Uptake, distribution, clearance, and toxicity of iron oxide nanoparticles with different sizes and coatings

Iron oxide nanoparticles (IONPs) have been increasingly used in biomedical applications, but the comprehensive understanding of their interactions with biological systems is relatively limited. In this study, we systematically investigated the in vitro cell uptake, cytotoxicity, in vivo distribution, clearance and toxicity of commercially available and well-characterized IONPs with different sizes and coatings. Polyethylenimine (PEI)-coated IONPs exhibited significantly higher uptake than PEGylated ones in both macrophages and cancer cells, and caused severe cytotoxicity through multiple mechanisms such as ROS production and apoptosis. 10 nm PEGylated IONPs showed higher cellular uptake than 30 nm ones, and were slightly cytotoxic only at high concentrations. Interestingly, PEGylated IONPs but not PEI-coated IONPs were able to induce autophagy, which may play a protective role against the cytotoxicity of IONPs. Biodistribution studies demonstrated that all the IONPs tended to distribute in the liver and spleen, and the biodegradation and clearance of PEGylated IONPs in these tissues were relatively slow (>2 weeks). Among them, 10 nm PEGylated IONPs achieved the highest tumor uptake. No obvious toxicity was found for PEGylated IONPs in BALB/c mice, whereas PEI-coated IONPs exhibited dose-dependent lethal toxicity. Therefore, it is crucial to consider the size and coating properties of IONPs in their applications.

Magnetic iron oxide nanoparticles (IONPs) have been used for a wide range of biomedical applications such as drug delivery, magnetic resonance imaging (MRI), thermal ablation therapy, in vivo cell tracking, and magnetic separation of cells or molecules 1 . In recent years, more and more nanomedicines have been already approved by the U.S. Food and Drug Administration (FDA) for human use, and some others are undergoing clinical trials 2 .
For example, Feridex ® (ferumoxides) is the dextran-coated IONPs approved as a imaging contrast agent for the detection of liver lesions 3 . Feraheme ® (ferumoxytol) has been approved for the treatment of iron deficiency anemia in adult patients with chronic kidney disease 4 . The increasing applications have raised public concerns about the biosafety, long-term distribution, and clearance of IONPs.
Most IONPs introduced to the bloodstream are usually subjected to opsonization (adsorption of plasma proteins on the particles surface), followed by subsequent recognition and uptake by macrophages residing in the organs of the mononuclear phagocytic system (MPS), ultimately resulting in the elimination from the blood circulation 5 . It is generally believed that the interactions with biological components (e.g., proteins), cellular uptake, in vivo fate and toxicity of IONPs are strongly correlated with their physicochemical characteristics. For example, hydrodynamic size is one of the most important factors in determining the distribution and clearance of IONPs. It has been reported that IONPs larger than 100 nm in diameter are rapidly trapped in the liver and spleen through macrophage phagocytosis, whereas IONPs smaller than 10 nm in diameter are likely to be eliminated through renal clearance 6 . According to the information provided by the manufacture, the hydrodynamic size of IONPs is about 5-10 nm larger than their inorganic core size measured by TEM, due to the attachment of organic layers (PEG or PEI). DLS measurement (Fig. 1D,E and F) demonstrated that the hydrodynamic size of SEI-10, SMG-10 and SMG-30 were 17.2 ± 5.0 nm, 16.5 ± 4.7 nm, and 35.8 ± 10.3 nm, respectively. The polydispersity index (PDI) of SEI-10, SMG-10 and SMG-30 were 0.192, 0.163, and 0.105, respectively, and their zeta potential were +29.28 mV, −0.52 mV, and −0.52 mV, respectively ( Table 1). There was no apparent aggregation when these IONPs were incubated with Borate, Tris, HEPES or PBS buffer (data not shown), indicating that they were stable in the buffer solution. To predict the colloidal stability of IONPs in vivo, their size distribution in the presence of human plasma (50% in volume) was monitored by DLS measurement over time. As shown in Supplemental Fig. 1A, the particle size of SMG-10 and SMG-30 in human plasma were uniform during the first 24 h, although a small portion of aggregates was found at 48 h. In contrast, SEI-10 displayed scattered size distribution after mixing with human plasma, and more than 50% population had the average size of thousands of nanometers. Obvious precipitation was found in SEI-10 plasma samples, but not in SMG-10 and SMG-30 samples at 48 h (Supplemental Fig. 1B). These results suggest that SMG-10 and SMG-30 were relatively stable in in vitro cell culture medium and in vivo blood circulation, and SEI-10 tended to form large aggregates in such conditions. Particle uptake and localization. The uptake of IONPs in RAW264.7 macrophages and SKOV-3 cells were qualitatively visualized by Prussian blue staining ( Fig. 2A). The results demonstrated that SEI-10 had higher uptake, shown as more blue deposition, than SMG-10 and SMG-30 in both macrophages and cancer cells after 4-h incubation. In addition, the uptake of these IONPs in RAW264.7 macrophages appeared to be higher than that in SKOV-3 cancer cells. The particles uptake in macrophages was further quantitatively measured by ICP-MS. As shown in Fig. 2B, IONPs were internalized in macrophages in a concentration-dependent manner. Among these IONPs, SEI-10 had the highest cellular uptake in macrophages, followed by SMG-10, and finally SMG-30, which was consistent with Prussian blue staining results. No proportional increase of cellular uptake was found for all these IONPs when the exposure time increased from 1 h to 4 h (Fig. 2C), indicating the rapid phagocytosis of IONPs by macrophages. Intracellular particle localization of IONPs upon their uptake was further investigated using TEM images. After 2-h exposure, a large number of individually dispersed SEI-10, less SMG-10 and little SMG-30 were internalized and mainly localized within membrane-bound structures (e.g., endosome, lysosome) in both RAW264.7 macrophages and SKOV-3 cancer cells (Fig. 3). Interestingly, it appeared that the uptake of SEI-10 occurred after membrane accumulation (Supplemental Fig. 4), suggesting that the adsorptive endocytosis is the main mechanism of cell uptake of positively charged SEI. However, PEG-coated IONPs were independent from membrane accumulation, and seemed to follow common endocytic internalization processes and receptor-mediated endocytosis 12 . In addition, no signs of nuclear localization were found for all these IONPs. Fig. 4A and B, SEI-10 induced dose-dependent cytotoxicity against both RAW264.7 macrophages and SKOV-3 cancer cells at the test concentrations (3.125-100 µg/mL), and SKOV-3 cells were relatively more susceptible to SEI-10 toxicity than RAW264.7 macrophages. In contrast, no appreciable cytotoxic effects were observed for both SMG-10 and SMG-30 at the concentration of 25 µg/mL and below in both cells, although slight cytotoxicity was shown at the concentration above 50 µg/mL. The cell death induced by IONPs was further confirmed by Hoechst 33342 and PI staining. Hoechst 33342 stains the chromatin in living cells while PI is only permeable to dead cells. After 16-h exposure to 5 µg/mL SEI-10, the majority of nuclei in SKOV-3 cells displayed positive PI staining, which was the sign of cell death (Fig. 4C). In contrast, only a small portion of nuclei of cells treated with SMG-10 showed positive PI staining at the extremely high concentration (400 µg/mL), and the PI staining of cells treated with SMG-30 was almost negative even at the concentration up to 400 µg/mL. Mechanisms underlying the cytotoxicity of IONPs. Cell membrane integrity. The effect of IONPs on the cell membrane integrity was evaluated by LDH assay, which determines the amount of LDH enzyme released into the culture medium. As shown in Fig. 5A, SEI-10 treatment led to the dose-dependent release of LDH in SKOV-3 cells in at concentrations ranging from3.9 to 125 µg/mL. In contrast, only minimal or negligible LDH release was found in cells treated with SMG-10 or SMG-30 at the tested concentrations.

Cytotoxicity assays. As illustrated in
The hemolytic properties of IONPs were evaluated in mouse erythrocyte suspension after 4-h incubation at 37 °C. The results (Fig. 5B) demonstrated that both SMG-10 and SMG-30 had almost no hemolytic activity at the concentrations ranging from 15.6 to 125 µg/mL, whereas SEI-10 showed severe dose-dependent hemolysis.
Western blot. It was observed that the levels of anti-apoptotic protein Bcl-2 and pro-apoptotic protein Bax were down-regulated in SKOV-3 cells treated with SEI-10 in a dose-dependent manner when compared with control group, but there was no obvious difference in the Bcl-2/Bax ratio. Meanwhile, SEI-10 treatment resulted in a dose-dependent decrease in the protein expression of cyclin D, which is involved in the cell cycle regulation (Fig. 6F). Interestingly, the treatment of cells with SMG-10 and SMG-30, but not SEI-10, significantly increased the expression of autophagy marker LC3B-II (2-3 fold increase, Supplemental Fig. 5). To confirm the autophagy formation induced by PEGylated IONPs, TEM imaging was further performed to identify autophagosomes in IONPs treated cells. As shown in Supplemental Fig. 6, double-or multiple-membraned autophagosomes were observed in cells treated with SMG-10.
In vivo biodistribution. We next investigated the in vivo accumulation, biodistribution, degradation, and clearance of the IONP formulations in SKOV-3 tumor bearing mice. Excess iron distributed in the tumors obtained at 24 h post-injection using ICP-MS measurement, demonstrated that all the IONPs were able to accumulate in the tumor sites to varying extent via the enhanced permeability and retention (EPR) effects, and the tendency was: SMG-10 > SMG-30 > SEI-10 ( Fig. 7A). Besides, these IONPs tended to be trapped in the MPS organs such as liver and spleen (Fig. 7B), with a little less amount distributed in other organs such as lung, heart, and kidney. Interestingly, the uptake of SEI-10 in the kidney was significantly higher than that of SMG-10 and SMG-30, indicating its faster clearance. At 24 h post-injection, the iron concentration in the serum was almost undetectable. Tissue samples dissected at different time points (6 h and 2 weeks) after administration were stained with Prussian blue (iron) to observe time-dependent degradation and clearance of the IONPs accumulated in the liver and spleen (Fig. 7C,D). In particular, most of the IONPs were found in the Kupffer cells located in the linings of the hepatic sinusoids, as shown in the enlarged images. Consistent with the ICP-MS assay results, the iron staining of SEI-10 in the hepatic tissues at 6 h post-injection appeared weaker than that of SMG-10 and SMG-30. At 2 weeks post-injection, almost no staining was observed in the liver tissue of mice injected with SEI-10, whereas significant staining remained at the liver tissue of animals injected with SMG-10 or SMG-30. In the spleen, substantial iron staining was found in the tissue samples for all the IONPs formulations at 6 h post-injection. At 2 weeks post-injection, only minimal iron staining was left in the spleen tissue samples of SEI-10, whereas the majority of iron staining signals still remained in those of SMG-10 and SMG-30.
In vivo toxicity. In the preliminary study, all mice in 5 mg/kg SEI-10 group and a quarter of mice in 2.5 mg/kg SEI-10 died within 24 h post-injection (Fig. 8A), and the maximum tolerated dose (MTD) of SEI-10 in BALB/c mice was approximately 1.5 mg/kg. In contrast, no animal deaths were found for SMG-10 and SMG-30 at dose up to 5 mg/kg (data not shown). Since we observed significant amount of PEGylated IONPs remaining in the liver and spleen two weeks post-injection, we searched for signs of in vivo toxicity by monitoring body weight, hematology and blood biochemical indexes. The mice in all the IONPs treatment groups underwent transient slight body weight loss after administration (1.5 mg/kg), but they recovered gradually during one week (Fig. 8B), and no signs of distress such as behavioral changes were observed. The hematology (Supplemental Table 1) and blood chemistry (Supplemental Table 2) results on day 7 post-injection showed that blood cell counts and four indexes (AST, total bilirubin, BUN, and creatinine) out of the five hepatic and renal function panels were within the normal range, except that the level of ALT enzyme in mice treated with SMG-10 slightly increased compared to PBS control. On day 14 post-injection, the increased ALT level in SMG-10 treated mice returned to normal, and all other serum chemistry indexes were also in the normal ranges (data not shown). At 2 weeks post-injection, the liver and spleen tissues were harvested for histopathological examination. As shown in Fig. 8C, slight mononuclear cell infiltration in the portal area of the liver was identified in mice treated with SMG-10 and SMG-30. Splenic plasmacytosis was also noted in mice treated with SMG-30, which may represent a response to antigenic stimulation, presumably due to the uptake of exogenous IONPs.

Discussion
It is generally believed that the size of nanoparticles for cancer therapy should be in the range of 5-100 nm, in which the blood circulation time and EPR effect are maximized [13][14][15][16] . If the particle size is too small, it can be rapidly cleared through the kidneys, but if too large, it is efficiently trapped by cells in the reticuloendothelial system (RES) organs 17 . Meanwhile, appropriate surface coating is essential for biomedical applications of IONPs. For example, PEG is usually used to prevent the aggregation and opsonization of nanoparticles, reduce the macrophage uptake, and increase the blood circulation time 18 ; PEI is utilized to deliver DNA/siRNA due to its inherent ability to condense genetic material 19 . In this study, our aim was to understand the influence of particle size and surface coating on the biological distribution of IONPs and their biological effects in vitro and in vivo.
Commercially available IONPs with different particle sizes (10 nm or 30 nm) and surface coatings (PEG or PEI) were employed in the study. Their physicochemical properties including particle size, PDI, surface charge, and stability were well-characterized. After surface coating, the hydrodynamic sizes of SEI-10, SMG-10, and SMG-30 were approximately 17 nm, 17 nm, and 36 nm, respectively, and their PDI were less than 0.2, indicating their monodisperse distribution. The surface charge of IONPs with PEI coating (SEI-10) was positive, while IONPs with PEG coating (SMG-10 and SMG-30) were almost neutral. IONPs with PEI coating tended to rapidly form large aggregates in the presence of plasma, whereas PEG coating dramatically prevented the aggregation of IONPs, showing much better colloidal stability in the physiological condition. In the in vitro cellular uptake studies, both Prussian blue staining and ICP-MS measurement demonstrated that PEI-coated IONPs were internalized more efficiently than PEGylated IONPs with the same particle size, which may be due to the affinity of cationic nanoparticles to the negative phospholipid head groups or protein domains on cellular membranes 20 . Besides, small PEGylated IONPs (10 nm) showed relatively higher cellular uptake than large ones (30 nm), which is consistent with other previously reported nanoparticles 7 . TEM observation demonstrated that all the IONPs were distributed primarily in endosomes and/or lysosomes in both macrophages and cancer cells upon their uptake. However, SEI-10 was likely to be internalized through adsorptive endocytosis, while PEGylated IONPs seemed to be internalized through common endocytic internalization processes and receptor-mediated endocytosis 12 . In addition, no particles were found in the nuclei of cells treated with all the IONPs. Therefore, the observed genotoxic effects (mutagenicity) in our previous studies are presumably not due to the direct interaction of particles with DNA in the nucleus of cells 21,22 . In the in vitro cytotoxicity study, PEI-coated IONPs exhibited severe cytotoxicity against both macrophages and cancer cells at low concentrations, while no obvious cytotoxicity was observed for PEGylated IONPs at even much higher concentrations. In addition, SKOV-3 cancer cells seemed to be more susceptible to the cytotoxicity of SEI-10 than RAW264.7 macrophages. Owing to its special immune characteristics, macrophages show specific phagocytic ability, and tend to engulf the exogenous IONPs as demonstrated in our study. On the other hand, it has been reported that macrophages are highly tolerant to the cytotoxicity of anticancer drugs and nanoparticles 23 . Multiple molecular mechanisms were found to be involved in the cytotoxicity of PEI-coated IONPs, including the disruption of cell membrane integrity, ROS generation, apoptosis, and G2-phase cell cycle arrest. We speculate that the cytotoxicity of PEI-coated IONPs may originate from the PEI coating material, which was reported to be cytotoxic via two different mechanisms including the disruption of the cell membrane leading to necrotic cell death, and disruption of the mitochondrial membrane after internalization leading to apoptosis 24 . To overcome the toxicity issue and improve the biocompatibility of PEI-coated IONPs in biomedical applications such as gene delivery and MRI imaging, various optimization strategies have been adopted. For example, researchers utilized lactose to modify amphiphilic low molecular weight polyethylenimine (C12-PEI2K) at different lactosylation degree and demonstrated that the lactose modification can considerably reduce the cytotoxicity of PEI-coated IONPs without compromising their labeling efficacy as well as MR imaging capability 25 . It is also reported serum proteins could mask the cationic PEI surface of IONPs in a dose-dependent manner, leading to concurrent incremental decreases in cationic IONPs cytotoxicity 26 . In addition, it was interesting to find that IONPs with PEG coating but not PEI coating were able to induce autophagy, which may play a protective role against the cytotoxicity of IONPs.
Because of the high affinity with plasma proteins and rapid clearance of macrophages, positively charged nanoparticles are generally considered to have faster blood clearance than neutral particles. As expected, SEI-10 showed the lowest uptake in tumor tissue, whereas SMG-10 achieved the highest tumor uptake, followed by SMG-30. Meanwhile, it was not surprising that IONPs, irrespective of size and charge, mainly accumulated in the RES such as liver and spleen, where most macrophages reside. Prussian blue staining of liver and spleen sections showed that the degradation and removal of PEI-coated IONPs was faster than that of PEGylated IONPs, and PEGylated IONPs remained in the liver and spleen after two weeks of administration. Previous studies have demonstrated that IONPs are typically metabolized in the liver and other MPS organs. They are degraded in various ferrous Fe(II) and ferric Fe(III) products, which will be instead recycled and eventually incorporated into either the storage or utilization pathways (e.g. hemoglobin, ferritin, and transferrin) 27 . Toxicity study demonstrated that SEI-10 at the dose of 2.5 mg/kg and above may lead to animal death, which was likely due to the rapid hemolysis or capillary blockage as a result of large aggregates formation. Transient increase of ALT enzyme and slight histopathological changes in the liver and spleen were found in mice treated with IONPs with PEG coating. Long-term in vivo toxicity of IONPs warrants further investigation in consideration of their slow rate of degradation and clearance.
In summary, we have demonstrated both size and coating have remarkable impact on the cellular uptake, cytotoxicity, distribution and clearance of INOPs. When compared with PEGylated IONPs, PEI-coated ones had higher cellular uptake, stronger cytotoxicity, faster clearance, and less tumor distribution. Among PEGylated INOPs, 10 nm ones exhibited relatively higher cellular uptake and tumor accumulation than 30 nm ones. Therefore, the critical effects of particle size and surface coating have to be carefully considered in the biomedical applications of IONPs in order to improve their biocompatibility, prevent their undesirable clearance, and facilitate their targeted delivery to the disease site. Particle characterization. The morphology of IONPs was observed on a transmission electron microscope (TEM, Philips CM-120) at an acceleration voltage of 80 kV. The hydrodynamic particle size and zeta potential of IONPs were measured by the dynamic light scattering (DLS) technique using Zetatrac (Microtrac) as described previously 11 . The stability of these NPs in 50% human plasma was evaluated by monitoring the particle size of NPs over time. Particle location by TEM. RAW264.7 macrophages and SKOV-3 cells were incubated with IONPs at the concentration of 100 µg/mL for 2 h. Cells were fixed with 2% paraformaldehyde and 1% glutaraldehyde, dehydrated in graded dilutions of ethanol, embedded in artificial resin and processed for electron microscopy. Subcellular localization of IONPs was monitored by Philips CM-120 TEM.

Materials
Cytotoxicity assays. MTS assay. RAW264.7 macrophages and SKOV-3 cells were incubated with IONPs of different concentrations for 48 h. The cytotoxicity of IONPs was measured by MTS assay. Cell viability was expressed as the percentage of viable cells in the treated group compared to the untreated controls.
Hemolysis. Red blood cells (RBCs) were prepared using freshly harvested mouse blood. 200 μL of RBCs suspension (2% in PBS) was mixed with different concentrations of IONPs and incubated at 37 °C for 4 h. Following the incubation, the tubes were centrifuged at 1000 × g for 5 min, and the supernatant was collected for the detection of free hemoglobin at the absorbance of 540 nm. PBS was used as a negative control and Triton-100 (2%) was uses as a positive control. The percent hemolysis was calculated using the following formula: (OD sample − OD negative control )/(OD positive control − OD negative control ) × 100%. Cell cycle analysis. SKOV-3 Cells were treated with IONPs at indicated concentrations at 37 °C for 24 h. Cells were harvest, washed with PBS, fixed in 75% ethanol overnight at 4 °C. Then, cells were washed with PBS, treated with ribonuclease (RNase, 100 μg/mL), and stained with PI (10 μg/mL) for 30 min in dark. The samples were analyzed by flow cytometry under a 543-nm argon ion laser. Untreated cells were used as negative controls. Approximately 1 × 10 4 counts were collected for each sample.
Western blot. SKOV-3 cells were incubated with IONPs at indicated concentrations for 18 h. Cells were then washed with ice-cold PBS and lysed in RIPA buffer. Samples were run on sodium dodecyl sulfate-polyacrylamide gels (SDS-PAGEs), transferred onto PVDF membranes, and incubated with appropriate primary antibodies (Bcl-2, Bax, Cyclin D, LC3B, and actin). Blots were then incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies and visualized by enhanced chemiluminescence (GE Healthcare, Little Chalfont, Buckinghamshire, UK).
In vivo biodistribution. Female nude mice were injected subcutaneously with SKOV-3 cells (5 × 10 6 cells in PBS) on the right flank to establish ovarian cancer xenograft mouse model. When the tumors size reached an approximate diameter of 8-10 mm, the mice were randomly divided into four groups, with 4 mice in each group. Tumor bearing mice were intravenously injected with PBS or various IONPs (1.5 mgFe/kg), respectively. At 24 h post-injection, tumor tissues, blood and main organs (liver, spleen, lung, heart, and kidney) were collected, weighed and then digested in aqua regia. The iron concentration in the samples was determined by ICP-MS.
To evaluate their biodegradation and clearance, various IONPs were intravenously injected into BLAB/c mice at a dose of 1.5 mgFe/kg (n = 4). Liver and spleen were collected at 6 h and 2 weeks after injection, fixed in 10% neutral buffered formalin, embedded in paraffin, cut into 10 μm-thick slices. The sections were processed with Prussian blue and nuclear fast red staining.
In vivo toxicity. BALB/c mice (n = 4) were intravenously injected with various IONPs at different dose (1.5, 2.5, or 5 mg/kg), respectively. The body weight of mice was weighed twice per week for 2 weeks. On day 7, blood samples were obtained from all mice for hematological and serum chemistry analysis, including complete blood count (CBC), alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin, blood urea nitrogen (BUN) and creatinine. After 2 weeks, the mice were sacrificed, and liver and spleen tissues were harvested for histological analysis. The sections were stained with hematoxylin and eosin and then examined by a pathologist.
Statistical analysis. Data was analyzed in GraphPad Prism (version 6.01) by one-way analysis of variance (ANOVA) followed by Dunnett's multiple comparison, respectively. P values lower than 0.05 were considered statistically significant. Data are expressed as the mean ± standard deviation (SD). Data availability. No datasets were generated or analyzed during the current study.