Arsenic exposure is associated with DNA hypermethylation of the tumor suppressor gene p16

We conducted a paired case–control study approved by the Chinese PLA General Hospital review board. 40 case subjects were recruited from villages in Bameng, Inner Mongola of China where the drinking water is contaminated with high levels of arsenic (>0.05 mg/L, average = 0.6 mg/L). The selection criteria include a clear drinking history and a positive diagnosis of arsenic poisoning (arsenicosis). Two cohorts of control subjects, matched with respect to age, sex, ethnicity and socioeconomical status, were recruited including: 1) individuals from the same villages with high arsenic exposure but showed no clinical signs of arsenicosis; and 2) individuals from the neighboring villages where the drinking water is not contaminated with arsenic (<0.05 mg/L, average = 0.02 mg/L). Table 1 summarizes the descriptive characteristics of the study participants. None of the parameters is statistically significantly different between case and control groups.

Interviews were conducted between trained practitioners and study subjects with a structured epidemiological questionnaire including details about individual’s demographic factors, drinking water history and condition, smoking index, alcohol consumption, past medical history and current illness. 3 ml of vein blood were drawn from study subjects. Whole blood leukocytes were isolated through centrifugation and stored at −20°C. DNA was extracted for methylation-specific (MS) PCR reactions to determine the DNA methylation levels of p16 gene as described previously [16]. In brief, each sample was treated with bisulphite to convert unmethylated but not methylated cytosine to uracil. The samples were then amplified with two sets of primers targeting p16 promoter regions, one for unmethylated DNA (Forward 5′-TTATTAGAGGGTGGGGTGGATTGT-3′; Reverse 5′-CAACCCCAAACCACAACCATAA-3′) and one for methylated DNA (Forward 5′- TTATTAGAGGGTGGGGCGGATCGC-3′; Reverse 5′- GACCCCGAACCGCGACCGTAA-3′). The PCR products were analyzed by polyacrylamide gel electrophoresis and ethidium bromide staining. The appearance of a band of ~150 bp indicates the presence of p16 methylation in the blood. A representative result of the MS-PCR assay with positive and negative controls is shown in Additional file 1: Figure S1.

As shown in Table 2, 26 of 40 (65%) subjects in high arsenic exposure with arsenicosis group had p16 hypermethylation, while 19 of 40 (47.5%) subjects in high arsenic exposure without arsenicosis group were p16 methylation positive. In sharp contrast, only 9 of 40 (20%) subjects in low arsenic exposure group were p16 methylation positive (P < 0.01, Fisher’s Exact test). These results suggest that there is a significant correlation between p16 hypermethylation and arsenic exposure independent of the development of arsenicosis.

To extend our analysis, variables collected through interviews and MS-PCR results were assigned with quantified value as shown in Additional file 2: Table S1. Crude odds ratios (OR) and estimates of relative risk were calculated by univariate analysis. To identify variables that were independently associated with arsenic exposure, multivariate analyses were performed using conditional logistic regression methods. All P-values resulted from two-sided statistical tests. The FREQ and PHREG procedures in Statistical Package SAS 6.12 were employed.

We first compared case subjects (high exposure to arsenic with arsenicosis) to control subjects with minimal exposure to arsenic. 12 variables were included in the univariate logistic regression analysis and Table 3 summarizes the results of five of them. DNA hypermethylation of p16 showed a highly significant association with case subjects (OR = 10.0,P = 0.0019), while none of the other variables reached statistical significance.

We next compared case subjects to control subjects who are exposed to high arsenic levels without arsenicosis. 13 variables were included in the univariate logistic regression analysis and Table 4 summarizes the results of six of them. The only variables displaying significant association with case subjects were years of drinking contaminated water (OR = 1.192, P = 0.0154) and arsenic concentration in drinking water (OR = 4.2,P = 0.0039). None of the other variables, including p16 DNA methylation (OR = 2, P > 0.05), reached statistical significance.

Despite significant efforts to eliminate arsenic from industrial processing and agricultural use, occupational and environmental exposures to toxic levels of arsenic are still common in developing countries. Therefore understanding the link between arsenic intake and adverse health effects such as cancer will exert significant public health benefit. The carcinogenic mechanism(s) of arsenic have been extensively studied in laboratories using animal and cell culture models. Intriguingly, in contrast to most chemical carcinogens, arsenic only weakly induces genetic mutation [7]. On the other hand, multiple groups have reported that arsenic can induce significant changes in DNA cytosine methylation. For example, Zhao et al. reported that chronic arsenic treatment in rat liver epithelial cells induced malignant transformation which was accompanied by global DNA hypomethylation [10]. Mass and Wang showed that exposing human lung A549 cells to sodium arsenite produced significant dose-responsive hypermethylation at the promoter of p53 tumor suppressor gene [9]. Furthermore, recently Cui et al. demonstrated that giving mice arsenic in the drinking water induced formation of lung adenocarcinoma with DNA hypermethylation of p16[11]. In the current study, we present data that human exposed to high levels of arsenic are significantly more likely to have p16 DNA hypermethylation. Our findings as well as others’ support the hypothesis that arsenic may act as a potent “epi-mutagen” to epigenetically modify key tumor suppressor genes. As DNA methylation at promoters can silence gene expression in a heritable manner, this could be a critical mechanism by which cells exposed to arsenic overcome barriers to malignant transformation.

Previously Chanda et al. examined human subjects exposed to arsenic-contaminated drinking water in West Bengal in India and found that DNA hypermethylation of p53 and p16 positively correlated with arsenic levels in drinking water [17]. Studies of individuals using arsenic-rich coal with indoor unventilated stoves in Guangzhou, China also found trends of increased p16 methylation and reduced protein expression [18]. Our results are consistent with these findings. In addition, we for the first time showed that p16 DNA methylation levels are not significantly different between subjects with high exposure to arsenic and arsenicosis and subjects with similar arsenic exposure without arsenicosis. This result excludes the possibility that p16 DNA hypermethylation is a secondary pathological consequence of arsenic poisoning. Furthermore, the result suggests that p16 hypermethylation precedes the development of clinical symptoms and may be an important risk factor for arsenic-related diseases such as cancer. This is consistent with clinical observations that loss of p16 through genetic or epigenetic inactivation is an early event in tumor progression [19]. A weakness of our study was the use of whole blood leukocytes for p16 methylation analysis. Although this approach has been successful to study disease-related epigenetic changes [20], samples from more relevant tissue types such as skin would potentially provide more information.

There are several limitations of current study. First, although the high sensitivity of MS-PCR assay allows us to measure p16 methylation with a limited amount of human blood sample, this assay is qualitative rather than quantitative. Future study with bisulfite sequencing will be required to validate our findings and to provide a quantitative assessment of the correlation between arsenic exposure and p16 methylation. Second, because our sampling sites are located in remote areas with no laboratory access which makes the preservation of RNA from blood very challenging, our study did not provide a measurement of p16 RNA and protein expression. Although the association between p16 promoter methylation and gene repression has been firmly established [21],[22], it would be important in the future to demonstrate that the epigenetic silencing of p16 affects its expression. In conclusion, our study demonstrates that arsenic exposure is associated with p16 DNA hypermethylation in human, which may be an important risk factor for arsenic-induced tumor development [23],[24]. Further investigations are needed to reveal how arsenic metabolism specifically alters cellular epigenetic state. As clinically approved drugs targeting DNMTs are available, our results suggest that pharmacological reversal of DNA hypermethylation at tumor suppressor genes may offer therapeutic benefits to arsenic exposed individuals.

Written informed consent was obtained from the patient for the publication of this report and any accompanying images.