UNDERSTANDING SOURCES OF PERFLUORINATED ACIDS TO BIOLOGICAL SYSTEMS

by

Craig Michael Butt

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Chemistry University of Toronto

© Copyright by Craig M. Butt, 2010

Understanding Sources of Perfluorinated Acids to Biological Systems Doctor of Philosophy Degree, 2010 Craig M. Butt Department of Chemistry, University of Toronto

ABSTRACT

The overall aim of this thesis was to investigate the fate of perfluorinated alkyl compounds (PFCs) in biological systems. During the past several years, it has been shown that wildlife are ubiquitously contaminated with two classes of PFCs, the perfluoroalkyl carboxylates (CxF2x+1C(O)OH, PFCAs) and sulfonates (CxF2x+1SO3H, PFSAs). However, there is still considerable uncertainty regarding how wildlife are accumulating these PFCs, particularly in remote areas such as the Canadian arctic.

The potential for fluorotelomer acrylate monomers (CxF2x+1CH2CH2OC(O)CH=CH2, FTAcs) to act as precursors to PFCAs through atmospheric oxidation was investigated using smog chamber experiments. FTAc atmospheric fate is determined by OH radical reaction with a lifetime of approximately 1 day. The sole primary product of this reaction was the 4:2 fluorotelomer glyoxylate which is expected to undergo further atmospheric oxidation or photolysis to ultimately yield PFCAs. Temporal and spatial trends of PFCs in arctic ringed seals and seabirds were investigated to assist in understanding PFC transport mechanisms to remote regions. In ringed seals, perfluorooctane sulfonate (PFOS) levels decreased rapidly, coinciding with the phase out by the major manufacturer. These findings are consistent with volatile precursors as the dominant source of PFCs to arctic wildlife. The bioaccumulation and biotransformation of the 8:2 FTAc was investigated in two complimentary studies with rainbow trout. During the in vivo dietary exposure study, fish rapidly accumulated and biotransformed the 8:2 FTAc, with intermediate metabolites observed within 1 hour of dosing. Perfluorooctanoate (PFOA), perfluorononanoate (PFNA) and perfluoroheptanoate (PFHpA) were formed and accumulated in low yields. The carboxylesterase activity in the trout liver and stomach was investigated using in vivo sub- cellular (S9) incubations. Very high esterase activities were shown with approximately equal efficiency in the stomach and liver.

ii The metabolic pathway of the 8:2 fluorotelomer alcohol (8:2 FTOH) was investigated by separately dosing whole rainbow trout with three intermediate metabolites that represented important branching points. The 7:3 fluorotelomer saturated carboxylate (FTCA) did not form PFOA, but formed PFHpA and the 7:3 fluorotelomer unsaturated carboxylate (FTUCA). The 8:2 FTCA and 8:2 FTUCA did form PFOA, confirming a “beta-like-oxidation” mechanism.

iii ACKNOWLEDGEMENTS

A few moments for the author to express his humility.

I have had the fortune of being mentored by two inspiring supervisors, Scott Mabury and Derek Muir. They have served as excellent role models, each with their own strengths and bias, guiding me through my progression as a graduate student. The research opportunities have been unparalleled. Derek – I am grateful for your supportive nature, extremely quick turnaround on manuscripts and unequaled accessibility. Scott – thank you for pushing my development beyond an “X and Y” chemist, I will miss our late night and weekend chats.

Thank you also to Frank Wania for serving as the chair of my PhD committee and to Jamie Donaldson for acting as the final committee member.

I have also been able to collaborate with several excellent scientists during my tenure. Rosanna Bossi (University of Aarhus, Denmark), Urs Berger (Stockholm University, Sweden)) and Gregg Tomy (Department of Fisheries and Oceans) were co-authors for the arctic review paper. Tim Wallington and Mike Hurley at the Ford Motor Company provided project guidance, technical assistance and expertise during the 4:2 acrylate atmospheric study. Birgit Braune at the National Wildlife Research Centre (Environment Canada) was the project investigator for the temporal seabird study.

Thank you to all past and present members of the Mabury group – Dave, Jon, Xinghua, Suzanne, Jules, Monica, Naomi, Uli, Joyce, Erin, Sarah, Anne, Pablo, Amy (my swim coach), Holly and Derek (“We know what we did”). Whether by luck, or by vision, I am grateful to have worked with an incredibly talented and inspiring group. In particular I have enjoyed our non-academic times together – if this group wasn’t so fun, I would have graduated years ago! I have learned something from each one of you, including those who have taught me the virtue of patience. My hope is that you also learned something from me.

I have also been fortunate to mentor several undergraduate students throughout my tenure as a PhD student. Many thanks to Yan Li, Helen Sun, Clara Chan, Rodolfo Gomez, Alex Tevlin and Rob Di Lorenzo for assistance with sample preparation.

iv Many thanks also to the team at Canadian Centre for Inland Waters (Environment Canada) under the supervisor of Derek Muir. Xiaowa Wang provided assistance in the management of the arctic biota samples. Christine Spencer ensured that the LC-MS/MS was in good working order and free from contamination. Jeff Small assisted with the operation of the LC-MS/MS and, along with Colin Darling, provided much appreciated friendship.

I could not have completed this degree without the friendship of four truly amazing people, in no particular order – Amila De Silva, Jessica D’eon, Zamin Kanji and Cora Young. You’ve been wonderful companions on this crazy ride. Your contributions are beyond words. We’ve travelled together, competed together, laughed together, drank together and vented our frustrations together. Your (planned) interventions were appreciated. You opened your doors to me when I needed it the most. How is it that you know me better than I know myself? Of all the stories I tell, the most cherished are those that involve you.

To all the individuals I have known outside of the academic “bubble”, you have helped me achieve balance in life. You have managed to keep me sane, and for some, coincidently driven me insane. Rock climbing, hockey, running, friendship and patio beers. Thank you for nodding your head, smiling and acting interested when I explained my research. In particular, thank you to Jeff Rands of VWR Canada – my stomach has greatly benefited from your expense account.

To Sparky the Cat (aka – Sparkles MacPuss), I am thankful for your big meows and furry kisses.

I am grateful to my family for their everlasting love and support throughout my many years as a graduate student. Thank you for your constant encouragement and allowing me to trust my instincts. After years of asking, “When will you be finished?” I can confidently and definitively answer, “Now”.

v TABLE OF CONTENTS

Chapter One – Introduction to Poly- and Perfluorinated Compounds 1

1.1. Overview 2 1.2. Perfluorooctane Sulfonyl Fluoride-Based Compounds 4 1.3. Fluorotelomer-Based Compounds 6 1.4. Sources of Poly- and Perfluorinated Acids to the Environment 8 1.4.1. Direct Sources 8 1.4.2. Indirect Sources 8 1.4.2.1 Atmospheric Reactions 8 1.4.2.2 Polyfluorinated Compound Biotransformation 11 1.4.2.2.1 Fluorotelomer-based Compounds 11 1.4.2.2.1.1. Microbial 11 1.4.2.2.1.2. Rats and Mice 15 1.4.2.2.1.3. Biotransformation of Fluorotelomer-based 20 Compounds: Summary and Conclusions 1.4.2.2.2. Polyfluorinated Sulfonamide Compounds 23 1.4.2.2.2.1. N-ethyl perfluorooctane sulfonamide (N- 23 EtFOSA) 1.4.2.2.2.2. N-ethyl perfluorooctane sulfonamide ethanol 24 (N-EtFOSE) 1.4.2.2.2.3. Biotransformation of Polyfluorinated 25 Sulfonamide Compounds: Summary and Conclusions 1.5. Bioaccumulation of organic compounds 27 1.5.1. Definitions 27 1.5.2.Contaminant Uptake and Elimination in Fish 28 1.5.3. Xenobiotic Metabolism in Fish 29 1.5.3.1. Extrahepatic Metabolism in Fish 30 1.5.4. Quantifying Metabolism in Fish 30 1.5.5. Quantifying Bioaccumulation 32 1.5.5.1. Laboratory-based Experiments 32 1.5.5.2. Models to Predict Bioaccumulation 32 1.5.5.2.1. Empirical Correlation Models 32 1.5.5.2.2. Mechanistic Models 33 1.5.5.2.2.1. Extrapolating In Vitro Metabolic Data to 36 Predict In Vivo Bioaccumulation 1.6. Literature Cited 38

Chapter Two – Levels and Trends of Poly- and Perfluorinated Compounds in 45 the Arctic Environment

Butt, C.M.; Berger, U.; Bossi, R.; Tomy, G.T. Submitted to The Science of the Total Environment

2.1. Abstract 48 2.2. Introduction 50 2.3. Transport Pathways 51 2.4. Biotic Measurements 59 2.4.1. Trends 59 2.4.1.1. Food Web Studies 59 2.4.1.2 Spatial Studies 64 2.4.1.3 Temporal Trends 72 vi 2.4.1.3.1. North American Arctic 72 2.4.1.3.2. Greenland 80 2.4.1.3.3. Norway 84 2.4.2. PFC profiles 85 2.4.3. Animal Body Burdens 87 2.4.4. “Neutrals” and Precursors 88 2.5. Abiotic Measurements 90 2.5.1. Atmospheric Measurements 90 2.5.2. Snow 92 2.5.2.1. Canadian Arctic 92 2.5.2.2. Greenland 93 2.5.3. Lake Water & Sediments 93 2.5.3.1. Amituk, Char & Resolute lakes on Cornwallis Island, Canadian 93 Arctic 2.5.3.2. Isomers in Char Lake sediments; Surface Water from Char Lake 95 and Amituk Lake 2.5.4. Seawater and Marine Sediments 96 2.5.4.1. Greenland Sea 96 2.5.4.2. Labrador Sea 96 2.5.4.3. Canadian Arctic 97 2.5.4.4. Iceland & Faroe Islands 97 2.5.4.5. Russian Arctic 98 2.5.5. Sewage sludge & effluent 98 2.6. Conclusions and Research Needs 100 2.7. Literature Cited 105

Chapter Three – Thesis Goals and Hypothesis 113

Chapter Four - Atmospheric Chemistry of 4:2 Fluorotelomer Acrylate 117 (C4F9CH2CH2OC(O)CH=CH2): Kinetics, Mechanisms and Products of Chlorine Atom and OH Radical Initiated Oxidation

Butt, C.M.; Young, C.J.; Mabury, S.A.; Hurley, M.D.; Wallington, T.J. Journal of Physical Chemistry A 2009, 113, 315-3161

4.1. Abstract 118 4.2. Introduction 119 4.3. Experimental Section 120 4.3.1. Kinetic Experiments 120 4.3.2. Product and Mechanistic Experiments 121 4.4. Results and Discussion 122 4.4.1. Kinetics of the Cl + 4:2 FTAc reaction 122 4.4.2. Kinetics of the OH + 4:2 FTAc reaction 123 4.4.3. Products and Mechanism of the Cl + 4:2 FTAc reaction in the presence and 127 absence of NOx 4.4.4. Products and Mechanism of the OH + 4:2 FTAc reaction in the presence of 132 NOx 4.5. Implications for Atmospheric Chemistry 136 4.5.1. Atmospheric lifetime 136 4.5.2. Contribution to PFCA burden in remote locations 137 4.6. Acknowledgement 138 4.7. Literature Cited 139

vii Chapter Five - Rapid Response of Arctic Ringed Seals to Changes in 143 Perfluoroalkyl Production

Butt, C.M.; Muir, D.C.G.; Stirling, I.; Kwan, M.; Mabury, S.A. Environmental Science & Technology 2007, 41, 42-49

5.1. Abstract 144 5.2. Introduction 145 5.3. Materials and Methods 146 5.3.1. Standards and Reagents 146 5.3.2. Sample Collection 147 5.3.3. Extraction and Clean-up Methods 147 5.3.4. Instrumental Analysis 148 5.3.5. Statistical Analyses and Data Treatment 149 5.3.6. Quality Control and Quality Assurance 149 5.4. Results and Discussion 151 5.4.1. Contaminant Profiles and Concentrations 151 5.4.2. Age and Sex Trends 152 5.4.3. Temporal Trends 152 5.5. Acknowledgements 157 5.6. Literature Cited 161 5.7 Supporting Information 165

Chapter Six - Prevalence of long-chained perfluorinated carboxylates in 175 seabirds from the Canadian Arctic between 1975 and 2004

Butt, C.M.; Mabury, S.A.; Muir, D.C.G.; Braune, B.M. Environmental Science & Technology 2007, 41, 3521-3528

6.1. Abstract 176 6.2 Introduction 177 6.3. Materials and Methods 179 6.3.1. Standards and Chemicals 179 6.3.2. Sample Collection 179 6.3.3. Sample Extraction and Instrumental Analysis 179 6.3.4. Statistical Analysis and Data Treatment 180 6.3.5. Quality Control and Quality Assurance 181 6.4. Results and Discussion 181 6.4.1. Overall Concentrations and Contaminant Profiles 181 6.4.2. PFC-Sex Trends 185 6.4.3. Temporal Trends in Thick-Billed Murres and Northern Fulmars 185 6.5. Acknowledgements 191 6.6. Literature Cited 192

Chapter Seven – Spatial Trends of Perfluoroalkyl Compounds in Ringed Seals 199 (Phoca hispida) from the Canadian Arctic

Butt, C.M.; Mabury, S.A.; Kwan, M.; Wang, X.; Muir, D.C.G. Environmental Toxicology & Chemistry 2008, 27, 542-553

7.1 Abstract 200 7.2. Introduction 201 viii 7.3. Materials and Methods 203 7.3.1. Standards and chemicals 203 7.3.2. Sample collection 203 7.3.3. Sample extraction and instrumental analysis 203 7.3.4. Stable isotope analysis 204 7.3.5. Statistical analysis and data treatment 205 7.3.6. Quality control and quality assurance 206 7.4. Results and Discussion 207 7.4.1. Influence of age and sex on PFCs 207 7.4.2 General PFC concentrations in ringed seal 207 7.4.3. Spatial trends in concentration and PFC profiles 212 7.4.4. Stable isotopes of nitrogen and carbon 217 7.4.5. Relationships between stable isotope ratios and PFC concentrations 222 7.4.6. Ringed seal - polar bear biomagnification factors 227 7.5. Conclusion 229 7.6. Acknowledgements 229 7.7. Literature Cited 230

Chapter Eight – Bioaccumulation and Biotransformation of the 8:2 FTOH 237 Acrylate in Rainbow Trout

Butt, C.M.; Muir, D.C.G.; Mabury, S.A. In preparation for Environmental Science & Technology

8.1. Abstract 238 8.2. Introduction 239 8.3. Materials and Methods 241 8.3.1. Standards and Reagents 241 8.3.2. In vivo 8:2 FTAc Dietary Exposure 242 8.3.2.1. Food Preparation 242 8.3.2.2. Fish Care and Sampling 242 8.3.2.3. Extraction and Instrumental Analysis 243 8.3.2.4. Analytical Methods 244 8.3.2.5. Statistical Analysis and Data Treatment 245 8.3.3. In vitro S9 incubation experiments 246 8.3.3.1. S9 fraction preparation 246 8.3.3.2. Para-nitrophenyl Acetate (PNPA) Assay 246 8.3.3.3. 8:2 FTAc Incubations 246 8.3.3.4. Data Analysis 247 8.4. Results and Discussion 248 8.4.1. In Vivo 8:2 FTAc Dietary Exposure 248 8.4.1.1. Physical Indices 248 8.4.1.2. 8:2 FTAc and 8:2 FTOH 248 8.4.1.3. Intermediate and Terminal Metabolites 251 8.4.1.3.1. Uptake Phase – Liver, kidney and blood 251 8.4.1.3.2. Elimination Phase 253 8.4.1.3.3. Bile and Feces 258 8.4.2. In Vitro S9 Incubation Experiments 259 8.4.2.1. Para-nitrophenyl Acetate (PNPA) Assay 259 8.4.2.2. 8:2 FTAc Incubations 261 8.4.3. Mechanism of Biotransformation 264 8.4.4. Environmental Implications 265 8.5. Acknowledgements 267

ix 8.6. Literature Cited 268 8.7. Supporting Information 275

Chapter Nine – Elucidating the Mechanism of Poly- and Perfluorinated Acid 295 Production in Rainbow Trout

Butt, C.M.; Muir, D.C.G.; Mabury, S.A. In preparation for Environmental Science & Technology

9.1 Abstract 296 9.2. Introduction 297 9.3. Materials and Methods 299 9.3.1. Standards and Reagents 299 9.3.2. Synthesis of 8:2 FTCA and 8:2 FTUCA 299 9.3.3. Food Preparation 299 9.3.4. Fish Care and Sampling 300 9.3.5. Extraction and Clean-up Methods 301 9.3.6. Instrumental Analysis 301 9.3.7. Statistical Analyses and Data Treatment 302 9.4. Results and Discussion 303 9.4.1. Physical Indices 303 9.4.2. 7:3 FTCA Exposure 304 9.4.3. 8:2 FTCA Exposure 306 9.4.4. 8:2 FTUCA Exposure 308 9.4.5. Elimination Half-lives: Comparison of Intermediate and Terminal Metabolites 311 9.4.6. Environmental Implications 313 9.4.6.1. Formation of 7:3 FTCA 313 9.4.6.2. Formation of PFNA and PFOA 314 9.5. Acknowledgements 317 9.6. Literature Cited 319 9.7. Supporting Information 323

Chapter Ten – Conclusions and Future Directions 327

x LIST OF TABLES

CHAPTER ONE

Table 1.1 List of acronyms, common names and chemical structure 3

CHAPTER TWO

Table 2.1. Correlation coefficient (r2) of linear regression between adjacent 68 chain length PFCAs in polar bear liver from the North American and European Arctic.

CHAPTER FIVE

Table 5.1. Doubling times and disappearance half-lives (years ± 95% 158 confidence interval) for Arviat (1992-2005) and Resolute Bay (1972-2005) ringed seals. Doubling times and disappearance half- lives excluding 2005 samples are shown in parenthesis. Table S5.1. Multiple reaction monitoring (MRM) transitions of target analytes. 166

Table S5.2. Spike and recovery data for individual analytes spiked into ~1g of 169 ringed seal liver (n=10). Note: recoveries for PFTrA and PFPA could not be evaluated since commercial standards are not available.

Table S5.3. Comparison of Resolute Bay 2000 ringed seal liver samples: 170 Homogenates versus Sub-samples. “Homogenates” refers to ~0.5 g sample of homogenized liver (original mass ~6-8g). “Sub- sample” refers to ~0.5 g sample of unhomogenized liver. Table S5.4. Mean (minimum-maximum) concentration (ng/g ww) of 174 perfluorinated acids in ringed seals from Arviat and Resolute Bay. “nd” signifies analyte not detected, “na” signifies analyte not analyzed, “nq” signifies analyte detected but not quantified.

CHAPTER SIX

Table 6.1. Doubling times (years ± 95% confidence interval), geometric 190 mean (range) concentration (ng/g ww) in livers of thick-billed murres (1975-2004) and northern fulmars (1975-2003) from Prince Leopold Island, Nunavut. Table S6.1. Geometric mean concentrations and ranges (ng/g ww) of 198 perfluorinated acids in thick-billed murres and northern fulmars from Prince Leopold Island, Nunavut.

CHAPTER SEVEN

Table 7.1. Geometric mean concentrations (ng/g wet wt) and ranges for 209 perfluorinated carboxylates and sulfonates in ringed seals from the Canadian Arctic. Table 7.2. Adjusted mean concentrations (ng/g wet wt) of selected 221 perfluorinated acids and mean carbon and nitrogen stable isotope ratios (±95% confidence intervals). Perfluoroalkyl compound

xi (PFC) concentrations adjusted to overall mean δ13C value of - 19.2‰ using the analysis of covariance model, [lnPFC] = δ13C + site + δ13C⋅site. Adjusted means not presented for Sachs Harbour and Arctic Bay since these populations were excluded from the analysis of covariance model. Table 7.3. Geometric mean for regionally-based ringed seal-polar bear 228 biomagnification factors for perfluorinated carboxylates and sulfonates.

CHAPTER EIGHT Table 8.1. First-order elimination half-lives (d) for liver, kidney and blood 257 from 8:2 FTAc dietary exposure. Table 8.2. Michaelis constant (KM) and maximum reaction rate (Vmax) 261 parameters in fish using PNPA as the substrate for carboxylesterase activity. Table S8.1. Common name, acronym and structure for analytes monitored. 275 Table S8.2. “Active” 8:2 FTAc incubations with liver S9 fractions. 283 Table S8.3. Heat-inactivated control 8:2 FTAc incubations with liver S9 287 fractions. Table S8.4. “Active” 8:2 FTAc Incubations with stomach S9 fractions. 289 Table S8.5 Heat-inactivated control 8:2 FTAc incubations with stomach S9 293 fractions.

CHAPTER NINE Table 9.1. Liver Somatic Index (%) of control and dosed treatments at 12 and 303 120 hours of the uptake phase and 72 and 240 hours of the elimination phase. Table 9.2. Elimination half-lives (days) in blood and liver for polyfluorinated 313 and perfluorinated carboxylates in rainbow trout. Table S9.1. Mean ± standard error (n=3) concentrations (ng/g ww) in blood 323 and liver from 7:3 FTCA dietary exposure in rainbow trout. Table S9.2. Mean ± standard error (n=3) concentrations (ng/g ww) in blood 324 and liver from 8:2 FTCA dietary exposure in rainbow trout. Table S9.3. Mean ± standard error (n=3) concentrations (ng/g ww) in blood 325 and liver from 8:2 FTUCA dietary exposure in rainbow trout.

xii LIST OF FIGURES

CHAPTER ONE

Figure 1.1. Estimated global POSF production volumes (1970-2002). 5 Figure 1.2. Scheme of fluorotelomer-based chemical production. 7

Figure 1.3. Atmospheric reaction scheme for the production of PFCAs (boxes with 10 dashed outlines) from fluorotelomer-based compounds (boxes with solid outlines). Figure 1.4. 8:2 FTOH biodegradation in mixed microbial system as proposed by 21 Dinglasan et al. Structures in brackets were not determined in the study. Figure 1.5. 8:2 FTOH biodegradation in rats as proposed by Fasano et al. 22 Structures labeled in parenthesis are hypothesized intermediates and were not observed in the experiment. Note that glutathione conjugates can undergo further metabolism not presented in this figure. Figure 1.6. Proposed biotransformation pathways of N-EtFOSE in rat liver 26 microsomes, cytosol and slices. Figure 1.7. Conceptual diagram showing major routes of contaminant uptake and 28 elimination in fish. Figure 1.8. Schematic representation of a physiologically based bioaccumulation 35 model for fish. Figure 1.9. The influence of liver metabolism on the steady-state bioaccumulation 37 as predicted with a one-compartment model. Simulations obtained using in vitro intrinsic clearance values of 0.0 (1), 0.1 (2), 1.0 (3), 10.0 (4) and 100 (5) µl/min.mg protein. The bioaccumulate factor is normalized to the whole-body lipid content.

CHAPTER TWO

Figure 2.1. Major transport pathways of PFCs to the Arctic. 52 Figure 2.2. Modeled PFOA concentrations in ocean water from northern 54 hemisphere for period 1950-2050. Vertical bars represent annual emissions, solid line represents model concentrations in the Northern Temperate zone, dotted line represents model concentrations in the Northern Polar zone (arctic region). Figure 2.3. Modeled PFOA deposition fluxes (solid lines and bands) to the Arctic 56 (65oN to 90oN) resulting from FTOH atmospheric degradation (red) and FOSE atmospheric degradation (blue). Crosses indicate results from other models (red) and fluxes extrapolated from surface snow measurements (black). Figure 2.4. Mean (±1 SE) PFOS concentrations (ng/g wet wt) – trophic level 60 relationship for the eastern Arctic food web. BLKI = black-legged kittiwakes; GLGU = glaucous gulls. Figure 2.5. Relationship between PFOS concentration (ng/g wet wt) and trophic 62 level, as quantified by δ15N for Barents Sea food web. In the figure legend “polar cod” is identified as “arctic cod” in the text. Arrow indicates one ice amphipod sample was below the range displayed in the figure.

xiii Figure 2.6. Geometric mean concentrations (ng/g ww) of PFCs in polar bear liver 67 from the North American and European Arctic. Error bars represent 95% confidence intervals. Figure 2.7. Geometric mean concentration (ng/g ww) of PFOA, PFNA, PFDA, 69 PFUnA and PFOS in ringed seal liver from Canadian Arctic. Error bars indicate one standard error. Figure 2.8. Geometric mean concentration (ng/g ww) of selected PFCs in ringed 70 seals from the Canadian Arctic. Error bars represent 95% confidence intervals. Figure 2.9. Geometric mean concentrations (ng/g ww) of PFOS, PFNA, PFDA and 74 PFUnA in ringed seals from Arviat, Nunavut, Canada (1992-2005). Error bars indicate 95% confidence interval. Figure 2.10. Geometric mean concentrations (ng/g ww) of PFOS, PFNA, PFDA and 75 PFUnA in ringed seals from Resolute Bay, Nunavut, Canada (1972- 2005). Error bars indicate 95% confidence interval. Figure 2.11. Geometric mean concentration (ng/g ww) of PFCAs and fluorotelomer 77 acids in (a) thick-billed murres and (b) northern fulmars from Prince Leopold Island, Nunavut, Canada. Errors bars indicate 95% confidence intervals. “*” indicates that all samples were below MDL or were not detected for that time point. Figure 2.12. PFOS temporal trends in polar bear livers from near northern Baffin 78 Island, Canada (east) and near Barrow, Alaska (west) between 1972 and 2002. Vertical bars indicate 95% confidence intervals. Figure 2.13. PFOA, PFNA, PFDA and PFUnA temporal trends in polar bear livers 79 from near northern Baffin Island, Canada (east) and near Barrow, Alaska (west) between 1972 and 2002. Vertical bars indicate 95% confidence intervals. Figure 2.14. Temporal trends in PFOS, PFDA and PFUnA in ringed seal liver from 81 Ittoqqortoormiit (East Greenland), 1986-2006, and Qeqertarsuaq (West Greenland), 1982-2006 (Riget, unpublished). Red circles represent median concentrations, red line represents significant log-linear regression, black line represents non-significant log-linear regression. Figure 2.15. Temporal of PFCs in East Greenland polar bear liver from 1984 to 83 2006. Filled points represent log-linear regression lines or LOESS smoother lines. Broken lines represent 95% confidence limits. Figure 2.16. Temporal trends of PFOS, PFDS, PFUnA and PFTrA in herring gull 84 eggs from the Hornøya and Røst, Norwegian Arctic. PFDcS and PFTriA are identified as PFDS and PFTrA, respectively, in the manuscript. Figure 2.17. Total air concentrations (sum of gas-phase and particle-phase) of 91 individual FTOHs and FOSEs from North Atlantic and Canadian Archipelago. Figure 2.18. PFC concentrations (ng/g dry weight) in sediment core slices from 95 Resolute, Char and Amituk Lake on Cornwallis Island, Nunavut, Canada.

CHAPTER FOUR

Figure 4.1. Loss of 4:2 FTAc versus C2H4 (●) and C3H6 (▲) following UV 125 irradiation of 4:2 FTAc/reference/Cl2 mixtures in 700 Torr of N2. Figure 4.2. Loss of 4:2 FTAc versus C2H4 (●) and C3H6 (▲) following UV 126 irradiation of 4:2 FTAc/reference/CH3ONO mixtures in 700 Torr of air. Figure 4.3. FTIR spectra obtained before (A) and after (B) 85 seconds UV 129

xiv irradiation of a mixture of 7.1 mTorr 4:2 FTAc and 105 mTorr Cl2 in 700 Torr of air diluent. Panel C shows the product spectrum obtained by subtracting the IR features of the reactant from the spectrum in B. Panel D is a reference spectrum of C4F9CH2C(O)H. Panel E is the product spectrum obtained by subtracting IR features of C4F9CH2C(O)H from the spectrum in C. Figure 4.4. Yield of C4F9CH2C(O)H (4:2 FTAL), unknown residual product (panel 130 “E” in figure 3) and HC(O)Cl versus depletion of 4:2 FTAc after irradiation of a mixture of 7.1 mTorr 4:2 FTAc and 105 mTorr Cl2 in 700 Torr of air diluent (identified as “without NO”) and 7.4 mTorr of 4:2 FTAc, 101 mTorr Cl2 and 57 mTorr NO in 700 Torr of air diluent (identified as “with NO”). Lines are linear regressions to the 4:2 FTAL data. Figure 4.5. Simplified mechanism of Cl atom oxidation of 4:2 FTAc. 131 Figure 4.6. Simplified mechanism of OH radical oxidation of 4:2 FTAc. 134 Figure 4.7. Formation of HCHO versus loss of 4:2 FTAc, normalized to the initial 135 4:2 FTAc concentration following the UV irradiation of 4:2 FTAc/CH2DONO/NO mixtures in 700 Torr of air at 296 K. The curve is a fit of equation I to the data.

CHAPTER FIVE

Figure 5.1. Geometric mean concentrations (ng/g ww) of C9-C11 PFCAs and 159 PFOS from 1992-2005 Arviat ringed seals. Error bars indicate 95% confidence interval. Figure 5.2. Geometric mean concentrations (ng/g ww) of C9-C11 PFCAs and 160 PFOS from 1972-2005 Resolute Bay ringed seals. Error bars indicate 95% confidence interval. Note: Error bars not calculated for 1972 samples due to small sample size. Figure S5.1. Ringed seal sample locations (sample years). 165 Figure S5.2. Comparison of concentrations (ng/g ww) using Method A & Method B. 167 Figure S5.3. Comparison of Resolute Bay 2000 ringed seal liver samples: 170 Homogenates versus Sub-samples. “Homogenates” refers to ~0.5 g sample of homogenized liver (original mass ~6-8g). “Sub-sample” refers to ~0.5 g sample of unhomogenized liver. Figure S5.4. Statistically significant age-concentration trends. 172

CHAPTER SIX

Figure 6.1. Percent composition of total PFCAs (C7-C15 PFCAs), total PFSAs 184 (PFHxS, PFOS, PFDS, PFOSA), total telomer acids (8:2 FTCA & FTUCA, 10:2 FTCA & FTUCA). For calculation of means, concentrations less than the MDL or were not detected were replaced by a random number less than half of the MDL. Figure 6.2. Geometric mean concentration (ng/g ww) of perfluorinated carboxylic 188 acids and fluorotelomer acids in a) thick-billed murres and b) northern fulmars from Prince Leopold Island, Nunavut. Error bars indicate 95% confidence intervals. For calculation of means, concentrations less than the MDL or were not detected were replaced by a random number less than half of the MDL. “*” indicates that all samples were below MDL or were not detected for that time point. Figure 6.3. Geometric mean concentration (ng/g ww) of perfluorinated sulfonates 189 and PFOSA in thick-billed murres and northern fulmars from Prince

xv Leopold Island. Error bars indicate 95% confidence intervals. For calculation of means, concentrations less than the MDL or were not detected were replaced by a random number less than half of the MDL. “*” indicates that all samples were below MDL or were not detected for that time point. Figure S6.1. Location of sample collection (Prince Leopold Island Migratory Bird 197 Sanctuary, Lancaster Sound, Nunavut, Canada).

CHAPTER SEVEN Figure 7.1. Geometric mean concentration (ng/g wet wt) of perfluorooctanoate 215 (PFOA), perfluorononanoate (PFNA), perfluorodecanoate (PFDA), perfluoroundecanoate (PFUnA), and perfluorooctane sulfonate (PFOS) in ringed seals from individual locations from the Canadian Arctic. Error bars indicate one standard error. Figure 7.2. Geometric mean concentration (ng/g wet wt) of selected perfluorinated 216 carboxylates and sulfonates in ringed seals from the Canadian Arctic. Individual populations have been grouped into four geographic regions. Error bars indicate 95% confidence intervals. Acronyms defined in footnote from Table 7.1. Figure 7.3. Mean stable isotope ratios of carbon and nitrogen for ringed seals from 220 nine locations in the Canadian Arctic. Error bars indicate 95% confidence intervals. GH=Gjoa Haven, SH=Sachs Harbour, RB=Resolute Bay, AV=Arviat, AB=Arctic Bay, QT=Qikiqtarjuaq, PI=Pond Inlet, LN=Nain. Stable isotope data not available for Pangnirtung and Inukjuak. Figure 7.4. Correlations between ln[perfluoroalkyl compound (PFC)] (ng/g wet wt) 226 and δ13C value (‰) for (a) perfluorononanoate (PFNA) and (b) perfluoroundecanoate (PFUnA) for various ringed seal populations in the Canadian Arctic. Lines represent regression equations for individual populations. Inukjuak and Pangnirtung populations excluded since stable isotope data was not available. Sachs Harbour and Arctic Bay populations excluded since these population had positive ln PFC- δ13C correlations. Symbol legend is applicable to both (a) and (b). z Arviat, S Grise Fiord, T Gjoa Haven, U Nain, „ Pond Inlet, Qikiqtarjuaq, ‘ Resolute Bay.

CHAPTER EIGHT

Figure 8.1. Mean 8:2 FTOH and 8:2 FTAc concentrations (ng/g ww) in feces. 251 Error bars indicate 1 standard error. Figure 8.2. Liver concentrations (ng/g ww) of 8:2 FTCA, 8:2 FTUCA, 7:3 FTCA, 256 PFOA and PFNA from 8:2 FTAc dietary exposure in rainbow trout. Figure 8.3. Michaelis-Mention plot of para-nitrophenyl acetate carboxylesterase 260 activity (nmol/min.mg protein) in rainbow trout liver and stomach S9 fractions. Michaelis constant (KM) and maximum reaction rate (Vmax) obtained from nonlinear regression analysis. Error bars represent one standard error. Figure 8.4. Michaelis-Menten plot of 8:2 FTAc biotransformation (nmol/min.mg 263 protein) in rainbow trout liver and stomach S9 fraction. Michaelis constant (KM) and maximum reaction rate (Vmax) obtained from nonlinear regression analysis. Error bars represent one standard error. Figure 8.5. Kidney concentration (pmol/g ww) of 8:2 FTAc and metabolites during 267 uptake phase of dietary exposure to 8:2 FTAc. “Metabolites” represents

xvi sum of 8:2 FTOH, 8:2 FTCA, 8:2 FTUCA, 7:3 FTCA, 7:3 FTUCA, PFHpA, PFOA and PFNA. Error bars indicate 1 standard error. Figure S8.1. Whole fish weight (g) during uptake and elimination phases for dosed 277 (□) and control (▲) treatments. Data points for the dosed treatment represent the arithmetic mean (n=3) and errror bars indicate 1 standard error. Figure S8.2. Liver somatic index (LSI, %)) during uptake and elimination phases for 278 dosed (□) and control (▲) treatments. Data points for the dosed treatment represent the arithmetic mean (n=3) and error bars indicate 1 standard error. Figure S8.3. Kidney concentrations (ng/g ww) of 8:2 FTCA, 8:2 FTUCA, 7:3 FTCA, 279 PFOA and PFNA from 8:2 FTAc dietary exposure in rainbow trout. Figure S8.4. Blood concentrations (ng/g ww) of 8:2 FTCA, 8:2 FTUCA, 7:3 FTCA, 280 PFOA and PFNA from 8:2 FTAc dietary exposure in rainbow trout. Figure S8.5. Bile concentrations (ng/g ww) of 8:2 FTCA, 8:2 FTUCA, 8:2 FTOH- 281 glucuronide, 7:3 FTCA, PFOA and PFNA from 8:2 FTAc dietary exposure in rainbow trout. Figure S8.6. Feces concentrations (ng/g ww) of 8:2 FTCA, 8:2 FTUCA, 8:2 FTOH- 282 glucuronide, 7:3 FTCA, PFOA and PFNA from 8:2 FTAc dietary exposure in rainbow trout.

CHAPTER NINE Figure 9.1. Mean blood and liver concentrations (ng/g ww) of 7:3 FTCA, PFHpA 305 and 7:3 FTUCA resulting from 7:3 FTCA dietary exposure. Error bars represent 1 standard error. Figure 9.2. Mean blood and liver concentrations (ng/g ww) of 8:2 FTCA, 7:3 308 FTCA, 8:2 FTUCA, PFOA and PFNA resulting from 8:2 FTCA dietary exposure. Error bars represent 1 standard error. Figure 9.3. Mean blood and liver concentrations (ng/g ww) of 8:2 FTUCA, 7:3 311 FTCA, 8:2 FTCA and PFOA resulting from 8:2 FTUCA dietary exposure. The y-axis scale is equivalent for all graphs. The 8:2 GSH is plotted in arbitrary units. Error bars represent 1 standard error. Figure 9.4. Proposed biotransformation mechanism for the 8:2 FTOH. 318

xvii PREFACE This thesis is comprised of a series of manuscripts that have been published in or are in preparation for submission to peer-reviewed scientific journals. Therefore, repetition of introductory material and experimental details was unavoidable. Chapters One and Two together comprise the introduction to this thesis and for brevity, Chapter Two was condensed from the original manuscript. All manuscripts, with the exception of sections noted in Chapter Two, were written by Craig M. Butt with critical comments provided by Scott Mabury and Derek Muir. The contributions of the co-authors are detailed below.

Chapter One – Introduction of Poly- and Perfluorinated Compounds Contributions – Prepared by Craig Butt with editorial comments provided by Scott Mabury and Derek Muir

Chapter Two – Levels and Trends of Perfluorinated Compounds in the Arctic Environment Published In – The Science of the Total Environment (accepted pending minor revisions) Author List – Craig M. Butt, Urs Berger, Rossana Bossi and Gregg T. Tomy Contributions – Prepared by Craig M. Butt with exception of “Introduction” (Rossana Bossi) and “Spatial Trends” (Gregg Tomy) sections. Critical comments were provided by Urs Berger, Rossana Bossi and Gregg T. Tomy

Chapter Three – Thesis Goals and Hypothesis Contributions – Prepared by Craig Butt with editorial comments provided by Scott Mabury and Derek Muir

Chapter Four – Atmospheric Chemistry of 4:2 Fluorotelomer Acrylate

(C4F9CH2CH2OC(O)CH=CH2): Kinetics, Mechanisms and Products of Chlorine Atom and OH Radical Initiated Oxidation. Published In – Journal of Physical Chemistry A 2009, 113, 3155-3161 Author List – Craig M. Butt, Cora J. Young, Scott A. Mabury, Michael D. Hurley, Timothy J. Wallington Contributions – Smog chamber experiments were performed by Craig Butt with the assistance of Cora Young and Michael Hurley. Data interpretation and manuscript preparation was conducted by Craig Butt with critical comments provided by Cora Young, Michael Hurley and

xviii Timothy Wallington. The research and manuscript preparation was conducted under the guidance of Scott Mabury.

Chapter Five – Rapid Response of Arctic Ringed Seals to Changes in Perfluoroalkyl Production Published In – Environmental Science & Technology 2007, 41, 42-49 Author List – Craig M. Butt, Derek C.G. Muir, Ian Stirling, Michael Kwan and Scott A. Mabury Contributions – Sample extractions, instrumental analysis and data analysis was performed by Craig Butt. Ian Stirling coordinated sampling for archived Arviat ringed seal samples and Michael Kwan coordinated the sampling and liver sub-sectioning of recent Resolute Bay and Arviat collections. Craig Butt prepared all versions of the manuscript with critical comments provided by Scott Mabury and Derek Muir.

Chapter Six – Prevalence of Long-Chained Carboxylates in Seabirds from the Canadian Arctic between 1975 and 2004. Published In – Environmental Science & Technology 2007, 41, 3521-3528 Author List – Craig M. Butt, Scott A. Mabury, Derek C.G. Muir and Birgit M. Braune Contributions – Craig Butt performed liver extraction, instrumental analysis and data interpretation. Birgit Braune coordinated liver sample collection and sub-sampling. The manuscript was prepared by Craig Butt with the critical comments provided by Scott Mabury, Derek Muir and Birgit Braune.

Chapter Seven – Spatial Trends of Perfluoroalkyl Compounds in Ringed Seals (Phoca hispida) from the Canadian Arctic. Published In – Environmental Toxicology & Chemistry, 2008, 27, 542-553 Author List – Craig M. Butt, Scott A. Mabury, Michael Kwan, Xiaowa Wang and Derek C.G. Muir Contributions – Extraction of liver samples, instrumental analysis and data interpretation were performed by Craig Butt. Michael Kwan coordinated the sampling and liver sub-sectioning of recent ringed seal harvests. Xiaowa Wang coordinated the sampling handling at the Canadian Centre for Inland Waters. The manuscript was prepared by Craig Butt with critical comments provided by Scott Mabury and Derek Muir. This research was performed under the guidance of Derek Muir and Scott Mabury. xix

Chapter Eight – Bioaccumulation and Biotransformation of the 8:2 FTOH Acrylate in Rainbow Trout In preparation for submission to Environmental Science & Technology Author List – Craig M. Butt, Derek C.G. Muir and Scott A. Mabury Contributions – Fish care and sampling, preparation of dosed food, fish dissection, tissue extraction, synthesis of glucuronide and sulfate conjugates, S9 incubation experiments, instrumental analysis, data interpretation and manuscript preparation were performed by Craig Butt under the guidance of Derek Muir and Scott Mabury.

Chapter Nine – Elucidating the Mechanism of Poly- and Perfluorinated Acid Production in Rainbow Trout In preparation for submission to Environmental Science & Technology Author List – Craig M. Butt, Derek C.G. Muir and Scott A. Mabury Contributions – Fish care and sampling, synthesis of parent compounds, preparation of dosed food, fish dissection, tissue extraction, instrumental analysis, data interpretation and manuscript preparation were performed by Craig Butt under the guidance of Derek Muir and Scott Mabury.

Chapter Ten – Conclusions and Future Directions Contributions – Prepared by Craig Butt with editorial comments provided by Scott Mabury and Derek Muir

xx Other Publications During PhD:

Tseng, P.J.; Dinglasan-Panlilio, J.; Butt, C.M.; Mabury, S.A. Fluorotelomer acrylate polymer transformation in the environment. Manuscript in preparation.

Keller, J.M.; Calafat, A.M.; Kato, K.; Ellefson, M.E.; Reagen, W.K.; Strynar, M.; O’Connell, S.; Butt, C.M.; Mabury, S.; Small, J.; Muir, D.; Leigh, S.D.; Schantz, M.M. Determination of perfluorinated alkyl acid concentrations in human serum and milk standard reference materials. Anal. Bioanal. Chem. Accepted (October 7, 2009).

Furdui, V.I.; Stock, N.L.; Ellis, D.A.; Butt, C.M.; Whittle, D.M.; Crozier, P.W.; Reiner, E.J.; Muir, D.C.G.; Mabury, S.A. Spatial distribution of perfluoroalkyl contaminants in lake trout from the Great Lakes. Environ. Sci. Technol. 2007, 41, 1554-1559.

xxi

CHAPTER ONE

Introduction to Poly- and Perfluorinated Compounds

Contributions: Prepared by Craig Butt with critical comments provided by Scott Mabury and Derek Muir

1 2 Chapter One – Introduction to Poly- and Perfluorinated Compounds

1.1. Overview

The family of poly- and perfluorinated compounds (PFCs) comprises a large number of chemicals that are have wide use in commerce and industry [1]. The acronym, common name and chemical structure of the PFCs mentioned in this thesis are presented in Table 1.1. PFCs are attractive due to their stain repellency properties. Our knowledge of the fate and disposition of PFCs in the environment is largely due to recent developments in analytical capability, namely the development of liquid chromatography tandem mass spectrometry (LC-MS/MS) in the late 1990s. Two groups of PFCs, the perfluorinated carboxylates (PFCAs) and sulfonates (PFSAs) have received the bulk of scientific interest due to their widespread occurrence in humans [2], wildlife [3] and the abiotic environment [4]. The most widely known PFCs are probably perfluorooctane sulfonate (PFOS) and perfluorooctanoate (PFOA), although it has been shown that longer-chain PFCAs typically dominant PFCA profiles in wildlife [5]. The PFCs are unique as compared to many chlorinated and brominated organic compounds in that what is used in commercial products (e.g. fluorinated polymers) is generally not what is actually detected in the environment (i.e. PFCAs and PFSAs). With the exception of PFOA, and to a lesser degree perfluorononanoate (PFNA) and PFOS, the PFCAs and PFSAs were not directly produced in large quantities [6]. As will be discussed in more detail later in this chapter, the PFSAs and PFCAs are the degradation products, resulting from atmospheric and biological reactions, of commercial products (e.g. fluorinated phosphate surfactants) and compounds used in the manufacture of commercial products (e.g. fluorinated alcohol and acrylates).

3 Table 1.1. List of acronyms, common names and chemical structure.

Acronym Name Structure - Perfluorinated Carboxylates (PFCAs) CF3(CF2)xCOO - PFHxA perfluorohexanoate CF3(CF2)4COO - PFHpA perfluoroheptanoate CF3(CF2)5COO - PFOA perfluorooctanoate CF3(CF2)6COO - PFNA perfluorononanoate CF3(CF2)7COO - PFDA perfluorodecanoate CF3(CF2)8COO - PFUnA perfluoroundecanoate CF3(CF2)9COO - PFDoA perfluorododecanoate CF3(CF2)10COO - PFTrA perfluorotridecanoate CF3(CF2)11COO - PFTA perfluorotetradecanoate CF3(CF2)12COO - PFPA perfluoropentadecanoate CF3(CF2)13COO

- Perfluorinated Sulfonates (PFSAs) CF3(CF2)xSO3 - PFBS perfluorobutane sulfonate CF3(CF2)3SO3 - PFHxS perfluorohexane sulfonate CF3(CF2)5SO3 - PFOS perfluorooctane sulfonate CF3(CF2)7SO3 - PFDS perfluorodecane sulfonate CF3(CF2)9SO3

Telomer-Based Precursor Compounds

FTAc fluorotelomer acrylate CF3(CF2)xCH2CH2OC(O)CH=CH2

FTOH fluorotelomer alcohol CF3(CF2)xCH2CH2OH

FTO fluorotelomer olefin CF3(CF2)xCH=CH2

FTI fluorotelomer iodide CF3(CF2)xCH2CH2I

FTAL fluorotelomer aldehyde CF3(CF2)xCH2C(O)H

PFAL perfluorinated aldehyde CF3(CF2)xC(O)H

FTCA fluorotelomer saturated carboxylate CF3(CF2)xCH2C(O)OH

FTUCA fluorotelomer unsaturated carboxylate CF3(CF2)xCF=CHC(O)OH

POSF-Based Precursor Compounds

N-MeFOSE N-methyl perfluoroocane sulfamide ethanol CF3(CF2)7SO2N(CH3)CH2CH2OH

N-EtFOSE N-ethyl perfluoroocane sulfamide ethanol CF3(CF2)7SO2N(CH2CH3)CH2CH2OH

N-MeFOSA N-methyl perfluoroocane sulfamide CF3(CF2)7SO2NH(CH3)

N-EtFOSA N-ethyl perfluoroocane sulfamide CF3(CF2)7SO2NH(CH2CH3)

PFOSA N-ethyl perfluoroocane sulfamide CF3(CF2)7SO2NH2

4 1.2. Perfluorooctane Sulfonyl Fluoride-Based Compounds

Perfluorooctane sulfonyl fluoride-based (POSF) compounds are chemicals produced using the electrochemical fluorination (ECF) process [7]. As emphasized later in this section, on a global scale the 3M company was the major producer of POSF-compounds until 2002. The ECF process was used to manufacture POSF through the reaction,

C8H17SO2F + 34 HF → C8F17SO2F + 17 H2

POSF is a usable commercial product itself, although the dominate use was to synthesize other fluorinated sulfonyl compounds. POSF can be hydrolyzed to directly form PFOS. PFOS was used in aqueous film forming foams (AFFF) for fire fighting applications as well as in coating additives. However, the vast majority of POSF was used to manufacture intermediates for high molecular weight polymers and phosphate esters [7]. POSF is reacted with methyl or ethyl amine to yield the N-methyl and N-ethyl perfluorooctane sulfonamide (N-MeFOSA, N- EtFOSA). The N-MeFOSA and N-EtFOSA is reacted with ethylene carbonate to yield the N- methyl and N-ethyl perfluorooctane sulfonamide ethanols (N-MeFOSE, N-EtFOSE). The FOSA and FOSE compounds were the key building blocks in the manufactured commercial products [7]. The ECF process was also used to make PFOA until 2002.

The historical production volumes and estimated environmental releases of POSF-based compounds was recently reviewed by Paul et al. [8] (Figure 1.1). Estimates were derived using publically available information and “POSF” represents all POSF-based compounds and potential PFOS precursors. As stated earlier, on a global scale, 3M was the major producer of POSF-based chemicals, accounting for ~78% of the global POSF production in 2000. POSF was produced using the electrochemical fluorination process, predominately in Decatur, Alabama and Antwerp, Belgium. POSF compounds were manufactured by 3M starting in 1949, although it has been stated that production volumes were minimal prior to1970 [8]. In 2000, 3M announced the voluntary phase out of POSF and associated compounds, although actual production apparently did not cease until 2002. The historical POSF production volume estimated by Paul et al. [8] was 96 000 tonnes, including ~470 tonnes of direct PFOS production. These results indicate that historical direct production of PFOS was ~0.5% of POSF production. The POSF production estimates of Paul et al. [8] were similar to those of Prevedouros et al. [6], but 2-fold higher than Smithwick et al. [9]. Despite the voluntary phase- out by 3M, production of POSF in China has occurred since 2003, although exact production

5 volumes unclear. (http://chm.pops.int/Portals/0/Repository/ comments_draftRME2008/UNEP- POPS-POPRC-DRME-08-CHI-SCCP.English.PDF). Paul et al. [8] assumes this value to be approximately 1000 tonnes per year from 2002 onward.

In recent years the 3M Company has released new products based on a four-carbon compound (i.e. perfluorobutane sulfonyl fluoride) [10]. PFBS has been detected in wastewater treatment plant effluent and river water [11], river sediments [12] and house dust [13]. However, Olsen et al. [14] showed the perfluorobutane sulfonate (PFBS) had much shorter half- lives in rats, monkeys and humans as compared to PFOS. For example, in humans the half-life for PFBS was 26 days as compared to 2662 and 1751 days for PFHxS and PFOS, respectively.

5000 Smithwick et al. Prevedouros et al. 3M Global Production (Paul et al.) 4000 Total Global Production (Paul et al.)

3000

2000

1000 POSF ProductionPOSF Volume (tonnes)

0 1970 1975 1980 1985 1990 1995 2000 2005

Year

Figure 1.1. Estimated global POSF production volumes (1970-2002). Data taken from Paul et al. [8].

6 1.3. Fluorotelomer-Based Compounds

Fluorotelomer-based compounds have been produced since the 1970s and have similar applications to that of POSF-based compounds [6]. The fluorotelomer production scheme is shown in Figure 1.2. The telomerization process was developed by the Du Pont Company and involves the reaction of a taxogen (CF2=CF2) and a telogen (perfluoroalkyl iodide, “telomer A”,

F(CF2CF2)nI), yielding a longer perfluoroalkyl iodide. The perfluoroalkyl iodides themselves cannot be converted to produce intermediates for fluorinated surfactants and are thus reacted with ethylene to yield the fluorotelomer iodide (FTI). The FTIs can further be reacted to produce the fluorotelomer alcohols (FTOHs), fluorotelomer olefins (FTOs) and fluorotelomer acrylates (FTAcs). Fluorotelomer compounds have the general formula of F(CF2)nCH2CH2X, where n is an even number and X=OH, I, OC(O)CH=CH2 with the exception of the FTOs in which the formula is simply F(CF2)nCH=CH2. In addition, the telomerization process is used to produce PFOA and to a lesser degree, PFNA.

Similar to POSF-based compounds, the majority of fluorotelomer-based compounds have been used to manufacture high molecular weight polymers. Of these, FTAc-based polymers comprise the largest commercial category of polyfluorinated products [15]. The FTAc monomers have been shown to be residuals in a FTAc-based polymer [15], and may be released to the atmosphere from commercial fluorotelomer-based polymer products in a similar manner to FTOHs [16].

7

CF2=CF2 TFE

F(CF2CF2)nI Telomer A

n=3-8

Fluorotelomer F(CF2CF2)nCH2CH2I Iodide (FTI)

F(CF CF ) CH CH OH Fluorotelomer 2 2 n 2 2 Alcohol (FTOH)

F(CF2CF2)nCH=CH2 Fluorotelomer F(CF CF ) CH CH OC(O)CHR=CH Olefin (FTO) 2 2 n 2 2 2 Fluorotelomer Acrylate Monomer (FTAc)

Sales Products

Figure 1.2. Scheme of fluorotelomer-based chemical production. Figure from Prevedouros et al. [6].

8 1.4. Sources of Poly- and Perfluorinated Acids to the Environment

As mentioned earlier, PFCAs and PFSAs are widely distributed in the environment. Despite their apparently ubiquitous presence, the sources of PFC contamination are not well understood, particularly in remote environments such as the Canadian Arctic. Two mechanisms have been proposed, the “direct” and “indirect” sources. The relevance of each mechanism to overall human and wildlife exposure is unclear and is of intense scientific interest. The hypothesized transport pathways specific to the arctic environment are thoroughly described in Chapter 2, Section 2.3. In this section, the generalized source pathways are described with particular emphasis on the biotransformation of polyfluorinated compounds to yield PFCAs and PFSAs.

1.4.1. Direct Sources Direct sources refer to PFCAs and PFSAs that originate from releases during the production, use and disposal of fluorochemicals. The relevance of direct sources to the environmental burden of PFCs is described in Chapter Two. In addition, the magnitude of direct sources have been described by Paul et al. [8] for POSF-based compounds and Prevedouros et al. [6] for telomer-based compounds.

1.4.2. Indirect Sources Indirect sources refer to those that yield PFCAs and PFSAs through the degradation of precursor compounds (“precursors”). The two main classes of degradation reactions are atmospheric oxidation and biotransformation, and will be examined separately in this section. In addition, it has also been shown that the indirect photolysis of FTOHs can yield PFCAs [17], although the relevance of this mechanism is unclear considering the low water solubility of FTOHs.

1.4.2.1 Atmospheric Reactions The atmospheric chemistry that results in PFCA formation from precursor compounds has recently been reviewed in detail by Young & Mabury [18]. There are many compounds (e.g. HCFCs, fluorinated anesthetics) that form short-chain PFCAs, but these compounds are not likely to be bioaccumulative and thus are not of interest as potential human and wildlife sources. However, laboratory experiments have shown that FTOHs [19, 20], fluorotelomer olefins [21, 22] and iodides [23] can form PFCAs via atmospheric reactions. In addition, a FBSE has been

9 shown to form PFBS and PFCAs [24] and a FBSA can form PFCAs [25]. Presumably the FOSE and PFOSA would react in analogous mechanisms to form PFOS and the longer-chain PFCAs. Therefore, it has been shown that most of the major building blocks used in the manufacture of fluorinated commercial products can degrade in the atmosphere to yield PFCAs. However, presently it is not known if the fluorotelomer-based acrylates (FTAcs) are PFCA precursors via atmospheric reactions. This is relevant since FTAcs have been shown to be residuals in commercial fluorotelomer-based polymers [15] and are widely detected in the atmosphere [26-30].

The mechanism for PFCA formation from fluorotelomer-based precursors is thoroughly described by Young & Mabury [18]. As shown in Figure 1.3, the mechanism proceeds through common intermediates. The atmospheric degradation of FTOHs, FTIs and FTOs yields the perfluorinated aldehyde (PFAL) as the ultimate precursor to PFCAs. In addition, the FTOHs and FTIs form the fluorotelomer aldehyde (FTAL) as another common intermediate. Formation of the FTAL precedes that of the PFAL and can also lead to the formation of FTCAs. Therefore, any compound that degrades to form the FTAL or PFAL will ultimately yield PFCAs in the atmosphere.

In contrast to the telomer-based compounds, considerably less is known regarding the formation of PFSAs from volatile precursors. However, the production of PFSAs from the sulfonamide alcohols was shown by D’eon et al. [24]. This mechanism involves addition of OH radical to the sulfone double bond, resulting in S-N cleavage of the intermediate radical, ultimately forming the sulfonate. It was shown that the yield of PFCAs was greater than that of PFBS, implying that cleavage of the S-N bond was less favourable than that of the S-C bond.

The reaction mechanisms described above for the formation of PFCAs are favoured in a low NOx environment. In the presence of NOx, these compounds primarily unzip to form carbonyl fluoride. These mechanisms depend on reactions with HO2 and RO2 which are more likely in remote low NOx locations. The atmospheric lifetimes of the FTOHs, FTIs and FTOs and sulfonamide alcohols are sufficient to allow transport to remote regions where this chemistry can occur [18].

10

Fluorotelomer Alcohol CF3(CF2)nCH2CH2OH

1. +OH/-H2O 2. +O 2 3. +RO2/-RO/-O2 4. +O2/-HO2 1. hv/-I 3. +RO2/-RO/-O2 2. +O 4. +O /-HO Fluorotelomer Iodide 2 2 2 CF3(CF2)nCH2C(O)H CF3(CF2)nCH2CH2I

+OH/-H2O

+O hv/-HCO +HO /-O 2 2 3 CF3(CF2)nCH2C(O)OO• CF3(CF2nCH2C(O)•

+RO2/-RO/-O2 CF3(CF2)nCH2C(O)OH +HO2/-O2 OR -CO +NO/-NO2 -CO2 CF (CF ) CH C(O)O• CF (CF ) CH • CF (CF ) CH C(O)OOH 3 2 n 2 3 2 n 2 3 2 n 2 1. +O2 2. +RO2/-RO/-O2 1. +OH/-H2O 3. +O2/-HO2 2. +RO2/-RO/-O2 3. -CH OH Fluorotelomer Olefins 2 CF3(CF2)nC(O)H CF3(CF2)nCH=CH2

+OH/-H2O +O 2 hv/-HCO CF3(CF2)nC(O)OO• CF3(CF2)nC(O)• +HO /-O 2 3 +RO2/-RO/-O2 OR -CO +NO/-NO2

-CO2 CF3(CF2)nC(O)OH CF (CF ) C(O)O• CF (CF ) • 3 2 n 3 2 n

+O2

-HF +RCHR’O2/-RCOR’ CF3(CF2)yC(O)F CF3(CF2)xOH CF (CF ) OO• 3 2 x +RO2/-RO/-O2 Hydrolysis +O2 OR +NO/-NO2 -COF CF (CF ) C(O)OH 2 3 2 y CF3(CF2)y• CF3(CF2)xO•

Figure 1.3. Atmospheric reaction scheme for the production of PFCAs (boxes with dashed outlines) from fluorotelomer-based compounds (boxes with solid outlines). Figure adapted from Young et al. [31].

11 1.4.2.2 Polyfluorinated Compound Biotransformation

1.4.2.2.1 Fluorotelomer-based Compounds In 1981, Hagen et al. [32] published the original study investigating the biotransformation of FTOH-based compounds. Research was stagnant for over 20 years until Dinglasan et al. [33] published the biotransformation of 8:2 FTOH in a mixed microbial system. Since that time many studies have been conducted investigating the metabolism of FTOH-based compounds. At present, studies have utilized the 8:2 FTOH as the parent compound with the exception D’eon & Mabury [34] who investigated the biological fate of polyfluoroalkyl phosphates, which were shown to be direct precursors to the FTOHs. Thus, there is a need for additional studies on the biotransformation of compounds which are potential FTOH precursors, such as the FTAcs. The overwhelming majority of metabolism studies have been performed with either microbial systems or rats and mice. Presently, there has been very limited study of FTOH-based compounds in fish with exception of Nabb et al. [35] who examined FTOH biotransformation in rainbow trout hepatocytes, and liver cytosol and microsomes.

1.4.2.2.1.1. Microbial Dinglasan et al. [33] investigated the biotransformation of 8:2 FTOH in sewage sludge under aerobic conditions. The parent compound and volatile metabolites were monitored by GC-MS and nonvolatile metabolites monitored by LC-MS/MS. It was shown that 85% of the parent 8:2 FTOH was degraded by day 7. No degradation was observed in the sterile control vessels, indicating that microbial degradation was primarily responsible for the observed degradation. Concurrent with the depletion of 8:2 FTOH was the production of the 8:2 FTCA and 8:2 FTUCA. The 8:2 FTAL was also identified as a transient intermediate between the 8:2 FTOH and 8:2 FTCA, but levels of this metabolite could not be confidently quantified. The 8:2 FTCA was also a transient metabolite and depletion of this intermediate was coincident with the increase in 8:2 FTUCA levels. The authors suggested that the 8:2 FTCA degradation could have proceeded via abiotic hydrolysis or by enzymatic mechanisms. However, it was noted that the abiotic degradation of FTCAs proceeds at a slower rate (half-lives greater than 1 week) than observed in the microbial study. PFOA was formed during the experiment, reaching approximately 3% of the total mass by day 81. PFNA was not observed indicating that α- oxidation did not occur in the microbial system. Overall, a poor mass balance was achieved with only ~55% of the products accounted for at the end of the experiment. The poor mass

12 balance may have been the result of the inability to quantify the 8:2 FTAL, the production of other unidentified metabolites, and the potential production of non-extractable metabolites that were covalently bound to biological macromolecules. A biodegradation scheme was developed, building upon the experimental results as well as the findings of Hagen et al. [32] (Figure 1.4). It was proposed that the first biotransformation step was the oxidation of 8:2 FTOH, by an alcohol dehydrogenase enzyme, to form the 8:2 FTAL which is subsequently oxidized by an aldehyde dehydrogenase to yield the 8:2 FTCA. It was proposed that the 8:2 FTCA can enter β- oxidation, ultimately yielding PFOA. As stated above, the 8:2 FTCA could degrade to the 8:2 FTUCA through either abiotic or biotic mechanisms. It was postulated that the formation of PFOA from the 8:2 FTUCA would be slow due to the high thermodynamic cost of oxidizing the FTUCA β-carbon.

Wang et al. [36] examined the biodegradation of 14C-labelled 8:2 FTOH 14 [CF3(CF2)6 CF2CH2CH2OH] in activated sewage sludge under aerobic conditions for 28 days. The use of the radiolabelled parent compound allowed the researchers to obtain mass balance. At the end of the 28 day experiment, 16% of the parent 8:2 FTOH was unchanged. Further, a significant portion (41%) of the total 8:2 FTOH was adsorbed to the PTFE septa. The metabolites observed in the active vessels were the 8:2 FTCA, 8:2 FTUCA, PFOA and the novel metabolite, 7:3 FTCA. At the conclusion of the experiment, 27% of the mass balance was 8:2 FTCA, 6.0% was 8:2 FTUCA and 2.3% was 7:3 FTCA. PFOA accounted for 2.1% of the total mass balance. Several other unidentified metabolites were also observed, but these compounds each contributed <1% of the total mass balance. Similar to Dinglasan et al. [33], PFNA was not observed indicating that α-oxidation of the 8:2 FTCA did not occur. No metabolites were identified in the abiotic control vessels. Based on the experimental results, the authors proposed an 8:2 FTOH biotransformation reaction scheme. It was postulated that the first degradation step is the formation of 8:2 FTAL via alcohol dehydrogenase, although this analyte was not observed during the experiment. This is followed by formation of the 8:2 FTCA which may be mediated by aldehyde dehydrogenase. The 8:2 FTCA can react to yield the 8:2 FTUCA which can undergo monooxygenase-mediated reactions to yield PFOA. It was suggested that the proton deficiency of the 8:2 FTCA and 8:2 FTUCA effectively prevents these compounds from entering the β-oxidation cycle. A novel mechanism for PFOA formation was proposed. It was

13 postulated that the 8:2 FTCA can form the 7:3 FTCA, through an unknown mechanism, which can subsequently act as a substrate for β-oxidation and ultimately form PFOA.

Wang et al. [37] extended their earlier sewage sludge experiments by investigating the biotransformation of 14C-8:2 FTOH in mixed bacterial cultural and activated sludge. The mixed bacterial culture collected had previously been exposed FTOHs and thus was presumably “conditioned” to degrade telomer-based compounds. Bottles were subject to either closed or continuous air flow conditions for up to 4 months. No degradation was observed in the sterile controls. Similar to their previous experiment [36], significant 8:2 FTOH sorption to the PTFE septa in the bottles was observed (5.2% - 17.4% of mass balance). Metabolites identified included 5 previously reported compounds (8:2 FTAL, 8:2 FTCA, 8:2 FTUCA, 7:3 FTCA and PFOA) as well as 3 novel metabolites (7:2 sFTOH, 7:3 FTCA and 7:3 unsaturated amide). In addition, PFHxA was formed in low yields, ~1% of the initial 8:2 FTOH concentration, which the authors state is evidence for degradation of the fluorinated chain. In the mixed bacterial culture, the 7:2 sFTOH was the most abundant metabolite observed, comprising 14% of the mass balance at day 90. PFOA was formed in comparatively lower yields, accounting for 6% of the mass balance by the conclusion of the experiment. Consistent with their previous study, PFNA was not detected in the active vessels, suggesting that α-oxidation of the 8:2 FTCA did not occur. Based on the newly identified metabolites, a novel series of degradation mechanisms was proposed [37]. The initial degradation steps were consistent with the previous study [36], that is the oxidation of the 8:2 FTOH to 8:2 FTAL by alcohol dehydrogenase, followed by oxidation to 8:2 FTCA then conversion to the 8:2 FTUCA by HF elimination. It was postulated that the 8:2 FTUCA represents a key branching point in the microbial degradation of 8:2 FTOH. In the first branch, the 8:2 FTUCA is enzymatically decarboxylated to form the 8:1 olefin which is defluorinated to yield the 7:2 olefin. The 7:2 olefin could be oxidized through P450 reactions to form the 7:2 sFTOH. It was proposed that the 7:2 sFTOH could yield PFOA, although the mechanism is uncertain. The second branch involves defluorination of the 8:2 FTUCA to the 7:3 FTUCA which can then yield the 7:3 unsaturated amide via transaminase. It was proposed that the 7:3 unsaturated amide could form PFOA, however, the mechanism was not explained. The third branch postulates that after the 8:2 FTUCA forms the 7:3 FTUCA, the 7:3 FTUCA is reduced to the 7:3 FTCA which acts as substrate for β-oxidation, ultimately yielding PFOA.

14 Consistent with their previous mechanism [36], it was stated that the 8:2 FTCA and 8:2 FTUCA could not proceed through β-oxidation.

Wang et al. [38] examined the biodegradation of 14C-8:2 FTOH in aerobic soils under closed conditions and continuous air flow. It was shown that between 10-35% of the mass balance was irreversibly bound to the soil and could not be solvent extracted. The 8:2 FTOH was degraded in the soil and at least 11 metabolites were observed, 7 of which had been previously reported. Three metabolites were identified that were previously not reported for soil 8:2 FTOH degradation and including the 3-OH-7:3 FTCA, the 7:2 ketone and 2H-PFOA. PFOA and the 7:3 FTCA were formed in relatively high yields, averaging 25% and 11% of the mass balance, respectively. It should be noted that these yields are at least two orders of magnitude greater than other reported yields. The 7:2 sFTOH, 8:2 FTUCA and 7:3 FTCA were transient metabolites. Interestingly, the 8:2 FTCA was not observed which was attributed to the rapid degradation of this compound. The formation of PFNA was not observed, however, PFHxA was observed in the 8:2 FTOH soil incubations. To investigate the 8:2 FTOH biodegradation pathway, several identified or proposed intermediate metabolites were incubated and soil and activated sludge for 90 days, including the 8:2 FTUCA, 7:3 FTCA, 7:3 FTUCA and 7:2 sFTOH. It is not known why the 8:2 FTCA was not dosed since represents a potential branching point in the biotransformation mechanism. Incubations with the 7:2 sFTOH resulted in the formation of PFOA which was hypothesized to be produced through several monooxygenase mediated reactions. Therefore, it was postulated that the 7:2 sFTOH was the ultimate precursor to PFOA. The 7:2 sFTOH was formed during incubations with the 8:2 FTUCA but not the 7:3 FTUCA. Incubations with the 7:3 FTCA did not yield any metabolites, including PFOA, indicating that this compound is a terminal metabolite in soil systems. However, 7:3 FTUCA incubations did yield the 7:3 FTCA. Based on the experimental findings and building upon previous 8:2 FTOH metabolism studies in microbial systems and mammals performed by this group, a series of degradation pathways was proposed. Again, the initial steps of biotransformation are consistent with previous studies (8:2 FTOH > 8:2 FTAL > 8:2 FTCA > 8:2 FTUCA) and the 8:2 FTUCA represents a key branching point. In the first branch, it was proposed that the 8:2 FTUCA can be defluorinated to yield the 7:3 FTUCA which then be degraded to the PFHxA, reduced to the 7:3 FTCA, or oxidized to the 3-OH-7:3 FTCA. In the second branch, the 8:2 FTUCA forms the 7:2 sFTOH which, as stated above, was hypothesized

15 to be the ultimate precursor to PFOA. Alternatively the 7:2 sFTOH can be oxidized to the 7:2 ketone, then defluorinated to the 3-H, 7:2 ketone and ultimately the 2H-PFOA.

Liu et al. [39] investigated the aerobic biotransformation of 8:2 FTOH and metabolite profile in a clay loam soil as well as with two isolated soil bacterial cultures. In addition, the influence of three carrier solvents (ethanol, octanol, 1,4-dioxane), which may act as carbon sources, was investigated. In the soil studies, the fastest 8:2 FTOH biotransformation rate was observed when 1,4-dioxane was the carrier solvent, followed by octanol. Assuming a first-order reaction, the 8:2 FTOH biotransformation rates were 0.28 day-1 with 1,4-dioxane, 0.18 day-1 with ethanol, and 0.13 day-1 for octanol. Some loss of 8:2 FTOH (12%) was observed in the sterile controls. The authors attributed this loss to irreversible sorption since only trace levels of metabolites were detected. Metabolites formed from the 8:2 FTOH biotransformation included 8:2 FTCA, 8:2 FTUCA, 7:3 FTCA, 7:3 FTUCA, 7:2 sFTOH, PFHxA, PFHpA and PFOA. The metabolite profile was consistent between the three carrier solvents. At the end of the experiment (day 7), the dominant metabolites were 8:2 FTUCA (6.1-8.4 mol %, relative to applied 8:2 FTOH) followed by 8:2 FTCA (2.3-2.9 mol %) and 7:2 sFTOH (2.3-3.5 mol %). Formation of PFOA was comparatively low (0.6-0.8%). The mass balance decreased with time to which the authors contributed to irreversibly bound metabolites and unmonitored metabolites such as fluorotelomer aldehydes. It was shown that the two soil bacterial strains investigated, Pseudomonas species OCY4 and OCW4 were capable of biotransforming 8:2 FTOH without prior exposure or acclimation to 8:2 FTOH. The bacteria did not appear to use 8:2 FTOH as an energy source. It was shown the degradation yielded all of the metabolites observed in the soil incubations, with the exception of 7:2 sFTOH in the open system. Similar to the soil incubations, mass balance was not achieved potentially due to unknown or unmonitored metabolites. The authors postulated that a partial β-oxidation mechanism was responsible for the production of PFOA from 8:2 FTCA or 8:2 FTUCA.

1.4.2.2.1.2. Rats and Mice The original experiment on FTOH biotransformation in mammals was performed by Hagen et al. [32] in the early 1980s. In this study, Hagen et al. [32] dosed male rats via one single oral dose and periodically sacrificed up to 20 days post-dosage. Plasma samples were solvent extracted, diazomethane derivitized and analyzed by GC coupled with a microwave- sustained helium plasma detector (MPD). A detailed time course of 8:2 FTOH depletion and

16 metabolite formation was not presented by the authors. Further, the 8:2 FTOH half-life was not given. However, PFOA, the 8:2 FTCA and an unidentified metabolite were detected 2 hours post-treatment. The 8:2 FTUCA was also formed. Plasma levels of the 8:2 FTCA and the unidentified metabolite were transient, decreasing with time from an initial peak early in the experiment, ultimately reaching very low levels by 48 hours. In contrast, PFOA levels increased throughout the experiment. A detailed mechanism was not presented, however, it was suggested that β-oxidation was involved.

Approximately 25 years after the initial work by Hagen et al. [32], Martin et al. [40] performed a 8:2 FTOH dosing study using similar experimental conditions. The purpose of the study was to replicate the original findings of Hagen et al. [32], while employing improved analytical techniques. In the Martin et al. [40] study, the animal was euthanized at 6 hours post- dosage and blood, liver and kidney were analyzed. The 8:2 FTCA, 8:2 FTUCA, PFOA and PFNA were detected in the tissue. PFNA was formed at levels approximately 10-fold lower than PFOA. In addition, the conjugated metabolites, the O-glucuronide and O-sulfate, were identified via comparison to expected mass spectra. 8:2 FTOH time trends were not reported. In a second experiment, Martin et al. [40] investigated FTOH biotransformation in hepatocytes isolated from male Sprague-Dawley rats. Hepatocytes were incubated with the 4:2, 6:2, 8:2 and 10:2 FTOH, although discussion of study results focused on the 8:2 FTOH experiments. In the 8:2 FTOH incubation, it was shown that 78% of the parent compound had been depleted by 4 hours post-treatment although poor mass balance was obtained. The formation of the 7:3 FTCA and 7:3 FTUCA was also observed during these incubations. The O-glucuronide and O-sulfate conjugated metabolites were detected. Three GSH conjugated were also detected, the GSH- FTUAL, GSH-FTUCA and a reduced GSH-FTUAL conjugate. As well, the formation of aldehyde metabolites was investigated by derivatizing the hepatocyte medium with DNPH at four time points (30 min, 1 hour, 2 hours and 4 hours). The formation of the 8:2 FTAL and 8:2 FTUAL was observed by comparison to an authentic standard product spectrum (8:2 FTAL) and expected spectrum (8:2 FTUAL). The quantified metabolites accounted for only 8.5% of the transformed products: 8:2 FTCA (2.9%), 8:2 FTUCA (4.1%), PFOA (1.4%) and PFNA (<0.2%). These results indicate an overall low PFOA yield from 8:2 FTOH biotransformation. It was suggested that the unaccounted mass balance may be partially explained by the formation of the conjugated metabolites (three glutathione conjugates, one glucuronide and one sulfate) and to a lesser extent, the formation of 7:3 FTCA and 7:3 FTUCA. The contribution of these

17 metabolites to the total mass balance could not be determined since authentic standards were not available.

Martin et al. [40] further investigated the FTOH degradation mechanism by incubating the hepatocytes with some of the identified intermediate metabolites. When hepatocytes were incubated with the 8:2 FTAL for 2 hours, no 8:2 FTAL or 8:2 FTUAL was detected, but the formation of small quantities of PFOA, PFNA, 8:2 FTCA and 8:2 FTUCA was observed. However, a low mass balance was obtained to which the authors suggested the formation of acid metabolites was not the primary fate of the FTAL. Individual incubations with 8:2 FTCA and 8:2 FTUCA showed these compounds biotransformed much more slowly than the 8:2 FTOH with less than 10% of parent compound biotransformed 2 hours post-treatment. In these incubations, a much better mass balance was obtained.

Kudo et al. [41] investigated the biotransformation of 8:2 FTOH in male mice dosed via intraperitoneal injection. Animals were euthanized at periodic intervals up to 72 hours post- treatment. The serum and liver were analyzed for suspected 8:2 FTOH metabolites but the parent compound itself was not monitored. PFOA and two unidentified metabolites were observed in the liver within 2 hours post-treatment in the serum and liver, implying rapid biotransformation of the 8:2 FTOH. The 8:2 FTCA and the two unidentified metabolites peaked around 6 hour post-treatment in the liver, implying that these compounds were intermediate metabolites. In contrast, PFOA levels continued to increase throughout the experiment. In a second experiment, Kudo et al. [41] continuously exposed male mice via the diet and animals were sacrificed at weekly intervals up 28 days. Similar to the intraperitoneal exposure, the 8:2 FTOH was not monitored. PFOA and PFNA levels were shown to increase in a dose- and time- dependent manner. The formation of PFNA was approximately 10-fold lower than that of PFOA. One of the unidentified metabolites in the intraperitoneal exposure (“metabolite a”) was low throughout the experiment but the other metabolite (“metabolite b”) was not detected. In both experiments, the authors suggested that the first metabolite formed was the 8:2 FTCA.

Fasano et al. [42] investigated the biotransformation of radiolabelled [3-14C] 8:2 FTOH in male and female rats exposed via a single oral dose. 8:2 FTOH levels in plasma peaked at 1 hour post-treatment and were rapidly eliminated. The half-life was 1-5 hours for 8:2 FTOH and was 4-18 hours for 8:2 FTCA. PFOA half-lives were much longer, 112-217 hours in males and

18 5.6-17.5 hours in females. The majority of the 14C activity (>70%) was excreted in the feces, primarily as the parent 8:2 FTOH (34-50%). Bile elimination was also significant, whereas <4% of the activity was eliminated in the urine. Bile metabolites were predominately glucuronide and glutathione conjugates. It was shown that overall only a minor portion of the parent FTOH biotransformed to PFCAs with the majority metabolized to 8:2 FTUCA-GSH and 8:2 FTUAL-GSH conjugates. Interestingly, several PFCAs with chain-lengths smaller than

PFOA were observed, suggesting the potential for CF2 degradation. Based on their findings, Fasano et al. [42] proposed a metabolic degradation pathway for 8:2 FTOH in rats. The first step involves the series of oxidations of 8:2 FTOH > 8:2 FTAL > 8:2 FTCA. It was proposed that the 8:2 FTCA can undergo dehydrofluorination to yield 8:2 FTUCA, or α-oxidation to yield PFNA. Again, similar to other FTOH biotransformation pathways proposed by this group, it was suggested that the 8:2 FTUCA was an important branching point. One branch involves the reduction and dehydrofluorination of 8:2 FTUCA to form the 7:3 FTUCA which can be subsequently reduced to yield the 7:3 FTCA. It was postulated that both the 7:3 FTCA and 7:3 FTUCA can undergo β-oxidation to yield PFOA. The second branch involves the formation of the 8:1 olefin from the 8:2 FTUCA. The 8:1 olefin subsequently loses fluorine to yield the 7:2 olefin and ultimately form the 7:2 secondary FTOH. It was proposed that the 7:2 secondary FTOH could then react to from PFOA, although the mechanism is unclear. Finally, the FTUAL and FTUCA can form GSH conjugates, and the 8:2 FTOH and 7:2 secondary FTOH can form glucuronide and sulfate conjugate metabolites.

Henderson & Smith [43] investigated 8:2 FTOH biotransformation in time-pregnant mice (gestational day 8) exposed via a single oral gavage dose of 30 mg 8:2 FTOH/kg body weight. The 8:2 FTOH and intermediate metabolites, 8:2 FTCA and 8:2 FTUCA, were not detected in maternal serum or liver 24 hours post-treatment (the first sampling point after dosing). However, PFOA and PFNA were detected in the maternal serum and liver 24 hours post-treatment. These results are consistent with rapid biotransformation of the 8:2 FTOH observed in previous studies.

Nabb et al. [35] investigated the in vitro metabolism of [3-14C] 8:2 FTOH in rat, mouse, trout and human hepatocytes and in rat, mouse and human liver microsomes and cytosol fractions. In addition, incubations were performed with the 8:2 FTCA, 8:2 FTUCA, 7:3 FTCA and 7:3 FTUCA to further investigate the mechanism of FTOH metabolism. Differences

19 between the species were observed with 8:2 FTOH clearance rates following the rank order of rat > mouse > human ≥ trout. The yield of PFOA was low and represented between 0.02-0.47% of the initial 8:2 FTOH dosed, on a molar basis. Several novel metabolites were identified including the 7:2 ketone, 7:3 β-keto acid, 7:3 FTAL, 7:3 FTUAL and 7:3 FTCA taurine conjugate. A series of biotransformation mechanisms were proposed, building on previous studies as well as the study observations. Once again, the initial biotransformation steps are consistent with previous studies which comprise the series of oxidations of 8:2 FTOH > 8:2 FTAL > 8:2 FTCA > FTUCA. The 8:2 FTAL can be oxidized to form the 8:2 FTUAL and the 8:2 FTCA can undergo α-oxidation to yield PFNA and lower odd-chain PFCAs. However, in the mechanism proposed by Nabb et al. [35] the branching points are at both the 8:2 FTUAL and 8:2 FTUCA. It was postulated that the 8:2 FTUAL will form the 7:3 β–hydroxy unsaturated aldehyde followed by the 7:3 β-keto aldehyde, ultimately yielding PFOA. An additional fate of the 8:2 FTUAL is formation of the 7:3 FTUAL, ultimately yielding the 7:3 FTCA and 7:3 FTCA-TA conjugate. As well, two branches were proposed for 8:2 FTUCA biotransformation. One branch will ultimately yield the 7:2 sFTOH and 7:2 sFTOH glucuronide conjugate, whereas the second branch will yield the 7:3 FTUCA followed by the 7:3 FTCA and finally the 7:3 FTUCA-TA conjugate. Finally, Fasano et al. [44] recently published a 8:2 FTOH biotransformation scheme that is essentially a combination of the mechanisms previously proposed by Fasano et al. [42] and Nabb et al. [35] (Figure 1.5).

D’eon and Mabury [34] investigated the biotransformation of 8:2 fluorotelomer based phosphates (mono- and di-PAPs) in male Sprague-Dawley rats. Animals were dosed by oral gavage and blood levels were monitored for the parent compounds and metabolites for 15 days post-treatment. It was hypothesized that the PAPs could be biotransformed through the cleavage of the phosphate ester linkage via phosphatase enzymes, yielding the 8:2 FTOH which would be subsequently oxidized to form PFOA. In both the mono- and di-PAPs exposure, PFOA was formed and accumulated at low ng/g levels. Also, PFHpA was formed in comparatively lower levels. Further, several transient intermediate metabolites (8:2 FTCA, 8:2 FTUCA and 7:3 FTCA) were observed. PFNA was not detected above blank levels. The formation of these intermediate and terminal metabolites is consistent 8:2 FTOH biotransformation, although the 8:2 FTOH was not monitored in the PAPs experiments.

20 1.4.2.2.1.3. Biotransformation of Fluorotelomer-based Compounds: Summary and Conclusions Numerous studies have shown that 8:2 FTOH metabolism results in the formation of PFOA, and to a smaller fraction PFNA and lower chain-length PFCAs. It is clear that the overall yield of PFOA is low, presumably due to the fact that conjugation reactions with the 8:2 FTOH occur as well as the apparent branching in the overall biotransformation mechanism. There has been considerable evolution of the hypothesized mechanisms for biotransformation since the seminal work by Hagen et al. [32]. In particular, with the observation of novel metabolites, the proposed mechanisms have become increasingly complex. Presumably much of this development is result of improved analytical capabilities, synthesis of novel potential metabolites and the use of 14C-labeled compounds. More recent studies have shown that the key branching point in the mechanism is at 8:2 FTUCA, and possibly the 8:2 FTUAL. Several researchers have attempted to interrogate the biotransformation mechanism through dosing experimental systems with previously identified intermediate metabolites. However, uncertainty and contradictory mechanisms still exist in the literature. Nonetheless, it is evident that the formation of PFOA proceeds through a mechanism that is similar to that of β-oxidation, although it may not involve the same enzyme cofactors. Therefore, it is clear that additional mechanistic studies are warranted.

F F F F F F F F O F F F F F F F F O F F - O S-CoA F F F F F F F F F F F F F F F F 8:2 FTCA

minor pathway -HF -HF

F F F F F F F F F O F F F F F F F F O F F F F F F F F O F O- S-CoA F F F F F F F F F F F F F F F F F F F F F F F 8:2 FTAL 8:2 FTUCA

F F F F F F F O F S-CoA F OH F F F F F F F F F F F F F F F F OH F F F F F F F F F F F F F F O O 8:2 FTOH F S-CoA F F F F F F F F

F F F F F F O F O C O- + F S-CoA F F F F F F F

PFOA

Figure 1.4. 8:2 FTOH biodegradation in mixed microbial system as proposed by Dinglasan et al. [33]. Structures in brackets were not determined in the study. Figure adapted from Dinglasan et al. [33]. 21

F F F F F F F F COOH O O F OH F F F F F F F F OH OH F F F F F F F F F F F F F F F F F F F F F F F F O F F F PFNA 8:2 FTOH-Gluc - OH O O F F F PFHpA F F F F F F F F F F F F F F F F F F F F F PFPeA 8:2 FTOH 8:2 FTAL 8:2 FTCA F F F F F F F O F O S O- F O F F F F F F F F F F F F F F F F F F F F F F O F F 8:2 FTOH-Sulf O O- PFHxA I F F II F F F F F F F F F F F F F F 8:2 FTUAL 8:2 FTUCA

8:2 uFTOH-GSH 8:2 FTUAL-GSH 8:2 FTUCA-GSH

I II I II

F F F F F F F F F F F F O OH OH O F F F F F F F F F F F F F F F F O O- O- F O F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F 7:3 FTUAL 7:3 FTUCA (7:3 β-OH FTUAL) (7:3 β-OH FTUCA)

O O F F F F F F O O F F F F F F F F F F F F F F F F F F F F F F - O O- O O F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F (7:3 β-keto acid) 7:3 FTAL 7:3 FTCA 7:3 β-keto aldehyde

F F F F F F OH F F F F F F F O F F F F F F F F O F F F F F F F 7:3 FTCA-TA F - C O F 7:2 sFTOH F F F F F F F F F F F F F F F 7:2 ketone PFOA 7:2 sFTOH-Gluc

Figure 1.5. 8:2 FTOH biodegradation in rats as proposed by Fasano et al. [44]. Structures labeled in parenthesis are hypothesized intermediates and were not observed in the experiment. Note that glutathione conjugates can undergo further metabolism not presented in this figure. Adapted from Fasano et al. [44]. 22

23 1.4.2.2.2. Polyfluorinated Sulfonamide Compounds There has been comparatively less attention concerning the biotransformation of perfluorinated sulfonamide compounds. Studies are limited to the ethyl-substituted perfluorooctane sulfonamide and sulfonamide ethanol and have primarily been performed using rat models with one exception that utilized rainbow trout liver microsomes. As a result, there have been only two studies [45, 46] that have examined the biotransformation mechanism of polyfluorinated sulfonamide compounds in detail. These studies have shown the polyfluorinated sulfonamide biotransformation will yield PFOS but not PFOA or lower chain- length PFCAs.

1.4.2.2.2.1. N-ethyl perfluorooctane sulfonamide (N-EtFOSA) Arrendale et al. [47] analyzed whole blood from Sprague-Dawley rats and dogs that been orally or intravenously dosed with N-EtFOSA, commonly known as the insecticide, Sulfuramid. It was shown that the N-EtFOSA was rapidly biotransformed in both animals, with PFOSA detected within 1 hour post-treatment. PFOSA obtained higher blood concentrations than N- EtFOSA.

Grossman et al. [48] examined the biotransformation of N-EtFOSA in male Sprague- Dawley rats that were continuously exposed via the diet for 56 days. The parent compound was not detected in any tissue or blood samples, indicating rapid biotransformation. The suspected metabolite PFOSA was detected in all tissues and blood on the first sampling point and throughout the experiment. PFOSA concentrations at the conclusion of the uptake phase were greatest in fat, followed by lung and liver. The PFOSA levels were at steady-state throughout the experiment. These results may be suggestive of the rapid elimination of PFOSA or the fast biotransformation of PFOSA to PFOS, although PFOS was not monitored during the experiment.

Tomy et al. [49] investigated the biotransformation of N-EtFOSA in rainbow trout liver microsomes. Incubations were analyzed for loss of the parent compound as well as formation of PFOS, PFOA and PFOSA. The overall depletion of N-EtFOSA did not follow first-order reaction kinetics but showed two distinct features, an initial fast depletion followed by a second much slower depletion. PFOS and PFOSA levels were shown to increase throughout the experiment. The overall increasing PFOSA levels observed by Tomy et al. [49] may suggest a

24 low biotransformation rate of PFOSA to PFOS in the rainbow trout microsomes. PFOA was also detected in the incubation mixtures, but the levels did not increase during the experiment and were attributed to background contamination. Interestingly, a consistent mass balance throughout the experiment, suggesting that PFOSA and PFOS were the only metabolites formed during N-EtFOSA biotransformation. Xu et al. [45, 50] showed that PFOSA can react to form the N-glucuronide conjugate in human, rat, dog and monkey liver tissues. The results of Tomy et al. [49] may suggest that these pathway is not important in rainbow trout, or may highlight differences between metabolic pathways between microsomes and more complex tissues. Three biotransformation pathways were proposed by Tomy et al. [49]. First, the direct deethylamination of N-EtFOSA followed by oxidation of the sulfone to sulfonate. Second, the deethylation of N-EtPFOSA to form PFOSA followed by the deamination to yield PFOS. Third, the direct hydrolysis of N-EtFOSA to form PFOS.

1.4.2.2.2.2. N-ethyl perfluorooctane sulfonamide ethanol (N-EtFOSE)

Xu et al. [45] investigated the biotransformation pathway of N-EtFOSE and identified metabolites in rat liver microsomes, cytosol and slices. Results from the experiments were compiled to elucidate a biotransformation pathway for N-EtFOSE (Figure 1.6). The N-EtFOSE was biotransformed via N-deethylation in rat liver microsomes to yield the FOSE alcohol which was subsequently dealkylated to form PFOSA. It was shown that P450 enzymes were involved in the two dealkylation reactions. The biotransformation of the FOSE alcohol to PFOSA was faster than that of N-EtFOSE to the FOSE alcohol, perhaps indicating a rate-limiting step in N- EtFOSE metabolism [45]. The formation of N-EtFOSA from N-EtFOSE was not empirically observed, but it was shown that N-EtFOSA was rapidly biotransformed to PFOSA, consistent with previous studies [47-49]. The N-EtFOSE and FOSE alcohol were shown to form O- glucuronide conjugates, and PFOSA was shown to form the N-glucuronide conjugate. Also, it was shown that the N-EtFOSE and FOSE alcohol could be oxidized to form the N-EtFOSAA and FOSAA which were not further biotransformed. PFOSA was biotransformed to PFOS in the rat liver slices, but at a low rate. However, PFOS formation from PFOSA was not observed in the microsomal, cytosolic or S9 fractions. It was suggested that the degradation of PFOSA to PFOS could be via abiotic hydrolysis and not enzyme mediated. These observations appear to contradict the findings of Tomy et al. [49] and warrant further investigation.

25 Rhoads et al. [46] examined the aerobic biodegradation of N-EtFOSE in activated sludge. To investigate the mechanism of biotransformation, separate incubation studies were performed with previously reported N-EtFOSE metabolites, including N-EtFOSAA, N-EtFOSA, PFOSAA, PFOSA and PFOSi. PFOS was detected as the terminal product in the incubations of N- EtFOSE and potential intermediates. PFOA was not detected in any of the incubation treatments. Based on the observed products of the incubation experiments, a biotransformation pathway of N-EtFOSE was proposed. It was postulated that the N-EtFOSE is initially oxidized to N-EtFOSAA which is dealkylated to yield N-EtFOSA. This is contrary to the mechanism propose by Xu et al. [45] who showed that the N-EtFOSAA is not biotransformed in rat liver. PFOSA is formed either through the direct dealkylation of N-EtFOSA or indirectly through oxidation of N-EtFOSA to PFOSAA and subsequent dealkylation. PFOSA undergoes deamination to yield the PFOSi which is oxidized to from PFOS.

1.4.2.2.2.3. Biotransformation of Polyfluorinated Sulfonamide Compounds: Summary and Conclusions The biotransformation of polyfluorinated sulfonamide compounds is apparently much less complex than fluorotelomer compounds, presumably as the result of fewer branching points in the overall mechanism. PFOS is formed, although it is unclear whether the final degradation step (PFOSA > PFOS) is enzyme mediated or the result of abiotic hydrolysis. Interestingly, no degradation of the fluorinated tail is observed. These results indicate that sulfonamide group serves as a barrier to further degradation. Further, PFOA is not formed during biotransformation, indicating that degradation of the sulfonate functional group does not occur.

F F F F F F F F O F S N O-Glucuronide F O F F F F F F F OH O-Glucuronide N-EtFOSE X ?

F F F F F F F F O F F F F F F F F O F F F F F F F F O F F F S N OH S NH S NH X F O F O OH F O F F F F F F F F F F F F F F F F F F F F F O FOSE Alcohol N-EtFOSA N-EtFOSAA

F F F F F F F F O F F F F F F F F O ? F F F F F F F F O F OH F F - S NH X S NH2 S O F O O F O F O F F F F F F F F F F F F F F F F F F F F F PFOSAA PFOSA PFOS

Major Pathways: Observed Pathways: Unconfirmed ? Observed only in Pathways: Not Detected: X liver slices:

Figure 1.6. Proposed biotransformation pathways of N-EtFOSE in rat liver microsomes, cytosol and slices. Figure adapted from Xu et al. [45]. 26

27 1.5. Bioaccumulation of organic compounds

1.5.1. Definitions Knowledge of bioaccumulation is necessary to assess overall chemical risk. Since toxic effects are exerted on a cellular level, uptake of a chemical from the environment into the organism (i.e. bioaccumulation) is required. The literature contains many different and sometimes contradictory definitions of bioaccumulation. In this thesis, the widely accepted nomenclature, as defined by Gobas and Morrison [51], has been adopted.

Bioaccumulation in aquatic species refers to the process in which an organism achieves a chemical concentration greater than that of the water when all routes of exposure are considered (uptake via the gills, through the skin or from the diet). It is distinguished from bioconcentration in that it considers dietary exposure. The bioaccumulation factor (BAF) is defined as the ratio of chemical concentration in the organism (CB) to the concentration in the water. The water chemical concentration may be defined as the total chemical concentration

(CWT) or the freely dissolved chemical concentration (CWD), however the use of CWD is generally preferred since this represents the biologically available fraction. The BAF is expressed as,

CB BAF = WT or CC WD

Bioconcentration in aquatic organisms refers to the process by which the chemical concentration in the organism exceeds that in the water when considering uptake from the water only, via the gills and/or through the skin. Similar to the BAF, the bioconcentration factor (BCF) is expressed as,

CB BCF = CWT or CWD

Biomagnification occurs when the chemical concentration in the organism exceeds that of the diet (CA). By definition, this process considers dietary uptake only. The biomagnification factor (BMF) is defined as,

CB BMF = CA

28 1.5.2.Contaminant Uptake and Elimination in Fish A simple one-compartment model for contaminant uptake and elimination in aquatic organism was described by Arnot & Gobas [52] (Figure 1.7). In this model, the steady-state concentration in the fish (CB) is described by

CB = [ k1 · (mO · φ · CWT,O + mP · CWD,S) + kD · Σ Pi · CD,i ]

/ (k2 + kE + kG + kM)

Where CB = the chemical concentration in the fish, k1 = gill uptake rate constant, mO = fraction of the respiratory ventilation that involves overlying water, φ = fraction of total chemical concentration in overlying water that is freely dissolved and can be absorbed via membrane diffusion, CWT,O = total chemical concentration in water column above sediments, mP

= fraction of respiratory ventilation that involves sediment-associated pore water, CWD,S = freely dissolved chemical concentration in sediment pore water, kD = dietary uptake rate constant, Pi = fraction of diet consisting of prey item, CD,i = concentration of chemical in prey item, k2 = gill elimination rate constant, kM = metabolic transformation rate constant, kE = fecal egestion rate constant, kG = growth dilution rate constant.

k1 kM

k kD G

k2 kE Figure 1.7. Conceptual diagram showing major routes of contaminant uptake and elimination in fish. Terms defined in text. Figure adapted from Arnot & Gobas [52].

The simple model by Arnot & Gobas [52] serves to illustrate the relevance of metabolism (kM) in determining the magnitude of contaminant bioaccumulation, and is a focus of this thesis. Although the model described above is based on aquatic species, metabolism similarly influences the bioaccumulation of all organisms. Overall, the bioaccumulation of metabolically-labile compounds is not frequently investigated in whole-body fish experiments.

29 In the model by Arnot & Gobas [52], only nonmetabolizable chemicals were considered since it is was stated that empirically-derived kM data was not sufficiently available. Further, it has been stated by Cowan-Ellsberry et al. [53] that reliable and accepted methods for estimating biotransformation is fish do not exist. As will be reviewed later in this chapter, the consideration of metabolism in determining xenobiotic bioaccumulation is a very active area of current research. Of relevance to poly- and perfluorinated compounds, the PFCAs and PFSAs are known to be biologically persistent and not subject to biotransformation. However, as reviewed earlier in this chapter, several polyfluorinated compounds can act as precursors to PFCAs and PFSAs through biotransformation.

1.5.3. Xenobiotic Metabolism in Fish The general outcome of xenobiotic metabolism is to convert a bioaccumulative, hydrophobic compound to a molecule that is more polar and thus more readily excretable. However, there are cases in which the metabolites are equally or more bioaccumulative than the parent compound. Examples include oxychlordane [54], dieldrin [55], DDT [55] and benz(a)acridine [56]. As illustrated in the simple model of Arnot & Gobas [52], metabolism represents an elimination pathway. As such, compounds that are metabolizable show lower BAF/BCF values as compared with non-metabalizable compounds of similar hydrophobicity [57]. Fish are capable of a wide variety of biotransformation reactions which are similar in scope to mammals, although it has been shown that the overall enzyme activities are lower [58- 60]. The types of biotransformation reactions have been summarized by several authors [61-65] and thus are only briefly described here. Biotransformation is comprised of two types of reactions, termed phase I and phase II metabolism. Phase I consists of reactions that either reveal or introduce polar functional groups to the molecule. These may include oxidation, N- and O-dealkylation, hydrolysis and reduction reactions. Perhaps the most common phase I reaction is oxidation that is mediated by cytochrome P450 monooxygenase enzymes. Phase II metabolism comprises conjugation reactions with endogenous molecules in the body forming a compound that is more polar than the original chemical. Conjugation substrates include glucuronic acid, sulfonates, glutathione and taurine. Phase I metabolism typically precede phase II, although phase II reactions can occur directly on the parent compound.

30 1.5.3.1. Extrahepatic Metabolism in Fish In fish, as with mammals, it is assumed that the liver is the primary site for biotransformation. However, as noted by Nichols et al. [66], extrahepatic metabolism has been demonstrated in the gut [67-69] and gills [70]. The few studies available have shown that activities of monooxygenase enzymes and glucuronidase in the intestine are lower than those of the liver [67, 71]. Specifically, in rainbow trout it was shown that 7-ethoxycoumarin-O- dethylase (EROD) activity in the intestine was ~10-fold lower than in liver, whereas, glucuronidation activity was ~60% lower in the intestine [72]. Studies of esterase activities in fish extrahepatic tissues are very limited but do show comparable activity between the liver, stomach and intestine [73-75]. Further, Barron et al. [76] showed, using 4-nitrophenyl acetate as the substrate, that rainbow trout had somewhat higher carboxylesterase activity in the sera as compared to the liver. Esterase activity was observed in gill microsomes, although the activity was ~10-fold lower than measured in serum [76]. The potential for extrahepatic metabolism is important since it represents an additional elimination process that may not be captured in models that assume hepatic metabolism only. This may be particularly relevant for compounds such as esters that apparently show widespread activity throughout the body.

1.5.4. Quantifying Metabolism in Fish Metabolism rates and the influence of metabolism on the BAF have been investigated in vivo through the use of chemical inhibitors of biotransformation. Sijm et al. [77] investigated the bioconcentration of several chlorinated dioxins and furans in rainbow trout with the presence and absence of piperonyl butoxide (PBO), a broad P450 enzyme inhibitor. PBO-exposed fish had slower elimination rate constants for T4CDD and P5CDD, presumably due to the inhibition of metabolism. Biotransformation was estimated to comprise 50% and 36% of the total elimination rate constants for T4CDD and P5CDD, respectively. Lu et al. [78] showed that the BCF of benzo(a)pyrene in PBO treated mosquito fish was higher (BCF=22) as compared to the control fish (benzo(a)pyrene was not detected in control fish tissues). Barron et al. [79] used bis-(p-nitrophenyl)phosphate (BNPP) to inhibit the biotransformation of triclopyr BEE in coho salmon (water borne exposure). It was shown that metabolism was inhibited by approximately 86%. Karara & Hayton [80] showed that levels of di-2-ethylhexyl phthalate (DEHP) in sheepshead minnow were nearly 2-fold greater when treated with BNPP. Further, the levels of metabolites formed in the BNPP-treated fish was 23% of that formed in the control. Stehly & Hayton [81] investigated the bioconcentration of pentachlorophenol (PCP) in rainbow in the

31 presence and absence of salicylamide, an inhibitor of glucuronidation. The empirically calculated BCF of the salicylamide-inhibited fish increased nearly 50% from 193 to 278, relative to the controls.

Van der Linde [82] quantified KM as the difference between the empirically derived overall elimination rate and the model-estimated elimination rates in absence of biotransformation (depuration to water, feces excretion and growth dilution). This approach was also recently used by Arnot et al. [83].

Weisbrod et al. [84] have recently reviewed the in vitro methods for quantifying biotransformation in fish. These techniques include perfused tissues, tissue slices, primary cell cultures (e.g. hepatocytes) and subcellular fractions (e.g. cytosol and microsomes). Perfused livers have been used to study xenobiotic biotransformation in fish and are considered to be the closest representation to whole animal but also the most technically demanding [84]. Isolated cells from liver, intestine and gill have also been used to study fish metabolism. Whole cell isolates reflect both the biotransformation as well as uptake and elimination processes. However, these are also difficult to prepare and maintain viable, and generally cannot be frozen for future use [84]. Subcellular fractions (cytosol and microsomal fractions) are technically the easiest to prepare, maintain and they will hold their enzyme activity for a long period (>1 yr) if stored at -80oC. The cytosol fraction (more commonly known as the S9 fraction) is obtained by centrifuging the crude tissue homogenate at 9,000 to 12,000 g for approximately 20 minutes, the S9 fraction is supernatant [84]. Further purification of the S9, by centrifuging at 100,000 g, will yield the microsomal fraction. The microsomal fraction is primarily comprised of the endoplasmic reticulum and has been shown to be enriched in phase I enzymes, whereas, the S9 fraction is generally enriched in the phase II enzymes [84]. It has been stated that the S9 and microsomal fractions are the optimal in vitro test system for studying metabolic pathways and screening level biotransformation assessments [84]. This is due to their ease of use and efficient use of animal resources. As such, the majority of fish biotransformation rates have been generated using S9 and microsomal fractions [85]. As discussed later in the chapter, an active area of research is the incorporation of in vitro metabolism data into bioaccumulation models.

32 1.5.5. Quantifying Bioaccumulation 1.5.5.1. Laboratory-based Experiments Several standardized methods exist for the quantification of bioaccumulation in fish. These include protocols from the Organization for Economic Cooperation and Development (OECD 305: Bioconcentration – Flow-through fish test), the American Society for Testing and Materials (ASTM E1022-94: Standard guide for conducting bioconcentration tests with fishes and saltwater bivalve mollusks) and United States EPA (OPPTS 850.1730: Fish BCF). Note that these methods all consider bioconcentration and that presently no standardized methods exist for bioaccumulation. During laboratory BCF experiments, the fish are exposed to the test compound for approximately 30 days, typically under flow-through conditions. The fish are then subjected to a depuration phase in which they are exposed to clean water. During the uptake and depuration phases, fish and water are periodically collected and analyzed. The uptake (k1) and depuration (k2) rate constants are determined, and the BCF is calculated as k1/k2.

As noted by Weisbrod et al. [84], the OECD 305 test takes about 4 months to perform and requires a minimum of 108 fish. Further, each test costs approximately $125,000 due to the extensive analytical requirements and/or the synthesis of radiolabeled parent compounds. There is a need to develop alternative methods.

1.5.5.2. Models to Predict Bioaccumulation As mentioned, empirically evaluating the BAF/BCF for the many thousands of commercially-used chemicals is a process that is expensive, with regards to both monetary and animal resources, and is very time consuming. Therefore, predictive measures are attractive, particularly for new chemicals in which very little environmental information exists. There are two general methods for modeling and ultimately predicating bioaccumulation potential, empirical correlation models and mechanistic models. These models have been reviewed by several authors including Mackay and Fraser [86], Arnot and Gobas [52] and more recently Barber [87].

1.5.5.2.1. Empirical Correlation Models The first approach employs the relationship between a series of an empirically derived

BCF values and a measure of hydrophobicity (i.e. octanol-water partition coefficient (KOW)). Such relationships are generally expressed as

33

log BCF = a log KOW + b where a and b are empirically determined constants. Therefore, using the experimentally- derived relationship, the KOW of a chemical is used predict its bioconcentration potential. For example, Neely et al. [88] and Veith et al. [89] showed that the BCF could be predicted by the

KOW. The relationship was later extended to a larger data set by Mackay [90]. It was demonstrated that these linear log BCF – log KOW relationships resulted in good predictions of the BCF with log KOW values ranging from 1-6. However, it was shown that the relationship breaks down for chemicals that readily metabolized [79, 91]. Further, it was shown that for

“superhydrophobic” compounds, those with log KOW values > 6 [92], log BCF – log KOW relationship begins to show a negative slope, presumably as a result of reduced bioavailability, with the overall shape resembling that of a “horse shoe”. Several researchers have used polynomial equations to better describe the log BCF – log KOW relationship for the wider range of log KOW values [93, 94]. A major criticism of these models is that they are based are a relatively small training set, with no attempt incorporate variances due to molecular structure

[95]. Meylan et al. [95] investigated log BCF – log KOW relationships with a much larger data set than previously examined (694 chemicals). Improved correlations were developed by grouping the data set into log KOW ranges as well as group by structural features such as ionic compounds and compounds with long alkyl chains. A number of additional developments have been proposed in recent years. Dearden & Shinnawei [96] incorporated aqueous solubility, polarity, polarizability, hydrogen bond donor ability and molecular size. Jackson et al. [97] developed a quantitative structure-activity relationship based on molecular structure to predict the BCF for a series of pesticides. Fu et al. [98] recently developed improved an BCF relationship for ionic organic compounds.

1.5.5.2.2. Mechanistic Models Mechanistic-based models which include clearance-volume (CVPK) and physiologically based pharmacokinetic models (PBPK) [86]. Clearance-volume models (also known as compartment models) divide the fish into one or more compartments, with each compartment comprising tissues or organs that have similar kinetic properties. These models are advantageous in that limited data is required as compared to PBPK models. However, these models are limited in their ability to extrapolate between fish species [99].

34 The basic design and characteristics of PBPK models are described by Gobas & Morrison [51] and a schematic for a PBPK model developed by Nichols et al. [66] is shown in Figure 1.8. In PBPK models, the fish is separated into sections representing the different organs and tissues. The chemical is distributed throughout the body via the blood, and exchanged to the organs and tissues by equilibrium partitioning. Elimination from each compartment is through equilibrium partitioning, organ specific elimination (i.e urine excretion from the kidney, respiratory via the gills), biotransformation (i.e. in the liver and gut) and fecal excretion. PBPK are advantageous since they can assess concentrations in different tissues, which is important for readily metabolized chemicals or scenarios of short exposure [51]. However, PBPK models often are data intensive, requiring inputs such as metabolic transformation rates, chemical partitioning between blood and tissues and physiological data [51].

35

Gills Q Q W Effective W

CEXP Respiratory Volume CINSP Q C QC Cardiac Output C C VEN ART

QFAT Fat Tissue Group C C VFAT ART

QCARC Carcass Tissue Group C C VCARC ART

K , V M MAX

QLIV Liver C VLIV CART C QBILE CBILE VGUT

KM, VMAX QGUT Gut Tissue CART QDIET QDIGESTA Gut Lumen C C DIET DIGESTA Gut

Figure 1.8. Schematic representation of a physiologically based bioaccumulation model for fish. Figure adapted from Nichols et al. [66] to include gut metabolism. QLIV, QCARC, QFAT and QGUT are blood flows to the liver, carcass, fat and gut as a proportion of the cardiac output, QC. QW is rate of flow in water that exchanges with blood in gills, CINSP in the chemical concentration in the water. CVLIV, CVCARC, CVFAT and CVGUT are chemical concentrations in the venous blood leaving these tissues. CVEN and CART are chemical concentrations in the mixed venous and arterial blood. QDIET and QDIGESTA are bulk flow rates for food and feces. CDIET and CDIGESTA are chemical concentrations in the food and feces. KM and VMAX are kinetic rate and capacity parameters in the liver and stomach .

36 1.5.5.2.2.1. Extrapolating In Vitro Metabolic Data to Predict In Vivo Bioaccumulation The majority of research concerning xenobiotic biotransformation in fish has been carried out using in vitro systems with the emphasis on characterizing reaction mechanisms [100]. Nichols et al. [100] recently reviewed methods to extrapolate in vitro liver biotransformation data to predict whole body bioaccumulation in fish. As noted by Nichols et al. [100], this area of research is considerably advanced for mammals. However, until recently, there has been little effort to quantitatively relate in vivo biotransformation data to in vivo whole body rates in fish. The few examples include studies by Law et al. [101], Schultz and Hayton [102], Cowan-Ellsberry et al. [53] and Dyer et al. [103]. At present, studies extrapolating actual in vitro biotransformation rates to whole body bioaccumulation are limited to the consideration of liver biotransformation only. Therefore, there is a need for additional models that consider extrahepatic metabolism.

Two PBTK models that incorporated gut metabolism were developed Nichols et al. [66, 104]. However, these studies did not directly extrapolate in vitro parameters and used hypothetical gut metabolism rates only. Nichols et al. [104] developed a PBTK dietary exposure model that segmented the gastrointestinal tract into four sections (stomach, pyloric ceca, upper intestine and lower intestine) and incorporated gut metabolism. The model was developed using previously determined fish physiological parameters [105] and data obtained from a previous in vivo dosing study with PCB 52 [106]. Using hypothetic metabolic loss rates, it was shown by incorporating gut metabolism the whole body PCB 52 half-life was reduced from 220 days from 700 day (liver metabolism only).

As well, Nichols et al. [66] modeled the influence of liver metabolism on bioaccumulation as a function of log KOW. Hypothetical in vitro biotransformation rates were used to extrapolate to in vivo whole-body metabolism rates in both a one-compartment “clearance” and a PBTK model. Simulations obtained using the one-compartment model are shown in Figure 1.9. As expected, hepatic metabolism results in an overall decrease in BAF, although the magnitude of reduction was not consistent across all log KOW values. Further, it was shown that metabolism was not significant for chemicals with log KOW < ~3, presumably because depuration dominants the elimination kinetics [107]. Nichols et al. [66] also incorporated gut metabolism into their model simulations. As expected the influence of this

37 additional loss mechanism was to further reduce the BCF as compared to liver metabolism only, although this was not quantitatively described.

Figure 1.9. The influence of liver metabolism on the steady-state bioaccumulation as predicted with a one-compartment model. Simulations obtained using in vitro intrinsic clearance values of 0.0 (1), 0.1 (2), 1.0 (3), 10.0 (4) and 100 (5) µl/min.mg protein. The bioaccumulate factor is normalized to the whole-body lipid content. Figure taken from Nichols et al. [66].

38 1.6. Literature Cited

1. Kissa, E., Fluorinated Surfactants and Repellents. 2nd ed.; Marcel Dekker: New York, 2001.

2. Olsen, G. W.; Huang, H.-Y.; Helzlsouer, K. J.; Hansen, K. J.; Butenhoff, J. L.; Mandel, J. H., Historical comparison of perfluorooctanesulfonate, perfluorooctanoate, and other fluorochemicals in human blood. Environ. Health Perspect. 2005, 113, 539-545.

3. Houde, M.; Martin, J. W.; Letcher, R. J.; Solomon, K.; Muir, D. C. G., Biological monitoring of polyfluoroalkyl substances: A review. Environ. Sci. Technol. 2006, 40, 3463- 3473.

4. Yamashita, N.; Taniyasu, S.; Petrick, G.; Wei, S.; Gamo, T.; Lam, P. K. S.; Kannan, K., Perfluorinated acids as novel chemical tracers of global circulation of ocean waters. Chemosphere 2008, 70, 1247-1255.

5. Martin, J. W.; Smithwick, M. M.; Braune, B. M.; Hoekstra, P. F.; Muir, D. C. G.; Mabury, S. A., Identification of long-chain perfluorinated acids in biota from the Canadian Arctic. Environ. Sci. Technol. 2004, 38, 373-380.

6. Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H., Sources, fate and transport of perfluorocarboxylates. Environ. Sci. Technol. 2006, 40, 32-44.

7. 3M Company. 1999. Fluorochemical use, distribution and release overview. US EPA Docket: AR226-0550.

8. Paul, A. G.; Jones, K. C.; Sweetman, A. J., A first global production, emission, and environmental inventory for perfluorooctane sulfonate. Environ. Sci. Technol. 2009, 43, 386- 392.

9. Smithwick, M.; Norstrom, R. J.; Mabury, S. A.; Solomon, K.; Evans, T. J.; Stirling, I.; Taylor, M. K.; Muir, D. C. G., Temporal trends of perfluoroalkyl contaminants in polar bears (Ursus maritimus) from two locations in the North American arctic, 1972-2002. Environ. Sci. Technol. 2006, 40, 1139-1143.

10. Brezinski, D., Laying the foundation for new technologies: 3M creates a new building block for its fluorosurfactants. Paint & Coatings Industry January, 2003.

11. Ahrens, L.; Felizeter, S.; Sturm, R.; Xie, Z.; Ebinghaus, R., Polyfluorinated compounds in waste water treatment plant effluents and surface waters along the River Elbe, Germany. Mar. Pollut. Bull. 2009, 58, 1326-1333.

12. Bao, J.; Jin, Y.; Liu, W.; Ran, X.; Zhang, Z., Perfluorinated compounds in sediments from the Daliao River system of northeast China. Chemosphere 2009, 77, 652-657.

13. Kato, K.; Calafat, A. M.; Needham, L. L., Polyfluoroalkyl chemicals in house dust. Environ. Res. 2009, 109, 518-523.

39 14. Olsen, G. W.; Chang, S.-C.; Noker, P. E.; Gorman, G. S.; Ehresman, D. J.; Lieder, P. H.; Butenhoff, J. L., A comparison of the pharmacokinetics of perfluorobutanesulfonate (PFBS) in rats, monkeys, and humans. Toxicology 2009, 256, 65-74.

15. Russell, M. H.; Berti, W. R.; Szostek, B.; Buck, R. C., Investigation of the biodegradation potential of a fluoroacrylate polymer product in aerobic soils. Environ. Sci. Technol. 2008, 42, 800-807.

16. Dinglasan-Panlilio, M. J. A.; Mabury, S. A., Significant residual fluorinated alcohols present in various fluorinated materials. Environ. Sci. Technol. 2006, 40, 1447-1453.

17. Gauthier, S. A.; Mabury, S. A., Aqueous photolysis of 8:2 fluorotelomer alcohol. Environ. Toxicol. Chem. 2005, 24, 1837-1846.

18. Young, C. J.; Mabury, S. A., Atmospheric perfluorinated acid precursors: Chemistry, occurrence and impacts. Rev. Environ. Contam. Toxicol. submitted.

19. Ellis, D. A.; Martin, J. W.; De Silva, A. O.; Mabury, S. A.; Hurley, M. D.; Sulbaek Andersen, M. P.; Wallington, T. J., Degradation of fluorotelomer alcohols: A likely atmospheric source of perfluorinated carboxylic acids. Environ. Sci. Technol. 2004, 38, 3316-3321.

20. Hurley, M. D.; Wallington, T. J.; Sulbaek Andersen, M. P.; Ellis, D. A.; Martin, J. W.; Mabury, S. A., Atmospheric chemistry of fluorinated alcohols: reaction with Cl atoms and OH radicals and atmospheric lifetimes. J. Phys. Chem. A 2004, 108, 1973-1979.

21. Vésine, E.; Bossoutrot, V.; Mellouki, A.; Le Bras, G.; Wenger, J.; Sidebottom, H., Kinetic and mechanistic study of OH- and Cl-initiated oxidation of two unsaturated HFCs: C4F9CH=CH2 and C6F13CH=CH2. J. Phys. Chem. A 2000, 104, 8512-8520.

22. Nakayama, T.; Takahashi, K.; Matsumi, Y.; Toft, A.; Sulbaek Andersen, M. P.; Nielsen, O. J.; Waterland, R. L.; Buck, R. C.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of CF3CH=CH2 and C4F9CH=CH2: Products of the gas-phase reactions with Cl atoms and OH radicals. J. Phys. Chem. A 2007, 111, 909-915.

23. Young, C. J.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A., Atmospheric chemistry of 4:2 fluorotelomer iodide (n-C4F9CH2CH2I) and 6:2 fluorotelomer iodide (n-C6F13CH2CH2I): Kinetics and products of photolysis and reaction with OH radicals and Cl atoms. J. Phys. Chem. A 2008, 112, 13542-13548.

24. D'eon, J. C.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A., Atmospheric chemistry of N-methyl perfluorobutane sulfonamidoethanol, C4F9SO2N(CH3)CH2CH2OH: Kinetics and mechanism of reaction with OH. Environ. Sci. Technol. 2006, 40, 1862-1868.

25. Martin, J. W.; Ellis, D. A.; Mabury, S. A.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of perfluoroalkanesulfonamides: Kinetic and product studies of the OH radical and Cl atom initiated oxidation of N-ethyl perfluorobutanesulfonamide. Environ. Sci. Technol. 2006, 40, 846-872.

26. Dreyer, A.; Temme, C.; Sturm, R.; Ebinghaus, R., Optimized method avoiding solvent- induced response enhancement in the analysis of volatile and semi-volatile polyfluorinated

40 alkylated compounds using gas chromatrography-mass spectrometry. J. Chromatogr. A 2008, 1178, 199-205.

27. Oono, S.; Matsubara, E.; Harada, K. H.; Takagi, S.; Hamada, S.; Asakawa, A.; Inoue, K.; Watanabe, I.; Koizumi, A., Survey of airborne polyfluorinated telomers in Keihan area, Japan. Bull. Environ. Contam. Toxicol. 2008, 80, 102-106.

28. Piekarz, A. M.; Primbs, T.; Field, J. A.; Barofsky, D. F.; Simonich, S., Semivolatile fluorinated organic compounds in Asian and western U.S. air masses. Environ. Sci. Technol. 2007, 41, 8248-8255.

29. Dreyer, A.; Weinberg, I.; Temme, C.; Ebinghaus, R., Polyfluorinated compounds in the atmosphere of the Atlantic and South oceans: Evidence for a global distribution. Environ. Sci. Technol. 2009, 43, 6507-6514.

30. Dreyer, A.; Matthias, V.; Temme, C.; Ebinghaus, R., Annual time series of air concentrations of polyfluorinated compounds. Environ. Sci. Technol. 2009, 43, 4029-4036.

31. Young, C. J.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A., Atmospheric chemistry of 4:2 fluorotelomer iodide (n-C4F9CH2CH2I): Kinetics and products of photolysis and reaction with OH radicals and Cl atoms. J. Phys. Chem. A 2008, 112, 13542-13548.

32. Hagen, D. F.; Belisle, J.; Johnson, J. D.; Venkateswarlu, P., Characterization of fluorinated metabolites by a gas chromatographic-helium microwave plasma detector - the biotransformation of 1H,1H,2H,2H-perfluorodecanol to perfluorooctanoate. Anal. Biochem. 1981, 118, 336-343.

33. Dinglasan, M. J. A.; Ye, Y.; Edwards, E. A.; Mabury, S. A., Fluorotelomer alcohol biodegradation yields poly- and perfluorinated acids. Environ. Sci. Technol. 2004, 38, 2857- 2864.

34. D'eon, J. C.; Mabury, S. A., Production of perfluorinated carboxylic acids (PFCAs) from the biotransformation of polyfluoroalkyl phosphate surfactants (PAPs): Exploring routes of human contamination. Environ. Sci. Technol. 2007, 41, 4799-4805.

35. Nabb, D. L.; Szostek, B.; Himmelstein, M. W.; Mawn, M. P.; Gargas, M. L.; Sweeney, L. M.; Stadler, J. C.; Buck, R. C.; Fasano, W. J., In vitro metabolism of 8-2 fluorotelomer alcohol: Interspecies comparisons and metabolic pathway refinement. Toxicol. Sci. 2007, 100, 333-344.

36. Wang, N.; Szostek, B.; Folsom, P. W.; Sulecki, L. M.; Capka, V.; Buck, R. C.; Berti, W. R.; Gannon, J. T., Aerobic biotransformation of 14C-labeled 8-2 telomer B alcohol by activated sludge from a domestic sewage treatment plant. Environ. Sci. Technol. 2005, 39, 531-538.

37. Wang, N.; Szostek, B.; Buck, R. C.; Folsom, P. W.; Sulecki, L. M.; Capka, V.; Berti, W. R.; Gannon, J. T., Fluorotelomer alcohol biodegradation - direct evidence that perfluorinated carbon chains breakdown. Environ. Sci. Technol. 2005, 39, 7516-7528.

41 38. Wang, N.; Szostek, B.; Buck, R. C.; Folsom, P. W.; Sulecki, L. M.; Gannon, J. T., 8-2 Fluorotelomer alcohol aerobic soil biodegradation: Pathways, metabolites, and metabolite yields. Chemosphere 2009, 75, 1089-1096.

39. Liu, J.; Lee, L. S.; Nies, L. F.; Nakatsu, C. H.; Turco, R. F., Biotransformation of 8:2 fluorotelomer alcohol in soil and by soil bacteria isolates. Environ. Sci. Technol. 2007, 41, 8024- 8030.

40. Martin, J. W.; Mabury, S. A.; O'Brien, P. J., Metabolic products and pathways of fluorotelomer alcohols in isolated rat hepatocytes. Chem. Biol. Interact. 2005, 155, 165-180.

41. Kudo, N.; Iwase, Y.; Okayachi, H.; Yamakawa, Y.; Kawashima, Y., Induction of hepatic peroxisome proliferation by 8-2 telomer alcohol feeding in mice: Formation of perfluorooctanoic acid in the liver. Toxicol. Sci. 2005, 86, 231-238.

42. Fasano, W. J.; Carpenter, S. C.; Gannon, S. A.; Snow, T. A.; Stadler, J. C.; Kennedy, G. L.; Buck, R. C.; Korzeniowski, S. H.; Hinderliter, P. M.; Kemper, R. A., Absorption, distribution, metabolism, and elimination of 8-2 fluorotelomer alcohol in the rat. Toxicol. Sci. 2006, 91, 341-355.

43. Henderson, W. M.; Smith, M. A., Perfluorooctanoic acid and perfluorononanoic acid in fetal and neonatal mice following in utero exposure to 8-2 fluorotelomer alcohol. Toxicol. Sci. 2007, 95, 452-461.

44. Fasano, W. J.; Sweeney, L. M.; Mawn, M. P.; Nabb, D. L.; Szostek, B.; Buck, R. C.; Gargas, M. L., Kinetics of 8-2 fluorotelomer alcohol and its metabolites, and liver glutathione status following daily oral dosing for 45 days in male and female rats. Chem. Biol. Interact. 2009, 180, 281-295.

45. Xu, L.; Krenitsky, D. M.; Seacat, A. M.; Butenhoff, J. L.; Anders, M. W., Biotransformation of N-ethyl-N(2-hydroxyethyl)perfluorooctanesulfonamide by rat liver microsomes, cytosol, and slices and by expressed rat and human cytochromes P450. Chem. Res. Toxicol. 2004, 17, 767-775.

46. Rhoads, K. R.; Janssen, E. M.-L.; Luthy, R. G.; Criddle, C. S., Aerobic biotransformation and fate of N-ethyl perfluorooctane sulfonamidoethanol (N-EtFOSE) in activated sludge. Environ. Sci. Technol. 2008, 42, 2873-2878.

47. Arrendale, R. F.; Stewart, J. T.; Manning, R.; Vitayavirasuk, B., Determination of GX- 071 and its major metabolite in rat blood by cold on-column injection capillary GC/ECD. J. Agric. Food Chem. 1989, 37, 1130-1135.

48. Grossman, M. R.; Mispagel, M. E.; Bowen, J. M., Distribution and tissue elimination in rats during and after prolonged dietary exposure to a highly fluorinated sulfonamide pesticide. J. Agric. Food Chem. 1992, 40, 2505-2509.

49. Tomy, G. T.; Tittlemier, S. A.; Palace, V. P.; Budakowski, W. R.; Braekevelt, E.; Brinkworth, L.; Friesen, K., Biotransformation of N-ethyl perfluorooctanesulfonamide by rainbow trout (Onchorhynchus mykiss) liver microsomes. Environ. Sci. Technol. 2004, 38, 758- 762.

42 50. Xu, L.; Krenitsky, D. M.; Seacat, A. M.; Butenhoff, J. L.; Tephly, T. R.; Anders, M. W., N-glucuronidation of perfluorooctanesulfonamide by human, rat, dog, and monkey liver microsomes and by expressed rat and human UDP-glucuronosyltransferases. Drug Metab. Dispos. 2006, 34, 1406-1410.

51. Gobas, F. A. P. C.; Morrison, H. A., Bioconcentration and biomagnification in the aquatic environment. In Handbook of Property Estimation Methods for Chemicals, Boethling, R. S.; Mackay, D., Eds. CRC Press: Boca Raton, FL, 2000; pp 189-231.

52. Arnot, J. A.; Gobas, F. A. P. C., A food web bioaccumulation model for organic chemicals in aquatic ecosystems. Environ. Toxicol. Chem. 2004, 23, 2343-2355.

53. Cowan-Ellsberry, C. E.; Dyer, S.; Erhardt, S.; Bernhard, M. J.; Roe, A. L.; Dowty, M. E.; Weisbrod, A. V., Approach for extrapolating in vitro metabolism data to refine bioconcentration factor estimates. Chemosphere 2008, 70, 1804-1817.

54. Kenaga, E. E., Correlation of bioconcentration factors of chemicals in aquatic and terrestrial organisms with their physical and chemical properties. Environ. Sci. Technol. 1980, 14, 553-556.

55. Mackay, D.; Shui, W. Y.; Ma, K.-C.; Lee, S. C., Handbook of physical-chemical properties and environmental fate for organic chemicals, Second Edition. In CRC Press: Boca Raton, FL, 2006; Vol. 18. Insecticides.

56. Southworth, G. R.; Keffer, C. C.; Beauchamp, J. J., The accumulation and disposition of benz(a)acridine in the fathead minnow, Pimephales promelas. Arch. Environ. Contam. Toxicol. 1981, 10, 561-569.

57. Oliver, B. G.; Niimi, A. J., Bioconcentration factors of some halogenated organics for rainbow trout: Limitation in their use for prediction of environmental residues. Environ. Sci. Technol. 1985, 19, 842-849.

58. Han, X.; Nabb, D. L.; Yang, C.-H.; Snajdr, S. I.; Mingoia, R. T., Liver microsomes and S9 from rainbow trout (Oncorhynchus mykiss): Comparison of basal-level enzyme activities with rat and determination of xenobiotic intrinsic clearance in support of bioaccumulation assessment. Environ. Toxicol. Chem. 2009, 28, 481-488.

59. Han, X.; Nabb, D. L.; Mingoia, R. T.; Yang, C.-H., Determination of xenobiotic intrinsic clearance in freshly isolated hepatocytes from rainbow trout (Oncorhynchus mykiss) and rat and its application in bioaccumulation assessment. Environ. Sci. Technol. 2007, 41, 3269-3276.

60. Nabb, D. L.; Mingoia, R. T.; Yang, C.-H.; Han, X., Comparison of basal level metabolic enzyme activities of freshly isolated hepatocytes from rainbow trout (Oncorhynchus mykiss) and rat. Aquat. Toxicol. 2006, 80, 52-59.

61. Lech, J. J.; Bend, J. R., Relationship between biotransformation and the toxicity and fate of xenobiotic chemicals in fish. Environ. Health Perspect. 1980, 34, 115-131.

62. Cravedi, J. P., Role of biotranformation in the fate and toxicity of chemicals: consequences for the assessment of residues in fish. Revue Méd. Vét. 2002, 153, 419-424.

43 63. Kleinow, K. M.; Melancon, M. J.; Lech, J. J., Biotransformation and induction: Implications for toxicity, bioaccumulation and monitoring of environmental xenobiotics in fish. Environ. Health Perspect. 1987, 71, 105-119.

64. Sijm, D. T. H. M.; Opperhuizen, A., Biotransformation of organic chemicals by fish: enzyme activities and reactions. In The Handbook of Environmental Chemistry. Reactions and Processes, Part E, vol. 2, Hutzinger, O., Ed. Springer: Berlin, 1989.

65. Binder, R. L.; Melancon, M. J.; Lech, J. J., Factors influencing the persistence and metabolism of chemicals in fish. Drug Metab. Dispos. 1984, 15, 697-724.

66. Nichols, J. W.; Fitzsimmons, P. N.; Burkhard, L. P., In vitro-in vivo extrapolation of quantitative hepatic biotransformation data for fish. II. Modeled effects on chemical bioaccumulation. Environ. Toxicol. Chem. 2007, 26, 1304-1319.

67. Kleinow, K. M.; James, M. O.; Tong, Z.; Venugopalan, C. S., Bioavailability and biotransformation of benzo(a)pyrene in an isolated perfused in situ catfish intestinal preparation. Environ. Health Perspect. 1998, 106, 155-166.

68. Van Veld, P. A., Absorption and metabolism of dietary xenobiotics by the intestine of fish. Rev. Aquat. Sci. 1990, 2, 185-203.

69. Van Veld, P. A.; Patton, J. S.; Lee, R. F., Effect of preexposure to dietary benzo[a]pyrene (BP) on the first-pass metabolism of BP by the intestine of toadfish (Opsanus tau): In vivo studies using portal vein-catheterized fish. Toxicol. Appl. Pharmacol. 1988, 92, 255-265.

70. Barron, M. G.; Schultz, I. R.; Hayton, W. L., Presystemic branchial metabolism limits di-2-ethylhexyl phthalate accumulation in fish. Toxicol. Appl. Pharmacol. 1989, 98, 49-57.

71. James, M. O.; Altman, A. H.; Morris, K.; Kleinow, K. M.; Tong, Z., Dietary modulation of phase 1 and phase 2 activities with benzo(a)pyrene and related compounds in the intestine but not the liver of the channel catfish, Ictalurus punctatus. Drug Metab. Dispos. 1997, 25, 346-354.

72. Lindström-Seppä, P.; Koivusaari, U.; Hänninen, O., Extrahepatic xenobiotic metabolism in North-European freshwater fish. Comp. Biochem. Physiol. 1981, 69C, 259-263.

73. Li, S.-N.; Fan, D.-F., Activity of esterases from different tissues of freshwater fish and responses of their isoenzymes to inhibitors. J. Toxicol. Environ. Health 1997, 51, 149-157.

74. Hirji, K. N.; Courtney, W. A. M., Non-specific carboxylic esterase activity in the digestive tract of the perch, Perca fluviatilis L. J. Fish Biol. 1983, 22, 1-7.

75. Salamastrakis, S. S.; Haritos, A. A., Physicochemical characterization and tissue distribution of multiple molecular forms of fish (Trachurus trachurus) esterases. Comp. Biochem. Physiol. 1988, 91B, 741-750.

76. Barron, M. G.; Charron, K. A.; Stott, W. T.; Duvall, S. E., Tissue carboxylesterase activity of rainbow trout. Environ. Toxicol. Chem. 1999, 18, 2506-2511.

44 77. Sijm, D. T. H. M.; Yarechewski, A. L.; Muir, D. C. G.; Webster, G. R. B.; Seinen, W.; Opperhuizen, A., Biotransformation and tissue distribution of 1,2,3,7-tetrachlorodibenzo-p- dioxin, 1,2,3,4,7-pentachlorodibenzo-p-dioxin and 2,3,4,7,8-pentachlorodibenzofuran in rainbow trout. Chemosphere 1990, 21, 845-866.

78. Lu, P. Y.; Metcalf, R. L.; Plummer, N.; Mandel, D., The environmental fate of three carcinogens: Benzo(a)pyrene, benzidine, and vinyl chloride evaluated in laboratory model ecosystems. Arch. Environ. Contam. Toxicol. 1977, 6, 129-142.

79. Barron, M. G.; Mayes, M. A.; Murphy, P. G.; Nolan, R. J., Pharmacokinetics and metabolism of trichlopyr butoxyehtyl ester in coho salmon. Aquat. Toxicol. 1990, 16, 19-32.

80. Karara, A. H.; Hayton, W. L., Metabolic inhibition and di-2-ethylhexyl phthalate pharacokinetics in fish. Drug Metab. Dispos. 1988, 16, 146-150.

81. Stehly, G. R.; Hayton, W. L., Disposition of pentachlorophenol in rainbow trout (Salmo gairdneri): Effect of inhibition of metabolism. Aquat. Toxicol. 1989, 14, 131-147.

82. Van der Linde, A.; Jan Hendriks, A.; Sijm, D. T. H. M., Estimating biotransformation rate constants of organic chemicals from modeled and measured elimination rates. Chemosphere 2001, 44, 423-435.

83. Arnot, J. A.; Mackay, D.; Bonnell, M., Estimating metabolic biotransformation rates in fish from laboratory data. Environ. Toxicol. Chem. 2008, 27, 341-351.

84. Weisbrod, A. V.; Sahi, J.; Segner, H.; James, M. O.; Nichols, J.; Schultz, I.; Erhardt, S.; Cowan-Ellsberry, C.; Bonnell, M.; Hoeger, B., The state of in vitro science for use in bioaccumulation assessments for fish. Environ. Toxicol. Chem. 2009, 28, 86-96.

85. Fitzsimmons, P. N.; Lien, G. J.; Nichols, J. W., A compilation of in vitro rate and affinity values for xenobiotic biotransformation in fish, measured under physiological conditions. Comp. Biochem. Physiol. 2007, 145C, 485-506.

86. Mackay, D.; Fraser, A., Bioaccumulation of persistent organic chemicals: Mechanisms and models. Environ. Pollut. 2000, 110, 385-391.

87. Barber, M. C., Dietary uptake models used for modeling the bioaccumulation of organic contaminants in fish. Environ. Toxicol. Chem. 2008, 27, 755-777.

88. Neely, W. B.; Branson, D. R.; Blau, G. E., Partition coefficient to measure bioconcentration potential of organic chemicals in fish. Environ. Sci. Technol. 1974, 8, 1113- 1115.

89. Veith, G. D.; DeFoe, D. L.; Bergstedt, B. V., Measuring and estimating the bioconcentration factor of chemicals in fish. J. Fish Res. Board Canada 1979, 36, 1040-1048.

90. Mackay, D., Correlation of bioconcentration factors. Environ. Sci. Technol. 1982, 16, 274-278.

45 91. de Wolf, W.; de Bruijn, J. H. M.; Seinen, W.; Hermens, J. L. M., Influence of biotransformation on the relationship between bioconcentration factors and octanol-water partition coefficients. Environ. Sci. Technol. 1992, 26, 1197-1201.

92. Gobas, F. A. P. C.; Clark, K. E.; Shui, W. Y.; Mackay, D., Bioconcentration of polybrominated benzenes and biphenyls and related superhydrobic chemicals in fish: Role of bioavailability and elimination into the feces. Environ. Toxicol. Chem. 1989, 8, 231-245.

93. Connell, D. W.; Hawker, D. W., Use of polynomial expressions to describe the bioconcentration of hydrophobic chemicals by fish. Ecotoxicol. Environ. Saf. 1988, 16, 242-257.

94. Bintein, S.; Devillers, J.; Karcher, W., Nonlinear dependence of fish bioconcentration on n-octanol/water partition coefficient. SAR QSAR Environ. Res. 1993, 1, 29-39.

95. Meylan, W. M.; Howard, P. H.; Boethling, R. S.; Aronson, D.; Printup, H.; Gouchie, S., Improved method for estimating bioconcentration/bioaccumulation factor from octanol/water partition coefficient. Environ. Toxicol. Chem. 1999, 18, 664-672.

96. Dearden, J. C.; Sinnawei, N. M., Improved prediction of fish bioconcentration factor of hydrophobic chemicals. SAR QSAR Environ. Res. 2004, 15, 449-455.

97. Jackson, S. H.; Cowan-Ellsberry, C. E.; Thomas, G., Use of quantitative structural analysis to predict fish bioconcentration factors for pesticides. J. Agric. Food Chem. 2009, 57, 958-967.

98. Fu, W.; Franco, A.; Trapp, S., Methods for estimating the bioconcentration factor of ionizable organic chemicals. Environ. Toxicol. Chem. 2009, 28, 1372-1379.

99. Kleinow, K. M.; Nichols, J. W.; Hayton, W. L.; McKim, J. M.; Barron, M. G., Toxicokinetics in Fishes. In The Toxicology of Fishes, Di Giulio, R. T.; Hinton, D. E., Eds. CRC Press: Boca Raton, 2008.

100. Nichols, J. W.; Schultz, I. R.; Fitzsimmons, P. N., In vitro-in vivo extrapolation of quantitative hepatic biotransformation data for fish. I. A review of methods, and strategies for incorporating intrinsic clearance estimates into chemical kinetic models. Aquat. Toxicol. 2006, 78, 74-90.

101. Law, F. C. P.; Abenini, S.; Kennedy, C. J., A biologically based toxicokinetic model for pyrene in rainbow trout. Toxicol. Appl. Pharmacol. 1991, 110, 390-402.

102. Schultz, I. R.; Hayton, W. L., Interspecies scaling of the bioaccumulation of lipophilic xenobiotics in fish: An example using trifluralin. Environ. Toxicol. Chem. 1999, 18, 1440-1449.

103. Dyer, S.; Bernhard, M. J.; Cowan-Ellsberry, C.; Perdu-Durand, E.; Demmerle, S.; Cravedi, J.-P., In vitro biotransformation of surfactants in fish. Part II - Alcohol ethoxylate (C16EO8) and alcohol ethoxylate sulfate (C14EO2S) to estimate bioconcentration potential. Chemosphere 2009, 76, 989-998.

46 104. Nichols, J. W.; Fitzsimmons, P. N.; Whiteman, F. W., A physiologically based toxicokinetic model for dietary uptake of hydrophobic organic compounds by fish. II. Simulation of chronic exposure scenarios. Toxicol. Sci. 2004, 77, 219-229.

105. Nichols, J. W.; McKim, J. M.; Andersen, M. E.; Gargas, M. L.; Clewell III, H. J.; Erickson, R. J., A physiologically based toxicokinetic model for the uptake and disposition of waterborne organic chemicals in fish. Toxicol. Appl. Pharmacol. 1990, 106, 433-447.

106. Nichols, J. W.; Fitzsimmons, P. N.; Whiteman, F. W.; Dawson, T. D., A physiologically based toxicokinetic model for dietary uptake of hydrophobic organic compounds in fish. I. Feeding studies with 2,2'5,5'-tetrachlorobiphenyl. Toxicol. Sci. 2004, 77, 206-218.

107. Gobas, F. A. P. C.; Opperhuizen, A.; Hutzinger, O., Bioconcentration of hydrophobic chemicals in fish: Relationship with membrane permeation. Environ. Toxicol. Chem. 1986, 5, 637-646.

CHAPTER TWO

Levels and Trends of Poly- and Perfluorinated Compounds in the Arctic Environment

Craig M. Butt, Urs Berger, Rossana Bossi and Gregg T. Tomy

Submitted to: The Science of the Total Environment (accepted pending minor revisions)

Contributions: Prepared by Craig Butt with exception of the “Introduction” (Rossana Bossi) and “Spatial Trends” (Gregg Tomy) sections. Critical comments were provided by Urs Berger, Rossana Bossi and Gregg Tomy.

This chapter is condensed from the original manuscript. Specifically, the introduction has been edited to reduce redundancy with Chapter 1 and sections describing PFC levels in individual wildlife species and isomer trends have been removed.

47 48 Chapter Two – Levels and Trends of Poly- and Perfluorinated Compounds in the Arctic Environment

2.1. Abstract

Poly- and perfluorinated organic compounds (PFCs) are ubiquitous in the Arctic environment. Several modeling studies have been conducted in attempt to resolve the dominant transport pathway of PFCs to the arctic – atmospheric transport of precursors versus direct transport via ocean currents. These studies are generally limited by their focus on perfluorooctanoate (PFOA) fluxes to arctic seawater and thus far have only used fluorotelomer alcohols (FTOHs) and sulfonamide alcohols as inputs for volatile precursors. There have been many monitoring studies from the North American and European Arctic, however, almost nothing is known about PFC levels from the Russian Arctic. The bulk of the monitoring efforts in biological samples have focused on the perfluorinated carboxylates (PFCAs) and sulfonates (PFSAs), although there are very few measurements of PFC precursors. The marine food web has been well studied, particularly the top predators. In contrast, freshwater and terrestrial ecosystems have been poorly studied. Studies show that in wildlife perfluorooctane sulfonate (PFOS) is generally measured in the highest concentration, followed by either perfluorononanoate (PFNA) or perfluoroundecanoate (PFUnA). However, some whale species show relatively high levels of perfluorooctane sulfonamide (PFOSA) and seabirds are typically characterized by high proportions of the C11-C15 PFCAs. PFOA is generally infrequently detected and is present in low concentrations in arctic biota. Food web studies show high bioaccumulation in the upper trophic-level animals, although the mechanism of PFC biomagnification is not understood. Spatial trend studies show some differences between populations, although there are inconsistencies between PFC trends. The majority of temporal trend studies are from the Northern American arctic and Greenland. Studies show generally increasing levels of PFCs from the 1970s, although some studies from the Canadian arctic show recent declines in PFOS levels. In contrast, ringed seals and polar bears from Greenland continue to show increasing PFOS concentrations. The inconsistent temporal trends between regions may be representative of differences in emissions from source regions. In general, there are few measurements of PFCs from the abiotic environment. Atmospheric measurements show the widespread occurrence of PFC precursors, FTOHs and perfluorinated sulfonamide alcohols. Further, PFCAs and PFSAs have been detected on atmospheric particles. The detection of PFCAs and PFSAs in snow deposition is consistent with the volatile precursor transport

49 hypothesis. There are limited seawater measurements of PFCs. PFOA is generally detected in the greatest levels. Additional seawater measurements are needed to validate existing model predications.

50 2.2. Introduction The sources of PFCs to the arctic are not well understood and the proposed transport pathways will be discussed in detail in this review. PFCs may be released into the environment by direct discharge (“direct emissions”) from the production of fluorochemicals and disposal of products containing fluorochemicals. In addition, the degradation of “precursor compounds” such as FTOHs and PFOSF-based chemicals has been identified as an “indirect source” of PFCs to the environment. The sources of some PFCAs to the environment and emission estimates have been reviewed by Prevedourous et al. [1]. A list of fluorinated compounds that may degrade to PFCAs has been recently compiled by OECD [2].

There is clear evidence that many PFSAs and PFCAs are globally distributed. This has prompted regulations on the production and uses of several PFCs by national and international regulatory agencies, such as U.S. EPA, Environment Canada and European Union. PFOS has recently been added by the OSPAR Commission to the list of chemicals for priority action, has been included by the Stockholm Convention as a candidate persistent organic pollutant (POPs) and is on the list of ‘new contaminants’ being monitored by the Arctic Monitoring and Assessment Program (AMAP). The detection of some PFCs in human blood from arctic regions [3, 4] has also raised concerns about their potential toxicity. Several studies have been published on the toxicological effects of PFCs. A recent review by Lau et al. [5] summarizes the advances in understanding the toxicological mode of action of PFCs.

The potential for long range transport of PFSAs and PFCAs, their tendency to bioaccumulate and to induce toxic effects are characteristics of POPs. Unlike legacy POPs - which accumulate in lipid rich tissues - PFSAs and PFCAs bind to blood proteins and accumulate mainly in the liver, kidneys and bile secretions [6]. Similar to legacy POPs, PFSAs and PFCAs are transported to remote regions such as the circumpolar Arctic.

In general, the effects of PFCs on wildlife are not known, in particular for Arctic biota. In a recent study, Sonne et al. [7] investigated the potential impact from exposure to PFSAs and PFCAs on liver lesions in East Greenland polar bears (Ursus maritimus). An assessment of the effects of POPs, including PFCs, on Arctic wildlife is presented in this issue [8].

51 Since 2001 there has been considerable progress in the assessment of the environmental levels and potential transport pathways of PFCs to arctic regions. The intent of this review is to provide a state of the science summary of the PFC monitoring data in the arctic environment. The scope of this review will include the PFCAs and PFSAs, as well as their known precursor compounds. Specific areas of discussion include the current understanding of PFC transport pathways, overall levels and, spatial and temporal trends in the biotic and abiotic environments, as well as a discussion of data-gaps and future research needs.

2.3. Transport Pathways

The source of PFCs to the Arctic environment is complex and has been the subject of considerable scientific interest. While two major potential transport mechanisms of PFCs to the arctic have been postulated, the relative contribution of each pathway remains unresolved (Figure 2.1). One pathway involves the transport of volatile precursors via the atmosphere, degradation by atmospheric oxidation to PFCAs and PFSAs and subsequent wet and dry deposition. Volatile precursors include the FTOHs which have been shown to degrade to PFCAs [9, 10] and the perfluorinated sulfonamide alcohols (FOSEs) which have been shown to degrade to PFCAs and PFSAs [11, 12]. In addition, recent research has shown that the fluorotelomer olefins [13, 14], iodides [15] and acrylates [16] may form PFCAs via atmospheric oxidation. The second pathway involves the transport of directly emitted PFCAs and PFSAs via oceanic currents to the Arctic marine environment [17, 18]. PFCAs and PFSAs may be emitted during fluorochemical manufacturing processes and as residuals in consumer products. In addition, they have been intentionally added in some products (e.g. aqueous film forming foams,

AFFF). Although there is some uncertainty as to the actual pKa values, PFCAs and PFSAs are expected to be found primarily as anions [19-21] in aquatic environments, particularly ocean waters. As such, they will be relatively water soluble and amenable to oceanic transport. In addition, PFCAs and PFSAs are extremely persistent under ambient environmental conditions.

Local inputs may be another important source of PFCs to arctic regions. For example, Stock et al. [22] detected comparatively elevated levels of perfluorohexane sulfonate (PFHxS), PFOS, perfluoroheptanoate (PFHpA) and PFOA in the water and sediment of Resolute Lake near Resolute Bay on Cornwallis Island, Nunavut, Canada. The authors attributed the elevated PFC levels to AFFF contamination and sewage runoff from the local airport. Although Resolute

52 Lake flows directly into the Barrow Strait, ringed seals (Phoca hispida) from the region did not show comparatively elevated PFC levels [23]. Therefore, it is unclear if local PFC sources will significantly influence PFC concentrations in the arctic regional marine environment.

Figure 2.1. Major transport pathways of PFCs to the Arctic.

Several models have been developed to quantify the relevant contribution of the “indirect” and “direct” transport pathways. A brief description of each model is presented, followed by a discussion of supporting evidence for each pathway.

Wallington et al. [24] estimated the formation of PFCAs from the atmospheric oxidation of 8:2 FTOH using the IMPACT model. Although the focus of the study was PFOA, the formation of other PFCAs (PFNA and lower chain-length PFCAs) was also investigated. The global emission of 8:2 FTOH was estimated as 1000 tonnes per year, the upper range predicted in order to maintain the observed atmospheric FTOH concentrations [25]. The model predicted that 8:2 FTOH would be globally distributed, consistent with the measured half-life of approximately 20 days [25]. Arctic air concentrations were predicted to be 5-fold lower than from the southern source regions. PFOA was also predicted to be ubiquitously formed in the atmosphere, but with concentrations frequently higher in remote regions (such as the Arctic) due to the presence of low NOx concentrations. The molar yield of PFOA from 8:2 FTOH

53 degradation was estimated to be 3-6% in the Northern Hemisphere and similar yields of PFNA were also predicted. When integrated over the latitudes of 65-90oN, the estimated PFOA deposition flux from 8:2 FTOH was 0.4 tonnes per year. Young et al. [26] showed that this deposition flux was similar to that extrapolated from surface snow measurements in the Canadian Arctic.

Armitage et al. [17] examined the PFOA flux from direct emissions to the Arctic via oceanic currents using the Globo-POP model. PFOA emission estimates between 1950-2004 were taken from Prevedouros et al. [1]. In addition, projected PFOA emission estimates (2005- 2050) were assumed based on recent emission reductions and from reduction commitments made by industry. The model predicted surface ocean water concentrations to range from 25-90 pg/L in the Northern Polar zone (arctic region) during 2005. The authors noted that these results agreed well with ocean measurements from the Greenland Sea [27]. The estimated net flux of PFOA to the Northern Polar zone was 8-23 tonnes per year. The authors note that this flux is 20 to 60-fold greater than that predicted from FTOH degradation by Wallington et al. [24]. It was suggested that these modeling results support the study hypothesis that direct emissions are responsible for the PFOA burden in the arctic surface water. The model predicted doubling times of approximately 7.5-10 years for PFOA surface water concentrations between 1975 and 2004. It was noted that these doubling times agreed well with those observed in arctic wildlife [28]. Despite the estimated downturn in direct PFOA emissions in the early 2000s, arctic seawater levels were predicted to increase until about 2030 and then gradually decline (Figure 2.2).

54

Figure 2.2. Modeled PFOA concentrations in ocean water from northern hemisphere for period 1950-2050. Vertical bars represent annual emissions, solid line represents model concentrations in the Northern Temperate zone, dotted line represents model concentrations in the Northern Polar zone (arctic region). Reprinted with permission from Armitage et al. [17].

Wania [18] extended the work of Armitage et al. [17] by using the Globo-POP model to compare PFOA fluxes to the arctic surface waters (defined as the “Northern Polar zone” in the model) resulting from both the direct transport via ocean currents as well as that generated by the atmospheric oxidation of FTOHs. In addition, the Arctic Contamination Potential (ACP) was calculated as an estimate of the relative efficiency of each transport pathway. FTOH production values from the mid-1970s to 2005 were taken from Prevedourous et al. [1, 29] [30]and the DuPont Global PFOA Strategy (2005)[31]. The model assumed 2% of FTOH production was emitted to the atmosphere. The 2005 FTOH emission rate was 200 tonnes per year which was considerably lower than that used by Wallington et al. [24]. PFCA yields from FTOH atmospheric oxidation were assumed to be between 3-10% as estimated from Wallington et al. [24]. Direct emissions of PFOA were taken from Armitage et al. [17]. The cumulative emission rate for FTOHs was approximately 1500 tonnes from 1974 to 2005, whereas, the cumulative PFOA emission rate from direct releases between 1950 to 2005 was 5000 tonnes. The model predicted that, after 10 years of continuous release, the ACP of directly emitted PFOA is about 16-fold greater than that of PFOA generated by the atmospheric degradation of FTOHs. This suggests that oceanic transport is a more efficient process to deliver PFOA to the arctic surface waters. The difference was largely due to the relatively low yield of PFCAs from

55 FTOH atmospheric oxidation. The model results predicted the flux to the Arctic by directly emitted PFOA to be 9-20 tonnes/year for the years 2000-2005. In comparison, the deposition flux of all PFCAs generated from FTOH atmospheric degradation was predicted to be 110 kg/yr in 2005. This suggests that oceanic transport is about 1-2 orders of magnitude more significant than atmospheric deposition for PFOA. Further, it was noted that PFCA flux from FTOH degradation was 5 to 10-fold less than predicted for PFOA deposition alone by Wallington et al. [24], which is consistent with lower FTOH emission estimates. Similar to Armitage et al. [17], it was shown that model predictions for arctic seawater concentrations resulting from direct PFOA emissions were consistent with measured levels from the Greenland Sea [27]. In contrast, arctic seawater concentrations resulting solely from FTOH degradation were about 2 orders of magnitude lower. It was suggested that this implies the quantity of PFCAs generated from FTOH atmospheric degradation is too low to explain measured seawater levels. Interestingly, the model predicted that the bulk of PFCAs generated from FTOH degradation would be deposited in the mid-latitude Northern hemisphere oceans and subsequently transported to the Arctic via ocean currents. Therefore, a lag in arctic seawater levels was predicted, similar to Armitage et al. [17], in response to reduced direct PFOA and FTOH emissions.

Schenker et al. [32] estimated the PFOA deposition to the Arctic derived from the degradation of FTOHs and perfluorooctyl sulfonamide ethanols between 1998 and 2005 using the CliMoChem model. PFOA emission estimates were taken from Prevedouros et al. [1] whereas precursor emission estimates were derived from information presented by Prevedouros et al. [1] and Wania [18]. The model explicitly included the volatile precursors as well as intermediates. The model assumed a 5% yield of PFOA from FTOHs, as modeled by Wallington et al. [24]. A similar PFOA yield from PFOSF-based alcohols was assumed based on the similarity in the reaction pathways and recent smog-chamber results [11, 12]. The model predicted similar PFOA deposition fluxes from the degradation of FTOHs and PFOSF-based compounds until 2002 (Figure 2.3). After 2000, PFOA deposition from FTOH degradation dominated the overall deposition profile, presumably as a result of the cessation of PFOSF production by the major manufacturer. However, the authors note that recent atmospheric measurements of PFOSF-based compounds were in fact much larger than those predicted by the model. Thus, it is possible that PFOA fluxes derived from PFOSF-based compounds may be under-predicted by the model. Further, the authors note that the actual PFOA yield may be

56 lower than the 5% assumed in the model. It was shown that fluxes from direct PFOA emissions, via the ocean currents, were about 2 orders of magnitude greater than those from precursor degradation. However, it was shown that the deposition fluxes from the model agreed well with those derived by ice core measurements from the Canadian Arctic [26].

Figure 2.3. Modeled PFOA deposition fluxes (solid lines and bands) to the Arctic (65oN to 90oN) resulting from FTOH atmospheric degradation (red) and FOSE atmospheric degradation (blue). Crosses indicate results from other models (red) and fluxes extrapolated from surface snow measurements (black). Reprinted with permission from Schenker et al. [32].

The modelling studies published to date have primarily focused on PFOA fluxes to the arctic seawater. It is important to note that PFOA is generally only infrequently detected, and at low levels, in arctic wildlife. In fact, the PFC profiles in arctic biota are typically dominated by PFOS and the long-chain PFCAs (ie. PFNA, perfluoroundecanoate (PFUnA) or perfluorotridecanoate (PFTrA)). In contrast to PFOA, no or comparatively little direct production of the long-chain PFCAs (eg- perfluorodecanoate (PFDA), perfluorododecanoate (PFDoA)) [1] has been reported. Therefore, there is an immediate need for models that include PFOS and the long-chain PFCAs. In addition, other volatile PFC precursor compounds have been identified and should be included in future modelling endeavours. Perfluorotelomer-based olefins have been shown to degrade to perfluoroaldehydes, and thus will form PFCAs, via atmospheric oxidation [13, 14]. In addition, fluorotelomer-based iodides [15] and acrylates [16] have recently been shown to form PFCAs by atmospheric oxidation. Further, the fluorotelomer- based phosphates (PAPs), which are used in commercial products for paper treatment and floor

57 waxes (Zonyl RP and Masurf FS-115 technical information)[33, 34], are non-volatile precursors that have been shown to metabolise to PFCAs [35]. However, considering the labile nature of PAPs and potential for sorption to organic matter, it is unclear whether PAPs would be transported to the Arctic via ocean currents.

There is considerable empirical support for the “indirect” transport pathway. FTOHs, FOSEs and FOSAs have sufficient atmospheric lifetimes to undergo long-range transport [11, 12, 25] and have been detected in the arctic atmosphere [22, 36]. As well, intermediate (8:2 and 10:2 fluorotelomer unsaturated carboxylate (FTUCA)) and terminal degradation products (PFOA and long-chain PFCAs) have been detected on atmospheric particles in the Arctic [22]. PFOS and PFCAs have been detected in snow cores from remote ice caps in the Canadian Arctic [26]. Deposition fluxes extrapolated from surface snow measurements [26] are consistent with those from models [24, 32]. Further, PFCs have been detected in surface water [22], sediment [22] and fish [37] from arctic lakes that are primarily influenced by atmospheric deposition. It was noteworthy that ratios of PFOA:PFNA and PFDA:PFUnA in the lake water and sediment were consistent with those measured from the ice cap snow [22, 26]. Further, it has been suggested that the PFC doubling times observed in arctic wildlife, as well as the apparent rapid reduction in PFOS levels in some species [38, 39], is too short to be explained by oceanic transport [28] which show long delays (~30 years) in response to emission changes. Finally, the predominance of certain PFCs, in arctic wildlife, that are known to have insignificant levels of direct production (PFOS and some long-chain PFCAs) is suggestive of volatile precursors as the source of these PFCs. It should be noted that PFC profiles in wildlife are not directly comparable to those measured in sources, such as seawater or industrial emissions. Wildlife profiles are complicated by compound specific bioaccumulation potentials that may vary between species.

Support for the “direct” transport pathway largely comes from results obtained from global transport models that indicate the yield of PFOA from FTOH atmospheric degradation is insufficient to account for arctic seawater concentrations. Further, it is noted that predicted arctic seawater levels of PFOA are consistent with measured values. There have been some efforts to characterize the spatial distribution of PFCs in Canadian Arctic and sub-arctic seawater [40]. This study, in addition to the study of PFCs in the Greenland Sea [27] and Labrador Sea [41], represent the only measurements on PFCs in Arctic seawaters. Further

58 measurements of Arctic seawater levels, from various regions in the circumpolar Arctic are needed to confirm model predictions. As well, the models are ultimately sensitive to emission levels [17, 18] and estimated yields of PFCAs from precursors.

It has recently been suggested that PFCAs may be transported to the Arctic via the gas- phase in their protonated form [42]. The authors noted that while the PFOA anion is known to have a negligible vapour pressure and Henry’s Law constant, these properties are appreciable in the protonated PFOA. Also, it was suggested that the formation of marine aerosols, such as during wave breaking, enhances the formation of gas-phase PFOA. Obviously, the fraction of PFCAs that will be in their protonated form, and thus subject to volatilization, under ambient environmental conditions is a function of the pKa value. The relevance of this transport pathway is unclear since there is uncertainty regarding their pKa values [20, 21, 42]. A full discussion regarding the PFCA pKa values is beyond the scope of this review. However, in a recent study by Cheng et al. [43] it was shown that in a negative electrospray ionization mass spectrometer, the normalized molecular ion ratios of both PFOS and PFOA were independent of pH under the range studied (pH 1-6). These findings confirm the lack of formation of the protonated species under these pH conditions. Since the pKa of PFOS is not disputed, and is considerably <1, it was concluded that the pKa of PFOA is also <1. The authors note that in ocean water at pH ~8.1, the ratio of protonated-PFOA to unprotonated-PFOA should be well below 10-7. These findings provide support for the hypothesis that PFCAs will predominately be in the anionic form in the aqueous environment.

There are limited empirical reports of gas-phase PFCAs in the atmosphere. Kim &

Kannan [44] reported levels of C7-C12 PFCAs, PFHxS, PFOS and PFOSA in the gas-phase from Albany, New York. In fact, it was shown that the gas-phase levels of some PFCs exceeded that in the particle phase. However, traditional experimental techniques may not be appropriate for the collection of gas-phase PFCAs and PFSAs. Arp & Goss [45] suggest that gas-phase PFCAs may irreversibility bind to the glass-fiber and quartz-fiber media typically used as particulate filters in high-volume air samplers, thus preventing their collection on gas-phase sorbents. Therefore, the relevance of the atmospheric transport of gas-phase PFCAs as an important pathway to the Arctic remains to be elucidated.

59 2.4. Biotic Measurements 2.4.1. Trends 2.4.1.1. Food Web Studies Eastern Canadian Arctic A marine food web from the eastern Canadian Arctic, was analyzed for PFOS, PFOA, PFOSA and N-EtFOSA [46]. Samples were collected from various locations from 1996-2002, which may confound trend interpretation due to spatial and temporal variation. Stable isotopes of nitrogen were analyzed to assess relative trophic level. In approximate order of trophic level, the food web consisted of clams (Mya truncate, Serripes groenlandica), various zooplankton species, shrimp (Pandalus borealis, Hymenodora glacialis), walrus, arctic cod, redfish, narwhal, beluga whale, black-legged kittiwake and glaucous gull. Whole body homogenate was analyzed in the arctic cod, clams and zooplankton samples, whereas, liver samples were analyzed for the remaining organisms. PFOS and PFOA were detected in low ng g-1 concentrations with PFOS levels generally greater than PFOA. PFOS was shown to biomagnify through the entire food web (Figure 2.4). A significant linear relationship (p<0.0001) was found between ln (natural logarithm) PFOS concentration and trophic level and followed the equation ln PFOS (ng/g wet wt) = -3.285 + (1.14 * trophic level). The PFOS trophic magnification factor was calculated as 3.1 and was generally lower than those calculated for persistent organochlorines. In contrast, PFOA did not biomagnify through the food web as a whole, but did biomagnify between certain individual feeding relationships (e.g. cod to beluga). Interestingly, PFOS precursor compounds, N-EtFOSA and PFOSA were detected in some food web organisms, sometimes at concentrations much greater than PFOS. The potential metabolism of neutral volatile PFOS precursors likely represents a source of PFOS to marine biota especially at higher trophic levels.

60

Figure 2.4. Mean (±1 SE) PFOS concentrations (ng/g wet wt) – trophic level relationship for the eastern Arctic food web. BLKI = black-legged kittiwakes; GLGU = glaucous gulls. Reprinted with permission from Tomy et al. [46].

Western Canadian Arctic (Banks Island) Powley et al. [47] investigated PFCs in a marine food web located near Sachs Harbour on Banks Island in the western Canadian Arctic. Samples were collected in June 2004. The food web samples consisted of 3 species of zooplankton (Calanis hyperboreus, Themisto libellula, Chaetognatha), arctic cod, ringed seal and bearded seal. Whole body samples were analyzed for zooplankton (n=1 homogenate per species) and Arctic cod (n=5), whereas liver, blood and blubber were analyzed in the ringed seal (n=5) and bearded seal samples (n=1). PFC concentrations were dominated by PFOS. PFOS concentrations ranged from ND-0.2 ng/g for zooplankton and 0.3-0.7 ng/g ww for Arctic cod. The rank order of PFOS in tissues of ringed seal were liver (18-34 ng/g ww) >> blood (2.5-8.6 ng/g ww)> blubber (0.4-0.9 ng/g ww). PFOS ringed seal liver levels were similar to those reported from other arctic regions. Although only 1 individual was analyzed, PFOS levels were much lower in the bearded seal as compared to the ringed seals. Similar trends were observed with the PFCAs. This may represent differences in feeding ecology between these seal species, although nitrogen and carbon stable isotopes were not measured in this study to confirm this hypothesis. PFOS concentrations in the bearded seal tissues were ND in blubber, 1.3 ng/g ww in blood and 2.6 ng/g ww in liver. PFDS was detected

61 only in 1 sample (ringed seal liver) at 0.05 ng/g ww which was near the limits of detection. PFDoS was not detected in any sample. PFCAs with less than eight carbons were not detected in any of the food web samples. PFCA profiles were dominated by PFUnA and the usual “odd- even” pattern was observed in which odd-numbered PFCAs had greater levels than adjacent even-numbered PFCAs. The 7:3 fluorotelomer saturated carboxylate (FTCA), a product from 8:2 FTOH degradation, was only observed in ringed seal liver at low ng/g levels. The detection of the 7:3 FTCA is indicative of 8:2 FTOH degradation, via abiotic or biotic mechanisms, as a source of at least some portion of the PFCA body burden.

The PFC patterns varied between trophic levels with PFOS generally becoming enriched, as compared to the PFCAs, with increasing trophic level. For example, PFOS constituted the lowest proportion of PFCs in the zooplankton, but was found in the greatest proportion in ringed seal liver. Biomagnification factors (BMFs) were calculated for Arctic cod/zooplankton and seal blood/Arctic cod. BMFs for Arctic cod/zooplankton were less than 1 for PFDA and PFUnA, but was 8.7 for PFOS. This suggests biomagnification only for PFOS from zooplankton to Arctic cod. For Arctic cod to seal blood, BMFs were greater than 1 with the exception of PFDoA.

Barents Sea (Norwegian Arctic) PFCs were examined in a marine food web from the Barents Sea, located east of Svalbard (77-79 oN, 30oE) [48]. Samples were collected in May-July 2004. The food web consisted of sea ice amphipod (n=6 pools), arctic cod (n=16 pools, consisting of 50 individuals total), black guillemot (n=18) and glaucous gull (n=9). Whole body homogenates of the amphipods were analyzed, whereas, liver was analyzed in arctic cod, black guillemot and glaucous gull. Stable isotopes of nitrogen were determined to assess relative trophic level. Mean trophic levels were 2.0 for ice amphipod, 3.7 for arctic cod, 4.3 for black guillemot and 4.5 for glaucous gull. Trophic levels values were statistically different for all species with the exception of black guillemot and glaucous gull. BMFs were calculated based on trophic level and weighted based on the presumed diet. The authors noted that interpretations of PFC trophic transfer may be complicated by the fact that arctic cod are unlikely to consume ice amphipods, as well as the glaucous gull consumes more prey species than just polar cod and black guillemot. In addition, trend interpretations may be complicated by the migration of the

62 glaucous gull to more industrial areas which presumably have higher PFC levels than the Barents Sea.

PFOS was the dominant PFC analyzed and constituted 52%, 41%, 80% and 91% of total PFCs for ice amphipod, polar cod, black guillemot and glaucous gull, respectively. The mean PFOS concentrations generally increased with trophic level, although PFOS concentrations were statistically similar between ice amphipods and arctic cod (Figure 2.5). Thus, a non-linear relationship was shown between PFOS concentration and trophic level. The relationship between PFOS and trophic level was significantly linear when ice amphipods were excluded. Within species there was no correlation between PFOS concentration and trophic level.

Interestingly, the 6:2 FtS was only detected in the ice amphipod (3 out of 6 pools) and 1 black guillemot individual. The 6:2 FtS may be a precursor to PFHxA [49]. Further, PFOA contributed about 50% of the ΣPFC concentration in the ice amphipod, much higher than the other species. In fact, PFOA was infrequently detected in other species, but was detected in all 6 ice amphipod pools. It was suggested that the presence of the more hydrophilic PFOA and 6:2 FtS in ice amphipods may be due to partitioning from surrounding water.

Trophic level-corrected biomagnification factors > 1 were observed for PFHxS, PFNA, PFOS and ΣPFC in most predator-prey relationships except for polar cod-ice amphipods. The highest BMFs were observed for PFOS.

63

Figure 2.5. Relationship between PFOS concentration (ng/g wet wt) and trophic level, as quantified by δ15N for Barents Sea food web [48]. In the figure legend “polar cod” is identified as “arctic cod” in the text. Arrow indicates one ice amphipod sample was below the range displayed in the figure. Reprinted with permission from Haukäs et al. [48].

Greenland, various locations PFOS, PFOSA, PFOA and PFHxS was analyzed in liver samples of polar bear, minke whale, ringed seal, black guillemot and shorthorn sculpin from various locations around coastal Greenland [50]. Two samples (pools of 5 individuals each) were collected from northeast Greenland (Avanersuaq), west Greenland (Qeqertarsuaq) and east Greenland (Ittoqqortoormiit) from 1998-2002. Stable isotopes of nitrogen were not measured and thus trophic level was not determined. PFOS was the dominant PFC detected with the exception of the minke whale in which PFOSA levels were greater than PFOS. PFOA, PFHxS and PFOSA were not detected or below the LOQ for all samples with the exception of the minke whale and 1 shorthorn sculpin sample. PFOS levels in the east Greenland species showed trends of shorthorn sculpin < ringed seal < polar bear. These results indicate PFOS biomagnification although trophic level was not determined on these samples. West Greenland polar bears were not analyzed.

Food Web Studies: Conclusions In conclusion, studies of PFCs in marine ecosystems have generally shown that there can be trophic-level biomagnification within a food web, especially for PFOS and some long-chain

64 PFCAs. However, there have been limited studies of PFCs in arctic food webs. In fact, only marine food webs have thus far been examined. To date, there have been no studies of PFCs in freshwater or terrestrial food webs. This represents a significant knowledge gap in our understanding of the trophic transfer of PFCs in arctic food webs. Further, the food web studies published to date are generally not spatially or temporally integrated but rather may incorporate samples collected over several years and from varying regions.

It should be noted that positive correlation between trophic position and PFC concentration does not necessarily imply that biomagnification is occurring. There are significant uncertainties regarding the mechanism of bioaccumulation and biomagnification for PFCs. Unlike other “legacy” halogenated organic contaminants (e.g. PCBs and PBDEs), PFCs appear to bind to proteins rather than partition into lipid. As such, PFCs are transported in the body through the blood and preferentially accumulate in protein-rich tissues such as the liver and kidney. Therefore, biomagnification may be related to the quantity and composition of proteins in specific tissues and organs, as well as the protein elimination ability of the organism. At present, PFC “protein normalized” biomagnification factors do not exist and thus comparison of PFC levels among tissues and between species is difficult. Calculation of PFC biomagnification factors using a single tissue (e.g. liver) may be erroneous since this may not accurately represent consumption trends (e.g. polar bears primarily consume the skin and blubber tissue of ringed seals). In addition, there is the potential formation of recalcitrant PFCs (i.e. PFSAs and PFCAs) from the metabolism of precursor compounds within the body. This is further complicated by potentially differing metabolic capabilities between species. Additional research is needed to elucidate the mechanisms of PFC biomagnification.

2.4.1.2. Spatial Studies Bossi et al. [50] assessed the geographic distribution of PFOS, PFOA and PFOSA in biota (fish, birds and mammals) from Greenland and the Faroe Islands. Individual species were pooled according to age and sex to obtain representative samples (n=5) for each of three sampling locations in Greenland. For PFOS, there was a general overall trend of concentrations being greater in animals from east Greenland (Ittoqqortoormiit) than in west Greenland (Qeqertarsuaq). For ringed seal collected in 2002, PFOS concentrations ranged from 52-67 ng/g ww (n=2, all males) from Ittoqqortoormiit, <10-10 ng/g ww (n=2, all males) in animals from Qeqertarsaaq and 27 ng/g ww (n=2, all females) from Avanersuaq. Similar PFOS

65 concentrations were measured in animals collected from Qeqertarsaaq in 2000 (13 ng/g ww). PFOA and PFOSA concentrations were too small to make any meaningful spatial comparisons. In a more recent study, Bossi et al. [51] again examined the spatial distribution of PFOS and PFOSA along with a more expanded suite of PFCAs in juvenile ringed seals from Qeqertarsaaq and Ittoqqortoormiit collected over a 20 year time-span. Consistent with their previous study, PFOS concentrations were greater (ANOVA, p<0.0001) in animals from Ittoqqortoormiit for all the sampling years. A similar trend was shown for PFHxS, PFOSA, PFNA, PFDA and PFUnA in which concentrations were greater in Ittoqqortoormiit animals as compared to Qeqertarsaaq.

In black guillemot collected in 2000 from Ittoqqortoormiit, PFOS concentrations in females (13 ng/g ww) were only slightly smaller than those in males (16 ng/g ww) and similar to females from Qeqertarsauq (14 ng/g ww) [50]. Interestingly, PFOS was not detected in animals collected in 2002. PFOS was also detected in shorthorn sculpin from Ittoqqortoormiit (range: 13-18 ng/g ww, all females) but was undetectable in animals from Qeqertarsauq.

Verreault et al. [52] reported on the spatial trends of sixteen PFCs in whole eggs of herring gulls from two isolated colonies in northern Norway. Thirty samples of freshly laid eggs were randomly collected in 1983, 1993 and 2003 from Hornøya, situated in the north eastern part of northern Norway and Røst (this site consisted of samples from Røst and Hekkingen which were in close proximity to each other and were therefore considered one colony) in the southern part along the west coast. Both sites were thought to represent two distinct population exposure scenarios. Concentrations of PFHxS, PFOA and PFNA in samples from 1993 were greater in the southern most colony; no other statistically significant difference in PFC concentration were observed for the other sampling years. The contribution of the individual PFCA to the total PFCA burden in the eggs were compared between colonies and sampling years. PFOA concentrations were greater in eggs from Røst sampled in 1993 and 2003 than those from Hornøya sampled in the same year. The authors suggest that this might be due to proximal sources of PFOA in the coastal region of northern Norway and/or to enhanced contribution of PFOA to this region due to oceanic transport. There were also greater proportions of PFTeA and PFPeA in eggs collected in 1993 from Røst than Hørnøya.

Löfstrand et al. [53] also used seabird eggs as biomonitors to assess the spatial distribution of PFCs in five locations from West Nordic countries: Vestmannaeyjar (Iceland),

66 Sandøy (the Faroe Islands), Sklinna (Norway), Hjelmsøya (Norway) and Stora Karlsö (Sweden). The site from Sweden was thought to represent a sampling location close to known sources of PFCs but is not actually located in the arctic region as defined by AMAP. In general, different spatial patterns were found for PFOS, PFCAs, PFOSA and N-EtFOSA. The greatest concentration of PFOS was detected in eggs from Sweden (mean: 400 ng/g ww) which was statistically different to concentrations from all the other sites. PFOS concentrations were lowest in eggs from Iceland and the Faroe Islands (mean: 16 and 15 ng/g ww). Samples from Norway contained PFOS concentrations (mean of both sites: 85 ng/g ww) that were about 5 times lower than that of the Swedish samples. PFOSA and N-EtFOSA were detected less frequently than PFOS and in contrast to PFOS, PFOSA concentrations were greatest in eggs from Sklinna, Norway. N-EtFOSA was detected in only ten animals with concentrations ranging from 0.77 ng/g ww in eggs from Iceland to 2 ng/g ww in samples from Hjelmsøya. The PFCA spatial pattern was different to that of PFOS. PFOA was not detected in the samples and PFNA was only detected in eggs from Sweden. PFUnA was the most abundant PFCA detected and the rank order of concentrations were Sweden (mean: 82 ng/g ww) > Faroe Islands (mean: 57 ng/g ww) > Norway (mean of both sites: 30 ng/g ww) > Iceland (mean: 18 ng/g ww).

Smithwick et al. [54] examined PFCs concentrations in polar bear liver tissue from five locations in North America and two in the European Arctic collected between 1999 and 2002. North American samples were from Nunavut (n=26) that was subdivided into south Baffin Island (consisting of animals from Pangnirtung, Qikiqtarjuaq, Iqaluit and Kimmirut) and the High Arctic (consisting of animals from Resolute, Grise Fjord and Pond Inlet), Northwest Territory (n=7), Northwestern Alaska (n=10, consisting of animals from Chukchi Sea and Bering Sea) and South Hudson Bay (Sanikiluaq). European samples were from Eastern Greenland (n=29, Scoresby Sound) and Svalbard. (It should be noted that because only blood plasma was available for samples from Svalbard, the authors used a conversion factor to estimate PFC concentrations in liver.) PFOS concentrations were greater than any other PFC examined in the study (Figure 2.6). There was a significant geographic trend for PFOS with animals from south Hudson Bay and Greenland having significantly greater concentrations than Svalbard, High Arctic and the Northwest Territory (p<0.05). This was attributed to closer proximity of these sites to possible sources in Europe and eastern North America. PFOS concentrations in animals from the Chukchi Sea were smaller than those from any other region.

67

Figure 2.6. Geometric mean concentrations (ng/g ww) of PFCs in polar bear liver from the North American and European Arctic. Error bars represent 95% confidence intervals. Reprinted with permission from Smithwick et al. [54].

For the PFCAs, PFDoA, PFTriA and PFPA showed a similar trend to that of PFOS. Other PFCs, namely, PFNA, PFDA, PFUnA, PFTA, PFHxS and PFOSA did not show distinguishable overall geographic trends. Interestingly, PFOS and PFCAs with greater than 10 carbon atoms had similar geographic distributions while PFOA, PFNA, PFDA, PFHxS and PFOSA were more evenly distributed or showed greater concentrations in the western North American Arctic.

For each location, there were some statistically significant correlations found for some of the PFCA homologues (Table 2.1). The strongest correlation was consistently observed between PFDA and PFUnA. This was attributed to a common source of PFCAs at each location. This was further addressed by calculating ratios of adjacent-chain-length PFC concentrations. Chukchi Sea samples were found to have a much greater proportion of PFNA to

68 PFOS than those in the eastern sampling locations. Similar relationships were also found between PFUnA to PFDA, PFDA to PFNA and PFDoA to PFUnA. Further, the proportion of PFNA was much greater than PFOA in Chukchi Sea samples but smaller in the eastern sampling areas. Differences in the sources of PFCs to the eastern and western locations were suggested to explain the patterns.

Table 2.1. Correlation coefficient (r2) of linear regression between adjacent chain length PFCAs in polar bear liver from the North American and European Arctic. Reprinted with permission from Smithwick et al. [54]

chain length 9:10 10:11 12:13 9:11 location r2 P r2 P r2 P r2 P Chukchi Sea 0.65 <0.01 0.78 <0.01 0.38 0.06 0.28 0.12 NWT 0.88 <0.01 0.80 0.01 0.81 0.01 0.57 0.08 High Arctic 0.98 <0.01 0.83 <0.01 0.83 <0.01 0.78 <0.05 South Baffin Island 0.66 <0.01 0.95 <0.01 0.62 0.02 0.57 0.03 South Hudson Bay 0.76 0.03 0.97 <0.01 0.38 0.38 0.72 0.07 Greenland 0.90 <0.01 0.75 <0.01 0.55 <0.01 0.79 <0.01 Svalbard 0.77 <0.01 0.71 <0.01 0.30 0.56 0.37 0.02

Tomy et al. [55] examined the spatial distribution of PFCs in beluga from the Canadian Arctic. Ten animals were collected from each of the following locations: Arviat (2003), Sanikiluaq (2003), Kimmirut (2003), Pangnirtung (2002) and Hendrickson Island (2005). Although the authors reported on some regional differences in PFC concentrations, there were no broad geographic trends that were discernable. PFCA concentrations were greatest in animals from Kimmirut (216.9 ± 14.2 ng/g ww, geometric mean ± 1 SE) and the smallest at Pangnirtung (25.6 ± 4.7 ng/g ww). The rank order of PFOS in liver was Sanikiluaq (47.7 ± 8.2 ng/g, ww) > Pangnirtung (22.6 ± 2.5 ng/g, ww) > Hendrickson Island (11.9 ± 1.8 ng/g, ww) > Arviat (7.9 ± 3.6 ng/g, ww) > Kimmirut (5.4 ± 0.7 ng/g, ww). ΣPFCA concentrations were significantly different (one-way ANOVA) in Kimmirut animals relative to those from Pangnirtung and Arviat (p<0.05); ΣPFCA concentrations in Sanikiluaq animals also differed to those at Arviat and Pangnirtung (p<0.05). PFOS concentrations in animals from Sanikiluaq were also different to those from Kimmirut, Hendrickson Island and Arviat; PFOS concentrations in Pangnirtung animals were different to those from Kimmirut (p<0.05). Like PFCAs, PFOSA concentrations was greatest in animals from Kimmirut (305.6 ± 25.3 ng/g ww) followed by Arviat (188.3 ± 36.3 ng/g ww). Spatial trends of PFOSA were different to that of PFOS. PFOSA concentrations were statistically different in animals from Kimmirut compared with those from

69 Pangnitung, Sanikiluaq and Hendrickson Island (p<0.05); differences were also evident between animals from Arviat and those from Sanikiluaq and Pangnirtung (p<0.05).

Spatial concentrations of PFCs were examined in ringed seals from 11 locations in the Canadian Arctic collected between 2002 and 2005 [23] (Figure 2.7). Ten individuals were collected from each of the following locations: Sachs Harbour (2005), Gjoa Haven (2004), Resolute Bay (2005), Arviat Bay (2004), Arctic Bay (2004), Grise Fjord (2003), Inukjuak (2002), Pond Inlet (2004), Qikiqtarjuaq (2005), Pangnirtung (2002) and Nain (2005). The authors found statistically significant differences in PFC concentrations among the 11 locations. However, these differences were driven largely by elevated concentrations at Gjoa Haven and Inukjuak, and by smaller concentrations at Pangnirtung. Geometric mean concentrations of PFNA, PFDA and PFOS at Gjoa Haven were about 8, 4 and 2.6-fold greater than those of the other 10 ringed seal populations.

Figure 2.7. Geometric mean concentration (ng/g ww) of PFOA, PFNA, PFDA, PFUnA and PFOS in ringed seal liver from Canadian Arctic. Error bars indicate one standard error. Reprinted with permission from Butt et al. [23].

70 The authors also used stable isotopes of nitrogen (δ15N) and carbon (δ13C) to discern differences in the trophic level and carbon sources, respectively. Mean δ15N data suggested that all the animals were from the same trophic level. Interestingly, the Gjoa Haven animals had a δ13C value that was significantly depleted as compared to the other populations suggesting a more carbon-rich terrestrial source. After adjusting their dataset for δ13C values, concentrations of most PFCs were generally greater in the Grise Fjord, Qikiqtarjuaq, Arviat and Nain populations.

Regional comparisons of PFC concentrations were then explored by grouping the ringed seals into four broad regions (excluding Gjoa Haven and Inukjuak): southeast Beaufort Sea, Hudson Bay, South Baffin Island & Labrador and High Arctic (Figure 2.8). With the exception of PFUnA and PFTrA which were statistically greatest in the Hudson Bay population, all the other PFCs analyzed were statistically similar across the regions. The authors caution that temporal variation in the dataset especially for PFOS and PFOSA might confound interpretation of spatial trends.

Figure 2.8. Geometric mean concentration (ng/g ww) of selected PFCs in ringed seals from the Canadian Arctic. Error bars represent 95% confidence intervals. Reprinted with permission from Butt et al. [23].

71 Spatial Trend Summary In summary, there is still a paucity of information on the geographic distribution of PFCs in the circumpolar Arctic. Thus far, only marine ecosystems have been investigated and no spatial studies exist on freshwater or terrestrial wildlife. The studies done to date in North America do capture wider and more large-scale geographic regions. Results indicate some spatial differences between populations, however, the origins of these trends have generally not been investigated. Further, geographic trends across the regions were not the same for all PFCs, confounding spatial trend interpretation. Overall the selection of biological species used as biomonitors are prudent as they represent an integrated and realistic measure of exposure to PFCs. More conclusive and large-scale datasets exist for other emerging halogenated compounds, such as the brominated diphenyl ethers (BDEs). This is perhaps not too surprising because existing sample extracts from research on more studied halogenated compounds like PCBs are suitable for BDE research. The scenario of course is quite different for the PFCs where different extraction and sometimes tissue compartments are needed.

72 2.4.3.1. Temporal Trends 2.4.1.3.1. North American Arctic Burbot PFC temporal trends (1986, 1999, 2003, 2006) were examined in burbot liver (n=10 per time point) from Fort Good Hope, Northwest Territories, Canada [56]. PFOS levels were steady from 1986 to 2003 but a large decrease was shown from 2003 (mean ± standard deviation = 9.88 ± 10.16 ng/g ww) to 2006 (1.93 ± 0.78 ng/g ww). PFOA exhibited somewhat similar trends with steady concentrations from 1985-1999 and consistent decreases from 1999-2003 and 2003- 2006. In contrast, PFDA levels were steady from 1985-1999 with consistent increases from 1999-2003 and 2003-2006. PFNA and PFUnA both showed increases from 1985-2003 with noticeable decreases from 2003-2006.

Lake Trout Temporal trends (1999, 2000, 2001, 2002, 2004, 2005, 2006) of PFOS, PFNA, ΣPFSAs and ΣPFCAs were reported in lake trout muscle from Lutsel K’e (eastern arm of Great Slave Lake), Northwest Territories, Canada [57]. PFC concentrations showed increasing concentrations from 1999 to 2001, followed by a marked decrease from 2001 to 2006. For example, PFOS concentrations increased from 1.9 to 4.7 ng/g ww between 1999 and 2001, decreasing to 0.04 ng/g ww in 2006. Similarly, PFNA concentrations increased from 0.4 to 1.6 ng/g ww between 1999 and 2001, decreasing to 0.02 ng/g ww in 2006. Statistical analysis was not reported.

Northern Sea Otter PFC temporal trends were investigated in livers of male northern sea otters from south- central Alaska from 1992-2007 [39]. Samples were collected from Prince William Sound (n=36), Resurrection Bay (n=7) and Kachemak Bay (n=34) but were grouped together for the temporal analysis. Samples were analyzed from every year with the exception of 1995 and the sample size ranged from 1 to 11 individuals per year. For the temporal trends analysis, only adults and sub-adults were included and samples were grouped into three time periods: “1992- 1997” (n=18), “1998-2001” (n=24) and “2002-2007” (n=26). PFOS showed an overall decline during the study period with a statistically significant decrease from the 1992-1997 period to the 2002-2007 period. In addition PFOS levels showed a significant decrease from the 1998-2001

73 period to the 2002-2007 period. Considering individual years, PFOS levels peaked in 2001 at 21.2 ± 22.7 ng/g ww (mean ± standard deviation). PFOS concentrations showed similar levels at the start (1992) and end (2007) of the study period. Similar temporal trends were observed for PFOSA with an approximately 8-fold decrease from the 1998-2001 period to the 2002-2007 period. PFOSA peaked in 1999 at 15.2 ± 8.6 ng/g ww. In contrast to PFOS and PFOSA, PFNA showed increasing levels from the 1998-2001 period to the 2002-2007 period. PFNA concentrations peaked in 2007, increasing from <2 ng/g ww in 2004 to 9.4 ± 10.4 ng/g ww in 2007. In addition, the PFC profiles shifted during the study period. PFOS and PFOSA dominated the earlier time periods, contributing 61% and 71% of the ΣPFC profile in the 1992- 1997 and 1998-2001 periods, respectively. The PFOS and PFOSA contribution decreased to 37% of the ΣPFC profile in the 2002-2007 period while the PFCA contribution increased 2-fold to 66%. PFOA, PFDA and PFUnA were below the LOQ.

Ringed Seal Temporal trends were examined in ringed seal liver from two populations in the Canadian Arctic, Arviat (western Hudson Bay) (1992, 1998, 2004, 2005) and Resolute Bay (Lancaster Sound) (1972, 1993, 2000, 2004, 2005) [38]. PFCs analyzed included C7-C15 PFCAs, 8:2 FTCA & FTUCA, 10:2 FTCA & FTUCA, C4, C6, C8, C10 perfluorinated sulfonates and PFOSA. PFHpA, PFBS, PFHxS and PFDS were not detected in any seal sample. The 8:2 FTCA & FTUCA were detected but were below the method detection limits (defined as the mean blank levels plus three times the standard deviation of the blanks). The 10:2 FTCA was detected but concentrations were not reported due to quantification problems. PFOA was infrequently detected above the MDL and thus not included in temporal trend analysis.

The C9-C15 PFCAs showed overall increasing levels during the time interval investigated. Calculated doubling times ranging from 19.4 ± 1150 (95% confidence interval) years to 15.8 ± 12.2 years for PFDoA to 10.0 ± 7.2 to 7.7 ± 2.0 years for PFNA in the Arviat and Resolute Bay populations, respectively (Figures 2.9 & 2.10). However, the later time points (1998, 2003 and 2005 for Arviat, and 2000, 2004 and 2005 for Resolute Bay) for the PFCAs were not statistically different from each other, implying the PFCA levels may be levelling off in ringed seals from these locations. In contrast to the PFCA trends, PFOS showed maximum concentrations during 1998 and 2000 at Arviat and Resolute Bay, respectively. Both

74 populations show statistically significant decreases from their maximum to 2005. In the Arviat population, two consecutive statistically significant decreases were observed, initially 1998 to 2003 and also from 2003 to 2005. In the Resolute Bay population, PFOS levels declined from 2000 to 2004 but were not statistically significant. However, the overall PFOS decline from 2000 to 2005 was significant. Apparent PFOS disappearance half-lives were 3.2 ± 0.9 (95% confidence interval) years and 4.6 ± 9.2 years for Arviat and Resolute Bay, respectively.

Figure 2.9. Geometric mean concentrations (ng/g ww) of PFOS, PFNA, PFDA and PFUnA in ringed seals from Arviat, Nunavut, Canada (1992-2005). Error bars indicate 95% confidence interval. Reprinted with permission from Butt et al. [38].

75

Figure 2.10. Geometric mean concentrations (ng/g ww) of PFOS, PFNA, PFDA and PFUnA in ringed seals from Resolute Bay, Nunavut, Canada (1972-2005). Error bars indicate 95% confidence interval. Reprinted with permission from Butt et al. [38]. Copyright 2007 American Chemical Society.

Beluga Whale

Temporal trends in PFOS, PFOSA and C8-C12 PFCAs were investigated in beluga whale liver from Hendrickson Island, Northwest Territories (1984, 1993, 1995, 2001, 2005, 2006 and 2007) and Pangnirtung, Baffin Island, Nunavut (1982, 1986, 1992, 1995, 2002, 2005, 2006 and 2007) in the Canadian Arctic [58]. Ten individuals were analyzed per time point with the exception of 2006 Pangnirtung samples (n=5). Temporal trends were not consistent between the two populations. For PFOS, an overall statistically significant increase was shown from 1984 to 2005 in the Hendrickson Island population with an apparent levelling off after 2005. In the Pangnirtung population, PFOS levels increased linearly from 1982 to 2002, increasing at a rate of ~0.5 ng/g ww per year, with an apparent decline shown after 2002. However, these latter decreases were not statistically different from the 2002 maximum. PFOSA temporal trends were similar to PFOS. For ΣPFCAs, the Hendrickson Island population showed a linear

76 decrease over the study period, decreasing from a maximum of 157 ± 15 ng/g (geometric mean ± SE) in 1984 to a minimum of 9.7 ± 1.5 ng/g ww in 2006. In contrast, ΣPFCA levels in the Pangnirtung population increased from 1982 (9.5 ± 2.4 ng/g ww) to 2002 (23.2 ± 4.7 ng/g ww) with an apparent levelling off after 2002.

Seabirds PFC temporal trends were examined in thick-billed murre (1975, 1993, 2004) and northern fulmar (1975, 1987, 1993, 2003) liver samples from Prince Leopold Island (Lancaster Sound) in the Canadian Arctic [59]. Between 8 and 10 individuals were analyzed per time point. PFCs analyzed included C7-C15 PFCAs, 8:2 FTCA & FTUCA, 10:2 FTCA & FTUCA, C6, C8, C10 perfluorinated sulfonates and PFOSA. In general, PFCs showed increasing levels over the entire study period in both species (1975 to 2003/2004 in thick-billed murres and northern fulmars, respectively) (Figure 2.11). Considering PFCAs, the concentration increases over both time periods (1975→1993 and 1993→2004) were significant for thick-billed murres but the trends were not as clear for the northern fulmars. For northern fulmars, PFCAs showed either maximum concentrations in 1993 or statistically similar concentrations in 1987, 1993 and 2003, suggesting a levelling off in PFCA levels beginning in the early 1990s. Doubling times could be calculated for both species since the increase over the entire study period was significant. Doubling times in thick-billed murres ranged from 2.3 yrs for PFPA to 9.9 yrs for PFDoA, and from 2.5 yrs for PFPA to 11.7 yrs for PFDA in northern fulmars. PFOS levels were statistically similar in both populations between 1993 and 2003/2004. The authors caution that the temporal trends should be viewed as representative of long-term trends due to the large interval between sampling periods.

77

Figure 2.11. Geometric mean concentration (ng/g ww) of PFCAs and fluorotelomer acids in (a) thick-billed murres and (b) northern fulmars from Prince Leopold Island, Nunavut, Canada. Errors bars indicate 95% confidence intervals. “*” indicates that all samples were below MDL or were not detected for that time point. Reprinted with permission from Butt et al. [59].

Polar Bear Temporal trends in PFCs were examined in polar bear livers from eastern and western populations in the North American Arctic between 1972-2002 [28]. The “eastern” population comprised samples collected near northern Baffin Island, Canada, with samples collected in 1972, 1975, 1982, 1984, 1993 and 2002. The “western” population comprised samples collected near Barrow, Alaska with samples collected in 1972, 1982 and 2002.

PFOS, PFNA, PFDA and PFUnA showed significant increases over the study period in both the eastern and western populations (Figures 2.12 & 2.13). Significant increases were also observed for PFHxS in the western group only and for PFOA in the eastern group only. PFOSA showed decreasing temporal trends in both groups, but was statistically significant in the eastern population only. PFDoA levels were near detection limits in all samples. PFTrA and PFTA were only detected in the 2002 samples, and PFPA, 8:2 FTCA & FTUCA, and 10:2 FTCA &

78

FTUCA were not detected in any sample. Doubling times were calculated for C8-C12 PFCAs, PFHxS, PFOS and PFOSA. Doubling times ranged from 3.6 ± 0.9 years for PFNA in the eastern group to 13.1 ± 4.0 years in the western group. The mean doubling times (mean of C8-

C11 PFCAs and PFOS) were much shorter in the eastern population (5.8 years) as compared to the western population (9.1 years). Considering individual PFCs, the rate of increase with time was significantly greater for PFNA and PFDA in the eastern population as compared to the western population. The rate of increase was statistically similar for PFOS and PFUnA between both populations. It was shown that the PFOS doubling times in polar bears (9.8 and 13 years for the eastern and western populations, respectively), was in good agreement with the PFOSF production doubling time of about 11 years. It was suggested that the PFOS doubling times observed in the polar bears was too short to account for transport via ocean currents.

Figure 2.12. PFOS temporal trends in polar bear livers from near northern Baffin Island, Canada (east) and near Barrow, Alaska (west) between 1972 and 2002. Vertical bars indicate 95% confidence intervals. Reprinted with permission from Smithwick et al. [28].

79

Figure 2.13. PFOA, PFNA, PFDA and PFUnA temporal trends in polar bear livers from near northern Baffin Island, Canada (east) and near Barrow, Alaska (west) between 1972 and 2002. Vertical bars indicate 95% confidence intervals. Reprinted with permission from Smithwick et al. [28].

80 2.4.1.3.2. Greenland Ringed Seal Temporal trends of PFC concentrations were investigated in ringed seal livers from two locations in Greenland [51]. Seals were collected from Ittoqqortoormiit (East Greenland) in 1986, 1994, 1999 and 2003, and from Qeqertarsuaq (West Greenland) in 1982, 1994, 1999 and 2003. Increasing PFOS levels over the study period were shown for the Ittoqqortoormiit and Qeqertarsuaq populations, although the regression was not significant for the Ittoqqortoormiit seals. Using a log-linear regression of the median concentration, PFOS annual increases were 8.2% (standard error = 3.9%) for Ittoqqortoormiit and 4.7% (SE=1.1%) for Qeqertarsuaq. Similarly, PFDA annual increases were 3.3% and 1.7% in Ittoqqortoomiit and Qeqertarsuaq populations respectively; and PFUnA annual increases were 6.8 and 5.9%.

PFC temporal trends in Greenland ringed seals were recently updated with collections in 2006 (Riget, unpublished) (Figure 2.14). Levels of PFOS, PFDA and PFUnA in the 2006 samples were greater than in 2003 in both populations. Annual increases over the entire study period (1986-2006 for Ittoqqortoormiit and 1982-2006 for Qeqertarsuaq) were greater than those measured over the 1986/1982-2003 time period. Annual increases in the Qeqertarsuaq seals were 10.7%, 5.7% and 7.6% for PFOS, PFDA and PFUnA, respectively. Similarly, annual increases in the Ittoqqortoormiit seals were 12.1%, 6.4% and 7.8% for PFOS, PFDA and PFUnA, respectively. However, the only significant regression was for PFUnA in both the Ittoqqortoormiit and Qeqertarsuaq populations.

81 Ittoqqortoormitt

PFOS PFDA PFUnA ng/g ww 51015202530 200 600 1000 0510 0

1985 1990 1995 2000 2005 1985 1990 1995 2000 2005 1985 1990 1995 2000 2005

Year Year Year Qeqertarsuaq PFOS PFDA PFUnA 600 800 400 ng/g ww 0246 024681012 0200

1985 1990 1995 2000 2005 1985 1990 1995 2000 2005 1985 1990 1995 2000 2005 Year Year Year

Figure 2.14. Temporal trends in PFOS, PFDA and PFUnA in ringed seal liver from Ittoqqortoormiit (East Greenland), 1986-2006, and Qeqertarsuaq (West Greenland), 1982-2006 (Riget, unpublished). Red circles represent median concentrations, red line represents significant log-linear regression, black line represents non-significant log-linear regression.

Polar Bear Temporal trends were investigated in polar bear liver samples from Ittoqqortoormiit in East Greenland between 1984-2006 [60] (Figure 2.15). Samples were analyzed for 19 out of the 21 years (n=128 subadults). Yearly increases were investigated by a log-linear regression of the median and in some instances a LOESS smoothing equation was applied to obtain a better fit. By using the log-linear model, significant annual increases were shown for all PFCs investigated with the exception of PFOSA, which did not show any significant trend. Yearly annual increases were shown for PFOS (4.7% per year), PFOA (2.5%), PFNA (6.1%), PFDA (4.3%), PFUnA (5.9%) and PFTrA (8.5%). Using a nonlinear LOESS smoother model it was shown that PFOS, PFOSA, PFDA and PFTrA showed steeper linear increases after 1990 or 2000. For

82 PFOSA, annual increases were 9.2%. After 2000, annual yearly increases were 19.7% for PFOS, 18.6% for PFDA and 27.4% for PFTrA. The dramatic increase in recent sampling years warrants further investigation to determine if these trends will continue.

It was suggested that the continued PFOS increase observed in Greenland polar bears and ringed seals may be indicative of other primary sources or different pathways to this region [60]. It was further speculated that the temporal trends in Greenland wildlife may be representative of a time lag in PFOS transport to East Greenland.

83

Figure 2.15. Temporal of PFCs in East Greenland polar bear liver from 1984 to 2006. Filled points represent log-linear regression lines or LOESS smoother lines. Broken lines represent 95% confidence limits. Reprinted with permission from Dietz et al. [60].

84 2.4.1.3.3. Norway Seabirds Temporal trends (1983, 1993, 2003) were investigated in herring gull eggs from two colonies located on the northern coast of Norway, Hornøya and Røst (n=5 per colony per time point) [52]. PFOS concentrations from both colonies increased nearly 2-fold (statistically significant increase) from 1983 to 1993 (Figure 2.16). PFOS levels appeared to level off between 1993 and 2003. PFHxS levels showed similar trends to PFOS in both colonies. In contrast, PFDS showed increasing levels throughout the entire study period (1983 to 2003). PFCA levels showed significant increases between 1983 and 1993 followed by either nonsignificant increases from 1993 to 2003 (C8-C11 PFCAs) or levelling off (C12 and C13 PFCAs). Variation in the PFCA composition was observed between sampling years, suggesting the PFCA sources may have changed over time. For example, eggs from Røst in 1993 and 2003 had significantly higher proportions of PFOA as compared to 1983. Further, Røst eggs collected from 1993 had higher proportions of PFTA and PFPA. Doubling times were not calculated due to the limited time points.

Figure 2.16. Temporal trends of PFOS, PFDS, PFUnA and PFTrA in herring gull eggs from the Hornøya and Røst, Norwegian Arctic. PFDcS and PFTriA are identified as PFDS and PFTrA, respectively, in the manuscript. Reprinted with permission from Verreault et al. [52].

85 Temporal Trend Conclusions In summary, there are numerous studies examining temporal trends of PFCs in arctic wildlife. Studies thus far have been exclusively from marine ecosystems (marine mammals, seabirds and polar bears) with the exception of burbot and lake trout from the Northwest Territories, Canada. Most temporal trend studies have been from the North American Arctic and Greenland.

Temporal trends between Arctic regions are not consistent. For example, declining PFOS concentrations have been shown in sea otter [39], ringed seal [38] and beluga whale [58] from the Canadian Arctic, whereas, ringed seals [51] and polar bears [60] from Greenland continue to show increasing PFOS levels from the 1980s to 2006. Some temporal studies may be confounded by the relatively large temporal intervals. Further, temporal trends within species from different regions are inconsistent. For example, Tomy et al. [58] observed declining ΣPFCAs trends in Hendrickson Island beluga but increasing ΣPFCA trends in Pangnirtung beluga. The inconsistencies observed between temporal studies may be due to differences in emissions from source regions, although spatially resolved temporal emission data is not presently known. Disparate regions of the Arctic are influenced by air currents from different regions [61], and thus could receive unique temporal patterns of volatile precursors. In addition, it has been suggested that the North American and European Arctic are influenced by dissimilar ocean waters [62]. For example, it has been suggested that surface seawater in the Canadian Archipelago and northern Hudson Bay is entirely of Pacific origin [63]. In contrast, the European Arctic is primarily influenced by Atlantic Ocean waters.

2.4.2. PFC profiles General In general, PFOS levels are shown to dominant the PFC profiles in arctic wildlife. Except for some whale species, PFOSA is usually detected at lower levels than PFOS. PFOSA has been to shown to be metabolically labile [64, 65]. Individual PFCA levels are typically lower than PFOS, although the ΣPFCA levels may be comparable in magnitude to PFOS. The PFCA profiles are usually dominated by either PFNA or PFUnA with a distinctive “odd-even” pattern in which concentrations of odd chain-length PFCAs are greater than adjacent even chain-length PFCAs. It has been suggested that these trends are the result of FTOHs as the dominant source of PFCAs [66]. For example, it has been shown that the atmospheric oxidation

86 of 8:2 FTOH yields approximately equal proportions of PFOA & PFNA [9]. Similar patterns are expected for other FTOHs (ie. atmospheric oxidation of 10:2 FTOH yields equal amounts of PFDA & PFUnA). Bioaccumulation has been shown to increase with increasing chain-length [67, 68], thus it would be expected that PFNA concentrations would be greater than PFOA, assuming exposure concentrations at the base of food web are equal. Alternatively, it has been reported that a PFNA commercial product (Surflon S-111) contains significant quantities of PFUnA and PFTrA [1]. However, the contribution of this commercial product to the observed PFCA patterns is unclear.

PFOA and PFCAs of lower chain lengths are typically not detected, or are in low concentrations. This is despite the fact that global PFOA seawater concentrations are generally similar, or greater than PFOS [41]. These trends are presumably due to the low bioaccumulation potential of PFOA and lower chain-length PFCAs [67, 68]. These trends are also observed in wildlife from non-Arctic regions.

The mechanisms explaining the apparently greater bioaccumulation potential of PFSAs as compared to PFCAs have not been fully elucidated. Several researchers have investigated structure-activity relationships that demonstrate differences in the behaviour of biological binding between the carboxylates and sulfonates and a full discussion of these trends is beyond the scope of this review. Gender differences in elimination rates are observed for PFOA, but generally not for PFOS [5] and may represent differences in the mechanism of protein binding. In addition, PFOS has been shown to bind more strongly than PFOA to liver-fatty acid proteins [69]. Chen and Guo [70] showed differences in binding affinity to human serum albumin between various PFCAs and PFSAs. These observed differences in protein binding likely influences overall clearance and accumulation potential.

Seabirds and Freshwater Birds

Longer-chain PFCAs (ie. C11-C15 PFCAs) have been shown to dominant PFCA profiles in some seabirds [53, 59, 71] as well as the common loon [66]. In contrast, PFNA and PFUnA are the dominant PFCAs in most wildlife. It is unclear why the long-chain PFCAs dominate seabird profiles.

87 Whales Several whale species have been shown to have PFOSA concentrations that are greater than, or approximately equal to, PFOS [50, 58, 72]. Bossi et al. [50] noted that cetaceans from the Mediterranean Sea showed PFOSA levels that were 1 to 5-fold greater than PFOS [73]. Further, melon-headed whales from the Japanese coast were shown to have comparable levels of PFOS and PFOSA [74]. These trends are in contrast to the majority of wildlife species that show relatively very low PFOSA levels. It has been shown that PFOSA is a metabolic precursor to PFOS [64, 65], and thus the relatively high levels of PFOSA in some cetaceans may be representative of a diminished metabolic capacity.

2.4.3. Animal Body Burdens There are few reports of PFC body burden in arctic wildlife. Liver and blood are the most frequently analyzed tissue in wildlife due to the tendency of PFCs to accumulate in the enterohepatic system [68]. However, predator animals may eat either the whole animal (e.g. seals eating fish) or blubber (e.g. polar bears eating seals). Thus, it may be useful to analyze PFC total body burden to achieve a more realistic assessment of biomagnification potential.

PFC concentrations in blubber, blood and liver were investigated in ringed seal (n=5) and bearded seal (n=1) collected from near Sachs Harbour, North West Territories in the Canadian Arctic [47]. Total animal body burdens were not calculated. Considering individual PFCs, concentrations were generally greatest in the liver followed by the blood and finally blubber. In fact, only PFOS (range: 0.4-0.9 ng/g ww) and PFDA (ND-0.2 ng/g ww) were detected in the blubber. In comparison, PFOS levels ranged between 2.5-8.6 ng/g ww and 18- 34 ng/g ww in the blood and liver, respectively. Similarly, PFDA levels ranged between 0.4-1.1 ng/g ww and 2.0-3.3 ng/g ww in the blood and liver, respectively. PFCs were not detected in the bearded seal blubber. It is unclear whether these trends represent actual partitioning of the PFCs into the seal blubber or PFCs in the blood vessels are found within the blubber.

PFCs were investigated in the plasma, liver, brain and eggs of glaucous gulls from the Norwegian Arctic [71]. PFOS levels showed plasma > liver ~ egg > brain. Similarly, ΣPFCA levels showed plasma > egg > liver > brain. There were few PFCAs detected in the brain samples. While the long-chain PFCAs dominated all tissues, specifically PFUnA and PFTrA, the individual dominant PFCA did vary by tissue type. PFUnA was the predominant PFCA in

88 plasma and eggs, whereas, PFTrA was predominant in liver. As well, the ΣPFCA: PFOS ratio varied among tissues, representing variation in the relative distribution of PFOS and ΣPFCA in the body. The ΣPFCA: PFOS ratios were significantly greater in plasma as compared to eggs, liver and brain, which were similar to each other.

2.4.4. “Neutrals” and Precursors Several polyfluorinated compounds have been shown to form PFSAs and PFCAs through metabolic processes. The presence of metabolically active “precursor” compounds within the animals may represent reservoirs of PFSAs and PFCAs. The production of PFOS from the biotransformation of N-ethyl perfluorooctane sulfonamide ethanol (N-EtFOSE) [65], N-EtFOSA [64] and PFOSA [75] has been observed in liver microsomal fractions of several species. Also, numerous studies have shown the formation of PFCAs from the metabolism of FTOHs [76-80] in whole animal and microsomes, with FTCAs and FTUCAs as intermediates. The 6:2 FtS may be a precursor to PFHxA [49]. More recently, PFHpA and PFOA were shown as metabolites in the biotransformation of 8:2 analogues of polyfluorinated phosphate surfactants [35]. The actual exposure of arctic wildlife to PFSA and PFCA precursors is not known. Many of these compounds possess high Henry’s Law constants [81] which effectively precludes exposure via water breathing organisms. With the exception of PFOSA, there are very few reports of PFSA and PFCA precursors in arctic wildlife. As such, to avoid repetition, discussion of precursors levels in this section does not include PFOSA.

Tittlemier et al. [82] reported N,N-Et2FOSA and N-EtFOSA levels in beluga and narwhal liver from Canadian Arctic. Mean N,N-Et2FOSA and N-EtFOSA concentrations were 1.2 and 3.6 ng/g ww and 3.3 and 11 ng/g ww for beluga and narwhal liver, respectively. PFCAs and PFSAs were not reported in these samples.

Tomy et al. [46] reported N-EtFOSA levels in organisms from an eastern Canadian Arctic food web. Levels of N-EtFOSA ranged from 0.39 ± 0.07 ng/g ww (mean ± standard error) for mixed zooplankton to 92.8 ± 41.9 for arctic cod. Interestingly, N-EtFOSA levels in some species were greater than PFOS and PFOA. N-EtFOSA was not detected in the redfish samples. Walrus, black-legged kittiwake and glaucous gull samples were not analyzed for N- EtFOSA.

89

Levels of N-EtFOSA were reported guillemot eggs from Iceland, Faroe Islands, Norway and Sweden [53]. Mean N-EtFOSA levels were 0.77, 0.98 and 2.0 ng/w ww in the Vestmannaeyjar (Iceland), Sklinna (Norway) and Hjelmsøya (Norway) populations, respectively. N-EtFOSA was not detected in the Faroe Island guillemot eggs. In general, N- EtFOSA levels were similar in magnitude to PFOSA but much lower than those of PFOS and the C10-C12 PFCAs.

Haukås et al. [48] reported 6:2 FtS concentrations in organisms from a Barents Sea food web. The 6:2 FtS was detected in 3 out of 6 ice amphipod samples (mean ± standard error: 0.48 ng/g ww ± 0.24) and 1 black guillemot sample (3.4 ng/g ww). The 6:2 FtS was not detected in the arctic cod and glaucous gull samples. As well, Verreault et al. [52] et al. reported 6:2 FtS levels in herring gull eggs from Hornøya and Røst in Northern Norway. Levels of 6:2 FtS were below the LOQ (0.16 ng/g ww) in samples from both colonies.

Several studies have monitored for but did not detect the 8:2 and 10:2 FTCAs & FTUCAs [47, 53, 71]. In a survey of 11 ringed seal populations across the Canadian Arctic [23], FTCAs and FTUCAs were detected in low levels. The 8:2 and 10:2 FTUCAs were detected in all populations. However, with the exception of Grise Fiord seals (8:2 FTUCA geometric mean = 6.0 n/g ww) and Pond Inlet seals (4.0 ng/g ww), levels were predominantly less than the MDL. Fluorotelomer saturated and unsaturated acids were also detected in temporal ringed seals from Resolute Bay and Arviat in the Canadian Arctic [38]. The 8:2 FTCA and FTUCA levels were below method detection limits. The 10:2 FTUCA levels ranged from <0.75 to 9.6 ng/g ww in the Arviat population and from <0.75 to 1.3 ng/g ww in the Resolute Bay population. The 8:2 and 10:2 FTUCAs were also detected in thick-billed murres and northern fulmars from Prince Leopold Island in the Canadian Arctic [59]. In the most recent samples, the geometric mean concentrations of 8:2 FTUCA were 0.02 and 0.01 ng/g ww for thick-billed murres and northern fulmars, respectively. The geometric mean 10:2 FTUCA concentrations were <0.20 and 0.48 ng/g ww for thick-billed murres and northern fulmars, respectively.

90 2.5. Abiotic Measurements

2.5.1. Atmospheric Measurements FTOHs and sulfonamide alcohols were measured in air from the North Atlantic and Canadian Archipelago [36]. The samples (n=20) were collected during a cruise in July 2005. FTOHs and sulfonamide alcohols were detected in all Arctic air samples, confirming their extensive occurrence in the Arctic atmospheric environment (Figure 2.17). These findings were consistent with models that predict the long-range atmospheric transport and widespread distribution of FTOHs [24] in arctic regions. The 8:2 FTOH was the dominant FTOH measured, representing between 50-70% (sum of gas- and particle- phases) of the total FTOH concentration, followed by the 10:2 FTOH and 6:2 FTOH. FTOH concentration ranges (sum of gas- and particle- phases) were: 5.8-26 pg/m3 for the 8:2 FTOH, 1.9-17 pg/m3 for the 10:2 FTOH, and

Spatial variation in the relative proportion of the volatile fluorinated compounds was observed. Air samples collected in the eastern region of the North Atlantic were dominated by N-MeFOSE. Back-trajectory analysis showed that these samples were representative for air originating from the North Atlantic. In contrast, air samples collected between western Greenland and the Canadian Archipelago showed a dominance of 8:2 FTOH. Air from these samples was representative of the Canadian Arctic Archipelago and Beaufort Sea region.

91

Figure 2.17. Total air concentrations (sum of gas-phase and particle-phase) of individual FTOHs and FOSEs from North Atlantic and Canadian Archipelago. Reprinted with permission from Shoeib et al. [36].

Stock et al. [22] reported neutral precursors and degradation products (PFCAs, telomer acids & PFSAs) in air from Resolute Bay, Nunavut, Canada. Samples were collected during the 2004 summer. Gas- and particle-phase were analyzed separately for the neutral precursors. FTOHs were detected in 50% of the samples and were almost exclusively in gas-phase. The mean concentration of ΣFTOHs (gas- and particle phase) was 28 pg/m3 with individual FTOH mean levels ranging from 2.8 pg/m3 for 10:2 FTOH to 14 pg/m3 for 8:2 FTOH. The mean ΣPFSAms (FBSEs + FBSAs + FOSEs + FOSAs + PFOSA) level was 112 pg/m3, 4-fold greater than ΣFTOHs. Mean concentrations of individual PFSAms ranged from 11 pg/m3 (N-EtFOSA) to 29 pg/m3 (N-MeFOSE).

PFCAs, telomer acids and PFSAs were measured in the filter samples only. PFOS (mean = 5.9 pg/m3) was the major compound with concentrations 1-2 orders of magnitude greater than most PFCAs and PFSAs. PFHxS (0.2 pg/m3) and PFDS (0.2 pg/m3) were also

92 detected. PFOA (1.4 pg/m3) was the dominant PFCA. Longer-chain PFCAs were detected less frequently and at comparatively lower levels. Mean concentrations of PFNA and PFDA were both 0.4 pg/m3 and mean concentrations of PFUnA, PFTrA and PFTA ranged from 0.02-0.06 pg/m3. PFHpA and PFDoA were not detected. The FTOH intermediate degradation compounds, 8:2 FTUCA (0.06 pg/m3) and 10:2 FTUCA (0.07 pg/m3), were detected at levels similar to the longer-chain PFCAs.

PFCs were measured in the particle phase of air samples from Zeppelinstasjonen, Svalbard collected in 2006 and 2007 [83, 84](Norwegian Institute for Air Research, 2007a; 2007b). PFOS and PFOSA were monitored in 2006 samples collected between late September to early December 2006. PFOS and PFOA were monitored in 2007 samples collected between early August and late December 2007. In the 2006 samples, mean PFOS levels were 0.11 pg/m3 (range: 0.03-0.50 pg/m3) and mean PFOSA concentrations were 0.07 pg/m3 (0.01-0.22 pg/m3). In the 2007 samples, mean PFOS levels were 0.18 pg/m3 (range: 0.02-0.97 pg/m3) and mean PFOA concentrations were 0.44 pg/m3 (0.15-1.51 pg/m3).

2.5.2. Snow 2.5.2.1. Canadian Arctic

Perfluorinated carboxylates (C8-C11 PFCAs) and PFOS were analyzed in snow from the Agassiz, Devon, Meighen and Melville ice caps in the Canadian arctic [26]. The sampling locations were located near the summit of the ice caps and were not thought to be significantly influenced by blowing snow. Therefore, due to the physical location of the sampling sites on the ice caps, the snow samples acted as a surrogate for atmospheric deposition. Surface snow samples were collected from the 4 ice caps during the spring of 2005 and 2006, representing deposition from 2004 and 2005. Perfluorinated carboxylates and PFOS were observed in surface snow samples from all ice caps with concentrations in the low pg/L levels. Specifically, surface concentration ranges were: 1.4-4.6 pg/L for PFOS, 13.1-53.7 pg/L for PFOA, 5.0-12.1 pg/L for PFNA, 1.5-4.5 pg/L for PFDA and 1.1-5.1 pg/L for PFUnA. Devon Island ice cap concentrations were approximately one order of magnitude greater than the other locations, possibility due to its relatively more southern latitude. PFC deposition fluxes were calculated by correcting the snow concentrations for density. Fluxes from each ice cap were extrapolated across the entire area north of 65oN in order to calculate the total PFC deposition to the Arctic. Calculated 2005 Arctic deposition fluxes were 114-587 kg/yr (arithmetic mean=271 kg/yr) for

93 PFOA, 73-860 kg/yr (295 kg/yr) for PFNA, 16-84 kg/yr (38 kg/yr) for PFDA, 26-62 kg/yr (46 kg/yr) for PFUnA, 250-1593 kg/yr (651 kg/yr) for total PFCAs, and 18-48 kg/yr (33 kg/yr) for PFOS. The PFOA and PFNA deposition fluxes were in good agreement with the 400 kg/yr calculated in a modeling study by Wallington et al. [24].

In addition, snow depth samples were collected from the Devon ice cap, representing atmospheric deposition from 1996 to 2005. Temporal trends of PFCAs showed overall relatively constant deposition fluxes. In contrast, PFOS deposition fluxes, after increasing from 1996 to 1998, were shown to decrease from 1998 to 2001 with relatively constant fluxes from 2001 onwards. These temporal trends reflect the changes in PFOSF production, although it is noted that the peak deposition flux occurred several years prior to the phase-out reported by industry. These temporal trends are consistent with PFOSF production trends. The source of the PFCAs on the ice caps was inferred to be from the atmospheric oxidation of FTOHs. Ratios of PFC to sodium (marker for seawater) concentrations were not correlated, suggesting that marine aerosols were not a significant source of PFCs to the ice caps. Further, ratios of adjacent PFCAs (eg- PFOA and PFNA) were approximately unity and significantly correlated. Finally, the detection of PFDA and PFUnA, compounds which are not directly commercially produced, suggests an indirect source such as FTOHs.

2.5.2.2. Greenland Surface snow from an ice flow east of Greenland was analyzed for PFC levels [27]. The major PFCs measured were PFOS (range: 25.2-137 pg/L), PFOA (50.9-520 pg/L) and PFDA (110-149 pg/L). PFHxS (8.2-40.2 pg/L), PFOSA (24.2-39.4 pg/L), PFHxA (<10-34.8 pg/L), PFHpA (12.1-85.4 pg/L) and PFNA (<30-76.6 pg/L) were measured in comparatively lower concentrations. It was noted that surface snow concentrations were much greater than seawater levels from the region.

2.5.3. Lake Water & Sediments 2.5.3.1. Amituk, Char & Resolute lakes on Cornwallis Island, Canadian Arctic Stock et al. [22] reported PFC levels in surface water from four lakes (Amituk Lake, Char Lake, Resolute Lake and Meretta Lake) on Cornwallis Island, Nunavut, Canada. Char, Resolute and Meretta Lake are located nearby the hamlet of Resolute Bay, whereas Amituk Lake is located approximately 40 kilometres north. Samples were collected during 2003 and

94 2005 with the exception of Amituk Lake (2003 only). PFC profiles and concentrations were similar in Amituk Lake and Char Lake. Mean PFOS concentrations ranged from 1.2-1.8 ng/L while PFHxS and PFDS were not detected. The C7-C12 PFCAs were also detected in water from Amituk Lake and Char Lake. Ranges of mean concentrations were 0.3-0.6 ng/L for PFHpA, 0.9-4.1 ng/L for PFOA, 0.3-1.5 ng/L for PFNA, 1.1-10.1 ng/L for PFDA, 2.5-4.9 ng/L for PFUnA and nd-0.4 ng/L for PFDoA. In contrast, the PFC profiles and concentrations were noticeably different in Resolute Lake and Meretta Lake. PFHxS, PFOS, PFHpA and PFOA levels were up to 60-fold higher as compared to Amituk Lake and Char Lake. For example, mean PFOS concentrations ranged from 23-69 ng/L and mean PFOA concentrations ranged from 5.6-14 ng/L. Concentrations of the C9-C12 PFCAs were similar to those in Amituk Lake and Char Lake. Meretta Lake serves as the inflow to Resolute Lake and thus both lakes appear to be contaminated by the same source. Intermediate FTOH degradation products, 8:2 FTUCA and 10:2 FTUCA, were also detected in all lakes with mean concentrations ranging from nd-1.9 ng/L for 8:2 FTUCA and nd-6.4 ng/L for 10:2 FTUCA.

As well, PFCs in sediment from Amituk, Char and Resolute Lake were reported [22]. Surface and depth samples were analyzed, allowing for temporal trend analysis (Figure 2.18). PFCAs greater than PFDA were not detected and levels of 8:2 FTUCA were

95 The elevated PFHxS, PFOS, PFHpA and PFOA levels measured in the surface water of Meretta and Resolute Lake, and the sediment of Resolute Lake, was attributed to AFFF contamination from the local airport and sewage runoff. As indicated in an earlier section, Arctic Char from Resolute Lake also showed elevated levels of some PFCs relative to Char and

Amituk Lake. The authors noted that the long-chain PFCAs (C10-C12 PFCAs) were not elevated in the Resolute Lake water or sediment which was suggestive of an atmospheric source of these compounds. It was also noted that the ratios of PFOA:PFNA and PFDA:PFUnA in Amituk and Char Lake sediment and water were generally consistent with those in Arctic glacial ice caps [26].

Figure 2.18. PFC concentrations (ng/g dry weight) in sediment core slices from Resolute, Char and Amituk Lake on Cornwallis Island, Nunavut, Canada [22]. Reprinted with permission from Stock et al. [22].

2.5.3.2. Isomers in Char Lake sediments; Surface Water from Char Lake and Amituk Lake De Silva et al. [85] reported PFCA isomers in sediment and surface water from Char Lake, and sediment from Amituk Lake on Cornwallis Island, Nunavut, Canada. Both lakes are thought to primarily receive PFCAs from atmospheric deposition [22].

96 PFOA profiles in Char Lake sediment showed a predominance of the linear isomer. The isopropyl (iso-) isomer (iso-:n- ratio = 2-3%) and the 5m- isomer (1-2%) were also detected in the sediment. In Char Lake and Amituk Lake, the iso- and 5m-PFOA isomers were detected at very low levels. Branched isomer proportions in Char Lake were 0.3% and 0.6% for the 5m- and iso-PFOA isomers, and in Amituk Lake were 0.2% and 0.5% for the 5m- and iso-PFOA.

PFNA profiles in Char Lake sediment were dominated by the linear isomer (96-97%). All four branched PFNA isomers (iso-, 1-, 3-, 4-) were detected and the iso-PFNA (2-3%) was the major branched isomer detected. The linear isomer was also the major isomer detected in the Char Lake and Amituk Lake surface waters. However, the only branched isomer detected was iso-PFNA with iso-:n-PFNA ratios in the two lakes ranging from 0.8-1%.

In the Char sediment, the iso- isomer was the only branched isomer detected in the longer-chain PFCA profiles. The iso-:n-PFCA ratios were 1-4% for PFDA, 1-2% for PFUnA and 4-8% for PFDoA. In contrast, branched isomers of PFDA, PFUnA or PFDoA were not detected in the lake surface waters.

2.5.4. Seawater and Marine Sediments 2.5.4.1. Greenland Sea PFC concentrations in surface seawater samples from the Greenland Sea have been reported [27]. Twenty-one samples were collected between eastern Greenland and Tromsö, Norway including locations near Svalbard. The major PFCs detected were PFOA (range: <30- 111 pg/L) and PFOS (<10-90 pg/L). PFHxS (<6-19 pg/L), PFHxA (10.2-37.6 pg/L), PFHpA (<12-31 pg/L) and PFNA (<30-55 pg/L) were measured in comparatively lower concentrations. PFOSA was detected at very low levels (<2-3.2 pg/L). PFDA levels were <20 pg/L. PFCs concentrations were greatest west of Norway. It was suggested that these samples may have been influenced by the Gulf Stream, although it was noted that the number of samples was too low to make definitive conclusions.

2.5.4.2. Labrador Sea Yamashita et al. [41] reported PFBS, PFOS and PFOA concentrations in seawater from the Labrador Sea in the North Atlantic Ocean. Three surface water samples (0-2 m) and two depth profiles were collected during September 2003 and 2004, respectively. Depth samples

97 were collected from south-western Greenland (“AO1”, 11 depth samples between 45-3500m) and from south-eastern Greenland (“AO2”, 13 depth samples between 15-2750m). In the “AO1” water column, PFOA was the dominant PFC measured. Surface water concentrations of PFOS and PFOA were 20 pg/L and 55 pg/L, respectively. PFBS, PFOS and PFOA concentrations were relatively constant with depth until 2000 m, after which PFBS and PFOA levels increased. In the “AO2” water column, PFBS was the dominant PFC measured. PFBS, PFOS and PFOA were elevated in the surface waters followed by uniform concentrations down to 2000 m. PFC depth profiles were consistent with temperature and salinity measurements which suggested a well-mixed water column down to 2000 m. Similar to the AO1 column, PFBS and PFOA levels increased below 2000 m. It was suggested that the increase in PFCs below 2000 m in both water columns was due to the influence of a deep water current, specifically the “Denmark Strait Overflow Water”.

2.5.4.3. Canadian Arctic Rosenberg et al. [40] assessed the spatial and vertical distribution of PFCs in seawater from Arctic and sub-arctic seawater in the Canadian Arctic archipelago. While C6 to C11 PFCAs as well as PFBS, PFHxS and PFOS were detected in almost all samples, PFOA and PFNA were the dominant PFCAs in seawater accounting for 60% of the ΣPFCA concentration while PFOS accounted for over 75% of ΣPFSAs. Mean PFOA concentration in water from the Labrador Sea at the Makkovik Margin (n=2, 182 pg/L) were ~3-fold greater than those measured by Yamashita et al. [41] for a site (AO1) in the central Labrador Sea but similar to those measured further south off Newfoundland. Concentrations of PFOS in seawater ranged between ~10 pg/L from the McClintock Channel to 424 pg/L from Kuujjuarapik.

2.5.4.4. Iceland & Faroe Islands PFCs in seawater from Iceland (n=1, 4 replicates) and the Faroe Islands (n=3) were reported by Kallenborn et al. [86]. PFOA was the major PFC measured with concentration ranges of 3.53-4.02 ng/L in Iceland and 3.62-7.24 ng/L in the Faroe Islands. PFHxA was the next highest PFC with concentration ranges of 0.63-0.73 ng/L and 0.59-1.85 ng/L in Iceland and the Faroe Islands, respectively. PFBS, PFHxS, PFOS (Faroe Islands only) and PFNA (Iceland only) were also detected at levels generally <1 ng/L. PFOSA levels were

98 Kallenborn et al. [86] reported marine sediment PFC levels from Gufunes Bay, Iceland (n=1) and from Torshavn, Vagsbotni and Fjardakanningar in the Faroe Islands (n=1 per location). In the Iceland sample, all PFCs measured (PFHxS, PFOS, PFOSA, PFHxA, PFOA, PFNA) were

2.5.4.5. Russian Arctic PFCs were monitored in ice cores sampled from Baydaratskaya Bay in the Russian Federation during May 2007 [87]. The samples represent frozen seawater with some snow deposition on the surface. Samples were collected from different depths, ranging from the surface to 300 cm. A pooled sample, comprising samples from various depths was reported. The most abundant analyte was PFOSA (mean ± standard deviation = 824 ± 592 pg/L), followed by PFOA (131 ± 77.2 pg/L). PFHxS was below the limit of detection. The remaining

PFCAs (C4, C6-C12) and PFSAs (PFBS, PFOS) were measured in comparatively lower levels with mean concentrations ranging from 3.6 ± 5.0 pg/L for PFDoA to 37.4 ± 39.2 pg/L for PFNA.

2.5.5. Sewage sludge & effluent PFCs in sewage sludge from Iceland (n=2) and the Faroe Islands (n=1) were reported by Kallenborn et al. [86, 88]. PFOA was the dominant PFC measured in both the Iceland (range: 0.25-0.40 ng/g ww) and Faroe Island (1.08 ng/g ww) samples. PFHxS and PFOS were detected at comparatively lower levels. PFHxS levels were 0.01-0.02 ng/g ww and 0.02 ng/g ww in the Iceland and Faroe Islands, respectively. PFOS levels were 0.07-0.22 ng/g ww and 0.24 ng/g ww in the Iceland and Faroe Islands, respectively. PFHxA was

Sewage effluent from the Faroe Islands (n=1) was reported by Kallenborn et al. [86]. PFC profiles were dominated by PFHxA (1.61 ng/L), PFOA (1.26 ng/L) and PFOS (1.22 ng/L). PFBS (0.20 ng/L), PFHxS (0.26 ng/L) and PFNA (0.44 ng/L) were also detected at comparatively lower levels. PFOSA levels were

99

The sewage sludge and effluent concentrations are representative of anthropogenic discharges to the Arctic environment and as such may represent point sources of PFCs. The relevance of these sources to the receiving environments is not known. Presumably the levels are representative of PFC consumer products and consumer applications and thus will vary by community.

2.5.6. Abiotic Environment Conclusions In summary, there have been very limited PFC measurements in the abiotic environment. To date, the majority of abiotic measurements are from the Canadian Archipelago and the North Atlantic. There have been some air measurements of neutral precursors (FTOHs, FOSEs and FOSAs) and degradation compounds (telomer acids, PFCAs, PFSAs), mainly from the Canadian Archipelago and the North Atlantic. Although limited, current studies have shown that FTOHs and sulfonamide alcohols are ubiquitous in the Arctic environment, confirming that these compounds are subject to long range transport. The detection of telomer acids, PFCAs and PFSAs on particulate-associated fractions is supportive of atmospheric oxidation of precursors as the source of these compounds. Spatial variation in the atmospheric profile of neutral precursors may be indicative of continental emission trends, although data is very limited. Snow has been used as a surrogate for atmospheric deposition of PFCs. Deposition fluxes of PFCs are consistent with some model predications. There is also very limited data on PFCs in lake water and sediment. There are limited measurements of PFCs in Arctic seawaters and studies to date have only been from the Canadian Arctic archipelago, the Labrador and Greenland Seas in the North Atlantic, and the Russian Federation. However, the detection of PFCs in Arctic seawaters confirms that direct transport via ocean currents also occurs. Additional seawater measurements are critical to validate existing model predications, to assess the relative importance of direct versus indirect long-range transport (see section 2), as well as to elucidate the spatial and temporal trends observed in some wildlife species. Finally, there have been limited temporal trend studies in abiotic media. Present studies either show poor temporal resolution (lake sediment) or extend over only a few years (snow core).

100 2.6. Conclusions and Research Needs To date, the bulk of the monitoring efforts in the arctic environment have concentrated on the PFCAs and PFSAs. The PFCAs and PFSAs are unique from the legacy POPs in that they are potential degradation compounds of commercial products (e.g. fluorinated polymers, fluorinated phosphate surfactants) and of compounds used in the manufacture of commercial products (e.g. fluorinated alcohols and acrylates). Furthermore, some PFCAs and PFSAs are also produced and applied directly. However, with the exception of PFOA, PFNA and PFOS, PFCAs and PFSAs were not industrially produced in large quantities.

The sources and transport routes of PFCs to the arctic are not well understood and have recently been the subject of considerable research effort. The two major postulated pathways are: the atmospheric transport and oxidation of volatile precursors, and the direct transport of PFCAs and PFSAs via ocean currents. Local inputs do not appear to significantly influence regional PFC concentrations; however, research is very limited. Several modeling studies have attempted to quantify the relevance of the two major postulated transport pathways. Thus far, models have used FTOHs and sulfonamide alcohols as inputs for volatile precursors. Future modeling studies should include additional precursors such as the fluorinated olefins, iodides and acrylates, as well as, the fluorinated phosphates. Model results are inherently sensitive to emission estimates. For example, Wallington et al. (2006), using an emission rate of 1000 tonnes per year, showed the FTOH degradation could explain environmental levels of PFCAs. In contrast, several others studies, using lower FTOH emissions rates, suggest that direct emissions of PFCAs and subsequent ocean transport is the dominant transport pathway. Modeling studies published to date have primarily focused on PFOA fluxes to arctic seawater, presumably due to the general dominance of PFOA in water samples. However, considering that PFOA is generally only infrequently detected in arctic wildlife, and at low levels, future modeling efforts should also include the PFSAs and longer-chain PFCAs.

PFCs appear to be ubiquitously detected in arctic biota. The marine food web has been well studied, in particular top predators such as seabirds, ringed seals and polar bears. There have been few reports of PFCs in marine zooplankton and fish. Earlier monitoring efforts generally only reported levels of PFOS, PFOSA, PFOA and occasionally a small set of other PFCAs such as PFNA. However, in recent years there are more frequent reports of the longer-

101 chain PFCAs. The importance of monitoring for longer-chain PFCAs is demonstrated by the fact that these compounds typically dominate PFCA profiles.

The freshwater ecosystem has been poorly examined, and in fact reports are limited to freshwater fish. Similarly, the terrestrial ecosystem has also been poorly investigated. The only reports of PFCs in terrestrial wildlife are from the Canadian Arctic. The limited reports of PFCs in freshwater and terrestrial ecosystems represent significant knowledge gaps in our understanding of PFCs in the arctic. PFCs have been detected in the surface water and sediments of freshwater lakes in the Canadian Arctic. There is a need for additional studies examining lake water and sediment levels in other regions of the Arctic. Sediment core analysis shows the general increase in PFC levels over the past 50 years.

Food web studies are limited and have been restricted to the marine environment. As such, there is an immediate need to study freshwater and terrestrial food webs. The marine food web studies published to date have generally shown high bioaccumulation in species at the top of the food web, particularly for PFOS and some long-chain PFCAs. However, the positive correlation between trophic position and PFC concentration is not necessarily indicative of biomagnification since there are significant uncertainties regarding the mechanism of PFC bioaccumulation and biomagnification. For example, biomagnification factors are typically calculated based on single organ concentrations (i.e. liver) that may not accurately represent consumption trends.

Similarly, there are few spatial trend studies and those published to date have only examined marine ecosystems. There is a requirement for spatial trend studies on freshwater and terrestrial wildlife. Further, there is a need for spatial studies that are larger in scale, preferably encompassing the entire circumpolar arctic. In particular, there is almost nothing known about PFC levels in the Russian Arctic. Understanding spatial patterns may assist in our interpretation of emission patterns and transport pathways. The more large-scale geographic studies have generally been from the North American arctic. Although some studies have reported spatial differences between populations, in general the causes of these discrepancies have not been investigated. In addition, interpretation of PFC spatial patterns is confounded by the fact that trends between PFCs are generally not consistent.

102 With few exceptions, most temporal trend studies are from marine ecosystems in the North American arctic and Greenland. There is a need for temporal trend studies from other arctic regions. In addition, temporal trend studies from freshwater and terrestrial ecosystems are needed. To date, temporal studies have shown inconsistent trends between regions, possibly due to differences in emissions from source regions, although spatially resolved temporal emission data is not presently available. Declining PFOS levels have been observed in sea otter, ringed seal and beluga whale from the Canadian Arctic, however, ringed seals and polar bear from Greenland show continued increasing PFOS levels from the 1980s to 2006. Given these inconsistencies, continued monitoring in existing studies is needed to confirm trends.

There are very few measurements of PFC precursors, such as the FOSEs, FOSAs and FTOHs, in wildlife. As such, trends are difficult to interpret. Given that these compounds have been shown to metabolically degrade to form PFSAs and PFCAs, the presence of “precursors” may represent reservoirs of these compounds. Additional measurements of PFC precursors in wildlife are needed. However, given the labile nature of these compounds in wildlife, it is unclear whether will wildlife will accumulate significant quantities of PFC precursors.

Overall, there are few measurements of PFCs in abiotic media from the arctic. The majority of the abiotic measurements are from the Canadian arctic and the North Atlantic. Atmospheric measurements from the North Atlantic and Canadian Arctic show the presence of PFC precursors, specifically the FTOHs and sulfonamide alcohols. PFC degradation products (PFCAs and PFSAs) and precursor intermediates (fluorotelomer carboxylates) have also been detected on atmospheric particles. Presumably these compounds originate from the atmospheric oxidation of precursor compounds. Spatial variation in the relative fraction of FTOHs and sulfonamide alcohols was observed between the North Atlantic and Canadian Archipelago, and may be representative of continental emission patterns. Expanded atmospheric monitoring is required to confirm these trends. In addition, atmospheric monitoring should include recently identified potential PFCA precursors such as the fluorinated olefins, iodides and acrylates. Finally, temporal analysis of arctic air samples would assist in understanding wildlife temporal trends.

Snow samples act as a surrogate for atmospheric deposition and have been analyzed from the Canadian arctic and Greenland. The presence of PFOS and PFCAs in the snow

103 samples confirmed that PFCs can be transported to the arctic environment via the atmosphere, presumably as volatile precursors with subsequent atmospheric oxidation and deposition. Depth samples, representing deposition from 1996 to 2005, in a snow pit from the Canadian arctic showed relatively constant deposition fluxes of PFCAs, but decreasing fluxes of PFOS from 1998 to 2001 followed by constant fluxes thereafter. These trends, taken together with some of the wildlife temporal studies, may indicate that the arctic environment is responding to changes in PFOSF-production, although additional studies in other regions of the arctic are needed.

There are very limited measurements of PFCs in seawater. Studies to date are relatively spatially constrained and are primarily from the Canadian Archipelago, the Labrador and Greenland Seas in the North Atlantic, and the Russian Federation. PFOA is generally detected in the greatest concentrations. Levels of long-chain PFCAs have not been reported. The detection of PFCs in Arctic seawater confirms that direct transport via ocean currents occurs. However, it is unclear whether the PFCs in the seawater have originated from direct emissions or from the atmospheric oxidation of precursors and subsequent deposition to the ocean surface. There is an immediate need for additional seawater measurements to validate existing model predications. Such measurements will greatly assist in assessing the relative importance of direct versus indirect long-range transport.

There are few measurements of the C8 PFOSF-chemistry replacement compounds such as FBSE (butyl sulfonamide alcohol). PFBS and PFBA are expected to be formed from FBSE and derivatives through metabolic and abiotic mechanisms analogous to the FOSEs. However, it is unlikely that PFBS and PFBA will accumulate in arctic wildlife considering their short biological half-life [67, 89].

There is increasing evidence of recent climate change in the Arctic environment. The magnitude of warming is variable across the Arctic, but overall it is nearly twice that of the global average [90, 91]. A changing climate has been linked to dramatic ecosystem changes in many global regions [92] including the arctic [93]. Climate change may influence contaminant accumulation through altering transport pathways and changing food web dynamics [94]. These processes are complex and it remains unclear if or how PFC levels will be specifically influenced. Therefore, it is suggested that future research focus on the influence of climate change on PFC levels in the arctic.

104

In conclusion, PFCs were first reported as contaminants in arctic wildlife by Giesy & Kannan in 2001. Since that time, our understanding of the levels and trends of PFCs in the biotic and abiotic environment of the circumpolar arctic has improved drastically. As with many other persistent organohalogenated compounds, we now appreciate that the arctic environment is ubiquitously contaminated with PFCs, however, many data gaps still exist. The effects of PFCs on arctic wildlife are not yet known. In addition, there is relatively limited understanding of PFCs in the abiotic environment, as well as in terrestrial and freshwater biota. Some PFCs appear to biomagnify in arctic food webs; and in a global context some of the highest known PFC concentrations are measured in polar bears. Further, potential transport pathways have been postulated in the literature, and several models have been developed in attempt to understand the relevance of these pathways.

Additional monitoring efforts should be designed in a deliberate manner to assist in our understanding on the fate and disposition of PFCs in the arctic. As mentioned above, additional seawater samples are needed. In particular, seawater samples should spatially distributed such that the potential influence of the major oceanic currents (and thus related to source region) can be identified. Additional air monitoring of precursors, including newly identified chemicals, and degradation compounds are necessary to understand fluxes to the terrestrial and marine environments. Similar to seawater collection, air monitoring should be spatially distributed to investigate contributions from different source regions. Additional monitoring of temporal trends in biota is needed to confirm observed trends. Archived biota samples should be chosen such that there is a long time course with sufficient temporal resolution to establish trends and distinguish the influence of inter-year variability. Given the observed differences in temporal trends in ringed seals from two arctic regions, spatially variable sample locations are needed. As well, temporal trends in additional wildlife species are needed.

2.8 Acknowledgements The authors wish to thank Annika Jahnke for the PFC transport pathway figure (Figure 2.1). We are grateful to James Armitage (Stockholm University) for useful discussions regarding PFC modeling. Thank you to Marlene Evans (Environment Canada) and Frank Riget (National Environmental Research Institute, Denmark) for providing previously unpublished data.

105 2.9. Literature Cited 1. Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H., Sources, fate and transport of perfluorocarboxylates. Environ. Sci. Technol. 2006, 40, 32-44.

2. OECD. Lists of PFOS, PFAS, PFOA, PFCA, related compounds and chemicals that may degrade to PFCA. Environment Directorate, Joint meeting of the chemicals committee and the working party on chemicals, pesticides and biotechnology, ENV/JM/MONO(2007)15, August 21, 2007. http://www.oecd.org/LongAbstract/0,3425,en_2649_34375_39160347_119666 _1_1_1,00.html.

3. Dallaire, R.; Dewailly, É.; Pereg, D.; Dery, S.; Ayotte, P., Thyroid function and plasma concentrations of polyhalogenated comounds in Inuit adults. Environ. Health Persp. 2009, 117, 1380-1386.

4. Weihe, P.; Kato, K.; Calafat, A. M.; Nielsen, F.; Wanigatunga, A. A.; Needham, L. L.; Grandjean, P., Serum concentrations of polyfluoroalkyl compounds in Faroese whale meat consumers. Environ. Sci. Technol. 2008, 42, 6291-6295.

5. Lau, C.; Anitole, K.; Hodes, C.; Lai, D.; Pfahles-Hutchens, A.; Seed, J., Perfluoroalkyl acids: A review of monitoring and toxicological findings. Toxicol. Sci. 2007, 99, 366-394.

6. Jones, P. D.; Hu, W.; De Coen, W.; Newsted, J. L.; Giesy, J. P., Binding of perfluorinated fatty acids to serum proteins. Environ. Toxicol. Chem. 2003, 22, 2639-2649.

7. Sonne, C.; Bossi, R.; Dietz, R.; Leifsson, P. S.; Riget, F. F.; Born, E. W., Potential correlation between perfluorinated acids and liver morphology in East Greenland polar bears (Ursus maritimus). Toxicol. Environ. Chem. 2008, 90, 275-283.

8. Letcher, R. J.; Gabrielsen, G. W.; Bustnes, J. O.; Dietz, R.; Jenssen, B. M.; Jorgensen, E. H.; O'Hara, T.; Sonne, C.; Verreault, J.; Strand, J., A state-of-the-science assessment of the effects and implications of persistent organic pollutants in circumpolar arctic wildlife. Sci. Total Environ. In press.

9. Ellis, D. A.; Martin, J. W.; De Silva, A. O.; Mabury, S. A.; Hurley, M. D.; Sulbaek Andersen, M. P.; Wallington, T. J., Degradation of fluorotelomer alcohols: A likely atmospheric source of perfluorinated carboxylic acids. Environ. Sci. Technol. 2004, 38, 3316-3321.

10. Hurley, M. D.; Wallington, T. J.; Sulbaek Andersen, M. P.; Ellis, D. A.; Martin, J. W.; Mabury, S. A., Atmospheric chemistry of fluorinated alcohols: reaction with Cl atoms and OH radicals and atmospheric lifetimes. J. Phys. Chem. A 2004, 108, 1973-1979.

11. Martin, J. W.; Ellis, D. A.; Mabury, S. A.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of perfluoroalkanesulfonamides: Kinetic and product studies of the OH radical and Cl atom initiated oxidation of N-ethyl perfluorobutanesulfonamide. Environ. Sci. Technol. 2006, 40, 846-872.

12. D'eon, J. C.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A., Atmospheric chemistry of N-methyl perfluorobutane sulfonamidoethanol, C4F9SO2N(CH3)CH2CH2OH: Kinetics and mechanism of reaction with OH. Environ. Sci. Technol. 2006, 40, 1862-1868.

106 13. Vésine, E.; Bossoutrot, V.; Mellouki, A.; Le Bras, G.; Wenger, J.; Sidebottom, H., Kinetic and mechanistic study of OH- and Cl-initiated oxidation of two unsaturated HFCs: C4F9CH=CH2 and C6F13CH=CH2. J. Phys. Chem. A 2000, 104, 8512-8520.

14. Nakayama, T.; Takahashi, K.; Matsumi, Y.; Toft, A.; Sulbaek Andersen, M. P.; Nielsen, O. J.; Waterland, R. L.; Buck, R. C.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of CF3CH=CH2 and C4F9CH=CH2: Products of the gas-phase reactions with Cl atoms and OH radicals. J. Phys. Chem. A 2007, 111, 909-915.

15. Young, C. J.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A., Atmospheric chemistry of 4:2 fluorotelomer iodide (n-C4F9CH2CH2I) and 6:2 fluorotelomer iodide (n-C6F13CH2CH2I): Kinetics and products of photolysis and reaction with OH radicals and Cl atoms. J. Phys. Chem. A 2008, 112, 13542-13548.

16. Butt, C. M.; Young, C. J.; Mabury, S. A.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of 4:2 fluorotelomer acrylate [C4F9CH2CH2OC(O)CH=CH2]: Kinetics, mechanisms, and products of chlorine atom- and OH radical-intiated oxidation. J. Phys. Chem. A 2009, 113, 3155-3161.

17. Armitage, J.; Cousins, I. T.; Buck, R. C.; Prevedouros, K.; Russell, M. H.; MacLeod, M.; Korzeniowski, S. H., Modeling global-scale fate and transport of perfluorooctanoate emitted from direct sources. Environ. Sci. Technol. 2006, 40, 6969-6975.

18. Wania, F., A global mass balance analysis of the source of perfluorocarboxylic acids in the Arctic Ocean. Environ. Sci. Technol. 2007, 41, 4529-4535.

19. Grondin, J.; Sagnes, R.; Commeyras, A., Perfluorosulfonic acids. 3. Hammett acidity functions of perfluoroalkanesulfonic acids and of their mixtures with SBF5. B. Soc. Chim. Fr. 1976, 1779-1782.

20. Goss, K.-U., The pKa values of PFOA and other highly fluorinated carboxylic acids. Environ. Sci. Technol. 2008, 42, 456-458.

21. Brace, N. O., Long chain alkanoic and alkenoic acids with perfluoroalkyl terminal segments. J. Org. Chem. 1962, 27, 4491-4498.

22. Stock, N. L.; Furdui, V. I.; Muir, D. C. G.; Mabury, S. A., Perfluoroalkyl contaminants in the Canadian Arctic: Evidence of atmospheric transport and local contamination. Environ. Sci. Technol. 2007, 41, 3529-3536.

23. Butt, C. M.; Mabury, S. A.; Kwan, M.; Wang, X.; Muir, D. C. G., Spatial trends of perfluoroalkyl compounds in ringed seals (Phoca hispida) from the Canadian Arctic. Environ. Toxicol. Chem. 2008, 27, 542-553.

24. Wallington, T. J.; Hurley, M. D.; Xia, J.; Wuebbles, D. J.; Sillman, S.; Ito, A.; Penner, J. E.; Ellis, D. A.; Martin, J.; Mabury, S. A.; Nielsen, O. J.; Sulbaek Andersen, M. P., Formation of C7F15COOH (PFOA) and other perfluorocarboxylic acids during the atmospheric oxidation of 8:2 fluorotelomer alcohol. Environ. Sci. Technol. 2006, 40, 924-930.

107 25. Ellis, D. A.; Martin, J. W.; Mabury, S. A.; Hurley, M. D.; Sulbaek Andersen, M. P.; Wallington, T. J., Atmospheric lifetime of fluorotelomer alcohols. Environ. Sci. Technol. 2003, 37, 3816-3820.

26. Young, C. J.; Furdui, V. I.; Franklin, J.; Koerner, R. M.; Muir, D. C. G.; Mabury, S. A., Perfluorinated acids in Arctic snow: New evidence for atmospheric formation. Environ. Sci. Technol. 2007, 41, 3455-3461.

27. Theobald, N.; Gerwinski, W.; Caliebe, C.; Haarich, M. Development and validation of a method for the determination of polyfluorinated organic substances in sea water, sediments and biota. Occurrence of these compound in the North and Baltic Seas.; Report 41/07, ISSN-1862 4804. http://www.umweltdaten.de/publikationen/fpdf-l/3319.pdf; Umweltbundesamt: 2007.

28. Smithwick, M.; Norstrom, R. J.; Mabury, S. A.; Solomon, K.; Evans, T. J.; Stirling, I.; Taylor, M. K.; Muir, D. C. G., Temporal trends of perfluoroalkyl contaminants in polar bears (Ursus maritimus) from two locations in the North American arctic, 1972-2002. Environ. Sci. Technol. 2006, 40, 1139-1143.

29. DuPont Global PFOA Strategy - Comprehensive Source Reduction; U.S. Environmental Protection Agency public docket AR226-1914; U.S. EPA: Washington, D.C., January 31, 2005.

30. DuPont Global PFOA Strategy - Comprehensive Source Reduction; presented to the U.S. EPA OPPT, January 31, 2005; U.S. Environmental Protection Agency public docket AR226-1914; U.S. EPA: Washington, D.C. In.

31. DuPont Global PFOA Strategy - Comprehensive Source Reduction. U.S. Environmental Protection Agency public docket AR226-1914. U.S. EPA: Washington, D.C.; U.S. Environmental Protection Agency public docket AR226-1914.

32. Schenker, U.; Scheringer, M.; MacLeod, M.; Martin, J. W.; Cousins, I. T.; Hungerbuhler, K., Contribution of volatile precursor substances to the flux of perfluorooctanoate to the Arctic. Environ. Sci. Technol. 2008, 42, 3710-3716.

33. Dupont Zonyl RP paper fluoro-protectant; technical information.

34. Masurf FS-115; technical information.

35. D'eon, J. C.; Mabury, S. A., Production of perfluorinated carboxylic acids (PFCAs) from the biotransformation of polyfluoroalkyl phosphate surfactants (PAPs): Exploring routes of human contamination. Environ. Sci. Technol. 2007, 41, 4799-4805.

36. Shoeib, M.; Harner, T.; Vlahos, P., Perfluorinated chemicals in the arctic atmosphere. Environ. Sci. Technol. 2006, 40, 7577-7583.

37. Muir, D.; Köck, G.; Gantner, K.; Iqaluk, D.; Iqaluk, P.; Wang, X.; Evans, M.; Reist, J.; Babaluk, J.; Sardella, G.; Sverko, E.; Barresi, E., Temporal trends of persistent organic pollutants and mercury in landlocked char in the High Arctic. In Synopsis of research conducted under the 2007-2008 Northern Contaminants Program, Indian and Northern Affairs Canada: Ottawa, ON, Canada, 2008.

108 38. Butt, C. M.; Muir, D. C. G.; Stirling, I.; Kwan, M.; Mabury, S. A., Rapid response of Arctic ringed seals to changes in perfluoroalkyl production. Environ. Sci. Technol. 2007, 41, 42- 49.

39. Hart, K.; Gill, V. A.; Kannan, K., Temporal trends (1992-2007) of perfluorinated chemicals in Northern Sea Otters (Enhydra lutris kenyoni) from south-central Alaska. Arch. Environ. Contam. Toxicol. 2009, 56, 607-614.

40. Rosenberg, B.; DeLaronde, J.; MacHutchon, A.; Stern, G.; Spencer, C.; Scott, B.; Lopez, E.; Muir, D. C. G.; Tomy, G. T., Spatial and vertical distribution of perfluorinated compounds in Canadian Arctic and sub-arctic ocean water. Organohalogen Compounds 2008, 70, 386-389.

41. Yamashita, N.; Taniyasu, S.; Petrick, G.; Wei, S.; Gamo, T.; Lam, P. K. S.; Kannan, K., Perfluorinated acids as novel chemical tracers of global circulation of ocean waters. Chemosphere 2008, 70, 1247-1255.

42. McMurdo, C. J.; Ellis, D. A.; Webster, E.; Butler, J.; Christensen, R. D.; Reid, L. K., Aerosol enrichment of the surfactant PFO and mediation of the water-air transport of gaseous PFOA. Environ. Sci. Technol. 2008, 42, 3969-3974.

43. Cheng, J.; Psillakis, E.; Hoffmann, M. R.; Colussi, A. J., Acid dissociation versus molecular association of perfluoroalkyl oxoacids: Environmental implications. J. Phys. Chem. A 2009, 113, 8152-8156.

44. Kim, S.-K.; Kannan, K., Perfluorinated acids in air, rain, snow, surface runnoff, and lakes: Relative importance of pathways to contamination of urban lakes. Environ. Sci. Technol. 2007, 41, 8328-8334.

45. Arp, H. P. H.; Goss, K.-U., Irreversible sorption of trace concentrations of perfluorocarboxylic acid to fiber filters used for air sampling. Atmos. Environ. 2008, 42, 6869- 6872.

46. Tomy, G. T.; Budakowski, W.; Halldorson, T.; Helm, P. A.; Stern, G. A.; Friesen, K.; Pepper, K.; Tittlemier, S. A.; Fisk, A. T., Fluorinated organic compounds in an eastern arctic marine food web. Environ. Sci. Technol. 2004, 38, 6475-6481.

47. Powley, C. R.; George, S. W.; Russell, M. H.; Hoke, R. A.; Buck, R. C., Polyfluorinated chemicals in a spatially and temporally integrated food web in the Western Arctic. Chemosphere 2008, 70, 664-672.

48. Haukås, M.; Berger, U.; Hop, H.; Gulliksen, B.; Gabrielsen, G. W., Bioaccumulation of per- and polyfluorinated alkyl substances (PFAS) in selected species from the Barents Sea food web. Environ. Pollut. 2007, 148, 360-371.

49. Key, B. D.; Howell, R. D.; Criddle, C. S., Defluorination of organofluorine sulfur compounds by Pseudomonas sp. strain D2. Environ. Sci. Technol. 1998, 32, 2283-2287.

50. Bossi, R.; Riget, F. F.; Dietz, R.; Sonne, C.; Fauser, P.; Dam, M.; Vorkamp, K., Preliminary screening of perfluorooctane sulfonate (PFOS) and other fluorochemicals in fish,

109 birds and marine mammals from Greenland and the Faroe Islands. Environ. Pollut. 2005, 136, 323-329.

51. Bossi, R.; Riget, F. F.; Dietz, R., Temporal and spatial trend of perfluorinated compounds in ringed seal (Phoca hispida) from Greenland. Environ. Sci. Technol. 2005, 39, 7416-7422.

52. Verreault, J.; Berger, U.; Gabrielsen, G. W., Trends of perfluorinated alkyl substances in herring gull eggs from two coastal colonies in northern Norway: 1983-2003. Environ. Sci. Technol. 2007, 41, 6671-6677.

53. Löfstrand, K.; Jörundsdóttir, H.; Tomy, G.; Svavarsson, J.; Weihe, P.; Nygård, T.; Bergman, Å., Spatial trends of polyfluorinated compounds in guillemot (Uria aalge) eggs from North-Western Europe. Chemosphere 2008, 72, 1475-1480.

54. Smithwick, M.; Mabury, S. A.; Solomon, K. R.; Sonne, C.; Martin, J. W.; Born, E. W.; Dietz, R.; Derocher, A. E.; Letcher, R. J.; Evans, T. J.; Gabrielsen, G. W.; Nagy, J.; Stirling, I.; Taylor, M. K.; Muir, D. C. G., Circumpolar study of perfluoroalkyl contaminants in polar bears (Ursus maritimus). Environ. Sci. Technol. 2005, 39, 5517-5523.

55. Tomy, G.; Pleskach, K.; Ismail, N.; Stern, G., Time-trend on Hendrickson Island beluga and spatial-trend studies on Hendrickson Island, Kimmirut, Arviat, Sanikiluaq and Pangnirtung beluga for new and emerging persistent halogenated compounds. In Synopsis of research conducted under the 2005-2006 Northern Contaminants Program, Indian and Northern Affairs Canada: Ottawa, ON, 2006.

56. Stern, G. A.; Tomy, G., Temporal trend studies of trace metals and halogenated organic compounds (HOCs), including new and emerging persisstent compounds, in Mackenzie River burbot, Fort Good Hope, NWT. In Synopsis of research conducted under the 2006-2007 Northern Contaminants Program, Indian and Northern Affairs Canada: Ottawa, ON, 2007.

57. Evans, M. S.; Muir, D.; Low, G.; Boucher, M.; Krieger, M.; Boniface, R.; Letcher, L.; Keating, J.; Wang, X., Spatial and long-term trends in persistent organic contaminants and metals in lake trout and burbot in Great Slave Lake, NT. In Synopsis of research conducted under the 2005-2006 Northern Contaminants Program, Indian and Northern Affairs Canada: Ottawa, ON, 2006.

58. Tomy, G.; Pleskach, K.; Rosenberg, B.; Stern, G., Temporal trends of emerging halogenated persistent compounds in belgua from Hendrickson Island and Pangnirtung. In Synopsis of research conducted under the 2007-2008 Northern Contaminants Program, Indian and Northern Affairs Canada: Ottawa, ON, 2008.

59. Butt, C. M.; Mabury, S. A.; Muir, D. C. G.; Braune, B. M., Prevalence of long-chained perfluorinated carboxylates in seabirds from the Canadian Arctic between 1975 and 2004. Environ. Sci. Technol. 2007, 41, 3521-3528.

60. Dietz, R.; Bossi, R.; Riget, F.; Sonne, C.; Born, E. W., Increasing perfluoroalkyl contaminants in East Greenland polar bears (Ursus maritimus): A new toxic threat to the arctic bears. Environ. Sci. Technol. 2008, 42, 2701-2707.

110 61. Macdonald, R. W.; Barrie, L. A.; Bidleman, T. F.; Diamond, M. L.; Gregor, D. J.; Semkin, R. G.; Strachan, W. M. J.; Li, Y. F.; Wania, F.; Alaee, M.; Alexeeva, L. B.; Backus, S. M.; Bailey, R.; Bewers, J. M.; Gobeil, C.; Halsall, C. J.; Harner, T.; Hoff, J. T.; Jantunen, L. M. M.; Lockhart, W. L.; Mackay, D.; Muir, D. C. G.; Pudykiewicz, J.; Reimer, K. J.; Smith, J. N.; Stern, G. A.; Schroeder, W. H.; Wagemann, R.; Yunker, M. B., Contaminants in the Canadian Arctic: 5 years of progress in understanding sources, occurrence and pathways. Sci. Total Environ. 2000, 254, 93-234.

62. Jones, E. P.; Anderson, L. G.; Swift, J. H., Distribution of Atlantic and Pacific waters in the upper Arctic Ocean: Implication for circulation. Geophys. Res. Lett. 1998, 25, 765-768.

63. Jones, E. P.; Swift, J. H.; Anderson, L. G.; Lipizer, M.; Civitarese, G.; Falkner, K. K.; Kattner, G.; McLaughlin, Tracing Pacific water in the North Atlantic Ocean. J. Geophys. Res. 2003, 108(C4), doi:10.1029/2001JC001141.

64. Tomy, G.; Tittlemier, S. A.; Palace, V. P.; Budakowski, W. R.; Braekevelt, E.; Brinkworth, L.; Friesen, F., Biotransformation of N-ethyl perfluorooctanesulfonamide by rainbow trout (Onchorohynchus mykiss) liver microsomes. Environ. Sci. Technol. 2004, 38, 758-762.

65. Xu, L.; Krenitsky, D. M.; Seacat, A. M.; Butenhoff, J. L.; Anders, M. W., Biotransformation of N-ethyl-N(2-hydroxyethyl)perfluorooctanesulfonamide by rat liver microsomes, cytosol, and slices and by expressed rat and human cytochromes P450. Chem. Res. Toxicol. 2004, 17, 767-775.

66. Martin, J. W.; Smithwick, M. M.; Braune, B. M.; Hoekstra, P. F.; Muir, D. C. G.; Mabury, S. A., Identification of long-chain perfluorinated acids in biota from the Canadian Arctic. Environ. Sci. Technol. 2004, 38, 373-380.

67. Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G., Dietary accumulation of perfluorinated acids in juvenile rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 189-195.

68. Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G., Bioconcentration and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 196-204.

69. Luebker, D. J.; Hansen, K. J.; Bass, N. M.; Butenhoff, J. L.; Seacat, A. M., Interactions of fluorochemicals with rat liver fatty acid-binding protein. Toxicology 2002, 176, 175-185.

70. Chen, Y.-M.; Guo, L.-H., Fluorescence study on site-specifc binding of perfluoroalkyl acids to human serum albumin. Arch. Toxicol. 2009, 83, 255-261.

71. Verreault, J.; Houde, M.; Gabrielsen, G. W.; Berger, U.; Haukås, M.; Letcher, R. J.; Muir, D. C. G., Perfluorinated alkyl substances in plasma, liver, brain, and eggs of glaucous gulls (Larus hyperboreus) from the Norwegian Arctic. Environ. Sci. Technol. 2005, 39, 7439- 7445.

72. Muir, D.; Alaee, M.; Butt, C.; Braune, B.; Helm, P.; Mabury, S.; Stock, N.; Tomy, G.; Wang, X., New contaminants in Arctic biota. In Synopsis of research conducted under the 2003-

111 2004 Northern Contaminants Program, Indian and Northern Affairs Canada: Ottawa, ON, Canada, 2004; pp 139-148.

73. Kannan, K.; Corsolini, S.; Falandysz, J.; Oehme, G.; Focardi, S.; Giesy, J. P., Perfluorooctanesulfonate and related fluorinated hydrocarbons in marine mammals, fishes, and birds from coasts of Baltic and the Mediterranean Seas. Environ. Sci. Technol. 2002, 36, 3210- 3216.

74. Hart, K.; Kannan, K.; Isobe, T.; Takahashi, S.; Yamada, T. K.; Miyazaki, N.; Tanabe, S., Time trends and transplacental transfer of perfluorinated compounds in melon-headed whales stranded along the Japanese coast in 1982, 2001/2002, and 2006. Environ. Sci. Technol. 2008, 42, 7132-7137.

75. Xu, L.; Krenitsky, D. M.; Seacat, A. M.; Butenhoff, J. L.; Tephly, T. R.; Anders, M. W., N-glucuronidation of perfluorooctanesulfonamide by human, rat, dog, and monkey liver microsomes and by expressed rat and human UDP-glucuronosyltransferases. Drug Metab. Dispos. 2006, 34, (1406-1410).

76. Hagen, D. F.; Belisle, J.; Johnson, J. D.; Venkateswarlu, P., Characterization of fluorinated metabolites by a gas chromatographic-helium microwave plasma detector - the biotransformation of 1H,1H,2H,2H-perfluorodecanol to perfluorooctanoate. Anal. Biochem. 1981, 118, 336-343.

77. Martin, J. W.; Mabury, S. A.; O'Brien, P. J., Metabolic products and pathways of fluorotelomer alcohols in isolated rat hepatocytes. Chem. Biol. Interact. 2005, 155, 165-180.

78. Kudo, N.; Iwase, Y.; Okayachi, H.; Yamakawa, Y.; Kawashima, Y., Induction of hepatic peroxisome proliferation by 8-2 telomer alcohol feeding in mice: Formation of perfluorooctanoic acid in the liver. Toxicol. Sci. 2005, 86, 231-238.

79. Nabb, D. L.; Szostek, B.; Himmelstein, M. W.; Mawn, M. P.; Gargas, M. L.; Sweeney, L. M.; Stadler, J. C.; Buck, R. C.; Fasano, W. J., In vitro metabolism of 8-2 fluorotelomer alcohol: Interspecies comparisons and metabolic pathway refinement. Toxicol. Sci. 2007, 100, 333-344.

80. Fasano, W. J.; Carpenter, S. C.; Gannon, S. A.; Snow, T. A.; Stadler, J. C.; Kennedy, G. L.; Buck, R. C.; Korzeniowski, S. H.; Hinderliter, P. M.; Kemper, R. A., Absorption, distribution, metabolism, and elimination of 8-2 fluorotelomer alcohol in the rat. Toxicol. Sci. 2006, 91, 341-355.

81. Lei, Y. D.; Wania, F.; Mathers, D.; Mabury, S. A., Determination of vapor pressures, octanol-air, and water-air partition coefficients for polyfluorinated sulfonamide, sulfonamidoethanols, and telomer alcohols. J. Chem. Eng. Data 2004, 49, 1013-1022.

82. Tittlemier, S. A.; Pepper, K.; Edwards, L.; Tomy, G., Development and characterizationof a solvent-extraction-gas chromatographic/mass spectrometric method for the analysis of perfluorooctanesulfonamide compounds in solid matrices. J. Chromatogr. A 2005, 1066, 189-195.

112 83. Nye miljøgifter i luft - 2007. Bromerte flammehemmere og perfluorerte stoffer i luft. OR28/2008. Norwegian Institute for Air Research. ISBN 978-82-425-1984-9. Available at: www.nilu.no.

84. Nye miljøgifter i luft - 2007. Bromerte flammehemmere og perfluorerte stoffer i luft. OR16/2008. Norwegian Institute for Air Research. ISBN 978-82-425-1962-7. Available at: www.nilu.no.

85. De Silva, A. O.; Muir, D. C. G.; Mabury, S. A., Distribution of perfluorocarboxylate isomers in select samples from the North American environment. Environ. Toxicol. Chem. 2009, 28, 1801-1814.

86. Kallenborn, R.; Berger, U.; Järnberg, U.; Dam, M.; Glesne, O.; Hedlund, B.; Hirvi, J.-P.; Lundgren, A.; Mogensen, B. B.; Sigurdsson, A. S. Perfluorinated Alkylated Substances (PFAS) in the Nordic Environment.; Nordic Council of Ministers.: 2004.

87. Saez, M.; Moreno, D. V.; Begoña, J.; van Leeuwen, S., Uncommon PFC-profile in arctic ice samples from Russia. Organohalogen Compounds 2008, 70, 1870-1873.

88. Kallenborn, R., Berger, U., Järnberg, U., Dam, M., Glesne, O., Hedlund, B., Hirvi, J.-P., Lundgren, A., Mogensen, B.B., Sigurdsson, A.S. 2004. Perfluorinated Alkylated Substances (PFAS) in the Nordic Environment. Nordic Council of Ministers.

89. Chang, S.-C.; Das, K.; Ehresman, D. J.; Ellefson, M. E.; Gorman, G. S.; Hart, J. A.; Noker, P. E.; Tan, Y.-M.; Lieder, P. H.; Lau, C.; Olsen, G. W.; Butenhoff, J. L., Comparative pharmacokinetics of perfluorobutyrate in rats, mice, monkeys, and humans and relevance to human exposure via drinking water. Toxicol. Sci. 2008, 104, 40-53.

90. Johannessen, O. M.; Bengtsson, L.; Miles, M. W.; Kuzmina, S. I.; Semenov, V. A.; Alekseev, G. V.; Nagurnyi, A. P.; Zakharov, V. F.; Bobylev, L. P.; Pettersson, L. H.; Hasselmann, K.; Cattle, H. P., Arctic climate change: Observed and modelled temperature and sea-ice variability. Tellus A 2004, 56, 328-341.

91. Graversen, R. G.; Mauritsen, T.; Tjernström, M.; Källén, E.; Svensson, G., Vertical structure of recent Arctic warming. Nature 2008, 541, 53-56.

92. Parmesan, C.; Yohe, G., A globally coherent fingerprint of climate change impacts across natural systems. Nature 2003, 421, 37-42.

93. Post, E.; Forchhammer, M. C.; Bret-Harte, M. S.; Callaghan, T. V.; Christensen, T. R.; Elberling, B.; Fox, A. D.; Gilg, O.; Hik, D. S.; Høye, T. T.; Ims, R. A.; Jeppesen, E.; Klein, D. R.; Madsen, J.; McGuire, A. D.; Rysgaard, S.; Schindler, D. E.; Stirling, I.; Tamstorf, M. P.; Tyler, N. J. C.; Van Der Wal, R.; Welker, J.; Wookey, P. A.; Schmidt, N. M.; Aastrup, P., Ecological dynamics across the arctic associated with recent climate change. Science 2009, 325, 1355-1358.

94. Macdonald, R. W.; Harner, T.; Fyfe, J., Recent climate change in the Arctic and its impact on contaminant pathways and interpretation of temporal trend data. Sci. Total Environ. 2005, 342, 5-86.

CHAPTER THREE

Thesis Goals and Hypothesis

113 114 Chapter Three – Thesis Goals and Hypothesis

Perfluorinated alkyl compounds (PFCs) are ubiquitously detected in humans and wildlife, including remote regions such as the Canadian Arctic. The two dominant classes of PFCs detected in biota are the perfluorinated carboxylates (PFCAs) and sulfonates (PFSAs). However, the sources of these compounds to the environment remains unclear. With the exception of PFOA and PFOS, and to a lesser degree PFNA, these chemicals were not directly produced in large volumes. These compounds are primarily degradation products of intermediates used in the manufacture of commercial fluorinated products.

Two mechanisms have been postulated to explain the global dissemination of PFCAs and PFSAs. One mechanism (“direct sources”) is through the direct release of PFCAs and PFSAs during product manufacture, use and disposal. The PFCAs and PFSAs are quite water soluble and can be transported throughout the globe through oceanic currents. The other mechanism (“indirect sources”) proposes that precursor compounds may ultimately form PFCAs and PFSAs through atmospheric and biological reactions. As discussed in the introductory chapters, our research group and several others have investigated various atmospheric and biological reactions that contribute to the indirect sources of PFCAs and PFSAs.

The overall aim of this thesis was to further our understanding of the sources of PFCAs and PFSAs to humans and wildlife. To achieve this aim, four specific goals were developed.

It has been shown that the atmospheric oxidation of some volatile intermediates used in the manufacture of fluorinated commercial products results in the formation of PFCAs and PFSAs. These compounds include the fluorotelomer alcohols, iodides and olefins, which yield the PFCAs. However, it is not known if the fluorotelomer acrylate monomer (FTAc), the key building block in fluorotelomer-based polymers, can also form PFCAs through atmospheric oxidation. Therefore, the goal of Chapter 4 was to investigate the potential of the FTAc to act as a PFCA precursor through atmospheric degradation. In addition, kinetics with respect to reaction chlorine atom and hydroxyl radical were measured, with the objective of calculating atmospheric lifetime. Further, the reaction products were investigated with the objective of understanding the atmospheric degradation mechanism.

115 As thoroughly described in Chapter Two, the arctic environment is widely contaminated with PFCs. In fact, of all global regions, the arctic is probably the most extensively studied. In particular, this includes the Canadian Arctic. It is apparent that additional general monitoring studies will contribute little scientific value to our understanding of PFCs sources to remote environments. However, monitoring studies can be designed in an appropriate and strategic manner to investigate PFC transport pathways, processes and dynamics. As such, three monitoring studies were performed with this goal in mind. Chapters 5 and 6 investigated the PFC temporal trends in two arctic animals: ringed seals and seabirds (northern fulmars and thick-billed murres). The response of animals to changes in PFC production (i.e. 3M’s phase- out of POSF-chemistry in 2002) can provide insight as to the dominant transport mechanism to the arctic. The animals serve as biomonitors for general changes in the environmental levels of contaminants. Chapter 7 investigated the spatial trends of PFCs in ringed seals in 11 populations across the Canadian Arctic. Differences in PFC levels and patterns across the region can be indicative of different source regions (i.e. Asian versus North American).

The biotransformation of telomer-based compounds has been well studied with the majority of the research on microbial systems and mammalian models, using the 8:2 fluorotelomer alcohol (8:2 FTOH) as the substrate. There has been very little attention to fluorotelomer biotransformation in aquatic species such as fish. Whole animal experiments show that the majority of the dosed 8:2 FTOH is excreted as products of phase II metabolites (e.g. glucuronide, sulfate and glutathione conjugates). There is presently no knowledge on the fate of other fluorotelomer compounds, such as the FTAc monomer. Therefore, the third goal of this thesis was to investigate the bioaccumulation and biotransformation of the 8:2 FTAc in aquatic species, using rainbow trout as the animal model. This goal was achieved through complimentary in vivo and in vitro experiments described in Chapter 8. The objective of the in vivo experiment was to investigate the bioaccumulation of the 8:2 FTAc, identify and quantify biotransformation products and examine elimination kinetics. It was hypothesized that the 8:2 FTAc would initially undergo phase I metabolism to yield the 8:2 FTOH which would subsequently degrade through oxidative reactions to yield the fluorotelomer intermediates and PFCAs. The in vitro experiments were performed with sub-cellular fractions isolated from the liver and stomach tissues with the objective of assessing the overall rate of 8:2 FTAc metabolism via esterase enzymes, as well as comparing the relative kinetics between the tissues. Although the metabolism of 8:2 FTOH has been well studied, there exists a considerable

116 difference in the postulated biotransformation mechanisms. Therefore, the fourth goal of this thesis was to further elucidate the 8:2 FTOH biotransformation mechanism. To achieve this goal whole rainbow trout with exposed via the diet to three key intermediates in the 8:2 FTOH metabolism – 8:2 fluorotelomer saturated and unsatured carboxylates (8:2 FTCA and 8:2 FTUCA) and the 7:3 FTCA. These results of this experiment are presented in Chapter 9.

CHAPTER FOUR

Atmospheric Chemistry of 4:2 Fluorotelomer Acrylate (C4F9CH2CH2OC(O)CH=CH2): Kinetics, Mechanisms and Products of Chlorine Atom and OH Radical Initiated Oxidation

Craig M. Butt, Cora J. Young, Scott A. Mabury, Michael D. Hurley, Timothy J. Wallington

Published In: Journal of Physical Chemistry A 2009, 113, 3155-3161

Contributions: Smog chamber experiments were performed by Craig Butt with the assistance of Cora Young and Michael Hurley. Data interpretation and manuscript preparation was conducted by Craig Butt with critical comments provided by Cora Young, Michael Hurley and Timothy Wallington. The research and manuscript preparation was conducted under the guidance of Scott Mabury.

Reproduced with permission from Journal of Physical Chemistry A Copyright American Chemical Society 2009

117 118 Chapter Four – Atmospheric Chemistry of 4:2 Fluorotelomer Acrylate (C4F9CH2CH2OC(O)CH=CH2): Kinetics, Mechanisms and Products of Chlorine Atom and OH Radical Initiated Oxidation

4.1. Abstract

Relative rate techniques were used to measure k(Cl + C4F9CH2CH2OC(O)CH=CH2) = -10 -11 3 (2.21 ± 0.16) x 10 and k(OH + C4F9CH2CH2OC(O)CH=CH2) = (1.13 ± 0.12) x 10 cm -1 -1 molecule s in 700 Torr of N2 or air diluent at 296K. The atmospheric lifetime of

C4F9CH2CH2OC(O)CHCH2 (4:2 FTAc) is determined by its reaction with OH radicals and is approximately 1 day. The chlorine atom initiated oxidation of 4:2 FTAc in 700 Torr of air at

296 K gives C4F9CH2C(O)H in molar yields of 18% and 10% in the absence and presence of NO, respectively. The OH radical initiated oxidation of 4:2 FTAc in 700 Torr in the presence of

NO gives HCHO in a molar yield of 102 ± 7% with C4F9CH2CH2OC(O)C(O)H (4:2 fluorotelomer glyoxylate) as the expected co-product. The atmospheric fate of the 4:2 fluorotelomer glyoxylate will be photolysis and reaction with OH radicals which will lead to the formation of C4F9CH2C(O)H and ultimately perfluorinated carboxylic acids. The atmospheric oxidation of fluorotelomer acrylates is a potential source of perfluorinated carboxylic acids in remote locations.

119 4.2. Introduction

Perfluorinated carboxylic acids (PFCAs) are global environmental pollutants and are distributed widely in humans [1, 2], wildlife [3, 4], and abiotic media such as seawater [5]. Despite their apparently ubiquitous presence, the sources of PFCA contamination are not well understood, particularly in remote environments. It has been suggested that an indirect mechanism may explain the widespread distribution of PFCAs in the environment. The indirect mechanism involves the release and transport of volatile precursor compounds that degrade via atmospheric oxidation to PFCAs which subsequently undergo wet and dry deposition. Smog chamber studies in our laboratory and elsewhere have shown that oxidation of fluorotelomer alcohols (FTOHs) [6, 7], fluorosulfonamido alcohols [8, 9], fluorotelomer olefins [10, 11] and fluorotelomer iodides [12] in low NOx environments forms PFCAs.

Alternatively, it has been suggested that the observation of PFCAs in marine wildlife in remote locations reflects historical direct emission of PFCAs, mainly PFOA, during fluorochemical manufacturing processes and as residuals in consumer products [13], followed by global transport in ocean currents [14, 15]. Empirical support for the indirect mechanism comes from the fact that deposition fluxes extrapolated from arctic surface snow measurements[16] are consistent with those estimated from atmospheric models [17]. The relative importance of the indirect and direct routes is unclear but is likely to depend on the location, PFCA, and time period under consideration.

Fluorotelomer acrylates (general formula: CxF2x+1CH2CH2OC(O)CH=CH2, FTAc) are monomers used in the manufacture of fluorotelomer-based polymers. Fluorotelomer-based polymers have wide commercial uses, primarily for their stain repellency. It has been reported that FTAc-based polymers are the largest commercial category of polyfluorinated products [18]. Fluorotelomer acrylates have been shown to be significant residuals in a FTAc-based polymer [18], and may be released to the atmosphere from commercial fluorotelomer-based polymer products in a similar manner to FTOHs [19]. The 6:2, 8:2 and 10:2 FTAc have been detected in the atmosphere at generally low (<1-6 pg/m3, but up to 3000 pg/m3 near a suspected source in Japan) levels [20-22].

120 There are no reported studies of the atmospheric chemistry of FTAcs and the contribution of these compounds to PFCA burdens observed in remote locations is unclear. Given their commercial importance and presence in the atmosphere, information on the atmospheric chemistry of fluorotelomer acrylates is needed. The goal of the present work was to improve our understanding the atmospheric chemistry of FTAcs using

C4F9CH2CH2OC(O)CHCH2 (4:2 FTAc) as a representative compound. We report here the results of studies of the kinetics and mechanisms of the chlorine atom and OH radical initiated oxidation of 4:2 FTAc.

4.3. Experimental Section

4.3.1. Kinetic Experiments Experiments were performed in a 140 liter Pyrex reactor interfaced to a Mattson Sirus 100 FTIR spectrometer [23]. The reactor was surrounded by 22 fluorescent blacklamps (GE F15T8-BL), which were used to photochemically initiate the experiments. Chlorine atoms were produced by photolysis of molecular chlorine:

Cl2 + hv → Cl + Cl (1)

OH radicals were produced by the photolysis of CH3ONO in air:

CH3ONO + hv → CH3O + NO (2)

CH3O + O2 → HO2 + HCHO (3)

HO2 + NO → OH + NO2 (4)

In the relative rate experiments the following reactions take place:

Cl + reactant → products (5) Cl + reference → products (6) OH + reactant → products (7) OH + reference → products (8)

121 Assuming that the reactant and reference compounds are lost solely via reaction with Cl atoms or OH radicals and that neither the reactant, nor reference, are formed in any processes, then the decays of the reactant and reference can be plotted using the expression,

⎛[reactant]to ⎞ kreactant ⎛ ][reference to ⎞ ln⎜ ⎟ = ln ⎜ ⎟ (9) ⎝ [reactant]t ⎠ kreference ⎝ ][reference t ⎠

where [reactant]t0, [reactant]t, [reference]t0, [reference]t are the concentrations of reactant and reference at time “t0” and “t”, kreactant, kreference are the rate constants for reaction of Cl atoms or

OH radicals with the reactant and reference. Plots of ln([reactant]t0/[reactant]t) versus ln([reference]t0/[reference]t) should be linear, pass through the origin and have a slope of kreactant/kreference. Unless stated otherwise, quoted uncertainties represent the precision of the measurements and include two standard deviations from regression analyses and uncertainties in the IR analysis (typically ±1% of the initial reactant concentrations).

CH3ONO was synthesized by the drop-wise addition of concentrated sulfuric acid to a saturated solution of NaNO2 in methanol. The reagent, 4:2 FTAc, was obtained from a commercial source (Oakwood Products Inc., 97%, West Columbia, SC) and subject to repeated freeze/pump/thaw cycling to remove volatile impurities before use. Experiments were conducted in 700 Torr total pressure of N2 or air diluent at 296 K. Ultra-high-purity synthetic air and nitrogen from Michigan Airgas were used as diluent gases. Concentrations of reactants and products were monitored by FTIR spectroscopy. IR spectra were derived from 32 coadded interferograms with a spectral resolution of 0.25 cm-1 and an analytical path length of 27.1 m.

4.3.2. Product and Mechanistic Experiments Products of the atmospheric oxidation of 4:2 FTAc with Cl atoms and OH radicals were monitored using FTIR spectroscopy in 700 Torr of air diluent. Mixtures used to study the Cl atom initiated oxidation consisted of either 7.1 mTorr 4:2 FTAc and 105 mTorr Cl2, or 7.4 mTorr 4:2 FTAc, 101 mTorr Cl2, and 57 mTorr NO. Mixtures used to study the OH radical initiated oxidation consisted of 3.8-10.6 mTorr 4:2 FTAc, 27-100 mTorr CH3ONO, and 8.4-15 mTorr NO. An additional experiment was conducted using deuterated methyl nitrite to investigate the mechanism of OH oxidation. In these experiments we used mixtures consisting of 7.7 mTorr 4:2 FTAc, 27 mTorr CH2DONO, and 28 mTorr NO.

122 4.4. Results and Discussion

4.4.1. Kinetics of the Cl + 4:2 FTAc reaction The rate of reaction (10) was measured relative to reactions (11) and (12):

Cl + 4:2 FTAc → products (10)

Cl + C2H4 → products (11)

Cl+ C3H6 → products (12)

Reaction mixtures consisted of 7.29 - 7.48 mTorr 4:2 FTAc, 98.4 - 101.5 mTorr Cl2, and either 7.35 - 7.39 mTorr C2H4, or 5.67 - 7.18 mTorr C3H6, in 700 Torr of N2 diluent. The observed loss of 4:2 FTAc versus those of the reference compounds is plotted in Figure 4.1.

Linear least squares analysis of the data in Figure 4.1 gives k10/k11 = 2.36 ± 0.15 (n=16) and -11 -10 k10/k12 = 0.84 ± 0.05 (n=18). Using k11 = 9.29 x 10 [24] and k12 = 2.64 x 10 [25] we derive -10 -10 3 -1 -1 k10 = (2.19 ± 0.14) x 10 and (2.22 ± 0.13) x 10 cm molecule s . We choose to cite a final value which is the average of the individual determinations together with error limits which -10 3 -1 encompass the extremes of the determinations, hence k10 = (2.21 ± 0.16) x 10 cm molecule s-1. We estimate that uncertainties in the reference rate constants contribute an additional approximately 10% uncertainty to k10.

While there have been no previous studies of k10, we can compare our result with the reactions of the corresponding fluorotelomer alcohol, iodide, and olefin: k(Cl + -11 -12 C4F9CH2CH2OH) = (1.61 ± 0.49) x 10 [26], k(Cl + C4F9CH2CH2I) = (1.25 ± 0.15) x 10 -11 3 -1 -1 [12], k(Cl + C4F9CH=CH2) = (8.9 ± 1.0) x 10 cm molecule s [10]. The acrylate is more reactive than the alcohol and iodide by factors of 14 and 180, respectively. By analogy to unsaturated hydrocarbons [27] we expect that the reaction of Cl atoms with the acrylate will proceed predominately via addition to the >C=C< double bond and will be more rapid than H- abstraction from the –CH2- groups in the alcohol and iodide. The fluorotelomer acrylate is approximately 2.5 times more reactive than the fluorotelomer olefin. The reactions of chlorine atoms with the acrylate and olefin proceed via electrophilic addition to the >C=C< double bond.

The electron withdrawing effect of the C4F9- group is deactivating and will reduce the reactivity of both molecules towards Cl atoms. The ester functional group of the 4:2 FTAc increases the

123 distance between fluorinated tail and the double bond, mitigating the influence of the fluorinated tail.

4.4.2. Kinetics of the OH + 4:2 FTAc reaction The rate of reaction (13) was measured relative to reactions (14) and (15):

OH + 4:2 FTAc → products (13)

OH + C2H4 → products (14)

OH + C3H6 → products (15)

Initial reaction mixtures consisted of 6.46 - 7.52 mTorr 4:2 FTAc, 98.1 - 106.7 mTorr

CH3ONO, and either 4.31 mTorr C2H4 or 0.31 - 0.47 mTorr C3H6 in 700 Torr total pressure of air diluent. Figure 4.2 shows the loss of 4:2 FTAc plotted versus loss of the reference compounds. The lines through the data in Figure 4.2 are linear least squares fits which give k13/k14 = 1.37 ± 0.10 (n=8) and k13/k15 = 0.42 ± 0.03 (n=16). Using recommended values of k14 -12 -11 -11 = 8.52 x 10 [28] and k15 = 2.63 x 10 [28] we derive k13 = (1.17 ± 0.08) x 10 and k13 = (1.10 ± 0.09) x 10-11 cm3 molecule-1 s-1. We choose to report an average with uncertainties - which encompass the extremes of the individual determinations, hence, k13 = (1.13 ± 0.12) x 10 11 cm3 molecule-1 s-1. We estimate that uncertainties in the reference rate constants contribute an additional approximately 10% uncertainty to k13.

While there have been no previous studies of k13, we can compare our result with the reactions of the corresponding fluorotelomer alcohol, iodide, and olefin: k(OH + -12 -12 C4F9CH2CH2OH) = (1.2 ± 0.6) x 10 [26], k(OH + C4F9CH2CH2I) = (1.17 ± 0.57) x 10 [12], -12 3 -1 -1 and k(OH + C4F9CH=CH2) = (1.3 ± 0.2) x 10 cm molecule s [10]. With respect to reaction with OH radicals, the acrylate is approximately an order of magnitude more reactive than the alcohol, iodide, and olefin. As was the case for chlorine atoms, the reaction of OH radicals with the acrylate is expected to proceed predominately via electrophilic addition to the >C=C< double bond. Addition to >C=C< double bonds proceeds faster than H-abstraction from

–CH2- groups [28]. The ester functional group shields the double bond from the deactivating effect of the C4F9- group. The observation that the fluorotelomer acrylate is more reactive than the alcohol, iodide, and olefin is consistent with expectations.

124

We can also compare our measured value of k(OH + C4F9CH2CH2OC(O)CH=CH2) = (1.13 ± 0.12) x 10-11 to kinetic data for analogous non-halogenated unsaturated esters, k(OH + -11 CH3OC(O)CH=CH2) = (1.3 ± 0.1) x 10 [29], k(OH + CH3CH2OC(O)CH=CH2) = (1.6 ± 0.2) x -11 -11 3 -1 -1 10 [29] and k(OH + CH3CH2CH2OC(O)CH=CH2) = (1.8 ± 0.3) x 10 cm molecule s

[30]. The reactivity of C4F9CH2CH2OC(O)CH=CH2 is indistinguishable, within the experimental uncertainties, from that of CH3OC(O)CH=CH2 suggesting that the electron withdrawing effect of the C4F9- group on the reactivity of OH radicals towards the double bond is small.

125

2.5

C2H4 ) t 2.0

1.5 /[4:2 FTAc] t0

C3H6 1.0

0.5 Ln ([4:2 FTAc]

0.0 0.0 0.5 1.0 1.5 2.0 2.5 Ln ([Reference] /[Reference] ) t0 t

Figure 4.1. Loss of 4:2 FTAc versus C2H4 (●) and C3H6 (▲) following UV irradiation of 4:2

FTAc/reference/Cl2 mixtures in 700 Torr of N2.

126

2.0

) C H

t 2 4 1.5 /[4:2 FTAc]

t0 1.0

C3H6 0.5 Ln ([4:2 FTAc]

0.0 0.0 0.5 1.0 1.5 2.0 2.5 Ln ([Reference] /[Reference] ) t0 t

Figure 4.2. Loss of 4:2 FTAc versus C2H4 (●) and C3H6 (▲) following UV irradiation of 4:2

FTAc/reference/CH3ONO mixtures in 700 Torr of air.

127 4.4.3. Products and Mechanism of the Cl + 4:2 FTAc reaction in the presence and absence of NOx Figure 4.3 shows IR spectra acquired before (A) and after (B) 85 second UV irradiation of a mixture of 7.1 mTorr 4:2 FTAc and 105 mTorr Cl2 in 700 Torr of air diluent. Subtraction of IR features attributable to 4:2 FTAc from panel B gives the residual spectrum in panel C. A reference spectrum of C4F9CH2C(O)H (4:2 fluorotelomer aldehyde, 4:2 FTAL) is shown in panel D. Comparing panel D with panel C shows that C4F9CH2C(O)H is formed as a product in the system. The formation of C4F9CH2C(O)H is plotted versus the loss of 4:2 FTAc as the circles in Figure 4.4. As seen from Figure 4.4, the amount of C4F9CH2C(O)H in the chamber increased linearly with 4:2 FTAc loss consistent with its formation as a primary product. The

C4F9CH2C(O)H formation yield was 18 ± 0.6% (n=6) and 10 ± 0.9% (n=7) in the absence and presence of NO, respectively. Subtraction of IR features attributable to C4F9CH2C(O)H from the product spectrum (C) gives the residual spectrum (E). The residual product spectra did not change shape significantly throughout the time course of the experiment, suggesting the formation of one, or more, stable product(s). The magnitude of the residual absorption increased linearly with the consumption of 4:2 FTAc suggesting that the compound(s) responsible for the absorption feature shown in panel (E) is a primary product(s).

As discussed in section 4.4.1, the bulk of the reaction of chlorine atoms with

C4F9CH2CH2OC(O)CH=CH2 is expected to proceed via addition to the >C=C< double bond (Figure 4.5). Addition of chlorine to the terminal carbon leads to a more stable radical and is thermodynamically favored. Reaction of the carbon radical with oxygen yields the chloroperoxy radical followed by reaction with NO or RO2 to yield the chloroalkoxy radical shown in Figure 4.5.

There are three potential fates of the chloroalkoxy radical intermediate. In pathway “A”

(see Figure 4.5) the chloromethyl radical, CH2Cl, is eliminated giving

C4F9CH2CH2OC(O)C(O)H (4:2 fluorotelomer glyoxylate, 4:2 FTGly). In 700 Torr of air chloromethyl radicals react to give HC(O)Cl which has a characteristic absorption feature at 1783 cm-1. Examination of the product spectra revealed that only trace quantities (approximately 3% yield) of HC(O)Cl were formed. We conclude that pathway A is not of major importance.

128 Decomposition via pathway “C” (see Figure 4.5) gives chloroacetaldehyde,

CH2ClC(O)H. Features attributable to this product were sought but not found in the product spectra and we conclude that pathway C is not of major importance. Having excluded pathways A and C as playing major roles, we conclude that reaction with oxygen (pathway “B”) is the main fate of the chloroalkoxy radical intermediate formed via addition of Cl to 4:2 FTAc. This behavior is similar to those of α-chloro alkoxy radicals formed following reaction of chlorine atoms with other alkenes (e.g., ethene [31]). We conclude that the residual product spectrum in panel E of Figure 4.3 reflects the formation of C4F9CH2CH2OC(O)C(O)CH2Cl in the system.

The reaction of chlorine atoms with 4:2 FTAc may also occur by H-abstraction from the

C4F9CH2CH2OC(O)- portion of the compound (pathway “D”). Abstraction most likely occurs from the carbon adjacent to the ester function. Reaction of this radical intermediate, with oxygen followed by NO or RO2 will give the alkoxy radical, C4F9CH2C(O•)HOC(O)CH=CH2.

The likely fate of this alkoxy radical would be decomposition to yield C4F9CH2C(O)H and α- ester rearrangement to give C4F9CH2C(O) radicals and the acid HOC(O)CH=CH2. A reference spectrum for HOC(O)CH=CH2 was not available and thus it was not possible to determine if this compound was formed in our experiments. The observed formation of C4F9CH2C(O)H in yields of 18% and 10% in the absence and presence of NO, respectively, may reflect the contribution of pathway D.

Finally, we note that the above is a simplified discussion of the mechanism of the chlorine atom initiated oxidation of 4:2 FTAc. In the absence of NO the self- and cross- reactions of peroxy radicals will be important. In addition to the radical channel of these reactions leading to alkoxy radicals there will be molecular channels which will give a variety of alcohol and carbonyl compounds [32]. Furthermore, HO2 radicals formed in the system (e.g., in pathway B, see Figure 4.5) will react with peroxy radicals to give hydroperoxides [32]. In the presence of NO, the HO2 radicals will be converted into OH radicals and the oxidation of 4:2 FTAc will be initiated by a mixture of Cl atom and OH radical attack. The decreased yield of

C4F9CH2C(O)H in the presence of NO probably reflects the contribution of OH radical initiated oxidation. The goal of the present study was to elucidate the atmospheric oxidation mechanism of fluorotelomer acrylates. In light of the complexity and the minor atmospheric importance of chlorine atom initiated oxidation, further experiments and analysis were not pursued.

129

0.4

0.2 A: Before Irradiation

0.0 0.4

0.2 B: 85s Irradiation

0.0

0.2 C: B - 0.24*A

0.0

0.2 Absorbance D: C4F9CH2C(O)H

0.0

0.2 E: C - 0.35*D

0.0

1500 1600 1700 1800 1900 2000 -1 Wavenumber (cm )

Figure 4.3. FTIR spectra obtained before (A) and after (B) 85 seconds UV irradiation of a mixture of 7.1 mTorr 4:2 FTAc and 105 mTorr Cl2 in 700 Torr of air diluent. Panel C shows the product spectrum obtained by subtracting the IR features of the reactant from the spectrum in B.

Panel D is a reference spectrum of C4F9CH2C(O)H. Panel E is the product spectrum obtained by subtracting IR features of C4F9CH2C(O)H from the spectrum in C.

130

1.0 1.2 4:2 FTAL (without NO) Unknown (without NO) 1.0 0.8 HC(O)Cl (without NO) 4:2 FTAL (with NO) Unknown (with NO) 0.8 0.6

0.6

0.4 0.4

0.2 [Unknown] (arbitary units) 0.2 [4:2 FTAL], [HC(O)Cl] (mTorr)

0.0 0.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0

[Δ 4:2 FTAc] (mTorr)

Figure 4.4. Yield of C4F9CH2C(O)H (4:2 FTAL), unknown residual product (panel “E” in figure 4.3) and HC(O)Cl versus depletion of 4:2 FTAc after irradiation of a mixture of 7.1 mTorr 4:2 FTAc and 105 mTorr Cl2 in 700 Torr of air diluent (identified as “without NO”) and

7.4 mTorr of 4:2 FTAc, 101 mTorr Cl2 and 57 mTorr NO in 700 Torr of air diluent (identified as “with NO”). Lines are linear regressions to the 4:2 FTAL data.

131 O Cl Cl

CH2 H O C4F9 CH2 O + • C H C O • CH H H O C4F9 CH2 O 4:2 fluorotelomer glyoxylate

D A

O O Cl + O2, CH2 + Cl• CH2 H C4F9 CH2 O + NO, -NO2 C4F9 CH2 O H 4:2 FTOH Acrylate O• B + O2, -HO2 C

O Cl O CH2 H O C4F9 CH2 O CH2 C • H + Cl C4F9 CH2 O O H C H H

Figure 4.5. Simplified mechanism of Cl atom oxidation of 4:2 FTAc.

132

4.4.4. Products and Mechanism of the OH + 4:2 FTAc reaction in the presence of NOx To investigate the products and mechanism of OH radical initiated oxidation a series of experiments were conducted using 4:2 FTAc/CH3ONO/NO/air mixtures. The UV irradiation of these mixtures lead to the formation of one, or more, carbonyl containing product(s) with a broad absorption feature centered at approximately 1750 cm-1. The absorption feature did not match those for any compounds in our reference library. However, it is possible to formulate a reaction mechanism based our findings and those from other laboratories [30, 33].

As discussed in section 4.4.2, the reaction of OH radicals with 4:2 FTAc proceeds mainly via addition to the >C=C< double bond. Addition to the terminal carbon leads to a more stable radical and is thermodynamically favored. Addition of oxygen to the carbon radical gives a hydroxy peroxy radical which then reacts with NO to give the hydroxy alkoxy radical shown in Figure 4.6.

As with the chloro alkoxy radical discussed in the previous section, there are three potential fates of the hydroxy alkoxy radical intermediate. Elimination of a CH2OH radical in pathway A (see Figure 4.6) gives the 4:2 fluorotelomer glyoxylate. The CH2OH radical will then react with O2 to give HCHO. Reaction with oxygen in pathway B gives the hydroxyl dicarbonyl compound shown in Figure 4.6. Finally, decomposition via the C-C bond scission shown in pathway C gives glycolaldehyde (HOCH2CHO) and the C4F9CH2CH2OC(O) radical.

Using our calibrated reference spectrum of HOCH2CHO we searched for IR product features that could be attributed to glycolaldehyde, but none were found. Similarly, we looked for characteristic -OH stretch features in the 3500-3230 cm-1 region in the product spectra which could be attributed to hydroxyl dicarbonyl compound, but none were found. We conclude that pathways B and C (see Figure 4.6) are not significant.

The formation of HCHO is a marker for the importance of pathway A. The chemical system used to generate OH radicals (photolysis of CH3ONO) generates substantial amounts of HCHO (via reactions 2 and 3, see section 4.3.1) and complicates quantification of HCHO formation during 4:2 FTAc oxidation. To investigate the formation of HCHO during the oxidation of 4:2 FTAc an additional series of experiments was performed using deuterated methyl nitrite (CH2DONO) as a source of OH radicals. The formaldehyde originating from the

CH2DONO is mainly HCDO:

133

CH2DONO + hv → CH2DO + NO (16)

CH2DO + O2 → HCDO + HO2 (17)

CH2DO + O2 → HCHO + DO2 (18)

The branching ratio k17:k18 is 0.88:0.12 (± 0.011) [34]. The observed formation of

HCDO was combined with the rate constant ratio k17:k18 to calculate, and correct for, the HCHO resulting from photolysis of CH2DONO.

HCHO reacts with OH radicals,

HCHO + OH → products (19)

-12 3 -1 - The rate constant for reaction 19 has been determined; k19 = 8.5 x 10 cm molecule s 1 [35].

The concentration profile of HCHO can be described[36] by the expression

[HCHO] α −− xx kk −1)13/19( − ]1)1)[(1( = (I) FTA c] 2:[4 FTAc] o − kk 1319 )]/(1[

where x = 1 – ([4:2 FTAc]/[4:2 FTAc]o) is the fractional consumption of 4:2 FTAc and α is the molar yield of HCHO in the OH radical initiated oxidation of 4:2 FTAc. Figure 4.7 shows a plot of [HCHO]/[4:2 FTAc]o versus Δ[4:2 FTAc]/[4:2 FTAc]o. The best fit line gives α = 1.02 -12 ± 0.07 and k19/k13 = 0.70 ± 0.05. We can combine k19/k13 = 0.70 ± 0.07 with k19 = 8.5 x 10 to -11 3 -1 -1 give a value of k13 = (1.21 ± 0.12) x 10 cm molecule s which is indistinguishable from that derived in section 4.4.2. The fact that the HCHO yield is indistinguishable from 100% indicates that nitrate formation in the RO2 + NO reactions in the system is not substantial and we conclude that the oxidation of 4:2 FTAc can be represented in atmospheric models by expression (20):

OH + C4F9CH2CH2OC(O)CH=CH2 → C4F9CH2CH2OC(O)CHO + HCHO (20)

134

This conclusion is consistent with the recent finding by Blanco and Teruel [30] that butyl glyoxylate is the main product of the OH initiated oxidation of butyl acrylate.

O OH O CH2 H C4F9 CH2 O + • C H C H

H H O 4:2 fluorotelomer glyoxylate

A

O O OH + O2, CH2 + OH• CH2 H C4F9 CH2 O + NO, -NO2 C4F9 CH2 O H 4:2 FTOH Acrylate O• B + O2, -HO2 C

O OH O CH2 H O C4F9 CH2 O CH2 C • H + OH C4F9 CH2 O O H C H H

Figure 4.6. Simplified mechanism of OH radical oxidation of 4:2 FTAc.

135

0.5

0.4 o

0.3

0.2 [HCHO] / [4:2 FTAc] 0.1

0.0 0.0 0.2 0.4 0.6 0.8 1.0 Δ [4:2 FTAc] / [4:2 FTAc] o

Figure 4.7. Formation of HCHO versus loss of 4:2 FTAc, normalized to the initial 4:2 FTAc concentration following the UV irradiation of 4:2 FTAc/CH2DONO/NO mixtures in 700 Torr of air at 296 K. The curve is a fit of equation I to the data.

136 4.5. Implications for Atmospheric Chemistry

4.5.1. Atmospheric lifetime The atmospheric oxidation of 4:2 FTAc is initiated via reaction with OH radicals. Chlorine atoms are present in the atmosphere at levels typically several orders of magnitude lower than those for OH and chlorine atom initiated oxidation will not be significant. Loss by photolysis is considered negligible since unsaturated esters do not photolyze in the actinic region [29]. Assuming that 4:2 FTAc and methyl acrylate [37] have a similar reactivity towards -18 3 -1 -1 ozone we can estimate k(O3 + 4:2 FTAc) = 1 x 10 cm molecule s . Taking a typical value of [O3] = 50 ppb in urban areas then the lifetime of 4:2 FTAc with respect to reaction with ozone is approximately 9 days. From k(OH + 4:2 FTAc) = 1.1 x 10-11 cm3 molecule-1 s-1 (section 4.4.2) and an OH radical concentration of 1 x 106 molecules cm-3 [38], the atmospheric lifetime of 4:2 FTAc with respect to reaction with OH radicals is estimated to be approximately 1 day. We conclude that the atmospheric lifetime of 4:2 FTAc is determined by reaction with OH radicals and is approximately 1 day. It should be noted that the lifetime of 4:2 FTAc will vary seasonally and geographically with local OH radical concentrations.

The reactivity towards OH radicals, and hence the atmospheric lifetime of FTOHs, is not dependent on chain length [26]. It seems reasonable to assume that the same will be true for fluorotelomer acrylates and the kinetic data reported here for 4:2 FTAc will be applicable to the more commercially relevant longer chain fluorotelomer acrylates of the general formula,

CxF2x+1CH2CH2OC(O)CH=CH2. Our results show that fluorotelomer acrylates will be oxidized into fluorotelomer glyoxalates close to the emission sources on a time scale of approximately 1 day.

A recent analysis of residuals from a commercial fluorotelomer-based polymer showed similar levels of FTAcs and FTOHs [18]. Residuals from commercial products may be significant emission sources of FTAcs and FTOHs in urban areas. The limited available data suggest that the concentrations of FTAcs in the atmosphere are substantially lower than those of FTOHs [20-22]. The low levels of FTAcs are consistent with their short atmospheric lifetimes and may also reflect lower rates of emission for these compounds. Further work is needed to better define the emission fluxes of fluorotelomer compounds into the atmosphere.

137 4.5.2. Contribution to PFCA burden in remote locations

The atmospheric oxidation of fluorotelomer acrylates (CxF2x+1CH2CH2OC(O)CH=CH2) is initiated by reaction with OH radicals and gives fluorotelomer glyoxylates

(CxF2x+1CH2CH2OC(O)C(O)H, FTGly) in a molar yield which is indistinguishable from 100%. The atmospheric fate of fluorotelomer glyoxylates will be photolysis and reaction with OH radicals. It is expected that the reaction of OH radicals with fluorotelomer glyoxylates will proceed predominately by hydrogen abstraction from the aldehyde group and hence the rate constant is probably similar to that for reaction with CH3C(O)C(O)H. Further work is needed to confirm this expectation, especially in light of the approximately linear yield plot of methyl glyoxylate in the chlorine atom initiated oxidation of methyl propionate reported by Cavalli et al. [39] which suggests that glyoxylates may be somewhat less reactive than glyoxals. -11 3 Proceeding on the assumption that k(OH + FTGly) = k(OH + CH3C(O)C(O)H) = 1.5 x 10 cm molecule-1 s-1 [28] and taking a diurnal average [OH] = 106 molecule cm-3 leads to an estimate of 18 hours for lifetime of FTGly with respect to reaction with OH radicals. The dicarbonyl o chromophores in FTGly and CH3C(O)C(O)H are similar. For a solar zenith angle of 45 the -4 -1 photolysis rate of CH3C(O)C(O)H is 10 s at 0.5 km altitude for a cloudless sky [28]. This photolysis rate is equivalent to a lifetime of 3 hours. It is probable that FTGlys have a similarly short lifetime with respect to photolysis. We conclude that photolysis is probably the major atmospheric fate of FTGlys. By analogy to CH3C(O)C(O)H [28], photolysis will occur via rupture of the CxF2x+1CH2CH2OC(O)–C(O)H bond. Photolysis will initiate a sequence of reactions leading to the formation of C4F9CH2CHO:

C4F9CH2CH2OC(O)C(O)H + hν → C4F9CH2CH2OC(O) + HCO (21)

C4F9CH2CH2OC(O) + O2 → C4F9CH2CH2OC(O)OO (22)

C4F9CH2CH2OC(O)OO + NO → → C4F9CH2CH2O+ CO2 + NO2 (23)

C4F9CH2CH2O + O2 → C4F9CH2CHO + HO2 (24)

Whether initiated by photolysis, or via reaction with OH, the oxidation of FTGlys will proceed to give fluorotelomer aldehydes (CxF2x+1CH2CHO) on a time scale of probably less than

1 day. As discussed in detail previously [17], the oxidation of CxF2x+1CH2CHO proceeds on a time scale of approximately 6 days [40, 41] and gives CxF2x+1CHO which in turn is oxidized to

COF2 (on a time scale of approximately 1-2 days [42]) as the major product with

138 perfluorocarboxylic acids as minor products (in approximately 1-10% molar yield) [17]. Hence, the atmospheric oxidation of fluorotelomer acrylates is expected to lead to the formation of perfluorocarboxylic acids in an approximately 1-10% molar yield. The time scale of conversion of FTAcs into PFCAs is sufficiently slow (approximately 10 days) to allow transport of FTAcs and their oxidation products over large distances (3500 km at the global average wind speed of approximately 4 m s-1). We show here that the atmospheric degradation of FTAcs has the potential to contribute to the observed burden of PFCA pollution in remote locations. To assess the magnitude of this contribution estimates for the flux of FTAcs into the atmosphere are required. Further work is needed to provide such estimates.

4.6. Acknowledgement Financial support provided by the Natural Sciences and Engineering Research Council (NSERC) of Canada (Mabury). C.M.B. also appreciates the support of NSERC through a Post- Graduate Scholarship.

139 4.7. Literature Cited

1. Hansen, K. J.; Clemen, L. A.; Ellefson, M. E.; Johnson, H. O., Compound-specific, quantitative characterization of organic fluorochemicals in biological matrices. Environ. Sci. Technol. 2001, 35, 766-770.

2. Olsen, G. W.; Huang, H.-Y.; Helzlsouer, K. J.; Hansen, K. J.; Butenhoff, J. L.; Mandel, J. H., Historical comparison of perfluorooctanesulfonate, perfluorooctanoate, and other fluorochemicals in human blood. Environ. Health Perspect. 2005, 113, 539-545.

3. Giesy, J. P.; Kannan, K., Distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 2001, 35, 1339-1342.

4. Houde, M.; Martin, J. W.; Letcher, R. J.; Solomon, K.; Muir, D. C. G., Biological monitoring of polyfluoroalkyl substances: A review. Environ. Sci. Technol. 2006, 40, 3463- 3473.

5. Yamashita, N.; Taniyasu, S.; Petrick, G.; Wei, S.; Gamo, T.; Lam, P. K. S.; Kannan, K., Perfluorinated acids as novel chemical tracers of global circulation of ocean waters. Chemosphere 2008, 70, 1247-1255.

6. Ellis, D. A.; Martin, J. W.; De Silva, A. O.; Mabury, S. A.; Hurley, M. D.; Sulbaek Andersen, M. P.; Wallington, T. J., Degradation of fluorotelomer alcohols: A likely atmospheric source of perfluorinated carboxylic acids. Environ. Sci. Technol. 2004, 38, 3316-3321.

7. Hurley, M. D.; Wallington, T. J.; Sulbaek Andersen, M. P.; Ellis, D. A.; Martin, J. W.; Mabury, S. A., Atmospheric chemistry of fluorinated alcohols: reaction with Cl atoms and OH radicals and atmospheric lifetimes. J. Phys. Chem. A 2004, 108, 1973-1979.

8. Martin, J. W.; Ellis, D. A.; Mabury, S. A.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of perfluoroalkanesulfonamides: Kinetic and product studies of the OH radical and Cl atom initiated oxidation of N-ethyl perfluorobutanesulfonamide. Environ. Sci. Technol. 2006, 40, 846-872.

9. D'eon, J. C.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A., Atmospheric chemistry of N-methyl perfluorobutane sulfonamidoethanol, C4F9SO2N(CH3)CH2CH2OH: Kinetics and mechanism of reaction with OH. Environ. Sci. Technol. 2006, 40, 1862-1868.

10. Vésine, E.; Bossoutrot, V.; Mellouki, A.; Le Bras, G.; Wenger, J.; Sidebottom, H., Kinetic and mechanistic study of OH- and Cl-initiated oxidation of two unsaturated HFCs: C4F9CH=CH2 and C6F13CH=CH2. J. Phys. Chem. A 2000, 104, 8512-8520.

11. Nakayama, T.; Takahashi, K.; Matsumi, Y.; Toft, A.; Sulbaek Andersen, M. P.; Nielsen, O. J.; Waterland, R. L.; Buck, R. C.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of CF3CH=CH2 and C4F9CH=CH2: Products of the gas-phase reactions with Cl atoms and OH radicals. J. Phys. Chem. A 2007, 111, 909-915.

12. Young, C. J.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A., Atmospheric chemistry of 4:2 fluorotelomer iodide (n-C4F9CH2CH2I) and 6:2 fluorotelomer iodide (n-C6F13CH2CH2I):

140 Kinetics and products of photolysis and reaction with OH radicals and Cl atoms. J. Phys. Chem. A 2008, 112, 13542-13548.

13. Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H., Sources, fate and transport of perfluorocarboxylates. Environ. Sci. Technol. 2006, 40, 32-44.

14. Armitage, J.; Cousins, I. T.; Buck, R. C.; Prevedouros, K.; Russell, M. H.; MacLeod, M.; Korzeniowski, S. H., Modeling global-scale fate and transport of perfluorooctanoate emitted from direct sources. Environ. Sci. Technol. 2006, 40, 6969-6975.

15. Wania, F., A global mass balance analysis of the source of perfluorocarboxylic acids in the Arctic Ocean. Environ. Sci. Technol. 2007, 41, 4529-4535.

16. Young, C. J.; Furdui, V. I.; Franklin, J.; Koerner, R. M.; Muir, D. C. G.; Mabury, S. A., Perfluorinated acids in Arctic snow: New evidence for atmospheric formation. Environ. Sci. Technol. 2007, 41, 3455-3461.

17. Wallington, T. J.; Hurley, M. D.; Xia, J.; Wuebbles, D. J.; Sillman, S.; Ito, A.; Penner, J. E.; Ellis, D. A.; Martin, J.; Mabury, S. A.; Nielsen, O. J.; Sulbaek Andersen, M. P., Formation of C7F15COOH (PFOA) and other perfluorocarboxylic acids during the atmospheric oxidation of 8:2 fluorotelomer alcohol. Environ. Sci. Technol. 2006, 40, 924-930.

18. Russell, M. H.; Berti, W. R.; Szostek, B.; Buck, R. C., Investigation of the biodegradation potential of a fluoroacrylate polymer product in aerobic soils. Environ. Sci. Technol. 2008, 42, 800-807.

19. Dinglasan-Panlilio, M. J. A.; Mabury, S. A., Significant residual fluorinated alcohols present in various fluorinated materials. Environ. Sci. Technol. 2006, 40, 1447-1453.

20. Piekarz, A. M.; Primbs, T.; Field, J. A.; Barofsky, D. F.; Simonich, S., Semivolatile fluorinated organic compounds in Asian and western U.S. air masses. Environ. Sci. Technol. 2007, 41, 8248-8255.

21. Oono, S.; Matsubara, E.; Harada, K. H.; Takagi, S.; Hamada, S.; Asakawa, A.; Inoue, K.; Watanabe, I.; Koizumi, A., Survery of airborne polyfluorinated telomers in Keihan area, Japan. Bull. Environ. Contam. Toxicol. 2008, 80, 102-106.

22. Dreyer, A.; Temme, C.; Sturm, R.; Ebinghaus, R., Optimized method avoiding solvent- induced response enhancement in the analysis of volatile and semi-volatile polyfluorinated alkylated compounds using gas chromatography-mass spectrometry. J. Chromatogr. A 2008, 1178, 199-205.

23. Wallington, T. J.; Japar, S. M., Fourier transform infrared kinetic studies of the reaction of HONO with HNO3, NO3 and N2O5 at 295 K. J. Atmos. Chem. 1989, 9, 399-409.

24. Wallington, T. J.; Andino, J. M.; Lorkovic, I. M.; Kaiser, E. W.; Martson, G., Pressure dependence of the reaction of chlorine atoms with ethene and acetylene in air at 295K. J. Phys. Chem. 1990, 94, 3644-3648.

141 25. Ezell, M. J.; Wang, W.; Ezell, A. A.; Soskin, G.; Finlayson-Pitts, B. J., Kinetics of reactions of chlorine atoms with a series of alkenes at 1 atm and 298 K: Structure and reactivity. Phys. Chem. Chem. Phys 2002, 4, 5813-5820.

26. Ellis, D. A.; Martin, J. W.; Mabury, S. A.; Hurley, M. D.; Sulbaek Andersen, M. P.; Wallington, T. J., Atmospheric lifetime of fluorotelomer alcohols. Environ. Sci. Technol. 2003, 37, 3816-3820.

27. Ragains, M. L.; Finlayson-Pitts, B. J., Kinetics and mechanism of the reaction of Cl atoms with 2-methyl-1,3-butadiene (isoprene) at 298K. J. Phys. Chem. A 1997, 101, 1509-1517.

28. Calvert, J. G.; Atkinson, R.; Kerr, J. A.; Madronich, S.; Moortgat, G. K.; Wallington, T. J.; Yarwood, G., The mechanisms of atmospheric oxidation of the alkenes. Oxford University Press: Oxford, 2000.

29. Teruel, M. A.; Lane, S. I.; Mellouki, A.; Solignac, G.; Le Bras, G., OH reaction rate constants and UV absorption cross-sections of unsaturated esters. Atmos. Environ. 2006, 40, 3764-3772.

30. Blanco, M. B.; Teruel, M. A., Photodegradation of butyl acrylate in the troposphere by OH radicals: Kinetic and fate of 1,2-hydroxyalcoxy radicals. J. Phys. Org. Chem. 2008, 21, 397- 401.

31. Orlando, J. J.; Tyndall, G. S.; Wallington, T. J., The atmospheric chemistry of alkoxy radicals. Chem. Rev. 2003, 103, 4657-4689.

32. Wallington, T. J.; Dagaut, P.; Kurylo, M. J., Ultraviolet absorption cross sections and reaction kinetics and mechanisms for peroxy radicals in the gas phase. Chem. Rev. 1992, 92, 667-710.

33. Saunders, S. M.; Baulch, D. L.; Cooke, K. M.; Pilling, M. J.; Smurthwaite, P. I., Kinetics and mechanisms of the reactions of OH with some oxygenated compounds of importance in tropospheric chemistry. Int. J. Chem. Kinet. 1994, 26, 113-130.

34. Nilsson, E. J. K.; Johnson, M. S.; Taketani, F.; Matsumi, Y.; Hurley, M. D.; Wallington, T. J., Atmospheric deuterium fractionation: HCHO and HCDO yields in the CH2DO + O2 reaction. Atmos. Chem. Phys. 2007, 7, 5873-5881.

35. Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, J., R.F.; Kerr, J. A.; Rossi, M. J.; Troe, J., Evaluated kinetic and photochemical data for atmospheric chemistry: Supplement VII, Organic Species. J. Phys. Chem. Ref. Data 1999, 28, 191-393.

36. Meagher, R. J.; McIntosh, M. E.; Hurley, M. D.; Wallington, T. J., A kinetic study of the reaction of chlorine and fluorine atoms with HC(O)F at 295 +/- 2K. Int. J. Chem. Kinet. 1997, 29, 619-625.

37. Grosjean, E.; Grosjean, D., Rate constants for the gas phase reaction of ozone with unsaturated oxygenates. Int. J. Chem. Kinet. 1998, 30, 21-29.

142 38. Prinn, R. G.; Huang, J.; Weiss, R. F.; Cunnold, D. M.; Fraser, P. J.; Simmonds, P. G.; McCulloch, A.; Harth, C.; Salameh, P.; O'Doherty, S.; Wang, R. H. J.; Porter, L.; Miller, B. R., Evidence for substantial variations of atmospheric hydroxyl radicals in the past two decades. Science 2001, 292, 1882-1888.

39. Cavalli, F.; Barnes, I.; Becker, K. H.; Wallington, T. J., Atmospheric oxidation mechanism of methyl propionate. J. Phys. Chem. A 2000, 104, 11310-11317.

40. Sellevåg, S. R.; Kelly, T.; Sidebottom, H.; Nielsen, C. J., A study of the IR and UV-Vis absorption cross-sections, photolysis and OH-initiated oxidation of CF3CHO and CF3CH2CHO. Phys. Chem. Chem. Phys 2004, 6, 1243-1252.

41. Chiappero, M. S.; Argüello, G. A.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of C8F17CH2CHO: Yield from C8F17CH2CH2OH (8:2 FTOH) oxidation, kinetics and mechanisms of reactions with Cl atoms and OH radicals. Chem. Phys. Lett. 2008, 461, 198-202.

42. Chiappero, M. S.; Malanca, F. E.; Arguello, G. A.; Wooldridge, S. T.; Hurley, M. D.; Ball, J. C.; Wallington, T. J.; Waterland, R. L.; Buck, R. C., Atmospheric chemistry of perfluoroaldehydes (CxF2x+1CHO) and fluorotelomer aldehydes (CxF2x+1CH2CHO): Quantification of the important role of photolysis. J. Phys. Chem. A. 2006, 110, (43), 11944- 11953.

CHAPTER FIVE

Rapid Response of Arctic Ringed Seals to Changes in Perfluoroalkyl Production

Craig M. Butt, Derek C.G. Muir, Ian Stirling, Michael Kwan and Scott A. Mabury

Published In: Environmental Science and Technology 2007, 41, 42-49.

Contributions: Data analysis and interpretation was conducted by Craig Butt. All versions of the manuscript were prepared by Craig Butt with critical comments provided by Derek Muir and Scott Mabury.

Reproduced with permission from Environmental Science and Technology Copyright American Chemical Society 2007

143 144 Chapter Five – Rapid Response of Arctic Ringed Seals to Changes in Perfluoroalkyl Production

5.1. Abstract 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 & 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 LC-MS/MS. C9-C15 PFCAs showed statistically significant increasing concentrations 1992-2005 and 1993-2005 at Arviat and Resolute Bay, respectively. Doubling times ranged from 19-16 years for PFDoA to 10-8 years for PFNA at Arviat and Resolute Bay, but were shorter when excluding the 2005 samples. Conversely, 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 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.

145 5.2. Introduction Perfluoroalkyl compounds (PFCs) have recently garnered intense scientific and regulatory interest due to their widespread detection in fish and wildlife (1,2) and humans (3,4), and their potential toxicological effects (5). Two such classes of PFCs include the perfluorinated carboxylic acids (PFCAs) and 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 (US EPA) launched a global stewardship program to reduce PFOA emissions and its presence in products by 95 percent 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 the PFCAs may also reach the Arctic through oceanic transport (13). The detection of only linear isomers of PFOA and >99% linear isomers of PFNA and 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). PFOSF based 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

146 manufacturers and production volumes have increased ~2-fold to 11-14 x 106 kg yr-1 from 2000- 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 PFOS (annual increase = 4.7-8.2%), PFDA (1.7-3.3%) and PFUnA (5.9-6.8%) between 1982-2003 (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 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 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, Nunavut and Arviat,

Nunavut. Seal liver samples were analyzed for C7-C15 PFCAs, the suspected PFCA precursors

8:2 & 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.

5.3. Materials and Methods

5.3.1. 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 13 13 13 13 (Oakville, ON, Canada). Stable isotope standards ( C4-PFOA , C5-PFNA, C2-PFDA, C4- 13 13 PFOS, 8:2 C2-FTUCA, 10:2 C2-FTUCA) were provided by Wellington Laboratories 13 (Guelph, Ontario). C2-PFOA was purchased from PerkinElmer Life and Analytical Sciences Canada.

147 5.3.2. 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 (74o42’N, 94o49’W) in 2000 (n=9), 2004 (n=9) and 2005 (n=10) and Arviat, Nunavut (61o7’N, 94o4’W) in 1998 (n=10), 2003 (n=10) and 2005 (n=10). All individuals were collected during the annual spring hunt, from approximately May-June. Samples collected in 2000-2005 were shipped to the Nunavik Research Centre (Kuujjuaq) for sub-sampling and tooth aging. Archived livers tissues from Resolute Bay: 1972 (n=2), 1993 (n=9) and Arviat 1992 (n=6) were generously provided by the National Wildlife Research Centre - Canadian Wildlife Service Specimen Bank, Ottawa, ON, Canada. Liver samples were shipped to the Canada Centre for Inland Waters (Burlington, ON, Canada) and stored at -20 oC. Ages were determined by longitudinal thin sectioning a lower canine tooth and counting the annual growth layers in the dentine using transmitted light (23).

5.3.3. Extraction and Clean-up Methods Liver samples were liquid-liquid extracted using methods similar to Hansen et al. (3) with the addition of a novel clean-up step (24). The purpose of the fluorosolvent clean-up 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 clean-up step as compared to the traditional methods (24). Sub-samples (~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 MTBE for 5 min, and then centrifuged for 10 min (3300 rpm). The organic layer was transferred to a clean centrifuge tube; liver homogenate extracted again, centrifuged and organic layers combined. The combined MTBE extract was reduced to 0.5 ml, an equal volume of hexafluoropropanol (HFP) was added, solution shaken for 1 min and 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 30s and filtered. Internal standards were added immediately prior to analysis.

148 5.3.4. 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 x 2.1 mm, 3 μm particle size, Aberdeen, UK), preceded by a C18 guard column (4.0 x 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 (method 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, UK) with the mobile phases delivered at 300 μl/min using a Waters 600S controller and samples injected (20 μl) with a Waters 717 plus autosampler (Waters, Milford, USA). 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 oC. 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 & 2005 and all Arviat samples) was run with method B.

To compare the two methods, a sub-set of the extracts analyzed in method A (n=11, including 2-3 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.

149 5.3.5. Statistical Analyses and Data Treatment The instrument detection limits (IDL) were defined as concentration that produced a peak with a signal-to-noise ratio of three. Method detection limits (MDL) were defined as the mean blank response (procedural blank, 1 ml of 18MΩ water carried through extraction procedure) plus three times the standard deviation of the blank response. For calculating the MDL, ND values were replaced by one-half of the IDL. Concentrations that were below the method detection limit (MDL) were reported as

Prior to statistical analysis, data normality was tested using the Shapiro-Wilk test and homogeneity of variance was tested using the Levene’s test. Data were natural-logarithm transformed prior to statistical analysis to meet assumptions of normality and homogeneity of variances. After transformation, ANOVA tests were conducted to determine statistically significant differences between the different years for each analyte separately. Tukey’s honestly significant difference (HSD) test was used post-hoc to assess differences between the years (SPSS for Windows, 2001, Chicago, Illinois). Due to the small sample size of the 1972 Resolute Bay data set (n=2), these samples were not included in the ANOVA and post-hoc tests.

Sex was known for 33 out 36 individuals and 37 out of 39 individuals in the Arviat and Resolute Bay samples, respectively. Specifically, male: female counts were 1M:5F (1992), 2M:7F (1998), 2M:7F (2003) and 3M:5F (2005) for Arviat; and were 8M:1F (1993), 8M:1F (2000), 6M:3F (2004) and 7M:3F (2005) for Resolute Bay. Sex was not known for the 1972 Resolute Bay samples. Within each year, differences between sexes were assessed using the t- test. Age was known for the 1992, 1998 and 2003 Arviat samples, and 2000 and 2004 Resolute Bay samples only. Age trends were assessed using linear regression.

Doubling times were calculated using the equation, t1/2 = ln(2) / m, where m represents the slope of the natural logarithm transformed liver concentration versus time.

5.3.6. Quality Control and Quality Assurance Data quality control and assurance included instrumental blanks, procedural (method) blanks, matrix spikes and duplicate analysis. Generally, the C8-C12 PFCAs and 10:2 FTCA and FTUCA were detected in most procedural blanks at concentrations ranging from 10 pg/g for PFTA to 1500 pg/g for 10:2 FTCA. Analyte recovery was evaluated by spiking a mixed

150 standard (target concentration ~40 ng/g) into 10 sub-samples (~1 g) of a single ringed seal liver. Full extraction, clean-up and instrumental analysis details are presented in the supporting information. Mean analyte recoveries ranged from 76-145% with the exception of PFBS which was 56%. Recoveries could not be assessed for PFTrA and PFPA since analytical standards were not available.

Two replicate samples were analyzed per year for the Arviat and Resolute Bay 1972 & 2005 samples, whereas one replicate per year was analyzed for the Resolute Bay 1993, 2000 & 2004 samples. Replicates were generally in good agreement with values within 20% of each other for all analytes. Between-day variation was assessed by analyzing Arviat samples during two different runs for PFNA, PFDA, PFUnA, PFDoA and PFOS using method B. Concentrations between the two runs were highly correlated (p<0.01) and values were generally within 20%.

To test PFC variability within the liver, larger portions (~6-8 g) of nine Resolute Bay 2000 samples were individually homogenized and a 0.5 g sample was extracted and analyzed using method B (“homogenate”). Results were compared to those obtained from the 0.5 g sub- samples of the liver (“sub-sample”). The mean ratio between the “homogenate” and “sub- sample” concentrations varied between 0.82 for PFUnA and 1.2 for PFOS, and no systematic bias was observed between the samples. Approximately 91% of the “homogenate” and “sub- sample” concentrations were within a factor of 2. Thus, the fairly typical procedure of analyzing small sub-samples of large whole tissues may result in potential bias of ~2-fold. Complete results are presented in the supplemental information.

Responses of internal standards spiked into calibration standards, method blanks and sample extracts showed little variation for most mass-labeled internal standards. For example, the average difference in internal standard response between the calibration standards and 13 13 samples varied between 18% for C2-8:2 FTUCA and 44% for C4-PFOA. This suggests that there was no significant signal enhancement or suppression.

151 5.4. Results and Discussion

5.4.1. Contaminant Profiles and Concentrations At both locations PFC profiles were dominated by PFOS, although the percent composition generally decreased with time (table of summarized results presented in supporting information). Odd number chain-length PFCAs were generally greater than the corresponding even number chain-length PFCA (ie- PFNA>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 PFOA and PFNA (10). However, longer-chain PFCAs have been shown to be more bioaccumulative (25,26) (ie-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.

PFHpA, PFBS, 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, <0.85- 3.6 ng/g ww at Arviat and <3.6-6.2 ng/g ww at Resolute Bay. This is consistent with other reports of ringed seals (8,22) and likely indicative of the relatively low bioaccumulation and concentration potential of PFOA (25,26).

Fluorotelomer saturated and unsaturated acids were detected in some individuals from both ringed seals populations (summarized results presented in supporting information). The 8:2 FTCA and FTUCA was detected but concentrations were below the method detection limits. The 10:2 FTCA was detected at both sites, but concentrations could not be calculated due to problems with quantification. However, the 10:2 FTUCA concentrations ranged from <0.75-9.6 ng/g ww and <0.75-1.3 ng/g ww in Arviat and Resolute Bay samples, respectively. The detection of fluorinated telomer acids in ringed seals supports the hypothesis of FTOHs as a source of PFCAs to the environment (10,11). FTOHs have been shown to degrade via atmospheric oxidation to FTCAs and FTUCAs (10,11). FTUCAs and FTCAs are known degradation products of FTOHs in microbial degradation (27,28), sewage sludge (29) and rat metabolism (30,31). It is unclear whether the FTCAs and FTUCAs measured within the ringed

152 seals were from the direct exposure of telomer acids or through the uptake and subsequent metabolism of FTOHs. There are few published reports of fluorinated telomer acids in the environment. The 8:2 & 10:2 FTCAs and FTUCAs have been detected in North American rainwater samples from urban and rural locations at concentrations generally less than 1 ng/l (32,33). The 8:2 & 10:2 FTUCAs were measured in bottlenose dolphins at concentrations ~1 ng/g ww (34).

5.4.2. Age and Sex Trends No significant differences (p<0.05) between sex were identified for any analyte within any of the Arviat data sets, and 1993, 2000, 2004 and 2005 Resolute Bay data sets. Sex was not known for the two 1972 Resolute Bay samples. No significant age trends were identified in the 1992 Arviat data set. Significant positive relationships (p<0.05) between age and analyte concentration were identified for PFDoA and PFOS in the 1998 Arviat samples, and PFDA and PFUnA in the 2003 Arviat samples. However, the significant trends observed in the 1998 and 2003 samples were biased by the oldest individual in the data (35 yrs and 12 yrs in the 1998 and 2003 data sets, respectively) and these trends were not significant when those individuals were removed. The occurrence of age-analyte concentration trends was rare, occurring in only 4 out of 45 cases, consistent with a general lack of age-analyte concentration trends in the literature (2). It should be noted that the oldest individuals from the 1998 and 2003 Arviat data sets did not bias the temporal trend analysis with regards to comparison between years. This was determined by performing the ANOVA and post-hoc tests with and without their presence in the data set; identical results were obtained. Figures of the significant age-concentration trends are presented in the supporting information. No significant age-concentration relationships were observed for PFNA, PFDA, PFUnA, PFDoA, PFTrA and PFOS in any of the Resolute Bay data sets.

5.4.3. Temporal Trends Arviat ringed seal samples showed an increase in PFCA concentration from 1992-2005, whereas PFOS and PFOSA concentrations increased from 1992-1998 and subsequently declined from 1998-2005 (Figure 5.1, table in supporting information). Mean concentrations between the sampling years were statistically different (p<0.05) for PFNA, PFDA, PFUnA, PFDoA, PFTrA, PFTA, PFPA, PFOSA and PFOS. PFOA concentrations were detected in 72% of the samples, but below the MDL (0.85 ng/g ww) in 42% of the samples and thus temporal trends were not

153 investigated. Concentrations increased from 1992-1998 for the long-chained PFCAs (PFNA, PFDA, PFUnA, PFDoA, PFTrA, PFTA, PFPA), but later sampling points (1998, 2003, 2005) were not statistically different from each other. Overall, PFCAs have increased by 117% for PFTA to 310% for PFNA between 1992 and 2005. Conversely, mean PFOS concentrations increased from 1992 (22.7 ng/g ww) to 1998 (91.6 ng/g ww), but decreased in 2003 (35.1 ng/g ww) and again in 2005 (19.6 ng/g ww). PFOS concentrations in 1992 and 2005 were not statistically different, indicating that current PFOS concentrations in Arviat ringed seals are equivalent to 1992 levels.

PFOSA levels were very low, ~25- to 230-fold lower than PFOS concentrations, consistent with other reports for Arctic ringed seals (8,22). PFOSA has been shown to be to a metabolic precursor of PFOS in rat liver slices (35). PFOSA temporal trends were similar to those observed with PFOS in Arviat samples and PFOS was statistically correlated (r2=0.67) to PFOSA concentrations in these samples. Relatively high blank PFOSA levels prevented similar comparisons with the Resolute Bay seals.

Similar temporal trends were observed in Resolute Bay ringed seals. The low sample size of 1972 data set prevented its inclusion in the statistical comparison between years. Overall, concentrations of PFCAs increased from 1993-2005, while PFOS showed maximum concentrations in 2000 followed by declining concentrations to 2005 (Figure 5.2). However, the increases in PFCAs observed in the most recent years were not statistically significant. For example, PFNA concentrations increased significantly from 1993 to 2004, but were statistically unchanged from 2004 to 2005. Similarly, concentrations of PFDA, PFUnA and PFDoA increased significantly from 1993 to 2000 but there were not statistically significant differences in the mean concentrations between the 2000, 2004 and 2005 populations. Conversely, PFOS concentrations increased from 1993 to 2000. Mean 2004 PFOS concentrations were lower than 2000, but the measured decrease from 2000 to 2004 is not statistically significant. However, the overall PFOS decrease from 2000 to 2005 is significant. Similar to Arviat data, 2005 concentrations are statistically equal to early 1990s levels.

Even though this study measured the highest PFOS concentrations during 1998 (Arviat) and 2000 (Resolute Bay), it is reasonable to assume that the actual PFOS concentrations peaked shortly after the voluntary withdrawal of PFOSF production in 2001. The exact year of

154 maximum PFOS concentration could not be determined since samples were not available between 1998 & 2003 (Arviat) and 2000 & 2004 (Resolute Bay). Since this study showed a decrease in PFOS concentrations 3-4 years after the production termination, the sampling frequency is an important factor in the ability to detect trends. Similar PFOS trends were not observed in polar bears from the eastern and western North American Arctic collected between 1972-2002 (19), likely because the most recent samples analyzed were from 1993 & 2002 only and 1992 & 2002 only for the “eastern” and “western” groups, respectively. Thus, similar ringed seal and polar bear trends were not observed presumably because the sampling intervals were too far apart to capture the PFOS maximum in the polar bear samples. Ringed seals are important prey for polar bears and thus similar temporal trends would be expected.

Although the most recent data sets (late 1990s and early 2000s) were not statistically different for the majority of PFCAs, regression analysis revealed that the slope of ln concentration (ng/g ww) versus year, when considering all of the time points, was statistically significant (p<0.05) (Table 5.1). Since PFOS and PFOSA concentrations showed peaking concentrations in the late 1990s & early 2000s, doubling times were not calculated over all time periods. Instead, doubling times were calculated during the increasing intervals and disappearance half-lives were calculated during the decreasing intervals. Arviat doubling times (1992-1998) for PFOS and PFOSA were 3.0±1.3 and 3.4±1.7 years, whereas, disappearance half-lives (1998-2005) for PFOS and PFOSA were 3.2±0.9 and 2.0±0.6 years. Resolute Bay doubling times (1972-2000) and disappearance half-lives (2000-2005) for PFOS was 7.1±3.9 and 4.6±9.2 years, respectively.

PFCA doubling times in ringed seals were slightly longer but within the range reported for polar bears from the Canadian arctic (19). Conversely, PFOS doubling times were shorter than reported for polar bears (9.8 ± 5.1 to 13 ± 4.0 years). Reported PFOS doubling times in guillemot eggs from the Baltic Sea were 7-10 years (36). Smithwick et al. noted that the doubling time for PFOSF production was ~11 years within the range measured for polar bears (19).

The mean concentration of some PFCAs in the 2005 Arviat and Resolute Bay samples were lower than measured in 2003 (Arviat) and 2004 (Resolute Bay), however, in all cases this

155 difference was not statistically significant. These results may suggest a recent decrease in PFCA emissions, such as FTOHs which are presumably the main source of PFCAs to the Arctic. Although these results are not consistent with the reported 2-fold increase in FTOH production from 2000-2004 (21), it may in fact be representative of a decline in fugitive emissions during industrial production. Unfortunately, detailed production and emission estimates for FTOHs are not known. In addition, the apparent PFCA decline may also be due to between-year variability, a factor which has been shown to be significant in other temporal trend studies of wildlife (37,38). In contrast, the successive years of PFOS decline suggests that the observed PFOS decrease is not due to between-year variability. These studies also highlight the importance of long-term monitoring for the assessment of temporal trends. We are continuing to collect ringed seal liver samples from these locations in order to follow future trends.

The observation of two consecutive time-points of declining concentrations provides compelling evidence that PFOS levels in Arctic ringed seals are declining. These findings are consistent with a reduction of PFOS and PFOS precursors as a result of the PFOSF chemistry phase-out by 3M in 2001. Further, these findings imply a coincident PFOS decrease in both the abiotic media (eg -air, seawater) as well as the lower trophic levels of the ringed seal foodweb (eg -arctic cod, polar cod, crustaceans), however, there have been no studies to confirm this trend. PFOS elimination rates for ringed seals are not known, but the rapid elimination of PFOS and PFCAs in mammals has been inferred from the general lack of correlation between concentration and age (2,39).

Ultimately, the decrease in ringed seal PFOS concentrations with time is most probably the result of declining concentrations in seawater. There are no known relevant environmental degradation mechanisms of PFCs and thus the decline is presumably not due to chemical or biological degradation. However, it is known that the majority of marine biological processes occur in the Arctic Ocean surface layer (~10 m), particularly during spring ice melt when productivity is elevated (40). In the absence of major riverine inputs, the top of this surface layer (maximum ~2 m) would be highly influenced by ice melt and thus would represent PFC levels resulting from the current year’s atmospheric deposition. The input of freshwater would result in stratification, preventing mixing with the saline lower layer. Therefore, it is possible that the base of food web (phytoplankton and zooplankton) are being exposed to the current year’s PFC atmospheric levels, which would be expected to rapidly respond to changes to the

156 PFC atmospheric flux to the Arctic, as well as historical PFC inputs. Upon ice formation in the winter, the surface layer becomes more saline and drives thermohaline circulation and mixing with the lower layer.

Environmental changes, such as climate warming, reduced sea ice thickness and ice coverage in the past 20 years could influence trends in contaminants via effects on the marine food web thereby influencing the trophic transfer of contaminants to higher-order marine mammals, such as ringed seals (41). A detailed investigation of such environmental factors as related to PFC temporal trends in ringed seals is complex and beyond the scope of this paper. Various biological parameters, such as reproductive status and time of sample collection, have been found as covariates in temporal studies of organochlorinated contaminants in ringed seals and polar bears (42,43). Two such factors, age and sex, were examined in the present study and not found to significantly influence temporal trends. Further, all seal samples were collected during annual spring hunts. However, it should be noted that if such environmental and biological factors were responsible for the observed PFOS decline, then it would be expected that the PFCAs would experience similar trends. But, this is not the case as PFCAs did not show a peak concentration in 1998/2000, and in contrast they show a general increasing trend with time.

The observed rapid PFOS decline in ringed seals is supported by empirical evidence with other persistent organic pollutants, particularly the polybrominated diphenyl ethers (PBDEs). Exponential increases of PBDEs between 1981-2000 in ringed seals from the Canadian Arctic (Holman Island) were found to closely match trends in global penta-BDE production (44), indicating both the rapid atmospheric transport and bioaccumulation of PBDEs within an arctic food web. However, recent results suggest a stabilization of PBDE levels in ringed seals between 2000-2003 (45), consistent with the decreased production of the penta- and octa-BDE formulations in North America (46).

In contrast to PFCAs, levels of “legacy” persistent organic pollutants such as PCBs and organochlorinated pesticides show decreasing temporal trends from the 1980s to 2004 in Resolute Bay and Arviat ringed seals (47). Total DDT compounds (ΣDDT, includes DDT and degradation compounds) declined significantly from 1984-2004 with a disappearance half-life of 11.7 yrs (47). Similarly, the sum of 10 major PCB congeners (ΣPCB10) decreased with a

157 disappearance half-life of 22 yrs. Disappearance half-lives for Arviat ringed seals were 9.3 yrs and 15 yrs for ΣDDT and ΣPCB10, respectively.

5.5. Acknowledgements

Wellington Laboratories (Guelph, Ontario, Canada) is thanked for donation of the mass- labeled internal standards. We thank the Hunters and Trappers associations of Arviat and Resolute Bay for conducting the collection of seal samples. The project would not have been possible without their cooperation. Yan Li is thanked for assistance with sample preparation. Christine Spencer at the NWRI is thanked for LC-MS/MS assistance. Xiaowa Wang is thanked for assistance with sampling storage and sub-sampling at the NWRI. We thank Steve Ferguson (Department of Fisheries and Oceans, Winnipeg) and Magaly Chambellant (University of Manitoba) for help with collection of 2005 ringed seals from Arviat. This paper benefited from discussions with Don Mackay (Trent University) and Robbie Macdonald (Institute of Ocean Sciences, Department of Fisheries and Oceans). The Northern Contaminants Program (Indian and Northern Affairs Canada) provided financial support for the project.

Table 5.1. Doubling times and disappearance half-lives (years ± 95% confidence interval) for Arviat (1992-2005) and Resolute Bay (1972- 2005) ringed seals. Doubling times and disappearance half-lives excluding 2005 samples are shown in parenthesis. Arviat (1992-2005) Resolute Bay (1972-2005)

Analyte r2 P Doubling Time (yr) r2 P Doubling Time (yr)

PFNA 0.305 <0.001 10 ± 7.2 (7.6 ± 4.8) 0.668 <0.0001 7.7 ± 2.0 (7.1 ± 1.8)

PFDA 0.244 <0.01 11 ± 11 (7.0 ± 4.5) 0.625 <0.0001 8.7 ± 2.4 (8.1 ± 2.2)

PFUnA 0.230 <0.001 11 ± 12 (6.2 ± 3.1) 0.734 <0.0001 6.9 ± 1.5 (6.2 ± 1.1)

PFDoA 0.110 0.048 19 ± 1100 (8.7 ± 6.3) 0.270 <0.001 16 ± 12 (13 ± 9.5)

PFTrA 0.236 <0.01 12 ± 13 (7.3 ± 4.5) 0.786 <0.0001 6.9 ± 1.5 (7.0 ± 1.6)

PFTA 0.131 0.033 17 ± 82 (9.2 ± 12) 0.344 <0.001 12 ± 7.1 (9.3 ± 3.8)

PFPA 0.173 0.012 12 ± 21 (6.5 ± 4.5) na na naa

PFOS

1992-1998/ 0.706 <0.001 3.0 ± 1.3 0.556 <0.001 7.1 ± 3.9 1972-2000

1998-2005/ 0.824 <0.001 3.2 ± 0.9 b 0.210 0.014 4.6 ± 9.2 b 2000-2005

PFOSA

1992-1998/ 0.655 <0.001 3.4 ± 1.7 na na naa 1972-2000

1998-2005/ 0.677 <0.001 2.0 ± 0.6 b na na naa 2000-2005 a Doubling times and disappearance half-lives could not be calculated since analytes were not detected (PFPA) or concentrations were below the MDL (PFOSA) b Values in italics represent a disappearance half-life since a significant decline in concentration with time was observed

158

140 8 PFOS PFNA 120 7 6 100 5 80 4 60 3 40 2

20 1

0 0

7 25

PFDA 6 PFUnA 20 5

15 4 ConcentrationConcentration (ng/g (ng/g ww) ww) 3 10 2 5 1

0 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 1990 1992 1994 1996 1998 2000 2002 2004 2006

Year

Figure 5.1. Geometric mean concentrations (ng/g ww) of C9-C11 PFCAs and PFOS from 1992-2005 Arviat ringed seals. Error bars indicate 95% confidence interval. 159

35 10 PFOS 30 PFNA 8 25

20 6

15 4 10

5 2 0

0 6 14 PFUnA PFDA 12 5 10 4 8

ConcentrationConcentration (ng/g (ng/g ww) ww) 3 6

4 2 2 1 0 0 1970 1975 1980 1985 1990 1995 2000 2005 2010 1970 1975 1980 1985 1990 1995 2000 2005 2010

Year

Figure 5.2. Geometric mean concentrations (ng/g ww) of C9-C11 PFCAs and PFOS from 1972-2005 Resolute Bay ringed seals. Error bars indicate 95% confidence interval. Note: Error bars not calculated for 1972 samples due to small sample size. 160

161 5.6. Literature Cited

(1) Giesy, J. P.; Kannan, K. Distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 2001, 35, 1339-1342.

(2) Houde, M.; Martin, J. W.; Letcher, R. J.; Solomon, K.; Muir, D. C. G. Biological monitoring of polyfluoroalkyl substances: A review. Environ. Sci. Technol. 2006, 40, 3463- 3473.

(3) Hansen, K. J.; Clemen, L. A.; Ellefson, M. E.; Johnson, H. O. Compound-specific, quantitative characterization of organic fluorochemicals in biological matrices. Environ. Sci. Technol. 2001, 35, 766-770.

(4) Olsen, G. W.; Huang, H.-Y.; Helzlsouer, K. J.; Hansen, K. J.; Butenhoff, J. L.; Mandel, J. H. Historical comparison of perfluorooctanesulfonate, perfluorooctanoate, and other fluorochemicals in human blood. Environ. Health Perspect. 2005, 113, 539-545.

(5) Lau, C.; Butenhoff, J. L.; Rogers, J. M. The developmental toxicity of perfluoroalkyl acids and their derivatives. Toxicol. Appl. Pharmacol. 2004, 15, 231-241.

(6) Renner, R. Canada bans fluoropolymer stain repellents. Policy News. December 15, 2004. http://pubs.acs.org/subscribe/journals/esthag-w/2004/dec/policy/rr_canada.html

(7) United States Environmental Protection Agency. 2006. 2010/15 PFOA Stewardship Program. http://www.epa.gov/oppt/pfoa/pfoastewardship.htm

(8) Martin, J. W.; Smithwick, M. M.; Braune, B. M.; Hoekstra, P. F.; Muir, D. C. G.; Mabury, S. A. Identification of long-chain perfluorinated acids in biota from the Canadian Arctic. Environ. Sci. Technol. 2004, 38, 373-380.

(9) Smithwick, M.; Mabury, S. A.; Solomon, K. R.; Sonne, C.; Martin, J. W.; Born, E. W.; Dietz, R.; Derocher, A. E.; Letcher, R. J.; Evans, T. J.; Gabrielsen, G. W.; Nagy, J.; Stirling, I.; Taylor, M. K.; Muir, D. C. G. Circumpolar study of perfluoroalkyl contaminants in polar bears (Ursus maritimus). Environ. Sci. Technol. 2005, 39, 5517-5523.

(10) Ellis, D. A.; Martin, J. W.; De Silva, A. O.; Mabury, S. A.; Hurley, M. D.; Andersen, M. P. S.; Wallington, T. J. Degradation of fluorotelomer alcohols: A likely atmospheric source of perfluorinated carboxylic acids. Environ. Sci. Technol. 2004, 38, 3316-3321.

(11) Hurley, M. D.; Wallington, T. J.; Sulbaek Andersen, M. P.; Ellis, D. A.; Martin, J. W.; Mabury, S. A. Atmospheric chemistry of fluorinated alcohols: reaction with Cl atoms and OH radicals and atmospheric lifetimes. J. Phys. Chem. A 2004, 108, 1973-1979.

(12) Ellis, D. A.; Martin, J. W.; Mabury, S. A.; Hurley, M. D.; Sulbaek Andersen, M. P.; Wallington, T. J. Atmospheric lifetime of fluorotelomer alcohols. Environ. Sci. Technol. 2003, 37, 3816-3820.

(13) Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H. Sources, fate and transport of perfluorocarboxylates. Environ. Sci. Technol. 2006, 40, 32-44.

162 (14) D'eon, J. C.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A. Atmospheric chemistry of N-methyl perfluorobutane sulfonamidoethanol, C4F9SO2N(CH3)CH2CH2OH: Kinetics and mechanism of reaction with OH. Environ. Sci. Technol. 2006, 40, 1862-1868.

(15) Martin, J. W.; Ellis, D. A.; Mabury, S. A.; Hurley, M. D.; Wallington, T. J. Atmospheric chemistry of perfluoroalkanesulfonamides: Kinetic and product studies of the OH radical and Cl atom initiated oxidation of N-ethyl perfluorobutanesulfonamide. Environ. Sci. Technol. 2006, 40, 846-872.

(16) Martin, J. W.; Muir, D. C. G.; Moody, C. A.; Ellis, D. A.; Kwan, W. C.; Solomon, K. R.; Mabury, S. A. Collection of airborne fluorinated organics and analysis by gas chromatography/chemical ionization mass spectrometry. Anal. Chem. 2002, 74, 584-590.

(17) Stock, N. L.; Lau, F. K.; Ellis, D. A.; Martin, J. W.; Muir, D. C. G.; Mabury, S. A. Polyfluorinated telomer alcohols and sulfonamides in the North American troposphere. Environ. Sci. Technol. 2004, 38, 991-996.

(18) De Silva, A. O.; Mabury, S. A. Isolating isomer of perfluorocarboxylates in polar bears (Ursus maritimus) from two geographical locations. Environ. Sci. Technol. 2004, 38, 6538- 6545.

(19) Smithwick, M.; Norstrom, R. J.; Mabury, S. A.; Solomon, K.; Evans, T. J.; Stirling, I.; Taylor, M. K.; Muir, D. C. G. Temporal trends of perfluoroalkyl contaminants in polar bears (Ursus maritimus) from two locations in the North American arctic, 1972-2002. Environ. Sci. Technol. 2006, 40, 1139-1143.

(20) 3M. Phase-Out Plan for POSF-Based Products; USEPA Docket ID OPPT-2002-0043; 3M, Specialty Materials Markets Group: St. Paul, MN, 2000.

(21) DuPont Global PFOA Strategy - Comprehensive Source Reduction; presented to the USEPA OPPT, January 31, 2005; U.S. Environmental Protection Agency public docket AR226- 1914

(22) Bossi, R.; Riget, F. F.; Dietz, R. Temporal and spatial trend of perfluorinated compounds in ringed seal (Phoca hispida) from Greenland. Environ. Sci. Technol. 2005, 39, 7416-7422.

(23) Stewart, R. E. A.; Lavigne, D. M. Proceedings of an International Workshop on the Biology and Management of Northwest Atlantic Harp Seals 1979.

(24) Furdui, V. I.; Stock, N. L.; Ellis, D.; Butt, C. M.; Whittle, D. M.; Crozier, P. W.; Reiner, E. J.; Muir, D. C. G.; Mabury, S. A. Spatial distribution of the perfluoroalkyl contaminants in lake trout from the Great Lakes. Environ. Sci. Technol. 2007, 41, 1554-1559.

(25) Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G. Dietary accumulation of perfluorinated acids in juvenile rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 189-195.

163 (26) Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G. Bioconcentration and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 196-204.

(27) Dinglasan-Panlilio, M. J. A.; Mabury, S. A. Significant residual fluorinated alcohols present in various fluorinated materials. Environ. Sci. Technol. 2006, 40, 1447-1453.

(28) Wang, N.; Szostek, B.; Buck, R. C.; Folsom, P. W.; Sulecki, L. M.; Capka, V.; Berti, W. R.; Gannon, J. T. Fluorotelomer alcohol biodegradation - direct evidence that perfluorinated carbon chains breakdown. Environ. Sci. Technol. 2005, 39, 7516-7528.

(29) Wang, N.; Szostek, B.; Folsom, P. W.; Sulecki, L. M.; Capka, V.; Buck, R. C.; Berti, W. R.; Gannon, J. T. Aerobic biotransformation of 14C-labeled 8-2 telomer B alcohol by activated sludge from a domestic sewage treatment plant. Environ. Sci. Technol. 2005, 39, 531-538.

(30) Hagen, D. F.; Belisle, J.; Johnson, J. D.; Venkateswarlu, P. Characterization of fluorinated metabolites by a gas chromatographic-helium microwave plasma detector - the biotransformation of 1H,1H,2H,2H-perfluorodecanol to perfluorooctanoate. Anal. Biochem. 1981, 118, 336-343.

(31) Martin, J. W.; Mabury, S. A.; O'Brien, P. J. Metabolic products and pathways of fluorotelomer alcohols in isolated rat hepatocytes. Chem. Biol. Interact. 2005, 155, 165-180.

(32) Loewin, M.; Halldorson, T.; Wang, F.; Tomy, G. T. Fluorotelomer carboxylic acids and PFOS in rainwater from an urban center in Canada. Environ. Sci. Technol. 2005, 39, 2944-2951.

(33) Scott, B. F.; Spencer, C.; Mabury, S. A.; Muir, D. C. G. Poly and perfluorinated carboxylates in North American precipitation. Environ. Sci. Technol. 2006, 40, 7167-7174.

(34) Houde, M.; Wells, R. S.; Fair, P. A.; Bossart, G. D.; Hohn, A. A.; Rowles, T. K.; Sweeney, J.; Solomon, K. R.; Muir, D. C. G. Polyfluorinated compounds in free-ranging bottlenose dolphins (Tursiops truncatus) from the Gulf of Mexico and the Atlantic Ocean. Environ. Sci. Technol. 2005, 39, 6591-6598.

(35) Xu, L.; Krenitsky, D. M.; Seacat, A. M.; Butenhoff, J. L.; Anders, M. W. Biotransformation of N-ethyl-N(2-hydroxyethyl)perfluorooctanesulfonamide by rat liver microsomes, cytosol, and slices and by expressed rat and human cytochromes P450. Chem. Res. Toxicol. 2004, 17, 767-775.

(36) Holmström, K. E.; Järnberg, U.; Bignert, A. Temporal trends of PFOS and PFOA in Guillemot eggs from the Baltic Sea, 1968-2003. Environ. Sci. Technol. 2005, 39, 80-84.

(37) Hebert, C. E.; Weseloh, D. V. C. Assessing temporal trends in contaminants from long- term avian monitoring programs: The influence of sampling frequency. Ecotoxicology 2003, 12, 141-151.

(38) Bignert, A.; Riget, F.; Braune, B. M.; Outridge, P. M.; Wilson, S. Recent temporal trend monitoring of mercury in Arctic biota - how powerful are the existing data sets? J. Environ. Monit. 2004, 6, 351-355.

164 (39) Houde, M.; Balmer, B. C.; Brandsma, S.; Wells, R. S.; Rowles, T. K.; Solomon, K. R.; Muir, D. C. G. Perfluoroalkyl compounds in relation with life-history and reproductive parameters in bottlenose dolphins (Tursiops truncatus) from Sarasota Bay, Florida, USA. Environ. Toxicol. Chem. 2006, 25, 2405-2412.

(40) Macdonald, R. W.; Barrie, L. A.; Bidleman, T. F.; Diamond, M. L.; Gregor, D. J.; Semkin, R. G.; Strachan, W. M. J.; Li, Y. F.; Wania, F.; Alaee, M.; Alexeeva, L. B.; Backus, S. M.; Bailey, R.; Bewers, J. M.; Gobeil, C.; Halsall, C. J.; Harner, T.; Hoff, J. T.; Jantunen, L. M. M.; Lockhart, W. L.; Mackay, D.; Muir, D. C. G.; Pudykiewicz, J.; Reimer, K. J.; Smith, J. N.; Stern, G. A.; Schroeder, W. H.; Wagemann, R.; Yunker, M. B. Contaminants in the Canadian Arctic: 5 years of progress in understanding sources, occurrence and pathways. Sci. Total Environ. 2000, 254, 93-234.

(41) Macdonald, R. W.; Harner, T.; Fyfe, J. Recent climate change in the Arctic and its impact on contaminant pathways and interpretation of temporal trend data. Sci. Total Environ. 2005, 342, 5-86.

(42) Henriksen, E. O.; Wiig, Ø.; Skaare, J. U.; Gabrielsen, G. W.; Derocher, A. E. Monitoring PCBs in polar bears: Lessons learned from Svalbard. J. Environ. Monit. 2001, 3, 493-498.

(43) Muir, D.; Riget, F.; Cleemann, M.; Skaare, J.; Kleivane, L.; Nakata, H.; Dietz, R.; Severinsen, T.; Tanabe, S. Circumpolar trends of PCBs and organochlorine pesticides in the arctic marine environment inferred from levels in ringed seals. Environ. Sci. Technol. 2000, 34, 2431-2438.

(44) Ikonomou, M. G.; Rayner, S.; Addison, R. F. Exponential increases of the brominated flame retardants, polybrominated diphenyl ethers, in the Canadian Arctic from 1981 to 2000. Environ. Sci. Technol. 2002, 36, 1886-1892.

(45) Ikonomou, M. G.; Kelly, B. C.; Stern, G. A. Spatial and temporal trends of PBDEs in biota from the Canadian Arctic marine environment. Organohalogen Compounds 2005, 67, 950- 953.

(46) Renner, R. In U.S., flame retardants will be voluntarily phased out. Environ. Sci. Technol. 2004, 38, 14A.

(47) Muir, D.; Kwan, M.; Evans, M.; Butt, C.; Mabury, S.; Moore, S.; Sverko, E.; Williamson, M.; Wang, X. Temporal trends of persistant organic pollutants and metals in ringed sleas from the Canadian Arctic. 2005. Synopsis of research conducted under the 2004-2005 Northern Contaminants Program. Ottawa, ON, Indian and Northern Affairs Canada.

165 5.7. Supporting Information

Resolute Bay: 1972, 1993, 2000, 2004, 2005

Arviat: 1992, 1998, 2003, 2005

Figure S5.1. Ringed seal sample locations (sample years).

166 Table S5.1. Multiple reaction monitoring (MRM) transitions of target analytes. Analyte Parent Ion Daughter Ion PFHpA 363 319 PFOA 413 369 PFNA 463 419 PFDA 513 469 PFUnA 563 519 PFDoA 613 569 PFTrA 663 619 PFTA 713 669 PFPA 763 719 PFBS 299 99 PFHxS 399 99 PFOS 499 99 PFDS 599 99 PFOSA 497.7 78 8:2 FTCA 477 393 8:2 FTUCA 457 393 10:2 FTCA 577 493 10:2 FTUCA 557 493 13 C2-PFOA 414.8 369.9 13 C4-PFOA 417 372 13 C5-PFNA 467.8 422.9 13 C2-PFDA 514.8 469.9 13 C4-PFOS 503 99 13 C2-8:2 FTUCA 459 394 13 559 494 C2-10:2 FTUCA

13 13 13 Perfluoro-n-[1,2,3,4- C4]octanoic acid ( C4-PFOA), perfluoro-n-[1,2,3,4,5- C5]nonanoic acid 13 13 13 ( C5-PFNA), perfluoro-n-[1,2- C2]decanoic acid ( C2-PFDA), sodium perfluoro-1-[1,2,3,4- 13 13 13 13 C4]octanesulfonate ( C4-PFOS), 2H-perfluoro-[1,2- C2]-2-decenoic acid (8:2 C2-FTUCA), 13 13 2H-perfluoro-[1,2- C2]-2-dodecenoic acid (10:2 C2-FTUCA) were provided by Wellington 13 13 Laboratories (Guelph, Ontario). Perfluoro-n-[1,2- C2]octanoic acid ( C2-PFOA) was purchased from PerkinElmer Life and Analytical Sciences Canada.

13 Analyte responses were normalized to internal standard responses. In method A, C2-PFOA 13 was used as the internal standard for all analytes. In method B, C4-PFOS was used for PFOS, 13 13 13 C4-PFOA for PFHpA and PFOA, C5-PFNA for PFNA, C2-PFDA for PFDA, PFUnA, 13 PFDoA, PFTrA, PFTA and PFPA, 10:2 C2-FTUCA for 10:2 FTCA and 10:2 FTUCA, and 8:2 13 C2-FTUCA for 8:2 FTCA and 8:2 FTUCA. In both methods, PFTrA was quantified using the average response factor for PFDoA and PFTA, and PFPA was quantified using the response factor for PFTA since analytical standards were not available.

In method B, potential PFCA contamination from polytetrafluorethylene (PTFE) components in the liquid chromatograph were reduced by replacing any PTFE components. These included bypassing the degasser, replacing PTFE solvent lines with Peek tubing, inserting Peek seals, and replacing glass inlet solvent filters with stainless steel filters. Consequently, injections of blank solvent did not contain any of the monitored analytes. This suggests that the liquid chromatograph did not contribute PFC contamination.

167 Figure S5.2. Comparison of concentrations (ng/g ww) using Method A & Method B.

PFNA 9

Method B = 0.67 * Method A R2 = 0.9131 6 Method B Method 3

0 02468101214

Method A

PFDA 6

Method B = 0.72* Method A 5 R2 = 0.9313

4

3 Method B 2

1

0 01234567 Method A

PFUnA 16 Method B = 1.03 * Method A R2 = 0.9377 12

8 Method B Method

4

0 02468101214 Method A

168

PFDoA 3 Method B = 0.69 * Method A R2 = 0.8496 2

Method B Method 1

0 01234 Method A

PFOS 25

Method B = 0.79 *Method A 20 R2 = 0.9787

15

Method B 10

5

0 0 5 10 15 20 25 Method A

169

Table S5.2. Spike and recovery data for individual analytes spiked into ~1g of ringed seal liver (n=10). Note: recoveries for PFTrA and PFPA could not be evaluated since commercial standards are not available. Analyte Mean Recovery % RSD PFHpA 145.1 18 PFOA 93.4 23 PFNA 98.4 25 PFDA 94.0 20 PFUnA 96.7 22 PFDoA 100.6 20 PFTrA n/a n/a PFTA 110.6 20 PFPA n/a n/a PFBS 56.2 21 PFHxS 76.0 22 PFOS 85.9 31 PFDS 86.2 23 PFOSA 126.5 18 8:2 FTCA 81.7 19 8:2 FTUCA 85.8 19 10:2 FTCA 76.5 74 10:2 FTUCA 82.4 21

To evaluate analyte recovery, 13 sub-samples (~1 g) of a single ringed seal liver were individually homogenized with sodium carbonate and TBAS. Ten homogenates were spiked with ~40 ng of a mixed stock solution, vortexed for 30s and kept cool for 7 days. Three unspiked sub-samples were also analyzed to correct for background concentrations in the liver. Extraction, clean-up and instrumental analysis was performed following method B.

170 Table S5.3 & Figure S5.3. Comparison of Resolute Bay 2000 ringed seal liver samples: Homogenates versus Sub-samples. “Homogenates” refers to ~0.5 g sample of homogenized liver (original mass ~6-8g). “Sub-sample” refers to ~0.5 g sample of unhomogenized liver.

PFOS Sub- Homogenate: PFOS Homogenate sample Sub-sample ratio 60

RB00-06 18.1 4.4 4.1 50 RB00-08 12.4 10.7 1.2 RB00-02 14.4 15.8 0.9 40 RB00-11 15.0 25.7 0.6 30 RB00-14 24.0 49.7 0.5 Sub-sample RB00-15 13.0 22.1 0.6 20 RB00-10 12.0 11.0 1.1 RB00-18 19.9 27.7 0.7 10 RB00-21 40.8 31.8 1.3 0 Mean 0 1020304050 Ratio 1.2 Homogenate PFNA Sub- Homogenate: Homogenate sample Sub-sample ratio PFNA RB00-06 3.3 2.1 1.6 10 RB00-08 3.9 3.3 1.2 RB00-02 1.7 2.1 0.8 8 RB00-11 3.0 4.3 0.7 6 RB00-14 2.3 4.1 0.6 RB00-15 1.4 2.3 0.6 4 RB00-10 1.5 1.6 1.0 Sub-sample RB00-18 2.7 4.8 0.6 2 RB00-21 12.1 8.7 1.4 Mean 0 Ratio 0.9 02468101214 Homogenate

PFDA Sub- Homogenate: Homogenate sample Sub-sample ratio

RB00-06 2.9 1.3 2.3 PFDA 6 RB00-08 3.3 2.9 1.1 RB00-02 2.5 2.0 1.3 5 RB00-11 2.4 2.7 0.9 4 RB00-14 3.1 4.7 0.7

RB00-15 1.6 2.1 0.8 3 RB00-10 1.4 1.1 1.2 Sub-sample RB00-18 2.8 3.2 0.8 2

RB00-21 6.2 5.6 1.1 1 Mean Ratio 1.1 0 01234567 Homogenate

171 PFUnA Homogenate: Homogenate Sub-sample Sub-sample ratio PFUnA 16 RB00-06 4.8 2.8 1.7 RB00-08 5.1 5.2 1.0 14

RB00-02 3.6 4.5 0.8 12

RB00-11 3.7 7.4 0.5 10 RB00-14 6.2 13.4 0.5 8 RB00-15 2.9 5.2 0.5 RB00-10 2.6 3.4 0.8 Sub-sample 6 RB00-18 3.8 6.7 0.6 4

RB00-21 9.2 9.4 1.0 2 Mean Ratio 0.8 0 0246810 PFDoA Homogenate Homogenate: Homogenate Sub-sample PFDoA Sub-sample ratio 3.0

RB00-06 1.1 0.5 2.4 2.5 RB00-08 1.1 0.8 1.2 RB00-02 1.1 1.2 0.9 2.0 RB00-11 0.9 1.4 0.7 1.5 RB00-14 1.8 2.4 0.7 Sub-sample RB00-15 0.8 1.2 0.6 1.0 RB00-10 0.9 0.9 1.0 RB00-18 0.9 1.0 0.9 0.5 RB00-21 2.1 1.4 1.5 0.0 Mean Ratio 1.1 0.00.51.01.52.02.5 Homogenate PFTrA PFTrA Homogenate: Homogenate Sub-sample 6.0 Sub-sample ratio 5.0 RB00-06 1.6 0.8 2.1

RB00-08 1.0 1.2 0.9 4.0 RB00-02 1.2 1.8 0.7 RB00-11 1.2 1.7 0.7 3.0 Sub-sample RB00-14 2.1 5.4 0.4 2.0 RB00-15 0.9 2.2 0.4 RB00-10 1.2 1.5 0.8 1.0 RB00-18 1.0 2.1 0.5 0.0 RB00-21 2.4 2.0 1.2 0.00.51.01.52.02.53.0 Mean Ratio 0.9 Homogenate

172 Figure S5.4. Statistically Significant Age-Concentration Trends.

Note: Statistically significant (p<0.05) age-concentration correlations were observed for PFDoA and PFOS in the Arviat 1998 data sets and PFDA and PFUnA in the Arviat 2003 data sets. These trends became insignificant when the oldest individual was removed from the data set.

Arviat 1998 - PFOS 5.5

5.0

4.5

4.0 y = 0.04x + 3.68 ln Conc (ng/g ww) (ng/g ln Conc 3.5 R2 = 0.52, p=0.01

3.0 0 10203040 age (yr)

Arviat 1998 - PFDoA 1.4

1.2

1.0

0.8

0.6 y = 0.02x + 0.64 2 0.4 R = 0.57, p=0.01 ln Conc. (ng/g ww) (ng/g ln Conc. 0.2

0.0 0 10203040 age (yr)

173

Arviat 2003 - PFDA 2.5

2.0

1.5

1.0 y = 0.08x + 0.94 ln Conc. (ng/g ww) (ng/g Conc. ln 0.5 R2 = 0.46, p=0.04

0.0 02468101214 age (yr)

Arviat 2003 - PFUnA 4.0

y = 0.07x + 2.28 3.5 R2 = 0.51, p=0.03

3.0

ln Conc. (ng/g ww) (ng/g ln Conc. 2.5

2.0 02468101214 age (yr)

Table S5.4. Mean (minimum-maximum) concentration (ng/g ww) of perfluorinated acids in ringed seals from Arviat and Resolute Bay. “nd” signifies analyte not detected, “na” signifies analyte not analyzed, “nq” signifies analyte detected but not quantified.

Arviat Resolute Bay 1992 (n=6) 1998 (n=10) 2003 (n=10) 2005 (n=10) 1972 (n=2) 1993 (n=9) 2000 (n=9) 2004 (n=9) 2005 (n=9) <1.6 <1.6 <1.6 PFHpA nd nd nd nd nd nd (nd-<1.6) (nd-<1.6) (nd-<1.6) <0.85 1.67 0.98 4.5 6.2 ( PFOA 1.13 1.1 (0.97-1.2) 3.9 (<3.6-4.1) <0.85 (nd-<0.85) (0.98-3.61) (0.96-1.01) (<3.6-4.5) <3.6-6.2) 0.33 4.8 PFNA 1.7 (1.2-2.5) 5.0 (2.0-8.1) 5.1 (3.0-12.0) 5.1 (1.4-9.1) 1.9 (0.8-3.2) 3.7 (1.6-8.7) 6.8 (2.0-9.3) (0.15-0.51) (2.4-10.9) 0.43 PFDA 1.6 (0.9-2.6) 4.9 (2.2-9.4) 4.7 (3.0-10.2) 3.9 (1.8-6.4) 1.2 (0.7-1.9) 2.8 (1.1-5.6) 4.3 (2.0-5.4) 3.4 (1.1-5.8) (0.14-0.72) 5.5 15.6 17.6 12.0 0.34 11.0 7.5 PFUnA 2.8 (0.8-4.0) 6.4 (2.8-13.4) (2.4-12.1) (8.2-31.5) (12.4-36.0) (5.9-22.4) (0.20-0.48) (7.0-13.6) (2.8-14.8) 0.55 0.47 0.96 PFDoA 1.4 (0.5-3.6) 2.8 (1.8-3.7) 2.9 (2.1-5.8) 1.9 (1.2-3.1) 1.2 (0.47-2.4) 1.4 (1.1-1.8) (0.17-0.93) (0.10-0.85) (0.40-2.0) 0.11 0.54 PFTrA 2.0 (0.6-6.4) 4.0 (2.5-8.1) 4.5 (3.4-6.5) 3.3 (2.1-4.9) 2.1 (0.76-5.5) 2.0 (1.2-2.6) n/a (nd-0.11) (0.23-0.78) 0.3 0.49 0.51 0.37 0.04 0.16 0.39 0.27 0.21 PFTA (0.06-1.05) (0.28-1.02) (0.35-0.79) (0.26-0.49) (0.01-0.06) (0.08-0.29) (0.14-0.76) (0.09-0.51) (0.06-0.46) 0.15 0.27 0.29 0.19 PFPA nd nd nd nd n/a (0.06-0.40) (0.17-0.49) (0.18-0.46) (0.10-0.34)

PFBS nd nd nd nd nd nd nd nd nd

PFHxS nd nd nd nd nd nd nd nd nd 22.7 91.6 35.1 19.6 7.0 22.1 16.8 8.1 PFOS 1.8 (0.58-3.0) (11.7-41.6) (36.3-177.0) (20.8-74.1) (8.0-44.1) (1.6-14.7) (4.4-49.7) (5.9-27.7) (2.0-17.0) PFDS nd nd nd nd nd nd nd nd nd

0.40 1.48 0.48 0.15 0.04 0.15 PFOSA <9.9 <9.9 <9.9 (0.20-0.61) (0.71-3.62) (0.19-0.93) (0.05-0.64) (0.03-0.06) (0.05-0.26) <4.6 <4.6 <4.6 8:2 FTCA nd nd nd nd nd nd (nd-<4.6) (nd-<4.6) (nd-<4.6) <0.04 <0.04 <0.04 <0.04 0.10 <0.2 <0.2 8:2 FTUCA <0.2 <0.04 (nd-<0.04) (nd-<0.04) (nd-<0.04) (nd-<0.04) (0.06-0.13) (nd-<0.2) (nd-<0.2)

174

CHAPTER SIX

Prevalence of long-chained perfluorinated carboxylates in seabirds from the Canadian Arctic between 1975 and 2004

Craig M. Butt, Scott A. Mabury, Derek C.G. Muir, Birgit M. Braune

Published In: Environmental Science & Technology 2007, 41, 3521-3528

Contributions: Craig Butt performed liver extraction, instrumental analysis and data interpretation. Birgit Braune coordinated liver sample collection and sub-sampling. The manuscript was prepared by Craig Butt with the critical comments provided by Scott Mabury, Derek Muir and Birgit Braune

Reproduced with permission from Environmental Science and Technology Copyright American Chemical Society 2007

175 176 Chapter Six – Prevalence of long-chained perfluorinated carboxylates in seabirds from the Canadian Arctic between 1975 and 2004

6.1. Abstract

Temporal trends in perfluoroalkyl compounds (PFCs) were investigated in liver samples from two seabird species, thick-billed murres (Uria lomvia) and northern fulmars (Fulmaris glacialis) from Prince Leopold Island in the Canadian Arctic. Thick-billed murres samples were from 1975, 1993 and 2004, whereas, northern fulmars were from 1975, 1987, 1993 and 2003.

Between 8-10 individuals were analyzed per year. Analytes included C7-C15 perfluorinated carboxylates (PFCAs) and their suspected precursors, the 8:2 & 10:2 fluorotelomer saturated and unsaturated carboxylates (FTCAs, FTUCAs), C6, C8 (perfluorooctane sulfonate, PFOS), 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 LC-MS/MS. Overall, concentrations in seabirds were lower than other marine animals that occupy similar or higher trophic positions. In contrast to most other wildlife samples, PFC profiles were dominated by the PFCAs which comprised 81% and 93% of total PFC profiles in the 2004 thick-billed murre and 2003 northern fulmar samples, respectively. As well, the PFCA profiles were mainly comprised of the C11-C15 PFCAs, which appears to be unique among other wildlife species. PFC concentrations were found to increase significantly from 1975 to 2003/2004. Doubling times in thick-billed murres ranged from 2.3 yrs for perfluoropentadecanoate (PFPA) to 9.9 yrs for perfluorododecanoate (PFDoA), and from 2.5 yrs for PFPA to 12 yrs for perfluorodecanoate (PFDA) in northern fulmars. PFCA concentration increases in thick-billed murres were significant both time periods (1975→1993, 1993→2004), but in northern fulmars appeared to remain steady after 1993. Differences in the temporal trends observed may be the result of differing migratory patterns of the seabirds. Finally, the detection of the 8:2 and 10:2 FTUCAs in seabirds is suggestive of fluorotelomer alcohols as a source of some PFCAs.

177 6.2 Introduction

The occurrence of perfluorinated alkyl compounds (PFCs) in wildlife and humans, including remote regions such as the Arctic, has been extensively reported (1-3). The most commonly monitored compounds comprise two PFC classes, the perfluorinated carboxylates (PFCAs) and the perfluorinated sulfonates (PFSAs). Although most studies have only analyzed for the eight-carbon perfluorinated acids, perfluorooctanoate (PFOA) and PFOS, there has been increasing attention concerning the longer-chain PFCAs. In fact, in most wildlife tissues the overall PFCA profile is dominated by the long-chain PFCAs (4) due to the greater bioaccumulation factors of these compounds (5, 6). There is some evidence of increased toxicological effects associated with increasing fluorinated chain-length, such as the inhibition of gap-junction intercellular communication (7).

In addition, there have been several reports of known PFCA and PFSA precursors in wildlife tissues, such as the perfluoroalkyl sulfonamido alcohols, N-EtFOSA, and the FTCAs & FTUCAs (8-10). Fluorotelomer alcohols (FTOHs) have been shown to degrade to PFCAs, through FTCAs and FTUCAs, via atmospheric oxidation (11, 12) and biotic processes such as microbial degradation (13, 14) and rat metabolism (15). Perfluoroalkyl sulfonamido alcohols have been shown to degrade via atmospheric oxidation to PFCAs and PFSAs, in a mechanism analogous to FTOH oxidation (16, 17)

Two mechanisms have been proposed to elucidate the transport of PFCs to the arctic environment; indirectly through atmospheric transport and directly via the ocean currents. The atmospheric mechanism involves the atmospheric transport of volatile precursor chemicals, such as FTOHs and polyfluorinated sulfonamido alcohols that degrade to PFCAs and PFSAs. FTOHs and polyfluorinated sulfonamido alcohols have been detected in the atmosphere (18-20). Further, it has been shown that the atmospheric lifetime of FTOHs with respect to hydroxyl radical reaction is ~20 days, and is sufficiently long to permit transport to the Arctic (21). The other mechanism involves the direct transport of PFCAs themselves to the Arctic through oceanic transport (22, 23). PFCs are presumably transported to the Arctic through a combination of both mechanisms. However, the short PFOS disappearance half-life in ringed seals following 3M’s PFOSF phase-out (10) as well as the detection of almost exclusively linear

178 isomers of several PFCAs (24) is consistent with the atmospheric transport of PFC precursors as a significant transport mechanism to the Arctic wildlife.

Seabirds have previously been used for the biomonitoring of temporal trends in “legacy” (ie- PCBs, organochlorinated pesticides, mercury) and “new” (ie- PBDEs) contaminants in the Arctic (25-27). The contaminant body burden in seabirds represents an integration of exposure from both the breeding grounds and overwintering locations. Thus, interpretation of temporal trends may be confounded by the migratory behaviour of seabirds that overwinter in differing locations in addition to potentially differing diets and metabolic capacities. This may result in differing temporal trends of contaminants between seabird species (27).

There have been relatively few reports of PFCs in seabirds from arctic regions. Previous studies include northern fulmars (4), black-legged kittiwakes (8) and glaucous gulls (8) from the Canadian Arctic, and glaucous gulls from the Norwegian Arctic (28). However, only two of the studies analyzed for the longer-chain PFCAs (4, 28). Not surprisingly, PFC concentrations in seabirds from these remote regions are lower than measured in species from industrial regions (3). PFC concentrations in fish-eating birds are generally lower than in fish-eating mammals, which has been suggested to be due a shorter elimination half-life in birds (3).

Temporal studies of PFCs in wildlife are limited, particular those from arctic regions. In fact, we are not aware of any previous studies of PFC temporal trends in arctic seabirds. Overall, there appears to be a general increasing trend of PFCAs and PFOS concentrations in wildlife over time (3). However, a recent decline in PFOS levels in arctic ringed seals, from the early 2000s through 2005, has been observed (10). Regarding seabirds from non-arctic regions, white-tailed sea eagles from eastern Germany and Poland showed significant PFOS increases from 1979-1999 (29). Also, guillemot eggs from the Baltic Sea showed a 30-fold increase in PFOS concentrations between 1968-2000 (30). However, PFOS concentrations in these samples appeared to decline post-1997, although this trend was observed only in the pooled egg samples.

This paper presents temporal trends of PFCs in two seabird species, thick-billed murres and northern fulmars, from Prince Leopold Island in the Canadian Arctic. The influence of overwintering location on temporal trends is described. Seabird liver samples were analyzed for

179

C7-C15 PFCAs, the suspected PFCA precursors 8:2 and 10:2 saturated and unsaturated fluorotelomer acids, and C6, C8 and C10 PFSAs.

6.3. Materials and Methods

6.3.1. Standards and Chemicals Standards and reagents used were identical to those previously reported by our laboratory (10).

6.3.2. Sample Collection Adult northern fulmars (Fulmarus glacialis) and thick-billed murres (Uria lomvia) were collected from the Prince Leopold Island (PLI) Migratory Bird Sanctuary (74o02’ N, 90o05’ W) in Lancaster Sound, Nunavut, Canada (figure S6.1 in supporting information). Northern fulmars were collected in 1975 (n=9), 1987 (n=8), 1993 (n=10) and 2003 (n=9), whereas thick-billed murres were collected in 1975 (n=8), 1993 (n=10) and 2004 (n=10). All birds were collected using quick-kill techniques. With the exception of the 2003 fulmars, birds were shipped to the Canadian Wildlife Service (CWS) at the National Wildlife Research Centre (NWRC) in Gatineau, Quebec, where the birds were dissected and livers removed under chemically clean conditions. Livers were homogenized, transferred to acetone-hexane rinsed glass vials and stored at -40oC. These storage conditions have been shown to avoid changes over time in the concentration of various organochlorine contaminants as well as moisture content (31, 32). Livers from the 2003 fulmars were excised in the field, transferred to a Whirlpak bag, placed in a second Whirlpak bag and kept on ice until shipment back to the NWRC. Specific ages could not be determined, but all individuals were classified as adult birds.

6.3.3. Sample Extraction and Instrumental Analysis Liver samples were extracted using MTBE and hexafluoropropanol (HFP) as described in detail elsewhere (10, 33). Briefly, the MTBE extract was blown-down to ~0.5ml under a gentle stream of nitrogen gas, an equal volume of HFP was added and the precipitated components were removed by filtering. The MTBE/HFP mixture was evaporated to dryness and reconstituted with 500 µl of methanol. The suite of internal standards was added immediately prior to instrumental analysis.

180

Instrumental analysis was performed by liquid chromatography with negative electrospray tandem mass spectrometry (LC-MS/MS) under conditions previously described (33). Analytes were detected using an API 4000 Q Trap (Applied Biosystems/MDS Sciex, Concord, ON, Canada) with samples injected with an Agilent 1100 autosampler (injection volume = 100 μl, flow rate = 250 μl/min). Chromatography was performed using a Genesis C18 column (50 mm x 2.1 mm, 4 μm particle size, Chromatographic Specialties, Brockville, ON, Canada), preceded by a C18 guard column (4.0 x 2.0 mm, Phenomenex, Torrance, CA, USA). The mobile phase was methanol (MeOH):water (0.01 M ammonium acetate) and analytes were separated using gradient conditions. Initial conditions were 75:25 MeOH:water, increasing to 90:10 MeOH:water over 3 minutes, followed by a 3 minute hold and then reverting to initial conditions.

Target analytes included perfluoroheptanoate (PFHpA), PFOA, perfluorononanoate (PFNA), PFDA, perfluoroundecanoate (PFUnA), PFDoA, perfluorotridecanoate (PFTrA), perfluorotetradecanoate (PFTA), PFPA, 8:2 & 10:2 FTCA, 8:2 & 10:2 FTUCA, perfluorohexane sulfonate (PFHxS), PFOS, perfluorodecane sulfonate (PFDS) and PFOSA.

Analyte responses were normalized to internal standard responses. The internal standard mix was added to give a final concentration of approximately 250 ng/L for all internal standards. 13 13 C2-PFDA was used for all PFCAs, PFSAs and PFOSA; 10:2 C2-FTUCA for 10:2 FTCA and 13 10:2 FTUCA, and 8:2 C2-FTUCA for 8:2 FTCA and 8:2 FTUCA. PFTrA was quantified using the response factor for PFDoA, and PFPA was quantified using the response factor for PFTA since analytical standards were not available. Concentrations were not corrected for recovery.

6.3.4. Statistical Analysis and Data Treatment Instrumental detection limits (IDL) were determined as the standard deviation from nine injections of the lowest calibration standard. Method detection limits (MDL) were determined as three times the standard deviation of the procedural blanks. The IDL was used if analytes were not detected in the blanks. Concentrations less than the IDL were reported as non-detect (nd), whereas concentrations less than the MDL were reported as

181 were blank corrected. For calculation of means, concentrations less than the MDL or non-detect were replaced by a random number less than half of the MDL.

Statistical comparison of means and temporal trend analysis was not performed for analytes in which greater than 50% of the samples were not detected or had concentrations

6.3.5. Quality Control and Quality Assurance Results from a detailed quality control and quality assurance program using ringed seal livers has previously been reported (10). Data QC/QA included the analysis of instrumental blanks, procedural (method) blanks, matrix spikes, duplicate analysis and assessment of between-day variation. Mean analyte recoveries using ringed seal liver ranged from 76 to 145% (10).

6.4. Results and Discussion

6.4.1. Overall Concentrations and Contaminant Profiles Perfluorinated acids were detected in all thick-billed murre and northern fulmars samples from all time points (full data summarized in table S6.1 of supporting information). However, in the case of the 1975 murre and fulmar samples, concentrations of most analytes were below the MDL. PFHxS was either not detected or levels were

182 exception of the 2003 fulmar samples. Similarly, quantifiable levels of PFHpA were observed in the 1993 and 2003 fulmars only. The 8:2 and 10:2 FTCAs were not detected in any sample. Considering the most recent seabird samples (i.e. thick-billed murres from 2004 and northern fulmars from 2003), the geometric mean total (Σ) PFCA concentrations were 23.9 ng/g (range: 8.9-34.2 ng/g) and 12.4 ng/g (4.2-34.5 ng/g) for thick-billed murres and northern fulmars, respectively. PFOS concentrations were much lower with geometric mean concentrations of 0.76 ng/g (0.13-6.1 ng/g) and 0.41 ng/g (0.07-0.85 ng/g) in the murres and fulmars, respectively.

Overall, the 2003/2004 PFC concentrations in the seabirds were lower than other wildlife that occupy similar or higher tropic levels, such as ringed seals and polar bears (3). Biomagnification of PFCs through food chains has been shown in other arctic (4, 8) and temperate food webs (34, 35). For example, PFOS concentrations in the seabirds were ~10-fold lower than ringed seals (4) and 1000-fold lower than polar bears (36) from the Canadian Arctic. However, total PFCA (ΣPFCA) concentrations were similar in seabirds and ringed seals and ~100-fold higher in polar bears. This may represent a higher biomagnification potential for PFCAs in the seabirds, although adequate field data is not available to test this for the higher chain-length PFCAs.

Compared with other seabird species from arctic regions, PFOS concentrations were ~20-fold lower than those measured in glaucous gulls from the eastern Canadian Arctic (8), presumably due to the higher trophic position of the glaucous gulls. It has been shown through stable isotope analysis that within seabird species at Prince Leopold Island, murres and fulmars both occupy intermediate trophic levels whereas glaucous gulls occupy higher tropic levels (37, 38). However, PFOS levels in the black-legged kittiwake from the eastern Canadian Arctic were ~10-fold lower than in murres and fulmars at Prince Leopold Island, although all three species occupy similar trophic levels (37, 38). This discrepancy is presumably not due to regional differences in PFC levels, since ringed seals from 11 locations across the Canadian Arctic showed overall similar PFC concentrations (39). Most likely, the observed trends can be attributed to species differences in PFC elimination. Metabolic variations have been observed in various seabird species from the Northwater Polynya (40, 41) and Barents Sea (42, 43) with respect to several organochlorine contaminants including PCBs and pesticides.

183 Interestingly, the PFCAs dominated the seabird PFC contaminant profiles (Figure 6.1). This trend is in contrast to most other wildlife samples in which PFOS typically dominates the PFC profiles (3). In murre samples from 2004 and fulmar samples from 2003, ΣPFCAs represented 81% and 93% of the total PFC profile, respectively. These findings suggest a greater exposure of PFCAs and their precursors, particularly the longer chain-length PFCAs, to the seabirds, or an enhanced capacity for PFOS elimination. Similar to seabirds, mink from the Yukon were found to have PFNA concentrations ~2-fold higher than PFOS (4).

Within the PFCAs, seabird contaminant profiles were dominated by the C11-C15 carboxylates with PFTrA (C13) being the predominant compound. In the most recent murre (2004) and fulmar (2003) samples, PFTrA comprised 30% and 24% of the ΣPFCA, respectively, followed by PFUnA (20%) in the murres and PFTA (21%) in the fulmars, respectively. The predominance of the higher chain-length PFCAs in the seabirds is also unique since the major PFCA compound in most wildlife samples is typically either PFNA or PFUnA (3). Few studies have reported the longer chain-length PFCAs in seabirds, thus it is difficult to determine if this trend is characteristic of seabird species. These findings, however, are consistent with those from glaucous gull livers in which PFTrA was the dominant PFCA (28). Within the glaucous gull livers, PFTA and PFPA were not detected, whereas these compounds comprised a significant portion of the PFCA profiles in the northern fulmars and thick-billed murres. A previous study of PFCs in northern fulmars and black guillemots from Prince Leopold Island found detectable but not quantifiable levels of C8-C15 PFCAs (4), although these analytes were found in quantifiable levels in the present study. The apparent discrepancy is presumably due to the much lower detection limits of the present study, as a result of employing a more sensitive LC-MS/MS instrument. MDLs ranged from 0.03-2.3 ng/g in the present study whereas MDLs were 0.5 ng/g (with the exception of PFOA that was 2.0 ng/g) in Martin et al. (4).

The 8:2 and 10:2 FTUCAs were detected in some individuals at low concentrations (supporting information). The 8:2 and 10:2 FTCAs were not detected in any sample. The detection of telomer acids in seabirds supports the hypothesis of FTOHs as a source of PFCAs to the Arctic (11, 12). In addition to their formation from FTOHs via atmospheric oxidation (11, 12), FTCAs and FTUCAs may be formed as intermediate products from FTOHs in rat and fish metabolism (15, 44) and microbial degradation (13, 14). There are few reports of

184

Northern Fulmars 100

80 ΣPFCAs ΣPFSAs ΣTAs

60

40 Percent Composition

20

0 1975 1987 1993 2003

Thick-Billed Murres 100

80

60

40 Percent Composition Percent

20

0 1975 1993 2004

Figure 6.1. Percent composition of total PFCAs (C7-C15 PFCAs), total PFSAs (PFHxS, PFOS, PFDS, PFOSA), total telomer acids (8:2 FTCA & FTUCA, 10:2 FTCA & FTUCA). For calculation of means, concentrations less than the MDL or were not detected were replaced by a random number less than half of the MDL.

185 fluorinated telomer acids in the environment, but they have been detected in North American rainwater samples (45, 46), bottlenose dolphins from the Gulf of Mexico (9) and ringed seals from the Canadian Arctic (10).

6.4.2. PFC-Sex Trends Significant sex-PFC concentration trends were only observed for PFTA in the 2003 fulmars (higher levels in males) and PFPA in the 1993 murres (higher levels in males). These findings are consistent with a general lack of sex differences for PFC concentrations in wildlife and humans (3).

6.4.3. Temporal Trends in Thick-Billed Murres and Northern Fulmars PFC concentrations showed overall increasing concentrations between 1975-2003/2004 in both thick-billed murres and northern fulmars (Figures 6.2 & 6.3). These findings are consistent with the general trend of increasing PFC concentrations in wildlife (3). Analytes were not detected or were

Thick-billed murres showed increasing concentrations of PFCAs through 1975-2004 with doubling times ranging from 2.3 yrs for PFPA to 9.9 yrs for PFDoA (Table 6.1). Concentration increases over both time periods (1975→1993 and 1993→2004) were statistically significant for PFUnA, PFTrA, PFTA and PFPA. The concentration increase from 1993→2004 was statistically significant for PFDoA, but 1975 and 1993 levels were statistically similar since most of these individuals were

186 The differences in PFCA temporal trends between the seabird species may be the result of differing migratory patterns. The contaminant body burden in seabirds represents the integration of exposure from their breeding grounds in the Arctic as well as from overwintering locations. There is evidence that northern fulmars from the Canadian Arctic undertake transoceanic migration to open seas of Northeast Atlantic in winter (51), thus exposure to contaminants may also come from European sources. Data from band return studies indicate that thick-billed murres from Prince Leopold Island overwinter in open water off southwestern Greenland (52), representing more of a North American source. Unfortunately, spatially-based emission trends of PFCAs and their precursors are not known. Similar trends have also been observed for other contaminants in seabird species at Prince Leopold Island (27). For example, mercury concentrations in black-legged kittiwake eggs did not change significantly with time, whereas, levels in northern fulmar and thick-billed murre eggs increased significantly. These differences were attributed to the fact that kittiwakes overwinter in more southern locations where mercury concentrations have decreased with time.

A recent drop in PFOS concentrations in ringed seals from the Canadian Arctic has been reported (10). Similar trends were not observed in the seabirds presumably due the large sampling time intervals, particularly within the late 1990s and 2000s. PFOS concentrations in the murres have increased 39% from 1993 to 2004, however this increase is not statistically significant. In the fulmars, PFOS levels decreased 32% from 1993 to 2003, although this decrease was not statistically significant. These findings suggest the PFOS concentrations remained steady in thick-billed murres and northern fulmars between 1993 and 2003/2004, consistent with trends in ringed seals in which concentrations in 2005 were statistically similar to early 1990s levels (10). It is interesting to note that the overall (from 1975 to 2003/2004) PFOS doubling times for thick-billed murres (10 ± 5.2 yrs) and northern fulmars (9.8 ± 9.8 yrs) were similar to those of 7-10 yrs in guillemot eggs from the Baltic Sea from 1968-2003 (30).

The overall increasing concentrations of PFCs in the seabirds is in contrast to temporal trends of most legacy POPs, whose uses have been largely discontinued in circumpolar countries in the 1970s and 1980s (25, 53). For example, levels of most organochlorine pesticides, PCBs and PCDD/Fs show decreasing trends in seabird eggs from 1975-2003. Notable exceptions include total-chlordane, dieldrin and mirex which have not changed significantly. Further, increases of mercury and β-HCH have been observed. However, some

187 chemicals with widely industrial and consumer use appear to be increasing in seabirds. For example, levels of total PBDEs in fulmar and murre eggs from Prince Leopold Island have increased 9.1- and 4.4-fold, respectively, between 1975-1998 (25). The largely increasing concentrations of PFCs in the seabirds is consistent with the overall increase in the production of known PFC precursors (54). Production volumes of fluorotelomer-based products, precursors to the PFCAs, have increased ~2-fold from 2000 to 2004 (55). Similarly, PFOSF- based products, precursors to PFOS and PFOA, have been produced since the 1950s and reached maximum production in 2000 prior to their withdrawal from production in 2001 (56). As mentioned above, PFOS levels in the seabirds did not decrease following 2001 presumably because of large sampling intervals. Similar to the PFCs, temporal trends of other predominantly transported organohalogen contaminants (PCBs, DDT, PBDEs, α-HCH) generally show good agreement with emission/production estimates (54).

Shifts in diet, as indicated by changes in trophic position, may confound interpretation of contaminant temporal trends (57, 58). For example, the temporary decrease in PFOS concentrations in Lake Ontario lake trout from 1989-1995 has been attributed to food web disruptions resulting from the invasion of zebra mussels (34). Previous analysis of nitrogen stable isotopes in murre and fulmar eggs from Prince Leopold Island showed no consistent shift in trophic position from 1975-2003, suggesting no significant dietary changes (27). Thus, it is most likely that the temporal changes observed are representative of the increases in the PFCs and their precursors in the arctic environment. This also has been observed in sediments from lakes on Cornwallis Island, Nunavut, which have shown increasing PFC concentrations from the 1950s (59).

10 1975 1993 2004 a) 1

0.1

0.01

0.001

* * *** * ** *** ** * *** 0.0001

1975 10 1987 1993 2003 b) 1

0.1 PFCA & Acid Concentration (ng/g Telomer ww) 0.01

0.001

*** *** * * ** ** ****** 0.0001 A A A A A A A A A A A A A Hp FO FN FD Un Do FTr FT FP TC UC TC UC PF P P P PF PF P P P :2 F FT :2 F FT 8 8:2 10 10:2

Figure 6.2. Geometric mean concentration (ng/g ww) of perfluorinated carboxylic acids and fluorotelomer acids in a) thick-billed murres and b) northern fulmars from Prince Leopold Island, Nunavut. Error bars indicate 95% confidence intervals. For calculation of means, concentrations less than the MDL or were not detected were replaced by a random number less than half of the MDL. “*” indicates that all samples were below MDL or were not detected for that time point. 188

Thick-billed Murres Northern Fulmars 3.0

1975 1975 1987 2.5 1993 2004 1993 2003

2.0

1.5

1.0

0.5

PFSA & PFOSA Concentration (ng/g ww) (ng/g Concentration PFOSA & PFSA *** *** *** *** * 0.0 PFHxS PFOS PFDS PFOSA PFHxS PFOS PFDS PFOSA

Figure 6.3. Geometric mean concentration (ng/g ww) of perfluorinated sulfonates and PFOSA in thick-billed murres and northern fulmars from Prince Leopold Island. Error bars indicate 95% confidence intervals. For calculation of means, concentrations less than the MDL or were not detected were replaced by a random number less than half of the MDL. “*” indicates that all samples were below MDL or were not detected for that time point. 189

Table 6.1. Doubling times (years ± 95% confidence interval), geometric mean (range) concentration (ng/g ww) in livers of thick-billed murres (1975-2004) and northern fulmars (1975-2003) from Prince Leopold Island, Nunavut.

Thick-Billed Murres (1975-2004) Northern Fulmars (1975-2003) Doubling 1993 2004 Doubling Analyte r2 P 1975 (n=8) r2 P 1975 (n=9) 1987 (n=8) 1993 (n=10) 2003 (n=9) Time (yr) (n=10) (n=10) Time (yr) 0.08 0.09 0.47 0.18 0.36 0.46 PFDA a na na na 0.309 <0.001 12 ± 8.5 <0.19 (0.01-0.27) (0.01-0.76) (0.11-2.2) (0.03-1.2) (0.04-0.82) (0.22-0.55) 0.66 4.6 0.62 2.5 1.4 PFUnA 0.866 <0.001 4.6 ± 1.1 <0.31 0.549 <0.001 5.5 ± 1.9 <0.31 (0.29-1.7) (2.1-9.4) (0.12-2.2) (0.86-6.8) (0.69-2.8) 0.34 0.15 3.7 0.05 0.38 13.4 0.91 PFDoA 0.489 0.008 9.9 ± 14 0.36 <0.001 5.6 ± 3.3 (0.03-1.6) (0.03-0.87) (1.1-9.7) (0.01-0.28) (0.12-2.1) (8.1-20.4) (0.27-1.7) 0.04 0.40 7.1 1.0 4.0 3.8 PFTrA 0.952 <0.001 3.9 ± 0.5 0.511 <0.001 3.2 ± 1.2 <0.37 (0.02-0.10) (0.14-0.95) (2.7-12.1) (0.43-3.5) (1.9-12.8) (1.3-10.4) 0.16 4.5 0.14 1.2 3.6 2.9 PFTA 0.889 <0.001 3.8 ± 0.8 <0.01 0.627 <0.001 6.4 ± 1.9 (0.01-0.68) (1.9-9.0) (0.02-0.46) (0.28-2.9) (1.9-10.1) (1.1-9.4) 0.01 2.0 0.01 0.77 2.4 2.3 PFPA 0.831 <0.001 2.3 ± 0.7 nd 0.658 <0.001 2.5 ± 0.7 (0.01-0.37) (0.83-3.2) (0.01-0.17) (0.27-2.9) (0.92-5.5) (0.41-10.7) 0.55 0.76 0.12 0.60 0.41 PFOS 0.693 <0.001 10 ± 5.2 <0.40 0.242 0.002 9.8 ± 9.8 <0.40 (0.29-1.3) (0.13-6.1) (0.02-0.51) (0.01-1.9) (0.07-0.85) 0.01 0.01 0.01 0.65 0.13 PFDS a na na na 0.454 <0.001 3.1 ± 1.4 nd 0.07 (0.01-0.53) (0.01-0.38) (0.01-0.61) (0.20-2.4) (0.01-1.6) a Doubling times were not calculated for PFDA and PFDS in the thick-billed murres since >50% of samples had concentrations

191 6.5. Acknowledgements

Thanks to D.N. Nettleship, A.J. Gaston and all of the field crews for their collection of the seabird samples over the years. In particular, we thank K. Allard, A. Fontaine, R. Forbes, A. Greene, E. Greene, D. Jeffrey, K. O’Donovan, I. Storm and P. Taylor for assistance in the field, and M. Asrat, B. Dodge, J. Learning and I. Price for sample handling and preparation in the laboratories at the National Wildlife Research Centre (NWRC). X. Wang of the Aquatic Ecosystem Protection Research Division of Environment Canada is thanked for logistical help with transfer of the samples from NWRC to the analytical laboratory. We are grateful to V. Furdui and E. Reiner (Ontario Ministry of the Environment) for assistance with LC-MS/MS operation. Funding was provided by Environment Canada and the Northern Contaminants Program of Indian and Northern Affairs Canada. Logistical support out of Resolute Bay was provided by the Polar Continental Shelf Project, Natural Resources Canada.

192 6.6. Literature Cited

1. Giesy, J. P.; Kannan, K., Distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 2001, 35, 1339-1342.

2. Olsen, G. W.; Huang, H.-Y.; Helzlsouer, K. J.; Hansen, K. J.; Butenhoff, J. L.; Mandel, J. H., Historical comparison of perfluorooctanesulfonate, perfluorooctanoate, and other fluorochemicals in human blood. Environ. Health Perspect. 2005, 113, 539-545.

3. Houde, M.; Martin, J. W.; Letcher, R. J.; Solomon, K.; Muir, D. C. G., Biological monitoring of polyfluoroalkyl substances: A review. Environ. Sci. Technol. 2006, 40, 3463-3473.

4. Martin, J. W.; Smithwick, M. M.; Braune, B. M.; Hoekstra, P. F.; Muir, D. C. G.; Mabury, S. A., Identification of long-chain perfluorinated acids in biota from the Canadian Arctic. Environ. Sci. Technol. 2004, 38, 373-380.

5. Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G., Dietary accumulation of perfluorinated acids in juvenile rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 189-195.

6. Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G., Bioconcentration and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 196-204.

7. Upham, B. L.; Deocampo, N. D.; Wurl, B.; Trosko, J. E., Inhibition of gap junctional intercellular communication by perfluorinated fatty acids is dependent on the chain length of the fluorinated tail. Int. J. Cancer 1998, 78, 491-495.

8. Tomy, G. T.; Budakowski, W.; Halldorson, T.; Helm, P. A.; Stern, G. A.; Friesen, K.; Pepper, K.; Tittlemier, S. A.; Fisk, A. T., Fluorinated organic compounds in an eastern arctic marine food web. Environ. Sci. Technol. 2004, 38, 6475-6481.

9. Houde, M.; Wells, R. S.; Fair, P. A.; Bossart, G. D.; Hohn, A. A.; Rowles, T. K.; Sweeney, J.; Solomon, K. R.; Muir, D. C. G., Polyfluorinated compounds in free-ranging bottlenose dolphins (Tursiops truncatus) from the Gulf of Mexico and the Atlantic Ocean. Environ. Sci. Technol. 2005, 39, 6591-6598.

10. Butt, C. M.; Muir, D. C. G.; Stirling, I.; Kwan, M.; Mabury, S. A., Rapid response of Arctic ringed seals to changes in perfluoroalkyl production. Environ. Sci. Technol. 2007, 41, 42-49.

11. Ellis, D. A.; Martin, J. W.; De Silva, A. O.; Mabury, S. A.; Hurley, M. D.; Andersen, M. P. S.; Wallington, T. J., Degradation of fluorotelomer alcohols: A likely atmospheric source of perfluorinated carboxylic acids. Environ. Sci. Technol. 2004, 38, 3316-3321.

12. Hurley, M. D.; Wallington, T. J.; Sulbaek Andersen, M. P.; Ellis, D. A.; Martin, J. W.; Mabury, S. A., Atmospheric chemistry of fluorinated alcohols: reaction with Cl atoms and OH radicals and atmospheric lifetimes. J. Phys. Chem. A 2004, 108, 1973-1979.

193 13. Dinglasan, M. J. A.; Ye, Y.; Edwards, E. A.; Mabury, S. A., Fluorotelomer alcohol biodegradation yields poly- and perfluorinated acids. Environ. Sci. Technol. 2004, 38, 2857- 2864.

14. Wang, N.; Szostek, B.; Folsom, P. W.; Sulecki, L. M.; Capka, V.; Buck, R. C.; Berti, W. R.; Gannon, J. T., Aerobic biotransformation of 14C-labeled 8-2 telomer B alcohol by activated sludge from a domestic sewage treatment plant. Environ. Sci. Technol. 2005, 39, 531-538.

15. Martin, J. W.; Mabury, S. A.; O'Brien, P. J., Metabolic products and pathways of fluorotelomer alcohols in isolated rat hepatocytes. Chem. Biol. Interact. 2005, 155, 165-180.

16. D'eon, J. C.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A., Atmospheric chemistry of N- methyl perfluorobutane sulfonamidoethanol, C4F9SO2N(CH3)CH2CH2OH: Kinetics and mechanism of reaction with OH. Environ. Sci. Technol. 2006, 40, 1862-1868.

17. Martin, J. W.; Ellis, D. A.; Mabury, S. A.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of perfluoroalkanesulfonamides: Kinetic and product studies of the OH radical and Cl atom initiated oxidation of N-ethyl perfluorobutanesulfonamide. Environ. Sci. Technol. 2006, 40, 846-872.

18. Martin, J. W.; Muir, D. C. G.; Moody, C. A.; Ellis, D. A.; Kwan, W. C.; Solomon, K. R.; Mabury, S. A., Collection of airborne fluorinated organics and analysis by gas chromatography/chemical ionization mass spectrometry. Anal. Chem. 2002, 74, 584-590.

19. Stock, N. L.; Lau, F. K.; Ellis, D. A.; Martin, J. W.; Muir, D. C. G.; Mabury, S. A., Polyfluorinated telomer alcohols and sulfonamides in the North American troposphere. Environ. Sci. Technol. 2004, 38, 991-996.

20. Shoeib, M.; Harner, T.; Wilford, B. H.; Jones, K. C.; Zhu, J., Perfluorinated sulfonamides in indoor and outdoor air and indoor dust: Occurrence partitioning, and human exposure. Environ. Sci. Technol. 2005, 39, 6599-6606.

21. Ellis, D. A.; Martin, J. W.; Mabury, S. A.; Hurley, M. D.; Sulbaek Andersen, M. P.; Wallington, T. J., Atmospheric lifetime of fluorotelomer alcohols. Environ. Sci. Technol. 2003, 37, 3816-3820.

22. Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H., Sources, fate and transport of perfluorocarboxylates. Environ. Sci. Technol. 2006, 40, 32-44.

23. Armitage, J.; Cousins, I. T.; Buck, R. C.; Prevedouros, K.; Russell, M. H.; MacLeod, M.; Korzeniowski, S. H., Modeling global-scale fate and transport of perfluorooctanoate emitted from direct sources. Environ. Sci. Technol. 2006, 40, 6969-6975.

24. De Silva, A. O.; Mabury, S. A., Isolating isomer of perfluorocarboxylates in polar bears (Ursus maritimus) from two geographical locations. Environ. Sci. Technol. 2004, 38, 6538- 6545.

25. Braune, B. M.; Outridge, P. M.; Fisk, A. T.; Muir, D. C. G.; Helm, P. A.; Hobbs, K.; Hoekstra, P. F.; Kuzyk, Z. A.; Kwan, M.; Letcher, R. J.; Lockhart, W. L.; Norstrom, R. J.; Stern,

194 G. A.; Stirling, I., Persistent organic pollutants and mercury in marine biota of the Canadian Arctic: An overview of spatial and temporal trends. Sci. Total Environ. 2005, 351-352, 4-56.

26. Braune, B. M.; Donaldson, G. M.; Hobson, K. A., Contaminant residues in seabird eggs from the Canadian Arctic. Part I. Temporal trends 1975-1998. Environ. Pollut. 2001, 114, 39- 54.

27. Braune, B. M., Temporal trends of organochlorines and mercury in seabird eggs from the Canadian Arctic, 1975 to 2003. Environ. Pollut. 2007, 148, 599-613.

28. Verreault, J.; Houde, M.; Gabrielsen, G. W.; Berger, U.; Haukås, M.; Letcher, R. J.; Muir, D. C. G., Perfluorinated alkyl substances in plasma, liver, brain, and eggs of glaucous gulls (Larus hyperboreus) from the Norwegian Arctic. Environ. Sci. Technol. 2005, 39, 7439-7445.

29. Kannan, K.; Corsolini, S.; Falandysz, J.; Oehme, G.; Focardi, S.; Giesy, J. P., Perfluorooctanesulfonate and related fluorinated hydrocarbons in marine mammals, fishes, and birds from coasts of Baltic and the Mediterranean Seas. Environ. Sci. Technol. 2002, 36, 3210- 3216.

30. Holmström, K. E.; Järnberg, U.; Bignert, A., Temporal trends of PFOS and PFOA in Guillemot eggs from the Baltic Sea, 1968-2003. Environ. Sci. Technol. 2005, 39, 80-84.

31. Norstrom, R. J.; Won, H. T., Long-term preservation of egg and tissue homogenates for determination of organochlorine compounds: freezing versus freeze-drying. J. Assoc. Off. Anal. Chem. 1985, 68, 129-135.

32. Heinz, G. H.; Miller, D. S.; Ebert, B. J.; Stromborg, K. L., Declines in organochlorines in eggs of red-breasted mergansers from Lake Michigan, 1977-1978 versus 1990. Environ. Monit. Assess. 1994, 33, 175-182.

33. Furdui, V. I.; Stock, N. L.; Ellis, D.; Butt, C. M.; Whittle, D. M.; Crozier, P. W.; Reiner, E. J.; Muir, D. C. G.; Mabury, S. A., Spatial distribution of the perfluoroalkyl contaminants in lake trout from the Great Lakes. Environ. Sci. Technol. 2007, 41, 1554-1559.

34. Martin, J. W.; Whittle, D. M.; Muir, D. C. G.; Mabury, S. A., Perfluoroalkyl contaminants in a food web from Lake Ontario. Environ. Sci. Technol. 2004, 38, 5379-5385.

35. Kannan, K.; Tao, L.; Sinclair, E.; Pastva, S. D.; Jude, D. J.; Giesy, J. P., Perfluorinated compounds in aquatic organisms at various trophic levels in a Great Lakes food chain. Arch. Environ. Contam. Toxicol. 2005, 48, 559-566.

36. Smithwick, M.; Mabury, S. A.; Solomon, K. R.; Sonne, C.; Martin, J. W.; Born, E. W.; Dietz, R.; Derocher, A. E.; Letcher, R. J.; Evans, T. J.; Gabrielsen, G. W.; Nagy, J.; Stirling, I.; Taylor, M. K.; Muir, D. C. G., Circumpolar study of perfluoroalkyl contaminants in polar bears (Ursus maritimus). Environ. Sci. Technol. 2005, 39, 5517-5523.

37. Hobson, K. A.; Welch, H. E., Determination of trophic relationships within a high Arctic marine food web using σ13C and σ15N analysis. Mar. Ecol. Prog. Ser. 1992, 84, 9-18.

195 38. Hobson, K. A., Trophic relationships among high Arctic seabirds: insights from tissue- dependent stable-isotope models. Mar. Ecol. Prog. Ser. 1993, 95, 7-18.

39. Muir, D.; Kwan, M.; Evans, M.; Butt, C.; Mabury, S.; Moore, S.; Sverko, E.; Williamson, M.; Wang, X. Temporal trends of persistent organic pollutants and metals in ringed seals from the Canadian Arctic. 2006. Synopsis of research conducted under the 2005-2006 Northern Contaminants Program. Ottawa, ON, Indian and Northern Affairs Canada.

40. Fisk, A. T.; Moisey, J.; Hobson, K. A.; Karnovsky, N. J.; Norstrom, R. J., Chlordane components and metabolites in seven species of Arctic seabirds from the Northwater Polynya: Relationships with stable isotopes of nitrogen and enantiomeric fractions of chiral components. Environ. Pollut. 2001, 113, 225-238.

41. Buckman, A. H.; Norstrom, R. J.; Hobson, K. A.; Karnovsky, N. J.; Duffe, J.; Fisk, A. T., Organochlorine contaminants in seven species of Arctic seabirds from northern Baffin Bay. Environ. Pollut. 2004, 128, 327-338.

42. Borgå, K.; Gabrielsen, G. W.; Skaare, J. U., Biomagnification of organochlorines along a Barents Sea food chain. Environ. Pollut. 2001, 113, 187-198.

43. Borgå, K.; Wolkers, H.; Skaare, J. U.; Hop, H.; Muir, D. C. G.; Gabrielsen, G. W., Bioaccumulation of PCBs in Arctic seabirds: Influence of dietary exposure and congener biotransformation. Environ. Pollut. 2005, 134, 397-409.

44. Hagen, D. F.; Belisle, J.; Johnson, J. D.; Venkateswarlu, P., Characterization of fluorinated metabolites by a gas chromatographic-helium microwave plasma detector - the biotransformation of 1H,1H,2H,2H-perfluorodecanol to perfluorooctanoate. Anal. Biochem. 1981, 118, 336-343.

45. Loewen, M.; Halldorson, T.; Wang, F.; Tomy, G. T., Fluorotelomer carboxylic acids and PFOS in rainwater from an urban center in Canada. Environ. Sci. Technol. 2005, 39, 2944-2951.

46. Scott, B. F.; Spencer, C.; Mabury, S. A.; Muir, D. C. G., Poly and perfluorinated carboxylates in North American precipitation. Environ. Sci. Technol. 2006, 40, 7167-7174.

47. Smithwick, M.; Norstrom, R. J.; Mabury, S. A.; Solomon, K.; Evans, T. J.; Stirling, I.; Taylor, M. K.; Muir, D. C. G., Temporal trends of perfluoroalkyl contaminants in polar bears (Ursus maritimus) from two locations in the North American arctic, 1972-2002. Environ. Sci. Technol. 2006, 40, 1139-1143.

48. Bossi, R.; Riget, F. F.; Dietz, R., Temporal and spatial trend of perfluorinated compounds in ringed seal (Phoca hispida) from Greenland. Environ. Sci. Technol. 2005, 39, 7416-7422.

49. Bignert, A.; Riget, F.; Braune, B. M.; Outridge, P. M.; Wilson, S., Recent temporal trend monitoring of mercury in Arctic biota - how powerful are the existing data sets? J. Environ. Monit. 2004, 6, 351-355.

50. Hebert, C. E.; Weseloh, D. V. C., Assessing temporal trends in contaminants from long- term avian monitoring programs: The influence of sampling frequency. Ecotoxicology 2003, 12, 141-151.

196 51. Hatch, S.A.; Nettleship, D.N. In The Birds of North America. No. 361; Poole, A., Gill, F., Academy of Natural Sciences: Philadelphia, American Ornithologists' Union: Washington, DC, 1998.

52. Donaldson, G.M.; Gaston, A.J.; Chardine, J.W.; Kampp, K.; Nettleship, D.N.; Elliot, R.D. Canadian Wildlife Service Occasional Paper No. 96; Canadian Wildlife Service: Ottawa, ON, Canada, 1997.

53. Braune, B. M.; Simon, M., Dioxins, furans and non-ortho PCBs in Canadian Arctic seabirds. Environ. Sci. Technol. 2003, 37, 3071-3077.

54. Muir, D.; Braune, B.; Li, Y.-F.; Sverko, E.; Butt, C.; Mabury, S. Can temporal trends of legacy POPs in arctic biota help infer pathways and future trends of new chemicals? 232nd ACS National Meeting, September 10-14, 2006. San Francisco, California.

55. DuPont Global PFOA Strategy - Comprehensive Source Reduction; presented to the USEPA OPPT, January 31, 2005; U.S. Environmental Protection Agency public docket AR226- 1914.

56. 3M. Environmental and health assessment of perfluorooctane sulfonic acid and its salts. USEPA Docket ID AR226-1486; 2003.

57. Hebert, C. E.; Shutt, J. L.; Norstrom, R. J., Dietary changes cause temporal fluctuations in polychlorinated biphenyl levels in herring gull eggs from Lake Ontario. Environ. Sci. Technol. 1997, 31, 1012-1017.

58. Hebert, C. E.; Hobson, K. A.; Shutt, J. L., Changes in food web structure affect rate of PCB decline in herring gull (Larus argentatus) eggs. Environ. Sci. Technol. 2000, 34, 1609-1614.

59. Stock, N. L.; Furdui, V. I.; Muir, D. C. G.; Mabury, S. A., Perfluoroalkyl contaminants in the Canadian arctic: Evidence of atmospheric transport and local contamination. Environ. Sci. Technol. 2007, 41, 3529-3536.

197 6.7. Supporting Information

Prince Leopold Island

Greenland

Northeast Atlantic

Canada

Figure S6.1. Location of sample collection (Prince Leopold Island Migratory Bird Sanctuary, Lancaster Sound, Nunavut, Canada).

Table S6.1. Geometric mean concentrations and ranges (ng/g ww) of perfluorinated acids in thick-billed murres and northern fulmars from Prince Leopold Island, Nunavut.

Thick-Billed Murres Northern Fulmars 1975 1993 2004 1975 1987 1993 2003 (n=8) (n=10) (n=10) (n=9) (n=8) (n=10) (n=9)

0.50 0.02 PFHxA <0.06 <0.06 <0.06 nd <0.06 (0.17-1.3) (0.01-0.13) 0.17 0.09 PFOA <0.33 <0.33 <0.33 <0.33 <0.33 (0.04-0.34) (0.01-1.7) 0.05 0.05 0.16 0.11 0.21 0.05 PFNA <0.19 (0.02-0.16) (0.01-0.24) (0.07-0.60) (0.02-0.97) (0.03-0.44) (0.01-0.09) 0.08 0.09 0.47 0.18 0.36 0.46 PFDA <0.31 (0.01-0.27) (0.01-0.76) (0.11-2.2) (0.03-1.2) (0.04-0.82) (0.22-0.55) 0.66 4.6 0.05 0.62 2.5 1.4 PFUnA <0.31 (0.29-1.7) (2.1-9.4) (0.01-0.28) (0.12-2.2) (0.86-6.8) (0.69-2.8) 0.34 0.15 3.7 0.38 13.4 0.91 PFDoA <0.37 (0.03-1.6) (0.03-0.87) (1.1-9.7) (0.12-2.1) (8.1-20.4) (0.27-1.7) PFTrA 0.04 0.40 7.1 0.01 1.0 4.0 3.8 (0.02-0.10) (0.14-0.95) (2.7-12.1) (0.01-0.17) (0.43-3.5) (1.9-12.8) (1.3-10.4) 0.16 4.5 0.14 1.2 3.6 2.9 PFTA <0.01 (0.01-0.68) (1.9-9.0) (0.02-0.46) (0.27-2.9) (1.9-10.1) (1.1-9.4) 0.01 2.0 0.01 0.77 2.4 2.3 PFPA nd (0.01-0.37) (0.83-3.2) (0.01-0.05) (0.37-1.8) (0.92-5.5) (0.41-10.7) 0.01 PFHxS nd nd nd nd nd nd (0.01-0.05) 0.55 0.76 0.12 0.60 0.41 PFOS <0.40 <0.40 (0.29-1.3) (0.13-6.1) (0.02-0.51) (0.01-1.9) (0.07-0.85) 0.01 0.01 0.01 0.65 0.13 PFDS nd 0.07 (0.01-0.53) (0.01-0.38) (0.01-0.61) (0.20-2.4) (0.01-1.6) 1.3 PFOSA <2.3 <2.3 <2.3 <2.3 <2.3 <2.3 (0.18-5.3) 8:2 FTCA nd nd nd nd nd nd nd 0.01 0.01 0.02 0.01 0.01 0.01 8:2 FTUCA <0.02 (0.001-0.08) (0.002-0.03) (0.008-0.07) (0.003-0.005) (0.001-0.006) (0.001-0.03) 10:2 FTCA nd nd nd nd nd nd nd 0.03 0.04 0.48 10:2 FTUCA <0.20 <0.20 <0.20 <0.20 (0.01-0.09) (0.01-0.22) (0.09-24.6)

198

CHAPTER SEVEN

Spatial Trends of Perfluoroalkyl Compounds in Ringed Seals (Phoca hispida) from the Canadian Arctic

Craig M Butt, Scott A. Mabury, Michael Kwan, Xiaowa Wang and Derek C.G. Muir

Published In: Environmental Toxicology & Chemistry 2008, 27, 542-553

Contributions: Extraction of liver samples, instrumental analysis and data interpretation were performed by Craig Butt. Michael Kwan coordinated the sampling and liver sub-sectioning of recent ringed seal harvests. Xiaowa Wang coordinated the sampling handling at the Canadian Centre for Inland Waters. The manuscript was prepared by Craig Butt with critical comments provided by Scott Mabury and Derek Muir. This research was performed under the guidance of Derek Muir and Scott Mabury

Reproduced with permission from Environmental Toxicology & Chemistry Copyright Allen Press Publishing Services 2008

199 200 Chapter Seven – Spatial Trends of Perfluoroalkyl Compounds in Ringed Seals (Phoca hispida) from the Canadian Arctic

7.1 Abstract The present study examined spatial trends of perfluoroalkyl compounds (PFCs) in liver samples from eleven populations of ringed seals (Phoca hispida) in the Canadian Arctic from 2002 to 2005. Trophic position and relative carbon sources were compared by analyzing stable nitrogen and carbon isotopes in muscle samples. Geometric mean concentrations of total C9-C15 perfluorinated carboxylates (ΣPFCAs) ranged from 8.8 to 84 ng/g wet weight and C9-C11 PFCAs predominated. Perfluorooctane sulfonate (PFOS) was the dominant PFC measured and concentrations ranged from 6.5 to 89 ng/g wet weight, contributing between 29 to 56% of the total PFC concentration. Overall, mean PFC concentrations were similar between populations and differences were largely attributed to elevated levels in the Gjoa Haven (Rae Strait, central Canadian Arctic archipelago) and Inukjuak populations (eastern Hudson Bay), and lower concentrations at Pangnirtung (Cumberland Sound, Baffin Island). Mean stable nitrogen isotope ratios (± 95% confidence intervals) ranged from 14.7‰ (±0.3‰) at Nain (Labrador) to 17.9‰ (±0.7‰) at Gjoa Haven, suggesting all populations were within the same trophic level. Stable carbon isotope ratios varied widely between the seal populations, ranging from -22.9‰ (±0.2‰) at Gjoa Haven to -17.7‰ (±0.4‰) at Nain. The δ13C ratios from Gjoa Haven were significantly more depleted than other populations and may suggest a terrestrially-based carbon source. The depleted stable carbon ratio may explain the elevated PFC concentrations in the Gjoa Haven population. Analysis of covariance (ANCOVA) indicated that δ13C was a significant covariable for seven of nine seal populations for which δ13C values were available. After adjusting for δ13C values, concentrations of most PFCs were generally statistically greater in the Grise Fiord, Qikiqtarjuaq, Arviat and Nain populations.

201 7.2. Introduction

Perfluoroalkyl compounds (PFCs) are a class of contaminants that are globally- distributed in wildlife, including remote regions such as the Canadian Arctic [1]. Perfluorooctanoate (PFOA), perfluorooctane sulfonate (PFOS) and perfluorooctane sulfonamide (PFOSA) were first reported in wildlife from remote regions by Giesy and Kannan [2], followed by Martin et al. [3] who reported long-chain (greater than eight carbons) perfluorinated carboxylates (PFCAs). Since that time, our understanding of the transport pathways to remote regions has increased substantially. Two possible transport mechanisms have been proposed. One potential pathway is the transport of volatile precursors via the atmosphere with subsequent degradation to PFCAs and perfluorinated sulfonates (PFSAs). Volatile precursors include the fluorotelomer alcohols (FTOHs), which degrade via atmospheric oxidation to PFCAs [4, 5], and perfluorinated sulfonamido alcohols, which degrade to both PFCAs and PFSAs in an analogous mechanism [6, 7]. An additional mechanism is the direct transport of PFCAs and PFSAs to remote regions via oceanic transport [8, 9]. While PFCs are presumably transported to the Arctic via a combination of both transport mechanisms, current evidence indicates atmospheric transport as the dominant pathway influencing concentrations in Arctic marine mammals. Fluorotelomer alcohols and sulfonamido alcohols have been detected in the arctic atmosphere [10, 11]. The relatively short doubling times of PFCs in polar bears, as well as the short disappearance half-lives for PFOS in ringed seals following 3M’s PFOS phase-out, support atmospheric transport as the main pathway [12, 13]. Further, time trends of PFCs in Canadian Arctic ice cores, whose PFC levels are derived from atmospheric deposition, were similar to those observed in the ringed seals [14]. The close association between wildlife and atmosphere concentrations provides further support of atmospheric transport as the main pathway influencing PFC concentrations in Arctic marine biota.

There are few large-scale spatial studies of PFCs in a single species [15-17]. Previous studies have included polar bears from the circumpolar Arctic [15], bottlenose dolphins from the Gulf of Mexico and western Atlantic ocean [16], and sea turtles from coastal eastern United States [17]. Differences in contaminant profiles between regions can be indicative of differing source regions. For example, Smithwick et al. [15] observed west-east differences in polar bears from the circumpolar Arctic. Concentrations of C11-C13 PFCAs and PFOS were higher at

202 eastern sites (South Hudson Bay, East Greenland and Svalbard), whereas C9-C10 PFCAs were greatest in western sites (Alaska and western Canadian Arctic). Similar regional disparities in PFC trends may also be expected for ringed seals, a key prey item for polar bears.

Ringed seals (Phoca hispida) are an ideal organism for the examination of spatial trends due to their wide-spread distribution throughout the circumpolar Arctic [18]. Further, ringed seals occupy a relatively small home range [18], and contaminant concentrations are presumably representative of regional marine food web contamination and are not confounded by factors such as accumulation during migration. In addition, ringed seals comprise an important prey source for polar bears [19]; a wildlife species with the highest known PFC concentrations [1].

Stable isotopes of nitrogen and carbon have been used to interpret contaminant trends in food webs. In general, stable isotopes of nitrogen show an enrichment of approximately +3.8‰ per trophic level [20, 21] and act as a surrogate for trophic position. Non-metabolizable, bioaccumulative contaminants have been shown to be positively correlated with trophic position, inferred through stable nitrogen isotope ratios [22, 23]. Further, differences in contaminant levels among wildlife populations have been related to relative trophic position [23]. Conversely, carbon isotopes show relatively little enrichment (<1‰) with increasing trophic position [24] and thus have been used to infer carbon sources. For example, carbon stable isotope ratios have been used to evaluate relative contributions of inshore versus offshore and pelagic versus benthic carbon sources [25].

In the present study spatial trends of PFCs were investigated in ringed seals from 11 locations in the Canadian Arctic collected between 2002 and 2005. Perfluoroalkyl compounds analyzed included C7-C15 PFCAs and their suspected precursors, the 8:2 and 10:2 fluorotelomer saturated carboxylates (FTCA) and fluorotelomer unsaturated carboxylates (FTUCA) as well as

C4, C6, C8 and C10 PFSAs and the PFOSA, a known PFOS precursor. Correlations between PFC concentration and age, sex and stable isotopes were also investigated.

203 7.3. Materials and Methods

7.3.1. Standards and chemicals Standards and reagents used were identical to those previously reported by our laboratory [13].

7.3.2. Sample collection Ringed seal liver samples (n=10 per site) were collected by local subsistence hunters and trappers during annual hunts from 11 locations (year) in the Canadian Arctic; Inukjuak (2002), Pangnirtung (2002), Grise Fjord (2003), Gjoa Haven (2004), Pond Inlet (2004), Arctic Bay (2004), Sachs Harbour (2005), Qikiqtarjuaq (2005), Nain (2005), Arviat (2005), and Resolute Bay (2005). Sample locations are shown in Figure 7.1. All individuals were collected during annual spring hunts from April through June. Samples were sent to the Nunavik Research Centre (Kuujjuaq) for sub-sampling and tooth aging. Ages were determined by longitudinal thin sectioning a lower canine tooth and counting the annual growth layers in the dentine using transmitted light [26]. For samples in which age was not known, ages were estimated using length-age correlations established from a data set of individuals of known ages. Estimated ages were required for only 3 out 10 samples in the Inukjuak population, and 1 out of 10 individuals for Grise Fiord population. Sex was established for most individuals through visual observation in the field. In addition, sex was determined for the Sachs Harbour, Qikiqtarjuaq, Nain, Arviat and Resolute Bay populations using a specific DNA gender marker (Wildlife Genetics International, Nelson, BC, Canada).

7.3.3. Sample extraction and instrumental analysis Samples were extracted using methods similar to the ion-pairing method of Hansen et al. [27] but with the addition of a fluorosolvent clean-up step described elsewhere [13, 28]. Briefly, the methyl tert-butyl ether (MTBE) extract was blown-down to approximately 0.5 ml under a gentle stream of nitrogen gas, an equal volume of hexafluoropropanol (HFP) was added and the precipitated components were removed by filtering. The MTBE/HFP mixture was evaporated to dryness and reconstituted with 500 μl of methanol. The suite of 13C- mass labeled internal standards (see below) was added immediately prior to instrumental analysis.

204 Instrumental analysis was performed by liquid chromatography with negative electrospray tandem mass spectrometry under conditions previously described [13]. Analytes were detected using an API 4000 Q Trap (Applied Biosystems/MDS Sciex, Concord, ON, Canada) with samples injected using an Agilent 1100 autosampler (injection volume = 10 μl, flow rate = 300 μl/min). Chromatography was performed using an ACE C18 column (50 mm x 2.1 mm, 3 μm particle size, Aberdeen, UK), preceded by a C18 guard column (4.0 x 2.0 mm, Phenomenex, Torrance, CA, USA) and the column oven was set to 30 oC.

13 Analyte responses were normalized to internal standard responses. C4-PFOS was used for perfluorobutane sulfonate, perfluorohexane sulfonate (PFHxS), PFOS, perfluorodecane 13 13 sulfonate and PFOSA; C4-PFOA for perfluroheptanoate (PFHpA) and PFOA, C5- 13 perfluorononanoate (PFNA) for PFNA’ C2-perfluorodecanoate (PFDA) for PFDA, perfluoroundecanoate (PFUnA), perfluorododecanoate (PFDoA), perfluorotridecanoate 13 (PFTrA), perfluorotetradecanoate (PFTA) and perfluoropentadecanoate (PFPA); 10:2 C2- 13 FTUCA for 10:2 FTCA and 10:2 FTUCA, and 8:2 C2-FTUCA for 8:2 FTCA and 8:2 FTUCA Perfluorotridecanoate was quantified using the average response factor for PFDoA and PFTA, and PFPA was quantified using the response factor for PFTA since analytical standards were not available. Concentrations were not corrected for recovery. The 10:2 FTCA was not reported due to quantification problems.

7.3.4. Stable isotope analysis Stable isotope ratios of 15N/14N and 13C/12C were determined on individual samples of ringed seal muscle obtained from the lumbar or dorsal region (Environmental Isotope Lab, University of Waterloo, Waterloo, ON, Canada). Samples (~1.0-1.5 mg) were initially freeze- dried and ground to a fine powder with a mortar and pestle. Samples were not pre-extracted to remove lipids. Samples were run on an Isochrom continuous flow stable isotope mass spectrometer (GVInstruments/Micromass, Manchester, UK) coupled to a Carlo Erba Elemental Analyzer (CHNS-O EA 1108, Milan, Italy). Stable isotope ratios were expressed as the difference between the isotope ratio of a sample and standard material:

δx (parts per mil, ‰) = [(Rsample/Rstandard) – 1] ⋅ 1000

205 where x is 13C or 15N and R represents isotopic ratios in the sample and standard (13C/12C, 15N/14N). Standard materials used were Peedee Belemnite formation for carbon and atmospheric air for nitrogen. Approximately 10% of the samples were analyzed in duplicate for nitrogen and carbon analysis. The mean difference between duplicates was 0.10‰ for δ13C and 0.06‰ for δ15N.

7.3.5. Statistical analysis and data treatment Instrument detection limits (IDL) were defined as the concentration that produced a peak with a signal-to-noise ratio of at least three. Method detection limits (MDL) were determined as three times the standard deviation of the procedural blanks. One-half of the IDL was used as the MDL if analytes were not detected in the blanks. Concentrations less than the IDL were reported as nondetect (ND), whereas concentrations less than the MDL were reported as

Prior to statistical analysis, data normality was tested using the Shapiro-Wilk test and homogeneity of variance was tested using the Levene’s test. Perfluoroalkyl compound concentrations were natural-logarithm transformed prior to statistical analysis to meet assumptions of normality and homogeneity of variances. Stable isotope ratios were normally distributed, and thus no transformation was necessary. After transformation, analysis of variance tests were conducted to compare mean PFC concentrations between the sites for each analyte separately. Differences between the sites were assessed post-hoc using the Bonferroni test. An analysis of covariance model, [lnPFC analyte] = δx + site + δx ⋅ site (where x represents either 13C or 15N), was used to analyze for interactions between stable isotope values and location. When nonsignificant interactions were found, post-hoc comparisons were performed with the Bonferroni test. Correlations between stable isotope values and PFC concentration were tested using linear regression. Within each population, differences between sexes were assessed using the t test. Age trends were assessed using linear regression. Analysis of variance and Pearson’s correlation tests were performed only for PFNA, PFDA, PFUnA, PFDoA, PFTrA, PFTA, PFPA, PFOS, and PFOSA since other analytes were not detected or had concentrations that

206 were below the MDL in greater than 50% of individuals. All statistical analysis was performed using SPSS® for Windows (2001, Chicago, IL, USA).

Regional polar bear-ringed seal biomagnification factors (BMF) were calculated from similarly located sample collection sites using liver based PFC concentrations. Biomagnification factors were calculated as,

bear polar bear concentrat (ng/gion wet wt) BMF ringed-bearpolar seal = concentrat seal ringed seal concentrat (ng/gion wet wt)

Geometric mean PFC polar bear concentrations were obtained from Smithwick et al. [15]. For the polar bear data, reported means less than the method detection limit were replaced by one-half of the MDL value. The following polar bear sample sites, as defined by Smithwick et al. [15], were associated with ringed seal populations: Northwest Territories polar bears with Sachs Harbour ringed seals (referred to as Southeast Beaufort Sea); South Hudson Bay polar bears with Arviat ringed seals (referred to as Hudson Bay); South Baffin Island polar bears with Pangnirtung, Qikiqtarjuaq and Nain ringed seals (referred to as South Baffin Island and Labrador); High Arctic polar bears with Resolute Bay, Arctic Bay, Pond Inlet and Grise Fjord ringed seals. The same groupings were applied for the regional comparison of PFC levels in ringed seals. Gjoa Haven and Inukjuak ringed seals were omitted from BMF and regional PFC level calculations since these populations exhibited abnormally elevated concentrations of some PFCs.

7.3.6. Quality control and quality assurance Quality control and assurance procedures included instrumental blanks, procedural blanks (18 MΩ water), matrix spike and recovery samples, and duplicate sample analysis. In addition, comparison between individual subsample and homogenate samples was conducted for nine livers. Details of the quality assurance and quality control program are described elsewhere [13]. Mean analyte recoveries (n=10) were typically greater than 76% with the exception of perfluorobutane sulfonate (56%). Recoveries for PFTrA and PFPA could not be assessed since analytical standards were not available. Twenty percent of the samples were extracted and analyzed in duplicate, and replicate values were generally within 20%.

207 7.4. Results and Discussion

7.4.1. Influence of age and sex on PFCs Sex was known through either field observations or DNA marker testing for 108 out of 110 individuals. In cases when there were inconsistencies between the two methods (two individuals in both the Nain and Resolute Bay populations), the result obtained by DNA marker testing was considered as the true result. Of the 108 individuals, 56 were males and 52 were females, however, the sex ratio within individual populations varied. Overall mean concentrations of PFCs did not differ between sexes. The exceptions were higher concentrations of PFUnA in females from Pond Inlet and higher concentration of PFDA, PFUnA, PFDoA, PFTrA, and PFOS in males from Grise Fiord. These findings are consistent with a general lack of gender differences for PFC concentrations in wildlife and humans [1]. Perfluoroalkyl compounds have been detected in the milk of bottlenose dolphins [29], however, the lack of gender-PFC concentration trends observed in the present study suggests that this is not a significant depuration pathway in arctic ringed seals.

No statistically significant age-PFC concentration correlations were observed in any of the ringed seal populations with the exception of PFOSA in the Gjoa Haven seals (p=0.030, r2 =0.44). The lack of age-PFC concentration trends is suggestive of relatively rapid depuration rates. Fast depuration rates of PFCs have been inferred through the rapid decline of PFOS levels in arctic ringed seals after the termination of PFOS production [13]. Similarly, the half- life of PFOS in bottlenose dolphins has been estimated as 21 weeks [29].

7.4.2. General PFC concentrations in ringed seals Perfluoroalkyl compounds were detected in all ringed seal samples from all 11 populations in the Canadian Arctic (Table 7.1). Perfluorooctanoate was infrequently observed (39 out of 110 individuals) above the MDL level (0.78 ng/g wet wt). Further, PFOA concentrations that were above the MDL were generally low, <2 ng/g wet weight. These findings are consistent with other reports of Arctic wildlife [3, 30] and are presumably due to the low bioaccumulation potential of PFOA [31, 32]. Therefore, these results suggest that, as compared to the long-chain PFCAs and PFOS, PFOA is not a significant PFC contaminant in

Arctic wildlife. Concentrations of PFNA, PFDA, and PFUnA (C9-C11 PFCAs) in the ringed

208 seals generally ranged between 1 to 10 ng/g wet weight, whereas PFDoA, PFTrA, PFTA, and PFPA (C12-C15 PFCAs) were generally less than 1 ng/g wet weight. Perfluorooctane sulfonate concentrations were mainly between 10 to 40 ng/g wet weight, although higher concentrations were measured in the Gjoa Haven and Inukjuak populations. Perfluoroalkyl compound levels were within range previously reported for ringed seals from the Canadian Arctic [3] and Greenland [33]. However, PFOS levels were approximately one order of magnitude lower than those reported in ringed seals from Baltic Sea [34]. Perfluorobutane sulfonate was not detected while PFHpA, PFHxS, and perfluorodecane sulfonate were measured infrequently (<25%) above the MDL.

Saturated and unsaturated fluorotelomer acids were detected in some individuals at low levels. The 8:2 FTCA was not detected in any sample. The 8:2 and 10:2 FTUCAs were detected in all populations but concentrations in most individuals were below the MDL. However, quantifiable levels of 8:2 FTUCA were observed in Grise Fiord seals (geometric mean=6.0 ng/g wet wt) as well, 10:2 FTUCA was observed in Pond Inlet seals (4.0 ng/g wet wt). The detection of fluorotelomer acids in ringed seals supports the hypothesis of FTOHs as a source of PFCAs to the environment [4, 35]. Fluorotelomer saturated carboxylates and FTUCAs have been observed as degradation products of FTOHs in microbial degradation [36, 37], sewage sludge [38] and rat metabolism [39, 40]. Fluorotelomer unsaturated carboxylates have also been measured in bottlenose dolphins [16] and lake trout [28].

Table 7.1. Geometric mean concentrations (ng/g wet wt) and ranges for perfluorinated carboxylates and sulfonates in ringed seals from the Canadian Arctic a. PFHpA PFOA PFNA PFDA PFUnA PFDoA PFTrA PFTA PFPA ΣPFCA Sachs Harbour mean <0.18 <0.78 3.7 2.4 4.9 0.88 1.4 0.21 0.10 14.4 min 0.91 0.75 1.5 0.87 0.37 0.10 0.04 4.2 max 12.9 8.3 14.6 2.2 3.1 0.39 0.17 39.6 Gjoa Haven mean ND 1.07 47.5 15.7 13.4 2.0 2.3 0.25 0.10 83.5 min <0.78 8.9 4.0 4.3 0.68 0.94 0.13 0.05 19.4 max 3.5 100.0 42.5 35.3 4.0 4.2 0.52 0.18 189.0 Resolute Bay mean <0.18 <0.78 4.1 2.8 6.6 0.61 0.68 0.11 0.05 15.8 min 2.2 0.88 2.8 0.55 0.22 0.10 0.03 7.1 max 10.8 5.6 14.7 1.9 1.4 0.43 0.12 30.4 Grise Fjord mean 0.10 1.99 7.5 4.8 7.5 1.9 1.9 0.13 0.02 28.0 min ND 0.93 0.48 1.3 3.8 0.80 0.89 0.19 0.06 8.4 max 7.1 13.9 22.0 12.2 15.3 4.1 4.5 0.81 0.16 74.3 Arctic Bay mean 0.04 0.22 5.9 2.9 4.2 0.71 0.93 0.13 0.04 15.6 min ND <0.78 1.74 0.75 1.84 <0.42 0.48 <0.10 ND 5.7 max 0.19 2.40 14.7 7.59 9.64 1.50 1.69 0.31 0.08 35.3 Pond Inlet mean 0.15 0.22 3.7 1.7 2.7 0.40 0.76 0.11 0.02 11.0 min ND ND 2.0 0.87 1.7 <0.42 0.45 <0.10 ND 6.1 max 9.8 1.6 8.5 4.3 6.9 1.1 1.5 0.28 0.13 22.3 Qikiqtarjuaq mean ND <0.78 9.2 4.9 9.0 1.4 2.2 0.30 0.14 27.8 min 6.7 3.0 5.3 0.73 1.3 0.19 0.10 18.2 max 16.4 7.3 12.2 2.4 3.4 0.39 0.24 36.2 Pangnirtung mean 0.13 1.2 1.8 1.1 2.2 0.32 0.93 0.23 0.03 8.8 min ND <0.78 0.94 0.36 1.0 0.54 0.48 0.16 ND 4.4 max 1.2 2.9 5.1 2.9 5.6 0.94 2.1 0.38 0.16 17.6 Arviat mean ND <0.78 4.4 3.3 10.9 1.7 3.1 0.33 0.17 24.7 min 1.3 1.6 5.8 1.1 2.1 0.23 0.09 13.7 max 9.0 6.2 22.3 3.0 4.9 0.46 0.33 42.3 Inukjuak mean ND 1.5 12.7 10.3 26.2 4.9 5.3 0.65 0.02 64.6 min ND 7.6 4.5 13.0 2.6 2.4 0.30 ND 37.3 max 3.7 26.3 25.7 71.6 8.1 9.3 1.4 0.35 137.5 Nain mean ND 0.19 6.0 3.0 6.7 1.1 2.3 0.32 0.18 20.1 min <0.78 3.12 1.81 3.42 0.45 1.17 0.14 0.08 10.4 max 1.23 17.4 7.90 13.6 1.90 4.05 0.59 0.34 44.7 209

PFHxS PFOS PFDS PFOSA ΣPFSA 8:2 FTUCA 10:2 FTUCA Sachs Harbour mean <0.73 9.6 ND 0.14 11.5 <4.0 <3.1 minimum 0.89 0.03 3.8 maximum 31.3 2.5 31.7 Gjoa Haven mean <0.73 12.8 ND 0.16 61.7 <4.0 <3.1 minimum 61.2 0.07 13.0 maximum 108.4 0.45 108.9 Resolute Bay mean <0.73 6.5 0.02 0.09 6.9 <4.0 <3.1 minimum 1.7 ND 0.02 1.8 ND maximum 16.6 0.07 0.23 17.2 <4.0 Grise Fjord mean <0.73 37.4 ND 2.7 41.0 6.5 <3.1 minimum 13.2 1.2 14.6 ND maximum 96.9 5.6 99.6 11.3 Arctic Bay mean <0.73 10.4 ND 0.15 10.9 <4.0 <3.1 minimum 5.07 0.05 5.4 maximum 18.36 0.46 18.9 Pond Inlet mean <0.73 11.1 ND 0.08 11.5 <4.0 4.0 minimum 5.8 0.03 6.1 <3.1 maximum 20.4 0.24 20.8 4.0 Qikiqtarjuaq mean 0.57 22.1 ND 0.20 23.4 <4.0 <3.1 minimum <0.73 15.0 0.09 16.1 maximum 1.69 44.4 0.37 45.8 Pangnirtung mean 0.18 10.7 ND 0.20 11.3 <4.0 <3.1 minimum ND 3.5 0.10 3.9 maximum 0.80 26.6 0.49 27.1 Arviat mean ND 16.8 ND 0.05 17.2 <4.0 <3.1 minimum 7.6 0.02 7.8 ND maximum 43.7 0.61 44.7 <4.0 Inukjuak mean 2.5 88.8 0.01 1.7 94.9 <4.0 <3.1 minimum 2.0 40.1 ND 0.42 50.8 ND maximum 3.0 189.0 3.3 8.2 198.5 <4.0 Nain mean 0.81 21.9 ND 0.11 23.2 <4.0 <3.1 minimum <0.73 9.7 0.05 11.0 maximum 1.98 48.4 0.31 50.5 a Perfluorobutane sulfonate (PFBS) and 8:2 fluorotelomer saturated acid (8:2 FTCA) were not included since these analytes were not detected in any individual, and 10:2 fluorotelomer saturated acid (10:2 FTCA) was excluded due quantification problems. For calculation of 210

means, concentrations that were less than the method detection limit (MDL) or not detected were replaced by a random number less than half of the MDL. PFHpA=perfluoroheptanoate, PFOA=perfluorooctanoate, PFNA=perfluorononanoate, PFDA=perfluorodecanoate, PFUnA=perfluoroundecanoate, PFDoA=perfluorododecanoate, PFTrA=perfluorotridecanoate, PFTA=perfluorotetradecanoate, PFPA=perfluoropentadecanoate, ΣPFCA=total perfluorocarboxylates PFHxS=perfluorohexane sulfonate, PFOS=perfluorooctane sulfonate, PFOSA=perfluorooctane sulfonamide, ΣPFSA=total perfluorosulfonates. 211

212 7.4.3. Spatial trends in concentration and PFC profiles Overall, statistically significant differences in mean PFC concentrations were observed among the 11 individual ringed seal populations for all analytes statistically tested (C9-C15 PFCAs, PFOS, and PFOSA) but these differences were largely driven by elevated concentrations at Gjoa Haven and Inukjuak, and lower concentrations at Pangnirtung. Concentrations at Gjoa Haven were anomalously higher for some, but not all PFCs (Figure 7.1, full data presented in Table 7.1). Specifically, geometric mean concentrations of PFNA, PFDA, and PFOS at Gjoa Haven were 8.1-fold, 4.2-fold, and 2.6-fold higher than the geometric mean of the other 10 ringed seal populations. Perfluorodecanoate, PFUnA, and PFOS levels were also either the highest or second highest at Gjoa Haven although these mean concentrations were not significantly different from Inukjuak (PFDA, PFUnA, PFOS) and Grise Fiord (PFOS).

The elevated PFC concentrations measured at Gjoa Haven and Inukjuak are likely not the result of the direct contamination from nearby local sources, such as use of aqueous film- forming foams (AFFF). Aqueous film-forming foam contamination is generally associated with high levels of PFSAs (PFHxS, PFOS) and short-chain PFCAs (C6-C8 PFCAs) [11, 41, 42]. Also, a recent study showed enhanced concentrations of these PFCs as well as several long- chain PFCAs (PFNA, PFDA, and PFUnA) [43] in Tomakomai Bay, Japan, a site known to be contaminated with AFFF. Stock et al. [11] observed elevated levels of PFHxS, PFOS, PFHpA, PFOA, and PFNA in lake water and sediment from Resolute Lake, downstream of a small airport at Resolute Bay, which they attributed to AFFF contamination and sewage runoff. Although the outflow of the lake flows directly into Barrow Strait near the hamlet of Resolute Bay, ringed seals collected from the nearby marine waters did not show elevated PFC levels. These results illustrate that potential local contamination from small Arctic communities is unlikely to influence PFC concentrations in wildlife in the marine environment.

To examine geographic trends in concentration, individual seal populations were grouped into four broad regions, excluding Gjoa Haven (see Materials and Methods section for details). Concentrations of most PFCs statistically tested (C9-C15 PFCAs, PFOS, and PFOSA) were statistically similar across the regions with the exception of PFUnA and PFTrA that were statistically greatest in the Hudson Bay population (Figure 7.2). These results are consistent

213 with PFC spatial trends in polar bears which showed elevated levels of PFOS and PFCAs with chain lengths greater than 10 in the Hudson Bay population [15].

Sample collection year between ringed seal populations varied from 2002 to 2005, and it is possible that interpretation of spatial trends may be confounded by temporal variations of PFC concentration within seal populations. Concentrations of PFCAs in ringed seal populations from Arviat and Resolute Bay, from 1998/2000 to 2005, were recently shown to be statistically unchanged, however, PFOS and PFOSA concentrations decreased significantly over this period [13]. Thus, temporal changes in PFC concentrations are presumably not important for PFCAs, but may be relevant for PFOS and PFOSA. For example, the highest PFOS concentrations were measured in the Inukjuak population (geometric mean = 88.8 ng/g wet weight), but these samples were collected in 2002 and thus may have been lower by 2005. Similarly, the lowest PFOS levels were measured in the Resolute Bay population (6.5 ng/g wet weight) and samples were collected during 2005.

Perfluorooctane sulfonate was the dominant PFC measured, contributing between 29 to 56% of total PFC concentration, and is consistent with other reports of PFCs in wildlife [1]. Among the PFCAs, odd number chain-length PFCAs were generally greater than their corresponding even number chain-length PFCA (PFNA>PFOA, PFUnA>PFDA, PFTrA>PFDoA). These odd-even PFCA trends have been observed in other arctic wildlife populations including ringed seals [3, 33] and polar bears [15], and are consistent with FTOH degradation as the source of PFCAs. For example, the atmospheric degradation of 8:2 FTOH yields similar quantities of PFNA and PFOA [4, 35]. This is empirically supported by similar PFOA and PFNA concentrations in North American rainwater [44]. However, the bioaccumulation of PFCAs increases with increasing chain-length [31, 32], resulting in higher biota concentrations of PFNA relative to PFOA. Perfluorinated carboxylate profiles were dominated by either PFNA or PFUnA, contributing between 18 to 58% and 16 to 45% of the total PFCA profile, respectively. These compounds have also been found to predominate in ringed seal populations from Greenland [33] and other regions of the Canadian Arctic [3]. Statistically greater proportions of PFUnA compared to PFNA (paired t test, p>0.05) were measured in the two Hudson Bay seal populations (Arviat and Inukjuak) and Resolute Bay. Perfluorononanoate was statistically greater in the Gjoa Haven, Arctic Bay and Pond Inlet seal

214 populations. Perfluoroundecanoate was predominant in the Sachs Harbour, Pangnirtung and Nain populations although these differences were not statistically significant. Similarly, PFNA was predominant but not statistically greater than PFUnA at Grise Fiord and Qikiqtarjuaq.

Figure 7.1. Geometric mean concentration (ng/g wet wt) of perfluorooctanoate (PFOA), perfluorononanoate (PFNA), perfluorodecanoate (PFDA), perfluoroundecanoate (PFUnA), and perfluorooctane sulfonate (PFOS) in ringed seals from individual locations from the Canadian Arctic. Error bars indicate one standard error.

215

216 Figure 7.2. Geometric mean concentration (ng/g wet wt) of selected perfluorinated carboxylates and sulfonates in ringed seals from the Canadian Arctic. Individual populations have been grouped into four geographic regions. Error bars indicate 95% confidence intervals. Acronyms defined in footnote from Table 7.1.

100

Southeast Beaufort Sea Hudson Bay South Baffin Island & Labrador 10 High Arctic

1

Concentration (ng/g ww) Concentration (ng/g 0.1

0.01 PFOA PFNA PFDA PFUnA PFDoA PFTrA PFTA PFPA PFOS PFOSA

217 7.4.4. Stable isotopes of nitrogen and carbon A significant negative correlation (p<0.001, r2 =0.27) between δ13C and δ15N was observed among the populations (Figure 7.3 and Table 7.2). In general, greater δ15N and more depleted (more negative) δ13C ratios were associated with the western/northern locations (Gjoa Haven, Sachs Harbour, Resolute Bay), whereas the smaller δ15N and more enriched δ13C ratios were associated with the southern/eastern locations (Nain, Pond Inlet). Ratios of δ15N have been shown to be positively correlated with trophic level [20], whereas depleted δ13C ratios are associated with more pelagic or offshore food sources relative to benthic or inshore [45]. Thus, these results indicate that, among the ringed seal populations, a higher trophic position was associated with a greater proportion of pelagic or offshore food sources. Although a negative relationship between δ13C ratio and latitude has been observed in marine phytoplankton [46], differences in stable isotope ratios between seal populations are most likely related to diet variations (with the exception of Gjoa Haven, as explained in a later paragraph). Ringed seal populations, comprising a similar geographical area as the present study, showed unique regional fatty acid signatures indicative of large-scale regional variation in the diets [47].

The half-life of stable nitrogen isotopes in ringed seal tissues has not been measured, but has been observed to be approximately one month in mouse and gerbil muscle [48, 49]. Stable isotope turnover rates are directly proportional to metabolic rate, and metabolic rates are indirectly proportional to body mass [50]. Since ringed seals are considerably larger than mice and gerbils, the isotope turnover rate in ringed seals is presumably much greater than one month. Therefore it is very likely that the δ15N in seal muscle represents several months of dietary intake prior to sampling.

Mean (± 95% confidence intervals) stable nitrogen isotope ratios ranged from 14.7‰ (±0.3‰) at Nain to 17.9‰ (±0.7‰) at Gjoa Haven, and were within the range (16-18‰) reported for arctic ringed seals [20, 51, 52]. Statistical differences in δ15N ratios were observed between the populations and were grouped into five homogeneous but overlapping subsets. The highest δ15N ratios were observed in the Gjoa Haven, Resolute Bay and Sachs Harbour populations, whereas the lowest δ15N ratios were observed in the Nain population. A δ15N enrichment of +3.8‰ per trophic level was previously found in the Barrow Strait-Lancaster

218

Sound marine food web [20]. In the present study, δ15N ratios between populations ranged by a maximum of 3.2‰, suggesting that all seal populations were at approximately the same trophic level. However, it should be noted that direct comparison between seal populations is difficult since it has been shown that baseline δ15N values vary between aquatic systems [53, 54]. Thus, a more appropriate measure of trophic position would have been to calculate δ15N ratios relative to a site specific baseline organism, such as a long-lived primary consumer [53]. Unfortunately, baseline samples were not available due to the logistical challenges of obtaining samples. Similarly, there was small δ15N variation within the individual populations (1.3‰ at Pond Inlet to 3.6‰ at Gjoa Haven), indicating that all seals occupied the same trophic position at each location.

Stable carbon isotope ratios also varied between the seal populations with mean δ13C (± 95% confidence intervals) ranging from -22.9‰ (±0.2‰) at Gjoa Haven to -17.7‰ (±0.4‰) at Nain. Similar to δ15N ratios, the populations could be divided into five homogenous subsets with overlap between some groups. However, the most negative δ13C ratio was measured at Gjoa Haven and was statistically different from the other sites. In contrast to nitrogen isotope ratios, stable carbon isotope ratios typically show little enrichment (<1‰) with increasing trophic level in marine ecosystems [55, 56]. Assuming a similar primary carbon source, the moderate range of δ15N ratios observed among populations (up to 3.2‰) suggests that δ13C ratios between seal populations should show little variation. However, δ13C ratios showed much larger variation than δ15N, with mean δ13C ratios ranging by up to 5.2‰. This suggests differing carbon sources between the seal populations particularly at Gjoa Haven. Further, the mean δ13C ratio at the Gjoa Haven population was 2.5‰ lower than Sachs Harbour (- 20.4‰±0.5), the population with the second smallest δ13C ratio, and this difference was similar to the overall variation among the remaining populations (2.7‰). Small variations in δ13C values (1.2‰) were observed between two ringed seal populations from the North Water Polynya. Variations were attributed to dietary differences between the populations and was confirmed by stomach analysis [51]. However, in contrast to the present study, δ13C and δ15N were positively correlated between the two sites, suggesting a slightly higher trophic level attributed with a more pelagic prey source.

219

The anomalous δ13C ratios measured in the Gjoa Haven seals were much more depleted than is typically reported from their main prey. Ringed seals are known to feed mainly on arctic and polar cod whose reported δ13C values typically range from -21 to -19‰ [45, 57, 58]. Further, invertebrates, a generally smaller diet component, also show similar δ13C values [20, 59]. The depleted carbon signatures in the Gjoa Haven seals are more similar to freshwater or terrestrially-based aquatic systems [60]. There are two possible explanations for a terrestrial signal in the Gjoa Haven population. First, the seals could be feeding predominately on a terrestrially based fish, such as anadromous arctic char. It is anticipated that the char would have a terrestrial carbon signature when they enter the ocean in early summer, but would presumably lose at least some of this terrestrial signal after feeding in the ocean Alternatively, it is possible that the baseline of the Gjoa Haven food web has been shifted to more a terrestrially-based carbon signature as a result of freshwater input. Stable carbon isotope ratios in anadromous arctic char collected near Gjoa Haven in 2005 showed very similar values (- 23.0‰, [61]) to what was observed in the Gjoa Haven ringed seals. The arctic char were collected in early September after feeding in the ocean for one to two months and most probably would have lost their terrestrial carbon signature. These results suggest that the anomalous carbon signature observed in the Gjoa Haven seals is prevalent through the marine food web. Gjoa Haven is located within the Rae Strait, a shallow and relatively isolated region that receives significant freshwater input from the Back River (watershed area=106 500 km2, mean discharge=612 m3/s) [62]. Depleted δ13C ratios also have been observed in arctic cod (Boreogadus saida) from semi-enclosed lagoons in the eastern Beaufort Sea, depressing δ13C values to -24 to -26‰ as compared to -21.5‰ in the coastal Beaufort Sea [63]. Stomach content analysis of the ringed seals and stable carbon isotope analysis of lower food web organisms would help elucidate the source of depleted carbon in the Gjoa Haven population.

220 Figure 7.3. Mean stable isotope ratios of carbon and nitrogen for ringed seals from nine locations in the Canadian Arctic. Error bars indicate 95% confidence intervals. GH=Gjoa Haven, SH=Sachs Harbour, RB=Resolute Bay, AV=Arviat, AB=Arctic Bay, QT=Qikiqtarjuaq, PI=Pond Inlet, LN=Nain. Stable isotope data not available for Pangnirtung and Inukjuak.

19.0

GH 18.0 RB SH

17.0 AV AB

Nitrogen GF

15 16.0

δ QT

PI 15.0

LN

14.0 -24.0 -23.0 -22.0 -21.0 -20.0 -19.0 -18.0 -17.0 δ 13Carbon

221 Table 7.2. Adjusted mean concentrations (ng/g wet wt) of selected perfluorinated acids and mean carbon and nitrogen stable isotope ratios (±95% confidence intervals). Perfluoroalkyl compound (PFC) concentrations adjusted to overall mean δ13C value of -19.2‰ using the analysis of covariance model, [lnPFC] = δ13C + site + δ13C⋅site. Adjusted means not presented for Sachs Harbour and Arctic Bay since these populations were excluded from the analysis of covariance model. a

15 13 δ N δ C PFNA PFDA PFUnA PFDoA PFTrA PFTA PFPA PFOS Gjoa Haven, NU 17.9±0.7 -22.9±0.2 12.8 3.2 2.1 0.26 1.1 0.02 0.08 17.5 Resolute Bay, NU 17.5±0.3 -18.9±0.5 4.5 3.1 7.4 0.69 0.71 0.13 0.05 7.0 Grise Fiord, NU 16.5±0.5 -18.1±0.3 11.0 7.8 12.7 3.4 2.3 0.27 0.02 53.3 Pond Inlet, NU 15.3±0.3 -18.6±0.4 4.5 2.2 3.4 0.53 0.84 0.16 0.02 13.3 Qikitarjuaq, NU 16.4±0.3 -18.7±0.1 10.8 6.0 11.1 1.8 2.4 0.40 0.14 25.7 Arviat, NU 16.6±0.4 -19.1±0.2 4.5 3.4 11.4 1.8 3.2 0.34 0.17 17.3 Nain, NL 14.7±0.3 -17.7±0.4 9.7 5.4 13.3 2.2 2.9 0.82 0.20 34.9 Sachs Harbour, NT 17.3±0.5 -20.3±0.5 ------Arctic Bay, NU16.2±0.6-18.8±0.3------a Acronym definitions are provided in the table 7.1 footnote. δ13C represents the carbon stable isotope ratio; δ15N represents the nitrogen stable isotope ratio.

222 7.4.5. Relationships between stable isotope ratios and PFC concentrations Stable nitrogen isotope ratios have been used as a surrogate for relative trophic position and thus nonmetabolized, bioaccumulative contaminants typically show increasing concentrations with δ15N ratios [22, 23, 64]. Considering PFCs, PFOS levels have been positively correlated with δ15N ratios within marine food webs [30, 65] and among marine mammal species [66]. In addition, enriched δ13C ratios have been correlated with elevated PFOS levels in marine mammals from northern and southwest Norway [67], and from the southern North Sea [66]. These latter studies have associated elevated PFOS concentrations with coastal food sources and are presumably representative of greater exposure to PFOS contamination from continental Europe. Similar trends are not expected in the Canadian Arctic due its low population density (0.02-0.06 person/km2 in Yukon Territory, Northwest Territories and Nunavut, (Statistics Canada, 2007) and suspected lack of direct PFC sources.

Within individual ringed seal populations several statistically significant stable isotope- PFC concentration correlations were observed (p<0.05), although no overall trends were discernable. Statistically significant δ15N-PFC correlations were only observed in seal populations from Sachs Harbour (PFNA, PFDA, PFUnA, PFDoA, PFTrA, PFPA, PFOS, ΣPFC, ΣPFCA, ΣPFSA), Resolute Bay (PFNA, PFOS, PFOSA), Pond Inlet (PFTrA, PFOS, ΣPFSA) and Arviat (PFOS, PFOSA, ΣPFSA). All significant δ15N-PFC correlations were positive, consistent with greater PFC concentrations in those individuals that occupied trophic positions. Significant δ13C-PFC correlations were observed in seal populations from Gjoa Haven (PFNA, PFDA, PFOS, ΣPFC, ΣPFCA, ΣPFSA), Resolute Bay (PFDA, PFUnA, PFDoA, PFTA, PFOS, ΣPFC, ΣPFCA, ΣPFSA) and Qikiqtarjuaq (PFTA, PFPA). No significant stable isotope-PFC correlations were observed within the Grise Fiord, Arctic Bay and Nain seal populations. Thus, significant stable isotope-PFC correlations were observed in only four of nine populations for δ15N and three of nine populations for δ13C when analyzed by location. The low occurrence of statistically significant stable isotope-PFC concentration correlations was presumably due to the small variation in δ15N and δ13C ratios within the ringed seal populations (δ15N range: 1.3‰ at Pond Inlet to 3.6‰ at Gjoa Haven; δ13C range: 0.7‰ at Qikiqtarjuaq to 2.8‰ at Sachs Harbour). Correlations between stable isotopes values and organohalogen contaminant levels have been observed within other wildlife populations from the Canadian Arctic [68]. For

223 example, significant δ15N-Σhexachlorocyclohexane correlations within three Arctic fox populations, however, for this species δ15N ratios were shown to vary up to up to 10 ‰ [68].

When all ringed seal populations were combined, a significant positive correlation (p<0.05) was observed between δ15N and PFC concentration for PFNA, PFDA, PFUnA, PFDoA, and PFTrA but not for PFTA, PFPA, and PFOS. However, these trends appear to be biased by elevated PFC and δ15N values in the Gjoa Haven population since when these samples were removed from the data set, the only statistically significant correlation that remained was for δ15N-PFUnA (p=0.04, r2=0.053). Similarly for δ13C, when considering all sites significant negative δ13C-PFC concentration correlations (p<0.05) were observed for PFNA, PFDA, PFUnA, PFDoA, and PFOS (δ13C-PFNA and -PFUnA correlations shown in Figure 7.4). However, when Gjoa Haven was excluded the only remaining significant correlation was for δ13C-PFOS (p=0.018, r2=0.07). Interestingly, the statistically significant positive δ13C-PFOS correlation observed in the ringed seals, suggesting elevated PFOS levels associated with benthic/inshore carbon sources, is consistent with trends in marine mammals from Europe [66, 67]. However, in the case of the ringed seals, direct PFC sources from nearby small Arctic communities are not expected to be significant. Overall, these results suggest that relative trophic position and carbon source were not good predictors of PFC concentration among the seal populations. This is presumably because there was relatively little variation between seal populations with regards to stable isotope values.

Analysis of covariance was performed using either δ15N or δ13C as the covariate and location as the factor. Considering δ15N as the covariate, the interaction terms between location and δ15N were significant (p<0.05) for most PFCs (PFNA, PFUnA, PFDoA, PFTrA, PFTA, and PFOS). These results indicate that the slopes of the lnPFC concentration-δ15N correlations were significantly different between the populations and thus trophic position was not a suitable covariate in explaining differences between the seal populations. Using δ13C as the covariate, the interaction terms between location and δ13C were not significant (p>0.05) for all PFCs (excluding PFOS) when Sachs Harbour and Arctic Bay were removed from the data set. Ln[perfluoroalkyl compound] concentration-δ13C correlations within these populations were much different (generally positive rather than negative) than the other populations and thus

224 considered to be outliers. Adjusted least square means were compared post-hoc using the Bonferroni test. The adjusted mean concentrations are shown in Table 7.2 (normalized to the mean δ13C value of -19.2‰). Adjusting the PFC concentrations for δ13C had two major consequences. First, whereas elevated concentrations of PFNA, PFDA, and PFUnA were observed in the Gjoa Haven seals, after adjustment levels of these PFCs were statistically similar to the other populations. Second, after adjusting for δ13C values, concentrations of most PFCs were generally statistically greater in the Grise Fiord, Qikiqtarjuaq, Arviat and Nain populations. Thus, excluding Grise Fiord, the populations with the greatest PFC concentrations after adjusting for carbon source using δ13C were located in the southern and eastern regions of the Canadian Arctic. These results are consistent with PFOS spatial trends observed in polar bears [15]. In general, spatial trends for other halogenated organic compounds show little broad geographic trends within the Canadian Arctic Archipelago [69]. However, concentrations of some halogenated organic compounds, such as p,p-DDE have been shown elevated levels in Hudson Bay populations [69, 70]. The variation in individual PFCA and PFOS concentrations among the seal populations was generally greater than observed for other halogenated organic compounds such as PCBs [70].

As previously discussed, the Gjoa Haven population was characterized by elevated concentrations of some PFCs (notably PFNA, PFDA, and PFOS) and a remarkably depleted δ13C ratio. The depleted Gjoa Haven δ13C levels are suggestive of a predominately terrestrial carbon source. Stable nitrogen isotopes were also the highest in the Gjoa Haven seals. However, δ15N levels were statistically similar to the Sachs Harbour and Resolute Bay populations and PFC levels at these locations were not similarly elevated. This suggests that the elevated PFC levels in the Gjoa Haven seals were most probably related to terrestrial inputs as opposed to occupying a higher relative trophic position. Elevated levels of PFDA, PFUnA, and PFOS were also observed in the Inukjuak seals, and may be the result of similar terrestrial influences. It is unknown whether the Inukjuak seals analyzed in the present study are similarly characterized by a terrestrial carbon signature since muscle samples for stable isotope determination were not available. Near offshore to Inukjuak is a group of coastal barrier islands, the Hopewell Islands, and these islands may act to increase the residence time of riverine drainage, and thus terrestrial carbon, in the surrounding marine environment, similar to what has been observed in the Beaufort Sea [63]. However, ringed seals from Inukjuak that

225 were collected from 1989 through 1991 showed only slightly depressed δ13C values (mean= -19.7‰) [71] as compared to the seal populations analyzed in the present study (mean=-18.6‰ when excluding Gjoa Haven and Sachs Harbour). Further investigation of the influence terrestrial runoff on PFC levels in the ocean and marine wildlife is warranted.

The mechanism relating elevated PFC levels and terrestrial inputs is unclear. One potential explanation involves the atmospheric deposition of PFCs, formed by the atmospheric oxidation of FTOHs [4, 5] and sulfonamido alcohols [6, 7] onto the Arctic terrestrial surface, followed by runoff (e.g., of snow melt), eventually flowing into the Arctic Ocean. Perfluoroalkyl compounds would be expected to be fairly mobile within watersheds as a result of their moderate water solubility. Thus, a watershed that drains a significantly large surface area may act to focus contaminants. Interestingly, elevated concentrations of α-HCH (water solubility=1 g/m3 [72]), were observed in polar bears from the Gjoa Haven region [73], consistent with continental runoff as a significant source of moderately water soluble contaminants to marine environments.

226 Figure 7.4. Correlations between ln[perfluoroalkyl compound (PFC)] (ng/g wet wt) and δ13C value (‰) for (a) perfluorononanoate (PFNA) and (b) perfluoroundecanoate (PFUnA) for various ringed seal populations in the Canadian Arctic. Lines represent regression equations for individual populations. Inukjuak and Pangnirtung populations excluded since stable isotope data was not available. Sachs Harbour and Arctic Bay populations excluded since these population had positive ln PFC-δ13C correlations. Symbol legend is applicable to both (a) and (b). z Arviat, S Grise Fiord, T Gjoa Haven, U Nain, „ Pond Inlet, Qikiqtarjuaq, ‘ Resolute Bay.

5 Ln PFNA = -0.35*σ13C - 4.36 + population a) 4

3

2

1 Ln PFNA (ng/g ww) Ln PFNA

0

-1

4

Ln PFUnA = -0.49*σ13C - 6.73 + population b)

3

2 Ln PFUnA (ng/g ww) PFUnA Ln 1

0 -24 -23 -22 -21 -20 -19 -18 -17 -16 δ13C (‰)

227 7.4.6. Ringed seal - polar bear biomagnification factors Regionally based BMFs were calculated by grouping the ringed seal populations to corresponding similarly located polar bear populations as defined by Smithwick et al. [15]. Regional BMFs reflect spatial differences in PFC concentrations and thus are presumably more representative than compiled data from much larger areas. Biomagnification factors were calculated using liver concentrations in both species. Since PFCs have been shown to preferentially accumulate in proteinous tissues such as liver and plasma, biomagnification may have been overestimated. For example, in a bottlenose dolphin food web BMFs calculated using liver concentrations were up to 30-fold higher than when using whole body values [65]. Further, polar bears are known to primarily consume the skin and blubber tissues of ringed seals [19] and thus using liver PFC concentrations may not be accurate of consumption trends. Biomagnification factors are more appropriately calculated using whole body concentrations, however, in the present study only liver tissues were analysed in the ringed seals and polar bears.

Biomagnification factors were greater than 1 for C8-C14 PFCAs as well as PFOS and PFOSA indicating biomagnification of these PFCs in ringed seal-polar bear food webs of the Canadian Arctic (Table 7.3). Biomagnification of PFCs has previously been observed in a polar bear-ringed seal food web from Greenland [74], as well as aquatic food webs in the Great Lakes [75, 76], Canadian Arctic [30], and Atlantic Ocean [65]. Overall, BMFs were greatest for PFOS (geometric mean range 71-163). Biomagnification factors generally decreased with increasing PFCA chain-length, consistent with trends from a bottlenose dolphin food web [65]. Biomagnification factors were generally greatest in the Southeast Beaufort Sea populations, and to some extent in the South Baffin Island and Labrador populations, however most BMFs from the four regions were within a factor of 3. Perfluorooctane sulfonate BMFs from the Canadian Arctic were approximately three-fold greater than from Greenland [74]. As well, hepatic PFOS BMFs were much greater, approximately one order of magnitude, than lipid-based PCB BMFs [77].

Table 7.3. Geometric mean for regionally-based ringed seal-polar bear biomagnification factors for perfluorinated carboxylates and sulfonates a.

PFOA PFNA PFDA PFUnA PFDoA PFTrA PFTA PFPA PFHxS PFOS PFOSA Southeast Beaufort Sea 119 111 43 21 3.5 2.7 1.4 3.0 251 137 6.0 Hudson Bay 125 63 23 10 2.9 3.4 3.1 1.7 373 163 116 South Baffin Island & Labrador 107 40 17 8.8 3.6 2.0 8.5 3.3 163 80 36 High Arctic 45 35 21 7.1 2.3 1.6 3.0 10 285 91 3.6

Canadian Arctic Mean 79 56 23 11 2.8 2.2 3.8 5.5 199 95 16 a Regions similar to those defined by Smithwick et al. [15]. Individual ringed seal populations were grouped according to Northwest Territories polar bears with Sachs Harbour ringed seals (referred to as Southeast Beaufort Sea); South Hudson Bay polar bears with Arviat ringed seals (referred to as Hudson Bay); South Baffin Island polar bears with Pangnirtung, Qikiqtarjuaq and Nain ringed seals (referred to as South Baffin Island and Labrador); High Arctic polar bears with Resolute Bay, Arctic Bay, Pond Inlet and Grise Fjord ringed seals. Gjoa Haven ringed seals were omitted from biomagnification factor (BMF) calculations since this population exhibited abnormally elevated concentrations of some perfluoroalkyl compounds (PFCs). Acronym definitions are provided in the Table 7.1 footnote.

228

229 7.5. Conclusion

Perfluoroalkyl compounds were detected in all individuals from 11 ringed seals populations throughout the Canadian Arctic, confirming their ubiquitous presence in arctic wildlife. Results suggest that FTOHs were a significant source of PFCs to arctic ringed seals due to the predominance of long-chained PFCAs, greater portion of odd versus even PFCAs and the detection of unsaturated and saturated fluorotelomer acids. Overall, PFC concentrations were generally similar between seal populations. However, some statistically significant differences were observed between individual populations with variations largely attributed to elevated levels in the Gjoa Haven and Inukjuak populations, and lower concentrations at Pangnirtung. Stable nitrogen and carbon isotope trends indicated that all ringed seal populations occupied a similar trophic position, but that relative carbon sources varied greatly. The depleted stable carbon isotope ratio measured in the Gjoa Haven seals is suggestive of a terrestrially- based carbon source and may explain the elevated PFC concentration observed in this population. The analysis of covariance model indicated that carbon source (i.e. pelagic vs terrestrial carbon) was a significant covariable and after adjusting for δ13C values, concentrations of most PFCs were generally statistically greater in the Grise Fiord, Qikiqtarjuaq, Arviat and Nain populations.

7.6. Acknowledgements We are grateful to the Hunters and Trappers associations of Nunavut and Northwest Territories for the collection of seal samples. This project would not have been possible without their cooperation. We thank Steve Ferguson (Department of Fisheries and Oceans, Winnipeg) and Magaly Chambellant (University of Manitoba) for providing samples from Arviat and Lois Harwood (Department of Fisheries and Oceans, Yellowknife) for arranging collection of samples at Sachs Harbour. Yan Li is thanked for his assistance with sample processing. Christine Spencer (Environment Canada, Burlington, ON) provided support for use of the liquid chromatography with negative electrospray tandem mass spectrometry. The Northern Contaminants Program (Indian and Northern Affairs Canada) and Environment Canada provided financial support for the project. Wellington Laboratories (Guelph, ON, Canada) is thanked for donation of mass-labeled standards. Craig Butt also appreciates support of the Natural Sciences and Engineering Research Council through a Post Graduate Scholarship.

230 7,7. Literature Cited 1. Houde, M.; Martin, J. W.; Letcher, R. J.; Solomon, K.; Muir, D. C. G., Biological monitoring of polyfluoroalkyl substances: A review. Environ. Sci. Technol. 2006, 40, 3463- 3473.

2. Giesy, J. P.; Kannan, K., Distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 2001, 35, 1339-1342.

3. Martin, J. W.; Smithwick, M. M.; Braune, B. M.; Hoekstra, P. F.; Muir, D. C. G.; Mabury, S. A., Identification of long-chain perfluorinated acids in biota from the Canadian Arctic. Environ. Sci. Technol. 2004, 38, 373-380.

4. Ellis, D. A.; Martin, J. W.; De Silva, A. O.; Mabury, S. A.; Hurley, M. D.; Sulbaek Andersen, M. P.; Wallington, T. J., Degradation of fluorotelomer alcohols: A likely atmospheric source of perfluorinated carboxylic acids. Environ. Sci. Technol. 2004, 38, 3316-3321.

5. Hurley, M. D.; Wallington, T. J.; Sulbaek Andersen, M. P.; Ellis, D. A.; Martin, J. W.; Mabury, S. A., Atmospheric chemistry of fluorinated alcohols: reaction with Cl atoms and OH radicals and atmospheric lifetimes. J. Phys. Chem. A 2004, 108, 1973-1979.

6. D'eon, J. C.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A., Atmospheric chemistry of N-methyl perfluorobutane sulfonamidoethanol, C4F9SO2N(CH3)CH2CH2OH: Kinetics and mechanism of reaction with OH. Environ. Sci. Technol. 2006, 40, 1862-1868.

7. Martin, J. W.; Ellis, D. A.; Mabury, S. A.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of perfluoroalkanesulfonamides: Kinetic and product studies of the OH radical and Cl atom initiated oxidation of N-ethyl perfluorobutanesulfonamide. Environ. Sci. Technol. 2006, 40, 846-872.

8. Armitage, J.; Cousins, I. T.; Buck, R. C.; Prevedouros, K.; Russell, M. H.; MacLeod, M.; Korzeniowski, S. H., Modeling global-scale fate and transport of perfluorooctanoate emitted from direct sources. Environ. Sci. Technol. 2006, 40, 6969-6975.

9. Wania, F., A global mass balance analysis of the source of perfluorocarboxylic acids in the Arctic Ocean. Environ. Sci. Technol. 2007, 41, 4529-4535.

10. Shoeib, M.; Harner, T.; Vlahos, P., Perfluorinated chemicals in the arctic atmosphere. Environ. Sci. Technol. 2006, 40, 7577-7583.

11. Stock, N. L.; Furdui, V. I.; Muir, D. C. G.; Mabury, S. A., Perfluoroalkyl contaminants in the Canadian Arctic: Evidence of atmospheric transport and local contamination. Environ. Sci. Technol. 2007, 41, 3529-3536.

12. Smithwick, M.; Norstrom, R. J.; Mabury, S. A.; Solomon, K.; Evans, T. J.; Stirling, I.; Taylor, M. K.; Muir, D. C. G., Temporal trends of perfluoroalkyl contaminants in polar bears (Ursus maritimus) from two locations in the North American arctic, 1972-2002. Environ. Sci. Technol. 2006, 40, 1139-1143.

231 13. Butt, C. M.; Muir, D. C. G.; Stirling, I.; Kwan, M.; Mabury, S. A., Rapid response of Arctic ringed seals to changes in perfluoroalkyl production. Environ. Sci. Technol. 2007, 41, 42- 49.

14. Young, C. J.; Furdui, V. I.; Franklin, J.; Koerner, R. M.; Muir, D. C. G.; Mabury, S. A., Perfluorinated acids in Arctic snow: New evidence for atmospheric formation. Environ. Sci. Technol. 2007, 41, 3455-3461.

15. Smithwick, M.; Mabury, S. A.; Solomon, K. R.; Sonne, C.; Martin, J. W.; Born, E. W.; Dietz, R.; Derocher, A. E.; Letcher, R. J.; Evans, T. J.; Gabrielsen, G. W.; Nagy, J.; Stirling, I.; Taylor, M. K.; Muir, D. C. G., Circumpolar study of perfluoroalkyl contaminants in polar bears (Ursus maritimus). Environ. Sci. Technol. 2005, 39, 5517-5523.

16. Houde, M.; Wells, R. S.; Fair, P. A.; Bossart, G. D.; Hohn, A. A.; Rowles, T. K.; Sweeney, J.; Solomon, K. R.; Muir, D. C. G., Polyfluorinated compounds in free-ranging bottlenose dolphins (Tursiops truncatus) from the Gulf of Mexico and the Atlantic Ocean. Environ. Sci. Technol. 2005, 39, 6591-6598.

17. Keller, J. M.; Kannan, K.; Taniyasu, S.; Yamashita, N.; Day, R. D.; Arendt, M. D.; Segars, A. L.; Kucklick, J. R., Perfluorinated compounds in the plasma of loggerhead and Kemp's Ridley sea turtles from the southeastern coast of the United States. Environ. Sci. Technol. 2005, 39, 9101-9108.

18. Reeves, R. 1998. Distribution, abundance and biology of ringed seals (Phoca hispida): an overview. In Heide-Jørgensen MP, and Lydersen C, eds, Ringed seals in the North Atlantic, NAMMCO Scientific Publications, Vol. 1. The North Atlantic Marine Mammal Commission, Tromsø, Norway, pp 9-45.

19. Stirling, I.; Øritsland, N. A., Relationships between estimates of ringed seal (Phoca hispida) and polar bear (Ursus maritimus) populations in the Canadian Arctic. Can. J. Fish Aquat. Sci. 1995, 52, 2594-2612.

20. Hobson, K. A.; Welch, H. E., Determination of trophic relationships within a high Arctic marine food web using δ13C and δ15N analysis. Mar. Ecol. Prog. Ser. 1992, 84, 9-18.

21. Jardine, T. D.; Kidd, K. A.; Fisk, A. T., Applications, considerations, and sources of uncertainty when using stable isotope analysis in ecotoxicolgy. Environ. Sci. Technol. 2006, 40, 7501-7511.

22. Fisk, A. T.; Hobson, K. A.; Norstrom, R. J., Influence of chemical and biological factors on trophic transfer of persistent organic pollutants in the Northwater Polyna marine food web. Environ. Sci. Technol. 2001, 35, 73-738.

23. Hop, H.; Borgä, K.; Gabrielsen, G. W.; Kleivane, L.; Skaare, J. U., Food web magnification of persistent organic pollutants in poikilotherms and homeotherms from the Barents Sea. Environ. Sci. Technol. 2002, 36, 2589-2597.

24. Rau, G. H.; Mearns, A. J.; Young, D. R.; Olson, R. J.; Schafer, H. A., Animal 13C/12C correlates with trophic level in pelagic food webs. Ecology 1983, 64, 1314-1318.

232 25. Kelly, J. F., Stable isotopes of carbon and nitrogen in the study of avian and mammalian trophic ecology. Can. J. Zool. 2000, 78, 1-27.

26. Stewart, R. E. A.; Lavigne, D. M. In Proceedings of an International Workshop on the Biology and Management of Northwest Atlantic Harp Seals, Guelph, ON, Canada, 1979.

27. Hansen, K. J.; Clemen, L. A.; Ellefson, M. E.; Johnson, H. O., Compound-specific, quantitative characterization of organic fluorochemicals in biological matrices. Environ. Sci. Technol. 2001, 35, 766-770.

28. Furdui, V. I.; Stock, N. L.; Ellis, D.; Butt, C. M.; Whittle, D. M.; Crozier, P. W.; Reiner, E. J.; Muir, D. C. G.; Mabury, S. A., Spatial distribution of the perfluoroalkyl contaminants in lake trout from the Great Lakes. Environ. Sci. Technol. 2007, 41, 1554-1559.

29. Houde, M.; Balmer, B. C.; Brandsma, S.; Wells, R. S.; Rowles, T. K.; Solomon, K. R.; Muir, D. C. G., Perfluoroalkyl compounds in relation with life-history and reproductive parameters in bottlenose dolphins (Tursiops truncatus) from Sarasota Bay, Florida, USA. Environ. Toxicol. Chem. 2006, 25, 2405-2412.

30. Tomy, G. T.; Budakowski, W.; Halldorson, T.; Helm, P. A.; Stern, G. A.; Friesen, K.; Pepper, K.; Tittlemier, S. A.; Fisk, A. T., Fluorinated organic compounds in an eastern arctic marine food web. Environ. Sci. Technol. 2004, 38, 6475-6481.

31. Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G., Bioconcentration and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 196-204.

32. Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G., Dietary accumulation of perfluorinated acids in juvenile rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 189-195.

33. Bossi, R.; Riget, F. F.; Dietz, R., Temporal and spatial trend of perfluorinated compounds in ringed seal (Phoca hispida) from Greenland. Environ. Sci. Technol. 2005, 39, 7416-7422.

34. Kannan, K.; Corsolini, S.; Falandysz, J.; Oehme, G.; Focardi, S.; Giesy, J. P., Perfluorooctanesulfonate and related fluorinated hydrocarbons in marine mammals, fishes, and birds from coasts of Baltic and the Mediterranean Seas. Environ. Sci. Technol. 2002, 36, 3210- 3216.

35. Wallington, T. J.; Hurley, M. D.; Xia, J.; Wuebbles, D. J.; Sillman, S.; Ito, A.; Penner, J. E.; Ellis, D. A.; Martin, J.; Mabury, S. A.; Nielsen, O. J.; Sulbaek Andersen, M. P., Formation of C7F15COOH (PFOA) and other perfluorocarboxylic acids during the atmospheric oxidation of 8:2 fluorotelomer alcohol. Environ. Sci. Technol. 2006, 40, 924-930.

36. Dinglasan, M. J. A.; Ye, Y.; Edwards, E. A.; Mabury, S. A., Fluorotelomer alcohol biodegradation yields poly- and perfluorinated acids. Environ. Sci. Technol. 2004, 38, 2857- 2864.

233 37. Wang, N.; Szostek, B.; Buck, R. C.; Folsom, P. W.; Sulecki, L. M.; Capka, V.; Berti, W. R.; Gannon, J. T., Fluorotelomer alcohol biodegradation - direct evidence that perfluorinated carbon chains breakdown. Environ. Sci. Technol. 2005, 39, 7516-7528.

38. Wang, N.; Szostek, B.; Folsom, P. W.; Sulecki, L. M.; Capka, V.; Buck, R. C.; Berti, W. R.; Gannon, J. T., Aerobic biotransformation of 14C-labeled 8-2 telomer B alcohol by activated sludge from a domestic sewage treatment plant. Environ. Sci. Technol. 2005, 39, 531-538.

39. Hagen, D. F.; Belisle, J.; Johnson, J. D.; Venkateswarlu, P., Characterization of fluorinated metabolites by a gas chromatographic-helium microwave plasma detector - the biotransformation of 1H,1H,2H,2H-perfluorodecanol to perfluorooctanoate. Anal. Biochem. 1981, 118, 336-343.

40. Martin, J. W.; Mabury, S. A.; O'Brien, P. J., Metabolic products and pathways of fluorotelomer alcohols in isolated rat hepatocytes. Chem. Biol. Interact. 2005, 155, 165-180.

41. Moody, C. A.; Martin, J. W.; Kwan, W. C.; Muir, D. C. G.; Mabury, S. A., Monitoring perfluorinated surfactants in biota and surface water samples following an accidental release of fire-fighting foam into Etobicoke Creek. Environ. Sci. Technol. 2002, 36, 545-551.

42. Moody, C. A.; Hebert, G. N.; Strauss, S. H.; Field, J. A., Occurrence and persistence of perfluorooctanesulfonate and other perfluorinated surfactants in groundwater at a fire-training area at Wurstsmith Air Force Base, Michigan, USA. J. Environ. Monit. 2003, 5, 341-345.

43. Taniyasu, S.; Kannan, K.; So, M. K.; Gulkowska, A.; Sinclair, E.; Okazawa, T.; Yamashita, N., Analysis of fluorotelomer alcohols, fluorotelomer aicds, and short- and long- chain perfluorinated acids in water and biota. J. Chromatogr. A 2005, 1093, 89-97.

44. Scott, B. F.; Spencer, C.; Mabury, S. A.; Muir, D. C. G., Poly and perfluorinated carboxylates in North American precipitation. Environ. Sci. Technol. 2006, 40, 7167-7174.

45. Hobson, K. A.; Fisk, A. T.; Karnovsky, N. J.; Holst, M.; Gagnon, J.-M.; Fortier, M., A stable isotope (δ13C, δ15N) model for the North Water food web: Implications for evaluating trophodynamics and the flow of energy and contaminants. Deep-Sea Research II 2002, 49, 5131-5150.

46. Rau, G. H.; Sweeney, R. E.; Kaplan, I. R., Plankton 13C:12C ratio changes with latitude: differences between northern and southern oceans. Deep-Sea Research 1982, 29, 1035-1039.

47. Thiemann, G. W.; Iverson, S. J.; Stirling, I., Variability in the blubber fatty acid composition of ringed seals (Phoca hispida) across the Canadian Arctic. Mar. Mammal Sci. 2007, 23, 241-261.

48. Tieszen, L. L.; Boutton, T. W.; Tesdahl, K. G.; Slade, N. A., Fractionation and turnover of stable carbon isotopes in animal tissues: Implications for σ13C analysis of diet. Oecologia 1983, 57, 32-37.

49. MacAvoy, S. E.; Macko, S. A.; Arneson, L. S., Growth versus metabolic tissue replacement in mouse tissues determined by stable carbon and nitrogen isotope analysis. Can. J. Zool. 2005, 83, 631-641.

234 50. Gillooly, J. F.; Brown, J. H.; West, G. B.; Savage, V. M.; Charnov, E. L., Effects of size and temperature on metabolic rate. Science 2001, 293, 2248-2251.

51. Holst, M.; Stirling, I.; Hobson, K. A., Diet of ringed seals (Phoca hispida) on the east and west sides of the North Water Polynya, Nothern Baffin Bay. Mar. Mammal Sci. 2001, 17, 888-908.

52. Dehn, L.-A.; Follmann, E. H.; Thomas, D. L.; Sheffield, G. G.; Rosa, C.; Duffy, L. K.; O'Hara, T. M., Trophic relationships in an Arctic food web and implications for trace metal transfer. Sci. Total Environ. 2006, 362, 103-123.

53. Cabana, G.; Rasmussen, J. B., Comparison of aquatic food chains using nitrogen isotopes. Proc. Natl. Acad. Sci. 1996, 93, 10844-10847.

54. Jennings, S.; Warr, K. J., Environmental correlates of large-scale spatial variation in the δ15N of marine animals. Mar. Biol. 2003, 142, 1131-1140.

55. Michener, R. H.; Schell, D. M., Stable isoptope ratios as tracers in marine aquatic food webs. In Stable isotopes in ecology and environmental science, Lajtha, K.; Michener, R. H., Eds. Blackwell: Oxford, UK, 1994.

56. France, R. L.; Peters, R. H., Ecosystem differences in the trophic enrichment of 13C in aquatic food webs. Can. J. Fish Aquat. Sci. 1997, 54, 1255-1258.

57. Hoekstra, P. F.; Dehn, L.-A.; George, J. C.; Solomon, K. R.; Muir, D. C. G.; O'Hara, T. M., Trophic ecology of bowhead whales (Balaena mysticetus) compared with that of other arctic marine biota as interpreted from carbon-, nitrogen-, and sulfur-isotope signatures. Can. J. Zool. 2002, 80, 223-231.

58. Dunton, K. H.; Schell, D. M., Dependence of consumers on macroalgal (Laminaria solidungula) carbon in an arctic kelp community: δ13C analysis. Mar. Biol. 1987, 93, 615-625.

59. Dehn, L.-A.; Sheffield, G. G.; Follmann, E. H.; Duffy, L. K.; Thomas, D. L.; O'Hara, T. M., Feeding ecology of phocid seals and some walrus in the Alaskan and Canadian Arctic as determined by stomach contents and stable isotope analysis. Polar Biol. 2007, 30, 167-181.

60. Schoeninger, M. J.; DeNiro, M. J., Nitrogen and carbon isotopic composition of bone collagen from marine and terrestrial animals. Geochim. Cosmochim. Ac. 1984, 48, 625-639.

61. Evans MS, Muir D, Keating J, Wang X, Sardela G, Sverko E, Brozowski J, Gantner N. 2006. Temporal trends and spatial variations in persistent organic pollutants and metals in sea- run char from the Canadian Arctic. Synopsis of research conducted under the 2005-2006 Northern Contaminants Program. Ottawa, ON, Indian and Northern Affairs Canada.

62. Tews, J. 2005. Back River. In Nuttal M, ed, Encyclopedia of the Arctic. Volume 1: A-F. Routledge, New York, NY, USA.

63. Dunton, K. H.; Weingartner, T.; Carmack, E. C., The nearshore western Beaufort Sea ecosystem: Circulation and importance of terrestrial carbon in arctic coastal food webs. Prog. Oceanogr. 2006, 71, 362-378.

235 64. Rasmussen, J. B.; Rowan, D. J.; Lean, D. R. S.; Carey, J. H., Food chain structure in Ontario lakes determines PCB levels in lake trout (Salvelinus namaycush) and other pelagic fish. Can. J. Fish Aquat. Sci. 1990, 47, 2030-2038.

65. Houde, M.; Bujas, T. A. D.; Small, J.; Wells, R. S.; Fair, P. A.; Bossart, G. D.; Solomon, K. R.; Muir, D. C. G., Biomagnification of perfluoroalkyl compounds in the bottlenose dolphin (Tursiops truncatus) food web. Environ. Sci. Technol. 2006, 40, 4138-4144.

66. Van de Vijver, K. I.; Hoff, J. T.; Das, K.; Van Dongen, W.; Esmans, E. L.; Jauniaux, T.; Bouquegneau, J. M.; Blust, R.; De Coen, W. M., Perfluorinated chemicals infiltrate ocean waters: Link between exposure levels and stable isotope ratios in marine mammals. Environ. Sci. Technol. 2003, 37, 5545-5550.

67. Van de Vijver, K. I.; Hoff, P. T.; Das, K.; Van Dongen, W.; Esmans, E. L.; Siebert, U.; Bouquegneau, J. M.; Blust, R.; De Coen, W. M., Baseline study of perfluorochemicals in harbour porpoises (Phocoena phocoena) from Northern Europe. Mar. Pollut. Bull. 2004, 48, 992-997.

68. Hoekstra, P. F.; Braune, B. M.; O'Hara, T. M.; Elkin, B.; Solomon, K. R.; Muir, D. C. G., Organochlorine contaminant and stable isotope profiles in Arctic fox (Alopex lagopus) from the Alaskan and Canadian Arctic. Environ. Pollut. 2003, 122, 423-433.

69. Weis, I. M.; Muir, D. C. G., Geographical variation of persistent organochlorine concentrations in blubber of ringed seal (Phoca hispida) from the Canadian Arctic: Univariate and multivariate approaches. Environ. Pollut. 1997, 96, 321-333.

70. Muir, D. C. G.; Norstrom, R. J., Geographical differences and time trends of persistent organic pollutants in the Arctic. Toxicol. Lett. 2000, 112-113, 93-101.

71. Muir, D. C. G.; Segstro, M. D.; Hobson, K. A.; Ford, C. A.; Stewart, R. E. A.; Olpinski, S., Can seal eating explain elevated levels of PCBs and organochlorine pesticides in walrus blubber from eastern Hudson Bay (Canada)? Environ. Pollut. 1995, 90, 335-348.

72. Mackay D, Shui WY, Ma K-C, Lee, SC. 2006. Handbook of Physical-Chemical Properties and Environmental Fate for Organic Chemicals. 2nd ed, Volume IV - Nitrogen and Sulfur Containing Compounds and Pesticides. CRC Press, Boca Raton, FL, United States.

73. Norstrom, R. J.; Simon, M.; Muir, D. C. G.; Schweinsburg, R. E., Organochlorine contaminants in arctic marine food chains: Identification, geographic distribution, and temporal trends in polar bears. Environ. Sci. Technol. 1988, 22, 1063-1071.

74. Bossi, R.; Riget, F. F.; Dietz, R.; Sonne, C.; Fauser, P.; Dam, M.; Vorkamp, K., Preliminary screening of perfluorooctane sulfonate (PFOS) and other fluorochemicals in fish, birds and marine mammals from Greenland and the Faroe Islands. Environ. Pollut. 2005, 136, 323-329.

75. Martin, J. W.; Whittle, D. M.; Muir, D. C. G.; Mabury, S. A., Perfluoroalkyl contaminants in a food web from Lake Ontario. Environ. Sci. Technol. 2004, 38, 5379-5385.

236 76. Kannan, K.; Tao, L.; Sinclair, E.; Pastva, S. D.; Jude, D. J.; Giesy, J. P., Perfluorinated compounds in aquatic organisms at various trophic levels in a Great Lakes food chain. Arch. Environ. Contam. Toxicol. 2005, 48, 559-566.

77. AMAP. 2004. AMAP Assessment 2002: Persistent Organic Pollutants (POPs) in the Arctic. Arctic Monitoring and Asessment Programme (AMAP), Olso, Norway.

CHAPTER EIGHT

Bioaccumulation and Biotransformation of the 8:2 FTOH Acrylate in Rainbow Trout

Craig M. Butt, Derek C.G. Muir and Scott A. Mabury

In preparation for submission to Environmental Science & Technology

Contributions: Fish care and sampling, preparation of dosed food, fish dissection, tissue extraction, synthesis of glucuronide and sulfate conjugates, S9 incubation experiments, instrumental analysis, data interpretation and manuscript preparation were performed by Craig Butt under the guidance of Derek Muir and Scott Mabury.

237 238 Chapter Eight – Bioaccumulation and Biotransformation of the 8:2 FTOH Acrylate in Rainbow Trout

8.1. Abstract

The bioaccumulation and biotransformation of the 8:2 fluorotelomer acrylate

(C8F17CH2CH2OC(O)CH=CH2, 8:2 FTAc) was investigated using rainbow trout as the animal model. Two experiments were preformed, an in vivo dietary exposure and in vitro incubations with liver and stomach S9 fractions. The 8:2 FTAc is a monomer used in the manufacture of fluorinated polymers and has been detected in the atmosphere. During the in vivo study, concentrations of the parent 8:2 FTAc and suspected intermediate and terminal metabolites were monitored during the 5 day uptake and 8 day elimination phases using a combination of GC-MS and LC-MS/MS. Liver, blood and kidney were monitored as well as bile and feces. Very low levels of the 8:2 FTAc were detected in the internal tissues (kidney and liver only) and feces, suggesting that the 8:2 FTAc was rapidly biotransformed in the gut or liver. Similarly, low concentrations of the 8:2 FTOH – the first metabolite formed – were accumulated in the fish tissues. The 8:2 FTCA and 8:2 FTUCA were also formed quickly and reached steady-state concentrations by 24 hours. The 8:2 FTCA was formed in the highest concentration, reaching steady-state tissue concentrations of ~1000-1400 ng/g ww. The 8:2 FTUCA and 7:3 FTCA was also accumulated in high levels, at levels about 10-fold lower than the 8:2 FTCA. Both the 7:3 FTCA and PFOA showed increasing levels throughout the uptake phase and into the initial stages of the elimination phase, indicating continued formation through precursors still present in the body. PFNA was also formed in low ng/g ww levels. The intermediate and terminal metabolites were also detected in the bile and feces indicating an important elimination pathway for these compounds. In addition, the 8:2 FTOH glucuronide conjugate was measured in relatively high concentrations in the bile and feces. In the in vitro S9 incubation experiments, both the stomach and liver showed very high and apparently similar esterase activity towards the 8:2 FTAc. These results demonstrate the potential for considerable extrahepatic metabolism of the 8:2 FTAc prior to uptake into the internal tissues, ultimately limiting the overall bioaccumulation. Therefore, the results of the present study demonstrated a situation in which a biologically-labile compound is biotransformed to terminal metabolites that are potentially very biologically persistent.

239 8.2. Introduction

Perfluorinated carboxylic acids (PFCAs) are now recognized as global environmental pollutants and are ubiquitously detected in most environmental compartments including humans [1, 2] and wildlife [3-5] as well as abiotic media such as seawater [6]. Further, PFCA contamination has reached remote environments, including the Arctic [3, 4] and Antarctic [7]. Despite their widespread occurrence, the sources of PFCA contamination are not well understood. Early research efforts investigated the potential of “indirect” pathways as sources of PFCAs to the environment. This pathway postulates that precursor compounds, some of which have been shown to be present in commercial products [8, 9], degrade via abiotic and biotic reactions to PFCAs. Early work by our group showed that fluorotelomer alcohols (FTOHs) form PFCAs via atmospheric oxidation [10, 11]. Further research by our group and other laboratories have shown that fluorinated sulfonamide alcohols [12, 13], fluorotelomer olefins [14, 15], fluorotelomer iodides [16] and fluorotelomer acrylates [17] are also potential PFCA precursors through atmospheric oxidation. FTOHs may also undergo indirect photolysis in aqueous systems [18], although this loss mechanism is presumably of negligible importance given the high Henry’s Law constants of FTOHs [19]. The biotransformation of a FTOH compound to a PFCA was first reported by Hagen et al. [20] who investigated the fate of 8:2 FTOH in rats. Additional research has shown the formation of PFCAs from FTOHs in microbial systems [21-23], in vivo rat [24, 25] and mice [26, 27] experiments, and isolated hepatocytes of rat [24, 25, 28], mouse [28], rainbow trout [28] and humans [28]. Further, PFCAs have been shown to be formed during the biotransformation of fluorotelomer-based phosphates in rats [29].

In addition, it has been postulated that historical PFCA emissions have resulted from the “direct” emission of PFCAs during fluorochemical manufacturing processes and as residuals in consumer products [30]. Specifically, significant quantities of perfluorooctanoate (PFOA) have been directly manufactured. Perfluorononanoate (PFNA) has also been produced directly, with perfluoroundecanoate (PFUnA) and perfluorotridecanoate (PFTrA) detected as impurities in a single PFNA formulation [30].

Fluorotelomer acrylates (FTAc, general formula: CxF2x+1CH2CH2OC(O)CH=CH2) are monomers used in the manufacture of fluorotelomer-based polymers. Fluorotelomer-based

240 polymers have wide commercial uses, primarily for their stain repellency properties. It has been reported that the largest commercial category of polyfluorinated products is FTAc-based polymers [9]. FTAcs have been produced on the order of 100s of kilotonnes per year [31]. Fluorotelomer acrylates have been shown to be significant residuals in a FTAC-based polymer [9], and may be released to the environment as residuals from commercial fluorotelomer-based polymer products in a similar manner to FTOHs [8]. The 6:2, 8:2 and 10:2 FTAc have been detected in the atmosphere, including remote regions, at generally low pg/m3 levels [32-36].

The bioaccumulation of fluorotelomer-based precursors is of interest since their metabolism may form compounds (i.e. PFCAs) that are much more persistent and bioaccumulative than the parent compounds themselves. PFCAs with chain lengths greater than seven carbons have been shown to be bioaccumulative [37]. In addition, it has been shown that fluorotelomer acids (FTCAs), which are intermediate metabolites of FTOH biotransformation, are much more toxic than PFCAs [38]. Further, Martin et al. [39] showed the FTOH metabolites (i.e. fluorotelomer aldehydes, FTALs) were in fact responsible for observed toxic effects rather than the parent FTOH itself.

Xenobiotic metabolism is frequently studied in vitro using perfused tissues, tissue slices, primary cell cultures (e.g. hepatocytes) and sub-cellular fractions (i.e. cytosol and microsomes) [40]. In vitro techniques are attractive due to their minimal animal use and relative speed. Further, as compared to in vivo studies, metabolite quantification is simplified since excretion does not occur (i.e. parent compound and biotransformation product remain in the experimental vessel). Of the in vitro techniques, sub-cellular fractions are technically the easiest to prepare and use [40]. A crude purification of the tissue homogenate will yield the cytosol or “S9” fraction, whereas further purification will yield the microsomal fraciton. The S9 fraction contains the complement of phase I and II enzymes, whereas the microsomal fraction primarily contains the phase I enzymes.

The goal of this study was to investigate the biological fate of the 8:2 fluorotelomer acrylate (C8F19CH2CH2OC(O)CH=CH2, 8:2 FTAc), ultimately assessing its potential as a PFCA precursor through biotransformation. It was hypothesized that the 8:2 FTAc would undergo hydrolysis to the 8:2 FTOH, and subsequent oxidation, eventually forming PFCAs. In this study we chose to use rainbow trout (Oncorhynchus mykiss) as the animal model, performing two

241 complementary experiments. The first experiment consisted of the in vivo dietary exposure of the 8:2 FTAc with the objective of examining overall 8:2 FTAc disposition, identification of biotransformation products, relative metabolite yields and elimination kinetics. The second experiment consisted of in vitro 8:2 FTAc incubations in stomach and liver S9 fractions with the objective of quantifying 8:2 FTAc biotransformation kinetics.

8.3. Materials and Methods

8.3.1. Standards and Reagents The 8:2 FTAc (1H, 1H, 2H, 2H-perfluorodecyl acrylate, 97%), 8:2 FTOH (1H, 1H, 2H, 2H-perfluoro-1-decanol, 97%), 10:2 FTOH (1H, 1H, 2H, 2H-perfluoro-1-dodecanol, 97%), and 7:3 FTCA (2H, 2H, 3H, 3H-perfluorodecanoic acid, 97%) were obtained from Synquest Labs Inc. (Alachua, FL). Ethyl acetate and methanol (OmniSolv grade, 99.9%) were purchased from EMD Chemicals Inc. (Gibbstown, NJ). Standards for PFHpA, PFOA, PFNA, 8:2 FTCA and 8:2 13 13 13 13 FTUCA as well as the stable isotope standards ( C4-PFOA, C5-PFNA, C2-8:2 FTUCA, C2- 13 8:2 FTCA, C4-8:2 FTOH) were provided by Wellington Laboratories (Guelph, ON).

The 8:2 FTOH sulfate and 10:2 FTOH sulfate (CF3CFxCH2CH2OSO3H, where x=7 or 9) were synthesized following methods described by Chaudier et al. [41]. Briefly, the FTOH and 1.3 moles equivalent of sulfur trioxide pyridine complex (Sigma-Aldrich, Oakville, ON) were dissolved in a minimal quantity of dimethylformamide and heated at 50oC for 72 hours. The dimethylformamide was evaporated, crude mixture reconstituted in water, passed through sodium ion exchange resin (Amberlite IRA 200, Alfa Aesar, Ward Hill, MA), cleaned up on a C18 solid-phase extraction cartridge and lyophilized. Products were confirmed by 19F- and 1H-

NMR as well as LC-MS/MS analysis. The 8:2 FTOH glucuronide (CF3CF7CH2CH2O-gluc) was synthesized using methods adapted from Xu et al. [42]. Briefly, the 8:2 FTOH (0.35 mmol) and acetobromo-α-D-glucuronic acid methyl ester (0.75 mmol, Sigma-Aldrich) were dissolved in anhydrous chloroform (10 mL), stirred for 10 min at room temperature and then cooled to - 15oC. Silver triflate (0.50 mmol) was added and mixture was warmed to ambient temperature over 2 hours. Triethylamine (3 mL) and water (10 mL) were added, the mixture was shaken.

242 The chloroform phase was separated and extracted with water, followed by brine, dried with sodium sulfate and lyophilized.

8.3.2. In vivo 8:2 FTAc Dietary Exposure 8.3.2.1. Food Preparation Control and dosed food was prepared in a manner adapted from Stapleton et al. [43] in effort to minimize volatilization of the 8:2 FTAc. Briefly, commercial fish food (Martin’s floating feed, size 3, Martin Mills, Elmira, ON) was ground into a fine powder using a laboratory blender. The 8:2 FTAc was added to 20 ml of canola oil in a polypropylene centrifuge tube, gently inverting the tube for 1 minute. The 8:2 FTAc/canola oil mixture was added to 200 g of ground fish food powder and thoroughly mixed. To pelletize the food, 200 ml of a 5% gelatin mixture was added and the resulting paste was passed through a kitchen food mill. The reformed food was approximately the same size of the original food pellets. The food was allowed to set and dry in the fume hood for 12 hours. The 8:2 FTAc concentration in the dosed food was 93 ± 9.5 µg/g. The control food was prepared in an identical manner as the dosed food with the exception that the 8:2 FTAc was not added to the canola oil. The 8:2 FTOH was not detected in the FTAc dosed food, and the 8:2 FTAc and 8:2 FTOH was not detected in the control food. Food was kept at 4oC until immediately before feeding. All components used in the food preparation were initially thoroughly rinsed with methanol.

8.3.2.2. Fish Care and Sampling Juvenile rainbow trout were purchased from a local hatchery (Rainbow Springs Trout Hatchery, Orangeville, Ontario) and allowed to acclimate for 2 weeks prior to dosing. The initial fork length was approximately 6 inches and the initial weight was approximately 45 g. Trout were kept at the Aquatic Facility, Department of Cell and Systems Biology (University of Toronto) in 475 L fiberglass tanks under flow-through conditions (~10 L/min) using carbon- filtered, dechlorinated City of Toronto water. The water temperature was maintained at 18oC and a 12 hour daily photoperiod was used. The initial fish loadings were ~4 g/L. Control fish were kept in 1 tank, whereas, the “dosed” fish were divided equally between 2 tanks. Care and treatment of the fish was approved by the University of Toronto Animal Care Committee and was in compliance with the guidelines of the Canadian Council on Animal Care.

243 A preliminary range-finding study in our lab indicated that the 8:2 FTAc biotransformation occurred very rapidly (results not shown). As such, an intensive feeding and sampling schedule was designed, consisting of a 120 hour uptake phase and 192 hour elimination phase. Fish were starved 24 hrs prior to initial dosing. During the uptake phase fish were fed the dosed or control food twice per day at 1.5% of the average initial body weight per feeding. Fish were collected at -1 hour (predose), 1 hr, 2 hr, 4 hr, 6 hr, 8 hr, 12 hr, 18 hr, 24 hr, 36 hr, 48 hr, 60 hr, 72 hr, 96 hr and 120 hr. During the elimination phase, fish were fed unaltered commercial food once per day. Elimination samples were collected at 12 hr, 24 hr, 36 hr, 48 hr, 72 hr, 96 hr, 144 hr and 192 hr. During all time points, 3 dosed fish and 1 control were collected. Fish were euthanized by an overdose exposure to MS-222 (4 g/L, buffered to pH 7 with sodium carbonate) and blood was immediately drawn through cardiac puncture using heparin rinsed syringes. Fish were dissected in the laboratory and the liver, kidney and bile were removed. The feces were collected from the entire intestine. Blood, tissue and feces samples were kept frozen (-20 oC) until analysis. For the bile only, samples could not be collected from 6hr control, and only two “dosed” samples per time point could be collected from the predose, 4hr and 12hr-elimination samples.

8.3.2.3. Extraction and Instrumental Analysis Liver, kidney (~0.5 g each), bile (~0.05 g), feces (~0.25 g) and blood (300 µl) were sub- sampled and placed in 15 ml polypropylene centrifuge tubes. The suite of internal standards 13 13 13 13 13 ( C4-PFOA, C5-PFNA, C2-8:2 FTUCA, C2-8:2 FTCA, 10:2 FTOH sulfate, C4-8:2 FTOH) was added to each sample prior to extraction. Liver and kidney samples were homogenized for 1 min using a mechanical mixer (Tissue-Tearor™, Biospec Products, Bartlesville, OK) in 8 ml of ethyl acetate. Homogenization was not required for the bile and feces. The extracts were centrifuged and the solvent layer split equally into two fractions for LC-MS/MS and GC-MS analysis, respectively. The blood samples were extracted by gently shaking for 5 min with 10 ml of ethyl acetate. The extracts were centrifuged, split into equal fractions and the extraction was repeated. The extracts that were to be analyzed on the LC-

MS/MS were blown down to dryness under a gentle stream of N2 gas and reconstituted in 1 ml of methanol. Extracts for GC-MS analysis were reduced in volume to 1 ml and transferred to a GC vial for analysis.

244 8.3.2.4. Analytical Methods A table summarizing the common name, acronym and structure of the analytes monitored is presented in the supporting information. The 8:2 FTAc and 8:2 FTOH were monitored by gas chromatography-mass spectrometry using an Agilent 7890 gas chromatography coupled to an Agilent 5975 mass selection detector. The instrument was operated in positive chemical ionization under selected ion monitoring mode. Separation was achieved on a 30m RTX-Wax column (0.25 mm ID x 0.25 µm film thickness, Restek, Bellefonte, PA). The GC oven program employed was: initial temperature of 60oC held for 1 min, 5oC/min to 75oC, 10oC/min to 130oC, 50oC/min to 240oC, hold 1 min. Analytes were quantified using their molecular ion (M+1) with the stable isotope of the 8:2 FTOH (M+4) as the internal standard.

Instrumental analysis for non-volatile metabolites (C6-C9 PFCAs, 8:2 FTCA & FTUCA, 7:3 FTCA & FTUCA, 8:2 FTUCA-GSH conjugate, 8:2 FTOH sulfate and glucuronide conjugates) was performed by liquid chromatography with negative electrospray-tandem mass spectrometry under conditions described previously [44]. Analytes were detected using an API 4000 Q Trap (Applied Biosystems/MDS Sciex, Concord, ON, Canada) with samples injected using an Agilent 1100 autosampler (injection volume: 10 µl, flow rate: 300 µl/min). Chromatography was performed using an ACE C18 column (Advanced Chromatography Technologies, Aberdeen, UK; length, 50 mm; inner diameter, 2.1 mm; particle size, 3 µm), preceded by a C18 guard cartridge (length, 4.0 mm, inner diameter, 2.0 mm; Phenomenex, Torrance, CA, USA). The column oven temperature was set to 30oC.

Analytes were identified based on the MS>MS transitions and retention times from authentic and synthesized standards. The 8:2 FTUCA-GSH and 7:3 FTCA-taurine conjugate were monitored using the expected MS>MS transition since no standard was available. Analyte responses were normalized to responses of either stable isotope-labeled internal standard or 13 analogues of similar structure. C4-PFOA was used for PFHxA, PFHpA, PFOA and the 8:2 13 13 13 FTOH glucuronide, C5-PFNA for PFNA, C2-8:2 FTCA for 8:2 FTCA and 7:3 FTCA, C2- 8:2 FTUCA for 8:2 FTUCA and 7:3 FTUCA, 10:2 FTOH sulfate for 8:2 FTOH sulfate. The 7:3 FTUCA was quantified using the response factor for the 8:2 FTUCA since no commercial standard was available.

245 8.3.2.5. Statistical Analysis and Data Treatment The instrumental detection limits were determined as the concentration that produced a peak with a signal to noise ratio of at least three. Method detection limits (MDLs) were determined as threefold the standard deviation of the results for control fish samples analyzed (n=19). If analytes were not detected in the blanks, half the instrumental detection limit was used as the MDL. Concentrations less than the instrumental detection limit were reported as non-detect. Concentrations were blank corrected using the mean level in the control fish. After blank correcting, concentrations less than the MDL were reported as less than the MDL value.

Fish growth rates were determined by fitting measured body weights to the model, ln(mass) = a + b*t, where b is the growth rate (fish mass in grams/time in hours), t is the time in hours, and a is a constant [45]. All tissue concentrations were corrected for growth dilution by multiplying the tissue concentrations by (1 + b*t), where t is the time relative to t=0.

Elimination rates (kd) were calculated using growth-corrected tissues levels and fitting to the first-order elimination equation, ln(Cfish) = a + kd * t, where Cfish is the growth-corrected blood -1 or liver concentration, kd is the first-order elimination rate constant (hour ), t is time (hours), and a is a constant. Elimination half-lives (t1/2) were calculated as ln 2/kd. Statistical differences (α=0.05) between elimination half-lives of the 8:2 FTCA and 8:2 FTUCA was examined by testing the Student’s t-test test using SPSS ® for Windows (2001; SPSS, Inc., Chicago, IL, USA).

The liver somatic index (LSI) was calculated as,

(g) massliver massliver (g) (%) LSI (%) = x 100% (g) massfish whole massfish (g)

The formation efficiency (FE %) was defined as the PFOA or PFNA concentration in the tissues at the end of the uptake phase divided by the 8:2 FTAc concentration in the food, on a molar basis,

or PFOA or PFNA tissue concentrat (nmol/g)ion (%) FE (%) = x 100% concentrat food FTAc 2:8 FTAc food concentrat (nmol/g)ion

246 8.3.3. In vitro S9 incubation experiments 8.3.3.1. S9 fraction preparation Hepatic and stomach cytosol fractions (herein called S9) were prepared from juvenile rainbow trout previously used as control animals in our in vivo experiments. Fish were anaesthetized with 4 g/L tricaine methanesulfonate (MS-222, buffered to pH 7), liver and stomach were removed within 15 minutes and minced using a scalpel. The excised stomachs and livers were rinsed with 18MΩ water to remove any residual blood and remaining food particles (stomach). Livers (n=10) and stomachs (n=3) were homogenized in cold buffer (4oC, 0.1 M phosphate buffer at pH 7.4, 1 mM EDTA, 0.15 mM KCl) at a ratio of 1 g of liver per 4 ml of buffer using a mechanical homogenizer. The homogenate was centrifuged at 9,000 g for 20 minutes, maintained at 4oC. The supernatant (S9 fraction) was decanted, the fractions were combined, separated into 1 ml cryogenic vials and frozen at -80oC until use. For the in vitro experiments, the S9 fraction was diluted to the appropriate concentration using an incubation solution consisting of 0.1 M phosphate buffer (pH 7.4) and 1 mM EDTA. It has been reported that esterases do not require any additional cofactors [46]. Protein concentration was determined using the Bradford assay [47] using bovine serum as the standard.

8.3.3.2. Para-nitrophenyl Acetate (PNPA) Assay Prior to experiments with the 8:2 FTAc, general carboxylesterase activity was determined by measuring the hydrolysis of para-nitrophenyl acetate (PNPA), following the methods of Heymann and Mentlein [48] with slight modifications. The formation of para- nitrophenol (PNP) from PFPA was quantified by monitoring the reaction at an absorbance of 405 nm. Experiments were run for 2 minutes with an absorbance recorded every 5 seconds. The initial PNPA concentration varied between 5 and 450 µM and S9 protein content was 0.5 mg/mL and 0.28 mg/ml for liver and stomach, respectively. Three replicates per PNPA concentration were performed. In addition, the spontaneous hydrolysis of PNPA was monitored by measuring PNPA degradation in buffer only. Experiments were performed at room temperature (22 oC).

8.3.3.3. 8:2 FTAc Incubations Enzyme kinetic parameters were obtained by varying the 8:2 FTAc concentration from 15-1006 nM for the liver S9 experiments and 40-6370 nM for the stomach S9 experiments. The protein concentration was held at 0.5 mg/ml for both experiments. Experiments were performed

247 in 2 ml microcentrifuge tubes, and the total volume of incubation buffer/S9 fraction/8:2 FTAc was 1 ml and 0.5 ml for the liver and stomach experiments, respectively. Stock solutions of the 8:2 FTAc were prepared in acetonitrile such that the final percentage of acetonitrile in the incubation solutions was 1-2%. Spiking at lower fractions prevented the 8:2 FTAc from becoming fully dissolved into the bulk phase (data not shown), with 8:2 FTAc presumably sitting at the air-water interface. The 8:2 FTAc was combined with the incubation buffer, left to equilibrate for 30 s, and the experiment was initiated by addition of the S9 fraction. Three vials per time point were sacrificed every 0.5-1.0 min and the experiments ran a total of 2-4 min. Control experiments (n=1 per time point), using heat inactivated S9 fractions, were performed to account for background hydrolysis. To maintain consistency with the in vivo 8:2 FTAc study, experimental vials were immersed in a water bath maintained at 18oC. Further, the incubation buffer solution was pre-equilibrated at 18oC. Experiments were stopped by adding 1 ml of ethyl acetate, and analytes were extracted by vortexing for 2 min. Ethyl acetate extracts were analyzed for loss of the parent compound as well as formation of the 8:2 FTOH by GC-MS.

8.3.3.4. Data Analysis The 8:2 FTAc incubation studies, spike and recovery tests (n=3) using heat-inactivated S9 fractions showed that the mean recovery (RSD) was 112% (15%) and 115 (7%) for the 8:2 FTAc and 8:2 FTOH, respectively. Therefore, concentrations were not corrected for recovery.

The initial rate (vo) of 8:2 FTOH or PNP formation was calculated for each test concentration using only the linear portion of the [S] versus time plot. Line of best fit and enzyme kinetic parameters were obtained by nonlinear regression analysis using the Enzyme Kinetics module (v. 1.3) for SigmaPlot (v. 11.0, Systat Software Inc., Chicago, IL).

248 8.4. Results and Discussion

8.4.1. In Vivo 8:2 FTAc Dietary Exposure 8.4.1.1. Physical Indices No mortalities were observed during the experiment in either the dosed or control treatments. The growth rates were significant in both the dosed (p<0.001) and control (p<0.001) treatments (Figure S8.1 in supporting information). The overall growth rates were 0.64 and 1.9 g/day in the dosed and control fish, respectively. No significant growth was observed during the elimination phase (dose: p=0.79, control: p=0.26) when the feeding rate was reduced to 1.5% of the fish body weight per day.

The liver somatic index (LSI) is a measure of the liver mass relative to the whole body and is generally used as a measure of metabolic stress resulting from contaminant exposure. The LSI was shown to increase significantly from the beginning of the experiment to the end of the uptake phase in both the dosed and control treatments (Figure S8.2 in supporting information). The LSI trend during the uptake phase of the dose treatment was best described using a logarithmic model (dose: LSI=0.12*Ln(time, hours) + 1.0, r2=0.43), whereas, control treatments was best described using a linear model (LSI=0.005 * time (hours) + 0.50, r2=0.72). The coincident trends indicate that the 8:2 FTAc exposure is not responsible for the observed LSI increase. In fact, when considering all time points, the LSI in the dose and control treatments were high correlated (p<0.0001, r2=0.55). Presumably, the LSI increase is the result of the twice per day feeding schedule, or the canola oil that was amended to the food during preparation. During the elimination phase, when the feeding rate was reduced and the fish were fed unaltered commercial food (i.e. no addition of canola oil, gelatin or 8:2 FTAc), the LSI decreased and approached pre-dose levels by the end of the experiment. Therefore, under the dosing levels used in the present study, exposure to the 8:2 FTAc, and the resulting metabolites, did not result in metabolic stress to the fish.

8.4.1.2. 8:2 FTAc and 8:2 FTOH Overall, the parent 8:2 FTAc was accumulated in very low concentrations by the internal tissues of the dosed fish. The 8:2 FTAc was not detected in the tissues of the control fish. Quantifiable levels of the 8:2 FTAc were only measured in the kidney and liver, and at very low levels (mean concentrations are presented in the supporting information). Prior to 24hrs after

249 the initial dosing, 8:2 FTAc liver concentrations were either non-detect or below the MDL. The 8:2 FTAc liver concentrations were relatively stable throughout the remainder of the uptake phase, with mean concentrations ranging from 1.8 to 6.0 ng/g ww. In the kidney, the 8:2 FTAc was periodically detected during the uptake phase (26 out of 42 individuals) and showed very low but steady concentrations, ranging from 1.0 to 26 ng/g ww. Interestingly, the liver 8:2 FTAc concentrations were relatively steady during the first 36 hrs of the elimination phase, which may be the result of enterohepatic circulation from the intestine. For the remainder of the elimination phase, liver 8:2 FTAc concentrations were predominately

As discussed later in the chapter, we have confirmed, using in vivo S9 hepatic and stomach fractions, that the 8:2 FTOH is the first product of 8:2 FTAc biotransformation. Liver 8:2 FTOH levels were low and relatively stable through the uptake. Mean concentrations ranged from 15 to 290 ng/g ww with the overall mean concentration of 85 ng/g ww during the uptake phase. The 8:2 FTOH levels in the kidney were above the MDL (16 ng/g ww) beginning at 4 hours after the initial dosing. Concentrations were low and steady throughout the uptake phase with mean concentrations generally ranging from 20 to 60 ng/g ww. The 8:2 FTOH was quickly eliminated in both the liver and kidney, presumably through a combination of depuration and metabolism. The 8:2 FTOH levels were consistently below the MDL within 24 hr (liver) and 48 hrs (kidney) after starting the elimination phase.

The 8:2 FTAc and 8:2 FTOH were consistently detected in the feces throughout the uptake phase (Figure 8.1). The 8:2 FTOH feces levels were relatively high but steady during uptake with a mean concentration of 1240 ng/g ww (range: 380-6180 ng/g ww), which represented ~1.5% of the parent 8:2 FTAc food dose, on a mole basis. In comparison, the 8:2 FTAc feces levels were also steady but much lower with a mean concentration of 190 ng/g ww (range: 16-990 ng/g ww), representing only ~0.2% of the food concentration.

The very low concentrations of 8:2 FTAc measured in the fish tissues is suggestive of a high biotransformation rate within either the gut or internal tissues. Previous research has shown that fish can rapidly biotransform ester compounds, resulting in an overall decrease in

250 bioaccumulation [46, 49]. Also, it has been shown that the carboxylesterase activity in rainbow trout is comparable to glutathione-s-transferase acitivity, and much higher than several mixed- function oxygenases and glucuronidation activity [50]. The low tissue concentrations were not the result of poor uptake efficiency as evidence by the very low quantities of 8:2 FTAc detected in the feces, which represents primarily digested food. In a later section in this chapter, we showed quantitatively similar carboxylesterase activity towards the 8:2 FTAc in the stomach and liver rainbow trout S9 fractions. As well, although data are very limited, other studies have shown comparable esterase activity between the liver, stomach and intestine in fish [51-53]. Further, Barron et al. [50] showed, using 4-nitrophenyl acetate as the substrate, that rainbow trout had somewhat higher carboxylesterase activity in the sera as compared to the liver. These results lend support to the hypothesis of rapid biotransformation of the 8:2 FTAc, possibly occurring in the gut prior to uptake into the internal tissues. Presumably, the 8:2 FTOH formed by the 8:2 FTAc metabolism would subsequently be taken up into the internal tissues where further biotransformation would occur in the liver.

In addition to enzyme mediated biotransformation, the 8:2 FTAc may have been degraded abiotically in the stomach via acid catalyzed hydrolysis. Again, the expected first degradation product would be the 8:2 FTOH. However, the relevance of this degradation reaction has not been investigated.

The 8:2 FTOH was detected at low, but steady concentrations throughout the uptake phase in the kidney and liver. Similar to the 8:2 FTAc, these results suggest rapid biotransformation but may also represent rapid depuration. As mentioned earlier, the 8:2 FTOH was not detected in the bile suggesting elimination via the bile is not significant. Brandsma et al. [54] investigated the biotransformation of the 8:2 and 10:2 FTOH in rainbow trout. The parent FTOH compounds were not monitored, however, the corresponding FTCAs & FTUCAs were detected within 1 hr of initial dosing, suggesting rapid biotransformation. Further, whole animal 8:2 FTOH exposure studies with mammals show rapid biotransformation with peak concentrations occurring within several hours of dosing [24, 25]. For example, Fasano et al. [25], in a rat model, reported peak 8:2 FTOH concentrations in plasma 1 hour after oral dosing, with an observed half-life of 5 hours. Further, Henderson & Smith [27] reported the 8:2 FTOH was not detected in the plasma and liver of pregnant mice 24 hours after oral dosing. However, biotransformation of the 8:2 FTOH in fish may be slower. Nabb et al. [28] reported that in vitro

251 8:2 FTOH half-lives in trout hepatocytes and microsomes were about 9 times slower than in rodents.

10000 8:2 FTOH 8:2 FTAc 1000

100

10 Concentration (ng/g ww) Concentration (ng/g

Uptake Elimination 1 0 50 100 150 200 250 300 350

Time (hours)

Figure 8.1. Mean 8:2 FTOH and 8:2 FTAc concentrations (ng/g ww) in feces. Error bars indicate 1 standard error.

8.4.1.3. Intermediate and Terminal Metabolites Concentrations in the “predose” fish, collected from each tank 1 hr prior to dosing, were similar to those in the control fish. These results indicate that the trends observed were the result of 8:2 FTAc biotransformation and not due to background contamination.

8.4.1.3.1. Uptake Phase – Liver, kidney and blood Overall, the 8:2 FTCA, 8:2 FTUCA, 7:3 FTCA and PFOA were rapidly formed within the fish and were detected in the liver, kidney and blood within 1-4 hours of initial dosing. Uptake and elimination trends in the liver are shown in Figure 8.2, whereas kidney and blood trends are shown in the supporting information. In addition, tables showing the mean tissue concentrations for each time point are presented in the supporting information.

252 The PFHpA was not measured above MDL levels in liver or kidney, but was quantifiable in blood at very low levels. It has been shown that the PFHpA is formed from 7:3 FTCA biotransformation, but at very low yields [55]. The 7:3 FTUCA was also occasionally detected in the liver and kidney, but all measured concentrations were below MDL with the exception of blood which showed levels of ~1 ng/g ww. The 7:3 FTCA-taurine, 8:2 FTUCA glutathione and 8:2 FTOH sulfate conjugates were not detected in any of the tissue samples. However, authentic standards were not available for the 7:3 taurine and 8:2 FTUCA glutathione conjugates.

The 8:2 FTCA was the metabolite formed and accumulated in the highest concentrations. Similarly, in rainbow trout hepatocyte incubations with the 8:2 FTOH, Nabb et al. [28] also showed that the 8:2 FTCA was formed in the greatest proportion followed by much lower levels of the 8:2 FTUCA and 7:3 FTCA. The 8:2 FTCA levels were similar among the tissues and showed consistent time trends. Specifically, tissue concentrations were above the MDL within 1 hour or 2 hours (blood) of initial dosing and reached steady-state concentrations of ~1000-1400 ng/g ww within 24 hours.

The 8:2 FTUCA was also formed quickly and showed tissue levels above the MDL within 1-4 hours after initial dosing. The 8:2 FTUCA also showed rapid increases in tissue concentrations and steady-state concentrations were reached by 24 hours. However, there was some variability in the tissue levels. Steady-state concentrations were ~400 ng/g ww, ~65 ng/g ww and ~40 ng/g ww in the blood, kidney and liver respectively. The approximately 10-fold difference in 8:2 FTUCA concentrations between the blood and liver is consistent with a rainbow trout in vivo 8:2 FTUCA dosing experiment [55] and may be representative of enhanced biotransformation of this compound in the liver.

The 7:3 FTCA was also formed and accumulated in significant quantities, although the tissue levels were ~10-fold lower than the 8:2 FTCA. As with the 8:2 FTCA, 7:3 FTCA concentrations were similar between the tissues and showed consistent time trends. The 7:3 FTCA also showed rapid increase through the first 12 hours of dosing, but in contrast did not appear to reach steady state and tissue levels increased slowly until the end of the uptake phase, reaching levels of ~100 ng/g ww. The unsaturated analogue, the 7:3 FTUCA, was also detected

253 in the tissues based on the expected MS>MS transition. Tissues concentrations were very low and were ~0.01 ng/g ww in kidney and liver, and ~1 ng/g ww in the blood.

PFOA was also formed from the 8:2 FTAc biotransformation and tissue levels were above the MDL within 4-6 hours of initial dosing. PFOA concentrations were generally comparable between the tissues, within ~2-fold. Similar to the 7:3 FTCA, PFOA levels showed a rapid increase until ~24 hours, followed by a much slower but constant increase throughout the uptake phase. Mean PFOA concentrations at the end of the uptake phase were 123 ng/g ww, 74 ng/g ww and 51 ng/g ww in the blood, kidney and liver, respectively. In addition, PFNA was detected at very low concentrations in the liver and kidney. Mean PFNA levels ranged from 0.6-1.9 ng/g ww in the kidney and 0.3-1.5 ng/g ww in the liver, respectively. All PFNA blood concentrations during the uptake phase were below the MDL, however, this tissue had an unusually high MDL of 16 ng/g ww. The lower PFNA levels as compared to PFOA are consistent with several studies that have investigated 8:2 FTOH exposure in mammals [24, 26- 28].

8.4.1.3.2. Elimination Phase Among the intermediate and terminal metabolites, elimination trends were consistent between the liver, kidney and blood (Table 8.1). The 8:2 FTCA and 8:2 FTUCA were rapidly eliminated from the tissues and half-lives ranged from 1.1-1.2 days and 1.0-1.5 days for the 8:2 FTCA and 8:2 FTUCA, respectively. Elimination half-lives for the 8:2 FTCA and 8:2 FTUCA were statistically similar in the kidney and blood, and the 8:2 FTUCA half-life was slightly longer than the 8:2 FTCA in the liver. These results suggest generally similar elimination kinetics for the 8:2 FTCA and 8:2 FTUCA. Within each tissue the 8:2 FTCA and 8:2 FTUCA levels were highly correlated during the elimination phase (r2: 0.80-0.96), further suggesting similar elimination kinetics. In addition, the calculated 8:2 FTCA and 8:2 FTUCA elimination half-lives were consistent with those recently reported for a 8:2 FTCA dietary exposure study in rainbow trout [55]. In that study, calculated elimination half-lives in blood were 1.2 d (95% confidence interval: 1.1-1.3 d) and 1.3 d (1.1-1.5 d) for the 8:2 FTCA and 8:2 FTUCA, respectively, and in liver were 1.3 d (1.1-1.4 d) and 1.8 d (1.0-8.6 d) for the 8:2 FTCA and 8:2 FTUCA, respectively [55]. It should be noted that the two studies consisted of identical experimental conditions. However, the 8:2 FTUCA elimination half-lives in the present study were significantly longer than those obtained when the 8:2 FTUCA is used as the parent

254 compound [55]. In the 8:2 FTUCA dietary exposure study [55], the blood elimination half-life was 0.39 d (0.31-0.53 d) while the liver half-life could not be calculated due the very low concentrations. In rainbow trout, it has been shown that the 8:2 FTUCA is formed from the biotransformation of the 8:2 FTCA [55]. Therefore, the results from the present study, combined with those of the individual 8:2 FTCA and 8:2 FTUCA dosing experiments, suggest that the 8:2 FTUCA tissue concentrations are determined primarily by the 8:2 FTCA trends in the body.

In contrast to the 8:2 FTCA and 8:2 FTUCA trends, tissue levels of the 7:3 FTCA and PFOA were steady during the initial 72-96 hours of elimination. As such, the elimination curves for the 7:3 FTCA and PFOA were not statistically significant from zero over the initial period of the elimination phase. These results are suggestive of continued production of these metabolites from precursors still present in the body, despite the fact the fish were receiving clean food. In rainbow trout, it has been shown that the 7:3 FTCA and PFOA are produced from the biotransformation of 8:2 FTCA, via the 8:2 FTUCA [55]. Therefore, it is possible that the 7:3 FTCA and PFOA were continuing to be formed and accumulated from the residual levels of 8:2 FTUCA, and ultimately the 8:2 FTCA, present in the fish. Presumably, these metabolites only began showing decreasing levels once concentrations of their precursors had decreased below a threshold level.

Despite the initial delay in elimination, the 7:3 FTCA and PFOA elimination curves were statistically significant (p<0.01) during the final 96-120 hours. The calculated elimination half-lives ranges were 3.1-3.3 d for 7:3 FTCA and 1.5-2.0 d for PFOA. Presumably, the large uncertainty in the elimination half-lives was due to the limited number of data points used in their calculation. Surprisingly, the 7:3 FTCA elimination half-lives were much shorter than those reported from a 7:3 FTCA dietary dosing experiment with rainbow trout [55]. In the 7:3 FTCA dosing experiment, elimination half-lives were 5.1 d (3.1-14 d) and 10.3 d (6.4-26.0 d) for blood and liver, respectively. Similarly, the PFOA half-lives in the present study were somewhat shorter than, but within experimental error of the previously reported half-lives in rainbow trout of 3.7 to 5.2 days [37, 56, 57]. Collectively, these results suggest much different pharmacokinetics for compounds produced via biotransformation versus those directly dosed via the diet. Finally, given that the PFOA and 7:3 FTCA levels did not appear to decline until the

255 later stages of the experiment, a longer elimination period would have been warranted and may have reduced the uncertainty in elimination half-lives.

10000 100

8:2 FTCA 8:2 FTUCA

1000 10

100

1 10

Uptake Elimination Uptake Elimination 1 0.1 0 75 150 225 300 0 75 150 225 300

1000 1000 7:3 FTCA PFOA 100 PFNA 100

Concentration (ng/g ww) (ng/g Concentration 10 10 1

1 0.1 Uptake Elimination Uptake Elimination 0.1 0.01 0 75 150 225 300 0 75 150 225 300

Time (hours)

Figure 8.2. Liver concentrations (ng/g ww) of 8:2 FTCA, 8:2 FTUCA, 7:3 FTCA, PFOA and PFNA from 8:2 FTAc dietary exposure in rainbow trout. 256

Table 8.1. First-order elimination half-lives (d) for liver, kidney and blood from 8:2 FTAc dietary exposure.

Liver Kidney Blood 2 2 2 t1/2 (d) (95% CI) r t1/2 (d) (95% CI) r t1/2 (d) (95% CI) r 8:2 FTCA 1.2 (1.1-1.4) 0.92 1.1 (1.0-1.2) 0.93 1.1 (0.9-1.2) 0.92 8:2 FTUCA 1.5 (1.3-1.8) 0.86 1.0 (0.8-1.2) 0.83 1.0 (0.8-1.3) a 0.81 7:3 FTCA (0-96hr depuration) n.s. (p=0.92) b <0.01 n.s. (p=0.40) b 0.03 n.s (p=0.07) 0.16 (96-192hr depuration) 3.2 (2.0-8.1) c 0.58 3.1 (2.0-6.4) c 0.63 3.3 (1.8-24.1) 0.52 PFOA (0-96hr depuration) n.s. (p=0.98) b <0.01 n.s (p=0.65) 0.02 n.s (p=0.68) 0.01 c (96-192hr depuration) 2.0 (1.2-5.1) 0.58 1.5 (0.8-9.3) 0.52 1.7 (0.9-9.6) 0.54 a elimination half-lives calculated from 0-96 hr elimination since blood concentrations at 144 hr and 192 hr were below the MDL. b elimination half-lives calculated from 0-72 hr elimination. c elimination half-lives calculated from 72-192 hr elimination.

257

258 8.4.1.3.3. Bile and Feces The intermediate fluorotelomer carboxylates and terminal PFCAs were detected in the bile at concentrations that were generally within 2-fold of liver levels (Figure shown in the supporting information). In addition, the time trends for these metabolites were similar between the bile and liver. These trends are consistent with the liver being the site of bile formation. Steady- state concentrations in the bile were achieved for the 8:2 FTCA and 8:2 FTUCA and were ~1800 ng/g ww and ~20 ng/g ww, respectively. 7:3 FTCA concentrations increased throughout the uptake phase and were ~100 ng/g at the conclusion of the uptake phase. PFOA levels reached steady-state by 60 hrs and were ~85 ng/g ww. Interestingly, PFHpA levels in the bile were comparable to PFOA, whereas, PFHpA levels in the other tissues were either very low or below MDL. Bile PFHpA concentrations increased throughout the uptake phase reaching ~100 ng/g ww. Finally, relatively high levels of the 8:2 FTOH-glucuronide conjugate were measured in the bile. The bile 8:2 FTOH-glucuronide levels increased at a slightly slower rate than the other metabolites, reaching steady-state concentrations of ~250 ng/g ww after 48 hours.

The detection of the 8:2 FTOH glucuronide in the bile and feces was unique since this metabolite was not detected in the blood, liver or kidney. Glucuronide conjugates, including the 8:2 FTOH glucuronide [25], have been shown to be preferentially eliminated via the bile [58]. The relatively high concentrations, particularly in the feces, demonstrates an important elimination pathway for 8:2 FTOH metabolites. The 8:2 FTOH sulfate conjugate was not detected which is consistent with results from 8:2 FTOH incubations in rainbow trout hepatocytes by Nabb et al. [28]. However, it has been shown that sulfate conjugates resulting from pentachlorophenol metabolism are preferentially eliminated through renal and branchial pathways [58]. It is not known if these trends are applicable to 8:2 FTOH metabolism.

The intermediate and terminal metabolites were also measured in the feces. Since the fish were fasted 24 hrs prior to commencing the experiment, feces were not produced until 4 hr post initial dose. The rank order of the metabolite concentrations was consistent between the feces and bile, although feces levels were generally higher than bile. Specifically, the 8:2 FTCA and 8:2 FTUCA levels were at steady-state by 24 hrs at ~3400 ng/g ww and ~25 ng/g ww, respectively. The 7:3 FTCA, PFOA and PFHpA levels continued to increase throughout the uptake phase, reaching concentrations of ~400 ng/g ww, ~180 ng/g ww and 15 ng/g ww, respectively.

259 Interestingly, the 8:2 FTOH glucuronide steady-state feces concentration was ~2500 ng/g ww, approximately 10-fold higher than measured in the bile.

The high concentrations of intermediate and terminal metabolites measured in the feces indicate the potential for significant elimination via this excretion pathway. Alternatively, once in the intestine these metabolites may also undergo enterohepatic circulation and be reabsorbed, effectively re-entering the internal tissues. The relatively comparable levels (generally within 2 to 4-fold) in the bile suggest that excretion from the liver, via the bile, could be a source of these feces concentrations. In addition, some of the poly- and perfluorinated metabolites in the feces may originate from metabolism within the gut. As discussed in a later section in this chapter, we have shown high levels of carboxylesterase activity in the stomach which could explain the relatively high levels (~1200 ng/g ww) of 8:2 FTOH measured in the feces. Several studies have shown the potential for monooxygenase activity and glucuronidation in the intestine, although activity levels are generally lower than those of the liver [59, 60]. Specifically, in rainbow trout it was shown that 7-ethoxycoumarin-O-dethylase (EROD) activity in the intestine was ~10-fold lower than in liver, whereas glucuronidation activity was ~60% lower in the intestine [61].

8.4.2. In Vitro S9 Incubation Experiments 8.4.2.1. Para-nitrophenyl Acetate (PNPA) Assay The carboxylesterase activity of the hepatic and stomach S9 fractions, as determined using PNPA as the surrogate, showed typical Michaelis-Menten kinetics (Figure 8.3). The enzyme kinetic parameters for the hepatic fraction, obtained from nonlinear regression analysis were 80 ± 7.0 µM (standard error) for the Michaelis constant (KM) and 614 ± 18 nmol/min.mg protein for the maximum reaction rate (Vmax). These results are consistent with those previously reported for liver S9 and microsomes from rainbow trout [50], Chinook salmon (Oncorhynchus tshawytscha) [62], white grunt (Haemulon plumieri) [63], rabbitfish (siganus canaliculatus) and seabream (Acanthopagrus latus) [64] (Table 8.2). However, apparent KM values for the rabbitfish and seabream liver fractions were ~10- to 100-fold greater than in the present study.

The enzyme kinetic parameters for the stomach S9 fraction were 273 ± 61 µM for KM and 147 ± 16 nmol/min.mg protein for the maximum reaction rate (Vmax). These results show that the carboxylesterase activity in the stomach was only ~4-fold than that in the liver.

260

600 Liver 500

400

V = 614 +/- 18 nmol/min.mg 300 max KM = 80 +/- 7.0 μM

200

Initial Rate (nmol/min.mg) Rate Initial 100

0 0 100 200 300 400 500

120

100 Stomach

80

60

40 Vmax = 147 +/- 16 nmol/min.mg

Initial Rate (nmol/min.mg)Initial Rate K = 273 +/- 61 μM 20 M

0 0 100 200 300 400 500 [PNPA] (µM)

Figure 8.3. Michaelis-Mention plot of para-nitrophenyl acetate carboxylesterase activity

(nmol/min.mg protein) in rainbow trout liver and stomach S9 fractions. Michaelis constant (KM) and maximum reaction rate (Vmax) obtained from nonlinear regression analysis. Error bars represent one standard error.

261

Table 8.2. Michaelis constant (KM) and maximum reaction rate (Vmax) parameters in fish using PNPA as the substrate for carboxylesterase activity.

Vmax K (µm) Reference M (nmol/min.mg) Rainbow Trout Liver, S9 80 614 this study Rainbow Trout Liver, microsomes 28 672 Barron et al., 1999 Rainbow Trout Serum, microsomes 214 1,639 Barron et al., 1999 Chinnok Salmon Liver 154 86 Wheelock et al., 2005 (Oncorhynchus tshawytscha ), S9 White Grunt Liver (Haemulon not reported 191 Leticia & Gerardo, 2008 plumieri ), S9 Rabbitfish Liver (Siganus 320 5,220 Al-Ghais et al., 2000 canaliculatus ), S9 Rabbitfish Liver (Siganus 1,180 2,830 Al-Ghais et al., 2000 canaliculatus ), microsomes Seabream Liver (Acanthopagrus 3,900 670 Al-Ghais et al., 2000 latus ), S9 Seabream Liver (Acanthopagrus 2,430 480 Al-Ghais et al., 2000 latus ), microsomes

8.4.2.2. 8:2 FTAc Incubations The 8:2 FTAc was rapidly biotransformed by the rainbow trout liver and stomach S9, with formation of the 8:2 FTOH observed within 30 s-1 min of incubation (complete data shown in the supporting information). The coincident formation of the 8:2 FTOH was shown and mass balance (8:2 FTOH + 8:2 FTAc) was preserved throughout the experiment, indicating that the 8:2 FTOH was the sole product formed. No degradation of the 8:2 FTOH was observed in either liver or stomach treatments, presumably due to the relatively slower biotransformation rate of this compound [28] and the fact that the incubation system was not optimized for oxidative reactions (i.e. no NADPH). Background hydrolysis in the heat-inactivated S9 fractions was negligible, as evidence by the insignificant slopes (p>0.05 for all incubation concentrations) of 8:2 FTAc versus time relationship. In addition, the formation of 8:2 FTOH was observed only in the higher concentration incubations in the heat-inactivated liver fractions and only at trace levels (<1% of initial 8:2 FTAc level). The 8:2 FTOH formation was not observed in the heat-inactivated stomach fractions. These results indicate that the observed 8:2 FTAc degradation was due to the carboxylesterase activity of the hepatic and stomach S9 fractions.

The carboxylesterase activity of the liver and stomach S9 fractions, as determined using 8:2 FTAc as the surrogate, showed typical Michaelis-Menten kinetics (Figure 8.4). The

262 maximum reaction rates (Vmax) were 387 ± 64 pmol/min.mg protein and 315 ± 15 pmol/min.mg protein for the liver and stomach, respectively. The Michaelis constants (KM) were 231 ± 124 nM (standard error) and 788 ± 124 nM for the liver and stomach, respectively.

The incubation experiments were performed at 18oC to maintain consistency with the in vivo study. As such, no attempt was made to investigate the optimal temperature for the carboxylesterase activity. However, Barron et al. [50], using PNPA as the substrate, found little temperature variation (2 to 40oC) of carboxylesterase activity in rainbow trout microsomes with a temperature optimum of approximately 22 oC.

The results of the present study show that the carboxylesterase activity towards the 8:2 FTAc is similar between the stomach and liver. This trend is consistent with those using PNPA as the substrate which showed carboxylesterase activities in the stomach only 4-fold lower than in the liver fractions. Although data are limited, other studies have shown comparable esterase activity between the liver, stomach and intestine in fish [51-53]. Further, Barron et al. [50] showed that esterase activity towards PNPA was significantly greater in serum microsomes as compared to liver microsomes (Table 8.2). Overall, these studies, along with the results from the present study, suggest the widespread distribution of carboxylesterase activity throughout the body with the capacity for ester degradation prior to uptake into the internal organs. These results have important implications with regards to the fate and disposition of the 8:2 FTAc, and other esters, in fish tissues. Specifically concerning the 8:2 FTAc, the results suggest that the parent compound may be, at least partially, biotransformed to the 8:2 FTOH in the stomach and intestine, followed by uptake of the 8:2 FTOH into the internal organs. Subsequent metabolism of the 8:2 FTOH to polyfluorinated intermediates (i.e. 8:2 FTCA, 8:2 FTUCA, 7:3 FTCA) and terminal metabolites (i.e. PFOA and PFNA) presumably would occur in the liver, the main site for oxidative biotransformation reactions.

Nichols et al. [65] recently reviewed methods to extrapolate in vitro liver biotransformation data to predict whole body bioaccumulation in fish. Until recently, this area of research has received little attention. The few examples include studies by Law et al. [66], Schultz and Hayton [67], Cowan-Ellsberry et al. [46] and Dyer et al. [68]. At present, studies extrapolating actual in vitro biotransformation rates to whole body bioaccumulation are limited

263 to the consideration of liver biotransformation only. Two models were developed by Nichols et al. [69, 70], however, these models used estimated, hypothetical gut biotransformation rates. Therefore, there is a need for additional models that consider extrahepatic metabolism.

0.40

0.35 Liver

0.30

0.25

0.20

0.15 Vmax = 387 +/- 64 pmol/min.mg K = 231 +/- 124 nM 0.10 M Initial Rate Rate (nmol/min.mg) Initial 0.05

0.00 0 200 400 600 800 1000 1200 0.35

0.30 Stomach

0.25

0.20

0.15 Vmax = 315 +/- 15 pmol/min.mg KM = 788 +/- 124 nM 0.10 Initial Rate (nmol/min.mg) 0.05

0.00 0 1000 2000 3000 4000 5000 6000 7000 [8:2 FTAc] (nM)

Figure 8.4. Michaelis-Menten plot of 8:2 FTAc biotransformation (nmol/min.mg protein) in rainbow trout liver and stomach S9 fraction. Michaelis constant (KM) and maximum reaction rate (Vmax) obtained from nonlinear regression analysis. Error bars represent one standard error.

264 8.4.3. Mechanism of Biotransformation. As shown by the S9 incubations, the 8:2 FTAc is initially metabolized to the 8:2 FTOH via carboxylesterase enzymes. This initial biotransformation step may occur either within the gut prior to uptake or within the internal tissues, with what appears as reasonably equal efficiency. Several studies have investigated the in vivo metabolism of 8:2 FTOH in mammals. In contrast, only one study has examined 8:2 FTOH metabolism in fish [28] and that study only investigated in vitro biotransformation using hepatocytes and hepatic cellular fractions. A more complete discussion of the hypothesized 8:2 FTOH biotransformation pathways is discussed in Chapter 9, however, a brief discussion will be presented here for completeness. For simplicity, the following discussion only considers 8:2 FTOH mechanistic studies in animals. The initial 8:2 FTOH biotransformation steps are consistent among most literature published mechanisms. Briefly, the 8:2 FTOH undergoes a series of oxidation reactions, mediated via P450 mediated monooxygenase reactions, to form the 8:2 FTAL and then the 8:2 FTCA. Alternatively, the 8:2 FTOH may form glucuronide and sulfate conjugates which would presumably be excreted. The 8:2 FTAL, which was not monitored in the present study, may be further oxidized to the unsaturated analogue, the 8:2 fluorotelomer unsaturated aldehyde (8:2 FTUAL) that can form a glutathione conjugate. It has been postulated that the 8:2 FTUAL-GSH conjugate is reduced to form the 8:2 unsaturated FTOH-GSH conjugate [24, 25, 28] with further degradation to the 8:2 unsaturated FTOH 3-thiol [71]. In addition, Nabb et al. [28] have proposed that further biotransformation of the 8:2 FTUAL will form the 7:3 β–hydroxy unsaturated aldehyde followed by the 7:3 β-keto aldehyde, ultimately yielding PFOA. An additional fate of the 8:2 FTUAL suggested by Nabb et al. [28] is formation of the 7:3 FTUAL, ultimately yielding the 7:3 FTCA and 7:3 FTCA-TA conjugate.

The 8:2 FTCA can undergo α-oxidation to form PFNA. This mechanism has been suggested by several authors [24, 25, 28] and has been confirmed in our rainbow trout in vivo dosing experiments [55]. In addition, the 8:2 FTCA will be further oxidized to yield the 8:2 FTUCA. Similar to the 8:2 FTUAL, the 8:2 FTUCA will form the 8:2 FTUCA-GSH conjugate as was observed in our in vivo dosing experiment [55]. The fate of the 8:2 FTUCA is the main source of disagreement among the various proposed biotransformation reactions [24, 25, 28, 71]. Martin et al. [24] postulated that the 8:2 FTUCA will yield PFOA via β-oxidation. This mechanism was also proposed Hagen et al. [20] (in vivo study of male rats) and Dinglasan et al. [21] (microbial system). In contrast, Nabb et al. proposed that the 8:2 FTUCA is metabolized

265 through two branches. One branch will ultimately yield the 7:2 sFTOH and 7:2 sFTOH glucuronide conjugate, whereas the second branch will yield the 7:3 FTUCA followed by the 7:3 FTCA and finally the 7:3 FTUCA-TA conjugate. Fasano et al. [25] also proposed two branching pathways for the 8:2 FTUCA biotransformation. One branch will similarly yield the 7:2 sFTOH, although different intermediates were proposed. The second branch also yields the 7:3 FTUCA and 7:3 FTCA, however, Fasano et al. [25] postulated that the 7:3 FTCA will form PFOA via β-oxidation. In addition, Fasano et al. [25] proposed the direct formation of PFOA from the 7:3 FTUCA. The recent pathway put forth by Fasano et al. [71] is a synthesis of the Nabb et al. [28] and Fasano et al. [25] mechanisms.

The present study made no attempt to elucidate the mechanism of 8:2 FTOH biotransformation. This issue is addressed in Chapter 9. These in vivo experiments [55], in which the 8:2 FTCA, 8:2 FTUCA and 7:3 FTCA were dosed separately, showed that PFOA was formed from the 8:2 FTCA and 8:2 FTUCA but not the 7:3 FTCA. These findings directly contradict those of Nabb et al. [28] who showed that PFOA was only formed from 8:2 FTAL biotransformation and not through the 8:2 FTCA.

8.4.4. Environmental Implications Generally, the role of chemical biotransformation is to form a compound that is more readily eliminated than the parent compound. This is achieved by altering a chemical, either through introducing polar functional groups (phase I) or conjugation (phase II), such that it becomes more hydrophilic and thus more readily eliminated. For example, it has been shown that the biotransformation of labile hydrophobic organic compounds results in lower bioaccumulation than predicted through traditional BAF/BCF-log KOW relationships [46, 72]. Similar results have also been shown with more sophisticated pharmacokinetic models [73]. Further, it has been shown that suppressing biotransformation enzymes, using chemical inhibitors, will result in enhanced bioaccumulation of the parent compound and lower accumulation of metabolites [74]. However, the current study showed that the biotransformation a polyfluorinated telomer-based compound resulted in the formation of metabolites with much longer biological half-lives than the parent compound (Figure 8.5). As a result, metabolite concentrations reached levels that were up to 270-fold higher than the parent compound. These trends are expected to be exaggerated with the longer-chain precursors which would form metabolites with longer half-lives [37, 57]. Of particular significance to

266 fluorotelomer-based compounds is that the intermediate degradation products (i.e. FTALs and FTCAs) show greater toxicity than the parent FTOH or terminal PFCAs [38, 39]. The relatively longer biological half-life for poly- and perfluorinated compounds may partially be the result of their unique protein binding and enterohepatic circulation capability. In addition, there are no known biological degradation mechanisms for the perfluorinated carboxylates. Therefore, the current study demonstrated a situation where biotransformation can modify a readily labile compound into a biologically persistent compound. There are other examples of xenobiotics in which the metabolites have with similar or greater biological persistency as compared to the parent compound (e.g. DDE, dieldrin and oxychlordane [75-77]). However, the present study demonstrated an extreme case in which the metabolite levels were over two orders of magnitude greater than the parent compound. Another such example is the bioconcentration of benz(a)acridine in the fathead minnow [78], in which the sum of metabolites obtained concentrations ~10-fold greater than the parent compound.

The overall formation and accumulation of PFOA and PFNA was very low, which is consistent with other studies that have examined 8:2 FTOH biotransformation [28]. The PFOA formation efficiency (FE) was 0.17%, 0.07% and 0.10% in the blood, liver and kidney. The PFNA FE values were much lower, reflective of the relatively lower tissue concentrations, and were 0.002% and 0.003% for liver and kidney, respectively. It should be noted that these values are underestimated since both PFOA and PFNA did not reach steady-state concentrations at the conclusion of the uptake phase.

In the current study, we investigated the bioaccumulation and biotransformation of the 8:2 FTAc, using rainbow trout as the model organism. Complimentary in vivo and in vitro experiments were performed. We used the 8:2 FTAc since FTAc monomers are the primary building block used in the production of fluorinated polymers, which are widely used in commerce. In addition, FTAcs are widely detected in the outdoor atmosphere with presumably higher indoor levels. As we have shown in our S9 incubation experiments, the 8:2 FTAc can readily undergo phase I metabolism, with apparently equal efficiency in the stomach and liver, which ultimately limits its bioaccumulation potential. Additional research should investigate methods to extrapolate the observed in vitro kinetic data to predict whole body bioaccumulation, as has been recently performed for some organic compounds [46], with emphasis on incorporating gut metabolism.

267

10000 8:2 FTAc Metabolites 1000

100

10

1 Concentration (pmol/g ww) Concentration

0.1 0 20 40 60 80 100 120 140

Time (hours)

Figure 8.5. Kidney concentration (pmol/g ww) of 8:2 FTAc and metabolites during uptake phase of dietary exposure to 8:2 FTAc. “Metabolites” represents sum of 8:2 FTOH, 8:2 FTCA, 8:2 FTUCA, 7:3 FTCA, 7:3 FTUCA, PFHpA, PFOA and PFNA. Error bars indicate 1 standard error.

8.5. Acknowledgements Norman White and staff at the Aquatic Facility in the Department of Cell & Systems Biology (University of Toronto) are thanked for fish care and husbandry. Cora Young (Mabury group) provided assistance with fish dissection. Clara Chan, Helen Sun and Alex Tevlin provided assistance with sample preparation. Robert Di Lorenzo assisted with the stomach S9 incubation experiments. We are grateful to Wellington Laboratories for donation of mass- labeled standards. Project funding was provided by the Natural Sciences & Engineering Research Council of Canada (NSERC) (Mabury) and Environment Canada’s Chemical Management Plan (Muir). C.M.B. also appreciates the support of NSERC through a Post- Graduate Scholarship.

268 8.6. Literature Cited

1. Hansen, K. J.; Clemen, L. A.; Ellefson, M. E.; Johnson, H. O., Compound-specific, quantitative characterization of organic fluorochemicals in biological matrices. Environ. Sci. Technol. 2001, 35, 766-770.

2. Olsen, G. W.; Huang, H.-Y.; Helzlsouer, K. J.; Hansen, K. J.; Butenhoff, J. L.; Mandel, J. H., Historical comparison of perfluorooctanesulfonate, perfluorooctanoate, and other fluorochemicals in human blood. Environ. Health Perspect. 2005, 113, 539-545.

3. Giesy, J. P.; Kannan, K., Distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 2001, 35, 1339-1342.

4. Martin, J. W.; Smithwick, M. M.; Braune, B. M.; Hoekstra, P. F.; Muir, D. C. G.; Mabury, S. A., Identification of long-chain perfluorinated acids in biota from the Canadian Arctic. Environ. Sci. Technol. 2004, 38, 373-380.

5. Houde, M.; Martin, J. W.; Letcher, R. J.; Solomon, K.; Muir, D. C. G., Biological monitoring of polyfluoroalkyl substances: A review. Environ. Sci. Technol. 2006, 40, 3463- 3473.

6. Yamashita, N.; Taniyasu, S.; Petrick, G.; Wei, S.; Gamo, T.; Lam, P. K. S.; Kannan, K., Perfluorinated acids as novel chemical tracers of global circulation of ocean waters. Chemosphere 2008, 70, 1247-1255.

7. Tao, L.; Kannan, K.; Kajiwara, N.; Costa, M. M.; Fillmann, G.; Takahashi, S.; Tanabe, S., Perfluorooctanesulfonate and related fluorochemicals in albatrosses, elephant seals, penguins, and polar skuas from the south ocean. Environ. Sci. Technol. 2006, 40, 7642-7648.

8. Dinglasan-Panlilio, M. J. A.; Mabury, S. A., Significant residual fluorinated alcohols present in various fluorinated materials. Environ. Sci. Technol. 2006, 40, 1447-1453.

9. Russell, M. H.; Berti, W. R.; Szostek, B.; Buck, R. C., Investigation of the biodegradation potential of a fluoroacrylate polymer product in aerobic soils. Environ. Sci. Technol. 2008, 42, 800-807.

10. Hurley, M. D.; Wallington, T. J.; Sulbaek Andersen, M. P.; Ellis, D. A.; Martin, J. W.; Mabury, S. A., Atmospheric chemistry of fluorinated alcohols: reaction with Cl atoms and OH radicals and atmospheric lifetimes. J. Phys. Chem. A 2004, 108, 1973-1979.

11. Ellis, D. A.; Martin, J. W.; De Silva, A. O.; Mabury, S. A.; Hurley, M. D.; Sulbaek Andersen, M. P.; Wallington, T. J., Degradation of fluorotelomer alcohols: A likely atmospheric source of perfluorinated carboxylic acids. Environ. Sci. Technol. 2004, 38, 3316-3321.

12. D'eon, J. C.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A., Atmospheric chemistry of N-methyl perfluorobutane sulfonamidoethanol, C4F9SO2N(CH3)CH2CH2OH: Kinetics and mechanism of reaction with OH. Environ. Sci. Technol. 2006, 40, 1862-1868.

13. Martin, J. W.; Ellis, D. A.; Mabury, S. A.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of perfluoroalkanesulfonamides: Kinetic and product studies of the OH radical and Cl

269 atom initiated oxidation of N-ethyl perfluorobutanesulfonamide. Environ. Sci. Technol. 2006, 40, 846-872.

14. Nakayama, T.; Takahashi, K.; Matsumi, Y.; Toft, A.; Sulbaek Andersen, M. P.; Nielsen, O. J.; Waterland, R. L.; Buck, R. C.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of CF3CH=CH2 and C4F9CH=CH2: Products of the gas-phase reactions with Cl atoms and OH radicals. J. Phys. Chem. A 2007, 111, 909-915.

15. Vésine, E.; Bossoutrot, V.; Mellouki, A.; Le Bras, G.; Wenger, J.; Sidebottom, H., Kinetic and mechanistic study of OH- and Cl-initiated oxidation of two unsaturated HFCs: C4F9CH=CH2 and C6F13CH=CH2. J. Phys. Chem. A 2000, 104, 8512-8520.

16. Young, C. J.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A., Atmospheric chemistry of 4:2 fluorotelomer iodide (n-C4F9CH2CH2I): Kinetics and products of photolysis and reaction with OH radicals and Cl atoms. J. Phys. Chem. A 2008, 112, 13542-13548.

17. Butt, C. M.; Young, C. J.; Mabury, S. A.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of 4:2 fluorotelomer acrylate [C4F9CH2CH2OC(O)CH=CH2]: Kinetics, mechanisms, and products of chlorine-atom- and OH-radical-initiated oxidation. J. Phys. Chem. A 2009, 113, 3155-3161.

18. Gauthier, S. A.; Mabury, S. A., Aqueous photolysis of 8:2 fluorotelomer alcohol. Environ. Toxicol. Chem. 2005, 24, 1837-1846.

19. Lei, Y. D.; Wania, F.; Mathers, D.; Mabury, S. A., Determination of vapor pressures, octanol-air, and water-air partition coefficients for polyfluorinated sulfonamide, sulfonamidoethanols, and telomer alcohols. J. Chem. Eng. Data 2004, 49, 1013-1022.

20. Hagen, D. F.; Belisle, J.; Johnson, J. D.; Venkateswarlu, P., Characterization of fluorinated metabolites by a gas chromatographic-helium microwave plasma detector - the biotransformation of 1H,1H,2H,2H-perfluorodecanol to perfluorooctanoate. Anal. Biochem. 1981, 118, 336-343.

21. Dinglasan, M. J. A.; Ye, Y.; Edwards, E. A.; Mabury, S. A., Fluorotelomer alcohol biodegradation yields poly- and perfluorinated acids. Environ. Sci. Technol. 2004, 38, 2857- 2864.

22. Wang, N.; Szostek, B.; Folsom, P. W.; Sulecki, L. M.; Capka, V.; Buck, R. C.; Berti, W. R.; Gannon, J. T., Aerobic biotransformation of 14C-labeled 8-2 telomer B alcohol by activated sludge from a domestic sewage treatment plant. Environ. Sci. Technol. 2005, 39, 531-538.

23. Wang, N.; Szostek, B.; Buck, R. C.; Folsom, P. W.; Sulecki, L. M.; Capka, V.; Berti, W. R.; Gannon, J. T., Fluorotelomer alcohol biodegradation - direct evidence that perfluorinated carbon chains breakdown. Environ. Sci. Technol. 2005, 39, 7516-7528.

24. Martin, J. W.; Mabury, S. A.; O'Brien, P. J., Metabolic products and pathways of fluorotelomer alcohols in isolated rat hepatocytes. Chem. Biol. Interact. 2005, 155, 165-180.

25. Fasano, W. J.; Carpenter, S. C.; Gannon, S. A.; Snow, T. A.; Stadler, J. C.; Kennedy, G. L.; Buck, R. C.; Korzeniowski, S. H.; Hinderliter, P. M.; Kemper, R. A., Absorption,

270 distribution, metabolism, and elimination of 8-2 fluorotelomer alcohol in the rat. Toxicol. Sci. 2006, 91, 341-355.

26. Kudo, N.; Iwase, Y.; Okayachi, H.; Yamakawa, Y.; Kawashima, Y., Induction of hepatic peroxisome proliferation by 8-2 telomer alcohol feeding in mice: Formation of perfluorooctanoic acid in the liver. Toxicol. Sci. 2005, 86, 231-238.

27. Henderson, W. M.; Smith, M. A., Perfluorooctanoic acid and perfluorononanoic acid in fetal and neonatal mice following in utero exposure to 8-2 fluorotelomer alcohol. Toxicol. Sci. 2007, 95, 452-461.

28. Nabb, D. L.; Szostek, B.; Himmelstein, M. W.; Mawn, M. P.; Gargas, M. L.; Sweeney, L. M.; Stadler, J. C.; Buck, R. C.; Fasano, W. J., In vitro metabolism of 8-2 fluorotelomer alcohol: Interspecies comparisons and metabolic pathway refinement. Toxicol. Sci. 2007, 100, 333-344.

29. D'eon, J. C.; Mabury, S. A., Production of perfluorinated carboxylic acids (PFCAs) from the biotransformation of polyfluoroalkyl phosphate surfactants (PAPs): Exploring routes of human contamination. Environ. Sci. Technol. 2007, 41, 4799-4805.

30. Prevedouros, K.; Cousins, I. T.; Buck, R. C.; Korzeniowski, S. H., Sources, fate and transport of perfluorocarboxylates. Environ. Sci. Technol. 2006, 40, 32-44.

31. Muir, D. C. G.; Howard, P. H.; Meylan, W. Developing analytical methodology for PB&T substance - a systematic process for identification of important chemicals. An update. Final report to US EPA Great Lakes National Program Office. Environment Canada, Water Science & Technology Directorate, WSTD Contribution No. 09-045, 25 pp. + appendix.

32. Dreyer, A.; Temme, C.; Sturm, R.; Ebinghaus, R., Optimized method avoiding solvent- induced response enhancement in the analysis of volatile and semi-volatile polyfluorinated alkylated compounds using gas chromatrography-mass spectrometry. J. Chromatogr. A 2008, 1178, 199-205.

33. Oono, S.; Matsubara, E.; Harada, K. H.; Takagi, S.; Hamada, S.; Asakawa, A.; Inoue, K.; Watanabe, I.; Koizumi, A., Survey of airborne polyfluorinated telomers in Keihan area, Japan. Bull. Environ. Contam. Toxicol. 2008, 80, 102-106.

34. Piekarz, A. M.; Primbs, T.; Field, J. A.; Barofsky, D. F.; Simonich, S., Semivolatile fluorinated organic compounds in Asian and western U.S. air masses. Environ. Sci. Technol. 2007, 41, 8248-8255.

35. Dreyer, A.; Weinberg, I.; Temme, C.; Ebinghaus, R., Polyfluorinated compounds in the atmosphere of the Atlantic and South oceans: Evidence for a global distribution. Environ. Sci. Technol. 2009, 43, 6507-6514.

36. Dreyer, A.; Matthias, V.; Temme, C.; Ebinghaus, R., Annual time series of air concentrations of polyfluorinated compounds. Environ. Sci. Technol. 2009, 43, 4029-4036.

271 37. Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G., Bioconcentration and tissue distribution of perfluorinated acids in rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 196-204.

38. Phillips, M. M. M.; Dinglasan-Panlilio, M. J. A.; Mabury, S. A.; Solomon, K. R.; Sibley, P. K., Fluorotelomer acids are more toxic than perfluorinated acids. Environ. Sci. Technol. 2007, 41, 7159-7163.

39. Martin, J. W.; Chan, K.; Mabury, S. A.; O'Brien, P. J., Bioactivation of fluorotelomer alcohols in isolated rat hepatocytes. Chem. Biol. Interact. 2009, 177, 196-203.

40. Weisbrod, A. V.; Sahi, J.; Segner, H.; James, M. O.; Nichols, J.; Schultz, I.; Erhardt, S.; Cowan-Ellsberry, C.; Bonnell, M.; Hoeger, B., The state of in vitro science for use in bioaccumulation assessments for fish. Environ. Toxicol. Chem. 2009, 28, 86-96.

41. Chaudier, Y.; Barthelemy, P.; Pucci, B., Synthesis and preliminary assessments of hybrid hydrocarbon-fluorocarbon anionic and non-ionic surfactants. Tetrahedron Lett. 2001, 42, 3583-3585.

42. Xu, L.; Krenitsky, D. M.; Seacat, A. M.; Butenhoff, J. L.; Anders, M. W., Biotransformation of N-ethyl-N(2-hydroxyethyl)perfluorooctanesulfonamide by rat liver microsomes, cytosol, and slices and by expressed rat and human cytochromes P450. Chem. Res. Toxicol. 2004, 17, 767-775.

43. Stapleton, H. M.; Brazil, B.; Holbrook, R. D.; Mitchelmore, C. L.; Benedict, R.; Konstantinov, A.; Potter, D., In vivo and in vitro debromination of decabromodiphenyl ether (BDE 209) by juvenile rainbow trout and common carp. Environ. Sci. Technol. 2006, 40, 4653- 4658.

44. Butt, C. M.; Muir, D. C. G.; Stirling, I.; Kwan, M.; Mabury, S. A., Rapid response of Arctic ringed seals to changes in perfluoroalkyl production. Environ. Sci. Technol. 2007, 41, 42- 49.

45. Fisk, A. T.; Norstrom, R. J.; Cymbalisty, C. D.; Muir, D. C. G., Dietary accumulation and depuration of hydrophobic organochlorines: Bioaccumulation parameters and their relationship with the octanol/water partition coefficient. Environ. Toxicol. Chem. 1998, 17, 951- 961.

46. Cowan-Ellsberry, C. E.; Dyer, S.; Erhardt, S.; Bernhard, M. J.; Roe, A. L.; Dowty, M. E.; Weisbrod, A. V., Approach for extrapolating in vitro metabolism data to refine bioconcentration factor estimates. Chemosphere 2008, 70, 1804-1817.

47. Bradford, M. M., A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of protein dye binding. Anal. Biochem. 1976, 72, 248-254.

48. Heymann, E.; Mentlein, R., Carboxylesterases - Amidases. In Methods in Enzymology, Jakoby, W. B., Ed. Academic: New York, NY, USA, 1981; Vol. 77, Detoxication and Drug Metabolism: Conjugation and Related Systems, pp 333-341.

272 49. Barron, M. G.; Schultz, I. R.; Hayton, W. L., Presystemic branchial metabolism limits di-2-ethylhexyl phthalate accumulation in fish. Toxicol. Appl. Pharmacol. 1989, 98, 49-57.

50. Barron, M. G.; Charron, K. A.; Stott, W. T.; Duvall, S. E., Tissue carboxylesterase activity of rainbow trout. Environ. Toxicol. Chem. 1999, 18, 2506-2511.

51. Li, S.-N.; Fan, D.-F., Activity of esterases from different tissues of freshwater fish and responses of their isoenzymes to inhibitors. J. Toxicol. Environ. Health 1997, 51, 149-157.

52. Hirji, K. N.; Courtney, W. A. M., Non-specific carboxylic esterase activity in the digestive tract of the perch, Perca fluviatilis L. J. Fish Biol. 1983, 22, 1-7.

53. Salamastrakis, S. S.; Haritos, A. A., Physicochemical characterization and tissue distribution of multiple molecular forms of fish (Trachurus trachurus) esterases. Comp. Biochem. Physiol. 1988, 91B, 741-750.

54. Brandsma, S. H.; Smithwick, M.; Solomon, K.; Small, J.; Muir, D. C. G., Uptake, transformation and elimination of 8:2 and 10:2 fluorotelomer alcohols and PFOSA by rainbow trout from feed. Manuscript in prep.

55. Butt, C. M.; Muir, D. C. G.; Mabury, S. A., Elucidating the mechanism of per- and polyfluorinated acid production in rainbow trout. Environ. Sci. Technol. in preparation.

56. De Silva, A. O.; Tseng, P. J.; Mabury, S. A., Toxicokinetics of perfluorocarboxylate isomers in rainbow trout. Environ. Toxicol. Chem. 2009, 28, 330-337.

57. Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G., Dietary accumulation of perfluorinated acids in juvenile rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 189-195.

58. Stehly, G. R.; Hayton, W. L., Metabolism of pentachlorophenol by fish. Xenobiotica 1989, 19, 75-81.

59. James, M. O.; Altman, A. H.; Morris, K.; Kleinow, K. M.; Tong, Z., Dietary modulation of phase 1 and phase 2 activities with benzo(a)pyrene and related compounds in the intestine but not the liver of the channel catfish, Ictalurus punctatus. Drug Metab. Dispos. 1997, 25, 346-354.

60. Kleinow, K. M.; James, M. O.; Tong, Z.; Venugopalan, C. S., Bioavailability and biotransformation of benzo(a)pyrene in an isolated perfused in situ catfish intestinal preparation. Environ. Health Perspect. 1998, 106, 155-166.

61. Lindström-Seppä, P.; Koivusaari, U.; Hänninen, O., Extrahepatic xenobiotic metabolism in North-European freshwater fish. Comp. Biochem. Physiol. 1981, 69C, 259-263.

62. Wheelock, C. E.; Eder, K. J.; Werner, I.; Huang, H.; Jones, P. D.; Brammell, B. F.; Elskus, A. A.; Hammock, B. D., Individual variability in esterase activity and CYP1A levels in Chinook salmon (Oncorhynchus tshawytscha) exposed to esfenvalerate and chlorpyrifos. Aquat. Toxicol. 2005, 74, 172-192.

273 63. Leticia, A.-G.; Gerardo, G., -B., Determination of esterase activity and characterization of cholinestereases in the reef fish Haemulon plumieri. Ecotoxicol. Environ. Saf. 2008, 71, 787- 797.

64. Al-Ghais, S. M.; Ahmad, S.; Ali, B., Differential inhibition of xenobiotic-metabolizing carboxylesterases by organotins in marine fish. Ecotoxicol. Environ. Saf. 2000, 46, 258-264.

65. Nichols, J. W.; Schultz, I. R.; Fitzsimmons, P. N., In vitro-in vivo extrapolation of quantitative hepatic biotransformation data for fish. I. A review of methods, and strategies for incorporating intrinsic clearance estimates into chemical kinetic models. Aquat. Toxicol. 2006, 78, 74-90.

66. Law, F. C. P.; Abenini, S.; Kennedy, C. J., A biologically based toxicokinetic model for pyrene in rainbow trout. Toxicol. Appl. Pharmacol. 1991, 110, 390-402.

67. Schultz, I. R.; Hayton, W. L., Interspecies scaling of the bioaccumulation of lipophilic xenobiotics in fish: An example using trifluralin. Environ. Toxicol. Chem. 1999, 18, 1440-1449.

68. Dyer, S.; Bernhard, M. J.; Cowan-Ellsberry, C.; Perdu-Durand, E.; Demmerle, S.; Cravedi, J.-P., In vitro biotransformation of surfactants in fish. Part II - Alcohol ethoxylate (C16EO8) and alcohol ethoxylate sulfate (C14EO2S) to estimate bioconcentration potential. Chemosphere 2009, 76, 989-998.

69. Nichols, J. W.; Fitzsimmons, P. N.; Whiteman, F. W., A physiologically based toxicokinetic model for dietary uptake of hydrophobic organic compounds by fish. II. Simulation of chronic exposure scenarios. Toxicol. Sci. 2004, 77, 219-229.

70. Nichols, J. W.; Fitzsimmons, P. N.; Burkhard, L. P., In vitro-in vivo extrapolation of quantitative hepatic biotransformation data for fish. II. Modeled effects on chemical bioaccumulation. Environ. Toxicol. Chem. 2007, 26, 1304-1319.

71. Fasano, W. J.; Sweeney, L. M.; Mawn, M. P.; Nabb, D. L.; Szostek, B.; Buck, R. C.; Gargas, M. L., Kinetics of 8-2 fluorotelomer alcohol and its metabolites, and liver glutathione status following daily oral dosing for 45 days in male and female rats. Chem. Biol. Interact. 2009, 180, 281-295.

72. de Wolf, W.; de Bruijn, J. H. M.; Seinen, W.; Hermens, J. L. M., Influence of biotransformation on the relationship between bioconcentration factors and octanol-water partition coefficients. Environ. Sci. Technol. 1992, 26, 1197-1201.

73. Barron, M. G.; Mayes, M. A.; Murphy, P. G.; Nolan, R. J., Pharmacokinetics and metabolism of trichlopyr butoxyehtyl ester in coho salmon. Aquat. Toxicol. 1990, 16, 19-32.

74. Karara, A. H.; Hayton, W. L., Metabolic inhibition and di-2-ethylhexyl phthalate pharacokinetics in fish. Drug Metab. Dispos. 1988, 16, 146-150.

75. Mackay, D.; Shui, W. Y.; Ma, K.-C.; Lee, S. C., Handbook of physical-chemical properties and environmental fate for organic chemicals, Second Edition. In CRC Press: Boca Raton, FL, 2006; Vol. 18. Insecticides.

274 76. Kenaga, E. E., Correlation of bioconcentration factors of chemicals in aquatic and terrestrial organisms with their physical and chemical properties. Environ. Sci. Technol. 1980, 14, 553-556.

77. Kawano, M.; Inoue, T.; Wada, T.; Hidaka, H.; Tatsukawa, T., Bioconcentration and residue patterns of chlordane compounds in marine animals: Invertebrates, fish mammals, and seabirds. Environ. Sci. Technol. 1988, 22, 792-797.

78. Southworth, G. R.; Keffer, C. C.; Beauchamp, J. J., The accumulation and disposition of benz(a)acridine in the fathead minnow, Pimephales promelas. Arch. Environ. Contam. Toxicol. 1981, 10, 561-569.

275 8.7. Supporting Information

Table S8.1. Common name, acronym and structure for analytes monitored.

Common Name Acronym Structure

F F F F F F F F O 8:2 fluorotelomer acrylate 8:2 FTAc F O F F F F F F F F

F F F F F F F F 8:2 fluorotelomer alcohol 8:2 FTOH F OH F F F F F F F F

O 8:2 saturated F F F F F F F F 8:2 FTCA F fluorotelomer carboxylate O- F F F F F F F F

F O 8:2 unsaturated F F F F F F 8:2 FTUCA F fluorotelomer carboxylate O- F F F F F F F F

O 7:3 saturated F F F F F F 7:3 FTCA F fluorotelomer carboxylate O- F F F F F F F F

O 7:3 unsaturated F F F F F F 7:3 FTUCA F fluorotelomer carboxylate O- F F F F F F F F

7:2 secondary F F F F F F OH 7:2 sFTOH F fluorotelomer alcohol CH 3 F F F F F F F F

F F F F F F F F 8:2 fluorotelomer sulfate F O 8:2 FTOH-Sulf conjugate O S O- F F F F F F F F O

F F F F F F F F COOH 8:2 fluorotelomer O O 8:2 FTOH-Gluc F OH glucuronide conjugate F F F F F F F F OH OH

276

F F F F F F F F O- 8:2 unsaturated F 8:2 FTUCA- F F F F F F S O fluorotelomer carboxylate O GSH glutathione conjugate H2N NH COO H N H COOH O

F F F F F F O 7:3 saturated F NH fluorotelomer carboxylate 7:3 FTCA-TA F taurine conjugate F F F F F F F

HO 3 S

F F F F O F Perfluorohexanoate PFHxA C O - F F F F F F

F F F F F F Perfluoroheptanoate PFHpA F C O- F F F F F F O

F F F F F F O F Perfluorooctanoate PFOA C O - F F F F F F F F

F F F F F F F F Perfluorononanoate PFNA F C O- F F F F F F F F O

277

80

70

60

50

Whole Fish (g) 40 Dose Control 30 Uptake Depuration 20 0 50 100 150 200 250 300 350

Time (hours)

Figure S8.1. Whole fish weight (g) during uptake and elimination phases for dosed (□) and control (▲) treatments. Data points for the dosed treatment represent the arithmetic mean (n=3 for “dose” treatment” and n=1 for “control” treatment) and error bars indicate 1 standard error.

278

2.2

2.0 Dose Control 1.8

1.6

1.4

1.2

1.0 Liver Somatic Index (%) 0.8 Uptake Depuration 0.6 0 50 100 150 200 250 300 350

Time (hours)

Figure S8.2. Liver somatic index (LSI, %)) during uptake and elimination phases for dosed (□) and control (▲) treatments. Data points for the dosed treatment represent the arithmetic mean (n=3 for “dose” treatment” and n=1 for “control” treatment) and error bars indicate 1 standard error.

10000 1000 8:2 FTCA 8:2 FTUCA

1000 100

100 10

10 1

Uptake Elimination Uptake Elimination 1 0.1 0 75 150 225 300 0 75 150 225 300 1000 1000 7:3 FTCA PFOA 100 PFNA 100 Concentration (ng/g ww) (ng/g Concentration 10 10 1

1 0.1 Uptake Elimination Uptake Elimination 0.1 0.01 0 75 150 225 300 0 75 150 225 300

Time (hours)

Figure S8.3. Kidney concentrations (ng/g ww) of 8:2 FTCA, 8:2 FTUCA, 7:3 FTCA, PFOA and PFNA from 8:2 FTAc dietary exposure in rainbow trout. 279

10000 1000 8:2 FTCA 8:2 FTUCA

1000 100

100 10

10 1

Uptake Elimination Uptake Elimination 1 0.1 0 75 150 225 300 0 75 150 225 300

1000 1000 7:3 FTCA PFOA

100

Concentration (ng/g ww) (ng/g Concentration 100

10

10 1

Uptake Elimination Uptake Elimination 1 0.1 0 75 150 225 300 0 75 150 225 300 Time (hours)

Figure S8.4. Blood concentrations (ng/g ww) of 8:2 FTCA, 8:2 FTUCA, 7:3 FTCA, PFOA and PFNA from 8:2 FTAc dietary exposure in rainbow trout. 280

10000 1000 8:2 FTCA 8:2 FTUCA 8:2 Gluc 100 1000

10

100 1

Uptake Elimination Uptake Elimination 10 0.1 0 75 150 225 300 0 75 150 225 300

1000 1000 7:3 FTCA PFOA

100 100 Concentration (ng/g ww)Concentration

10

10 1

Uptake Elimination Uptake Elimination 1 0.1 0 75 150 225 300 0 75 150 225 300 Time (hours)

Figure S8.5. Bile concentrations (ng/g ww) of 8:2 FTCA, 8:2 FTUCA, 8:2 FTOH-glucuronide, 7:3 FTCA, PFOA and PFNA from 8:2 FTAc dietary exposure in rainbow trout. 281

10000 10000 8:2 FTCA 8:2 FTUCA 1000 8:2 Gluc 1000

100 100 10

10 1

Uptake Elimination Uptake Elimination 1 0.1 0 75 150 225 300 0 75 150 225 300

1000 1000 7:3 FTCA PFOA

100 100 Concentration (ng/g ww)Concentration

10

10 1

Uptake Elimination Uptake Elimination 1 0.1 0 75 150 225 300 0 75 150 225 300 Time (hours)

Figure S8.6. Feces concentrations (ng/g ww) of 8:2 FTCA, 8:2 FTUCA, 8:2 FTOH-glucuronide, 7:3 FTCA, PFOA and PFNA from 8:2 FTAc dietary exposure in rainbow trout. 282

Table S8.2. “Active” 8:2 FTAc Incubations with Liver S9 Fractions.

7.6 ppb (15 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 7.6 ppb (15 nM) 8:2 FTAc Incubation 0.0 15 0.0 15 20 0.0 13 0.0 13 0.0 17 0.0 17 0.5 9.6 5.5 15 15 0.5 11 6.8 18 0.5 12 4.2 16 10 1.0 9.0 10 19 1.0 8.1 6.2 14 5

1.0 7.5 9.4 17 Concentration (nM) 1.5 7.5 9.2 17 1.5 6.3 7.4 14 0 1.5 6.9 9.7 17 0.00.51.01.52.02.5 2.0 7.7 11 19 Time (min) 2.0 7.3 10 17 2.0 5.4 11 16 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM)

46 ppb (89 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 0.0 98 0 98 46 ppb (89 nM) 8:2 FTAc Incubation 0.0 90 0 90 100 0.0 79 0 79 0.5 62 20 82 0.5 54 22 77 75 0.5 63 23 86 1.0 52 30 82 50 1.0 44 37 82 1.0 39 35 74 25

1.5 31 45 76 (nM) Concentration 1.5 36 49 85 1.5 36 42 78 0 2.0 30 52 82 0.0 0.5 1.0 1.5 2.0 2.5 3.0 2.0 33 52 85 Time (min) 2.0 28 54 82 2.5 36 48 84 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 2.5 29 48 78 2.5 26 52 79 283

103 ppb (199 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 103 ppb (199 nM) 8:2 FTAc Incubation 0.0 195 0 195 250 0.0 208 0 208 0.0 196 0 196 200 0.5 131 59 190 0.5 158 67 224 150 0.5 142 58 200 1.0 102 100 202 100 1.0 90 92 183

1.0 123 100 222 Concentration (nM) 50 1.5 64 131 195 1.5 73 113 185 0 1.5 69 123 192 0.0 0.5 1.0 1.5 2.0 2.5 2.0 87 134 221 Time (min) 2.0 65 132 197 2.0 64 144 208 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM)

190 ppb (365 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 190 ppb (365 nM) 8:2 FTAc Incubation 0.0 378 0 378 500 0.0 351 0 351 0.0 367 0 367 400 0.5 314 82 397 0.5 306 73 379 300 0.5 327 77 404 1.0 301 168 469 200 1.0 250 140 390

1.0 245 129 374 Concentration (nM) 100 1.5 241 181 422 1.5 216 215 431 0 1.5 217 256 473 0.0 0.5 1.0 1.5 2.0 2.5 2.0 192 239 431 Time (min) 2.0 158 265 423 2.0 152 295 447 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM)

284

345 ppb (665 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 345 ppb (665 nM) 8:2 FTAc Incubation 0.0 688 0 688 900 0.0 671 0 671 0.0 634 0 634 750 1.5 387 288 675 600 1.5 378 303 681 1.5 327 313 640 450 1.0 443 194 637 1.0 436 176 611 300

1.0 411 179 591 Concentration (nM) 150 2.0 301 346 647 2.0 275 346 622 0 2.0 370 467 838 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 3.0 186 422 607 Time (min) 3.0 158 490 648 3.0 252 496 748 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM)

500 ppb (963 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 500 ppb (963 nM) 8:2 FTAc Incubation 0.0 991 0 991 1200 0.0 912 0 912 1000 0.0 985 0 985 1.0 772 124 896 800 1.0 790 143 933 1.0 933 171 1104 600 2.0 587 259 846 400 2.0 703 247 949 2.0 675 268 943 (nM) Concentration 200 3.0 560 340 901 3.0 563 291 854 0 3.0 518 298 816 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Time (min)

8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM)

285

569 ppb (1094 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 569 ppb (1094 nM) 8:2 FTAc Incubation 0.0 1146 0 1146 0.0 1074 0 1074 1400 0.0 1061 0 1061 1200 1.0 850 166 1016 1000 1.0 842 178 1020 1.0 869 206 1075 800 2.0 599 298 897 600 2.0 723 332 1055 2.0 701 321 1022 400 Concentration (nM) 3.0 561 480 1041 200 3.0 601 503 1104 3.0 546 437 983 0 4.0 332 525 857 0.0 1.0 2.0 3.0 4.0 5.0 4.0 462 566 1029 Time (min) 4.0 508 611 1118 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM)

286

Table S8.3. Heat-inactivated control 8:2 FTAc Incubations with Liver S9 Fractions.

7.6 ppb (15 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 25 0.0 25 1.0 22 0.0 22 2.0 21 0.0 21 3.0 21 0.0 21

46 ppb (89 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 108 0.0 108 1.0 107 0.0 107 2.0 124 0.0 124 3.0 107 0.0 107

103 ppb (199 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 299 0.0 299 0.5 257 0.0 257 1.0 269 0.0 269 1.5 263 0.0 263 2.0 285 0.0 285

190 ppb (365 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 617 0.0 617 0.5 567 0.0 567 1.0 542 0.0 542 1.5 543 0.0 543 2.0 537 0.0 537

345 ppb (665 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 955 0.0 955 1.0 917 0.0 917 2.0 810 0.0 810 3.0 813 0.0 813 287

500 ppb (963 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 1061 0.0 1061 1.0 1053 0.0 1053 2.0 963 5.1 968 3.0 967 8.6 976

569 ppb (1094 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 1439 0.0 1439 1.0 1313 0.0 1313 2.0 1419 5.5 1425 3.0 1166 7.8 1173 4.0 1193 8.7 1202

288

Table S8.4. “Active” 8:2 FTAc Incubations with Stomach S9 Fractions.

20 ppb (39 nM) 8:2 FTAc Incubation 20 ppb (39 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 60 0.0 35 0.0 35 0.0 42 0.0 42 50 0.0 39 0.0 39 0.5 37 7.5 44 40 0.5 33 10 43 30 0.5 38 7.0 45

1.0 27 13 40 (min) Time 20 1.0 29 13 42 1.0 31 11 42 10 2.0 23 20 43 2.0 25 19 44 0 2.0 20 19 39 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 3.0 25 22 47 Concentration (nM) 3.0 26 22 48 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 3.0 24 20 44

47 ppb (89 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 47 ppb (89 nM) 8:2 FTAc Incubation 0.0 94 0.0 94 120 0.0 72 0.0 72 0.0 102 0.0 102 100 1.0 56 19 74 80 1.0 54 22 77 1.0 53 19 72 60 2.0 39 33 72 2.0 42 37 80 (min) Time 40 2.0 58 27 85 3.0 35 47 83 20 3.0 34 48 82 0 3.0 40 42 83 0.0 1.0 2.0 3.0 4.0 5.0 4.0 33 60 92 Concentration (nM) 4.0 33 52 85 4.0 34 58 92 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM)

289

76 ppb (146 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 76 ppb (146 nM) 8:2 FTAc Incubation 0.0 162 0.0 162 250 0.0 134 0.0 134 0.0 143 0.0 143 200 1.0 102 26 128 1.0 94 26 120 150 1.0 103 27 130 2.0 87 58 145 100

2.0 87 61 148 (min)Time 2.0 93 62 155 3.0 89 84 174 50 3.0 100 97 197 3.0 80 96 176 0 4.0 75 95 170 0.0 1.0 2.0 3.0 4.0 5.0 4.0 63 99 162 Concentration (nM) 4.0 66 108 174 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM)

150 ppb (289 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 150 ppb (289 nM) 8:2 FTAc Incubation 0.0 309 0.0 309 400 0.0 274 0.0 274 0.0 284 0.0 284 300 1.0 272 42 314 1.0 212 34 247 1.0 284 47 330 200 2.0 183 79 262 2.0 200 90 289 Time (min) 2.0 198 91 289 100 3.0 181 128 309 3.0 155 125 281 0 3.0 181 115 295 0.0 1.0 2.0 3.0 4.0 5.0 4.0 146 171 317 Concentration (nM) 4.0 161 177 338 4.0 152 166 318 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM)

290

262 ppb (505 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 262 ppb (505 nM) 8:2 FTAc Incubation 0.0 498 0.0 498 700 0.0 523 0.0 523 0.0 493 0.0 493 600 1.0 404 46 450 500 1.0 379 57 436 1.0 380 50 431 400 2.0 414 123 537 300

2.0 451 112 563 (min) Time 2.0 440 117 557 200 3.0 369 219 587 100 3.0 376 179 556 3.0 365 185 550 0 4.0 365 277 643 0.0 1.0 2.0 3.0 4.0 5.0 4.0 315 201 517 Concentration (nM) 4.0 299 233 533 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM)

914 ppb (1757 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 914 ppb (1757 nM) 8:2 FTAc Incubation 0.0 1553 0.0 1553 0.0 1935 0.0 1935 2500 0.0 1782 0.0 1782 1.0 1545 102 1647 2000 1.0 1676 99 1775 1.0 1739 95 1834 1500 2.0 1408 227 1636 2.0 1401 242 1643 1000 Time (min)Time 2.0 1619 198 1817 3.0 1306 320 1626 500 3.0 1419 379 1798 3.0 1577 401 1978 0 4.0 1406 549 1955 0.01.02.03.04.05.0 4.0 1422 486 1908 Concentration (nM) 4.0 1152 467 1619 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM)

291

2130 ppb (4095 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 2130 ppb (4095 nM) 8:2 FTAc Incubation 0.0 3731 0.0 3731 6000 0.0 4242 0.0 4242 0.0 4313 0.0 4313 5000 1.0 4102 154 4255 1.0 4831 150 4980 4000 1.0 3617 97 3714 2.0 3059 181 3240 3000

2.0 3477 287 3764 Time (min) 2000 2.0 3884 286 4170 3.0 4466 571 5037 1000 3.0 4143 517 4659 3.0 4257 483 4739 0 4.0 4421 646 5067 0.01.02.03.04.05.0 4.0 3374 565 3939 Concentration (nM) 4.0 3597 547 4144 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM)

3310 ppb (6368 nM) 8:2 FTAc Incubation 2130 ppb (4095 nM) 8:2 FTAc Incubation Time (min) 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 0.0 7348 0.0 7348 8000 0.0 7193 0.0 7193 0.0 4563 0.0 4563 6000 1.0 5196 137 5333 1.0 5007 151 5157 1.0 5290 112 5402 4000 2.0 4806 295 5101 Time (min) Time 2.0 5747 282 6029 2000 2.0 4441 286 4728 3.0 5618 435 6053 3.0 5016 484 5500 0 3.0 4518 417 4934 0.0 1.0 2.0 3.0 4.0 5.0 4.0 5303 617 5921 Concentration (nM) 4.0 4306 588 4894 8:2 FTAc (nM) 8:2 FTOH (nM) FTAc+FTOH (nM) 4.0 5161 625 5787

292

Table S8.5. Heat-inactivated control 8:2 FTAc Incubations with Stomach S9 Fractions.

20 ppb (39 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 35 0.0 35 1.0 33 0.0 33 2.0 36 0.0 36 3.0 35 0.0 35 4.0 35 0.0 35

47 ppb (89 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 63 0.0 63 1.0 62 0.0 62 2.0 58 0.0 58 3.0 - 0.0 - 4.0 60 0.0 60

76 ppb (146 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 134 0.0 134 1.0 124 0.0 124 2.0 133 0.0 133 3.0 124 0.0 124 4.0 134 0.0 134

150 ppb (289 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 305 0.0 305 1.0 291 0.0 291 2.0 298 0.0 298 3.0 288 0.0 288 4.0 290 0.0 290

293

262 ppb (505 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 586 0.0 586 1.0 539 0.0 539 2.0 500 0.0 500 3.0 480 0.0 480 4.0 505 0.0 505

914 ppb (1757 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 1534 0.0 1534 1.0 1477 0.0 1477 2.0 1499 0.0 1499 3.0 1650 0.0 1650 4.0 1681 0.0 1681

2130 ppb (4095 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 5652 0.0 5652 1.0 4270 0.0 4270 2.0 4867 0.0 4867 3.0 4510 0.0 4510 4.0 4743 0.0 4743

3310 ppb (6368 nM) 8:2 FTAc Incubation Time (min) FTAc (nM) FTOH (nM) FTAc+FTOH(nM) 0.0 5537 0.0 5537 1.0 6210 0.0 6210 2.0 6126 0.0 6126 3.0 5060 0.0 5060 4.0 5678 0.0 5678

294

CHAPTER NINE

Elucidating the Mechanism of Poly- and Perfluorinated Acid Production in Rainbow Trout

Craig M. Butt, Derek C.G. Muir and Scott A. Mabury

In preparation for submission to: Environmental Science & Technology

Contributions: Fish care and sampling, synthesis of parent compounds, preparation of dosed food, fish dissection, tissue extraction, instrumental analysis, data interpretation and manuscript preparation were performed by Craig Butt under the guidance of Derek Muir and Scott Mabury.

295 296 Chapter Nine – Elucidating the Mechanism of Poly- and Perfluorinated Acid Production in Rainbow Trout

9.1. Abstract

Several studies, utilizing both mammals and fish, have shown that fluorotelomer-based compounds can be metabolized to poly- and perfluorinated carboxylates, such as perfluorooctanoate (PFOA). Thus far, research has predominately focused on the biological fate of the 8:2 fluorotelomer alcohol (8:2 FTOH). However, there is considerable uncertainty in the literature regarding the biotransformation pathway of FTOHs. Specifically, many proposed mechanisms deviate concerning the fate of the 8:2 fluorotelomer unsaturated carboxylate (FTUCA). In addition, there is ambiguity regarding the fate of the 7:3 fluorotelomer saturated carboxylate (FTCA). Therefore, the objective of this study was to further elucidate the mechanism for 8:2 FTOH biotransformation through dosing rainbow trout with three 8:2 FTOH metabolism intermediates, specifically the 7:3 FTCA, 8:2 FTCA and 8:2 FTUCA. This study represents the first investigation of these three reactive intermediate metabolites in an in vivo system. The parent compounds were dosed via the diet and levels of the parent compounds and intermediates were monitored in the blood and liver during the 7 day uptake phase and 10 day elimination phase. Exposure to the 7:3 FTCA did not result in the formation and accumulation of PFOA. Instead, 7:3 FTCA exposure resulted in the formation of low levels of the 7:3 FTUCA and perfluoroheptanoate (PFHpA), a novel finding. PFOA was formed in the 8:2 FTCA and 8:2 FTUCA dosing, confirming the “beta-like-oxidation” that has been previously proposed. In addition, the 7:3 FTCA was formed during exposure to both the 8:2 FTCA and 8:2 FTUCA. Elimination half-lives were 5.1 d (95% confidence interval: 3.1-14 d) for 7:3 FTCA, 1.2 d (1.1-1.3 d) for 8:2 FTCA and 0.39 d (0.31-0.53 d) for 8:2 FTUCA. The observed differences in the elimination half-life may be the result of differences in either the depuration or metabolism rate. Based on the findings of this study, and reported analogous literature mechanisms, we propose a “beta-like-oxidation” mechanism for PFOA formation proceeding from the 8:2 FTUCA > 7:3 β-keto acid > 7:2 ketone > PFOA.

297 9.2. Introduction

In 2001, Giesy & Kannan [1] reported the world-wide dissemination of perfluorooctanoate (PFOA), perfluorooctane sulfonate (PFOS) and perfluorooctane sulfonamide (PFOSA) in wildlife. Additional research by Martin et al. [2] demonstrated that wildlife were contaminated with a much broader suite of perfluoroalkyl compounds (PFCs), including long- chain perfluorinated carboxylates (PFCAs) of more than eight carbons. Further, PFCs are ubiquitously detected in humans [3-5]. However, since their initial identification in human and wildlife tissues, the source of PFCs has remained elusive and has been the domain of intensive scientific interest. Several studies have attempted to quantify human PFC exposure pathways [6-8], although these studies are limited by their failure to include newly identified indirect routes of exposure such as the polyfluoroalkyl phosphates [9]. In fact, D’eon & Mabury [10] showed the formation of PFOA from the metabolism of fluorotelomer-based phosphates (PAPs) in rats prior to publication of references 7 & 8.

Considerable effort has focused on the role of precursor compounds as potential sources of PFCs. One mechanism is through the atmospheric oxidation of precursors which form PFCAs that are subsequently deposited. Laboratory studies have shown that fluorotelomer- based compounds, including alcohols (FTOHs) [11, 12], olefins [13, 14], iodides [15] and acrylates [16] form PFCAs. As well, the atmospheric oxidation of fluorinated sulfonamide alcohols has been shown to form PFCAs [17, 18]. An additional mechanism, and the focus of this paper, is through the biotransformation of precursors. The seminal work by Hagen et al. [19] showed that the 8:2 FTOH was metabolized in rats to form intermediate metabolites, fluorotelomer saturated and unsaturated acids (FTCAs & FTUCAs), as well as PFOA. In recent years, several additional studies have shown the formation of PFCAs and intermediate metabolites from 8:2 FTOH biotransformation, including microbes [20-24], and in vivo studies of rats [25, 26] and mice [27, 28]. In addition, Nabb et al. [29] investigated the in vitro biotransformation of 8:2 FTOH in rat, mouse, trout and human hepatocytes and microsomes. Although the above studies investigated the biotransformation of the 8:2 FTOH, presumably analogous degradation mechanisms occur for other chain-length FTOHs. For example, Brandsma et al. [30] showed the formation of 10:2 FTCA, 10:2 FTUCA and PFDA from 10:2 FTOH dietary exposure in rainbow trout. An analogous series of metabolites was also detected

298 from 8:2 FTOH exposure [30]. As well, fluorotelomer-based compounds that are metabolized to FTOHs also have been shown to form PFCAs. As mentioned above, D’eon & Mabury [10] showed the formation of PFOA from PAPs metabolism in rats. Finally, Butt et al. [31] showed that the 8:2 FTOH acrylate was biotransformed by rainbow trout to form PFOA, PFNA and related intermediate metabolites.

There is considerable variability in literature regarding the proposed biotransformation pathway of FTOHs. It is commonly reported that the first degradation step is the oxidation of the alcohol group to form the fluorotelomer aldehyde (FTAL), followed by oxidation of the FTAL to yield the FTCA. It is at this step that most biotransformation pathways deviate. In particular, there is uncertainty regarding the fate of the 7:3 FTCA. This metabolite has been observed in several 8:2 FTOH biotransformation studies [21, 22, 24-26, 29]. Martin et al. [25], Fasano et al. [26] and earlier work by Wang et al. [21, 22] suggested that the 7:3 FTCA will form PFOA through β-oxidation. However, Nabb et al. [29] as well as the more recent work of Wang et al. [24] have shown that PFOA is not formed when 7:3 FTCA is used as the substrate.

The objective of this study was to further elucidate the biotransformation mechanism of 8:2 FTOH in rainbow trout by means of individual dosing with three reactive intermediate metabolites identified in previous 8:2 FTOH biotransformation studies – 8:2 FTCA, 8:2 FTUCA and 7:3 FTCA. This is the first investigation of these compounds in an in vivo system. The test compounds were identified in a recent rainbow trout biotransformation study [31] using the 8:2 FTOH acrylate as the parent compound, and represent important branching points in the 8:2 FTOH degradation mechanism in which there exists uncertainty in the literature. The animals were exposed via dietary exposure and levels of the parent compounds and intermediates were monitored in the blood and liver during the 7 day uptake phase and 10 day elimination phase.

299 9.3. Materials and Methods

9.3.1. Standards and Reagents The 7:3 FTCA (2H, 2H, 3H, 3H-perfluorodecanoic acid, 97%) was obtained from Synquest Labs Inc. (Alachua, FL). Ethyl 3-aminobenzoate methanesulfonate salt (MS-222) was purchased from Sigma-Aldrich (Oakville, Ontario). Standards for PFHxA, PFHpA, PFOA, 13 13 PFNA, 8:2 FTCA and 8:2 FTUCA as well as the stable isotope standards ( C4-PFOA, C5- 13 13 PFNA, C2-8:2 FTUCA, C2-8:2 FTCA) were provided by Wellington Laboratories (Guelph, ON).

9.3.2. Synthesis of 8:2 FTCA and 8:2 FTUCA The 8:2 FTCA and 8:2 FTUCA were synthesized using methods described by Achilefu et al. [32]. Briefly, the 8:2 FTOH was dissolved in acetone: ether (3:1 v/v) and Jones’ reagent was added dropwise until the solution remained reddish-brown. The solution was extracted with ether. The ether extract was washed with water, chemically dried using sodium sulfate and rotary evaporated to dryness. The crude product was recrystallized with toluene to yield the 8:2 FTCA. Analysis by 1H-NMR confirmed the absence of the 8:2 FTOH starting material. Further, the chemical shift characteristic of the vinyl group was not observed, confirming that the 8:2 FTUCA was not formed. The 8:2 FTUCA was synthesized by the reaction of 2M NaOH with 8:2 FTCA (synthesized as described above) dissolved in tetrahydrofuran. The reaction mixture was stirred at room temperature for 2 hours, followed by the addition of 2M HCl. The solution was extracted with ether, the ether extract chemically dried with sodium sulfate and evaporated to dryness. The crude product was recrystallized with toluene, yielding the 8:2 FTUCA. The product was confirmed by 1H- and 19F-NMR analysis.

9.3.3. Food Preparation Separate batches of 8:2 FTCA, 8:2 FTUCA and 7:3 FTCA spiked food (target concentration: 100 µg/g) were prepared by adding commercial fish food (Martin’s floating feed, size 3, Martin Mills, Elmira, ON) to a round-bottom flask and adding sufficient acetone to cover the pellets. The parent compounds were added in a small volume (~1 ml) of acetone and rotary evaporated to dryness. The spiked food was removed from the flask onto methanol-rinsed

300 aluminum foil and allowed to dry overnight in the fume hood. Control food was prepared in an identical manner without the addition of the fluoroalkyl compounds.

The spiked food was analyzed by LC-MS/MS to confirm dosing concentration and to identify any impurities. Three separate grabs of spiked food (0.5 g each) were crushed with a mortar and pestle and extracted in 10 ml of ethyl acetate. One ml of the ethyl acetate extract was reduced to dryness under a gentle stream of N2 gas and reconstituted in 10 ml of methanol. The mean 7:3 FTCA concentration in the 7:3 FTCA spiked food was 71.5 ± 10.0µg/g (arithmetic mean± standard error) and no other analytes were detected. The 8:2 FTCA spiked food had a mean 8:2 FTCA concentration of 103.2 ± 6.9 µg/g with the 8:2 FTUCA detected as an impurity at a mean level of 1.5 µg/g ± 0.14 µg/g. The synthesized 8:2 FTCA compound did not show any traces of 8:2 FTUCA, suggesting the impurity detected was formed sometime during the food preparation. In the 8:2 FTUCA spiked food, the mean 8:2 FTUCA concentration was 69.1 ± 1.2 µg/g. In addition, trace quantities of unreacted 8:2 FTCA (0.7 µg/g, ~1% of the 8:2 FTUCA concentration) were detected in the 8:2 FTUCA food. Perfluoroheptanoate, PFOA and PFNA were not detected in any of the food preparations.

9.3.4. Fish Care and Sampling Juvenile rainbow trout were purchased from a local hatchery (Rainbow Springs Trout Hatchery, Orangeville, Ontario) and allowed to acclimate for 2 weeks prior to dosing. The initial fork length was approximately 7 inches and initial weight was approximately 60 g. Trout were kept at the Aquatic Facility, Department of Cell and Systems Biology (University of Toronto) in 475 L fiberglass tanks under flow-through conditions (~1 L/min) using carbon- filtered, dechlorinated City of Toronto water. The water temperature was maintained at 18oC and a 12 hour daily photoperiod was used. Care and treatment of the fish was approved by the University of Toronto Animal Care Committee and was in compliance with the guidelines of the Canadian Council on Animal Care.

An intensive feeding and sampling schedule was designed, consisting of a 168 hour uptake phase and 240 hour elimination phase. During the uptake phase fish were fed the dosed or control food once per day at 1.5% of the average initial body weight per feeding. Fish were collected at -1 hour (pre-dosing), 12 hr, 24 hr, 48 hr, 72 hr, 120 hr and 168 hr. During the

301 elimination phase, fish were fed clean food once per day. Elimination samples were collected at 24 hr, 48 hr (8:2 FTUCA treatment only), 72 hr, 120 hr, 168 hr and 240 hr. During all time points, 3 dosed fish and 1 control were collected. The 8:2 FTUCA dosing experiment was conducted at a separate occasion and thus has a unique set of control samples (n=5) and method detection limits. Fish were euthanized by an overdose exposure to MS-222 (4 g/L buffered to pH 7 with sodium carbonate) and blood was immediately drawn through cardiac puncture using heparin rinsed syringes. Within 30 min of euthanization, fish were dissected in the laboratory and the liver was removed. Blood and liver samples were kept frozen (-20 oC) until analysis.

9.3.5. Extraction and Clean-up Methods Liver (~0.5 g) and blood (300 µl) were sub-sampled and placed in 15 ml polypropylene 13 13 13 centrifuge tubes. The suite of stable isotope internal standards ( C4-PFOA, C5-PFNA, C2- 13 8:2 FTUCA, C2-8:2 FTCA) was added to each sample prior to extraction. Liver was homogenized for 1 min using a mechanical mixer (Tissue-Tearor™, Biospec Products, Bartlesville, OK) in 8 ml of ethyl acetate. The extracts were centrifuged and the solvent decanted into a clean polypropylene tube. The blood samples were extracted by gently shaking for 5 min with 4 ml of ethyl acetate. The mixture was centrifuged, the solvent decanted into a clean polypropylene tube, the extraction repeated and solvent fractions combined. The liver and blood extracts were blown down to dryness under a gentle stream of N2 gas and reconstituted in 1 ml of methanol for LC-MS/MS analysis.

9.3.6. Instrumental Analysis Instrumental analysis was performed by liquid chromatography with negative electrospray-tandem mass spectrometry under conditions described previously [33]. Sample extracts were analyzed using an API 4000 Q Trap (Applied Biosystems/MDS Sciex, Concord, ON, Canada) coupled to an Agilent 1100 pump (injection volume: 10 µl, flow rate: 300 µl/min). Chromatography was performed using an ACE C18 column (Advanced Chromatography Technologies, Aberdeen, UK; length, 50 mm; inner diameter, 2.1 mm; particle size, 3 µm), preceded by a C18 guard cartridge (length, 4.0 mm, inner diameter, 2.0 mm; Phenomenex, Torrance, CA, USA).

302

13 Analyte responses were normalized to internal standard responses. C4-PFOA was 13 13 used for PFHxA, PFHpA and PFOA, C5-PFNA for PFNA, C2-8:2 FTCA for 8:2 FTCA and 13 7:3 FTCA, C2-8:2 FTUCA for 8:2 FTUCA, 7:3 FTUCA and 8:2 FTUCA-glutathione conjugate (8:2 FTUCA-GSH). Authentic standards were not available for the 7:3 FTUCA, 8:2 FTUCA-GSH and 7:3 FTCA taurine conjugate, and these analytes were identified based on expected MS/MS transitions. The 7:3 FTUCA was quantified using the response factor for the 8:2 FTUCA.

9.3.7. Statistical Analyses and Data Treatment The instrumental detection limits were determined as the concentration that produced a peak with a signal to noise ratio of at least three. Method detection limits (MDLs) were determined as threefold the standard deviation of the results in control fish samples analyzed (n=12 for 8:2 FTCA and 7:3 FTCA treatment, n=5 for 8:2 FTUCA treatment). If analytes were not detected in the blanks, half the instrumental detection limit was used as the MDL. Concentrations less than the instrumental detection limit were reported as nondetect. Concentrations were blank corrected using the mean level in the control fish. After blank correcting, concentrations less than the MDL were reported as less than the MDL value. For calculation of mean tissue concentrations, values

The liver somatic index (LSI) was calculated as,

(g) massliver massliver (g) (%) LSI (%) = x 100% (g) massfish whole massfish (g)

Fish growth rates were determined by fitting measured body weights to the model, ln(mass) = a + b * t, where b is the growth rate (fish mass in grams/time in hours), t is the time in hours, and a is a constant [34]. Blood and liver concentrations were corrected for growth dilution by multiplying the elimination phase concentrations by (1 + b * t), where t is the time since the start of elimination (hours). Elimination rates (kd) were calculated using growth- corrected tissues levels and fitting to the first-order elimination equation, ln(Cfish) = a + kd * t,

303 where Cfish is the growth-corrected blood or liver concentration, kd is the first-order elimination rate constant (hour-1), t is time (hours), and a is a constant. Regression analysis

(α=0.05) was used to test if the slope of the ln(Cfish) versus time relationship was significantly different from zero during the elimination phase. Elimination half-lives (t1/2) were calculated as ln 2/kd.

9.4. Results and Discussion

9.4.1. Physical Indices No mortalities were observed during the experiment in either the dosed or control treatments. The liver somatic index (LSI) is a measure of the liver mass relative to the whole body and is generally used as a measure of metabolic stress resulting from contaminant exposure (Table 9.1). The overall mean LSI was not statistically different (ANOVA, Tukey’s HSD post-hoc test) between the control and 7:3 FTCA (p=0.35), 8:2 FTCA (p=0.69) and 8:2 FTUCA (p=0.73) treatments.

The overall growth rate was significant for the 7:3 FTCA, 8:2 FTCA and 8:2 FTUCA treatments with fish weight increasing at ~0.9 g/day. Growth rate for the control treatment was ~1.0 g/day, but this growth rate was not statistically significant (p=0.44). We believe the discrepancy in statistical significance between the dosed and control treatments is due to the much larger sample size of the dosed groups.

Table 9.1. Liver Somatic Index (%) of control and dosed treatments at 12 and 120 hours of the uptake phase and 72 and 240 hours of the elimination phase. Treatment LSI (%) Control 7:3 FTCA 8:2 FTCA 8:2 FTUCA 12 hours of uptake 1.05 1.24 ± 0.13 1.18 ± 0.10 1.90 ± 0.22 120 hours of uptake 1.57 1.45 ± 0.02 1.14 ± 0.03 1.57 ± 0.05 72 hours of elimination 1.53 1.31 ± 0.03 1.13 ± 0.05 1.45 ± 0.10 240 hours of elimination 1.38 1.58 ± 0.30 1.51 ± 0.20 1.43 ± 0.11

304 9.4.2. 7:3 FTCA Exposure The parent 7:3 FTCA compound was rapidly accumulated from the dosed food with very high levels detected in the blood and liver within 12 hours after initial dosing (Figure 9.1). Mean tissue concentrations are presented in the Supporting Information. The 7:3 FTCA blood levels increased throughout the uptake phase, but appeared to reach steady-state by 120 hours at ~2150 ng/g ww. In contrast, the liver concentrations appeared to achieve steady-state concentrations within 12 hours of dosing at ~1500 ng/g ww. The only metabolites detected during 7:3 FTCA exposure were the unsaturated analogue, 7:3 FTUCA, and PFHpA. The 7:3 FTUCA was formed in very low yields (blood only), with steady-state concentrations reaching ~0.01-0.02% of the parent 7:3 FTCA. As a result of the low yields, 7:3 FTUCA levels were only above the MDL (0.01 ng/g ww) in the latter stage of the uptake phase and early during the elimination phase. Similarly, PFHpA was formed in very low yields, with mean steady-state levels of 3.0 and 2.4 ng/g ww in the blood and liver respectively. These concentrations represented ~0.1-0.2% of the parent 7:3 FTCA steady-state levels. The observation of PFHpA formation from 7:3 FTCA exposure is a novel finding and this pathway may represent the source of PFHpA observed in several 8:2 FTOH exposure studies [26, 29, 35].

PFOA formation was not observed in the blood or liver. These results do not support the earlier suggestions by Martin et al. [25], Fasano et al. [26] and Wang et al. [21, 22] that postulated the 7:3 FTCA may form PFOA through β-oxidation. These results are surprising since it was expected that one round of β-oxidation would yield PFOA. However, the findings of the present study are consistent with those of Nabb et al. [29] who did not observe PFOA formation during the incubation of rat, mouse, human and trout hepatocytes with 7:3 FTCA and 7:3 FTUCA. Nabb et al. [29] did not detect the formation of PFHpA from 7:3 FTCA biotransformation, as was observed in the present study. Instead, it has been postulated that the exclusive fate of the 7:3 FTCA is conjugation with taurine [29, 35]. The 7:3 taurine conjugate was not observed in the present study, although since an authentic standard was not available this analyte was only monitored using the expected MS>MS transition. Further, it is unknown if the 7:3 taurine conjugate is stable in the extraction and analysis procedure. The mechanism for PFHpA formation is not known and warrants further study.

305 The growth corrected 7:3 FTCA elimination half-life was 5.1 days (95% confidence intervals: 3.1-14 days) for blood and 10.3 days (6.4-26 days) for liver (Table 9.2). Elimination half-lives could not be calculated for PFHpA and 7:3 FTUCA since concentrations of these metabolites were below MDL almost immediately after elimination began. The relatively long half-life of the 7:3 FTCA, as compared to the 8:2 FTCA and 8:2 FTUCA, may suggest that this compound is a suitable biomarker for 8:2 FTOH exposure. Powley et al. [36] measured low ng/g ww levels of the 7:3 FTCA in ringed seal and bearded seal liver from the western Canadian arctic, but did not detect the 8:2 FTCA or FTUCA.

10000 Blood 1000

100

7:3 FTCA 10 PFHpA 7:3 FTUCA 1

0.1

0.01 10000 Liver Concentration (ng/g ww) 1000

100

10

1

0.1 0 100 200 300 400 500 Time (hours)

Figure 9.1. Mean blood and liver concentrations (ng/g ww) of 7:3 FTCA, PFHpA and 7:3 FTUCA resulting from 7:3 FTCA dietary exposure. Error bars represent 1 standard error.

306 9.4.3. 8:2 FTCA Exposure Similar to the 7:3 FTCA, the parent 8:2 FTCA was rapidly accumulated from the spiked food with very high concentrations measured in blood and liver within 12 hours after initial dosing (Figure 9.2). The 8:2 FTCA concentrations increased slightly during the experiment with apparent steady-state concentrations, ~5500 ng/g ww and ~2500 ng/g ww for blood and liver, respectively, reached within 72 hours.

In addition, the 8:2 FTCA was rapidly metabolized with levels of 8:2 FTUCA, 7:3 FTCA, PFOA and PFHpA observed within 12 hours of initial dosing. Levels of 7:3 FTUCA and PFNA were below the MDL at the 12 hour time point, but were quantifiable at 24 hours and throughout the remainder of the experiment. The metabolite accumulated in the highest quantity was the 7:3 FTCA, reaching 940 ng/g ww and 870 ng/g ww in the blood and liver, respectively, by the end of the uptake phase. The 7:3 FTCA concentrations did not reach steady-state, but continued to increase in both tissues throughout the uptake phase, presumably as the result of the relatively longer half-life of this compound as compared to the 8:2 FTCA and 8:2 FTUCA. The 8:2 FTUCA was also formed and accumulated in high levels, reaching steady-state concentrations of ~100 ng/g ww and ~200 ng/g ww in blood and liver, respectively.

Trace levels of the 8:2 FTUCA (1.5 µg/g) were detected in the 8:2 FTCA spiked food, and may be responsible for some of the 8:2 FTUCA observed in the fish tissues. However, this likely contributed only a trace amount to the total 8:2 FTUCA body burden. The average liver mass was ~1 g, and the average whole fish mass was ~75 g equating to a blood mass of ~6.6 g, assuming the blood represents 7.4% of total fish mass [37]. Since each fish was fed ~0.75 g of food per day, the daily dose of 8:2 FTUCA from the 8:2 FTCA food was 1.1 µg. This equates to a combined concentration increase in the blood and liver of ~150 ng/g ww per day if it is assumed that the 8:2 FTUCA is completely accumulated by the fish, with no depuration or biotransformation occurring, and the 8:2 FTUCA is partitioned only and equally into blood and liver. These conditions are highly conservative and not representative of pharmacokinetics. The mean blood and liver 8:2 FTUCA concentrations were 56 ng/g ww and 120 ng/g ww, respectively, 12 hour after initial dosing. Thus, it is doubtful that the 8:2 FTUCA contamination in the food contributed significantly to the 8:2 FTUCA body burden.

307 The terminal metabolites, PFOA and PFNA, were formed in low yields from the 8:2 FTCA biotransformation. Tissue PFOA concentrations increased throughout the uptake phase and reached levels of 8.8 ng/g ww and 56 ng/g ww in the blood and liver, respectively. Further, PFOA blood concentrations continued to increase for 5 days after the beginning of elimination, however, liver concentrations showed an apparent immediate decrease. These trends are suggestive of formation from PFOA precursors (i.e. FTCA & FTUCA) that were still present in the body. PFNA concentrations in the blood were predominately below the MDL during the uptake phase, but were above MDL during the elimination phase. Interestingly, PFNA blood concentrations continued to increase during the elimination phase until the latter stages of elimination. Finally, the 7:3 FTUCA and PFHpA were formed and accumulated in low concentrations (~1-2 ng/g ww). Drawing from the results of the 7:3 FTCA dosing, it is assumed that these metabolites were formed from the biotransformation of the 7:3 FTCA that was produced in high concentrations from the 8:2 FTCA metabolism.

The parent 8:2 FTCA compound and the metabolite 8:2 FTUCA were rapidly eliminated, presumably through a combination of depuration and biotransformation (Table 9.2). Growth corrected elimination half-lives in blood were 1.2 days (1.1-1.3 days) for 8:2 FTCA and 1.3 days (1.1-1.5 days) for 8:2 FTUCA, and in liver were 1.3 days (1.1-1.4 days) for 8:2 FTCA, 1.8 days (1.0-8.6 days) for 8:2 FTUCA (Table 9.2). The 7:3 FTCA showed an immediate decrease in the blood, whereas, 7:3 FTCA liver concentrations remained steady during the initial 5 days of elimination. These trends are reflected in the 7:3 FTCA elimination half-lives of 3.5 days (2.9-4.3 days) for blood and 6.8 days (4.4-15.0 days) for liver. Not surprisingly, a longer elimination half-life for 7:3 FTCA in blood was also shown in the 7:3 FTCA exposure. The opposite trend was shown for PFOA in which blood levels increased for the initial 5 days of elimination, but liver levels showed an immediate decrease. As such, the slope of the PFOA versus time relationship was not statistically different (p=0.41) from zero during elimination and thus the half-life could not be calculated. However, the PFOA elimination half-life in the liver was 4.1 days (2.8-7.7 days) which is within the range reported for PFOA when dosed as the parent compound [38, 39]. Blood PFNA concentrations were predominately steady during the elimination phase and the elimination half-life was not statistically significant from zero. These results are consistent with the relatively long half-life of PFNA in rainbow trout - 16 days in

308 blood and 6 days in liver [39]. The elimination half-life for PFNA in liver could not be calculated since most data points were below the MDL.

10000 Blood Blood 1000

100

10

1

8:2 FTCA 0.1 7:3 FTCA PFOA 10000 8:2 FTUCA PFNA Liver Liver 1000 Concentration (ng/g ww) 100

10

1 0 100 200 300 400 500 0 100 200 300 400 500 Time (hours) Time (hours) Figure 9.2. Mean blood and liver concentrations (ng/g ww) of 8:2 FTCA, 7:3 FTCA, 8:2 FTUCA, PFOA and PFNA resulting from 8:2 FTCA dietary exposure. Error bars represent 1 standard error.

9.4.4. 8:2 FTUCA Exposure In a similar manner to the 7:3 FTCA and 8:2 FTCA, the 8:2 FTUCA was rapidly accumulated within the blood and liver tissues (Figure 9.3). The 8:2 FTUCA appeared to reach steady-state concentrations within 12 hours of dosing at concentrations of ~1400 ng/g ww and ~70 ng/g ww in blood and liver, respectively. Interestingly, liver concentrations were ~70-fold lower as compared to the blood. The remarkable difference between 8:2 FTUCA tissue concentrations is suggestive of rapid biotransformation within the liver. These results are in contrast to the 7:3 FTCA and 8:2 FTCA dosing experiments in which levels of the parent compounds were within ~2-fold between blood and liver. The 7:3 FTCA was formed from the

309 8:2 FTUCA exposure and accumulated within 12 hours of dosing with levels initially increasing and appearing to reach steady-state near the end of the uptake phase at 150 ng/g ww and ~270 ng/g ww in blood and liver, respectively. The 7:3 FTUCA was also formed (data not shown), reaching concentrations of ~1 ng/g ww. Interpretation of PFOA trends was constrained by the relatively high PFOA blank levels resulting in MDL values of 5.4 and 9.6 ng/g ww in blood and liver, respectively. However, PFOA was formed and accumulated with levels increasing throughout the uptake phase, reaching concentrations of ~10 and 15 ng/g ww in blood and liver, respectively. In addition, the 8:2 FTUCA-GSH was detected in the liver using the expected MS-MS transition. Authentic standards for this compound were not available and thus were quantified using arbitrary units.

Detection of the 8:2 FTUCA-GSH conjugate in liver is consistent with several in vitro and in vivo studies examining 8:2 FTOH exposure [25, 26, 29, 35]. Martin et al. [40] showed that rat hepatocytes exposed to 8:2 FTOH resulted in an intracellular GSH depletion of 93%. Further, Fasano et al. [26] hypothesized that the FTUCA, the unsaturated fluorotelomer aldehyde and associated GSH conjugates may be responsible for liver lesions observed in 8:2 FTOH dosed rats. However, in whole rat experiments, Fasano et al. [35] showed no change in GSH liver store levels after oral dosing to 8:2 FTOH.

Trace quantities of 8:2 FTCA (0.7 µg/g, equivalent to 1% of the 8:2 FTUCA concentration) were detected in the 8:2 FTUCA dosed food, presumably as the result of unreacted starting material from the 8:2 FTUCA synthesis. The biotransformation of the residual 8:2 FTCA likely contributed to some of the 8:2 FTUCA body burden in this experiment.

The mechanism for the formation of the 7:3 FTCA is not fully known. Accumulation of high levels of the 7:3 FTCA in the 8:2 FTUCA dosing experiment is suggestive of formation from the 8:2 FTUCA, although it is possible that the 7:3 FTCA was formed directly from the 8:2 FTCA impurity in the 8:2 FTUCA food. To investigate this, we evaluated the accumulation ratio (defined as ratio of 7:3 FTCA tissue concentration: 8:2 FTCA food concentration at steady- state). The accumulation ratios for 7:3 FTCA from 8:2 FTCA exposure were 0.91 and 0.84 for blood and liver, respectively. However, these ratios are likely underestimated since the 7:3

310 FTCA did not reach steady-state in the 8:2 FTCA dosing experiment. The 8:2 FTCA level in the 8:2 FTUCA food was 0.7 µg/g, which would be expected to yield concentrations of 6.4 ng/g ww and 5.9 ng/g ww in the blood and liver, respectively. These concentrations are ~25- and ~45-fold lower than what was measured in the blood and liver, respectively. Thus, it is very likely that the 7:3 FTCA observed in the tissues originated from the 8:2 FTUCA dose and not the 8:2 FTCA impurities in the food.

The 8:2 FTUCA was eliminated very quickly with an elimination half-life in blood of 0.39 days (0.31-0.53 days) (Table 9.2). Liver concentrations were below the MDL (1.2 ng/g ww) within 24 hrs of commencing elimination and thus an elimination half-live could not be calculated. The elimination half-life of PFOA in blood was 0.54 days (0.3-6.1), whereas, liver PFOA concentrations were below the MDL (9.6 ng/g ww) in all of the elimination samples. Interestingly, the PFOA concentrations did not show an initial delay in elimination as observed in the 8:2 FTCA treatment. Presumably, this was due to the much faster elimination half-life of the 8:2 FTUCA as compared to the 8:2 FTCA.

311

10000 Blood Blood

1000 8:2 FTUCA

100

10

8:2 FTUCA 1 7:3 FTCA 8:2 FTCA 10000 Liver Liver 1000 PFOA 8:2 FTUCA-GSH Concentration (ng/g ww) 100

10

1 0 100 200 300 400 500 0 100 200 300 400 500 Tiime (hours) Tiime (hours)

Figure 9.3. Mean blood and liver concentrations (ng/g ww) of 8:2 FTUCA, 7:3 FTCA, 8:2 FTCA and PFOA resulting from 8:2 FTUCA dietary exposure. The y-axis scale is equivalent for all graphs. The 8:2 GSH is plotted in arbitrary units. Error bars represent 1 standard error.

9.4.5. Elimination Half-lives: Comparison of Intermediate and Terminal Metabolites The elimination half-lives for the intermediate metabolites (i.e. 8:2 FTCA, 8:2 FTUCA and 7:3 FTCA) investigated were generally much shorter compared to those of the terminal metabolites (i.e. PFOA and PFNA) that had been directly dosed [38, 39]. Presumably this trend is due to the fact that the intermediate metabolites are metabolized, increasing their overall elimination rate. Using a similar experimental system as the present study, De Silva et al. [39] reported linear isomer (n-) PFOA half-lives of 5.6 days (95% CI: 4.8-6.7 days) and 3.7 days (3.2-4.4 days) and n-PFNA half-lives of 15.9 days (13.5-19.6 days) and 6.0 days (5.2-7.1 days) for blood and liver, respectively. Similarly, Martin et al. [38], also through dietary exposure to rainbow trout, reported a liver half-life for the n-PFOA of 5.2 days ± 0.48 (±1 SE); PFNA was not investigated. In comparison the present study showed that the half-lives in blood and liver,

312 respectively, were 5.1 days (3.1-14 days) and 10.3 days (6.4-26.0 days) for 7:3 FTCA, 1.2 days (1.1-1.3 days) and 1.3 (1.1-1.4 days) for 8:2 FTCA, and 0.39 days (0.31-0.53 days) for the 8:2 FTUCA (blood only).

As stated earlier, the overall elimination for the intermediate metabolites is a combination of depuration and metabolism. Given that all three intermediate metabolites are the same chain-length and are generally structurally similar, they could be expected to exhibit similar hydrophobicity as well as comparable binding affinity to anion transporter or serum proteins; thus similar depuration rates among the compounds could be anticipated. Therefore, the observed differences in the calculated elimination half-lives is likely attributable to differences in metabolism rate (KM). The very short half-life of the 8:2 FTUCA is suggestive of a very high metabolic rate and is consistent with the comparatively low levels of this compound detected in the liver. Further, the relatively long half-life of the 7:3 FTCA is indicative of a much lower metabolic rate and is consistent with the overall lower concentrations of metabolites formed from this compound. The overall aim of this study was to elucidate the biotransformation mechanism for the polyfluorinated intermediate metabolites, rather than quantification of the elimination kinetics. However, future studies investigating the in vitro biotransformation of these and other intermediates, such as using S9 or microsomal fractions, would greatly assist in understanding the overall pharmacokinetics of PFCA precursors.

313 Table 9.2. Elimination half-lives (days) in blood and liver for polyfluorinated and perfluorinated carboxylates in rainbow trout a,b. Blood Liver Reference 7:3 FTCA Exposure 7:3 FTCA 5.1 (3.1-14) 10.3 (6.4-26.0) this study 8:2 FTCA Exposure 8:2 FTCA 1.2 (1.1-1.3) 1.3 (1.1-1.4) this study 7:3 FTCA 3.5 (2.9-4.3) 6.8 (4.4-15.0) this study PFOA n.s. c 4.1 (2.8-7.7) this study 8:2 FTUCA Exposure 8:2 FTUCA 0.39 (0.31-0.53) n/a this study PFOA 0.54 (0.3-6.1) n/a this study n-PFOA Exposure n-PFOA 5.6 (4.8-6.7) 3.7 (3.2-4.4) De Silva et al. n-PFOA n/a 5.2 ± 0.48 d Martin et al. n-PFNA Exposure n-PFNA 15.9 (13.5-19.6) 6.0 (5.2-7.1) De Silva et al. a Data cited from experiments in which the analyte was dosed via the diet. b Error represents 95% confidence intervals unless otherwise noted. c Elimination curve not statistically significant. d Error represents ± 1 standard error.

9.4.6. Environmental Implications 9.4.6.1. Formation of 7:3 FTCA The results from the present study showed that the 7:3 FTCA is formed from the biotransformation of the 8:2 FTCA and 8:2 FTUCA. Further, in separate dosing experiments it was empirically shown that the 8:2 FTCA is metabolized to the 8:2 FTUCA, and that the 8:2 FTUCA is not metabolized to the 8:2 FTCA. Therefore, the 7:3 FTCA appears to be formed from the direct biotransformation of the 8:2 FTUCA, which is formed via 8:2 FTCA biotransformation. These results are consistent with those of Nabb et al. [29] who observed the formation of the 7:3 FTCA and 7:3 FTUCA from hepatocyte incubations with both the 8:2 FTCA and 8:2 FTUCA. Nabb et al. [29] proposed the formation of the 7:3 FTCA from the 7:3 FTUCA, originating from the 8:2 FTCA (8:2 FTCA → 8:2 FTUCA → 7:3 FTUCA → 7:3 FTCA). In the hepatocyte incubations performed by Nabb et al. [29], 7:3 FTCA formation was observed when the 7:3 FTUCA was used as the substrate. However, the 7:3 FTUCA was also formed when the 7:3 FTCA was the substrate. Therefore, it appears that the 7:3 FTCA can both be the source and the metabolic product of the 7:3 FTUCA. We note that a similar reaction (i.e. the 8:2 FTUCA metabolically forming the 8:2 FTCA) is not feasible since this would involve the formation of a carbon-fluorine bond.

314 9.4.6.2. Formation of PFNA and PFOA The formation of PFNA was observed in the 8:2 FTCA exposure, but not during the 8:2 FTUCA or 7:3 FTCA exposure. These results are consistent with the suggested PFNA formation mechanism, through the direct α-oxidation of 8:2 FTCA [25, 26, 29]. Further, Kudo et al. [28] suggested that PFNA may be formed through metabolism of the 8:2 FTCA or an unidentified metabolite, but no mechanism was suggested. Several studies, investigating 8:2 FTOH exposure in mammalian systems, have shown that PFNA is formed in lower yields as compared to PFOA [25, 27-29]. In the present study, approximately equal concentrations of PFNA and PFOA were accumulated in blood samples of the 8:2 FTCA exposure. These results imply an overall lower yield of PFNA than PFOA from 8:2 FTCA biotransformation due to longer half-life of PFNA.

There is considerable variability in the literature regarding the hypothesized mechanisms of PFOA formation during 8:2 FTOH metabolism. Consistent with a recent study involving rainbow trout hepatic fractions [29], the present work empirically showed that the 7:3 FTCA is not a precursor to PFOA. Rather, it was shown that the PFHpA is formed from 7:3 FTCA biotransformation, a finding that has previously not been reported. The results from the present study demonstrate that PFOA is formed from the biotransformation of 8:2 FTCA and 8:2 FTUCA. Hagen et al. [19] originally postulated that β-oxidation was involved in the biotransformation of 8:2 FTOH in rats, while Dinglasan et al. [20] initially proposed the formation of PFOA from the β-oxidation of 8:2 FTCA in a microbial system. Further, Martin et al. [25] postulated PFOA formation from the “β-oxidation-like” metabolism of 8:2 FTUCA, which originates from the biotransformation of either the 8:2 FTCA or 8:2 FTUAL. This mechanism was supported by Nabb et al. [29], although the authors in fact proposed a unique mechanism for the predominant formation of PFOA. It was suggested by Wang et al. [21] that β-oxidation could not proceed through either the 8:2 FTCA or FTUCA since these compounds do not contain sufficient hydrogen atoms required to reduce NAD or FAD. However, the empirical observation of PFOA formation from the direct exposure of 8:2 FTCA and 8:2 FTUCA [25, 29] does suggest a β-like oxidation mechanism. As mentioned by Martin et al. [25], literature precedent does exist, specifically the dehydrofluorination of 2,2,- - - - - difluorosuccinate (CO2 CF2CH2CO2 ) to monofluorofumarate (CO2 CF=CHCO2 ) [41]. It was shown that this reaction is catalyzed by succinate dehydrogenase and is non-oxidative, thus does

315 not require the simultaneous reduction of FAD [42]. The second biotransformation step, the - - hydroxylation of monofluorofumarate to 2-fluoromalate (CO2 CF(OH)CH2CO2 ) is catalyzed by fumarate hydratase in the citric acid cycle [43], and is also NAD-independent. The 2- fluoromalate intermediate is unstable [44] and is rapidly dehydrofluorinated non-enzymatically - - to yield oxaloacetate (CO2 C(O)CH2CO2 ). In addition, the defluorination of methoxyflurane

(CHCl2CF2OCH3) has been shown in humans [45]. The ultimate step preceding the defluorination is the cleavage of the methyl ether (-OCH3) yielding the 2,2-dichloro-1,1- difluoroethanol (CHCl2CF2OH). Similar to 2-fluoromalate, this hydroxylated compound is chemically unstable and spontaneously decomposes, releasing HF, to the yield 2,2- dichloroacetyl fluoride (CHCl2C(O)F) which is hydrolyzed to the dichloroacetic acid

(CHCl2C(O)O2H). A similar series of reactions would occur for the dechlorinated molecule

(C(O)O2CF2OCH3). Therefore, as shown in the biotransformation of fluorinated succinate and methoxyflurane, hydroxylating the carbon containing fluorine will result in spontaneous non- enzymatic defluorination, resulting in either a carboxylic acid or ketone. The reactions described above are analogous to those in the β-oxidation mechanism and demonstrate the formation of a β-ketoacyl compound, from a 2,2-difluorinated carboxylic acid, that does not require NAD. As noted by Martin et al. [25], the β-ketoacyl analogue of 8:2 FTUCA would yield PFOA-S-CoA and acetyl CoA via thiolase.

As noted above, Nabb et al. [29] observed the formation of PFOA from incubations with the 8:2 FTCA and FTUCA, however, they postulated that the primary formation pathway of PFOA was from the 8:2 FTAL through the 8:2 FTUAL > 7:3 β-hydroxy unsaturated aldehyde > 7:3 β-keto aldehyde > PFOA. In a recent paper, Fasano et al. [35] proposed a combination of the Nabb et al. [29] pathway and a unique mechanism, analogous to the 8:2 FTUAL biotransformation previously proposed [29] but originating with the 8:2 FTUCA. In this mechanism, the 8:2 FTUCA is hydroxylated to yield the 7:3 β-hydroxyl carboxylate followed by oxidation to the 7:3 β-keto carboxylate which is then decarboxylated to form the 7:2 ketone. The 7:2 ketone is subsequently reduced to the 7:2 sFTOH which then forms PFOA. Interestingly, Nabb et al. [29] also proposed the formation of the 7:2 sFTOH from the 8:2 FTUCA but the formation of PFOA from the 7:2 sFTOH was not suggested. The formation of the 7:3 β-keto aldehyde and 7:3 β-keto carboxylate from the 8:2 FTUAL and 8:2 FTUCA are essentially analogous to the “β-like-oxidation” biotransformation mechanisms described for the

316 fluorinated succinate and methoxyflurane and thus appear reasonable. However, formation of the β-OH intermediates seems unlikely given the instability of the fluorohydroxy compound. Beta-keto acids are susceptible to decarboxylation, forming the methyl ketone. Thus, the decarboxylation of 7:3 β-keto acid to 7:2 ketone via decarboxylase is possible. It has been shown that methyl ketones undergo α-hydroxylation and subsequent oxidation to yield ketocarboxylic acids [46], ultimately forming aliphatic carboxylic acid through oxidative decarboxylation. Therefore, the formation of PFOA from the biotransformation of 7:2 ketone appears to be reasonable. Aldehyde reductase enzymes, capable of metabolizing xenobiotic compounds, are found in mammals [47] and fish [48]. Therefore, literature precedent exists for the reduction of the 7:2 ketone to the 7:2 sFTOH. However, we could not find any precedent for the formation of a carboxylic acid from a secondary alcohol and thus formation of PFOA from the 7:2 sFTOH, as postulated by Fasano et al. [35], does not appear likely. It is likely that the predominate biological fate of the 7:2 sFTOH, like the 8:2 FTOH, is glucuronidation. Regarding the 7:3 β-keto aldehyde, we could not find any literature precedent for the formation of carboxylic acids from β-keto aldehydes. Rather, we propose that this compound is oxidized to the 7:3 β-keto acid > 7:2 ketone > PFOA. Based on the findings of this study and reported analogous literature mechanisms, we expand on the “beta-like-oxidation” mechanism [25] for PFOA formation as proceeding from the 8:2 FTUCA > 7:3 β-keto acid > 7:2 ketone > PFOA (Figure 9.4).

The overall formation and accumulation of PFOA was very low, consistent with other studies examining 8:2 FTOH biotransformation [29]. To compare the PFOA yield between the 8:2 FTCA and 8:2 FTUCA doses, we calculated the formation efficiency, defined as the PFOA concentration in the tissues at the end of the exposure phase divided by the parent concentration in the blood (i.e. FE (%) = tissue PFOA * 100/ food8:2 FTCA or 8:2 FTUCA.). The FE for the 8:2 FTCA dosing was 0.009% and 0.054% in the blood and liver, respectively. The FE values for the 8:2 FTUCA were similar at 0.014% and 0.022% in the blood and liver, respectively. The FE values do not represent actual biotransformation yield, rather they are a function of yield and pharmacokinetics. Therefore, higher FE values are expected for the biotransformation of longer fluorotelomer compounds which would yield longer-chain PFCAs with comparability slower elimination half-lives.

317 The current study has built upon previous metabolic investigations of telomer-based polyfluorinated compounds and has provided additional insights into the biotransformation mechanism. Although this study specifically examined the 8:2 fluorotelomer-based compounds, it is expected that these results are applicable for all telomer-based compounds, of various chain lengths, which degrade to the FTOH, such as polyfluorinated phosphates [10] and polyfluorinated acrylates.

9.5. Acknowledgements Norman White and staff at the Aquatic Facility in the Department of Cell & Systems Biology (University of Toronto) are thanked for fish care and husbandry. Clara Chan, Helen Sun and Alex Tevlin provided assistance with sample preparation. We are grateful to Wellington Laboratories for donation of mass-labeled standards. Project funding was provided by the Natural Sciences & Engineering Research Council of Canada (NSERC) (Mabury) and Environment Canada’s Chemical Management Plan (Muir). C.M.B. also appreciates the support of NSERC through a Post-Graduate Scholarship.

FTOH Precursors F F F F F F F F COOH O O F OH F F F F F F F F OH F F F F F F F F OH F 8:2 FTOH-Gluc OH F F F F F F F F

F F F F F F F O 8:2 FTOH F 8:2 FTUAL-GSH O S O- F O F F F F F F F F F F F F F F F F 8:2 FTOH-Sulf F F F F F F F F F O O F F F F F F F F F F F F F F F F 8:2 FTAL 8:2 FTUAL

O F F F F F F F F F F F F F F F F F F F F F F F O F - F O- C O F F O F F F F F F F F F F F F F O F F F F F F F F F 8:2 FTCA PFNA 7:3 β-keto aldehyde

F F F F F F F F O F F F F F F F O- F F F F F F O O F F β-oxidation F - O F F F F F F S O - F O O F F F F F F F F F F F F F F H2N NH COOH F N H 7:3 β-keto acid 8:2 FTUCA COOH O

8:2 FTUCA-GSH F F F F F F O F F F F F F O F F O- F F F F F F F F F F F F F F F F 7:2 ketone 7:3 FTCA

O F F F F F F OH F F F F F F F F F F F F O F F F F F F F F C O- F F - - F O C O F F F F F F F F F F F F F F F F F F F F F F F F F F F F F O 7:2 sFTOH PFOA 7:3 FTUCA PFHpA

Figure 9.4. Proposed biotransformation mechanism for the 8:2 FTOH. 318

319 9.6. Literature Cited 1. Giesy, J. P.; Kannan, K., Distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 2001, 35, 1339-1342.

2. Martin, J. W.; Smithwick, M. M.; Braune, B. M.; Hoekstra, P. F.; Muir, D. C. G.; Mabury, S. A., Identification of long-chain perfluorinated acids in biota from the Canadian Arctic. Environ. Sci. Technol. 2004, 38, 373-380.

3. Hansen, K. J.; Clemen, L. A.; Ellefson, M. E.; Johnson, H. O., Compound-specific, quantitative characterization of organic fluorochemicals in biological matrices. Environ. Sci. Technol. 2001, 35, 766-770.

4. Olsen, G. W.; Mair, D. C.; Church, T. R.; Ellefson, M. E.; Reagen, W. K.; Boyd, T. M.; Herron, R. M.; Medhdizadehkashi, Z.; Nobiletti, J. B.; Rios, J. A.; Butenhoff, J. L.; Zobel, L. R., Decline in perfluorooctanesulfonate and other polyfluoroalkyl chemicals in American red cross adult blood donors, 2000-2006. Environ. Sci. Technol. 2008, 42, 4989-4995.

5. Kannan, K.; Corsolini, S.; Falandysz, J.; Fillmann, G.; Kumar, K. S.; Loganathan, B. G.; Mohd, M. A.; Olivero, J.; van Wouwe, N.; Yang, J. H.; Aldous, K. M., Perfluorooctanesulfonate and related fluorochemicals in human blood from several countries. Environ. Sci. Technol. 2004, 38, 4489-4495.

6. Tittlemier, S. A.; Pepper, K.; Seymour, C.; Moisey, J.; Bronson, R.; Cao, X.-L.; Dabeka, R. W., Dietary exposure of Canadians to perfluorinated carboxylates and perfluorooctane sulfonate via consumption of meat, fish, fast foods, and food items prepared in their packaging. J. Agric. Food Chem. 2007, 55, 3203-3210.

7. Fromme, H.; Tittlemier, S. A.; Völkel, W.; Wilhelm, M.; Twardella, D., Perfluorinated compounds - exposure assessment for the general population in western countries. Int. J. Hyg. Environ. Health 2009, 212, 239-270.

8. Vestergren, R.; Cousins, I. T., Tracking the pathways of human exposure to perfluorocarboxylates. Environ. Sci. Technol. 2009, 43, 5565-5575.

9. D'eon, J. C.; Crozier, P. W.; Furdui, V. I.; Reiner, E. J.; Libelo, E. L.; Mabury, S. A., Observation of a commercial fluorinated material, the polyfluoroalkyl phosphoric acid diesters, in human sera, wastewater treatment plant sludge, and paper fibers. Environ. Sci. Technol. 2009, 43, 4589-4594.

10. D'eon, J. C.; Mabury, S. A., Production of perfluorinated carboxylic acids (PFCAs) from the biotransformation of polyfluoroalkyl phosphate surfactants (PAPs): Exploring routes of human contamination. Environ. Sci. Technol. 2007, 41, 4799-4805.

11. Ellis, D. A.; Martin, J. W.; De Silva, A. O.; Mabury, S. A.; Hurley, M. D.; Sulbaek Andersen, M. P.; Wallington, T. J., Degradation of fluorotelomer alcohols: A likely atmospheric source of perfluorinated carboxylic acids. Environ. Sci. Technol. 2004, 38, 3316-3321.

320 12. Hurley, M. D.; Wallington, T. J.; Sulbaek Andersen, M. P.; Ellis, D. A.; Martin, J. W.; Mabury, S. A., Atmospheric chemistry of fluorinated alcohols: reaction with Cl atoms and OH radicals and atmospheric lifetimes. J. Phys. Chem. A 2004, 108, 1973-1979.

13. Vésine, E.; Bossoutrot, V.; Mellouki, A.; Le Bras, G.; Wenger, J.; Sidebottom, H., Kinetic and mechanistic study of OH- and Cl-initiated oxidation of two unsaturated HFCs: C4F9CH=CH2 and C6F13CH=CH2. J. Phys. Chem. A 2000, 104, 8512-8520.

14. Nakayama, T.; Takahashi, K.; Matsumi, Y.; Toft, A.; Sulbaek Andersen, M. P.; Nielsen, O. J.; Waterland, R. L.; Buck, R. C.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of CF3CH=CH2 and C4F9CH=CH2: Products of the gas-phase reactions with Cl atoms and OH radicals. J. Phys. Chem. A 2007, 111, 909-915.

15. Young, C. J.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A., Atmospheric chemistry of 4:2 fluorotelomer iodide (n-C4F9CH2CH2I) and 6:2 fluorotelomer iodide (n-C6F13CH2CH2I): Kinetics and products of photolysis and reaction with OH radicals and Cl atoms. J. Phys. Chem. A 2008, 112, 13542-13548.

16. Butt, C. M.; Young, C. J.; Mabury, S. A.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of 4:2 fluorotelomer acrylate [C4F9CH2CH2OC(O)CH=CH2]: Kinetics, mechanisms, and products of chlorine-atom- and OH-radical-initiated oxidation. J. Phys. Chem. A 2009, 113, 3155-3161.

17. Martin, J. W.; Ellis, D. A.; Mabury, S. A.; Hurley, M. D.; Wallington, T. J., Atmospheric chemistry of perfluoroalkanesulfonamides: Kinetic and product studies of the OH radical and Cl atom initiated oxidation of N-ethyl perfluorobutanesulfonamide. Environ. Sci. Technol. 2006, 40, 846-872.

18. D'eon, J. C.; Hurley, M. D.; Wallington, T. J.; Mabury, S. A., Atmospheric chemistry of N-methyl perfluorobutane sulfonamidoethanol, C4F9SO2N(CH3)CH2CH2OH: Kinetics and mechanism of reaction with OH. Environ. Sci. Technol. 2006, 40, 1862-1868.

19. Hagen, D. F.; Belisle, J.; Johnson, J. D.; Venkateswarlu, P., Characterization of fluorinated metabolites by a gas chromatographic-helium microwave plasma detector - the biotransformation of 1H,1H,2H,2H-perfluorodecanol to perfluorooctanoate. Anal. Biochem. 1981, 118, 336-343.

20. Dinglasan, M. J. A.; Ye, Y.; Edwards, E. A.; Mabury, S. A., Fluorotelomer alcohol biodegradation yields poly- and perfluorinated acids. Environ. Sci. Technol. 2004, 38, 2857- 2864.

21. Wang, N.; Szostek, B.; Folsom, P. W.; Sulecki, L. M.; Capka, V.; Buck, R. C.; Berti, W. R.; Gannon, J. T., Aerobic biotransformation of 14C-labeled 8-2 telomer B alcohol by activated sludge from a domestic sewage treatment plant. Environ. Sci. Technol. 2005, 39, 531-538.

22. Wang, N.; Szostek, B.; Buck, R. C.; Folsom, P. W.; Sulecki, L. M.; Capka, V.; Berti, W. R.; Gannon, J. T., Fluorotelomer alcohol biodegradation - direct evidence that perfluorinated carbon chains breakdown. Environ. Sci. Technol. 2005, 39, 7516-7528.

321 23. Liu, J.; Lee, L. S.; Nies, L. F.; Nakatsu, C. H.; Turco, R. F., Biotransformation of 8:2 fluorotelomer alcohol in soil and by soil bacteria isolates. Environ. Sci. Technol. 2007, 41, 8024- 8030.

24. Wang, N.; Szostek, B.; Buck, R. C.; Folsom, P. W.; Sulecki, L. M.; Gannon, J. T., 8-2 Fluorotelomer alcohol aerobic soil biodegradation: Pathways, metabolites, and metabolite yields. Chemosphere 2009, 75, 1089-1096.

25. Martin, J. W.; Mabury, S. A.; O'Brien, P. J., Metabolic products and pathways of fluorotelomer alcohols in isolated rat hepatocytes. Chem. Biol. Interact. 2005, 155, 165-180.

26. Fasano, W. J.; Carpenter, S. C.; Gannon, S. A.; Snow, T. A.; Stadler, J. C.; Kennedy, G. L.; Buck, R. C.; Korzeniowski, S. H.; Hinderliter, P. M.; Kemper, R. A., Absorption, distribution, metabolism, and elimination of 8-2 fluorotelomer alcohol in the rat. Toxicol. Sci. 2006, 91, 341-355.

27. Henderson, W. M.; Smith, M. A., Perfluorooctanoic acid and perfluorononanoic acid in fetal and neonatal mice following in utero exposure to 8-2 fluorotelomer alcohol. Toxicol. Sci. 2007, 95, 452-461.

28. Kudo, N.; Iwase, Y.; Okayachi, H.; Yamakawa, Y.; Kawashima, Y., Induction of hepatic peroxisome proliferation by 8-2 telomer alcohol feeding in mice: Formation of perfluorooctanoic acid in the liver. Toxicol. Sci. 2005, 86, 231-238.

29. Nabb, D. L.; Szostek, B.; Himmelstein, M. W.; Mawn, M. P.; Gargas, M. L.; Sweeney, L. M.; Stadler, J. C.; Buck, R. C.; Fasano, W. J., In vitro metabolism of 8-2 fluorotelomer alcohol: Interspecies comparisons and metabolic pathway refinement. Toxicol. Sci. 2007, 100, 333-344.

30. Brandsma, S. H.; Smithwick, M.; Solomon, K.; Small, J.; Muir, D. C. G., Uptake, transformation and elimination of 8:2 and 10:2 fluorotelomer alcohols and PFOSA by rainbow trout from feed. Manuscript in prep.

31. Butt, C. M.; Muir, D. C. G.; Mabury, S. A., Bioaccumulation and biotransformation of the 8:2 FTOH acrylate in rainbow trout. in preparation.

32. Achilefu, S.; Mansuy, L.; Selve, C.; Thiebaut, S., Synthesis of 2H,2H-perfluoroalkyl and 2H-perfluoroalkenyl carboxylic acid and amines. J. Fluor. Chem. 1995, 70, 19-26.

33. Butt, C. M.; Muir, D. C. G.; Stirling, I.; Kwan, M.; Mabury, S. A., Rapid response of Arctic ringed seals to changes in perfluoroalkyl production. Environ. Sci. Technol. 2007, 41, 42- 49.

34. Fisk, A. T.; Norstrom, R. J.; Cymbalisty, C. D.; Muir, D. C. G., Dietary accumulation and depuration of hydrophobic organochlorines: Bioaccumulation parameters and their relationship with the octanol/water partition coefficient. Environ. Toxicol. Chem. 1998, 17, 951- 961.

35. Fasano, W. J.; Sweeney, L. M.; Mawn, M. P.; Nabb, D. L.; Szostek, B.; Buck, R. C.; Gargas, M. L., Kinetics of 8-2 fluorotelomer alcohol and its metabolites, and liver glutathione

322 status following daily oral dosing for 45 days in male and female rats. Chem. Biol. Interact. 2009, 180, 281-295.

36. Powley, C. R.; George, S. W.; Russell, M. H.; Hoke, R. A.; Buck, R. C., Polyfluorinated chemicals in a spatially and temporally integrated food web in the Western Arctic. Chemosphere 2008, 70, 664-672.

37. Brown, R. P.; Delp, M. D.; Lindstedt, S. L.; Rhomberg, L. R.; Beliles, R. P., Physiological parameter values for physiologically based pharmacokinetic models. Toxicol. Ind. Health 1997, 13, 407-484.

38. Martin, J. W.; Mabury, S. A.; Solomon, K. R.; Muir, D. C. G., Dietary accumulation of perfluorinated acids in juvenile rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem. 2003, 22, 189-195.

39. De Silva, A. O.; Tseng, P. J.; Mabury, S. A., Toxicokinetics of perfluorocarboxylate isomers in rainbow trout. Environ. Toxicol. Chem. 2009, 28, 330-337.

40. Martin, J. W.; Chan, K.; Mabury, S. A.; O'Brien, P. J., Bioactivation of fluorotelomer alchohols in isolated rat hepatocytes. Chem. Biol. Interact. 2009, 177, 196-203.

41. Brodie, J. D.; Nicholls, P., Metabolsim and enzymology of fluorosuccinic acids. I. Interactions with the succinate oxidase system. Biochim. Biophys. Acta 1970, 198, (423-437).

42. Tober, C. L.; Nicholls, P.; Brodie, J. D., Metabolism and enzymology of fluorosuccinic acids. II. Substrate and inhibitor effects with soluble succinate dehydrogenase. Arch. Biochem. Biophys. 1970, 138, 506-514.

43. Clarke, D. D.; Nicklas, W. J.; Palumbo, J., Fluorofumarate - A substrate for fumarate hydratase. Arch. Biochem. Biophys. 1968, 123, 205-206.

44. Walsh, C., Fluorinated substrate analogs: Routes of metabolism and selective toxicity. In Advances in enzymology, Meister, A., Ed. John Wiley & Sons: Toronto, 1983; Vol. 55, pp 197- 285.

45. Mazze, R. I.; Trudell, J. R.; Cousins, M. J., Methoxyflurane metabolism and renal dysfunction: Clinical correlation in man. Anesthesiology 1971, 35, 247-252.

46. Gabriel, M. A.; Ilbawi, M.; Al-Khalidi, U. A. S., The oxidation of acetoin to CO2 in intact animals and in liver mince preparation. Comp. Biochem. Physiol. 1972, 41B, 493-502.

47. Bachur, N. R., Cytoplasmic aldo keto reductases: a class of drug metabolizing enzymes. Science 1976, 193, 4253.

48. Tulayakul, P.; Sakuda, s.; Dong, K. S.; Kumagai, S., Comparative activities of glutathione-S-transferase and dialdehyde reductase toward aflatoxin B1 in livers of experimental and farm animals. Toxicon 2005, 46, 204-209.

9.7. Supporting Information

Table S9.1. Mean ± Standard Error (n=3) concentrations (ng/g ww) in blood and liver from 7:3 FTCA dietary exposure in rainbow trout.

Blood PFHxA PFHpA PFOA PFNA 8:2FTUCA 7:3FTUCA 8:2FTCA 7:3 FTCA 12 hr nd <0.03 <0.77 <2.8 <0.27 <0.01 nd 660 ± 300 24 hr nd 0.66 ± 0.40 <0.77 <2.8 <0.27 <0.01 nd 1090 ± 120 48 hr nd 0.42 ± 0.18 <0.77 <2.8 <0.27 <0.01 nd 1440 ± 110 72 hr nd 1.0 ± 0.22 <0.77 <2.8 <0.27 0.19 ± 0.05 nd 1810 ± 160 120 hr nd 1.8 ± 0.68 <0.77 <2.8 <0.27 0.39 ± 0.12 nd 2190 ± 240 168 hr nd 3.0 ± 1.3 <0.77 <2.8 <0.27 0.43 ± 0.13 nd 2130 ± 50 24 hr (dep) nd 0.66 ± 0.32 <0.77 <2.8 <0.27 0.15 ± 0.05 nd 570 ± 230 72 hr (dep) nd <0.03 <0.77 <2.8 <0.27 0.11 ± 0.08 nd 480 ± 34 120 hr (dep) nd <0.03 <0.77 <2.8 <0.27 <0.01 nd 630 ± 21 168 hr (dep) nd <0.03 <0.77 <2.8 <0.27 <0.01 nd 370 ± 38 240 hr (dep) nd <0.03 <0.77 <2.8 <0.27 <0.01 nd 280 ± 33

Liver PFHxA PFHpA PFOA PFNA 8:2FTUCA 7:3FTUCA 8:2FTCA 7:3 FTCA 12 hr nd 1.2 ± 0.35 <3.9 <4.6 <13.2 <0.16 nd 1060 ± 480 24 hr nd 1.8 ± 0.19 <3.9 <4.6 <13.2 <0.16 nd 1850 ± 230 48 hr nd 0.83 ± 0.33 <3.9 <4.6 <13.2 <0.16 nd 1310 ± 52 72 hr nd 1.5 ± 0.56 <3.9 <4.6 <13.2 <0.16 nd 1200 ± 98 120 hr nd 2.0 ± 0.35 <3.9 <4.6 <13.2 0.11 ± 0.08 nd 1750 ± 80 168 hr nd 2.4 ± 0.11 <3.9 <4.6 <13.2 0.1 ± 0.5 nd 1310 ± 92 24 hr (dep) nd <0.77 <3.9 <4.6 <13.2 <0.16 nd 1320 ± 4.2 72 hr (dep) nd <0.77 <3.9 <4.6 <13.2 <0.16 nd 890 ± 210 120 hr (dep) nd <0.77 <3.9 <4.6 <13.2 <0.16 nd 1130 ± 130 168 hr (dep) nd <0.77 <3.9 <4.6 <13.2 <0.16 nd 1040 ± 210 240 hr (dep) nd <0.77 <3.9 <4.6 <13.2 <0.16 nd 560 ± 85

323

Table S9.2. Mean ± Standard Error (n=3) concentrations (ng/g ww) in blood and liver from 8:2 FTCA dietary exposure in rainbow trout.

Blood PFHxA PFHpA PFOA PFNA 8:2FTUCA 7:3FTUCA 8:2FTCA 7:3 FTCA 12 hr nd 0.57 ± 0.23 0.94 ± 2.3 <2.8 56 ± 15 <0.01 3200 ± 1150 66 ± 5.1 24 hr nd nd 1.3 ± 1.9 3.6 28 ± 16 0.03 ± 0.33 1250 ± 830 91 ± 57 48 hr nd 0.62 ± 0.21 2.8 ± 0.9 <2.8 94 ± 21 0.89 ± 0.11 5780 ± 990 490 ± 44 72 hr nd 0.74 ± 0.42 2.9 ± 1.3 <2.8 72 ± 27 0.78 ± 0.36 4160 ± 1640 350 ± 160 120 hr nd 0.39 ± 0.26 4.6 ± 2.0 <2.8 71 ± 30 1.2 ± 0.40 4560 ± 1900 690 ± 270 168 hr nd 1.1 ± 0.31 8.8 ± 0.7 5.2 ± 0.7 120 ± 4.1 2.8 ± 0.63 7390 ± 540 940 ± 140 24 hr (dep) nd 0.02 9.5 ± 3.7 5.0 ± 1.2 54 ± 14 2.5 ± 0.57 3290 ± 940 840 ± 170 72 hr (dep) nd <0.03 7.9 ± 1.6 4.8 ± 1.7 36 ± 3.7 2.0 ± 0.31 1780 ± 220 510 ± 22 120 hr (dep) nd 1.6 ± 1.6 13.8 ± 6.8 6.3 ± 1.9 2.9 ± 2.9 1.2 ± 0.19 350 ± 31 350 ± 43 168 hr (dep) nd <0.03 5.1 ± 0.9 5.1 ± 0.7 1.8 ± 0.56 0.62 ± 0.12 97 ± 30 230 ± 30 240 hr (dep) nd nd 2.4 ± 2.0 3.9 0.65 ± 0.12 0.16 ± 0.08 25 ± 1.9 140 ± 32

Liver PFHxA PFHpA PFOA PFNA 8:2FTUCA 7:3FTUCA 8:2FTCA 7:3 FTCA 12 hr nd <0.77 23 ± 6.9 <4.6 120 ± 32 <0.16 1550 ± 380 105 ± 17 24 hr nd <0.77 33 ± 10 <4.6 77 ± 27 0.19 ± 0.04 750 ± 370 150 ± 77 48 hr nd <0.77 23 ± 1.7 <4.6 200 ± 47 0.80 ± 0.16 2500 ± 520 330 ± 69 72 hr nd <0.77 36 ± 12 <4.6 260 ± 89 1.4 ± 0.48 2030 ± 1200 420 ± 240 120 hr nd <0.77 40 ± 12 <4.6 150 ± 59 1.0 ± 0.37 2060 ± 800 510 ± 110 168 hr nd 1.3 ± 0.08 56 ± 15 <4.6 300 ± 43 2.5 ± 0.47 3550 ± 490 870 ± 150 24 hr (dep) nd <0.77 49 ± 12 <4.6 130 ± 24 2.0 ± 0.45 1410 ± 130 1080 ± 150 72 hr (dep) nd <0.77 31 ± 11 7.6 ± 2.3 98 ± 44 1.8 ± 0.77 930 ± 420 830 ± 170 120 hr (dep) nd <0.77 24 ± 2,2 <4.6 <13.2 0.69 ± 0.13 160 ± 17 980 ± 300 168 hr (dep) nd <0.77 16 ± 1.1 13 ± 3.8 <13.2 0.38 ± 0.13 56 ± 14 670 ± 59 240 hr (dep) nd <0.77 8.4 <4.6 <13.2 0.12 ± 0.07 15 ± 2.8 310 ± 68

324

Table S9.3. Mean ± Standard Error (n=3) concentrations (ng/g ww) in blood and liver from 8:2 FTUCA dietary exposure in rainbow trout.

Blood PFHxA PFHpA PFOA PFNA 8:2FTUCA 7:3FTUCA 8:2FTCA 7:3 FTCA 12 hr nd <0.13 <5.4 <1.7 1030 ± 460 0.10 ± 0.05 36 ± 13 13 ± 4.8 24 hr nd <0.13 <5.4 <1.7 1300 ± 180 0.40 ± 0.11 62 ± 13 59 ± 8.8 48 hr nd 1.5 ± 1.4 5.5 ± 3.4 1.1 ± 0.37 1070 ± 210 0.69 ± 0.17 63 ± 6.9 94 ± 16 72 hr nd 0.13 ± 0.06 7.5 ± 0.33 <1.7 1930 ± 630 1.5 ± 0.29 120 ± 17 130 ± 11 120 hr nd <0.13 9.3 ± 1.6 <1.7 1660 ± 760 1.6 ± 0.63 100 ± 39 160 ± 60 168 hr nd <0.13 10 ± 0.48 <1.7 1160 ± 120 1.3 ± 0.10 89 ± 6.0 140 ± 5.8 24 hr (dep) nd <0.13 3.8 ± 2.4 <1.7 104 ± 27 0.59 ± 0.13 38 ± 8.4 58 ± 12 48 hr (dep) nd <0.13 2.2 ± 1.9 <1.7 27 ± 13 0.60 ± 0.09 19 ± 0.6 72 ± 13 72 hr (dep) nd <0.13 <5.4 <1.7 1.8 ± 1.4 0.20 ± 0.04 5.6 ± 1.8 32 ± 5.0 120 hr (dep) nd <0.13 <5.4 <1.7 0.58 ± 0.42 0.07 ± 0.03 <4.6 26 ± 6.0 168 hr (dep) nd <0.13 <5.4 <1.7 0.66 ± 0.09 0.06 ± 0.03 <4.6 17 ± 4.4 240 hr (dep) nd <0.13 <5.4 <1.7 <0.20 <0.03 <4.6 22 ± 3.8

Liver PFHxA PFHpA PFOA PFNA 8:2FTUCA 7:3FTUCA 8:2FTCA 7:3 FTCA 12 hr nd <0.74 <9.6 <5.1 21 ± 16 0.12 ± 0.07 22 ± 3.1 59 ± 18 24 hr nd <0.74 4.9 ± 2.7 <5.1 19 ± 7.9 0.28 ± 0.14 24 ± 2.9 150 ± 24 48 hr nd 0.87 ± 0.43 7.5 ± 4.9 <5.1 15 ± 4.8 0.49 ± 0.11 24 ± 2.0 220 ± 41 72 hr nd 1.7 ± 0.53 11 ± 3.4 <5.1 15 ± 6.4 0.86 ± 0.14 29 ± 0.89 240 ± 25 120 hr nd 1.2 ± 0.16 16 ± 6.1 <5.1 45 ± 35 0.73 ± 0.15 28 ± 10 270 ± 76 168 hr nd 0.95 ± 0.21 5.6 ± 3.1 <5.1 5.1 ± 0.37 0.71 ± 0.05 49 ± 2.6 330 ± 30 24 hr (dep) nd 0.37 ± 0.24 <9.6 <5.1 <1.2 0.39 ± 0.08 20 ± 2.6 160 ± 6.9 48 hr (dep) nd <0.74 <9.6 <5.1 <1.2 0.26 ± 0.12 24 ± 3.1 170 ± 42 72 hr (dep) nd <0.74 <9.6 <5.1 <1.2 <0.20 18 ± 6.2 93 ± 22 120 hr (dep) nd <0.74 <9.6 <5.1 <1.2 <0.20 nd 68 ± 16 168 hr (dep) nd <0.74 <9.6 <5.1 <1.2 <0.20 nd 25 ± 3.9 240 hr (dep) nd <0.74 <9.6 <5.1 <1.2 <0.20 nd 31 ± 5.3

325

CHAPTER TEN

Conclusions and Future Directions

Contributions: Prepared by Craig Butt with critical comments provided by Scott Mabury and Derek Muir

326 327 Chapter Ten – Conclusions and Future Directions

This thesis broadly investigated the sources of perfluorinated alkyl compounds (PFCs) to biota. A specific focus was the role of the fluorotelomer acrylate monomer (FTAc) as an indirect source of PFCAs through atmospheric and biological reactions. In addition, three monitoring studies of arctic wildlife were performed to provide further insight into the transport pathways and dynamics of PFCs to remote environments.

While this thesis focused on poly- and perfluorinated alkyl compounds, there are broader applications to environmental chemistry. In recent years there has been increased interest in the fate of metabolically-labile compounds in fish. The biotransformation studies of Chapters 8 and 9 illustrate examples where the metabolic products are far more bioaccumulative than the parent compound. In addition, recent research has investigated the use of in vitro biotransformation data to quantify whole body transformation. The use of such data could be used as an alternative to whole body bioaccumulation studies. The sub-cellular tissue incubations of Chapter 8, in combination with the in vivo results, provide a useful data set to investigate these models.

Chapter 4 investigated the potential of the FTAc to act as a PFCA precursor through atmospheric degradation. It was shown that the lifetime of the FTAc, towards reaction with hydroxyl radical, was approximately 1 day. These results indicate that the FTAc itself will primarily be a local pollutant. However, it was shown that the fluorotelomer aldehyde (FTAL) was directly formed during reaction with chlorine radical. The FTAL and their degradation products have lifetimes that are sufficient to permit long-range transport. Further, formation of the fluorotelomer glyoxylate was tentatively identified through reaction of the FTAc with hydroxyl radical. The fluorotelomer glyoxylate will degrade through either photolysis or hydroxyl radical reaction to yield the FTAL. The FTAL has been shown to be a precursor to PFCAs. Therefore, this work demonstrated that FTAcs are a potential source of PFCAs to the environment, including remote regions. Although the fluorotelomer glyoxylate was tentatively identified during the experiments, its presence could not be confirmed since an authentic standard was not available. Further research is needed to quantify atmospheric degradation rates (i.e. photolysis and reaction with chlorine atoms and hydroxyl radical) as well as to confirm products.

328

Chapter 5 investigated the temporal trends of PFCs in ringed seals from two locations in the Canadian Arctic. It was shown that concentrations of perfluorooctane sulfonate (PFOS) reached maximum concentrations during 1998 and 2000 in the Arviat and Resolute Bay populations, respectively, followed by statistically significant decreases to 2005. These trends were consistent with 3M’s phase-out of perfluorooctane sulfonyl fluoride (POSF) chemistry in 2002. In contrast, levels of the PFCAs mainly showed increasing or steady concentrations in recent years. The results of this work lend support to the hypothesis of atmospheric transport and degradation of precursors as an important transport mechanism to the arctic. Additional monitoring is needed to confirm the trends observed in this study. There is evidence that climate change is occurring in the arctic with the potential to significant disturb ecosystems and contaminant dynamics. The additional years of monitoring will aid in our understanding of this potential impact. Also, this work postulated that the annual melting of the sea ice strongly influences the PFC dynamics in the arctic marine ecosystem. The potential relevance of this process needs to be explored further. For example, depth samples should be collected to investigate the stratification of PFCs in the seawater. In addition, novel foodweb models are needed, specific to PFC biomagnification, that account for this mechanism.

The PFC temporal trends in seabirds from the Canadian Arctic were investigated in Chapter 6. It was shown that the temporal resolution was too coarse to discern a potential response to 3M’s POSF-chemistry phase out; similar to what was performed in Chapter 5. Additional monitoring, with a much shorter temporal resolution, is recommended. In addition, it was shown that the PFC profiles in the seabirds were dominated by the C11-C15 PFCAs. This pattern is in contrast to other wildlife which shows a dominance of the C9 or C11 PFCAs. Additional monitoring of other seabird species is suggested to confirm if the observed patterns are common among seabirds. Further, the mechanism responsible for this pattern should be further explored. This may include feeding studies with captive birds to determine if the PFCA pharmacokinetics differs between seabirds and other arctic species.

The PFC spatial trends in 11 populations of ringed seals across the Canadian Arctic were investigated in Chapter 7. Seal samples were also analyzed for stable isotopes of nitrogen and carbon to assess relative trophic position and carbon sources. The overall mean PFC levels were similar between the populations but differences were observed. Most notably were the

329 comparatively elevated concentrations of some PFCs in the Gjoa Haven population (Rae Strait, central Canadian Arctic archipelago). The δ15N levels were relatively similar between the populations, indicating a similar trophic position. Interestingly, the δ13C ratios showed significantly depleted δ13C levels in seals from Gjoa Haven compared with the other populations and may suggest a terrestrial carbon source. The relevance of this apparent enhancement of terrestrial carbon is unclear. The hypothesis that terrestrial runoff may explain the elevated PFC levels in Gjoa Haven seals could be tested by measuring the mass flow of PFCs from the Back River into the Rae Strait.

The bioaccumulation and biotransformation of the 8:2 FTAc were explored through two complementary experiments in Chapter 8. The in vivo dietary study showed that the 8:2 FTAc was rapidly biotransformed in the fish with intermediate metabolites detected within 1 hr post- dosing. The metabolites observed were consistent with those previously reported for 8:2 FTOH biotransformation. Interestingly, levels of PFOA and the 7:3 FTCA did not show declines until approximately 4 days after depuration started. These results suggest that the PFOA and 7:3 FTCA were formed from precursors that still remained in the body. In addition, very low levels of the 8:2 FTAc were detected in the liver and kidney only. As a result, the tissue levels of the metabolites were more than two orders of magnitude greater than the parent compound. It was postulated that these trends were due to high activities of esterase enzymes in the fish. To address this hypothesis, in vitro sub-cellular incubations were performed to obtain enzyme kinetics. As compared to literature reports of other phase I and phase II enzymes, very high esterase activities were shown with apparently equal efficiency in the stomach and liver. Recently there have been efforts to extrapolate in vitro biotransformation data to predict whole body bioaccumulation. With few exceptions, present models only consider hepatic metabolism. Future research should develop pharmacokinetic models that incorporate in vivo gut metabolism data. The results from the current study provide an excellent data set for such models.

The biotransformation of 8:2 FTOH has been extensively studied in microbes and mammals, with little attention to fish species. All proposed mechanisms show the formation of PFOA, however, there is considerable variability in the actual details of the biotransformation scheme. Therefore, Chapter 9 described in vivo dietary dosing experiments that were performed with the objective to further elucidate the 8:2 FTOH biotransformation mechanism. Fish were

330 dosed with food that was separately spiked with the 8:2 FTCA, 8:2 FTUCA and 7:3 FTCA. These compounds represented important branching points in the 8:2 biotransformation mechanism. The 7:3 FTCA did not form PFOA, but instead formed the 7:3 FTUCA and perfluoroheptanoate (PFHpA), a novel finding. Both the 8:2 FTCA and 8:2 FTUCA formed PFOA which confirmed a “beta-like-oxidation” mechanism. Based on the results of the study, a biotransformation scheme was proposed, refining existing hypothesized pathways, with PFOA formation proceeding from the 8:2 FTUCA > 7:3 β-keto acid > 7:2 ketone > PFOA. Other research groups have postulated that the 7:2 secondary FTOH (7:2 sFTOH) is a direct precursor to PFOA formation. Based on current knowledge of enzyme reactions, this mechanism does not seem reasonable. However, this hypothesis could be tested using a similar experimental system as in the current study, but using the 7:2 sFTOH as the substrate. In addition, the half-lives of the parent compounds investigated show significantly different half-lives. These trends could be due to either differences in depuration or metabolism rates, or a combination of both. However, the present study was not designed to differentiate between these processes. The metabolism rates could be explored using in vitro sub-cellular incubation experiments, similar to that performed in Chapter 8.