Metabolism of 3-Chlorobiphenyl (PCB 2) in a Human-Relevant Cell Line: Evidence of Dechlorinated Metabolites

Abstract

Lower chlorinated polychlorinated biphenyls (LC-PCBs) and their metabolites make up a class of environmental pollutants implicated in a range of adverse outcomes in humans; however, the metabolism of LC-PCBs in human models has received little attention. Here we characterize the metabolism of PCB 2 (3-chlorobiphenyl), an environmentally relevant LC-PCB congener, in HepG2 cells with in silico prediction and nontarget high-resolution mass spectrometry. Twenty PCB 2 metabolites belonging to 13 metabolite classes, including five dechlorinated metabolite classes, were identified in the cell culture media from HepG2 cells exposed for 24 h to 10 μM or 3.6 nM PCB 2. The PCB 2 metabolite profiles differed from the monochlorinated metabolite profiles identified in samples from an earlier study with PCB 11 (3,3'-dichlorobiphenyl) under identical experimental conditions. A dechlorinated dihydroxylated metabolite was also detected in human liver microsomal incubations with monohydroxylated PCB 2 metabolites but not PCB 2. These findings demonstrate that the metabolism of LC-PCBs in human-relevant models involves the formation of dechlorination products. In addition, untargeted metabolomic analyses revealed an altered bile acid biosynthesis in HepG2 cells. Our results indicate the need to study the disposition and toxicity of complex PCB 2 metabolites, including novel dechlorinated metabolites, in human-relevant models.

Keywords: ADMET predictor; HepG2 cells; MetaDrug; Nt-HRMS; PCB 2 metabolites; dechlorination; untargeted metabolomics.

Figure 1

Metabolism of PCB 2 by HepG2 cells results in complex PCB metabolite mixtures. Representative metabolites include methoxylated metabolites [(a) MeO-OH-PCB 2 (m/z 233.03694), (b) MeO-PCB 2 sulfate (m/z 312.99377), (c) MeO-PCB 2 glucuronide (m/z 409.06906), and (d) MeO-OH-PCB 2 sulfate (m/z 328.98869)] and potential dechlorinated metabolites [(e) OH-BP sulfate (m/z 265.01710), (f) MeO-BP sulfate (m/z 279.03275), (g) MeO-BP glucuronide (m/z 375.10804), (h) MeO-OH-BP sulfate (m/z 295.02767), and (i) OH-BP cysteine (m/z 288.06945)]. For each metabolite class, up to three isomers were observed (Table 1). LC-Orbitrap MS analyses were performed in the negative mode. The extracted ion chromatograms are based on the calculated accurate masses of each metabolite class, with a mass window of 10 ppm. They are presented for samples from the high-PCB 2 concentration group. For the relative levels of PCB 2 or BP metabolites between the high- and low-PCB 2 concentration groups, see Table 1. For selected MS and MS/MS spectra, see Figures S1–S13. ND, not detected.

Figure 2

Differences in (a) retention times and the relative metabolite levels [expressed as molecular response-independent (MRI) percentages] and (b) cos θ similarity coefficients of the metabolites suggest the formation of different monochlorinated PCB metabolites in HepG2 cells exposed to PCB 2 (high and low concentrations) or PCB 11 containing a PCB 2 impurity. The retention times are corrected for the internal standard, 3-F,4′-PCB 3 sulfate, to account for batch-to-batch variability in the retention times. The analyses were performed in negative mode on an LC-Orbitrap MS (see the Experimental Section for additional details). (c) The extracted ion chromatograms with a mass window of 10 ppm and (d) the MS or MS/MS data confirm the formation of putative PCB 2 metabolites in incubations of HepG2 cells with PCB 11 containing a PCB 2 impurity. The studies with the PCB 2-contaminated PCB 11 have been described previously. For the MS and MS/MS spectra of other PCB 2 metabolites from PCB 2-contaminated PCB 11 incubations listed in Table S5, see Figures S16–S20.

Figure 3

(a) Extracted ion chromatograms (EICs) at m/z 203.02637 and 219.02129 (with a mass window of 10 ppm) and MS spectra showing the measured accurate masses of the molecular ion and the isotopic pattern of chlorine support the formation of mono- and dihydroxylated PCB 2 metabolites (OH-PCB 2 and di-OH-PCB 2, respectively) in HLM incubations with 3.6 nM or 10 μM PCB 2. For the time course of the formation of these metabolites, see Figure S21. (b) EICs at m/z 185.06027 and 219.02129 show the presence of a dihydroxylated biphenyl metabolite (di-OH-BP, a dechlorinated metabolite eluting at 5.62 min) and dihydroxylated PCB metabolites (di-OH-PCB 2 eluting at 6.17 min and di-OH-PCB 3 eluting at 6.34 min) in HLM incubations with 10 μM 4-OH-PCB 2 or 3-OH-PCB 3 for both 5 and 15 min. The featured fragment ions measured in the MS/MS spectra support the identification of (c) di-OH-BP (m/z 185.06027) in the HLM incubation with 10 μM 4-OH-PCB 2 or 3-OH-PCB 3 and (d) di-OH-PCB 2 (m/z 219.02129, eluting at 6.17 min) in the HLM incubation with 10 μM 4-OH-PCB 2. The MS/MS spectrum of di-OH-PCB 3 was not collected in the data-dependent MS/MS measurement due to its low intensity. The MS spectra showing the accurate mass of the molecular ions and the isotopic pattern of chlorine of the di-OH-PCB metabolites identified in the HLM incubations with 4-OH-PCB 2 or 3-OH-PCB 3 are provided in Figure S22. For more information regarding the metabolites identified in HLM incubations, see Table S6.

Figure 4

Proposed metabolic pathway showing the potential dechlorination of PCB 2 or, on the basis of our earlier work, PCB 11 in HepG2 cells. Dechlorination (blue arrow) occurs either via a 3,4-arene oxide or by reductive dechlorination of hydroxylated PCB metabolites with the hydroxyl group ortho to the chlorine substituent. Structures shown in green boxes were detected in this metabolic study with PCB 2. Monochlorinated metabolites on an orange background were detected in our earlier metabolic study with PCB 11 but were not identified as PCB 11 metabolites due to the limited experimental evidence. The placement of the functional groups is for the purpose of illustration only and does not indicate their actual position. The classical metabolic scheme (conventional pathways) of PCB 2 is depicted in Figure S23. Abbreviations: P450, cytochrome P450 enzyme; Glc, glucuronide; DHDH, dihydrodiol dehydrogenase; SULT, sulfotransferase; EH, epoxide hydrolase; UGT, uridine 5′-diphospho-glucuronosyltransferase; COMT, catechol-O-methyltransferase; GST, glutathione S-transferase; SG, glutathione; SCys, cysteine; GGT, γ-glutamyl transpeptidase; CysGly DP, cysteinylglycine dipeptidase; MAP, mercapturic acid pathway.

Figure 5

Untargeted metabolomic analysis of medium samples revealed concentration-dependent effects of PCB 2 on the metabolome in HepG2 cells exposed to 3.63 nM PCB 2, 10 μM PCB 2, or vehicle (DMSO). (a) Univariate statistical analyses identified 170 and 787 metabolic features above a p = 0.05 threshold when comparing low and high PCB 2 exposures, respectively, to controls. In this analysis, 32 features were above the FDR = 0.05 threshold when comparing high PCB 2 exposures and control groups. (b) Pathway enrichment analyses using feature lists with raw p values from the linear regression with concentration and univariate statistical analyses of low PCB 2 vs control and high PCB 2 vs control identified four, three, and 12 significantly affected pathways, respectively (p < 0.05). Bile acid biosynthesis was the top enriched pathway in all three pathway enrichment analyses. The number of features altered by PCB 2 in these pathways is depicted as overlap/total features. (c) Several metabolites in the bile acid biosynthesis pathway were significantly affected by PCB 2 exposure. (d) Levels of these metabolites, plotted as normalized raw intensities, increased for trihydroxycholestanal (p = 0.0373), cholate (p = 0.0232), and glycochenodeoxycholate (p = 0.0165) and decreased for taurocholate (p = 0.0262) and taurochenodeoxycholate (p = 0.0601). The accurate m/z values, retention times, adducts, significances, and confidence scores of the metabolite annotations in the bile acid biosynthesis pathway are listed in Table S7.