• arsenic;
  • human bladder cells;
  • gene expression;
  • benchmark dose


  1. Top of page
  2. Abstract
  7. Author Contributions
  8. Acknowledgements
  10. Supporting Information

Gene expression changes in primary human uroepithelial cells exposed to arsenite and its methylated metabolites were evaluated to identify cell signaling pathway perturbations potentially associated with bladder carcinogenicity. Cells were treated with mixtures of inorganic arsenic and its pentavalent or trivalent metabolites for 24 hr at total arsenic concentrations ranging from 0.06 μM to 18 μM. One series (five samples) was conducted with arsenite and pentavalent metabolites and a second (10 samples) with arsenite and trivalent metabolites. Similar gene expression responses were obtained for pentavalent or trivalent metabolites. A suite of eight gene changes was consistently identified across individuals that reflect effects on key signaling pathways: oxidative stress, protein folding, growth regulation, metallothionine regulation, DNA damage sensing, thioredoxin regulation, and immune response. No statistical significance of trend (NOSTASOT) analysis of these common genes identified lowest observed effect levels (LOELs) from 0.6 to 6.0 μM total arsenic and no observed effect levels (NOELs) from 0.18 to 1.8 μM total arsenic. For the trivalent arsenical mixture, benchmark doses (BMDs) ranged from 0.13 to 0.92 μM total arsenic; benchmark dose lower 95% confidence limits (BMDLs) ranged from 0.09 to 0.58 μM total arsenic. BMDs ranged from 0.53 to 2.7 μM and BMDLs from 0.35 to 1.7 μM for the pentavalent arsenical mixture. Both endpoints varied by a factor of 3 across individuals. Thisstudy is the first to examine gene expression response in primary uroepithelial cells from multiple individuals and to identify no effect levels for arsenical-induced cell signaling perturbations in normal human cells exposed to a biologically plausible concentration range. © Environ. Mol. Mutagen., 2013. © 2012 Wiley Periodicals, Inc.


  1. Top of page
  2. Abstract
  7. Author Contributions
  8. Acknowledgements
  10. Supporting Information

It has been suggested that the dose–response for the carcinogenicity of inorganic arsenic (iAs) is highly nonlinear, with an effective threshold, due to an apparent mode of action involving inhibition of DNA repair [Li and Rossman, 1989; Snow et al., 2005; Calabrese et al., 2007; Straif et al., 2009; Gentry et al., 2010]. Cytoxicity has also been proposed as a nonlinear mode of action for the bladder carcinogenicity of arsenic [Wei et al., 2005; Cohen et al., 2006, 2007; ATSDR, 2007]. Arsenic is not a direct-acting mutagen as observed in a number of mutation assays [ATSDR, 2007; Klein et al., 2007; Kitchin and Wallace, 2008; EFSA, 2009], but does react directly with cellular proteins [Hartwig et al., 2002; Kitchin and Wallace, 2005, 2008; Benton et al., 2011; Zhou et al., 2011].

Trivalent arsenic species play an indirect role in damaging DNA by induction of reactive oxygen species (ROS) resulting in oxidative DNA base damage when studied using in vitro systems [Kasai, 1997; Helbock et al., 1999; Mei et al., 2002; Rossman, 2003; Ding et al., 2005]. Arsenic-induced oxidative stress can be attenuated, however, by modulation of cellular thiols [Flora, 2011]. Arsenite has been shown to inhibit poly(ADP-ribose) polymerase-1 (PARP-1), a major enzyme involved in DNA base excision repair by excising 8-hydroxyyl-2′-deoxyguanosine (8-OHdG) DNA lesions [Yager and Wiencke, 1993, 1997; Hartwig et al., 2002; Qin et al., 2008; Zhou et al., 2011]. At exposures ≤3μM for 24 hr to AsIII alone in a human keratinocyte cell line (HaCat), no significant increase of ROS or DNA strand breaks was observed and PARP-1 activity was not detected [Qin et al., 2008]. At 10 μM AsIII in HaCat cells, ROS and DNA strand breaks were detected; PARP-1 activity was not assayed at this concentration, however, dose-dependence of ROS induction by arsenite was shown. At 15 weeks exposure to 1.0 μM AsIII, HaCat cells show increased proliferation, DNA double strand breaks and anchorage independent growth. It is speculated that the repressive effect of NFκB on TP53 function in these cells leads to genomic instability [Qin et al., 2008]. Involvement of NRF2 in the arsenic-induced malignant transformation of keratinocytes has also been suggested [Pi et al., 2008]. In contrast to NF-κB, however, the activation of the NRF2 pathway provides protection against the effects of arsenic trioxide induced oxidative stress in keratinocytes [Pi et al., 2008] and immortalized human bladder urothelial cells [Wang et al., 2007a]. Tumor cells that are deficient for NRF2 display increased sensitivity to arsenic trioxide induced apoptosis [Liu et al., 2010].

Standard animal cancer bioassays of exposure to iAs alone have not shown significantly increased tumor incidence in the bladder [Tokar et al., 2010, 2011; IARC, 2012]. However, exposure to DMAV at ≥50 mg/L in the drinking water did induce urinary bladder cancer in rats; the mode of action is considered to involve a nonlinear process involving cytotoxicity and sustained increased cell proliferation [Wei et al., 2005; Cohen et al., 2006, 2007; ATSDR, 2007]. Urinary bladder hyperplasia was observed in both wild type and arsenic (+3 oxidation state) methyltransferase knockout mice following exposure to concentrations greater than 1 ppm arsenite in drinking water [Yokohira et al., 2011].

In vitro studies using a single dose of trivalent arsenicals in SV40 transformed human uroepithelial cells (UROTsa cells) [Petzoldt et al., 1995] have been employed to explore the effect of chronic low exposure to arsenic. Activation of NRF2 protects these cells from toxicity due to AsIII and MMAIII exposure [Wang et al., 2007a]. DNA strand breaks persisted up to 52 weeks in transformed UROTsa cells in the presence of a single dose of 50 nM MMAIII; PARP-1 activity increased significantly in the absence of MMAIII [Wnek et al., 2009]. More recently, in transformed UROTsa cells exposed at 50 nM AsIII and MMAIII, sustained production of inflammatory cytokines IL-1, IL-6, IL-8, and TNF was observed as well as an increase in NF-κB, c-jun, p38MAPK protein, and ERK [Escudero-Lourdes et al., 2010]. Anchorage-independent growth was observed at 24 weeks of exposure to 50 nM MMAIII; at 52 weeks exposure, cells were injected into immunocompromised nude mice to produce squamous cell tumors. However, exposure to arsenic in drinking water has been most closely associated with transitional cell carcinoma of the bladder in humans [Chiou et al., 2001; Karagas et al., 2004; Eble et al., 2004] therefore the human health implications of squamous cell tumors produced in immunocompromised nude mice by in vitro arsenic transformed UROTsa cells remain unclear.

Fifty-six proteins displayed significant changes in the TK-6 lymphoblastoid cell line exposed to a single dose of 0.1 μM As III with 21 proteins being up regulated and 35 down-regulated [Benton et al., 2011]. Modulation of gene transcription associated with proteins regulating cell proliferation, inflammation, and cell death was reported. Histone proteins were among those up-regulated perhaps related to arsenite's induction of chromosome damage and modulation of DNA methylation in mammalian cells [Lerda, 1994; Kitchin, 2001]. Among down-regulated proteins were CNR2 that inhibits cell proliferation and attenuates inflammatory processes and IL1RN that inhibits activity of IL-1, a cytokine implicated in inflammatory response.

In summary, the foregoing studies provide insight into induction of gene transcription changes at a single low dose of AsIII but do not provide information regarding potential transitions over a biologically plausible range of exposure concentrations.

To more comprehensively evaluate the available evidence regarding the concentration-response relationship for arsenic-gene interactions in mammalian cells, we conducted an extensive literature review on in vitro studies of gene expression responses to inorganic arsenic exposure in tumor-derived cell lines, immortalized cells and primary cells [Gentry et al., 2010]. A hierarchy of responses related to exposure concentration was observed across an array of mammalian cell types. Responses at concentrations ≤0.1 μM arsenic indicated adaptive responses while those studies in which exposures were between 0.1 μM and 10 μM resulted in responses related to oxidative stress, proteotoxicity, inflammation, proliferative signaling, as well as cell cycle G2/M checkpoint control, inhibition of DNA repair, and induction of apoptosis. At in vitro exposures >10 μM, changes in apoptotic genes prevailed. Immortalized and primary cells appeared to respond similarly, however, gene expression in tumor-derived cell lines may be altered due to inactivation or over expression of key genes. These results taken together with the known metabolism and protein binding activities of arsenic support a carcinogenic mode of action hypothesis in which a cascade of biological responses progressing from adaptive (interaction with critical genes or proteins related to DNA repair and oxidative stress) to proliferative (cell cycle control, proteotoxicity) occurs across a concentration-response continuum. At even higher arsenic exposure concentrations, apoptosis related to toxicity may overtake other responses, thereby leading to suppression of malignant processes. Findings of dose-dependent transitions in the induction of specific genes related to adaptation, cell proliferation control and toxicity provide reasonable support for the hypothesis that an exposure-response relationship for carcinogenicity is very likely nonlinear.

In order to explore these concepts in vivo, we conducted a 12-week exposure of mice to arsenate in drinking water in which gene expression changes in the urinary bladder were determined at 1 and 12 weeks of exposure [Kenyon et al., 2008; Clewell et al., 2011]. The direction of gene expression over time changed principally from down-regulation at 1 week to up-regulation at 12 weeks. These results are consistent with findings from in vitro studies indicating an early adaptive response followed by suppression of gene expression upon longer exposure. Morphogenesis, cell cycle control, inflammation (immune response), apoptosis/survival, and DNA damage response were the principal pathways affected over the exposure concentration range. A nonmonotonic concentration-response was observed at both 1 and 12 weeks exposure wherein many fewer gene expression changes were observed at 2 mg/L As than at lower or higher concentrations providing evidence of a dose-dependent transition at 2 mg As/L in drinking water.

Given earlier results of our comprehensive literature review of gene expression changes in primary mammalian cells as well as concordant results in the in vivo mouse bladder study, the present study was undertaken to assess genomic response in human primary uroepithelial cells. Cells were exposed to mixtures of inorganic arsenic and its methylated metabolites reflective of their proportion in urine of humans ingesting arsenic in drinking water. While it is well understood that microarray analysis does not provide information about the translational regulation of gene expression to protein levels or their functional capabilities, discovery of distinct gene sets related to similar biological pathways provides a means to evaluate the mode of action of arsenicals across a relevant concentration gradient. We here also focus on an important aspect of human health risk estimation—interindividual response variability—by examining induction of significant common and dissimilar pathways in cells from a number of different individuals over a biologically plausible concentration range.


  1. Top of page
  2. Abstract
  7. Author Contributions
  8. Acknowledgements
  10. Supporting Information


Sodium (meta)arsenite (AsIII) ≥99% purity (CAS 7784-46-5), methylarsonic acid (MMAV) ≥98% purity (CAS 124-58-2), and sodium cacodylate trihydrate (DMAV) >98% purity (CAS 6131-99-3) were purchased from Sigma Chemical (St. Louis, MO). Monomethyl arsenic diiodide (MMAIII) >98% purity (CAS 7207-97-8) and dimethylarsenic iodide (DMAIII) >98% purity (CAS 676-75-5) were purchased from Argus Chemicals (Vernio, Italy) and stored according to manufacturer specifications. Stock solutions of AsIII, MMAV, and DMAV were prepared in sterile distilled water; stock solutions of MMAIII and DMAIII were prepared in 0.1% dimethylsulfoxide (DMSO). Dosing solutions were prepared in complete Keratinocyte-SFM (Invitrogen, Carlsbad, CA) culture medium on the day of cell culture treatment.

Cell Culture and Arsenic Treatment Protocol

Human primary uroepithelial cell cultures were derived from small pieces of excess distal ureter tissue acquired during kidney transplant surgery under informed consent from normal human kidney donors. The protocol was reviewed and approved by the Institutional Review Board at the University of Nebraska Medical Center. Tissue was immediately transferred to standard custodial HTK (histidine-tryptophan-ketoglutarate) preservation solution and shipped refrigerated overnight to the receiving laboratory where it was placed into a clean sterile Petri dish containing 5 mL Keratinocyte-SFM culture medium with supplements. Cells were isolated by cutting the ureter tissue open and gently scraping the interior. The cells were passed through a 100 μm nylon filter (BD Falcon Cell Strainer #352360) to remove debris and rinsed with an additional 5 mL. medium. All cells from each individual donor sample (Tables I and II) were pooled into a 50 mL centrifuge tube. Viability was measured by trypan blue dye exclusion. Cells were centrifuged for 5 min at 200g, resuspended in 5 mL cell culture medium, seeded at a density of 50,000 cells per well into 96-well BD BioCoat plates (collagen I-coated; BD # 354649) and incubated at 37°C in a 5% CO2 atmosphere. RNA was extracted from the cells using the Invitrogen PureLink 96 kit with PureLink DNase (Invitrogen, Carlsbad, CA). Cells were analyzed for the expression of keratin 10, PPAR-γ, and uroplakin 2 mRNA using qRT-PCR (Roche Lightcycler) in a single-step reaction using the QuantiTect SYBR Green RT-PCR kit (Qiagen, Valencia, CA). The GAPDH housekeeping gene was used as an internal control and results expressed as a ratio of target gene: GAPDH as a fold change relative to control. The expression of keratin 10, PPAR-γ, and uroplakin 2 mRNA provide confirmation of epithelial cell growth and differentiation [Roop et al., 1983; Southgate et al., 1994, 2007; Varley et al., 2003; Kong et al., 2004].

Table I. Sex, Age, and Smoking Status of 5 Individual Subjects Treated With Arsenite Plus Pentavalent Metabolites Protocol
Microarray sample numberSexAgeSmoker (Y/N)
Table II. Sex, Age, and Smoking Status of 10 Individual Subjects Treated With Arsenite Plus Trivalent Metabolites Protocol
Microarray sample numberSexAgeSmoker (Y/N)
  1. Microarray sample numbers listed in Tables I and II, represent cells from each of 15 different individuals. Thus, samples UC1–UC5 in Table I represent five different individuals as distinguished from the five samples listed UC1–UC5 in Table II.


Following the initial incubation period, cells were treated for 24 hr with five-dose levels of combinations of arsenite, MMA, and DMA in a 1:1:4 proportion to mimic approximate ratios of these metabolites measured in urine of humans exposed to arsenic in drinking water (Tables III and IV). Experiments were conducted on samples from five individual subjects with combinations of arsenite, MMAV, and DMAV (Table III) and on samples from ten individual subjects with combinations of arsenite, MMAIII, and DMAIII (Table III). Thus, uroepithelial cells from a total of 15 different individuals were treated with combinations of arsenite and its methylated metabolites as described. Compound solubility in culture was assessed by nephelometry immediately after dosing and at the end of treatment just before cell harvest.

Table III. Arsenic In Vitro 24-Hr Treatment Protocols With Arsenite Plus Pentavalent Metabolites
TreatmentiAsIII (μM)MMAV (μM)DMAV (μM)Total mixture (μM)
Table IV. Arsenic In Vitro 24-Hr Treatment Protocols With Arsenite Plus Trivalent Metabolites
TreatmentiAsIII (μM)MMAIII (μM)DMAIII (μM)Total mixture (μM)

Biochemical Assays

Adenosine Triphosphate (ATP) Cytotoxicity Assay

To assess cytotoxicity after treatment, medium was removed from cells and 100 μL ATP cell lysis buffer was added. Plates were mixed for 10 min and stored at −80°C. On the day of analysis, samples were thawed and an ATP calibration curve was prepared in cell lysis buffer. ATP was determined using a luciferase-based assay (ATP-Lite, Perkin-Elmer).

Intracellular Glutathione

Intracellular glutathione levels were determined as described [Griffith, 1980] with modifications. At the end of the culture treatment period, medium was removed and 200 μL metaphosphoric acid (MPA) was added to each well with or without the addition of 1-methyl-2-vinyl-pyridinium trifluoromethane sulfonate (M2VP) to determine reduced to oxidized glutathione ratio (GSH:GSSG). Plates were shaken for 5 min at room temperature and stored at −20°C until analysis. Samples were thawed just before analysis and centrifuged at >2,000g for at least 2 min. Sample aliquots were transferred to a clean 96-well plate along with GSSG standard curve samples. Sample pH was adjusted to 7.0 and each well received an aliquot of sodium phosphate reaction buffer containing EDTA, DTNB, NADPH, and glutathione reductase. Plates were mixed for 15 to 30 min at room temperature. Glutathione content was determined colorimetrically at 415 nm. Standard curve GSSG concentrations were multiplied by 2 to convert values to glutathione equivalents (GSX). GSX concentration (pmoles/well) was determined by regression analysis using a standard curve. GSH:GSSG ratios were calculated for each treatment concentration and statistically evaluated using one-way analysis of variance.

Dihydro-2′,7′-dichloroflorescin-diacetate (DCFDA) Analysis

Cellular oxidative stress (peroxide formation) was measured using the DCFDA assay. DCFDA (Sigma C-6827) was dissolved in DMSO at a concentration of 205 mM. DMSO stock was diluted to 10 μM in Hanks balanced salts solution. Cells were pre-exposed to DCFDA for 1 hr before treatment. DCFDA was then removed before adding treatment solutions. Freshly prepared t-butyl hydroperoxide was added to positive control cell wells. Plates were read at 0, 1, 3, and 6 hr after addition of t-butyl hydroperoxide at 485 nm/530 nm in a Packard Fusion plate reader. Results were corrected for outliers (±2 standard deviations) and expressed as fold change relative to untreated controls.

Speciation of Arsenic in Cell Culture Supernatants

Aliquots of cell culture supernatants were analyzed for presence of inorganic As (AsIII and AsV), and the major methylated metabolites (monomethylarsonic acid (MMA) and dimethylarsenic acid (DMA)) in order to ascertain the stability of these compounds in culture and to note deviations, if any, in compound proportions since major changes in compound proportions from that of the dosing solution could indicate methylating capacity in these cells. Samples were stored in the dark at −80°C until analysis. Samples were thawed immediately before analysis. Reagent grade deionized water (>18.0 MΩ) and high purity chemicals were used for all analytical work. Commercial stock solutions of AsIII and AsV (1,000 μg AsIII mL−1 in 2% HCl and 1,000 μg AsV mL−1 in 2% NaOH, respectively) were purchased from High-Purity Standards (Charleston, SC). Standard solutions of MMA and DMA were prepared with disodium methylarsonate hexahydrate and dimethylarsenic acid, respectively, obtained from ChemService. NIST SRM 1640 was used to evaluate recovery of AsV. A high-performance liquid chromatograph (HPLC) was interfaced with an inductively coupled plasma mass spectrometer (ICP-MS) to accomplish species separation followed by quantification [Branch et al., 1994; Heitkemper et al., 2001; Mandal et al., 2001]. The chromatographic system consisted of a Dionex GP 50 pump with a 200 μL sample loop, a Dionex AS 50 autosampler, and a Hamilton PRP-X100 anion exchange column [Heitkemper et al., 2001; Wang et al., 2007b; Ketavarapu et al., 2008]. At the beginning of each chromatographic run 50 μL of a 10 μg AsV L−1 solution was injected to correct for any signal sensitivity drift that occurred over the duration of sample batch time. The ICP-MS system consisted of a Perkin Elmer SCIEX Elan DRC-e equipped with a dynamic reaction chamber [Wang et al., 2007b; Bednar et al., 2009]. Oxygen was used as the reaction gas at a flow rate of 0.2 mL min−1 and a rejection parameter q of 0.4. Method limits of detection for AsIII, AsV, MMA, and DMA were 0.0047 μM, 0.0056 μM, 0.0033 μM, and 0.0031 μM, respectively.

Gene Expression Microarray Measurements

Microarray measurements were made separately on control and five treatment levels for each individual's cell culture sample. Total RNA was isolated using the Invitrogen PureLink 96 kit with Invitrogen PureLink DNAse (Invitrogen, Carlsbad, CA). The quantity and integrity of the RNA was verified spectrophotometrically and with the Agilent Bioanalyzer (Palo Alto, CA). Double-stranded cDNA was synthesized from 150 ng of total RNA using the 3′ IVT Express Kit (Affymetrix, Santa Clara, CA). Biotin-labeled cRNA was transcribed from the cDNA and 15 μg of labeled cRNA was fragmented and hybridized to Human Genome U133 Plus 2.0 Arrays for 16 hr at 45°C. The hybridized arrays were washed using the Gene Chip Fluidics Station 450 and scanned using a GeneChip 3000 scanner.

Statistical Analysis of Gene Expression Microarray Data

Initial population level analysis was performed with Partek® (Partek Genome Suite 6.5; Partek Incorporated, St. Louis, MO). Arrays were robust multi-chip analysis (RMA) normalized [Irizarry et al., 2003] and a standard two-way analysis of variance (ANOVA) was performed with factors being the arsenic treatment dose and the subject's treated cell sample (designated as a random factor). Linear contrasts were used to assess significance of differential gene expression at each experimental treatment dose.

Principal components analysis (PCA) was performed with Partek® to assess differences between the subjects within each experimental dose range. PCA uses orthogonal transformation to convert a set of possible correlated values into linear noncorrelated variables known as principal components. Each axis is one of the principal component variables with x-axis always PC1, the variable that explains the single largest proportion of the variance, y-axis is PC2, or the second largest component of variance, and z-axis is PC3 or the third largest variance. PCA analysis was performed to determine the variability between responses for individual subjects across the dose range.

In order to perform analysis on differential gene expression within individual subjects, Affymetrix array files were analyzed in R (release 2.11.1) [R Core Team, 2012] and BioConductor (release 2.6) (Bioconductor; available at: Using the package affycoretools, fold change was calculated on RMA normalized intensity values using only those probe sets with at least 25% of probes displaying a normalized intensity value greater than or equal to log2(100) [Irizarry et al., 2003]. Statistical significance testing of individual comparisons of treatment versus control array pairs was performed using the SScore package [Kennedy et al., 2006]. The SScore (significance score) algorithm [Zhang et al., 2002; Kerns et al., 2003] uses raw (i.e., non-normalized) probe level data directly to determine differential gene expression. By analyzing probe level data, SScore is able to compute probabilities of differential gene expression in the absence of replicate samples. SScore relies on the initial assumption of an error model where raw detected signal intensity is assumed to be proportional to the probe pair intensity for genes which are highly expressed, and should approach background array intensity for genes expressed at low levels. The SScore is calculated from the probe pair error estimates and is summed over the probe pairs to derive the SSscore of an entire probe set. Assuming no differential gene expression, the SScore will follow a standard normal distribution. This allows conventional P values to be derived for each probe set and hence each gene represented on the array. The SScore algorithm inherently adjusts for multiplicity when comparing only a single pair of expression arrays [Zhang et al., 2002; Kennedy et al., 2006]. The results were used to identify single genes with statistically significant changes in expression for each individual subject.

Gene Enrichment Analysis

Gene enrichment analysis was conducted using the GeneGo pathway maps and process networks in the Metacore database (version 6.2, GeneGo, St. Joseph, MI). The enrichment P values were calculated based on a hypergeometric distribution and significant enrichment was defined as a false discovery corrected P value <0.05. GeneGo enrichment analysis was performed separately for each individual sample treated with AsIII plus trivalent methylated metabolite treatment mixtures (10 samples) and with AsIII plus pentavalent methylated metabolite treatment mixtures (five samples). The analysis was then used to identify the top pathways as indicated by P value at each dose. The individual GeneGo pathway maps were reviewed in order to recognize the smallest dose at which the lowest significant dose level was observed.

Benchmark Dose (BMD) Analysis

The continuous linear model USEPA BMDS software version (2.2) was used to calculate BMDs. A constant variance, no restrictions on the parameters and a benchmark response (BMR) of 1 standard deviation difference from control were applied. BMD lower 95% confidence limits (BMDLs) were also calculated from the continuous linear models fit to the data.


To determine a lowest significant dose level, the no statistical significance of trend method (NOSTATOT) [Tukey et al., 1985] was applied. This procedure typically has greater power for determining dose relationships than do multiple pairwise tests and was used to determine a lowest observed effect level (LOEL) and a no observed effect level (NOEL). A suitable trend test is first selected; for this analysis, linear regression was used. The trend test is then applied using all dose groups. If the trend test indicates no significant trend, the highest dose may be considered a no observed effect level (NOEL) or the NOSTASOT dose. If the trend test applied to all groups detects a significant trend, the highest dose group is deleted from consideration and the test is repeated. This process is repeated until there is no statistically significant trend. The outcome is thus based on lack of acceptance of the null hypothesis. The lowest dose for which there is a statistically significant trend is the lowest observed effect level (LOEL).


  1. Top of page
  2. Abstract
  7. Author Contributions
  8. Acknowledgements
  10. Supporting Information

Primary uroepithelial cell cultures were established from a small segment of normal distal ureter tissue of 15 individual kidney transplant donors. Cell morphology and marker expression indicated that uroepithelial cells were well differentiated after approximately 2 to 3 weeks.

The majority of donors were female (12 of 15) nonsmokers; the average donor age was 44 years (Tables I and II). Cell culture treatments were carried out on cells from five individual subjects using a protocol that involved the application of a mixture of iAsIII and pentavalent methylated metabolites protocol (Table III). Similarly, cell culture treatments were carried out on cells from 10 different individual subjects using the protocol that involved the application of a mixture of AsIII plus trivalent methylated metabolites (Table IV).

Cytotoxicity as reflected by measurement of ATP at 24 hr of treatment for iAsIII plus MMAV and DMAV (Fig. 1a) and iAsIII plus MMAIII and DMAIII treatment (Fig. 1b) showed a statistically significant change only at the highest exposure concentration combination of 1 μM iAsIII, 1 μM MMAIII, and 4 μM DMAIII (Fig. 1b). For all pentavalent and trivalent treatment concentrations, cellular viability was >80% of control.

thumbnail image

Figure 1. Mean concentrations of ATP viability assessment in samples treated with five increasing treatment concentrations of a combination of (a) arsenite and pentavalent (iAsIII + MMAV + DMAV) or (b) trivalent (iAsIII + MMAIII + DMAIII) methylated compounds per treatment protocol at 24 hr in culture. No significant difference between treatment groups (one-way ANOVA, P > 0.05) or between samples (two-way ANOVA, P > 0.05) for pentavalents and trivalents with the exception of paired t-test between control and the highest trivalent treatment indicated by asterisk (P < 0.05). N = 3. Error bars indicate standard error of mean (SEM) of three replicates. [Color figure can be viewed in the online issue, which is available at]

Download figure to PowerPoint

No significant changes were observed in cellular oxidative stress as measured by DCFDA analysis at 1, 3, and 6 hr of exposure (Figs. 2a and 2b). There was little variation in response between different subject's cell samples.

thumbnail image

Figure 2. Mean concentrations of DCFDA oxidative stress response for samples treated with five increasing concentrations of a combination of (a) arsenite and pentavalent (iAsIII + MMAV + DMAV) or (b) trivalent (iAsIII + MMAIII + DMAIII) arsenical compounds per treatment protocol at 1, 3, and 6 hr in culture. Error bars indicate standard error of mean (SEM) of three replicates. (a). DCFDA oxidative stress response—pentavalent treatment. No significant difference between treatment groups (two-way ANOVA, P > 0.05). (b). DCFDA oxidative stress response—trivalent treatment. No significant difference between treatment groups (two-way ANOVA, P > 0.05). [Color figure can be viewed in the online issue, which is available at]

Download figure to PowerPoint

Cellular oxidative stress was additionally assessed by determination of the oxidized to reduced glutathione ratio (GSH: GSSG) after 3 hr of exposure to iAsIII plus trivalent methylated compounds for three subject's cells only. No significant changes were observed in ratios over the dose range (see Supporting Information Fig. S1).

Chemical speciation analysis conducted on cell culture supernatants at 0 and 24 hrs of treatment for each arsenic compound and each sample indicated that arsenic compounds were stable over the 24 hr exposure period (see Supporting Information Fig. S2). No significant increase in MMA and DMA with either the trivalent or the pentavalent mixture was observed over time, thereby providing some indication that under these experimental conditions the cells do not possess substantial arsenic methylating capacity.

Population level statistical analyses were conducted using two-way ANOVA with contrasts on samples from five individual subjects that were treated with combinations of iAsIII and MMAV and DMAV. The number of genes significantly differentially expressed showed no significant change in expression in any gene until the total arsenic concentration reached 1.8 μM (Table V). Similarly, for the population level analysis using two-way ANOVA with contrasts of samples from 10 individual subjects that were exposed to a combination of iAsIII, MMAIII, and DMAIII, the number of genes significantly differentially expressed showed a significant change only when the total arsenic concentration in the mixture reached 1.8 μM (Table VI). In both cases, the genes associated with statistically significant express changes at the population level were related primarily to the cellular oxidative stress response (data not shown).

Table V. Number of Genes Significantly Differentially Expressed After 24 Hr In Vitro Treatment of Primary Human Uroepithelial Cells From Five Individuals With Arsenite Plus MMAV and DMAV per Treatment Protocol
Dose levelTotal arsenic (μM)Genes significant by FDR <0.05Genes significant by FDR <0.05 AND FC >±1.5
 Overall ANOVA significant by dose4,799n/a
 ANOVA linear contrasts  
Table VI. Number of Genes Significantly Differentially Expressed After 24 Hr In Vitro Treatment of Primary Human Uroepithelial Cells From 10 Individuals With Arsenite Plus MMAIII and DMAIII per Treatment Protocol
Dose levelTotal arsenic (μM)Genes significant by FDR <0.05Genes significant by FDR <0.05 and FC >±1.5
  1. Population level results showing the number of genes significantly differentially expressed from analysis using two-way ANOVA with linear contrasts over the population of (Table V) five subject's samples exposed to arsenite plus MMAV and DMAV, and (Table VI) 10 subject's samples exposed to arsenite plus MMAIII and DMAIII for each control and exposure dose. FDR is false discovery rate [Benjamini and Hochberg, 1995]. FC is simple numeric fold change applied as either an up or a down-regulation of 1.5-fold or greater.

 Overall ANOVA significant by dose903n/a
 ANOVA linear contrasts  

Principal component plots illustrating dose–response for each of the individual samples treated with either the pentavalent arsenical mixture (Fig. 3a) or trivalent arsenical mixture (Fig. 3b) provide some insight into the variation in response across individual samples. The size of the ellipse around each individual's cells' response to varying doses of arsenicals reflects deviation in response within that individual. However, the angle of view also affects perception of relative interindividual and intraindividual deviation. Nonetheless, this plot indicates that interindividual variability in genetic background is likely greater than intraindividual variability in response to different doses of arsenic compounds. In addition, the direction of gene changes for different individuals appears to deviate from parallel, suggesting a different qualitative response as well. These observations were confirmed by additional analyses discussed below.

thumbnail image

Figure 3. Interindividual variability in response as reflected in three-dimensional principal components analysis (PCA) plots of dose-response (ellipses set at standard deviation of 2) to in vitro treatment with increasing concentrations of arsenite plus metabolites. (a). PCA plot of response to in vitro treatment with increasing concentrations of arsenite plus pentavalent metabolites in cells from five different individuals. (b). PCA plot of response to in vitro treatment with increasing concentrations of arsenite plus trivalent metabolites in cells from 10 different individuals.

Download figure to PowerPoint

In order to evaluate the cellular effects of arsenic at lower concentrations than those observed at the population level, individual level gene pathway analysis (Metacore from GeneGo, Inc.) was conducted on each of the 10 samples that were treated with combinations of iAsIII, MMAIII, and DMAIII (Table VII). The lowest concentration at which a significant change in gene expression was observed (LOTEL, Table VII) based on the SScore from single individuals was at a total arsenic concentration below 0.1 μM. Genes showing significant expression change were highly diverse among different individuals. Gene expression pathway enrichments for individuals at these LOTELs were related to inflammation, epithelial-to-mesenchymal transition, protein folding, oxidative stress, cell cycle, and DNA damage response. A similar result was obtained for the five individual samples that were treated with combinations of iAsIII, MMAV, and DMAV (see Supporting Information, Table SI).

Table VII. Individual Level Statistically Significant Genomic Pathway Enrichment at the Lowest Observed Transcriptional Effect Level (LOTEL)a for Uroepithelial Cells From Each Individual Exposed to a Trivalent Arsenite Mixture
  • An empty space indicates no statistically significant response observed at the LOTEL in that category.

  • a

    a[Ankley et al., 2006].

  • b

    bMissing data at 0.06 µM and 0.18 µM total arsenic.

  • c

    cMissing data at 0.06 µM total arsenic.

 DNA damage      X XX
 p53     X   X
 Cell cycle     X XXX
 Inflammation  XXX XXXX
 WNT     X   X
 Akt      X X 
 Apoptosis         X
 AP1/NRF2XX X      
 Protein folding     XX  X

Subsequently, genes showing a ≥1.5-fold change in expression were identified for each of the 10 individual samples treated with arsenite and the trivalent methylated compounds in order to explore the range of interindividual differences in specific gene expression responses (Table VIII). The selection of a ≥1.5-fold change cut-off point provides opportunity to detect small changes in key hub genes (e.g., p53) that may be highly significant biologically, whereas large changes in genes at pathway termini (e.g., HMOX1) may be easily detectable, but large expression changes in these genes may have relatively small import with regard to sustained toxicological impact.

Table VIII. Genes Showing ≥1.5-Fold Change at the LOTEL for Each Subject After 24 Hr Treatment of Primary UroepithelialCells Exposed to Increasing Concentrations of 1:1:4 Mixtures of Arsenite, MMAIII, and DMAIII
Gene symbolGene nameIndividual
  1. An empty space indicates the specific gene was not observed in the individual at the LOTEL).

AKT1v-akt Murine thymoma viral oncogene homolog 1    X     
CASP3Caspase 3, apoptosis-related cysteine peptidase         X
CDKN1A (p21)Cyclin-dependent kinase inhibitor 1A (p21, Cip1)     X    
EGR1Early growth response 1 X        
ESR1Estrogen receptor 1 X        
FosFBJ osteosarcoma oncogeneXX   X  X 
HIF1Hypoxia inducible factor 1, alpha subunit (basic helix-loop-helix transcription factor)     X XX 
Hsp07Heat shock cognate 70-kd protein      X   
JUNjun Proto-oncogene  X    XX 
MAPK1 (Erk2)mitogen-activated protein kinase 1         X
MDM2Mdm2 p53 Binding protein homolog (mouse) X   X   X
MYCv-myc Myelocytomatosis viral oncogene homologXXXX XXXX 
NFKB1Nuclear factor of kappa light polypeptide gene enhancer in B-cells 1X      X  
Raf1v-raf-1 Murine leukemia viral oncogene homolog 1X         
RELAv-rel Reticuloendotheliosis viral oncogene homolog A (avian)         X
SRCv-src Sarcoma (Schmidt-Ruppin A-2) viral oncogene homolog (avian)      XX  
STAT3Signal transducer and activator of transcription 3 (acute-phase response factor)        X 
TGFB1Transforming growth factor, beta 1        X 
TP53Tumor protein 53XX  XXXX X
WNT1Wingless-type MMTV integration site family, member 1      X   

Genes were identified in individual samples that exhibited a statistically significant dose–response trend in at least 8 out of 10 subjects for the trivalent treatments and at least 3 out of 5 subjects for the pentavalent treatments (Table IX). The BMD and the BMDL95 for each gene for each individual was then calculated using the continuous linear model USEPA Benchmark dose software (BMDS) version 2.2. The lowest BMD for any gene was 0.13 μM for changes in expression of HMOX1 in an individual sample treated with the trivalent mixture expressed as total arsenic concentration in the mixture. Examination of the BMD range for these commonly expressed genes showed an approximately threefold variation in response between subject cells treated with the arsenical mixtures (Table IX).

Table IX. Benchmark Dose Ranges for Genes With a Statistically Significant Dose–Response Trend in Primary Uroepithelial Cells of Most Subjects after Treatment with Arsenite, MMAIII, and DMAIII (Trivalent) Mixtures (At Least 8 out of 10 Subjects) and Arsenite, MMAV, and DMAV (Pentavalent) Mixtures (At Least 3 out of 5 Subjects)
Gene nameGene probe IDDescriptionNo. subjects expressing the gene/total subjectsBMD range (μM)BMDL range (μM)No. subjects expressing the gene/total subjectsBMD range (μM)BMDL range (μM)
  1. The continuous linear model USEPA Benchmark dose software (BMDS) version 2.2 was used to calculate the BMDs. A constant variance, no restrictions on the parameters and a benchmark response (BMR) of 1 standard deviation difference from control were used. BMD lower 95% confidence limits (BMDLs) were calculated from the continuous linear models fit to the data. Doses are expressed as total arsenic concentration (μM) in trivalent and pentavalent treatment mixtures.

HMOX1203665_atOxidative stress response10/100.13–0.500.09–0.333/51.6–2.71.1–1.7
FKBP5224840_atProtein folding9/100.36–0.920.24–0.584/51.0–2.20.66–1.4
TXNRD1 201266_atThioredoxin reductase9/100.32–0.750.21–0.483/52.3–2.61.5–1.7
MT1E212859_x_atMetallothionine regulation8/100.24–0.770.16–0.494/50.53–1.70.35–1.1
DDB2203409_atDNA damage sensing8/100.30–0.880.20–0.564/50.67–2.30.44–1.5
LGALS8208933_s_at; 208934_s_at; 208935_s_at; 208936_x_at; 210731_s_at; 210732_s_atCell adhesion, growth regulation8/100.16–0.920.11–0.584/51.0–2.30.69–1.5
THBD203888_at; 203887_s_at; 237252_atImmune response8/100.32–0.900.20–0.573/50.55–2.70.37–1.7

Gene expression response (log2 normalized gene expression value) for each of eight individual genes identified in primary uroepithelial cells from at least 8 of 10 individuals following 24 hr treatment with increasing concentrations of a 1:1:4 mixture of arsenite, MMAIII, and DMAIII was plotted against the concentration of total arsenic in each mixture (Fig. 4). Genes that exhibited a change in expression were those for oxidative stress response (Fig. 4a), protein folding (Fig. 4b), thioredoxin reductase (Fig. 4c), metallothionine regulation (Fig. 4d), DNA damage sensing (Fig. 4e), thioredoxin (Fig. 4f), cell adhesion, growth regulation (Fig. 4g), and immune response (Fig. 4h). Note that y-axis indicating gene expression responses are not plotted on the same scale for each gene since the magnitude of response was highly variable among the genes. Results for cells from five individuals following 24 hr treatments with mixtures of arsenite, MMAV, and DMAV were highly concordant since the most common genes identified were oxidative stress response (HMOX1), protein folding (FKBP5), cell adhesion, growth regulation (LGALS8), metallothionine regulation (MT1E), and DNA damage sensing (DDB2).

thumbnail image

Figure 4. Dose–response for expression of genes most commonly expressed among 10 subjects. Results are shown for individual gene expression in primary uroepithelial cells in at least 8 out of 10 individuals following 24 hr treatment with increasing concentrations of a 1:1:4 mixture of arsenite, MMAIII, and DMAIII. x-Axis: dose (μM) expressed as concentration of total arsenic. y-Axis: Log2 normalized gene expression value. (a). Oxidative stress response. (b). Protein folding. (c). Thioredoxin reductase. (d). Metallothionine regulation. (e). DNA damage sensing. (f). Thioredoxin. (g). Cell adhesion; growth regulation. (h). Immune response.

Download figure to PowerPoint

In general, these plots indicate a monotonic response with increasing gene expression response with increasing dose for a number of regulatory genes e.g., oxidative stress response, cell adhesion, growth regulation, thioredoxin reductase, and thioredoxin. In addition, limited response variability is observed among subjects for the foregoing gene expression responses, whereas responses for metallothionine regulation expression appear to be more variable. Likewise, decreasing gene expression for protein folding with increasing dose appears to show relatively greater interindividual variation in response. For two genes, DDB2 (DNA damage sensing) and THBD (immune response), there is a qualitative suggestion of a nonmonotonic response with a change in direction of response around 0.1 μM arsenite.

In order to further explore these observations statistically, the NOSTATOT statistical method was applied to these individual data (Table X). As earlier stated, the NOSTATOT method is based on lack of acceptance of the null hypothesis. In order to evaluate the direction of expression change for a specific gene, refer to Figure 4. Excellent concordance of the lowest significant dose levels identified by SScore and NOSTATOT methods was observed (see Supporting Information Fig. S3). Significant gene expression response for all eight genes was detected consecutively at the highest two total arsenic doses (1.8 μM, 6.0 μM) for all subjects and occasionally at 0.6 μM. For at least two individuals, a bimodal or nonmonotonic response is suggested (UC6 for LGALS8 and UC2 for MT1E gene expression response, respectively).

Table X. NOSTASOT Analysis of Individual Genes Significantly Differentially Expressed (log2 Normalized Gene Expression Value) at Treatment Dose of Total Arsenic (μM) that are also those Genes Expressed in at least 8 of 10 Subjects after 24 Hr Treatment with 1:1:4 Mixtures of Arsenite, MMAIII, and DMAIII
ProbeSymbolLabelControl to 0.18 μMControl to 0.6 μMControl to 1.8 μMControl to 6 μMLOEL (μM)NOEL (μM)
  1. Numbers in bold reflect statistically significant dose coefficient in linear regression (P < 0.01). Probe number and gene symbol originated from Affymetrix Gene Chip Human Genome U133 Plus 2.0. LOEL is lowest observed effect level. NOEL is no observed effect level.aThese genes had multiple probes associated with them. The probe with the most subjects expressing the gene is presented here.

203665_atHMOX1UC10.88910.56210.0105 0.000061.8
UC20.51380.1020 0.0015 0.00001.80.6
UC30.56370.5763 0.0099 0.00011.80.6
UC40.66160.14020.0668 0.000261.8
UC50.23400.2622 0.0066 0.00001.80.6
UC60.2720 0.0075 0.0001 0.00000.60.18
UC70.64330.1229 0.0041 0.00001.80.6
UC80.29540.4796 0.0090 0.00001.80.6
UC90.14920.76940.2393 0.001061.8
UC110.26430.37270.1966 0.000761.8
224840_atFKBP5UC10.84960.04540.3058 0.002861.8
UC20.25220.58570.2877 0.001361.8
UC30.38110.48410.2217 0.001561.8
UC40.44400.83140.1084 0.000461.8
UC60.82700.98160.0830 0.000361.8
UC70.02470.87510.1132 0.000861.8
UC80.39200.01400.3117 0.001861.8
UC90.10540.59500.9675 0.009761.8
201266_atTXNRD1UC10.09310.0197 0.0002 0.00091.80.6
UC20.26180.0210 0.0026 0.00131.80.6
UC30.87100.0690 0.0011 0.00081.80.6
UC40.10540.0155 0.0006 0.00021.80.6
UC50.41370.84780.0377 0.000261.8
UC60.22430.1094 0.0053 0.00231.80.6
UC70.35980.0929 0.0011 0.00191.80.6
UC90.23480.48580.0251 0.001161.8
UC110.16580.97440.2898 0.004761.8
212859_x_atMT1EUC10.54760.58480.6978 0.005261.8
UC2 0.00570.33240.0256 0.000261.8
UC30.19270.09070.0569 0.000961.8
UC40.19390.1569 0.0021 0.00011.80.6
UC50.09940.94280.0843 0.000361.8
UC60.28720.50990.0248 0.000161.8
UC70.02750.69460.0478 0.000161.8
UC110.04620.45150.1413 0.000761.8
203409_atDDB2UC10.97510.07710.0909 0.000761.8
UC20.69380.3340 0.0061 0.00021.80.6
UC30.74230.58080.0508 0.000161.8
UC40.45720.38310.1494 0.002261.8
UC60.70560.99570.1307 0.001661.8
UC70.33430.41770.1952 0.001761.8
UC80.35300.52110.5719 0.008361.8
UC110.60540.98640.2889 0.001561.8
208864_s_atTXNUC10.03930.1251 0.0063 0.00491.80.6
UC20.48440.12470.0178 0.001361.8
UC30.1098 0.00640.0187 0.003361.8
UC40.01320.05880.0257 0.000161.8
UC50.0787 0.0038 0.0001 0.00020.60.18
UC60.14440.44070.0116 0.003261.8
UC70.25430.0154 0.0028 0.00281.80.6
UC90.25080.60910.2808 0.001361.8
208933_s_atLGALS8aUC10.97870.75420.0651 0.000461.8
UC20.90840.48290.0782 0.000361.8
UC30.73660.35190.0190 0.000361.8
UC40.05010.0473 0.0005 0.00001.80.6
UC50.58900.18690.3402 0.002661.8
UC6 0.00490.05920.5146 0.005861.8
UC70.1224 0.0026 0.0017 0.00020.60.18
203887_s_atTHBDaUC10.91610.50650.0412 0.000261.8
UC20.26770.0382 0.0032 0.00161.80.6
UC30.43940.11430.4166 0.008761.8
UC40.34860.84110.0373 0.000261.8
UC50.19750.86010.0516 0.009261.8
UC60.93410.63420.1430 0.004761.8
UC70.56640.16620.0174 0.009261.8
UC90.0176 0.0004 0.0029 0.00390.60.18


  1. Top of page
  2. Abstract
  7. Author Contributions
  8. Acknowledgements
  10. Supporting Information

This study represents the first time that primary uroepithelial cells from 15 individual normal human donors have been cultured and treated with a wide range of arsenical concentrations in a 1:1:4 (iAs, MMA, DMA) ratio based on that typically observed in urine from arsenic-exposed humans. With the exception of the highest trivalent mixture dose, no significant effect of exposure concentration for the arsenical mixtures was shown for cytotoxicity (ATP levels) with >80% cell viability at all doses. No effect on phenotypic indicators of oxidative stress or change in cellular glutathione ratio was shown. Taken together, these results indicate that cells were in an optimal physiological state during arsenic treatment. Limitations of an in vitro system that may serve to reduce or increase interindividual variability in response are recognized in that the magnitude of effects seen may be different from those that may occur in vivo. Speciation of arsenic in control and treated culture supernatants at 0 and 24 hr showed that arsenic concentrations were stable over time and suggested that significant oxidation of arsenite to arsenate had not taken place. These findings also imply that under the conditions of this study primary uroepithelial cells do not exhibit significant methylating capacity. Since speciation of MMA and DMA at these concentrations was not feasible, cells were treated with known concentrations of either the pentavalent or trivalent methylated compounds along with arsenite. However, results of numerous biochemical and gene expression analyses as illustrated herein were highly similar for treatments with either the trivalent or pentavalent methylated arsenicals accompanied by iAsIII in both instances.

When gene expression results from all subjects were combined in a population level analysis using two-way ANOVA with contrasts relative to treatment doses for either trivalent arsenicals or pentavalent arsenicals, no significant gene expression effects were observed until a total arsenic dose of 1.8 μM was attained (Tables V and VI). For both trivalent and pentavalent arsenicals, the gene changes identified at the population level were related primarily to oxidative stress. Individual level single replicate analysis was then carried out to evaluate the dose–response for pathway enrichments (Table VII). These analyses revealed a remarkable similarity to previous findings in the mouse [Clewell et al., 2011] in that inflammatory and epithelial-to-mesenchymal transition responses were seen at relatively low concentrations.

Principal component analysis indicated that variation in gene expression between individuals was, in general, greater than the changes in expression elicited by arsenic treatment (Fig. 3). The direction of gene changes for different individuals does not appear to be parallel suggesting a qualitative difference as well. It is unlikely that pharmacokinetic differences can account for this degree of difference in response. Pathway analysis confirmed that interindividual variability dominates gene expression response compared with response to arsenic treatment concentrations. Highly diverse gene expression responses for single individuals were observed at arsenic concentrations as low as 0.1 μM total arsenic.

The extensive variability between individuals in different genes expressed in response to the same in vitro treatment concentration(s) also suggests that treatment effects imposed on individual genetic backgrounds result in a unique pattern and magnitude of gene expression for each individual (Table VIII). Sex, age, and smoking status are likely to have contributed to the differences observed in measured outcomes; however, statistical power to detect the potential influence of these variables was limited. Given large differences in the number and type of unique genes significantly expressed between individuals, it was not possible to combine all gene expression data in order to completely characterize global response at the population level.

Eight genes were identified that met criteria of being significantly differentially expressed by SScore analysis and showing a ≥1.5-fold change and occurring in at least 8 of 10 subjects treated with trivalent arsenical mixtures and in at least 3 of 5 subjects treated with pentavalent arsenical mixtures (Tables IX and X). BMDs for the eight genes most commonly expressed in cells of subjects ranged from 0.13 to 0.92 μM total arsenic in trivalent mixtures and from 0.53 to 2.7 μM for total arsenic in pentavalent mixtures (Table IX). These results appear to be consistent with those changes noted in primary cells in vitro, as well as in mouse bladder cells in vivo reported by Clewell et al. [2011], providing evidence of a threshold or transition concentration critical to the potential carcinogenicity of arsenic compounds.

Gene expression changes in genes primarily related to oxidative stress (HMOX1, TXN, TXNRD1, MT1E) were observed at LOELs as low as 0.6 to 1.8 μM (Table X) and were up-regulated with increasing arsenic concentrations. In contrast, gene expression changes related to inflammatory and immune response (THBD), protein folding (FKBP5), and DNA damage sensing (DDB2) exhibited LOELs in the range of 6 μM and were down-regulated with increasing concentrations of arsenic (Fig. 4). Specifically, the DNA repair gene, DDB2, is down-regulated at higher concentrations of As and the transition noted at a concentration of approximately 0.1 μM is one from up-regulation to down-regulation. This is consistent with a carcinogenic mode of action in which higher doses inhibit DNA repair, thus increasing misreplication and mutagenesis. FKBP5, a gene involved in protein folding, is also down-regulated by As, but only at relatively high concentrations. As with DDB2, FKBP5 shows evidence of up-regulation at doses below 0.1 μM As. In contrast, LGALS8, a gene involved in cell adhesion, is up-regulated by As, but with a fairly broad range of BMD and BMDL values.

Oxidative stress response is characteristic of in vitro arsenic exposure [Kasai, 1997; Helbock et al., 1999; Mei et al., 2002; Rossman, 2003; Ding et al., 2005; Gentry et al., 2010] and of in vivo exposure at higher arsenic concentrations [Kitchin and Ahmad, 2003]. Results in the present study are concordant with these findings. In a previous comprehensive review of the available data from both in vitro and in vivo studies on gene expression changes [Gentry et al., 2010], clear evidence of concentration and time dependence was demonstrated on the expression of various genes or proteins following exposures to inorganic arsenic compounds. In primary cells incubated for 24 hr with arsenite, up-regulation of genes associated with oxidative stress (i.e., superoxide dismutase 1 and NADPH quinine oxidoreductase) were reported at the lowest concentrations tested (approximately 0.01 μM). With increasing concentrations of arsenic, a transition in the expression of genes related to DNA repair, cell cycle control, and apoptosis was noted in primary cells at concentrations between 0.1 and 1.0 μM, suggesting a transition or threshold from expression changes in pathways associated with adaptive responses to those potentially relevant to the mode of action for inorganic arsenic carcinogenicity.

Results from this present study are coincident with those of the short term mammalian cell in vitro exposure studies described above. Here, we detected gene expression responses for multiple genes at LOELs of 0.6 to 6.0 μM with respect to total arsenic concentration. Most noteworthy, we were also able to quantitatively describe NOELs in the range of 0.18 to 1.8 μM total arsenic concentration for these same genes.

Dose–response for pathway enrichments in the present study was remarkably similar to that observed in previous studies. In an earlier in vivo drinking water study conducted in mice exposed to arsenate in drinking water [Clewell et al., 2011], significant changes in the expression of genes associated with similar pathways as those observed in in vitro studies [Gentry et al., 2010] were noted in mouse bladder cells following 1 or 12 weeks of exposure. The changes in gene expression were bimodal in nature, with significant changes in expression following exposure to the lowest concentration (0.5 mg As/L) and the two highest concentrations (10 and 50 mg As/L), but no significant changes observed following exposure to 2 mg As/L. Based on the arsenite measured in the urine of the treated mice (∼0.05 to 16 μg/L), the bladder cells in vivo would be expected to be exposed to similar concentrations in vivo as those evaluated in vitro (0.01–10 μM) by Gentry et al. [2010]. However, cells in vivo would be expected to be exposed to a mixture of inorganic arsenic and methylated metabolites, making direct comparisons difficult. The mouse study does provide additional evidence for a transition in gene expression comparable to that noted in the review of the in vitro data in primary cells [Gentry et al., 2010], with a transition or threshold concentration on the order of 0.1 μM.

As mentioned previously, the most common pathways affected across individuals were similar to those noted in both the literature review and the mouse study, including several genes related to oxidative stress response (HMOX1), thioredoxin reductase (TXNRD1), thioredoxin (TXN), metallothionine regulation (MT1E), and to genes related to protein folding (FKBP5), DNA damage sensing (DDB2), cell adhesion, growth regulation (LGALS8), and immune response (THBD). BMDs based on the concentration of total arsenic were in the approximate range of 0.1 to 1.0 μM (0.13–0.92 μM for trivalent; 0.53–2.7 μM for pentavalent), consistent with a critical concentration in this range.

BMDs showed an approximate threefold difference for each gene across all individuals. A much larger suite of dissimilar genes were also simultaneously expressed thus variation in response may be greater than the estimates based on these eight commonly significantly expressed genes. Results from previously reviewed studies combined with evidence herein provide strong evidence for a threshold and in some cases reversal in direction of effect in the vicinity of 0.1 to 1.0 μM as examination of NOSTASOT results (Table X) indicate. While a BMD of 0.1 μM is a reasonably high concentration, it is nonetheless in the range of exposure concentrations for highly exposed human populations.

Suggestion for reversal in direction of effect, although not statistically significant, is seen when comparing response at the lowest treatment concentration of 0.18 μM total arsenic with gene expression response at 0.6 μM total arsenic for individual genes (Table X, Figs. 5e and 5h.). Up-regulation of DNA repair gene expression in the presence of low arsenite concentrations has been observed in in vitro studies [Snow et al., 2005; Wnek et al., 2011] and a population study in which up-regulation of DNA repair genes was positively associated with arsenic concentration in drinking water [Mo et al., 2009]; the authors interpreted these findings as indicating an adaptive (protective) response.

BMDLs for the eight gene biomarkers were all in the range of 0.09 to 0.58 μM for total arsenic for trivalent treatments and 0.35 to 1.7 μM based on the concentration of total arsenic for pentavalent treatments (Table IX). The lowest BMDL value in both instances is above the lowest treatment concentration. Therefore, the range of concentrations employed in these experiments is informative to estimate concentrations at which no gene transcriptional effects would be expected.

In addition to a transition in expression with increasing concentrations of arsenic, previous studies suggest changes in expression with duration of exposure. Studies conducted by Chouchane and Snow [2001], Hu et al. [2002], and Sykora and Snow [2008] evaluated changes in gene expression over exposure duration in GM847 fibroblast cells. Cells were incubated with a concentration of approximately 0.1 μM arsenite for 24 to 72 hr or for 7 to 150 days. Following the acute exposures, increases in gene or protein expression associated with oxidative stress (i.e., AP-1, NF-κB) were noted. However, as duration of exposure increased (greater than 10 weeks), a transition in changes in gene expression from increases to decreases were noted in most of the genes/proteins evaluated. The authors suggested that the effects observed following both short-term and long-term administration of arsenic indicate an impact on redox-sensitive signal transduction pathways. Induction of transcription factors following low-dose short-term exposure suggests an early response to stressors on the cellular system. Transition from induction to regression for selected genes/proteins indicates a progression that suggests adaptation over time.

Exposure duration related changes in gene expression were also noted in mice following drinking water exposure for 1 or 12 weeks [Clewell et al., 2011]. Expression of genes following 1 week of arsenite exposure was primarily down-regulated, while at 12 weeks expression was primarily up-regulated. Down-regulation was greatest at the highest concentration at week 1, and up-regulation was greatest at the highest concentration at week 12. There was little overlap in the genes affected at the two time-points; the number of genes in common between 1 and 12 weeks were 1, 0, 47, and 34 at 0.5, 2, 10, and 50 mg As/L, respectively.

Recent studies demonstrate the importance of consideration of not only concentration, but also duration of exposure in the application of changes in gene expression in evaluating the potential mode of action for the carcinogenicity of arsenic compounds. While this current study in human primary uroepithelial cells suggests concentrations that may be associated with a threshold or transition of effects, cell treatment was limited to a 24 hr duration. Results from cited studies provide evidence of changes in gene expression with duration of exposures; therefore studies of gene expression changes in human uroepithelial cells following chronic exposure over time are needed to determine changes in gene expression under such exposure conditions.

A significant quantity of research has been completed or is underway in both animals and in vitro to investigate the potential carcinogenic mode of action of arsenic. When assimilated, these results, taken as whole, provide support for a shift in the approach for conducting a cancer risk assessment for arsenic that is in contrast to the standard risk assessment approach for carcinogens. The data strongly support a nonlinear dose response. Further, these data strengthen the concept of a potential threshold or a concentration demonstrating a dose-dependent transition in response from those representing adaptive change to those that may be key events in the development of cancer itself. Integration of in vitro and in vivo experimental research results support a threshold or dose-dependent transition concentration of approximately 0.1 μM arsenic. While additional information is nevertheless needed, particularly on the effect of exposure duration on gene expression changes, integration of these research results provides considerable insight for a new quantitative approach to cancer risk assessment as an alternative to other methods such as application of a biologically-based dose response (BBDR) arsenic cancer model.

A 2007 NRC-NAS report [NRC, 2007] suggested that population variability linearizes the dose response; however, it is more likely that variation in sensitivity among the target population broadens, but does not linearize the dose–response relationship. The NRC-NAS report also pointed out that interindividual variability is an area that is not being extensively investigated. This study in which primary uroepithelial cells from multiple normal human donors were treated with a wide range of arsenical concentrations does provide initial insight into interindividual variability in gene expression response. Here we show that for genes expressed in nearly all individuals, interindividual response varies by a factor of approximately 3.

Dose-dependent transitions in the vicinity of 0.1 μM arsenic were observed in the expression of key genes in primary normal human uroepithelial cells treated in vitro with biologically plausible concentrations of an arsenical mixture representative of that observed in urine of individuals exposed to arsenic in drinking water. Interindividual variability in gene expression response was elucidated. These data strengthen the concept of a potential threshold or a concentration demonstrating a dose-dependent transition in response and provide additional support for an alternative approach to arsenic cancer risk assessment that is in contrast to the standard method.

Author Contributions

  1. Top of page
  2. Abstract
  7. Author Contributions
  8. Acknowledgements
  10. Supporting Information

Drs. Yager, Gentry, and Clewell designed the study and provided overall study management. Dr. Thomas provided key expertise in molecular toxicology; Ms. Pluta conducted all microarray analyses. Dr. Yager prepared the manuscript draft with important intellectual input from Drs. Gentry, Clewell, and Thomas. Mr. Black conducted statistical analyses of microarray data. Ms. Efremenko carried out statistical and enrichment analyses of microarray data and prepared graphical and tabular display of results. Ms. Arnold coordinated attainment and transport of ureter tissue under informed consent. Dr. McKim and Mr. Wilga established human primary ureter cell culture conditions and carried out cell treatment protocols and biochemical analyses. Drs. Gill and Choe contributed arsenic analytical chemistry capability and carried out arsenic speciation analyses of cell culture supernatants. All authors approved the final manuscript.


  1. Top of page
  2. Abstract
  7. Author Contributions
  8. Acknowledgements
  10. Supporting Information

The authors thank Drs. Samuel M. Cohen and Brian Stevens, and Anna M. Kellogg, R.N., B.S.N., M.S., Nicole Miller, and Kristie DeHaai for their invaluable help in obtaining ureter tissue. The authors are indebted to Cynthia Van Landingham, M.S. and Bruce Allen, Ph.D. of ENVIRON, International for expert data analyses. The authors thank Drs. Samuel M. Cohen, Toby Rossman, Aaron Barchowsky, and Barbara Beck for helpful review of original study protocols.


  1. Top of page
  2. Abstract
  7. Author Contributions
  8. Acknowledgements
  10. Supporting Information
  • AnkleyGT,DastonGP,DegitzSJ,DenslowND,HokeRA,KennedySW,MiracleAL,PerkinsEJ,SnapeJ,TillittDE,TylerCR,VersteegD. 2006.Toxigenomics in regulatory ecotoxicology. Environ Sci Technol 40:40554065.
  • ATSDR. 2007.Toxicological Profile for Arsenic.Atlanta, GA: United States Department of Health and Human Services. Public Health Service. Agency for Toxic Substances and Disease Registry. Available at:
  • BednarAJ,KirganRA,JonesWT. 2009.Comparison of standard and reaction cell inductively coupled plasma mass spectrometry in the determination of chromium and selenium species by HPLC-ICP-MS. Anal Chem Acta 632:2734.
  • BenjaminiY,HochbergY. 1995.Controlling the false discovery rate: A practical and powerful approach to multiple testing.J R Stat Soc Ser B 57 760:289300.
  • BentonMA,RagerJE,SmeesterL,FryRC. 2011.Comparative genomic analyses identify common molecular pathways modulated upon exposure to low doses of arsenic and cadmium.BMC Genomics 12:173183.
  • BranchS,EbdonL,O'NeillP. 1994.Determination of arsenic species in fish by directly coupled high-performance liquid chromatography-inductively coupled plasma mass spectrometry.J Anal At Spectrom 9:3337.
  • CalabreseEJ,BachmannKA,BailerAJ,BolgerPM,BorakJ,CaiL,CedergreenN,CherianMG,ChiuehCC,ClarksonTW,CookRR,DiamondDM,DoolittleDJ,DoratoMA,DukeSO,FeinendegenL,GardnerDE,HartRW,HastingsKL,HayesAW,HoffmannGR,IvesJA,JaworowskiZ,JohnsonTE,JonasWB,KaminskiNE,KellerJG,KlaunigJE,KnudsenTB,KozumboWJ,LettieriT,LiuSZ,MaisseuA,MaynardKI,MasoroEJ,McClellanRO,MehendaleHM,MothersillC,NewlinDB,NiggHN,OehmeFW,PhalenRF,PhilbertMA,RattanSI,RiviereJE,RodricksJ,SapolskyRM,ScottBR,SeymourC,SinclairDA,Smith-SonnebornJ,SnowET,SpearL,StevensonDE,ThomasY,TubianaM,WilliamsGM,MattsonMP. 2007.Biological stress response terminology: integrating the concepts of adaptive response and preconditioning stress within a hormetic dose-response framework.Toxicol Appl Pharmacol 222:122128.
  • ChiouHY,ChiouST,HsuYH,ChouYL,TsengCH,WeiML,ChenCJ. 2001.Incidence of transitional cell carcinoma and arsenic in drinking water: A follow-up study of 8,102 residents in an arseniasis-endemic area of northeastern Taiwan.Am J Epidemiol 153:411418.
  • ChouchaneS,SnowET. 2001.In vitro effect of arsenical compounds on glutathione-related enzymes.Chem Res Toxicol 14:517522.
  • ClewellHJ,ThomasRS,KenyonEM,HughesMF,AdairBM,GentryPR,YagerJW. 2011.Concentration- and time-dependent genomic changes in the mouse urinary bladder following exposure to arsenate in drinking water for up to 12 weeks.Toxicol Sci 123:421432.
  • CohenSM,ArnoldLL,EldanM,LewisAS,BeckBD. 2006.Methylated arsenicals: The implications of metabolism and carcinogenicity studies in rodents to human risk assessment.Crit Rev Toxicol 36:99133.
  • CohenSM,OhnishiT,ArnoldLL,LeXC. 2007.Arsenic-induced bladder cancer in an animal model.Toxicol Appl Pharmacol 222:258263.
  • DingW,HudsonLG,LiuKJ. 2005.Inorganic arsenic compounds cause oxidative damage to DNA and protein by inducing ROS and RNS generation in human keratinocytes.Mol Cell Biochem 279:105112.
  • EbleJN, SauterG, EpsteinJI, SesterhennIA, editors. 2004.WHO Classification of Tumours: Tumours of the Urinary System and Male Genital Organs.Lyon, France: International Agency for Research on Cancer, IARC Press.
  • EFSA. 2009.Scientific opinion on arsenic in food. EFSA Panel on Contaminants in Food Chain (CONTAM). European Food Safety Authority, Parma, Italy.EFSA J 7:1351.
  • Escudero-LourdesC,MedeirosMK,Cardenas-GonzalezMC,WnekSM,GandolfiJA. 2010.Low level exposure to monomethylarsonous acid induced the over-production of inflammation-related cytokines and the activation of cell signals associated with tumor progression in a urothelial cell model.Toxicol Appl Pharmacol 244:162173.
  • FloraSJ. 2011.Arsenic-induced oxidative stress and its reversibility.Free Radic Biol Med 51:257281.
  • GentryPR,McDonaldTB,SullivanDE,ShippAM,YagerJW,ClewellHJ III. 2010.Analysis of genomic dose-response information on arsenic to inform key events in a mode of action for carcinogenicity.Environ Mol Mutagen 51:114. Review.
  • GriffithOW. 1980.Determination of glutathione and glutathione disulfide using glutathione reductase and 2-vinylpyridine.Anal Biochem 106:207212.
  • HartwigA,AsnaussM,EhlebenI,HerzerU,KostelacD,PelzerA,SchwerdtleB,BurkleA. 2002.Interference by toxic metal ions with DNA repair processes and cell cycle control: Molecular mechanisms.Environ Health Perspect Suppl 5:797799.
  • HeitkemperDT,VelaNP,StewartKR,WestphalCS. 2001.Determination of total and speciated arsenic in rice by ion chromatography and inductively coupled plasma mass spectrometry.J Anal At Spectrom 16:299306.
  • HelbockHJ,BeckmanKB,AmesBN. 1999.8-Hydroxy-deoxyguanosine and 8-hydroxyguanine as biomarkers of oxidative DNA damage.Methods Enzymol 300:156166.
  • HuY,JinX,SnowET. 2002.Effect of arsenic on transcription factor AP-1 and Nf-kappaB DNA binding activity and related gene expression.Toxicol Lett 133:3345.
  • IARC. 2012.Review of Human Carcinogens. Part C. Arsenic, Metals, Fibres and Dusts. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 100. Lyon, France: International Agency for Research on Cancer. pp 79.
  • IrizarryRA,BolstadBM,CollinF,CopeLM,HobbsB,SpeedTP. 2003.Summaries of Affymetrix GeneChip probe level data.Nucleic Acids Res 31:e15.
  • KaragasMR,TostesonTD,MorrisJS,DemidenkoE,MottLA,HeaneyJ,SchnedA. 2004.Incidence of transitional cell carcinoma of the bladder and arsenic exposure in New Hampshire.Cancer Causes Control 15:465472.
  • KasaiH. 1997.Analysis of a form of oxidative DNA damage, 8-hydroxy-2'-deoxyguanosine, as a marker of cellular oxidative stress during carcinogenesis.Mutat Res 387:147163.
  • KennedyRE,KernsRT,KongX,ArcherKJ,MilesMF. 2006.SScore: An R package for detecting differential gene expression without gene expression summaries.Bioinformatics 22:12721274.
  • KenyonEM,HughesMF,AdairBM,HighfillJM,CreceliusEA,ClewellHJ,YagerJW. 2008.Tissue distribution and urinary excretion of inorganic arsenic and its methylated metabolites in C57BL6 mice following subchronic exposure to arsenate in drinking water.Toxicol Appl Pharmacol 232:448455.
  • KernsRT,ZhangL,MilesMF. 2003.Application of S-score algorithm for analysis of oligonucleotide microarrays.Methods 31:274281.
  • KetavarupuS,YathavakillaV,FrickeM,CreedPA,HeitkemperDT,ShockeyNV,SchegelC,CarusoJA. CreedJT. 2008.Arsenic speciation and identification of monomethylarsonous acid and monomethylthioarsonic acid in a complex matrix.Anal Chem 80:775782.
  • KitchinKT. 2001.Recent advances in arsenic carcinogenesis: Modes of action, animal model systems, and methylated arsenic metabolites.Toxicol Appl Pharmacol 172:249261.
  • KitchinKT,AhmadS. 2003.Oxidative stress as a possible mode of action for arsenic carcinogenesis.Toxicol Lett 137:313.
  • KitchinKT,WallaceK. 2005.Arsenite binding to synthetic peptides based on the Zn finger region and the estrogen binding region of the human estrogen receptor-alpha.Toxicol Appl Pharmacol 206:6672.
  • KitchinKT,WallaceK. 2008.Evidence against the nuclear in situ binding of arsenicals—oxidative stress theory of arsenic carcinogenesis.Toxicol Appl Pharmacol 232:252257.
  • KleinCB,LeszczynskaJ,HickeyC,RossmanTG. 2007.Further evidence against a direct genotoxic mode of action for arsenic-induced cancer.Toxicol Appl Pharmacol 222:289–297.
  • KongXT,DengFM,HuP,LiangFX,ZhouG,AuerbachAB,GenieserN,NelsonPK,RobbinsES,ShapiroE,KacharB,SunTT. 2004.Roles of uroplakins in plaque formation, umbrella cell enlargement, and urinary tract diseases.J Cell Biol 167:11951204.
  • Lerda, D. 1994.Sister chromatid exchange (Sce) among individuals chronically exposed to arsenic in drinking water.Mutat Res 312:111120.
  • LiJH,RossmanTG. 1989.Inhibition of DNA ligase activity by arsenite: A possible mechanism of its comutagenesis.Mol Toxicol 2:19.
  • LiuQ,ZjangH,SmeesterL,ZouF,KesicM,JaspersI,PiJ,FryRC. 2010.The NRF2-mediated oxidative stress response pathway is associated with tumor cell resistance to arsenic trioxide across the NCI-60 panel.BMC Med Genomics 3:37.
  • MandalBK,OgraY,SuzukiKT. 2001.Identification of dimethylarsenious and monomethylarsonous acids in human urine of the arsenic-affected areas in west Bengal, India.Chem Res Toxicol 14:371378.
  • MeiN,KunugitaN,HiranoT,KasaiH. 2002.Acute arsenite-induced 8-hydroxyguanine is associated with inhibition of repair activity in cultured human cells.Biochem Biophys Res Commun 297:924930.
  • MoJ,XiaY,NingZ,WadeTJ,MumfordJ. 2009.Elevated ERCC1 gene expression in blood cells associated with exposure to arsenic from drinking water in Inner Mongolia.Anticancer Res 29:32533259.
  • NRC-NAS. 2007.Toxicity Testing in the 21st Century: A Vision and a Strategy.Washington, DC: National Research Council, National Academy of Sciences.
  • PetzoldtJL,LeighIM,DuffyPG,SextonC. 1995.Immortalization of human urothelial cells.Urol Res 23:377380.
  • PiJ,DiwayBA,SunY,LiuJ,QuW,HeY,StybloM,WaalkesMP. 2008.Arsenic-induced malignant transformation of human keratinocytes: Involvement of Nrf2.Free Radic Biol Med 45:651658.
  • QinX-J,HudsonLG,LiuW,DingW,CooperKL,LiuKJ. 2008.Dual actions involved in arsenite-induced oxidative DNA damage.Chem Res Toxicol 21:18061813.
  • The R Core Team. 2012.R: A Language and Environment for Statistical Computing.Vienna, Austria: R Foundation for Statistical Computing. ISBN 3-900051-07-0. Available at: http://www.R-
  • RoopDR,Hawley-NelsonP,ChengCK,YuspaSH. 1983.Keratin gene expression in mouse epidermis and cultured epidermal cells.Proc Natl Acad Sci USA 80:716720.
  • RossmanTG. 2003.Mechanisms of arsenic carcinogenesis: an integrated approach.Mutat Res 533:3765.
  • SnowET,SykoraP,DurhamTR,KleinCB. 2005.Arsenic, mode of action at biologically plausible low doses: What are the implications for low dose cancer risk?.Toxicol Appl Pharmacol 207 (2 Suppl):557564.
  • SouthgateJ,HuttonKA,ThomasDF,TrejdosiewiczLK. 1994.Normal human urothelial cells in vitro: proliferation and induction of stratification.Lab Invest 71:583594.
  • SouthgateJ,VarleyCL,GarthwaiteMAE,HinleyJ,MarshF,StahlschmidtJ,TrejdosiewiczLK,EardleyI. 2007.Differentiation potential of urothelium from patients with benign bladder dysfunction.Br J Urol 99:15061516.
  • StraifK,Benbrahim-TallaaL,BaanR,GrosseY,SecretanBE,GhissassiF,BourardV,GuhaN,FreemanC,GalichetL,CoglianoV, the Working Group WHO/IAEC. 2009.A review of human carcinogens. Part C. Metals, arsenic, dust and fibres.Lancet Oncol 10:453454.
  • SykoraP,SnowET. 2008.Modulation of DNA polymerase beta-dependent base excision repair in cultured human cells after low dose exposure to arsenite.Toxicol Appl Pharmacol 228:385394.
  • TokarEJ,Benbrahim-TallaaL,WardJM,LunnR,SamsRLII,WaalkesMP. 2010.Cancer in experimental animals exposed to arsenic and arsenic compounds.Crit Rev Toxicol 40:912927.
  • TokarEJ,DiwanBA,WardJM,DelkerDA,WaalkesMP. 2011.Carcinogenic effects of whole life exposure to inorganic arsenic in CD1 mice.Toxicol Sci 119:7383.
  • TukeyJW,CimineraJL,HeyseJF. 1985.Testing the statistical certainty of a response to increasing doses of a drug.Biometrics 41:295301.
  • VarleyCL,StahlschmidtJ,LeeW-C,HolderJ,DiggleC,SelbyPJ,TrejdosiewiczLK,SouthgateJ. 2003.Role of PPARγ and EGFR signaling in the urothelial terminal differentiation programme.J Cell Sci 117:20292036.
  • WangX-J,SunZ,ChenW,EblinKE,GandolfiJA,ZhangDD. 2007a. Nrf2 protects human bladder urothelial cells from arsenite and monomethylarsonous acid toxicity.Toxicol Appl Pharmacol 225:206213.
  • WangR-Y,HsuY-L,ChangL-F,JiangS-J. 2007b. Speciation analysis of arsenic and selenium compounds in environmental and biological samples by ion-chromatography-inductively coupled plasma dynamic reaction cell mass spectrometer.Anal Chem Acta 590:239244.
  • WeiM,ArnoldLL,CanoM,CohenSM. 2005.Effect of co-administration of antioxidants and arsenicals on the rat urinary bladder epithelium.Toxicol Sci 83:237245.
  • WnekSM,MedeirosMK,EblinKE,GandolfiAJ. 2009.Persistence of DNA damage following exposure of human bladder cells to chronic monomethylarsonous acid.Toxicol Appl Pharmacol 241:202209.
  • WnekSM,KuhlmanCL,CamarilloJM,MedeirosMK,LiuKJ,LauSS,GandolfiAJ. 2011.Interdependent genotoxic mechanisms of monomethylarsonous acid: role of ROS-induced DNA damage and poly (ADP-ribose) polymerase-1 inhibition in the malignant transformation of urothelial cells.Toxicol Appl Pharmacol 257:113.
  • YagerJW,WienckeJK. 1993.Enhancement of chromosomal damage by arsenic: Implications for mechanism.Environ Health Perspect 101(Suppl 3):7982.
  • YagerJW,WienckeJK. 1997.Inhibition of poly (ADP-ribose) polymerase by arsenite.Mutat Res 386:345351.
  • YokohiraM,ArnoldLL,PenningtonKL,SuzukiS,Kakiuchi-KiyotaS,Herbin-DavisK,ThomasDJ,CohenSM. 2011.Effect of sodium arsenite dose administered in the drinking water on the urinary bladder epithelium of female arsenic (+3 oxidation state) methyltransferase knockout mice.Toxicol Sci 121:257266.
  • ZhangL,WangL,RavindranathanA,MilesMF. 2002.A new algorithm for analysis of oligonucleotide arrays: Application to expression profiling in mouse brain regions.J Mol Biol 317:225235.
  • ZhouX,SunX,CooperKL,WangF,LiuKJ,HudsonLG. 2011.Arsenite interacts selectively with zinc finger proteins containing C3H1 or C4 motifs.J Biol Chem 286:2285522863.

Supporting Information

  1. Top of page
  2. Abstract
  7. Author Contributions
  8. Acknowledgements
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

EM_21749_sm_SuppFig1.TIF62KSupporting Information Figure 1
EM_21749_sm_SuppFig2.TIF514KSupporting Information Figure 2
EM_21749_sm_SuppFig3.TIF27KSupporting Information Figure 3
EM_21749_sm_SuppInfo.docx12KSupporting Information
EM_21749_sm_SuppTab1.docx12KSupporting Information Table 1

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.