当前位置: 首页 > 期刊 > 《毒物学科学杂志》 > 2005年第2期 > 正文
编号:11294860
Butylhydroquinone Protects Cells Genetically Deficient in Glutathione Biosynthesis from Arsenite-Induced Apoptosis Without Significantly Cha
http://www.100md.com 《毒物学科学杂志》
     Center for Environmental Genetics and Department of Environmental Health, University of Cincinnati Medical Center, Cincinnati, Ohio 45267–0056

    Shriners Hospital for Children, Cincinnati, OH 45229

    ABSTRACT

    Arsenic, first among the top environmentally hazardous substances, is associated with skin, lung, liver, kidney, prostate, and bladder cancer. Arsenic is also a cardiovascular and a central nervous system toxicant, and it has genotoxic and immunotoxic effects. Paradoxically, arsenic trioxide is used successfully in the treatment of acute promyelocytic leukemia and multiple myeloma. Arsenic induces oxidative stress, and its toxicity is decreased by free thiols and increased by glutathione depletion. To further characterize the role of glutathione and oxidative stress in the toxicity of arsenic, we have used fetal fibroblasts from Gclm–/– mice, which lack the modifier subunit of glutamate-cysteine ligase, the rate-limiting enzyme in glutathione biosynthesis. Gclm–/– mouse embryo fibroblasts (MEFs) are eight times more sensitive to arsenite-induced apoptotic death. Because of a dramatic decrease in glutathione levels, Gclm–/– MEFs have a high prooxidant status that is not significantly relieved by treatment with the phenolic antioxidant tBHQ; however, tBHQ blocks arsenite-induced apoptosis in both Gclm+/+ and Gclm–/– cells, although it raises a significant antioxidant response only in Gclm+/+ cells. Global gene expression profiles indicate that tBHQ is significantly effective in reversing arsenite-induced gene deregulation in Gclm+/+ but not in Gclm–/– MEFs. This effect of tBHQ is evident in the expression of metalloproteases and chaperones, and in the expression of genes involved in DNA damage and repair, protein biosynthesis, cell growth and maintenance, apoptosis, and cell cycle regulation. These results suggest that regulation of glutathione levels by GCLM determines the sensitivity to arsenic-induced apoptosis by setting the overall ability of the cells to mount an effective antioxidant response.

    Key Words: arsenic; oxidative stress; apoptosis; glutathione; global gene expression analysis; Gclm knockout.

    INTRODUCTION

    Arsenic has been known for more than 100 years to be a potent human carcinogen (Wang and Rossman, 1996). Environmental arsenic contamination occurs from industrial smelting of metals, from power generation with coal, and from applications of pesticides and herbicides. A major route of human exposure is the accumulation of soil arsenic into grains, vegetables, fish, and meats. Arsenic enters the organism by dermal contact, inhalation, or ingestion of contaminated drinking water or foodstuffs. The tissues most often affected by exposure are skin, nasal passages, lungs, gastrointestinal tract, and liver. Because of urinary excretion, the kidneys are also important targets for arsenic toxicity (Nieboer and Fletcher, 1996). The toxicity of arsenite is reduced by addition of free thiols, such as glutathione (GSH), and is increased by GSH depletion. Binding to vicinal dithiol groups is much more avid than to GSH, requiring treatment with other dithiols for reversal (Clewell et al., 1999). For this reason, it is believed that many of the effects of arsenite result from its ability to inactivate sulfhydryl-containing enzymes through binding to critical vicinal dithiols.

    Arsenate (As5+) is the most common environmental form of inorganic arsenic, although arsenite (As3+) is more toxic and the most likely carcinogenic species (Tinwell et al., 1991). At the cellular level, arsenite induces the production of reactive oxygen species. Arsenic-induced oxidative stress has been shown to cause DNA damage through the production of superoxide and hydrogen peroxide (Kitchin, 2001) and to disrupt mitosis and promote apoptosis (States et al., 2002). Paradoxically, arsenic trioxide is an effective chemotherapeutic agent for certain types of leukemias because of its cytotoxic properties, especially its ability to induce apoptosis (Evens et al., 2004; Munshi et al., 2002). Further understanding of the cytotoxic mechanisms of arsenic may provide more effective management of environmental exposure while maximizing its chemotherapeutic potential.

    Inorganic arsenate induces a rapid burst of oxidative stress in mammalian cells as a result of the repetitive reduction of pentavalent to trivalent arsenic followed by the oxidative methylation of trivalent arsenic (Aposhian, 1997). This oxidative stress induces a compensatory increase in the levels of cellular GSH by increasing cystine uptake (Deneke et al., 1992), which is reduced intracellularly to cysteine and further utilized in GSH biosynthesis through stimulation of glutamate cysteine ligase (Ochi, 1997). It has been proposed that the genotoxic effects of arsenite may be mediated by the generation of reactive oxygen species (ROS) (Nesnow et al., 2002; Shi et al., 2004). In fact, addition of ROS scavengers to the culture medium protects against arsenite toxicity (Hei et al., 1998; Nordenson and Beckman, 1991). Glutathione depletion increases the toxic effects of arsenite (Li et al., 2002b; Oya-Ohta et al., 1996; Sakurai et al., 2002, 2004a), suggesting that the oxidative effects of arsenite might be amplified by GSH depletion. Cells, however, contain millimolar amounts of GSH and, under normal conditions, changes in homeostatic GSH levels are quickly restored by de novo synthesis. On the one hand, depletion of GSH might take place only if very high arsenite concentrations were maintained for long periods of time, which would be lethal. On the other hand, small changes in redox status resulting from physiological arsenic exposure might not be sufficient to produce the effects at the molecular level associated with long-term toxicity. Trivalent arsenic metabolites are potent inhibitors of glutathione reductase (Styblo et al., 1997) and thioredoxin reductase (Lin et al., 1999), and they deplete intracellular GSH, altering cellular redox status and leading to changes in cell signaling and lipid peroxidation, and ultimately causing DNA damage. Glutathione depletion further potentiates arsenic toxicity by reducing the scavenging of arsenicals into GSH conjugates (Sakurai et al., 2004b).

    To address the role of GSH depletion in arsenite toxicity, we used mice embryo fibroblasts (MEFs) from Gclm-knockout mice lacking the modifier subunit of glutamate-cysteine ligase (GCL), the rate-limiting enzyme in glutathione biosynthesis (Yang et al., 2002). These mice have normal levels of the catalytic glutamate cysteine ligase catalytic (GCLC) subunit, but in the absence of the glutamate cysteine ligase modifier (GCLM) subunit, GCLC is catalytically inefficient (Anderson, 1998), leading to a decrease in GSH levels to 9–16% of the levels in their Gclm+/+ littermates (Yang et al., 2002). Gclm–/–mice are viable, but their lower GSH levels makes them more sen sitive than wild-type mice to chemical oxidants such as H2O2 (Yang et al., 2002). We find that Gclm–/– MEFs are also significantly more sensitive to arsenite exposure than Gclm+/+or Gclm+/– MEFs. Treatment with the phenolic antioxidant tBHQ, which is considerably more effective in reversing arsenite-induced gene deregulation in Gclm+/+ than in Gclm–/– MEFs, blocks arsenite-induced apoptotic death, although it does not restore GSH levels or raise a significant antioxidant response.

    MATERIALS AND METHODS

    Cell lines and treatments.

    MEFs from Gclm+/+, Gclm+/– and Gclm–/– mice (Yang et al., 2002) were prepared by standard techniques from 14.5-day-old fetuses and grown in -MEM medium (Gibco BRL, Grand Island, NY) supplemented with 10% fetal bovine serum and 1% antibiotics (Antibiotic Antimycotic Solution, Sigma Aldrich Co., St. Louis, MO). Gclm–/– MEFs cease to proliferate within 5–6 passages and were used at passage 2–4. Cells were incubated at 37°C with 5% CO2 and grown to 90–100% confluence before treatment. As indicated in each individual experiment, cultures were treated with sodium arsenite, tBHQ, arsenite plus tBHQ, or control vehicle (referred to also as untreated). Unless otherwise stated, whenever sequential treatments were performed, the growth medium containing the first additive, usually tBHQ, was removed and replaced with growth medium containing only the second additive, usually arsenite. Aqueous solutions of NaAsO2 (hereinafter referred to as arsenite) (Sigma) were prepared as 1000x stocks. tBHQ (Fluka Chemical Co.) was dissolved in DMSO also as a 1000x stock solution. Specific concentrations and time points of treatment are indicated in the text and in the figure legends.

    Viability assays and apoptosis determination.

    Viability was measured using the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay (Denizot and Lang, 1986). Briefly, a fixed number of cells were seeded in 24-well plates and incubated for varying lengths of time in complete medium plus the indicated treatments. After treatment, the medium was changed and the cells were incubated in complete medium for 24–36 h after exposure to arsenite to allow for maximal expression of the cytotoxic phenotype (Komissarova et al., 2005), followed by addition of 100 μl of MTT reagent (Promega BioTech) to each well. The plates were incubated for 2 h at 37°C and read on a Wallac Victor2 plate reader at 544 nm. Alternatively, cells were counted under the microscope after staining with 0.4% trypan blue.

    Annexin-V staining was used to measure apoptosis. Cells were harvested, washed twice with phosphate buffered saline (PBS) and suspended in binding buffer at a density of 106 cells/ml. After counterstaining with propidium iodide, cells were analyzed by flow cytometry using an Annexin-V FITC Apoptosis Detection Kit (Calbiochem).

    Western blot analysis.

    To prepare whole cell extracts for protein analyses by Western immunoblotting, cells were washed twice with ice-cold PBS, harvested by scraping and collected by centrifugation. Pelleted cells were resuspended in 100 mM NaCl, 20 mM Tris-HCl pH8.0, 1 mM EDTA, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 1 mM PMSF, 0.1 mM NaVO4, 10 mM NaF, 10 mM NaP2O7, 0.5% NP-40, and lysed for 5 min on ice. After lysis, cells were briefly sonicated to disrupt the nuclei, and debris was removed by centrifugation at 12,000 rpm for 15 min. Protein concentration of the supernatants was determined and the supernatants were stored at –80°C until ready to use. Ten micrograms of protein were resolved by electrophoresis in 15% sodium dodecyl sulfate (SDS)-polyacrylamide gels and transferred to PVDF or nitrocellulose membranes. Membranes were blocked with 5% low-fat milk in PBST (0.1 M PBS with 0.1% Tween 20) and incubated for 1 h at room temperature with rabbit anti-PARP antibody (Cell Signaling Technology, Beverly, MA) at 1:1000 dilution in PBST containing 3% BSA. After washing with PBST, the membrane was incubated for 1 h in 1:10,000 goat anti-rabbit HRP-conjugated antibody (Santa Cruz) PBST containing 5% milk. After washing, bands were visualized using PicoWest Chemiluminescent Super Signal (Pierce Rockford, IL). The same procedure was used for detection of -actin using an anti--actin antibody generously provided by Dr. James L. Lessard (Cincinnati Children's Hospital Medical Center).

    Electrophoretic mobility shift assays.

    Nuclear extracts were prepared by procedures described previously (Carrier et al., 1992; Puga et al., 2000) with minor modifications. Nuclear extracts were obtained in a final volume of 100:μl buffer containing 2 mM EDTA, 2 mM dithiothreitol, 0.4 M KCl, 10% glycerol, and 25 mM HEPES, pH 7.9, at a protein concentration of 10–20 μg/μl. DNA-binding reactions were performed in a 20-μl reaction volume with 10,000 dpm (approximately 0.1 ng) of double-stranded EpRE or nuclear factor kappa B (NFB) probe and 5–15 μg nuclear protein, in a buffer containing 1 mM EDTA, 1 mM dithiothreitol, 80 mM KCl, 10% glycerol, 1 μg poly(dI-dC)-poly(dI-dC) carrier, and 20 mM HEPES, pH 7.8. One strand of each complementary pair of oligonucleotides was end-labeled with T4 polynucleotide kinase and [-32P]-ATP, and annealed to an excess of the unlabeled complementary oligonucleotide. Binding reactions were allowed to proceed for 20 min at room temperature, and samples were loaded onto non-denaturing 4% polyacrylamide gels. Following electrophoresis at 200 V for 2–3 h in 0.5x Tris-borate buffer, the gels were dehydrated and exposed to x-ray film. Oligonucleotide probes were (forward strand only) 5'-GATCGAGGGGACTTTCCCTAGC-3' and 5'-GAAATGACATTGCTAATGGTGACAAAGCAAC-3' for nuclear factor kappa B (NFB) and Nrf2, respectively.

    Analysis of glutathione levels.

    Glutathione and glutathione disulfide (GSSG) levels in whole cell extracts were determined spectrophotofluorometrically using o-phthalaldehyde (OPA, phthalic dicarboxaldehyde), as previously described (Senft et al., 2000). OPA reacts with GSH and has a high quantum yield but does not react with GSSG. To measure GSSG levels, GSSG was first reduced to GSH with dithionite and GSSG determined by subtraction. Redox potentials at pH 7.0 were calculated as described (Dalton et al., 2004) by inserting molar GSH and GSSG concentrations into the Nernst equation,

    RNA isolation, fluorescent labeling of target cDNAs, and high-density microarray hybridization.

    Total RNA from cells treated with sodium arsenite, tBHQ, sodium arsenite plus tBHQ, or from vehicle-treated control cells, was isolated using Tri Reagent (Invitrogen) according to the manufacturer's instructions, with additional purification steps applied to RNA samples used for microarray analysis. To verify RNA quality prior to labeling for microarray analyses, samples were analyzed with an Agilent 2100 Bioanalyzer. Labeling of cDNAs, preparation of microarrays, and hybridization reactions were performed by the University of Cincinnati Functional Genomics Core and are briefly described here. Fluorescence-labeled cDNAs were synthesized from 20 μg of total RNA using an indirect amino allyl labeling method (DeRisi et al., 1996). The cDNA was synthesized by an oligo(dT)-primed, reverse transcriptase reaction, and the cDNA was labeled with monofunctional reactive Cytidine-3 and Cytidine-5 dyes (Cy3 and Cy5; Amersham; Piscataway, NJ). Specific details of the labeling protocols may be found at http://microarray.uc.edu.

    The hybridization probes were from arrayed mouse oligonucleotide microarrays derived from the Operon/Qiagen Verified Libraries currently containing 31,769 sequences from annotated mouse genes, affixed each in a 100 μm diameter spot to polylysine-treated microscope slides. The hybridization targets were the paired Cy3– and Cy5–labeled control and test cDNAs, which were mixed in approximately equal proportions and applied to the microarray for hybridization under high stringency conditions. After hybridization and washing unhybridized targets, Cy3 (green) and Cy5 (red) fluorescent channels were simultaneously scanned with independent lasers at 10 μm resolution. Each comparison was done with quadruplicate biological replicates using flipped dye arrays to allow for the removal of gene-specific dye effects. Each comparison consisted of four microarray slides; in two, the cDNA was labeled with one fluorescent dye, and in the other two it was labeled with the other dye.

    Data analysis and normalization.

    Microarray hybridization data representing raw spot intensities generated by the GenePix Pro v. 5.0 software were analyzed to identify differentially expressed genes under different experimental conditions. Data normalization was performed in three steps for each microarray separately. First, channel-specific local background intensities were subtracted from the median intensity of each channel (Cy3 and Cy5). Second, background-adjusted intensities were log-transformed, and the differences (R) and averages (A) of log-transformed values were calculated as R = log2(X1) – log2(X2) and A = [log2(X1) + log2(X2)]/2, where X1 and X2 denote the Cy5 and Cy3 intensities after subtracting local backgrounds, respectively. Third, data centering was performed by fitting the array-specific local regression model of R as a function of A (Dudoit et al., 2002). The difference between the observed log-ratio and the corresponding fitted value represented the normalized log-transformed gene expression ratio. Normalized log-intensities for the two channels were then calculated by adding half of the normalized ratio to A for the Cy5 channel and subtracting half of the normalized ratio from A for the Cy3 channel.

    Identification of differentially expressed genes.

    The statistical analysis was performed for each gene separately by fitting the following mixed-effects linear model (Wolfinger et al., 2001): Yijk = μ + Ai + Sj + Ck+ ijk, where Yijk corresponds to the normalized log-intensity on the ith array (i = 1, ..., 52), with the jth treatment/genotype combination (j = 1, ... ,14), and labeled with the kth dye (k = 1 for Cy5, and 2 for Cy3). μ is the overall mean log-intensity, Ai is the effect of the ith array, Sj is the effect of the jth treatment/genotype combination and Ck is the effect of the kth dye. Assumptions about model parameters were the same as described elsewhere (Wolfinger et al., 2001), with array effects assumed to be random, and treatment and dye effects assumed to be fixed. Statistical significance of the differential expression between different treatment combinations was assessed by calculating p values and adjusting for multiple hypotheses testing by calculating false discovery rates (FDR) (Benjamini and Hochberg, 1995; Reiner et al., 2003). Genes used in Tables 1, 2, and 3 had fold-ratio 2 or –2 and FDR < 0.1 for those changes considered to be significantly different in the comparison. Given the lower number of genes in these tables, we do not expect more that 20–25 genes to be included by chance. Data normalization and statistical analyses were performed using the SAS statistical software package (SAS Institute Inc., Cary, NC).

    Hierarchical clustering of genes and samples was based on Bayesian mixture based clustering procedures for clustering gene expression data with experimental replicates (Medvedovic et al., 2004). In this approach, the statistical distribution of microarray data is described by a Bayesian mixture model. Clusters of co-expressed genes are created from the posterior distribution of clusterings, which is estimated by a Gibbs sampler. Clustering was performed with the Windows-based program GIMM (Gaussian Infinite Mixture Modeling) (Medvedovic et al., 2004), and results were displayed using the freely available TreeView computer program (Eisen et al., 1998). To minimize the number of false positives in these analyses, we chose a fold-ratio 2 or –2 and FDR < 0.05. As an alternative to hierarchical clustering, we organized the data at the level of known biological processes and metabolic pathways, which identifies those groups of biologically related genes in which a significantly larger number of gene expression changes have taken place. To this end we used GenMAPP (Gene MicroArray Pathway Profiler) (Dahlquist et al., 2002), which exploits the defined vocabulary of terms describing biological processes, cellular components, and molecular functions of all genes available from the Gene Ontology (GO) Consortium, in addition to contributed pathway MAPPS from other sources (Ashburner et al., 2000). To link gene expression data to the GO hierarchy, we used MAPPFinder (Doniger et al., 2003), a program that, for each specific GO node, calculates the percentage of the genes that were measured that meet a user-defined criterion and ranks them using this percentage and a standardized difference score. MAPPFinder information was tabulated by accessing each specific pathway reported by the program and extracting the relevant changes. Those results are presented in tabular form. For the statistical analyses presented later in Table 5, the absolute magnitudes of the gene expression changes in each of the 13 Gene Ontology groups were log(2) transformed and the resulting values were compared using two-tailed paired t-tests. Statistical significance was determined using an adjusted p value cutoff of 0.00385 based on a strict Bonferroni adjustment for testing 13 categories.

    RESULTS

    GCLM-null MEFs Are Extremely Sensitive to Arsenite

    It has been well documented that redox balance within the cell is regulated to a large extent by GSH and that arsenite promotes apoptosis in numerous cell types and across a wide range of doses (Pulido and Parrish, 2003) through the production of ROS (Busciglio and Yankner, 1995; Chen et al., 1998). Gclm–/– mice have recently been generated, and mouse embryonic fibroblasts (MEFs) from these animals have greatly reduced GSH levels (Yang et al., 2002). Availability of these mice and of cells derived from them allowed us to study the role of genetically induced GSH depletion in arsenic sensitivity without the problems intrinsic in studies where changes in GSH levels are induced chemically, with subsequent fast restoration of homeostatic levels by de novo synthesis. Arsenite dose–response survival rates were compared for MEFs from Gclm–/–, wild-type Gclm+/+, and hemizygous Gclm+/– mice. As expected, all three genotypes showed decreased viability as the arsenite dose increased. The dose response for cell killing of wild-type and Gclm+/– MEFs did not differ significantly, but the viability of the Gclm–/– cells decreased at a far greater rate than that of the other two (Fig. 1A). The ED50 for Gclm–/– MEFs was 11 μM, whereas for Gclm+/+ and Gclm+/– MEFs it was 86 μM, approximately 8 times higher. These data indicate that Gclm–/– MEFs are also highly sensitive to arsenite and that arsenite toxicity in these cells is a function of the depletion of intracellular GSH.

    To determine whether the decrease in viability resulting from arsenite treatment was due to apoptosis we used annexin-V staining. Gclm–/– MEFs were treated with 25 μM arsenite for 8 h or 24 h, allowed to recover for an additional 24 h in complete medium, harvested, stained with annexin V-FITC and propidium iodide, and scored by flow cytometry analysis. Control untreated cells were greater than 90% viable; in contrast, 8 h of arsenite treatment resulted in only 40% cell viability, with most of the remaining cells staining positively for both annexin-V and propidium iodide, indicative of late apoptosis. Treatment for 24 h resulted in the near elimination of the viable cell population, with 83% of the cells in late apoptosis (Fig. 1B). Interestingly, while these results clearly indicate that arsenite-promoted apoptosis is time-dependent and progressive, it also shows that there is no population-wide transition from early to late apoptosis, which is consistent with the notion that arsenic-treated cells become apoptotic as they progress through G2/M (States et al., 2002).

    tBHQ Protects Against Arsenite Toxicity

    Arsenite toxicity is partially mediated by the generation of ROS (Nesnow et al., 2002; Shi et al., 2004). To determine whether the phenolic antioxidant tBHQ, a potent activator of Nrf2-mediated protective responses (Nguyen et al., 2000), would inhibit arsenite toxicity in the sensitive Gclm–/– MEFs, we pretreated them for 24 h with varying concentrations of tBHQ prior to an 8-h treatment with different doses of arsenite. This regimen was followed by a 24h culture period in medium without additives. Cell viability increased with tBHQ concentration, being maximal at doses of 25 μM and 50 μM, at which point survival of Gclm–/– reached 90%, even at the highest arsenic dose tested, 64 μM (Fig. 2A).

    Not only the dose but also the duration of the tBHQ pretreatment had a major effect on the survival of arsenic-treated cells. Gclm–/– and Gclm+/– MEFs were pretreated with 10 μM tBHQ for 0, 12, or 24 h followed by doses of arsenite increasing from 4 to 64 μM. After 8 h, the medium was replaced with complete growth medium, and viability was scored 24 h later. Increasing the time of the tBHQ pretreatment also increased the cell viability for both genotypes, although the effect was considerably stronger for the Gclm+/– cells (Fig. 2B and 2C). These results indicate that protection by tBHQ is both time and concentration dependent.

    Annexin-V staining showed that tBHQ acted by blocking arsenite-induced apoptosis. Gclm+/+ MEFs were pretreated for 8 h with 50 μM tBHQ, followed by 12 h in 25 μM arsenite and 24 h in complete medium. As controls, cells treated for 12 h with 25 μM arsenite or with 50 μM tBHQ or untreated cells were used. After exposure to 25 μM arsenite, cell viability dropped to approximately 45% of the untreated controls. In contrast, pretreatment with 50 μM tBHQ provided enough protection from subsequent arsenic exposure to maintain 90% viability; these cells showed little loss of viability, with values statistically similar to those found for untreated cells or for cells treated with tBHQ alone (Fig. 3A).

    To confirm the annexin-V data, we examined the effect of tBHQ on a second diagnostic apoptotic event, namely poly-ADP ribose polymerase (PARP) cleavage. Activation of caspase-3 during apoptosis leads to the cleavage of the 116-kDa PARP into two diagnostic fragments of 85 kDa and 31 kDa (Chen et al., 1998; Pulido and Parrish, 2003). Gclm+/+ MEF cells were treated for 8 h with 50 μM tBHQ, with 25 μM arsenite, or with 25 μM arsenite added simultaneously with 50 μM tBHQ or after a 16-h pretreatment with tBHQ. Whole cell extracts were prepared and analyzed for PARP expression by Western immunoblotting. Arsenite treatment led to a significant level of PARP cleavage, as detected by the presence of a large amount of the 85-kDa cleavage product. Simultaneous co-treatment with 25 μM arsenite and 50 μM tBHQ significantly reduced the amount of cleaved product, while tBHQ pretreatment for 16 h prior to arsenite addition completely prevented PARP cleavage (Fig. 3B).

    tBHQ Blocks the Arsenite-Mediated Inhibition of NFB and Promotes Nrf2 Nuclear Translocation

    One of the best characterized effects of arsenite is the inhibition of NFB activation. High concentrations of arsenic inhibit NFB nuclear translocation and expression of NFB-dependent genes by mechanisms involving the reaction with thiol groups of specific cysteines in the IB kinases and (Kapahi et al., 2000; Roussel and Barchowsky, 2000). To determine whether tBHQ-mediated protection of arsenite toxicity involved effects on NFB, wild-type MEF cells were treated for 2 h with 25 μM arsenite, arsenite plus 50 μM tBHQ, or 50 μM tBHQ alone and examined for NFB nuclear translocation by electrophoretic mobility shift assay. As a positive control for the induction of NFB, we used cells treated with TNF-, and a fifth culture was left as the untreated control. The TNF- and tBHQ treatments visibly elevated the level of nuclear NFB relative to the untreated control (compare Fig. 3C, top panel lanes 3 and 4 to lane 2), including both p65/p50 heterodimers and p50/p50 homodimers. Extracts from cells treated with arsenite showed no NFB translocation and were indistinguishable from control untreated cells (compare lanes 5 and 2 in Fig. 3C, top panel). Simultaneous treatment with tBHQ and arsenite showed the formation of NFB-DNA complexes almost as intense as those found in tBHQ or TNF treatments, suggesting that one of the effects of tBHQ was to reverse the arsenite-mediated inhibition of NFB translocation.

    Exposure to tBHQ has been shown to trigger the nuclear accumulation of the transcription factor Nrf2 and to be critical in generating an antioxidant response and inducing the expression of several phase II detoxification genes (Dhakshinamoorthy and Jaiswal, 2001; Lee et al., 2001). When the same nuclear extract samples were analyzed with a DNA probe bearing the antioxidant response element (ARE) or the electrophile response element (EpRE), we found that arsenite treatment caused a small but significant level of Nrf2 translocation above untreated control cells (Fig. 3C, bottom panel, lanes 5 and 2). Induction was much greater in cells treated with, TNF or tBHQ, alone or in simultaneous combination with arsenite, which induced Nrf2 nuclear translocation to comparable levels (Fig. 3C, bottom panel, lanes 3, 4, and 6). These results suggest that one potential mechanism for the cytoprotective effects of tBHQ is likely to be its ability to induce gene expression changes through the activation of at least NFB and Nrf2.

    Gclm–/– MEFs Have a High Level of Oxidative Stress that Cannot be Significantly Decreased by tBHQ

    Arsenite has been shown to trigger apoptosis by its interaction with flavoprotein-containing superoxide-producing enzymes like NADPH-oxidase (Chen et al., 1998), an effect that is compounded by the ability of arsenite to deplete GSH. To determine if the protective properties of tBHQ directly affected the levels of intracellular glutathione, GSH and GSSG levels were determined in Gclm+/+ and Gclm–/– MEFs treated for 8 h with various concentrations of arsenite, with or without a prior 24-h pretreatment with 10 μM tBHQ. Pretreatment with tBHQ approximately doubled the intracellular GSH levels, regardless of genotype or arsenite treatment, although the GSH levels of Gclm–/– cells were about one-fourth the levels in wild-type Gclm+/+ cells (Fig. 4 A, B), as previously shown (Yang et al., 2002). Gclm+/+ MEFs, which have a complete GCL enzyme, maintained relatively constant GSH levels, between 3 and 4 mM for cells not pretreated with tBHQ and about 8 mM for pretreated cells, regardless of arsenite concentration. In contrast, treatment of Gclm–/– cells with arsenite concentrations of 10 μM or greater caused a significant GSH reduction from 1 to 0.1 mM, and this reduction could not be rescued by tBHQ pretreatment, even though tBHQ helped raise the initial GSH level in these cells to 2 mM. The levels of GSSG were also examined after similar pretreatments and arsenite doses. Glutathione disulfide increased with increasing arsenite nearly twofold in Gclm+/+ cells and threefold in Gclm–/– cells. tBHQ pretreatment reduced GSSG levels in both cases, causing them to plateau relative to arsenite concentration; however, the plateau value for the Gclm–/– cells was 0.7–0.8 mM, double the value for Gclm+/+ cells (Fig. 4C and 4D).

    With the [GSSG]/[GSH] ratio as a measure of the response of the redox environment in these two cell types to arsenite exposure, it became evident that the wild-type cells maintained steady [GSSG]/[GSH] ratios with little overall change regardless of tBHQ pretreatment. In contrast, [GSSG]/[GSH] ratios in Gclm–/– cells increased with arsenite dose and, although pretreatment with tBHQ was effective in partly attenuating this effect, the ratios for these cells remained significantly above those of the wild-type cells (Fig. 4E).

    Although there is no single method to determine cellular oxidative stress, the measure of both oxidized and reduced forms of a reversible redox couple like GSSG/2GSH provides a valid determination of the cellular oxidative (or reductive) state (Dalton et al., 2004). In addition, depletion of GSH by electrophiles, and/or increases in GSSG by oxidants, provides a sound estimate of the cellular oxidative stress response. To get a quantitative estimate of the redox state of these cells, we used the Nernst equation to calculate the reducing power of the GSSG/2GSH couple in these cells. Wild-type cells, without tBHQ pretreatment, maintained a mean E' of –150 mV almost independent of arsenite concentration, although at the highest concentration tested, the oxidative potential increased to –110 mV. Pretreatment with tBHQ increased the reducing potential of these cells to –200 mV regardless of arsenite concentration (Fig. 4F). In sharp contrast with these results, the reduction potential of Gclm–/– cells was severely compromised, even in the absence of arsenite treatment, which, by depleting GSH beyond its naturally low levels in these cells, generated positive E' values for doses greater than 5 μM arsenite. Pretreatment with 10 μM tBHQ did confer a measure of recovery, although E' levels never fell below –25 mV (Fig. 4F). These results clearly show that Gclm–/– cells are naturally under a very potent oxidative stress status that becomes stronger after arsenite exposures even as low as 5 μM. tBHQ is helpful in reducing to some extent the oxidative potential of these cells, but it is far from reaching the reductive levels homeostatically maintained by wild-type cells.

    Gene Expression Analyses of Arsenite- and tBHQ-Treated Cells

    Given the observation that arsenite induced cellular changes that could be reversed or blocked altogether by tBHQ pretreatment, it seemed reasonable to hypothesize that at least a fraction of the tBHQ-induced gene expression changes would protect cells from arsenite toxicity by blocking gene expression changes associated with arsenite exposure (Andrew et al., 2003; Rea et al., 2003). We approached the test of this hypothesis in three steps: first, we determined how different were the global gene expression levels in untreated Gclm+/+ and Gclm–/– MEFs; second, we examined their responses to treatment with 50 μM tBHQ; and third, we analyzed gene expression profiles in cells treated with vehicle control, with 20 μM arsenite for 8 h, and with 50 μM tBHQ for 8 h followed by removal of tBHQ and treatment with 20 μM arsenite for another 8 h. The results from the first experimental question are shown in Table 1. It is evident that ablation of the Gclm gene did not result in major changes in gene expression patterns. At a ratio 2 or –2, only 22 genes were significantly changed between Gclm+/+ and Gclm–/– cells (Table 1). Of these, expression of several genes involved in lipid metabolism and three insulin growth factor–binding proteins was decreased in the knockout cells, whereas expression of caspase-1, aconitase, and three small GTPases was increased. Hence, ablation of Gclm does not cause drastic changes in gene expression profiles. In contrast, treatment with 50 μM tBHQ caused distinct responses in the two cell lines. In Gclm+/+ MEFs, tBHQ induced global gene expression changes typical of an antioxidant response, in good agreement with what has already been well defined by other labs (Li et al., 2002a). In addition to phase II detoxification genes like Gsta and Nqo1, many genes responsive to oxidative stress, like Hmox1, Txnrd1, and many oxidoreductases were also upregulated in Gclm+/+ MEFs, including both genes, Gclc and Gclm, coding for the two subunits of the GCL holoenzyme. By comparison, except for Gclc, Hmox1, and Txnrd1, tBHQ treatment of Gclm–/– cells was of no consequence to the expression of genes coding for the oxidoreductases and phase II enzymes affected in Gclm+/+ cells (Table 2), suggesting that the ability of cells to mount a tBHQ-dependent antioxidant response is somehow dependent on cellular GSH levels or on other potential functions of GCLM.

    Treatment with arsenite alone or with consecutive additions of tBHQ and arsenite led to significant deregulation of 5014 genes (ratio –2 or 2; FDR < 0.05) from the 31,769 sequences contained in the array library. Hierarchical clustering of these changes revealed that the response to 20 μM arsenite treatment was very similar in both Gclm+/+ and Gclm–/– MEF cells and that pretreatment with tBHQ blocked the deregulation of a much larger number of genes in wild-type Gclm+/+ cells than in their knockout counterparts (Fig. 5A). The Venn diagrams in Figure 5B provide numerical values for these changes. Of the genes deregulated by arsenite in wild-type cells, 524 were changed in both arsenite-treated and tBHQ plus arsenite-treated cells, whereas 2647 and 562 were specific for each treatment, indicating that tBHQ blocked the deregulation by arsenite of 2647 genes, although it affected the expression of 562 genes that were not significantly changed by arsenite alone. A similar result, with different numbers of genes, was found for the knockout cells. A large number of genes, 1748, were affected by arsenite alone in both wild-type and knockout MEFs. Of these, 1267 were also affected by tBHQ pretreatment in either cell line or in both. Ninety-six percent of these genes (343 + 868 = 1211) were still significantly changed in tBHQ pretreated knockout cells, whereas only 31% were significantly changed in tBHQ pretreated wild-type cells. Taking into account that deregulation by arsenite of 343/1267 = 27% of the genes is common to the two cell lines, the number of genes significantly deregulated in both cell lines that cannot be blocked by tBHQ in Gclm–/– is 868, or 68%, whereas the equivalent number for Gclm+/+ cells is only 56, or 4%. This is a vast difference that suggests that the ability of tBHQ to block arsenite toxicity is largely dependent on GSH levels or on other functions that the GCLM protein might have.

    To identify groups of biologically related genes in which a significant number of gene expression changes had taken place, we used GenMAPP (Dahlquist et al., 2002), which searches all genes available from the Gene Ontology (GO) Consortium and from pathway MAPPS contributed from other sources and organizes the data at the level of known cellular processes and metabolic pathways. Relevant GenMAPP data was extracted and the most salient information is shown in Table 3. In addition, Table 3 includes genes upregulated or downregulated by 8-h exposure of both Gclm+/+ and Gclm–/– MEFs to two concentrations, 2 μM and 20 μM, of arsenite with and without a prior treatment with 50 μM tBHQ also for 8 h. Exposures of Gclm+/+ MEFs to 20 μM arsenite and of Gclm–/– MEFs to 2 μM are equitoxic, because both lead to approximately 20% cytotoxicity (Fig. 1A); however only the group of oxidative stress genes show some similarity of response in the two cells at equitoxic treatments.

    Most genes in the other gene groups showed significant changes only at the high concentration treatment. Consistent with the prooxidant effect of arsenite, several oxidative stress-inducible genes were highly upregulated by 20 μM arsenite in both cells. For the most part, tBHQ did not reverse this effect. Many of these genes are the same ones that were upregulated by tBHQ treatment in the absence of arsenite (see Table 2), including Hmox1, Txnrd1, Blvrb, and Gsta1. In wild-type cells, Gclm was induced by arsenite, and both Gclm and Gclc were induced by tBHQ + arsenite. In knockout cells, only Gclc was induced, as could be expected.

    A large number of DNA repair genes were also deregulated by arsenite, an effect reversed in almost all cases by tBHQ in wild-type MEFs, but not as well in knockout MEFs. Arsenite also downregulated the expression of a mitotic checkpoint protein, a few chromatin remodeling proteins and several histones, including histone H2A.X, whose modification is an early marker of DNA fragmentation (Rogakou et al., 2000). Histone methyltransferases, on the other hand, were upregulated by arsenite. For the most part, these changes were blocked by tBHQ in both cell lines. Many genes involved in TGF- signaling and in integrin-mediated cell adhesion were downregulated by arsenite and this effect was largely insensitive to tBHQ pretreatment. Expression of matrix metalloproteinases, proteasome subunits, and protease inhibitors was also altered by arsenite. Approximately half as many metalloproteinases were upregulated as were downregulated, although Mmp10, the gene coding for stromelysin-2, was upregulated as much as 44-fold and 25-fold in arsenite-treated wild-type and knockout cells, respectively. Peculiarly, Mmp11, coding for stromelysin-3, was downregulated by almost as much, 13-fold and 6-fold, respectively. Many mitochondrial protein synthesis genes were downregulated by arsenite, an effect that was blocked by tBHQ, suggesting that tBHQ was able to reverse the mitochondrial damage associated with apoptosis. The effects of arsenic on protein homeostasis were also evident on the deregulation of chaperone proteins. Eighteen different heat shock proteins and the HSF1 transcription factor that controls their transcription were upregulated by arsenite in both cells, and, consistent with the lack of effect on the expression of heat shock proteins shown in Table 2, tBHQ had no effect on their deregulation by arsenite.

    Major discrepancies in the responses of Gclm+/+ and Gclm–/– MEFs were found on the expression of genes involved in cell cycle regulation, cell growth and maintenance, and apoptosis. In general, both types of cells responded very similarly to arsenite treatment, showing upregulation or downregulation of the same genes to comparable magnitudes; however, significant differences were evident in the reversal of arsenite effects by tBHQ, with deregulation of many more genes being reversed in Gclm+/+ than in Gclm–/– cells. For example, among the anaphase-promoting complex components, Anapc-2 and -4 were downregulated in both cells, but Anapc-1 was downregulated only in Gclm+/+ MEFs. Deregulation of Anapc-1 and -4, but not -2, was reversed by tBHQ. A similar picture emerges from the expression of genes involved in the G1/S transition, with some genes, like Rb1 (p105) and Rbl1 (p107) being deregulated in opposite directions by arsenite with no reversal by tBHQ, while others, like p53 or Mcm7, were downregulated by arsenite and reversed by tBHQ.

    Expression of many transcription factor encoding genes and genes that regulate protein translation was also significantly altered by arsenite, including genes like Tsc1 and Pten that regulate mTOR and maintain proper size and growth patterns of cells (Tee and Blenis, 2005). Most changes induced by arsenite in the expression of genes in this group associated with cell growth and maintenance were reversed by tBHQ in wild-type cells, but not to a comparable extent in knockout cells. This divergent response was more evident in the responses of genes in the apoptosis group. Pro-apoptotic effectors, such as NFB inhibitors, death-domain proteins, and Fas, among others, were upregulated in both cell types, but this effect was downregulated by tBHQ only in wild-type cells. Both pro-and anti-apoptotic effects were upregulated and downregulated by arsenite; the overall direction of these changes can best be appreciated with the summary shown in Table 4. Of the 30 genes included in the apoptosis group, 14 changes in arsenite-treated Gclm+/+ MEFs were proapoptotic, 11 were antiapoptotic, and the remaining 5 genes did not change. Pretreatment with tBHQ reversed the expression changes of 25 of these genes, leaving only one antiapoptotic change and four proapoptotic changes. In contrast, arsenite treatment of Gclm–/– MEFs led to approximately the same number of gene expression changes in all three groups, with 11 being proapoptotic, 11 being antiapoptotic, and 8 unchanged. Pretreatment with tBHQ did not appreciably change this distribution.

    To determine whether the degree of arsenic-mediated gene expression change was statistically different between the cell types and treatments, we used the log(2)-transformed values of the absolute value of the gene expression changes in each of the 13 Gene Ontology groups and compared them as groups using two-tailed paired t-tests. Statistical significance was determined based on an adjusted p value cutoff of 0.00385 determined from the Bonferroni adjustment for testing multiple groups. Of the 13 GO categories, the response of Gclm+/+ and Gclm–/– cells to treatment with 20 μM arsenite was significantly different only for the Oxidative Stress gene group; differences in the other 12 groups were not statistically significant (Table 5). Pretreatment with 50 μM tBHQ was ineffective in reversing the gene deregulation caused by arsenite in the Gclm–/– cells, whereas it significantly reversed the effect of arsenite in seven of the gene groups for Gclm+/+ cells, including DNA Damage and Repair, Matrix Metalloproteases, Protein Biosynthesis, Cell Growth and Maintenance, Apoptosis, Chaperone Activity, and Cell Cycle Regulation.

    DISCUSSION

    The results presented in this article show that embryonic fibroblasts from Gclm–/– mice, genetically impaired in GSH biosynthesis, are considerably more sensitive to sodium arsenite toxicity than fibroblasts from Gclm+/+ or Gclm+/– mice. Arsenite dose– and time–response experiments indicate that the sodium arsenite ED50 for null cells is 8x lower, at approximately 10 μM, than the ED50 for wild-type cells. This concentration range is within the limits where effects on signal transduction pathways and cytotoxicity have been observed in rodent and human cells (Bode and Dong, 2002) or used to induce complete remission in acute promyelocytic leukemia patients (Shen et al., 1997). The concentrations are one order of magnitude above acceptable environmental levels in drinking water, although in many highly contaminated parts of the world, humans are exposed to more than 10x the concentration used in these experiments (Duker et al., 2005; Guha Mazumder, 2005). For cells of both genotypes, toxicity leads to death, and both outcomes may be blocked by pretreatment with the phenolic antioxidant tBHQ. tBHQ, however, does not regenerate GSH in null cells to levels comparable to those in wild-type cells, nor does it reduce the high oxidative stress levels found in Gclm–/– cells. Protection by tBHQ was also observed at the gene expression level. The DNA binding activities of NFB and Nrf2 were not induced in cells treated with arsenite alone, but they were when the cells were treated with tBHQ or with tBHQ plus arsenite. In addition, deregulation of gene expression resulting from arsenite exposure was almost completely eliminated in wild-type cells by pretreatment with tBHQ, but not so in Gclm–/– cells. In summary, just as it has been shown for H2O2-induced oxidative stress and apoptosis (Li et al., 2002a), tBHQ also protects against arsenite-induced apoptosis.

    tBHQ can redox cycling to the quinone and the semiquinone and form reactive oxygen, thus acting as an oxidant as well as an antioxidant. The t-butylquinone, being an electrophile, can conjugate with GSH or with protein thiols. It could be argued that the difference in GSH levels between Gclm+/+ and Gclm–/– cells are a major determinant of the difference in response of the two cells to tBHQ. Although possible, we think that this is unlikely, because the stoichiometric GSH-tBHQ conjugates have been shown to lead to intracellular accumulation in the nanomolar to micromolar GSH range (Nakamura et al., 2003b), well below the low GSH concentration in the Gclm–/– cells.

    The majority of the apoptotic cells in our experiments are positive for both annexin-V and propidium iodide, indicative of externalized phosphatidylserine and of a fairly permeable membrane. Time-course analysis indicates that the number of annexin-V–positive, propidium iodide–negative cells, diagnostic of early stage apoptosis, does not change significantly from 8 h to 24 h after arsenite treatment, whereas the number of late apoptotic cells doubles in this time period. These results suggest that once apoptosis has been initiated, arsenic induces the rapid degeneration of the membrane, in good agreement with previous data from States and colleagues suggesting that arsenic induces the continued generation of apoptotic cells that subsequently lose membrane integrity and degenerate (States et al., 2002). In this context, it is worth noting that several of the genes found to be deregulated by arsenite exposure are involved in membrane structure, extracellular matrix formation, and cell adhesion. Excessive propidium iodide staining is also consistent with the apoptotic cells having a 4 N DNA content, as might be expected from arsenite inducing apoptosis at the G2/M phases of the cell cycle (States et al., 2002). Deregulation of genes coding for the anaphase-promoting complex is also consistent with arsenite-induced apoptosis occurring in G2/M.

    Although the GSSG/2GSH ratio is not directly involved in intermediary metabolism, this redox couple provides an estimate of the thiol redox status of the cell under equilibrium conditions where GSH and GSSG do not change rapidly. Furthermore, depletion of GSH by electrophiles or increases in GSSG by oxidants provide a sound estimate of the cellular oxidative stress response, even though simple changes in GSH or GSSG are not sufficient to estimate the quantitative impact of a change in thiol redox state on specific cellular functions. Nonetheless, the large differences in homeostatic GSH levels between Gclm+/+ and Gclm–/– cells are a good indication that, because they lack a functional GCLM subunit, the latter are under significant oxidant conditions even without an arsenite challenge. Gclm–/– cells have twice the oxidizing potential (or half the reducing potential) of Gclm+/+ cells. By 10 μM arsenite, Gclm–/– cells have a E' of +10 mV, whereas Gclm+/+ cells are still at –146 mV. Higher arsenite concentrations narrow somewhat the gap between the two cells, mostly because the oxidant potential of the Gclm–/– cannot get much worse than the highly positive value of +24 mV already reached by 20 μM arsenite, whereas the increases in arsenite concentration decrease the reducing potential of the Gclm+/+ cells to –108 mV. Pretreatment with 10 μM tBHQ raises GSH and lowers GSSG levels in both cells, although the differences between the two cell lines remain; however, wild-type cells maintain GSH levels at all arsenite concentrations, whereas GCLM-null cells do not respond to the arsenite challenge. Treatment with tBHQ reduces the high oxidative potential of GCLM-null cells, but it does not counteract significantly the effect of arsenic in these cells. In contrast, tBHQ-pretreated GCLM-positive cells show a reducing potential in the –200 mV range. It is reasonable to conclude that these differences between Gclm+/+ and Gclm–/– cells, both intrinsic as well as in response to arsenite and tBHQ, occur because, even though they retain a functional Gclc gene, the ability of the null cells to restore GSH levels is impaired by the ablation of the GCLM subunit.

    These differences in redox potential between Gclm+/+ and Gclm–/– cells do not translate into major changes in gene expression profiles in unchallenged cells. Very few genes seem to be deregulated by Gclm ablation, including downregulation of several sterol metabolism genes, the LDL receptor, two collagen genes, and three insulin growth factor binding proteins, and upregulation of three extracellular matrix proteins and three small GTPase genes. These genes might be indicative of possible problems in cholesterol transport and metabolism and membrane physiology, consistent with the shorter life span of these cells (Dalton and Chen, unpublished).

    Our global profiling data indicate that arsenite treatment deregulates the expression of a large number of genes in both Gclm+/+ and Gclm–/– cells. Interestingly, most parallel expression changes in the two cell lines happen at equimolar rather than at equitoxic arsenite concentrations, suggesting that perhaps these gene-regulatory events induced by arsenite are involved with adaptive rather than with toxic changes. Many changes taking place in expression of DNA damage and repair genes are consistent with the known genotoxic effects of arsenic (Rossman, 2003). Not surprisingly, these changes extend to genes coding for proteins necessary to maintain chromosome structure, including histones and chromatin remodeling factors. Posttranslational changes in histone H2A.X and H2B.A are diagnostic of DNA damage (Rogakou et al., 2000), and it would be interesting to determine whether arsenite causes both transcriptional and posttranslational changes in their expression and whether one of the main effects of arsenic might be to upset the histone code. Other significant changes are largely consistent with the suppression of TGF- signals, inhibition of integrin-mediated cell adhesion, induction of multiple transcription factors, repression of co-repressors, and derailment of cell cycle regulatory functions. Arsenite exposure also brings about profound changes in protein levels in what appear to be conflicting regulatory changes. These changes go hand in hand with massive upregulation of heat shock proteins, metalloproteinases, and proteasome components, suggesting that arsenite induces critical changes in protein folding and structure and that the cells mount a major effort to properly refold misfolded proteins or to eliminate them altogether. Concomitantly, positive regulators of translation, such as Tuberous sclerosis-1, are upregulated and negative regulators, such as pTen, are downregulated. Indicative of distress in the mitochondria, many genes coding for mitochondrial ribosomal proteins are also downregulated. Paradoxically, several protease inhibitors are also upregulated. Upregulation of heat shock proteins appears to be a molecular signature of exposure to oxidants, because it has been observed in all other arsenite global profiling experiments (Andrew et al., 2003; Rea et al., 2003), as well as in expression profiles of H2O2, another strong oxidant (Li et al., 2002a). Arsenite also deregulates genes involved in the induction of apoptosis as well as in protection from apoptosis. As with the Gene Ontology processes discussed above, arsenite induced conflicting changes in apoptosis-related genes, with roughly half of them being proapoptotic and the other half antiapoptotic. The expression changes described do not appear to be exclusive of fibroblast cells; quite to the contrary, they may constitute a molecular signature of arsenic exposure. Genes in many of these same processes and functions are also deregulated by arsenite in lung cell lines as well as in keratinocyte cell lines (Andrew et al., 2003; Rea et al., 2003).

    Arsenite induces (or reinforces) the generation of oxidative stress in both GCLM-positive and -negative cells, but this arsenite-generated oxidative stress and the one generated by tBHQ seem to have critically different properties. Whereas tBHQ induces the expression of many oxidoreductases capable of transferring electrons between oxidized and reduced substrates, arsenite does not. Consequently, the antioxidant response raised by tBHQ is not elicited by arsenite. As a result, tBHQ causes a major reversal of the gene expression changes induced by arsenite. Clustering analysis of these changes (Fig. 5) unambiguously demonstrates that this reversal requires GCLM functions because it takes place in wild-type cells to a much larger extent than in knockout cells. This is the same conclusion that can be drawn from the data in Table 2, which shows that cells need either GSH or GCLM to mount a tBHQ-mediated antioxidant response.

    Induction of oxidative stress, possibly by GSH depletion, together with inhibition of NFB (Kapahi et al., 2000; Roussel and Barchowsky, 2000) are believed to enhance stress-induced JNK expression and promote apoptosis (Davison et al., 2004). Mechanistically, how these molecular events lead to apoptosis is still unknown. tBHQ is a potent activator of Phase II detoxification genes mediated by the transcription factor Nrf2 (Li et al., 2002a). Nrf2 has been shown to be activated by arsenite with subsequent induction of Gclc and Nqo1 (Pi et al., 2003), although nuclear translocation, a measure of its activation, was maximal 12 h post-treatment (Pi et al., 2003). It is not surprising that within the 2-h treatment period of our mobility shift experiments we detected only a small yet significant increase in nuclear Nrf2; however, arsenite pretreatment did not protect cells against apoptosis induced by a second arsenite challenge (Kann and Puga, unpublished), suggesting that induction of the Nrf2 repertoire of antioxidant genes is not sufficient to protect cells from arsenic toxicity. Consistent with this conclusion, our results in Gclm–/– cells strongly indicate that the protection from apoptosis provided by tBHQ does not include a significant change in the oxidant status of the cells. By a concentration of 10 μM arsenite, practically 100% of the Gclm–/– cells survive if pretreated with tBHQ, even though their oxidative potential is still very high (Fig. 4F). This is not to say that oxidant status is irrelevant to survival. It is clear that a high oxidative potential, as is the case in GCLM-null cells, sets the overall sensitivity to arsenic. Surviving the arsenic challenge does not seem to depend strongly on the oxidative level of the cells, but on their ability to respond either by de novo synthesizing more GSH or by performing some other as yet uncharacterized GCLM-dependent function.

    Treatment with arsenic trioxide has proven to be a successful chemotherapeutic agent for certain types of leukemias because of its cytotoxic properties, especially its ability to induce apoptosis (Evens et al., 2004; Munshi et al., 2002). It goes without saying that simultaneous administration of antioxidants might interfere with the treatment. On the other hand, a polymorphism in the human GCLM gene promoter that shows impaired promoter activity, is associated with coronary vasomotor function and myocardial infarction (Nakamura et al., 2000, 2003a). Those individuals may be likely to benefit from antioxidant therapies.

    NOTES

    1 These two authors contributed equally to the work in this article.

    2 Present address: Grünenthal GmbH, 52099 Aachen, Germany.

    ACKNOWLEDGMENTS

    We thank Daniel Nebert for the use of Gclm knockout mice and James Lessard for a gift of anti-actin antibody. This research was supported by NIEHS grants R01 ES10807, Center for Environmental Genetics grant P30 ES06096, Superfund Basic Research Program grant P42 ES04908 and by a grant from Phillip Morris USA. J.F.R. is a Postdoctoral Trainee partly supported by NIEHS T32 ES07250, Environmental Carcinogenesis and Mutagenesis Training Grant. Gene array data has been submitted to the MIAMEXPRESS public repository at http://www.ebi.ac.uk/miamexpress/ with E-MEXP-371. Conflict of interest: none declared.

    REFERENCES

    Anderson, M. E. (1998). Glutathione: An overview of biosynthesis and modulation. Chem. Biol. Interact. 111–112, 1–14.

    Andrew, A. S., Warren, A. J., Barchowsky, A., Temple, K. A., Klei, L., Soucy, N. V., O'Hara, K. A., and Hamilton, J. W. (2003). Genomic and proteomic profiling of responses to toxic metals in human lung cells. Environ. Health Perspect. 111, 825–835.

    Aposhian, H. V. (1997). Enzymatic methylation of arsenic species and other new approaches to arsenic toxicity. Annu. Rev. Pharmacol. Toxicol. 37, 397–419.

    Ashburner, M., Ball, C. A., Blake, J. A., Botstein, D., Butler, H., Cherry, J. M., Davis, A. P., Dolinski, K., Dwight, S. S., Eppig, J. T. et al. (2000). Gene ontology: Tool for the unification of biology. The Gene Ontology Consortium. Nat. Genet. 25, 25–29.

    Benjamini, Y., and Hochberg, Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing. J. R. Stat. Soc. B 57, 289–300.

    Bode, A. M., and Dong, Z. (2002). The paradox of arsenic: Molecular mechanisms of cell transformation and chemotherapeutic effects. Crit Rev. Oncol. Hematol. 42, 5–24.

    Busciglio, J., and Yankner, B. A. (1995). Apoptosis and increased generation of reactive oxygen species in Down's syndrome neurons in vitro. Nature 378, 776–779.

    Carrier, F., Owens, R. A., Nebert, D. W., and Puga, A. (1992). Dioxin-dependent activation of murine Cyp1a-1 transcription requires protein kinase C–dependent phosphorylation. Mol. Cell. Biol. 12, 1856–1863.

    Chen, Y. C., Lin-Shiau, S. Y., and Lin, J. K. (1998). Involvement of reactive oxygen species and caspase 3 activation in arsenite-induced apoptosis. J. Cell Physiol. 177, 324–333.

    Clewell, H. J., Gentry, P. R., Barton, H. A., Shipp, A. M., Yager, J. W., and Andersen, M. E. (1999). Requirements for a biologically realistic cancer risk assessment for inorganic arsenic. Int. J. Toxicol. 18, 131–147.

    Dahlquist, K. D., Salomonis, N., Vranizan, K., Lawlor, S. C., and Conklin, B. R. (2002). GenMAPP, a new tool for viewing and analyzing microarray data on biological pathways. Nat. Genet. 31, 19–20.

    Dalton, T. P., Chen, Y., Schneider, S. N., Nebert, D. W., and Shertzer, H. G. (2004). Genetically altered mice to evaluate glutathione homeostasis in health and disease. Free Radic. Biol. Med. 37, 1511–1526.

    Davison, K., Mann, K. K., Waxman, S., and Miller, W. H., Jr. (2004). JNK activation is a mediator of arsenic trioxide–induced apoptosis in acute promyelocytic leukemia cells. Blood 103, 3496–3502.

    Deneke, S. M., Lawrence, R. A., and Jenkinson, S. G. (1992). Endothelial cell cystine uptake and glutathione increase with N,N-bis(2-chloroethyl)-N-nitrosourea exposure. Am. J. Physiol. 262, L301–L304.

    Denizot, F., and Lang, R. (1986). Rapid colorimetric assay for cell growth and survival. Modifications to the tetrazolium dye procedure giving improved sensitivity and reliability. J. Immunol. Methods 89, 271–277.

    DeRisi, J., Penland, L., Brown, P. O., Bittner, M. L., Meltzer, P. S., Ray, M., Chen, Y., Su, Y. A., and Trent, J. M. (1996). Use of a cDNA microarray to analyse gene expression patterns in human cancer. Nat. Genet. 14, 457–460.

    Dhakshinamoorthy, S., and Jaiswal, A. K. (2001). Functional characterization and role of INrf2 in antioxidant response element-mediated expression and antioxidant induction of NAD(P)H:quinone oxidoreductase1 gene. Oncogene 20, 3906–3917.

    Doniger, S. W., Salomonis, N., Dahlquist, K. D., Vranizan, K., Lawlor, S. C., and Conklin, B. R. (2003). MAPPFinder: Using Gene Ontology and GenMAPP to create a global gene-expression profile from microarray data. Genome Biol. 4, R7.

    Dudoit, S., Yang, Y. H., Callow, M. J., and Speed, T. P. (2002). Statistical methods for identifying differentially expressed genes in replicated cDNA microarray experiments. Statist. Sinica 12, 111–139.

    Duker, A. A., Carranza, E. J., and Hale, M. (2005). Arsenic geochemistry and health. Environ. Int. 31, 631–641.

    Eisen, M. B., Spellman, P. T., Brown, P. O., and Botstein, D. (1998). Cluster analysis and display of genome-wide expression patterns. Proc. Natl. Acad. Sci. U.S.A. 95, 14863–14868.

    Evens, A. M., Tallman, M. S., and Gartenhaus, R. B. (2004). The potential of arsenic trioxide in the treatment of malignant disease: past, present, and future. Leuk. Res. 28, 891–900.

    Guha Mazumder, D. N. (2005). Effect of chronic intake of arsenic-contaminated water on liver. Toxicol. Appl. Pharmacol. 206, 169–175.

    Hei, T. K., Liu, S. X., and Waldren, C. (1998). Mutagenicity of arsenic in mammalian cells: Role of reactive oxygen species. Proc. Natl. Acad. Sci. U.S.A. 95, 8103–8107.

    Kapahi, P., Takahashi, T., Natoli, G., Adams, S. R., Chen, Y., Tsien, R. Y., and Karin, M. (2000). Inhibition of NF-kappa B activation by arsenite through reaction with a critical cysteine in the activation loop of Ikappa B kinase. J. Biol. Chem. 275, 36062–36066.

    Kitchin, K. T. (2001). Recent advances in arsenic carcinogenesis: Modes of action, animal model systems, and methylated arsenic metabolites. Toxicol. Appl. Pharmacol. 172, 249–261.

    Komissarova, E. V., Saha, S. K., and Rossman, T. G. (2005). Dead or dying: The importance of time in cytotoxicity assays using arsenite as an example. Toxicol. Appl. Pharmacol. 202, 99–107.

    Lee, J. M., Hanson, J. M., Chu, W. A., and Johnson, J. A. (2001). Phosphatidylinositol 3-kinase, not extracellular signal-regulated kinase, regulates activation of the antioxidant-responsive element in IMR-32 human neuroblastoma cells. J. Biol. Chem. 276, 20011–20016.

    Li, J., Lee, J. M., and Johnson, J. A. (2002a). Microarray analysis reveals an antioxidant responsive element-driven gene set involved in conferring protection from an oxidative stress–induced apoptosis in IMR-32 cells. J. Biol. Chem. 277, 388–394.

    Li, M., Cai, J. F., and Chiu, J. F. (2002b). Arsenic induces oxidative stress and activates stress gene expressions in cultured lung epithelial cells. J. Cell Biochem. 87, 29–38.

    Lin, S., Cullen, W. R., and Thomas, D. J. (1999). Methylarsenicals and arsinothiols are potent inhibitors of mouse liver thioredoxin reductase. Chem. Res. Toxicol. 12, 924–930.

    Medvedovic, M., Yeung, K. Y., and Bumgarner, R. E. (2004). Bayesian mixture model based clustering of replicated microarray data. Bioinformatics 20, 1222–1232.

    Munshi, N. C., Tricot, G., Desikan, R., Badros, A., Zangari, M., Toor, A., Morris, C., Anaissie, E., and Barlogie, B. (2002). Clinical activity of arsenic trioxide for the treatment of multiple myeloma. Leukemia 16, 1835–1837.

    Nakamura, S., Kugiyama, K., Sugiyama, S., Miyamoto, S., Koide, S., Fukushima, H., Honda, O., Yoshimura, M., and Ogawa, H. (2002). Polymorphism in the 5'-flanking region of human glutamate-cysteine ligase modifier subunit gene is associated with myocardial infarction. Circulation 105, 2968–2973.

    Nakamura, S., Sugiyama, S., Fujioka, D., Kawabata, K., Ogawa, H., and Kugiyama, K. (2003a). Polymorphism in glutamate-cysteine ligase modifier subunit gene is associated with impairment of nitric oxide–mediated coronary vasomotor function. Circulation 108, 1425–1427.

    Nakamura, Y., Kumagai, T., Yoshida, C., Naito, Y., Miyamoto, M., Ohigashi, H., Osawa, T., and Uchida, K. (2003b). Pivotal role of electrophilicity in glutathione S-transferase induction by tert-butylhydroquinone. Biochemistry 42, 4300–4309.

    Nesnow, S., Roop, B. C., Lambert, G., Kadiiska, M., Mason, R. P., Cullen, W. R., and Mass, M. J. (2002). DNA damage induced by methylated trivalent arsenicals is mediated by reactive oxygen species. Chem. Res. Toxicol. 15, 1627–1634.

    Nguyen, T., Huang, H. C., and Pickett, C. B. (2000). Transcriptional regulation of the antioxidant response element. Activation by Nrf2 and repression by MafK. J. Biol. Chem. 275, 15466–15473.

    Nieboer, E., and Fletcher, G. G. (1996). Determinants of reactivity in metal toxicology. In Toxicology of Metals, pp. 113–132. CRC Press, Boca Raton, FL.

    Nordenson, I., and Beckman, L. (1991). Is the genotoxic effect of arsenic mediated by oxygen free radicals Hum. Hered. 41, 71–73.

    Ochi, T. (1997). Arsenic compound–induced increases in glutathione levels in cultured Chinese hamster V79 cells and mechanisms associated with changes in gamma-glutamylcysteine synthetase activity, cystine uptake and utilization of cysteine. Arch. Toxicol. 71, 730–740.

    Oya-Ohta, Y., Kaise, T., and Ochi, T. (1996). Induction of chromosomal aberrations in cultured human fibroblasts by inorganic and organic arsenic compounds and the different roles of glutathione in such induction. Mutat. Res. 357, 123–129.

    Pi, J., Qu, W., Reece, J. M., Kumagai, Y., and Waalkes, M. P. (2003). Transcription factor Nrf2 activation by inorganic arsenic in cultured keratinocytes: involvement of hydrogen peroxide. Exp. Cell Res. 290, 234–245.

    Puga, A., Barnes, S. J., Chang, C., Zhu, H., Nephew, K. P., Khan, S. A., and Shertzer, H. G. (2000). Activation of transcription factors activator protein-1 and nuclear factor-kappaB by 2,3,7,8-tetrachlorodibenzo-p-dioxin. Biochem. Pharmacol. 59, 997–1005.

    Pulido, M. D., and Parrish, A. R. (2003). Metal-induced apoptosis: Mechanisms. Mutat. Res. 533, 227–241.

    Rea, M. A., Gregg, J. P., Qin, Q., Phillips, M. A., and Rice, R. H. (2003). Global alteration of gene expression in human keratinocytes by inorganic arsenic. Carcinogenesis 24, 747–756.

    Reiner, A., Yekutieli, D., and Benjamini, Y. (2003). Identifying differentially expressed genes using false discovery rate controlling procedures. Bioinformatics 19, 368–375.

    Rogakou, E. P., Nieves-Neira, W., Boon, C., Pommier, Y., and Bonner, W. M. (2000). Initiation of DNA fragmentation during apoptosis induces phosphorylation of H2AX histone at serine 139. J. Biol. Chem. 275, 9390–9395.

    Rossman, T. G. (2003). Mechanism of arsenic carcinogenesis: An integrated approach. Mutat. Res. 533, 37–65.

    Roussel, R. R., and Barchowsky, A. (2000). Arsenic inhibits NF-kappaB–mediated gene transcription by blocking IkappaB kinase activity and IkappaBalpha phosphorylation and degradation. Arch. Biochem. Biophys. 377, 204–212.

    Sakurai, T., Kojima, C., Ochiai, M., Ohta, T., Sakurai, M. H., Waalkes, M. P., and Fujiwara, K. (2004a). Cellular glutathione prevents cytolethality of monomethylarsonic acid. Toxicol. Appl. Pharmacol. 195, 129–141.

    Sakurai, T., Ochiai, M., Kojima, C., Ohta, T., Sakurai, M. H., Takada, N. O., Qu, W., Waalkes, M. P., and Fujiwara, K. (2004b). Role of glutathione in dimethylarsinic acid–induced apoptosis. Toxicol. Appl. Pharmacol. 198, 354–365.

    Sakurai, T., Qu, W., Sakurai, M. H., and Waalkes, M. P. (2002). A major human arsenic metabolite, dimethylarsinic acid, requires reduced glutathione to induce apoptosis. Chem. Res. Toxicol. 15, 629–637.

    Senft, A. P., Dalton, T. P., and Shertzer, H. G. (2000). Determining glutathione and glutathione disulfide using the fluorescence probe o-phthalaldehyde. Anal. Biochem. 280, 80–86.

    Shen, Z. X., Chen, G. Q., Ni, J. H., Li, X. S., Xiong, S. M., Qiu, Q. Y., Zhu, J., Tang, W., Sun, G. L., Yang, K. Q. et al. (1997). Use of arsenic trioxide (As2O3) in the treatment of acute promyelocytic leukemia (APL): II. Clinical efficacy and pharmacokinetics in relapsed patients. Blood 89, 3354–3360.

    Shi, H., Hudson, L. G., Ding, W., Wang, S., Cooper, K. L., Liu, S., Chen, Y., Shi, X., and Liu, K. J. (2004). Arsenite causes DNA damage in keratinocytes via generation of hydroxyl radicals. Chem. Res. Toxicol. 17, 871–878.

    States, J. C., Reiners, J. J., Jr., Pounds, J. G., Kaplan, D. J., Beauerle, B. D., McNeely, S. C., Mathieu, P., and McCabe, M. J., Jr. (2002). Arsenite disrupts mitosis and induces apoptosis in SV40-transformed human skin fibroblasts. Toxicol. Appl. Pharmacol. 180, 83–91.

    Styblo, M., Serves, S. V., Cullen, W. R., and Thomas, D. J. (1997). Comparative inhibition of yeast glutathione reductase by arsenicals and arsenothiols. Chem. Res. Toxicol. 10, 27–33.

    Tee, A. R., and Blenis, J. (2005). mTOR, translational control and human disease. Semin. Cell Dev. Biol. 16, 29–37.

    Tinwell, H., Stephens, S. C., and Ashby, J. (1991). Arsenite as the probable active species in the human carcinogenicity of arsenic: Mouse micronucleus assays on Na and K arsenite, orpiment, and Fowler's solution. Environ. Health Perspect. 95, 205–210.

    Wang, Z., and Rossman, T. G. (1996). The carcinogenicity of arsenic. In Toxicology of Metals, pp. 221–229. CRC Press, Boca Raton, FL.

    Wolfinger, R. D., Gibson, G., Wolfinger, E. D., Bennett, L., Hamadeh, H., Bushel, P., Afshari, C., and Paules, R. S. (2001). Assessing gene significance from cDNA microarray expression data via mixed models. J. Comput. Biol. 8, 625–637.

    Yang, Y., Dieter, M. Z., Chen, Y., Shertzer, H. G., Nebert, D. W., and Dalton, T. P. (2002). Initial characterization of the glutamate-cysteine ligase modifier subunit Gclm(-/-) knockout mouse. Novel model system for a severely compromised oxidative stress response. J. Biol. Chem. 277, 49446–49452.(Simone Kann, Cameron Este)