TiO2 Nanoparticle Exposure and Illumination during Zebrafish ...

21 downloads 0 Views 1MB Size Report
Mar 19, 2013 - Richard E. Peterson,. †,§ and Warren Heideman ..... addition, TiO2NPs were frequently found in the liver (part E of. Figure 7). Finally, whereas ...
Article pubs.acs.org/est

TiO2 Nanoparticle Exposure and Illumination during Zebrafish Development: Mortality at Parts per Billion Concentrations Ofek Bar-Ilan,† Connie C. Chuang,† Denise J. Schwahn,‡ Sarah Yang,§ Sanjay Joshi,# Joel A. Pedersen,§,∥ Robert J. Hamers,⊥ Richard E. Peterson,†,§ and Warren Heideman†,§,* †

Pharmaceutical Sciences Division, School of Pharmacy, ‡Research Animal Resources Center, Chemistry Department, §Molecular and Environmental Toxicology Center, ∥Soil Sciences Department, and ⊥Department of Chemistry, University of Wisconsin, Madison, Wisconsin 53706, United States # Deptartment of Industrial and Manufacturing Engineering, Penn State University, University Park, Pennsylvania 16802, United States S Supporting Information *

ABSTRACT: Photoactivation of titanium dioxide nanoparticles (TiO2NPs) can produce reactive oxygen species (ROS). Over time, this has the potential to produce cumulative cellular damage. To test this, we exposed zebrafish (Danio rerio) to two commercial TiO2NP preparations at concentrations ranging from 0.01 to 10 000 ng/mL over a 23 day period spanning embryogenesis, larval development, and juvenile metamorphosis. Fish were illuminated with a lamp that mimics solar irradiation. TiO2NP exposure produced significant mortality at 1 ng/mL. Toxicity included stunted growth, delayed metamorphosis, malformations, organ pathology, and DNA damage. TiO2NPs were found in the gills and gut and elsewhere. The two preparations differed in nominal particle diameter (12.1 ± 3.7 and 23.3 ± 9.8 nm) but produced aggregates in the 1 μm range. Both were taken up in a dose-dependent manner. Illuminated particles produced a time- and dose-dependent increase in 8-hydroxy-2′-deoxyguanosine DNA adducts consistent with cumulative ROS damage. Zebrafish take up TiO2NPs from the aqueous environment even at low ng/mL concentrations, and these particles when illuminated in the violet-near UV range produce cumulative toxicity.



INTRODUCTION Nanoscale materials are being put to use in countless biomedical and environmental applications.1 These uses bring increased exposure of humans and wildlife to these materials. The unique physicochemical properties that make nanomaterials so desirable can in some cases cause adverse biological effects.2−6 Titanium dioxide nanoparticles (TiO2NPs) are photocatalysts. Photons at and just beyond the violet end of the visible spectrum excite electrons from the valence to the conduction band generating reactive electron−hole pairs that can in turn interact with oxygen and water to form reactive oxygen species (ROS).7−10 TiO2NPs exist in several lattice forms. The two most common crystalline forms, anatase and rutile, have similar band gap energies that govern photoactivation (rutile 3.23 eV/λ = 385 nm, anatase 3.06 eV/λ = 400 nm).8,10,11 This photocatalytic activity can be useful but also represents a potential hazard. We previously found that exposure to TiO2NPs caused increased ROS and toxicity in developing zebrafish (Danio rerio) embryos. This toxicity was completely dependent on illumination.12 The previous study lasted only 5 days and used TiO2NPs at 100−1000 μg/mL. However, the damage to macromolecules produced by ROS can be continuous and cumulative. For this reason, we expected that, given sufficient time, lower, more environmentally relevant concentrations of TiO2NPs could produce significant toxicity. © 2013 American Chemical Society

To test this, we examined growth and survival of zebrafish exposed to aqueous suspensions of two different commercial preparations of TiO2NPs (0.01−10 000 ng/mL) from 0 to 23 days post fertilization (dpf) with illumination. This period of exposure encompasses embryogenesis, larval development, and metamorphosis from the larval into the juvenile/adult form. Both preparations produced significant mortality at 1 ng/mL (1 ppb).



MATERIALS AND METHODS Aeroxide P25 TiO2NPs (Evonik Degussa; Essen, Germany; reported diameter: 21 nm; 3:1 anatase/rutile): referred to as Degussa. Sun Innovations TiO2 NPs (Fremont, CA; reported size: 5−10 nm; anatase): abbreviated as Sun Innovations. Anatase/rutile ratios were confirmed by X-ray diffraction to show that the Degussa particles were 78% anatase/22% rutile (similar to the nominal 75:25 ratio). The Sun Innovations particles were all in the anatase form as reported by the manufacturer. Stock suspensions (2 mg/mL, 40 mL in aquarium water) were made every 24 h by sonication for 60 min in a G112SP1T Received: Revised: Accepted: Published: 4726

November 20, 2012 March 18, 2013 March 19, 2013 March 19, 2013 dx.doi.org/10.1021/es304514r | Environ. Sci. Technol. 2013, 47, 4726−4733

Environmental Science & Technology

Article

spectroradiometer was programmed to take the average of three consecutive readings and to record the data automatically. The IBM PC Interface program (LICOR proprietary software) was used to analyze the data and report values on Spectral Irradiance, Correlated Color Temperature, and Photosynthetic Photon Flux Density. Data was exported to Excel for further analysis and plotting the spectral power distribution graph. This lamp is specifically manufactured for the husbandry of photodependent organisms such as corals and some marine clams (http://www.xmlighting.com/index.aspx). A comparison between the spectral intensity of the metal halide lamp illumination at 45 cm and sunlight (ASTM G-173-03, rredc.nrel.gov/solar/spectra/am1.5/ASTMG173/ ASTMG173.html) is provided in Figure S1 of the Supporting Information. While far brighter than the ambient fish facility lighting, the experimental lighting does not approach the intensity of daylight. Toxicity Assessment: Survival. We considered n equal to 1 set of fish exposed together in a beaker starting with 16 fish per beaker. For each data point, mean survival was based on data from 39 to 54 independently treated and cared for beakers of fish (n = 39−54). To obtain sufficient numbers of surviving fish for analysis, more beakers were required at higher exposure concentrations. Toxicity Assessment: Growth, Development, and Malformation. Fish were immobilized in 3% methylcellulose at 23 dpf and photographed live in a lateral orientation at 2× magnification with a MicroFire camera (Optronics, Goleta, CA, USA) mounted onto a Leica MZ16 stereomicroscope (Meyer Instruments, Houston, TX, USA). Micrographs were analyzed for: presence of fin rays, fin buds, or a bilobed swim bladder.13 Incidence of jaw and caudal fin malformations were also recorded. For each of 10 replicate experiments, 6−10 fish were assessed: n = 10. Inductively Coupled Plasma-Optical Emission Spectrometry. Samples were digested in 5 M NaOH at 60 °C for 24 h producing a greater than 95% yield, diluted, and brought to pH 2.12,14 Samples were analyzed by inductively coupled plasma-optical emission spectrometry (ICP-OES) along with a Ti standard curve using a Varian Vista-MPX spectrometer (Varian, Inc., Walnut Creek, CA, USA) and default instrument conditions (0.75 L/min nebulizer flow, 15.0 L/min plasma flow). Emission was recorded at 334.188, 334.941, 336.122, and 337.280 nm and averaged. Data were fit to a quadratic curve. Each sample consisted of 10 pooled fish that were euthanized, collected, and washed with aquarium water, and 4 aliquots from each sample were submitted for ICP-OES measurement to produce an average value for that sample. The exposure, collection, and analysis experiment was replicated 3 times, and the values from each experiment were averaged (n = 3). TEM. Fish were washed in water, euthanized, washed in 1× phosphate buffer saline (PBS), and fixed in 2.5% glutaraldehyde/2% paraformaldehyde in PBS overnight at 4 °C. Samples were postfixed in 1% osmium tetroxide in PBS for 2 h at 25 °C, dehydrated in a graded ethanol and propylene oxide series, and embedded in Spurr’s epoxy resin. Samples were sectioned (60− 80 nm) using a Leica EM UC6 ultramicrotome, collected onto pioloform coated slot grids, and micrographs were captured with a Philips CM120 EM (80 kV; Philips, Eindhoven, The Netherlands) and SIS MegaView III digital camera (Olympus, Munster, Germany). Samples were left unstained to reveal TiO2NPs. In each of four replicate experiments, sections from four individual fish were examined per treatment (n = 4).

bath sonicator (Laboratory Supplies Co., Hicksville, NY) running at 300 W and 80 kHz. After dilution and before any use, each stock suspension was sonicated for 3 min. Tricaine was from Sigma Aldrich (St Louis, MO). Physicochemical Characterization of TiO2NPs. For DLS measurements, autocorrelation functions were accumulated and particle diffusivities were calculated using a Nanosizer ZS (Malvern Instruments Ltd., Worcestershire, UK). For each TiO2NP concentration and treatment condition tested, 3 samples were collected and 10 DLS measurements were made on each sample. Particles were collected immediately after preparation and deposited on carbon-coated copper grids for transmission electron microscopy (TEM) with a Philips CM120 EM (80 kV; Philips, Eindhoven, The Netherlands) and SIS MegaView III digital camera (Olympus, Munster, Germany). Zebrafish Embryo Husbandry. Zebrafish embryos (AB strain) were kept at 27−30 °C in buffered water consisting of reverse osmosis purified water to which 60 mg/L Instant Ocean Salts were added (Aquarium Systems, Mentor, OH) with a 14 h/10 h light/dark cycle, in all experiments throughout the study. Fish were fed AP100 (Gardners, PA) and Artemac (Aquafauna Bio-Marine, Inc.; Hawthorne, CA); dead fish were removed twice per day and recorded. Water was changed daily. Fish were euthanized in Tricaine (1.67 mg/mL). The protocol for zebrafish use and maintenance was approved by the Research Animal Resources Center of the University of Wisconsin-Madison (protocol # M00489). Fish were bred in 10 gallon aquaria using 10 males and 10 conditioned females per aquarium. Each of these spawning tanks produced approximately 1000 eggs per spawn. Exposure. AB zebrafish embryos were collected immediately after fertilization, sorted to remove feces and infertile eggs, placed into 24-well plates (4 embryos in 1 mL water/well), and treated with graded concentrations of waterborne Degussa or Sun Innovations TiO2NPs. At 5 dpf they were transferred to 100 mL dosing suspension in beakers with 16 fish per beaker. Because water was renewed daily to remove waste, the dosing suspensions were renewed every day. No dispersants were used at any point. Fish were housed and fed as described above in all experiments. Illumination. The illuminated samples were placed under a 250 W blue-spectrum metal halide lamp (XM 250 W, 10 000 K; electronic ballast; hellolights.com) designed to replicate the slightly blue sunlight in shallow (∼1 m) water. Fish were placed 45 cm below this lamp for 14 h each day. Temperatures were maintained at 27−30 °C for both illuminated and nonilluminated groups. The spectral irradiance of the mogul base XM 250 W 10 000 K metal halide lamp was measured using a LiCOR LI-1800/12 portable spectroradiometer with a standard cosine receptor. It is capable of measurement of spectral power from 300 to 850 nm at 2 nm intervals. The specific instrument used was calibrated for use from 310 to 850 nm by the manufacturer prior to use. The lamp was mounted in a fixture that positions the center of the lamp’s arc tube 18 in. from the cosine receptor. To eliminate interference from stray incident light and reflector related bias, the lamp was measured without reflectors in a dark room with the walls painted in dull black. A magnetic ballast (ANSI M58) manufactured by Magnetek (part #1110-247SCTC) was used to fire the lamp. The lamp was allowed to warm up for 20 min before recording the readings. The 4727

dx.doi.org/10.1021/es304514r | Environ. Sci. Technol. 2013, 47, 4726−4733

Environmental Science & Technology

Article

Dynamic Light Scattering (DLS). Particle aggregation was analyzed immediately after preparation of the TiO2NP dosing suspension and at the end of 24 h when exposure suspensions were replaced with fresh suspensions made from new master stocks. Therefore, the DLS data reflects the condition of the material immediately after preparation for use and at the end of its use as an exposure suspension. For DLS measurements, autocorrelation functions were accumulated and particle diffusivities were calculated. The diffusivities were used to calculate hydrodynamic diameters of TiO2NPs in the water prepared for fish husbandry by DLS using a Nanosizer ZS (Malvern Instruments Ltd., Worcestershire, UK). For each TiO2NP concentration and treatment condition, 3 samples were collected and 10 DLS measurements were made on each sample. Whereas number-weighted DLS measurements can overestimate the size of aggregates,15 the primary goal of these experiments was to confirm that the individual particles do form aggregates in water suspension. 8-Hydroxy-2′-deoxyguanosine ELISA. The concentration of 8-hydroxy-2′-deoxyguanosine (8-OHdG) DNA adducts was measured by ELISA. Fish were collected at 23 dpf, and DNA was extracted according to a standard protocol listed in The Zebrafish Book,16 digested using nuclease P1, and then used with a 96-well competitive assay kit according to the manufacturer’s instructions (StressMarq Biosciences, Victoria, B.C., Canada; two dilutions run in duplicate). Plates were read using a SpectraMax M2 microplate reader (Molecular Devices, Sunnyvale, CA). For each experiment, 9−12 pooled fish were used per extract, each extract was measured in quadruplicate by ELISA, and the values were averaged. These average values were then averaged over 6 replicate experiments (n = 6). Data Analysis. Statistica 7 software was used for all statistical analyses. Data were evaluated for homoscedasticity using Levene’s median test, and one-way analysis of variance was employed to identify effects associated with exposure conditions and treatments. Tukey’s test was used for pairwise multiple comparisons. Level of significance for all analyses was p < 0.05 unless otherwise stated.

Figure 1. Survival of zebrafish exposed to Degussa and Sun Innovations TiO2NPs. Zebrafish were exposed to the indicated concentrations of (A) Degussa or (B) Sun Innovations TiO2NPs with illumination and daily water exchanges for 23 days as described in Materials and Methods, and survival was monitored. Results are presented as mean ± standard error of the mean. For both panels, the survival for all treated groups was significantly lower than the control group (p ≤ 0.05) at the 8 day point and thereafter. For these experiments, n is equal to an individual group of 16 fish exposed to the condition, so for each point n = 39−54 groups.



RESULTS TiO2NP Exposure and Survival. We selected a metal halide lamp doped to provide light that mimics the spectrum of solar illumination as it occurs approximately 1 m below the surface of clear water. Zebrafish were continuously exposed to graded concentrations of TiO2NP suspensions and illumination (14 h per day) for 23 days spanning initial embryogenesis through larval development. Survival over time is shown in Figure 1. All concentrations tested ranging from 1 to 10 000 ng/mL produced significant mortality compared to the control. In the absence of illumination, none of the concentrations produced significant toxicity (Figure S2 of the Supporting Information). The period of embryogenesis lasts until the fish hatch at 2−3 days. After this, the fish enter the larval phase. In the larval phase, individuals grow at different rates, based in part on a positive feedback mechanism in which the strongest fish are able to collect more food, and thus become stronger yet. The timing of metamorphosis is therefore variable, occurring at over an approximately one week period, dependent on the rate of growth of the individual fish. Metamorphosis is generally complete at approximately 3 weeks. During metamorphosis, numerous changes in morphology occur to make the fish

appear less tadpolelike and more like the adult fish. During this period, many fish die off despite ideal husbandry. Those fish that do survive metamorphosis become outwardly like small adults but remain as juveniles until sexual maturity at approximately 3−4 months of age. We found that the exposure to TiO2 nanoparticles and illumination accelerated the die-off during the metamorphosis phase in which the fish are vulnerable. Interestingly, in this and other experiments, we find that the bright illumination simulating daylight consistently reduces die-off during metamorphosis by approximately 10%. Figure 2 shows dose response plots for mortality in fish exposed for 23 days. The data is presented in two different ways. The open symbols show the mortality data as it was presented in Figure 2 without correction. Because in these experiments by 23 dpf we had significant loss of fish in all groups, including the controls, we also plotted the data after subtracting the numbers of fish dying in the control groups. This allows the graph to start at zero, which is necessary for most types of dose−response analysis. 4728

dx.doi.org/10.1021/es304514r | Environ. Sci. Technol. 2013, 47, 4726−4733

Environmental Science & Technology

Article

Figure 3. TiO2NP exposure causes photodependent delays in development, decreased growth, and tissue malformation. Zebrafish were exposed to TiO2NPs and illumination as described in the Materials and Methods. Representative micrographs of 23-dpf zebrafish are shown. Fish were exposed to illumination in all cases along with (A) water as a control, (B) Degussa TiO2NPs (1 ng/mL), (C) Sun Innovations TiO2NPs (1 ng/mL). Representative effects of the TiO2NPs are indicated. The red arrows show the normal bilobed swim bladder in the control and the single-lobed swim bladder in the treated fish. The small blue points indicate the row of developing melanophores making an unbroken line in the normal fish, and underdeveloped in the treated fish. Fin rays in the normal fish are indicated by a white arrow, and the angle of the notochord as it intersects the caudal fin is indicated by a black line and blue arrow. The black arrow indicates the snout shortening and related craniofacial malformations caused by treatment.

Figure 2. Dose−response curves for zebrafish exposed to illuminated TiO2NP. Zebrafish were exposed to the indicated concentrations of (A) Degussa or (B) Sun Innovations TiO2NPs with illumination and daily water exchanges for 23 days as described in the Materials and Methods, and survival was counted at 23 dpf to produce dose− response curves. In these experiments, normal control survival was approximately 60%, so the data is presented as two lines. The uncorrected data from each group is shown as open squares. The black filled boxes indicate the mortality data from which control mortality has been subtracted. Results are presented as mean ± standard error of the mean. For these experiments, n is equal to an individual group of 16 fish exposed to the condition, so for each point n = 39−54 groups, each beginning with 16 fish.

TiO2NPs with illumination and the controls receiving illumination alone. The fin rays were not developed, the somites at the tail were indistinct, the pronounced angle in the notochord reinforcing the tail had not developed, and the pigmentation was either lacking or poorly organized. The rows of melanocytes along the spinal column showed obvious gaps in the treated fish. During metamorphosis, the oval shaped larval tail is transformed into a highly organized, muscular tail with an array of rays capable of sophisticated water movement. The tails of the treated fish resembled larval tails. Despite the die-off during metamorphosis, control larvae did not show malformations, and the images in Figure 3 are representative of what we observed. At the higher concentrations, most of the fish showed at least one of these malformations and up to half of the fish displayed all of them, as shown. The fish exposed to 1 ppb TiO2NPs and illumination also tended to have a single-lobed swim bladder rather than developing the bilobed swim bladder typical of juveniles and adults. The mature form can be clearly seen in the control (part A of Figure 3). The swim bladders in the TiO2NP-exposed and illuminated fish in parts B and C of Figure 3, while clearly present, had only a single lobe. Craniofacial structures were also a site of malformation: the sizes of both the upper and lower

We observed dose-dependent mortality for both the Degussa and Sun Innovations particles. At the highest dose, the Degussa particles produced more than 90% mortality, whereas the Sun Innovations particles produced approximately 80% mortality. However, we did not observe the sigmoid plot often seen in dose−response experiments. Instead, toxicity steadily increased with concentration without reaching a true plateau. For the Degussa particles, the 1 ng/mL dose produced a degree of mortality that was approximately halfway between control and 100%. The Sun Innovations particles consistently appeared to be slightly less potent. At 1 ppb, both preparations produce significant mortality compared to controls with p < 10−10. Effects on Growth and Development. Exposure to the TiO2NPs and illumination produced a characteristic pattern of stunted growth and malformation. Representative images of surviving fish exposed to illumination and 1 ng/mL TiO2NPs show a reduction in overall size, failure to progress through metamorphosis, and specific defects (Figure 3). The caudal fin showed striking differences between the groups exposed to 4729

dx.doi.org/10.1021/es304514r | Environ. Sci. Technol. 2013, 47, 4726−4733

Environmental Science & Technology

Article

jaw structures were reduced. We also frequently observed degradation of the caudal fin. TiO2NP exposure appeared to delay metamorphosis from the larval form to the juvenile/adult morphology. To better assess this, we developed a system for scoring metamorphosis. This was a conservative scoring system because, if the fish displayed any trait indicative of metamorphosis at 23 dpf, it was scored as positive for metamorphosis. These traits were: presence of fin rays, fin buds, or a bilobed swim bladder. The metamorphosis scores for the dose response experiment are shown in Figure 4. Both the Sun Innovations and Degussa

Figure 4. TiO2NP exposure causes photodependent delays in development. Zebrafish were exposed to Degussa or Sun Innovations TiO2NPs as described in Figure 1, and samples were scored for metamorphosis as described in the Materials and Methods. Any fish that showed one or more signs of metamorphosis was scored as positive. The averages for the treatment groups were normalized to the control values and are presented as a percent of controls. Results are presented as mean ± standard error of the mean. Asterisks indicate significant differences between illuminated TiO2NP treatment groups and control groups (p < 0.01). For these experiments, n is equal to an individual group of fish exposed to the condition in the same beaker, so for each point n = 39−54 groups.

Figure 5. Aggregation of TiO2NPs. TiO2NP samples were prepared by sonication as described in the Materials and Methods. Samples were prepared in water and aliquots were split into samples and incubated with or without illumination, and the NP preparations were examined either immediately after preparation or at 24 h. Number-weighted hydrodynamic diameters were measured in water by DLS. (A) TEM image of Degussa TiO2NPs immediately after sonication. (B) Average hydrodynamic diameters for Degussa TiO2NPs at the indicated concentrations (Open bars, 0 h samples; shaded bars, 24 h without illumination; dark bars, 24 h with illumination). (C) TEM image of Sun Innovations TiO2NPs. (D) Average hydrodynamic diameters for Sun Innovations TiO2NPs (symbols are the same as in B). Results are presented as mean ± standard error of the mean. Ten DLS measurements were made on each sample, and 3 independent samples were run for each point (n = 3).

forms delayed metamorphosis in a dose-dependent fashion. Because a fish with only one of the three juvenile traits is scored as positive, the results underestimate potency. TiO2NP Aggregation and Zebrafish Ti Body Burden. Transmission electron microscopy (TEM) and measurement of individual particle sizes showed that the average primary particle size of Sun Innovations was 12.1 ± 3.7 nm, whereas Degussa was 23.3 ± 9.8 nm (Figure 5). The maintenance of the fish required daily water changes. We therefore also made new stock solutions daily so that the water renewals could be made with fresh dosing suspensions. With this schedule in mind, we investigated particle size and aggregation during the 24 h that the particles were in solution.

In aqueous suspension, we always observed some degree of particle aggregation. DLS measurements showed a trend to increased aggregate size with increased concentration (Figure 5). Full number distributions of the hydrodynamic diameters can be found in Figure S3 of the Supporting Information. The Degussa particles showed a tendency to increase aggregation state over 24 h in suspension, but this modest change was not consistently observed in the Sun Innovations particles. The time points chosen for parts B and D of Figure 5 correspond to 4730

dx.doi.org/10.1021/es304514r | Environ. Sci. Technol. 2013, 47, 4726−4733

Environmental Science & Technology

Article

exposure conditions in which fresh TiO2NP suspensions were prepared each day to replace the previous day’s suspension. Thus, the fish were exposed to TiO2NPs in the aggregated state, and the aggregation did not change dramatically during the time between suspension changes. ICP-OES measurements showed a dose-dependent increase in Ti body burden (Figure 6). Illumination had no apparent

Figure 7. TEM images illustrating distribution of TiO2NPs in zebrafish. Fish were exposed as described in Figure 1, and samples were collected for ultrathin sectioning and TEM. The examples shown are from fish exposed to 10 000 ng/mL TiO2NPs with illumination. Arrows indicate electron-dense Ti-rich nanoparticles. (A) Gut lumen from 23 dpf fish exposed to Degussa particles. (B) Gut lumen from 12 dpf fish exposed to Sun Innovations particles. (C) Gill tissue from 23 dpf fish exposed to Degussa particles. (D) Higher magnification of gill from 23 dpf fish exposed to Degussa particles. (E) Liver from 23 dpf fish exposed to Degussa particles. (F) Muscle from 12 dpf fish exposed to Degussa particles. Figure 6. TiO2NP Uptake. Fish were exposed to TiO2NPs as described for Figure 1 with (open circles) and without (filled circles) illumination. Samples were removed at 23 dpf for Ti content analysis with ICP-OES. Average Ti body burden is shown for (A) Degussa or (B) Sun Innovations TiO2NPs. Results are presented as mean ± standard error of the mean from 3 independent experiments (n = 3). In each experiment, 10 fish were pooled per sample, and 4 aliquots from each sample were submitted for ICP-OES measurement to produce an average value for that sample. The asterisks indicate significant differences between TiO2NP treatment groups and water controls (p < 0.05).

associated with gut microvilli (part B of Figure 7). Another common location was the gill, where particles were associated with the gill surface and within gill cells (part C of Figure 7) as well as within cells of the gill lamellae (part D of Figure 7). In addition, TiO2NPs were frequently found in the liver (part E of Figure 7). Finally, whereas TiO2NPs were distributed to the gastrointestinal tract, gills, and liver, they were also widely distributed throughout the fish to various tissues. As an example, TiO2NP aggregates were found in skeletal muscle tissue of the exposed fish (part F of Figure 7). The examples shown were from fish exposed at the higher (10 μg/mL) concentrations because more particles are evident in a given field; however, the general distribution was unchanged by concentration. Neither uptake, nor location was photodependent, and the distribution pattern did not differ between the formulations. We previously used energy dispersive X-ray analysis to show that the small electron dense spots observed are indeed Ti-based material.12 Subchronic Exposure to TiO2NPs and Illumination Produces Oxidative Stress. A consequence of increased ROS levels in vivo is macromolecular damage.17,18 A sign of this damage is the formation of 8-hydroxy-2′-deoxyguanosine (8OHdG) DNA adducts that interfere with transcription and replication.19 ELISA measurements showed significant increases in 8-OHdG levels compared to controls in TiO2NPexposed zebrafish (Figure 8). The degree of 8-OHdG levels

effect on body burden. It should be noted that the TiO2NP suspension concentrations on the x axis are plotted on a logarithmic scale, and the Ti measured in the 1 ng/mL samples approached the limit of detection of the instrument. These experiments showed TiO2NPs associated with the fish but did not distinguish between Ti in tissues and Ti on the surface. We conclude that Ti found on or in the fish is proportional to the suspension concentration of TiO2 nanoparticles. We used ultrathin tissue sections and TEM to locate TiO2NPs within the fish (Figure 7). We observed TiO2NPs as aggregates in all tissues sampled and at all concentrations tested. However, TiO2NP aggregates were most commonly observed in sections of the intestinal tract associated with gut contents that included spherical food particles eaten by the fish (part A of Figure 7). Micrographs also showed TiO2NPs 4731

dx.doi.org/10.1021/es304514r | Environ. Sci. Technol. 2013, 47, 4726−4733

Environmental Science & Technology

Article

with elevated levels of 8-OHdG, a sign of oxidative damage. As previously demonstrated, we saw no signs of toxicity without illumination. Our hypothesis predicts widespread intracellular damage of a slowly accumulating nature. This is consistent with a struggle to grow and mature. We speculate that the slowed development and inhibition of metamorphosis is due to a diversion of resources toward repair and maintenance in favor of growth. The observed phenotype is reminiscent of the stunting caused by malnutrition. The slowed growth, delayed development, and progressive organ and DNA damage are all consistent with cumulative cellular damage. Nanoparticle Properties and Implications. We selected two readily available preparations of TiO2NPs as models to study phototoxicity in exposures that spanned development and occurred at concentrations that approach those that might be encountered in the environment. This allowed us to examine preparations with different mixtures of crystal forms and different proprietary formulations. It was not our intention to model any specific environment other than a littoral aquatic environment in which nanomaterials might settle and accumulate in sunlit waters. We chose these particles because they are already being produced, readily available, and suitable for testing a biological prediction based on a physical property. Because we predicted long-term accumulative phototoxicity, we wished to determine whether our prediction was correct and if it was correct make this information available. Our results cannot be directly used for risk assessment, but they do suggest areas of research in that area. It is possible to make photoabsorbant TiO2 nanoparticles that do not react with water and oxygen. Our results suggest that this might be a productive area of investigation. During the time in which the embryos were in the eggs, the particle aggregates would tend to settle, undisturbed by movement of the fish protected in the chorion. After hatching, the fish would be on the bottom, potentially exposed to higher local concentrations of settled aggregates. However, when moving they would be above the bottom and they would resuspend sediments. Finally, as the fish begin to feed, uptake with the food and through the gills becomes probable. Therefore, the exposure conditions, even in a polished plastic well, are quite complex. With that said, there is no reason to suppose that these processes of sedimentation and resuspension, with high and low local concentrations, are not found in nature. The prey for small fish would be filter feeders and sediment grazers, so bioconcentration could be a factor as well. These factors are beyond the scope of our study and do not alter our findings. Increased concentration produced increases in aggregation, uptake, and toxicity. Aggregates of particles were found inside fish organs and tissues; indeed, isolated single particles were rarely if ever observed within the tissues (this study and in ref 12). The increased uptake and toxicity at higher concentrations indicate that aggregation does not prevent exposure or toxicity per se. Other conditions changed over the course of the experiment. The early embryos and larvae are almost completely transparent, and, even after pigment spots appear, all of the organs remain clearly visible and transparent to light. As the fish reach the end of larval development, additional pigments appear that shield the internal organs from light and give the fish their characteristic silver appearance. As they grow, the developing fish tend to segregate into size cohorts representing small initial

Figure 8. TiO2NP exposure produces signs of in vivo oxidative stress. Fish were exposed as described in Figure 1, and samples were collected for ELISA of 8-OHdG DNA adducts at 23 dpf. For each assay, 9−12 fish were pooled for DNA extraction. Only results from illuminated samples are shown, the nonilluminated samples were at baseline for all concentrations. The colorimetric signal from untreated controls was set at 100% to produce an arbitrary scale, and the optical density readings were converted to percent of the average control values. Three independent samples of 9−12 pooled fish were collected for each treatment (n = 3). Results are presented as mean ± standard error of the mean. The asterisks indicate significant difference between treatment group and control (p < 0.01).

increased with increased TiO2NP concentrations and, as expected, was photodependent (not shown). The Degussa preparation elicited a larger oxidative stress response than the Sun Innovations particles.



DISCUSSION Our initial experiments showed that TiO2NPs produce toxicity through photodependent production of ROS. This suggested that low TiO2 nanoparticle concentrations might produce cumulative toxicity that would escape detection in short studies. Our findings confirm this: over many repeated experiments, we observed significant increases in mortality in fish exposed to TiO2NPs at only 1 ng/mL. Toxicity and Light. Using a model system, we were careful to note that the results would be irrelevant if we used conditions not encountered in nature. The most important of these was lighting. It is generally not appreciated how much less intense interior lighting is than daylight. Our animal facility has low-level lighting to satisfy rodent accreditation standards, and we found no difference in toxicity or ROS production between samples shielded from all light with foil and those exposed to the low-level animal facility lighting.12 We chose an illumination source that has been developed to duplicate sunlight for aquatic organisms. Two important points cannot be overstressed: First, while bright to the eye, our lighting was not nearly as intense as sunlight on a cloudless day; we did not use an extreme form of illumination. Second, while our source of lighting enters the near UV, as does sunlight, the experimental system was not designed to produce UV illumination; our experiments were not directed at UV illumination per se. Any photoactivation observed in our experiments would also occur with daylight. Illumination by itself produced no adverse effects. Our experiments were repeated many times and consistently showed increased mortality in the groups exposed to TiO2 nanoparticles at 1 ppb (p < 10−10). This toxicity was associated 4732

dx.doi.org/10.1021/es304514r | Environ. Sci. Technol. 2013, 47, 4726−4733

Environmental Science & Technology

Article

changes in growth rate that become amplified as the larger fish feed and grow faster. Although many fish species can delay metamorphosis to suit conditions, it is likely that zebrafish must grow to a certain size by a certain time to survive metamorphosis. We speculate that the brighter light in our illuminated cohorts allowed the control fish to more readily identify and consume food leading to the approximately 10% higher survival rate that we see in the illuminated compared to the ambient light controls. Although the ceiling fluorescent lighting we have in our fish facility is on par with lighting used around the globe, we have begun using brighter illumination to raise fish when we need the highest possible survival to adulthood. Prediction of TiO2NP concentrations in the aquatic environment is difficult. Lighting, flow, temperature, depth, organisms, and natural organic matter make real ecosystems more complex than our conditions. However, it is estimated that consumer products will release enough TiO2NPs into the environment to produce nanogram per milliliter water concentrations in some locations.5,20 In general, zebrafish are adaptable and hardy aquarium fish. Our results, using these hardy zebrafish, and an exposure of less than 1 month in the nanogram per milliliter range, showed significant mortality and toxicity. The environmental implications of this result are worth consideration.



Nanomaterials in the environment: behavior, fate, bioavailability, and effects. Environ. Toxicol. Chem. 2008, 27, 1825−51. (5) Menard, A.; Drobne, D.; Jemec, A. Ecotoxicity of nanosized TiO2. Review of in vivo data. Environ. Pollut. 2011, 159, 677−84. (6) Roco, M. C. Environmentally responsible development of nanotechnology. Environ. Sci. Technol. 2005, 39, 106a−112a. (7) Chatterjee, D.; Mahata, A. Visible light induced photodegradation of organic pollutants on dye adsorbed TiO2 surface. J. Photochem. Photobiol., A 2002, 153, 199−204. (8) Dunford, R.; Salinaro, A.; Cai, L.; Serpone, N.; Horikoshi, S.; Hidaka, H.; Knowland, J. Chemical oxidation and DNA damage catalysed by inorganic sunscreen ingredients. FEBS Lett. 1997, 418, 87−90. (9) Rao, M. V.; Rajeshwar, K.; Paiverneker, V. R.; Dubow, J. Photosynthetic production of H2 and H2O2 on semiconducting oxide grains in aqueous-solutions. J. Phys. Chem. 1980, 84, 1987−1991. (10) Wamer, W. G.; Yin, J. J.; Wei, R. R. Oxidative damage to nucleic acids photosensitized by titanium dioxide. Free Radical Biol. Med. 1997, 23, 851−8. (11) Bickley, R. I.; Gonzalezcarreno, T.; Lees, J. S.; Palmisano, L.; Tilley, R. J. D. A structural investigation of titanium dioxide photocatalysts. J. Solid State Chem. 1991, 92, 178−190. (12) Bar-Ilan, O.; Louis, K. M.; Yang, S. P.; Pedersen, J. A.; Hamers, R. J.; Peterson, R. E.; Heideman, W. Titanium dioxide nanoparticles produce phototoxicity in the developing zebrafish. Nanotoxicology 2012, 6, 670−679. (13) Parichy, D. M.; Elizondo, M. R.; Mills, M. G.; Gordon, T. N.; Engeszer, R. E. Normal table of postembryonic zebrafish development: Staging by externally visible anatomy of the living fish. Dev. Dyn. 2009, 238, 2975−3015. (14) Yang, S.; Bar-Ilan, O.; Peterson, R. E.; Heideman, W.; Hamers, R. J.; Pedersen, J. A. Influence of humic acid on titanium dioxide nanoparticle toxicity to developing zebrafish. Environ. Sci. Technol. 2013, DOI: 10.1021/es3047334. (15) Ma, X.; Benson, C. H.; McKenzie, D.; Aiken, J. M.; Pedersen, J. A. Adsorption of pathogenic prion protein to quartz sand. Environ. Sci. Technol. 2007, 41, 2324−30. (16) Westerfield, M. The Zebrafish Book. A Guide for the Laboratory Use of Zebrafish (Danio rerio), 4th ed.; University of Oregon Press: Eugene, OR, 2000. (17) Digiulio, R. T.; Washburn, P. C.; Wenning, R. J.; Winston, G. W.; Jewell, C. S. Biochemical responses in aquatic animals - a review of determinants of oxidative stress. Environ. Toxicol. Chem. 1989, 8, 1103−1123. (18) Kelly, S. A.; Havrilla, C. M.; Brady, T. C.; Abramo, K. H.; Levin, E. D. Oxidative stress in toxicology: Established mammalian and emerging piscine model systems. Environ. Health Perspect. 1998, 106, 375−384. (19) Petersen, E. J.; Nelson, B. C. Mechanisms and measurements of nanomaterial-induced oxidative damage to DNA. Anal. Bioanal. Chem. 2010, 398, 613−50. (20) Mueller, N. C.; Nowack, B. Exposure modeling of engineered nanoparticles in the environment. Environ. Sci. Technol. 2008, 42, 4447−53.

ASSOCIATED CONTENT

S Supporting Information *

Comparison between the spectral intensity of the metal halide lamp illumination at 45 cm and sunlight, survival of zebrafish exposed to Degussa and Sun Innovations TiO2NPs without illumination, and full number distributions of the hydrodynamic diameters. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*Fax: (608) 265-3316; e-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank Dorothy Nesbit, Myzoom Yang, Bert Jansch, and B. Hunter for technical assistance, Ben August and Amanda Thomas for TEM support (UW Madison Medical School Imaging Center), Jackie Bastyr-Cooper for ICP-OES support, Ralph Albrecht for his sage advice, and the Heideman/Peterson groups for insightful discussion. This work was funded by NSEC (NSF Grant No. DMR-0425880) and EPA Star Program (RD-83386001-0).



REFERENCES

(1) Buzea, C.; Pacheco, I. I.; Robbie, K. Nanomaterials and nanoparticles: Sources and toxicity. Biointerphases 2007, 2, MR17−71. (2) Farre, M.; Gajda-Schrantz, K.; Kantiani, L.; Barcelo, D. Ecotoxicity and analysis of nanomaterials in the aquatic environment. Anal. Bioanal. Chem. 2009, 393, 81−95. (3) Handy, R. D.; Owen, R.; Valsami-Jones, E. The ecotoxicology of nanoparticles and nanomaterials: Current status, knowledge gaps, challenges, and future needs. Ecotoxicology 2008, 17, 315−25. (4) Klaine, S. J.; Alvarez, P. J.; Batley, G. E.; Fernandes, T. F.; Handy, R. D.; Lyon, D. Y.; Mahendra, S.; McLaughlin, M. J.; Lead, J. R. 4733

dx.doi.org/10.1021/es304514r | Environ. Sci. Technol. 2013, 47, 4726−4733