4. Results and Discussion
4.2 Part 2. Metabolomics of Children and Adolescents Exposed to Industrial
Based on the age-dependent findings from our Part 1 study, we decided further metabolomics studies for the two age groups should be conducted separately. At the same time, Yuan et al. and Chen et al. reported increased risk of all cancers for residents living near No. 6 Naphtha Cracking Complex (Chen et al. 2018; Yuan et al. 2018). Since children exposed to carcinogenic pollutants at critical periods of development have more time for chronic adverse health effects such as cancer to manifest, we decided to first focus on our children participants in order to identify the association between industrial carcinogenic pollutants exposure, serum metabolic changes, and cancer-related early health effects including oxidative stress and serum acylcarnitines. We used serum samples in our Part 2 study because blood circulates the body covering every tissue and organ, carrying all the molecules that are secreted, excreted, or discarded in response to physiological needs and stress, and alterations in blood metabolite profile could reflect pathological states and the body’s attempt to maintain homeostasis (Psychogios et al.
2011).
4.2.1 Results
Table 8 showed the comparison of basic characteristics, external and internal exposure levels, and urine oxidative stress biomarker levels between high and low exposure groups of the 107 children participants with serum samples available for metabolomics analysis. High exposure group lived 10.57±2.52 km and 10.02±2.73 km away from the main emission points of coal-fired power plant and oil refineries, respectively, while low exposure group lived 21.81±5.71 km and 20.91±5.44 km away,
51
respectively. High and low exposure group showed no significant difference in age, sex distribution, systolic blood pressure (SBP), smoking history, alcohol history, and betelnut history. However, high exposure group had higher BMI compared to low exposure group.
Road area surrounding homes showed no significant difference between the two exposure groups at neither 25 m or 500 m buffer. Urine concentrations of exposure biomarkers As, Cd, Cr, Ni, 1-OHP, V, and Hg were increased in high exposure group compared to low exposure group, with As, Cd, 1-OHP, V, and Hg reaching statistical significance. The difference was most profound in 1-OHP and V. Urine concentrations of oxidative stress biomarkers showed all four biomarkers were increased in high exposure group compared to low exposure group, but the differences were more statistically significant for lipid peroxidation biomarkers HNE-MA (p=0.006) and 8-isoPF2α (p=0.09) than DNA damage biomarkers 8-OHDG (p=0.21) and 8-NO2Gua (p=0.11). Only 99 out of the 107 subjects had available data for urine oxidative stress biomarkers concentrations.
Table 8. Comparison of basic characteristics, carcinogens exposure levels, and oxidative stress biomarker levels in 107 study subjects.
High (n=37) Low (n=70) pa
Basic characteristics
Age, mean±SD 13.67 ±0.92 13.70 ±0.90 0.84
Male, n(%) 20 (54.1) 38 (54.3) 0.98
Systolic blood pressure (SBP), mean±SD 118.7 ±12.96 116.2 ±13.73 0.37
Body Mass Index (BMI), mean±SD 21.67 ±3.41 20.15 ±3.48 0.04
Smoke history, n(%) 5 (13.5) 5 (7.1) 0.28
Drink history, n(%) 5 (13.5) 3 (4.3) 0.12
Betelnut history, n(%) 1 (2.7) 3 (4.4) 1.00
External exposuresb, mean±SD
Distance to coal-fired power plant 10.57 ±2.52 21.81 ±5.71 <0.0001
Distance to oil refinery 10.02 ±2.73 20.91 ±5.44 <0.0001
Road area surrounding homes
25 m buffer 304.1 ±211.4 329.4 ±204.7 0.58
500 m buffer 70938.4 ±26594.8 64120.1 ±20016.4 0.18
Internal exposuresc, mean±SD
Group 1 carcinogen
Arsenic 60.27 ±42.16 39.62 ±30.18 0.01
Cadmium 0.34 ±0.34 0.19 ±0.15 0.02
Chromium 3.24 ±2.96 2.14 ±1.63 0.10
Nickel 6.69 ±8.72 3.89 ±2.85 0.31
Group 2A carcinogen
Lead 0.64 ±0.64 0.66 ±0.65 0.80
1-OHP 0.19 ±0.14 0.03 ±0.01 <0.0001
Group 2B carcinogen
Vanadium 2.46 ±1.64 0.24 ±0.10 <0.0001
Mercury 3.13 ±2.89 1.86 ±1.88 0.04
Oxidative stressd, mean±SD
8-OHDG 3.01 ±2.15 2.63 ±2.86 0.21
HNE-MA 2.00 ±2.58 1.30 ±2.20 0.006
8-isoPF2α 3.27 ±3.54 2.06 ±2.18 0.09
8-NO2Gua 6.49 ±11.77 2.54 ±3.05 0.11
a Comparison of basic characteristics between the high and low exposure groups for continuous variables was performed using Student’s t-test, and for discrete variables, Chi-squared test or Fisher’s exact test. Urinary exposure biomarker concentrations are log-transformed, high and low exposure groups compared by ANCOVA test adjusting
53
age, sex, smoking, alcohol consumption, betel nut chewing, fish consumption, and BMI with a post comparison by Scheffe test. Urinary oxidative stress biomarker concentrations are log-transformed, high and low exposure groups compared by Student’s t-test.
b Distance to source: Average of home-to-coal-fired power plant and home-to-oil refinery distance, unit: km; Road area surrounding homes unit: m2.
c For urine 1-OHP, unit: µmol/mol-creatinine; for urine heavy metals, unit: µg/g-creatinine
d Urine oxidative stress biomarkers unit: µg/g-creatinine. High exposure group N=34, low exposure group N=65.
Figure 8 showed serum levels of six acylcarnitines that were significantly different in high exposure group compared to low exposure group in 107 study subjects. Samples are in columns and arranged according to high exposure (red) and low exposure (green) groups. Acylcarnitines are in rows and were arranged according to hierarchical clustering using Euclidean distance measure and Ward algorithm. The colors vary from deep blue to dark brown to indicate data values change from down-regulation (blue) to up-regulation (brown). We found long-chain acylcarnitines were clustered together and down-regulated in high exposure group compared to low exposure group (Dodecanoylcarnitine, C12; Tetradecanoylcarnitine, C14; Tetradecenoylcarnitine, C14:1;
Hexadecenoylcarnitine, C16:1; Pristanoylcarnitine, C19), while short-chain acylcarnitine (Hexanoylcarnitine, C6) was up-regulated in high exposure group compared to low exposure group.
Figure 8. Heat map of serum acylcarnitine levels in 107 study subjects
PCA showed clear clustering of pooled QC on the score plots shown in Figure 9.
Metabolomics identified 84 potential metabolite features in study subjects serum samples after removing features missing in more than 50% of the samples, 80 of which had available HMDB ID number as shown in Appendix 2. 84 potential metabolite features were put through PLS-DA analysis, and results showed metabolic profiles between high and low exposure groups could be significantly separated by two components that accounted for 5.8 % and 9.0 % of variability of metabolic profiles between high and low exposure groups, respectively (Accuracy=0.78, R2=0.53, Q2=0.23) (Figure 10A).
Permutation test was performed to confirm the validity of DA model (p=0.01). PLS-DA and ANCOVA analysis adjusting for age, sex, and BMI found 11 exposure-related potential metabolite features (average variable importance in projection (VIP) score >1, ANCOVA p <0.05), which through in house library search was identified as 10 potential metabolites. Two potential metabolites, one detected under positive mode and one under negative mode of UHPLC-qTOFMS analysis, were both identified as pyroglutamic acid.
Figure 10B showed the up- and down-regulation of exposure-related potential metabolites in high and low exposure groups. Samples are in columns arranged according to high exposure (red) and low exposure (green) groups. Potential metabolites are in rows and were arranged according to hierarchical clustering using Euclidean distance measure and
55
Ward algorithm. The colors vary from deep blue to dark brown to indicate data values change from down-regulation (blue) to up-regulation (brown). We found potential metabolites up-regulated in high exposure group compared to low exposure group were clustered together, including ketoleucine, carnitine, isovalerylcarnitine, aspartic acid, and octenoyl-L-carnitine, while down-regulated potential metabolites were also clustered together, including pyroglutamic acid, adenosine monophosphate (AMP), inosinic acid (inosine monophosphate, IMP), oxoglutaric acid, and malic acid (Figure 10B). Pathway analysis results showed purine metabolism was the main biological pathway affected by multiple exposures (p < 0.05, Impact > 0.1) (data not shown). We identified two exposure-related potential metabolites involved in purine metabolism, nucleotides AMP and IMP (Figure 10B) (Simoni et al. 2007). Through the Comparative Toxicogenomics Database (CTD), we found that group 1 carcinogens As, Cd, Cr, and Ni were significantly associated with purine metabolism pathway (Bonferroni adjusted p < 0.01).
Figure 9. Principle component analysis score plot of 61 pooled quality control (QC) samples data from 11 batches detected under (A) negative mode and (B) positive mode of UHPLC-qTOFMS metabolomics analysis
Figure 10. Comparison of serum metabolic profile in 107 study subjects using (A)
PLS-DA score plot (Accuracy=0.78, R
2=0.53, Q
2=0.23, Permutation p=0.01) and
(B) heat map of exposure-related potential metabolite levels (average VIP score >1,
ANOVA p <0.05 are shown adjusted for sex, age, and BMI)
57
Table 9 showed the association between eight individual carcinogens with four oxidative stress biomarkers in 99 study subjects. Individually, amongst four group 1 carcinogens, urinary levels of Cd was positively associated with urinary concentrations of 8-NO2Gua (p=0.029). The two group 2A carcinogens was not associated with any of the four oxidative stress biomarkers. For the two group 2B carcinogens, V was associated with HNE-MA (p=0.003), 8-isoPF2α (p=0.023), and 8-NO2Gua (p=0.004), and Hg was associated with 8-OHDG (p=0.008) and HNE-MA (p=0.012). Figure 11A to 11D showed the WQS regression analysis of the association of combined eight carcinogens exposure with four oxidative stress biomarkers, respectively. Association with all four oxidative stress biomarkers were positive and statistically significant with 8-OHDG, HNE-MA, and 8-NO2Gua, while association with 8-isoPF2α was borderline significant. For 8-OHDG, group 2B carcinogen Hg predominated in the mixture index (49.7%) and group 1 carcinogens Ni, As, and Cd also contributed to the association (p=0.002) (Figure 11A).
Figure 11B showed group 2B carcinogens Hg (43.3%) and V (31.1%) contributed to over half of the mixture index positively associated with HNE-MA levels (p=0.0006), and group 1 carcinogens As, Cd, Ni, and Cr also showed contribution. Associations with 8-isoPF2α was predominated by group 2B carcinogens Hg (36.9%) and V (31.6%), followed by group 1 carcinogen Ni (19.4%), with contribution from Cr and As (p=0.08) (Figure 11C). In Figure 11D, we can see in the mixture index positively associated with 8-NO2Gua (p=0.0001), group 2B carcinogen V contributed to half of the association (50.1%), followed by group 1 carcinogen Ni (32.0%), with contributions from Cd, Cr, and As.
Table 9. Individual association between urine carcinogens and oxidative stress biomarkers in 99 study subjects.
Linear regression analysis adjusted for sex, age, and BMI
8-OHDG HNE-MA 8-isoPF2α 8-NO2Gua
Estimate 95% CI p value Estimate 95% CI p value Estimate 95% CI p value Estimate 95% CI p value
Group 1 carcinogen
Arsenic 0.049 (-0.165, 0.263) 0.653 0.092 (-0.163, 0.347) 0.476 -0.096 (-0.321, 0.130) 0.402 0.016 (-0.324, 0.356) 0.925 Cadmium -0.025 (-0.252, 0.202) 0.831 0.097 (-0.173, 0.366) 0.480 -0.141 (-0.379, 0.097) 0.243 0.391 (0.040, 0.743) 0.029 Chromium -0.186 (-0.380, 0.007) 0.059 -0.016 (-0.251, 0.219) 0.892 -0.061 (-0.269, 0.147) 0.562 0.120 (-0.192, 0.432) 0.448 Nickel 0.018 (-0.172, 0.209) 0.848 0.033 (-0.194, 0.260) 0.775 0.126 (-0.073, 0.326) 0.213 0.224 (-0.075, 0.523) 0.140 Group 2A carcinogen
Lead 0.115 (-0.112, 0.341) 0.318 0.158 (-0.112, 0.429) 0.249 -0.135 (-0.374, 0.104) 0.265 -0.033 (-0.397, 0.331) 0.858 1-OHP 0.094 (-0.057, 0.244) 0.219 0.146 (-0.032, 0.325) 0.107 0.073 (-0.087, 0.232) 0.369 0.188 (-0.050, 0.425) 0.121 Group 2B carcinogen
Vanadium 0.132 (-0.013, 0.276) 0.073 0.254 (0.087, 0.422) 0.003 0.176 (0.025, 0.327) 0.023 0.330 (0.107, 0.554) 0.004 Mercury 0.246 (0.064, 0.427) 0.008 0.279 (0.062, 0.496) 0.012 0.119 (-0.078, 0.316) 0.233 0.005 (-0.293, 0.304) 0.973
59
Figure 11. Combined associations between internal exposure levels and (A) 8-OHDG (p=0.002), (B) HNE-MA (p=0.0006), (C) 8-isoPF
2α(p= 0.08), and (D) 8-NO
2Gua (p=0.0001) levels based on weighted quantile sum (WQS) regression analysis in 99 study subjects. (Adjusted for sex, age, and BMI)
“Meet-in-the middle” approach identified eight potential metabolites that were both carcinogens exposure-related and associated with biomarkers of early health effects.
Table 10 and 11 showed the level of association between carcinogens exposure-related potential metabolites (in rows) and biomarkers of early health effects (in columns). For oxidative stress biomarkers, 8-OHDG was significantly associated with pyroglutamic acid and inosinic acid, HNE-MA was significantly associated with ketoleucine, octenoyl-L-carnitine, pyroglutamic acid, AMP, and IMP, 8-isoPF2α was significantly associated with octenoyl-L-carnitine, and 8-NO2Gua was not significantly associated with any exposure-related potential metabolites (Table 10). Long-chain acylcarnitines C14 and C19 were associated with the most number of exposure-related potential metabolites, five for C14 (carnitine, octenoyl-L-carnitine, pyroglutamic acid detected in both positive and negative modes, IMP) and five for C19 (carnitine, octenoyl-L-carnitine, pyroglutamic acid, AMP, IMP). C16:1 was associated with carnitine and octenoyl-L-carnitine. Short-chain acylcarnitine C6 was associated with four exposure-related potential metabolites
including ketoleucine, isovalerylcarnitine, and pyroglutamic acid detected in both positive and negative modes (Table 11). Overall, for the exposure-related potential metabolites, octenoyl-L-carnitine and pyroglutamic acid were associated with the most number of biomarkers of early health effects. Octenoyl-L-carnitine was associated with two oxidative stress biomarkers and three long-chain acylcarnitines, and pyroglutamic acid was associated with two oxidative stress biomarkers, short-chain acylcarnitine, and two long-chain acylcarnitines. Aspartic acid, oxoglutaric acid, and malic acid were not significantly associated with any of the biomarkers of early health effects.
61
Table 10. Association between exposure-related potential metabolites and oxidative stress biomarkers in 99 study subjects
(Estimates of linear regression analysis are shown. 95% CI are in brackets. *p<0.05 †p<0.01 ‡p<0.001).
8-OHDG HNE-MA 8-isoPF2α 8-NO2Gua
Ketoleucine 0.03 0.10* 0.05 0.09
(-0.05, 0.11) (2.19E-03, 0.19) (-0.04, 0.13) (-0.03, 0.22)
Carnitine 2.79E-03 0.05 0.01 -0.02
(-0.08, 0.09) (-0.05, 0.14) (-0.08, 0.10) (-0.15, 0.11)
Isovalerylcarnitine -0.05 -0.01 0.03 0.03
(-0.13, 0.03) (-0.11, 0.09) (-0.05, 0.12) (-0.10, 0.16)
Aspartic acid -0.03 -0.03 2.84E-04 0.11
(-0.11, 0.04) (-0.12, 0.07) (-0.08, 0.08) (-0.01, 0.24)
Octenoyl-L-carnitine 0.05 0.09* 0.08* 0.11
(-0.03, 0.13) (2.00E-03, 0.18) (4.28E-03, 0.17) (-0.01, 0.23)
Pyroglutamic acida -0.04 -0.09 0.02 -0.07
(-0.12, 0.04) (-0.19, 1.89E-03) (-0.07, 0.10) (-0.20, 0.05)
Pyroglutamic acidb -0.08* -0.09* 0.01 -0.04
(-0.16, -3.36E-03) (-0.19, -3.48E-04) (-0.07, 0.09) (-0.16, 0.09)
Adenosine Monophosphate -0.05 -0.11* 0.02 -0.11
(-0.13, 0.03) (-0.20, -0.02) (-0.07, 0.10) (-0.24, 0.01)
Inosinic acid -0.08* -0.13† -0.04 -0.08
(-0.16, -9.14E-04) (-0.23, -0.04) (-0.12, 0.05) (-0.21, 0.05)
Oxoglutaric acid -0.04 -0.01 1.48E-03 -0.09
(-0.12, 0.04) (-0.11, 0.08) (-0.08, 0.09) (-0.22, 0.03)
Malic Acid -0.01 -0.03 0.05 0.01
(-0.09, 0.07) (-0.12, 0.07) (-0.03, 0.14) (-0.11, 0.14)
Linear regression model adjusted for age, sex, and BMI.
Pyroglutamic acid was detected in both a negative and b positive mode.
Table 11. Association between exposure-related potential metabolites and acylcarnitines in 107 study subjects
(Estimates of linear regression analysis are shown. 95% CI are in brackets. *p<0.05 †p<0.01 ‡p<0.001).C6 b C14:1 b C12 b C19 b C14 b C16:1 b
Ketoleucine 0.31 (0.12, 0.49)† -0.01 (-0.20, 0.17) 0.01 (-0.18, 0.19) -0.04 (-0.23, 0.15) -0.04 (-0.23, 0.15) 4.66E-03 (-0.19, 0.20)
Carnitine 0.11 (-0.09, 0.31) -0.13 (-0.32, 0.05) -0.16 (-0.34, 0.02) -0.19 (-0.38, -3.39E-03)* -0.25 (-0.43, -0.06)* -0.23 (-0.42, -0.04)*
Isovalerylcarnitine 0.20 (1.39E-03, 0.40)* 0.08 (-0.11, 0.28) 0.02 (-0.17, 0.21) -0.09 (-0.29, 0.10) -0.14 (-0.34, 0.05) -0.09 (-0.29, 0.11)
Aspartic acid 0.04 (-0.15, 0.24) -0.09 (-0.27, 0.09) -0.14 (-0.31, 0.04) -0.02 (-0.21, 0.16) -0.10 (-0.28, 0.11) -0.08 (-0.27, 0.11)
Octenoyl-L-carnitine 0.13 (-0.06, 0.32) -0.12 (-0.30, 0.06) -0.12 (-0.30, 0.05) -0.32 (-0.49, -0.14)‡ -0.21 (-0.39, -0.03)* -0.19 (-0.38, -0.01)*
Pyroglutamic acida -0.39 (-0.57, -0.32)‡ 0.17 (-0.01, 0.36) 0.13 (-0.05, 0.32) 0.27 (0.09, 0.46)† 0.21 (0.02, 0.39)* 0.19 (-0.01, 0.38)
Pyroglutamic acidb -0.32 (-0.50, -0.13)† 0.13 (-0.05, 0.31) 0.07 (-0.11, 0.25) 0.17 (-0.01, 0.35) 0.20 (0.02, 0.39)* 0.14 (-0.05, 0.33)
Adenosine Monophosphate 0.03 (-0.17, 0.22) 0.08 (-0.10, 0.27) 0.15 (-0.02, 0.33) 0.34 (0.17, 0.52)‡ 0.18 (-2.75E-03, 0.36) 0.09 (-0.10, 0.28)
Inosinic acid -0.06 (-0.25, 0.13) 0.01 (-0.17, 0.19) 0.05 (-0.13, 0.23) 0.48 (0.31, 0.64)‡ 0.20 (0.02, 0.38)* 0.05 (-0.15, 0.24)
Oxoglutaric acid 0.05 (-0.14, 0.25) 0.13 (-0.05, 0.32) 0.08 (-0.10, 0.26) 0.13 (-0.06, 0.31) 0.09 (-0.09, 0.28) 0.12 (-0.07, 0.31)
Malic Acid 0.06 (-0.14, 0.25) 0.14 (-0.04, 0.32) 0.14 (-0.04, 0.32) 0.17 (-0.01, 0.36) 0.15 (-0.03, 0.34) 0.08 (-0.11, 0.27)
Linear regression model adjusted for age, sex, and BMI.
Pyroglutamic acid was detected in both a negative and b positive mode.
63
4.2.2 Discussion
Previous studies have reported exposure to individual carcinogens As, Cd, Cr, Ni, Pb, PAHs, V, and Hg can induce oxidative stress through production of reactive radicals and/or depletion of anti-oxidants (Fu et al. 2012; Jomova and Valko 2011; Valko et al.
2005). However, these studies mostly focused on the association between single carcinogen exposure and oxidative stress, and only occupational exposure studies in adults reported the association between multiple heavy metals exposure and oxidative stress (Ko et al. 2017). Our subjects were exposed to multiple carcinogens and therefore it was difficult to find one-to-one association between specific carcinogens and oxidative stress. The level of which each carcinogen induced oxidative stress may also vary, especially in a mixture. The strength of our study is that we applied WQS regression analysis and showed in children and adolescents, exposure to a mixture of eight environmental carcinogens was positively associated with four oxidative stress biomarkers, and both group 1 and group 2 carcinogens contributed to this association.
Our study is the first to report multiple carcinogens exposure could be associated with alterations in serum acylcarnitine levels in children and adolescents. Our findings support a previous study of adults occupationally exposed to metal-containing welding fumes who had significant decrease in short- and long-chain acylcarnitines (Shen et al.
2018). Previous studies have suggested acylcarnitines to be suitable candidates for cancer diagnosis (Ni et al. 2016). Interestingly, the up- and down-regulation of short- and long-chain acylcarnitines vary by cancer types and studies. Ni et al. reported both short-long-chain and long-chain acylcarnitines were significantly increased in lung cancer patients compared to healthy control subjects (Ni et al. 2016). Another study showed significant
decrease of short-chain acylcarnitines in early stage non-small-cell lung cancer patients (Klupczynska et al. 2017). In hepatocellular carcinoma patients, short-chain acylcarnitines were decreased and long-chain acylcarnitines were increased compared to control subjects (Chen et al. 2013; Zhou et al. 2012b). We identified serum acylcarnitine deregulations in children and adolescents exposed to multiple carcinogens that has been reported in cancer patients, which may imply the possibility of increased cancer risk.
Deregulation in purine metabolism has been associated with early stage cancer development and cancer progression (Bester et al. 2011). Purine metabolism is involved in energy production and signal transduction, and the enzymes and metabolites from this pathway can mediate oxidative stress through reactive species and anti-oxidant productions (Cantu-Medellin and Kelley 2013; Maiuolo et al. 2016; Pedley and Benkovic 2017). Our findings suggest multiple carcinogens exposure can induce perturbations in purine metabolism and link to increased oxidative stress and altered serum acylcarnitine levels.
Multiple carcinogens exposure were also associated with several potential metabolites in this study which could not be summarized in pathway analysis, but are involved in important biological mechanisms and have been reported in cancer studies.
These potential metabolites included aspartic acid, an amino acid that has been reported to be involved in oxidative stress regulations (Sivakumar et al. 2011). Carnitine and citrate cycle-related metabolites malic acid and oxoglutaric acid were also identified, and carnitine was associated with acylcarnitines. Carnitine cooperates with acylcarnitines transporting fatty acids into mitochondria for β-oxidation, forming acetyl-CoA that enters the citrate cycle (Semba et al. 2017). These findings suggest multiple carcinogens exposure in children and adolescents may affect fatty acid oxidation and energy
65
production mechanisms leading to deregulation of acylcarnitines. Multiple carcinogens exposure in children and adolescents also affected pyroglutamic acid, an intermediate metabolite of anti-oxidant glutathione, and was linked to oxidative stress biomarkers and acylcarnitines (Kumar and Bachhawat 2012). Interestingly, four of the exposure-related potential metabolites we identified also showed similar patterns of alteration in early stage non-small cell carcinoma patients, including increased serum aspartic acid and carnitine, and decreased serum malic acid and pyroglutamic acid (Klupczynska et al. 2016a;
Klupczynska et al. 2016b; Klupczynska et al. 2017). Our findings suggest multiple carcinogens exposure may have diverse effects on children and adolescents, causing disruptions in various biological mechanisms such as fatty acid oxidation, energy production, oxidative stress, and amino acid metabolism.
In this study, we found children and adolescents living near a petrochemical complex had increased exposure to multiple carcinogens which induced metabolic changes associated with early health effects including increased oxidative stress and altered serum acylcarnitines, both of which may lead to increased cancer risk. Our findings may provide an explanation for increased cancer incidence among adult residents living near the same petrochemical complex reported in previous studies (Chen et al. 2018; Yuan et al. 2018).
There are limitations to this study. Firstly, we analyzed metabolomics using single analytical platform which limited the number of potential metabolite features detected, and cannot provide a comprehensive view of the metabolome. Secondly, we applied in-house library match using m/z for metabolite identification and therefore could not rule out the possibility of inaccurate metabolite identification and could not provide exact quantification of potential metabolites. Thirdly, our sample size was limited, which could possibly explain why three of the four oxidative stress biomarkers were increased but did
not reach statistically significant difference between high and low exposure groups. Lastly, this is a cross-sectional study using single urine and serum samples, and therefore we could not confirm biomarker stability and could not be certain if the potential metabolites we identified can serve as life-long indicators of increased cancer risk.
Our findings imply multiple carcinogens exposure during critical periods of childhood and adolescence development induce metabolic perturbations in children and adolescents linking to early health effects that may contribute to cancer risk later in life.
This indicates significant reduction of toxic emissions from the complex could decrease carcinogens exposure and metabolic abnormities, which may potentially reduce cancer risks in children and adolescents living nearby. We recommend longitudinal epidemiological studies in this area to follow up on children and adolescents’ health if carcinogens emission continues in the near future.
67