Carcinogenicity of dimethylarsinic acid in Ogg1‐deficient mice

Abstract

Oxidative stress to DNA is recognized as a mechanism underlying carcinogenic effects of some environmental agents. Here, we hypothesized that dimethylarsinic acid (DMA^V), an organic metabolite of inorganic arsenic in humans, might exert carcinogenic potential in a mouse line carrying a mutant Mmh allele of the Mmh/OGG1 gene encoding the enzyme 8‐hydroxyguanine DNA glycosylase 1 (OGG1). Ogg1 mutant and wild type mice were treated with DMA^V in their drinking water at a dose of 200 p.p.m. for up to 72 weeks. All DMA^V‐treated Ogg1 ^−/–animals developed tumors, with a tendency for lower total incidences in the Ogg1 ^+/+ cases. Lung tumors in particular were induced as compared to the lack in non‐carcinogen controls and were significantly more frequent in the homozygotes. At week 4, the levels of DNA 8‐OH‐dG and cell proliferation were significantly elevated in the lungs of non‐treated Ogg1 ^−/– as compared to Ogg1 ^+/+ mice and were strongly enhanced by DMA^V treatment. Marked induction of Pola1, Cyp7b1, Ndfua3, Mmp13 and other genes specific to cell proliferation, cell signaling and xenobiotic metabolism in the lungs of DMA^V‐treated Ogg1 ^−/– mice was found. Electron microscopic examination revealed the growth of microvilli, with increased numbers of mitochondria only in lungs and lung tumors of DMA^V‐exposed Ogg1 ^−/– mice. Therefore, we strongly suggest that DMA^V exerts carcinogenicity in the lungs of Ogg1 ^−/– mutant mice, with a possible role for persistent accumulation of DNA oxidative adducts. (Cancer Sci 2007; 98: 803–814)

DNAdamage and defective DNA repair are considered to be essential for the initiation of carcinogenesis and age‐related changes in both the efficacy and the rate of DNA repair in mammals to modify susceptibility to exogenous or endogenous carcinogens.^{( 1 )} Lesions in genomic DNA caused by reactive oxygen species (ROS) are closely associated with aging and various diseases, including cancer, a typical major example being 8‐hydroxy‐2′‐deoxyguanosine (8‐OH‐dG). This is produced by exposure to various carcinogens and is considered to be a critical mutagenic adduct.^{( 2 )} Mispairing of this oxidized base with adenine results in GC→TA transversion mutations found in mammalian cells.^{( 3 )} In Saccharomyces cerevisiae and various bacteria, 8‐hydroxyguanine (8‐OH‐G) is excised from DNA by DNA glycosylases (catalyzed by MutM and Ogg1, respectively) with associated lyase activity for chain cleavage. A mammalian homolog of 8‐OH‐G glycosylase/apurinic, apyrimidinic lyase (AP lyase; MutM homolog, MMH) has already been identified and cloned and spontaneously developing tumors in the lungs and ovaries, as well as lymphomas, were recently detected in Myh and Ogg1 knockout mice.^{( 4} , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ^{11 )} Furthermore, chronic ultraviolet (UV) B irradiation of Ogg1 knockout mice have been shown to result in skin cancer development.^{( 12 )} However, it still remains unclear how deletion of this enzyme affects the carcinogenicity of chemicals in vivo.

The Mmh/OGG1 homozygous mutant mice used in the present study have a physically normal appearance, but lack nicking activity in liver extracts for substrate DNA containing 8‐OH‐dG, exhibiting a 3‐ and 7‐fold increased accumulation of this adduct at 9 and 14 weeks of age, respectively, as compared with wild type or heterozygous mice.^{( 13 )} Substantial elevation in the spontaneous mutation frequency has been identified in Mmh/OGG1 mutant mice bearing transgenic gpt genes and Arai et al. recently reported that administration of potassium bromate (KBrO₃) to Ogg1‐deficient mice resulted in a tremendous increase (∼70‐fold) in 8‐OH‐dG levels in kidney DNA, with consequent GC→TA transversions and deletions.^{( 13} , ^{14 )} Furthermore, administration of KBrO₃ to Ogg1 mutant mice followed by partial hepatectomy also led to a 3.5‐fold increase in mutation frequency during liver regeneration.^{( 15 )}

Arsenics are of world‐wide concern as environmental carcinogens and epidemiological studies have shown that chronic exposure of humans to inorganic arsenic compounds (trivalent arsenite (As³⁺) or pentavalent arsenate (As⁵⁺)) increases incidences of cancer of the lung, skin, liver, and urinary bladder, also being associated with liver injury and peripheral neuropathy.^{( 16} , ^{17 )} Although there is strong evidence for the carcinogenicity of inorganic arsenic compounds in humans, the molecular mechanisms remain incompletely defined. It is reasonable to suspect that more than one mechanism may be involved in arsenicals carcinogenesis because they can be carcinogenic in many diverse tissues in both humans and animals.^{( 18 )} Defining these mechanisms is critical to determining the nature and extent of the human health hazard presented by environmental arsenic exposure and to the development of intervention and prevention strategies. In this regard, it has become very important to develop the mouse models of arsenic‐induced carcinogenesis. Recent data from biochemical and carcinogenic studies using laboratory animals suggest that methylation of inorganic arsenic to form both dimethylarsinic acid (DMA^V) and methylarsonic acid (MMA^V), which are the major pathways of biotransformation and detoxification requiring reduced cellular glutathione (GSH), may paradoxically also result in toxification.^{( 18 )} DMA^V, one of the organic methylated arsenics, has been used as a herbicide and is also a major metabolite formed after exposure to As³⁺ or As⁵⁺ inorganic arsenic via ingestion or inhalation in both humans and most rodents.^{( 19 )} DMA^V has been considered to be over 10‐fold less acutely toxic than inorganic arsenic.^{( 20 )}

Initiation‐promotion studies have demonstrated that DMA^V acts as a multiorgan tumor promoter in rats and mice and A/J mice treated with 400 p.p.m. DMA^V in their drinking water for 50 weeks developed more pulmonary tumors than untreated animals.^{( 21} , ²² , ²³ , ²⁴ , ^{25 )} A significant dose‐dependent increase in preneoplastic lesions, papillomas and carcinomas in the urinary bladders of rats was also found with DMA^V treatment after initiation.^{( 21 )} Furthermore, in a long‐term carcinogenicity test, DMA^V application in the drinking water resulted in induction of bladder carcinomas in rats.^{( 26 )} It has been suggested that in mice this damage might be due primarily to peroxyl radicals and production of ROS in pulmonary tissues.^{( 27 )} Tumor promotion might thus be related to oxidative stress.^{( 28} , ^{29 )} However, until now, DMA^V has not been found to exert initiation activity for mouse tumorigenesis and this is feasible given differences in metabolism from the rat case.^{( 18 )}

In the present study, to cast light on the relationship between human carcinogenic arsenics and oxidative stress, the carcinogenic effects of DMA^V were examined in mutant Ogg1 mice. At the end of the treatment period (72 weeks), multi‐organ histopathological, immunohistochemical, biochemical, genomic, proteomic and electron microscopic analyses were conducted with a particular focus on cell proliferation, apoptosis, formation of oxidative DNA modifications and alterations to cell ultrastructure in mouse tissues targeted by DMA^V. The findings indicate that lung carcinogenesis in Ogg1 ^−/– homozygotes is associated with marked elevation of DNA 8‐OH‐dG levels and cellular proliferation.

Materials and Methods

Chemicals.  DMA^V (purity 99%) was purchased from Sigma (St. Louis, MO, USA). Other chemicals were from Sigma or Wako Pure Chemical Industries (Osaka, Japan).

Maintenance of mice.  Animals were housed in an animal facility maintained on a 12 h (07:00–19:00) light/dark cycle, at a constant temperature of 22 ± 1°C, with a relative humidity of 44 ± 5%, and given free access to water and food ad libitum. Throughout the experiment they were administered a common basal pellet diet (CE2; Clea Japan, Tokyo, Japan). In the present study, 14‐week‐old mice at the commencement were used, as our previous data demonstrated that chronic DMA^V administration is particularly toxic in young animals.^{( 19} , ^{30 )}

Experimental protocol.   Experiment 1. 14‐week‐old male (20) and female (20) Ogg1 ^−/–, and male (22) and female (22) Ogg1 ^+/+mice of a C57Bl/6 J background were divided into two groups each (Ogg1 ^−/– male and female: five mice per group; Ogg1 ^−/– male and female: five or six mice per group) and administered DMA^V at a concentration of 200 p.p.m. (Ogg1 ^−/– and Ogg1 ^+/+ treatment groups) ad libitum in drinking water for 72 weeks or tap water as controls. Their body weight, water intake and food consumption were recorded every week for the first 12 weeks of the study and subsequently once every 4 weeks. Mice were observed daily, killed when becoming moribund during the study or at the end of the experiment at week 72, and autopsied for macroscopic, histopathological (H&E), immunohistochemical, biochemical and electron microscopic examinations.

At death, lungs were removed and the left lobe was fixed in 10% buffered formalin or Bouin solution for histological and immunohistochemical assessment. Part of the right lungs were treated with 2% glutaraldehyde and 2% paraformaldehyde in 0.12 mol/L phosphate buffer (PB) (pH 7.2) for subsequent electron microscopic analysis and the remainder were frozen in liquid nitrogen and stored at −80°C for molecular assessment. Lung nodules were accurately separated from normal lung and divided into separate pieces fixed in 10% buffered formalin, Bouin or 2% glutaraldehyde and 2% paraformaldehyde in 0.12 mol/L PB (pH 7.2).

Experiment 2.

14‐week‐old male and female Ogg1 ^−/– (40) and Ogg1 ^+/+ (40) mice on a C57Bl/6 J background (10 animals/group) were treated with DMA^V at a concentration of 0 or 200 p.p.m. for 4 weeks. Mice lungs were used for the analysis of 8‐OH‐dG levels, proliferating cell nuclear antigen (PCNA), apoptosis indices and alteration to gene expression using the Affimetrix oligonucleotide microarray system.

Analysis of 8‐OH‐dG formation.  In experiment 2, mice lung DNA 8‐OH‐dG levels were determined by a high performance liquid chromatography (HPLC)‐electrochemical detection (ECD) method.^{( 2 )} Furthermore, in the long term experiment 1, immunohistochemical staining of 8‐OH‐dG was carried out as described previously.^{( 31 )}

Immunohistochemistry for PCNA and apoptosis.

In experiments 1 and 2, immunohistochemical staining for PCNA and apoptosis was carried out using an anti‐PCNA rabbit polyclonal (PC‐10, IgG; Santa Cruz Biotechnology, Santa Cruz, CA, USA; 1:500) and single‐stranded DNA ((ssDNA); IgG, 100 µg/mL, DAKO JAPAN, Kyoto, Japan; 1:400) antibodies, respectively, on sections of lung and bladder as described previously.^{( 31 )} PCNA and apoptosis indices were estimated for epithelial cells of lung alveoli with counts of clearly brown staining positive nuclei per 1000 cells.

Transmission electron microscopy.  In long‐term experiment 1, portions of the harvested lungs and urinary bladders were minced into 1‐mm³ pieces, fixed in 0.12 mol/L PB (pH 7.2), containing 2% (v/v) glutaraldehyde and 2% paraformaldehyde (v/v) for 1.5 h, washed in 0.12 mol/L PB (pH 7.2) and 7% sucrose in PB for 7 min twice and postfixed in 2% osmium tetroxide. They were then dehydrated through ascending concentrations of ethanol and propylene oxide, embedded in Durcupan (TAAB Laboratories, Aldermaston, UK), and polymerized at 60°C overnight. Ultrathin sections exhibiting a pale gold interference color from selected blocks were then cut with a diamond knife, mounted on grids, and stained with ethanol uranyl acetate and lead citrate before transmission electron microscopy (TEM) examination (Hitachi H‐7500, Tokyo, Japan).

Affymetrix oligonucleotide microarray analysis.  In experiment 2, total RNA was isolated from mouse lung tissue (separate lung lobe without bronchi; pieces <5 mm in diameter) as described previously.^{( 31 )} Development of lung nodules was not observed at week 4. Therefore, only normal‐appearing lungs were submitted to the microarray analysis. Eight‐microgram pooled aliquots of total RNAs from five mice in each group were treated with DNase 1 and processed for PolyA⁺ RNA enrichment and generation of cDNA probes using an Affymetrix GeneChip T7‐Oligo(dT) Promoter Primer Kit (Affymetrix) according to the manufacturer's protocol. Biotin‐labeled antisense cRNA was synthesized by in vitro transcription reaction (IVT) using an RNA Transcript Labeling Kit (Affymetrix, P/N 900182), purified and fragmented, and hybridized to Mouse E430 Plus 2.0 arrays (MOE430 Plus 2.0), with 45 101 probe sets, according to the manufacturer's instructions (Affymetrix). Affymetrix GCOS software version 1.0 was used for normalization and for monitoring specific hybridization and microarray data were analyzed using GeneSpring software version 7.2 (Silicon Genetics, Redwood City, CA, USA). Each array was normalized to the 50th percentile and each gene was normalized to the control (non‐treated Ogg1 ^−/– or Ogg1 ^+/+ groups). Log2 ratios of the hybridization intensities of mouse lung samples were used to represent the relative gene expression level. To minimize the effects of measurement variation introduced by artificial sources during the experiments, we only included spots with up‐regulation or down‐regulation at least twofold. Microarray analysis was repeated twice to check the reproducibility of the data and mean values for gene expression was calculated. One‐way‐anova was applied to compare replicates mean values of control and experimental groups (Gene Spring). The microarray set for statistical analysis was three. Comparisons of gene expression across treatment groups were carried out using Venn Diagrams and clustering was carried out with the Condition Tree algorithm. In addition, the Gene Ontology grouping was used in conjunction with Gene Spring to identify pathways and functional groups of genes. The Ingenuity program (Ingenuity Systems, Mountain View, CA, USA) was also used to identify networks of interacting genes and other functional groups.

Measurement of mRNA in mouse lung tissues by real‐time quantitative‐polymerase chain reaction.  For the confirmation of Affimetrix microarray analysis results, real‐time quantitative‐polymerase chain reaction (Q‐PCR) was carried out using TaqMan probes for 490 C. Primer sequences were designed with Primer Express software (Applied Biosystems, USA): Mm00447106_m1 for polymerase (DNA‐directed), alpha 1 (NM_008892); Mm00484157_m1 for cytochrome P450, 7b1 (NM_007825); Mm00772875_m1 for nicotineamide adenine dinucleotide reduced form (NADH) dehydrogenase (ubiquinone) 1 alpha subcomplex, 3 (NM_025348); Mm00439491_m1 for matrix metalloproteinase 13 (NM_008607); Mm00834384_g1 for heat shock protein 1 (NM_013560); and Mm00607939 for beta‐actin, cytoplasmic (NM_007393). The cDNA product from each sample was used to carry out Q‐PCR according to the manufacturer's instructions. We used beta‐actin mRNA as an internal control for comparison and to normalize the data. Assays were carried out in triplicate using Applied Biosystems Model 7700 instruments.

Statistical analysis.  The significance of differences between mean values was analyzed using the Dunnet's two‐tailed test and Fisher's protected least significant difference (PLSD) method using the StatView–J 5.0 program (Berkeley, CA, USA). Kaplan‐Meier analysis was conducted to determine the survival rates for Ogg1 knockout, heterozygous and wild type mice. One‐way‐anova was applied in Affymetrix oligonucleotide microarray analysis to compare replicates mean values of control and experimental groups (Gene Spring). The significance of differences in lesion incidences between groups was assessed by χ² test or Fisher's exact probability test.

Results

General observations (Experiment 1).  During the experiment no toxicity of DMA^V in mice was apparent and no significant differences were found among the groups in the quantities of food and water consumed or body weight gained (data not shown). Total DMA^V intake was comparable among Ogg1 homozygous and wild type mice.

Significant increase in relative lung, liver and spleen weights was found in the group of Ogg1 ^−/– mice treated with DMA^V (male lung: 0.017 ± 0.003 (P < 0.01); female lung: 0.012 ± 0.005 (P < 0.5); male liver: 0.07 ± 0.02 (P < 0.05); female liver: 0.082 ± 0.025 (P < 0.05); male spleen: 0.006 ± 0.006, female spleen 0.019 ± 0.013 (P < 0.05); as compared to corresponding Ogg1 ^−/– controls (male lungs: 0.012 ± 0.002; female lungs: 0.007 ± 0.001; male liver: 0.042 ± 0.004; female liver: 0.049 ± 0.003; male spleen: 0.002 ± 0.001; female spleen: 0.003 ± 0.001). Furthermore, lungs weight of control Ogg1 ^−/– animals was significantly higher (P < 0.05) than that of control Ogg1 ^+/+ (male and female lungs: 0.005 ± 0.001); relative liver, kidney and spleen weights did not differ between Ogg1 ^+/+ DMA^V‐treated and control groups.

Survival curves for Ogg1 homozygous and wild type age‐matched littermates are shown in Fig. 1a–c, respectively. At week 38, the first Ogg1 ^+/+ DMA^V‐treated male mouse was found dead with a renal adenocarcinoma. The number of surviving Ogg1^−/– mice administered DMA^V started to decrease at week 58 due to early tumor development as compared to Ogg1 ^+/+ mice (Fig. 1). The development of various spontaneous tumors in wild type mice was the reason for their mortality, while Ogg1^−/– mice (less induction of spontaneous tumors compared to Ogg1 ^+/+) died predominantly due to the occurrence of malignant lymphomas.

Figure 1.

Survival curves for Ogg1 ^−/– (a) and Ogg1 ^+/+ (b) mice. Non‐treated Ogg1 ^−/– mice were found to be more healthy and long‐lived as compared to the Ogg1 ^+/+ control mice. The number of surviving Ogg1 ^−/– mice administered dimethylarsinic acid (DMA^V) started to decrease at week 58 due to tumor development.

Lesion incidence, multiplicity and histopathology.  Data for tumor incidences, multiplicities and organ distributions of tumors in DMA^V‐treated and non‐treated mice with reference to the Ogg1 ^−/– and Ogg1 ^+/+ mice genotypes are presented in Table 1. Because of highly similar tumor induction in male and female mice, combined data are shown. Ogg1 ^−/– mice treated with DMA^V were more susceptible to the induction of tumors as compared to wild type littermates, although no tumors were found in untreated Ogg1 ^−/– control mice, this being the reason for the high survival rate. Significant increases in total tumor incidence (P < 0.0001) and multiplicity (P < 0.001) were observed in the DMA^V‐applied Ogg1 ^−/– group as compared to the control Ogg1 ^−/– or DMA^V‐treated or non‐treated wild type animals.

Table 1.

Tumor incidence, multiplicity and histopathology

Site and type of tumor†	Incidence (%)		(Multiplicity (no. tumors/mouse)
	DMAV, 200 p.p.m.	Control	DMAV, 200 p.p.m.	Control
Ogg1 −/–
No. mice	10	10	10	10
No. tumor bearing mice (%)	10 (100%)**, ***	0
M. lymphoma/leukemia	4 (40)	0	0.4 ± 0.5	0
Liver
HCC	1 (10)	0	0.1 ± 0.31	0
Lung
Hyperplasia	10 (100)	10 (100)	5.0 ± 1.0*, ****	6.8 ± 1.09
Adenoma	2 (20)	0	0.2 ± 0.42	0
Adenocarcinoma	3 (30)	0	0.3 ± 0.48	0
Total tumors	5 (50)*, ***	0	0.5 ± 0.52*, ***	0
Mammary gland	1 (10)	0	0.1 ± 0.3	0
Adenocarcinoma
Ogg1 +/+
No. mice	12	10	12	10
No. tumor bearing mice (%)	6 (50%)	5 (50%)
M. lymphoma/leukemia	2 (16.7)	2 (20)	0.17 ± 0.38	0.2 ± 0.3
Liver
HCC	1 (8.3)	0	0.08 ± 0.28	0
Lung
Hyperplasia	10 (100)	2 (20)	3.00 ± 2.76*	0.33 ± 0.10
Adenoma	0	1 (10)	0	0.10 ± 0.31
Adenocarcinoma	0	0	0	0
Total tumors	0	1 (10)	0	0.10 ± 0.31
Mammary gland
Adenocarcinoma	1 (8.3)	0	0.08 ± 0.28	0
Renal
Adenocarcinoma	1 (8.3)	0	0.08 ± 0.28	0
Subcutis
Fibrosarcoma	0	1 (10)	0	0.1 ± 0.31
Skin
Sarcoma	1 (8.3)	0	0.08 ± 0.28	0

^† Only organs with neoplastic lesions are listed.

^* P < 0.05,

^** P < 0.0001 versus corresponding non‐treated controls;

^*** P < 0.05,

^**** P < 0.0001 versus Ogg1 ^+/+ DMA^V‐treated or control groups. DMA^V, dimethylarsinic acid; HCC; hepatocellular carcinoma; M. lymphoma, malignant lymphoma.

The tumors induced in the DMA^V‐treated group of Ogg1 ^−/– mice were mainly in the lungs (50%) and malignant lymphomas/leukemias (40%) (Table 1, Fig. 2a,b). Significant elevation of lung tumor multiplicity was apparent in DMA^V‐treated Ogg1 homozygous mice when compared to their control Ogg1 ^–/– and DMA^V‐treated Ogg1 ^+/+ counterparts (P < 0.05). Total incidence of lung adenoma and adenocarcinoma in male Ogg1 ^–/– mice was 60%, a little higher than in females (40%). Both adenomas and carcinomas were induced at the same rate (1:1). On the contrary, no lung tumors were observed in DMA^V‐treated male Ogg1 ^{+ /+} animals, and only one female Ogg1 ^+/+ mouse in the control group developed spontaneous adenocarcinoma (20%). Furthermore, a significant increase of hyperplastic epithelial lesion multiplicity was observed in the lungs of non‐treated Ogg1 ^–/– mice as compared to non‐treated wild type mice (Table 1). The trend for increase of lung hyperplasia was found in the lungs of Ogg1 ^{+ /+} animals treated with DMA^V. A slight decrease in the number of hyperplastic lesions but an increase of tumor development was observed in the lungs of DMA^V‐treated Ogg1 ^–/– animals as compared to non‐treated Ogg1 ^–/– mice, indicating that DMA^V exerted a promoting effect on lung carcinogenesis in Ogg1 ^–/– mice (Table 1). Malignant lymphomas/leukemias were mainly observed in Ogg1 ^−/– mice (40%) treated with DMA^V, although intergroup differences in incidence and multiplicity were not significant. Interestingly, no tumors were found in non‐treated Ogg1 ^−/– animals (Table 1). In contrast, 50% of DMA^V‐treated or control wild type littermates developed tumors, which included malignant lymphomas/leukemias, mammary tumors, renal adenocarcinomas, subcutis fibrosarcomas and skin sarcomas (Table 1). Spontaneous tumors of non‐treated Ogg1 ^+/+ mice were malignant lymphomas, lung adenomas and subcutis fibrosarcomas.

Figure 2.

Results of histopathological (H&E) staining and transmission electron microscopy (TEM) in the lungs of Ogg1 ^−/– mice maintained on 200 p.p.m. dimethylarsinic acid (DMA^V). (a) (× 400) and (b) (× 400) Adenocarcinoma and adenoma, respectively, in the lung of the DMA^V‐treated Ogg1 ^−/– mouse (H&E). (c) TEM of alveoli from a non‐tumorous area in a Ogg1 ^−/– animal treated with DMA^V and (d) alveoli of a non‐treated Ogg1 ^−/– animal. (e, f) TEM of a lung adenocarcinoma from an Ogg1 ^−/– animal treated with DMA^V. Note the numerous mitochondria, microvilli and abundant vesicles. Those changes were absent in Ogg1 ^+/+ mice (data not shown). Co, collagen fibres; CV, clatrin vesicles; D, desmosomes; Ec, euchromatin; GA, Golgy apparatus; Hc, heterochromatin; LB, lamella bodies; Mi, mitochondria; Mv, microvilli; N, nucleus; Nc, nucleolus; RER, rough endoplasmic reticulum; Vs, vesicles. Original magnifications: × 400 (a); × 400 (b); × 25 000 (c–f), Scale bar: 860 nm (c); 1.2 µm (d); 670 nm (e); 565 nm (f).

Extensive examination of the urinary bladders of all treated and non‐treated mice revealed simple hyperplasia only in DMA^V‐administered homozygous animals (data not shown). No urinary calculi were found in Ogg1 ^−/– and Ogg1 ^+/+ mice after DMA^V application.

Changes of cell ultrastructure detected in the lungs and lung tumors (Experiment 1).  Increased numbers of mitochondria, amounts of endoplasmic reticulum, and collagen fibers, with elongated microvilli only in the lungs of Ogg1 ^−/– mice were detected by TEM after DMA^V treatment for 72 weeks. Ten sections from each animal were analyzed to ensure that there real quantitative differences existed. Those changes were mostly observed in Type II pneumocytes of Ogg1 ^−/– mice but not wild type animals treated with DMA^V (Fig. 2c). TEM examination of lung tumors developing in Ogg1 ^−/– animals after DMA^V treatment, revealed an increased number of mitochondria and lamella bodies and markedly developed rough endoplasmic reticulum (Fig. 2e,f) and large numbers of pleomorphic microvilli (data not shown), in line with changes in Type II pneumocytes of non‐tumorous areas (Fig. 2c).

Lung 8‐OH‐dG, cellular proliferation and apoptosis indices (Experiments 1 and 2).  Elevated levels of DNA 8‐OH‐dG, detected by HPLC‐ECD analysis, were found in the lungs of non‐treated Ogg1 ^−/– animals as compared to their wild‐type counterparts (Fig. 3a). Furthermore, treatment with DMA^V caused a significant increase of 8‐OH‐dG levels in Ogg1 ^−/– mouse lungs. Immunohistochemical determination of 8‐OH‐dG also showed a strong overexpression in the lungs of Ogg1 ^−/– mice due to the DMA^V treatment (Fig. 3b).

Figure 3.

Formation of 8‐hydroxy‐2′‐deoxyguanosine (8‐OH‐dG) and cellular proliferation and apoptosis indices in the lungs of Ogg1 ^−/– and Ogg1 ^+/+ mice treated with dimethylarsinic acid (DMA^V) at 200 p.p.m. for 72 weeks. (a) 8‐OH‐dG (c) proliferating cell nuclear antigen (PCNA) and (d) apoptosis indices. (b) 8‐OH‐dG staining pattern in DMA^V‐treated and control Ogg1 ^−/– mouse. †Normal‐appearing area; *Significant at P < 0.05 versus Ogg1 ^−/– control group; **Significant at P < 0.001 versus Ogg1 ^−/– control group; ***Significant at P < 0.0001 and ****Significant at P < 0.05 versus Ogg1 ^+/+ control group.

Proliferating cell nuclear antigen positive indices for the lungs of non‐treated Ogg1 ^−/– mice were found to be significantly higher than in control Ogg1 ^+/+ animals. These data were confirmed in a short experiment (Experiment 2), in which Ogg1 ^−/– and wild type mice were administered DMA^V for 4 weeks. DMA^V treatment resulted in further significant elevation of PCNA positive indices in the normal‐appearing areas of lungs and urinary bladder mucosa (data not shown) of homozygous mice as compared to control Ogg1 ^−/– or their wild‐type age‐matched littermates (Fig. 3c).

Apoptosis indices tended to be elevated only in the lungs of Ogg1 ^−/– animals administered DMA^V but without significance (Fig. 3d). Tumors were found to be negative for apoptosis. ssDNA staining patterns in serial sections were similar to those for 8‐OH‐dG and increase of the apoptotic indices paralleled elevation in 8‐OH‐dG and the PCNA indices (data not shown).

Changes in gene expression triggered by DMA^V (Experiment 2).   The results of the Affymetrix oligonucleotide microarray analysis of differentially expressed genes in the lungs of mice, obtained in the short‐term experiment after 4 weeks of DMA^V administration are presented in Table 2. The results were compared with Gene Spring software and a strong concordance was found. Furthermore, the data obtained by microarrays on the expression of Pola1, Cyp7b1, Ndfua3, Mmp13, and Hspb1 were checked by real‐time Q‐PCR analysis, which showed the same patterns (Fig. 4a–e).

Table 2.

Differentially expressed genes in the lungs of Ogg1 ^−/– and Ogg1 ^+/+ mice treated with dimethylarsinic acid (DMA^V) for 4 weeks

Animals	Function Gene	Accession No.	Ratio	P‐value
Male Ogg1 −/–	Cell proliferation		DMA, 200 p.p.m. vs vehicle
	Polymerase (DNA directed), alpha 1	NM_008892	2.268	0.011
	Mitogen activated protein kinase kinase 6	BB540608	4.46	0.001
	Topoisomerase (DNA) II alpha	BB749838	2.922	0.006
	Heat shock protein 1A	AW763765	2.547	0.009
	Heat shock protein 1B	M12573	2.344	0.009
	Vascular endothelial growth factor C	AW228853	2.687	0.006
	Sulfotransferase, hydroxysteroid preferring 2	NM_009286	5.381	0.001
	Phospholipase A2, group VII	AK005158	3.321	0.003
	Extracellular matrix
	Catenin delta 2	BB431091	3.561	0.002
	Fibroblast growth factor 23	BB840359	0.292	0.015
	Stress response
	Cytochrome P450, family 7, subfamily b, polypeptide 1	NM_007825	2.318	0.012
	NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 3	AA060162	3.345	0.002
Female Ogg1 −/–	Cell proliferation
	Polymerase (DNA directed), alpha 1	NM_008892	2.021	0.018
	Polymerase (DNA directed), gamma	AV351762	2.023	0.019
	Polymerase (DNA directed), iota	NM_011972	2.193	0.011
	Polymerase (DNA directed), epsilon 2 (p59 subunit)	AF036898	2.255	0.022
	Topoisomerase (DNA) I	BB127876	2.015	0.029
	Mitogen‐activated protein kinase kinase kinase kinase 5	AW555664	2.356	0.027
	Mitogen‐activated protein kinase kinase kinase kinase 14	BG072756	2.867	0.02
	Ornithine decarboxylase antizyme inhibitor	BI155585	0.301	0.043
	Stress response
	Heat shock protein 4	NM_015765	2.281	0.028
	Heat shock 70 kDa protein 12A	BQ17716	2.245	0.023
	Cytochrome P450, family 7, subfamily b, polypeptide 1	NM_007825	2.318	0.03
	NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 3	AA060162	4.282	0.004
	Superoxide dismutase 1, soluble	BB530180	0.308	0.042
	Matrix metalloproteinase 13	NM_008607	2.564	0.029
	Carbonic anhydrase 13	AK010166	2.431	0.033
	Guanine nucleotide binding protein, gamma 5 subunit	BC010725	2.062	0.042
	Apoptosis
	Transforming growth factor, beta receptor I	NM_009370	0.197	0.027
	Ubiquitin specific protease 7	C77542	0.0581	0.001
Male Ogg1 +/+	Stress response
	Metallothionein 1	BC027262	5.491	0.005
	Cytochrome b‐245, beta polypeptide	AV373944	2.105	0.045
	NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 3	AA060162	3.282	0.009
	Matrix metalloproteinase 9	D12712	2.604	0.011
	Adipsin	NM_013459	2.464	0.021
	Cell signaling
	Mitogen activated protein kinase kinase 6	BB540608	0.444	0.034
	Mitogen activated protein kinase kinase kinase 12	AW556977	0.422	0.018
	Extracellular matrix
	Procollagen, type XVII, alpha 1	NM_007732	0.182	0.039
Female Ogg1 +/+	Stress response
	Matrix metalloproteinase 17	NM_011846	2.478	0.043
	Cytochrome P450, family 7, subfamily b, polypeptide 1	NM_007825	2.214	0.002
	Superoxide dismutase 1, soluble	BB530180	0.428	0.043
	Carbonic anhydrase 3	AK003671	0.266	0.017
	Extracellular matrix
	Titin	BC025840	2.783	0.042
Male Ogg1 −/– vs Ogg1 +/+			Vehicle vs vehicle
	Xenobiotic metabolism
	Cytochrome P450, family 7, subfamily b, polypeptide 1	NM_007825	2.032	0.001
	Cell proliferation
	Polymerase (DNA‐directed), delta 4	BB825816	2.065	0.002
	Metabolism
	Phosphoenolpyruvate carboxykinase 1, cytosolic	AW106963	10.75	0.033
	Amine oxidase, copper containing 3	NM_009675	2.179	0.048
	Extracellular matrix, intracellular transport
	Karyopherin (importin) beta 1	BC004096	2.483	0.037
	Cadherin 4	NM_009867	2.784	0.028
	Stress response
	Osmotic stress protein	NM_011020	0.371	0.041
	Heat shock protein 1	NM_013560	0.364	0.046
	Heat shock protein 1A	BC006722	0.228	0.022
	Heat shock protein 1B	M12573	0.256	0.033
	Heat shock protein 105	D67017	0.251	0.02
	Heat shock protein 8	AW763765	0.211	0.011
	Adipsin	NM_013459	5.094	0.001
Female Ogg1 ‐/‐ vs Ogg1 +/+	Xenobiotic metabolism
	Cytochrome P450, family 7, subfamily b, polypeptide 1	NM_007825	2.032	0.046
	Cell proliferation and angiogenesis
	Cyclin C	AK014079	2.049	0.049
	Mitogen activated protein kinase 8	BB184171	3.009	0.013
	Matrix metalloproteinase 13	NM_008607	11.08	0.001
	Matrix metalloproteinase 14 (membrane‐inserted)	AI325305	2.121	0.045
	Matrix metalloproteinase 17	NM_0118	2.585	0.035
	Stress response
	Heat shock protein 1	AF047377	0.469	0.041
	Heat shock protein 1A	BC006722	0.301	0.021
	Heat shock protein 1B	M12573	0.409	0.04
	Extracellular matrix
	Protocadherin beta 16	BB027682	4.741	0.004

^† Genes, that were up‐regulated or down‐regulated versus vehicle control more than two‐fold in Ogg1 ^−/– and Ogg1 ^+/+ animals. The significance of differences in gene expression was analyzed by statistical analysis anova, one‐way test (GeneSpring, n = 3). NADH, nicotineamide adenine dinucleotide reduced form.

Figure 4.

Gene Spring clustering analysis of the Affymetrix oligonucleotide microarray data (a) Ingenuity pathway analysis (b) and conformation of microarray results by real‐time quantitative polymerase chain reaction (Q‐PCR) (c–f). (c) Polymerase (DNA‐directed) alpha 1 (d) Mmp13 (e) nicotineamide adenine dinucleotide reduced form (NADH) dehydrogenase (ubiquinone) 1 alpha, and (f) CYP7b1 mRNA expression. *P < 0.05 and **P < 0.0001 versus Ogg1 ^+/+ control and dimethylarsinic acid (DMA^V)‐treated groups.

Expression of 165 genes in Ogg1 ^−/– males and 182 genes in Ogg1 ^−/– females was changed in lungs following treatment with the 200 p.p.m. carcinogenic dose of DMA^V. Among them we observed up‐regulation of metabolism associated genes, including cytochrome P450 family 7, subfamily b, polypeptide 1 (CYP7b1), NADH dehydrogenase (ubiquinone) 1 alpha, cell signaling and cell proliferation related polymerase (DNA‐directed) alpha 1, gamma, iota, epsilon 2 (p59 subunit), MAPKKK6, MAPKKK5 and MAPKKK14, topoisomerase (DNA) II alpha, stress proteins like heat shock protein 1A and 1B, vascular endothelial growth factor C, sulfotransferase and phospholipase A2, group VII (Table 2 and Fig. 4).

In the lungs of animals in the Ogg1 ^−/– control group, when compared with wild type mice, overexpression of steroid and lipid metabolizing mono‐oxygenase CYP7b1, cell proliferation regulating polymerase (DNA‐directed), delta 4 (Pold4), and cytosolic phosphoenolpyruvate carboxykinase 1 (Pck1) was found in males and cyclin C, mitogen activated protein kinase 8 (MAPK8), and involved in angiogenesis matrix metalloproteinases 13, 14, 17 (Mmp13, Mmp14 and Mmp17) in females (Table 2). Furthermore, in control male and female Ogg1 ^−/– mutant mice, genes including stress proteins like heat shock proteins (HSP1, 1 A, 1B, 8 and 105) were down‐regulated.

DMA^V administration to Ogg1 ^+/+ male mice resulted in up‐regulation of stress response‐related genes including metallothionein, genes encoding xenobiotic metabolizing enzymes and matrix metalloproteinase nine. In females Mmp17, CYP7b1, SOD1 and titin were found to be induced (Table 2, Fig. 4).

A total of more than two‐fold of up‐ or down‐regulated genes were used for further comparative analyses (e.g. Gene Spring and Ingenuity Pathway analyses). When the microarray data from the treatment groups of combined replicates of male and female mice were clustered into a condition tree by Gene Spring analysis in a standard correlation using the average replicates for signature genes, the relationship of one treatment group to another was as expected. The Ogg1 ^−/– and Ogg1 ^+/+ male and female DMA^V‐treated groups were clustered together (Fig. 4). The data suggest that there are patterns of early gene expression changes that clearly distinguish DMA^V‐treated Ogg1 ^−/– with large numbers of overexpressed genes from all other groups, as well as the non‐treated Ogg1 ^−/– and Ogg1 ^+/+ groups.

Analysis of the microarray averaged data by uploading to Ingenuity Pathway Analysis revealed suppression of drug, lipid metabolism and endocrine system development and function in the non‐treated Ogg1 ^−/– mice lung comparing to the Ogg1 ^+/+ case. However, genes related to cancer, cellular growth, proliferation and the cell cycle were found to be up‐regulated (data not shown). Expression of genes interfering with cancer, cell cycle, cellular proliferation, DNA replication, recombination and repair, lipid metabolism, cell–cell signaling and interaction, molecular transport and tumor morphology pathways was found to be induced, but the immune response pathway was suppressed when Ogg1 ^−/– was administered by DMA^V. When DMA^V‐treated wild type and Ogg1 mutant mice were compared, it was found that genes involved in cancer, cellular growth and proliferation, lipid metabolism and cell–cell signaling and interaction were overexpressed in Ogg1 ^−/– animals. Furthermore, free radical scavenging was not found to be activated in Ogg1 ^−/– mice.

Discussion

The present investigation revealed that long‐term administration of DMA^V in drinking water exerts carcinogenicity in the lung of homozygous Ogg1 mutant mice that are deficient in the capacity to repair oxidative DNA base modifications. To our knowledge, this is the first evidence that DMA^V, which is negative in the Ames test for mutagenicity, exerts carcinogenic activity in a mouse lung associated with high levels of 8‐OH‐dG. The mechanism of DMA^V carcinogenicity in the lungs of Ogg1 ^−/– mice might thus be an accumulation of non‐repaired 8‐OH‐dG with high levels of cell proliferation, possibly leading to mutations and further elevation in cell proliferation, thus, resulting in initiation and promotion of lung carcinogenesis in those animals.^{( 15 )} It was earlier proposed that promotion by DMA^V might be due mostly to the induction of enzymes involved in xenobiotic metabolism, peroxyl ((CH₃)₂AsOO) radicals, and additional damage to DNA and cell proliferation.^{( 32 )} Constitutively high levels of cell proliferation in the lung tissue of Mmh/Ogg1‐deficient mice might cause accumulation of DNA damage as the new cell cycle begins with already damaged and mutated DNA.

Interestingly, no spontaneously developed tumors were found in non‐treated Ogg1 homozygous mice, which might be due to the activation of backup DNA repair or other adaptation mechanisms to high levels of oxidative stress. Our data differ from those recently obtained by Xie et al. and Sakumi et al. reporting spontaneously developed lymphomas, lung and ovary tumors in Ogg1 knockout mice.^{( 10} , ^{11 )} This is possibly due to the differences in the method of Ogg1 knockout mice generation. Variation in the levels of cell proliferation in the lungs of Ogg1‐deficient and wild type mice are in line with previously reported data demonstrating differential expression of DNA repair enzymes in animal tissues and by the high production of reactive oxygen species in the lung due to respiration.^{( 33 )}

The results of the present study are in concordance with our recent data obtained on investigation of the effects of potassium bromate in Ogg1 ^−/– mouse liver after partial hepatectomy, causing an increase in the mutation frequency.^{( 15 )} Recently, carcinogenicity of DMA^V has been accessed in Muta and p53 heterozygous knockout mice.^{( 34} , ^{35 )} When DMA^V was intraperitoneally injected into Muta mice at a dose of 10.6 mg/kg for five consecutive days it caused only a weak 1.3‐fold increase in the mutation frequency of the lacZ gene in the lung, and the authors suggested that the mutagenicity of DMA^V is too low to be detected in Muta mice.^{( 34 )} However, administration of DMA^V in drinking water to p53 heterozygous knockout mice at a dose of 200 p.p.m. led to significantly early induction of more numerous tumors in p53^+/– as compared to wild type mice.^{( 35 )} However, no evidence for organ‐tumor specificity was observed. Similar to the study with p53 heterozygous knockout mice, we observed a tendency for increase in the development of malignant lymphomas in female Ogg1 ^−/– mice, but did not find significant differences in the numbers of developing tumors between DMA^V‐treated and control wild type mice. Furthermore, DNA damage was not observed in the lungs of wild type and heterozygous animals, possibly because they were more resistant to DMA^V treatment than their Ogg1 ^−/– counterparts.

Recently, oxidative stress has been postulated as a novel mode of arsenic carcinogenicity and it has been shown that DMA^V‐induced lung‐specific DNA damage in mice can be attributed to free radicals, particularly peroxyl, superoxide or hydroxyl radicals, arising from the reaction of DMA^V with molecular oxygen in vivo.^{( 18} , ^{26 )} Depletion in cellular glutathione may be correlated with oxidative stress mediated by reactive oxygen species and induction of nitric oxide. The reaction and interaction of these reactive species with target molecules lead to oxidative stress, lipid peroxidation, DNA damage, and activation of signaling cascades associated with tumor promotion and/or progression.^{( 36 )} Antioxidants that can inhibit, reduce, or scavenge the production of ROS and reactive nitrogen species (RNS) induced by arsenic can not only decrease direct cellular damage such as lipid peroxidation, enzyme inactivation and DNA oxidation caused by arsenic, but also ameliorate cell injuries or death by redox signaling pathways activated by arsenic exposure.^{( 36 )} The interaction of arsenic with essential metals is not clear. It has been suggested that peroxyl radicals may play an important role in DNA damage and in tumor promotion.^{( 23} , ^{37 )} Apurinic/apyriminidic sites may form where initial damage occurs, and altered DNA may undergo beta‐elimination to form DNA single‐strand scissions or Schiff‐base reactions to form DNA‐protein cross‐links.^{( 38 )} Those forms of DNA damage are also specifically produced in mouse and rat lungs after acute oral administration of DMA^V, but are not observed in other tissues.^{( 39} , ^{40 )} Mice treated with DMA^V in drinking water excrete DMA^V primarily in the urine, though it may be in the form of an oxidizable complex,^{( 41 )} and approximately 3–4% of an oral dose of DMA^V (40 mg As/kg) administered to mice is excreted as trimethylarsine oxide (TMAO) over 48 h.^{( 42 )} No demethylated products of DMA are observed in urine or tissues of mice administered DMA.^{( 42 )} For rodents the putative pathway of arsenic metabolism has been proposed (Fig. 5). When B6C3F1 mice were administered 1.11 or 111 mg/kg of DMA^V, kidney had the highest concentration of DMA^V of all tissues up to 60 min postadministration. It has been also reported that gavage and i.v. administration of DMA^V in mice results in dose‐dependent prolonged retention in the lung as compared with liver and kidney and the dose‐dependent distribution of DMA^V in the lung was suggested to have a role in the toxic effects DMA^V elicits in this organ.^{( 41} , ^{42 )} DMA could be binding to a critical macromolecule and this binding is related to its toxic effects. The present observation that the administration of DMA^V to Ogg1 ^−/– mice caused significant elevation of 8‐OH‐dG in lung DNA is in line with data obtained by other authors with regard to the promotion of lung carcinogenesis.^{( 23 )} Furthermore, recent results implicated 8‐OH‐dG repair defects in certain lung cancers.^{( 43 )}

Figure 5.

The putative pathway of arsenic metabolism in rodents. As^III, arsenite; As^V, arsenate; DMA^III, monomethylarsinous acid; DMA^V, dimethylarsinic acid; hGSTO1‐1, human glutathione‐S‐transferase omega 1‐1; MMA^III, monomethylarsonous acid; MMA^V, monomethylarsonic acid; SAHC, S‐adenosyl‐L‐homocysteine; SAM, S‐adenosyl‐L‐methionine; TMAO, trimethylarsine oxide.

The results of the oligonucleotide microarray analysis on uploading to Ingenuity Pathway Analysis indicated that DMA^V treatment induced the expression of genes involved with cancer, the cell cycle, cellular proliferation, DNA replication, recombination and repair, lipid metabolism, cell–cell signaling and interaction, molecular transport and immunological disease pathways in Ogg1 ^−/– mice. Those alterations, including up‐regulation of Pola1, Cyp7b1, Ndfua3, Mmp13, and Hspb1 are most likely to contribute to the lung carcinogenicity of DMA^V. To our knowledge these are the first data for alterations in such gene expression in mouse lung by DMA^V. Recently, in utero treatment with arsenic of pregnant mice was shown to result in CYP7B1 overexpression in their adult offspring.^{( 44 )}

In line with our results, it has recently been shown that short‐term transplacental inorganic arsenic exposure in C3H mice increases lung adenocarcinoma incidence from 0% in controls to over 20%.^{( 45 )} Furthermore, gestational inorganic arsenic exposure in C3H mice have been shown to initiate pulmonary carcinogenesis that was promoted by the phorbol ester, 12‐O‐tetradecanoyl phorbol‐13‐acetate (TPA) in male or female offspring.^{( 44 )} Arsenic was accumulated dose‐dependently in the lung tissues of iAs^V‐exposed mice. Recent results demonstrated an increase in lung tumor number and lung tumor size and the rate of poorly differentiated lung adenocarcinoma in iAs^V‐exposed mice compared to the control. Methylation rates of p16INK4a and RASSF1A appeared to be higher in a dose‐related tendency in lung tumors from iAs^V‐exposed mice compared to the control.^{( 46 )} Altogether these results are in accord with human studies that show the lung is a target of arsenic carcinogenesis.^{( 47} , ^{48 )} Characterization of new mouse models, will stimulate further research aimed at a better understanding of proliferative lesions of the lung.^{( 49 )}

In non‐treated control Ogg1 ^−/– mice lungs, analysis of all the microarrays averaged data revealed lower expression levels than normal for genes whose products contribute to drug, lipid metabolism and endocrine system development and function. In a large number of heat shock protein cases, down‐regulation was evident. This might be a possible explanation for why the spontaneous tumors were less evident in control Ogg1 ^−/– comparing to Ogg1 ^+/+ animals. On the other hand, genes related to the cancer, cellular growth and proliferation and cell cycle were shown to be up‐regulated in the lung, which might explain the increased levels of cellular proliferation, as assessed in terms of PCNA positive indices, in Ogg1 ^−/–. When Ogg1 ^−/– mice were administered DMA^V, Ingenuity pathway analysis revealed overexpression of genes involved with cancer, cell cycle, cellular proliferation, DNA replication damage and repair, lipid metabolism, cell–cell signaling and interaction, tumor morphology, molecular transport protein synthesis but suppression of immune response pathways. Alterations to cell proliferation included up‐regulation of CYP7b1, NADH dehydrogenase (ubiquinone 1 alpha), HSP1A, 1B, 4 and 70 kDa protein 12 A, polymerases (DNA‐directed) alpha 1, gamma, iota, epsilon 2, MAPKKK 5, 6 and 14. Suppression of the immune response pathway detected in Ogg1 ^−/– mice by DMA^V might explain the increased malignant lymphoma development in female Ogg1 ^−/– mice. When DMA^V‐treated wild type and Ogg1 mutant mice were compared, it was found that genes involved in cancer, cellular growth and proliferation, lipid metabolism and cell–cell signaling and interaction were overexpressed in Ogg1 ^−/– animals. The fact that the free radical scavenging pathway was not activated in Ogg1 ^−/– mice might be the reason for massive 8‐OH‐dG formation in the lung.

Interestingly, immunohistochemical examination of normal‐appearing tissue of Ogg1 ^−/– mice lungs induced by DMA^V treatment, tumors showed high levels of concordance between induction of 8‐OH‐dG and apoptosis (ssDNA). The observation that oxidative DNA damage in Ogg1 ^−/– mouse lung is mostly observed in apoptotic cells could mean that strong oxidative degradation of DNA is the direct cause.

Electron microscopic and immunohistochemical examinations revealed significant elevation of cellular proliferation in the lungs of Ogg1 ^−/– mice as compared to control Ogg1 ^−/– and Ogg1 ^+/+ animals. Both in the non‐neoplastic lungs and lung tumors of type II origin of Ogg1 ^−/– mice treated with DMA^V microvilli, increased and enlarged mitochondria were observed. These results indicated that changes of ultrastructure pointing to the induction of cell proliferation in pneumocyte type II cells induced by DMA^V are related to the development of lung tumors.

In conclusion, DMA^V, an organic methylated arsenical, here exerted carcinogenicity in the lungs of homozygous Mmh/Ogg1‐deficient mice defective in repair of oxidative DNA base modifications. Our data indicate that significant persistent elevation of 8‐OH‐dG, with induction of Pola1, Cyp7b1, Ndfua3, Mmp13, and Hspb1 genes expression induced by DMA^V with high background levels of cell proliferation might be responsible for DMA^V carcinogenicity in the Ogg1 ^−/– mutant mouse lung.