Document B5MqkwrKvnYr1wNBn66aZRe3E

Vol. 3, 583-590, October/November 1994 Cancer Epidemiology, Biomarkers & Prevention 583 Increased Micronuclei in Exfoliated Bladder Cells of Individuals Who Chronically Ingest Arsenic-contaminated Water in Nevada' Marcella L. Warner, Lee E. Moore, Martyn T. Smith,2 David A. Kalman, Elinor Fanning, and Allan H. Smith Division of Environmental Health Sciences, School of Public Health, University of California, Berkeley, California IM. L. W., L. E. M., M. T. S., E. F., A. H. 5.1; and Department of Environmental Health, School of Public Health and Community Medicine, University of Washington, Seattle, Washington ID. A. K.l Abstrad It is well established that inorganic arsenic is causally associated with lung cancer via inhalation and skin cancer via ingestion. Epidemiobogical evidence based on studies in Taiwan suggests that ingestion of inorganic arsenic may also cause other more fatal internal cancers, with the highest relative risks reported for bladder cancer. Here, we have used a biological marker of response, the micronucleus assay in exfoliated bladder cells, to evaluate the possible genotoxic effeds of chronic arsenic ingestion on the bladder. The overall objedive of this study was to compare the frequency of micronucleated cells in exfoliated bladder and buccal cells between a group of 1 8 individuals in Nevada who chronically ingested high levels of inorganic arsenic from their well water (average level, 1 ,31 2 pg/liter) and an individually matched control group with low exposure to arsenic (average level, 1 6 pg/liter). A 1 .8-fold increase (90% confidence interval, 1 .06 -2.99) was observed in the weighted mean frequency of micronucleated bladder cells in the exposed group (2.79 per 1 000 cells) compared with the unexposed group (1 .57 per 1 000 cells). In addition, the frequency of micronucleated bladder cells was positively associated with the urinary concentration of inorganic arsenic plus its methybated metabolites (Spearman correlation = 0.33; P = 0.03). In contrast, there was no increase in micronucleated buccal cells associated with arsenic ingestion (frequency ratio = 1 .0; 90% confidence interval, 0.65-1 .53). The results of this study provide evidence that chronic ingestion of high levels of inorganic arsenic in drinking water is associated with an increased frequency of micronucleated bladder cells. These findings are consistent with a genotoxic effed of arsenic on bladder cells, but a larger study is needed to confirm them. Received 2/1/94; revised 7/20/94; accepted 7/20/94. 1 This work was supported by Grants P30-ES01896 and P42-ES04705 from the National Institute of Environmental Health Sciences. Additional support was received from the Center for Occupational and Environmental Health. M. L. W., L. E. M., and E. F. are trainees of the Health Effects Component of the University of California Toxic Substances Program. 2 To whom requests for reprints should be addressed, at 747 University Hall, School of Public Health, University of California, Berkeley, CA 94720. Introduction Background. Arsenic is a naturally occurring element throughout the environment. Like some metals, arsenic can form both inorganic and organic compounds. The inorganic forms are considered to be the most toxic species (1). It is well established that inorganic arsenic is causally associated with lung cancer via inhalation and with skin can- cer via ingestion (2). A review ofthe epidemiobogical evidence suggests that ingestion of inorganic arsenic may also cause other more fatal internal cancers, including bladder cancer (3). The main source of epidemiobogical evidence comes from studies of a large population in southwest Taiwan, where residents were exposed to high levels of inorganic arsenic in drinking water from the 1920s until an alternate water supply was provided in the 1960s (4-8). Significant dose-response cancers were ranging from increased risks for mortality from a variety of reported for inorganic arsenic concentrations 1 70 to 800 pg/liter (6-8). The highest risks were reported for bladder cancer, with mortality rate ratios of 5.1 , 1 2.2, and 28.7 for males, and 1 1 .9, 25.1 , and 65.4 for females (6). Additional evidence for an association between arsenic ingestion and bladder cancer is provided by a study of persons who ingested medicinal arsenicals (9). The results of risk extrapolation from the Taiwanese findings suggest that the relative risk of bladder cancer arising from lifetime consumption of 1 biter of drinking water pg/liter) when ground per day at the current U.S. MCL' for arsenic (50 could be about 2.6 for females and 1 .4 for males, compared to ingestion of water containing back- levels of arsenic (<5 pg/liter) (10). Although estimates of human cancer risk exceed that estimated for other chemicals regulated by the Safe Drinking Water Act, the association between arsenic ingestion and internal cancers has not been investigated in the U.S. until recently. It has been estimated that more than 300,000 people have drinking water supplies with arsenic bevels in excess of the MCL (1 1 ). Many of these people reside in less populated parts of the western U.S. and are frequently served by private water supply wells. In this study, we used a biological marker of response, the micronucleus assay in exfoliated bladder cells, to eval- uate the possible carcinogenic effects of chronic arsenic ingestion on the bladder. Micronucbei are extranuclear bodies in the cytoplasm of a cell that form when acentnic chromosomal incorporated fragments on whole chromosomes are not into daughter nuclei during mitosis (1 2). The presence of micronuclei in a population of cells is an mdi- cation that chromosome damage has occurred as a result of 3 The abbreviations used are: MCL, maximum micronucleated cells; CI, confidence interval; InAs, inorganic arsenic; DMA, dimethylarsinic contaminant level; MMA, methylarsonic acid; FR, frequency. MNC, acid; 584 Arsenic.induced Bladder Genotoxicity exposure to a genotoxic agent that causes either clastogenic effects on spindle dysfunction (1 3). The mechanism of arsenic carcinogenicity is not well understood; however, the results of short-term tests indicate that arsenic does not induce point mutations but rather acts as a clastogen, inducing the formation of chromosomal aberrations and micronuclei in human cells (14-16). Thus, arsenic is an appropriate agent to be evaluated using the micronucleus assay. Additional support for examining mi- cnonuclei in bladder cells derives from the fact that cigarette smoking and schistomiasis, both associated with increased bladder cancer risks, have been found to increase the frequency of micronuclei in bladder cells (1 7). Exfoliated bladder cells are epithelial cells sboughed from the surface of the genitouninary tract. Epithelial cells turnover in the bladder every 1 to 2 weeks and micro- nucleated cells do not accumulate oven time (1 8). Because epithelial cells are derived from basal cells, damage in- duced by a genotoxic agent that acts as a clastogen at the basal cell layer should theoretically be reflected in the fre- quency of micronuclei in exfoliated cells (1 2). The presence of micronuclei in exfoliated bladder cells may serve as a way to measure the extent to which arsenic is associated with DNA damage to the bladder and thus potentially provide evidence for increased risk of bladder cancer development (19). Purpose of Study. To test the hypothesis that chronic in- gestion of high bevels of inorganic arsenic is associated with a detectable increase in the frequency of micnonuclei in ex- foliated bladder cells, a cross-sectional study was conducted in a small county in Nevada. According to historical water data maintained by the state of Nevada, Division of Water Planning, inorganic arsenic concentrations measured in pni- vate wells in that area have ranged from <1 0 to over 3000 pg/liter and have remained generally consistent within a given well. The high levels of arsenic in the ground water of this county are derived from the sobubilization of arsenic-beaning minerals in geological formations (20). Some ofthese drinking water arsenic levels exceed those measured in Taiwan and are perhaps some of the highest reported in the world. The overall objective of this study was to compare the frequency of micronucleated cells in exfoliated bladder and buccal cells between a group of individuals who chroni- cabby ingest high levels of inorganic arsenic and a matched control group with little exposure to arsenic. Bladder cells were selected because they represent cells from the target organ of concern, the bladder. Buccal cells were also cob- lected for comparability with other studies. An increase in micnonucbeated buccab cells was not expected, however, because there is no epidemiobogical evidence that arsenic is associated with oral cancer. Materials and Methods Seledion of Study Subjeds. The study population included residents of a county in Nevada with private water supply wells. For this study, exposed subjects were defined as individuals with well water arsenic levels 500 pg/biter As, more than 1 0 times the U.S. MCL. Unexposed subjects were defined as individuals with well water arsenic levels 10 pg/liter As. To ensure chronic exposure, the study was restricted to individuals who had resided in their home for at least 1 year and used no additional water source such as filtered on bottled water. A letter describing the study and the participation cmi- tenia was first sent to the 57 households in the county whose well water reportedly was 500 pg/liter As. Follow-up contact was made via telephone. Of the 1 3 eligible house- holds identified, 1 1 agreed to participate (85%) and information about the ages and sex of household members was obtained. The primary reason for ineligibility of households was consumption of filtered or bottled water at home. The final exposed group consisted of 1 8 subjects from 1 1 house- holds with well water arsenic levels over 500 pg/liter. To reduce the potential for confounding, each subject in the exposed group was individually matched on age (4 years), sex, and smoking status to a similarly identified subject in the unexposed group. A letter describing the study and the participation holds in the community : 1 0 pg/liter As. Follow-up criteria was sent to 70 housewhose well water reportedly was contact was made via telephone and at that time information about the ages and gender of household members was obtained. If a household member matched an exposed subject profile, that member was invited to participate in the study. All 1 8 eligible subjects agreed to participate (1 00%). Thus, thefinal study population comprised 1 8 matched pains of exposed and unexposed participants. Exposure Assessment. After giving written informed con- sent, each study subject was interviewed at home using a questionnaire regarding demographic information, daily fluid consumption, occupation, smoking status, and recent diet. To confirm the well water arsenic levels from the histor- ical data, a sample oftap water was collected from each home and analyzed for arsenic by Environmental Protection Agency Method 601 0 (21 ). As an additional measure of arsenic expo- sure, a spot urine sample was collected from each subject and analyzed for inorganic arsenic and its urinary metabolites, methybarsonic acid and dimethylarsinic acid. The method used for arsenic speciation was a modifi- cation of the method of Cnecebius (22) and has been previously reported (23). Inorganic and methylated arsenic species in urine were converted to their respective arsines by treatment with sodium bonohydnide under acidic conditions and were then collected by sparging and cryogenic trapping. Following the collection of arsine vapors, the trap was allowed to warm and the arsine species were sequentially volatilized and detected by atomic absorption spectroscopy using a microbunner combustion cell. Detection limits for each of the arsenic species were 1-5 pg/liter with coeffi- cients of variation of 0.1-0.1 5. Exfoliated Cell Colledion. After the interview, buccal cell samples were obtained by gently rubbing the inside of the subject's cheek with a premoistened wooden tongue de- pressor dipped in water. Cells were washed twice in a buffer solution of pH 7.0 that contained 0.1 M EDTA, 0.01 M Tnis HCI, and 0.02 M NaCI. Gentle pipetting of cells in the buffer solution reduced clumping and lysed broken cells. Cells were dropped on a coded slide, air-dried, and fixed in 80% methanol at 0#{176}C. To obtain bladder cell samples, each subject was asked to provide a urine sample from the second and third void of the day. The first void of the day was not collected because exfoliated bladder cell degradation occurs when the cells have been in contact with urine overnight. Because females generally provide more cells per void than males, a total of two urine samples were obtained from females and four from males. Participants were supplied with precoded polypropylene bottles and instructions for urine collection. Bladder cells were collected from the urine specimens within 2 hours via centnifugation and the cell pellet was Cancer Epidemiology, Biomarkers & Prevention 585 washed with 0.9% NaCI. After centnifugation, cells were dropped on a coded slide, air-dried, fixed in 80% methanol at 0#{176}aCn,d stored in a nitrogen atmosphere at -20#{176}Cuntil use for the micronucleus assay. Micronucleus Assay. A new version of the micronucleus assay that uses the fluorescent dye propidium iodide and fluorescent in situ hybridization with a biotin-labeled probe for all human centromeres was used (24). This procedure is easier and more reliable than the previous method of Feulgen-Fast-Green staining and also allows for the mech- anism of micronucleus formation to be determined. Briefly, the slides were preheated for 30 mm on a slide warmer at 63#{176}tCo fully attach the cells to the slide. They were then treated with 300 pg/mb pepsin for 30 mm at 37#{176}C to penmeabilize the cells (25). The slides were subsequently rinsed twice in phosphate-buffered saline and fixed in buff- ered 4% parafonmaldehyde for 20 mm at 0#{176}ACf.ter wash- ing, the slides were baked for 20 mm at 63#{176}CT.hey were then hybridized with a biotin labeled alpha-satellite probe for all human centromenes (Oncor) as described by Titenko- Holland eta!. (24). The fluorescent dye, propidium iodide at 1 pg/mb in antifade solution, was used to counterstain the DNA. Scoring Procedure and Criteria. All slides were first exam- med with low power magnification to observe the quality of the slide and the presence of polymorphonuclear leuko- cytes, bacteria, and fungi since heavy infections may inter- fere with scoring. Slides were then scored using a Nikon microscope equipped with epifluonescent illumination, a 1 OOx oil immersion lens, and a filter for fluonscein and propidium iodide (excitation at 450-490 nm, dicroic at 510 nm, and emission at 520 nm). Between 500 and 2,900 cells were scored for each sub- ject. Only cells that were not smeared, clumped, or oven- lapped and that contained intact nuclei were included in the analysis. Cells undergoing abnormal cell division and degen- erative processes such as karyorrhexis, karyolysis, nuclear fragmentation, or pyknosis (26) were recorded separately. The frequency of micronucleated cells was estimated based on the number of normal exfoliated cells scored. Micronuclei had to: (a) be less than 1/3 the diameter of the main nucleus; (b) be in the same plane of focus; (c) have the same color, texture, and refraction as the main nucleus; (d) have a smooth oval on round shape; and (e) be clearly separated from the main nucleus. All micnonuclei were photographed and cross-checked by two observers. Any questionable micronuclei were disregarded. Statistical Analyses. The data were analyzed using PC SAS software (27). The analysis focused on the effect of arsenic exposure on the frequency of cells with micronuclei. All counts were, therefore, converted to frequency of MNC per 1000 normal exfoliated cells. The mean frequency of MNC for the exposed and unexposed groups was computed with and without weighting the frequency of MNC for each subject by the number of normal cells scored. By including a weighting term, it was possible to account for the variation in the number of normal cells scored pen subject and thus reduce the overall variance of the mean frequency for each group. The general formula for computing the weighted mean frequency is: lii F = w, x f where i = 0 for unexposed and 1 for exposed; f., = mnc/n, = frequency of MNC for subject ij; w1 = n/N, = weight factor; mnc,#{14=9} micronucleated cells scored for subject ij; n = number o normal cells scored for subject ij; and N = total number of normal cells scored for group i. The frequency ratio, contrasting the mean frequency (weighted or unweighted) of MNC among the exposed and unexposed groups, was selected as the measure of effect. The frequency ratio was computed: FR = F,/FO where FR = frequency ratio; F, = mean frequency of MNC for exposed; and F0 = mean frequency of MNC for unexposed. The 90% Cl were calculated for the frequency ratios: 90% CI = exp [lnFR 1 .645[var(lnFR)]2] where var(lnFR) = [v,/ F,2 + v0/ F021; and v = [standard error (F,)b2 (28). Because the data were not normally distributed, statis- tical significance of the effect measure was assessed by the Wilcoxon sign-rank test. It was hypothesized a priori that arsenic exposure would be associated with an increase in the frequency of MNC, so one-tailed tests were used. In addition to dichotomous exposure status, the follow- ing arsenic exposure indices were investigated: (1 ) Exposure Index 1 (pg As/liter) = tap water arsenic (pg As/liter); (2) Exposure Index 2 (pg As/day) = tap water arsenic (pg As/biter) x liters As-fluid consumed pen day (biters/day); (3) Exposure Index 3 (pg As/liter fluid) = Index 2 (pg As/day)/totab biters fluid consumed per day (liters fluid/ day); (4) Sum of urine arsenic species (Sum) (pg/liter) = InAs + MMA + DMA; (5) InAs in urine (pg/liter); (6) MMA in urine (pg/liter); and (7) DMA in urine (pg/liter). Exposure index 1 reflects the household tap water an- senic concentration. Exposure indices 2 and 3 were esti- mated by incorporating fluid consumption data obtained during the interview to provide a more refined measure of individual exposure. Exposure index 2 reflects individual daily arsenic exposure. Exposure index 3 reflects individual daily arsenic exposure relative to total daily fluid consump- tion, thus enabling consideration of the possibility that the volume of other fluids consumed may dilute the effect of arsenic on the bladder. The dose-response relationship between each arsenic exposure measure and frequency of MNC was assessed by regression analysis. Analysis of covaniance was conducted to examine the relation between the other covaniates and frequency of MNC while adjusting for arsenic exposure. Results Descriptive data comparing the 1 8 exposed and 1 8 unex- posed subjects are summarized in Table 1 . The exposed and unexposed groups were each composed of eight males and 10 females. Four of the females in each group were current smokers. All men in the study were either non- smokers or ex-smokers. The average age of the exposed group was 37.5 years (range, 14-74 years), which was similar to that of the unexposed group (37.0 years; range, 586 Arsenic-induced Bladder Genotoxicity Table 1 Descriptive characteristics of study participants Exposed In = 18 8 Males, 10 Females) Mean SD Range Not exposed In = 1 8; 8 Males, 1 0 Females) Mean SD Range Age Iyrsl Education (yrs) Duration of residence Estimated total fluid intake (liters/day) Vegetables" (servings/week) Dark green vegetables" (servings/week) Carrots (servings/week) 37.5 13 4 3.8 9.2 1 .9 1 .3 14.2 2 3 1 .5 5.2 2.0 1 .8 (14-74) (8-17) (1-13) 12.4-6.9) (3-20) (0-7) 10-7) 37.0 13.8 (16-70) 13 2 (11-17) 54 (1-13) 3.6 1 .3 (1 .7-6.4) 13.0 6.7 (1-27) 2.1 1 .8 (0-6) 2.9 2.9 (0-10) Frequency (%) Frequency 1%) Alcohol Any None consumption 6 12 Coffee consumption 1+cup/day <1 cup/day 10 8 Smoking Current Ex Never 4 3 11 Average cigarettes/day (Current smokers) 27.5 Oral snuff Current Never use a Vegetables include broccoli, spinach, dark green I, Dark green vegetables include broccoli, spinach, 1 17 lettuces, corn, tomatoes, and dark green lettuces. (33.3) (66.7) (55.6) (44.4) (22.2) (16.7) (61.1) (5.6) (94.4) peas, green beans, cabbage, 9 9 8 10 4 4 10 27.5 1 17 and carrots. (50.01 (50.0) (44.4) 155.6) (22.2) (22.2) (55.6) (5.6) (94.4) Table 2 Summary of arsenic exposure measures Imean ISE)J' Exposure index 1 (pg As/liter) Exposure index 2 (pg As/day) Exposure index 3 (pg As/liter fluid) Sum As Urine arsenic species Ipg/literl InAs MMA DMA Exposed (n= 18) NotExposed (n= 18) 1,310 (350) 16 (7.2) 2,260 (570) 36 (15) 630 (160) 12 (5.6) 750 (150) 68 (23) 170 (47) 9 (2.5) 190 145) 14 (4.8) 390 (66) 44 (16) Mean Difference 1,296 P' (one tail)" <0.0001 a SE, standard error; Sum As = InAs + MMA sumed per day; Index 3 = tap water arsenic bWilcoxon Sign Rank Test. 2,224 <0.0001 + DMA; Index 1 = tap level x (liters As-containing 621 <0.0001 water arsenic level; fluid consumed 683 158 <0.0001 <0.0001 Index 2 = tap water arsenic level pen day/total liters fluid consumed 177 <0.0001 x liters As-containing per day). 348 <0.0001 fluid con- 1 6-70 years). The average age difference for the 1 8 pairs of subjects was 0.6 years. The descriptive data show that the matching criteria (age, sex, smoking status) were well met and that the exposed and unexposed groups do not differ signifi- cantly with respect to the other covaniates measured. None of the subjects were occupationally exposed to genotoxic agents. The mean, median, and range of each measure of arsenic exposure is presented for the exposed and unex- posed groups in Table 2. Although we attempted to include in the unexposed group only individuals with well water arsenic levels <1 0 pg/biter, one individual had a level of 100 pg/liter. As a result, the average exposure bevel to arsenic for the unexposed group was 1 6 pg/liter (mean) and 5 pg/liter (median). Nonetheless, by any measure of arsenic expo- sure, nanging from household tap water arsenic concentra- tion to a more refined estimate of arsenic consumption or a spot urine arsenic concentration, the exposed group is very highly exposed compared to the controls. A comparison of the weighted mean frequencies of MNC in bladder and buccal cells between the exposed group and the unexposed group is presented in Table 3. There was a 1 .8-fold increase in the weighted mean fne- quency of MNC in bladder cells of the exposed group compared with the unexposed group (90% Cl, 1 .06 -2 .99). The weighted mean frequency in the exposed group was 2.79/1 000 cells as compared with 1 .57/1 000 cells in the unexposed group. In contrast, the frequency ratio for buccal cells was 1 .0 (90% Cl, 0.65-1 .53). In Table 4, a comparison of the weighted mean fre- quencies of MNC in bladder and buccal cells between the exposed and unexposed groups after stratification on sex is presented. The weighted mean frequency of MNC in blad- Cancer Epidemiology, Biomarkers & Prevention 587 Table 3 Compar ison of MNC frequency in e xfoliated bladder and buccal cells by ars enic exposure status Exfoliated bladder cells Exfoliated buccal cells MNC frequency/ 1 000 cells Mean (SE) Frequency ratio (90% CI) P-Value" lone-tail) MNC frequency/ 1000 cells Mean (SE) Frequency ratio (90% CI) P-Value" (one-tail) Exposed 2.79 (0.73) (n= 18) 1 .78 11.06, 2.99) 0.09 2.49 (0.42) In= 161 1 .00 (0.65, 1.53) 0.5 Notexposed 1.57(0.28) (n=18) 2.50 10.50) (n=16( a Wilcoxon Sign Rank Test. Note: The apparent discrepancy between a P value of 0.09 and a confidence being nonparametric whereas confidence limits were derived with parametric statistics. interval excluding 1 is due to the test of significance Ta ble 4 Comparison o f MNC frequency in exfolia ted bladder and b uccal cells by arsenic exposure status and sex MNC frequency/ 1000 cells Mean (SE) Exfoliated bladder cells Frequency ratio 190% CI) P-Value" (one tail) MNC frequency/ 1000 cells Mean (SE) Exfoliated buccal cells Frequency ratio (90% CI) Males Exposed In = 8) Not Exposed In = 8) Females Exposed In = 10) Not Exposed (n= 10) 5.00 (1 .50) 2.14 (0.46) 1 .82 (0.53) 1.28 (0.31) 2.34 (1 .27, 4.29) 1 .42 (0.76, 2.65) 0.07 0.38 1 .36 (0.44) 1 .53 (0.48) 3.47 (0.48) 3.50 (0.72) 0.89 (0.42, 1 .86) 1 .00 (0.66, 1.49) a Wilcoxon Sign Rank Test. P-Value" (one-tail) 0.34 0.40 den cells was cleanly elevated among exposed males corn- pared to unexposed males (FR = 2.34; 90% Cl, 1.27-4.29); however, there was little difference between exposed fe- males and unexposed females (FR = 1.42; 90% CI, 0.76- 2.65). The frequency ratio for buccal cells remained close to one for both males (FR = 0.89; 90% Cl, 0.42-1 .86) and females (FR = 1 .00; 90% Cl, 0.66-1 .49). The unweighted results were not significantly different from the weighted results presented in Tables 3 and 4. There was a two-fold increase in the unweighted mean frequency of MNC in bladder cells of the exposed group (3.36/1 000 cells) compared with the unexposed group (1 .67/1000 cells). The frequency ratio for buccal cells was 1 .1 (90% Cl, 0.73-1 .70). After stratifying on sex, the unweighted and weighted results remained similar. The unweighted mean frequency of MNC in bladder cells was elevated among exposed males compared to unexposed males (FR = 2.49; 90% CI, 1 .32-4.70); but there was little difference between exposed and unexposed females (FR = 1 .37; 90% Cl, 0.67- 2.77). The frequency ratio for buccal cells remained close to one for both males (FR = 0.83; 90% Cl, 0.40-1 .74) and females (FR = 1.18; 90% Cl, 0.76-1.84). Fig. 1 presents a scatter plot ofthe relationship between the frequency of MNC in bladder cells and the sum of arsenic species in urine for males and females. The figure suggests that an increase in arsenic exposure is accompa- nied by an increased frequency of micronucleated bladder cells. In addition, the apparent increase appears to be stron- gem for males compared to females. Associations between each measure of arsenic expo- sure and the frequency of MNC in bladder cells are sum- manized in Table 5 with Speamman correlation coefficients and one-way Pvalues. Adjusting for age, sex, and smoking status did not change the findings; therefore, only the un- adjusted results are presented. Each arsenic exposure index is positively associated with the outcome, with index 3 reflecting the strongest correlation. Each urine arsenic spe- cies measure is also significantly positively associated with the outcome. In contrast, all arsenic exposure measures were neither significantly associated with the frequency of MNC in buccal cells nor were they consistent in terms of direction of association. After adjusting for arsenic exposure and sex, no signif- icant on suggestive associations were found between the weighted mean frequencies of MNC in bladder cells or buccal cells and the other covamiates measured. Theme was no evidence of any association between the frequency of cells undergoing abnormal cell division and degenerative processes and arsenic exposure. Discussion The results of the exposure assessment portion of this study confirm that residents of the U.S. can be very highly ex- posed to inorganic arsenic through ingestion of arsenic- contaminated drinking water. In this study, due to the very high exposure concentrations (average level, 1 31 2 pg/liter As), subjects consumed an average of 2260 pg As/day, much more than the 100 pg As/day estimated for individ- uals who consume 2 liters of water/day containing arsenic at the MCL (50 pg/liter). The major route for inorganic arsenic detoxification involves methylation of arsenic by methyltransfenase fob- 588 Arsenic-induced Bladder Genotoxicity 14 . Cl) C) a) V 12 . V 0 Female . 0 0 0 10 . S Male . a- Cl) ci 8. 0 V a) a) C) 6. : C 2. C) S 4. 0 > C) C a) p 0 50 .. 2. a) S 00 0 S 00 LI 00 0 0 0. a I II liii #{149}0 It I I I liii 0 l 10 100 1,000 I I I I iitI 10,000 Fig,'. 1. Relationship Sum of Arsenic Species in Urine (pg/I) between the frequency of micronucleated cells/1000 exfoliated bladder cells and sum of arsenic species in urine. Table 5 Summary of spearman corre lation coefficients Exposure status (yes/no) Exposure index 1 (pg As/liter) Frequency of MNC/ 1000 bladder cells P-value (one-wayl 0.209 (0.11) 0.258 (0.06) for arsenic exposur e indices and frequency Exposure index 2 (pg As/day) Exposure index 3 (pg As/liter fluid) 0.222 0.284 of MNC in bladder and buccal cells" Urine Sum As arsenic InAs species (pg/Iiter( DMA MMA 0.326 0.283 0.388 0.296 10.10) 10.05) 10.03) 10.05) (0.01) 10.04) Frequency of MNC/ 1000 buccal cells P-value (one-way) 0.068 (0.36) 0.106 10.28) 0.083 (0.33) 0.089 10.32) " Sum As = InAs + MMA + DMA; Index 1 = tap water arsenic level; Index 2 = tap water arsenic Index 3 = tap water arsenic level x (biters As-containing fluid consumed per day/liters fluid consumed -0.054 -0.056 (0.39) (0.38) level X liters As-containing per day). -0.025 -0.075 (0.45) fluid consumed 10.34) per day; bowed by elimination via urinary excretion. The sum of the urinary concentration of inorganic arsenic and its methyl- ated metabolites, MMA and DMA, is therefore considered to be a good biological measure of arsenic exposure. After exposure, inorganic arsenic and its metabolites can be mea- sumed in the urine in the following approximate proportions: InAs, 20%; MMA, 20%; and DMA, 60% (29). It has been proposed that a threshold exposure concentration exists at which arsenic methylation becomes saturated (30). Above the threshold concentration arsenic methylation activity would become saturated, and as a consequence, the pro- portion of methylated urinary metabobites (MMA and DMA) would decrease. In this study, the mean proportions of InAs, MMA, and DMA measured in the urine were 1 9.5, 22.0, and 58.4% for the exposed group, and 1 8.6, 21 .7, and 59.6% for the unexposed group, respectively. Urine arsenic metabolite profiles for both the exposed and unexposed groups do not differ significantly from each other or from that reported by others (31-33). This finding suggests that with chronic ex- Cancer Epidemiology, Biomarkers & Prevention 589 posune, there is no threshold exposure concentration and the ability to detoxify inorganic arsenic by methylation does not diminish even at very high exposure concentrations. The findings are consistent with data from many other population studies (29). Although other studies have been conducted to con- firm exposure to arsenic in the U.S. (34-39), only one study has utilized a biological marker of response to assess pos- sible genotoxic effects of arsenic. No differences in the frequencies of chromosome aberrations or sister chromatid exchanges were found in lymphocytes of individuals who were exposed to 1 00 pg/liter arsenic in drinking water compared to unexposed controls (40). However, the pop- ulation studied was exposed to much lower levels of arsenic than the current study population and arsenic has not been shown to be associated with cancer in blood-forming tissue. The results of this study provide evidence that chronic ingestion of high levels of inorganic arsenic in drinking water is associated with an increased frequency of MNC in exfoliated bladder cells. The weighted mean frequency of MNC in bladder cells of the exposed group is increased approximately 1 .8-fold compared to the unexposed group. The weighted mean frequency mate of MNC in bladder cells among the unexposed group is similar to the frequency mates reported in other studies of control populations (24, 41-43). The finding that each of the arsenic exposure indices was significantly positively associated with the frequency of MNC in bladder cells is suggestive of a dose-related effect (Table 5). It is noteworthy that after stratification on sex (Table 4), the findings are less consistent. The frequency ratio in blad- den cells remained elevated among exposed males corn- pared to unexposed males; however, theme was little differ- ence between exposed females and unexposed females. This apparent lack of consistency may be due to one of three possible explanations. First, perhaps male bladder cells are more susceptible to genotoxic damage caused by chronic ingestion of high bevels of inorganic arsenic than female bladder cells. This explanation is consistent with the approximately three-fold increase in the incidence of blad- den cancer observed in males relative to females that cannot be explained entirely by differences in exposure to known risk factors for bladder cancer (44). Second, it is possible that theme is no difference in the risk of genotoxic damage to the bladder cells of males and females from chronic ingestion of high bevels of inorganic arsenic, but due to the small sample size of this study, the power to detect an effect after stratification on sex is very small. Third, the apparent back of consistency may be due to the fact that males exfoliate almost exclusively transitional bladder cells in the urine, while females primarily exfoliate squamous cells from the tnigone of the bladder and only a small percentage of transitional bladder cells in the urine (45, 46). In addition, the mate of exfoliation of squamous cells varies with the female hormonal cycle. Since arsenic exposure has been shown to be associated with transitional cell carcinoma, the exfoliation of squamous cells by fe- males (which may not be affected by arsenic exposure) without the ability to differentiate them from transitional cells could be expected to dilute any real association be- tween chronic ingestion of inorganic arsenic and bladder cell micronuclei. The occurrence of such bladder cell mis- classification among females but not males could be ex- pected to bias the results among females towards no effect. A larger study must be conducted to confirm these findings and determine which explanation is correct. It is noteworthy that as hypothesized a priori, the me- suIts provide no evidence that the frequency of MNC in exfoliated buccal cells is associated with chronic arsenic ingestion. As observed for the bladder cells, the weighted mean frequency mate of MNC in buccal cells among the unexposed group falls within the mange of frequency mates reported in other studies of control populations (24, 41, 47-51). In contrast to the significantly positive correlation observed between each arsenic exposure index and the frequency of micronucleated bladder cells, none of the exposure indices was associated with the frequency of mi- cmonucleated buccal cells. In summary, this is the first study to show genotoxic effects of chronic ingestion of high bevels of inorganic am- senic in the U.S. The results of this study provide evidence that chronic ingestion of high levels of inorganic arsenic is associated with an increased frequency of MNC in exfoli- ated bladder cells. Although there is no documented direct association between micronuclei formation and develop- ment of cancer, an increased frequency of MNC in exfobi- ated bladder and buccal cells has been reported in studies of populations with exposures to known risk factors for bladder cancer (41 -43, 52) and oral cancer (1 9, 41 , 48, 50, 53-55). The significance of these findings with respect to blad- den cancer risk for people ingesting high concentrations of arsenic in drinking water remains to be determined. How- ever, it is becoming increasingly apparent that carcinogen- esis requires the progressive accumulation of numerous genetic alterations in a target tissue (56, 57). An observed increase in the amount of genetic damage occurring in a tissue should be associated with an elevation in risk for development of cancer at that site. Such a situation would increase the probability of the tissue acquiring the genetic changes necessary for carcinogenesis to occur (58). Acknowledgments The authors wish to thank Dr. Nina Titenko-Holland and Claudia Hopenhayn-Rich for much help and advice, Jill Dale for her assistance in sample collection, and the study participants for their cooperation. References 1 . US EPA. Health assessment document for inorganic arsenic (EPA-600/8- 83-0218). Washington, DC: U.S. Environmental Protection Agency, 1984. 2. IARC. ARC Monographs on the Evaluation of the Carcinogenic Chemicals to Man: Some Metals and Metallic Compounds, Vol. France: International Agency for Research on Cancer, 1980. Risk of 23. Lyon, 3. Bates, M. N., Smith, A. H., and Hopenhayn-Rich, and internal cancers: a review. Am. I. Epidemiol., C. Arsenic 135: 462-476, ingestion 1992. 4. Chen, C. J., Chuang, Y. C., Lin, T. M., and Wu, H. Y. Malignant among residents ofa blackfoot disease-endemic area in Taiwan: artesian well water and cancers. Cancer Res., 45: 5895-5899, neoplasms high arsenic 1985. 5. Chen, C. J., Chuang, Y. C., You, S. L., Lin, T. M., and Wu, H. Y. A retro- spective study on malignant neoplasms of bladder, lung, liver in blackfoot disease endemic area in Taiwan. Br. J. Cancer, 53: 399-405, 1986. 6. Chen, C. J., Kuo, T. L., and Wu, M. M. Arsenic 414-415, 1988. and cancers. Lancet, 7. Chen, C. J., and Wang, C. J. Ecological in well water and age-adjusted mortality Res., 50:5470-5474, 1990. correlation between arsenic level from malignant neoplasms. Cancer 8. Wu, relation cancers M. M., Kuo, T. L., Hwang, Y. H., and Chen, C. J. Dose-response between arsenic concentration in well water and mortality from and vascular diseases. Am. I. Epidemiol., 130: 1 1 23-1 1 32, 1989. 9. Cuzick, J., Sasieni, P., and Evans, S. Ingested arsenic, bladder cancer. Am. J. Epidemiol., 136:417-421, 1992. keratoses, and 590 Arsenic-induced Bladder Genotoxicity 10. Smith, A. H., Hopenhayn-Rich, C., Bates, M. N., Goeden, H. M., Hertz- Picciotto, I., Duggan, H. M., Wood, R., Kosnett, M. J., and Smith, M. T. Cancer risks from arsenic in drinking water. Environmental Health Perspect., 97:259-267, 1992. 1 1 . Science Applications International Corporation. currence and exposure to arsenic in public drinking draft). Prepared for the US Environmental Protection D.C. (EPA Contract 68-01-7166), 1987. Estimated national ocwater supplies Irevised Agency, Washington, 1 2. Vine, M. F. Micronuclei. (eds.), Biological Markers Press, 1990. In: B. S. Hulka, T. C. Wilcosky, and J. D. Griffith in Epidemiology. New York: Oxford University 13. Schmid, W. The micronucleus test. Mutation Res., 31:9-15, 1975. 14. Jacobson-Kram, D., and Montalbano, sessment group's report on the mutagenicity mental Mutagenesis, 7: 787-804, 1985. D. The reproductive effects as- of inorganic arsenic. Environ- 1 5. IARC. IARC Monographs Chemicals to Humans, Suppl. Research on Cancer, 1987. on the Evaluation 6. Lyon, France: of Carcinogenic Risk of International Agency for 1 6. Mass, M. J. Human carcinogenesis istry Health, 14:49-54, 1992. by arsenic. Environmental Geochem- 17. Smith, A. H., Hopenhayn-Rich, C., Warner, M. L., Biggs, M. L., Moore, L. E., and Smith, M. T. Rationale for selecting exfoliated nuclei as potential biomarkers for arsenic genotoxicity. bladder cell micro- J. Toxicol. Environ- mental Health, 40: 223-234, 1993. 18. Stich, H. F., San, R. H. C., and Rosin, M. P. Adaptation ofthe and micronucleus tests to human cell suspensions and exfoliated NY Acad. Sci., 407:93-105, 1983. DNA repair cells. Ann. 19. Stich, H. F., and Rosin, M. P. Micronuclei in exfoliated human cells as an internal dosimeter for exposures to carcinogens. In: H. F. Stich led.), Carcinogens and Mutagens in the Environment, Vol II. Naturally Occurring Compounds: Endogenous Formation and Modulation. Boca Raton, FL: CRC Press, 1983. 20. Glancy, P. A. Geohydrobogy of the basalt and unconsolidated tary aquifers in the Falbon area, Paper 2263. Churchill County, Geological Survey Water-Supply, 1986. sedimenNV: US 21 . US EPA. Test Methods for Evaluating tory Manual Physical/Chemical Methods Environmental Protection Agency, 1986. Solid Waste. Volume (SW-846). Washington, 1A: LaboraDC: U.S. 22. Crecebius, E. A. Modification of the arsenic speciation hydride generation. Anal. Chem., 50: 826-827, 1978. technique using 23. Kalman, D. A. Quantification of arsenic speciation in urine for exposure assessment studies. J. Res. NatI. Bureau Stand., 93: 315-318, 1988. 24. Titenko-Holland, N., Moore, L. E., and Smith, M. T. Measurement and characterization of micronuclei in exfoliated human cells by fluorescence in situ hybridization with a centromeric probe. Mutation Res., 312: 39-50, 1994. 25. Moore, L. E., Titenko-Holland, N., and Smith, M. T. Use of fluorescence in situ hybridization to detect chromosome-specific changes in exfoliated human bladder and oral mucosa cells. Environment. Mol. Mutagen., 22: 130-137, 1993. 26. Tolbert, P. E., Shy, C. M., and Allen, anomalies in buccal smears: methods 69-77, 1992. J. W. Micronuclei and other nuclear development. Mutation Res., 271: 27. SAS Institute Inc. SAS/STAT Institute Inc., 1988. User's Guide Release 6.03. Cary, NC: SAS 28. Cochran, Inc., 1962. W. G. Sampling Techniques. New York: John Wiley & Sons, 29. Hopenhayn-Rich, C., Smith, do not support the methylation inorganic arsenic. Environmental A. H., and Goeden, H. M. Human studies threshold hypothesis for the toxicity of Res., 60: 161-177, 1993. 30. US EPA. Risk Assessment Forum. Special Report on Ingested Inorganic Arsenic: Skin Cancer; Nutritional Essentiality IEPN625/3-87/01 3). Washing- ton DC: U.S. Environmental Protection Agency, 1988. 31 . Tam, G., Charbonneau, S., Bryce, F., Pomroy, C., and Sandi, E. Metab- olism of inorganic arsenic in humans following oral ingestion. Toxicology Applied Pharmacol., 50: 3 1 9 -322, 1979. 32. Crecelius, E. Changes in the chemical speciation of arsenic following ingestion by man. Environmental Health Perspect., 19: 147-150, 1977. 33. Buchet, J., Lauwerys, R., and Reels, H. Comparison of the urinary excretion of arsenic metabolites after a single oral dose of sodium arsenite, monomethylarsonate, or dimethylarsinate in man. Intern. Archives Occupa- tional Environmental Health, 48: 71 -79, 1 981. 34. Valentine, blood, urine, mental Res., I. L., Kang, H. K., and Spivey, and hair in response to exposure 20:24-32, 1979. G. Arsenic levels in human via drinking water. Environ- 35. Kreiss, K., Zack, M. W., Feldman, R. G., Niles, C. A., Sax, D. S., Chirico-Post, I., Landrigan, P. J., Boyd, M. H., and Cox, D. H. Neurologic evaluation of a population exposed to arsenic in Alaskan chives Environmental Health, 38: 116-121, 1983. well water. Ar- 36. Harrington, J. M., Middaugh, survey of a population exposed water in Fairbanks, Alaska. Am. J. P., Morse, D. L., and Houseworth, to high concentrations of arsenic J. Epidemiol., 108: 377-385, 1978. J. A in well 37. Whanger, P. D., Weswig, P. H., and Stoner, J. C. Arsenic levels in Oregon waters. Environmental Health Perspect., 19: 139-143, 1977. 38. Southwick, J. W., Western, A. E., Beck, M. M., Whitley, T., Isaacs, R., Petajan, drinking J., and Hansen, C. D. An epidemiological study of arsenic in water in Millard County, Utah. In: W. H. Lederer, and R. J. Fensterheim, beds.), Arsenic: Industrial, Biomedical, Environmental Perspec- tives, pp. 210-225. New York: Van Nostrand Reinhold, 1983. 39. Goldsmith, J. R., of health implications 1133-1136, 1972. Deane, M., of elevated Thom, arsenic J., and Gentry, in well water. G. Evaluation Water Res., 6: 40. Vig, B. K., Figueroa, M. L., Cornforth, M. N., and mosome studies in human subjects chronically exposed ing water. Am. J. Indust. Med., 6: 325-338, 1984. Jenkins, S. H. Chroto arsenic in drink- 41 . Fontham, E., Correa, P., Rodriguez, E., and Lin, Y. Validation of smoking history with the micronuclei test. In: D. Hoffman and C. C. Harris beds.), Mechanisms in Tobacco Carcinogenesis. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1986. 42. Reali, D., DiMarino, F., Bahramandpour, S., Carducci, and Loprieno, N. Micronuclei in exfoliated urothelial cells genicity in smokers. Mutat. Res., 192: 145-149, 1987. A., Barale, R., and urine muta- 43. Rosin, M. P., and Anwar, from Egyptians with chronic Cancer, 50:539-543, 1992. W. Chromosomal damage schistosoma haematobium in urothelial infections. cells Int. J. 44. Silverman, D. T., Hartage, P., Morrison, obogy ofbladder cancer. Hematology/Oncology A. S., and Devesa, S. S. Epidemi- Clinics N. Am., 6: 1-30, 1992. 45. Schulte, J., King, C., MacDonald, for recognizing abnormal epithelial 615-625, 1963. D., and Jassie, M. A simple technique cells in urinary sediment. J. Urol., 89: 46. Tyler, D. Stratified squamous urethra: findings correlated with Anatomy, 1 1 1: 319-325, 1962. epithelium in the vesical trigone and the menstrual cycle and age. Am. J. 47. Rosin, M. P., and Gilbert, A. M. Modulation humans. Mutat. Environ., 351-359, 1990. 48. Livingston, G., Reed, R., Olson, B., and Lockey, aberrations by smokeless tobacco in epithelial cells Environ. Mob. Mutagenesis, 15: 136-144, 1990. of genotoxic effects in J. Induction of nuclear of human oral mucosa. 49. Rosin, M. P., and German, J. Evidence vivo in Bloom syndrome: increased numbers cells. Human Genetics, 71: 187-191, 1985. for chromosome of micronuclei instability in in exfoliated 50. Tolbert, P. E., Shy, C. M., and Allen, J. W. Micronuclei and other nuclear anomalies in buccal smears: a field test in snuff users. Am. I. Epidemiol., 134: 840-850, 1991. 51 . Stich, H. F., and Rosin, a tool for studies in cancer 241-253, 1984. M. P. Micronuclei risk and cancer in exfoliated human cells as intervention. Cancer Left., 22: 52. Raafat, M., El-Gerwazi, urotheliab cells of bilharzial Na,tl. Cancer Inst., 1: 63-73, S., and Stich, H. Detection patients by the micronucleus 1984. of mutagenicity in test. J. Egyptian 53. Stich, H., Curtis, J., and Panda, B. Application of the micronucleus test to exfoliated cells of high cancer risk groups: tobacco chewers. Int. J. Cancer, 30:553-559, 1982. 54. Sarto, F., Finotto, S., Giacomelli, L., Mazzotti, D., Tomannin, Lewis, A. The micronucleus assay in exfoliated cells of the human mucosa. Mutagenesis, 2: 11-17, 1987. E., and buccal 55. Stich, H., Stich, W., and Panda, B. Elevated frequency of micronucle- ated cells in the buccal mucosa of individuals at high risk for oral cancer: betel quid chewers. Cancer Lett., 17: 125-134, 1982. 56. Tsai, Y., Nichols, P., Hiti, Allelic losses of chromosomes CancerRes.,50:44-47,1990. A., Williams, 9, 1 1 , and Z., Skinner, 1 7 in human D., and Jones, P. bladder cancer. 57. Olumi, Jones, P. low grade 7081-7083, A., Tsai, Y., Nichols, P., Skinner, Allelic loss of chromosome 1 7p transitional cell carcinomas of 1990. D., Cain, distinguishes the bladder. D., Bender, L., and high grade from Cancer Res., 50: 58. Shields, P. G., and Harris, C. C. Molecular epidemiology genetics of environmental cancer. JAMA, 266: 681 -687, 1 991. and the