Tamara S.
Galloway
*a,
Yuktee
Dogra
a,
Natalie
Garrett
b,
Darren
Rowe
a,
Charles R.
Tyler
a,
Julian
Moger
b,
Eva
Lammer
c,
Robert
Landsiedel
d,
Ursula G.
Sauer
e,
Gertrud
Scherer
f,
Wendel
Wohlleben
g and
Karin
Wiench
*h
aCollege of Life and Environmental Sciences, University of Exeter, EX4 4QD, UK. E-mail: t.s.galloway@exeter.ac.uk
bCollege of Mathematical and Physical Sciences, University of Exeter, EX4 4QD, UK
cBASF Schweiz AG, Klybeckstrasse 141, CH-4057 Basel, Switzerland
dBASF SE, GB/TB - Z470, D-67056 Ludwigshafen, Germany
eScientific Consultancy – Animal Welfare, Hallstattfeld 16, D-85579 Neubiberg, Germany
fBASF SE, E-EDE/QP, D-67056 Ludwigshafen, Germany
gBASF SE, RAA/OR - B7, D-67056 Ludwigshafen, Germany
hBASF SE, FEP/PC - Z470, 67056 Ludwigshafen, Germany
First published on 22nd August 2017
Nanoparticle-containing polymer dispersions are widely used, but little is known of their environmental effects. We studied the bioavailability, uptake, tissue localisation and effects of nanoparticle-containing acrylic copolymer (ACP) dispersions (mean nanoparticle sizes: 80 nm and 110 nm) in aquatic invertebrates (Thamnocephalus platyurus; fairy shrimp) and Danio rerio zebrafish embryos after aquatic exposures. Dietary exposure tests were enabled using Casper zebrafish that lack skin pigmentation allowing for bio-imaging of uptake and internal distribution. Aqueous exposures of 1000 and 2500 mg L−1 80 nm-ACP or 110 nm-ACP showed no acute toxicity in fairy shrimp or zebrafish, constituting a non-toxic classification according to the United Nations Globally Harmonised System of Classification and Labelling of Chemicals threshold (100 mg L−1). Similarly, dietary exposures resulted in no ecotoxicological effects. In Casper zebrafish fed with 80 nm-ACP-spiked food, hyperspectral signals derived using coherent Raman scattering (CRS) indicated that test material was present in the intestine, and possibly in the liver, but not in other organs. CRS imaging indicated that the chemical composition of the yolk sac of an 80 nm-ACP exposed zebrafish (aquatic exposure) was altered, attributed to a change in lipid metabolism, although we could not confirm with certainty that the test material was physically present in the yolk sac. These results illustrate how CRS microscopy can be used to investigate the bioaccumulation of organic nanomaterials, provided that they induce hyperspectral profiles distinct from the biological samples. In conclusion, both 80 nm- and 110 nm-ACP dispersions are internalised through dietary exposure, but are not associated with significant toxic effects.
Environmental significanceFew studies are available on potential ecotoxicological effects of organic nanomaterials. We examined acute aquatic effects of widely used, environmentally relevant nanoparticle-containing acrylic polymer dispersions. In fairy shrimp and zebrafish embryos, the test materials showed virtually no acute aquatic toxicity. Further, we investigated the uptake, bioavailability, and biodistribution of the acrylic polymers using coherent Raman scattering (CRS) microscopy. When zebrafish were fed with acrylic polymer-containing food, test material was present in the intestine, possibly in the liver, but not in other organs. CRS microscopy proved to be a promising method for ecotoxicological investigations, not requiring extensive preparation to the biological sample. In conclusion, we provide insight on ecotoxicological effects of organic nanomaterials and on suitable tools for their further investigation. |
In the few available studies, ACP dispersions have been shown to possess low toxicity in mammals. Oral administration of 2 g per kg body weight ACP dispersions to rats did not result in any signs of acute toxicity, mutagenicity or sensitising activity.1 In a short-term inhalation toxicity study in rats (5 days of exposure; 6 h per day), two ACP dispersions (the one containing approx. 35% nanoparticles in the aerosol and the other no appreciable amount of nanoparticles) did not elicit any adverse effects up to aerosol concentrations of 10 mg m−3. Hence, the higher proportion of nanoparticles in these polymer dispersions did not affect the outcome of the study.1
Hazard assessment of nanomaterials and nanotechnology-enabled products should take into account all relevant aspects of their life cycle and biological pathways from their initial synthesis, to commercial use and disposal.4–7 The potential ecotoxicology of nanoparticle-containing polymer dispersions is of particular concern for aquatic organisms, since aquatic systems are the ultimate repository for released anthropogenic substances.8 In receiving waters, the ecotoxicology of nanomaterials, including polymers, is largely governed by their intrinsic properties and the nature of the local environment, which influences interactions with suspended particulate matter and the extent of biological and abiotic degradation, agglomeration and aggregation.9,10 An enhanced understanding of the potential effects of nanomaterials to aquatic species requires a comprehensive determination of their system-dependent properties in relevant environmental matrices, coupled with detailed information of their bioavailability, uptake and distribution within the biological tissues of appropriate species.
Very few studies are available addressing the potential environmental effects of nanoparticle-containing polymer dispersions released into the environment.11 In tests performed for regulatory purposes, a spectrum of nanoparticle-containing polymer dispersions did not show ecotoxicological effects in fish, daphnia or algae upon short-term exposure (unpublished in-house data). The published data available for polymer nanocomposites do not indicate appreciable nanoparticle release or (eco-)toxicity of released fragments.12–14 When solid acrylate polymers with embedded quantum dots, used in the lighting industry, were exposed to different environmental and biological simulant fluids, the leachates contained soluble metals derived from the quantum dots, but no free quantum dots or quantum dot-containing polymeric debris.15 Commercial nanofiller composites submitted to chemical stresses as relevant in outdoor aging and/or mechanical forces mimicking sanding can release fragment polymer particles, with a fraction as colloidal nanoparticles,16,17 but these did not lead to significant toxicity.12,13,18
The present study investigated the bioavailability, uptake, body tissue distribution and effects of nanoparticle-containing ACP dispersions to an aquatic invertebrate Thamnocephalus platyurus (fairy shrimp) and Danio rerio (zebrafish) after aquatic and dietary exposures. Test methods adopted were the International Standardisation Organisation 14380:2011 Water quality determination of the acute toxicity to Thamnocephalus platyurus Crustacea; Anostraca, (fairy shrimp)19 and the Organisation for Economic Co-operation and Development (OECD) Test Guideline (TG) 236 Fish Embryo Toxicity (FET) test20 using Danio rerio (zebrafish) embryos.
Thereby, the present study served to provide first insight on the potential acute aquatic toxicity of nanoparticle-containing polymer dispersions. Assessment of short-term aquatic toxicity in invertebrates and fish is a mandatory component of the base set of data requirements for ecotoxicity testing. In Regulation EC No. 1907/2006 on the Registration, Evaluation, Authorisation, and Restriction of Chemicals (REACH),21 information on short-term aquatic toxicity in invertebrates is generally required for substances manufactured or imported in quantities exceeding 1 ton per year. Information on acute fish toxicity is generally required if the annual tonnage exceeds 10 tons. As compared to the fish acute toxicity test (OECD TG 203),22 the FET test serves the 3Rs principle to replace, reduce and refine animal testing23 that has been implemented in Directive 2010/63/EU on the protection of animals used for scientific purposes.24 Therefore, the FET test was selected for acute aquatic fish toxicity testing in the present study. Since the zebrafish embryos used in the FET test are not considered as ‘independently feeding larval forms’, the test, as far as the embryos are concerned, does not fall within the scope of Directive 2010/63/EU.24,25 In zebrafish embryos the egg shell, the chorion, is transparent26,27 and the developmental stages are well characterised and completed rapidly, i.e. within 120 h.28–30 The fairy shrimp were selected for acute aquatic toxicity testing in invertebrates, since they lack a carapace and have an open circulatory system, allowing an unimpeded exchange of oxygen or chemicals, and they are sensitive indicators of environmental pollution.31,32
Further, dietary exposure studies were conducted using zebrafish fry. These studies served to investigate the uptake and bioavailability of ACP dispersions in fish. Information on the toxicokinetics of substances are equally relevant for regulatory hazard and risk assessment. However, to data there are no standardised test methods to investigate test material uptake or biodistribution in environmentally relevant species. In the present study, the dietary exposure studies used a mutant line of D. rerio, known as Casper, whose embryos and adults lack all melanocytes and iridophores33 avoiding problems associated with pigments that can interfere with the microscopic detection of internalised test materials.33
In addition, two coherent Raman scattering (CRS) microscopy techniques (stimulated Raman scattering (SRS) and coherent anti-Stokes Raman scattering (CARS)) were applied which allow for test material-specific identification without the need for extraneous dyes or fluorophores and provide sub-cellular resolution and three-dimensional sectioning capability for visualising the structural features of the test organisms with which the particles interact.34–36 To date, the CARS technique has been applied to image the uptake, distribution and bioavailability of different nanomaterials (metal oxides, metals, polymers) in human tissues and ecological model species.34–39 See section 1 of the ESI† for further details on the CRS techniques.
In these pre-tests, 110 nm-ACP was best suited for CRS microscopy in terms of its differentiation from fairy shrimp or zebrafish tissues. The 110 nm-ACP exhibited a strong CH peak at 3055 cm−1, indicative of a benzene ring, which is typically not a pronounced peak in biological samples. By contrast, in a comparative assessment of the spectral profiles of 110 nm-ACP and unexposed fairy shrimp, the majority of the other pronounced peaks were recorded in either spectral profile (Fig. 1A; cf. ESI,† Fig. SI-1).
To verify that the spectral profile of 110 nm-ACP exhibited the same peaks in spontaneous Raman scattering and CRS, a hyperspectral SRS profile was acquired of 110 nm-ACP applied to a glass coverslip (cf. ESI,† section 2). Using SRS, a strong peak at 3055 cm−1 was also obtained. This confirmed both the suitability of 110 nm-ACP as test material and of the CRS microscopy (Fig. 1B). Further, a hyperspectral profile was obtained from an unexposed wild-type zebrafish embryo (Fig. 1C). There was negligible contribution to the CH stretch region at 3060 cm−1. Consequently, the wavenumber of 3055 cm−1 was identified and selected as an ideal Raman shift to detect the ACP test materials in biological samples.
Overall the experiments were performed in compliance with the relevant national and European Union laws and institutional guidelines, and the responsible institutional committees approved the experiments. Of note, the tests on fairy shrimp are not covered by Directive 2010/63/EU on the protection of animals used for scientific purposes.24 Of the tests using wild-type or zebrafish embryos, only the study parts involving independently feeding forms are covered by this Directive.
The preliminary set of five different polymers (i.e. 110 nm-ACP; ACP-2; AAECP; SACP; and PPE) was submitted to range finding studies that were conducted applying the same protocols as the ones described for the main acute aquatic toxicity tests in fairy shrimp and zebrafish. Test material concentrations of 1, 10, 100, 1000 and 10000 mg L−1 (corresponding to 0.0001–1% (v/v)) were assessed in fairy shrimp (24 h exposure; counting 30 animals/concentration) and zebrafish (24, 48, 72 and 96 hpf; 30 animals/concentration). Subsequently, test material concentrations of 1000 and 2500 mg L−1 (corresponding to 0.1 and 0.25% (v/v)) were selected for the main tests assessing 80 nm-ACP and 110 nm-ACP. These very high test material concentrations were selected to account for the very low ecotoxicity of the nanoparticle-containing polymer dispersions observed in the range finding studies.
The ecotoxicological relevance of the results from the acute toxicity tests with fairy shrimp and wild-type zebrafish embryos was assessed using the LC50 value threshold laid down in Regulation (EC) No 1272/2008 on classification, labelling and packaging of substances and mixtures (CLP)44 and the United Nations Globally Harmonised System of Classification and Labelling of Chemicals (GHS):45 LC50 values exceeding 100 mg L−1 recorded after 48 h exposure to crustaceans or after 96 h exposure to fish (or failure to calculate LC50 values at all due to a lack of mortality) do not result in substance classification for acute aquatic toxicity.
To prepare food containing polystyrene beads, the surfactant present on the polystyrene beads (proportion of the surfactant: 0.1–0.5%; as specified by the supplier Sigma-Aldrich) was first removed by centrifugation (Beckman Avanti J25, Beckman Coulter, UK; 10 min at 17000 rpm). Next, 4 mL beads were re-suspended in 4 mL standardised dilution water, and added to 4 g ZM-000 food to yield 9 × 1012 particles per gram of food. Food and beads were mixed thoroughly to form a paste, spread on baking sheets and dried for 24 h at 60 °C and ground to a fine powder.
Food containing the ACP test material was prepared by adding equal volumes of the test material preparations in standardised dilution water (0–20%) to equal weights of food to yield final test material concentrations of 0.01, 0.1, 1 and 10% in the food (corresponding to 0.1, 1, 10 and 100 g per kg food). Control food was prepared by using only standardised dilution water.
After mixing, the test material-supplemented food was further processed by oven drying (60 °C; 24 h) or by air drying (∼21 °C; 48 h). Oven-dried and air-dried food showed no difference in hyperspectral profile, hence oven drying was used forthwith.
Embryos were collected from a breeding stock of Casper zebrafish and individual embryos placed in single wells of a 24-well plate with 2 mL of standardised dilution water, changed every 24 h. For each exposure concentration, 15 embryos were used. The plate was maintained at 26 ± 1 °C, and embryos kept under a constant artificial dark/light cycle of 8/16 h. Ninety-six h post fertilisation (hpf) and 120 hpf, embryos were fed with a fine pipette tip of control food or food containing either the polystyrene beads or the ACP test material (two feedings, each, on either day; for technical reasons, it was not possible to calculate the specific amount fed to the embryos). After each feeding, the water in the wells was changed. At 120 hpf, the feeding studies were terminated immediately after the fry stopped feeding.
Operational parameters of the combined SRS–CARS microscopy that was applied to produce hyperspectral images and also pre-tests to determine the laser power tolerance of the biological samples are documented in section 2 of the ESI.† Briefly, scan times of approx. 14 and 20 seconds were necessary to obtain a good signal-to-noise ratio in the CARS and SRS images, respectively. At these scan times, locomotion of the live fairy shrimp resulted in blurred images. Therefore, both the fairy shrimp and the zebrafish were fixed in paraformaldehyde (Sigma-Aldrich) prior to CRS microscopic imaging and mounted on their left sides between glass coverslips using low melting point agarose gel (Thermo Fisher Scientific, UK).
Settings that are very specific to the specimen of this study and to the interpretation of microscopy results (i.e. the determination of the limit of detection and pre-tests to optimise the SRS and CARS techniques) are presented below.
Further, pre-tests were conducted to optimise test material detection during CARS imaging using fairy shrimp exposed to 1000 mg L−1 110 nm-ACP. CARS images were acquired with the pump and Stokes beams tuned to excite contrast from the CH stretch at 2870 cm−1 that yielded a peak in the spectral profile of this animal (Fig. 1C). For comparative reasons, a transmitted light image was acquired simultaneously.
Property | Method | Units | 80 nm-ACP | 110 nm-ACP |
---|---|---|---|---|
Abbreviations: AUC: analytical ultracentrifugation; DMEM/FCS: Dulbecco's modified Eagle medium supplemented with 10% fetal calf serum; GC: gas chromatography; HPLC: high performance liquid chromatography; LD: laser diffraction; TEM: transmission electron microscopy; XPS: X-ray photoelectron spectroscopy. Not determined (organic material): crystallite size and crystallite phase; water solubility. Not determined (suspension): surface area by method of Brunauer Emmett and Teller; dustiness (not expected for suspension).a As described by Wohlleben and co-workers,71 the VSSA is an integral property of materials, and it is obtained by dividing the samples' external surface (S) by its solid volume (V) or by multiplying the specific surface area (SSA, surface per mass) by the materials skeletal density (ρ). It is conventionally stated in units of m2 cm−3.b Near perfect stability in serum-containing media; complete stability (dispersed to individual particles) in water. | ||||
Physical appearance | Aqueous dispersion of acrylic copolymer; stable, white, milky; glass temperature: 15.4 °C | |||
Volatile organic components | HPLC//GC | mg kg−1 | Acrylamide: 190; acrylate: 470//n-butyl acrylate: 410; styrole: 10; acetone: 110; n-butanol: 210; n-butyl propionate: 50; t-butanol: 180; dibutyl ether: 230; total (GC): 1500 | Acrylamide: 8; acrylate: 700//n-butyl acrylate: 50; n-butanol: 280; n-butyl propionate: 20: dibutyl ether: 90; total (GC): 550 |
Primary particle size (mean) | TEM | D 50 (nm) | 80 ± 5.8 | 112 ± 8.1 |
Volume specific surface area | Derived from AUC in H2O | m2 cm−3a | 49 | 56 |
Particle size/dispersibility (H2O)b | AUC | D 50 (nm) | 77.4 | 107 |
Particle size/dispersibility (DMEM/FCS)b | AUC | D 50 (nm) | 95.4 | 123 |
Photocatalytic activity | Methylene blue test | without unit | 5.5 × 10−4 | 5.0 × 10−4 |
Surface chemistry | XPS with carbon line shape analysis | Atom-% | CC, CH: 62.9; CN, C–O: 16.5; COOR: 5.6; O: 14.2; N: 0.3; Si: 0.1; S: 0.1 | CC, CH: 62.8; CN, C–O: 16.4; COOR: 5.3; O: 14.6; N: 0.6; Si: 0.1; S: 0.1; Na: 0.1 |
Iso-electric point | Electrophoretic mobility titration | pH | <3 | <3 |
Zeta-potential at pH 7.4 | mV | −60 | −56 |
In the range finding studies with fairy shrimp (ESI,† Fig. SI-2A–SI-2E), none of the initial five test materials showed any toxicity after 24 h exposure at concentrations between 1 and 100 mg L−1, i.e. up to the GHS/CLP threshold of 100 mg L−1 for acute aquatic toxicity.44,45 The 110 nm-ACP and AAECP caused no mortality at either 1000 or 10000 mg L−1. The PPE induced significant mortality at 10000 mg L−1 (p < 0.01) and ACP-2 and SACP at 1000 and 10000 mg L−1 (p < 0.5 and < 0.01). In the main experiments, no fairy shrimp (0 of 90 animals) died following exposure to 1000 mg L−1 80 nm-ACP, and 2 of 90 animals died following exposure to 2500 mg L−1 80 nm-ACP. Inter-assay variability is shown in the data (ESI,† Fig. SI-3). Following exposure to 1000 and 2500 mg L−1 110 nm-ACP, a similarly low number of fairy shrimp died (0–7 animals/90 animals at either concentration; ESI,† Fig. SI-4). Thus, for both 80 and 100 nm-ACP, it was not possible to calculate an LC50 value (ESI,† Fig. SI-3 and SI-4). Inter-assay variability was higher for the 110 nm-ACP than for the smaller sized 80 nm-ACP.
In range finding studies with zebrafish embryos (ESI,† Fig. SI-5), 110 nm-ACP, AAECP, and PPE showed no effects on survival after 24-, 48-, 72- or 96 h exposure at 1–10000 mg L−1. ACP-2 and SACP caused significant mortality in zebrafish after 24-, 48-, 72- or 96 h exposure at concentrations of 10000 mg L−1 (100-fold higher than the GHS/CLP threshold of 100 mg L−1 for acute aquatic toxicity).44,45 In the main experiments, neither 1000 nor 2500 mg L−1 80 nm-ACP or 110 nm-ACP caused mortality in zebrafish embryos up to 96 hpf, and it was not possible to calculate an LC50 (data not shown). At all test concentrations, flocs of particulate test material were recorded surrounding the chorion, but the embryos developed normally.
Fig. 4 CARS and transmitted light images of fairy shrimp exposed to 1000 mg L−1 110 nm-ACP. Left image: Epi-detected CARS images were acquired with the pump and Stokes beams tuned to excite contrast from the CH stretch at 2870 cm−1. For Fig. 5, a second set of images was acquired from the region marked with a yellow square to determine whether signals from 110 nm-ACP were detectable at the higher magnification. Middle image: Simultaneously acquired transmitted light image of fairy shrimp exposed to 1000 mg L−1 110 nm-ACP (both: 20-fold magnification). Right image: The region of interest was selected, as indicated by the black box on the fairy shrimp image. |
Fig. 5 CARS images of fairy shrimp exposed to 1000 mg L−1 110 nm-ACP (60-fold magnification). These images were obtained from the yellow insert indicated in the left-hand image of Fig. 4. The pump and Stokes beams were tuned to excite contrast from (left image) the aromatic CH stretch at 3060 cm−1 and (right image) the CH stretch at 2870 cm−1. |
SRS hyperspectral images generated from the yolk sac of an unexposed wild-type zebrafish (Fig. 6) and from the yolk sac of a wild-type zebrafish embryo exposed to 1000 mg L−1 80 nm-ACP (Fig. 7), generated at a depth of approx. 3 μm, clearly differ with respect to the spectral features of the 2820–2986 cm−1 CH-stretch region. For the zebrafish embryo exposed to 1000 mg L−1 80 nm-ACP, the hyperspectral profile in Fig. 7 exhibits a peak at 3005 cm−1 that is attributed to the CH stretch of CC groups, a marker for the degree of unsaturation of lipids. In the false colour rendering of the hyperspectral data set, this peak appears as discrete droplets with diameters ranging from 2–5 μm which can be interpreted to be lipid droplets that are more unsaturated than the surrounding features. Further, the lipid-rich regions of the yolk sac of the zebrafish embryos that were exposed to 80 nm-ACP (Fig. 7) exhibited a higher degree of unsaturation than the corresponding regions of the yolk sac of the unexposed zebrafish (Fig. 6).
The hyperspectral profile in Fig. 7 also shows a very small, broad peak at 3055 cm−1 corresponding to water-rich regions of the sample. Such a peak may be attributable to endogenous amino acids, such as tyrosine or phenylalanine, although these amino acids are more commonly associated with protein-rich regions (such as cell nuclei), and are not commonly seen within lipid droplets. The hyperspectral profile also shows an elevation at 3055 cm−1. Whilst this could be attributed to the benzene CH stretch of 80 nm-ACP, an elevation at 3055 cm−1 was also observed in the hyperspectral data set of the control animal. The elevation at 3055 cm−1 is more pronounced in the hyperspectral profile of the exposed animal, but it is clearly also present in the control animal (Fig. 6). Hence, based upon the 3055 cm−1 aromatic CH stretch, we could not say for certain that 80 nm-ACP was taken up into the yolk sac.
Fig. 8 SRS hyperspectral profile of polystyrene beads taken of the CH stretch region from 2800–3120 cm−1. |
Fig. 9 SRS hyperspectral profiles of fish food without polystyrene beads (left); and of fish food spiked with polystyrene beads (right). |
Hyperspectral SRS data sets acquired from Casper zebrafish fed with food spiked with polystyrene beads or control food are provided in Fig. 10. The hyperspectral profile generated from the fish fed with polystyrene bead-containing food exhibits the characteristic CH stretch peak in the yolk region of the fry, indicating the presence of polystyrene beads (Fig. 10A). This peak was not observed in the fish that were fed with the control food (Fig. 10B). The hyperspectral SRS profile of fish food spiked with 0.1 g kg−1 110 nm-ACP or 80 nm-ACP showed a prominent peak located at 3055 cm−1, which correlates well with the CH stretch peak from benzene rings in the spectral profile of the test materials (Fig. 11). This reveals that the test material was present in the fish food. However, it could not be ascertained that it continued to prevail in the form of nanoparticles after the drying procedure. When dried, polymer chains from neighbouring particles, due to their Tg below room temperature, start to inter-diffuse, first forming necks between particles and ultimately forming a continuous film.2
Light microscopic images of Casper zebrafish fed with food containing 0.1–100 g kg−1 80 nm-ACP showed that the fry fed freely on the supplemented food irrespective of the test material concentrations (Fig. 12). Similar results were obtained when the food was supplemented with 110 nm-ACP (data not shown). Hyperspectral SRS data of Casper zebrafish fed with food dosed with an extremely high concentration of 10 g kg−1 80 nm-ACP indicated the characteristic test material peak at around 3055 cm−1 in the gut region, albeit only for a total of 5 pixels (Fig. 13; cf. Fig. SI-8 and SI-9† for details on the identification of regions of the Casper zebrafish that contain the test material signal).
Hyperspectral data sets were acquired from the liver, muscle tissue, intestines, and the surrounding agarose gel of Casper zebrafish fed with food spiked with 1 g kg−1 80 nm-ACP or 110 nm-ACP. Signals relating to 80 nm-ACP (i.e. the characteristic benzene stretch at 3055 cm−1) were recorded in the liver and intestines, but not in the muscle tissue (Fig. 14). Similarly, 110 nm-ACP (i.e. the characteristic benzene stretch at 3055 cm−1) was presumably recorded within the liver (quantity unknown), but not in the muscle tissue or agarose gel surrounding the sample (Fig. 15).
Neither 80 nm-ACP, nor 110 nm-ACP showed any appreciable toxicity in fairy shrimp or zebrafish, and it was not possible to calculate a LC50 value. This confirms that the ACP dispersions show no acute aquatic toxicity under the standardised experimental conditions adopted. It is possible that at higher exposure concentrations, the test materials might cause toxic effects by physically adhering to the chorion,47–50 but such physical effects are most probably not relevant for environmental exposures.
A rapidly increasing number of studies have investigated the potential human health effects of engineered nanomaterials. There have been far fewer studies of their potential ecotoxicological effects and these have generally focused on inorganic nanomaterials, e.g. metals and metal oxides51,52 or metalloid oxides, such as amorphous silica nanomaterials.53,54 Only a few publications consider the ecotoxicology of organic (polymer) nanomaterials, e.g. the studies from Naha and co-workers11 and Besseling and co-workers.55
Generally, different nanomaterials induce effects on cells, tissues or organisms through different modes-of-action. For instance, the toxicity of metal and metal oxide nanomaterials is mainly attributable to the effects of the solubilised ions, and effects caused by high-aspect ratio carbon nanotubes follow a fibre toxicity paradigm.5,6 Evidence is currently lacking for ‘nanospecific’ effects, which can be attributed to the smallness of the particles alone.56,57
This study illustrates that the 24 h acute lethal toxicity test using the Thamnotoxkit F™ (ISO 14380:2011)19 and the zebrafish FET test (OECD TG 236)20 are suitable to assess the potential acute aquatic toxicity of organic nanomaterials and nanoparticle-containing polymer dispersions. During regulatory environmental hazard assessment addressing the aquatic compartment, e.g. in accordance with the REACH Regulation,21 this information is relevant together with information on degradation (e.g. sediment simulation testing) and fate and behaviour in the environment. Further, information on potential ecotoxicological effects has to be available before test substance uptake and biodistribution can be assessed, because such testing has to be performed at non-toxic concentrations. Nanoparticle agglomeration or dispersibility may alter the effective concentration of nanomaterials within the test systems or target organs in toxicity tests.58,59 Hence, specific dispersion protocols should be used during nanomaterial preparation to reduce agglomeration. Different methods used to prepare dispersions of the same nanomaterial may account for variations in the outcomes of (eco)toxicity studies.43
In the few studies published investigating the effects of inorganic or organic nanomaterials on fairy shrimp, they have proven to be sensitive indicators of ecotoxicity. In a comparative assessment of different metal oxide nanomaterials (primary particle sizes (PPS): 50–70 nm ZnO; 30 nm CuO; 25–70 nm TiO2), fairy shrimp (T. platyurus) were more sensitive than water fleas (Daphnia magna).60 TiO2 nanomaterials did not show any toxicity up to an extremely high concentration of 20 g L−1, and although toxicity occurred at higher exposure concentrations for zinc oxide and copper oxide nanoparticles this appeared mainly to be driven by the effects of the solubilised metal ions.60 Exposure (1 h) of T. platyurus to aqueous suspensions of 3 and 6 mg L−1 fullerene C60 and C70, resulted in their ingestion leading to fullerene C60 and C70 burdens of 2.7 ± 0.4 and 6.8 ± 1.5 mg per mg wet weight, respectively, but without any observable signs of toxicity.61
Poly N-isopropylacrylamide and N-isopropylacrylamide/N-tert-butylacrylamide copolymer nanoparticles were virtually non-toxic in fairy shrimp, yielding a LC50 value of 943 mg L−1.11 In a test battery of aquatic organisms representing different trophic levels, amorphous silica nanomaterials (PPS: 50 and 100 nm) showed no effects up to 1.0 mg mL−1, whereas polyethyleneimine polystyrene nanomaterials (PPS: 55 and 110 nm) caused acute toxicity at 0.0004–0.42 mg mL−1, with effects generally more pronounced for the 110 nm-, than for the 55 nm test material.62 In this study, T. platyurus was less sensitive than D. magna, but more sensitive than Vibrio fischeri bacteria.62
In brine shrimp (Artemia franciscana) larvae, 48 h exposure to 40 nm anionic carboxylated polystyrene nanomaterials or 50 nm cationic amino polystyrene nanomaterials (for both up to 0.1 mg mL−1) did not induce mortality, but larvae exposed to amino polystyrene nanoparticles underwent multiple molting events compared with the control larvae, suggesting a stress response.63 At 48 h exposure, the test materials were heavily sequestrated in the gut and some material was also adhered to the antennae and appendages, potentially hampering larvae motility.63
A broad spectrum of different nanomaterials encompassing metals, metal oxides, metal compounds as well as amorphous silica, calcium and carbon compounds has already been tested in the zebrafish FET test.52,64 The most pronounced effects have been observed for metal and metal oxide nanomaterials that shed toxic (metal) ions.52 Generally, the chorion that protects the developing zebrafish embryos is not a barrier for most manufactured non-nanosized or nanosized substances.65–68 Substances that form complexes or large structures with other similar molecules through chemical interactions, such as ionic bonding (e.g. cationic polymers); substances which interact with sulfhydryl groups (e.g. heavy metals); or large molecules (e.g. polymers) can be blocked by the chorion.65 By comparison, small 5–46 nm silver (Ag) nanomaterials (at concentrations >0.19 nM) can cross chorion pore channels that have diameters of 0.5–0.7 μm, via passive diffusion.69 Exposure to these small Ag nanomaterials can lead to fin fold abnormalities, tail flexure, cardiac oedema, yolk sac oedema, head oedema, and eye abnormalities.69 By comparison, studies on fluorescent core–shell silica nanoparticles (PPS: 60 nm and 200 nm) exposed at concentrations between 0.0025 and 200 mg L−1 were found to adsorb onto the chorion of zebrafish embryos (at 96 hpf), but were not taken up by the embryos.53 These materials did not cause mortality or deformities53 (of note, 200 nm particles are not covered by the definition of ‘nanomaterial’ as laid down in an EU Commission recommendation70).
In another study, exposure to 25–200 μg ml−1 amorphous silica nanomaterials (62 nm; without core shell) resulted in a dose-dependent decrease in the hatching rate of zebrafish embryos, and increase in mortality and embryonic malformations, including pericardial oedema, yolk sac oedema, and head malformation.54 Duan and co-workers54 did not report if or how the applied amorphous silica interacted with the chorion. Thus, while further investigations are necessary to corroborate our findings, the nanoparticle-containing ACP dispersions we tested do not appear to show any toxic effects of significance in either fairy shrimp or zebrafish embryos. Further, uptake of the test materials into the liver of the Casper zebrafish could not be identified with certainty. If such biodistribution is confirmed during future research work, specific investigations of the hepatic metabolism of the test materials would be merited. The major metabolic pathway for the monomers after oral uptake is ester hydrolysis to the alcohol and acrylic acid finally resulting in CO2. The monomers are eliminated quickly.
This study presents the first example to our knowledge of the application of CRS microscopy to explore the potential ecotoxicology of nanoparticle-containing ACP dispersions and to infer on the uptake and biodistribution of these test materials. Therefore, our study included pre-tests to determine the limit of detection of the test materials using CRS (50 mg L−1 110 nm-ACP), and to optimise the imaging procedure (i.e. fixation of the animals in paraformaldehyde; 14–20 seconds scan times; generation of overlapping depth stacks combined with lower magnification images to localise regions of interest within the animals). An advantage of CRS microscopy is that it does not require extensive preparation of the biological samples that may affect the nanomaterials, as is the case for, e.g., TEM or scanning electron microscopy.36,46 The applicability of CRS microscopy for investigating organic nanomaterials (or larger sized organic materials) is however restricted to organic materials whose spectral profile is clearly distinguishable from the spectral profile of the biological structures under investigation. In the present study, only 1 of 5 originally selected organic polymers had such a distinct spectral peak, i.e. the 3055 cm−1 aromatic CH stretch.
The CRS images yielded preliminary information on the uptake and biodistribution of nanoparticle-containing ACP dispersions. The hyperspectral profiles of control wild-type zebrafish embryos and embryos exposed to 1000 mg L−1 80 nm-ACP both showed peaks at 3055 cm−1, although the peak was weaker in the control animal. This suggests that endogenous features, such as the amino acids tyrosine and phenylalanine were also contributing to this peak. We could not determine with certainty that 80 nm-ACP penetrated into the yolk sac of zebrafish embryos, and we could not quantify the tissue burden in the test animals. There were nonetheless clear differences between control and test samples within the 2820–2986 cm−1 CH stretch region. There are numerous CH stretch peaks that contribute to this profile, overlapping with each other to generate these complex spectral features. It would be interesting to obtain additional spectra to de-convolute these peaks from the measured spectra (e.g. by curve fitting) to determine more clearly which components of the CH stretch region change after exposure to 80 nm-ACP. Such investigations could also investigate how this test material could be affecting lipid composition within the embryo's yolk sac, as is indicated here. This could plausibly be due to physical adherence of the test material to the chorion when applied at high concentrations that may affect (for example) oxygen exchange.
The results from the feeding studies using Casper mutant zebrafish provide preliminary results on the uptake, bioavailability and organ distribution of ACP dispersions when spiked in fish food. We show that the ACP dispersions were taken up by the Casper zebrafish, with strong indications for their presence in the intestines, as we would expect for dietary expose to any material, and also with preliminary evidence for a presence in the liver. The latter would indicate uptake across the gut and into the circulation, a topic worthy of further investigation. Since ACP are known to fuse when dried, future research should also explore how the intrinsic material properties of the ACP dispersions are affected by different processing techniques, e.g. oven heating versus air drying.
Footnote |
† Electronic supplementary information (ESI) available: Supplementary information, SI-tables and SI-figures. See DOI: 10.1039/c7en00385d |
This journal is © The Royal Society of Chemistry 2017 |