2019 Volume 44 Issue 11 Pages 737-751
Industry demand for nanomaterials is growing, but metal nanoparticle toxicity is not fully understood. For example, nickel nanoparticles (NiNPs) are used in electric capacitors, and their consumption is increasing, but there have been few reports of their toxicity and environmental effects. To elucidate the toxicological characteristics of NiNPs, we investigated their effects on the histopathology and oxidative states of zebrafish (Danio rerio) and compared the results with those of ionic nickel. Zebrafish exposed to four different concentrations of NiNPs or NiCl2 for 72 hr or 7 days were subjected to histopathological analysis, and tissue samples were subjected to analyses for oxidative stress and gene expression. High concentrations of both NiNPs and NiCl2 caused tissue damage in the gills, digestive tract, and liver. The damage was typically characterized by epithelial degeneration and necrosis in the gills, esophagus, and intestines, as well as by lipid loss and palisade pattern degradation in the liver. The damages to the gills, esophagus, and intestines were more severe after exposure to NiNPs, but exposure to NiCl2 led to more severe liver damage. Exposure to NiNPs increased lipid peroxidation in the skin but decreased it in the liver and intestines; exposure to NiCl2 increased lipid peroxidation in the intestines. Only exposure to NiCl2 changed antioxidative responses, enzymatic antioxidant activities, and metallothionein gene expression. These results indicate that NiNPs, which are highly adsorptive, cause severe damage to the epithelium by physical contact with the cell surface and production of reactive oxygen spices, whereas ionic nickel, which is absorptive, affects cellular antioxidative responses by absorption into the body and delivery to the liver.
Nickel is a relatively abundant metal in Earth’s crust. Because of its unique combination of hardness, toughness, and stability (which is comparable to that of iron), nickel has been used in many applications such as implants, batteries, metallic catalysts, and the constituent of coins and stainless steel. Many anthropogenic activities involved in nickel production such as mining, smelting, refining, and processing increase nickel emissions into the environment (Muñoz and Costa, 2012). Nickel is also an essential nutrient for animals and plants. Though nickel is stable like iron and acts as a catalyst in some organisms, excessive discharges of nickel into the environment may harm wildlife and ecosystems. Many studies have shown that nickel induces pathological changes in mammals, including lung cancer, inflammation, and allergic reactions (Muñoz and Costa, 2012). In addition, Kong et al. (2014) have shown that nickel may induce reproductive and developmental abnormalities in rats. However, compared with other metals, relatively few studies have examined nickel toxicity, and most have focused on its toxicity to mammals. Therefore, further studies are needed to evaluate the toxicity of nickel to other organisms (Kubrak et al., 2013).
Nanoparticles are increasingly being used because of their useful and unique properties, but their high rate of consumption and disposal may be exerting adverse effects on wildlife and ecosystems. Nickel nanoparticles (NiNPs) are one kind of metallic nanoparticle, and are widely used in electric capacitors and medical implants (Jayaseelan et al., 2014). Widespread use of these products could increase the risk of environmental exposure to NiNPs and their toxic effects. However, regardless of the increasing global demand for NiNPs, assessments of NiNP safety have been inadequate, and their chemical and toxicological characteristics are not fully understood.
Nanoparticles can penetrate the pulmonary system, digestive tract, or skin, and can reach secondary organs through the bloodstream (Fushimi et al., 2008). The target organs of nanoparticles have often been investigated in mice and rats, but there have been few such studies in fish or wildlife. Some toxic pollutants like heavy metals are known to induce oxidative stress by generating reactive oxygen species (ROS) in animals, including fish. ROS react with many endogenous substances such as proteins, nucleic acids, and lipids. By oxidizing these substances they impair the functionality of proteins, DNA, and membrane lipids (Leonard et al., 2004). To deal with oxidative stress, organisms use a defense system that includes antioxidant enzymes such as catalase (CAT), superoxide dismutase (SOD), glutathione peroxidase, glutathione reductase, and glutathione S-transferase (GST). In addition to these defense systems, metallothionein acts as another antioxidant by binding and detoxifying heavy metals and ROS. Past studies have demonstrated that oxidative stress can be caused by the production of ROS induced by ionic metals, and recent studies have reported that nanoparticles also induce oxidative stress (Ahamed, 2011). However, it is unknown whether NiNPs induce ROS production, and there is incomplete understanding of the differences between NiNPs and ionic nickel with respect to the mechanisms of toxicity and target organs.
The zebrafish is a vertebrate animal model widely used in toxicity tests of chemicals. The zebrafish is especially useful in ecotoxicological studies for estimating the influence of chemical emissions to the environment. Zebrafish have been used in many studies to evaluate the toxicity of nanoparticles. For example, Duan et al. (2013) have shown that exposure to silica nanoparticles produces adverse effects, including ROS production, lactate dehydrogenase release, and head deformities in embryos. ZnO nanoparticles have also been shown to increase ROS production, resultant antioxidant responses, and DNA damage (Zhao et al., 2013).
The objectives of the present study were to assess the toxicity of NiNPs in zebrafish and to differentiate the toxicological properties of NiNPs and ionic nickel. We observed histopathological changes in zebrafish exposed to NiNPs or NiCl2 to determine the target organs of each substance. We then explored the mechanistic basis for the toxic effects of NiNPs and ionic nickel by examining oxidative stress as measured by lipid peroxidation, several antioxidant enzymatic activities, and metallothionein gene expression.
Zebrafish (Danio rerio) were chosen as the test organism. The fish were obtained from a local aquatic breeder (Masuko Suikei, Saitama, Japan) and acclimated to laboratory conditions for one week prior to the experiments. The water was maintained at a temperature of 28 ± 1°C, and light was provided on a 14 hr light/10 hr dark cycle. Experimentation protocols in this study followed the rules set by Institutional Animal Care and Use Committee of Kitasato University under the Act on Welfare and Management of Animals in Japan.
NiNPs (> 99.9%, 40 nm) and NiCl2 were purchased from EM Japan (Tokyo, Japan) and Tokyo Chemical Industry (Tokyo, Japan), respectively. NiNP solutions were prepared according to the method of Jayaseelan et al. (2014). NiNPs and NiCl2 were dissolved with ultrapure water to prepare their respective stock solutions of 1.0 mg/mL. NiNPs in the stock solution were dispersed by sonication for 6 hr in a bath-type sonicator (100 W, 42 kHz). Four different concentrations of NiNP and NiCl2 test solutions were prepared by diluting the respective stock solutions with breeding water and adjusting the concentrations to 0, 0.1, 2.5, and 10 mg/L (NiNPs: 0, 1.7, 42.6, and 170 μM; NiCl2: 0, 0.77, 19.3, and 77 μM). NiNP test solutions were sonicated with a hand sonicator for 10 min prior to exposure.
Seven adult fish selected randomly from acclimated fish were maintained in a 1-L glass aquarium and exposed to test solutions for 72 hr or 7 days under aeration. The fish were not fed from the day before the onset of the exposure period to the end of the exposure period to avoid Ni ingestion from food. Static conditions were applied for the exposures, and test solutions were completely replaced every day to maintain test substance concentrations.
To confirm changes in test substance concentrations during exposure, nickel concentrations were measured in test solutions before and after exposure using an inductively coupled plasma atomic emission spectrometer (ICPE-9000, Shimadzu Corporation, Kyoto, Japan). The measurements were performed twice for NiNP and NiCl2 test solutions at each concentration. First, the supernatants from the respective test solutions were collected to detect dissolved and dispersed nickel. Second, test solutions were dispersed by sonication prior to collection to determine the total nickel content, including nickel particles precipitated and adsorbed to the aquarium wall. Test solutions (20 mL) were mixed with 4.52 mL of 60% nitric acid and 3.43 mL of 35% hydrochloric acid and diluted to 40 mL with ultrapure water. The mixtures were heated at 90°C for 2 hr, diluted to 50 mL with 5% nitric acid, and then put into the ICPE-9000. Standard solutions (0, 1.0, 3.0, 5.0, and 7.0 mg/L) of NiNPs (0, 17, 51, 85, and 119 μM) and NiCl2 (0, 7.7, 23, 38.6, and 54 μM) were also measured in the same way, and NiNP and NiCl2 concentrations in the test solutions were calculated from standard curves derived from their respective standard solutions.
After 7 days of exposure, fish were euthanized and fixed in Davidson’s solution (Kumagai et al., 2006; Miwa, 2000; Wang et al., 2016) for 64 hr. The fish were dehydrated, cleared, and then embedded in paraffin by routine methods. The paraffin blocks were cut into thin sections (2 μm) and stained with hematoxylin and eosin. Pathological changes were observed with a light microscope.
Gills, intestines, livers, and skins from fish exposed to test solutions for 72 hr or 7 days were sampled and homogenized in 100 μL of phosphate buffer (50 mM, pH 7.0). The tissue homogenates were then centrifuged for 10 min at 10,000 × g (4°C), and the supernatants were collected and used in the following assays for oxidative stress. Tissue samples were homogenized in 150 μL of glycine-HCl buffer (0.5 M, pH 3.6) for lipid peroxidation. The homogenates were centrifuged for 10 min at 10,000 × g (4°C), and the supernatants were used for the lipid peroxidation assay.
Lipid peroxidation was measured with the thiobarbituric acid reactive substances (TBARS) assay according to the method of Asakawa et al. (1975). Aldehydic secondary products of lipid peroxidation are generally accepted as markers of oxidative stress. A lipid peroxidation product, malondialdehyde (MDA), forms an adduct with thiobarbituric acid (TBA), and the MDA-TBA adduct can be measured colorimetrically. The prepared samples (50 μL) were mixed with 150 μL of 0.36% TBA solution and 80 μL of 95% ethanol. The mixtures were heated in boiled water for 15 min and then mixed with 100 μL of 35% trichloroacetic acid and 200 μL of chloroform. The absorbance of the resultant mixtures was measured at 535 nm using a spectrophotometer, and TBARS levels in tissue samples were determined from an MDA equivalence standard and expressed as μmol MDA/mg protein.
Superoxide dismutase (SOD) activity was measured with a SOD Assay Kit (WST kit, Dojindo Laboratory, Kumamoto, Japan) according to the manufacturer’s instructions. The SOD activity was measured in serially diluted samples, and one unit of SOD activity (U) was defined as the amount of the enzyme in 20 μL of diluted sample showing 50% inhibition of the enzyme activity. Then the SOD activity was calculated from the dilution ratio and expressed as U/mg protein.
Catalase (CAT) activity was measured by the method of Aebi (1983). Sample solutions (20 μL) in phosphate buffer were mixed with 1 mL of 30 mM H2O2 and 1 mL of phosphate buffer. Changes in the absorbance of the mixture were measured at 240 nm for 2 min using a spectrophotometer to determine reduction of H2O2 concentration. Catalase activity in samples was calculated as nmol H2O2 consumed/min/mg protein.
Glutathione S-transferase (GST) catalyzes glutathione (GSH) conjugation to xenobiotic substances, including peroxides, for detoxification. A GST substrate, 1-chloro-2,4-dinitrobenzene (CDNB), is conjugated efficiently to GSH, and the resultant conjugate exhibits maximum absorption at 340 nm. Sample solutions (20 μL) were added to 1 mL of phosphate buffer and mixed with 120 μL of 2 mM GSH and 60 μL of 1 mM CDNB. Changes in the absorbance of the mixture were measured at 340 nm for 5 min using a spectrophotometer to determine GSH-CDNB conjugate production. GST activity in samples was expressed as μmol CDNB conjugated/min/mg protein.
The total protein content of samples was measured for the specific activities of the respective enzymes using a BCA Protein Assay Kit (Takara Bio Inc., Shiga, Japan) according to the manufacturer’s instructions. Samples (10 μL) were added to 200-μL working solutions and incubated at 37°C for 30 min. The absorbance of samples at 562 nm was measured using a spectrophotometer, and protein levels were determined from a protein standard curve using bovine serum albumin.
Livers sampled from fish exposed to test solutions for 72 hr were maintained in RNAlater™ solution (Thermo Fisher Scientific Inc., Waltham, MA, USA) at −30°C prior to RNA extraction. Total RNA was extracted from liver samples using a ReliaPrep™ RNA Tissue Miniprep System (Promega, Madison, WI, USA) according to the manufacturer’s instructions. The total RNA (400 ng) with 0.5 μg oligo (dT)15 primer (Promega) and 40 units RNasin® Plus RNase Inhibitor (Promega) was incubated at 70°C for 5 min in a total volume of 10 μL. After incubation, 1.25 μL of 10 mM dNTP Mixture (Promega), 5 μL of M-MLV 5X Reaction Buffer (Promega), and 200 units M-MLV Reverse Transcriptase (Promega) were added to the mixtures. Reverse transcription was performed at 37°C for 60 min and 95°C for 5 min in a total volume of 25 μL.
Real-time PCR was performed with a PowerUp™ SYBR® Green Master Mix (Thermo Fisher Scientific). One microliter of cDNA (16 ng) was amplified with 500 nM each of the specific primer combinations for target genes, metallothionein-2 (mt2) and ribosomal protein L7 (rpl7), and 5 μL 2 × PowerUp™ SYBR® Green Master Mix in a total volume of 10 μL. PCR conditions were as follows: 40 cycles of 95°C for 3 sec and 60°C for 30 sec. Real-time PCR analysis of each sample was carried out according to a relative standard curve method using a StepOnePlus™ Real-Time PCR System (Thermo Fisher Scientific). Standard curves were obtained from serially diluted sample mixtures, and the expression levels of samples were measured using the standard curves. The rpl7 was chosen as an endogenous control, because its expression level is stable between fish and is unrelated to treatments (Lang et al., 2016). Primers for the respective target genes (listed in Table 1) were designed to anneal at 60°C.
Statistical analysis was carried out using Microsoft Excel 2013 (Microsoft Corporation, Redmond, WA, USA) and freeware analysis tools (MEPHAS, Osaka University, Osaka, Japan). Differences in the proportion of observed histopathological changes between control and exposed groups were tested using Fisher’s exact test. For the oxidative stress and metallothionein gene expression results, differences between control and exposed groups were tested using Dunnett’s multiple comparison test. The antioxidant analysis data were not normally distributed, and statistical analysis was therefore carried out using Steel’s test. Data were expressed as means ± SDs. P values less than 0.05 were considered statistically significant.
NiNP concentrations in test solution supernatants, which were set at 0.1, 2.5, and 10 mg/L, were determined to be 0.11, 1.91, and 9.07 mg/L, respectively, before exposure and 0.04, 1.46, and 4.95 mg/L, respectively, after exposure. NiCl2 concentrations in supernatants were 0.12, 2.37, and 10.53 mg/L before exposure and 0.12, 2.05 and 9.59 mg/L after exposure. When test solutions were collected after sonication, the NiNP concentrations were 0.13, 2.21, and 9.46 mg/L before exposure and 0.11, 1.79, and 10.42 mg/L after exposure. The corresponding NiCl2 concentrations were 0.18, 2.52, and 10.70 mg/L before exposure and 0.11, 2.17, and 8.33 mg/L after exposure. Because NiNP concentrations were considerably lower in unsonicated supernatants after exposure but not in sonicated test solutions, these reductions could be attributed to NiNP precipitation and adsorption to the aquarium walls. Consequently, the actual NiNP concentrations were probably much lower than the predetermined concentrations and may have been comparable to the concentrations measured here. The NiNP concentration measured in the sonicated test solution with the highest predetermined concentration (10 mg/L) was higher after exposure than before exposure. This increase may have been due to the NiNPs being dispersed inhomogeneously in the solution by sonication because of the physical form of the NiNPs. The NiCl2 concentrations in most cases remained at predetermined levels during exposure, although the NiCl2 concentration decreased somewhat in the sonicated test solution with the highest predetermined concentration.
Tissue damage caused by zebrafish exposure to NiNPs and NiCl2 was observed in the gills, esophagus, intestines, and liver. Table 2 summarizes the observed histopathological changes and test solution concentrations.
Asterisks indicate significant differences in the proportion of observed histopathological changes from controls (p < 0.05).
Cell swelling and necrosis in secondary lamellae as specific alterations in gill morphology were observed in both NiNP-exposed and NiCl2-exposed fish (Fig. 1). NiNP-exposed fish experienced exfoliation in the secondary lamellae. NiCl2-exposed fish experienced hyperplasia in the primary and secondary lamellae. NiNPs caused significant damage at lower concentrations (2.5 and 10 mg/L) than did NiCl2 (10 mg/L).
Representative histological changes in zebrafish gills. Cell swelling and necrosis (white arrow) in secondary lamellae were observed in gills from fish exposed to 10 mg/L of either NiNPs or NiCl2. Exfoliation (black arrow) was observed in gills from fish exposed to NiNPs. Hyperplasia in primary and secondary lamellae (arrowheads) was observed in fish exposed to NiCl2.
In both exposure groups, we observed major alterations of the esophagus, including degeneration and exfoliation of epithelium (Table 2, Fig. 2). Because mucosa cell swelling and collapse were severe in NiNP-exposed fish, the esophageal lumen became obstructed by mucus at concentrations of 2.5 and 10 mg/L. Exposure to NiCl2 resulted in degeneration and exfoliation of epithelium, except for obstruction of the lumen, at concentrations of 0.1, 2.5, and 10 mg/L. However, the extent of damage was more severe in fish exposed to NiNPs versus NiCl2 at concentrations of 2.5 and 10 mg/L.
Representative histological changes in zebrafish esophagi. Epithelial degeneration and exfoliation (black arrow) were observed in the esophagus of fish exposed to 10 mg/L of either NiNPs or NiCl2. In addition, mucosa cell swelling and collapse in fish exposed to NiNPs led to obstruction of the esophageal lumen (white arrow).
In the intestines, epithelial cells had irregular shapes and were partially degenerated in fish exposed to 2.5 and 10 mg/L of NiNPs (Table 2, Fig. 3). In addition, black NiNPs accumulated in the intestinal tracts of NiNP-exposed fish. NiCl2 at the highest concentration (10 mg/L) damaged intestines, and hyperplasia of intestinal villi was observed (Table 2, Fig. 3). An increase of goblet cells was only observed in one fish exposed to 2.5 and 10 mg/L NiCl2, respectively.
Representative histological changes in zebrafish intestines. Irregularly shaped, partially degenerated epithelial cells (black arrow) and collections of black NiNPs in the intestinal tract (white arrow) were observed in fish exposed to 10 mg/L of NiNPs. Villous hyperplasia and an increase in goblet cells were observed in fish exposed to NiCl2.
Both NiNP- and NiCl2-exposed fish exhibited a lack of hepatic lipid storage and degeneration of the palisade pattern throughout the liver. NiCl2-exposed fish exhibited eosinophilic cytoplasms (Table 2, Fig. 4). Hepatic damage was observed at lower NiCl2 concentrations (2.5 and 10 mg/L) compared with NiNPs (10 mg/L).
Representative histological changes in zebrafish livers. Hepatocyte lipid loss and palisade pattern degeneration were observed in fish exposed to 10 mg/L of either NiNPs or NiCl2. Eosinophilic cytoplasms (black arrow) were observed in the livers of fish exposed to NiCl2.
After 72 hr of exposure, MDA, a lipid peroxidation product, significantly increased in the intestines of fish exposed to 2.5 and 10 mg/L of NiCl2 and in skin at the highest NiNP concentration (10 mg/L) compared with the vehicle control (Fig. 5a). However, NiNPs significantly inhibited lipid peroxidation in liver at 0.1 mg/L and in intestines at 10 mg/L. After 7 days of exposure, the concentration of MDA significantly increased in the skin of fish exposed to 0.1 and 10 mg/L of NiNPs, but it decreased in liver at 2.5 mg/L and in intestines at 0.1 and 10 mg/L of NiNPs (Fig. 5b). Exposure to NiCl2 did not lead to a statistically significant difference in MDA concentration in any tissues, although it showed a tendency to decrease MDA in liver and intestines.
Lipid peroxidation in the liver, gills, intestines, and skin of zebrafish exposed to either NiNPs or NiCl2 for 72 hr (a) and 7 days (b). Lipid peroxidation was quantified in terms of malondialdehyde (MDA) concentrations. X-axis values represent NiNP or NiCl2 concentrations. Error bars show standard deviations. Asterisks indicate significant differences from controls (p < 0.05).
After 72 hr of exposure, SOD activity was significantly inhibited in the intestines of fish exposed to 10 mg/L of NiCl2 (Fig. 6a). After 7 days of exposure, SOD activity was inhibited in the liver of fish exposed to 2.5 and 10 mg/L of NiCl2 compared with controls and in the intestines of fish exposed to the highest NiCl2 concentration (Fig. 6b). However, SOD activity increased in the skin of fish exposed to 10 mg/L of NiCl2. Exposure to NiNPs did not change SOD activity in any tissues.
Superoxide dismutase activity (U) in the liver, gill, intestines, and skin of zebrafish exposed to either NiNPs or NiCl2 for 72 hr (a) and 7 days (b). Elements of the box plot indicate the following: upper horizontal line, 75th percentile; middle horizontal line, median; lower horizontal line, 25th percentile; vertical line extent, minimum to maximum. X-axis values represent concentrations of NiNPs or NiCl2. Asterisks indicate significant differences from controls (p < 0.05).
After 72 hr of exposure, CAT activity was significantly inhibited in the liver of fish exposed to 10 mg/L NiCl2 and in the skin of fish exposed to 0.1, 2.5, and 10 mg/L NiCl2 (Fig. 7a). However, SOD activity was accelerated in the intestines of fish exposed to 2.5 mg/L NiCl2. After 7 days of exposure, CAT activity was inhibited in the skin of fish exposed to 2.5 and 10 mg/L NiCl2 (Fig. 7b). Exposure to NiNPs did not change CAT activity in any tissues.
Catalase activity in the liver, gills, intestines, and skin of zebrafish exposed to NiNPs or NiCl2 for 72 hr (a) and 7 days (b). Elements of the box plot indicate the following: upper horizontal line, 75th percentile; middle horizontal line, median; lower horizontal line, 25th percentile; vertical line extent, minimum to maximum. X-axis values represent concentrations of NiNPs or NiCl2. Asterisks indicate significant differences from controls (p < 0.05)
No significant differences were observed in the GST activities of any tissues at any NiNP or NiCl2 concentrations (data not shown).
Expression of the mt2 gene increased in the liver of fish exposed to 0.1 and 10 mg/L NiCl2 for 72 hr compared with the control (Fig. 8). Exposure to NiNPs did not change mt2 gene expression.
Changes in metallothionein-2 (mt2) gene expression in zebrafish livers associated with exposure to NiNPs or NiCl2. Gene expression levels, detected by real-time RT-PCR and normalized against rpl7 expression, is presented relative to a vehicle control. X-axis values represent concentrations of NiNPs or NiCl2. Values are presented as means ± SDs. Asterisks indicate significant differences from control (p < 0.05).
Nanoparticles discharged by human activities are now present throughout much of the environment and may have adverse effects on wildlife and ecosystems. Thorough studies are therefore needed to ascertain nanoparticle toxicities and chemical characteristics. It is known that nickel harms the gills and liver of fish (Jayaseelan et al., 2014), but few studies have revealed where and how nickel damages tissues in the bodies of fish. The present study quantified histopathological changes, lipid peroxidation, antioxidant responses, and metallothionein gene expression in zebrafish exposed to NiNPs or NiCl2 to characterize the absolute and relative toxicities of these two substances.
The histopathological analyses revealed that both NiNPs and NiCl2 affected zebrafish gills, esophagus, intestines, and livers. All of these organs are involved in pollutant uptake and excretion. Gills, for example, are not only essential for fish respiration, but are also important for controlling osmotic pressure, acid–base regulation, and the metabolism and excretion of xenobiotics and endogenous substances. The gill is among the principal organs that first encounter xenobiotics, and its role, inter alia, is to prevent invasion by them. Similarly, the esophagus is the site of first encounter for swallowed pollutants, and mucosal epithelial cells defend the organism from the damaging effects of toxic pollutants. Because the intestine is an absorptive organ, orally ingested pollutants are likely to be absorbed there. Any pollutants absorbed from the intestines are delivered to the liver and undergo hepatic metabolism. The liver also plays a role in the metabolism and storage of energy reserves such as glycogen and fat. Normal liver cells form a palisade structure, and their cytoplasm is filled with glycogen and lipid (Brusle and Anadon, 1996). The histopathological changes observed in this study were closely connected to the structural and functional characteristics of these organs in zebrafish, and also to the physical and chemical properties of the tested materials.
Damage to gills and esophagus included epithelial degeneration and exfoliation and was more severe in fish exposed to NiNPs than to NiCl2. In addition, irregularly-shaped intestinal epithelia occurred only in fish exposed to NiNPs. Considering that NiNPs are particles (not ions) and tend to adsorb to surfaces, it is likely that the damage caused by exposure to NiNPs occurred as a result of physical contact with epithelial cells rather than particle absorption into the cells. Nickel ions are easily taken into cells (Edwards et al., 1998), where Zn transporters, perhaps members of the Zrt- and Irt-like protein family, are thought to regulate Ni influx (Onodera et al., 2018). NiCl2 is easily absorbed into the epithelium and therefore has less influence on the surface of epithelia. The degrees of degeneration and exfoliation caused by NiCl2 were therefore mild compared with those caused by NiNPs. Damage in the liver was observed at lower concentrations and was more severe in fish exposed to NiCl2 versus NiNPs. Ionic nickel absorbed and delivered to the liver appears to injure hepatocytes directly. Though exposure to NiNPs also caused liver damage, this damage may have been due to a decline in the ability of the fish to absorb and use nutrients as a result of epithelial degeneration and necrosis in their intestines. In other words, an insufficient nutrient supply may have caused mitochondrial dysfunction and resulted in hepatic degeneration.
The toxicity of inorganic nickel has been thoroughly studied, and it is possible that nickel ions released from NiNPs caused the tissue damage observed in this study. We did not measure the amount of nickel ions released from NiNPs into the breeding water or digestive tract. However, NiNPs are considered to be insoluble in water. Experimental studies have shown that little dissolution of NiNPs occurs in aqueous media over time periods comparable to the duration of this study; Kanold et al. (2016) reported nickel ion dissolution amounts from 3 mg/L of NiNPs (< 100 nm diameter) of less than 2% after 24 hr in artificial seawater, and Latvala et al. (2016) found that only about 1–3% of the nickel contained in 10 μg/L of NiNPs (< 100 nm) was released into a cell culture medium (supplemented Dulbecco’s Minimal Essential Medium, pH 7.4) after 24 hr. Because the concentration of nickel ions resulting from a release of 3% of the nickel in the NiNP test solutions at the highest concentration studied here (10 mg/L) would be only 5.1 μM—equivalent to the nickel concentration in a solution of 0.66 mg/L NiCl2—it is unlikely that nickel ions were primarily responsible for the tissue damage and other changes observed in NiNP-exposed fish. Although Latvala et al. (2016) reported that substantial NiNP dissolution can occur at low pH (4.5) in the presence of complexing agents, zebrafish lack a stomach and gastric glands, and the pH of the zebrafish intestines therefore never becomes as low as 7.5 (Brugman, 2016; Nalbant et al., 1999). It thus appears that the concentration of nickel ions released from NiNPs into the esophagus and intestines of zebrafish would be very low, but with the caveat that there is no information about the presence of complexing agents in the digestive tracts of these fish. It is reasonable to assume that the damage due to exposure to NiNPs was caused mainly by physical contact with the nanoparticles and only slightly, if at all, by nickel ions.
Oxidative stress can be induced by ROS overproduction. ROS are produced during energy production by oxidative phosphorylation in mitochondria. ROS are also produced to maintain physiological functions when inflammation occurs because of exposure to chemicals and infection. However, if severe inflammation provokes an overproduction of ROS, the excess ROS target their host cells and lead to cellular dysfunction. Aerobic organisms generally have antioxidants to prevent oxidative stress, and the cellular oxidant/antioxidant balance is maintained. If this balance is disrupted, the cytotoxicity of ROS can trigger various disorders and diseases, including neurodegeneration and diabetes (Niki, 1992; Pisoschi and Pop, 2015). Lipid peroxidation is a process in which fatty acids are oxidized by ROS, and the resulting lipid peroxides have oxidizing activity and cytotoxicity comparable to those of ROS. Therefore, lipid peroxide generation secondarily injures cells (Fujita, 2002). Past studies have reported that heavy metals, including nickel, increase ROS production and lipid peroxidation in the liver of Carassius auratus (Zheng et al., 2014), and NiNPs also enhance lipid peroxidation in human epithelial cells (Siddiqui et al., 2012). The present study shows that both NiNP and NiCl2 exposure affects lipid peroxidation. NiNPs accelerated lipid peroxidation quantified in terms of the MDA concentration in skin at both 72 hr and 7 days, but inhibited lipid peroxidation in the liver and intestines. Increased lipid peroxidation in the skin is consistent with the strong adsorption affinity of NiNPs noted above. Because the skin is exposed to the external environment, NiNPs can easily adsorb onto the surface of the skin and could enhance lipid peroxidation. In contrast, the decreased lipid peroxidation in the liver and intestines of NiNP-exposed fish may have been caused by tissue damage and an imbalance of the oxidant/antioxidant system. Because the supply of nutrients to the liver is reduced by intestinal tissue damage caused by NiNPs, an insufficient nutrient supply may lead to mitochondrial dysfunction and a resultant reduction of ROS production and lipid peroxidation. Another explanation for the reduction of lipid peroxidation in the liver is that NiNPs might function as radical scavengers in the liver. A previous study has indicated the possibility that other metals like chromium possess the ability to inhibit lipid peroxidation (Ueno et al., 1988). Seventy-two hours of exposure to NiCl2 accelerated lipid peroxidation in the intestines, but conversely, exposure for 7 days tended to inhibit lipid peroxidation in the liver and intestines. Because the intestine is an absorptive organ, NiCl2 appears to be efficiently absorbed in the intestine, where it increases ROS production and lipid peroxidation at early stages of exposure. The resultant continuous overproduction of ROS may damage the intestines, deteriorate liver function, and eventually reduce lipid peroxidation in both the intestines and liver at the end of the exposure.
Antioxidants are indicators of oxidative stress as well as lipid peroxidation. One antioxidant, SOD, has three isoenzymes and plays a role in suppressing ROS cytotoxicity by efficiently converting superoxides to H2O2. CAT and GST are also antioxidants and protect cells against oxidative stress. CAT contributes to detoxification by decomposing the H2O2 produced by SOD into oxygen and hydrogen. GST plays a role in detoxification and excretion by catalyzing the conjugation of xenobiotic substances, including peroxides, into an antioxidant, GSH, via a sulfhydryl group (Eguchi et al., 2009). As with lipid peroxidation, antioxidant-mediated detoxification is affected by exposure to both heavy metals and nanoparticles (Jayaseelan et al., 2014). In the present study, significant differences in antioxidant enzymatic activities were observed only in the tissues of fish exposed to NiCl2. SOD activity increased in skin at a high NiCl2 concentration (10 mg/L) after 7 days of exposure, but decreased in the liver and intestines after both 72 hr and 7 days of exposure. CAT activity also decreased in the liver at 72 hr and skin at both 72 hr and 7 days. However, CAT activity was accelerated in the intestines at 72 hr. The decreased SOD activity observed in the liver and intestines may have resulted from tissue damage, as described for the results of lipid peroxidation. Because NiCl2 did not apparently affect intestinal morphology, the cytotoxicity of NiCl2 may have caused the general deterioration of cellular physiology, including the inhibition of SOD and lipid peroxidation activities before morphological changes were apparent. The activity patterns of SOD and CAT were inconsistent in the intestines and skin of NiCl2-exposed fish. NiCl2 tended to increase lipid peroxidation in these tissues, even though there were no significant differences in lipid peroxidation in the skin. Therefore, although oxidative stress occurs in NiCl2-exposed tissues, the antioxidant response appears to differ between SOD and CAT. In contrast, NiNPs did not noticeably affect antioxidant enzymatic activities. The histopathological changes caused by NiNPs reflect the fact that their strong adsorption affinity can lead to damage to the epithelium of tissues they contact. NiNPs cannot therefore penetrate cells or reach the liver, nor can they cause changes in antioxidant production, which is performed in organelles like the mitochondrion or peroxisome.
Metallothionein is a stable and cysteine-rich protein that is found mainly in the liver, kidneys, intestines, and pancreas. Because of its strong affinity for many metals, metallothionein efficiently binds to heavy metals, regulates their concentrations in the body, and therefore plays an important role in their detoxification. Because metallothionein production is induced by heavy metal exposure and oxidative stress (Kito et al., 1984; Min et al., 2005), metallothionein genes and proteins are used as biomarkers for the presence of heavy metals and oxidative stress and for induction of a defensive response to metals. In the present study, metallothionein synthesis increased at the gene level in the livers of fish exposed to 0.1 and 10 mg/L NiCl2 but not to NiNPs. These results indicate that metallothionein synthesis is more easily induced by NiCl2 than by NiNPs, and that NiCl2 or secondarily generated oxidative stress induces this detoxification response. Boran and Şaffak (2018) have reported that NiCl2 and NiNPs activate methallothionein-2 gene expression in zebrafish larvae and that NiCl2 is a more potent inducer of methallothionein-2 expression than NiNPs. Metallothionein appears to be able to bind to and detoxify most of the ROS induced by NiCl2 at 72 hr. However, the severity of damages at 7 days suggests that the binding capacity of metallothionein may have been exceeded, thereby inhibiting SOD and lipid peroxidation. Because metallothionein is not abundant in the skin, it has a limited ability to reduce oxidative stress in skin tissue. Antioxidant enzymes such as SOD may be activated as part of a compensatory response. As mentioned above, NiNPs have strong adsorption affinity but cannot penetrate cells. They therefore do not appear to enhance metallothionein gene expression or antioxidant activities.
Our study reveals that NiNP exposure is harmful to fish. Comparison with the effects of NiCl2 exposure revealed differences between NiNPs and NiCl2 in the distribution of damaged tissues and the degree of damage. We also attempted to identify the toxic mechanisms of NiNPs and NiCl2 and their detoxification mechanisms. Although we succeeded in elucidating the toxicological properties of these forms of nickel in terms of oxidative stress, more work remains to be done. Because physical properties are important when evaluating nanoparticle toxicity (Hirose, 2013; Utembe et al., 2015), the present study, which revealed the unique characteristics of NiNPs, has indicated that it may be necessary to reevaluate the methodologies and results of other toxicological assessments. Appropriate and safe nanoparticle use will require thorough toxicity studies, the results of which should be incorporated into policy.
The authors thank Dr. Rieko Takamatsu and Dr. Akira Kubota for their skillful technical assistance and helpful comments.
The authors declare that there is no conflict of interest.