Emission-related Heavy Metal Associated with Oxidative Stress in Children: Effect of Antioxidant Intake

Heavy metals, the common pollutants emitted from industrial activities, are believed to cause harmful effects, partially through the mechanism of elevated oxidative stress, and antioxidant intake has been hypothesized to provide a potential protective effect against oxidative stress. This study aims to investigate the heavy metal exposure and the associated oxidative damage of young children living near a petrochemical complex and to assess the protective effect of antioxidant intake. There were 168 children recruited from the kindergartens near a huge petrochemical complex, with 87 as the high exposure group and 81 as the low exposure group. Urinary concentrations of eleven metals were detected by inductively coupled plasma mass spectrometry, and four biomarkers of oxidative stress were measured in urine by liquid chromatography-tandem mass spectrometry. The food frequency questionnaire was collected to assess participants’ intake of antioxidants. Multiple linear regression was performed to determine the predictors of metals for oxidative stress and to measure the beneficial effect of antioxidants. Weighted quantile sum regression was performed to determine the contributors among metals to the oxidative stress. Results showed that high exposure group had significantly higher concentrations of chromium, manganese, nickel, arsenic, strontium, cadmium, and lead when compared to those in low exposure group. There was no obviously difference on the total antioxidant intake and dietary profile between two groups. The elevated levels of two oxidative stress markers were significantly associated with most of the urinary metal concentrations in all study subjects after adjusting confounders, while no significant association was found between oxidative stress and antioxidant intake. Among the metals, mercury and strontium showed the dominated contributions for elevated levels of oxidative stress. It concluded that higher metal exposure was associated with elevated oxidative stress but with no protective effect by antioxidant intake among the young children residents near a petrochemical industry.


Introduction
Environmental pollution, mainly from various industries and motor vehicles, was understood to be a health hazard for humans, with its negative effects spanning a wide range of diseases. Heavy metals are one source of environmental pollution that are of interest to researchers due to their potential for long-term negative health effects. One mechanism through which heavy metals cause damage to human health is oxidative stress. Many heavy metals have oxidation-reduction (redox) properties, which can contribute to the generation and overproduction of reactive oxygen species (ROS) when the antioxidant defenses are not sufficient to prevent this from occurring [1].
Consequently, the overproduction of ROS leads to oxidative damage. In recent years, many studies have found exposure to heavy metals to be associated with higher oxidative stress [2]. Industrial exposure to heavy metals has consistently been found to be associated with increased oxidative stress of the populations living in areas of high exposure [3,4,5]. And, epidemiological studies have linked proximity to industries to oxidative stress and related diseases [6,7,8].
High levels of oxidative stress during childhood may be a risk factor for various adult diseases. Studies have found that asthma, obesity, hypertension, severe disability, ADHD, and acute brain damage are associated with higher oxidative stress, and these diseases often begin developing during childhood [9,10,11,12,13,14]. Some studies have focused on even earlier stages of development, examining the effect of neonatal oxidative stress on later-life diseases [15,16]. Since many later life diseases have been linked to oxidative stress that occurred during early childhood, it is important to determine the causes of oxidative stress on young children and limit potential exposures. Young children are particularly susceptible to the effects of oxidative stress and environmental exposures due to their developing nervous, immune, digestive, respiratory, antioxidant, and reproductive systems [17]. During these developmental stages, harmful environmental exposures have potential to cause irreversible, life-long damage to cells [18]. Various observational studies have examined the ways heavy metal exposure affects young children. These studies have concluded that heavy metal exposure is a consistent predictor of urinary oxidative stress among children [19,20,21]. Other studies have reached similar conclusions after examining the effects of heavy metal exposure on infants and their mothers and adolescents [22,23].
In recent years, several experimental and observational studies have aimed to address the issue about the relationship between dietary antioxidant intake and oxidative stress. As children reach preschool age, their diets begin to more similarly resemble adult diets, allowing researchers to observe differences in dietary patterns among children. Studies have shown that for all ages of people, diets low in antioxidants have been linked to increased disease, especially when coupled with exposure to heavy metals [24].
Antioxidant status is of great importance because low antioxidant intake is often found to be associated with certain diseases. Previous study examined young children's dietary antioxidant intake and exposure to environmental chemicals, concluding that the relationship between antioxidant intake, environmental exposure to chemicals, and other physiological factors interact in a complex way [25]. Other studies have linked oxidative stress and dietary antioxidant intake to diseases, with many of them concluding that dietary supplementation of antioxidants may provide a beneficial effect on childhood developmental diseases such as asthma and neurological disorders [10,11,13,26,27].
Given the complex relationship between heavy metal exposure through industrial complexes, oxidative stress, and dietary antioxidant intake, the primary objective of this study was to investigate the heavy metal exposure and oxidative stress levels among young children living in the vicinity of a big petrochemical complex, and to determine if heavy metal exposure is associated with oxidative stress among this population of young children. And, another objective was to explore if dietary antioxidant intake provided a protective effect against oxidative stress related to environmental metal exposure.

Study Area
The study area selected for this study was in Central Taiwan near a large petrochemical complex, which was built in 1998. There are 53 plants in the complex including one thermal power plant with the capacity of 1.8 million kW of electricity, three oil refineries, two naphtha cracking plants, three cogeneration plants with the generation of 2.82 million kW of electricity, and other related plants. And, the production capacity of this complex has expanded to 540,000 barrels of oil per day and 2.9 million tons of ethylene per year [28].
Previous studies have concluded that various pollutants from the complex are a possible health risk for the local residents [29,30]. Among the petrochemical emission-related pollutants, the effect of toxic metal exposure was observed on residents and environment in the vicinity areas. For ambient air, the contents of many metals in PM 10 samples were higher during the downwind season in the two townships nearby the complex [31], and the obviously increasing ambient concentrations of vanadium (V) was found in the closer areas of the complex [32]. For internal exposure biomarkers, urinary V and As levels displayed a concentration gradient in accordance with the distance-to-source gradient of V and As exposure [32], and the significant association between proximity to the petrochemical complex and heavy metals and oxidative stress biomarkers was found in teenagers and elders [6,33].

Study Subject
In this study, the study subjects selected were kindergarten children, ages 4-8, from four townships located in vicinity of the petrochemical complex. Initially, there were 104 children recruited from the two kindergartens in two townships closest to the petrochemical complex as the high exposure group, and there were 96 children recruited from the two kindergartens in two townships located farther from the petrochemical complex as the low exposure group. Geological information system (GIS) software (ArcGIS) was used to determine the distances from the petrochemical complex to the study subject's home address. A weighted average of the geographical exposure for each study subject was calculated using the number of hours the children spent at home and at school.
The informed consent of these 200 children were provided by their participants' guardians and the food frequency questionnaire (FFQ) investigation and urine sample collection were conducted at the four kindergartens. There were 32 participants excluded because of incomplete procedures, including returning both a morning urine sample and a completed food frequency questionnaire (FFQ), and therefore the total study subjects were 168 in further analysis. On the other hand, there were 24 samples with the urinary creatinine concentration below 30 mg/dL excluded from the study for urinary analysis in accordance with the World Health Organization (WHO) standards. The flowchart of this cross-sectional study was shown in Fig. 1, and the Institutional Review Board (IRB) approval (201312017RIND) was obtained prior to initiation of the study.

Analysis of Exposure Biomarkers
A morning spot urine sample was collected by the guardians of the participants and then stored in a -20 C freezer until analysis was performed. From the urine samples, levels of eleven heavy metals including vanadium (V), chromium (Cr), manganese (Mn), nickel (Ni), copper (Cu), arsenic (As), strontium (Sr), cadmium (Cd), mercury (Hg), thallium (Tl), and lead (Pb) were determined by inductively coupled plasma mass spectrometry (ICP-MS). To ensure the accurate measurements, the urinary metal levels of standard reference materials (SERO, Billingstad, Norway) analyzed by our method were all within acceptable ranges provided by the standard reference materials. And, the relative error of the ten spiked samples for each batch of the experiment was below 10% for these urinary chemicals. In addition, the measurement data was statistically analyzed when the recovery rate of each batch of the experiment higher than 85%. One half of the method detection limit was used to represent the urinary metals level for samples below the method detection limits. Urinary metal levels were adjusted using urinary creatinine concentrations, and these levels were log-transformed to fit a normal distribution for further statistical analysis.

Analysis of Antioxidant Intake
Total antioxidant intake of the participants was determined using a FFQ. FFQs are a tool used by nutritionists to determine food patterns and habits among study subjects. To calculate total antioxidant intake per week per participant, each food in the questionnaire was matched to its corresponding antioxidant intake measured by the FRAP value in mmol/100 g using data from previous studies [35,36]. Most of the foods' antioxidant content could be found through the previous two sources. Average portion sizes for children ages 4-8 were calculated using Taiwan's National Health Survey data.
Portion sizes were then multiplied by the antioxidant intake to get an approximate antioxidant intake per one-time consumption of each food. These numbers were then multiplied by the frequency with which the participant consumed a given food in a oneweek period to find the total antioxidant intake per week. This FFQ included 67 specific food groups.

Statistics
Basic characteristics and antioxidant intake between high and low exposures were compared using Student's t-tests for continuous variables. For discrete categorical variables, the Chi-squared tests were performed. After adjusting for age, gender, household smoking, and parental work history at the petrochemical plant, differences on the levels of urinary metal and oxidative stress biomarkers between the high and low exposure participants were compared using analysis of covariance (ANCOVA). Multiple linear regression analysis was performed to determine the relationship between environmental exposure biomarkers, dietary intake, and oxidative stress. All linear regression models were with antioxidant intake and metal exposure biomarkers individually as independent variables and oxidative stress biomarkers individually as dependent variables adjusting for confounding factors. Weighted quantile sum (WQS) regression was performed to determine highest contributors among metal exposure biomarkers to each oxidative stress biomarker. A p-value of < 0.05 was considered significant. All tests were performed using R Studio 3.2. WQS regression was performed using the gWQS package for R 3.5.1. Table 1 showed the basic characteristics of participants between the high and low exposure groups. Gender was significantly different between the two groups, with the high exposure and low exposure groups comprised of 57.41% males and 40.74% males, respectively. Age, household smoking, and percent of single parent households were not significantly different, but parental work history at the petrochemical plant was different between the two groups. Of the parents of high exposure participants, 33.33% were permanent workers at the plant, compared with 17.28% from the low exposure group.

Basic Characteristics
Although between-group differences in socioeconomic markers, such as income, education, and occupation, were not statistically significant, there was a slight trend of higher exposure parents having higher socioeconomic status, with higher average salaries and higher attained education than the low exposure group parents (data not shown). Table 1 Comparison of basic characteristics, urinary heavy metal exposure levels, and urinary oxidative stress biomarker levels between high and low exposure groups e High exposure group, n = 76; low exposure group, n = 68.

Exposure Status
For the overall external exposure, the high exposure group were with significantly closer distance of an average of 6.33 kilometers from the plant than the low exposure group with an average of 13.16 kilometers from the plant (

Antioxidant Intake and Nutritional Patterns
In Table 2, it showed the comparison of total antioxidant intake and the top five antioxidant ingestion food between high and low exposure groups. Total antioxidant intake per week varied between the exposure groups, with means of 43.18 and 34.20 mmol/week for the high and low exposure groups, respectively. However, this difference was not statistically significant. And, the dietary amount of the highest antioxidant intake foods were similar between the two groups with no statistically significant differences, with the exception of intake of berries per week. Table 2 The comparison of total antioxidant intake and dietary pattern between high and low exposure groups.

Effect of Antioxidant Intake on the Association between Metal
Exposure and Oxidative Stress Table 3 presented the associations between each heavy metal and each oxidative stress biomarker for all participants pooled together coupled with the coefficient of antioxidant 13 intake. It showed the obviously associations between metal exposure and oxidative stress.    Figure 2B shows that Sr was the highest contributor to 4-HNE-MA at 44.7%. The highest contributor to 8-isoPF 2α was Cu for 49.1% in Fig. 2C. And, V, Hg, and Sr were the highest contributors to 8-NO 2 Gua at 29.1, 25.0, and 22.7 percent, respectively (Fig. 2D). All four WQS regression models were statistically significant and showed positive associations between the heavy metals and the oxidative stress biomarker outcomes.

Discussion
For the petrochemical-related metal exposure, the present study indicated that the study subjects lived in the areas nearby were with significantly higher urinary metal levels except only for Cu with the reverse result (Table 1), and there were limited research revealed the internal metal exposure dose in young children like the present study. In addition, our previous studies investigated the obviously elevated urinary metal levels in all of the residents with different age groups including elders, adults, and teenagers [6,32]. And, we further conducted the distance-to-source analysis and it revealed that the increased concentrations of most urinary heavy metals in study subjects were associated with the decreased distance from the plant, with the only exception of Cu (data not shown). Among these metals, the Cr, Mn, Ni, and As showed the big different exposure levels to response the possible emission pollution by this petrochemical complex because these metals all were suggested be the key pollutants by petrochemical industry previously [37,38]. On the other hand, there was only the Cu with non-significant finding in this study, and it also indicated the accuracy of the metal exposure representative for the study subjects in this areas because of the main source of Cu exposure are from natural sources (decaying vegetation, forest fires, and sea spray) and anthropogenic emission sources (nonferrous metal production, wood production, iron and steel production) not from petrochemical industry [39,40]. In the past, the collected air samples in high exposure areas have found that the contents of many metals in PM 10 were higher during the downwind season to provide the external metal exposure from the petrochemical complex [31]. Nevertheless, the findings in the present study implied the emission-related metal exposure existed even in the children with the kindergarten age, and it should pay more attention on the potential adverse health effects of children in this polluted area in the future.
In previous study conducted in this study area, it found that teenagers and elders who lived in the high exposure areas were with significantly higher levels of urinary oxidative stress markers [6]. However, the present study did not show any obvious differences in the oxidative stress marker levels of young children between high and low exposure groups (Table 1). One possible reason is due to the different exposure definitions of these two studies. The participants of previous study were selected from the extreme highest and lowest exposure status with about a two-fold difference in most of the urinary metal levels, but the participants in high and low exposure in the present study were only with relatively slight differences in urinary metal levels. Nevertheless, it showed the significant associations between urinary levels of metals and oxidative stress markers for all subjects in this study (Table 3). Among these four oxidative stress markers, we found 8-OHdG and 4-HNE-MA were more sensitive to the metal exposure, and the past studies indicated the consistent findings for the application of these two markers in the prediction of the oxidative stress caused by metal exposure [6,41]. In addition, the levels of these two markers in young children in the present study, even with lower urinary metal levels, were obviously higher than those in the teenagers and elders in past study [6]. And, previous studies supported that the inverse age-oxidative stress relationship could due to the naturally low glutathione levels at children which means the ability to detoxify reactive oxygen is limited making the younger children more susceptible to oxidative stress [42].
On the other hand, the Hg and Sr showed the dominated contributions for most of the oxidative stress markers in this study (Fig. 2). These two metals were main petrochemicalrelated emission pollutants [37,38], and several studies have confirmed their effects on the increasing of oxidative stress [43,44,45]. According to the above-mentioned results, young children might be with relatively higher oxidative stress when exposure to metal emission from the petrochemical industry, especially for Hg and Sr, and the subsequent effects of oxidative stress in these young children require further research to clarify.
The differences in dietary antioxidant intake varied slightly between the two groups in the present study, although the high exposure group having a higher average total antioxidant intake (Table 1). Currently, only a few studies have suggested dietary recommendations for vitamin C (the best-known antioxidant), but it has not been clearly defined and it was limited for adults [46,47]. In fact, there was no well-established dietary recommendation for intake of antioxidants per week, especially for children. Limited studies have measured the intake of antioxidants in children populations. One Swedish study analyzed the associations between antioxidant intake and allergic disease on 8-year-old children by applying a food-frequency questionnaire. The result found that the intake of antioxidant, like β-carotene and magnesium in food, had inverse association with allergic disease such as rhinitis, atopic sensitization, and asthma [48]. However, most of the previous studies just estimated the single antioxidant not for the total intake of antioxidants in food. On the other hand, there was recently no clear definition on the amount of antioxidant intake enough to achieve the obvious antioxidant effects because of the difficulty to define the level of significant antioxidant protective effect. Therefore, it might be one important reason that we did not observe any significant association between antioxidant intake and oxidative stress (Table 3), even with no contribution for the oxidative stress levels when conducting the WQS model including the antioxidant intake levels (data not shown).
On the other hand, previous research has shown that foods with primarily higher antioxidant intake including fruits and vegetables (particularly strawberries, citrus, kiwi), soybeans, nuts, spices, herbs, yam, mackerel, and so on [35,49]. Among them, soybeans, nuts, strawberries, and kiwi are the more common food in Taiwan Availability of data and materials: The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests:
The authors declare they have no actual or potential competing financial interests.  Taiwan (grant number NTU-107L9003 and NTU-108L9003).
Authors' contributions: BK collected the questionnaire, detected the urinary metal levels, analyzed the study results, and wrote the main part of whole manuscript. TH was the key person to assist and revise the whole process for this study. CH contacted the kindergartens and recruited the study subjects. Tina HT designed the dietary questionnaire and estimated the antioxidant intake levels. YH assisted the writing and revision of the manuscript. CC was the PI for this study project, and managed the whole process of this study.  Figure 1 The study design and flow chart. The study design and flow chart.