Applications of Gene Arrays in Environmental Toxicology: Fingerprints of Gene Regulation Associated with Cadmium Chloride, Benzo(a)pyrene, and Trichloroethylene.

Toxicity testing of unknown chemicals currently uses a number of short-term bioassays. These tests are costly and time consuming, require large numbers of animals, and generally focus on a single end point. The recent development of DNA arrays provides a potential mechanism for increasing the efficiency of standard toxicity testing through genome-wide assessments of gene regulation. In this study, we used DNA arrays containing 148 genes for xenobiotic metabolizing enzymes, DNA repair enzymes, heat shock proteins, cytokines, and housekeeping genes to examine gene expression patterns in the liver in response to cadmium chloride, benzo(a)pyrene (BaP), and trichloroethylene (TCE). Dose-response studies were carried out in mice for each chemical; each produced a unique pattern of gene induction. As expected, Cd[Cl.sub.2] markedly up-regulated metallothionine I and II (5- to 10,000-fold at the highest doses) and several of the heat shock/stress response proteins and early response genes. In contrast, administration of BaP upregulated only Cyp1a1 and Cyp1a2 genes and produced no significant increases in any of the stress response genes or any of the DNA repair genes present on the array. Likewise, TCE-induced gene induction was highly selective; only Hsp 25 and 86 and Cyp2a were up-regulated at the highest dose tested. Microarray analysis with a highly focused set of genes is capable of discriminating between different classes of toxicants and has potential for differentiating highly noxious versus more subtle toxic agents. These data suggest that use of microarrays to evaluate the potential hazards of unknown chemicals or chemical mixtures must include multiple doses and time points to provide effective assessments of potential toxicity of these substances. Key words: benzo(a)pyrene, cadmium chloride, DNA arrays, gene expression, trichloroethylene. Environ Health Perspect 109:71-74 (2001). [Online 15 December 2000]

http://ehpnet1.niehs.nih.gov/docs/2001/109p71-74bartosiewicz /abstract.html

DNA microarrays are a powerful, high-throughput tool for monitoring the expression of thousands of genes simultaneously (1-3). The application and use of DNA arrays in drug discovery and development (4), in identifying changes in gene expression associated with various disease processes, and in screening populations for allelic variants (single nucleotide polymorphisms) (5-8) has received considerable attention. Recently, there has been interest in using arrays in toxicology to quickly classify toxicants based on characteristic expression profiles and to use these profiles as a means of identifying putative mechanisms of action (9-11), as well as for identifying genes that are under similar regulatory control. Thus, the expression profiles associated with a heavy metal, a polycyclic aromatic hydrocarbon, and an oxidant stressor are likely to differ considerably and could be useful not only in assessing the potential toxicity of new synthetic chemical entities but in monitoring animal populations in the environment for potential point sources of environmental contamination. Finally, the application of this technology to address the potential human health relevance of environmental disasters such as the recent tire fires in California (12) may provide a rapid means of assessing the bioavailability and potential toxicity of complex mixtures of chemicals released into the air and into groundwater. In this case, changes in gene regulation in tissues from sentinel animals placed at the site could form the basis for these assessments.

In this paper we describe the application of a simple toxicology chip to study gene expression in mice treated with one of three distinct environmental/occupational contaminants: cadmium chloride, benzo(a)pyrene (BaP), and trichloroethylene (TCE). These chemicals were selected to meet the following criteria: each chemical was in a different chemical class [i.e., heavy metal, polycyclic aromatic hydrocarbon (PAH), chlorinated hydrocarbon]; each has high environmental/ occupational relevance; each has potential to affect human health (13-16); and data sets for changes in the regulation of selected genes were available in the literature for comparison. Cd[Cl.sub.2] is primarily an industrial by-product that humans are exposed to in a variety of media (13). Contamination of the environment with PAHs such as BaP occurs from natural and anthropogenic sources (i.e., mining, coal burning, oil refineries, etc.), and TCE contamination is a result of its widespread use as an industrial solvent (14-16). A number of genes have been reported to be up-regulated in response to both CD[Cl.sub.2] (17) and BaP (18,19) (i.e., metallothionein, cytochrome P-450 1a1, Cyp1a2, etc.). Although changes in gene expression in response to TCE are not as well documented, the widespread contamination of soil and groundwater with this solvent support the need for a more global approach to examining changes in gene expression associated with both acute and more chronic exposures (16).

In our previous work (20) we focused on the validation of the use of microarrays in an in vivo model. These earlier studies used a wide range of doses of [Beta]-naphthoflavone, a known inducer of Cyp1a1 and Cyp1a2 genes in murine liver, to address issues of sensitivity and dynamic range of microarray analysis in comparison with accepted Northern blot assays as well as to investigate sources of variability in the assay (slide to slide, spot to spot, and animal to animal). The data derived from this work demonstrated that the dynamic range and sensitivity of DNA microarrays is comparable to Northern blotting analysis and that there is far more variability introduced in the data through interanimal differences than through variations related to spotting and hybridization.

The chip used for these studies contained 148 genes coding for both phase I and II metabolizing enzymes, DNA repair enzymes, stress proteins, and cytokines as well as housekeeping genes. Studies examining gene regulation in the liver in response to varying doses of Cd[Cl.sub.2], BaP, and TCE clearly indicate a unique fingerprint of gene-level alterations resulting from exposure to the different chemical classes. Although the arrays used in these studies were limited to a highly focused set of genes, this work provides proof of the concept that patterns of gene regulation assessed by array technology are sufficiently unique that they may be useful in whole-animal studies to classify unknowns (i.e., heavy metal, PAH, chlorinated hydrocarbon, alkylating carcinogen, etc.).

Materials and Methods

Animals. Male Swiss Webster mice (20-25 g) were obtained from Charles River Breeding Laboratories (Portage, MI) and were housed in animal care facilities at the University of California, Davis, which are accredited by the American Association of Laboratory Animal Care. Mice received food and water ad libitum and were housed in HEPA-filtered racks for at least 1 week before use.

Animal treatment and isolation of mRNA. Mice (four to five per group) were treated intraperitoneally (ip) with either Cd[Cl.sub.2] (Sigma, St. Louis, MO; dissolved in isotonic saline, 10 mL/kg), BaP (Sigma), or TCE (Sigma; dissolved in corn oil such that 10 mL was administered per kilogram body weight). Doses of each toxicant are indicated in Figure 1. Controls received either isotonic saline or corn oil only, as appropriate. We killed mice with an overdose of pentobarbital at 6 hr (Cd[Cl.sub.2] and TCE) or at 6 and 24 hr (BaP) and isolated hepatic RNA from individual animals as described previously (20). Briefly, RNA was isolated as described by Chomczynski and Sacchi (21) using Trizol (GIBCO BRL, Bethesda, MD). We used DNase-treated total RNA (500 [micro]g) to isolate mRNA using Oligotex dt resin (Qiagen, Valencia, CA) following the manufacturer's instructions.

Figure 1. Patterns of gene expression in livers of mice treated with
Cd[Cl.sub.2], BaP, or TCE. Up- or down-regulation of gene expression
is denoted with a color pattern (scale at top); the numbers refer to
the expression level as a percent control (control = 100%). All
colored boxes represent gene expression levels that were significantly
different from control. All values are the mean of four to five mice
at each dose and time point studied. Standard deviations are omitted
for clarity but were generally less than 15% of the mean value.

5,001-   2,501-   1,001-   501-    251-   126-   76-   51-   25-
10,000   5,000    2,500    1,000   500    250    125   75    50

                             Cd[Cl.sub.2](mg/kg)

Gene                  10      2.5    1.0   0.25   0.1   ctrl

                                  Stress genes

Metallothionein I    2,671   1,726   254    354   246    126

Metallothionein II   6,805   2,077   368    423   283    115

Hsp 105              2,276     210   120    119   117    107

Hsp 25               3,720     450    80     38    57     98

Hsp 86               1,329     493   113    105   112    119

GFIP                 1,089     105   141    116   103    101

chop 10                387     132    95     90    98     81

c-jun                  743      95    96     87    99     98

jun-b                  327     165   115     97   100     61

                           Xenobotic metabolizing enzymes

Cyp2F2                  24     142   116    140   149    122

CYP3A11                 30     125   152    131   161     99

Methyl TX 111           34      94   109    104   110     96

AcetylTx               424     240   116    107   133    117

Cyp1a2

Cyp1a1

Cyp2a

                                BaP(mg/kg)

Gene                  100    10     1    0.1   ctrl

                               Stress genes

Metallothionein I

Metallothionein II

Hsp 105

Hsp 25

Hsp 86

GFIP

chop 10

c-jun

jun-b

                     Xenobotic metabolizing enzymes

Cyp2F2

CYP3A11

Methyl TX 111

AcetylTx

Cyp1a2                 438   222   102    95    105

Cyp1a1               3,596   765   123    90    122

Cyp2a

                            TCE(mg/kg)

Gene                 1000   100   10    ctrl

                          Stress genes

Metallothionein I

Metallothionein II

Hsp 105

Hsp 25                232    79    76     98

Hsp 86                237   104   109    103

GFIP

chop 10

c-jun

jun-b

                      Xenobotic metabolizing
                              enzymes

Cyp2F2

CYP3A11

Methyl TX 111

AcetylTx

Cyp1a2

Cyp1a1

Cyp2a                 281   150    41    108

Preparation of DNA chips. Chips containing 148 unique genes involved in phase I and II metabolism, heat shock, DNA repair, inflammation, transcription, and housekeeping were used in the current studies. The methods for preparation and analysis of these, along with applications to whole animal studies, have been presented in a previous publication (20) and are described only briefly here. We purchased cDNAs, corresponding to the 3' region of the respective genes, from the Image Consortium cDNA mouse libraries through suppliers (Research Genetics, Huntsville, AL, or American Type Culture Collection, Bethesda, MD). NCBI BLAST (Basic Local Alignment Tool) was used to determine which fragments, corresponding to desired genes, contained the least homology to other known genes. For the fragments that contained areas of high homology (i.e., Cyp1a1 and Cyp1a2), polymerase chain reaction (PCR) primers were designed to amplify unique regions. A complete list of genes used in these studies is available on our web site (22).

Sequence-verified PCR products, prepared as described in detail earlier (20), varied in size from approximately 500 bp to 1,200 bp. These were purified according to the manufacturer's protocol by using Qiaquick columns (Qiagen), and the concentrations were determined by absorbance at 260 nm.

Spotting cDNA arrays. We mixed each purified PCR product (10 [micro]L at concentrations [is greater than] 200 ng/[micro]L) with 10 [micro]L of 8 M NaSCN in 96-well plates and spotted the DNA using the Molecular Dynamics Microarray Spotter Gen II (Molecular Dynamics, Sunnyvale, CA), onto slides treated by vapor deposition of 3-aminopropyltrimethoxysilane (23). Each gene was spotted in eight separate quadrants of the slide. The slides were baked at 80 [degrees] C for 2 hr and stored at room temperature in a desiccator until use. Before hybridization, the slides were washed in isopropanol for 10 min and then boiled in water for 5 min.

Preparation of cDNA probes labeled with Cy3 or Cy5. Cy3- and Cy5-labeled probes were prepared using 1 [micro]g mRNA from control or treated animals, respectively. Synthesis of the labeled first strand was conducted using Superscript II (GIBCO), and template RNA was removed by incubation with RNase as described earlier (20). We purified single-stranded cDNA probes using a PCR purification kit (Qiagen) and determined the absorbance of each probe preparation at 260 nm and either 550 nm (Cy3) or 650 nm (Cy5). The total dye content (picomoles), amount of probe (nanograms), and specific activity [(number of Cy molecules incorporated)/(number of bases)] were calculated. Probe mixtures were evaporated in a vacuum centrifuge.

Hybridization. Hybridization solution was filtered through a 0.1-[micro]m filter. Poly A (Amersham Life Sciences) and mouse Cot DNA (GIBCO) were added to the hybridization solution at a final concentration of 16 ng/[micro]L and 40 ng/[micro]L, respectively. Probes were resuspended in hybridization solution containing 50% formamide, 5x standard saline citrate (SSC), and 0.1% SDS. Cy3 (control) and Cy5 (treated) probes were matched based on total dye content (picomoles) and were then mixed. The probe mixture was placed on a previously prepared array and the slides were hybridized overnight at 42 [degrees] C in a humid chamber. After hybridization, slides were placed in a wash solution (2x SSC, 0.1% SDS) for 5 min at 37 [degrees] C with gentle shaking. Cover slips floated off the slides during this initial wash. Slides were then washed once in 0.1x SSC, 0.1% SDS at room temperature for 5 min and twice in 0.1x SSC at room temperature for a total time of 5 min. The slides were then rinsed briefly in water and dried with a gentle stream of nitrogen.

Analysis of fluorescence spots. We scanned slides, prepared as described above, on the Molecular Dynamics MicroArray Scanner (23). Both Cy3 and Cy5 channels were scanned at a photomultiplier tube setting of 750 v. We analyzed the data sets using ArrayVision Software (Imaging Research, St. Catherines, Ontario, Canada). We calculated the fluorescence intensity of each spot using local median background subtraction. The relative fluorescent units (RFUs) were then normalized to the median signal of probe (Cy3 and Cy5) for that slide. The change in gene expression for each spot was calculated as [RFU.sub.treated]/[RFU.sub.control]. This value was then compared to its duplicate--that is, [RFU.sub.treated]/[RFU.sub.control] for spot 1 (left side of the slide) versus spot 5 (right side of the slide). If the ratio of the value for these two spots was between 0.67 and 1.5, the spot and duplicate were accepted, and an average value of induction was calculated. If the value was not within 50% of its replicate, we rejected the spot from the set. This criterion tended to exclude those spots that were contaminated with dust or those slides that had variable and high backgrounds. We rejected 10-20% of the values obtained during this work using these criteria. The data on those genes up- or down-regulated by chemical treatment are presented as the percentage of the Cy-3 controls. Significant differences between control and treatment doses were determined using a two-tailed t-test (SigmaStat, SPSS, Chicago, IL). In those cases where data used in the t-test were not normally distributed, a Mann-Whitney Rank Sum test was performed. We considered p-values [is less than] 0.05 statistically significant.

Results

Comparison of genes up-regulated by chemicals. Figure 1 presents a visual comparison for all genes that were influenced significantly at any dose of Cd[Cl.sub.2], BaP, or TCE studied. Even with the highly focused set of genes used in these studies, there are clear differences in the pattern of genes up- or down-regulated by the three chemicals.

Changes in gene expression in the liver by Cd[Cl.sub.2]. Ten genes were significantly induced over controls by ip administration of the highest dose of Cd[Cl.sub.2] tested (10 mg/kg; Figure 1). Induction of metallothionein-I was noted at doses of Cd[Cl.sub.2] as low as 0.25 mg/kg; at 1 mg/kg and higher metallothionein-II was up-regulated significantly. Up-regulation of heat shock proteins (Hsp 105 and 86) occurred at 2.5 mg/kg, and Hsp 25 was elevated significantly only at 10 mg/kg. Several immediate early genes were induced as well. c-jun showed significant induction over the control at 1 mg/kg Cd[Cl.sub.2], whereas increases in jun-b were significant at the 2.5 mg/kg dose of CD[Cl.sub.2]. Other hepatic genes that were induced by administration of Cd[Cl.sub.2] included growth-factor-induced protein, chop 10 (gadd 153), and acetyltransferase 96. In addition to gene up-regulation noted with Cd[Cl.sub.2] treatment, the expression levels of a number of genes were repressed upon treatment with Cd[Cl.sub.2]. Methyltransferase 111, Cyp2f2, and Cyp3a11 were all repressed significantly at the highest dose of Cd[Cl.sub.2] (10 mg/kg). Repression was not observed at lower doses.

Induction of genes by BaP. BaP was administered ip to mice at doses of 0.1-100 mg/kg. Animals treated with 100 mg/kg were killed at either 6 or 24 hr. Two genes, Cyp1a1 and Cyp1a2, showed significant induction over controls, and both genes were up-regulated at the 10 and 100 mg/kg doses of BaP. The pattern of gene expression noted in the 6-hr samples (i.e., significant elevation in mRNA for Cyp1a1 and Cyp1a2) did not change in the 24-hr samples (data not shown). However, none of the other genes previously reported to be up-regulated by TCDD, a ligand with very high affinity for the aryl hydrocarbon (Ah) receptor, were upregulated by the doses of BaP used in these studies (18,19). Moreover, none of the genes reported to be up- or down-regulated in response to BaP in other systems [p53 in A549 or 3T3 cells (24), c-myc and transforming growth factor (TGF)-[Beta] in a placental choriocarcinoma cell line (25), and glutathione-S-transferase (GST)-Ya in HEPG2 cells (26)] responded to BaP at any of the doses or times studied.

Induction of genes by TCE. Groups of four to five mice were treated (ip) with TCE at doses varying from 10 mg/kg to 1,000 mg/kg. Only three genes (Cyp2a, Hsp 25, and Hsp 86) were significantly induced at the highest dose of TCE tested (1,000 mg/kg). Cyp2a appeared to be repressed at the 10 mg/kg dose of TCE, but this effect was not statistically significant.

Discussion

The introduction of hundreds of new chemical entities to the market each year presents a substantial challenge for determining the potential adverse health and environmental effects associated with their use. Currently, toxicologic screening of these chemicals relies heavily on the use of a combination of short-term assays. Separate bioassays are implemented to determine immunotoxicity, neurotoxicity, and genetic toxicity end points. These assays, however, are costly, require many animals, and may take several years and millions of dollars to complete [see Gelbke et al. (27) for review]. Similarly, for chemical contaminants in the environment, the Environmental Protection Agency uses three-species bioassays to determine the relative toxicity in water of specific chemicals or of specific effluents to wildlife (28). Because these tests are limited to lethality, growth, and reproductive end points, considerable attention has focused on optimization of these processes by developing methodologies that are more efficient, informative, and cost effective.

As outlined in a recent symposium (9) and as described previously by Nuwaysir et al. (10), the use of microarrays to measure changes in gene expression in response to chemicals may improve the efficiency of standard toxicity bioassays. In environmental monitoring, microarrays could provide not only a method to quickly categorize chemicals and assign a mode of toxic action but also allow more sensitive end points to be addressed--specifically, gene expression. Moreover, as studies are conducted with larger gene chips, we will gain a more complete understanding of similarities in regulatory sequences, which control the expression of a variety of gene products. Our immediate goals have been to understand the potential limitations of this technology (dynamic range, sensitivity, and sources of variability) (20) and to determine whether microarrays offer a useful alternative approach in toxicity testing and environmental monitoring.

Important questions need to be addressed if microarrays are to be used in toxicity testing and environmental monitoring: a) Will microarrays provide distinct patterns of gene expression for separate chemical classes? b) How will this chemical signature change with dose and tissue studied, and, in those cases where the chemical species is unknown, will the study of gene regulation in multiple tissues lead to clearer understanding of the class of chemicals being studied? c) Are the methods sufficiently robust to provide reliable assessments? and d) Will these assays provide an improved method for assessing the potential synergistic effects of chemical mixtures? Although these questions will require substantial long-term efforts to fully address, our current work provides data supporting proof of concept; namely, that three environmental contaminants (Cd[Cl.sub.2], BaP, and TCE) belonging to different chemical classes, elicit unique patterns of gene expression over the doses tested in an in vivo model (Figure 1). Although limited in scope, the data presented here on gene induction by BaP are consistent with the patterns of gene induction noted in earlier studies on [Beta]-naphthoflavone (20). There are, however, several additional genes, all of which were present on our chip but were not up-regulated significantly at either of the doses or times studied with BaP but are under control of the Ah receptor (18,19). These include p53 in A549 or 3T3 cells (24), c-myc and TGF-[Beta] in a placental choriocarcinoma cell line (25), and GST-Ya in HEPG2 cells (26). At this point, it is not clear whether these differences are primarily a result of the different systems used (i.e., an in vivo system with liver tissue compared to cell lines from various sources including liver) or whether these additional genes would be upregulated in other tissues at different doses and times after BaP administration.

For the doses tested and genes examined, Cd[Cl.sub.2] was the only chemical treatment where the expression pattern changed appreciably with dose. This was readily apparent when the Cd[Cl.sub.2] concentration was increased from noncytotoxic to probable cytotoxic doses (17). At low concentrations of Cd[Cl.sub.2] (0.25-1.0 mg/kg) up-regulation of metallothionine occurred (Figure 1). As the concentration was increased, the magnitude of induction for metallothionine increased, and several other genes were induced or repressed including heat shock proteins. The dose-response data are consistent with earlier studies in which gadd 153 (chop 10) is upregulated when Cd[Cl.sub.2] is present at cytotoxic concentrations (29). Although gene expression changed with dose, the signature response for Cd[Cl.sub.2] was discernable from BaP and TCE (i.e., metallothionine remained elevated in Cd[Cl.sub.2] treated animals throughout the dose range). Additionally, the graded response of gene induction observed with Cd[Cl.sub.2] treatment illustrates the potential use of microarrays to assess a degree of exposure for specific chemicals or classes of chemicals.

Currently, we are testing the DNA arrays with additional chemicals from the same chemical classes and are including work on additional tissues to determine whether this provides a further mechanism for distinguishing chemical classes. We will use this as a basis for studying chemical mixtures to assess whether the arrays can provide information about the additive or subtractive effects of mixtures. This information will be helpful in assessing whether DNA arrays can be used to screen effluents and other complex mixtures. Finally, a larger mouse array is under construction to encompass more biochemical pathways and responses. With these larger arrays, it may be possible to begin making associations between pathways that act in concert during periods of chemical stress.

REFERENCES AND NOTES

(1.) Shalon D, Smith SJ, Brown PO. A DNA microarray system for analyzing complex DNA samples using two-color fluorescent probe hybridization. Genome Res 6:639-645 (1996).

(2.) Schena M, Heller RA, Theriault TP, Konrad K, Lachenmeier E, Davis RW. Microarrays: biotechnology's discovery platform for functional genomics. Trends Biotechnol 16:301-306 (1998).

(3.) Khan J, Bittner M, Chen Y, Meltzer P, Trent J. DNA microarray technology: the anticipated impact on the study of human disease. Biochim Biophys Acta 1423:M17-M28 (1999).

(4.) Debouck C, Goodfellow PN. DNA microarrays in drug discovery and development. Nat Genet 21(1 suppl):48-50 (1999).

(5.) DeRisi J, Penland L, Brown PO, Bittner ML, Meltzer PS, Ray M, Chert Y, Su YA, Trent JM. Use of a cDNA microarray to analyze gene expression patterns in human cancer. Nat Genet 14:457-460 (1996).

(6.) Heller RA, Schena M, Chai A, Shalon D, Bedilion T, Gilmore J, Woolley DE, Davis RW. Discovery and analysis of inflammatory disease-related genes using cDNA microarrays. Proc Natl Acad Sci USA 94:2150-2155 (1997).

(7.) Wang DG, Fan JB, Siao CJ, Berno A, Young P, Sapolsky R, Ghandour G, Perkins N, Winchester E, Spenser J, et al. Large-scale identification, mapping and genotyping of single-nucleotide polymorphisms in the human genome. Science 280:1077-1082 (1998).

(8.) Winzeler EA, Richards DR, Conway AR, Goldstein AL, Kalman S, McCullough MJ, McCusker JH, Stevens DA, Wodicka L, Lockhart DJ, et al. Direct allelic variation scanning of the yeast genome. Science 281:1194-1197 (1998).

(9.) Rockett JC, Dix DJ. Application of DNA arrays to toxicology. Environ Health Perspect 107:681-685 (1999).

(10.) Nuwaysir EF, Bittner M, Trent J, Barrett JC, Afshari CA. Microarrays and toxicology: the advent of toxicogenomics. Mol Carcinog 24:153-159 (1999).

(11.) Medlin J. Innovations: Timely toxicology. Environ Health Perspect 107:A256-A258 (1999).

(12.) Bowman, C. Tire fire sets off flow of toxic oil: smoke fills San Joaquin sky. Sacramento (CA) Bee, 25 September 1999;A1.

(13.) IARC. Beryllium, cadmium, mercury and exposure in the glass manufacturing industry. IARC Monogr Eval Carcinog Risk Hum 58:41-117 (1993).

(14.) Dipple A. Polycyclic aromatic hydrocarbon carcinogenesis: an introduction. In: Polycylic Hydrocarbons and Carcinogenesis (Harvey RG, ed). Washington, DC: American Chemical Society, 1985;1-17.

(15.) Davidson IWF, Beliles RP. Consideration of the target organ toxicity of trichloroethylene in terms of metabolite toxicity and pharmacokinetics. Drug Metab Rev 23:493-599 (1991).

(16.) NTP. Carcinogenesis Studies of Trichloroethylene (without Epichlorohydrin) in F344/N Rats and B6C3F1 Mice. TR 243. Research Triangle Park, NC: National Toxicology Program, 1990.

(17.) Beyersmann D, Hechtenberg S. Cadmium, gene Regulation, and cellular signalling in mammalian cells. Toxicol Appl Pharmacol 144:247-261 (1997).

(18.) Schmidt JV, Bradfield C. Ah receptor signaling pathways. Ann Rev Cell Dev Biol 12:55-89 (1996).

(19.) Denison MS, Heath-Pagliuso S. The Ah receptor: a regulator of the biochemical and toxicological actions of structurally diverse chemicals. Bull Environ Contam Toxicol 61:557-568 (1998).

(20.) Bartosiewicz M, Trounstine M, Barker D, Johnston R, Buckpitt A. Development of a toxicological gene array and quantitative assessment of this technology. Arch Biochem Biophys 376:66-73 (2000).

(21.) Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162:156-159 (1987).

(22.) Buckpitt Lab: List of genes-microarray. Available: http://faculty. vetmed.ucdavis.edu/faculty/arbuckpitt/BuckpittLab/ genetemplate.htm [last update 1 December 2000].

(23) Worley J, Bechtol K, Penn S, Roach D, Hanzel D, Trounstine M, Barker D. A systems approach to fabricating and analyzing DNA microarrays. In: Microarray Biochip Technology (Schena M, ed). Natick, MA:BioTechniques Books, 2000;65-85.

(24.) Pei XH, Nakanishi Y, Takayama K, Bai F, Hara N. Benzo(a)pyrene activates the human p53 gene through induction of nuclear factor kappa B activity. J Biol Chem 274:35240-35246 (1999).

(25.) Zhang L, Shiverick KT. Benzo(a)pyrene, but not 2,3,7,8-tetrachlorodibenzo-p-dioxin, alters cell proliferation and c-myc and growth factor expression in human placental choriocarcinoma JEG-3 cells. Biochem Biophys Res Commun 231:117-1120 (1997).

(26.) Chen Y, Ramos KS Negative regulation of rat GST-Ya gene via antioxidant/-electrophile response element is directed by a C/EBP-like site. Biochem Biophys Res Commun 265:18-23 (1999).

(27.) Gelbke HP, Hildebrand B, Kerfoot, EJ. Evaluation of methods for the registration of chemicals. In: Toxicology (Marquardt H, Schafer SG, McClellan RO, Welsch F, eds). New York:Academic Press, 1999.

(28.) Barbour M, Gerritsen J, Snyder B, Stribling J. Rapid Bioassesment Protocols for Use in Streams and Wadeable Rivers: Periphyton, Benthic Macroinvertebrates and Fish. EPA 841-B-99-002. Washington, DC:U.S. Environmental Protection Agency, Office of Water, 1999.

(29.) Leunthy JD, Holbrook NJ. Activation of the gadd 153 promotor by genotoxic agents: a rapid and specific response to DNA damage. Cancer Res 52:5-10 (1992).

Matthew Bartosiewicz,(1) Sharron Penn,(2) and Alan Buckpitt(1)

(1) Department of Molecular Biosciences, School of Veterinary Medicine, University of California, Davis, California, USA; (2) Molecular Dynamics, Sunnyvale, California, USA

Address correspondence to M. Bartosiewicz, Department of Molecular Biosciences, School of Veterinary Medicine, Haring Hall, UC Davis, Davis, CA 95616 USA. Telephone: [phone]. Fax: [phone]. E-mail: mjbartosiewicz@ ucdavis.edu

This research was supported by NIEHS P42 04699 and R01-ES09681. UC Davis is an NIEHS Environmental Health Center (ES 05707), and the use of core facilities to conduct this work is gratefully acknowledged.

Received 5 May 2000; accepted 27 September 2000.