3.1 Occurrences of NEOs and their metabolites in the blood
The distribution of the levels (median, GM, mean, minimum, and maximum) of p-NEOs and m-NEOs in blood samples were presented in Table 1. NEOs were widely detected in blood samples collected from 100 healthy individuals from South China, and the detection rates ranged from 61–94% (Table 1). Among the p-NEOs, DIN had the highest median blood concentration (0.62 ng/mL), followed by CLO (0.31 ng/mL), ACE (0.29 ng/mL), and THI (0.29 ng/mL; Table 1). Moreover, DIN was the predominant p-NEOs, accounting for 32.4% (Fig. 1a). Our results indicated that the population collected for this study had a high concentration of exposure to DIN, CLO, ACE, and THI. A similar pattern was also found in previous study, which suggested that high level of exposure to CLO, ACE, and DIN was observed in Chinese population (Xu et al. 2021). Among the m-NEOs, Of-IMI (1.28 ng/mL) had the highest median concentration in blood samples, followed by ACE-dm (0.79 ng/mL) and UF (0.73 ng/mL; Table 1). In addition, Of-IMI was the most abundant m-NEOs (37.6%; Fig. 1a). Meanwhile, Of-IMI was dominant NEOs in blood samples, accounting for 24.1%, followed by ACE-dm (14.9%), UF (13.7%), DIN (11.7%), and 5-OH-IMI (6.85%). To our knowledge, this study is the first work to report the concentrations of DN and 5-OH-IMI in blood samples.
NEOs and their metabolites were also commonly measured in 274 liver cancer patients, and the detection frequencies ranged from 61–91% (Table 1). The median concentrations of p-NEOs in blood samples from the liver cancer population were in the following order: DIN (0.69 ng/mL), IMI (0.56 ng/mL), ACE (0.49 ng/mL), and CLO (0.47 ng/mL; Table 1). This pattern was similar to those found in the healthy population. And consistent with the results detected in the healthy population, Of-IMI had the highest median blood level in the liver cancer population (2.03 ng/mL, accounting for 20.7%), followed by UF (1.88 ng/mL, 19.2%) and ACE-dm (1.61 ng/mL, 16.4%; Table 1 and Fig. 1b). These findings suggested that the concentrations of m-NEOs (i.e., Of-IMI, 5-OH-IMI, ACE-dm, and UF) in blood samples were relatively higher than those of their corresponding p-NEOs (i.e., IMI, ACE, and DIN; Table 1 and Fig. 1). A similar trend was also found in previous studies reported in blood (Xu et al. 2021) and serum samples (Zhang et al. 2021c). Previous studies widely measured m-NEOs in foodstuffs (Chen et al. 2020; D. Li et al. 2020) and environmental matrices (Wan et al. 2019; Wan et al. 2020; Wang et al. 2019; Zhang et al. 2021a). In addition, NEOs were easily metabolized through phase I metabolic biotransformation in mammal (Casida 2011; Ford and Casida 2006a, b). This might be the reason for the high m-NEO concentrations found in blood samples from both populations. Interestingly, the median blood concentrations of ΣNEOs/IMIeq in the liver cancer population (23.0/44.4 ng/mL) were significantly higher (p < 0.05) than those in the healthy population (7.75/23.9 ng/mL; Table 1). Furthermore, the median level of Σm-NEOs in the liver cancer population (12.3 ng/mL) was approximately fourfold higher than that in the healthy population (3.69 ng/mL; Table 1). Our findings indicated that the liver cancer population was highly exposed to NEOs and their metabolites.
Limited studies have reported the concentrations of NEOs and their metabolites in blood samples (Fuke et al. 2014; Proença et al. 2005; Xu et al. 2021; Yeter and Aydın 2014). The median blood concentrations of NEOs and their metabolites in the liver cancer population in this study were at least onefold higher than those reported in South China (Table 2) (Xu et al. 2021). Moreover, the median levels of most NEOs and their metabolites (except for IMI and UF) in blood samples collected from the healthy population were considerably higher than those reported by Xu et al. (2021) (Table 2). With the comparison of previous studies reported on serum samples, the median blood concentrations of ACE (0.49 and 0.29 ng/mL), THI (0.21 and 0.29 ng/mL), and THIX (0.20 and 0.19 ng/mL) in this study are relatively higher than those reported in South China (ACE: 0.04 and 0.05 ng/mL; THI: 0.06 and 0.01 ng/mL; THIX: 0.16 and 0.16 ng/mL) (Zhang et al., 2021), East China (ACE: < 0.005 ng/mL; THI: < 0.018 ng/mL; THIX: < 0.023 ng/mL) (Chen et al. 2021), and Saudi Arabia (ACE: < 0.002 ng/mL; THI: not detected; THIX: < 0.061 ng/mL; Table 2) (Li et al. 2020a). However, the median blood levels of UF (1.88 and 0.73 ng/mL) and Of-IMI (2.03 and 1.28 ng/mL) were comparable to those reported in South China (UF: 1.27 and 0.73 ng/mL; Of-IMI: 3.08 and 1.41 ng/mL), but the median concentration of ACE-dm (1.61 and 0.79 ng/mL) was significantly lower (5.99 and 2.02 ng/mL; Table 2) (Zhang et al. 2021c). These differences might be explained by the different sample size, NEO usage pattern, human bodily fluids, such as blood and serum, and populations.
Table 2
Summary of median concentration (ng/mL) of NEOs and their metabolites in blood samples from this study with those reported for other countries.
countries | sampling date | age | n a | ACE | CLO | DIN | IMI | THI | THIX | DN | UF | ACE-dm | 5-OH-IMI | Of-IMI | references |
China | 2018–2019 | 11–88 | 274 | 0.49 | 0.47 | 0.69 | 0.56 | 0.21 | 0.20 | 0.87 | 1.88 | 1.61 | 0.79 | 2.03 | This study |
China | 2018–2019 | 22–91 | 100 | 0.29 | 0.31 | 0.62 | 0.21 | 0.29 | 0.19 | 0.24 | 0.73 | 0.79 | 0.36 | 1.28 | This study |
China | 2018 | 20–27 | 196 | 0.13 | 0.22 | 0.12 | 0.29 | 0.08 | 0.08 | NR b | 0.8 | 0.58 | NR | 0.78 | (Xu et al., 2021) |
China d | 2019 | 52–89 | 120 | 0.04 | 1.40 | 0.68 | 1.00 | 0.06 | 0.16 | NR | 1.27 | 5.99 | NR | 3.08 | (Zhang et al., 2021) |
China d | 2019 | 50–89 | 80 | 0.05 | 0.71 | 0.24 | 0.76 | 0.01 | 0.16 | NR | 0.73 | 2.02 | NR | 1.41 | (Zhang et al., 2021) |
China d | NA c | 9–80 | 120 | < 0.005 | < 0.016 | < 0.035 | < 0.018 | < 0.006 | < 0.023 | NR | NR | NR | NR | NR | (Chen et al., 2021) |
Saudi Arabia d | 2017 | 41–78 | 25 | < 0.002 | < 0.025 | < 0.035 | 0.04 | ND e | < 0.061 | NR | NR | 0.03 | NR | NR | (Li et al., 2020) |
Turkey | NA | 7–29 | 3 | 2700 | NR | NR | NR | NR | NR | NR | NR | NR | NR | NR | (Yeter and Aydın, 2014) |
Japan | NA | 70 | 1 | NR | NR | NR | 105,000 | NR | NR | NR | NR | NR | NR | NR | (Fuke et al., 2014) |
Sri Lanka f | 2002–2007 | < 14 | 68 | NR | NR | NR | 10.6 | NR | NR | NR | NR | NR | NR | NR | (Mohamed et al., 2009) |
Portugal | NA | 33–66 | 2 | NR | NR | NR | 7280 g | NR | NR | NR | NR | NR | NR | NR | (Proença et al., 2005) |
a the number of collected samples; b NR represents that this chemical is not reported in this reference; c “NA” represents that information is not available; d serum samples are used in this reference; e ND represents that this chemical is not detected in this reference; f plasma samples are used in this reference; g this value is calculated by two levels of IMI (2.05 and 12.5 µg/mL) in this reference. |
3.2 Correlations and source analysis
Associations among the concentrations (log-transformed) of individual p-NEO, m-NEOs, and between m-NEOs and their corresponding p-NEOs in blood samples were examined through Pearson correlation analysis, and all sampling populations were considered collectively (Fig. 2).
In general, significant positive relationships (p < 0.01) were found among most p-NEO concentrations in the blood, and the correlation coefficients ranged from 0.213 (IMI vs. THIX) to 0.459 (IMI vs. THI; Fig. 2). Our findings were similar to those of previous studies, which suggested significant positive correlations among the concentrations of individual p-NEOs in urine (Zhang et al. 2019), saliva and periodontal blood (Zhang et al. 2021b), and tooth (N. Zhang et al. 2021) samples. NEOs were commonly detected in foodstuffs (Chen et al. 2020; D. Li et al. 2020; Lu et al. 2018; Song et al. 2018) and environmental matrices (i.e., water and indoor dust) (Mahai et al. 2021; Wan et al. 2019; Wang et al. 2019; Zhang et al. 2021a) in China. Thus, these results suggest that the sources of these p-NEOs are common or related.
Importantly, the relationships between p-NEOs (i.e., ACE, DIN, and IMI) and their corresponding m-NEOs (i.e., ACE-dm, DN, UF, 5-OH-IMI, and Of-IMI) were positively correlated (p < 0.05), with correlation coefficients ranging from 0.173 (DIN vs. DN) to 0.338 (DIN vs. UF; Fig. 2). Our results were consistent with those of the previous reports on urinary (Song et al. 2020; Zhang et al. 2021a) and blood (Xu et al. 2021) concentrations. As mentioned above, p-NEOs could be easily metabolized into characteristic metabolites through phase I reaction in mammals (Casida 2011; Ford and Casida 2006a, b). The significant positive relationships between p-NEOs and m-NEOs might result from the metabolic transformation of p-NEOs in humans. However, m-NEOs could be directly detected in water, dust, soil, tea, fruits, and vegetable samples from China (Chen et al. 2020; S. Li et al. 2020; Li et al. 2020; Mahai et al. 2021; Wan et al. 2019; Wan et al. 2020; Zhang et al. 2020; Zhang et al. 2021a), which indicated that exogenous m-NEOs might be an important source of dietary m-NEO exposure. Furthermore, significant positive correlations were found between ACE-dm and 5-OH-IMI (r = 0.273, p < 0.01), ACE-dm and DN (r = 0.182, p < 0.05), ACE-dm and UF (r = 0.335, p < 0.01), 5-OH-IMI and DN (r = 0.298, p < 0.01), 5-OH-IMI and UF (r = 0.271, p < 0.01), and Of-IMI and DN (r = 0.593, p < 0.01; Fig. 2). Therefore, the sources of m-NEOs might be not only be from the metabolic biotransformation of p-NEOs in the human body but also from exogenous sources.
3.3 Association of blood levels with demographic factors
In the present study, age-related patterns of the levels of NEOs and their metabolites in all donors were examined using Pearson’s rank correlation. No significant correlations (p > 0.05) were found between age and concentrations of target analytes. We further categorized the age into three groups, on the basis of age histogram (Figure S1): < 40 years, 40–60 years, and > 60 years. No significant differences (p > 0.05, Mann–Whitney U test) in blood NEOs and their metabolites concentrations were found among these three age groups both in health and liver cancer populations (Fig. 3). A similar pattern was also found in previous reports on the levels of NEOs and their metabolites in human urine, saliva, periodontal blood, tooth, and blood samples collected from China (Song et al. 2020; Xu et al. 2021; Zhang et al. 2019; Zhang et al. 2021b; N. Zhang et al. 2021). These results indicate that human exposure to NEOs and their metabolites might not relatively differ in these populations.
Subsequently, gender-related differences in the levels of NEOs and their metabolites among health and liver cancer populations were examined. In the healthy population, the median concentrations of most target analytes (except for ACE, IMI, 5-OH-IMI, and ACE-dm) in females were higher than those in males (Fig. 3a). Moreover, significant differences (p < 0.05) in the median concentrations (ng/mL) of CLO (0.39 vs. 0.23), THI (0.46 vs. 0.26), DIN (0.76 vs. 0.50), UF (0.98 vs. 0.48), and IMIeq (28.8 vs. 20.7) were observed between females and males (Fig. 3a). Similarly, females from the liver cancer population had higher median levels of most NEOs and their metabolites (except for ACE, IMI, THIX, DN, and ACE-dm) than males (Fig. 3b). However, no gender-related differences were found in liver cancer population (Fig. 3b). Such trend was also found in previous studies reported in human matrices (i.e., urine, saliva, periodontal blood, tooth, and blood), which indicated that females had higher concentrations of NEOs and their metabolites than males (Ospina et al. 2019; Song et al. 2020; Tao et al. 2019; Xu et al. 2021; Zhang et al. 2021a, b, c; N. Zhang et al. 2021). Although the mechanism of gender-related pattern is still unclear, the reason might be that females have smaller body sizes, higher metabolic capability, and higher consumption of foodstuffs than males (Chen et al. 2020; Steer et al. 2006; Tipton 2001).
3.4 Association of blood levels of NEOs and their metabolites with liver cancer
The associations between NEOs and their metabolites concentrations and liver cancer were examined. The crude and adjusted ORs for liver cancer diagnosis and 95% CIs for target analytes in blood samples were presented in Table 3. In addition, the correlations between exposure to these target compounds and α-fetoprotein (AFP) values, a biomarker commonly used in the management of population with liver cancer (Giannini et al. 2014), were analyzed (Table S 5).
Table 3
Odds of liver cancer diagnosis by blood concentrations of NEOs and their metabolites presented in a liver cancer population (n = 274)-health population (n = 100) study from South China.
Compounds (ng/mL) | Crude | p-trend | Adjusted a | p-trend |
| ORs (95% CI) | | ORs (95% CI) | |
ACE | 0.42 (0.12–1.13) | > 0.05 | 0.31 (0.09–0.98) | > 0.05 |
ACE-dm | 3.45 (0.36–7.13) | < 0.01 | 3.02 (0.22–6.65) | < 0.05 |
DIN | 0.78 (0.16–1.88) | > 0.05 | 0.43 (0.10–1.58) | > 0.05 |
DN | 2.33 (0.31–4.20) | < 0.05 | 2.25 (0.25–3.02) | > 0.05 |
UF | 2.64 (0.47–4.33) | < 0.05 | 2.31 (0.13–4.15) | > 0.05 |
THI | 0.72 (0.24–2.88) | > 0.05 | 0.48 (0.13–2.05) | > 0.05 |
CLO | 1.88 (0.76–3.77) | > 0.05 | 1.27 (0.44–2.83) | > 0.05 |
THIX | 0.99 (0.31–2.75) | > 0.05 | 0.63 (0.24–1.96) | > 0.05 |
IMI | 2.24 (0.75–5.86) | > 0.05 | 1.66 (0.47–4.37) | > 0.05 |
Of-IMI | 3.31 (1.68–8.56) | < 0.05 | 2.53 (1.01–6.09) | > 0.05 |
5-OH-IMI | 2.55 (0.52–5.54) | < 0.05 | 2.26 (0.41–5.07) | > 0.05 |
Σp-NEOs | 5.71 (3.14–12.8) | > 0.05 | 4.89 (2.89–10.8) | > 0.05 |
Σm-NEOs | 8.93 (4.37-19.0) | < 0.01 | 7.58 (3.72–16.5) | < 0.05 |
ΣNEOs | 8.67 (3.02–18.6) | < 0.01 | 6.24 (2.83–13.7) | < 0.05 |
IMIeq | 9.02 (5.77–22.7) | < 0.01 | 7.63 (3.99–17.9) | < 0.05 |
a Adjusted for age and gender. | | |
As shown in Table 1, the median concentrations of target analytes (except for THI) in blood samples collected from the liver cancer population were higher than those from the healthy population. In addition, significant liver cancer-related differences (p < 0.05) were found in the blood levels of ACE, IMI, 5-OH-IMI, Of-IMI, ACE-dm, DN, UF, Σm-NEOs, ΣNEOs, and IMIeq (Fig. 5). Furthermore, the associations between blood concentrations of 5-OH-IMI (OR = 2.55, 95% CI = 0.52–5.54, p < 0.05), Of-IMI (OR = 3.31, 95% CI = 1.68–8.56, p < 0.05), ACE-dm (OR = 3.45, 95% CI = 0.36–7.13, p < 0.01), DN (OR = 2.33, 95% CI = 0.31–4.20, p < 0.05), UF (OR = 2.64, 95% CI = 0.47–4.33, p < 0.05), Σm-NEOs (OR = 8.93, 95% CI = 4.37–19.0, p < 0.01), ΣNEOs (OR = 8.67, 95% CI = 3.02–18.6, p < 0.01), and IMIeq (OR = 9.02, 95% CI = 5.77–22.7, p < 0.01) and liver cancer were significant (Table 3). After the multivariate analyses were adjusted by age and gender, associations between blood levels of ACE-dm (adjusted OR = 3.02, 95% CI = 0.22–6.65, p < 0.05), Σm-NEOs (adjusted OR = 7.58, 95% CI = 3.72–16.5, p < 0.05), ΣNEOs (adjusted OR = 6.42, 95% CI = 2.83–13.7, p < 0.05), and IMIeq (adjusted OR = 7.63, 95% CI = 3.99–17.9, p < 0.05) and liver cancer still showed statistical significance (Table 3). Meanwhile, the AFP values and blood concentrations of ACE-dm (r = 0.453, p < 0.01), Σm-NEOs (r = 0.532, p < 0.01), ΣNEOs (r = 0.428, p < 0.05), and IMIeq (r = 0.601, p < 0.01) were positively correlated (Table S5). These findings indicated that human exposure to NEOs and their metabolites was associated with liver cancer, and the blood levels of NEOs and their metabolites were correlated with increased odds of liver cancer prevalence in this population from China. Our results were similar to recent reports, which showed that human exposure to NEOs and their metabolites would increase the odds of disease prevalence (i.e., osteoporosis and periodontitis) (Li et al. 2020a; N. Zhang et al. 2021; Zhang et al. 2021c).
The metabolism and elimination of pollutants are conducted in the liver. In addition, the liver is a critical target organ that is damaged by NEOs (Han et al. 2018). Previous studies reported that NEOs and their metabolites could cause hepatoxicity effect on zebrafish birds, rabbits, mice, and even human beings (Alarcan et al. 2020; Arfat et al. 2014; Chang et al. 2020; El Okle et al. 2018; Emam et al. 2018; Green et al. 2005a,b; Swenson and Casida 2013). Furthermore, several animal reports demonstrated that NEOs could penetrate the organ (i.e., liver) through the production of reactive oxygen species, causing remarkable oxidative damage (El Okle et al. 2018; Iturburu et al. 2018; Vieira et al. 2018; Wang et al. 2018). In addition, more recent studies have shown a positive correlation between NEO exposure and oxidative stress in humans (Li et al. 2020b; Zhang et al. 2021b). As mentioned above, the toxicity of m-NEOs was higher than that of their corresponding p-NEOs (Casida 2011; Honda et al. 2006; Marfo et al. 2015; Suchail et al. 2001; Suchail et al. 2003; Zhang et al. 2021b). This might explain the relationships between the blood levels of NEOs and their metabolites (especially for m-NEOs) and liver cancer. However, given that other pollutants could also cause liver cancer (Parrón et al. 2014; Vopham et al. 2017; Yueh et al. 2015), co-exposure assessment is needed in future studies.