Polychlorinated Biphenyl Congener Patterns in Fish near the Hanford

Jan 26, 2015 - Six factors were resolved with PMF2 software. Depletion and enhancement of PCB congeners in factors, relative to Aroclor 1254, suggeste...
0 downloads 10 Views 3MB Size
Download PDF
Article pubs.acs.org/est

Polychlorinated Biphenyl Congener Patterns in Fish near the Hanford Site (Washington State, USA) Lisa A. Rodenburg,† Damon Delistraty,*,‡ and Qingyu Meng§ †

Department of Environmental Science, Rutgers University, 14 College Farm Road, New Brunswick, New Jersey 08901, United States Washington State Department of Ecology, 4601 North Monroe Street, Spokane, Washington 99205-1295, United States § School of Public Health, Rutgers University, Piscataway, New Jersey, United States ‡

S Supporting Information *

ABSTRACT: It is well-known that absorption, distribution, metabolism, and excretion (ADME) processes in fish can alter polychlorinated biphenyl (PCB) congener patterns in fish, but these patterns have never been investigated using an advanced source-apportionment tool. In this work, PCB congener patterns in freshwater fish were examined with positive matrix factorization (PMF). PCB congeners were quantified via EPA Method 1668 in fillet and carcass of six species in four study areas in the Columbia River near the Hanford Site. Six factors were resolved with PMF2 software. Depletion and enhancement of PCB congeners in factors, relative to Aroclor 1254, suggested biotransformation (via cytochrome P450) and bioaccumulation in fish, respectively. Notable differences were observed among species and across study locations. For example, sturgeon and whitefish exhibited congener patterns consistent with Aroclor weathering, suggesting potential PCB metabolism in these species. In terms of location, average concentration of total PCBs for all species combined was significantly higher (P < 0.05) at Hanford 100 and 300 areas, relative to upriver and downriver study sites. Furthermore, a distinct PCB signature in sturgeon and whitefish, collected at Hanford study areas, suggests that Hanford is a unique PCB source.

■

INTRODUCTION PCBs are persistent organic pollutants (POPs) that have been detected in the Columbia River environment at the Hanford Site (Figure 1).1 The Hanford Site was acquired by the United States government in 1943 in order to produce plutonium for some of the nuclear weapons tested and used in World War II.2 The site is located in southeastern Washington State and includes a section of the Columbia River that provided a source of clean water to cool nine plutonium production reactors. Historic operations at Hanford resulted in the release of both radiological and nonradiological wastes to the environment. Some of these wastes have entered the Columbia River and have accumulated to varying extents in aquatic biota, including fish. The sources of PCBs in the Columbia River system are likely tied to both Hanford and non-Hanford activities. For example, Hanford PCBs were associated with use of electrical devices (e.g., transformers and capacitors) as well as paint and sealants.3 In contrast, PCBs can be transported from more distant environmental reservoirs (e.g., soils) over long distances via successive volatilization/condensation cycles in the global atmospheric circulation.4 PCBs can also be dechlorinated by bacteria in anaerobic environments such as river sediment and even landfills, contaminated groundwater, and sewers.5−8 These processes © XXXX American Chemical Society

can alter PCB congener patterns in characteristic ways, such that examination of congener patterns can suggest likely sources of PCBs. In addition, it is well-known that absorption, distribution, metabolism, and excretion (ADME) processes alter PCB congener patterns in fish.9−12 Although ADME processes have been extensively studied, usually these studies are carried out on single congeners and/or under laboratory conditions. Although controlled laboratory studies can indicate what can happen, a major advantage of factor analysis is that it can indicate what does actually happen to complex mixtures in real environments. From a toxicokinetic perspective, the tissue concentration of an environmental contaminant is a function of the composite action of ADME processes.13 Due to their lipophilicity, PCBs are readily absorbed but not easily eliminated, leading to their bioaccumulation. Furthermore, biotransformation of PCBs may result in either detoxification (via subsequent conjugation reactions) or activation to toxic metabolites.12 In particular, cytochrome P450 monooxygenases are induced in fish exposed to PCBs.14 Cytochrome P450 enzymes Received: October 14, 2014 Revised: January 9, 2015 Accepted: January 26, 2015

A

DOI: 10.1021/es504961a Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

of PCBs (which typically facilitates elimination), toxicity has been shown to be enhanced (relative to that of parent compounds) in rainbow trout with respect to endocrine disruption.26 Overall, metabolic pathways involving P450 enzymes have been shown to alter the PCB congener profile in fish, which in turn, modify associated toxicity.27,28 Factor analysis is an important tool that can be used to discern the sources of contaminants such as PCBs via the examination of congener patterns.29 The most popular among the first generation of factor analysis tools was principle components analysis (PCA), which has been used to investigate PCB congener patterns in fish30 and many other references. However, second generation tools such as positive matrix factorization (PMF) and polytopic vector analysis (PVA) have, to our knowledge, never been applied to fish tissue data. As part of a recent risk assessment supporting Hanford cleanup, PCB congener data were collected in several species of fish in the Columbia River near Hanford.31 This data set has previously been evaluated with respect to potential ecotoxicity and risk to human fish consumers.32 The purpose of the present study was to identify PCB congener patterns in these fish using PMF methods to determine the likely sources of PCBs as well as the importance of ADME processes occurring in the fish.

■

METHODS Study Area. The study area includes a section of the Columbia River in southeastern Washington State, because of its proximity to the Hanford Site (Figure 1). This area begins upriver from the Hanford Site above the Wanapum Dam and continues to the McNary Dam, the first dam downriver from the Hanford Site. The study area was divided into four subareas, bounded by river mile (RM) markers. River miles are shown in Figure 1 because these are a conventional measure (in the USA) of distance upriver from the river mouth. Subareas included Upriver (RM 440−388), 100 Area (RM 387−366), 300 Area (RM 365−340), and Lake Wallula (RM 339−292).33 Fish Collection and Sample Preparation. Six fish species were collected from the study area from 2009 to 2010, including common carp (Cyprinus carpio), mountain whitefish (Prosopium williamsoni), walleye (Stizostedion vitreum), smallmouth bass (Micropterus dolomieu), bridgelip sucker (Catostomus columbianus), and white sturgeon (Acipenser transmontanus). These species are resident fish and are consumed by the local human population.33 For nonsturgeon species, fish were grouped into composites by species, tissue, and location. In general, five composite samples (each consisting of 5−7 individual fish) were obtained for each species from each subarea. Fillets for each composite sample were combined, homogenized in a food grinder, and frozen. Carcass composite samples were processed in a manner similar to fillet composites. Fillet consisted of muscle, skin, and scales, whereas carcass consisted of bones, head, and fins. For sturgeon, samples were from 30 individual fish (rather than composites) and grouped by tissue and location. Five to ten sturgeon were collected from each subarea. Fillet, carcass, liver, and viscera samples were obtained by dissection and frozen. Fillet consisted of muscle, whereas carcass consisted of bones, head, fins, and skin. Viscera consisted of entrails (internal organs less the stomach, liver, and kidney). Details of fish collection and sample preparation are described elsewhere.34 Length and weight of fish species are shown in Table S-1. Ages were determined only for sturgeon,

Figure 1. Columbia River study area near the Hanford Site, divided into four subareas (dashed rectangles).32 Numbers along the river refer to river miles (see text) and reactor areas (100-B/C, 100-K, 100-N, 100-D, 100-H, and 100-F; small boxes).

(designated with the abbreviation CYP, followed by a number indicating gene family and an uppercase letter indicating subfamily) catalyze a wide variety of phase I reactions and are expressed in nearly all tissues.15 Biotransformation of PCBs is dependent on the chlorine substitution pattern. It appears that greater metabolism of nonplanar PCBs by CYP2B occurs with vicinal hydrogen atoms in the meta and para positions.9−12 In addition, metabolism of planar PCB congeners by CYP1A occurs with vicinal hydrogen atoms in the ortho and meta positions in combination with ≤1 ortho chlorine atom.10,12 Planar PCBs include those congeners with dioxin-like toxicity mediated via the Ah receptor, whereas nonplanar PCBs congeners exhibit a range of toxicities independent of the Ah receptor.16 Adverse effects attributed to nonplanar PCB congeners are diverse (e.g., disruption of calcium homeostasis, changes in protein kinase C translocation, inhibition of dopamine uptake, and formation of reactive oxygen species).17−20 In fish, increased activity of ryanodine receptors after exposure to nonplanar PCBs has been observed in skeletal muscle of rainbow trout.21 Because the first step in P450 biotransformation is the insertion of oxygen into the PCB molecule10 and formation of arene oxide intermediates,16,22,23 hydroxylated and methylsulfonyl PCBs in fish may be derived from P450 activity.12,24 Some of these PCB derivatives may also exhibit toxicity.25 For example, although hydroxylation increases the water solubility B

DOI: 10.1021/es504961a Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 2. Six resolved factors. Fraction of total PCBs for each factor is shown on the y axis, and PCB congener number is displayed on the x axis. The contribution of each factor to the total PCB mass in the data set is given in the label in each panel. Error bars (mostly invisible) represent one standard deviation based on nine seed runs.

0.25 mm internal diameter, 0.25 μm film thickness). PCB concentrations in fish tissue were expressed in wet weight for individual congeners and total PCBs (e.g., mg/kg) or as a fraction or percent of total PCBs. In addition to fish tissue, a limited set of samples from abiotic media in the four study areas (i.e., soils, n = 8; surface water, n = 8; and sediment, n = 34)

ranging from 6 to 25 years (mean = 15.1 years, SD = 4.7 years, n = 28). PCB Congener Analysis. PCB congeners were quantified in fish tissues by high-resolution gas chromatography/highresolution mass spectrometry (HRGS/HRMS) with USEPA Method 1668A,35 using an SPB-octyl column (30 m length, C

DOI: 10.1021/es504961a Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

LOD. As a result, BDL values were replaced with one-quarter of the lowest detected concentration. Other Statistical Tests. A series of one-way analysis of variance (ANOVA) tests were carried out to assess differences in the distributions of PCB factors among tissues, species, and locations. Least significant difference (LSD) tests were applied as a post-hoc analysis when ANOVA revealed a significant effect. Cosine theta (cos Θ) values were calculated to assess the association between each PMF factor versus individual Aroclor patterns, using data from Rushneck et al.40 Multiple linear regressions were carried out to evaluate the contribution of four standard Aroclors in combination (i.e., Aroclors 1242, 1248, 1254, and 1260) to each PMF factor. Partial regression coefficients (bk) for a set k of independent variables and cos Θ values were calculated in multiple regression analyses. The Bonferroni inequality was used to maintain the overall α at 0.05 in each set of analyses.41,42 Calculations were carried out within Microsoft Excel as well as with Statgraphics software.43

was collected and analyzed for PCB congeners with the same analytical method to assist in interpretation of congener patterns. Quality assurance and quality control (QA/QC) methods evaluated a subset of PCB congener data analyzed with Method 1668A in soil, surface water, sediment, and fish tissue.36 QA/QC information is summarized in Supporting Information. PMF Modeling. PCB concentration data (wet weight) were analyzed using PMF2 software.37 The master data set consisted of 293 samples in which 209 PCB congeners were measured in 160 peaks. In this data matrix, about 34% of all the data points were below detection limit (BDL). This matrix was reduced by discarding congeners that were BDL in more than 25% of samples, with the exception that PCB 4 was retained. PCB 4 was retained even though it was BDL in 154 samples (53%) because it is a marker for microbial dechlorination of PCBs.7,38 Because of the low concentrations and high number of values below detection for this congener, the PMF results (i.e., the assignment of the mass of this congener across the different factors) may be less reliable than for the other congeners. This resulted in 76 congeners (or coeluting congener groups) in the data matrix. For each number of factors, the model was run nine times using nine different seed values. In essence, this means originating the model in nine different “places” to ensure that they all converge on the same solution, suggesting that the solution is robust and represents a global, not local, minimum. After several preliminary runs, it was determined that two samples were causing instability in the PMF model. (The relative standard deviation (RSD) of the nine seed runs was high, and the nine runs generally fell into two groups of solutions.) These two samples had the highest concentrations in the data matrix and were outliers in the plots of measured versus predicted concentrations for several congeners. This was confirmed via Grub’s test of outliers, using the log-transformed ΣPCB data (because it was log-normally distributed) with an alpha (α) of 0.05.39 These two samples represent the fillet and carcass from one whitefish composite obtained from the Upriver location. In some cases, the influence of high concentrations (outliers) in the data matrix can be overcome by normalizing the data (i.e., expressing all congeners as a percent of the sum of the concentrations of all congeners). In the present case, the PMF2 solution from the normalized data was far inferior to the solution obtained when the two outliers were discarded. Thus, the final data matrix consisted of 76 congeners (or coeluting congener groups) in 291 samples. Of the remaining 22 116 data points, 901 (4.1%) were BDL, as a result of data reduction described above. The uncertainty matrix took the form of (x,3x), where x is the uncertainty in the detected concentrations and was the same for all measurements of a single congener, and three times this uncertainty (3x) was applied to values BDL. The RSD of the percent recoveries of the surrogate 13C-labeled PCB congeners was used to calculate the uncertainty for the detected concentrations (x), which ranged from 6.8 to 17.3%. Where Method 1668A requires the use of the average recovery of several surrogates, the uncertainty was propagated from the uncertainties (RSDs) for the individual surrogates. Because the limits of detection (LODs) were not reported, the LOD matrix was constructed using one-half of the lowest detected concentration for each congener, whereas BDL values were replaced with one-half the

■

RESULTS AND DISCUSSION PMF Model. A critical step in any factor analysis is determining the correct number of factors that provide clear and physically meaningful results and that at the same time reduce matrix dimensionality. The correct number of factors was determined to be six, on the basis of a weight of evidence approach (see Supporting Information). The six factors resolved by the PMF2 model are shown in Figure 2. Most of the PCB mass in the data set (about 76%) resides in factors 2, 3, and 6. Because the congeners were listed in order of increasing molecular weight (MW) in the input matrix, the average MW of the factors increased with factor number from 311 (factor 1) to 369 g/mol (factor 6). Comparison to Aroclors, Sediment, and Potential Biotransformation by Fish. Factors were compared with each individual Aroclor congener pattern (Table 1), using data from Rushneck et al.40 Some factors strongly resembled Table 1. Cosine Theta (cos Θ) Values for Comparisons of the Six Resolved Factors (n = 76) with the Four Main Aroclors and Partial Regression Coefficients (bk) for the Description of Each Factor as a Mixture of Aroclors single Aroclors cos Θ factor

1242

F1 F2 F3 F4 F5 F6

0.60 0.16 0.17 0.10 0.06 0.04

1248

1254

1260

0.87 0.74 0.37 0.94 0.36 0.90 0.21 0.62 0.14 0.61 0.08 0.44 best-fit mixture of Aroclors

0.28 0.58 0.63 0.59 0.91 0.79 cos Θ

bk

a

D

factor

1242

1248

1254

1260

mixture of Aroclors

F1 F2 F3 F4 F5 F6

0 0 0 0 0 0

0.64a 0 0 0 0 0

0.31a 0.93a 0.80a 0.61a 0.23a 0.12

0.07 0.18a 0.28a 0.54a 0.77a 1.01a

0.93 0.96 0.93 0.71 0.93 0.79

P < 0.05 for nonzero bk. DOI: 10.1021/es504961a Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Figure 3. Congener pattern of Aroclor 1254 subtracted from the congener patterns of the six resolved factors. Percent of total PCBs for these differences (i.e., factor − Aroclor) is shown on the y axis, and PCB congener number is displayed on the x axis. Positive-value bars indicate higher congener proportions in the factor than in the Aroclor (i.e., enhancement), and negative-value bars indicate lower proportions in the factor than in the Aroclor (i.e., depletion).

Aroclors. Factor 1 resembled Aroclor 1248 (cos Θ = 0.87), factor 2 strongly resembled Aroclor 1254 (cos Θ = 0.94), factor 3 resembled Aroclor 1254 (cos Θ = 0.90), and factor 5 resembled Aroclor 1260 (cos Θ = 0.91). For the other factors,

the resemblance to unweathered Aroclors was less strong. We interpreted factor 4 to be weathered beyond recognition, whereas factor 6 resembled weathered Aroclor 1260 (cos Θ = 0.79). With the exception of a weak signal from factor 1 (Table E

DOI: 10.1021/es504961a Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

Table 2. Depletion and Enhancement of PCB Congeners in Resolved Factors, Relative to Aroclor 1254, in Relation to Cytochrome P450 Biotransformation depleted congeners (single or coeluters)a

enhanced congeners (single or coeluters)a 129 + 138* + 160 + 163 153* + 168 180* + 193 187

52 44* + 47* + 65 61 + 70* + 74* + 76 86 + 87* + 97 + 108 + 119 + 125 90 + 101* + 113 110* + 115 118

Cl substitution patternb 22′344′5′ 22′44′55′ 22′344′55′ 22′34′55′6 22′55′ 22′35′(44), 22′44′ (47) 23′4′5(70), 244′5 (74) 22′345′ 22′455′ 233′4′6 23′44′5

number pairs of ortho and meta vicinal H atomsb 1 0 0 0 0 1(44), 2(47) 1(70), 2(74) 1 0 1 1

number pairs of meta and para vicinal H atomsb 0 0 0 0 2 2(44), 0(47) 1(70), 0(74) 1 1 1 0

number ortho-Cl atomsb

metabolic groupb,c

2 2 2 3 2 2(44), 2(47) 1(70), 1(74) 2 2 2 1

2 1 1 1 4 4(44), 2(47) 4(70), 3(74) 4 4 4 3

a

Predominant congeners (in a coeluted group) marked with an asterisk (*), on the basis of Frame et al.46 bApplies to single or predominant congeners. cCriteria for metabolic groups:10 1-congeners without any vicinal H atoms, 2- congeners with vicinal H atoms exclusively in the ortho and meta positions in combination with ≥2 ortho-Cl substituents, 3-congeners with vicinal H atoms in the ortho and meta positions in combination with ≤1 ortho-Cl, 4-congeners with vicinal H atoms in the meta- and para-positions in combination with ≤2 ortho-Cl, and 5-congeners with vicinal H atoms in the meta and para positions in combination with ≥3 ortho-Cl.

from 100 Area and one from Upriver, more closely resembled Aroclor 1260. To determine which congeners were most affected by fish metabolism, the congener pattern of Aroclor 1254 was subtracted from the congener patterns of the resolved factors (Figure 3). In comparison with Aroclor 1254 (Figure 3), congeners 129 + 138 + 160 + 163 (mostly PCB 138), 153 + 168 (mostly 153), 180 + 193 (mostly 180), and 187 are enhanced in the factors, whereas congeners 52, 44 + 65 + 47 (mostly 44 and 47), 61 + 70 + 74 + 76 (mostly 70 and 74), 86 + 87 + 97 + 108 + 119 + 125 (mostly 87), 90 + 101 + 113 (mostly 101), 110 + 115 (mostly 110), and 118 are depleted. (The designation of which congeners are dominant in each coeluting congener group was made on the basis of the analyses of Aroclors by Frame et al.46 and Rushneck et al.40) Relative to Aroclor 1254, enhanced and depleted congeners have been assigned one of five metabolic groups (Table 2). In particular, metabolic groups 1 and 2 are generally more resistant to metabolism, whereas groups 3−5 are more readily metabolized by CYP1A (group 3) and CYP2B (groups 4 and 5).10,12 Notably, depleted congeners (PCBs 52, 44 + 47, 70 + 74, 87, 101, 110, and 118) are not substituted at adjacent positions, often the meta/para or ortho/meta positions, where cytochrome P450 oxidation may occur. In particular, PCB 52 is known to be metabolized in fish, relative to other PCB congeners.12 Alternatively, enhanced PCBs (PCBs 153, 180, and 187) have substituted adjacent sites where oxidation does not occur. It is not surprising that congeners such as PCB 153 are enhanced in the fish because this congener is well-known for preferentially bioaccumulating.47 PCBs 138 and 163 are exceptions. Although they have an open ortho/meta configuration, they are also enhanced relative to Aroclor 1254. None of the PCB congeners in our study were assigned to metabolic group 5. Overall, however, these results are consistent with the metabolic groupings specified by Boon et al.10 Comparison to Non-Aroclor Sources. Notably absent from this data matrix are factors related to non-Aroclor sources of PCBs such as pigments (e.g., PCB 11) and titanium dioxide (e.g., PCB 209). These congeners are often associated with

1), any congener pattern similar to Aroclor 1242 is minimal, despite the fact that Aroclors 1242 and 1016, which have similar chlorine content, together made up about 64% of Aroclor production in the United States.44 In thePMF analysis of other data sets involving water column and sediment concentrations of PCBs, we have similarly been unable to identify a factor corresponding to Aroclor 1242/1016.7,29 To determine whether any of the factors represented mixtures of Aroclors,45 a multiple linear regression was carried out in which a congener pattern was calculated that represented a linear combination of the four main Aroclors, represented as Cf = aC1242 + bC1248 + cC1254 + dC1260

where C is the concentration of the resolved factor (f) or individual Aroclor and a, b, c, and d are partial regression coefficients (bk). Coefficients were constrained to be positive, and cos Θ values for this best-fit composite Aroclor-congener pattern versus factor-congener pattern were calculated (Table 1). Factors 1−3 and 5 are fairly well-described as a combination of Aroclors (cos Θ > 0.9), suggesting they have undergone relatively little weathering. Factors 4 and 6 are less welldescribed as a combination of Aroclors (cos Θ 0.05), owing to relatively large standard deviations in these tissues. Differences across Fish Species. There are clear differences in PCB congener patterns across fish species. Interspecies results are shown for carcass tissue (Figure S-8). As noted above, a similar pattern was observed in fillet (Figure S9). Factor 2 is dominant in bass, carp, sucker, and walleye (P < 0.05). Sturgeon contained proportionately and significantly

more of factors 3 and 6 (P < 0.05). Factor 4 dominates in whitefish (P < 0.05). Factor 2 is thought to represent unweathered Aroclor 1254, so its dominance in these species may indicate a reduced ability to metabolize PCBs. In contrast, factors 3, 4, and 6 are the most weathered, so the abundance of these factors may indicate greater metabolism of PCBs in sturgeon and whitefish. As noted above, factors 3, 4, and 6 also contain relatively high proportions of dioxin-like congeners, suggesting that these species do not efficiently metabolize dioxin-like congeners and/or they bioaccumulate them. Moreover, this analysis suggests that metabolic pathways may be different because whitefish accumulate factor 4, whereas sturgeon accumulate factors 3 and 6 (Figure S-8). Alternatively, different congener patterns in factors 3 or 4 versus factor 6 could be explained by similar metabolism of two distinct Aroclor precursors: Aroclor 1254, leading to factors 3 and 4, and Aroclor 1260, leading to factor 6. This would imply that whitefish exhibit a PCB congener pattern derived predominantly from Aroclor 1254, whereas sturgeon display a hybrid pattern derived from Aroclors 1254 and 1260. Interspecies differences in PCB congener patterns and concentrations may also relate to varying diets and habitats.55 For example, fish species exhibiting a greater degree of piscivory (e.g., adult sturgeon) may be exposed via diet to more altered PCBs than species ingesting predominantly benthic invertebrates (e.g., sucker), given the higher biotransformation capacity of prey fish versus invertebrates.14 Although sturgeon are typically anadromous, dams along the Columbia River impede their migration, limiting their range and thereby influencing their exposure to contaminants, including PCBs. Finally, it is possible that demersal species (e.g., sturgeon) would contact more sediment, which contains a greater concentration of PCBs relative to the water column. In discussing differences across fish species, it is important to remember that two samples, the carcass and fillet from one whitefish composite, were discarded from the data matrix. These two samples had virtually identical congener patterns (R2 = 0.99), which is in agreement with our assessment above (i.e., that congener patterns of fillet showed little difference versus those of carcass from the same species). Comparing these two samples to the resolved factors suggests that they are most similar to factor 4, which dominates in whitefish (Figure S-8). Thus, these two samples are similar to other whitefish samples and support our observation that whitefish display distinct congener patterns relative to other fish species. Differences by Location. For some fish species, obvious differences across the four study locations were observed. PCB concentration and factor differences by location are shown in Figures S-10−S-12. Notably, for all fish species combined, the average total PCB concentrations in carcass were significantly higher (P < 0.05) in 100 and 300 Areas relative to those of Upriver and Lake Wallula locations (Figure S-11). A similar relationship for these locations was reported for total PCB, dioxin-like PCB, and nondioxin-like PCB concentrations in sturgeon liver with this same data set.32 Use of Aroclors 1248, 1254, and 1260 has been reported at the Hanford Site,3 and evidence for all three of these Aroclors is reflected in resolved factors in fish-tissue data (Figure 2). However, linkage to specific Hanford sources remains uncertain because of fish mobility, as well as PCB input by possible non-Hanford sources. In terms of individual species, total PCB concentrations in carcass generally differed by less than 2-fold across locations G

DOI: 10.1021/es504961a Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

(3) Hermann, R. M. Polychlorinated biphenyl presence in the Columbia River corridor; Technical report prepared for the U.S. Department of Energy, Richland Operations Office, report no. WCH-208; Washington Closure Hanford: Richland, WA, 2007. (4) Simonich, S. L.; Hites, R. A. Global distribution of persistent organochlorine compounds. Science 1995, 269, 1851−1854. (5) Bzdusek, P. A.; Christensen, E. R.; Lee, C. M.; Pakdeesusuk, U.; Freedman, D. C. PCB Congeners and Dechlorination in Sediments of Lake Hartwell, South Carolina, Determined from Cores Collected in 1987 and 1998. Environ. Sci. Technol. 2006, 40, 109−119. (6) Bzdusek, P. A.; Lu, J.; Christensen, E. R. PCB Congeners and Dechlorination in Sediment of Sheboygan River, Wisconsin, Determined by Matrix Factorization. Environ. Sci. Technol. 2006, 40, 120−129. (7) Rodenburg, L. A.; Du, S.; Fennell, D. E.; Cavallo, G. J. Evidence for Widespread Dechlorination of Polychlorinated Biphenyls in Groundwater, Landfills, and Wastewater Collection Systems. Environ. Sci. Technol. 2010, 44, 7534−7540. (8) Rodenburg, L. A.; Du, S. Y.; Lui, H.; Guo, J.; Oseagulu, N.; Fennell, D. E. Evidence for Dechlorination of Polychlorinated Biphenyls and Polychlorinated Dibenzo-p-Dioxins and -Furans in Wastewater Collection Systems in the New York Metropolitan Area. Environ. Sci. Technol. 2012, 46, 6612−6620. (9) Matthews, H. B.; Dedrick, R. L. Pharmocokinetics of PCBs. Annu. Rev. Pharmacol. Toxicol. 1984, 24, 85−103. (10) Boon, J. P.; van der Meer, J.; Allchin, C. R.; Law, R. J.; Klunsøyr, J.; Leonards, P. E. G.; Spliid, H.; Storr-Hansen, E.; McKenzie, C.; Wells, D. E. Concentration-dependent changes of PCB patterns in fish-eating mammals: Structural evidence for induction of cytochrome P450. Arch. Environ. Contam. Toxicol. 1997, 33, 298−311. (11) Warner, N. A.; Norstrom, R. J.; Wong, C. S.; Fisk, A. T. Enantiomeric fractions of chiral polychlorinated biphenyls provide insights on biotransformation capacity of arctic biota. Environ. Toxicol. Chem. 2005, 24, 2763−2767. (12) Buckman, A. H.; Wong, C. S.; Chow, E. A.; Brown, S. B.; Solomon, K. R.; Fisk, A. T. Biotransformation of polychlorinated biphenyls (PCBs) and bioformation of hydroxylated PCBs in fish. Aquat. Toxicol. 2006, 78, 176−185. (13) Lehman-McKeeman, L. D. Absorption, distribution, and excretion of toxicants. In Casarett and Doull’s toxicology: The basic science of poisons; Klaassen, C. D. E., Ed.; McGraw-Hill Medical: New York, 2008; pp 131−159. (14) Stegeman, J. J.; Lech, J. J. Cytochrome P-450 monooxygenase systems in aquatic species - carcinogen metabolism and biomarkers for carcinogen and pollutant exposure. Environ. Health Perspect. 1991, 90, 101−109. (15) Guengerich, F. P. Cytochrome P450 Enzymes. Am. Sci. 1993, 81, 440−447. (16) Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological profile for polychlorinated biphenyls (PCBs); U.S. Department of Health and Human Services, Public Health Service: Atlanta, GA, 2000. (17) Simon, T.; Britt, J. K.; James, R. C. Develoment of a neurotoxic equivalence scheme of relative potency for assessing the risk of PCB mixtures. Regul. Toxicol. Pharmacol. 2007, 48, 148−170. (18) Hamers, T.; Kamstra, J. H.; Cenijn, P. H.; Pencikova, K.; Palkova, L.; Simeckova, P.; Vondracek, J.; Andersson, P. L.; Stenberg, M.; Machala, M. In Vitro Toxicity Profiling of Ultrapure Non-Dioxinlike Polychlorinated Biphenyl Congeners and Their Relative Toxic Contribution to PCB Mixtures in Humans. Toxicol. Sci. 2011, 121, 88−100. (19) Al-Salman, F.; Plant, N. Non-coplanar polychlorinated biphenyls (PCBs) are direct agonists for the human pregnane-X receptor and constitutive androstane receptor, and activate target gene expression in a tissue-specific manner. Toxicol. Appl. Pharmacol. 2012, 263, 7−13. (20) Wigestrand, M. B.; Stenberg, M.; Walaas, S. I.; Fonnum, F.; Andersson, P. L. Non-dioxin-like PCBs inhibit H-3 WIN-35,428 binding to the dopamine transporter: A structure-activity relationship study. Neurotoxicology 2013, 39, 18−24.

(Figure S-11). Exceptions included whitefish, which varied by approximately 3-fold, and walleye, which varied over 8-fold. However, neither of these differences was significant (P > 0.05), which is again due to relatively large standard deviations for these species. Distribution of the six factors by location for carcass in each fish species is shown in Figure S-12. Spatial differences in congener patterns were most pronounced for sturgeon and whitefish. Sturgeon contained more of factor 6 in the 100 and 300 Areas, relative to Upriver and Lake Wallula (P < 0.05). Whitefish were dominated by factor 4 in the 300 Area and Lake Wallula (P < 0.05), whereas Upriver whitefish contained more of factor 2 (P < 0.05). Factors 4 and 6 resemble weathered Aroclors, indicating potential metabolism or dietary differences in these species at these locations. Spatial differences were least pronounced for bass, which contained a fairly constant congener pattern across all locations. Viewing spatial distributions of factors overall (Figure S-12), a distinction in PCB congener patterns for Hanford locations (100 and 300 Areas) versus offsite study areas (Upriver and Lake Wallula) appears discernible. This observation is most prominent in sturgeon and whitefish. For at least some fish species, therefore, a distinct signature suggests that Hanford is a unique PCB source, consistent with reported Aroclor use onsite.3

■

ASSOCIATED CONTENT

S Supporting Information *

Quality assurance/quality control (QA/QC) information, justification for choosing six factors, tables on fish morphometrics and Aroclor versus sediment comparisons, and 12 figures on distributions of PCBs and factors by species, tissue, and location. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: (509)329-3547. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Elizabeth A. Rochette (Washington State Department of Ecology , WDOE) for constructive comments and Larry Hulstrom (Washington Closure Hanford) for providing PCB fish-tissue data. We thank Lisa Brown and Cheryl Whalen (WDOE) for supporting research activities. Although research for this study was supported by Rutgers and WDOE, results and conclusions solely represent the authors’ views and are not to be construed as endorsement or policy of the sponsoring organizations.

■

REFERENCES

(1) Riley, R. G.; Bean, R. M.; Fitzner, R. E.; Neitzel, D. A.; Rickard, W. H. A preliminary survey of polychlorinated biphenyls (PCBs) in aquatic habitats and Great Blue Herons on the Hanford Site; Technical report prepared for the United States Department of Energy, accession no. D195063320A; report nos. PNL-5685, UC-11; Pacific Northwest Laboratory: Richland, WA, 1986. (2) Hanford site environmental report for calendar year 2011; Technical report prepared for the U.S. Department of Energy, Richland Operations Office, report no. DOE/RL-2011-119; Washington Closure Hanford: Richland, WA, 2012. H

DOI: 10.1021/es504961a Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology (21) Fritsch, E. B.; Pessah, I. N. Structure-activity relationship of noncoplanar polychlorinated biphenyls toward skeletal muscle ryanodine receptors in rainbow trout (Oncorhynchus mykiss). Aquat. Toxicol. 2013, 140, 204−212. (22) Klasson-Wehler, E.; Bergman, A.; Athanasiadou, M.; Ludwig, J. P.; Auman, H. J.; Kannan, K.; Van Den Berg, M.; Murk, A. J.; Feyk, L. A.; Giesy, J. P. Hydroxylated and methylsulfonyl polychlorinated biphenyl metabolites in albatrosses from Midway Atoll, North Pacific Ocean. Environ. Toxicol. Chem. 1998, 17, 1620−1625. (23) Polychlorinated biphenyls (PCBs), Chapter 5.10 in Air Quality Guidelines for Europe; WHO Regional Publication, European Series, No. 91; World Health Organization Regional Office for Europe: Copenhagen, Denmark, 2000. (24) Stapleton, H. M.; Letcher, R. J.; Baker, J. E. Metabolism of PCBs by the deepwater sculpin (Myoxocephalus thompsoni). Environ. Sci. Technol. 2001, 35, 4747−4752. (25) Zhang, Y.; Wu, J.-P.; Luo, X.-J.; She, Y.-Z.; Mo, L.; Mai, B.-X. Methylsulfonyl polychlorinated biphenyls in fish from an electronic waste-recycling site in South China: Levels, congener profiles, and chiral signatures. Environ. Toxicol. Chem. 2012, 31, 2507−2512. (26) Carlson, D. B.; Williams, D. E. 4-hydroxy-2′,4′,6′-trichlorobiphenyl and 4-hydroxy-2′,3′,4′,5′-tetrachlorobiphenyl are estrogenic in rainbow trout. Environ. Toxicol. Chem. 2001, 20, 351−358. (27) Brown, J. F. Metabolic alterations of pcb residues in aquatic fauna: distributions of cytochrome P4501A-like and P4502B-like activities. Mar. Environ. Res. 1992, 34, 261−266. (28) White, R. D.; Shea, D.; Stegeman, J. J. Metabolism of the Aryl Hydrocarbon Receptor Agonist 3,3′,4,4′-Tetrachlorobiphenyl by the Marine Fish Scup (Stenotomus chrysops) In Vivo and In Vitro. Drug Metab. Dispos. 1997, 25, 564−572. (29) Praipipat, P.; Rodenburg, L. A.; Cavallo, G. J. Source Apportionment of Polychlorinated Biphenyls in the Sediments of the Delaware River. Environ. Sci. Technol. 2013, 47, 4277−4283. (30) Greenfield, B. K.; Allen, R. M. Polychlorinated biphenyl spatial patterns in San Francisco Bay forage fish. Chemosphere 2013, 90, 1693−1703. (31) Columbia River component risk assessment; Technical report prepared for the United States Department of Energy, report no. DOE/RL-2010-117; United States Department of Energy: Richland, WA, 2011. (32) Delistraty, D. Ecotoxicity and risk to human fish consumers of polychlorinated biphenyls in fish near the Hanford Site (USA). Sci. Total Environ. 2013, 445, 14−21. (33) Hulstrom, L. C. Data summary report for the Remedial Investigation of Hanford Site releases to the Columbia River, Hanford Site, Washington; Technical report prepared for the U.S. Department of Energy, Richland Operations Office, report no. WCH-398; Washington Closure Hanford: Richland, WA, 2011. (34) Hulstrom, L. C.; Tiller, B. L. Field summary report for Remedial Investigation of Hanford Site releases to the Columbia River, Hanford Site, Washington: Collection of fish tissue samples; Technical report prepared for the U.S. Department of Energy, Richland Operations Office, report no. WCH-387; Washington Closure Hanford: Richland, WA, 2010. (35) Method 1668, Revision A: Chlorinated Biphenyl Congeners in Water, Soil, Sediment, and Tissue by HRGC/HRMS; EPA Report No. 821-R-00-002; United States Environmental Protection Agency: Washington, DC, 2003. (36) Hulstrom, L. C. Data quality assessment report for the Remedial Investigation of Hanford Site releases to the Columbia River, Hanford Site, Washington; Technical report prepared for the U.S. Department of Energy, Richland Operations Office, Report No. WCH-381, Rev. 1; Washington Closure Hanford: Richland, WA, 2010. (37) Paatero, P.; Tapper, U. Positive Matrix Factorization: a Nonnegative Factor Model with Optimal Utilization of Error Estimates of Data Values. Environmetrics 1994, 5, 111−126. (38) Brown, J. F.; Feng, H.; Bedard, D. L.; Brennan, M. J.; Carnahan, J. C.; May, R. J. Environmental dechlorination of PCBs. Environ. Toxicol. Chem. 1987, 6, 579−593.

(39) Grubbs, F. E. Sample criteria for testing outlying observations. Ann. Math. Stat. 1950, 21, 27−58. (40) Rushneck, D. R.; Beliveau, A.; Fowler, B.; Hamilton, C.; Hoover, D.; Kaye, K.; Berg, M.; Smith, T.; Telliard, W. A.; Roman, H.; Ruder, E.; Ryan, L. Concentrations of dioxin-like PCB congeners in unweathered Aroclors by HRGC/HRMS using EPA Method 1668A. Chemosphere 2004, 54, 79−87. (41) Stevens, J. Applied multivariate statistics for the social sciences; Erlbaum: Hillsdale, NJ, 1986. (42) Suter, G. W. Abuse of hypothesis testing statistics in ecological risk assessment. Hum. Ecol. Risk Assess. 1996, 2, 331−347. (43) StatPoint Technologies. Statgraphics Centurion XVI, version 16.1.17; StatPoint Technologies: Warrenton, VA, 2011. (44) Brown, J. F. Determination of PCB metabolic, excretion, and accumulation rates for use as indicators of biological response and relative risk. Environ. Sci. Technol. 1994, 28, 2295−2305. (45) Kostyniak, P. J.; Hansen, L. G.; Widholm, J. J.; Fitzpatrick, R. D.; Olson, J. R.; Helferich, J. L.; Kim, K. H.; Sable, H. J. K.; Seegal, R. F.; Pessah, I. N.; Schantz, S. L. Formulation and characterization of an experimental PCB mixture designed to mimic human exposure from contaminated fish. Toxicol. Sci. 2005, 88, 400−411. (46) Frame, G. M.; Cochran, J. W.; Boewadt, S. S. Complete PCB congener distributions for 17 Aroclor mixtures determined by 3 HRGC systems optimized for comprehensive, quantitative, congenerspecific analysis. J. High Resolut. Chromatogr. 1996, 19, 657−668. (47) Oliveira Ribeiro, C. A.; Vollaire, Y.; Coulet, E.; Roche, H. Bioaccumulation of polychlorinated biphenyls in the eel (Anguilla anguilla) at the Camargue Nature Reserve - France. Environ. Pollut. 2008, 153, 424−31. (48) Rodenburg, L. A.; Guo, J.; Du, S.; Cavallo, G. J. Evidence for Unique and Ubiquitous Environmental Sources of 3,3′-dichlorobiphenyl (PCB 11). Environ. Sci. Technol. 2010, 44, 2816−2821. (49) Hu, D.; Hornbuckle, K. C. Inadvertent Polychlorinated Biphenyls in Commercial Paint Pigments. Environ. Sci. Technol. 2010, 44, 2822−2827. (50) Guo, J.; Capozzi, S. L.; Kraeutler, T. M.; Rodenburg, L. A. Global Distribution and Local Impacts of Inadvertently Generated Polychlorinated Biphenyls in Pigments. Environ. Sci. Technol. 2014, 48, 8573−8580. (51) Zhu, Y.; Mapuskar, K. A.; Marek, R. F.; Xu, W.; Lehmler, H.-J.; Robertson, L. W.; Hornbuckle, K. C.; Spitz, D. R.; Aykin-Burns, N. A New Player in Environmentally Induced Oxidative Stress: Polychlorinated Biphenyl Congener, 3,3-Dichlorobiphenyl (PCB11). Toxicol. Sci. 2013, 136, 39−50. (52) Abramowicz, D. A. Aerobic and anaerobic biodegradation of PCBs: a review. Critical Reviews in Biotechnology 1990, 10, 241−251. (53) Seeger, M.; Zielinski, M.; Timmis, K. N.; Hofer, B. Regiospecificity of dioxygenation of di- to pentachlorobiphenyls and their degradation to chlorobenzoates by the bph-encoded catabolic pathway of Burkholderia sp strain LB400. Appl. Environ. Microbiol. 1999, 65, 3614−3621. (54) Commandeur, L. C. M.; May, R. J.; Mokross, H.; Bedard, D. L.; Reineke, W.; Govers, H. A. J.; Parsons, J. R. Aerobic degradation of polychlorinated biphenyls by Alcaligenes sp JB1: Metabolites and enzymes. Biodegradation 1997, 7, 435−443. (55) Froese, R.; Pauley, D., Eds. FishBase, version 10/20/13; 2013. http://www.fishbase.org/home.htm

I

DOI: 10.1021/es504961a Environ. Sci. Technol. XXXX, XXX, XXX−XXX