Distribution of methyl sulfone metabolites of polychlorinated biphenyls and p,p'-DDE in human tissues.

We determined methylsulfonyl metabolites of chlorinated biphenyls (MeSO2-CBs) and 1,1-bis(4-chlorophenyl)-2,2-dichloroethene (p,p'-DDE) in human adipose, liver, brain, and lung tissues obtained from 11 Belgian individuals (9-62 years of age). The total concentration of MeSO2-CBs (lipid weight basis) decreased in the following order: liver (mean, 9.30 ng/g; range, 1.68-27.03 ng/g lipid) > lung [mean, 2.72 ng/g; range, not detected (ND) to 11.54 ng/g lipid] > adipose tissue (mean, 1.57 ng/g; range, 0.33-4.33 ng/g lipid) > brain (mean, 0.24 ng/g; range, ND-0.56 ng/g lipid). The profiles of MeSO2-CBs and MeSO2-DDE in each tissue were similar for all 11 subjects. In adipose, brain, and lung tissues, 4'-MeSO2-CB87, 4'-MeSO2-CB101, and 3-MeSO2-CB149 (except brain) occurred at higher concentrations than did other MeSO2-CBs. However, 3'-MeSO2-CB132 was by far the most abundant congener in liver, contributing on average to approximately 60% of the sum of MeSO2-CBs. The concentrations of 3-MeSO2-DDE in different tissues were at the same or lower levels than the total concentrations of MeSO2-CBs. This study suggests that the distribution patterns of MeSO2-CBs and MeSO2-DDE in humans differ between liver and other tissues. Moreover, these profiles differ from those found in other mammals, such as polar bears, porpoises, and otters.

VOLUME 111 | NUMBER 9 | July 2003 • Environmental Health Perspectives Research | Article Polychlorinated biphenyls (PCBs) and 1,1-bis (4-chlorophenyl)-2,2-dichloroethene (p,p´-DDE) are well-known pollutants that have been found in many environmental matrices and in humans (Kimbrough and Jensen 1989;Turusov et al. 2002). Despite their persistence in the environment, PCBs and p,p´-DDE are susceptible to enzyme-mediated biotransformation in biota. Generally, an initial step in the PCB metabolism is the generation of an arene oxide intermediate. Further, two main metabolic pathways in mammals have been reported: the mercapturic acid pathway to give methyl sulfone (MeSO 2 ) metabolites and hydroxylation to yield biphenylols (Letcher et al. 2000a). Although the major metabolic route is hydroxylation generally followed by excretion, significant amounts of methylsulfonyl metabolites of chlorinated biphenyls (MeSO 2 -CBs) and of DDE (MeSO 2 -DDE) may be accumulated in tissues because of their lipophilic character and specific protein-binding properties (Jönsson et al. 1994;Lund et al. 1985Lund et al. , 1988. MeSO 2 -CBs and MeSO 2 -p,p´-DDE have been mostly quantified in fat and liver tissues (Letcher et al. 2000a(Letcher et al. , 2000b, whereas hydroxylated PCBs were mostly determined in blood (Letcher et al. 2000a).
It was previously shown that the distribution of methylsulfonyl metabolites depends on both the congener structure and the speciesspecific metabolic capacity (Letcher et al. 1998;Stapleton et al. 2001). In experimental animals, it was demonstrated that MeSO 2 -CBs might also have specific protein-binding properties reflected by specific retention in the body, such as adrenal, lung, kidney tissues, and uterine fluid (Bergman et al. 1979;Brandt and Bergman 1987;Jönsson et al. 1994;Lund et al. 1985Lund et al. , 1988. Although specific accumulation in selected organs has been demonstrated in experimental animals dosed for a short period (no more then few weeks) (Brandt et al. 1982;Jönsson et al. 1992), it is difficult to relate it to the distribution of methylsulfonyl metabolites in humans. Although several MeSO 2 -CBs have been recently observed to be selectively and strongly retained in human liver (Norén et al. 1999;, the knowledge of accumulation and distribution of methylsulfonyl metabolites in other human tissues is still scarce.
The occurrence of MeSO 2 -CBs and MeSO 2 -DDE in wildlife and human samples and the demonstrated accumulation of certain MeSO 2 -CBs and MeSO 2 -DDE in specific tissues of mammals have made the investigation of these compounds in the environment important. In the present study, we have determined MeSO 2 -CB and MeSO 2 -DDE levels, together with PCBs and p,p´-DDE in different human tissues from Belgian individuals, to investigate their congener distribution and correlation between metabolites and parent compounds.
Extraction and cleanup. The sample preparation and analysis of MeSO 2 metabolites were as reported elsewhere (Chu et al. 2002) and are briefly described here. The tissue samples (0.5 g adipose tissue or 5 g of other tissues) were thawed, quantitatively weighed, and dehydrated by grinding with anhydrous Na 2 SO 4 until a free-flowing powder was obtained. The samples were spiked with 100 µL × 10 pg/µL 3-MeSO 2 -4-Me-CB as internal standard for MeSO 2 -CBs and MeSO 2 -DDE, and 100 µL × 100 pg/µL 2,2´,3,6´-tetra-CB (CB46) and 2,2´,3,4,5,6´-hexa-CB (CB143) as internal standards for PCBs and organochlorine pesticides. A Soxhlet system (B-811; BUCHI, Brussels, Belgium) was used in hot extraction mode for 4 hr with 100 mL hexane/acetone (3/1, vol/vol). After the extract was concentrated to approximately 10 mL and transferred to a centrifuge tube, an accurate portion (~10%) was removed for the gravimetric determination of lipid concentration. The remaining extract was concentrated, and the solvent was exchanged to hexane and reduced to a volume of 2 mL.
MeSO 2 -CBs and MeSO 2 -DDE were separated from other organochlorine compounds by liquid-liquid partitioning with concentrated H 2 SO 4 . The hexane solution was extracted twice with 2 mL concentrated H 2 SO 4 , and the acid phase was separated by centrifugation (4,000 rpm). The acid phases were combined and used for the determination of MeSO 2 -CBs and MeSO 2 -DDE, and the remaining hexane layer was used for the determination of PCBs and organochlorine pesticides.
For the determination of MeSO 2 -CBs and MeSO 2 -DDE, the combined sulfuric acid phase (~4 mL) was diluted with 4 mL cold distilled water. The solution was cooled, and the analytes were twice back-extracted with 2 mL hexane. The resulting organic layer was separated and washed with 2 mL NaCl solution (10%, wt/wt). The organic phase containing MeSO 2 -CBs and MeSO 2 -DDE was concentrated and loaded on a column packed with 1 g basic silica (33% KOH) and 5 g anhydrous Na 2 SO 4 and eluted with 10 mL DCM. The solvent was exchanged to hexane, concentrated to approximately 1 mL, and applied to a column (30 cm × 1 cm) filled with 8 g Florisil. The column was first eluted with 40 mL DCM/hexane (1:1, vol/vol), and 10 mL DCM, which were discarded. The MeSO 2 metabolites were eluted with 40 mL DCM, and this fraction was collected and concentrated to approximately 2 mL. The purified extract was evaporated to dryness by a gentle stream of nitrogen at ambient room temperature. The residue was dissolved in 100 µL isooctane and then transferred to an injection vial.  PCBs and organochlorine pesticides were determined as previously described . The remaining hexane phase from the partitioning with concentrated H 2 SO 4 was further cleaned on a column filled with 10 g acid silica gel (45%, wt/wt), and the analytes were eluted with 30 mL hexane. After evaporation to dryness, the residue was dissolved in 100 µL isooctane and transferred to vials for gas chromatography/electron-capture detection (GC-ECD) and gas chromatography/mass spectroscopy (GC/MS) analysis.

Determination of MeSO 2 -CBs and MeSO 2 -DDE by GC/electron-capture negative ionization MS.
An Agilent 6890 gas chromatograph (Agilent, Palo Alto, CA, USA) equipped with a 5973 quadrupole MS detector was used for the determination of MeSO 2 -CBs and MeSO 2 -DDE. The GC column was an AT-5MS capillary column 50 m × 0.18 mm internal diameter with a film thickness of 0.25 µm (Alltech, Lokeren, Belgium). Helium was used as carrier gas at a constant flow of 0.5 mL/min. Extract (2 µL) was injected in solvent vent mode. The injector temperature was held at 90°C for 0.3 min, heated at 600°C/min to 300°C, and then held for 20 min. The vent flow was 100 mL/min, and vent and purge times were 0.2 and 2 min, respectively. The oven temperature was programmed as follows: 80°C, held for 2.5 min, then at 20°C/min to 250°C, then at 5°C/min to 290°C, and finally held for 40 min. The transfer line and quadrupole temperatures were 250°C and 150°C, respectively. Ionization was performed in electroncapture negative ionization (ECNI) mode, using methane as reagent gas.
The fraction containing MeSO 2 -CBs and MeSO 2 -DDE was analyzed in the selected ion-monitoring mode, and the identification of analytes was based on their retention times and the ratios of monitored ions (Chu et al. 2002). Three most intense ions from Mand/or [M-Me]clusters were monitored (Chu et al. 2002). Twenty-six MeSO 2 -CB congeners and MeSO 2 -DDE, representing the majority of MeSO 2 metabolites identified so far in biota (Letcher et al. 2000a), were chosen for the method validation. Quantitative determination was done using a five-level calibration curve spanning the range of the analyte concentrations in real samples.

Determination of PCBs and organochlorine pesticides by GC-ECD and GC/MS.
An Agilent 6890 GC with µ-cell ECD was used for the determination of PCBs and pesticides. The separation was performed on an HT-8 capillary column (25 m × 0.22 mm, 0.25 µm film thickness) from SGE (Darmstadt, Germany). The carrier and make-up gas were helium (1 mL/min) and argon/methane 95%/5% (40 mL/min), respectively. The oven temperature was programmed as follows: 90°C, held for 1 min, then to 180°C at 15°C/min, held for 1 min, to 250°C at 3°C/min, and then to 290°C at 15°C/min, held for 6 min. The injector and detector temperatures were 270°C and 320°C, respectively. One microliter of the extract was injected in pulsed splitless mode, and the purge time was 1 min. The identification was based on the retention time of individual standard congeners, and a multicalibration curve was used for quantitative determination.
For confirmation, all samples were analyzed by GC/MS with electron ionization source and selected ion-monitoring mode. An HP-1MS capillary column (30 m × 0.25 mm, 0.25 µm film thickness) was used for the separation of analytes. The oven temperature was programmed as follows: 90°C, held for 2 min, then to 200°C at 3°C/min, then to 290°C at 3°C/min, held for 20 min. The injector temperature was 300°C. One microliter of extract was injected in splitless mode, and the purge time was 1 min. Although GC/MS (electron ionization) analysis can give additional information concerning the analyte identification, its sensitivity was lower than that of GC-ECD. Therefore, the quantitative results were based on the ECD determination, except for a few pairs of PCB congeners and pesticides (e.g., Article | Chu et al. 1224 VOLUME 111 | NUMBER 9 | July 2003 • Environmental Health Perspectives CB132 and p,p´-DDD) coeluting on the column used on GC-ECD. Quality control. The limit of detection (LOD) was defined as 3 times the standard deviation of measured values for samples (n = 7) spiked with the analyte amount that produced a signal approximately equal with three times the noise response in a blank sample. For MeSO 2 -CB and MeSO 2 -DDE determination, the limits of detection ranged from 0.06 to 0.10 ng/g lipid weight. Average recoveries of individual MeSO 2 -CB congeners and MeSO 2 -DDE from vegetable oil spiked with standards (1 ng) ranged between 84% and 99%, with a mean value of 89%. For PCB and organochlorine pesticide determination, the LOD ranged from 0.40 ng/g (CB149) to 24.6 ng/g (CB28), and the recovery ranged from 84 to 111%. The recoveries of internal standards (CB46 and CB143) in the samples ranged from 60 to 97%. No recovery corrections were applied to the final results. Quality control was ensured by regular analysis of procedural blanks and of certified reference material (CRM 350, PCBs in mackerel oil; Bureau of Community Research, Geel, Belgium), for which all measurements were within the range of certified values.
All statistical analyses were completed with STATISTICA for Windows, version 5.1, from StatSoft. Inc. (Tulsa, OK, USA). Concentrations of pollutants in tissues are summarized using arithmetic means along with minimum and maximal values (Tables 2 and 3).

Results
The individual concentrations of 26 MeSO 2 -CB congeners and 3-MeSO 2 -DDE, together with the sum of MeSO 2 -CBs, measured in four different tissues from 11 Belgian subjects are presented in Table 2. Individual and total concentrations of 29 PCB congeners, together with concentrations of selected organochlorine pesticides (p,p´-DDE, p,p´-DDD, and p,p´-DDT), are shown in Table 3. All concentrations listed in Tables 2 and 3 are expressed in nanograms per gram lipid weight, to conveniently compare them with previously reported data. Concentrations of CB18, CB28, and CB31 were below the LOD in all samples and are omitted from Table 3.
The sum of determined PCBs and p,p´-DDE (expressed per lipid weight) was similar in adipose tissue and liver from the same subject. However, higher concentrations of MeSO 2 -CBs, ranging from 1.68 to 27.03 ng/g lipid weight, were found in liver compared with other tissues ( Table 2). The ratios between the sum of MeSO 2 -CBs and the sum of PCBs and between MeSO 2 -DDE and p,p´-DDE, as calculated from the individual values, were 4-28 and 2-11 times higher in liver than in adipose tissue, respectively. When comparing levels of contaminants (expressed on wet weight basis) between the tissues from the same subject, the highest concentrations were found in adipose tissue. The total concentration of MeSO 2 -CBs decreased in the following order: adipose (mean, 1.30 ng/g wet weight; range, 0.25-3.62 ng/g) > liver (mean, 0.79 ng/g; range, 0.14-3.78 ng/g) > lung [mean, 0.03 ng/g; range, not detected (ND) to 0.14 ng/g] ≈ brain (mean, 0.02 ng/g; range, ND-0.05 ng/g) tissue.
The data on chemical residues were not normally distributed (Liliefors test), and therefore, all data were log-transformed before statistical analysis. Values below the LOD were set to zero. Pearson's correlation coefficients between the sum of MeSO 2 -CBs or MeSO 2 -DDE and the sum of PCBs or DDE were not high (Tables 4 and 5) and sometimes not statistically significant, whereas they were very low for most individual MeSO 2 -CBs and their corresponding PCB precursors. Significant correlation (p < 0.05) was found between age and sum of PCBs in lung, adipose tissue, and liver, whereas age was correlated only with sum of MeSO 2 -CBs in adipose tissue (Table 4).
The investigated tissues had characteristic profiles for the MeSO 2 metabolites. In adipose, brain, and lung tissues, 4´-MeSO 2 -CB87 and 4´-MeSO 2 -CB101 occurred at higher concentrations than did other MeSO 2 -CBs. These two congeners contributed together, on average, 55, 67, and 50% to the sum of MeSO 2 -CBs in adipose, brain, and lung, respectively. In adipose tissue, the concentration of 3´-MeSO 2 -CB132 was below the LOD or < 10% of the sum of MeSO 2 -CBs. In contrast, 3´-MeSO 2 -CB132 was the most abundant MeSO 2 -CB congener in liver, contributing on average 59% to the sum of MeSO 2 -CBs. In liver, the concentration of the sum of 3-MeSO 2 -CBs was higher than for the sum of 4-MeSO 2 -CBs, whereas in other tissues a reversed condition was observed. The ratio between the concentrations of the sum of 3-MeSO 2 -CBs and the sum of 4-MeSO 2 -CBs was 9.3, 0.3, 0.1, and 0.5 in liver, adipose, brain, and lung, respectively.

Discussion
The concentrations of MeSO 2 -CB congeners and MeSO 2 -DDE varied in various tissues from the subjects. Most of the lower chlorinated MeSO 2 -CB congeners were not detected in human tissues, probably because of a reduced abundance of lower chlorinated PCBs in the tissues (Table 3). In contrast, congeners with four to six chlorine atoms predominated in accordance with the longer biologic half-lives of the higher chlorinated MeSO 2 metabolites (Letcher et al 2000a).  have reported that the concentration of MeSO 2 -CBs in Swedish subjects ranged from 2.0 to 9.0 ng/g lipid weight in adipose tissue and from 11.8 to 358 ng/g lipid weight in liver. The concentrations of MeSO 2 -CBs are similar for adipose tissue in the Swedish and the present study, whereas the concentrations of MeSO 2 -CBs in Swedish liver are higher than the levels determined in the Belgian subjects, resulting from the wider concentration range determined in the Swedish liver samples. A possible explanation might be the difference in the age of subjects in both studies (34 years for Belgian subjects and 68 years for Swedish individuals), although only low correlation coefficients between MeSO 2 -CBs and age could be computed (Table 4). The difference in the diet might have an impact, as well. In addition, the PCB and p,p´-DDE levels in the Belgian subjects investigated in this study were lower than in the Swedish samples collected in the 1990s, probably in accordance with the decrease in PCB contamination during the last 10 years (Norén and Meironyté 2000). Ellerichmann et al. (1998) determined the enantiomers of eight MeSO 2 -CBs atropisomers in human liver and lung and found that the concentration of MeSO 2 -CBs present in Article | Methyl sulfone PCBs in human tissues Environmental Health Perspectives • VOLUME 111 | NUMBER 9 | July 2003 lung was substantially lower than in liver tissue. This result is in accordance with the present study, where MeSO 2 -CBs and MeSO 2 -DDE are also preferentially stored in the liver relative to other tissues. Calculated on a wet-weight basis, the concentrations of MeSO 2 -CBs in the lung were 1-2 orders of magnitude lower than those in adipose or liver, because of a much lower lipid percentage (Ellerichmann et al. 1998). In an early investigation (Haraguchi et al. 1986), MeSO 2 -CBs concentrations in lung and adipose tissue from a patient who died from causes unrelated to the Yusho accident are similar to those found in the present study. However, in the Yusho patient, the highest concentration of MeSO 2 -CBs was found in lung (15.9 ng/g wet weight). The Yusho patient had ingested rice oil contaminated with a large quantity of Kanechlor 400 (a PCB mixture) in a short period (Haraguchi et al. 1986), whereas the normal exposure to PCBs (as in the present study) is through the daily intake of foods with low levels of contamination.
There are no previously reported data for MeSO 2 -CB or MeSO 2 -DDE residues in the human brain. In the present investigation, only a few MeSO 2 -CB congeners (Table 2) could be quantitatively determined from brain tissues, although the lipid concentration in brain is similar to that in liver. Phospholipids, the dominating lipids in brain (Aguilar 1985;Tilbury et al. 1997), are less able to dissolve lipophilic compounds compared with triglycerides, which are present only in a small percentage (3-5%) (Kawai et al. 1988). However, the distribution of MeSO 2 -CBs in biota tissues can not be totally explained by a simple physical partition theory, and in this case, protein binding may play an important role in the tissue-specific retention (Letcher et al. 2000a).
For monitoring of MeSO 2 -CB residues in humans, blood and milk were more often investigated (Newsome and Davies 1996;Norén et al. 1996). It was reported that the mean concentration for the sum of MeSO 2 -CBs was 1.7 ng/g lipid in Canadian human milk, and the major congeners were 3´-MeSO 2 -CB101, 4´-MeSO 2 -CB87, and 4´-MeSO 2 -CB151 (Newsome and Davies 1996). In Swedish human milk, the concentration of MeSO 2 -CBs decreased from 9 to 2 ng/g lipids between 1972 and 1992, and the concentrations of 4-MeSO 2 -CBs were higher than the corresponding 3-MeSO 2 -CBs, with the major congeners being 4´-MeSO 2 -CB87 and 4-MeSO 2 -CB149 (Norén et al. 1996). Norén et al. (1999) reported that the concentration of the sum of MeSO 2 -CBs in Swedish human plasma ranged from 1 to 5 ng/g lipids, and that 4-MeSO 2 -CB149, 4´-MeSO 2 -CB87, and 4´-MeSO 2 -CB101 were the dominant congeners. These congener profiles seem to be similar to those in the adipose tissue investigated in the present study, although we did not determine the presence of 4´-MeSO 2 -CB151.
3´-MeSO 2 -CB132 is the most abundant MeSO 2 -CB congener we found in liver, contributing an average of 59% to the sum of MeSO 2 -CBs (Figure 1). Similar results were found in previous investigations Norén 1997, 1998), where 3´-MeSO 2 -CB132 was present in higher concentration than other MeSO 2 metabolites in Swedish human liver, whereas 4´-MeSO 2 -CB87 and 4-MeSO 2 -CB149 were the predominant MeSO 2 -CBs in adipose samples. In the present study, the concentration of the sum of 3-MeSO 2 -CBs in liver was higher than the sum of 4-MeSO 2 -CBs, whereas in other tissues a reversed condition was observed. A specific affinity for 3-MeSO 2 -CBs in the liver has been observed in gray seals and otters (Bergman et al. 1994), whereas a specific retention of 4-MeSO 2 -CBs has been observed in the lungs of rats and mice (Bergman et al. 1979;Haraguchi et al. 1999). This phenomenon indicates a different binding affinity (probably also species dependent) of 3-MeSO 2 -CB and 4-MeSO 2 -CB metabolites in various tissues (Haraguchi et al. 1999).
In the present study, it is not possible to clearly determine the relative roles of bioaccumulation versus bioformation of MeSO 2 metabolites, but it is useful to indicate that Article | Chu et al. 1226 VOLUME 111 | NUMBER 9 | July 2003 • Environmental Health Perspectives both may contribute. This could be part of the reason why there is such poor or no correlation between individual MeSO 2 -CBs and their precursors (Table 4). Letcher et al. (1998) have reported a clear indication of formation of 3´and 4´-MeSO 2 -CB87, 3-and 4-MeSO 2 -CB91, and 3´-and 4´-MeSO 2 -CB141 and 4´-MeSO 2 -CB49 in polar bears. The bears bioaccumulated all of the MeSO 2 -CB132 because the precursor was absent in the diet (seal). Most of the remaining congeners (3and 4-MeSO 2 -CB31, 3´-and 4´-MeSO 2 -CB49, 3-and 4-MeSO 2 -CB64, 3-and 4-MeSO 2 -CB70, 3´-and 4´-MeSO 2 -CB101, and 3-and 4-MeSO 2 -CB110 and 3-MeSO 2 -CB149) were likely bioaccumulated. The reason why patterns of MeSO 2 -CBs differ between wildlife and humans could be because the relative importance of bioaccumulation may be higher in some wildlife species and because of a very different diet. MeSO 2 -CBs, MeSO 2 -DDE, and their precursor compounds (PCBs and p,p´-DDE) are all lipophilic contaminants, but their bioaccumulation efficiency is not similar. MeSO 2 -CBs and MeSO 2 -DDE are selectively retained in liver tissues because of specific protein binding (Brandt and Bergman 1987). This phenomenon was also observed in the present study. In paired samples of human adipose tissue and liver, the concentrations of PCBs and DDE were similar on the lipid weight basis, but the concentrations of methylsulfonyl metabolites were different. The average ratios of the sum MeSO 2 -CBs to the sum PCBs were 1/306 and 1/39 in adipose tissue and liver, respectively. In this study, only 3-MeSO 2 -DDE was determined, because other MeSO 2 -DDE standards were unavailable. The ratios of 3-MeSO 2 -DDE to p,p´-DDE, calculated from average values, were 1/421 and 1/100 in adipose tissue and liver, respectively.
For the precursor compounds of the investigated MeSO 2 -CBs, only low levels of CB52, CB87, CB101, and CB110 were found in the tissues. To minimize the variations among individual subjects, the concentration of MeSO 2 -CBs and their precursors was normalized to the concentration of CB153 in individual tissues (Figure 1). It is well known that CB153 is resistant to biotransformation in many organisms, and it is assumed that CB153 represents the maximum bioaccumulation potential for slowly metabolized lipophilic compounds (Boon et al. 1992(Boon et al. , 1997Tanabe et al. 1994). A different distribution of metabolites was observed between the tissues (Figure 1). The ratios between MeSO 2 -CB and the parent CB decreased in the following order: liver > brain ~ lung > adipose tissue. The highest normalized concentration of MeSO 2 -CBs and the highest ratio of MeSO 2 -CB congener concentration versus its corresponding precursor CB congener were found for 3´-MeSO 2 -CB132 in liver. The reasons for these ratio variations are complex and depend not only on the congener metabolism but also on the preferential affinity of the metabolite for tissues.
In conclusion, the present results show that MeSO 2 -CB congeners have a different distribution, both as concentration and as profiles, in human tissues. Further studies are needed to reveal details of their mechanisms and their selective protein-binding affinity in different tissues.