Research Rapid Response of Arctic Ringed Seals to Changes in Perfluoroalkyl Production CRAIG M. BUTT,† DEREK C. G. MUIR,‡ IAN STIRLING,§ MICHAEL KWAN,| AND S C O T T A . M A B U R Y * ,† Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario M5S 3H6, Canada, Environment Canada, Water Science and Technology Directorate, 867 Lakeshore Road, Burlington, Ontario L7R 4A6, Canada, Environment Canada, Canadian Wildlife Service, Edmonton, Alberta T6H 3S5, Canada, and Nunavik Research Centre, Kuujjuaq, Quebec J0M 1C0, Canada
Temporal trends in perfluoroalkyl compounds (PFCs) were investigated in liver samples from two ringed seal (Phoca hispida) populations in the Canadian Arctic, Arviat (Western Hudson Bay) (1992, 1998, 2004, 2005) and Resolute Bay (Lancaster Sound) (1972, 1993, 2000, 2004, 2005). PFCs analyzed included C7-C15 perfluorinated carboxylates (PFCAs) and their suspected precursors, the 8:2 and 10:2 fluorotelomer saturated and unsaturated carboxylates (FTCAs, FTUCAs), C4, C6, C8, C10 sulfonates, and perfluorooctane sulfonamide (PFOSA). Liver samples were homogenized, liquid-liquid extracted with methyl tert-butyl ether, cleaned up using hexafluoropropanol, and analyzed by liquid chromatography with negative electrospray tandem mass spectrometry (LC-MS/MS). C9-C15 PFCAs showed statistically significant increasing concentrations during 19922005 and during 1993-2005 at Arviat and Resolute Bay, respectively. Doubling times ranged from 19.4 to 15.8 years for perfluorododecanoate (PFDoA) to 10.0-7.7 years for perfluorononanoate (PFNA) at Arviat and Resolute Bay but were shorter when excluding the 2005 samples. Conversely, perfluorooctane sulfonate (PFOS) and PFOSA concentrations showed maximum concentrations during 1998 and 2000 at Arviat and Resolute Bay, with statistically significant decreases from 2000 to 2005. In the case of Arviat, two consecutive decreases were measured from 1998 to 2003 and from 2003 to 2005. PFOS disappearance half-lives for seals at Arviat and Resolute Bay were 3.2 and 4.6 years. These results indicate that the ringed seals and their food web are rapidly responding to the phase out of perfluorooctane sulfonyl fluoride based compounds by 3M in 2001. Further, the relatively short doubling times of the PFCAs and PFOS disappearance half-lives support the hypothesis of atmospheric transport as the main transport mechanism of PFCs to the arctic environment.
* Corresponding author phone: (416)978-1780; fax: (416)978-3596; e-mail:
[email protected]. † University of Toronto. ‡ Water Science and Technology Directorate. § Canadian Wildlife Service. | Nunavik Research Centre. 42
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 1, 2007
Introduction Perfluoroalkyl compounds (PFCs) have recently garnered intense scientific and regulatory interest because of their widespread detection in fish and wildlife (1, 2) and in humans (3, 4) and because of their potential toxicological effects (5). Two such classes of PFCs include the perfluorinated carboxylic acids (PFCAs) and the perfluorinated sulfonic acids (PFSAs). In 2004, Environment Canada issued a temporary 2-year ban on four fluorotelomer polymers used as stain repellents (6). Further, on January 26, 2006, the United States Environmental Protection Agency (U.S. EPA) launched a global stewardship program to reduce perfluorooctanoate (PFOA) emissions and its presence in products by 95% by 2010, aiming for complete elimination by 2015 (7). The detection of perfluorinated carboxylates and sulfonates in remote regions, such as the Arctic (8, 9), has raised questions as to their transport mechanism. One possible mechanism is through the atmospheric transport of volatile precursors that degrade to PFCAs and PFSAs. Fluorotelomer alcohols (FTOHs) are one candidate and have been shown to degrade via atmospheric oxidation to PFCAs (10, 11). The atmospheric lifetime of FTOHs with respect to hydroxyl radical reaction is ∼20 days and is sufficiently long to permit transport to the Arctic (12). FTOHs have been manufactured since the early 1970s and are used in the synthesis of many fluorosurfactants and fluorinated polymers (13). Similarly, perfluorinated sulfonamido alcohols have been shown to degrade to both PFCAs and PFSAs via atmospheric oxidation, in a mechanism analogous to PFCA formation from FTOHs (14, 15). Both FTOHs and sulfonamido alcohols have been widely observed in the North American troposphere (16, 17). An alternative hypothesis has suggested that the PFCAs may also reach the Arctic through oceanic transport (13). The detection of only linear isomers of PFOA and >99% linear isomers of perfluorononanoate (PFNA) and perfluorotridecanoate (PFTrA) in polar bears from the Canadian Arctic strongly suggests FTOHs are the source of PFCAs to the arctic environment (18), via atmospheric transport. Further, Smithwick et al. (19) have shown that the doubling times of PFCs observed in polar bears are too short to support the oceanic transport pathway. The major manufacturer of perfluorinated sulfonamides, 3M, voluntarily ceased production of perfluorooctane sulfonyl fluoride (PFOSF) based compounds in 2001 (20). PFOSFbased products have been produced since the 1950s and reached maximum production in 2000 (13, 19). In contrast, fluorotelomer-based products continue to be produced by several manufacturers, and production volumes have increased ∼2-fold to 11-14 × 106 kg yr-1 from 2000 to 2004 (21). Studies investigating temporal trends of PFCs, particularly in the arctic environment, are limited. Ringed seals from two locations in Greenland were found to have increasing concentrations of perfluorooctane sulfonate (PFOS) (annual increase ) 4.7-8.2%), perfluorodecanoate (PFDA) (1.73.3%), and perfluoroundecanoate (PFUnA) (5.9-6.8%) between 1982 and 2003 (22). Recently, Smithwick et al. (19) reported significant increases of PFOS, PFNA, PFDA, and PFUnA in polar bears from the eastern and western Canadian Arctic regions. Further, it was shown that the PFOS doubling time (9.8-13.1 years) in the polar bears was similar to the doubling time of PFOSF production (21). However, presently there are no published reports of temporal trends in arctic biota that extend beyond 2003, potentially missing the 10.1021/es061267m CCC: $37.00
2007 American Chemical Society Published on Web 11/30/2006
influence of the PFOS-related chemistry phase out. Further, these studies are limited by the relatively large interval between sampling times during the period 1995-2005. This paper presents temporal trends of archived ringed seal livers collected between 1972 and 2005 at two remote locations in the Canadian Arctic: Resolute Bay and Arviat, Nunavut. Seal liver samples were analyzed for C7-C15 PFCAs, the suspected PFCA precursors 8:2 and 10:2 saturated and unsaturated fluorotelomer acids, and C4, C6, C8, and C10 PFSAs. In this paper, we show, for the first time, recent significant decreases of PFOS in biological tissues.
Materials and Methods Standards and Reagents. Standards and reagents used were identical to those previously reported by our laboratory (8, 14). 1,1,1,3,3,3-Hexafluoro-2-propanol (HFP, 99+%) was purchased from Sigma-Aldrich (Oakville, ON, Canada). Stable isotope standards (13C4-PFOA, 13C5-PFNA, 13C2PFDA, 13C4-PFOS, 8:2 13C2-FTUCA (fluorotelomer unsaturated carboxylates), 10:2 13C2-FTUCA) were provided by Wellington Laboratories (Guelph, Ontario). 13C2-PFOA was purchased from Perkin-Elmer Life and Analytical Sciences Canada. Sample Collection. Ringed seal liver samples were collected by local subsistence hunters and trappers during annual hunts from two locations in the Canadian Arctic: Resolute Bay, Nunavut (74°42′ N, 94°49′ W) in 2000 (n ) 9), 2004 (n ) 9), and 2005 (n ) 10) and Arviat, Nunavut (61°7′ N, 94°4′ W) in 1998 (n ) 10), 2003 (n ) 10), and 2005 (n ) 10). All individuals were collected during the annual spring hunt, from approximately May to June. Samples collected in 2000-2005 were shipped to the Nunavik Research Centre (Kuujjuaq) for subsampling and tooth aging. Archived livers tissues from Resolute Bay 1972 (n ) 2) and 1993 (n ) 9) and from Arviat 1992 (n ) 6) were generously provided by the National Wildlife Research Center, Canadian Wildlife Service Specimen Bank, Ottawa, ON, Canada. Liver samples were shipped to the Canada Center for Inland Waters (Burlington, ON, Canada) and were stored at -20 °C. Ages were determined by longitudinal thin sectioning a lower canine tooth and counting the annual growth layers in the dentine using transmitted light (23). Extraction and Cleanup Methods. Liver samples were liquid-liquid extracted using methods similar to Hansen et al. (3) with the addition of a novel cleanup step (24). The purpose of the fluorosolvent cleanup step was to reduce matrix effects resulting from co-extractable substances, thereby increasing data quality. In addition, greater analyte recoveries were found when incorporating the cleanup step as compared to the traditional methods (24). Subsamples (∼0.7-1.0 g) were taken from the interior of whole or partial livers. Liver samples were homogenized in 15-mL plastic centrifuge tubes containing 4 mL of 0.25 M sodium carbonate and 1 mL of 0.5 M tetrabutylammonium hydrogen sulfate (pH 10) using a mechanical tissue homogenizer. Liver homogenates were extracted by vigorously shaking with 5 mL of methyl tert-butyl ether (MTBE) for 5 min and then were centrifuged for 10 min (3300 rpm). The organic layer was transferred to a clean centrifuge tube; liver homogenate was extracted again, was centrifuged, and organic layers were combined. The combined MTBE extract was reduced to 0.5 mL, an equal volume of hexafluoropropanol (HFP) was added, and the solution was shaken for 1 min and was centrifuged for 10 min. The precipitated proteins and lipids were separated from the MTBE/HFP solvent by filtering with a 0.2-µm nylon syringe filter. The MTBE/HFP solvent mixture was blown to dryness under nitrogen, reconstituted with 500 µL of methanol, vortexed for 30 s, and filtered. Internal standards were added immediately prior to analysis. Instrumental Analysis. Instrumental analysis was performed by liquid chromatography with negative electrospray
tandem mass spectrometry (LC-MS/MS). Chromatography was performed using an ACE C18 column (50 mm × 2.1 mm, 3-µm particle size, Aberdeen, United Kingdom), preceded by a C18 guard column (4.0 × 2.0 mm, Phenomenex). Initial conditions were 50:50 methanol:water (10 µM ammonium acetate) increasing to 95:5 at 7 min, decreasing to 5:95 at 7.25 min, and held to 9.0 min, reverting to initial conditions over 1 min and equilibrating for 4 min. Samples were analyzed using two different methods (methods A and B), differentiated by the LC-MS/MS instrumentation and internal standard suite used. In method A, samples were analyzed using a Micromass Ultima (Micromass, Manchester, United Kingdom) with the mobile phases delivered at 300 µL/min using a Waters 600S controller and the samples injected (20 µL) with a Waters 717 plus autosampler (Waters, Milford, United States). In method B, analytes were detected using an API 4000 Q Trap (Applied Biosystems/MDS Sciex) with samples injected with an Agilent 1100 autosampler (injection volume ) 10 µL, flow rate ) 300 µL/min). The column oven was set to 30 °C. In both methods, data were acquired under multiple reaction monitoring (MRM) conditions using optimized conditions (MRM transitions presented in Supporting Information). Analyte responses were normalized to internal standard responses. Details about internal standards used, analyte transitions, and instrumental conditions are presented in the Supporting Information. Concentrations were not corrected for recovery. The initial set of samples received were run with method A (Resolute Bay 1993, 2000, and 2004), whereas the latter set (Resolute Bay 1972 and 2005 and all Arviat samples) was run with method B. To compare the two methods, a subset of the extracts analyzed in method A (n ) 11, including two to three samples from each year) were analyzed using method B. Concentrations obtained using the two methods were not equal, but the regressions were significant (r2 > 0.85, p < 0.05). For example, method A concentrations were 1.5-fold greater than method B for PFNA, whereas, PFUnA concentrations were approximately equal for the two methods. Therefore, the remainder of concentrations obtained using method A was corrected to the method B results by applying a correction factor. Results are shown in the Supporting Information. Statistical Analyses and Data Treatment. The instrument detection limits (IDL) were defined as the concentrations that produced a peak with a signal-to-noise ratio of 3. Method detection limits (MDL) were defined as the mean blank response (procedural blank, 1 mL of 18 MΩ water carried through extraction procedure) plus 3 times the standard deviation of the blank response. For calculating the MDL, not detected (ND) values were replaced by one-half of the IDL. Concentrations that were below the method detection limit (MDL) were reported as PFOA, PFUnA > PFDA). Similar odd > even PFCA patterns have been observed in other arctic ringed seal samples (8, 22) as well as polar bears (9), and such PFCA profiles support the hypothesis of FTOH degradation as the source of PFCAs. For example, the atmospheric degradation of 8:2 FTOH has been shown to yield similar quantities of 44
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 1, 2007
PFOA and PFNA (10). However, longer-chain PFCAs have been shown to be more bioaccumulative (25, 26) (i.e., PFNA > PFOA), resulting in higher biota concentrations of PFNA relative to PFOA. A similar mechanism is expected for the production of PFUnA and PFDA from 10:2 FTOH. Perfluoroheptanoate (PFHpA), perfluorohexane sulfonate (PFBS), perfluorodecane sulfonate (PFHxS), and PFDS were not detected in any samples. Further, PFPA was not detected in the Resolute Bay 1972-2004 samples. PFOA concentrations were low,
