Selective retention of hydroxylated PCB metabolites in blood.

Polychlorinated biphenyls (PCBs) are important environmental contaminants, and their toxicity to wildlife and humans are of major concern. PCBs form persistent and abundant metabolites, PCB methyl sulfones, that accumulate in biota. We now report that certain hydroxylated PCB metabolites show a strong and selective accumulation in mammalian blood. Plasma from experimentally PCB-dosed rats and blood from environmentally exposed grey seals (Halichoerus grypus) and humans were analyzed. Among all possible hydroxylated metabolites of PCB that may be formed, only a few, dominated by 4-OH-2,3,5,3',4'-pentachlorobiphenyl and 4-OH-2,3,5,6,2',4',5'-heptachlorobiphenyl, were found in the blood samples. All identified compounds have a structure with the hydroxy group in a para or meta position, with chlorine atoms on vicinal carbon atoms. The concentrations of hydroxylated PCB in the blood were almost in the same range as the most persistent PCB congeners both for seals and humans.

The environmental persistence of polychlorinated biphenyls (PCBs) is well known, and it has been proposed that their presence in organisms adversely affects a number of biological systems. Yet little is known about the mechanisms or the agents responsible for these effects, except for PCB congeners with dioxinlike effects that are mediated through binding to the aryl hydrocarbon hydroxylase (Ah) receptor (14). The initial step in the biotransformation of PCB involves cytochrome P450 (CYP lAl, 1A2, and CYP2B1/2B2)-mediated oxidation to arene oxides-intermediates with a limited half-life (5,6). Arene oxides are mainly transformed to hydroxylated aromatic compounds but also to sulfurcontaining metabolites via the mercapturic acid pathway (MAP) (5,7). Halogenated aromatic compounds such as chlorinated biphenyls (CBs) may, depending on the number of halogen substituents and position of the substituents, form more than one arene oxide isomer from each compound.
The metabolism of all different CBs present in the environment will thus result in the formation of a large number of hydroxylated PCB metabolites. Normally, the hydroxylated metabolites are excreted in feces and/or in urine, as such or conjugated to glucuronic acid or sulfate (8). Hydroxylated PCBs were also excreted in feces from environmentally PCB-exposed seals (9). Not all of the phenolic compounds are always excreted but may instead be retained in the body (10,11), either due to their high lipophilicity or reversible binding to proteins. Pentachlorophenol is retained in the blood of mammals and binds to a thyroxin-transporting protein, transthyretin (TTR) (11), and hydroxylated metabolites of 2,5,4'-triCB* have been shown to be localized to intraluminal uterine fluid of pregnant mice (13).
Metabolism studies of individual CBs, e.g., 2,5,4'-triCB (CB-31, rat), 3,4,3',4'-tetraCB (CB-77, rat and mouse), and 2,3,4,3',4'-pentaCB (CB-105, mouse and mink) have shown that several isomers of hydroxylated metabolites, including dihydroxylated and dechlorinated metabolites, are formed (14)(15)(16)(17)(18). Metabolites of the two latter CBs, formed after a 1,2-shift of a chlorine atom, have also been identified. These metabolites were 4-OH-3,5,3',4'-tetraCB from CB-77 (17,19) and 4-OH-2,3,5,3',4'-pentaCB formed from CB-105 (16). In both cases, these metabolites were shown to be retained in the blood. However, CB-77 is present only in trace amounts in commercial PCBs: 0.45% in Aroclor 1242 and not detected in Aroclor 1254 (20). As this CB is rapidly metabolized to several other hydroxylated metabolites, it is doubtful if it is a quantitatively important metabolite. Other PCB congeners present in higher concentrations, such as CB-105 [3.7% in Aroclor 1254 (18)], may be toxicologically more important. There are also a number of CBs that have similar structures as CB-77 and CB-105 and therefore may be metabolized similarly. The major CBs of this type are 2,4,5,3',4'-pentaCB (CB-i 18), 2,3,4,5,3',4-hexaCB (CB-156), and 2,3,4,2',3',4'-hexaCB (CB-128), which are present in Aroclor 1254 in amounts of 6.4%, 1.6%, and 2.1%, respectively (20). This study investigated the general retention pattern of hydroxylated PCB metabolites in blood of rats at three time points after an oral dose ofa commercial PCB *The numbering of the chlorine atoms is not according to the IUPAC rules but was chosen to facilitate understanding of the structures for the reader. The numbering system introduced by Ballschmiter et al. (12) is used for the PCB congeners. product (Aroclor 1254). Blood samples from Baltic grey seals and human plasma samples were also analyzed for possible content of OH-CB. The present work is primarily aimed at structural identification of the hydroxylated CBs retained but includes data on the quantification of a major hydroxylated PCB metabolite in the blood. Ratios between the major OH-CB and a major CB are given.

Materials and Methods
Twelve male Sprague-Dawley rats (150-200 g) were divided into four groups. One group was kept as a control group and dosed with corn oil only. The rats in the other three groups were dosed orally with Aroclor 1254 (25 mg/kg body weight dissolved in 0.2 ml peanut oil) once a day for 3 days. The rats were kept on a 12 hr/ 12 hr light/dark cycle and given food and water ad libitum. The rats in one group were killed 24 hr after the last gavage; the rats in the second group were killed after 7 days, and the rats in the last group were killed after 14 days. We collected blood (plasma), lungs, livers, kidneys, and adipose tissue from all the rats and analyzed them for PCB and potential OH-CBs.
Human plasma samples (six samples) were kindly donated by Danderyds Hospital Blood Donor Centre. The samples were randomly selected and contained plasma from females (20, 20, and 42 years old) and males (23,48, and 53 years old).
Aroclor 1254, a commercial PCB product from Monsanto (Washington, DC USA), was used for the animal experiments. We synthesized 48 methoxy-chlorobiphenyls (MeO-CB) (unpublished), but the majority of MeO-CBs were prepared according to the Cadogan diaryl coupling reaction (21). A few of the MeO-CBs were synthesized via the Ullman diaryl coupling reaction (22) with subsequent chlorination of the MeO-CB product obtained in this coupling reaction (23). Previous synthesis of   (25) and was used as internal standard for the PCB.
MeO-CBs has been described by Jansson and Sundstr6m (24). The standards used are listed in Table 1. Hexane, pesticide grade, was purchased from Fison (Leicestershire, England), analysis-grade methanol from Merck (Darmstadt, Germany), and analysis-grade methyl tertbutyl ether from Ratborm (Walkerburn, Scotland). The sulfuric acid was purchased from BDH (Poole, England) and all other chemicals and solvents were of analysisgrade quality. Diazomethane was used for derivatization and was synthesized as described by Fieser and Fieser (26).
Gas chromatography with electron capture detection (GC-ECD) was performed using a Varian 3400 gas chromatograph on a DB-5 capillary column (30 m x 0.25 mm i.d., 0.25 pm film thickness; J&W Scientific Inc., Folsom, CA). The temperature program was 80°C for 2 min at 10°C/ min to 300°C. Injector and detector temperatures were 250°C and 360°C, respectively. Hydrogen was used as carrier gas, and the injections were made in the splitless mode. The same GC was used with a cyano-derivatized fused silica capillary column (SP-2331, 0.25 mm i.d., Supelco Inc.), and the oven temperature was programmed 80°C for 2 min, 20°C/min to 150°C, 8°C/min to 280°C and hold for 10 min. The injector temperature was 250°C, and the detector temperature 360°C.
Gas chromatography/mass spectrometry was performed on a Finnigan 4021 instrument upgraded with a 4500 ion source connected to an Incos data system. GC was performed on an Ultra 2 fused silica capillary column (50 m X 0.2 mm i.d., 0.33 pm film thickness; Hewlett Packard, Hoofddorp, the Netherlands) with helium as the carrier gas. Injections were made in the splitless mode at an injector temperature of 2600C. The oven temperature was programmed as follows: 70°C for 2 min; 30°C/min to 22°C; and 4°C/min to 300°C. The ion source temperature was 1000C. The MS was operated in the Negative Ion Chemical Ionization (NICI) mode, scanning from 250 to 500 amu and with an electron energy of 125 eV. Methane (>99.95% pure, with <100 ppm 02) was used as the reagent gas.
We diluted the plasma (or blood) with one volume of water and one volume of methanol. The samples were acidified with sulfuric acid (0.5 M) and then extracted with hexane: methyl tert-butyl ether (MTBE; 1:1, one volume) three times. The combined organic phases were concentrated and the lipid content determined gravimetrically. The extracts were redissolved in hexane and partitioned with potassium hydroxide (1 M in 50% ethanol). The alkaline phase was acidified, the phenolic compounds re-extracted in hexane:MTBE, the solvent evaporated, and the residue dissolved in hexane before treatment with diazomethane. The coextracted lipids were removed by treatment with concentrated sulfuric acid. Finally, the samples were purified on an alumina oxide column (3 g, neutral, washed with 30 ml hexane before use) using hexane as mobile phase (30 ml). The hexane phase from the partitioning with alkali, contain-ing neutral compounds such as PCB, was also treated with concentrated sulfuric acid and purified on an alumina oxide column. We analyzed the samples by GC-ECD individually, but by GC/MS as pooled samples. Internal standards, 4-OH-2,3,5,6,3',4',5'-heptaCB and CB-189, were added to the plasma and blood samples before extraction and to tissues after extraction. The cleanup of tissues was performed as described by Bergman et al. (27. The methylated phenolic samples were analyzed for the presence of potential MeO-CBs and compared to the authentic standards. The neutral fraction was analyzed for PCB and the total PCB concentration (determined by quantification of the major PCB congeners; 2,4,5,2',4',5'-hexaCB, 2,3,4,2',4',5'-hexaCB, 2,3,4,5, 2',4',5'-heptaCB and 2,3,4,5,2',3',4'-heptaCB) was determined in the human and seal samples and the persistent congener 2,4,5,2',4',5'-hexaCB determined in all samples for comparison with individual MeO-CBs.
The recovery of 4-OH-3,5,2',3',4'-pentaCB and 4-OH-2,3,5,6,3',4',5'-heptaCBs was determined by adding these compounds to human plasma (5 ml) to give concentrations at 5 ng/ml plasma and 40 ng/ml plasma in five parallel experiments. The samples were treated by the same method as described above and the compounds quantified by calculating the ratio to an added volumetric internal standard and comparing to a standard mixture. The blank plasma samples contained no peaks at the retention times for the tested MeO-CBs or for the surrogate. The concentration of the added 4-OH-3,5,2',3',4'-pentaCB was so much higher (250 times) than OH-CBs in the sample that it was negligible for quantification purposes. We checked the recovery of CB-189 (10 ng was added) in a similar way.
In the rat plasma, a total of 13 OH-CBs were determined by GC/MS with the major compound identified as the methyl derivative of 4-OH-2,3,5,3',4'-pentaCB. The 4-MeO-2,3,5,3',4'-pentaCB and 4-MeO-3,5,2',3',4'-pentaCB, have identical mass spectra obtained by electron ionization (El) and NICI and similar retention times on a DB-5 capillary column, but they do separate on the cyano-derivatized column also used for analyses. The major OH-CB in the rat plasma is therefore identified as 4-OH-2,3,5,3',4'-pentaCB. After comparison of the samples with the synthesized MeO-CB standards on two different GC columns, seven metabolites were identified. The structure of the identified compounds are shown in the chromatogram in Figure la.
In Table 2, the concentrations of 4-OH-2,3,5,3',4'-pentaCB and CB-153 in the tissues are given, together with the ratio of OH-pentaCB/CB-153. Hydroxylated PCB metabolites were also determined in rat tissues (lung, liver, and kidney). The chromatographic (GC-ECD) pattern was similar to the pattern in plasma. The concentrations of the major compound, 4-OH-2,3,5,3',4'-pentaCB, were generally lower in these tissues than the concentration of CB-153 ( Table 2). The ratio between the major OH-CB and CB-153 increased with time and on day 14 was about 1 in both liver and lung. No hydroxylated PCB metabolites were detected in rat adipose tissue.
The structures of the MeO-CBs in human plasma after methylation are shown in the chromatogram in Figure Ic.
The concentration of total PCB in the human plasma samples was 3.6 ± 1.6 pg/g lipid weight (range 2.0-5.3 jig/g lipid weight). The mean CB-153 concentration was 2.2 ± 0.9 pg/g (range 0.74-2.8 pg/g), making it a major constituent of the total PCB in human plasma. The lipid content in human plasma varies depending on the diet and how soon after a meal the blood sample was taken, and it is therefore more correct to express the concentrations on a fresh-weight basis. However, in order to compare these results with those from the environmentally exposed grey seals, where whole blood coagulates were analyzed, the concentrations are expressed on a lipidweight basis. By calculating a ratio between one of the OH-CBs and CB-153, it is possible to get indications of relative abundancy and to compare both species. The concentrations of the OH-CBs in human plasma were about 10% compared to the CB-153 concentration. The mean concentration of 4-OH-2,3,5,3',4'-pentaCB was 0.36 ± 0.2 pg/g lipid weight, giving a ratio of 0.18 (range 0.08-0.35) between the 4-OH-2,3,5,3',4'-pentaCB (including 4-OH-3,5,2',3',4'-pentaCB) and CB-153.

Discussion
In the present study, a crucial first experiment was performed by gavage treatment of rats with a technical PCB product and subsequent analysis of PCB and OH-CBs in their blood plasma. Only 13 OH-CBs were present in the rat plasma, a significantly lower number than expected, as many of the individual CBs in the PCB product may be transformed to several different hydroxylated metabolites. The estimated number of potential OH-CB metabolites formed from PCB may be as many as 200 individual compounds. No OH-CBs were detected in adipose tissue at any survival time, although they were present in the other tissues analyzed. Their presence in these tissues may partially be explained by blood residues in the tissues. It is striking that the concentration of the hydroxylated metabolite of CB-105, for example, is higher even than one of the most persistent PCB congeners, CB-153.
Because of the strong and specific retention of OH-CBs in the plasma from rats dosed with PCB (Aroclor 1254), we considered the possibility that OH-CBs might also be detected in environmental samples. Grey seals from the Baltic are known to have high levels of PCB in their blubber (28), and seals were therefore chosen for blood analysis. It was not possible to obtain fresh blood samples from any seals, but blood coagulates were collected from dead grey seals during autopsy. It is thus not possible to report the OH-CBs concentrations in plasma, so the concentrations are given on a lipid-weight basis in whole blood instead.
The grey seal blood also contained only a limited number of OH-CBs. Several of the methylated derivatives show similar retention times as the methylated OH-CBs identified in rat serum. Thus, the major peak in the chromatogram has the same retention time as 4-MeO-2,3,5,3',4'-pentaCB, but this peak also contains a minor amount of an isomeric compound, 4-MeO-3,5,2',3',4'-pentaCB. The ratio between these two CB-105 metabolites and CB-153 is 0.1:1.7, indicating similar concentrations of these metabolites as one of the most persistent PCB congeners.
It is notable that the major metabolite in rat plasma and a dominating OH-CB in both seal blood and human plasma is a compound that to a major extent is formed after a 1,2-shift of a chlorine in the para position in the 2,3,4-trichlorinated phenyl ring. A similar rearrangement is also observed to occur in the 3,4-dichloro-substituted phenyl rings of CB-77, CB-105, CB-118, CB-156 (18,19). Thus, all the major I-ortho-CBs can be transformed to 4-OH-CB metabolites that are retained in plasma or blood (Fig. 1).
A considerable variation in the relative amounts of the different OH-CBs is observed in the different species. In the rat, 4-OH-2,3,5,3',4'-pentaCB dominates, whereas the higher chlorinated OH-CBs are only minor components. In the human plasma, the heptaand hexachlorinated OH-CBs are much more abundant compared to the rat. The seal pattern shows some similarities with both human and rat. The reason for these dissimilarities may be due to different exposure situations. The rats were given a high dose 3 days in a row, whereas humans are exposed to a low dose of PCB during a life span. The higher chlorinated PCBs are often slowly metabolized [e.g., CB-153 (30)] and may therefore not have been formed in detectable amounts in the rat under the experimental conditions used.
A possible explanation for the highly selective retention of the OH-CBs in the blood samples may be their structural resemblance with thyroxin. In a previous study on the metabolism of CB-77, blood was shown to contain 4-OH-3,5,3',4'-tetraCB in a concentration 15 times higher than the parent compound, 5 days after oral exposure in mice (19). This metabolite was bound to a thyroxin-transporting protein (transthyretin) in the blood (19,31,32). Another PCB congener, 2,3,4,3',4'-pentaCB, has also been shown to be metabolized to a hydroxylated metabolite, 4-OH-2,3,5,3',4'-pentaCB, which was retained in blood after oral dose to mice (18). This metabolite is quantitatively the most important OH-CB identified in the blood samples of rats and seals in the present study.
In vitro binding studies between synthetic OH-CBs and TTR have shown that, for example, 4-OH-3,5,2',3',4'-pentaCB competes for the thyroxin (T4) binding site six times more efficiently than T4, the endogenous ligand, and twice as well as 4-OH-3,5,3',4'-tetraCB (33,34). 4-OH-2,3,5, 3',4'-pentaGB has not yet been tested for its binding capacity to TTR. The competitive binding of OH-GB congeners relative to T4 has been reported also by Rickenbacher et al. (35). In that study, computer modeling showed that OH-CBs with the substituents in meta or para positions were much more effective competitors for T4 than if the substituents were bound in an ortho position. Differences in the results from in vitro and in vivo binding studies have been observed (18,19,33,34) that indicate a more complex situation for binding than currently can be described by modeling or by in vitro studies. Thus, 4-OH-2,3,5-trichloroor 4-OH-3,5-dichloro-substituted OH-CBs compete with T4 both in vivo and in vitro, whereas 5-OH-3,4-dichloro-substituted OH-CBs have only been observed to compete in vitro.
The presence of selected OH-CBs in high concentrations in the blood of humans and seals gives cause for concern. PCB has been reported to inhibit the transport of thyroid hormones in rat plasma by competing with T4 for the binding site on the thyroxin-transporting protein TTR (31,32,36). Subchronic exposure to low doses of technical-grade PCB was reported to cause reproductive and thyroid effects (37). PCB congeners have also been shown to interfere with hepatic and brain thyroid hormone metabolism in fetal and neonatal rats after subchronic exposure in utero (38). Perinatal exposure to specific PCB congeners (CB-1 18 and CB-153) markedly decreased serum T4 in pups but not in dams, whereas 2,4,4'-triCB did not show such effects (39). The two former CBs are both transformed to hydroxylated metabolites present in blood, whereas CB-28 is not likely to form metabolites that fulfill the structural requirements.
PCB has also been reported to have antiestrogenic effects (40), but in another report both antiestrogenic and estrogenic effects were claimed (41). Korach et al. (42 showed that ortho-substituted PCB congeners had a higher affinity for the estrogen receptor than CBs without ortho-chlorine atoms. In the same report, it was also shown that a 4-OH-2',4',6'-triCB had affinity for the estrogen receptor (42). The OH-CBs identified in the present study must be further investigated in regard to their potential estrogenic activities.
The importance of the retention of a few OH-CBs in blood at levels similar to the persistent PCB congeners has hitherto been unknown, but the toxicological implications must be clarified. Studies on the toxicity of the major OH-CBs in blood must be initiated. Also, further development of the analytical methods to improve quantification of MeO-CBs must be carried out. Environmental Health Perspectives