Carcinogenesis vol.23 no.8 pp.1387–1397, 2002

Carcinogenicity of dimethylarsinic acid in male F344 rats andgenetic alterations in induced urinary bladder tumors

Min Wei1, Hideki Wanibuchi1, Keiichirou Morimura1,Shuji Iwai1, Kaoru Yoshida2, Ginji Endo2, Dai Nakae3

and Shoji Fukushima1,4

1Department of Pathology, Osaka City University Medical School, 1-4-3,Asahi-machi, Abeno-ku, Osaka 545-8585, 2Department of PreventiveMedicine and Environment Health, Osaka City University Medical School,1-4-3, Asahi-machi, Abeno-ku, Osaka 545-8585 and 3Department ofOncological Pathology, Cancer Center, Nara Medical University, Nara 634,Japan4To whom correspondence should be addressedEmail: fukuchan@med.osaka-cu.ac.jp

Arsenic is a well-documented human carcinogen, and con-tamination with this heavy metal is of global concern,presenting a major issue in environmental health. However,the mechanism by which arsenic induces cancer isunknown, in large part due to the lack of an appropriateanimal model. In the present set of experiments, we focusedon dimethylarsinic acid (DMA), a major metabolite ofarsenic in most mammals including humans. We provide,for the first time, the full data, including detailed pathology,of the carcinogenicity of DMA in male F344 rats in a2-year bioassay, along with the first assessment of thegenetic alteration patterns in the induced rat urinarybladder tumors. Additionally, to test the hypothesis thatreactive oxygen species (ROS) may play a role in DMAcarcinogenesis, 8-hydroxy-2�-deoxyguanosine (8-OHdG)formation in urinary bladder was examined. In experiment1, a total of 144 male F344 rats at 10 weeks of age wererandomly divided into four groups that received DMA atconcentrations of 0, 12.5, 50 and 200 p.p.m. in the drinkingwater, respectively, for 104 weeks. From weeks 97–104,urinary bladder tumors were observed in 8 of 31 and 12of 31 rats in groups treated with 50 and 200 p.p.m. DMA,respectively, and the preneoplastic lesion, papillary ornodular hyperplasias (PN hyperplasia), was noted in 12and 14 rats, respectively. DMA treatment did not causetumors in other organs and no urinary bladder tumors orpreneoplastic lesions were evident in the 0 and 12.5 p.p.m.-treated groups. Urinary levels of arsenicals increasedsignificantly in a dose-responsive manner except for arseno-betaine (AsBe). DMA and trimethylarsine oxide (TMAO)were the major compounds detected in the urine, withsmall amounts of monomethylarsonic acid (MMA) andtetramethylarsonium (TeMa) also detected. Significantlyincreased 5-bromo-2�-deoxyuridine (BrdU) labeling indiceswere observed in the morphologically normal epithelium

Abbreviations: AP-1, activating protein-1 transcription factor; AsBe, arseno-betaine; DMA, dimethylarsinic acid; BBN, N-butyl-N-(4-hydroxybutyl)nitrosamine; BrdU, 5-bromo-2�-deoxyuridine; CDK, cyclin dependent kinase;COX-2, cyclooxygenase-2; IC-ICP-MS, ion chromatography with inductivelycoupled plasma mass spectrometry; MMA, monomethylarsonic acid; MSI,microsatellite instability; NF-κB, nuclear factor-κB; 8-OHdG, 8-hydroxy-2�-deoxyguanosine; PN hyperplasia, papillary or nodular hyperplasia; ROS,reactive oxygen species; TCC, transitional cell carcinoma; TeMA, tetramethyl-arsonium; TMAO, trimethylarsine oxide.

© Oxford University Press 1387

of the groups treated with 50 and 200 p.p.m. DMA.Mutation analysis showed that DMA-induced rat urinarybladder tumors had a low rate of H-ras mutations (2 of20, 10%). No alterations of the p53, K-ras or β-cateningenes were detected. Only one TCC (6%) demonstratednuclear accumulation of p53 protein by immunohistochem-istry. In 16 of 18 (89%) of the TTCs and 3 of 4 (75%)of the papillomas, decreased p27kip1 expression could bedemonstrated. Cyclin D1 overexpression was observed in26 of 47 (55%) PN hyperplasias, 3 of 4 (75%) papillomas,and 10 of 18 (56%) TCCs. As a molecular marker ofoxidative stress, increased COX-2 expression was noted in17 of 18 (94%) TCCs, 4 of 4 (100%) papillomas, and 39of 47 (83%) PN hyperplasias. In experiment 2, 8-OHdGformation in urinary bladder was significantly increasedafter treatment with 200 p.p.m. DMA in the drinking waterfor 2 weeks compared with the controls. The studiesdemonstrated DMA to be a carcinogen for the rat urinarybladder and suggested that DMA exposure may be relevantto the carcinogenic risk of inorganic arsenic in humans.Diverse genetic alterations observed in DMA-induced urin-ary bladder tumors imply that multiple genes are involvedin stages of DMA-induced tumor development. Further-more, generation of ROS is likely to play an importantrole in the early stages of DMA carcinogenesis.

Introduction

Arsenic is a known human carcinogen with the developmentof neoplasms in humans exposed to arsenic in the generalenvironment and in industry (1,2). Unfortunately, acute andchronic arsenic exposure still remains a major public healthproblem in many countries. In particular, drinking watercontamination results in increased incidences of cancers atmultiple organ sites, especially the skin and urinary bladder,but possibly also in the liver, kidney, lung, nasal cavity, andprostate, in exposed populations from Taiwan, China, EasternEurope, Southwestern United States, and South America (3,4).However, unlike most substances classified as carcinogens,carcinogenesis in laboratory animals by the metalloid arsenichas proved elusive (5). In fact, attempts to induce tumors inexperimental animals with inorganic arsenic compounds havemostly failed, except for a few studies in which animals weregiven arsenic trioxide by intratracheal instillation (6,7). Due tocertain deficiencies, the findings were not considered sufficientevidence of carcinogenicity in animals (1,2). The lack of aproperly designed bioassay was of integral importance in thisrespect (8).

We have focused on dimethylarsinic acid (DMA), a majormetabolite of arsenic in most mammals, including humans(9–12). Humans have continuous exposure to DMA fromarsenic in their drinking water, production or use of arsenic-containing herbicides and ingestion of foods contaminatedwith these herbicides. We previously have demonstrated that

M.Wie et al.

DMA promotes carcinogenesis in the urinary bladder, kidney,liver, and thyroid gland of F344 rats in an in vivo multi-organcarcinogenesis bioassay (13). Wanibuchi et al.(14,15) indicatedthat DMA exerts promoting potential in a dose-dependentmanner with regard to urinary bladder and liver carcinogenesisin F344 rats, possibly via a mechanism involving stimulationof cell proliferation and DNA damage caused by oxygenradicals. Of particular importance, the findings from our animalstudies revealed effects at sites where human arsenic-associatedcancers develop, such as the urinary bladder, liver and kidney,suggesting that DMA exposure may be relevant to the carcino-genic risk of arsenic to humans. This led us to conduct a2-year bioassay to determine the carcinogenicity of DMA inmale F344 rats. Previously, we reported that administration ofDMA induced urinary bladder tumors in male F344 rats atdoses of 50 and 200 p.p.m. in drinking water for 2 years(16). However, so far no detailed information for DMAcarcinogenicity in long-term studies is available, especially inorgans other than urinary bladder. A brief report of a 2-yearcarcinogenicity study in F344 rats fed DMA in the diet showedthat it produced bladder tumors at doses of 40 or 100 p.p.m.,and the effect was greater in female rats compared with males(17). In contrast, no bladder tumors or treatment-related tumorsin other tissues were observed in mice treated with DMA inthe diet for a 2-year period (17). Administration of DMA inthe drinking water to mice caused an increase in the totalnumbers of spontaneous tumors in wild type mice and signific-antly accelerated tumor induction in both p53 knockout andwild type mice when all neoplastic lesions were combined ina one and a half years study, while there was no evidence ofDMA-related tumors (18). Considering the multi-site promo-tional effects of DMA, one might question logically whetherDMA treatment induces tumors or not in other organs in rats (8).

A better understanding of the carcinogenic mechanisms ofarsenic will facilitate not only cancer risk assessment but alsocancer prevention strategies. The possibility exists that DMAplays a role in arsenic carcinogenicity in humans and therefore,elucidation of the kinds of genetic alterations involved inDMA-induced rat urinary bladder carcinogenesis is of interest.So far, a variety of mechanisms have been proposed by whicharsenic may exert carcinogenic activity. Based on evidencelargely from in vitro studies, Goering et al. (5) have suggestedthat induction of chromosome abnormalities, altered DNArepair, altered DNA methylation, oxidative stress, and increasedcell proliferation could be involved. Although DMA, arseniteand other arsenic compounds are not believed to directlybind to DNA (19), DNA single-strand breaks and crosslinkformation between DNA and protein have been observed(20,21). DMA is considered to be a clastogenic agent (22).Increasing in vivo evidence, furthermore, indicates that itstimulates cell proliferation (15,23,24) while generating ROS(14,25).

The first objective in the present study was to determine thecarcinogenicity of DMA in F344 rats when orally administratedin the drinking water for a 2-year period. The pathology oftumor development at a number of sites is provided in detailfor the first time.

The second objective was to elucidate possible mechanismsinvolved in DMA bladder carcinogenesis by analyzing DMA-induced urinary bladder tumors for mutations of p53, ras, andβ-catenin genes. As described above, the effect of DMA mightbe mediated as a result of interference with multiple pathwaysencompassing the cell cycle, DNA damage/repair, and genera-

1388

tion of reactive oxygen species (ROS). Protein expression ofcell cycle regulatory factors p53, p27kip1, and cyclin D1 innormal epithelium, preneoplastic and neoplastic lesions of raturinary bladder was assessed by immunohistochemistry. Totest the hypothesis that ROS may play a role in DMA bladdercarcinogenesis, we evaluated formation of 8-hydroxy-2�-deoxyguanosine (8-OHdG) in urinary bladder after short-termDMA treatment, this lesion being most commonly used as amarker for evaluation of oxidative DNA damage (26,27).Moreover, cyclooxygenase-2 (COX-2) expression was exam-ined in DMA-induced rat bladder tumors as it has been shownto play a role in development of preneoplastic and neoplasticlesions in the human and rat urinary bladder (28,29). ROS isknown to play a crucial role in the expression of COX-2through activating nuclear factor-κB (NF-κB), which acts asa positive regulatory element of expression (30,31). We alsoexamined the possible role of defective DNA mismatch repairactivity in DMA-induced urinary bladder tumors by analyzingmicrosatellite instability (MSI), reported in various humanmalignant tumors including human urinary bladder tumors(32,34). Finally, we discuss the possible implications of theresults in DMA urinary bladder carcinogenesis.

Materials and methods

Chemical

DMA was purchased from Wako Pure Chemical Industries, Japan (purity, 99%).

Experiment 1

The animal experimental protocol was previously described in detail (16). Atotal of 144 male, 6 week old, F344/DuCrj rats were obtained from CharlesRiver Japan, Hino, Japan, and divided randomly into four groups, eachconsisting of 36 rats. At the age of 10 weeks, rats received DMA atconcentrations of 0, 12.5, 50 and 200 p.p.m., respectively, in the drinkingwater for 104 weeks. Urine was collected by forced urination at the end ofweeks 30, 60 and 100; pH was immediately measured with a pH meter(Horiba model F-15, Tokyo, Japan), and then samples were stored at –80°Cuntil analyzed for urine chemistry and DMA metabolites. Rats that had diedor were killed under ether anesthesia when becoming moribund during thestudy or killed at the end of the study at week 104 were autopsied formacroscopic and histopathological examination. Blood samples were takenfrom all surviving rats from each group at week 104 for biochemical andhematological analysis. Ten rats from each group received an i.p. injection of100 mg/kg body weight of BrdU (Sigma Chemical Co., St Louis, MO) 1 hbefore autopsy. All major organs were excised and fixed in 10% bufferedformalin. After adequate fixation, they were cut and processed for paraffinembedding and routinely stained with hematoxylin and eosin for histologicalexamination.

Experiment 2

Forty 10-week-old male F344 rats were used for detection of 8-OHdGformation in urinary bladder DNA after DMA treatment. They were dividedinto two equal groups, and given DMA at concentrations of 0 and 200 p.p.m.,respectively, in the drinking water for 2 weeks. The rats were killed underether anesthesia and their urinary bladder removed, immediately frozen inliquid nitrogen and stored at –80°C until used for analysis. Tissues from pairsof urinary bladders were pooled as samples for detection of 8-OHdG formationin nuclear DNA.

Urinary analysis

Urinary samples from the rats at weeks 30, 60 and 100 were used forassessment of urine chemistry, including sodium, potassium, chloride, andcalcium (Hitachi-710 Electrolyte Analyzer, Tokyo Japan). Combined ionchromatography (IC, model 7000, Yokogawa Analytical Systems, Tokyo,Japan) with inductively coupled plasma mass spectrometry (ICP-MS, model4500; Hewlett-Packard, DE, USA) were used for the determination of arsenicspecies in the urine samples. Analysis conditions of the IC-ICP-MS systemwere described previously (35).

DNA extraction

In the present study, 20 tumors (18 carcinomas and two papillomas) fromdifferent rats were available for mutation analyses of p53, H- and K-ras, andthe β-catenin genes. Sixteen carcinomas were used for analysis of MSI. DNA

Dimethylarsinic acid carcinogenicity

was extracted from formalin-fixed, paraffin embedded urinary bladder tumorswith microdissection. DNA concentrations were determined with a spectropho-tometer (Ultraspec 3000, UV/Visible Spectrophotometer; Pharmacia Biotech,Tokyo, Japan) and adjusted to a final concentration of 50 ng/µl for PCR-SSCP and MSI analyses.

PCR-SSCP analysis of p53, H-ras, K-ras, and β-catenin

Exons 5–8 of p53, exons 1 and 2 of the H-ras and K-ras, and the β-cateningenes were analyzed by the PCR-SSCP method. For the β-catenin gene, theprimers were designed to cover the 5� terminal region of the gene correspondingto functionally important phosphorylation sites. Sequences of primers andconditions used for PCR were described previously (36,37). PCR-SSCP wasrepeated at least twice to confirm the results. If shifted bands were observedon PCR-SSCP analysis, they were cut out from the acrylamide gel to besequenced. All mutations were confirmed by repeating the PCR from freshmicrodissected tumor tissues to avoid PCR-related artifacts.

DNA sequencing

DNA sequences were determined with an ABI PRISM dye terminator cyclesequencing ready reaction kit with AmpliTaq DNA polymerase (Perkin Elmer,Applied Biosystems, Foster City, CA) on the ABI PRISM 310 genetic analyzer(Perkin Elmer, Applied Biosystems).

Microsatellite instability analysis

Analysis of MSI was performed by PCR using 18 microsatellite loci inter-spersed throughout the rat genome including D2MIT2 and D2MIT12 onChr.2; D3MGH9 on Chr.3; D6MGH3, D6MGH7 and IGHE on Chr.6;D8MGH3 on Chr.8; D9MIT1 and D9MGH1 on Chr.9; D13UWM1 onChr.13; D14MGH2 on Chr.14; D15MGH14 on Chr.15; D16MGH3 on Chr.16;D18MGH3 on Chr.18; TAT on Chr.19; D20MIT1 and D20UW1 on Chr.20;and DXMGH1 on Chr.X. Primer sets were chosen based on previouspublications (38,39) and obtained from Research Genetics (Huntsville, AL).Hot start PCR was performed in a 5 µl reaction volume containing 50 ngDNA, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.2 mM MgCl2, 50 µM eachof dATP, dGTP, dTTP, and dCTP, 0.2 µCi of [32P]dCTP, 4 pM of each primer,and 0.5 U of AmpliTaq Gold™ polymerase. Reactions were carried out in athermal cycler (MJ Research, Inc.) at 95°C for 10 min, followed by 38 cycles(95°C for 30 s, 52–56°C for 30 s, 72°C for 30 s), and final incubation forelongation at 72°C for 12 min. The PCR products were diluted with 20 µlstop solution, denatured, separated in two kinds of 6% polyacrylamide gelcontaining either 5 M urea or 30% formamide. Comparing tumors and normalliver tissues, a tumor was diagnosed as positive for MSI if one or moreabnormal band-shifts were observed.

Immunohistochemical analysis

Urinary bladders from a total of 40 rats from the groups in which bladdertumors were observed (50, 200 p.p.m. DMA-treated groups), as well as 10rats from the 12.5 p.p.m. DMA and control groups, respectively, wereexamined for p53, p27kip1, cyclinD1 and COX-2 by immunohistochemistry.BrdU incorporation in morphologically normal bladder epithelium was exam-ined in 10 rats of each group, respectively. Serial sections (4 µm) were cutfrom paraffin-embedded urinary bladder tissues and mounted on poly-L-lysine-coated slides. Established procedures for the immunohistochemical stainingwith the avidin–biotin–peroxidase complex (ABC) method were used withminor modification (36,40,41). Paraffin sections were deparaffinized andrehydrated through graded alcohols. Endogenous peroxidase activity wasblocked with 0.3% H2O2 in distilled water for 5 min, and then antigen retrievalwas performed by microwaving at 98°C for 20 min in 0.01 M citrate buffer(pH 6.0). For Cox-2 immunohistochemistry, sections were trypsinated at 37°Cin 0.4% trypsin, 0.01% CaCl2 in Tris-buffered saline for 30 min beforemicrowave. After blocking non-specific binding with serum at 37°C for30 min, sections were incubated with mouse monoclonal anti-BrdU antibody(M0744, Dako. A/S, Denmark) at 1:500 dilution, rabbit polyclonal p53antibody (FL-393, Santa Cruz Biotechnology, Santa Cruz, CA) at 1:1000dilution, rabbit polyclonal anti- p27kip1 antibody (C-19, Santa Cruz Biotechno-logy, Santa Cruz, CA) at 1:2000 dilution, mouse monoclonal cyclin D1antibody (A-12, Santa Cruz Biotechnology, Santa Cruz, CA) at 1:1000 dilution,or mouse monoclonal COX-2 antibody (Transduction Laboratories, Lexington,KY) at 1:500 dilution, overnight at 4°C. Immunoreactivity was detected usinga Vectastain Elite ABC Kit (PK-6102; Vector Laboratories, Burlingame, CA)and 3,3�-diaminobenzidine hydrochloride (Sigma Chemical Co., St Louis,MO) followed by counterstaining with Mayer’s hematoxylin. A negativecontrol was included with each staining procedure by omitting the primaryantibody. All specimens were evaluated independently by two of the authors(M.Wei and H.Wanibuchi).

Assessment of staining patterns

For BrdU labeling index, 3000–5000 urothelial cells in morphologicallynormal urothelia for each rat were counted. The BrdU labeling indices were

1389

scored as the number of positive cells per 100 urothelial cells. Overexpressionof p53 positive or cyclin D1 in lesions was defined as positive when nuclearstaining was evident in �5% of the cells (36,40). The level of p27kip1

immunoreactivity was high in nuclei of the transitional epithelial cells innormal specimens and preneoplatic lesions, providing a positive internalcontrol for each specimen. The specimens with p27kip1 expression �50% ofthe neoplastic cells were considered to have decreased p27kip1 expression(42,43). For COX-2 staining, normal urothelium consistently demonstratedno COX-2 immunoreactivity, providing a negative internal control for eachspecimen. Lesions with increased expression of COX-2 included tumors withweak immunoreactivity (cytoplasmic staining for �5% cells).

Detection of 8-OHdG formation in nuclear DNA of the urinary bladder

DNA of the urinary bladders from experiment 2 was extracted by chaotropicNaCl isolation method using a DNA Extractor WB kit (Wako Pure ChemicalIndustries, Kyoto, Japan). DNA hydrolysis and microfiltration of the resultantsamples were subsequently conducted by the method of Helbock et al (44,45).The levels of 8-OHdG were then determined by the HPLC-ECD method asdescribed elsewhere (27). Peaks detected with electrochemical (for 8-OHdG)and UV (for dG) detectors were integrated with a background noise correctionloaded on an integrator. Values for 8-OHdG per 105 cells were obtained bycalibration against curves from runs of standard samples containing knownamounts of authentic 8-OHdG (Wako Pure Chemical Industries, Kyoto, Japan)and dG (Sigma Chemical Co., St Louis, MO). During the assays, light andair contamination were avoided as strictly as possible.

Statistical analysis

Statistical comparisons between the different groups were completed withStatView J-5.0 software (Abacus Concepts, Berkeley, CA) for the Macintoshcomputer. Inter-group relationships for lesion incidences and immunohisto-chemical variables were determined by χ2 probability analysis or Fisher’sexact probability test. For assessment of the mean values of the BrdU labelingindex and 8-OHdG formation, Dunnett analysis and Welch’s t-test were used,respectively. A result was only considered as statistically significant if P � 0.05.

Results

General conditionsThere was no evidence of increased morbidity or mortalityafter DMA treatment. There were no statistically significantdifferences for the overall survival rates between groups duringthe experiment. The most common cause of death in all groupswas leukemia, with grossly enlarged spleen and usually diffuseinfiltrates of leukemic cells in the liver and spleen. DMA didnot have adverse effects on food consumption, but caused anincrease in water consumption at 50 and 200 p.p.m.. The finalbody weights were not significantly different between thegroups. No treatment-related adverse effects were apparent inthe hematological and serum biochemical data among groups(data not shown).

Tumorigenicity of DMAData for the numbers and incidences of rats with urinarybladder lesions are summarized in Table I. The reported ratesare for animals surviving until at least week 97, when the firsturinary bladder tumor was found. A total of 22 urinary bladdertumors including 18 TCCs and four papillomas were observedin 8 of 31 and 12 of 31 animals in 50 and 200 p.p.m. DMAgroups, respectively. Two rats in 200 p.p.m. DMA group withbladder papillomas also had carcinomas. PN hyperplasias,preneoplastic lesions in the urinary bladder, were observed in12 and 14 rats in 50 and 200 p.p.m. DMA groups, respectively.No such lesions were observed in groups treated with 0 or12.5 p.p.m. DMA.

As Table II shows, in all groups, the incidences of tumorsexcept for those in the urinary bladder were typical for maleF344 rats and did not exceed the historical control incidences/range of the NTP (46). There were no significant differencesin the incidences of these tumors at any dose level, althoughDMA-treated rats tended to have more fibromas than controls.

M.Wie et al.

Table I. Incidences of urinary bladder lesions in rats treated with DMAa

DMA (p.p.m.) Effective no. of ratb PN hyperplasia Papilloma (%) TCC (%) No. of tumor-bearing rats (%)

Total no. counted Incidence (%)

0 28 0 0 0 0 012.5 33 0 0 0 0 050 31 22 12 (39)c 2 (6) 6 (19)d 8 (26)c

200 31 25 14 (45)e 2 (6) 12 (39)e 12f (39)e

aPart of data were reported previously (ref. 14).bThe effective number of rats was indicated as the number alive at week 97, when the first bladder tumor was found.cP � 0.01, dP � 0.05, eP � 0.001 (vs. 0 p.p.m. group)fTwo rats with bladder papillomas also had carcinomas.

Table II. Incidences of tumors other than urinary bladder tumors in F344male rats treated with DMA

Site and type of tumor DMA (p.p.m.)

0 12.5 50 200(n � 28a) (n � 33) (n � 31) (n � 31)

Whole bodyMalignant lymphoma- 4b (14)c 7 (21) 5 (16) 4 (13)leukemia

Thyroid glandFollicular adenoma 1 (4) 0 0 0C-cell adenoma 3 (11) 3 (9) 4 (13) 3 (10)C-cell carcinoma 1 (4) 1 (3) 0 1 (3)

LungAdenoma 1 (4) 0 2 (6) 1 (3)

LiverNeoplastic nodular 1 (4) 2 (6) 1 (3) 0Hepatocellular carcinoma 1 (4) 0 0 1 (3)Cholangioma 1 (4) 0 0 0Lipoma 0 0 0 1 (3)

EsophagusSquamous cell carcinoma 0 0 0 0Large intestine/rectumAdenocarcinoma 1 (4) 1 (3) 1 (3) 0

Pancreas 1 (4) 2 (6) 1 (3) 0Acinar dell adenoma 0 0 1 (3) 0Islet cell adenoma 1 (4) 2 (6) 0 0

KidneyAdenoma 1 (4) 1 (3) 0 0

SkinSquamous cell papilloma 0 0 1 (3) 0Squamous cell carcinoma 1 (4) 0 0 1 (3)

SubcuitsFibroma 1 (4) 4 (12) 5 (16) 5 (16)Fibrosarcoma 1 (4) 2 (6) 0 1 (3)Malignant fibrous histocytoma 2 (7) 1 (3) 0 1 (3)Lipoma 0 0 1 (3) 0Liposarcoma 0 0 2 (6) 0

TestisInterstitial cell tumor 20 (71) 26 (79) 24 (77) 25 (81)

aThe effective number of rats was indicated as the number alive at week 97,when the first bladder tumor was found.bNumber of rats with tumors.cNumbers in parentheses are percentage of rats with tumors.

Urinary analysisThe urinary pH did not differ significantly among groups. Theconcentrations of sodium, potassium, chloride and calciumwere decreased in the rats treated with DMA in a dose-dependent manner, with statistical significance being reachedin the 50 and 200 p.p.m. DMA groups (data not shown). Thedecrease in urine electrolytes was presumably due to increased

1390

urinary volume. These results confirmed previous observationsshowing the carcinogenic action of DMA in rat urinary bladderis not correlated with urine pH or sodium concentration (15).

The urinary concentrations of arsenic metabolites at week100 are indicated in Table III. Arsenic compound levelsincreased in a dose-dependent manner except for AsBe. Majorcompounds were DMA itself and TMAO, with small amountsof MMA and TeMa also detected. As described in previousreports, two unidentified metabolites, peak 1 and peak 2(15,35), were also found in DMA-treated groups but notthe controls.

Mutations of the H-ras, p53, K-ras and β-catenin genes inDMA-induced urinary bladder tumorTwenty paraffin-embedded DMA-induced rat urinary tumors(18 TCCs and two papillomas) were examined for mutations inexons 1 and 2 of the H- and K-ras oncogenes, in exons 5–8 ofthe p53 tumor suppressor gene and the β-catenin gene usingPCR-SSCP and direct sequencing techniques. Mutations in exon1 of H-ras were demonstrated in two TCCs (2 of 20, 10%).Direct sequencing of the two samples indicated one G→T trans-version at codon 13 (GGC→GTC) and one G→T transversionand a C→A transversion at the second and third bases of codon13 (GGC→GTA), respectively. Both mutations resulted in anidentical amino acid substitution of glycine for valine. Nomutations were found in p53, K-ras, and β-catenin genes.

MSI analysisEighteen microsatellite loci interspersed throughout the ratgenome were analyzed in 16 DMA-induced urinary bladdercarcinomas, but no alterations were detected.

BrdU labeling indexUrinary bladder epithelial cell proliferation following treatmentwith DMA were determined by BrdU incorporation. The dataare presented as BrdU labeling indices for morphologicallynormal bladder epithelium. As shown in Figure 1, a significantincrease was noted for the groups treated with 50 and 200 p.p.m.DMA when compared with controls.

ImmunohistochemistryResults of immunohistochemical assessment of p53, p27kip1,cyclin D1 and COX-2 expressions in the rat urinary bladderlesions are shown in Figure 2.

Positive p27kip1 staining was noted within the nuclei of theurothelial cells in the control group and in morphologicallynormal appearing urothelium and simple hyperplasia in DMA-treated groups. Typical positive staining patterns for p27kip1

are shown in Figure 3. The majority of PN hyperplasias alsostained intensely; in marked contrast, almost all TCC and

Dimethylarsinic acid carcinogenicity

papillomas demonstrated a heterogeneous pattern of signific-antly reduced p27kip1 immunoreactivity. Thus 16 of 18 (89%)TTCs, and 3 of 4 (75%) papillomas demonstrated decreasedp27kip1 expression. It is worth noting that p27kip1 is significantlydownregulated in papillomas and TCCs when compared withPN hyperplasias. In fact, no staining was observed in onepapilloma and 10 TCCs.

Fig. 1. BrdU labeling indices for morphologically normal bladderepithelium of rats treated with DMA for 104 weeks. Significantly differentfrom controls (DMA, 0 p.p.m.) at *P � 0.05; **P � 0.01. Bars indicate theSD.

Fig. 2. Immunohistochemical assessment of p53, p27kip1, cyclin D1 and COX-2 expression in rat urinary bladder lesions induced by DMA. Significantlydifferent from PN hyperplasia at *P � 0.05; **P � 0.001. The assessment criteria are described in Materials and methods.

Table III. Urinary concentrations of arsenic compounds at week 100 in male F344 rats treated with DMA

DMA (p.p.m.) No. of samples Concentration (µg/ml)

MMA DMA TMAO TeMA AsBe Peak 1 Peak 2

0 10 �0.01 0.08 � 0.03 0.04 � 0.01 �0.01 0.37 � 0.11 0 012.5 10 0.01 � 0.01 3.1 � 1.7 2.7 � 1.6 0.01 � 0.01 0.32 � 0.16 0.08 � 0.07 0.23 � 0.1350 10 0.02 � 0.02 20.3 � 8.5 9.5 � 2.5 0.04 � 0.02 0.34 � 0.11 2.01 � 0.68 2.14 � 0.76

200 10 0.07 � 0.01 44.1 � 5.4 36.6 � 9.1 0.27 � 0.04 0.45 � 0.06 5.45 � 1.81 6.99 � 0.13

1391

All 10 control specimens of the control group did notshow positive cyclin D1 nuclear staining in the urothelium(Figure 4A). In contrast, all 40 specimens from the DMA-treated group displayed occasional nuclear cyclin D1 stainingin small stretches of normal appearing epithelium, includingall 10 bladders examined from rats given 12.5 p.p.m. DMAin which no bladder lesion was observed (Figure 4B). Thecyclin D1 overexpression phenotype, defined as positive immu-noreactivity in the nuclei of �5% of neoplastic cells(Figure 4D–F), was found in 26 of 47 (55%) PN hyperplasias,3 of 4 papillomas (75%), and 10 of 18 TCCs (56%). Therewas no significant difference in incidences of cyclin D1overexpression between the bladder lesion types.

COX-2 staining was localized to the cytoplasm of tumor orpreneoplastic cells but was not present in normal urothelialcells (Figure 5). Increased expression was noted in 17 of 18(94%) TCCs, 4 of 4 (100%) papillomas, and 39 of 47 (83%)PN hyperplasias. Normal epithelial cells in the control groupand morphologically normal urothelium of the DMA-treatedgroups commonly did not show immunoreactivity for COX-2, providing a negative internal control for each specimen.However, positive COX-2 staining also occasionally wasnoted in the simple hyperplasia and morphologically normalepithelium adjacent to tumors.

Only one TCC (6%) demonstrated nuclear accumulation ofp53 protein in �5% of the nuclei (Figure 6A). We did notfind any statistically significant association between p27kip1,cyclin D1 and COX-2 expression in the DMA-induced urinarybladder lesions although most bladder lesions showeddecreased p27kip1 expression and increased cyclin D1 andCOX-2 expression simultaneously. However, as shown in

M.Wie et al.

Fig. 3. Immunohistochemical assessment of p27kip1 expression. A high level of p27kip1 immunoreactivity is apparent in urothelial cell nuclei of the controlgroup and in morphologically normal appearing epithelium and simple hyperplasia in the DMA-treated groups. (A) Normal bladder mucosa from a controlrat; (B) morphologically normal mucosa following 200 p.p.m. DMA; (C) simple and PN hyperplasia; (D) papilloma with nuclear p27kip1 expression �50%;(E) TCC with nuclear p27kip1 expression �50%.

Figure 6, no mutually exclusive distribution pattern wasobserved.

Increased formation of 8-OHdG in urinary bladder DNA8-OHdG formation was significantly increased in DMA-treatedrats (1.76 � 0.59/105dG) after treatment with 200 p.p.m. DMAin the drinking water for 2 weeks compared with the controls(1.21 � 0.13/105dG).

Discussion

The present study demonstrated that DMA is carcinogenic forthe urinary bladder of F344 male rats, but lacks other organ-specific carcinogenic effects, particularly regarding the liverand kidney in which promoting effects on rat carcinogenesishave been shown and arsenic-associated tumors have beenreported in humans. We also found that DMA-induced raturinary bladder tumors had a low rate of H-ras mutations andno mutations in p53, K-ras or β-catenin genes. Furthermore,we demonstrated that induction of cell proliferation mightcontribute to the carcinogenicity of DMA via mechanismsinvolving oxidative stress and/or alterations in cell cycleregulatory protein, p27kip1 and cyclin D1. DMA is a majormetabolite of arsenic in most mammals (9–12). Thus, theresults from the present study would appear directly relevantto the carcinogenic risk of arsenic.

1392

The bladder-specific carcinogenic effect of DMA in rat mayindicate: (i) longer exposure to DMA in the urinary bladderthan in other organs due to urinary retention; (ii) promotingactivities of DMA on multiple organs need to be consideredfor assessing carcinogenic effects of arsenic in humans; (iii)humans may be more sensitive than experimental animals tocancer induction by arsenic (5,8); or (iv) DMA may not bethe only carcinogenic compound involved in arsenic carcino-genesis. An ongoing 2-year carcinogenicity study of MMAand TMAO in our laboratory should clarify whether this isindeed the case.

p53 is the most frequently mutated tumor suppressor genedescribed so far in human and experimental animal cancersincluding urinary bladder cancer (47,48). Examination of themolecular changes in the p53 tumor suppressor gene cancontribute to our understanding of the nature of carcinogenicactivity. In contrast to our previous finding of frequent p53mutations in BBN-induced rat urinary bladder cancer (49), nosuch lesions were found in DMA-induced rat urinary bladdertumors examined in the present study. Our results indicatedthat DMA differs from BBN, which is genotoxic, and pathwaysother than the p53 pathway must be involved in the etiologyof the DMA-induced rat urinary bladder tumors. Two previousstudies reported high frequency of p53 mutations in arsenic-related bladder and skin tumors from the endemic area of

Dimethylarsinic acid carcinogenicity

Fig. 4. Immunohistochemical detection of cyclin D1 expression in morphologically normal bladder mucosa and neoplastic lesions following DMA treatment.(A) Normal bladder mucosa from a control rat; (B) morphologically normal bladder mucosa following 12.5 p.p.m. DMA; (C) simple hyperplasia from a200 p.p.m. DMA rat; (D) PN hyperplasia with increased cyclin D1 expression; (E) TCC with increased cyclin D1 expression; (F) higher magnification of theTCC in (E).

black foot disease in Taiwan (50,51). However, the discrepancymight be partly explicable by the fact that development ofmalignancy in humans is a complex multistep process, andmany factors may affect the likelihood that cancer will develop.Therefore, it is reasonable to hypothesize that genetic altera-tions found in human cancers are results of factors such assmoking and exposure to other arsenicals. DMA is consideredto be a clastogenic agent (22) and negative in most mutagenicitystudies (19). In light of these actions, mutations of H-ras intwo TCCs could occur indirectly by oxidative damage orcytotoxicity of DMA (23,2,52).

It is well established that disruption of the normal cell cycleis a critical step in cancer development (53). Because alterationsin the p53 gene were lacking, we therefore focused ourattention on the expression of the tumor suppressor genep27kip1. p27kip1 functions as a p53 independent negative cellcycle regulator involved in G1 arrest, and the reduction in theprotein level of p27kip1 have been reported in a variety ofhuman cancers and is likely to provide a selective growthadvantage (54). The strong p27kip1 staining in the majority ofpreneoplastic lesions noted here is consistent with the estab-lished role of p27kip1 as a tumor suppressor gene counteractingproliferative signals generated by DMA exposure. The reducedexpression of p27kip1 protein in almost all TCCs and papillomassuggests a role in malignant progression in DMA-induced ratbladder tumors.

1393

Overexpression of cyclin D1 in human urinary bladdertumors could be a key regulatory event leading to cellproliferation and tumorigenesis (36). In the present study, wealso found that positive cyclin D1 nuclear staining appearedin small stretches of histologically normal appearing epitheliumfollowing DMA treatment, even in animals in which no bladderlesions were observed histologically, as well as in most locallesions. The present data strongly suggested that cyclin D1induction is one of the early events in DMA-induced ratbladder carcinogenesis. However, the fact that failure to inducebladder tumors at a dose of 12.5 p.p.m. DMA suggests thatincreased cyclin D1 associated with such a low dose may beinsufficient for DMA bladder carcinogenesis under the presentconditions. In addition, the observed existence of tumorswithout increased cyclin D1 expression but featuring down-regulation of p27kip1 expression may mean that it is no longernecessary for at least a subset of tumors in the later stages.The actions of cyclin D1 are regulated by CDK inhibitorssuch as p27kip1, which control its ability to activate CDK4 andCDK6. Therefore, those tumors are likely a result, at least inpart, of either increased degradation or transcription of p27kip1.The lack of any exclusive distribution pattern indicates thatalterations may occur independently and there might existother mechanisms by which DMA affects cell cycle regulation.Defects in a cell cycle check point may be responsible forthe genomic instability (53). We can conclude that such

M.Wie et al.

Fig. 5. Increased COX-2 expression in DMA-induced bladder lesions. Note increase in cytoplasm. (A) Normal bladder mucosa from a control rat; (B) nodularhyperplasia; (C) advanced nodular hyperplasia; (D) TCC; (E) higher magnification of the TCC in (C).

abnormalities are frequent in DMA bladder carcinogenesis andmight induce genomic instability despite the rarity of mutationsin the present study.

Increasing evidence supports the hypothesis that ROS mayplay a role in DMA carcinogenesis. It is reported that metabol-ism of substances by the P-450 enzyme system can generateoxygen free radicals (55,56). Our recent finding that an increasein hepatic P450 levels, especially in CYP2B1 protein in ratlivers after treatment with 100 p.p.m. DMA suggests thatDMA are metabolized by P450 in rat liver and could representa mechanism by which DMA generate ROS (57). Alternatively,the possibility also exists that cytotoxicity of DMA mayinvolve the generation of ROS since xenobiotic chemicals canproduce ROS by either direct or indirect means (58,59).Yamanaka et al. (60,61) demonstrated production of oxygenradicals in the metabolism of DMA, such as the superoxideanion radical and the dimethylarsenic peroxyl radical, andmight have a role in DNA damage in the lungs of mice andrats. Among ROS-induced forms of DNA damage, 8-OHdGis typical and most commonly used as a marker for quantitativeanalysis (26,62). The finding of a significant increase in DMA-treated rats in the present study suggests that DMA treatmentcauses DNA damage via ROS generation, as shown earlier forthe rat liver (14).

COX-2 expression, shown to be involved with developmentof preneoplastic and neoplastic lesions in the human and rat

1394

bladder (28,29), was also diffuse in the majority of TCCs,papillomas and PN hyperplasias in the present study. Theoccasional positive COX-2 staining noted in the morpholo-gically normal epithelium adjacent to tumor, observed also inhuman invasive TCC of the bladder, may indicate neoplasticcells can exert paracrine effects through the release of cytokinesand/or growth factors. ROS are known to play a crucial rolein the expression of COX-2, so that this may also be amolecular marker of oxidative stress (63).

Microsatellite instability was absent in the available DMA-induced rat bladder cancers in this study, although this couldbe due to the low number of markers, only one to threemarkers for each chromosome. Thus, further study with alarger number of microsatellites is necessary for clarification.

Based on the observations in the present experiment andthe results from the literature, potential modes of action forDMA with regard to rat urinary bladder carcinogenesis aregiven in Figure 7. We propose two possible mechanisms bywhich ROS could be involved in DMA carcinogenesis in rats:(i) DMA-initiated ROS may cause specific molecular changesresulting in the activation of transcription factors such asAP-1 and NF-κB. Although AP-1 and NF-κB were notinvestigated here, ROS has been shown to cause synthesis ofAP-1, and activation of AP-1 or NF-κB promotes carcino-genesis (64,65). DMA induces cell proliferation and geneexpression in the bladder epithelium associated with AP-1 in

Dimethylarsinic acid carcinogenicity

Fig. 6. Immunohistochemistry for p53, p27kip1, cyclin D1 and COX-2 in serial sections. Note that multiple alternative expressions existed concurrently intumors but without a mutually exclusive distribution pattern. (A) The only TCC with p53 overexpression; (B) serial section showed negligible p27kip1 nuclearstaining; (C) serial section showing diffusely increased cyclin D1 expression; (D) serial section showed increased COX-2 expression.

Fig. 7. Potential modes of action underlying DMA carcinogenesis in raturinary bladder.

mice (24). It should be remembered that ROS are known toplay a crucial role in the expression of COX-2 throughactivating nuclear factor-κB. (30,31) (ii) DMA could causechromosomal abnormalities by generation of ROS and resultant

1395

DNA single strand breaks and DNA-protein crosslinks(21,60,66). It is generally accepted that tumor developmentoccurs as the result of accumulation of genetic alterations(67,68). The various genetic alterations induced by DMA maynot be the result of independent mechanisms. Some modesmay be operating concurrently or sequentially. DMA-initiateddefects in cell cycle checkpoints could give rise to genomicinstability, while DMA-induced DNA damage would beexpected to affect the expression of cell cycle regulators.Moreover, the fact that oxygen radicals may participate inthe carcinogenic process, including the stages of initiation,promotion, and progression (69,71), suggest a reasonablemechanism by which DMA may act in all three phases. Inaddition, an alternative or complementary mechanism sug-gested by Cohen et al. for rat bladder carcinogenesis, iscytoxicity and regeneration (23,52). The ultimate relationshipsbetween oxidative stress, toxic stress and genetic alterationsin arsenic carcinogenesis remain to be determined.

In conclusion, the present work provides unequivocalevidence that DMA may be a carcinogen for the rat urinarybladder, supporting the epidemiological data that inorganicarsenic is a human bladder carcinogen and suggesting thatDMA may be relevant to the carcinogenic risk of inorganicarsenic exposure in humans. DMA-induced urinary bladdertumors occur as the result of an accumulation of diverse

M.Wie et al.

genetic alterations. The present elucidation of the cellularand molecular pathways involved in DMA carcinogenesisin rats suggest that particular attention should be paid tooxygen stress in the human populations at risk from arseniccarcinogenesis.

Acknowledgements

This work was supported in part by Core Research for Evolutional Scienceand Technology (CREST) grant from Japan Science and Technology Corpora-tion (JST), and Grant in-Aid for Cancer Research of Arsenics from theEnvironment Agency, Japan, and Fund for Medical Research from Osaka CityUniversity Medical Research Foundation. We are grateful to Kawakami Emiand Touma Kaori (Department of Pathology, Osaka City University MedicalSchool, Japan) for assistance in the immunohistochemistry studies.

References

1. IARC (1980) Monographs on the Evaluation of the Carcinogenic Risk ofChemicals to Humans, Vol. 23: Some Metals and Metallic Compounds.IARC, Lyon, France, pp. 39–141.

2. IARC (1987) Monographs on the Evaluation of the Carcinogenic Risk ofChemicals to Humans, Suppl. 7, Overall Evaluation of the Carcinogenicity:An Updating of IARC. Monographs, Vols. 1–40. IARC, Lyon, France, pp.100–106.

3.Chen,C.J., Chen,C.W., Wu,M.M. and Kuo,T.L. (1992) Cancer potential inliver, lung, bladder and kidney due to ingested inorganic arsenic in drinkingwater. Br. J. Cancer, 66, 888–892.

4.Smith,A.H., Hopenhayn,Rich,C., Bates,M.N., Goeden,H.M., Hertz,P.I.,Duggan,H.M., Wood,R., Kosnett,M.J. and Smith,M.T. (1992) Cancer risksfrom arsenic in drinking water. Environ. Health Perspect., 97, 259–267.

5.Goering,P.L., Aposhian,H.V., Mass,M.J., Cebrian,M., Beck,B.D. andWaalkes, M.P. (1999) The enigma of arsenic carcinogenesis: role ofmetabolism. Toxicol. Sci., 49, 5–14.

6. Ishinishi,N., Mizunoe,M., Inamasu,T. and Hisanaga,A. (1980)[Experimental study on carcinogenicity of beryllium oxide and arsenictrioxide to the lung of rats by an intratracheal instillation (author’s transl)].Fukuoka Igaku Zasshi, 71, 19–26.

7.Pershagen,G., Nordberg,G. and Bjorklund,N.E. (1984) Carcinomas of therespiratory tract in hamsters given arsenic trioxide and/or benzo[a]pyreneby the pulmonary route. Environ. Res., 34, 227–241.

8.Huff,J., Chan,P. and Nyska,A. (2000) Is the human carcinogen arseniccarcinogenic to laboratory animals? Toxicol. Sci., 55, 17–23.

9.Bertolero,F., Marafante,E., Rade,J.E., Pietra,R. and Sabbioni,E. (1981)Biotransformation and intracellular binding of arsenic in tissues of rabbitsafter intraperitoneal administration of 74As labelled arsenite. Toxicology,20, 35–44.

10.Buchet,J.P., Lauwerys,R. and Roels,H. (1980) Comparison of severalmethods for the determination of arsenic compounds in water and in urine.Their application for the study of arsenic metabolism and for the monitoringof workers exposed to arsenic. Int. Arch. Occup. Environ. Health., 46,11–29.

11.Tam,G.K., Charbonneau,S.M., Bryce,F., Pomroy,C. and Sandi,E. (1979)Metabolism of inorganic arsenic (74As) in humans following oral ingestion.Toxicol. Appl. Pharmacol., 50, 319–322.

12.Vahter,M. (1981) Biotransformation of trivalent and pentavalent inorganicarsenic in mice and rats. Environ. Res., 25, 286–293.

13.Yamamoto,S., Konishi,Y., Matsuda,T., et al. (1995) Cancer induction byan organic arsenic compound, dimethylarsinic acid (cacodylic acid), inF344/DuCrj rats after pretreatment with five carcinogens. Cancer Res.,55, 1271–1276.

14.Wanibuchi,H., Hori,T., Meenakshi,V., et al. (1997) Promotion of rathepatocarcinogenesis by dimethylarsinic acid: association with elevatedornithine decarboxylase activity and formation of 8-hydroxy-deoxyguanosine in the liver. Jpn. J. Cancer Res., 88, 1149–1154.

15.Wanibuchi,H., Yamamoto,S., Chen,H., Yoshida,K., Endo,G., Hori,T. andFukushima,S. (1996) Promoting effects of dimethylarsinic acid on N-butyl-N-(4-hydroxybutyl)nitrosamine-induced urinary bladder carcinogenesis inrats. Carcinogenesis, 17, 2435–2439.

16.Wei,M., Wanibuchi,H., Yamamoto,S., Li,W. and Fukushima,S. (1999)Urinary bladder carcinogenicity of dimethylarsinic acid in male F344 rats.Carcinogenesis, 20, 1873–1876.

17.van Gemert,M. and Eldan,M. (1998) Chronic carcinogenicity assessmentof cacodylic acid. 3 rd International Conference on Arsenic Exposure andHealth Effects, Book of Abstracts, 113 pp.

1396

18.Fukushima,S., Wanibuchi,H., Wei,M., Salim,E.I., (2000) Carcinogenicityof dimethylarsinic acid in rats and mice. 3 rd International Conference onArsenic Exposure and Health Effects, Book of Abstracts, 110pp.

19.U.S. Environmental Protection Agency. (1997) Report on the ExpertPanel on Arsenic Carcinogenicity. National Center for EnvironmentalAssessment, U.S. Environmental Protection Agency, Washington D.C.

20.Dong,J.T. and Luo,X.M. (1993) Arsenic-induced DNA-strand breaksassociated with DNA-protein crosslinks in human fetal lung fibroblasts.Mutat. Res., 302, 97–102.

21.Yamanaka,K., Hasegawa,A., Sawamura,R. and Okada,S. (1989) DNAstrand breaks in mammalian tissues induced by methylarsenics. Biol.Trace. Elem. Res., 21, 413–417.

22.ATSDR (1999) Toxicological profile for arsenic (update). Agency forToxic Substances and Disease Registry, Atlanta, GA.

23.Arnold,L.L., Cano,M., St John,M., Eldan,M., van Gemert,M. andCohen,S.M. (1999) Effects of dietary dimethylarsinic acid on the urineand urothelium of rats. Carcinogenesis, 20, 2171–2179.

24.Simeonova,P.P., Wang,S.Y., Toriuma,W., et al. (2000) Arsenic mediatescell proliferation and gene expression in the bladder epithelium: associationwith activating protein-1 transactivation. Cancer Res, 60, 3445–3453.

25.Hei,T.K., Liu,S.X. and Waldren,C. (1998) Mutagenicity of arsenic inmammalian cells: Role of reactive oxygen species. Proc Natl Acad Sci.USA, 95, 8103–8107.

26.Floyd,R.A. (1990) The role of 8-hydroxyguanine in carcinogenesis.Carcinogenesis, 11, 1447–1450.

27.Nakae,D., Kobayashi,Y., Akai,H., Andoh,N., Satoh,H., Ohashi,K.,Tsutsumi,M. and Konishi,Y. (1997) Involvement of 8-hydroxyguanineformation in the initiation of rat liver carcinogenesis by low dose levelsof N-nitrosodiethylamine. Cancer Res., 57, 1281–1287.

28.Mohammed,S.I., Knapp,D.W., Bostwick,D.G., Foster,R.S., Khan,K.N.,Masferrer,J.L., Woerner,B.M., Snyder,P.W. and Koki,A.T. (1999)Expression of cyclooxygenase-2 (COX-2) in human invasive transitionalcell carcinoma (TCC) of the urinary bladder. Cancer Res., 59, 5647–5650.

29.Kitayama,W., Denda,A., Yoshida,J., Sasaki,Y., Takahama,M.,Murakawa,K., Tsujiuchi,T., Tsutsumi,M. and Konishi,Y. (2000) Increasedexpression of cyclooxygenase-2 protein in rat lung tumors induced by N-nitrosobis(2-hydroxypropyl)amine. Cancer Lett., 148, 145–152.

30.Kosaka,T., Miyata,A., Ihara,H., Hara,S., Sugimoto,T., Takeda,O.,Takahashi,E. and Tanabe,T. (1994) Characterization of the human gene(PTGS2) encoding prostaglandin-endoperoxide synthase 2. Eur. J.Biochem., 221, 889–897.

31.Sen,C.K. and Packer,L. (1996) Antioxidant and redox regulation of genetranscription [see comments]. FASEB J., 10, 709–720.

32.Gonzalez Zulueta,M., Ruppert,J. M., Tokino,K., et al. (1993) Microsatelliteinstability in bladder cancer. Cancer Res., 53, 5620–5623.

33.Honchel,R., Halling,K.C. and Thibodeau,S.N. (1995) Genomic instabilityin neoplasia. Semin. Cell Biol., 6, 45–52.

34.Mao,L., Schoenberg,M.P., Scicchitano,M., Erozan,Y.S., Merlo,A.,Schwab,D. and Sidransky,D. (1996) Molecular detection of primary bladdercancer by microsatellite analysis. Science, 271, 659–62.

35.Yoshida,K., Inoue,Y., Kuroda,K., Chen,H., Wanibuchi,H., Fukushima,S.and Endo,G. (1998) Urinary excretion of arsenic metabolites after long-term oral administration of various arsenic compounds to rats. J. Toxicol.Environ. Health Part A, 54, 179–192.

36.Lee,C.C., Yamamoto,S., Morimura,K., et al. (1997) Significance of cyclinD1 overexpression in transitional cell carcinomas of the urinary bladderand its correlation with histopathologic features. Cancer, 79, 780–789.

37.Dashwood,R.H., Suzui,M., Nakagama,H., Sugimura,T. and Nagao,M.(1998) High frequency of beta-catenin (ctnnb1) mutations in the colontumors induced by two heterocyclic amines in the F344 rat. Cancer Res.,58, 1127–1129.

38. Jacob,H.J., Brown,D.M., Bunker,R.K., et al. (1995) A genetic linkage mapof the laboratory rat, Rattus norvegicus. Nat. Genet., 9, 63–69.

39.Toyota,M., Ushijima,T., Weisburger,J.H., Hosoya,Y., Canzian,F.,Rivenson,A., Imai,K., Sugimura,T. and Nagao,M. (1996) Microsatelliteinstability and loss of heterozygosity on chromosome 10 in rat mammarytumors induced by 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine.Mol. Carcinog., 15, 176–182.

40.Lee,C.C., Yamamoto,S., Wanibuchi,H., Wada,S., Sugimur,K., Kishimoto,T.and Fukushima,S. (1997) Cyclin D1 overexpression in rat two-stage bladdercarcinogenesis and its relationship with oncogenes, tumor suppressor genes,and cell proliferation. Cancer Res., 57, 4765–4776.

41.Lee,T.C., Tanaka,N., Lamb,P.W., Gilmer,T.M. and Barrett,J.C. (1988)Induction of gene amplification by arsenic. Science, 241, 79–81.

42.Porter,P.L., Malone,K.E., Heagerty,P.J., Alexander,G.M., Gatti,L.A.,Firpo,E.J., Daling,J.R. and Roberts,J.M. (1997) Expression of cell-cycleregulators p27Kip1 and cyclin E, alone and in combination, correlate with

Dimethylarsinic acid carcinogenicity

survival in young breast cancer patients [see comments]. Nat. Med., 3,222–225.

43.Esposito,V., Baldi,A., De Luca,A., et al. (1997) Prognostic role of thecyclin-dependent kinase inhibitor p27 in non-small cell lung cancer.Cancer Res., 57, 3381–3385.

44.Helbock,H.J., Beckman,K.B., Shigenaga,M.K., Walter,P.B., Woodall,A.A.,Yeo,H.C. and Ames,B.N. (1998) DNA oxidation matters: The HPLC-electrochemical detection assay of 8-oxo-deoxyguanosine and 8-oxo-guanine. Proc. Natl. Acad. Sci. USA, 95, 288–293.

45.Helbock,H.J., Beckman,K.B. and Ames,B.N. (1999) 8-hydroxy-deoxyguanosine and 8-hydroxyguanine as biomarkers of oxidative DNAdamage. Oxid. Antioxid. Part B, 300, 156–166.

46.Haseman,J.K., Arnold,J., and Eustis,S. L. (1990) Tumor Incidences inFischer 344 Rats: NTP Historical Data. In: Boorman,G.A., Eustis,S.L.,Elwell,M.R., Montgomery,C.A. and Mackenzie,W. F., (eds) Pathology ofthe Fischer Rat, Reference and Atlas. Academic Press, San Diego, CA.,pp. 555–564.

47.Sidransky,D., Von Eschenbach,A., Tsai,Y.C., et al. (1991) Identification ofp53 gene mutations in bladder cancers and urine samples. Science, 252,706–709.

48.Hollstein,M., Sidransky,D., Vogelstein,B. and Harris,C.C. (1991) p53mutations in human cancers. Science, 253, 49–53.

49.Masui,T., Dong,Y., Yamamoto,S., Takada,N., Nakanishi,H., Inada,K.,Fukushima,S. and Tatematsu,M. (1996) p53 mutations in transitional cellcarcinomas of the urinary bladder in rats treated with N-butyl-N-(4-hydroxybutyl)-nitrosamine. Cancer Lett., 105, 105–12.

50.Hsu,C.H., Yang,S.A., Wang,J.Y., Yu,H.S. and Lin,S.R. (1999) Mutationalspectrum of p53 gene in arsenic-related skin cancers from the blackfootdisease endemic area of Taiwan. Br. J. Cancer, 80, 1080–1086.

51.Shibata,A., Ohneseit,P.F., Tsai,Y.C., Spruck,C.H., 3rd, Nichols,P.W.,Chiang,H.S., Lai,M.K. and Jones,P.A. (1994) Mutational spectrum in thep53 gene in bladder tumors from the endemic area of black foot diseasein Taiwan. Carcinogenesis, 15, 1085–1087.

52.Cohen,S.M., Yamamoto,S., Cano,M. and Arnold,L.L. (2001) Urothelialcytotoxicity and regeneration induced by dimethylarsinic acid in rats.Toxicol. Sci., 59, 68–74.

53.Hartwell,L. (1992) Defects in a cell cycle checkpoint may be responsiblefor the genomic instability of cancer cells. Cell, 71, 543–546.

54.Steeg,P.S. and Abrams,J.S. (1997) Cancer prognostics: past, present andp27 [news; comment]. Nat. Med., 3, 152–154.

55.Parke,D.V. and Ioannides,C. (1990) Role of cytochromes P-450 in mouseliver tumor production. Prog. Clin. Biol. Res., 331, 215–230.

56.Klaunig,J.E., Xu,Y., Isenberg,J.S., Bachowski,S., Kolaja,K.L., Jiang,J.,Stevenson,D.E. and Walborg,E.F., Jr. (1998) The role of oxidative stressin chemical carcinogenesis. Environ. Health. Perspect., 106 (Suppl. 1),289–295.

1397

57.Nishikawa,T., Wanibuchi,H., Ogawa,M., et al. (2002) Promoting effectsof monomethylarsonic acid, dimethylarsinic acid and trimethylarsine oxideon induction of rat liver preneoplastic glutathione S-transferase placentalform positive foci: A possible reactive oxygen species mechanism. Int. J.Cancer (In press).

58.Trush,M.A. and Kensler,T.W. (1991) An overview of the relationshipbetween oxidative stress and chemical carcinogenesis. Free Radic. Biol.Med., 10, 201–209.

59.Halliwell,B. (1996) Mechanisms involved in the generation of free radicals.Pathol Biol (Paris), 44, 6–13.

60.Yamanaka,K., Hasegawa,A., Sawamura,R. and Okada,S. (1989)Dimethylated arsenics induce DNA strand breaks in lung via the productionof active oxygen in mice. Biochem. Biophys. Res. Commun., 165, 43–50.

61.Yamanaka,K., Hoshino,M., Okamoto,M., Sawamura,R., Hasegawa,A. andOkada,S. (1990) Induction of DNA damage by dimethylarsine, a metaboliteof inorganic arsenics, is for the major part likely due to its peroxyl radical.Biochem. Biophys. Res. Commun., 168, 58–64.

62.Dizdaroglu,M. (1991) Chemical determination of free radical-induceddamage to DNA. Free. Radic. Biol. Med., 10, 225–242.

63.Romanenko,A., Morimura,K., Wanibuchi,H., Salim,E.I., Kinoshita,A.,Kaneko,M., Vozianov,A. and Fukushima,S. (2000) Increased oxidativestress with gene alteration in urinary bladder urothelium after the Chernobylaccident. Int. J. Cancer, 86, 790–798.

64.Schenk,H., Klein,M., Erdbrugger,W., Droge,W. and Schulze,O.K. (1994)Distinct effects of thioredoxin and antioxidants on the activation oftranscription factors NF-kappa B and AP-1. Proc. Natl Acad. Sci. USA,91, 1672–1676.

65.Kerr,L.D., Inoue,J. and Verma,I.M. (1992) Signal transduction: the nucleartarget. Curr. Opin. Cell. Biol., 4, 496–501.

66.Rin,K., Kawaguchi,K., Yamanaka,K., Tezuka,M., Oku,N. and Okada,S.(1995) DNA-strand breaks induced by dimethylarsinic acid, a metaboliteof inorganic arsenics, are strongly enhanced by superoxide anion radicals.Biol. Pharm. Bull., 18, 45–48.

67.Weinberg,R.A. (1991) Tumor suppressor genes. Science, 254, 1138–1146.68.Bishop,J.M. (1987) The molecular genetics of cancer. Science, 235,

305-311.69.Copeland,E.S. (1983) A National Institutes of Health Workshop report.

Free radicals in promotion: a chemical pathology study section workshop.Cancer Res., 43, 5631–5637.

70.O’Connell,J.F., Klein Szanto,A.J., DiGiovanni,D.M., Fries,J.W. andSlaga,T.J. (1986) Enhanced malignant progression of mouse skin tumors bythe free-radical generator benzoyl peroxide. Cancer Res., 46, 2863–2865.

71.Slaga,T.J. (1983) Overview of tumor promotion in animals. Environ.Health Perspect., 50, 3–14.

Received December 21, 2001; revised April 16, 2002; accepted May 1, 2002