Micronuclei as Biomarkers of Carcinogen Exposure in Populations Exposed to Arsenic Through Drinking Water in West Bengal, India: A Comparative Study in Three Cell Types

Arsenic is a recognized human multisite carcinogen (1), presently affecting more than 19 countries as an environmental contaminant. At present, almost 6 million people are endemically exposed to inorganic arsenic in West Bengal by drinking heavily contaminated groundwater through hand-pumped tube wells (2). The concentration of arsenic in these water samples ranges from 60 to 560.23 μg/l (3), thus greatly exceeding the current maximum contamination level laid down by both WHO (4) and U.S. Environmental Protection Agency (5). All the hallmark signs of chronic hydroarsenicism such as hyperpigmentation, raindrop pigmentation, keratosis of skin, anemia, burning sensation of the eyes, solid edema of legs, liver fibrosis, chronic lung disease, gangrene of the toes, and neuropathy have been reported for inhabitants of the nine arsenic-affected districts of West Bengal (6). Because arsenic-caused skin lesions typically manifest after a latency period ranging from 10 months to 10 years of exposure (7) it is likely that over a period of decades, a larger number of people who are subclinically affected will also progress to various diseases including cancer. Thus, monitoring of this population is of considerable importance.

Arsenic contents in urine, hair, and nails have been considered to be meaningful measures of the absorbed dose of inorganic arsenic and thus serve as reliable biomarkers of arsenic exposure. Arsenic accumulation in nails and hair is attributable to the keratin-rich composition of these tissues, which reflects a cumulative exposure to arsenic over a long period (8). Estimation tests on hair and fingernails can measure exposure to high levels of arsenic or arsenic exposure over the past 6–12 months, whereas urinary arsenic level is regarded as the best biological indicator for assessing current arsenic exposure because more than 60% of the ingested arsenic are excreted through urine (9, 10). Genetic toxicology end points have also been used as biomarkers as these are considered to be markers of early biological effects of carcinogen exposure (11). Exfoliated epithelial cells have traditionally been used for cancer screening and biomonitoring of genotoxic effects in humans (12). The frequencies of micronuclei observed in the exfoliated cells of oral mucosa and urinary bladder serve as an appropriate index to monitor the genotoxicity induced by arsenic because these cells are in direct contact with the carcinogen (13). Urothelial cell micronuclei reflect damage to the bladder epithelial tissue, which occurs ∼1–3 weeks prior to the exfoliated cells appearing in urine (14). The cytokinesis block micronuclei technique in lymphocyte culture is widely regarded as a sensitive and reliable method for assessing chromosome damage (15).

Arsenic-induced genotoxic effects are implicated in carcinogenic outcomes (16). The putative genotoxic effects of arsenic both in vivo and in vitro have been investigated. Higher incidences of micronuclei, chromosomal aberrations, sister chromatid exchanges, and aneuploidy have been reported from the human populations exposed to arsenic through drinking water in various countries such as Mexico (17, 18), Finland (19), Argentina (20, 21), and Taiwan (11). The exact mechanism of arsenic-induced carcinogenicity still remains elusive; however, short-term assays indicate that arsenic does not induce point mutations but rather acts as a clastogen, inducing the formation of chromosomal aberrations and micronuclei in animal and human systems (22). Thus, arsenic is an ideal genotoxicant to be evaluated using the micronuclei assay. To our knowledge, there was no genetic monitoring study of the residents of arsenic endemic districts of West Bengal until 2002, when we reported an elevated frequency of micronuclei in lymphocytes, oral mucosa cells, and urothelial cells in a pilot study of individuals from different arsenic-affected districts of West Bengal (23). This report was followed by our second study on chromosomal aberrations and sister chromatid exchange from the exposed individuals of North 24 Parganas district (3). In our first study, the 45 selected subjects were inhabitants of different districts with a wide range of arsenic exposure. Some of them had switched over to safer sources of drinking water also. Hence, correlation between the arsenic content in drinking water sample and the corresponding micronuclei frequencies was lacking in some cases. To bring homogeneity and to establish association between arsenic exposure and micronuclei incidence, we extended the study using samples from individuals inhabiting the same area and still exploiting the contaminated water source.

In this study, we have conducted a cross-sectional biomarker survey of a large population from North 24 Parganas district, West Bengal using micronuclei assay. The aim of this study was to compare the micronuclei frequencies in three cell types of 163 residents of the above-mentioned district, chronically exposed to high levels of arsenic in drinking water and a matched unexposed group (154 subjects) with little exposure to inorganic arsenic.

RPMI 1640, newborn calf serum, phytohemagglutinin (M form), l-glutamine, and penicillin-streptomycin were purchased from Invitrogen Life Technologies, Inc. Cytochalasin B, trizma hydrochloride, and EDTA were purchased from Sigma Chemical Co. (St. Louis, MO).

The study sites were four administrative blocks (designated as 1–4 in Fig. 1; i.e., 1 Gaighata, 2 Habra, 3 Deganga, and 4 Baduria) located within 20-km radius in the arsenic-affected North 24 Parganas district of West Bengal. This district was chosen for its proximity to Calcutta, the consequent convenience of transporting samples from the study site to the laboratory and also because it is reported to be severely affected by arsenic (24). The field survey was designed to reduce subjectivity. It was known if the site of sample collection was in the affected or unaffected district because the nine arsenic-affected districts of West Bengal are well documented (25). However, we did not know the arsenic concentration of the particular tube well. Arsenic concentrations in water samples vary widely as contaminated tube wells are scattered irregularly through out the study site. Each subject was first asked to complete a questionnaire that elicited information on demographics, life-style, occupation, diet, and addiction, medical, and residential histories. Then, physicians and dermatologists examined the study participants. Water and other biosamples were collected from the subjects on the same day and carried code numbers. Information from questionnaire-sourced data on the subjects was not revealed before the arsenic analyses were completed. Physical examination of the subjects and micronuclei assay were performed blind as to the arsenic concentrations in water and other biosamples. A large number of individuals were examined; the selected subjects were a consecutive convenient group of individuals who provided informed consent to participate and fulfilled the inclusion criteria.

One hundred sixty-three (77 females and 86 males) subjects from the four above-mentioned blocks, who were exposed to high levels of arsenic in drinking water and manifested cutaneous signs of arsenicism like hyperpigmentation, hypopigmentation, raindrop pigmentation, palmoplanter hyperkeratosis, or ulcerative lesions, were recruited as the exposed group. One hundred fifty-four (66 females and 88 males) subjects from Contai subdivision of East Midnapur district (designated as 5 in Fig. 1) with very little or no arsenic exposure were selected to form the unexposed group. The unexposed were matched to the exposed cases by age, sex, and socioeconomic status. This study was conducted in accord with the Helsinki II Declaration and approved by the institutional ethics committee.

In the present study, due to the large sample size, some participants were found to be mildly addicted in both exposed and unexposed groups. Based on addiction habits, both the exposed and the unexposed groups were further subdivided into two categories: addicted and nonaddicted. The nonaddicted subjects were never addicted, while the addicted included the current users. In the addicted category, we included subjects who were mildly addicted to betel quid and/or bidi. Betel quid chewing is a widespread habit in India and exists in several forms (26). Occupationally, the majority of the study participants were farmers, porters, or daily wage earners. Because arsenic-containing pesticides were not used and arsenic mining was not detected in this region, occupational exposure to arsenic was ruled out. Seafood, which may be a source of arsenic contamination, was not available for consumption. Therefore, arsenic in drinking water was the principal source of exposure in this region.

Collection of Water and Biosamples for Arsenic Estimation The samples collected for arsenic estimation include drinking water (∼100 ml), urine (100 ml), nails (∼250–500 mg), and hair (∼300–500 mg). The samples were analyzed mostly at the Institute of Wetland Management and Ecological Design, Salt Lake, Calcutta, and some at the School of Environmental Studies, Jadavpur University, Calcutta. Study participants were provided with acid-washed [nitric acid-water (1 + 1)] plastic bottles for collection of drinking water samples into which nitric acid (1.0 ml/l) was added later on as preservative (27). Ceramic blade cutters were used to collect nail and hair samples. Both samples were thoroughly cleaned following the method of Curatola et al. (28) and Agahian et al. (8) for removal of exogenous arsenic. Hair samples were of similar size and were taken from more or less similar region of head (close to the scalp behind the ear with a diameter of about 1 cm; Ref. 19). First morning voids were collected in precoded polypropylene bottles for arsenic estimation as these give the best measure of the recent arsenic exposure (9). Immediately after collection, the samples were stored in salt-ice mixture and brought to the laboratory where they were kept at −20°C until estimation was carried out. Concentrated HCl (1 ml/100 ml urine) was added in the urine samples to prevent bacterial growth (29).

Oral Mucosa Cell Collection Oral mucosa cell samples were collected from each subject using a soft toothbrush to scrape cells gently from the oral mucosa (inside of both cheeks). The brush was then swirled into a centrifuge tube containing a buffer solution [0.1 m EDTA, 0.01 m Tris-HCl, 0.02 m NaCl (pH 7.0); Ref. 30], thereby creating a cell suspension. The cell suspensions were stored at 2–4°C in a cooling device and brought to the laboratory within 2 h of sample collection.

Urothelial Cell Collection To collect urothelial-exfoliated cells, each subject was asked to provide ∼50 ml of the urine samples from the second and third voids of the day. As females generally exfoliate more cells per void than males (30), a total of four urine samples were collected from males and two from females. The urine samples were coded, kept at 2–4°C in a cooling device, and brought to the laboratory within 2 h of sample collection. First morning voids were not used for micronuclei assay because exfoliated cells tend to degrade from overnight exposure to urine (31).

Blood Sampling Blood samples (5–7 ml) were obtained from each individual by venipuncture. Whole blood (0.5 ml) was added to 6 ml RPMI 1640 containing 20% newborn calf serum, 2% phytohemagglutinin, 100 IU/ml penicillin, 100 μg/ml streptomycin, and 2 mm l-glutamine directly at the time of sample collection. The media containing blood samples were coded, kept at 4°C in a cooling device, and brought to the laboratory within 2 h of collection.

Micronuclei Assay in Exfoliated Epithelial Cells Oral mucosa cells were obtained by simply centrifuging the cell suspension at 1500 rpm for 10 min. The supernatant was discarded and cell pellets were resuspended in fresh buffer solution. Cells were washed thrice with the buffer solution. Gentle pipetting of cells in the buffer solution reduced clumping and lysed broken cells. Volumes of 25 ml of the buffer solution in 50 ml conical tubes were used in every washing step.

Urothelial cells were recovered by centrifuging urine samples (2000 rpm for 15 min) and washing the cell pellet with 0.9% NaCl (30).

Cell suspension of both cell types (50 μl) was laid and spread well on clean, preheated (40°C) glass slides and allowed to air-dry for 5–10 min. Cell density was checked with a phase-contrast microscope. The cell solution was either concentrated by centrifugation or diluted in the buffer solution (for oral mucosa cells) or 0.9% NaCl (for urothelial cells) as required. Once the desired cell density (no overlapping cells) was reached, more slides were prepared. The slides were fixed in methanol (80% v/v) at 0°C for 20 min and air-dried (32). Micronuclei in oral mucosa cells were scored in accordance with the criteria reported by Tolbert et al. (33), while urothelial cells were analyzed following the method of Reali et al. (34). At least 3000 oral mucosa cells and 1000 urothelial cells were scored per individual.

Micronuclei Assay in Lymphocytes Lymphocyte cultures were carried out for micronuclei analysis following the standard protocol of Fenech (15) and Migliore et al. (35). Replicate cultures were established at the time of sample collection. The whole blood cultures were brought to the laboratory and incubated for 44 h at 37°C. Cytochalasin B (a cytokinesis blocker) was added to each culture to give a final concentration of 6 μg/ml and the culture was incubated at 37°C for an additional 28 h to induce binucleated cell formation. After a total of 72 h incubation, the cells were centrifuged at ∼1000 rpm for 5 min. Supernatant was discarded and cell pellets were treated with a weak hypotonic solution (0.075 m KCl/saline, 1:9) for 5 min. After centrifugation, the cells were fixed in fresh fixative (methanol/glacial acetic acid, 3:1). Fixative was removed by centrifugation and two more changes of fixative were performed. The cells were dropped onto wet clean slides and the slides were air-dried and stained with 5% Giemsa in phosphate buffer (pH 6.8). At least 2000 binucleated cells from each subject were examined for micronuclei under the microscope.

Scoring Procedure All slides were first examined with low-power (20×) magnification using an Olympus BX40 microscope to discard those infected with bacteria, fungi, and polymorphonuclear leukocytes as these may interfere with scoring. Slides were then scored at 100× (oil immersion lens). Smeared, clumped overlapped or necrotic cells or those without intact nuclei were not recorded. Only those micronuclei were noted which were (a) rounded or oval shaped; (b) less than one-third the diameter of the main nucleus; (c) in the same focal plane as the nucleus; (d) of the same color, texture, and refraction as the main nucleus; and (e) clearly separated from the main nucleus. Two trained research fellows cross-checked all micronuclei scores to obviate the risk of bias. The values so obtained were averaged (36). Variability of repeated scoring of the slides by the same scorer was extremely low. The scorers were highly consistent on repeat counts and the concurrence between two scorers was good. All questionable micronuclei were additionally assessed by a third scorer and discussed until a consensus was reached.

Estimation of Arsenic in Nails, Hair, Water, and Urine Before estimation, the nail and hair samples were digested with 5 ml concentrated nitric acid and 3 ml concentrated sulfuric acid following the method of Agahian et al. (8). Water and urine samples were analyzed following the method of Guha Mazumder et al. (6) using an alkali-induced sample digestion procedure. Flow injection-hydride generation-atomic absorption spectrometry was used for estimation of arsenic in the collected biosamples. A Perkin-Elmer Model 3100 spectrometer equipped with a Hewlett-Packard Vectra computer with GEM software, Perkin-Elmer EDL System-2, arsenic lamp (lamp current 400 mA) was used for the purpose.

Statistical analysis focused on measuring the effect of arsenic exposure in the three cell types of exposed and unexposed individuals. Fisher's t test was performed to assess the difference in mean arsenic contents in drinking water, urine, nail, and hair between the exposed and the unexposed groups.

All micronuclei scores were converted to frequency of micronuclei/1000 cells. Due to the absence of normal distribution of micronuclei frequencies, the nonparametric Wilcoxon rank sum test (with correction for large samples) was used to assess the statistical significance of micronuclei frequency in three cell types between exposed and unexposed individuals. Statistical significance of micronuclei prevalence was also evaluated within the exposed group after stratification based on sex (between male and female) and addiction (between addicted and nonaddicted). As it was hypothesized a priori that exposure to arsenic would cause an increase in micronuclei frequency, so one-tailed test (P) was used.

Cochran-Armitage trend test was conducted to compare the micronuclei prevalence in three cell types with levels of arsenic in drinking water following the method of Armitage (37) and Cochran (38) as suggested by Hothorn (39). The test of significance was carried out at both 5% and 1% levels based on χ ²_1,0.05 = 3.848 and χ ²_1,0.01 = 6.035, respectively.

Descriptive characteristics of the exposed and unexposed subjects are summarized in Table 1. The mean (SE) arsenic contents of drinking water, urine, nail, and hair samples were significantly high (P < 0.01) in the exposed individuals when compared with the arsenic contents in the unexposed subjects. Majority (59%) of the subjects in the exposed addicted category were addicted to betel quid (fresh betel leaf, fresh areca nut, slaked lime, and catechu with or without tobacco), 13% were addicted to bidi (a dried rectangular piece of temburni leaf rolled into a conical shape containing finely cut tobacco), and 28% were addicted to both betel quid and bidi. The unexposed addicted category included 44% subjects addicted to betel quid, 31% with bidi addiction, and 25% subjects addicted to both betel quid and bidi. The average number of bidis smoked/day by the addicted subjects was <8, while the number of betel quids consumed was <5. In the exposed group, majority of the male study participants were occupationally farmers (44.19%); others were porters (18.6%), students (4.65%), schoolteachers (2.32%), shop owners (4.65%), rickshaw pullers (3.49%), and daily wage earners (22.09%). The majority of exposed female participants were housewives (84.42%); others were tailors (12.98%), students (2.6%), etc. The unexposed male participants included 48.86% farmers, 10.22% porters, 11.36% students, 4.54% schoolteachers, 7.95% shop owners, and 17.04% daily wage earners. The unexposed female participants included 74.24% housewives, 13.63% tailors, 7.57% students, and 4.54% schoolteachers. Due to the absence of arsenic-containing pesticides and arsenic mining in this region, it can be stated that none of the study subjects were occupationally exposed to arsenic.

The mean of micronuclei frequencies for the exposed and unexposed groups is presented in Table 2, where a comparison of these values for the three cell types for all subjects combined, for nonaddicted and addicted, males and females is made. Exposed individuals, on an average, exhibited a 5.63-fold increase in the proportion of micronuclei in lymphocytes compared with that observed in unexposed lymphocyte cultures, while the frequencies of micronuclei in oral and urothelial cells of exposed individuals was higher by 4.64 and 4.71 times, respectively, of that observed in unexposed. As far as nonaddicted exposed participants were considered, there was a 5.83-fold increase in the mean frequency of micronuclei in lymphocytes. The increase observed for oral mucosa and urothelial cells was 4.40-fold over unexposed and 4.71-fold over unexposed, respectively, in the nonaddicted group. Among the addicted, the increases in the micronuclei frequencies were less for all the three cell types (5.08-, 4.33-, and 4.60-fold over unexposed for lymphocytes, oral mucosa cells, and urothelial cells, respectively) than that of nonaddicted category. micronuclei incidence in urothelial cells was found to be slightly elevated (5.02-fold) among exposed males compared with the unexposed males. This difference was less (4.36-fold) between exposed and unexposed females. A comparative analysis of micronuclei by cell type, sex, and addiction within the exposed group shows that addicted subjects exhibit a significantly higher micronuclei incidence (P < 0.01) over the nonaddicted category in oral mucosa cells only. The results of the same analysis within the unexposed group show that addicted subjects have higher micronuclei rates in all three cell types (P < 0.01) when compared with the nonaddicted category. Exposed males manifest enhanced micronuclei frequencies over exposed females in oral mucosa cells (P < 0.01) and lymphocytes (P < 0.05). Unexposed males display increased micronuclei frequencies in oral mucosa cells only when compared with the unexposed females. No sex-based difference is observed in lymphocytes of the unexposed group.

Associations between distribution of micronuclei frequencies in lymphocytes and exfoliated epithelial cells of exposed individuals and arsenic content in drinking water are shown in Table 3. The exposed group has been categorized into three subgroups based on the levels of arsenic exposure. There is a high-level (>250 μg/l) exposure group, a middle-level (151–250 μg/l) group, and a low-level (51–150 μg/l) exposure group. In all the three cell types, positive trend effects of micronuclei prevalence in relation to the three levels of arsenic in drinking water were investigated and found to be statistically insignificant at 1% and 5% levels.

Our study describes the comparison of micronuclei frequencies in three cell types between arsenic-exposed and unexposed subjects. The results of this comprehensive cross-sectional biomarker study provide evidence that arsenic is responsible for the increased micronuclei prevalence in lymphocytes, oral mucosa cells, and urothelial cells of the residents of the North 24 Parganas district. The exposure assessment results confirm that the participants are very highly exposed to inorganic arsenic through ingestion of arsenic-contaminated drinking water (Table 1). We are unaware of any other reported assessment of micronuclei frequencies in the three cell types considered together in a population in relation to arsenic exposure. Our data are consistent with levels of micronuclei induction reported for lymphocytes and exfoliated epithelial cells in populations from other parts of the world with prolonged arsenic exposure (18, 30, 31, 40). All the three types of cells analyzed from arsenic-exposed individuals manifested significant increases over micronuclei frequencies of unexposed individuals (Table 2). Unlike the study by Warner et al. (30) in which they reported absence of significant micronuclei incidence in oral mucosa cells, our surveillance is supportive of significantly higher micronuclei frequencies in exposed oral mucosa cells. This observation is in agreement with several earlier reports (18, 40).

Our results indicate that genotoxic effects show variation among different tissues. The tissue response variability may be due to difference in levels of direct/metabolic arsenic exposure to cells and/or to different cellular kinetics that influence the relative efficiencies of the cells (41). A slightly greater increase in micronuclei prevalence was observed in lymphocytes compared with the exfoliated epithelial cells. However, due to the longer life span of lymphocytes compared with epithelial cells, the micronuclei observed in this tissue should not necessarily be correlated with the micronuclei observed in cells that have different turnover rates (13). Urothelial cells turnover in the bladder every 1–2 weeks, so micronuclei cells do not accumulate over time. At times, correlation may be lacking between urothelial and oral mucosa cells due to different target tissue and individual sensitivity (18). The performance of micronuclei as a biomarker in three cell types within the exposed group based on low (51–150 μg/l), middle (151–250 μg/l), and high (>250 μg/l) levels of arsenic content in drinking water shows absence of positive trend effect (Table 3). The three different levels of arsenic in drinking water did not affect the micronuclei proportions in the three subgroups within the exposed group. This could probably be due to differences in quantity of water intake and durations of arsenic exposure in the study participants as well as interindividual variation in susceptibility to arsenic.

Several variables such as habitual factors (exercising, drinking, and smoking), dietary factors (folate deficiency and plasma levels of vitamin B12 and homocysteine), and demographic factors (age and gender) are supposed to affect micronuclei frequency (42). However, age and gender did not influence our results because the arsenic-exposed and unexposed groups were age and sex matched. As both groups had similar occupations and socioeconomic status, their diet and level of physical exertion were also more or less similar. Thus, these factors do not impact the usefulness of micronuclei as a biomarker of arsenic exposure and the overall conclusion of the study. Because addiction is an important effect modifier and betel quid chewing with or without tobacco has been associated with oral cancer (26), data from addicted and nonaddicted individuals were analyzed separately although the participants were mildly addicted. All the addicted subjects kept away from alcohol consumption, cigarette smoking, and pan masala (a popular Indian addiction containing a dry mixture of areca nut, lime, catechu, condiments, and tobacco) chewing. Majority of the study subjects were addicted to betel quid. The high carotene content of betel leaf, low dry weight, and less tobacco content of betel quid make it less toxic and carcinogenic than cigarette smoking or other forms of addiction (43). The data analysis of nonaddicted category revealed greater differences for micronuclei induction between exposed and unexposed subjects. When exclusively unexposed subjects are considered, it is observed that addicted subjects have higher micronuclei frequencies in all three cell types (P < 0.01) than the nonaddicted category. However, within the exposed group, addicted individuals exhibited significantly increased micronuclei frequency (P < 0.01) over the nonaddicted ones in oral mucosa cells. This may be due to the synergistic effect of arsenic with the betel quid chewing habit.

There is a dearth of information on the mechanism by which arsenic exerts its carcinogenic effect. Usefulness of micronuclei assay as a screening and early detection technique for cancer susceptibility has been suggested before (42). Most in vivo human studies using centromere-specific probes report predominance of centromere negative micronuclei, thereby indicating that arsenic is more clastogenic than aneugenic (31). The present study adds to the increasing volume of literature suggesting that chronic environmental exposure to arsenic causes genotoxic effects, which may be implicated in the increased carcinogenesis incidence. To summarize, we believe that micronuclei assay is an effective technique adopted for rapid risk assessment of arsenic-induced cancer in large human populations exposed to arsenic. Our study has demonstrated the comparative analyses of micronuclei as biomarkers in lymphocytes, oral mucosa cells, and urothelial cells in arsenic-exposed individuals. Among the three cell types, lymphocytes display slightly higher micronuclei frequency. Further investigations on the mechanism of arsenic action and possible indicators of susceptibility and effect would have public health implications.

We thank Drs. J. N. Sarkar (Head, Department of Dermatology, School of Tropical Medicine, Calcutta, India), A. Mukherjee (Department of Neurology, Vivekananda Institute of Medical Sciences, Ramakrishna Mission Seva Pratishthan, Calcutta, India), and A. K. Sarkar (Peerless Hospital and B. K. Roy Research Centre, Calcutta, India) for much help in clinical examination of the study participants and the Council of Scientific and Industrial Research (Government of India) for sanctioning the Mission Mode Project on Toxicogenomics.