Acute toxicity and genotoxicity of silver nanoparticle in rats

Objective The potential risk of a nanoparticle as a medical application has raised wide concerns, and this study aims to investigate silver nanoparticle (AgNP)-induced acute toxicities, genotoxicities, target organs and the underlying mechanisms. Methods Sprague-Dawley rats were randomly divided into 4 groups (n = 4 each group), and AgNP (containing Ag nanoparticles and released Ag+, 5 mg/kg), Ag+ (released from the same dose of AgNP, 0.0003 mg/kg), 5% sucrose solution (vechicle control) and cyclophophamide (positive control, 40 mg/kg) were administrated intravenously for 24 h respectively. Clinical signs and body weight of rats were recorded, and the tissues were subsequently collected for biochemical examination, Ag+ distribution detection, histopathological examination and genotoxicity assays. Results The rank of Ag detected in organs from highest to lowest is lung>spleen>liver>kidney>thymus>heart. Administration of AgNP induced a marked increase of ALT, BUN, TBil and Cre. Histopathological examination results showed that AgNP induced more extensive organ damages in liver, kidneys, thymus, and spleen. Bone marrow micronucleus assay found no statistical significance among groups (p > 0.05), but the number of aberration cells and multiple aberration cells were predominately increased from rats dosed with Ag+ and AgNP (p < 0.01), and more polyploidy cells were generated in the AgNP group (4.3%) compared with control. Conclusion Our results indicated that the AgNP accumulated in the immune system organs, and mild irritation was observed in the thymus and spleen of animals treated with AgNP, but not with Ag+. The liver and kidneys could be the most affected organs by an acute i.v. dose of AgNP, and significantly increased chromosome breakage and polyploidy cell rates also implied the potential genotoxicity of AgNP. However, particle-specific toxicities and potential carcinogenic effect remain to be further confirmed in a chronic toxicity study.

Introduction Silver nanoparticle (AgNP), having potent broad-spectrum antibacterial properties, strong permeability and little drug resistance, was used to produce a range of antibacterial medical products, such as, toothpaste, gynecologic suppository and wound dressing [1].Potential adverse effects consequentially associated with exposure to AgNPs are of concern. The most prominent characteristic of a metal nanoparticle is that, as a carrier, it could enhance the organ enrichment of ions [2], which also allows its extensive application in targeted cancer treatment and biomedical imaging technology. On the other hand, nanoparticles could take a long period to clear once they accumulated in the organs, and they may have a toxic effect in persistence [3,4]. Thus, the concerns have been raised on the potential risk of using nanoparticles in medical applications. In recent years, accumulating evidence has shown toxicities induced by AgNP in various in vitro experimental models, such as alveolar macrophages [5], neutrophils [6] and also sertoli and granulosa cells [7]. These results need to be confirmed in the in vivo system.
Currently, however, the information of AgNPs' toxicities based on in vivo studies is very limited and often controversial. A recent study [8] suggested that short-term oral administration of high doses of AgNP (5 to 100 mg/kg) could significantly increase ROS, ALT, AST, ALP, and lipid hydroperoxide, and cause DNA breakage. By contrast, in a 28-day inhalation toxicity study (1.32× 10 6 AgNP /cm 3 ), no changes on bodyweight, hematology and blood biochemical parameters of Sprague-Dawley (SD) rats were observed [9]. Another study also suggested that SD rat oral gavage with up to 36 mg/kg AgNP for 13 weeks showed no obvious change in histopathology, hematology, serum chemistry, micronuclei, and reproductive system parameters [10]. The toxicities that resulted from the different administration routes often varied due to the subsequent distribution patterns. For example, in a single-dose oral administration study [11], the tissue distribution of Ag in the liver, kidneys, and lungs was higher when Ag + was administered compared with AgNP. Whereas intravenously administered AgNP predominantly accumulated in the liver and spleen, and the free Ag + were subsequently released and excreted, and most of which were deposited in the kidneys, lungs, and brain [12]. Therefore, it is crucial to investigate the distribution pattern of AgNPs vs. Ag + and to understand their toxic effects.
Although the genotoxicity of nanoparticles and the underlying mechanisms have been widely studied, most of the results were obtained from cell lines [1,11,[13][14][15]. For instance, AgNP induced dose-dependent DNA damage was measured by single cell gel electrophoriesis and cytokinesis blocked micronucleus assay in human lung fibroblas cells. Our previous in vitro micronucleus test and golden hamster embryo cells transformation data also suggested the potential genotoxicity and carcinogenicity of AgNP [16]. The persistency of metal nanoparticles in biological systems increases the risk of carcinogenicity, and thus it is important to investigate the tissue distribution, toxicity, and genotoxicity of AgNP in vivo.
In this study, SD rat as a popular model was adopted for the in vivo bio-distribution and toxicities including genotoxicities triggered by AgNP intravenous administration. In addition, AgNP-released Ag + from the same dose AgNP were also used to compare the toxicological differences between AgNP and AgNP released ion.This research will provide a comprehensive insight about AgNP bio-distribution and AgNP-induced genotoxicity in vivo, which will help us have a better understanding of the potential risk of AgNP containing medical devices.

Materials and methods
Characterization of AgNP and Ag + AgNP solution (2000 μg/mL) was purchased from Nanux (SL1105001, Korea) and the characteristics of AgNP used in this study were described in our previous study [16] The morphology was observed by transmission electron microscopy (TEM, Tecnai G2 20 S-TWIN, America). The size distribution of AgNP suspended in deionized water and the Z-potential were determined using a Malvern Zeta Sizer Nano ZS (Malvern Instruments, Worcestershire, U.K.). AgNP was suspended in a 5% sucrose solution, and the sizes of AgNP were 6.3-629 nm (61.1% ranges from 27.3-106.2 nm. Fig 1). While Ag + were prepared from the supernatant obtained by centrifuging AgNP (1 mg/mL, suspended in a 5% sucrose solution) at 20,000 rpm for 2 h at room temperature after the supernatant was left at 37˚C for 24 h, and the concentration of Ag + was detected as 60 ng/mL by atomic absorption spectroscopy (AAS, MKII M6, U.S.).

Ethics statement
All animal experiments and sample collections were performed within the barrier system and a necropsy room at the National Center for Safety Evaluation of Drugs (NCSED). The protocols were approved by the Institutional Animal Care and Use Committee (IACUC) at NCSED and conducted in compliance with China's national ethical standards to minimize the suffering of animals (see S1 and S2 Files for animal research description).

Animal grouping and administration
16-week-old wild type specific pathogen free (SPF) female Sprague-Dawley rats (with body weights between 280 g and 310 g) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd.(Beijing, China; Animal Quality Certificate No: SCXK(Jing)2007-0001). The rats were housed in polycarbonate cages in a barrier system maintained at 20-25˚C with 40-70% relative humidity, a 12 h light-dark cycle, and a room air exchange of 10-20 times per hour. The feeding density was three rats per cage after dosing. Rats had ad libitum access to the certified rodent diet, and sterilized municipal tap water was given ad libitum through water bottles. Each rat was given a unique number and identified by its ear tag and animal number. All rats neither had previous procedures nor abnormal clinical conditions before the study.
SD rats were randomly divided into four groups, including vehicle control, AgNP, Ag + , and CPA positive control group (n = 4 for each group). 5% sucrose solution was used as vehicle control. CPA was single-dosed i.p. at 16 h before sacrifice, while the sucrose (vehicle control), AgNP, and Ag + were administered i.v. at 24 h before sacrifice. The administration doses are 40 mg/kg for CPA, and 0, 5, and 0.0003 mg/kg for AgNP and Ag + respectively. The administration route and doses for the AgNP were chosen to reach the maximum blood concentration possible for genotoxicity. Colchicine was dosed i.p. at 4 h before sacrifice at 4 mg/kg to maximize the metaphase cells. All animals were anesthetized by CO 2 inhalation to minimize suffering, then sacrificed by abdominal venesection method.

Clinical signs and body weight measurement
Clinical signs of animals before and 0-3 h after dosing were carefully observed, including the appearance, activities, hair, possible trauma, feces, and death. Weight change, an important toxicity index for rats, was measured before grouping, dosing, and sacrifice for statistical analysis. No differences were observed within dosing groups as compared to control.

Ag detection in organs
The heart, lungs, liver, spleen, kidneys, and thymus (n = 3-4) were collected and weighed during necropsy. The organ tissues were digested in 6 mL concentrated nitric acid using the microwave digestion system (MARS, CEM, USA). Then, the content of Ag (μg/g) in organs was detected through inductively coupled plasma mass spectrometry (ICP-OES, Optima 5300DV, America).

Histopathological examination
The heart, lungs, liver, spleen, kidneys, and thymus of animals (n = 3) were collected. Tissues were subsequently fixed, dehydrated, paraffin embedded, sliced (about 3 μm thick) and stained with hematoxylin eosin for histological observation under a light microscope.

Bone marrow micronucleus assay
The unilateral femur of each animal was removed and the bone marrow was washed using fetal bovine serum for harvesting cell suspensions. The cell suspension was centrifuged at 1000 rpm for 5 min, then most of the supernatant was removed. Cells in the remaining supernatant were resuspended, and bone marrow smears were prepared on clean slides (3-4 per animal). Cells were stained by 5% Giemsa to calculate the number of polychromatic erythrocytes (PCE) and normochromatic erythrocytes (NCE) in a total of 200 erythrocytes (ERY) in each animal; 1% acridine orange was used to discriminate and calculate the micronucleated reticulocytes frequencies in 2000 PCEs in each animal.

Bone marrow chromosome aberration test
Bone marrow cells were harvested from the other femur of each rat for the chromosome abbreviation test. The femur was washed using 5 ml sodium chloride solution. Cell suspension was centrifuged at 1000 rpm for 5 min, and the supernatant was removed subsequently. 0.075M potassium chloride (7 ml) was added to the cells, and incubated at 37˚C for 30-60 min for hypotonic treatment. Fixative solution (3 ml, glacial acetic acid: absolute methanol = 1:3) was added to the cells, centrifuged at 1000 rpm for 5 min, and the supernatant was discarded. Then 10 ml of fixative solution was added and cells were incubated at room temperature for 20 min before centrifuging at 1000 rpm for 5 min to remove the supernatant. The fixing was repeated for 2 more times, and fresh fixative solution was added to mix well with the cells. The cells were prepared on clean slides (2-3 slides per animal) using dripping method and dried at room temperature. The slides were stained using 5% Giemsa staining buffer for about 30 min and rinsed with tap water. For each animal, 200 metaphase cells with good dispersion were analyzed to calculate the occurrence rate of cells with chromosomal aberrations.

Data analysis and statistics
All data are shown as the mean±SD of n values, where n corresponds to the number of rats used. The distribution of all continuous parameters was assessed using Fisher's exact test. Statistical analysis for Ag concentrations in organs and biochemical examinations calculated by one-way ANOVA, followed by Dunnett's test for comparisons against control; the micronucleus and chromosome abbreviation assays were analyzed using chi-square test. The figures were prepared using GraphPad Prism 5 for Windows (GraphPad Software, San Diego, CA, USA), and the statistical significance was determined using SPSS (ver.12), as values were considered significantly different when p < 0.05.

Clinical signs
No clinical symptom was observed in the vehicle control, Ag + , and CPA groups. In contrast, increased breathing rate and decreased activity were observed in all animals in the AgNP group immediately after administration. Brown-red color urine appeared (2 of 4 animals) within 8 h after administrated with AgNP. The brown-red color urine was further investigated in our another study (S1 Text, S1 Table). The red color and massive precipitates were observed in 3 out of 3 urine samples from SD rats on the day administrated with 5 mg/kg AgNP (Nanux) i.v., and erythrocytes were only clearly observed in one sample under microscope. This phenomenon was disappeared in all animals after 24h. These results together suggested the brownish red colored AgNP in the circulatory system could stain the urine, and the Ag + would acutely excreted from the body through the urinary system. Purpura and congestion were observed around the injection sites of animals administered with AgNP, suggesting vascular irritation and injury. Difficulty in blood collection was experienced in animals administered with Ag + and AgNP.

Organ distribution of Ag
Bio-distribution of Ag was investigated after a single-dose intravenous administration as shown in Fig 2 (n = 3). The rank of Ag detected in organs in AgNP group through ICP-MS from highest to lowest is lung>spleen>liver>kidney>thymus>heart. In contrast, Ag content is undetectable in all organs of animals dosed with the released Ag + . Compared with the released Ag + group, the AgNP dosed group demonstrated a different distribution pattern, showing the lungs, spleen, and liver enriched with Ag content.

Biochemical examination
Intravenous administration of AgNP induced marked increase of ALT, BUN, TBil and Cre, and the reduction of P (Table 1, n = 3-4, samples from 4 aminals in the control group and 3 animals in the Ag + group and AgNP group were analyzed), whereas, only decreasing of BUN and P was found in Ag + group. This may be due to the much lower Ag + plasma concentration in rats dosed with Ag + .

Histopathological examination
In contrast to control, AgNP induced more extensive organ damages (Fig 3), such as multifocal liver cell degeneration, necrosis and hemorrhage (3 of 3  Critical cell degeneration and necrosis in the liver and kidneys observed in the AgNP group is

Fig 2. Ag dstribution in organs for SD rats intravenously single-dosed with AgNP (n = 3-4).
Concentrations of Ag in organs collected from rats administrated with AgNP was analysised, as lung, spleen and liver showed as the major targeted organs. consistent with the bio-distribution and biochemical results, and due to the high Ag concentrations that accumulated in these organs.

Genotoxicity
The data and representative micronuclei (MN) observed in the bone marrow micronucleus assay are shown ( Table 2, Fig 4). No obvious myelosuppression compared to the vehicle control group was observed. Although the MN‰ of samples collected from the Ag + and AgNP groups were relatively higher than those from the vehicle control group (Control: 5.13 ± 0.85‰; AgNP: 12.63 ± 6.93‰; Ag + : 9.25 ± 2.99‰ and CPA: 21.13 ± 5.54‰ respectively), no statistical significance was found (p > 0.05). By contrast, a chromosome aberration test revealed the damages caused by Ag + and AgNP in more detail (Table 3, Fig 5). The number of aberration cells (AC) and multiple aberration cells (MAC) were predominately higher in the bone marrow samples from animals dosed with Ag + and AgNP, and the data range were, for AC%, AgNP: 14.3% and Ag + : 21.3%; for ACG%, AgNP: 15.1% and Ag + : 23.6%; for MAC%, AgNP: 7.1% and Ag + : 7.8%. Thesepresented significant differences (p < 0.01) when compared with the vehicle control. Although the Ag + generated more structural aberration cells than AgNP (p < 0.01), there was no difference between the counts of multiple aberration cells between the AgNP and Ag + groups (p > 0.05). The most interesting phenomenon was that more polyploidy cells were generated in the AgNP group (4.3%), while much lesser polyploidy cells can be observed in the Ag + group (0.1%) compared to control.

Discussion
In this study, acute toxicity and genotoxicity of AgNP (containing Ag nanoparticles and released Ag + ) or released Ag + (generated from same dose AgNP) after a single i.v. injection  was studied. Biochemical markers and histopathological changes were observed in the liver and kidneys. AgNP accumulated in main immune system organs including the thymus and spleen. Histopathology results also showed that mild irritations were observed in the thymus and spleen only in the AgNP-treated group rather than the Ag + -treated group. Furthermore, chromosome breakage and polyploidy cells rates were significantly higher, implying the potential genotoxic and carcinogenic effects caused by AgNP. The small size of nanoparticles defines their distinct bio-distribution pattern and target organs. The impact of particle size to the AgNP's toxicity on cell death and cell cycle progression has been reported previously [17,18]. Particle size plays an important role on the uptake kinetics of NPs in the cells [19]. Our results showed that the Ag were predominantly accumulated in the lungs, spleen, liver, and kidneys of rats dosed with AgNP, suggesting the AgNP transferred and accumulated into specific target organs where they may further generate Ag + . However, for the animals dosed with the Ag + alone, the concentration of Ag was undetectable in all organs, probably due to the lower dose of Ag + administration. As reported previously, AgNPs were prone to accumulate in the liver, lungs, and kidneys [20]. After a single i.v. injection, AgNP distributed into the pulmonary circulation system, and thus predominately accumulated in the lungs, which could potentially lead to chronic lung toxic effects with extended administration period. Due to the short study period, however, we didn't observe acute lung toxicity induced by AgNP. A recent study [21], nevertheless, has showed evident vascular injury could be induced by a 7 day repeated pulmonary exposure of AgNP (20 nm). The liver and kidneys were the targets of AgNPs in rats via oral administration [22][23][24][25], and the TEM data demonstered that the AgNPs were able to penetrate into the liver cells, with the AgNPs tended to bind to proteins around 25-70 kDa. The potential systemic toxicity of AgNPs  remains controversial. In general, oral administration of AgNP showed no obvious organ toxicity in subacute toxicity studies [26][27][28]. However, high concentrations of AgNP in the circulatory system may cause severe toxicities [29]. For instance, intravenously administration of AgNP at 20 mg/kg and above in Wistar rats exhibited significant changes in WBC count, platelet count, hemoglobin, and RBC count, and the levels of liver function enzymes (including ALT, AST, ALP, GGT and TBil) were elevated when 40 mg/kg of AgNP were injected [29].
The maximum blood concentration of AgNP was previously measured as 1 μg/ml with a bioavailability of 4.2% while orally administrated at 10 mg/kg in rats [20]. As such, in our study, the plasma concentration of AgNP after a single i.v. dose with 5 mg/kg of AgNP is estimated to be about 125 μg/ml. Similarly, the study showed that the toxic effects in the liver and kidneys were observed 48 h after i.v., and the increased ALT, BUN, TBil and Cre levels implied the appearance of acute liver and kidney injuries in SD rats. Moreover, lymphocyte and macrophage infiltration were revealed as the key observations for AgNP induced morphological changes [24,30], in addition to cell degeneration, regeneration and necrosis. Nanoparticles introduced inflammation and immune response while interacting with the tissues, which is also a key mechanism of nanotoxicity. Therefore, the mild irritation observed in the spleen and thymus was proposed to be associated with the immune recognition process. AgNP could enter the human body by digestive tract, respiratory tract, skin or blood vessel, and introduce injuries in liver, kidneys [31], lungs [32]and central nervous system [33]. It was highly persistence in rats and difficult to be excreted from the body, and thus could exerted its toxicity in a chronic mode. The toxicities of AgNP and Ag + at similar concentration levels were compared and different toxiciological mechanisms were exhibited. For example, in a single i.v. exposure study [34], AgNP demonstrated significant spleen and liver toxicities, while the silver acetate (AgAc) at the same concentration (10 mg/kg) mostly distributed in the kidneys and therefore exhibited a distinct toxic manner. Another study focused on the potential cardiac toxicity of nano materials and demonstrated that both AgNP and Ag + could trigger severe myocardial conditions in mice at dosages above 6 mg/kg. However, AgNP induced a sinus bradycardia and complete atrio-ventricular conduction block, whereas the Ag + led to multifocal ventricular arrhythmias [35].
Whether the toxicities caused by AgNPs were completely attributed to the released ions are yet an elusive question, andour study attempted to discriminate the toxicities produced between AgNP and Ag + liberated immediately from the same concentration of AgNP (estimated to be 0.006% of AgNP). Some previous studies suggested that the Ag + , not the nanoparticles, were responsible for the major toxic effects of AgNPs [36,37]. Most of the toxic effects caused by AgNP were contributed by the dissolved Ag + [38], and the possible mechanisms include activation of lysosomal acid phosphatase activity, disruption of actin cytoskeleton and stimulation of phagocytosis, increase of MXR transport activity, inhibition of Na-K-ATPase, etc. An exquisitely designed study [39] ruled out the particle-specific effect of AgNP, since the AgNP showed negligible cytotoxicity to bacteria when synthesized and tested under strict anaerobic conditions to preclude the release of Ag + . Recent in vitro studies further pointed out that the toxicity of AgNP is mainly depend on the intracellular release, but not the silver ions liberated in the culture medium [40,41]. However, Lin et al's study reported that the lethal bradyarrhythmias could be generated at the presence of AgNP, and further suggested that AgNP was the one contribute to their gross acute toxic effect on myocardial I Na and I K1 channels, as the released Ag + was estimated to be less than 0.02% [35]. In addition, the organismspecific immune response induced by nanoparticles should not be underestimated. Our recent KEGG pathway analysis implied that inflammatory signal pathways in rats can be affected by AgNP but not by ionic Ag alone. It further suggested that inflammatory response may be important for AgNP induced toxicity [42]. The enhanced toxicity produced by AgNP observed in our study, by contrast with the toxicity generated by Ag + , is in accordance with both the higher dose and the organ enrichment feature of AgNP.
The genotoxicity of AgNP has been universally evaluated in vitro [1,11]. The Ames test result of AgNP was negative, which might be due to the bacterial being is incapable of endocytosis, whereas the effects on inducing micronuclei formation and DNA breakage was discovered. Few studies have examined the in vivo genotoxicity of AgNP so far, and it is innovative to focus on the extra effects of the nanoparticle itself. Our data demonstrated that both AgNP and Ag + produced certain chromosome damages to the bone marrow cells, which were mainly in the form of chromosome or chromatid breakages. Although it is thousands of times less concentrated, such effect is more significant in those injected with Ag + , implying Ag + , besides the nanoparticle is the fundamental cause of the gap, breakage, and exchange in chromatid and chromosome. The AgNP-induced delay of the cell cycle from G0/G1 to S phase using golden hamster embryo cell model has been previsouly observed [16]. These data altogether suggested that the AgNP might reinforce the chromosomal damage of Ag + together with its effects on the cell cycle [13]. The possible mechanism of AgNP-induced genotoxicity was involved with the interruption of ATP synthesis, subsequent to the disruption of the mitochondrial respiratory chain and excess production of serum reactive oxygen species (ROS) [13].The genetic materials is susceptible to oxidative free radical attacks induced by the metal nanoparticles [2]. Elevated serum ROS can be measured in Wistar rats treated with AgNP, and the single-cell gel electrophoresis data had a significant tail migration [29]. In addition, evidence showed that the AgNP might interact with the DNA directly by disrupting the hydrogen bonding between DNA double strand, and affect its conformation change in calf thymus [43]. Furthermore, the occurrence rate of polyploidy was significantly higher in AgNP group suggested additional genotoxicity induced by AgNP, which was very likely associated with the nanoparticle itself but not the ions. In contrast, the reduced polyploidy rate in Ag + group might be due to the increased chromosome fragmentation. AgNP-induced increased polyploidy rate has been reported previously [44,45], but has not been highlighted in studies with other metal nanoparticles. Kim et al. showed that AgNP participates in the ROS-induced genotoxicity, which plays important role in mediating DNA and chromosome instability, as well as mitosis inhibition [14]. The increase of polyploidy cells was also suggested to be associated with the activation of the G 2 /M DNA damage checkpoint and ATR/p53/p21 signaling, which can be rescued by an antioxidant treatment [46]. Hence, AgNP-promoted polyploidy formation may relate to the oxidative stress triggered by its interaction with DNA or chromosomes in the cells.
Although we failed to detect the differences of micronucleus rates between the control and AgNP group, a single intravenous exposure to AgNP (5 or 10 mg/kg bw) produced significantly increased micronuclei frequency at 24 h after exposure, and this increase can also be observed 1 and 4 weeks later [3], implying the potential risk of AgNP continuously present in the body. A recent study also showed that SD rats orally administered with AgNP for consecutive 5 days at a range of 5 to 100 mg/kg demonstrated a significant increase of the frequency of micronuclei formation [47]. The negative results of the micronucleus test might relate to the pretreatment of COL and few animal numbers.

Conclusions
The toxicities of nanoparticle-containing medical devices in vivo is currently gain extensive attention, and more knowledge on AgNP is required for safety evaluation and risk management. Taken our data and previous studies together, it is rational to speculate that a lower dose of AgNP with extended exposure period could essentially accumulated in targeted organs and produce chronic toxicity. Furthermore the risk of carcinogenicity may also increase. Major concerns on AgNP's safety assessment at present are its persistence and disposition in the targeted organs and the subsequent toxicities, which necessitates a chronic toxicity study or carcinogenicity study to follow. The distribution of AgNPs in animals can be identified by, for instance, a recently developed method using gold nanocluster as fluorescence probes [48]. In conclusion, this study sheds light on the underlying distribution, targeted organs and genotoxicities particularly generated by AgNP in SD rat models. The specific toxicities and potential carcinogenic effect induced by nanoparticles need to be further investigated in a chronic toxicity study. For example, a AgNP carinogenic test using C57-ras transgenic mouse model, which are currently performing by our team, might be included. The authors gratefully acknowledge Chao Wang, Yanwei Yang and Xin Wang for their