Effects of particle composition and species on ... - Wiley Online Library

2 downloads 0 Views 288KB Size Report
Jan 3, 2008 - analysis in the linear potion of the graph to determine the median effective concentration (EC50) according to the fol- lowing equation: EC50.
Environmental Toxicology and Chemistry, Vol. 27, No. 9, pp. 1972–1978, 2008 䉷 2008 SETAC Printed in the USA 0730-7268/08 $12.00 ⫹ .00

Nanomaterials in the Environment EFFECTS OF PARTICLE COMPOSITION AND SPECIES ON TOXICITY OF METALLIC NANOMATERIALS IN AQUATIC ORGANISMS ROBERT J. GRIFFITT,† JING LUO,§ JIE GAO,‡ JEAN-CLAUDE BONZONGO,‡ and DAVID S. BARBER*† †Center for Environmental and Human Toxicology, ‡Department of Environmental Engineering Sciences, University of Florida, Gainesville, Florida 32911, USA §Hospital of Shanxi Medical University, Taiyuan, Shanxi 030001, China ( Received 3 January 2008; Accepted 8 April 2008) Abstract—Metallic nanoparticles are among the most widely used types of engineered nanomaterials; however, little is known about their environmental fate and effects. To assess potential environmental effects of engineered nanometals, it is important to determine which species are sensitive to adverse effects of various nanomaterials. In the present study, zebrafish, daphnids, and an algal species were used as models of various trophic levels and feeding strategies. To understand whether observed effects are caused by dissolution, particles were characterized before testing, and particle concentration and dissolution were determined during exposures. Organisms were exposed to silver, copper, aluminum, nickel, and cobalt as both nanoparticles and soluble salts as well as to titanium dioxide nanoparticles. Our results indicate that nanosilver and nanocopper cause toxicity in all organisms tested, with 48-h median lethal concentrations as low as 40 and 60 ␮g/L, respectively, in Daphnia pulex adults, whereas titanium dioxide did not cause toxicity in any of the tests. Susceptibility to nanometal toxicity differed among species, with filter-feeding invertebrates being markedly more susceptible to nanometal exposure compared with larger organisms (i.e., zebrafish). The role of dissolution in observed toxicity also varied, being minor for silver and copper but, apparently, accounting for most of the toxicity with nickel. Nanoparticulate forms of metals were less toxic than soluble forms based on mass added, but other dose metrics should be developed to accurately assess concentration–response relationships for nanoparticle exposures. Keywords—Nanotoxicology

Nanometals

Nanoparticles

Zebrafish

Daphnia

the role of dissolution in NM toxicity. Exposures were performed on adult and juvenile life stages of the zebrafish (Danio rerio), adult Daphnia pulex, Ceriodaphnia dubia neonates, and algae (Pseudokirchneriella subcapitata). These taxa were chosen to cover a range of trophic and taxonomic levels. Danio rerio is a widely used model organism in biological research and is increasingly becoming a model species for xenobiotic testing and toxicogenomics. This species is small, easily cultured, and has a diverse background of research that makes them an attractive model species. Daphnia pulex and C. dubia are freshwater cladocerans that naturally feed on particles, including bacteria, algae, and yeast. They are widely used as test organisms to assess the acute toxicity of environmental contaminants either dissolved or dispersed in water [8]. Because daphnids are filter feeders, they may be more susceptible than fish to nanometal exposure. The chlorophyte alga P. subcapitata was included to test the response of primary producers. The results of the present study demonstrate that significant differences in toxicity of nanometals to aquatic organisms exist and that these differences appear to be driven by nanoparticle composition and species of organism.

INTRODUCTION

Recently, increasing production rates and utilization of nanoparticles have raised concerns that releases of engineered nanomaterials (NMs) may pose a serious environmental threat [1]. Because aquatic ecosystems likely will serve as terminal sinks for NMs introduced to natural systems [2], research investigating the potential for environmental toxicity as a result of NMs has become a priority. Although still in its infancy, research concerning the environmental implications of NMs is expanding rapidly [3–7], but so far, much of the current research has focused on carbonaceous NMs (carbon nanotubes and fullerenes). Metallic NMs, referred to hereafter as nanometals, are being used in industrial settings and have diverse applications that likely will result in release to the environment (antimicrobial coatings, fuel cells, water electrolysis, air and water purification, and biomedical imaging contrast agents). Given the likelihood of exposure and that many of the metals used to formulate nanometals are toxic to aquatic species, a need exists for research evaluating the potential toxic effects of nanometals to aquatic species. The goal of the present research was to assess the toxicity of a variety of nanometals in model aquatic organisms to determine how particle composition and species affect toxicity. The toxicity of NMs and dissolved metal solutions were compared to determine the relative toxicity of NMs and to identify

MATERIALS AND METHODS

Materials used Copper, silver, nickel, and cobalt nanopowders were provided gratis by Quantum Sphere (Santa Ana, CA, USA). These particles were produced by gas-phase condensation and are coated with thin layers (thickness, ⬃2–5 nm) of metal oxide. Nominal ranges of particle diameters as provided by the manufacturer were as follows: Silver, 20–30 nm; copper, 15–45 nm; nickel, 5–20 nm; and cobalt, 10–20 nm. Nanoaluminum

* To whom correspondence may be addressed ([email protected]). The research has not been reviewed by the National Science Foundation and does not necessarily represent the views of the funding agency. Published on the Web 5/20/2008. 1972

Environ. Toxicol. Chem. 27, 2008

Toxicity of metal nanoparticles in aquatic organisms

powder was purchased from NovaCentrix (Austin, TX, USA) and had a nominal particle diameter of 51 nm. Nanoparticulate TiO2 (P-25) was obtained from Degussa (Essen, Germany). The P-25 particles had a nominal diameter of 30 nm and were comprised of 20% rutile and 80% anatase TiO2. Metal salts used to compare the toxicity of nanoparticulate and soluble metals were CuCl2, AgNO3, CoCl2, NiCl2, and AlCl3. These metal salts were purchased from Sigma Chemical (St. Louis, MO, USA) and were of the highest purity available (⬎99%). For soluble metals, all concentrations are reported as the mass of metal added.

Nanometallic particle characterization Before use, nanometallic particles were first characterized as received in the Particle Engineering Research Center at the University of Florida (Gainesville, FL, USA). For each nanometal, the specific surface area, particle size distribution, polydispersity, zeta potential, and particle dissolution were measured. Primary particle sizes were determined from scanning-electron micrographs of each particle type. A minimum of 50 individual particles were used to estimate diameter. Specific surface area was measured with the Brunauer, Emmett, and Teller method. Density and specific surface area were measured using a Nova 1200 (Quantachrome, Syossett, NY, USA). Zeta potential was measured with a Zeta Reader Mk 21-II (Zeta Potential Instruments, Bedminster, NJ, USA) using particles suspended in moderately hard freshwater (pH 8.2) for toxicity evaluation (see below for water chemistry). To approximate the particle size distributions of nanoparticle suspensions in toxicity tests, particles were suspended in moderately hard freshwater at 270 mg/L and dispersed with a probe sonicator having an output of 6 W at 22.5 kHz. Particle size distributions of these suspensions were determined with a Coulter LS 13 320 (Beckman Coulter, Fullerton, CA, USA). Measurements were made at high concentrations to allow accurate assessment of particle size distributions. Dissolution in the absence of organisms was determined by suspending the NMs in moderately hard water and stirring gently for 48 h. Dissolved metals were defined as those present in the supernatant of samples centrifuged at 100,000 g for 30 min. Metal concentrations were determined by inductively coupled plasma–optical emission spectroscopy using a PerkinElmer 3200 (Waltham, MA, USA). Three repetitions were measured per sample, and the average was used to calculate the metal concentration from a seven-point standard curve.

1973

studies, additional experiments were conducted spanning the relevant concentration range. For definitive experiments, five concentrations were tested: The estimated median lethal concentration (LC50) from range-finder tests, and 0.6-, 0.36-, 1.67-, and 2.78-fold the estimated LC50. For each test, dosing was performed by preparing stock suspensions of nanometals (except silver) by adding 10 mg of nanometal powder to 10 ml of ultrapure Milli-Q威 water (Millipore, Billerica, MA, USA) and then sonicating with a probe sonicator for six halfsecond pulses at an output of 6 W and 22.5 kHz. Silver nanoparticle suspensions were prepared in a similar manner, except that a 0.5% sodium citrate solution was used to help stabilize the suspensions. Next, required amounts of nanometal suspensions were added to test chambers by pipetting. The volume of stock solution never exceeded 0.5% of total exposure volume, and addition of sodium citrate had no effect on viability of the organisms being tested. Test solutions were prepared immediately before initiation of the exposures to minimize particle dissolution, aggregation, and sedimentation. Static toxicity tests were performed to generate environmental indices of interest to environmental scientists. For each of the tested nanometals, the LC50 was calculated using the trimmed Spearman–Karber method. Concentrations in soluble metal salt exposures were calculated based on the mass of the metal ion rather than the metal salt. For all exposures, control survival was required to exceed 90%; if control survival was less than 90%, the test was repeated.

Zebrafish bioassays Zebrafish assays were performed as 48-h static renewal bioassays following American Society for Testing and Materials guidelines [9]. Wild-type zebrafish were purchased from Ekk-Will (Gibsonton, FL, USA) and were maintained in the Aquatic Toxicology Laboratory at the University of Florida. Adult female zebrafish (four individuals per replicate, with three replicates per concentration) were placed in large beakers containing 2 L of filtered (pore size, 0.45 ␮m) test water with constant, gentle aeration. Exposures were monitored constantly, and dead individuals removed every 12 h. Zebrafish fry (age, ⬍24 h) were exposed in 12-well plates containing 4 ml of test solution and one fry per well. Exposures were performed in triplicate and monitored every 12 h for survival. Survival was assessed by direct observation under light microscopy; death was defined as lack of a visible heartbeat for 30 s and lack of response to physical prodding.

Test water Unless otherwise noted, toxicity tests were conducted in moderately hard freshwater from the University of Florida Aquatic Toxicology Facility (Gainesville, FL, USA). This facility uses dechlorinated municipal water and forced air aeration. Test water had the following characteristics: Dissolved oxygen concentration, 8.5 to 8.9 mg/L; pH 8.2 ⫾ 1; hardness, 142 ⫾ 2 mg/L as CaCO3; conductivity, 395 ␮S; measured total un-ionized ammonia, less than 0.5 mg/L.

Toxicity testing Initially, range-finding experiments were performed with a gradient of concentrations up to 10 mg/L. Essentially no data exist regarding actual environmental concentrations of NMs, and it is unclear how much of a given NM will eventually enter the environment. Therefore, environmental concentrations are unlikely to exceed 10 mg/L. Based on results of initial

Daphnia pulex bioassays These exposures were performed as 48-h static renewal bioassays following American Society for Testing and Materials guidelines using adult daphnids [9]. Cultures were maintained continuously for six months, and daphnids were fed YCT (yeast, cerophyll, trout chow) media ad libitum daily before use. For each concentration, five adult daphnids were placed into 500-ml beakers containing 200 ml of test solution in filtered (pore size, 0.45 ␮m) test water (see above). Negative treatment controls were performed simultaneously. Each concentration contained four replicate exposures. Exposures were performed at 25⬚C with an ambient photoperiod (14:10-h light: dark). The beakers were monitored daily, and dead individuals were removed. Death was assessed by lack of movement or response to gentle prodding. Daphnia pulex were not fed for the duration of the exposures.

1974

Environ. Toxicol. Chem. 27, 2008

R.J. Griffitt et al.

Table 1. Characterization of metallic nanoparticles used in toxicity researcha Specific surface area (m2/g)

Primary particle size distribution (nm)

Zeta potential (mV)

Polydispersity

Silver

14.53

26.6 ⫾ 8.8b

⫺27.0

0.161

Copper Aluminum

30.77 27.26

26.7 ⫾ 7.1 41.7 ⫾ 8.1

⫺0.69 ⫹18.2

0.179 0.224

Cobalt

36.39

10.5 ⫾ 2.3

⫹17.8

0.238

Nickel

50.56

6.1 ⫾ 1.4

⫹21.9

0.292

TiO2

45.41

20.5 ⫾ 6.7

⫺25.1

0.197

Nanometals

Major particle diameters observed in suspension (nm) 44.5 216 94.5 447.1 4442 223.9 742 44.9 206.1 446.1 220.8 687.5

Dissolution 48 h after resuspension (% of mass) 0.07 0.03 0.05 1.2 3.7 NM

a

Primary particle sizes were determined from scanning-electron micrographs of materials. Following resuspension in the moderately hard freshwater used for toxicity testing, the zeta potential, polydispersity index, particle size dispersion, and particle dissolution were measured. All particles exhibited aggregation in suspension, as indicated by the increases in particle diameters following suspension. NM ⫽ not measured. b Mean ⫾ standard deviation.

Ceriodaphnia dubia assays

Exposure characterization for toxic NMs

These assays were performed as 48-h static renewal bioassays using neonates (age, ⬍24 h). Before the start of each assay, the neonates were separated from the adult daphnids and fed a mixture of 7 ml of YCT and 7 ml of algae (P. subcapitata) for every liter of invertebrate culture. After feeding, 10 neonates were transferred to each test container (30ml plastic cups) using a small, wide-mouth, plastic pipette to minimize the transfer of culture water. The sample dilutions, if necessary, were prepared with filtered test water, and 20 ml of the sample or its dilutions were added to cups containing the neonates. Filtered test water was used as the negative control, and all test containers were then placed in a water bath at 20 ⫾ 2⬚C for 48 h while exposing the neonates to ambient lighting. The assay endpoint was death/immobilization.

For nanometals demonstrating acute toxicity, water samples were taken from the center of the water column in exposures at the LC50 to measure total and dissolved metal after 48 h. Zebrafish were exposed to nanocopper and nanosilver at the calculated LC50s for 48 h. Daphnia pulex was exposed to nanocopper, nanosilver, and nanonickel for 48 h at the calculated LC50s. Total metal content was determined in water samples that were digested with nitric acid as described previously [11]. To determine soluble metal concentrations, 10 ml of test solution were filtered (pore size, 20 nm; Anotop 25; Whatman, Maidstone, UK), and ultrapure nitric acid (Optima; Fisher Scientific, Waltham, MA, USA) was added to the filtrate to a final concentration of 2%. Metal concentrations in these samples were determined by inductively coupled plasma–mass spectrometry (X Series; Thermo Electron, Waltham, MA, USA) using a seven-point standard curve and iridium as an internal standard.

Pseudokirchneriella subcapitata growth inhibition assay The chronic toxicity of metallic nanoparticle suspensions to algae was evaluated using a 96-h P. subcapitata growth inhibition assay described by the U.S. Environmental Protection Agency [10]. For each tested nanometal, algal growth media were prepared to produce a concentration gradient before being inoculated with a similar volume of algal culture. These experiments were conducted in triplicate and run in parallel with three negative controls (algal medium without nanometals). Samples were incubated for 96 h in a water bath at room temperature and under controlled light intensity (86 ⫾ 8.6 ␮E/m/s). The algal growth was assessed by measurement of chlorophyll a, and the percentage inhibition was determined using the following equation: % Inhibition ⫽ 1 ⫺ (As /A0 ) · 100 where As is the sample absorbance and A0 is the negative control absorbance. Next, a plot of sample concentrations versus the percentage inhibition was used to perform a regression analysis in the linear potion of the graph to determine the median effective concentration (EC50) according to the following equation: EC50 ⫽ (50 ⫺ yi)/S where yi is the intercept and S the slope of the above-mentioned regression.

RESULTS

Nanometallic particle characterization Specific surface areas, primary particle sizes, zeta potentials, and polydispersity for the different nanometal powders are listed in Table 1. All particles exhibited some degree of aggregation following suspension in moderately hard freshwater. Particle size distributions for suspensions of the six nanometals tested are shown in Figure 1. Whereas aggregation increased mean particle diameter and much of the particle mass is present as larger aggregates, the nanocopper, nanosilver, and nanocobalt still had significant numbers of particles less than 100 nm in diameter. Particle dissolution in the absence of organisms for the two most toxic nanometals, silver and copper, was very low (Table 1). Less than 1% by mass of the original dose was present in the dissolved form after 48 h. Nanonickel and nanoaluminum showed slightly higher dissolution rates; approximately 1 and 4%, respectively, of the particle mass had dissolved after 48 h.

Toxicity testing Nanocopper and nanosilver were toxic to all organisms tested, with calculated LC50s ranging from 0.04 mg/L (D. pulex)

Environ. Toxicol. Chem. 27, 2008

Toxicity of metal nanoparticles in aquatic organisms

1975

tions of nanoparticulate silver (⬎90%) and copper (50%) were removed from the water column within 48 h. These results also indicate that the toxicity of nanocopper and nanosilver in both zebrafish and D. pulex is unlikely to be attributable to particle dissolution (Fig. 3). For zebrafish, the soluble concentrations of both silver and copper present in a LC50 exposure were well below the 48-h LC50s for the soluble metals (Fig. 3). In D. pulex exposures, silver dissolution was, again, much lower than concentrations required to produce acutely lethal effects, whereas dissolution of copper nanoparticles appears to produce sufficient soluble ions to account for 10 to 15% of the observed mortality (Fig. 3). In contrast, nickel was present largely in the dissolved form in the D. pulex exposures and reached approximately 70% of the soluble nickel LC50. It is possible that dissolution of nanonickel is responsible for a majority of the toxicity of nanonickel to D. pulex. DISCUSSION Fig. 1. Particle size distribution of the nanometals used in the present study. Nanometal powders were suspended in moderately hard freshwater at a concentration of 270 mg/L, and size distributions were measured using a Coulter LS instrument (Beckman Coulter, Fullerton, CA, USA). Data are presented as the differential volume percentage (dV%).

to 7.2 mg/L (D. rerio fry) for nanosilver and from 0.06 mg/L (D. pulex) to 0.94 mg/L (D. rerio adults) for nanocopper. Daphnia pulex also was acutely susceptible to nanonickel (LC50, 3.89 mg/L) (Tables 2 and 3). All the nanometals tested caused toxicity in C. dubia and P. kirchneriella (Table 3) at 48 and 96 h, respectively. Based on observed LC50s, invertebrates (daphnids and algae) were markedly more susceptible to toxicity from nanometals compared with either stage of zebrafish. Overall, nanometal suspensions generally were less toxic than solutions of metal salts on the basis of mass of metal added, although D. rerio fry were more sensitive to nanoparticulate forms of copper and silver than to the soluble forms of the same metal (Table 2).

Exposure characterization Organism/nanometal combinations demonstrating significant toxicity were further investigated to determine the extent to which dissolution of the nanoparticle suspensions were mediating the observed toxic effects. At the end of the exposure, water samples were taken and analyzed as described above to determine total and soluble metal concentrations remaining in the water column (Fig. 2). The results indicate that large frac-

The results of the present study demonstrate that nanometals are capable of causing acute toxicity in multiple aquatic species; however, toxicity differs significantly with the particle composition and the species tested. The toxicity of nanometals does not appear to be a generic response to exposure to nanosized particles; rather, it seems that particular nanometals have an intrinsic property that confers toxicity. No relationship was apparent between size, surface area, or zeta potential (Table 1) and toxicity for any organism surveyed. For example, nanosilver and nanocopper produced toxicity in all species. They differ substantially in zeta potential, however, and although their sizes are similar, nanotitania of a similar size produced no toxicity. Within a given species, chemical composition of a particle appears to be the most important factor in toxicity. In the species tested, nanosize copper and silver were consistently the most toxic, with 48-h LC50s of less than 10 mg/L in zebrafish (both life stages) and less than 1 mg/L for invertebrates (daphnids and algae) (Table 3). Additionally, however, silver and copper are the most toxic when present in a soluble form, and toxicity because of dissolution must be differentiated from particle-mediated toxicity. The role of dissolution in toxicity caused by nanometals also varies with particle composition and species. Dissolution of nanosilver and nanocopper during exposures is relatively low, and the observed mortality is unlikely to be attributable solely to particle solubilization (Fig. 3). Additionally, we have shown previously [12] that soluble copper accounts for only approximately 15% of the observed toxic responses in adult zebrafish (Fig. 3b). In contrast, the toxicity of nanonickel in daphnids appears to result largely from the presence of dis-

Table 2. Toxicity of metallic nanoparticles and soluble metals in zebrafish (Danio rerio) using 48-h static bioassaysa 48-h LC50 of nanoparticles (mg/L) Metal Silver Copper Aluminum Cobalt Nickel TiO2 a

Adult 7.07 (6.04–8.28) 0.94 (0.75–.17) ⬎10 ⬎10 ⬎10 ⬎10

Juvenile 7.20 (5.9–8.6) 0.71 (0.54–0.93) ⬎10 ⬎10 ⬎10 ⬎10

48-h LC50 of soluble metal (mg/L) Adult 0.0222 (0.0195–0.026) 0.13 (0.11–0.15) 7.92 (7.11–8.81) ⬎10 ⬎10 NM

Juvenile ⬎10 1.78 (1.00–3.14) ⬎10 ⬎10 ⬎10 NM

Juvenile zebrafish were younger than 24 h at the initiation of experiments. Values in parentheses are the 95% confidence interval. LC50 ⫽ median lethal concentration; NM ⫽ not measured.

1976

Environ. Toxicol. Chem. 27, 2008

R.J. Griffitt et al.

Table 3. Toxicity of metallic nanoparticles and soluble metals in aquatic invertebratesa LC50 of soluble metals (mg/L)

LC50 of nanoparticles (mg/L) Metal Silver Copper Aluminum Cobalt Nickel TiO2 a

Daphnia pulex adults 0.040 (0.030–0.050) 0.060 (0.050–0.070) ⬎10 ⬎10 3.89 (1.93–7.43) ⬎10

Ceriodaphnia dubia Pseudokirchneriela neonates subcapitata 0.067 0.419 3.99 1.67 0.674 ⬎10

0.19 0.54 8.30 NM 0.35 NM

Daphnia pulex adults 0.008 0.009 3.65 9.72 1.48

(0.007–0.009) (0.007–0.011) (2.49–5.35) (7.94–11.89) (1.01–2.19) NM

Ceriodaphnia dubia neonates 0.16 0.58 153.44 94.66 19.64

(0.11–0.23) (0.42–0.80) (117.26–200.80) (61.94–144.67) (2.68–143.84) NM

Daphnid exposures were performed as static 48-h bioassays. Neonates were younger than 24 h at the start of tests. Algal assays were performed as 96-h growth inhibition tests in Pseudokirchneriella subcapitata. Values in parentheses are the 95% confidence interval. LC50 ⫽ medial lethal concentration; NM ⫽ not measured.

solved nickel (Fig. 3c). Similar results have been reported for nanoparticulate zinc oxide, in which toxicity could be ascribed almost completely to the release of soluble zinc [13]. It is important to note that the presence of organisms can affect particle dissolution. Substantially larger amounts of copper dissolve in incubations with test organisms compared with the absence of test organisms, which is consistent with a previous report [12] and likely results from dissolution by the organisms following ingestion. This highlights the need for careful exposure characterization in experiments conducted with nanometals to clearly define the effects that result from nanoparticulate exposure. The precise mechanisms of toxicity for nanoparticles are largely unknown. Previous research has indicated that copper nanoparticles are very highly reactive, and the toxic effects of nanocopper may result from this reactivity, possibly through metabolic alkalosis or intracellular dissolution of copper nanoparticles leading to very high local ionic copper concentrations [14]. Other researchers have found that certain nanoparticles are capable of producing high levels of different reactive oxygen species (ROS) levels in water [15–17], although it is not clear if the metallic nanoparticles studied here are capable of ROS generation. One exception is nanoparticulate TiO2, which

Fig. 2. Water-column concentration and dissolution during 48-h exposures of daphnids and zebrafish to metallic nanoparticles at the median lethal concentration. For each combination, the initial concentrations of particles and the concentration of total and soluble metal remaining in the water column after 48 h are plotted. Bars represent the mean ⫾ standard deviation (n ⫽ 3).

has been shown to generate high levels of peroxide, although these ROS levels have not been shown to be toxic [18] and nanoparticulate TiO2 was not toxic to any of the organisms in the present study. In developing methods for testing the toxicity of NMs, it is important to identify sensitive species, life stages, and endpoints. The results of the present study demonstrate that susceptibility to acute toxicity of nanometals differs among species (Fig. 3a). Daphnid species tend to be more susceptible to nanometal exposure compared with either zebrafish or algae. It seems likely that this results from the separate feeding strategies employed. Daphnids are particulate filter feeders, and they would be expected to be intimately exposed to large numbers of nanoparticles over the course of the exposure. Whereas the concentrations used in the present study were too low to accurately measure particle count in the exposures, we can attempt to estimate particle number in these exposures. If we assume that the nanocopper particles are 30 nm in diameter, are spherical, and have a density of 8,920 kg/m3, then a 60 ␮g/L monodispersed suspension of nanocopper would contain approximately 400 million nanoparticles per milliliter of test solution. Because D. pulex filter approximately 0.4 ml of water per hour depending on environmental conditions [19], an individual daphnid could be exposed to greater than 190 million nanoparticles per hour. This value is highly theoretical and does not account for particle aggregation and sedimentation among other issues, but it illustrates why daphnids and other filter feeders may be especially susceptible to toxicity from nanometals. Pseudokirchneriella subcapitata, however, was the organism most susceptible to nanonickel, demonstrating the importance of identifying appropriate test organisms for specific particles. Nanoparticulate forms of metals, with the exception of copper in zebrafish fry, were less toxic than soluble forms of metals based on the mass of metal added to each exposure. This suggests that existing regulations based on soluble metals may be adequate to protect aquatic life from nanoparticulate forms of metals. Nanomaterials in aqueous suspensions, however, are dynamic systems undergoing simultaneous dissolution, aggregation, and sedimentation [20,21]. During a 48-h exposure, from 50 to 90% of the initial mass of nanometal added to an exposure may be lost through aggregation and sedimentation (Fig. 2). The amount of material lost from the water column is affected by the species of organism present, as evidenced by differences in daphnid and zebrafish exposures to copper (Fig. 2), perhaps because of differences in organic matter excreted by various organisms. This results in constantly changing exposure parameters that differ between test

Toxicity of metal nanoparticles in aquatic organisms

Environ. Toxicol. Chem. 27, 2008

1977

systems and makes calculating appropriate dose metrics for aquatic exposures difficult. The problem is compounded by the lack of techniques to accurately characterize particles in aqueous suspensions at the concentrations that cause toxicity. Currently, we are aware of no instruments capable of accurately calculating particle size distributions at the low-␮g/L concentrations that are toxicologically relevant. It is critical that the scientific community identifies accurate means of expressing exposure of aquatic organisms to NMs, either by accurate measures of external exposure or by developing measures of internal or effective dose. Therefore, the data of the present study are nominal, and the actual concentration to which the organisms were exposed may be less than reported, with the nanometals being correspondingly more toxic. CONCLUSION

From the results of the present study, we conclude that certain nanometals have the potential to cause toxicity in aquatic organisms. Susceptibility to toxicity from exposure to a given nanometal varies with species, and given the susceptibility of algae and daphnids to nanometals, the potential for ecological impacts likely is greater in the invertebrate and filter-feeding communities compared with that in larger, more visible vertebrate species. Particle dissolution appeared to explain some, but not all, of the observed mortality, indicating that using soluble metal toxicity data to predict nanometal toxic responses may not always be warranted. Toxicity of nanometals appears to be lower than that of soluble forms of the metals, but accurately describing exposures of aquatic organisms to NMs is a challenge that must be addressed. Acknowledgement—The authors gratefully acknowledge Paul Martin and Gil Brubaker of the University of Florida Particle Engineering Research Center for their assistance with particle characterization. This research was funded by grants from the National Science Foundation (BES-0540920) and the University of Florida School of Natural Resources and the Environment. REFERENCES

Fig. 3. Concentration–response curves for toxicity of nanoparticulate and soluble metal species. Data in this figure illustrate the toxicity of nanoparticulate copper and silver in Danio rerio and Daphnia pulex (a) as well as the contribution of dissolution to nanoparticulate toxicity in D. rerio (b) and D. pulex (c). In b and c, the survival curves for the soluble metal are plotted along with the amount of dissolved metal released by dissolution over 48 h. Short dashed lines represent calculated median lethal concentrations (LC50s) for the soluble metals; long dashed lines indicate dissolved concentrations released from nanometals. On each graph, blue lines represent silver responses, red lines copper responses, and black lines nickel responses. Values are presented as the mean ⫾ standard error (n ⫽ 4–5).

1. Oberdorster G, Oberdorster E, Oberdorster J. 2005. Nanotoxicology: An emerging discipline evolving from studies of ultrafine particles. Environ Health Perspect 113:823–839. 2. Moore MN. 2006. Do nanoparticles present ecotoxicological risks for the health of the aquatic environment? Environ Int 32:967– 976. 3. Oberdorster E. 2004. Manufactured nanomaterials (fullerenes, C60) induce oxidative stress in the brain of juvenile largemouth bass. Environ Health Perspect 112:1058–1062. 4. Kashiwada S. 2006. Distribution of nanoparticles in the seethrough medaka (Oryzias latipes). Environ Health Perspect 114: 1697–1702. 5. Lovern SB, Klaper R. 2006. Daphnia magna mortality when exposed to titanium dioxide and fullerene (C60) nanoparticles. Environ Toxicol Chem 25:1132–1137. 6. Zhu S, Oberdorster E, Haasch ML. 2006. Toxicity of an engineered nanoparticle (fullerene, C60) in two aquatic species, Daphnia and fathead minnow. Mar Environ Res 62(Suppl):S5–S9. 7. Templeton RC, Ferguson PL, Washburn KM, Scrivens WA, Chandler GT. 2006. Life-cycle effects of single-walled carbon nanotubes (SWNTs) on an estuarine meiobenthic copepod. Environ Sci Technol 40:7387–7393. 8. Martins J, Oliva Teles L, Vasconcelos V. 2007. Assays with Daphnia magna and Danio rerio as alert systems in aquatic toxicology. Environ Int 33:414–425. 9. American Society of Testing and Materials. 2002. Standard guide for conducting acute toxicity tests on test materials with fishes, macroinvertebrates and amphibians. In Annual Book of ASTM Standards, Vol. 11.06. Philadelphia, PA, pp 249–269. 10. U.S. Environmental Protection Agency. 2002. Short-term meth-

1978

11.

12.

13.

14.

15.

Environ. Toxicol. Chem. 27, 2008

ods for estimating the chronic toxicity of effluents and receiving waters to freshwater organisms. EPA-821-R-02-013, 4th ed. Washington, DC. Barber DS, Ehrich MF, Jortner BS. 2005. The effect of stress on the temporal and regional distribution of uranium in rat brain after acute uranyl acetate exposure. J Toxicol Environ Health A 68: 99–111. Griffitt R, Weil R, Hyndman K, Denslow N, Powers K, Taylor D, Barber D. 2007. Exposure to copper nanoparticles causes gill injury and acute lethality in zebrafish (Danio rerio). Environ Sci Technol 41:8178–8186. Franklin NM, Rogers NJ, Apte SC, Batley GE, Gadd GE, Casey PS. 2007. Comparative toxicity of nanoparticulate ZnO, bulk ZnO, and ZnCl2 to a freshwater microalga (Pseudokirchneriella subcapitata): The importance of particle solubility. Environ Sci Technol 41:8484–8490. Meng H, Chen Z, Xing G, Yuan H, Chen C, Zhao F, Zhang C, Zhao Y. 2007. Ultrahigh reactivity provokes nanotoxicity: Explanation of oral toxicity of nanocopper particles. Toxicol Lett 175:102–110. Badireddy AR, Hotze EM, Chellam S, Alvarez P, Wiesner MR. 2007. Inactivation of bacteriophages via photosensitization of fullerol nanoparticles. Environ Sci Technol 41:6627–6632.

R.J. Griffitt et al. 16. Pickering KD, Wiesner MR. 2005. Fullerol-sensitized production of reactive oxygen species in aqueous solution. Environ Sci Technol 39:1359–1365. 17. Zhu X, Zhu L, Li Y, Duan Z, Chen W, Alvarez PJ. 2007. Developmental toxicity in zebrafish (Danio rerio) embryos after exposure to manufactured nanomaterials: buckminsterfullerene aggregates (nC60) and fullerol. Environ Toxicol Chem 26:976– 979. 18. Xia T, Kovochich M, Brant J, Hotze M, Sempf J, Oberley T, Sioutas C, Yeh JI, Wiesner MR, Nel AE. 2006. Comparison of the abilities of ambient and manufactured nanoparticles to induce cellular toxicity according to an oxidative stress paradigm. Nano Lett 6:1794–1807. 19. Buikema AL. 1973. Filtering rate of the cladoceran, Daphnia pulex as a function of body size, light, and acclimation. Hydrobiologia 41:515–527. 20. Brant J, Lecoanet H, Wiesner MR. 2005. Aggregation and deposition characteristics of fullerene nanoparticles in aqueous systems. Journal of Nanoparticle Research 7:545–553. 21. Murdock RC, Braydich-Stolle L, Schrand AM, Schlager JJ, Hussain SM. 2008. Characterization of nanomaterial dispersion in solution prior to in vitro exposure using dynamic light scattering technique. Toxicol Sci 101:239–253.