Dimethylarsinic Acid in Drinking Water Changed the Morphology of Urinary Bladder but Not the Expression of DNA Repair Genes of Bladder Transitional Epithelium in F344 Rats

Abstract

Inorganic arsenic increases urinary bladder transitional cell carcinoma in humans. In F344 rats, dimethylarsinic acid (DMA[V]) increases transitional cell carcinoma. Arsenic-induced inhibition of DNA repair has been reported in cultured cell lines and in lymphocytes of arsenic-exposed humans, but it has not been studied in urinary bladder. Should inhibition of DNA damage repair in transitional epithelium occur, it may contribute to carcinogenesis or cocarcinogenesis. We investigated morphology and expression of DNA repair genes in F344 rat transitional cells following up to 100 ppm DMA(V) in drinking water for four weeks. Mitochondria were very sensitive to DMA(V), and swollen mitochondria appeared to be the main source of vacuoles in the transitional epithelium. Real-time reverse transcriptase polymerase chain reaction (Real-Time RT PCR) showed the mRNA levels of tested DNA repair genes, ataxia telangectasia mutant (ATM), X-ray repair cross-complementing group 1 (XRCC1), excision repair cross-complementing group 3/xeroderma pigmentosum B (ERCC3/XPB), and DNA polymerase β (Polβ), were not altered by DMA(V). These data suggested that either DMA(V) does not affect DNA repair in the bladder or DMA(V) affects DNA repair without affecting baseline mRNA levels of repair genes. The possibility remains that DMA(V) may lower damage-induced increases in repair gene expression or cause post-translational modification of repair enzymes.

Introduction

People, especially smokers, exposed to inorganic arsenic in drinking water have a higher risk of urinary bladder transitional cell carcinoma than the general population (Karagas et al. 2004; Steinmaus et al. 2003). While inorganic arsenite (As[III]) and arsenate (As[V]) are the main forms of arsenic in drinking water, humans are also exposed to methylated arsenic generated through the metabolism of inorganic arsenic in the body. These metabolites include monomethylarsonic acid (MMA[V]), monomethylarsonous acid (MMA[III]), dimethylarsinic acid (DMA[V]), and dimethylarsinous acid (DMA[III]). The use of herbicides containing DMA(V) and the consumption of seafood containing DMA(V) or arsenosugars that can be metabolized to DMA(V) are also potential sources of human exposure to DMA(V) (Adair et al. 2007). Arsenic is mainly excreted in urine, and DMA(V) is a major metabolite detected in the urine of humans drinking arsenic-contaminated water (Vahter and Concha 2001).

The only laboratory animal model in which arsenic acts as a complete urinary bladder carcinogen is F344 rats exposed to DMA(V) (Cohen et al. 2007; Wang et al. 2002). After two years of exposure to DMA(V) at 40 ppm or higher concentrations via diet or drinking water, F344 rats showed increased incidences of urinary bladder transitional cell carcinoma (Arnold et al. 2006; van Gemert and Eldan 1998; Wei et al. 2002). Studies using F344 rats and DMA(V) showed that the mode of action for arsenic-induced urinary bladder transitional cell carcinoma involves regenerative proliferation of transitional cells following DMA(III)-induced cell death in the urothelium (Cohen et al. 2007). Other factors that increase the chance of stable genetic errors in normal proliferating cells may be involved as well. For example, DMA(V)-induced oxidative stress could damage DNA in the bladder and increase the chance of mutation and cancer in the long term. Similarly, inhibition of DNA damage repair could contribute to genomic instability and carcinogenesis.

Arsenic-induced decreases in DNA repair and in the expression of DNA repair genes have been reported in cultured cells, mouse skin, and lymphocytes collected from arsenic-exposed humans (Andrew et al. 2003; Hartwig et al. 2003; Wu et al. 2005). However, whether arsenic affects DNA repair in the urinary bladder is not clear, because arsenic effects on DNA damage and repair appear to be cell-type specific (Evans et al. 2004; Fischer et al. 2005), and arsenic effects on DNA repair have not been previously studied in an animal model that develops bladder cancer after arsenic exposure.

In vitro studies showed that base excision repair, nucleotide excision repair, and the repair of DNA double-strand breaks were all affected by arsenic (Hartwig et al. 1997; Lynn et al. 1997; Maier et al. 2002), and several mechanisms of arsenic-induced inhibition on DNA repair have been proposed. Possible mechanisms include: (1) direct inhibition of DNA repair enzymes by trivalent methylated arsenic (Schwerdtle et al. 2003; Walter et al. 2007); (2) indirect inhibition of DNA repair enzymes through reactive oxygen or nitrogen species (Bau et al. 2001; Chien et al. 2004); and (3) alterations in the expression of DNA repair genes (Andrew et al. 2006; Hamadeh et al. 2002). Alterations in signal pathways (Vogt and Rossman 2001) and post-translational modification on proteins involved in DNA repair (Yuan et al. 2002) also have been reported.

The first objective in the present study was to determine whether DMA(V) in drinking water causes ultrastructural changes in the urothelium of F344 rats. Previously published studies showed hyperplasia, necrosis, and exfoliation in the urothelium of rats exposed to DMA(V) in their feed (Cohen et al. 1998; Cohen et al. 2002), but morphological changes were not evaluated at the organelle level. Recently, mitochondria have been shown to play an important role in arsenic-induced toxicity, including oxidative stress, apoptosis, and genotoxicity (Liu et al. 2005; Pourahmad et al. 2003; Santra et al. 2007). With a focus on the mitochondria, the present study investigated DMA(V) effects on the organelles in the urothelium using transmission electron microscopy, light microscopy, and scanning electron microscopy. The ultrastructural changes of the urothelium are provided in detail for the first time.

The second objective was to test the hypothesis that DMA(V) affects the expression of DNA repair genes in rat urinary bladder transitional cells. We harvested mRNA from urinary bladder transitional cells (the target cell of DMA[V] carcinogenesis in the rat) and measured the expression of DNA repair genes by real-time reverse transcriptase polymerase chain reaction (real-time RT PCR). Ataxia telangectasia mutant (ATM), X-ray repair cross-complementing group 1 (XRCC1), and excision repair cross-complementing group 3/xeroderma pigmentosum B (ERCC3/XPB) were tested in this study because their expressions were affected by inorganic arsenic exposure in other cell types (Andrew et al. 2003; Bae et al. 2002). DNA polymerase β (Polβ) was tested because the ligation step in DNA repair was inhibited by inorganic arsenic, but isolated Polβ protein was insensitive to arsenic (Hu et al. 1998). The present study adds understanding of DMA(V) effects in the urinary bladder, particularly on DNA damage repair inhibition as a critical feature of carcinogenesis.

Materials and Methods

Chemicals

Dimethylarsinic acid, in the form of sodium cacodylatetrihydrate [(CH₃)₂AsO₂Na·3H₂O], (purity of 99.52%, with inorganic arsenic 0.0030%) (CAS no. 6131-99-3) was purchased from Electron Microscopy Science (Fort Washington, PA, USA). Trizol solution was from Invitrogen (Carlsbad, CA, USA).

Animals and Treatments

Prior to the initiation of the study, animal use and procedures were approved by the Virginia Tech Institutional Animal Care and Use Committee for Animals Used in Research and Testing. Female Fischer 344 rats were purchased from Harlan (Madison, WI, USA) at the age of six weeks. Female rats were used because they are more sensitive than males to DMA(V)-induced urothelium toxicity (Arnold et al. 2006; Cohen et al. 2001; Cohen et al. 2002; Shen et al. 2006). Rats were uniquely identified by ear tags after one week of quarantine. The rats were single-housed in polycarbonate shoebox-style cages with long-fiber paper bedding. The temperature and humidity were monitored continuously, and a twelve-hour light-dark cycle was maintained. Teklad 2018 SC diet and water in plastic water bottles with stainless steel sipper tubes with stoppers was available ad libitum. Since chlorination byproducts have been suspected as a human carcinogen for the urinary bladder, chlorinated tap water, instead of deionized or distilled water, was used to simulate human exposure (Huff et al. 1998) for all controls and DMA(V)-exposed rats. Arsenic was not detected in either the diet or tap water (detection limits: 0.10 ppm in diet and 0.002 ppm in water). Selenium, an antagonist to arsenic toxicity (Wang et al. 2006; Zeng et al. 2005), was 0.16 ppm in diet and was not detected in water (detection limit: 0.01 ppm).

After one week of acclimation, rats were assigned into experiments by body weight stratification. Twenty rats were assigned to the morphology study, and thirty rats were assigned to the study of expression of DNA repair genes. The body weight stratification was done by randomly assigning two of the five heaviest rats for morphology, and the same process was repeated with the next five heaviest rats until all rats were assigned. Within each experiment, rats were completely randomly divided into five treatment concentrations: 0, 1, 4, 40, or 100 ppm of DMA(V).

Rats were given DMA(V) in drinking water for four weeks. While the DMA(V) was administered in the drinking water, the concentrations in our study corresponded to concentrations that were administered in the diet in a two-year study (van Gemert and Eldan 1998), in which 40 and 100 ppm produced bladder hyperplasia and tumors in rats. The concentrations in our study also overlapped the concentration range of DMA(V) in drinking water in another two-year study (Wei et al. 1999), in which 50 and 100 ppm increased bladder cancer in rats. The DMA(V) solution prepared with tap water once a week and stored at room temperature. Fresh water with or without DMA(V) was provided three times a week, and water consumption was measured three times a week at each water change. Food consumption and body weight were measured weekly when the rats were transferred to clean cages with fresh bedding.

Terminal Necropsy

All rats survived the four-week treatment without overt signs of toxicity. Prior to the necropsy, food and water were available until the rats were removed from the animal room to the necropsy room. Moving the rats occurred shortly before the beginning of sacrifice at 9:00 AM, which was two hours after the light was turned on in the animal room. Rats were euthanized by carbon dioxide asphyxia, and the bladders were removed within two minutes after death. (Note that the bladder should ideally be fixed in situ while the rat is under anesthesia, because electron microscopic changes of the bladder epithelium can occur within sixty seconds of the death of a rat.) Bladders from rats assigned for the gene expression study were stripped by Trizol solution for mRNA extraction immediately after the removal of the bladders. Bladders from rats assigned for the morphology study were cut and fixed within two minutes after the removal of the bladder.

Morphological Examination for Urinary Bladder

Each bladder was cut into approximate halves longitudinally, bisecting the trigone, at the time of removal. One half (for light microscopy) was fixed in 10% neutral buffered formalin, and the other half (for transmission electron microscopy [TEM] and scanning electron microscopy [SEM]) was dissected and fixed in a cold general fixative for electron microscopy samples (4.4% formaldehyde/5% glutaraldehyde/2.75% picric acid, in a 0.05M sodium cacodylate buffer at pH 7.4).

Each half bladder fixed in 10% neutral buffered formalin was trimmed and submitted to Carilion Histology Laboratory (Roanoke, VA) to be processed into slides for routine histology. Briefly, the samples were dehydrated through a graded series of alcohols, infiltrated with paraffin polymer, and then sectioned at 3 μm. Sections were stained with hematoxylin and eosin on automated equipment.

For diagnostic purposes, the degrees of hyperchromatination and cytoplasmic vacuolation were scored separately on each slide. The slides were randomized for scoring sequence, and one person scored all slides without the knowledge of treatments at the time of scoring. A set of standard slides of various scores was used as a reference to ensure consistency. A score of 0 indicated no hyperchromatic cells or vacuolation at 400X magnification, and the score increased with the presence of hyperchromatic cells or vacuolation. A score of 1 indicated minimal hyperchromatination or vacuolation (few tiny vacuoles). A score of 2 indicated mild hyperchromatination or vacuolation (frequent tiny vacuoles). A score of 3 indicated mild hyperchromatination or vacuolation (few large vacuoles also present). A score of 4 indicated diffused hyperchromatination or vacuolation (frequent large vacuoles). A score of 5 indicated profound hyperchromatination or vacuolations observed throughout the whole epithelium.

The other halves of the bladders, designated for ultrastructural studies under TEM and SEM, were fixed in the general fixative for at least twenty-four hours, washed in 0.1M sodium cacodylate, post-fixed in 0.1M sodium cacodylate containing 1% osmium tetroxide (OsO₄), and washed in 0.1M sodium cacodylate. Dehydration was performed by placing samples in increasing concentrations of graded ethanol (15%, 30%, 50%, 70%, 95%, and 100%) for ten minutes each, followed by submersion in propylene oxide for fifteen minutes.

For TEM, samples were embedded in a 50:50 solution of propylene oxide:Poly/Bed 812. To determine the areas for thin sectioning, 1.0 μm sections were cut and tri-stained (methylene blue stain in sodium borate, Azure II stain in water, and basic fuchsin in sodium borate) (Humphrey and Pittman 1974). These sections were observed under a light microscope to identify areas with transitional epithelium. Selected areas were cut into ultrathin sections (60 nm thick), stained with uranyl acetate and lead citrate, and then examined with a Zeiss 10CA transmission electron microscope (Zeiss, Germany). Some TEM images were globally sharpened using Adobe Photoshop.

For SEM, samples were dried using a Ladd Critical Point Dryer, model 28000 (Ladd Research Industries, Burlington, VT, USA), mounted onto an SEM specimen stub, and sputter-coated with gold in an SPI-Module Sputter Coater (SPI Supplies, West Chester, PA, USA). The specimens were examined in a Philips Scanning Electron Microscope 505 (Philips, Eindhoven, Netherlands).

Trizol Stripping and mRNA Extraction of Urinary Bladder Transitional Cells

Selective collection of urinary bladder transitional cells for RNA extraction has been described (Sen et al. 2005). Within two minutes of the death of the rat, the bladder was ligated, rinsed, and then filled with cold Trizol solution for ten minutes. The cell lysate was aspirated, flash frozen, and stored at −70°C. The stripped bladder was fixed in formalin for light microscopic confirmation that only the urothelium was removed. Total RNA was extracted from the Trizol/cell lysate according to Trizol manufacturer’s protocols. Total RNA yield from each bladder ranged from 21.7 μg to 90.8 μg, based on absorbance at the 260 nm wavelength on a spectrophotometer.

Real-time RT PCR for Gene Expression

The ATM, ERCC3/XPB, XRCC1, and Polβ genes were chosen for expression study because either their expressions or their functions have been affected by arsenic in other cell types. Ataxia telangectasia mutant (involved in DNA double strand break repair) was critical in cellular response to As(III) in primary porcine aortic endothelial cells (Tsou et al. 2006). The mRNA levels of ERCC3/XPB (in nucleotide excision repair) and XRCC1 (in base excision repair) were reported to be affected by inorganic arsenic exposure (Andrew et al. 2003; Andrew et al. 2006; Bae et al. 2002; Bae et al. 2003). The expression of Polβ was suspected to be affected by arsenic because the ligation step of base excision repair was inhibited by As(III) exposure, whereas purified Polβ protein was insensitive to inorganic arsenic treatments (Hu et al. 1998; Lynn et al. 1997).

Real-time RT PCR was performed using the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA) according to the manufacturer’s instructions. In a 1-step RT PCR reaction, 200 ng of total RNA were subjected to cDNA synthesis and subsequently amplified during forty PCR cycles (48°C for thirty minutes, 95°C for ten minutes, followed by forty cycles of 95°C for fifteen seconds and 60°C for sixty seconds). Amplification reactions were carried out in triplicate using TaqMan One-Step RT-PCR Master Mix (ABI). Primer and probe sequences are listed in Table 1. TATA-box binding protein (TBP) was chosen as a reference gene because it demonstrated consistent expression across all treatment groups. The probe for TBP was labeled with the VIC reporter dye, and probes for the four genes of interest (ATM, ERCC3/XPB, Polβ, and XRCC1) were labeled with FAM (6-carboxyfluorescein). The primers and probes for TBP were designed using Primer Express version 2.0 software (Applied Biosystems). The primers and probes for XRCC1 and Polβ were from TaqMan Pre-Developed Assays for Gene Expression–Rat (ABI), and all other primers and probes were ABI Assays-on-Demand Gene Expression products.

Standard curves for each gene were constructed using three serial dilutions of starting samples (the relative standard curve method in ABI PRISM 7900HT Sequence Detection System User Bulletin #2). From the standard curves, the relative input for each gene was determined (i.e., normalized to the reference gene, which formerly was called a “housekeeping gene”). The average relative input from triplicates was designated as the expression level of each gene.

Statistical Analysis

For water consumption, food consumption, and body weight gain, the numbers were adjusted by body weight. The effects of arsenic concentration, time, and interaction of arsenic concentration and time on group means of body weight, body weight gain, food consumption, and water consumption were evaluated by repeated measure analysis and Wilk’s λ approach. The histological changes (vacuolation and hyperchromatin) of urinary bladder were analyzed using the Cochran-Mantel-Haenszel method, in which both lesion scores and DMA(V) treatments were ranked.

The effect of DMA(V) on gene expression was tested by multiple linear regression analysis with a maximal random likelihood model. To verify the association between accumulated arsenic exposure (expressed as DMA[V] intake per kg body weight) and gene expression levels, the expression of the reference gene (TBP) and the starting RNA amount for RT PCR were controlled in the model. The assumptions of multiple linear regression analysis were verified.

A p value less than .05 was considered significant. All statistical analyses were performed using a SAS computer program version 8.02 or 9.1 (SAS Institute Inc., Cary, NC, USA).

Results

General Conditions

All animals survived the treatment, and there were no clinical signs of intoxication. Up to 100 ppm DMA(V) did not affect body weight, body weight gain, or food consumption (data not shown). Fischer 344 rats exposed to 40 and 100 ppm DMA(V) increased their water consumption (Table 2 and Figure 1). For the whole four-week exposure time, the average daily water intakes were 15.0, 16.2, 15.2, 17.8, and 20.0 g/rat/day for rats exposed to 0, 1, 4, 40, and 100 ppm DMA(V), respectively. The daily water intake levels in our study were comparable to the reported average daily water intake of 18.4 to 20.7 g/rat/day for rats exposed to up to 200 ppm DMA(V) in water for two years (Wei et al. 1999). As a result of increased water intake, rats in the 100 ppm group had 4.6 times greater total arsenic intake as compared to those of the 40 ppm group, and 169 times those of the 1 ppm group (Table 2). For the whole four weeks of exposure, the estimated average daily DMA(V) intakes were 0, 1.0, 4.0, 46.7, and 130 μm/kg BW/day for rats exposed 0, 1, 4, 40, and 100 ppm DMA(V), respectively, and were comparable to that in the two-year study (Wei et al. 1999).

Histological Changes of Urinary Bladder

Dimethylarsinic acid exposure increased cytoplasmic vacuolation (Table 2) and nuclear hyperchromatin in the rat transitional epithelium. Each bladder was given two scores, one each for vacuolation or hyperchromatin, and a higher score indicated a higher quantity of the vacuolation or hyperchromatin. Although vacuolation and hyperchromatin were present in all DMA(V) treatment groups, the scores of vacuolation and hyperchromatin were highest in rats exposed to 100 ppm DMA(V) ( Figure 3). Furthermore, the scores were increased with DMA(V) concentration in a dose-dependent manner, as tested by the Cochran-Mantel-Haenszel method.

Ultrastructural Changes of Transitional Epithelium

Scanning Electron Microscopy:

Normal transitional epithelium with large, flat, polygonal superficial cells covered in micro-ridges were present in control rats and rats treated with up to 40 ppm DMA(V) (Figures 4A and 4C). In rats exposed to 100 ppm DMA(V), transitional epithelium had widely distributed sites of necrotic and exfoliated cells (Figure 4B), consistent with morphologic evidence of cytotoxicity. Superficial transitional cells from rats treated with 100 ppm DMA(V) were round, with variable size, and covered in polymorphic microvilli (Figure 4D). Although the bladders were fixed at various stages of expansion, which prevented the use of cobblestone-like appearance and elevation of cells as indicators of hyperplasia, the surface structures (microridges and microvilli) were not affected. Based on the polymorpholic microvilli observed, rats exposed to 100 ppm DMA(V) had signs of regeneration and hyperplasia (Fukushima et al. 1981).

Transmission Electron Microscopy:

The transitional epithelium is typically composed of one layer of superficial cells facing the bladder lumen, one to three layers of intermediate cells (which appear six to eight cells thick in a fully contracted bladder), and one layer of basal cells situated on top of the basement membrane.

Transitional epithelium from rats exposed to 1 or 4 ppm DMA(V) was not different from control (data not shown), but rats exposed to 40 or 100 ppm DMA(V) had increased vacuolation and decreased numbers of mitochondria in the transitional epithelium (Figure 5). Vacuolation was more prominent in the superficial cells than in the intermediate or basal cells.

Morphological changes to the mitochondria were variable and appeared to be the main source of the vacuoles identified by light microscopy. Mitochondria in transitional cells from rats treated with 40 ppm DMA(V) were swollen, with loss of cristae (Figures 5C and 5D). Transitional cells from rats treated with 100 ppm had almost no recognizable mitochondria in the superficial cells (Figures 5E and 5F).

Gene Expression

The expressions of the four tested target genes (ATM, ERCC3, Polβ, and XRCC1) were calculated by relative quantification using standard curves of the same primer and probe set. The relative expression of each gene to TBP, the reference gene, was plotted against DMA(V) intake, and the average of each DMA(V) treatment group was also graphed (Figure 6).

After controlling for TBP expression and total RNA, there was no association between gene expression and DMA(V) intake adjusted by body weight, as tested by multiple linear regression analysis. This finding indicated that the expression of ERCC3, ATM, Polβ, and XRCC1 in the bladder transitional cells was not altered by DMA(V).

Discussion

At concentrations that produced no apparent clinical toxicity, dimethylarsinic acid induced morphological changes in urinary bladder transitional epithelium, a target of arsenic carcinogenesis, but no alterations in the expression of DNA repair genes. Although vacuolation, hyperchromatin, and hyperplasia in the transitional epithelium have been reported (Cohen et al. 2001; Sen et al. 2005; Shen et al. 2006), the present study is the first to show organelle damage, particularly mitochondrial damage, after DMA(V) exposure in a dose-response manner. The mRNA levels of DNA repair genes involved in excision repair or double-strand break repair (ATM, ERCC3/XPB, Polβ, and XRCC1) were not altered by DMA(V) in transitional cells.

Rats exposed to up to 100 ppm DMA(V) showed no clinical signs of intoxication; however, water consumption was increased in rats exposed to 40 or more ppm of DMA(V). This finding is consistent with previous studies using F344 rats and DMA(V) (Arnold et al. 2006; Nishikawa et al. 2002; Shen et al. 2006; Wei et al. 1999; Wei et al. 2002). In addition to increases in water consumption and urine volume, increases in calcium excretion in the urine and calcification in the kidney (but not in the bladder) were reported in F344 rats given DMA(V) in Purina diet (Arnold et al. 1999). In the present study, as was true of rats given Altromin diet (Arnold et al. 1999), rats given Teklad 2018 SC diet showed increased water consumption and no calcification in the kidney after DMA(V) exposure (unpublished data). In contrast to F344 rats, Sprague-Dawley rats did not increase their water consumption when given DMA(V) in drinking water (unpublished data). Fischer 344 rats also did not increase water consumption when given MMA(V) or trimethylarsine oxide in drinking water (Nishikawa et al. 2002; Shen et al. 2006). The cause of DMA(V)-specific increases in water consumption in F344 rats remains unclear.

Exposure to DMA(V) increased vacuolation, hyperchromatin, and mitochondrial swelling in the rat urothelium. Superficial cells were more severely damaged than the intermediate or basal cells, suggesting that urine in the bladder lumen, rather than blood from the submucosa or muscle, may be the main source of DMA(V) exposure. Alternatively, superficial cells may be more sensitive than other cell layers to DMA(V) effects. The mitochondrial swelling in the bladder supports the hypothesis that arsenic specifically targets mitochondria, as in other cell types and purified mitochondria (Bustamante et al. 2005; Liu et al. 2005; Miller et al. 2007; Santra et al. 2007). Low concentrations of As(III) stimulated cytochrome c release from mitochondria and led to apoptosis, whereas high concentrations of As(III) directly targeted the mitochondrial respiratory chain and led to necrosis (Bustamante et al. 2005). In the present study, 100 ppm DMA(V) induced severe damage to mitochondria, likely rendering these organelles unable to support energy-dependent apoptosis. This situation could lead to necrosis, rather than apoptosis, consistent with previous reports of urothelial toxicity from high concentrations of DMA(V) (Cohen et al. 2001; Shen et al. 2006). Rats exposed to low concentrations of DMA(V) have increased apoptosis indices in the urothelium (Kinoshita et al. 2007). These data are consistent with previous descriptions of gene expression from rats similarly exposed (Sen et al. 2005), in that the animals treated with 100 ppm DMA(V) did not have prominent differential gene expression compared to lower doses due to greater toxicity and cell death.

Vacuoles formed from various membranous components, namely, mitochondria, endoplasmic reticula, and Golgi apparatus, can be indistinguishable. It is possible that other membranous components, in addition to mitochondria, also contributed to DMA(V)-induced vacuoles. For example, dilated rough endoplasmic reticula tend to degranulate and appear similar to dilated smooth endoplasmic reticula, vacuoles with no specific morphological characteristics. Because no intermediate degrees of swelling were observed in the endoplasmic reticula or the Golgi apparatus, it is unlikely that they were major sources of vacuoles in the present study. It is possible that postmortem changes in the bladder epithelium contributed to the observed morphological changes under TEM, because electron microscopic changes of the bladder epithelium can occur within sixty seconds of the death of a rat. The lack of mitochondrial changes in the control rats suggested that the electron microscopic changes of the bladder epithelium were likely to be from DMA(V) treatments in this study.

Increased hyperchromatin was present in the transitional epithelium of F344 rats exposed to DMA(V) in a dose-response manner. Because hyperchromatin indicates active production of DNA, it can be a sign of proliferation, which is consistent with our SEM observation. In mice, DMA(V)-induced hyperchromatin has been observed in the lung, but not in the liver (Nakano et al. 1992). The observed hyperchromatin in the lung was suspected to have been caused by DMA(V)-induced DNA-protein crosslinks (Nakano et al. 1992), which has been reported in DMA(V)-treated cultured cells (Kato et al. 1994; Yamanaka et al. 1993; Yamanaka et al. 1995). Arsenic-induced DNA-protein crosslinks in the urinary bladder have been reported in As(III)-exposed mice (Tice et al. 1997), but DNA-protein crosslinks were not observed by us in the urinary bladders of rats sacrificed one day after one week of exposure to 100 ppm DMA(V) in drinking water (Wang, in press).

Arsenic has been shown to inhibit DNA damage repair in vivo (Tran et al. 2002) and in vitro, including in cultured lung and skin cells (Schwerdtle et al. 2003; Wu et al. 2005). The decreased DNA repair may be due to decreased DNA repair gene expression. Recently, decreases in both DNA repair and expression of nucleotide excision repair genes (mRNA concentrations of ERCC1, XPF, and XPB; protein concentration of ERCC1) were observed in the lymphocytes of humans who drink arsenic-contaminated water in the United States of America (Andrew et al. 2003; Andrew et al. 2006). Decreased DNA repair gene expression was also seen in As(III)-treated human keratinocytes at the presence of oxidative stress and possibly oxidative DNA damage (Bae et al. 2002; Hamadeh et al. 2002).

To investigate whether DMA(V) targets DNA repair genes in the urinary bladder, we used real-time RT PCR to measured mRNA of ATM, ERCC3/XPB, XRCC1, and Polβ in transitional cells of DMA(V)-exposed rats. These genes were chosen because their expressions or functions were affected by arsenic in other cell types (see Materials and Methods). Our results showed that DMA(V) did not affect the mRNA levels of ATM, ERCC3/XPB, XRCC1, or Polβ in the transitional cells. The differences in arsenic effects on DNA repair gene expression between the present and previous studies may be due to the forms of arsenic and cell-type specific response to arsenic.

It is possible that DMA(V) does not affect excision repair in the rat urinary bladder. Cell-type specific responses to arsenic effects on DNA repair have been reported in As(III)-exposed mice. When mice were given As(III) in drinking water and benzo[a]pyrene topically on the skin, benzo[a]pyrene-induced DNA adducts were increased by As(III) exposure in the lung, but not in the skin (Evans et al. 2004). Moreover, when FVB/N mice carrying the G11 PLAP transgene were used to detect frame-shift mutations, mutation was increased in the skin (but not in the lung or the urinary bladder) of mice receiving both As(III) in water and benzo[a]pyrene on the skin (Fischer et al. 2005).

Alternatively, DMA(V) may affect DNA repair in the urinary bladder without affecting baseline mRNA expression of repair genes. For example, DMA(V) may only lower damage-induced increases in DNA repair gene expression in the bladder, similar to As(III) in keratinocytes (Bae et al. 2002). Furthermore, arsenic may affect DNA repair without altering gene expression at the mRNA level. Possible nontranslational influence of arsenic on DNA repair includes (1) indirect inhibition of repair enzymes through oxidative stress; (2) direct inhibition of enzymes by crosslinking or binding with enzymes, or releasing zinc from enzymes; and (3) post-translational modification of enzymes (see online supplemental material).

Exposure to DMA(V) in drinking water did not alter baseline mRNA levels of DNA repair genes in the transitional cells of F344 rats. Although these results did not provide supporting evidence for the hypothesis that DNA repair inhibition is a mode of action for DMA(V) carcinogenesis in the urinary bladder, our data cannot exclude the possibility that DMA(V) affects DNA repair in the bladder through other mechanisms, such as alterations in damage-induced DNA repair gene expression and modifications on repair enzymes. However, DMA(V) in drinking water induced vacuolation, hyperchromatin, and mitochondrial damage in rat urinary bladder transitional epithelium in a dose-response manner, suggesting that mitochondria are a target of DMA(V) toxicity.