Investigating the association between urinary levels of acrylonitrile metabolite N-acetyl-S-(2-cyanoethyl)-L-cysteine and the oxidative stress product 8-hydroxydeoxyguanosine in adolescents and young adults

Chien-Yu Lin, Hui-Ling Lee, Fung-Chang Sung, Ta-Chen Su
a Department of Internal Medicine, En Chu Kong Hospital, New Taipei City, 237, Taiwan
b School of Medicine, Fu Jen Catholic University, New Taipei City, 242, Taiwan
c Department of Chemistry, Fu Jen Catholic University, New Taipei City, 242, Taiwan
d Department of Health Services Administration, College of Public Health, China Medical University, Taichung, 404, Taiwan
e Department of Internal Medicine, National Taiwan University Hospital, Taipei, 10002, Taiwan
f Institute of Occupational Medicine and Industrial Hygiene, National Taiwan University College of Public Health, Taipei, Taiwan

Acrylonitrile is a colorless volatile liquid mostly present in tobacco smoke. Acrylonitrile exposure has shown to increase oxidative stress in animal studies; however, there was no previous research in human epidemiology. In this study, 853 subjects were recruited from a cohort of Taiwanese adolescents and young adults to investigate the association between urinary concentrations of the acrylonitrile metab- olite N-acetyl-S-(2-cyanoethyl)-L-cysteine (CEMA), the oxidative stress product 8- hydroxydeoxyguanosine (8-OHdG) and cardiovascular disease (CVD) risk factors. The geometric mean (SD) of CEMA and 8-OHdG concentrations were 4.67 (8.61) mg/L and 2.97 (2.14) mg/L, respectively. 10% elevated in CEMA (mg/L) was positively correlated with the change of 8-OHdG levels (mg/L) (b ¼ 0.325, SE ¼ 0.105, P ¼ 0.002) in multiple linear regression analyses. The urinary CEMA was not related to other CVD risk factors. In subpopulation analyses, the association between CEMA and 8-OHdG was evident in all genders, adolescents, homeostasis model assessment of insulin resistance score ≥0.89, and envi- ronmental tobacco smokers. In this study, we observed that higher levels of CEMA levels were correlated with increased levels of 8-OHdG in this cohort. Future research on exposure to acrylonitrile and oxidative stress was warranted.

1. Introduction
Acrylonitrile, a high-volume commodity chemical known as vinyl cyanide (CH2 CHC^N), is used in clothing, carpeting, food containers, computers, and medical tubing (Watcharasit et al., 2010), with an estimated yearly production of more than 10 billion pounds (Neal et al., 2009). The possible ways of human occupational contact to acrylonitrile are primary by inhalation and dermal contact (Marsh and Zimmerman, 2015). In addition to the occupational exposure, acrylonitrile exposure is limited to residual acrylonitrile in commercial material, fires, auto exhaust, and to- bacco smoke in the general population (Leonard et al., 1999). To- bacco smoke is the most important source of acrylonitrile exposure, where it is present at concentrations of 1e2 mg acrylonitrile per 100 cigarettes (WHO, 2000). If only 10% were absorbed, smoking each cigarette would result in 1e2 mg of acrylonitrile intake (WHO, 2000).
In humans, acrylonitrile is metabolized by two main pathways. The first is by microsomal cytochrome P450 to form epoxide cya- noethyl oxide (glycidonitrile, 2-cyanoethylene oxide). Second, acrylonitrile and glycidonitrile are converted to mercapturic acids by conjugation with glutathione and voided into urine (WHO, 2000). N-acetyl-S-(2-cyanoethyl)-L-cysteine (CEMA) is the main urinary excretory product after acrylonitrile exposure. In one study with six male volunteers, after exposure to acrylonitrile at 5 or 10 mg/m3 for 8 h, the respiratory system retained 52% of the acry- lonitrile, and CEMA composed of 21.8% retained acrylonitrile and voided in urine with half-life of 8 h (Jakubowski et al., 1987). Uri- nary CEMA has been regarded as a suitable short-term biomarker for acrylonitrile exposure (H.W, 2007; Jakubowski et al., 1987; Minet et al., 2011).
In experimental animals, acrylonitrile had little evidence to cause reproductive or developmental toxicity (H.W, 2007). How- ever, the evidence for the carcinogenicity of acrylonitrile in chronic toxicity is sufficient, while the predominant type is central nervous system tumors in rats (De Smedt et al., 2014). In humans, acute exposure to acrylonitrile is toxic at relatively low levels. Exposure between 20 and 150 ppm may cause skin and eye irritation, nausea, and dizziness and exposure above 500 ppm is lethal (Cole et al., 2008). Moreover, reduced hemoglobin, erythrocyte, and white blood cell counts have also been reported (H.W, 2007). The refer- ence concentration for acrylonitrile in the air is 0.002 mg/m3 ac- cording to animal study (EPA, 2000). In chronic exposure, the toxicity of acrylonitrile exposure in the reproduction and devel- opment in humans is unavailable (H.W, 2007; Neal et al., 2009). Since the cancer incidence in human epidemiological findings was inconsistent (Blair et al., 1998; Cole et al., 2008), acrylonitrile has been classified by the International Agency for Research on Cancer as group 2B carcinogen (WHO, 1999).
It has been reported that exposure to acrylonitrile increased oxidative stress in animal studies, and this has been proposed a possible model of the carcinogenic process in rat brain tumors (El- Sayed el et al., 2008; Pu et al., 2016; Rongzhu et al., 2009). An in- crease in oxidative stress is thought to be involved a vital role in cardiovascular disease (CVD) in recent studies (Csanyi and Miller, 2014). The oxidized nucleoside 8-hydroxydeoxyguanosine (8- OHdG) is developed during the repair of oxidative damage to DNA and is eliminated by excretion in the urine (Wu et al., 2004). It has been suggested that urinary 8-OHdG is a suitable oxidative stress biomarker (Sajous et al., 2008).
Given the findings from the in vitro and animal studies mentioned above, it is rational to question if acrylonitrile exposure is correlated with increased oxidative stress and CVD risk factors in humans. However, the relationship had never been studied in epidemiology. To test our hypothesis, a cross-sectional study was designed in a cohort composed of adolescents and young adults in Taiwan. We used urinary CEMA as a biomarker of acrylonitrile exposure and urinary 8-OHdG as a biomarker of oxidative stress.

2. Materials and methods
2.1. Study population and data collection
From 2006 to 2008, we established a cohort (the Young Taiwanese Cohort Study) composed of 886 subjects selected from the 1992e2000 mass urine screening program in Taiwan (Lin et al., 2016; Wei et al., 2003). Physical check-ups were performed after written informed consent. All methods in this study were per- formed in accordance with the relevant guidelines and approved by the Research Ethics Committee of at the National Taiwan University Hospital. A flow chart of this study is shown in Fig. 1. Of the 886 subjects, urine samples for testing CEMA were unavailable in 33 participants. Finally, 853 subjects were recruited in the present study. The detailed information was available in the supplemental section.

2.2. Measurement of urinary levels of CEMA
LC-MS/MS analysis to measure CEMA was performed as described by Chiang et al. (2015). The limit of detection for CEMA was 0.077 ng/mL. The methods were detailed in the supplemental section.

2.3. Measurement of urinary levels of 8-OHdG
LC-MS/MS analysis to measure 8-OHdG was performed as described by Chen et al. (2016). The limit of detection for 8-OHdG was 0.02 ng/mL. The methods were detailed in the supplemental section.

2.4. Covariates
Age and gender were recorded during the interview. Household income was divided as above or below 50,000 New Taiwan Dollars (NTD) per month. The body mass index (BMI) was measured as weight (in kg) divided by the square of the height (in m). For par- ticipations 20-year-old, BMI z-score was calculated by equation (BMI of each participant – mean of BMI)/(standard deviation of BMI) while aged 12e19 years were based on WHO anthropometric calculator (WHO, 2018). Smoking status was categorized as non- exposed, environmental tobacco smoker, <10 cigarettes/day, 10e19 cigarettes/day, and 20 cigarettes/day. The detection limit of serum cotinine was below 0.05 ng/mL. Subjects with detectable serum cotinine and who did not report smoking were defined as exposure to environmental tobacco smoke. Those with undetect- able serum cotinine and no self-reported smoking were defined as non-exposed (Weitzman et al., 2005). Alcohol consumption was divided into current alcohol consumers or non-current consumers. Blood pressure measurements were recorded twice after 3 min of rest, at least 1 min apart, using a mercury manometer. In adults, hypertension was determined by the self-reported current use of anti-hypertensive medication or an average blood pressure 140/ 90 mmHg. Childhood elevated blood pressure was defined as either systolic blood pressure or diastolic blood pressure, or both, that were greater than or equal to the modified sex- and age-specific criteria for blood pressure values (The fourth report on the diag- nosis, evaluation, and treatment of high blood pressure in children and adolescents) (Battistoni et al., 2015). Diabetes mellitus was defined as a fasting serum glucose level 126 mg/dL or the self- reported current use of oral hypoglycemic agents or insulin. The levels of urine creatinine and serum creatinine, uric acid, total cholesterol, low- and high-density lipoprotein cholesterol (LDL-C and HDL-C), triglyceride and glucose were quantified with an autoanalyzer (Technician RA, 2000 Autoanalyzer, Bayer Diagnostic, Mishawaka, IN) and are expressed as mg/dL. Serum insulin con- centrations were analyzed by the kit IMMULITE 2000 (Siemens Healthcare Diagnostics, Tarrytown, NY). Serum cotinine concen- tration was measured by the DRI Cotinine Assay for urine (Micro- genics Corp., Fremont, CA) on a Dimension RXL analyzer (Siemens Healthcare Diagnostics, Inc.,Tarrytown, NY). 2.5. Statistical analysis The unit of CEMA was expressed as mg/L or mg/g creatinine after corrected for urine creatinine in the study. The concentration of CEMA was described by the exponential mean and SD calculated from natural log CEMA in different subgroups and tested by Stu- dent's two-tailed t-test and one-way analysis of variance. In order to investigate the relationships between urinary CEMA level (mg/L) and CVD risk factors (glucose homeostasis, lipid profiles, oxidative stress, uric acid, BMI z score), we used an extended model approach to test the associations. Besides CVD risk factors we interested, we used age and gender as covariates of Model 1 while age, gender, smoking, drinking status, BMI z score and urinary creatinine as covariates of model 2. We use urinary creatinine as a separate in- dependent variable instead of an adjustment for hydration in multiple regression analysis based on previous literature (Barr et al., 2005; O'Brien et al., 2016). An association was considered to be significant only when both models to be statistically signifi- cant. To further evaluate the relationship between CEMA and 8- OHdG in multiple linear regression analysis, we used gender, age, smoking, and drinking status, BMI z score, systolic blood pressure, LDL-C, HOMA-IR and urinary creatinine as covariates of model 3. The CEMA concentrations were stratified into quartiles to evaluate the association with 8-OHdG to evaluate dose-response relation- ship. The linear association between 8-OHdG and CEMA in the different subgroup of participants was also shown in the results. The natural log transformation was conducted for CEMA, 8-OHdG, insulin, and HOMA-IR because of divergence from the normal dis- tribution. Statistical significance was defined as P < 0.05.

3. Results
The geometric mean (SD) of CEMA and 8-OHdG concentrations were 4.67 (8.61) mg/L and 2.97 (2.14) mg/L, respectively. The urine levels of CEMA were below LOD in 11.6% of study subjects. The demographics of the study participants are demonstrated in Table 1. There were 337 males and 516 females with a mean age (range) were 21.3 (12e30) years in this study. CEMA levels were higher in participants with male gender, older age ( 20 years old), active smokers and drinkers. The levels of CEMA were not different between no exposure and exposure to environmental tobacco smoke by analysis of covariance (P 0.994). The mean CEMA con- centrations were not different between subgroup of household income, BMI, HOMA-IR, hypertension, or diabetes mellitus.
The association between CEMA levels and CVD risk factors after adjustment for two models is listed in Table 2. One-unit elevated in natural log-transformed CEMA (mg/L) was positively correlated with urinary 8-OHdG levels in the two models (b 0.034, SE 0.011, p 0.003 in the model 2). There was no significant correlation be- tween CEMA concentrations and other CVD risk factors. The corre- lation between quartiles of CEMA and 8-OHdG in multiple linear regression models is shown in Table 3. After controlling for model 3, urinary 8-OHdG increased significantly with increasing quartiles of CEMA (P for trend 0.003) with highest mean levels in the third quartile. The levels of 8-OHdG were not different between third and fourth quartiles of CEMA by analysis of covariance (P 0.713). When comparing the third to first quartiles, acrylonitrile exposure was associated with a 1.23-fold increase in 8-OHdG levels.

4. Discussion
In this study, we showed that elevated urinary CEMA concen- trations are positively correlated with urinary 8-OHdG in this Taiwanese cohort composed of adolescents and young adults. To the best of our knowledge, this finding is the first to connect urinary CEMA levels to oxidative stress in humans. When comparing the highest to lowest quartiles, acrylonitrile exposure was associated with a 1.4-fold increase in 8-OHdG levels in the present study. Since urinary CEMA has been regarded as a biomarker of short-term acrylonitrile exposures (H.W, 2007; Jakubowski et al., 1987; Minet et al., 2011), low-dose acrylonitrile exposure may increase oxida- tive stress in this study population.
Tobacco smoke is the most important source of acrylonitrile exposure in the general community (WHO, 2000). In this research, the urine levels of CEMA were detectable in 88.4% of the study subjects and 86.4% of non-exposed participants (who did not report smoking and undetectable serum cotinine). The source of acrylonitrile exposure in non-exposed subjects is not known. In- door and outdoor air is the source of acrylonitrile contamination and may be present indoors at lower concentrations due to its reactivity (WHO, 2000). In the general population of the Netherlands, around the 1980s, the average annual ambient air concentration of acrylonitrile was estimated to be as low as 0.01 mg/ m3. Many reports of poisoning have been published in the U.S. and Germany that have been fumigated with mixtures of acrylonitrile and carbon tetrachloride or dichloromethane in homes (WHO, 2000). Accidental fires, residual acrylonitrile in commercial mate- rial, and auto exhaust may be the sources of acrylonitrile other than tobacco smoke in this study (Leonard et al., 1999).
Instead of direct DNA reactivity, acrylonitrile exposure was found to increase the oxidative DNA damage in recent animal studies (El-Sayed el et al., 2008; Pu et al., 2016; Rongzhu et al., 2009). The conjugation of acrylonitrile and glutathione by gluta- thione-S-transferase is a major route for the detoxification of acrylonitrile. Moreover, acrylonitrile is oxidized to glycidonitrile, which also undergoes conjugation with glutathione (WHO, 2000). In rat brain, acrylonitrile induces oxidative stress through decreasing glutathione levels and the activity of superoxide dis- mutase (Pu et al., 2016). However, the dose of acrylonitrile exposure is high (ranging from 10 mg/kg to 100 mg/kg) in animal studies (Jiang et al., 1998; Pu et al., 2009, 2016), which is much higher than in the general population. No previous study has investigated low- dose acrylonitrile exposure and oxidative stress. Our cross- sectional study provided the first evidence that low-dose acrylo- nitrile exposure may increase oxidative stress in humans.
Environmental tobacco smoke was classified as a group A carcinogen and had been a risk factor for CVD (Raupach et al., 2006). In a smoky room, sidestream smoke accounts for 85 percent of the environmental tobacco smoke (Columbia University, 2015) and yields more acrylonitrile than mainstream smoke: 76e85 and 3e15 mg/cigarette, respectively. In indoor air, environ- mental tobacco smoke-related acrylonitrile concentrations were low (0.6e0.8 mg/m3) (H.W, 2007). Increased oxidative stress has been found in active and passive smokers in previous studies (Kosecik et al., 2005; Raupach et al., 2006; Zhou et al., 2000). In this study, the concentrations of CEMA were higher in smokers than non-smokers, but there was no difference between those not exposed and those exposed to environmental tobacco smoke. However, the association between 8-OHdG and CEMA was only significant in subjects exposed to environmental tobacco smoke. One possible explanation is that the number of active smokers in this study is too small to reduce the power. The other explanation is that there are synergistic effects between acrylonitrile and other chemicals in sidestream smoke on oxidative stress, but differences in the effects were not seen between acrylonitrile and mainstream smoke. Finally, it is possible that the effect of other chemicals on oxidative stress in mainstream smoke is much stronger than the effect of acrylonitrile, and the association between CEMA and 8- OHdG is too small to become statistically significant in active smokers.
In multiple regression analysis of this study, the correlation between CEMA and oxidative stress was more obvious in younger subjects, although the concentration of CEMA was higher in the older group. Recent evidence has found that oxidative stress is one of the important determinants of aging (Kregel and Zhang, 2007). It is possible that exposure to acrylonitrile may have a lesser influ- ence on oxidative stress than age. A positive association between CEMA and oxidative stress in higher HOMA-IR subjects was also found in this study. Chronic oxidative stress is particularly dangerous for b-cells, leading to insulin resistance and type 2 Diabetes Mellitus (Evans et al., 2003). There might be synergistic effects of acrylonitrile exposure and insulin resistance on oxidative stress.
There are several limitations of our study. First, we can’t conclude the causal inference due to the cross-sectional design. Second, it is possible that other components from tobacco smoke may have an influence on oxidative stress (Li et al., 2015). Third, our study population was composed of adolescents and young adults, and therefore, we cannot apply the conclusion to other populations. Finally, the activities enzymes involved in the metabolism of N-acetylcysteine are different in different individuals due to gene polymorphisms. This may influence the urine levels of CEMA (Schettgen et al., 2012) and not considered in this study.
In conclusion, we found significant associations between urine CEMA, and 8-OHdG in a Taiwanese cohort consisting of adolescents and young adults. These associations appear significant after we enrolled CVD risk factors as covariates. We found low-level acry- lonitrile exposure may increase oxidative stress. Future research on exposure to acrylonitrile and oxidative stress was warranted.