Insights into Free and Conjugated Forms of Neonicotinoid Insecticides in Human Serum and Their Association with Oxidative Stress

Abstract

Following exposure, neonicotinoid insecticides (NEOs) can be metabolized by both Phase I and Phase II reactions catalyzed by human cytochrome P450 enzymes. However, toxicities of parent NEOs and their metabolites are still unclear, and little is known about biotransformation rates and pathways of NEOs in humans. In this study, 98 serum samples collected in China were analyzed for free, conjugated and total forms of six parent NEOs (i.e., acetamiprid (ACE), imidacloprid (IMI), clothianidin (CLO), thiacloprid (THD), thiamethoxam (THM), and dinotefuran (DIN)) and four metabolites (i.e., N-desmethyl-acetamiprid (N-dm-ACE), 1-methyl-3-(tetrahydro-3-furylmethyl) (DIN-U), 5-hydroxy-imidacloprid (5-OH-IMI), olefin-imidacloprid (Of-IMI)). NEOs and their metabolites were detected in all serum samples, and the total median concentrations of free, conjugated, and total forms of 10 NEOs were 2.04, 2.01, and 5.12 ng/mL, respectively. Conjugated forms of NEOs accounted for only half (53%) of the total forms of NEOs. Based on the profiles of Phase I and Phase II metabolites of NEOs in serum, it was found that age is a determinant in Phase I metabolism of DIN and Phase II metabolism of IMI. The Phase II metabolites of NEOs are associated with oxidative DNA damage, and the conjugated forms of IMI, DIN, and 5-OH-IMI in serum were significantly positively correlated with oxidative stress. Overall, the amount of NEOs present in conjugated forms in human serum was determined to document the existence of a considerable proportion of free forms of these insecticides.

Introduction

Neonicotinoids (NEOs) are a new generation of insecticides that are widely used to kill insects by blocking and paralyzing the acetylcholine receptor transmission of insects.¹ Since the introduction of imidacloprid (IMI) in 1991, NEOs have rapidly dominated the insecticide market, accounting for approximately 24% of the global insecticide market share in 2018.² China is one of the largest producers and exporters of NEOs globally, with IMI production reaching 2.3 million tons in 2016.^3,4 As a consequence of the widespread application of NEOs in agricultural production, they are found in foods, in environmental samples, and even in human tissues.⁵⁻⁸ A recent study reported that NEO metabolites (m-NEOs) may increase the prevalence of liver cancer.⁹ NEOs and their metabolites can pass through the human placenta and may accumulate in the fetus.¹⁰ In vivo studies have found that some NEOs affect mammalian reproductive health, from decreased sperm quality in males and infertility in females as well as increased embryonic lethality and preterm birth.¹¹⁻¹³

Following ingestion or inhalation pathways of exposure to NEOs, these compounds are metabolized by Phase I and Phase II pathways (Figure 1), which result in conjugation in human bodies. The water-soluble and conjugated metabolites (i.e., Phase I and Phase II) of parent NEOs (p-NEOs) tend to be less toxic and excreted quickly from the body. However, in certain cases, metabolism results in reactive molecules that are more toxic than parent compounds.¹⁴ Plant and animal exposure studies showed that the metabolite of dinotefuran (DIN), 1-methyl-3-(tetrahydro-3-furylmethyl) urea (DIN-U), is more highly toxic than the parent;^15,16 for humans, Zhang et al. showed positive correlations between serum concentrations of N-desmethyl-acetamiprid (N-dm-ACE) or olefin-IMI (Of-IMI) and the prevalence of osteoporosis in the elderly.¹⁷ However, little is known about the metabolic transformation rates and pathways of NEOs in humans.

Figure 1.

Metabolic pathways of ACE (a), DIN (b), and IMI (c) under Phase I and Phase II metabolic reactions and metabolic pathways (d) of CLO, THD, and THM under Phase II reactions.

8-Hydroxydeoxyguanosine (8-OHdG), an oxidative adduct produced by reactive oxygen species (ROS), affects the eighth carbon atom of guanine and has been used as a marker of DNA oxidative damage.¹⁸ Animal and cell-based studies suggested that oxidative stress is one of the mechanisms of toxicity of NEOs, which may be a trigger for hepatotoxicity, neurotoxicity, reproductive toxicity, and immunotoxicity.^19,20 However, very few studies have been performed on humans, and only two studies have shown an association between NEO exposure and 8-OHdG. Li et al. found a significant correlation between IMI and 8-OHdG in urine,²¹ and Zhang et al. found a significant correlation between DIN and 8-OHdG in periodontal blood.²² However, little is known with regard to the correlation between 8-OHdG and free and conjugated forms of NEOs in human serum.

In this study, concentrations of free, conjugated, and total forms of six p-NEOs, four m-NEOs, and 8-OHdG were analyzed in 98 serum samples collected from the general population in China. The relative percentages of Phase I and Phase II metabolites of NEOs and sex- and age-related differences in metabolite profiles were examined. The main objectives of this study were to (i) analyze the concentrations and profiles of NEOs and their metabolites in human serum from China, (ii) investigate the metabolite profiles of NEOs in human serum, and (iii) determine the association between free and conjugated NEOs and 8-OHdG in human serum. To the best of our knowledge, this is the first study to investigate the occurrence and forms of free and conjugated NEOs and their association with oxidative stress in human serum. These data are the basis for health risk assessment from human exposure to NEOs.

Materials and Methods

Standards and Reagents

Native standards of acetamiprid (ACE), IMI, clothianidin (CLO), thiacloprid (THD), thiamethoxam (THM), and DIN and internal standards (IS) such as ACE-d₃, IMI-d₄, CLO-d₃, THD-d₄, THM-d₃, and DIN-d₃ were obtained from Sigma-Aldrich (St. Louis, MO, USA). Native standards for the three NEO metabolites of N-dm-ACE, DIN-U, and Of-IMI were obtained from Dr. Ehrenstorfer (Augsburg, Germany), and the native standard of 5-hydroxy-imidacloprid (5-OH-IMI) was obtained from Cambridge Isotope Laboratories (Tewksbury, MA, USA). The purity of the above standards were >97%. β-Glucuronidase used for the deconjugation of analytes and the native standard of 8-OHdG (>97% purity) were purchased from Sigma-Aldrich (St. Louis, MO, USA). ¹³C₃-8-OHdG (98% purity) was purchased from Cambridge Isotope Laboratories (Andover, MA, USA). Formic acid (purity >95%), HPLC-grade methanol, acetonitrile, and ethyl acetate were supplied by Merck (Damstadt, Germany). Milli-Q water was obtained with a Millipore system (Billerica, MA, USA).

Sample Collection

A total of 98 volunteers (age range: 14–86 yrs), from Guangzhou, China, provided blood samples during June to July 2021. A questionnaire survey of sociodemographic characteristics and lifestyle was conducted for all donors randomly selected in the city. The participants were healthy males (n = 24) and females (n = 74) and did not report occupational exposure to NEOs. Detailed information regarding the age and gender of the participants is shown in Table S1. Blood samples were collected through venipuncture by a registered nurse from a local hospital. After coagulation, the blood sample was centrifuged at 1200 rpm for 10 min and the serum was transferred into a new polypropylene (PP) tube. All serum samples were frozen at −20 °C until sample extraction. The collection and analysis of samples were approved by the Institutional Review Board of Sun Yat-Sen University, Guangzhou, China.

Extraction of Free Form NEOs in Serum Samples without Hydrolysis

Serum samples were extracted by following the liquid–liquid extraction (LLE) method described in our previous studies, with some modifications.^44,45 Briefly, 0.5 mL of serum samples was transferred to a 15 mL PP tube and spiked with the IS solution (mixture of ACE-d₃, CLO-d₃, DIN-d₃, IMI-d₄, THD-d₄, THM-d₃, and ¹³C₃-8-OHdG). Subsequently, 3.0 mL of ethyl acetate was added to the samples and vigorously shaken for 30 min. The mixture was centrifuged at 3000 rpm for 5 min. The supernatants were transferred into a new 15 mL PP tube. The extraction was repeated twice. The combined extracts were concentrated to near dryness and redissolved in 0.5 mL of methanol and placed into LC vials for instrumental analysis.

Extraction of Total Form NEOs in Serum Samples with Hydrolysis

Serum samples were treated with β-glucuronidase to hydrolyze the conjugated NEOs using the method described in previous studies.²³ In detail, 0.5 mL of serum samples (spiked with 1.0 ng of ACE-d₃, CLO-d₃, DIN-d₃, IMI-d₄, THD-d₄, THM-d₃, and ¹³C₃-8-OHdG) was mixed with 175 μL of β-glucuronidase and 125 μL of 0.1 mol/L acetic acid-sodium acetate (HAC-NaAC) buffer to adjust the pH to 5.5 in 15 mL PP tubes and then vortexed for 1 min. Finally, the mixtures were incubated for 12 h at 37 °C, allowing the NEO conjugates to be hydrolyzed to free-form NEOs. After that, ethyl acetate was used to extract NEOs in serum samples after hydrolysis, as described above.

Instrumental Analysis

The concentrations of target analytes in sample extracts were analyzed using an Agilent 1290 series HPLC system (Agilent Technologies, CA, USA) coupled to an Applied Biosystems SCIEX 5500 triple quadrupole mass spectrometry (ESI-MS/MS; Applied Biosystems, CA, USA). An analytical column (Zorbax SB-C18; 3.5 μm, 100 × 2.1 mm; Agilent) was used for the chromatographic separation of the analytes. The oven temperature was set at 25 °C. The sample injection volume was 2 μL. The mobile phases were acetonitrile and water (containing 0.1% formic acid) at a flow rate of 0.3 mL/min with the following gradient: 0–3 min, 5–30% acetonitrile; 3–6 min, 30–45% acetonitrile; 6–9 min, 45–5% acetonitrile; 9–12 min, 5% acetonitrile. The positive-ion multiple reaction monitoring mode was applied to record the spectral data. The curtain gas, collision gas, nebulizing gas, and heater gas were set at 20, 12, 50, and 60 psi, respectively. The ionization voltage was 3.5 kV, and the source heater temperature was 300 °C. The precursor (Q1), product ion (Q3), declustering potential (DP), and collision energy (CE) for NEOs and 8-OHdG quantification are described in Table S2.

Quality Assurance and Quality Control

As a check for background levels of contamination, instrumental blanks, procedural blanks, and matrix blanks were analyzed with every 20 samples. IMI was detected in instrumental blanks at 0.03 ng/mL. The highest concentration measured in the blanks was subtracted from the corresponding concentrations in the samples. Ten-point solvent (i.e., methanol) calibration standards (0.05–50.0 ng/mL) were prepared to quantification. The regression coefficient (r²) of each standard calibration curve was >0.99. The matrix effects of the target analytes ranged from −6.4 to 8.2%, indicating that there were no significant enhancing or inhibiting effects in human serum. The limits of quantification (LOQs) of target analytes were derived from peak values with signal-to-noise ratios (S/N) of 10, and these values ranged from 0.005 (ACE and IMI) to 0.05 (DIN and DIN-U) ng/mL (Table 1). The matrix spiked recoveries of individual analytes were determined in randomly selected serum samples (n = 5) fortified with analyte standards at a concentration of 10 ng/mL, and the average matrix spike recoveries of target analytes ranged from 79% ± 7% to 106 ± 19% for serum samples without enzymatic hydrolysis and from 90 ± 18% to 115 ± 7% for serum samples with enzymatic hydrolysis. The internal standards ACE-d₃, CLO-d₃, DIN-d₃, IMI-d₄, THD-d₄, THM-d₃, and ¹³C₃-8-OHdG were spiked into all serum samples (1.0 ng in each) before extraction. The recoveries of internal standards in serum samples ranged from 64 ± 9% to 119 ± 14% without enzymatic hydrolysis and from 56 ± 10% to 130 ± 18% with enzymatic hydrolysis.

Table 1. Concentrations (ng/mL) of Free, Conjugated, and Total Forms Neonicotinoid Insecticides and Their Metabolites in Serum Samples Collected from Participants in China.

	ACE	IMI	CLO	THD	THM	DIN	Σp-NEOsa	N-dm-ACEb	DIN-U	5-OH-IMI	Of-IMI	Σm-NEOsc	Σ10NEOsd
LOQ	0.005	0.005	0.05	0.01	0.02	0.05	0.005	0.01	0.05	0.01	0.02	0.01	0.005
free form concentration (ne = 98)
DR (%)f	100	86	88	65	94	61	100	100	63	65	86	100	100
min	0.01	<LOQg	<LOQ	<LOQ	<LOQ	<LOQ	0.07	0.02	<LOQ	<LOQ	<LOQ	0.13	0.46
mean	0.09	0.16	0.32	0.05	0.15	0.19	0.95	0.29	0.63	0.07	0.11	1.11	2.21
median	0.05	0.13	0.15	0.02	0.12	0.11	0.72	0.16	0.59	0.03	0.10	0.88	1.87
max	1.16	1.05	2.25	0.92	1.11	1.43	6.82	2.38	3.04	1.02	0.45	3.64	10.1
conjugated form concentration (n = 98)
DR (%)	68	83	87	59	68	78	89	64	51	69	65	71	97
min	<LOQ	<LOQ	<LOQ	<LOQ	<LOQ	<LOQ	<LOQ	<LOQ	<LOQ	<LOQ	<LOQ	<LOQ	<LOQ
mean	0.04	0.08	1.14	0.02	0.07	0.52	1.84	0.03	0.35	0.07	0.08	0.48	3.01
median	0.04	0.06	0.83	0.01	0.08	0.37	1.51	0.05	0.27	0.07	0.05	0.27	2.12
max	0.39	0.74	4.04	0.33	0.89	3.28	6.12	0.85	5.61	0.62	1.01	6.79	12.9
total form concentration (n = 98)
DR (%)	98	86	98	81	95	90	100	100	73	81	99	100	100
min	<LOQ	<LOQ	<LOQ	<LOQ	<LOQ	<LOQ	0.16	0.04	<LOQ	<LOQ	<LOQ	0.19	1.59
mean	0.13	0.23	1.46	0.06	0.22	0.69	2.79	0.32	0.94	0.14	0.19	1.59	4.95
median	0.12	0.21	1.26	0.04	0.20	0.48	2.54	0.20	0.67	0.11	0.16	1.27	4.09
max	0.51	1.25	5.26	0.36	0.94	3.58	7.22	2.72	5.61	0.62	1.01	7.04	13.4

^aSum of concentrations of six parent NEOs.

^bNumber participants of four NEO metabolites (N-dm-ACE, DIN-U, 5-OH-IMI, and Of-IMI), Σm-NEOs, and Σ₁₀NEOs was 78.

^cSum of concentrations of four NEO metabolites.

^dSum of concentrations of NEOs and their metabolites.

^en: number of donors.

^fDR: detection rate.

^g<LOQ: the values below than LOQ.

Statistical Analysis

Data analysis and graphing programs were performed with IBM SPSS version 27.0 and GraphPad Prism version 9.0. The concentrations below the LOQ were assigned a value equal to half of the LOQ for the calculation of the median and mean values. The sum concentrations of six p-NEOs were denoted as Σp-NEOs, the sum concentrations of four m-NEOs were denoted as Σm-NEOs, and the sum concentrations of six p-NEOs and four m-NEOs were denoted as Σ₁₀NEOs. The raw and log-transformed concentrations of analytes were tested for normality by using the Kolmogorov–Smirnov test. Pearson correlation coefficients were used to test the associations between variables when the data were distributed normally; otherwise, Spearman’s rank correlation was performed. Furthermore, one-way ANOVA was used to investigate the differences between groups when the data were distributed normally; otherwise, Mann–Whitney U test was used. The statistical significance was set at p < 0.05.

Results and Discussion

Concentrations of Free, Conjugated and Total Forms of NEOs

Free Form of NEOs

The serum samples were extracted without enzymatic hydrolysis for the quantification of the free form of NEOs (NEOs__free) concentrations (Table 1 and Figure S1). Six p-NEOs__free (i.e., ACE__free, IMI__free, CLO__free, THD__free, THM__free, and DIN__free) and four m-NEOs__free (i.e., N-dm-ACE__free, DIN-U__free, 5-OH-IMI__free, and Of-IMI__free) were detected with detection rates (DRs) ranging from 61 (DIN__free) to 100% (ACE__free and N-dm-ACE__free). The median (range) concentration of Σ₁₀NEOs__free (i.e., the sum of NEOs__free) was 1.87 ng/mL (0.46 to 10.12) across all participants. Among six p-NEOs__free, CLO__free had the highest median serum concentration (0.15 ng/mL), followed by IMI__free (0.13), THM__free (0.12), and DIN__free (0.11) (Table 1). DIN-U__free had the highest median concentration at 0.59 ng/mL among all m-NEOs__free, followed by N-dm-ACE__free (0.16), Of-IMI__free (0.10), and 5-OH-IMI__free (0.03).

Conjugated Form of NEOs

Serum specimens were enzymatically deconjugated with β-glucuronidase to hydrolyze conjugated species, which enabled measurement of total concentrations (i.e., free + conjugated forms) of NEOs (i.e., NEOs__total). The concentrations of the conjugated form of NEOs (i.e., NEOs__conjugated) were calculated as total concentration subtracted from the free form. As shown in Table 1, all NEOs__conjugated were detected in 51 (DIN-U__conjugated) to 87% (CLO__conjugated) of serum samples, and the median concentration of Σ₁₀NEOs__conjugated was 2.12 ng/mL (range: <LOQ to 12.9 ng/mL). CLO__conjugated had the highest median serum concentration (0.83 ng/mL) among p-NEOs__conjugated (Figure S1), followed by DIN__conjugated (median: 0.37 ng/mL). THM__conjugated, IMI__conjugated, ACE__conjugated, and THD__conjugated had low median concentrations (<0.10 ng/mL). DIN-U__conjugated was also the abundant m-NEO__conjugated with a median value of 0.27 ng/mL, followed by 5-OH-IMI__conjugated (0.07 ng/mL), N-dm-ACE__conjugated (0.05), and Of-IMI__conjugated (0.05). These striking results of NEOs__conjugated indicated that p-NEOs and their metabolites²⁴⁻²⁶ are subjected to Phase II metabolic transformation to produce glucuronide conjugates in humans (Figure 1).^14,27

Total NEOs (Free + Conjugated)

NEOs__total in serum samples was directly measured by employing enzymatic deconjucation as the sum concentration of NEOs__free and NEOs__conjugated. The DRs of NEOs__total in all serum samples ranged from 73% (DIN-U__total) to 100% (N-dm-ACE__total), and the median concentration of Σ₁₀NEOs__total was 4.09 (range: 1.59 to 13.4) ng/mL (Table 1 and Figure S1). CLO__total was the abundant p-NEO__total with a median value at 1.26 ng/mL, followed by DIN__total (0.48 ng/mL), IMI__total (0.21), THM__total (0.20), ACE__total (0.12), and THD__total (0.04). Among m-NEOs__total, DIN-U was abundant at 0.67 ng/mL, followed by N-dm-ACE (0.20), Of-IMI (0.16), and 5-OH-IMI (0.11).

Limited studies have reported the serum concentrations of p-NEOs and m-NEOs in humans.^9,39−41 Our results on serum levels of NEOs agree with those monitored previously in whole blood samples (median Σ₁₁NEOs: 7.75 ng/mL) and in periodontal blood samples (median Σ₉NEOs: 6.87 ng/mL) collected from Guangzhou.^9,28 However, the median sum concentration of seven p-NEOs (including ACE, IMI, CLO, THD, and THM) in human serum samples collected from Wuxi city in China was 0.026 ng/mL,²⁹ which is 2 orders of magnitude lower than that of Σp-NEOs__total (median: 2.54 ng/mL) found in our study, mostly because our study used enzymatic hydrolysis while the other study did not. Therefore, our study highlights the significance of quantification of free and conjugated forms of NEOs. Similarly, most of the studied NEOs (i.e., ACE, IMI, CLO, THD, THM, DIN, and N-dm-ACE) were rarely detected in serum samples from Saudi Arabia, and only IMI was detected with a low median concentration at 0.04 ng/mL,³⁰ which was also much lower than that of Σ₁₀NEOs__total (median: 4.09 ng/mL) observed in our results (Table 1). Although conjugated NEOs were detected by enzymatic hydrolysis in the above literature,³⁰ it was mainly due to the lower use of pesticides with nonagricultural predominance and less agricultural production activities in Saudi Arabia.³¹ In contrast, China is the largest user and producer of NEOs,^32,33 with an annual production of 1.2 million tons of IMI in 2012, accounting for two-thirds of the total global production.⁴

Free versus Conjugated Forms of NEOs

The median serum concentrations of Σ₁₀NEOs__conjugated were significantly higher (p < 0.05) than that of Σ₁₀NEOs__free (2.12 vs. 1.87 ng/mL) (Table 1 and Figure 2a); the median Σp-NEOs__conjugated (1.51) concentration was significantly higher (p < 0.001) than that of Σp-NEOs__free (0.72), while the median level of Σm-NEOs__conjugated (0.27) was significantly lower (p < 0.001) than that of Σm-NEOs__free (0.88). CLO and DIN were dominated p-NEOs in conjugated forms (total accounting for 86% of Σp-NEOs__conjugated), and the median concentration of CLO__conjugated (0.83 ng/mL) and DIN__conjugated (0.37) were significantly higher (p < 0.0001) than that of CLO__free (0.15) and DIN__free (0.11) (Figure S1), respectively. Among all analyzed m-NEOs, DIN-U was abundant as both free and conjugated forms; interestingly, the median concentration of DIN-U__conjugated was much lower (p < 0.05) than that of DIN-U__free (0.27 vs 0.59 ng/mL); similarly, N-dm-ACE__conjugated (0.05) and Of-IMI__conjugated (0.05) were also significantly lower than those of N-dm-ACE__free (0.16) (p < 0.05) and Of-IMI__free (0.10) (Figure S1), respectively. It is worthy to note that DIN-U is a specific metabolite of DIN, and DIN is mainly present as conjugated form, while its metabolite (i.e., DIN-U) is mainly exists as free forms.

Figure 2.

Median concentration distributions (a) and composition profiles (b) of the free form and conjugated form of parent neonicotinoids and their metabolites in human serum samples.

The contribution of median Σ₁₀NEOs__free and Σ₁₀NEOs__conjugated concentrations to Σ₁₀NEOs__total were about 47 and 53%, respectively, suggesting that NEOs__conjugated, like NEOs__free, had an equal and non-negligible proportion in human serum (Figure 2b); in other words, NEOs__conjugated cannot be ignored when evaluating human exposure levels of NEOs through blood biomonitoring. Furthermore, compound-specific variations in the composition profiles of NEOs__conjugated and NEOs__free were observed (Figure 2b). The contributions of each p-NEOs_{_conjugated} to p-NEOs_{_total} ranged from 35 (IMI__conjugated) to 83% (CLO__conjugated and DIN__conjugated). For m-NEOs, 5-OH-IMI__conjugated and Of-IMI__conjugated accounted for 87 and 36% of 5-OH-IMI__total and Of-IMI__total, respectively. This result suggests that 5-OH-IMI tends to be present as conjugated forms, while Of-IMI exhibited an opposite trend. Because the structure of 5-OH-IMI contains an −OH function group that may be glucuronidated easily to produce conjugated forms (Figure 1).³⁴ Furthermore, N-dm-ACE__conjugated and DIN-U__conjugated only accounted for 21 and 46%, respectively, of total concentrations, and their structures are less susceptible to conjugation reactions; therefore, they are mainly present as free forms. Based on available structure–activity information and binding site models, glucuronidation of NEOs and their metabolites was probably a detoxification process in organisms; however, few studies indicated that some conjugated NEOs may be more toxic than p-NEOs. For example, glucuronidation of sulfoxaflor metabolites caused metabolic disturbances in the liver and bile.³⁵ Therefore, from a toxicological point of view, it is significant to consider the serum levels of p-NEOs__conjugated and m-NEOs__conjugated, although limited data are available on the toxicokinetics of Phase II metabolic reactions of NEOs.^14,34,36

Characteristics of Phase I and Phase II Metabolic Reactions

The metabolism of NEOs follows the Phase I and Phase II reactions (Figure 1). Phase I metabolic reactions are mainly redox and hydrolysis. NEOs are first metabolized by Phase I enzymes, generating multiple metabolic sites for conjugation reactions. Furthermore, Phase II reactions are mainly conjugation reactions, often interacting with the p-NEOs themselves and their metabolites, ultimately forming a series of conjugated NEOs for excretion.^14,37 In this study, therefore, the serum levels of p-NEOs__free represent the untransformed p-NEOs; m-NEOs__total represents the Phase I metabolism of p-NEOs, and p-NEOs__conjugated and m-NEOs__conjugated represent the Phase II metabolism of p-NEOs and m-NEOs, respectively.

Phase I Metabolic Reactions

Significant positive correlations between serum levels of p-NEOs_{_total} and the corresponding m-NEOs_{_total} were observed (Table S3), with regression coefficients in the range of 0.233 to 0.687 (p < 0.05). This suggests that the four m-NEOs analyzed in this study originated from the Phase I metabolism of corresponding p-NEOs. In order to assess the differences in Phase I metabolic transformations of three p-NEOs (i.e., ACE, DIN, and IMI) selected in this study, a coefficient (f_{i_I}) was used to elucidate the distribution of each metabolite (N-dm-ACE_{_total}, DIN-U_{_total}, and 5-OH-IMI_{_total}) in relation to the total concentrations of p-NEO_{_total} and m-NEO_{_total}, calculated using the following eq 1:

1

where C_1,i (ng/mL) and C_2,j (ng/mL) are the individual serum concentrations of total form p-NEOs (i: ACE, DIN, or IMI) and their corresponding total form m-NEOs (j: N-dm-ACE, DIN-U, or 5-OH-IMI), respectively, for each sample. In vitro and in vivo studies identified Of-IMI as a less important metabolite of IMI.^38,39 Therefore, we excluded this m-NEO for the estimation of f_{IMI_I}. Furthermore, the serum samples that contained undetectable concentrations (<LOQ) of m-NEOs_{_total} or corresponding p-NEOs_{_total} were also excluded from f_{i_I} analysis. DIN exhibited the highest median f_{DIN_I} value (median: 0.62, range: 0.25–0.79), followed by f_{ACE_I} (median: 0.59, range: 0.13–0.96) and f_{IMI_I} (median: 0.34, range: 0.05–0.68), indicating considerable differences in the proportions of m-NEOs and corresponding p-NEOs in serum. Gender- and age-related differences in f_{i_I} values were also estimated. The f_{i_I} values of p-NEOs did not exhibit significant gender-related differences (p > 0.05). Interestingly, a significantly negative correlation was observed between log f_{DIN_I} and log age (r = −0.384, p < 0.01) (Figure 3). Although the number of samples analyzed is limited, the age-related differences in log f_{DIN_I} suggested that the Phase I metabolic reaction of DIN is higher in the younger age group than in the older group.

Figure 3.

(a) Pearson correlation between log f_DIN__I (f_DIN__I: ratio between serum concentrations of DIN-U and sum serum concentrations of DIN and DIN-U) and log age of donors who had detectable concentrations (>LOQ) of both DIN and DIN-U in serum. (b) Pearson correlation between log f_IMI__II (f_IMI__II: ratio between serum concentrations of IMI__conjugated and serum concentrations of IMI__total) and log age of donors who had detectable concentrations (>LOQ) of both IMI__conjugated and IMI__total in serum.

Phase II Metabolic Reactions

To estimate the Phase II metabolic transformations of p-NEOs and m-NEOs, a coefficient (f_{k_II}) was defined as the ratio of the concentrations of NEOs__conjugated (C_{k, conjugated}) to NEOs__total (C_k, total) and was calculated as shown in eq 2:

2

where k represents individual NEOs (i.e., ACE, IMI, CLO, THD, THM, DIN, N-dm-ACE, 5-OH-IMI, Of-IMI, and DIN-U). Furthermore, the serum samples which contained undetectable concentrations (<LOQ) of NEOs__conjugated or NEOs__total were excluded from f_{k_II} analysis. The median f_{IMI_II} value was the lowest (0.35), followed by f_{THM_II} (0.39) and f_{ACE_II} (0.48), indicating that they are mainly present in serum as free forms. While the median f_{THD_II}, f_{CLO_II} and f_{DIN_II} were 0.67, 0.83, and 0.83, respectively, suggesting that these p-NEOs are more prone to Phase II metabolic reactions to generate conjugated forms. The f_{k_II} was plotted against the logarithm of the total serum concentration (log C_k, total) (Figure 4). A significantly positively correlation between f_{k_II} and log C_k, total was found for all six p-NEOs (r = 0.293–0.704, p < 0.05; Figure 4), indicating that Phase II metabolic rates increased with increased concentrations of p-NEOs. However, among the four m-NEOs, the median f_{k_II} values of N-dm-ACE, Of-IMI and DIN-U were relatively low at 0.21, 0.36, and 0.46, respectively, indicating that these metabolites are mainly exist in free forms. However, the median f_{5-OH-IMI_II} was high at 0.87, and this can be explained by the structure of 5-OH-IMI containing −OH groups, which makes it easier to bind to glucosinolates. Significant positive correlations were found between the corresponding f_{k_II} and log C_k, total for three m-NEOs (DIN-U, 5-OH-IMI and Of-IMI) (r = 0.416–0.687, p < 0.01; Figure S2), indicating that the rates of Phase II metabolism of these metabolites increases with increasing concentrations. Nevertheless, no significant correlations were found between f_{N-dm-ACE_II} and log C_{N-dm-ACE,total}.

Figure 4.

Pearson correlation between the fraction (f_{k_II}) of conjugated form p-NEOs and logarithmic concentration of total form p-NEOs (log C_k,total) for ACE (a), IMI (b), CLO (c), THD (d), THM (e), and DIN (f), in donors who had detectable concentrations (>LOQ) of both p-NEOs__conjugated and p-NEOs__total in serum.

Gender- and age-related differences in the f_{k_II} values were calculated among participants. Our results found that the median f_{ACE_II} was significantly higher (p < 0.05) for males (0.74) than that for females (0.53), most likely due to the higher metabolic capacity of males compared to females.⁴⁰ But the reasons for gender-related differences still need to be further investigated. In addition, a significant age-related negative correlation was found in log f_{IMI_II} (r = −0.283, p < 0.05) (Figure 3), indicating that the Phase II metabolic reaction of IMI gradually decreases with increasing age, which can be explained by a relatively slower metabolic rate in elderly.

Associations between Oxidative Stress and Serum NEOs Levels

In this study, 78 of 98 serum samples were used for correlation analysis between oxidative stress and serum levels of NEOs_{_free} and NEOs_{_conjugated}, respectively (Table 2). This is the first study to examine differences in correlations of an oxidative stress marker between free and conjugated forms of NEOs in humans.

Table 2. Pearson Correlation Analysis between the Log-Transformed Serum Concentrations of Individual NEOs and 8-OHdG.^a.

	free form		conjugated form
	r	p	r	p
ACE	–0.016	0.900	0.081	0.610
IMI	0.093	0.477	0.350*b	0.021
CLO	–0.011	0.936	0.188	0.162
THD	–0.032	0.831	0.161	0.297
THM	–0.087	0.504	0.262	0.090
DIN	0.342	0.055	0.390**	0.005
N-dm-ACE	–0.081	0.534	0.238	0.151
DIN-U	0.041	0.806	0.282	0.132
5-OH-IMI	0.081	0.637	0.363*c	0.017
Of-IMI	0.052	0.713	0.234	0.152

^a“r” represents the correlation coefficients, and “p” represents significance (two-tailed); * represents p < 0.05.

^b** represents p < 0.01; concentration values less than LOQ were excluded from these analyses; log-transformed concentrations were used for the analysis.

^cBold font indicates significant associations (p < 0.05 or p < 0.01).

No significant correlations were found between NEOs__free and 8-OHdG (p > 0.05; Table 2) for all studied NEOs. However, 8-OHdG was significantly positively correlated with IMI__conjugated (r = 0.350, p = 0.021), DIN__conjugated (r = 0.390, p = 0.005), and 5-OH-IMI__conjugated (r = 0.363, p = 0.017) (Table 2). This suggests differences in toxicity between the NEOs__free and NEOs__conjugated forms of NEOs. Similar correlations were also observed in a study with periodontal blood.²² No earlier study reported the toxicity of conjugated NEOs in humans, and the available reports are all based on the toxicity of NEOs__free or NEOs_{_total}.^9,22 Our results suggested that IMI__conjugated, DIN__conjugated, and 5-OH-IMI__conjugated contribute to DNA damage in humans in comparison with those of IMI__free, DIN__free, and 5-OH-IMI__free, respectively.

Several studies have presented evidence that conjugated forms of chemicals may elicit greater toxicity. Kauffman et al. summarized that some chemicals may generate more potent carcinogenic compounds via a Phase II conjugation reaction.⁴¹ For example, the metabolic transformation of 2-naphthylamine produces N-hydroxy-N-glucuronide that decomposes in urine to hydroxylamine and its protonated nitrogen ions, which easily react with DNA to cause bladder cancer.⁴² Furthermore, Feng et al. investigated possible impacts of free and conjugated forms of phthalate metabolites on male fertility and they showed that in the fertile group, the ratios of free forms to conjugated forms of monoethyl phthalate and mono-2-ethyl-hexyl phthalate were higher than those in the infertile group, indicating a higher efficiency of Phase II metabolic transformation in the infertile group.^43,44 In this study, the correlation of IMI__conjugated and 5-OH-IMI__conjugated with 8-OHdG also suggested oxidative stress. IMI-induced oxidative DNA damage has been reported in many studies.^35,45−47 In vivo, IMI exposure revealed that it increased ROS level and elevated activities of sulfoxide dismutase (SOD), catalase (CAT) and glutathione-S-transferase (GST); ROS attacked cell membranes resulting in increased malondialdehyde (MDA), which cause DNA oxidative damage.^35,46 Moreover, IMI-induced DNA oxidative damage is dose- and time-dependent.⁴⁶In vitro, human prostate epithelial WPM-Y.1 cell line exposed to low concentrations of IMI also reported an increase in MDA levels, further explaining the failure of antioxidant defense mechanisms and thus the induction of oxidative damage.⁴⁷ In addition, although the serum level of 5-OH-IMI__conjugated was low (median: 0.07 ng/mL), it accounted for up to 87% of the total forms. 5-OH-IMI was associated with elevated oxidative stress, but our results imply that 5-OH-IMI__conjugated may be involved in oxidative stress rather than 5-OH-IMI__free. Oxidative stress may be a relevant mechanism of IMI toxicity, and metabolites are more toxic than the parent IMI.⁴⁸⁻⁵⁰ A study on the metabolism of IMI in honeybees (Apis mellifera) showed that 5-OH-IMI contributed to prolonging the effects and was involved in honeybee mortality.⁵⁰ On the other hand, although there is no evidence that DIN__free directly causes oxidative stress, it is strongly speculated that the Phase II metabolic reaction of DIN (e.g., DIN__conjugated) may be an important factor to induce oxidative stress. Many experimental animal studies revealed that DIN induces oxidative stress.⁵¹⁻⁵³ The mechanism of DIN-induced oxidative stress was also similar to that of IMI, and elevated levels of SOD and CAT, MDA and GSH were observed in fish (Gobiocypris rarus).⁵² Moreover, DIN can induce hepatotoxicity and carcinogenicity via oxidative stress.⁹

However, it should be noted that a limited sample size may affect the statistical power of our correlations and may bias the results. Therefore, a large sample size is needed to further augment these findings.

In summary, this study is the first to report the presence of both free and conjugated forms of NEOs in human serum. Human exposure to NEOs should take into account the different forms of NEOs, particularly conjugated NEOs, which account for about half of the total NEO concentrations present in human serum. In the general population, both Phase I metabolism and Phase II metabolism of NEOs seem to be effective, and certain NEOs may exhibit age-dependent metabolic rates. Nevertheless, further studies are needed to determine the biotransformation pathway of NEOs. Moreover, some conjugated NEOs may be more toxic than freely available NEOs in serum. For instance, the conjugated forms of IMI, DIN and 5-OH-IMI were correlated with oxidative stress but not the aglycones. These findings provide novel insights into the evaluation of toxicological risks associated with free and conjugated NEOs. Since the enzymatic hydrolysis method used in this study mostly releases NEO-glucuronide and fewer other forms of conjugates, the total amount of conjugated NEOs in human serum is likely underestimated, and further research is needed.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/envhealth.3c00023.

• Additional detailed information included characteristics of all participants, instrumental analysis parameters for all analytes, concentrations of different forms of NEOs in human serum samples, Pearson correlation analysis between p-NEOs and corresponding m-NEOs, fraction of conjugated m-NEOs and logarithmic concentration of total m-NEOs (PDF)

The authors declare no competing financial interest.