Nephrotoxicity and genotoxicity of silver nanoparticles in juvenile rats and possible mechanisms of action

The silver nanoparticles (Figure 1) used in this research were obtained from the University of Hertfordshire, England, UK. They were suspended in deionised water at a concentration of 1 mg/mL and dispersed with ultrasonic vibration for 20 min before everyday injection. The mean hydrodynamic particle size in suspension was 98.3 nm (range 30.3–178.1 nm due to aggregation) as measured with dynamic light scattering (DLS) using a Zeta-PALS+BI-90 Plus (Brookhaven Instruments Corporation, Hotsville, NY, USA) at a wave length of 659 nm and the scattering angle of 90 °. Surface area, measured with the Brunauer Emmett and Teller (24) method was 114 m²/g. The Zeta potential of -36.58 was measured in the suspension with a combination of laser Doppler velocimetry and phase analysis light scattering (PALS) using Zeta-PALS+BI-90 Plus.

Figure 1

As for other materials we used the reactive oxygen species assay kit (Nanjing Jiancheng, Nanjing, China), rabbit polyclonal anti-TRPC6 antibodies, rabbit polyclonal anti-NADPH oxidase 4 (primary antibody) (Abcam, Cambridge, MA, USA), Alexa 488-conjugated goat anti-rabbit IgG antibodies (secondary antibody, working dilution 1:1000) (Invitrogen, San Diego, CA, USA), rabbit polyclonal anti-β-actin IgG (primary antibody, working dilution 1:1000) (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), chemiluminescent HRP substrate (Immobilon Western, Millipore Corporation, Billerica, MA, USA), haematoxylin and eosin (H&E) staining kit (Solarbio, Beijing, China), creatinine (Cr) assay kit (Nanjing Jiancheng), microalbumin assay kit (Nanjing Jiancheng), total protein assay kit (Nanjing Jiancheng), enzyme-linked immunosorbent assay (ELISA) kit for nicotinamide adenine dinucleotide phosphate oxidase 4 (NOX4) (Shanghai Enzyme-linked Biotechnology Co., Ltd. Shanghai, China), rat IL-6 ELISA kit and rat TNF-α ELISA kit (NeoBioscience, Guangdong, China), rat IL-1β ELISA kit (BioGems, New Jersey, NJ, USA), GelRed^TM nucleic acid gel stain (Biotium, San Francisco, CA, USA), and radioimmunoprecipitation assay (RIPA) lysis buffer (Beyotime, Shanghai, China).

Figure 2 shows the experimental design for this study, in which we used 32 Wistar male rats. The pups were housed with the dam and their litter until weaning on postnatal day 21, after which they were randomly assigned to either the control or one of the three treatment groups, eight to each, and moved to respective cages. All three treatment groups received intraperitoneal injection of AgNP solution with 3 mg of AgNP per kg of body weight (bw) every day. This dose was based on our previous findings of its effects (lower doses showed none) (7) and preliminary analysis. The first group was treated with AgNPs for one week (1-wk-AgNP), the second for two weeks (2-wk-AgNP), and the third for three weeks (3-wk-AgNP). The control group received sterile saline alone.

Figure 2

We observed specific pathogen-free conditions throughout the experiment. The environmental conditions were kept at 22±1 °C and 45–55 % relative humidity, with a 12:12 h light/dark cycle. Clean water and conventional diet were provided ad libitum.

When the rats reached adulthood at eight weeks, they were killed under isoflurane anaesthesia and their blood and urine samples collected and kidneys removed.

All the protocols were approved by the Committee for Animal Care of Tianjin Medical University. All animal experiments complied with the EU Directive 2010/63/EU (25). The experiment was designed to minimise the number of animals used and their suffering.

Depending on the treatment group, the animals were given time to recover from AgNP exposure until they reached adulthood at 8 weeks. Overnight urine samples were collected in metabolic cages and blood samples drawn from the tail vein with a syringe. Urinary microalbumin was detected with the microalbumin assay kit. Serum was harvested after blood centrifugation at 1902 ×g for 10 min and creatinine measured with the creatinine assay kit. All procedures were performed following the kit instructions.

The micronucleus and the single-cell gel electrophoresis (comet) are the most common tests used in vitro or in vivo to investigate genotoxicity of nanoparticles (26, 27, 28). We performed the comet assay as described by Tice et al. (29) and Machado et al. (30). Cell viability was above 90 %, according to the trypan blue dye exclusion method. Samples of peripheral blood were immediately mixed with 0.5 % of low melting point agarose dissolved in phosphate-buffered saline (PBS) and spread on microscope slides precoated with 1.5 % (w/v) of normal melting point agarose. After solidification of the gel (within about an hour), the coverslips were gently removed and the slides immersed in cold (4 °C) lysis solution (2.5 mol/L NaCl, 100 mmol/L EDTA, 1 % (v/v) triton X-100, and 10 mmol/L Tris at pH 10) for 24 h. After that, the slides were placed in a horizontal electrophoresis unit containing freshly prepared electrophoresis buffer (300 mmol/L NaOH and 1 mmol/L EDTA) for 20 min at an electric field strength of 0.78 V/cm (25 V and 300 mA). When electrophoresis was over, the slides were immersed in a neutralisation buffer (0.4 mol/L Tris–HCl, pH 7.5) for 5 min and stained with GelRed^TM (1:10,000). All slides were randomly coded for five animals per group. For each slide (two per rat) we scored 50 cells using the Comet Assay IV^® software (Perceptive Instruments Ltd., Suffolk, UK), and determined tail intensity (% DNA in tail) as the genotoxicity parameter.

The frequency of micronuclei is a characteristic parameter indicative of chromosomal loss and breakage (20). Bone marrow cells of five animals per group were harvested following the method described by Salamone et al. (31) and the micronucleus test performed as described by Schmid (32). After being homogenised and centrifuged at 400×g for 10 min, the cells were resuspended and layered on a microscopic slide. The slides were randomly coded, fixed in absolute methanol for 10 min, and stained with Giemsa. Two slides were prepared per animal. Two thousand polychromatic erythrocytes (PCE) were scored per animal (10,000 per group) for the occurrence of micronuclei. Normochromatic erythrocytes (NCE) were counted in parallel to calculate the PCE to NCE ratio.

After blood and urine sample collection, the rats were killed and their kidneys removed and prepared for single-cell suspensions. Half of the suspension volume was used to determine the levels of ROS, glutathione (GSH), and superoxide dismutase (SOD) following kit instructions. The other half was used to determine protein concentrations.

Inflammatory cytokines (IL-1β, IL-6, and TNF-α) were determined in kidney tissue homogenates with the ELISA kits following kit instructions.

Renal cortex was isolated, minced into fragments, and lysed in a lysis buffer containing a protease inhibitor cocktail (1:1000 dilutions). After preparation of protein extracts, western blotting was conducted as described elsewhere (33). For the analysis, we quantified the intensity of hybridisation signals. Image J program (https://imagej.nih.gov/ij/) was used to evaluate differences between the samples of interest and respective β-actin.

Kidney samples were embedded in the optimum cutting temperature (OCT) formulation (Tissue-Tek, Torrance, CA, USA) after being fixed in 4 % paraformaldehyde for 24 h and dehydrated in 30 % sucrose solution at 4 °C overnight. Seven micron thick slices were cut with a freezing microtome (Leica CM 1850, Leica Microsystems Nussloch, Wetzlar, Germany) and washed in PBS three times for 5 min and then incubated in a blocking buffer with 10 % goat serum for 1 h. After that, the slices were incubated in primary antibody (anti-TRPC6, 1:1000) at 4 °C overnight. Sections were protected from light, washed in PBS three times for 10 min to get rid of non-specific binding, and incubated in Alexa 488-conjugated anti-rabbit IgG. When all the steps were finished, fluorescent images were inspected with a Leica TCS SP5 laser-scanning confocal microscope (40x magnification). Image-Pro Plus software (Version VI; Media Cybernetics, Silver Springs, MD, USA) was used to quantify intensity in the renal slices. For TRPC6 expression analysis, three sections were selected randomly from one kidney sample, and there were eight samples for each group.

Rat kidneys for histological analysis were removed carefully and fixed in a 10 % formalin solution containing neutral PBS. After dehydration with graded ethanol and vitrification with dimethylbenzene, the organs were embedded in paraffin, cut into slices, H&E-stained following the kit instructions, and examined under a light microscope (40x magnification, Nikon, Tokyo, Japan).

Means ± standard errors (SEM) of all data were analysed using one-way analysis of variance (ANOVA) followed by a Newman-Keuls post hoc test. Animal body weights were analysed with two-way ANOVA with groups as between-subject factor and time as repeated measure. All analyses were run with statistical software SPSS v. 17.0. (IBM, Armonk, NY, USA). P<0.05 was considered significant.

No mortality related to AgNP administration was observed during this study. There were no significant differences in body weight between the control and AgNP-treated groups (Figure 3A).

Figure 3

Urinalysis results are shown in Figure 3B. Urinary microalbumin was significantly higher in the 3-wk-AgNP group than controls [F(3, 28)=3.83, P<0.05], but it did not differ significantly from the other two treatment groups. Serum creatinine showed no statistical differences between any of the four groups, although it was slightly higher in the 2- and 3-wk-AgNP group than controls [F(3, 28)=1.67, P>0.05] (Figure 3C).

Significantly higher tail intensity in the 2- and 3-wk-AgNP groups than controls [F(3, 28)=3.07, P<0.05] suggests that the DNA damage depended on the length of nanoparticle exposure.

Figure 3E shows the frequency of micronucleated PCEs in bone marrow, while Figure 3F shows the PCE to NCE ratio. The 3-wk-AgNP group had a significantly higher micronucleus frequency than controls [F(3,28)=3.18, P<0.05], but there was no statistical difference in the PCE to NCE ratio between any of the groups [F(3,28)=2.32, P>0.05].

The 2- and 3-wk-AgNP groups had significantly higher ROS [F(3,28)=3.31, P<0.05] (Figure 4A) and significantly lower GSH and SOD levels in the kidney tissue than control [F(3,28)=3.36, P<0.05] (Figure 4B and 4C).

Figure 4

Figure 4D–F shows inflammatory cytokine levels in the four groups. As expected, IL-6 and IL-1β were significantly higher in the 2- and 3-wk-AgNP groups than control [F(3,28)=3.72 and 3.28, respectively, P<0.05] (Figure 4D and 4E), while the level of TNF-α was significantly higher only in the 3-wk-AgNP group [F(3,28)=3.16, P<0.05] (Figure 4F). We also observed a coincidence in trends between inflammatory cytokines and ROS levels, which suggests a link between inflammation and oxidative stress.

The 2- and 3-wk-AgNP groups showed significantly higher expression of NOX4 than controls [F(3,28)=3.09, P<0.05] (Figure 5B), which suggests that higher ROS production was at least partly owed to the mobilisation of NOX4. At the same time, TRPC6 expression was significantly higher only in the 3-wk-AgNP group [F(3,28)=2.98, P<0.05] (Figure 5D).

Figure 5

Figure 6A shows TRPC6 in the glomeruli and tubules of the kidney cortex and its fluorescent intensity (Figure 6B), which was significantly higher only in the 3 -wk-AgNP group compared to control [F(3,28)=3.23, P<0.05].

Figure 6

H&E staining revealed pathological changes of several glomeruli in the 3-wk-AgNP group, which included glomerular cell proliferation, thickened basement membrane, fusion of glomerulus and renal capsule, and proliferation in the mesangial cell and mesangial matrix. The interstitial areas were infiltrated with inflammatory cells. Renal tubules did not show pathological changes (Figure 7).

Figure 7