Using dietary strategies to explore mechanisms of hepatic toxicity caused by 3,3',4,4',5-Pentachlorobiphenyl (PCB 126) in an animal model

Transcription

1 University of Iowa Iowa Research Online Theses and Dissertations Summer 2011 Using dietary strategies to explore mechanisms of hepatic toxicity caused by 3,3',4,4',5-Pentachlorobiphenyl (PCB 126) in an animal model Ian Kwan-Tai Lai University of Iowa Copyright 2011 Ian Lai This dissertation is available at Iowa Research Online: Recommended Citation Lai, Ian Kwan-Tai. "Using dietary strategies to explore mechanisms of hepatic toxicity caused by 3,3',4,4',5-Pentachlorobiphenyl (PCB 126) in an animal model." PhD (Doctor of Philosophy) thesis, University of Iowa, Follow this and additional works at: Part of the Toxicology Commons

2 USING DIETARY STRATEGIES TO EXPLORE MECHANISMS OF HEPATIC TOXICITY CAUSED BY 3,3,4,4,5-PENTACHLOROBIPHENYL (PCB 126) IN AN ANIMAL MODEL by Ian Kwan-Tai Lai An Abstract Of a thesis submitted in partial fulfillment of the requirements for the Doctor of Philosophy degree in Human Toxicology in the Graduate College of The University of Iowa July 2011 Thesis Supervisor: Professor Larry W. Robertson

3 1 ABSTRACT This doctoral dissertation work strived to contribute to the ever expanding knowledge about the mechanisms of polychlorinated biphenyl (PCB) toxicity using dietary strategies. PCBs are a family of persistent environmental pollutants with a wide range of toxicity. The toxicity of PCBs is largely dependent on the congener s chlorination pattern. Of particular interest to this work was 3,3,4,4,5- pentachlorobiphenyl (PCB 126), the most potent of the dioxin-like PCB congeners. I hypothesized that in vivo PCB 126 toxicity would be ameliorated by dietary selenium supplementation, lowered dietary copper, and dietary N-acetylcysteine (NAC) supplementation. Dioxin-like PCBs are known for diminishing hepatic selenium and seleniumdependent glutathione peroxidase (SeGPx), an antioxidant enzyme. In the first study, PCB 126 caused a dose-dependent decrease in hepatic selenium and SeGPx. Supplemental dietary selenium significantly increased hepatic selenium and SeGPx, and decreased incidence of liver apoptosis in these rats. The results from this study support the concept that selenium plays a protective role, and differences in liver injuries of these rats may be reflected in their selenium status. The dose-dependent increase in hepatic copper caused by PCB 126 was a subject of interest and concern in the next study. Lowering dietary copper levels without negatively affecting the function of the essential antioxidant enzyme copper zinc superoxide dismutase did not result in reduction of PCB 126-induced toxicity. Copper metabolism was unlikely a main target of PCB 126 toxicity as increasing dietary copper did not significantly increase hepatic copper levels. Hepatic copper is highly regulated and likely does not play a significant role in PCB 126-induced toxicity. The effectiveness of NAC on restoring glutathione status and reducing PCB 126 toxicity was tested in the final study. While NAC did not restore glutathione status, NAC

4 2 supplemented rats had significantly reduced severity of PCB 126-induced liver steatosis. The results of this study are consistent with the theory that NAC has a glutathioneindependent effect in improving mitochondrial energy metabolism. It also suggests that PCB 126-induced mitochondrial metabolic disruption of the liver is of concern in addition to oxidative stress. Abstract Approved: Thesis Supervisor Title and Department Date

5 USING DIETARY STRATEGIES TO EXPLORE MECHANISMS OF HEPATIC TOXICITY CAUSED BY 3,3,4,4,5-PENTACHLOROBIPHENYL (PCB 126) IN AN ANIMAL MODEL by Ian Kwan-Tai Lai A thesis submitted in partial fulfillment of the requirements for the Doctor of Philosophy degree in Human Toxicology in the Graduate College of The University of Iowa July 2011 Thesis Supervisor: Professor Larry W. Robertson

6 Graduate College The University of Iowa Iowa City, Iowa CERTIFICATE OF APPROVAL PH.D. THESIS This is to certify that the Ph.D. thesis of Ian Kwan-Tai Lai has been approved by the Examining Committee for the thesis requirement for the Doctor of Philosophy degree in Human Toxicology at the July 2011 graduation. Thesis Committee: Larry W. Robertson, Thesis Supervisor Gabriele Ludewig Michael Duffel Douglas Spitz Kyle Brown

7 To Mom and Dad, Thank You. ii

8 ACKNOWLEDGMENTS First and foremost I would like to thank my supervisor Dr. Robertson, without his guidance and patience over the years none of this work would have been possible. Thank you for encouraging me to explore my ideas and for your support of them. Dr. Ludewig, thank you for the help and support of my work. Their kindness and dedication was instrumental in allowing me to achieve my goals for this work. The animal studies in this work were large studies that required the help of my lab mates, for which I am grateful. I would especially like to thank Brittany Prather, Bingxuan Wang, Miao Li, and Kiran Dhakal, for devoting their time into helping me with these animal studies. I was fortunate to have worked Dr. Yingtao Chai, Dr. Brian Wels, and Dr. Don Simmons from the State Hygienic Lab, and thankful for their immense contributions towards the selenium and copper studies. Dr. Wanda Haschek and Dr. Alicia Olivier contributed significantly to the histological sections of this work. Dr. Kai Wang, thank you for your time in answering the many questions I had in regards to statistical analysis. Portions of this work were performed with the assistance of Dr. Spitz and his lab. Thank you for opening your laboratory and protocols to me, and specifically Dr. Michael McCormick for helping me learn many of the assays that I used for this work. Dr. Duffel, your enzymology class opened my eyes into the world of enzymes, and thank you for your help on the subject. Dr. Brown, your expertise of the liver was valuable in many ways and thank you for always being willing to answer my questions. Last but not least, thank you to all my friends and family who have supported me through this entire process. Dr. Jim Jacbous, your willingness to help me both as a colleague and as a friend from the beginning is much appreciated. Finally, Mom and Dad, without your endless support and sacrifices I would never have gotten to where I am today. Thank you for always believing in me. iii

9 ABSTRACT This doctoral dissertation work strived to contribute to the ever expanding knowledge about the mechanisms of polychlorinated biphenyl (PCB) toxicity using dietary strategies. PCBs are a family of persistent environmental pollutants with a wide range of toxicity. The toxicity of PCBs is largely dependent on the congener s chlorination pattern. Of particular interest to this work was 3,3,4,4,5- pentachlorobiphenyl (PCB 126), the most potent of the dioxin-like PCB congeners. I hypothesized that in vivo PCB 126 toxicity would be ameliorated by dietary selenium supplementation, lowered dietary copper, and dietary N-acetylcysteine (NAC) supplementation. Dioxin-like PCBs are known for diminishing hepatic selenium and seleniumdependent glutathione peroxidase (SeGPx), an antioxidant enzyme. In the first study, PCB 126 caused a dose-dependent decrease in hepatic selenium and SeGPx. Supplemental dietary selenium significantly increased hepatic selenium and SeGPx, and decreased incidence of liver apoptosis in these rats. The results from this study support the concept that selenium plays a protective role, and differences in liver injuries of these rats may be reflected in their selenium status. The dose-dependent increase in hepatic copper caused by PCB 126 was a topic of interest and concern in the next study. Lowering dietary copper levels without negatively affecting the function of the essential antioxidant enzyme copper zinc superoxide dismutase did not result in reduction of PCB 126-induced toxicity. Copper metabolism was unlikely a main target of PCB 126 toxicity as increasing dietary copper did not significantly increase hepatic copper levels. Hepatic copper is highly regulated and likely does not play a significant role in PCB 126-induced toxicity. The effectiveness of NAC on restoring glutathione status and reducing PCB 126 toxicity was tested in the final study. While NAC did not restore glutathione status, NAC iv

10 supplemented rats had significantly reduced severity of PCB 126-induced liver steatosis. The results of this study are consistent with the theory that NAC has a glutathioneindependent effect in improving mitochondrial energy metabolism. It also suggests that PCB 126-induced mitochondrial metabolic disruption of the liver is of concern in addition to oxidative stress. v

11 TABLE OF CONTENTS LIST OF TABLES... ix LIST OF FIGURES...x CHAPTER I POLYCHLORINATED BIPHENYLS: PERSISTENT ENVIRONMENTAL POLLUTANTS...1 Introduction: Origins of PCBs...1 PCBs as Public Health Threat...4 Routes of PCB Exposure...4 PCB Toxicity...5 PCBs and Carcinogenesis...6 Non-carcinogenic Effects...7 Respiratory and Immunologic Effects...7 Reproductive, Endocrine, and Developmental Effects...8 Neurologic Effects...9 Hepatic Effects...10 Mechanisms of PCB Toxicity ,3,4,4,5-Pentachlorobiphenyl...13 Dietary Supplementation Approaches...16 Selenium...17 Selenium Metabolism...18 Glutathione Peroxidases...20 Thioredoxin Reductase...21 Selenium in Disease...21 Selenium and Carcinogenesis...22 Copper...24 Copper Metabolism...25 Ceruloplasmin...26 Cytochrome C Oxidase...27 Superoxide Dismutase...27 Copper in Disease...28 N-Acetylcysteine...31 NAC Metabolism...32 NAC in Disease...33 Specific Aims of Thesis...37 CHAPTER II ACUTE TOXICITY OF 3,3,4,4,5-PENTACHLOROBIPHENYL (PCB 126) IN MALE SPRAGUE DAWLEY RATS: EFFECTS ON HEPATIC OXIDATIVE STRESS, GLUTATHIONE AND METALS STATUS...39 Abstract...39 Introduction...40 Materials and Methods...41 Chemicals...41 Animals...42 Glutathione (GSH/GSSG) Analysis...43 Trace Elements Determination...43 Histology...44 Measurement of CYP1A Activity...44 vi

12 Measurement of Glutathione Peroxidase (GPx) activities...44 Statistics...44 Results...45 Effects on Growth and Organ Weights...45 Effects on Total Hepatic GSH and GSSG...45 Effects on Trace Elements and GPx Activities...45 Histology...45 Effects on EROD and MROD Activities...46 Discussion...46 CHAPTER III THE EFFECTS OF DIETARY SELENIUM SUPPLEMENTATION ON ANTIOXIDANT STATUS DURING PCB 126 TOXICITY...61 Abstract...61 Introduction...62 Materials and Methods...63 Chemicals...63 Animals, Diet and PCB 126 Exposure...64 Preparation of Hepatic Subcellular Fractions...64 Measurement of Cytochrome P450 (CYP1A1) Activity...65 Activity Determination of Superoxide Dismutases (SOD)...65 Trace Elements Determinations...65 Activity Determination of Glutathione Peroxidases (SeGPx, Total GPx, GST)...66 Measurement of Thioredoxin Reductase (TrxR) Activity...66 Measurement of the Redox State of Thioredoxin-1 (Trx1) and Thioredoxin-2 (Trx2)...66 Total Glutathione (GSH and GSSG) Analysis...67 Histology...67 Statistics...68 Results...68 Growth and Organ Weights...68 Effects on CYP 1A1 (EROD) Activity...69 Effects on Hepatic Se...69 Effects on Hepatic Copper, Iron, Manganese, and Zinc...69 CuZnSOD, and Total SOD Activities...70 Effects on Glutathione Peroxidase Activities...70 Effects on Thioredoxin Reductase Activity and Thioredoxin Oxidation States...71 Effects on Hepatic Glutathione...72 Histology...72 Discussion...73 CHAPTER IV THE DISPOSITION AND ROLE OF COPPER IN RODENT LIVER TOXICITY FOLLOWING EXPOSURE TO PCB Abstract...94 Introduction...95 Materials and Methods...96 Chemicals...96 Animals...97 Hepatic subcellular fractions preparation...97 Measurement of CYP1A1 activity...97 vii

13 Total Glutathione (GSH) analysis Hydroxynonenal (4-HNE) determination...98 Measurement of Superoxide Dismutase (SOD) activities...98 Ceruloplasmin determination...99 Trace elements determination...99 Histology...99 Statistics Results Effects on growth, feed consumption, and organ weights Effects on EROD activity Effects on liver, kidney, and blood Cu Effects on liver, kidney, and blood iron Effects on liver, kidney, and blood selenium Effects on liver, kidney, and blood zinc Effects on liver, kidney, and blood manganese Effects on liver, kidney, and blood molybdenum Effects on CuZnSOD, MnSOD, and Total SOD activities Effects on total hepatic glutathione Effects on Liver 4-HNE adducts Effects on serum ceruloplasmin Histology Discussion CHAPTER V N-ACETYCYSTEINE DIMINISHES THE SEVERITY OF PCB 126 INDUCED FATTY LIVER IN MALE RODENTS Abstract Introduction Methods and Materials Chemicals Animals Hepatic subcellular fractions preparation Glutathione analysis Glutathione-S-transferase (GST) activity Histology and Special Stains Lipid Staining and Quantification Statistics Results Effects on growth, feed consumption, and organ weights Effects on EROD activity Effects on total glutathione and oxidized glutathione (GSSG) Effects on hepatic glutathione transferase (GST) activity Histology Discussion CHAPTER VI SUMMARY, CONCLUSIONS, AND FUTURE STUDIES PCBs and Diet Overview of Studies Future Directions REFERENCES viii

14 LIST OF TABLES Table 1.1.Comparison of composition of Aroclor mixtures...3 Table 2.1.AIN-93 Diet Composition...51 Table 2.2.Body Weight, Liver Weight, Thymus Weight, and Ratio...52 Table 2.3.GSH & GSSG...53 Table 2.4.SeGPx and total GPx activities, (A) and CYP 1A activity (B)...54 Table 3.1.Composition of AIN-93M modified selenium diets...80 Table 3.2.Two-way ANOVA analysis of the effects of PCB 126, dietary selenium, and interaction...81 Table 3.3.Growth (%) and relative liver weights (%) (A), and relative lung weights (%) and relative thymus weights (%) (B)...82 Table 3.4.EROD activity (nmol/min/mg protein)...84 Table 3.5.Liver copper ( g/g) and iron ( g/g) (A), and manganese ( g/g) and zinc ( g/g) (B)...85 Table 3.6.CuZnSOD activity (U/mg protein)...87 Table 3.7.Glutathione transferase ( mol/min/mg protein) and total glutathione peroxidase activities ( mol/min/mg protein)...88 Table 3.8.GSH (nmoles GSH/mg liver wet weight), GSSG (nmoles GSSG/mg liver wet weight), and GSSG/GSH ratio...89 Table 4.1.Composition of AIN-93G modified copper diets Table 4.2.Two-way ANOVA analysis of the effects of dietary copper, PCB 126, and interaction Table 4.3.Liver ( g/g), kidney ( g/g), and blood ( g/l) copper (A), iron (B), selenium (C), zinc (D), manganese (E), and molybdenum (F) Table 4.4.CuZnSOD, MnSOD, and Total SOD activities (U/mg protein) Table 5.1.Composition of AIN-93G and modified NAC supplemented diets Table 5.2.Two-way ANOVA analysis of the effects of NAC and PCB ix

15 LIST OF FIGURES Figure 1.1.General chemical structure of polychlorinated biphenyls...2 Figure 1.2.Catalytic cycle of cytochrome P450 enzymes...12 Figure 1.3.Comparison of 3,3',4,4',5-pentachlorobiphenyl (PCB 126) and 2,3,7,8- tetrachlorodibenzene-p-dioxin (TCDD) Figure 1.4.Metabolic pathways of selenium...18 Figure 1.5.Reaction catalyzed by selenium-dependent glutathione peroxidase (SeGPx)...20 Figure 1.6.Cellular reactions of thioredoxin reductase...21 Figure 1.7.Fenton-like reactions of free copper...26 Figure 1.8.Reactions catalyzed by CuZnSOD...27 Figure 1.9.Glutathione (GSH) synthesis from N-acetylcysteine (NAC)...32 Figure 1.10.Acetaminophen metabolism Figure 2.1.Growth curve of vehicle- (control) and PCB 126-treated rats...55 Figure 2.2.Liver and kidney selenium (A), copper (B), and zinc (C) levels of vehicle- (control) and PCB 126-treated rats...58 Figure 2.3.Histopathology of liver from vehicle- (control) (A) and PCB 126-treated (B) rats...59 Figure 2.4.Histopathology of thymus from vehicle- (control) (A and B) and PCB 126-treated (C and D) rats...60 Figure 3.1.Liver selenium ( g/g tissue)...90 Figure 3.2.Se-dependent glutathione peroxidase activity ( mol/min/mg protein)...91 Figure 3.3.Thioredoxin reductase activity (U/ml)...92 Figure 3.4.Thioredoxin-1 and Thioredoxin-2 oxidation state...93 Figure 4.1.Growth (A) and feed consumption (B) of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals Figure 4.2.Relative liver (A) and kidney (B) weights of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals Figure 4.3.Liver ethoxyresorufin-o-deethylase (EROD) activity of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals x

16 Figure 4.4.Liver total Glutathione (GSH) of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals Figure 4.5.Liver 4-HNE levels of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals Figure 4.6.Serum ceruloplasmin of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals Figure 4.7.Percentage of lipids in liver tissue of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals Figure 4.8.Electron micrographs from hepatocytes of rats fed adequate (6 ppm) copper and treated with corn oil (vehicle) (A) or and PCB 126 (1 mol/kg and 5 mol/kg) (B and C) Figure 5.1.Growth (A) and feed consumption (B) of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals Figure 5.2.Relative liver (A), lung (B), and thymus (C) weights of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals Figure 5.3.Liver ethoxyresorufin-o-deethylase (EROD) activity of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals Figure 5.4.Liver total glutathione (A), oxidized glutathione (GSSG) (B), ratio of liver oxidized glutathione (GSSG) (C) to liver total glutathione of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals Figure 5.5.Liver glutathione-s-transferase (GST) activity of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals Figure 5.6.Osmium tetroxide staining for lipid (stains lipid black) of liver from control and PCB treated rats fed a control (AIN-93G) and NACsupplemented diet Figure 5.7.Histological examination of the liver from control and PCB treated rats fed the control (AIN-93G) and NAC supplemented diet xi

17 1 CHAPTER I POLYCHLORINATED BIPHENYLS: PERSISTENT ENVIRONMENTAL POLLUTANTS Introduction: Origins of PCBs Polychlorinated biphenyls (PCBs) are a class of synthetic organic chemical compounds containing one to ten chlorines attached to a structure containing two benzene rings, called a biphenyl. The variable chlorination pattern gives rise to 209 distinct congeners each with distinct chemical, physical, and biological properties. These congeners were not manufactured individually but rather as mixtures that can have more than a hundred congeners (Mayes et al. 1998). PCB mixtures were synthesized by a reaction between chlorine gas and biphenyl in a reactor under extremely hot and pressurized conditions. Under these conditions and in the presence of a catalyst, usually an electron transporter such as iron, chlorine atoms replace the hydrogen atoms attached to the biphenyl. This method of batch synthesis produced the mixtures that were widely used, and the physical characteristics of these mixtures ranged from liquid to resin and colorless to yellow, depending on the congeners found in the mixtures (Silberhorn et al. 1990). PCB mixtures were largely chemically inert and stable under heat, making them an ideal alternative to the flammable solvents used for a variety of industrial and commercial applications. The stability of PCBs saw them gain widespread usage in dielectrics, as transformer and capacitor oils and cooling fluids for hydraulic systems. They were also used as organic diluents, plasticizers, pesticides, flame-retardants, carbonless copy paper and sealants (Safe 1993). Unfortunately, it was these same characteristics of stability and persistence that eventually brought PCBs into the limelight as a public health threat.

18 2 Figure 1.1. General chemical structure of polychlorinated biphenyls. Substitution with 1-10 chlorines in the biphenyl gives rise to 209 possible congeners with various biological, physical, and chemical properties. PCBs mixtures were sold commercially in mass quantities under various trade names. In the United States, PCB mixtures were manufactured by the Monsanto Company under the trade name Aroclor from 1929 to 1977, with production peaking in the United States in 1970 (ATSDR 2000). Aroclor mixtures were given four-numbered names, with the first two defining them as being refined and the last two based on their chlorine composition in percent weight (Erickson and Kaley 2011). For example, Aroclor 1242 consists mainly of the lighter tri- and tetrachlorobiphenyls that are more likely to become airborne, while Aroclor 1254 consists mainly of the higher-chlorinated and heavier penta- and hexchloro biphenyls (Table 1.1). Most of the Aroclor mixtures produced were between 20 to 60 percent chlorine by weight, and the industrial or commercial usages of these mixtures were also largely dependent upon the chlorination of these mixtures.

19 3 Table 1.1. Comparison of composition of Aroclor mixtures Aroclor No a 1254b 1260 % U.S. Production Usage Test article Test article Commercial product Test article Test article Monochloro-BPs Dichloro-BPs Trichloro-BPs Tetrachloro-BPs Pentachloro-BPs Hexachloro-BPs Heptachloro-BPs Octachloro-BPs Nonachloro-BPs Decachloro-BPs Source: (Mayes et al. 1998)

20 4 PCBs as Public Health Threat Concerns regarding the long term usage of PCBs were mounting by the late 1960s and early 1970s. While there were occasional reports of PCBs toxicity in the 1930s and 1940s from occupational exposure, these were considered to be rare events given the widespread usage (Golden and Kimbrough 2009). As PCBs have no known taste or smell, these minor incidences did not lead PCBs to becoming a public health threat. However, two large scale incidences of accidental poisoning with PCBs shifted public opinion against PCBs, and combined with the widespread discovery of PCBs outside of their intended usage increasing, eventually led to the decline of PCBs. Despite these concerns leading to the cessation of PCB production in the United States and other developed countries in 1977, such was the popularity of PCBs that 1.1 billion pounds of PCBs were estimated to have been in the United States alone (Ross 2004). Ultimately their useful properties means PCBs are still in present day use, limited to closed dielectrics systems. Routes of PCB Exposure Before the ban on PCB manufacturing and decline in usage, large amounts of PCBs entered the environment simply through the routine manufacture and usage processes. PCBs also entered the environment from contaminated waste buried in landfills, as well as from accidental spills and leaks. While the release of PCBs into the environment has been limited since the manufacturing ban, releases from usage or leaks of existing dielectrics still occur. Illegal or improper dumping of PCB-contaminated wastes and poorly maintained PCB waste sites are also ongoing concerns (Hafferty et al. 1977). Recent outbreaks of contaminated feed for animals, such as the Irish Pork crisis (2008), have once again raised concerns about exposure of PCBs to the general public (White and Birnbaum 2009).

21 5 Because of their stable nature, PCBs that are released would enter the environment and remain for a long time. While chemically stable, PCBs can move about ubiquitously in air, water, and soil, further distributing them around the world. The lower-chlorinated PCBs are more prone to becoming airborne as a particle or vapor and traveling greater distances due to their light weight (Fox et al. 1994). In addition to occupational inhalation of airborne PCBs, recent dredging of PCB-contaminated sediments from heavy industrial areas has raised new concerns about exposure to residents in nearby areas (Martinez et al. 2010). In contrast, the higher-chlorinated PCBs are more likely to settle into the ground and water, where they remain resistant to degradation or become taken up into the food chain (Yakushiji 1988). PCB uptake into the food chain is of real concern, because PCBs do not degrade once they enters the food chain, instead remaining in animal fats and accumulating as they moves up the food chain. Waterborne PCBs taken up by fishes and other small aquatic organisms, where they then continue to bioaccumulate as the smaller organisms with PCBs are consumed by larger predators (La Rocca and Mantovani 2006). In a similar manner, exposure of soil-dwelling land organisms to PCB-contaminated soil also results in the bioaccumulation up the food chain. In both cases, by the time PCBs reach humans at the top of the food chain, their levels are at their highest (Hansen 1987). PCB Toxicity The toxicity of PCBs was first brought to public attention by two accidental poisoning incidents, eventually leading to the decline of PCBs. The Yusho incident in Japan (1968) and the Yu-Cheng incident in Taiwan (1979) both had highly acute toxic effects. These incidences occurred when oils intended for use with food were accidentally contaminated with highly concentrated PCB mixtures, which are highly lipophilic making contamination hardly detectable. While PCBs are not likely to react directly with human tissues even after acute exposure to large doses, acute PCB toxicity still can cause

22 6 serious long-term consequences. The majority of these acute PCB poisonings resulted in severe dermatologic effects, presenting as chloracne (Aoki 2001), lasting as long as fifteen years (Masuda 2001). Because the culprits in these incidences were PCB-mixtures or possibly their heat-degraded byproducts, polychlorinated dibenzofuans (PCDFs), exact mechanisms for the effects seen in acute human PCB toxicity remain in question (ATSDR 2000). While the exposure was acute, due to the persistence of PCBs in the body, the actual exposure can become chronic due to their ability to persist. Chronic exposure to PCBs has been linked to a battery of health effects, including carcinogenesis. PCBs and Carcinogenesis PCBs have been labeled as probable carcinogens by the U.S. Environmental Protection Agency (EPA) and International Agency for Research of Cancer (IARC) (IARC 1987; IRIS 2006). Epidemiological studies have associated the Yusho and Yucheng poisonings with increased hepatocarcinogenesis (Silberhorn et al. 1990). Other epidemiological studies involving capacitor and transformer workers exposed to PCBs were associated with higher rates of liver cancer, gastrointestinal cancer, and skin melanomas (Gustavsson and Hogstedt 1997; Kimbrough et al. 1999). Subsequent animal studies with chronic dosing of Aroclors, particularly the mixtures with higher chlorinated PCBs, have supported the findings in humans that PCBs result in benign liver tumor formation in rats that may eventually progress to malignant carcinomas (Mayes et al. 1998). Carcinogenesis is a multi-stage process requiring an initiating event causing irreversible genetic mutations to a cell, followed by promotion to a benign tumor, and finally progression into carcinoma. PCBs are complete carcinogens with effects in all three stages of carcinogenesis (Pereg et al. 2002). Because they are more likely to be biotransformed, lower chlorinated PCBs have higher initiating potential, while repeated animal studies have shown PCBs to be promoters in initiated cells of the liver and lungs

23 7 exposed to PCB mixtures and or individual higher-chlorinated congeners (Beebe et al. 1993; Hemming et al. 1993). Non-carcinogenic Effects Because the 209 individual PCBs congeners each have varying chemical, physical, and biological properties, PCB mixtures can have a wide range of toxicity depending on the congeners involved. While dermatologic effects were the most obvious of the non-carcinogenic effects, other more serious effects were discovered during epidemiological studies of populations exposed to PCBs. The following is a brief review of non-carcinogenic health effects associated with PCB exposure: Respiratory and Immunologic Effects In addition to dermatologic disorders, respiratory disorders were also widely associated with the populations exposed to PCBs, especially from the Yusho and Yu- Cheng poisonings. Clinically these patients presented symptoms similar to chronic bronchitis with persistent coughing and sputum production (Shigematsu et al. 1971). Lowered pulmonary function following chronic PCB exposure was observed even in nonsmokers. Capacitor workers chronically exposed to PCBs were reported to have respiratory distress and reduced pulmonary function (Fischbein et al. 1979; Warshaw et al. 1979). However, it should be noted that these workers are commonly exposed to other agents that may also have cardiovascular effects, such that it cannot be concluded that PCBs were solely responsible for these effects (ATSDR 2000). Epidemiological and animal studies have also reported immunosuppression as an effect of PCB exposure (Kunita et al. 1985; Nakanishi et al. 1985). Those who were already suffering from respiratory distress were more susceptible to bacterial or viral infections. Lu (1985) reported humoral immunity to be lower in Yu-Cheng patients. These patients were found to have significantly lower immunoglobulin levels, changes to T-lymphocytes, and delayed skin hypersensitivity to antigens. Particularly troubling is the

24 8 finding that immune-sensitivity is increased in infants exposed to PCBs in utero and/or during breast feeding (Weisglas-Kuperus et al. 2000). In contrast to the respiratory findings, PCB immunotoxicity has been widely reported in animal studies. Morphologically, thymic and splenic atrophy are commonly observed in animals exposed to PCBs. Similar to the effects observed in humans, animals exposed to PCBs also showed altered sensitivity to antigens and increased susceptibility to infections. Reproductive, Endocrine, and Developmental Effects Reproductive effects of PCBs have been reported as disturbances in female menstruation and male fertility. Menstrual disturbances were observed in females from Yusho accidental contamination and cohorts chronically exposed to PCBs from dietary sources (Kusuda 1971; Mendola et al. 1997). Exposure during pregnancies was associated with higher rates of miscarriage, and those pregnancies that do come to term have been associated with developmental effects (Gerhard et al. 1998). Although decreased sperm motility in infertile men was associated with significantly higher levels of blood PCBs (Bush et al. 1986; Pines et al. 1987), results from human epidemiological studies have been inconclusive. Results have also been inconclusive in animal studies, however, significant effects were observed on the seminal vesicles and sperm count in weanling rats exposed to PCBs (Gray et al. 1993). The greatest effects on reducing fertility were observed in animals exposed via lactation, suggesting that exposure during developmental periods may be more critical for toxicity (Sager et al. 1987). PCB-induced endocrine disruption is not limited to reproduction. Thyroid hormone production is susceptible to disruption, with the results dependent on the mixture of PCBs to which an individual is exposed to. Thyroid hypertrophy was observed in workers exposed to PCBs, while, in severe cases such as Yusho, patients developed goiters (Guo et al. 1999). Epidemiological studies of serum thyroxine levels in cohorts exposed to PCBs have been inconclusive, however, in many cases circulating thyroid

25 9 hormone levels were disrupted but without a clear trend observed in animals studies. Animal studies have provided a clearer picture of PCB s role in endocrine disruption, as studies have shown PCBs to disrupt thyroid hormone production, interfere with transport, and accelerate elimination from the body (Fisher et al. 2006; Liu et al. 1995; Vansell and Klaassen 2001). Endocrine disruption was linked to developmental effects by Goldey and Crofton (1998), showing that animals with thyroxine levels lowered by PCBs were more susceptible to developing neurodevelopmental deficits. Decreased birth weight of about 15% and slowed growth were first noticed in children who were exposed to PCBs in utero from the Yusho and Yu-cheng incidents (Rogan et al. 1988). Height and weight gains of children exposed directly to PCBs were significantly decreased (Masuda 2003). Another concern with chronic exposure during development is largely due to the lipophilicity of PCBs, making adipose tissue, including the mammary tissues, a specific target for accumulation and toxicity. Animal studies have shown that lactational exposure to PCBs significantly decreased body weight (Lundkvist 1990). In addition, developmental toxicity of PCBs has been widely associated with neurologic effects, as neurodevelopment is critical during childhood. Neurologic Effects Extensive studies of the neurologic effects of PCBs have shown an association with neurological alterations. Although adult workers exposed to PCBs rarely exhibit clinical neurological deficits, particular attention has been paid to mothers who may indirectly expose their offspring to PCBs. Epidemiological studies of children exposed to PCBs have shown that PCB-exposed children consistently scored lower in motor function tests (ATSDR 2000). Repeated findings of hypoactive reflexes, motor immaturity, and poor cognitive functions in humans shows that PCBs can cause irreversible damage during these critical growth and development periods. Children exposed prenatally to PCBs were also reported to perform poorly at standardized exams such as the IQ test

26 10 (Jacobson and Jacobson 1996). Results from extensive animal studies have supported both of these findings. Rats exposed to PCBs were found to have decreased motor skills and decreased performance in learning spatial and visual tasks (Nishida et al. 1997). The most consistent finding in animal studies is the decrease in dopamine concentrations in the many areas of the brain, which may be associated with the clinical symptoms observed (ATSDR 2000). Hepatic Effects The liver plays a vital role in xenobiotic biotransformation and detoxification, making it an intensely studied target of PCB toxicity. PCBs not taken up into adipose tissue are taken to the liver where attempts at biotransformation take place. This process has been widely studied and associated with a large number of hepatic effects. Induction of xenobiotic enzymes is commonly seen as the first step towards severe liver injury, although it should be noted that congener-specificity plays a large role in this induction (Parkinson et al. 1983a; Van den Berg et al. 1994). However, research into PCB-induced human hepatic enzyme induction has been limited. Clinical assays have mainly been limited to liver function tests, such as AST, ALT, and GGT. Elevated levels of these enzymes have been positively correlated with hepatomegaly and blood PCB levels (Brown et al. 1994). Liver tissues from Yusho and Yu-cheng patients were observed to have distinct morphological changes consistent with that of microsomal enzyme induction or liver damage (Kuratsune 1989). Because of the difficulty in congener specificity, controlling for confounding variables, and limited sample populations, researchers have largely focused their efforts on animal studies in studying the hepatic effects of PCBs. The most common findings from animal studies are hepatomegaly, hepatic microsomal enzyme induction, and elevated levels of serum hepatic enzymes (Pond 1982). The most distinct morphologic change is the proliferation of rough endoplasmic

27 11 reticulum, consistent with the induction of microsomal enzymes, specifically the cytochrome P450 isoenzymes (CYP). CYP was named because interruption of its catalytic cycle with carbon monoxide resulted in an absorbance peak at 450 nm. These enzymes have diverse roles, including catalyzing the oxidative biotransformation of PCBs, which will be reviewed in further detail in the following section. Lipid droplet accumulation may also occur in a similar manner to the early stages of non-alcoholic fatty liver disease (NAFLD) and indicative of altered lipid metabolism (Emond et al. 2005). Early research focused on enzyme induction and activation of specific cellular receptors by PCBs that fit physically and chemically for these receptors. Mechanisms of PCB Toxicity Receptor-mediated toxicity has been an intense area of research into determining the mechanisms of PCB toxicity. Of particular interest are the aryl hydrocarbon (AhR) and constitutive androstane (CAR) receptors, the activation of which leads to the transcription of CYP1A1/2 and CYP2B1/2 enzymes, respectively. In lower chlorinated PCBs, CYP catalyzes the biotransformation of PCBs to arene oxides. They can then undergo further biotransformation to hydroxylated PCBs. PCBs with hydroxyl groups in the para or ortho positions to each other can undergo further oxidation to reactive quinone metabolites (Amaro et al. 1996; Oakley et al. 1996; Safe 1980). Because unsubstituted meta and para carbons are the preferred sites of oxidation, higher chlorinated PCBs are more resistant to biotransformation by CYPs. However, this does not prevent the binding of PCBs to the active site of CYPs, resulting in the disruption of the catalytic cycle of the enzyme. At the active site of CYP is a heme group that binds to oxygen before oxygenation of the substrate (Figure 1.2). Interruption or incompletion of this cycle can result in the incomplete reduction of oxygen, causing release of reactive intermediates superoxide and hydrogen peroxide (Shertzer et al. 2004).

28 Figure 1.2. Catalytic cycle of cytochrome P450 enzymes. Complete oxygenation by CYP requires oxygen and two electrons, with water released as a byproduct. Source: (Parkinson and Ogilvie 2010). Higher-chlorinated PCBs interrupt the cycle during oxygenation (C to D), causing the release of superoxide and hydrogen peroxide (Safe 1980). 12

29 13 The most intense research has focused on the AhR agonists, which have been widely suspected of being the culprit of the toxicity caused by Aroclors and other PCB mixtures. These so-called dioxin-like PCBs, known for their resemblance to 2,3,7,8- tetrachlorodibenzo-p-dioxin (TCDD), originally an unintentional by-product of organochlorine manufacturing. Acute exposure to small doses of TCDD was shown to inhibit antioxidant enzymes and cause lipid peroxidation (Pohjanvirta et al. 1990; Stohs et al. 1986). Chronic exposure to TCDD has been shown to produce liver and lung carcinogenesis in animals (Knerr and Schrenk 2006). The dioxin-like PCBs include 3,3,4,4 -tetrachlorobiphenyl (PCB 77), 3,3,4,4,5-pentachlorobiphenyl (PCB 126), and 3,3,4,4,5,5 -hexachlorobiphenyl (PCB 169). The lack of ortho substitution of the carbon in either ring allows these PCBs take assume a more co-planar formation similar to that of TCDD, making it an ideal ligand for the AhR. The AhR, a cytosolic receptor, translocates to the nucleus upon activation by dioxin or a dioxin-like ligand. It heterodimerizes with the AhR nuclear translocator protein (ARNT) and binds to the dioxin response element (DRE) sequences in the DNA, inducing transcription of many genes including CYP1A1, which is commonly used as a biomarker for AhR agonist activity. CYP1A1 can also act as a carcinogenic initiator, by catalyzing the biotransformation of a xenobiotic such as benzo-[a]-pyrene into a mutagen (Shi et al. 2009). Recent evidence also suggests that AhR deregulates cell-cell contact resulting in tumor promotion and progression, findings that further point to toxicity of dioxin-like PCBs (Dietrich and Kaina 2010). Because humans are usually exposed to mixtures, the toxicity of these dioxin-like PCBs is classified using the toxic equivalency factor (TEF) potency scheme, with TCDD given the value of 1 as the most potent dioxin (Birnbaum and DeVito 1995). 3,3,4,4,5-Pentachlorobiphenyl

30 14 Figure 1.3. Comparison of 3,3',4,4',5-pentachlorobiphenyl (PCB 126) and 2,3,7,8- tetrachlorodibenzene-p-dioxin (TCDD). PCB 126 co-planar-like conformation and physical size closely resembles that of TCDD, making it a potent AhR agonist. One of the well-studied dioxin-like PCB congeners is 3,3,4,4,5- pentachlorobiphenyl (PCB 126). PCB 126 was one of the congeners produced in Aroclor 1254, a commercially produced PCB mixture. Despite being a higher-chlorinated PCB and thus not expected to be airborne, a recent study found PCB 126 in the Chicago air profile (Zhao et al. 2010). PCB 126 is the most potent of these dioxin-like PCBs, with a TEF of 0.1 relative to TCDD. This TEF value indicates that PCB 126 has one-tenth the dioxin toxicity of TCDD, much higher than those of the other dioxin-like PCBs, because PCB 126 is the most similar in size to TCDD (Figure 1.3). PCB 126 is potent ligand for the AhR, and thus an efficacious inducer of CYP1A1/2 as has been shown in animal studies. In the studies published by the National Toxicology Program (2006), rats chronically exposed to low doses PCB 126 for 2-years developed a variety of adverse health effects in multiple organs. These included liver, lung, and mouth cancer; liver hypertrophy, hyperplasia, and fibrosis; lung metaplasia; adrenal gland atrophy or hypertrophy; pancreas inflammation and atrophy; kidney nephropathy; cardiomyopathy; thymus and spleen atrophy; and mesentery inflammation.

31 15 Oxidative stress has long been suspected as a contributing factor towards PCB 126-induced liver toxicity (Hassoun et al. 2002), specifically carcinogenesis. One possible source of ROS release caused by PCB 126 is from the uncoupling of the catalytic cycle of CYP1A1. Although PCB 126 can bind to CYP1A1, it is not easily biotransformed, interrupting the catalytic cycle of the enzyme. However, cells have developed adaptive protective responses against unwanted ROS release, better known as antioxidants. The best-known of these is glutathione (GSH), a tripeptide containing a thiol from the amino acid cysteine, which can donate electrons to ROS and prevent them from reacting with the cellular membranes and genome. This process causes glutathione to become oxidized forming a disulfide bond (GSSG), with the enzyme glutathione reductase (GR) converting GSSG back to reduced GSH at the expense of NADPH. It is of great concern that dioxin-like compounds are known for diminishing GSH, despite the induction of GR (Senft et al. 2002a). GSH is also a cofactor for other antioxidant enzymes, the glutathione peroxidases (GPx), which detoxify lipid hydroperoxides and hydrogen peroxide to water. Another group of enzymes with peroxidase activity is the glutathione-s-transferases (GST) that catalyzes the conjugation of potentially toxic electrophiles with GSH for elimination from the body. Extensive research has been focused on a specific form of GPx, the selenium-dependent GPx (SeGPx), which was shown to be significantly diminished by dioxin-like PCBs (Twaroski et al. 2001b). The exact mechanisms causing these cellular changes and their contribution to toxicity observed are currently under investigation. Because steatosis, or accumulation of lipids, is a commonly observed noncarcinogenic injury in PCB 126-induced toxicity, an alternative theory posits that PCB 126 and the AhR play significant roles in altering energy and lipid metabolism. One recent animal study by Moffat et al. (2010) using AhR-splice variants rats found that transactivation of the AhR affects genes involved in lipid metabolism, cellular membrane function, and energy metabolism. As the major site of lipogenesis, the liver is especially

32 16 susceptible to these effects. Lipid metabolism is a complex but essential process that produces adenosine triphosphate (ATP), the main source of cellular energy. ATP is generated from the metabolism of triglycerides, a process that begins with the transport of free fatty acids into the mitochondria for conversion into triglycerides. Kawano et al. (2010) showed that the AhR via the PPAR upregulates fatty acid translocase (FAT), increasing the transport of free fatty acids into the liver. This is followed by -oxidation to produce acetyl-coa, a process that was shown to be disrupted by the AhR (Tessari et al. 2009). Acetyl-coA is then used in the citric acid cycle to produce the cofactors necessary for the electron transport chain. Disruption of the electron transport chain by AhR agonists resulted in diminished ATP levels in the liver (Forgacs et al. 2010; Senft et al. 2002b). These findings point to hepatic microvesicular steatosis caused by AhR activation is a symptom of metabolic disruption. Dietary Supplementation Approaches Because PCB 126 toxicity involves redox and metabolic disruptions, maintaining homeostasis of physiological functions in the event of exposure becomes critical in preventing or reducing toxicity (Goldhaber 2003). Diet and nutrition is an important part of maintaining this homeostasis, as our diets contain many micronutrients that have antioxidant properties. With nutritional awareness increasing in modern society, so has the use of micronutrient supplementation to prevent deficiencies. Many of these nutritional supplements can be found in over-the-counter products to be purchased without the need for a prescription from a doctor. Supplementation of these micronutrients beyond recommended therapeutic levels can cause toxicity. Because of these concerns, it is necessary to study the benefits and drawbacks of dietary modulations and supplementation. Results from both human and animal dietary studies have been used in risk assessment studies by government agencies to determine the therapeutic ranges of micronutrients.

33 17 To prevent dietary deficiencies, the US Food and Drug Administration (FDA) has developed Reference Daily Intake (RDI) levels for various micronutrients. The RDI values are based on data gathered by the U.S. Food and Nutrition Board of the Institute of Medicine. On the opposite end of the spectrum, Reference Doses (RfD) for toxicity for certain micronutrients can be found in the US EPA s risk assessment studies of chemicals. RfD is an estimate of the daily exposure to which the human population may be continually exposed over a lifetime without an appreciable risk of deleterious effects. When used in their therapeutic range, micronutrient modulation or supplementation has potential to help maintain homeostasis, and prevent or reduce toxicity. The micronutrients of particular interest to this current work are the essential trace elements selenium and copper, and the cysteine derivative N-acetylcysteine (NAC). Selenium Selenium is an essential trace element chemically closely related to sulfur. Selenium was named after the Greek goddess of the moon, Selene, after its discovery as a byproduct of sulfur production in 1818 by Swedish chemist Jons Jakob Berzelius. Selenium is a chalcogen falling in between sulfur and tellurium the Group 16 of the Periodic Table, and is considered a metalloid because it has both metallic and nonmetallic properties. Selenium rarely occurs in elemental form in nature, as it is usually found in semiconductor form, replacing sulfur in pyrite ores (Wiberg et al. 2001). While no longer popular as a semiconductor in electronics due to the use of silicon, selenium is still commercially used as a coloring dye in the glassmaking industry (Brown 1998), and also in shampoos for its antidandruff properties. Commercial production of selenium occurs as a byproduct during the process of copper electrolyte refining (Haygarth 1994). Selenium can be found in soil at different levels depending on the region. Plants such as garlic and yeast grown in the soil take up selenium and consumption of these plants by humans is

34 18 the major source of organic selenium. Inorganic selenium also is readily available as an over-the-counter supplement. Selenium Metabolism Figure 1.4. Metabolic pathways of selenium. Selenium is metabolized to selenide before incorporation into proteins as selenocysteine through an unusual mechianism. Selenide not incorporated into selenocysteine is methylated and excreted (Combs 2004).

35 19 Because selenium has multiple oxidation states, its metabolism can play an important role in physiological functions. It is unusual in that it can be taken up into amino acids replacing sulfur, forming biologically relevant organic compounds. Dietary selenium, both inorganic and organic, is readily absorbed in the gastrointestinal tract, where it is then rapidly distributed to various organs. As seen in Figure 1.4, inorganic selenium taken up as selenate (SeO 2-4 ) or selenite (SeO 2-3 ) undergoes reductive metabolism catalyzed by GSH to selenodiglutathione (GSSeSG) and selenoglutathione (GSSeH), then to selenide (H 2 Se) in a reaction catalyzed by GR. Selenide is then converted to either selenocysteine or conjugated for excretion. Selenide can also be formed from the -lyase metabolism selenomethionine, commonly found in food sources of plant origin, although inorganic selenium is more readily available for incorporation into selenoproteins. Selenide not incorporated into selenoproteins is methylated for excretion via breath (CH 3 SeCH 3 ) or urine ((CH 3 ) 3 Se + ). For incorporation into proteins, selenide is metabolized to selenophosphate by selenophosphate synthetase, which is then incorporated into selenocysteine through an unusual mechanism (Tamura et al. 2004). Selenocysteine is coded for by a RNA stemloop sequence called the selenocysteine insertion sequence (SECIS) element. The SECIS element is found immediately following what would normally be the terminating UGA codon in the 3 untranslated region of the selenoprotein mrna. This is then recognized by a specialized trna charged with the amino acid serine (seryl-trna). The seryl residue in the trna is then converted to selenocysteine residue by the enzyme selenocysteine synthase, which is then translated into selenocysteine. Selenocysteine, commonly referred to as the 21 st amino acid, is incorporated into the active sites of several antioxidant enzymes, including the selenium dependent form of glutathione peroxidase (SeGPx) and thioredoxin reductase (TrxR) (Berry et al. 2001).

36 20 Glutathione Peroxidases (1) 2 GSH + H 2 O 2 GS SG + 2H 2 O (2) GS SG + NADPH + H + 2 GSH + NADP + Figure 1.5. Reaction catalyzed by selenium-dependent glutathione peroxidase (SeGPx). SeGPx (1) catalyzes the reduction of hydrogen peroxide to water with the oxidation of glutathione as a cofactor. Glutathione is returned to its reduced state by a reaction (2) catalyzed by glutathione reductase. The most studied selenoproteins are a family of enzymes called the glutathione peroxidases (GPx). The selenium-dependent GPxs (SeGPx) have selenocysteine at the active site, where it is oxidized in a reaction with hydrogen peroxide before being reduced by glutathione (Figure 1.5). This reaction catalyzes the reduction of hydrogen peroxide to water preventing reactions with and damage to cellular lipids, while releasing oxidized glutathione as a byproduct (Rotruck et al. 1973). Of the eight GPx isoforms currently known, five are known to be selenium-dependent. GPx1, also called the classical or cytosolic GPx, was the first selenoprotein identified and the most studied of these isoforms, is ubiquitously found in tissues including the liver (Brown and Arthur 2001). The others are the GPx2, known as gastrointestinal GPx; GPx3, an extracellular GPx found in plasma and the renal proximal tubular epithelial cells in the kidneys (Avissar et al. 1994); GPx4, a phospholipid hydroperoxide GPx with a preference for lipid hydroperoxides found in spermatogenic cells (Calvin et al. 1987; Weitzel et al. 1990); GPx6, an olfactory GPx found only in humans but has a cysteine-homologue in animals. During selenium deficiency, a hierarchy of GPx exists, in which GPx2 remains the most stable, followed by GPx4, with GPx1 and GPx3 most affected by selenium deficiency (Brigelius-Flohe 2006). While only GPx4 is essential during embryonic development, loss of GPx activity has been associated with increased oxidative stress and carcinogenesis (Khan et al. 2010; Muller et al. 2007).

37 21 Thioredoxin Reductase Figure 1.6. Cellular reactions of thioredoxin reductase. With an electron donated from NADPH, TrxR reduces Trx, which functions to keep cellular proteins and redox status in a reduced state. Source: Holmgren and Lu (2010). Another selenoprotein with antioxidant properties is the thioredoxin reductase (TrxR). As shown in Figure 1.6, TrxR catalyzes reduction of oxidized thioredoxin with an electron from NADPH, with flavin adenine dinucleotide (FAD) serving as a cofactor. Thioredoxins (Trx) are cysteine-containing proteins with the function of keeping cellular proteins, including ribonucleotide reductases, peroxiredoxins, and methionine sulfoxide reductases, in a reduced state (Arner and Holmgren 2000). Because of the presence of thiols, Trx has also been associated with antioxidant activities against ROS. Three forms of TrxR exist with its own gene, functioning to reduce a corresponding Trx isoform: TrxR1, the cytosolic TrxR that reduces Trx1; TrxR2, the mitochondrial TrxR that reduces Trx2; TrxR3, the testes-specific TrxR that reduces Trx3 in the mitochondria (Arner 2009). Unlike GPx, TrxR and Trx are essential for keeping redox status in a reduced state. Selenium in Disease Selenium was initially thought to be a toxin due to toxicity from contaminations (Lenz and Lens 2009). Populations in selenium-rich or contaminated areas are susceptible to selenosis caused by consumption of plants with high selenium concentrations. Acute selenosis, while rarely fatal, can lead to nausea, vomiting and diarrhea, and

38 22 cardiovascular effects (USHHS 2003). Chronic selenosis can result in dermal and neurological symptoms as severe as paralysis (Fordyce 2007). Selenium contamination also poses a threat to wildlife, as selenium can bioaccumulate up the food chain (Wu 2004). However, selenium s role as an antioxidant and its deficiency in diseases eventually led to its recognition as an essential trace element. The distribution of selenium is unusual, such that particular regions in Brazil and China are both selenium deficient and seleniferous (Dhillon and Dhillon 2003). First discovered in farm animals, selenium deficiency and mutations or polymorphisms in selenoprotein genes and synthesis cofactors have been implicated in a variety of diseases, including muscle and cardiovascular disorders, immune dysfunction, cancer, neurological disorders and endocrine function (Bellinger et al. 2009; Burk et al. 1980). The most widely-known consequence of severe selenium deficiency is a potentially fatal form of cardiomyopathy, known as Keshan disease, named after a selenium-deficient province in China where these effects were first discovered (Thomson 2004). Selenium deficiency also causes immunodeficiency such that susceptibility to viral infection is increased (Beck et al. 2003). Fortunately these effects are reversible by supplementation. However, caution must be exercised with regards to supplementation due to the narrow therapeutic range of selenium, estimated to be between 50 g to 350 g per day for an average adult. In healthy adult humans, selenium is found in the liver at about g/g wet weight (Zachara et al. 2001). Selenium and Carcinogenesis Extensive research has been focused on selenium and its effects on carcinogenesis and chemoprevention. Because SeGPx and TrxR are selenoproteins functioning as antioxidants, the potential of selenium supplementation has been investigated in chemoprevention. However, this topic remains to be one of great controversy largely due to studies yielding inconclusive results and the complexities of carcinogenesis. That

39 23 being said, a large number of epidemiological studies have found an inverse relationship between selenium and cancer risks (Glauert et al. 2010). In populations from regions with high soil selenium levels, mortality rates of various cancers, including gastrointestinal, peritoneum, lung, breast, and even lymphomas were found to be lower (Shamberger 1970; Shamberger et al. 1973). Based on the known antioxidant properties of liver selenoproteins (Combs et al. 2001), studies have been carried out on the effectiveness of selenium supplementation on reducing liver cancer, with some success. Supplementation trials in China using inorganic selenium as a source saw the most success in reducing both hepatocellular carcinoma (HCC) and viral hepatitis, the major precursor to HCC (Yu et al. 1999; Yu et al. 1991). Supplementation studies with organic selenium found success in protecting against colon, lung, and prostate cancers (Clark et al. 1996). However, studies of serum selenium levels have so far been inconclusive. While low serum selenium levels have been associated with increased risk of certain cancers, including GI, thyroid, prostate, cervical, and skin (Brooks et al. 2001; Glattre et al. 1989; Willett 1986), several studies found no significant association of serum selenium with cancer risk (Coates et al. 1988; Nomura et al. 1987). Numerous in vivo animal studies have examined the potential of dietary selenium in chemoprevention, and while the results have been largely positive, the mechanisms by which this occurs remain unclear. Early selenium supplementation studies showed promise in reducing incidences of liver carcinogenesis (Combs and Gray 1998; Nyandieka and Wakhisi 1993). However, studies of selenium on the three stages of carcinogenesis (initiation, progression, and promotion) individually have yielded inconclusive results. Studies into selenium s effects on promotion with a known carcinogenic initiator, such as dimethylbenz(a)anthracene (DMBA) and diethylnitrosamine (DEN), have found selenium supplementation to have limited ability in inhibiting DNA adduct formation (Glauert et al. 2008; Ip and Daniel 1985; Thirunavukkarasu et al. 2004) In contrast, another study using DEN as the initiator and

40 24 phenobarbital as the promoter found selenium supplementation during either of these stages had no effects on carcinogenesis (Aquino et al. 1985; Dorado et al. 1985). Making matters even more unclear, increased carcinogenesis was associated with higher selenium levels in some studies (LeBoeuf et al. 1985; Stemm et al. 2008; Wycherly et al. 2004). With oxidative stress implicated in carcinogenesis, studies have attempted to determine the effects of selenium supplementation on the activities of selenoenzymes and carcinogenesis. Glauert et al. (1990) found that while selenium supplementation did increase serum and liver GPx activity in initiated rats, this had no effect on oxidative stress markers TBARS and conjugated dienes. In a later study, selenium supplemention either before or during initiation and promotion stages of hepatocarcinogenesis was found to be effective in reducing hepatic lipid peroxidation either in the hepatoma or in the normal liver tissues (Thirunavukkarasu and Sakthisekaran 2001). Selenium supplementation has also been shown to increase hepatic TrxR activity, but without an increase in TrxR protein synthesis (Berggren et al. 1999). However, despite its known antioxidant properties, there is also increasing evidence to suggest that TrxR actually plays a role in the promotion of carcinogenesis (Jackson-Rosario and Self 2010). These findings suggest that while selenium supplementation has shown promising results in chemoprevention, these results are largely dependent on the exact mechanisms of carcinogenesis, and the exact role of selenium in these mechanisms remains unclear. Copper Copper is a transition metal with a distinct red-brown color, with recorded use dating as far back as 9000 BC. Copper belongs in Group 11 of the Periodic Table, and is chemically related to gold and silver, two highly valued transition metals. In modern times, copper is highly valued because of its high malleability and ductility, and used commonly as piping and wiring for heat and electricity conduction. Copper is also popular as an alloying material, where a combination of copper and tin forms bronze.

41 25 While copper can occur in multiple oxidation states, it is rare in that it can be found in large amounts in its native form without being part of a compound. Copper production from copper ores goes through a process of roasting, converting or leaching, and electrolytic refining (Ellingsen et al. 2007). Alternatively, copper recycling has also become popular in recent years due to rising costs of production. The essentiality of copper as a micronutrient was first demonstrated in an animal study (Hart et al. 1928), where anemia caused by a copper-deficient diet was reversed by copper supplementation. Copper can be found a variety of sources in the human diet, including organ meats, fish, fruits, cereal, nuts, and vegetables (WHO 1998). Copper Metabolism In humans, copper from dietary sources is absorbed in the small intestines, where in mucosal cells it is sequestered in metallothionein or bound to glutathione and albumin in blood for distribution (Pena et al. 1999; Tapiero et al. 2003). This process requires ATP7A, an extrahepatic copper transporting ATPase. Blood copper is primarily transported to the liver, which controls the homeostasis of copper. Most of the copper transported to the liver is then incorporated into ceruloplasmin and transported back into the blood for distribution to extrahepatic organs (Gitlin 1998). Copper remaining in the liver is taken up into cells and shuttled to appropriate intracellular compartments for incorporation into copper proteins. Copper is required for several enzymes essential for proper metabolic function, including cytochrome c oxidase (complex IV of the electron transport chain in the mitochondria), the iron and copper transporter ceruloplasmin, and the antioxidant superoxide dismutase (SOD) (Jaiser and Winston 2010). Copper is delivered to those intracellular compartments of these proteins by highly specific copper chaperones (Field et al. 2003). Copper not incorporated into proteins is bound quickly by metal scavengers such as glutathione, metallothionein, and metallochaperones (Huffman and O'Halloran 2001). Intracellular copper levels are tightly controlled by these

42 26 scavengers, keeping free copper at low levels due to their redox reactive nature (Rae et al. 1999). As seen in Figure 1.7, cuprous (Cu +1 ) and cupric (Cu +2 ) can participate in a Fenton-like redox reaction, which in the presence of hydrogen peroxide can result in the generation of reactive hydroxyl radicals (Gaetke and Chow 2003). (1) Cu 1+ + H 2 O 2 Cu 2+ + OH + OH (2) Cu 2+ + H 2 O 2 Cu 1+ + OOH + H+ Figure 1.7. Fenton-like reactions of free copper. Free cuprous (1) and cupric (2) copper reacts with hydrogen peroxide, resulting in the formation of reactive oxygen species. As the liver maintains primary homeostatic control of copper levels, the majority of excess copper bound for excretion is mainly eliminated in feces through bile, and a small amount in urine through the kidneys (Wijmenga and Klomp 2004). Copper efflux into the bile requires a hepatic transporter ATPase, ATP7B. In the presence of excess copper, ATP7B is moved from the trans-golgi network to a post-golgi compartment closer to the biliary canaliculi, where it facilitates the excretion of excess copper into the bile (Langner and Denk 2004). Hepatocytes can also sequester excess copper in lysosomes, where it can be exocytosed and excreted into bile (Gross et al. 1989). Excess copper that reaches the kidneys is usually reabsorbed following ultrafiltration in the renal tubuli, and thus rarely excreted through urine (Linder and Hazegh-Azam 1996). Ceruloplasmin Ceruloplasmin is the major copper-carrier in the blood, bound to approximately 70% of total copper in human plasma, for distribution from the liver to extrahepatic organs (Waldmann et al. 1967). Ceruloplasmin is synthesized in the liver, each containing six atoms of copper. Copper is required for the stability of the enzyme as apoceruloplasmin is mainly degraded in the liver or released into circulation with a much shorter half-life (de Bie et al. 2007). Aside from copper distribution, ceruloplasmin is also

43 27 a ferric-oxidase playing a key role in iron metabolism. Ferrous iron (Fe 2+ ) in mucosal or parenchymal cells is oxidized by ceruloplasmin to ferric iron (Fe 3+ ), in a reaction dependent on copper (Roeser et al. 1970). This step in iron metabolism is necessary for iron transport, as the iron carrier transferrin can only carry ferric iron. Cytochrome C Oxidase Cytochrome c oxidase (COX), also known as Complex IV, is a mitochondrial integral membrane protein and the last enzyme of the electron transport chain, essential for proper mitochondrial and metabolic functions. COX is comprised of four subunits, with electron transport occurring from subunit II to subunit I. Electrons from cytochrome c are transferred to oxygen, resulting in the formation of water and the translocation of protons across the membrane. This process generates a transmembrane difference of proton electrochemical potential that is required for the synthesis of ATP by ATP synthase. COX is a large protein comprised of 13 subunits from the mitochondria and nucleus, including two copper centers, Cu A and Cu B (Tsukihara et al. 1995). An electron is passed to the Cu A binuclear center from cytochrome c, where it is then passed through cytochrome a to the Cu B binuclear center with cytochrome a 3 where oxygen reduction occurs. Superoxide Dismutase (1) (Cu 2+ )SOD + O 2 (Cu 1+ )SOD + O 2 (2) (Cu 1+ )SOD + O 2 + 2H + (Cu 2+ )SOD + H 2 O 2 Figure 1.8. Reactions catalyzed by CuZnSOD. Superoxide is dismutated to oxygen (1) or hydrogen peroxide (2) by SOD, depending on the redox state of metal cation at the active site. Superoxide Dismutase (SOD) is a metal-containing antioxidant found in nearly all cells that consume oxygen. SOD catalyzes the dismutation of reactive superoxide radicals

44 28 to oxygen and hydrogen peroxide (Figure 1.8). The reaction catalyzed by SOD is extremely rapid such that it outcompetes other reactions of superoxide, preventing it from reacting with cellular membranes and making it a key cellular antioxidant. SOD is the most catalytically efficient of any known enzyme (~7 x 10 9 M 1 s 1 ), with the reaction only limited by the amount of superoxide coming in contact with the enzyme (Löffler et al. 2006). Multiple forms of SOD exist, and they are identified according to the metal they are associated with. The most common SOD used by eukaryotes is SOD1, known as copper-zinc SOD (CuZnSOD). CuZnSOD is a cytosolic-soluble enzyme, found mainly in the cytoplasmic and mitochondrial intermembrane space. Another vital SOD is SOD2, commonly known as manganese SOD (MnSOD), is found in the mitochondria of eukaryotes. Because the electron transport chain and cellular respiration occurs in the mitochondria, it is the main site of cellular oxygen consumption and thus susceptible to superoxide generation from incomplete reduction of oxygen. The third form of SOD is SOD3, is the extracellular form of CuZnSOD in eukaryotes. While loss of SOD3 results in increased sensitivity to hyperoxic injury it rarely results in increased mortality, loss of SOD1 results in oxidative stress and increases likelihood of developing hepatocellular carcinoma, and loss of SOD2 is generally fatal during development (Elchuri et al. 2005; Li et al. 1995; Sentman et al. 2006). Copper in Disease Because copper deficiency or toxicity can result in severe disease, copper homeostasis is tightly regulated by the body through absorption, distribution, storage, and excretion. In healthy adult humans, copper is found at about 3-10 g/g tissue wet weight in the liver (Ellingsen et al. 2007). Copper deficiency, while a rare disease, has significant consequences. Because of the role ceruloplasmin plays in iron metabolism, copper deficiency causes impaired iron transport and results in a variety of hematological diseases. The most common of these are anemia, leucopenia, and neutropenia

45 29 (Halfdanarson et al. 2008). Neurological effects, including myelopathy, peripheral and optic neuropathy, are suspected to be caused by disruption of the electron transport chain (Jaiser and Winston 2008). Copper deficiency can have a variety of causes, from competitive inhibition by zinc to genetic disorders. In particular, research has focused on the disruption of the ATP7A, also known as the Menkes disease gene (Menkes et al. 1962). Hereditary X-linked recessive mutation of this gene prevents proper transport and distribution of copper, and subsequently preventing the proper synthesis and function of copper-containing enzymes (Kim et al. 2002). Menkes disease is generally fatal in early development, because copper cannot cross the blood-brain barrier, preventing the proper function of essential enzymes including cytochrome c oxidase and SOD in the brain (Horn et al. 1992). Menkes disease also has the paradoxical effect of causing copper accumulation in organs other than liver and brain (Horn and Tumer 2002). Treatment options for Menkes disease are limited due to the lethality of the disease, and only in recent times have direct subcutaneous or intravenous injection of copper salts shown some promise (Kaler et al. 2008). Another well-studied genetic disease disrupting copper homeostasis is Wilson s disease, named after Samuel Alexander Kinnier Wilson (Compston 2009), who first described the pathological changes associated with the disease that was eventually linked to a mutation of ATP7B causing copper accumulation. In contrast to Menkes disease, Wilson s disease primarily affects the liver and the central nervous system. ATP7B is involved in the incorporation of copper into ceruloplasmin, and in Wilson s disease this causes apoceruloplasmin to be released. Because excess copper cannot be effectively effluxed from liver to bile for excretion, it accumulates to levels of toxicity, resulting in severe liver pathology. The most common is chronic active hepatitis, which eventually progresses to fibrosis and cirrhosis, which in severe cases can cause hepatocellular carcinoma (Ala et al. 2007). Liver pathology from Wilson s disease is attributed mainly to the Fenton-like reactions of free copper that results in the generation of ROS.

46 30 Neuropathology has also attributed to the redox reactive nature of free copper, which can enter the blood and reach the brain in the absence of ceruloplasmin. Free copper deposition in the basal ganglia results in the degeneration of cognitive and motor functions, and is associated with psychobehavioral changes (Lorincz 2010). With neurologic effects present, Wilson s disease can be diagnosed by Kayser Fleischer rings caused by copper deposition around the iris (Merle et al. 2007). Renal and endocrine disruption has also been observed in Wilson s disease (Pandit et al. 2002; Prohaska 1986). Multiple treatment options have been used for Wilson s disease. For managing the disease, the simplest treatment is by using diets low in copper, while chelators such as penicillamine have also been used in conjunction with zinc, a competitive inhibitor of copper (Walshe 1996). Liver transplantation remains an option as a cure for the disease as well. Copper toxicity can also occur as a result of non-genetic causes. These incidences usually occur as a result of using uncoated copper cooking utensils, or excess copper in drinking water or other environmental sources. Acute copper toxicity effects are mainly hematological, such as vomiting blood, hypotension, and hemolysis (Mendel et al. 2007). Chronic non-genetic copper toxicity exhibits similar liver, kidney, and neurologic symptoms as Wilson s disease with the free copper suspected as the culprit. Increased free copper levels in the brain have also been associated with increased risk of Alzheimer s disease (Hureau and Faller 2009). There is also evidence suggesting that free copper can act as a mutagen due to its ability to cause oxidative stress. Animal studies have shown an increase in genomic instability (micronuclei and chromosomal aberrations) after the administration of copper (Agarwal et al. 1990; Bhunya and Jena 1996). These results have been supported by in vitro studies, where DNA strand breaks and synthesis disruption have been observed following copper administration (Sideris et al. 1988; Sirover and Loeb 1976). Few epidemiological studies involving copper exposure are available, and carcinogenesis observed in those studies is generally

47 31 attributed to other causes (Lightfoot et al. 2010; Verma et al. 1992). As a result, neither the EPA nor WHO currently considers there to be adequate evidence to link copper to carcinogenesis (IRIS 2011). N-Acetylcysteine N-Acetylcysteine (NAC) is an analogue of the amino-acid cysteine, so named because of the acetyl group attached to the nitrogen atom. In medicine, NAC is widely used to stimulate production of cysteine and more importantly glutathione (GSH). GSH is an essential endogenous cellular antioxidant made up of the amino acids cysteine, glutamate, and glycine, found ubiquitously in the human body (Ballatori et al. 2009). The thiol group in cysteine acts as an electron donor, keeping cytoplasmic proteins in a reduced state. GSH can react directly with reactive intermediates such as ROS, or act as a cofactor in biotransformation, such as GPx. This process results in the oxidation of GSH, which forms a disulfide bridge with another molecule of GSH to form glutathione disulfide (GSSG). Under normal conditions, most cellular GSH is kept in its reduced state by glutathione reductase (GR). A high level of GSSG relative to GSH is often an indicator of oxidative stress (Pastore et al. 2003). GSH can also be conjugated to electrophilic compounds, in either non-enzymatic reactions or ones catalyzed by glutathione-s-transferases (GSTs), with these GSH-conjugates excreted. GSH is replenished by de novo synthesis, with cysteine, an essential amino acid, as the limiting factor (Jahoor et al. 1999; Taniguchi et al. 1989). However, direct supplementation of GSH proved to be troublesome, due to difficulty in absorption in the GI tract. NAC was introduced in the 1960s as a mucolytic agent in treating chronic pulmonary diseases (Flanagan and Meredith 1991). The purpose of NAC was to split disulfide bonds of glycoproteins proteins in mucus, thus reducing mucus viscosity (Aitio 2006). At that point in time, there were reports of liver injury from overdose of acetaminophen, also known as paracetamol outside the United States (Davidson and

48 32 Eastham 1966). The cause of these injuries was determined to be caused by depletion of liver GSH, due to its conjugation to a toxic metabolite of acetaminophen called N-acetylp-benzoquinone imine (NAPQI) (Mitchell et al. 1973). Subsequent treatment with compounds known for replenishing GSH, methionine and cysteamine, were successful in preventing these injuries (Prescott et al. 1974; Prescott et al. 1976). However, the side effects of these treatments, including flushing and vomiting, led to the search for an alternative. By 1977, that alternative was NAC (Prescott et al. 1977). NAC can be administered orally, intravenously, or as an inhalable aerosol (Atkuri et al. 2007). NAC Metabolism Figure 1.9. Glutathione (GSH) synthesis from N-acetylcysteine (NAC). GSH synthesis from NAC requires the conversion of NAC to cysteine in the liver, the addition of glutamate catalyzed by glutamate cysteine ligase, and the addition of glycine catalyzed by glutathione synthatase. Since cysteine occurs at the lowest intracellular concentration, its availability is the rate limiting factor in glutathione synthesis. Direct supplementation with cysteine is not a viable option, mainly due to its reactivity. Because of the acetyl group, not only is NAC far less reactive and toxic, it is also more soluble in water making it a more suitable supplement (Bonanomi and Gazzaniga 1980). The majority of NAC administered ends up being supplied to the liver and kidneys, where it is eventually deacetylated to cysteine and used for GSH synthesis (Holdiness 1991). As seen in Figure 1.9, GSH synthesis is a multi-step process that occurs in the liver, beginning with cysteine from endogenous

49 33 sources or NAC (Meister 1974). Gamma-glutamylcysteine is formed from cysteine and glutatmate in an ATP-dependent reaction catalyzed by glutamate cysteine ligase (GCL), also known as gamma-glutamylcysteine synthetase. In another ATP-dependent reaction catalyzed by glutathione synthetase, glycine is added to the C-terminus of gammaglutamaylcysteine. NAC in Disease Figure Acetaminophen metabolism. Acetaminophen that undergoes Phase II biotransformation form metabolites that are excreted, while Phase I biotransformation by CYP results in formation of N-acetyl-pbenzoquinonemine (NAPQI), a toxic metabolite that will cause toxicity if not conjugated by GSH for excretion. Adapted from James et al. (2003). Clinically, NAC was first used as a mucolytic agent in treating respiratory obstructions, from chronic bronchitis to chronic obstructive pulmonary disorders (COPD) (Millar et al. 1985; Sadowska et al. 2006). Mucus is secreted by mucus glands to protect epithelial cells from pathogens. Mucus is a viscous colloid containing antiseptic enzymes

50 34 and disulfide bond-containing glycoproteins. Increased mucus production in the respiratory system is a common protective reaction to respiratory infections such as the common cold or influenza. While this generally causes only mild discomfort, more serious respiratory illnesses involving inflammation, such as allergies, asthma, and chronic bronchitis, cause hypersecretion of mucus that can cause respiratory obstruction. Treatment with NAC was shown to prevent the deterioration of lung function, particularly in elderly patients (Lundback et al. 1992). NAC has also been used in treating the respiratory symptoms of the genetic disease cystic fibrosis with some success (Tirouvanziam et al. 2006). Lungs from those suffering from cystic fibrosis have reduced mucociliary clearance, resulting in mucus buildup and inflammation, which if left untreated will likely result in death (Flume et al. 2010). While NAC is useful as mucolytic agent, it is far better known as a GSH precursor and an antidote in acetaminophen poisoning. When used in therapeutic doses, acetaminophen is an over-the-counter drug commonly used for pain relief. Acetaminophen overdose is the most common cause of acute liver failure in the United States (Chun et al. 2009). The toxicity of acetaminophen lies in its different metabolic pathways (Figure 1.10). At the therapeutic dose, the majority of acetaminophen undergoes biotransformation by glucuronidation or sulfation conjugation to non-toxic metabolites that are excreted. Approximately 5% of acetaminophen undergoes biotransformation by CYPs 2E1 and 3A4 to the reactive metabolite NAPQI (James et al. 2003). With adequate GSH present, NAPQI is quickly detoxified to an inactive conjugate bound for excretion. During overdose, the sulfation and glucuronidation pathways become saturated, leaving CYP biotransformation to NAPQI. When de novo synthesis of GSH cannot keep up with its depletion, it results in unconjugated NAPQI reacting with cellular proteins and forming adducts. Without NAC supplementation, hepatotoxicity occurs when GSH is depleted 70% below normal physiological levels, which can happen

51 35 within 48 hours (Richardson 2000). Patients treated with NAC within 48 hours are usually expected to recover fully. Recent research has focused on the nephroprotective properties of NAC. Diagnostic procedures often require the use of iodinated contrast media which can result in acute nephropathy. The mechanisms behind this are currently under investigation, but it has been linked to oxidative injury caused by ischemia. Nephropathy resulting from the use of contrasting agent in radiographic examination is of real concern in patients with predisposing risk factors. These are patients with conditions such as diabetes that compromised renal function and have increased creatinine levels. In clinical trials, patients with compromised renal function have been supplemented with NAC have had some success in reducing nephropathy (Anderson et al. 2011; Stenstrom et al. 2008). However, there have also been studies that found NAC to be ineffective in reducing nephropathy (Fishbane 2008). Although NAC continues to be used in the high risk patients with pre-existing compromised renal function, the inconclusive results and the poorly understood mechanisms have prevented NAC from being declared a cure for this condition. NAC also shown promise in prolonging the lifespan of those infected with human immunodeficiency virus (HIV). HIV-positive patients generally have lower cysteine and GSH status, and thus increasing their risk for oxidative stress (Buhl et al. 1989; Halliwell and Cross 1991). Oral NAC supplementation of these patients replenished their GSH status (Herzenberg et al. 1997). In vitro studies suggest that NAC is effective in preventing cytotoxicity, likely caused by ROS, to immune cells from HIV-positive patients (Roberts et al. 1995). These results were supported by clinical trials where NAC supplementation slowed the decline of lymphocyte count of HIV-positive patients (Akerlund et al. 1996). Studies have also shown NAC to slow the transcription HIV, through NAC s ability to regulate nuclear transcription factor kappa B (NFKB) (Staal et

52 36 al. 1990). It is likely that NAC supplementation prolongs the lifespan of HIV-positive patients (Droge et al. 1992). The scavenging ability of the thiol group in NAC is also of interest. Like the thiol group in cysteine and GSH, there is evidence suggesting that the thiol group in NAC is capable of directly facilitating ROS detoxification. While radical scavenging is naturally of interest, there is also evidence that NAC is capable of improving cellular energy metabolism. Zwingmann and Bilodeau (2006) showed in an animal study that NAC can act independently of GSH. NAC was found to stimulate flux of pyruvate through pyruvate dehydrogenase instead of pyruvate carboxylase. The result of this prevents the do novo synthesis of GSH and increases acetyl-coa production, resulting in more efficient mitochondrial energy metabolism. In addition, NAC increased the formation of hypotaurine, a precursor of taurine. In the liver, taurine is known to protect against fatty liver and reduce cirrhosis (Kerai et al. 1998). These findings are promising for the use of NAC in treatment of non-acetaminophen induced liver injury. However, like any drug, NAC is not free of adverse side effects. NAC is a compound with an unpleasant smell and taste, such that oral supplementation commonly results in vomiting. About 5% of those reactions to oral NAC supplementation are so severe such it would require a switch to intravenous administration (Yip et al. 1998). Anaphylactic reactions, including rash, pruritus, angioedema, bronchospasm, tachycardia, and hypotension, are common in intravenous NAC administration, occurring in about 15% of patients within two hours (Heard 2008; Kerr et al. 2005). However, these reactions are rarely serious enough to discontinue NAC administration, although in extremely severe cases can cause fatal respiratory arrest (Reynard et al. 1992). Although adults can be safely given a dose up to 8000 mg/day, errors in dosing children for intravenous NAC can have serious consequences, such as cerebral edema and hyponatremia (Bailey et al. 2004; De Rosa et al. 2000; Sung et al. 1997). While NAC as an over-the-counter supplement can be poses many benefits, in vivo studies of otherwise

53 37 healthy animals given high doses of NAC have increased risk of heart and lung damage caused by hypertension (Palmer et al. 2007). Despite these adverse effects, NAC is largely considered a safe drug when used in its therapeutic range given its many benefits. Specific Aims of Thesis The overall working hypothesis of this work is that PCB 126 exerts its toxic effects by altering the hepatic homeostasis of micronutrients and key antioxidant enzymes, and that these adverse effects will be protected against by dietary selenium supplementation, lowered dietary copper, and NAC supplementation. The next chapter in this thesis shows that PCB 126 can cause significant biochemical and toxic changes at a low dose. This first aim of this these is to determine the hepatic effects that PCB 126 exposure has on selenium and selenoproteins, and the ability of selenium supplementation reduce liver injury caused by PCB 126. Potent AhR agonists such as TCDD and PCB 126 have been shown to diminish the selenium levels in rat livers. In the liver, PCB 126 is expected to shift the balance towards oxidative stress, by increasing production of proxidants and diminishing antioxidants. In chapter two, it was demonstrated that PCB 126 caused significant decreases of hepatic selenium and SeGPx activity. Dietary selenium supplementation is hypothesized to restore hepatic selenium levels and ameliorate the toxicity caused by PCB 126. The second aim is to determine the effects of dietary copper on hepatic toxicity and the role of copper in hepatic toxicity following PCB 126 exposure. In previous experiments with PCB 126, hepatic copper levels were observed to increase a dosedependent manner. Because these findings were not accompanied by increases in CuZnSOD activity, suggesting that increased hepatic copper may be non-enzymatic. Although copper also has a role in vital cellular processes and antioxidant enzymes, high levels of free copper can become a strong prooxidant by participating in Fenton-like

54 38 reactions that result in the generation of hydroxyl radicals. It is hypothesized that reduced dietary copper will protect against PCB 126-induced toxicity by diminishing the ability of the liver to accumulate free copper. The final aim is to examine the ability of NAC to ameliorate the hepatic toxicity following exposure to PCB 126, and to determine the effects of the alteration to the redox status on hepatic toxicity. NAC is a precursor to GSH, a key intracellular antioxidant. Because GSH is poorly absorbed when orally ingested, alternatives such as NAC have been successfully used to increase intracellular GSH levels instead. In addition, NAC can also function directly as a sulfhydryl group donor. Exposure to AhR agonists have been shown to reduce hepatic GSH levels, indicating disruption to the hepatic redox status. NAC supplementation is hypothesized to restore hepatic glutathione status and prevent toxicity caused by PCB 126 exposure.

55 39 CHAPTER II ACUTE TOXICITY OF 3,3,4,4,5-PENTACHLOROBIPHENYL (PCB 126) IN MALE SPRAGUE DAWLEY RATS: EFFECTS ON HEPATIC OXIDATIVE STRESS, GLUTATHIONE AND METALS STATUS 1 Abstract It is well-documented that PCBs may induce drug metabolizing enzymes leading to the bioactivation of PCBs themselves. Alternatively they may lead to oxidative events within the cell or other biochemical and histological changes. The goal of this present study was to evaluate the effects of a single, very low dose of PCB 126 (3,3,4,4,5- pentachlorobiphenyl), a coplanar, dioxin-like PCB congener and aryl hydrocarbon receptor (AhR) agonist, on the redox status of hepatic glutathione (GSH), metals homeostasis, antioxidant and drug metabolism enzymes, and histology. To examine these parameters, male Sprague-Dawley rats were fed a defined AIN-93 basal diet containing 0.2 ppm selenium for two weeks, then administered a single i.p. injection of corn oil (5 ml/kg) or 1 µmol PCB 126/kg (326 µg/kg) in corn oil two weeks before being euthanized. PCB 126 significantly increased liver weight (42%) and hepatic microsomal cytochrome P-450 (CYP1A) enzyme activities (10-40-fold increase). Hepatic zinc, selenium, and GSH levels were significantly decreased 15%, 30%, and 20%, respectively, by PCB 126. These changes were accompanied by a 25% decrease in selenium-dependent glutathione peroxidase (SeGPx) activity. In contrast, hepatic copper 1 Reproduced with permission from Lai I, Chai Y, Simmons D, Luthe G, Coleman MC, Spitz D, Haschek WM, Ludewig G, Robertson LW. Environment International Nov; 36(8): Copyright Elsevier, Agreement license number: Glutathione analysis was contributed by M.C. Coleman. Trace elements determination was contributed by Y. Chai and D.L. Simmons. Histology was contributed by W.M. Haschek.

56 40 levels were increased 40% by PCB 126. PCB 126-induced pathology was characterized by hepatocellular hypertrophy and mild steatosis in the liver and a mild decrease in thymic cortical T-cells. This controlled dietary study shows that even a single, very low dose of PCB 126 may significantly perturb metal homeostasis and antioxidant and enzyme levels in the rat liver. Introduction Polychlorinated biphenyls (PCBs) were commercially manufactured as industrial mixtures that were known for their stability under a broad range of chemical, thermal, and electrical conditions (Safe 1994). Although production of these chemicals was halted in the 1970s, their resistance to breakdown and their lipophilicity allows them to biomagnify in the food chain and persist in the environment (Hansen 1987). Among the 209 possible PCB congeners, biological and chemical properties vary greatly depending on the number and placement of chlorine atoms on the biphenyl ring (Ludewig et al. 2007). Congeners with no chlorine atoms adjacent to the biphenyl bridge (in positions 2, 2, 6, 6 ) may assume a more co-planar configuration and are known as dioxin-like PCBs because of their ability to mimic 2,3,7,8-tetrachlorodibenzodioxin (TCDD) and bind to the aryl hydrocarbon receptor (AhR) (Bandiera et al. 1982). Especially higher chlorinated PCBs selectively induce cytochrome P-450 monooxygenases (Parkinson et al. 1983a) that catalyze the oxidation of a broad range of endogenous and exogenous substances but may also result in the generation of reactive oxygen species (ROS) via an uncoupling of the catalytic cycle and a partial reduction of oxygen (Schlezinger et al. 2000; Schlezinger et al. 1999). Cytochrome P-450 (CYP) 1A proteins are induced by dioxin-like PCBs, such as PCB 126, that are potent AhR agonists. It has been reported that a dose of less than 1 mol/kg (275 g/kg) PCB 126 was able to maximally induce CYP1A activity, a biomarker of AhR activation and dioxin-like toxicity (Fisher et al. 2006; Hu et al. 2007). The high efficacy of binding of PCB 126 to

57 41 the AhR (Bandiera et al. 1982) is reflected in its Toxic Equivalency Factor (TEF) of 0.1 relative to 2,3,7,8-tetrachlorodibenzodioxin (TCDD) (Yoshizawa et al. 2007). It has been reported that the PCB mixture Aroclor 1254 (Schramm et al. 1985), as well as congeneric PCBs, decrease the levels of the selenium-dependent form of the antioxidant enzyme glutathione peroxidase (SeGPx). This decrease has been shown to be caused by coplanar chlorinated biphenyls, including PCB 77 and PCB 126 (Hori 1997). The decrease in SeGPx is accompanied by a decrease in message RNA for the enzyme as well as a decrease in hepatic Se levels (Twaroski et al. 2001b). Also, in addition to increasing the production of reactive oxygen species (ROS), PCBs also decrease the antioxidant defenses that are capable of detoxifying ROS. Therefore, one goal of this study was to identify indices as biomarkers of any oxidative stress using the redox couple of glutathione (GSH/GSSG). Selenium (Se) is an essential trace element in the antioxidant defense system with the ability to be encoded into selenocysteine (SeCys), an amino acid incorporated into a key antioxidant enzyme, glutathione peroxidase (SeGPx). SeGPx plays an important role in detoxifying lipid peroxidation-forming hydrogen peroxides, and loss of SeGPx has been shown to be associated with increased oxidative stress (Condell and Tappel 1983; McCray et al. 1976; Muller and Pallauf 2002; Pohjanvirta et al. 1990). Here we examined the congener specific toxicity of a dioxin-like PCB congener, PCB 126, by examining its ability to alter hepatic redox status. We hypothesized that a low dose of PCB 126 (1 μmol/kg) will disrupt GSH, Se, and SeGPx homeostasis, increasing the risk of oxidative stress. Materials and Methods Chemicals All chemicals were obtained from Sigma-Aldrich Chemical Company (St. Louis, MO) unless otherwise stated. PCB 126 (3,3,4,4,5-pentachlorobiphenyl) was prepared by

58 42 an improved Suzuki-coupling method of 3,4,5-trichlorobromobenzene with 3,4- dichlorophenyl boronic acid utilizing a palladium-catalyzed cross-coupling reaction (Luthe et al. 2009). The crude product was purified by aluminum oxide column, and flash silica gel column chromatography and recrystallized from methanol. The final product purity was determined by GC-MS analysis >99.8% and identity confirmed by 13C NMR. Caution: PCBs and their metabolites should be handled as hazardous compounds in accordance with NIH guidelines. Animals This animal experiment was conducted with approval from the Institutional Animal Care and Use Committee of the University of Iowa. Male Sprague-Dawley rats ( g) (Harlan Sprague Dawley, Indianapolis, IN) were housed in hanging metal mesh cages in a controlled environment maintained at 22 o C with a 12 hour light-dark cycle. Rats were fed a defined AIN-93 basal diet containing 0.2 ppm sodium selenate (Table 2.1; Harland Teklad, Madison, WI) and water ad libitium. After two weeks on the diet, rats were administered a single i.p. injection of vehicle (stripped corn oil; Acros Chemical Company, Pittsburgh, PA) or vehicle with PCB 126 (1 µmol/kg body weight; 326 µg/kg body weight), and were euthanized 2 weeks later by carbon dioxide asphyxiation and cervical dislocation. Livers were excised and apportioned as follows: 1) Approximately 0.5 g liver was immediately homogenized in 5% 5-sulfosalicylic acid (5-SSA) (w/v) (Anderson 1985; Meister and Anderson 1983) and was used for the determination of GSH and GSSG as described below. 2) Representative slices of liver, spleen, and thymus were placed in 10% buffered formalin for histologic analysis as described below. 3) Portions of liver and kidney, approximately 2.0 g, were taken and frozen in acidwashed HDPE scintillation vials for metal analysis.

59 43 4) The remaining liver tissue was homogenized in ice-cold 0.25 M sucrose solution containing 0.1 mm EDTA, ph 7.4. The homogenate was centrifuged at 10,000g for 20 min. The resulting supernatant was then centrifuged at 100,000g for 1 h. The supernatant containing the cytosolic fractions were aliquoted. Microsomal pellets were resuspended in ice-cold sucrose/edta solution. Protein concentrations were determined by the method of Lowry et al. (1951). Glutathione (GSH/GSSG) Analysis GSH and GSSG levels were determined in hepatic liver tissue by the methods of Griffith (1980) and Anderson (1985), based on an enzyme recycling assay using glutathione reductase and 5,5'-dithio-bis-(2-nitrobenzoic acid) (DTNB) and NADPH to continuously oxidize and reduce the GSH. Absorbance change at 412 nm was followed in a Beckman DU-670 spectrophotometer for 5 min. The rate of yellow color accumulation is the result of 2-nitro-5-thiobenzoate (TNB) forming from DTNB. This rate is proportional to the amount of total glutathione in the sample. GSSG was measured independently by incubating the tissue in 2-vinylpuridine, which conjugates GSH prior to the measurement of the remaining GSSG. Reduced glutathione is determined by subtracting GSSG from total glutathione. GSH levels are expressed as per mg protein. Trace Elements Determination Metal concentrations in rat tissues were quantitatively determined by Inductively Coupled Plasma Mass Spectrometry (ICP-MS). ICP-MS is selected due to its low detection limits and multi-element capacity (Entwisle and Hearn 2006). Liver tissue was acid-digested with HNO 3 in a closed Teflon vessel in a CEM MARS5 Microwave Digestion System prior to ICP-MS measurement. Metal concentrations in the digested samples were then determined in an Agilent 7500ce ICP-MS equipped with a CETAC AS520 auto sampler.

60 44 Histology Formalin fixed tissues were processed routinely, embedded in paraffin, sectioned at 3-4 m and stained with hematoxylin and eosin for light microscopic examination. Frozen sections of selected liver samples were stained with oil-red-o for the evaluation of lipids. Measurement of CYP1A Activity Cytochrome P450 1A1 and 1A2 activities were estimated in hepatic microsomes by measuring the activities of ethoxyresorufin deethylase (EROD) and methoxyresorufin deethylase (MROD), respectively, as previously reported (Burke and Mayer 1974). Briefly, using ethoxy- or methoxyresorufin as substrate, the monooxygenase reaction by CYP1A results in the formation of the fluorescent resorufin, which was detected spectrofluorometrically using an excitation wavelength of 550 nm and emission wavelength of 585 nm. Measurement of Glutathione Peroxidase (GPx) activities The GPx activities in the cytosolic fraction were determined using the method of Lawrence and Burke (1976). Briefly, SeGPx and total GPx activities were determined using hydrogen peroxide and cumene hydroperoxide as the substrates, respectively. Absorbance change at 340 nm was followed in a Beckman DU-650 for 5 min. One unit of enzymatic activity is defined as the amount of protein that oxidizes 1 μm of NADPH per min, expressed as milliunits per mg protein. Statistics Statistical analysis was performed using ANOVA followed by Bonferroni post hoc to test for differences in treatment means. Treatment groups were considered statistically different at p < 0.001, 0.01, or 0.05.

61 45 Results Effects on Growth and Organ Weights The growth rate of PCB 126-treated rats did not differ significantly from that of vehicle-treated rats (Figure 2.1). In PCB 126-treated rats, both absolute and relative liver weights increased significantly (38% and 42%, respectively), while both absolute and relative thymus weights decreased significantly (46% and 44%, respectively) (Table 2.2). Effects on Total Hepatic GSH and GSSG As indicators of oxidative stress, glutathione levels in the liver were determined. Hepatic levels of both oxidized (GSSG) and reduced glutathione (GSH) were reduced by 20% in PCB 126-treated rats, relative to control levels, reflecting an overall decrease in hepatic total glutathione in the PCB 126-treated rats (Table 2.3). Effects on Trace Elements and GPx Activities PCB 126 caused a 30% decrease in hepatic Se (Figure 2.2) accompanied by a decrease (60%) in hepatic cytosolic SeGPx activity (Table 2.4a). Total GPx activity was only decreased 25% by PCB 126 (Table 2.4a). In the liver of PCB 126-treated rats, copper levels were increased 40% (Figure 2.3), while zinc levels were decreased by 15% (Figure 2.4). By comparison, no significant effects on target metals were observed in the kidneys of PCB 126-treated rats. Histology Histologic examination of livers indicated an increase in size of hepatic lobules due to hypertrophy (enlargement) of hepatocytes in PCB 126-treated rats; mild capsular irregularity was also present. Hepatocellular hypertrophy was characterized by large areas of cytoplasmic pallor (hydropic degeneration), small poorly defined vacuoles and occasional peripheralization of nuclei (Figure 2.5). With Oil-Red-O stain, small lipid globules were present in hepatocytes (mild steatosis) from PCB 126-treated rats but

62 46 virtually absent in controls. In addition, there was mild thymic atrophy in PCB 126- treated rats characterized by decreased cortical to medullary width (Figure 2.6). The cortex had decreased density of T-lymphocytes with scattered macrophages indicating T- lymphocyte apoptosis. No changes were observed in the spleen of PCB 126-treated rats. Effects on EROD and MROD Activities The increase in liver mass in PCB 126 treated rats was accompanied by increase in the hepatic microsomal activities of CYP 1A (ethoxyresorufin O-deethylase, EROD; methoxyresorufin O-deethylase, MROD). EROD and MROD activities were significantly increased 43-fold and 10-fold, respectively, in PCB 126-treated rats (Table 2.4b). Discussion PCBs are known for their variety in mechanisms of toxicity (Ludewig et al. 2007; Silberhorn et al. 1990). PCB toxicity can vary depending upon chlorination pattern and metabolic activation, causing various levels of genotoxicity and oxidative stress (Sadeghi-Aliabadi et al. 2007; Slim et al. 1999). PCBs induce mixed function oxidases and redox reactive oxygen, resulting in oxidative stress (Brown et al. 2007). Oxidative stress caused by PCBs results in lipid peroxidation in hepatocytes and rat liver (Banudevi et al. 2006; Fadhel et al. 2002; Thome et al. 1995), and increased DNA binding activity of the oxidative stress-induced transcription factors, NF- and AP-1 (Lu et al. 2003), leading to oxidative DNA damage (Hassoun et al. 2001). PCB-induced oxidative stress has been shown to promote carcinogenesis in rat livers (Glauert et al. 2008). The toxicity of a specific congener depends upon its chlorination pattern and potential for metabolic activation or detoxification. The chlorination pattern determines the binding activity to various cellular receptors. The most studied PCB congeners are the co-planar, non-ortho substituted PCBs, known as dioxin-like PCBs, because of their structural resemblance to TCDD and ability to bind to the AhR and induce CYP1A enzymes (Hestermann et al. 2000). Dioxin-like PCBs have also been shown to decrease

63 47 the activity of the antioxidant selenoenzyme, glutathione peroxidase, and decrease the mrna for the enzyme and total selenium within the liver (Schramm et al. 1985; Twaroski et al. 2001b). PCB 126 is known as the most toxic of these congeners because of its chlorination pattern and size similarities to TCDD. Thus, the goal of this study was to elucidate the congener-specific potency of a dioxin-like PCB-induced oxidative stress related effects on the hepatic redox status of glutathione (GSH/GSSG) and on hepatic selenium status in the liver. Indeed, PCB 126-treatment produced a significant decrease in total liver glutathione levels. This was not accompanied by an increase in GSSG, the oxidized form of glutathione, suggesting that either the reduction of total GSH is due to covalent binding, for example to a PCB metabolite, or the GSSG was very efficiently reduced back to GSH or excreted from the liver into the bile. Treatment of rats with up to 600 µmol/kg PCB 77 over a 3 week period did not cause a decrease in total GSH (Twaroski et al. 2001b). This indicates again the much stronger liver toxicity of PCB 126 compared to PCB 77. In the above mentioned experiment a highly significant increase in GSH transferase and reductase was observed (Twaroski et al. 2001b). These changes would work towards protection against GSH oxidation and possibly increase the likelyhood of GSH loss by covalent binding. The significant decreases in hepatic levels of zinc and selenium, metals which are considered protective, and components of antioxidant enzymes, such as glutathione peroxidase and thioredoxin reducase, certainly signify an alteration in the redox balance of the cell and perhaps an inability to cope with increased oxidative events associated with PCB exposure, including lipid peroxidation (Fadhel et al. 2002). Likewise the increase in copper may indicate an increased production of ROS, as has been seen in TCDD-treated rats (Elsenhans et al. 1991; Wahba et al. 1988). Notably in the current study, these PCB 126-induced alterations in hepatic levels of zinc, copper and selenium occurred at a dose so low as to not alter the growth of the rats (Figure 2.1).

64 48 This is the first known report about increased copper levels in the liver of rats after PCB-exposure. Nishimura and coworkers reported increased hepatic copper in rats after a single dose of 1 g/kg TCDD and increase 8-oxo-dG levels after 2 g/kg, and reduced GSH levels after 4 g/kg (Nishimura et al. 2001). This increase in Cu level may be due to impaired biliary excretion of Cu, the normal way to maintain Cu homeostasis, based on the impaired biliary flow and excretion by TCDD (Elsenhans et al. 1991; Mahoney et al. 1955). Previous reports about an effect of PCBs on zinc levels in the liver were not found, however, Wahba et al. reported that liver zinc levels were not changed after treatment with TCDD, indicating a possible difference in the mode of action (MOA) between TCDD and PCB 126 (1988). The mechanism resulting in the zinc decrease is not clear, but oxidative stress seems to be the causative agent. Zinc decrease may then be partially responsible for the decrease in antioxidant enzymes such as SeGPx, but again the mechanism is unknown. The results from this study confirm the strong effect of dioxin-like compounds on the liver: two weeks after even a single injection of 1 µmol/kg PCB 126 an enlarged liver, most likely due to hypertrophic hepatocytes and formation of small lipid droplets were observed. This increased liver weight and histologic evidence of hepatocellular degeneration and lipid accumulation point to significant changes taking place in the liver after PCB 126 exposure. Hepatocellular vacuolization has also been described in immature ovariectomized female C57BL/6 mice treated with PCB 126 (Kopec et al. 2008). In the latter study, complementary microarray and clinical chemistry data suggested the lipid accumulation resulted from the disruption of hepatic lipid uptake and metabolism. The observed thymic change in consistent with early atrophy which is a well-known response to PCB exposure; the immune system is among the most sensitive of all organ systems to PCBs.

65 49 Based on our biochemical and histological data, liver hypertrophy, besides the lipid deposition, observed in the PCB 126-treated rats is likely associated with the induction of CYPs. Over 40-fold increase in CYP1A1 (EROD) and ~10-fold increase in CYP1A2 (MROD) were measured after PCB 126 treatment. Compared to this a twiceweekly i.p. injection of 100 µmol/kg of the dioxin-like congener PCB 77 resulted in a similar increase in liver weight, but increased EROD activity of ~35-fold after the first week and ~18-fold after the second week (Twaroski et al. 2001a). The WHO suggested TEF of PCB 126 and PCB 77 are 0.1 and , respectively (Van den Berg et al. 1998). By comparison, 3 nmol/kg TCDD was able to cause a 36-fold and 11-fold induction in EROD and MROD activities, respectively (Santostefano et al. 1999). Transcriptional activation of the AhR gene battery includes the induction of a broad range of phase I and phase II enxymes. These changes in gene transcription may disrupt antioxidant homeostasis (Dostalek et al. 2008; Pereg et al. 2006). Notably CYP 1A, a key enzyme in the biotransformation of xenobiotics, may cause toxicity by producing toxic xenobiotic metabolites, or may increase ROS production through the uncoupling of the CYP catalytic cycle (Lewis and Pratt 1998; McLean et al. 2000; Pereg et al. 2002). Dioxin-like PCBs have also been shown to decrease the activity of the antioxidant selenoenzyme, glutathione peroxidase, and decrease the mrna for the enzyme and total Se within the liver (Schramm et al. 1985; Twaroski et al. 2001b). The results from this study show that PCB 126 may increase oxidative stress by diminishing hepatic SeGPx activity, reducing the liver s capacity to remove hydrogen peroxide. The PCB 126- induced decrease in SeGPx was very pronounced, down by about 60% after only 1 µmol/kg PCB 126. Total GPx activity, which consists of both SeGPx and glutathione transferase (GST) activities, was diminished, but at a lower magnitude to that of SeGPx, suggesting an induction of GST (Prohaska 1980). In contrast, much higher doses of PCB 77 reduced SeGPx activity only by ~20 to 30%, showing the potent toxicity of PCB 126.

66 50 Similar findings of GPx activity decrease have also been observed in rats treated with TCDD with a dose as low as 3 nmol/kg (Stohs et al. 1986; Twaroski et al. 2001b). The loss of SeGPx activity can be linked to the loss of hepatic Se and GSH (Chen et al. 1990), as exposure to PCB 126 significantly diminished both of these two key components of the enzyme reaction. Reduction in hepatic Se levels was also reported previously after treatment with PCB 77 and TCDD (Hassan et al. 1985; Stemm et al. 2008). The mechanism by which hepatic Se is lost following exposure to AhR agonists is currently unknown and under investigation. As noted by Hassan et al. (1985), selenium is thought to be necessary to partially protect from the toxic effects of TCDD, as rats exposed to TCDD that were deficient in selenium had higher levels of hepatic lipid peroxidation. Dietary feeding studies are difficult to compare with those employing other routes of administration. Even if we assume that 100% of the oral dose is absorbed, it is clear that the current study was conducted well above the stated NOAEL of 0.01 µg/kg body weight/day of PCB 126 (Chu et al. 1994), and above the stated LOAEL of 0.74 µg/kg body weight/day (ATSDR 2000). This indicates a broad range of doses, insufficient to cause impairment in growth rates, but sufficient to significantly alter the redox status of the liver, a state reflected by reductions in GSH, selenium and zinc, with concomitant increases in copper. These findings may well prove to be useful biomarkers for PCB exposure.

67 51 Table 2.1. AIN-93 Diet Composition* Constituent g/kg Casein L-Cystine 1.8 Corn Starch Maltodextrin Sucrose Soybean Oil 40.0 Cellulose 50.0 Mineral Mix, AIN-93-MX 35.0 Sodium Selenate (0.1% in sucrose) 0.12 Vitamin Mix, AIN-93-VX 10.0 Choline Bitartrate 2.5 TBHQ, antioxidant *: Se concentration is the diet was confirmed with ICP-MS as 0.21 ppm.

68 52 Table 2.2. Body Weight, Liver Weight, Thymus Weight, and Ratio Final Body Weight (g) Raw Liver Weight (g) Liver/Body X 100 (%) Thymus Weight (g) Thymus/Body X 100 (%) Control 261 ± ± ± ± ± PCB ± ± 0.8* 7.1 ± 0.2* 0.32 ± 0.01* ± 0.004* * p < Results expressed as mean ± SE with n = 4 rats receiving 1 µmol/kg/injection of PCB 126. * signifies a statistically significant change as compared to the vehicle control.

69 53 Table 2.3. GSH & GSSG nmoles GSH / mg tissue (wet weight) nmoles GSSG / mg tissue (wet weight) Control 7.44 ± ± PCB ± 0.32* ± ** * p < 0.05 ** p < Results expressed as mean ± SE of GSH or GSSG / whole liver weight (mm/mg) with n = 6 rats, receiving 1 µmol/kg PCB 126.

70 54 Table 2.4. SeGPx and total GPx activities (A), and CYP 1A activity (B) A Treatment Se-dependent GPx Total GPx (nmol/mg protein/min) Corn Oil 480 ± ± 71 PCB ± 17* 510 ± 28** * p < 0.01 ** p < Results expressed as mean ± SE of SeGPx or Total GPx (nmol/mg protein/min) with n = 6 rats, receiving 1 µmol/kg PCB 126. B Treatment EROD Activity (nmol/min/mg protein) MROD Activity (nmol/min/mg protein) Corn Oil ± ± PCB ± 0.151** ± 0.013** ** p < Results expressed as mean ± SE of EROD or MROD (nmol/min/mg protein) with n = 6 rats, receiving 1 µmol/kg PCB 126.

71 55 Day of Injections Day 0 Injection Day Average Weights (g) Control 184 ± 6.0 PCB ± 4.0 Figure 2.1. Growth curve of vehicle- (control) and PCB 126-treated rats. Growth of rats was not significantly affected by PCB 126 treatment.

72 56 A Figure 2.2. Liver and kidney selenium (A), copper (B), and zinc (C) levels of vehicle- (control) and PCB 126-treated rats. (A) Hepatic selenium was significantly diminished in PCB 126 treated rats. Kidney selenium was not significantly affected by PCB 126. (B) Hepatic copper was significantly increased in PCB 126 treated rats. Kidney selenium was not significantly affected by PCB 126. (C) Hepatic zinc was significantly diminished in PCB 126 treated rats. Kidney selenium was not significantly affected by PCB 126.

73 57 B Figure 2.2. Continued.

74 58 C Figure 2.2. Continued.

75 Figure 2.3. Histopathology of liver from vehicle- (control) (A) and PCB 126-treated (B) rats. (CL = centrilobular vein) H&E stain. (A) Control: Mild cytoplasmic clearing without vacuolization and with centrally located nuclei, consistent with glycogen. (B) PCB 126: Hepatocellular hypertrophy characterized by large areas of cytoplasmic pallor (hydropic degeneration), poorly defined vacuoles (lipid) and occasional peripheralized nuclei. 59

76 Figure 2.4. Histopathology of thymus from vehicle- (control) (A and B) and PCB 126-treated (C and D) rats. (C = cortex, M = medulla) H&E stain. (A) Control: The thymic cortex is densely cellular with clear delineation from the medulla. (B) Control: At higher magnification, the T-lymphocytes are closely packed. (C) PCB 126: The thymic cortex is thinner than that in controls with scattered vacuoles. (D) PCB 126: At higher magnification, the cortex consists of a mixed population of T-lymphocytes and macrophages surrounded by a clear space (vacuoles from C). 60

77 61 CHAPTER III THE EFFECTS OF DIETARY SELENIUM SUPPLEMENTATION ON ANTIOXIDANT STATUS DURING PCB 126 TOXICITY 2 Abstract Selenium is a critical antioxidant that is incorporated into amino acids and enzymes. Positive correlation between selenium supplementation and cancer prevention has been observed in both animal models and humans. Here the importance of dietary selenium was examined in preventing the toxicity of the most toxic congener of the polychlorinated biphenyl (PCB) family of environmental pollutants, 3,3,4,4,5- pentachlorobiphenyl (PCB 126), a potent AhR agonist. Male Sprague-Dawley rats were fed a modified AIN-93 diet with different dietary selenium levels (0.02, 0.2, 2 ppm). Following three weeks of acclimatization, rats from each dietary group were given a single ip injection of corn oil (control), 0.2, 1, or 5 mol/kg body weight PCB 126, followed by euthanization two weeks later. PCB 126 caused dose-dependent increases in liver wet weights (20-80%) and at the highest PCB 126 dose decreases in whole body weight gains. Hepatic cytochrome P450 (CYP1A1) activity was induced up to 27-fold even at the lowest dose of PCB126, indicating potent AhR activity. PCB exposure diminished hepatic selenium (8-35%) in a dose-dependent manner, and this was accompanied by diminished selenium-dependent glutathione peroxidase (SeGPx) activity (20-45%). Both of these effects were partially mitigated by selenium supplementation. 2 Activity determination of superoxide dismutases were contributed by Bingxuan Wang, Human Toxicology Program, University of Iowa. Trace elements determination was contributed by Drs. Yingtao Chai and Don L. Simmons, Iowa State Hygienic Laboratory. Glutathione analysis was contributed by Dr. Michael McCormick, Free Radicals and Radiation Biology, University of Iowa. Determination of thioredoxin redox state was contributed by Dr. Walter H. Watson, Johns Hopkins University. Histology was contributed by Wanda M. Haschek, University of Illinois. Statistical analysis formulas were contributed by Dr. Kai Wang, University of Iowa.

78 62 Conversely, thiorexodin reductase (TrxR) activity and thioredoxin oxidation state, while significantly diminished in the lowest dietary selenium groups, were not affected by PCB exposure. These results demonstrate that while supplemental dietary selenium was not able to completely prevent the toxicity caused by PCB 126, it was able to moderately increase the levels of several key antioxidants, hepatic selenium levels and SeGPx, thereby maintaining them roughly at the normal level. Introduction Selenium (Se) is an essential trace element, a key component of the antioxidant defense against reactive oxygen species (Rayman 2000). Se is unusual in that it is cotranslationally incorporated into the amino acid selenocysteine (SeCys) (Papp et al. 2007). SeCys becomes part of the active site of antioxidant enzymes, including the Sedependent family of glutathione peroxidases (SeGPx) responsible for reducing free hydrogen peroxide to water (Flohe et al. 1973; Rotruck et al. 1973) and the thioredoxin reductase (TrxR) responsible for reducing thiol-donating thioredoxin (Trx). Se deficiency has been associated with congestive cardiomyopathy and osteoarthropathy in humans (Brenneisen et al. 2005), white muscle disease in animals (Oldfield 1987) and downregulates antioxidant defense genes in rats (Fischer et al. 2002). In recent studies, supplemental Se was shown to reduce oxidative stress in animals (Menendez-Carreno et al. 2008; Singh et al. 2006) and in humans (Bardia et al. 2008; Seyedrezazadeh et al. 2008). Polychlorinated biphenyls (PCBs), a family of 209 different congeners, have been produced commercially as mixtures and were widely used in industrial settings because of their stability under a broad range of chemical, thermal, and electrical conditions (Safe 1994). However, that same stable nature of PCBs also allows them to persist in the environment, despite the decline in production since the 1970s. Because of their lipophilic nature, PCBs bioaccumulate and biomagnify in the food chain (Evans et al.

79 ; Galassi et al. 1994; Hansen 1987; La Rocca and Mantovani 2006; Metcalfe and Metcalfe 1997). Biologically individual PCB congeners were observed to have very different toxic effects, necessitating studies of individual congeners (Safe 1984). Of the 209 congeners, one of the most studied group of PCBs are the non-ortho substituted co-planar PCBs, which are able to bind to the arylhydrocarbon receptor (AhR) and induce cytochrome P450 (CYP) isoenzymes. The most potent of these is PCB 126, with a TEF of 0.1 relative to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) (Bandiera et al. 1982; Birnbaum and DeVito 1995; Leece et al. 1985; Mason and Safe 1986; Sawyer and Safe 1982). These dioxins and dioxin-like PCBs have been shown to disrupt the homeostasis of antioxidant enzymes, including the Se-dependent glutathione peroxidase (Se-GPx) (Hassan et al. 1985; Twaroski et al. 2001b). Severe liver injury can occur if in a situation of increased oxidative stress for example due to exposure to various dietary factors or environmental contaminants such antioxidant enzymes as GPx are not sufficiently present to remove the toxic intermediates (Polavarapu et al. 1998). Supplemental dietary Se is hypothesized to mitigate some or all of the hepatic alterations in enzyme activities and oxidative stress caused by a PCB congener of potent toxicity, PCB 126, while on the other hand low dietary intake of Se should increase the negative effects of PCB 126-exposure. Male rats were fed Se-controlled diets with low (0.02 ppm), adequate (0.2), or supplemental (2 ppm) levels of Se, administered a single low (0.2), medium (1), or high (5 mol/kg) dose of PCB 126, and their hepatic activity of CYP1A1, GPx, and TrR, as well as the hepatic levels of Se, Fe, Cu, Zn, and GSH were determined two weeks after the application of PCB 126 to test this hypothesis. Materials and Methods Chemicals All chemicals were obtained from Sigma-Aldrich Chemical Company (St. Louis, MO) unless otherwise stated. PCB 126 (3,3,4,4,5-pentachlorobiphenyl) was prepared by

80 64 an improved Suzuki-coupling method of 3,4,5-trichlorobromobenzene with 3,4- dichlorophenyl boronic acid utilizing a palladium-catalyzed cross-coupling reaction (Luthe et al. 2009). The crude product was purified by aluminum oxide column and flash silica gel column chromatography and recrystallized from methanol. The final product purity was determined by GC MS analysis to be > 99.8% and its identity confirmed by 13 C NMR. Caution: PCBs and their metabolites should be handled as hazardous compounds in accordance with NIH guidelines. Animals, Diet and PCB 126 Exposure Male Sprague-Dawley rats weighing grams from Harlan Laboratories (Indianapolis, IN) were housed in individual wire cages in a controlled environment maintained at 22 ºC with a 12 h light-dark cycle and water ad libitum. Animals were randomly divided into three dietary groups, and were fed ad libitum an AIN-93 based diet (Table 3.1) containing 0.02 ppm, 0.2 ppm, or 2 ppm sodium selenate obtained from Harland Teklad (Madison, WI). Following three weeks of acclimatization, animals were given a single i.p. injection of vehicle (stripped corn oil; 5 ml/kg b.w.; Acros Chemical Company, Pittsburgh, PA), or vehicle with a low (0.2 mol/kg, 65 g/kg), medium (1 mol/kg; 326 g/kg), or high (5 mol/kg; 1.63 mg/kg) dose of PCB 126. Feed consumption was monitored by determining the weight of the remaining food per cage every other day. Two weeks after the injection animals were euthanized using carbon dioxide asphyxiation followed by cervical dislocation. Liver and other organs were excised, weighed, and further processed as described below. All animal experiments were approved by the Institutional Animal Care and Use Committee of the University of Iowa. Preparation of Hepatic Subcellular Fractions Parts of the liver tissues were immediately minced and homogenized in ice-cold 0.25 M sucrose solution containing 0.1 mm EDTA, ph 7.4. The homogenates were centrifuged at 10,000g for 20 min. The resulting supernatants were then centrifuged at

81 65 100,000g for 1 h. These supernatants, containing the cytosolic fractions were dispensed and aliquoted. The pellets, containing the microsomes, were washed twice with cold sucrose/edta solution, resuspended in this solution and aliquoted. Protein concentrations of the microsomal solutions were determined by the method of Lowry et al. (1951). All tissue fractions were frozen and stored at -80 C until further analysis. Measurement of Cytochrome P450 (CYP1A1) Activity CYP1A1 activity in hepatic microsomal fractions was determined according to a slightly modified method of Burke and Mayer (1974), by measuring the ethoxyresorufin deethylase (EROD) activity with ethoxyresorufin as the substrate. The resulting fluorescent resorufin product from the monooxygenase reaction was detected using a Perkin-Elmer LS 55 spectrofluorometer at excitation wavelength of 550 nm and emission wavelength of 585 nm. Activity Determination of Superoxide Dismutases (SOD) SOD activities were determined following the method of Spitz and Oberley (1989). Briefly, to obtain Total SOD activity the scavenging of xanthine/xanthine oxidase-generated superoxide anion radical by SOD was monitored by measuring nitroblue tetrazolium reduction at 560nm. Manganese SOD (MnSOD) activity was determined in the presence of cyanide. CuZnSOD activity was then determined by subtracting the MnSOD activity from the Total SOD activity. SOD activities are expressed as units of SOD activity per milligram protein. Trace Elements Determinations Metal concentrations in different rat tissues were quantitatively determined with an elemental mass spectrometer by Inductively Coupled Plasma Mass Spectrometry (ICP-MS). ICP-MS was selected because of its low detection limits and multi-element capacity (Entwisle and Hearn 2006). Liver tissues were pretreated with HNO 3 acid

82 66 digestion prior to ICP-MS measurement, and metal concentrations in the treated tissues were determined in an Agilent 7500ce ICP-MS equipped with a CETAC AS520 auto sampler. Activity Determination of Glutathione Peroxidases (SeGPx, Total GPx, GST) SeGPx and total GPx activities were measured in hepatic cytosolic fractions by the methods of Lawrence and Burk (1976), using hydrogen peroxide and cuemene hydroperoxide as the substrates, respectively. GST activity was determined by the method of Habig et al. (1974), using 1-chloro-2,4-dinitrobenzene (CDNB) as the substrate. An absorbance change at 340 nm caused by the conjugation of CDNB to reduced glutathione (GSH) was followed in a Beckman DU-650 spectrophotometer for 5 min. One unit of enzymatic activity is defined as the amount of protein that oxidizes 1 µm of NADPH per min, expressed as milliunits per mg protein. Measurement of Thioredoxin Reductase (TrxR) Activity TrxR activity was measured in hepatic cytosolic fractions by the methods of Holmgren and Bjornstedt (1995), using 5,5'-dithio-bis (2-nitrobenzoic acid) (DTNB) as the substrate. The formation of the resulting product, 2-nitro-5-thiobenzoate (TNB), was determined spectrometrically at an absorbance of 412 nm. TrxR activity was estimated by subtracting the time-dependent increase in absorbance in the presence of a TrxR activity inhibitor. One unit of activity was defined as 1 μm TNB formed (min mg protein). Measurement of the Redox State of Thioredoxin-1 (Trx1) and Thioredoxin-2 (Trx2) The redox state of Trx1 was determined by the redox western blot technique, as described previously (Watson et al. 2003). Briefly, cells were lysed in the presence of 50 mm iodoacetic acid to carboxymethylate-reduced cysteines. Proteins were separated by

83 67 native PAGE, and oxidized and reduced forms of Trx1 were detected by western blotting using a commercially available antibody (American Diagnostica, Stamford, CT). Densitometry was performed using LI-COR imaging software. The redox state of Trx1 was calculated by dividing the sum of the 1 disulfide and 2 disulfide forms by the sum of all forms (oxidized and reduced) and expressed as percent oxidized. The Trx2 redox state was measured as described in (Halvey et al. 2005). After derivatization of reduced Trx2 with 4-acetoamido-4-maleimidylstilbene-2,2-disulphonic acid (AMS; Molecular Probes), oxidized and reduced forms were separated by SDS- PAGE and detected by western blotting using a Trx2-specific antibody (a kind gift from Dean P. Jones, Emory University, GA). Total Glutathione (GSH and GSSG) Analysis For the determination of glutathione liver samples were homogenized in 5% salicylic acid. Reduced (GSH) and oxidized (GSSG) glutathione levels were determined by the methods of Griffith (1980) and Anderson (1985), using glutathione reductase as the substrate. Absorbance change at 412 nm was followed in a Beckman DU-670 spectrophotometer for 5 min. The rate of yellow color accumulation is the result of 5- thio-2-nitrobenzoate (2TNB) formation from 5,5'-dithio-bis-(2-nitrobenzoic acid) proportional to the amount of total glutathione in the sample. GSSG was measured independently by incubating the tissue in 2-vinylpyridine (2-VP), which conjugates GSH. Reduced glutathione was determined by subtracting GSSG from total glutathione. GSH levels are expressed as per mg protein. Histology Formalin-fixed tissues (liver, spleen, and thymus) were processed routinely, embedded in paraffin, sectioned at 3-4 micrometers and stained with hematoxylin and eosin for light microscopic examination. Frozen sections were prepared from selected liver samples and stained with Oil-Red-O for determination of the presence of lipid.

84 68 Statistics The effect of PCB 126-treatment and dietary Se level on various responses was studied using ANOVA analysis via procedure PROC GLM in the statistical analysis package SAS (version 9.2). The Dunnett's test was used to compare PCB 126-treatment with the corn oil control and other Se levels with Se level 0.2 ppm. This comparison was conducted separately to PCB 126-treatment and Se level (one-way ANOVA) and also jointly (two-way ANOVA). In two-way ANOVA, the interaction term was removed if it was not significant at level The effect of Se level is controlled when applying Dunnett's test to PCB 126-treatment by using lsmeans statement in PROC GLM. The same was done when applying the Dunnett's test to Se level. Results Growth and Organ Weights The growth rate, indicated by weight gain during the week after PCB 126- injection, was not significantly changed in rats treated with low (0.2 mol/kg) and mid (1 mol/kg) doses of PCB 126 in any of the Se groups (Table 3.3). However, growth was significantly slowed and some animals even lost weight at the high (5 mol/kg) dose of PCB 126. This effect was slightly ameliorated by high dietary Se levels. Feed consumption was also significantly affected overall by PCB 126-exposure and correlated with the growth effects (Table 3.2). Relative liver weight (as percentage of body weight) was increased in a dosedependent manner by PCB 126 from 11-35% with 0.2 mol/kg PCB 126 to over 70% in all three Se groups with 5 mol/kg PCB 126 (Table 3.3). Dose-dependent thymic involution of 20-29% at the low dose and around 80% at the high dose of PCB 126 was observed (Table 3.3). Low (0.02 ppm) dietary Se had an overall significant reducing effect on thymic involution. All doses of PCB 126 significantly increased relative lung weights, but not in a dose-dependent manner nor was an effect of Se diets apparent

85 69 (Table 3.3). No consistent significant effect on relative kidney weight was seen by PCB 126 or dietary Se (Table 3.2). The relative testes weight was significantly affected overall only by the high dose of PCB 126, where an increase compared to the vehicle control was seen (Table 3.2). Effects on CYP 1A1 (EROD) Activity As depicted in Table 3.4, EROD activity was significantly induced (17 to 27-fold) by low (0.2 mol/kg) and medium (1 mol/kg) doses of PCB 126. At a 5 mol/kg dose of PCB 126, EROD activity was also significantly induced, but at a lower magnitude (8 to 13-fold) compared to the two lower doses. The EROD activity was not significantly affected by different dietary levels of Se. Effects on Hepatic Se Hepatic Se levels were significantly (65-75%) lower in the low dietary Se groups compared to the adequate Se groups (Figure 3.1). Rats receiving supplemental (2 ppm) dietary Se resulted in significantly 50-80% increased hepatic Se levels compared to the rats receiving adequate (0.2 ppm) dietary Se (Figure 3.1). PCB 126 produced no significant change in hepatic Se in the rats receiving low (0.02 ppm) or adequate Se. However, overall PCB 126 significantly diminished hepatic Se by 8-35% at all doses, and this reduction reached statistical significance with the mid (1 mol/kg) and high (5 mol/kg) doses in the supplemental dietary Se groups relative to the corn oil control (Figure 3.1). This resulted in a highly significant (p < 0.004) interaction effect of PCB 126 and dietary Se (Table 3.2). Effects on Hepatic Copper, Iron, Manganese, and Zinc The liver tissue levels of several metals were analyzed to identify the effects of PCB 126 and/or different dietary Se levels. Hepatic iron was diminished by PCB 126, significantly at the mid and high PCB 126 doses and also overall (Table 3.5). This effect

86 70 was most pronounced in the supplemental (2 ppm) Se group where the reduction reached 49% and least pronounced in the low (0.02 ppm) Se group, where it only reached a 15% reduction at high (5 µmol/kg) dose PCB 126. However, overall the dietary level of Se had no effect on hepatic iron. Copper levels were also not affected by dietary Se, but increased in a dosedependent manner by PCB 126 treatment which reached significance at the mid (1 mol/kg) and high dose groups of PCB 126 in all dietary Se groups (Table 3.5). PCB 126 reduced hepatic Zn levels by 5-20%, which was significant overall at high dose PCB 126 (Table 3.5). No significant differences in zinc levels were seen among the different dietary Se groups. Hepatic manganese was diminished 8-33% by PCB 126 and this effect was dosedependent, although statistically significant only in all mid dose PCB 126 groups and at all doses overall and in the supplemental (2 ppm) dietary Se groups. Although different Se levels did not cause a significant difference at the individual treatment points, a significant overall higher manganese level was seen in the low Se group compared to the adequate Se group (Table 3.5). CuZnSOD, and Total SOD Activities The cytoplasmic superoxide dismutase has Cu and Zn in the active center and the activities could potentially be influenced by changes in the cellular level of these metals. CuZnSOD and total SOD activities were not significantly affected by PCB 126 treatment or dietary Se levels (Table 3.6). Effects on Glutathione Peroxidase Activities SeGPx activities were strongly influenced by the dietary Se levels. In the vehicle control treated animals, low (0.02 ppm) dietary Se resulted in 92% diminished SeGPx activity compared to adequate (0.2 ppm) dietary Se, whereas a supplementation of the diet with 2 ppm Se resulted in a 41% increase of SeGPx (Fig 3.2). In the adequate and

87 71 supplemental dietary Se groups PCB 126 caused a dose-dependent reduction of SeGPx activity by 25-40% (Fig 3.2). No effect of PCB 126 treatment was apparent in the low dietary Se groups. Nevertheless, overall the effects of dietary Se and PCB 126 were strongly interactive (p < 0.005; Table 3.2). Glutathione transferases activities were not affected by dietary Se, but were increased in a dose-dependent manner (1.4 to 2.5-fold) by PCB 126 in all three dietary Se groups (Table 3.7). Total GPx activity was 73% lower in low dietary Se vehicle control group and 24% higher in the supplemental dietary Se vehicle control group compared to the adequate dietary Se vehicle control group (Table 3.7). PCB 126 exposure caused a dosedependent (1.3 to 2-fold) increase in total GPx activity in the lowest dietary Se groups, but dose-dependent increase was not observed in the adequate and supplemental dietary Se groups. In the adequate Se group total GPx activities were diminished by low (0.2 mol/kg) dose PCB 126, reached control level in the mid (1 mol/kg) dose of PCB 126, and was increased, although non-significantly, in the high (5 mol/kg) dose group (Table 3.7). With the supplemental dietary Se, PCB 126 caused a non-significant reduction in total GPx activity at the two lower doses and a return to control levels was seen with the high dose of PCB 126. Overall both, dietary Se levels and PCB 126, had a significant effect on total GPx and also together a significant interaction effect (p < 0.05, Table 3.2). Effects on Thioredoxin Reductase Activity and Thioredoxin Oxidation States Thioredoxin reductase (TrxR) activity was only influenced by low dietary (0.02 ppm) Se levels (Figure 3.3) which caused significantly (76-89%) diminished TrxR activity compared to the adequate (0.2 ppm) dietary Se groups. No overall effect on TrxR was observed in rats fed supplemental (2 ppm) Se (Figure 3.3). Different concentrations of PCB 126 did not have a significant effect on TrxR activity at any of the dietary Se

88 72 levels, but an overall significant reduction of TrxR activity was seen in the low dietary Se groups. This effect resulted in a significant interaction effect between PCB 126 and Se (Table 3.2) Thioredoxin-1 (Trx1) and thioredoxin-2 (Trx2) were significantly more oxidized, 79% and 107%, respectively, in the vehicle controls with low dietary Se compared to the adequate Se group (Figure 3.4). PCB 126 had no significant effect on the oxidation state of Trx1 or Trx2. Effects on Hepatic Glutathione No significant effects by PCB 126 or dietary Se on hepatic GSH levels were observed, although Se had an overall significant effect on hepatic GSSG and the GSSG/GSH ratio, both of which were diminished by low (0.02 ppm) dietary Se (Table 3.8). Histology Dose-related changes were present in all PCB 126-treated rat livers. PCB 126 caused an increase in the amount and density of cytoplasm in centrilobular hepatocytes, suggestive of smooth endoplasmic reticulum induction. No additional histological changes were observed in rats exposed to low (0.2 mol/kg) dose PCB 126. In rats treated with higher doses of PCB 126 (1 and 5 mol/kg), dose-dependent cytoplasmic vacuolation and degeneration were present in all groups except for the group receiving low dietary (0.02 ppm) Se and exposed to mid (1 mol/kg) dose PCB 126. In addition, dose-dependent scattered apoptosis and karyomegaly were observed in all mid and high PCB dose groups, but severity was found to be lower in rats receiving supplemental (2 ppm) dietary Se. Oil-Red-O staining revealed that the vacuolation was due to lipid accumulation (lipidosis/steatosis).

89 73 Dose-related thymic atrophy was observed in rats given higher doses (1 and 5 mol) of PCB 126, independent of dietary Se level. No changes were observed in the spleen. Discussion Se and Se-containing enzymes are important for health and for antioxidant defense (Oldfield 1987; Stadtman 2000). Se supplementation was shown to inhibit high fat-induced serum cholesterol oxidation (Menendez-Carreno et al. 2008), to reduce the toxicity of cadmium in rats (Banni et al. 2010), and to be promising in protecting against cancer in animals (Menendez-Carreno et al. 2008; Singh et al. 2006) and humans (Bardia et al. 2008; Seyedrezazadeh et al. 2008). Even modest deficiency in Se may increase the risk of diseases (McCann and Ames 2011). Thus a decrease in Se and Se-containing enzymes such as previously observed with PCB 126 (Lai et al. 2010), PCB 77 (Schramm et al. 1985; Stemm et al. 2008; Twaroski et al. 2001b), and TCDD (Hassan et al. 1985) may have unexpected negative health consequences through this unrecognized mechanism. On the other hand, dietary Se levels beyond the therapeutic range are toxic and may increase the risk of type II diabetes and cancer, including PCB-induced hepatocarcinogenesis (Oldfield 1987; Stemm et al. 2008; Stranges et al. 2007). Therefore, to examine the effect of different dietary Se levels, rats were fed for 5 weeks with a diet containing low (0.02 ppm), adequate (0.2 ppm) or supplemented (2 ppm) levels of Se and at the end of the third week on the diets with one ip injection of corn oil or PCB 126 (0.2, 1, and 5 µmol/kg). All Se diets were very well tolerated by the animals with respect to growth and organ weights. PCB 126 on the other hand produced reduced growth and even weight loss at the highest (5 mol/kg) dose. This was most likely due to decreased food intake, or wasting syndrome, similar to the effects of acute TCDD toxicity (Seefeld et al. 1984). Even though Se alone had no effect on feed intake, a significant interaction was seen with

90 74 PCB 126 and dietary Se. This could indicate that Se ameliorates PCB 126-induced wasting, but more experiments are needed to confirm the small effect seen in these studies. As expected, relative liver weights were dose-dependently increased by PCB 126, an effect that was also seen to a smaller extent in the lungs of the animals (Table 3.3). PCBs induce liver cancer in rodents and possibly in humans (Mayes et al. 1998; Ward et al. 2010). Although their mechanism of carcinogenicity is not known, it is assumed that AhR activation may be involved, at least in the promoting activity of certain PCBs (Brown et al. 2007). PCB 126 is by far the most potent AhR agonist of all 209 PCB congeners. Increase in liver weight is a well-known consequence of AhR mediated increase in endoplasmic reticulum. Histological analysis confirmed this effect of all PCB doses on the liver. Se had no effect on these parameters either alone or in combination with PCB 126. Thymus involution is another well-known effect of AhR agonists. Macroscopic and histologic evaluation confirmed the dose-related thymic toxicity of PCB 126. Animals receiving the low Se diet had overall slightly higher thymus weights (Table 3.3), not enough, however, to produce visible histological differences. The mechanism or consequence of this small protective effect of the low Se diet is not known. The 5 weeks on low or supplemented Se diets produced a significant 70% reduction and 80% increase in hepatic Se levels, respectively, compared to the adequate Se diet. These changes in Se were more pronounced than in our previous 10 week promotion study on the same Se diets (Stemm et al. 2008). It is possible that the longer time period on the diets allowed for adaptations in the Se kinetics in the body. As before (Lai et al. 2010), PCB 126 reduced hepatic Se levels in the adequate dietary Se group, and this effect was dose-dependent (Figure 3.1). While supplemental dietary Se was unable to prevent the PCB 126-induced loss of Se from the liver, the remaining hepatic Se levels were still significantly higher than those in rats receiving adequate Se. Thus

91 75 supplementing the diet with Se was chemoprotective, since it prevented hepatic Se to fall below the normal tissue level. This is similar to the protective effect seen in cadmium exposed rats on a Se supplemented diet (Banni et al. 2010). Rats fed the low Se diet showed hepatic Se levels that were already very low and PCB 126 did not produce a significant further reduction. Interestingly, this level (~ ug/g tissue) is the same as the one observed with PCB 77-treatment of rats after 10-weeks on a low Se diet (Stemm et al. 2008). This may be the lowest hepatic Se level that the rat physiology will tolerate. Overall the interaction between PCB 126 and dietary Se levels was highly statistically significant (Table 3.2), confirming the observed effect of this strong AhR agonist, but also the modulation of this effect by different dietary Se levels. PCB 126 is known to induce cytochrome P450s (CYP), particularly CYP1A. CYP1A is not constitutively expressed, but its synthesis can be greatly enhanced by AhR driven up-regulation of gene expression (Parkinson et al. 1983a; Parkinson et al. 1983b). Here PCB 126 caused an increase in CYP1A1 activity (Table 3.4), consistent with the higher liver weight due to induction of smooth endoplasmic reticulum. The lower CYP activity observed in the high PCB 126 dose groups may be due to oxidative inactivation of the enzyme by reactive oxygen species or diminished cellular resources available to support protein synthesis, since the synthesis of other proteins in the liver is also reduced. Important in this study is that the different Se diets did not influence this effect of PCB 126 on CYP1A1 activity in any way. CYPs are membrane bound enzymes that play key part of phase I biotransformation, catalyzing the metabolism of many endogenous and exogenous compounds. Low basic enzyme levels are protective, since uncoupling of their catalytic cycle can cause a release of superoxide and hydrogen peroxides (Schlezinger et al. 1999). Thus this increase in CYP activity by PCB 126 is believed to increase oxidative stress in the liver. Surprisingly, in this study no effect of hepatic glutathione (GSH) was observed with PCB 126 as was seen with PCB 77 (Twaroski et al. 2001b). Even more surprising

92 76 was that the low Se diet slightly increased the hepatic total GSH and oxidized glutathione (GSSG) levels and also the GSSG/GSH ratio, suggesting reduced oxidative stress. In wild birds high tissue Se levels were shown to correlate with high GSSG levels and GSSG/GSH ratios and the authors suggest that the increased SeGPx activity was responsible for increased oxidative stress (TBARS) and oxidized glutathione (Franson et al. 2002; Hoffman 2002). Thus high Se levels may cause the opposite of the desired effect. However, Se-related toxicity in birds was seen at tissue levels of 3 ppm, 200% higher than those achieved in the rats on supplemented Se diet. A metal with possible chemoprotective activity is zinc (Zn). Hepatic Zn levels were slightly but significantly reduced by exposure to the high (5 µmol/kg) dose and slightly but not significantly with 1 µmol/kg PCB 126. This is in agreement with previous findings with PCB 126 (Lai et al. 2010), but different from the results with TCDD which produced no change (Wahba et al. 1988) or increases (Nishimura et al. 2001) in hepatic Zn levels. The reason for this difference between PCB 126 and TCDD is not known. The Se diets did not influence the Zn levels in any group, indicating that the PCB-induced changes in Se levels were not involved in lowering of the hepatic Zn content. In addition, hepatic Mn levels were reduced by all PCB 126 treatments in all Se groups. Liver manganese levels were slightly but overall significantly higher in the low Se group compared to adequate or supplemented Se. This could indicate a compensatory mechanism. High Se levels on the other hand did not reduce Mn levels, indicating that this dietary level may be safe to use. Despite the induction of CYP, a hemoprotein, hepatic iron (Fe) levels were decreased as the dose of PCB 126 increased (Table 3.5). This is most likely a diet effect, since the exposure time to PCB 126 (2 weeks) was too short to produce uroporphyria. TCDD was reported to increase hepatic Fe levels (Nishimura et al. 2001). Similar to CYP induction, dietary Se levels did not change hepatic iron levels, indicating that these parameters are indeed independent from each other.

93 77 The other major transition metal, besides Fe, is copper (Cu). Hepatic Cu levels were significantly increased by PCB 126 in a dose-dependent manner (Table 5a), which is in agreement with previous findings and similar reports after TCDD exposure of rats (Elsenhans et al. 1991; Nishimura et al. 2001; Wahba et al. 1988). This effect was dosedependent: even a single injection of 0.2 µmol/kg PCB 126 produced an apparent, although not significant, increase in Cu, while the high dose doubled hepatic copper levels. The mechanism of hepatic Cu by PCB 126 or TCDD is not known, but an impairment of biliary excretion was suggested (Elsenhans et al. 1991; Mahoney et al. 1955). Cu is a very efficient Fenton reagent which can convert H 2 O 2 to the highly reactive hydroxyl radical which immediately oxidizes proteins and cellular DNA (Buettner and Jurkiewicz 1996). Cu supplementation was shown to reduce Se levels in muscles of cattle (Garcia-Vaquero et al. 2011), indicating that these two trace metals influence each other. Our major question in this study was whether a high or low Se diet would influence Cu levels. As shown in Table 3.5 and 3.2, Se did not affect hepatic Cu levels alone or in combination with PCB 126. Cu is a constituent of the CuZnSOD, an important cytoplasmic antioxidant enzyme. However, the higher availability of Cu as a result of PCB exposure did not cause an increase in CuZnSOD activity, and neither did any dietary level of Se (Table 3.6). Changes in dietary Se have been reported to influence other antioxidant enzymes, notably the selenoenzymes glutathione peroxidase (SeGPx) and thioredoxin reductase (TrxR) (Stadtman 2000). SeGPx is a major peroxidase that detoxifies H 2 O 2 and cytosolic hydroperoxides. Rats receiving the supplemented Se diet had higher SeGPx activity, while animals on the low Se diet had significantly diminished SeGPx activities compared to the adequate group (Figure 3.2). In addition, PCB 126 decreased SeGPx activity in a dose dependent manner; this effect was more pronounced when the hepatic Se level was higher. As a consequence even the low dose of PCB 126 significantly reduced SeGPx in the supplemental Se group while the mid PCB 126 dose was needed in the adequate Se

94 78 group and no statistical significance effect of PCB 126 was apparent in the low Se group. However, in Se supplemented rats the SeGPx activity after PCB 126-treatment remained in the baseline range of the control animals on the adequate Se diet. Thus supplementation of the diet with Se prevented a significant reduction of this important antioxidant enzyme below normal levels, a chemoprotective effect against PCB 126 toxicity. Although the low Se diet had a stronger effect on SeGPx than any dose of PCB 126, it was hypothesized that even a small reduction in SeGPx may cause aging related diseases such as cancer, loss of bone density and resistance to infections, heart diseases and others (McCann and Ames 2011). In addition, PCB 126 and Se diet had an interaction effect (Table 3.2), raising the concerns that even slight dietary deficiency of Se and exposure to AhR agonists like PCB 126 and TCDD together may be enough to cross the threshold of NOEL. Interestingly, PCB 126 had no effect on the total GPx activity. Low and supplemented dietary Se reduced and increased, respectively, the total GPx activity level, but treatment with PCB 126 did not have a significant effect in the adequate and supplemental Se groups. In contrast, PCB 126 caused an increase in Total GPx in the low Se groups. As a consequence total GPx activity was significantly affected by PCB 126, dietary Se levels, and in interaction between these two factors. GST is one component of Total GPx activity, and although PCB 126 s effect on Total GPx activity was partially Se-dependent, the increases observed was Se-independent (Table 3.2). Indeed a PCB dose-dependent increase in GST (Table 3.6), similar to the one reported previously (Schramm et al. 1985) was observed, suggesting crosstalk between nuclear factor erythroid-2 p45-related factor-2 (Nrf2)/antioxidant response element (ARE) with the AhR (Anwar-Mohamed et al. 2011). The second major Se-dependent antioxidant system includes the thioredoxin reductase (TrxR). TrxR activity was influenced by dietary Se levels. The low Se diet caused an about 75-85% reduction in TrxR activity (Figure 3.3). A similar effect of Se

95 79 deficient diets had been reported by others (Hill et al. 1997). A small, non-significant increase in TrxR was seen in the Se supplemented group. Unlike SeGPx, the TrxR activity was not affected by PCB 126 (Figure 3.3). This is in agreement with reports that SeGPx activity declines before TrxR activity, possibly because the GPx mrna loses stability during Se deficiency (Bermano et al. 1996). However, a small, but significant overall interaction effect between Se and PCB 126 was observed by ANOVA analysis (Table 3.2). Several PCB-treatment groups showed higher oxidized Trx levels than the controls, but this did not reach significance and now overall or combined effect of PCB 126 on oxidized Trx was seen. However, a significant increase in the oxidation level of Trx1 and Trx2 in the low Se group compared to the adequate and supplemental dietary groups was seen (Figure 3.4a), most likely a consequence of the strongly reduced TrxR levels in this group, indicate an increase in mixed protein disulfides. Overall these data indicate that both, dietary Se and contamination with AhR agonists should be analyzed together. These results demonstrate that while supplemental dietary Se did not prevent Se depletion from the liver following PCB 126 exposure it was able to increase hepatic Se levels and SeGPx activity moderately, keeping either above or within the range of the untreated control. Thus considering the bioaccumulation of ubiquitous environmental contaminants like PCBs, dioxins and furans and the possibility that even a small reduction of selenoproteins like SeGPx may increase the risk of age-related diseases like cancer, more emphasis should be placed in understanding the complex interaction between these contaminants and our diet.

96 80 Table 3.1. Composition of AIN-93M modified selenium diets Low Adequate Supplemental 0.02 ppm Se 0.2 ppm Se 2 ppm Se Constituent g/kg g/kg g/kg Torula Yeast DL-Methionine Corn Starch Dextrose, monohydrate Soybean Oil Cellulose Se Deficient Mineral Mix Calcium Carbonate Sodium Selenate (0.1% in sucrose) Vitamin Mix, AIN-93-VX Choline Bitartrate

97 81 Table 3.2. Two-way ANOVA analysis of the effects of PCB 126, dietary selenium, and interaction PCB 126 Dietary Selenium Interaction Effect Growth Rate (%) < Feed Consumption < Relative Liver Weight (%) < Relative Thymus Weight (%) < Relative Lung Weight (%) < Relative Kidney Weight (%) Relative Testes Weight (%) EROD Activity < MnSOD Activity < CuZnSOD Activity Total SOD Activity Hepatic Selenium <0.001 < Hepatic Copper < Hepatic Iron < Hepatic Manganese < Hepatic Zinc SeGPx Activity <0.001 < GST Activity < Total GPx Activity <0.001 < TrxR Activity - < Trx1 Oxidation % Trx2 Oxidation % - < GSH (protein) GSSG (protein) GSSG/GSH (protein) Only p-values <0.05 are reported

98 82 Table 3.3. Growth (%) and relative liver weights (%) (A), and relative lung weights (%) and relative thymus weights (%) (B) Growth (%) Relative Liver Weight (%) Dietary Selenium Level Dietary Selenium Level Low Adequate Supplemental Overall Low Adequate Supplemental Overall Treatment (0.02 ppm) (0.2 ppm) (2 ppm) (0.02 ppm) (0.2 ppm) (2 ppm) Corn Oil 21.1 ± ± ± ± ± ± 0.10 (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) mol/kg PCB ± 1.4 (--, --) 16.1 ± 1.9 (--, --) 16.9 ± 2.0 (--, --) * 5.00 ± 0.19 (*, --) 5.57 ± 0.17 (*, --) 4.84 ± 0.17 (--, *) * 1 mol/kg PCB ± 2.1 (--, --) 15.7 ± 2.2 (--, --) 14.1 ± 1.2 (*, --) * 6.02 ± 0.10 (*, --) 6.27 ± 0.17 (*, --) 6.15 ± 0.17 (*, --) * 5 mol/kg PCB ± 2.5 (*, --) -4.2 ± 1.9 (*, --) -0.6 ± 1.1 (*, --) * 6.91 ± 0.43 (*, --) 7.01 ± 0.31 (*, --) 7.52 ± 0.14 (*, --) * Overall Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine for statistically significant differences (p < 0.05) between each PCB 126 level and the corn oil treatment (*) and between low or supplemental and the adequate dietary selenium level (#). Significance for each factor based on two-way ANOVA is indicated in the bottom margin (#, Se) and in the right margin (*, PCB).

99 83 Table 3.3. Continued. Relative Lung Weight (%) Relative Thymus Weight (%) Dietary Selenium Level Dietary Selenium Level Low Adequate Supplemental Overall Low Adequate Supplemental Overall Treatment (0.02 ppm) (0.2 ppm) (2 ppm) (0.02 ppm) (0.2 ppm) (2 ppm) Corn Oil ± ± ± ± ± ± (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) mol/kg ± ± ± ± ± ± PCB 126 (--, --) (*, --) (--, --) * (*, --) (*, --) (*, --) * 1 mol/kg ± ± ± ± ± ± PCB 126 (*, --) (*, --) (--, --) * (*, --) (*, --) (*, --) * 5 mol/kg ± ± ± ± ± ± PCB 126 (*, --) (*, --) (--, --) * (*, --) (*, --) (*, --) * Overall # Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine for statistically significant differences (p < 0.05) between each PCB 126 level and the corn oil treatment (*) and between low or supplemental and the adequate dietary selenium level (#). Significance for each factor based on two-way ANOVA is indicated in the bottom margin (#, Se) and in the right margin (*, PCB).

100 84 Table 3.4. EROD activity (nmol/min/mg protein) Dietary Selenium Level Treatment Low (0.02 ppm) Adequate (0.2 ppm) Supplemental (2 ppm) Overall Corn Oil 0.10 ± 0.02 (--, --) 0.13 ± 0.02 (--, --) 0.11 ± 0.02 (--, --) mol/kg PCB ± 0.12 (*, --) 2.70 ± 0.08 (*, --) 2.72 ± 0.09 (*, --) * 1 mol/kg PCB ± 0.22 (*, --) 2.29 ± 0.16 (*, --) 2.28 ± 0.10 (*, --) * 5 mol/kg PCB ± 0.17 (*, --) 1.08 ± 0.07 (*, --) 1.22 ± 0.12 (*, --) * Overall Results are expressed as mean ± SEM. Each group contained 4-9 animals. One-way ANOVA was used to examine the difference between each PCB 126 level and the corn oil treatment (*). No significant difference between low or supplemental and the adequate dietary selenium level were observed. Significance for each factor based on two-way ANOVA is indicated in the bottom margin (Se,#) and in the right margin (PCB,*). The level for significance is 0.05.

101 85 Table 3.5. Liver copper ( g/g) and iron ( g/g) (A), and manganese ( g/g) and zinc ( g/g) (B) Liver Copper ( g/g) Liver Iron ( g/g) Dietary Selenium Level Dietary Selenium Level Low Adequate Supplemental Overall Low Adequate Supplemental Overall Treatment (0.02 ppm) (0.2 ppm) (2 ppm) (0.02 ppm) (0.2 ppm) (2 ppm) Corn Oil 5.41 ± ± ± ± ± ± 14 (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) mol/kg PCB ± 0.46 (--, --) 5.81 ± 0.19 (--, --) 5.93 ± 0.32 (--, --) ± 14 (--, --) 158 ± 14 (--, --) 150 ± 12 (*, --) -- 1 mol/kg PCB ± 0.90 (*, --) 7.31 ± 0.62 (*, --) 7.99 ± 0.36 (*, --) * 137 ± 11 (--, --) 119 ± 7 (--, --) 125 ± 3 (*, --) * 5 mol/kg PCB ± 0.88 (*, --) ± 0.60 (*, --) 9.67 ± 0.46 (*, --) * 141 ± 14 (--, --) 119 ± 11 (--, --) 98 ± 10 (*, --) * Overall Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine for significant differences between each PCB 126 level and the corn oil treatment (*) and between low or supplemental and the adequate dietary selenium level (#). Significance for each factor based on two-way ANOVA is indicated in the bottom margin (#, Se) and in the right margin (*, PCB). The level for significance is 0.05.

102 86 Table 3.5. Continued Liver Manganese ( g/g) Liver Zinc ( g/g) Dietary Selenium Level Dietary Selenium Level Low Adequate Supplemental Overall Low Adequate Supplemental Overall Treatment (0.02 ppm) (0.2 ppm) (2 ppm) (0.02 ppm) (0.2 ppm) (2 ppm) Corn Oil 3.00 ± ± ± ± ± ± 1.9 (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) mol/kg PCB ± 0.26 (--, --) 2.29 ± 0.18 (--, --) 2.26 ± 0.16 (*, --) * 35.5 ± 2.9 (--, --) 34.3 ± 2.7 (--, --) 34.6 ± 2.1 (--, --) -- 1 mol/kg PCB ± 0.18 (*, --) 1.83 ± 0.18 (*, --) 2.09 ± 0.13 (*, --) * 35.7 ± 2.2 (--, --) 31.8 ± 1.2 (--, --) 35.1 ± 1.5 (--, --) -- 5 mol/kg PCB ± 0.20 (--, --) 2.17 ± 0.14 (--, --) 1.95 ± 0.15 (*, --) * 31.7 ± 2.6 (--, --) 29.8 ± 1.6 (--, --) 31.0 ± 1.5 (*, --) * Overall # Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine for significant differences between each PCB 126 level and the corn oil treatment (*) and between low or supplemental and the adequate dietary selenium level (#). Significance for each factor based on two-way ANOVA is indicated in the bottom margin (#, Se) and in the right margin (*, PCB). The level for significance is 0.05.

103 87 Table 3.6. CuZnSOD activity (U/mg protein) under each experiment condition and significance of various comparisons (adjusted using Dunnett s test) Dietary Selenium Level Treatment Low (0.02 ppm) Adequate (0.2 ppm) Supplemental (2 ppm) Overall Corn Oil ± 48.9 (--, --) ± 33.0 (--, --) ± 28.3 (--, --) mol/kg PCB ± 33.2 (--, --) ± 30.5 (--, --) ± 15.4 (--, --) -- 1 mol/kg PCB ± 29.8 (--, --) ± 16.5 (--, --) ± 9.6 (--, --) -- 5 mol/kg PCB ± 33.0 (--, --) ± 29.0 (--, --) ± 43.9 (--, --) -- Overall Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine the difference between each PCB 126 level and the corn oil treatment and between low/supplemented Se vs adequate Se. Significance for each factor overall based on two-way ANOVA is indicated in the bottom margin and in the right margin. The level for significance is No significance was detected.

104 88 Table 3.7. Glutathione transferase ( mol/min/mg protein) and total glutathione peroxidase activities ( mol/min/mg protein) under each experiment condition and significance of various comparisons (adjusted using Dunnett s test) Glutatione Transferase Activity ( mol/min/mg protein) Total GPx Activity ( mol/min/mg protein) Dietary Selenium Level Dietary Selenium Level Low Adequate Supplemental Overall Low Adequate Supplemental Overall Treatment (0.02 ppm) (0.2 ppm) (2 ppm) (0.02 ppm) (0.2 ppm) (2 ppm) Corn Oil 211 ± ± ± ± ± ± 25 (--, --) (--, --) (--, --) -- (--, #) (--, --) (--, #) mol/kg PCB ± 27 (*, --) 301 ± 32 (--, --) 305 ± 20 (*, --) * 323 ± 28 (--, #) 640 ± 48 (--, --) 723 ± 35 (--, --) -- 1 mol/kg PCB ± 22 (*, --) 382 ± 23 (*, --) 331 ± 22 (*, --) * 384 ± 23 (*, #) 707 ± 41 (--, --) 719 ± 20 (--, --) -- 5 mol/kg PCB ± 41 (*, --) 562 ± 68 (*, --) 461 ± 45 (*, --) * 500 ± 40 (*, #) 830 ± 80 (--, --) 821 ± 55 (--, --) * Overall # Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine for statistically significant differences (p < 0.05) between each PCB 126 level and the corn oil treatment (*) and between low or supplemental and the adequate dietary selenium level (#). Significance for each factor based on two-way ANOVA is indicated in the bottom margin (#, Se) and in the right margin (*, PCB).

105 89 Table 3.8. GSH (nmoles GSH/mg liver wet weight), GSSG (nmoles GSSG/mg liver wet weight), and GSSG/GSH ratio under each experiment condition and significance of various comparisons (adjusted using Dunnett s test) GSH (nmoles GSH/mg liver wet weight) GSSG (nmoles GSSG/mg liver wet weight) GSSG/GSH ratio (%) Dietary Selenium Level Dietary Selenium Level Dietary Selenium Level Low Adequate Supplemental Ov. Low Adequate Supplemental Ov. Low Adequate Supplemental Ov. Treatment (0.02 ppm) (0.2 ppm) (2 ppm) (0.02 ppm) (0.2 ppm) (2 ppm) (0.02 ppm) (0.2 ppm) (2 ppm) Corn Oil 8.15 ± ± ± ± ± ± ± ± ± (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- 1 mol/kg 8.57 ± ± ± ± ± ± ± ± ± PCB 126 (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- 5 mol/kg 8.45 ± ± ± ± ± ± ± ± ± PCB 126 (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- Overall # # Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine for statistically significant differences (p < 0.05) between each PCB 126 level and the corn oil treatment (*) and between low or supplemental and the adequate dietary selenium level (#). Significance for each factor based on two-way ANOVA is indicated in the bottom margin (#, Se) and in the right margin (*, PCB).

106 Figure 3.1. Liver selenium ( g/g tissue) under each experiment condition and significance of various comparisons (adjusted using Dunnett s test). Results are expressed as mean ± SEM (n = 4-6 animals/group). One-way ANOVA was used to examine the difference between each PCB 126 level and the corresponding corn oil treatment (a) and low (0.02 ppm) or supplemented (2 ppm) Se level vs. adequate (0.2 ppm) Se level (b). p <

107 Figure 3.2. Se-dependent glutathione peroxidase activity ( mol/min/mg protein) and significance of various comparisons (adjusted using Dunnett s test). Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine for significant difference (p < 0.05) between each PCB 126 level and the corresponding corn oil treatment (a) and low (0.02 ppm) or supplemented (2 ppm) Se vs. adequate (0.2 ppm) dietary selenium. 91

108 Figure 3.3. Thioredoxin reductase activity (U/ml) under each experiment condition and significance of various comparisons (adjusted using Dunnett s test). Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine for significant difference (p < 0.05) between each PCB 126 level and the corresponding corn oil treatment (a) and low (0.02 ppm) or supplemented (2 ppm) Se vs. adequate (0.2 ppm) dietary selenium. 92

109 Figure 3.4. Thioredoxin-1 (top) and Thioredoxin-2 (bottom) oxidation state and significance of various comparisons (adjusted using Dunnett s test). Results are expressed as mean ± SEM. Each group contained 3 animals. One-way ANOVA was used to examine the difference between each PCB 126 level and the corn oil treatment (a) and low/supplemented Se vs. 0.2 ppm dietary selenium (b). p <

110 94 CHAPTER IV THE DISPOSITION AND ROLE OF COPPER IN RODENT LIVER TOXICITY FOLLOWING EXPOSURE TO PCB Abstract In its free form, copper can also participate in Fenton-like reactions that result in the production of reactive hydroxyl radicals. In an effort to assess the effect of dietary copper on the toxicity of 3,3,4,4,5-pentachlorobiphenyl (PCB 126), male Sprague- Dawley rats were fed an AIN-93G diet with one of three dietary copper levels: low (2 ppm), adequate (6 ppm), and high (10 ppm). After three weeks, rats from each dietary group were given a single ip injection of corn oil (control), 1, or 5 µmol/kg body weight PCB 126 in corn oil, followed two weeks later by euthanization. Growth rate was slowed by PCB 126 in a dose-dependent manner, significantly at the highest dose (40-75%). Relative liver weight was increased in a dose-dependent manner (30-65%) by PCB 126, while there was no effect on liver weight by dietary copper. Hepatic cytochrome P450 activity was maximally-induced by 1 mol/kg PCB 126. Increasing dietary copper and increasing dose of PCB 126 both increased hepatic copper levels. Blood copper and serum ceruloplasmin levels were diminished by increasing PCB 126 dose at low dietary copper, while increased by high dietary copper. Hepatic CuZnSOD activity was not significantly diminished by low dietary copper. Liver lipids were increased in a dosedependent manner by PCB 126, while lowered by both low and high dietary copper. These results suggest that copper does not play a significant role in PCB 126-induced 3 Trace elements determination was contributed by Drs. Brian Wels and Don L. Simmons, Iowa State Hygienic Laboratory. Glutathione analysis was contributed by Dr. Michael McCormick, Free Radicals and Radiation Biology, University of Iowa. Ceruloplasmin determination was contributed by Miao Li, Human Toxicology Program, University of Iowa. Histology was contributed by Drs. Wanda M. Haschek, Pathology, University of Illinois and Alicia K. Olivier, Pathology, University of Iowa. Statistical analysis formulas were contributed by Dr. Kai Wang, University of Iowa.

111 95 toxicity. However, we showed that modulating dietary copper levels did not significantly diminish key antioxidant enzymes or increase liver injury. Introduction Copper (Cu), an essential trace element with multiple biological roles, can be found in trace concentrations (3-10 g/g wet weight) in vital organs of the human body, including the liver, brain, heart, and kidneys (Ellingsen et al. 2007). Cu is necessary for cellular respiration and energy production, since the Cu-containing cytochrome c oxidase is part of the electron transport chain that reduces oxygen. Cu also functions as a key component of the antioxidant enzyme copper zinc superoxide dismutase (CuZnSOD), which detoxifies reactive superoxide radicals. However, Cu in excess can also act as a pro-oxidant, where it participates in a Fenton-like reaction that generates reactive hydroxyl radicals (Brewer 2008). 3,3,4,4,5-pentachlorobiphenyl (PCB 126) is the most toxic congener (Yoshizawa et al. 2007) of the polychlorinated biphenyls (PCBs) family of persistent organic pollutants, which were commercially produced as industrial mixtures. PCBs were initially valued for their stability for use in a wide range of chemical, thermal, and electrical conditions (Safe 1994). They were, however, also stable when released into the environment, bioaccumulating and biomagnifying despite a massive decline in production since the 1970s (Hansen 1987). Toxicity of PCBs varies greatly depending on the structure of the 209 congeners. Biological and chemical properties are dependent upon the number and placement of the chlorines (Silberhorn et al. 1990). PCB 126, a congener with no chlorine atoms adjacent to the biphenyl bridge in the 2, 2, 6, or 6 position, can assume a more co-planar conformation similar to that of 2,3,7,8- tetrachlorodibenzo-p-dioxin (TCDD). These congeners, known as dioxin-like PCBs, are potent aryl hydrocarbon receptor (AhR) agonists, inducing cytochrome P-450 (CYP) 1A enzymes (Bandiera et al. 1982; Parkinson et al. 1983a; Silberhorn et al. 1990).

112 96 Changes seen following exposure of rodents to potent AhR agonists, like TCDD and PCB 126, include increased hepatic Cu (Elsenhans et al. 1991; Lai et al. 2010; Wahba et al. 1988), diminished hepatic glutathione (Nishimura et al. 2001; Stohs 1990), diminished hepatic selenium, selenium-dependent glutathione peroxidase activity and mrna, and altered lipid metabolism (Huster et al. 2007; Schramm et al. 1985; Twaroski et al. 2001b). Other than the observation of altered distribution of Cu, little is known about the role Cu plays in PCB 126-induced toxicity, if any. Here we investigate Cu s involvement in PCB 126-induced alterations to hepatic redox status and toxicity. We hypothesized that by reducing the dietary intake of Cu, resulting in the reduction of the amount of Cu available for Fenton-like reactions following PCB 126 exposure, will result in reduced toxicity, as measured by biochemical changes and morphology. To this end, male rats, placed on purified diets containing 2 (low), 6 (adequate), or 10 (high) ppm Cu, were subsequently administered a single dose of 1 (low), or 5 (high) mol/kg of PCB 126. Materials and Methods Chemicals All chemicals were obtained from Sigma-Aldrich Chemical Company (St. Louis, MO) unless otherwise stated. PCB 126 (3,3,4,4,5-pentachlorobiphenyl) was prepared by an improved Suzuki-coupling method of 3,4,5-trichlorobromobenzene with 3,4- dichlorophenyl boronic acid utilizing a palladium-catalyzed cross-coupling reaction (Luthe et al. 2009). The crude product was purified by aluminum oxide column and flash silica gel column chromatography and recrystallized from methanol. The final product purity was determined by GC MS analysis to be > 99.8% and its identity confirmed by 13 C NMR. Caution: PCBs and their metabolites should be handled as hazardous compounds in accordance with NIH guidelines.

113 97 Animals This animal experiment was conducted with approval from the Institutional Animal Care and Use Committee of the University of Iowa. Male Sprague-Dawley rats weighing grams from Harlan Sprague-Dawley (Indianapolis, IN) were housed in individual wire cages in a controlled environment maintained at 22 C with a 12 h lightdark cycle and water ad libitum. Animals were randomly divided into three dietary groups, and were fed ad libitum an AIN-93G based diet (Table 1) containing low (2 ppm), adequate (6 ppm), or high (10 ppm) copper level obtained from Harland Teklad (Madison, WI). Following three weeks of acclimatization, time for the rats to adjust to the dietary copper levels (Roughead et al. 1999), animals were given a single i.p. injection of vehicle (stripped corn oil; Acros Chemical Company, Pittsburgh, PA), or vehicle with a 1 mol/kg (326 g/kg) or 5 mol/kg (1.63 mg/kg) dose of PCB 126. Two weeks following the injection, a time frame sufficient for liver pathology to manifest itself (Lai et al. 2010), rats were euthanized using carbon dioxide asphyxiation followed by cervical dislocation, and livers and other organs were excised, weighed, and further processed as described below. Hepatic subcellular fractions preparation Liver tissues were excised immediately following sacrifice, and homogenized in ice-cold 0.25 M sucrose solution, ph 7.4. The homogenate was centrifuged at 10000g for 20 min. The resulting supernatant was then centrifuged at g for 1 h. The supernatant containing the cytosolic fractions were dispensed and aliquoted. Microsomal pellets were washed twice with cold sucrose solution and resuspended in that solution. Protein concentrations were determined by the methods of Lowry et al. (1951). Measurement of CYP1A1 activity CYP1A1 activity was determined in hepatic microsomal fractions by the methods of Burke and Mayer (1974) with slight modifications, measuring the ethoxyresorufin

114 98 deethylase (EROD) activity and using ethoxyresorufin as the substrate. The resulting fluorescent resorufin product from the monooxygenase reaction was detected using a Perkin-Elmer LS 55 spectrofluorometer at excitation wavelength of 550 nm and emission wavelength of 585 nm. Total Glutathione (GSH) analysis GSH levels were determined in hepatic liver tissue homogenized in 5% salicylic acid based on the methods of Griffith (1980) and Anderson (1985), using 5,5 -dithio-bis- (2-nitrobenzoic acid) (DTNB) as the substrate. The reaction catalyzed by glutathione reductase and NADPH results in the formation of 2-nitro-5-thiobenzoate (TNB) and a yellow color in that can be measured at 412 nm. Absorbance change at 412 nm was followed in a Beckman DU-670 spectrophotometer for 5 min. The rate of yellow color accumulation is proportional to the amount of total glutathione in the sample. Total glutathione levels are expressed as per mg protein. 4-Hydroxynonenal (4-HNE) determination Liver 4-HNE levels were determined using an ELISA kit purchased from Cell Biolabs, Inc. (San Diego, CA). Briefly, homogenized liver tissue samples are incubated with an anti-hne-his antibody, and then followed by incubation with a HRP-conjugated secondary antibody. Incubation with the provided substrate solution results in an absorbance change at 450 nm read in a Molecular Devices Spectra Max well plate reader. The quantity of HNE-His adduct in the tissue homogenates is determined by comparing the recorded absorbance with that of a known HNE-BSA standard curve. Measurement of Superoxide Dismutase (SOD) activities SOD activities were determined by the methods of Spitz and Oberley (1989), measuring the reduction of nitroblue tetrazolium (NBT) colormetrically, using xanthinexanthine oxidase as a source of superoxide anion radical. Absorbance change at 560 nm

115 99 was followed in a Beckman DU-670 spectrophotometer for 5 min. To determine manganese SOD (MnSOD) activity, the assay was performed in the presence of 5mM cyanide inhibiting copper zinc SOD (CuZnSOD) activity. CuZnSOD activity was subsequently determined by subtracting MnSOD activity from total SOD activity. One unit of SOD activity is defined as the amount of protein that yields 50% of maximal inhibition of NBT reduction by superoxide anion radicals. Ceruloplasmin determination Serum ceruloplasmin was determined using an ELISA kit purchased from ALPCO (Salem, NH). Briefly, ceruloplasmin present in the serum binds to anticeruloplasmin antibodies. Anti-ceruloplasmin antibodies labeled with horseradish peroxidase are added to the ceruloplasmin-bound antibodies forming complexes. The addition of a chromogenic substrate, 3,3,5,5 -tetramethylbenzidine, causes an absorbance change at 450 nm, which is used as a measure of the concentration of ceruloplasmin in the test sample. The quantity of bound horseradish peroxidase correlates directly with the concentration of ceruloplasmin in the sample tested. Trace elements determination Metal concentrations in different rat tissues were quantitatively determined with an elemental mass spectrometer: the Inductively Coupled Plasma Mass Spectrometry (ICP-MS). ICP-MS is selected due to its low detection limits and multi-element capacity (Keen et al. 2010). Liver tissue was pretreated with HNO 3 acid digestion prior to ICP- MS measurement, and metal concentrations in the treated tissues were determined in an Agilent 7500ce ICP-MS equipped with a CETAC AS520 auto sampler. Histology Liver tissues were fixed in formalin and glutaraldehyde for light microscopic and transmission electron microscopic examinations, respectively. Formalin-fixed tissues

116 100 were routinely processed, embedded in paraffin, sectioned at 4 m and stained with hematoxylin and eosin. Glutaraldehyde-fixed tissues were stained with osmium tetroxide for lipid evaluation, embedded in epoxy resin, sectioned at nm and stained with uranyl acetate and lead citrate. Frozen sections of selected liver samples were sectioned at 8 m and stained with Oil-Red-O for lipid evaluation. Oil-Red-O liver sections were imaged at 400X magnification (DP72, Olympus) and analyzed using ImageJ software. The images were converted to an RGB stack and staining was thresholded in the blue channel. The percent staining was calculated by dividing the stained area by the total parenchymal area. Two images were analyzed per tissue section. Statistics The effects of PCB 126 and dietary Cu on various responses were studied using ANOVA analysis via procedure PROC GLM in statistical analysis package SAS (version 9.2). Dunnett's test was used to compare PCB 126 with the corn oil vehicle control and dietary Cu. This comparison was conducted separately to PCB 126 treatment and dietary Cu levels (one-way ANOVA) and also jointly (two-way ANOVA) (Table 4.2). In twoway ANOVA, the interaction term was removed if it is not significant at level The effect of dietary Cu was controlled when applying Dunnett's test to PCB 126 by using lsmeans statement in PROC GLM. The same is done when applying Dunnett's test to dietary Cu levels. Results Effects on growth, feed consumption, and organ weights Growth rates of rats were slowed by treatment with PCB 126, significantly at the high (5 mol/kg) dose (Figure 4.1a). Amongst the high dose PCB 126 groups, the rats receiving low (2 ppm) dietary Cu had the highest growth rate. Both doses of PCB 126 significantly reduced feed consumption, with no effect by dietary Cu (Figure 4.1b).

117 101 Relative liver weight was significantly increased by both dietary Cu and both doses of PCB 126 (Figure 2a). Relative kidney weight was not significantly affected by dietary Cu levels or PCB 126 (Figure 2b). Effects on EROD activity EROD activity was significantly induced by treatment with PCB 126. At a high (5 mol/kg) dose of PCB 126, EROD activity was also significantly induced, but at a lower magnitude compared to the low (1 mol/kg) doses (Figure 3). EROD activity was not affected overall by dietary Cu levels, however, the rats receiving high (10 ppm) dietary Cu treated with low dose PCB 126 had a significantly higher induction of EROD activity. Effects on liver, kidney, and blood Cu Liver Cu was significantly affected by both dietary Cu and PCB 126 with an interaction effect. Low (2 ppm) dietary Cu significantly diminished hepatic Cu levels, with a dose-dependent increase in hepatic Cu in the rats treated with PCB 126, significantly at the high (5 mol/kg) dose. In contrast, high dietary Cu did not significantly increase hepatic Cu levels, while significant dose-dependent increases in hepatic Cu following PCB 126 treatment was observed consistent with the rats receiving adequate (6 ppm) dietary Cu. Kidney Cu was also significantly affected by both dietary Cu and PCB 126. Low and high dietary Cu significantly increased and decreased kidney Cu levels, respectively. High (5 mol/kg) dose PCB 126 significantly increased kidney Cu levels compared to corn oil-treated controls. Similarly, blood Cu was also significantly affected by both dietary Cu and PCB 126 with an interaction effect. Interestingly, blood Cu was dose-dependently diminished by PCB 126 in rats receiving low dietary Cu. See Table 4.4a for organ Cu data.

118 102 Effects on liver, kidney, and blood iron Liver iron (Fe) was significantly diminished by PCB 126 at all doses, but only in the adequate (6 ppm) and high (10 ppm) dietary Cu groups. Blood Fe was significantly affected by dietary Cu, but not PCB 126 overall. No significant overall effects were observed in kidney Fe levels. See Table 4.4b for organ Fe data. Effects on liver, kidney, and blood selenium Liver selenium (Se) was significantly affected overall by both dietary Cu and PCB 126 with an interaction effect. PCB 126 caused a decrease in a dose-dependent manner. Kidney Se was also significantly affected overall by both dietary Cu and PCB 126, but without interaction. In contrast to the liver, kidney Se levels were increased by PCB 126, significantly overall at the high (5 mol/kg) dose. Blood Se was only significantly affected overall by dietary Cu. Rats receiving low (2 ppm) dietary Cu had significantly higher blood Se levels. See Table 4.4c for organ Se data. Effects on liver, kidney, and blood zinc Liver zinc (Zn) was significantly affected overall by both dietary Cu and PCB 126 without interaction. PCB 126 caused significant decreases in a dose-dependent manner. High dietary Cu increased Zn significantly at the high (5 mol/kg) dose PCB 126. Kidney Zn was significantly affected overall by dietary Cu but not PCB 126. In contrast, blood Zn was significantly decreased overall by PCB 126 but not affected by dietary Cu. See Table 4.4d for organ Zn data. Effects on liver, kidney, and blood manganese Liver manganese (Mn) was significantly affected overall by both dietary Cu and PCB 126 with no interaction effect. PCB 126 caused significant decreases in a dosedependent manner. Kidney Mn was also significantly affected overall by dietary Cu and PCB 126, but not in a dose-dependent manner. Blood Mn was not affected by PCB 126,

119 103 but was significantly affected overall by dietary Cu. Rats receiving low (2 ppm) dietary Cu had significantly higher blood Mn levels than rats receiving adequate (6 ppm) and high (10 ppm) dietary Cu. See Table 4.4e for organ Mn data. Effects on liver, kidney, and blood molybdenum Liver molybdenum (Mo) was not significantly affected overall by dietary Cu or PCB 126. In contrast, kidney Mo was significantly affected overall by both dietary Cu and PCB 126, with no interaction effect. Both dietary Cu and PCB 126 caused increases in a dose-dependent manner. Blood Mo was significantly affected overall by dietary Cu but not PCB 126. See Table 4.4f for organ weight data. Effects on CuZnSOD, MnSOD, and Total SOD activities No significant effects were observed on CuZnSOD activity by dietary Cu. MnSOD activity was diminished by PCB 126, significantly at the high (5 mol/kg) dose in the high dietary Cu group, however this has no effect on the total SOD activity (Table 4.4). Effects on total hepatic glutathione Total glutathione was diminished in a dose-dependent manner by PCB 126, while rats fed low dietary Cu had the highest total glutathione levels (Figure 4.4). Total glutathione was diminished in a dose-dependent manner by PCB 126 and dietary Cu levels. Effects on Liver 4-HNE adducts Liver 4-HNE in the high (5 mol/kg) dose PCB 126-treated animals was significantly lower in the low (2 ppm) dietary copper group compared to the adequate (6 ppm) dietary copper group. PCB 126 had a significant overall effect only in the low dietary copper groups, but low (1 mol/kg) and high doses PCB 126 had opposite effects.

120 104 In the adequate and high (10 ppm) dietary copper groups, animals treated with PCB 126 had lower levels of 4-HNE, but these effects were not statistically significant. Effects on serum ceruloplasmin Serum ceruloplasmin was significantly diminished by low (2 ppm) dietary Cu (Figure 4.6). PCB 126 caused an increase at both doses, significantly at the high (5 mol/kg) dose in the rats receiving high (10 ppm) dietary Cu. Histology On histologic examination, PCB 126 treated livers had hepatocellular enlargement due to cytoplasmic lipid accumulation with a mild increase in cytoplasmic density. Additionally periportal hepatocytes had cytoplasmic clearing consistent with hydropic change. There was a PCB 126 dose-response in severity of these hepatocyte alterations. Apoptotic hepatocytes were identified in rats treated with high dose PCB 126 in the adequate (6 ppm) and high (10 ppm) dietary copper groups although only 1-2 rats were affected in each group. Evaluation and quantification of Oil-Red-O staining resulted in a dose dependent increase in liver lipid accumulation in PCB 126-treated rats (Figure 4.7). Liver sections were examined by transmission electron microscopy examination to determine if ultrastructural differences could be detected with PCB 126-treatment and dietary copper effect. The ultrastructure changes were consistent with PCB 126 exposure including increased cytoplasmic lipid and sections with increased smooth endoplasmic reticulum (SER). There was no evidence of lipid within the SER. There were no significant differences between dietary copper groups exposed to the same PCB 126 dose. Discussion Copper s (Cu) ability to act as both an essential antioxidant and a prooxidant makes it an intriguing target in toxicity. Cu absorption occurs mainly through the

121 105 stomach and small intestine, from dietary sources such as shellfish, meats, grains, and fruits. In trace amounts, Cu is delivered via serum proteins in the blood to vital organs and incorporated into essential metalloenzymes (van den Berghe and Klomp 2009). Complete cellular respiration requires the Cu-containing enzyme cytochrome c oxidase, the last step of oxygen reduction in the mitochondrial electron transport chain (Horn and Barrientos 2008). Incomplete reduction of oxygen results in the generation of reactive superoxide anions (Rigoulet et al. 2011). Essential defenses against superoxide include the Cu-containing form of superoxide dismutase (CuZnSOD). However, Cu in excess of the required trace amounts can result in toxicity. Because of the small therapeutic range, Cu metabolism and excretion is tightly controlled and regulated (Stern et al. 2007). Cu absorbed in the body is mostly bound to a transporter protein, ceruloplasmin, or sequestered into metallothionein. Excess Cu not bound to metalloproteins are highly reactive. Cu has the ability to accept and donate electrons making it ideal for use in the catalytic reactions of enzymes. In the absence of an enzyme, free Cu can react with hydrogen peroxide and generate highly reactive hydroxyl radicals in a Fenton-like reaction. These reactive oxygen species (ROS) can bind to lipids, proteins, and DNA, and cause covalent modifications (Klaunig et al. 2011). Acute Cu toxicity can result in various pathologies, while chronic Cu toxicity is can cause severe hepatic and neurologic damage. (Butterworth 2010; Uriu-Adams and Keen 2005). By modulating dietary Cu, the role of liver copper in PCB 126 toxicity was examined. Although toxicity as a result of direct disruption of Cu homeostasis has been well studied, little is known about the effects of indirect disruption of Cu homeostasis caused by xenobiotics. Polychlorinated biphenyls (PCBs), a family of persistent environmental pollutants, can cause ROS generation and disrupt copper homeostasis (Hennig et al. 2005; Ohchi et al. 1987). The most potent dioxin-like PCB congener, 3,3 4,4 4,5- pentachlorobiphenyl (PCB 126), is known for its acute effects on the liver (Yoshizawa et al. 2007). Consistent with previous findings, we observed that high (5 mol/kg) dose

122 106 PCB 126 significantly slowed the growth and feed consumption of the rats, while rats receiving low (2 ppm) dietary Cu had higher growth and feed consumption rates that were not significant compared to controls (Figure 4.1). Liver weights were significantly increased by PCB 126, reflecting liver hypertrophy, while dietary Cu did not have any significant effects (Figure 4.2). As expected, CYP1A1 activity, a biomarker of AhR activity, was significantly induced by PCB 126 (Figure 4.3). Interestingly, rats receiving low dietary Cu had higher total liver glutathione levels (Figure 4.4), suggesting these rats have increased synthesis of glutathione and/or lower levels of oxidative stress in the liver. Past studies have shown that dioxin-like compounds, including PCB 126, caused significant increases in hepatic Cu levels (Lai et al. 2010). Consistent with those observations, there was a significant dose-dependent increase in hepatic Cu levels in all rats exposed to PCB 126 (Table 4.3a). This observation may be explained by an induction of metallothionein or disrupted biliary transport (Fletcher et al. 2005). Although low dietary copper significantly lowered hepatic copper levels, there were no effects on hepatic CuZnSOD activity (Table 4.4), nor were these effects on MnSOD or total SOD activities. These results support that the low dietary Cu level used in this study did not reduce liver Cu below a physiologically relevant level (Harris 1992). In contrast to Cu, liver iron was significantly diminished by PCB 126 (Table 4.3b). It is unlikely that iron absorption into the liver was affected by changes in Cu levels, as neither low nor high dietary Cu has any significant effects on liver iron. However, serum ceruloplasmin, a Cu carrier involved iron-metabolism, was significantly diminished by low dietary Cu (Figure 4.6). PCB 126 also significantly diminished blood iron only in the low dietary Cu group. Consistent with previous observations (Lai et al. 2010), liver selenium was significantly diminished by PCB 126, while dietary Cu had no significant effects (Table 4.3c). In contrast, kidney selenium was unexpectedly affected by dietary Cu in a dosedependent manner. Kidney selenium was also significantly increased by high dose PCB

123 These observations would support that PCB 126 caused an increase in selenium excretion, a result of conjugation to electrophilic compounds (Muttenthaler and Alewood 2008). Selenium is incorporated into several antioxidant enzymes, glutathione peroxidase and thioredoxin reductase. The removal of selenium from the liver results in the loss of activity of those enzymes, and increases the risk of oxidative stress (Ganther 1999; Steinbrenner and Sies 2009). Zinc, manganese, molybdenum are elements that can affect with Cu absorption. Liver zinc was significantly diminished by PCB 126, while dietary Cu had a dosedependent effect (Table 4.3d). Blood zinc was also significantly diminished by PCB 126. Liver manganese was also significantly diminished by PCB 126 and affected in a dosedependent manner by dietary Cu (Table 4.3e). These observations suggest that Cu is preferentially absorbed (Aggett 1985). As expected, large cytoplasmic lipid vacuoles were observed in the liver of PCB 126-treated rats (Figure 4.8). We observed an increase in the lipid synthesizing smooth endoplasmic reticulum (SER), no lipid accumulation was observed in either smooth or rough endoplasmic reticulum (Figure 4.8). Oil-Red-O quantification confirmed an increase in lipid accumulation, suggesting an increase in lipid synthesis and/or disruption of lipid metabolism (Figure 4.7). Interestingly, the rats fed low and high dietary Cu had lower liver lipid levels than those fed the control diet. The significance of this observation is unknown. There is evidence to suggest that alterations to hepatic lipid metabolism may be linked to complexed copper (Kennedy et al. 2009), although the mechanism of this change is unknown. It is possibly similar to early stages of Wilson s Disease, a genetic disease that causes accumulation of Cu in the liver (Lalioti et al. 2010). Steatosis observed in PCB 126-treated rats is also observed in the early Wilson s Disease model, as a result of oxidative damage caused by excess Cu and Fenton chemistry. If left untreated, this can result in inflammation, necrosis, and finally cirrhosis (Huster 2010). Steatosis

124 108 may also be a result of Cu accumulation but independent of oxidative stress. AhR agonists have been observed to disrupt the electron transport chain and diminish ATP levels in the liver (Forgacs et al. 2010; Senft et al. 2002b). It is likely that this disrupts the hepatic efflux of excess Cu into the bile, as this process requires ATP hydrolysis (Prohaska 2008). This suggests that increased hepatic Cu is a byproduct of metabolic disruption caused by PCB 126 toxicity. Little is understood as to the role of copper in PCB 126-induced toxicity. Knowing that AhR agonists cause increases in hepatic Cu levels, the goal of this study was to determine mechanisms of PCB 126-induced liver toxicity related to Cu levels. While lowering dietary Cu did result in lower liver Cu and lipid levels following PCB 126 exposure in comparison to rats receiving control diets, dose-dependent effects were not observed in regards to dietary Cu. It is unclear whether increased hepatic Cu directly causes toxicity or is a byproduct of toxicity. Further studies will be needed to fully elucidate the role of Cu following PCB 126 exposure.

125 109 Table 4.1. Composition of AIN-93G modified copper diets Low Adequate Supplemental 2 ppm Cu 6 ppm Cu 10 ppm Cu Constituent g/kg g/kg g/kg Casein, low Cu & Fe L-Cystine Corn Starch Maltodextrin Sucrose Soybean Oil Cellulose Cu Deficient Mineral Mix Cupric Carbonate Vitamin Mix, AIN-93-VX Choline Bitartrate THBQ, antioxidant

126 110 Table 4.2. Two-way ANOVA analysis of the effects of dietary copper, PCB 126, and interaction Dietary Copper PCB 126 Interaction Effect Growth Rate (%) - < Feed Consumption - < Relative Liver Weight (%) < Relative Kidney Weight (%) EROD Activity - < MnSOD Activity CuZnSOD Activity Total SOD Activity GSH (protein) HNE Liver Copper < < Liver Selenium < Liver Iron Liver Zinc < < Liver Manganese < Liver Molybdenum Kidney Copper < Kidney Selenium < < Kidney Iron Kidney Zinc Kidney Manganese Kidney Molybdenum Blood Copper < Blood Selenium Blood Iron Blood Zinc Blood Manganese Blood Molybdenum Serum Ceruloplasmin < Only p-values <0.05 are reported.

127 111 Table 4.3. Liver ( g/g), kidney ( g/g), and blood ( g/l) copper (A), iron (B), selenium (C), zinc (D), manganese (E), and molybdenum (F) A Liver Copper ( g/g tissue wet weight) Kidney Copper ( g/g tissue wet weight) Blood Copper ( g/l) Dietary Copper Level Dietary Copper Level Dietary Copper Level Low Adequate High Overall Low Adequate High Overall Low Adequate High Overall Treatment (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) Corn Oil 3.12 ± ± ± ± ± ± ± ± ± 34 (--, *) (--, --) (--, --) -- (--, *) (--, --) (--, --) -- (--, *) (--, --) (--, --) -- 1 mol/kg 3.51 ± ± ± ± ± ± ± ± ± 41 PCB 126 (--, *) (*, --) (*, *) * (--, *) (--, --) (--, --) -- (--, *) (--, --) (--, --) -- 5 mol/kg 4.33 ± ± ± ± ± ± ± ± ± 56 PCB 126 (*, *) (*, --) (*, *) * (--, *) (--, --) (*, --) * (--, *) (*, --) (*, --) * Overall * -- * * -- * * -- * Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine the difference between each PCB 126 level and the corn oil treatment. The first character in the parentheses is * if it is significant. Similarly, the second character in the parentheses is * if the difference between low or supplemental and the adequate dietary selenium level is significant. Significance for each factor based on two-way ANOVA is indicated in the bottom margin and in the right margin. The level for significance is 0.05.

128 112 Table 4.3. Continued. B Liver Iron ( g/g tissue wet weight) Kidney Iron ( g/g tissue wet weight) Blood Iron (mg/l) Dietary Copper Level Dietary Copper Level Dietary Copper Level Low Adequate High Overall Low Adequate High Overall Low Adequate High Overall Treatment (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) Corn Oil 115 ± ± ± ± ± ± ± ± ± 22 (--, *) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- 1 mol/kg PCB ± 12 (--, --) 107 ± 6 (*, --) 106 ± 5 (*, --) * 71.7 ± 3.9 (--, --) 78.1 ± 4.8 (--, --) 82.5 ± 8.8 (--, --) ± 30 (--, --) 513 ± 35 (--, --) 547 ± 24 (--, --) -- 5 mol/kg PCB ± 10 (--, --) 107 ± 3 (*, --) 100 ± 7 (*, --) * 71.5 ± 5.3 (--, *) 87.7 ± 5.2 (--, --) 78.1 ± 2.5 (--, --) ± 11 (*, *) 522 ± 26 (--, --) 470 ± 32 (--, --) * Overall * Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine the difference between each PCB 126 level and the corn oil treatment. The first character in the parentheses is * if it is significant. Similarly, the second character in the parentheses is * if the difference between low or supplemental and the adequate dietary selenium level is significant. Significance for each factor based on two-way ANOVA is indicated in the bottom margin and in the right margin. The level for significance is 0.05.

129 113 Table 4.3. Continued. C Liver Selenium ( g/g tissue wet weight) Kidney Selenium ( g/g tissue wet weight) Blood Selenium ( g/l) Dietary Copper Level Dietary Copper Level Dietary Copper Level Low Adequate High Overall Low Adequate High Overall Low Adequate High Overall Treatment (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) Corn Oil ± ± ± ± ± ± ± ± ± 23 (--, *) (--, --) (--, *) -- (--, --) (--, --) (--, *) -- (--, --) (--, --) (--, --) -- 1 mol/kg ± ± ± ± ± ± ± ± ± 15 PCB 126 (*, --) (*, --) (*, --) * (--, --) (--, --) (--, --) -- (--, *) (--, --) (--, --) -- 5 mol/kg ± ± ± ± ± ± ± ± ± 10 PCB 126 (*, --) (*, --) (*, *) * (*, *) (*, --) (--, --) * (--, *) (--, --) (--, --) -- Overall * * -- * * Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine the difference between each PCB 126 level and the corn oil treatment. The first character in the parentheses is * if it is significant. Similarly, the second character in the parentheses is * if the difference between low or supplemental and the adequate dietary selenium level is significant. Significance for each factor based on two-way ANOVA is indicated in the bottom margin and in the right margin. The level for significance is 0.05.

130 114 Table 4.3. Continued. D Liver Zinc ( g/g tissue wet weight) Kidney Zinc ( g/g tissue wet weight) Blood Zinc (mg/l) Dietary Copper Level Dietary Copper Level Dietary Copper Level Low Adequate High Overall Low Adequate High Overall Low Adequate High Overall Treatment (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) Corn Oil 27.0 ± ± ± ± ± ± ± ± ± 0.31 (--, *) (--, --) (--, --) -- (--, --) (--, --) (--, *) -- (--, --) (--, --) (--, --) -- 1 mol/kg PCB ± 0.5 (--, --) 25.8 ± 0.4 (*, --) 29.2 ± 1.6 (--, *) * 19.5 ± 0.4 (--, --) 20.8 ± 0.4 (--, --) 21.2 ± 1.2 (--, --) ± 0.17 (--, --) 5.68 ± 0.26 (--, --) 5.95 ± 0.20 (--, --) * 5 mol/kg PCB ± 1.1 (*, --) 25.1 ± 1.1 (*, --) 26.9 ± 1.0 (--, --) * 20.2 ± 0.3 (--, *) 22.6 ± 0.5 (*, --) 21.5 ± 0.5 (--, --) ± 0.28 (*, --) 6.08 ± 0.25 (--, --) 6.07 ± 0.28 (--, --) * Overall * * Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine the difference between each PCB 126 level and the corn oil treatment. The first character in the parentheses is * if it is significant. Similarly, the second character in the parentheses is * if the difference between low or supplemental and the adequate dietary selenium level is significant. Significance for each factor based on two-way ANOVA is indicated in the bottom margin and in the right margin. The level for significance is 0.05.

131 115 Table 4.3. Continued. E Liver Manganese ( g/g tissue wet wt.) Kidney Manganese ( g/g tissue wet wt.) Blood Manganese ( g/l) Dietary Copper Level Dietary Copper Level Dietary Copper Level Low Adequate High Overall Low Adequate High Overall Low Adequate High Overall Treatment (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) Corn Oil 2.15 ± ± ± ± ± ± ± ± ± 0.50 (--, *) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- (--, *) (--, --) (--, --) -- 1 mol/kg 1.67 ± ± ± ± ± ± ± ± ± 0.62 PCB 126 (*, --) (*, --) (*, --) * (--, --) (--, --) (--, --) -- (--, *) (--, --) (--, --) * 5 mol/kg 1.55 ± ± ± ± ± ± ± ± ± 0.49 PCB 126 (*, --) (*, --) (*, --) * (--, --) (--, *) (--, --) -- (--, --) (--, --) (--, --) -- Overall * * Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine the difference between each PCB 126 level and the corn oil treatment. The first character in the parentheses is * if it is significant. Similarly, the second character in the parentheses is * if the difference between low or supplemental and the adequate dietary selenium level is significant. Significance for each factor based on two-way ANOVA is indicated in the bottom margin and in the right margin. The level for significance is 0.05.

132 116 Table 4.3. Continued. F Liver Molybdenum ( g/g tissue wet weight) Kidney Molybdenum ( g/g tissue wet weight) Blood Molybdenum ( g/l) Dietary Copper Level Dietary Copper Level Dietary Copper Level Low Adequate High Ov. Low Adequate High Ov. Low Adequate High Ov. Treatment (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) Corn Oil ± ± ± ± ± ± ± ± ± 0.8 (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- 1 mol/kg ± ± ± ± ± ± ± ± ± 0.5 PCB 126 (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- (--, *) (--, --) (--, --) -- 5 mol/kg ± ± ± ± ± ± ± ± ± 0.9 PCB 126 (*, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) * (--, --) (--, --) (--, --) -- Overall Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine the difference between each PCB 126 level and the corn oil treatment. The first character in the parentheses is * if it is significant. Similarly, the second character in the parentheses is * if the difference between low or supplemental and the adequate dietary selenium level is significant. Significance for each factor based on two-way ANOVA is indicated in the bottom margin and in the right margin. The level for significance is 0.05.

133 117 Table 4.4. CuZnSOD, MnSOD, and Total SOD activities (U/mg protein) CuZnSOD Activity (U/mg protein) MnSOD Activity (U/mg protein) Total SOD Activity (U/mg protein) Dietary Copper Level Dietary Copper Level Dietary Copper Level Low Adequate High Overall Low Adequate High Overall Low Adequate High Overall Treatment (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) (2 ppm) (6 ppm) (10 ppm) Corn Oil 256 ± ± ± ± ± ± ± ± ± 36 (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- (--, --) (--, --) (--, --) -- 1 mol/kg PCB ± 26 (--, --) 270 ± 32 (--, --) 284 ± 61 (--, --) ± 9.9 (--, --) 74.5 ± 7.6 (--, --) 82.6 ± 6.5 (--, --) * 308 ± 29 (--, --) 344 ± 32 (--, --) 367 ± 60 (--, --) -- 5 mol/kg PCB ± 27 (--, --) 296 ± 40 (--, --) 283 ± 33 (--, --) ± 8.3 (--, --) 81.8 ± 8.1 (--, --) 61.7 ± 5.5 (*, --) * 293 ± 26 (--, --) 378 ± 47 (--, --) 344 ± 32 (--, --) -- Overall Results are expressed as mean ± SEM. Each group contained 4-6 animals. One-way ANOVA was used to examine the difference between each PCB 126 level and the corn oil treatment. The first character in the parentheses is * if it is significant. Similarly, the second character in the parentheses is * if the difference between low or supplemental and the adequate dietary selenium level is significant. Significance for each factor based on two-way ANOVA is indicated in the bottom margin and in the right margin. The level for significance is 0.05.

134 118 A Figure 4.1. Growth (A) and feed consumption (B) of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals. (A) Growth is defined as the weight gained relative to initial weight. High (5 mol/kg) dose PCB 126 significantly slowed growth relative to vehicle-treated control indicating acute toxicity. (B) Feed consumption is defined as total feed consumed in grams following injection. PCB 126 diminished feed consumption in a dose-dependent manner. Error bars represent SEM. p < 0.05 as compared to Corn Oil vehicle control.

135 119 B Figure 4.1. Continued.

136 120 A Figure 4.2. Relative liver (A) and kidney (B) weights of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals. (A) Relative liver weight is defined as final liver weight as a percentage of total body weight. PCB 126 significantly increased liver weight in a dose-dependent manner. (B) Relative kidney weight is defined as final liver weight as a percentage of total body weight. Error bars represent SEM. p < 0.05 as compared to Corn Oil vehicle control.

137 121 B Figure 4.2. Continued.

138 Figure 4.3. Liver ethoxyresorufin-o-deethylase (EROD) activity of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals. PCB 126 significant induced EROD activity. Error bars represent SEM. * p < 0.05 as compared to adequate (6 ppm) dietary copper. p < 0.05 as compared to Corn Oil vehicle control. 122

139 123 Figure 4.4. Liver total Glutathione (GSH) of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals. Error bars represent SEM. * p < 0.05 as compared to adequate (6 ppm) dietary copper. p < 0.05 as compared to Corn Oil vehicle control.

140 Figure 4.5. Liver 4-HNE levels of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals. Error bars represent SEM. * p < 0.05 as compared to adequate (6 ppm) dietary copper. p < 0.05 as compared to Corn vehicle control. 124

141 Figure 4.6. Serum ceruloplasmin of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals. Error bars represent SEM. * p < 0.05 as compared to adequate (6 ppm) dietary copper. p < 0.05 as compared to Corn Oil vehicle control. 125

142 126 Figure 4.7. Percentage of lipids in liver tissue of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals. Error bars represent SEM. * p < 0.05 as compared to adequate (6 ppm) dietary copper. p < 0.05 as compared to Corn Oil vehicle control.

143 Figure 4.8. Electron micrographs from hepatocytes of rats fed adequate (6 ppm) copper and treated with (A) corn oil (vehicle) or (B and C) PCB 126 (1 mol/kg and 5 mol/kg). N, nucleus; M, mitochondria; R, rough endoplasmic reticulum; S, smooth endoplasmic reticulum; L, lipid vacoules. 127

144 Figure 4.8. Continued. 128

145 Figure 4.8. Continued. 129

146 130 CHAPTER V N-ACETYCYSTEINE DIMINISHES THE SEVERITY OF PCB 126 INDUCED FATTY LIVER IN MALE RODENTS 4 Abstract Due to difficulty in direct uptake of glutathione, an alternative cysteine source, NAC, has been widely used as a supplement. Strong aryl hydrocarbon receptor agonists, including the most potent PCB congener, PCB 126, cause liver pathology and diminish hepatic glutathione. Male Sprague-Dawley rats were fed a standard AIN-93G diet or a modified diet supplemented with 1.0% NAC. After one week, rats from each dietary group were exposed to 0, 1, or 5 mol/kg body weight PCB 126 by i.p. injection followed two weeks later by euthanization. Growth rate was slowed dose-dependently (20%) by PCB 126, while relative liver weight was increased (42-52%) and thymus weight diminished (40-85%). Total hepatic glutathione was diminished dose-dependently (4-34%) by PCB 126. Histologic examination of liver tissue showed PCB 126-induced hepatocellular steatosis in centrilobular to mid-zonal hepatocytes (1 mol/kg dose) to all zones (5 mol/kg dose). Hepatocellular lipid was diminished in NAC-supplemented rats. These results demonstrate that NAC supplementation in PCB 126-treated rats show a protective effect against accumulation of hepatic lipids. Introduction The beneficial effects of N-acetylcysteine (NAC) as an antioxidant and exogenous source of cysteine are well documented (Atkuri et al. 2007; Dodd et al. 2008). NAC is a naturally occurring amino acid precursor of the tripeptide glutathione, the major thiol 4 Histology and lipid quantification were contributed by Dr. Alicia K. Olivier, Pathology, University of Iowa. Statistical analysis formulas were provided by Dr. Kai Wang, University of Iowa.

147 131 group donor in the body known for its antioxidant activities (Meister 1994). In addition to its role as a free radical scavenger, glutathione also serves as the substrate for conjugation by glutathione transferases, an important phase II detoxification reaction. Depletion of glutathione exposes cellular protein thiols to oxidation and covalent modification (Comporti 1987). Dietary supplementation with glutathione is hampered due to its the poor intestinal absorption. Therefore, NAC has been widely used as a supplement in lieu of direct glutathione administration (Witschi et al. 1992). Polychlorinated biphenyls (PCBs), originally manufactured commercially for industrial applications, were known for their insulating and flame resistant properties (Safe 1994). Industrial PCBs mixtures were widely used in manufacturing beginning in the 1930s, which continued until the 1970s at which time their manufacture as commercial products was discontinued due to increasing environmental and health concerns. PCBs lipophilicity and persistence in the environment resulted in their bioaccumulation and biomagnification, effects of which are still felt to the present day (Hansen 1987). The bioaccumulative and toxic effects of PCBs can vary greatly depending on the chlorination patterns of the 209 congeners. One approach to evaluating the toxicity of complex PCB mixtures was to identify the spectra of adverse effects and biochemical changes elicited by individual PCB congeners (Ludewig et al. 2007; Silberhorn et al. 1990). Research with rodents demonstrated that, like 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), 3,3,4,4,5-pentachlorobiphenyl (PCB 126) binds with high avidity to the aryl hydrocarbon receptor (Bandiera et al. 1982), induces cytochromes P-450 (CYP), namely CYP1A1/2 (Parkinson et al. 1983a), and elicits these effects at much lower doses than other PCB congeners. Aside from the plethora of changes in gene expression, PCB 126 causes wasting syndrome, severe thymic involution with the loss of cortical lymphocytes, and liver enlargement with fatty liver changes (Lai et al. 2010; Parkinson et al. 1983a).

148 132 The overexpression of hepatic CYP1A, especially in the presence of reducing equivalents and absence of an oxidizable substrate, is thought to be related to the toxic sequelae seen, in that during the catalytic cycle, CYP1A releases reactive oxygen species (ROS) as oxygen is only partially reduced (Schlezinger et al. 2006). Another line of reasoning posits that the mitochondria are the source of ROS. This argument is buttressed by the observation that PCBs increased steady-state levels of superoxide that were found by confocal microscopy to be primarily located in the mitochondria (Zhu et al. 2009). Several attempts were undertaken to ameliorate the adverse effects of halogenated biphenyls with dietary interventions, for example with fat substitutes like olestra (Jandacek et al. 2010), minerals like selenium (Stemm et al. 2008), various antioxidants (Robertson et al. 1983; Tharappel et al. 2008), or phytochemicals (Glauert et al. 2008). Generally these manipulations resulted in only marginal success. A universal finding is that PCBs, especially PCB 126, diminish liver total glutathione levels (Lai et al. 2010). Regardless of whether the origin of ROS is mitochondrial or microsomal, the effects on glutathione are consistent, and one may pose that an exogenous source of cysteine in the form of NAC may well protect against these adverse effects, a level of protection already seen in vitro (Zhu et al. 2009). Therefore, it is hypothesized that dietary NAC supplementation will reduce the toxicity caused by PCB 126 in vivo, by increasing the bioavailability of thiols and reducing oxidative stress. Methods and Materials Chemicals All chemicals were obtained from Sigma-Aldrich Chemical Company (St. Louis, MO) unless otherwise stated. PCB 126 (3,3,4,4,5-pentachlorobiphenyl) was prepared by an improved Suzuki-coupling method of 3,4,5-trichlorobromobenzene with 3,4- dichlorophenyl boronic acid utilizing a palladium-catalyzed cross-coupling reaction (Luthe et al. 2009). The crude product was purified by aluminum oxide column and flash

149 133 silica gel column chromatography and recrystallized from methanol. The final product purity was determined by GC MS analysis to be > 99.8% and its identity confirmed by 13 C NMR. Caution: PCBs and their metabolites should be handled as hazardous compounds in accordance with NIH guidelines. Animals This animal experiment was conducted with approval from the Institutional Animal Care and Use Committee of the University of Iowa. Male Sprague-Dawley rats weighing grams from Harlan Sprague-Dawley (Indianapolis, IN) were housed in individual wire cages in a controlled environment maintained at 22 C with a 12 h lightdark cycle and water ad libitum. Animals were randomly divided into two dietary groups, and were fed ad libitum an AIN-93G diet or an AIN-93G based diet supplemented with 1.0% NAC (Table 5.1) purchased from Harland Teklad (Madison, WI). After one week, animals were given a single i.p. injection of vehicle (stripped corn oil; 5 mls/kg body weight; Acros Chemical Company, Pittsburgh, PA), or vehicle with 1 mol/kg body weight (326 g/kg body weight) or 5 mol/kg body weight (1.63 mg/kg body weight) of PCB 126. These doses were chosen based on a previous study in which a 1 µmol/kg dose of PCB 126 was shown to elicit mild fatty liver (Lai et al. 2010). Animals were weighed and feed consumption determined two times per week. Two weeks following the PCB treatment rats were euthanized using carbon dioxide asphyxiation followed by cervical dislocation. The two week time period was shown to be sufficient for development of pathology in PCB 126-treated rats (Lai et al. 2010). Livers and other organs were excised, weighed, and further processed as described below. Hepatic subcellular fractions preparation Liver tissues were excised immediately following euthanization, and homogenized in ice-cold 0.25 M sucrose solution, adjusted to ph 7.4. The homogenates were centrifuged at 10,000g for 20 min. The resulting supernatants were then centrifuged

150 134 at 100,000g for 1 h. These supernatants, which contain the cytosolic fractions, were dispensed and aliquoted. The microsomal pellets were washed twice with cold sucrose solution and resuspended in that solution. Protein concentrations were determined by the method of Lowry et al. (1951). Glutathione analysis Total glutathione levels in liver supernatant were determined in hepatic liver tissue by the methods of Griffith (1980) and Anderson (1985), using glutathtione reductase as the substrate. Absorbance change at 412 nm over 5 minutes was measured followed in a Beckman DU-670 spectrophotometer for 5 min. The rate of yellow color accumulation is the result of thionitrobenzoate formation from 5,5'-dithio-bis-(2- nitrobenzoic acid), proportional to the amount of total glutathione in the sample. Oxidized glutathione (GSSG) wasis measured independently by incubating the supernatants tissue in the presence of 2-vinylpyridine, which conjugates reduced glutathione (GSH), followed by the determination of the remaining glutathione equivalents as described above. Glutathione levels are expressed as per mg protein. Glutathione-S-transferase (GST) activity GST activity was determined in hepatic cytosolic fractions by the method of Habig et al. (1974), using 1-chloro-2,4-dinitrobenzene (CDNB) as the substrate. The absorbance change at 340 nm caused by the conjugation of CDNB to reduced glutathione was followed in a Beckman DU-650 for 5 min. Histology and Special Stains Liver sections were fixed in 10% neutral buffered formalin, processed routinely, embedded in paraffin, and stained with hematoxylin and eosin (H&E). Additionally sections were stained with Rhodamine for copper and periodic acid-schiff (PAS) for glycogen (Sheehan and Hrapchak 1987). Sections were immunostained for

151 135 myeloperoxidase (MPO) with a rabbit polyclonal antibody (DAKO A0398) to detect neutrophils. Briefly, liver sections were cut at 4 m and antigen unmasking was performed in citrate buffer (ph 6.0) for 3 x 4 min in the microwave (1000 watts). Endogenous peroxidase activity was quenched with 3% hydrogen peroxide and nonspecific background staining was blocked using background buster reagent (Innovex Biosciences, Richmond, CA). Slides were incubated with the primary antibody (1:1000) for 30 min at room temperature. The slides were then washed with buffer followed by application of DAKO rabbit Envision HRP System reagent for 30 min, washed again and then developed with DAKO DAB Plus for 5 min. Slides were counterstained with Surgipath hematoxylin, dehydrated and coverslipped. Lipid Staining and Quantification Formalin fixed liver sections were stained for lipid using osmium tetroxide (Luna 1992). Samples were placed in a potassium dichromate (5%)/Osmium Tetroxide (2%) solution in water overnight. Samples were washed for 2 hours in running tap water and then processed normally and embedded in paraffin. Sections were cut at 4 m and baked in a 60 C oven overnight. Slides were cooled, deparaffinized and counterstained with nuclear fast red for 5 minutes. Slides were then dehydrated and coverslipped. Osmium stained slides were examined with a high resolution microscope (BX51, Olympus), digital images collected at 100X (DP72, Olympus) and analyzed using Image J software (Image J, NIH). Images were converted to an RGB stack and thresholded in the red channel. The percentage of cellular staining was calculated by dividing the stained area by the total parenchymal area. Statistics The effect of PCB 126 treatment and NAC on various responses was studied using ANOVA analysis via procedure PROC GLM in statistical analysis package SAS (version 9.2). Dunnett's test was used to compare PCB 126 treatment with the corn oil

152 136 control and NAC. This comparison was conducted separately for PCB 126 treatment and NAC level (one-way ANOVA) and also jointly (two-way ANOVA) (Table 5.2). In twoway ANOVA, the interaction term was removed if it was not significant at level The effect of NAC was controlled when applying Dunnett's test to PCB 126 exposure by using lsmeans statement in PROC GLM. The same was done when applying Dunnett's test to NAC level. Results Effects on growth, feed consumption, and organ weights Growth was significantly slowed by PCB 126 at the high (5 mol/kg) dose. NAC supplementation had no overall significant effect. However, the growth rate in the low (1 mol/kg) dose-treated rats supplemented with NAC was significantly lower (Figure 5.1a). Feed consumption was decreased by both PCB 126 doses, but in the low PCB 126- dose this effect was statistically significant only in the NAC-supplemented group (Figure 5.1b). NAC alone did not influence feed consumption (Figure 5.1b). Relative liver and lung weights were significantly increased in a dose-dependent manner by PCB 126, but not by NAC (Figures 5.2a and 5.2b). Relative thymus weight was significantly decreased in a dose-dependent manner by PCB 126, and there was a significant overall decrease in relative thymus weight caused by NAC supplementation (Figure 5.2c and Table 5.2). Effects on EROD activity PCB 126 is a potent inducer of CYP1A1. EROD activity was significantly increased by PCB 126 with the highest induction observed at the lower PCB dose (1 µmol/kg). NAC neither increased nor diminished the CYP 1A1 activity (Figure 5.3).

153 137 Effects on total glutathione and oxidized glutathione (GSSG) Hepatic total glutathione was significantly diminished in a dose-dependent manner by PCB 126, and in addition NAC supplementation also significantly diminishing total glutathione overall, although this effect was not significant at the level of individual PCB-treatment groups (Figure 5.4a and Table 5.2). The effects on GSSG were similar but more pronounced for NAC treatment, with a significant dose-dependent decrease in the PCB 126 groups and additional significant decreases with NAC supplementation (Figure 5.4b and Table 5.2). The GSSG-to-GSH ratio, an indicator of oxidative stress, was significantly lower in NAC supplement rats (Figure 5.4c and Table 5.2). No effect was seen for PCB 126-treatment. Effects on hepatic glutathione transferase (GST) activity Hepatic GST activity was significantly increased in a dose-dependent manner by PCB 126 (Figure 5.5). NAC supplementation had no significant overall effect (Table 5.2). However, the NAC-supplemented rats treated with corn oil had significantly lower GST activity compared to those receiving the control diet. In contrast, in the PCB 126- treated rats, NAC receiving animals had higher hepatic GST activities, although this increase was not significant compared to those receiving the control diet (Figure 5.5). Histology Livers from rats treated with PCB 126 had hepatocellular vacuolation that varied in distribution and severity. All rats treated with PCB 126 had well defined cytoplasmic vacuoles of varying sizes confirmed to be lipid with osmium tetroxide staining (Figures 5.6 and 5.7). In the low (1 mol/kg) dose group, lipid was predominantly found within centrolobular to midzonal hepatocytes while in the high (5 mol/kg) dose group lipid was present in all zones (Figures 5.6 and 5.7). Occasionally periportal hepatocytes in the high dose PCB treated livers were enlarged and characterized by cytoplasmic clearing with

154 138 peripheralization of nuclei (hydropic degeneration). Necrosis, inflammation and copper accumulation were not significant features of any group. Discussion In addition to N-acetylcysteine s (NAC) well-known role as the antidote to acetaminophen toxicity, there are a wide range of scientific and therapeutic applications for NAC. For example, NAC is widely used to treat chronic obstructive pulmonary disorder, pulmonary fibrosis, and contrast-induced nephropathy (Millea 2009). Although NAC is only a precursor of the main cellular antioxidant glutathione (GSH), its relative ease of uptake as compared to GSH has supported its wide use a supplement. NAC provides a source of cysteine, an essential amino acid, contributing critical thiols for the cellular antioxidant. With reactive oxygen species (ROS) linked to carcinogenesis, increasing research emphasis has been placed on NAC in chemoprevention, with promising results (Balansky et al. 2010; Hanczko et al. 2009). Pretreatment with NAC has also been shown to reduce radiation-induced oxidative stress and injury (Mansour et al. 2008). Thus, this study set out to determine the efficacy of NAC in reducing the toxicity of a specific polychlorinated biphenyl (PCB) congener of a family of persistent organic pollutants, labeled as probable carcinogens by the EPA (IRIS 2006). PCBs were produced for industrial purposes, but have persisted in the environment despite declines in their usage since the 1970s (Safe 1994). Research into the mechanisms of PCB toxicity and carcinogenesis has proven to be difficult due to the varied effects attributed to the structural and chemical differences of the 209 congeners. Studies have focused on the non-ortho substituted congeners, which can assume a more co-planar conformation similar to dioxin (TCDD). Although these dioxin-like congeners, PCBs 77, 126, and 169, are known for their binding to the aryl hydrocarbon receptor (AhR), the exact mechanisms of toxicity of these PCBs are currently under investigation (Bock and Kohle 2009; Janosek et al. 2006). Various PCBs, including the dioxin-like

155 139 congeners, have been linked to the generation of ROS (Hennig et al. 2002; Schlezinger and Stegeman 2001). Dioxin-like PCBs may generate ROS through their efficacious induction of CYP1A1/2, and their interference with the catalytic cycle of these enzymes (De et al. 2010). Because cellular injury from oxidative stress can participate in all stages of carcinogenesis (Klaunig et al. 2011), glutathione homeostasis becomes critical in preventing oxidative stress. There was a significant induction of EROD (CYP1A1) activity following PCB 126 exposure, confirming its potent binding to the AhR (Figure 5.3). Liver hypertrophy observed in rats exposed to PCB 126 (Figure 5.2a) is consistent with the proliferation of endoplasmic reticulum caused by CYP1A1 induction and steatosis. Growth rate was significantly slowed by PCB exposure (Figure 5.1a), concomitant with significantly reduced feed consumption (Figure 5.1b), indicating acute toxicity. Consistent with previous studies with PCB 126 (Lai et al. 2010), total liver glutathione levels were significantly diminished in a dose-dependent manner by PCB 126 (Figure 5.4a). However, this was also accompanied by diminished levels of the oxidized glutathione (GSSG) in a dose-dependent manner (Figure 5.4b). Interestingly, NAC did not increase the levels of GSH or GSSG. Contrary to expectations, rats supplemented with NAC had overall lower levels of total hepatic glutathione. One previous experiment had a similar observation in which no change in total glutathione level in the liver and a decrease in the bone marrow of mice was observed after administration of NAC (McLellan et al. 1995). In this study, animals fed NAC also had significantly lower levels of GSSG, indicating an increase in mixed protein disulfides. The lower GSSG-to-GSH ratio in the NAC supplemented rats suggests that NAC has a protective, antioxidant effect (Figure 5.4c). However, this was independent from the PCB 126-treatment, since PCB 126 did not affect the GSSG/GSH ratio. These unexpected results may be due to an increase in mixed protein disulfides in the NAC supplemented rats. PCB 126 increased GST activity, while NAC caused a slight but not significant induction of GST activity in PCB 126-treated

156 140 animals (Figure 5.5). Induction of GST activity by NAC has been previously observed after a single or repeated injection of rats with NAC (Arfsten et al. 2007). Histologically, PCB 126-induced liver toxicity presents as steatosis. Steatosis is a feature of non-alcoholic fatty liver disease (NAFLD) in human patients. NAFLD includes a spectrum of changes from steatosis to non-alcoholic steatohepatitis (NASH). NASH is characterized by steatosis with inflammation, fibrosis, and necrosis (Begriche et al. 2006). In animal models of NASH, NAC has been used effectively to attenuate oxidative stress and reduce liver injury in NASH (Thong-Ngam et al. 2007). More promisingly, NAC along with metformin has shown promise as treatment for NASH in humans by reducing hepatocellular lipid (de Oliveira et al. 2008). The most promising of this study s observations was the reduction of lipid deposition in the livers of rats supplemented with NAC (Figures 5.6 and 5.7), despite the lack of change in hepatic glutathione levels. Based on these results, it is likely that NAC is acting as a protective agent independent of liver glutathione levels. PCBs are known for causing changes to gene expression, some of which have been linked to disruptions in lipid metabolism (Arzuaga et al. 2009). Past studies have shown that activation of the AhR increases expression of CD36, a scavenger receptor involved in fatty acid uptake, and tumor necrosis factor alpha (TNF ), a mediator of inflammation (Lee et al. 2010; Vondracek et al. 2011). Both genes have been implicated in liver steatosis (Greco et al. 2008; Sundaresan et al. 2010). These genes are possibly targets of NAC, which has been shown to reduce both liver CD36 and TNF expression in animals models of NASH (Baumgardner et al. 2008; Ronis et al. 2011). More intriguing is NAC s ability to improve mitochondrial metabolic energy production, which may improve lipid metabolism and reduce steatosis (Zwingmann and Bilodeau 2006). The authors suggest that NAC has limited capacity in de novo glutathione synthesis in non-acetaminophen liver injuries. Rather, NAC favors the formation of hypotaurine, an intermediate of taurine synthesis and potent radical scavenger that may have protective properties of its own (Acharya and Lau-Cam 2010;

157 141 Messina and Dawson 2000). In addition, NAC also increased flux through the pyruvate dehydrogenase pathway, increasing mitochondrial ATP production (Zwingmann and Bilodeau 2006). Taken together, these observations suggest the detoxification by NAC is likely due to modulation of gene expression and/or direct interaction with mitochondrial metabolic pathways. This was the first known in vivo study using NAC as a therapeutic agent against dioxin-like PCB toxicity. While steatosis was still a pathological feature in PCB 126- induced toxicity, NAC was found to be very promising in its ability to reduce steatosis. Steatosis is a prominent feature of PCB toxicity (Robertson et al. 1991). Exposure to persistent compounds that continuously produce oxidative stress and release of proinflammatory cytokines may provide the second hit that may transform simple steatosis into NASH and even hepatocellular carcinoma (Cave et al. 2007). These results demonstrate that while NAC is a promising therapeutic agent in mitigating PCB 126- induced alterations to lipid metabolism, further studies will be needed to determine its exact mechanisms and efficacy in protection against PCB toxicity.

158 142 Table 5.1. Composition of AIN-93G and modified NAC supplemented diets AIN93-G AIN-93G w/ 1% NAC Constituent g/kg g/kg Casein, low Cu & Fe L-Cystine Corn Starch Maltodextrin Sucrose Soybean Oil Cellulose Mineral Mix, AIN-93G-MX Vitamin Mix, AIN-93-VX Choline Bitartrate THBQ, antioxidant N-acetylcysteine

159 143 Table 5.2. Two-way ANOVA analysis of the effects of NAC and PCB 126 NAC PCB 126 Growth Rate (%) - <.0001 Feed Consumption - <.0001 Relative Liver Weight (%) - <.0001 Relative Lung Weight (%) - <.0001 Relative Thymus Weight (%) <.0001 EROD Activity - <.0001 Total GSH <.0001 GSSG <.0001 <.0001 GSSG/GSH Ratio < GST Activity - <.0001 Liver Lipid % <.0001 Only p-values <0.05 are reported.

160 144 A Figure 5.1. Growth and feed consumption of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals. Growth is defined as the weight gained relative to initial weight. High (5 mol/kg) dose PCB 126 significantly slowed growth relative to vehicle-treated control indicating acute toxicity. Animals fed NAC treated with low (1 mol/kg) PCB 126 had significantly slowed growth. Feed consumption is defined as total feed consumed in grams following injection. Error bars represent SEM. * p < 0.05 as compared to Corn Oil vehicle control, p < 0.05 as compared to AIN-93G control diet.

161 145 B Figure 5.1. Continued.

162 146 A Figure 5.2. Relative liver (A), lung (B), and thymus (C) weights of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals. (A) Relative liver weight is defined as final liver weight as a percentage of total body weight. PCB 126 significantly increased liver weight in a dose-dependent manner. (B) Relative lung weight is defined as final lung weight as a percentage of total body weight. PCB 126 significantly increased lung weight in a dose-dependent manner. (C) Relative thymus weight is defined as final thymus weight as a percentage of total body weight. PCB 126 caused significant dose-dependent thymic involution. Error bars represent SEM. * p < 0.05 as compared to Corn Oil vehicle control.

163 147 B Figure 5.2. Continued.

164 148 C Figure 5.2. Continued.

165 Figure 5.3. Liver ethoxyresorufin-o-deethylase (EROD) activity of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals. PCB 126 significant induced EROD activity. Error bars represent SEM. * p < as compared to Corn Oil vehicle control. ** p < as compared to Corn Oil vehicle control. 149

166 150 A Figure 5.4. Liver total glutathione (A), oxidized glutathione (GSSG) (B), ratio of liver oxidized glutathione (GSSG) to liver total glutathione (C) of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals. (A) PCB 126 decreased total GSH in a dose-dependent manner, significantly at the high (5 mol/kg) dose. (B) PCB 126 decreased GSSG in a dose-dependent manner, significantly at the high (5 mol/kg) dose. GSSG was significantly in animals fed NAC-supplemented diets. (C) Liver GSSG to GSH ratio was diminished significantly in animals fed NACsupplemented diets and treated with PCB 126. Error bars represent SEM. * p < 0.05 as compared to Corn Oil vehicle control. p < 0.05 as compared to AIN-93G diet.

167 151 B Figure 5.4. Continued.

168 152 C Figure 5.4. Continued.

169 Figure 5.5. Liver glutathione-s-transferase (GST) activity of vehicle- (Corn Oil) and PCB 126- (1 mol/kg and 5 mol/kg) treated animals. GST activity was significantly induced in a dose-dependent manner by PCB 126. * p < 0.05 as compared to Corn Oil vehicle control. p < 0.05 as compared to AIN-93G diet. 153

170 Figure 5.6. Osmium tetroxide staining for lipid (stains lipid black) of liver from control and PCB treated rats fed a control (AIN-93G) and NAC-supplemented diet. There is little hepatocellular lipid in control treated rats fed the control diet (A) and the NAC-supplemented diet (B). Livers from rats treated with 1 mol/kg PCB fed the control (C) and NAC-supplemented diet (D) had increased hepatocellular lipid in centrolobular to midzonal hepatocytes. Livers from rats treated with 5 mol/kg PCB fed the control (E) and NAC-supplemented diet (F) had lipid in all zones. The percent parenchymal lipid staining was quantified and showed that PCB 126 treatment increased liver lipid in a dose-dependent manner (G). Rats treated with 5 mol/kg PCB and fed the NACsupplemented diet had significantly lower percent liver lipid than rats fed the control diet. Error bars represent SEM. * p < 0.05 as compared to Corn Oil vehicle control. p < 0.05 as compared to AIN-93G diet. 200X 154

171 155 G Figure 5.6. Continued.

172 Figure 5.7. Histological examination of the liver from control and PCB treated rats fed the control (AIN-93G) and NAC-supplemented diet. Representative liver sections taken from control (A, B), 1 µmol/kg PCB treated (C, D) and 5 mol/kg PCB treated rats. The severity of hepatocellular vacuolation is dose dependent. All livers had PAS positive material consistent with glycogen. Left panels: H&E, Right panels: Periodic Acid Schiff (PAS). 200X 156