Arsenic methylation, GSTO1 polymorphisms, and metabolic syndrome in an arseniasis endemic area of southwestern Taiwan

Arsenic methylation, GSTO1 polymorphisms, and metabolic

syndrome in arseniasis endemic area of southwestern Taiwan

Jein-Wen Chen1, PhD; Wan-Fen Li1, PhD; Ya-Hui Wang1, MSc; Yeou-Lih Huang2, PhD; Chien-Wen Sun1, MSc; Chien-Jen Chen3, PhD; Shu-Li Wang1,4*, PhD

1_{Division of Environmental Health and Occupational Medicine, National Health} Research Institutes, Taiwan;

2_{Faculty of Biomedical Laboratory Science, Kaohsiung Medical University;} 3_{Genomic Research Center, Academia Sinica, Taipei, Taiwan;}

4_{Department of Public Health, China Medical University, Taichung, Taiwan}

* Address correspondence to:

Dr. Shu-Li Wang,

Division of Environmental Health and Occupational Medicine, National Health Research Institutes, 35, Keyan Road, Zhunan Town, Miaoli County 350, Taiwan, ROC.

Tel.: +886 (0) 37 246 166 extn. 36509. Fax: +886 (0) 37-587 406. E-Mail: [email protected].

Running head: Arsenic and metabolic syndrome

Abstract 1

Previous studies have shown that hair arsenic (As) levels are associated with an 2

increased prevalence of metabolic syndrome (MetS), which is a strong predictor for 3

type 2 diabetes. The objective of this study was to evaluate whether urinary arsenic 4

methylation is related to MetS in the arseniasis endemic area of southwestern 5

Taiwan, taking genetic factors into account. Subjects were from a community-based 6

cohort recruited in 1990 from three villages in Putai township. In 2002-2003, 247 7

subjects were successfully followed, including urinary arsenic speciation 8

measurements and single-nucleotide polymorphism determinations. We found that 9

subjects with MetS drank from well water with higher As concentrations. Multiple 10

logistic regression analysis showed that the odds ratio (OR) of MetS significantly 11

increased with increasing dimethylarsinic acid ratios (dimethylarsinic acid 12

levels/Total As) and decreasing monomethylarsonic acid ratios (monomethylarsonic 13

acid levels/Total As). The highest OR (4.65, 95% CI: 2.22-9.73, p<0.001) was found 14

in the higher secondary methylation index group with the GSTO1 AA genotype 15

compared to the reference group with lower secondary methylation index and the 16

GSTO1 AA genotype. Our findings suggested that a decreasing monomethylarsonic 17

acid ratio is associated with an increased risk of MetS and that this risk might be 18

marginally modified by the GSTO1 genotype among As-exposed subjects. 19

20

Keywords: arsenic, metabolic syndrome, type 2 diabetes mellitus, risk factor, 21

methylation 22

1. Introduction 24

Due to the lack of treated tap water and the high salinity of shallow well water 25

during 1910-1960s, the local residents (before the 1970s) in an arseniasis endemic 26

area of southwestern Taiwan typically consumed artesian well water with arsenic 27

(As) concentrations as high as 0.7-0.93 mg/L (Kuo, 1964). Several studies have 28

shown a dose-response relationship between As in drinking water and the prevalence 29

of diabetes mellitus in southwestern Taiwan (Wang et al., 2003). 30

As is abundant in the environment (Smith et al., 2002) and its threat in the 31

Western Pacific region has been noted (Suk et al., 2003). A great deal of new 32

information is emerging from the extensive research on the health effects of chronic 33

As toxicity (Guha Mazumder, 2008). Chronic exposure to high levels of inorganic 34

As (iAs) is associated with a wide range of human ailments including cancer, 35

arteriosclerosis, hypertension, and type 2 diabetes (T2D) (Fu et al., 2010). Recently 36

it has been shown that even low levels of exposure to iAs in drinking water may 37

play a role in the prevalence of T2D (Navas-Acien et al., 2008). Groundwater and 38

industrial materials are important sources of As (Nordstrom, 2002). Epidemiological 39

studies carried out in Taiwan (Chiou et al., 2006), Bangladesh (Nabi et al., 2005), 40

and Mexico (Coronado-Gonzalez et al., 2007) have shown a strong diabetogenic 41

effect of As in humans. 42

It is recognized that iAs contributes to oxidative stress in several organs and 43

systems through generation of reactive oxygen species, which can function as 44

signaling molecules to activate a number of cellular stress-sensitive pathways linked 45

to insulin resistance and decreased insulin secretion (Izquierdo-Vega et al., 2006). 46

T2D accounts for 90-95% of all cases of diabetes and is a major public health 47

problem (Wild et al., 2004). The prevalence of T2D has been increasing 48

substantially worldwide, including in Taiwan (Wang et al., 2005). Conventional risk 49

factors have been identified, but they have not yet been sufficiently explained (Wang 50

et al., 1997). Metabolic syndrome (MetS) has been demonstrated as a strong 51

predictor for T2D incidence in middle-aged individuals (Wannamethee et al., 2005). 52

Established risks factors for T2D include older age, obesity, physical inactivity, 53

family history, and genetic polymorphisms (Navas-Acien et al., 2006). Moreover, As, 54

an environmental toxicant, has been suggested to play an etiologic role in the 55

development of diabetes (Tseng, 2004). Wang et al. (2007) has also reported a 56

relationship between hair arsenic levels and MetS in a cross-sectional study in 57

Taiwan. 58

The toxicity of As greatly differs among their chemical species, and methylated 59

arsenicals are considered to be less toxic than iAs. Methylation is the major 60

metabolic pathway for iAs in humans, and most of the iAs (arsenite [AsIII] and 61

arsenate [AsV]) are metabolized to monomethylarsonic acid (MMAV) and 62

dimethylarsinic acid (DMAV) before excretion in the urine. Methylation of As 63

involves a two-electron reduction of pentavalent (e.g., AsV and MMAV) to trivalent 64

(e.g., AsIII and MMAIII [monomethylarsonous acid]) As species followed by the 65

transfer of a methyl group from a methyl donor, such as S-adenosylmethionine 66

(Thompson, 1993). In the biotransformation process of As, As (+3 oxidation state) 67

methyltransferase (AS3MT) and glutathion S-transferase ω (GSTO) are required in a 68

variety of animals including humans (Aposhian and Aposhian, 2006). GST is a 69

phase II enzyme that can detoxify xenobiotics by catalyzing their conjugation with 70

reduced glutathione. GSTO1 is involved in the reduction activities of AsV, MMAV, 71

and DMAV (Agusa et al., 2010). Human AS3MT is known to catalyze the 72

methylation of arsenite (Fujihara et al., 2010). However, methylation of iAs is not 73

necessarily a detoxification process, in that methylation is paradoxically both a 74

detoxification and activation process (Thomas et al., 2007). 75

In order to evaluate the changes in disease risk, we follow a community-based 76

population with documented previous ingestion of As from artesian well water 77

(Chen et al., 1995). We aimed to assess if GSTO1 and AS3MT affected the As 78

methylation patterns and evaluated the association among As methylation patterns, 79

gene polymorphisms, and the occurrence of MetS. 80

81

2. Materials and Methods 82

2.1 Study site and subjects 83

Subjects were form a community-based cohort recruited in 1990 (Chen et al., 84

1995) consisting of 1297 men and women over 40-years-old. All subjects were from 85

Putai village, a high arsenic exposure township in southwestern Taiwan, where the 86

prevalence of black foot disease (BFD) was the highest and a median As 87

concentration in the artesian well water ranging from 700 to 930 μg/L (Kuo, 1964). 88

The soil and water from shallow wells in this area has a high salt content. The 89

residents, therefore, have used artesian well water since the 1900s. During 1972 and 90

thereafter, a tap water supply system was implemented in this BFD endemic area, 91

which almost completely supplied the entire area with municipal water. In 1997 and 92

2002-3, we successfully followed 287 subjects from the residents, and 247 subjects 93

with urinary arsenic speciation measurements and single-nucleotide polymorphism 94

determinations. 95

96

2.2 Data and specimen collections 97

Informed consent from each participant and approval by the Human Subjects 98

Review Board of National Health Research Institutes in Taiwan was obtained prior 99

to data collection. Demographic factors, artesian well water usage, personal and 100

family disease history of diabetes mellitus, dietary habits, and life style information 101

were acquired using standardized questionnaires at both baseline and follow-up. 102

Blood pressure was measured using a mercury sphygmomanometer with a suitable 103

cuff. Two measurements were carried out at least 30 min apart while subjects sat 104

calmly. If the difference between the two measurements exceeded 5% (i.e., >5 105

mmHg out of 100 mmHg), blood pressure was measured a third time. The average 106

of the two closest values was used for data analysis. The fasting blood samples were 107

stored frozen after collection and then keep in a -700_{C freezer until analysis.} 108

109

2.3 Specimen analysis 110

We quantified four As species in urine collected at the 2002-2003 follow-up: 111

arsenite (AsIII), arsenate (AsV), monomethylarsonic acid (MMA), and 112

dimethylarsinic acid (DMA). Urinary As species were analyzed using 113

high-performance liquid chromatography (HPLC) coupled with flow injection 114

atomic absorption spectrometry. The HPLC system consisted of a solvent delivery 115

pump (PU-1580; Jasco, Tokyo, Japan) and a silica-based anion-exchange column 116

(Nucleosil 10 SB, 250 mm × 4.6 mm; Phenomenex, Torrance, CA, USA) with a 117

guard column packed with the same material. A flow injection analysis system 118

(FIAS-400; PerkinElmer, Waltham, MA, USA) was designed as the on-line interface 119

to the continuous hydride generation system (Analyst 100; PerkinElmer, Waltham, 120

MA, USA). The within-day and between-day precision (coefficient of variation, 121

CV%) for AsIII, AsV, MMA, and DMA determinations ranged from 1.0 to 3.7%. 122

The recoveries for AsIII, AsV, MMA, and DMA were 99.0, 98.9, 99.0, and 99.0%, 123

while the detection limits were 0.75, 1.47, 1.19, and 0.76 μg/L, respectively. 124

Fasting plasma was analyzed in a central laboratory of the Kaohsiung Medical 125

Center for blood glucose, cholesterol, triglycerides, low- and high-density 126

lipoproteins, urine acid, and urine creatinine using a Beckmen Synchron LX20 127

System (Beckman Coulter, Brea, CA, USA). The CV% ranged between 0.4% and 128

3.4% with a mean of 2.8% and 2.2% for day-to-day and within-run measurements, 129 respectively. 130 131 2.4 Genotype Analysis 132

Genomic DNA was extracted from buffy coat using a commercial kit (Gentra 133

Puregene; Qiagen, Hilden, Germany). The genotype of GSTO1 A140D was 134

determined using polymerase chain reaction (PCR) followed by restriction enzyme 135

digestion as described previously (Marahatta et al., 2006). For the GSTO1 A140D 136

genotype, a PCR containing 100 ng of DNA was incubated at 930_{C for 5 min} 137

followed by 33 cycles at 930_{C for 60 sec, 51}0_{C for 60 seconds, and 72}0_{C for 30} 138

seconds, with a final 4-min extension at 720_{C. The A140D polymorphism was} 139

detected by Cac 8I digestion, with the result of a single 254 bp band indicating the 140

presence of the D allele, while double bands (186 and 68 bp) indicating the A allele. 141

For the GSTO2 N142D genotype, PCR conditions included an initial step at 950_{C for} 142

5 min followed by 35 cycles of 930_{C for 60 sec, 62}0_{C for 60 sec, and 72}0_{C for 30 sec,} 143

with a final 4-min extension at 720_{C. A fragment of 185 bp was amplified and} 144

analyzed using Mbo I. The presence of the variant D allele resulted in digested 145

bands of 122 and 63 bp. All PCR products were analyzed using 3.5% NuSieve 146

(Cambrex, Charles City, IA, USA) agarose gel electrophoresis. Primers for GSTO1 147

A140D were 5’-GAACTTGATGCACCCTTGGT-3’ (forward) and 148

5’-TGATAGCTAGGAGAAATAATTAC-3’ (reverse). Primers for GSTO2 N142D 149

were 5’-AGGCAGAACAGGAACTGGAA-3’ (forward) and 150

5’-GAGGGACCCCTTTTTGTACC-3’ (reverse). 151

The AS3MT M287T genotype was determined using a commercial TaqMan 152

SNP Genotyping Assay (Assay ID C_31979150_10; Applied Biosystems, Foster 153

City, CA, USA). Reactions were conducted following the instructions of the 154

manufacturer. 155

2.5 Data analysis 157

Adult Treatment Panel III (2001) (Alberti et al., 2006; Saely et al., 2006) 158

criteria were adopted to define MetS: fasting plasma glucose (≥ 110 mg/dL), 159

triglycerides (≥ 150 mg/dL), high density lipoprotein (≤ 40 mg/dL for men and ≤ 50 160

mg/dL for women), increased systolic (≥ 130 mmHg) or diastolic (≥ 85 mmHg) 161

blood pressure, and waist girth (≥ 90 cm for men and ≥ 80 cm for women). MetS 162

was defined as the presence of three or more of the risk factors mentioned above. We 163

also calculated insulin sensitivity (Si), which was the inverse of the homeostasis 164

model assessment index, defined as 22.5EXP[-ln(glucose)]/insulin (Cohen et al., 165

2006). 166

The total urine As (TotAs) in the present study was defined as the sum of AsIII, 167

AsV, MMA, and DMA. The cumulative arsenic exposure (CAE) (mg/L-y) was 168

defined as the sum of the products, derived by multiplying the As concentration in 169

the well water by the duration of water consumption during consecutive periods of 170

living at different villages (Chen et al., 1995). Primary methylation index (PMI) and 171

secondary methylation index (SMI) were defined as MMA divided by (AsIII+ AsV) 172

and as DMA divided by MMA, respectively. Because of the distribution of As was 173

skewed to the right, As concentrations were log-transformed and calculate as 174

geometric means (GM). We examined the differences between (Student’s t-test) and 175

among the groups (analysis of variance, ANOVA) using GM. A chi-square test was 176

performed for categorical variables. The univariate and multivariate logistic 177

regression analyses were used to evaluate MetS status in relation to risk factors, with 178

the multivariate model adjusted for significant risk factors identified by univariate 179

analysis. All analyses were performed using SPSS v15 (SPSS, Chicago, IL, USA). 180

181

3. Results 182

Among the 247 followed subjects, 111 were found to have MetS. Table 1 183

shows the occurrence of MetS according to demographic factors, life style, and 184

MetS-related risk factors. The mean age of the group with MetS (64.34 ± 7.93) was 185

significantly higher than that of the group without MetS (60.36 ± 9.00) (p < 0.001). 186

Insulin sensitivity showed a significant difference the groups without and with MetS 187

(0.95 ± 0.47 vs. 0.51 ± 0.37, respectively; p < 0.001). We also found that the As 188

concentration in the well water and betel nut chewing significantly correlated with 189

the risk of MetS after adjustment for age. Table 2 shows MetS according to As 190

metabolites and genotypes. MMA levels were significantly lower in the group with 191

MetS than those without MetS. Subjects without MetS tended to have a higher 192

incidence of the AA genotype, with borderline significance after adjustment for age. 193

Table 3 shows the Pearson correlation coefficients for Si and As metabolites. We 194

found that Si was associated with MMA, MMA%, PMI, and SMI. TotAs showed a 195

high correlation with other As metabolites, but not with CAE. Table 4 shows the 196

results of multiple logistic regression analysis for MetS in relation to tertiles of As 197

methylation patterns by three models. The upper groups of MMA% (MMA/TotAs), 198

DMA% (DMA/TotAs), PMI, and SMI showed significant differences as compared 199

to their reference groups. MMA% and PMI were negatively associated with MetS 200

(OR = 0.35, 95% CI: 0.18-0.66 and 0.39, 95% CI: 0.20-0.76 for MMA% > 0.11 and 201

PMI > 0.91 compared to MMA% < 0.06 and PMI < 0.44, respectively). DMA% and 202

SMI were also associated with increased risk of MetS (OR = 2.01, 95% CI: 203

1.05-3.86 and 2.61, 95% CI: 1.35-5.08 for DMA% > 0.82 and SMI > 12.03, 204

respectively). Table 5 shows the arsenic methylation patterns linked to MetS. Three 205

models for logistic regression were performed and four As methylation patterns, 206

MMA%, DMA%, PMI, and SMI, were considered. We found the MetS OR for 207

subjects with a lower PMI and higher SMI was significantly higher (3.71, 95% CI: 208

1.90-7.21) than that for subjects with a higher PMI and lower SMI. Consistently, the 209

MetS OR for subjects with a lower MMA% and higher DMA% was significantly 210

higher (3.63, 95% CI: 1.90-6.96) than that for subjects with a higher MMA% and 211

lower DMA%. Table 6 shows the arsenic methylation patterns and GSTO1 212

genotypes linked to MetS. The analysis of the four As methylation patterns and 213

GSTO1 genotypes were performed using three logistic regression models. Subjects 214

with a lower PMI, lower MMA%, and the AD+DD genotype showed a significantly 215

higher MetS risk of around 3–4 fold (ORs were 4.00 and 3.24 for lower PMI and 216

lower MMA%, respectively). Subjects with both a higher SMI and higher DMA% 217

showed a significantly higher MetS risk (ORs were 3.87 and 3.54 for higher SMI 218

and higher DMA%, respectively). 219

220

4. Discussion 221

The positive association between arsenic exposure and increased rates of 222

diabetes has been shown in Taiwan; however, its association with MetS, an 223

important risk factor and predictor for cardiovascular diseases including T2D 224

(especially for middle-age subjects), is seldom studied. A cross-sectional study in 225

Taiwan (Wang et al., 2007) indicated the increasing prevalence of MetS from the 2nd 226

tertile (0.034 μg/g) of hair arsenic levels (OR = 2.54, p = 0.015) after adjustment for 227

age, gender, occupation, and life style. To our knowledge, the present study is the 228

first to report the relation between arsenic methylation patterns and MetS, with 229

genotype and other risk covariates taken into account. 230

The excretion of methylated As metabolites in urine is the result of a number of 231

factors, though our knowledge of the importance of these factors is limited. 232

According to a report from the National Research Council (1999), it can be 233

concluded that the level of arsenic has little influence on methylation efficiency. The 234

rates of As metabolites in our study are close to those of a previous study (Vahter, 235

2000), which indicated that most individuals have an average of 10–30% iAs, 236

10–20% MMA, and 60–70% DMA (for this study, 17%, 10%, and 73% for iAs, 237

MMA, and DMA, respectively). In one study, a 500-μg/L increase in the 238

concentration of As metabolites in urine corresponded to a 2% increase in urine 239

MMA and a 3% decrease in DMA (Hopenhayn-Rich et al., 1996). We found an 240

association between AS3MT and MMA that was similar to the results of a recent 241

study (Ahusa et al., 2010). MMA%, PMI, and SMI are affected by AS3MT (MM vs. 242

MT genotype: 0.10 vs. 0.22, 0.82 vs. 1.75, and 18.25 vs. 3.45, respectively; data not 243

shown). Some studies have shown that MMA is more cytotoxic (Petrick et al., 2000) 244

and genotoxic (Mass et al., 2001) than AsIII and AsV, suggesting that the oxidation 245

state of methylated arsenicals is important for the manifestation of their toxic and/or 246

genotoxic effects. A recent study (Chung et al., 2009) showed that MMM% might be 247

a potential predictor of cancer risk and that changes in MMA% are linked to 248

individual cancer susceptibility related to AS3MT. In the present study, however, we 249

found that Si increased with increasing MMA, MMA%, PMI, and SMI. We also 250

found that lower PMI and MMA% and higher SMI and DMA% seemed to be risk 251

factors for MetS. When looking at the As methylation pattern and the genotypes of 252

GSTO1 (Table 5), As methylation patterns seems to be dominate factors of MetS 253

risk, as apposed to GSTO1. Up to now, a clear cut-off level for As exposure or 254

As-induced health effects has not been established. The biological exposure index of 255

35 μg/g creatinine is based on an estimated risk for lung cancer. The thresholds of 256

different As metabolites for developing MetS are not clear and await further 257

investigation. 258

Several factors can influence the toxicity of As towards organs involved in 259

glucose metabolism and can determine the progression of insulin resistance, such as 260

personal susceptibility with respect to genetics, nutritional uptake, health status, 261

detoxification capability, interactions with other trace elements, and the other 262

well-recognized risk factors for diabetes mellitus. Although drinking water is the 263

main source of As uptake, up to 44% of ingested As may come from food sources 264

(Del Razo et al., 2002). Another study has indicated that diet is the largest exposure 265

source for most individuals, with an average intake of about 50 μg/d from food 266

(Tchounwou et al., 2003). Other uptake routes are usually much smaller, but may 267

become significant in areas of As contamination. For accumulative As, we measured 268

individual As exposure and then multiplied the number of years lived in a specific 269

area times the average As level in the well water. The measurement, however, was 270

based on one measurement for the period and assumed constant water As 271

concentrations throughout the period. Das et al. (1989) proposed that selenium (Se) 272

and As counteract each other in regards to glucose metabolism and that the joint 273

effect of high As and low Se could play a role in developing T2D. The variation of 274

exposure to Se, consumption of As via food, and other nutrients were not considered 275

in this study. 276

Induction of oxidative stress and interference of signal transduction or gene 277

expression by As or by its methylated metabolites are the most possible causes to 278

As-induced T2D through mechanisms of induced insulin resistance (Tseng, 2004). 279

In subjects with chronic As exposure, oxidative stress is increased. The expression 280

levels of tumor necrosis factor α (TNFα) and interleukin-6 (IL-6), both cytokines, 281

are well known for their effect on the induction of insulin resistance. Fu et al. (2010) 282

indicated that low levels of As provoke a cellular adaptive oxidative stress response 283

that increases antioxidant levels, dampens reactive oxygen species signaling 284

involved in glucose-stimulated insulin secretion, and thus disturbs β-cell function. It 285

was reported that the methylation of As in individuals via drinking water (in Taiwan) 286

is associated with genetic polymorphism in GST (Chiou et al., 1997). Between 287

individuals, there are a lot of differences in As biomethylation. As such, it is likely 288

that there are genetic polymorphisms affecting the regulation of As 289

methyltransferases. It is possible that As metabolism is also affected by the 290

documented polymorphism in enzymes involved in the remethylation of 291

homocysteine. 292

Current available evidence (including in vitro, in vivo, and animal studies) are 293

inadequate to establish the cause-effect relationship between As and MetS. Due to 294

the short biological half-life of arsenic, hair or nails may be a more valid source for 295

estimating individual As exposure. Further studies should focus on the measurement 296

of appropriate samples (e.g., hair) that more accurately represent long-term As 297 exposure. 298 299 5. Conclusions 300

We found that the group of subjects with MetS had a higher As concentration in 301

their well water. AS3MT polymorphisms were associated with MMA%, PMI, and 302

SMI. Multiple logistic regression analysis showed that the OR of MetS significantly 303

decreased with increasing MMA% and decreasing DMA%. We found that subjects 304

with lower PMI and higher SMI levels had a significantly increased OR of MetS. 305

The high OR (4.65, 95% CI: 2.22-9.73, p<0.001) was found in the higher SMI group 306

with the AA genotype compared to the reference group with lower SMI and the AA 307

genotype. Our findings suggested that decreasing MMA% and PMI as well 308

increasing DMA% and SMI are associated with an increased risk of MetS and that 309

this risk might be marginally modified by the GSTO1 genotype among As-exposed 310 subjects. 311 312 Acknowledgements 313

The authors thank Ms. HY Chen for support with sample collection and 314

management. The statistical assistance from Mr. KH Chang is also acknowledged. 315

This study was supported by grant EO-094-PP-11 and EO-095-PP-07 from the 316

National Health Research Institutes (NHRI), Taiwan. 317

REFERENCES 319

Agusa, T., Iwata, H., Fujihara, J., Kunito, T., Takeshita, H., Minh, T.B., Trang, P.T., 320

Viet, P.H., Tanabe, S., 2010. Genetic polymorphisms in glutathione S-transferase 321

(GST) superfamily and arsenic metabolism in residents of the Red River Delta, 322

Vietnam. Toxicol Appl Pharmacol 242, 352-362. 323

Alberti, K.G., Zimmet, P., Shaw, J., 2006. Metabolic syndrome--a new world-wide 324

definition. A Consensus Statement from the International Diabetes Federation. 325

Diabet Med 23, 469-480. 326

Aposhian, H.V., Aposhian, M.M., 2006. Arsenic toxicology: five questions. Chem 327

Res Toxicol 19, 1-15. 328

Chen, C.J., Hsueh, Y.M., Lai, M.S., Shyu, M.P., Chen, S.Y., Wu, M.M., Kuo, T.L., 329

Tai, T.Y., 1995. Increased prevalence of hypertension and long-term arsenic 330

exposure. Hypertension 25, 53-60. 331

Chiou, H.Y., Hsueh, Y.M., Hsieh, L.L., Hsu, L.I., Hsu, Y.H., Hsieh, F.I., Wei, M.L., 332

Chen, H.C., Yang, H.T., Leu, L.C., Chu, T.H., Chen-Wu, C., Yang, M.H., Chen, 333

C.J., 1997. Arsenic methylation capacity, body retention, and null genotypes of 334

glutathione S-transferase M1 and T1 among current arsenic-exposed residents in 335

Taiwan. Mutat Res 386, 197-207. 336

Chiu, H.F., Chang, C.C., Tsai, S.S., Yang, C.Y., 2006. Does arsenic exposure 337

increase the risk for diabetes mellitus? J Occup Environ Med 48, 63-67. 338

Chung, C.J., Hsueh, Y.M., Bai, C.H., Huang, Y.K., Huang, Y.L., Yang, M.H., Chen, 339

C.J., 2009. Polymorphisms in arsenic metabolism genes, urinary arsenic 340

methylation profile and cancer. Cancer Causes Control 20, 1653-1661. 341

Cohen, D., Stolk, R.P., Grobbee, D.E., Gispen-de Wied, C.C., 2006. Hyperglycemia 342

and diabetes in patients with schizophrenia or schizoaffective disorders. Diabetes 343

Care 29, 786-791. 344

Coronado-Gonzalez, J.A., Del Razo, L.M., Garcia-Vargas, G., Sanmiguel-Salazar, F., 345

Escobedo-de la Pena, J., 2007. Inorganic arsenic exposure and type 2 diabetes 346

mellitus in Mexico. Environ Res 104, 383-389. 347

Das, P.M., Sadana, J.R., Gupta, P.K., Kumar, K., 1989. Experimental selenium 348

toxicity in guinea pigs: biochemical studies. Ann Nutr Metab 33, 57-63. 349

Del Razo, L.M., Garcia-Vargas, G.G., Garcia-Salcedo, J., Sanmiguel, M.F., Rivera, 350

M., Hernandez, M.C., Cebrian, M.E., 2002. Arsenic levels in cooked food and 351

assessment of adult dietary intake of arsenic in the Region Lagunera, Mexico. 352

Food Chem Toxicol 40, 1423-1431. 353

Fu, J., Woods, C.G., Yehuda-Shnaidman, E., Zhang, Q., Wong, V., Collins, S., Sun, 354

G., Andersen, M.E., Pi, J., 2010. Low-level arsenic impairs glucose-stimulated 355

insulin secretion in pancreatic beta cells: involvement of cellular adaptive 356

response to oxidative stress. Environ Health Perspect 118, 864-870. 357

Fujihara, J., Soejima, M., Yasuda, T., Koda, Y., Agusa, T., Kunito, T., Tongu, M., 358

Yamada, T., Takeshita, H., 2010. Global analysis of genetic variation in human 359

arsenic (+3 oxidation state) methyltransferase (AS3MT). Toxicol Appl 360

Pharmacol 243, 292-299. 361

Guha Mazumder, D.N., 2008. Chronic arsenic toxicity & human health. Indian J 362

Med Res 128, 436-447. 363

Hopenhayn-Rich, C., Biggs, M.L., Smith, A.H., Kalman, D.A., Moore, L.E., 1996. 364

Methylation study of a population environmentally exposed to arsenic in 365

drinking water. Environ Health Perspect 104, 620-628. 366

Izquierdo-Vega, J.A., Soto, C.A., Sanchez-Pena, L.C., De Vizcaya-Ruiz, A., Del 367

Razo, L.M., 2006. Diabetogenic effects and pancreatic oxidative damage in rats 368

subchronically exposed to arsenite. Toxicol Lett 160, 135-142. 369

Kuo, T.L., 1964. Arsenic content of artesian well water in endemic area of chronic 370

arsenic poisoning, in Reports of insulin of pathology. National Taiwan 371

University College of Medicine, Taipei, Taiwan 7-13. 372

Marahatta, S.B., Punyarit, P., Bhudisawasdi, V., Paupairoj, A., Wongkham, S., 373

Petmitr, S., 2006. Polymorphism of glutathione S-transferase omega gene and 374

risk of cancer. Cancer Lett 236, 276-281. 375

Mass, M.J., Tennant, A., Roop, B.C., Cullen, W.R., Styblo, M., Thomas, D.J., 376

Kligerman, A.D., 2001. Methylated trivalent arsenic species are genotoxic. 377

Chem Res Toxicol 14, 355-361. 378

Nabi, A.H., Rahman, M.M., Islam, L.N., 2005. Evaluation of biochemical changes 379

in chronic arsenic poisoning among Bangladeshi patients. Int J Environ Res 380

Public Health 2, 385-393. 381

National Research Council, 1999. Risk assessment of arsenic in drinking water. 382

Subcommittee on arsenic in drinking water. National Academy Press, 383

Washington DC, pp. 263. 384

Navas-Acien, A., Silbergeld, E.K., Pastor-Barriuso, R., Guallar, E., 2008. Arsenic 385

exposure and prevalence of type 2 diabetes in US adults. JAMA 300, 814-822. 386

Navas-Acien, A., Silbergeld, E.K., Streeter, R.A., Clark, J.M., Burke, T.A., Guallar, 387

E., 2006. Arsenic exposure and type 2 diabetes: a systematic review of the 388

experimental and epidemiological evidence. Environ Health Perspect 114, 389

641-648. 390

Nordstrom, D.K., 2002. Worldwide occurrences of arsenic in ground water. Science 391

296, 2143-2145. 392

Petrick, J.S., Ayala-Fierro, F., Cullen, W.R., Carter, D.E., Vasken Aposhian, H., 2000. 393

Monomethylarsonous acid (MMA(III)) is more toxic than arsenite in Chang 394

human hepatocytes. Toxicol Appl Pharmacol 163, 203-207. 395

Saely, C.H., Koch, L., Schmid, F., Marte, T., Aczel, S., Langer, P., Hoefle, G., Drexel, 396

H., 2006. Adult Treatment Panel III 2001 but not International Diabetes 397

Federation 2005 criteria of the metabolic syndrome predict clinical 398

cardiovascular events in subjects who underwent coronary angiography. 399

Diabetes Care 29, 901-907. 400

Smith, A.H., Lopipero, P.A., Bates, M.N., Steinmaus, C.M., 2002. Arsenic 401

epidemiology and drinking water standards. Science 296, 2145-2146. 402

Suk, W.A., Ruchirawat, K.M., Balakrishnan, K., Berger, M., Carpenter, D., Damstra, 403

T., de Garbino, J.P., Koh, D., Landrigan, P.J., Makalinao, I., Sly, P.D., Xu, Y., 404

Zheng, B.S., 2003. Environmental threats to children's health in Southeast Asia 405

and the Western Pacific. Environ Health Perspect 111, 1340-1347. 406

Tchounwou, P.B., Patlolla, A.K., Centeno, J.A., 2003. Carcinogenic and systemic 407

health effects associated with arsenic exposure--a critical review. Toxicol Pathol 408

31, 575-588. 409

Thomas, D.J., Li, J., Waters, S.B., Xing, W., Adair, B.M., Drobna, Z., Devesa, V., 410

Styblo, M., 2007. Arsenic (+3 oxidation state) methyltransferase and the 411

methylation of arsenicals. Exp Biol Med (Maywood) 232, 3-13. 412

Thompson, D.J., 1993. A chemical hypothesis for arsenic methylation in mammals. 413

Chem Biol Interact 88, 89-14. 414

Tseng, C.H., 2004. The potential biological mechanisms of arsenic-induced diabetes 415

mellitus. Toxicol Appl Pharmacol 197, 67-83. 416

Vahter, M., 2000. Genetic polymorphism in the biotransformation of inorganic 417

arsenic and its role in toxicity. Toxicol Lett 112-113, 209-217. 418

Wang, S.L., Chang, F.H., Liou, S.H., Wang, H.J., Li, W.F., Hsieh, D.P., 2007. 419

Inorganic arsenic exposure and its relation to metabolic syndrome in an 420

industrial area of Taiwan. Environ Int 33, 805-811. 421

Wang, S.L., Chiou, J.M., Chen, C.J., Tseng, C.H., Chou, W.L., Wang, C.C., Wu, T.N., 422

Chang, L.W., 2003. Prevalence of non-insulin-dependent diabetes mellitus and 423

related vascular diseases in southwestern arseniasis-endemic and nonendemic 424

areas in Taiwan. Environ Health Perspect 111, 155-159. 425

Wang, S.L., Pan, W.H., Hwu, C.M., Ho, L.T., Lo, C.H., Lin, S.L., Jong, Y.S., 1997. 426

Incidence of NIDDM and the effects of gender, obesity and hyperinsulinaemia in 427

Taiwan. Diabetologia 40, 1431-1438. 428

Wannamethee, S.G., Shaper, A.G., Lennon, L., Morris, R.W., 2005. Metabolic 429

syndrome vs Framingham Risk Score for prediction of coronary heart disease, 430

stroke, and type 2 diabetes mellitus. Arch Intern Med 165, 2644-2650. 431

Wen, C.P., Cheng, T.Y., Tsai, S.P., Hsu, H.L., Wang, S.L., 2005. Increased mortality 432

risks of pre-diabetes (impaired fasting glucose) in Taiwan. Diabetes Care 28, 433

2756-2761. 434

Wild, S., Roglic, G., Green, A., Sicree, R., King, H., 2004. Global prevalence of 435

diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care 27, 436

1047-1053. 437

Table 1. Occurrence of metabolic syndrome according to demographic factors, life style, and metabolic syndrome parameters (n = 247)

Characteristics Metabolic Syndrome p-valuea p-valueb No ( n = 136 ) Yes ( n = 111 ) Mean ± SD Mean ± SD Continuous variable Age (y) 60.36 ± 9.00 64.34 ± 7.93 < 0.001***

Residing yrs (y) 42.29 ± 14.08 46.66 ± 12.46 0.011* 0.97

Drinking yrs (y) 19.95 ± 12.90 25.31 ± 12.32 0.001** 0.13

As concentration in well water (μg/L)c 569.94 ± 321.51 684.39 ± 245.93 0.005** 0.03*

CAE (mg/L-y) 13.96 ± 9.03 17.71 ± 8.62c 0.001** 0.13

Glucose (mg/dL) 97.77 ± 42.64 127.39 ± 56.85 < 0.001*** < 0.001*** Waist girth (cm) 80.73 ± 9.18 89.77 ± 10.56 < 0.001*** < 0.001*** Systolic blood pressure (mmHg) 128.42 ± 19.51 146.32 ± 20.24 < 0.001*** < 0.001*** Diastolic blood pressure (mmHg) 77.39 ± 12.65 84.78 ± 14.09 < 0.001*** < 0.001*** Triglycerides (mg/dL) 98.40 ± 49.27 181.34 ± 102.80 < 0.001*** < 0.001*** High density lipoprotein (mg/dL) 45.11 ± 12.57 33.57 ± 7.85 < 0.001*** < 0.001*** Insulin sensitivityd 0.95 ± 0.47 0.51 ± 0.37 < 0.001*** < 0.001*** Frequency (%) Frequency (%) Categorical variable Gender Female 79 (58.09) 62 (55.86) _{0.72 0.81} Male 57 (41.91) 49 (44.14) Alcohol drinking No 126 (92.65) 98 (89.09) _{0.33 0.27} Yes 10 (7.35) 12 (10.91)

Betel nut chewing

No 126 (95.45) 98 (89.91) _{0.09 0.045}*

Yes 6 (4.55) 11 (10.09)

Smoking

No 105 (80.15) 78 (72.90) _{0.19 0.28}

Yes (including ever) 26 (19.85) 29 (27.10)

Secondary Smoking No 92 (71.32) 81 (75.00) 0.52 0.69 Yes 37 (28.68) 27 (25.00) Education No 36 (26.47) 40 (36.04) 0.11 0.46 Primary 71 (52.21) 57 (51.35)

High school or higher 29 (21.32) 14 (12.61) Exercise

No 75 (55.15) 59 (53.15) _{0.75 0.47}

Yes 61 (44.85) 52 (46.85)

a_{Student’s T test for continuous variables, chi-square test for categorical variables} b_age-adjusted

c_{102 subjects had records available}

d_{insulin sensitivity, which was the inverse of the homeostasis model assessment index, was defined as}

22.5EXP[-ln(glucose)]/insulin

Table 2. Metabolic syndrome according to arsenic patterns and genotypes (n = 247) Characteristics Metabolic Syndrome p-valueb _p-valuec No ( n = 136 ) Yes ( n = 111 ) Mean ± SD Mean ± (SD) Arsenic metabolitesa iAs% 17.2 ± 14.4 16.8 ± 16.2 0.84 0.98 oAs% 82.6 ± 14.1 83.4 ± 15.8 0.84 0.98 AsIII% 10.1 ± 11.6 10.4 ± 13.1 0.73 0.98 AsV% 7.3 ± 7.3 6.9 ± 6.5 0.85 0.99 MMA% 10.6 ± 9.2 9.4 ± 8.8 0.08# 0.06# DMA% 72.3 ± 18.4 73.5 ± 20.0 0.49 0.37 PMI 0.92 ± 0.75 0.75 ± 0.71 0.08# 0.06# SMI 16.12 ± 31.21 19.98 ± 32.71 0.35 0.42 AsIII (μg/g creatinine) 3.79 ± 4.60 4.12 ± 6.15 0.64 0.79 AsV (μg/g creatinine) 1.78 ± 1.93 1.47 ± 1.14 0.11 0.29 iAs (μg/g creatinine) 5.57 ± 4.84 5.58 ± 6.08 0.98 0.97 MMA (μg/g creatinine) 4.51 ± 5.92 3.08 ± 2.95 0.02* 0.03* DMA (μg/g creatinine) 32.79 ± 28.94 35.52 ± 36.98 0.53 0.42 oAs (μg/g creatinine) 37.30 ± 30.92 38.60 ± 38.26 0.77 0.63 TotAs (μg/g creatinine) 42.87 ± 33.06 44.19 ± 40.94 0.78 0.66 Frequency (%) Frequency (%) Genotypes GST01 A140D AA 92 (68.15) 66 (60.00) 0.08 0.06 AD 37 (27.41) 31 (28.18) DD 6 (4.44) 13 (11.82) GST02 N142D NN 68 (50.75) 53 (47.75) 0.84 0.61 ND 53 (39.55) 48 (43.24) DD 13 (9.70) 10 (9.01) AS3MT M287T MM 130 (97.74) 109 (99.09) 0.63d 0.29 MT 3 (2.26) 1 (0.91)

a_{iAs: inorganic arsenic, AsIII+AsV; oAs: organic arsenic, MMA+DMA; TotAs: total arsenic, iAs+oAs; PMI:}

MMA/(AsIII+AsV); SMI: DMA/MMA; (As metabolites)%: (As metabolites)/TotAs

b_{Student’s T test for continuous variables, chi-square test for categorical variables} c_age-adjusted

d_{Fisher's Exact Test} #_p<0.1, *_p<0.05

Table 3. Pearson correlation coefficients for insulin sensitivity and arsenic metabolites (n = 247)

Si CAE TotAs MMA DMA iAs MMA% DMA% PMI SMI

Si 1.0 CAEa -0.03 1.0 TotAs 0.05 0.11# 1.0 MMA 0.14* 0.09 0.40*** 1.0 DMA 0.04 0.09 0.97*** 0.26*** 1.0 iAs 0.03 0.18** 0.49*** 0.26*** 0.36*** 1.0 MMA% 0.14* 0.02 0.23*** 0.62*** -0.33*** -0.08 1.0 DMA% -0.03 -0.002 0.37*** -0.19*** 0.48*** -0.23*** -0.64*** 1.0 PMI 0.14* 0.04 0.11# 0.66*** 0.06 -0.22*** 0.60*** -0.03 1.0 SMI 0.14* -0.03 0.45*** -0.23*** 0.52*** 0.15** -0.41*** 0.35*** -0.30*** 1.0 a_{n = 238} #_p<0.1, *_p<0.05, **_p<0.01, ***_p<0.001

Table 4. Multiple logistic regression analysis for metabolic syndrome in relation to tertiles of previously exposed arsenic methylation patterns (n = 247)

Characteristics OR 95% CI p-value ORa _{95% CI p-value OR}b _{95% CI p-value}

As concentration in well water (μg/L)c

< 700 1.00 1.00 1.00 700 - 767.65 1.56 (0.85 - 2.84) 0.15 1.35 (0.72 - 2.53) 0.34 1.25 (0.66 - 2.39) 0.49 > 767.65 1.51 (0.81 - 2.80) 0.19 1.33 (0.71 - 2.51) 0.38 1.24 (0.65 - 2.37) 0.52 CAE < 12.60 1.00 1.00 1.00 12.60 - 18.90 1.56 (0.83 - 2.93) 0.17 1.04 (0.51 - 2.15) 0.91 1.01 (0.48 - 1.89) 0.98 > 18.90 2.71 (1.42 - 5.17) 0.003** 1.96 (0.83 - 4.65) 0.13 1.73 (0.72 - 4.19) 0.22 AsIII% < 5.2 1.00 1.00 1.00 5.2 - 8.9 1.02 (0.54 - 1.91) 0.95 1.05 (0.55 - 2.00) 0.89 1.09 (0.56 - 2.11) 0.80 > 8.9 0.94 (0.52 - 1.71) 0.84 0.88 (0.48 - 1.63) 0.69 0.78 (0.41 - 1.49) 0.45 AsV% < 3.1 1.00 1.00 1.00 3.1 - 6.4 0.87 (0.46 - 1.63) 0.66 0.91 (0.47 - 1.77) 0.78 0.86 (0.43 - 1.71) 0.67 > 6.4 0.81 (0.45 - 1.46) 0.48 0.80 (0.43 - 1.46) 0.46 0.80 (0.43 - 1.49) 0.48 MMA% < 5.8 1.00 1.00 1.00 5.8 - 11.3 0.96 (0.51 - 1.80) 0.89 0.97 (0.50 - 1.89) 0.92 0.91 (0.45 - 1.81) 0.78 > 11.3 0.36 (0.19 - 0.68) 0.001** _{0.34 (0.18 - 0.64) 0.001}** _{0.35 (0.18 - 0.66) 0.001}** DMA% < 72.4 1.00 1.00 1.00 72.4 - 81.9 1.57 (0.84 - 2.94) 0.16 1.59 (0.84 - 2.99) 0.15 1.60 (0.83 - 3.08) 0.17 > 81.9 1.83 (0.99 - 3.38) 0.06# _{1.94 (1.03 - 3.65) 0.04}* _{2.01 (1.05 - 3.86) 0.04}* PMI < 0.44 1.00 1.00 1.00 0.44 - 0.91 0.86 (0.47 - 1.60) 0.64 0.83 (0.44 - 1.58) 0.56 0.82 (0.42 - 1.60) 0.57 > 0.91 0.40 (0.21 - 0.76) 0.005** 0.39 (0.20 - 0.74) 0.004** 0.39 (0.20 - 0.76) 0.006** SMI < 6.06 1.00 1.00 1.00 6.06 - 12.03 1.23 (0.65 - 2.33) 0.52 1.29 (0.68 - 2.45) 0.44 1.23 (0.63 - 2.39) 0.54 > 12.03 2.44 (1.31 - 4.56) 0.005** _{2.71 (1.41 - 5.21) 0.003}** _{2.61 (1.35 - 5.08) 0.005}** a_{adjusted by age}

b_{adjusted by age and betel nut chewing} c_{n = 238}

Table 5. Arsenic methylation patterns and metabolic syndrome (n = 247)

Metabolic syndrome

OR 95% CI p-value ORa 95% CI p-value ORb 95% CI p-value

No (n = 136) Yes (n = 111) Frequency (%) PMI SMI High Low 58 (67.44) 28 (32.56) 1.00 1.00 1.00 Low Low 23 (65.16) 14 (37.84) 1.26 (0.57 - 2.82) 0.57 1.30 (0.58 - 2.93) 0.52 1.41 (0.61 - 3.26) 0.43 High High 21 (56.76) 16 (43.24) 1.58 (0.72 - 3.48) 0.26 1.7 (0.76 - 3.84) 0.19 2.17 (0.92 - 5.09) 0.08# Low High 34 (39.08) 53 (60.92) 3.23 (1.73 - 6.02) <0.001*** _{3.41 (1.79 - 6.50) <0.001}*** _{3.71 (1.90 - 7.21) <0.001}*** MMA/TotAs DMA/TotAs High Low 64 (68.09) 30 (31.91) 1.00 1.00 1.00 Low Low 169 (61.54) 10 (38.46) 1.33 (0.54 - 3.28) 0.53 1.35 (0.55 - 3.32) 0.52 1.27 (0.51 - 3.18) 0.61 High High 16 (51.61) 15 (48.39) 2.00 (0.88 - 4.57) 0.10 2.05 (0.89 - 4.70) 0.09# _{1.96 (0.82 - 4.65) 0.13} Low High 40 (41.67) 56 (58.33) 2.99 (1.65 - 5.41) <0.001*** _{3.32 (1.78 - 6.19) <0.001}*** _{3.63 (1.90 - 6.96) <0.001}*** a_{adjusted by age}

b_{adjusted by age and betel nut chewing}

Table 6. Arsenic methylation patterns and GSTO1 genotypes linked to metabolic syndrome (n = 247)

Metabolic syndrome

OR 95% CI p-value ORa 95% CI p-value ORb 95% CI p-value

No (n = 135) Yes (n = 110) Frequency (%) PMI GSTO1 High AA 48 (66.67) 24 (33.330 1.00 1.00 1.00 High AD + DD 30 (60.00) 20 (40.00) 1.33 (0.63 - 2.82) 0.45 1.29 (0.60 - 2.76) 0.52 1.32 (0.61 - 2.90) 0.48 Low AA 44 (51.16) 42 (48.84) 1.91 (1.00 - 3.65) 0.05# _{1.94 (1.00 - 3.77) 0.05}* _{2.03 (1.03 - 4.01) 0.04}* Low AD + DD 13 (35.14) 24 (64.86) 3.69 (1.60 - 8.50) <0.001*** 3.96 (1.65 - 9.47) 0.002** 4.00 (1.65 - 9.71) 0.002** SMI GSTO1 Low AA 57 (74.03) 20 (25.97) 1.00 1.00 1.00 Low AD + DD 23 (51.11) 22 (48.89) 2.73 (1.26 - 5.92) 0.01* _{2.78 (1.27 - 6.07) 0.01}* _{2.91 (1.29 - 6.57) 0.01}* High AA 35 (43.21) 46 (56.79) 3.75 (1.91 - 7.34) <0.001*** 4.25 (2.10 - 8.60) <0.001*** 4.65 (2.22 - 9.73) <0.001*** High AD + DD 20 (47.62) 22 (52.38) 3.14 (1.42 - 6.92) 0.005** _{3.15 (1.39 - 7.14) 0.006}** _{3.87 (1.62 - 9.26) 0.002}** MMA% GSTO1 High AA 55 (71.43) 22 (28.57) 1.00 1.00 1.00 High AD + DD 24 (51.06) 23 (48.94) 2.40 (1.13 - 5.10) 0.02* 2.40 (1.12 - 5.14) 0.02* 2.35 (1.07 - 5.15) 0.03* Low AA 37 (45.68) 44 (54.32) 2.97 (1.54 - 5.75) 0.001** _{3.28 (1.65 - 6.52) <0.001}*** _{3.36 (1.66 - 6.79) <0.001}*** Low AD + DD 19 (47.50) 21 (52.50) 2.76 (1.25 - 6.11) 0.01* _{2.87 (1.26 - 6.53) 0.01}* _{3.24 (1.37 - 7.66) 0.007}** DMA% GSTO1 Low AA 57 (71.25) 23 (28.750 1.00 1.00 1.00 Low AD + DD 22 (56.41) 17 (43.59) 1.92 (0.86 - 4.25) 0.11 1.93 (0.87 - 4.30) 0.11 1.86 (0.83 - 4.18) 0.13 High AA 35 (44.87) 43 (55.13) 3.05 (1.58 - 5.88) <0.001*** _{3.50 (1.75 - 6.99) <0.001}*** _{3.44 (1.69 - 6.99) <0.001}*** High AD + DD 21 (43.75) 27 (56.25) 3.19 (1.51 - 6.73) 0.002** _{3.32 (1.53 - 7.19) 0.002}** _{3.54 (1.59 - 7.87) 0.002}** a_{adjusted by age}

b_{adjusted by age and betel nut chewing}