Fullerene C₆₀ exposure elicits an oxidative stress response in embryonic zebrafish

Abstract

Due to its unique physicochemical and optical properties, C₆₀ has raised interest in commercialization for a variety of products. While several reports have determined this nanomaterial to act as a powerful antioxidant, many other studies have demonstrated a strong oxidative potential through photoactivation. To directly address the oxidative potential of C₆₀, the effects of light and chemical supplementation and depletion of glutathione (GSH) on C₆₀-induced toxicity were evaluated. Embryonic zebrafish were used as a model organism to examine the potential of C₆₀ to elicit oxidative stress responses. Reduced light during C₆₀ exposure significantly decreased mortality and the incidence of fin malformations and pericardial edema at 200 and 300 ppb C₆₀. Embryos co-exposed to the glutathione precursor, N-acetylcysteine (NAC), also showed reduced mortality and pericardial edema; however, fin malformations were not reduced. Conversely, co-exposure to the GSH synthesis inhibitors, butathionine sulfoximine (BSO) and diethyl maleate (DEM), increased the sensitivity of zebrafish to C₆₀ exposure. Co-exposure of C₆₀ or its hydroxylated derivative, C₆₀(OH)₂₄, with H₂O₂ resulted in increased mortality along the concentration gradient of H₂O₂ for both materials. Microarrays were used to examine the effects of C₆₀ on the global gene expression at two time points, 36 and 48 hours post fertilization (hpf). At both life stages there were alterations in the expression of several key stress response genes including glutathione-S-transferase, glutamate cysteine ligase, ferritin, α-tocopherol transport protein and heat shock protein 70. These results support the hypothesis that C₆₀ induces oxidative stress in this model system.

Introduction

Carbon fullerenes, otherwise known as Buckminster fullerenes, or “bucky balls”, have unique physicochemical properties that may be exploited for use in consumer products such as cosmetics, lubricants, food supplements, building materials, clothing treatment, electronics and fuel cells (Loutfy et al., 2002). While fullerenes are currently produced by the ton every year, actual commercialization is still mostly under development (Robichaud et al., 2005). Due to the potential for widespread applications and consequently, widespread exposures, evaluation of the biological effects of these important nanomaterials is warranted. C₆₀ has been shown to induce toxicity in numerous cell culture and whole animal systems (Oberdorster, 2004; Sayes et al., 2005; Isakovic et al., 2006; Usenko et al., 2007). However, contradictory reports in the literature make it difficult to interpret the mechanism by which C₆₀ toxicity is induced.

Fullerene C₆₀ has been described as a free radical scavenger by some (Dugan et al., 1996; Wang et al., 1999; Mori et al., 2006) yet others have reported it to generate oxygen free radicals within and outside of biological systems (Yamakoshi et al., 2003; Isakovic et al., 2006). With few exceptions, pristine C₆₀ and suspensions of C₆₀ clusters (nC₆₀) are reportedly more toxic than their derivatives (Isakovic et al., 2006). In cell culture, nC₆₀ synergized the effects of oxidative stress-inducing agents and elicited cytotoxic action through cell membrane lipid peroxidation (Sayes et al., 2005; Isakovic et al., 2006). In vivo, nC₆₀ induced oxidative stress and lipid peroxidation in the brain of juvenile largemouth bass (Oberdorster, 2004). However, others have reported C₆₀ to act as a powerful antioxidant in vivo in rats with no acute or subacute toxicities (Gharbi et al., 2005). Certain surface modifications (e.g. hydroxylation) have been shown to impart cytoprotective activity by eliminating reactive oxygen species (ROS) and antagonizing the effects of the oxidative stress-dependent cytotoxicity (Dugan et al., 2001; Bogdanovic et al., 2004; Isakovic et al., 2006). In fact, carboxylated-C₆₀ has been patented as a method to increase a metazoan’s lifespan presumably through these same mechanisms (Dugan et al., 2003). Despite disagreement amongst researchers as to the oxidative potential of C₆₀, oxidative stress is currently the leading proposed mechanism through which fullerenes may induce toxicity.

Oxidative stress is a common pathway of toxicity and disease. An organism can undergo oxidative stress through several different mechanisms. First, it may be directly induced by an oxidizing agent, such as H₂O₂. Second, it may be produced through the induction of cytochrome P450, as is the case with many polycyclic aromatic hydrocarbons (PAHs) (Wassenberg and Di Giulio, 2004). Third, a xenobiotic may inhibit the production of antioxidant molecules, such as glutathione (GSH), that function to maintain oxidative balance (Anderson, 1998). GSH is an endogenous tripeptide enzyme and known free radical scavenger, thus, it is important for detoxification of metabolites and ROS associated with chemical exposure and disease (Asikainen et al., 2002; Franklin et al., 2002; Suh et al., 2004). It has yet to be determined through which, if any, of these pathways fullerenes may act.

The physicochemical properties of C₆₀ support the hypothesis that this nanomaterial may induce oxidative stress following photoactivation. Due to its unique spherical structure, C₆₀ has the ability to accept up to 6 electrons (Jensen et al., 1996). These electrons essentially race around the structure through dipole moments. When C₆₀ is acted upon by light, it is raised to a higher energy level producing singlet C₆₀, which reacts with O₂ to form singlet oxygen (¹O₂) (Hirsch and Brettreich, 2005). The amount of UV absorption necessary to raise the molecule to the triplet state varies with surface functionalization (Prat et al., 1999). In general, more functional groups added to the fullerene requires more energy to move to the excited state, resulting in lower triplet and singlet oxygen quantum yields (Prat et al., 1999). In the presence of C₆₀, both visible and ultraviolet light can generate ROS, particularly as singlet oxygen and superoxide (Pickering and Wiesner, 2005). These byproducts can induce oxidative stress leading to a variety of detrimental downstream effects such as lipid peroxidation, DNA and protein adduction and cellular death (Kamat et al., 2000; Dugan et al., 2001; Dhawan et al., 2006).

In these studies, we exploit the numerous advantages of the embryonic zebrafish model to elucidate the mechanism by which C₆₀ induces its toxic actions. This model has been used extensively for drug discovery and chemical or nanomaterial screening to rapidly evaluate integrated system effects and subsequently identify mechanisms of toxic action (Usenko et al., 2007). Here, we tested the effect of light activation and antioxidant environment on C₆₀ toxic potential, and used zebrafish microarrays to evaluate global gene expression following C₆₀ exposure. Further, we evaluated both C₆₀ and hydroxylated C₆₀ for antioxidative protection from H₂O₂ exposure. The results presented herein strongly indicate oxidative stress as a pathway of C₆₀-induced toxicity in this experimental model.

Methods

Solution Preparation

C₆₀ was obtained from Sigma Aldrich, WI (99+%) and dissolved in 100% dimethyl sulfoxide (DMSO). The stock solution of 50 ppm C₆₀ was sonicated for 1.5 hours to ensure a uniform dispersion and size distribution. C₆₀(OH)₂₄ (MER Corp, Arizona) was dissolved in DMSO and sonicated for 5 minutes to ensure uniform distribution. The size range of C₆₀ and C₆₀(OH)₂₄ agglomerates following sonication were measured using photon correlation spectroscopy and previously reported in Usenko et al., 2007. H₂O₂ was purchased from VWR International (Brisbane, CA) and was diluted to make exposure solutions of 0.5 mM and 1.0 mM. Buthionine sulfoximide (BSO), diethyl maleate (DEM), N-acetylcysteine (NAC), and L-glutathione (GSH) were purchased from Sigma Aldrich (St. Louis, MO).

General Exposure Protocol

Embryonic zebrafish were obtained from the AB strain of zebrafish (Danio rerio) reared in the Sinnhuber Aquatic Research Laboratory (SARL) at Oregon State University. Adults were kept at standard laboratory conditions of 28 °C on a 14 h light/10 h dark photoperiod with a conductivity of 500 Siemens per meter. Water consisted of reverse osmosis (RO) water supplemented with a commercially available salt solution (0.6% Instant Ocean®). Zebrafish were spawned and embryos were collected and staged (Reimers et al., 2006). The chorion, an acellular envelope surrounding the embryo, was removed via enzymatic reaction with pronase at 24 hours post fertilization (hpf) as described previously (Usenko et al., 2007). At 24 hpf, embryos were placed individually into wells of a 96-well plate, so that there was one embryo per well, and exposed to 100 μl of nanomaterial solution.. A 1% DMSO in water solution was used as a control as this concentration was previously found to have no effect on the embryos. Embryos were evaluated every 24 hours for 5 days for morphological malformations, mortality and behavioral abnormalities.

Dark Exposure

To determine if light influenced the toxic potential of C₆₀, the stock solution was prepared and kept in the dark and the 96-well plates were protected from the light after exposure. Embryos were dechorionated at 24 hpf and placed individually in wells and exposed to 1% DMSO controls, 100, 200, 300, or 500 ppb C₆₀ (equivalent to 0, 0.05, 0.1, 0.15, 0.25 ng/mg as reported in Isaacson et al., 2007). Embryos were analyzed daily until 120 hpf for physiological malformations and mortality. The cumulative results were reported for effects. Sublethal effects were scored against number of surviving animals at the initial assessment (24 hours of exposure).

Co-incubation with GSH level-modifying Chemicals

The concentration-response of each of the test chemicals was conducted to determine the levels at which there was an adverse effect. Embryos were dechorionated at 24 hpf and were scored daily until 120 hpf. Embryos were co-incubated with either 100 μM GSH, 50 μM NAC, 250 μM ascorbic acid, 5 μM BSO, or 50 nM DEM. Previous reports found the maximum tolerable concentration (MTC) ascorbic acid to be 250 μM and GSH to be 100 μM (Hirsch and Brettreich, 2005). A concentration gradient was used for other co-incubations in this study. NAC is the rate-limiting substrate in GSH synthesis, and thus was used to potentially increase GSH production. BSO and DEM were used to inhibit GSH synthesis, thus making the embryos more sensitive to oxidative stressors. All chemicals were co-incubated with 50–300 ppb C₆₀ or 1% DMSO and evaluated as described above. The pH of exposure solutions was monitored and buffered appropriately. H₂O₂ is a known oxidant and was co-incubated with 10 ppb C₆₀ or 100 ppb C₆₀(OH)₂₄ to determine if either fullerene shifted the concentration-response curve to the right or left. H₂O₂ was co-incubated at 0.5 mM and 1.0 mM based on initial concentration-response of H₂O₂. At 1.0 mM, H₂O₂ induced 100% mortality.

Cellular Death Assays

Cellular death was evaluated in embryos co-exposed to C₆₀ and NAC, DEM, BSO or ascorbic acid. This endpoint serves as an early indicator of perturbation based on previous evaluations that showed significant cell death could be detected after only 12 hours of exposure to C₆₀ (Reimers et al., 2006). Since cellular death was evaluated in whole embryos, potential targets of toxi city could be identified. Embryos were exposed to 100 ppb C₆₀ or 1% DMSO and were co-exposed to 50 μM NAC, 50 nM DEM, 5 μM BSO, or 250μM ascorbic acid at 24 hpf. At 36 hpf, embryos were rinsed with water and then incubated in 100 μl of 5 μg/ml acridine orange for one hour in the dark at 28 °C. Embryos were rinsed again, mounted in low melt agarose (1% w/v, Promega, Madison, WI) and imaged using an Axiovert 200M Zeiss microscope (Carl Zeiss International, Germany) with a 546 nm filter. Fluorescence in the head region was measured and quantified using ImagePro Plus software (Media Cybernetics, Inc., Silver Spring, MD).

Global Gene Expression Analysis

Custom Array Chip

Oligonucleotides (50 mer) were purchased from MWG (High Point, NC). Zebrafish custom arrays of 14,000 genes were spotted by Eric Johnson at University of Oregon. Epoxy-coated slides were cross-linked and were kept at room temperature in a dessicator until use.

Isolation of RNA

Embryos were exposed to 200 ppb C₆₀ or 1% DMSO at 24 hpf until 36 or 48 hpf and then pooled into three groups (biological replicates) of 40 embryos. Embryos were euthanized with tricaine methanesulfonate and rinsed thoroughly. Excess water was removed from the samples and 100 μl of TRIzol Reagent was added to extract RNA. Embryos were homogenized using a pestle on ice and stored at −80 °C until processing. Once all samples were collected, homogenates were thawed and 900 μl of TriReagent was added and each sample was vortexed. Next, the homogenates were centrifuged at 12,000 g at 4 °C, and then 200 μl chloroform was added. The samples were centrifuged again and the supernatants aliquoted into new vials. 500 μl of isopropanol was added to each sample and the sample was centrifuged. All liquid was removed and the pellet was washed several times with 70% ethanol: H₂O. The pellets were air dried and then resuspended in 30 μl diethylpyrocarbonate-water. The extractions were flash frozen using liquid nitrogen and stored at −80 °C. RNA was quantified using NanoDrop sensor and the quality verified using Agilent’s BioAnalyzer 2100 (Palo Alto, CA) at the Center for Genome Research and Biocomputing (CGRB) at OSU.

Processing

Labeling, hybridization and scanning were performed at the CGRB. A Genisphere 950 labeling kit (Hatfield, PA) was the hybridization label for the samples. A dye swap for each sample was done to eliminate variability due to dye affinities. One microgram of RNA of each sample was reverse-transcribed with Superscript II (Invitrogen) using the Genisphere olido d(T) primer containing a capture sequence for the Cy3 or Cy5 labeling reagents. Slides were scanned using a GenePix Scanner at 543 nm for Cy3 and 633 for Cy5 at 80% power. Data was compiled using Gene Pix Software.

Data Analysis

Values from corresponding dye swaps were averaged between individual samples to obtain a single value for each sample. Treatments were conducted in triplicate for significance validation. Image files were quantified using QuantArray (PerkinElmer) and the raw data was imported into BASE software. Raw mean backgrounds were subtracted using LOWESS in order to eliminate dye-related artifacts. Genes induced or repressed by greater than 2-fold in all three replicates were selected and placed into categories based on function.

qPCR

Oxidative stress response genes were validated using quantitative real time polymerase chain reaction (qRT-PCR). cDNA was made from the RNA isolated for the microarrays. Briefly, cDNA was synthesized from 1 μg of total RNA per group using Superscript II (Life Technologies, Gaithersburg, MD) and oligo(dT) primers in a 20 ml volume. Primers were designed and ordered from MWG (High Point, NC) for several targeted genes (sequences given in table 1). Quantitative PCR was conducted using gene specific primers with the Opticon-2 real-time PCR detection system (MJ Research, Waltham, MA). For each PCR reaction, 1 μl of cDNA was used in the presence of SYBR Green, using DyNAmo SYBR green qPCR kit according to the manufacturer’s instructions (Finnzymes, Espoo, Finland). A temperature gradient (54 °C–58 °C) was used to determine the appropriate annealing temperature for each primer set. All primers had an annealing temperature of 58 °C and ran for 35 cycles. The PCR product was separated using gel electrophoresis to ensure only one product was made. The PCR product was mixed with 6x SDS dye and loaded onto a 1.5% agarose gel with 0.05% ethidium bromide. The gel electrophoresis was run at 110 V and 80 mA for 1 hour. The gel was viewed on an Ultraviolet Transilluminator and imaged with a Polaroid photo documentation camera. PCR data was analyzed against an adult cDNA calibrator using the GAPDH primer set by the Opticon Monitor 3.1 software (MJ Research, Waltham, MA). Each sample PCR product was analyzed against a glyceraldehydes-3-phosphate dehydrogenase (GAPDH) control for that sample.

Table 1.

Primer sequences used in quantitative PCR. All products are between 180 and 300 mer.

Gene	Forward Primer	Reverse Primer
Ferritin heavy chain	5′-AGACACACTACTTGGACGAG	5′-AACAAGCTAGGAGGTTCTGC
HSP70	5′-GACCAAAGACAACAACCTGC	5′-ATGTTGAAGGCGTAAGACTCC
GCLc	5′-CTATCTGGAGAACATGGAGG	5′-CATTTTCCTCTGTTGACCGG
GAPDH	5′-GAATTCTGGGATACACGGAG	5′-AAAGGGGTCACATCTACTCC
GST-pi	5′-TTCAGTCCAACGCCATGC	5′-ATGAGATCTGATCGCCAACC
Tocopherol transport protein	5′-GTGTTTTGCTCATGCTCTGC	5′-ACTTCATCTACGCTGGGTCC

Statistical Analysis

All statistics were compiled using SigmaStat and plotted using Sigma Plot (SPSS Inc, Chicago, IL). Two-way analysis of variance (ANOVA) was used to detect significant differences between control and treated groups at a p < 0.05. All exposure studies had an N = 24 with 80% confidence interval. Significance was determined for cellular death assays using two-way ANOVA (p < 0.05), N = 12.

Results

Light activation of C₆₀

Oxidative stress was probed at the physical-chemical level since ROS generation by C₆₀ has been shown to be photoactivated (Wang et al., 1999). Although exposure in the dark was not completely protective, it did reduce the physiologic response of C₆₀ exposure. There was a significant reduction in fin malformations, pericardial edema and mortality in the 200 and 300 ppb exposure groups when exposed in the dark (figure 1). Mortality was reduced by 30% at 200 and 300 ppb; however, 500 ppb C₆₀ still induced 100% mortality within the first 24 hours of exposure (figure 1 A). Fin malformations were reduced by approximately 40% and pericardial edema was reduced by 85% at 200 ppb (figure 1 B, C). The presence or absence of light did not have an effect on development of the embryos.

Figure 1. C₆₀ concentration-response: dark vs. light exposure.

24 hpf embryos were dechorionated and exposed in the dark to graded concentrations of C₆₀. Mortality (A), fin malformations (B), and pericardial edema (C) were scored daily for 5 days post exposure. Significant difference was determined using two-way ANOVA (*p < 0.05), N = 24, compared to C₆₀ effect in the light at that concentration. Error bars represent the standard error of means (SEM).

Chemical supplementation of GSH and ascorbic acid

ROS-induced toxicity should be reduced by an increase in intracellular antioxidant levels. To determine if oxidative stress is a potential mechanism of C₆₀-induced responses, chemical alterations in GSH levels were evaluated. Co-incubation of 100 μM GSH with a concentration gradient of C₆₀ did not alter the responses elicited by C₆₀ alone (data not shown). Due to the low membrane permeability of GSH, embryos were co-exposed to NAC rather than GSH to determine if favoring GSH synthesis would offer protection from C₆₀-induced responses. NAC can induce toxicity at high concentrations >300 μM (data not shown) so a low concentration (50 μM) was co-incubated with a concentration gradient of C₆₀ (50–300 ppb C₆₀) (figure 2 A–C). NAC co-exposure with C₆₀ reduced mortality by approximately 35% at 200 and 300 ppb and pericardial edema by 50% at 200 ppb C₆₀ (figure 2 A,C). However, at concentrations of 300 ppb, the incidence of pericardial edema was significantly increased, most likely due to improved survival. NAC co-exposure to 1% DMSO did not have an effect compared to the 1% DMSO controls. Interestingly, co-exposure with NAC did not reduce the incidence of fin malformations at any concentration (figure 2 B).

Figure 2. NAC co-incubation with C₆₀.

Embryonic zebrafish were co-incubated with 50 μM NAC and graded concentrations of C₆₀ from 24 to 120 hpf. (A) Cumulative mortality by 120 hpf for C₆₀ + NAC (●) and C₆₀ only (▽). (B) Percentage of embryos with fin malformation including those that died prior to 120 hpf. (C) Percentage of embryos with pericardial edema including those later scored for mortality. Significance was determined using two-way ANOVA, (*p < 0.05, N = 24), and error bars represent the +/− the SEM.

An alternative approach to block ROS-induced toxicity is to increase intracellular antioxidant levels. Ascorbic acid (vitamin C) has been shown to protect zebrafish from ethanol-induced toxicity at 250 μM (Reimers et al., 2006). Embryos were co-exposed to 250 μM ascorbic acid and 50, 100, 200, and 300 ppb C₆₀. At 200 ppb C₆₀, there was a significant reduction in mortality; however at 300 ppb, 100% mortality was observed (figure 3 A). There was not a statistically significant difference between ascorbic acid controls (with 1% DMSO) and 1% DMSO controls. Unlike NAC, ascorbic acid reduced incidence of fin malformations at 200 ppb C₆₀ (figure 3 B). At 200 ppb C₆₀, there was nearly complete protection from pericardial edema compared to background levels (figure 3 C).

Figure 3. C₆₀ co-incubation with ascorbic acid.

Embryos were co-exposed to 250 μM ascorbic acid (AA) and graded concentrations of C₆₀ at 24 hpf until 120 hpf. Ascorbic acid decreased (A) mortality, (B) fin malformations (FM), and (C) pericardial edema (PE) at 200 ppb C₆₀. Significance was determined by two-way ANOVA (*p < 0.05, N = 24), and error bars represent +/− the SEM.

Chemical depletion of GSH

BSO and DEM, known inhibitors of glutathione synthesis, were used to determine if C₆₀ exerts a toxic response through depletion of glutathione. DEM was found to have a maximum tolerable concentration (MTC) of 50 nM and BSO had a MTC of 5 μM. These concentrations of DEM and BSO were used for each co-incubation with graded concentrations of C₆₀ (50–300 ppb) and 1% DMSO controls. Both DEM and BSO increased the embryo’s sensitivity to C₆₀, with increased mortality within the first 24 hours of exposure (figure 4). Embryos were more sensitive to DEM than BSO. There was not a significant increase in mortality between controls for either DEM or BSO and control embryos.

Figure 4. DEM and BSO concentration response.

DEM and BSO were used to block glutathione production. Embryos were co-exposed to 5 μM BSO or 50 nM DEM and graded concentrations of C₆₀ at 24 hpf for 24 hours. DEM shifted the LC₅₀ to 50 ppb rather than 200 ppb. BSO induced 100% mortality at 200 ppb, but no significant mortality at 100 ppb. Significance was determined using two-way ANOVA (*p < 0.05, N = 24). Error bars represent +/− the SEM.

To more closely examine the C₆₀-elicited physiologic response following exposure to NAC or DEM, cellular death was evaluated in whole embryos. Embryos were co-incubated with 100 ppb C₆₀ and 50 nM DEM, 5 μM BSO, 250 μM ascorbic acid or 100 μM NAC. Control embryos were incubated with 1% DMSO, and chemical controls were also co-incubated with 1% DMSO. DEM, BSO, and NAC did not induce cellular death when co-incubated with 1% DMSO. However, DEM and BSO significantly increased cell death when co-exposed to C₆₀ compared to C₆₀ exposed embryos (figure 5). NAC exposure offered only partially protected the embryo from C₆₀-induced cellular death. Ascorbic acid had no effect on cell death despite its protection from physiologic responses (data not shown).

Figure 5. Cell death in co-exposures.

Cell death was measured as a result of co-incubation with either NAC or DEM and 100 ppb C₆₀. Embryos were exposed at 24 hpf and cellular death was determined at 36 hpf using acridine orange. There was not a statistically significant difference between (A) control and (C) DEM, (E) BSO, or (G) NAC only. (D,E) DEM and BSO co-incubated with C₆₀ significantly increased cell death compared to (B) C₆₀ only. (H) NAC decreased cell death compared to C₆₀ alone; however cell death was significantly higher than controls. Significance was determined using two-way ANOVA (*p < 0.05, N = 12).

Antioxidative Potential of Fullerenes

A study by Wang et al. (1999) found C₆₀ to be a powerful antioxidant, even more powerful than vitamin E. So far, high levels of C₆₀ have been investigated for oxidative potential; however, antioxidant properties may be observed only at very low concentrations. Low concentrations of C₆₀ and C₆₀(OH)₂₄ (10 ppb and 500 ppb, respectively) were co-incubated with two concentrations of H₂O₂ (500 μM and 1 mM) to determine if C₆₀ offered protection from H₂O₂-induced oxidative stress and mortality. These fullerene concentrations were previously reported to be 10-fold below the no observable adverse effect level (NOAEL) (Kamat et al., 1998; Kamat et al., 2000). Neither co-exposure offered embryonic protection, but instead significantly increased mortality compared to embryos exposed only to H₂O₂ (figure 6).

Figure 6. H₂O₂ and C₆₀ co-incubation.

Embryos were exposed to H₂O₂ at 24 hpf for 48 hours to determine the concentration-response. Embryos were co-exposed to H₂O₂ and either 10 ppb C₆₀ or 500 ppb C₆₀(OH)₂₄ for 48 hours. Both significantly increased mortality at 0.5 mM H₂O₂. Significance was determined using two-way ANOVA (*p < 0.05, N = 24). Error bars represent +/− the SEM.

Global Gene Expression

Custom zebrafish microarrays were used to identify the early genomic responses to C₆₀ exposure. The microarray data was analyzed to determine the transcripts that were mis-regulated by more than 2 fold following exposure. Of the 234 and 95 genes that were up-regulated by C₆₀ exposure at 36 hpf and 48 hpf, respectively, 66 genes were up-regulated at both time points. Only 15 genes were commonly down-regulated between the two time points. Genes were grouped by function and time of exposure (figure 7 A–D). Noteworthy, 24% and 19% of genes up-regulated at 36 and 48 hpf, respectively, were associated with a response to stress. These genes include GST-pi, GCLc, heat shock proteins and ferritin. Furthermore, there was up-regulation of development, cell cycle (including induction of apoptosis), signal transduction, and cytoskeleton and cytoplasmic transfer genes (figure 7). Examples of regulated genes from each category are given in table 2. See supporting information for complete list of genes. It is important to point out that for several of the genes the MWG oligo set contained multiple oligos for each target gene covering different regions of the transcript, and we found very little variation in intensity across array elements which further increased the confidence of the raw array data.

Figure 7. Pie chart of categories of gene regulation alterations.

Differentially regulated genes were put into broad categories by stage and regulation: (A) 36 hpf up, (B) 36 hpf down, (C) 48 hpf up, (D) 48 hpf down. Stress response genes were only found to be up-regulated at both stages.

Table 2.

Examples of misregulated genes. Several genes that were significantly up- or down-regulated at 36 hpf or 48 hpf compared to 1% DMSO controls were selected.

Reporter ID	Gene Name	36 hpf Average	36Standard Deviation	48 hpf Average	48 hpf Standard Deviation
Cell cycle/chromatin/RNA/DNA processing
mwgzebrafish#04109	Biglycan-like protein 3	11.585	3.942	9.412	0.904
mwgzebrafish#05567	DCMP deaminase	5.986	1.124	3.403	1.000
mwgzebrafish#04079	Mob1/phocein	8.900	0.726	7.296	2.019
mwgzebrafish#04110	Wings apart	5.195	1.279	4.688	2.163
Cytoskeleton/Cytosolic Transport
mwgzebrafish#05549	E-cadherin binding protein E7	10.391	3.233	5.756	1.575
mwgzebrafish#02873	Keratin 18	3.726	0.862	3.134	0.539
mwgzebrafish#06894	Beta-actin			0.280	0.023
Immune Function
mwgzebrafish#04096	CTAGE family, member 5	4.665	0.864	3.875	0.448
mwgzebrafish#04064	Tec-family kinase	4.956	0.705	3.487	1.024
Metabolism
mwgzebrafish#04095	Angiotensin converting enzyme	7.947	0.739	6.020	0.822
mwgzebrafish#00919	Cbr1l protein	8.035	1.971	5.063	0.540
mwgzebrafish#11641	Inosine triphosphate pyrophosphatase	5.625	3.587	2.715	0.156
mwgzebrafish#13913	Transketolase	3.112	0.280	3.040	0.350
Protein Processing/Degradation
mwgzebrafish#12741	Cathepsin B	2.627	0.208	3.031	0.294
mwgzebrafish#04081	Esrom	8.228	0.201	7.710	2.500
mwgzebrafish#08710	Ubiquitin C	3.708	1.163	2.859	0.759
Signal Transduction
mwgzebrafish#07570	CaM kinase	18.887	4.605	6.425	1.833
mwgzebrafish#01242	Neurogenin 1	10.825	4.724	5.662	0.877
mwgzebrafish#05534	RNA-binding region RNP-1	8.312	0.999	5.003	0.656
Stress Response
mwgzebrafish#01262	Appb protein	6.920	0.908	5.730	0.667
mwgzebrafish#11036	Ferritin	15.994	1.393	7.482	1.976
mwgzebrafish#05998	Ferritin subunit 1	18.070	1.657	9.912	1.624
mwgzebrafish#00125	Glutathione S-transferase pi	11.527	1.276	9.723	0.781
mwgzebrafish#13567	Major vault protein	2.865	0.421	2.977	0.711
mwgzebrafish#08871	Sulfotransferase	4.616	0.953	3.931	1.624
mwgzebrafish#03954	Thioredoxin peroxidase	13.581	1.584	5.747	0.680
mwgzebrafish#00078	Hsp70	6.991	1.046	2.984	0.515
mwgzebrafish#07774	Glutamate-cysteine ligase, catalytic subunit alpha-tocopherol	4.926	0.677
mwgzebrafish#05558	transfer protein	4.173	1.006
mwgzebrafish#00120	Ferritin heavy chain	2.716	0.112	2.750	0.359

PCR

Five target genes that were found to be differentially expressed by microarray analysis were validated using qPCR: GST-pi, GCLc, ferritin heavy chain, α-tocopherol transport protein (TTP), and heat shock protein 70kD (HSP70). All genes chosen were significantly up-regulated over 2-fold. GAPDH was used as a loading control and adult zebrafish cDNA was used as a calibrator. These stress response genes were significantly up-regulated (figure 8). The abundance of GCLc was very low, resulting in a large fold induction. GST-pi had strong results with similar fold induction at both 36 and 48 hpf (figure 8 A). The results from PCR validated the directional regulation observed in the microarray analyses.

Figure 8. qRT-PCR results of gene expression.

Genes involved in oxidative stress responses that were identified as mis-regulated by the microarray were further validated using qRT-PCR. Significance was determined using one-way ANOVA between controls and 200 ppb C₆₀ treated (*p < 0.05, N = 3). Error bars represent the SEM.

Discussion

In vivo evaluations conducted using embryonic zebrafish provide strong evidence indicating C₆₀ is a powerful oxidant in the absence of functionalization. Four lines of evidence presented herein point to oxidative stress as a primary pathway of C₆₀-induced toxicity. First, physiologic responses and cellular death were inhibited by the absence of light during solution preparation and throughout the exposure period. A reduction in light resulted in reduced mortality and malformations, except at the highest concentration. While photoactivation strongly indicates oxidative stress as a mechanism of action, the lack of complete protection when exposed in the dark indicates this may not be the only mechanism of action. Lee et al. (2007) recently found that agglomeration size of C₆₀ affects the degree to which C₆₀ is able to mediate energy and electron transfer (Lee et al., 2007). The large agglomeration sizes previously measured in our system indicate there could be a decrease in photoirradiation potential of C₆₀ at higher concentrations (Kamat et al., 2000; Yamakoshi et al., 2003). Kamat and others demonstrated that when photoexcited, both C₆₀ and hydroxylated C₆₀ can induce membrane damage in rat hepatic and tumor microsomes by generation of ROS in a time- and concentration-dependent manner (Prat et al., 1999). Although these photoactivation characteristics may be undesirable for healthy cells, this activity could be exploited for applications in biomedical science (Zhu et al., 2007). For example, targeted drug delivery to tumors may be achieved via encapsulation of chemotherapeutics in C₆₀ particles and subsequent illumination localized at the target site to activate C₆₀ and trigger release of the drug. Additionally, surface groups could be used to alter the photochemical properties of fullerenes since they are apparently less responsive to light with increased functionalization (Kamat et al., 1998; Sayes et al., 2005).

A second line of evidence is provided by the change in sensitivity of embryonic zebrafish to C₆₀ exposure that resulted from altered levels of the antioxidants GSH and ascorbic acid. Although administration of GSH itself did not have an effect on embryonic susceptibility, uptake of GSH into cells is thought to be limited requiring ATP for transport (Wang et al., 1999). However, another study using embryonic zebrafish revealed that GSH co-exposure with tetrahydrofuran-C₆₀ (THF-C₆₀) increased survival when compared to THF-C₆₀ alone (Bogdanovic et al., 2004). In this case, it may be that solubilization of C₆₀ with THF changed its physicochemical properties such that GSH interacted directly with THF-C₆₀ in the exposure solution and reduced C₆₀ uptake. Internal levels of GSH would need to be measured before and after waterborne GSH exposure to determine its uptake into biological systems.

Uptake limitations and uncertainties of GSH were overcome by using chemical that could deplete internal antioxidant pools. Depletion of antioxidants is one pathway through which nanomaterials could induce oxidative toxicity. When DEM and BSO were used to chemically inhibit glutathione production in embryonic zebrafish, the concentration-response curve shifted to the left; that is, a lower concentration of C₆₀ was required to induce the same effects observed at higher concentrations. This not only implicates oxidative stress as a mechanism of C₆₀ toxicity but also the highlights the importance of GSH in mediating the response. While DEM and BSO potentiated the effects of C₆₀, the addition of NAC offered protection from C₆₀-induced toxicity. NAC had been used in previous studies to increase GSH levels and protect organisms from oxidative stress (Kamat et al., 1998; Sayes et al., 2005). Similarly, our results demonstrated a significant reduction in mortality and pericardial edema in NAC co-incubated embryos. Fin malformations, however, were still present in all embryos exposed to concentrations of 200 ppb and 300 ppb. Co-exposure of embryos to C₆₀ and ascorbic acid resulted in a significant reduction in mortality, pericardial edema and fin malformations, though at 300 ppb C₆₀, vitamin C could no longer provide protection and 100% of the embryos died. Our results concur with previous studies that found ascorbic acid to protect against C₆₀-induced cell death and lipid peroxidation. Collectively, these results suggest C₆₀ has oxidative potential since depletion of antioxidants increased the sensitivity of zebrafish embryos to C₆₀ exposure while antioxidant enhancement decreased their sensitivity.

The lack of antioxidative function observed from both C₆₀ and C₆₀(OH)₂₄ when co-exposed with H₂O₂ provides the third line of evidence supporting an oxidative stress echanism of toxic action. Cell culture evaluations suggest that C₆₀ acts as a stronger antioxidant than vitamin E, protecting against 1 mM H₂O₂. Derivatives of C₆₀ have also been shown to exhibit antioxidant properties in numerous cell culture and whole animal studies (Sayes et al., 2004; Kamat et al., 1998; Yamakoshi et al., 2003; Sayes et al., 2005). Despite previous reports of antioxidant capabilities, neither C₆₀ nor C₆₀(OH)₂₄ demonstrated antioxidant properties in the embryonic zebrafish model. In fact, rather than offering protection from H₂O₂ exposure, both C₆₀ and its hydroxylated derivative significantly increased the deleterious oxidative effects. Delineation of those that possess antioxidant potential and those that do not should be addressed systematically since ‘fullerenes’ are already being marketed to consumers as antioxidants in cosmetics and night creams. Additionally, identification of fullerenes with antioxidant potential could be developed as therapeutic applications for central nervous system neurodegenerative diseases (i.e. Parkinson’s disease, Alzheimer’s disease, multiple sclerosis, amyotrophic lateral sclerosis, or Huntington’s disease), stroke, atherosclerosis, myocardial ischemia, myocardial reperfusion, diabetes, complications of diabetes, circulatory impairment, retinopathy, blindness, kidney disease, pancreas disease, neuropathy, gum disease, cataracts, skin disease, skin damage, radiation damage, damage caused by tobacco use, excessive angiogenesis, insufficient angiogenesis, hearing loss, collateral damage of chemotherapy, mucositis, senescence, hemorrhagic shock, distributive shock, septic shock, heat stroke, severe burn shock, or non-hemorrhagic trauma shock (Hartnagel et al., 2006).

A final line of evidence indicates an oxidative stress mechanism of C₆₀ toxicity. A zebrafish microarray developed from commercially available oligonucleotides was used to determine important genes and pathways that play a role in the response to C₆₀ exposure. By analyzing two early time points during development, we were able to identify impacts that persist over multiple stages of development. Many of the genes known to be involved in an oxidative stress response were upregulated in embryonic zebrafish after exposure to C₆₀. In particular, two genes that were significantly up-regulated are directly related to glutathione activity: GST-pi and GCLc. GCL is important for increasing GSH levels during oxidative stress, while GST is involved in the phase II metabolism conjugation of GSH to electrophilic xenobiotics. A study by Henry et al. also found GST-pi to be up-regulated; however, they attributed the up-regulation to the THF vehicle (Henry et al., 2007). In our study, 1% DMSO was used as the vehicle, not THF, so the common alterations in gene expression were likely due to the C₆₀ and not the solvent vehicle.

Microarray results revealed increased expression of additional oxidative stress-responsive genes. An increase in α-tocopherol transfer proteins (TTP) transcript levels further implicates an oxidative stress response and is in accordance with the previous studies which found α-tocopherol protected cells in culture from C₆₀-induced injury (Kamat et al., 1998; Sayes et al., 2005). Ferritin is a critical protein in detoxifying oxidative stress that has been linked to an antioxidant response element (ARE) (Iwasaki et al., 2006). Up-regulation of ferritin in response to C₆₀ exposure indicates a disruption in iron homeostasis and/or oxidative stress. Ferritin is regulated by levels of iron, cytokines, hormones and oxidative stress (Zandman-Goddard and Shoenfeld, 2007). There is a chance that C₆₀ had direct effects on these factors and subsequent indirect effects on ferritin transcription; however, it is more likely that ferritin was up-regulated to bind iron and prevent propagation of the oxidative stress cascade. The role of iron and ferritin for oxidative stress responses is an important area of research, particularly with regards to their interbalance following a stroke and in neurodegenerative disorders such as Alzheimer’s disease (Quitana et al., 2006).

Disparity among reports on C₆₀ oxidative potential could be attributed to the wide range of methods for solution preparation which likely have an impact on the biological activity and toxic potential of fullerenes. Our experimental design called for sonication of C₆₀ in 100% DMSO with dilution to final exposure concentrations of 1% DMSO. Granted, the use of such a solvent would increase potential uptake but we were interested in specifically detailing the interactions of fullerenes within biological systems, not assessing uptake from the environment into the system. Other methods of solubilization include the use of THF, ethanol, methanol or stirring for prolonged periods in direct sunlight (Yamago et al., 1995; Oberdorster, 2004; Sayes et al., 2004; Dhawan et al., 2006; Levi et al., 2006; Lovern and Klaper, 2006; Zhu et al., 2006). It is important to recognize that extensive methods of solubilization may also alter the physicochemical properties of C₆₀, thereby altering the interaction within the biological system. For example, with increased hydroxylation, the photosensitivity of C₆₀ decreases. This could be of potential concern for protocols that require prolonged stirring of C₆₀ in direct sunlight.

Results presented herein are in agreement with the studies that show C₆₀ can act as a pro-oxidant and elicit a toxic response via oxidative stress (Kamat et al., 1998; Yamakoshi et al., 2003; Oberdorster, 2004; Sayes et al., 2005). Chemically increasing antioxidant levels in vivo reduced adverse effects, while chemically inhibiting glutathione production increased sensitivity to C₆₀ exposure. Functionalization of C₆₀ may alter the biological response; however, there was no indication in this study to suggest that hydroxylation of C₆₀ imparts antioxidant properties. Future studies should be focused on the role surface functionalization and methods of C₆₀ preparation have on oxidative potential. The embryonic zebrafish has once again proven to be a valuable model for studying nanomaterial-biological interactions at multiple levels of biological organization, i.e. whole animal, cellular and molecular.