Distinct gene expression profiles in immortalized human urothelial cells exposed to inorganic arsenite and its methylated trivalent metabolites.

Inorganic arsenic is an environmental carcinogen. The generation of toxic trivalent methylated metabolites complicates the study of arsenic-mediated carcinogenesis. This study systematically evaluated the effect of chronic treatment with sodium arsenite ([iAs.sup.III]), monomethylarsonous acid ([MMA.sup.III]), and dimethylarsinous acid ([DMA.sup.III]) on immortalized human uroepithelial cells (SV-HUC-1 cells) using cDNA microarray. After exposure for 25 passages to [[iAs.sup.III]] (0.5 [micro]M), [MMA.sup.III] (0.05, 0.1, or 0.2 [micro]M), or [DMA.sup.III] (0.2 or 0.5 [micro]M), significant compound-specific morphologic changes were observed. A set of 114 genes (5.7% of the examined genes) was differentially expressed in one or more sets of arsenical-treated cells compared with untreated controls. Expression analysis showed that exposure of cells to [DMA.sup.III] resulted in a gene profile different from that in cells exposed to [iAs.sup.III] or [MMA.sup.III], and that the [iAs.sup.III]-induced gene profile was closest to that in the tumorigenic HUC-l-derived 3-methylcholanthrene-induced tumorigenic cell line MC-SV-HUC T2, which was derived from SV-HUC-1 cells by methylcholanthrene treatment. Of the genes affected by all three arsenicals, only one, that coding for interleukin-1 receptor, type 11, showed enhanced expression, a finding confirmed by the reduced increase in NF-[kappa]B (nuclear factor kappa B) activity seen in response to interleukin-1[beta] in [iAs.sup.III]-exposed cells. The expression of 11 genes was suppressed by all three arsenicals. 5-Aza-denxycytidine partially restored the transcription of several suppressed genes, showing that epigenetic DNA methylatinn was probably involved in arsenical-induced gene repression. Our data demonstrate that chronic exposure to [iAs.sup.III]. [MMA.sup.III], or [DMA.sup.III] has different epigenetic effects on urothelial cells and represses NF-[beta]B activity. Key words: eDNA microarray, immortalized urothelial calls, real-time polymerase chain reaction, toxicogenomics, trivalent arsenite, trivalent methylated arsenic compounds. Environ Health Perspect 114:394-403 (2006). doi:10.1289/ehp.8174 available via http://dr.doi.org/[Online 17 August 2005]

**********

Arsenic intoxication is a worldwide problem. Billions of people in India, Bangladesh, Inner Mongolia, Taiwan, and North and South America have drunk arsenic-contaminated water for many years (National Research Council 1999) and suffer from a variety of arsenic-induced diseases, such as cancer, diabetes, hypertension, and hyperkeratosis (Chen et al. 1992; Haque et al. 2003; Liu et al. 2002). During the past few decades, arsenic carcinogenesis has been extensively studied using a variety of in vitro molecular cytogenetic approaches and in vivo animal models (Chen et al. 2004; Kitchin 2001; Rossman 2003; Waalkes et al. 2004). The carcinogenesis-associated effects of arsenic involve genotoxic damage such as chromosomal abnormalities and oxidative stress (Basu et al. 2001; Hei and Filipic 2004; Rossman 2003). In addition to genetic alterations, arsenic exposure was shown recently to induce both global hypomethylation (Xie et al. 2004) or specific hypomethylation of the cyclin D1 and estrogen receptor-[alpha] genes (Chen et al. 2004) and hypermethylation of the p53 (tumor suppressor protein p53) gene (Mass and Wang 1997). Epigenetic alterations caused by modification of DNA methylation are therefore considered to play crucial roles in arsenic carcinogenesis (Sutherland and Costa 2003).

Microarray technology, which measures changes in gene expression at the transcriptional level, is a powerful tool for studying global cellular responses to toxicants (Waters and Fostel 2004). Numerous reports have demonstrated that genes showing aberrant expression after exposure to inorganic trivalent arsenic are involved in signal transduction, cell proliferation, oxidative stress responses, and DNA repair in a variety of cell systems (Bae et al. 2003; Chen et al. 2001a; Hamadeh et al. 2002; Pea et al. 2003; Yih et al. 2002; Zheng et al. 2003) as well as in liver tumors in mice (Chen et al. 2004; Liu et al. 2004). We recently examined gene expression profiles in lymphocytes from an arsenic-exposed human population and found that the expression of several inflammatory molecules is increased after prolonged exposure to arsenic (Wu et al. 2003). Furthermore, we and others have shown that long-term exposure to low concentrations of arsenic causes increased neoplastic transformation of a variety of cells (Achanzar et al. 2002; Chien et al. 2004; Mure et al. 2003; Zhao et al. 1997). These studies strongly suggest that chronic exposure to low levels of arsenic results in epigenetic alterations that may promote arsenic-induced neoplastic transformation or tumor development.

Ingested inorganic arsenic compounds are metabolized by oxidative methylation (Kitchin 2001; Styblo et al. 2002; Thomas et al. 2001). Inorganic pentavalent arsenate ([iAs.sup.V]), trivalent arsenite ([iAs.sup.III]), and the intermediate metabolites of monomethylarsonic acid ([MMA.sup.V]), monomethylarsonous add ([MMA.sup.III]), dimethylarsinic acid ([DMA.sup.V]), and dimethylarsinous acid ([DMA.sup.III]) have been identified in the urine of arsenic-exposed subjects (Mandal et al. 2004; Wang et al. 2004). [MMA.sup.v] and [DMA.sup.v] are nontoxic, but [MMA.sup.III] and [DMA.sup.III] are more toxic than [iAs.sup.III] for a variety of cell lines (Styblo et al. 2000; Thomas et al. 2001; Vega et al. 2001). [MMA.sup.III] and [DMA.sup.III] are genotoxic for human lymphocytes (Kligerman et al. 2003) and Chinese hamster ovary cells (Dopp et al. 2004) and may interfere with DNA repair to a greater extent than [MMA.sup.v] and [DMA.sup.v] (Schwerdtle et al. 2003). The realization that [MMA.sup.III] and [DMA.sup.III] also cause injury has led to a better understanding of arsenic-mediated carcinogenesis, but their roles in arsenic carcinogenesis remain to be clarified.

To explore the effects of long-term exposure of human urothelial cells to [iAs.sup.III] and its toxic trivalent metabolites, [MMA.sup.III] and [DMA.sup.III], we initiated a systematic study of gene expression changes using cDNA microarray in an SV40-immortalized human urethra-derived urothelial cell line, SV-HUC-1 (HUC-1) (Christian et al. 1987). An HUC-1--derived 3-methylcholanthrene-induced tumorigenic cell line, MC-SV-HUC T2 (MC-T2) (Reznikoff et al. 1988) was also included to examine the relationship between the changes caused by arsenic exposure and tumorigenicity. In this report, we show that chronic exposure to trivalent arsenicals induced compound-specific cell morphologic changes. Different gene expression profiles were produced by exposure to these three trivalent arsenicals, and that caused by [iAs.sup.III] most closely resembled the profile seen in MC-T2 cells. A reduction in the increase in NF-[kappa]B (nuclear factor kappa B) activity in response to interleukin 1[beta] (IL1B) treatment was seen in arsenical-exposed cells. 5-Aza-deoxycytidine (5-aza-dC) restoration experiments showed that DNA hypermethylation was involved in the changes in gene expression caused by exposure to trivalent arsenicals.

Materials and Methods

Cell culture and arsenic compounds. HUC-1 cells (CRL-9520) and MC-T2 cells (CRL-9519) were purchased from American Type Culture Collection (Manassas, VA, USA) and grown in F-12 nutrient mixture (Ham) medium (Gibco Invitrogen Life Technologies, Carlsbad, CA, USA) supplemented with 10% fetal calf serum (Hyclone Laboratory, Logan, UT, USA), 2 mM L-glutamine, 100 U/mL of penicillin, and 100 [micro]g/mL streptomycin sulfate at 37[degrees]C in a humidified atmosphere containing 5% C[O.sub.2]. Both lines were shown to be mycoplasma-free by polymerase chain reaction (PCR) analysis. In nude mice, HUC-1 cells are nontumorigenic, whereas MC-T2 cells form tumors, [iAs.sup.III] (sodium arsenite, Art. 6287) was purchased from Merck KGaA (Darmstadt, Germany). [MMA.sup.III] and [DMA.sup.III] were synthesized in our laboratory, and their chemical structure and purity were verified by nuclear magnetic resonance, mass spectrophotometry, and dement analysis.

Cytotoxicity assay and long-term exposure to [iAs.sup.III], [MMA.sup.III], or [DMA.sup.III]. The cytotoxicity of the three trivalent arsenic compounds for HUC-1 cells was determined using the sulforhodamine B (SRB) colorimetric assay (Skehan et al. 1990). Briefly, 1 x 104 cells in 0.1 mL of F12 medium were seeded in each well of a 96-well plate and incubated for 72 hr with various concentrations of [iAs.sup.III], [MMA.sup.III], or [DMA.sup.III]. The cells were then fixed with 50% trichloroaceric acid, stained for 30 min at room temperature with 0.4% SRB in 10% acetic acid, and extracted for 5 min at room temperature with unbuffered 10 mM Tris-base solution, and the absorbance of the extract was measured at 540 nm. The growth-inhibiting and cytotoxic effects of the arsenicals were calculated using the formula [([A.sub.treated] - [A.sub.zero])/ ([A.sub.control] - Azero)] x 100%, where [A.sub.zero] is the absorbance at the time of drug addition and [A.sub.control] and [A.sub.treated] are the absorbance of untreated and treated wells, respectively, at the end of the experiment. [IC.sub.50] (median inhibitory concentration) values were calculated using CalcuSyn software (Biosoft, Cambridge, UK).

To establish arsenic long-term exposed cell lines, we maintained HUC-1 cells in the presence of trivalent arsenicals at doses giving a 90% cell survival rate, namely, 0.5 [micro]M [iAs.sup.III], 0.2 or 0.5 [micro]M [DMA.sup.III], or 0.05, 0.1, or 0.2 [micro]M [MMA.sup.III]. The trivalent arsenical-treated and untreated HUC-1 cells were subcultured twice weekly for at least 20-25 passages before being subjected to microarray assay.

Preparation of total and [poly(A).sup.+] RNA. Total cellular RNA was extracted using TRI reagent (Molecular Research Center, Cincinnati, OH, USA) according to the manufacturer's instructions. [Poly(A).sup.+] RNA was then isolated using an Oligotex mRNA midi kit (Qiagen, Hilden, Germany).

cDNA microarray analysis. The microarray chips were made in-house from a collection of approximately 2,000 expressed gene sequences, including genes of potential significance in arsenic toxicity (Wu et al. 2003) and genes selected from subtraction libraries established from oral carcinoma, nasopharyngeal carcinoma, colon cancer, and hepatoma (provided by K.W. Chang, National Yang-Ming University); detailed information on these genes is available from the Institute of Biomedical Sciences, Academia Sinica (Taipei, Taiwan; http://www.ibms.sinica.edu.tw/~bmtcl/As-chip-TCL04.xls). Sources of gene information are GenBank (http://www.ncbi. nih.gov/genbank/) and CGAP (Cancer Genome Anatomy Project; http://cgap.nci.nih. gov/Genes). The cDNA microarray hybridization experiments were performed using a colorimetric detection method as described previously (Wu et al. 2003; Yih et al. 2002). Briefly, cDNA targets were prepared from 2 [micro]g of mRNA using Superscript II reverse transcriptase (Gibco Invitrogen Life Technologies) by incorporating biotin-16-2'-deoxyuridine-5'-triphosphate (Roche Diagnostics, Mannheim, Germany). The labeled cDNA targets were hybridized to the membrane array overnight at 65[degrees]C, then the arrays were washed and incubated for 60 min at room temperature with streptavidin-conjugated alkaline phosphatase for chromagen development. After extensive washes to remove any unbound conjugate, substrate solution (0.1 M Tris, pH 9.5, 0.1 M NaCl, 5 mM Mg[Cl.sub.2], 165 [micro]g/mL nitrotetrazolium blue chloride, and 87 [micro]g/mL bromo-4-chloro-3-indolyl phosphate p-toluidine salt) was added to the array, which was then incubated in the dark at room temperature for 15-20 main with occasional shaking. The intensity of the spots on the arrays was measured using a scanner at an appropriate optical resolution. Quantitative results were obtained using GenePix (version 4.0; Axon Instruments, Union, CA, USA), normalized by a linear regression method, then analyzed and clustered using Spotfire (Spotfire Inc., Somerville, MA, USA). For each treatment, the microarray analysis was repeated 4 times.

Quantitative real-time reverse-transcriptase PCR We used total RNA as the template in the synthesis of first-strand cDNA using an oligo(dT) primer and the AMV reverse transcriptase system (Roche Diagnostics). The PCR primers used to detect the genes of interest and the internal control gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), are listed in Table i. Quantitative real-time RT-PCR (Q-PCR) was performed using an ABI PRISM 7700 sequence detection system (Applied Biosystems, Foster City, CA, USA). Thermocycling was performed in a final volume of 10 [micro]L containing 4 [micro]L of cDNA sample, 200 nM of each of the primers, and 5 [micro]L of SYBR green PCR master mix (Applied Biosystems). The relative differences in expression level between genes were expressed using cycle time ([C.sub.t]) values as follows: the [C.sub.t] value of the gene of interest was first normalized to that for GAPDH in the same sample, then the difference between the treatment and control group was calculated and expressed as an increase or decrease in cycle numbers compared with the control.

Semiquantitative RT-PCR and agarose gel electrophoresis. Total RNA was isolated and converted into first-strand cDNA as described above. The cDNA for the test gene was amplified by PCR using the Ampli Taq gold system (Applied Biosystems). The PCR conditions were 94[degrees]C for 10 min; 35 cycles of 94[degrees]C for 30 sec, 55[degrees]C for 30 sec, and 72[degrees]C for 30 sec; and 72[degrees]C for 7 min. The PCR products were analyzed on a 1.5% agarose gel. GAPDH RNA levels were analyzed in parallel to ensure that equal amounts of cDNA were loaded for each reaction

Immunoblotting analysis. Cells were harvested in ice-cold phosphate-buffered saline, pH 7.4, using a rubber policemen, then lysed by addition of an equal volume of 2x sodium dodecyl sulfate (SDS) lysis buffer (100 mM Tris HCl, pH 6.8, 200 mM [beta]-mercaptoethanol, 4% SDS, 0.2% bromophenol blue, and 20% glycerol). The protein concentration was determined using a Bio-Rad Protein Assay system (Bio-Rad Laboratory, Hercules, CA, USA) with bovine serum albumin as standard. Equal amounts (20 lag of protein) of each sample were electrophoretically separated on a 10% SDS polyacrylamide gel and electro-transferred to a PVDF (polyvinylidene difluoride) membrane. After blocking with 4% skim milk, the membranes were probed with primary mouse antibodies against heparin-binding epidermal growth factor-like growth factor (AF-259-NA; R&D System, Inc., Minneapolis, MN, USA), E-cadherin (sc-8426) or keratin 10/13 (sc-6258) (both from Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA), or integrin [beta]3 (611141; BD Transduction Laboratories, Franklin Lakes, NJ, USA) or goat antibody against [beta]-actin (sc-1616; Santa Cruz Biotechnology, Inc.), followed by secondary horseradish peroxidase--conjugated antibodies against goat or mouse IgG (AP106P or AP124P; Chemicon International Inc., Temecula, CA, USA). Bound antibody was visualized using enhanced SuperSignal West Pico chemiluminescent substrate (Pierce Biotechnology, Inc., Rockford, IL, USA). [beta]-Actin was used as the loading control.

Transaction and NF-[kappa]B activity assay. To measure NF-[kappa]B activity, we used the pNF-[kappa]B-Luc reporter (Stratagene, La Jolla, CA, USA) containing five NF-[kappa]B enhancer motifs and the coding sequence for firefly luciferase. The pRL-TK plasmid containing Renilla luciferase driven by thymidine kinase promoter (Promega Corp., Madison, WI, USA) was used as the internal control. Cells were seeded at 1.5 x [10.sup.5] per well in 24-well plates the day before and transfected with 0.25 [micro]g of pNF-[kappa]B-Luc reporter and 0.05 [micro]g pRL-TK with 0.9 [micro]L FuGene 6 reagent (Roche Diagnostics) following the manufacturer's instructions. The cells were incubated in the presence of serum-containing medium for 16 hr followed by serum starvation for 24 hr. Human recombinant IL1B (201-LB, R&D System, Inc.) was added as indicated concentrations, and incubation continued for 5 hr. The cells were then harvested, and firefly and Renilla luciferase activities were measured using the Dual-Luciferase assay system (E1910, Promega Corp.) according to the manufacturer's protocol. The NF-[kappa]B activity associated with the firefly activity was normalized to the Renilla luciferase activity.

5-Aza-dC treatment HUC-1 cells exposed to 0.2 [micro]M [MMA.sup.III] for 25 passages were incubated for 72 hr with 0, 0.5, or 2.0 [micro]M 5-aza-dC (A3656, Sigma-Aldrich Co.). Passage-matched HUC-1 cells not treated with [MMA.sup.III] or 5-aza-dC were used as controls. The expression level of the test gene in the control HUC-1 cells was taken as 100%.

Results

Morphologic changes induced in HUC-1 cells by long-term exposure to [iAs.sup.III], [MMA.sup.III], or [DMA.sup.III]. To establish an in vitro culture system for studying the chronic effect of arsenicals on urothelial cells, we determined the toxicity of three trivalent arsenicals [iAs.sup.III], [MMA.sup.III], and [DMA.sup.III] for nontumorigenic HUC-1 cells using the SRB colorimetric assay (Figure 1) and found the [IC.sub.50] values to be 2.91 [+ or -] 0.58, 0.46 [+ or -] 0.11, and 1.59 [+ or -] 0.26 [micro]M, respectively. For long-term treatment, HUC-1 cells were continuously grown in the presence of trivalent arsenicals at a dose lower than that giving 90% cell survival, that is, 0.05, 0.1, or 0.2 [micro]M [MMA.sup.III]; 0.2 or 0.5 [micro]M [DMA.sup.III]; and 0.5 [micro]M [iAs.sup.III]. Control HUC-1 cells were processed in parallel without arsenicals.

[FIGURE 1 OMITTED]

After in vitro maintenance for 25 passages, the morphology of the untreated HUC-1 cells was unchanged, but long-term exposure to arsenicals resulted in significant changes. HUC-1 cells exposed to [iAs.sup.III] or [MMA.sup.III] tended to aggregate, were smaller in size than untreated cells (Figure 2), and required longer trypsin digestion to detach them from the surface (data not shown), indicating increased adhesion activity. Long-term exposure to [DMA.sup.III] induced morphologic changes in a concentration-dependent manner, and these differed from those induced by [iAs.sup.III] and [MMA.sup.II]; [DMA.sup.III] did not induce significant cell aggregation, but the size of the cells was reduced by exposure to 0.2 [micro]M [DMA.sup.III], and this effect was greater using 0.5 [micro]M [DMA.sup.III]. These results show that chronic exposure to trivalent arsenic compounds induced differential morphologic changes in urothelial cells and that none of the induced morphologies resembled that of tumorigenic MC-T2 cells. The proliferation activity of cells cultured in the presence of each arsenical was 70-80% that of untreated controls except for cells exposed to 0.5 [micro]M [DMA.sup.III] for which the proliferation rate was 40-50% that of untreated controls. The tumorigenicity assay showed that HUC-1 cells exposed long term to these trivalent arsenicals were still nontumorigenic in nude mice after subcutaneous injection of 5 x [10.sup.6] cells, which is in contrast to the results of our previous study (Chien et al. 2004) showing that HaCaT cells (immortalized but non-tumorigenic human keratinocytes) become tumorigenic in nude mice after long-term exposure to low-dose [iAs.sup.III].

[FIGURE 2 OMITTED]

Induction of different gene profiles in HUC-1 cells after long-term exposure to [iAs.sup.III], [MMA.sup.III], or [DMA.sup.III]. To evaluate global changes in gene expression induced by long-term exposure to low-dose [iAs.sup.III], [MMA.sup.III], or [DMA.sup.III], we used a colorimetric cDNA array (Wu et al. 2003; Yih et al. 2002). We used passage-matched untreated HUC-1 cells as the negative control, and included tumorigenic MC-T2 cells to examine the relationship between changes induced by arsenic exposure and tumorigenesis. As shown in Figure 3A, the intensity (arbitrary units) of the signal for the internal control gene GAPDH was similar in untreated and arsenical-treated cells (p = 0.3008 by analysis of variance), and similar results were obtained for two other commonly used human housekeeping genes, the transferrin receptor (p90, CD71) (TFRC) and hypoxanthine phosphoribosyltransferase 1 (HPRT1), except that TFRC levels in [iAs.sup.III]-treated HUC-1 cells were slightly, but significandy, lower than those in untreated HUC-1 cells. These results showed that the intraassay variation was acceptable.

[FIGURE 3 OMITTED]

We selected a set of 114 differentially expressed genes (5.7% of the tested genes) (Supplementary Material, Table 1; http:// ehp.niehs.nih.gov/members/2005/8174/ suppl.pdf) on the basis of the criteria that there was a significant difference (p < 0.05, Student's t-test, n = 4) between the expression of the gene in untreated HUC-1 cells and in treated HUC-1 cells and/or MC-T2 cells, and that the ratio for the expression in one or more sets of treated HUC-1 cells and/or MC-T2 cells compared with the expression in untreated HUC-1 cells was > 1.9 or < 0.25. A hierarchical dendrogram of the 114 selected genes is shown in Figure 3B. The expression profiles in cells exposed to different concentrations of [MMA.sup.III] or [DMA.sup.III] were clearly clustered as separate groups. The gene expression pattern most closely resembling that in MC-T2 cells was seen in the [iAs.sup.III]-exposed HUC-1 cells. The profile of down-regulated genes in HUC-1 cells exposed to [MMA.sup.III], especially at 0.2 [micro]M, also shared similarities with the MC-T2 cell profile. [DMA.sup.III]-treated HUC-1 cells showed a profile distinct from those in the other cells. Together, these gene expression profiles show that these three different trivalent arsenic compounds had different effects on the cells.

To characterize the trivalent arsenical-responsive genes, the selected 114 genes were subjected to Gene Ontology-based biologic property analysis (FatiGO; http://www.fatigo. org/) (Al-Shahrour et al. 2004). Ninety-eight genes of known annotation were submitted, of which 85 were acceptable for level 3 biologic process analysis (Figure 3C). The major gene population (80%) was categorized as "cellular physiologic processes" covering the processes of cell growth and maintenance, cell death, mobility, and transport. In addition, 61.8% of the genes were associated with metabolism (metabolism of nucleic acids, proteins, fatty acids, and vitamin B6; protein phosphotylation and targeting; RNA processing; ubiquitin-dependent proteolysis; and regulation of transcription), 36.5% with cell communication (cell adhesion, cell-cell signaling, and signal transduction), 25.9% with responses to stimuli (endogenous and exogenous), and 23.5% with morphogenesis (cellular morphogenesis and organogenesis).

Confirmation of gene expression profiles by Q-PCR and immunoblotting assays. The arsenical-induced changes in gene expression identified by microarray analysis (Supplementary Material, Table 1; http:// ehp.niehs.nih.gov/members/2005/8174/ suppl.pdf) were checked by Q-PCR. Thirteen of the 114 genes were tested. These 13 included 2 genes [S100 calcium binding protein A8 (S100A8) and E-cadherin (CDH1)] showing no change or enhanced expression in all sets of arsenical-treated HUC-1 cells in the microarray assay, 4 genes [connexin 43 (GJA1), matrix metalloproteinase type 2 (MMP2), protease, serine 11 (IGF binding) (PRSS11), and insulinlike growth factor binding protein 5 (IGFBP5)] showing decreased expression in all sets of arsenical-treated HUC-1 cells, and 7 genes [thrombomodulin (THBD), interleukin-8 (IL8), heparin-binding epidermal growth factor-like growth factor (HBEGF), Vav 3 oncogene (VAV3), E2F transcription factor 1 (E2F1), keratin 13 (KRTI3), and interleukin-1 receptor, type II (IL1R2)] showing up- or down-regulation of expression in some, but not all, sets of arsenical-treated HUC-1 cells. Figure 4A and B shows the transcriptional differences between untreated HUC-1 cells and arsenical-treated HUC-1 cells or MC-T2 cells measured by Q-PCR assay. To examine the correlation between gene expression changes measured by microarray and by Q-PCR, we analyzed 91 pairs of data for 13 genes in HUC-1 cells exposed to the different concentrations of the three arsenicals and in MC-T2 cells. As shown in Figure 4C, in most cases, gene expression levels measured by the two methods showed a positive correlation, being either up- or down-regulated in both cases (quadrants I and III). Inconsistent results were seen in quadrant II for VAV3 in 0.05 [micro]M [MMA.sup.III]-exposed cells, KRT13 in 0.1 [micro]M [MMA.sup.III]-exposed cells, E2F1 in 0.1 and 0.2 [micro]M [MMA.sup.III]-exposed cells, and IL1R2 in 0.2 [micro]M [DMA.sup.III]-exposed cells; and in quadrant IV S100A8 in 0.5 [micro]M [DMA.sup.III]-exposed cells, IL8 in 0.2 [micro]M [DMA.sup.III]-exposed cells, and IL1R2 and IGFBP5 in MC-T2 cells.

[FIGURE 4 OMITTED]

The results of immunoblotting analysis confirmed the changes in gene expression at the protein level because protein levels of CDH1, HBEGF, and KRT13 in arsenical-treated cells and MC-T2 cells (Figure 5) agreed well with the mRNA levels measured by Q-PCR assay (Figure 4A). In addition, protein expression of integrin 133 (ITGB3) (Figure 5) was consistent with transcript levels determined by microarray.

[FIGURE 5 OMITTED]

Distinct gene alterations in HUC-1 cells after long-term exposure to [iAs.sup.III], [MMA.sup.III] or [DMA.sup.III] To determine which genes were affected similarly by these three arsenicals and which responded to specific arsenic compounds, for further comparative analysis, we selected those genes showing at least a 1.5-fold higher or 2-fold lower expression in HUC-1 cells exposed to 0.5 [micro]M [iAs.sup.III], 0.2 [micro]M [MMA.sup.III], or 0.5 [micro]M [DMA.sup.III] compared with untreated HUC-1 cells.

Enhanced expression of 29, 9, and 28 genes was seen, respectively, in [iAs.sup.III]--, [MMA.sup.III]-, and [DMA.sup.III]-exposed cells (Table 2). Of these, IL1R2 was the only gene showing enhanced expression in response to all three trivalent arsenic compounds (Figure 6A). Of thegenes showing enhanced expression in [iAs.sup.III]-, [MMA.sup.III]-, or [DMA.sup.III]-exposed cells, the expression of 15 of 29 (51.7%), 4 of 9 (44.4%), and 18 of 28 (64.3%), respectively, was enhanced uniquely by this one agent (Figure 6A). A hierarchical dendrogram of all 51 genes affected by one or other of the arsenicals is shown in Figure 6B. Clearly, these three trivalent arsenicals induced the expression of distinct patterns of genes.

[FIGURE 6 OMITTED]

In addition, suppressed expression was seen for 20, 19, and 29 genes in [iAs.sup.III]-, [MMA.sup.III]-, and [DMA.sup.III]-exposed cells, respectively (Table 3). The number of genes uniquely suppressed in [iAs.sup.III]-, [MMA.sup.III]-, or [DMA.sup.III]-exposed cells was 5 (25%), 6 (31.6%), and 14 (48.3%), respectively (Figure 6C). The higher percentage of genes uniquely suppressed in [DMA.sup.III]-exposed cells shows that exposure to [DMA.sup.III] may confer a more diverse gene pattern than exposure to the other two arsenic compounds, consistent with the hierarchical cluster analysis results (Figure 3B). However, in contrast to the single gene showing enhanced expression on exposure to all three agents, 11 genes showed suppressed expression on exposure to all three agents (Figure 6C). Ten are of known function [GJA1, ITGB3, collagen type I alpha 1 (COL1A1), IGFBP5, MMP2, myosin light polypeptide kinase (MYLK), GATA binding protein 6 (GATA6), chemokine (C-C motif) ligand 2 (CCL2), protein kinase C alpha (PRKCA), and nudeoporin (NUP88)], and one has unknown function. The hierarchical dendrogram for this set of 41 genes (Figure 6D) showed that the genes suppressed by each of these three trivalent arsenicals shared greater similarity than the enhanced genes (Figure 6B).

Repression of the IL1B-induced increase in NF-[kappa]B activity in [iAs.sup.III]-exposed HUC-1 cells. Venn diagram analysis showed that IL1R2was the only gene to show enhanced expression after exposure to the three test arsenicals. IL1R2 is one of the receptors of inflammatory cytokine ILl; however, because of the lack of a cytoplasmic signaling domain, binding to IL1R2 attenuates ILl-mediated signal transduction, such as the IL1B-induced increase in NF-[kappa]B activation (Re et al. 1996). To investigate the effect of overexpression of IL1R2 on NF-[kappa]B signaling, we examined the transgenic NF-[kappa]B-dependent promoter activity of arsenical-exposed cells in response to ILl stimulation. As shown in Figure 7, in untreated HUC-1 cells, the activity of the NF-[kappa]B-dependent promoter was increased by treatment with IL1B, and the effect was dose-dependent, with increases of 5.3-fold at 0.05 ng/mL and 8.6-fold at 0.2 ng/mL, then plateauing. In contrast, in cells exposed long-term to [iAs.sup.III], the transgenic NF-[kappa]B activity showed only a slight increase in response to IL1B. These results demonstrate that the ILl-mediated NF-[kappa]B signaling cascade was impaired in HUC-1 cells exposed long term to [iAs.sup.III].

[FIGURE 7 OMITTED]

A methyltransferase inhibitor reactivates arsenic-suppressedgenes. Gene expression can be inactivated through a wide spectrum of mechanisms. To determine whether DNA hypermethylation silencing was involved in arsenic-induced gene suppression, we tested a methyltransferase inhibitor, 5-aza-dC, for its ability to restore IGFBP5 and MMP2 mRNA levels (measured by semiquantitative RT-PCR and Q-PCR) in HUC-1 cells treated with arsenicals. As shown in Figure 8, chronic exposure to 0.2 [micro]M [MMA.sup.III] significantly reduced the expression of IGFBP5 and MMP2 to 17.5 and 8.2% of control levels, respectively. However, after 72 hr treatment with 0.5 or 2.0 [micro]M 5-aza-dC, the expression of both genes increased. Using 2 [micro]M 5-aza-dC, IGFBP5, and MMP2 gene expression increased 3.4- and 2.2-fold, respectively (Figure 8). These results suggest that at least part of the [MMA.sup.III]induced repression of genes is dependent on DNA methylation.

[FIGURE 8 OMITTED]

Discussion

Toxicogenomics is a powerful tool for studying the interaction between genes and environmental stress in disease causation (Waters and Fostel 2004). The results of a microarray-based study performed to give a comprehensive idea of the effects of trivalent arsenic compounds on human urothelial cells showed that long-term exposure of HUC-1 cells to nonlethal doses of [iAs.sup.III], [MMA.sup.III], or [DMA.sup.III] had distinct effects on a variety of cellular responses. On the basis of the morphology (Figure 2), the dendrogram (Figure 3A), and the gene profile analyses (Figure 6), the [DMA.sup.III]-induced cellular and molecular changes differed from those induced by [iAs.sup.III] or [MMA.sup.III]. In addition, the gene expression pattern induced by [iAs.sup.III] most closely resembled that in tumorigenic MC-T2 cells. However, each of these three compounds can cause DNA damage and cytogenetic alterations in mammalian cells (Dopp et al. 2004; Kligerman et al. 2003), inhibit nucleotide excision repair (Schwerdtle et al. 2003), and induce morphologic transformation in Syrian hamster embryo cells (Lee et al. 1985; Ochi et al. 2004). The effects on cancer development of the distinct changes in gene expression caused by low-dose long-term exposure to these three trivalent arsenic compounds require further clarification.

It is important to note that the HUC-1 cells used in the present study were derived by SV40 transformation (Christian et al. 1987). The transforming ability of SV40 is mediated by a large T-antigen oncoprotein binding to, and inactivating, gatekeeper tumor suppressor proteins, such as p53 and retinoblastoma (RB) (Pipas and Levine 2001). p53 protein is a crucial mediator of cellular responses to DNA damage and can initiate a program of cell-cycle arrest, cellular senescence, or apoptosis (Harris and Levine 2005). In the absence of functional p53 protein, cells lose genomic integrity and are prone to cancer development, p53 functions as a transcription factor up-regulating many target genes, such as p21 (cydin-dependent kinase inhibitor 1A, Cipl), MDM2 (transformed 3T3 cell double minute 2), and GADD45 (growth arrest and DNA-damage-inducible, alpha) (Levine 1997). A recent report (Liu et al. 2003) has shown that myeloma cells lacking p53 are more sensitive to arsenic trioxide (ATO), which induces apoptosis and arrest of the cells in G2/M phase of the cell cycle, whereas cells with wild-type p53 are relatively resistant to ATO and ate arrested in G1 phase by ATO. Thus, cells containing a wild-type or mutated/null p53 may show different cellular responses to arsenicals. Our findings showing transcriptional alterations mediated by chronic exposure to arsenic species in cells lacking functional p53 protein might therefore differ from those seen in cells with functional p53.

Our microarray profiling showed that several of the genes responding to chronic exposure to all three arsenicals were involved in the regulation of signal transduction pathways. Of the significantly affected genes in the total of 2,000 tested, that coding for IL1R2 was the only one to show enhanced expression after long-term exposure to each of the three trivalent arsenicals. IL1R2 is a nonsignaling decoy receptor of the strong proinflammatory cytokine IL1. Binding of IL1 to IL1R2 attenuates IL1-induced inflammatoty responses, such as NF-[kappa]B activation (Dower 2000; Wald et al. 2003). Our results (Figure 7) showed that the increased IL1R2 expression caused by the trivalent arsenicals disrupted signal transduction mediated by IL1B. Alterations of cellular signal transduction are suggested to play crucial roles in arsenic-mediated carcinogenesis (Qian et al. 2003). Arsenic has been shown to activate the stress-induced kinase pathway of MAPK p38, ERK (extracellular signal-regulated kinase), and JNK (c-Jun N-terminal kinase) (Cavigelli et al. 1996; Liu et al. 1996) and to repress the STAT (signal transducers and activators of transcription) signaling pathway (Hayashi et al. 2002). It can also either stimulate or inhibit NF-[kappa]B activation depending on the concentration, exposure time, and cell type (Harris and Shi 2003). For instance, arsenite or ATO has been shown to induce apoptosis of melanoma and lymphoma cells by inhibiting NF-[kappa]B activity (Ivanov and Hei 2004; Mathas et al. 2003); however, ATO enhances CD95/Fas-induced apoptosis through NF-[kappa]B activation (Woo et al. 2004). Further studies are required to characterize the biologic significance of the overexpressed IL1R2 and suppression of NF-[kappa]B activity in urothelial cells chronically exposed to arsenic.

Unexpectedly, the expression of genes that are negatively associated with carcinogenesis was also enhanced by arsenical exposure. For instance, expression of CDH1, a major component involved in cell-cell adhesion and frequently negatively associated with metastatic and invasive carcinomas (Thiery 2002), was enhanced in [iAs.sup.III]-exposed cells, and expression of its precursor protein, protocadherin 1, was enhanced in DMAm-exposed cells. THBD is a cell-surface-expressed glycoprotein playing a role in anticoaguladon (Van de Wouwer et al. 2004), and higher THBD expression in tumors correlates with better survival and reduced metastatic activity (Weiler and Isermann 2003). THBD expression was enhanced in [MMA.sup.III]- or [iAs.sup.III]-exposed HUC-1 cells. The present results suggest that a wide spectrum of cellular responses are induced to cope with chronic nonlethal toxic stresses.

In contrast to the single gene activated by all three trivalent arsenicals, 11 genes were suppressed by all three. Several of these are also involved in signal transduction. For instance, integrins, heterodimeric transmembrane receptors composed of an [alpha]-chain and a [beta]-chain, are the major component of the focal adhesive complex that regulates stress-fiber formation, cell shape migration, and signal transduction via molecules such as phosphatidylinositol 3-kinase and NF-[kappa]B (Filardo et al. 1995; Guo and Giancotti 2004; Hood and Cheresh 2002). A recent report has shown that sodium arsenite reduces focal adhesions and the phosphorylation of focal adhesion kinase and paxillin (Yancy et al. 2005), supporting the idea that focal adhesions and the downstream signal transduction pathways may be targets of chronic arsenical exposure. GJA1 is a gap-junction channel protein involved in cell communication (Aasen et al. 2003). A recent study has revealed the novel paradigm that GJA1 acts as a positive regulator of the NF-[kappa]B signaling cascade and that GJA1 not only is crucial in cell communication but also plays a role in signal transduction regulation (Matsuda et al. 2003). MYLK is one of the cytoskeleton kinases involved in focal adhesion and is also associated with NF-[kappa]B activity induced by tumor necrosis factor-[alpha] (Wadgaonkar et al. 2005). Taken together, our results suggest that many signal transduction cascades may be modulated by chronic exposure to trivalent arsenicals. Imbalanced expression of genes in response to stress has been suggested to be involved in carcinogenesis (Hofseth 2004). The present study revealed that arsenic induced toxicopathologic injuries and that cellular adaptation to arsenic was associated with alterations in the expression of a complex of genes.

Generation of oxidative stress has been proposed to be involved in the toxicity induced by low-dose arsenic (Schoen et al. 2004). The expression of many oxidative stress-responsive genes is induced by arsenic exposure; examples include superoxide dismutase I and NAD(P)H quinone oxidoreductase, which respond to oxidative stress generated by arsenic metabolic methylation (Hamadeh etal. 2002); heme oxygenase 1, a marker for arsenic-induced stress (Liu et al. 2001b); and glutathione (Brambila et al. 2002), heat-shock protein (Barnes et al. 2002), and multidrug-resistant efflux transporters (Liu et al. 2001a), which are induced during the acquisition of tolerance to arsenic. These genes showed no major changes in our gene profiles, although they were included on the array. The differences in expression profiles obtained in our system and in others may be due to tissue-specific variation but are more likely due to different dosages and exposure times. In our system, we applied [iAs.sup.III], [MMA.sup.III], and [DMA.sup.III] at doses of 0.05 to 0.5 [micro]M for more than 3 months (~25 passages), whereas the other studies mentioned above involved the use of 0.05-25 [micro]M [iAs.sup.III] for periods of only several hours to 2 weeks. Our results therefore suggest that these oxidative stress-related genes may be involved in the carcinogenic process at a period earlier than the window of this study.

Toxicant-induced gene expression changes and cancer have been shown to be associated with epigenetic modifications (Sutherland and Costa 2003). Methylation of CpG islands appears to be responsible for the transcriptional silencing of critical genes in various types of cancer (Egger etal. 2004; Jones and Baylin 2002). Exposure of cells to arsenic results in global or specific hypomethylation (Chen et al. 2001b, 2004; Xie et al. 2004) or hypermethylation of specific genes such as tumor suppressor p53 (Mass and Wang 1997). Our microarray analysis showed that the number of genes suppressed by all three agents was much higher than the number activated by all three, suggesting that down-regulation of genes might be an effect common to all three of these trivalent arsenicals. Furthermore, the methyltransferase inhibitor 5-aza-dC partially restored transcription of IGFBP5 and MMP2, again implying that epigenetic gene regulation may be involved in chronic arsenical-mediated gene modulation.

The various forms of arsenicals to which humans are exposed complicate studies of the mechanism of arsenic-mediated carcinogenesis. In addition, the lack of appropriate animal models and sensitive biomarkers for monitoring the adverse effects of arsenic has also limited progress. However, recent investigations have revealed that arsenic may exert its adverse effects via epigenetic pathways. Because biologic pathways are connected through a complicated network, the ability to identify and confirm the roles of arsenic-induced differentially expressed genes during cancer development would help us understand how a normal cell is transformed into a neoplasm and to search for new intervention regimes. The characterization of the gene expression profiles induced by arsenicals may therefore provide an understanding of their roles in carcinogenesis.

Received 5 April 2005; accepted 17 August 2005.

REFERENCES

Aasen T, Hodgins MB, Edward M, Graham SV. 2003. The relationship between connexins, gap junctions, tissue architecture and turnout invasion, as studied in a novel in vitro model of HPV-16-associated cervical cancer progression. Oncogene 22:7969-7980.

Achanzar WE, Brambila EM, Diwan BA, Webber MM, Waalkes MP. 2002. Inorganic arsenite-induced malignant transformation of human prostate epithelial cells. J Natl Cancer lest 94:1888-1891.

Al-Shahrour F, Diaz-Uriarte R, Dopazo J. 2004. FatiGO: a web tool for finding significant associations of Gone Ontology terms with groups of genes. Bioinformatics 20:578-580.

Bae D-S, Hands R J, Yang RSH, Campain JA. 2003. Gone expression patterns as potential molecular biomarkers for malignant transformation in human kerstinocytes treated with MNNG, arsenic, or a metal mixture. Toxicol Sci 74:32-42.

Barnes JA, Collins BW, Dix DJ, Allen JW. 2002. Effects of heat shock protein 70 (Hsp70) on arsenite-induced genotoxicity. Environ Mol Mutagen 40:236-242.

Basu A, Mahata J, Gupta S, Girl AK. 2001. Genetic toxicology of a paradoxical human carcinogen, arsenic: a review. Mutat Res 488:171-194.

Brambila EM, Achanzar WE, Qu W, Webber MM, Waalkes MR. 2002. Chronic arsenic-exposed human prostate epithelial cells exhibit stable arsenic tolerance: mechanistic implications of altered cellular glutathione and glutathione S-transferase. Toxicol Appl Pharmacol 183:99-107.

Cavigelli M, Li W, Lin A, Su B, Yoshioka K, Karin M. 1996. The tumor promoter arsenite stimulates AP-1 activity by inhibiting a JNK phosphatase. EMB0 J 15:6269-6279.

Chen CJ, Chen CW, Wu MM, Kuo TL. 1992. Cancer potential in liver, lung, bladder and kidney due to ingested inorganic arsenic in drinking water. Br J Cancer 66:888-892.

Chen H, Li S, Liu J, Diwan BA, Barrett JC, Waalkes MP. 2004. Chronic inorganic arsenic exposure induces hepatic global and individual gone hypomethylation: implications for arsenic hepatocarcinogenesis. Carcinogenesis 25:1779-1786.

Chen H, Liu J, Merrick BA, Waalkes MP. 2001a. Genetic events associated with arsenic-induced malignant transformation: applications of eDNA microarray technology. Mol Carcinog 30:79-87.

Chen H, Liu J, Zhao CQ, Diwan BA, Merrick BA, Waalkes MP. 2001b. Association of c-myc overexpression and hyparproliforation with arsenite-induced malignant transformation. Toxicol Appl Pharmacol 175:260-268.

Chien C-W, Chiang M-C, Ha I-C, Lee T-C. 2004. Association of chromosomal alterations with arsenite-induced tumorigenicity of human HaCaT keratinocytes in nude mice. Environ Health Perspect 112:1704-1710.

Christian BJ, Loretz LJ, Oberley TD, Reznikoff CA. 1987. Characterization of human uroepithelial cells immortalized in v/fin by simian virus 40. Cancer Res 47:6066-6073.

Dopp E, Hartmann LM, Rares AM, van Recklinghausen U, Pieper R, Shekouhi B, et el. 2004. Uptake of inorganic and organic derivatives of arsenic associated with induced cytotoxic and genotoxic effects in Chinese hamster ovary (CHO) cells. Toxicol Appl Pharrnacol 201:156-165.

Dower SK. 2000. Cytekines, virokines and the evolution of immunity. Nat Immunol 1:367-368.

Egger G, Liang G, Aparicio A, Jones PA. 2004. Epigenetics in human disease and prospects for epigenetic therapy. Nature 429:457-463.

Filardo E, Brooks P, Deming S, Damsky C, Cheresh D. 1995. Requirement of the NPXY motif in the integrin beta 3 subunit cytoplasmic tail for melanoma cell migration in vitro and in vivo. J Cell Biol 130:441-450.

Fujikawa K, Inoue Y, Sakai M, Koyama Y, Nishi S, Funada R, at el. 2002. Vav3 is regulated during the cell cycle and effects cell division. Proc Natl Acad Sci USA 99:4313-4318.

Gun W, Gianeotti FG. 2004. Integrin signalling during tumour progression. Nat Rev Mol Cell Biol 5:816-826.

Hamedeh HK, Trouba KJ, Amin RP, Afshad CA, Germolec D. 2002. Coordination of altered DNA repair end damage pathways in arsenite-expesed keratinocytes. Toxicol Sci 69:306-316.

Haque R, Mazumder DN, Samanta S, Ghosh N, Kalman D, Smith MM, et el. 2003. Arsenic in drinking water and skin lesions: dose-response data from West Bengal, India. Epidemiology 14:174-182,

Harris GK, Shi X. 2003. Signaling by carcinogenic metals and metal-induced reactive oxygen species. Mutat Res 533:183-200.

Harris SL, Levine AJ. 2005. The p53 pathway: positive and negative feedback loops. Oncogene 24:2899-2908.

Hayashi T, Hideshima T, Akiyama M, Richardson P, Schlossman RL, Chauhan D, et al. 2002. Arsenic trioxide inhibits growth of human multiple myeloma cells in the bone marrow microenvironment. Mol Cancer Thor 1:851-860.

Hei TK, Filipic M. 2004. Role of oxidative damage in the genotoxicity of arsenic. Free Radio Biol Med 37:574-581.

Hofseth LJ. 2004. The adaptive imbalance to genotoxic stress: genome guardians rear their ugly heads. Carcinogenesis 25:1787-1793.

Hood JD, Cheresh DA. 2002. Role of integrins in cell invasion and migration. Nat Rev Cancer 2:91-100.

Inoue K, Slaton JW, Kim S J, Perrotte P, Eve BY, Bar-Eli M, et el. 2000. Interleukin 8 expression regulates tumorigenicity and metastasis in human bladder cancer. Cancer Res 60: 2296-2299.

Ivanov VN, Hei TK. 2004. Arsenite sensitizes human melanomas to apoptosis via tumor necrosis factor {alpha}-mediated pathway. J Biol Chem 279:22747-22758.

Iwamoto R, Mekada E. 2000. Heparin-binding EGF-like growth factor: a juxtacrine growth factor. Cytokine Growth Factor Rev 11:335-344.

Jones PA, Baylin SB. 2002. The fundamental role of epigenetic events in cancer. Nat Rev Genet 3:415-428.

Kitchin KT. 2001. Recent advances in arsenic carcinogenesis: modes of action, animal model systems, and methylated arsenic metabolites. Toxicol Appl Pharmacol 172:249-261.

Kligerman AD, Doerr CL, Tennant AH, Harrington-Brock K, Allen JW, Winkfield E, et al. 2003. Methylated trivalent arsenicals as candidate ultimate genotoxic forms of arsenic: induction of chromosomal mutations but not gone mutations. Environ Mol Mutagen 42:182-205.

Lee TC, Oshimura M, Barrett JC. 1985. Comparison of arsenic-induced cell transformation, cytotoxicity, mutation and cytogenetic effects in Syrian hamster embryo cells in culture. Carcinogenesis 6:1421-1426.

Levine AJ. 1997. p53, the cellular gatekeeper for growth and division. Cell 88:323-331.

Liu J, Chen H, Miller DS, Saavedra JE, Keefer LK, Johnson DR, at el. 2001a. Overexpression of glutathione S-transferase II and multidrug resistance transport proteins is associated with acquired tolerance to inorganic arsenic. Mol Pharmacol 60:302-309.

Liu J, Kadiiska MB, Liu Y, Lu T, Qu W, Waalkes MP. 2001b. Stress-related gene expression in mice treated with inorganic arsenicals. Toxicol Sci 61:314-320.

Liu J, Xie Y, Ward JM, Diwan BA, Waalkes MP. 2004. Toxicogenomic analysis of aberrant gene expression in liver tumors and nontumorous livers of adult mice exposed in utero to inorganic arsenic. Toxicol Sci 77:249-257.

Liu J, Zheng B, Aposhian HV, Zhou Y, Chen ML, Zhang A, et al. 2002. Chronic arsenic poisoning from burning high-arsenic-containing coal in Guizhou, China. Environ Health Perspect 110:119-122.

Liu Q, Hilsenbeck S, Gazitt Y. 2003. Arsenic trioxide-induced apoptosis in myeloma cells: p53-dependent [G.sub.1] or [G.sub.2/]M cell cycle arrest, activation of caspase-8 or caspase-9, and synergy with AP02/TRAIL Blood 101:4078-4087.

Liu Y, Guyton KZ, Gorospe M, Xu Q, Lee JC, Holbrook NJ. 1996. Differential activation of ERK, JNK/SAPK and p38/CGBP/RK map kinase family members during the cellular response to arsenite. Free Radic Riol Med 21:771-781.

Mandal GK, Ogra Y, Anzai K, Suzuki KT. 2004. Speciation of arsenic in biological samples. Toxicol Appl Pharmacol 198:307-318.

Mass MJ, Wang L. 1997. Arsenic alters cytosine methylation patterns of the promoter of the tumor suppressor gene p53 in human lung cells: a model for a mechanism of carcinogenesis. Mutat Res 386:263-277.

Mathas S, Lietz A, Janz M, Hinz M, Jundt F, Scheidereit C, et el. 2003. Inhibition of NF-kO essentially contributes to arsenic-induced apoptosis. Blood 102:1028-1034.

Matsuda A, Suzuki Y, Honda G, Muramatsu S, Matsuzaki O, Nagane Y, et al. 2003. Large-scale identification and characterization of human genes that activate NF-kappaB and MAPK signaling pathways. Oncogene 22:3302-3318.

Mure K, Uddin AN, Lopez LC, Styblo M, Rossman TG. 2003. Arsenite induces delayed mutagenesis and transformation in human osteosarcoma cells at extremely low concentration. Environ Mol Muta 41:322-331.

National Research Council. 1999. Arsenic in Drinking Water. Washington, DC:National Academy Press, 229-301.

Ochi O, Suzuki T, Barrett JC, Tsutsui T. 2004. A trivalent dimethylarsenic compound, dimethylarsine iodide, induces cellular transformation, aneuploidy, centrosome abnormality and multipolar spindle formation in Syrian hamster embryo cells. Toxicology 203:155-163.

Ongusaha PP, Kwak JC, Zwible AJ, Macip S, Higashiyama S, Taniguchi N, et al. 2004. HB-EGF Is a potent inducer of tumor growth and angiogenesis. Cancer Res 64:5283-5290.

Pipas JM, Levina AJ. 2001. Role of T antigen interactions with p53 in tumorigenesis. Semin Cancer Biol 11:23-30.

Qian Y, Castranove V, Shi X. 2003. New perspectives in arsenic-induced cell signal transduction. J Inorg Biochem 96:271-278.

Re F, Sironi M, Muzio M, Matteucci C, Introna M, Orlando G, et al. 1996. Inhibition of interleukin-1 responsiveness by type II receptor gene transfer: a surface "receptor" with anti-interleukin-1 function. J Exp Med 183:1841-1850.

Rea MA, Gregg JP, Qin Q, Phillips MA, Rice RH. 2003. Global alteration of gene expression in human keratinocytes by inorganic arsenic. Carcinogenesis 24:747-756.

Reznikoff CA, Loretz L J, Christian B J, Wu SQ., Meisner LF. 1988. Neoplastic transformation of SV40-immortalized human urinary tract epithelial cells by in vitro exposure to 3-methyl-cholanthrene. Carcinogenesis 9:1427-1436.

Rossman TG. 2003. Mechanism of arsenic carcinogenesis: an integrated approach. Murat Res 533:3745.

Schoen A, Beck B, Sharma R, Dube E. 2004. Arsenic toxicity at low doses: epidemiological and mode of action considerations. Toxicol Appl Pharmacol 198:253-267.

Schwerdtle T, Walter I, Hartwig A. 2003. Arsenite and its biomethylated metabolites interfere with the formation and repair of stable BPDE-induced DNA adducts in human cells and impair XPAzf and Fpg. DNA Repair (Amst) 2:1449-1463.

Skehan P, Storeng R, Scudiero D, Monks A, McMahon J, Vistica D, et al. 1990. New calorimetric cytotoxicity assay for anticancer-drug screening. J Natl Cancer Inst 92:1107-1112.

Styblo M, Del Raze LM, Vega L, Germolec DR, LeCluyse EL, Hamilton GA, et el. 2000. Comparative toxicity of trivalent and pentavalent inorganic and methylated arsenicals in rat and human cells. Arch Toxicol 74:289-299.

Styblo M, Drobna 7., Jaspers I, Lin S, Thomas DJ. 2002. The role of biomethylation in toxicity and carcinogenicity of arsenic: a research update. Environ Health Perspect 110:767-771.

Sutherland JE, Costa M. 2003. Epigenetics and the environment Ann NY Acad Sci 983:151-160.

Thiery JP. 2002. Epithelial-masenchymal transitions in tumour progression. Nat Rev Cancer 2:442-454.

Thomas D J, Styblo M, Lin S. 2001. The cellular metabolism and systemic toxicity of arsenic. Toxicol Appl Pharmacol 176:127-144.

Van de Wouwer M, Collen D, Conway EM 2004. Thrombomodulin-protein C-EPCR system: integrated to regulate coagulation and inflammation. Arteriosclar Thromb Vase Biol 24:1374-1383.

Vega L, Styblo M, Patterson R, Cullen W, Wang C, Germolec D. 2001. Differential effects of trivalent and pentavalent arsenicals on cell proliferation and cytokine secretion in normal human epidermal keratinocytes. Toxicol Appl Pharmacol 172:225-232.

Waalkes MP, Ward JM, Diwan BA. 2004. Induction of tumors of the liver, lung, ovary and adrenal in adult mice after brief maternal gestational exposure to inorganic arsenic: promotional effects of postnatal phorbol ester exposure on hepatic and pulmonary, but not dermal cancers. Carcinogenesis 25:133-141.

Wadgaonkar R, Linz-McGillem L, Zaiman AL, Garcia JG. 2005. Endothelial cell myosin light chain kinase (MLCK) regulates TNFalpha-induced NF-kB activity. J Cell Biochem 94:351-364.

Wald D, Qin J, Zhao Z, Qian Y, Naramura M, Tian L, et al. 2003. SIGIRR, a negative regulator of Ton-like receptor-interleukin 1 receptor signaling. Nat Immunol 4:920-927.

Wang Z, Zhou J, Lu X, Gong Z, Le XC. 2004. Arsenic speciation in urine from acute promyelocytic leukemia patients undergoing arsenic trioxide treatment. Chem ResToxicol 17:95-103.

Waters MD, Festal JM. 2004. Toxicoganomics and systems toxicology: aims and prospects. Nat Rev Genet 5:930-948.

Weiler H, Isermann BH. 2003. Thrombomodulin. J Thromb Haemost 1:1515-1524.

Woo GH, Park IC, Park MJ, An S, Lee HC, Jin HO, et al. 2004. Arsenic trioxide sensitizes CD95/Fas-induced apoptosis through ROS-mediated upregulation of CD95/Fas by NF-kappaB activation. Int J Cancer 112-596-606.

Wu MM, Chiou HY, He IC, Chen CJ, Lee TC. 2003. Gene expression of inflammatory molecules in circulating lymphocytes from arsenic-exposed human subjects. Environ Health Perspect 111:1429-1438.

Xie K. 2001. Interleukin-8 and human cancer biology. Cytokine Growth Factor Rev 12:375-391.

Xie Y, Trouba K J, Liu J, Waalkes MP, Germolec DR. 2004. Biokinetics and subchrenic toxic effects of oral arsenite, arsenate, monomethylarsonic acid, and dimethylarsinic acid in v-Ha-ras transgenic (Tg.AC) mice. Environ Health Perspect 112:1255-1263.

Yancy SL, Shelden EA, Gilmont RR, Welsh MJ. 2005. Sodium arsenite exposure alters cell migration, focal adhesion localization and decreases tyrosine phosphorylation of focal adhesion kinase in H9C2 myoblasts. Toxicol Sci 84:278-286.

Yih L-H, Peck K, Lee T-C. 2002. Changes in gene expression profiles of human fibroblasts in response to sodium arsenite treatment Carcinogenesis 23:897-876.

Zeng L, Sachdev P, Yen L, Chan JL, Trenkle T, McClelland M, et al. 2000. Vav3 mediates receptor protein tyrosine kinase signaling, regulates GTPase activity, modulates cell morphology, and induces cell transformation, Mol Cell Biol 20:9212-9224.

Zhao CO., Young MR, Diwan BA, Coogan TP, Waalkes MR. 1997. Association of arsenic-induced malignant transformation with DNA hypomethylation and aberrant gene expression. Proc Natl Acad Sci USA 94:10907-10912.

Zheng XH, Watts GG, Vaught S, Gandolfi AJ. 2003. Low-level arsenite induced gene expression in HEK293 cells. Toxicology 197:39-48.

Pei-Fen Su, (1) Yu-Jie Hu, (2) I-Ching Ho, (1) Yang-Ming Cheng, (1) and Te-Chang Lee (1)

(1) Institute of Biomedical Sciences, Academia Sinica, Taipei, Taiwan, Republic of China; (2) Institute of Biopharmaceutical Sciences, National Yang Ming University, Taipei, Taiwan, Republic of China

Address correspondence to T-C. Lee, Institute of Biomedical Sciences, Academia Sinica. Nankang 115, Taipei, Taiwan, ROC. Telephone: [phone]. Fax: [phone]. E-mail: bmtcl@ ibms.sinica.edu.tw

Supplementary Material is available online (http:// ehp.niehs.nih.gov/members/2005/8174/suppl.pdf).

This work was supported by National Research Program for Genomic Medicine grants NSC91-3112-B-010-006, NSC92-3112-B-010-019 and NSC93-3112-B-010-005 from the National Science Council, Taiwan, ROC.

The authors declare they have no competing financial interests.

Table 1. Primer sequences used for Q-PCR and RT-PCR. (a)

Gene           Forward primer               Reverse primer

CDH1      AACGCATTGCCACATACACTCT      CATTCTGATCGGTTACCGTGATC
E2F1      GAGCAGATGGTTATGGTGATCAAA    AGGGCACAGGAAAACATCGA
GAPDH     GGAGTCCCTGCCACACTCA         GCCCCTCCCCTCTTCAAG
GJA1      CCAAAGACTGTGGGTCTCAA        CTCACTTGCTTGCTTGTTGTA
HBEGF     AGGACTTCTGCATCCATGGA        TGGTCATAGGTATATAAGCGAT
IGFBP5    CTCACTTGCTTGCTTGTTGTA       GCACTGCTTTCTCTTGTAGAA
IL1R2     TGTGCTGGCCCCACTTTC          GCACAGTCAGACCATCTGCTTT
IL8       GTGCAGTTTTGCCAAGGA          TTCTGTGTTGGCGCAGTG
KRT13     GCGGGACTACAGCCCCTACTAC      CCAGCCTGGCATTGTCAAT
MMP2      GCACTGCTTTCTCTTGTAGAA       GATCTGAGCGATGCCATCAA
PRSS11    CCGTGGTTCATATCGAATTGT       CCGTGGTTCATATCGAATTGT
S100A8    TGGAGAAAGCCTTGAACTCTATCA    GGTCTCTAGCAATTTCTTCAGGTCAT
THBD      GAGGACGTGGATGACTGCATACT     GGTACTCGCAGTTGGCTCTGA
VAV3      GAACAAGGGACACTCAAACTACCA    GTTCCTAATGACCTGCATCTTTGG

(a) Primer sequences were designed by Primer Express 2.0
(Applied Biosystems) according to their sequence ID shown in
Supplementary Material, Table 1
(http://ehp.niehs.nih.gov/members/2005/8174/suppl.pdf).

Table 2. Genes showing enhanced expression in HUC-1 cells
exposed long-term to [MMA.sup.III], [DMA.sup.III], or [iAs.sup.III].

Cluster and gene name          Gene symbol  Function

[MMA.sup.III],
    [DMA.sup.III], and
    [iAs.sup.III]
  Interleukin-1                IL1R2        Interleukin-1, type II,
  receptor, type II                           blocking receptor
                                              activity
[MMA.sup.III] specific

  Arachidonate 5-              ALOX5AP      Leukotriene biosynthesis
    lipoxygenase-
    activating protein
  Thrombomodulin               THBD         Receptor activity

[DMA.sup.III] specific
  A kinase (PRKA) anchor       AKAP5        Signal transduction
    protein 5
  Ataxin 1                     ATXN1        RNA binding
  CD22 antigen                 CD22         Cell adhesion
  Connective tissue            CTGF         Regulation of cell
    growth factor                             growth, cell adhesion
  RNA, U70 small nucleolar     RNU70        Unknown
  Integrin beta 1 binding      ITGB1BP1     Cell adhesion
    protein 1
  E2F transcription            E2F1         [G.sub.1] phase of
    factor 1                                  mitotic cell cycle
  Guanine nucleotide           GNB2           signaling pathway
    binding protein
    (G protein), beta
    polypeptide 2
  Keratin 13                   KRT13        Intermediate filament
  Keratin 15                   KRT15        Intermediate filament
  Laminin, alpha 3             LAMA3        Cell surface receptor
  Phosphoinositide-            PIK3R1       Phosphatidylinositol
    3-kinase, regulatory                      3-kinase activity
    subunit, polypeptide 1
  Protocadherin 1              PCDH1        Cell-cell signaling
  Transducin (beta)-like       TBL1XR1      Regulation of
    1X-linked receptor 1                      transcription
  Vav 3 oncogene               VAV3         Small GTPase-mediated
                                              signal transduction
  V-myc myelocytomatosis       MYCL 1       Transcription factor
    viral oncogene homolog 1                  activity
  Xeroderma pigmentosum,       XPA          Nucleotide-excision
    complementation group A                   repair
[iAs.sup.III] specific
  E-cadherin                   CDH1         Cell-cell adhesion
  CDC2-related protein         CRK7         Protein kinase activity
    kinase 7
  Tubulin-specific             TBCD         Beta-tubulin folding
    chaperone d
  Cisplatin resistance         CRR9         Integral to membrane
    related protein CRR9p
  Cyclin D2                    CCND2        Regulation of cell cycle
  Polymerase IDNA directed),   POLB         DNA replication
    beta
  Ring finger protein 141      RNF141       Ubiquitin ligase complex
  Nuclear factor of            NFATC2       Transcription factor
    activated T-cells,                        activity
    cytoplasmic,
    calcineurin-dependent 2
  S100 calcium binding         S100P        Calcium ion binding
    protein P
  Fem-1 homolog c              FEM1C        Unknown
    (C. elegans)
  Transforming growth          TGFB1I4      Transcription factor
    factor beta 1 induced                     activity
    transcript 4
[MMA.sup.III] and
    [iAs.sup.III]
  Diphtheria toxin receptor    HBEGF        Growth factor activity
    (heparin-binding
    epidermal growth
    factor-like
    growth factor)
[DMA.sup.III] and
    [iAs.sup.III]
  Glucose phosphate            GPI          Glucose-6-phosphate
    isomerase                                 isomerase activity,
                                              cytokine activity
  Growth factor, augmenter     GFER         Cell proliferation
    of liver regeneration
  Hect domain and RLD 6        HERC6        Ubiquitin cycle
  Homeodomain interacting      HIPK3        Protein kinase activity
    protein kinase 3
  Interleukin-8                IL8          Cytokine activity
  Involucrin                   IVL          Keratinocyte
                                              differentiation
  SRY (sex determining         SOX15        Transcription factor
    region Y)-box 15                          activity
  Mitogen-activated protein    MAP3Kl1      JNK cascade
    kinase kinase kinase 11
  DC0717-S-0282 (a)
  S100 calcium binding         S100A8       Inflammatory response
    protein A8

Gene names and symbols are from UniGene (http://www
.ncbi.nlm.nih.gov/entrez/query.fcgi?db=unigene).

(a) Clone ID of original subtraction libraries.

Table 3. Genes suppressed in HUC-1 cells after long-term exposure
to [MMA.sup.III], [DMA.sup.III], or [iAs.sup.III].

                                      Gene
Cluster and gene name                 symbol     Function

[MMA.sup.III], [DMA.sup.III],
 and [iAs.sup.III]
  Chemokine (C-C motif) ligand 2      CCL2       Cell-cell signaling
  Collagen, type I, alpha 1           COL1A1     Extracellular matrix
  Connexin 43                         GJA1       Cell-cell signaling
  GATA binding protein 6              GATA6      Transcription factor
                                                   activity
  Insulin-like growth factor          IGFBP5     Regulation of
    binding protein 5                              cell growth
  Integrin, beta 3                    ITGB3      Cell-matrix adhesion
  Matrix metalloproteinase 2          MMP2       Metallopeptidase
                                                   activity
  Myosin, light polypeptide kinase    MYLK       Protein kinase
                                                   activity
  Nucleoporin 88 kDa                  NUP88      Nuclear pore
  Protein kinase C, alpha             PRKCA      Cell proliferation
  Homo sapiens PAC clone RP5-1057Ml              Unknown
    from 7, complete sequence

[MMA.sup.III] specific
  A kinase (PRKA) anchor protein 5    AKAP5      Signal transduction
  CDC-like kinase 3                   CLK3       Protein kinase
                                                   activity
  0C0717-S-0282 (a)
  V-myc myelocytomatosis viral        MYCL1      Transcription factor
    oncogene homolog 1                             activity

[DMA.sup.III] specific
  CCAAT/enhancer binding protein      CEBPB      Transcription factor
    (C/EBP), beta                                  activity
  Cyclin A1                           CCNA1      Regulation of cell
                                                   cycle
  Heparin-binding epidermal growth    HBEGF      Growth factor
    factor-like growth factor                      activity
  DNA-damage-inducible transcript 3   DDIT3      Cell cycle arrest
  Fibroblast growth factor            FGFR1      MAPKKK cascade,
    receptor 1                                     protein kinase
                                                   activity
  RNA pseudouridylate synthase        RPUSD4     RNA processing
    domain containing 4
  Proteasome (prosome, macropain)     PSMA3      Ubiquitin-dependent
    subunit, alpha type, 3                         protein catabolism
  Hypoxia-inducible factor 1,         HIF1A      Response to stress,
    alpha subunit                                  signal
                                                   transduction,
                                                   transcription
                                                   factor activity
  Chromosome 21 open reading          C21orf4    Unknown
    frame 4
  N-Ethylmaleiinide-sensitive         NSF        ATP-dependent
    factor                                         peptidase activity
  Ornithine decarboxylase 1           ODC1       Catalytic activity
  Thioredoxin-like 5                  TXNL5      Electron transporter
                                                   activity
  Thrombospondin 1                    THBSI      Cell adhesion,
                                                   signal
                                                   transduction,
                                                   coagulation

[iAs.sup.III] specific
  3-Hydroxy-3-methylglutaryl-         HMGCS1     Acetyl-CoA
    coenzyme A synthase 1                          metabolism
  Myristoylated alanine-rich          MARCKS     Actin binding,
    protein kinase C substrate                     cell motility
  Nedd4 family interacting            NDFIP2     Signal transducer
    protein 2                                      activity
  Proteasome subunit, beta type, 1    PSMB1      Ubiquitin-dependent
                                                   protein catabolism
  Peroxisomal biogenesis factor 14    PEX14      Integral to
[MMA.sup.III] and [iAs.sup.III]                    peroxisomal
                                                   membrane
  Protease, serine,                   PRSS11     Serine-type
    11 (IGF binding)                               endopeptidase
                                                   activity,
                                                   regulation of
                                                   cell cycle

[DMA.sup.III] and [iAs.sup.III]
  Arachidonate 5-lipoxygenase-        ALOX5AP    Leukotriene
    activating protein                             biosynthesis
  SMAD, mothers against DPP           SMAD7      Transforming growth
    homolog 7 (Drosophila)                         factor beta
                                                   receptor,
                                                   inhibitory
                                                   cytoplasmic
                                                   mediator activity
  Transcription factor 12             TCF12      Regulation of
                                                   transcription

Gene names and symbols are from UniGene (http://www.ncbi.nlm.nih.gov/
entroz/query.fcgi?db=unigene).

* Clone ID of original subtraction libraries.