Analysis of Hydroxylated Metabolites of PCBs (OH-PCBs) and Other Chlorinated Phenolic Compounds in Whole Blood from Canadian Inuit.

In this study, we identified the main hydroxylated polychlorinated biphenyls (OH-PCBs) and other chlorinated phenolic compounds and we determined their relative concentrations in whole blood from 13 male and 17 female Inuit from northern Quebec, Canada, and from a pooled whole blood sample from southern Quebec. We also determined concentrations of polychlorinated biphenyls (PCBs). Total OH-PCB concentrations were variable among the Inuit samples, ranging over 2 orders of magnitude (0.117-11.6 ng/g whole blood wet weight). These concentrations were equal to and up to 70 times those found for the southern Quebec pooled whole blood sample. Geometric mean concentrations of total OH-PCBs were 1.73 and 1.01 ng/g whole blood for Inuit men and women, respectively, and 0.161 ng/g whole blood for the southern population pool. There are limited data available for comparison, but the levels of OH-PCBs in Inuit are higher than those previously reported in the literature for other populations. There was a significant correlation (p [is less than] 0.005) between OH-PCBs and PCBs (r = 0.84) and both correlated significantly (p [is less than] 0.005) with age (r = 0.68 and 0.78, respectively). The ratio of OH-PCBs to PCBs was lower in Inuit (0.11) than in the southern Quebec pool (0.33). There is no apparent explanation for the difference. There was considerable variability in the congener pattern of the identified OH-PCBs. The main metabolite, 4-OH-CB109 (4-OH-2,3,3',4',5-pentachlorobiphenyl), constituted 12-62% of the total OH-PCBs in the samples. Pentachlorophenol (PCP) was the dominant phenolic compound in blood, constituting 46% (geometric mean) of the total quantitated chlorinated phenolic compounds. PCP concentrations in Inuit blood ranged from 0.558 to 7.77 ng/g on a wet weight basis. All but two Inuit samples had lower concentrations than the southern Quebec pool (6.29 ng/g). The possible role of OH-PCBs in mediating PCB-induced adverse effects needs to be investigated further. Key words: Canadian Inuit, chlorinated phenolic compounds, hydroxylated metabolites, OH-PCBs, PCBs, pentachlorophenol, polychlorinated biphenyls. Environ Health Perspect 108:611-616 (2000). [Online 25 May 2000]

http://ehpnet1.niehs.nih.gov/docs/2000/108p611-616sandau /abstract.html

Polychlorinated biphenyls (PCBs) have been extensively studied since their discovery as environmental pollutants over 30 years ago (1). Their persistence in biota is well known but the mechanism of their adverse effects on biologic systems is still not completely understood. Some of their toxicity may be linked to the biotransformation products of PCBs (2). PCBs are biotransformed by a diverse enzyme system, the cytochrome P450 monooxygenases. Most of the known metabolic pathways involve the initial formation of hydroxylated metabolites. Even one of the more recalcitrant PCBs, CB153, is metabolized in vitro and in vivo to a number of phenolic metabolites (3-5). Phenolic metabolites can be excreted unchanged, as they were first discovered in the excreta of Baltic seals and Guillemots (6), or further conjugated with glucuronic acid or sulfate (7). The introduction of a hydroxyl group increases the polarity of the PCB and facilitates excretion. If the hydroxyl group is para to the phenyl-phenyl bond and has adjacent chlorine atoms, the structure resembles the prohormone, thyroxin ([T.sub.4]) (8). This structural similarity allows hydroxylated polychlorinated biphenyls (OH-PCBs) to bind with high affinity to one of the thyroid hormone transport proteins, transthyretin (TTR) (9-12). For example, 4-OH-3,3',4,5-'tetrachlorobiphenyl, a metabolite of CB77, has a binding affinity to TTR 4 times stronger than that of [T.sub.4] (9), and can disrupt thyroid hormone and retinol (vitamin A) transport (13). This is the presumed mechanism by which OH-PCBs are selectively retained in plasma and is a possible mechanism of PCB toxicity (14). This type of interaction is not limited to OH-PCBs. Pentachlorophenol (PCP) binds with 2 times the affinity of [T.sub.4] to TTR (15), indicating that other chlorinated phenolic compounds may also be interfering in thyroid hormone transport.

Some Inuit consume traditional foodstuffs consisting of fatty tissues from sea mammal species such as ringed seal and beluga (16-18). Ringed seal blubber, beluga skin, and beluga blubber in northern Canada contain average total PCB concentrations of 1,283, 145, and 5,000 ng/g wet weight, respectively (19,20), which is much higher than that in the diet of the general population (21). Thus, the Inuit population may be exposed to large doses of PCBs. In one study, mean PCB blood levels in Inuit were 30 times those of a southern population (22). An increased PCB body burden could result in the increased formation of metabolites due to induction of the P450 system. Increased levels of metabolites may be significant enough to disrupt thyroid hormone transport. Alterations of thyroid hormone status have been proposed as a mechanism of action by which PCBs would induce adverse neurodevelopmental effects (23). It is therefore of interest to study the concentrations of OH-PCB metabolites in the Inuit population as a first step to evaluate their possible implications mediating PCB-induced health effects. Hydroxylated PCBs have never been studied in the Arctic environment and little is known about levels and patterns in humans. There are only two other published studies on OH-PCB metabolite concentrations in human blood [from Sweden and Latvia (14,24)].

Materials and Methods

During fall 1992, 499 Inuit adults living in Nunavik (Northern Quebec, Canada) participated in the Sante Quebec Health Survey. After they signed an informed consent form, we drew a 30-mL blood sample by venous puncture for organochlorine determination. We randomly selected 30 subsamples of these 499 samples for chlorinated phenolic compound and additional PCB residue analysis. The Laval University Medical Centre (Sainte-Foy, Quebec) donated a southern Quebec general population sample of pooled whole blood for comparison to the Inuit population.

The chosen nomenclature for numbering hydroxylated PCBs or their derivatized analogs, methoxylated PCBs, is slightly different from general International Union of Pure and Applied Chemistry (IUPAC) guidelines (25). The OH- or MeO- groups are not given numbering priority. Rather, the chlorine pattern on the biphenyl rings determines the congener number according to the IUPAC PCB numbering rules (25), with corrections by Guitart et al. (26), and the OH- or MeO- groups are numbered thereafter. This facilitates direct structural comparison with PCBs because the IUPAC numbering system is so well established and familiar.

We used the following [sup.13][C.sub.2]-labeled standards [acquired from Wellington Laboratories (Guelph, Ontario, Canada)] as recovery internal standards for OH-PCB determination: 4'-OH-CB120 (4'-OH-2,3',4,5,5'-pentachlorobiphenyl); 4'-OH-CB159 (4'-OH-2,3,3',4,5,5'-hexachlorobiphenyl); 4'-OH-CB172 (4'-OH-2,2',3,3',4,5,5'-heptachlorobiphenyl); and 4-OH-CB187 (4-OH-2,2',3,4',5,5',6-heptachlorobiphenyl). We purchased PCP ([sup.13][C.sub.6]) from Cambridge Isotope Laboratories (Andover, MA) and used it for PCP quantitation. We added a synthesized performance standard, 4'-Me-4-MeO-2,3,3',5,6-pentachlorobiphenyl, to all phenolic compound fractions before mass spectral analysis. We used labeled PCBs ([sup.13][C.sub.12]-CB28, -52, -118, -153, - 180, and - 194) as recovery standards and [sup.13][C.sub.12]-CB138 as the performance standard for PCB analysis by the external standard method. The [sup.13][C.sub.12]-PCB standards were purchased from Cambridge Isotope Laboratories.

Whole blood samples (mass range 1.54-5.76 g) were spiked with [sup.13][C.sub.12] recovery standards before extraction. The standards consisted of a PCB mixture ([sup.13][C.sub.12]-CB28, -52, -118, -153, -180, and -194; 10 [micro]L, 2.5 ng/[micro]), a OH-PCB mixture ([sup.13][C.sub.12]-4'-OH-CB120, 4'-OH-CB159, 4'-OH-CB172, and 4-OH-CB187; 20 [micro]L, 100 pg/[micro]L), and PCP ([sup.13][C.sub.6]; 100 [micro]L, 100 pg/[micro]L). The samples were then extracted using the Wallenberg plasma extraction method, as described elsewhere (14,27), with the following modifications. The organic phase from the potassium hydroxide partitioning, which contains the neutrals, was reduced in volume and applied to a Florisil column (8 g, 1.2% deactivated). The PCBs were collected in one fraction of 75 mL dichloromethane (DCM):hexanes (1:1) then reduced in volume and applied to a silica/sulfuric acid (3 g, 22%) column to remove coextracted biogenic components. The PCBs were eluted in 50 [micro]L DCM:hexanes, reduced to final volume (50 [micro]L) by rotoevaporation, and spiked with performance standard ([sup.13][C.sub.12]; CB138, 5 [micro]L, 2.0 ng/[micro]L) for mass spectral analysis.

The partitioned phenolic compounds were acidified and back extracted with hexanes, dried over sodium sulfate, and derivatized with diazomethane. We generated diazomethane as needed from nitrosomethylurea precursor, as described elsewhere (28). The derivatized compounds were then cleaned up on a silica/sulfuric acid (5 g, 22%) column and eluted with 50 mL DCM:hexanes. We reduced the samples in volume by rotoevaporation and brought the samples down to final volume (25 [micro]L) by a gentle stream of nitrogen. All samples were then spiked with a performance standard (4'-Me-4-MeO-2,3,3',5,6-pentachlorobiphenyl, 5 [micro]L, 200 pg/[micro]L before analysis.

Analyses were completed on a Hewlett Packard 5890A Series II gas chromatograph Hewlett Packard, Atlanta, GA) equipped with an HP 7673A automatic injector and a Hewlett Packard 5988A mass spectrometer. We used helium as the carrier gas; we set the head pressure at 80 kPa. All injections (2 [micro]L) were made in splitless mode onto a DB-5 column [(5%-phenyl)-methylpolysiloxane, 30 m x 0.25 [micro]m i.d., 0.25-[micro]m film thickness; J&W Scientific, Inc., Folsom, CA). The injector temperature was set at 250 [degrees] C. We performed mass spectrometry with selected ion monitoring in electron capture negative ionization (ECNI) mode using methane (99.99% pure) as the reagent gas for the chlorinated phenolic fraction. The reagent gas source pressure was 2.5 x [10.sup.-4] torr at the inlet and the source temperature was 140 [degrees] C. The GC temperature program for the derivatized phenolic compounds was as follows: 80 [degrees] C for 2 min, ramp 10 [degrees] C/min to 250 [degrees] C, hold for 5 min, then ramp 5 [degrees] C/min to 300 [degrees] C. We analyzed the phenolic compound fraction using ECNI because of the presence of residual biogenic material that interferes with the analysis in electron impact (EI) mode. The PCB fraction was free of biogenic contamination, allowing analysis of the fraction using selected ion monitoring in EI mode. The electron energy was 70 eV and the source temperature was 200 [degrees] C. The GC temperature program for the PCB analysis was 100 [degrees] C for 3 min, 10 [degrees] C/min to 180 [degrees] C, and then 2.5 [degrees] C/min to 280 [degrees] C.

Because analysis was carried out on methylated OH-PCBs, we used a standard mixture of MeO-PCBs. The MeO-PCB mixture included 4'-MeO-2,3',4,5,5'-pentachlorobiphenyl (4'-MeO-CB120);4'-MEO-2,3,3',4,5'-pentachlorobiphenyl (4'-MeO-CB107); 4-MEO-2,2',3,4',5,6,6'-heptachlorobiphenyl (4-MeO-CB188); 3-MeO-2,2',4,4',5,5'-hexachlorobiphenyl (3-MeO-CB153); 4-MEO-2,2',3,4',5,5'-hexachlorobiphenyl (4-MeO-CB146);3'-MeO-2,2',3,4,4',5'-hexachlorobiphenyl (3'-MeO-CB138); 4'-MEO-2,2',3,3',4,5'-hexachlorobiphenyl (4'-MeO-CB130); 3'-MeO-2,2',3,4',5,5',6-heptachlorobiphenyl (3'-MeO-CB187); 4'-MEO-2,2',3,3',4,5',6-heptachlorobiphenyl (4'-MeO-CB175);4'-MEO-2,3,3',4,5,5'-hexachlorobiphenyl (4'-MeO-CB159); 3'-MEO-2,2',3,4,4',5,5'-heptachlorobiphenyl (3'-MeO-CB180);4'-MeO-2,2',3,3 ,4,5,5'-heptachlorobiphenyl (4'-MeO-CB172); and 4-MeO-2,3,3',4',5,5',6-heptachlorobiphenyl (4-MeO-CB193). The MeO-PCBs were supplied by A. Bergman (Wallenberg Laboratories, Stockholm, Sweden), and synthesized as described elsewhere (29). We used the MeO-PCBs for quantitation by the external standard method.

We used a serial dilution of the MeO-PCB mixture for quantitation. We generated response factors relative to the performance standard (RRFs) for each of the compounds in the standard mixture. The RRFs were then used to quantitate all identified and unidentified OH-PCBs in each sample by the external standard method, with volume correction using the performance standard. Identified compounds were quantitated using their RRF. Structures were considered confirmed if they had identical spectra and matching retention times on three GC columns with varying polarities (DB5, DB1701, and DB210; J&W Scientific, Inc.). We characterized unidentified compounds by full scan mass spectrometry to determine the chlorination pattern. The average RRFs for that chlorination pattern were used for quantitation for the unidentified compounds. The RRFs for compounds with a given chlorination and methoxy group substitution were comparable. For example, the RRFs for the para-methoxylated heptachlorobiphenyls were all within 10% of the average RRF. We determined recoveries using the [sup.13][C.sub.12] OH-PCB standards and 4'-Me-4-MeO-2,3,3',5,6-pentachlorobiphenyl as the performance standard. PCP was quantitated by isotope dilution using the [sup.13][C.sub.6] internal recovery standard correcting for the 3.4% contribution of native PCP (m/z = 280 amu) to the main ion of the [sup.13][C.sub.6] standard (m/z = 286 amu) duster.

PCBs were quantitated on a congener-specific basis by the external standard method using an Aroclor 1242:1254:1260 (1:1:1) secondary quantitation standard solution calibrated against primary standard PCB congener solutions provided by the National Research Council of Canada (Marine Analytical Chemistry Standards Program, Halifax, Nova Scotia, Canada). We determined concentrations of congeners for which standard solutions were not available in the Aroclor mixture by GC using flame ionization detection (30). We purchased Aroclor from Monsanto (St. Louis, MO).

PCBs were previously determined in these samples by the Quebec Toxicology Centre (Sainte-Foy, Quebec, Canada) using plasma as the tissue for analysis. Because of Aroclor 1254 contamination in one of our rotoevaporators, 15 of the 30 whole blood samples were irreclaimable for PCB quantitation. We compared the 15 whole blood samples not contaminated with Aroclor 1254 to the previous plasma analyses. Concentrations of all previously quantitated congeners (CB28, CB52, CB99, CB101, CB105, CB118, CB128, CB138, CB153, CB156, CB170, CB180, CB183, and CB187) and the sum of congeners in plasma were highly correlated (r [is greater than] 0.97) to the whole blood concentrations using least-squares regression analysis. The regression analysis for CB153 and total PCBs are given as examples in Figure 1. Thus, the previous PCB plasma concentrations for the 15 samples not quantitated in our lab were adjusted by applying a correction factor from the regression analysis. We used the adjusted PCB concentrations and the values determined in the 15 whole blood samples in the present study for comparison to OH-PCB concentrations and for subsequent statistical analysis. Aroclor contamination had no effect on the quantitation of OH-PCBs in the contaminated samples because PCBs elute earlier and have no fragment ions that interfere with the detection of OH-PCBs.

[Figure 1 ILLUSTRATION OMITTED]

All solvents were residue analysis grade and were purchased from EM Science (Gibbstown, NJ). Florisil (Pesticide Analysis Residue grade, 60-100 mesh) was purchased from BDH, Inc. (Toronto, Ontario, Canada). Merck Silica gel (Grade 60, 70-230 mesh, 60 [Angstrom]) was purchased from Aldrich Chemical Company, Inc. (Milwaukee, WI). We purchased trace-metal-grade sulfuric acid from Fisher Scientific (Pittsburgh, PA).

Results

Whole blood recoveries of OH-PCBs were variable, ranging from 50 to 105%. The mean recoveries for [sup.13][C.sub.12]- 4'-OH-CB120, -4'-OH-CB159,-4'-OH-CB172 and -4-OH-CB187 were 82, 69, 72, and 76%, respectively. This was the first study to use [sup.13][C.sub.12]-labeled OH-PCBs for accurate OH-PCB quantitation. PCB recoveries had a slightly lower mean recovery of 60%. All PCB and OH-PCB concentrations were consequently recovery corrected.

An example chlorinated phenolic ECNI-MS full scan chromatogram is given in Figure 2. The major peaks are identified when known. The OH-PCB fraction in Inuit whole blood contained [is greater than] 30 congeners, of which 11 were positively identified with authentic standards. Positive identification is based on identical mass spectra using full scan ECNI mass spectrometry and matching retention times using three GC columns with varying polarities (results not shown). All identified congeners are listed in Table 1. The congeners constituted between 59 and 81% of the total OH-PCBs in the samples (mean = 70%).

[Figure 2 ILLUSTRATION OMITTED]

Table 1. Concentration of PCP and hydroxylated metabolites (ng/g whole blood wet weight x [10.sup.3]) in Inuit whole blood.(a)

                         Males (n = 13)

                                 Range

                     GM      Min     Max

Age (years)            38      18       66
PCP                 2,740   1,350    7,770
4'-OH-CB120            12       3       67
4-OH-CB109(b)         314      39    2,550
3-OH-CB153(b)          85       5      537
4-OH-CB146(b)         219      10    1,750
3'-OH-CB138            37       3      225
4'-OH-CB130            15       1      506
3'-OH-CB187            17       1      213
4-OH-CB187(b)         293      22    1,840
3'-OH-CB180            17       2      134
4'-OH-CB172(b)         74       5      443
4-OH-CB193             41       3      659
Sum congeners(b)    1,040      88    6,740
Sum identified      1,210     119    7,520
Sum all OH-PCBs     1,730     162   10,100
Sum PCBs           12,900   2,070   65,900
CB-153              3,120     460   13,900

                      Female (n = 17)

                                Range        Southern
                                            population
                    GM      Min     Max        pool

Age (years)           38      18       72         -
PCP                1,590     558    7,510     6,290
4'-OH-CB120            7     < 1       39         3
4-OH-CB109(b)        234      15    1,470        25
3-OH-CB153(b)         49       4      830         2
4-OH-CB146(b)        111      10    1,880         8
3'-OH-CB138           20       2      401         2
4'-OH-CB130            4     < 1       47       < 1
3'-OH-CB187           11     < 1      167       < 1
4-OH-CB187(b)        152      27    2,260        31
3'-OH-CB180            8       1      239       < 1
4'-OH-CB172(b)        38       4      740        11
4-OH-CB193            26       2       95        18
Sum congeners(b)     614      59    7,070        76
Sum identified       712      87    8,060        82
Sum all OH-PCBs    1,010     117   11,600       161
Sum PCBs           7,940   1,190   38,100       488
CB-153             1,960     263    9,510        74

Abbreviations: GM, geometric mean; max, maximum; min, minimum.

(a) Total OH-PCBs, sum five main congeners, and sum identified congeners are given for comparison with previous studies.

(b) Sum five main congeners: 4-OH-CB109 + 3-OH-CB153 + 4-OH-CB146 + 4-OH-CB187 + 4'-OH-CB172.

We did not determine the lipid content of the whole blood samples. Therefore, all concentrations hereafter are expressed on a whole blood wet weight basis. The chemical residue data were not normally distributed. Log-transformed data approached normal distribution; therefore, log-transformed data were used in all statistical analyses and geometric mean values were used unless otherwise stated. The geometric mean concentrations in men and women of all identified congeners and their range of values are listed in Table 1. Women consistently had lower geometric mean concentrations of all the chlorinated phenolic compounds quantitated, but the difference was statistically significant between men and women using Student's t-test only for PCP.

The concentrations in the pooled sample are equivalent to arithmetic mean concentrations. The arithmetic mean of the total OH-PCBs in Inuit whole blood was 1.89 times greater than the geometric mean. Therefore, the difference in geometric mean values between populations is probably an overestimate of the geometric mean and the differences in concentrations are even greater than indicated in Table 1.

PCP was the main chlorinated phenolic compound in the Inuit whole blood. It contributed between 14 and 89% of the total chlorinated phenolic compounds quantitated. The concentration of PCP ranged from 0.558 to 7.77 ng/g with a geometric mean of 2.02 ng/g wet weight. The pooled whole blood sample from the southern population had higher PCP levels than all but two Inuit samples analyzed, with a concentration of 6.29 ng/g. PCP composed 97% of the total quantitated chlorinated phenolic compounds in the southern population.

The major OH-PCB congeners varied considerably among individuals. The mean ratios of the five main congeners to total OH-PCBs are shown in Figure 3, with bars representing the range. The main congener identified in Inuit whole blood (21 of 30 samples) was 4-OH-CB109, with a geometric mean concentration of 0.266 ng/g and a range of 0.015-2.55 ng/g. This congener is probably a mixture of two congeners, 4-OH-CB109 (4-OH-2,3,3',4',5-pentachlorobiphenyl) and 4'-OH-CB107, because they coelute and because separation was not achieved with the DB-5 column used in this study. This peak was quantitated using only 4'-OH-CB107 as the quantitation standard because 4-OH-CB109 was not commercially available. The peak is assumed to be predominantly 4-OH-CB109 because it is the dominating compound of the mixture in human plasma at a 5:1 excess (14,31). However, considering pattern variability this would need to be confirmed for each sample in the future.

[Figure 3 ILLUSTRATION OMITTED]

In eight of the samples, 4-OH-CB187 was the dominating congener. It ranged from 0.022 to 2.26 ng/g with a geometric mean of 0.202 ng/g for all samples. This was the main metabolite in plasma determined in previous studies (14,32,33). The main metabolite in the remaining sample was 4-OH-CB146. The geometric mean concentration for 4-OH-CB146 in all samples was 0.149 ng/g and ranged from 0.010 to 1.88 ng/g. The metabolite may be formed from CB146 via a direct oxygen insertion or from CB153 or CB138 via a National Institutes of Health (NIH) shift of Cl in the hydroxylation step. If a metabolite is formed by a direct oxygen insertion, it would retain the precursor PCB number. If the metabolite is produced via an NIH shift of Cl, the metabolite PCB number will be different from that of the parent PCB.

Total quantitated OH-PCBs ranged from 0.117 to 11.6 ng/g with a geometric mean value of 1.27 ng/g. Total OH-PCBs in individuals were equal to and up to 70 times that of the southern Quebec pooled sample (0.161 ng/g). Geometric mean CB153 and geometric mean total PCB concentrations in combined Inuit samples were 2.37 ng/g (range 0.263-13.9 ng/g) and 15.2 ng/g (range 1.19-65.9 ng/g), respectively. Concentrations in the southern Quebec pooled sample were 0.074 ng/g for CB153 and 0.488 ng/g for total PCBs. The mean ratio of total OH-PCBs to total PCBs was 0.11 (range 0.02-0.45) for combined Inuit and 0.33 for the southern Quebec pooled sample, demonstrating the importance of OH-PCBs in the total PCB-related compounds in blood. The concentration of the five main OH-PCB metabolites relative to total OH-PCBs and CB153 for both Inuit and the southern Quebec pooled sample are shown graphically in Figure 3.

Plots of log-normalized concentrations versus age are shown in Figure 4. Both PCBs (r = 0.78) and OH-PCBs (r = 0.68) were significantly (p [is less than] 0.005) correlated with age. As seen by the similar slopes, total OH-PCBs and total PCBs were also significantly correlated (r = 0.84, p [is less than] 0.005).

[Figure 4 ILLUSTRATION OMITTED]

Discussion

The only human blood analyzed previously for OH-PCBs and other chlorinated phenolic compounds was plasma from the Swedish population (14,31). PCP was described as the dominant compound in these studies, but levels were not given. The geometric mean PCP concentration in the Inuit samples was 2.02 ng/g, approximately 3 times lower than in the southern Quebec pooled sample. Geyer et al. (34) measured PCP in two different general populations in Germany and found average concentrations of approximately 20 ng/g wet weight. The study did not indicate if the samples analyzed were whole blood or plasma, so the comparison of levels with the present results may be incorrect. However, a recent study documented levels of PCP in men from Sweden and Latvia between 170 and 1,800 ng/g on a plasma lipid weight basis (24). When the Swedish data are approximated to plasma equivalents, these levels are the same order of magnitude and have a similar range as those for the Inuit and the southern Quebec pooled whole blood samples. The plasma equivalent was estimated by assuming that whole blood is approximately half plasma by weight and that the average plasma lipid levels in the population are approximately 1%.

Although the concentration of PCP is lower in Inuit than that measured in the southern Quebec pooled whole blood sample, it may still play a role in the disruption of thyroid hormone transport. PCP binds to TTR with twice the affinity of the native hormone (15) and is responsible for 14-89% of the total phenolic compounds quantitated. Geyer et al. (34) calculated the half-life of PCP as approximately 19 days and concluded that the German cohort's calculated intake of PCP exceeded the elimination rate and approximated a bioconcentration factor of 1.4 for blood. This is likely due to binding with TTR and explains the high levels of PCP found in blood. Considering the significance of PCP in the chlorinated phenolic fraction, further investigation of PCP and its possible effects on thyroid hormone transport is required, especially for southern populations, where levels of PCP far exceed OH-PCBs and other chlorinated phenolic compounds.

Another major compound in the Inuit whole blood chromatogram is an unidentified chlorinated phenolic compound that elutes just before the OH-PCBs. The major ion cluster for this compound, using the given mass spectrometry conditions, has an isotope pattern indicative of a hexachlorinated compound. Therefore, the levels were approximated using the C16 response factor from the OH-PCB standard mixture. Based on approximate concentrations, the compound accounts for 1-17% of total chlorinated phenolic compounds in the Inuit. Levels occasionally exceeded those of the main OH-PCBs. Further work on the identification and accurate quantitation of this unknown compound is currently underway.

The chromatogram of the phenolic compound fraction consists of [is greater than] 30 peaks that were identified as hydroxylated metabolites of PCBs based on their mass spectra. These represent 20% of the total chlorinated phenolic compounds, on average, in Inuit whole blood (range 2-56%). OH-PCBs constitute a significantly more important proportion of the total chlorinated phenolic compounds in Inuit as compared to the general population pooled sample, in which OH-PCBs represent only 1.2% of the estimated total.

The mean concentration of the five main OH-PCBs (1.47 ng/g) in Inuit samples is slightly higher than that of a Swedish general population cohort (1.0 ng/g) (31). The Swedish data have been adjusted to whole blood wet weight equivalents. It was expected that levels of OH-PCBs in Inuit would be higher than those found because the mean PCB levels in plasma were 3 times that of this particular Swedish population, and an average of 30 times (range 2.5-133) that of the southern Quebec pooled sample. The range of the sum of these five congeners in the Swedish study (0.35-1.65 ng/g) is small as compared to concentrations in the Inuit population, which range from 0.059 to 7.07 ng/g plasma wet weight levels of OH-PCBs. The higher levels of OH-PCBs are likely due to the higher levels of parent PCBs. OH-PCBs were significantly correlated with PCBs (r = 0.84, p [is less than] 0.005), as expected.

The strong linear relationship observed between concentrations of PCBs and their metabolites suggests that no changes in enzyme activity affecting the rate of formation or saturation of OH-PCBs binding to plasma proteins are taking place for this range of PCB exposure. The concentration of total quantitated chlorinated phenolic compounds was [is less than] 2% of the theoretical molar circulating concentration of TTR in an average human (35). Thus, it is not likely that the concentration of total chlorinated phenolic compounds reached the carrying capacity of the blood, assuming that Inuit have similar TTR concentrations as the published values from other populations (35).

Women had lower mean concentrations of all phenolic compounds quantitated than men. The generally lower levels in women may result from the loss of OH-PCBs and PCBs through lactation because both have been identified in milk (36-38). Both OH-PCBs and PCBs were significantly correlated with age (r = 0.68 and 0.78, respectively; p [is less than] 0.05). The increase of both PCB and OH-PCB concentrations with age may be due to the increased exposure with age (elders' preference for traditional foods), a slow excretion rate of PCBs that prevents steady state from being achieved, or both. A positive association between age and PCB levels in breast milk was reported previously in southern Quebec women (39) and in Inuit adults from Nunavik (22).

The concentration of omega-3 fatty acids in plasma phospholipids is a good indicator offish intake (40). In the course of the Sante Quebec Health Survey, polyunsaturated fatty acid analysis was performed on plasma samples obtained from 492 Inuit adults. There was a strong association between omega-3 fatty acid content in plasma phospholipids and the age of the participants, with mean (geometric) concentrations of 6.4% [95% confidence interval (CI), 5.9-6.9]; 8.7% (CI, 8.2-9.3); and 12.2% (11.5-13.0) for the 18-24 year, 25-44 year, and 45-74 year age groups, respectively (Fisher exact test: p [is less than] 0.0001) (41). Thus, an increase in the intake of traditional foods with age is indicated, and is probably partly responsible for the increase in PCBs with age. However, indications of increasing concentration of PCBs with age in other populations suggest that slow excretion of PCBs is also important.

The main OH-PCBs identified in previous studies were also the dominating congeners in the Inuit population (Figure 2). The major metabolite in 21 of the 30 samples analyzed was 4-OH-CB109. This was not the major metabolite found in the study by Bergman et al. (14), but it was later identified by Klasson-Wehler et al. (31) as the dominant OH-PCB in another study of human plasma. The probable mechanism of formation of this metabolite is direct oxygen insertion into CB-109 or an NIH shift of Cl through an arene oxide intermediate of CB118 or CB105. CB118 and CB105 are major congeners of Aroclor mixtures, whereas CB-109 represents only a fraction of a percent of the Aroclor mixture; therefore, the NIH shift mechanism is more likely, and occurs in the mink and the mouse (42). The majority of metabolites were highly correlated to all PCB congeners, making it impossible to determine the most probable precursors by correlation analysis. For example, 4-OH-CB109 was significantly correlated to both CB105 and CB118 (r = 0.67 and 0.72, p [is less than] 0.005). The correlation coefficients for 4-OH-CB109 and unrelated PCBs, such as CB153 and CB187, were equally high (0.77 and 0.71, p [is less than] 0.005). This was true for the majority of the identified metabolites and the major PCBs in the Inuit whole blood samples.

The main metabolite in most of the remaining samples was 4-OH-CB187, which was previously identified as the major metabolite in a Swedish population (14). This compound was also identified as the main metabolite for the southern Quebec pooled sample and in other species including polar bears (33) and albatrosses (32). This metabolite is most likely formed from the metabolism of CB183 and/or CB187, which represent 2 and 5%, respectively, of the Aroclor 1254 mixture (43).

The OH-PCB pattern present in the chlorinated phenolic compound fraction in plasma is complex and it is difficult to generalize patterns in humans from such a small data set. The ratio of 4-OH-CB109 to total OH-PCBs ranged from 0.12 to 0.62 (mean = 0.34) (Figure 3). The remaining congeners in the Inuit all composed similar proportions of total OH-PCBs as compared to the southern population. The ratio of the five main congeners to CB153 in the southern pooled samples was within the range of those determined for the Inuit samples, except for 4-OH-CB187, where the southern pooled sample was higher than all of the Inuit samples.

A number of factors, which include exposure to different proportions of precursor PCBs, alteration of metabolism rates by induction of hepatic enzymes, inhibition of the metabolizing enzymes and the protein-binding specificity of the plasma, may all influence the retention of hydroxylated metabolites. Genetic diversity among individuals and populations may also influence the metabolism rates and binding specificity.

To evaluate the toxicologic significance of the phenolic fraction, the main compounds must be identified, accurately quantitated, and toxicologic studies must be undertaken to evaluate the effects of these compounds on thyroid hormone and retinol homeostasis and binding at receptor sites. PCP remains the dominating phenolic compound in Inuit whole blood, even more so for the southern Quebec pooled sample. Concentrations of PCP in most of the samples far exceeded the main OH-PCBs. Other compounds, such as the unknown chlorinated compound, will be researched further to elucidate structure and possible roles in disruption of thyroid hormone transport.

In conclusion, total OH-PCB concentrations were 11 and 33% of total PCB concentrations in Inuit and a southern Quebec pooled sample of whole blood, respectively. Both total PCB and total OH-PCB geometric mean concentrations were higher for Inuit than for the southern Quebec pooled sample and the current literature values (31). Increased concentrations resulted in increased OH-PCB levels in Inuit blood, but the ratio of metabolites to PCBs was not as high as that found for the southern Quebec pooled sample. Unlike PCBs, the pattern of OH-PCBs was not consistent among individuals. Three different congeners alternated as the dominant metabolite of PCBs in the 30 samples analyzed. A larger study would help elucidate the reasons for the pattern variability, the lower ratio of OH-PCBs to PCBs in Inuit, and the possible differences between the sexes. PCP was the dominant chlorinated phenolic compound in the majority of the samples and must be included in future studies that may be evaluating the effects of phenolic compounds on circulating levels of thyroid hormones and vitamin A.

REFERENCES AND NOTES

(1.) Jensen S. The PCB story. Ambio 1:123-131 (1972).

(2.) Letcher RJ, Klasson-Wehler E, Bergman A. Methyl sulfone and hydroxylated metabolites of polychlorinated biphenyls. In: The Handbook of Environmental Chemistry--New Types of Persistent Halogenated Compounds (Paasivirta J, ed). Berlin:Springer-Verlag, 1999;317-359.

(3.) Sipes IG, Slocumb ML, Perry DF, Carter DE. 2,4,5,2',4',5'-Hexachlorobiphenyl: distribution, metabolism, and excretion in the dog and monkey. Toxicol Appt Pharmacol 65:264-272 (1982).

(4.) Schnellman RG, Putnam CW, Sipes IG. Metabolism of 2,3',3,3',6,6'-hexachlorobiphenyl and 2,2',4,4',5,5'-hexachlorobiphenyl by human hepatic microsomes. Biochem Pharmacol 32:3233-3239 (1983).

(5.) Ariyoshi N, Koga N, Oguri K, Yoshimura H. Metabolism of 2,4,5,2',4',5'-hexachlorobiphenyl with liver microsomes of phenobarbital-treated dog: the possible formation of PCB 2,3-arene oxide intermediate. Xenobiotica 22:1275-1290 (1992).

(6.) Jansson B, Jensen S, Olsson M, Renberg L, Sundstrom G, Vaz R. Identification by GC-MS of phenolic metabolites of PCB and p,p'-DDE isolated from Baltic guillemot and seal. Ambio 6:93-97 (1975).

(7.) Caldwell J. Xenobiotic Conjugation Chemistry. Washington DC:American Chemical Society, 1986.

(8.) Rickenbacher U, McKinney JD, Oatley SJ, Blake CCF. Structurally specific binding of halogenated biphenyls to thyroxine transport protein. J Med Chem 29:641-648 (1986).

(9.) Brouwer A, Klasson-Wehler E, Bokdam MM. Competitive inhibition of thyroxin binding to transthyretin by monohydroxy metabolites of 3,4,3'4'-tetrachlorobiphenyl. Chemosphere 20:1257-1262 (1990).

(10.) Lans MC, Klasson-Wehler E, Brouwer A. Thyroid hormone binding proteins as targets for hydroxylated PCB, PCDD and PCDF metabolites; an overview. Organohalogen Compounds 20:481-485 (1994).

(11.) Lans MC, Klasson-Wehler E, Willemsen M, Meussen E, Safe S, Brouwer A. Structure-dependent, competitive interaction of hydroxy-polychlorobiphenyls, -dibenzo-p-dioxins and -dibenzofurans with human transthyretin. Chem Biol Interact 88:7-21 (1993).

(12.) Damerund PO, Morse D, Klasson-Wehler E, Brouwer A. Binding of a 3,3',4,4'-tetrachlorobiphenyl (CB-77) metabolite to fetal transthyretin and effects on fetal thyroid hormone levels in mice. Toxicology 106:105-114 (1996).

(13.) Brouwer A, van den Berg KJ. Binding of a metabolite of 3,4,3',4'-tetrachlorobiphenyl to transthyretin reduces serum vitamin A transport by inhibiting the formation of the protein complex carrying both retinol and thyroxin. Toxicol Appl Pharmacol 85:301-312 (1986).

(14.) Bergman A, Klasson-Wehler E, Kuroki H. Selective retention of hydroxylated PCB metabolites in blood. Environ Health Perspect 102:464-469 (1994).

(15.) van den Berg KJ. Interaction of chlorinated phenols with thyroxine binding sites of human transthyretin, albumin and thyroid binding globulin. Chem Biol Interact 76:63-75 (1990).

(16.) Kuhnlein HV, Soueida R, Receveur O. Baffin Inuit food use by age, gender and season. J Can Diet Assoc 56:175-183 (1995).

(17.) Kuhnlein HV, Kubow S, Soueida R. Lipid components of traditional Inuit foods and diets of Baffin Island. J Food Comp Anal 4:227-236 (1991).

(18.) Kuhnlein HV. Benefits and risks of traditional food for indigenous peoples: focus on dietary intakes of Arctic men. Can J Physiol Pharmacol 73:765-771 (1995).

(19.) Cameron M, Weis IM. Organochlorine contaminants in the country food diet of the Belcher Island Inuit, Northwest Territories, Canada. Arctic 46:42-48 (1993).

(20.) Stern GA, Muir DCG, Segstro M, Dietz R, Heide-Jorgenson M. PCBs and other organochlorine contaminants in white whales (Delphinapterus leucas) from West Greenland: variations with age and sex. Medd Gronl Biosci 39:245-259 (1994).

(21.) Newsome WH, Davies DJ, Sun WF. Residues of polychlorinated biphenyls (PCB) in fatty foods of the Canadian diet. Food Addit Contam 15:19-29 (1998).

(22.) Ayotte P, Dewailly E, Ryan JJ, Bruneau S, LeBel G. PCBs and dioxin-like compounds in plasma of adult Inuit living in Nunavik (Arctic Quebec). Chemosphere 34:1459-1468 (1997).

(23.) Seegal RF. Epidemiological and laboratory evidence of PCB-induced neurotoxicity. Crit Rev Toxicol 26:709-737 (1996).

(24.) Sjodin A, Hagmar L, Klasson-Wehler E, Bjork J, Bergman A. Unpublished data.

(25.) Ballschmiter K, Zell M. Analysis of polychlorinated biphenyls (PCB) by glass capillary gas chromatography. Composition of technical Aroclor- and Clophen-PCB mixtures. Fresenius Z Anal Chem 302:20-31 (1980).

(26.) Guitart R, Puig P, Gomez-Catalan J. Requirement for a standardized nomenclature criterium for PCBs: computer-assisted assignment of correct congener denomination and numbering. Chemosphere 27:1451-1459 (1993).

(27.) Hovander L, Athanasiadou M, Asplund L, Jensen S, Klasson-Wehler E. Extraction and cleanup methods for analysis of phenolic and neutral organic halogenated substances in plasma. J Anal Toxicol (in press).

(28.) Blatt AH. Organic Syntheses--Collective Volume 2. New York:John Wiley & Sons, Inc., 1943.

(29.) Bergman A, Klasson-Wehler E, Kuroki H, Nilsson A. Synthesis and mass spectrometry of some methoxylated PCB. Chemosphere 30:1921-1938 (1995).

(30.) Norstrom RJ, Belikov SE, Born EW, Garner GW, Malone B, Olpinski S, Ramsay MA, Schliebe S, Stirling I, Stishov MS, et al. Chlorinated hydrocarbon contaminants in polar bears from eastern Russia, North America, Greenland, and Svalbard: biomonitoring of Arctic pollution. Arch Environ Contain Toxicol 35:354-367 (1998).

(31.) Klasson-Wehler E, Hovander L, Bergman A. New organohalogens in human plasma--identification and quantification. Organohalogen Compounds 33:420-425 (1997).

(32.) Klasson-Wehler E, Bergman A, Athanasiadou M, Ludwig JP, Auman HJ, Kannan K, van den Berg M, Murk AJ, Feyk LA, Giesy JP. Hydroxylated and methylsulfonyl polychlorinated biphenyl metabolites in albatrosses from Midway atoll, North Pacific Ocean. Environ Toxicol Chem 17:1620-1625 (1988).

(33.) Sandau CD, Norstrom RJ. Comparison of methylation methods for the determination of hydroxy-PCBs and preliminary results for polar bear blood plasma. Organohalogen Compounds 29:412-417 (1996).

(34.) Geyer H J, Scheunert I, Korte F. Distribution and bioconcentration of the environmental chemical pentachlorophenol (PCP) in different tissues of humans. Chemosphere 16:887-899 (1987).

(35.) Robbins J. Thyroid hormone transport proteins and the physiology of hormone binding. In: Werner and Ingbar's The Thyroid--A Fundamental and Clinical Text, Vol 6 (Braverman LE, Utiger RD, eds). Philadelphia, PA:J.B. Lippincott Company, 1991;111-143.

(36.) Newsome WH, Davies D. Determination of PCB metabolites in Canadian human milk. Chemosphere 33:559-565 (1996).

(37.) Dewailly E, Nantel A, Weber JP, Meyer F. High levels of PCBs in breast milk of Inuit women from arctic Quebec. Bull Environ Contain Toxicol 43:641-646 (1989).

(38.) Dewailly E, Weber J-P, Gingras S, Laliberte C. Coplanar PCBs in human milk in the province of Quebec, Canada: are they more toxic than dioxin for breast fed infants? Bull Environ Contain Toxicol 47:491-498 (1991).

(39.) Dewailly E, Ayotte P, Laliberte C, Weber J-P, Gingras S, Nantel AJ. Polychlorinated biphenyls (PCBs) and dichlorodiphenyl dichloroethylene (DDE) in breast milk of women living in the province of Quebec, Canada. Am J Public Health 86:1241-1246 (1996).

(40.) Silverman DI, Reis GJ, Sacks FM, Boucher TM, Pasternak RC. Usefulness of plasma phospholipid N-3 fatty acid levels in predicting dietary fish intake in patients with coronary artery disease. Am J Cardiol 66:860-862 (1990).

(41.) Sante Quebec. A Health Profile of the Inuit: Report of the Sante Quebec Health Survey Among the Inuit of Nunavik, Vol 1 (Jette M, ed). Montreal, Quebec, Canada:Ministere de la Sante et des Service Sociaux, Gouvernement du Quebec, 1994.

(42.) Klasson-Wehler E, Lindberg L, Jonsson C-J, Bergman A. Tissue retention and metabolism of 2,3,4,3',4'-pentachlorobiphenyl in mink and mouse. Chemosphere 27:2397-2412 (1993).

(43.) Frame GM, Cochran JW, Bowadt SS. Complete PCB congener distributions for 17 Aroclor mixtures determined by 3 HRGC systems optimized for comprehensive, quantitative, congener specific analysis. J High Resolut Chromatogr 19:657-668 (1996).

Courtney D. Sandau,(1),(2) Pierre Ayotte,(3) Eric Dewailly,(3) Jason Duffe,(2) and Ross J. Norstrom(1),(2)

(1) Centre for Analytical and Environmental Chemistry, Department of Chemistry, Carteton University, Ottawa, Ontario, Canada; (2) Environment Canada, Canadian Wildlife Service, National Wildlife Research Centre, Hull, Quebec, Canada; (3) Unite de recherche en sante publique, Centre de Recherche du CHUL-Centre Hospitalier Universitaire de Quebec, Beauport, Quebec, Canada

Address correspondence to R. Norstrom, Environment Canada, Canadian Wildlife Service, National Wildlife Research Centre, 100 Gamelin Boulevard, Hull, Quebec, K1A OH3, Canada. Telephone: [phone] Fax: [phone]. E-mail: [e-mail]

We thank A. Bergman and E. Klasson-Wehler for their contributions toward this publication. We obtained informed consent from all individuals before their participation in this study.

The project was funded by the Canadian Chlorine Coordinating Committee (C4) and the Canadian Chemical Producers Association.

Received 20 September 1999; accepted 17 December 1999.