microvascular endothelial cell function and viability. The particles were internalized but did not exert cytotoxic effects (MTS assay). However, cells showed impaired proliferative activity and pro-inflammatory stimulation. Napierska et al. [118] reported a dose-dependent cytotoxicity (by MTT and LDH assay) of monodisperse amorphous SNPs (16-335 nm) in a human endothelial cell line. The toxicity of the particles was strongly related to particle size; smaller particles showed significantly higher toxicity and also affected the exposed cells faster. Ye et al. [119] evaluated the toxicity of amorphous SNPs (21 and 48 nm) towards rat myocardial cells. Exposure to the SNPs for up to 48 h resulted in size-, dose- and time-dependent cytotoxicity, smaller particles again showing higher toxicity.
Barnes et al. [120] reported no detectable genotoxic activity (by Comet assay) of amorphous SNPs (20 nm to < 400 nm) in 3T3-L1 fibroblasts at 4 or 40 μg/ml silica for 24 h. The particle dispersions were carefully characterized and the results were independently validated in 2 separate laboratories. In a recent review, Gonzalez et al. [121], in a literature review, compared 2 genotoxicity tests -- the alkaline Comet assay and the micronucleus test - in terms of chemical composition and size of engineered SNPs: engineered SNPs did not seem to induce DNA strand breakage. However, when monodisperse amorphous SNPs of 3 different sizes (16, 60 and 104 nm) were selected to assess the genotoxic potential of these particles in A549 lung carcinoma cells with a well-validated assay (the in vitro cytochalasin-B micronucleus assay), at non-cytotoxic doses, the smallest particles showed an apparently higher-fold induction of micronucleated binucleated (MNBN) cells [122]. When considering the 3 SNPs together, particle number and total surface area accounted for MNBN induction because they were significantly associated with the amplitude of the effect.
Wang et al. [99] investigated cytotoxicity (by MTT assay) and genotoxicity of ultrafine crystalline SiO 2 particulates (UF-SiO2) in cultured human lymphoblastoid cells. A 24-h treatment with 120 μg/ml UF-SiO2 produced a fourfold increase in MNBN cells, with no significant difference as measured by the Comet assay. However, the ultrafine crystalline silica used was extracted from commercially available crystalline silica and the particle sizes were not uniform [99].
The cytoxicity of amorphous mesoporous SNPs (MSNs) was recently studied intensively because they are promising materials for drug delivery systems and cell markers [8, 123, 124]. Several studies have demonstrated that efficient cellular uptake of MSNs could be achieved at concentrations < 50 μg/ml, with no cytotoxic effects observed up to 100 μg/ml in different mammalian cells [125130]. Lu et al. [128] reported on the optimal size of ~50 nm MSNs for cell uptake. Slowing et al. [131] reported that, contrary to the known cytotoxicity of amorphous SNPs toward red blood cells, mesoporous SNPs exhibit high biocompatibility at concentrations adequate for potential pharmacological applications.
However, studies have reported cytotoxicity of mesoporous silica nanomaterials. Tao et al. [132] investigated the effects of two types of MSNs (pore diameters of 31 and 55 Å) on cellular bioenergetics (cellular respiration and ATP content) in myeloid and lymphoid cells and isolated mitochondria. Only cells exposed to MSNs with larger size and larger pores showed concentration- and time-dependent inhibition of cellular respiration, and both nanoparticles were toxic to the isolated mitochondria. Di Pasqua et al. [133] reported that the toxicity of MSNs towards human neuroblastoma cells was related to the adsorptive surface area of the particle. However, the nature of the functional groups playing a role could not be excluded. Vallhov et al. [134] investigated the effects of mesoporous SNPs of different sizes (270 nm and 2.5 μm) on human dendritic cells and found viability, uptake and immune regulatory markers affected by increasing size and dose. He et al. [135] evaluated the influence of size and concentration of mesoporous SNPs (190, 420 and nm) on cytotoxicity in human breast cancer cells and monkey kidney cells. The cytotoxicity of the particles was associated with particle size: silica of 190 and 420 nm in diameter showed significant cytotoxicity at concentrations > 25 μg/ml; whereas particles of nm in diameter showed slight cytotoxicity at 480 μg/ml. The smaller particles were suggested to be more easily endocytosed and consequently located within lysosomes [135].
Brown et al. [136] attempted to evaluate the role of shape in particle toxicity in the lung; the authors compared the response of rod-shaped and spherical amorphous silica particles (Stöber), not coated or coated with fibronectin or polyethylene glycol (PEG), under stretched and static conditions. The dosimetric comparison of materials with different shapes (e.g., needle-shaped or acicular and isotropic) was not straightforward. Non-coated particles induced an increase in IL-8 and LDH release, whereas a surface modification with PEG mitigated this effect, which suggested the significance of adhesive interactions for membrane binding/signal transduction, for example [136].
Diaz et al. [137] described the interactions of two amorphous silica particles - a pristine particle, without any coating, and PEGylated silica particles (average size 130 and 155 nm), as well as an iron oxide particle with a silica shell (80 nm) -- with different human peripheral blood cells, several human tumor cell lines and mouse peritoneal macrophages. The effects depended on the cell analyzed: although all particles were phagocytosed and were able to induce ROS expression in mouse macrophages, they differentially affected the human cell lines and peripheral blood cells, both in terms of internalization and ROS induction. The availability of the particles to be internalized by the cells seemed to strongly depend on aggregation, especially on the size and morphology of the aggregates [137].
Almost all of the existing cytotoxicity studies of SNPs involved monocultures of cells that are organ specific. The exception is the study by Wottrich et al. [103], in which co-cultures of epithelial cells (A549) and macrophages (THP-1, Mono Mac 6) exposed to 60- and 100-nm amorphous SNPs showed an increased sensitivity to the cytokine release as compared with monocultures of each cell type. The enhanced responses to nanoparticles in different contact and non-contact co-cultures were reported in studies by Herseth et al. [138, 139] with micron-sized crystalline silica, showing that more realistic models should be applied to study interactions between nanoparticles and cells or organs of interest.
Few recently published studies have systematically investigated nanomaterial properties in terms of the degree and pathways of cytotoxicity. Sohaebuddin et al. [140] selected nanomaterials of different composition, including silica, to analyze the effects of size and composition on 3 model cell lines: fibroblasts, macrophages and bronchiolar epithelial cells. The authors concluded that the physico-chemical properties of size and composition both determined the cellular responses and induced cell-specific responses. In another recent study, Rabolli et al. [141] studied the influence of size, surface area and microporosity on the in vitro cytotoxic activity of a set of 17 stable suspensions of monodisperse amorphous SNPs of different sizes (2-335 nm) in 4 different cell types (macrophages, fibroblasts, endothelial cells and erythrocytes). The response to these nanoparticles was governed by different physico-chemical parameters that varied by cell type: in murine macrophages, the cytotoxic response increased with external surface area and decreased with micopore volume; in human endothelial cells and mouse embryo fibroblasts, the cytotoxicity increased with surface roughness and decrease in diameter; and in human erythrocytes, the hemolytic activity increased with the diameter of the SNP [141].
Overall, most of these in vitro studies involving different SNPs documented the cytotoxic effects of these nanomaterials. The determinants of the observed cytotoxicity seem to be complex and vary with the particles used and cell type tested. Unfortunately, for many published studies, adequate material characterization is still missing. The mere cytotoxicity reported with some particles does not strictly imply hazard. However, this observation indicates that a proactive development of nanomaterials should consider physical, chemical and catalytic properties of nanoparticles.
Along with particle size, surface area and particle number appear to be integral components contributing to the mechanisms of lung toxicity induced by nano-sized particles. The high deposition rate of ultrafine particulates is a result of a small aerodynamic diameter and is assumed to be important in the lung inflammatory process. Some evidence suggests that inhaled nanoparticles, after deposition in the lung, largely escape from alveolar macrophage clearance and gain greater access to the pulmonary interstitium via translocation from alveolar spaces through epithelium [3, 142]. A summary of the in vivo responses to SNPs can be found in Table 3.
Table 3 In vivo studies on nanosilica particles (SNPs) toxicityFull size table
In , Warheit et al. [143] performed a rat inhalation study (nose-only) with an aerosol of colloidal silica (mass median aerodynamic diameter 2.9, 3.3 and 3.7 μm) for 2 or 4 weeks at concentrations up to 150 mg/m3, and some groups of rats were allowed to recover for 3 months. The inflammatory responses, mainly seen as increased numbers of neutrophils in bronchoalveolar lavage fluid (BALF), with the 2 and/or 4 weeks of exposure were evident at 50 mg/m3 concentration. Three months after exposure, most biochemical parameters returned to control values [143].
Lee and Kelly [52] studied the effects of repeated inhalation (6 h/day, 5 days/week for 4 weeks) of an aerosol of colloidal silica (mass median aerodynamic diameter 2.9, 3.3 and 3.7 μm; concentration up to 150 mg/m3) in rats. The authors reported a dose-dependent alveolar macrophage response, polymorphonuclear leukocytic infiltration, and type II pneumocyte hyperplasia in alveolar duct regions. Lung-deposited nanosilica were cleared rapidly from the lungs, with half-times of approximately 40 and 50 days for the 50 and 150 mg/m3 treatment groups, respectively. The lungs did not show formation of fibrotic scar tissue or alveolar bronchiolarization [52].
Cho et al. [144] investigated inflammatory mediators (24 h, and 1, 4 or 14 weeks after exposure) induced by intratracheal instillation in mice of up to 50 mg/kg of ultrafine amorphous silica with a primary particle diameter of 14 nm. The authors observed significantly increased lung weights, total cell numbers and levels of total protein in BALF up to 1 week after treatment. The histopathological examination revealed acute inflammation, with neutrophils and chronic granulomatous inflammation. The expression of cytokines (IL-1β, IL-6, IL-8, and TNF-α) and chemokines (monocyte chemoattractant protein 1 and macrophage inflammatory protein 2) was significantly increased during the early stages, with no changes after week 1 [144].
Chen et al. [145] studied age-related differences in response to amorphous SNPs (average size 38 nm). Changes in serum biomarkers, pulmonary inflammation, heart injury and pathology were compared in young (3 weeks), adult (8 weeks) and old (20 months) rats that inhaled tested nanoparticles for 4 weeks (40 min/day). Old animals appeared to be more sensitive to nanoparticle exposure than were young and adult rats. The risk of pulmonary damage was old > young > adult, but the risk of cardiovascular disorder was observed only in old animals [145].
Kaewamatawong et al. [98] compared acute pulmonary toxicity induced in mice by ultrafine colloidal silica particles (UFCSs; average size 14 nm) or fine colloidal silica particles (FCSs; average size 213 nm) after intratracheal instillation of 3-mg particles. Histopathological examination with both sizes revealed bronchiolar degeneration, necrosis, neutrophilic inflammation, alveolar type II cell swelling and alveolar macrophage accumulation. However, UFCSs induced extensive alveolar hemorrhage, more severe bronchiolar epithelial cell necrosis and neutrophil influx in alveoli as compared with FCSs. Electron microscopy showed UFCSs and FCSs on the bronchiolar and alveolar wall surface and in the cytoplasm of alveolar epithelial cells, alveolar macrophages and neutrophils. The findings suggest that UFCSs (possibly linked to size and/or larger surface area) have a greater ability to induce lung inflammation and tissue damage than do FCSs [98]. The same research group reported acute and subacute pulmonary toxicity of low-dose UFCS particles in mice after intratracheal instillation [146]. Exposure of up to 100 μg UFCSs produced moderate to severe pulmonary inflammation and tissue injury 3 days after exposure. Mice instilled with 30 μg UFCSs and sacrificed at intervals from 1 to 30 days after exposure showed moderate pulmonary inflammation and injury on BALF indices at the acute period; however, these changes gradually regressed with time. Concomitant histopathological and laminin immunohistochemical results were similar to BALF data. The authors reported a significant increase in the apoptotic index (TUNEL) in lung parenchyma at all observation times. The findings suggest that instillation of a small dose of UFCSs causes acute but transient lung inflammation and tissue damage in which oxidative stress and apoptosis may be involved [146].
In a study of fibrogenesis, Wistar rats were intratracheally instilled with silica (of unknown composition) nano- (10 ± 5 nm) and microparticles (0.5-10 μm), and were sacrificed 1 and 2 months after dosing [147]. One month after instillation, cellular nodules (Stage I silicosis) were found in the nano-sized SiO2 group, whereas more severe lesions were found in the micron-sized SiO2 treatment group (Stage II and Stage II+ of silicotic nodules). One month later, the nano-sized SiO2 group still showed only Stage I silicotic nodules, whereas the micron-silica group showed disease progression and Stage II+ and III silicotic nodules. Therefore, in rats, the effect of nano-SiO2 on fibrogenesis might be milder than that of micron-SiO2. Nanoparticles, because of their size, probably diffuse more easily to other pulmonary compartments than do microparticles [147].
Warheit et al. [148] compared the toxicity of synthetic nanoquartz particles (12 and 50 nm) to mined Min-U-Sil quartz (500 nm) and synthetic fine-quartz particles (300 nm) and (2) evaluated the surface activity (hemolytic potential) of the different samples in terms of toxicity. Rats were instilled with the different particle types (1 or 5 mg/kg), and pulmonary toxicity was assessed with BALF biomarkers, cell proliferation, and histopathological evaluation of lung tissue at 24 h, 1 week, 1 month, and 3 months after exposure. Exposure to the quartz particles of different sizes produced pulmonary inflammation and cytotoxicity, with nanoscale quartz of 12 nm and Min-U-Sil quartz being more toxic than fine quartz and nanoscale quartz of 50 nm. The pulmonary effects were not consistent with particle size but were associated with surface activity, particularly hemolytic potential [148].
In a recent work by Sayes et al. [149], rats inhaled freshly generated aerosolized amorphous SNPs of 37 and 83 nm for a short-term period. In contrast to previous studies' measurements, particle number rather than particle mass was chosen as dose metrics (3.7 × 107 or 1.8 × 108 particles/cm3) for 1- or 3-day exposure. Pulmonary toxicity (cell counts, differentials, enzymatic activity of LDH and alkaline phosphatase (ALP) in BALF) and genotoxicity endpoints (micronuclei induction) were assessed from 24 h up to 2 months after exposure. One- or 3-day aerosol exposure produced no significant pulmonary inflammatory, genotoxic or adverse lung histopathological effects in rats exposed to very high particle numbers in a range of mass concentrations (1.8 or 86 mg/m3).
Recently, airway irritants were suggested to facilitate allergic sensitization [150152]. Arts et al. [153] examined the effect of pre-exposure to synthetic (fumed) amorphous SNPs (14 nm) on elicitation of airway hypersensitivity reactions by the low-molecular-weight allergen trimellitic anhydride (TMA). Brown Norway rats were topically sensitized with TMA, exposed (head or nose only) to SNPs for 6 h/day for 6 days and then challenged by inhalation with a minimally irritating concentration of TMA. One day later, breathing parameters, cellular and biochemical changes in BALF, and histopathological airway changes were studied. Exposure to SNPs alone resulted in transient changes in breathing parameters during exposure and in nasal and alveolar inflammation with neutrophils and macrophages. Exposure to particles before a single TMA challenge resulted in only a slightly irregular breathing pattern during TMA challenge. Interestingly, pre-exposure to particles diminished the effect of TMA on tidal volume, laryngeal ulceration, laryngeal inflammation, and the number of BALF eosinophils in most animals. When an additional group of animals was exposed to nanosilica before a second challenge to TMA, the pulmonary eosinophilic infiltrate and edema induced by a second TMA challenge in control animals was diminished by the preceding silica exposure, but the number of lymphocytes in the BALF was increased. The authors concluded that SNPs could reduce as well as aggravate certain aspects of TMA-induced respiratory allergy [153].
As mentioned, next to inhalation, ingestion is considered a major route for the uptake of nanoparticles in the human body. So et al. [154] studied the effects on mice fed nano- and micron-sized amorphous silica particles (30 nm and approximately 30 μm, respectively). After feeding the animals for 10 weeks (total amount of 140 g silica/kg mouse), blood was tested biochemically and hematologically. The group fed SNPs showed higher serum values of alanine aminotransferase as compared with the other groups (both control and micron-silica treated). Although the contents of Si in the livers of the groups were almost the same, hematoxylin and eosin staining revealed a fatty liver pattern in the group treated with SNPs [154].
The successful use of nanoparticles in the clinic requires exhaustive studies on the behavior of these particles in vivo. Unfortunately, biocompatibility, biodistribution and clearance studies of silica-based nanoparticles are sparse. Kumar et al. [155] used nanoparticles of organically modified amorphous silica (ORMOSIL; amino-terminated; 20-25 nm) to study biodistribution, clearance and toxicity in a mouse model. Particles conjugated with fluorophore and radiolabeled were injected systemically in mice. Biodistribution studies showed a greater accumulation of nanoparticles in liver, spleen and stomach than in kidney, heart and lungs. Over 15 days, almost 100% of the injected nanoparticles were effectively cleared out of the animals via hepatobiliary excretion, without any sign of organ toxicity. Hudson et al. [156] examined the biocompatibility of mesoporous silica particles (150 nm, 800 nm and 4 μm) after injection in rats and mice. When the particles were injected subcutaneously in rats, the amount of residual material decreased progressively over 3 months, with no significant injury to surrounding tissues. Subcutaneous injection of the same particles in mice produced no toxic effects. In contrast, intra-peritoneal and intra-venous injection in mice resulted in death; microscopic analysis of the lung tissue of the mice indicated that death might have been due to pulmonary thrombosis. Nishimori et al. [157] evaluated the acute toxicity of amorphous silica particles (70, 300 and nm) after a single intravenous injection in mice and reported that 70-nm silica injured the liver but not the spleen, lung or kidney. Moreover, chronic administration of 70-nm nanoparticles (injections every 3 days for 4 weeks) caused liver fibrosis. Cho et al. [158] examined the impact of the size of amorphous SNPs on toxicity, tissue distribution and excretion. Fluorescence dye-labeled 50-, 100- and 200-nm silica particles were intravenously injected in mice. The incidence and severity of inflammation with the 100- and 200-nm SNPs was significantly increased in the liver at 12 h; the 50-nm particles induced a slight but nonsignificant inflammatory response. The tissue distribution and excretion of the injected particles differed depending on particle size. With increasing particle size, more particles were trapped by macrophages in the liver and spleen. All particles were cleared via urine and bile; however, the 50-nm SNPs were excreted faster than were the other 2 particle sizes [158].
Park and Park [159] performed in vitro and in vivo studies to investigate oxidative stress and pro-inflammatory responses induced by amorphous SNPs (average primary size 12 nm). RAW 264.7 cells derived from mouse peritoneal macrophages were exposed to SNPs (5-40 ppm) in vitro and showed ROS generation and decreased intracellular GSH levels, as well as increased levels of nitric oxide released from the cultured macrophage cell line. In vivo, mice were treated with a single intraperitoneal dose of 50 mg/kg of nanosilica. The treatment produced activated peritoneal macrophages, increased blood level of IL-1β and TNF-α, and increased level of nitric oxide released from peritoneal macrophages. Ex vivo, cultured peritoneal macrophages harvested from the treated mice showed the expression of inflammation-related genes (IL-1, IL-6, TNF- α, inducible nitric oxide synthase, cyclooxygenase 2). In the spleen, the relative distribution of natural killer cells and T cells was increased 184.8% and 115.1%, respectively, as compared with control animals, and that of B cells was decreased to 87.7% [159].
Kim et al. [160] addressed the toxicity of nano- and micron-sized silica particles (14 nm and 1-5 μm, respectively) in vitro and in vivo. In vitro, RAW 264.7 cells were exposed to both particle sizes for 24 h, and the cell viability was decreased in dose-dependent manner; however, apoptosis was observed only after treatment with nanoparticles. In vivo, mice received up to 5 mg/kg silica particles via oropharyngeal aspiration. Again, size-dependent toxicity of silica was observed; pulmonary injury and neutrophilic infiltration were greater after treatment with nano-sized SiO2 particles than with micron-sized silica [160].
Sayes et al. [161] assessed the capacity of in vitro screening studies to predict in vivo pulmonary toxicity of several fine or nanoscale particle types in rats. For the in vitro component of the study, rat lung epithelial cells, primary alveolar macrophages and alveolar macrophages-lung epithelial cell co-cultures were incubated with quartz particles and precipitated amorphous silica. In the in vivo component of the study, rats were exposed by intratracheal instillation to the same particles. In vivo, pulmonary toxicity studies demonstrated that crystalline silica particles produced sustained inflammation and cytotoxicity, whereas amorphous silica particles produced reversible and transient inflammatory responses. Ex vivo, pulmonary inflammation studies showed that crystalline and amorphous silica-exposed rat lung epithelial cells did not produce MIP-2 cytokines, but alveolar macrophages and, to a lesser degree, co-cultures secreted this chemotactic factor into the culture media. In vitro cytotoxicity studies demonstrated a variety of responses to the different particle types, primarily at high doses. When considering the range of toxicological endpoints, comparisons of in vivo and in vitro measurements revealed little correlation, particularly when considering the many variables assessed in this study such as cell types used, culture conditions and time course of exposure, as well as measured endpoints.
To summarize, extrapolating (or comparing) the results obtained in vitro to the in vivo situation is difficult and applies not only to toxicity studies with nanoparticles -- any existing in vitro test system lacks the complexity of animal models or the human body. However, considering the number of particles and the number of possible properties of these particles that may vary (size, shape, coating, etc.), clearly, not all can be evaluated in in vivo studies, and scientists have been striving to determine the correlation between the results obtained from in vitro and in vivo toxicity assessments. Although little correlation has been found in these studies with nanosilica [159161], Lu et al. [162] tested a panel of metal oxide nanoparticles and could predict the inflammogenicity of tested nanomaterials with a battery of simple in vitro tests. Similar conclusions were drawn in a recent study by Rushton et al. [163]; the authors could predict the acute in vivo inflammatory potential of nanoparticles with cell-free and cellular assays by using NP surface area-based dose and response metrics. The authors also found that a cellular component was required to achieve a higher degree of predictive power.
Established and validated co-culture systems may provide a tool to better mimic the in vivo system. Using recently developed 3-D cell cultures and improving the exposure system (likewise exposure at the air-liquid interface of a human epithelial airway model reported by Brandenberger et al. [164]), could substantially improve the outcome from in vitro studies with nanomaterials.
Want more information on high-purity silica nano powder? Feel free to contact us.