Introduction

Seventy eight of the 209 possible polychlorinated biphenyl (PCB) congeners display axial chirality and exist as rotational isomers, or atropisomers, that are non-superimposable mirror images of each other (Fig. 1) (Lehmler et al. 2009). As with other axially chiral biaryls, a PCB congener is chiral when both phenyl rings have an unsymmetrical substitution pattern relative to the axis formed by the central C–C bond of the biphenyl system. The rotational free energy barriers of PCB congeners with three or four chlorine substituents ortho to the central C–C bond are ∆ǂG = 176.6 kJ/mol and ∆ǂG = 246 kJ/mol, respectively (Harju and Haglund 1999; Schurig and Reich 1998). These free rotational energy barriers are sufficient to prevent a racemization of chiral PCBs with three or four ortho chlorine substituents, even at the relatively high temperatures used for gas chromatographic separations (Harju and Haglund 1999). In contrast, atropisomers of chiral PCB congeners with none, one, or two ortho chlorine substituents readily racemize and, therefore, do not form stable atropisomers. Out of the 78 chiral PCB congeners, 19 PCB congeners, or C-PCBs, form stable atropisomers under ambient conditions. Thus, when the number of C-PCBs is taken into account, PCBs really make up a class of compounds consisting of 228, and not 209, individual chemicals.

Fig. 1
figure 1

Structure of the two rotational PCB 136 isomers (or atropisomers) that are non-superimposable mirror images of each other. Note: Although the optical rotation (Haglund 1996b) and the elution orders of both atropisomers on different enantioselective gas chromatography columns (Haglund and Wiberg 1996) have been established, the absolute configuration of PCB 136 (and most other C-PCBs) has not been established

Full size image

PCBs were manufactured by batch chlorination of biphenyl and used as complex mixtures containing over 150 of the 209 possible PCB congeners (Boonyathumanondh et al. 1995; Driss et al. 1989; Falandysz et al. 2004; Frame 1997; Schulz et al. 1989; Taniyasu et al. 2003). Technical PCB mixtures were sold under many different trade names around the world, for example as Aroclor in the USA or as Kaneclor in Japan (US Environmental Protection Agency 2013). The USA (de Voogt and Brinkman 1989), Slovakia (Holoubek 2006), and Russia (Arctic Monitoring and Assessment Programme 2003) were major manufacturers of PCBs for a large range of consumer and industrial applications, followed by Germany (Fiedler 1997) and Japan (de Voogt and Brinkman 1989). China (Xing et al. 2005) and Poland (Sulkowski et al. 2003) also produced considerable amounts of PCBs. Many other countries, including France and North Korea, produced PCBs; however, total production numbers for these countries are not available. An estimated 500,000 tons of PCBs were manufactured worldwide (estimated based on the following references: Arctic Monitoring and Assessment Programme 2003; de Voogt and Brinkman 1989; Fiedler 1997; Holoubek 2006; Sulkowski et al. 2003; US Environmental Protection Agency 2013; Xing et al. 2005). In addition, PCBs can be formed inadvertently by industrial processes, which represent a contemporary source of certain PCB congeners. For example, PCBs are present in paint pigments from several manufactures in the USA and Japan (Anezaki and Nakano 2014; Hu et al. 2010).

Based on published production data (Arctic Monitoring and Assessment Programme 2003; de Voogt and Brinkman 1989; Fiedler 1997; Holoubek 2006; Sulkowski et al. 2003; US Environmental Protection Agency 2013; Xing et al. 2005) and published congener profiles (Boonyathumanondh et al. 1995; Driss et al. 1989; Falandysz et al. 2004; Frame 1997; Schulz et al. 1989; Taniyasu et al. 2003), C-PCBs constitute 6 %, or 29,500 tons, of the total amount of PCB manufactured worldwide. Depending on the level of chlorination, PCB mixtures can contain 0.15–33 % by weights of C-PCBs. PCB 95 and PCB 149 are the major C-PCB congeners present in technical PCB mixtures, followed by PCB 132, PCB 174, PCB 84, and PCB 45 (Fig. 2). C-PCBs, such as PCB 95, are also present in paint pigments from several manufactures in the USA and Japan (Anezaki and Nakano 2014; Hu et al. 2010). The highest concentration of PCB 95 found in paint pigment was 2.3 ng/g (Anezaki and Nakano 2014). PCB 132 and PCB 149 were also detected in at least one paint pigment sample (Hu and Hornbuckle 2010). PCBs in paint and other building materials, such as caulking and fluorescent light ballasts (US Environmental Protection Agency 2014a, b), contribute to indoor and outdoor air pollution with PCBs (Klosterhaus et al. 2014; Robson et al. 2010). For example, some C-PCBs (i.e., PCBs 84, 91, 95, 132, 149, and 174) have been detected in indoor air in New York City school (Thomas et al. 2012). PCB 95 was found to be one of the highest congeners present in the indoor air, at a concentration of 50 ng/m3. However, diet remains the major route of human PCB exposure (Bergkvist et al. 2014; Chan-Hon-Tong et al. 2013; Kostyniak et al. 2005; Schecter et al. 2010; Su et al. 2012; Voorspoels et al. 2008). Unfortunately, data on C-PCB levels in foodstuff are almost non-existent. For example, a recent study of a wide range of foods from the Dallas metro area reported seven marker PCBs (i.e., PCBs 28, 52, 101, 118, 138, 153, and 180), but no C-PCB levels (Schecter et al. 2010).

Fig. 2
figure 2

Total production numbers for C-PCBs by a country and b C-PCB congener. Data are based on the available production data for the USA, Germany, Russia, Japan, Slovakia, and Poland and published PCB congener profiles for the respective technical mixtures (Arctic Monitoring and Assessment Programme 2003; de Voogt and Brinkman 1989; Fiedler 1997; Holoubek 2006; Sulkowski et al. 2003)

Full size image

C-PCBs present in technical PCB mixtures and synthetic C-PCBs are essentially racemic (i.e., are formed as a 1:1 mixture of atropisomers) (Kania-Korwel et al. 2006; Pakdeesusuk et al. 2003). However, pure atropisomers of C-PCBs can be prepared in a multi-step synthesis (Püttmann et al. 1989; Püttmann et al. 1986) or obtained after enantioselective separation of the racemic mixture using liquid chromatography (Haglund 1996a, b; Mannschreck et al. 1985; Schurig 2001). The availability of small quantities of pure PCB atropisomers allowed the determination of the optical rotation (i.e., (+)- vs. (-)-enantiomer) (Haglund 1996b), the elution order on different enantioselective gas chromatography columns (Haglund and Wiberg 1996), and the absolute configuration (i.e., R- vs. S-enantiomer) of some, but not all, C-PCBs (Toda et al. 2012). Moreover, the metabolites of C-PCBs are also chiral due to the hindered rotation around the central C–C bond (Nezel et al. 1997), and—as with the parent compounds—can be separated into their atropisomers using liquid chromatography (Pham-Tuan et al. 2005; Zhai et al. 2013) or gas chromatography (Kania-Korwel et al. 2011; Kania-Korwel et al. 2008c). The absolute configuration of several C-PCB metabolites has been determined using a combination of experimental and computational techniques (Pham-Tuan et al. 2005).

C-PCBs can display considerable atropisomeric enrichment in the environment (Lehmler et al. 2009). The atropisomeric enrichment of C-PCBs, especially at higher trophic levels, raises the question if C-PCBs display atropselective biological effects in ecosystems and, ultimately, humans. A particular concern is the developmental neurotoxicity of PCBs, which has been documented in epidemiological (Carpenter 2006; Korrick and Sagiv 2008; Schantz et al. 2003; Seegal 1996; Winneke 2011) and laboratory studies (Mariussen and Fonnum 2006; Seegal 1996; Tilson and Kodavanti 1998). Mechanistic studies suggest that PCB developmental neurotoxicity is mediated by mechanisms independent of aryl hydrocarbon (AhR) activation and likely involves impaired Ca2+ signaling (reviewed in: Kodavanti and Curras-Collazo 2010; Mariussen and Fonnum 2006; Pessah et al. 2010). Recent studies with pure C-PCB atropisomers showed that (-)-PCB 136 activates ryanodine receptors (RyRs) (Pessah et al. 2009), microsomal Ca2+ channels that are widely expressed in the mammalian brain and skeletal muscle. Moreover, (-)-PCB 136 enhances dendritic growth in primary cultured rat hippocampal neurons by mechanisms involving RyRs (Yang et al. 2014). In contrast, (+)-PCB 136 lacked activity towards RyRs and had no effect on dendritic growth. PCB 84 also atropselectively impaired Ca2+ signaling, as determined using [3H]-phorobol ester binding and microsomal 45Ca2+ uptake (Lehmler et al. 2005). In addition to their effects on Ca2+ signaling, C-PCBs atropselectively alter the expression of hepatic xenobiotic-metabolizing enzymes, including cytochrome P450 enzymes (Püttmann et al. 1990; Püttmann et al. 1989; Rodman et al. 1991). According to recent experimental evidence, hydroxylated PCB (HO-PCB) metabolites of C-PCBs cause muscle dysfunction in cells in culture by mechanisms involving RyRs and, thus, are also toxic (Niknam et al. 2013; Pessah et al. 2006).

The available evidence suggests that exposure to atropisomerically enriched C-PCBs and/or their atropselective disposition plays a poorly understood role in the adverse effects on PCBs in wildlife, laboratory animals, and, ultimately, humans. Therefore, the objective of this review is to summarize our current knowledge of the processes involved in the absorption, metabolism, and excretion of C-PCBs and their metabolites in laboratory animals and humans. Although C-PCBs and their metabolites are the primary focus of this review, findings from studies of structurally related PCB congeners are discussed when no or limited data for C-PCBs and/or their metabolites are available.

Absorption of C-PCBs

Oral absorption is an important and well-understood route for systemic uptake of PCBs following PCB ingestion because diet, in particular fish, represents a major route of human exposure to PCBs (Bergkvist et al. 2014; Chan-Hon-Tong et al. 2013; Kostyniak et al. 2005; Schecter et al. 2010; Su et al. 2012; Voorspoels et al. 2008). To enter the systemic circulation, PCBs must diffuse from the bulk intestinal content to the intestinal wall (for a review of intestinal absorption and biomagnification models, see Kelly et al. 2004). Since PCBs enter the systemic circulation of several species, including ring doves (Drouillard and Norstrom 2000), dogs (Busbee et al. 1985), and humans (Kuwabara et al. 1979), at rates similar to those observed for dietary lipids, PCBs are thought to move to the intestinal wall dissolved in mixed micelles (i.e., micelles containing bile salts, lipids and lipid digestive products) (Dulfer et al. 1998; Dulfer et al. 1996; Oomen et al. 2001) prior to entering the enterocyte by passive diffusion (Dulfer et al. 1998). As a result, oral absorption of PCBs is highly efficient; however, higher chlorinated congeners are less efficiently absorbed than lower chlorinated ones in most species, including fish (Gobas et al. 1988), ring doves (Drouillard and Norstrom 2000), rats (Matthews and Tuey 1980; Tanabe and Tatsukawa 1981), cows (McLachlan 1993), and humans (Jödicke et al. 1992), probably due to their lower solubility in the interior of mixed micelles (Dulfer and Govers 1995). Although this has not been investigated, the oral absorption efficiency of C-PCBs may, therefore, decrease with increasing degree of chlorination.

Dermal exposure represents an important occupational exposure, especially in workers directly exposed to technical PCB mixtures (Lees et al. 1987). This route of exposure contributes to increased PCB levels in monkeys (Mayes et al. 2002) and humans (Lees et al. 1987; Maroni et al. 1981a; Maroni et al. 1981b) and causes adverse human health effects (Maroni et al. 1981a; Maroni et al. 1981b) due to PCBs’ high affinity for skin (Garner et al. 2006). For example, PCB 136, an important C-PCB, readily diffuses into rat skin slices, with almost 90 % of PCB 136 being absorbed by skin slices within 8 h (Garner et al. 2006). Toxicokinetic studies comparing the dermal vs. oral routes of exposure demonstrate that dermal exposure increases the time to reach the maximum PCB concentration with increasing degree of chlorination (Garner et al. 2006). In contrast, the maximum PCB concentration in blood occurs within 3–4 h after oral administration of PCBs to rats, independent of the PCB congener. This is due to the retention of higher chlorinated and more lipophilic PCB congeners in fat depots of the skin.

There is currently growing concern about inhalation as a route of PCB exposure, especially in susceptible populations such as children, because of PCBs’ presence in indoor and outdoor air (Currado and Harrad 1998; Currado and Harrad 2000; Harrad et al. 2006; Hu et al. 2012a; Jamshidi et al. 2007; Robson and Harrad 2004). The observation that certain C-PCBs have considerable concentrations in indoor air from US schools represents a particular public health concern (Thomas et al. 2012). Recent nose-only inhalation studies demonstrate the PCBs, including some C-PCBs, are absorbed in the lung and distributed systemically in rats (Hu et al. 2010; Hu et al. 2012b). The uptake of PCBs in the lung appears to be highly efficient, with a 99.8 % pulmonary uptake of an inhaled, radiolabeled PCB reported in rats (Hu et al. 2014). Together, these studies demonstrate that inhalation is an important, but currently overlooked, route of exposure to PCBs.

Experimental evidence suggests that the absorption of PCB, including C-PCBs, following oral, dermal, and inhalation exposure occurs by passive diffusion and, thus, should not be atropselective. It is possible that, in addition to passive diffusion, active transport processes play a role in the absorption of C-PCB in the gastrointestinal tract and, thus, contribute to the atropisomeric enrichment of C-PCBs observed in vivo; however, animal studies do not support the hypothesis that PCBs are substrates for active transporters, such as multidrug resistance transporters (Milanowski et al. 2010; Tampal et al. 2003). It is more likely that atropselective metabolism at the site of PCB absorption contributes to the atropisomeric enrichment of C-PCBs. Indeed, skin is able to rapidly metabolize C-PCBs, such as PCB 136 (Garner et al. 2006). Moreover, the intestine, lung, and other extrahepatic tissues express P450 enzymes in different species (Ding and Kaminsky 2003; Renaud et al. 2011) and, thus, may be an important site of atropselective PCB metabolism. Further studies are therefore needed to assess the contribution of extrahepatic tissues to the disposition of C-PCBs.

Metabolism of C-PCBs to HO-PCBs

Structure-activity relationships and HO-PCB profiles of C-PCB metabolism

PCBs are typically thought of as persistent organic pollutants; however, many PCB congeners, including C-PCBs, are biotransformed through complex, species-dependent metabolic pathway to hydroxylated, methylsulfonylated, sulfated, glucuronidated, and other metabolites (James 2001; Letcher et al. 2000) (Fig. 3). A convenient nomenclature for PCB metabolites has been proposed by Maervoet et al. and, as shown for PCB 136 in Fig. 3, will be used throughout this manuscript (Maervoet et al. 2004). The oxidation of C-PCBs, in particular those with a 2,3,6-trichloro substitution pattern in one phenyl ring, to HO-PCBs has been studied extensively using recombinant enzymes (Lu et al. 2013; Lu and Wong 2011; Waller et al. 1999; Warner et al. 2009), hepatic microsomes (Kania-Korwel et al. 2011; Kania-Korwel and Lehmler 2013; Schnellmann et al. 1983; Wu et al. 2014; Wu et al. 2011), isolated hepatocytes (Vickers et al. 1986) and, liver, hippocampus, and skin slices (Garner et al. 2006; Wu et al. 2013a; Wu et al. 2013b) obtained from mammalian species. To the best of our knowledge, the oxidation of C-PCBs in non-mammalian species, such as amphibians, fish, or avian species, has not been investigated to date. Indirect evidence, for example from toxicokinetic studies, suggests that C-PCBs are also subject to biotransformation in many non-mammalian species.

Fig. 3
figure 3

Simplified metabolism scheme of PCB 136 atropisomers (only one atropisomer is shown for clarity reasons): PCB 136 can be oxidized by direct insertion of oxygen into a meta C–H bond or form epoxide intermediates. PCB 136 epoxides can either rearrange to form HO-PCB, undergo 1,2-shift to form 3-HO-PCB 150, from an dihydrodiol epoxide and react with cellular nucleophiles, such as glutathione (GSH). HO-PCBs can be further oxidized to dihydroxylated metabolites, such as 4,5-HO-PCB 136. Alternatively, mono- and dihydroxylated PCB 136 metabolites may undergo phase II metabolism to glucuronide or sulfate conjugates. Similar to other dihydroxylated PCB metabolites, 4,5-HO-PCB 136 can likely be oxidized via a semiquinone radical to PCBs quinones, which subsequently will react with cellular nucleophiles via a chlorine displacement reaction. Glutathione conjugates formed from PCB 136 epoxides will be metabolized in several steps to the corresponding methylsulfonyl metabolites. Metabolites shown in black have been detected in in vitro and/or in vivo studies, whereas experimental evidence suggests that metabolites shown in blue are likely formed in mammals. For details regarding the PCB metabolite nomenclature, see Maervoet et al. (2004). Glc glucuronide, GS glutathione residue, GST glutathione transferase, Nuc cellular N- or S-nucleophile, P450 cytochrome P450 enzyme, SULT sulfotransferase, UGT UDP glucuronosyl transferase

Full size image

Detailed kinetic studies for the oxidation of C-PCBs by P450 enzymes have been reported only for racemic PCB 136 in pooled human liver microsomes (HLMs; pool of three) (Schnellmann et al. 1983) and rat hepatocytes (Vickers et al. 1986). Schnellmann et al. reported that PCB 136 is metabolized by HLMs with a K m of 20.2 mM and a V max of 19.3 pmoles/min/nmoles P450 (Schnellmann et al. 1983). Metabolism studies with PCB 136 in isolated rat hepatocytes suggested the presence of two different metabolism reactions (Vickers et al. 1986). The high-affinity reaction had an apparent K m of 0.48 μM and a V max of 3.7 pmoles/min/mg hepatocyte. The low-affinity reaction had an apparent K m of 645 μM and a V max of 367 pmoles/min/mg hepatocyte. Several recent studies using pure PCB 132 and 136 atropisomers suggest that the respective atropisomers interfere with each other’s metabolism by CYP2B1 (Lu et al. 2013; Lu and Wong 2011), and further studies of the enzyme kinetics of pure PCB atropisomers are needed.

A simplified outline for the metabolism of PCB 136 is shown in Fig. 3. Briefly, the oxidation reaction occurs either by direct insertion of oxygen into an aromatic C–H bond (Preston et al. 1983) or via an epoxide intermediate (Forgue and Allen 1982; Forgue et al. 1979). The 4,5-epoxide intermediate of C-PCBs with a 2,3,6-trichloro substitution pattern can rearrange to form both 4- or 5-hydroxylated PCB metabolites. In addition, a 1,2-shift (or NIH shift) (Guroff et al. 1967; Jerina and Daly 1974) of the 3-chloro substituent can lead to the formation of HO-PCBs with a 2,4,6-trichloro-3-hydroxy substitution pattern as minor metabolites (Kania-Korwel et al. 2012; Kania-Korwel et al. 2011; Wu et al. 2013a; Wu et al. 2014). Studies with recombinant P450 enzymes demonstrate that CYP2B isoforms, such as rat CYP2B1 (Lu et al. 2013; Waller et al. 1999; Warner et al. 2009), human CYP2B6 (Warner et al. 2009), and dog CYP2B11 (Waller et al. 1999), are involved in the formation of meta-hydroxylated C-PCBs. The P450 isoforms responsible for the formation of para-hydroxylated metabolites of C-PCBs have not been identified to date. Many of the mono- and dihydroxylated metabolites of C-PCBs reported in vitro are also detected in vivo (Table 1).

Table 1 Hydroxylated PCB metabolites detected in different animal species
Full size table

Species differences in the metabolism rates and metabolite profiles have been observed in many studies investigating the P450 enzyme-mediated oxidation of C-PCBs to HO-PCBs (Ohta et al. 2013; Ohta et al. 2005; Waller et al. 1999; Warner et al. 2009; Wu et al. 2014). Several factors contribute to these differences, for example species differences in the structure of relevant P450 enzymes (Waller et al. 1999) and the hepatic P450 enzyme and isoform composition (Lewis et al. 1998). Studies with liver microsomes reveal that 5-HO-PCB 136 is formed by liver microsomes with a rank order of rat > dog ~ guinea pig ~ hamster > human > monkey ~ mouse ~ rabbit (Fig. 4) (Wu et al. 2014). In the same study, 4-HO-PCB 136 is formed in the order human > monkey > rabbit ~ rat > dog ~ guinea pig ~ mouse ~ hamster. 5-HO-PCB 136 is the major metabolite formed in HLMs (pool from 50 individual donors), with the 5-HO-PCB 136/4-HO-PCB 136 metabolite ratio ranging from 0.4:1 to 0.8:1. An earlier study reports a different metabolite profile in incubations with HLMs from a pool of three donors, with 4-HO-PCB 136 being the major metabolite (5-HO-PCB 136/4-HO-PCB 136 metabolite ratio = 1.3) (Schnellmann et al. 1983). The differences in the HO-PCB profiles are likely due to differences in the hepatic P450 enzyme composition. Consistent with this explanation, Ohta et al. report that the formation of HO-PCBs by individual HLMs depends on the P450 isoform composition and correlates with hepatic CYP2B6 activity (Ohta et al. 2013).

Fig. 4
figure 4

The P450 enzyme-mediated formation of 5-HO-PCB 136 and 4-HO-PCB 136 from PCB 136 is highly species dependent (Wu et al. 2014). Liver microsomes prepared from different species were incubated with racemic PCB 136, and relative rates of metabolism, adjusted for total P450 content in the respective microsomal preparation, were determined for 5-HO-PCB 136 and 4-HO-PCB 136. Different letters indicate significant differences in the relative rates of metabolism for 5-HO-PCB 136 or 4-HO-PCB 136 (p < 0.05). Data represents the mean ± one standard deviation of three microsomal incubations. For details regarding the PCB metabolite nomenclature, see Maervoet et al. (2004). Reprinted with permission from Wu et al. (2014). Copyright 2014, American Chemical Society

Full size image

In addition to species differences, mutations of a P450 isoform can affect the rates and regioselectivity of the oxidation of C-PCBs. A study with E. coli-expressed rat CYP2B1 and dog CYP2B11 mutants demonstrates that the metabolism of PCB 136 to 5-HO-PCB 136 by P450 mutants is slower compared to the wild-type P450 enzyme (Waller et al. 1999). The I480V mutant is an exception and increases the 5-HO-PCB 136 formation approximately 1.7-fold compared to the CYP2B1 wild type. Interestingly, CYP2B1 mutant I114V and the CYP2B11 mutant L363V display an altered regioselectivity and also form 4-HO-PCB 136. The differences in the metabolite profiles and the altered metabolism rates in the CYP2B1 and CYP2B11 mutants can be attributed to a different orientation adopted by PCB 136 in the active site of the respective mutant. These observations are important because human CYP2B6, the human orthologue of CYP2B1 and CYP2B11, is a highly polymorphic enzyme (Zanger et al. 2007). Similar to chlorpyrifos (Crane et al. 2012), C-PCB metabolism rates and metabolite profiles may therefore depend on the CYP2B6 genotype. Since PCB 136 binds atropselectively to rat microsomal CYP2B enzymes (Kania-Korwel et al. 2008b), site mutants of P450 enzymes may also metabolize PCB 136 atropisomers with different selectivity compared to the wild-type enzyme. Further studies are needed to assess if chiral signatures of PCBs and HO-PCBs are affected by P450 polymorphisms.

Atropselective oxidation of C-PCBs and their metabolites

Several studies using recombinant P450 enzymes (Lu et al. 2013; Warner et al. 2009), liver microsomes (Kania-Korwel et al. 2011; Kania-Korwel and Lehmler 2013; Wu et al. 2014; Wu et al. 2011), or liver tissue slices (Wu et al. 2013a; Wu et al. 2013b) demonstrate that PCBs are atropselectively metabolized to HO-CPBs by P450 enzymes, thus resulting in an atropisomeric enrichment of PCBs and HO-PCBs in the microsomal incubation (Table 2). The atropselective biotransformation of C-PCBs to HO-PCB metabolites by recombinant rat and human cytochrome P450 enzymes has been studied in some detail (Lu et al. 2013; Warner et al. 2009). Incubations of individual, racemic PCB congeners or a mixture of several C-PCBs with rat CYP2B1 show preferential transformation of E2-PCB 45, (-)-PCB 84, E1-PCB 91, E2-PCB 95, (-)-PCB 132, (+)-PCB 136 and (-)-PCB 149 (E1 and E2 denotes the atropisomers eluting first and second on the respective enantioselective column). The same direction of atropisomeric enrichment is observed in rat liver microsomal incubations (Kania-Korwel et al. 2011; Kania-Korwel and Lehmler 2013; Wu et al. 2011) and, for PCB 136, in rat liver tissue slices (Wu et al. 2013b). At the same time, the formation of HO-PCB metabolites by rat P450 enzymes is atropselective, thus resulting in an atropisomeric enrichment of the HO-PCBs. Although all C-PCBs investigated are atropselectively metabolized by rat P450 enzymes to HO-PCBs, the PCB metabolism rates and HO-PCB formation rates are highly congener specific, which may, at least in part, explain the differences in the extent and direction of the atropisomeric enrichment of different C-PCBs in wildlife (Lehmler et al. 2009).

Table 2 Direction of the atropisomeric enrichment of HO-PCBs formed from C-PCBs with a 2,3,6-trichloro substitution pattern
Full size table

The chiral signatures observed in in vitro studies predict the direction of the atropisomeric enrichment of C-PCBs in vivo. Specifically, the observation that (+)-PCB 136 is more rapidly biotransformed by CYP2B1 is in agreement with a study showing an enantiomeric enrichment of (-)-PCB 136 in male and female Sprague-Dawley rats (Kania-Korwel et al. 2008c). Moreover, the atropisomeric enrichment of E2-PCB 95 in the in vitro metabolism experiments is consistent with the atropisomeric enrichment of PCB 95 observed in rats treated with an Aroclor 1254 mixture (Kania-Korwel et al. 2006). Similarly, in vitro metabolism studies (Wu et al. 2013a) accurately predict the direction of the atropisomeric enrichment of PCB 136 (Kania-Korwel et al. 2010; Kania-Korwel et al. 2008a; Kania-Korwel et al. 2007; Kania-Korwel et al. 2008d; Milanowski et al. 2010) and PCB 95 (Kania-Korwel et al. 2012; Kania-Korwel et al. 2010; Milanowski et al. 2010) in mice in vivo. Incubations with human CYP2B6 result in the biotransformation of PCBs 45, 91, and 132; however, only the biotransformation of PCBs 45 and 132 is enantioselective. Interestingly, (-)-PCB 132 is more rapidly biotransformed by CYP2B6, which is consistent with the enrichment of (+)-PCB 132 in human liver (Chu et al. 2003) and breast milk (Bordajandi et al. 2008).

Hepatic microsomes prepared from male Sprague-Dawley rats pretreated with phenobarbital (PB, CYP2B inducer) were used to identify which HO-PCB atropisomers form which PCB 136 atropisomer based on their gas chromatographic elution order (Fig. 5) (Wu et al. 2011). (-)-PCB 136 elutes first on Chirasil-Dex and Cyclosil-B columns (Kania-Korwel and Lehmler 2013). Both HO-PCB metabolites of (-)-PCB 136 also elute first (E1) on both enantioselective columns. Vice versa, metabolites formed from (+)-PCB 136 corresponded to the second eluting atropisomer (E2) of 5-HO-PCB 136 and 4-HO-PCB 136. In parallel incubations with racemic PCB 136, the second elution 5-HO-PCB 136 atropisomer (E2-5-136; Chirasil-Dex column) and the first eluting 4-HO-PCB 136 atropisomer (E1-4-HO-PCB 136; Cyclosil-B column) are enriched in microsomal incubations using racemic PCB 136, independent of the microsomal formulation, PCB 136 concentration, and time point analyzed. These observations demonstrate that the metabolism of (+)-PCB 136 to E2-5-HO-PCB 136 in microsomal incubations is faster compared to the metabolism of (-)-PCB 136 to E2-5-HO-PCB 136. Since 5-HO-PCB 136 is the major metabolite, this observation is consistent with an atropisomeric enrichment of (-)-PCB 136 in in vitro (Lu et al. 2013; Wu et al. 2013b; Wu et al. 2011) and in vivo metabolism studies in rats (Kania-Korwel et al. 2008c).

Fig. 5
figure 5

Metabolism studies with pure PCB 136 atropisomers demonstrate that, relative to a the racemic standard, b (-)-PCB 136 is metabolized by microsomal P450 enzymes to the first eluting atropisomers (E1) of 5-HO-PCB 136 and 4-HO-PCB 136, whereas c (+)-PCB is metabolize to the second eluting atropisomers (E2) of 5-HO-PCB 136 and 4-HO-PCB 136 (Wu et al. 2011). Microsomal incubations of racemic PCB 136 result in the preferential formation on E2-5-HO-PCB 136 and E1-4-HO-PCB 136 using rat liver microsomes from male rats pretreated with d phenobarbital (PB), e dexamethasone (DEX), or f corn oil alone (VEH). Microsomal incubations were optimized for protein, time, and NADPH concentration and performed with 50 μM (-)-, (+)- or racemic PCB 136 at 37 °C for 10 min. Atropisomers of 5-HO-PCB 136 and 4-HO-PCB 136 were separated on commercially available Chirasil-Dex and Cyclosil-B columns, respectively, as described previously (Kania-Korwel et al. 2008c). For details regarding the PCB metabolite nomenclature, see Maervoet et al. (2004). Reprinted with permission from Wu et al. (2011). Copyright 2011, American Chemical Society

Full size image

Several recent metabolism studies demonstrate that the atropselective formation of HO-PCBs is also species dependent (Kania-Korwel et al. 2008b; Wu et al. 2013a; Wu et al. 2014; Wu et al. 2013b; Wu et al. 2011). For example, both 5-HO-PCB136 and 4-HO-PCB136 are atropselectively formed by liver microsomes from humans, dogs, monkeys, guinea pigs, mice, hamsters, and rabbits (Fig. 6). While the extent of the atropisomeric enrichment of both HO-PCB 136 metabolites differs significantly between species, the direction of the atropisomeric enrichment is similar in many species. Specifically, E2-5-HO-PCB136 and E2-4-HO-PCB136 are formed preferentially from racemic PCB 136 in incubations using liver microsomes from different, toxicologically relevant species. Both metabolites are formed from (+)-PCB 136. The mice is a notable exception, with E1-5-HO-PCB136 and E1-4-HO-PCB136 formed preferentially in vitro (Wu et al. 2013a; Wu et al. 2014). In agreement with this finding, mice display an enrichment of (+)-PCB 136 in all tissues and excreta (Kania-Korwel et al. 2010; Kania-Korwel et al. 2008a; Kania-Korwel et al. 2007; Kania-Korwel et al. 2008d; Milanowski et al. 2010).

Fig. 6
figure 6

The direction and extant of the atropselective metabolism of PCB 136 by cytochrome P450 enzymes is highly species dependent (Wu et al. 2014). Representative chromatograms of a 5-HO-PCB 136 and b 4-HO-PCB 136 (analyzed as the corresponding methoxy derivatives) formed by incubations of liver microsomes from several toxicologically relevant species, including microsomes from humans (pooled HLMs), dogs, monkeys, guinea pigs, mice, hamsters, and rabbits. The racemic composition of the respective authentic standard is shown for comparison. 5-HO-PCB 136 and 4-HO-PCB 136 were separated on Chirasil-Dex and Cyclosil-B columns, respectively. For details regarding the PCB metabolite nomenclature, see Maervoet et al. (2004). Reprinted with permission from Wu et al. (2014). Copyright 2014, American Chemical Society

Full size image

Atropselective metabolism of HO-PCBs

HO-PCBs themselves can be further metabolized to a myriad of different metabolites (Fig. 3). For example, P450 enzymes can oxidize monohydroxylated C-PCBs to dihydroxylated C-PCBs (Lu et al. 2013; Waller et al. 1999). Studies with recombinant rat CYP2B1 reveal that the formation of 4,5-HO-PCB 95 from 5-HO-PCB 95 and 4-HO-PCB 95 is both regio- and atropselective (Fig. 7) (Lu et al. 2013). Specifically, the 5-HO-PCB 95 regioisomer has a higher relative metabolism rate than the 4-HO-PCB 95 regioisomer. Moreover, the extent, but not the direction of the atropisomeric enrichment of 4,5-HO-PCB 95, is different between 5-HO-PCB 95 and 4-HO-PCB 95, with a more pronounced enrichment of 4,5-HO-PCB 95 observed in incubations with 5-HO-PCB 95. In contrast, an in vivo disposition study in mice reports near racemic chiral signatures for 4,5-HO-PCB 95 (Kania-Korwel et al. 2012). As shown in Fig. 3, dihydroxylated C-PCBs may subsequently undergo autooxidation and enzymatic oxidation to PCB quinones (Amaro et al. 1996). Afterwards, the reactive quinone intermediates can react with cellular nucleophiles in a non-enzymatic chlorine displacement reaction (Song et al. 2009). In addition to the corresponding PCB 136 epoxides, the proposed PCB 136 quinone intermediate may contribute to the microsomal protein binding observed in HLM incubations with racemic PCB 136 (Schnellmann et al. 1983).

Fig. 7
figure 7

Monohydroxylated metabolites of C-PCBs, such as 4-HO-PCB 95 and 5-HO-PCB 95, are oxidized stereoselectively to dihydroxylated metabolites, such as 4,5-HO-PCB 95, by recombinant rat CYP2B1 (Lu et al. 2013). The different biotransformation rates result in an atropisomeric enrichment of the parent HO-PCB congener and the 4,5-HO-PCB 95 metabolites. A1 Biotransformation activities of 4-HO-PCB 95 and 5-HO-PCB 95, A2 enantiomeric fractions of 4-HO-PCB 95 and 5-HO-PCB 95, B1 formation activities of 4,5-HO-PCB 95 from 4-HO-PCB 95 and 5-HO-PCB 95, and B2 enantiomeric fraction of 4,5-HO-PCB 95 formed from 4-HO-PCB 95 and 5-HO-PCB 95, respectively, in the incubation of rat CYP2B1 after a 60-min incubation time. Values are the mean ± standard deviation of three replicates. *p < 0.05, **p < 0.01, unpaired, one-tailed t test. For details regarding the PCB metabolite nomenclature, see Maervoet et al. (2004). Reprinted with permission from Lu et al. (2013). Copyright 2013, American Chemical Society

Full size image

There is some experimental evidence that C-PCBs are metabolized to glucuronide conjugates in rats. For example, the disposition of chiral catechol PCB metabolites differs between Wistar and Gunn rats (Haraguchi et al. 2004). Specifically, the dihydroxylated metabolites are more persistent in Gunn than Wistar rats, most likely due to the uridine diphosphate glucuronosyltransferase (UGT) deficiency in Gunn rats. There is also evidence that PCB 136 glucuronides accumulate in the fetal compartment of pregnant rats (Lucier et al. 1978). Moreover, a growing number of in vitro and in vivo studies demonstrate that HO-PCBs can be further metabolized to PCB sulfates; however, the formation of C-PCB sulfates in vivo has not been investigated to date. It is also unknown if the metabolism of HO-PCBs to sulfate or glucuronide conjugates, or other metabolites, is atropselective. Unfortunately, analytical standards and straightforward analytical methods for studies of C-PCB sulfate or glucuronide conjugates in vitro or in vivo are currently not available.

Atropselective formation of methylsulfone metabolites of C-PCBs

In addition to the direct insertion of an oxygen atom in the meta position, PCBs can also form reactive PCB epoxide intermediates that undergo glutathione conjugation, followed by stepwise transformation to mercapturic acid and, ultimately, methylsulfone derivatives of PCBs (MeSO2-PCBs) (Bakke et al. 1982; Letcher et al. 2000). Para-substituted MeSO2-PCBs, such as 4-MeSO2-PCB 91, 4’- MeSO2-PCB 132, and 4-MeSO2-PCB 149, are typically present at higher concentrations in the liver, lung, and adipose tissues of wildlife and laboratory animals than the corresponding meta-substituted MeSO2-PCBs (Karasek et al. 2004; Larsson et al. 2002). Like the parent C-PCBs, MeSO2-PCB metabolites and the respective intermediates also display axial chirality, can be separated into the respective atropisomers (Pham-Tuan et al. 2005), and undergo atropisomeric enrichment in vivo. Some chiral MeSO2-PCB congeners, such as 4-MeSO2-PCB 149, 5-MeSO2-PCB 149, 4′-MeSO2-PCB 149, and 5′-MeSO2-PCB 132, display atropisomeric enrichment in human postmortem liver (Ellerichmann et al. 1998) and wildlife tissues (Karasek et al. 2004; Karasek et al. 2007; Larsson et al. 1999; Larsson et al. 2002). The atropisomeric enrichment of para-substituted MeSO2-PCBs is typically not as pronounced as that of meta-substituted MeSO2-PCBs (Ellerichmann et al. 1998; Larsson et al. 1999). In addition to these environmental studies, the R-enantiomer of 4′- and 5′- MeSO2-PCB 132 is formed almost atropselectively in male rats following a single oral administration of racemic PCB 132 (Norström et al. 2006). The processes resulting in the atropisomeric enrichment of MeSO2-PCBs are poorly understood. Limited evidence demonstrating the atropselective biotransformation of PCB 149 to 5-S-MeSO2-PCBs 149 by rat hepatocytes suggests a role of P450 enzymes in the atropisomeric enrichment of MeSO2-PCBs in vivo (Hühnerfuss et al. 2003). Moreover, the toxicological relevance of the atropisomeric enrichment of MeSO2-PCBs remains unexplored.

Excretion of C-PCBs and their metabolites

Feces and urine are only minor routes of excretion for parent PCBs. For example, in the case of PCB 136, only a small percentage of the parent compound is excreted with feces and urine in the rat (Birnbaum 1983), dog (Sipes et al. 1982), and monkey (Sipes et al. 1982). Instead, fecal and urinary excretion of PCB 136 occurs primarily in form of its metabolites (Birnbaum 1983; Sipes et al. 1982). Two processes contribute to the fecal excretion of PCBs. A small percentage of unabsorbed PCBs are excreted within the first 24 h after PCB administration in rats (Tanabe and Tatsukawa 1981) and mice (Kania-Korwel et al. 2008a). The contribution of unabsorbed PCBs to fecal PCB levels is directly related to the fecal fat content, with higher fecal fat contents increasing the PCB partitioning capacity of feces and, thus, fecal PCB levels (Drouillard and Norstrom 2003; Gobas et al. 1988). A second relevant process is the intestinal elimination of the parent PCB, which also is related to the fecal fat content. Biliary excretion of PCB 136 apparently does not contribute to the excretion of the parent compound (Sipes et al. 1982). These excretion processes occur by passive diffusion and, therefore, are not atropselective. Although PCBs apparently are not substrates for active transporters, such as multidrug resistance transporters (Milanowski et al. 2010; Tampal et al. 2003), in in vivo studies, the possibility that atropselective active transport processes contribute to the excretion of C-PCBs cannot be completely ruled out. Moreover, it is currently unknown if active transporters are involved in the excretion of PCB metabolites.

Several studies have investigated the enantiomeric enrichment of C-PCBs in feces. Mice fed a high-fat diet excreted near racemic PCB 136 in the feces within the first 24 h after oral administration of a bolus dose of racemic PCB 136 (Kania-Korwel et al. 2008a). Twenty-four hours after PCB administration, a significant enrichment of (+)-PCB 136 was observed in the feces from the same animals due to the elimination of PCB 136 from the systemic circulation. The fact that little-to-no enantiomeric enrichment was observed within the first 24 h is due to the 10- to 100-times higher levels of unabsorbed PCBs in feces, which mask the chiral signatures of the PCB eliminated from the systemic circulation. Similarly, Norström and co-workers reported that the enantiomeric enrichment of PCB 132 in feces collected over a 4-day period from male Wistar rats was much less pronounced compared to tissue EF values due to non-absorbed PCB 132 (Norström et al. 2006). Chiral signatures of hydroxylated C-PCBs and their conjugates in excreta have not been investigated to date.

Harrad et al. investigated the enantiomeric enrichment of PCBs 95 and 149, two environmentally relevant PCB congeners, in feces from volunteers eating a diet containing trace levels of racemic PCBs 95 and 149 (Harrad et al. 2006). Only a slight enantiomeric enrichment of PCB 95 was observed in two out of ten feces samples. PCB 149 was only detected in one feces sample. This is a surprising observation because based on microsomal metabolism studies with HLMs (Wu et al. 2014), disposition studies in laboratory animals (Kania-Korwel et al. 2008a; Norström et al. 2006), and chiral signatures in human liver (Chu et al. 2003) and breast milk (Glausch et al. 1995) samples, a more pronounced atropisomeric enrichment would be expected in human feces. Therefore, unabsorbed PCBs are one possible explanation for the near racemic PCBs 95 and 149 residues in the study by Harrad and co-workers.

Conclusion

C-PCBs represent an important group of PCBs that are thought to play an important role in the adverse effects of PCBs, in particular in PCB developmental neurotoxicity. An intriguing yet unanswered question is how the atropselective disposition of C-PCBs and their metabolites affects adverse outcomes in laboratory animals and humans. Specific questions that warrant future attention include the following:

  • The absorption of C-PCBs in the gastrointestinal tract, lung, and skin occurs by passive transport processes and, thus, makes no contribution to the atropisomeric enrichment of C-PCBs observed in vivo. However, it is currently unclear if atropselective metabolism at the site of absorption or in other, extrahepatic tissues contributes to the atropisomeric enrichment of C-PCBs.

  • Hepatic P450 enzymes atropselectively metabolize C-PCBs to HO-PCBs and other metabolites. However, only limited information about the P450 isoforms involved in the atropselective metabolism of C-PCBs is currently available. Moreover, it is unknown how P450 polymorphisms affect the atropselective metabolism of C-PCBs.

  • In vitro studies demonstrate that pure PCB atropisomers can interfere with each other’s metabolism. Further studies are needed to assess how exposure to complex mixtures affects the atropselective disposition of C-PCBs in vivo. Such studies will make a fundamental contribution to our understanding of the chiral signatures observed in wildlife and human samples.

  • C-PCBs can be metabolized via the corresponding HO-PCBs or PCB epoxides to a myriad of other metabolites (Fig. 3). Similar to other PCB congeners, chiral HO-PCBs can be further oxidized to dihydroxylated metabolites and/or metabolized to sulfate or glucuronide conjugates. Moreover, PCB epoxides can form glutathione conjugates that are converted via the mercapturic acid pathway to chiral methylsulfonyl PCBs. Unfortunately, only limited information about the phase II metabolites of C-PCBs is available. Furthermore, it is unknown which atropselective biological processes play a role in the formation, distribution, and elimination of these phase II metabolites.

  • Active transport processes may play a role in atropselective distribution and elimination of C-PCBs and their metabolites and, therefore, warrant further investigation.

  • Metabolites of C-PCBs, such as their HO-PCB and MeSO2-PCB metabolites, are potentially toxic themselves. Because C-PCB metabolites are also chiral, future studies need to asses if—like the parent C-PCBs—their toxicity is atropselective.

  • Authentic standards and validated analytical methods are typically not available to systematically address the above-mentioned questions. Additional efforts are needed to develop novel synthetic strategies for C-PCB metabolites and established methods for their atropselective analysis in different biological matrices.