Environ. Sci. Technol. 2007, 41, 3769-3774
Classification of Heavy-Metal Toxicity by Human DNA Microarray Analysis KOJI KAWATA, HIROYUKI YOKOO, RYUHEI SHIMAZAKI, AND SATOSHI OKABE* Department of Urban and Environmental Engineering, Graduate School of Engineering, Hokkaido University, North-13, West-8, Kita-ku, Sapporo 060-8628, Japan
Microarray technology is proving to be a useful tool to classify undefined environmental toxicants, to investigate underlying mechanisms of toxicity, and to identify candidate toxicant-specific genetic markers by examining global effects of putative toxicants on gene expression profiles. The aim of this study was to evaluate the toxicities of six heavy metals through the comparison with gene expression patterns induced by well-known chemicals. For this purpose, we first identified the genes altered specifically in HepG2 under the exposure of 2,3-dimethoxy-1,4naphthoquinone (DMNQ), phenol, and N-nitrosodimethylamine (DMN), which were selected as the model chemicals, using DNA microarray. On the basis of the expression profiles of these genes, toxicities of six heavy metals, arsenic, cadmium, nickel, antimony, mercury, and chromium, were evaluated. The specific gene alteration and hierarchical clustering revealed that biological action of six heavy metals was clearly related to that of DMNQ which has been reported to be a reactive oxygen species (ROS) generating chemical and which induced the genes associated with cell proliferative responses. These results suggest that cell proliferative responses which are probably caused by ROS are a major apparent biological action of high-dose heavy metals, supporting the previous reports. Overall, a mechanism-based classification by DNA microarray would be an efficient method for evaluation of toxicities of environmental samples. DNA microarray is a powerful technology that enables the examination of the expression of thousands of genes simultaneously and has a major impact on many different areas such as pharmacology and oncology. In particular, there has been interest in using arrays in toxicology to discriminate and classify toxicants on the basis of unique gene expression profiles induced by putative toxic actions (1-4). Recently, this concept has been applied to evaluate the putative toxicity of environmental pollutants, and some chemical-specific gene expression patterns in animal tissues, cultured cells, and yeast Saccharomyces cerevisiae have been reported (58). Until now, several databases of gene expression patterns induced by chemicals have been established (9). However, for evaluation of potential hazards of environmental samples including a variety of chemicals, it is necessary to accumulate gene expression data of a large number of chemical species. * Corresponding author phone: +81-11-706-6266; fax: +81-11707-6266; e-mail:
[email protected]. 10.1021/es062717d CCC: $37.00 Published on Web 04/17/2007
2007 American Chemical Society
Hamadeh et al. have demonstrated that structurally unrelated compounds with similar toxic mechanism produced similar gene expression profiles in cultured cells or animal tissues (4). This result suggests that it would be possible to define a toxicity of samples as gene expression pattern similarity to model chemicals whose toxicities are defined well. We are focusing on the possibility of using human DNA microarray analysis for evaluating a toxicity of environmental samples on the basis of well-defined chemical properties. Among various environmental pollutants, mutagens and carcinogens can cause serious life-threatening conditions. In addition, toxic materials may cause metabolic damage, including oxidative stress. Therefore, we selected 2,3-dimethoxy-1,4-naphthoquinone (DMNQ) and phenol, which have been reported to generate free radicals (10, 11), and N-nitrosodimethylamine (DMN) known as carcinogen or mutagen, and gene expression patterns of these chemicals were analyzed and used as criteria for evaluation of apparent toxicity of samples. For the estimation of this method, we selected heavy metals as model pollutants. Toxic heavy metals such as arsenic, cadmium, and chromium are well-known pollutants which accumulate in groundwater and soil through release from industrial practices and even natural sources. Increasing evidence indicates that the toxicity of heavy metals has various potential mechanisms, such as oxidative stress, interference with essential metals, and interactions with cellular macromolecules (12, 13). Thus, the mechanisms that are responsible for heavy-metal toxicity are multifactorial. Hence, heavy metals would be suitable for evaluating a usefulness of DNA microarray analysis. In addition, this analysis would reveal a major biological action of heavy metals at a given exposure dose and time, which is a useful index to the presence of these chemicals in environmental samples. In this paper, we performed DNA microarray analyses for human hepatoma derived cell line HepG2 following exposure of three model chemicals mentioned above and six heavy metals; arsenic, cadmium, nickel, antimony, mercury, and chromium. Through the comparison of gene expression patterns induced by these chemicals, major toxicities of high dose and short-term exposure of six heavy metals were evaluated.
Materials and Methods Cell Culture and Treatments. In this study, we used human hepatoma HepG2 cells, which retain the activities of phase I and II enzymes and can induce activation and detoxification reactions (14). Furthermore, these cells exhibit the DNA damage indexes (micronuclei and sister chromatid exchange) following exposure of different classes of mutagens (15). The cells obtained from the Riken Cell Bank (Tsukuba, Japan) were cultured in Eagle’s minimal essential medium (MEM) (Nissui, Tokyo, Japan) supplemented with 1% nonessential amino acid (Invitrogen, Carlsbad, CA), 10% fetal bovine serum, and 60 mg/mL kanamycin at 37 °C and 5% CO2. Cells were grown to 70% confluency in 60 mm culture dishes and were exposed to three model chemicals, 10 µM 2,3dimethoxy-1,4-naphthoquinone (DMNQ), 5.4 mM N-nitrosodimethylamine (DMN), and 10 mM phenol, and to six heavy metals, 20 µM arsenic(III) oxide (As), 2 µM cadmium chloride (Cd), 6.5 mM nickel(II) chloride hexahydrate (Ni), 200 µM bis[(+)-tartrato]diantimonate(III) dipotassium trihydrate (Sb), 20 µM mercury(II) chloride (Hg), and 20 µM potassium dichromate (Cr), by adding each chemical compound to the culture medium. The chemical compounds VOL. 41, NO. 10, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3769
used in this study were purchased from Wako Pure Chemical Industries (Osaka, Japan), Sigma-Aldrich Japan (Tokyo, Japan), or Sigma (St. Louis, MO). To detect early biological responses, which would be useful indexes for rapid detection of the toxicities of chemicals, we selected 6 h exposure, and doses were chosen as maximum concentrations at which significant cytotoxicity was not observed (more than 80% cell viability was still determined by neutral red uptake) (16). Exceptionally, DMN does not exhibit apparent cytotoxicity in high concentrations, so that the cells were treated for 48 h by DMN. Following exposure, cells were washed with phosphate-buffered saline (PBS) and were immediately subjected to RNA extraction. Three independent cultures were used for each treatment or control group. Microarray Experiment. The cells were lysed directly in culture dishes and total RNA was extracted using RNeasy Mini Kit (Qiagen, Hilden, Germany). Target preparation and hybridization were performed according to One-Cycle Eukaryotic Target Labeling Assay protocols described in Affymetrix technical manual (Affymetrix, Santa Clara, CA). cDNA was synthesized from the total RNA by using OneCycle cDNA Synthesis kit (Invitrogen, Carlsbad, CA) with a T7-(dT)24 primer incorporating a T7 RNA polymerase promoter. cRNA was synthesized from the cDNA and biotinlabeled by in vitro transcription using IVT Labeling kit (Affymetrix, Santa Clara, CA). Labeled cRNA was fragmented by incubation at 94 °C for 35 min in the presence of 40 mM Tris acetate, pH 8.1, 100 mM potassium acetate, and 30 mM magnesium acetate. Ten micrograms of fragmented cRNA was hybridized to a Human Genome Focus array (Affymetrix, Santa Clara, CA) containing probes for 8795 human genes for 16 h at 45 °C. After hybridization, the microarrays were automatically washed and stained with streptavidin-phycoerythrin by using a fluidics station (Affymetrix, Santa Clara, CA). Finally, probe arrays were scanned with the Genechip System confocal scanner (Affymetrix, Santa Clara, CA). Microarray Data Analysis. Expression data stored as “CEL file” in the Gene Chip Operating Software (GCOS) (Affymetrix, Santa Clara, CA) were transferred into the Avadis 4.2 prophetic (Strand Genomics, Redwood City, CA). Signal intensity of probes was scaled and normalized by MAS5 algorism. These summarized data have been deposited to the National Center for Biotechnology Information (NBCI) Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo) and are accessible through GEO Series accession number GSE6907. From the results of detection call analyses, the genes with “Present calls” in all 30 samples (10 treatments, n ) 3) were selected and used in the subsequent steps. To identify differentially expressed genes, the unpaired t-tests for control and respective treatment groups (n ) 3) were performed for each gene. From the results of these analyses, the genes with p < 0.05 and g2.0 fold change in either direction were identified as being differentially expressed. Furthermore, the genes that differentially expressed among three model chemicals were selected and functionally classified on the basis of Gene Ontology categories by using web-based gene ontology program Fatigo (http:// fatigo.bioinfo.cipf.es). The hierarchical clustering (using the Euclidean distance metric) analysis and principal component analysis (PCA) were performed on these selected genes by using Avadis 4.2.
Results and Discussion Commonly Induced Genes among Three Model Chemicals. DMNQ, DMN, or phenol treatment altered the expression levels of 167, 328, or 160 genes, respectively. In these genes, 64 genes were altered in two or more chemical treatments (Supporting Information 1 and 2). Particularly noteworthy was the up-regulation of oxidative stress inducible genes: 3770
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 10, 2007
TXNRD1 coding thioredoxin reductase 1 (17) in three chemical exposures, HMOX1 coding heme oxygenase 1 (18) in DMNQ and phenol exposure, and the genes associated with glutathione biosynthesis (GCLC and GSR) (19) in DMN exposure. Three model chemicals used in this study have been reported to generate free radicals during the processes of metabolism (10, 11, 20) and cause oxidative stress. The gene expressions mentioned above are consistent with the previously reported biological actions of three model chemicals. Selection of Genes Altered Specifically by the Model Chemical Treatments. The aim of this study was to evaluate the toxicities of six heavy metals through the comparison of gene expression patterns in HepG2 following exposure of heavy metals and three model chemicals. For this purpose, we identified the genes altered specifically in HepG2 under the exposure of DMNQ, DMN, or phenol. To identify genes that would discriminate among the biological actions of DMNQ, phenol, and DMN, the genes overlapping among these chemical treatments were excluded. This resulted in a total of 481 genes consisting of 103 DMNQ specific, 241 DMN specific, and 137 phenol specific genes, which were used for evaluation of the heavy-metal toxicities. Genes Specifically Altered in DMNQ Treatment. Of the 103 genes selected to discriminate DMNQ from DMN and phenol, 51 were up-regulated and 52 were down-regulated. For further analyses, we functionally classified these genes on the basis of gene ontology (GO) terms. As shown in Figure 1, 37 up-regulated genes could be annotated and 21 biological processes containing at least three gene hits were found. Particularly, an important finding was that DMNQ treatment induced the genes classified in “M phase”, “mitotic cell cycle”, “regulation of cell cycle”, and “regulation of cell proliferation”. These classes contained eight genes (UBE2C, CDC25B, CDKN3, KIF22, H2AFX, SMC4L1, RAE1, and CCNB2), most of which have been reported to be up-regulated in association with acceleration of cell division and proliferation (21-25). Of these genes, CCNB2, UBE2C, SMC4L1, and CDKN3 exhibited remarkably high fold increases (Supporting Information 3). On the other hand, functional classification of the annotatable 40 genes repressed by DMNQ exposure revealed 30 biological processes including “immune response”, “lipid metabolism”, “amino acid and derivative metabolism”, and “programmed cell death” (Figure 1). These processes contained a total of eight genes (CLU, AZGP1, BF, TNFSF10, PCYOX1, BCAT1, ALDH6A1, PSPH, ITIH3, and ST6GACNAC4), all of which exhibited the higher fold transcription decreases (Supporting Information 3). Quinones including DMNQ generate reactive oxygen species (ROS) such as hydrogen peroxide, the hydroxyl radical, and singlet oxygen via the redox cycle (10). These ROS cause lipid peroxidation, protein oxidation, and DNA damage, which lead to cell injury, whereas they stimulate the expression of early growth related genes such as c-fos and c-jun in mammalian cells and might have a function as mitogenic stimuli through biochemical processes common to natural growth factors (26). The expressions of genes associated with cell proliferation and the repressions of genes functionally classified in programmed cell death were presumably consequent on signaling function of ROS promoting cell growth responses. Genes Specifically Altered in DMN Treatment. Of the 241 genes specifically altered in DMN exposure, 114 were up-regulated and 127 were down-regulated. As shown in Figure 1, these genes were classified into a variety of biological processes by functional classification. Annotatable 85 upregulated genes were classified into 43 biological processes. Of these processes, programmed cell death and regulation of cell cycle containing 12 genes (PHLDA2, DFFA, BCL10,
FIGURE 1. Comparison of the gene ontology (GO) categories among the genes specifically altered by DMNQ (103 genes), phenol (137 genes), and DMN (241 genes). GO terms of biological processes were represented at level 5 using the web-based gene ontology program Fatigo (http://fatigo.bioinfo.cipf.es). CIDEC, TGFB1, CDKN1A, NCKAP1, FAF1, YARS, BUB3, CHES1, and SHC1), which have been reported to associate with apoptosis or cell growth arrest (27-35), could be considered to reflect the DNA damaging action of DMN. In these genes, CDKN1A (p21/Waf1), CIDEC, and PHLDA2 were remarkably highly up-regulated (Supporting Information 4). Particularly, CDKN1A has been shown previously to be up-regulated in rat liver treated by DNA damaging agent such as diethylnitrosamine and related to growth arrest following DNA damage (36, 37). It has been reported that the DMN treatment increased the frequencies of appearance of micronuclei and sister chromatid exchange in HepG2 cells and delayed cell cycle (15). Our results are consistent with this report and
suggest that the pathways associated with cell growth arrest caused by DNA damage would be activated in the cells by DMN exposure. Of the down-regulated 127 genes, 91 were annotatable and could be classified into 40 biological processes. From these results, we found that the genes involved in lipid metabolism including cellular lipid metabolism and lipid biosynthesis were distinctively down-regulated by DMN exposure (Figure 1). These processes contained 20 genes, of which 15 are associated with steroid and cholesterol biosynthesis. Moreover, 10 of these genes (EBP, LSS, HSD17B2, FDPS, HMGCR, SQLE, FDFT1, DHCR7, HSD17B7, and HMGCS1) exhibited appreciably high transcription level VOL. 41, NO. 10, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3771
TABLE 1. Number of Specific Genes Altered by Six Heavy Metalsa no. of seleted genes
As
Cd
Cr
Ni
Hg
Sb
DMNQ induction: 51 repression: 52 total: 103
33 18 51
7 9 16
23 12 35
11 9 20
33 20 53
24 24 48
Phenol induction: 63 repression: 74 total: 137
1 0 1
0 1 1
4 10 14
9 10 19
2 1 3
13 21 34
DMN induction: 114 repression: 127 total: 241
4 2 6
5 3 8
9 12 21
16 17 33
9 16 25
9 32 41
a Number of genes with g2-fold increase or decrease, p < 0.05 is shown.
decreases appearing on the list of the 25 genes most strongly repressed (Supporting Information 4). Genes Specifically Altered in Phenol Treatment. Of the 137 genes specifically altered in phenol exposure, 63 were up-regulated and 74 were down-regulated. Functional classification of annotatable genes (up-regulated: 43 and downregulated: 55) revealed 14 and 22 biological processes containing at least three gene hits. As shown in Figure 1, the genes involved in regulation of cellular metabolism (19 upregulated and 12 down-regulated) were distinctively altered in phenol exposure. This ontology term includes any processes that modulate the frequency, rate, or extent of the chemical reactions and pathways by which individual cells transform chemical substances. In the genes highly altered, up-regulated 10 genes (ING3, KLF10, MYC, ELL2, PER2, PPARG, TCFL5, HES1, NFYA, and TERF2) and down-regulated 3 genes (TAF7, RFP, and NR0B2) participated in this process (Supporting Information 5). These genes are associated with various processes such as cell proliferation, apoptosis, transcription, abiotic stimulus, and lipid metabolism. Furthermore, the genes associated with cell growth arrest or induction of apoptosis (ING3, TNFRSF21, PPM1D, and FAS) (38-41) were highly up-regulated (Supporting Information 5). These results suggest that phenol exposure may affect various cellular functions and particularly may induce cell growth inhibition and apoptosis strongly. Expression Level Changes of Model Chemical Inducible Genes in Six Heavy Metal Exposures. On the basis of the expression profiles of the genes altered by three model chemical exposures, the toxicities of As, Cd, Cr, Ni, Hg, and Sb were evaluated. The genes associated with oxidative stress responses that were up-regulated in model chemical exposures were induced also in heavy-metal exposures. TXNRD1 up-regulated in three chemical exposures exhibited 1.7 (p ) 0.098), 2.7 (p ) 0.0007), 1.9 (p ) 0.0124), 2.8 (p ) 0.0013), 3.4 (p ) 0.0009), and 3.2 (p ) 0.0005) fold change in As, Cd, Cr, Ni, Hg, and Sb, respectively. In addition, HMOX1 induced by DMNQ and phenol was up-regulated in heavy-metal exposures except for Cr with significantly high-fold changes [As: 15 fold (p ) 0.0021), Cd: 15 (p ) 0.0021), Ni: 6.3 (p ) 0.008), Hg: 19 (p ) 0.0014), and Sb: 24 (p ) 0.0011)]. Oxidative stress would be a common biological action in three model chemicals and high-dose heavy metals. Furthermore, six heavy-metal toxicities were evaluated on the basis of expression alterations of 481 genes specific to three model chemical exposures. In Table 1, we summarized the expression of model chemical-specific genes in six heavy-metal exposures with the alteration threshold at 2-fold and p < 0.05. Of 103 DMNQ specific genes, 16-53 3772
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 10, 2007
(16-51%) showed expression level change. In particular, As, Cr, Hg, and Sb up-regulated a large number (45-66%) of DMNQ inducible genes. The genes most strongly induced or repressed by DMNQ were altered by most of the heavy-metal exposures in the same direction (Supporting Information 3). Evidently, induction of CCNB2, MCM2, UBE2C, ZWINT, HIST1H4C, RNASEH2A, and ITGB3BP (1.7-6.3 folds, p < 0.0001-0.0465) and repression of CLGN, OS-9, SERPINC1, and TNFSF10 (1.4-6.3 folds, p ) 0.0001-0.0069) were observed in five or more heavy-metal treatments. In these, expression level increases of CCNB2 and UBE2C are related to cell proliferation as mentioned earlier and suggest that heavy metals might stimulate growth-related responses as DMNQ does. On the other hand, limited numbers of DMN and phenolspecific genes were affected by heavy metals except for Ni and Sb. In As, Cd, Cr, and Hg exposures, only 6-21 (2-10%) of DMN specific genes were altered, while Ni and Sb modified the expression of 33 and 41 (13 and 17%) genes in the specific gene set (Table 1). These results suggest that these heavy metals share certain characteristics with DMN. However, only a few of the genes that were considered to clearly reflect the DNA damaging action of DMN were induced in these two heavy metals. The only exception was that a DNA damage responsive gene (CDKN1A) was up-regulated (2.4-6.1 folds, p ) 0.0019-0.0288) in Cr, Ni, and Sb exposures (Supporting Information 4). Similarly, 19 and 34 (14 and 25%) of phenol-specific genes associated with various biological functions were altered by more than 2-fold in Ni and Sb exposures, respectively (Table 1). In particular, ATF3, which has been reported to respond to a variety of stress signals (42), exhibited a remarkably higher (20 and 21 folds, p < 0.0001 and p ) 0.0002) transcriptional level increase in both heavy metals. Evaluation of Heavy-Metal Toxicity on the Basis of Expression Patterns of Selected Genes. To visualize the differences in gene expression patterns of all chemical treatments, hierarchical clustering analysis was performed on the selected 481 genes. The result is presented as a tree dendrogram (Figure 2). In this tree, gene expression patterns of six heavy metals formed a cluster with DMNQ and were distinct from DMN and phenol branches. Furthermore, gene expression patterns of As, Cd, Cr, and Hg were recognized as being more closely related to each other, forming a distinct subcluster from Ni and Sb branches. In addition, the validity of this cluster analysis was verified utilizing principal component analysis (PCA). On the basis of this technique, each treatment sample was classified into a smaller dimensional space than represented by the original 481 dimensions. The two-dimensional scatter plot on the basis of the first two principal components is shown in Figure 3. The gene expression patterns of As, Cd, Cr, and Hg were closely located with DMNQ while Ni and Sb were slightly distinct from the other four metals, supporting the result of hierarchical clustering analysis (Figure 2). Thus, overlapping gene alteration and similarity of those patterns indicated that biological actions of six heavy metals were closely related with that of DMNQ. As mentioned previously, DMNQ has been reported to generate ROS. Coincidentally, recent studies indicated that transition metals caused an increase in production of ROS, undergoing redox cycle or depleting antioxidants (12). Our results strongly support this speculation and suggest that ROS induction and activation of cell proliferation are common biological actions to six heavy metals tested in this study. Besides, carcinogenesis is also a well-known heavy-metal toxicity. However, we could not find certain relationships between the gene expression profiles of DMN and heavy metals in our experimental conditions. Expression of the genes associated with carcinogenesis might be dependent
FIGURE 2. Hierarchical clustering analysis of three model chemicals and six heavy-metal exposures on the basis of 481 selected genes. Each row represents one single gene, and each column represents one chemical treatment. Red spectrum colors indicate up-regulation, while green spectrum colors indicate down-regulation. oxidative stress is well-known, however, the use of microarrays would be expected to provide more detailed information on these mechanisms. Our method would be applicable to evaluate more various environmental toxicants and to give an insight into development of toxicogenomics-based bioassay for evaluation of toxicities of environmental samples.
Acknowledgments This research was carried out as a part of the 21st Century COE Program “Sustainable Metabolic System of Water and Waste for Area-Based Society”. This study was also supported partially by a grant-in-aid (no. 17360250 and 17780241) for Developmental Scientific Research from the Ministry of Education, Science, and Culture of Japan.
Supporting Information Available Additional details shown in one figure and four tables. This material is available free of charge via the Internet at http:// pubs.acs.org.
FIGURE 3. Principle component analysis (PCA) based on the expression of selected genes. The variation in expression levels of 481 genes is reduced to two-dimensional space. Variance of component 1 and 2 was 32.1 and 25.4%, respectively. upon the exposure time and dose. It has been reported that cadmium chloride, benzo[a]pyrene, and trichloroethylene produced different patterns of gene expression in the livers of exposed mice (5). Likewise, cadmium chloride, sodium dichromate, and nickel subsulfide altered only a few genes that overlapped with DNA-cross-linker mitomycin C in human lung cells (43). Carcinogenic action of heavy metals might be multifactorial and might not be characterized on the basis of only the genes specifically altered by genotoxic carcinogens such as DNA-alkylating and -cross-linking agents. In conclusion, oxidative stress characterized by induction of TXNRD1 and HMOX1 could be detected in three model chemicals and six heavy metals in the present study. Furthermore, the specific gene alteration and hierarchical clustering based on the selected 481 genes, which might be specific to exposure of DMNQ, DMN, and phenol, revealed that biological actions of six heavy metals were clearly related to those of DMNQ and were distinguishable from the other model chemicals. These results suggest that stimulation of cell proliferative responses that are probably caused by ROS is a major apparent biological action of high-dose heavy metals and might be a useful fingerprint for rapid detection of these heavy-metal pollutants. The fact that metals cause
Literature Cited (1) Rockett, J. C.; Dix, D. J. Application of DNA arrays to toxicology. Environ. Health Perspect. 1999, 107, 681-685. (2) Nuwaysir, E. F.; Bittner, M.; Trent, J.; Barrett, J. C.; Afshari, C. A. Microarrays and toxicology: the advent of toxicogenomics. Mol. Carcinog. 1999, 24, 153-159. (3) Lovett, R. A. Toxicogenomocs. Toxicologists brace for genomics revolution. Science 2000, 289, 536-537. (4) Hamadeh, H. K.; Bushel, P. R.; Jayadev, S.; Martin, K.; DiSorbo, O.; Siber, S.; Bennett, L.; Tennant, R.; Stoll, R.; Barrett, J. C.; Blanchard, K.; Paules, R. S.; Afshari, C. A. Gene expression analysis reveals chemical-specific profiles. Toxicol. Sci. 2002, 67. 219-231. (5) Bartosiewicz, M.; Penn, S.; Buckpitt, A. Applications of gene arrays in environmental toxicology: fingerprints of gene regulation associated with cadmium chloride, benzo(a)pyrene, and trichloroethylene. Environ. Health Perspect. 2001, 109, 7174. (6) Puga, A.; Maier, A.; Medvedovic, M. The transcriptional signature of dioxin in human hepatoma HepG2 cells. Biochem. Pharmacol. 2000, 60, 1129-1142. (7) Momose, Y.; Iwahashi, H. Bioassay of cadmium using a DNA microarray: genome-wide expression patterns of Saccharomyces cerevisiae response to cadmium. Environ. Toxicol. Chem. 2001, 20, 2353-2360. (8) Kitagawa, E.; Momose, Y.; Iwahashi, H. Correlation of the structures of agricultural fungicides to gene expression in Saccharomyces cerevisiae upon exposure to toxic doses. Environ. Sci. Technol. 2003, 37, 2788-2793. (9) Boverhof, D. R.; Zacharewski, T. R. Toxicogenomics in risk assessment: applications and need. Toxicol. Sci. 2005, 89, 352360. (10) Kappus, H.; Sies, H. Toxic drug effects associated with oxygen matabolism: redox cycling and lipid peroxidation. Experientia 1981, 37, 1233-1241. VOL. 41, NO. 10, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3773
(11) Shvedova, A. A.; Kommineni, C.; Jeffries, B. A.; Castranova, V.; Tyurina, Y. Y.; Tyurin, V. A.; Serbinova, E. A.; Fabisiak, J. P.; Kagan, V. E. Redox cycling of phenol induces oxidative stress in human epidermal keratinocytes. J. Invest. Dermatol. 2000, 114, 354-364. (12) Snow, E. T. Metal carcinogenesis: mechanistic implications. Pharmacol. Ther. 1992, 53, 31-65. (13) Ercal, N.; Gurer-Orhan, H.; Aykin-Burns, N. Toxic metals and oxidative stress part I: mechanisms involved in metal induced oxidative damage. Curr. Top. Med. Chem. 2001, 1, 529-539. (14) Darroudi, F.; Natarajan, A. T. Metabolic activation of chemicals to mutagenic carcinogens by human hepatoma microsomal extracts in Chinese hamster ovary cells (in vitro). Mutagenesis 1993, 8, 11-15. (15) Natarajan, A. T.; Darroudi, F. Use of human hepatoma cells for in vitro metabolic activation of chemical mutagens/carcinogens. Mutagenesis 1991, 6, 399-403. (16) Borenfreund, E.; Puerner, J. A. A simple quantitative procedure using monolayer cultures for cytotoxicity assays (HTD/NR-90). Methods Cell Sci. 1985, 9, 7-9. (17) Powis, G.; Kirkpatrick, D. L.; Angulo, M.; Baker, A. Thioredoxin redox control of cell growth and death and the effects of inhibitors. Chem. Biol. Interact. 1998, 112, 23-34. (18) Ryter, S. W.; Tyrrell, R. M. The heme synthesis and degradation pathways: role in oxidant sensitivity. Heme oxygenase has both pro- and antioxidant properties. Free Radical Biol. Med. 2000, 28, 289-309. (19) Krzywanski, D. M.; Dickinson, D. A.; Iles, K. E.; Wigley, A. F.; Franklin, C. C.; Liu, R. M.; Kavanagh, T. J.; Forman, H. J. Variable regulation of glutamate cysteine ligase subunit proteins affects glutathione biosynthesis in response to oxidative stress. Arch. Biochem. Biophys. 2004, 423, 116-125. (20) Floyd, R. A.; Soong, L. M.; Stuart, M. A,; Reigh, D. L. Spin trapping of free radicals produced from nitrosoamine carcinogens. Photochem. Photobiol. 1978, 28, 857-862. (21) Townsley, F. M.; Aristarkhov, A.; Beck, S.; Hershko, A.; Ruderman, J. V. Dominant-negative cyclin-selective ubiquitin carrier protein E2-C/UbcH10 blocks cells in metaphase. Proc. Natl. Acad. Sci. 1997, 94, 2362-2367. (22) Karlsson, C.; Katich, S.; Hagting, A.; Hoffmann, I.; Pines, J. Cdc25B and Cdc25C differ markedly in their properties as initiators of mitosis. J. Cell Biol. 1999, 146, 573-584. (23) Luch, A. Cell cycle control and cell division: implications for Chemically induced carcinogenesis. ChemBioChem 2002, 3, 506-516. (24) Hirokawa, N.; Takemura, R. Kinesin superfamily proteins and their various functions and dynamics. Exp. Cell Res. 2004, 301, 50-59. (25) Schmiesing, J. A.; Ball, A. R., Jr.; Gregson, H. C.; Alderton, J. M.; Zhou, S.; Yokomori, K. Identification of two distinct human SMC protein complexes involved in mitotic chromosome dynamics. Proc. Natl. Acad. Sci. 1998, 95, 12906-12911. (26) Burdon, R. H. Superoxide and hydrogen peroxide in relation to mammalian cell proliferation. Free Radical Biol. Med. 1995, 18, 775-794. (27) Lee, M.; Feinberg, A. Genomic imprinting of a human apoptosis gene homologue, TSSC3. Cancer Res. 1998, 58, 1052-1056. (28) Nagata, S. Apoptoic DNA fragmentation. Exp. Cell Res. 2000, 256, 12-18.
3774
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 10, 2007
(29) Barkett, M.; Gilmore, T. D. Control of apoptosis by Rel/NF-B transcription factors. Oncogene 1999, 18, 6910-6924. (30) Markowitz, S. D.; Roberts, A. B. Tumor suppressor activity of the TGF-beta pathway in human cancers. Cytokine Growth Factor Rev. 1996, 7, 93-102. (31) Bunz, F.; Dutriaux, A.; Lengauer, C.; Waldman, T.; Zhou, S.; Brown, J. P.; Sedivy, J. M.; Kinzler, K. W.; Vogelstein, B. Requirement for p53 and p21 to sustain G2 arrest after DNA damage. Science 1998, 282, 1497-1501. (32) Suzuki, T.; Nishiyama, K.; Yamamoto, A.; Inazawa, J.; Iwaki, T.; Yamada, T.; Kanazawa, I.; Sakaki, Y. Molecular cloning of a novel apoptosis-related gene, human Nap1 (NCKAP1), and its possible relation to Alzheimer disease. Genomics 2000, 63, 246-254. (33) Chu, K.; Niu, X.; Williams, L. T. A Fas-Associated Protein Factor, FAF1, potentiates Fas-Mediated apoptosis. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 11894-11898. (34) Wakasugi, K.; Schimmel, P. Two distinct cytokines released from a human aminoacyl-tRNA synthetase. Science 1999, 284, 147151. (35) Pati, D.; Keller, C.; Groudine, M.; Plon, S. E. Reconstitution of a MEC1-independent checkpoint in yeast by expression of a novel humen fork head cDNA. Mol. Cell Biol. 1997, 17, 30373046. (36) Lee, Y. S.; Kim, W. H.; Yu, E. S.; Kim, M. R.; Lee, M. J.; Jang, J. J. Time course of cell cycle-related protein expression in diethylnitrosamine-initiated rat liver. J. Hepatol. 1998, 29, 464469. (37) Waring, J. F.; Jolly, R. A.; Ciurlionis, R.; Lum, P. Y.; Praestgaad, J. T.; Morfitt, D. C.; Buratto, B.; Roberts, C.; Schadt, E.; Ulrich, R. G. Clustering of hepatotoxins based on mechanism of toxicity using gene expression profiles. Toxicol. Appl. Pharmacol. 2001, 175, 28-42. (38) Nagashima, M.; Shiseki, M.; Pedeux, R. M.; Okamura, S.; Kitahama-Shiseki, N.; Miura, K.; Yokota, J.; Harris, C. C. Anovel PHD-finger motif protein, p47ING3, modulates p53-mediated transcription, cell cycle control, and apaptosis. Oncogene 2003, 22, 343-350. (39) Ashkenazi, A.; Dixit, V. M. Apoptosis control by death and decoy receptors. Curr. Opin. Cell Biol. 1999, 11, 255-260. (40) Lu, X.; Nannenga, B.; Donehower, L. A. PPM1D dephosphorylates Chk1 and p53 and abrogates cell cycle checkpoints. Genes Dev. 2005, 19, 1162-1174. (41) Fulda, S.; Los, M.; Friesen, C.; Debatin, K. M. Chemosensitivity of solid tumor cells in vitro is related to activation of the CD95 system. Int. J. Cancer 1998, 76, 105-114. (42) Hai, T.; Wolfgang, C. D.; Marsee, D. K.; Allen, A. E.; Sivaprasad, U. ATF3 and stress responses. Gene Expression 1999, 7, 321335. (43) Andrew, A. S.; Warren, A. J.; Barchowsky, A.; Temple, K. A.; Klei, L.; Soucy, N. V.; O’hara, K. A.; Hamilton, J. W. Genomic and proteomic profiling of responses to toxic metals in human lung cells. Environ. Health Perspect. 2003, 111, 825-838.
Received for review November 14, 2006. Revised manuscript received February 21, 2007. Accepted February 27, 2007. ES062717D
