Nathaniel J.
Clark
a,
Robert
Clough
b,
David
Boyle
a and
Richard D.
Handy
*a
aSchool of Biological and Marine Sciences, University of Plymouth, Plymouth, UK. E-mail: r.handy@plymouth.ac.uk
bAnalytical Research Facility, School of Geography, Earth and Environmental Sciences, University of Plymouth, Plymouth, UK
First published on 11th October 2019
To determine bioaccumulation potential, suitable methods to extract and quantify engineered nanomaterials (ENMs) from fish tissues are required. The aim was to develop single particle inductively coupled plasma mass spectrometry (spICP-MS) protocols for liver tissue from rainbow trout and to assess the suitability of enzymatic (proteinase K) and alkali (tetramethylammonium hydroxide; TMAH) methods of extraction. A total of four different extractants were used: proteinase K or TMAH both with/without CaCl2. Spike recovery tests using equal mass concentrations (50 ng L−1 of Ag) as silver nanoparticles (Ag NPs) or AgNO3 were conducted. Extractants alone spiked with Ag NPs showed recovery similar to ultrapure deionised water (95–105%). However, the TMAH alone caused AgNO3 to precipitate and was not a suitable extractant. A second series of experiments looked at spike recovery on liver tissue samples. Proteinase K, with or without CaCl2, failed to completely digest the tissues. Only TMAH + CaCl2 demonstrated the ability to solubilise the liver. Ag NPs spiked onto liver tissues and analysed 24 h later, also showed no significant change in particle size distribution or particle mass concentration compared to those freshly spiked without liver present. The particle number concentration fell significantly to around 80% of the freshly spiked Ag NPs. Samples from an in vivo dietary study where fish were fed nominally 100 mg kg−1 Ag as either AgNO3 or Ag NPs were analysed to demonstrate the utility of the method. There was no significant difference between the particle number concentration, mean particle size or particle mass concentration between the in vivo AgNO3 and Ag NP treatment liver tissues. For example, the particle number concentrations were 68.3 ± 33.1 and 76.9 ± 51.6 × 109 particle per g dw liver in the AgNO3 and Ag NPs, respectively. In conclusion, a TMAH + CaCl2 extraction method was developed with good recovery and utility for detecting Ag NPs from in vivo exposures of trout liver.
Environmental significanceRobust methods to quantitatively detect particulate forms of engineered nanomaterial in fish tissues are needed for environmental risk assessment and food safety. This study describes an extraction method for fish liver that enables the detection of silver nanoparticles without altering the dissolved or particulate fractions in the sample. Of the extraction methods explored, only an alkali extraction using tetramethylammonium hydroxide (TMAH) with 5 mmol L−1 calcium chloride did not cause artefacts or dissolved silver interferences. This method was then applied to quantify biologically incorporated Ag in fish livers. Particles were readily detected in the livers of fish previously exposed to silver nitrate or silver nanoparticles via the diet in vivo. |
The available methods for detecting and characterising ENMs in environmental samples have been documented (reviews,6–8). For tissue samples, the approaches for metallic particles have been mainly concerned with the in situ detection of the presence or absence of particles, such as scanning electron microscopy coupled with energy dispersive X-ray measurements (EDX) for elemental composition.9 Coherent anti-Stokes Raman scattering (CARS) microscopy has also been used to identify metallic ENMs in or on fish gills.10 While such methods can successfully identify ENMs in cells, microscopy is often only semi-quantitative and laborious. Alternative approaches include attempts to extract the ENM from the tissue into liquid samples. For example, toluene extraction of C60 from Daphnia magna,11 or acid-extraction of (acid resistant) TiO2 NPs from fish tissues.12 However, any subsequent effort to quantify the particle number concentration or size distribution in the extracted sample by light scattering methods (e.g., nanoparticle tracking analysis, NTA) is problematic because of the corrosive nature of the samples, interferences from other colloids in the sample, and/or the modest detection limits for light scattering methods (around 10 mg L−1, see ref. 9).
The development of single particle inductively coupled plasma mass spectrometry (spICP-MS), has allowed the measurement of both particle number concentration and particle size distributions in liquid samples. The approach relies on aspirating a dispersion of intact particles into the plasma of the ICP-MS instrument. The extremely high temperature (5000–10000 K) of the plasma atomises each particle to produce an ion ‘cloud’, which in theory, arrives at the detector together (i.e., within one time period or dwell time). The resulting signal is proportional to the size of the particle atomised, and the signal frequency magnitude is proportional to the number of particles in the sample.13 The spICP-MS technique was first demonstrated in ultrapure water14 and has been used to demonstrate its capability to detect silver nanoparticles in ultrapure water.15 The technique has also been applied to tissues of animals16 and plants materials (tomato plants,17 rice plants,18).
However, these methods have not been applied to fish tissue. The determination of ENMs in fish is crucially important for environmental regulation, and currently there is no validated method for ENM detection to support mandatory bioaccumulation testing,5 for environmental monitoring, or the food safety of edible fish. Furthermore, to gain regulatory acceptance, any method would need to be robust and simple, preferably without multiple steps in digestion procedures or expensive reagents. There are two approaches that have been reported for animal tissues, and neither has been applied to trout. The first is a strong alkali digestion, using tetramethylammonium hydroxide (TMAH) to extract Ag NPs and Au NPs from Daphnia after waterborne exposure to these ENMs followed by the measurement of particle size distributions.19 The second approach used an enzymatic digestion with proteinase K to extract Ag NPs which had been spiked into a chicken meat matrix; with subsequent analysis by spICP-MS.20 However, organisms contain both particulate and dissolved fractions of metals in their tissues following an ENM exposure,1,21 but the reports so far on tissue extraction methods for ENMs have not assessed the suitability of extractants to remove the dissolved metal in the tissue, without causing any changes to the sample. For example, the precipitation of ‘dissolved’ metal species to make new particulates in the tissue, and/or altered dissolution of the original ENM. Such changes could result in false positives, where particles from the dissolved metal are apparently detected, and not part of the original ENM dose inside the tissue. However, there are some additional technical challenges to overcome to routinely apply either method to determine Ag NPs in fish tissues. Firstly, exposure to Ag NPs at environmentally-relevant concentrations result in a modest bioavailability (typically a few percent of the exposure dose,22), and a key challenge is to routinely detect low concentrations (e.g., ng g−1) of Ag NPs in fish without suffering interferences from the chemical matrix of the sample. Secondly, fish have a ‘naturally’ occurring background of dissolved Ag in the tissue of around 1 μg g−1 dry weight or less,23 and this will influence both the detection limit and the minimum particle size that can be determined by spICP-MS. Finally, since there are no certified reference tissue samples for ENMs in fish tissues, spike recovery tests must be used for quality control; but spike-recovery test with dissolved silver (e.g., silver nitrate) may not represent the recovery of the nano form.
The present study aimed to tackle these challenges and to develop a robust, but simple method to extract Ag NPs from fish tissues, using modified versions of the proteinase K or TMAH digestions outlined above to determine the particle number concentration, particle mass concentration and particle size distributions in liver tissue from rainbow trout by spICP-MS. The specific objectives included spiking different extractants with either Ag NPs or AgNO3 to determine the effect of the digestion matrix on the detection of each form of Ag. Then, experiments were performed by spiking normal liver tissues prior to extraction in order to investigate the procedural recovery of the entire protocol. Finally, attempts were made to establish the dissolved versus nanoparticulate fractions of Ag from fish livers that had been previously exposed in vivo to either AgNO3 or Ag NPs via the diet.24 These latter measurements aimed at determining if the method could detect ‘biologically incorporated’ Ag NPs in a tissue sample compared to artificial processes of directly spiking liver samples with Ag NPs. It also enabled some appreciation of the likely ratio of dissolved to particulate silver in exposed fish in vivo.
Trace amounts of Ag was present in the normal animal diet. Batches of the food pellets were digested in primer plus grade nitric acid, diluted and analysed using ICP-MS,5 with a total Ag concentration in the animal feed of 0.52 ± 0.02 μg g−1 dw (mean ± S.E.M., n = 5). Livers were also collected from juvenile rainbow trout that had been subject to a dietary exposure in vivo to food containing no added Ag (control) or 100 mg kg−1 Ag as either AgNO3 or Ag NPs (see ref. 24). The animals were fed a 2% ration and the Ag was applied to the food as a ‘top dressing’ to a commercially available diet (i.e., the control diet).24 Livers were collected at week 2 of the exposure for spICP-MS.
The next series of experiments were conducted to assess the stability of either AgNO3 or Ag NPs in solutions of the two extractants under study, namely proteinase K or TMAH. The proteinase K extractant was prepared as previously reported,20 but with minor modifications. A digestion buffer was made containing 10 mmol L−1 Tris buffer, 1% Triton X-100 and 1 mmol L−1 calcium acetate and adjusted to pH of 9.5 with NaOH. A working stock solution of proteinase K (one unit will hydrolyse urea-denatured haemoglobin to produce colour equivalent to 1.0 μmole of tyrosine per min at pH 7.5 and at 37 °C) was also prepared by dissolving 6 mg of proteinase K in 1 mL of HPW. Subsequently, 0.57 mL of this solution was added to 4 mL of the digestion buffer, giving a final proteinase K activity of 3.74 units per mL. The manufacturers recommend a different approach which was also followed. Therefore, a solution of proteinase K (1 mg mL−1) was added to 5 mmol L−1 CaCl2 with 50 mmol L−1 Tris-HCl and the pH adjusted to pH 8.
To assess the stability of AgNO3 or Ag NPs in TMAH, 2 mL of 25% TMAH was added to a 15 mL centrifuge tube and 50 μL of 100 μg L−1 Ag NP stock added to give a final total Ag concentration of 2.5 μg L−1. The effect of added CaCl2 on the stability was also examined by adding 0.4 mL of a 25 mmol L−1 CaCl2 solution to separate 1.6 mL aliquots of TMAH. The samples of Ag-spiked matrix were immediately vortexed for 1 min and then diluted 50-fold using HPW to a final concentration of 50 ng L−1 before analysis by spICP-MS. Separate samples were also spiked with 50 μL of 100 μg L−1 Ag as silver nitrate for comparison in exactly the same way. To measure the within sample variability, a relevant selected sample was measured five consecutive times, and to determine between sample variability, measurements on at least five independently prepared samples were made.
The suitability of proteinase K, proteinase K + CaCl2, TMAH and TMAH + CaCl2 solutions (prepared as above) for digesting fish livers was also assessed. Using approximately 50 mg (dry weight) of fish liver accurately weighed into a 15 mL centrifuge tube. For both proteinase K extraction protocols, 4 mL of extraction buffer was added to the tissue. The samples were vortexed for 1 min, sonicated for 5 min (FB15048 ultrasonic bath, 35 W, Thermo Fisher) and then incubated in a water bath at 35 °C for 3 hours.20 For the TMAH-based extractions, 2 mL of extractant was added to the fish liver and vortexed for 1 min and the samples were left in the dark at room temperature for 24 h.
To investigate the overall procedural recovery of the entire digestion plus the measurement protocol, normal liver tissues were spiked with either AgNO3 or Ag NPs. Only the TMAH + CaCl2 extractant was used for these experiments as the results obtained for the three other extractants were unsatisfactory. The 50 mg dry weight of liver tissue, plus the 2 mL of extraction solution was fortified with 50 μL of 100 μg L−1 Ag as Ag NPs or AgNO3. The samples were left in the dark at room temperature for 24 hours before being diluted 50-fold with HPW and analysed using spICP-MS (see below).
The final experiment involved liver tissue from week 2 of an in vivo dietary exposure.24 Fish (n = 6/treatment) were euthanised with an overdose of buffered (NaHCO3) MS222 and pithing to destroy the brain (schedule 1), following which the liver was carefully removed, weighed, and then freeze dried. The resulting tissue was cut in half (as reported in Clark et al.24) and the sub-sample weighed (∼25 mg) and digested using the optimised TMAH + CaCl2 protocol, and the extract analysed using spICP-MS. This latter experiment was aimed at determining if the method could detect ‘biologically incorporated’ Ag NPs in a tissue sample compared to the artificial process of directly spiking liver samples with Ag NPs. It also provided an opportunity to measure the dissolved versus nanoparticulate fractions of Ag from fish livers that had been previously exposed in vivo.
Statistical analysis was performed in SigmaPlot 13.0, unless otherwise specified. Data were checked for normality and equal variance (Shapiro–Wilk and Brown Forsythe tests, respectively). Statistical differences were highlighted using either a one-way ANOVA (between day variability and in vivo data) or a two-way ANOVA (between/within sample variation and matrix). The normality of the size distributions was checked using a Kolmogorov–Smirnov test (with and without the presence of liver, and spiked liver versus in vivo data) and was conducted in SPSS software.
The recovery of Ag NPs in the proteinase K extractant without additions of 5 mmol L−1 CaCl2 showed no appreciable loss of particle mass concentration, particle number concentration or mean particle size (96–103%; Table 1, Fig. 1). However, proteinase K in the presence of 5 mmol L−1 CaCl2 caused a significant increase in the apparent particle mass concentration (140–150%) and mean particle size (111–112%) compared to suspensions in deionised ultrapure water for within and between sample variation. Despite these changes, there was no change in the particle size distribution except more particles were found in higher bin sizes. The likely cause of this is particle agglomeration, resulting in multiple particles in each dwell time, which would affect the particle mass concentration and size, but not necessarily the particle number concentration. This effect was not observed in the proteinase K extractant in the absence of 5 mmol L−1 CaCl2 (Fig. 1). When AgNO3 was added to proteinase K in either the presence or absence of CaCl2, there was no unwanted reaction from the reagents (Fig. 2). Therefore, proteinase K in the presence or absence of 5 mmol L−1 CaCl2 is a suitable extractant that does not alter the Ag NPs or AgNO3.
Extraction method | Form of silver | Sample type | Particle mass concentration (ng L−1) | % recovery | Coefficient of variation (%) | Mean particle number concentration L−1 (×106) | % recovery | Coefficient of variation (%) | Mean particle size (nm) | % recovery | Coefficient of variation (%) |
---|---|---|---|---|---|---|---|---|---|---|---|
Data are means ± S.D., n = 5. Recovery was based on the particle number concentration as measured in ultrapure deionised water at the same concentration suspension (50 ng L−1). The limit of detection was 12 particles per scan, which equates to 0.38 × 106 particles per L. N/A indicates does not apply due to no particles detected (ND), or cannot be calculated as was a false positive (e.g. AgNO3 in TMAH only). Different uppercase letters denotes significant difference between same sample and silver type compared to different extraction method (two-way ANOVA or Kruskal–Wallis). Different lower case letters denote significant difference between sample types within the same form of silver (two-way ANOVA or Kruskal–Wallis). There was no significant difference in any measurements of the AgNO3 in TMAH (one-way ANOVA or Kruskal–Wallis). | |||||||||||
High purity water | Ag NPs | Within sample | 47.66 ± 2.06Aa | 100 ± 4.32 | 4.32 | 45.1 ± 0.9Aa | 100 ± 1.98 | 1.98 | 57.21 ± 0.57Aa | 100 ± 1.00 | 1.00 |
Between samples | 51.83 ± 2.90Ab | 100 ± 5.59 | 5.59 | 48.5 ± 1.9Aa | 100 ± 3.98 | 3.98 | 57.65 ± 0.53Aa | 100 ± 0.92 | 0.92 | ||
AgNO3 | Within sample | ND | N/A | N/A | ND | N/A | N/A | ND | N/A | N/A | |
Between samples | ND | N/A | N/A | ND | N/A | N/A | ND | N/A | N/A | ||
Proteinase K (without CaCl2) | Ag NPs | Within sample | 49.48 ± 1.38Aa | 103.84 ± 2.89 | 2.78 | 45.4 ± 1.3Aa | 100.67 ± 2.86 | 2.84 | 58.29 ± 0.23Ba | 101.90 ± 0.41 | 0.40 |
Between samples | 50.74 ± 1.66Aa | 97.90 ± 3.20 | 3.27 | 46.6 ± 1.3Aa | 96.15 ± 2.62 | 2.73 | 58.09 ± 0.20Ba | 100.00 ± 0.35 | 0.35 | ||
AgNO3 | Within sample | ND | N/A | N/A | ND | N/A | N/A | ND | N/A | N/A | |
Between samples | ND | N/A | N/A | ND | N/A | N/A | ND | N/A | N/A | ||
Proteinase K (with 5 mmol L−1 CaCl2) | Ag NPs | Within sample | 70.79 ± 3.67Ba | 148.54 ± 7.71 | 5.19 | 47.0 ± 2.1Aa | 104.19 ± 4.59 | 4.40 | 64.58 ± 0.33Da | 112.89 ± 0.58 | 0.52 |
Between samples | 72.87 ± 3.54Ba | 140.58 ± 6.83 | 4.86 | 48.7 ± 1.7Aa | 100.52 ± 3.42 | 3.40 | 64.38 ± 0.39Da | 111.68 ± 0.67 | 0.60 | ||
AgNO3 | Within sample | ND | N/A | N/A | ND | N/A | N/A | ND | N/A | N/A | |
Between samples | ND | N/A | N/A | ND | N/A | N/A | ND | N/A | N/A | ||
TMAH (without CaCl2) | Ag NPs | Within sample | 50.16 ± 1.27Aa | 105.25 ± 2.66 | 2.66 | 44.3 ± 1.3Aa | 98.26 ± 2.84 | 2.89 | 59.06 ± 0.22Ca | 103.24 ± 0.38 | 0.37 |
Between samples | 54.72 ± 3.75Ab | 105.58 ± 7.24 | 7.24 | 48.6 ± 4.6Aa | 100.32 ± 9.56 | 9.53 | 58.98 ± 0.51Ca | 102.30 ± 0.88 | 0.86 | ||
AgNO3 | Within sample | 0.85 ± 0.11Aa | N/A | 13.52 | 3.0 ± 0.3Aa | N/A | 10.54 | 34.34 ± 0.25Aa | N/A | 0.73 | |
Between samples | 0.67 ± 0.27Aa | N/A | 39.58 | 2.3 ± 0.7Aa | N/A | 33.25 | 34.04 ± 0.72Aa | N/A | 2.12 | ||
TMAH (with 5 mmol L−1 CaCl2) | Ag NPs | Within sample | 51.04 ± 0.78Aa | 107.09 ± 1.65 | 1.54 | 44.8 ± 0.5Aa | 99.24 ± 1.12 | 1.13 | 59.05 ± 0.17Ca | 103.22 ± 0.31 | 0.30 |
Between samples | 53.30 ± 2.08Aa | 102.83 ± 4.02 | 3.91 | 47.0 ± 1.6Aa | 96.86 ± 3.23 | 3.33 | 59.09 ± 0.11BCa | 102.50 ± 0.20 | 0.19 | ||
AgNO3 | Within sample | ND | N/A | N/A | ND | N/A | N/A | ND | N/A | N/A | |
Between samples | ND | N/A | N/A | ND | N/A | N/A | ND | N/A | N/A |
For the TMAH digestion (without CaCl2), the particle mass concentration and particle number concentration recovery was within generally accepted limits of 100 ± 5%, and did not significantly differ from those measurements in ultrapure deionised (Table 1). However, there was a slight but significant increase in the mean particle size to 59 nm (two-way ANOVA, P < 0.001). When AgNO3 was spiked into TMAH, particle-like events were observed in the time scan (Table 1, Fig. 2). Of these particle-like events, there was no difference observed between the within and between sample variation (one-way ANOVA or Kruskal–Wallis, P > 0.05). Therefore, TMAH alone is not a suitable extractant when there is a potential background of dissolved Ag in the original sample (e.g. from particle dissolution or the incidental occurrence of dissolved silver salts in the original sample). The cause of this apparent particulate signal with TMAH is likely to be precipitation as Ag2O under the high pH conditions.29,30 The mechanism by which the addition of CaCl2 to the TMAH solution prevented particle like events from AgNO3 requires further investigation. The presence of TMAH at high pH with AgNO3 without added CaCl2 may promote the formation of AgOH and/or Ag2O species. However, the additional of more dissolved Cl− ions as CaCl2 might drive dissolved Ag+ to form complexes with the Cl− anion instead (e.g. AgCl2−), or perhaps complexes with TMAH such as [AgN(CH4)4]2+. Any competition between complex formations at high pH would depend on the reaction rate and stability constants for the Ag species involved. Further experiments, such as stripping voltammetry, may help verify this hypothesis, and is beyond the scope of the present study.
For the TMAH digestion in the presence of 5 mmol L−1 CaCl2, there was no effect on the between and within sample variation particle mass or particle number concentrations (Table 1). However, similarly for TMAH alone, there was a slight but statistically significant increase (∼103%) in the mean size particle, both between and within samples (two-way ANOVA, P < 0.001). This is within acceptable limits for the analysis of even dissolved metals in complex matrices (i.e., within 10%), and while acceptable limits of deviation for measurements with ENMs are not generally agreed,6 the percent deviation for an ‘ideal’ suspension by spICP-MS has been suggested to be 100 ± 25% or less.19 The results here (Table 1) for TMAH + CaCl2 are all within the suggested limit offered by Gray et al.19 When AgNO3 was spiked into the sample, there was no presence of particle-like events (Fig. 2, Table 1); therefore, TMAH in the presence of 5 mmol L−1 CaCl2 is a suitable extractant that will not alter the form of silver within the tissue.
From the results presented, it is apparent that TMAH in the presence of 5 mmol L−1 CaCl2 was the only suitable extractant for Ag NPs contained within/on liver tissues. Spiking Ag NPs onto liver tissue that were left in the dark at room temperature for 24 h caused no appreciable change in particle size distribution compared to freshly spiked samples (Fig. 3). The results of the liver tissue spiking tests, where TMAH + CaCl2 was spiked with Ag NPs or AgNO3 in either the presence or absence of liver tissue and left for 24 h are shown (Table 2). The presence of liver tissue did not alter the particle mass concentration (two-way ANOVA, P > 0.05), or size distribution (Fig. 3). However, the presence of the liver did cause a significant decrease in mean particle number concentration and an increase in mean particle size, which suggests some settling and/or agglomeration occurred. For example, the between sample particle number concentration decreased from 47 to 36 × 106 particle per L in the absence and presence of liver, respectively (Table 2). Nonetheless, the coefficient of variation between samples was only 5% and within acceptable limits and the particle mass recovered was 100% (Table 2). Gray et al.19 spiked beef and Daphnia magna samples with Ag NPs in a TMAH matrix, and the recoveries found are also comparable to those reported here for fish.
TMAH + CaCl2 | Sample type | Particle mass concentration (ng L−1) | % recovery | Coefficient of variation (%) | Mean particle number concentration L−1 (×106) | % recovery | Coefficient of variation (%) | Mean particle size (nm) | % recovery | Coefficient of variation (%) |
---|---|---|---|---|---|---|---|---|---|---|
Data are means ± S.D., n = 5. Recovery was based on the particle number concentration as measured in freshly spiked TMAH + 5 mmol L−1 CaCl2 at the same Ag concentration solution (50 ng L−1). No particles were detected in both AgNO3 treatments (ND). Different upper case letters denote statistical difference between the presence and absence of liver tissue with TMAH + 5 mmol L−1 CaCl2 matrix within the same sample type (two-way ANOVA or Kruskal–Wallis). Different lower case letters denote statistical difference between sample types (two-way ANOVA or Kruskal–Wallis). | ||||||||||
Spiked with Ag NPs | Within sample | 49.07 ± 2.49Aa | 96.14 ± 4.88 | 5.07 | 41.2 ± 1.8Aa | 91.93 ± 4.06 | 4.42 | 59.30 ± 0.18Aa | 100.43 ± 0.30 | 0.30 |
Between samples | 52.98 ± 2.36Aa | 99.40 ± 4.43 | 4.45 | 46.8 ± 4.9Ab | 99.57 ± 10.39 | 10.44 | 59.26 ± 0.45Aa | 100.28 ± 0.77 | 0.77 | |
Spiked with AgNO3 | Within sample | ND | N/A | N/A | ND | N/A | N/A | ND | N/A | N/A |
Between samples | ND | N/A | N/A | ND | N/A | N/A | ND | N/A | N/A | |
Liver spiked with Ag NPs | Within sample | 54.60 ± 2.03Aa | 106.98 ± 3.98 | 3.72 | 37.6 ± 1.2Ba | 83.96 ± 2.95 | 3.15 | 63.42 ± 0.65Aa | 107.41 ± 1.10 | 1.03 |
Between samples | 53.38 ± 4.85Aa | 100.15 ± 9.11 | 9.09 | 36.0 ± 2.0Ba | 76.77 ± 5.58 | 5.58 | 63.46 ± 1.81Ba | 107.39 ± 3.06 | 2.85 | |
Liver spiked with AgNO3 | Within sample | ND | N/A | N/A | ND | N/A | N/A | ND | N/A | N/A |
Between samples | ND | N/A | N/A | ND | N/A | N/A | ND | N/A | N/A |
The particles found in the control livers did not significantly differ in mean size compared to both of the dietary Ag treatments (Table 3). But in the Ag NP treatment, there was a significantly elevated number of particles and particle mass concentration in the livers compared to that of the control fish (Fig. 4 and Table 3). The particles in the livers of fish from the dietary Ag NP treatment were very similar to those found in livers from the dietary AgNO3 treatment (Fig. 4), as there was no significant difference between any particle-based measurements between the two treatments (Table 3). The presence of Ag-containing particles in the liver of fish exposed to dietary AgNO3in vivo is interpreted as biogenic particle formation in the living tissue. The normal fish livers simply spiked with AgNO3 solution showed no particle formation (Table 2). Trout are known to form metal storage granules in vivo (e.g., dietary Cu exposures,31). Liver from humans exposed during silver mining also contain Ag granules.32 In fish it is also expected that metal homeostasis will inevitably vary between animals in vivo and around the homeostatic set point;33 and this is reflected in the relative standard deviation of the in vivo livers that were at least 6-fold higher compared to spike tests in the same matrix (Table 2) or in the presence of uncontaminated liver tissue (Table 3). If fish can synthesise biogenic Ag-containing particles from dissolved Ag exposure in vivo, then there is also the possibility that fish might also modify Ag NPs taken up in vivo. In the dietary study with fish here,24 all of the particle size distributions from the in vivo livers were significantly different from those of Ag NPs spiked into the matrix (Kolmogorov–Smirnov, P < 0.001, Tables 2 and 3); suggesting in vivo modification of the particles. Whether or not this modification occurred in the intestine during uptake or in the liver itself requires further investigation.
Measurement | Treatment | Liver 1 | Liver 2 | Liver 3 | Liver 4 | Liver 5 | Liver 6 | Mean ± S.D | Coefficient of variation (%) |
---|---|---|---|---|---|---|---|---|---|
Data are n = 5/6. (*) denotes a significant outlier from Grubbs test and removed from further calculations. Different upper case letters denotes significant difference between treatments (columns; one-way ANOVA or Kruskal–Wallis). The LOD for particle number concentration was 1.5 × 106 particles per g tissue dw. | |||||||||
Tissue particle mass concentration (μg g−1 dw) | Control | 0.002 | 0.005 | 0.002 | 0.006 | 0.002 | 0.004 | 0.004 ± 0.002A | 50.3 |
AgNO3 | 43.8 | 200.0 | 309.9 | 23.1 | 196.9 | 57.9 | 138.6 ± 114.3B | 82.5 | |
Ag NP | 84.5 | 0.3* | 184.8 | 155.7 | 111.3 | 54.1 | 118.1 ± 52.8B | 44.7 | |
Particle number concentration (×109 g−1 dw) | Control | 0.001 | 0.009 | 0.00 | 0.01 | 0.00 | 0.01 | 0.007 ± 0.004A | 60.9 |
AgNO3 | 32.4 | 67.4 | 51.1 | 53.2 | 12.8 | 78.0 | 68.3 ± 33.1B | 48.4 | |
Ag NP | 39.2 | 26.6 | 65.8 | 14.0 | 92.1 | 12.2 | 76.9 ± 51.6B | 67.0 | |
Mean particle size (nm) | Control | 61.2 | 44.7 | 43.7 | 43.5 | 42.8 | 39.0 | 45.8 ± 7.8A | 17.0 |
AgNO3 | 61.8 | 76.6 | 93.6 | 42.7 | 53.2 | 49.2 | 62.9 ± 19.1A | 30.4 | |
Ag NP | 66.1 | 26.7 | 72.8 | 54.0 | 54.7 | 41.5 | 52.7 ± 16.7A | 31.7 |
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9en00547a |
This journal is © The Royal Society of Chemistry 2019 |